Step 3 of 4
57%
Where the engineering decisions live
The forge die build is a chain of decisions, not a pile of line items. This lesson maps the decisions to where they live in the build path and which ones move die life by the largest margin.
Step 3 of 4The relaxation traps
Some decisions look like small spec relaxations and are not. The relaxation reads as a contained tweak to one cycle, and the consequence propagates through every decision downstream.
Single temper instead of double on a 5 percent Cr steel. The hardness certificate reads in spec, the cycle is shorter, and the substrate carries fresh untempered martensite plus 12-18 percent retained austenite into nitride, polish, and service. The downstream nitride cycle distorts more than it should because the substrate is dimensionally unstable. The polish removes a thicker layer because the surface dimensions drift during nitride. The die enters service brittle and grows dimensionally over the first few thousand hits. The single-cycle relaxation costs roughly 20-40 percent of die life.
Open-loop ammonia bleed instead of closed-loop Kn control on nitride. The certificate reads a case depth number, often without a measurement standard, and the compound layer is whatever the retort produced that day. The downstream polish operates on an inconsistent compound layer thickness. The PVD (if specified) deposits onto a varying surface. The die enters service with a case structure that does not match the print, and the failure mode in service may not match the design intent. The cycle looks identical on the PO and the cert. The consequence shows up in scatter on die life.
Skipping the witness coupon at any stage. The certificate becomes testimony rather than evidence. A failed die six months later cannot be diagnosed because the case structure at ship date is unknown. The relaxation looks like a contained inspection cut. The consequence is that every other inspection artifact in the chain loses its anchor.
Skipping stress relief between rough and semi-finish machining. The part finishes machining on dimension, goes to heat treat, and warps to a new equilibrium shape during the quench. The downstream EDM, polish, and case treatments inherit a part whose actual geometry differs from the print. The relaxation looks like a saved cycle. The consequence is unpredictable distortion that cannot be ground out without compromising case depth.
The common pattern is that the relaxation is invisible on the certificate and visible only in service. The earlier the decision lives in the build chain, the more downstream decisions inherit the consequence. Material-zone relaxations propagate through every later zone. Surface-zone relaxations propagate only through final inspection and service.
Quick check
A nitride vendor proposes substituting an open-loop ammonia bleed cycle for the specced closed-loop Kn-controlled cycle on a duplex die that will get AlTiN PVD downstream. The vendor argues the case depth and surface hardness numbers will still hit the print. Which downstream decisions inherit the relaxation, and what is the in-service consequence?