Motor power suddenly drops followed by crash landing

🧪 Flight Testing & Log Analysis
1

While flight testing one of our drones, it had a sudden throttle loss for only about a second, and it crash landed. I put it in STABILIZE mode as quickly as possible, but it was not very high, so I didn’t have enough time to avoid a crash landing.

https://review.px4.io/plot_app?log=752459b8-adc8-41cd-b81e-9ad967fb156f

If you look carefully at time 9:05 - 9:06 there is a momentary drop in current to zero amps. I took a look at the csv files for PWM actuator outputs and saw that the PWM value for throttle dropped from about 1600 usec to 1075 usec during this interval.

During this time the drone was in DESCEND mode, so it started to come down slowly, which is the expected behavior.

What would cause the PWM values to drop like that? The battery voltage is good. The other parameters look OK. This appears to be some kind of momentary glitch, but what would cause it?? Very disconcerting. Never saw this before.

Any thoughts?
Thanks,
Joe

2

I agree with you, so:

  1. GPS is lost (something must be up with the cabling or receiver)
  2. It goes into descend mode and first descends slowly.
  3. Then suddenly it goes to almost minimal throttle.

All I can see is that the estimated altitude suddenly goes up instead of down, so this would hint at an estimator problem.

@Paul_Riseborough could you have a look here? Thank you.

3

The large GPS vertical velocity innovations indicate GPS data is poor quality throughout flight - what receiver model is this?

Innovations are the difference between the inertial calculation and the sensor. The baro height innovations are small which rules out a IMU data quality issue.

35%20pm
35%20pm1716×1620 461 KB

After loss of GPS for the period, when it comes back for the first few measurements it has a large error as indicated by the baro height reducing slightly, but the GPS vertical velocity measurement showing a climb rate of 3.5 m/s. This degrades the height rate estimate of the EKF. This effect is made worse by the fact that the receiver does not report a velocity accuracy so the EKF uses a parameter defined default value instead.

The GPS in its current state is not safe for flight.

1 Like