The overall current drops to nothing, while the command to the rear motors is going up, and the quad is pitching back (the command is correct for the pitch change, but there is no sign of response). The fact the current drops off says the ESC is off line.
ESC Overheat - very probable, especially if ESC works fine after the incident.
Verify Settings of the Temperature protection. Possibly setup ESC temperature telemetry. If that is too complicated, fly for a few minutes and use your finger (on the ESC)…
I agree - it appears that the ESC shut down due to temperature.
I’ve built several other copters with this enclosed frame using other ESC’s, but haven’t ever had an issue. Maybe there’s a reason that this ESC has a big heatsink.
I went to check the temp data from my other copters like this using BLHeli_32 ESCs. Turns out that on BDshot my BLHeli_32 ESC’s only report RPM. On the other had, AM32 reports many more parameters over BDshot - including temperature.
So now the question is how to enable a proper failsafe. According to what I’ve read, this requires the Ardupilot AFS module. (Advanced Failsafe System) This is included in “plane” but requires a custom compile to get it in “copter”.
Seems to me this would be a no brainer to include in the Copter BDshot firmware. It’s nice to be able to use RPM for a notch filter - but if someone is going to the trouble of using BDshot, I’d guess they’d like to make good use of as much of this data as possible.
Of course for ESC’s with a telemetry port that’s connected, this data would be useful without BDshot.
I hope you’re aware that a failsafe function of any kind doesn’t eliminate the cause of the overheating, but only serves as an emergency brake. However, in normal operation, it shouldn’t even get that far.
Therefore, in your case, I think it’s much more important to find and fix the real cause.
I have no experience with drones and ESCs like this. Is the behavior shown in your last picture completely normal, or are the motors being controlled back and forth too intensively? Could this be causing the overheating, or are you already reaching the limits of the ESCs with your drone hovering?
The last time you pointed out rcout oscillations for me it was because they needed to be reduced to get proper performance when enabling the rpm notch filter.
Last time I followed these instructions. I reduced the PID gains in half - and then a second time half again.
So today I just dropped the PID gains by 75% and performed a test hover flight. I landed after 2 minutes because the copter was so “mushy” I didn’t feel like I had proper control.
So then I re-loaded my original parameters and this time reduced the PID gains by 50% and performed an 8 minute hover flight. (it took a little over 10 minutes for the ESCs to quit last time)
The rcout chart looks about the same - and the temp comes up to about 120C again. When the ESC’s quit before the showed a max temperature of 128, 129, 124 and 118.
I need to contact Skystars to find out if they’ll tell me what temperature the ESC is set to shutdown.
Regarding PIDs to start with - I have to start somewhere with the PIDs. So and again I started with the initial tuning facility in Mission Planner.
If you can suggest better starting PID gains, I’ll try them. Or I could try the PID gains from my previous build that had a successful autotune - although it uses 2213 motors instead of 2216, and 9.5" props instead of 10". And the arms were about 30mm shorter.
By the way, I made a start at using the Methodic Configurator and I sort of stalled out at the IMU calibration step. I’ll work on it some more later - but it’s a lot of work - particularly for a new user.
But just as we use a failsafe at a battery voltage or amperage level that prevents the vehicle from totally exhausting the battery, I believe it would be helpful to have a failsafe to prevent the temperatures from exceeding a level that causes the ESC to shut down.
This failsafe (for ESC overheating) is absolutely useless. You must make sure that the ESCs operate in normal conditions, i.e. max 70…80 C.
I am not sure why you have an overheat, your copter seems to be small from the video, the 55 A ESC can easily sustain a flying weight of 3…4 kg with temperatures not exceeding 50 C.
Unless you have a complete enclosure which prevents wind cooling, you should not have these problems.
I just checked your current, less than 10 A. Your MUST get rid of enclosure completely! That is the most stupid thing to have. Your ESC, totally exposed, should run at 30…40 C max.
Yes, I can imagine it was. I would set the Rate Pitch/Roll PID’s to .05/.05/.0015 and improve the cooling situation as @Michail_Belov is suggesting. Configure the notch filter and then run a Quicktune before trying Auto Tune.
Look, I have to see your enclosure to be certain, but if you know better, than so be it. There is virtually no way an ESC can overheat at 8 A unless you encapsulated it well. If you have other copters flying “well” that does not mean anything.
Now, there is a remote possibility that the ESC heat shield is detached from the Mosfets. Since we know that your ESC had 130 C from the telemetry data, you have to measure the exterior of the ESC with a thermometer , if it is close to 110 C, that means heat is correctly dissipated.
Now, since you were flying for 10 minutes, I can guarantee you that your ESC would last for about 1…2 minutes if it had bad thermal contact with heat shield. So I doubt you have any problems with the ESC heat dissipation.
I have a copter with this frame that has flown 20 minute mapping missions without a problem. It uses a T-Motor F45A Mini ESC. I just connected it’s telemetry port so I could get temperature data from it for comparison. Unfortunately I’m only receiving Voltage and TPM on the F45A’s telemetry port - no Temp or Current data.
My F45A ESC is running version 3.9 of the BLHeli_32 firmware - maybe version 3.10 fixes this - but unless I find an ESC that has it installed, I’m stuck.
You make a good point about the thermal past on the Skystars heatsink - I’ll consider checking that in my options. But the fact that Skystarts includes a heatshink and it’s the only 20x20 ESC I recall using one - it’s suspicious.
Given that there are dozens of this frame flying with other 20x20 ESCs (FETTEC, SpeedyBee, T-Motor, HolyBro) that haven’t had any issues - the easiest thing to do is just ditch this SkyStars ESC. But who knows - maybe I’ve just been lucky on the other copters.
What I’d like to do is find another 20x20 ESC using AM32 and give that a try. ESCs are in short supply right now - so it may take a while.
