Troubleshooting motors/ESC/Revolution
« on: September 16, 2019, 06:07:48 pm »
I hope someone can help!

Quad was running fine, all the motors spin.  I hadn't run it for a couple of weeks but came back to start it up and had a few problem.  I had it powered by a battery and was messing with the GC software.  I was connected wirelessly to the drone but not actually doing anything.  Smelled smoke and look over and the ESC on position 1 is sparking and flaming!  Unplug battery and everything stops.  I had a spare ESC and swapped it in.  Went back through hardware setup and had the following issues:
1. Motor 1 will not spin. Nothing. When it first powered it smoked a little.
2. ESC 1 gets warm. Not hot.
3. Motor 2 just freaks out, moving back and forth, with not spin.
4. Motor 3 spins normally.
5. Motor 4 freaks out like motor 2.
6. GPS no longer lights up and seeks a signal.  As far as I can tell it is dead.

I replaced motor 1. No spin.  Connected that motor to position 3 ESC and it spun. Connected back to position 1 ESC, nothing.  Replaced all Simok ESCs.  Still same behavior.

At this point I suspect that something is wrong with the Revolution FC.  Any idea how I would trouble shoot that? Any idea of what could have gone wrong in the first place?

Ed

Re: Troubleshooting motors/ESC/Revolution
« Reply #1 on: September 16, 2019, 07:56:21 pm »
If either the motor or ESC is bad or of bad design it can cause them both to burn.  Flashing the wrong ESC firmware can also do this.

Sometimes both will burn at the same time.  Swapping parts around without using a current limited power supply or battery with a light bulb in series can cause you to burn more good components.

One symptom of a partially burned motor (still unusable) or motor with bad connector (or even a partially burned out ESC) is that it just jumps back and forth and will not start, but ESC/motor incompatibilities will make it hard starting too.  If you see this, most people will try to give it more power to get it to start, but that can turn an incompatibility into a burned motor or ESC.

99.9% of the time, if you see smoke, that part (at least) is ruined.

Once you have a bad motor or ESC, you must be very careful to avoid causing more damage to good components by swapping stuff around and burning more things out.  To safely test ESCs and motors insert a large 12V automotive bulb such as #1156 in series with your battery to power the system so that even if the ESC shorts out, that the light bulb glows then and keeps max current from getting very high.  You can even test motors for continuity with this.  You should be able to start and run motors at slow speed (faster without props).  When you see the bulb start to glow you are getting close to the fastest it will run without resetting.

to esc (-) <------------- (-) battery (+) ----- bulb ----------> to esc (+)

If you have very tiny motors (max current less than 2 amps), you may want to find a bulb with a smaller current draw for testing.  High current bulb for really big motors.

You can test a motor directly this way:  Unplug it from the ESC and use an Ohm meter set on lowest scale (like x1 ohm) (or a light bulb and battery) to test continuity.  Each of the 3 leads should appear directly connected to the other two.  If you find any with no connection, mark that motor as bad and keep it for parts.  If motor wires are labeled as A, B, and C, then you only need to test A-B, A-C, and B-C.

Due to ESC manufacturing tolerances, the BECs inside the ESCs produce slightly different voltages.  Like 5.05V, 5.09V, 5.10V and 5.15V.  In this case the 5.15V will be supplying all the power (for FC, receiver, and servos) and gets a little warmer than the others.  This is normal.

GPS will only light up when it is plugged into the FC and the FC is powered with a battery via the ESC/BECs.  Also, unplug it and use a magnifying glass to look at the FC socket where it is plugged in.  Look at the tiny pins.  Are any bent?  Be aware that what ever voltage you power the FC with, it directly connects that to power the ports.  The FC may be able to accept higher voltage than 5V or 6V, but some of the things plugged in the FC may not be able to handle it.  You should always power the FC with 5V or 6V and never connect the flight battery directly to the FC.

