Copter-4.7.0-beta7 has been released for beta testing and can be downloaded using Mission Planner, QGC or directly downloaded from firmware.ardupilot.org.
WARNING: TradHeli’s arming behaviour changes to require collective stick full down regardless of using stick or switch arming. In RC_OPTIONS parameter, uncheck option for “Arming check throttle for 0 input” to allow switch arming with a collective raised from bottom. This option is not recommended for stick arming as it allows arming at non zero collective setting.
As with earlier, versions, if you’re using QGC, please use the daily build which has been updated for the parameter changes in 4.7.
Seems so, but they went just dead silent. I suppose this is expected reaction of ardupilot trying to keep altitude as hard as it can. Too bad this ESC has not serial port telemetry.
I wonder what is ESC[n].Err as that spiked up all at once on all channels when motors stopped.
Yes. It was in loiter, so 1500 was when I released stick to mid position after motors went silent. Throttle channel in loiter is not directly translated to motors output. AP keeps some limits of velocity and acceleration too, so this was barely 10m/s ascending. I tested it way harder before (with 4.6.3) up to 25m/s vertical ascending and over 70A current which is still in specs of ESC (theoretically 4x70A), so 70A total is just 25% load, and 40A reached in this test is just around 15% max of ESC capabilities.
Thanks again for the report. I had a bit of a look at the log (not complete yet) and it looks like the vehicle is using:
BLHeli / Dshot 300 with ESC telemetry arriving on Serial7
Harmonic notch is setup to use ESC telemetry
The ESC log message has errors telling us that the ESC telemetry stopped arriving from the ESCs. This in turn leads to the FCN log message displaying NaNs because the harmonic notch feature no longer had a reference for where to filter.
You said you haven’t seen this since 4.5.7 so I guess you have seen this before? If you have any additional background on that, that would be great.
I think an ESC failure is probably the most likely cause but we’d like to be sure especially at this critical phase of the 4.7 release. Can I ask which ESCs are being used? Is it a 4-in-1 ESC? If “yes” then a sudden failure of all motors seems possible. If they’re independent ESCs then that’s less likely
On that board PWM1/2 and 3/4 are on different timers so unlikely to be DMA issue.
Did you make any changes to the version of AM32? I know there have been some questions over the version being used on that board, there is also a thread where it has burnt out.
That is correct, but the ESC i am using now has no serial telemetry actually and relays only on bdshot to get RPMs.
I did not. 4.5.7 is just version when I started using ardupilot as FC firmware.
ESC in this setup is Miicoair 70A 4in1 ESC (surprisingly currently not available in micoair’s store) - MicoAir 70A AM32 | Rotorama
No hardware or settings changes. AM32 version 2.20
What concerns ESC failure - I tested it post crash and it works normally, nothing unusual, motors are spinning as expected, also did some load test on ground with revesed props and everything works fine.
Not sure how much this ESC is affected but I have Micoair AIO 45A that went into flames, and has issues with F4A chip also used in this 70A ESC.
That beeing said, I decided to change ESC to something I consider more reliable and with serial telemetry supported - Skystars AM60. Not sure if I want to re-test this beta version again, but probably will do it. This ESC I have already tested to about 100A indicated.
Despite all positive impressions at the beginning I am starting to think micoair hardware is not as high quality as it appears to be. At least ESC.
Thanks for all analysis and feedback. I have currently very stormy weather but will try to retest with new ESC as soon as possible.
That was on a small side (only 470uF - I belive that was value that came with stack) from unknown brand changed to nichicon. This time I put 2200uF for Panasonic Should be plenty
Pro TIP: Always make sure new ESC has + and - in same orientation as the old one… Somehow everything survived. Nothing better that adrenaline spike at the end of rebuild when You try to pull apart connectors that decided to weld themself. At lease I know my welds on 21700 pack are done well
The critical thing we need to know is - what are the odds this is a regression rather than just a once-off event which could have happened at any time.
Two questions, really.
were any hardware modifications made between flying whatever was on their before the beta and flying the beta?
how much flying has this vehicle had - in the same hw config - on previous firmwares
No hardware changes (in terms of FC+ESC) since July 2025 when I installed micoair’s stack. Only changes are like: GPS module, camera but not recently. Also with 4.7.0 beta5 did endurance flight about 30 minutes to check if battery failsafe works OK.
Total at least 100h flight time including short dynamic flights and longer range ~7km total flights.
BTW - I had a little chat with alkamotors dev about this.
Should be plenty 
