Pixhawk Quad PWM issues (motors blocking)


we are running a standard 3DR DIY quad configuration with 3DR Pixhawk. When we rotate the motors, cracking can be heard and it increases with motor speed. We investigated the problem and discovered that every 230 ms, there are a few PWM pulses with the length of 1000 us, as can be seen in the picture. The Saleae Logic captured data is also available.


The same problem can also be observed in these videos I found on Youtube (not my videos):

Can anyone please provide us with guidance on how to fix this problem?

Kind regards,
Blaž Modic

Which version of PX4/Firmware is this, and what hardware?

FW: 1.4.4dev, HW: 3DR Pixhawk 1, vehicle: 3DR DIY quad, airframe type: quadrotor x

Thank you and please let me know if you need any additional information.

I can’t reproduce this on a 3DR Pixhawk using v1.4.4 or master. Logic analyzer and digital scope show no glitches at all. Also, I’ve never observed this in flight. Have you made any changes to the firmware or customized the startup?

We haven’t made any changes to the firmware or customised the startup - all stock configuration. We only optimised the attitude stabilisation control parameter values.

On firmware version 1.3 the error is not present, on 1.4 it is. We also ran the firmware on our second Pixhawk hardware and the problem is the same. Otherwise, Board ID is 9.

Any ideas on what to try next?

Thanks, Marko

Are you flashing the firmware via QGC or building yourself? Maybe a bad build?

Via QGC. Are there different 3DR Pixhawk 1 hardware versions? Otherwise I am clueless why couldn’t you reproduce the error.

I doubt that any likely hardware revisions could cause the behavior you’re observing and I don’t think there were any significant revisions to the Pixhawk schematic during production.
Try disconnecting everything but the safety switch and see if the PWM outputs are still flaky.

Even without any peripheral components the outputs are still flaky. It’s interesting that with the 1.3 version everything is OK, but with the 1.4 the outputs become corrupted. To be clear, flaky outputs are only observed when the output PWM pulse width is greater than 1000us.

Do you have flight logs which show the problem behavior?

Do you have thumb sticks enabled in QGroundControl?

Thank you, that was it! If you disable the Virtual Joystick the problem is gone. Is the Virtual Joystick enabled by default in QGroundControl? I can’t remember setting it up. Thank you once more!