There have been so many reports of ESC catching fire the past 3, 4 months, it’s unreal! And I burnt an ESC yesterday, there was so much smoke and flame it was horrifying! I am trying to find out why this is happening so frequently.
A bit of background from yesterday: I was experimenting with different PID. The quad flew happily for 30 mins, but after a sharp turn, the quad just fell out of the sky. As I walked closer to the quad I saw lots and lots of smoke, and fire! I rushed to the crash site, immediately unplug the lipo battery, few seconds later the fire went out. Luckily no damaged was done to the quad.
This mini quad flew reliably and beautifully for the last 2 months, and you know what I did before that crash? I almost doubled my P gain for both roll and pitch (I was testing the effect of high P gain, it was changed from 3.8 to 7)
I also written a guide on how to choose ESC you should check out.
So here are my theory what could potentially cause ESC set on fire:
Running Higher Voltages
More and more people run 6S LiPo. Although some ESC’s are rated for 6S, and a good proportion of them are reliable, there is a higher chance of burning the ESC than 4S. Higher voltage is just more risky.
Excessive PID value, especially P gain?
When your P is high, the quadcopter will try to change motor speed more rapidly and stabilize the aircraft harder, therefore motors will draw more current.
This is what I think was responsible for my last ESC fire. Luckily I was running blackbox and I got some data of the flights. I am comparing the data of two different flights, first one with P gain of 2.0, and the other with P gain of 7.0 (both roll and pitch), all other PID values and rates are identical.
What I found interesting in those blackbox graphs was the motor outputs (The lines at the bottom). The flight with higher P gains, you can clearly see the change of motor speed is more violent, even when the quadcopter isn’t doing much the lines are still jumping quite a bit.
(If you are interested in how blackbox works, stay tune I will put a tutorial on my blog tomorrow night!)
Here I was just flying forward with medium throttle.
And here I was doing a sharp yaw turn.
And here is a roll. Notice there are two rapid changes of motor speed, First one is the actual roll, the second part is the overshoot (as known as “bounce”). The overshoot is huge with high P gain.
And finally here is a flip. The overshoot is dance all over the place! Must be painful for the motors!
ESC setting – Damped Light / Motor Timing / Demag
Pretty sure motor timing has an effect on how the motor runs. I have heard when motor timing is set to too high, some motors just draw excessive amount of current, gets hot or simply don’t work well. I think for most motors the default “medium” is pretty good. Some prefer “med-high” for Cobra 2204 2300KV motors, and seem to have good result.
Does Damped light put more pressure on the motors and ESCs? I think it does, as it uses more power to actively brakes the prop and motor from spinning when you slow down. But this feature really improves flight performance.
BLHeli in general?
I have heard more than once that people accuse BLHeli firmware for burning ESC in general, and said SimonK doesn’t have this issue as often. Hope someone with experience with these firmware to shed some light on it!
Bad Air Flow to Cool ESC Down
Whatever you do, always have good air flow to help ESCs cool down. I personally prefer mounting ESCs on the arms, cheapest and most effective way to avoid overheat.
Simply Bad ESC
Badly designed or over-rated ESCs are likely to catch fire. Also I think the reason for miniture ESCs like DYS SN20A and KISS 18A to catch fire so often is due to the tiny size, insufficient surface area will make heat dissipation slower, and they don’t even come with heat-sink. You probably don’t want to push the limit with these ESCs.
What else could set your ESCs on Fire?
Let me know if you think some the above factors are the ESC killer, or anything else I have not mentioned.
I just burnt esc on my wizard x220 and i replaced it and it works but after giving throtle and also it doesnt accelerate as fast as it used too. Any help would be helpful
today i plugged in my quad, it played its little tune and one of my spedix gs35A esc’s burst into flames….scared the shit out of me, still have no idea why it happened… ?
so say i landed upside down and killed the throttle, yet the motors still wana spin as they do when its armed even at 0 throttle.
now i know i need to disarm as fast as possible to avoid that but say i dont, how long can it take to burn something out?
today was the first day i flew a brushed quad (sailfly-x). second fpv drone that i had (beta65x lite was first)
i got into the habbit of disarming but sometimes there would be a second or 2 were it was in the ground still armed. i wanted to know aproximitly how much time i have to disarm (assuming throttle is all the way down) before something catches fire
I made some tests with a bleheli ESC after a copter catched fire.
