I don’t think that’s to do with this PR, the comment says:
// protected_switch_full_sector is protected by switch_full_sector to
// avoid an infinite recursion problem; switch_full_sectory calls
// write() which can call switch_full_sector. This has been seen in
// practice, and while it might be caused by corruption… corruption
// happens.
So you might need to reset you flash
Flash Rover then Copter again?
Yes, exactly - you can also set the format version parameter to 0
Really strange: after reset flash now all combination of rate and pwm type gives me only 3 motors working.
Without bdshot all working.
Is this with the config above? I can try again to see if I can reproduce
I’m afraid I can no longer reproduce with your config - all works perfectly. Can you try 32.8? I am using that now - just in case.
Also the calibration step happens at power-up, if you power your ESCs after the FC they will miss it
I am using 32.8. Battery power ESC and from ESC go to the FC.
I just tried this on a different build - all worked perfectly 
I think my ESC has something wrong and is getting worse, even with an old release (from 3 Feb) that never gives me problems in bdshot now gives me only 3 motors working.
I have ordered a new one that is on the way to my home.
That is my suspicion - you definitely have found issues with your testing, now I am not so sure.
About what?
[more char]
I’m not so sure that the problems you are now seeing are down to software - I think you are right there is hardware at fault here.
Hey, so I’ve just found a fix from betaflight that makes two of the motors work on the BeastH7 - can you see if this affects your results at all? It’s specific to the H7.
I am sorry, I tried to re-flash 32.8, one of the ESC gives error and now is definitely dead
In the next week I should have the new ESC.
My fix makes dshot work but telemetry is busted
I have backed it out - does not seem to help in any way that I can detect
Hi @andyp1per I mounted a new BLHeli_32 ESC on my 3 inch with Matek H743 mini, for now tested on master w/o bidir and is all OK. Which PR (which branch on your repo) would you like I test?
Tested many, many times with SERVO_DSHOT_RATE = 0, 1, 2 and MOT_PWM_TYPE = 4, 5, 6, 7 and only one time with SERVO_DSHOT_RATE = 0 and MOT_PWM_TYPE = 5 (Dshot300) started with 3 motors, but I was not able to reproduce this negative result.
other relevant params:
SERVO_BLH_AUTO = 1
SERVO_BLH_BDMASK = 15
SERVO_BLH_TRATE = 100