MikroKopter motor failure and fire

Discussion in 'Electronics' started by Charlie Cushing, Nov 4, 2014.

  1. Charlie Cushing

    Joined:
    Jan 22, 2014
    Messages:
    56
    Likes Received:
    2
    Was flying this past weekend with my Cinestar 8, T Motors, MikroKopter electronics version 2.04, MOVI 5 with a Canon T3i in 46 degree weather with a 10-15MPH wind in GPS HOLD.

    I had one 10 minute flight and everything was going well, I landed, changed batteries (6S 10,000mAh) and went up again immediately, I started getting a motor error warning while hovering in one place after about 1 to 2 minutes. I was between 1/4 and 1/2 throttle when this error started. I then lost a motor and decided to immediately land where I was.

    I was able to gently land the chopper and ran over to it and it was smoking, got the battery unplugged. The pictures below will illustrate what happened. Somehow the solder on one of the motors got too hot to the point the solder blew right off of the MK Board and my best guess is it then started an ark somewhere inside that started the fire. Good news is the physical chopper appears to be in perfect condition, the electronics, do not appear savable at all, burned up pretty bad.

    There are pictures and GPX files below, if anyone can take a look and tell me what they see it would be greatly appreciated. I was monitoring voltages during the flight and I thought everything was normal until this motor error started. Thanks to anyone that can help me out! -Charlie IMG_0106.JPG
    Chopper fire1.jpg
     

    Attached Files:

  2. Steve Maller

    Steve Maller UAV Grief Counselor

    Joined:
    Oct 30, 2012
    Messages:
    3,979
    Likes Received:
    807
    That's awful! Was there a short on the main lipo leads? I am mobile and can't look at the gpx files but maybe somebody else can. Thank goodness for telemetry or you could have lost everything!
     
  3. Charlie Cushing

    Joined:
    Jan 22, 2014
    Messages:
    56
    Likes Received:
    2
    No short on the main leads that I can see, when that motor quit my best guess is the leads for the motor number 3 blew off and that's where the short came from that started the fire.
     
  4. Steve Maller

    Steve Maller UAV Grief Counselor

    Joined:
    Oct 30, 2012
    Messages:
    3,979
    Likes Received:
    807
    Hmm...the motor leads wouldn't cause a short if they came off...they're on the wrong side to do that. It seems like maybe you had a catastrophic failure of the PDB that caused a cascading failure on the board itself. That's disturbing. I'm getting more and more wary of the all-in-one design of the MK PDB and BLs. Engineering is good, but it's engineered to succeed, not to fail. It should fail better than it does.
     
  5. Steve Maller

    Steve Maller UAV Grief Counselor

    Joined:
    Oct 30, 2012
    Messages:
    3,979
    Likes Received:
    807
    I just took a quick look at your GPX files from before and after the crash, and there are a couple of question marks:
    1. your motor currents are all over the place, which seems wrong. There's a 250%+ variation, so maybe something was wrong with your config in the first place
    2. the flight before the crash had a number of "Motor Restart" errors. this shouldn't happen. that should have grounded you.
    I wonder if one or more of your motors went bad.
    Sorry, don't have much time to dig any deeper than that right now.
     
  6. Andy Johnson-Laird

    Andy Johnson-Laird Administrator
    Staff Member

    Joined:
    Jul 31, 2012
    Messages:
    10,369
    Likes Received:
    1,162

    Steve: Why would the motor leads not cause a short if they came off? They would be shorting out the output from the MOSFETs which are providing quite a lot of current to the motors?

    Charlie: I'll have a look at the GPX file later today -- not that I doubt what Steve found, but more to see if there's anything else that might be wrong.

    Can you post any GPX files from your previous day's flying? Just want to see what the track record is.

    Andy.
     
  7. Steve Maller

    Steve Maller UAV Grief Counselor

    Joined:
    Oct 30, 2012
    Messages:
    3,979
    Likes Received:
    807
    The current flowing off the pads to the motor wires would cease to go anywhere, right? Or do we think he was using bullets which came loose? I think the motor errors are an indication of a lurking problem that eventually cascaded. I had motor errors at one point, and it was a combination of a bad firmware update, incorrect motor timing, and a couple of marginal bullets.

