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 mount models

  • This topic is locked This topic is locked
723 replies to this topic

#701 mccomiskey

mccomiskey

    Explorer 1

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

Posted 03 January 2021 - 04:43 PM

Equally, I no longer plate solve. Focal lengths from 530 to 1090 and after model building target always centered within a few pixels. I’ve done 8 hour sessions with half a dozen targets while experimenting, and the mount has been dead on every time.
 

#702 EFT

EFT

    Vendor - Deep Space Products

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

Posted 04 January 2021 - 02:21 AM

You should generally not need to plate solve your final position unless you are using an extremely high magnification on a very distant object or you have an unstable imaging system.  The specification for pointing accuracy with the mount you ordered is below 20".  

 

At 200mm, you won't be plate solving.  There are people doing survey work which needs to be far more accurate with long focal lengths that can shoot that quantity of points while never plate solving.  In general, I would expect this level of performance, or near to it, from any premium mount.  The difference is how the different mounts get to that point.


 

#703 Raginar

Raginar

    Cosmos

  • *****
  • Posts: 9,781
  • Joined: 19 Oct 2010
  • Loc: Pensacola, FL

Posted 06 January 2021 - 10:26 AM

I have ordered a 10micron!   It  won't be here for awhile, and in the meantime I have a model building / using question.

 

How accurate is slewing with a 10micron and a good model?

 

Most mounts have uncorrected errors in tracking, so one must use a closed loop control - i.e. guiding.  The approach for 10micron is to make a pointing model, and that allows accurate tracking without using the closed loop control of guiding.

 

Most mounts also have some error in slewing to a position, and conventionally one uses a different closed loop control in that case:  take an image, plate solve it, compare to the desired location, then iterate correcting the slew.   NINA or SGP or other programs will take a list of slew positions and do that for you automatically, but it is time consuming.

 

I am wondering whether the 10micron models, coupled with the absolute encoders, are accurate enough that you don't have to do plate solving etc.

 

The answer probably depends on the field of view.  If you have very long focal length and are looking at a 

 

If you are spending the whole night on a few positions it would not matter, but I do quite a bit of mosaic work in which case I will have 100 or more positions in a night.  This is typically with 200mm focal length, which has 10 degree field of view (long axis of frame). 

 

If the RMS from 10micron models reported in this thread is correct, then I could save doing the plate solves and iteration.

First night I setup, I did 10 stars and it centered the moon and kept it there the entire night (lunar eclipse).  None of my cheaper mounts ever could center and track accurately a solar/lunar body.

 

You're going to love it!


 

#704 Tonk

Tonk

    Cosmos

  • *****
  • topic starter
  • Posts: 9,453
  • Joined: 19 Aug 2004
  • Loc: Leeds, UK, 54N

Posted 06 January 2021 - 06:45 PM

If the RMS from 10micron models reported in this thread is correct, then I could save doing the plate solves and iteration.


In seven years of using these mounts I have never needed to do either. Even on multiple night acquisitions the mount goes to the same target position each time. The RMS is an overall measure of pointing accuracy over available sky - but for a specific position the pointing is near reproducible for a run of consecutive nights.


Edited by Tonk, 06 January 2021 - 06:47 PM.

 

#705 Raginar

Raginar

    Cosmos

  • *****
  • Posts: 9,781
  • Joined: 19 Oct 2010
  • Loc: Pensacola, FL

Posted 08 January 2021 - 12:32 PM

In seven years of using these mounts I have never needed to do either. Even on multiple night acquisitions the mount goes to the same target position each time. The RMS is an overall measure of pointing accuracy over available sky - but for a specific position the pointing is near reproducible for a run of consecutive nights.

Same.  I used to do it just because and realized how much time I was wasting.


 

#706 Whichwayisnorth

Whichwayisnorth

    Gemini

  • *****
  • Posts: 3,333
  • Joined: 04 Jul 2011
  • Loc: Southern California

Posted 08 January 2021 - 03:13 PM

I don't plate solve any more with the exception of doing the model itself.  The pointing is so accurate it is rather pointless(hehe) to plate solve.

