×

INDI Library v2.0.6 is Released (02 Feb 2024)

Bi-monthly release with minor bug fixes and improvements

Trying to debug astrometry.net problem (cross-posted from Cloudy Nights)

  • Posts: 222
  • Thank you received: 20
Used to plate-solve just fine on my Mac; now it doesn't. Running it out of KStars 3.1.1 (i.e. I'm using what KStars installed, not a freestanding install). Signs & symptoms:
  • Solving works fine for my ASI120MC-S
  • Used to work for the 183MM, but on the Night of Disasters, it totally didn't
  • Spins endlessly with "failed to solve" messages going by, times out after 3 minutes. Used to work in 5-30 seconds.
  • I tried downloading some smaller-footprint index files, which may have been the issue. But removing them (I have autoindexing on) makes no diff.
  • Turning up downsampling (e.g. from 2 to 16) makes no difference.

FOV with my SV70t-IS, FF/R, and 183MM is 125' x 84'; I have 4205 up through 4215 index files, and tried 4203 and 4204.

Lacking access to stars, I have been mucking around with Ekos's "Load and Slew" function, loading a previously-shot sub to solve. Same FITS files solve just fine at nova.astrometry.net, of course.

At one point, I thought I had it narrowed down to "working with visual filter, broken with Ha", which seemed like progress, even though I know for a fact that it worked fine through the Ha filter in the field. I wondered if the amp glow from the 183 was causing a problem, so I tried it with a dark-calibrated sub. Yay, it worked! Then it stopped working. I changed nothing -- honest! But now I can't solve a sub with the blue filter, the Ha, or the dark-calibrated Ha. It was literally "I'll run that again just to make sure" and it stopped solving.

Any ideas at all? Something broken with how the index files are set up? If so, how on earth would I fix that? If anyone has confronted a similar problem, I'd love to hear about it.

In my wackadoodle setup, the guidescope isn't collimated with the main, so having Ekos zero in on the target with the guidescope doesn't help me too much. But if I have to, I guess I'll collimate as best I can and calibrate the difference, then try manually nudging with the hand paddle.

BUT IT USED TO WORK!


Tia (sigh. AGAIN.)
4 years 9 months ago #39691

Please Log in or Create an account to join the conversation.

  • Posts: 2876
  • Thank you received: 809
Please try a newer version of KStars and see if it works then. I had to completely rewrite my build script after 3.0.0 due to some changes in QT and there were some issues with 3.1 that I had to work out. They should all be sorted out now. Please try either 3.2.2 or 3.2.3.

Also, I highly highly recommend testing inside with the Simulators to make sure everything is working with your install/setup before going out in the field. You can simulate everything with fake stars. Just install GSC. I left instructions in the PDF that I include in the dmg. I made an installer so that all you have to do is click a button. It is a 300MB install, but it is highly worth it because you can test plate solving, focusing, sequences, slewing, basically everything you can do under the stars.

But please just ask me if you are having some sort of problem with the Mac build, since I built it.

Thanks,

Rob
The following user(s) said Thank You: Rick Wayne
4 years 9 months ago #39696

Please Log in or Create an account to join the conversation.

  • Posts: 222
  • Thank you received: 20
Nope. 3.2.2 running now. The "Simulators" profile fails to start with "Connection to INDI server locally on port 7624 failed." That's a different problem that I've also posted about here. But if I run things with, say, INDIGO Control Panel so that I can bring up the Alignment tab, I get the identical result.
4 years 9 months ago #39702

Please Log in or Create an account to join the conversation.

  • Posts: 2876
  • Thank you received: 809
Ok but to know why it failed on your system I would need to know more. On my system it is running fine. What are your profile settings. Are you running locally inside kstars or trying to connect to the indigo server? Those are very different profile settings. I would need to see your settings to look for an error in them. And if that is all correct I would next need to see your setup in the INDI settings. You can upload screenshots.

Thanks

Rob
The following user(s) said Thank You: Rick Wayne
4 years 9 months ago #39703

Please Log in or Create an account to join the conversation.

I've heard using binning 4x4 (or whatever maximum binning your camera supports) can work wonders. Also, change the downscale scale to 4 or 6..etc
The following user(s) said Thank You: Rick Wayne
4 years 9 months ago #39705

Please Log in or Create an account to join the conversation.

  • Posts: 222
  • Thank you received: 20
Thanks guys. I did manage to get Ekos running using the INDIGO Control Panel, and I can connect to and run my actual equipment using my Stellarmate appliance. So I'll table the INDI Server connection problem for right now -- if the weather ever clears, I know I can get images, but I'll be pretty lost without plate-solving!

Using the entire simulator stack, I could easily do a "live" plate solve. And when I saved an image to disk, slewed the telescope simulator elsewhere, and then tried "Load and Slew" on the saved image, that worked perfectly too. So it's just my actual live 183MM images that it's failing on. Any suggestions for the next step?

Index files, maybe? According to the guidelines for the index files, the simulator (63.3' x 50.6') should need index files down to 5.0', right? That would be 4203 or even 4202, and I only have 4205 installed right now and it worked fine. ??? Unless I've done the math wrong my 183MM on my scope is 125' x 83'. So, um, get rid of 4205 and try again? That seems crazy, but this is a pretty crazy problem.

[Later: That didn't work. No big surprise there.]
Last edit: 4 years 9 months ago by Rick Wayne.
4 years 9 months ago #39727

Please Log in or Create an account to join the conversation.

  • Posts: 222
  • Thank you received: 20
OK, seems to be working now. For the record, if anyone else is in this jam, here's how I got it to work:

1) Turned off all the options in the "solver options" panel besides "downsample", which I had set to 8. (That's the Alignment tab, Options button, Solver Options).
2) Copied all the index files I'd downloaded to an external drive, so that I could play around with deleting them without fear of Eternal Download Times.
3) Unclicked all the "recommended" index files on either end of the range of the "required" range. (There was one "recommended" *inside* that range, which I elected to leave alone.) I also ensured that the required files were there for both the Tycho2 and 2Mass catalogs.

I haven't tried it in the field, but it's now solving with the "Load and Slew" button, so I'm reasonably confident. Both the simulator (63' wide) and my images (125' wide) are solving with the 4110 index, for whatever that's worth.

Rob, Jasem, thanks for the suggestions. I might very well try binning when I get to the field, if the 183 driver allows that.

NOW I can start debugging the internal-INDI-server problem! :-)
4 years 9 months ago #39744

Please Log in or Create an account to join the conversation.

Time to create page: 2.893 seconds