Repeatedly, my Pixhawk equipped Rover will stop responding. The telemetry link will go down, and RC commands will be ignored. It does NOT, however, go into HOLD mode - the servos stay at the positions they were in when the Pixhawk froze. As you can imagine, this is not great at anything other than zero throttle. I’ve seen it hang with the steering turned sideways, so I know it’s holding its position.
On a crash, the following happens:
The telemetry radio loses the red flashing light, but holds the green.
The Spektrum satellite receiver stays solid. If I turn off the tx, it flashes, then goes solid when I turn the tx back on. The vehicle doesn’t respond, but I’m confident the satellite receiver is reacting properly.
The FMU PWR indicator stays green, and the B/E error light is off. The big light on the Pixhawk stays green.
The I/O side shows the following:
PWR: Solid green
B/E: Flashing red (?4Hz?) - seemingly failure of FMU input?
ACT: Flashing blue
Here’s what it looks like: drive.google.com/file/d/0B9AHb7 … sp=sharing
The IO B/E light goes off after I power cycle the Pixhawk and everything comes back. Interestingly, the Pixhawk restarts already ARMED.
I’ve seen one or two similar reports of planes doing this, but the advantage I seem to have is that my Rover is already on the ground, so I don’t crash and I can get repeatable results with the same config.
I’ve checked the power supply logs and it’s consistenly above 5V.
I’m running a beta EKF code from 08 April with two GPS, set up on a post. I found the same problem occurred, albeit only once, with the 2.45 release code. By the end of today, it wouldn’t last more than about 5 minutes, even sitting still in MANUAL mode, without crashing.
There’s nothing that jumps out at me from the logs - everything seems stable right to the end of the log. The log appears to be truncating when the Pixhawk crashes as I’ve had one or two runs where the log ends showing a non-zero speed.
Suggestions? Ideas? Things to test?