Trouble with very first tuning stages for 10" Quad

ArduCopter Copter 4.5
1

Update:
I thought I had bdshot firmware flashed but that was not the case. I’m sure my HNTCH filter did not like that. Upgraded to 4.5.5 bdshot and getting ready to retest.

I have a GEPRC EF10 that I reconfigured for Arducopter 4.5.4.

I’m working through the Arducopter 7" Long-Range Standard Build guide as well as the official Ardupilot Initial Tuning Flight documentation.

First flight with 10" defaults from MP: I engaged POSHOLD and the quad began slowly sinking towards the ground at approx 0.5 m/s and I can hear motor small oscillations at about 1.5-2 Hz.

I then reduced PSC_POSZ_P and PSC_VELZ_P by 50% and tried again. This time there were no audible oscillations but the quad still sank slowly into the ground at about 0.5 m/s.

I haven’t gotten to AUTOTUNE yet because I need to perform it from POSHOLD due to persistent light winds where I live (and there’s a pretty significant yaw instability on this quad with default PIDs that is a real pain to manually counter constantly).

Any suggestions would be welcome. Attached video, logfile, and param files are from the latest config where I reverted back to defaults, with audible vertical oscillations and slow sink into the ground from POSHOLD.

Thanks!

logfile
video
paramshttps://drive.google.com/file/d/1aKstpbBd04AiDIJCl7pewZ0MVULhDrwZ/view?usp=sharing

2

Most of the dev team will tell you to use the methodic configurator, and post a config and bin file.

3

Links to those files are posted in the OP

4

You should not start with POSHOLD. If you are not confident in stabilize then start in AltHold. Get AltHold working well before you move to modes that the EKF is controlling more strongly.

1 Like
5

Thanks @andyp1per. After fixing my firmware to the correct version (bdshot) I ended up cutting the default PID values in half and same for PSC_POSZ_P and PSC_VELZ_P and that fixed the althold problems. I was finally able to get through an initial autotune and now the quad is handling much better. Ready to progress.

1 Like
6

The methodic calculator did cue me in to the fact that I didn’t have a BDSHOT_MASK available to set, so that was good advice. Thanks!

1 Like