Every machine in every lab falls into one of three categories. The category tells you who has to be standing there before you switch it on.
Three categories, one question
Three categories, one question
Every piece of equipment in the lab is Category 1, 2 or 3. The
category answers one question: who has to be present before you switch it on?
Permission from an instructor is required in every case. What changes between categories is
when that permission can be given, and who has to be in the room while the machine is running.
Category 1
Permission may be given in advance. Work alone once you have it.
Category 2
Permission at the time of use. A peer must be present throughout.
Category 3
Permission at the time of use. A peer and an instructor must be present throughout.
Power down safely — turn off and disconnect the equipment from its power source.
Report issues immediately — notify the instructor of any broken or malfunctioning equipment.
Return and organize — put all equipment and tools back in their designated locations.
Restore to shutdown state — return the equipment to its original, safe configuration.
Clean thoroughly — remove all dust, chips, liquids and debris from the equipment and the work area.
The habit behind the last two is 5S — the same discipline,
named and made routine.
Safety/Universal Rules
Personal Protective Equipment
What you wear, and when. Eye protection is mandatory in the lab spaces at all times; the rest depends on the task.
Personal Protective Equipment
Personal Protective Equipment
Five topics. One of them — eye protection — applies the moment you walk
in. The rest are set by what you are doing.
Eyes
Always, in the lab spaces. No exceptions.
Hearing
Prolonged exposure to loud equipment.
Feet
Closed-toe shoes, always.
Respiratory
Any grinding or sanding; hazardous dust or fumes.
Clothing & hair
Always, near anything that rotates.
Eye protection; Watch out
Eye protection
Safety glasses are MANDATORY in the lab spaces at all times.
If you wear prescription glasses
Side shields that attach to your existing frames.
Oversized safety glasses that fit over corrective glasses.
Prescription safety glasses are also acceptable.
Lens requirements
Safety glasses must be clear or yellow tinted only.
Dark or shaded safety glasses are not acceptable for indoor lab use.
Must meet ANSI Z87.1 safety standards.
Hearing protection; Watch out
Hearing protection
Industry standard: shop tools nearly always require hearing protection in
professional settings. In a teaching lab the exposure is typically shorter than in production
work, but the protection still matters.
When to use it
Required for prolonged exposure to loud equipment — grinders, sanders, planers and similar.
The rule of thumb: if you have to raise your voice to be heard by someone at arm’s
length, it is loud enough to need protection. You do not need a meter, and most shop
tools fail this test the moment they start.
The numbers behind it
OSHA measures noise as an average across the working day, in A-weighted decibels
(dBA). Two figures matter:
85 dBA
The action level, averaged over 8 hours — the point at which a workplace must run a hearing conservation programme.
90 dBA
The permissible exposure limit, averaged over 8 hours.
Above 90 dBA the allowed time drops fast — every extra 5 dBA halves it:
90 dBA
8 hours
95 dBA
4 hours
100 dBA
2 hours
105 dBA
1 hour
110 dBA
30 minutes
115 dBA
15 minutes or less
Impact noise should not exceed 140 dB peak — a single hammer blow or
nail gun can reach it, and that damage is immediate rather than cumulative.
These are OSHA workplace limits (29 CFR 1910.95). NIOSH recommends a stricter
85 dBA limit that halves the allowed time every 3 dBA rather than every 5. Hearing loss
from noise does not come back, so treat these as ceilings, not targets.
Two types available
Earplugs
Small foam plugs inserted into the ear canal. Disposable or reusable. Available on request from the instructor.
Ear muffs
Over-ear cups covering the whole ear. Reusable and adjustable. Often preferred for intermittent use.
Both types significantly reduce harmful noise levels.
Foot protection
Foot protection
Minimum requirement: closed-toe shoes are MANDATORY — no sandals, no flip-flops, no open-toe footwear.
Recommended upgrades
Slip-resistant soles, for traction on surfaces that may be slippery.
Puncture-resistant soles, against sharp metal debris.
Steel-toe or composite-toe protection — recommended, not required.
Avoid
Canvas shoes or thin-soled footwear.
High heels or unstable footwear.
Anything that does not fully enclose your feet.
You will be standing for most of the lab period — choose something
supportive and comfortable.
Respiratory protection; Stop
Respiratory protection
MANDATORY when using any grinder or sander, and when working with materials
that produce hazardous dust.
Grinding galvanized material is PROHIBITED. Zinc fumes from galvanized steel
are toxic when heated.
Health hazards
Metal dust and particles cause respiratory irritation and long-term lung damage.
Wood dust can cause allergic reactions.
Dust and fumes are not the same problem, and they do not take the
same protection. The next section is the one that decides which you need.
Respiratory protection — what to wear; Stop; Watch out
Respiratory protection — what to wear
A dust mask does nothing against fumes or vapours. An N95 is a
particle filter. It does not stop gases, welding or brazing fume, solvent vapour, or
anything else you can smell. Wearing one while you can smell the air is not protection.
For dust — grinding, sanding, sawing
N95 mask
Filters 95% of airborne particles. The minimum for dusty work.
Half- or full-face respirator
Fitted with a particulate filter, for heavy or extended dust.
For fumes and vapours
You need an approved respirator with the right chemical cartridge, not a
dust mask. Ask the instructor which cartridge before you start.
A respirator only works if it seals. A loose mask, or a beard along the
seal, defeats it.
If there is no correct cartridge available, the job does not happen until
there is.
If you can smell fumes or taste particles, your protection is inadequate.
Stop work immediately and upgrade it — or stop the job.
Loose clothing and accessories catch in rotating machinery. That is how this
causes serious injury or death — it is not a dress code.
Prohibited
Loose-fitting clothing, or garments with dangling features — drawstrings, ties and similar.
Long sleeves, hoodies or jackets near rotating equipment.
ALL jewelry, except silicone wedding rings.
Neckties, scarves and lanyards.
Hair, and what to wear instead
Hair, and what to wear instead
Hair
Long hair must be secured back and kept away from equipment.
Use hair ties, clips or caps to prevent entanglement.
Recommended
Cotton or fire-retardant materials — preferred, not required.
Close-fitting, short-sleeved shirts.
Pants without excessive loose material.
Safety/Universal Rules
Lifting and Carrying
Team lifting above 40 lbs, and the technique that keeps you in control of the load. What usually defeats a lift is the size of the thing, not its weight.
Lifting Rules; Watch out
Lifting Rules
Get a second person for anything over 40 lbs. A gym weight plate is
45 lbs and nobody lifts one in pairs. Size — not weight — is what makes shop
materials hard to carry.
Forty pounds of aluminum bar is about four inches square and two feet long: you can hold it
against your chest, see over it, and set it down where you meant to. A twenty-foot piece of tube
can weigh less than that and still put half its length out in front of you, and a sheet of
plywood fills your arms and hides the floor you are walking on. Same weight, different job.
OSHA sets no weight limit — no standard says how much a person may
lift, only that a workplace be kept free of recognized hazards. The number everyone borrows
comes from NIOSH, the federal institute that researches workplace injury:
51 lbs with two hands under ideal conditions, cut for every way a real
lift is worse. Hold that same load at arm's length instead of against you and it rates about
20 lbs. Turning a quarter circle as you lift takes off another 30%, and
nothing good to grab takes off 10% more. A load much wider than your body is outside what the equation covers
at all — used there, it underestimates the hazard.
Team lifting is required for
Anything over 40 lbs.
Items lighter than 40 lbs but awkward in shape, size or balance.
Any object that feels unstable or is difficult to grip.
The 40 lb line is as much about dropping it as about your back.
A load you cannot fully control is a hazard to your feet, the floor, the workpiece and anyone
walking past — well before it is a hazard to your spine.
If you cannot hold it against your body and keep your back straight, it is too heavy
for you to lift alone — whatever the number says. Weight still wins in the end;
that point just sits somewhere different for every person and every shape.
Proper lifting technique; Checklist
Proper lifting technique
Plan the lift — clear your path and know where you are going.
Get close to the object before lifting.
Squat down using your leg muscles; keep your back straight.
Grip firmly with both hands; avoid awkward hand positions.
Lift smoothly using your legs, not your back.
Keep the load close to your body throughout the lift.
Setting it down
Use the same technique in reverse — squat down, do not bend over.
Place objects gently. Do not drop them.
Safety/Universal Rules
Electrical Safety
Why 50 V is a workplace control threshold, why 5 V can still burn you, the inspection required before every use, cords on the floor, and what to do when someone is shocked.
Electrical Sources
Electrical Sources
What hurts you at 5 V is not what hurts you at 240 V; one is a burn hazard
while the other is a shock hazard.
5 V / 12 V
The power supply on the bench, batteries, and the battery
pack out of a cordless drill. Lower shock risk, not a magic cutoff. Not below the
burn threshold.
120 V
Wall outlets. Dangerous — can cause serious injury or
death.
240 V / 480 V
Industrial equipment. Extremely dangerous — often
fatal.
Voltage tells you whether a circuit can push current through you. It does
not tell you what that circuit can do to a ring, a dropped tool or a shorted cell.
Those are two different hazards with two different thresholds. The next two sections are
about how each hazard changes.
Electrical — Shock Hazard; Stop
Electrical — Shock Hazard
There is no magic voltage at which shock suddenly becomes possible. A shock can be
painful without being life-threatening, and its severity depends on current, path, duration, and
your body’s resistance. Higher voltage can drive more current through the body, so the risk of
serious injury rises with it. In U.S. workplace rules, 50 volts is a common
control threshold: at or above it, live parts must be guarded and work near exposed live parts
requires specific controls.
OSHA requires live parts to be de-energized before you work on or near them — and then
says that “live parts that operate at less than 50 volts to ground need not be
deenergized if there will be no increased exposure to electrical burns or to explosion due to
electric arcs” (29 CFR 1910.333(a)(1)). Read that second
half again; the rest of this page hangs on it.
At or above the line, live parts must be guarded against accidental contact.
The rule is written for “electric equipment operating at 50 volts or more”
(29 CFR 1910.303(g)(2)(i)).
That number applies to DC as well as AC. The standard does not distinguish
between them, and OSHA has stated it “considers all voltages of 50 volts or above to be
hazardous” (interpretation letter, 4 September 2015).
NFPA 70E draws the same line. Below 50 V you may work live only where the
capacity of the source and any overcurrent protection have been considered and found
not to add burn or arc exposure (2024 edition, 110.2(B) Exception No. 5).
Supporting detail: Addendum 1 — “It’s not the voltage that gets you, it’s the current”.
Supporting detail: Addendum 2 — The international standards are not one number.
In this lab, use 50 V — AC or DC. It is the number OSHA enforces,
it is the most conservative of the figures above, and it assumes nothing about your skin being
dry or your hands being clean. If a circuit is at or above 50 V, it is not something you
work on live.
Electrical — Burn Hazard; Stop
Electrical — Burn Hazard
Below 50 V, the primary concern often shifts from shock to
short-circuit current. This is the half of the page people get wrong, because the
voltage people assume is harmless is the one nobody is careful with.
Look at what the exemption above is actually conditioned on. Not shock — burns
and arcs. NFPA 70E names the same condition as the capacity of the source.
Both standards are telling you that once you are under that workplace control threshold, the
question changes from how many volts to how many amps can this thing deliver into
a dead short.
Where the current comes from
A bench supply set to 5 V will deliver its full rated current into a
short without complaint. That is what the current limit is for — and if you have wound the
limit up to get a motor turning, you have wound the protection off.
A lithium cell or pack has almost no internal resistance. Shorted, a single
cell can dump tens of amps within milliseconds. A pack does considerably worse.
A lead-acid battery is rated in hundreds of cold-cranking amps — and
that is a 30-second load test, not a ceiling. A dead short pulls far more.
None of these can push meaningful current through your body at 12 V. All of them can push
hundreds of amps through a piece of metal — and metal carrying current it was never sized
for gets hot, fast.
Metal jewelry is the classic way this happens
ALL jewelry is already PROHIBITED here — see
clothing and jewelry. That rule was written for rotating machinery. It earns its keep a
second time on the bench, and OSHA writes it a second time for exactly this reason: conductive
jewelry “may not be worn if they might contact exposed energized parts”
(29 CFR 1910.333(c)(8)) — a rule that states no voltage
floor. The same sentence gives the way out: such articles may be worn if they are
“rendered nonconductive by covering, wrapping, or other insulating means.”
A ring or a metal watch band bridging a low-voltage source is a near-zero-resistance path
across something that can supply hundreds of amps. The ring becomes the heating element. This is
documented injury, not a story: a mechanic’s gold ring touched the positive terminal of a
12 V car battery while the wrench in his hand bridged the ring and the
negative terminal, and the result was a deep circumferential burn around the base of the finger
that needed a skin graft. A review of nine ring burns at one burn center found the same pattern:
the small size of the burn hides how bad it is, because a burn that wraps a
finger is far more likely to need grafting and to leave lasting loss of function than a burn of
the same area anywhere else.
You can drop a wrench. You cannot drop a ring.
The silicone wedding ring that the PPE rule allows is also the right answer
electrically: silicone does not conduct, so it cannot become the short.
A shorted lithium cell does not just get warm
An external short is one of the recognized triggers for thermal runaway
— the cell generates heat faster than it can shed it, and the outcome is
“extremely high temperatures, violent cell venting, smoke and fire” (UL Research
Institutes). Once it has started, disconnecting it does not stop it. The cell is
now its own energy source.
NEVER leave loose cells or packs where metal can bridge the terminals
— a drawer with tools in it, a pocket with keys in it, a bin with offcuts in it. Federal
transport rules require lithium cells and batteries to be packed so their terminals cannot short
(49 CFR 173.185). The reason that rule exists is the reason to do the same
thing on your own bench.
A cell that has been shorted, dropped, crushed, punctured or has swollen is out of
service. Tell the instructor. Do not put it back on charge to find out. See
Fire Safety.
The sentence to take away: low voltage does not mean low energy.
Voltage helps determine whether a circuit can shock you. Current decides whether it can burn
you — and current is set by what the source can deliver and what is in the path,
not by the number printed on the front of the supply.
Electrical Conductors; Stop; Watch out; Checklist
Electrical Conductors
Inspect every cord, every time, before it goes into the receptacle.
Never run your hand along a cord while it is plugged in; that is a dangerous way to find fraying.
It takes about five seconds, and it is the one check that catches most of this page before it
happens rather than after.
With the cord unplugged, run it through your hand end to end. Look for frayed,
cracked, cut or exposed insulation, and for flat spots where something has been driven
over it.
Check the plug: blades all present, straight, tight in the body, not
discolored or pitted. Check the ground pin is still there — a missing
ground pin means the cord is out of service, not that you have found a workaround.
Check where the cord enters the plug and enters the tool. That is where
flexing breaks conductors, and it is where the damage hides.
Inspect extension cords for cuts, crushing and wear, and check the receptacle end for
scorching or melted plastic.
Damage like this means the cord is finished, not that it needs watching.
If you find damage, stop, unplug it, and notify the instructor. Not
“just for this one cut”. Tag it or hand it over, so the next person does not find it
the hard way.
Tape is not a repair. OSHA allows electrical tape over superficial scuffs in
the outer jacket, and has stated just as plainly that “tape repairs of the jacket may not
be used to bring a worn or frayed flexible cord into compliance” (interpretation letter,
16 December 1998). If conductors or their insulation are showing, the cord is done.
Connections and plugs
Always disconnect power before performing any maintenance or adjustment.
NEVER yank or pull a cord — grasp the plug, not the cord. Pulling by
the cord wrecks the strain relief where you cannot see it, and it is how a plug ends up
half out of the receptacle with the blades exposed.
Push a plug fully home, and look at it. A plug that will not seat is a
fault to report, not a fit problem to force.
Keep electrical connections away from water and wet surfaces. Wet skin has on the order of
1,000 ohms of resistance where dry skin has 100,000 ohms or
more, so water moves you roughly a hundredfold closer to the current every rule on this
page is trying to keep out of you.
Be aware of overhead wiring when you are handling long stock, moving a
ladder or lifting.
A partly unplugged appliance is a short waiting for something to land on it
The hazard is the gap, and it is there before anything lands in it.
Exposed blades sit at full line voltage with nothing between them but air. Anything
conductive that falls across them completes the circuit straight across the supply — and
in a shop, the thing that falls is a metal chip. What follows is not a shock.
It is a dead short: a bang, a hard flash, and enough heat at the contact point to weld and
pit the blades. It will very likely trip the breaker and it can wreck the outlet or whatever
is plugged into it. The breaker protects the wiring, not your face, and it does not act until
after the flash.
Plugs are all the way in or all the way out. Nothing in between, and nothing
left in between.
Chips travel further than you expect and they conduct — see
Metal Chips and Swarf. That is the direct line between the housekeeping rules and this one.
Cords on the floor — the one most likely to get you; Stop; Checklist
Cords on the floor — the one most likely to get you
Nothing else on this page is as likely to actually injure someone as a cord
lying across a walkway. It is not a shock hazard at all. It is a trip hazard
— and a trip in a shop is not a trip onto carpet. You land on concrete, or into a machine, or on something sharp.
The regulation treats this as a floor problem rather than an electrical one:
“walking-working surfaces are maintained free of hazards such as sharp or protruding
objects, loose boards, corrosion, leaks, spills, snow, and ice”
(29 CFR 1910.22(a)(3)), and passageways and work areas are kept “in a
clean, orderly, and sanitary condition” (29 CFR 1910.22(a)(1)). A cord
across an aisle is an obstruction in an aisle.
A trip in a shop is not a trip onto carpet. You land where the machines are.The correct route is longer and costs nothing. That is the whole argument.
What to do, in order of preference
Do not cross the walkway. Move the work to the outlet, or use a different
outlet. This is almost always available, and it is almost never the first thing anyone tries.
Route along the wall or the bench line rather than across open floor. Take
the long way around the edge of the room.
Go overhead where a drop or a reel is provided. That is what they are
for.
If it genuinely must cross, cover or anchor it. OSHA’s own guidance is
to “ensure that temporary electrical cords that cross aisles are taped or anchored to the
floor”. A cord ramp or cord cover does the same job better and protects the cord as well
as the person.
Use the shortest cord that reaches. Slack on the floor is the hazard, and a
coiled surplus in a walkway is the same hazard in a tighter package.
Unplug it and put it away when you are done, in the same movement as
shutting the tool down. The cord left out after the job is finished is the one nobody is
expecting.
PROHIBITED — running a flexible cord through a doorway, a window, or a hole in
a wall, ceiling or floor, and attaching it to building surfaces. Those are named
prohibitions in the electrical standard (29 CFR 1910.305(g)(1)(iv)),
along with using flexible cord as a substitute for the fixed wiring of a building. Flexible cord
is temporary by definition. A cord treated as permanent gets pinched, crushed and forgotten.
NEVER walk past a cord lying across a walkway and leave it because it is not
yours. If you can see it, you can move it or say something. The person who trips over it will
not have seen it.
Read from the cord’s side, the rule is the same: portable equipment must be
handled so it is not damaged, and every cord set must be inspected before use on any shift for
cuts, crushing and a pinched jacket (29 CFR 1910.334(a)). A cord that gets
walked on is a cord that fails the inspection above. This section and that one are the same
problem twice.
Keeping walkways clear is the Set in Order and Shine part of
5S. The safety case and the lean case are the same case.
If someone is shocked; Stop; Watch out
If someone is shocked
Do NOT touch them directly. If current is still flowing they are part of the
circuit, and grabbing them puts you in it as well.
Kill the power first — the machine switch, the disconnect, the
breaker, or pull the plug.
If you cannot kill it, use a dry, non-conductive object to separate them
from the source, and stand on something dry while you do it.
A small mark is not evidence of a small injury. Electrical and electrothermal
burns run deep with very little to see on the surface, and a burn that wraps a finger can cost
the use of it. Any shock, and any electrical burn, gets reported and gets looked at
— however fine the person says they are.
Safety/Universal Rules
Chemicals and Safety Data Sheets
Every chemical and liquid in the shop has a Safety Data Sheet. Read it before you use the material, not after.
What is a Safety Data Sheet (SDS)?
What is a Safety Data Sheet (SDS)?
A Safety Data Sheet is the manufacturer's written answer to
what their product can do to you, what to wear, and what to do when it goes wrong.
There are sixteen sections, always in the same order. You do not need to read a SDS front to
back but you should read the following three sections before you use a product.
Section 2
Hazard(s) identification. What the chemical can do
to you, in the manufacturer's own words — a signal word, the hazard statements, the
red-bordered pictograms.
Section 4
First-aid measures. On skin, in eyes, breathed in,
swallowed.
Section 8
Exposure controls / personal protection. Gloves,
eye protection, ventilation — the PPE answer for that specific chemical, rather
than the general one.
None of this is a local rule. OSHA requires a sheet for every hazardous chemical, and requires
the sixteen sections and the order they come in — 29 CFR 1910.1200(g).
Supporting detail: Addendum 3 — All sixteen sections, and the four OSHA does not enforce.
Supporting detail: Addendum 4 — Signal words, pictograms and H-numbers — how Section 2 is written.
Where is SDS binder located?
Where is SDS binder located?
All Safety Data Sheets are organized in a designated binder for easy reference.
Review before use — check the sheet for any unfamiliar chemical or liquid.
Ask the instructor if you cannot locate the sheet for a material.
The binder is one way of satisfying the rule underneath it.
1910.1200(g)(8) requires an employer to keep the sheets "readily accessible
during each work shift to employees when they are in their work area(s)", and allows
electronic copies in place of paper "as long as no barriers to immediate employee access in
each workplace are created by such options." Binder or screen, the test is the same: the
sheet is something you can reach now, not something you have to go and ask for.
Your responsibility: know the hazards of any chemical you are using, and
follow the proper handling procedures.
Supporting detail: Addendum 5 — Telling a current sheet from an old one.
Material Storage and incompatible materials
Material Storage and incompatible materials
Flammable liquids must be stored in an approved flammables cabinet when not in use.
NEVER store incompatible chemicals together.
All containers must be properly labeled.
“Incompatible” is not a judgement call, and answering it needs no
chemistry. It is written on the sheet, by name, in two places:
Section 10
Stability and reactivity. One of its required
entries is "incompatible materials" — what this chemical must not contact.
Beside it sit "conditions to avoid" and "hazardous decomposition products",
which is what the two of them would make.
Section 7
Handling and storage. Required to give
"conditions for safe storage, including any incompatibilities" — the same
question answered for the shelf rather than for the reaction.
So the test before two things share a cabinet: open both sheets and read Section 10 on
each. The entry may name a specific substance or a whole class of them — oxidizers,
strong acids. If either sheet names the other material, they do not go together. Read both,
because the pair may only be called out on one of them.
Supporting detail: Addendum 6 — What makes a flammables cabinet “approved”, and how to spot one.
Safety/Universal Rules
Machine Hazards and Housekeeping
The two ways a machine injures you - it catches hold and pulls you in, or it closes on you - plus the environment around it and the habits that keep both off you.
Rotating machinery; Stop; Watch out
Rotating machinery
Rotating machinery injures you by catching hold of something and pulling
it in. It does not have to catch your hand. It catches a sleeve, a glove, a lanyard or a
strand of hair, and your hand goes wherever that goes.
Gears, pulleys and belts, chains and sprockets are extremely dangerous.
Where a rotating part meets another surface there is a wedge that closes to nothing — an
in-running nip point, the place the machine feeds things into itself. It is
where the belt runs onto the pulley, where the chain runs onto the sprocket, where the gear
teeth mesh.
The contact and the injury happen in different places. You touch the belt out
in the open, where it looks like a moving strap; the belt carries you to the pulley, where there
is no gap at all. That distance is why staying clear of the nip point is not enough on its own
— anything the machine can grip is a handle it uses to bring you there.
Gloves. Gloves or rags are for handling sharp objects such as cutters, and
only once the machine is shut down. NEVER wear gloves at a running
machine — a glove catches where bare skin would pull free, and takes your hand
with it. See chips and gloves.
It does not take teeth, and it does not take speed. A plain smooth shaft turning
slowly will still wind hair or cloth onto itself, and once it has a turn or two it does not let
go.
Contact happens where the belt is open. The injury happens where it closes.
Supporting detail: Addendum 7 — The word for this, and where it comes from.
The rotating parts that have no cover
Belts, gears and chains belong behind covers. The working end of a machine does
not. The spindle — the driven shaft the motor turns — the
chuck that grips the tool or the work, and the stock (the raw
material) held in it all turn out in the open. Nothing is between you and them, because you have to
reach them to do the job.
So everything above about what gets caught applies here with nothing in the way — and it
applies to every machine that turns, not only the loud ones. A drill press is
quiet, it is slow, and it looks like the harmless one. It is still an open chuck spinning, and it
will wind hair, a sleeve or a glove onto itself like anything else. Slow is how long it
takes, not whether it happens.
The sleeve tore before the arm did. That is the shirt’s doing, not the machine’s.
The lathe is where this goes furthest, and the reason is what a lathe is. Most
machines hold the work still and move a cutter across it; a lathe does the opposite and
turns the work. What is spinning in the open is then not a bit the width of a
pencil — it is the whole part, as long and as heavy as that part happens to be. That is why
the fatality reports for lathes describe something other than a hand injury: a person wound onto
their own workpiece.
Wound on, cloth is a rope. Each revolution takes up another turn, and bracing changes nothing.
The question, before anything is switched on: what is turning, and what of mine can
reach it? A sleeve, a glove, loose hair and a strip of abrasive cloth held against the
work have each pulled somebody into a machine. Every one of them is easier to deal with before
the spindle is turning than after.
None of this is a ranking. The mechanism is the same on every machine that turns,
and the one that can catch you is whichever one you are standing at.
Supporting detail: Addendum 8 — What the investigations actually describe.
Moving machinery and pinch points; Watch out
Moving machinery and pinch points
Moving machinery injures you without ever catching hold of you.
Two surfaces close on each other with you in between. Nothing pulls you anywhere — you run
out of room.
What makes a pinch point
A pinch point is a gap that is closing: two parts, at least one of them
moving, with space between them that is on its way to nothing. That is the whole of it, and it is
why pinch points are hard to spot. There are no teeth, no edge and often no noise. A wide gap and
a narrow gap look the same until the last inch, and the gap does not care what is in it.
Two kinds of movement, and both make one:
Back and forth, up and down
A press ram, a drill head coming down, a saw guide
being lowered, a clamp closing. You are caught between the part that moves and the part that
does not.
Straight along
A table, a carriage or a saddle travelling. It closes the gap
against whatever waits at the end of its travel — a stop, a column, the machine’s
own frame, or the hand resting there.
Where they hide
Between the work and whatever holds it. A vise jaw, a clamp, a fixture: the
hand steadying the workpiece is inside the gap that is closing on the workpiece.
At the end of a table’s travel. Wind it far enough and the space
between the table and the frame goes to zero, with whatever was resting there in between.
Under anything that comes down. A head, a ram, a guard, a guide.
On hinges. Covers, doors and lids close on their own frames — at the
free edge, where a hand is holding them, and again right at the hinge, where the gap closes over
almost no distance at all.
A pinch point needs no power, no speed and no noise. A vise turned by hand
multiplies what you put into the handle many times over, and it closes slowly and quietly enough
that nothing warns you. Switching a machine off removes the part that would have closed the gap
for you. It does not remove the gap.
The question to ask
Before moving anything — a handwheel, a lever, a clamp, a guard —
what closes, and what is in the way? Follow the moving part to the end of its
travel and find what it meets there. Then look at where your hands are, and where they will still
be once you are watching the cut instead of watching your hands.
A pinch point is a gap on its way to zero. It looks safe for most of the trip.
Supporting detail: Addendum 9 — Why belts and chains count as rotating, not pinching.
Guards stay on, and taking one off is not your call; Stop
Guards stay on, and taking one off is not your call
A guard is the thing standing between you and everything the two sections above
describe. It is not an accessory, and it is not in your way — it is doing the one job
nothing else on the machine does.
No one removes a guard in this lab. A machine that needs a guard off is the
instructor’s to deal with, out of class.
That is the whole rule, and it is deliberately not a judgement call. You do not have to decide
whether this particular setup is the exception, whether the cut is short enough, or whether you
will be careful. If the guard is in the way, the answer is to stop and say so — not to
work out whether you can manage without it.
It also means a machine you find with a guard already missing is not ready to use.
Say something. You have not done anything wrong by finding it that way, and the person who uses it
after you has no way of knowing.
Supporting detail: Addendum 10 — Why the rule has no exceptions in it.
Lockout/Tagout — before anyone works on a machine; Stop
Lockout/Tagout — before anyone works on a machine
Everything on a machine that turns or travels does so because something is
powering it. Lockout/Tagout is how that power is taken away and kept away while somebody
works on the machine.
NEVER work on equipment with rotating or moving parts without first
disconnecting power.
Lockout/Tagout procedures must be followed for maintenance.
The two words
Lockout
A lock on the device that isolates the energy — the disconnect,
the breaker, the valve — so it physically cannot be switched back on.
Tagout
A tag on that same device saying who put it there and why, so nobody
removes it wondering.
Off is not disconnected, and disconnected is not still
Off is not disconnected. A machine’s own on/off switch is a control.
Someone can press it back on, and either way the machine is still connected to its supply. The
disconnect is the thing that gets locked.
Disconnected is not still. A spindle coasts. A raised head or ram drops. A
spring, a charged capacitor or air still in a line will all move something after the power is
gone. Energy already inside the machine is part of what the procedure has to deal with.
Supporting detail: Addendum 11 — What the standard actually covers.
The environment around the machine
The environment around the machine
Slip hazards
Clean dust, metal chips, water and coolant off the floor immediately.
Lighting
Ensure lighting is adequate before operating any machine.
Walkways
Keep aisles and emergency exits unobstructed.
Operating habits
Operating habits
The chuck key never leaves your hand. It is not released until it is out of
the chuck. There is no moment where the key is in the chuck and not in your grip — not to
reach for something, not to answer someone, not for a second.
Stop machines completely before taking measurements or making adjustments.
Secure all workpieces before operation — use clamps, vises or fixtures.
Never leave tools on or near operating machinery.
Supporting detail: Addendum 12 — Why the chuck key rule is worded as a grip and not as a reminder.
When to fight a fire and when to leave, and how to use an extinguisher.
Decide before you reach for anything; Stop; Checklist
Decide before you reach for anything
An extinguisher is for a small fire you caught early.
Fight it only if all four are true
The fire is small and contained — a bin, a rag, a puddle. Not spreading.
You know what is burning, and you have the right extinguisher for it.
You have a clear exit behind you and the fire is not between you and it.
Someone else has already raised the alarm.
Otherwise: get out. Pull the alarm, leave, call 911 from
outside, and tell the instructor and the fire crew what was burning.
If you try and it does not go out in the first few seconds, stop and leave.
A hand extinguisher lasts roughly 10–20 seconds. That is your whole attempt.
Using an extinguisher — PASS
Using an extinguisher — PASS
Remember the acronym: P–A–S–S.
When it looks out, back away watching the fire, exit behind
you. Do not assume it is over — a knocked-down fire can reignite.
Fire classes, and what puts them out
Fire classes, and what puts them out
Fires are classed by what is burning. An extinguisher is rated
for the classes it can handle, and using the wrong one ranges from useless to dangerous.
Most extinguishers on a wall are ABC dry chemical — one
unit covering the first three classes. Check the label; the classes it is rated for are printed
on it.
Metal fires, and what it takes to start one; Stop; Watch out
Metal fires, and what it takes to start one
Metal burns only when it is divided finely enough, and then only when several
conditions line up at once. Ordinary aluminum chips and swarf will not do this,
and neither will the quantities of fine material a lab space like this one generates. Solid stock
does not burn at all. The cases below are the ones that are genuinely different.
The form matters more than the metal
Magnesium is the real Class D case. Thin sections, chips and turnings
ignite readily and burn with intense heat and a brilliant white light, and dull tools alone can
produce enough heat during machining to start it.
Titanium turnings can ignite, and titanium fines and grinding dust are
worse — fine enough that a static discharge is a credible ignition source.
Aluminum fines — the dust off sanding, grinding, buffing and
polishing — are a genuine hazard, but only through the specific mechanisms below.
Aluminum chips and swarf are hard to ignite. Aluminum carries heat away
from a hot spot fast, it wears a tough oxide skin that holds together while the metal
underneath melts, and a chip has little surface area for its mass. Put a torch to a pile of
chips and they melt.
Solid stock does not burn. A block of aluminum is not fuel.
Magnesium and titanium are not common in student labs. A senior project
might bring one in. If you are about to cut, grind or sand either one, ask
first — every time.
Size is the usual line — shape is the exception
The line in the standard is about half a millimeter, measured across a
particle’s smallest dimension — which different editions write as
500 microns or as 420 microns (fine enough to pass a 40-mesh sieve), not two separate
limits. Below it, metal dust can carry an explosion. Above it, generally it cannot — and
generally is the word the standard uses, not never.
Flakes are the exception, and buffing and polishing make flakes. A flake is
wide but thin, so it is too big for the sieve and still burns like dust. Treat anything off a
buffing or polishing wheel as fines however big it looks. Machining chips are genuinely coarse
— that is the difference between a chip tray and a dust collector.
The ways fine metal actually ignites
A suspended dust cloud meeting an ignition source. Fines sitting in a tray
are not the hazard. Fines thrown into the air, at the right concentration, reaching a flame, a
torch, a spark or a static discharge, are — burning travels from one particle to the next
with extreme speed. It is why you never blow fine dust around, and why you do not raise a cloud
of it cleaning up.
Fines heavily mixed with rust or grinding scale. Aluminum plus iron
oxide — rust, or the scale off steel grinding — is a thermite reaction: the
aluminum strips the oxygen off the iron, so the mixture burns without needing air from outside.
That is what makes it a metal fire rather than an ordinary one, and why water and an ABC
extinguisher do nothing to it. It takes a real mixture to go — something near
three parts iron oxide to one part aluminum by weight, and an ignition source
far hotter than a spark. That is a rust-caked pile, not a trace. See Metal Chips and Swarf for how far
apart that threshold and the recycling one actually sit.
Wet fines. Aluminum fines in contact with water slowly give off hydrogen.
That is a reason not to seal damp fines into a closed container, and part of the reason water is
the wrong thing to put on a metal fire.
If fine metal does ignite
Do not fight it. Most general-purpose shops have no Class D
extinguisher — assume you do not have one unless somebody has shown you
where it is. Pull the alarm, get out, call 911, and tell the fire crew
it is a metal fire. That changes how they fight it, and they need to know
before they arrive with a hose.
PROHIBITED — the ABC extinguisher off the wall. “Dry
powder” (Class D) and “dry chemical” (ABC) are not the same thing. An ABC
unit throws an incompatible agent at pressure: it scatters burning fines and spreads the
fire.
NEVER put water on a metal fire. The stream throws burning material, can
raise an explosive dust cloud, and reacts with the metal to release hydrogen.
Alarm, evacuate, report.
Supporting detail: Addendum 13 — Why dry sand is on the list, and why it is probably not on yours.
What that means day to day
Do not let fine dust accumulate. Empty trays, vacuums and collectors rather than letting
them fill — see Machine Hazards and Housekeeping for housekeeping.
Chips and fines are disposed of differently, and the difference matters — see
Metal Chips and Swarf.
Let hot chips and debris cool before they go anywhere near a bin of paper or rag.
Ask before you machine an unfamiliar metal. If you do not know what it is,
that is the answer: ask.
Hot work and the things that start fires
Hot work and the things that start fires
Grinding throws sparks a surprising distance. Clear rag, sawdust, solvent
and packaging out of the area before you start, not after it smolders.
Oily rags — it depends entirely on which oil.Drying oils — linseed, tung, oil-based stains, varnishes and finishes
— cure by reacting with oxygen, and that reaction gives off heat. Bunched in a pile the
rags hold the heat in, and the pile can reach ignition on its own with nothing to light it.
Those rags NEVER go in a pile: spread them flat to dry, drop them in water, or
use the designated container.
Petroleum oils do not do that. Way oil, cutting fluid, hydraulic oil and
ordinary shop oil do not cure by oxidation, so a rag wiped with them will not self-heat and will
not spontaneously combust. It is still combustible — a spark or a hot
chip will light it like any oily cloth — so keep it away from hot work and out of a bin of
paper or sawdust. If you do not know which oil is on the rag, treat it as a drying
oil.
Chips come off hot, razor sharp and fast. How to handle them, how to clean them up, and the rules on compressed air.
Four reasons a chip is dangerous
Four reasons a chip is dangerous
A chip is not scrap until it has cooled and stopped moving. Straight off the
tool it is every one of these at once.
Hot
Where an aluminum chip forms, the cut runs roughly
250–650 °C (500–1200 °F), depending on the alloy
and the speed. Chips do not look it.
Sharp
The leading edge of a chip is always razor sharp. It is
a freshly cut edge and it will open a hand.
Strong
A long continuous chip is a steel ribbon. If it wraps a sleeve or a
limb it can pull you into the machine.
Slippery
Chips on the floor roll underfoot. This is one of the most common
ways people get hurt without the machine touching them.
Never clear chips with your hands; Stop
Never clear chips with your hands
Do not reach into a machine to clear chips while it is running, and do not
pick up a long stringy chip by hand at any time. Stop the spindle, then deal with it.
Gloves
Gloves may only be used for cleanup after the machine is shut
down.
Never wear gloves at a running machine. A glove catches where bare skin
would pull free, and takes your hand with it — see Machine Hazards and Housekeeping.
Brush
A brush may be used while the machine is in operation — your hands stay
outside the machine. If clearing the chip means reaching in, stop the spindle first.
On a CNC machine, take particular care: the tool can move without warning, and the
enclosure invites you to reach further in than you should.
Compressed air — there are limits on this; Watch out; Checklist
Compressed air — there are limits on this
Compressed air is convenient and it is the option people reach for first. It is also
regulated, because a chip in a moving air stream is a projectile and air
forced against skin can be driven into the body.
The rule
OSHA permits compressed air for cleaning only where it is
reduced to less than 30 psi, and only then with
effective chip guarding and personal protective equipment
(29 CFR 1910.242(b)). All three conditions, together — not the
pressure alone.
Chip guarding is anything that stops a chip reaching eyes or unbroken skin: a
protective cone nozzle covers the operator, and a screen or barrier is what covers everyone
else.
In practice
The 30 psi is measured at the nozzle, dead-ended — blocked by the
work or a hand — and is met by a limited air nozzle or a regulator, not by the supply
pressure upstream of it.
Eye protection, every time. Everyone nearby too — you do not get to decide the
range.
Never at skin, yours or anyone else’s.
Never toward another person, and check who is behind the machine first.
Never into an enclosed space.
If air would just relocate the mess into the air or across the shop, use a brush or the
vacuum instead.
Cleaning up; Stop; Checklist
Cleaning up
Use cleaning equipment or gloves to handle chips for disposal — the sharp edge does
not dull because the machine stopped.
Clear the machine so only the residue in the bottom tray remains.
Sweep the floor, and mop if the floor is wet with coolant.
If cleaning equipment such as a vacuum gets wet, open it up and let it dry
before it is put away.
Chips and fines do not go in the same place
Chips and swarf go in the recycling container. That container is for
chips only.
Fines go in the trash. Dust from sanding, grinding, buffing and polishing,
and what comes out of a dust collector — never the recycling container.
Why — two reasons, and the second one is the safety reason
The recycler pays for clean aluminum. A clean load gets the aluminum price;
mix another metal in and the whole load drops to a much cheaper one. Fines pick iron up —
off a shared belt or wheel, off steel work at the next bench — so fines do not go in.
Compare how often each container actually leaves the building. A chip
recycling container fills slowly and can stand most of a semester between pickups; the trash
goes out daily. Fines belong in the stream that leaves, not the one that stands in the corner
for months quietly collecting more of them. Anything whose problem is accumulation does
not go in the container designed to accumulate.
Two thresholds, same observation. “There is iron in the
fines” is a real problem for the recycling price and not a fire problem — and the
reason both are true at once is that the two thresholds are orders of magnitude apart. The price
threshold is a trace: a few steel chips are enough to drop the whole load to the
cheaper price. A thermite reaction needs iron oxide — rust or grinding
scale, not bright steel — at something near
three parts oxide to one part aluminum by weight, plus an ignition source far
hotter than a spark. That is most of the pile, not a trace. Most rules in a shop have a number
behind them; knowing which number you are being measured against, and how far you
actually are from it, is the difference between following a rule and understanding it.
Fine dust burns; chips do not. Aluminum chips and swarf are hard to ignite.
The fire hazard is fines — and then only when a cloud of them is thrown
into the air and meets an ignition source, or when a great deal of rust or steel grinding scale
is mixed into them. Do not let fine dust accumulate, and do not raise a cloud of it
cleaning up.
If fine metal debris does ignite, do not fight it. The ABC extinguisher on
the wall is the wrong tool and will scatter it. Alarm, evacuate, and tell the fire crew it is a
metal fire. See Fire Safety.
Safety/Universal Rules
Lab Conduct and Emergencies
What is never acceptable in a lab, what alertness actually requires, and what to do when something goes wrong.
Prohibited behavior; Stop
Prohibited behavior
NEVER operate equipment under the influence of drugs, alcohol, or medication that causes drowsiness.
NO horseplay, pranks or roughhousing in any lab area.
Do not distract others who are operating equipment.
No unauthorized experiments or modifications to equipment.
Staying alert
Staying alert
Stay focused — maintain concentration while operating equipment.
Take breaks when tired or distracted.
Ask questions if you are unsure about any procedure.
Report unsafe conditions immediately to the instructor.
Emergency procedures; Checklist
Emergency procedures
Know the location of emergency stops, fire extinguishers, first aid kits and exits.
Call 911 for serious injuries or fires.
Notify the instructor immediately of any incident, injury, or near-miss.
Do not move a seriously injured person unless they are in immediate danger.
For electrical shock specifically, the first action is different — see
If someone is shocked.
Workspace responsibility
Workspace responsibility
Clean as you go — do not wait until the end.
Report equipment problems before they become safety hazards.
Lab Conduct and Emergencies
Lean and Quality/Lean and Organization
5S Workplace Organization
The five-step workplace organization method. It is a lean manufacturing practice, and it is also why a tidy shop is a safer one.
Why 5S sits in a safety module
Why 5S sits in a safety module
5S is a lean manufacturing method for organizing a workplace. It
earns its place here because most of what it asks for is also what keeps people from getting hurt.
A cluttered bench hides a hazard. An unlabeled container is an unknown chemical. A floor with
chips on it is a slip. Every one of those is a 5S failure before it is an incident — which is
why the shutdown steps in After operating any equipment are 5S in
everything but name.
This page is the basics. 5S will get its own expanded treatment later.
Sort — Seiri
Sort — Seiri
Separate necessary items from unnecessary ones.
Remove or relocate items that do not belong in the work area.
Keep only what you need for the task in front of you.
Set in Order — Seiton
Set in Order — Seiton
Organize and label the remaining items clearly, for easy identification.
A place for everything, and everything in its place.
Create logical, efficient storage systems.
Shine — Seiso
Shine — Seiso
Clean work areas to maintain high hygiene standards.
Reduce accident risk through cleanliness and organization.
Inspect while cleaning — that is how problems get found early.
Check the setup and the material extending through the spindle; Checklist
Check the setup and the material extending through the spindle
In our labs, check the setup before starting the lathe.
Check that the chuck or other workholding grips the part securely and that the cutting tool is secured.
The chuck key never leaves your hand. It is not released until it is out of the chuck (see the chuck-key rule).
Check the support for long material, including material extending through the back of the spindle. In our labs, never run long, unsupported material through the spindle.
Check the planned movement of the tool and carriage against the chuck, workpiece and supports.
Locate the emergency stop and foot brake, and confirm how to operate them before starting.
A long unsupported bar can whip as it rotates. Support at the cutting end does not establish that the part behind the spindle is safe.
The whole workpiece rotates; chip strings can catch and cut
The whole workpiece rotates; chip strings can catch and cut
The chuck, workpiece and any material projecting from the spindle are rotating hazards. Apply the entanglement rules to the entire setup, including its back side.
In our labs, stop and clear long chip strings when they become a hazard. Never pull a chip string by hand. Use Metal Chips and Swarf for removal and cleanup. If long strings keep forming, have the instructor review the cutting setup before continuing.
If the work shifts or the setup behaves unexpectedly, stop safely
If the work shifts or the setup behaves unexpectedly, stop safely
Do not try to steady rotating work or catch a loose part. Use the stopping controls you were shown if you can reach them safely, keep clear, and notify the instructor. Use Lab Conduct and Emergencies for an injury or emergency.
Check workholding, cutter clearance and cutting settings; recognize the different categories for manual and CNC mills.
Manual mill: Category 3; CNC mill: Category 2 in the Metals lab
Manual mill: Category 3; CNC mill: Category 2 in the Metals lab
A vertical mill uses a rotating cutter to remove material held in a vise or clamps. See Vertical Mill Anatomy and Motion for the machine's parts.
The manual mill is Category 3 in the Metals lab; the CNC mill is Category 2 there. Use Equipment Categories and Supervision for the corresponding permission and supervision requirements.
Check clamping, cutter travel and cutting settings before the cut; Checklist
Check clamping, cutter travel and cutting settings before the cut
In our labs, check and recheck the workpiece clamping before beginning an operation.
Check the vise or clamps and the cutter's retention. Confirm the tool can follow the planned path without striking the workholding.
Identify the closing gaps between the moving table, workholding and machine frame. Keep the planned travel clear.
Check spindle speed, feed, depth of cut and width of cut for the planned operation.
Check coolant and slide lubrication as directed for the installed mill. Report low levels, damage or a maintenance need rather than adjusting the machine's sliding fits yourself.
A rotating cutter and a moving table create different hazards
A rotating cutter and a moving table create different hazards
The cutter can cut or catch what touches it; the table can close a gap against the column, vise or clamps. Keep clear of both the cutting area and the table's travel. CNC motion follows the program, so a pause in motion is not permission to reach in.
Use Machine Hazards and Housekeeping for stopping before measurements and for isolation before adjustments or maintenance. Use Metal Chips and Swarf for chip removal, including the required controls if compressed air is used.
If a tool breaks or the work comes loose, step away
If a tool breaks or the work comes loose, step away
In our labs, step away if a tool breaks or the workpiece becomes loose. Stop the mill if it is safe to do so; do not endanger yourself to reach a control. Notify the instructor in every case.
Before starting: secure the bit, work and vise; Checklist
Before starting: secure the bit, work and vise
Select a bit and spindle speed suitable for the material and hole diameter. Have the instructor confirm an unfamiliar setup; a list of bit types is not approval to use every type in this press.
With the machine isolated for setup, seat and tighten the bit using the installed chuck's procedure. The chuck key never leaves your hand. It is not released until it is out of the chuck. Do not count on a key ejecting itself (see the chuck-key rule).
Clamp wood and metal to the table, or grip the work in a vise secured to the table. Secure small parts in a suitable vise or fixture; clamp larger parts with adequate support. Handholding is not workholding.
Check that the bit's full travel clears the table, vise and clamps. For a through-hole, provide clearance beneath the work or suitable sacrificial backing.
Secure the table at the required height and check that the work cannot rotate or lift when the bit catches.
Feed steadily and keep hands away from the rotating work
Feed steadily and keep hands away from the rotating work
A bit can catch and turn an unsecured part into a spinning lever. That is why a vise must be secured as well as tightened around the work.
Start with the bit clear of the work. Feed steadily so it cuts without being forced.
Ease the feed as the bit breaks through the underside. Keep hands clear rather than reaching underneath to support the exit.
Withdraw the bit under control, then stop the spindle before releasing or moving the work.
Supporting detail: Addendum 14 — Why larger drills usually need fewer revolutions per minute.
If the bit breaks or the part comes loose: stand clear
If the bit breaks or the part comes loose: stand clear
Do not grab a spinning part. Step clear and stop the drill press using a safely reachable stop control. For a broken bit, use the emergency stop if safely reachable. Wait for motion to cease and notify the instructor; do not restart until the fault and workholding have been checked.
Keep hands away from the cutting nozzle and clear of its travel. The jet cuts material, and the moving head can trap a hand. See Machine Hazards and Housekeeping for movement, guards and stopping. OMAX safety information is further reading for the installed-machine training.
Prepare the file and toolpath: lab steps 1–10
Prepare the file and toolpath: lab steps 1–10
This is our lab's OMAX operating sequence. Follow it with the machine-specific instruction you received at checkout.
Export the CAD file to DXF or DWG, as for a laser-cut part.
Turn on the OMAX machine.
Turn on the water: the top water line under the sink.
Open ProtoMAX Layout software.
Click File → “Open file for other CAD software.”
Use the move, rotate and duplicate tools to lay out the parts. One square is equal to one inch.
Click Autopath to create toolpaths. Check that the path cuts on the correct side of each line and cuts inside items before outside items, so the part is not cut loose before its interior features are cut.
Click Post.
Select the starting position, normally the bottom left.
Save the post. It saves as an .omax file.
Verify settings, follow the prompts and dry-run: lab steps 11–15
Verify settings, follow the prompts and dry-run: lab steps 11–15
Right-click on Post and open the .omax file in ProtoMAX Make.
When Make asks for cutting parameters, select the appropriate material and enter its thickness. The lab procedure says to oversize the thickness “by just a bit”; confirm that allowance with the instructor.
If the machine was just turned on, Make shows a red homing button at the top of the screen. Make sure nothing obstructs the nozzle's movement before clicking it to home the machine.
Follow all the on-screen steps. These include, but are not limited to:
Attaching the abrasive hose.
Testing the nozzle for abrasive flow.
Placing and securing the material.
Homing the nozzle to the starting position on the material.
Adjusting the water height if necessary.
Dry-running the program.
Flipping down the splash guard.
Run the program.
In step 14, “homing” to the material means the job's starting position; it is distinct from the machine-homing step in step 13. Have both positions demonstrated during checkout.
A dropped part or a fault needs a safe stop, not a reach into the tank
A dropped part or a fault needs a safe stop, not a reach into the tank
Do not reach into the cutting area to retrieve a part while the machine is operating. If the material moves or the cut behaves unexpectedly, keep clear and use the stopping procedure demonstrated during checkout. Notify the instructor; follow Lab Conduct and Emergencies for an emergency.
The lab sequence above ends at running the program. Have the instructor demonstrate shutdown and safe part recovery before using the machine; a part falling into the tank is not a recovery procedure. Use Equipment Operating Procedure for the general shutdown and reporting requirements.
Machining/Production and Nontraditional Processes
Material-Removal Process Families
Sort machine tools by how they remove material, then follow the process that fits the feature.
Start with what removes the material
Start with what removes the material
A machine tool controls motion between a tool and the workpiece, the part being made. In material removal, the final shape is what remains after unwanted material is taken away.
This is a teaching map of removal mechanisms. Abrasives also make small chips; the branches distinguish the kind of tool or energy doing the work.
CNC is a control method, not a removal family. Computer numerical control can guide a mill, lathe, grinder, EDM machine or cutting jet. See G-code and Toolpaths. Forming moves material without removing it; see Manufacturing Process Categories for the wider manufacturing map.
Conventional cutting uses defined cutting edges
Conventional cutting uses defined cutting edges
Turning rotates the work past a cutting tool; milling rotates the cutter and feeds it relative to the work.
Drilling makes holes; sawing separates stock, the starting material, with successive teeth.
Broaching uses a sequence of teeth to progressively cut a profile, such as a keyway. Dedicated tooling often suits repeated production, but broaching is not limited to large runs.
Shaping and planing use a back-and-forth straight cutting stroke. A shaper normally moves the tool; a planer normally moves the work.
A process name describes the action, not every possible shape. Milling can make curved contours as well as flat faces.
Abrasive processes use many small cutting grains
Abrasive processes use many small cutting grains
Grinding wheels hold abrasive grains that remove material. Precision grinding controls size and surface texture, often on hardened parts. Grinding can also remove substantial stock; it is not restricted to light finishing.
Honing commonly uses abrasive stones with combined motions to refine a bore's size, form and surface. Lapping uses fine abrasive between a lap and the work to refine a surface. Neither is simply another name for grinding.
Nontraditional processes change the removal mechanism
Nontraditional processes change the removal mechanism
Thermal:EDM, laser and plasma remove material with concentrated heat. EDM's energy supply is electrical, but its removal mechanism is thermal.
Electrochemical:ECM uses electric current to dissolve metal.
Match the tool to the feature, the allowed size range, and the measurement you need to trust.
Choose from the tolerance and the shape of the feature
Choose from the tolerance and the shape of the feature
Start with what must be measured: an outside size, an opening, a depth, a height, or a small change from a reference.
As a rule of thumb, choose a tool that reads about ten times finer than the total tolerance. This is a starting point, not a standard or a guarantee that the measurement is good enough. Check its specified accuracy, condition, measuring range, and the way its contacts reach the part.
