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

Building 10Micron models and the things that affect models

  • Please log in to reply
160 replies to this topic

#126 Tonk

Tonk

    Cosmos

  • *****
  • topic starter
  • Posts: 9,529
  • Joined: 19 Aug 2004
  • Loc: Leeds, UK, 54N

Posted 18 April 2021 - 07:57 PM

I see you explicitly mentioned that number for the first model but didn't for the second model.


I called it a "small" model on both occasions - therefore same rules - its in the name (I'm a programmer : ) ). You are just iteratively repeating the process until "convergence". When your PA error is <1 arc min your are basically converged enough - and unless you have a floppy mirror scope that will happen in two passes if it hasn't happened on the first
 

 

For the final model, how many points do you recommend for 600s unguided with a rigid fsq-85? I was thinking between 50 and 60.


You can't recommend - its a function of many factors to do with you specific mount and scope as what you can get away with. Easiest answer is always as many as you can do - mine are always 80+ - and that is for a Tak TOA130. Experiment - try small and large and see what you get. With my TV85 I can use a 25 point model for 10 minutes unguided but that has a wider FOV. Note your points should be evenly spread of course.


Edited by Tonk, 18 April 2021 - 08:03 PM.

  • psandelle, durak and Raginar like this

#127 vdb

vdb

    Surveyor 1

  • -----
  • Posts: 1,615
  • Joined: 08 Dec 2009

Posted 19 April 2021 - 02:42 AM

ModelCreator and 2 iterations of 12 points, 35 Arcsec polar aligned.

It really is that easy.

I finally had a couple of hours clear sky, so I ran a 12 point model and the person running the remote centred the mount, the mount was of by almost 2 degrees, luckily this was a wide field imaging setup. After that initial run the next model reported 3 minutes of the pool. Quick 12 point run and it brought it further down to 35 seconds ...

With a 12 point model I can easily do 5 minute unguided imaging.

 

And each slew lands the object dead center.

 

/Yves


Edited by vdb, 19 April 2021 - 02:48 AM.

  • Tonk, psandelle and yzhzhang like this

#128 SimonIRE

SimonIRE

    Surveyor 1

  • -----
  • Posts: 1,562
  • Joined: 16 Dec 2018
  • Loc: South East England

Posted 19 April 2021 - 03:00 AM

ModelCreator and 2 iterations of 12 points, 35 Arcsec polar aligned.
It really is that easy.
I finally had a couple of hours clear sky, so I ran a 12 point model and the person running the remote centred the mount, the mount was of by almost 2 degrees, luckily this was a wide field imaging setup. After that initial run the next model reported 3 minutes of the pool. Quick 12 point run and it brought it further down to 35 seconds ...
With a 12 point model I can easily do 5 minute unguided imaging.

And each slew lands the object dead center.

/Yves


Yep - told ya! ;)

#129 SimonIRE

SimonIRE

    Surveyor 1

  • -----
  • Posts: 1,562
  • Joined: 16 Dec 2018
  • Loc: South East England

Posted 19 April 2021 - 04:22 AM

Incidentally, if you want to guide it's incredibly easy. 

 

I went to bed at 21:00 last night. I had Voyager set to start imaging at 21:30. 

 

Dual tracking on, exposures in PHD2 set to 10s. I didn't fiddle with any of the other PHD2 settings. 

Image scale was 0.65"/pixel

Full frame sensor (QHY600)

One push of a button

 

Here was my guiding-

 

Screenshot 2021 04 19 At 09.02.45
 
so if you want to guide, it's easy to do so. 

  • durak likes this

#130 Tonk

Tonk

    Cosmos

  • *****
  • topic starter
  • Posts: 9,529
  • Joined: 19 Aug 2004
  • Loc: Leeds, UK, 54N

Posted 19 April 2021 - 05:41 AM

Also note that if you came from the Losmandy world and have used either the Gemini I or Gemini II controllers, then the model building and polar aligning is *exactly* the same procedures as used in the 10Micron. I came this route and hence found using the 10Micron incredibly easy.

The only thing that flumoxed me when I got my first 10M in 2013 was the lack on PEC on the handset. I had no idea what absolute encocders were then - so it wasn't the selling point at all - it was seeing that the model building and PA procedure were familiar to me and it didn't need me to cart around a laptop. It was my uber mobile setup!