Obiously the motor can gain a very high voltage on they input if Damped Light is on and the motor is breaking suddenly.
The voltage peak is very strong especially if you have big motors/props.
SimonK ESC does not have active breaking and do not cause this peaks.
The Voltage can be more than twice the input voltage.
The amplitude depends especially on battery health(resistance) and capacitor capacity and connector resistance.
Also other electronics connected to the copter could get harmed or disturbed.
All blheli ESCs show this problem in newer ESC its possible to disable damping.
I hope I can help to avoid fire and electronic issues in the future.
I am experiencing a twitchy motor. After a crash with my E-maxx Hawk. The following day after I repaired the superficial damage I attempt to arm the motors and I got three spinning and one just kind of twitching. I’d like to troubleshoot this but I’m not sure where to begin I checked the solder joints on the leads to the motor in question. All my previous quads have been Horizon/Blade Pnp so I had don’t have any familiarity with cleanflight/betaflight or any of the other moddable type firmware/software
Not sure it will help much, and you probably already solved it, but just in case:
In my experience, twitchy motors are most of the time due to bad ESC.
The best troubleshooting method imo is to do a cross check:
Invert the motor with the one going in the same way (invert 4 and 1 or 3 and 2 and viceversa) and see if you get the same behavior.
A bit tedious to do, but at least you’ll be sure.
And to avoid scars, blood and tears, be sure you remove all your props while testing ;-)
Bad ESC give motors a pretty unpredictable behaviour…
PS: I just smoked a 4 in 1 ESC tonight (shortcut due to badly insulinated wire running against the frame) and one of the motor gets twitchy (going on and off, on and off) then… full RPM.
Props weren’t off but I luckily enough held it firmly.
Scary enough ^^”
SIR can you please please help me……
I have build just now A F450 Quadcopter
10×4.5 Inch propellers
30A simonK ESC
3000mah 3s 30c Battery
And Flysky Fs-Ct6b transmitter and receiver
This is my quad specification….
My issue is that my ESC burns almost 2ESC i have burned…
As propeller of quad made collide with wall it was burned….
I don’t understand that why this happening….
All wire connections and battery is good…
But i have doubt that as you mentioned in this blog about P gain that how much P gain is better for Quadcopter???
I have same prob. Pls help . Burnt many esc’s and the battery became thick . I think it is because of i did not use any balance charger. It was a compact charger
Smoke and fire from my Walkera f210 today. Hopefully can salvage with just an ESC replacement. flames and all was pretty spectacular.
I have bought LHI 220 with ESC Littlebee 20a and 2205 2203 motors. I am new at this. After putting together and notice one of ESC was hot than other three. I haven’t fly it yet but I do not want to fly it if it continuous getting hot. Is there a reason why one get hot and others don’t? Another thing is the motor that is wire to hot ESC does not work.
Did anyone mention a short in the wiring? If I did sorry. I too just started getting this issue. I don’t use a pdb instead a wiring harness. I am on my 3rd esc and second motor and a 4th flight controller. I think this is going to be a rebuild to include a pdb.
Did you find a solution for the problem. I am searching vor the web and just you have the same problem . My emax 12a blheli esc burn up immediately after a soft crash in grass . That strange but i thing as you that the software is killing them when the copter is correcting the crash and putting a strong value.
Did you find a solution ? It would help as a lot ?
Obstructed propellers after crashing will cause ESC overheating and fried.
How to avoid this?
don’t crash? or dis-arm right after the crash.
I have the Spedix s250 converted from the KK2 to the CC3D. (Just wasn’t satisfied with KK2) Today, after setting up the quad thru the wizard, with the battery plugged in, motor #3 began to twitch as if I were bumping the throttle, which I wasn’t, and furthermore the quad hadn’t yet been armed. I did a short test hop, which was very unstable and to me favored the suspect motor, so I landed, reconnected to the computer and restarted the wizard. While setting the low throttle point POOF, smoked! Idk what could’ve happened.
short somewhere? check if the motor screws are touching the wilding inside the motor? or could it just simply be a bad ESC?
lastly, i would really appreciate it if you could post your questions on the forum in the future: http://intoFPV.com … I don’t want to miss your comment, because I only check my blog comments once a week, but I use the forum daily!
