Relative Pixhawk noob here as I think I have the right log file attached.. I flew my second flight with my newly built Tarot TL8X - Octocopter. During the first flight which lasted about 8 minutes, one of the ESCs failed and the octo hovered just fine. I had to double take that it was still just hovering without one prop running. After the ESC swap out and a fresh set of batteries I headed back out the next day. The second flight only lasted a few seconds. I'm still not sure why it happened. It took off, then a couple motors went balls to the walls and the octo entrenched itself in the turf. I suspect that there was some serious vibration, which set off the flight control in some way. I'm not sure though. I had the pixhawk mounted using double stick tape on foam (thin foam) You can see some vibes in the third segment of the video which is off the gimbal cam. All of the props were carefully balanced, however they are on prop adapters from Foxtech.
Here's the video:
https://www.youtube.com/watch?v=qtg-6hHHWpQ&feature=youtu.be
Here's the bird:
Frame: Tarot TL8X
Motors: Foxtech 4016
ESC: Foxtech 40a Multipal
Controller: Pixhawk PX4: I took off in loiter since I haven't really tuned the PID values yet. (that was to be next flight)
Radio: Taranis with X8R receiver
Gimbal: 3 axis Shaoyun Pro with Sony A6000
Props: Foxtech 15.5 supreme lite
Any ideas on what might have caused this? Fortunately the damage was somewhat minor, couple props cracked, one arm bent (good luck getting a new one), and some dirt buried in a motor.
Replies
Did you find a supplier for the arms. I found Tarot-RC has the arms. They come from China so it will take a while to arrive but at least it is a source for parts. Tarot X8 Metal Machine arm TL8X014.
Doug, I think there have been some similar situations that others are having. Check this comment out from Andrew Tridgel. http://diydrones.com/forum/topics/pixhawk-bad-accel-health-apm-3-2-.... Until this is figured out you may want to use an older firmware like 3.15. I may do the same thing. Btw how cold was it when your Tarot crashed?
Doug, Below are graphs of accX, accY and accZ for IMU1, IMU2 and both. IMU2 looks like correct. Had you calibrated your copter with the pixhawk installed? Did you wait before moving the copter after you powered up so the inertia sensors could center? Or you may have a bad IMU1. Your vibrations are actually very good according to IMU2. Finally I put up a graphs of my Hexacopter with a Pixhawk. If you notice the IMU values for each axis is close and not magnatutes or negative numbers apart.
IMU1
Both IMU's
Good IMU graph
To see if your IMU is bad I would do a bench test. Take your props off and arm your copter and give it a little throttle. Then download that dataflash and check both IMUs. If they don't match close I would contact 3DR if your Pixhawk is from them.
Thanks Rob for your detailed reply. I'll check it out on the bench and go from there.
Doug, sorry to here and watch your crash. I did look at your log and it appears the first IMU is way off. Not sure what caused it. If you don't get any other response to your posting you should repost it in the Pixhawk category here http://diydrones.com/forum/categories/px4-and-pixhawk-hardware-1/li....
Pixhawk has two IMUs, IMU and IMU2. What I did using mission planner is create a graph comparing IMU(acclX, acclY, acclZ) to IMU2(acclX, acclY, acclZ). IMU 2 gave results I would expect. IMU results where way off. Later when I get a chance I'll redo the graph and post a screen capture. Someone like Randy McKay will probably be able to tell you what happened.
From memory the vibrations on IMU2 did not look bad. I think it was something else.
Thanks Rob, looks like it moved over that way. With respect to the IMU, which line in the log are you looking at, and at what point is it way off? I noticed when I connected via mavlink just prior to the flight, that the computer took about three times the usual amount of time to connect, but everything else seemed to progress ok.