For a size of 0.625 ± 0.001 in, the total tolerance is 0.002 in. A starting resolution target is 0.002 / 10 = 0.0002 in. An outside micrometer reading 0.0001 in is a candidate for an accessible outside diameter. A 0.001 in caliper is too coarse for that rule of thumb.
Resolution is the increment; accuracy is agreement with a reference
Resolution is the increment; accuracy is agreement with a reference
Resolution
The smallest indicated change or scale increment. A display that changes by 0.0005 in has that resolution; its last digit is not an accuracy certificate.
Accuracy
How closely a measurement agrees with the reference value. Use the instrument's error specification and checks against suitable standards to judge its capability.
Precision
How closely repeated measurements agree under stated conditions. A dirty measuring face can give repeatable readings that are all wrong.
Extra digits cannot remove jaw wear, a tilted contact, dirt, or heat. Record the instrument and units along with the result when the measurement matters.
Use a steel rule for layout and quick size checks
Use a steel rule for layout and quick size checks
A steel rule is useful for stock lengths, approximate sizes, and setting a divider. Read the graduation pattern on the edge you are using: fractional edges may divide an inch into 8, 16, 32, or 64 parts; decimal edges may use 0.1, 0.05, 0.02, or 0.01 in increments. A decimal rule does not ordinarily have 0.001 in graduations.
Place the rule along the distance being measured and look straight at the marks. A damaged end is a poor zero. If you start at the 1 in line instead, subtract 1 in from the far-end reading. Do not use a rule to accept a feature with a tight size tolerance.
Calipers cover outside, inside, depth, and step measurements
Calipers cover outside, inside, depth, and step measurements
Dial, digital, and vernier calipers use a sliding jaw. The larger jaws measure outside sizes; the smaller jaws measure openings. A depth rod reaches down from the end of the beam. On tools with step faces, the back shoulders measure a step.
The readout changes how you read the tool, not how carefully you must position it.
Readout
Use and reading
Watch for
Dial
Quick mechanical reading: beam scale plus dial remainder.
Check the marked value per revolution. Dirt in the rack can disturb the reading.
Digital
Direct numerical display; often offers inch/mm and relative zero.
Confirm units and zero mode. Resolution varies by model; a common display increment is 0.0005 in / 0.01 mm.
Vernier
Main scale plus the aligned vernier division; no battery or dial mechanism.
Read the correct scale and its least count. Inch 0.001 in tools can use different main-scale spacings.
Use light contact and keep the jaws square. Measure near the beam where access allows. Long jaws can flex; small inside jaws may not reach the true diameter of a small hole. A caliper's versatility does not make it the best tool for every tolerance.
Micrometers specialize in outside sizes, inside sizes, or depths
Micrometers specialize in outside sizes, inside sizes, or depths
An outside micrometer measures between a fixed anvil and a moving spindle. Choose a range that contains the dimension: a 1–2 in instrument does not start from zero like a 0–1 in instrument. A conventional inch scale reads 0.001 in; a sleeve vernier can add 0.0001 in readings. A conventional metric scale reads 0.01 mm. Digital models differ.
The sleeve supplies the coarse reading; turning the thimble moves the spindle. The ratchet or friction device helps control contact force.
An inside micrometer spans an opening using internal contacts. Rod and two-point types require careful alignment; three-point types have a different contact arrangement. Check the reference and any extension length for the actual tool.
A depth micrometer rests its flat base on the reference face and extends a rod to the lower surface. Keep the base seated. Check the installed rod's range and the scale direction: many depth scales run opposite to an outside micrometer. Do not add a rod length without checking how that instrument is marked.
Indicators show change from a reference
Indicators show change from a reference
A dial indicator commonly has a plunger moving along its axis. A test indicator has a pivoting lever stylus and usually a shorter measuring range. Both are useful for alignment, comparing heights, and observing runout as a part turns.
Read the marked graduation and count revolutions where needed. Zeroing the bezel sets a reference; it does not turn the indicator into an absolute diameter tool. Total indicator reading, or TIR, is the maximum indication minus the minimum during a sweep. From −0.001 to +0.002 in, TIR is 0.003 in.
Use a rigid stand. A magnetic base needs a suitable magnetic support; it will not attach magnetically to granite. For plunger indicators, align motion with the plunger. For lever test indicators, follow the maker's stylus-angle instructions; the lever is normally close to parallel with the surface. These are different geometries.
Supporting detail: Addendum 15 — What a sweep can and cannot establish.
Gauge blocks supply a reference; transfer gauges carry a size; Practice
Gauge blocks supply a reference; transfer gauges carry a size
Gauge blocks have accurately finished opposing faces and a known length. Use suitable blocks to set a height, compare a measurement, or check an instrument. Clean and inspect the faces before wringing blocks together; follow their care instructions. A zero check alone is not a full calibration.
Build stacks from sizes actually present in the set. For example, 1.0000 + 0.1340 + 0.1005 = 1.2345 in uses three blocks if those sizes are available. Do not assume a set contains separate 0.0040 or 0.0005 in blocks.
A telescoping gauge transfers a bore size to an outside micrometer. With light locking friction, set the contacts slightly across the bore at an angle and sweep gently through the diameter; withdraw without changing the setting, then measure across the contacts. Follow the tool's instructions for locking and repeat until the technique is consistent.
A small-hole gauge expands rounded contacts inside a smaller opening. Adjust for light drag, withdraw as its design permits, and measure across the contacts with a micrometer. Neither transfer tool has its own numerical readout. Take readings at several depths and directions; keep them separate to reveal taper or out-of-roundness rather than averaging the shape away.
Read rules, micrometers, calipers, and verniers by separating the coarse scale from the remainder.
Read the unit and the graduation before the number
Read the unit and the graduation before the number
The same-looking line can mean 0.025 in, 0.5 mm, or something else. Identify the instrument and its scale before adding anything.
On a fractional rule, count spaces from the last whole inch using that edge's denominator. Seven 1/16 in spaces beyond 2 in gives 2 + 7/16 = 2.4375 in. That exact fraction-to-decimal conversion does not make the rule accurate to four decimal places.
On a decimal rule with 0.01 in spaces, eight spaces beyond 1.20 in gives 1.28 in. On a digital caliper, read the displayed unit and confirm whether the tool shows an absolute size or a deviation from a stored zero.
In inch shop speech, a thou is 0.001 in and a tenth often means 0.0001 in, one tenth of a thousandth. Write units and decimals to avoid confusing a shop tenth with 0.1 in.
An inch micrometer adds sleeve plus thimble
An inch micrometer adds sleeve plus thimble
For the conventional 0.001 in micrometer, one turn advances the spindle 0.025 in. The sleeve has a line every 0.025 in; each fourth line marks another 0.100 in. The thimble divides one turn into 25 spaces of 0.001 in.
Read the last sleeve line reached by the thimble edge. Count the 0.100 in intervals and the extra 0.025 in lines.
Read the thimble line at the sleeve's horizontal datum line.
Add the two. Include the range offset for a larger micrometer according to its markings.
Example: sleeve 0.300 in plus two 0.025 in lines, with thimble 18 at the datum:
0.300 + 0.050 + 0.018 = 0.368 in
Watch the rollover. Sleeve 0.375 and thimble 24 gives 0.399 in. The next 0.001 in step gives sleeve 0.400 and thimble 0: 0.400 in. Do not add the old 0.025 in interval again.
A vernier micrometer adds the fourth decimal place
A vernier micrometer adds the fourth decimal place
The sleeve vernier divides a thimble space into ten parts. Read the sleeve and the whole 0.001 in thimble amount first, using the thimble graduation just passed when the datum falls between lines. Then find the sleeve vernier line that aligns with a thimble line.
Example: the last sleeve line is 0.450 in, the datum is between thimble 12 and 13, and sleeve vernier line 7 aligns:
0.4500 + 0.0120 + 7 × 0.0001 = 0.4627 in
The matching vernier line adds 0.0007 in, not 0.007 in. Do not round the thimble to 13 and then add the vernier. If the zero line aligns, add zero; the end line represents the same alignment at the next interval.
A metric micrometer includes the exposed half millimeter
A metric micrometer includes the exposed half millimeter
On a conventional 0.01 mm micrometer, one turn moves the spindle 0.5 mm. The sleeve shows whole millimeters and half-millimeter marks; 50 thimble spaces divide the turn into 0.01 mm increments.
If 8 mm and the following 0.5 mm mark are exposed, and thimble 23 aligns, the reading is 8 + 0.5 + 0.23 = 8.73 mm. If that half-millimeter mark is still covered, the same thimble position gives 8.23 mm.
At the rollover, 8.49 mm is sleeve 8.0 plus thimble 49; 8.50 mm is sleeve 8.5 plus thimble 0. Use the marked unit; these increments are not the inch micrometer's increments converted by eye.
A 0.001 in vernier caliper adds the aligned division
A 0.001 in vernier caliper adds the aligned division
For the 25-division inch vernier used here, each main-scale space is 0.025 in. Twenty-five vernier spaces occupy the length of 24 main-scale spaces, so each vernier space is 0.024 in. Their 0.001 in difference makes the fine reading possible.
The shortened diagram shows the first few lines of a 25-division scale. The mismatch accumulates by 0.001 in per division.
Read the last main-scale line at or before the vernier zero.
Find the vernier division that lines up with a main-scale line.
Add its division number × 0.001 in.
Vernier zero just beyond 1.325 in, with division 17 aligned, gives 1.325 + 0.017 = 1.342 in. It is the vernier zero, not the right end of the slider, that locates the coarse reading.
Supporting detail: Addendum 17 — Other 0.001 in verniers can have different spacing.
A dial caliper adds the beam reading and the dial remainder
A dial caliper adds the beam reading and the dial remainder
On an inch dial caliper marked 0.100 in per revolution, the dial's 100 divisions each represent 0.001 in. Read whole inches and completed 0.100 in intervals on the beam, then add the dial reading.
Beam 1.200 in and dial 47 gives 1.200 + 0.047 = 1.247 in. Beam 1.200 plus dial 99 is 1.299 in. At the next thousandth the beam reaches 1.300 and the dial returns to zero.
Some dial calipers use a different travel per revolution. Check the dial markings rather than assuming this example applies. Clean the jaws, close gently, check zero, and confirm the reading again after measuring.
Practice reading a new micrometer or vernier scale; Interactive; Practice
Practice reading a new micrometer or vernier scale
The trainer draws ideal scales with zero error. It teaches reading, not instrument accuracy. Micrometer examples stay within 0–1 in; caliper examples stay below 2 in. Enter the exact indicated reading in inches.
Read these scale positions before uncovering the answers. These examples also serve as the print version of the trainer.
Inch micrometer: sleeve 0.275 in, thimble 24 at the datum.
Vernier micrometer: sleeve 0.275 in, datum between thimble 24 and 0, vernier line 9 aligned.
25-division vernier caliper: zero past main line 1.150 in, division 8 aligned.
Readings as sums
1. 0.275 + 0.024 = 0.299 in. 2. 0.2750 + 0.0240 + 0.0009 = 0.2999 in. 3. 1.150 + 0.008 = 1.158 in.
Control zero, contact force, alignment, cleanliness, and temperature before trusting the last digit.
Clean, check, measure, and check again; Checklist
Clean, check, measure, and check again
A good reading starts before the scale is read. Use this sequence with the instrument's own instructions.
Identify the feature, its limits, and its reference surfaces. Choose a suitable measuring tool.
Clean the contact faces and the part. Look for chips, dirt, raised burrs, and damage that would hold the contacts apart.
Let the part and instrument settle to the measuring environment. Avoid a freshly cut hot part or a tool warmed by long handling.
Check zero or the correct setting reference. Close a 0–1 in outside micrometer gently; use the appropriate setting standard for larger ranges or inside tools.
For a critical size, check a suitable known reference near the working size as well. Confirm the instrument is within its required calibration status.
Measure with the intended contact force and alignment. Release and repeat to check that you can reproduce the reading.
Record the result with units, location, and enough method information to repeat it. Check zero again afterward; investigate any change.
Zero adjustment removes an offset at one position. It cannot repair worn jaws, a damaged screw, or an error that changes along the range. If a tool fails its checks, have it evaluated rather than hiding the error with a new zero.
Use consistent light contact instead of squeezing out a number
Use consistent light contact instead of squeezing out a number
With a micrometer, bring the faces close, then use the ratchet or friction thimble as its maker directs. These devices help repeat the measuring force. They do not make dirt, tilt, or a flexible part disappear.
Use a smooth, slow approach. The required clicks or turns depend on the mechanism; a friction thimble may slip without clicking. Do not force the main thimble after the force-limiting device operates. The spindle lock holds a setting; it is not a clamp for tightening the measurement.
With a caliper or transfer gauge, feel means consistent light contact or drag, not a hard squeeze. Too much force bends jaws and compresses thin parts. Repeated readings should agree because the setup is repeatable, not because the tool has been forced to the expected number.
Keep contacts square and look straight at the scale
Keep contacts square and look straight at the scale
A tilted instrument can span a diagonal rather than the intended distance. For an outside diameter, keep the measuring direction perpendicular to the cylinder's axis. For a two-point inside measurement, the contacts must pass through the bore center in the cross-section while staying square along its length. A single instruction to “always find the largest” or “always find the smallest” does not cover both alignment directions.
Parallax is a reading error from looking at separated scales at an angle: the marks appear to shift against each other. Put your eye square to the reading line. Good lighting helps distinguish a line from a reflection.
For an indicator, keep its intended motion and contact geometry aligned. Preload enough to maintain contact throughout the expected movement, while staying within its usable travel. Do not apply a universal number of needle turns to both plunger and lever instruments.
Abbe error grows when tilt acts through an offset
Abbe error grows when tilt acts through an offset
The scale on a caliper runs along the beam, while the part often touches farther down the jaws. If the moving jaw tilts a little, its tip moves more than the scale location does. The scale misses that extra motion. This is Abbe error.
Picture a door: a small rotation moves the handle much farther than a point near the hinge. Greater distance between the measurement line and the scale line gives tilt more room to create error.
Keep the part near the beam when access permits, use light contact, and avoid loose or worn sliders. An ordinary outside micrometer puts its measurement and screw axes in line, reducing this particular source of error. Alignment still matters.
Supporting detail: Addendum 19 — The offset-and-angle relationship.
Temperature and surface condition can move the result; Practice
Temperature and surface condition can move the result
Parts and instruments change size with temperature. Let them approach the same stable measuring conditions; follow the reference temperature and corrections required by the inspection procedure. Avoid direct sun, drafts, and prolonged hand contact with precision tools.
Clean contact faces matter because the instrument measures what touches them. A chip can enlarge an apparent outside size; a raised edge can hold a depth-micrometer base above its reference. Surface peaks can affect contact too. See Surface Texture and Finish for surface condition rather than treating every last-digit change as a size change.
Measure at several positions when taper or out-of-roundness matters. Keep the individual values. An average can hide an oversize region and does not prove the whole feature meets its limits.
In this invented classroom example, each value is a local diameter reading in a 20 mm-deep bore. Near, middle, and far name axial stations 2, 10, and 18 mm from the entry face. X and Y are two marked perpendicular diametral directions at each station.
Invented classroom local-diameter readings; treat the stated values as exact
Axial station from entry face
Diametral direction
Local diameter (mm)
Near — 2 mm
X
20.000
Near — 2 mm
Y
20.002
Middle — 10 mm
X
20.001
Middle — 10 mm
Y
20.003
Far — 18 mm
X
20.004
Far — 18 mm
Y
20.026
For stated limits of 19.980 to 20.020 mm, the six-reading mean is 120.036 / 6 = 20.006 mm, inside the limits. The far-depth Y reading is 20.026 mm, above the stated upper limit. Keep that location with the value: the in-limit average does not make the sampled high reading disappear. These six local readings do not establish the size or form of the unsampled bore.
Carry dimensions from a reference onto the work with clear lines, checked centers, and correctly read angles.
Start layout from a reference you can return to
Start layout from a reference you can return to
Layout places lines, arcs, and points on a part to guide cutting or locate features. A neat line is useful only when it is in the right place.
Identify the reference surfaces from which dimensions are taken. Keep using those references rather than measuring each new line from the last one. A layout mark guides the work; it does not prove the finished part is within tolerance.
A surface plate supplies a reference plane for layout and inspection. Clean the plate, the part's supporting face, and the tool base. Support the part so it cannot rock. Use an appropriate angle plate or V-block when the shape needs support.
Protect the plate from impact and abrasive work. Its flatness depends on its size, grade, condition, and calibration; “a surface plate” does not imply one universal accuracy. Follow its specified support arrangement. Level and flat are different: leveling a plate does not repair a worn surface.
Dye makes a fine scribed line easier to see
Dye makes a fine scribed line easier to see
Layout dye adds contrast. Prepare the surface, apply a thin coating as directed, and let it dry before marking. Use Chemicals and Safety Data Sheets for chemical handling guidance.
A scriber scratches a narrow line through the coating. Hold the guide firmly and draw the point along it with light, consistent pressure. Repeated passes can make a wide, ambiguous mark. Scribe only where the drawing and process allow a permanent scratch.
A surface gage carries an adjustable scriber at a fixed height above its base. Set it from a separate height reference, then move the base along the plate to scribe a line parallel to that plane. A surface gage ordinarily has no direct height scale.
A height gage adds a scale or readout. Set its reference correctly, seat the base flat, set and lock the height as intended, and make the line. With a suitable probe it can compare or measure heights. It is not simply another name for a depth gage.
Supporting detail: Addendum 20 — Finding the middle of two heights.
Dividers and trammels carry a distance into an arc
Dividers and trammels carry a distance into an arc
Dividers have two pointed legs. Set their spacing from a suitable scale, hold one point at the center, and swing the other to mark a circle or arc. They can also transfer or step off equal distances.
Trammels put adjustable points on a beam for radii too large for ordinary dividers. Lock the points after setting, support the beam, and check the spacing again before marking.
The point spacing is the radius, not the diameter. For a 3.00 in diameter circle, set 1.50 in between points. Recheck transferred dimensions against the original reference; repeated stepping can accumulate error.
Prick-punch the location; center-punch after checking it
Prick-punch the location; center-punch after checking it
A prick punch has a sharper point for a small, precise location mark at intersecting layout lines. A center punch makes a wider depression to help start a drilling operation. Point angles vary by tool; the important distinction here is the fine locating mark versus the wider starting mark.
Locate the point carefully, bring the punch upright, and make a light mark. Check the location before enlarging it. A heavy first mark is difficult to correct. Support the work on a suitable punching support, away from the precision surface plate.
A punch mark is not a substitute for the machine setup or spotting method needed to locate a hole accurately. Check the finished feature with a suitable measuring method. Use Personal Protective Equipment and Equipment Operating Procedure for the applicable shop rules.
Supporting detail: Addendum 21 — Punch types and their uses.
A combination set uses one blade with three different heads
A combination set uses one blade with three different heads
Square head
Provides 90° and 45° reference faces. With the blade it can lay out right angles, check squareness, or establish an offset or depth. Seat the reference face fully before reading or scribing.
Center head
Its two faces seat on round stock, positioning the blade along a diameter. Scribe one line, rotate the work or head, and scribe another; their intersection locates the center on the end face.
Protractor head
Sets the blade to an angle on a graduated head. Read the scale in the correct direction, lock the setting, and recheck contact with the reference.
Clean the mating faces and blade groove so the head seats properly. Pick the tool's accuracy for the task: a combination set is useful for layout, but it does not automatically replace a precision square or a finer angle instrument.
A universal bevel transfers an angle; a vernier protractor reads it; Practice
A universal bevel transfers an angle; a vernier protractor reads it
A universal bevel has an adjustable blade and stock that can fit against two surfaces. Lock it to transfer or compare an angle. A plain bevel has no numerical scale; obtain the angle from a separate reference.
A vernier bevel protractor combines the adjustable blade with degree and vernier scales. On a 5-minute instrument, each matching vernier division adds 5 arcminutes. 60 minutes = 1 degree, so 5 minutes is 1/12 degree. This is the reading increment, not a universal accuracy claim.
Seat the stock and blade against the intended surfaces. Identify the angle being measured.
Read the whole degrees between the main zero and the vernier zero.
Use the vernier on the same side and in the same reading direction as those degrees. Find the matching division.
Add the matching division number × 5 minutes to the whole degrees.
Example: 32 whole degrees and the seventh matching vernier division give 32° + 7 × 5′ = 32°35′. That is 32 + 35/60 = 32.5833…°, not 32.35°.
Check the contact geometry before reporting an obtuse angle. If the instrument has measured the acute supplement, a reading of 32°35′ corresponds to 180° − 32°35′ = 147°25′ for the required obtuse angle. Do not automatically subtract every reading from 180°.
Drawing views and dimensioning conventions belong in the Engineering Graphics collection. Layout follows those requirements; it does not redefine them.
Machining/Cutting Fluids
Cutting Fluids and Application Methods
Cutting fluids cool, lubricate, carry chips away and protect surfaces; match the fluid and its delivery to the tool, material and operation.
Cool the cut, reduce friction and carry chips away
Cool the cut, reduce friction and carry chips away
Coolant does more than cool. A cutting fluid removes heat from the cutting zone, lubricates contact between the tool, chip and workpiece, flushes loose chips away, and helps protect the machine and work against corrosion.
Lubrication can reduce material sticking to the cutting edge, called built-up edge. Together with cooling and chip removal, that can improve surface finish and dimensional control. Fluid cannot make up for a dull tool, poor workholding or unsuitable feeds and speeds.
The fluid must reach the cutting zone to do its job. A full tank with a blocked filter or a nozzle aimed away from the cut is not effective delivery.
Choose a fluid family for the operation
Choose a fluid family for the operation
Use a product approved for the machine, work material and tooling. These are general tendencies, not a product selection chart for a particular shop. Check compatibility with seals, coatings and the metals being cut.
Family
Strength and typical use
Tradeoff
Straight oil (neat oil)
Used without water dilution. Strong lubrication suits operations where sliding contact dominates, such as some tapping, broaching and slow cutting.
Less cooling than water-based fluids; oil mist, smoke and fire risk depend on the product and process. Not suitable for every machine.
Soluble oil (emulsion)
Concentrate disperses oil droplets in water. Combines lubrication with water's cooling for many general machining operations.
Needs concentration and contamination control. Water quality affects stability; residue and foam can develop.
Semi-synthetic
A water-diluted formulation with less mineral oil than a conventional soluble oil. Often balances cooling, lubrication and cleanliness in general machining.
Still needs water-quality and sump maintenance. Foam, residue and compatibility depend on the formulation.
Synthetic
A water-diluted formulation without mineral oil. Cooling and visibility can suit grinding and suitable machining operations.
Lubrication depends on its additives; do not assume all synthetics suit heavy cutting. Concentration, corrosion protection and contamination still need attention.
Concentrate and working fluid are different. A supplier's oil-content figure for the concentrate is not the percentage to put in the tank. The working concentration comes from that product's technical data sheet.
Supporting detail: Addendum 23 — Why the family name does not settle performance.
Run dry, mist or flood according to the cutting setup
Run dry, mist or flood according to the cutting setup
Dry: use when the tool supplier's guidance for that material and operation calls for it. Some carbide milling benefits from avoiding repeated heating and sudden cooling of the edge. Dry does not mean chips or dust need no control.
Mist or minimum-quantity lubrication: a designed system meters a small amount of fluid toward the cut when lubrication and access matter more than bulk cooling. It does not provide flood-level heat removal or washing. Use the specified enclosure and mist control.
Flood: a continuous flow can remove heat and wash chips from the cut when the setup calls for wet machining. Supply enough flow to reach the edge consistently; a weak, intermittent splash is not the same thing.
Oil mist is liquid droplets carried by air, not a residue-free gas. Neither “all carbide runs dry” nor “more coolant is always better” is a useful rule. Follow the tool and machine guidance for the actual cut.
Supporting detail: Addendum 24 — Why some milling runs dry while other milling needs fluid.
Read Brix, apply the product factor, and balance concentrate and water when mixing a fresh charge or correcting a partly full tank.
Zero the refractometer, then multiply Brix by the product factor
Zero the refractometer, then multiply Brix by the product factor
A refractometer measures how a sample bends light. On an analog Brix instrument, read where the light and dark areas meet. Brix is a scale reading; it is not automatically coolant percent.
Find the product's refractometer factor and working concentration range in its technical data sheet. Use the specified method for that product.
Clean the prism. Calibrate to zero with water, using the water and temperature procedure specified for the instrument and fluid. Fluid suppliers commonly specify the water used for mixing.
Remove the calibration water. Apply a representative, well-mixed coolant sample, close the cover and read the light/dark boundary. Avoid sampling a floating oil layer.
Calculate concentration (%) = Brix reading × product factor. Record the result and compare it with the product's range for the operation.
Clean the prism after use. Recheck a doubtful reading; contamination can blur the boundary or bias the result. Use the supplier's alternative test if the sample cannot be read reliably.
Read the boundary: 3.2 Brix. With an invented teaching factor of 1.5, concentration is 3.2 × 1.5 = 4.8%. That factor is not a shop recommendation.
There is no universal factor or target range. The SDS explains hazards and handling; the product's technical data sheet supplies the mixing and measurement information. See Chemicals and Safety Data Sheets.
Add concentrate to water, and state what the ratio means
Add concentrate to water, and state what the ratio means
Water first; add concentrate while mixing. Follow the product's mixing instructions in a clean container or approved mixing system. For emulsions, this lets the emulsifier disperse oil into the water. Adding water to a body of concentrate can form an unstable or inverted mixture that does not disperse properly.
Write the order of a ratio. 24 parts water to 1 part concentrate contains 25 total parts, so its concentrate percentage is 100 × 1 ÷ (24 + 1) = 4.00%. It is not 100 ÷ 24. For water:concentrate = W:1, use C = 100 ÷ (W + 1).
To make 12 L of that 4.00% premix, measure 0.48 L concentrate and 11.52 L water. Put the water in first. This is a volume-based teaching example with additive volumes, not a product recommendation.
Supporting detail: Addendum 26 — Why mixing order matters.
A falling level does not tell you how strong the coolant is
A falling level does not tell you how strong the coolant is
Measure both level and concentration before topping up. Evaporation mainly removes water, raising concentration. Coolant carried out on parts and chips removes both water and concentrate. Adding plain water lowers concentration.
A tank below the target is lean; one above it is rich. Compare with the actual permitted range before deciding it needs adjustment. A tank can be low in volume and rich in concentrate at the same time.
Top up fluid that is still serviceable. Persistent odor, separation, sludge or contamination calls for the evaluation and cleaning decisions in Coolant Condition and Tank Maintenance. New premix does not make spoiled fluid sound.
Calculate the premix needed to reach both target level and concentration; Interactive
Calculate the premix needed to reach both target level and concentration
Use the same volume unit throughout. Let V be the current tank volume, C its measured concentration, F the intended final volume, and T the target concentration. Concentrations below are entered as percent numbers, such as 6 for 6%.
Premix volume A = F − V
Premix concentration P = (F × T − V × C) ÷ A
Concentrate to measure = A × P ÷ 100. Water to measure = A − concentrate volume.
Example: a hypothetical tank has 36 L at 4.0%. Bring it to 48 L at an assumed target of 5.5%. Add A = 48 − 36 = 12 L. The premix must be (48 × 5.5 − 36 × 4.0) ÷ 12 = 10.0%: 1.2 L concentrate added to 10.8 L water.
Check the balance: the tank contains 1.44 L concentrate, the addition supplies 1.20 L, and 2.64 ÷ 48 × 100 = 5.5%. These assumed values teach the calculation; the product must permit the proposed makeup mix.
Stop when the result cannot be made. A must be positive. A negative P means even plain water cannot achieve the target at that final level; P above 100% asks for more concentrate than the addition can contain. A full tank has no top-up space. Seek an approved partial replacement or other correction instead of overfilling. Even a mathematical result between 0 and 100% may be outside the product's permitted mixing method.
Mix the approved addition separately, add it, circulate as specified, then measure again. The calculation assumes accurate volumes, uniform mixing and no losses during the adjustment.
5.00% concentrate 5 parts concentrate in 100 total parts. 95:5 = 19:1 water : concentrate.
Brix needs the product factor
6 Brix × an invented teaching factor of 1.0 = 6.00%. Use the actual product data sheet factor.
Correct a lean sump: account for the added volume
Teaching example: 100 L at 3%, target 5%, capacity 120 L. With x litres of neat concentrate, the balance is 3 × 100 + 100x = 5(100 + x).
x = 100(5 − 3) ÷ (100 − 5) = 2.1053 L. Final volume = 102.1053 L, within capacity. Check: (3 + 2.105263…) ÷ 102.105263… × 100 = 5.00%.
This is a concentrate balance, not an instruction to pour neat concentrate into a sump. Use the product-approved makeup method described on this page.
Predict what water replaces
Start with 100 L at 5%. Losing 20 L of well-mixed coolant leaves 80 L at 5%. Replace it with water: 4 L concentrate ÷ 100 L = 4%. Replace it with 5% premix: (4 + 1) ÷ 100 × 100 = 5%.
If instead only 20 L of water evaporates, all 5 L of concentrate remains: 5 ÷ 80 × 100 = 6.25%. Replacing that water restores 5%; adding 5% premix raises it to 6%. The loss mechanism matters.
Supporting detail: Addendum 27 — Derive the balance, or solve for an unknown addition volume.
Practice the reading, the ratio and the top-up balance; Practice
Practice the reading, the ratio and the top-up balance
Use fresh hypothetical values to check the method. Every question states its rounding and assumptions; none supplies a shop concentration recommendation.
The practice catalog contains the full Coolant Concentration and Mixing bank.
Machining/Cutting Fluids
Coolant Condition and Tank Maintenance
Recognize concentration drift and contamination, decide whether to top up or replace, and work through a general tank-cleaning sequence.
Look for oil, odor, solids and foam—not just a low level
Look for oil, odor, solids and foam—not just a low level
The sump is the reservoir that collects coolant for reuse. It also collects material that does not belong in the cutting fluid.
Tramp oil: unwanted lubricant leaking into the coolant, such as hydraulic oil or oil from sliding machine surfaces. Skim the floating layer and investigate its source; it can interfere with fluid condition and concentration readings.
Bacteria and odor: water-mixed fluids can support microbial growth. New or persistent bad odor is a reason to investigate, not a test that identifies the organism. Lack of odor does not prove the fluid is healthy.
Concentration drift: evaporation tends to make the tank richer; repeated water additions can make it lean. Measure using Brix and the product factor.
Chips and fines: chips settle and fine particles collect as sludge. They occupy tank space and can obstruct screens, filters and flow paths.
Foam: check concentration, water quality, contamination and air entering the delivery system. Foam is not a reliable concentration measurement.
Skin irritation: an overly strong mix can irritate skin; an overly weak mix can lose protection against microbial growth. Irritation is a reason to report exposure and investigate, not to guess a dilution.
Top up a serviceable tank; replace fluid that cannot be restored
Top up a serviceable tank; replace fluid that cannot be restored
Use a top-up when the existing coolant is fit for continued use and the problem is level or a correctable concentration. Choose the addition from the measured condition, not from the fresh-fill recipe. A rich tank may need a weaker makeup mix; a lean tank may need a stronger one.
Arrange a cleanout and replacement when contamination, unstable mixing or persistent odor cannot be brought under control by the approved maintenance process, or when the machine/fluid maintenance instructions require it. Investigate leaks and other causes so the fresh charge does not immediately inherit the problem.
Straining removes larger solids; it does not remove dissolved contamination or establish that drained fluid is reusable. Do not return rejected fluid just because it looks cleaner. Do not add household disinfectant or an unapproved chemical to mask odor.
Use the machine and fluid suppliers' maintenance intervals and condition checks. There is no universal calendar changeout interval or guaranteed sump life.
Clean the tank and its return path before refilling; Checklist
Clean the tank and its return path before refilling
Use this as a general sequence with the machine's approved maintenance procedure. Stop the machine and follow its isolation requirements before accessing the tank, pumps or chip-handling equipment. Only run pumps for a cleaning or circulation step when that procedure calls for it and the system is safely reassembled.
Removing old fluid without removing deposits leaves contamination ready to enter the fresh charge.
Stop and isolate the equipment as required by its maintenance procedure. Prepare suitable labeled receiving containers before opening or draining the system.
Skim accessible floating tramp oil. Drain or transfer coolant through a suitable strainer to catch larger chips, keeping the destination consistent with the reuse or waste decision.
Remove accumulated sludge and chips with the approved tools. Clean accessible screens, filters and return passages; inspect for blocked flow and leaks.
Clean dried coolant residue with a compatible cleaner diluted to its labeled instructions. Use the specified contact time and cleaning method; do not invent a stronger mixture.
Remove the spent cleaning solution. Rinse and drain as the machine, cleaner and fluid suppliers specify; capture those liquids for the approved waste route. Account for retained liquid before refilling.
Reinstall components and guards. Prepare fresh coolant by adding concentrate to water at the product's specified concentration, then fill to the machine's operating level.
Restore and circulate the system according to its procedure. Check flow and leaks, measure concentration after mixing, and record the fluid condition, result and service performed.
Use the safety pages for contact, PPE and waste handling
Use the safety pages for contact, PPE and waste handling
Minimize skin contact with coolant and cleaning liquid. Use the SDS for the actual products and the PPE guidance for exposure control. Use Metal Chips and Swarf for removing and handling solids; this maintenance page does not replace those rules.
Manage spent coolant, removed oil, sludge and cleaning liquid through the approved regulated-waste process. Do not pour them down a drain or onto the ground. The responsible waste handler must determine the applicable classification and disposal route from the fluid, contaminants and local requirements. “Water-based” or “biodegradable” is not disposal permission.
Supporting detail: Addendum 30 — Why the waste route depends on what is in the tank.
Machining/Sawing and Grinding
Cutoff Sawing
Choose a saw for the blank, hold the material securely, and leave enough material for the next operation.
Saw a blank with room for finishing
Saw a blank with room for finishing
Sawing often turns a long bar or a large plate into a manageable blank: the starting piece for machining. Plan both the material left for finishing and the kerf, the slot removed by the saw. Kerf is material lost between pieces; it is not finishing allowance.
Choose allowance from the required finished dimensions, cut quality, and later operations. A saw cut is not automatically a finished datum surface. Check dimensions and squareness before using it to locate the next cut. See Choosing Shop Measuring Tools and Shop Layout and Angle Measurement.
Match the saw's motion to the cut
Match the saw's motion to the cut
Saw
How it removes material
Typical use
Power hacksaw
A straight blade moves back and forth, cutting on one stroke and relieving on the other. Cutting-stroke direction depends on the machine.
Occasional straight cutoff work; the return stroke does no cutting.
Horizontal bandsaw
A continuous toothed loop advances through material clamped in a vise.
Bar, tube, and structural sections cut to length. Miter capability depends on the head or vise design.
Vertical bandsaw
The band passes through a table while the work is fed along a path.
Repeatable cutoffs where squareness and surface finish matter. Inspect for burrs even when the cut looks clean.
Abrasive cutoff saw
A thin abrasive wheel grinds through the section.
Cutoff work suited to the wheel's material rating, including some hardened materials. Heat and cut quality can require further finishing.
“Cold” distinguishes chip cutting from abrasive or friction cutting; it does not mean zero heat. Use Saw Blade Selection for toothed blades. An abrasive wheel has its own selection rules.
Hold every piece, including pieces in a bundle; Checklist
Hold every piece, including pieces in a bundle
NEVER hand-hold material for a horizontal saw cut. Use the machine's vise or a suitable fixture. Support long material at the saw's working height so it does not tilt the clamped piece. Arrange support for the cutoff without trapping the blade as the cut closes.
Nesting means arranging several pieces for one cut. Clamping the outside of a bundle does not prove that its inner pieces are held. Use a nesting arrangement and clamping system approved for the machine, and select the blade for the resulting section.
Identify the material, section shape, finished size, and finishing allowance.
Confirm that the vise or fixture grips each piece and that long stock is supported.
Account for the kerf when locating the cut or setting a length stop.
Plan how the cutoff will be supported as the blade breaks through.
After cutting, inspect the blank and remove the burr before the next setup.
Band speed is the distance the blade travels per minute, usually stated in feet/min or m/min. Feed is the rate or force that advances the blade into the work. They are different adjustments. Material grade and hardness, blade material, tooth geometry, and section shape all affect the starting settings.
Use the blade maker's cutting chart for the actual material, then follow the machine manual for setting it. Follow the blade maker's break-in procedure for a new blade. There is no single speed table or fixed percentage adjustment that fits every saw. Connect these terms to Feeds and Speeds.
On a saw designed to use cutting fluid, the fluid helps lubricate the cut, carry away heat, and clear chips. Use the approved fluid, concentration, and delivery method. Some abrasive cutoff systems are dry and others are wet; do not add coolant to a dry-only setup.
Supporting detail: Addendum 31 — Why finer teeth do not fix every bad cut.
Machining/Sawing and Grinding
Saw Blade Selection
Use section shape, teeth in the cut, tooth form, set, and blade material to narrow the choice before consulting the maker's chart.
Pitch spaces the teeth; gullets carry the chips
Pitch spaces the teeth; gullets carry the chips
TPI means teeth per inch along the blade. A higher TPI is a finer pitch. For evenly spaced teeth, pitch in inches is 1 ÷ TPI. The gullet is the hollow between teeth that provides chip space. Set is the sideways displacement of teeth, making clearance for the blade body in the kerf.
Pitch is measured along the blade. Set is seen from a different direction. Shapes are schematic.
Estimate teeth in the cut before choosing TPI; Interactive; Practice
Estimate teeth in the cut before choosing TPI
As a rule of thumb, aim to keep at least about three teeth engaged. For a uniform pitch crossing one continuous thickness of material:
Estimated teeth engaged = contact length in inches × TPI Minimum TPI for three teeth = 3 ÷ contact length in inches
Use thickness measured along the tooth line through solid material, not automatically the height of the part. For 0.25 in of continuous contact at 12 TPI, the estimate is 0.25 × 12 = 3 teeth. At the same thickness, the three-tooth calculation gives 3 ÷ 0.25 = 12 TPI.
Same tooth spacing, different contact length. The thinner section needs a finer pitch to meet the three-tooth estimate.
A calculated minimum is a screening value, not a blade specification. If it falls between available pitches, select a suitable finer option from the maker's chart rather than rounding down. Excessively fine teeth can run out of chip space in thick stock.
Supporting detail: Addendum 32 — Why the answer can be a fraction of a tooth.
Predict first: at mid-depth in the thin-wall tube below, would doubling TPI halve, double, or leave the estimated engagement unchanged? It doubles: the same contact length contains twice as many tooth spacings.
Dashed line = about three teeth; shading = below three. Each plot uses its own vertical scale.
Thin-wall tube, coarse blade
50.80 mm outside diameter; 1.00 mm wall; 6 TPI.
At depth 25.40 mm: 0.47 teeth across 2.00 mm of material. Across the cut: minimum 0.00; maximum 3.33 teeth.
Below the common rule of thumb of about three teeth.
Contact tends to zero at entry and exit. At mid-depth only the two walls count. The peaks occur where the blade touches the inside circle.
Thick solid bar, fine blade
50.80 mm contact thickness; 25.40 mm feed height; 24 TPI.
At depth 12.70 mm: 48.00 teeth across 50.80 mm of material. Across the cut: minimum 48.00; maximum 48.00 teeth.
At or above about three teeth here; this is not a blade approval.
Gullets may fill with chips; check the blade maker's guidance.
Estimated teeth = solid contact length ÷ tooth spacing; mm per tooth = 25.4 ÷ TPI. Readouts round to two decimal places; calculations keep the unrounded, real-valued estimate. Highlighted tooth tips illustrate one blade position, not a guaranteed integer count.
Round-section minima include the exact entry and exit tangencies, where contact is zero. A flat bar has constant contact through its feed height, including the limiting faces. Rotate the flat bar 90° to swap contact thickness and feed height; rotating a round bar or round tube about its axis changes neither. This model is a straight tooth line crossing an ideal section, with uniform pitch and no blade tilt.
The thick-bar, fine-blade preset highlights the full contact region as a reminder: gullets may fill with chips; check the blade maker's guidance. It sets no numeric upper limit. Tooth count alone cannot predict chip filling. For shop practice and hazards, see Cutoff Sawing and Safety collection.
A tube's outside diameter is not its solid thickness; Practice
A tube's outside diameter is not its solid thickness
A solid bar offers continuous material across the blade's path. A tube, angle, or channel presents walls and gaps, and the contact changes as the cut progresses. Select from a chart for tubes or structural sections, using wall thickness, overall dimensions, and orientation as that chart requires.
Thin walls generally need finer teeth than a solid bar of the same outside size. Do not multiply tube outside diameter by TPI and count the empty center as cutting engagement. Nor does simply adding two wall thicknesses describe every stage of a round tube cut.
Variable pitch alternates tooth spacing within a repeating group. It can reduce repeated tooth-impact vibration and suit changing engagement, but it still needs to match the material and section.
Common general-purpose form, typically with a zero rake face. Fine versions suit thinner sections.
Skip
Widely spaced teeth with generous gullets, typically zero rake; useful when softer materials and larger chips need clearance.
Hook
A positive rake face takes a more aggressive bite. Use where the maker recommends it for the material and the machine can support the cut.
Rake describes the angle of the tooth face that meets the material. Exact angles vary by blade; “hook” does not specify one universal angle.
Set pattern
Arrangement and use
Raker
A common sequence is left, right, straight. A general-purpose arrangement; the straight tooth is unset.
Alternate
Successive teeth are set left and right.
Wavy
Groups of teeth vary their sideways set in a wave. Common on fine-pitch blades for thin sections.
Set pattern does not replace pitch selection. Choose the complete blade specification, not a tooth-form name alone.
Choose blade material and dimensions for the machine
Choose blade material and dimensions for the machine
Carbon steel: an economical choice for suitable easy-to-machine materials and lighter service.
Bimetal: high-speed-steel cutting teeth joined to a flexible alloy-steel back; widely used for metal cutoff work.
Carbide-tipped: carbide cutting tips for applications needing greater wear resistance, with a compatible machine and cutting conditions.
Match blade length and thickness to the machine. Blade width, measured from tooth tip to back, also affects contouring; see Vertical Band Machining. Material name alone does not establish a blade's hardness limit or speed. Use the maker's product chart for the actual alloy and condition.
Supporting detail: Addendum 33 — Manufacturer resources for the next selection step.
Machining/Sawing and Grinding
Vertical Band Machining
Plan contour cuts around blade width and turning radius, and recognize the purpose of blade welding, band filing, and polishing attachments.
A vertical band follows a contour through the work
A vertical band follows a contour through the work
A vertical bandsaw supports the work on a table while a continuous band cuts through it. Moving the work along a marked path allows curved outlines as well as straight cuts. This can remove excess material before milling a profile.
A narrow blade can turn a tighter curve; a wide blade better resists wandering in a straight cut. Find the smallest radius on the intended saw path, then use the blade maker's width-versus-minimum-radius chart. Choose a width that fits both that path and the machine.
Radius is measured from the center of a curve to its edge; diameter is twice the radius. A chart's minimum radius is not a guarantee for every blade with that width: blade thickness and kerf clearance matter too. Do not force a wide blade around a tighter corner than its chart permits.
Supporting detail: Addendum 34 — Why relief cuts help with some outlines.
Blade welding closes the loop for some internal cuts
Blade welding closes the loop for some internal cuts
Blades may arrive as ready-welded loops. On equipment designed for it, a suitable open band can pass through a starting hole in the work and then be joined, making an internal contour accessible without a slot from the outside edge.
In general, blade joining involves preparing and aligning the ends, making a butt weld, heat-treating the joint as specified, finishing the weld flush, and inspecting it. The joint must flex around the wheels and pass through the guides. Follow the welder and blade maker's instructions for the exact sequence and settings; not every blade construction is suitable for a shop blade welder.
Supporting detail: Addendum 35 — Why joining includes more than making a weld.
Filing and polishing require their own bands and guides
Filing and polishing require their own bands and guides
Some band machines accept a file band to remove small amounts of material and refine a contour, or an abrasive band to smooth a surface. These are dedicated attachments with matching guides and operating instructions, not ordinary saw blades used differently.
Hold a part in a bench vise, saw by hand, choose and clean files, and remove burrs without losing the intended shape.
Use the vise to make the part steady
Use the vise to make the part steady
A bench vise holds the part while both hands control the tool. Clamp near the working area with enough material between the jaws to resist movement. Excessive projection makes the part flex and chatter. Use suitable soft jaw covers when the ordinary jaws would mark a finished surface, and support long pieces.
A conventional hand hacksaw cuts on the push stroke; Practice
A conventional hand hacksaw cuts on the push stroke
Install the blade with its teeth pointing away from the handle for the conventional push-cut setup, and tension it according to the frame's instructions. Choose pitch for the material actually crossed by the teeth: thin sections need finer teeth. Apply the three-teeth rule of thumb from Saw Blade Selection.
Start a shallow groove with controlled strokes. Once established, use long, even strokes that spread the work over the blade. Apply cutting pressure on the forward stroke and ease it on the return. Keep the blade aligned with the cut instead of twisting it to steer back to the line. Ease pressure as the cut nears completion and plan support for the separated piece.
File cut and coarseness are separate choices; Practice
File cut and coarseness are separate choices
A file removes small chips with many teeth. Choose its shape to reach the surface: flat for a flat face, round for a round opening, or half-round for both flat and curved areas. A safe edge has no cutting teeth, allowing work close to an adjoining surface that should remain uncut.
Cut pattern is not a coarseness grade. The drawing enlarges the tooth spacing for comparison.
Single-cut files have one family of parallel tooth cuts; they are useful for controlled finishing and draw filing. Double-cut files have crossing cuts that divide the teeth into smaller cutting points; they commonly suit faster stock removal.
Coarseness describes tooth spacing. Common grades run from rough, coarse, bastard, second-cut, smooth, to dead smooth, from coarser to finer. Availability varies. Compare grades within the same file family and length; the name alone is not a universal tooth-spacing measurement. “Second-cut” is a coarseness grade, not another name for double-cut.
Fit a suitable handle to a tanged file. For ordinary forward filing, guide the file with both hands, keep it level, apply pressure on the cutting stroke, and ease pressure on the return. Rocking rounds the edges of a surface intended to stay flat. Check progress frequently rather than filing past size.
For draw filing, hold a suitable single-cut file crosswise to the work and move it along the surface, with its length across the direction of travel. Use light, controlled pressure and orient it so the teeth cut on the working stroke. This helps blend marks and refine a flat surface.
Chips lodged in the teeth are called pinning; they can scratch the work. Clean with a file card, brushing with the tooth grooves to clear trapped chips. Store files so their teeth do not rub against other tools.
Supporting detail: Addendum 36 — File selection and care resources.
Remove the burr without changing the required edge
Remove the burr without changing the required edge
Deburring removes the thin unwanted lip left by cutting. Use a file for accessible straight edges, a suitable swivel-blade deburring tool for edges it can follow, or a suitable countersinking tool for a hole entrance. Match the cutter to the material and edge shape.
Take light passes and inspect the result. Removing a burr does not authorize a large chamfer or a rounded edge. Preserve the drawing's dimensions and edge requirements; if an edge break is specified, make that feature deliberately.
Scribers mark; hammers and punches move or indent material
Scribers mark; hammers and punches move or indent material
A scriber makes a fine line with a hard point. Use it for the marking method described in Shop Layout and Angle Measurement, rather than as a substitute for a punch.
A ball-peen hammer supplies controlled blows for appropriate metalworking tasks. A soft-face hammer can reduce surface marking when positioning a part; it does not guarantee that a delicate surface will be unmarked.
A center punch makes an indentation for locating a drill start. A pin punch drives a pin with a flat-ended tip sized for that pin. Choose the tool for the task: the pointed center punch is not a pin driver. Detailed measuring and layout procedures belong in Shop Layout and Angle Measurement.
Machining/Milling
Vertical Mill Anatomy and Motion
Identify what moves on a knee mill or bed mill, and distinguish spindle rotation, quill travel, head tram and part coordinates.
The cutter rotates; feed changes its position relative to the work
The cutter rotates; feed changes its position relative to the work
A vertical mill has a spindle whose axis is vertical in its normal setup. The spindle rotates the tool. The head carries the spindle and its drive; the column supports and guides the machine assemblies. The table carries the workholding.
Do not picture the workpiece as fixed in space. Depending on the machine, the table, head, or both move to feed the rotating cutter through the work. On a typical three-axis vertical mill, X is left–right, Y is front–back, and Z is along the vertical spindle axis.
A CNC position describes the tool relative to the work. A table moving left can therefore produce tool motion to the right relative to the part. Confirm the machine's jog directions and readout before locating a feature. Continue with Milling Operations for what those movements cut.
A knee raises the table; a bed mill supports it without a rising knee
A knee raises the table; a bed mill supports it without a rising knee
Assembly
Typical knee mill
Typical bed mill
Table
Travels along X on the saddle.
Supported by the bed and slides; often travels along X.
Saddle
Travels along Y on the knee.
May provide Y travel beneath the table; layouts vary.
Vertical positioning
The knee raises or lowers the saddle and table.
The head commonly travels vertically on the column.
Quill
A sliding sleeve gives the spindle a shorter axial feed, useful for drilling and fine depth adjustment.
May be absent; the head's Z motion provides the feed.
Identify the actual moving assemblies. “Bed mill” does not mean that every machine assigns Y travel to the head. CNC describes how motion is controlled, not whether the machine has a knee.
The setup checklist follows a Tormach workflow. Use the manual for the installed machine to identify controls and travel; a familiar brand name does not establish its spindle interface, options or limits.
Head tram and vise squareness check two different alignments
Head tram and vise squareness check two different alignments
Tram is the spindle axis's perpendicular alignment to the table in both left–right and front–back directions. An indicator carried around the spindle axis compares points on the table at the same sweep radius. This is an alignment check with spindle power off; head adjustment belongs to the machine-specific procedure.
Squaring the vise makes its fixed jaw parallel to an axis of table travel. It uses a straight sweep along the jaw, not a circular sweep over the table. A square vise does not prove the head is in tram, and a trammed head does not prove the vise is square. See Milling Vises and Strap Clamps.
Supporting detail: Addendum 37 — Why facing can reveal a tram problem.
Machining/Milling
Milling Vises and Strap Clamps
Seat and align a milling vise, support work on parallels or soft jaws, and understand how strap-clamp position divides the clamping force.
Clean contact surfaces before tightening anything
Clean contact surfaces before tightening anything
A vise locates the part against its fixed jaw; the moving jaw supplies the grip. Clean the table and underside of the vise before mounting it. Clean the jaw faces, vise bed, supports and part before loading. A chip trapped underneath changes the position you are trying to establish. Remove raised burrs without rounding away a locating surface.
Choose a grip that leaves access for the cutter while supporting the cut. A part perched high above shallow jaw contact is a different setup from a part seated low in the vise. For the operating rules that govern this work, use Equipment Operating Procedure and Machine Hazards and Housekeeping.
Sweep the fixed jaw, adjust, tighten, then sweep again; Interactive; Practice
Sweep the fixed jaw, adjust, tighten, then sweep again
Mount the vise with its fasteners engaged. Snug one mounting point enough to act as a pivot and leave the other adjustable.
Mount an indicator rigidly with the spindle stopped. Contact a clean, undamaged part of the fixed jaw's locating face with some preload. See indicator setup.
Choose a convenient reference near one end; zero there if it makes comparison easier. Traverse so the contact point follows the jaw toward the other end, keeping contact height and indicator orientation unchanged. Alignment comes from the differences across the sweep, whether or not you re-zero.
Read the difference, then tap the adjustable end of the vise body lightly with a suitable soft mallet. Sweep again. Use the demonstrated indicator response to choose the direction.
Tighten the mounting fasteners to the vise/mounting instructions and repeat the full sweep. Accept the setup against the job's alignment requirement, not an invented universal tolerance.
First use: apply the same small corrective rotation about different illustrative pivots. The tilt can disappear while both ends settle at different common offsets; re-zeroing afterward is optional convenience.
Rotate about a chosen illustrative pivot
Before: +0.060 mm tilt, old zero at the left
Measuring span 0–100 mm; 1 mm face marginsStored zero at 0.000 mmNext pivot P=(25.000, 0) mmAccumulated |d| 0.0000 / 0.0040 mm/mmLeft reference 0.000 mmCenter reference +0.030 mmRight reference +0.060 mmRight − left +0.060 mmModel summary — full sweep range 0.060 mm
After −0.0006 about 25 mm: both ends +0.015 mm
Measuring span 0–100 mm; 1 mm face marginsStored zero at 0.000 mmNext pivot P=(25.000, 0) mmAccumulated |d| 0.0006 / 0.0040 mm/mmLeft reference +0.015 mmCenter reference +0.015 mmRight reference +0.015 mmRight − left 0.000 mmModel summary — full sweep range 0.000 mm
Optional convenience: re-zero at 50 mm
Measuring span 0–100 mm; 1 mm face marginsStored zero at 50.000 mmNext pivot P=(25.000, 0) mmAccumulated |d| 0.0006 / 0.0040 mm/mmLeft reference 0.000 mmCenter reference 0.000 mmRight reference 0.000 mmRight − left 0.000 mmModel summary — full sweep range 0.000 mm
The rotation removes the linear tilt while the chosen pivot determines the common offset. Re-zeroing changes the displayed references, not the alignment.
