Up until two nights, my sesions were going very well. A few months ago, I experimenced unexplained issues with Eqmod during autocenter after platesolving. I switched to GS server and everithing worked perfectly until two days agosto, when the same problem recurred without me having changed a thing.
Could you helpme if we take a look at. Thank you so much.
Link to Logs
Approx time of issue: 3:00
Useful Info
OS: Microsoft Windows 10 Pro Ver: 4.5.0.1652 (64-bit) .NET: 4.8 ASCOM: 7.1.3.4851
I have been looking into the issues shown in the logs (for the past week) and just forgot to report back on a status update. Without getting into the minutiae, I can say that the major issue here seems to be that, after a successful solve, the EQMOD server was in a state that rejected external sync attempts by SGPro in favor of its own models.
Detected SGPro changes incoming:
Despite not being able to control the issues above, SGPro can still make some changes to be better here:
The tolerance for detecting “bad solves” in this part of the sky at your image scale was buggy and did not apply the projection math properly. This will be fixed in the next release of SGPro 4.5 but note that the sequence still would have failed because EQMOD was not accepting external syncs.
Second, there is an error handling flow in SGPro that was bugged allowing the centering attempt to continue despite the mount sync attempt failing. This alone will not really affect anything except issues a more accurate error to the SGPro notification center.
A concise summary:
With the same target and same image scale you should no longer receive erroneous “bad solve” errors and, once EQMOD is reconfigured to accept external syncs yoy should be good to go.
Hi Ken.
It’s good news that you know this is an error that can be fixed with an upcoming update.
I used EQMOD for many years and even opened support tickets for this exact issue. I was very accustomed to EQMOD, but given the problems I was having with this specific matter, I decided to switch to different software—GsServer (Green Swamp Server). Everything was going well and the problem was gone, but suddenly—and without changing anything (perhaps due to SGP updates)—I started running into this wretched issue again.
Do you have any recommendations regarding the troubleshooting workflow? Also, do you have any advice on external synchronization when using GsServer?
Thanks.
I dont have any experience with GS, but an easy way to test before you start a sequence would be to just run “Solve and Sync” manually and check for success. This action uses the same code for telescope sync that is used during target centering. As for exactly why GS would suddenly start rejecting external sync attempts, I am not sure.
I’ve already run those tests, and everything works fine. I tested various objects before starting the sequence, and both software programs synchronize their data correctly. It’s only during the meridian flip, Ken, that it fails after plate solve. I suspect there’s an issue with one of SGP’s algorithms. It performs poorly with EQMOD and also with GSServer. Many of us are complaining about these issues; I hope the new update resolves the problem.
I understand the frustration, but I am just not seeing that in the logs. I see SGPro make a request to GS and I see it rejected. I am not saying SGPro is not at fault (nor am I saying it is), but I just don’t have anything to go on.
Can you direct me to these other complaints? Maybe they will have additional information we can use for troubleshooting.
Also, maybe GS has its own logs that can be used for troubleshooting? It is extremely difficult to troubleshoot server communications when we only have access to the client’s (SGPro’s) perspective.
I’ve been dealing with this issue for years—first with EQMOD and now with GSserve. I’ve searched for “autocenter” and read through various posts looking for information, but I haven’t found anything concrete. I can get the GS log, and I even tried cross-referencing the data with ChatGPT to see if it could offer a solution, but I don’t fully trust this “thing” they call artificial intelligence… I’d much rather have an answer from a professional like you.