Maiden flight = almost instant ESC burnout

[hogster]

Hi axelneid,

Thanks for your reply

I was literally looking at the video again just now, and I'm wondering whether the props are just starting too quickly ... does anyone know if the 'starting PWM' defines the starting speed of the motors? Eg. does it go from motors-off to 1125us instantly?

Re: the motors, they're all T-Motors and feel very smooth to spin, so I'm not sure they're at fault. My thoughts keep going back to the ESCs trying to start them too abruptly, and failing as a result ...

Many thanks, David

[Max]

[JussiH]

I would say that based on watching the video, there is something wrong with the range of the esc´s - some of them looks to start later than the others, or is not starting at all. It could also be that the MOT_START value is too low.

I am totally not up to speed on SimonK based ESC´s, but I do believe that they still support stick calibration.

But you need to make sure that you are calibrating them to the right range. So you need to set up your throttle channel to have a 1000-1950 uS range, and then connect escs to the throttle channel and calibrate them.

I have a servo tester that allows me to define top and bottom pulse width, but if you dont, you can use multiwii on a arduino to tune the throttle channel to the exact range you need before calibrating.

Then to determine min PWM widt to start the motors, hook up one ESC (the one that seems to start late) and the Multiwii on a Y-cable to see the actual PWM width that starts the motor. Use this value as MOT_START in AQ.

Test the individual motors on a servo tester to see if any of them has startup problems or run suspiciously.

I would look into a BL-Heli ESC. There must be a 6S capable Silabs based ESC out there, and with BL-Heli suite its a breeze to flash and configure them to match your motors exactly.

http://www.rcgroups.com/forums/showthread.php?t=2136895

[LPR]

David

I also had a problem with ESC getting warm without even running the motors.
You might have the same problem. On the AQ board you can set the middle pin on the headers for the ESC signal wires so it will give 5 volts out. This feature is meant to power the ESC32. I forgot to look and see the bridge was soldered when the headers were solder on. After I noticed the ESC were warm I looked and found that they were soldered on my new AQ with DIMU. Most ESCs have a bec to power servos and when I connected a standard ESC to an AQ board that is also producing 5volt out I found the ESC getting hot.

To fix the heating problem all I did is remove the center or red wire from the ESC signal wire plug and then tape the wire back to the signals wires. You could just cut off the red wire but I wanted to be able to use the ESC for other aircraft.

Larry

[hogster]

Many thanks for the replies everyone

I'm going to get hold of a servo tester so I can more accurately see what my AQ is outputting, and also how I setup those ESCs.

In terms of replacement ESCs I'm in a bit of a quandary - there are so many "dedicated" multirotor ESCs like the DYS40, the HobbyKing Afro ESCs, that I'm baffled which to go for! I'm still considering these:

http://www.electricwingman.com/tiger-mo ... simonk-esc

if I can find somewhere with stock.

Jussi - I have dabbled with flashing ESCs in the past and had great success with it (flashing Turnigy F-30A ESCs if i remember correctly). Is there a database showing ESCs compatible with this BL-Heli code?

Larry - The ESCs I'm currently using are opto-isolated, so they needed the 5V line connected to them

Many thanks,

David

[joebar.rc]

[hogster]

Hi Menno,

Thanks for the guidance

I've been having a closer look at the X8 this evening.

Firstly, I think the ESC burnout has one of two causes. As the motor was running quite freely up until the point of shorting, I can only assume that either
– Debris entered the motor during spool-up, cut through some of the windings and caused the short
– Somehow the ESC was driving the motor incorrectly, causing it to stall and draw a large current

I thought I spotted some burned windings in the motor which abruptly stopped running (see 1m05s in the - password: hogstercortex), but I'm not so sure now. The motor doesn't smell at all of burned wire enamel, and it certainly would if the windings really had shorted / burned!

So I'm now wondering how an ESC which was working initially can sudden short out and release its magic gas. Either the ESCs are just poor quality (this is the second ESC I've had short out – the first was being powered by a PSU at the time so no smoke was released - the motor wasn't even running when it shorted out!), or the AQ is somehow driving the ESCs incorrectly. I've read in the AQ documentation that the PWM motor outputs update at 400Hz by default – is that correct? I've asked the X8 manufacturer if his ESCs are 400Hz compatible – if not do you think this might be the cause of them failing? The ones on his site are shown as 400Hz, but he can no longer find stock of the previous ESC (the ones I'm using) ... will see what he says.

On a similar note, I've been looking at the log files for the "flights", and I'm puzzled why the motor demands are so horrendously noisy. If you look at the log file itself (), motor outputs (LOG_MOT_MOTOR) 0, 1, 3, 4, 7, 8, 11, and 12 are showing activity, whereas in the 'Mixing & Output' tab, I've only selected port numbers: 1, 2, 4, 5, 8, 9, 12 and 13. Could this be a sign of a flashing / programming error? If you look at the video again, the front left (rear right from the camera's point of view) motors seem to be much more reluctant to start than the other motors and I can only think it's a motor port mapping issue.

As always, your advice is much appreciated.

Many thanks, David

[Max]

[hogster]

Thanks for the confirmation Max I was quite tired last night when I wrote that and I didn't realise that both sets of numbers were exactly one unit apart!

However something is still not right with the motor outputs. I've just had the AQ and Rx powered up and I was trying to watch the motor demands in the MAVLink Data Plot tab. Firstly there only seems to be 8 'servo_raw' channels, so I couldn't watch all the motor demands. Secondly, after arming and giving some throttle, only channels 1, 4 and 5 were responding immediately to my throttle stick movements. The mean values of channels 2 and 8 were rising much more slowly - shouldn't they all rise / fall together when giving large changes in throttle demand?

I've tried to use the AQ Diagnostic Telemetry tab (which has 14 motor outputs), but every time I've pressed 'Start Telemetry', nothing happens except the words "FAILURE: Temporarily rejected CMD: 2" appear right at the top of the screen next to the battery voltage readout bar. Is there something I have to do differently to use this live Telemetry feed?

Many thanks for your help guys, it's very much appreciated

David

[Max]