Basic questions about auto focus


#1

Hello all,

So I’ve been struggling with getting my auto focus routine to consistently produce nice results for a couple of months already and no luck. I won’t mention equipment because I have been trying with different setups, including an ES 127 with Feather Touch, Edge 800 with Feather Touch and WO Star 71 with stock focuser. I have 2 steppers, the Microtouch Wireless and the Rigel Nstep.

So I’ve been reading posts about the subject and trying to identify possible errors I am making and I have 2 basic questions. First, how do I determine my fine and coarse focus? As of right now I just have some logical numbers like 10-100 but I am not sure if those are the correct numbers and how I would get to that number. Will those numbers influence anything in the auto focusing routine?

My next questions is regarding step sizes. What I’ve been doing is check my focus position, check the HFR number, go up 3-5 times, note focus position again, subtract small number from large number and divide by 9 (focus points). But today I was reading a post here that this process should be done differently by achieving focus with a Baht mask, going out 3-5 HFR then back to perfect focus then in another 3-5 HFR from perfect focus than subtracting the smaller number from the large and doing the math above, if this is correct I’ve been doing it wrong all this time, so which one is right?

Thank you!


#2

Hi Ladyhawke,

Your method was the right one as this setting is used only for the manual focus. Just try the values and adjust them until they work right.

This is almost OK, except dividing by 9: The nine steps are distributed left/right around the best focus, so you have only 4 steps on each side to reach the 3-5 times HFR number. You should divide by 4. Anyway this is a very forgiving setting.

Regards,
Horia


#3

I was a little bit confused, now I am REALLY confused :joy:


#4

Hi Ladyhawke,

Here’s how I went about getting my focuser setup and working for the first time. First off you need to understand your equipment. Mine is as follows so your results may, and I’m sure, differ. My imaging train is 8” SCT, followed by Optec TCF-S focuser, Optech Lepus 0.62x focus reducer, Optech Pyxis Rotator, 55mm T-adapter, Atik 383L+OSC camera.
All that being said, the Optec focuser consist of 7,000 steps, midpoint being at 3,500. Given the Lepus 0.62X reducer and the back focus length to camera and my camera pixel count/size gives me an image scale of 0.79”.

  1. Load camera information into SGPro as needed including your image scale.
  2. Set the focuser to the midway point, in my case it was 3,500.
  3. Perform a manual focus so the best focus is obtained with the focuser set at the midpoint. I used Nebulosity to perform this first manual focus. Once I felt I was in reasonably good focus using the focuser on the optical tube I fired up SGPro
  4. In SGPro focuser settings I set the following
    a. Bin 3x3
    b. Exposer time = 5 seconds
    c. Data points = 9
    d. Step size
  5. I started off using 500 for my step size (this will add 2,000 steps to the 3,500 to obtain the first image shot at 5,500 steps with every subsequent shot being 500 less than the previous until obtaining the required 9 data points .
  6. I found for the minimum star diameter 4 works best for my setup
  7. I crop the image at 10%
    Once I have the above loaded into the focuser settings I run the auto focus and watch the graph. It should form a V or U curve with the lowest point on the curve being about 3 or 4 times lower than the highest point. For example if my lowest point was say 2 HFR I’m looking for my highest point to be approximately 6 or7 HFR.
    If after the focus run the spread was not big enough I would increase the step size. If the spread was too much I would decrease the step size and rerun. I found for my setup a step size of 325 gives me very good and repeatable results with 9 data points. I’m seeing typical spreads from about 1.5 HFR low to 4 HFR high on a very good V-plot. If your seeing isn’t real great the V-curve might look more like a U-Curve.

One note – the auto focus may actually take more than 9 data points if your steps are set to low and it needs to take additional data point to get enough point to calculate the focus point. Don’t worry about that for now. You can fine tune once you get a better understanding on how your focuser works with SGPro.

Hope this helps,
mahaffm


#5

how to set up and calculate your step size for AF is explained on pages 123-131 of the user manual.


#6

Thank you. You gave me a few ideas!


#7

Thank you, Dave. Of course I read the manual many times. Still not working as it should, that’s why I am asking questions.


#8

Once you visualize it, it will snap into place. You said that you want to use 9 points. That’s a good number. When AF is working properly, those 9 points will form a nice v-shaped curve (you have probably seen the curves posted here). The “wings” of the “V” will have four points each, and the last point will be at the vertex of the “V” is everything works well. That’s what a good AF curve should look like.

To set your step size, you want the tips of the “V” to have a HFR around 3 times the HFR at perfect focus. Now, remember that the 9 points are spread around the “V” with four on each side and one in the middle. That means that the tip of the “V” is 4 points away from perfect focus.

So you were on the right track. Get good focus and note the HFR and focuser position. Then move focus until the HFR is about 3 times bigger and note the focuser position, just as you described above. But, instead of dividing the difference in focus position by 9, you should divide by four. That should be your step size.

Tim


#9

You’re doing it slightly different. I go from ‘in focus’ out to HFR 3 and note my ‘out’ distance. Then I go back to ‘infocus’. Form there I go ‘in’ to HFR3 and note my ‘in’ distance. I take my ‘out’ distance from my ‘in’ distance and divide that number by however many data points I want. I use 9, but you could use as many as you want (9 is recommended by Jared/Ken). That gives me my step size. All of these are measured in ‘steps’.

Her problem is she’s using an SCT and has backlash. You need to calculate your backlash setting. Start at ‘in focus’ while moving ‘out’ to arrive at ‘in focus’ and turn on HFR measurements via the ‘star’ icon on the tool bar at the top of the screen. Put your step size for ‘fine’ focus to something small like 10 or 20. Then, click once, and take a picture. Note the HFR. Continue this process until you finally see the HFR change significantly like you would on an Autofocus run. My scope is like HFR 1.8 then 1.2… you just have to know how your scope works. Subtract your ‘in focus’ number from the ‘HFR changed’ number (in steps) and that is your backlash. You’ll put that in as backlash compensation (per the manual).

That is the hardest part.


#10

PROBLEM SOLVED!

I’m happy to report that my auto focus issues have been resolved, I am now getting beautiful V or U graphics with perfectly focused stars on 2 of my telescopes. I am not sure exactly what the issue was because there were too many things that I changed. My main suspicion lies on a bad physical connection with stepper’s gears, then I was running a Rigel stepper with a Microtouch controller, I changed to the Rigel but still need to test back on the Microtouch now that I have everything running just to make sure that changes anything or not. The third thing that I changed was the backlash settings which I was not aware I had to use (I read the manual many times, I just don’t think it stands out enough to get my attention so I somehow missed that).

Anyway, thank you so much everyone who took the time to reply!


www.mainsequencesoftware.com