CoatingIQ
← Course index

Wrap-up

86%

Lesson 09·Running Forge Dies

Diagnostic workflow when a die starts losing tolerance

The triage order that keeps a shop from replacing a working die because the spray bar moved. Part first, die surface second, process third, press last.

6 min readLesson 9 of 13

Tying it together

The triage order

Four steps, in order. The order is the lesson. Skipping ahead is how a working die gets pulled.

  1. The part. Which dimension drifted, in which direction, on how many parts, and is the drift uniform across the impression or localized? The part is the cheapest diagnostic the shop has. It is already in your hand.
  2. The die surface. Visual inspection without a pull. Photos at the same angle and lighting as the last inspection. Heat-check pattern, lube coverage, evidence of adhesive pickup or washout. Mid-cost in time, low-cost in tooling.
  3. The process. Lube system function and pattern, preheat verification, billet temperature trace, alignment check, press tonnage signature. Mid-cost. Touches several systems but does not pull the die.
  4. The press. Ram alignment, parallelism, tonnage monitor calibration, slide condition. Last, because the diagnostic work is expensive in downtime and the answer is rarely the press if steps one through three have not produced a finding.

Walk the steps in order. Document each one. If step one resolves it, you never touch the rest. If step three resolves it, you saved the cost of a press service call. If step four is the answer, you have the documentation to prove every other variable was checked first.

Worked example

The same die at 60K hits from the opening. Parts are 0.05 mm thick on the same dimension across a 200-part lot, uniformly.

Step 1 · Part. Wall thickness is over by 0.05 mm. Uniform across the lot, not localized to one feature on the part. Step change at the start of second shift. Other dimensions in tolerance. Finding: uniform thickening, step change, single dimension. Consistent with die opening up or process change, not with a discrete die fracture.

Step 2 · Die surface. Photo at the same angle and lighting as the inspection from 10K hits ago. Heat-check network looks unchanged. No new cracks. Washout signature on the high-flow region looks unchanged. Wipe test on the die face after a lube cycle shows uniform graphite coverage except one corner near the dimension that drifted, where the wipe comes back clean. Finding: lube coverage gap on the corner of the impression that corresponds to the dimension that drifted.

Step 3 · Process. Spray bar inspection. The nozzle on that corner is partially clogged with graphite buildup. Spray pressure on the bar is 65 psi (on spec) but the corner nozzle is delivering an estimated half the volume of the adjacent nozzles. Tank concentration is on target. Preheat record is clean. Billet temperature is on target. Alignment is within spec. Tonnage trace shows a small increase in peak over the last 500 hits, consistent with a slightly larger contact area from accelerated washout in the under-lubricated corner. Finding: clogged spray nozzle on the corner that drifted, secondary effect of localized washout from under-cooling and under-lubrication.

Fix. Pull the nozzle, clean it, reinstall, verify spray pattern on a paper target before restarting. Run 100 hits and re-measure the dimension. The drift either resolves (the corner is now cooling and lubricating correctly) or it does not (the washout that opened the impression in the clogged-nozzle period is permanent and the die is now slightly oversized on that corner). Either outcome is information. If the dimension comes back, the diagnosis is closed and the die continues. If it does not, the next decision is whether the localized washout justifies a pull-and-polish or whether the dimension can be held by adjusting the process around it for the remaining die life.

Cost of the right diagnosis. Maybe two hours of shop-floor time across steps one through three. Cost of the wrong diagnosis (pull the die and replace it): the downtime impact of a die change plus the larger downtime impact of bringing a new die online plus the carrying cost of a working die now sitting in the rack with a tag that says "worn out at 60K" when it was not.

What this means on the shop floor

Closed-die forge with mid-life tolerance drift: the triage order resolves most findings at step three. The lube system is the single most common root cause of part-dimensional drift on a die that was running well a shift ago. Walk the spray pattern, the nozzles, and the tank before discussing anything downstream.

Upsetter on long-running die sets: the diagnostic that catches most issues is the photographic record from the prior inspection. Without the photo, the die surface check at step two collapses to "looks about the same as I remember," which is not a diagnostic.

Hot-header with frequent die changes: the most common cause of "this die is running out of tolerance" on a recently installed die is the install itself. Alignment check at step three with a dial indicator catches a die that was bolted in slightly off-center before any other variable is investigated.

Ring-roll with large dies: the press tonnage signature has more diagnostic weight on ring-roll than on closed-die because the load path is longer and slight press condition changes show up sooner. The tonnage trace check in step three is worth doing earlier in the walk on this equipment class.

Internal pushback questions

These are for the supervisor and the maintenance crew, not the vendor.

  1. What is the documented triage order on this floor when a die starts losing tolerance, and is it written down where the shift supervisor can find it at 2 a.m.? If the answer is "we just pull the die," the order does not exist as a process.

  2. When was the last time a die was pulled and the post-pull inspection showed nothing wrong with the die? If the answer is "never" or "we don't track it," the shop is pulling working dies and burning the cost.

  3. Who has the authority to hold a die in the press while step three runs against shift schedule pressure? If the answer is the operator alone, the schedule wins and the diagnosis loses.

  4. What photographic record exists from the last inspection of this die, and at what angle and lighting? If the answer is "we'd have to dig through phones," step two has no baseline and the visual inspection is impressionistic.

Common confusions

A die that "looks worn" on a casual visual is not a finding. Step two requires a comparison against a prior photograph at the same angle and lighting. Without that, the inspection is an opinion.

A part that is uniformly out of tolerance in one direction is rarely a fractured or cracked die. Crack-driven failures show up as localized defects, surface marks, or sudden one-off rejects, not as a clean uniform shift across a lot. The triage order respects that distinction.

The tonnage trace is a process diagnostic, not a press diagnostic. A change in tonnage signature can come from the die, the lube, the billet, or the press. Use it in step three to point at the right next variable, not as evidence the press is the cause.

"We replaced the die and the problem went away" is not proof the die was the cause. Replacing the die also reset the lube nozzle alignment, the preheat sequence, the operator attention, and the inspection cadence. The new die is running because all of those got fixed at the same time. The old die may have been fine.

Up next: maintenance options when triage points at the die itself.

Sources