Automatic Refocus Troubleshooting

[gtdulaney]

I am getting hit or miss success, mostly miss when using automatic refocus.

I've recently switched from a RASA8 to a William Optics Cat 91 WIFD. I have no experience with refractors.

The WO Cat 91 is paired with a Pegasus Astor FocusCube V3 on the course focus knob.

I ran a sequence with autofocus only to establish the baseline. This first Autofocus worked. Since this first sequence autofocus I've been trying to use automatic refocus with Multi-star FWHM Measurement. My success rate is probably 30%, the remainder of the attempts fail. When I do get success in refocus the calculated best focus varies substantially upwards of 200 steps on the Focus cube, this converts to 8 to 9 degrees of focus knob rotation. The results bounce around alot, 9657, 9695, on up to 9800. The FocusCube is 0.0625 degrees per step.

I switched to the Bahtinov mask similar results, sometimes Sharpcap gives a "best" focus point, other times it fails. When it gets results again they don't seem repeatable.

I've tried finding the logs for these attempts but have failed to find them. The focus log window shuts down before I can try and copy and paste.

I don't know if this is the way refactors work. But does not seem correct to me. I need help getting collecting the data from the focus runs for troubleshooting. Any advice would be appreciated. I'm concerned my new Cat 91 is a lemon.

Greg Dulaney

[admin]

Hi Greg,

the trick to troubleshooting is to bring up the multi-star FWHM focus score *first*, then use the 'Refocus' button in that panel (under the Goto 'Best Focus' button). That will run exactly the same refocus procedure, but it leaves the focus tool open afterwards so that you can inspect the graph and see how good it was. It's the details of that graph that will determine success or failure and accuracy of your focusing...

The things to look out for in the graph are that there should be *at least* 3 points on each side of the best focus point, and that the 'U' shape should be significant - that is that the height of the U is significantly more than the random variations in measurement, so that you get a fairly clear U with a little randomness, not a 'just about might be a U if you squint at it' with a lot of randomness.

Things that can cause a poor graph include

* not scanning over a wide enough distance so that the edges of the scan are still pretty close to best focus. When using the focus scan, the stars should be significantly out of focus at each end of the scan - maybe if you are aiming for a best FWHM near 3 or 4, the values at the outside edges of the scan should be 10-15 or even a bit more

* Not having dealt with the backlash in the focuser movement - this is usually best dealt with by setting the 'overshoot' option in the focuser hardware settings and then picking an overshoot value that is a bit more than the amount of backlash in your focuser movement. If the backlash is not being dealt with then the adjustments that SharpCap is making to the focus position could just be moving around inside the backlash range, meaning that there would be no signal in the measurements of FWHM, just noise...

Both of these can apply to the other focus tools too, as well as FWHM.

If you are still stuck, grab a screenshot of the graph after the refocus fails and post here and I will be able to advise further.

cheers,

Robin

[gtdulaney]

Thanks. Clear skies gone for the next few days, which is good because my patience has departed as well.

I will incorporate your recommendations and grab a couple of screen shots.