Cabin Lighting That Talks to the VHF Radio
After fitment of a new cabin LED lighting controller, the crew report occasional brief clicks and a wavering ident on VHF COM 1 (118-137 MHz) during taxi, and one unexplained autopilot disengagement during climb. The controller's power return is a 0.6 m wire bonded to a nearby composite panel with a single 6 mm bolt through paint; the harness to the lights runs parallel to the VHF antenna feeder for 3 m at 0.15 m spacing. The lightning/EMC manual requires bond resistance of 2.5 milliohms or less and separation of 0.5 m from antenna feeders, or shielded twisted pairs with the shield grounded at both ends. The complaint appears only with the lights above 80% brightness (PWM dimming).
1. Classify the coupling mechanisms at work and explain the significance of the brightness threshold.
Read the worked answer
Two mechanisms, one root cause.
- VHF clicks and wavering ident: the dimmer's PWM switching edges are the emissions source. A 3 m parallel run at 0.15 m from the antenna feeder, on an unshielded harness, is a textbook radiated near-field coupling path into the victim feeder.
- Autopilot disengagement: low-level signalling is normally a common-mode/conducted problem rather than a radiated one. The 0.6 m return bonded through paint on a single bolt is an inductive, high-resistance path, far above the 2.5 milliohm requirement, so PWM return current flows through structure of unknown impedance and develops a common-mode voltage on nearby wiring sharing that structure.
Brightness threshold. Interference appearing only above 80% brightness ties the source to the dimmer duty cycle, not to the lamps. As brightness rises the dimmer's pulse width and edge rate change, raising both the amplitude and the harmonic content in the VHF band. This is a diagnostic gift: it identifies the dimmer as source and lets you prove the fix.
Single explanation for both symptoms: the bad bond. The parallel run explains why VHF is the most sensitive victim.
2. Rank the corrective actions in order and state how you would verify each one.
Read the worked answer
1. Restore the bond (primary, addresses both symptoms)
- Bond to bare metal structure with a proper bonding jumper or stud, remove and protect paint at the joint, then measure with a four-wire bond meter: target 2.5 milliohms or less, recorded in the log.
- This is a continuing-airworthiness requirement, not an optional improvement; a painted, single-bolt joint is not a bond.
- Verify by repeating the 100% brightness ground run and re-measuring the bond after disturbance.
2. Restore separation or shielding (addresses the VHF path)
- Re-route the light harness to 0.5 m or more from the VHF feeder.
- Where that is impossible, use shielded twisted pair with the shield grounded at both ends, as the manual permits. Single-end grounding is wrong for RF: it leaves the shield acting as an antenna.
- Verify by checking ident quality and background noise at several frequencies across 118-137 MHz, at full brightness, with the engines and other systems running.
3. Supplementary measures
- Ferrite at the controller, better filtering on the dimmer, or a shielded controller enclosure - only if residual emissions remain after steps 1 and 2. Treat these as refinements, never as substitutes.
- Re-run the affected-system tests (autopilot) after any wiring change, since a common-mode path may have been disturbed.
3. The shop proposes to retune the VHF squelch and clip a ferrite onto the light harness, then release the aircraft. Give your judgement with reasoning.
Read the worked answer
Reject as a release basis.
Why the squelch retune is unacceptable
- It raises the threshold at which audio is passed, so it masks the symptom by reducing usable receiver sensitivity. It does not change the emissions or the coupling path.
- Reduced sensitivity can suppress weak but genuine transmissions - a safety consequence far worse than a click.
- The autopilot disengagement is untouched by a radio setting, and a flight-control anomaly is not a nuisance item.
Why the ferrite alone is weak
- A clip-on ferrite is a supplementary measure: it saturates at high current, is effective over a limited band, and cannot compensate for a bad bond. It does not affect the common-mode voltage developed by return current flowing through a painted joint.
What to do instead
- Correct the bond to 2.5 milliohms or less and restore separation or shielding first; then test.
- Do not release for flight with an unexplained autopilot disengagement attributable to a recent modification. If the root cause cannot be corrected in the available time, remove or isolate the new system.
- Record the defect and raise an occurrence report; a known EMI path that caused a flight-control event is reportable material.