×

INDI Library v2.0.7 is Released (01 Apr 2024)

Bi-monthly release with minor bug fixes and improvements

Plate Solver fails to solve

  • Posts: 249
  • Thank you received: 62
wait, i'm getting lost. Are you testing plate solving for simulators only (as in first post and attached screenshots) or actual setup with your asi guide and atik primary camera?
5 years 8 months ago #27917

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

Can you solve _away_ from the pole (30+ degrees away) like at any star? Try that first, get your FOV right then try to do polar alignment. If it fails, attach the logs.
5 years 8 months ago #27918

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

  • Posts: 1119
  • Thank you received: 182


Select ALL the Index files in the Astrometry module. I also use an RC8". same FOV and I have no trouble solving near the pole or anywhere. But I am also using an external hard drive to hold the 25GB worth of Index files.

One other thing I saw is that you have set a chip size of 1280x1024 at 5.2µm pixel size. That will not translate to a FOV of 38'x28' in an RC8". That FOV corresponds to a a much larger sensor size like the one in the ASI1600. I know, because that is my FOV with my RC8" (focal length 1600mm) and an ASI1600.

Also, after how many seconds have you set the solver to time out?

Jo
5 years 8 months ago #27919

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

  • Posts: 1957
  • Thank you received: 420
When you start Ekos you can specify log levels for all sorts of Ekos modules. Please enable verbose logging to file, try again and post the logs here.
5 years 8 months ago #27921

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

  • Posts: 71
  • Thank you received: 2

Replied by kostas on topic Plate Solver fails to solve

we're getting there. i double checked the camera settings and i corrected the FOV for the Atik/Rc8 simulator profile, i missed at telescope tab the camera settings, after that the online solver worked after 105 secs of solving. It seems i need to have installed all the index files for the offline solver to worksince after 200sec of solving nothing happens.

I have also created a second simulator profile with an Asi120m/finder scope combo i'll give that a try later
HEQ5 Pro Rowan Belt Mod, GSO RC8'', SW ED80, Atik 383L+, Atik EFW2, Atik OAG, SX Lodestar, ZWO ASI 120MC
Last edit: 5 years 8 months ago by kostas.
5 years 8 months ago #27922

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

  • Posts: 1119
  • Thank you received: 182

Are you using a Raspberry Pi for plate solving offline? That may be your bottleneck, too, since having such an extraordinarily small FOV with that little sensor and an RC8" will require a great deal of number crunching. The Pi may well time out. I am looking at about 90 s on a Pi3 with an RC8" and the larger ASI1600 sensor. I estimate the sensor area to be approximately 1/6 of the ASI1600 in your case, so the corresponding FOV will be that much smaller, too. That will require the solver to dig much deeper into the larger data sets. You might need a more powerful CPU and more RAM to handle that task.

But in reality you are not going to use that sensor for imaging with an RC8", right? It would be a very small sliver of the sky you would be able to record that way. I would suggest using a DSLR for imaging with an RC8" or at least an ASI1600 size sensor.
5 years 8 months ago #27934

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

  • Posts: 71
  • Thank you received: 2

Replied by kostas on topic Plate Solver fails to solve

I am not following very well your answer el corazon. My imaging setup is an RC8''' with an Atik 383L+ and i guide with an OAG and a Lodestar. My goal is to setup a system that i can control remotely from my tablet and capture my images

Since i am setting up an RPi 2 and i have little time available for field tests i have made the following simulation profiles in order to learn the software, one is with the RC8/atik 383l+ specs and the second simulation profile is with an ASI120M/60mm finder scope specs (which i also have), the purpose of these simulation profiles is to learn the polar alignment procedure that Ekos has and then move to the capture procedure before going out on the filed and figuring things out as i go. If what you say is true and the offline solving is not possible for either of these two profiles with the RPi what other options for polar alignment are available for the RPi setup? I have a skywatcher mount and with the use of EQMod's polar alignment tool and a very well aligned polar scope i can get a very good PA within minutes and adjust with drift align. Is there any equivalent utility for EQMod's polar alignment procedure?

Isn't this also a big limitation for plate solving images in order to frame them? (for the RPi setup that is) In the field we can't always have internet connection and this seems a huge drawback for using an RPi as a remote system to do astrophotography.

