SCREWKING
Head forming limits: why small heads are harder
The short version: there is a practical ceiling on how much material one blow can upset. Common guidance for steel is an upset ratio around 2.25–2.3 in one blow and roughly 4.5 in two — guidance, not a standard requirement, and it moves with die support, free length, material condition, friction and preform geometry. It is also a macro-scale figure. Microforming work reports the achievable ratio falling to around 1.6. Small parts are not the same problem scaled down; the limit itself tightens.
What upset ratio means
Cold heading forms a head by shortening and thickening the length of wire protruding from the die. That protruding length divided by the wire diameter is the upset ratio.
upset ratio = length being upset ÷ original wire diameter
Exceed the workable range and one of two things happens. Too much free length and it buckles — once it leans, the head forms off-centre. Or the material is displaced too abruptly and folds or cracks at the head-to-shank transition.
So larger heads need more stages. A head that is much wider than its shank usually needs a preform before the finish blow — though how many stages depends on head height, profile, recess volume and support, not on the diameter ratio alone. This is why changing a head style can mean one new die, or a rebuilt sequence.
Why the limit tightens at small sizes
Intuitively, forming is geometry and should scale. It does not. Microforming calls this the size effect: as a part approaches the material’s grain size, the material stops behaving like a continuum.
- Fewer grains in the section. A macro cross-section contains thousands and averages out. A miniature one may contain dozens, so the orientation and strength of individual grains start to show in the result.
- A larger share of surface grains. Grains at a free surface are less constrained than interior ones and flow differently. The smaller the part, the greater their proportion.
- The consequence is scatter. Same wire, same die, less consistent outcome than at macro scale.
So treating the macro 2.3 as usable at M1.4 is risky. Microforming studies report achievable values falling to around 1.6. But 1.6 is not a universal ceiling for M1.4 either — it is one study’s result under its own material and conditions. The direction is the usable finding: less margin, more scatter. What your part can actually do has to be established on your wire and geometry.
Simulation compares designs. It does not promise one.
The Cockroft-Latham ductile fracture criterion is commonly used with finite element simulation to anticipate cracking before a die is cut. One study on an M2.6×5 TORX screw in SUS304 reports that reshaping the protruding top of the first-stage preform punch lowered the damage value at the crack location from 0.594 to 0.392, about 34%.
The authors are explicit that these values serve relative comparison only. No experimental critical value was established, so the number cannot decide whether a part will crack.
At micro scale there is a further problem: classical ductile fracture criteria do not account for grain size or surface-grain effects, so extrapolating them to miniature parts without calibration is unreliable.
In practice: simulation is good for deciding whether design A beats design B. It is not a guarantee that either will not crack. That still comes from samples.
Tooling is not consumed by one order
The usual worry about custom work is the tooling charge. The number that matters is not what a die costs but how many parts it makes.
| Tool | Life |
|---|---|
| Flat dies (thread rolling) | average over 2,000,000 pieces |
| Standard die, straight shank | around 200,000 pieces |
| Carbide punch | 1,500,000 – 3,000,000 strokes |
| High-speed steel punch, same part | around 150,000 pieces |
On those figures, ten thousand pieces uses roughly 0.3% to 6.7% of a tool’s life.
So tooling is generally not consumed by a single order. Whether a repeat order is charged again, however, is a contract term rather than a technical consequence — ownership, storage, repair and replacement are handled differently by different suppliers. The question to ask is not what a die costs, but whether it is charged once or amortised, whether a repeat is charged again, how long it is retained, and who carries a failure.
⚠️ The table above is general magnitude, not a universal figure for every tool, material and geometry; the carbide and HSS values come from a single M6 DIN 912 / SCM435 case. Treat a supplier’s measured data as authoritative.
What to ask
- How many stages does this head need at this size? Stage count drives both tooling and lead time.
- What upset ratio do you work to — a macro figure, or one adjusted for small sizes?
- Do you simulate forming? If so, is it used to compare options or offered as assurance?
- Carbide or high-speed steel punches? A ten- to twenty-fold difference in life implies a different cost structure.
- Are dies changed on schedule or on failure, and are strokes tracked? Wear degrades gradually and can be planned around; fatigue fracture stops a line.
References
- Padfield & Bhupatiraju, “Cold Heading,” ASM Handbook, Vol. 14A, 2005
- Talangkun, “Design of the semi-closed die for shaping thick coin-like carbon steel parts in a single operation,” SN Applied Sciences 5, 176 (2023)
- Seo et al., “Design of Cold Heading Process of a Screw for Storage Parts,” Transactions of Materials Processing 20(1), 2011
- Ran, Fu, Chan & Kwan, “The influence of size effect on the ductile fracture in micro-scaled plastic deformation,” International Journal of Plasticity 41 (2013)
- Nickel Institute — Cold Forming Stainless Steel Bar and Wire
Some sources are paywalled. Values quoted are each study’s result under its own conditions and are not general design values; confirm process parameters through samples and with your supplier.
Enquiries
Changed a head style, with the size between M1.0 and M5.0, and want to know whether that is a die adjustment or a new tool? Send the drawing or a sample and we will come back on stage count and what it means for lead time.