CoatingIQ
← Course index

Wrap-up

88%

Lesson 10·Running Forge Dies

Maintenance options: polish, repolish, weld repair, re-nitride, re-coat, scrap

Match a wear mode to a maintenance action. Know when polish buys real life and when it hides a crack, when TIG weld is the right call versus a coin flip versus never, what re-nitride and re-coat actually recover, and when scrap is the right answer.

7 min readLesson 10 of 13

Tying it together

When never to weld

Four conditions where the answer is no, regardless of how the rest of the die looks.

High-alloy powder-metallurgy grades. PM tool steels (CPM-3V, CPM-10V, ASP-class) carry carbide volume fractions and alloy chemistries that do not behave like H13 under a weld cycle. The carbide network around the weld pool dissolves and re-precipitates unpredictably, the HAZ is brittle in ways that no shop-floor PWHT recipe reliably fixes, and the documented weld-repair literature on these grades is sparse. If the die is in CPM-9V or comparable, weld repair is research-grade work, not production maintenance. Replace the die.

Fully nitrided cavities without strip capability. A weld laid into a nitrided surface fuses into a chromium-and-nitrogen-saturated zone that produces porosity, hot cracks, and a heat-affected zone full of brittle nitrides. The only legitimate path is strip the entire nitrided case off the repair region first, weld onto clean H13 substrate with the right preheat and interpass, PWHT the whole die, then re-nitride the repair area to match the surrounding cavity. If the shop cannot do all four steps in sequence, do not start.

Multiple prior weld repairs on the same die. A die with one prior weld has a single HAZ to manage and a known metallurgy. A die with two prior welds has two HAZs intersecting somewhere in the structure and a substrate that has been through three PWHT cycles. A die with three is a metallurgical specimen, not a production die. The third weld is where the cumulative residual stress, the cumulative tempering, and the cumulative repair-zone fatigue add up to a failure that the maintenance crew cannot predict from inspection. Scrap before the third weld, not after.

Geometry that blocks uniform PWHT. Large dies with deep cavities, thin sections next to heavy bosses, or geometry that produces sharp thermal gradients in the PWHT furnace cannot be reliably stress-relieved. The PWHT soak hits the surface temperature on the readable thermocouple while parts of the die never reach setpoint. The result is a die that is stress-relieved in some regions and as-welded in others, with an unmapped boundary between the two. Either build PWHT fixturing that guarantees uniform heating, or do not weld this die.

What this means on the shop floor

For closed-die forge in H13, the typical maintenance ladder is polish at break-in scuffs, repolish at the first stable heat-check pattern, re-nitride somewhere between the first and second life cycle when surface hardness drops below process minimum, possibly add re-coat on the re-nitrided case if the application has documented high-flow wear, and scrap at the third repair decision or the first propagating crack the substrate cannot recover from.

For upsetters and hot-headers, the punch faces over-temper and wear through the case first. Re-nitride is the workhorse repair. Weld repair on punch geometry is rarely worth the turnaround because the punch is usually small enough that replacement is faster than repair.

For ring-roll work-rolls and mandrels, weld repair is more common because the parts are large, expensive to build, and have well-defined wear regions where filler can be laid down and machined back. PWHT capability is usually available because the parts justify the heat-treat fixturing.

For PM tool-steel dies anywhere, the maintenance ladder is shorter. Polish and re-coat where the substrate allows. Weld repair is not on the menu. Scrap earlier than on an H13 equivalent.

Internal pushback questions

  1. For each active die in the shop, what is the maintenance history, and is the next proposed repair the second, third, or fourth intervention on that die?
  2. When the maintenance lead proposes a weld repair, is preheat plus interpass plus PWHT plus re-nitride quoted as one job, or is the PWHT step missing from the quote?
  3. Who in the shop is responsible for the dye-pen on a polish candidate, and does that person see the die before or after the polish?
  4. What is the threshold, written down, at which a die moves from "repair again" to "scrap and order replacement," and is anyone tracking the cumulative engineering effort across a die's maintenance history?

Common confusions

A polish is not a repair. It is a finish-restoration step that only works on intact surfaces. Treating polish as a repair option for a cracked or washed die hides the failure mode for one more inspection cycle.

Weld repair is not optional PWHT. The 2025 Lachowicz work made the gradient measurable: as-welded HAZ hardness varies by 200 to 300 HV across a few millimeters; post-PWHT converges to a uniform tempered martensite. The PWHT step is the repair. The weld itself is just deposition.

Re-nitride is not a "refresh." It is a strip plus reprocess plus finish cycle on a substrate that has to be sound. If the strip step is skipped, the new case grows into a contaminated surface and fails early. If the substrate is unsound, the new case hides the underlying failure for a few thousand hits and then the die fails through the case.

Re-coat is not "the coating cycle." It is the prep work plus the coating cycle, and the prep is usually where the engineering effort lives. A re-coat quote that does not specify prep is a quote for half a job.

Scrap is not failure of maintenance. Scrap on the right schedule is good maintenance. Carrying a die through a third repair because the original build effort feels sunk is what fails.

Up next: die change-out economics. When pushing a tired die one more shift costs more than pulling it now.

Sources