It should be impossible for the FC to make motors or ESC's burn out.  The FC only tells the ESC how fast to run.  Use PwmSync or PWM@490Hz for your ESC protocol (as set in the FC on the Output page) until you get it all working and actually have a real reason to change it.  Or unless you know for absolute sure that your ESCs will not run these protocols.

https://forum.librepilot.org/index.php?topic=2152.msg17429;topicseen#msg17429
https://forum.librepilot.org/index.php?topic=3434.msg23481;topicseen#msg23481
https://forum.librepilot.org/index.php?topic=3795.msg25971#msg25971
https://forum.librepilot.org/index.php?topic=3060.msg21284#msg21284
https://forum.librepilot.org/index.php?topic=4533.msg30874;topicseen#msg30874

and a general search for ESC and bulb which turned up these other posts :)
https://forum.librepilot.org/index.php?action=search2;params=eJwtjMsOgjAURH_FuHEzC-6lCHxN01cEA9S0RWPSj_di2M05Mxnj32ZzwddbpXqtNv0TGC0UOtzRY8AIakAEYlALUqAONIBGcAOWNYNbsAL34EFu8hQ_2sX1tYQSzuu822dwRcdt-Yo6TExF-zkJ-ZDdaYRmr9f8ODiY5CYxUl_svtgfe_o4Rg..

Re: Troubleshooting motors/ESC/Revolution
« Reply #2 on: September 17, 2019, 03:44:28 am »
OK...update

Checked that each ESC, gforce K-Series 30A ESC BEC:3A@5V, had 12v to them.
All motors have continuity between all leads. 
I took a know good motor (runs) and hooked to each ESC
Position 1 no movement.
Position 2,3,4 all ran (with the good motor).
I selected the "lowest" selection in LP possible. 
I actually have a mix of motors.  All are 2212-920kv.  2 motors are LHI and 2 are Ready Tosky. 

I had 2 extra motors and put those in positions 2 and 4.  Now 2,3,4 work as expected.  Position 1 still does nothing. After all this testing the ESC 1 is really warm.  I guess I will order another ESC?

Can I mix ESC with another makers?

Could the software in the ESC be a problem and could you reset/flash them?

Re: Troubleshooting motors/ESC/Revolution
« Reply #3 on: September 17, 2019, 04:46:27 am »
Maybe you plugged the ESC#1 signal/servo connector in backwards?

What happens when you plug ESC#1 into known working signal connector #2?

You don't want lowest possible ESC protocol.  Lowest is PWM@50Hz.  You should run either PWM@490 or PwmSync.  You should probably first calibrate the ESCs using 1000-1900 (not 1000-2000).
https://librepilot.atlassian.net/wiki/spaces/LPDOC/pages/12058743/ESC+Calibration

Notice that on Output page, you should make sure that all banks that have outputs you are using, should all be set correctly and the same way.  That is usually banks 1 (outputs 1-3) and 2 (only output 4) (from what I recall).  One easy way (if you are not using outputs for other things, like servos) is to just set all banks the same way.

I would allow using slightly different ESCs together on the same quad if they were running the same firmware brand and version number, and after doing ESC calibration and motor neutral on all ESCs.

Re: Troubleshooting motors/ESC/Revolution
« Reply #4 on: September 17, 2019, 04:49:17 pm »
Cliff, as always thanks for the help!!!

See answers below in red.

-----------------------------------------------

Maybe you plugged the ESC#1 signal/servo connector in backwards? I tried that.  Flipped it around, no luck.  I've also connected 1,2,4 (not working) to output 3 (working) with no luck.

What happens when you plug ESC#1 into known working signal connector #2? See above

You don't want lowest possible ESC protocol.  Lowest is PWM@50Hz.  You should run either PWM@490 or PwmSync.  You should probably first calibrate the ESCs using 1000-1900 (not 1000-2000).  I am a little confused on this point.  I run the PPM at the lowest (50hz I think).  I also calibrate the ESC.  I am testing the motors during the hardware setup phase where you set the minimum turn (motor neutral) setting.  They do calibrate at 1000-1900.
https://librepilot.atlassian.net/wiki/spaces/LPDOC/pages/12058743/ESC+Calibration

Notice that on Output page, you should make sure that all banks that have outputs you are using, should all be set correctly and the same way.  That is usually banks 1 (outputs 1-3) and 2 (only output 4) (from what I recall).  One easy way (if you are not using outputs for other things, like servos) is to just set all banks the same way. I have upgraded the firmware to LP and erased all settings.  As above I don't get that far in the setup.

