Why is this news when it comes to ZWO? :-) :-)
BTW, after buying some EOS and Nikon couplers that can use filter drawers/filter wheels, I just figured out that there is no need for filter wheels and filter drawers, if I can measure the difference in EAF step from a Bahtinov focus between a star on the optical axis and a star on the corner. You do need a halfway decent electronic focuser. The high end QHY Q-focuser is probably an overkill, even. The ZWO EAF probably suffices.
You only need one filter with a known glass thickness! Together with a filter holder that is empty (no glass).
Why?
OK, take a look at the numbers from last night. Assuming that the EAF step sizes for no-filter and 2mm filter are accurate (fuggetabout ASIAIR; it simply is useless for anything to do with accurate focusing).
I had measured EAF steps of 21248 for no filter (glass thickness == 0), and 21700 for 2mm filter. That is a difference of 452 EAF steps (for my 0.8MOD geared EAF) for 2mm glass (or 0.667mm of mechanical backfocus). This comes to 1.476 µm per EAF step (0.001476 mm, but easier to think in terms of µm at these kinds of scale).
What we do is focus a bright star at the optical axis really well with no filter; note down its EAF step size. Now find a bright star at the corner (say 12mm away on an APS-C frame; an APS-C frame is about 28mm from one corner to the diagonally opposed corner, so corner is about 14mm from optical center).
Refocus on this bright corner star (this is the part I am not sure I can do with the ZWO EAF's sloppiness, will have to test during clear nights). This is where you want to use the best Bahtinov mask that money can buy.
If this is do-able, note down the EAF step difference between the center star and the corner star. Make sure to record the sign of the delta. This is the EAF delta for no-filter case.
Repeat the above with a 2mm filter. This is the delta for the 2mm filter case.
Now you have two EAF deltas, one for a 2mm filter, and one for no-filter. You can now interpolate (or extrapolate if the backfocus is more) for the equivalent filter thickness that produces an EAF delta of precisely 0.
Remember, the whole exercise of backfocus adjustments comes down to finding where the image plane is flat (i.e., no difference between focus of center star and focus of corner star). Forget about the "stars have radial pattern means backfocus is too short" nonsense -- we go back to the original mathematical definition, and not some hobbyist rule of thumb.
I.e., if the corner-to-center EAF is +100 steps for the no filter case, and the corner-to-center EAF is -200 steps for the 2mm filter case, using linear interpolation, the filter that produces +0 EAF steps would have a thickness of 0.333*2mm.
Go back to our scale above to now convert this to the number of µm of mechanical backfocus.
Done. Punkt. All with one filter thickness (plus empty filter holder), an accurate Bahtinov mask, and an electronic focuser that has no slippage and as small backlash as possible.
No need to add shims, rotate camera angles, reinspect the corner stars, and all that nonsense. Just use good old mathematical linear interpolation of what we see as field flatness of the corner stars. Just four measurements: EAF for focused center star with empty filter holder, EAF for focused corner star with empty filter holder, EAF for focused center star with 2mm filter, EAF for focused corner star with 2mm filter.
I'll bet the four refocusing and EAF measurements take less than 15 minutes, and just use a slide rule after that.
And it works especially well for camera lenses where you cannot insert any mechanical backfocus tools (most of them will rotate camera angle too, although there is at least one that does not).
You may have to make a couple of sets of measurements to average out the errors from ZWO EAF's sloppiness.
Chen
