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
118 replies to this topic

#1 Tonk

Tonk

    Cosmos

  • *****
  • topic starter
  • Posts: 9,407
  • Joined: 19 Aug 2004
  • Loc: Leeds, UK, 54N

Posted 23 January 2021 - 07:08 PM

Please continue to ask about 10Micron model building and related stuff on this thread.

The old thread can be found here ....https://www.cloudyni...n-mount-models/

 

So we were last talking about the meaning of the model RMS values and how that might affect tracking as well as pointing accuracy ....

So topics would include:

The process of building models - either using support software, or doing it manually; the meaning of modeling terms; periodic errors; the effects on non-rigidity in your system; the orgins of non-rigidity - tripods, cables and felt etc.; refractors versus reflectors; refraction data - and where to source it; timing clocks - gps versus inet sourced atomic clocks, model effects on tracking (includes refraction compensation); guided versus unguided tracking - and how the model might help or hinder; hybrid guiding - for exposures exceeding the 10Micron quoted limits; how dithering is supported by 10M etc etc; plus any issue that you think relates to model building / model use - we can help decide if it related or not.

All the above stuff and more works its way back to the use of absolute encoders and the pointing models built on top

 


Edited by Tonk, 24 January 2021 - 06:24 AM.

  • psandelle, RazvanUnderStars, RossW and 1 other like this

#2 nathanm

nathanm

    Vostok 1

  • -----
  • Posts: 186
  • Joined: 28 Jun 2020

Posted 23 January 2021 - 09:14 PM

As per previous comments, my surmise is that the RMS value is the modeling error between predicted mount performance, and the set of trial positions in the model.   

 

If that is true, then it ought to be relevant to GoTo error - i.e. if you issued N commands to GoTo coordinates of each modeled point, then it is the RMS error across those commands and the actual positions;  under the assumption that the mount can execute its model perfectly.

 

My guess is that this is nearly true - i.e. that the model parameters are generally very very very close matches for what the hardware can do.

 

Pointing error is not necessarily the same as tracking - we don't issue N different Go-tos to track.   However, it seems likely that the primary errors in tracking are also potential errors in pointing so the two are probably very closely linked.   



#3 RazvanUnderStars

RazvanUnderStars

    Apollo

  • *****
  • Posts: 1,135
  • Joined: 15 Jul 2014
  • Loc: Toronto, Canada

Posted 23 January 2021 - 11:04 PM

I think https://www.10micron...ron-technology/ will be of interest to readers of this thread, it's an overview of their tech. One interesting comment is " if you allow for an error of one arcsecond during the exposure of an image, a little calculation will show that this error can be introduced by a deformation of a leg of the typical tripod of just 5 thousandths of a millimeter! A deformation like this can be easily caused just by moving around the barycenter of a non-perfectly balanced telescope." I haven't done the math and I'll presume the excerpt is applicable to larger focal lengths, but nonetheless, it highlights the need for a rigid setup.


  • psandelle likes this

#4 Tonk

Tonk

    Cosmos

  • *****
  • topic starter
  • Posts: 9,407
  • Joined: 19 Aug 2004
  • Loc: Leeds, UK, 54N

Posted 24 January 2021 - 06:38 AM

if you allow for an error of one arc second during the exposure of an image, a little calculation will show that this error can be introduced by a deformation of a leg of the typical tripod of just 5 thousandths of a millimeter!


Likewise I did a calculation that if a pair of felt lined scope rings had a differential compression of the felt of only 10 thousandths of a millimeter between the two rings**  after slewing the scope to a new position (especially on the other side of the meridian), then during the time the scope settled on the felt (up to 5 minutes!!), the image of a star would shift by a few pixels on my camera (creating a short trail).

Effects like this will distort built models unless you a) eliminate the effect b) put in enough of a time delay between slewing and plate solving such that you are not sampling a intermediate position during the settling period.

In my case I went for a) and modified the felt with a plasticiser compound to make it rigid enough.

** happens when the scope is unbalanced across the two scope rings - easily solvable when using a single scope but far more complex when using stacked scopes and stacked scope rings where the goal is to initially balance the whole assembly - then getting the sub-assemblies balanced across different points becomes near impossible


Edited by Tonk, 24 January 2021 - 06:40 AM.

  • psandelle likes this

#5 SimonIRE

SimonIRE

    Apollo

  • -----
  • Posts: 1,159
  • Joined: 16 Dec 2018
  • Loc: South East England

Posted 26 January 2021 - 02:44 AM

Hi All, 

 

 

I am having some difficulty building models with a new 10 Micron mount I received last week. Clear skies are few and far between here in the UK and last night I had a chance to set up. Here is the issue:

 

The mount is big (GM4000) so I ran a crude polar alignment with polemaster (twice) to get close enough to the Pole. My plan was to then run a couple of models with 15-20 stars (base and refinement) to tighten up the polar alignment using my small 430mm FL refractor. Then I was going to run 99 star models on each of the 3 scopes on the rig. 

 

The problem was platesolving. No matter what I did, I couldn't solve using Platesolve 2. I was able to blind solve images (took about 30 secs) and the coordinates indicated that I was about 20 arcseconds off in DEC and about 4-5 arcminutes off in RA. I also ran PHD2 to see what my PAE was and it settled on about 2.7 arcminutes. But if I slewed to a star it was nowhere to be seen on the image (at an image scale of 1.82 "/px: 430mm FL WO Zenithstar 73, ASI1600mm). One thing I forgot to do was simply build a model using an eyepiece on individual stars (Tonk-style); it would have been interesting to see if it was simply a case of my PAE making Platesolve2 timeout. By 2am I was too tired and cold to keep going. 

 

My feeling is that I have something fundamental wrong in the mount set up (rather than the Platesolve2 software). The mount is getting info from GPS - the location and time settings are correct. It is also receiving refraction data. Dual tracking is OFF, although I don't think in this particular instance this is going to make a significant difference.

 

So this is a 10 Micron model building question - any suggestions as to:

 

a) what I could try 

 

b) what might I have forgotten to check

 

Many thanks for reading

 

Simon



#6 Tapio

Tapio

    Skylab

  • -----
  • Posts: 4,077
  • Joined: 24 Sep 2006
  • Loc: Tampere, Finland

Posted 26 January 2021 - 02:58 AM

What software are you using to make model ?

Have you set focal length of the scope and the pixel size of the camera sensor ?


  • SimonIRE likes this

#7 SimonIRE

SimonIRE

    Apollo

  • -----
  • Posts: 1,159
  • Joined: 16 Dec 2018
  • Loc: South East England

Posted 26 January 2021 - 03:05 AM

What software are you using to make model ?

Have you set focal length of the scope and the pixel size of the camera sensor ?

 

Thanks for reading.

 

Modelcreater for model making. I have the FL and pixel size set up in this and it calls SGP to do the capture and platesolve. I have the image scale set up correctly in SGP and when I look at Platesolve2 when it tries to solve, the image field parameters look right (and the image scale is correctly saved in the image headers). I usually bin 2x2 when making models but to avoid complications or something I might be missing with the parameters I didn't bother with this. 

 

If I am honest, I am not 100% sure the issue ISN'T my parameter set up for solving. I just can't see where the issue might be. 



#8 MikiSJ

MikiSJ

    Surveyor 1

  • *****
  • Posts: 1,726
  • Joined: 24 Feb 2006
  • Loc: San Jose, CA

Posted 26 January 2021 - 03:44 AM

I have a CGX and have difficulty in getting models to last for more than a couple of sessions. I recently acquired TheSKYX Pro and started to use the Closed Loop Slew to find an object. It worked, just simply worked. 

 

I contacted the SGPro folk and they also have a similar capability. You might want to look into using something like either the Closed Loop Slew or the Center on Target with SGPro to find your modeling stars.



#9 SimonIRE

SimonIRE

    Apollo

  • -----
  • Posts: 1,159
  • Joined: 16 Dec 2018
  • Loc: South East England

Posted 26 January 2021 - 07:06 AM

I think I know what might be the issue.

I had daylight saving time set to ON (and the correct time set in the HC).

I can’t test it now, but it seems obvious that I should have daylight saving time OFF. Foolish error.

I remember fiddling with this with my 2000; there are a lot of people that seem to get confused by this - as they say, it pays to read the manual properly.

Wish I had a clear night to check this though.

Edited by SimonIRE, 26 January 2021 - 07:51 AM.