 

Every person I know that has had the opportunity to use a 10Micron understands.  I'd have two 10u mounts but I couldn't afford the 2000.  


 

#707 mccomiskey

mccomiskey

    Explorer 1

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

Posted 20 January 2021 - 06:29 PM

After 2 weeks of poor weather, the clouds have lifted for a few hours.  Having set up again and built a new model (RMS 3.9"), I thought I would post a sub here to see if it can be reverse engineered for any flaws in the model and/or tracking.  15 minute unguided Ha Sub unprocessed, taken at 3.5" sampling:

 

https://1drv.ms/u/s!...wsqXqw?e=j0oanS

 

On a related note, is the RMS attainable be a model limited by the sampling, or is the plate solving sufficient to achieve subpixel accuracy, given a sufficient number of points in the model?


 

#708 Tonk

Tonk

    Cosmos

  • *****
  • topic starter
  • Posts: 9,453
  • Joined: 19 Aug 2004
  • Loc: Leeds, UK, 54N

Posted 20 January 2021 - 06:56 PM

On a related note, is the RMS attainable be a model limited by the sampling, or is the plate solving sufficient to achieve subpixel accuracy, given a sufficient number of points in the model?


The RMS is a measure of pointing accuracy over the whole sky when asked to go to a specific location - its not a measure of tracking accuracy.

You will have to measure the latter directly from your images. Go to a chosen target then take a series of images at known times over say an hour and plate solve the image centers and plot a graph.


Edited by Tonk, 20 January 2021 - 06:58 PM.

 

#709 dhaval

dhaval

    Vanguard

  • *****
  • Posts: 2,236
  • Joined: 21 Jul 2008
  • Loc: Round Rock, TX

Posted 20 January 2021 - 09:49 PM

Is there a need to plate solve if one is using an EdgeHD scope? I should also add - image scale of 0.28"/px.

 

CS!


Edited by dhaval, 20 January 2021 - 09:50 PM.

 

#710 C8er

C8er

    Mariner 2

  • -----
  • Posts: 259
  • Joined: 15 Sep 2012

Posted 21 January 2021 - 01:35 AM

Is there a need to plate solve if one is using an EdgeHD scope? I should also add - image scale of 0.28"/px.

 

CS!

I've recently done some imaging with a C11 Edge HD at full 2800 mm focal length (with mirror locked of course) on my GM1000, at .34"/pixel image scale, and with a 53 point model of 7.8" RMS gotos I did were accurate enough without any plate solving. I didn't try sub-exposures beyond 3 minutes unguided so far at that focal length, but the unguided subs at that exposure lengths had round enough stars for me. I was pleasantly surprised at how well it worked and didn't expect that. Normally I would have guided at that focal length, and maybe I would need to at longer exposure times than 3 minutes.


 

#711 dhaval

dhaval

    Vanguard

  • *****
  • Posts: 2,236
  • Joined: 21 Jul 2008
  • Loc: Round Rock, TX

Posted 21 January 2021 - 08:53 AM

I've recently done some imaging with a C11 Edge HD at full 2800 mm focal length (with mirror locked of course) on my GM1000, at .34"/pixel image scale, and with a 53 point model of 7.8" RMS gotos I did were accurate enough without any plate solving. I didn't try sub-exposures beyond 3 minutes unguided so far at that focal length, but the unguided subs at that exposure lengths had round enough stars for me. I was pleasantly surprised at how well it worked and didn't expect that. Normally I would have guided at that focal length, and maybe I would need to at longer exposure times than 3 minutes.

It is good to know that go-tos are accurate. On the other hand, 3 minute subs sounds a bit on the lower side to me. When I get the mount, I am hoping I can get at least 10 minutes unguided from the mount at that image scale. I may be being ambitious, but one of the major challenges with the C11 is that there are quite a few targets where I can't find a guidestar. ONAG is obviously an option as is the usage of a rotator to frame the target in a manner that may allow picking some guidestars, but I am more inclined to try and solve the issue via unguided imaging. 

 

CS!


