I have GSO 10RC, and hang STT8300 + self guiding filter wheel + AO8T. Quite a bit of weight.
I found the same flex mentioned above as being corrected with the truss tube focuser/image train directly to the back plate. To check the flex I used focuser + Howie Glatter laser. Pulling down on body of focuser approximately the weight of my image train, the laser dot moved from inside to almost outside the secondary donut. When pull down same amount on draw tube it definitely moved outside the donut. I think collimation will be changing with different OTA pointing positions.
To correct this:
1. I designed and built an isolator part the bolts directly to the back plate and goes around (1/16" clearance all around) the original tube the focuser screws to. The isolator has same threads, so extensions/focuser attach nicely.
2. The GSO 3.3" focuser is nice for visual or DSLR AP, but not up to the task I need. So I replaced it with Moonlite CSL focuser with 68mm threaded connection to STL thread of AO8T. Now everything rock steady... Same tests as above (Glatter install to focuser), and the laser dot stays inside the donut. Maybe moves a slight amount when pull down on draw tube. Push on the image train, and it feels rock steady.
Now the system pieces can be independently collimated.
1. Use Glatter in 2" adapter in Moonlite focuser and rotate Glatter only to check if it is collimated (was perfect)
2. Rotate focuser (it has collimatable rotation rings) to different positions and adjust the focuser collimation so the laser dot is not drawing it's own donut on the secondary.
3. Shim the new isolator part such that the laser dot is in the middle of secondary. (Now the axis of focuser is pointing directly at donut of secondary).
4. Check step 2 and then 3 again.
At this point I can confidently rotate the focuser, and everthing stays aligned.
5. Adjust secondary so laser reflects directly back on center of the Glatter. (Use your trusty LED flashlight to light up the end of the Glatter and see the reflected beam better).
6. Rough check primary collimation with Glatter concentric ring hologram.
7. Star test using eyepiece (if doing visual), or add image train and star test collimate secondary and primary. (Deep Sky method)
I am waiting to do Step 7, but that nasty rain from 2 days ago hurricane (typhoon over here) keeps coming and coming.