Well, I've come to the conclution I need to bring forth some blue smoke, hopefully datalogged by our schools 8 channel oscilloscope. As some people here state, one prototype is better than 1000 simulations.
Couple of troublesome points:
- Lazy driver, 0,5A max. Solution: Low gate charge on lower fet. I'll limit on my upper fets at 200nC, (these can be low RDSon) (400ns Rise time). The bottom fets will be something with low gate charge (sub 100nC (200ns)) and preferably some rise and fall time - as it would otherwise contribute greatly to noise.
- Capacitance. I'll just start with 200uF of MLCCs, to see if they are useable. PWM will be turned way up - 80kHz start.
This Mosfet from fairchild is a strong candidate as a bottom fet.
Even tho I like the differentiating of fets, I'm not sure how easy it is to implement FOC with only pwm on lower fets. I'll look more into it later