I’ve got a Flite Test Explorer running ArduPlane on an H7 (with analog servos, GPS/compass, telemetry radio, but no airspeed sensor) that I just maidened a few days ago. (The autopilot was very cool.) I’ve got some questions below, but I’m basically looking for any and all suggestions and tips to get it flying as good as it can.
I’m a reasonable RC pilot, have a half dozen scratchbuilt foamboard planes, but this is the first with any sort of stabilization.
The dataflash logs are available at Dropbox - Ardupilot - Simplify your life
The first thing I noticed was that it had a tendency to turn left that didn’t seem to be dealt with by the SERVO_AUTO_TRIM adjustment. I see the automatic trim docs don’t mention yaw, is it not included? Do I need to recalibrate my compass? Plotting ATT.Roll, ATT.DesYaw and ATT.Yaw shows this drift clearly at around 19:06:01.
I did a couple of autotune sessions at level 6 in FBWA, but even after that when I compared it with MANUAL mode it was much less maneuverable. Are there any benefits to bumping up AUTOTUNE_LEVEL to 7 or 8 since the plane is physically capable of more aggressive maneuvers?
The autotune docs talk about graphing ATRP.Demanded and ATRP.Achieved but I can’t figure out where those are in the Log Browser in Mission Planner. Do I need to turn on more logging in the config or am I just not looking in the right place?
On the first flight at around 19:10 I noticed the tail of the plain wagging up and down, and looking at the logs I think I see pitch oscillations around 19:08:13. What do I need to adjust to get rid of those?
I’m also curious about my TRIM_PITCH_CD value. In the first flight there are minimal RC inputs between 19:05:20 and 19:06:20, and it seems like the angle of attack is more like 3.5 degrees. Should I bump TRIM_PITCH_CD slightly?
I’m curious, can I tell anything about CG from these logs?
One final question, I set up for flaperons as per the docs with one channel left as “aileron” and two channels as “flaperon left” and “flaperon right”. Plotting RCOU.C2 (aileron) it seems like it only shows half the range of motion, though RCOU.C6 and RCOU.C7 seem okay. Are the weird RCOU.C2 values just an artifact of internal calculations or something?