Flying 2.023 Beta in X configuration - yaw issues - Discussions

I've really enjoyed my Arducopter flying NG RC2 since December and decided to take a plunge on the 2.xxx betas. I reconfigured my stock Arducopter into X configuration (APM pointed between arms) and loaded 2.023 Beta yesterday and flew several flights. Magnetometer, gps (1.6), sonar, installed as well as an xbee and had no issues with the configuration using the Mission Planner. (Did all steps including radio calibration, enabling sonar, magnetometer, and autocalibrated my stock ESC's. I enabled both stabilize and simple modes on my gear switch to start with and flew around first flight to auto calibrate the magnetometer in stabilize, but noticed right away drifting of the yaw axis to the right (clock wise) that I can counter with the stick. I thought it might be magnetometer cal issues, so I flew around the back yard trying to hit all headings to a full 360 for about 8 minutes. Sometimes the yaw locked nicely and sometimes after a small yaw to the right, it would just start a slow spin around in a clockwise circle again until I countered with full yaw to the left. I switched battery packs and after checking that the heading looked correct over the telemetry (which it did after the cal), I flew again with pretty much the same twitchiness in the yaw. I also tried simple mode and that works really well for me which confirms to me that the magnetometer is set right and in the right configuration. A little hard to get use to and trust, but it was working very well. In fact I could still cause the copter to spin with a small right yaw command, but it flew around nicely while spinning slowly in a clockwise direction.

I've got 3 logs from my 3 flights yesterday that I'm looking at now to understand what the software is doing (for a learning experience) as well as some static bench tests to understand what the log data is. Anyway, the Arducopter flys great in simple mode which I really like - is nice and stable otherwise, but just a little twitchy in yaw and requires some finesse to keep it stable in rotation. I'm flying a Futaba 12FGH radio with the 12 channel receiver and I'm wondering if the center position of the yaw channel (rudder) may not be exactly centered at 1.500 msecs. Anyway - since there is talk of a release soon, I wanted to give my experience with the latest.

Thanks to the team for all the hard work on the software development and I'm having a blast and learning a bunch.

Heino

Replies

Paul June 6, 2011 at 6:20am

jason, how would you tweak you motor angles? im just not sure as to why my would also seem to yaw naturally too, am i naive to think that the ac2 software brains is cleaver enough to correct itself? ive just got a standard jdrones setup with my rx mounted on the top to the left, so surely that wouldnt be throwing it out of balance?
Developer

Jason Short June 4, 2011 at 9:32pm

If you can post a log and point out exactly where the Yaw issue happened I would really appreciate it.

Jason
Paul June 4, 2011 at 7:22pm

hey heino,

i was always trying to fight the yaw (rudder stick) to keep it stable, as i can recall i think i was holding it a little to the left, and was trying it in both stable and simple modes, also trying out altitude hold but didnt seem to work??

after some little back and forwards action getting used to the quad (however a little annoying that wouldnt be stable and just going on its on) - the quad decided to topple over (literally, done a 360 flip) and crash! haha, broke a couple of motor mounts. im interested though as to why and how this happened? how can i download the logs to post here to possible help you guys identify what it was/is.
3D Robotics

Chris Anderson June 4, 2011 at 3:44pm

Thanks for the helpful test report. Some users are reporting similar yaw action. Although we don't see it on our own quads, it does seem to crop up with certain radio/airframe combination. Jason is looking into this, so look for an update to yaw in the next rev.

This reply was deleted.