Cope with negative ESC rpms in notch filter #28309
Conversation
|
@andyp1per The test flight that caused me to open the issue was done on Copter 4.5.6. Master has a different behavior: the logged filter frequency is negative for notches that are tracking negative RPM. I haven't read through the code enough yet to know what the filter is actually doing in this case. This is what it looks like on master: With your changes: |
|
It doesn't look quite right, does it. We have made some effort to properly log what values are actually being used. |
|
Actually, maybe it does look right. I'm just a little thrown by the weird symmetry of the ESC values - I wonder if there is something else going on there. Unless this is just split yaw. |
|
Yeah I think it is fine. The symmetry is because the drone was yawing at the time (uncommanded; bad tune because this was default tuning params). Here are the logs from the above screenshots. |
|
great! |
|
My preference would be to just apply the abs in the ESC RPM method. Just so that -1 for a static notch still results in off rather than 1hz. -1 makes no sense for many of the sources. |
|
IDK, this feels safer preventing negative numbers ever reaching the notches |
|
@IamPete1 i'm OK with this change, but will delay merge for you to expand on your comment |
|
discussed on the EU dev call, decided this is good for now |
No description provided.