Tracking
Tracking of stars was smooth in the original commercial design. The encoders on the motors theoretically provided sub-arcsecond resolution which permitted fine control of the tracking velocity, but gear imperfections limited the absolute pointing accuracy.
With my new encoders on the axes, I had a resolution of only 34 arc seconds per encoder count. They provided absolute pointing knowledge, but at the price of limited position resolution. Because I was only aiming for 2 arc minute accuracy in my final catalog (similar to Tycho's), the 34 arcsec/count was not a real limitation for position determination. However, the earth rotation rate is 15 arcsec/sec so this implied updating the motor control loops as rarely as every 2 seconds. This would induce a jitter in the pointing of about one-half an arc minute which would ruin any imagery, but I had no intention of collecting images. In operation, the inertia of the scope smoothed out the az tracking to the point that it was unnoticeable. The position jumps remained noticeable in the altitude tracking.
For meridian altitude measurements, the star would need no alt tracking as it passed the meridian. Similarly, any jitter in the az axis would be east-west and thus not affect the altitude measurement.
For separation measurements, the two star images on my detector would jitter around in unison, so I could manipulate their relative spacing unambiguously. If either star was rising/setting rapidly, the altitude position steps were quite noticeable.
I considered mixing both the axis and the motor encoders together in my control loop, but I was not OCD enough to do it. ('better is the enemy of good')