Update to my MicroCopter Leaning Issue - Solved? (nope..)

I’ve been posting off and on for years now about my 3" prop and smaller copters leaning. Been through vibrations, FC damping/mounting, various settings, different hardware, etc.

I tried ekf3 yesterday and it seemed to fix it. I flew a few packs again today w/ ekf3, and again, no problems. I’ve tried two copters now… my worst two. And on both, ekf3 does not cause the leaning issue. The issue was somewhat random, but more often than not it leaned. I think I’ve done enough flights w/ ekf3 at this point to say yes, it has solved it.

Pinging @andyp1per, just because he’s been looking at my logs with me about this issue for at least the past year.

That’s great - would be good to see some logs of the copter working! I’m not sure what would be different about EKF3 but leaning is definitely due to bad data in the EKF so it probably copes better (i.e. I suspect the underlying physical issue is the same).

Agreed… I do wish I knew what was causing it. I think one of your suggestions months ago to try a 3s battery is something I do need to do at some point. I’m starting to wear out some of my 2s bats, so maybe soon I’ll tear a few apart and make a 3s bat w/ the good cells.

I haven’t looked at the logs yet… I’ll download them later and post one.

ARG, never mind… I flew at least 10 packs yesterday and this morning… No problems. I’ve never had so many non leaning flights in a row on these copters.

Decided to take a flight now to get a new log, no changes, took off and it’s leaning. not as bad… but leaning.
I initially thought ekf3 completely eliminated the issue, but apparently not.

Since the issue is somewhat intermittent, I’ll need to take a lot more flights to see if it’s actually helping at all… I think it is. I’m just back to not being confident in my results since every flight is a bit different.

I did notice, GPS signal was great yesterday and this morning… Got 10 satellites inside in just a minute or two. Now it’s taking 10 minutes to get 10 satellites. Also, the good flights didn’t have the IMU-GX offset, where it centers above zero. Good flights were centered on zero. This latest flight, it’s centered on zero at boot, but around .03 at the end of the flight.
Back to testing…

Is this with the omnibus nano?

Well, these two I’ve flown the past couple days are kakute mini’s. But, I do also have two omnibus nano’s w/ the same issue.
All 4 of my microcopters do this. (It’s why I have four… I built #2 and #4 because I thought I had hardware issues on #1 then #3 and wanted to rule that out… They’re good for spare parts, too, if the need arises)

For what it’s worth I had the same issue on the kakute mini and omnibus nano on a 7" build. The nano seemed to fly better overall.
I switched it to an mRo control zero and it’s been rock-solid.

I’m working on a new 3" quad and its working great apart I still not able to finish the autotune procedure due to a similar leaning issue. Today I’m going to fly a new 3" unit I made to try to figure out this and I’ll share here my fly log. Also I’ll trying to use the EKF3. I really want to figure out if its the same issue as yours or not, if we have a bug we have to solve it!

What’s IMU-GX offset ?

I guess it does not to be affected by GPS signal quality, otherwise why without GPS we have not this issue in our 5" quads and bigger one?

@wicked1 @andyp1per I’m not an expert when it comes to log analisys, could you suggest me the right log bitmask to use?

P.S.: another test could be to disable all notch filters and use the old low pass filter more agressive and come back to the standard main loop at 400, just to prove nothing change.

The issue with the logging is that generally you have a FRAM based log on these smaller FC’s and so they get large quite quickly. If you use mission planner it will give you all of the log options as a list of checkboxes. Ideally we would see nearly all plus fast or even raw IMU to see the gyro noise. Also since this may have some vibration impact setting INS_LOG_BAT_MASK = 255 would let us see overall frequency spread, but you wouldn’t be able to use this with IMU_RAW

Thanks! yes I’m using MP and in my micro 3" I have an sd card integrated, if I’ll solve this last issue I hope to have a little surprise to show you soon :slight_smile: . So is 255 one of the best settings to start logging?

start with LOG_BITMASK = 655358 that will give you all logging. Then try with IMU_RAW switched off and set INS_LOG_BAT_MASK.

https://ardupilot.org/copter/docs/parameters.html#log-bitmask-log-bitmask
https://ardupilot.org/copter/docs/parameters.html#ins-log-bat-mask-sensor-bitmask

Thanks… Good to know you had this issue on a 7" prop build… One of my theories was that something about the frequency of vibrations from the 3" props was unique and causing the issue.
And yes, I agree the omnibus nano’s seem better than the kakutes.

I doubt I could get the control zero to fit in a little 3" build… I’d pay $350 for a high quality FC that fit standard 20x20 mount pattern :). I’ve spent that much on these chap FC’s anyway, trying to find one that actually works properly.

One way I’ve worked around the issue is by hovering or flying for a minute for it to develop the lean. Then trim the copter so it’s level. Land, and save trim. (I have “Save Trim” as an option on my transmitter). It doesn’t “fix” it, but it will at least let it hover level, and maybe you can do an auto-tune.
When you first take off, it might pull to the side you have it trimmed to, but then it will develop the lean, and end up level.

FWIW I autotuned my 3" without issue (Kakute F7 Mini), but it was on 4S. Have yet to try the new build - the weather simply isn’t playing ball.

Reading the evolution of these boards its clear that the IMUs are very sensitive to electrical noise.

I’ve now tried the new build. It flies without issue and without leaning. In fact it flies incredibly well. Didn’t manage to get up to anything over about 50% throttle as it is incredible overpowered, but everything looks good. Here is an example log file.

1 Like

Thanks for a good log to review… I’ll compare to some of my bad ones and see if I can figure anything out.
EKF3 really does seem to have helped. I flew a lot yesterday w.out any problems.

I put my 2.4 VTX on one. I have no limit to video range, but lost a minute of flight time due to its size/weight. Not a great compromise when my flight time was only 6 minutes to begin with…

There’s a new F7 Mini with an MPU6000 - http://www.holybro.com/product/kakute-f7-mini/ might be worth a try

Same target as the current F7 Mini right Andy? That discussion was covered with the F7 V1.5 relative to the IMU.

Should be the same

more characters

Well… Good, but frustrating… Being on the cutting edge has a cost, I guess.