The following is a collimation technique for SCT's described by Jeff Barton Astronomy Director for the 3 Rivers Foundation. He very successfully taught me this technique after replacing the adjustment scresss with knobs on the back of hte secondary mirror. Source Robert Piekerial book on Collminating SCT's.
To begin with, we're talking here about SCTs secondary collimation where the secondary pivots on a center post and there are three adjustment screws. Some older models have a different setup, but this covers most Meade and Celestron SCTs produced since about 1995, and many earlier ones.
You use a slightly-defocused image of a mid-bright star to show a "donut" (bright ring with dark center "hole") and adjust the screws to bring the hole exactly into the center of the donut and have both hole and bright ring(s) concentric.
There are some things to remember:
A) Always re-center the out-of-focus image in the eyepiece after each time you turn an adjusting screw.
B) Use moderate to high magnification. The higher the magnification, the "touchier" this will be, but also the more accurate it will be.
There are a number of variations to this process, but the one I find easiest to remember is this one:
1) Center a mid-bright star in the field of view, defocus it until you see a "donut"-like image, using moderate to high power (50X to 100X is plenty for starters).
2) If the dark "hole" is centered in the donut, you're good to go. Do a star test. If the stars are nice and sharp, and round while very slightly defocused, you're good to go.
3) If the hole is not centered in the donut, use the drive controls to move the star to the edge of the field of view in the direction of the "shadow" ... that is, if the donut is displaced from the center to the right, then move the mount so the star is at the right-hand edge of the field of view. I like to leave the entire donut in the field of view.
4) Find the adjustment screw that moves the donut back to the center of the field of view. Get it as near centered as possible turning just THAT screw and no other.
6) Use the drive control to exactly re-center the star in the field of view. Go to step 20 or 3 and repeat as necessary until you have the dark hole centered in the donut and your scope passes a star test.
That's it. After a few successful tries, you likely won't need much adjustment and generally tweaking just one screw will do it.
I should have mentioned that your success depends on using a diagonal that's accurately collimated, if you use a diagonal. If it's off, then things are gonna get screwy. IF you notice that you have a nice, centered and concentric donut BUT you still have slightly seagulled stars, THEN your diagonal may be off.
Checking the collimation of a diagonal is easy. Recollimating one can be hard, especially if the mirror is held in place with springs and there are no collimation screws. Diagonals from AP, Televue, Lumicon, and Stellarvue are pretty good. Vixen diagonals vary, depending on whether they're shipped with their less expensive scopes. I've found that some Orion diagonals are made by Vixen (you can tell because the bodies are rounded and have access screws behind the mirror-mount panel.
To check a diagonal, you need a cheshire eyepiece and a flat mirror (be careful about using the rear-view mirror on your car or truck as it may be curved ("Objects in mirror are closer than they appear").
Place the cheshire in the diagonal and hold the focuser tube end of the diagonal firmly against the flat mirror. If enough light is getting into the Cheshire through the side cut-out, you'll notice two images of the cross-hairs. One is the direct view and the other is the reflection from the mirror. The closer these two images are to being coincident, the nearer the diagonal is to being collimated.
In most cheap diagonals I've tested (and this includes practically EVERY diagonal I've seen that's delivered with a scope as part of the basic package, except from the companies listed above), the collimation is off. To remedy this, open the body of the diagonal and turn ONE adjustment screw no more than 1/4 of a turn, re-test the diagonal, and note whether it's improved and in what direction.
Get the images as near as possible to coincident, with as little rotation as possible. It is extremely difficult to make it "perfect". Then re-check your SCT collimation.