I have Compass1 in GPS external and Compass2 in Pixhawk internal. They are recently calibrated.
Its normal to have yaw reset once, this is set up by parameter EK3_MAG_CAL 3 
Tnx Michal
Yes, you should delete them. Bad coefficients are worse than no coefficients.
Tnx concider it done
What about that external compass yaw? Not sure but most of my compasses is yaw 45 selected butit depends how was chip orientated on pcb. Could be 0 too.
I dont think 45deg will help here, I have also experienced to use angled compass location settings on other builds. This one is inside my GPS, and its a sealed box. I have also checked my compass headings with a true magnetic compass and north is north. Hope to go without a Pixhawk Powermodule so there will not be current readings. Do you think I can gain something by adjust some parameters since we know approx the current 20A on hoover and 30 on max ?
Maybe get one of those
Or use some pixhawk power module instead.
1 Like
Hi Dave, How to set this to 0 ?just do it or
COMPASS_MOTCT, 2 0:Disabled 1:Use Throttle 2:Use Current Set motor interference compensation type to disabled, throttle or current. Do not change manually.
After I input the new parameters from magfit I get lot of : PreArm: Compasses inconsistent
Then if I turn of the internal, COMPASS_USE2, 0 then no PreArm messages as above
If I recalibrate, I still got the PreArm messages and magfit settings will be lost.
Is it ok just to use the external or do you have some other things to test / check?
Regards Espen
Typically you do want to disable the internal compass. I have on most craft that have one. And yes the Magfit data will be overwritten with a compass cal you will have to repeat it. But the same log you used before is OK so no big deal.
Well, I did go for only external and then added old magfit. Did testrun, felt fine. Have to study log.
Log here: https://drive.google.com/file/d/106qARFJ6kJrZaK15IYsnTbfwx4hVdIOQ/view?usp=sharing
Parameters:
Tali Pixhawk1-1M AC V4.2.2 magfit compassmot adj fine 17.08.2022.param (15.4 KB)
I see that running magfit again on this log gives different results, is it worth to test the new values for mag(0) ?
That’s looking really good! I would say at this point fly the hell out of it 
1 Like
Tnx a lot! I will proceed by testing various FlightModes 
Tnx to Dave, Allister and Michal for helping me with this 