  • psandelle likes this

#10 Panotaker

Panotaker

    Viking 1

  • *****
  • Posts: 611
  • Joined: 12 Sep 2003
  • Loc: Austin, Texas

Posted 26 January 2021 - 07:24 AM

I can't help you with your plate solving because I own a Paramount MX, but I am wondering why you have to use the Polemaster twice, to get a crude polar alignment. I use a Polemaster on my MX, and it does a perfect polar alignment on the first try. I have double checked after doing a 50 star Tpoint run on TheSkyX, and it says the polar alignment is perfect, and needs no further adjustment.  It should do it perfect on the 10 Micron mount too on the first try.



#11 SimonIRE

SimonIRE

    Apollo

  • -----
  • Posts: 1,159
  • Joined: 16 Dec 2018
  • Loc: South East England

Posted 26 January 2021 - 07:37 AM

I can't help you with your plate solving because I own a Paramount MX, but I am wondering why you have to use the Polemaster twice, to get a crude polar alignment. I use a Polemaster on my MX, and it does a perfect polar alignment on the first try. I have double checked after doing a 50 star Tpoint run on TheSkyX, and it says the polar alignment is perfect, and needs no further adjustment. It should do it perfect on the 10 Micron mount too on the first try.

[I agree - you are correct. You could it once and my use of it here is simply to get close so it was a bit pointless doing it twice. More generally I have found that with my seeing, sometimes doing it twice improves things a little. The last step can be jumpy due to conditions where I am. I should add that polemaster is not, in my experience, as precise as a decent sized model with a 10 Micron for getting PAE]

We better stay on topic... (the last 10 Micron thread got locked for this very reason!) - thanks for the comment.

Edited by SimonIRE, 26 January 2021 - 07:47 AM.


#12 Tapio

Tapio

    Skylab

  • -----
  • Posts: 4,077
  • Joined: 24 Sep 2006
  • Loc: Tampere, Finland

Posted 26 January 2021 - 08:00 AM

When I used Model Creator last time I used SGP as plate solver (and there is Platesolve2 as plate solve).

Don't know if that makes any difference.

But you should have Platesolve2 in SGP working first.



#13 SimonIRE

SimonIRE

    Apollo

  • -----
  • Posts: 1,159
  • Joined: 16 Dec 2018
  • Loc: South East England

Posted 26 January 2021 - 08:09 AM

When I used Model Creator last time I used SGP as plate solver (and there is Platesolve2 as plate solve).
Don't know if that makes any difference.
But you should have Platesolve2 in SGP working first.

Yes. I’m using the same settings I had for my 2000 a few weeks ago with the same scope so it should be ok.

I do need to change the DST; that’s obviously wrong and hopefully is the only issue. It makes sense too; polemaster and PHD2, two applications independent of the mount computer suggest arc minute PAE which should be amenable to plate solving (albeit slow); unless the mount computer doesn’t know where it’s at due to an incorrect internal setting (ie DST on when it should be off) and is pointing way off.

Edited by SimonIRE, 26 January 2021 - 08:10 AM.

  • psandelle likes this

#14 dhaval

dhaval

    Vendor

  • *****
  • Vendors
  • Posts: 2,235
  • Joined: 21 Jul 2008
  • Loc: Round Rock, TX

Posted 26 January 2021 - 10:57 AM

Question for those that have been using 10Micron mounts, especially if you are using one with either a SCT or a CDK/iDK and running it unguided (or may have tried doing it) - 

I have heard that SCTs are generally an issue when it comes to a model's accuracy, just because of the fact that they have moving mirrors. I assume this extends to EdgeHD scopes as well? Does this also extend to CDKs as well? I don't think that it should, but want some form of confirmation. That obviously also plays a role in having the ability to go unguided, which is really something that I want to explore with a 10Micron mount.

 

CS!



#15 mccomiskey

mccomiskey

    Explorer 1

  • -----
  • Posts: 57
  • Joined: 17 Apr 2017

Posted 26 January 2021 - 11:32 AM

I was hoping to pick up the topic I had raised on the prior version of this thread.  Now that I have my modelling pointing accuracy reliably at or below 4", I am focused on tracking accuracy to achieve the longest unguided subs possible and want understand how this relates to the model. 

 

I have a GM2000, and the specs say that it can do ~0.7" of drift per 15 minute period.

 

So, taking as a given a reasonably accurate pointing model (say RMS 10" or less), is the model a significant factor in tracking accuracy?  I would assume not actually, but want to make sure my understanding is correct.

