Common Mistakes

Common Mistakes and Troubleshooting in Space Payload and Avionics Manufacturing

The eight errors that most often blow schedules and yield in payload and avionics manufacturing, each with the symptom, the root cause, and a numbers-backed fix.

The most expensive scheduling error in payload environmental test is sizing a campaign by chamber count instead of dwell hours. Symptom: a qual sequence that looked like three weeks runs six. Root cause: a MIL-STD-1540 or GEVS profile at 8 cycles between -34 C and +71 C with 1 hour soaks and a 3 C per minute ramp burns roughly 5 to 6 hours per cycle, so 8 cycles is 40 plus hours per unit, not per chamber. Fix: drive the Thermal Cycle Test Capacity calculator off soak plus ramp time and transition losses, then add 15 percent for chamber recovery. That correction moved one program from 60 percent to 92 percent schedule adherence.

Vibration profiles get botched when engineers add PSD breakpoints as if they were g levels. Symptom: a random vibration run that overtests by 3 to 6 dB and cracks solder joints that would have flown fine. Root cause: overall grms is the square root of the area under the g squared per Hz curve, not the arithmetic sum of breakpoints. A 0.04 g squared per Hz plateau from 20 to 2000 Hz is about 8.0 grms, but reading the plateau as 0.04 g gives nonsense. Fix: integrate the PSD before you load the Vibration Test Schedule calculator, and confirm the controller shows the same grms within 0.5 before you release the shaker.

Conformal coating yield collapses when masking labor is left out of the first-pass model. Symptom: quoted yield of 95 percent, actual 78 percent, with rework on connectors and test points flooded by acrylic or silicone. Root cause: keep-out violations and bubbles, not coating chemistry. On a board with 40 masked features, even a 1 percent miss-mask rate per feature compounds to roughly a 33 percent board-level defect chance. Fix: feed real feature counts and operator masking time into the Conformal Coating Yield calculator, and hold coating thickness to spec, typically 25 to 75 microns for Type AR, verified under UV before cure locks the defect in.

Component derating gets treated as a checkbox instead of a temperature-dependent calculation. Symptom: parts pass at ambient, then fail after months at hot orbit dwell. Root cause: engineers derate voltage to 50 or 60 percent but ignore junction temperature, where a capacitor rated 105 C run at 90 C junction has a fraction of the life it shows at 40 C. Every 10 C rise roughly halves electrolytic life. Fix: run the Component Derating Risk calculator with worst-case hot-side junction temperatures, not board ambient, and hold resistor power dissipation under 50 percent and semiconductor junctions at least 25 C below max rated for Class B hardware.

Burn-in capacity is routinely overstated because oven slot count is confused with usable channel count. Symptom: a 168 hour burn-in promised in one week actually queues for three. Root cause: a 500-part order at 168 hours each needs 84000 part-hours, and a 96-channel oven delivers only 16128 part-hours per week, so the real answer is over five weeks. Fix: size against powered channels and duty cycle in the Burn-In Capacity calculator, not shelf positions, and reserve 10 to 15 percent of channels for infant-mortality retests so a 2 to 3 percent early-fail rate does not cascade into missed ship dates.

As-built traceability blows up late when serialized part logging is estimated per board instead of per component. Symptom: a delivery held at the customer gate because the as-run record is missing lot dates and date codes. Root cause: a single avionics slice can carry 800 to 1500 serialized or lot-controlled parts, and logging 30 seconds each is 7 to 12 hours of pure documentation per unit that nobody budgeted. Fix: size the Traceability Workload calculator on component count and required attributes per part, then automate capture at kitting so the record is born complete instead of reconstructed from memory at three times the labor.

Nonconformance cost is chronically undercounted because teams log only the rework hours. Symptom: an MRB disposition that reads 2 hours of labor but actually consumed a week. Root cause: the hidden load is engineering disposition, retest, requalification, and schedule slip, which on flight hardware routinely runs 5 to 20 times the visible touch labor. A single escaped part can trigger a 40 hour failure review board plus re-screen of the whole lot. Fix: capture full disposition, retest, and containment time in the Nonconformance Cost calculator, and track cost of poor quality as a percent of program value, where anything above 5 percent signals a process out of control.

Cleanroom assembly plans fail when gowning and airlock time are ignored in station loading. Symptom: an ISO 7 line planned for 8 operators that only fits 5 without particle counts drifting past 352000 particles per cubic meter at 0.5 micron. Root cause: each body adds particle load and each entry cycles the airlock, so overpacking the room defeats the class rating. Fix: load the Cleanroom Assembly Load calculator with air changes per hour, typically 60 to 90 for ISO 7, plus person count, and stage harness work with the Harness Routing Labor calculator outside the critical zone so only final integration happens under the tightest particle budget.

Published 2026-07-02.