When I did finally twig that it had absolute encoders and what it did, I realised I'd bought a gem of a mount. smile.gif







PS I love unintentional puns.


Edited by Tonk, 19 April 2021 - 05:42 AM.

  • psandelle, Raginar, Greg F and 1 other like this

#131 SimonIRE

SimonIRE

    Surveyor 1

  • -----
  • Posts: 1,562
  • Joined: 16 Dec 2018
  • Loc: South East England

Posted 19 May 2021 - 12:56 PM

Bump.

Any interesting new 10 Micron model building news or observations?

#132 FredOS

FredOS

    Ranger 4

  • -----
  • Posts: 347
  • Joined: 16 Feb 2017

Posted 20 May 2021 - 05:38 PM

More of a question but as I see your example with 10s guiding and using voyager, I was wondering how you manage to dither as it probably takes a lot of time to settle with 10s guide frames ?


Edited by FredOS, 21 May 2021 - 04:18 AM.


#133 Raginar

Raginar

    Cosmos

  • *****
  • Posts: 9,805
  • Joined: 19 Oct 2010
  • Loc: Pensacola, FL

Posted 20 May 2021 - 06:01 PM

More of a question but as I see your example with 10s guiding and using voyager, I was wondering how you manage to dither as it proabaly takes a lot of time to setlte with 10s guide frames ?

Never thought about that.  I'm interested to hear how they solve that.



#134 SimonIRE

SimonIRE

    Surveyor 1

  • -----
  • Posts: 1,562
  • Joined: 16 Dec 2018
  • Loc: South East England

Posted 23 May 2021 - 06:35 AM

Never thought about that. I'm interested to hear how they solve that.


I’ve only started dithering recently but I understand the question and I’m not sure I have an answer...

#135 FredOS

FredOS

    Ranger 4

  • -----
  • Posts: 347
  • Joined: 16 Feb 2017

Posted 24 May 2021 - 04:58 PM

I have been going unguided which facilitated dithering. I plan to go with OAG guiding which leaves two options : (1) shorter guide frames of 4/5 s, (2) drop dithering. With an ASI 6200 I’m unclear what you lose we no dithering.

#136 Lead_Weight

Lead_Weight

    Vanguard

  • *****
  • Posts: 2,270
  • Joined: 21 Jul 2016
  • Loc: Houston

Posted 24 May 2021 - 05:38 PM

I'm sure there are still dead/bad pixels which could cause fixed pattern noise. So dithering is still good to do. There's a bunch of imaging apps out there that can do non-guide dither pulses for dithering without guiding. I use EKOS and it has that feature in the guiding preferences.



#137 FredOS

FredOS

    Ranger 4

  • -----
  • Posts: 347
  • Joined: 16 Feb 2017

Posted 25 May 2021 - 12:05 PM

Agree. Non-guide is managed by several applications (I use SGPro and Voyager). Question is more about what to do when guiding.



#138 alphatripleplus

alphatripleplus

    World Controller

  • *****
  • Moderators
  • Posts: 125,333
  • Joined: 09 Mar 2012
  • Loc: Georgia

Posted 25 May 2021 - 02:24 PM

Everyone please remember this is not an imaging forum. If you have questions regarding the pros/cons of various general imaging decisions unrelated to 10Micron mount models, please ask them in an imaging forum. Thanks.



#139 Raginar

Raginar

    Cosmos

  • *****
  • Posts: 9,805
  • Joined: 19 Oct 2010
  • Loc: Pensacola, FL

Posted 26 May 2021 - 11:44 AM

https://www.cloudyni...o=new_post&f=76

 

Since we have issues with this every 37 days or so, I created a thread in the imaging forum linked above.



#140 bortle2

bortle2

    Ranger 4

  • -----
  • Posts: 313
  • Joined: 18 Sep 2019

Posted 28 May 2021 - 05:08 AM

https://www.cloudyni...o=new_post&f=76

 

Since we have issues with this every 37 days or so, I created a thread in the imaging forum linked above.

The above link just tries to create a new topic, unfortunately.

 

Correct link is https://www.cloudyni...hread-catchall/

 

... and BTW thanks for that new thread.



