DShot and BLHeli ESC Telemetry support

Yeah, I might be able to work on that today. https://shop.holybro.com/c/esc_0483 appear like a good replacements for you.

1 Like

I ended up testing something similar, the Airbot Wraith32 Metal v2. Same MOSFETs and driver, anyway.

It took some tweaking to get it to drive a low-kV motor, but it did have working telemetry on a PixRacer . . . for a while.

I think somehow either the PixRacer RX pin burned out, or the ESC TX pin burned out, because the telemetry stopped working. The ESC still works, though. I liked this one because I figured the current would be better calibrated thanks to the Allegro hall effect sensor.

@Naterater Have you found an ESC that seems to work reliably? Or not since the trouble with MultiStars?

I’m putting more eggs in the TEKKO 32 metal (very similar to the wraith 32 metal) category. They’re not fully tested though.

Your burnt out telemetry sounds like it could be from this problem.

I had two crashes (second one yesterday) due to midair freezed F405-Wing Controler with the 41 ampere version of this ESC associated with DSHOT + Telemetry.

This is a bigger problem than I thought. I’ll bring this on the Dev Call tonight to see if anything can be done. At minimum, a warning should be put on the Wiki.

If it is of interest, here is the video of the “freeze” flight:

That’s about exactly what happened to me, except both on takeoff and immediately after commanding a turn (so control surfaces were stuck at extremes). The FC won’t initialize unless it is still, so you never get servo outputs to control where it is going.

I’m so glad you guys talked this through. I bought three of the 41 amp versions, and after reading this, I bench-tested one at full throttle (33 amps) and control movements periodically, and it got somewhat hot, but lasted all the way through the first battery. Heaving a sigh of relief, I went to bed. This morning, just before the maiden, I tried a thee-minute full throttle test on the bench, and it failed. crashed my FC as well.
It’s of course in the bottom of everything.

Follow-up: no freeze anymore with exchanged ESC

After two midair freezings with the Turnegy MultiStar 41A BlHeli32 (occured midair without warning after 2,5h resp. 1,2 h of total flying time) I threw the multistar where it belongs into the trash and took the wraith32 metal.

After now 7.5 hours of trouble-free flight time with the Wraith and otherwise identical aircraft / electrical system (ASW 28, Dshot-150, ESC telemetry enabled, F405-Wing) I do not assume that a freezing occurs again.


1 Like

Just discovered this thread. Has there been any discussion or code changes to allow a mid-air reboot under conditions like this?


I’ve brought it up, but it’s never been a priority for the dev team.


Better wording might be, “The dev team is incredibly stretched so while handling mid-air reboots has been discussed, it’s never reached the top of the to-do list”. BTW, Tridge actually has been looking into mid-air reboots recently for Plane. It’s tricky though and it’s very very rarely required.

In any case, I suspect we are talking about different things here. Handling a mid-air reboot of the autopilot is not what’s being requested here is it? The discussion about a lockup is related to these specific ESCs locking up not ArduPilot.

I respect that the dev team is working on other issues, but we are talking about autopilots rebooting here… NOT just ESC’s.

In addition, there’s language to suggest that manual control of servos is possible on planes in the event that the main thread stops (that’s the whole point of the pixhawk’s co-processor). Unfortunately that functionality is unavailable during the reboot sequence that can never complete (because the plane is in motion). To have ANY servo outputs, the plane must be still to complete the initialization phase.

I’ve lost thousands of dollars of commercial equipment due to this, and I can’t afford it again. I’ve had two specific instances where this happened, and it’s been corroborated by many people including Tridge on the octo porter.

1 Like

Sorry to be slow, but what is this week’s Plane 3.9.8 release about? Isn’t it midair reboot improvements? Main loop freezes, watchdog timer detects, and resets everything, bypassing standard pre-flight requirements?


1 Like

Hi Nate,
We never had the flight controller on the porter reset in flight. We did lose two ESCs in one flight, but the FC kept running.
I do believe it can happen though, as I have seen large voltage spikes on the telemetry line when I was investigating this with a scope. I don’t know if all ESCs have this property or only some.
One thing that would be good to investigate is if a resistor placed in series with the telemetry wire from the ESC to the FC provides any reasonable degree of protection against these spikes. To determine that we need two things 1) someone to test it and see that ESC telemetry still works 2) someone with the right sort of electrical engineering knowledge to tell us that it is a good idea
Would you like to do bench tests with a resistor in series with the telem line and see if it still functions for you?

As a side note, I’ve updated our DShot wiki page to replace the seemingly problematic MultiStar ESCs (discussed above) with the HolyBro Tekka32. These are the ESCs that I have although I haven’t actually used them much yet.

1 Like

@rmackay9 digging this up, why can’t the cube do dshot on the 8 main outputs? I’ve got an octo-quad plane and was hoping to try some apd escs…

nm… from http://ardupilot.org/copter/docs/common-dshot.html

DShot and BLHeli_32 features are currently only supported on the “FMU” outputs of your flight controller. Boards with IO coprocessors like Pixhawk and The Cube provide DShot and BLHeli_32 support only on the AUX OUT ports that are directly driven by the board’s main processor. For Pixracer and other boards without a separate IO coprocessor, all PWM outputs can be used.

On that: Mirko has done some work on supporting DShot on the iomcu https://github.com/ArduPilot/ardupilot/pull/12604/files


Are there any news on the investigation if a resistor placed in series with the telemetry wire is solving the spike issue for ESC telemetry? Or was the discussion moved?