over the last few days, I found some time and got my dome set up properly for testing this stuff. In the process, I found an initial small error in the indi calcs as compared to the spreadsheet ferran has posted earlier, pull request on github now with a correction. But it's not completely right now either.

I did a bit of a write up on our blog, with photos and a description of the setup. What I have done is mount a webcam under the telescope at the mount, then put some alligator clips on the telescope to provide a target aligned with the scope. Then I re-did my optical offset measurements to line up with the aligator clip rather than the center of the telescope. I have also marked the center of the slit on our dome, so now I have a setup in place where i can slew the telescope anywhere during the day, and just look at the webcam to see how it lines up with the dome slit.

My write-up with photos is here:- astro.rozehaven.ca/astro/?p=197

As you can see, it's better than it was, but, not yet perfect.

Our dome has pointing accuracy of about 1 degree, tested over many slews in various directions. It's an NexDome, driven by a stepper motor system for which I wrote the firmware. I've done a lot of testing using various slews all over, then back to the home sensor to see how much slippage there is in the system, and it's not much. Rarely is the calcuated position off by more than a degree when it comes back to the home sensor.

In doing my initial testing with this, one thing I did learn, measurements of the offsets inside the dome are very important to be very accurate. Changes of just a couple cm can go from 'telescope lands in slit every time' to 'telescope often looking at the dome panels'. In our case, I'm using a 13cm telescope and pointing at a 60cm slit with a dome radius of 1.18m. The setup now is essentially 'good enough' that it seems to get the telescope into the slit every time, but, the second dome is going up next week, has the same geometry, but it's going to have a 30cm telescope pointing at that 60cm slit. For that one, I'll have to find the issues in the math and get pointing dead accurate, it has little extra room for error.