WARNING: Both PR #3139 and PR #3126 are sending dshot 0000 disarm command prior to ELRS link!

This PR (#3139) does allow user to send the disarmed/zero throttle command which is required for use of AM32 ESCs. And I am now able to use 3d mode with an AM32 esc by setting the RX to use normal DSHOT output and using an EdgeTX curve to trick sending 0000 at center with inverted reverse throttle and normal forward throttle ranges. However, it seems there is something weird with the throttle scale conversions in general that has been causing unexpected output behavior and in the case of dshot3d mode it will complete stop/hang an AM32 ESC.

dshot data from boot->low->high->low test :

dshot3d data from boot->mid>low->high->mid test :

draft EdgeTX dshot to dshot3d test curve:
EdgeTX-dshot-to-dshot3d-test-curve.txt

*Updated to remove invalid data due to a dirty pot and loose connection in my test setup. -fixed now :-)

New test of PR#3139 normal dshot along with the RX set to mirror the same channel out on another PWM pin captured with pulseview.
PR3139-PulseView-DSHOT-vs-PWM.zip

This data set clearly shows that the RX is outputting the dshot 0000 disarm command immediately on power up without awaiting link and user input to send dshot 0000 while the PWM pin does not output any signal until after link has been established as to allow the user to safely arm a PWM ESC.

I do notice the output I collected for this test is much cleaner. I am starting to suspect my MT12 P2 knob tainted my prior tests and P2 knob just needed a bit more exercise to clean up the potentiometer wipe path.

When I have more time I will:
-re-test PR #3139 dshot3d mode on AM32 ESC
-refine my EdgeTX test curve to convert dshot to dshot3d

I just ran a quick Dshot-3D test and find this does behave as expected.

Currently I think this Just need to leave the dshot pins floating at boot and wait for ELRS link to send dshot packets.

Commenting this single line will restore arming safety to user control.

that was there before I started modification, I thought that was the desired behavior

the original code would've armed BLHeli32 ESCs, but not AM32

I'm fine with removing the line that you are pointing out as long as everybody is in agreement

Summary understanding:
In ELRS releases versions DShotRMT had nullified 0 command but that has since been fixed in master PR #3126 which has now resulted in the RX removing arming safety from the user. Commenting this line restores arming safety to the user to result in a behavior similar to a PWM ESC.