#141 Tonk

Tonk

    Cosmos

  • *****
  • topic starter
  • Posts: 9,529
  • Joined: 19 Aug 2004
  • Loc: Leeds, UK, 54N

Posted 09 August 2021 - 04:00 PM

I'm right now commissioning a pair of telescopes in Chile in a stacked configuration on the same GM2000HPS mount. I wanted to answer the question about how good or otherwise the model building can be and the calculated polar alignment error. To do this I built a series of models using Mount Wizard 4 using increasing numbers of points. All models have the initial 3 alignment points and a series of different sized altitude/azimuth grids arranged within 30 to 75 degrees altitude.

First the bottom telescope (a TOA130) nearest the RA axis
 

Points |   RMS  |   Polar Alignment Error  |  Orthogonal Alignment Error  |  # Terms
------------------------------------------------------------------------------------
3          8.3"     427"                      407"                            2
7         19.2"     409"                      396"                            6
10        17.7"     382"                      396"                            9
14         7.4"     381"                      396"                           11
20         9.4"     378"                      400"                           11
26         9.3"     373"                      397"                           11
33        10.2"     374"                      399"                           11 
39        11.1"     375"                      400"                           11
92        10.7"     379"                      400"                           11

Next the FSQ106 telescope furthest above the RA axis
 

Points |   RMS  |   Polar Alignment Error  |  Orthogonal Alignment Error  |  # Terms
------------------------------------------------------------------------------------
3        17.2"      391"                      -926"                           2
7        14.5"      392"                      -926"                           6
9        12.3"      369"                      -928"                           8
15       10.2"      370"                      -926"                          11 
20       11.3"      366"                      -924"                          11
25        4.3"      365"                      -922"                          22
32        5.2"      366"                      -921"                          22
38        5.7"      366"                      -919"                          22
49        6.6"      364"                      -922"                          22
59       12.9"      365"                      -920"                          11
71       11.9"      362"                      -922"                          11
84        6.6"      359"                      -920"                          22
94        7.4"      361"                      -920"                          22
96       11.6"      361"                      -922"                          11

The same FSQ106 results grouped for matching number of terms (all-sky versus split East and West models)
 

Points |   RMS  |   Polar Alignment Error  |  Orthogonal Alignment Error  |  # Terms
------------------------------------------------------------------------------------
Single all-sky model (11 terms)

15       10.2"      370"                      -926"                          11 
20       11.3"      366"                      -924"                          11
59       12.9"      365"                      -920"                          11
71       11.9"      362"                      -922"                          11
96       11.6"      361"                      -922"                          11

Separate East and West hemisphere models (11 + 11 terms)

25        4.3"      365"                      -922"                          22
32        5.2"      366"                      -921"                          22
38        5.7"      366"                      -919"                          22
49        6.6"      364"                      -922"                          22
84        6.6"      359"                      -920"                          22
94        7.4"      361"                      -920"                          22

Note that these are un-pruned (non-optimised) models

Some features to take note of:

  1. The two scopes on the same mount converge to different polar alignment solutions - 361" (FSQ106) and 379" (TOA130). This needs explaining (I don't have one)
     
  2. When the number of terms is below the needed 11 coefficients of  the pointing model fitting polynomial, then polar alignment error (PAE) solution is far from convergence and the result is inaccurate. A sharp inflection occurs around about when the 11th term is solved when the PAE solution levels out and is essentially stable.
     
  3. This suggests that a minimum of a 12 point model is needed to have a good handle on the polar alignment error but 20 points is certainly better and avoids the need to generate large models when polar aligning and checking polar alignment. Note that if you use manual point selection via a preferred bright star list provided by the keypad then it is only possible to build 20 to 25 star models - so the list runs out around about the start of the zone of good solutions.
     
  4. The generated model is either an all-sky (11 term) solution, or if the mount algorithm decides, a separate pair of East and West sky models (22 terms) for use on either side of the meridian. Notably when the model is split in two the RMS fit can be substantially better. Note you don't get to choose an all-sky or split models.
     
  5. The convergence of the polar alignment error and the orthogonal alignment error is not affected by model splitting - essentially same result of all-sky or split when using a large enough number of points
     
  6. The solution for the orthogonal alignment error of the scope is quite robust requiring only around 7 points (possibly less but more than 3) to provide an essentially stable solution

Edited by Tonk, 09 August 2021 - 04:51 PM.

