I forgot to mention another thing that may be confusing to users. The setting of the backstop has no effect, none whatsoever, on the axis backlash. All that the backstop does is prevent the gear teeth of the worm and worm wheel from backing off when large amounts of unbalanced load is present on the axis. This is to prevent the teeth from fully separating and possibly running across each other when the unbalance pressure is too high.

Periodic error is caused by slight eccentricity of the worm gear and also the eccentricity of the spur gear on the end of the worm shaft. This is a fixed mechanical thing that cannot be changed by mount crashes or software issues or anything else you can think of. Even if you dropped the mount on concrete, you would not disturb the natural PE of the mount. 

When evaluating tracking performance, you first have to evaluate the seeing in the air above you before trying to chase down blind alleys after dead fish. You simply need to run your guider program for about 1 minute to evaluate the seeing. Turn off all guide pulses so that the program is just recording the motion of the guide star on your guider graph. Set the guide exposure to 2 seconds. Then look at the Dec axis to see how far the star moves around as each exposure position is shown on your guide graph. What you want to see is the peak excursion of the guide star in both axes, ignoring the steady drift that may be occurring. If they are similar, then the RA is tracking properly and not throwing excess variations into the movement of the RA axis. 

It also tells you what the actual seeing is, and you will not improve on it no matter what. This is the baseline seeing that will determine what your guided rms values will be. The Dec axis is the key to knowing the seeing conditions because it simply does not move, so it is a perfect determinant of your local seeing at any time.

Rolando