×

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

Bi-monthly release with minor bug fixes and improvements

My first successful (but painful) Ekos/Stellarmate session

  • Posts: 910
  • Thank you received: 86
Hello,
Just wanted to share a newb experience.
The setup:
-- RPI with StellarMate OS, indi server, kstars and ekos.
-- Skywatcher AZ-GTI mount in EQ mode
-- Nikon DSLR with 180 mm manual antique lens.
-- Everything is connected to StellarMate
-- Using VNC (or Windows client) to control things
-- It's a mobile setup - I need to get the tripod outside every time, do polar alignment, etc.

A couple of weeks ago I tried my setup for the first time.
Everything worked well for a while (I was slewing/plate-solving to different targets) until the plate-solver stopped working. It would just try to solve for long time with no success (for more that 30 min). So I aborted the session.

The weather turned bad and I was playing with the mount and CCD simulator for a while. Updated kstars to the latest nightly and the plate-solver started to work.

Yesterday, the skies cleared and I gave the setup another try.
Installed the tripod, connected everything up to find that the (real) camera was recognized but could not be connected. Spent a couple of hours trying to make it work.
Gave up and restored a backup from two weeks ago to find that the camera now works but the plate-solver doesn't!

Installed the latest updates and boom! everything started to work as supposed to.

Spent more than 4 hours on chasing problems but in the end I got a nice set of pictures.
The target was dead in the center of the frame for an hour. Pretty happy with the outcome.

My only gripe is that the plate-solving is very slow when done internally on RPI (5-6 min) - is this normal?
Today, the sky is clear again and I am shooting the same target but now with Windows KStars client connected to StellarMate server - on-line solver takes 70-90 seconds.
My FOV is 450x300.

Very pleased with KStars/Ekos/Stellarmate.
Thank you very much,
Max
-- Max S
ZWO AM5. RST-135. AZ-GTI. HEQ5. iOptron SkyTracker.
TPO RC6. FRA400. Rokinon 135 and other lenses.
ZWO ASI2600MC. D5500 modified with UVIR clip-in filter.
ZWO ASI120MM Mini x 2. ZWO 30F4 guider. Orion 50mm guider.
ZWO EAF x 2.
The following user(s) said Thank You: Jasem Mutlaq
Last edit: 5 years 3 months ago by maxthebuilder.
5 years 3 months ago #32409

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

Max,

Thank you for sharing your experience with us. Was the FOV specified in the align module when you tried solving using "remote" on the StellarMate? The FOV is the key here. If it was 0x0 initially, then the first solve would take quite a bit of time, but otherwise, I'm using remote astrometry all the time here with QSI583wsg camera and typical solving times are less than 10 seconds.
The following user(s) said Thank You: Craig
5 years 3 months ago #32411

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

  • Posts: 910
  • Thank you received: 86
It was 0x0 initially but the first solve (and subsequent solves) gave me 449x299 (I calculated 300x180 with some on-line service - could this be better?).
Solver takes about 5 min and then some more time for WCS.
-- Max S
ZWO AM5. RST-135. AZ-GTI. HEQ5. iOptron SkyTracker.
TPO RC6. FRA400. Rokinon 135 and other lenses.
ZWO ASI2600MC. D5500 modified with UVIR clip-in filter.
ZWO ASI120MM Mini x 2. ZWO 30F4 guider. Orion 50mm guider.
ZWO EAF x 2.
5 years 3 months ago #32412

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

Can you share a link to one your captured FITS images?
5 years 3 months ago #32413

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

  • Posts: 910
  • Thank you received: 86
I didn't save any (are they stored somewhere automatically?) I have NEF files which I can convert to jpg.
-- Max S
ZWO AM5. RST-135. AZ-GTI. HEQ5. iOptron SkyTracker.
TPO RC6. FRA400. Rokinon 135 and other lenses.
ZWO ASI2600MC. D5500 modified with UVIR clip-in filter.
ZWO ASI120MM Mini x 2. ZWO 30F4 guider. Orion 50mm guider.
ZWO EAF x 2.
5 years 3 months ago #32414

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

  • Posts: 910
  • Thank you received: 86
Here's one (jpg) (don't laugh - my place is very light polluted).
-- Max S
ZWO AM5. RST-135. AZ-GTI. HEQ5. iOptron SkyTracker.
TPO RC6. FRA400. Rokinon 135 and other lenses.
ZWO ASI2600MC. D5500 modified with UVIR clip-in filter.
ZWO ASI120MM Mini x 2. ZWO 30F4 guider. Orion 50mm guider.
ZWO EAF x 2.
Last edit: 5 years 3 months ago by maxthebuilder.
5 years 3 months ago #32415
Attachments:

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

  • Posts: 1957
  • Thank you received: 420
With my Canons EOS 700D camera and WilliamOptics 71 mm refractor (F=427 mm) I get a field of view of 180x119 arc minutes. The Pleiades look much bigger in my pics than in this one so I think 499x299 is probably correct.


