I guess I have a question on the relevance of the autocollimator given the discussion in this thread. Unfortunately this will be somewhat long.
I have "quality" tools including the Glatter laser with the 1mm aperture stop and barlow attachments. I have a Catseye sight tube, the Blackcat cheshire, and the old, one pupil Catseye autocollimator.
Skipping the sight tube adjustments, my understanding for rough collimation, is that you align the laser to the center spot on the primary by using the knobs on the secondary mirror. Then you adjust the primary mirror knobs until the return hotspot image is centered in the barlowed laser aperture white face, or the white cheshire ring. For finer collimation you could then iterate the cheshire and autocollimator to stack the hotspot or triangle reflections keeping the hotspot centered in the cheshire.
I think if I understood Vic correctly, he often tweaks the secondary so that the return beam of the laser with the 1mm aperture stop coincides such that you see the laser line up and also see the diffraction rings on the face of the laser's white face. I took this as an alternative to using the autocollimator in the middle of the night when one is too tired to get the red flashlight to tweak the secondary.
There has been skepticism expressed that star collimation is largely useless given good tools or is beyond the skill set that most of us have. Jason gave some examples of trying to utilize star collimation and then check the results using the Blackcat cheshire. His first night out indicated that star collimation was hard and his images illustrated that he got close to decent collimation, but certainly not perfect collimation.
Mike Lockwood then chimed in for Jason to follow his method as outlined on his website. Mike argues that star collimation is not as difficult as it seems once learned and that it leads to better results. He states that his method should be used after rough collimation with decent tools, and that he mostly uses a laser (assuming Glatter) and does not own a "quality" cheshire. He also said that he tweaks the collimation with a coma corrector in place under a high power eyepiece by adjusting the primary mirror knobs. Because of the various registration errors (paracorr tightened in the focuser, the eyepiece in the paracorr, the quality of the focuser and structure, etc.), I believe he is saying that final star collimation will give the best results essentially incorporating all the "issues" of the optical train into the final collimation.
In Mike's words, "Lasers have improved greatly (thank you Howie Glatter and others), but still there are errors introduced by imperfect tools, the fit of the tool in the focuser, and other small errors. No tool is perfect, but some are better than others. So, the only method that avoids all of the errors is final collimation using a star."
Jason then re-read Mike's treatise and tested his skills yet again. I think his take was that he understood star collimation better and his results after comparing to the Blackcat cheshire were indeed better. Further he indicated that his test verified that his hotspot was placed correctly and the optical center of his premium mirror was consistent with the mechanical center . He indicated that the tools (Blackcat cheshire) and Mike's method were consistent with each other. But the tools are easier to use. I don't believe Jason used a laser or autocollimator to check for end collimation in these tests, rather just the Blackcat cheshire.
Nils says, "I have said it before - the only meaningful purpose of star collimation that I can think of is to verify the placement of the primary's spot. You might be the first (bold mine) to find that a spot carefully centered on the blank is not at the optical center."
Nils also asked a star tester whether he checks his star collimation afterwards with a quality cheshire for consistency.
Don P says. "Jason, I could be wrong, but I don't think he (Mike Lockwood) uses Catseye tools, especially the autocollimator. He did say he didn't have a good cheshire. When I follow up the Glatter laser tools with a Catseye AC, I always find residual errors. It seems that Mike may be using a star to do the same thing."
Nil's responds, "Don. If instead you follow up just with the Blackcat, what do you find? The limitation of the AC is that it is pretty useless for collimating the primary! It may be a bit better in showing the focuser axis collimation, but not really that great)."
So here we have very experienced folks not necessarily in agreement on some of the finer points. Nils seems skeptical of using star collimation to collimate and seems to indicate that the Blackcat cheshire and the hotspot would do the job for the most part.
Don P is skeptical that most of us can accurately collimate using a star and that the tools are better/easier. Further he fine tunes his collimation to eliminate residual errors with the autocollimator.
Mike Lockwood indicates that only collimation on a star will give perfect collimation because the tools/scope mechanics are not perfect which is not to say that quality tools are not desirable or that you don't want a Feathertouch focuser. Collimating on a star with the end optical train as one would use it would at least adjust for the errors in the train (at least as I interpret it).
That would also imply to me that tweaking collimation with an autocollimator might be somewhat futile in the sense that once you take the AC out and put a Paracorr and eyepiece in the focuser, you have re-introduced errors that mess up your tweaking. I know from personal experience using an AC that you just need to put mild pressure on any one component in the system and your stacks will become un-stacked. I further assume this is why Nils is a little dubious of the value of the autocollimator versus just using the Blackcat cheshire.
Of course, Jason perfected the AC to its current state. He was dubious that you could get accurate results by star collimation and is likely a little less dubious now having tried and tested Mike's methods. But in practice Jason says quality tool based collimation is consistent with and easier than star based collimation. Not sure at this point whether Jason thinks that the AC would add anything over star collimation although I would assume that he would say to use an AC if the seeing supports it.
Of course, the title of this thread is, "Common star collimation mistakes" for beginners who are attempting star collimation.
Personally I am not sure I want to go through the process of using and implementing star collimation. I find the process of aligning the secondary with a laser simple to implement. I prefer the cheshire to the barlowed laser as I can read and interpret it better (as I said I own the Glatter barlow attachment and the Tublug). And I then tweak it with the laser and 1mm aperture stop. But sometimes I am a little over obsessed with collimation and after reading this thread am wondering if the autocollimator offers much value, especially given what I have pointed out above. And especially that the folks I have quoted are all gurus in the art of collimation. (I apologize if I misrepresented someone's point of view.)
So is the AC necessary from a practical perspective?