log help for issue with loiter

I believe you'll find that if you select 'alt-hold' mode, you'll get the same effect. I would suggest holding on to your quad, engaging the motor, and tighten your grip, then change to the alt-hold flight mode, and I suspect it'll throttle up and try to pull up and away from you.

I have been told by people here that is caused by 'too much vibrations' on the APM, and that the APM needs to be isolated. I haven't dug through the 2.9.1x code yet to determine why excessive, presumably cyclical (no directional bias), noise in the Z axis acceleration would result in the craft thinking it needed to climb. But I'm told it does right now.

Current diagnostic ideas include during on 'raw' mode and dump the logs after landing and specifically look at the z acceleration when you have the craft reliably hovering in stablize mode (not alt-hold, but when you're moderating the throttle yourself to maintain hover).

Replies

T3

Richard Boyhan May 12, 2013 at 6:20pm

zAccel is primary sensor used for Alt Hold. Alt Hold is the second major part of loiter. The other major part is GPS for position hold. Don't start with a good GPS lock you get a fly away. zAccel measures your z axis or up and down movement. Bad vibration messes up zAccel. You need values between -5 and -15. -10 is basic gravity pulling us down. Balance props and vibration damping on the corners of your APM help. I also recommend a strap over damping material on top of the APM holding it down.
The next release of code will use x and y Accel for a much better position hold. It's good practice to check all three Accel values and make sure vibration is minimal.
Good luck and fly safe.
James Small May 12, 2013 at 9:13am

I believe you'll find that if you select 'alt-hold' mode, you'll get the same effect. I would suggest holding on to your quad, engaging the motor, and tighten your grip, then change to the alt-hold flight mode, and I suspect it'll throttle up and try to pull up and away from you.

I have been told by people here that is caused by 'too much vibrations' on the APM, and that the APM needs to be isolated. I haven't dug through the 2.9.1x code yet to determine why excessive, presumably cyclical (no directional bias), noise in the Z axis acceleration would result in the craft thinking it needed to climb. But I'm told it does right now.

Current diagnostic ideas include during on 'raw' mode and dump the logs after landing and specifically look at the z acceleration when you have the craft reliably hovering in stablize mode (not alt-hold, but when you're moderating the throttle yourself to maintain hover).

This reply was deleted.

Discussions

log help for issue with loiter

Replies

Activity