Jump to content

  •  

CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.

Photo

Is it possible to make plate solving work at long focal lengths without internet access?

  • Please log in to reply
10 replies to this topic

#1 tsk1979

tsk1979

    Mariner 2

  • -----
  • topic starter
  • Posts: 272
  • Joined: 02 Aug 2017

Posted 19 October 2020 - 05:17 PM

I have always struggled to make ASTAP work at Long focal lengths, eg 4000mm+

 

ASTAP requires focal length, and focal length for something like Edge HD 8" changes with focus. So you can never be sure its 2000mm or 2100mm, and that can apparently break ASTAP.

Do there exist plate solvers which can do plate solving with a rough focal length?

 

The max I have been able to solve is full frame DSLR on a 2000mm, but when I go to a 1" sensor on the same scope, even with 2000mm put in it just failed, both primary and secondry ASTAP with N.I.N.A

 

I am wondering is there a way to do plate solving over such focal lengths. Any software etc., when there is no internet access?



#2 rgsalinger

rgsalinger

    Fly Me to the Moon

  • *****
  • Moderators
  • Posts: 7,366
  • Joined: 19 Feb 2007
  • Loc: Carlsbad Ca

Posted 19 October 2020 - 05:23 PM

ANSVR and The SKYX Image Link worked fine for me at 2.5 meters of focal length Those are the only ones that I know of that are also trivial to set up. There are others but those two worked right out of the box for me. Of course they have limits on how small the detector needs to be (field of view). 

 

Rgrds-Ross



#3 Midnight Dan

Midnight Dan

    Voyager 1

  • *****
  • Posts: 13,828
  • Joined: 23 Jan 2008
  • Loc: Hilton, NY, Yellow Zone (Bortle 4.5)

Posted 19 October 2020 - 05:32 PM

First thing is the focal length.  Yes it does change if you change focus, but with the Edge 8 you should definitely not be doing that.

 

The Edge series has a built in corrector lens inside the baffle tube.  As with any corrector, there is a specific distance at which you need to have your sensor for the field to be flat and fully corrected.  Any other position will give you significantly worse results. 

 

For the Edge 8, without a focal reducer, that distance is specified as 133.35mm.  You should measure carefully so that your sensor is 133mm from the rear of the scope, then use your focus knob to bring your scope to focus at that point.  If you do that, you will always have the same focal length.

 

And by the way, while the "nomimal" focal length of the 8" SCT design is 2032mm, the focal length with the built in corrector, and the sensor at the proper distance, is actually 2125mm. See page 13 in this Celestron document:

https://tinyurl.com/y38xb6tf

 

If you're imaging at 4000mm, you're likely oversampled by a significant amount, depending on your camera.  If you can set up your software to take your plate solving images at 2x or 3x binning, that would likely help.

 

-Dan


Edited by Midnight Dan, 19 October 2020 - 05:33 PM.

  • RossW likes this

#4 kathyastro

kathyastro

    Vanguard

  • *****
  • Posts: 2,395
  • Joined: 23 Dec 2016
  • Loc: Nova Scotia

Posted 19 October 2020 - 05:33 PM

If you can do a blind plate-solve one time, perhaps using astrometry.net, you can find out the image scale (as/px).  Knowing the pixel density, you can calculate the actual focal length.  It should not change much with temperature.  With a refractor or Newt, it wouldn't change at all with temperature, not 100% sure about an SCT.



#5 Peregrinatum

Peregrinatum

    Surveyor 1

  • *****
  • Posts: 1,535
  • Joined: 27 Dec 2018
  • Loc: South Central Valley, Ca

Posted 19 October 2020 - 05:35 PM

blind solve at 2350mm at F10 with ANSVR no problems, the catalogs are download on to the hard drive... not sure what 4000mm would do


  • wrnchhead likes this

#6 rgsalinger

rgsalinger

    Fly Me to the Moon

  • *****
  • Moderators
  • Posts: 7,366
  • Joined: 19 Feb 2007
  • Loc: Carlsbad Ca

Posted 19 October 2020 - 05:50 PM

The issue is field of view. That's why it's really hard if not impossible to get a blind solve on a guide camera but easy to do with a large chip. I also use a 25" 4.5 meter FL RC and we have no trouble with the SKYX but the chip is a 16803 - really big. I've never tried it on the second camera on the mount - a QHY 290 but I think that a colleague did a while back and failed. 

Rgrds-Ross



#7 BrendanC

BrendanC

    Explorer 1

  • -----
  • Posts: 73
  • Joined: 31 Oct 2014

Posted 21 October 2020 - 04:16 AM

I can't remember if it's All Sky Plate Solver or PlateSolve, but one of them takes you through a wizard whereby you tell it the focal length and it downloads the necessary files. So you could do this for all the focal lengths you think you might be spanning. Wish I could recall which one! Hope this helps.

