Finally, here it is the last DSI image. Essentially the above image defocused to show you all the donuts, how they looked at the end of collimation.
I must admit that I found interpreting these donuts to be difficult. Perhaps that was in part because I was initially using too low a resolution (binned 2x2). Moving the binning back to 1x1 for these extended stars really hepled, especially on the brighter stars where I could see the spider viens and more details on the primary mirror, some concentric rings etc. Still they move around a lot under the seeing and transparency changes during the night.
So I now pronounce this telescope collimated. I hope some of you got some insights from these posts. Good Luck and regards
Congratulations for that kind of end result!
I think you could stop reading here and skip the rest of my reply, because it could spoil your day...
Anyway, here comes some explation for the DSI guide and those defocussed star donuts.
They say this in the DSI guide:
In a well collimated, rotationally symmetrical optical system, any resulting images should also be rotationally symmetrical. We call this “balanced”. In a balanced system, off‐axis stars will converge to
focus uniformly to produce pinpoint star images.
All of the perimeter star images on the right are roughly round and similar. They are rotationally symmetrical. That is, they all look about the same from the perspective of the center of the image indicating good balance. The on‐axis star image 5 is uniformly illuminated indicating no obvious on‐axis coma. This system is well collimated.
You can watch all that happening in this DSI image below:
(I think your image was not unfocussed enough to see all that, but let's forget that at this stage)
I have drawn some lines into your DSI image (attached here as the last image). You could draw (virtually in your own mind) similar lines into image above and notice how those are rotationally symmetrical and the center area stars are uniformly illuminated (just like DSI guide tells us). You don't need a high resolution image to see this (the above image is 4x4 binned only, because I had a very slow image downloading QSI 583ws CCD camera over then and 4x4 binned images downloaded quickly enough for my collimation purposes). What I have done for that image is this: I have altered it's screen stretch function too so that I could see only brighter stars and their unfocussed areas more easily from those oval stars' peripheries. By unfocussing more and altering screen stretch in my imaging software, that kind of DSI image becomes easier to understand.
(BTW, you can see there the center area star's Arago spot too.)
In practise, there is a radical difference between Neil's image and its' center area (latest image here) and my image above for collimated secondary/primary mirror. In my image all unfocussed center stars over there are round, but in Neil's image those are slightly oval. That tells us that his RCT scope is not fully collimated by the principles of DSI method. I'm sorry to tell you this, but DSI method is quite ruthless in this perpective and it will tell every tiny error in RCT collimation. This gives us also more understanding to akulapanam's comment here saying that "GSO scope collimating that way is virtually impossible".
As a result I could say quite frankly that Neil's scope is not a rotationally symmetrical optical system like they describe there in DSI guide.
What explains this? For example your focuser axis might not be perfectly aligned with your RCT scope's optical axis and light goes into zig-zag ways, before it gets to your camera in this RCT focuser. Or your primary could need some minor adjustment combined with secondary re-adjustment etc. etc. etc. (which could never be fully completed with DSI principles, unless he does some minor adjustment to his RCT's mechanics (link)). So its' quite likely that Neil has run here to the fact that GSO RCT scope's mechanics might be just a tiny little bit deficient (incomplete) for this DSI method to work out as it supposed to work. I don't want to spoil Neils party here, but he's quite right there saying that his RCT is now collimated up to 96% (=not fully 100%, which is a very rare achievement with these RCT scopes).
I'd suggest you should just continue using your RCT as it is right now, because it can give you such good looking stars all over your imaging field.
Forget what I have told you above and enjoy from your scope, because going forward from here is going to be very time consuming and extremely arduous task. And this will require doing more mechanical checks for all optical parts ie. perhaps disassembling your RCT scope. Come back to DSI method, when you change your focuser for this scope, unless you're willing to spend more night hours maybe fruitlessly jogging around with DSI method. I have been there, so please take my advise.
I'll show below an image, which shows different star shapes from my own RCT scope during several years of collimation adjustments for it (stars are in no particular order there):
When someone gets' his RCT scope into over 98% perfect collimation, he could achieve such pinpoint stars as seen in the lower part of that image (seeing is one variable there together with RCT collimation accuracy).
Second row of stars shows good examples from a similar RCT stars we have seen in this thread. But chasing that kind of perfect pinpoint collimation from a GSO scope could literally spoil your astro life, so I do not suggest this to anyone wanting to use their RCT into normal imaging (etc.) activities. If you want to get your RCT into better than your current collimated status, that could mean that your imaging hobby becomes a hobby of collimating a RCT scope...
I have been there and I could say, that it really gives you pleasure to watch those 96% accurate RCT star shapes suddenly settle for over 98% accurate RCT stars with a less than tiny adjustment to some random collimation screw in your scope. In that situation your stars almost suddenly "improve your seeing" and snap into focus. But depending on your RCT mechanics, this does not necessarily mean that all four corner areas in your imaging field will have round stars...
Says a man who has achieved this 98% collimation status for his RCT scope, but never that 100%
PS. My desire for perfect stars was achieved by getting a high cost TMB 152/1200 APO LZOS refractor with its' really pinpoint stars.