We were doing a demo for some water resources groups here in the SF north Bay area along the delta showing salinity gradients at specific GPS locations. The protocol was to take off in auto mode and loiter at 9 meters. From there we would enter guided mode where we navigate to a specific location where a "land" flight mode was evoked from the radio. We would sit in the water and collect data for a time before repeating again. We did this several times.
After lunch we wanted to repeat the experiment when we encountered the problem. The copter did not seem to be responding to radio commands. This should have been the moment we stopped the experiment and checked all the equipment. We went ahead. After launching in auto mode and achieving altitude, the copter took off towards the east. I immediately returned to "Stabilise" mode where I had a tenuous connection at best. I managed to get the craft heading back towards us when it seemed to loose throttle to the point of crash landing in the marsh. In about an acre of blackberry briers to be more exact.
I have looked through the files to see that there was not a goosd connection from the radio as the "roll in", "pitch in" and "throttle in" show. What I don't understand is why it took off towards the east while in a loiter mode and if there is any more info that could help to describe the radio failure.
We were able to connect to the MavLink several hundred yards away in the marsh and recovered the aircraft.Other than one of the ESCs becoming submerged in the landing and burning out the aircraft was no worse for wear.
Any help looking at these files would be appreciated.
Replies
It's also possible that the dock might have thrown off the compass. We've seen one report from a developer of this happening.
The best way to know is, if you remember which way the vehicle was facing when it took off, you could compare this to what the mission planner shows when you replay the tlog. The redline projecting from the front of the copter in the mission planner shows what it thought it's heading was.
Randy,
Thank you for taking a look and explaining what you saw. The dock may have had some part. It is aluminum and it seemed like the copter didn't want to arm at first as i mentioned. Maybe "bad compass health" pre arm check. There are also a number of power lines not too far east of us carrying a tremendous amount of power from a large ) wind farm (probably 20-30 turbines) to our west. We ran similar tests in a much calmer environment with the same copter today with new ESCs. The motors and props are just as they where when we found the copter after the crash. So i am pretty sure the motors and props were operating properly. Earlier in the day we ran a mapping mission of the area and found the 25 MPH gusts and accompanying variations in air pressure to cause extreme drops in altitude. Since I was in Stabilise at this point and giving it quite a bit of throttle, I had discounted this.
I will go through the logs again with the things you described to try to nail it down a bit more but I think the compass was likely the culprit here.
Thanks Again
This is probably a good time to ask this question. Would marine radar cause such issues? It operates in the spectrum we use for Tx. Recently, I was told about a film crew losing a craft while doing a job for a small cruise company while photographing their boat. They put it down to the radar on the ship. Is that something to be aware of? TIA.
Darrell,
I don't really know but I guess it's possible for interference if it's operating at the same frequency. Interference of the TX/RX signal is pretty rare but it might happen some times.
Sean,
Really sorry about your troubles. I've had a look at the logs and it looks like a few things went wrong.
The final crash looks like some kind of power failure. Not a brownout but perhaps a loose prop or a failing battery. You can see in the graph below from 1400 the altitude starts falling even though the throttle (both in and out) goes to maximum. The battery voltage looks ok (drops from 16.6V to 15.4V) but strangely the current also drops. I'd guess the battery or ESCs weren't able to supply the 50amps being requested while the vehicle was leaning over and requesting full power.
I think the reason it flew off in the wrong direction is because there was some kind of compass problem. It's not a GPS glitch because the number of satellites and hdop values seem fine while the vehicle was in Auto. I'm assuming you're using a Ublox?
The reason I think it's a compass issue is that the desired vel and actual velocity don't track well. I see that compassmot was done, I was wondering if you remembered what % interference it showed when that was configured?Looks like you're using an external compass and it's orientation is ok so that part looks ok.
The last issue is tuning. During AUTO it's attitude control is ok because it's not asking for large changes in attitude but when in stabilize mode it's control is pretty terrible. The copter's actual attitude diverges from the desired by 20degrees. The Rate Roll/Pitch P values look reasonable (0.13) but I don't know what this copter is like. Maybe they're too high, maybe too low, hard to be sure. It's also possible that some mechanical issue was causing it to not be able to maintain it's attitude properly.