    I bet the motor wires coming loose happened after the crash because of the heat. But I could be wrong.
     
  8. Andy Johnson-Laird

    Andy Johnson-Laird Administrator
    Staff Member

    Joined:
    Jul 31, 2012
    Messages:
    10,369
    Likes Received:
    1,162

    If one motor wire partially unsolders and moves sideways it could bridge over to the next MOSFET pad. Also, bear in mind that there are three sets of coils, and only three wires, so each wire is actually connected to two coil sets -- it has to be so that, depending on which pair of wires is energized, it can energize the corresponding coil set.

    So one wire breaking free could cause Very Bad Things to happen. Check out the scorch marks on the center MOSFET pad on the edge nearest the camera in the second image above.

    My theory would be the other way around....one solder joint starts to develop high resistance, which causes a thermal runaway (as in the heat increases the resistance which increases the heat....) until one wire comes loose. The motor restart errors occur as a result because one coil set in the motor is not working right, so the Brushless Controller fails to detect the zero crossing point and cannot commutate the motor correctly, so it calls a motor restart error and tries to get back into sync with the motor....

    Charlie: Were you motor wires soldered directly to the power distribution board pads, please?

    Thanks
    Andy.
     
  9. Andy Johnson-Laird

    Andy Johnson-Laird Administrator
    Staff Member

    Joined:
    Jul 31, 2012
    Messages:
    10,369
    Likes Received:
    1,162
    Charlie:
    I'm looking at file 14110207.GPX (the one for the crash and burn) and what I'm seeing is:
    Errors / warnings:
    I2C errors: 001
    Error "ERR: FC I2C" (17) occured 2 times!
    Error "ERR: Bl Missing" (18) occured 6 times!
    Error "Motor restart" (23) occured 36 times!
    Error "compass not calibrated" (32) occured 2 times!
    FC-Flag "LowBat" (5) occured 9 times!

    That first Motor Restart, occurs at row 82, so, given your log file is written at one row every 500 milliseconds, that's 41 seconds into the flight. From the motor currents, it looks like it might be motor #7. According to the log file, if you have audio, you should have heard the Motor Restart warning.

    The second Motor Restart, which seems to be the precursor to all the other problems, doesn't happen until row 146 of the data file (the log file is being written one row per 500 milliseconds, so you can calculate that that is 73 second into the flight).

    The GPX file from the preceding flight, file 14110206, also has:

    Errors / warnings:
    Error "Motor restart" (23) occured 18 times!

    These occurred at row 656 of the flight, 328 seconds into the flight. All 18 occurrences are a contiguous block....the flight continued for 15 seconds after the last occurrences. It also looks like it might have been motor #7 causing the problem.

    And finally, the flight before *that* (the first of the three), was error free.

    I realize the above doesn't tell you the cause(s) of the problem, but that's what the evidence says, as far as I can see with my limited analysis.

    Andy
     
  10. Andy Johnson-Laird

    Andy Johnson-Laird Administrator
    Staff Member

    Joined:
    Jul 31, 2012
    Messages:
    10,369
    Likes Received:
    1,162
    Charlie:
    I was looking closely at the first image that you posted.

    Could you visually inspect a couple of areas more closely?
    In the image below, where the red arrow is pointing, it looks like the outer insulation has been abraded off and there might be wire strands visible. The margin around the wear area is smooth and doesn't look like it is the result of melting. I'm curious therefore as to whether the wire, when it was connected to the Double Quadro, might have been in contact with the heat sink or some carbon fiber.

    Same question for the green arrow. Is that wire strands that I'm seeing? What might that wire have been in contact with, assuming those really are wire strands.

    Thanks
    Andy

    Charlie Cushing 2014-11-04.png
     
  11. Charlie Cushing

    Joined:
    Jan 22, 2014
    Messages:
    56
    Likes Received:
    2
    Andy, the motor wires had wire with bullet connectors soldered to the main board and then connected to the motors. It' s a QC build.

