How would you troubleshoot intermittent APU voltage fluctuations?

[Richard92]

I'm an A&P mechanic working at a regional airline, and we're starting to see recurring intermittent faults on a specific model's auxiliary power unit that are proving difficult to diagnose. The logs show random voltage fluctuations, but ground runs and standard checks don't replicate the issue. For other maintenance technicians who have dealt with elusive electrical gremlins, what's your systematic troubleshooting approach when the fault isn't present? How do you effectively use historical data and collaborate with engineering without excessive downtime, and what non-obvious wiring or connector inspection points would you prioritize? Are there any specific diagnostic tools or techniques for tracking down intermittent faults that have proven invaluable in your experience?

[Elizabeth36]

You're not alone—intermittent electrical faults are notoriously hard. Start with a structured problem statement and data-driven timeline. Build a fault history by correlating flight logs, ground-test results, environmental conditions, and recent maintenance. Frame plausible root causes (transient electrical noise, grounding issues, connector failures) and design a minimal test plan that isolates variables without forcing downtime.

[BrandonCS]

Leverage historical data: set up a fault-tracking board to log every incident, sampling rate, phase of flight, temperature, vibration, etc. Use reliability data analysis (counting occurrences per cycles, time-between-failure) to identify patterns. Work early with engineering to access telemetry and trending to confirm hypotheses, so you can plan tests during scheduled maintenance or ground runs with minimal impact.

[Stella92]

Non-obvious wiring/inspection: prioritize harness routing and clamp integrity, bulkhead connectors at the APU interface, and grounds/bonding. Look for signs of chafing, insulation wear, discoloration from heat, and moisture ingress in connector shells. Check for damaged shielding or EMI filters and ensure connectors are seated and locked. Inspect penetrations where harnesses pass through metal, as corrosion can be invisible inside.

[Donald60]

Diagnostics toolkit: use a portable data logger or scope to capture line voltage/current waveforms during a representative ground run, if allowed, and compare to baseline. Use differential probes for noisy lines, and consider thermal imaging to spot overheated connectors or components. Maintain a library of 'fault signatures' from past incidents to accelerate triage. And ensure you’re compliant with MEL and OEM test procedures.

[Andrew10]

Downtime-conscious workflow: coordinate with operations to schedule trials during planned outages; use parallel testing where possible; implement a staged, time-boxed test plan with clear pass/fail criteria. Document each test with time stamps and results so engineers can review without re-running full tests. Keep spare parts or test patches ready to reduce repeated downtime.

[VioletJH]

Communication and collaboration: set up a small cross-functional task group with maintenance, reliability, and engineering; share concise incident briefs; request access to real-time data streams when available; maintain a centralized log and a simple 'issue card' for each fault. Ensure documentation is auditable and safe to share with regulators if needed.