This was a trial run on a new guiding system that cancels out most of the differential flexure between the C8 HyperStar imaging system and the Televue Pronto guide system. Interesting results. I'd say an improvement, but this particular region was more difficult to image than I had anticipated. The region is to the South West overlooking downtown Tucson, about 5 miles away. So lots of light pollution. Imaging occurred over elevations from 40 deg and downward.

This was about 1.4 hours at F/2! Offhand, it looks like it needs even more integration time.

---

Let's do something about all that noise in the background. The first two wavelet layers are adding essentially nothing to the image, and the 3rd can be pulled back a bit. Less noise is the result, and now you can clearly see another edge-on spiral galaxy in the SW (lower right) sector.

---

So... autoguiding. With the C8 HyperStar imaging system there is nowhere to place an OAG. So I have to use a separate guidescope and camera. My system has an ancient Televue Pronto 70 mm F/6.5 telescope put to use as a guide scope with a Lodestar X2 camera. There turns out to be significant differential flexure in the system, that produces a fairly steady and predictable rate of changing flexure during the course of a one hour set.

So I wrote a little side utility that runs in a forever loop, kicking the autoguider every so often to tell it that the target has moved. I make sure that I never kick more than 1/2 guider pixel at a time, and let the autoguiding glide smoothly to the new target position.

By taking two exposures of a bright star separated in time by 150 sec, or 300 sec, all the while under autoguiding, we can measure the centroid of a star in the two images and from that difference in position, determine how much extra we need to move the autoguider target in the opposite direction to compensate for the differential flexure.

There are two modes for using this: (A) taking long exposures, and (B) taking better quality shorter exposures. Both work and are useful. But in this image test of the Witch Head Nebula, I decided to go for shorter exposures under more accurate guidance (choice B).

During the run, I can update the amount of differential movement in each axis and the utility will use those new values compounded on what it is already using, compute a new update interval (typically every 50-70 sec), and continue on making guider corrections. That's how I made the image above, keeping the guider errors in the actual image plane to about 1 pixel in each axis. That's a fraction of the FWHM for the imaging train.

WARNING: Once you have more accurate guiding, the stars stack up better, and start showing the ugly truth about your collimation.