§1Forming, not cutting
Press a hardened die carrying the thread’s negative against a rotating blank and the profile is extruded into the surface: nothing removed, everything relocated.
The tapping page cut the groove away and threw it in the chip bin; rolling keeps it. As the die ridges sink in, the displaced metal has exactly one place to go — up, flowing plastically into the die’s grooves to form the crests — so the finished thread stands partly below and partly above the original blank surface, and the blank itself must be sized between minor and major diameter for the volumes to balance (§2 proves it lands precisely on d2). The process is cold working end to end: the same mechanism as the thread-forming screws and forming taps met earlier in the section, scaled to production and turned outward onto the bar. Three consequences define everything that follows. The metal is strain-hardened and grain-flowed along the profile rather than sheared across it (§4’s fatigue story). The economics invert — no swarf, no cutting time, a cycle measured in fractions of a second (§3). And the process inherits cold forming’s constraints: it needs ductility, die-quality tooling and an accurate blank, because every error in the blank is conserved, not cut away (§5).
Contents§2Why the blank is the pitch diameter
The blank size is not a table value to memorise — it is the definition of d2 doing manufacturing work: at the pitch line, ridge and groove are equal, so what sinks exactly fills what rises.
| Quantity | Reasoning | Value |
|---|---|---|
| Blank diameter | ridge = groove at d2 → volumes balance | d2 = 9.026 ≈ 9.03 mm |
| Crest rises above blank | (d − d2)/2 | 0.487 mm |
| Root sinks below blank | (d2 − d3)/2 | 0.433 mm |
| Stock saving vs Ø10 bar | 1 − (9.026/10)² | 18.5% |
| The first row is the calculating page’s definition of pitch diameter — the cylinder where thread and space are equal — read as a conservation law: metal displaced from the grooves below that cylinder is exactly the metal needed for the ridges above it, so a blank turned (or drawn) to d2 rolls to a full thread with nothing left over and nothing missing. The last row is the purchasing department’s version of the same fact: rolled threads start from wire already at 9.03 mm, and the 18.5% of bar a cut M10 turns into swarf is simply never bought. Multiply that across the world’s fastener tonnage and §2 is arguably the most economically consequential arithmetic in this whole section. | ||
§3The machines
Four machine layouts carry the same die principle from the fastener plant to the toolroom, differing in how the negative form is presented to the blank.
The hero’s flat-die machine is the fastener classic: two rectangular dies faced with inclined thread ridges, one fixed, one reciprocating, with the blank fed between — one stroke rolls the blank along, sinking the form progressively, and drops a finished screw off the end in a fraction of a second. Header upstream, roller here: it is the line that makes the machine screws page’s subject by the billion. Cylindrical (two- or three-roll) machines mount the negative on round dies that turn with the work while feeding radially inward — slower per piece, but gentle, precise and adjustable, the layout of precision and large-diameter work, and of the toolroom attachment that rolls a thread on a lathe. Through-feed versions skew the rolls slightly so the work screws itself axially past them, rolling threaded rod and long leadscrew blanks continuously by the metre. And the planetary machine fixes one die as a curved segment while a central round die spins blanks around it — continuous feed with no reciprocating stroke, the highest-rate layout of all. One negative form, four presentations; the choice is pure production engineering, and the metallurgy of §4 comes out the same from each.
Contents§4What rolling buys
Static strength barely changes — the stress area is the stress area — but at the root, where fatigue lives, the rolled thread is a different and better object than the cut one.
Three effects stack, all at the one place the fatigue pages identified as a threaded part’s Achilles heel. The rolled root is burnished, not sheared: die-pressed to a mirror finish with its radius formed smoothly, where a cut root carries tool marks — each a micro-notch — across the very radius doing the stress-concentrating. The cold work leaves compressive residual stress locked into the root surface, a pre-load the service tension must first unwind before the surface ever feels tensile — the same mechanism shot peening buys, delivered free by the process. And the grain flow follows the profile, fibres sweeping continuously around the root like the timber page’s unbroken grain, instead of being truncated across by a cutting edge. Together they raise fatigue performance decisively — most decisively when the sequence is right: roll after heat treatment, so hardening cannot relax the residual compression, which is exactly the rule aerospace writes into its rolled-radius-root UNJ fastener practice. Add the process gifts — burnished flanks that lower thread friction scatter (a nod back to the torque-and-tension page), and work-hardened surfaces that wear better in service — and rolling stops being merely the cheap method. For a fatigue-loaded fastener it is the correct one.
Contents§5What rolling demands
A process that relocates metal instead of removing it inherits three strict conditions: the metal must flow, the blank must be right, and the form must be rollable.
Ductility first. The blank material must sustain severe cold flow without cracking — the everyday steels, stainless grades, brass and aluminium of fastener work all qualify — while very hard or brittle stock tears at the root instead of flowing, which sets a practical hardness ceiling on what the dies can form (and prices the dies themselves, which must stay far harder than anything they roll). Blank discipline second, and it is the sharp one: cutting corrects a slightly wrong bar by removing the error, but rolling conserves it — §2’s volume balance means every micrometre of blank diameter error reappears directly as pitch-diameter error, so the blank is turned, ground or drawn to a tolerance the finished 6g band can absorb. Blank preparation, not die setting, is where rolled-thread accuracy is actually made. Form third: shallow, rounded, generously radiused profiles — the whole 60° fastening family, knuckle forms, the rounded closure threads — flow beautifully; deep, sharp or square-rooted forms demand flows the metal resists, which is one more quiet reason the standard forms look the way they do. Finally the detailing: rolling wants run-out room, so threads-to-a-shoulder and other tight-corner features stay easier for the single-point tool — the small territory §6 leaves to cutting.
Contents§6The rolled world
Rolling owns commercial threading so completely that the cut fastener is the exception — and a rolled thread announces itself if you know where to look.
Essentially every machine screw, bolt, self-tapper and threaded rod in commerce is rolled; the Edison cap of the other-threads page is rolled in sheet; forming taps roll internally; even precision work increasingly finish-rolls for §4’s fatigue case. The tells are legible on any screw from the bench. The plain shank of a simple rolled screw runs at blank diameter — visibly slimmer than the thread crests, which stand proud of it by §2’s 0.487 mm-scale margin, where a cut thread can never overhang the bar it was carved from. The crests may show the faint seam where the two risen flows met, the flanks carry a die-burnished sheen quite unlike tool-mark lay, and there is no centre-drill or chuck evidence of a turning operation. Cutting keeps what rolling cannot reach: one-offs and repairs at the bench, hard and brittle materials, threads dead up to shoulders, interrupted and internal large forms, and the toolroom’s any-pitch-today flexibility — which is precisely the division of labour this half of the section describes, with grinding (next page) arriving where neither cutting’s reach nor rolling’s economy can deliver the accuracy.
Contents§7Quick reference
The working core of the page on one card rack.
Principle
displace, never remove
groove metal becomes crest
Blank
Ø = d2 = 9.03 mm on M10
crest +0.487 · root −0.433
Economics
18.5% stock saved
a thread per die stroke
Fatigue
burnished root · compression
grain flow — roll after HT
Demands
ductile stock · exact blank
rounded, shallow forms