ECSS Harness Manufacturing — Complementary Crimping Requirements (space-systems/ecss/q2030-comp-crimping)
Use when the task is the complementary crimping layer of ECSS-Q-ST-20-30C clause 7.4 -- the small set of space-specific crimp requirements that sits on top of the commercial workmanship chapter the standard delegates the bulk of crimp acceptance to, and that wins wherever the two overlap.
Domain quick reference
- The standard does not restate crimp workmanship; it adopts an external workmanship document chapter by chapter and then adds a complementary clause. Reading only the complementary clause misses most of the crimp criteria; reading only the adopted chapter misses the deltas that make the termination a space termination.
- Precedence runs by SOURCE, not by tightness. Where the complementary clause and the adopted chapter both speak to a characteristic, the complementary value governs even when it is the looser of the two. That is a deliberate engineering decision, so the correct behaviour is to apply it and report the relaxation, not to quietly substitute the tighter number and call the result compliant.
- Crimp height is the process witness, not a cosmetic dimension. It is the one measurement that ties a finished crimp back to the qualified tool-and-die setting, so it is graded against a window keyed to the contact part number and the wire gauge together -- the same contact on a different gauge is a different window.
- The tensile check is destructive and therefore sampled. A sample is sized from the lot, never fixed at a habitual three, and every pull in it has to reach the minimum its gauge carries. A sample mean is not the criterion: one low pull is a failed sample however high the others ran.
- Two constructions the complementary layer rules out outright are a second wire pushed into a barrel qualified for one, and solder anywhere in a crimp. Both produce a joint that passes a pull test on the day and fails later, the first by unequal strand compression and the second by a wicked, stiff transition that cracks under vibration.
- A crimp is only as traceable as the tool that made it. A lot crimped with a tool whose calibration has lapsed has no valid process record, and that is a lot finding rather than a per-crimp one.
Workflow
- Validate the lot: each crimp record carries an identifier, a gauge, a measured height and the nominal-plus-tolerance window it is graded against. Reject a duplicate identifier rather than folding two records into one.
- For every characteristic that both sources speak to, resolve the governing limit by precedence and record whether the complementary value relaxed the delegated one.
- Grade each measured crimp height against its window, absorbing representation error at the window edge with a named tolerance instead of widening the window.
- Examine each record for the ruled-out constructions: an unqualified multiple-wire barrel, solder in the crimp, an insulation support the contact provides but the crimp left unengaged.
- Size the tensile sample from the lot size, the sampling fraction and the sample floor, and take the governing minimum force as the toughest gauge present in the lot.
- Grade every pull against that minimum, absorbing an exact equality with a tolerance, and treat a short sample as a finding in its own right rather than as a pass on the pulls that were run.
- Confirm the tool calibration interval, then roll every per-crimp and lot-level finding into one verdict with the findings named.
Pitfalls
- Substituting the tighter of the two sources. Precedence is by source; silently applying the delegated number where the complementary clause has spoken hides a relaxation the project accepted on purpose.
- Grading a crimp height against the contact alone. The window depends on the contact and the gauge together, so a gauge change inside the same connector is a new window and usually a new die setting.
- Reporting a tensile sample by its mean or its median. Every pull owes the minimum; a mean hides the one strand-damaged crimp the sample was taken to find.
- Running a habitual three pulls whatever the lot size. The sample is sized from the lot; a fixed count is a coverage finding on a large lot and wasteful destruction on a small one.
- Accepting a double crimp because the pull test passed. The pull test measures the barrel, not the two wires inside it; unequal compression shows up as an intermittent months later, so the qualification of the multiple-wire process is what licenses it, not the force reading.
- Treating a lapsed tool calibration as paperwork. Without a valid tool record the height measurements have nothing to tie back to, so the whole lot loses its process evidence, not just its signature.
- Failing a crimp whose height lands exactly on a window edge. The equality is a representation question, settled with a tolerance inside the comparison and never by moving the window.
Behavior contract (gate 3)
The source normalization, the precedence resolution and relaxation flag, the crimp-height window and edge grading, the gauge-keyed tensile minimum, the sample sizing, the per-pull grading, the ruled-out construction screens, the tool-calibration currency check and the whole lot verdict are exercised by the gate 3 contract test: scripts/test_q2030_comp_crimping.py against scripts/q2030_comp_crimping_logic.py (stdlib unittest, offline). Run: python3 scripts/test_q2030_comp_crimping.py
Compliance
- ECSS standards are freely downloadable (ESA); cite the source and paraphrase per standards-map.yaml.
- compliance: STANDARDS-REF, gated: false.
