This is the first night I tried using the Deep Sky Instruments guide to collimate my AT6RC.
This technique proceeds in two steps:
In this step you first defocus a star to produce a "donut" and then study it's symmetry. If the primary is out of position then one part of the donut will be brighter than the rest. Once you identify which part of the donut is brighter then you adjust the corresponding primary mirror screw. The DSI document describes how to rotate your CCD camera so it is easy to figure out which screw to turn. This was my first time doing this and what I found is I had to adjust the opposite screw from what the DSI documentation suggested. This might be due to the fact the adjustment screws on a Deep Sky Instrument RC scope work differently that on the AT6RC? I'll have to keep this in mind in the future when collimating this scope.
I setup on a bright star and defocused until I had a donut around 64 pixels across. I am using an SXV-H9 CCD camera and Maxim DL. I set Maxim DL to take continuous focus frames and then started adjusting the primary screws until I had a reasonably symmetric looking donut:
For the second step the DSI guide suggests taking a long exposure of a part of the sky with lots of stars. Then you choose stars around the periphery of the image and evaluate the shape and orientation of the donuts. I took a different approach. I wrote a Python script that interacts with Maxim DL and the telescope ASCOM driver to move a bright star around on the CCD frame and take exposures. This gives me consistent brightness donuts to look at all over the frame. For now I am using a 3x3 array of star images. I have Maxim DL stack the 9 frames to get a single frame like this:
You can see the 9 images of the bright star. The center image looks reasonably symmetric so we have maintained the on-axis alignment we setup in Step 1. The 8 images around the edges are not symmetric, but this is OK according to the DSI guide. What is important is that these images are "balanced", using the term from the DSI guide. In this case all the images appear to be pointing towards the center of the image, indicating the image is balanced and the secondary mirror is now aligned properly.
At this point the collimation procedure is complete. If in the process of balancing the outside images the center image became un-symmetric then you have to goto Step 1 again and fix the primary mirror, then check the outside images again. This process is iterated until you are satisfied with both the central and outside star images.
The sky conditions were not good the night I attempted this collimation. There were thin clouds and seeing was predicted to be average at best. Nonetheless I took some test frames of NGC 2158 before shutting down for the night. Here is some CCDInspector analysis of the test frame:
The FWHM was around 2.7 pixels or 4 arc-seconds. I'm reserving judgement at this time because the sky conditions were marginal. The off-center minimum shown in the CCDInspector graph seems to indicate I haven't quite got thing right but it appears I'm on the right track.
Here is a stack of 7 frames.
My next step is to work on a Python script that will analyze the frame showing the 9 star images and overlays arrows indicating the orientation of each star image. This should give nice feedback for which secondary screw(s) to adjust.