CoatingIQ
← Course index

Wrap-up

86%

Lesson 04·Running Forge Dies

Thermal cycling and heat checking: where the cracks come from and how to slow them

Why every forge die heat-checks, which of the three contributors a shop can actually move, and how to tell a healthy crazing pattern at 40K hits from a runaway network at 25K.

9 min readLesson 4 of 13

Tying it together

What this means on the shop floor

For closed-die forge, the high-heat regions are the impression corners and the gutter just outboard of the parting line. The crazing network appears there first and stays denser there throughout the die's life. The early-warning shape is a crack leaving that network and running into the impression face. Photograph the same regions on the same die every inspection and the trend is visible.

For upsetters and hot-headers, the punch face crazes before anything else because it sees contact on every cycle without much radiative relief between hits. The punch is also where parallel-line patterns appear, because the mechanical load is more uniaxial than on a closed-die. A parallel-line shape on a punch is the call to dye-pen earlier than the calendar would suggest.

For ring-roll dies, the work-roll face and the mandrel craze on different schedules because they see different ΔT cycles. The mandrel sustains contact and runs hotter; the work-roll face cycles faster but cooler. Treat them as two different patterns on the same die, not as one pattern.

Across all of them, the shift-to-shift photograph of a defined region at a defined inspection point is the most useful diagnostic the shop will keep. The trend tells the story the absolute pattern cannot.

Internal pushback questions

  1. When the same job runs on two sister dies and one heat-checks twice as fast as the other, who in the shop is responsible for finding the process difference, and what is the first variable they check?
  2. Is the inspection log capturing whether the heat-check network is uniform or whether cracks are leaving it, or is it only capturing "heat check present" as a yes-or-no field?
  3. When a die is polished as part of routine maintenance, is the polish removing only surface oxide or is it grinding into a heat-check pattern in a way that hides the network without removing the cracks underneath?
  4. If a die at 30K hits has a denser network than a sister die at 60K hits on the same job, what process variable is being investigated this week, and who is doing the investigation?

Common confusions

Heat checking is not a single failure mode. The network itself is normal at maturity. The propagating cracks that leave the network are the failure mode. Inspecting for "heat check" without distinguishing the two is inspecting for the wrong thing.

A polish does not slow heat checking. A polish removes the surface signature and can hide propagating cracks under a smooth finish. The right time to polish a die is when the network is the only pattern present and there is no propagation. The wrong time is when polishing is the response to finding cracks that are leaving the network.

A thin compound layer is not always the right answer for heat-check resistance. A compound-layer-free recipe (Course #2 Lesson 5) gives the diffusion zone the full thermal cycle without an ε-rich brittle skin on top, but the surface microhardness is lower and adhesive wear can dominate. A thin γ'-dominant compound layer balances the two for impact-loaded forge dies. Course #2 Lesson 3 covered the Floe cycle that produces it.

Cycle frequency is not the first lever to pull. Throughput costs more than the heat-check rate it buys in most jobs. The realistic shop-floor lever is ΔT per cycle, through preheat discipline and thermal management across the shift.

Up next: lubrication fundamentals.

Sources