Went to go fly the Grinderino frame I just built (Swapping electronics from my other 3.5 inch, everything working fine before the swap), and it spazzed out - After checking the calibration, the image of the quad keeps spinning! Any ideas what could be the issue?

Video of issue: https://drive.google.com/file/d/1rRlY5bl...sp=sharing

If you haven't already done so, while it is doing that, go to the Receiver tab and check that the yaw channel bar is perfectly centred with a value of 1500. If it's not then depending on how far off the 1500 centre value it is, that could be the cause of the yaw creep.

(16-Dec-2023, 11:28 PM)SnowLeopardFPV Wrote: If you haven't already done so, while it is doing that, go to the Receiver tab and check that the yaw channel bar is perfectly centred with a value of 1500. If it's not then depending on how far off the 1500 centre value it is, that could be the cause of the yaw creep.

That was one of the first things I checked - Yaw, pitch and roll are all steady at 1500 when centered; I've also upgraded to the most recent BF, and also tried downgrading to 4.3, didn't make any difference - In almost 60 builds I've never had this issue before, and it's driving me nuts!

Did you try moving the board around, are the actual movements registered by the gyro? If you go into the BF motors tab you will see the actual gyro sensor values displayed there. The yaw is typically z-axis and should read near zero. If you spin your FC back and forth you should see the graph oscillate up and down. If you don't see this, maybe your gyro is broken. Wasn't there some gyro bricking issue on earlier versions of BF, don't know if that applies here.

(17-Dec-2023, 05:18 PM)mstc Wrote: Did you try moving the board around, are the actual movements registered by the gyro? If you go into the BF motors tab you will see the actual gyro sensor values displayed there. The yaw is typically z-axis and should read near zero. If you spin your FC back and forth you should see the graph oscillate up and down. If you don't see this, maybe your gyro is broken. Wasn't there some gyro bricking issue on earlier versions of BF, don't know if that applies here.

The gyro recognizes the actual movements - In BF the graphic moves exactly with the FC, but as soon as I stop moving the FC the graphic starts rotating without any external forces - I tried taking it for a flight, and as soon as it is armed it starts yawing

[EDIT]
Just checked the graphic in BF as I moved the FC around - It seems that I have pitch and yaw, but roll movements are not recognized - I'm assuming the FC is toast?

So what is the reading on the z-axis? If is not near zero then maybe this will help?
https://betaflight.com/docs/development/gyro-offset-yaw

Edit: Or maybe not, it seems this setting is only for headsfree mode...

Tried to get a better video:

Do you have a capacitor soldered to the battery pads?

(23-Dec-2023, 04:25 PM)Marcelo Wrote: Do you have a capacitor soldered to the battery pads?

Yes I do, 35A 1000uf Panasonic cap

Does this symptom persist in another area??, not near the Computer tower and all of its fields?   I would not have noticed it, but for the lighted cooling fans

Just a wild guess, but calibrate again when free of external influence.  I once had a magnetic direction indicator on a large quad, it had to be removed, but that FC had a magnetometer (compass)

It may be worth a quick test to set the board yaw orientation back to zero if you've changed it, or just test temporarily changing it to something else to see if it makes any difference.

I had a recent build that for some reason began to have a small yaw rotation on the screen similar to yours, after I changed the board orientation to match how I installed it.

On mine, the issue seems to have gone away after I installed a capacitor.

I just plugged in the quad again, and apparently on mine the issue is not 100% gone after all, it just got reduced. It still appears to drift on the yaw axis, but much less than in your video here. Mine wasn't noticeable at first because it changes about 1 degree every once in a while, sometimes 10 seconds or so. Mine is a Happymodel CrazyF405 with BMI270 gyro.

I'm going to test to see if gyro_offset_yaw makes any difference, even if it's intended for headfree mode, and also test calibrating outside somewhere farther from any interference as suggested.

I've also already checked using BF 4.3, and 4.2, both have the same slight drift on yaw.

Maybe not a huge issue for me as the quad is very much flyable, but it would be good to find a cause or fix.

I couldn't test outside yet since it's raining here lately.

I had inconsistent results with using gyro_offset_yaw.  I followed the instructions from the betaflight page about how to do this initial coarse setup, and at first it seemed to be helping.  However, after the next unplug of the battery and re-plug it went back to drifting on the yaw axis at a very slow rate for me.

So that made me think maybe it's something happening during the gyro calibration.  Since it happens when you first plug in the battery, I tried forcing a gyro calibration using the betaflight stick command (throttle down/yaw left, pitch stick down) after the quad was completely still on a table.

This actually seemed to make it almost stop drifting for me.  After 10 minutes of letting it sit still, my yaw axis heading had changed by only 4 degrees.  Before, I had gotten anywhere from 10-40 degrees of yaw drifting over a 10 minute period.

If you try this, you can do this while on the betaflight setup page that shows the image of the quad, and should see a brief message on the right indicating that it's calibrating.

I think I figured out what happened to the gyro on this board - I had it sitting on top of my work bench while I had my ultrasonic cleaner running, and I think the vibrations travelled through the top of the bench into the board, knocking the gyro out of whack - I've tried everything to get the gyro to stop drifting (Even brought the FC outside with my laptop to get away from all other electrical influences), and it still drifts - I'm thinking at this point I completely killed it