First I used 3 sec unbinned exposures and wondered how the tracking error was about 0.6 to 0.8 pixels. Then I realized that one noisy pixel on the edge of the guide star made the difference. After that I changed to 8 sec exposures and noticed that the error graph changed dramatically with lower noise (or with better SNR). After some fine adjustments of the guider aggressiveness I got the unbinned guider error below 0.5 pixels which is enough for my setup (STL pixel size vs. internal guide chip pixel size).
Finally I changed 2*2 binning to the guider and started to use 5-6 sec exposures and got tracking error to about 0.2 pixels.
This kind of graph I've never seen on my previous EQ-5/6 mounts... So no worries of the mounts performance for this small scope!!
