The motors are attached to the worm. The assembly tilts around the axis on the motor side. The other side is clamped down with a spring, in my case a piece of bicycle tube rubber. So, it is a motor-integrated one-piece worm assembly with SLW (spring-loaded worm). The pivot axis is tightened at the bottom with 2 small Belleville washers for some springiness. It was quite easy to assemble; one aluminum edge and 2 small aluminum plates with some holes drilled, glued together with metal epoxy.
I have been imaging with it for quite a while and it seems to work OK but not great. The best I can do is usually 0.7" RMS total but that is rare. About a week ago, when the seeing was good I got 0.5" RMS total - just where I want to be, and it was stable! Being inside an observatory makes it so much nicer to test than on the driveway out in the open.
So, I set out to do some testing. First off, I noticed that my calibration curves looked horrible mostly because the step size was too small. I adjusted that in Ekos and the back-and-forth curves had different angles. Looking at the assembly, I found that I had been way too careful with tightening the pivot axis down. Tightening it to the point where the friction started to set in, I finally got nice calibration curves.
I need to add that I had been working with a slightly off-balance
Another thing I noticed is that my ASI120MM shows a jumpy behavior when sampling at 2 seconds. Short jump, large jump and repeat. At one second it does not show that behavior. When I started sampling at one second with 2x2 binning and it looked a lot better. I really need a better guide camera that supports USB3 not just USB2. Last week I ordered a QHY 5-III 200M that will arrive today.
Altogether I think I am close to reaching my 0.5" total RMS. Much more cannot be expected because of the direct worm drive with a 0.9 degree stepper at 32 microsteps. One worm rotation of 360/0.9 = 400 steps is 1 degree for the G11S, so that is 3600/(400*32) = 0.28" resolution. That calculation should hold for