§1When cutting is not enough
Two customers exhaust what the tap, die and single point can give: threads in steel already hardened past cutting, and threads whose accuracy budget is measured in micrometres.
The first customer is metallurgical. The heat-treatment pages established the sequence problem: hardening distorts, so a part finished soft and then hardened is no longer the part that was finished — yet a file-hard surface turns cutting edges into scrap. The second customer is the accuracy chain this whole section quietly depends on: the taps and dies of the tapping page, the GO and NOT-GO gauges and setting masters of the measuring page, the leadscrews whose pitch the lathe copies into every thread it cuts, and the ball-screw and worm hardware of the section ahead. Every one of those is a hardened, precision thread — which means every one is a ground thread, and thread grinding is revealed as the root of the section’s family tree: the process that makes the tools and references by which all the other processes make and judge their threads. Its own reference, closing the regress, is the diamond and the dresser geometry of §2, which is where absolute thread form actually enters the world.
Contents§2The wheel is the tool
A thread grinder’s wheel rim carries the thread profile in negative — formed into the abrasive by a diamond dresser, and renewable to fresh perfection whenever it dulls or wears.
That dressability is the process’s quiet superpower: a cutting tool’s form degrades from the moment it first cuts, but a grinding wheel’s form is re-created on demand — the diamond traverses the rim under precise control and leaves a new, exact vee (or rib set) in seconds, so the tool’s geometry is a machine setting rather than a wearing asset. Two working styles use it. The single-rib wheel carries one thread profile and traverses the work exactly as the lathe’s single point did — one pitch of axial travel per work revolution, pass after pass — the flexible style for long leadscrews and one-offs of any pitch the dresser can form. The multi-rib wheel carries a row of profiles, each rib ground a share deeper than its neighbour like the tap’s chamfer translated into abrasive: plunged to depth, it finishes a thread in barely more than one work revolution plus run-out — the production style, and the standard way short threads are ground from solid in fully hardened blanks, no prior cutting at all. Either way the wheel is doing what §1 promised: carrying form into material that would destroy any edged tool asked to do the same.
Contents§3The geometry
One angle governs the setup: the wheel must lean over by the thread’s own lead angle, or its flat rib will foul the helical groove it is trying to follow.
The wheel is a disc; the groove is a helix. Run the disc square to the work’s axis and its sides interfere with the climbing flanks — the same clash the helical-milling and gear pages met wherever a straight tool follows a helix — so the wheel head is swivelled through the lead angle λ, laying the rib’s plane along the groove’s local direction. The systems page’s numbers set the dial: about 3.03° for the house M10×1.5, 4.05° for the Tr30×6 — small angles, but at micrometre stakes an unswivelled wheel measurably fattens the groove and thins the thread, and multi-start work with its doubled lead angles doubles the care. Two more geometric facts shape practice. Wheel diameter is a trade: large wheels hold form longer and run truer, but internal thread grinding forces small wheels into small bores at furious spindle speeds to keep §4’s surface speed — which is why ground internal threads cost what they cost. And the interference logic sets a floor under fine pitches from large wheels: the wheel’s curvature must be able to live inside the groove’s helix, a constraint the setup tables encode and the operator inherits. The angle itself, though, is the whole secret: set λ, and the disc becomes an honest helix-follower.
Contents§4Numbers of the trade
Grinding’s energy budget is extreme and local — a rim doing tens of metres per second against work doing centimetres — and every practice at the machine exists to manage the heat that budget implies.
That 33.5 m/s — motorway pace at the rim — meets work turning at a crawl of a few tens of revolutions per minute, so essentially all the relative speed, and all the specific energy of abrasion, arrives as heat in a contact patch fractions of a millimetre wide. Unmanaged, the consequences carry names every grinding hand knows: burn, the straw-to-blue temper colours of a surface that has locally un-hardened itself, and grinding cracks in a root that was supposed to be the fatigue-safe feature of the part. The management is threefold and non-negotiable: flood coolant delivered into the arc of contact, not near it; conservative infeeds with the multi-rib’s work shared across its chamfered ribs exactly as the tap shared its cut; and spark-out — final revolutions at zero added infeed, the grinding twin of the lathe’s spring passes, letting machine deflection relax and the wheel cut its way to true size and a burnished finish. The reward for the discipline is the point of the whole page: hardened threads whose pitch, form and finish sit at the accuracy summit of everything this section has described.
Contents§5After hardening
Grinding’s place in the process chain is fixed by one sequencing rule: it is the operation that comes after the furnace, so the furnace’s distortion dies with the finishing cut.
The canonical route for a precision hardened thread reads: turn or cut the thread a grinding allowance oversize in the soft state — or leave the blank plain for from-solid work — then harden and temper, then grind. Everything the heat treatment bent, grew or shifted is inside the allowance, and the ground surface that emerges is both the final geometry and the final metallurgy, untouched by any later process that could disturb either. Down that route travel the parts §1 listed: leadscrews and ball-screw tracks whose lead accuracy every machine axis inherits; taps, whose ground thread is the master their every tapped hole copies; thread gauges and setting plugs, ground to grades that make the measuring page’s arithmetic mean something; worm shafts and precision studs. The regress the measuring page raised — who measures the measurers? — genuinely terminates here: master threads are ground on machines whose own leadscrews were ground, under laboratory measurement, against the diamond-dressed form of §2, and calibration chains the world over bottom out on such artefacts. It is a satisfying structural fact: the humblest tapped hole in the section stands, three or four removes back, on a ground thread.
Contents§6Cut, rolled, ground
Three ways to make a thread, three different bargains — and the section’s production story compresses into one comparison.
| Process | Accuracy & finish | Root & fatigue | Hardened work | Economics |
|---|---|---|---|---|
| Cut (tap · die · point) | good; tool-mark finish | sheared root, tool marks | no — soft state only | flexible, any pitch today; slow |
| Rolled | good; die-burnished | best — compressed, grain-flowed | limited by die life | fastest and cheapest at volume |
| Ground | best — micrometre class | excellent radius; burn risk if abused | yes — its whole point | slowest, dearest; unmatched where it matters |
| Read as a supply chain rather than a contest: rolling makes the world’s fasteners, cutting makes everything one-off, odd or in-place, and grinding makes the hardened, micrometre-true minority — including the taps, dies, gauges and leadscrews the other two columns depend on. A fatigue-critical hardened fastener takes the best of two columns at once: rolled after heat treatment, on blanks and dies that grinding made possible. | ||||
§7Quick reference
The working core of the page on one card rack.
Tool
diamond-dressed negative
form renewed on demand
Styles
single-rib traverse
multi-rib plunge, from solid
Setup
swivel wheel by λ
3.03° / 4.05° examples
Energy
rim 33.5 m/s vs slow work
flood coolant · spark-out
Place
after the furnace, always
makes the tools and masters