    In regards to the wires being slightly worn, I looked closer and washed it off, there is no bare wires showing, it was slightly melted and discolored from the fire, but no bare wires that I can find. Here are a few more pictures and GPX files from the previous day.
    6.jpg 1.jpg 2.jpg 3.jpg 4.jpg 5.jpg
     

    Attached Files:

  12. Andy Johnson-Laird

    Andy Johnson-Laird Administrator
    Staff Member

    Joined:
    Jul 31, 2012
    Messages:
    10,369
    Likes Received:
    1,162
    Thanks for the information Charlie.

    Do you happen to use Graupner Tx/Rx with audio? That motor restart on the flight before the crash would have caused shivers up my spine and damp underwear ;) and I would have grounded the aircraft until I found the cause.

    Also, the firmware you were using for the Flight Control and Navigation Controller boards are old versions.
    The first GPX file shows: FC HW:2.5 SW:2.04a + NC HW:2.0 SW:2.04a

    The current version for the FC is 2.08.

    I have also been in email contact with MK and I believe the BL-Ctrls on your Double Quadro board were running version 1.02 firmware -- which was known to have problems with motor restarts and voltage spikes with certain motors. Can you confirm what version of firmware you were running on the BL-Ctrls, please? I may have got that wrong.

    For those of you you have not upgraded, yeah, it's a bit of a fiddly process as you have to update each one of the eight Bl-Ctrl's separately, but the BL-Ctrls must be upgraded to the current version 1.08.

    Charlie: I'm mindful that this info is too late to be of much use to you -- and for that I apologize, but for others flying similar MK boards it is highly relevant.

    Andy.
     
  13. MIke Magee

    MIke Magee Active Member

    Joined:
    Jun 29, 2012
    Messages:
    422
    Likes Received:
    103
    Andy, Not that **I** haven't upgraded my BL's - I'm asking for a friend here... Do you have the url for the excellent instructions that I've seen posed here handy? I have My friend looked for them and couldn't find them, and believes that now would be a good time to do it.

    tx,m


     
  14. Charlie Cushing

    Joined:
    Jan 22, 2014
    Messages:
    56
    Likes Received:
    2
    Yes, I have a Graupner MX-20, and I was monitoring the voice commands for voltages, etc. I've owned this chopper for roughly 6 months so I'm somewhat new to how MK electronics work and had never received that error before so I really wasn't sure how serious or what it meant. I was aware that I was two generations behind on firmware updates, but the advice I've received was don't rush to update the firmware until the bugs are worked out. It was flying well prior to this happening, so I wasn't in a hurry to do it right before a shoot. I'm not sure what I was running on the BL-Ctrls, where can I find that information?

    The good news is my frame is intact with virtually no damage, but the electronics are garbage, the next question is if I clean up the Flight Control board, which outside of smoke discoloration appears to be undamaged is that still operational? The GPS should be ok as its on the boom away from where the fire was, but I ultimately won't know until I install the new parts.

    Needless to say this whole thing is ridiculously frustrating.....
     
  15. Andy Johnson-Laird

    Andy Johnson-Laird Administrator
    Staff Member

    Joined:
    Jul 31, 2012
    Messages:
    10,369
    Likes Received:
    1,162
    Cute, Mike. Very cute!
    Please tell yourself your friend that I will try and find them for him.
     
  16. Andy Johnson-Laird

    Andy Johnson-Laird Administrator
    Staff Member

    Joined:
    Jul 31, 2012
    Messages:
    10,369
    Likes Received:
    1,162
    MIke Magee likes this.
  17. Andy Johnson-Laird

    Andy Johnson-Laird Administrator
    Staff Member

    Joined:
    Jul 31, 2012
    Messages:
    10,369
    Likes Received:
    1,162
    I hear ya, Charlie. Trust me. There is that fine balance between updating too soon and updating too late. (The saying in the USA is that the "early settlers got all the arrows" :) ).

