DM542 driving Nema17 controlled by a Teensy 4.1, random stepper stalls

Question

I am trying to drive 4 Nema17 stepper motors using 4 DM542 drivers, controlling them using a teensy4.1. The step and direction signal is level shifted to 5V, the DM542 are powered using a separate 48V supply. In general this setup works great, however sometimes randomly one of the steppers refuses to turn. The corresponding driver is powered, since the motor is applying a holding torque. This issue shows up randomly and fixes itself randomly too. This is not a satisfying answer however, so I am looking for clues on what the underlying issue might be. What have I tried so far:

1. Wiring issue: I have checked whether one the the leads wiggles itself loose or something and was not able to find any issues in the wiring.
2. Broken stepper or driver: I have changed the drivers and steppers around and was not able to detect any pattern. However lets say, if the issue occurs with stepper 1 connected to driver 1. Now connecting a different stepper 2 to driver 1 will cause stepper 2 to stall. Switching the teensy pins connected to driver 1 to driver 2, will cause the motor connected to driver 2 to stall. This suggests that the problem is on the teensy side of things. But:
3. Check for a software issue: I am quite sure that this is not a software issue on the teensy side because like I said, generally this setup is working great. Of course there is a chance of a random software bug, however if this issue occurs, resetting the teensy does not fix it. Rather waiting for a few minutes does.
4. Checking the signals with an oscilloscope: Interestingly I was not able to check the signals to a stalling driver because 1. the issue does not occur that frequently and 2. the one time I was actually ready with a scope on hand, the issue fixed itself immediately as soon as I placed the scope probes on the signal pins.

I have to admit that I am out of my depth right now, so any hints on what might cause this problem or what steps I can take for a proper diagnosis are appreciated.

Edit Reacting to comments made by Tony Stewart EE75: The level shifter being used is a BSS138 (https://www.sparkfun.com/products/12009). The DM542 is setup to provide a max peak current of 2.37A (RMS of 1.69). Microsteps are set to 25600. I struggle to provide a max accleration of the steppers, or comment on the Friction on the shafts. The application is a build similar to https://github.com/T-Kuhn/HighPrecisionStepperJuggler .

Answer 1

With this high voltage, I think it is a driver thermal problem with OTP protection.

The driver must use current limiting to use this high voltage and may need a fan to keep cool 35 mm type or more.

Otherwise , you have to reduce supply voltage. Acceleration parameter is also critical for slipping. Since a=F/m this is also proportionally to current and heat rise with your present thermal resistance.

So you either adjust current limit, reduce V+ or add force air cooling or reduce max acceleration in steps/second motion or examine friction loss and improve all/any until it has enough margin to NEVER fail under ALL environmental and load conditions.

OTP inside switching or linear regulators is thermally sensed and self resetting, normally.