Equal endpoints can hide a local high spot
Measuring span 0–100 mm; 1 mm face marginsStored zero at 0.000 mmNext pivot P=(25.000, 0) mmAccumulated |d| 0.0000 / 0.0040 mm/mmLeft reference 0.000 mmCenter reference +0.080 mmRight reference 0.000 mmRight − left 0.000 mmModel summary — full sweep range 0.080 mm
The endpoints both read 0.000 mm, while the center reads +0.080 mm. Traverse the whole jaw instead of checking only its ends.
Model: physical jaw position is a + m·x + b·4(x/L)(1 − x/L). Displayed reading = response × physical position − the raw reading stored as zero. A small illustrative rotation d about world pivot P=(p, 0) updates m to m + d and a to a − p·d; the stored zero is unchanged. The fixed 100 mm measuring span is inset 1 mm from each end of the finite modeled face. This is a first-order small-angle model with |d| ≤ 0.001 mm/mm, 0.0001 mm/mm control steps, and total rotation travel ≤ 0.004. The step is model resolution, not real-vise quantization. It does not model adjustment force. The drawing exaggerates deviation and is not a literal scale.
Worked example: over 100 mm, start with left 0.000 and right +0.060 mm. Apply d = −0.0006 about 25 mm: both ends become +0.015 mm, so the tilt is gone but the old zero reveals a common offset. About 50 mm both become +0.030 mm; about 0 mm both become 0.000 mm. Re-zeroing is optional convenience, not an alignment requirement. A real setup is accepted against the job requirement, not a universal tolerance from this model.
Alignment procedure: MIT, Introduction to the Mill, pp. 2–3. The comparison geometry and examples here are original.
Prove the sign first. Check which way the reading changes when the contact is displaced a known small amount toward the indicator body. With the response shown above, a higher right-end reading means that end is nearer the indicator: move it away. A lower right-end reading needs the opposite correction. Reverse the contact arrangement and the tap direction can reverse.
Supporting detail: Addendum 38 — Read the whole sweep, not just two endpoints.
Parallels establish height; soft jaws establish a shaped seat
Parallels establish height; soft jaws establish a shaped seat
Parallels are matched support bars placed beneath the part. Choose a height that gives cutter clearance and enough jaw engagement. Keep the planned drilling and milling paths clear of them.
Snug the part, seat it with light downward taps from a dead-blow mallet, then tighten and check the seating again. Check that both supports are contacting; a loose parallel or visible gap means the part may have lifted. A mallet cannot make a chip underneath disappear.
Soft jaws are machinable replacement jaws. A step or pocket cut into them can support a thin part or fit a contour. Machine the pocket with the jaws preloaded in the intended clamping direction, following the vise maker's method. Clean the pocket and check full seating when loading each part; the jaw shape does not excuse insufficient grip.
Put the clamp stud nearer the part than the heel support; Practice
Put the clamp stud nearer the part than the heel support
A strap clamp bridges the part and a heel support. A step clamp has a toothed heel that mates with a step block. Set the heel to the clamp maker's specified height, normally level with or slightly above the part contact. Locate the part against suitable stops and support it beneath the clamping points; a strap pressing on an unsupported span can bend the work.
Place the stud toward the part end of the strap. Keep adequate contact on the part and full seating at the heel. The stud must lie between these contacts. Plan the straps, studs and supports so the tool and holder can pass them.
A short distance from the part to the stud gives the part a larger share of the same stud load.
In the ideal two-contact model, part force = P × b / (a + b). Here P is the downward force from the nut, a is part-contact-to-stud distance, and b is stud-to-heel-contact distance. With P = 6,000 N, a = 20 mm and b = 60 mm, the part receives 4,500 N and the heel receives 1,500 N.
Supporting detail: Addendum 39 — Why the force divides this way—and what the model cannot tell you.
Choose an indicator for alignment, account for edge-finder radius, and keep machine home, work zero and tool length separate.
A test indicator sweeps with a lever; a dial indicator measures plunger travel
A test indicator sweeps with a lever; a dial indicator measures plunger travel
Instrument
Contact motion
Setup use
Test indicator
A short stylus pivots through a small angle.
Convenient for sweeping a vise jaw or a bore where access is tight.
Plunger dial indicator
The spindle, or plunger, moves straight through a fixed mounting stem into the body.
Useful for displacement checks and alignment when the plunger can face the measured motion.
Digital plunger indicator (shown below)
A spindle/plunger still moves linearly; an LCD replaces the pointer and dial.
Useful for the same kinds of displacement checks when its range, contact geometry and specifications fit the task.
Lever test indicator. The short stylus and contact point pivot at the instrument body. The post, arm and clamps are parts of a separate stand; its magnetic base is only partly visible at the frame edge. Photo: David English (Fixerdave), Wikimedia Commons, CC BY-SA 4.0. Downscaled and converted to WebP; the adapted image remains CC BY-SA 4.0. No endorsement implied.Analog plunger indicator. The body carries the dial, bezel and pointer; the straight spindle/plunger reaches the contact point. The stand is separate from the instrument, and its base is not fully shown. Photo: David English (Fixerdave), Wikimedia Commons, CC BY-SA 4.0. Downscaled and converted to WebP; the adapted image remains CC BY-SA 4.0. No endorsement implied.Digital plunger indicator. The LCD and buttons change the readout, not the contact mechanism. Below the body, the fixed mounting stem surrounds the moving spindle/plunger, which ends at the rounded contact point. No stand is shown. Photo: Huugo, Wikimedia Commons, CC0 1.0. Downscaled and converted to WebP. The manufacturer mark identifies the photographed instrument; no endorsement implied.
These displays do not define the contact mechanism: the digital instrument shown here is still a plunger indicator. Test and plunger indicators can both compare alignment. Use the maker's stylus angle for a test indicator; on a common lever design the stylus lies nearly parallel to the measured surface. Align a plunger with the displacement being measured. Incorrect contact angles change the scale factor.
Keep the mounting rigid, preload within the instrument's range and confirm repeatability. During a fixed-jaw sweep, keep the spindle stopped and the contact geometry unchanged. For a bore-centering sweep, turn the spindle by hand using the approved setup method. Follow Equipment Operating Procedure for the machine state during setup.
Supporting detail: Addendum 40 — Contact angle affects the reading.
An edge finder locates the spindle center one tip radius from the edge
An edge finder locates the spindle center one tip radius from the edge
A mechanical edge finder has a spring-loaded contact tip that becomes concentric and then kicks sideways at contact. Use its specified spindle speed and approach method. Electronic edge finders and probes use different signals and procedures; do not assume they rotate like a mechanical edge finder.
Use a clean, defined edge, repeat the approach and account for the contact-tip radius. If the tip is 10 mm in diameter, the spindle center is 5 mm outside the edge at the trigger point. Setting that center position to zero without the radius correction moves every feature by 5 mm.
Example: the left face is X = 0 and positive X goes into the part to the right. At contact from the left with a 10 mm tip, the spindle center is X = −5 mm. Enter that position in the work-coordinate setting method for the control. On approach from a right-hand face, the sign changes. Sketch the spindle center and edge before choosing a sign.
Supporting detail: Addendum 41 — Check both repeatability and the offset sign.
Find an edge, choose the sign, and check where the hole lands; Interactive
Find an edge, choose the sign, and check where the hole lands
Jog a mechanical edge finder toward any of the four edges, watch the tip wobble, centre and kick, then decide what the readout should read at that position. A check hole shows where your zero really is. The radius used is half the contact diameter, which can differ from the body diameter. Re-approach from the same direction each time; reversing a feed screw adds backlash, which this model does not include.
Wobble, centre, kick
Free: off-centre wobble
0.200 mm before the trigger: nearly concentric
At the trigger: the tip kicks sideways
The offset tip orbits while free, is pushed concentric by the edge, then kicks when the spindle centre is one contact radius from the edge. The wobble and kick are drawn for illustration, not as contact mechanics.
Where the check hole lands
Distance from the datum edge, magnified (mm); the dashed line is the drawing’s 20.000
Check holes: drawing says 20.000 mm from the left edge
Hole
Readout set at kick (mm)
Lands from edge (mm)
Off by (mm)
Your hole
−5.000
20.000
0.000
Radius forgotten
0.000
15.000
5.000 toward edge (1 r)
Wrong sign
+5.000
10.000
10.000 toward edge (2 r)
The sign depends on the side approached (+X right, +Y toward the back)
Edge
Spindle approaches moving
Edge position
Set readout at the kick (10 mm contact)
Left edge
+X
X = reading + r
−5.000 mm
Right edge
−X
X = reading − r
+5.000 mm
Front edge
+Y
Y = reading + r
−5.000 mm
Back edge
−Y
Y = reading − r
+5.000 mm
Worked example (left edge, 10 mm contact, r = 5 mm): at the kick the X readout shows +32.460 mm in its old, arbitrary frame. The edge is at reading + r = +32.460 + 5.000 = +37.460 mm. Setting the readout to −5.000 mm at that position makes the edge read 0, and a hole programmed at X 20.000 lands 20.000 mm from the edge. Setting 0 instead (radius forgotten) puts the hole at 15.000 mm: off by r. Setting +5.000 (wrong sign) puts it at 10.000 mm: off by 2r.
Model: top view; positions are the spindle centre relative to the part, whichever of them actually moves. The trigger occurs when the spindle centre is exactly one contact radius (half the contact diameter, not the body diameter) from the edge, on the approach side. Lengths are exact to 0.001 mm; jog steps are 1, 0.1 and 0.01 mm. A step that crosses the trigger leaves the readout past it by up to one step. The other axis is assumed already located. Backlash, tip wear, edge condition and speed are not modelled; use the edge finder’s specified speed.
Sources: MIT 2.670, Milling Machine, “Using an Edge Finder”, p. 6; Starrett 827MB and 827A product pages (how to use; 6 mm and 0.200 in contact diameters). The part, readings and drawings are original illustrations.
Home the machine, locate the part, then account for each tool's length
Home the machine, locate the part, then account for each tool's length
Machine home establishes the machine's reference positions. It does not locate the part.
Work zero is the origin chosen on the part or fixture. Its work offset relates that origin to the machine coordinates.
Tool length offset accounts for where a particular tool tip lies relative to the control's tool reference.
Choose the origin used by the drawing and program, locate X and Y, and establish Z using the approved touch-off or probing method. Account for any gauge or touch plate thickness. Confirm the active work offset, tool number and tool length before a trial path. Recheck after moving the part or changing how a tool is seated.
Two tools touch the same surface at different readings; Interactive
Two tools touch the same surface at different readings
Once X and Y are located, Z still depends on which tool is in the spindle. Predict the readout for two tools of different lengths touching the same part top, with ideal contact or through a block of stated thickness. CNC Axes and Offsets shows how a control applies a stored length value.
Two tools touched off on the same surface
Touch-off through a 10.000 mm block
Tool
Length, gauge line to tip (mm)
Z reading at touch-off (mm)
Length value vs T1 (mm)
T1 · 10 mm end mill (reference)
82.000
+10.000
0.000
T2 · 6 mm drill
109.500
+37.500
+27.500
Touch-off with ideal zero-thickness contact
Tool
Length, gauge line to tip (mm)
Z reading at touch-off (mm)
Length value vs T1 (mm)
T1 · 10 mm end mill (reference)
82.000
0.000
0.000
T2 · 6 mm drill
109.500
+27.500
+27.500
Worked example: the readout follows the spindle and reads T1’s tip height, with the part top at Z 0. T2 is 27.500 mm longer than T1. Touching T2 onto a 10.000 mm block, the readout shows Z = 0 + 10.000 + (109.500 − 82.000) = +37.500 mm. T2’s length value is that reading minus the block: +27.500 mm. With that value applied, the tip Z is the reading minus 27.500. Without it, a move to Z +2.000 with T2 would put its tip at 2.000 − 27.500 = −25.500 mm: 25.500 mm below the part top.
Model: Z reading at touch-off = surface Z + gauge thickness + (Ltool − LT1). Lengths run from the spindle gauge line to the tip and are illustrative. The gauge is either an ideal zero-thickness contact or a stated 10.000 mm block. Controls store and apply tool length in different ways; follow the control’s documentation for the actual offset entry.
Machining/Milling
Mill Toolholders and Collets
Match the spindle interface, compare tool-gripping mechanisms by five job criteria, and assemble ER collets with the required grip and reach.
The spindle connection and the tool grip are separate choices
The spindle connection and the tool grip are separate choices
A holder has a machine end and a cutting-tool end. “ER” describes a collet system at the tool end; it does not tell you which spindle the holder fits.
Interface
How to recognize the idea
R8
A taper seats in the spindle and a drawbar threads into the collet or holder to pull it in. A direct R8 collet grips a matching tool shank.
CAT / BT
Steep-taper holder families with flanges and retention knobs, also called pull studs, for compatible spindle retention and tool changing. Similar-looking tapers do not establish interchangeability: family, size, flange and retention knob must match the machine.
TTS
The Tormach Tooling System uses a straight holder shank gripped in a compatible collet and a shoulder that seats against the spindle face. TTS is a toolholding system, not a CAT/BT-style taper. Its seated shoulder provides a repeatable axial reference.
Identify the installed interface before selecting a holder. A Tormach name alone does not identify the spindle system or drawbar arrangement.
Supporting detail: Addendum 42 — Public toolholder references.
Compare what actually closes around the tool shank
Compare what actually closes around the tool shank
Ordinary keyed or keyless drill chuck: adjustable jaws grip drilling-tool shanks. A key or a hand-operated sleeve closes the jaws. Jacobs and Albrecht are brands, not gripping-mechanism categories. Do not assume an ordinary drill chuck or its mounting is suitable for milling side loads.
Weldon / side-lock / set-screw end mill holder: a fitted bore locates the shank and a screw bears on its designated flat. Match the bore, shank and flat. This retention feature does not guarantee that a tool cannot move under every load.
ER collet: the nut draws a split collet into a taper, closing it around the shank. Match the series and stated gripping range. A direct R8 collet is a different system.
Shrink fit: controlled heating expands the holder bore; cooling around the matching shank produces an interference grip. Tool changes need the compatible heating/cooling equipment and the maker's process.
Hydraulic expansion chuck: an actuator pressurizes fluid inside the holder; a sleeve deforms against the shank. The actuating screw pressurizes the system rather than bearing directly on a Weldon flat.
Mechanical milling chuck: a mechanical drive compresses the gripping bore or a special collet. Designs differ: some use a nut and needle bearings; Albrecht APC uses gearing and a tapered collet. APC is a milling-chuck product, distinct from Albrecht's jaw drill chucks.
REGO-FIX powRgrip: a dedicated unit presses a matching tapered collet into the holder. The retained grip is a mechanical press fit, not an ER nut, thermal shrink fit or hydraulic-expansion sleeve.
Grip the intended shank surface, not the cutting flutes. Keep the tool's specified engagement in the holder while leaving only the reach needed to cut and clear the part.
Different force paths can produce similar-looking holders. Schematics identify mechanisms, not assembly instructions, dimensions or performance rankings.
Supporting detail: Addendum 43 — Manufacturer examples and the drill-chuck exception.
Choose against runout, balance, grip, clearance and total system cost
Choose against runout, balance, grip, clearance and total system cost
Runout concerns the tool's rotating geometry; balance concerns the distribution of rotating mass. Grip must resist the required turning and axial loads. Clearance and reach must work throughout the path. Cost includes the holder, collets or sleeves, setup equipment, inspection, maintenance and change time.
Use this matrix to shortlist actual products. Each cell names a useful check, not a universal quality grade or approval to run at a particular speed. On a narrow screen, scroll the table sideways; keyboard users can focus it and use the arrow keys.
Compare the assembled setup, not just the holder family
Mechanism
Runout / high-RPM balance
Grip / clearance and reach
System cost and upkeep
Ordinary drill chuck
Check jaws, mounting and the exact speed rating.
Drilling use; body and jaws need clearance.
Adjustable diameter coverage; inspect jaws and mounting.
Side-lock
Verify the screw/flat assembly's runout and balance.
Collets cover sizes; correct nut, wrench and torque equipment.
Shrink fit
Check the assembled tool, not concentric appearance.
Slim-nose options; verify shank match and engagement.
Heating/cooling equipment, compatible tooling and change cycle.
Hydraulic
Check the sleeve's condition and product specifications.
Check the exact milling application and engagement limits.
Approved actuation and sleeves; inspection/service requirements.
Mechanical milling chuck
Clamping configuration can change grip and runout.
Roughing candidates; compare the actual body/nut envelope.
Maker's wrench and approved parts; mechanism-specific care.
Pressed collet / powRgrip
Check the matched collet, holder and complete assembly.
Correct collet diameter; compare actual nose profiles.
Matching collets and a dedicated press-in/out unit.
Supporting detail: Addendum 44 — A real trade-off within one product family.
Snap the ER collet into its nut before threading the nut onto the holder
Snap the ER collet into its nut before threading the nut onto the holder
Match the ER series, nut, collet and tool shank. Read the collet's actual range; the ER series number is not the tool diameter. Do not force an oversized shank into it or collapse it below its rated range.
Clean and inspect the holder socket, collet, nut and tool shank. Follow the manufacturer's lubrication instructions for the named surfaces; do not treat every taper and gripping surface alike.
Engage the collet's extraction groove with the nut's eccentric retaining ring. Tilt and snap it into place as instructed by the maker.
Thread the nut-and-collet assembly loosely onto the holder.
Insert the tool with the required shank engagement and planned stickout. Tighten to the specified torque using the proper wrench and holding fixture.
The nut captures the collet before either goes on the holder. Do not place the collet in the holder first and try to force the nut over it.
Range example: a collet explicitly rated 9–10 mm can grip a 9.5 mm shank; it cannot grip a 10.2 mm shank. This is a reading exercise, not a universal collapse range for every ER collet.
Supporting detail: Addendum 45 — Why the snap matters.
A low runout reading does not establish high-speed balance
A low runout reading does not establish high-speed balance
Runout is the variation seen as a tool or reference shank rotates about the spindle axis. If one cutting edge projects farther than the others, it can carry more of the load. Dirt, damaged seating surfaces, a poor collet fit, the tool, holder or spindle can contribute. Clean and reseat before blaming cutting data; measure against the job and tool requirements.
A geometrically straight assembly can still have uneven mass around its axis. A holder's balance marking may apply without the tool, collet or other fitted parts. Check the complete rotating assembly, the stated measurement conditions and every applicable operating-speed limit. Balancing does not raise a component's permitted speed.
Supporting detail: Addendum 46 — Why adding the tool can change the answer.
Measure the reach you need without losing the required grip
Measure the reach you need without losing the required grip
Physical tool stickout is the unsupported tool length beyond the effective grip. For the bending model, identify the clamp-to-load distance. Holder projection or gauge length refers to dimensions on the holder drawing, often measured from the spindle's gauge reference to a specified holder feature. Flute length is the length of the tool's cutting portion. These are different measurements; read the endpoints on the actual drawing.
A longer holder can move the grip closer to a deep feature while the cutter keeps adequate engagement. It also adds holder projection to the system, so check stiffness and the complete collision envelope. Do not gain reach by pulling a tool out below its required clamping length. Explore Tool Stickout and Deflection for unsupported length and Milling Chatter for the wider cutting system.
Practice choosing the governing constraint before naming a holder
Practice choosing the governing constraint before naming a holder
Shallow roughing: start with a milling-rated assembly's torque transfer and axial retention. A matching side-lock holder is one candidate when the tool has the designated flat; compare other approved systems against the job rather than declaring one family the strongest.
Deep, narrow feature: compare nose/body clearance and required tool reach. A slim holder may avoid a collision that a large collet nut would create. Verify engagement and the holder's projection as well as the exposed cutter.
High-speed finishing with a small cutter: prioritize measured runout, assembly balance and the actual speed limits. A holder family name, shiny finish or high gripping-force claim settles none of those checks.
Supporting detail: Addendum 47 — Check your selection reasoning without a video.
Use short Haas video stops to test the mechanism comparisons
Use short Haas video stops to test the mechanism comparisons
The native explanations and prompts below work without video or network access. The links open selected moments in two Haas videos; they are study references, not this shop's operating procedures or equipment inventory.
Tool Holder Essentials — compare the choices
1:21 — the five criteria. Before watching, choose which criterion governs each of the three situations above. Check: the job determines the priority; one overall winner is not established.
15:50 — torque guidance. Identify where the correct value comes from. Check: the manufacturer of the actual nut, collet and holder combination, not a number generalized from a demonstration.
Machining/Milling
End Mill Anatomy and Cutter Selection
Read end mill geometry, compare flute counts and tool materials, and choose a cutter for the material and operation.
End Mill Anatomy
End Mill Anatomy
An end mill has cutting teeth separated by recessed flutes. Read the side and the cutting end together: the flute hollow, the cutting edge and the relieved face behind it are different features.
Oblique view: the flutes blend out into a plain shank. This representative tool has no reduced neck.
Side view: OAL 76 mm, functional LOC 26 mm; the runout finishes at 33 mm. Runout is not extra functional cutting length.
End detail: two long lips and two short lips meet real recessed gash pockets. This is the cutting end, not a section through the fluted body.
Original Blender-authored representative geometry. Use the lab below to orbit the same model, locate the named features and inspect the end. These are teaching dimensions, not a manufacturer part specification or setup recommendation.
Flutes are grooves beside the teeth that provide room for chips. The sharp edges bordering them do the cutting.
Helix angle describes the spiral relative to the tool axis. It affects how the edge enters the cut and the direction of cutting forces; it is not the spindle's rotation direction.
Cutting diameter is the diameter swept by the cutting edges. Shank diameter fits the holder and may differ from cutting diameter.
Overall length (OAL) runs from tool tip to shank end. Length of cut (LOC) is the axial cutting-edge length, not permission to cut that entire depth in one pass.
Length below shank (LBS) runs from the start of a reduced neck to the tip, so it includes both relieved neck and LOC. It is not the neck-only length.
Square end makes a flat-bottom form with a sharp nominal corner. A corner-radius or bull-nose end rounds that corner. A ball end has a rounded tip for curved surfaces.
Center-cutting geometry can cut at the tool center, allowing axial entry when the manufacturer permits it. Flute count alone does not prove a tool can plunge; check its entry method and limits.
Peripheral and end cutting edges
The sharp intersections of the rake and relieved faces cut on the side and the end respectively. An edge is not the same thing as the whole flute groove.
Rake face
The face ahead of the cutting edge, along which the newly formed chip flows.
Primary relief and land
The narrow ground surface just behind an edge clears the newly cut surface. Further relief behind it provides additional clearance; the exact grind varies with the tool.
Flute or gullet
The hollow beside a tooth provides chip space. The material between the flutes forms the teeth and the central core.
End gash and end relief
A gash is a recessed pocket at the end near the lips; the relieved face behind an end edge provides clearance. Neither is a decorative line drawn on a flat cap.
Core, runout and reduced neck
The core is the central material remaining between the flutes. Runout is where a groove blends into the shank. A reduced neck is a separate relieved reach region; the displayed four-flute model does not have one.
Helix and radial pitch
The helix angle is measured relative to the tool axis on its cylindrical surface. Radial pitch is angular spacing between teeth: four equally spaced flutes give 90°. These describe different directions.
Holder-face projection is the mounted distance from holder face to tip. Pulling the same tool farther out changes projection and insertion, not OAL, LOC or LBS. Follow the tool and holder makers' gripping requirements. Use Mill Toolholders and Collets for the holder system, Tool Stickout and Deflection for the cost of extra reach and Milling Operations for the shapes these cutters produce.
Supporting detail: Addendum 48 — Public anatomy reference.
End Mill Anatomy Lab; Interactive
End Mill Anatomy Lab
Approved fixed model: 12 mm diameter, four flutes, 30° helix, LOC 26 mm and OAL 76 mm; no reduced neck. The same geometry is shown in the live viewer.
Side and oblique surfaces show the flute hollows and cutting teeth.
End gashes and relieved faces are real geometry, not markings on a flat cap.
Separate catalog comparison: reach and mounting
The following two schematic catalog examples differ from the fixed 3D model. Their 100 mm OAL, LOC and LBS stay fixed when the holder moves.
First use: compare one reduced-neck tool mounted at 75 and 85 mm projection. Its OAL, LOC and LBS do not stretch when the mounting changes.
LBS includes LOC plus any relieved neck. It is neither neck-only length nor holder projection.
Model boundary: the approved fixed mesh represents a square-end cutter with end gashes and relief. The separate catalog sketches are dimension illustrations, not manufacturing CAD or proof of a permitted plunge, cutting depth or grip.
The 3D lab uses the approved fixed four-flute model. The separate static catalog comparisons explain reduced-neck and long-flute examples; they are not selectable shapes of this GLB. A reduced neck provides clearance, not proof of strength or permission to use a particular depth or holder projection.
Choose flute count for chip space and engagement, not material alone
Choose flute count for chip space and engagement, not material alone
Flutes
Useful starting point
Check before choosing
2
Generous chip space, often useful for aluminum slots.
Geometry, edge strength and the maker's slotting recommendation.
3
Common aluminum-specific choice balancing chip space with another cutting edge.
That the tool is designed for the material and engagement; three flutes alone are not a material specification.
4
Common for steel and for cuts where chip space is less restrictive.
A full-width slot can pack the smaller flute spaces. Do not assume more teeth always improve removal rate.
A side cut and a full-width slot load the same cutter differently. Select the actual tool's data for the operation, then use Feeds and Speeds for the calculations and Climb and Conventional Milling for feed direction.
Supporting detail: Addendum 49 — Why the two-for-aluminum, four-for-steel rule is only a starting point.
Match HSS, carbide and coating to the job and machine; Practice
Match HSS, carbide and coating to the job and machine
High-speed steel (HSS) is relatively tough and often economical. It can be a useful choice at lower cutting speeds, especially where the setup cannot take full advantage of carbide. Cobalt HSS is an HSS family with improved hot hardness.
Cemented carbide retains hardness at higher cutting temperatures and is stiffer than HSS, but is less tolerant of impact and poor support. A carbide tool does not make a loose setup rigid. Select its grade and geometry as well as its material.
Coatings change wear, friction and heat behavior. TiN and TiCN are familiar options; AlTiN/TiAlN are common on tools for hot-running ferrous applications. For aluminum, choose an uncoated polished tool or a coating explicitly recommended for that alloy, such as TiB2 where specified. Color is not a reliable selection method, and no coating guarantees a fixed multiple of tool life.
Job examples: for an aluminum slot, start by comparing aluminum-specific two- and three-flute tools with enough chip space. For steel side milling on a rigid setup, a steel-rated four-flute carbide tool may be a good starting point. For a modest-speed, interrupted job, a suitable HSS tool may be more forgiving. Check the manufacturer's recommendations before selecting the cutting data.
Supporting detail: Addendum 50 — Check the coating's material recommendation.
Face mills and fly cutters cover broad flat surfaces
Face mills and fly cutters cover broad flat surfaces
A face mill uses several cutting edges, often replaceable inserts, to machine a broad face. Match its diameter, insert geometry, holder and power demand to the machine and part.
A fly cutter carries a single cutting point that sweeps a circle. It can make a broad finishing pass, but its single edge and interrupted loading need suitable rigidity, balance and cutting data. It is not simply a face mill that can use the same feed setting.
What milling chatter is, why it damages finish, accuracy and tools, and how to diagnose the force path before changing one permitted variable at a time.
Chatter is a self-reinforcing vibration during cutting
Chatter is a self-reinforcing vibration during cutting
Milling chatter is vibration that grows through the cutting
process instead of simply following a steady outside disturbance. The changing motion changes
chip thickness and cutting force, which can feed more energy into the motion. Left uncontrolled,
it can spoil surface finish and dimensional accuracy, increase tool deflection and accelerate
cutting-edge wear or failure.
A common mechanism, regenerative chatter, is self-excited: one cutting edge leaves a waviness that changes the chip thickness seen by the next edge, and the changing force can reinforce the motion. That mechanism does not describe every vibration or finish defect. A repeating surface pattern, a loud cut or another forced vibration can be clues; they do not prove regenerative chatter by themselves.
Supporting detail: Addendum 51 — Several faults can look or sound like chatter.
Stop the cut before inspecting or changing the setup; Watch out
Stop the cut before inspecting or changing the setup
Do not tune around a part, tool or holder that moved. A shifted workpiece, incomplete seating, damaged edge or visibly slipping tool is a mechanical problem to correct or report before another cutting trial.
Record the evidence before trying another cut
Record the evidence before trying another cut
Record where the symptom began, where it appears on the part, which operation and tool were active, and the actual speed, feed, engagement and holder projection. This creates a comparison for the next check instead of relying on sound or memory.
Trace force through the complete cutting system
Trace force through the complete cutting system
The cut closes the loop. A stiff cutter cannot compensate for a moving part, and rigid workholding cannot correct a damaged holder interface.
Route suspected machine or spindle faults to authorized staff
Route suspected machine or spindle faults to authorized staff
After the cutter, holder, workholding, path and chips have been checked, a persistent symptom is evidence worth giving to authorized staff. Abnormal motion or a visibly damaged or contaminated interface also calls for that route. It does not prove one component failed. Do not remove covers, adjust bearings or gibs, or service a drawbar; manufacturer guidance treats those as separate checks for trained personnel.
Check the holder, collet and cutter as one assembly
Check the holder, collet and cutter as one assembly
Confirm the correct holder and collet, complete seating, clean contact surfaces under the approved procedure, adequate shank engagement, and no visible damage. Compare tool and holder projection with the last stable setup. Inspect for wear, chipping, built-up material and chip damage. Check runout only with the approved instrument and method; compare it with the tool, holder and job requirements. See Mill Toolholders and Collets, End Mill Anatomy and Cutter Selection and Tool Stickout and Deflection.
Check the fixture, workpiece and support
Check the fixture, workpiece and support
Check that the fixture is seated and secured, the work is against its locators, clamps direct force into supports, and the part is not perched or springing. A symptom confined to an unsupported end, thin wall or one clamping orientation points first toward support and force direction. Do not keep tightening until the part distorts. See Milling Vises and Strap Clamps.
Check engagement, path and chip evacuation
Check engagement, path and chip evacuation
Compare where the pattern begins with changes in radial or axial engagement, entry and exit, cornering, cutter engagement, and chip evacuation. A clean straight wall followed by chatter in an inside corner is evidence to inspect the corner path and the sudden engagement change. Packed or re-cut chips call for the approved chip-control method and a review of the operation, not an automatic speed change. See Milling Operations and Cutting Fluids and Application Methods.
It changed with a longer tool or holder
It changed with a longer tool or holder
Check: tool projection, holder projection, shank engagement and both interfaces. Next: restore the shortest permitted assembly that reaches the feature, then repeat the same cut conditions.
Straight walls are clean; an inside corner chatters
Straight walls are clean; an inside corner chatters
Check: the engaged arc, chip load through the corner and programmed path. Next: use an approved path or engagement change, then compare the same corner again.
Only an unsupported or thin area vibrates
Only an unsupported or thin area vibrates
Check: workpiece support, clamp-force path and local flexibility. Next: improve support or revise force direction without distorting the work; verify the setup before cutting.
One edge is chipped or marks repeat once per revolution
One edge is chipped or marks repeat once per revolution
Check: edge condition, chip re-cutting, holder seating and runout. Next: correct the physical cause or replace the damaged tool before tuning speed.
It persists after the student-controlled setup checks; Practice
It persists after the student-controlled setup checks
Check: after the cutter, holder and workholding are verified, ask whether the symptom still follows the same path, engagement or chip condition. Next: if it persists across different permitted cutting conditions, stop and give the complete evidence to authorized staff. It is evidence for inspection, not proof of a spindle fault.
Clear mechanical causes and unnecessary leverage first
Clear mechanical causes and unnecessary leverage first
Clear mechanical causes first. Verify seating, condition, projection, support and chip clearance. Do not use a parameter change to hide motion or damage.
Reduce unnecessary leverage. Shorten the permitted tool and holder projection, improve support, or choose a more suitable assembly. Keep the required shank engagement and feature clearance.
Review engagement and cutter geometry after the setup is sound
Review engagement and cutter geometry after the setup is sound
Review engagement and force direction. A smaller radial or axial engagement, different entry, or revised corner path may reduce the force variation. Stay inside the approved process and tool limits.
Reconsider cutter geometry. Diameter, flute count and pitch, helix, edge geometry and material suitability can change force and excitation. Record the selected tool and why it was changed.
Run one permitted speed trial
Run one permitted speed trial
A stable speed can be higher or lower. Change to one permitted speed, then compare the same feature. Do not assume that slowing down is always the answer.
Review feed when speed changes
Review feed when speed changes
If the trial is intended to preserve feed per tooth, calculate the matching table feed rather than changing chip load accidentally. Too little feed can rub; excessive feed can overload the edge. Use Feeds and Speeds and the tool maker's data.
Record the controlled result
Record the controlled result
Record the observation, the one intentional change and the result. If two settings move together to preserve a stated quantity such as feed per tooth, record both and the quantity held constant. That is one controlled trial, not two unexplained changes.
Manufacturer references behind this diagnostic order
Manufacturer references behind this diagnostic order
A Tormach-based sequence for preparing the machine, locating the part and tools, checking a path, and closing down.
Use the checklist with the installed machine's procedure
Use the checklist with the installed machine's procedure
This checklist follows a Tormach vertical-mill workflow. Match each control, inspection method and tool-change step to the installed machine and controller. It is a setup reminder for trained users, not a substitute for instruction or the machine manual.
Confirm the machine and controller have power available. Turn on the controller computer, then the mill as directed by the installed-machine procedure.
Reset depressed emergency-stop controls after completing the prescribed checks, then use the machine's Start/enable control to activate its motor controllers.
When the controller is ready, reference (home) all installed axes by the machine's procedure. Confirm that referencing completed; homing has not set the part zero.
Part and tool setup: seat, grip, locate and record; Checklist
Part and tool setup: seat, grip, locate and record
Plan the datum faces, work zero, tool access and fixture clearance for the intended operations.
For a vise setup, clean and inspect the vise and part, remove raised burrs, and check the fixed-jaw alignment. See Milling Vises and Strap Clamps.
Fit parallels if needed. Snug the part, seat it with light dead-blow taps, tighten the vise and recheck seating.
For table clamping, have the support, locating and clamp arrangement checked against the job and installed procedure; confirm every intended contact is seated and every clamp is firm. Use strap-clamp geometry to review placement.
Choose the cutter and matching holder. For ER tooling, capture the collet in its nut, thread the assembly loosely, insert the shank with the planned reach, and tighten as specified. For a set-screw holder, align the designated shank flat and follow its seating method. See Mill Toolholders and Collets.
Install the holder using the procedure for the actual spindle and drawbar. For a manual drawbar arrangement, verify completion of its spindle-lock release and spindle-cover closure steps before operation; do not apply that sequence to a different retention system.
Set and verify tool lengths and the work offset using Mill Indicators and Part Zero. Confirm the active tool number, units and work coordinate system.
Before cutting: verify the program or manual movement plan; Checklist
Before cutting: verify the program or manual movement plan
Confirm cutter, spindle direction, speed, feed, width and depth of cut against the tool data and Feeds and Speeds.
For programmed work, load the intended program and check the displayed path, tools, offsets and limits. See G-code and Toolpaths.
Perform the approved air-cut or dry-run procedure with verified clearance. If that method changes offsets or modes, restore and verify them before cutting the part.
For manual movement, confirm the selected axis, direction, jog rate and increment. Use a low approach rate near a surface and incremental motion for fine positioning, as directed in the machine procedure. Select cutting feed from the tool data rather than treating a low approach rate as a cutting-feed prescription.
Complete the cutting-readiness checks in Equipment Operating Procedure, then perform the operation and its planned measurements.
Shutdown: end operation, shut down the controller and restore the work area; Checklist
Shutdown: end operation, shut down the controller and restore the work area
Follow the installed Tormach shutdown sequence. In this workflow: depress emergency stop, shut down the controller computer normally, then remove its power and turn off the mill.
Use the machine-specific instructions for any required parking position, tool removal, lubrication or power sequence. These details vary with the machine and its options.
Machining/Milling
Milling Operations
Make a part on a simulated vertical mill: choose each tool, move it through the metal, and check what is left against the drawing. Facing, shoulders, pockets, slots, drilled, tapped and bored holes, countersinks and chamfers.
A vertical mill shapes a block with a turning tool moved in X, Y and Z
A vertical mill shapes a block with a turning tool moved in X, Y and Z
On a vertical mill the cutting tool turns in a spindle above the work. The work is
clamped on the table, the table moves it front to back and side to side, and the tool comes down
into it. Every shape the mill makes is the space the turning tool swept through.
X and Y
The two directions across the table. In this lesson X
runs left to right as you face the machine and Y runs away from you, toward the
back.
Z
Up and down, along the tool. Z+ is up, away from the work, so
negative Z is below the finished top face. The raw top is at Z +1, so facing can remove
metal at zero or positive Z while taking off that allowance.
Readout
The display that shows where the tool is, in X, Y and Z. A readout on a
manual mill is often called a DRO, for digital readout. Here it shows the position of
the tool’s tip.
Work zero
The point the readout counts from. In this lesson X 0, Y 0 is
the front left corner of the block and Z 0 is the finished top face. Every dimension on the
drawings is measured from those same edges and that same face.
The setup used in every job: a block 100 × 60, finished 20 thick, held by its
front and back faces in a vise and resting on two parallels, bars of equal height
that lift it off the floor of the vise. The raw block stands 1 mm proud of the finished top. Below
Z −8 the vise jaws are in the way at the front and back, and a hole drilled through
along those edges would run into a parallel.
On a real knee mill the table carries the work under a spindle that stays put in X and Y. In
the simulator you move the tool instead. The readout tells the same story either way: where the
tool’s tip is, measured from work zero.
Each kind of tool cuts a shape of its own, and some cannot cut sideways
Each kind of tool cuts a shape of its own, and some cannot cut sideways
The tools in the simulator, side on and at one scale, each with its tip on the same
line. The darker part of each is the part that cuts.
Face mill
A wide body with cutting inserts round its rim. It is fed
across the work to leave a broad flat face. The one in the simulator takes at most
4 mm in one pass. The inserts sit at the rim, not across the middle, so it is brought down beside the work
and then fed in, never pushed straight down into the metal.
End mill
Cuts on its sides and its end. The end mills here are
centre-cutting: their end teeth reach the middle, so they can be fed straight
down into the metal and then moved across to cut pockets, slots, steps and walls. An inside
corner they leave is rounded to the cutter’s own radius.
Spot and chamfer tool
A pointed tool with a 90° point. Fed down, it cuts a
cone: the start of a hole, or a countersink round the mouth of one. Run along an
edge, it bevels the edge, which is a chamfer.
Drill
Cuts only at its point, as it feeds down along its own axis. It leaves a
round hole its own size, with a cone at the bottom from its point. It must never be pushed
sideways in the metal.
Tap
Cuts a screw thread into the wall of a hole that is already there. It only
goes straight in and straight back out.
Boring head
An adjustable single cutting point on an offset bar. Set to a
diameter and fed down a hole, it opens the hole to that size. Its bar has no cutting edge at the
centre, so it needs a hole to start in.
This is a teaching rack, not a list of what to buy or use. The sizes are
illustrative. A real cutter’s length, number of teeth and depth of cut come from its own
data and from the person responsible for the machine.
The tool’s centre runs one radius away from the wall it cuts
The tool’s centre runs one radius away from the wall it cuts
The readout shows where the centre of the tool is. The metal is cut by its
edge. To leave a wall at a given place, put the centre one radius away from it, on
the side being cut away.
Worked example. A shoulder’s wall belongs at X 75, with the metal to
the right of it cut away. A Ø12 end mill has a 6 mm radius, so its centre runs along
X 75 + 6 = X 81. Run it along X 75 instead and the
wall ends up at X 69, 6 mm too far into the part. That metal cannot be put back.
Inside corners take the cutter’s radius. A round tool cannot reach into a
sharp inside corner, so a pocket’s corners come out rounded to the radius of the cutter that
made them. A drawing that allows R4 at most calls for a cutter no bigger than Ø8.
Outside corners stay sharp. Going round the outside of a corner, the tool
swings about it and never cuts into it.
A slot is the width of the cutter that made it in one pass. A cutter moved in a
straight line leaves a groove exactly its own width, with round ends of its own radius.
When a cutter cuts with its side, the metal is on one side of it or the other. With the
cutter turning clockwise seen from above, metal on the cutter’s right as it moves is
climb milling, and metal on its left is conventional milling. The
simulator says which each side cut was. Climb and Conventional Milling explains the difference.
Holes: drilled to size, threaded by a tap, opened up by a boring head
Holes: drilled to size, threaded by a tap, opened up by a boring head
A drill leaves a hole of its own diameter. A hole’s depth on a drawing is measured to where
the hole stops being full size, not to the tip of the cone the drill point leaves below that.
A tapped hole starts as a smaller drilled hole
A tap cuts its thread into the wall of a hole, so the hole has to be drilled smaller than the
thread first, to the tap drill size. For the common metric coarse threads that
size is close to the thread’s outside diameter minus its pitch, the distance from one thread
to the next:
Drilled too small, the tap has to cut too much metal and can break. Drilled too big, the thread is
shallow. The simulator checks the hole against a simplified minor-diameter band and refuses
a severely undersize pilot. This does not establish functional thread fit: pitch diameter,
thread form and a real thread gauge are separate checks.
Drill deeper than the thread you need
The first few threads of a tap are ground away to a taper so that it can start in the hole. Those
threads cut only part of a thread, so full thread stops short of the tap’s tip. In a hole
with a bottom, the tap has to go deeper than the full thread you need, and the drilled hole deeper
again, past the drill’s point. This simulated tap has a 4-pitch cutting chamfer; real taps
vary. The job calls for full thread reaching Z −10 from the finished face. Its
countersink removes the first part of the thread, so that is not 10 mm of surviving full-thread length.
Sizes no drill gives
Drills come in fixed sizes. A boring head is set to the size you need and fed down a hole already
big enough for its bar, which makes it the tool for a hole of an exact, larger size.
Simulator: make the part on the drawing, then inspect it; Interactive
Simulator: make the part on the drawing, then inspect it
Job 1: Face the top, then cut a shoulder. Face the top flat at Z 0.00, then mill a shoulder 25 mm wide and 6 mm deep across the right-hand end.Job 2: A pocket with tight corners, and two clearance holes. Face the top, mill a 40 × 24 pocket 6 deep whose inside corners are no rounder than R4, and drill two Ø6.6 holes right through.Job 3: A slot and two tapped holes. Face the top, mill a slot 10 wide with full-round ends, and make two countersunk M6×1 holes whose full thread reaches Z −10 or deeper, measured from the finished face.Job 4: A boss with a bored hole and chamfered edges. Face the top, mill away everything round a 50 × 30 boss so it stands 5 high, bore a Ø20.00 hole through it, and chamfer its top edges 1 × 45°.
Start by examining the finished Target, then choose a tool and cut Your work.
In Compare, gold is metal still to remove and striped red-orange is metal cut past the
drawing. The inspection measures the part that is left
after every move. Nothing checks which tools you used or in what order: any way of making the right
part passes.
How to drive it.Target shows the finished nominal part;
Your work shows the block you are making. Compare marks what still needs work,
using the drawing’s allowed limits. In Inspect / orbit, dragging and the arrow keys
move only the camera. Home restores the view; the View list also offers top, front,
below and 2D sections.
Choose a tool and its size, then choose Feed X, Feed Y or Feed Z
to drag its tip along that machine axis. The path previews while you drag; releasing starts the
feed. Place XY places the centre at the current Z. In these feed modes the arrow keys
move the selected machine axis. To go to an exact place, type X, Y and Z under Move to and press Move; the
tool goes up before it goes across, and across before it goes down. The handwheel buttons and the
selected-axis keys move one step at a time. A move that would break a tool or hit the vise stops where it
would happen, and says why. Undo takes back the last move: the one thing a real mill
cannot do.
Things to try in the simulator; Checklist
Things to try in the simulator
In job 1, load the face mill, move it over the block and try to lower it straight onto the
metal. Read why it stops.
Lower the face mill to Z 0.00 beside the block and drag it across. Watch the top view and
the sections: the raw top is 1 mm above Z 0. Overlap a second pass until the top face
check matches.
Cut the shoulder with the Ø12 end mill with its centre on X 75 instead of X 81.
Read the inspection, then use Undo and cut it again in the right place.
In job 2, clear the pocket with the Ø12 end mill and read what the inspection says about
its corners. Undo, and clear it with a smaller cutter.
Watch the gold as you cut: it is metal the drawing does not keep. Cut one pass a little too
deep and see where the striped red-orange appears in the top view and the section.
In job 3, try to tap a hole you drilled with the Ø6.0 drill, then one you drilled with the
Ø4.2. Then drill Ø5.0 only 10 mm deep and tap it as deep as it will go. How much
full thread do you get?
In job 4, try the boring head before there is any hole. Then drill one first.
Make one part two different ways, for example job 4 with the face mill doing the area round the
boss, and again with an end mill. Both pass.
What the simulator leaves out
What the simulator leaves out
The 3D target is a nominal guide sampled at 0.25 mm in X and Y. Its flat floors have the
drawing’s depth; its small curved outlines are approximations. The cut surface is derived
separately from the analytic tool paths. Neither picture is a dimensional gauge: inspection uses
the drawing’s allowed limits, including smaller corners where the drawing says R4 max.
Complex conical histories or very long cut histories switch to 2D sections to keep the controls
responsive. The work is kept, and Undo or a fresh block can restore the 3D view.
The simulator is a model of where metal is and where tools can reach. It does not check speeds,
feeds, forces or finish, and passing a job here says nothing about whether a real setup could take
the same cuts. Feeds and Speeds is about those numbers.
Only straight down and straight across. The tool axis is always vertical and
nothing is undercut, so every shape is a height above each point of the table.
Tools are exact and rigid. A hole is exactly the drill’s size, a cutter
never bends or wears, and edges come out without burrs.
Threads are recorded, not drawn. The simulator keeps the size of the hole a
thread was cut in and how deep its full thread runs, and checks that later cuts have not taken the
metal it is in. It does not model the thread’s form or fit.
No chips, coolant or heat, and no limit on depth of cut except where a
tool’s cutting edges run out.
Supporting detail: Addendum 52 — How the inspection decides, and what its numbers are.
Machining/Milling
G-code and Toolpaths
How a CNC mill reads a program line by line: rapid and feed moves, absolute and incremental coordinates, units, arcs and a drilling cycle. Build a program from blocks and watch it cut a simulated block.
What is G-code?
What is G-code?
A CNC mill moves its tool by reading a program from the top down, one line at a
time. The language most mills read is called G-code. Each line holds a few
words, and together they say where the tool goes next, how fast, and what the spindle and
tool are doing.
Start with CNC Control Fundamentals for the control and machine types,
CNC Axes and Offsets for coordinate references, and CNC Program Structure for common
Fanuc-style milling codes. G- and M-code meanings vary by control. The simulator below uses
its own documented subset; check the machine's manual for real programs.
Word
A letter and a number, such as X20, G1 or
F300. The letter says what the number means: X, Y and
Z are positions, F is the feed rate, S the spindle speed,
T a tool number. G and M words are codes that choose what
kind of move or machine action the line is.
Block
One line of the program. Machinists often call a line a block; in
the simulator below, each block you snap into place is one line of G-code, with the same
number.
Work zero
The point the program’s numbers count from. Here, as in
Milling Operations, X 0, Y 0 is the front left corner of the block and Z 0 is its
finished top face. The uncut stock starts at Z +1. Positions refer to the tip of the tool,
on its centre line: Z −3 is 3 mm below work zero and 4 mm below that raw top.
A pointed drill reaches its programmed Z at its tip; the full-diameter hole is shallower.
Modal
Most codes stay in force until another code from their group replaces
them. After a G1, a line with only X80 is still a G1 move; a
feed rate set with F remains available until changed. Not every word is modal.
For ordinary G0/G1 moves, an omitted axis stays in place; it does not repeat the preceding
incremental displacement.
A first cut: one slot
(A first cut: a closed-ended slot)
G21 G90 G17
T1 M6
S2500 M3
G0 X10 Y30
G0 Z2
G1 Z-3 F100
G1 X90 F300
G0 Z20
M5
M30
A first cut: one slot. Mill a closed-ended slot 10 mm wide to Z−3 with a Ø10 end mill. The raw top is Z+1, so the cut is 4 mm deep. Lift clear and stop. The program on the left, numbered line by line, and the block after it has run. Solid lines are feed moves, dashed lines are rapids; the darker the cut, the deeper it goes.
Read the program above from the top. Line 2 sets millimetres, absolute positions and the XY
plane. Line 3 puts tool 1, a Ø10 end mill, in the spindle, and line 4 starts the
spindle at 2500 rpm. Lines 5 and 6 move the tool through the air to just above the start of
the slot. Line 7 feeds it to Z −3, cutting 4 mm below the raw top at
Z +1. Line 8 feeds it along the closed-ended slot, and
line 9 lifts it clear. Line 10 stops the spindle and line 11 ends the program.
Rapid moves put the tool in place through the air; feed moves cut
Rapid moves put the tool in place through the air; feed moves cut
G0 is a rapid move. The control uses its rapid positioning rate, subject to machine limits and applicable overrides. It is
for getting the tool to where a cut starts, through the air, never for cutting.
G1 is a feed move. The tool moves in a straight line at the feed
rate set by F, here in millimetres per minute. It can cut or feed through air.
The spindle has to be turning first.S sets its speed in
revolutions per minute, M3 starts it turning clockwise seen from above, and
M5 stops it. T1 M6 changes to tool 1.
A move goes in one straight line.G1 X80 Y40 Z-2 moves all three
axes together and they all arrive at once. An axis left out of the line stays where it is:
G1 X80 moves along X only.
A rapid into metal is a crash: the tool arrives at full traverse speed with nothing to say it
should be cutting. The simulator stops a rapid at the point where it would touch the block and
says so. It does the same if a feed move reaches the metal with the spindle stopped.
This simulator moves a rapid in one straight line, the way the
LinuxCNC G-code reference
describes G0. Controls differ in how they move during a rapid, so check how the one you
use moves before you count on the path a rapid takes.
G90 and G91, G20 and G21: two modes that change what every number means
G90 and G91, G20 and G21: two modes that change what every number means
G90, absolute
A number is a position, measured from work zero.
With the tool at X 10, G1 X30 moves it to X 30.
G91, incremental
A number is a distance from where the tool is
now. With the tool at X 10, the same G1 X30 moves it 30 mm, to X 40.
G21, millimetres
X30 means 30 mm and F300 means
300 mm per minute with the simulator's G94 feed mode.
G20, inches
The same X30 means 30 inches, which is 762 mm,
and, with G94, the feed rate is in inches per minute. A millimetre program read in inches asks for moves 25.4
times as long.
Because these modes change the meaning of every line after them, a program sets both near its
start, before the first move. The simulator insists on it: a move before G20 or
G21 and G90 or G91 have been set stops the program, and so
does a change of units after the machine has started moving.
Work through a point list in Absolute and Incremental Positioning. Units and feed mode
are separate settings; G95, where supported, uses feed per spindle revolution.
G2 turns clockwise and G3 counterclockwise, seen from above; Practice
G2 turns clockwise and G3 counterclockwise, seen from above
An arc move cuts along part of a circle in the XY plane (G17) at the feed rate. Its
direction is the way it turns as seen looking down on the work from above: G2
clockwise, G3 counterclockwise. The line gives the arc’s end
point with X and Y, and its centre in one of two ways.
I and J say where the centre is, measured in X and Y from the point where the
arc starts. That is the usual mode, G91.1, and it holds even while
G90 makes every X and Y a position. G90.1 makes I and J a position
too in this simulator. Those decimal codes are not universal; check the control's centre-mode rules.
R gives the radius instead, and the control works out the centre. A positive
R selects an arc of up to half a turn, a negative R the longer way round in the convention used here. An R arc cannot be a
full circle, because a circle through one point has no single centre; a full circle uses I and J
with its end point the same as its start.
Worked example. The tool is at X 72, Y 15, moving right along the
bottom of an outline. G3 X80 Y23 I0 J8 puts the centre 0 in X and 8 in Y from the start,
at X 72, Y 23. Turning counterclockwise about it, the tool rounds a corner of radius 8 and
ends heading up the right-hand side at X 80, Y 23. The end point has to lie on the circle:
change J8 to J6 and the start radius is 6 mm while the end radius is
√(8² + 2²) = √68, about 8.246 mm, so
the simulator stops at that line instead of guessing.
In the diagram, absolute endpoints and start-relative centres give
G03 X22 Y18 I0 J10 from A to B, then G02 X12 Y8 I-10 J0 back to A.