  • lambermo, psandelle and RazvanUnderStars like this

#142 RazvanUnderStars

RazvanUnderStars

    Apollo

  • *****
  • Posts: 1,347
  • Joined: 15 Jul 2014
  • Loc: Toronto, Canada

Posted 09 August 2021 - 04:41 PM

Thanks Tonk, very interesting.

 

A few times when I built the models I did it twice with the same ~20 points and without changing anything. There was a 2" difference between them in RMS, suggesting there is some variability in the measurements. This doesn't explain the difference in the PA error that you have seen, particularly that the two numbers clearly converge to different values. This reminds me of the discrepancy in measuring the Hubble constant which keeps cosmologists awake at night nowadays.



#143 Tonk

Tonk

    Cosmos

  • *****
  • topic starter
  • Posts: 9,529
  • Joined: 19 Aug 2004
  • Loc: Leeds, UK, 54N

Posted 11 August 2021 - 06:33 AM

OK - there is a very important Blog post from 10Micron that explains a number of things about model building and polar alignment

You can find it at https://www.10micron...php?f=16&t=1397

 

 

 

My results (above - 2 posts back) suggest that the advice to use at least 10 stars to obtain a good polar alignment error value should be amended to be between 12 and 20 points


  • psandelle likes this

#144 SimonIRE

SimonIRE

    Surveyor 1

  • -----
  • Posts: 1,562
  • Joined: 16 Dec 2018
  • Loc: South East England

Posted 11 August 2021 - 07:16 AM

OK - there is a very important Blog post from 10Micron that explains a number of things about model building and polar alignment

You can find it at https://www.10micron...php?f=16&t=1397


 

My results (above - 2 posts back) suggest that the advise to use at least 10 stars to obtain a good polar alignment error value should be amended to be between 12 and 20 points

 

This is gold. 



#145 Tonk

Tonk

    Cosmos

  • *****
  • topic starter
  • Posts: 9,529
  • Joined: 19 Aug 2004
  • Loc: Leeds, UK, 54N

Posted 11 August 2021 - 08:18 AM

This is gold.


Sorry I'm not allowed to re-quote the info. You will have to register with 10Micron to read it

#146 alphatripleplus

alphatripleplus

    World Controller

  • *****
  • Moderators
  • Posts: 125,333
  • Joined: 09 Mar 2012
  • Loc: Georgia

Posted 11 August 2021 - 08:18 AM

Moderator Note:

 

Everyone please note that under the CN Terms of Service you may include brief quotes of textual material only in your posts. If you have any further questions on the Terms of Service, please contact a moderator directly. Thanks.

 

 

b. Short (one or two line) quotes of textual material may be posted, as long as the quote is properly attributed to the author and a link to the material is posted if available.



#147 Tonk

Tonk

    Cosmos

  • *****
  • topic starter
  • Posts: 9,529
  • Joined: 19 Aug 2004
  • Loc: Leeds, UK, 54N

Posted 11 August 2021 - 08:30 AM

which begs the question as to why people get away with re-quoting an entire earlier post on the topic thread including embedded pictures. Its about time that was stopped!


  • psandelle likes this

#148 alphatripleplus

alphatripleplus

    World Controller

  • *****
  • Moderators
  • Posts: 125,333
  • Joined: 09 Mar 2012
  • Loc: Georgia

Posted 11 August 2021 - 08:38 AM

Let's get back on topic please.



#149 SimonIRE

SimonIRE

    Surveyor 1

  • -----
  • Posts: 1,562
  • Joined: 16 Dec 2018
  • Loc: South East England

Posted 11 August 2021 - 08:42 AM

Perhaps the most pertinent, on topic post in the thread is not allowed smile.gif

Back to Chile…

Fascinating information on model building and the SH. I’m very much looking forward to seeing how this thread develops.


Edited by SimonIRE, 11 August 2021 - 09:57 AM.


#150 Tonk

Tonk

    Cosmos

  • *****
  • topic starter
  • Posts: 9,529
  • Joined: 19 Aug 2004
  • Loc: Leeds, UK, 54N

Posted 11 August 2021 - 10:30 AM

Back to Chile…


suspect thread confusion - this is model building - Chile here - https://www.cloudyni...ern-hemisphere/


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