One thing I have learned from Luxo ii is you start the collimation process with a large doughnut like you have pictured in post No. 6, but then to fine tune you make the doughnut small.
The thing with a large doughnut is whatever collimation error remains the large doughnut will be evening out any error. By then making the doughnut small, these wee errors are then much easier to spot and fix.
For those of you not too familiar with the collimation process, regardless of the type of reflector you have (Newt, SCT, Mak, whatever) this collimation process is done using high magnification. And when it is done using a star, the final check is known as star testing and should be done ALL the time if you use a laser to collimate with as only this process will truly verify of your work with a laser. We should take a leaf out of professional observatories where initial collimation is done using lasers, but verified by star testing, which after all it is stars we are looking at, not a laser dot.