Each start point needs its own centre offsets. Radius format can instead use
G03 X22 Y18 R10 for the first arc under the stated convention.
G17 uses X/Y and I/J; G18 uses X/Z and I/K;
G19 uses Y/Z and J/K. I, J and K correspond to centre components along X, Y and Z.
Judge arc direction looking from the positive normal axis toward the plane: +Z for XY,
+Y for XZ, +X for YZ. The simulator supports XY arcs only.
Arcs: a rounded outline. With a Ø6 end mill, follow a 60 × 30 rounded rectangle with R8 corners at Z−1.5. The groove is 2.5 mm below the raw top, Z+1. Four G3 arcs turn the corners. The program on the left, numbered line by line, and the block after it has run. Solid lines are feed moves, dashed lines are rapids; the darker the cut, the deeper it goes.
A drilling cycle packs a whole hole into one line
A drilling cycle packs a whole hole into one line
G81 X15 Y15 Z-8 R2 F120 drills a hole at X 15, Y 15. The tool moves there at
rapid, comes down at rapid to R, the height just above the work where feeding begins,
feeds down to Z, the drill-tip depth, and rises back up at rapid.
How high it rises. In this simulator G99 rises to R between
holes, and G98 rises to the height the tool was at when the cycles began, or to R if
that is higher. The simulator needs one of the two before the first G81.
The cycle is modal. After a G81, a line with only X or Y drills
another hole with the same depth, R and feed. G80 ends the cycle; after it, a line
with only X or Y is an error until a G0 or G1 says how to move.
Without cutter compensation, program the tool centre path
Without cutter compensation, program the tool centre path
With cutter compensation off, as in this simulator, X and Y describe where the centre of the tool goes. The metal is
cut by its edge, one radius further out. A Ø10 end mill fed along Y 30 cuts a slot from
Y 25 to Y 35. To leave a wall at X 80 with the metal beyond it cut away, the centre runs
along X 85. Milling Operations works through the same rule for a shoulder and a pocket.
Controls can do that offset for you with cutter radius compensation, G41 and
G42. The simulator does not model it, so every program here is written as the path of
the centre, and a line with G41 or G42 stops with a message.
CNC Axes and Offsets explains left/right compensation and tool-length offsets.
Simulator: build a program from blocks and watch it run; Interactive
Simulator: build a program from blocks and watch it run
Pick a program, change it, and run it on a simulated block of metal in a vise. Every
block is one line of G-code; the same interpreter reads it whether you build it from blocks or type
it, and the machine runs exactly what it reads.
How to drive it. Press a block in the program to open it and change its
numbers; leave a number empty to leave that word out of the line. Add a block from the row above
the program, either by pressing it, which puts it after the open block, or by dragging it into
place. Drag a block by its handle to move it; with the keyboard, focus the handle and press Alt with
↑ or ↓. Play runs the program, Step runs one block and stops,
Back goes back one block and Restart starts again with fresh stock. The block
that is running is outlined in gold, and its move is drawn in gold on the views; press a move in
the top view to open the block that makes it. Any change to the program puts the simulation back to its
start, so it always shows the program as it now is.
Holes: a drilling cycle
(Four holes with a drilling cycle)
G21 G90 G17
T3 M6
S1500 M3
G0 X15 Y15
G0 Z10
G98 G81 X15 Y15 Z-8 R2 F120
X85
Y45
X15
G80
G0 Z20
M5
M30
Holes: a drilling cycle. Drill four Ø6 holes with the tip reaching Z−8: 9 mm below the raw top, Z+1. Full-diameter depth is shorter because of the drill point. G81 drills the first hole; X or Y lines repeat it until G80. The program on the left, numbered line by line, and the block after it has run. Solid lines are feed moves, dashed lines are rapids; the darker the cut, the deeper it goes.
Things to try in the G-code simulator; Checklist
Things to try in the G-code simulator
Step through the slot program one block at a time. Watch the machine state change at blocks 2,
3 and 4 before anything moves.
Make the slot deeper and shorter, then run it again. Which numbers did you change?
Drag the Spindle on block below the first feed move, and read why the machine
stops.
Open block 7 of the slot program and switch it from G1 to G0. What
does a rapid do when it meets the metal?
Change the slot program to G91. Before you press Play, work out where its first
move will go.
In the arcs program, swap one G3 for a G2, and see which way that
corner turns now.
In the holes program, try G99 instead of G98, then add a fifth
hole.
In the G-code text, type a line the simulator does not read, such as G41 or
M7, and read what it says.
What the simulator reads, and what it leaves out
What the simulator reads, and what it leaves out
The simulator reads a small part of G-code, modelled on the way the LinuxCNC control reads it.
Any other code, or a line that is not G-code at all, stops the program at that line with a message
saying what is wrong. Nothing is skipped.
Motion
G0 rapid, G1 feed, G2 and
G3 arcs in the XY plane, G81 drilling cycle and G80 to end
it.
Modes
G17 XY plane, G20 inches, G21
millimetres, G90 absolute, G91 incremental, G90.1 and
G91.1 for I and J, G94 feed per minute, G98 and
G99 for the drilling cycle. G40, G49 and G54 are
read as the settings the simulator always has: no cutter compensation, no tool length offset, and
its one work offset.
Machine
M3 spindle on, M5 off, M6 tool
change, M8 and M9 coolant on and off (shown, with no effect on the cut),
M2 or M30 program end: stops spindle and coolant, restores G90 and
G1, and leaves the tool at its last position.
Words
X Y Z positions, I J arc centre, R
arc radius or drilling R plane, F feed rate, S spindle speed,
T tool, N line number, and comments in brackets or after a
semicolon.
Tools
T1 Ø10 end mill, T2 Ø6 end mill,
T3 Ø6 drill, T4 Ø12 spot and chamfer tool, T5
Ø50 face mill: the tools of Milling Operations, with the same rules about what each can
cut.
Some rules are the simulator’s own. It needs units and distance mode
set before the first move, a speed before M3, G98 or G99
before a drilling cycle, and, for a tool change, the spindle stopped and the tool at Z +20 or
higher. A real control may be less strict, or strict about different things.
Only straight down and straight across, through metal. The block is modelled
as a height above each point, so the tool can cut moving along Z alone, or in X and Y at one
height. A move in X or Y and Z at once is followed exactly through the air, but stops if it would
reach the metal: ramping and helical moves are not modelled.
Arcs are cut as short straight chords, each within 0.01 mm of the true
arc. They are drawn as true arcs.
Simple timing. Feed moves run at their F and rapids at 5000 mm per minute,
with no acceleration, no blending of one move into the next, and no time for tool changes.
Not a check of a real program. The speeds and feeds in the examples are round
teaching numbers, not cutting recommendations; Feeds and Speeds is about choosing them. A
program that runs here has met only this simulator’s rules. Controls differ in the codes they
accept, in their starting modes and in how they change tools, so a program for a real machine is
checked against the manual for its control. See CNC Programming and Prove-Out for graphics,
dry run, single block and overrides; follow Equipment Operating Procedure for machine use.
Supporting detail: Addendum 53 — How the simulator reads and runs a program.
Machining/Milling
Climb and Conventional Milling
Climb and conventional milling name the cutter's rotation relative to the feed; that relationship changes chip formation, force direction, chip evacuation and the importance of backlash.
Climb versus Conventional Milling
Climb versus Conventional Milling
A milling cutter turns while the work and cutter move past each other. At the
part of the cutter that is making the cut, climb milling means the cutting edge
moves in the same direction as that relative feed; conventional milling means it
moves against the feed. Those two relationships are the whole distinction—not a universal
left or right direction on the machine.
Climb milling
Also called down milling. Where a tooth is
cutting, it moves the same way the work is fed. Each chip starts
thick and ends thin.
Conventional milling
Also called up milling. Where a tooth is
cutting, it moves against the feed. Each chip starts thin and
ends thick.
Feed per tooth, width of cut (radial engagement) and axial depth of cut are defined together
on Feeds and Speeds. This page uses them only to show what changes when the direction changes.
How to read the drawings on this page. Each drawing is a slice straight
across the fluted part of an end mill, seen looking down its axis from the spindle; the small
side view in the corner marks where the slice is taken, the way a section view does on a
drawing. This sketch places the cutter’s axis horizontally, with the
table below it; screen-down is toward that table. It is not a top view of a vertical spindle.
From the spindle end a standard right-hand cutter turns clockwise. The cutter
stays where it is and the work is fed past it, the way a mill table carries the work. Every
Work feed arrow is the work’s motion. On a machine that moves the
cutter instead, nothing changes: the cutter simply travels the opposite way to that arrow.
The cutter is the same in both modes and spins the same way. What switches climb to
conventional is the direction of feed relative to the side of the cutter that is doing the
cutting. The down/up force arrows belong to this horizontal arrangement. A top view of a
vertical end mill would show forces across the table, not axial force into or away from it.
Climb and Conventional Milling Lab; Interactive
Climb and Conventional Milling Lab
The dashed gold outline is the slice of metal the cutting tooth has taken so far,
drawn where it was. The solid gold shape is that metal as a chip, curling off the cutting edge
into the flute. The curled chip is scaled to fit the flute at this exaggerated feed; its drawn
size is not a thickness measurement. Play it, or drag the tooth through the cut, and use the
dashed slice to compare thickness.
Both cutters turn clockwise, and each cutting tooth is about a third of the way
through its cut. In climb milling the work feeds left and the tooth has
already taken the thick end of its chip. In conventional milling the work
feeds right and the tooth is still in the thin end. The graph says the same thing as a line:
climb runs down from its peak, conventional rises to it. The feed per tooth is drawn many
times larger than it would be on a machine so the chip can be seen at all. The dashed outline
is the slice as the tooth cut it; the solid chip is the same metal curled into the
flute.
Six things to try in the demo; Checklist
Six things to try in the demo
Choose Compare both and step until each tooth has only just entered the
work. Which chip is already thick, and which is barely there?
Drag Tooth position in the cut from 0% to 100%, once in each mode. Watch
the uncut-thickness readout: in climb it runs down from the peak, in conventional it runs
up to it.
Set the width of cut to 50%. The tooth now cuts through a quarter of a
turn and the chip reaches the full feed per tooth. Drop it to 10%: the peak falls to
60% of the feed. At the same feed, a narrower cut makes a thinner chip.
Raise the feed per tooth and watch the dashed uncut slice
grow thicker. Its thick-to-thin trend stays the same. The curled chip is scaled to fit the
flute and may stop growing on screen; compare thickness using the dashed slice.
Follow one chip from start to finish. Pause just before a tooth leaves the work, then
step forward. In which mode is the thick end of the chip at the curled tip, and in which is it
still on the edge? Where does each chip go once it is free?
Watch the red-orange arrow, the pull of the cut on the work. In climb it pulls the work the
way it is already going and presses it down. In conventional it pushes back against the feed
and lifts. Keep that in mind for the card on backlash.
Chip Forming Characteristics
Chip Forming Characteristics
A chip does not stay where the slice was. The cutting edge shears it off the work,
it slides along the face in front of the edge and curls into the flute, and the flute carries it
round until it is thrown clear. The drawings compare chip paths in the cross-section of this
horizontal-cutter setup; they do not predict a universal chip-flight direction.
Swipe the four moments sideways on a narrow screen, or focus the
drawing and use the arrow keys to scroll.
One chip from each direction, at four moments. In climb milling
the chip is finished at the finished surface, where the tooth leaves the work, and the end that
curled furthest is its thick end. It is thrown out behind the cutter, over the surface that has
already been cut. In conventional milling the chip is finished at the top of
the cut with its thick end still on the edge, and this cross-section carries it toward the
uncut side.
Reading the cutter
Cutting edge: the tip of each tooth, the only part that touches the
finished surface. The tooth that is cutting is outlined in blue.
Rake face: the face just in front of the edge. The chip slides along it as
it forms.
Flute: the curved valley in front of each tooth. The chip curls in it and is
carried out of the cut in it, so a flute packed with chips cannot take the next one.
Core: the solid centre, hatched because the drawing cuts through it. In
this illustrated family, more flutes leave a thicker core and less room for chips. Actual
proportions depend on the tool design.
The chip is drawn shorter and thinner than the slice so it fits the flute at a feed drawn many
times too large. A continuous metal chip usually becomes shorter and thicker through
deformation; IIT Kharagpur’s chip-formation lesson
explains the change. The path it takes is illustrative. Actual chip motion depends on the
rotation/feed relationship, engagement, flute helix and chip space, material and chip-breaking
geometry, and coolant or air delivery. Helical flutes can also move chips along the cutter axis;
this in-plane drawing does not show that axial evacuation. The cutter’s flutes and chip space follow
Harvey Performance’s end mill anatomy.
Starting thin means rubbing; starting thick means a bite
Starting thin means rubbing; starting thick means a bite
A real cutting edge has a rounded tip. When the uncut layer is extremely thin, the edge can
push and rub the metal before a chip forms. The peripheral cuts shown here make that difference
visible.
Conventional milling starts every chip at zero, so every tooth rubs before
it bites. Rubbing makes heat and wears the edge. Look at where the tooth enters, too:
on the finished surface, the one you are keeping.
Climb milling starts the chips shown here at their thickest, so the tooth bites at
once, with a knock each time a tooth enters.
That is the main reason climb milling is often the preferred direction when the machine can
hold the work steady against it. Whether it can is the next card.
Supporting detail: Addendum 54 — When a rough or hard exterior can change the choice.
Which way the cut pulls on the work
Climb: the teeth drag the work along the way it is already being fed, and
tend to push it down onto the table in the peripheral arrangement shown.
Conventional: the teeth push back against the feed and tend to lift the
work in this arrangement, so whatever holds the work has to resist that lift.
Backlash: why climb milling depends on the machine; Watch out
Backlash: why climb milling depends on the machine
A feed screw moves the table by turning inside a nut. Unless the drive is built to
take it out, there is some clearance between the screw’s thread and the nut. That clearance
is called backlash, and it lets the table shift a little while the screw stands
still.
The clearance is drawn far larger than it is, and a real drive engages many turns of
thread, not one. What matters is which side of the gap the nut ends up on.
Conventional: the cut pushes the table back, against the feed. That holds
the nut against the flank of the thread the screw is driving with, so the gap stays on the far
side and the table moves only as fast as the screw turns.
Climb: the cut pulls the table forward, the way it is already moving. If
there is clearance, the table can slide across it ahead of the screw, and the cutter takes a
bigger bite than the feed was set for, all at once. That sudden, oversized bite is hard on the
cutter and on whatever is holding the work.
This is a question to ask about a machine, not a rule. How much clearance a
feed drive has, and whether it is built to take it out, differs from machine to machine. This
page does not tell you which direction to use on any machine you will meet. Ask the person
responsible for that machine.
Supporting detail: Addendum 55 — Why some machines climb mill without trouble.
A slot is both at once
A slot is both at once
When the cutter is buried across its whole width, cutting a slot, each tooth enters at one
wall and leaves at the other. Its chip starts thin, is thickest halfway across, and ends thin. The
first half of the sweep moves against the feed, like conventional milling, and the second half
moves with it, like climb milling, whichever way the work is fed. That is why the demo stops at
half the cutter’s diameter: past that point the two can no longer be pulled apart.
Engineering Fundamentals/Power Transmission
Backlash and Lost Motion
After a feed screw reverses, the handwheel turns through the thread clearance before the carriage moves. Predict that dead band from clearance and lead, and finish every approach in one direction.
The dial counts screw rotation; the indicator finds the carriage
The dial counts screw rotation; the indicator finds the carriage
A handwheel dial does not measure the carriage. It counts how far the screw has turned and multiplies by the lead. That is the carriage's travel only while the screw keeps pushing the nut in one direction.
A screw and its nut need some clearance between their thread flanks (a flank is the sloping side of a thread). After the screw changes direction, it must turn through that clearance before its other flank meets the nut and the carriage moves again. That clearance is backlash, the same idea defined for threads in Screw Threads and for gear teeth in Gear Drives and Tooth Geometry. The turn spent crossing it is lost motion: input that produces no output.
Dial reading: travel predicted from screw rotation, turns × lead. See Lead Screws.
Carriage position: where the carriage really is, as an indicator touching it would show.
The difference is zero after a forward approach (the direction the dial was zeroed in) and one full clearance after a reverse approach.
Supporting detail: Addendum 56 — Where the definition comes from.
Experiment: reverse the handwheel and watch the carriage wait; Interactive
Experiment: reverse the handwheel and watch the carriage wait
What the lab does: a handwheel turns an axially fixed lead screw. The dial reports travel calculated from screw rotation; an indicator on the carriage reports where the carriage really is. With a total axial clearance B between screw and nut flanks, the two agree only while the screw keeps pushing in the direction the dial was set in.
Dead band after a reversal = clearance ÷ lead, as a fraction of a turn:
angle = 360° × B ÷ Ldial divisions = B ÷ (L ÷ 100) on a dial with 100 divisions per turn.
Worked example: lead L = 5 mm/rev, so one division is 0.05 mm. Clearance B = 0.20 mm. Angle = 360° × 0.20 ÷ 5 = 14.4°; divisions = 0.20 ÷ 0.05 = 4. For four divisions after the reversal the dial changes and the carriage does not.
Driven forward to 1.00 mm, then reversed four divisions: the dial reads 0.80 mm (the shaded graduations passed under the index) while the indicator still reads 1.00 mm. Illustrative dial, lead 5 mm, clearance 0.20 mm.Forward 0 → 1.00 mm, back to 0, forward to 1.00 mm again. Each reversal runs along a flat step 0.20 mm long, so the path is a loop B wide measured along the dial axis.The same three moments in the thread, drawn to scale for lead 5 mm and clearance 1.0 mm (the exaggerated clearance of the 3D model). The heavy blue line marks the flank in contact. Reversing moves the contact from one flank to the other; between the two, nothing pushes the carriage.
Landing on 25.00 mm after overshooting to 25.60 mm (lead 5 mm, clearance 0.20 mm). The dial was zeroed after a forward approach, so it equals the carriage position only for forward approaches:
Same dial number, two approaches.
Method
Dial moves
Dial
Carriage
Error
Reverse straight to the number
25.60 → 25.00 mm
25.00 mm
25.20 mm
+0.20 mm
Back past it, approach forward again
25.60 → 24.50 → 25.00 mm
25.00 mm
25.00 mm
0.00 mm
Reversing straight to the number stops the carriage one full clearance away from the dial reading, still on the overshoot side: the dial shows 25.00 mm while the carriage is at 25.20 mm. Backing past the target by more than the clearance and finishing in the original direction puts the screw back on the flank the dial was set against.
Model limits: rigid parts and a kinematic dead band only. Friction is assumed to hold the carriage still inside the clearance, and nothing else pushes it along the axis; there is no force, stiffness, preload or stick-slip model. Lead, clearance and the 3D dimensions are illustrative and exaggerated, not a standard thread or a measured machine.
Predict the dead band: clearance divided by lead, as a fraction of a turn
Predict the dead band: clearance divided by lead, as a fraction of a turn
Once the screw reverses, the carriage stays still until the screw's helix has moved one full clearance B. One turn moves the helix one lead L, so the dead band is B ÷ L of a turn.
dead-band angle = 360° × B ÷ L
dead-band divisions = B ÷ (value of one division); with 100 divisions per turn, one division is L ÷ 100.
Same clearance, two leads (illustrative values, 100-division dial).
Lead L
One division
Clearance B
Dead band
5 mm/rev
0.05 mm
0.20 mm
4 divisions = 14.4°
2 mm/rev
0.02 mm
0.20 mm
10 divisions = 36°
The carriage error after a reverse approach is the same 0.20 mm in both rows. A finer lead only spreads that same clearance over more of the handwheel turn.
Land on a dimension: finish every approach in the direction the dial was set
Land on a dimension: finish every approach in the direction the dial was set
If you overshoot a target, turning straight back to the number leaves the carriage one clearance away from it, because the screw is now pushing on the other flank. Instead:
Back the handwheel past the target by more than the backlash.
Turn in the original direction again and stop on the number.
The same reasoning applies when setting a dial. MIT's 2.670 milling notes say to “be certain that the backlash in the mechanism driving the table is taken up prior to setting the dial indicator”. A dial set with the clearance taken up in one direction reads true for approaches in that direction. For edge finding, Mill Indicators and Part Zero repeats the touch from the same direction for the same reason.
Supporting detail: Addendum 57 — A readout that measures the slide, and when the clearance still matters.
What the lab model includes and leaves out
What the lab model includes and leaves out
Included: a rigid screw that turns but never slides along its axis, a nut fixed to the carriage, and a total axial clearance B. The carriage position c stays within B/2 of the screw's helical phase s = −Lφ/360° (right-hand thread, φ the screw rotation); inside that band it does not move.
Assumed: friction holds the carriage still inside the clearance, and no outside force pushes it along the axis.
Left out: forces, friction values, stiffness and elastic wind-up, preload, stick-slip and wear. Drives designed to remove the clearance, such as preloaded ball nuts or adjustable split nuts, are beyond this model.
Illustrative: the leads, clearances and the 3D screw (5 mm lead, 1.0 mm clearance, 30° flanks) are chosen to be seen, not taken from a standard thread or a measured machine.
For drive design in depth, Alexander Slocum's Precision Machine Design (1992) covers actuators and transmissions in Chapter 10, including section 10.8, Linear Power Transmission Elements.
Mill around an island: X and Y backlash change the material left behind; Interactive
Mill around an island: X and Y backlash change the material left behind
A perfect 4 mm diameter cutter makes one contour groove around a 24 mm circular island. The dashed outline is the intended island; span errors compare with zero backlash. Material in the middle and outside the groove remains: this is not a pocket-clearing path. Only X/Y drive clearance changes.
Cut as intendedExtra cutMissed cut
Zero backlash · X 0.00 / Y 0.00 mm
X span at Y = 024.000 mm0.000 mm error
Y span at X = 024.000 mm0.000 mm error
X only · X 0.50 / Y 0.00 mm
X span at Y = 023.500 mm−0.500 mm error
Y span at X = 023.989 mm−0.011 mm error
Y only · X 0.00 / Y 0.50 mm
X span at Y = 023.989 mm−0.011 mm error
Y span at X = 023.500 mm−0.500 mm error
Both axes · X 0.50 / Y 0.50 mm
X span at Y = 023.500 mm−0.500 mm error
Y span at X = 023.500 mm−0.500 mm error
All four start at the right side and run counterclockwise. Before the plunge, both axes approach the start in their positive directions and the coordinates are calibrated there. On reversal, the actual cutter can stay still along one axis while moving along the other. Its full circular footprint removes material along that path.
The spans above are sections through X = 0 or Y = 0, not a best-fit circle or the greatest overall width. They come from intersections with the swept cutter, not pixel counting. Changing the start, winding or initial seating can change the local shape; do not infer that every circle becomes the same oval.
Approximation: circular paths use 1° chords (under 0.001 mm sag here). These printed diagrams use 0.20 mm cells; the live area shading uses 0.10 mm cells. The millimetre span readings use the underlying swept chords.
The cutter removes the upper layer at one fixed depth. Grey is material still at its original height; the groove floor is not drawn. A boundary pass leaves both an island and the surrounding stock, so it is not a cleared pocket. The dashed line is the intended island boundary, not the cutter-centre path: that path is one cutter radius outside it.
Supporting detail: Addendum 58 — What the model means, and why it is not a prediction for every CNC.
Machining/Milling
Feeds and Speeds
Six key quantities connect cutting data to the machine and the cut: cutting speed, feed per tooth, spindle speed, table feed, radial engagement and axial depth.
Six quantities connect the cutting data, machine and cut
Six quantities connect the cutting data, machine and cut
Start by separating six quantities. Cutting speed and feed per tooth describe
what the cutting edges should experience. Spindle speed and table feed are the corresponding
machine settings. Width of cut and axial depth describe how much of the cutter is engaged. The
exact starting values come from cutting data for the tool, material and operation.
Cutting speed
How fast the cutting edge moves through the material: the speed
at the outside of the cutter. Written in surface feet per minute (SFM) or
metres per minute (m/min). Also called surface speed. Comes from cutting
data.
Feed per tooth
How far the work moves while the cutter turns from one tooth to
the next, in inches or millimetres per tooth. Also called chip load, because it sets the
size of the bite each tooth takes. Comes from cutting data.
Spindle speed
How many turns the cutter makes each minute, in revolutions per
minute (rev/min, usually said RPM). Set on the machine.
Table feed
How fast the work moves past the cutter, in inches or millimetres per
minute. Also called the feed rate. On a machine that moves the cutter instead, it is how
fast the cutter travels. Set on the machine.
Width of cut, ae
The cutter’s radial engagement in
the work, measured across the cutter in inches or millimetres. It is also called radial
depth of cut; those names describe the same quantity, not two separate settings.
Axial depth of cut, ap
The distance engaged along the
cutter’s axis, in inches or millimetres: the difference between the uncut and cut surfaces in
that axial direction.
"Speeds" are the cutter turning; "feeds" are the work moving. That is all
the phrase feeds and speeds means. The two are set separately on the machine, and the
rest of this page is about how they depend on each other, how width and axial depth set the
engaged cut, how much metal that cut removes, and the power it takes. Climb and Conventional Milling
explains how feed direction relative to rotation changes the chip from thin-to-thick or
thick-to-thin.
Spindle speed: the turns it takes to reach the cutting speed
Spindle speed: the turns it takes to reach the cutting speed
Cutting speed is a distance per minute, measured at the edge of the cutter. In one turn, a point
on that edge travels once around the cutter: its circumference, π × D.
So the spindle speed is the cutting speed divided by the distance one turn covers, once the two
are in the same units.
One turn of a 0.5 in cutter carries its edge 1.571 in. A cutting speed of 300 SFM
asks the edge to cover 300 ft, which is 3,600 in, every minute.
The 12 is there because cutting speed is in feet and the diameter is in inches:
12 inches to the foot. In metric units the same idea needs a 1000, millimetres to
the metre:
Worked example, with illustrative numbers. A 0.5 in cutter, with cutting data
that calls for 300 SFM: 12 × 300 ÷ (π × 0.5) = 2,292 rev/min.
The 300 SFM is there to make the arithmetic concrete. It is not a recommendation for any tool or
any material. The feed value below is illustrative too; use the later material guide for the
carbide starting bands.
Look at where the diameter sits: underneath. Halve the diameter and the spindle speed
doubles. A small cutter has to turn faster to give its edge the same cutting speed, which
is why small cutters so often ask for more speed than a machine has. That case has its own card
below.
Table feed: each tooth takes one chip load as it passes
Table feed: each tooth takes one chip load as it passes
Each time a tooth passes, the work should have moved one feed per tooth. The number of teeth
passing each minute is the spindle speed times the number of flutes, so:
Worked example, continued. A 4-flute cutter at 2,292 rev/min, with a feed per
tooth of 0.002 in: 2,292 × 4 × 0.002 = 18.3 in/min. The units work out
on their own: turns per minute, times teeth per turn, times inches per tooth, leaves inches per
minute.
Turned around, the same relation tells you what chip load a feed gives. That is the way to check
a feed someone else has set:
Flutes here means the teeth that actually take a chip in each turn. For an
ordinary end mill that is simply the number of flutes.
Removal rate: width × depth × table feed
Removal rate: width × depth × table feed
Two more numbers describe the cut itself. The width of cut is how far the
cutter reaches sideways into the work, across the direction of feed. The depth of
cut is how far down its side the cutter is buried, along its axis. A side cut can be no
wider than the cutter; a wider path takes more than one pass.
While the table feeds, the cutter sweeps through the work and leaves a channel behind it. In a
stretch of time the cutter travels table feed × time, and the solid metal it
takes out in that stretch is a slab as wide as the cut, as deep as the cut, and as long as that
travel. Per minute, that is the material removal rate:
Removal rate (in³/min) = width of cut (in) × depth of cut
(in) × table feed (in/min)
The front of the cut is curved, because it is the cutter’s own shape. The slab’s
two ends are the same curve, one travel apart, so whatever the curve, the slab holds exactly
width × depth × travel.
Worked example, continued. The same cutter, 0.1 in wide (20% of its diameter)
and 0.5 in deep, at 18.3 in/min: 0.1 × 0.5 × 18.3 ≈ 0.917
in³/min, or 15.0 cm³/min. In 5 seconds the cutter travels 1.53 in and removes
0.0764 in³.
That is the volume of solid metal. Once it is chips, the same metal lies loose
and takes up far more space, so the removal rate is not the size of the chip pile.
Cutting power: removal rate times the energy each cubic inch takes
Cutting power: removal rate times the energy each cubic inch takes
Cutting a cubic inch of a given material takes roughly a given amount of energy. That is its
specific cutting energy, and it is why the same cut needs more power in steel than
in aluminum. Written as power per unit of removal rate, it turns a removal rate straight into
power:
Cutting power (hp) = removal rate (in³/min) × unit power
(hp per in³/min)
Cutting power (W) = removal rate (mm³/min) × specific
energy (J/mm³) ÷ 60
The 60 turns minutes into seconds, because a watt is a joule per second. One
mechanical horsepower is 745.7 W.
Worked example, continued. For 6061-T6 aluminum the guide in the calculator
gives a band of 0.16 to 0.30 hp per in³/min. At 0.917 in³/min that is about
0.15 to 0.28 hp, or 110 to 210 W. The same cut in 304 stainless steel, whose band
is 0.65 to 1.3 hp per in³/min, needs about 0.60 to 1.2 hp.
The answer is usually a band. Some table rows give one factor; that still
produces an estimate, not a measured certainty. Published bands cover a family of materials over
a range of hardness, and the energy a real cut takes also depends on chip thickness and
cutting-edge geometry. Treat the band as the size of the job, not a
prediction to the watt.
Three different powers get called “the power”, and they are not the same:
Cutting power is what the cutting edge spends. That is the estimate
above.
Motor shaft power is more, because the belts, gears and bearings between the
motor and the cutter lose some on the way. If the drivetrain passes on 80% of what the motor
delivers, the motor must deliver the cutting power ÷ 0.8.
Electrical power drawn from the supply is more again, because the motor and
its drive lose some too.
None of these tells you whether a machine can take the cut. A spindle motor’s rating is
not necessarily available at every speed, and power is only one limit: the tool, the holding and
the machine’s stiffness are others. Use the estimate to compare cuts and to see why a cut
is heavy, not to approve one.
Calculator: from material guidance to removal and power; Interactive
Calculator: from material guidance to removal and power
Choose a material to see generic carbide starting ranges for it, then enter the
cutter, the cutting data and the cut. Nothing is filled in for you. Each result appears as soon as
the fields it needs are filled, and says what it is still waiting for until then. Use the middle
of these ranges fills in only the cutting speed and feed per tooth.
How to read the drawings. You are looking at a block of work from the front,
a little above and to the right. The cutter feeds to the right along the front top edge.
Gold is the solid metal removed in the time shown, drawn where it was before the
cutter took it. The dashed blue curve is where the front of the cut stood when that time began;
the outlined cylinder is the cutter now. The three labeled axes use the same length scale, and the
drawings in one comparison share that scale, so a slab twice as long is twice the volume. The
block around the cut is there only to hold it; its size means nothing. The cut is already
underway when the interval starts; only the gold volume counts toward that interval.
The comparison below the drawings prices the setup’s removal rate in every material.
The cut is the same in every row; only the energy per cubic inch changes.
Setup2,292 rev/min · 18.3 in/min
In 5.00 s: travel 1.53 in, removed 0.0764 in³Removal rate 0.917 in³/min · power ≈ 0.15–0.28 hp
In 5.00 s: travel 3.06 in, removed 0.153 in³Removal rate 1.83 in³/min · power ≈ 0.29–0.55 hp
MaterialEstimated cutting power at 0.917 in³/min
6061-T6 aluminum0.15–0.28 hp
7075-T6 aluminum0.15–0.28 hp
1018 mild steel0.56 hp one factor
1045 steel0.56–0.59 hp
4140 prehard steel0.72–0.83 hp
304 stainless steel0.60–1.2 hp
316 stainless steel0.60–1.2 hp
Class 30 gray cast iron0.40–0.60 hp
C360 brass0.28 hp one factor
Ti-6Al-4V titanium0.69–1.1 hp
01.02.0 hp
The worked example, 0.5 in cutter, 0.1 in wide and 0.5 in deep, three ways over
the same 5 seconds and at one scale. Doubling the spindle speed with the table feed held leaves
the slab unchanged: the same metal is removed, in chips half as thick. Doubling it with the chip
load held doubles the table feed, so the slab is twice as long and the removal
rate and power estimate double. Below, the setup’s 0.917 in³/min priced in each
material, from the guide’s cutting-energy bands.
Things to try in the calculator; Checklist
Things to try in the calculator
Before typing anything, read the results. Each one names the fields it is waiting for.
With 6061-T6 aluminum chosen, enter a 0.5 in diameter and
4 flutes, then press Use the middle of these ranges. Cutting speed and
feed per tooth fill in at 400 SFM and 0.00415 in/tooth. Nothing else changes.
Enter a width of 0.1 in and a depth of 0.5 in. Check the
removal rate against width × depth × table feed in the arithmetic.
Change the material to 304 stainless steel. The drawings, the table feed and
the removal rate stay exactly as they were; only the power estimate moves. The guide now marks
your cutting speed as above its range for stainless, but the calculator leaves your entries
alone.
Hold the table feed and drag the spindle speed to 200%. The slab does not
change and nor does the removal rate: the same metal comes off, in chips half as thick.
Now hold the chip load and do the same. The table feed doubles, the slab
grows twice as long in the same time, and the removal rate and power estimate double.
Switch to metric and back. The cut and every result stay the same, written
in other units, and any field you left empty stays empty.
Try a width larger than the diameter. The calculator refuses it: that is more than one
pass.
Enter a 0.125 in cutter with a 4,000 rev/min machine top
speed. Compare the speed the cutting speed asks for with the speed the machine can reach, and
where the cutting speed actually reached sits against the guide.
When the machine cannot turn as fast as the formula asks
When the machine cannot turn as fast as the formula asks
Because small cutters ask for high spindle speeds, the formula can easily give more than a
machine’s top speed. A 0.125 in cutter at 300 SFM asks for 9,167 rev/min. On a machine that
stops at 4,000, it runs at 4,000.
Work the feed from the speed the spindle really turns. At 4,000 rev/min, with
2 flutes and 0.0008 in per tooth, that is 4,000 × 2 × 0.0008 = 6.4 in/min. Each tooth
then takes the chip load you meant it to.
The cutting speed is what you give up. At 4,000 rev/min the edge of a 0.125 in
cutter moves at π × 0.125 × 4,000 ÷ 12 = 131 SFM, not the 300 the data
asked for.
Working the feed from the speed that was asked for, rather than the one the spindle is turning,
gets the chip load wrong by the same ratio. A feed worked from 9,167 rev/min would give each tooth
9,167 ÷ 4,000 = 2.3 times the chip load intended.
The same applies to a machine that offers only certain speeds, through belts or gears: set the
nearest speed it has, and work the feed from that one. The calculator’s machine top
speed box does this for a top speed.
Generic ranges are a starting point, not your tool’s data; Watch out
Generic ranges are a starting point, not your tool’s data
What determines the surface speed?
Start with the material and its hardness, then the cutting tool. Carbide is
a tool material, not one universal cutting edge: its grade, coating and geometry matter, as do
coolant and how steadily the setup can cut. A harder material generally calls for a lower
starting speed. Once that surface speed is chosen, the diameter determines the RPM;
a smaller cutter must turn faster to reach the same surface speed.
What determines the feed per tooth?
Choose a bite the cutting edge can carry and the flute can clear. That is why
the guide changes with cutter diameter: small cutters generally take smaller bites. Width and
depth of cut, a full-width slot, a long tool and a flexible setup change the loading and chip
clearance. The number of flutes then tells you how many of those bites happen per turn; it does
not by itself choose the size of each bite. Use the diameter-based range as a starting point,
with its stated engagement, and refine it from the actual tool’s cutting data.
The arithmetic on this page is exact. Its inputs are the part that takes
judgement. The ranges in the calculator are generic starting bands for solid-carbide end
mills, by material family, from published toolmaker tables. The cutting data for your
tool comes first: normally the toolmaker’s for that tool, that material and that kind of
cut, or the person responsible for the machine.
Cutting data accounts for things these formulas cannot see: the exact grade and hardness of the
material, the tool’s coating and geometry, how far it reaches out of the holder, coolant, and
how the setup is held. The guide’s feed-per-tooth ranges assume a side cut about 20% of the
cutter’s diameter wide; a full-width slot or a long reach needs the tool’s own data. A
calculator turns data into machine settings. It cannot tell you whether the data suit your cut.
Supporting detail: Addendum 59 — Feed per tooth is not always the thickness of the chip.
Machining/Milling
Tool Stickout and Deflection
A longer tool reaches farther, but bends much more under the same sideways load. Predict a change, test it, and compare the tool with a fixed baseline.
Experiment: how much does a shorter tool help?; Interactive
Experiment: how much does a shorter tool help?
Predict first: keep the same solid tool, force and material. If its unsupported length drops from 80 to 40 mm, will it bend half as much, one eighth as much, or the same amount?
Halving the clamp-to-load length, from 80 to 40 mm, cuts ideal tip bending from 57.947 to 7.243 µm: one eighth as much. Both drawings use the same geometry scale and ×100 bend magnification. Dashed outlines are the unloaded tools.In this ideal one-sided wall pass, the cutting edge bends away from the work. The short setup leaves 7.243 µm of extra material at the tip. This is a geometric consequence of the assumed load, not a prediction of a real finished part.
The length that matters starts at the clamp
The length that matters starts at the clamp
Stickout is unsupported length. Here it is the distance from the rigid clamp face to the point where the sideways force acts. It is not the total tool length, and it is not automatically the flute length.
A force bends the tool and moves its cutting edge away from its unloaded position. The clamp end stays in place and keeps its direction; the bend grows along the unsupported length. In a real cut, force acts over the engaged cutting edges rather than at one perfect point.
Reach is a tradeoff. The tool must reach the feature and clear the work and holder. Within those constraints, reducing unnecessary stickout improves bending stiffness. A larger diameter can also help if the feature has room for it.
Length is cubed; diameter is raised to the fourth power
Length is cubed; diameter is raised to the fourth power
For the ideal solid round tool in this lab, tip bending is proportional to F × L³ ÷ (E × d⁴). Change one input at a time to see which factor is responsible.
½ length → ⅛ bendHalf × half × half. Shortening the unsupported reach is a powerful change.
2× diameter → ¹⁄₁₆ bend2 × 2 × 2 × 2 goes into the denominator. This comparison is for solid round sections.
2× force → 2× bendForce acts linearly. Double E instead, with the other inputs fixed, and bending halves.
Try a combined change: start at L = 40 mm and d = 10 mm, save that baseline, then set L = 80 mm and d = 20 mm. The length contributes 8× and the diameter contributes 1/16×, leaving half the bending at the same load and modulus. The two changes compete.
Supporting detail: Addendum 60 — The calculation and its units.
Use the trend; know what this model leaves out
Use the trend; know what this model leaves out
This is a static, linear-elastic, slender-beam model: a uniform solid round tool, a perfectly rigid clamp, a steady sideways point load at the free end and deflections small compared with the length. Elastic modulus E measures stiffness, not hardness or breaking strength.
The lab starts at E = 600 GPa, a chosen teaching value within Hyperion's published 430–690 GPa range for cemented-carbide wear grades. It is not a universal value for carbide cutters. The force is an assumed input, not calculated from a milling operation.
Flutes and necks change stiffness. The holder, spindle and workpiece can move too, and a milling force changes as teeth engage. Static bending alone does not predict chatter thresholds, surface roughness, fracture, safe loads or guaranteed finished-part error.
The lab flags L/d below 4 and bending above 1% of L as prompts to question this simplification. Those are teaching checks, not certification or safety limits.
Identify lathe parts, choose workholding and tools, and explore turning operations in a simulator before studying tapers and threads.
Identify the parts that rotate, guide and feed
Identify the parts that rotate, guide and feed
Typical engine-lathe arrangement. Controls and drive arrangements vary by machine.
Headstock and spindle
The headstock supports the rotating spindle and its drive. The spindle carries the chuck, faceplate or other workholding. Its axis is the reference for turning.
Bed and ways
The bed supports the assemblies. Its guide surfaces, called ways, guide carriage travel and locate the tailstock.
Tailstock
Supports the far end of work with a center, or holds axial tools such as drills and machine reamers. Its quill extends along the spindle axis when aligned.
Carriage, cross slide and compound rest
The carriage moves along the bed. The cross slide moves across the axis. The compound sits above it and swivels to guide short angled movements. The tool post holds the cutting tool on the compound.
Apron
The assembly on the front of the carriage contains feed mechanisms and controls.
Feed rod and leadscrew
On machines with separate drives, the feed rod powers ordinary turning or facing feeds; the leadscrew synchronizes threading. Some lathes combine functions, so identify the actual drive arrangement before selecting a feed.
Half-nuts and quick-change gearbox
The half-nuts close around the leadscrew to drive the carriage during threading. The gearbox selects ratios for feeds or thread pitches; its chart tells you which control combinations produce each ratio.
Choose workholding for the shape and the surface that must run true
Choose workholding for the shape and the surface that must run true
Running true means the reference surface is centered on the spindle axis, with little measured runout. Gripping a part does not automatically establish that alignment.
One device can perform several roles, but the roles are not interchangeable. The drive path shown for work between ordinary centers runs from the drive plate through the dog; the centers support and locate the axis.
Read the role before choosing the device. A setup can combine location, grip, drive and support. Support is not the same as drive, and gripping a surface does not by itself establish that it runs true.
Three-jaw scroll chuck
Moves its jaws together for quick gripping of round or suitable hexagonal stock. Self-centering is convenient, but does not promise zero runout, especially after removing and replacing a part.
Four-jaw independent chuck
Moves each jaw separately. Useful for square or irregular work, intentional offsets, or indicating an existing surface true. Adjustment takes more time.
Collet
Grips around a matching stock shape over a limited size range. Useful for repeated work and small diameters; choose the correct collet and its permitted clamping range.
Centers and drive dog
Centers locate work through prepared center holes. A dog clamped to the work transmits torque from a drive plate; the centers provide support. Useful when several surfaces must share the center-hole axis or the work must be removed and returned.
Faceplate
Provides slots or mounting features for clamping work that does not suit chuck jaws. The fixture must locate the intended axis and account for the off-center mass.
The steady rest stays at one bed position. The follow rest moves with the carriage near the cutter. In both examples the chuck remains the primary drive; the rest contacts support the rotating work.
Steady rest
Mounts on the bed and supports a long workpiece at a fixed position. It can support work when the end must remain accessible for drilling or boring.
Follow rest
Mounts to the carriage and travels with the tool. It supports slender work near the cut to reduce bending under cutting force. Its support contacts must suit the diameter being supported.
Supporting detail: Addendum 62 — Reference examples for these mechanisms.
Match tool travel to the feature, and set the edge on center
Match tool travel to the feature, and set the edge on center
Facing feeds across an end to establish a plane perpendicular to the axis. Straight turning feeds parallel to the axis to make a cylinder. Shoulder turning joins a cylindrical diameter to a step face; the tool needs clearance to reach both surfaces without its holder or unused edge contacting the work.
Center drilling prepares the seat for a supporting center. It is not automatically the best starting method for every drill. Tailstock drilling makes an axial hole; reaming removes a controlled small allowance to improve hole size and finish. Select that allowance from the reamer maker. A reamer generally follows the existing hole, so it is not a reliable way to correct a misplaced hole. Boring uses a single-point tool to enlarge an existing hole and establish its surface relative to the spindle axis.
Grooving makes a recess; parting continues a similar radial cut until the part separates. Chamfering makes a bevel at an edge. Form knurling presses a pattern into the surface and increases the diameter over the raised points; cut knurling is a different process.
Use spindle center height as the starting reference for ordinary turning and threading tools, then follow any tool-specific instruction. An off-center edge changes the effective rake and clearance. In facing, a tool that cannot reach the axis leaves a small central nub. Boring also needs clearance between the tool and the curved hole wall.
Supporting detail: Addendum 63 — Rake and chipbreakers depend on the whole cutting setup.
A lathe spins the work and moves the tool on two slides
A lathe spins the work and moves the tool on two slides
In ordinary turning the work rotates while the tool feeds along it or toward its
center. This simulator models parts that are round about the spindle’s axis.
The shape comes from where the tool is put against the turning work, and in what
order.
Chuck
Grips the bar and turns with the spindle. The jaws hold one end; the tool
works on the part of the bar that sticks out.
Carriage (Z)
Carries the tool along the bed, parallel to the axis. On this page
Z minus is toward the chuck and Z plus toward the
tailstock.
Cross slide (X)
Moves the tool straight toward or away from the axis. The
simulator’s X readout is the diameter the tool would cut. On a real lathe,
find out whether a dial or readout shows diameter or radius before you trust it.
Tailstock
At the far end of the bed, on the centerline. It holds the center drill,
the drills and the countersink, which feed straight along the axis.
Z0
Where the drawing’s finished front face belongs. Every length on the
drawings is measured from the front face.
How to read the simulator’s picture. It is the lathe seen from the front,
with the whole work in one of two views; the switch under the picture changes view at any time
without changing the work. Section cuts the work in half along its centerline, the
way a section drawing shows a part: metal is hatched, a hole shows open, and a thread or knurl shows
as its profile. Normal shows the outside, as you would see the part on the machine,
with its thread and knurl; a hole is drawn as a dashed hidden line, and a tool working inside the
part is hidden by it. Tools on the cross slide come in from above, in the plane the section is cut
on, so a tool and the metal it cuts are in the same plane.
The dashed blue outline is the drawing, placed at Z0. Tinted metal is still to come off. A heavy
red-orange dashed line marks work that is already smaller than the drawing allows. On each tool, the
gold edge is the part that cuts and the small ring is the point the readout follows.
Before using a real lathe, read Machine Hazards and Housekeeping. The chuck key stays in your
hand until it is out of the chuck.
Lathe simulator: make the part on the drawing; Interactive
Lathe simulator: make the part on the drawing
The stepped pin part-way through. The facing blade made the end flat; the body
was rough-turned short of its shoulder. Its angled allowance still needs radial facing.
The rougher is halfway along the first pin pass. The tinted metal is still to come off.
Each target is shown in a complete Normal outside view and a
complete Section through its centerline. Dashed lines in Normal identify hidden
internal features. Section thread bands are schematic crest/root envelopes; the knurl remains a
surface callout. Dimensions are nominal inches; the simulator’s feature table gives the limits.
Stepped pin. Two turned diameters, a chamfer on the end and a head left at bar size. From 1.000 in bar. Facing, turning in passes, square shoulders, chamfering and parting off.
Bushing. A turned sleeve with a through bore, two outside chamfers and a chamfer at the bore mouth. From 1.250 in bar. Center drilling, drilling, boring, countersinking and chamfering a back edge before parting.
Thumb screw. A knurled head, a thread relief groove and a 16-per-inch thread. From 1.000 in bar. Knurling, grooving, external threading in passes, and why a thread needs somewhere to stop.
Knurled nut. A knurled ring with a 13-per-inch thread inside it. From 1.000 in bar. Drilling to the thread's inside size, internal threading and parting off a ring.
Pick a part and read its drawing (Drawing opens it). Then make it from bar: choose a
tool, start the spindle, and move the tool against the work. Measure the part against the drawing
whenever you stop.
Move the tool by dragging it in the picture, one slide at a time as if you were
turning one handwheel; with the handwheel buttons, which repeat while held; with the arrow keys once
the picture is selected; or by typing a position and pressing Feed to Z or Feed to
X. Stop feed, or Escape, stops a feed part-way.
The spindle has to be running for anything to cut, and stopped before you change
a tool, change the drill size or thread pitch, or measure the part. A caliper or micrometer never
goes on turning work.
Each tool does only its own kind of work. When a move would ask a tool for
something it cannot do, the simulator stops the tool there and says why. What the tool in use can
and cannot do is listed under the tool rack. Turn / face offers two actual tools in its
Cutting tool selector; stop the spindle before changing between them.
Undo takes back one cut at a time. New bar starts the part again,
and undo brings the old bar back.
Measure compares every feature with the drawing. The list under the simulator
says what each feature still needs, or whether it has already gone too far to save.
Each simulator tool has a defined job; its form knurl displaces metal
Each simulator tool has a defined job; its form knurl displaces metal
Facing, turning, drilling, threading and knurling are not the same action with a
different shape of tool. Each tool reaches different places and moves in different ways.
Turning tool
The roughing tool supplied here has a 10° lead and a
0.008 in nose radius. Its angled edge reaches ahead of its readout point. Stopping an axial
pass leaves an angled shoulder; leave allowance instead of running the nose to the finished
shoulder plane.
Facing tool
A square, 1/16 in wide finishing blade set for radial feed.
Put its left corner at the finished Z and feed X across the end or shoulder. Its actual sweep
removes the allowance and leaves the face flat. Retract clear before moving Z. These are the
simulator’s selected tools; other turning tools and settings can both turn and face.
Chamfer tool
A 90° point. Plunged at an edge, each side cuts a 45°
chamfer.
Grooving and parting blade
Cuts only on its front edge. Plunged part of the way it
makes a groove; plunged to the center, or into the hole, it parts the work off.
Center drill, drill, countersink
Fed from the tailstock along the axis. A center
drill makes a small pilot and a 60° seat for a lathe center; a spotting drill has different
starting geometry. The twist drill makes the hole and the countersink chamfers its mouth. Choose
a real drill’s starting method and spot angle from its maker’s guidance.
Boring bar
Turns the inside of a hole that is already there, the way the turning tool
works on the outside. Its bar has to fit in the hole.
Threading tools
Form tools with a 60° point, one for outside threads and one on a
bar for inside threads. For a threading pass the carriage is driven by the leadscrew, so the tool
moves one thread lead for every turn of the work; for the single-start threads here, lead equals pitch.
Knurling tool
A straddle head with two opposed forming wheels, with left-hand and
right-hand diagonal teeth. Their axles lie parallel to Z; together they roll a diamond pattern.
Gap Ø is a simulated gap/contact-diameter control. A real tool’s arm adjustment and
locking are separate from cross-slide infeed. The picture shows the barrels and axles in side
projection; the highlighted rims are the forming contact.
What makes the simulator stop a move. Each of these is something the real tool
cannot do, or should not be asked to:
The spindle is stopped, so the tool would only rub.
More of the cutting edge would be in the work than the tool takes in one pass. Take it in more
passes.
A part of the tool that does not cut would touch the work: the holder, the top of an insert, the
bar of a boring tool, the side of a blade.
The tool would hit the chuck jaws.
A threading tool is still in the thread when the carriage moves. Back it out first.
A knurl is pressed past full depth, or into a shoulder.
A drill is three of its own diameters into a cut without being backed out to clear its
chips.
The simulator’s limits are teaching values chosen for these four parts. A real
tool’s limits come from its maker and depend on the machine, the material and the setup.
Tool shape and setting matter:
Sherline’s tool-grinding guide
describes cutting edges, clearance and form tools. The particular blade widths, nose radius and
operating limits used here are declared teaching choices, not specifications for every lathe tool.
Most parts start with a facing cut and end with parting
Most parts start with a facing cut and end with parting
More than one order can make the same part, and the simulator accepts any order that
gets there. Some orders are still better than others, for reasons you can see in the work.
Face first. The faced end is the datum: every length on the drawing is measured
from it. A sawn end is neither flat nor square to the axis.
Rough, then finish. A large step comes off in several passes, and only the last
one goes to the size on the drawing. On the stepped pin, stop the roughing passes 0.040 in
short of each shoulder, retract clear and change to the facing tool. With its left corner at the
final shoulder Z, feed X to the smaller diameter. A deeper shoulder needs more clearance for the
angled edge; the facing blade can remove its allowance in slices up to 0.0625 in wide.
Drill, then bore to size. The boring bar needs a hole its bar fits in. A center
drill is available here to explore its pilot and center-seat shape, but it is not a universal
prerequisite for drilling. On a real machine, use the starting method recommended for the
following drill.
Heavy work before thin sections. Knurling pushes hard on the work. Knurling
before the thin sections next to it are turned keeps the work stiff while it happens.
Make room for the threading tool. A thread needs somewhere for the tool to stop:
a relief groove, or a stop short of the shoulder. Chamfer the end first so the thread starts on a
clean edge.
Chamfer the back edge before parting. Once the part is cut off, its back edge is
out of reach without a second setup.
Part off last. The part falls free, and anything not done to it by then needs
another setup.
The small Chuck end circle is a separate axial view, looking from the tailstock
toward the chuck. Its forward-rotation arrow and animation run counterclockwise in that view.
The main picture is longitudinal, not a view along the spindle. See the viewpoint diagram in
Haas’s
spindle-direction training, page 16.