On a side note i would to know if there is any torrent for the index files of astrometry, at random moments the download of the large files gets cut and is very irritating to waste time (i tried at my work's pc and i had the same issue)
HEQ5 Pro Rowan Belt Mod, GSO RC8'', SW ED80, Atik 383L+, Atik EFW2, Atik OAG, SX Lodestar, ZWO ASI 120MC
5 years 8 months ago #27942

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

  • Posts: 1119
  • Thank you received: 182

I am not sure I can answer your questions precisely, kookoo_gr. From the screenshots you pasted initially you had the sensor size of the 120MM but with the wrong vertical resolution (1024 instead of 960) and with the wrong pixel size (5.2 microm instead of 3.75 microm) entered instead of the Atik383L+ (3354x2529 and 5.4 microm) for your primary telescope. That resulted in the calculation of the wrong FOV and that is most likely why the solver failed (and worked after you corrected it). I know that with a Pi3 I can solve that FOV within about 90s offline with all the Index files installed. The Pi2 is slower, so that may not work with such a small field and with 200s time out (my timeout is set to 300s). I am also using a Zotac miniPC, which is about 6 times faster than the Pi3, mainly because it has 4 GB of RAM instead of 1GB. That makes all the difference when it comes to solving.

For polar alignment I am using my polar scope in the mount. I tried the polar alignment tool in Ekos once, but I am so much faster aligning with the polar scope and I have never had problems with it with exposures up to 5 min so I stick to that. You will find that you will have trouble using the polar alignment tool unless the optical axis of your RC8" is very close to the Northern Celestial Pole (i.e. minimal cone error). Otherwise, it will not work in the field and return an error. And if you align with a finder scope, you will find that flexure will introduce an error that is at least as large or larger than the alignment error with the polar scope. So I will stick to optical alignment for the time being. There is another gadget called Polemaster that does not have a problem with flexure and that is supposedly much better, but it is also cumbersome to use and you need to lug a PC around to run that software as it will not run on a Pi.

So my recommendations would be: 1) Polar align optically with a well aligned polar scope. 2) Use at least a Pi3. 3) Install all the Index files and use an external hard drive for faster access). 4) After you got solving to work in your simulator, try this out for real in your backyard or a park in the city. No need to travel to a dark site for testing it out. 5) Solving with a finder scope will not work for imaging UNLESS the optical axis of your finder scope is VERY closely aligned with the OA of your RC8". Again flexure becomes a problem here.

Also, I am not off-axis guiding, since that is very difficult to do if you want to use the scheduler in Indi, which really is the crown jewel of the entire suite.

Remote control from a tablet or laptop (even my iPhone) works fine using VNC.

I hope this helps somewhat.
The following user(s) said Thank You: Jasem Mutlaq, kostas
5 years 8 months ago #27943

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

  • Posts: 71
  • Thank you received: 2

Replied by kostas on topic Plate Solver fails to solve

Thanks for your reply. Right now i am downloading the index files but it takes forever with a limited bandwidth and i will give it another shot.
If the offline plate solve fails, and i believe it will from what you have written, my options are to use a windows mini-pc in order to do the initial polar alignment. I looked for a version of EQMod for ubuntu but i only managed to find the driver and not the graphic environment i am used to work with. A final option is to find an app for my cell phone that has a polar alignment tool (an app that shows the polar finder view in order to do a PA) for the skywatcher mount but so far i have managed to find for the new polar finders that come with the new mounts skywatcher has and at my HEQ5 pro i have the old one, does skywatcher make new polar finders for the old mounts? Are the new polar finders compatible with the old mounts? It's a shame that there is no visual tool to do polar alignment at Ekos

I recently bought a RPI 3 B+ but later on i found out that there is no compatible version of Ubuntu for that board in order to run KStars/Ekos or is there a way to run Ubuntu on that board that i missed?
HEQ5 Pro Rowan Belt Mod, GSO RC8'', SW ED80, Atik 383L+, Atik EFW2, Atik OAG, SX Lodestar, ZWO ASI 120MC
5 years 8 months ago #27960

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

  • Posts: 1119
  • Thank you received: 182

Let me first say that Indi and Ekos work exceptionally well once you have found out the limitations and have adapted to them. It was a steep learning curve for me as well, but now that I have adapted my hardware and my expectations to what the system can do I do not want to miss them for all the tea in China. Jasem and the team are constantly improving the software, so stay with it.

