I have autofocus set up perfectly, I think following the steps in the SG Pro help i.e. to first get perfect focus and note the focuser position then move the focuser 3.5 times out and set an odd number of data points etc. Generally I get good V-curves.
I do wonder though about the HFR achieved. On my system (APM 152mm triplet with QHY 12 OSC camera) the HFR during autofocus using 2x2 binning (I don’t know if 1x1 is better but the help says use 2x2 unless you have a good reason not to) is rarely better than about 1.6, which results of course in a 1x1 HFR of about 3 or so
My optics are perfect LZOZ glass and the refractor is small enough to be not too bothered by ‘seeing’ most nights I’d have thought perfect focus would be a lower number. The resulting images seem to be focused most nights but there is always the thought that you may be a smidgen off. Are there any guidelines on the types of number for the HFR one should be able to achieve?
The HFR reading from your AF routine is just a relative number and has no absolute meaning. When a frame is analyzed, the calculated HFR is a result of the star sizes used to make the HFR calculation. So your first AF frame might show an HFR of 3.5 and the final focus frame might show 1.1. The HFR of 1.1 means that during all the movement of the focuser, the “best” position (that is, smallest HFR) was chosen.
If the seeing is poor or there are mechanical issues, the lowest HFR reading might still produce a poorly focused image. You will also find that the AF routine might finalize at an HFR of 1.1 but the next full exposure frame may show an HFR of 1.8. That does not mean the frame drifted out of focus (but it could), its just means after 15 minutes of exposure the star images built up to be larger. Those larger, average star sizes increased the HFR calculation.
Finally, the HFR calculation will be affected by the focal length of your scope and the size of the pixels in your camera since those factor affect the size of the star images. If you are imaging at 2400mm your HFR will always be larger than if imaging at 780mm. In the end, the smallest HFR is the best focus regardless of the actual value of the HFR.
If you shoot binned exposures, the HFR will be about half what you will see if you switch to unbinned. For reference, my scope has a fl of 780mm and my camera has 5.4 micron pixels. On a good night I see well focused images at an HFR of 0.60 when shooting 2x2 binning. This is from the AF routine, not the final sub.
I certainly understand that HFR is relative and also that the lowest point is, by definition the best focus it could find and indeed that the number 10 mins later may not be the same either, but I’d have thought that with similar focal lengths and sensor pixel sizes there would be some similarity.
You, for example indicated 0.6 is normal for your 780mm focal length and pixel size of 5.4 microns. My system, however is only 900mm focal length (because of the reducer) and my pixel size is 5.12 microns so I would assume I should achieve some kind of HFR number not too far from yours at 2x2 binning, but mine is more than twice as high which makes me wonder if I am really as in focus as I could be?
I know without comparing RAW data that might be difficult to assess but those HFR numbers are not random and are arrived at via some scientific process so given the similar input data the result should be similar too, but it appears not to be?
There is another point I forgot to bring up-- the minimum HFR you are going to see will vary by target. Since every target is going to have a different mix of star sizes, the selected stars used in the HFR calculation will vary. Do I get 0.60 on every target - no.
Also, seeing is a big factor and it will affect your 152mm APO. At my dark sky site, seeing is typically 2 to 3 arc seconds. When its around 2, I can get HFRs as low as 0.60; when 3.0; I get 1.1 or so. On a night of poor seeing, the HFR will go to 2+. If it is a night when the AF routine doesn’t get below 2, I usually shut down or wait to see if it gets better because I know an actual sub is going to have an HFR twice the value of the AF routine HFR.
It is reasonable to wonder if the focus being achieved is the best possible. I suppose you could do a test such as running the AF routine and noting the focus position. Then put a Bahtinov mask on the scope and use it to focus and see how close the positions are. Or install FocusMax, spend a couple of nights calibrating it and then see how close the FM and SGP focus points are.
For me, I simply look at the sub at 2x or 3x to see if the stars are nice round points. If so, good enough.
Indeed. Totally relative and no simple answer. Ideally, one should have a seeing monitor. That way, if it says the seeing is good and you are noting higher numbers than usual, you can eliminate seeing. Conversely, if seeing is bad, you can figure that is the problem. I would not want to be w/o my seeing monitor as it is a great independent check on what is happening. In my experience, given the same equipment and location, seeing is by far the biggest factor.
