I have a Canon EOS 600D which I have modified for astrophotography by replacing the stock UV/IR filter with a Baader BCF filter (this extends the spectral response to >656nm so that Hydrogen-alpha emissions from nebulae can be captured). Unfortunately although the modification was a success in that I have a nice clean image and the camera functions correctly, the micrometer measurements I took of each of the 3 spring-loaded tilt adjustment screws for setting the tilt and level of the sensor assembly were out somehow. With hindsight I should have used the method of counting turns to completely tighten the screws and then loosened them off that much on reassembly, but it's too late for that now. I know the sensor has negligible zero tilt because by focusing on a piece of paper perpendicular to the camera using live view I can achieve an image that is focused throughout even at f/1.4, but it is either too far back or too far forward so that the focus is out if I focus using the standard auto-focus sensor. I figure I can work out whether the sensor is too far forward or backward by focusing on an object using the standard sensor and then without moving the camera or object seeing which way the focusing ring moves when I refocus using live view, but is there a way I can work out (say from the size of an airy disc of an image of a point source at infinity like a star) how far forward or backwards the sensor needs to be moved?
(Note, when answering you may assume I'm mathematically competent and (except for my stupid lapse here) sufficiently technically competent to make the necessary adjustments, I'd just rather figure out mathematically how much to adjust the plane of the sensor (and by extension how many turns of the screws that equates to) so I can perform surgery once, rather than many times in a kind of trial and error, as some of the ribbon conector seats feel somewhat delicate.)