A thread is cut in several passes along one helix
A thread is cut in several passes along one helix
For a threading pass the carriage is coupled to the leadscrew, so the tool advances
exactly one lead for every turn of the work. For the single-start threads here, that is one pitch: 1/16 in for 16 threads per inch. Every pass follows
the same helix, so every pass deepens the same groove.
The section uses a schematic thread envelope. Its unhatched bands lie between
the surviving crests and dashed roots; they are not solid annular thread teeth. A real single-start
helix intersects opposite sides half a pitch apart. The model samples a thread profile for depth
checks and preserves its material envelope for later cuts and parting. The carriage still stops at
the requested Z, even between full turns. Sandvik
explains the lead/pitch relation for single- and multi-start threads.
A little deeper each pass. The simulator allows 0.015 in of infeed per pass
outside and 0.012 in inside. The thread is done when its roots reach the root diameter on the
drawing.
Back out before you come back. At the end of a pass the tool is still in the
groove. Bringing the carriage back without backing the tool out would cut the crests off.
Keep the pitch. A pass at another pitch cuts a second helix across the first: a
crossed thread.
Leave somewhere to stop. The tool runs into a shoulder unless a relief groove
gives it room, which is why the thumb screw has one.
Supporting detail: Addendum 64 — How a manual lathe keeps every pass on one helix.
A form knurl moves metal instead of cutting it
A form knurl moves metal instead of cutting it
The knurling tool makes no chips. Its wheels press a pattern into the turning work,
and the metal they push aside has to go somewhere: it rises between their teeth. The diameter over
the points ends up larger than the blank was.
Blank
The diameter turned before knurling. On these drawings it is not given: the
knurled size is, and you work back to the blank.
Full points
The wheels pressed deep enough that the diamond points are fully
formed. Too shallow leaves flat tops; pressing past full depth crushes and flakes the
points.
Over the points
The knurled diameter, measured across the crests.
In the simulator the roots sink exactly as far as the crests rise, 0.010 in each
at full depth, so a full knurl adds 0.020 in to the diameter. A real knurl’s growth depends
on its pitch, the material and how hard it is pressed.
On the stepped pin, try to turn the body from 1.000 to 0.750 in in one pass.
Read why the tool stops, then do it in two.
Stop a roughing pass short of a shoulder and Measure. Choose Face without
cutting and measure again. Then make the radial facing cut: which geometry actually changed?
Face the end but stop the cut at X 0.200. Measure: what is left at the
center?
Face 0.010 in past Z0 and make the rest of the pin to the readout. Measure
the lengths. Which ones moved, and why?
On the bushing, bring the boring bar to the end before there is a hole. Then
drill a 1/4 in hole and try again. Which drill gives the bar room?
Drill the bushing without the center drill and read the note under the inspection.
On the thumb screw, try a threading pass up to the head before cutting the
relief groove.
Take one threading pass at 16 per inch, change to 13 and take another. Measure the
thread.
Knurl the head straight on the 1.000 in bar, without turning a blank first. Measure it
against the drawing.
On the knurled nut, part the ring off and watch where the blade breaks
through.
What this simulator leaves out
What this simulator leaves out
The simulator is about which tool does what, in what order, to what size. It leaves out:
Speed and feed. The spindle is on or off. Spindle speed, feed rate and what they
do to the cut are on Feeds and Speeds.
The cut itself. No tool wear, heat, chatter, deflection or surface finish. A
drill cannot wander, so a hole started without a center drill is only noted, not drawn off
center.
Setups. One setup only: no turning the part around, no second operation, no
tailstock center supporting long work.
Threads and knurls, simplified. Thread pitch, crests and roots are checked along
the required length; pitch diameter and functional thread fit are not verified. The form knurl
illustrates displaced material and diameter growth. Its patterned section is schematic, not an
exact volume-conservation calculation. Form
and cut knurling use different mechanisms.
Its numbers. Inch only. The tolerances, the tool limits and the parts are teaching
values, not tool-maker data and not a standard.
Tolerances on the drawings mean the same as on Tolerances and Fits: a size on a
limit passes.
Follow tooth counts through pairs, idlers and compound trains; multiply stage efficiencies and choose gears to meet speed and torque limits.
Experiment: trade speed for turning effort; Interactive
Experiment: trade speed for turning effort
Predict first: a 20-tooth gear turns at 120 rpm clockwise and drives a 40-tooth gear. Will the larger gear turn faster or slower? What changes if you add an idler between them?
One mesh: 2:1 reduction
B has 2× as many teeth as A. Each passing tooth advances B by one tooth space; B turns 0.5× as fast, in the opposite direction. At 120 rpm clockwise and 2 N·m input, output B is 60 rpm CCW and 4 N·m ideally. The dashed circles are pitch circles, where the speeds match.
Add an idler: still 2:1
The idler changes direction twice: A → I → B. Its 30 teeth cancel from (30 ÷ 20) × (40 ÷ 30). At 120 rpm clockwise and 2 N·m input, output B is 60 rpm CW and 4 N·m ideally. The dashed circles are pitch circles, where the speeds match.
Two stages: 2 × 2 = 4:1
B and C turn together. Multiply the two reductions: (40 ÷ 20) × (40 ÷ 20) = 4. At 120 rpm clockwise and 2 N·m input, output D is 30 rpm CW and 8 N·m ideally. The two front-view meshes are separated on the page; the dashed connector represents their common shaft.
All three examples use the same ideal input: 120 rpm and 2 N·m. Input power is 25.13 W, and ideal output power is also 25.13 W. More output torque comes with less speed.
With each actual mesh at 96% efficiency, the pair delivers 3.84 N·m and 24.13 W; the idler train delivers 3.6864 N·m and 23.16 W; the compound train delivers 7.3728 N·m and 23.16 W. Their output speeds remain 60, 60 and 30 rpm. This comparison charges a loss at both idler meshes.
Count the teeth to find the speed
Count the teeth to find the speed
The driver supplies motion; the driven gear receives it. For one external gear mesh, each tooth passing on the driver advances the driven gear by one tooth space. A gear with twice as many teeth needs twice as many passing teeth to complete a turn.
This lesson uses reduction ratio i = input speed ÷ output speed, with speed magnitudes. A 2:1 reduction means two input turns for one output turn. Ratios above 1 reduce speed; ratios below 1 increase it. At zero speed, use tooth counts to find the ratio; do not divide 0 by 0.
Notation map: this page's original equations use N for teeth and n for speed. The course uses n for teeth, ω for angular velocity and τ for torque. Thus page N = course n; page speed n is written ω in the additions below. GR and i name the same reduction ratio.
In course notation, using magnitudes: GR = nout/nin = dout/din = ωin/ωout = τout/τin for an ideal single pair. Here d is pitch diameter; the torque equality assumes no losses. For a compound train, multiply the stage ratios; the first and last tooth counts alone do not give the overall ratio.
Supporting detail: Addendum 65 — Why tooth counts work, and what the dashed circles mean.
An idler and a compound shaft do different jobs
An idler and a compound shaft do different jobs
One mesh reverses
A → B turns the output opposite to the input, viewed from the same end of each shaft.
An idler cancels
A → I → B reverses twice. I affects its own speed and the spacing, but its tooth count cancels from the end-to-end ratio.
Compound stages multiply
A → B, then C → D, with B and C fixed together: i = (NB/NA) × (ND/NC).
Same shaft means the same angular speed. A large B and a small C fixed to one shaft make the same number of turns in the same direction. Their rim speeds differ because their radii differ. They do not mesh with each other.
In the compound example, 20 → 40 teeth halves the speed. A second 20 → 40 stage halves it again: 120 → 60 → 30 rpm. The two external meshes make the final output turn in the same direction as the input.
An idler is not another reduction stage. Drop its tooth count when finding the end-to-end ratio, but count every external mesh for direction: odd means opposite; even means the same, viewed from the same end of the parallel shafts. Do not drop a gear that shares a shaft with a different gear in a compound train.
More torque costs speed; it does not create power
More torque costs speed; it does not create power
Torque is turning effort. In this ideal, lossless train at steady load, output torque magnitude = input torque magnitude × reduction ratio. A 2:1 reduction changes 120 rpm and 2 N·m into 60 rpm and 4 N·m. The 4:1 compound example gives 30 rpm and 8 N·m.
Rotational power is torque × angular speed: P = T × 2πn/60 when T is in N·m and n is in rpm. Both sides of either example carry 25.13 W ideally. Real friction and other losses leave less output power. Tooth strength, bearings, shafts, lubrication and the motor all limit usable torque.
At zero speed, mechanical power is zero. A loaded, stationary gear train can still carry torque. The lab shows ideal static torque balance in that case; it does not model motor stall behavior or promise a safe load.
Supporting detail: Addendum 66 — Primary references and model boundaries.
Each stage receives only the power that survived the stage before it. Use decimal efficiencies and multiply: ηtotal = η1 × η2 × η3. Never add or average them. See Power and Efficiency for the general meaning of efficiency.
Power-flow schematic, with gear pairs shown by pitch circles. The stages are separated for clarity; each connecting shaft turns its attached gears together.
At 98% per stage, three stages give 0.98³ = 0.941192: about 5.88% total loss, not 6%. Fewer, higher-quality stages often beat many cheap stages for efficiency, provided they can meet the ratio and load requirements.
Supporting detail: Addendum 67 — Why adding losses is only an approximation.
Positive drives keep the speed ratio; losses reduce torque
Positive drives keep the speed ratio; losses reduce torque
A positive drive keeps input and output linked in a fixed ratio without slipping. Gears, chains and timing belts are positive drives. A slipping flat belt or V-belt is not. Follow Rotation and Linear Speed for the connection from rotation to belt travel.
At a stated steady input speed and load, use magnitudes:
ωout = ωin/i τout = τin × i × ηtotal Pout = ηtotal × Pin
The teeth still set the speed exactly in this model. Efficiency reduces output torque from its ideal value. It does not reduce output rpm as well. Use matching torque units on both sides: lbf·ft or N·m. The course conversion is 1 rpm ≈ 0.1047 rad/s when angular speed in rad/s is needed.
Supporting detail: Addendum 68 — Follow both stages of one loaded train.
Count stages for ratio, meshes for direction and stated losses
Count stages for ratio, meshes for direction and stated losses
For a simple train A → I → B, the idler cancels from (NI/NA) × (NB/NI) = NB/NA. The two external meshes still reverse direction twice.
Follow the loss assumption in the problem. Real meshes lose some power, but a problem may explicitly ignore it. If every actual mesh is 96% efficient, A → I → B has ηtotal = 0.96². If a whole stage including its idler is assigned 96%, use that stage value once. If η = 1.00 is stated, ignore losses. Do not infer the loss count from the reduction-stage count when an idler is present.
Choose teeth, then check both output limits; Checklist; Practice
Choose teeth, then check both output limits
Record input speed and torque, the output speed ceiling and torque floor, available integer tooth counts, maximum ratio per stage, and the loss assumption.
Use the speed ceiling to find i ≥ ωin/ωceiling. With at most R:1 per stage, k reduction stages can provide at most Rk:1.
For each candidate stage count k, calculate ηtotal and check the torque bound i ≥ τfloor/(τin × ηtotal). Use the larger of the speed and torque bounds.
Choose whole-tooth pairs within the range and per-stage limit. Connect successive pairs by shared shafts. Multiply their actual ratios; do not assume a rounded target is achievable.
Calculate output speed and torque, keeping full precision for both limit checks. Check direction by counting external meshes. An added idler may require recalculating losses.
To choose the faster of feasible designs at the same input speed, compare their actual overall ratios: the smaller one is faster. A feasible design is not automatically the fastest possible design.
Supporting detail: Addendum 69 — Worked design: a winch with a speed ceiling and torque floor.
The permitted size ranges on a hole and a shaft decide whether the two have clearance, interference, or could have either.
A size tolerance gives two limits
A size tolerance gives two limits
Manufacturing produces variation. For each diameter in this lab, the drawing gives
a size between two limits. A measured diameter inside that interval passes this
size check.
Nominal size
The size a feature is named by, such as Ø0.7500 in.
Ø means diameter. Every limit is measured from this number.
Deviation
How far a limit sits above (+) or below
(−) nominal. +0.0010 puts a limit 0.0010 in above it.
Limits
Nominal plus each deviation: the largest and smallest sizes that pass
inspection. Ø0.7500 +0.0010 / 0.0000 means anything from 0.7500 in to
0.7510 in.
Tolerance
The width of that range: largest limit minus smallest limit. Here,
0.0010 in.
A size exactly on a limit passes. On this page the limits are part of the
allowed range: a 0.7500 in hole and a 0.7510 in hole are both good. A 0.7511 in hole
is not.
Clearance is the hole size minus the shaft size
Clearance is the hole size minus the shaft size
Put a shaft in a hole and the number that decides the fit is the
difference between their two diameters.
clearance = hole size − shaft size
Positive: the hole is bigger. There is a gap, and the shaft slides in.
Negative: the shaft is bigger. That is interference: on paper
the parts overlap, so the shaft has to be pressed in, or it does not go at all.
Zero: exactly the same size, called line-to-line. No gap and no
overlap.
Clearance is measured across the whole diameter. With the shaft centred in the hole, the gap you
would see on each side is half of it:
gap on each side = clearance ÷ 2
The gap is drawn hugely enlarged. At the size of this drawing, a
0.00025 in gap would be far too thin to see. Hatching marks the cut face of the
block.
Three kinds of fit, decided by the two worst cases
Three kinds of fit, decided by the two worst cases
You do not have to check every pair of parts a drawing allows. Two pairs settle it:
the tightest and the loosest.
Tightest legal pair
The smallest hole with the largest shaft. Its clearance is
the least any good pair can have: hole minimum − shaft maximum
Loosest legal pair
The largest hole with the smallest shaft. Its clearance is the
most any good pair can have: hole maximum − shaft minimum
The fit is named by where those two numbers land.
Fit
Tightest pair
Loosest pair
What it means
Clearance
gap, or exactly zero
gap
No permitted pair has dimensional overlap. Equality has no gap.
Transition
overlap
gap
Some good pairs slide together; some have to be pressed.
Interference
overlap
overlap, or exactly zero
No permitted pair has a positive gap. Equality has no overlap.
The zero cases are deliberate. When the tightest pair comes out at
exactly zero, every good pair either has a gap or just meets, so this page counts it as a clearance
fit and the lab marks it zero at the limit. The same goes the other way for interference.
Higher means bigger. Each box is the range of sizes one part may
have; the dashed line is the nominal size. The hatched band in the middle panel is the range of
sizes both parts are allowed to have, which is what makes an overlap possible.
Fit lab: make two parts, then try them together; Interactive
Fit lab: make two parts, then try them together
Pick a drawing, then set the size each part actually came out at. Watch whether
each part passes inspection, and whether the two go together.
The section view is the pair you made, cut in half. The gold and blue bands
are where each wall is allowed to be.
The pair map is every pair the drawing allows. Anything above the diagonal
has a shaft bigger than its hole.
The lab’s example drawings. All are made-up teaching limits on a
Ø0.7500 in nominal, not values from any standard.
Drawing
Hole limits
Shaft limits
Tightest pair
Loosest pair
Fit
Slides freely
0.7500 to 0.7510
0.7480 to 0.7490
+0.0010
+0.0030
Clearance
Could go either way
0.7500 to 0.7510
0.7498 to 0.7506
−0.0006
+0.0012
Transition
Pressed together
0.7500 to 0.7510
0.7515 to 0.7525
−0.0025
−0.0005
Interference
Two good parts can still make a bad pair; Watch out
Two good parts can still make a bad pair
Checking each diameter against its own limits does not by itself check the
clearance of the selected pair.
Take the transition drawing from the lab: holes from 0.7500 to 0.7510 in, shafts from 0.7498
to 0.7506 in.
Hole made at 0.7502
Inside 0.7500 to 0.7510. Passes.
Shaft made at 0.7505
Inside 0.7498 to 0.7506. Passes.
Together
0.7502 − 0.7505 = −0.0003 in. The shaft is
bigger than the hole. They overlap.
Neither part is wrong. The drawing allows that pair, and that is what
transition fit allows. To exclude dimensional overlap for every permitted pair, the
tightest pair must have zero clearance or more. To guarantee a positive dimensional gap, it
must be greater than zero. Neither check by itself guarantees assembly: shape, surface finish,
temperature and the assembly method also matter.
Call the fit from the drawing alone; Interactive
Call the fit from the drawing alone
Read the two callouts, work out the tightest and loosest pairs, and call it before
you look.
Five drawings, each on a Ø0.7500 in nominal. Work out the limits, then the tightest and
loosest pairs. The answer follows each one.
The lab compares two diameters and nothing else. A real assembly also depends on:
Shape. No hole is perfectly round or straight, and no shaft is either. A shaft
can measure in tolerance everywhere a micrometer touches it and still be bent too far to go
in.
Temperature. Parts grow when they warm up, so the same part can measure
differently straight off the machine and an hour later.
Surface finish and measurement. A rough surface and an imperfect measurement
both blur the last decimal place the drawing asks for.
Force. Whether an interference pair can be pressed together, how hard, and what
that does to the parts are questions about the material, not about the two sizes.
The limits in the lab are made-up teaching values. Real drawings often take limits
from published tables of standard fits; those tables are not reproduced here.
A machined surface has a lay, a waviness and a roughness; a stylus trace and a filter turn them into Ra, Rz and Rt, and a drawing symbol says what is required and how the surface may be made.
A surface has lay, waviness and roughness
A surface has lay, waviness and roughness
Every manufactured surface carries marks from the process that made it. Surface
texture describes those marks: which way they run, how far apart they are, and how tall they
stand.
Roughness
The finest, closely spaced marks, such as the feed marks a cutting tool
leaves at each pass or the scratches of abrasive grit.
Waviness
A longer, gentler undulation that the roughness sits on, for example from
vibration or deflection while the part was made.
Lay
The direction of the main surface pattern, set by how the tool or the part
moved.
Form
The overall shape, such as a face that is slightly curved. Size and geometric
tolerances control form; surface texture is what is left on top of it.
Heights drawn 50 times larger than widths. A made-up
5.6 mm patch of surface with parallel marks 0.2 mm apart running front to back,
on a gentle waviness. At true scale the marks would be far too small to see. The lines are the
crests of the marks; the blue line is a stylus trace across them.
A stylus traces one line, and a filter splits it; Watch out
A stylus traces one line, and a filter splits it
A contact instrument drags a fine stylus along a straight line and records its height.
That record is the measured profile. Everything else is worked out from it.
Trace direction
Usually across the lay, so the stylus crosses the marks. If the
drawing specifies a measuring direction, that direction governs.
Mean line
A smooth line the filter draws through the measured profile, a weighted
average of the heights around each point. This mean line is the waviness.
Cutoff λc
The wavelength where the filter splits the profile. Undulations
much longer than λc stay in the mean line; marks much shorter than λc are left over
for the roughness. A wave exactly λc long is shared half and half.
Roughness profile
The measured profile minus the mean line.
Evaluation length
The part of the trace the numbers come from. In this lab it is five
sections, each λc long. The trace runs one more λc at each end so the filter has heights
on both sides of every evaluated point.
The same trace, before and after the filter, with λc 0.8 mm.
Top: the measured profile and its gold mean line; the shaded ends are the extra trace. Bottom: the
roughness profile over the five sections, with each section’s peak (▼) and valley
(△), its height as a red bar, and ±Ra dashed. Heights are drawn far larger than lengths;
each plot gives its factor at the right.
The filter decides what counts as roughness, not the size of the undulation.
Change λc and the same undulation can move from the waviness into the roughness. That is why
a measurement result needs its cutoff, and why a drawing or its standard says which one applies.
Ra, Rz and Rt read the roughness differently
Ra, Rz and Rt read the roughness differently
All three come from the same roughness profile. They answer different questions about
it. This lab uses the older ISO 4287 convention with five sections for Rz; the drawing-symbol
comparison separately shows older and current graphics.
Ra
The average distance of the roughness profile from its mean line over the
evaluation length. Peaks and valleys count alike.
Rz
The average of the sections’ peak-to-valley heights. Each section counts once, whatever is in it.
Rt
The largest peak-to-valley height over the whole evaluation length: the single
highest peak to the single deepest valley.
Ra = average of |roughness| over the evaluation length
Rz = (Z1 + Z2 + Z3 + Z4 + Z5) ÷ 5
For the trace above: Ra 1.55 µm; Rz = (6.55 + 6.56 + 6.70 + 6.57 + 6.87) ÷ 5 = 6.65 µm;
Rt 7.10 µm.
The values are in micrometres (µm). They can also be given in microinches (µin): one inch is
exactly 25.4 mm, so 1 µin is exactly 0.0254 µm, and Ra 1.6 µm is
63.0 µin.
Name the standard. Current documents use the ISO 21920 series. Many existing
drawings and instruments use the older ISO 4287 parameters and ISO 1302 symbols. The names
Ra, Rz and Rt carried over, but the standards are not identical, so this page says which one each
symbol comes from and does not mix them. For a real part, check which standard the drawing
invokes.
Surface texture lab: trace a surface and hold it to a drawing; Interactive
Surface texture lab: trace a surface and hold it to a drawing
Change the surface, the trace and the drawing, and watch the numbers. Start with
experiment 1.
The block is a few millimetres of surface. Lines are the crests of the marks;
the blue line is the stylus trace, solid over the evaluation length.
The plots are that trace: the measured profile with its mean line, then the
roughness profile with its five sections.
The drawing holds one requirement against this one trace and writes it as a
callout in the view you choose.
Parallel marks 0.2 mm apart, mark height 6 µm, λc 0.8 mm unless
stated. Teaching surfaces, not predictions for any process.
Experiment
Setting
Ra, µm
Rz, µm
Rt, µm
Across or along the marks
Trace left to right, across
1.55
6.65
7.10
Trace front to back, along
0.328
1.59
1.70
Where waviness ends
12 µm waviness, λc 0.8 mm
1.57
6.75
7.43
Same surface, λc 2.5 mm
2.11
11.3
11.8
One deep scratch
No scratch
1.55
6.65
7.10
Scratch three times the mark height
1.88
10.5
23.4
The same-Ra experiment is the figure in the next section.
Same Ra, different surfaces
Same Ra, different surfaces
These three roughness profiles all have Ra 1.6 µm. They are not the
same surface. All three plots use the same height scale.
Rounded valleys, sharp crests. Ra 1.6, Rz 6.86,
Rt 7.32 µm.Flat tops, narrow grooves. Ra 1.6, Rz 10.8,
Rt 11.2 µm. Most of the profile sits above the mean line.Narrow peaks, flat valleys. Ra 1.6, Rz 10.8,
Rt 11.2 µm. Most of the profile sits below the mean line.
Ra alone cannot tell these apart.
Rz separates the rounded marks from the other two, but not flat tops from narrow
peaks.
Where the extremes sit, above or below the mean line, separates those two.
One number does not describe a surface. That is why a drawing names the parameter, not just a
value.
Reading a surface texture symbol
Reading a surface texture symbol
The symbol’s point touches the line that represents the surface in a drawing
view. Its shape says how the surface may be made; the text beside it says what texture is
required.
The three process requirements, in the older and the current graphics. Schematic drawings:
the standards dimension the symbols exactly.
Symbol
Older, ISO 1302
Current, ISO 21920-1
Meaning
Basic
Any manufacturing process is permitted. This does not forbid machining; it simply does not require it.
Bar
Material removal is required: the surface must be produced by removing material, for example by machining.
Circle
Material removal is prohibited: the surface must not be machined or otherwise have material removed to meet this requirement.
Current or older
An ISO 21920-1 symbol carries a short separate dash above the
short leg, touching neither leg. The older ISO 1302 graphic has none, so the familiar older marks
are not the current symbols unchanged.
The requirement
The parameter and its limit, such as Ra 3.2, sit below the
long line, to the right of the long leg, in both.
The lay
In an ISO 21920-1 callout the process and the lay sit above the long
line. In the older symbols the lay symbol sits beside the long leg, below the line.
A complete specification can state more than this page draws, such as the measuring
direction. The process mark is a requirement about how the surface is made; a profile measurement
checks the texture, not whether material was removed.
Lay symbols belong to the drawing view
Lay symbols belong to the drawing view
A lay symbol says which way the marks run. = and ⊥
are judged against the plane of projection of the view the symbol is drawn in, not against the page or
the screen.
=Parallel. Parallel to the plane of projection of the view in which the symbol is used.⊥Perpendicular. Perpendicular to the plane of projection of the view in which the symbol is used.XCrossed. Crossed in two oblique directions relative to the plane of projection of the view in which the symbol is used.MMultidirectional. Multidirectional.CCircular. Approximately circular relative to the centre of the surface to which the symbol applies.RRadial. Approximately radial relative to the centre of the surface to which the symbol applies.PParticulate. Particulate, non-directional or protuberant.
The first two figures are the same surface. Its marks run front to back: parallel to the right
side view’s plane of projection, so = there, and perpendicular to the front
view’s, so ⊥ there. The arrow shows which way each view looks.
Lay and trace direction are different things. The lay says how the marks run. The trace direction
says how they were measured; it is usually across the lay unless the drawing specifies one.
The lay symbols and meanings here are those of ISO 1302:2002.
Older notation: N grades and triangle marks; Interactive
Older notation: N grades and triangle marks
Older drawings may give a roughness grade number or a row of triangles instead of a
parameter and value.
Roughness grade numbers
Each grade number stands for one Ra value. A grade is an Ra value: it does not convert to Rz.
Grade numbers and the Ra values they stand for. Microinches are exact conversions, rounded
for display.
Grade
Ra, µm
Ra, µin
N1
0.025
1.0
N2
0.05
2.0
N3
0.1
3.9
N4
0.2
7.9
N5
0.4
15.7
N6
0.8
31.5
N7
1.6
63.0
N8
3.2
126
N9
6.3
248
N10
12.5
492
N11
25
984
N12
50
1969
Triangle marks
One catalog prints a comparison of downward triangle marks with Ra values, under the heading
“Extract from JIS B 0601-2001”. It calls the comparison approximate, not rigorous.
It gives no acceptance limits, no dates for the marks, and no Rz equivalents, so read it as a rough
historical guide.
The catalog’s comparison, as printed.
Marks
Ra values listed, µm
4 triangles
0.012, 0.025, 0.05, 0.1, 0.2
3 triangles
0.4, 0.8, 1.6
2 triangles
3.2, 6.3
1 triangle
12.5, 25
One requirement, four ways
Slide the Ra limit and compare the current symbol, the older symbol, the grade number and the
triangle marks for the same value.
Read the callout; Interactive
Read the callout
Ten short questions on the symbols, the lay, the older notations and what the numbers
respond to.
The older ISO 1302 symbol with a circle in its V, and Ra 3.2. May this surface be machined to reach Ra 3.2 µm? No: material removal is prohibited. The circle in the V prohibits material removal. Material removal is prohibited: the surface must not be machined or otherwise have material removed to meet this requirement.
A current symbol with an open V, no connecting material-removal crossbar or circle, a separate identifying dash above the short leg, and Ra 6.3. Does this callout forbid machining the surface? No, any process is permitted. The V has no connecting material-removal crossbar or circle: it is the basic symbol. The separate floating dash identifies ISO 21920; it does not require machining. Any manufacturing process is permitted. This does not forbid machining; it simply does not require it.
A symbol with a bar closing its V, a short separate dash above the short leg, and Ra 0.8. Which standard's graphic is this? ISO 21920-1 (current). The short separate dash above the short leg, touching neither leg, marks an ISO 21920-1 symbol. The older ISO 1302 graphic has no dash.
A block of surface whose marks run front to back. An arrow shows the right side view looking at its right side. The marks run front to back. Which lay symbol belongs on this surface in the right side view? =. The right side view looks along the arrow. Marks running front to back are parallel to that view's plane of projection, so the symbol is =. The same surface takes ⊥ in the front view: the symbol depends on the view, not on the screen or page.
A block of surface whose marks run front to back. An arrow shows the front view looking at its front. The marks run front to back. Which lay symbol belongs on this surface in the front view? ⊥. The front view looks along the arrow. Marks running front to back are perpendicular to that view's plane of projection, so the symbol is ⊥. The same surface takes = in the right side view: the symbol depends on the view, not on the screen or page.
The grade number N8. An older drawing gives grade N8. What does that stand for? Ra 3.2 µm. N8 stands for Ra 3.2 µm (126 µin). Each step of the grade number doubles or halves the Ra value, near enough.
Three downward triangles. In the catalog comparison, which Ra values sit beside three triangles? 0.4, 0.8, 1.6 µm. Three triangles sit beside Ra 0.4, 0.8, 1.6 µm. The catalog calls this comparison approximate, not rigorous. Read it as a rough historical guide, not as acceptance limits.
A current symbol with Rz 6.3. Which N grade means the same as this Rz 6.3 µm requirement? None: grades are Ra values. Grade numbers each stand for an Ra value; N9 is Ra 6.3 µm, not Rz 6.3 µm. Ra and Rz measure different things, and how they compare depends on the shape of the profile, as the same-Ra experiment in the lab shows.
One deep scratch crosses this otherwise even trace. Which number has the largest proportional change? Rt. In the lab, a scratch three times the mark height takes Rt from 7.10 to 23.4 µm, Rz from 6.65 to 10.5 µm and Ra from 1.55 to 1.88 µm. Rt spans the highest peak to the deepest valley over the whole length. Rz averages five section heights, diluting localized extremes; filtering can spread a scratch across section boundaries. Ra averages every point.
A block of surface with parallel marks running front to back, and a stylus trace running left to right across them. The trace on this surface crosses the marks. Traced along the marks instead, Ra would be: Much smaller. In the lab, Ra is 1.55 µm across the marks and 0.328 µm along them: the stylus misses the marks when it runs beside them. Tracing across the lay is the usual choice, but a direction the drawing specifies governs.
What this lab leaves out
What this lab leaves out
The lab is a teaching model of one straight trace. A real measurement and a real acceptance
decision involve more:
The instrument. A real stylus has a tip radius, and instruments apply a
short-wavelength filter and remove the part’s form before filtering. The lab’s stylus is
sharp and its part is flat.
The filter and parameters. This lab follows the older ISO 4287 teaching
convention of five sections for Rz, with a Gaussian mean line having 50% transmission at
λc. Dense samples resolve the narrow marks; numerical integration on a coarser grid computes
the smooth mean line. This is not a compliant instrument or a complete ISO 21920 inspection
setup, and a drawing or standard can call for other settings.
The decision. The lab compares one trace with one limit. The drawing and its
standard say how many traces are taken, where, and how the results are judged.
The process. No trace shows whether material was removed. The process
requirement is checked against how the part was made.
Scratches and pits. Whether an isolated flaw counts against a texture
requirement is for the drawing and its standard to say.
Symbols. The drawings are schematic. The lay symbols are those of
ISO 1302:2002; the lab writes the same symbols in the ISO 21920-1 lay position. Check the
current standard before relying on one there. Inch-drawing (ASME) practice is not covered here.
Trace torque through a keyed shaft-hub joint, identify common key forms, and distinguish key fit from axial retention.
A key joins the shaft slot to the hub slot to transmit torque
A key joins the shaft slot to the hub slot to transmit torque
A key is a separate piece fitted between a shaft and the bore of a hub, the part of a pulley, sprocket or gear that fits around the shaft. Contact between the key and slot faces transfers turning force. Here, keyseat means the recess in the shaft and keyway means the matching slot in the hub. Other references sometimes use keyway for both; always identify which part is being cut.
Schematic parallel-key joint. Clearances are exaggerated; obtain dimensions and fits from the drawing.
Trace the turning load through the connection: shaft slot face → key → hub slot face, or the reverse when the hub drives the shaft. At the rim, the pulley, sprocket or gear transfers the load through its belt, chain or mating gear teeth. Not every hub is keyed; check how the actual assembly connects the hub to the shaft.
Identify the key by its cross-section and seating shape; Practice
Identify the key by its cross-section and seating shape
Key
Recognizing feature
Square parallel key
Equal width and height in cross-section; uniform section along its working length.
Rectangular parallel key
Unequal width and height; uniform section along its working length.
Woodruff key
A segment of a disk, with a curved bottom seated in a matching circular recess in the shaft.
Gib-head key
A tapered key with a projecting head that provides a place to engage it for removal.
Supporting detail: Addendum 70 — Public reference for key forms.
Check key fit and axial retention as separate requirements
Check key fit and axial retention as separate requirements
For a parallel key, the side fit controls how the turning load is carried. Excessive side clearance permits movement; an overly tight fit can prevent assembly. Verify the specified top clearance too. A tapered key has a different seating action, so do not apply a parallel-key fit rule to it.
Use the drawing's key designation, dimensions and tolerances. Do not choose a key size from shaft diameter alone or assume that half the key height belongs in each part.
The key does not by itself define how the hub is held against movement along the shaft. Check the assembly drawing for axial retention as well as torque transmission. A hub that turns with the shaft is not necessarily secured against sliding along it.
Bearings support the shaft relative to the housing; the keyed hub transfers torque between the shaft and the attached rotating part. For the bearing arrangement and the shaft's axial location, see Shaft Support and Axial Adjustment.
Choose a shaft or hub cutting method, then check keyseat or keyway location, width, depth and fit.
A key joins the shaft slot to the hub slot to transmit torque
A key joins the shaft slot to the hub slot to transmit torque
Before choosing a cutting method, identify which part needs the recess: the keyseat in the shaft or the keyway in the hub. See Keys and Keyways in Shaft-Hub Connections for the joint's terminology, diagram and torque path.
Identify the key by its cross-section and seating shape
Identify the key by its cross-section and seating shape
Use the specified key form and drawing dimensions to choose the feature you need to make. Review key forms, fit and axial retention before planning the cut.
Mill a shaft keyseat with the cutter that matches its end and bottom shape
Mill a shaft keyseat with the cutter that matches its end and bottom shape
An end mill can make a straight, flat-bottom shaft keyseat that ends within the shaft's length. Its diameter leaves rounded ends in plan view; confirm that the key and drawing allow that shape. Use a permitted entry method for the selected cutter.
A Woodruff keyseat cutter is a disk-shaped cutter on a shank. Its diameter forms the curved recess for a Woodruff key and its cutting width establishes the slot width. Match both dimensions to the required recess. Not every tool sold as a keyseat cutter is the correct tool for every key.
Read the keyseat's width, depth, length or cutter diameter, and axial location from the drawing.
Align the shaft axis with the intended machine travel and locate its centerline using Mill Indicators and Part Zero. Set the cutter's cutting plane or path on that centerline, with the specified axial location and depth reference.
Supporting detail: Addendum 71 — Why cutter selection includes more than width.
Broach a through hub keyway; consider slotting when the geometry needs it; Practice
Broach a through hub keyway; consider slotting when the geometry needs it
A conventional keyway broach and guide bushing are useful for a straight keyway through a hub bore. See Broaching for the pass sequence. The same arrangement does not make a closed-end keyseat on the outside of a shaft.
Shaping or slotting uses a reciprocating single-point tool to remove material in successive strokes. A suitable internal slotting setup can make a hub keyway when a through broach is unsuitable. Blind work still needs a planned tool runout and chip space; the method must fit the actual geometry.
Check width, depth and location before using assembly as a fit check; Checklist
Check width, depth and location before using assembly as a fit check
Clean the feature and remove burrs without rounding the functional edges.
Check width at several accessible positions using a suitable slot gauge, gauge blocks or measuring contacts. Select a tool with enough access and resolution for the tolerance; see Choosing Shop Measuring Tools.
Measure depth from the drawing's stated reference. A depth instrument needs a stable reference bridge or fixture on a round shaft; resting it arbitrarily on the curved surface changes the result.
For a hub, establish the specified bore-to-keyway measurement with an appropriate internal/keyway gauge or fixture. A reading across the bore and slot includes the bore dimension; it is not automatically the radial slot depth.
Check length, axial position and centerline location. Compare actual measurements with the drawing limits.
Check the specified key fit and assembly clearance. Investigate interference or unwanted movement instead of forcing the assembly together.
A key that enters the slot is a useful functional check, but it does not prove depth, location or dimensional compliance.
Machining/Production and Nontraditional Processes
Broaching
See how progressive teeth cut a profile, how a guided keyway broach works, and where rotary broaching differs.
A broach divides a shaped cut among a row of teeth
A broach divides a shaped cut among a row of teeth
In conventional broaching, a tool moves in a straight line relative to the work. Successive cutting teeth reach farther into the material. Rise per tooth is that increase in cutting height between successive teeth; it is not the tooth spacing along the tool. The spaces between teeth hold chips.
Roughing teeth remove most of the allowance. Semi-finishing teeth approach the final profile with smaller cuts. Finishing teeth establish size and finish; several may have the same height. The exact arrangement depends on the tool.
Schematic side profile. Tooth heights are exaggerated; the small teeth enter first.
Supporting detail: Addendum 72 — Why the finishing teeth need not keep getting taller.
Push or pull describes the drive; internal or external describes the surface
Push or pull describes the drive; internal or external describes the surface
A push broach is driven from behind; a pull broach is drawn through by its leading end. An internal broach works inside a prepared opening. An external broach works on an outside surface, such as a flat or an external spline. These are separate distinctions: internal does not mean pull-only.
A suitable hole broach can produce a square, hexagon or internal spline; a keyway broach cuts a slot in an existing bore. Conventional through broaching needs an exit path. Choose the tool from the required finished profile and starting-hole specification, not just a nominal shape name.
Supporting detail: Addendum 73 — Production advantages depend on the job.
A bushing guides the keyway broach; shims move its cutting edge outward
A bushing guides the keyway broach; shims move its cutting edge outward
The hub is the part surrounding the shaft. A slotted bushing fits its bore and guides the broach. A shim is a thin spacer behind the broach that increases the depth reached on a later pass. Broach, bushing and shims must be a compatible set.
Conceptual later-pass setup, not a construction drawing or a statement of shop inventory. The shim is behind the broach, opposite its teeth.
Check the drawing, prepared bore and required keyway. Select the matching broach, bushing and specified shim schedule.
Seat the bushing in the hub. Insert the broach in its slot for the first pass, normally without a shim on a multi-pass set.
Support the hub squarely on the press bed with clearance below for the entire broach to pass through. Align the ram, tool and guide.
Apply the specified lubricant and press the broach through according to its instructions.
After the completed pass, remove the broach and clean its teeth. If another pass is specified, insert the prescribed shim behind it, reinsert, lubricate and check alignment again.
Repeat with the specified shim stack for each remaining pass. Stop at the prescribed final setup; deburr and inspect as described in Machining Keyseats and Keyways.
Do not add a shim after the final pass. One-pass tools and some small broaches require none. Adding arbitrary shims is not a way to correct an uncertain dimension.
Supporting detail: Addendum 74 — Manufacturer reference for the pass sequence.
Count the first pass separately from the added shim passes; Practice
Count the first pass separately from the added shim passes
Rise per tooth and depth per pass are different quantities. For a simple planning exercise, let d1 be the depth reached by the unshimmed pass and s the extra depth from each added shim increment. After N passes, depth is d1 + (N − 1)s.
For a target depth D greater than d1, the minimum count to reach at least the target is 1 + ceiling((D − d1)/s). Ceiling means round upward to a whole number. If the first pass already reaches the target, the count is one. All depths here are measured radially from the original bore surface.
Example: an invented first-pass depth of 1.0 mm and two 0.5 mm increments give 1.0, 1.5 and 2.0 mm: three passes. This is counting practice, not a real broach specification. If rounding upward exceeds the allowed finished depth, that shim schedule cannot make the part to size.
Rotary broaching can make polygon forms such as a hex socket using a special holder on a suitable lathe or mill. A small angular offset makes successive portions of the tool's end engage as the tool and work move relative to one another. It does not use the long row of progressively taller teeth shown above.
It needs a prepared opening for internal work. Tool form, available thrust, material, depth and chip clearance limit the application; a blind socket needs a planned place for the chips. This is a process overview, not a setup procedure.
Supporting detail: Addendum 75 — How the rotary holder produces the cutting motion.
Machining/Drilling and Threading
Twist Drills and Drill Sizes
Identify the parts that cut and guide a twist drill, compare point shapes, and read the four common drill size systems.
The lips cut; the flutes give chips a way out
The lips cut; the flutes give chips a way out
A twist drill makes a round hole with cutting edges at its point. The lips run from the center region toward the outside diameter. The helical grooves are flutes: they provide space for chips and a route for cutting fluid. The web is the central core between the flutes.
The chisel edge crosses the center of a conventional point. It pushes material aside with substantial thrust. Narrow strips called margins run along the outside of the drill body and help guide it in the hole. The relieved body behind them reduces rubbing. The shank is the portion held by the machine.
The point does the cutting. The body supports it and provides room for chips. Views are schematic, not sharpening templates.
Point angle and split point describe different features
Point angle and split point describe different features
118° and 135° are included point angles, measured across the point in an axial view. A 135° point is flatter and shorter than a 118° point of the same diameter. A 118° point is common on general-purpose drills; 135° points are common on drills intended for tougher materials. Select the complete tool for the material and setup, not by angle alone.
A split point has extra grinding near the center to reduce the chisel edge and improve starting behavior. “135°” does not automatically mean “split point.” A split point is also not a guarantee that the hole will be located accurately without a suitable setup.
Supporting detail: Addendum 76 — Why a damaged or uneven point makes a poor hole.
Read the diameter system before selecting the drill
Read the diameter system before selecting the drill
System
What the marking means
Fractional inch
A diameter in fractions of an inch, such as 1/4 in. Convert the fraction when a calculation needs a decimal.
Number
A size label, such as No. 7. In the common numbered series, a larger number means a smaller drill. The number itself is not a diameter.
Letter
A size label from A toward Z, increasing in diameter. Look up the actual diameter; letters are not equal increments.
Metric
A diameter in millimeters, such as 5.0 mm. It is not interchangeable with a nearby inch size just because both fit in the chuck.
Read the marking and confirm the cutting diameter when identification is uncertain. A drill index is storage, not proof that someone returned the drill to the correct slot. Use the specified diameter and a checked chart when selecting a tap drill.
Match the shank to the holder and the hole to the job
Match the shank to the holder and the hole to the job
A straight shank fits an appropriate chuck or collet. A reduced shank is smaller than the cutting diameter; it does not increase the machine's drilling capacity. A taper shank fits its matching taper socket or approved adapter, rather than being gripped as though it were a straight shank.
Choose enough flute length and reach for the hole, with as little unnecessary projection as the setup permits. A drill's nominal diameter does not guarantee a finished hole's size, location or surface finish. Plan the drilling sequence, then choose a finishing operation if the drawing calls for it.
Machining/Drilling and Threading
Drilling on a Drill Press or Mill
Plan hole location, workholding, starting, chip evacuation and depth before drilling, then inspect and deburr the result.
Locate and secure the work before starting a hole
Locate and secure the work before starting a hole
A drill press feeds a rotating drill along its spindle axis. A mill adds controlled table positioning, which helps place a pattern of holes from drawing datums. On either machine, the hole follows the relationship between spindle, tool and workholding.
Clamp the work or hold it in a suitable vise or fixture rather than by hand; see Machine Hazards for why. Round stock needs support that prevents rolling, such as a properly clamped V-block arrangement. Provide clearance for a through drill without drilling into the table, vise or supports.
A spot starts the hole; a pilot removes the center material
A spot starts the hole; a pilot removes the center material
Spot drilling makes a short starting feature to help locate the next drill. A center drill is primarily designed to produce a center hole for supporting work between centers; its small pilot and conical portion make it different from a spot drill. Select a starting tool compatible with the following drill, rather than treating the two tools as interchangeable.
A pilot hole is a smaller hole drilled before a larger one. It can reduce the thrust required at a conventional drill's chisel edge. It is not automatically required: some drill geometries are intended to enter solid material. Follow the drill maker's starting-hole and pilot recommendations.
Supporting detail: Addendum 77 — Why “always center drill first” is too broad.
Choose speed and feed for this drill and material; Practice
Choose speed and feed for this drill and material
Use Feeds, Speeds and Removal Rate for the surface-speed-to-RPM calculation. For drilling, its diameter input is the drill's cutting diameter. Surface speed is a cutting condition; RPM is the spindle setting. They are not the same quantity.
Feed per revolution describes how far the drill advances each turn. Select cutting data for the tool material, work material, diameter, depth and coolant arrangement. A generic material table is a starting reference only when its tool assumptions match. Use an available machine setting consistent with the tool and machine instructions.
Peck drilling advances in stages with retractions to break or evacuate chips. The needed stroke and retraction depend on the drill, material, hole depth and fluid delivery; there is no universal peck distance. Do not add a peck cycle to a tool that calls for continuous feed without checking its instructions.
For a blind hole, distinguish the depth reached by the tip from the depth that has the full drill diameter. The conical point extends beyond the full-diameter portion. A tapped blind hole also needs space for the tap's lead-in and chip clearance. For a through hole, account for breakthrough and the drill point below the work.
If chip shape, sound, thrust or finish changes unexpectedly, stop and inspect the setup before resuming; use Metal Chips and Swarf for chip handling.
Supporting detail: Addendum 78 — Why changes in the cut matter.
Compare advance, retracing and withdrawal paths
Compare advance, retracing and withdrawal paths
Open G-code and Toolpaths. In the playground's Program picker, choose these named examples in order: Drilling path: continuous advance, Drilling path: short in-bore retract, and Drilling path: withdrawal above the mouth. Use Step to follow each G0 and G1 move.
Program
Motion to watch
What the model reports
Continuous advance
G1 Z-8
One feed enters new modeled stock to final depth.
Short in-bore retract
G0 Z-1, then G1 Z-4 and G1 Z-8
The first feed back removes no additional modeled stock; the separate deeper feed does.
Withdrawal above the mouth
G0 Z2, then G1 Z-4 and G1 Z-8
The first feed back removes no additional modeled stock; the separate deeper feed does.
Mark where the drill enters material that has not yet been removed. That is new cutting in this geometric model.
Mark where the drill returns through the bore it already made. That travel retraces the prior path and removes no additional modeled stock; the separate deeper feed begins after the point reaches the previous bottom. The model does not show possible wall or chip contact, rubbing, force or heat during the return.
Compare the retraction heights with the raw top at Z+1. The short retraction stays inside the bore. The other withdrawal carries the point above the hole mouth, creating a clearer opportunity for chips to leave the hole.
Compare the finished holes. The three illustrative programs intentionally reach the same final tip depth and geometry even though their travel paths differ.
The playground shows programmed motion and removed geometry. It does not model chip transport or packing, cutting heat or force, coolant delivery, or whether chips actually cleared. A retraction creates an opportunity for chip escape; it is neither proof of chip clearance nor the only way chips can leave. Haas describes its G83 return as a chip-clearing motion and places the R plane above the cut so chips can get out, but these G0/G1 illustrations are not G73, G83 or Q implementations.
The coordinates, spindle speed and feed in these three examples are comparison values for the simulator, not operating recommendations. Select any real drilling cycle, retract amount, speed and feed from the drill maker's guidance, the machine/control manual and the approved shop procedure. Continue to follow the machine instructions and Metal Chips for observation and chip handling.
Read point depth and full-diameter breakthrough separately
Read point depth and full-diameter breakthrough separately
Open Milling Operations and its simulator. Choose Job 2 / Pocket and holes. From View, choose 2D sections. Under Drill: choose a size, choose Load the drill Ø6.6. With the tool clear above the block, use the Move to X and Y inputs to place it at X 15, Y 30, then use the Z input and Move to advance to each row below. Watch the Front section and DRO, then scroll to Inspection of the part as it is now and find Ø6.6 through at X 15, Y 30.
These are illustrative coordinates in this simulator: its raw top is Z +1 and its stock bottom is Z −20. They are not operating depths, safe overtravel or an allowance for a real part.
Drill-tip DRO
What the Front section represents
Read the individual checks
Z −20
The point reaches the stock bottom, but the full drill diameter stops higher.
Diameter can read Ø6.60 while THRU says Not through yet: full size only to Z −18.02.
Z −21
The point extends below the stock and leaves a partial bottom opening; the full diameter still has not reached the bottom.
Diameter can still read Ø6.60 while THRU says Not through yet: full size only to Z −19.02.
Z −22
The modeled opening at the bottom reaches the inspection's minimum allowed diameter.
THRU changes to Through at the minimum allowed diameter.
Why does a drill-tip DRO of Z −20 not mean the full Ø6.6 body has reached Z −20?
Which inspection result changes between Z −21 and Z −22: the measured diameter, the THRU subcheck, or both?
Why can the left hole's diameter subcheck pass before its THRU subcheck?
Why can the feature row or overall job remain incomplete even after this THRU subcheck passes?
For this feature, THRU checks the bottom opening against a minimum allowed diameter of 6.50 mm. Passing it does not guarantee an exact Ø6.6 opening, and the unfinished top face, pocket or right-hand hole can keep other checks incomplete. Read the Diameter and THRU subchecks separately; do not use the overall job result as the answer to this comparison.
Check the hole before calling it finished; Checklist
Check the hole before calling it finished
Confirm the hole location, required diameter and whether it is through or blind.
Check the tool, starting method, workholding, exit clearance and depth plan.
Set suitable speed, feed and chip-evacuation conditions.
After drilling and stopping the machine, inspect the hole and remove burrs without changing the specified feature.
Measure the required dimensions and complete any specified finishing operation.
Deburring removes the raised edge left by cutting. It does not automatically mean making a large countersink. Preserve the drawing's edge requirements and inspect the exit side too.
Machining/Drilling and Threading
Reaming, Boring and Hole Seats
Choose a finishing operation by the feature needed: a sized bore, a cylindrical recess, a conical seat or a flat bearing surface.
Choose the operation from the finished shape
Choose the operation from the finished shape
Section views through the hole axis. Gray is material; white is open space. A spot face provides a bearing surface; a counterbore provides a recess.
Needed feature
Operation
Improve the diameter and finish of a nearly correct hole
Ream
Enlarge an existing hole with control of its axis and diameter
Bore
Cylindrical recess with a flat shoulder for a fastener head
Counterbore
Conical seat for a matching flat-head fastener
Countersink
Local flat bearing surface around a hole
Spot face
Ream for size and finish; bore when the axis needs control
Ream for size and finish; bore when the axis needs control
A reamer removes a small amount of material with multiple cutting edges. Prepare an undersize hole with enough material for the reamer to cut, rather than rub. Too much allowance overloads the tool; too little may leave the previous surface. Use the reamer maker's allowance for that diameter and material, not one stock allowance for every hole.
A reamer tends to follow the existing hole. Do not rely on it to relocate a badly misplaced or crooked hole. Boring enlarges an existing hole with a cutting edge whose path can be set relative to the machine axis. With a suitable setup and enough material left to clean up, it can correct location and straightness as well as diameter.
Use Tolerances and Fits to interpret the size limits and select inspection methods. See Lathe Operations for boring and threading on the lathe.
Supporting detail: Addendum 79 — A nominal tool size is not a finished-hole guarantee.
Compare the old opening with the final bore envelope
Compare the old opening with the final bore envelope
A centered final bore can remove material around a smaller opening when that entire opening lies inside the final bore's removal circle. It cannot add material back where an earlier opening protrudes outside that circle. Setting the boring head to the drawing diameter does not erase the earlier geometry.
Open Milling Operations and choose Job 4 / Boss and bore. From View, choose 2D sections. Before each case, choose Start this job over, then Sure? Start over, so each comparison begins with a fresh block. Use the Move to X, Y and Z inputs and Move. Keep tool and boring-head changes at Z +20 by using Raise to Z +20.
Fresh-block case
Illustrative simulator moves
Individual bore inspection
Previous opening stays inside
Load the drill Ø10. At X 50.5, Y 30, move the tip to Z −24. Raise to Z +20. Load the boring head Ø20.00, move to X 50, Y 30, then bore to Z −21.
The modeled Ø20 bore contains the earlier opening. Its row reports measured Ø20.00 through, with no diameter-too-big, centre-position or CUT TOO FAR failure; the unfinished whole job does not need to pass.
Previous opening extends outside
Load the drill Ø10. At X 52, Y 30, move the tip to Z −24. Raise to Z +20. Load the boring head, enter 18 in Boring Ø and choose Set; bore at X 52, Y 30 to Z −21. Raise, set Ø20, then bore at X 50, Y 30 to Z −21.
CUT TOO FAR — Bore Ø20.00 ±0.05 through reports about Ø20.41, 0.41 too big, with the fitted centre about 0.38 from target. THRU also fails because inspection tests a full allowed diameter around the fitted irregular opening; it does not mean the centered boring-head move stopped short. The centered final pass cannot replace the material removed outside its circle.
In the first case, where is the edge of the Ø10 opening relative to the centered Ø20 removal circle?
In the second case, which individual checks show that the result is not simply a centered Ø20 hole: diameter, position, overcut, or all three?
Why does a final DRO of X 50, Y 30 not undo the earlier off-axis removal?
These coordinates, depths, offsets and fitted/grid inspection numbers belong only to this idealized simulator. They are not acceptable cleanup allowance, safe boring offset, rework instructions or physical gauge results. The model does not predict boring-bar entry or interrupted-cut behavior, rigidity, deflection, chatter, force, finish, alignment uncertainty or workholding. Geometric containment is necessary here, not sufficient for a real process: use the drawing, measured hole, complete setup, tool-maker guidance and responsible process authority.
