An old design is using UC2625 in a peculiar way: the pwm is generated externally and the RC-OSC is synchronised to it using a transistor and some passive components.
The problem is that the motor is running very unstable at medium speeds. Its ok at low and high speeds.
I have tried a number of things: filtering the Hall sensors, decreasing the rise-time on the Hall sensor signals, better decoupling of Vref etc. etc.
Could it be that the PWM/clocking principle causes a race between set and reset within the chip?
