[Fujiyama1]

Hi guys, another thought, the idea of PWM is to reduce volt drop from the system eliminating wasted power, so dont use a resistive current sensor , use a hall type, but ensure the frequency response is high enough for switching of the PWM rise times nS , and off cause the output of these device will not give you an accurate reading as it is not a dc current you are measuring but a square wave made up of a infinite amount of high frequency sine wave components of different phases and magnitudes, this will change of cause when the duty cycle changes, In industry we get around this using a true rms meter but these a expensive and ahve a sample time which will slow down the feedback response, I have considered several designs but could not guarantee to limit the current accurately to say 400 amps within milliseconds, whilst with a known duty cycle and battery voltage and motor (impedence) resistance I think you could calculate the max on peroid of the waveform for a given motor speed with some added safety margin
Barry