I have two quads with the same problem and entirely different hardware.
F450, 2830's, Full size APM, 10x4, BEC, 3.15, X, baro covered, (ESC, Radio, Acel and Compass cal'ed)
Blackout mini H, 1806, min Apm, 5x3, BEC, 3.15.X, baro covered,(ESC, Radio, Acel and Compass cal'ed)
Logging full, but happens with all logs off.
In hover after random time period quad just falls suddenly.
Studying the logs looks to me that:
Vibrations within spec
hdop 1.8/8 at illustrated event
acel ok, compass ok
Approaching an event, throttle-in constant in hover, throttle out normal
Motor 1 is turned off
Motor 2 goes full
Quad flips.
Graph of one event attached and logs with multiple repeatable events.
Thanks for your time.
Replies
Mike,
One more thing is that it looks like the board voltage is extremely low and perhaps this led you to disable the arming check which would have caught this and stopped the vehicle from being armed? If 'yes' then please remember the arming checks are there for a good reason, they've been added as a result of previous problems reported by users and it's much better to resolve the problem than to disable the checks.
The board voltage is very uneven (varies by 0.4 volts!) and low (regularly drops to just above 3.8V). I suspect your vehicle doesn't use a power module, if that's so, I'd highly recommend getting one! We saw power related failures drop massively when 3DR started shipping all flight controllers with power modules.
I disabled the arming checks because the GPS failsafe stuff. I was only flying in stabilize so I got tired of it not arming. My Frsky X6R was degrading my GPS performance, once I got rid of it things improved but I did not get around to enabling the checks.
The vcc is odd and I have not had time to troubleshoot but the output of the bec is correct, unless it falls down under load. So I am not quite sure where in the chain, from the ubec to the fc power connector, the voltage is dropping. My fc is encased and buried in the frame. I'll have to tear it apart and create a test bed to be able to see where it is falling down. Possibly a broken wire inside the insulation not visible or a connector from the bec to the fc I suppose but pretty weird. I can shotgun and replace the ubec too.
I have done my homework. I have read every document 3DR publishes and thousands of posts, literally. Doesn't mean I am doing it right yet or understand everything I read but I have made a great effort.
On one flight frame, I bought a brand new F550 from DJI.
Its a put together kit, and has no BECS in the ESCS. Its now in the trash. Several big crashes, lost 2 video transmitters, PTZ cameras, gimbals... not nice.
I ordered my first board for my F550 from 3DR and it came with a defective 3.3v regulator. Common problem, just google it. I replaced the regulator while waiting for the new unit I ordered (at my cost) and initially thought the problems were the result of my fix. New unit, new GPS and the new set had a problem with the GPS. Everytime I had bad GPS health. This was BEFORE it even took flight, and yes I read up before connecting the whole thing. Then 3DR didn't ship me a new one because of some trouble they were in.... hmmm
After I finally got another GPS from another company I was surprised to find problems in all modes, and after many (about 6 boards myself) and several buddies having what we call "Ardu moments" gave up.
I didn't buy this controller on price and my first purchase was direct from 3DR and had the power module. It was a disaster.
I would try another controller, just to confirm. Its quicker to confirm the controller at fault. Hell a KK board would be better.
Mike,
It looks like a mechanical failure possibly of the back-left motor (because the vehicle falls left and backwards). Motor 2 (back left) going to full while motor 1 (front right) goes to minimum is also consistent with this (motor numbers vs position can be seen on this wiki page). What we see here is the vehicle flips back and left even though the output from the controller to the back-left motor is at full... the motor can't be doing what it's been told.
I guess I'd recommend ensuring that the ESCs are calibrated, connections are good, no sync issues.
Thank you very much for your time, I know it is valuable. I came to that conclusion early but expected to see a pre-failure mode change in the z axis as the motor failed to respond. But Everything that changes seems to happen after the motor responses by the fc. Where the fc is trying to compensate for the left rear falling. The baro graph confused me to the point i figured I just did not understand what I was looking at.
I have calibrated everything over and over. Changed out the fc, the gps, the receiver so that follows your conclusion of a motor problem or I suppose the frame;). But the motors are pretty simple devices. They work everytime the quad is armed. The other odd data is that the previous quad the F450 did the same thing and all of the hardware was different. So I would have to conclude an intermittent esc but two entirely different intermittent esc's on two different quads? I'll change them out, only thing left to do.
Please keep us informed, if you can get to the bottom of the problem, your doing better than me.
I did appreciate finding your post because it made me feel like I am not the only one with wierd problems.
It is really frustrating. Especially since troubleshooting can mean completely tearing a quad apart over and over. Or waiting for a days for it to happen again. Partly comes with the territory of really inexpensive parts but, of course, I would not be doing it if it was 10x the price.
I have spent 30 hours of troubleshooting this week. If I did not have a year in learning the APM flight controller I would have given up. I also like and support open-source and appreciate what they have accomplished. Like yours my friend did give up and now flies a SuperX. But these are hand made DIY projects and it is really hard to separate a user build error and code bug. Especially since it may only happen at the conjunction of many variables. Maybe that only happens on your quad.
Beats working crossword puzzles in retirement tho:)
Time to get the soldering station cranked up and get that arm disassembled and the ESC removed.
Cheers
Hi.
Yes, me and 3 other buddies gave up on Arducopter.
The programmers really need to stop developing new features and come back and fix the old problems.
It WILL crash, its just a matter of time. You can have 10 good flights, then bang, it'll flip and crash, or do something else stupid.
I have swapped out for a NAZA, and its perfect. Same everything, just new board +compass.
Everyone I've spoken to here agrees that Arducopter is rubbish, which is a shame as I support open source as much as possible and I like the fact I can run MP on my linux box.
I'd use your board on a ground vehicles only, and a slow one at that. Better that then hurting someone.
I think you can see that the fc is telling the motors to change. The motors are not failing and then you see reactions to that failure. The motors always start right up after a fall. I can let the motors spin on the ground until the battery discharges.
The problem is not a solder joint. There are no connectors as well, I always solder everything.
Thank you for your thoughts.
-
1
-
2
of 2 Next