...I tried using a laser collimator, but found that the loose tolerances of the fit in the focuser creates...error...as much as two inches for the laser reflection on the primary.
Sure. First, I hope the loose fit is caused by the laser and not the focuser, because if it's the focuser, your eyepieces will be subjected to the same focuser axial misalignment--not good. Now, if it's the laser's fault (and it usually is), and the fit is really that bad, the best option, IMO, is to replace the laser. Some users advocate taping the laser barrel to get a tighter fit, but what you really need is a barrel that mimics an eyepiece barrel, that way when you secure the laser in the focuser, it's axis mimics an eyepiece's axis. I also suspect you're using a 2-inch focuser, with a 2- to 1.25-inch adapter, and a 1.25-inch laser. A better 2- to 1.25-inch adapter should improve the fit and the alignment consistency. How the eyepiece fits the adapter, and how the adapter fits the focuser drawtube, is called "registration". It's important to verify the focuser axial alignment with the locking screws secured, the same way you would use the locking screws with an eyepiece. This "fixes" the registration. If the misalignment is repeatably consistent when the locking screws are secured, your focuser axis is indeed, misaligned!
If you have a dial caliper or similar measuring tool, you can quickly determine the cause of your registration inconsistency.
...Tightening it down with the set screw shows the primary to be *way* out of alignment, despite the fact that I can visually confirm good alignment using the peephole in the collimation cap.
"Tightening it down" only secures the registration of the laser to the focuser. Remember, the laser beam defines the focuser
axis. The return beam is simply a reflection of the focuser axis. The only way the return beam can be used for primary mirror axial alignment is if the focuser axial alignment is perfect
. One half of any residual focuser axial error will be propagated forward to the laser beam emitter. I'm guessing this is what you're seeing when you say the primary is "*way* out of alignment". The collimation cap, which is relatively insensitive to focuser axial errors, shows the primary mirror alignment is in fact, "good", so you're back to where you started: good secondary mirror positioning, good primary mirror axial alignment, and unknown focuser axial alignment.
Of course, even without a laser, we know your focuser axis isn't grossly misaligned, even though it may be out of tolerance. As long as you're happy with your scope's performance, you shouldn't lose good observing hours worrying about what might be wrong. Learning how to collimate your scope is all about patience and persistence, an exercise best saved for rainy days and cloudy nights (pardon the pun).