Melting issues in FOC (VESC4.2)

Hello, I am not sure if this topic is already explained in another place, but I couldn't find specific info about it.

I am trying to get the best low-rpm torque for my e-bike application with a VESC4, but when doing tests in a dyno this is the second time I burn the VESC in FOC mode at low rpm.

I am testing in both sensored and sensorless mode:

In BLCD mode the test gave more torque than in FOC, but second time I was going to do the run the VESC burnt one of the tracks after the shunt. As a result: hi-side and lo-side MOSFETs are burnt and driver too. By now I didn't see more damage, but I couldn't test it yet, tomorrow a new driver should come. I think that somehow the hi-side fails, then it causes a short to the low-side and then something has to melt, as there is no fuse yet (with a fuse I don't know if it would survive or just less damage)

This is the 2nd VESC I burn in the dyno. First was a clone but second was original.

My questions are:

• Is this something related to VESC4 or to FOC in VESC4?
• Is this fixed in VESC6?

thanks

Here you can see the comparison of BLDC sensored (high torque peak) vs FOC sensorles (soft-start). The PCB melted when I was trying FOC sensored. 

Note that torque is measured as linear thrust (huge)

BTW: I also observed that battery current limit is applied when in BLDC but not if FOC (see flat section at red / power graph), is this possible?