Hi Oscar, your blog is a wealth of knowledge, one question, can a faulty battery cause esc’s to catch on fire?
personally i am not sure. But i don’t see how it would, since the voltage of a fault battery is only likely to drop rather than rise…
I am having some troubles with esc´s popping lately.
I have a ZMR 250 that i have modified and turned into a pusher prop design. The trouble is that when i do a hard yaw the esc´s loose sync and are prone to burning. When i dry run the motors in CF i have no trouble with desync as long as i calibrate the motors to 1848 in BLheli and use max throttle 1830 in CF.
So far i have tried three different esc´s (kiss 18, LittleBee, RCtimer NFS 30a) all the esc´s except the kiss 18 have been loaded with BLheli 14.1 and they have all had some sort of sync issue.
I was wondering if CF can go over the max throttle during hard yaw or if it should be locked to what i state in the configuration tab?
I am using a CC3d flashed with Betaflight but the same problem also occurs on stable CF, just to rule out problems with the FC i have tried a CC3d atom with no luck. Furthermore i have tried two different sets of motors 2206 2100kv Cobra and 2206 2000kv Tmotor but without any effect. I have had esc burn up both with and without propellers and the logs i have from blackbox do not show any hard oscillations in the pid loop. They do however show the yaw P going off the scale as the copter tries to yaw without propellers, witch leads me to suspect that it is a pid related issue.
Would be grateful for any suggestions to try.
Thanks for an awesome blog.
First thing I would try is, to calibrate your ESC the normal way. Set your min command and max throttle to 1000/2000, and then calibrate your ESCs again… don’t change anything and test fly it see if you still have any issue?
I tried resetting the max throttle command to 2000 and min to 1000 then i calibrated the esc´s in PWM mode as the calibration cant be done in oneshot. After calibrating i still have audible desync when running the motors without propellers. I have tried everything to get them up to 100% without losing sync but I cant find the problem. It seems the only solution is running maxthrottle lower than calibrated throttle but that means I am throwing away the final 10%. Been through 3 different esc´s and I have the same problem with them all. Thought it might be the FC but when i tried a different fc the problem was still evident. Tried switching my PDB and wires and had no effect. Tried disconnecting the FPVgear without any result. I have been over the whole copter with a multimeter without finding any shorts.
I was reading through your tutorial on flashing esc´s and it seems that the LittleBee 20a needs to be connected to GND and port 11 when flashing with Arduino Uno.
that’s very odd! i can’t believe you have the same problem with different ESC/FC/Motors… The only time i have desync issues was because of the bad motors I have, 1 Cobra 2204 and 2 Emax 1806… but replacing them fixed the issue…
Bang on about these small esc’s. I’m staying away from them now, especially these DYS SN20a’s. Got weird issues with them where the solder melted inside on just 3s power systems. Also bang on about high pids, they are sometimes the culprit.
Have a look at what happened to mine : https://www.youtube.com/watch?v=vwyAiAfuF2Y
Hi Oscar – interesting write up & plots. I’m surprised at the huge oscillation in the plots on the last two rows. Huge overshoot with such a high P rating, the ESCs must be switching from full on to full off that it’s putting a lot of strain on the motors. Even worse, if I understand damped light it will be actively slowing down the motors on the low points on those oscillations so the ESC is working twice as hard and not getting a break. So yes, damped light with high P is much more likely to put strain on the ESC and risk burning it out compared to SimonK which would free wheel.
I wonder if we might not see these strains on SimonK too since they also have Comp_pwm now.
Thanks Neil for the comment!
Hi Oscar :)
WoW !!! Extremely Usefull To know this !!
I will try to Log the same values to see how this works with an APM 2.6 . I’m curious :)
Thanks Garfield :)
Yes it’s fun to play with data!
if you really like fireworks, you should try kiss 18a. I saw the last one on friday. :P
No i installed the cleanflight from BorisB with some LPF on the gyro input and each PID output.
It should at least reduce the problem.
The github discussion:
thanks! that’s what I have been recommended many times and will try out this weekend!
What setup did you use with the Kiss 18A?