I have a couple of Tarot T18’s with pixhawk2’s that have been flying Copter 3.5.2 for a while. I’d like to upgrade to Copter 3.6.0 to fix an altitude drift issue (this PR), however on initially flying it the vehicle seemed less stable than it did on Copter 3.5.2.
It wasn’t unstable to the point of crashing or anything, but the gimbal which had been able to stabilize the video when the vehicle had Copter 3.5.2 was not able to do so with Copter 3.6.0. I tried briefly to adjust the gains (decreased roll / pitch rates PI from .2 to .15 and increased the roll / pitch rate D from .01 to .015) but that appeared to make things slightly worse). I was limited on time and wasn’t able to try adjusting the gains in the other direction, but hopefully will this week when I go out and fly. Is there anything between the firmware versions that could cause this?
I had a look at the logs and it’s flying pretty well. It’s got a slight overshoot in pitch but it’s not too bad.
Except for Loiter which has changed a lot, I’m not aware of any significant changes to the attitude controllers between Copter-3.5.x and 3.6.
The vibration levels are a bit high on this vehicle by the way. They are in the “grey zone” especially on the X-axis (forward-back) although like I said, it doesn’t seem to be affecting the vehicle’s ability to control itself.
Hi Randy,
Sorry for the late response–thanks for taking the time to look at the logs. I went out again yesterday with two T18’s (same as before, one with Copter 3.6.0 and one with Copter 3.5.2) and the vehicle with Copter 3.5.2 was definitely more stable (control gains on both vehicles were the same). Both vehicles were configured the same (same payload, weight, etc.) I guess it’s possible that esc’s on the two vehicles might be configured differently, so I think I’ll swap firmware on the autopilots (downgrade the autopilot with 3.6.0 to 3.5.2 and upgrade the autopilot with 3.5.2 to 3.6.0) and fly them both again. Any other tips on how to chase this down? I can upload additional logs if that would be helpful.
Also, the vibes–I also have an S1000 with a pixhawk 1 that is mounted on something similar to this that works really well. The T18’s have a pixhawk2 (cube) which I understand is already vibration damped so it just hard-mounted to a payload plate on the frame. I had considered mounting it similar to what I have on the S1000, I was just a little concerned that it would mask an underlying problem if say the props were out of balance or if it would overdamp the autopilot (since it should already be damped).
swapping the hardware around so that it’s very clear that it’s the software difference that is the root of the issue is a good idea.
comparing the parameters across the two boards to make sure they’re identical is also a good idea.
It’s also good to be precise about what is less stable. Is it’s attitude or position? is it a fast wobble? basically I guess it’s about making sure there’s a good objective test of what is “stable”.
It’s mainly attitude. It might wobble a little more on position, but that might just be a result of it being more twitchy in attitude. Heading appears to be spot on–no wash bowling or anything.