I recently went through a couple of sessions tuning a couple of 30K scanner pairs that I received from Omar. I referred to the thorough tutorial presented by Buffo which he posted aprox 4 years ago and based on a workshop by Bill Benner. I followed the full stool/stair analogy and this does take patience because the effects of the pot adjustments are not perfectly consistent or limited to just the effects described. But, it works. The first scanner had arrived fairly well tuned and was useable right out of the box, but was better after the tuning process. I then ran the scanner through about 10 additional test screens some provided with Beyond and some not and additional optimization was easy by adjusting the control that was the most directly related to the defect seen. An example of this is how the Laser Media test pattern revels a speed discrepancy between X and Y.

It occurred to me that as long as the scanners are useable ie. pretty close, then why run the pots so far down and up when the optimization may only require a NET change of less than a single turn of the pot. I am aware of the argument that the interaction of gain and damping may make the necessary correction counter-intuitive and so that is why I went ahead and tuned the first set the way I did. However, after seeing the substantially easier, safer and very effective method of using patterns that each highlight different defects, I decided to tune the second scanner set using only the patterns. After running through this list several times (about 5min.) the second set was every bit as well tuned as the first based on a comparison of the ILDA 12K and the L.M. test patterns and about 6-7 others as well.

I think this is a little like aligning a multiple diode laser within a projector. The very first time you need to start from diode centering through collimator focus and so on all the way to far field co-alignment. If nearly there, but not perfect a small adjustment of a sub set of the whole optical train is all it takes to dial it in and this is a lot faster not to mention less risky than disassembling the entire beam path.

Then I thought; Why stop there? (as I am wont to do). If the final optimization is so easy, then let's say you are going to run a couple of satellite projectors for only hot beams or sheets or...whatever. You might optimize the scan set for that particular projection type only. My guess is that smoothly scanning a graphic will not have exactly the same requirements as a rapid series of clean tight hot beams.