 

#712 mccomiskey

mccomiskey

    Explorer 1

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

Posted 21 January 2021 - 09:29 AM

Tonk,

 

Thanks.  I understand that RMS is a measure of pointing accuracy (at least I think I do!) and doesn't speak to tracking accuracy.  The question about sampling is a distinct one that has me curious: is pointing accuracy affected by the sampling accuracy?  That is, if I have a sampling of 0.75", will I get a higher pointing accuracy than if I have a sampling of 3.5"?

 

And thanks for the suggestion on how to check tracking accuracy.  I will set up a series of shots to do just that.

 

Regards,

 

Mark


 

#713 yzhzhang

yzhzhang

    Apollo

  • *****
  • Posts: 1,060
  • Joined: 13 Dec 2012
  • Loc: SoCal

Posted 21 January 2021 - 01:07 PM

The question about sampling is a distinct one that has me curious: is pointing accuracy affected by the sampling accuracy?  That is, if I have a sampling of 0.75", will I get a higher pointing accuracy than if I have a sampling of 3.5"?

I am guessing the question translates to if sampling affects plate-solving accuracy if you use automated model creating routine? I am guessing no, and would be interested to test out, simply bin your images and see if plate-solving results change?


Edited by yzhzhang, 21 January 2021 - 01:08 PM.

 

#714 EFT

EFT

    Vendor - Deep Space Products

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

Posted 21 January 2021 - 01:27 PM

After 2 weeks of poor weather, the clouds have lifted for a few hours.  Having set up again and built a new model (RMS 3.9"), I thought I would post a sub here to see if it can be reverse engineered for any flaws in the model and/or tracking.  15 minute unguided Ha Sub unprocessed, taken at 3.5" sampling:

 

https://1drv.ms/u/s!...wsqXqw?e=j0oanS

 

On a related note, is the RMS attainable be a model limited by the sampling, or is the plate solving sufficient to achieve subpixel accuracy, given a sufficient number of points in the model?

There is so little data that it is hard to get much of impression but a quick visual inspection only shows field curvature to my eye.  I don't see trails.


 

#715 mccomiskey

mccomiskey

    Explorer 1

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

Posted 21 January 2021 - 01:27 PM

yzhzhang,

 

Yours is the right initial question to ask, though in thinking it through it occurred to me that since a model is the result of a series of plate solves, any sampling-induced platesolve imprecision might "model out" by force of numbers.

 

I assume that sampling (within reasonable bounds) has a de minimis impact on plate solving accuracy, which would imply that there is a de minimis impact on the accuracy of a pointing model.  Having said that, on more than once occasion, my assumptions in this hobby have not been correct, so thought I would see if someone had already worked through this question.


 

#716 mccomiskey

mccomiskey

    Explorer 1

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

Posted 21 January 2021 - 01:34 PM

There is so little data that it is hard to get much of impression but a quick visual inspection only shows field curvature to my eye.  I don't see trails.

Ed,

 

Thanks for the response.  As someone who is still new to this hobby and learning, I appreciate any and all observations.  Just for my edification, when you say there is so little data, so you mean that more subs are necessary to make a proper evaluation?  Or are you saying there is insufficient data in this sub?  Just so I know which path to go down as I do my homework.

 

Regards,

 

Mark


 

#717 EFT

EFT

    Vendor - Deep Space Products

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

Posted 21 January 2021 - 01:44 PM

Tonk,

 

Thanks.  I understand that RMS is a measure of pointing accuracy (at least I think I do!) and doesn't speak to tracking accuracy.  The question about sampling is a distinct one that has me curious: is pointing accuracy affected by the sampling accuracy?  That is, if I have a sampling of 0.75", will I get a higher pointing accuracy than if I have a sampling of 3.5"?

 

And thanks for the suggestion on how to check tracking accuracy.  I will set up a series of shots to do just that.

 

Regards,

 

Mark

The controlling factor here is not how well the mount can produce a model with more or fewer points, it is whether your optical/imaging system is stable enough to do so.  The less stable the optical/imaging circle, the more model points you will want until you exceed the number of points you want to use (or the 100-point maximum) at which point you are looking at the need to guide.

