The problem has not been solved.
Here is a link to a video of the second engine not starting
https://drive.google.com/file/d/109ZLFyqs-xQccW4g6Ke6_-lbDrxLFlnh/view?usp=sharing
Any other ideas on how to fix this problem?
You will have to replace the ESC and maybe the motor.
You could try swappeing ESCs or motors and see which the problem follows.
If they are intergrated motors and ESCs you wont have a choice - just replace the whole unit.
Four drones with this problem.
The engine does not start randomly, not specifically the second one 
I still don’t understand how to solve it programmatically.
thank you Shawn
It’s likely not an issue with parameters if you have followed everything we say, like MissionPlanner motor test to determine MOT_SPIN_ARM.
It sounds like you have bad ESCs. Make one copter with all working ESCs and motors. Find out of it has continuing issues.
Exactly what ESCs and motors are they?
motors hobbywing x-rotor pro x6
Also just tested this on 4.3.3 and saw this error although it output successful calibration on the graphical interface.
The first time I tried I got nothing. It wasn’t until I pulled the throttle down as the GUI said “throttle not at zero” that it provided data. I have a Herelink HD so it always bounces back to the middle stick. Originally I was looking at throt-0 but had arming issues so set a 50% DZ on the RC and it seems to work in principle but I haven’t flown yet and don’t know if this is the way forward but arming works as expected.
I have a 25 KG Hexa Plus with HW X-rotor X9 plus. All the motors spin up. I’ve completed numerous motor tests individually and completed all the way from 15-95% throttle . They are set manually as the recommended built-in ESC’s are fixed at 1050-1950 μs same as the X6 Pro @IgorW.
At the top of range, it draws 251Amps at 95% Throttle which is eye-watering.
The interference number output is also very high (207 %)
I have the onboard Compass set to secondary and a here3 as the primary.
The next test is to recalibrate Accel and MagCal in 6 axis when I have a set of spare hands as it was so big initially I used Large Vehicle MagCal so it’s possible I didn’t do this correctly.
Might try to check Accel position is manually set before this as I read somewhere this can cause an issue.
Couldn’t view the video, but what happens if you swap power, and control plugs between a working and the non-working position? Does it ever make a difference to where the motor is located?
doesn’t matter, it happens randomly
Igor, I just thought of something and checked in your last log…
Nothing definite, but try this to use 5v signals on the PWM outputs
BRD_PWM_VOLT_SEL,1
This sets the voltage max for PWM output pulses. 0 for 3.3V and 1 for 5V output. On boards with an IOMCU that support this parameter this option only affects the 8 main outputs, not the 6 auxilliary outputs. Using 5V output can help to reduce the impact of ESC noise interference corrupting signals to the ESCs.
EDIT
The specs definitely say:
3.3V/5V (Compatible )
PWM 1100-1940 μs (Fixed or cannot be Programmed )
so you should have the correct parameters set already, and the 3v5/v setting is bit of a gamble. It wont make it worse.
I still get this error at the end of a CompassMot test however a bit of progress.
Redid the Accel in 3 axis, levelled and then run the MagCal with some assistance to move it about.
Not ideal given its size.
Disabled the 3rd compass referenced and achieved significantly better results. It’s 250A / 95% throttle / 59 interference so sitting in that grey area that’s referred to in the wiki.
I get another error off the back as well prearm ahrs: EKF3 Roll/Pitch inconsistent by 13 d which I’m investigating.
Ran CompassMot again today. Factory reset Orange Cube. updated firmware to 4.3.5.
Ran a single-axis Accel and did a full MagCal. 1 internal and 1 external Compass.
Says it’s completed Calibration but still see the error
3/16/2023 1:59:42 PM : PreArm: Internal errors 0x400 l:234 cnstring_nan
3/16/2023 1:59:23 PM : EKF3 IMU2 is using GPS
3/16/2023 1:59:17 PM : EKF3 IMU1 is using GPS
3/16/2023 1:59:17 PM : EKF3 IMU0 is using GPS
3/16/2023 1:59:11 PM : Internal Errors 400
3/16/2023 1:59:10 PM : Calibration successful
The interesting thing is I get more interference after running the MagCal than
setting the Large Vehicle Magcal manually with a heading.
50% Throttle
92A
79 Interference
I read at the top of the thread that this is a bug?
Is that still considered the case?
I know where this bug is coming from. This change fixes: Per-motor thrust-based compass compensation by andyp1per · Pull Request #22490 · ArduPilot/ardupilot · GitHub