Hello community, I am working on my personal UAV project in 10 Kg class Multirotor H frame configuration. Firmware version I am using is 4.3.5 on Pixhawk Cube orange plus. My UAV has Here3+ GPS and one Lightware SF45 lidar facing in the front. I was testing out object avoidance with Bendy Ruler + Dijkstra path planning alogirhtm in Auto mission mode and during this test my UAV crashed from around 9 meters height. This crash wasn’t due to any obstacle (Object avdoidance was working fine with Bendyruler Horizontal search configuration) but was in open space. I couldn’t get any reason for UAV just falling from sky midair. In logs I can cannot see any logs while UAV was crashing and log stopped exactly above the point of crash. So my initial guess is Pixhawk turned off midair since motors stopped spinning and pixhawk was off, batteries were also unplugged. Post crash we attached battery again and Pixhawk again turned on. If anyone can help me it would be really helpful. I have attached the logs as well.
@rishabsingh3003 Since you have been core developer for Avoidance part, I would really appreciate if you can look into this. Since this UAV has been tested to be flying fine multiple times without avoidance for 90 - 100 mins in single flight for multiple times in Autonomous missions.
It doesnt sound like it was anything to do with avoidance. It sounds more like main power faulure if everything just went off. Was the battery connector loose? Check the power supply to the pixhawk from the power module.
Does your rc controller have telemetry? Did it loose connection at the same time.
@geofrancis Thank you for your feedback. Our battery connector wasn’t loose, neither VCC to Pixhawk was going down. It shows in the graph to be consistently around 4.7 Volts. Our RC dosent have telemetry at all. Also we dont know weather RC lost the signal on crash point. We are unable to find any good evidence behind this free fall but this UAV is under test since many days and has been flying to its maximum potential nearly providing endurance of 1:30 hours.
Please let me know from logs if we can estimate anything.
post the log and we can look at it.
And update to ArduCopter 4.3.8 or 4.4.1-beta2 ASAP
@geofrancis Here’s the log file. Somehow I missed attaching the same. Please let me know your analysis.
@amilcarlucas Definately will update. Any specific issue you have noticed with 4.3.5 similar to my situation?
Log link : https://drive.google.com/file/d/1LLtSk-D3G9xruS3-4GPOE8zEYG6nGxze/view?usp=drive_link
@geofrancis Did you get a chance to check logs? Any help to boil down the problem would be highly appreciated.
I couldnt see anything specific, the only thing that looked a little unusual is your 5v VCC is a bit unstable and dropping down as far as 4.6v.
what do you have plugged into the pixhawk? are you powering your lidar from the flight controller serial port?
@geofrancis Thank you for your feedback. We have Mauch PDB supplying 5V supply to Pixhawk.
No we are not powering Lidar from serial port of Pixhawk and lidar gets power from a seperate Matek power supply of 5V.
What is the healthy VCC for pixhawk? Any range you can specify?
It should be at or very close to 5v, the input from the mauch should be around 5.2v, then it goes through a diode that drops it down a little.
Check the 5v from one of the serial ports with a multimeter to verify that it’s actually low.
@geofrancis Thanks. Will perform these checks.
What is the lowest accepted voltage for Pixhawk Cube orange Plus? Is there any possibility in our case Pixhawk Cube powered off midair causing the crash?
The way the logging just stops mid air I would say there is a high probability the flight controller reset mid air. If it was main battery voltage loss you usually still get a couple of log inputs before it goes offline as the capacitors will usually keep it going for a second or so.
4.8v from the pixhawk is the very lowest, but it shouldn’t be that low if you’re using a mauch power module. as it should be outputting 5.2v.