The top, boss and chamfers remain unfinished in both cases, so the overall feature row or job may still be incomplete. Read the bore's diameter, position and overcut checks for this comparison.
Match the head seat to the fastener drawing
Match the head seat to the fastener drawing
A counterbore has a cylindrical wall and a flat shoulder. A countersink has a conical seat. The screw head must match that seat's included angle; the thread diameter alone does not determine the head angle or the countersink diameter.
A spot face removes enough surface material to create a local flat seat for a head, nut or washer. It is useful on a rough or uneven surface. A spot face and a counterbore may use similar tools, but their required depth and purpose differ.
Read the specified diameters, depths and angle before cutting. A tool used only to break a sharp edge is not necessarily producing a fastener seat to size. Do not substitute a preferred cutter angle for the drawing requirement.
Test the choice before choosing the cutter; Practice
Test the choice before choosing the cutter
A smooth pin hole already on location suggests reaming. A hole whose axis must be corrected suggests boring, if there is enough material. A socket-head screw recessed into a cylindrical pocket suggests counterboring. A conical flat-head seat suggests countersinking.
Decode inch and metric thread names, distinguish thread fit from strength, and choose fastener features by their function.
Read diameter, spacing, series and class separately; Practice
Read diameter, spacing, series and class separately
In 1/4-20 UNC-2B, 1/4 is the nominal major diameter in inches, 20 is the number of threads per inch, UNC is the Unified coarse series, and 2B identifies the internal thread's fit class. The major diameter is the largest thread diameter; it is not the diameter of the hole drilled before tapping.
Read the whole callout. Matching only the nominal diameter is not enough.
UNF identifies the Unified fine series. At a given nominal diameter, fine threads have more threads per inch and therefore a smaller pitch than coarse threads. In M6 × 1, M identifies a metric thread, 6 is the nominal major diameter in millimeters, and 1 is the pitch in millimeters. It does not mean one thread per inch.
Unified external thread classes are 1A, 2A and 3A; internal classes are 1B, 2B and 3B. A means external, as on a screw; B means internal, as in a tapped hole. Class 1 permits a looser fit, class 2 serves general applications, and class 3 uses closer dimensional control.
The class number is not a material grade or a load rating. A class 3 thread does not by itself make a stronger bolt. Read the drawing for the required class and use the appropriate inspection limits. M6 × 1 alone does not specify a metric tolerance class; do not append an inch A/B class to it.
For the general distinction between a nominal dimension, its tolerance and a mating fit, see Tolerances and Fits. Exact thread acceptance limits require the applicable specification and gauges.
Pitch is spacing; lead is travel per revolution
Pitch is spacing; lead is travel per revolution
Pitch is the axial spacing between adjacent thread crests. For an inch thread, pitch in inches = 1/TPI. A 20-TPI thread has a pitch of 0.050 in. Lead is the axial travel in one complete relative revolution.
On a single-start thread, lead equals pitch. On a multiple-start thread, lead equals pitch times the number of starts. Right- and left-hand threads move in opposite axial directions for the same rotation. Look for a left-hand designation rather than forcing a fastener that does not start normally.
The lathe topic owns the process of single-point thread cutting. The tapping topic explains preparing and making internal threads with taps.
Choose the fastener by how it locates and clamps the parts
Choose the fastener by how it locates and clamps the parts
Part
Function and selection question
Bolt or screw
A bolt commonly passes through clearance holes and uses a nut; a screw commonly engages a tapped part. Names overlap in catalogs, so check the actual product specification.
Cap screw
A headed screw, often with a hex or socket drive. Choose the head clearance, bearing seat, thread and specified material properties, not just appearance.
Set screw
Its point bears on a mating part, such as a shaft, to resist relative motion. Point shape and the mating surface matter; it is not simply a small clamping bolt.
Stud
A headless threaded member, often installed in one part with a nut on the exposed end. Check which end engages the part and how much usable thread remains.
Nut
Provides the mating internal thread. Match thread size, pitch and required properties; head shape alone does not establish compatibility.
Washer
Provides a bearing surface beneath a head or nut. Choose its size, hardness and type for the joint rather than assuming every washer also locks it.
Threaded insert
Provides an internal thread in a parent material or repairs a damaged thread. Follow that insert system's hole, tap and installation requirements.
Engagement, grade and locking are separate decisions
Engagement, grade and locking are separate decisions
Engagement length is the axial length of mating threads carrying load. There is no single diameter multiplier that establishes adequate engagement for every combination of screw and parent material. Use the joint design or fastener-system requirements, and ensure the screw does not bottom before it clamps.
Strength identification must come from the specified fastener grade or property class and its traceable markings or documentation. Thread fit class and head style are not substitutes. Do not assign a strength grade to an unidentified fastener from its appearance.
A flat washer is not a locking device. A lock washer or thread-locking compound is not a guarantee against loosening in every joint. Use the specified locking method; for a compound, follow its compatibility, preparation, cure and removal instructions.
Supporting detail: Addendum 80 — Check the entire joint before substituting a fastener.
Machining/Drilling and Threading
Tap Drills and Tapping
Estimate a cutting-tap drill, distinguish thread percentage from engagement length, and plan tap geometry, alignment and depth.
Subtract the pitch to estimate a cutting-tap drill; Practice
Subtract the pitch to estimate a cutting-tap drill
A tap makes an internal thread in a prepared hole. Start by reading the full thread callout. For common straight Unified and metric cutting taps, a useful roughly 75% thread estimate is:
Inch: drill diameter ≈ nominal major diameter − 1/TPI. Both diameters are in inches; TPI means threads per inch.
Metric: drill diameter ≈ nominal major diameter − pitch. Both dimensions are in millimeters.
This is an estimate, not a final hole specification. Use the major diameter, not the minor or pitch diameter. Select the actual drill from the tap maker's chart for the material and required thread. Do not apply this cutting-tap estimate to forming taps, pipe threads or insert-installation taps.
Thread
Estimate
Cutting-tap chart comparison
1/4-20
0.250 − 1/20 = 0.200 in
No. 7, 0.2010 in
M6 × 1
6 − 1 = 5.00 mm
5.00 mm
3/8-16
0.375 − 1/16 = 0.3125 in
5/16 in, 0.3125 in
Supporting detail: Addendum 81 — Chart basis and the approximation's limits.
Thread percentage is not percent of bolt strength; Practice
Thread percentage is not percent of bolt strength
Percent thread describes how much thread height the hole permits relative to a reference profile. With the same cutting tap, a larger prepared hole leaves less thread height; a smaller hole leaves more material for the tap to remove and generally increases tapping torque.
It does not describe how far the screw enters the hole. That is engagement length. Nor does 75% thread mean 75% joint strength. Material, engagement length, thread geometry and loading all affect the joint. More thread height is not automatically a better choice.
Supporting detail: Addendum 82 — Keep percentage and fraction separate.
Taper, plug and bottoming describe the tap's lead-in
Taper, plug and bottoming describe the tap's lead-in
The chamfer is the gradual lead-in where the first teeth progressively cut the thread. A taper tap has a long lead-in that helps starting; a plug tap has a shorter lead-in; a bottoming tap has the shortest of the three and reaches full thread closer to the bottom of a blind hole.
A bottoming tap still has a lead-in. It cannot produce full thread all the way to the physical bottom. For hand tapping, start with a tap that can establish alignment, then use a bottoming tap when the required depth calls for it. Leave room for incomplete threads, chips and the drill point.
Chamfer style is separate from chip direction. A spiral-point cutting tap generally pushes chips ahead and needs exit space; a spiral-flute cutting tap can draw chips back toward the entrance. Select a tap intended for the material and through or blind hole.
Use a guide to start a hand tap square; Checklist
Use a guide to start a hand tap square
Confirm thread size, tap type, drill size, actual hole size and required full-thread depth.
Prepare the hole entrance as specified. Clamp the work or hold it in a vise rather than by hand; see Machine Hazards for why. Choose the tap maker's recommended lubricant for the material.
Align the tap with a suitable tapping guide. With a spindle used as a guide, keep it unpowered and follow the approved hand-tapping setup.
Turn with a suitable tap wrench while maintaining alignment. Follow the tap's chip-breaking instructions; do not force a tap whose resistance rises unexpectedly.
Keep within the planned travel, withdraw appropriately, clean and inspect the thread with the specified gauge or inspection method.
To help prevent broken taps, use the correct tap drill, maintain alignment, clear chips as the tap's instructions specify, use the recommended lubricant and do not force the tap. Reverse-to-break-chip routines depend on tap design; do not apply one hand-tapping rhythm to every tap.
Power tapping needs a machine and holder designed for it
Power tapping needs a machine and holder designed for it
Machine tapping coordinates spindle rotation with axial travel at the thread's lead. Depending on the machine, that may use synchronized rigid tapping or a suitable tapping attachment. The machine, holder, tap, reversal and depth control must work together. An ordinary drill-press setup is not automatically a power-tapping setup.
A cutting tap removes chips. A forming tap displaces suitable ductile material to make the thread without cutting chips. Forming taps require their own, generally larger, prepared-hole sizes and suitable lubrication. A cutting-tap drill can be much too small for a forming tap of the same thread size.
Use the machine's instructions and the approved shop procedure for tapping equipment.
Supporting detail: Addendum 83 — Why the local setup decides what is permitted.
Machining/Turning
Turning Tapers
Calculate taper rate, compound angle and tailstock offset, then choose a method and check both slope and size.
A taper changes diameter at a constant rate
A taper changes diameter at a constant rate
A straight taper is a cone, or part of one. Name the large diameter D, small diameter d and axial length of the tapered portion l. Use the same length unit for all three.
Taper rate = (D − d) / l. With inch dimensions, this is taper per inch, often abbreviated TPI. Here TPI means diameter change per inch of length; in Single-Point Threading, the same abbreviation means threads per inch. Write the units to distinguish them.
TPF = 12 × TPI, and TPI = TPF / 12. Taper per foot describes the diameter change over twelve inches; the actual tapered portion need not be a foot long.
Supporting detail: Addendum 84 — Check the rate with a fresh example.
Set the compound to the half-angle measured from the spindle axis
Set the compound to the half-angle measured from the spindle axis
The tool travels along one side of the cone. Its angle to the spindle axis is the half-angle α; the angle between the two sides is the included angle 2α.
α = atan((D − d) / (2l)) = atan(TPI / 2). Use degree mode for a setting in degrees. A rate of 0.050 in/in gives α = 1.432°, while the included angle is 2.864°.
Swivel the compound so its travel makes that half-angle with the spindle axis, then feed along the compound. The taper length is limited by compound travel. Check how the lathe labels its scale: a scale measured from cross-slide travel uses the complementary angle, 90° − α.
The radius changes by half the diameter change. The drawing exaggerates the slope.
Supporting detail: Addendum 85 — Why the factor of two belongs in the angle formula.
Tailstock offset uses the full supported work length
Tailstock offset uses the full supported work length
For a shallow external taper turned between centers, offsetting the tailstock tilts the work while the carriage still travels parallel to the bed. The usual small-angle setup estimate is offset ≈ TPI × L / 2 = (D − d)L / (2l).
L is the total work length between the effective center support points, approximated by the total part length in elementary problems. It is not just the length l of the tapered portion. Center-hole geometry matters in an actual setup. Use inches for L when the taper rate is in in/in; if given TPF, first divide it by 12.
Plan view: offset is transverse to the spindle axis, not a vertical change in center height.
For a rate of 0.050 in/in and L = 8.000 in, the estimate is 0.200 in even if only 3.000 in is tapered. Machine offset capacity and the center arrangement limit this method. Check a trial cut and restore center alignment before straight turning.
Supporting detail: Addendum 86 — Why this is an estimate rather than an exact setting.
Choose the taper method for length, access and support
Choose the taper method for length, access and support
Compound rest
Useful for short external or internal tapers. The compound guides the tool at the half-angle, but its limited travel restricts length.
Tailstock offset
Useful for shallow external tapers between centers. It changes the work axis and does not provide a method for boring an internal taper.
Taper attachment
A guide coordinates cross-slide motion with carriage travel. It can produce internal or external tapers within its travel and angle limits while the work stays aligned with the spindle. Connection and cross-feed arrangements depend on the attachment.
Form tool
A shaped cutting edge produces a short taper by feeding across the axis. A wide edge engages much of the cut at once, increasing cutting load and the tendency to chatter.
A matching taper needs both the right slope and the right size; Practice
A matching taper needs both the right slope and the right size
Measure diameters at two known axial positions to check the rate. A correct rate can still leave the entire taper too large or too small. A suitable taper plug or ring gauge checks the mating geometry; contact pattern and seating position answer different questions.
Morse tapers are commonly used for centers and tool shanks. Jacobs tapers commonly join a drill chuck to its arbor. They are separate families: identify the exact mating designation before selecting a tool or adapter. No dimensions or angles should be inferred from the family name alone.
No standard taper table is reproduced here. Use the specified component drawing and verified manufacturer or standard dimensions for a real fit; see Tolerances and Fits.
Read thread geometry, distinguish pitch from lead, understand compound infeed and check thread size beyond the outside diameter.
Pitch spaces the teeth; lead sets carriage travel per revolution
Pitch spaces the teeth; lead sets carriage travel per revolution
A thread follows a helix around a cylinder. Pitch P is the axial distance between corresponding points on adjacent teeth. For N threads per inch, P = 1/N inch. A 20-per-inch thread has a pitch of 0.050 in.
Lead is the axial advance along one helix in one complete revolution. A single-start thread has one helix, so lead equals pitch. A two-start thread has two interleaved helices and lead equals 2P. In general, lead = number of starts × pitch. Threading feed follows lead; the examples here use single-start threads.
The crest is the top of a tooth and the root is the bottom of the groove. The sides are flanks. The thread angle is the included angle between flanks in an axial section. This page discusses 60° profiles; other thread forms need other tools and geometry.
Schematic external thread, with only the upper half shown. Diameters span across the axis to the opposite side.
Major diameter alone cannot tell you whether a thread fits
Major diameter alone cannot tell you whether a thread fits
Major diameter
The largest thread diameter: across external crests or internal roots.
Minor diameter
The smallest thread diameter: across external roots or internal crests.
Pitch diameter
The diameter of an imaginary cylinder at which thread tooth width and groove width are equal for the ideal profile. It strongly affects flank clearance and fit.
The outside of a screw can measure correctly while its flanks are still too thick. Use the thread designation and its specified limits, not just nominal outside diameter. Tolerances and Fits explains limits; Lathe Operations explains what the simulator's simplified thread check measures.
The tool stays square while the compound feeds near one flank
The tool stays square while the compound feeds near one flank
For a 60° thread, use a tool with the required 60° profile and the specified tip shape. A center gauge, also called a fishtail gauge, checks a ground tool's angle and helps align its bisector perpendicular to the work axis. Set the cutting edge at spindle center height. Swiveling the compound does not mean turning the tool away from square.
A common manual arrangement sets compound travel about 29–30° from the radial cross-feed direction for this thread form. That is about 60–61° from the spindle axis. Check the scale's zero reference before setting it.
The angle describes the infeed path between passes, not the tool's included angle or the thread helix angle.
Near-flank infeed puts most of the cutting on one edge. Exactly 30° is parallel to a flank in the ideal axial profile; slightly less also gives the other edge a small cut. Tool-maker guidance may call for a different infeed strategy.
Supporting detail: Addendum 87 — Radial depth and compound travel are different quantities.
Keep every pass synchronized with the same helix; Checklist
Keep every pass synchronized with the same helix
The gearbox establishes the relationship between spindle rotation and leadscrew rotation. Closing the half-nuts connects the carriage to the leadscrew so that each spindle revolution advances the tool by the selected lead. An ordinary power-feed setting is not a substitute for a threading setting.
Check the machine's threading-dial chart before re-engaging the half-nuts. On an inch-leadscrew lathe, the engagement pattern depends on how the dial is graduated. With four numbered lines and unnumbered half-lines between them, common on older inch lathes with an 8-TPI leadscrew, an odd number of threads per inch (TPI) can usually be caught on any numbered line, but not on the half-lines (see LeBlond's four-numbered-line chart). On dials numbered 1–8, odd TPI are often restricted to lines of one parity (all odd or all even). Even TPI can usually be caught on any line, numbered or half. Fractional TPI may require returning to the same specified line or keeping the half-nuts engaged. These are examples, not universal dial settings: leadscrew pitch and dial gearing determine which marks work, and the machine's chart controls.
Supporting detail: Addendum 88 — Why only certain threading-dial positions return to the groove.
For some combinations, notably many metric threads on inch-leadscrew lathes, ordinary dial re-engagement will lose synchronization. Such setups commonly keep the half-nuts engaged and use the maker's prescribed return method. Confirm the machine and workholding permit that method; this page supplies no universal reversing procedure.
Identify the thread form, hand, pitch, number of starts and required limits on the drawing.
Check the machine chart for the gearing and permitted half-nut/dial procedure.
Check a light scratch pass with a pitch gauge before deepening the thread.
Plan the end of each pass and the return path before cutting to depth.
Measure as the thread approaches size; use the specified inspection method to decide when it is finished.
Calculate a profile depth only after naming the profile; Practice
Calculate a profile depth only after naming the profile
A 60° angle and a pitch do not, by themselves, specify a finished thread depth. Crest flats, root shape, tip geometry and dimensional limits all matter. Use the actual thread specification and tooling data for a production cut, then measure the result. Total depth is not a per-pass infeed instruction.
For the ideal sharp-V teaching profile used in the practice bank, radial depth is H = P / (2 tan 30°) = (√3 / 2)P ≈ 0.8660254P. This is the full sharp triangle height, not a finished Unified or metric thread depth. Actual truncated or rounded profiles remove less radial material than this sharp triangle.
Supporting detail: Addendum 89 — Verify the formula without a thread-depth table.
A pitch gauge checks spacing; wires help measure pitch diameter
A pitch gauge checks spacing; wires help measure pitch diameter
A pitch gauge has toothed leaves that compare with the thread spacing. It identifies pitch and helps detect a wrong gearbox selection, but it does not certify pitch diameter or fit.
In the three-wire method for an external thread, two calibrated wires sit in grooves on one side and a third on the opposite side. A micrometer measures over them. Because the wires contact the flanks, the measurement can be converted to pitch diameter using the correct relation for the thread form and wire size.
Use calibrated wire sizes and their applicable instructions or a verified table. Wire diameter, thread angle and lead-angle effects matter; this page does not supply an unchecked wire formula. A thread micrometer uses dedicated contacts for a similar purpose. Appropriate thread plug or ring gauges check functional acceptance against the specified gauge limits.
Trying a mating nut may reveal gross interference, but it does not establish conformance to a thread class.
Machining/Sawing and Grinding
Grinding Wheels and Abrasives
Match the abrasive to the material, distinguish grit from grade, and read the five main fields of a conventional wheel specification.
A grinding wheel cuts with many small grains
A grinding wheel cuts with many small grains
A bonded wheel holds abrasive grains in a supporting material called the bond. The grains act as small cutting edges. Spaces between them provide room for removed material and, in wet grinding, fluid. Grains can fracture or pull free as the wheel works; dressing restores a useful cutting surface when that renewal is insufficient.
Choose the complete wheel for the machine, material and operation. Abrasive type alone does not establish suitability: shape, dimensions, bond, speed rating and permitted grinding surfaces also matter. See Offhand Grinding and Precision Grinding.
Match the abrasive to the work material
Match the abrasive to the work material
Abrasive
General starting applications
Aluminum oxide
Steels, including high-speed steel (HSS) tool bits. Different formulations serve different grinding conditions.
Silicon carbide
Cast iron, some nonferrous materials, and brittle nonmetals; also conventional carbide-tool grinding with a suitable wheel.
Cubic boron nitride (CBN)
A superabrasive used for hardened steels and HSS.
Diamond
A superabrasive used for cemented carbide, ceramics and other hard nonferrous materials. It is not the general choice for steel grinding.
These are selection starting points, not permission to use any wheel containing that abrasive. In particular, a standard bench wheel for steel is not a wheel for aluminum. Confirm the manufacturer's application listing.
Supporting detail: Addendum 90 — Manufacturer resources for abrasive selection.
Grit is grain size; grade is how firmly grains are held
Grit is grain size; grade is how firmly grains are held
Grit size: within the same conventional grit-number system, a lower number means coarser grains and a higher number means finer grains. Coarse grit generally provides more space for chips and favors material removal; fine grit generally favors smaller scratches. Finish also depends on dressing, feeds, wheel condition and the work material.
Grade: the wheel's resistance to releasing abrasive grains. A hard grade holds grains more strongly; it does not mean that the abrasive mineral itself is harder. A hard work material may need a softer grade so dull grains can release.
Structure: the spacing of the grains. An open structure provides more space between grains than a dense structure. Do not confuse spacing with grain size.
Bond: vitrified bonds are glasslike; resinoid bonds use resin. Rubber and metal bonds serve other applications. Bond choice affects wheel behavior, dressing and permissible operating conditions; no bond name establishes a universal RPM rating.
Read the conventional marking from abrasive to bond; Practice
Read the conventional marking from abrasive to bond
The usual core sequence is abrasive → grit → grade → structure → bond. Prefixes and suffixes may identify a manufacturer's particular abrasive or bond. Some markings omit a structure field. Superabrasive specifications may include concentration and use a different format.
Read the meaning of each field before interpreting a manufacturer's code.
For a teaching example, suppose a supplier defines A 60 J 7 V as aluminum oxide, grit 60, grade J, structure 7 and vitrified bond. The 60 is not an RPM limit; the J is not the hardness of aluminum oxide. Use that supplier's key to interpret J and 7 rather than assuming all manufacturers' grades behave alike.
Supporting detail: Addendum 91 — A manufacturer's explanation of the marking.
Understand bench and pedestal grinder setup, wheel condition, tool-bit shaping and the difference between rim speed and a wheel's maximum RPM.
Offhand grinding guides the work by hand
Offhand grinding guides the work by hand
A bench grinder mounts on a bench; a pedestal grinder stands on a base. In offhand grinding, the operator guides the work against the wheel, usually supported by a tool rest. This suits shaping and sharpening; it does not provide the controlled workholding and feeds of Precision Grinding.
For a conventional straight bench wheel, the intended grinding surface is usually its outer circumference. Move the work across that surface to distribute wear. Use only the surfaces approved for that wheel and machine; a wheel designed for peripheral grinding is not automatically suitable for side grinding.
Wheel wear changes both rest and tongue-guard clearances
Wheel wear changes both rest and tongue-guard clearances
The tool rest supports the work near the wheel. The tongue guard is the adjustable part of the surrounding guard above the working opening; it is not the transparent eye shield. As the wheel becomes smaller, both clearances need checking and adjustment to the applicable machine instructions and requirements.
Positions only, not a scale drawing or a gap-setting gauge. The eye shield is omitted for clarity.
Make these adjustments with the wheel stopped and the machine secured according to its operating procedure. Recheck after dressing or wheel replacement. Use the specified clearance limits, not the proportions in the drawing.
Supporting detail: Addendum 92 — Where to check the clearance requirements.
Inspection, ring testing and dressing do different jobs
Inspection, ring testing and dressing do different jobs
Inspection checks condition and suitability before mounting. A ring test is an additional pre-mounting check for applicable vitrified wheels: a light nonmetallic tap produces a sound that a trained person evaluates for indications of cracking. Wheel size, shape and composition limit where the test applies. A clear ring does not replace the other checks.
Truing restores the required wheel shape and concentric running. Dressing renews the cutting surface by removing unwanted material and exposing useful abrasive. An appropriate dressing operation may do both. Use the dresser and method specified for the wheel.
Loading means work material fills the wheel's spaces. Glazing means dull abrasive remains at the surface and rubs instead of cutting effectively. Either condition calls for correction, not simply more pressure.
Supporting detail: Addendum 93 — Why a wheel can glaze, and where ring testing applies.
Shape an HSS tool bit without overheating its edge
Shape an HSS tool bit without overheating its edge
Grinding a high-speed steel tool bit creates the faces that meet at the cutting edge. Clearance behind the edge lets it cut without the adjacent tool face rubbing; rake influences chip flow. The required angles come from the intended tool and work material, not one universal sharpening angle.
Use light, controlled contact and check the shape frequently. Periodic dipping of a warm HSS bit in water can control temperature when the tool's instructions permit it. This is cooling during sharpening, not heat-treatment quenching; do not heat the bit severely and then plunge it into water.
Do not transfer that dipping method to a hot cemented-carbide tip. Sudden cooling can crack carbide. Use the specified dry method or a compatible continuous wet-grinding process. Intermittently quenching hot carbide is not equivalent to supplying coolant throughout the cut.
Supporting detail: Addendum 94 — Carbide needs consistent temperature control.
Aluminum and other soft, ductile metals can smear into and load a standard wheel intended for steel. Use equipment and abrasives specifically approved for the material. Changing from aluminum oxide to silicon carbide by itself does not establish that the whole setup is suitable.
Compare RPM ratings directly; calculate rim speed separately; Practice
Compare RPM ratings directly; calculate rim speed separately
The machine's spindle speed must not exceed the wheel's printed maximum operating RPM. A wheel's grade or grit marking is not that rating. Passing the RPM comparison is only one compatibility check; it does not approve the wheel's mounting, dimensions or application.
Rim speed is distance traveled at the wheel's outside diameter. With diameter D in inches and spindle speed N in revolutions per minute:
SFPM = π × D × N / 12, in surface feet per minute. The 12 converts inches to feet. Use diameter, not radius.
For an illustrative 7 in wheel at 1,750 RPM, π × 7 × 1,750 / 12 = 3,207 SFPM, rounded to the nearest whole foot per minute. This calculation is not a recommended operating speed. A smaller worn wheel has a lower rim speed at the same RPM; that does not authorize exceeding its marked maximum RPM.
Follow surface-grinder motions and workholding, plan small downfeed increments, and distinguish surface, cylindrical and centerless grinding.
Controlled motion lets many small cuts produce fine size and finish
Controlled motion lets many small cuts produce fine size and finish
Precision grinding combines an abrasive wheel with controlled workholding and machine feeds. Each active grain removes a small amount. A dressed wheel, stable support, small finishing increments and temperature control can produce accurate dimensions and fine surface texture, including on hardened materials.
A surface grinder separates table travel, crossfeed and downfeed
A surface grinder separates table travel, crossfeed and downfeed
On a horizontal-spindle reciprocating-table surface grinder, the table carries the work back and forth under the wheel. Crossfeed shifts the contact across the width of the work. Downfeed moves the wheel toward the work in small increments to remove thickness. Other surface grinders use different arrangements, including rotary tables.
Separate views keep the three directions distinct. Arrows show motion, not feed amounts.
After the finishing downfeed, sparking out continues passes without adding downfeed. The wheel can still remove a small amount as elastic deflection relaxes. Spark-out is not a substitute for measuring the part, and the number of passes is process-dependent.
A magnetic chuck holds suitable ferromagnetic work
A magnetic chuck holds suitable ferromagnetic work
A magnetic chuck uses magnetic attraction to hold suitable material such as many steels. It does not directly hold aluminum, brass or every stainless steel. Holding depends on material, contact area, thickness and the chuck's pole arrangement, not simply whether a magnet attracts one corner.
Clean mating surfaces and remove burrs that would lift the part. Use an approved fixture for parts that the chuck cannot hold adequately. A thin part can be pulled flat during grinding and spring back after release, so check the finished condition after removing the holding force.
Follow the chuck's release procedure after grinding. Residual magnetism may require demagnetizing so particles do not cling to the work. Demagnetizing is not the act that secures the work for grinding.
Supporting detail: Addendum 95 — Magnetic holding and release are equipment-specific.
Count downfeed increments separately from table strokes; Practice
Count downfeed increments separately from table strokes
For an ideal plan with constant downfeed, number of increments = total thickness to remove / downfeed per increment. One increment may require many table strokes and crossfeed steps to cover the entire face. State what a “pass” means before using the equation.
Example: removing 0.024 in from one face at an assumed 0.002 in per full-face layer takes 12 downfeed increments. These are arithmetic teaching values, not machine settings. If the division is not exact, use a smaller final increment rather than removing extra stock.
Supporting detail: Addendum 96 — Why a real grinding plan needs more than the quotient.
Coolant helps control heat and carry away grinding debris
Coolant helps control heat and carry away grinding debris
Grinding fluid can cool the work, lubricate contact and flush away removed particles. Heat changes dimensions during grinding and can alter the work surface. Delivery into the grinding zone, cleanliness and compatibility with the wheel and work material matter as much as having fluid in the tank.
Cylindrical and centerless grinding control round surfaces differently
Cylindrical and centerless grinding control round surfaces differently
Cylindrical grinding commonly rotates work held between centers or in a chuck while a grinding wheel removes material from an outside diameter. Traverse grinding moves along the length; plunge grinding feeds into a selected width. Internal grinding finishes bores and can use rotating work or other machine arrangements.
Centerless grinding supports the work on a blade between the grinding wheel and a regulating wheel. The regulating wheel controls work rotation and, in through-feed operation, helps advance the work axially. The work does not need center holes. Infeed arrangements serve shapes that cannot pass straight through.
Each process needs its own setup and wheel selection. A fine finish alone does not establish roundness, straightness or size.
Machining/CNC
CNC Control Fundamentals
How a control turns a program into motion, how feedback works, and what changes between milling and turning.
CNC repeats programmed motion; the operator still defines the process
CNC repeats programmed motion; the operator still defines the process
Computer numerical control (CNC) uses a program to command machine motion and functions. On a manual machine, the operator directs each move with handwheels or power-feed controls. On a CNC machine, the operator prepares the tools, workholding and coordinate references, then checks the programmed sequence.
CNC can repeat a complex path consistently. It does not remove variation from tool wear, deflection, temperature or setup errors. Programming and setup take time, so CNC is not automatically faster for every one-off job.
The machine control unit (MCU) reads program blocks, keeps track of active settings, calculates intermediate positions along a path, and sends commands to the axis drives and machine functions. The drives supply power to the motors. The control also presents positions, settings and alarms to the operator.
Feedback distinguishes open-loop and closed-loop motion
Feedback distinguishes open-loop and closed-loop motion
System
How position is controlled
What that means
Open loop
The drive sends motion commands without measuring actual position back into that positioning loop.
The control may not detect a missed step. A displayed commanded position is not proof of actual position.
Closed loop
A sensor reports position or motion; the drive/control compares feedback with the command and corrects error.
Excessive following error can produce an alarm. Feedback does not make the mechanics error-free.
A stepper motor advances through commanded angular steps and is often used in open-loop systems. A servo system uses feedback to regulate motion. Closed-loop stepper systems also exist: motor type alone does not tell you whether feedback is present.
Supporting detail: Addendum 97 — Where the sensor sits matters.
Milling rotates the cutter; turning rotates the work
Milling rotates the cutter; turning rotates the work
Basic machine
Cutting motion
Common programming concerns
Three-axis vertical mill
A rotating cutter moves relative to the work along X, Y and Z.
Tool lengths, cutter radius, XY contours and hole locations.
Two-axis lathe
The work rotates while a cutting tool moves along X and Z.
Diameter/radius convention, tool geometry and wear offsets, tool nose radius and changing work diameter.
These pages describe common Fanuc-style milling usage, with turning differences called out. G- and M-code meanings vary by control, machine type, options and configuration. Check the machine's own manual before applying a code or example. These examples teach interpretation; they are not production programs.
PathPilot is Tormach's machine control. Tormach publishes PathPilot programming documentation and post-processor guidelines. Select documentation for the actual machine and software version. A general description does not establish its installed tools, options or setup.
Identify mill and lathe axes, distinguish machine and work zero, and separate work offsets from tool compensation.
Describe tool motion relative to the work with a right-handed frame
Describe tool motion relative to the work with a right-handed frame
Use your right hand: thumb along +X, index finger along +Y and middle finger along +Z, with all three perpendicular. Equivalently, curling from +X toward +Y makes the thumb point along +Z. Z runs parallel to the principal spindle axis.
On a conventional vertical mill viewed from the front, +X is to the right, +Y is toward the back and +Z raises the tool away from the work. These are tool-relative-to-work directions. If the table moves instead of the tool, its physical motion is opposite the tool-relative motion.
For a basic lathe, Z is along the spindle and +Z is away from the headstock. X is radial; positive X values commonly describe the tool side of the centreline. A slant bed or rear turret changes what this looks like from the operator's position. Do not infer axis signs from screen orientation. Two-axis turning normally uses the XZ plane.
A lathe X value may describe diameter or radius; Practice
A lathe X value may describe diameter or radius
In diameter programming, X24 describes a surface 24 mm in diameter, 12 mm from the spindle centreline. In radius programming, the same surface is X12. Z distances are not halved.
For a change from diameter 30 mm to diameter 26 mm, the X coordinate changes by −4 mm in diameter mode, but the tool moves only 2 mm radially inward. Confirm the control's convention for absolute X, incremental X and arc-centre words separately; do not halve every number in a turning block.
Machine zero belongs to the machine; work zero belongs to the setup
Machine zero belongs to the machine; work zero belongs to the setup
Machine zero is the machine coordinate origin. A reference-return or homing procedure establishes the machine's position reference; the home position need not be the coordinate origin. Work zero is a chosen datum on the part or fixture, such as a corner, hole centre or faced end.
Common milling codes G54 through G59 select stored work coordinate systems. Selecting G54 does not measure the part, write the offset table or redefine machine zero. The setup procedure establishes the stored values.
With only a translation, machine position = work-zero location in machine coordinates + work position. If the work origin is (100,60) and the program point is (20,15), the machine point is (120,75). This example omits rotation, local shifts and tool offsets.
In common Fanuc-style milling, G53 selects machine coordinates for the block containing it. It does not erase G54 or force G54 to become the next active work system. The previously selected work system remains selected afterward. Tool-offset interaction and permitted motion modes are control-specific; check your control's manual. Machine-coordinate zero is not necessarily a clearance position.
Tool length and cutter radius compensate for different dimensions
Tool length and cutter radius compensate for different dimensions
In common Fanuc-style milling, G43 H... applies positive tool-length compensation: it adds the signed value stored in the selected H register to commanded Z positions so Z can refer to the intended tool-tip position. The offset's sign, established by the touch-off convention, determines the direction of that shift. G43 does not mean an unconditional upward move. Check your control's manual for its offset and touch-off conventions.
For example, H03 selects register 3; it does not mean a length of 3 mm. T selects a tool; H selects offset data. Their numbers may match by shop convention, but the programmer must verify that relationship. G49 commonly cancels length compensation.
Supporting detail: Addendum 98 — Why positive compensation can use a negative offset.
G41 offsets a cutter to the left of the programmed path looking in the direction of travel; G42 offsets it to the right. G40 cancels it. These are not clockwise/counterclockwise commands. Moving along +X in XY, left is +Y. On many mills a D word selects cutter compensation data; check whether that register stores radius, diameter or a wear adjustment.
Compensation needs suitable entry and exit moves. Do not add a full radius correction to a CAM path that already includes that correction. Turning tool nose compensation also depends on nose radius and tool orientation; it is not simply a milling H offset.
Convert between positions and signed moves, and trace the endpoint of a program as its distance mode changes.
G90 gives a destination; G91 gives a signed change
G90 gives a destination; G91 gives a signed change
In common milling usage, G90 makes axis values absolute coordinates in the active coordinate system. G91 makes them signed changes from the current position. Neither command moves an axis by itself. Units and work offsets still matter.
Increment = next position − current position. Reverse the calculation with next position = current position + increment. Keep a separate running total for every axis.
For an ordinary G00/G01 move, an omitted axis stays in place. In G90, X0 requests X zero; in G91, X0 requests zero X movement. Do not reuse an omitted incremental value from the preceding block.
These G90/G91 meanings must not be assumed on every lathe. Some turning controls use X/Z for absolute and U/W for incremental coordinates, and use G90 for a turning cycle. Check the control before translating a mill program to a lathe.
One path can be written as points or as moves; Practice
One path can be written as points or as moves
Start at A = (6,4) mm. Visit B = (26,4), C = (26,19), then D = (11,19). Z is unchanged; these are coordinate exercises with no cutting setup implied.
Move
Absolute destination, G90
Increment, G91
A → B
X26 Y4
X20 Y0
B → C
X26 Y19
X0 Y15
C → D
X11 Y19
X−15 Y0
Supporting detail: Addendum 99 — Check the conversion in both directions.
Trace the active mode before evaluating each block; Practice
Trace the active mode before evaluating each block
Assume millimetres, one unchanged work coordinate system, no compensation and an initial position (6,4,12). This is a paper trace, not a machine-ready program. F120 is an arithmetic example.
M30 ends this trace; it is not a command to move to zero. For arcs, the endpoint mode and centre mode are separate questions: see G-code and Toolpaths. For a broader view of modes use CNC Program Structure.
Read program identifiers, startup modes, machine functions, feed modes and drilling cycles in common milling G-code.
A block can combine motion, settings and machine functions
A block can combine motion, settings and machine functions
A word is a letter and value, such as X12 or F100. A block is a line of words. The control applies its execution rules; words are not necessarily executed in left-to-right typing order. N numbers label blocks and are often optional. Many Fanuc-style programs use an O number such as O2468 as the program identifier. File naming, percent delimiters and identifiers vary by control.
A modal command stays active until replaced or cancelled within its group. G01 can remain active while G90 changes to G91, because motion and distance mode are different groups. A startup line sets specific modes; it does not clear every stored setting or offset.
A safe-start line declares assumptions; it does not prove the setup
A safe-start line declares assumptions; it does not prove the setup
A common milling startup example is G17 G21 G90 G94 G40 G49 G80: XY plane, millimetres, absolute endpoints, feed per minute, cutter compensation cancelled, tool-length compensation cancelled and canned cycles cancelled. This is a reading example, not a universal restart line. Cancelling an offset changes the coordinate relationship; startup and restart sequences must match the actual control and tool position.
Identify the program and establish required units, plane and modes.
Select the intended work coordinate system and tool; apply the correct offset data.
Establish spindle, feed and coolant settings for the operation.
Perform positioning and machining blocks, including any cycle cancellation.
Use the machine-approved retract and shutdown sequence, then end the program.
Separate spindle, tool and coolant commands from motion; Practice
Separate spindle, tool and coolant commands from motion
These are common Fanuc-style meanings, not a promise for every machine.
Code
Common meaning
Read with it
M03 / M04 / M05
Spindle clockwise / counterclockwise / stop
S sets the speed value. Direction is conventionally viewed from the spindle toward the work; confirm machine orientation.
M06
Tool change on many mills
T selects a tool. The changer may be manual or automatic; offset activation is a separate issue.
M08 / M09
Flood coolant on / coolant off
Available coolant systems and auxiliary M codes depend on the machine.
M30
End program and commonly reset/rewind
It does not command a return to machine or work zero, and does not mean automatically run another part.
For example, S1800 M03 requests 1800 rpm in a fixed-rpm mode; it is not a cutting recommendation. A lathe using constant surface speed interprets S differently. See Feeds and Speeds.
Units, feed mode and plane answer different questions
Units, feed mode and plane answer different questions
Setting
Common milling interpretation
G20 / G21
Inch / millimetre length units.
G94 / G95
Feed in units per minute / units per spindle revolution, where supported.
G17 / G18 / G19
XY / XZ / YZ plane for plane-dependent operations such as arcs.
G00 / G01
Rapid positioning / straight-line motion at programmed feed.
G02 / G03
Clockwise / counterclockwise arc in the selected plane.
In G21 G94, F90 means 90 mm/min. In G21 G95, F0.15 means 0.15 mm/rev; at 600 rpm that corresponds to 90 mm/min. G20/G21 alone does not choose the time basis. Some lathes use G98/G99 for feed mode, while mills commonly use those codes for drilling return levels.
Rapid path shape depends on the control and can differ from G01. G01 may feed through air as well as cut. Learn the motion and arc geometry in G-code and Toolpaths, including the existing simulator; learn distance modes in Absolute and Incremental Positioning.
G81 and G83 repeat a drilling sequence at hole locations
G81 and G83 repeat a drilling sequence at hole locations
A canned cycle packages repeated motions. In common milling use, G81 feeds a drill to depth then retracts; G83 divides the depth into pecks with retracts for chip clearance. X/Y locate the hole, Z specifies the drill-tip depth, R specifies a return/approach plane, F sets feed, and Q commonly specifies peck depth for G83. Q is not the final hole depth.
With G90 on a conventional mill, Z and R are coordinates in the active work system. Incremental canned-cycle rules need separate study in the manual. G98 commonly returns to the initial level and G99 to R; choose a return level for the actual setup. G80 cancels the cycle. Modal hole cycles can drill again at subsequent position blocks.
Follow the CAD-to-CAM-to-post chain and use graphics, dry run, single block and overrides to check a program against its setup.
CAD describes the shape; CAM plans how to cut it
CAD describes the shape; CAM plans how to cut it
CAD supplies design geometry and drawings. CAM uses geometry plus the programmer's stock, tool, workholding, operation and cutting-data choices to calculate toolpaths. Depending on the operation, CAM may work from 2D profiles or a 3D model; a solid model is not always required.
A post-processor translates CAM output into the syntax and conventions for a particular machine/control combination. Selecting the wrong post can change tool calls, arc format, offsets or startup behavior. Review the posted code as well as the CAM preview.
Supporting detail: Addendum 100 — Other ways to create a program.
Each prove-out aid answers a different question
Each prove-out aid answers a different question
Prove-out compares the intended operation, posted code and actual setup. Follow Equipment Operating Procedure and Machine Hazards and Housekeeping for safety requirements; this page explains program-checking concepts, not a machine operating procedure.
Aid
Useful for
Limit
Graphics or simulation
Inspecting path shape, endpoints, sequence and modeled interference.
A missing clamp, wrong tool model or incorrect offset can make the preview disagree with the machine.
Dry run
Observing a trial cycle under the control's defined test conditions.
It may substitute feed rates; the name does not guarantee no motion or no contact. Use the manual's procedure.
Single block
Advancing through blocks while comparing position and active modes with expectations.
One canned-cycle block may contain several motions. Not every internal action stops separately.
Feed override
Adjusting eligible feed motions during observation.
It need not affect rapids, tool changes or synchronized operations. Rapid override may be a separate control.
Review the program against the setup before a supervised trial; Checklist
Review the program against the setup before a supervised trial
Match the posted file and revision to the intended part, machine and post-processor.
Identify the active units, plane, distance mode and feed mode at each operation.
Compare work offsets, tool numbers and offset registers with the documented setup.
Trace the first approach, tool changes, cycle returns and final retract against the modeled stock and fixtures.
Check that the actual setup matches what the preview includes; resolve any discrepancy with the instructor before proceeding.
Identify the manual's dry-run, single-block and override behavior before using those functions in a supervised trial.
After an edit, repeat the affected checks; compare the first part's measured features with the drawing.
Practise interpretation with G-code and Toolpaths and explore the existing 3D/WebGL milling lab in Milling Operations. A successful simulation only establishes behavior within that model. It does not approve a program for a real machine.
Machining/CNC
CNC Turning Setup and Offsets
Read lathe X and Z coordinates, identify tooling and workholding, and separate part location from tool geometry and wear.
Read X as a size and Z as a position along the spindle; Practice
Read X as a size and Z as a position along the spindle
Start with CNC Control Fundamentals for how CNC works and Lathe Operations for the cuts a lathe makes. Here the work rotates and the ordinary turning tool moves in the XZ plane. Z runs along the spindle; positive Z is away from the headstock in the basic arrangement shown. X moves radially, even when its displayed value is a diameter. A slant bed or rear turret changes the view, so check the machine's axis diagram.
Confirm diameter or radius programming before reading an X value. For a 28 mm diameter, diameter mode uses X28 and radius mode uses X14. Moving from X28 to X24 in diameter mode removes 2 mm radially. Z is not halved. Confirm incremental and arc-centre conventions separately.
Use Absolute and Incremental Positioning for coordinate arithmetic. On many Fanuc-style lathes X/Z are absolute endpoints and U/W are increments; G90 can be a turning cycle. Do not paste a mill's G90/G91 startup into a lathe program.
The tool arrangement determines how the next cutter reaches the work
The tool arrangement determines how the next cutter reaches the work
A turret indexes a selected tool station into working position. Gang tooling mounts several tools on a slide; moving the slide brings the selected tool into position without turret indexing. Each tool still needs the correct offsets and a checked approach path.
Live tooling powers a rotating cutter on the lathe for operations such as drilling or milling. A C axis controls the spindle's angular position, allowing the work to be indexed or rotated under position control. These are options, not features of every CNC lathe. Some machines also add a Y axis. Live-tool and spindle-mode commands belong to the machine's manual.
Supporting detail: Addendum 101 — Why an extra axis changes the work a turning center can do.
Match the workholding and automation to the part
Match the workholding and automation to the part
A chuck grips with jaws. A self-centering chuck moves its jaws together; independent jaws permit individual adjustment. Jaw style and setup determine which shapes can be held.
A collet closes around a matching stock size and shape. Its useful size range depends on the collet design.
A tailstock center supports the free end of suitable work. It does not replace the chuck or collet that drives the work.
A bar feeder advances bar stock for successive parts. A parts catcher receives a separated part. Their timing and interlocks are machine-specific; neither is assumed in an ordinary two-axis program.
Locate the part face once, then locate each tool relative to it
Locate the part face once, then locate each tool relative to it
Use CNC Axes and Offsets for machine zero, work zero and the general offset relationship. On a lathe, a common part datum is Z0 at the finished face and X0 on the spindle centreline. Establish the Z work shift using the control's documented reference-tool, probe or touch-off method. Allow for any material still to be faced away.
For X touch-off, the reference is a measured diameter. In a trial-cut method, a tool cuts a short cylindrical surface and that diameter is measured; the control's measurement function uses the known size to establish that tool's X geometry. Use the manual's procedure for the actual machine, including which table receives the value. Do not enter the measured diameter as though it were automatically the raw geometry-offset value.
Stored information
What it describes
Work shift or work offset
The part datum, especially its Z location.
X and Z tool geometry
The position of each tool's reference tip relative to the established reference system.
X and Z wear
Small dimensional corrections after measurement, separate from the basic tool setup.
Check whether an X wear entry is interpreted radially or diametrally, its sign, and whether the control adds to or replaces the existing value. A work-offset change affects the shared datum; a tool correction addresses one tool. Selecting G54 does not measure or store a new zero. G53 does not erase the offset table.
Supporting detail: Addendum 102 — Control example: separate geometry and wear fields.
Machining/CNC
CNC Turning Spindle and Feed Modes
Calculate constant-surface-speed RPM, apply an RPM ceiling, and distinguish feed per revolution from feed per minute.
G96 changes RPM with diameter; G97 holds a requested RPM; Practice
G96 changes RPM with diameter; G97 holds a requested RPM
In the Fanuc-style lathe convention used here, G96 selects constant surface speed (CSS). S gives cutting speed: surface feet per minute with inch units, or metres per minute with metric units. G97 cancels CSS and makes S an RPM value. Confirm the actual control's units and code meanings.
Use the work diameter at the cutting point, not the insert size. From Feeds and Speeds:
Inch: RPM = SFM × 12 / (π × diameter in inches).
Metric: RPM = cutting speed in m/min × 1000 / (π × diameter in mm).
At 180 SFM on a 1.5 in diameter, RPM = 180 × 12 / (π × 1.5) = 458 RPM, rounded to a whole RPM. At 90 m/min on a 30 mm diameter, the result is 955 RPM. These are arithmetic examples, not recommended cutting data.
Set the maximum RPM before enabling constant surface speed; Interactive; Practice
Set the maximum RPM before enabling constant surface speed
Establish an appropriate spindle-speed ceiling for the actual setup before G96. On the Fanuc-style lathe convention here, G50 S1800 limits requested speed to 1800 RPM. The number is only an illustration; determine the permissible limit from the machine, workholding, tooling and setup documentation. Other controls use different commands or parameters.
As diameter approaches zero during facing, CSS requests increasing RPM. Once the clamp is reached, RPM stays limited and surface speed falls. At the centre, a nonzero surface speed cannot be maintained at any finite RPM. Use CNC Programming and Prove-Out and the Safety collection for the required operating checks.
Supporting detail: Addendum 103 — Calculate where a speed ceiling starts to matter.
Worked example: 240 SFM, 2 in stock outside diameter (OD), and a 1000 RPM clamp. These are teaching values, not cutting recommendations. D* = 240 × 12 / (π × 1000) = 0.917 in.
The dashed horizontal line is the RPM ceiling. The dot shows D = 0.5 in.Below D*, RPM stays at 1000 and surface speed falls linearly toward zero.
One setpoint, three cutting diameters
D (in)
RPM
SFM
Clamped?
2
458.37
240
No
0.5
1000
130.9
Yes
0
1000
0
Yes
At 2 in: 240 × 12 / (π × 2) = 458.37 RPM. At 0.5 in, CSS would request 1833.46 RPM; the clamp gives 1000 RPM and π × 0.5 × 1000 / 12 = 130.90 SFM. At D = 0, the model commands the clamp RPM and surface speed is zero.
This ideal steady-state model uses π without rounding in the calculations. Displays round RPM and surface speed to two decimal places, diameters to three; trailing zeros are omitted. It assumes the commanded RPM is reached immediately and omits acceleration and other machine limits. Unit changes load a fresh teaching preset. For operating checks, see CNC Programming and Prove-Out and the Safety collection.
Read the feed mode before interpreting F; Practice
Read the feed mode before interpreting F
In this common Fanuc-style lathe convention, G99 means feed per spindle revolution and G98 means feed per minute. With inch units, G99 F0.006 means 0.006 in/rev; G98 F0.006 means 0.006 in/min. With metric units, the corresponding lengths are millimetres. Codes vary: the milling return-level meanings of G98/G99 do not apply here.
Feed per minute = feed per revolution × RPM. At 800 RPM, 0.006 in/rev gives 4.8 in/min. If CSS increases RPM while G99 holds F constant, the feed per minute increases too. If G98 holds feed per minute constant, feed per revolution changes with RPM.
Account for a rounded tool nose on tapers and arcs, then distinguish roughing, finishing, grooving and threading cycles.
A rounded tool nose does not cut at the imaginary sharp tip
A rounded tool nose does not cut at the imaginary sharp tip
A turning insert has a rounded corner, called its nose radius. Touching off along X and Z establishes an imaginary tip: the intersection of the two tangent reference lines. There is no cutting material at that intersection. The actual contact point moves around the nose as the direction of the surface changes.
Axis-parallel cuts can reach the measured diameter or face with the corresponding touch-offs. On a taper, chamfer or arc, simply driving the imaginary tip along the desired surface leaves a profile error. Account for the nose radius with a calculated toolpath or the control's tool nose compensation. See Turning Tapers for taper geometry.
Supporting detail: Addendum 105 — Why a good diameter measurement does not prove a taper is correct.
Compensation needs nose radius, tip orientation and a clear entry and exit
Compensation needs nose radius, tip orientation and a clear entry and exit
In common Fanuc-style usage, G41 selects compensation to the left of the programmed direction of travel and G42 to the right; G40 cancels it. Judge the side in the control's XZ-plane convention. Do not choose a code from the screen's apparent left or right, or assume one code means every outside cut.
Enter the actual nose radius and the control's tip orientation, which tells it where the imaginary tip sits relative to the nose centre. Select the correct tool-offset register, use the manual's lead-in and lead-out requirements, and check the resulting path with CNC Programming and Prove-Out. Tip numbers differ by control; there is no universal number table here.
Do not apply full nose-radius compensation again to a path already corrected for that radius. Check whether a CAM post expects full geometry compensation, wear-only compensation or no control compensation.
Supporting detail: Addendum 106 — Where to check the control behavior.
A roughing cycle removes layers while leaving stock for finishing
A roughing cycle removes layers while leaving stock for finishing
A G71-type roughing cycle uses a defined final profile to generate repeated stock-removal passes, commonly along Z for outside or inside turning. The programmer supplies the applicable cut depth, retract and finish allowance. A G70-type finishing cycle follows the defined profile to remove the remaining allowance.
Read the control's cycle definition before entering parameters. One-block and two-block formats differ. A depth word may be radial while an X allowance is diametral; the same letter can mean different things in different blocks. P and Q commonly identify the first and last numbered profile blocks. Check the referenced blocks, permitted profile direction changes, return path and compensation rules.
Supporting detail: Addendum 107 — Why no universal cycle template is shown.
Grooving repeats plunges; threading synchronizes feed with rotation
Grooving repeats plunges; threading synchronizes feed with rotation
A grooving or peck cycle divides a plunge into smaller advances and retracts. Some formats repeat the plunge at different Z positions to make a wider groove. G75 is one common outside/inside grooving family; face-grooving and peck-drilling families may use other codes. Check peck size, final depth, groove width and tool-width convention separately.
