VTOL motor rotation unstable when hovering

ArduPlane VTOL Plane
1

Hello, I am currently testing a quadplane and facing a serious issue during hover. Occasionally, one of the VTOL motors suddenly goes “idle” (cuts out for about 0.5–1 second) and then comes back on. This happens randomly—both after attempting transitions and even during a simple QLOITER hover—and has caused multiple uncontrolled drops because the aircraft instantly loses stability. I have attached the video and flight logs.

Summary of the flight tests:

  1. Test 1: Hover → attempted transition (failed) → switched back to QLOITER for landing. At ~15 m altitude, the front-left motor briefly idled, recovered, idled again, and the aircraft fell.

  2. Test 2: Transition succeeded normally, no issues observed.

  3. Test 3: Only QLOITER hover, but the same sudden motor idle occurred and the aircraft entered free fall.

A consistent observation after crashes: the input capacitor (Rubycon 63 V 650 µF) connected to the ESC is found physically broken (the capacitor body itself, not the solder joint). My setup uses T-Motor MN7005 motors, Flycolor XCross HV 80A ESCs, and a Tattu 10 Ah battery. I am unsure whether:
A) the capacitor broke because of the crash, or
B) the capacitor failed first, causing unstable ESC power and making the motor cut out.
The logs show irregular RCOUT values for the VTOL motors (ch9–13) before the fall.

Some friends suggested:
– The capacitor may be failing due to high vibration (it is mounted directly under the ESC/motor).
– A failed/missing input capacitor could cause voltage spikes/drops and trigger ESC protection, leading to brief motor cut-outs.
– Possible signal or ground reference issues on the PWM line.

I would really appreciate input from the community:

  1. From the attached logs (RCOUT + Vbatt), does the motor drop look like an ESC cut-out or a command from the flight controller?

  2. Can a failing/broken input capacitor realistically cause this intermittent motor idle behavior?

  3. Any recommendations for ESC mounting, capacitor placement, or additional tests to isolate the root cause?

Thank you — logs and video are attached.