language
English
中文
русский
Deutsch
português
français
الكتابة العربية
2026/09/01
How to Trace Fastener Defects Back to Tooling Root Causes in High-Speed Cold Heading
A hex flange die expected to run 2,000,000 pieces cracked at 600,000. The path from symptom to redesign that took it past the original target.
In this case study
Case at a glance
Part: M16 hex flange bolt, carbon-steel wire (boron steel, ~750 MPa tensile)
Tool: carbide die insert in a die holder, cold formed at multi-station
Target life: 2,000,000 pieces
Actual: cracked at ~600,000 pieces — the line stopped and no one could agree why
Root causes: a sharp internal radius (stress concentration) compounding with excess shrink-fit interference
Fix: opened the radius, adjusted the fit, stepped up to a tougher grade
Result: passed 2,000,000 pieces and kept running
This is a representative working example built from typical production conditions, to walk through the failure-analysis method in concrete terms. The process — and the numbers of the kind you would see on a real floor — are the point, not any single named customer.
The die set had been quoted and built for a long run: an M16 hex flange bolt in boron steel, expected to deliver 2,000,000 pieces between meaningful servicing. The carbide insert was the obvious choice for that volume — a high-wear forming cavity that a tool-steel die would not have carried.
Somewhere around 600,000 pieces — not at the start, and not near the end — the line operator heard the change in sound, felt the shift in press load, and pulled the die to find a crack running out of the flange-forming corner toward the outer wall of the insert. The line stopped. The maintenance team, the wire supplier and the machine crew each had a theory, and none of them agreed.
Premature cracking like this is one of the most common — and most frustrating — cold forming failures, because it is not a simple wear problem. Wear you can see coming. A crack that ends a die a third of the way through its expected life is a design or process problem, and reordering the identical die will reproduce it at the identical piece count.
The first instinct after a premature crack is to replace the die and move on. That instinct costs money, because it does nothing about why it cracked. If the cause is in the geometry, the fit, or the material selection, a new identical insert will fail in the same place at the same count — and you will have paid for the replacement plus another line stop to learn the lesson again.
The rule that applies here: replacing a failed tool only works if the root cause is fixed first. Otherwise you are paying to reproduce a known failure.
The disciplined path is a structured failure analysis. It does not need to be expensive or slow — it just needs to be methodical.
Failure analysis starts with where the crack is and how it looks, because that immediately narrows the possible causes:
| Crack location | What it usually points to |
|---|---|
| Starting at a sharp internal corner | Stress concentration from geometry |
| Running from the OD or along the bore wall | Interference / shrink-fit problem |
| Starting subsurface, propagating outward | Fatigue from cyclic load, or material/heat-treatment defect |
| Multiple cracks in the same zone | Load path concentrating on one area, or misalignment |
In this case the crack originated at the internal corner where the flange face meets the body-forming profile — the classic geometry-stress location — and ran toward the outer wall. That single observation told us to look hard at two things: the radius at that corner, and how much tensile stress the insert was carrying from its fit in the holder.
Where the crack starts is a map to the cause: this one started at the geometry-stress corner and ran toward the bore wall, pointing straight at two compounding causes.
Next, look at the part geometry against the direction of forming load. A carbide insert is strong in compression — that is why it is used for the forming cavity — but it is weak in tension. Any feature that turns part of the load into tension across the carbide is a candidate root cause.
The review found the first problem: the internal radius at the flange-forming corner was effectively sharp — about 0.2 mm, drawn that way to hold the part's corner definition. A sharp internal radius in a loaded cavity is a stress riser: it concentrates the cyclic forming load into a tiny zone. In a carbide insert already under hoop tension from its shrink fit, that concentrated zone sits right at the boundary between compressive and tensile stress — the exact place a crack wants to start.
Two more inputs needed checking: the carbide grade and the shrink-fit interference.
Material grade. The insert was a standard die-grade carbide. For a flange-forming cavity under high load with a sharp-cornered profile, a tougher, higher-cobalt grade can absorb the impact and local stress better than a harder, more brittle one. The question was not "is it carbide?" — it was whether the specific grade matched the specific load and geometry.