A G76-type threading cycle makes repeated synchronized passes with increasing depth. The thread lead determines the axial advance per spindle revolution. The cycle also needs control-specific information about final size, infeed, pass depth and runout. Use Single-Point Threading for pitch, lead and thread form; a threading cycle does not choose those dimensions for you.
Formats differ by control, so no complete parameter syntax is universal. See the Haas G75 reference and the G70/G76 entries in its lathe workbook as examples. CNC Turning Program Walkthrough spells out ordinary moves so the sequence can be read without decoding a cycle.
Machining/CNC
CNC Turning Program Walkthrough
Trace an original facing, outside roughing, chamfering, finishing and cutoff sequence with every block explained.
The invented part is 20 mm in diameter and 25 mm long, with a 1 mm × 45° front chamfer. The chamfer runs from X18 Z0 to X20 Z−1: the 2 mm diameter change is a 1 mm radial change. Stock is 24 mm diameter with its raw face at Z+0.5. G54 has Z0 at the finished face and X0 on the spindle centreline. Stock and clearance extend beyond Z−28; workholding is farther toward negative Z.
Assume diameter programming and absolute X/Z endpoints; this dialect does not need G90 for absolute moves. T0101 selects station 1 and offset 1: a right-hand OD/facing tool with a 0.4 mm nose radius, suitable for the illustrated cuts. Its imaginary tip and radius are set, and G42 is the correct side for this setup's finish contour. T0202 selects a 2 mm wide cutoff tool and offset 2. Its Z reference is the edge toward positive Z, so Z−25 puts the blade in the interval Z−27 to Z−25. The cutoff tool is capable of the illustrated small travel past centre.
The example assumes X60 Z10 is an already-verified tool-change clearance location for both tools and the turret sweep, and starts there with spindle and coolant off. These coordinates are not a general clearance rule. All speeds, feeds and the 1800 RPM ceiling are teaching values, not material recommendations. Verify the real setup, tool calls, compensation entry/exit, centre crossing and file delimiters using CNC Programming and Prove-Out and the Safety collection.
Face the end, then leave a small allowance on the outside diameter
Face the end, then leave a small allowance on the outside diameter
Every parenthesized comment explains its block. G01 remains active until G00 replaces it; X/Z values here are destinations. Facing travels slightly past the centre to account for the rounded nose. Roughing uses two explicit lengthwise passes rather than a control-specific canned-cycle format.
O2460 (IDENTIFY THIS EXAMPLE PROGRAM)
G18 G21 G40 G80 G99 (XZ PLANE, MM, COMP OFF, CYCLES OFF, MM/REV)
G54 (SELECT THE PRESET PART WORK OFFSET)
T0101 (SELECT OD TOOL 1 WITH OFFSET 1 AT VERIFIED CLEARANCE)
G50 S1800 (SET THE EXAMPLE MAXIMUM SPINDLE RPM)
G97 S600 M03 (FIXED 600 RPM, START SPINDLE IN SETUP DIRECTION)
G00 X26.0 (APPROACH OUTSIDE THE 24 MM STOCK AT Z10)
G00 Z0.0 (MOVE TO FINISHED FACE PLANE WHILE OUTSIDE STOCK)
M08 (TURN ON THE ASSUMED FLOOD COOLANT SYSTEM)
G96 S90 (SELECT 90 M/MIN CSS WITH THE RPM CEILING ACTIVE)
G01 X-0.8 F0.10 (FACE THROUGH CENTRE: 0.4 MM RADIAL OVERTRAVEL)
G00 Z2.0 (RETRACT AXIALLY BEYOND THE FACED END)
G00 X22.0 (POSITION FOR THE FIRST OD ROUGHING PASS IN AIR)
G01 Z-28.0 F0.20 (ROUGH LENGTHWISE TO 22 MM DIAMETER)
G00 X26.0 (RETRACT RADIALLY OUTSIDE THE ORIGINAL STOCK)
G00 Z2.0 (RETURN TO THE FRONT AT CLEAR X)
G00 X20.6 (POSITION FOR THE SECOND OD ROUGHING PASS)
G01 Z-28.0 F0.20 (LEAVE 0.6 MM ON DIAMETER FOR FINISHING)
G00 X26.0 (RETRACT OUTSIDE STOCK AGAIN)
G00 Z2.0 (RETURN AHEAD OF THE FACE FOR FINISHING)
Compensate the chamfer and finish, then cut off at fixed RPM
Compensate the chamfer and finish, then cut off at fixed RPM
The finish contour uses the same tool at a smaller feed. Nose compensation is active for the chamfer and OD; it is entered ahead of the face and cancelled after departing the finished contour. See CNC Turning Profiles and Cycles for why a sharp-point assumption would give the wrong chamfer.
G00 X16.0 (POSITION AHEAD OF THE FACE FOR THE COMPENSATED ENTRY)
G01 G42 X18.0 Z0.0 F0.10 (ENGAGE RIGHT COMP ON THE FACE APPROACH)
G01 X20.0 Z-1.0 (CUT THE 1 MM RADIAL BY 1 MM AXIAL CHAMFER)
G01 Z-26.0 (FINISH THE 20 MM OD PAST THE CUTOFF FACE)
G01 X22.0 Z-27.0 (FEED OUT FROM THE OD BEHIND THE FINISHED PART)
G01 G40 X26.0 (CANCEL COMPENSATION ON THE OUTWARD DEPARTURE)
G97 S600 (CANCEL CSS AND REQUEST FIXED 600 RPM)
G00 X60.0 (RETRACT RADIALLY TO THE ASSUMED CLEARANCE X)
G00 Z10.0 (RETURN AXIALLY TO THE VERIFIED TOOL-CHANGE LOCATION)
M09 (COOLANT OFF)
M05 (STOP THE SPINDLE BEFORE THE TOOL CHANGE)
T0202 (SELECT CUTOFF TOOL 2 AND ITS EDGE-REFERENCED OFFSET 2)
G97 S500 M03 (START AT FIXED 500 RPM FOR THIS CUTOFF EXAMPLE)
G00 X26.0 (POSITION OUTSIDE THE STOCK WITH THE NEW TOOL)
G00 Z-25.0 (LOCATE THE POSITIVE-Z BLADE EDGE AT THE REAR FACE)
M08 (COOLANT ON)
G01 X-0.2 F0.05 (PART OFF WITH THE ASSUMED 0.1 MM RADIAL OVERTRAVEL)
G00 X60.0 (WITHDRAW RADIALLY FROM THE CUTOFF SLOT)
G00 Z10.0 (RETURN TO THE ASSUMED CLEARANCE Z)
M09 (COOLANT OFF)
M05 (SPINDLE STOP)
M30 (END AND RESET THE PROGRAM; NO RETURN MOVE IS IMPLIED)
File-start/end percent delimiters, tool-call behavior and supported startup codes must be supplied or adapted as required by the target control. No bar-feeder or parts-catcher sequence is implied.
Trace the modal state and finished dimensions on paper; Checklist; Practice
Trace the modal state and finished dimensions on paper
Mark where S means RPM and where it means metres per minute; locate the preceding RPM clamp.
Confirm that every F value is interpreted in mm/rev under G99.
Calculate the radial depth of each OD pass and the two chamfer endpoints.
Locate the nose-compensation entry and departure, and identify the selected offset register.
Check that the cutoff blade's positive-Z edge leaves the intended 25 mm length.
The syntax conventions can be compared with the Haas lathe programming reference. This original example is not copied from a manufacturer program. A paper trace establishes intended geometry and modal state; actual prove-out still follows CNC Programming and Prove-Out.
Check the first part, watch the process during production, and use suitable measurements to decide whether parts meet requirements.
Inspect before a setup error becomes a batch of parts
Inspect before a setup error becomes a batch of parts
Quality control uses evidence to keep production meeting requirements. Inspection checks what was made; it does not repair a process or make a bad part good.
First-article inspection
Check the first part from a new or changed setup against the required features before releasing continued production. A good first part supports the setup decision; it does not prove that later parts will stay good.
In-process checks
Check selected features while production continues. Record results in production order so tool wear or a shifting setup is visible before many parts are affected.
Final inspection
Check the completed part or lot against the acceptance plan, including requirements that earlier operations could not verify. Keep its identity connected to its results.
If a check fails, identify and hold the affected material, report the result, and investigate before continuing under the applicable procedure. Consider parts made since the last known good check; checking only the next part leaves that interval unresolved.
A gauge gives a decision; a measurement gives a value
A gauge gives a decision; a measurement gives a value
Attribute gauging reports a category such as pass/fail. Variable measurement reports a number, such as 18.012 mm. A number can show how close the process is getting to a limit; a passing gauge alone cannot.
For a plain hole, the GO plug checks the small-hole limit and should enter over the specified inspection length. The NO-GO plug checks the large-hole limit and should not enter as defined by the gauging procedure. Do not force either end. For a plain shaft, a GO ring should pass over the specified length and a NO-GO ring should not pass: the GO ring checks the large-shaft limit.
Conceptual plain plug gauge. The diameter difference is exaggerated; these are not gauge manufacturing dimensions.
A GO plug that will not enter is a failed check, not proof of one particular cause: an undersize hole, burr, dirt, poor alignment or form error can obstruct it. Clean and investigate using the approved method. A NO-GO plug that enters beyond the permitted engagement is also a failed check.
Use the correctly identified gauge with current calibration status. Do not derive gauge sizes directly from workpiece limits. This plain-gauge explanation does not supply the engagement rules for threaded gauges.
Supporting detail: Addendum 109 — Why the gauge itself needs limits and wear checks.
Build a clean, repeatable surface-plate setup; Checklist
Build a clean, repeatable surface-plate setup
A surface plate provides a reference plane. A height gauge measures from that plane; an indicator can compare the part with a reference or show variation as a surface is swept. Neither tool corrects a rocking part or dirt under a support.
Read the required feature and reference surfaces on the drawing; select a suitable instrument and range.
Clean the plate, supports and contact surfaces. Address burrs under the approved preparation method without altering the feature being inspected.
Support the part stably on the intended reference surfaces and allow temperature differences to settle as the measurement procedure requires.
Check the instrument against the appropriate reference, set zero or reference height, and use consistent contact force and alignment.
Repeat the reading, record the value and units, and investigate disagreement before accepting the result.
A CMM relates measured points to the part's references
A CMM relates measured points to the part's references
A coordinate measuring machine (CMM) records positions in three dimensions using a probe or sensor. Software uses sampled points to evaluate features such as a plane, bore or hole location relative to the chosen part alignment.
The operator still needs a suitable setup, probe qualification, measurement strategy and drawing interpretation. A few points cannot reveal every possible surface defect. A CMM is a measurement system, not an automatic guarantee that every requirement has been checked.
Increase checks when the process is new or changing
Increase checks when the process is new or changing
Sampling means inspecting some parts to learn about a larger group or process. It reduces inspection effort but leaves a chance that unmeasured defects will escape. Select samples by a defined plan, not by choosing the nicest-looking parts.
A new setup, tool replacement, unexpected variation or a trend toward a limit calls for closer attention and potentially more frequent checks under the inspection plan. A demonstrated stable process may justify fewer checks; feature consequences and customer requirements still matter. Some features require every part to be inspected.
Keep the readings in order and compare their trend and spread, not just their pass/fail status. Process Capability and Variation explains that distinction. Do not repeatedly adjust a stable process just because its readings differ slightly.
Trust a result only as far as the measurement supports it; Practice
Trust a result only as far as the measurement supports it
Accuracy concerns agreement with a reference or true value; precision concerns agreement among repeated measurements under stated conditions. Readings can cluster tightly and still be consistently wrong.
Calibration compares instrument indications with reference values under specified conditions and establishes their relationship and uncertainty. It is not merely setting zero, and adjustment is a separate action. Traceability belongs to a measurement result supported by a documented chain of calibrations, each contributing uncertainty.
For classroom limit comparisons, treat the stated numbers as exact and include both endpoints. In real inspection, uncertainty matters near a limit: apply the specified acceptance decision rule rather than assuming the last displayed digit is exact.
Compare the allowed tolerance band with the process's centre and spread, without mistaking five good parts for proof of capability.
The drawing sets the limits; the process produces the spread
The drawing sets the limits; the process produces the spread
A tolerance band is the permitted interval from the lower size limit to the upper size limit. Process spread describes how much the manufactured values vary. Changing a machine setting may shift the centre; it does not necessarily reduce the spread.
A capable process consistently fits its variation within the required limits with useful margin. A narrow spread near a limit can still produce rejects if the process drifts. A wide spread cannot be fixed simply by moving its average to the middle.
Illustrative spreads, not probability curves or measured capability results. Both rows use the same size axis.
Consider five invented readings: 24.98, 24.99, 25.00, 25.01 and 25.02 mm, with limits of 24.95 to 25.05 mm.
Mean: add all five readings and divide by five: 125.00 / 5 = 25.00 mm.
Range: subtract the smallest from the largest: 25.02 − 24.98 = 0.04 mm.
Midpoint of the limits: (24.95 + 25.05) / 2 = 25.00 mm. This sample mean is centred on that midpoint, and all five values are inside the limits.
Now add 0.04 mm to every reading. The mean becomes 25.04 mm and the range remains 0.04 mm, but the largest reading becomes 25.06 mm and fails. Centre and spread answer different questions.
Supporting detail: Addendum 110 — Why a small sample range is not the full process spread.
Passing parts are evidence, not proof of a capable process; Interactive; Practice
Passing parts are evidence, not proof of a capable process
Capability concerns a process over time, not a single successful part. First establish that the measurement method is suitable and the process is stable enough to describe meaningfully. Collect representative readings across the conditions the conclusion is meant to cover.
A stable process can still be too wide for the tolerance. An unstable process can happen to produce five passing parts. Neither case is resolved by averaging away the problem. Investigate a trend or sudden change using the inspection plan.
Supporting detail: Addendum 111 — Where capability indices fit.
Capability explorer: assume a stable process with roughly normal output. μ is the process mean; σ (standard deviation) measures spread. These are chosen teaching inputs, not estimates from the five readings above. The five-readings example above explains why passing parts are evidence, not proof of a capable process.
Cp = (USL − LSL) / (6σ) compares tolerance width with spread, ignoring centring. Cpk = min(USL − μ, μ − LSL) / (3σ) uses the nearer limit. Cpk is negative when the mean lies outside the limits. LSL and USL are the lower and upper size limits.
Predict first: which case has the higher Cp? Does it also have fewer parts outside the limits?
Centring makes Cp and Cpk equal. The wide spread still puts parts beyond both limits.
Expected ppm means parts outside the limits per million under this normal model, not a measured reject count. Values are rounded; less than 1 ppm is not a guarantee of zero rejects. Requirements depend on the application; no universal pass threshold is used here.
Locate a part repeatably, clamp it against supports, and distinguish tool-guiding jigs from workholding fixtures.
A jig guides the tool; a fixture locates and holds the work
A jig guides the tool; a fixture locates and holds the work
A drill jig locates and holds the part and guides the drill with a bushing. A fixture locates and holds the part while the machine controls the tool path. A plate with clamps is not automatically a jig: ask whether it guides the cutting tool.
Both reduce the need to recreate the part position for every cycle. Neither removes the need for trained setup, secure restraint and inspection. Follow the approved mounting method; do not infer that a jig may be held by hand.
Three, two and one contacts establish six independent constraints
Three, two and one contacts establish six independent constraints
A free rigid body can translate along X, Y and Z and rotate about X, Y and Z: six degrees of freedom. For an ideal rectangular block seated against perpendicular locating planes, the 3-2-1 arrangement establishes its position without redundant locating contacts.
Contacts
Plane in this example
Additional motions constrained while seated
Three non-collinear supports
Primary, XY underside
Z translation, rotation about X and Y: three
Two separated side locators
Secondary, XZ side
Y translation and rotation about Z: two
One end locator
Tertiary, YZ end
X translation: one
Schematic contact locations, with the hidden contacts shown through the block. A, B and C name the locating planes in this example, not a replacement for the part drawing.
These contacts locate only while the work stays seated. Clamps maintain contact and resist forces that would lift or separate the part. Do not count positive and negative directions as twelve independent degrees of freedom.
Direct clamp force into supports and leave room for chips; Checklist
Direct clamp force into supports and leave room for chips
Place clamp forces toward locators and supported regions, not toward open space where the part can bend. Use enough restraint for the operation without distorting the part. Plan the load path from the cutting force through the work and supports into the machine.
A fourth fixed support under an imperfect surface may make the part rock instead of improving location. Extra supports can be useful for stiffness if they are adjusted to the already located part without shifting it. Avoid redundant locators that fight one another.
Foolproofing (poka-yoke) makes an incorrect loading orientation difficult or impossible—for example, a locating feature that accepts an asymmetric part in only one orientation. Keep chips away from locating contacts with accessible cleaning paths and relief spaces; a trapped chip can relocate every subsequent cut.
Identify the intended locating contacts and check that the part seats without rocking.
Check that clamps press into supports and that loading, cutting and unloading paths are clear.
Check that incorrect orientation is prevented or clearly detectable.
Verify the first part after setup before repeating the cycle.
Modular fixtures trade dedicated tooling for reusable components; Practice
Modular fixtures trade dedicated tooling for reusable components
A modular fixture combines a reusable base with locating pins, supports and clamps. It suits changing work when the arrangement can be documented and rebuilt repeatably. A dedicated fixture may reduce loading time for repeated production, but takes its own design and manufacturing effort.
A tombstone is an upright fixture carrier with mounting faces that can present several setups to a machining centre. It still needs suitable locating, clamping and tool clearance on each face.
Select fixture materials for stiffness, wear, temperature change and the required accuracy. Aluminium can be useful; thermal expansion is a design consideration rather than an automatic ban on close-tolerance work. Wear surfaces may need hardened inserts or replaceable components.
See how programmed machining, tool changes, pallet handling and robots work together—and what must be reliable before unattended production makes sense.
NC and CNC put machining motions under program control
NC and CNC put machining motions under program control
Numerical control (NC) directs machine actions using coded instructions. Computer numerical control (CNC) uses a computer-based controller to execute those instructions. This is a base for automation, not a promise that loading, inspection and fault recovery are automatic too.
Programs still depend on correct tools, work location, offsets and a verified setup. See G-code and Toolpaths and CNC Control Fundamentals for the controls and program details.
Tool changers and pallet changers automate different waits
Tool changers and pallet changers automate different waits
An automatic tool changer exchanges cutting tools so a program can perform several operations. Correct tool identification, dimensions and available tool life remain essential.
A pallet changer exchanges work carriers. In a suitable system, one pallet can be loaded at a designated station while another is being machined. The pallet locates the fixture; the fixture locates the part. Dirt at either interface can undo the expected repeatability.
Repeatable fixturing and planned inspection are part of the automation system, not tasks that disappear after a changer is installed.
A manufacturing cell coordinates machines and handling
A manufacturing cell coordinates machines and handling
A flexible manufacturing cell combines machining and handling resources arranged to produce a planned family of parts. Flexibility comes from suitable programs, tools, fixtures and scheduling; it does not mean any part can be run without preparation.
A loading robot uses a programmed manipulator and a gripping device to transfer parts between defined positions. The cell must confirm conditions such as part presence, clamp state and machine readiness before the next action. A robot's repeatable motion does not compensate for a wrongly seated part.
Supporting detail: Addendum 112 — Cell, system and integration describe different scopes.
Lights-out production requires a plan for faults and variation
Lights-out production requires a plan for faults and variation
Lights-out machining means planned operation without an operator continuously attending the machines. It relies on a validated process, sufficient material and tool life, dependable chip and coolant management, and a defined response to faults.
Monitoring may detect tool breakage, missing parts or measurement trends, but only for conditions the system is designed to detect. Automatic compensation and inspection of every part are options, not universal features of automation. Decide what is measured, what stops production and how affected parts are contained.
Unattended operation is a system-level decision under the applicable equipment and facility procedures. These concepts are not authorization to leave a machine running unattended or enter an automated cell. A stopped cycle alone does not establish safe access.
Materials and Manufacturing Processes/Metals
Metals and Their Properties
How bonding, grains and alloy condition affect the properties used to choose a metal.
Metal selection starts with the job the part must do
Metal selection starts with the job the part must do
A metal can carry a load, conduct current, transfer heat or resist wear. Those are different jobs. Start with the required behavior, then compare a particular alloy in a particular condition. An alloy combines a base metal with other elements to change its behavior.
Metallic bonding helps explain conduction and forming
Metallic bonding helps explain conduction and forming
In a metal, some electrons are shared throughout the structure rather than tied to one bond between two atoms. Their mobility helps metals conduct electricity and heat. Metallic bonding can remain intact while atoms rearrange during permanent deformation, which helps explain why many metals can be rolled or drawn.
Most engineering metals contain many tiny crystals called grains. Atoms have an ordered arrangement inside each grain; adjacent grains have different orientations. Grain size, alloy chemistry and processing history all influence properties.
Supporting detail: Addendum 113 — Why two pieces of the same alloy can behave differently.
Strength, hardness, ductility and toughness answer different questions
Strength, hardness, ductility and toughness answer different questions
Mechanical properties used when comparing metals
Property
Question it answers
Selection example
Strength
How much stress can the material withstand? Yield strength concerns the start of permanent deformation; tensile strength concerns the maximum stress in a tensile test.
A loaded bracket must resist permanent bending.
Hardness
How strongly does the surface resist local indentation?
A wear surface may need hardness, but hardness alone does not predict every kind of wear.
Ductility
How much permanent deformation can occur before fracture?
A formed part needs enough ductility to survive bending.
Toughness
How much energy can the material absorb before fracture?
A part struck repeatedly needs resistance to fracture, not simply a hard surface.
Strong does not mean stiff. Stiffness describes elastic deflection. Hardening a steel can raise its strength substantially without a comparable change in its elastic stiffness. Geometry also affects how far a part bends.
Pull a tensile specimen: strength, stiffness, ductility and toughness on one curve; Interactive
Pull a tensile specimen: strength, stiffness, ductility and toughness on one curve
A tension test pulls a specimen of known size until it breaks and records force against stretch. Dividing by the original area and gauge length turns that record into an engineering stress–strain curve, and one curve answers four different questions: how stiff, how strong, how far it stretches and how much energy it absorbs.
The curves below are synthetic teaching curves, shaped to resemble common behaviour. Read the method from them, not the numbers for any real alloy.
Tensile test: one curve, four different properties. This fallback shows the same synthetic curves and calculations as the live lab.
A round specimen with a chosen gauge length L0 = 50 mm and diameter 12.5 mm (original area A0 = 122.7 mm²) is pulled by a moving crosshead. Engineering stress σ = F/A0 (N/mm² = MPa). Engineering strain ε = ΔL/L0 (mm/mm, shown here in %). Both use the original dimensions, so engineering stress falls after the neck forms even though the force is concentrated on a shrinking section.
Synthetic specimen A (ductile, low-carbon-steel-like; illustrative, not a grade). Shaded: area under the curve to fracture. Dashed red line: unloading from 10% strain.
Yield strength (0.2% offset): draw a line with slope E starting at 0.2% strain; where it meets the curve, σy ≈ 250 MPa (at 0.325% strain). Specimen A then stretches along a level yield plateau (stress constant at 250 MPa up to 1.5% strain) before strain hardening makes the curve rise again.
Ultimate tensile strength (UTS): the highest engineering stress, 400 MPa at 20% strain. A neck begins to form after this peak.
Toughness (tensile): the area under the curve to fracture, ∫σ dε ≈ 107 MJ/m³ (1 MPa × 1 = 1 MJ/m³ = 1 J/cm³) of work per original volume. For this gauge volume that is about 659 J. This is not fracture toughness KIC, which is measured on a cracked specimen.
Worked example: unloading leaves permanent set. Stop at ε = 10.0%, where σ = 371 MPa. Removing the load follows a straight line parallel to the elastic slope, so the elastic part σ/E = 371/200 000 = 0.185% springs back. Permanent set = 10.0% − 0.185% = 9.815%, or 4.91 mm on the 50 mm gauge length. Below the yield point the same rule gives zero set: at 0.1% strain and 200 MPa, 0.1% − 200/200 000 = 0.
The same specimen at four moments. Length to scale; the neck shape is illustrative only.
Before loading. Gauge length 50.0 mm (original L₀ = 50.0 mm). Drawn to scale.
At the peak (UTS): stretched uniformly, no neck yet. Gauge length 60.0 mm (original L₀ = 50.0 mm). Drawn to scale.
After the peak: deformation concentrates in a neck. Gauge length 63.5 mm (original L₀ = 50.0 mm). Drawn to scale; neck shape illustrative.
Fractured; pieces fitted together. Broken pieces fitted back together: gauge length 64.9 mm (original L₀ = 50.0 mm). Drawn to scale; neck shape illustrative.
Compare the four synthetic specimens. Each ranks differently on each property: the stiffest is not always the strongest, and the most ductile is not the toughest.
All four synthetic curves on one set of axes; the right panel magnifies the first 1% of strain, where the elastic slopes differ.
A Ductile, low-carbon-steel-like
B Aluminium-alloy-like
C Brittle, gray-iron-like
D Polymer-like (cold drawing)
Read from the four synthetic curves (illustrative; not material data)
Specimen
Stiffness E (GPa)
0.2% offset yield (MPa)
UTS (MPa)
Elongation after fracture (%)
Area to fracture (MJ/m³)
A: Ductile, low-carbon-steel-like
200
250
400
29.8
107
B: Aluminium-alloy-like
70
266
310
12.6
38
C: Brittle, gray-iron-like
100
none: fractures first
215
0.1
0.42
D: Polymer-like (cold drawing)
2.5
36.7
48.0
53.2
21
Specimen C fractures before the 0.2% offset line reaches its curve, so it has no offset yield strength in this test: the value is undefined, not zero. Specimen D (polymer-like) reaches its peak stress early and then draws a neck along the specimen; its offset value is shown only for comparison.
Model, assumptions and sources
The four curves are synthetic: piecewise cubic curves through chosen points, shaped to resemble common behaviour. They are not measurements, grades, certification values or design allowables. Real results depend on composition, processing, specimen geometry, test speed and temperature, none of which are modelled.
Unloading and reloading follow a straight line with the initial elastic slope E (an ideal elastic approximation, no hysteresis). The load never goes below zero: once the unload line reaches zero load the specimen is slack. Permanent set = ε − σ/E, never negative. Elongation after fracture uses the same rule at the fracture point.
The specimen length is drawn to scale. Its thinning before the peak keeps volume constant; the neck and the fracture are drawn only to show where deformation concentrates, not calculated from a necking or fracture law. True stress, strain rate, temperature and hardness are outside this lab.
The area under the curve is work per original volume in this tension test (MPa = MJ/m³). Impact and fracture-toughness tests measure related but different behaviour.
Sources: ASTM E8/E8M, Standard Test Methods for Tension Testing of Metallic Materials (the test method and its definitions; consult the current edition). W. D. Callister Jr. and D. G. Rethwisch, Materials Science and Engineering: An Introduction, chapter “Mechanical Properties of Metals” (engineering stress and strain, modulus, 0.2% offset yield, tensile strength and necking, ductility, toughness as area under the curve, elastic recovery on unloading).
Density, conductivity and corrosion can decide the choice
Density, conductivity and corrosion can decide the choice
Density is mass per volume. Compare equal volumes when comparing densities; a large aluminum part can still weigh more than a small steel part. Use Geometry, Mass, and Force and Dimensional Analysis for mass, weight and unit calculations.
Conductivity describes how readily heat or current travels through a material. Metals vary widely: choosing a good conductor is different from choosing a high-strength alloy.
Corrosion resistance depends on the metal and its surroundings. Stainless steel and aluminum can corrode. Water, salts, trapped moisture and contact with other metals can change performance. A protective surface film helps only while the environment allows it to remain effective.
Supporting detail: Addendum 114 — Why connecting two different metals can change corrosion.
Ferrous means iron-based; a magnet does not identify the grade; Practice
Ferrous means iron-based; a magnet does not identify the grade
Ferrous metals have iron as their base: steels and cast irons belong here. Nonferrous metals have another base, such as aluminum, copper or titanium. Nonferrous does not mean that no trace of iron can be present.
Magnetic response is a clue, not a material certificate. Some stainless steels are weakly magnetic or effectively nonmagnetic in certain conditions. Appearance and a magnet cannot establish an alloy, its temper or its strength.
Carbon, alloy, stainless and tool steels, plus the difference between a steel family and a complete material specification.
Steel is iron-based, with carbon and other elements controlling behavior
Steel is iron-based, with carbon and other elements controlling behavior
Steel contains iron and carbon, often with additional alloying elements. Both composition and processing matter. The same grade can arrive softened for machining or hardened for service.
Use four practical groups: carbon steel, alloy steel, stainless steel and tool steel. These are useful buying categories, not four mutually exclusive chemical definitions: stainless and tool steels also contain alloying elements.
The four steel groups are a selection map. They do not replace grade and condition.
Carbon steel spans easy forming through hardenable cutting edges
Carbon steel spans easy forming through hardenable cutting edges
Relative carbon levels; actual grade limits come from the specification
Group
Typical behavior
Typical uses
Low carbon
Generally readily formed and welded; limited response to ordinary through-hardening.
Sheet parts, brackets and general fabrication.
Medium carbon
More potential for heat-treated strength; forming and welding require more consideration.
Shafts, axles and machinery parts.
High carbon
Can develop high hardness after suitable treatment; less forgiving of forming and impact.
Springs and some cutting tools.
These are trends, not a strength ranking of every item on a shelf. Compare the delivery condition and required properties before substituting one grade for another.
Alloy, stainless and tool steels solve different service problems
Alloy, stainless and tool steels solve different service problems
Choose a family, then qualify a specific grade
Family
Composition idea and purpose
Uses and limits
Alloy steel
Intentional additions such as chromium, nickel or molybdenum adjust properties and heat-treatment response.
Loaded shafts and gears. Alloy steel is not automatically corrosion resistant.
Stainless steel
Chromium supports a protective surface film; other elements adjust performance.
Food equipment, fittings and exposed hardware. Grade, cleaning and environment still matter.
Tool steel
Carbon and alloy additions support hardness, wear resistance or performance at elevated temperature.
Dies, punches and cutting tools. Heat treatment must balance wear resistance against chipping.
Use Heat Treating Metals to distinguish hardening from tempering. Hardenability means the ability to harden to depth under given cooling conditions; it is not the same as a measured surface hardness.
Grade labels identify a material; they are not a heat-treatment recipe
Grade labels identify a material; they are not a heat-treatment recipe
Common labels include 1018 and 1045 for carbon steels, 4140 for alloy steel, 304 and 316 for stainless steels, and O1, A2 and D2 for tool steels. Different designation systems have different rules. Do not apply a four-digit carbon-steel decoding rule to stainless or tool-steel labels.
On an order, identify the governing material specification, grade, delivery condition, form and dimensions. A short grade name alone does not specify strength, hardness, surface finish or dimensional tolerance. Ask for the required certification when traceability matters.
Use the current governing specification or the producer's grade sheet for composition limits and treatment requirements.
Decode carbon and alloy steel numbers; read stainless and tool grades separately
Decode carbon and alloy steel numbers; read stainless and tool grades separately
In the common AISI/SAE four-digit carbon and alloy steel system, the first two digits identify the steel series: 10 is plain carbon steel and 41 is chromium-molybdenum steel. The last two digits give nominal carbon content in hundredths of a percent.
Reading the series and nominal carbon content
Grade
Series
Nominal carbon
1018
10: plain carbon
18 → 0.18%
1045
10: plain carbon
45 → 0.45%
4140
41: chromium-molybdenum
40 → 0.40%
304 and 316 stainless: their chemistries are similar, but 316 has an added alloying element, molybdenum. This improves resistance to chloride exposure and urban or industrial pollution compared with 304.
Tool-steel groups: O1 is an oil-hardening grade, A2 is an air-hardening grade, and D2 is a high-carbon, high-chromium grade. These labels describe different tool-steel groups; they are not carbon-percentage codes.
Supporting detail: Addendum 115 — Sources and quoted designation meanings.
Cast iron is a separate iron-based material, and it can be machined
Cast iron is a separate iron-based material, and it can be machined
Cast irons generally contain more carbon than steels. Their properties depend strongly on how carbon appears in the structure. Gray iron contains graphite flakes and is useful for machine bases and housings where vibration damping matters; it is relatively brittle in tension. Ductile iron contains graphite nodules and can provide much greater ductility.
Casting creates the initial shape. Machining can then finish bores, faces and other accurate features. “Cast” and “machined” are not competing material identities. White cast iron contains substantial hard carbide and behaves differently from graphite-containing irons.
Watch the steel-family series at its original publisher; Practice
Watch the steel-family series at its original publisher
Aluminum, copper alloys, titanium, zinc and magnesium as choices for weight, conductivity, corrosion and manufacturing needs.
Choose the useful property, then check the tradeoff
Choose the useful property, then check the tradeoff
Nonferrous metals have a base other than iron. They are not all lightweight, corrosion-proof or easy to machine. A useful first comparison names both the benefit and the constraint.
Starting points for material selection, not design allowables
Metal
Why choose it?
Typical use and tradeoff
Aluminum
Low density, useful corrosion resistance and ready availability in extruded shapes.
Frames and housings. Lower elastic stiffness than steel; alloy and temper strongly affect strength.
Copper
High electrical and thermal conductivity.
Conductors and heat-transfer parts. Relatively high density and cost; alloying usually reduces conductivity.
Titanium
Useful strength relative to weight and corrosion resistance.
Aerospace and chemical-service parts. Material and processing costs can outweigh the weight saving.
Zinc
Useful casting behavior and ability to protect steel as a coating.
Die-cast hardware and galvanized products. Temperature and load requirements limit use.
Magnesium
Very low density among structural metals.
Light housings. Corrosion protection and suitable manufacturing procedures matter.
Aluminum needs both an alloy label and a condition
Aluminum needs both an alloy label and a condition
Aluminum is sold in many alloys and conditions. Wrought products are shaped by processes such as rolling or extrusion. Their familiar four-digit labels are organized into series, including 1xxx through 7xxx. A series is a family, not a hardness scale: a higher first digit does not guarantee a harder or better part.
6061 and 7075 are common alloy labels to look for in supplier literature. Compare data for the actual product form and temper rather than treating either number as a complete specification. Cast aluminum uses a different designation system; do not transfer wrought-number rules to cast products.
A suffix such as -O or -T6 identifies a condition as part of an alloy-temper designation. It must stay attached to the alloy on an order. See Heat Treating Metals for why processing history matters.
-O: “Annealed” — “Thermally treated to obtain the lowest strength temper.”
-T6: “Solution heat-treated and artificially aged.”
Use the governing product specification for an actual purchase.
Aluminum series 1xxx–7xxx identify purity or the main alloying additions
Aluminum series 1xxx–7xxx identify purity or the main alloying additions
Wrought aluminum series
Series
Principal alloying element or purity
1xxx
Commercially pure aluminum: at least 99% aluminum.
2xxx
Copper.
3xxx
Manganese.
4xxx
Silicon.
5xxx
Magnesium.
6xxx
Magnesium and silicon.
7xxx
Zinc.
Supporting detail: Addendum 116 — Source for the series table.
6061 favors fabrication; 7075 offers higher strength in the T6 condition
6061 favors fabrication; 7075 offers higher strength in the T6 condition
Compare like conditions: 7075-T6 has higher tensile strength than 6061-T6. The alloy number alone does not establish the strength of the material in hand.
6061 and 7075 at selection level
Alloy
Typical uses
Fabrication considerations
6061
Truck and marine frames; tanks.
Weldable and readily machined. Welding can change the temper in the weld area.
7075
Aircraft components, molds and industrial tooling.
Good machinability, but generally considered unsuitable for welding because of cracking susceptibility.
Supporting detail: Addendum 117 — Sources and comparison limits.
Brass and bronze trade some of copper's conductivity for other properties; Practice
Brass and bronze trade some of copper's conductivity for other properties
Brass is principally a copper-zinc alloy. Depending on grade, it can offer useful forming or machining behavior for fittings, hardware and decorative parts. Do not assume every brass has the same machinability or suitability for contact with drinking water.
Bronze traditionally means copper-tin alloy, but commercial names also include aluminum bronze and silicon bronze. Bronzes serve in bearings, bushings, gears and corrosion-resistant hardware, depending on grade. The word “bronze” alone is not enough to buy a bearing material.
For a conductor, begin with conductivity requirements. For a bearing, begin with load, speed, lubrication and wear. The same copper-based family will not automatically win both comparisons.
Choose a conditioning process by its goal: soften, refine structure, harden, improve toughness or reduce residual stress.
Heat treatment is conditioning: change structure to change properties
Heat treatment is conditioning: change structure to change properties
Heat treatment uses controlled heating, holding and cooling to change a metal's internal structure. It belongs to Conditioning in Manufacturing Process Categories. It can change properties without intentionally changing the overall shape, although distortion can occur.
The result depends on alloy, starting condition, section thickness and the complete cycle. Heating more or cooling faster is not automatically better. A treatment that strengthens one alloy can soften another.
Supporting detail: Addendum 118 — Steel microstructures are not the four steel families.
Anneal to soften; normalize suitable steels to refine structure
Anneal to soften; normalize suitable steels to refine structure
Annealing is a family of treatments used to soften material, restore ductility or adjust its structure. Full annealing of suitable steels includes slow cooling; annealing cold-worked metal can form new grains and restore formability. Not every anneal uses the same temperature or cycle.
Normalizing heats suitable steel into its austenitic range and then cools it in air. It can refine and make the structure more uniform after earlier processing. Air cooling and slow furnace cooling produce different results.
Choose the process for the desired condition. Do not apply one universal temperature offset to all steels.
Quench hardening usually needs tempering afterward; Practice
Quench hardening usually needs tempering afterward
Quench hardening heats suitable steel to form austenite, holds as required, then cools fast enough to obtain the intended hardened structure. It commonly raises hardness and strength but can reduce ductility and toughness. Cracking and distortion are possible.
Tempering reheats hardened steel below the temperature at which austenite begins to form. It adjusts the balance of hardness and toughness and reduces some quenching stresses. It is not simply heating “just below” one critical temperature; the required cycle depends on grade and target properties.
Separate illustrative cycles, not a furnace program. Peak temperatures, holds and cooling rates are grade-specific.
These concepts explain specifications; they are not operating instructions. Use the approved process for the identified material, equipment and quench medium.
Surface hardening puts wear resistance near the outside
Surface hardening puts wear resistance near the outside
A part may need a hard wear surface and a tougher interior. Carburizing adds carbon near the surface of suitable steel before hardening. Nitriding introduces nitrogen to develop a hard surface layer. These change surface chemistry.
Induction hardening heats a surface region electrically and then quenches it. It uses the steel's existing composition rather than adding carbon. It therefore needs a suitable starting steel. Case depth, surface hardness and core properties must be specified together.
Cold work strengthens; stress relief addresses locked-in stresses; Practice
Cold work strengthens; stress relief addresses locked-in stresses
Work hardening, also called strain hardening, occurs as metal is permanently deformed during cold working. It commonly raises strength and hardness while reducing remaining ductility. “Cold” here means below the relevant recrystallization range, not necessarily cold to the touch.
Stress relief reduces residual stresses left by processing. These are stresses that remain without an external load. The goal is often dimensional stability during later machining or service. Stress relief is not automatically a full softening anneal and does not guarantee zero distortion.
Aluminum strengthening does not use steel's martensite route
Aluminum strengthening does not use steel's martensite route
Some aluminum alloys can be strengthened by solution treatment, quenching and aging. Solution treatment puts selected elements into solid solution; subsequent aging develops small particles that obstruct deformation. Aging can occur naturally or through controlled heating, depending on the alloy and required condition.
Solid-solution strengthening and precipitation strengthening are different mechanisms. Dissolved atoms provide the first; particles of another phase provide the second. Other aluminum alloys obtain much of their strength through cold work. Annealing can soften cold-worked material.
An aluminum temper describes its processing condition. It is not synonymous with the steel process called tempering. Preserve suffixes such as -O and -T6 when reading or writing a material description; consult the relevant alloy-temper specification for their exact meanings.
Supporting detail: Addendum 119 — Further reading on conditioning processes.
Materials and Manufacturing Processes/Plastics
Cutting and Machining Plastics
Choose a separating process for the plastic and the feature, then control heat, chips and edge damage.
Choose the cut from the material and the feature
Choose the cut from the material and the feature
Separating removes material. A saw makes a blank, a drill makes a round hole, and a router or mill can make an outline, pocket or slot. Identify the resin and grade before choosing the process: two clear sheets can be acrylic and polycarbonate, with very different cutting behavior.
Process
Useful for
What controls the result
Sawing
Straight cuts or curved blanks on a band saw
Use a blade intended for the plastic and thickness. Teeth must cut without grabbing or packing the cut with chips; the finest tooth spacing is not always best.
Drilling
Holes through sheet or solid stock
Use suitable drill geometry, support the exit surface and provide chip clearance. Brittle acrylic can crack when the drill grabs at breakthrough.
Milling and routing
Accurate edges, slots and pockets
Sharp cutting edges, room for chips to escape and a supported workpiece reduce melting and chatter.
Laser cutting
Sheet profiles in an approved material
The resin, additives, thickness and laser wavelength matter. A glossy edge alone does not prove dimensional accuracy.
Hot-wire cutting
Contours in suitable thermoplastic foam
A heated wire softens a narrow path. This is not a general cutting method for solid plastic sheet or every foam formulation.
Heat builds up near the cut because plastics generally conduct heat poorly. Softened chips can stick to the tool or weld back onto the edge. The symptoms are smeared surfaces, strings of softened material and chips packed into the cutter.
Slowing the feed alone can make melting worse. Feed is how quickly the tool advances through the work; spindle speed is how quickly it turns. Too little advance per tooth lets the edge rub instead of taking a useful chip. Start with cutting data for the resin and tool, keep the tool sharp, and maintain chip clearance. Reducing excessive spindle speed may help, but feed, speed and cut depth must be considered together.
Deep holes need a planned way to clear chips. Coolants and cleaners must be compatible with the plastic; an unsuitable liquid can cause fine cracks called crazing. Use Feeds and Speeds for the relationships, not its metal cutting values as plastic settings.
Supporting detail: Addendum 120 — Why a slower cut can produce more heat.
Acrylic, PVC and polycarbonate behave differently under a laser
Acrylic, PVC and polycarbonate behave differently under a laser
Acrylic (PMMA) absorbs the infrared energy of a CO₂ laser. Concentrated heating removes material along the cut and can leave a smooth, glossy edge. That result depends on the sheet grade, thickness and process settings; it is not a promise for every clear plastic or every laser.
PVC is not a laser-cutting substitute for acrylic. Heating PVC can release hydrogen chloride; with moisture it forms hydrochloric acid, creating an exposure and corrosion concern. This is a chemical incompatibility, not simply a poor-looking edge.
Polycarbonate is often a poor choice for a clean CO₂-laser-cut sheet edge. It can yellow, char and leave a degraded edge instead of acrylic's clear finish. Specialized processes can cut some polycarbonate, so “polycarbonate cannot be laser processed” is too broad. Routing or sawing is often the better starting choice for a clear polycarbonate profile. Do not assign PVC's chlorine chemistry to polycarbonate.
Supporting detail: Addendum 121 — Manufacturer references and the limits of a material list.
Trim the molding excess without cutting into the part; Practice
Trim the molding excess without cutting into the part
A gate is the entry from a molding flow channel into the part cavity. Degating separates the part from that feed connection. Flash is a thin unwanted fin, often at a mold joint; trimming removes it. A burr is a raised edge left by cutting. These are different defects, even when the same trimming tool can remove them.
Support the part near the cut and leave the required contour intact. Cutting into the wall creates a notch that may start a crack. For batches, tumbling with suitable media can remove small burrs, but it can also round wanted edges or damage thin features. Inspect the dimensions and surface after the trial.
Sinker and wire EDM use controlled sparks to shape conductive work, with a gap between electrode and finished surface.
EDM needs an electrically conductive workpiece
EDM needs an electrically conductive workpiece
Electrical discharge machining (EDM) removes material with repeated electrical discharges across a small gap. Local heating melts and vaporizes tiny amounts of material. The electrode does not cut by rubbing against the work.
Check conductivity before considering EDM. Hardened tool steel is a candidate; ordinary nonconductive glass or ceramic is not. Hardness is less restrictive than in cutting with a conventional tool, but composition and electrical and thermal properties still affect machining behavior.
Supporting detail: Addendum 122 — Why the fluid and the gap matter.
Use sinker EDM for cavities and wire EDM for through profiles
Use sinker EDM for cavities and wire EDM for through profiles
Sinker, ram or die-sinking EDM feeds a shaped electrode toward the work to make a cavity. The cavity is larger than the electrode because removal extends across the spark gap. Mold cavities and hardened dies are typical uses. Electrode wear must be allowed for.
Wire EDM feeds a thin wire through the cutting zone while controlled motion traces a profile. The wire is continuously renewed. An enclosed opening normally needs a start hole for threading; the wire must pass through the work. A blind pocket calls for another approach, such as sinker EDM.
Wire EDM is useful for intricate die openings and thin features with low cutting force. Internal corners still have a radius: the wire radius plus the working gap limits how tight a corner can be. Specify a small radius rather than a mathematically sharp corner.
Offset the wire center by its radius plus the spark gap; Practice
Offset the wire center by its radius plus the spark gap
The wire runs perpendicular to this top view. Gap and wire are enlarged. The dashed line is the wire center path; the solid boundary is the finished wall.
Center-path offset = wire radius + spark gap. Measure this distance from the finished contour toward the material being removed. In a simple straight cut, the full cut width (kerf) is wire diameter plus twice the gap.
For an illustrative radius of 0.120 mm and a one-sided gap of 0.018 mm, the offset is 0.138 mm. These are teaching inputs, not machine settings. A control may apply compensation to a programmed part contour; do not also offset the geometry a second time.
Supporting detail: Addendum 123 — Why one offset does not describe every pass.
Low cutting force does not mean no thermal change
Low cutting force does not mean no thermal change
EDM can hold fine features and produce a fine finish with suitable finishing passes, but its surface contains small discharge craters. Melted material can solidify again as a recast layer. Heat can alter nearby material, so surface integrity and any required finishing belong in process selection.
Supporting detail: Addendum 124 — Further reading on EDM surfaces and wire systems.
ECM dissolves conductive metal through an electrolyte rather than removing it with sparks or a cutting edge.
ECM removes metal by electrochemical dissolution
ECM removes metal by electrochemical dissolution
In electrochemical machining (ECM), the metal workpiece is the positive electrode (anode), and the tool is the negative electrode (cathode). Direct current passes through a flowing, electrically conductive liquid called an electrolyte. Metal dissolves from the work surface while the tool advances without cutting contact.
Consider ECM for conductive metals, including hard alloys and complex contours. The tool shape, gap and electrolyte flow govern the resulting feature. A suitable electrolyte must be matched to the metal; a generic salt solution is not a universal process recipe.
ECM uses a conducting liquid; EDM uses controlled breakdown
ECM uses a conducting liquid; EDM uses controlled breakdown
Two electrically powered processes with different removal mechanisms
Question
ECM
EDM
What removes material?
Electrochemical dissolution
Heat from discharges
What does the liquid do?
Conducts current and carries reaction products away
Insulates between discharges; cools and flushes
Does the tool wear?
No removal-related tool wear in principle; real tooling still needs maintenance
Electrode wear occurs; wire is renewed
What surface results?
Can be smooth and burr-free, without a thermal recast layer
Discharge craters and possible recast material
ECM is not a melting process and normally avoids a heat-affected zone. That does not mean the electrolyte stays cold: current flow heats it. Flow, temperature, chemistry and gap control all matter.
Supporting detail: Addendum 125 — Why hardness alone does not set ECM performance.
Compare focused light, moving water and a plasma arc by material, thickness, edge quality and thermal effects.
Distinguish a thermal cut from an erosive cut
Distinguish a thermal cut from an erosive cut
Laser and plasma use concentrated heat. Waterjet uses mechanical energy. A heat-affected zone (HAZ) is material beside a cut whose properties or structure have changed from heating without being melted away.
Conceptual cross-sections, not a scale comparison of kerf or HAZ widths. Kerf means the width removed by a cut.
Laser cutting concentrates light into a small region
Laser cutting concentrates light into a small region
A focused laser beam supplies thermal energy; cutting gas commonly helps remove material. Laser cutting is a candidate for fine profiles in sheet material. Metals, wood and selected plastics require suitable laser systems; a machine that cuts acrylic is not automatically a metal cutter.
A narrow kerf and fine detail are possible, particularly in thinner material. Edge finish, taper and dimensional accuracy depend on material, thickness, focus, gas and cutting speed. Melting, discoloration or charring may affect the edge. Do not assume the HAZ is zero.
Supporting detail: Addendum 126 — Why the material and laser must be matched.
Waterjet cuts soft materials; added abrasive cuts hard materials
Waterjet cuts soft materials; added abrasive cuts hard materials
Pure waterjet uses a high-speed water stream for materials such as foam, rubber and other soft sheet. Abrasive waterjet adds hard particles to erode metals, stone, non-tempered glass and many ceramics. Electrical conductivity is not required.
Waterjet normally creates no thermal HAZ. It is a candidate for thick stock or a heat-sensitive edge, but thickness still affects speed and cut quality. Taper, where the cut width changes through the thickness, and striations, visible lines down the cut face, can limit accuracy and finish. Slower quality cuts may reduce these effects. Brittle edges may chip and need finishing.
Identify the glass type before choosing a process: tempered glass is not a candidate for ordinary waterjet cutting. A no-HAZ process is not a promise of no mechanical damage.
Supporting detail: Addendum 127 — Further reading on glass and waterjet limits.
Plasma is a thermal cutting candidate for conductive metal
Plasma is a thermal cutting candidate for conductive metal
A plasma cutter uses an electric arc through ionized gas to melt metal; the gas stream expels it. The work must conduct electricity. Steel, stainless steel and aluminum are candidates; ordinary glass and wood are not.
Plasma often suits productive plate cutting where its edge quality meets the drawing or later machining is planned. Compared with a fine laser cut in thin sheet, small details can be harder to hold. Edge bevel, attached solidified metal (dross) and a HAZ may need attention. Equipment and thickness determine actual capability; plasma is not automatically a rough-only process.
Supporting detail: Addendum 128 — Compare capability on the actual thickness.
A vibrating tool drives abrasive particles against hard, brittle work without requiring electrical conductivity.
Ultrasonic impact machining removes material with abrasive particles
Ultrasonic impact machining removes material with abrasive particles
A shaped tool vibrates rapidly with a small stroke, driving abrasive particles in a liquid slurry against the work. The removal energy at the work is mechanical, even though electricity powers the vibration system.
Consider this process for holes or shaped cavities in hard, brittle materials such as glass and ceramics. The work need not conduct electricity. It is generally less effective for soft, ductile materials that deform instead of breaking away under particle impacts.
No thermal HAZ does not mean a perfect surface
No thermal HAZ does not mean a perfect surface
Ultrasonic impact machining normally avoids a thermal heat-affected zone. Abrasive size, tool wear, slurry access and feature depth affect dimensional control and finish. Fine abrasive can improve the surface, but removal may be slow; edge chipping and subsurface damage remain possible in brittle work.
Distinguish this process from ultrasonic-assisted cutting, where vibration is added to another cutting or grinding process. The two names do not imply identical tooling or removal mechanisms.
Supporting detail: Addendum 129 — Why a shaped tool does not give an exact-size copy.
A mask protects selected areas while an etchant dissolves exposed metal; sideways etching limits the detail.
A mask defines where chemical removal occurs
A mask defines where chemical removal occurs
Chemical machining uses a chemical reaction to dissolve exposed material. A protective mask covers areas that should remain. Chemical milling can make shallow recesses or reduce a metal part's weight. Photochemical etching uses a patterned resist and can produce intricate thin-sheet parts, including through openings.
Unlike ECM, ordinary chemical etching does not use an external machining current between a tool and the work. Select an etchant compatible with the particular material; “chemical machining” does not name one chemistry that works on everything.
Allow for sideways etching beneath the mask
Allow for sideways etching beneath the mask
Etching can spread sideways beneath the mask edge as well as into the surface. This undercut limits narrow lands, corner sharpness and dimensional control, especially as etch depth increases. The mask design must compensate for the etching behavior. Surface finish depends on the material and chemistry. Chemical removal avoids a cutting-induced thermal HAZ, but that is not the same as saying the chemical bath has no temperature requirements.
Supporting detail: Addendum 130 — Why thin-sheet patterns are a useful application.
Mark the feature: through profile, blind cavity, shallow recess or finished surface. Identify tool access and any start-hole requirement.
Record thickness, minimum internal radius and the thinnest remaining feature.
Read the drawing's dimensional and surface requirements. Decide whether a thermal HAZ or recast layer is acceptable.
Shortlist processes, then check actual equipment capability, finishing needs and production quantity.
Hardness alone does not select a process. A hard nonconductive ceramic rules out ordinary EDM even though EDM can cut hardened steel. See Tolerances and Fits and Surface Texture and Finish for what the finished part must meet.
