Step 2 of 4
43%
Operator and crew procedures: the small habits that double die life
Ten shop-floor habits, grouped into four routines, that separate a 40K-hit shop from a 120K-hit shop on the same die. None of them require new equipment. All of them require that the crew actually do them, every shift, and log the result.
Step 2 of 4In-cycle habits
The shift is running. Billets are dropping, parts are coming off, the die is at temperature and inside the operating window. Three habits during the run that turn eight hours of production into eight hours of comparable, observable data.
Debris management in the flash gutter, every cycle. Flash and scale accumulate in the gutter and around the impression edges. Left for ten cycles, the debris bridges the parting line, locally insulates regions of the die, and on the next hit produces an off-axis load as the billet rocks on a scale chip the size of a coin. The shops that hit 120K wipe or air-blast the gutter every cycle, or have the automation do it. The shops that hit 40K wipe the gutter "when it looks bad," which is also the moment the load path on the die has been wrong for the last 50 hits.
Observation logging during the run. Not just defect logging. A column on the run sheet for the operator to mark anything that is different from the previous comparable point in the shift: a louder release, a part that hung longer than usual, a momentary change in spray sound, a visible flash burst, a billet that came out of the furnace looking a different color. Most of these observations turn out to be nothing. A small number of them turn out to be the first signal of a process drift that, three hours later, becomes a problem nobody can backtrack. The log is the backtrack.
First-piece check at shift start, and at any process change. Before the run is allowed to be "in production," the first part off the die at shift start gets a full dimensional check against the print, not just the operator's visual scan. Same after any change: new lube lot, new billet heat, spray bar tweak, preheat method change. The first-piece check is the place where a process drift is cheap to catch. The fiftieth-piece check is the place where the drift has already produced fifty marginal parts and the die has spent fifty cycles loading the wrong way.
The internal pushback is "we already inspect parts." The distinction matters. Inspecting one part in fifty for a dimensional sample is a quality check on the parts. The first-piece check is a process check on the die and the process. They are not the same and they do not substitute for each other.
Quick check
A die at 70K hits is still hitting tolerance on every part the QC sample catches. The operator notices the release sound on hits 320 through 340 of the shift is slightly louder and the parts in that range come off slightly hotter, but the next 50 parts pass the sample QC check normally and the operator does not log it. Four shifts later the die fails with adhesive pickup in one impression region. What did the observation log have been worth, and where does the failure trace back to?