for best control response (and best efficiency) the inductance would need to be minimised
(this has practical disadvantages and so is not for general purposes)
in this way the time for the MGU-K's full response of torque internally will be c. 5 millisec - the electromagnetic time constant
this response will be delivered externally if there is no rpm acceleration or acceleration torque is of external origin
eg wheelspin from K + ICE torque could allow K torque to reverse fully within 5 millisec
(accelerating itself the MGU-K's full response externally will be c. 40 millisec - the electromechanical time constant)
these values are taken from an older 5 kW machine - assuming the K's modernity and size give no net gain or loss
time constant was the time for 63% of a step demand to be realised - this for brushed machines
modern machines won't be given step demands - but the response is in principle the same as a brushed equivalent