I would allow using slightly different ESCs together on the same quad if they were running the same firmware brand and version number, and after doing ESC calibration and motor neutral on all ESCs. Is there a way to check the firmware on the ESC's? Change it?

Re: Troubleshooting motors/ESC/Revolution
« Reply #5 on: September 19, 2019, 10:26:38 pm »
ESC #1 smoked and burned.  Really no need to spend time on it.  It is dead.

FYI: PPM is a receiver signal protocol.  PWM can mean either a receiver or ESC signal protocol.  ESCs don't use PPM but this is not an issue for you.

You can calibrate and test with PWM@50Hz and it will work fine, but when you try to fly it, it will oscillate.  If you successfully calibrated to 1000-1900 using PWM@50Hz you do not need to recalibrate.  Just change to PwmSync or PWM@490Hz.

You should start by using a current limited power supply / light bulb for safety (so a bad ESC/motor doesn't fry yet another motor/ESC) and getting all your ESCs and motors tested and working.  It's hard to do a setup from scratch without working ESCs and motors.  Direcly connect the ESC to the RC receiver if you suspect the FC may be bad.

Run LibrePilot setup from the beginning and choose RapidESC from what I recall.  After running setup, you can look at the Output page and verify that you have PWM@490 or PwmSync for all motors.

Having very different ESCs or motors won't stop you from setting it up.  Having different ESCs or slightly different motors can cause slight bobbles when flying though.  Unless they are very different (like KV1000 and KV1500) you should at least be able to get it to hover well and start testing moving around.  You haven't got nearly that far yet.

Have you ever connected more than 5 volts to the FC?  I know it says it can take more, but for instance the signals going into the FC can never be higher than 5 volts.

ESC firmware, that is a whole 'nother big topic.  There are gadgets to buy to flash the firmware.  No need to spend more than $5 or so but there is a lot for sale for more than that.  There are two main brands of firmware (SimonK and BLHeli).  There are two main processor (CPU) types and different ways (adapter cables) of connecting to them.  Maybe research BLHeliSuite.  Also (be aware that you don't need an ESC protocol faster than PWM@490, but that is a HOTLY contested subject):
https://oscarliang.com/esc-firmware-protocols/

<soapbox>

For a middle of the road quad motor you are probably running say 6000 RPM at hover.  That is 100 revs per second and with PWM@490 you are sending "change RPM" commands about 500 times per second.  So even with PWM@490 you are changing the RPM 5 times per prop revolution.  Smaller quads and higher RPM motors and full throttle might be 30000 RPM is 500 RPS and you can change the RPM once per revolution.  Now I will tell you that ESC firmware smooths out the input variances which reduces the effective signal rate, but that is an issue to be addressed in the firmware if you think you actually need to change the speed of the prop 5 times in one revolution.  You have to remove the smoothing to get the faster protocols working faster too.

It's not very useful to decrease the control delay from .002 to .000000001 when the human adds .25 to the delay.  To actually significantly change the RPM several times per prop revolution would require huge current draws and generate torque that would snap a quad's prop and even snap it's arms off.  Maybe faster response will be more useful in the future with much better hardware and autonomous control that takes the human out of the loop so that the quad is flying stunts itself.

I will grant you that there are high end cases where one step up in protocol speed (say 8x) is useful today, and that changing to a bidirectional digital signal format is useful, but claiming to need to change the RPM 150 times per revolution even in this worst case (that would be 750 times per revolution in a more normal case)?  Or the other side:  Reducing a typical delay from 1/5 prop revolution to 1/750 prop revolution.

</soapbox>