I have had an extremely frustrating night (spending more time now getting things to work than actually imaging). Please remember that everything was fine when I was using my 800 mm astrograph sg_logfile_20190703234727.txt (817.0 KB)
The only plate solve that would work was remote astrometry blind solve (the local version wouldn’t work)
After plate solving and syncing the scope was going everywhere and anywhere when trying to centre on target (including pointing to the floor). It was going nowhere near where it was supposed to go. It was like the plate solve and sync where completely wrong!
I tried everything including factory resetting the Hand controller in case there was a model there that was interfering
I have attached my log If anyone can shed some light on what was happening
Mostafa,
I looked at your latest log file - quite a time consuming effort. It looks like you didn’t have any successful PlateSolve2 solver successes. It also looks like Astronomy.net also failed a couple of times (or you aborted it), but mostly succeeded. It also looks like solve hints were off. When PlateSolve2 failed SGP reverted to astronomy.net blind solve. The initial hints for RA and DEC for PlateSolve2 (subsequently used for Astronomy.net) were significantly different (way off) from the eventual Astronomy.net solve. Hints come from the mount reporting back RA and DEC it thinks it is at. Therefore the AVX thinks it is pretty far off from where it really is. Performing an initial sync doesn’t seem to get the mount pointing model straightened out sufficiently.
It looks like you tried an “offset” sync behavior initially and later changed to a “sync” sync behavior. This did not help either. From my experience with my AVX, I can sometimes get away with an initial sync - but only if I leave my AVX on the pier (properly aligned and modeled) from a previous session - and start the AVX with “Last Alignment”. So if I have dismounted the AVX from the pier, I always do a two star alignment with at least on additional alignment star. Also, your initial hint for scale was 76(!) and ultimately solved out at 0.8059 - which sounds approximately like a 2x2 bin from you native scale. So double check your FL and pixel size something maybe off by a factor of about 100.
It sounds to me (from a troubleshooting stand point) you want to make sure your AVX is reporting back reasonable RA and DEC so that when these values are used as initial hints for PlateSolve2, you have a good shot at solving and subsequent centering. Since you have devolved back to a really confusing and poor performing result, perhaps you should step back and see if you can get individual parts verified as performing as they should. Starting with the AVX. Here are two choices:
1.) Be sure the AVX is reasonably polar aligned - does not have to be super accurate - half a degree or better. Do a two star alignment and make sure that you are slewing to stars reasonably with the hand control. You can “Get RA/DEC” and see if the values are reasonable. Then move to SGPro. You should be able to get a PlateSolve2 solution right from the get go. Then knowing the AVX is tip-top, see if you can get back to where you were (150 pix error). Then you can try other SGP settings to see if you can improve.
2.) Instead of or in addition to point 1.) above, it might be that you want to get back to your 800 mm scope and make sure you can replicate the performance you were getting previously. Then cautiously proceed again with the 2032 mm scope.
Guess this is my best idea. I hope you can get this going.
I used to do a two star alignment when I did the polar alignment though the AVX mount but then started using a Polestar and so stopped doing the two start mount alignment. I was told that I could just point the telescope and do a plate solve sync which I suppose was wrong advice. So are you saying that without some sort of mount calibration, SGP sync doesn’t work?
I also thought that blind solve did not require any hints, is that incorrect?
Yes my FL was incorrect at one point from too much playing around trying to get things to work, but I corrected that
If you want to do a start from cold without an initial alignment then try setting the mount to the start position, with the alignment marks aligned and doing a quick align should get you close. This relies on the time and location being correct - within a minute - and the mount being polar aligned - to within a degree.
Then, before syncing, slew to a star that’s well away from the pole and do a gross error check. The mount should be pointed within a few degrees of the selected star. Now do an acquire and sync.
Don’t sync on the mount when it’s pointed at the pole region, tiny errors in pointing will generate huge errors when away from the pole.
If you get big errors on the first slew check the mount alignment data - the Ra and Dec offsets and the Cone error. Ra and Dec should be close to 90 and the Cone error small. If they aren’t then it may be worth doing a factory reset and entering all the parameters again.