First, the problem you are having with polar alignment is that you have a very narrow FOV with your RC8". That's why I was not successful aligning with it, either. The OA of my RC8" is not directed at the NCP due to cone error, therefore the alignment repeatedly failed, not because plate solving failed, that worked just fine and fast, too, but because the NCP must be within the FOV for the polar alignment tool to work. It worked when I used my guide scope (280mm focal length) and set that so that the NCP was within the initial FOV. So nothing wrong with the alignment tool, but you really have to fine tune your mount to work within the designated parameters and that becomes increasingly difficult the smaller the FOV is. The cone error of my RC8" on my mount is approximately 1.5 degrees after perfect collimation and I have not corrected that yet, because it really only comes into play anyway if I want to polar align using Ekos. Otherwise plate solving corrects for it within the first pass.
So I would assume that you will have the same problems with a Windows PC using the polar alignment tool. The narrow FOV is the main problem.

But perhaps I did something else wrong, so Jasem, please correct me, if I am writing nonsense here.

For you perhaps the Polemaster would be the right fit. That has a much larger FOV and it is firmly attached to the axis of your mount, so no flexure issues. It works, but as I wrote, I went back to optically polar aligning, because that is just very fast. I am usually done within 90 s at most, and I can do it about 10-15 min earlier while there is still twilight than using the Polemaster or Ekos, for which it needs to be much darker. Visually you can distinguish Polaris a lot sooner than the cameras can.

As for a Cell phone app for polar alignment, I am using PS Align on my iPhone. Simple, but perfectly capable of doing everything I need. Or what you would need, provided your polar scope has polar circle with 10' increments like the one on my Orion Atlas Pro mount.

www.google.com/url?sa=i&rct=j&q=&esrc=s&...ust=1532961577994469

The alignment I routinely get with this is so good, I see no field rotation in up to 5 min exposures and more is just not feasible from my location within a light polluted city. Maybe with narrowband imaging, but I am just getting into that myself.

No Ubuntu image yet for the 3+, but will surely come soon:


My favorite mini-PC for running Ubuntu and KStars is this: www.zotac.com/us/product/mini_pcs/pi332

That works like a charm, although I had to attach cooling fins to it, so it does not overheat in our brutal Texas summers. If you have $180 to spare, I would strongly recommend this mini-PC, as I can vouch that it works with Ubuntu MATE installed directly on the integrated 32 GB eMMC. The 5GHz wireless makes for very fast connectivity to your tablet or laptop for control by VNC and the 2 USB3 ports make for fast connection to your mount, camera and an external USB3 2.5" hard drive that I use for capturing planetary video using Planetary Imager (also run on this mini-PC in parallel with KStars. Makes planetary work simple and painless). It is insignificantly larger than a Pi3, but MUCH more powerful and plate solves about 5-10 times faster than the Pi3.

Check out the astrophotos under my profile, they were all collected using either a Pi3 or the Zotac.

The only thing I need to do these days is polar align my mount, and after that I will be inside the house running the entire session remotely from my desktop or iPhone using a VNC app.

This works really great! Don't give up! Making it work for you is part of the fun of this hobby.

Best

Jo
5 years 8 months ago #27961

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

  • Posts: 71
  • Thank you received: 2

Replied by kostas on topic Plate Solver fails to solve

i finally managed to download all the index files for the solver. In order to save space at the sd card i want to coy them at a usb stick and set the solver to look for the index files at the stick. Where do i do that i looked at every screen but i couldn't find anything
HEQ5 Pro Rowan Belt Mod, GSO RC8'', SW ED80, Atik 383L+, Atik EFW2, Atik OAG, SX Lodestar, ZWO ASI 120MC
5 years 8 months ago #28080

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

  • Posts: 1957
  • Thank you received: 420
The location of the index files is specified in the file /etc/astrometry.cfg via the add_path parameter.
The following user(s) said Thank You: kostas
5 years 8 months ago #28081

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

Time to create page: 0.795 seconds