    I would wash the FC board in isopropyl alcohol -- use a soft paint brush to remove the soot and debris -- you may need to do this a couple of times to get it clean. Then connect it up to your PC -- just free-standing with the MK USB adapter plugged directly into it -- nothing else connected to the FC board - the FC board can be just sitting of a piece of non-conductive cardboard!
    Install the small jumper on the USB adapter to provide +5v to the FC board and when you plug the USB cable into the PC, the FC should power up.

    Then start up MK tool and see if you can establish contact with the board and check if it seems to be working correctly, sensing motion correctly (use the 3D "simulator") and otherwise performing OK (e.g. can you read all the settings).

    If it all checks out OK then I would try a careful test flight with landing gear on the copter only - no gimbal and hover only three feet above ground -- preferably soft ground too!

    And you're correct, flying copters can/is ridiculously frustrating -- they are complex systems that marry a bevy of computers and aircraft that haul themselves into the air by brute force....

    Andy.
     
  18. Charlie Cushing

    Joined:
    Jan 22, 2014
    Messages:
    56
    Likes Received:
    2
    Going through this again for everyone. I was hovering and motor number 2 stopped which is the one that blew the solder right off of the board. When I landed is when the fire started from what I could see several hundred feet away. The only testing I've done on my motors so far is smelling them to see if they're burned up and none of them have the smell of burned electronics, what is the best way to test these to ensure they are all good without a board?

    What exactly is a MOTOR RESTART ERROR? It's not something I've seen in flight, or it just an electronic log?

    Steve, was the 250% motor variation on every flight, or just the last two and what does that mean? Wind, pilot, compass, electronic error?

    Quadrocopter has also been analyzing the data and they sent it off to MK and this was their analysis.

    "I heard back from the Germans, and they did not point out anything new. They did mention that you were running an older version of firmware on the BL's, and that a newer version would have been more helpful in diagnosing the problem. They also mentioned that you should do an inspection and diagnostic after a error warning.
    I also think this might have been something that could have happened because of something being loose or a bridged circuit. The fact that it happened quickly, also leads me in that direction."
    I have reviewed the files carefully, and have come to the conclusion that it was not your fault or the fault of the electronics, but from an unknown factor that could have bridged something on the board. "

    To all that have helped look at this data and analyze it, THANK YOU! Hopefully MK and other pilots can learn from what happened to me. -Charlie
     
  19. Andy Johnson-Laird

    Andy Johnson-Laird Administrator
    Staff Member

    Joined:
    Jul 31, 2012
    Messages:
    10,369
    Likes Received:
    1,162
    Charlie: The BL-Ctrls generate three phase power to make the motors spin in a particular direction and at the correct speeds. To do that involves a delicate electronic dance where two of three coils sets in the motor are receiving power and the third set is acting as a magnetic sensor, sensing the angular position of the motor as the magnets in the outer motor bell whiz by the third set of coils.

    Of course, as the motor rotates, which particular two coil sets are energized and which coil set is acting as a sensor changes rapidly -- this is called commutation. Occasionally, for particular motors, the timing gets out of kilter -- and the motor is said to have "decommutated." With many electronic speed controls, when that happens, the motor slows down and/or stops.

    With MK's BL-Ctrl V3, they contain a proprietary algorithm to sense the decommutation and quickly re-synchronize the BL-Ctrl to the motor so that the motor does not slow down or stop.

    You may find this table of interest: http://wiki.mikrokopter.de/en/ErrorCodes

    A good technique, when you have a Mikrokopter question is to do a Google search of the form:

    mikrokopter <insert your question here>​

    Putting mikrokopter as the first word "qualifies" your question so you'll get mikrokopter-related results first.

    So to find the error codes I did:

    mikrokopter error codes​

    If you *really* want to limit the search then do this:

    error codes site: mikrokopter.de​

    Note: the colon after the word site: -- to Google this means "confine your search just to this web site."

    Hope this helps
    Andy.
     
  20. Dave King

    Dave King Well-Known Member

    Joined:
    Dec 24, 2012
    Messages:
    2,711
    Likes Received:
    311
    wow, not sure how I missed this originally. I think this is the worst failure I have ever seen, even more than my board failure that caused a fire.
     

Share This Page