 

I ran a test a couple of nights ago, with a model that had an RMS of 3.9", using my basic setup (3.5" sampling use a Tak FSQ106), Polar alignment a little over 1 minute off (so well within the accuracy needed with this setup for long exposures), dual axis tracking turned on.  I had no star trailing at 15 mins, but significant star trailing at 30 mins (unfortunately, then the clouds rolled in, so the experiment came to and end).  In trying to understand what might have led to the trailing, I am ticking through potential issues, and wanted to see if modelling might have anything to do with it.  If not, I'd love some suggestions, but will post the question elsewhere, as I don't want to create any risk of this new thread being censured!


Edited by mccomiskey, 26 January 2021 - 11:33 AM.


#16 Arie

Arie

    Messenger

  • *****
  • Posts: 493
  • Joined: 01 Sep 2015
  • Loc: Netherlands

Posted 26 January 2021 - 12:16 PM

Well, you know, these are 10micron specific issues.

The 10M forum is dedicated for this.

Not every 10M user is following threads on CN, but do check the 10M forum.

There censuring is less of an issue.

Having said that, the moderators here at CN just try to keep the threads clean and efficient.



#17 Tapio

Tapio

    Skylab

  • -----
  • Posts: 4,077
  • Joined: 24 Sep 2006
  • Loc: Tampere, Finland

Posted 26 January 2021 - 01:43 PM

There are other mount specific threads here.

Why would 10Micron thread be any different ?


  • Tonk likes this

#18 mccomiskey

mccomiskey

    Explorer 1

  • -----
  • Posts: 57
  • Joined: 17 Apr 2017

Posted 26 January 2021 - 01:45 PM

Arie,

 

Thanks.  I read the 10Micron forum regularly, but it is a fraction as active as this forum, and I haven't found much in the way of useful information (though a few threads of unhappy people convinced their tracking issues are due to hardware/firmware issues). 

 

I think this is a 10Micron-specific thread, and in my question above, really just looking to get confirmation from the 10Micron user brain trust here that the model is not likely a contributing factor in any tracking issues.

 

And I didn't mean to imply anything negative about the censuring - I am a big fan of keeping threads on topic.  Makes the forums much more useful.



#19 Raginar

Raginar

    Cosmos

  • *****
  • Posts: 9,732
  • Joined: 19 Oct 2010
  • Loc: Pensacola, FL

Posted 26 January 2021 - 02:04 PM

Hi All, 

 

 

I am having some difficulty building models with a new 10 Micron mount I received last week. Clear skies are few and far between here in the UK and last night I had a chance to set up. Here is the issue:

 

The mount is big (GM4000) so I ran a crude polar alignment with polemaster (twice) to get close enough to the Pole. My plan was to then run a couple of models with 15-20 stars (base and refinement) to tighten up the polar alignment using my small 430mm FL refractor. Then I was going to run 99 star models on each of the 3 scopes on the rig. 

 

The problem was platesolving. No matter what I did, I couldn't solve using Platesolve 2. I was able to blind solve images (took about 30 secs) and the coordinates indicated that I was about 20 arcseconds off in DEC and about 4-5 arcminutes off in RA. I also ran PHD2 to see what my PAE was and it settled on about 2.7 arcminutes. But if I slewed to a star it was nowhere to be seen on the image (at an image scale of 1.82 "/px: 430mm FL WO Zenithstar 73, ASI1600mm). One thing I forgot to do was simply build a model using an eyepiece on individual stars (Tonk-style); it would have been interesting to see if it was simply a case of my PAE making Platesolve2 timeout. By 2am I was too tired and cold to keep going. 

 

My feeling is that I have something fundamental wrong in the mount set up (rather than the Platesolve2 software). The mount is getting info from GPS - the location and time settings are correct. It is also receiving refraction data. Dual tracking is OFF, although I don't think in this particular instance this is going to make a significant difference.

 

So this is a 10 Micron model building question - any suggestions as to:

 

a) what I could try 

 

b) what might I have forgotten to check

 

Many thanks for reading

 

Simon

Use ASTAP.  PS2 always gave me fits initially.



#20 EFT

EFT

    Vendor - Deep Space Products

  • *****
  • Vendors
  • Posts: 6,234
  • Joined: 07 May 2007
  • Loc: Phoenix, AZ

Posted 26 January 2021 - 10:39 PM

There are other mount specific threads here.

Why would 10Micron thread be any different ?