Shrink-fit interference. The carbide insert was retained in a steel holder by a shrink fit, as is standard for segmented and inserted die construction. The interference is what holds the insert, but it also puts the insert under permanent hoop tension. The measured interference was at the upper end of the working range. Under a sharp-cornered, high-load profile, that extra tension is the difference between a crack that never starts and one that propagates.
This is the compounding failure: neither the radius nor the fit alone would likely have killed the die so early. The sharp radius concentrated the load, and the high interference added tensile stress at exactly that spot. Together they crossed the carbide's fracture threshold at the worst point in the cycle.
Because the analysis had isolated two compounding causes, the fix was targeted rather than a blind "make it stronger":
None of the changes was dramatic or costly. Each was a deliberate, small correction aimed at the specific mechanism the analysis had identified.
The redesigned insert was manufactured — machined on wire EDM with 0.05 µm positioning scales and a five-axis center holding ±0.002 mm (calibrated per ISO 230-2) — and its dimensional inspection report confirmed the new radius and fit before it shipped. Then the real test: back on the line at full speed.
| Metric | Original die | Redesigned die |
|---|---|---|
| Target life | 2,000,000 pcs | 2,000,000 pcs |
| Life reached | ~600,000 pcs (cracked) | Passed 2,000,000 and still running |
| Critical radius at flange corner | ~0.2 mm (sharp) | ~0.8 mm (opened) |
| Shrink-fit interference | Upper range | Mid-range |
| Carbide grade | Standard die grade | Tougher high-cobalt grade |
| Line stops from this failure | 1 (emergency, at 600k) | 0 through 2,000,000 |
The cost of the analysis and redesign was a fraction of the value of the extra 1.4 million pieces of uninterrupted production — and it removed the recurring emergency-stop risk that the "replace with the same die" approach would have guaranteed.
Three lessons generalise to almost every premature cracking situation:
Hongli runs this structured failure-analysis service for fastener manufacturers. When a tool fails early, send us the failed sample or photos, the part drawing, and your production data — we examine the failure, review the design and fit, and return a concrete redesign with the corrected tooling built and validated.
A die that cracks at 600,000 pieces when it was quoted for 2,000,000 is not a wear problem and is not fixed by reordering the same die. It is a signal pointing at a specific mechanism — in this case a sharp internal radius compounding with excess shrink-fit tension in a carbide insert. Read the symptom, question the geometry against load direction, check the grade and the fit, make the small targeted changes, and validate on the line. Done methodically, the fix pays for itself in the first uninterrupted million pieces.
In the analyzed case, two factors compounded: a sharp internal radius in the flange-forming zone created a stress concentration, and the carbide insert ran under shrink-fit interference that added tensile stress right where that radius focused load. Neither alone would have failed so early; together they pushed the carbide past its fracture threshold at the worst point of the cycle.
Crack location is the first diagnostic clue. A crack starting at a sharp internal corner points to stress concentration from geometry. One at the OD or along a bore wall suggests interference or fit problems. One that starts subsurface and propagates suggests fatigue from cyclic loading or a material/heat-treatment defect. Reading where it starts — and its orientation — narrows the causes before you change anything.
Only if the root cause is fixed first. Reordering an identical die reproduces the identical failure at the same piece count because the geometry and fit that caused the crack are unchanged. The fix requires a design change — opening the critical radius, adjusting shrink-fit interference, changing the carbide grade, or a combination — then validating the redesigned tool on the line.
Even a small opening matters. In the studied case, increasing the critical radius from ~0.2 mm to ~0.8 mm, combined with a small reduction in interference and a tougher carbide grade, moved the die from 600,000 to well past 2,000,000 pieces. The exact change depends on the geometry and load, which is why structured analysis beats a blind 'bigger radius' fix.
Yes. Hongli runs a structured failure-analysis service: we examine the returned tool or failed samples, read the crack location and surface, review the design against load direction, and propose a concrete redesign — geometry, carbide grade or fit changes — then build and validate the corrected tooling. Send us the failed sample, the part drawing and your production data to start the analysis.
