Troubleshooting

Common Mistakes and Troubleshooting in Defense Electronics and Ruggedized Systems

The recurring, expensive errors in ruggedized defense electronics work and how to catch each one before it reaches a review board or a First Article.

Mistake one: sizing an Environmental Stress Screening chamber by unit count instead of thermal load. Symptom is a chamber that cannot hold its 5 C per minute ramp when the shelf is full, so a MIL-STD-810 profile that should run 10 cycles stretches from 40 hours to 62. Root cause is loading 220 boards at 0.18 kg each without adding their mass to the fixtures, ignoring the 55 kg total. Fix: enter mass and dwell into the Environmental Stress Screening Load calculator, then derate chamber capacity 20 percent for airflow blockage before you commit a schedule.

Mistake two: quoting conformal coat as one flat rate per board. Symptom is margin evaporating on the units with 40 masked connectors while thin-population boards look overpriced. Root cause is ignoring masking labor, which on a dense board runs 12 to 18 minutes of hand taping plus 8 minutes de-mask, dwarfing the 90 seconds of actual acrylic or urethane spray. Fix: separate masking count from coated area in the Conformal Coating Cost calculator. A board with 35 keep-out zones can carry 3 times the labor of one with 10, even at identical square inches.

Mistake three: treating obsolescence as a purchasing problem discovered at reorder. Symptom is a Diminished Manufacturing Sources notice landing mid-build with 14 units still to ship and an 18 week lead time on the replacement FPGA. Root cause is never scoring parts at design freeze. Fix: run every bill of material line through the Obsolescence Risk Score calculator at gate review, flag anything above 70, and size a Long-Life Component Buffer for those. A 7 year program with 2 percent annual attrition on a 400 part BOM needs buffer stock on roughly 56 parts, not the 3 anyone remembered.

Mistake four: unit confusion between shock g-levels and vibration Grms. Symptom is a Shock/Vibration Test Capacity plan that books the wrong slip table and blows the test window. Root cause is reading a 40 g peak shock spec and sizing a shaker for 40 Grms random, which is a completely different force demand. A 30 kg fixture at 14 Grms needs about 4,100 lbf of shaker force, while the same fixture at a 20 g half-sine shock is an entirely separate calculation. Fix: keep shock pulses and random PSD profiles on separate lines and never let a g and a Grms share a column.

Mistake five: underestimating setup on low-volume lots. Symptom is a 25 unit rugged enclosure build that loses money even though the per-hour machining rate was correct. Root cause is amortizing an 11 hour fixture and program setup across an assumed 200 units when the real order is 25, so setup per unit jumps from 3.3 minutes to 26 minutes. Fix: put the true lot size into the Low-Volume Setup Cost calculator and the Rugged Enclosure Machining Time calculator together. At 25 units, setup can be 60 percent of touch time, so a 15 percent quantity error swings the quote by double digits.

Mistake six: forgetting program documentation and secure configuration as billable, schedule-driving work. Symptom is a build that finishes machining and test on time but slips 3 weeks on first delivery. Root cause is treating the Contract Data Requirements List, the acceptance data package, and secure firmware provisioning as overhead rather than tracked hours. A single ruggedized LRU can carry 40 to 120 hours of documentation and 6 to 20 hours of secure configuration per lot. Fix: load both the Program Documentation Burden and Secure Configuration Workload calculators into the master schedule so the paper trail has a start date, not just a due date.

Mistake seven: planning the repair depot from optimistic mean time to repair. Symptom is a Repair Depot Throughput plan promising 30 units per week that actually clears 18, so returned assets age past their 45 day turnaround target. Root cause is using bench MTTR of 2.5 hours while ignoring diagnostic wait, awaiting-parts hold, and retest, which push real cycle time to 6 hours. Fix: model the full dwell, not just wrench time. If parts availability is 85 percent, roughly 15 percent of units stall in an awaiting-parts queue that no amount of technician staffing shortens.

The pattern behind all seven is the same: a number entered in the wrong unit, at the wrong quantity, or at the wrong point in the program. Catch them with two habits. First, freeze lot size, mass, and part counts as named inputs before any estimate, and re-run the affected calculator whenever one changes by more than 10 percent. Second, score obsolescence and documentation burden at design freeze, not at reorder or delivery. A 15 minute pass through the relevant calculators at gate review is cheaper than a First Article rejection that costs weeks.

Published 2026-07-01.