#8 RossW

RossW

    Mariner 2

  • -----
  • Posts: 282
  • Joined: 15 Jun 2018
  • Loc: Lake Biwa, Japan

Posted 21 October 2020 - 05:07 AM

I doubt ASTAP is failing due to a slightly off focal length, but as Kathy mentioned you can do a solve using Astronomy.net and confirm your OTA's exact focal length. It's then "set and forget". You're using NINA so I assume local ANSRV still doesn't work (did they fix that?). The most robust plate solver at long focal length among the usual free cohorts is PlateSolve2 IMHO. I've used it at 2.8 and 4m focal lengths and, at those focal lengths, it is far more robust than ASTAP in terms of successfully solving degraded or "difficult" images, i.e., low star count (ASTAP needs around 28 stars from memory), cloudy background, deformed stars, streaked stars due to the mount not settling, soft stars (high FWHM) etc. It is slow, but robust.

 

However, ASTAP has made a lot of improvements lately and I rarely have a solve failure now, even at 4 metres FL with minimal starcount. What I found helped greatly was several settings that needed to be adjusted for long focal length solves of images that have few stars and a lot of hot pixels or amp glow. In ASTAP, hit the "Stack" button and then the Alignment tab. Then try increasing "Hash code tolerance" (I'm at 0.007), increase "ignore stars less than [HFD] (start ot 1.5 or 2), and increase "Ignore stars less than ["]" (this one made a big difference, try 1.5 or more). Also you should use the G18 star database for long focal length solves (not G17). Keep "Downsample" at "0" to let ASTAP decide the downsampling (experiment with that setting if you still experience failed solves). Keep ASTAP up-to-date, because Han is regularly releasing improvements.

 

Finding the correct settings for your equipment will take some time (it's a process of iterative refinement), but you can do these tests in the daytime by throwing unsolved images at it while adjusting the above parameters.

 

Good luck.

 

Ross



#9 rnyboy

rnyboy

    Ranger 4

  • -----
  • Posts: 397
  • Joined: 29 Oct 2019
  • Loc: Rochester, NY (Bortle 7/8)

Posted 21 October 2020 - 10:23 AM

I have to echo what rgsalinger said above.  At longer FOVs with my 6SE and my small 385ms sensor, except in rather rare circumstances, I don't have enough stars in the FOV for a plate solve to be successful.  I've used plate solving in Sharpcap down to an experimental f/2.2 and as long as there are enough stars I get a solved plate nearly 100% of the time.  I get successful plate solves at f/3.3, 4, 5, and often enough at 6.3.   But pretty much never at native f/10 and most certainly not at f/25; and I have the plate files stored on my laptop.  So that being said, if you have enough stars in your FOV, I certainly don't know why you aren't able to plate solve without an internet connection as long as you have those plate files stored on your PC.

 

Now to let my plate solving ignorance, or poor memory, shine here... I don't remember ever having to input a focal length when I was setting up Sharpcap to work with AstroTortilla??  Maybe I did when first getting this to work early this year but I was doing so much at once to get everything controlled remotely that perhaps I simply forgot that I did have to input that.  But I don't tell Sharpcap what f-ratio I'm working at and it usually successfully plates solves, and quickly, at all the different faster f numbers I've worked at.  Maybe I'm confused about FL relative to focal ratios some how??  So I didn''t get the purpose of all the back and forth around FLs in the above discussions and I'm wondering what I missed or don't understand here.  From my limited experience if you have enough stars in the FOV then you get a successul plate solve nearly 100% of the time.


Edited by rnyboy, 21 October 2020 - 10:54 AM.


#10 rgsalinger

rgsalinger

    Fly Me to the Moon

  • *****
  • Moderators
  • Posts: 7,366
  • Joined: 19 Feb 2007
  • Loc: Carlsbad Ca

Posted 21 October 2020 - 11:10 AM

It's not just "enough" stars but that's unquestionably a consideration. It's also the field of view itself from what I've read about plate solving software. And, of course, that's another thing to check - are the correct files in place for the FOV that the OP is using. 

 

It's too bad that Ansvr apparently doesn't work with NINA. I had such good luck with it when I was using MaximDL and needed a fall back from time to time. It completely cured the occasional failures that I was getting. 

 

Rgrds-Ross


  • RossW likes this

#11 rnyboy

rnyboy

    Ranger 4

  • -----
  • Posts: 397
  • Joined: 29 Oct 2019
  • Loc: Rochester, NY (Bortle 7/8)

Posted 21 October 2020 - 03:51 PM

Yeah, when I set up Astro Tortilla I downloaded every FOV plate solve set they had.  WOW! on how big it all was.  I think I could now go back and trim that down a bit but had no idea at the time which FOVs I would be using.




CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.


Recent Topics






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics