We have that from 1500 us to 600us variation in pulse width causes 90degree of rotation.Thats not rotational velocity, thats maximum travel distance (divided by two) and the range of possible PWM.
So degrees/us = 90/(1500 - 600) = 0.1 degree/us change in pulse width.
I am not calculating the angular velocity, only calculating number of degree the servo will travel for a 1us change in pulse width
. Hence the number
0.1 Degrees/(us change in pulse width). Should have made that more clear.
Underlying assumption is that the variation is linear through out the range of motion.
But the servo has 8us dead-band, hence finest movement possible is 8*0.1 = 0.8 degrees.Dead-band is different for various servos. And if you apply any force, it will never be within that dead-band.
8us dead band is for HS422, the servo used by the original poster. I agree other servos will have different dead band.