I totally agree with Chris. Unless there is something fundamentally wrong with your setup or mount. As long as you are not too far off on your alignment, as per what Chis pointed out, a quick alignment should work, follow by a slew away from the pole and the plate solve should work. I do it this way all the time. In fact I pulled out my old Advance GT CG-5 mount last night with the 102 scope and set it up in my driveway to double check even with a mount less capable then the AVX and everything worked as expected. I did a Polar align with my Polemaster, Slewed to the area around Vega, did a focus run and did a Plate solve and sync. Next I slewed to the Dumbbell Nebula and did another Solve and Sync. This is the way I always do my field setups. Sometimes the PlateSove 2 fails to Solve and it fails over to the Local Astronomy Net solver.
Not sure if this is much help, but maybe just wanted to let you know what Chris suggest does work and it’s the way I ways do it every time I’m at my dark site. I never do any star alignment. If this is not working, as I mentioned earlier, I’m now wondering if there is some fundamental issue that is being overlooked or if the AVX has a fundamental issue. Ensuring you have your location and time correct if very important.
Mostafa,
I am not technically saying that sync won’t work - just that PlateSolve2 solving may fail and you have to fail over to Astro.net. Also on your latest log, it seemed that even after syncing to an Astro.net solve and subsequent slew, your AVX was seemingly reporting wildly wrong RA/DEC - so the next PlateSolve2 failed. Chris and Mark are explaining how to get the best result when all you want to do is Quick Align and sync. I know Quick Align / initial sync works for most. But I don’t want to fail over to Astro.net since I don’t have it (ANSVER) installed locally on on my field computers. I have no explanation for how your AVX got so twisted, but Chris and Mark outline how to get it back on track. For me a calibration works well and can be done quick. Those times I have tried to sync only with Last Alignment gave me poor results. (even though I am polar aligned to 10 arc minutes - mostly better - and time is set to 5 sec accuracy). I may try again after hearing Chris and Mark’s reasons for using Quick Align - but Mark does say this method sometimes fails PlateSole2 in his experience. So find what works for you and keep in mind that your 2032 mm FL will magnify centering errors (if error in pixels is your comparison without regard to image scale).
Blind Solve by definition means an image is solved without hints. I run ANSVER locally (via Cygwin and JS BASH) on by big machine (not my field laptops) and it does accept hints at the command line and works much faster. But no, you do not need hints for a blind solve. But I don’t know if SGP inputs hints (if available) on a blind solve and if wildly off, can these hints disrupt Astro.net. It did appear you had a few Astro.net fails from your log.
When I was first started imaging with my C8 on the CGE-DX mount. I did have some issues starting out. Newbie things. So what I did one night was took a lot of images around different parts of the sky both with 1x1 binning, 2x2 binning and 4x4 binning recorded the RA/DEC information and binning information on a notepad and saved them the images for later use. The next day I worked to ensure each image would solve using all three methods, PS2, Local astronomy net, and remote. Doing it this way I was able to narrow down the Astonomy local net catalogs that was used. by limiting the catalogs the local net solves faster since it not looking thought catalogs that are not used for my setup.
So it might be something to try to get the solving to work and to check to see if your coordinates match.
Thank you all for all these incredibly useful comments, really appreciated
The one thing I still don’t understand though, is given the fact the Blind solving does not require any hints, why can’t I just point the OTA assembly at any location (away from the pole and after polar alignment) and just do a sync using blind solve ? Why do I need any form of sync first? Also there have been several suggestions of plate solving several times to improve accuracy , why do that?
Why can’t blind solve just be used to point and sync the scope?
I have noticed that if I just point my telescope manually without slewing it using the motors (so just by undoing the clutches and pointing the OTA) subsequent pointing of the scope is way off ( might even point slew toward the floor) even after plate solving. I assume therefore that even though plate solving has synced the scope to what it thinks are the correct Dec and RA, the scope pointing will be incorrect because the gears are not set in the correct position . Is my assumption correct?
Well, I have never tried such an action so I don’t know, but it would seem like it could mess up the AVX’s sense of RA/DEC. You may have something there.
This might have caused your more radical malfunctions, but It doesn’t seem like it would create a mere 100 pix error on centering. But then again, maybe if there was some minor slippage or motion (a few 10’s of arc minutes) not driven by motors, the result of centering may be more subtle like you were experiencing.
Aside from this speculation, I still think that the increase in image scale (to the SCT) plays a role in centering accuracy.