He was speaking of the 10Micron forum, not a thread.  The 10Micron forum is dedicated to conversation between people who own or use 10Micron mounts.  There are very knowledgeable people there who do not bother to post on forums like CN.



#21 EFT

EFT

    Vendor - Deep Space Products

  • *****
  • Vendors
  • Posts: 6,234
  • Joined: 07 May 2007
  • Loc: Phoenix, AZ

Posted 26 January 2021 - 10:50 PM

So this is a 10 Micron model building question - any suggestions as to:

 

a) what I could try 

 

b) what might I have forgotten to check

I will occasionally have an outing where nothing will solve.  It frequently comes down to a wrong setting somewhere including easy things like using a different camera with the same model in ModelCreator while forgetting to change the profile so that the exposure times result in exposures too short to detect enough stars (those are real head slappers).  So the first thing to do is probably just take a look at an image (or a few occasionally) that is being solved to make sure there are actually stars there or not something rapidly drifting while the plate image is being taken.  Most of the time I find that I am just being dense.

 

The are sometimes though where whatever I do I can't get things to solve reliably.  In those situations, I find it helpful sometimes to just switch to one of the other plate solve programs.  I mostly use PS2, but sometimes I will switch to ASTPA when I can't identify the source of a problem.


  • psandelle likes this

#22 EFT

EFT

    Vendor - Deep Space Products

  • *****
  • Vendors
  • Posts: 6,234
  • Joined: 07 May 2007
  • Loc: Phoenix, AZ

Posted 26 January 2021 - 10:56 PM

When I used Model Creator last time I used SGP as plate solver (and there is Platesolve2 as plate solve).

Don't know if that makes any difference.

But you should have Platesolve2 in SGP working first.

I originally used Pinpoint via MaximDL.  I now use only one of the free solvers but I use it natively as opposed to through another program like SGP.  I keep the image capture separate from the plate solving so that imaging program captures the plate, ModelCreator sends the plate to the solver, and the solver shoots the results back to ModelCreator.  I don't know if this makes any difference, but it is how I have always done it.  These days, I have mostly been using TSX for image capture and PS2 for solving.


  • psandelle likes this

#23 Tonk

Tonk

    Cosmos

  • *****
  • topic starter
  • Posts: 9,407
  • Joined: 19 Aug 2004
  • Loc: Leeds, UK, 54N

Posted 27 January 2021 - 04:02 PM

Re plate solvers for model building. I've never had trouble with ANSVR (local running Astrometry.net engine) and its usable from SGPro. However the team I share my scope with have moved to ASTAP - which having used it a few times myself now appears to be a) bomb proof b) very fast - though we  have also moved to Voyager at this point, so I can't say if it works from SGPro myself.


  • psandelle likes this

#24 dhaval

dhaval

    Vendor

  • *****
  • Vendors
  • Posts: 2,235
  • Joined: 21 Jul 2008
  • Loc: Round Rock, TX

Posted 27 January 2021 - 04:44 PM

Re plate solvers for model building. I've never had trouble with ANSVR (local running Astrometry.net engine) and its usable from SGPro. However the team I share my scope with have moved to ASTAP - which having used it a few times myself now appears to be a) bomb proof b) very fast - though we  have also moved to Voyager at this point, so I can't say if it works from SGPro myself.

SGP works fine with ASTAP, in fact better than PinPoint or ANSVR. It is very fast too.

 

CS!



#25 Lead_Weight

Lead_Weight

    Surveyor 1

  • *****
  • Posts: 1,954
  • Joined: 21 Jul 2016
  • Loc: Houston

Posted 27 January 2021 - 05:31 PM

Question for those that have been using 10Micron mounts, especially if you are using one with either a SCT or a CDK/iDK and running it unguided (or may have tried doing it) - 

I have heard that SCTs are generally an issue when it comes to a model's accuracy, just because of the fact that they have moving mirrors. I assume this extends to EdgeHD scopes as well? Does this also extend to CDKs as well? I don't think that it should, but want some form of confirmation. That obviously also plays a role in having the ability to go unguided, which is really something that I want to explore with a 10Micron mount.

 

CS!

I use an Edge 11, and you have to lock down the mirror, and use an external focuser. This way the model doesn't have an issue when solving on either side of the meridian due to the mirror flop. I don't think this is an issue with CDKs or iDKs unless their main mirror isn't fixed.


Edited by Lead_Weight, 27 January 2021 - 05:34 PM.

  • psandelle and dhaval like this


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