 

When it comes to plate solving a specific image during modeling, as long as the image can be solved, it doesn't matter how long you go with exposures as this is completely dependent on the ability of the plate solving program to do its job accurately.  At some unreasonable length of time though, the accuracy of the plate solves will be impacted by the tracking of the mount without a proper model.  So, shorter exposures that show sufficient stars to get a plate solve, are better.  The longest I typically go is 2 to 3 seconds.  I allow the system to settle between slews for longer than the images.  Of course, all of this is dependent on the ability of the camera to collect sufficient data in the time allotted for the images.

 

The best thing for measuring tracking precision is going to be either single exposures long enough to show the tracking or a sufficient number of subs stacked or combined to show any drift.  Ideally, the easiest way to go is to see how long a single sub you can take without showing drift or blowing out the stars too much.  That will then be your upper limit on exposure time.

 

The C11HD will be able to go longer than 3 minutes unguided when the primary mirror is locked down (including turning for primary focuser a bit after engaging the mirror locks) and the image train is rigid.


 

#718 C8er

C8er

    Mariner 2

  • -----
  • Posts: 259
  • Joined: 15 Sep 2012

Posted 21 January 2021 - 01:56 PM

After 2 weeks of poor weather, the clouds have lifted for a few hours.  Having set up again and built a new model (RMS 3.9"), I thought I would post a sub here to see if it can be reverse engineered for any flaws in the model and/or tracking.  15 minute unguided Ha Sub unprocessed, taken at 3.5" sampling:

 

https://1drv.ms/u/s!...wsqXqw?e=j0oanS

 

On a related note, is the RMS attainable be a model limited by the sampling, or is the plate solving sufficient to achieve subpixel accuracy, given a sufficient number of points in the model?

I didn't have a problem seeing enough data. Like Ed I thought your stars near centre looked reasonably round indicating that your tracking was decent, but I think you do have some tilt or miscollimation somewhere in the system, from top left to bottom right of the sensor, as stars towards/at those two corners are elongated, but stars in the top right and bottom left corners, and the centre are reasonably round. CCD Inspector software shows the tilt. An image of the tilt from ccd inspector attached

 

ccdi01.jpg

 

Cheers.


 

#719 nathanm

nathanm

    Vostok 1

  • -----
  • Posts: 187
  • Joined: 28 Jun 2020

Posted 23 January 2021 - 01:17 PM

How does one handled dithering on 10micron mounts?   

 

I suppose you could add random noise to the Go To instructions for the mount, or is there is a specific way to dither?


 

#720 vehnae

vehnae

    Vostok 1

  • -----
  • Posts: 179
  • Joined: 17 May 2013
  • Loc: Finland

Posted 23 January 2021 - 04:36 PM

How does one handled dithering on 10micron mounts?   

Your imaging application needs to support dithering when running unguided. For example SGP has a "Direct Mount Guider" option that will nudge the mount a bit between each frame.

 

  ++ Jari


 

#721 RazvanUnderStars

RazvanUnderStars

    Apollo

  • *****
  • Posts: 1,162
  • Joined: 15 Jul 2014
  • Loc: Toronto, Canada

Posted 23 January 2021 - 04:40 PM

As Jari says. NINA also supports this, you need to enable the "Direct Guider" and then in the sequence you enable dithering. Works very well.


 

#722 nathanm

nathanm

    Vostok 1

  • -----
  • Posts: 187
  • Joined: 28 Jun 2020

Posted 23 January 2021 - 05:36 PM

great, thanks!


 

#723 EFT

EFT

    Vendor - Deep Space Products

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

Posted 23 January 2021 - 05:52 PM

You can set up dithering right in the mount if you want.  You would need it to coordinate with your imaging, but it can be done.  Dithering is covered in the mount's manual.


 

#724 alphatripleplus

alphatripleplus

    World Controller

  • *****
  • Moderators
  • Posts: 123,546
  • Joined: 09 Mar 2012
  • Loc: Georgia

Posted 23 January 2021 - 06:15 PM

Last warning everyone....

Okay everyone, we have strayed off topic yet again. lock.gif


 


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