GPS Loss and Crash in Loiter

ArduCopter Copter 4.4
, ,
1

There is a quadcopter with a CUAV X7 flight controller
The GPS CUAV 2HP module is installed and configured
Everything was fine for over 30 flights, I did all the basic settings, autotune, PIDs, etc.
In the last flight in Loiter mode, the number of satellites dropped to 0, HDOP increased to 170
The copter became at 70 degrees pitch and gave maximum to the motors. The drone flew to a nearby tree and crashed.
The maximum angle in Loiter is set to 15 degrees, the maximum speed limit in Loiter is set to 3 m/s.
Why did the copter react like that?

Log

Thanks for the answer!

2

At a time point of 157.4 seconds, the logs show that actual pitch started to diverge from the desired pitch by up to 8 degrees, which is way worse than in the preceding part of the flight. This means the copter started to incline forward.

(Actually pitch diverged up to 2 degrees earlier in the flight, unlike roll, which could also be an indication that something is not quite well. This is surely not a universal threshold for all copters, but for this one it might be already too much)

At the same time, the current consumption sharply increased from 40 to 54 amps.

The RPM from motor 3 (forward right, if I’m getting your frame configuration right) started to increase at the same moment. Something like this would normally result from a sudden undesired pitch forward, but a little effect from this indicates the forward right motor/propeller group has likely failed in some way. The nature of the failure is not known to me, and I cannot even guess because I don’t know the details of your design, in particular how exactly propellers are attached, and how motors are mounted to the frame.

Around that time, at 160.1 seconds, you switched to Stabilize, and current consumption got to 44 amps, already greater than 40. At this point, the RPMs from the motors still look pretty unbalanced compared to where they have been. The failure was evolving.

Why exactly the number of satellites dropped around the time when you switched back to Loiter (170.3 seconds), is not clear either. It could be a brownout, but it was not associated with any current spike. It could also be electric signal disturbance, especially if the GPS wiring comes close to power wires. It might even be GPS mount instability… Anyway, pitch was not healthy due to the previous failures and started to oscillate with an increasing amplitude, so the crash was imminent.

TL;DR: Looks like hardware failure to me. Check your propellers, motors and/or ESCs.

EDIT: after more details revealed, and after more investigation, this was found to be a non-obvious pilot error, which got worse due to poor tuning.

1 Like
3

Thank you for your reply.
There was a problem with GPS.
I am wondering why the flight controller ignored the maximum angle and maximum flight speed restrictions in Loiter mode
And why in Loiter mode, when there was a navigation error, the drone continued to fly in this mode, and did not change the flight mode to one that does not use navigation.

4

I only have to state it once again - the copter was already crashing regardless of GPS. A copter cannot adhere to any limits or follow any commands when motors or propellers are seriously misbehaving, (unless it’s a hexa or octacopter with a configuration tolerant to one motor failure). A quadcopter cannot really fly with only three motors alive. Have you actually read my post?

EDIT: it was difficult to understand what is actually happening without knowing one very important detail… but GPS is out of question anyway.

3 Likes
5

@PaulZP please read the posts from @MaxBuzz carefully and multiple times until you understand it.

6

Thank you, I read your first post carefully. Unfortunately, during the fall, one of the motor mounts was damaged and the propeller was broken. But with a probability of 99%, this happened on impact, and not in the air.

7

That does not change anything in the correct analysis that @MaxBuzz did.

You had an hardware failure before you had a gps glitch. The crash was not caused by the glitch.

2 Likes
8

Thanks @MaxBuzz
He helped me fully understand the situation!

9

To summarize the discussion (that happened in private messages) for future readers:

  • The GPS reception quality was indeed degrading a lot due to the environment, but this was not the root cause.
  • The non-obvious but important thing is that the copter’s job was to regularly contact walls, and the incident started when it attempted to detach a wall, which was unsuccessful.
  • The incident very likely started with a pilot error that led to an attitude/altitude conflict and high imbalance of motor outputs, which, at detach, resulted in a high undesired pitch action.
  • Pitch was sufficiently undertuned, so the copter could not react quickly when it detached.
1 Like
10

Yet another vehicle that could benefit from the ArduPilot methodic configurator software.

If you use it, follow all the steps and instructions. Skipping things gets you suboptimal results.

1 Like