CoatingIQ
← Course index

Wrap-up

86%

Lesson 11·Failure modes

The big four failure modes and how the build prevents them

Thermal fatigue, mechanical fatigue, plastic deformation, and abrasive or adhesive wear: how each shows up on a pulled die, what mechanism drives it, and which build decisions push the failure point out or shift the mode.

11 min readLesson 11 of 13

Tying it together

Shifting the failure mode

Sometimes the right engineering answer is not to eliminate the failure. It is to shift the die from one failure mode to a better one. Mechanical-fatigue corner cracking is the worst common mode because it is catastrophic and unpredictable. The die runs nominal for 30,000 cycles, the crack propagates silently through the bulk, and the die fails through-thickness on cycle 32,000 with no warning. Thermal fatigue is plannable. The heat-check network grows on a measurable schedule, the operator can see it on every shift inspection, and the die can be pulled, recut, or retired at a chosen point.

A connecting-rod forge die at one production shop ran on standard H13 at 48 HRC with R 1.5 mm at the base of a critical rib. Average life was 28,000 cycles, with high scatter (a few dies at 60,000, several below 20,000), and the failure mode was corner cracking through the rib base into the bulk. The shop redesigned to premium H13 at the same hardness, opened the radius to R 3 mm, and specified a thinner compound layer with controlled gamma prime phase. Average life moved to 82,000 cycles with much tighter scatter, and the failure mode shifted to thermal-fatigue craze across the working surface. The new mode allowed scheduled replacement at 75,000 cycles instead of unpredictable failure in service. Press downtime dropped, and the forged parts came out more consistent because the die ran in a stable geometric state for most of its life. The shift was the win, not the absolute life number.

This is the framing the build is for. Pick the failure mode the operation can manage, then choose the steel, the heat treat, the radii, the nitride recipe, and the polish that put the die in that mode.

Pushback questions for the operation

  1. Of the dies pulled in the last six months, what fraction failed in each of the four modes, and which mode dominates?
  2. Is the dominant mode the one the operation can plan around, or is it the unpredictable one that catches us mid-production run?
  3. If we shifted to the next steel grade up (premium H13, Dievar, W360 at appropriate hardness) and opened critical radii by 1 to 2 mm, which mode would the build move to and what would average life become?
  4. Are we accepting a worse mode because the build looks simpler on the spreadsheet, when the total impact of die replacement and unplanned downtime would favor shifting the mode?

Common confusions

The four modes are not mutually exclusive. A late-life die shows all four on one cavity, because once any mode is well underway the others accelerate. The diagnostic question is which mode initiated the failure, not which modes are visible at the end. Look at the earliest documented inspection record, not the pulled die on the bench.

Heat checking is not always thermal fatigue alone. A coarse heat-check pattern with deep cracks linked into a network and chunks spalling out is thermal fatigue past its useful life. A fine even network in a die that still runs is thermal fatigue in its working phase. Treat them as different operational states of the same mode.

Cavity wash is not always plastic deformation. Localized washout at a gate or flash land is abrasive or adhesive wear of the surface. Bulk cavity profile change across a working face is plastic deformation of the substrate. The distinction matters because the build decisions that fix wear are different from the build decisions that fix yield. Wear wants nitride and polish. Yield wants steel grade and case depth.

"Hydrogen embrittlement" surfaces in vendor explanations of mechanical-fatigue failures. Course 2 Lesson 10 covers why this is almost always wrong on a 48 to 50 HRC forge die. The actual cause is a stress concentration the geometry did not radius and the cycle count caught.

Up next: common build failures.

Sources