Use the feature to narrow the candidates
Use the feature to narrow the candidates
Starting candidates, not guaranteed process approvals
Part requirement
Candidate and reason
Check next
Hardened steel die with a through opening and small inside radii
Wire EDM: conductive work, low cutting force, fine profile
Wire radius plus gap, threading access and surface integrity
Keep conventional machining on the candidate list; Practice
Keep conventional machining on the candidate list
For an accessible flat face, hole or round shaft in a machinable material, conventional cutting may be the simplest route. Use the family map to reach those methods. A part may combine sawing, milling and grinding, or use waterjet for a blank followed by machining of critical faces.
Increasing thickness can increase cutting time, taper or flushing difficulty. A thin remaining wall may deflect or be damaged even when cutting forces are small. Ask for evidence of capability on the actual material and geometry rather than relying on a universal tolerance ranking.
Addendum 1. “It’s not the voltage that gets you, it’s the current”
The saying is true. What usually gets taken from it — that the voltage does not matter
— is not. Current is electricity actually moving through you, and it takes very little of
it to do real harm. Volts sitting on a wire nobody is touching do nothing at all.
Two things together decide how much current you get: how hard the electricity is being pushed,
which is what voltage measures, and how hard your body holds it back, which is called resistance.
Harder push, more current. More resistance, less.
Your resistance is not a fixed number — wet skin holds current back far less than dry
skin does, so the same push moves much more current through you. That half of it cannot be relied
on, which is why the line is drawn on the half that can be: the push.
Addendum 2. The international standards are not one number
On datasheets you will meet a different set of figures, and there the AC and DC limits
are not the same:
50 V AC / 120 V DC
Extra-Low Voltage under
IEC 61140 and IEC 60364-4-41 — 50 V AC rms, or 120 V
ripple-free DC. “Ripple-free” is a real condition, not a figure of speech:
past roughly 10 % rms ripple the supply is treated as AC and the lower number applies.
30 V AC / 60 V DC
Class ES1 in IEC 62368-1, the
standard behind most benchtop and consumer equipment. ES1 is the class that needs no safeguard
at all; ES2 runs up to 50 V AC / 120 V DC and requires one.
Addendum 3. All sixteen sections, and the four OSHA does not enforce
This is OSHA's own list, from Appendix D to 29 CFR 1910.1200. The numbers
and headings are fixed; a manufacturer cannot renumber them or reorder them.
Section 1
Identification — product identifier, recommended use and
restrictions on use, who made it, emergency phone number.
Section 2
Hazard(s) identification — classification, signal word,
hazard statement(s), symbol(s) and precautionary statement(s).
Section 3
Composition / information on ingredients — chemical name,
CAS number, and for a mixture, the ingredients classified as health hazards and how much of
each.
Section 4
First-aid measures — by route of exposure, plus the symptoms
that matter and when to get medical attention.
Section 5
Fire-fighting measures — suitable and unsuitable
extinguishing media, and what the chemical gives off when it burns.
Section 6
Accidental release measures — the spill response.
Section 7
Handling and storage — "conditions for safe storage,
including any incompatibilities".
Section 8
Exposure controls / personal protection — the OSHA
permissible exposure limit, the ACGIH threshold limit value, engineering controls, PPE.
Section 9
Physical and chemical properties — physical state, odor,
flash point, auto-ignition temperature, density, pH and the rest.
Section 10
Stability and reactivity — reactivity, conditions to avoid,
"incompatible materials", hazardous decomposition products.
Section 11
Toxicological information — the health effects and the data
behind them, including any carcinogen listings.
Section 12
Ecological information
Section 13
Disposal considerations
Section 14
Transport information
Section 15
Regulatory information
Section 16
Other information — "the date of preparation of the SDS
or the last change to it".
Twelve through fifteen are the odd ones out.
The headings have to be there: a note to the standard says an SDS "must also include the
headings" for those four, "in order", to stay consistent with the international
system the format comes from. What goes under them is a different question. Appendix D
says "Sections 12-15 may be included in the SDS, but are not mandatory", and a second
note to the same paragraph says "OSHA will not be enforcing information requirements in
sections 12 through 15, as these areas are not under its jurisdiction."
Those four cover the environment, waste, shipping and other agencies' regulations —
EPA and DOT ground, not OSHA's. So Section 13 tells you what the manufacturer suggests doing
with the leftovers. It does not tell you what happens to them here. Ask.
Two more things about the shape of the document. The section order is fixed
but the order within a section is not — preparers "are not required to
present the information in any particular order within each section", so two sheets can
look different inside the same numbered box. And a heading with nothing under it is not an
oversight: where there is no relevant information, "the SDS shall clearly indicate that no
applicable information is available." Read that as nobody has the data — not as
nothing to worry about.
Addendum 4. Signal words, pictograms and H-numbers — how Section 2 is written
There are exactly two signal words, and OSHA defines them in
1910.1200(c): a signal word indicates "the relative level of severity of
hazard", and "'Danger' is used for the more severe hazards, while 'warning' is used
for the less severe."
Severity is relative to the same kind of hazard, and that is the part that catches
people. Flammable liquids in OSHA's Appendix C are the clean example — all three
categories carry the same Flame pictogram, and the signal word is the only
thing separating them:
Category 1
Danger — "Extremely flammable liquid and vapor"
Category 2
Danger — "Highly flammable liquid and vapor"
Category 3
Warning — "Flammable liquid and vapor"
So “warning” does not mean mild. A Category 3
flammable liquid still catches fire. The signal word ranks a hazard against others of its own
kind; it is not a statement that the chemical is safe.
The pictograms are a fixed set of eight — a black symbol in a
red-bordered diamond standing on a point. OSHA names them: Flame, Flame Over Circle,
Exclamation Mark, Exploding Bomb, Corrosion, Gas Cylinder, Health Hazard, and Skull and
Crossbones.
The eight OSHA Hazard Communication pictograms
Flame
Flame Over Circle
Exclamation Mark
Exploding Bomb
Corrosion
Gas Cylinder
Health Hazard
Skull and Crossbones
The H-numbers are an index, not the warning. Real sheets often print a code
such as H225 beside a hazard statement. Those codes come from the international GHS
scheme this format is built on, which treats them as reference labels rather than as a
substitute for the statement itself. OSHA's Appendix C does not use them at all — it
specifies the sentence. Read the sentence.
Addendum 5. Telling a current sheet from an old one
Section 16 carries "the date of preparation of the SDS or the last
change to it." That is the only date on the document, and it is how you tell a live sheet
from one that has sat in a binder for a decade.
Keeping it current is the manufacturer's job, not yours. Under
1910.1200(g)(5), when a manufacturer or importer "becomes newly aware of
any significant information regarding the hazards of a chemical, or ways to protect against the
hazards, this new information shall be added to the safety data sheet within three
months."
Addendum 6. What makes a flammables cabinet “approved”, and how to spot one
OSHA calls it a storage cabinet and defines it by a fire test rather than by
how solid it looks. 29 CFR 1910.106(d)(3) requires cabinets to "limit the
internal temperature to not more than 325 °F. when subjected to a 10-minute fire test",
with all joints and seams staying tight and the door staying closed throughout.
The same paragraph requires the marking, which is the fastest way to recognise one:
"Cabinets shall be labeled in conspicuous lettering, 'Flammable - Keep Fire Away.'"
Addendum 7. The word for this, and where it comes from
“Nip point” is not shop slang. It is the term the machine-guarding standard uses.
29 CFR 1910.212(a)(1) requires guarding against “hazards such as
those created by point of operation, ingoing nip points, rotating parts, flying chips and
sparks.”
OSHA’s Machine Guarding eTool puts it plainly: “In-running nip point
hazards are caused by the rotating parts on machinery.” Its longer treatment,
Concepts and Techniques of Machine Safeguarding (OSHA 3067), gives three kinds:
“Parts can rotate in opposite directions while their axes are parallel to each other.
These parts may be in contact (producing a nip point) or in close proximity.” —
meshing gears.
“Nip points are also created between rotating and tangentially moving parts. Some
examples would be: the point of contact between a power transmission belt and its pulley, a
chain and a sprocket, and a rack and pinion.”
“Nip points can occur between rotating and fixed parts which create a shearing,
crushing, or abrading action.”
So belts, chains, sprockets and gears are rotating hazards by name. What makes them
dangerous is that they draw material in — not that two parts close on each other.
OSHA is just as direct about the plain shaft: “Rotating motion can be dangerous; even
smooth, slowly rotating shafts can grip hair and clothing, and through minor contact force the
hand and arm into a dangerous position.”
These parts are meant to be behind covers, and the standard is specific about it —
29 CFR 1910.219(f)(3): “All sprocket wheels and chains shall be
enclosed unless they are more than seven (7) feet above the floor or platform.”
Addendum 8. What the investigations actually describe
Oregon’s state fatality-investigation program, summarizing the record for this class of
machine: “The most common causes of death and injury from metal lathes include
entanglement of clothing in moving parts, being hit by loose objects on the lathe (e.g.,
chuck keys, tools, metal shavings) and being struck by a part that has not been adequately
secured.” (OR-FACE 2012-18-1, Experienced Journeyman Machinist Killed While
Operating an Engine Lathe, 2014.)
Four investigated deaths, all the same event — something attached to a person touched
something turning, and the machine did the rest:
Smoothing a part while it turned. A 69-year-old machinist with over 30
years’ experience “reached over the work piece to smooth out a radiused transition
with emery cloth and a file in hand” and “his loose clothing (left arm sleeve)
became entangled around the rotating work piece”. A co-worker 30 feet away heard “a
strange ‘winding-up’ sound”. Of that same part, the shop foreman had said that
“the rough surface could grab clothing or gloves”. (OR-FACE 2012-18-1.)
Stock sticking out the back of the spindle. A 52-year-old machinist was
set up on a 103-inch steel bar: “Due to the length of the material, 24 inches were
sticking out the back of the lathe, unguarded, and held in place with a clamping device that
rotates as the lathe rotates.” He reached across the running machine and “the sleeve
of his shirt became entangled”. (Kentucky FACE 21KY002, 2021.)
A burr on the end of the bar. A lathe operator in his twenties had a
1 7/8-inch stub of 1-inch bar standing past the tailstock. “The right pocket of his
shirt was caught by the rotating tailstock burr. His shirt wrapped around the stock and
tightened at the decedent’s neck and chest.” The listed contributing factor is
“walking to and/or reaching over unguarded rotating bar stock”. OSHA makes the same
point in general: the danger “increases when projections such as set screws, bolts, nicks,
abrasions, and projecting keys… are exposed on rotating parts.”
(Michigan FACE 14MI034, 2014; OSHA Machine Guarding eTool.)
A drill press. “As Employee #1 reached into the unguarded point
of operation, his shirt sleeve or gloves became entangled in the drill bit. Employee #1 was
pulled toward the machine, entangling Employee #1’s long hair. The drill press
continued to spin, completely pulling Employee #1 into the operating area.” She died
of her injuries. The same report records that the spindle “had speeds that ranged from 45
to 1115 rotations per minute”. (OSHA accident 114097.015, 20 February 2019.)
Why abrasive cloth keeps appearing. The UK Health and Safety Executive
publishes a sheet on this one task: “Every year there are serious accidents involving the
use of emery cloth on metalworking lathes. These accidents result in serious injuries such as
broken bones, dislocations, lacerations, amputations and occasionally death.” The mechanism
it names is the one above: “Accidents occur when hand-held emery cloth is entangled in the
rotating parts of the machine or where the emery cloth is snagged on the rotating component
dragging the operator into the danger zone.” Its own rules are that emery
cloth “should NEVER be applied directly by hand on manual metalworking lathes”, and
that gloves “increase the risk of entanglement and” are “never acceptable near
rotating parts of machinery”. (HSE Engineering Information Sheet No 2, rev 1,
2015 — that is HSE’s guidance; ask what the procedure is in the shop you are
standing in.)
And the drill press is not the exception it looks like. OSHA’s own
drill-press training script lists, among the ways accidents happen, “getting dangling
jewelry, loose clothing, or hair caught in the drill chuck (rotational hazard)”, and states:
“The primary hazards of drill presses are contact at the point of operation and
rotational hazards. An operator can risk serious hand injury when working too
close to the drilling area, wearing gloves, loose clothing, loose hair, or jewelry.”
Addendum 9. Why belts and chains count as rotating, not pinching
OSHA sorts machine hazards by the motion that causes them, and “pinch point” is
the vocabulary it attaches to parts that travel, not to parts that turn. From
Concepts and Techniques of Machine Safeguarding (OSHA 3067):
Reciprocating motion — “Reciprocating motions may be hazardous
because, during the back-and-forth or up-and-down motion, a worker may be struck by or caught
between a moving and a stationary part.”
Transverse motion, which means movement in a straight continuous line
— “Transverse motion (movement in a straight, continuous line) creates a hazard
because a worker may be struck or caught in a pinch or shear point by the
moving part.”
A belt on a pulley is a rotating hazard that draws material in; a table running up against a
column is a moving hazard that closes on it. Both can take a hand, and the vocabulary is worth
keeping straight because the two do not have the same tell. You avoid the first
by giving it nothing to grip. You avoid the second by knowing where the gap is going.
Guarding is not a local preference. 29 CFR 1910.212(a)(1) —
quoted already in Rotating machinery above for its definition of a nip point — is
a guarding requirement first: one or more methods of machine guarding shall be provided to
protect the operator and other employees from hazards including those created by point of
operation, ingoing nip points, rotating parts, flying chips and sparks.
The reason a rule like this is written without exceptions is that the exceptions are exactly
where people get hurt. A guard comes off for a legitimate setup, the setup takes longer than
expected, somebody else walks up to a machine that looks normal and is not. The standard's own
answer to “but this task genuinely needs the guard off” is not a more
careful operator — it is Lockout/Tagout, the next card on this page, which is
a procedure rather than a permission.
29 CFR 1910.147(a)(1)(i): “This standard covers the servicing
and maintenance of machines and equipment in which the unexpected energization or start
up of the machines or equipment, or release of stored energy could cause injury
to employees.”
Two things are worth reading twice. Unexpected is the whole point — the
hazard is not that the machine runs, it is that it runs while someone has a hand inside it. And
stored energy is named alongside energization, because a machine that has been
switched off and unplugged can still contain everything it needs to move once.
“Servicing and maintenance” is much broader than it sounds. The standard defines it
as “constructing, installing, setting up, adjusting, inspecting, modifying, and
maintaining and/or servicing machines or equipment”, and adds: “These
activities include lubrication, cleaning or unjamming of machines or equipment
and making adjustments or tool changes, where the employee may be exposed to the
unexpected energization or startup of the equipment or release of hazardous energy.” Setting
up, clearing a jam and changing a tool are all on that list.
Stored energy gets its own requirement — 1910.147(d)(5)(i):
“Following the application of lockout or tagout devices to energy isolating devices, all
potentially hazardous stored or residual energy shall be relieved, disconnected, restrained, and
otherwise rendered safe.” The lock comes first, and then the energy already in the machine
is dealt with. Both, in that order.
Addendum 12. Why the chuck key rule is worded as a grip and not as a reminder
A key left in a chuck is launched the instant the machine starts. Written as
“remember to remove the key” the rule depends on you remembering at the one
moment you are busy with something else; written as the key never leaves your hand there
is nothing to remember, because a key that is still in the chuck is still in your hand and you
cannot start the machine.
Do not let a spring be the thing that saves you. Some chuck keys are
spring-loaded and eject themselves when you let go — ours are, on the lathe. Not all keys
are, and our Jacobs chuck keys are not. So the habit has to be the same on every
machine, or it is a habit that works until you walk up to a different chuck.
A drill-press chuck key is lighter than a lathe chuck key and carries far less energy. It is
still dangerous: it is in the list of things that become projectiles in Rotating
machinery above, and “less mass” is not the same as “will not hurt you”
when it leaves a spinning chuck at the distance you are standing.
Addendum 13. Why dry sand is on the list, and why it is probably not on yours
Sand does work. The Aluminum Association’s Guidelines for Handling Aluminum
Fines (F-1) lists fine dry sand, “preferably finer than 20 mesh,” next to
approved dry powders as effective for isolating an aluminum fines fire. It also says
what has to be true first.
Dry. F-1 says dry sand. Damp sand is water, and water
“will react with the aluminum fines to form flammable hydrogen gas.”
Already in the room. “A plentiful supply should be easily available
in the operating area and stored in covered bins or drums.” Sand fetched after the fire
starts is not that.
With the right tool waiting beside it. “Long handled shovels should
be available at each bin or drum, and should be constructed of non-sparking metal.”
Laid, not thrown. The shovels “lay the sand or powder around the
edge or perimeter of the fire.” You are walling it in. Anything thrown at the pile
raises the dust cloud that is the actual explosion.
Decided before it started, by someone senior. “Qualified
supervisory personnel should decide in advance whether it is better to attempt to fight the
fire or to withdraw to a safe location and allow it to burn itself out.”
Those are preconditions, not tips. If you cannot point at the covered sand bin and
the non-sparking shovel beside it right now, you do not have this option — and
the call to fight it was never yours to make on the spot. Alarm, evacuate, report.
Addendum 14. Why larger drills usually need fewer revolutions per minute
At the same spindle speed, a larger bit's outer edge travels faster. Diameter, work material and bit material all affect the speed choice. “Larger or harder means slower” is a starting trend, not a speed setting.
Addendum 15. What a sweep can and cannot establish
A sweep combines the part's shape with its orientation and support. Runout is not a direct measurement of formal concentricity. A height sweep relative to a plate can reveal variation, but calling it a flatness inspection requires the proper setup and evaluation. Follow the drawing's specified inspection method.
Consult the actual tool's instructions and specifications. The Mitutoyo Quick Guide covers micrometers, calipers, height gages, gauge blocks, and indicators. The Starrett catalog separates graduations from accuracy and illustrates several instrument arrangements. The Mitutoyo ABSOLUTE caliper specifications provide examples of 0.0005 in / 0.01 mm display increments.
Addendum 17. Other 0.001 in verniers can have different spacing
A 50-division vernier can span 49 main-scale spaces of 0.050 in. Each vernier space is 0.049 in, again leaving a 0.001 in difference. Read the actual main scale; do not apply the trainer's 0.025 in spacing to that tool. The Starrett catalog, page 102, illustrates the 50-division arrangement.
For a small angle, the error is approximately the offset multiplied by the tilt in radians. An offset of 2 in and a tilt of 0.0005 rad give about 0.001 in of error. This is distinct from simply looking at the scale from an angle. See the Abbe discussion in the Mitutoyo Quick Guide, page 11.
If the lower edge of a slot is 0.600 in above the reference and the upper edge is 0.900 in above it, the center is (0.600 + 0.900) / 2 = 0.750 in above that reference. Half the difference, 0.150 in, is half the slot height; it still must be added to the lower height.
On the arrangement described in the Starrett catalog, page 306, 12 vernier spaces span 23 main-scale degrees. Each vernier space is 23/12 degrees, or 1°55′: 5 minutes short of two degrees. The line alignment reveals that difference. Page 308 distinguishes a plain universal bevel from the graduated protractor used with it.
Addendum 23. Why the family name does not settle performance
Base fluid, additives, dilution and water quality all affect the result. A neat oil can use mineral, vegetable or other base oils; “straight” describes using it undiluted. A synthetic is not automatically harmless or maintenance-free. See the fluid-family comparison on page 7 of Washington State Department of Ecology's metalworking-fluid guide and HSE's exposure overview.
Addendum 24. Why some milling runs dry while other milling needs fluid
A milling tooth repeatedly enters and leaves the material. Coolant can increase its temperature swing and promote thermal cracking in susceptible edges. Other operations need lubrication to limit adhesion, cooling to control distortion, or flow to clear chips. Sandvik Coromant's dry and wet milling guidance explains these competing demands.
Cimcool's refractometer FAQ covers zeroing, the boundary and factor conversion. Master Fluids' concentration guide discusses product factors and contamination. A refractometer is a concentration estimate, not a bacteria test or proof that the fluid is fit for reuse.
The aim is small oil droplets dispersed throughout the water, rather than water trapped inside an oil-rich mixture. Cimcool's mixing explanation describes the emulsifier's role. Mixing order is separate from the arithmetic: the right measured amounts still need the right preparation method.
Addendum 27. Derive the balance, or solve for an unknown addition volume
Concentrate already in the tank plus concentrate added equals concentrate in the final tank: VC/100 + AP/100 = FT/100. Multiply through by 100 and solve for P. This tracks concentrate, not the Brix number.
If a premix concentration P is already chosen, solve for its addition volume: A = V(T − C) ÷ (P − T). Check that A is positive and that V + A fits within the allowed fill level. If P = T and C differs from T, no finite addition reaches T exactly. If C = T already, adding premix at T preserves it.
Addendum 28. Why inspection and measurement belong together
Concentration alone does not describe contamination or biological condition. HSE's metalworking-fluid guidance calls for maintaining fluid quality and checking bacterial contamination. A supplier-directed fluid-management program may include pH and microbial checks as well as concentration; use its methods and action limits. Master Fluid Solutions' FAQ explains odor, residue and concentration-related skin irritation.
Addendum 29. Supplier instructions determine the cleaning treatment
HSE's sump-cleaning guidance emphasizes supplier-selected cleaner and circulation times. An extensively contaminated system may need treatment before draining and additional cleaning cycles; the general checklist is not a microbial decontamination recipe.
Addendum 30. Why the waste route depends on what is in the tank
Washington State Department of Ecology's metalworking-fluid guide, pages 12–13, illustrates how formulation and contamination change waste management. Its state-specific rules are not a disposal authorization elsewhere. EPA's used-oil guidance also distinguishes qualifying used oil from other wastes and notes that state requirements can be stricter. The product SDS is a starting point; used coolant also contains material picked up during machining.
Addendum 31. Why finer teeth do not fix every bad cut
More teeth share the load, but smaller gullets have less room for chips. Too much feed can overload the teeth; too little can make them rub. A wandering cut also calls for checking the blade condition, tension, guides, and workholding against the manuals. Changing speed alone cannot correct loose work or a worn guide.
Addendum 32. Why the answer can be a fraction of a tooth
The product counts tooth spacings across the contact length. Actual engaged tooth tips are discrete, and the number changes as the blade moves. Entry, exit, variable pitch, and curved sections also change contact. This estimate explains the selection principle; it does not guarantee a fixed count throughout a real cut.
Addendum 33. Manufacturer resources for the next selection step
Consult the chart supplied with the blade. Public starting points include Starrett's carbon blade range and DoALL's blade FAQ. Product-specific recommendations govern; the estimates here are teaching calculations.
Addendum 34. Why relief cuts help with some outlines
A relief cut runs in from a scrap edge toward the outline, allowing waste to separate in smaller pieces. Plan relief cuts on the waste side and preserve the finishing allowance. They can provide room to approach a contour from another direction; they do not make the blade itself capable of a smaller turning radius.
Addendum 35. Why joining includes more than making a weld
Alignment keeps the band running straight. The prescribed annealing or tempering cycle restores suitable flexibility at the joint, and finishing removes raised weld material. A sound-looking joint alone is not proof that the loop is ready to run. DoALL's blade welder range is a starting point for equipment-specific documentation.
A tilted spindle puts opposite sides of a face cutter at different heights. Adjacent passes may leave steps even with the vise correctly aligned. Surface finish alone does not identify the cause: check the geometry before changing cutting data.
Addendum 38. Read the whole sweep, not just two endpoints
The simplified direction diagram is another view of the linear preset. Jaw tilt is exaggerated; the traverse arrow follows the contact point and does not name physical table travel.
A smooth change suggests angular misalignment. A sudden jump at one spot suggests a chip, burr, damaged jaw or unstable contact. Tightening can move the vise, which is why the final sweep matters. Equal endpoint readings alone cannot reveal a bowed or damaged face between them.
Addendum 39. Why the force divides this way—and what the model cannot tell you
Take moments about the heel: Fpart(a + b) = Pb. Vertical balance gives Fpart + Fheel = P. These are upward reactions on the strap; the strap pushes downward on the part and heel with equal magnitudes.
This neglects strap weight, friction, deformation and additional contacts. It does not calculate stud force from wrench torque, allowable loads, sliding resistance or the number of clamps needed. Those depend on the hardware, workpiece and cutting forces. See Carr Lane's strap-clamp guide for the underlying geometry and hardware guidance.
Mitutoyo: Dial Test Indicators and the Cosine Effect shows the stylus geometry and correction. Its What is Cosine Error? video demonstrates the difference using a test indicator against a plunger indicator. Do not apply a plunger-indicator correction rule to a lever indicator without checking its instructions.
Addendum 41. Check both repeatability and the offset sign
Repeat the touch from the same direction. On manual feed screws, account for backlash when reversing; a handwheel reading can change while the table has not yet moved. Do not use a rough saw edge as though it were a finished datum. The radius correction is geometry, independent of the control's button sequence.
Addendum 44. A real trade-off within one product family
BIG DAISHOWA's mechanical-chuck guidance describes two clamping configurations for its slim HMC(S): the roughing configuration increases clamping force at the expense of greater runout. That is a qualified product example, not permission to tighten another chuck farther or a ranking of all milling chucks.
The eccentric ring retains and extracts the collet as the nut is loosened. Missing that engagement can prevent correct seating. Use REGO-FIX's ER assembly instructions for the mechanism and Haas: ER Collet Essentials for a public demonstration.
Addendum 46. Why adding the tool can change the answer
HAIMER's balancing guide (PDF) discusses assembly unbalance. Albrecht's APC catalog (PDF), Handling APC and Ultra, states that the marked balance condition excludes collet and tool and that shank slots affect the assembled result. A small runout reading is not a substitute for that separate balance and speed check.
Addendum 47. Check your selection reasoning without a video
A defensible answer identifies the limiting feature or load, names a candidate mechanism, and states what its manufacturer must confirm. Different products can satisfy the same job. If the only justification is a brand name or a universal star rating, the selection is incomplete.
Addendum 49. Why the two-for-aluminum, four-for-steel rule is only a starting point
Tool designers trade chip space against the core and cutting-edge arrangement. Purpose-built tools and controlled engagement can make other flute counts appropriate. See Harvey Performance: Why Flute Count Matters.
Addendum 50. Check the coating's material recommendation
Harvey Tool's coating guide distinguishes ferrous applications from aluminum applications. Its AlTiN and TiB2 entries illustrate why a coating name must be matched to the work material.
Addendum 51. Several faults can look or sound like chatter
Runout or one damaged edge can make one flute carry more load and leave uneven tooth marks or non-uniform edge wear.
Rubbing or built-up edge can smear material, generate heat and produce a poor finish without a regenerative vibration.
Chip re-cutting can scratch the finish and damage edges when chips remain in the cutting zone.
A forced vibration follows a repeating input from the cutter, interrupted cut, drive or surrounding system. Chatter is the self-reinforcing case; both still require finding the flexible path and the excitation.
After the approved safe stop, inspect the pattern, cutter and setup together. Sound is useful for noticing a change, not for identifying the failed component. Seco Tools’ vibration guidance connects insufficient stiffness and damping to self-excited chatter and identifies surface-finish and cutting-edge-wear consequences.
Addendum 52. How the inspection decides, and what its numbers are
The part is kept as a list of every cut: the space each tool swept through. The top of the metal
at any point is the raw top lowered by every cut that reaches that point. The inspection measures
walls, floors, holes and chamfers at sampled locations on that surface, with numbers rounded
to a hundredth of a millimetre. A display grid plus probes derived from each cut and drawing
boundary look for excess metal and gouges. Passing means the modeled checks pass; it does not
certify every point of a real surface. Removed volume is an estimate from the 0.25 mm grid;
very small cuts can change the analytic surface without a resolved grid volume.
Walls and holes pass within ±0.10 mm, the bore within ±0.05 mm, and floors and
faces within ±0.05 mm.
Tapped holes are accepted when their drilled diameter falls inside the minor-diameter limits for
a 6H internal thread: 4.917 to 5.153 mm for M6×1, and 6.647 to 6.912 mm for
M8×1.25. These are finished minor-diameter limits, used here as a teaching proxy for pilot
quality; they are not universal tap-drill recommendations or a complete 6H inspection.
Addendum 53. How the simulator reads and runs a program
Each line is read in the order a control carries out the parts of a line: feed rate, spindle
speed and tool first, then the tool change, the spindle, coolant, the modes, the move, and last the
program end. So S2500 M3 G1 X20 F100 starts the spindle before it moves. Two codes
from one group on a line, such as G0 G1, are an error.
The interpreter first reads the whole program, and a second machine runs it at once to find its
path, its time and the first place it would stop; that is the path drawn before you press Play.
Pressing Play runs the program again on the block, move by move. The metal, the vise and the tools
are those of the simulator in Milling Operations: every cut is kept as the space the tool swept,
and a move that would break a tool rule or hit the vise stops where it would happen.
An arc’s end point has to be on its circle to within 0.005 mm or 0.1 % of the
radius, whichever is larger. A rapid, or a feed with the spindle stopped, stops a hair short of the
first metal it would touch. The volume removed is an estimate from a 0.25 mm grid.
Addendum 54. When a rough or hard exterior can change the choice
Conventional milling can be useful in a qualified first pass on a rough surface because its
gradual entry can reduce the impact on a brittle cutting edge. Tooling guidance also documents
specialized exceptions such as machining through titanium alpha case before returning to climb
milling. That is not a rule to choose conventional whenever stock has rust or scale: the layer’s
hardness and abrasiveness, the work material, tool grade and geometry, engagement, fixturing and
the approved process all matter. Use the tooling manufacturer’s data and the method approved for
the machine.
Addendum 55. Why some machines climb mill without trouble
A feed drive can be designed to reduce backlash, for example with a preloaded ball nut.
The important question is whether the machine controls that movement, not simply whether it is called CNC.
The chip is the same shape either way; what changes is whether the machine can hold the table
against the pull of the cut.
Modern Machine Shop's column Eliminate Backlash From Machining Centers (Mike Lynch, 2000) defines backlash as “any non-movement that occurs during axis reversals.” This page uses the CourseWiki definition on Screw Threads: clearance that must be taken up after reversing before the load moves. In a real drive, elastic deflection can add lost motion as well; the lab below leaves it out.
Addendum 57. A readout that measures the slide, and when the clearance still matters
The MIT 2.670 milling machine notes (section “Using the Micrometer Dials”) also point out that a digital readout measures the bed position directly, so its reading is not thrown off by backlash. The clearance is still in the drive. A force that pushes the carriage along the screw can move it across the clearance without the handwheel turning; Climb and Conventional Milling explains when a cutting force does that.
Addendum 58. What the model means, and why it is not a prediction for every CNC
The coordinates describe the cutter relative to the workpiece, viewed from above: +X right, +Y up. A moving table travels in the opposite sense. The drive infers position from motor/screw motion, with no compensation or linear-position correction. The first approach seats both axes and calibrates the coordinates at the start, before the plunge. That reference is kept through the cut.
With positive initial seating, each actual coordinate c is updated as c = min(max(c, q), q + B), where q is the motor-inferred coordinate and B the total clearance. Negative seating uses the band [q − B, q]. These are shifted forms of the same play operator as the screw lab above. Each axis holds until a band edge reaches it. If the axes take up clearance at different moments within a command segment, the actual path has an extra bend; the cutter sweep includes it.
Every cut segment removes a disc swept along that actual segment, a capsule with the cutter's 2 mm radius. The area shading compares the actual and zero-backlash histories at the same command step. Span measurements intersect those capsules with the two centre sections. They are not a best-fit diameter, an overall size tolerance or a guarantee of the rest of the contour.
All other effects are idealized away: no runout, deflection, wear, servo lag, cutting-force pull, thermal error or compensation. Friction is assumed to hold a slide inside its clearance. The shapes, clearance settings and viewing speed are illustrative. Restart begins a fresh workpiece; changing a setting recomputes the experiment.
HEIDENHAIN describes direct linear-axis measurement as working “without intervening mechanical elements”. Such feedback changes what the controller can observe; this lab deliberately does not model that correction. Haas's ballbar interpretation guide distinguishes lost motion at reversal from controller-related reversal spikes. Only rigid lost motion is included here.
Addendum 59. Feed per tooth is not always the thickness of the chip
Feed per tooth is how far the work moves between teeth. The chip a tooth actually cuts is only
that thick when the cutter is at least half buried across its diameter. In a narrower cut, each
tooth meets the work at a slant and the chip comes out thinner. Climb and Conventional Milling
shows this: at a width of cut of 10% of the diameter, the thickest part of the chip is 60% of the
feed per tooth.
Some cutting data raises the feed per tooth for narrow cuts to make up for this. Read what your
data assumes before using its numbers, and ask if it does not say. Thinner chips also take more
energy per cubic inch, which is one reason a power estimate is a band.
Sandvik’s specific cutting-force explanation
shows why chip thickness and cutting-edge geometry affect the energy estimate.
Haas’s spindle power curves
illustrate why available power depends on RPM; read the machine’s own curve and duty
rating rather than treating its headline horsepower as available everywhere.
The solid round section has area moment of inertia I = πd⁴/64. With x measured from the clamp, displacement magnitude is y(x) = Fx²(3L − x)/(6EI). At the tip, x = L, so δ = FL³/(3EI). Both displacement and slope are zero at x = 0.
Use N for force, mm for length and diameter, and N/mm² for E. The controls show E in GPa: multiply by 1,000 to get N/mm². The answer is in mm; multiply by 1,000 for µm.
For F = 100 N, L = 40 mm, d = 10 mm and E = 600 GPa, I = 490.874 mm⁴ and δ = 0.007243 mm = 7.243 µm. Halfway along that tool, y(L/2) is 5/16 of the tip value, or 2.264 µm. These are calculated examples, not measured cutter data.
The steady-rest instructions show a rest fixed to the bed, while the follower-rest instructions show the rest attached to the saddle so it travels with the cutter. The drawings here explain those generic roles; they are not setup instructions or copies of manufacturer artwork.
Addendum 63. Rake and chipbreakers depend on the whole cutting setup
Rake describes the slope of the surface the chip flows over. Clearance keeps the tool below the edge from rubbing. Material matters, but hardness alone does not dictate positive, neutral or negative rake: tool material, edge strength, holder geometry and rigidity also matter. Use the tool maker's cutting geometry and range for the application.
A chipbreaker guides a chip into a tighter curl so it can break. It may be molded into an insert or ground into a tool. It works within a range of feeds and depths; a very thin chip can curl without breaking. A chipbreaker is not a guarantee of short chips at every setting.
Addendum 64. How a manual lathe keeps every pass on one helix
The carriage is coupled to the leadscrew by closing the half-nut. On a lathe
with an inch leadscrew cutting an inch thread, the thread dial shows when to close
it so that each pass drops into the same groove. Many machinists also feed the tool in with the
compound travel set near 30° from the radial cross-feed direction, so that most of the cut is taken by one flank of the tool.
The simulator simplifies all of this. It keeps every pass on one helix for you, feeds the tool
straight in, and checks a simplified 60° thread by its crest and root diameters. A real thread is
checked against a thread table with thread gauges or measuring wires.
Engineering Fundamentals/Motion and Rotation/Gear Ratios
Addendum 65. Why tooth counts work, and what the dashed circles mean
The gears here share one tooth size, called module, and a 20° pressure angle. Their pitch diameters are proportional to tooth count. The dashed pitch circles roll together at equal tangential speed, so nA × NA = −nB × NB when clockwise is positive on both shafts. Here n is signed rotational speed and N is tooth count.
Tooth flanks are involute curves. Actual tooth surfaces can slide during contact even though the pitch circles roll. The drawing samples those curves, simplifies the tooth roots and adds a small visual clearance. It is a motion model, not a manufacturing drawing.
The lab assumes rigid external spur gears, fixed shaft centers, matched tooth size and steady transmitted load. It starts lossless; its efficiency control applies the same specified loss to every actual mesh, including both meshes around an idler. Acceleration, elasticity, backlash dynamics, detailed friction, noise and tooth strength are omitted. Changing tooth counts moves the centers to maintain the correct mesh; it does not represent swapping a gear into a fixed gearbox housing.
Engineering Fundamentals/Motion and Rotation/Gear Ratios
Addendum 67. Why adding losses is only an approximation
For a 1000 W input and three 96% stages, the losses are 40 W, 38.4 W and 36.864 W. Each 4% applies to a smaller incoming power. Adding 4% + 4% + 4% incorrectly takes every loss from the original 1000 W.
Engineering Fundamentals/Motion and Rotation/Gear Ratios
Addendum 68. Follow both stages of one loaded train
A 24-tooth driver at 1440 rpm and 4 lbf·ft drives 48 teeth. That shaft also carries a 15-tooth driver for a 45-tooth output. Both stages are 96% efficient, with no other losses. After stage 1: 720 rpm and 4 × 2 × 0.96 = 7.68 lbf·ft. The shared shaft carries that same 720 rpm into stage 2. Final output: 1440/(2 × 3) = 240 rpm and 7.68 × 3 × 0.96 = 22.1184 lbf·ft, versus 24 lbf·ft ideally. Overall efficiency is 0.9216.
Engineering Fundamentals/Motion and Rotation/Gear Ratios
Addendum 69. Worked design: a winch with a speed ceiling and torque floor
A motor supplies 1680 rpm and 24 lbf·ft. The output must be at most 40 rpm and at least 950 lbf·ft. Available gears have 10–60 teeth, with unlimited copies; each reduction stage is at most 6:1 and 97% efficient. There are no idlers or other losses.
Speed requires i ≥ 1680/40 = 42. Two stages give at most 6² = 36, so try three.
Three stages give ηtotal = 0.97³ = 0.912673. Torque requires i ≥ 950/(24 × 0.912673) ≈ 43.3708. Torque controls.
Choose 12 → 48, 15 → 45, and 16 → 60 teeth: i = 4 × 3 × 3.75 = 45. Every tooth count and stage ratio meets its limit.
ωout = 1680/45 ≈ 37.33 rpm ≤ 40 rpm. τout = 24 × 45 × 0.912673 = 985.68684 lbf·ft ≥ 950 lbf·ft. Both pass. Three external meshes make the output opposite to the input.
This is one feasible train, not a claim of an optimum or a complete winch specification. For load force at a drum, continue to Torque; for travel speed, use Rotation and Linear Speed.
Addendum 71. Why cutter selection includes more than width
Reach, neck clearance and tooth arrangement affect access and cutting behavior. See Harvey Performance's keyseat cutter considerations. A nominal tool width does not certify the width of the finished slot.
Machining/Production and Nontraditional Processes/Broaching
Addendum 72. Why the finishing teeth need not keep getting taller
Once the intended profile is reached, further rise would make the feature larger. Repeated finishing teeth can maintain the final size as earlier teeth wear. See Nidec's precision cutting tools reference for broach terminology.
Machining/Production and Nontraditional Processes/Broaching
Addendum 73. Production advantages depend on the job
Tool cost, material, cut length, machine capacity and required finish determine the economics. Broaching is not automatically the cheapest process. Pilot Precision's broach overview includes both short-run sets and one-pass production tools. Cutting Fluids and Application Methods explains lubrication choices.
Addendum 76. Why a damaged or uneven point makes a poor hole
Unequal lip lengths or unequal angles make the two lips remove different amounts of material. The drill can cut oversize or wander. A rounded edge can rub instead of producing a chip. Inspect the point and replace or correctly sharpen a damaged drill before trying to compensate with more force.
Addendum 77. Why “always center drill first” is too broad
The starting recess determines which part of the next drill touches first. An unsuitable angle or an oversized pilot can load the drill's outer corners before its center is supported. Accurate starting also depends on tool rigidity, alignment and the surface being entered. A center-drilling step alone does not guarantee location.
Changes in chip shape, sound, thrust or finish can indicate rubbing, chip packing, a damaged point or unsuitable cutting data. Continuing to force the tool does not diagnose the cause.
Addendum 79. A nominal tool size is not a finished-hole guarantee
Tool condition, alignment, runout, temperature and workholding affect the result. A reamer stamped with the desired diameter still needs an appropriate prepared hole and cutting conditions. Inspect the actual bore rather than accepting it from the tool marking.
Addendum 80. Check the entire joint before substituting a fastener
Diameter, pitch, length and head clearance are only part of the selection. The nut or tapped material, coating, bearing surfaces, installation conditions and intended load also matter. A bolt that can be screwed into the hole has passed an assembly check, not a strength calculation.
Addendum 81. Chart basis and the approximation's limits
The three comparisons were checked against page 1, cutting-tap column, of North American Tool's Tap Drill Sizes — Inch & Metric with Drill Size Formulas, using the supplied four-page copy. Only these selected rows are used here; the chart is not reproduced. The estimate and a selected stock drill need not be identical. The chart's metric percentage formula gives about 77% when the entire pitch is subtracted, which is why “about 75%” must remain approximate.
Addendum 82. Keep percentage and fraction separate
For a metric cutting-tap estimate, the supplied chart gives hole diameter = major diameter − (percentage × pitch / 76.98), entering 75 for 75%. Equivalently, multiplying a reference thread depth by 75% requires the fraction 0.75. Mixing those conventions introduces a factor-of-100 error. This formula estimates geometry, not allowable joint load.
Addendum 83. Why the local setup decides what is permitted
Recognizing a production process does not authorize it on a particular machine. Spindle speed, tapping attachment and power-tapping procedure depend on the particular machine and setup.
A taper goes from 1.100 to 0.950 in over 3.000 in. Its diameter changes 0.150 in, so the rate is 0.050 in/in, or 0.600 in/ft. At half the length, the diameter has changed half as much: 0.075 in. These are teaching dimensions, not a standard machine taper.
Addendum 85. Why the factor of two belongs in the angle formula
Cut the cone through its axis. One side forms a right triangle whose rise is (D − d)/2 and whose run is l. Tangent is rise divided by run. Using the whole diameter change as the rise gives the wrong tool path.
Addendum 86. Why this is an estimate rather than an exact setting
The taper gives tan(α) = TPI/2. If L is the actual distance along the tilted work between support points, the ideal geometric offset is L sin(α). At shallow angles sin(α) and tan(α) are nearly equal. The usual formula substitutes L tan(α); real center contact and setup error add further differences. It is not a method for steep cones.
Addendum 87. Radial depth and compound travel are different quantities
If compound travel is c at angle β from radial, its radial component is h = c cos(β); therefore c = h / cos(β). The corresponding diameter reduction is 2h. Check whether a dial reads actual slide travel or diameter change before using it.
Addendum 88. Why only certain threading-dial positions return to the groove
The threading dial is geared to the leadscrew and carried with the carriage. Its rotation tracks their relative position. For a selected thread, certain dial positions repeat the relationship that lets the tool enter the same helical groove when the half-nuts close. Other positions can put the tool between previous passes, cutting a second groove.
Addendum 89. Verify the formula without a thread-depth table
Split the symmetric 60° tooth down the middle. The right triangle has axial width P/2 and radial height H. Its 30° angle gives tan(30°) = (P/2)/H. Rearranging gives the formula above. If a defined profile removes radial heights a at the crest and b at the root from this triangle, its height is H − a − b. Neither a nor b can be guessed from pitch alone.
For a hypothetical sharp-V, 20-per-inch thread, P = 0.050 in and H = 0.04330 in. At 29° from radial, ideal compound travel would be H/cos(29°) = 0.04951 in. These calculations isolate geometry; they are not settings for cutting a real 20-per-inch fastener.
Addendum 91. A manufacturer's explanation of the marking
Norton: Grinding Wheel Basics explains the five fields and manufacturer-specific variations. A wheel specification and its separately marked maximum operating speed answer different questions.
Addendum 93. Why a wheel can glaze, and where ring testing applies
Glazing depends on the wheel grade and operating conditions as well as the work material. Hard material alone does not diagnose it. Norton's ring-test guidance explains the method's application and limitations; follow the full manufacturer procedure before mounting.
Addendum 94. Carbide needs consistent temperature control
General Carbide: The Designer's Guide to Tungsten Carbide, page 50, explains thermal shock and the effects of interrupted coolant during grinding. Follow the tool manufacturer's cooling guidance for HSS as well; permission to cool a warm bit is not permission to quench an overheated one.
Addendum 96. Why a real grinding plan needs more than the quotient
The calculation assumes a flat starting face and no wheel wear or elastic error. A real plan may reserve material for lighter finish cuts, include dressing and spark-out, and stop to measure. Stock removal from two faces must be assigned to each face; do not silently count it twice.
A motor encoder measures the motor, while a linear scale can measure the slide directly. Screw error, backlash and elastic deflection between the sensor and the cutting point may remain. Homing establishes a reference; repeating homing is not a general cure for lost motion. See Oriental Motor's closed-loop stepper overview for an example of feedback used with steppers.
Addendum 98. Why positive compensation can use a negative offset
Positive compensation describes addition, not the sign of the stored number: adding a negative value shifts the compensated Z position in the negative direction. The Haas mill programming workbook's tool-length compensation section explains G43's addition of the selected offset to commanded positions.
Addendum 99. Check the conversion in both directions
Subtract endpoints: (26−6,4−4) = (20,0); (26−26,19−4) = (0,15); (11−26,19−19) = (−15,0). Add those changes to A: (6+20+0−15,4+0+15+0) = (11,19). The net move is (5,15), not the final absolute position.
A short program can be written directly in a text editor. Conversational programming asks for feature dimensions at the control. Manual data input (MDI) lets the operator enter commands at the control, often for immediate execution. These routes still require correct modes, geometry and setup data; they do not bypass prove-out. Tormach's post-processor guidelines identify control-specific requirements.
Addendum 101. Why an extra axis changes the work a turning center can do
Ordinary turning produces surfaces around the spindle axis. A powered tool plus controlled work angle can reach features at selected angular positions. See the Haas options programming reference for one manufacturer's C-axis implementation.
Addendum 102. Control example: separate geometry and wear fields
The Haas lathe manual, Operation, describes separate X/Z geometry wear, radius and tip-direction fields. Its button sequences and offset conventions must not be assumed on another control.
Addendum 103. Calculate where a speed ceiling starts to matter
With 90 m/min and a ceiling of 1800 RPM, the crossover diameter is 90 × 1000 / (π × 1800) = 15.92 mm. At 10 mm, unrestricted CSS would request 2865 RPM, so the simplified steady-state result is 1800 RPM. Acceleration and other machine limits can reduce actual speed further. See Haas's spindle-speed commands workbook for its G50/G96/G97 implementation.
Addendum 104. Reference for the code convention used here
The G96–G99 entries in the Haas lathe manual describe S in CSS versus RPM and F per minute versus per revolution. Consult the manual for your own control before transferring these meanings.
Addendum 105. Why a good diameter measurement does not prove a taper is correct
The two axis touch-offs locate tangent lines, not every possible contact point on the nose. The normal to an angled surface points in another direction, so its tangency requires another offset. Haas's tool nose compensation discussion describes the resulting undercut or overcut on tapered and curved profiles.
The Haas G41/G42 entry connects the compensation side to the selected tool offsets. The Haas SL-10 Programming Workbook (hosted by the University of Florida), printed pages 59–61, explains the imaginary tool tip, tip direction and approach/depart moves. Use the current manual for the installed control when programming.
Addendum 107. Why no universal cycle template is shown
The Haas G71 reference distinguishes path types and radial cut depth from diametral finish allowance. The Hurco ISNC G71 reference supplies its own two-block definition. A familiar cycle number does not establish compatible parameter meanings.
Addendum 108. Check roughing depth and remaining material
The first pass changes diameter from 24 to 22 mm, a 1 mm radial cut. The second changes 22 to 20.6 mm, a 0.7 mm radial cut. Finishing from 20.6 to 20 removes 0.3 mm radially. Extending the OD cuts beyond Z−25 leaves room for the cutoff blade and places the finishing departure on material that will not belong to the finished part.
Addendum 109. Why the gauge itself needs limits and wear checks
Gauges cannot be manufactured to an infinitely exact size. Their own tolerances and wear allowances are selected under the applicable gauging specification. Repeated use can shrink a plug or enlarge a ring and change its decision.
Addendum 110. Why a small sample range is not the full process spread
The range uses only the two extremes observed in this sample. More readings, different tools, temperature changes and later production may reveal variation not present in these five parts. Standard deviation is another measure of spread that uses every reading's distance from the mean; no standard-deviation calculation is needed for this example.
Indices such as Cp and Cpk compare estimated process variation and its position with specification limits. Their interpretation depends on the data, stability, distribution and application requirements. A single threshold is not a universal acceptance rule, and a five-reading classroom example is not a capability study.
Addendum 112. Cell, system and integration describe different scopes
A flexible manufacturing system may coordinate multiple machines or cells with material handling and scheduling. Computer-integrated manufacturing is the broader idea of connecting manufacturing information and activities. These terms overlap in use, but are not interchangeable names for one machine configuration.
Addendum 113. Why two pieces of the same alloy can behave differently
Permanent deformation involves movement of crystal defects called dislocations. Grain boundaries, dissolved alloying atoms and small particles can obstruct that movement. Cold working and heat treatment change these obstacles. Composition alone therefore cannot tell you the strength of a finished piece.
Addendum 114. Why connecting two different metals can change corrosion
When dissimilar metals are electrically connected in a conducting liquid, such as salty water, galvanic corrosion can accelerate attack on the more active metal. Material pairing, exposed areas and the environment matter. Isolation and drainage can be as important as the alloy choice.
Materials and Manufacturing Processes/Metals/Types of Steel
Addendum 115. Sources and quoted designation meanings
The Metals Service Center Institute identifies the first digit as “family of steel,” the second as “additional elements,” and the third and fourth as “nominal carbon content”; its example is 1018 at 0.18% carbon. Total Materia's SAE-AISI introduction gives the final two digits as “the carbon content in hundredths of a percent” and lists 41XX as “Chromium-molybdenum steels.”
The Aluminum Association's aluminum-alloy overview identifies these series and their principal additions. It describes 1xxx as “aluminum 99 percent or higher purity.” The series names do not list every element in an individual alloy; for example, the Association also identifies magnesium and possible copper and chromium additions in 7xxx alloys.
The Aluminum Association identifies 6061 use in truck and marine frames and 7075 use in aircraft. Metal Supermarkets' 6061/7075 comparison supports the T6 tensile-strength comparison, tank and tooling uses, machinability and welding differences. Its fabrication discussion notes that restoring the weld area's original temper in 6061 may require further heat treatment and aging. Compare the supplier's data for the actual temper and product form.
Addendum 118. Steel microstructures are not the four steel families
Ferrite and austenite are different iron-based crystal phases. Cementite is iron carbide, Fe₃C, a hard, brittle phase. Pearlite is a layered arrangement of ferrite and cementite. Martensite can form from austenite during sufficiently rapid cooling in suitable steel. These terms describe structure, not the buying categories carbon, alloy, stainless and tool steel.
Carbon can diffuse during slower transformations. Quenching can suppress those transformations and produce martensite. The outcome is not determined by cooling speed alone: chemistry and section size affect what forms throughout the part.
Addendum 119. Further reading on conditioning processes
Bodycote's heat-treatment overview provides entry points for annealing, hardening, tempering, stress relief and solution-and-age treatments. AEC's alloy and temper guide is a starting point for aluminum terminology. Neither an overview nor a qualitative sketch replaces a material-specific process specification.
Addendum 120. Why a slower cut can produce more heat
A tooth that advances too little spends more time sliding against the work. The next tooth may also recut trapped chips. Useful chips carry some heat away, so reducing chip thickness without reducing rubbing is not a reliable cooling strategy. ACRYLITE's drilling guidance identifies low feed or excessive RPM as causes of melted chips; its routing guidance treats feed, cutting edges and RPM together.
The dielectric fluid normally insulates the gap. A voltage pulse breaks down that insulation locally and permits a discharge. Between pulses, the fluid helps cool the region and flush debris. Debris removal and gap control help keep discharges stable. Sinker EDM commonly uses dielectric oil; wire EDM commonly uses controlled deionized water, though oil-based wire machines also exist.
Addendum 123. Why one offset does not describe every pass
Discharge energy, flushing and finishing passes change the working gap and surface. Actual compensation comes from the machine's process data and inspection. CHMER's EDM FAQ defines the offset as the wire radius plus the discharge gap, and notes that the gap changes with discharge energy.
Addendum 125. Why hardness alone does not set ECM performance
The reaction removes metal without requiring a harder cutting edge. Metal chemistry and current conditions still affect dissolution, so “all metals machine at the same rate” is not a valid selection rule. Tooling and electrolyte control can make ECM attractive for repeated parts, while adding development effort for a one-off feature.
Addendum 126. Why the material and laser must be matched
Different materials absorb different wavelengths differently. Power and beam delivery also limit the usable thickness and quality. See TRUMPF's laser-cutting overview.
Addendum 129. Why a shaped tool does not give an exact-size copy
Abrasive particles occupy space around the tool and also wear it. The finished cavity therefore depends on the tool, abrasive action and process control, not simply on copying the tool's nominal dimensions. Bullen describes loose-abrasive ultrasonic machining.
