Like I said, why not go to offsets? It’s repeatable and works. I use Astrodon 3-nm/Astrodon LRGB and I use filter offsets. It took me about an hour to figure out my offsets for my particular setup. I run 9 data points at 5-second exposures using my L filter. I think it takes about a minute to complete a focusing routine.
Do you have proof that your focus is more stable using a particular star versus taking a average of 100s of stars on a screen? I’m asking because I don’t know. I do know that moving to a new guide star takes time and that wrecks your reasoning above for why it’s important to focus your narrowband filters individually (time).
I understand the desire for the modeling aspect of TSX (if I owned a paramount).
Maybe re-evaluate what you need your software to do. I know SGP has issues focusing with long focal length obstructed telescopes (I’ve never used them yet) though so if that’s part of your problems with SGP, I completely understand.
Quitting cold turkey from MaximDL/FocusMAX/CCDAP was hard. Ultimately, SGP has proven to be significantly easier to use and gives me just as good of a product as far as I can tell. It took me a long time to accept what Ken/Jared always told me which was to fall in line with how SGP worked. I think if you gave Platesolve2/focus offsets/PHD2 a try, you’d find it works just as well as TSX.
I have a medium focal length astrograph with 50% CO. SGP fails 40-50% of the time even when start it at perfect focus manually. It usually gets a close approximation with the a weighted average of the low points. I have to guess that this not the best method, however. The slightest change in settings will generate false low points and then it is going to be way off or failed run.
Also for me the full field focus method is extremely slow, not fast. I have a 4K x 4k sensor with a 50 second download time at bin one. Bin two is going to be to loose accuracy. A center subframe might speed it up but you still have the known limitations above. We do need a method or improvement of this one that fills needs of scopes with big COs.
I don’t think TheSky X or any external software is necessary.
It ends up being just something else communicating unnecessarily with SGP. This creates a whole lot headaches in other automation packages. I think the SGP developers totally see this as a potential problem. The developers can speak for themselves but I am guessing that given enough time they would not want any code unless it is 100% internal SGP.
I would prefer an alternative focus implementation in SGPro as I have similar problems. I’m not sure but could the problems be due to the inaccurate measurement of the out of focus doughnuts which happen on scopes with a large CO ?
If whole frame focusing fails twice in SGPro then I switch to SkyX and focus then transfer the settings - it’s quicker (albeit a nuisance) than doing a third run for a NB filter in SGPro. My QSI 583 also has a long download time for the full frame which means focusing can take a while whereas The SkyX FWHM implementation using a bright star and use of subframes mean I can average a number for each step setting and the repeatability does seem very good.
Would FWHM subframe focusing be something to consider as an addition to the SGPro armoury as an alternative even though it might require a slew and return (slew prior).
I think the CO is the problem. I hear it works fine on refractors. I am not sure why whole image HFR fails with a CO. I see the number of stars recognized goes way down as it gets farther from focus.
I have a two degree field so there should be enough brighter stars to get a accurate measurement.
If it can’t be improved than another method needs to be added. One automation program required PinPoint to determine the targets. It must use a list of magnitudes somehow. Ideally you want brighter stars for narrow band. However, the most effective solution is to use one filter ( Lum etc) and do offsets. This way you just pick a magnitude range for the focusing filter.
This the only limitation of SGP for me currently. However, I am lucky focus is fairly stable with my scope. I can uses temperature compensation to tweak for the it without a full refocus.
So what is different about focus max that it handles CO so gracefully? Couldn’t we implement ROI to reduce download times for people with older cameras?
The issue with central obstructions is that SGP gets confused on the size of the star when stars go to donuts. It’s not a limitation of HFR but a limitation of our star detection algorithm. If/When we get that fixed then scopes with central obstructions will focus just as well as refractors.
If you already own PinPoint you can try setting the focus method to FWHM which will use PinPoint and may deal with donut stars better. Again, not because it uses FWHM but because PinPoint may be better at figuring out what to do with donuts. I would really like to hear some feedback on using PinPoint as the metric provider with a centrally obstructed scope. I’ve been recommending it for a while but have heard no feedback from it (and I don’t own a scope with a CO to test with).
You might seek some suggestions from the Pixinsight authors. Juan Conejero seems easy to approach for help.
(Aside: I know he would love to have a few acquisition programs implementing the XISF image format instead of FIT. Apparently, it can do a lot more than FITs and it now the default image format for Pixinsight. The specs on this are open source. PI users such as myself would like to have it added sooner or latter. )
Right, but nobody has been able to define any clear advantages from the capture perspective. I understand the advantages from a processing perspective and since PI will always open FITS images and allow saving as XISF, I am not sure how us spending time on this would help anyone. I’m not saying that it wouldn’t just that I don’t understand how it would.
Agree 100%. I was just thinking out loud . There should be away to make the focus recognize obstructed stars as well as unobstructed.
If your are hitting the wall on the focus routine, Juan would be a good resource for ideas.
Just saying he would be likely be extra helpful if he heard you eventually might add XISF.
It is not high my wish list either.
By the way, it is easy make a refractor into a obstructed scope. Create a 40-50% obstruction with cardboard and some wire. Use masking tape to hang it on the dew shield in front of the lens ( not touching the lens of course). Now your driving a convertible. Sounds goofy but for testing this should be equivalent to a SC or astrograph
That would be great. TSX plate solve is probably the best image link software there is. It never fails. People who own paramounts already have this installed in their computers. I’m looking to use SGP for my automation but I just don’t have room to install more software in my observatory computer.
For information, SGP has the ability to set backlash in its focus algorithm. The current version of TSX does not. I prefer SGP, it may be a bit slower, since it is not selecting bright stars and using short exposures, but with its classic graphs I have more confidence in the final result.
I’m not sure what you mean tsx doesn’t have backlash compensation for focus? It’s built in to the driver not the software. It will be in the settings of what ever focuser you are using.
As I mentioned, if SGP integrates with TSX it will only be as a solver. We will NOT control a camera, telescope (well maybe a telescope), focuser, filter wheel, etc, through TSX.
However this is not something we’re even considering doing in the upcoming 12 months…
Tolga - backlash can be implemented in the driver or in the focusing software. TSX relies upon the driver having it.
Not all drivers/focus modules have backlash, however, even some well known ones do not. Usefully, SGP has it. It is also more useful that SGP has it, rather than the driver, since if you have one module that you share between different motors/scopes, you can set the backlash value individually in the equipment profile, to match the mechanicals.
For what it’s worth, I use SGP for focusing both my Takahashi FSQ 106ED, and my Edge 9.25 at f10 2350mm. On both scopes I get beautiful repeatable v curves and crisp focus. Temperature compensation ensures I can go hours without rechecking focus. Donuts have become a non - issue with the latest focusing routine. One of the greatest things about SGP’s focus routine is that I don’t have to leave the target. That is a huge convenience and time saver to me. I can’t imagine the improvement that some other focusing software would have to offer to make me be willing to leave the target every time I wanted to check focus. The only improvement that I would be happy to see would be the image download speed of SGP in general.
Hi Ken,
I find even with image history turned off the download speed of images is still pretty slow. I am using a full frame DSLR, so maybe there isn’t much can be done about it.