HTH, Wouter
5 years 3 months ago #32420

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

  • Posts: 407
  • Thank you received: 74
Excepting the FOV set up via online plate solver is done first as per above suggestions.

1. Try turning on verbose for Plate solver and note the number of stars extracted.

2. I try for no more than 400 - this can be achieved by changing the "sampling" value in the options tab. Try changing the value to say 9 if you are getting 1000+ less if less than 1000 stars. if the extracted stars drop below 200 change the sampling value down towards 2 a couple of notches at a time. Practice makes perfect :-)

3. The number of stars extracted will varying so keep an eye on the extracted stars value - Another Wish list would be for "extracted stars" to be shown on std logging not just verbose :-)

4. Pay attention to the Index files recommended in the options window as unused index files will slow down the whole process -
Please moderators/experts correct me here as I know this is true on ASPS on windows but maybe it isn't on Kstars/Ekos.

You can create a manual "Index profile" for each camera/scope by using subdirectories below the normal index location (doesn't have to be could be located else where !) which you can set up a simple script to move the required indexes after deleting all the index's EXCLUDING the subdirectories (if they are below the normal index location)

Remember you are running on a £32 SBC not a PC so it wont be as fast but is idea for simple ligh weight mobile set ups. If you just have non fixed home set up then consider moving Kstars/ekos to a PC (Linux or Windows (64bit)) and just running Indiserver on the RPI that way you can platesolve on a much faster beast
RPI3 Ubuntu 16.04 / AMD desktop Kstars under Ubuntu 16.04 Mounts :azeq6 ,SWAZGoTo

RPI3 Fedora testing out on AMD desktop Fedpra 28 - running kstars 2.9.4 , Indilib 1.7.4 ?????
The following user(s) said Thank You: Craig
5 years 3 months ago #32422

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

  • Posts: 300
  • Thank you received: 57
@stash helped me with a very similar problem a few weeks ago, this advice is exactly right.

Use "verbose" option, adjust "downscaling" to keep "extracted stars" to around 100-200, and manage index files to just the appropriate scale for your rig.

WISH LIST: make these adjustments automatically!

I can be TAUGHT to manage image scale/index files and downscaling/extracted stars, because the algorithms are pretty simple. Therefore they should be in code instead of being passed down in the forum.

The image scale only needs to be determined once per session, then the algorithm should never use inappropriate index files again.

Downscaling is more dynamic. Images in rich Milky Way fields need more downscaling than those near the galactic poles. But solve-field provides plenty of feedback, which should be used to tweak the downscaling parameter on the next attempt.

In the meantime, we can turn on verbose output and watch intently as the "extracted stars" text scrolls by, then hit "Stop, then go into the Options tab to change the Downscaling parameter, then click "Solve and Sync" again and again until we get a good result.
The following user(s) said Thank You: Craig
5 years 3 months ago #32423

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

  • Posts: 1067
  • Thank you received: 140
So when you say you solve remotely, do you have the isolver ndex files on the rpi... as that would need a large card..?.

I have mine on my laptop which I control the rpi with remotely, I don’t use Kstars on the rpi, I use on laptop.. :)
5 years 3 months ago #32424

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

  • Posts: 407
  • Thank you received: 74
Not sure who you were asking but here's my take.

I do both or should I say I have the option of both so I can start Indiserver only on the RPI3 with say indi_synscan_telescope and indi_canon_ccd in the command line. Then I use Kstars/Ekos on a remote PC (I have used both Windows 10 and Linux - I prefer Linux mostly for less interuptions) - this enables me to solve on the PC.
But I can also just start Kstars/Ekos on the RPI3 and platesolve either online (if internet connected) or offline - yes you do need to watch the index's as some are quite large and the sum is larger but it depends on the FOV which the align module will provide - I do a online platesolve at home first (where I have Internet) this provides the FOV for the kit connected. The later is saved so should appear when you next start Kstars on the same computer. I use a Sandisk Ultra 32gb which provides enough for my mobile uses when I use a DSLR and a 18-55 std lens. Platesolving speed is acceptable and you could say not really needed as its a large FOV :-)
RPI3 Ubuntu 16.04 / AMD desktop Kstars under Ubuntu 16.04 Mounts :azeq6 ,SWAZGoTo

RPI3 Fedora testing out on AMD desktop Fedpra 28 - running kstars 2.9.4 , Indilib 1.7.4 ?????
The following user(s) said Thank You: AstroNerd
5 years 3 months ago #32425

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

  • Posts: 1067
  • Thank you received: 140
@stash
Yes I do it the same as you described first in your post.. :)
5 years 3 months ago #32426

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

Time to create page: 0.617 seconds