ECSS Electrical Test Setup — Tested-Unit Bonding (space-systems/ecss/e2007-tested-unit-bonding)
Use when the task is the ECSS-E-ST-20-07C clause 5.2.6.2 restriction that a tested unit is bonded to the test reference plane only by the bonding means already built into its own design -- categorizing every bond path by provenance, excluding test-added paths from the number, computing the effective bonding resistance of the design-provided set, and reconciling it with the bench measurement.
Domain quick reference
- The bond between a unit and the reference plane is part of what the measurement characterizes, not part of the instrumentation. A strap the laboratory adds for convenience lowers the impedance the flight configuration will actually have, so the measurement then describes a unit that will never fly.
- Every bond path is categorized by provenance before any arithmetic: design-provided (the unit design, its interface documentation, its qualified mounting interface) or test-added (a setup jumper, an auxiliary laboratory strap). A provenance outside that set is a record defect and is rejected rather than assumed benign.
- A test-added path is a non-conformance on its own, independent of what it measures. Its resistance is never folded into the result, because a parallel path can only lower the combined value and would make an out-of-family bond look compliant.
- The direct-current resistance of one path is the strap or foot resistance -- material resistivity times length divided by the cross-section -- in series with the contact resistance of every joint along it. Joints usually dominate a short, thick foot.
- Paths energized together carry current in parallel, so the unit's effective bonding resistance is the parallel combination of the design-provided paths only, and it is always lower than the lowest single path.
- Direct current is not the whole requirement. At radio frequency the strap inductance dominates, and that is controlled geometrically by holding the strap length-to-width ratio at or below a stated maximum (commonly five to one).
- The computed value is finally reconciled with the bench reading. A relative disagreement beyond a stated fraction means either the model or a joint is wrong; it is reported, not averaged.
Workflow
- Categorize every bond path by provenance as design-provided or test-added. Reject an unrecognized provenance, a path without a name, or duplicate path names before computing anything.
- Raise each test-added path as a setup non-conformance naming the path, and remove it from the resistance calculation entirely.
- For each design-provided path, compute the strap resistance from the material resistivity, length and cross-section, and add the series contact resistance of every joint.
- Combine the design-provided path resistances in parallel to get the effective bonding resistance, and compare it with the declared bonding-resistance requirement.
- Check each design-provided strap's length-to-width ratio against the low-inductance maximum; a compliant direct-current resistance on a long, thin strap is still a radio-frequency finding.
- Reconcile the computed effective resistance with the bench reading against a stated relative tolerance, and flag a missing bench reading as a finding in its own right.
- Aggregate: the bonding is as designed only when no test-added path exists, the effective resistance is within requirement, every strap geometry passes, and the bench reading agrees.
Pitfalls
- Adding a jumper because the bond "looked high" and re-measuring. That changes the configuration the clause exists to preserve; the correct move is to record the finding and restore the design bonding means.
- Folding a test-added path into the parallel combination. Parallel paths can only lower the result, so including the jumper turns a non-conformance into an apparent pass.
- Computing strap resistance and stopping there. The joints are in series with it and typically dominate a short, thick mounting foot by an order of magnitude.
- Accepting a compliant direct-current resistance on a long, thin strap. The resistance requirement and the length-to-width geometry rule are independent, and the geometry rule is the one that governs at radio frequency.
- Treating a missing bench measurement as a pass because the computed number is comfortable. The computation uses declared materials and nominal joints; the reading is the only evidence of the joint actually made.
- Reporting only the combined figure. A single bad joint can hide inside a parallel set, so each path resistance stays on the record beside the combined value.
- Letting a series sum that lands a few units in the last place above the requirement read as an exceedance. The logic absorbs representation error with a named tolerance far below any milliohm value; the requirement itself is never widened.
Behavior contract (gate 3)
The provenance categorization, strap and contact resistance, parallel
combination, strap-geometry, requirement-comparison and bench
reconciliation logic is exercised by the gate 3 contract test:
scripts/test_e2007_tested_unit_bonding.py against
scripts/e2007_tested_unit_bonding_logic.py (stdlib unittest, offline).
Run: python3 scripts/test_e2007_tested_unit_bonding.py
Compliance
- ECSS standards are freely downloadable (ESA); cite the source and paraphrase per standards-map.yaml.
- compliance: STANDARDS-REF, gated: false.
