When I calibrated Robofocus, I needed to move the focuser in as much as possible. I completely underestimated how much force Robofocus applies to the focuser. When I was done, the focus knob was completely locked! Neither with Robofucs nor manually could I unlock it (aka move it out) :-(
I asked at the Celestron Edge HD user group, Celestron support and also Robofocus support. Celestron and Robocus support didn't reply at all. From the user group I got two tips: a) loosen the mirror locks at the back of the scope, and b) contact Ed Thomas from Deep Space Products.
Ed Thomas gave me the additional tip to loosen the screw in the focuser itself.
I did both and then used Robofocus to try to move the focuser out. After the belt slipped a few times, the focus knob indeed came loose. I then removed the Robofocus knob and could turn the focus knob easily manually. Phew!!!
So, back to calibrating Robofocus. First, I set the backlash correctly. The Robofocus control program is a little weird in that it doesn't have a "Save" or "Set" button. Most settings seem to be set when tabbing out of the edit field (e.g. step size). But the backlash value wasn't. I used the "Data Traffic" window which shows the data traffic between computer and the control box. And indeed, nothing was sent when I tabbed out. After playing with the form, I figured out that one has to press <Enter> for setting the value. As recommended, I set it to 125 (1000 / step size) for my SCT scope. Then I moved the focuser all the way out again. And then Robofocus moves it in until you stop it. This time, I was on guard and stopped it immediately. Yes!
Unfortunately, the stepsize was too large (8) and I ended up with a max travel of 22198 (Robofocus can have up to 65000 max travel). So, I moved the focuser all the way out again. Set the step size to 4, backlash to 250 and started calibration again. I nervously observed what would happen when I hit position 0 again. But it seemed as if the focuser was indeed only half way in and Robofocus moved into negatives without any problem. At about -22000, the belt slipped again and I stopped the calibration run. Strangely enough, the max travel was still at 22198. It should have been roughly double that. I moved the focuser back to position 20000 and it was now roughly half in. So, I set the max travel position manually to 44000 and started to move the focuser further out - carefully observing if the belt slips again. At around 43600 the belt slipped again. I moved the focuser back to position 42000 and set the max travel position to 43000. This should be enough (I doubt that I will need the extreme in or out position anyway).
Now, I'm ready for the first real usage - the sky is still clear, so I'm hopeful that I can finally try it out tonight!