Replies to This Discussion

Permalink Reply by Vernon Barry on May 4, 2012 at 11:31am

More likely than not, your wiring is contributing to magenetic drift so the compass thinks the machine moved when in fact it didn't. There are a lot of threads about this but the biggest problem wires are the ones from the battery to the power distro since they are carrying serious amperage to power all the motors. The suggested fixes are twisting the wires or otherwise running them far away from the APM board so as to not create a strong magnetic field. The reason why it's slow is that the code primarily uses the gyros and acclerometers first, but then corrects them with the magnetometer, so when  you go to half power, that's enough to throw off the magnetometer which purposely corrects the assumed orientation slowly through each loop of the main software.

DeveloperPermalink Reply by Randy on May 4, 2012 at 6:55pm

Or if Maro is not using a compass at all then the yaw will naturally drift (because it doesn't have the compass to correct it) so he would see the above behaviour.

Permalink Reply by Vernon Barry on May 4, 2012 at 9:29pm

Possible, but then the drift is due to vibrations throwing off the gyros.

 

Possible, but again, I think my suggestion is more prevalent and with all the posts about magnetometer self calibration issues, fixes, and suggestions, I have a feeling this is related.

Permalink Reply by maroFX on May 5, 2012 at 1:09am

Compass is enabled and in use. The battery cables are situated as before and my quad could flight normally (sometimes I observed only very small motor breaks and peaks ). I only changed power system for camera servos and receiver (from other BEC).

I wonder if I correctly done whole sensors calibration and didn't skip anything. Auto leveling, sensor offset (imu test) are done but this time I didn't try with Live Calibration (manual frame rotation around). Is it necessery or not ?

Permalink Reply by maroFX on May 6, 2012 at 12:39am

1) Where can I find example of proper sensor calibration procedure , step by step ?

2) After many  "blind" attempts  all motors work correctly but I don't know exactly which procedure helped. Should I do any tests in Terminal Mode or it's now unnecessery for good calibration?