If only imaged on nights where my 2x2 HFR was lower than say 1.8 I would be lucky to do any imaging at all. In fact whilst I live in a light polluted suburb my seeing is fantastic evidenced by the fact that I can nearly always view planets at high magnification (e.g.400 x) with barely any degradation. My HFR is typically the same at the Queensland Astrofest which does have dark skies in comparison.
I think the images are in focus but can’t understand why the HFR number is so high compared too what I hear others mention. For the record I have NEVER seen an HFR lower than 1.4!
That is pretty much how I have treated it - after all I have had no other option like everyone else - it is what it is, but it still seems strange that using broadly similar equipment the HFR numbers are so different.
Do you find the exposure makes any difference? I used to do 10 seconds, but actually it seemed as though the HFR was lower if I changed to 20 seconds, but that may be coincidence. I use 11 steps and the step size is obviously derived partly by the number of steps. I have always left my minimum star diameter setting to 6. Perhaps that also affects the HFR?
Ultimately it is of course true that it is relative and for my system if I see an HFR of below 1.8 I assume focus is good. A higher number either means something went wrong or the seeing is rubbish. It would just give me more satisfaction if the numbers were lower, even though it may not actually matter.
Yes, I understand all these parameters can change the number. So why don’t we conduct an experiment?
I think you said your system is around 780mm and pixel size is 5.something. Mine is 900mm and 5.12 so really not too different. Our skies are different but can’t do anything about that but we can make everything as close as possible and test.
What is your minimum star size setting?
Exposure for focusing?
We can choose the same subject and post the HFR results. I’d say something at a good altitude - at this time of the year perhaps something like NGC253, but frankly it could be anything.
PS I didn’t suggest agreeing on the number of steps or step size because they will be different, but that shouldn’t matter as the point is we should achieve the best focus which will result in the lowest HFR.
We can do that experiment. My pixel size is 5.4 micron; my unbinned pixel scale is 1.43" so bigger than your 1.17" and that will reduce my HFR compared to you by a proportional amount.
My min star size setting is normally 3.
My exposure time is typically 10 seconds.
My binning is 2x2.
Temperature – all over the place. Right now the evenings will start around 20 C.
I will be at my observatory Sat for some maintenance. If the skies are predicted to be clear, I’ll stay and run the AF test Saturday night but NGC253 will only be at 30 degrees altitude. Need something further north for a decent AF test. Near the equator is preferable but not south of DEC -10.
I can run a series of AF tests with different min star size and exposure times and save the AF packs. Not binning won’t do anything but double the HFR reading; well, not quite since each pixel will have fewer electrons. I may need to use NB because of the Moon but I will try with LUM.
The Clear Sky Chart for my observing site (Stellar Skies) is showing clear for Sat night with 4/5 seeing. I have picked Barnard’s Galaxy (DEC -14) that transits around 9 PM Austin, TX time as a good AF target. Let me know if you prefer something else.
I’m actually away this weekend on the Sunshine Coast of Australia as I type this. I live in Brisbane! That explains my ‘mistake’ about NGC 253
Go ahead and gather your HfR and I’ll find a suitable similar target on my return and make my settings match yours in terms of exposure and star diameter. I suspect changing the latter from 6 to 3 will have some impact.
Hi Charlie, Did you manage to test the auto focus HFR? I arrived back yesterday but now we have a lot of cloud and rain so not sure yet when I will be able to test, but as soon as I can I’ll post the results I got here.
I was not able to do the HFR test, either. With our wide geographic separation, it probably won’t be possible to find a common target to co-test with. However, the reality is that for a test to be genuinely valid, we would both need to shoot the same target, at the same time from the same location.
A test that would show interesting results would be to run a series of AF tests where the various focusing parameters were varied (min star size, exposure lengths and steps between data points) in a systematic manner to see the effects on the HFR calculation. If know from experience that if the number of focuser steps between data points is too small, the V curve will often have “wings” that will skew the HFR calculation. However, the focus is typically still good.
With the current phase of the Moon, it is likely to be a week or more before I return to my observatory to run my tests.