Maybe because it could be destructive testing or even dangerous not knowing what could actually happen? I normally have limits set via my Cycle Analyst, so I guess just making a rule never to go unlimited unless in a safer controlled environment, or just keep limits in place all the time.
Some thoughts, a bit disjointed because I've been doing some involuntary napping....
When the previous one failed, was the bike going faster than it's normal no-load speed? If so, and if it's a DD motor, it oculd have generated voltages higher than the FETs could handle (especially if the battery has a common-port BMS and shut itself off for any reason), That could have caused the fet faulre.
If it was just during a startup from a stop, if it was high load, or high throttle on a non-FOC controller, so it doesn't control the phase current, just pwms the voltage to control the motor speed, and thus at zero or near zero speed it can use 'block mode" where it doesn't pwm anything and just commutates the
motor at very high current for some amount of time until the motor is above whatever speed the controller's designed for or the block time runs out.
it could be worse if the controller has been shunt modded as it then doesn't even know how much battery current is flowing, and if it uses that plus some math to guess phase currents for limiting motor draw then it can't do the right math so can't get the right answer and the fets can be overloaded and blow up.
The Cycle analyst can't see phase currents so it doesn't matter if it's got limits or not, or if it's even there, for this problem. it only sees battery current so if battery current never exceeds the limit, the ca won't roll throttle down.
An FOC controller that uses throttle to moduclate phase currents (most of them do) will probably not have this situation, unless you've tweaked the controller settings at or beyond whatever safe zones they have.
all controllers ahve some form of limiting if they haven't been hacked or whatever, or settings twaeked beyond the safe max limits or defaults. foc controllers monitor phase currents directly, some all three and some only two and do math for the third, almost no non-foc controllers measure phase currents at all and only measure battery curretns. so the limts for each apply to those measurements only, and can't protect against the other one it doesn't directly measure.
fwiw, a rear wheel lockup is much better than a front; at least you'll only lose rear traction or skid or brake hard or whatever, instead of potentially flipping over the bars from a sudden braking event or losing steering control.
the fuse idea should work as long as it the fuses are rated correctly; i'd use bolt on fuses to ring terminals crimped correclty to the phase wires. afvoids holder problems. just remember that as soon as they blow the phase generated voltages at whatever speed you're at will spike to the unloaded speed voltage right then, so you want to find out what the motor will generate (what it's kv is x the rpm) at the highest speed it will ever spin at worst case, and use fuses rated for higher than that voltage to ensure they can't arc across when they blow. and make sure the current rating is correct for that specific fuse per it's manufacturer chart to blow instantly when there's a short and you get phase lcokup, but not to blow for short regen events or startup events, etc.
of you areladly have a blown controller, you can measure what the phase currents are with an ammeter between the cotnroller and motor in any of the shorted hpase lines. then spin the motor up to speed (highest current will be at the speed it might be at the worst case when this could happen on the road, but you would need to test at the slwoest pseed you might hafve this problem at so you will know the lowest current the fuses must blow at), using a diferent external drive system, or if t will spin well enough to ride ,you could go down a steep enough hill to generate that speed. then once you know the current you need it to blow at
