One more thought. I really do not know what happens after the guide pulse. I guess encoders got updated, and the effect is the same as would be if we moved the mount with slew commands. But this is just my guess. It is possible that the MC hides this update and subtracts the guided angle from the reported position. In other words: if the position on the axis does not change after guide pulse, we are good. Otherwise, we need to do this subtraction in the driver by keeping track of the guiding shifts - this may be tricky but seems doable (the guiding pulses are simulated anyway, thus there is a strict relation between length of the pulse and actual number of steps shifted).