beachbum I do have it set for 50 steps, and I am using APT...

No, don't rely on the default step size.

Next time you have dark skies, try to do a manual focus sweep, and measure the average FWHM or HFD of stars in the sensor frame for each of the focuser positions.

When you plot the FWHM by hand as you change the EAF setting, you should see a V curve that looks like this (my own focusing program for macOS):

Try to get a sweep that starts with a FWHM of over 8, let it go through focus and then watch it rise towards 8 again. If it keeps rising when you start, you are moving in the wrong direction.

Notice the minima where the curve shallows out (the curve transitions from the geometric optics region, hyperbola focus region with straight line sides) to the Fourier optics region (diffraction limited) where the FWHM does not change very much as you move the focuser.

The focus of a star does not improve around the "critical focus" region (surrounding 53062 in the horizontal axis in above diagram). Find out how many steps are in this region. A step size that traverses the critical focus region with one or two steps should be sufficient.

Conversely, you can also change your auto focus step size so that there are about 8 to 12 points in the curve where it starts at about 8 and come back up to about 8 again. Use one of these step sizes as your initial autofocus step size. You can fine tune later "to taste" once you have it working.

The above curve is a bit more oversampled than necessary; which does not do much harm except to waste your time while it autofocuses.

You may need to rinse and repeat a few times to discover a good step size for your specific optics, specific rack and pinion focuser, etc.

You do not need to sample the critical focus region very precisely. The aim of the V curve is to measure the larger HFDs (where the walls of the hyperbola are close to a straight line) to establish the minima of the hyperbolic curve.

Be sure never to move the EAF back and forth when you measure the V curve. Always move the EAF in only one direction, otherwise you will be faced with the backlash. See here for how much backlash I measured a little while ago with the new version of the EAF:

https://bbs.astronomy-imaging-camera.com/d/12562-5v-eaf-linearity-and-backlash

To remeasure the FWHM for a V curve, always start at some place higher up in steps and move down and only down. Or start at a low number and move up and only up. Don't mix them.

Likewise, after you have figured out the correct focus location, first move the EAF to the same staring point and move down (or) up to the target EAF step.

Don't worry if the data is a little noisy. When a second order curve fit (I used minimum mean square for my program) is made with 8 or 12 points, there will be plenty of extra data points to average out the noisy data.

I had used the average HFD measured by the Star Detect tools in ASIAIR to measure the star sizes. You can use any tool to measure the average FWHM or HFD.

Looking at your curve, where the data points are just wandering around FWHM of 1 and not moving up past an FWHM of 6 or 8, I suspect that either (1) your step size is way too small, or (2) you are not exposing for long enough and you are actually measuring the FWHM of hot pixels instead the FWHM of stars, or (3) you are exposing for too long, so that all the stars are saturated (so the FWHM can only be driven from remaining warm pixels). Case (3) is usually not a problem since a good program should still see dimmer stars and use their FWHM.

Quite often, a belt driven EAF on a helical focuser like the Askar ACL200, and an EAF driving an SCT focuser both need larger step sizes to see a V curve without going through a very large number of steps.

Chen

Hi There
I am having EAF focus issues when changing filters from Ha to S2 the S2 after filter change will be out of focus when doing AF run and in Auto run.I have tried different settings but just cant get past the second filter change being in focus.I have followed advice and put the exposure times in for different filters in the Filter Wheel.
Using Astrodon NB 5nm and Astronon Next Gen LRGB have ASI1600MM Askar 500 with 0.7x and feel i have back lash set acurately 20
Probably user error
Thank you
Dave

davelrkn the S2 after filter change will be out of focus when doing AF run and in Auto run.

Hi Dave, you did not mention what software, but if "AF" refers to the auto focus routine in ASIAIR, then the backlash value is irrelavant. The ASIAiR always starts a high numerical EAF step, and sweep towards lower values; backlashes are cancelled out each time it resets to that high value.

Watch it try to form a U-shaped curve and check to see it resembles the plot that I showed above (v shaped).

If the curve is too flat or shallow, increase the step size in the Focuser Setting window Auto Focus setting (Step Size field). If the curve is too sharp, decrease the Step Size. This is not the Fine or Coarse steps, but something the ASIAIR uses only during its auto focus routine.

Try to change the step size so you only get somewhere between 2 and 5 steps or thereabouts inside the critical focus zone.

Chen

davelrkn They say the Astrodon filters are Parfocal but that a bid much to wish for with the EAF and the fine focus it can achieve .

Actually, a Bahtinov mask is much more accurate than the ASIAIR -- which more often that not, stups the adjustments just after it enters the critical focus zone, instead of at the center of the CFZ.

Whether the ASIAIR will hold focus will depend on which direction the focus needs to change when temperature changes. Generally, the focus plane is closer to the objective whe temperature drops and the tube length contracts. If you are right at the edge of the CFZ and the temperature drops 0.5ºC, you could fall completely out of focus -- it depends on whether the EAF steps should be smaller with a drop of temperature (i.e., depends o the Reverse switch of the EAF).

That being said, parfocal is difficult to attain if the filters are cut from different pieces of glass blanks. I do not know if they do that, or simply sort them after manufacturing.

My set of Chroma LRGB is barely within CFZ of one another, while the narrowband filters from the same brand that I bought separately are definitley far from parfocal.

An example is a cheap Antlia IR-pass filter that I use for guiding, where the maufacturer specified 2mm +/- 0.1mm. That may seem like good tolerance (range of 200 µm), while something like a 250mm/f 4.5 optics can have a CFZ that is less than 50 µm!

The good news is that they don't change measuably (except over centuries -- take a look at the glass of church windows; the bottom is thicker than the top :-). So, all you need to do is carefully measure how many EAF steps is away from another, and just apply the proper chnage to the EAF (taking care of backlash) when you change filters.

Most programs will let you enter these offsets and apply it automatically when you change filters (the ASIAIR is dumbed down, so does not support that function). Some programs will also let you enter a ∆EAF vs ∆T so that it will apply that when temperature changes, without needing to do a full focusing at least for a couple of degrees of temperature change.

I actually spent some time last night measuring the CFZ of a newly purchased Askar FRA300, instead of depending on formulas. Here is an example of one run:

The first column is EAF units, the secomd colum is the average star size, and the rest is some estimate of the CFZ (somewhere between 30 and 40 EAF units).

Chen

davelrkn O3 did two focus runs each time way out of focus

Hi Dave, I wonder if (even the Astrodon) OIII filter has a significant halo, which will cause problems in figuring out the star sizes (and ASIAIR depends on that for the U curve).

Take an image of a very over-exposed star with the Hα filter, then with the OIII filter; and compare them to see if there is any halo in the latter, that is not in the former.

Vega is should be available for such a test this season.

In any case, it looks too like the Astrodons are not perfectly parfocal. Except for SII, they almost seem to be monotonically related to wavelength.

But at least you now have a set of ∆ that you can reset the EAF to when changing filters, and not need to waste time refocusing (make sure to apply the countermeasures for the backlash -- I usually reference all EAF values by first moving it a couple of hundred steps high, and then ask it to go to the final target EAF value).

Chen