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

#626 mccomiskey

mccomiskey

    Explorer 1

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

Posted 15 December 2020 - 04:14 PM

Whichway is north,

 

The red bracket is a 3D print of a file that MOrion has posted on the MW4 download site.  Very cool and facilitates rapid set up and break down while keeping everything tight.

 

Mark


 

#627 mccomiskey

mccomiskey

    Explorer 1

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

Posted 15 December 2020 - 04:26 PM

Tonk,

 

I received an email notification of a reply from you to this thread, but for the life of me, cannot find the posting.  I completely agree that an 8" or 9" RMS is good, and, ultimately, given that it represents 30 or so pixels, likely all I need for pointing. 

 

I launched down this hunt largely because I am getting star eccentricities in the 0.6 range with unguided subs.  That is enough for me to see the distortion.  I had hoped that improving the pointing accuracy would somehow help this, but after some thought, realize that doesn't make much sense.  Pointing is pointing and tracking is tracking.  I am getting sub 50" on polar alignment for every imaging session, so field rotation is not an issue with my set up and even 15 minute subs. 

 

I suppose I must either have an issue with the mount not mechanically tracking accurately (which seems unlikely) or i have some kind of tilt issue (will be exploring with CCD Inspector; if anyone can steer me to a thread explaining how to do that, would be appreciated).  Open to any other thoughts on what might be wrong.

 

So far I have not been guiding, but will add that in soon.  Still, would like to get to 10 min subs and eccentricities below 0.5...

 

That said, I suppose that now that I have started to focus on the RMS issue, it has become something of a challenge for me, and I would like to nail it down to the point that I can get to 5" RMS (arbitrarily chosen) with my current setup, just as a matter of personal satisfaction.

 

And as I recall, you are the person who invented his own chemical solution to stiffen his ring felt!!!  This is a hobby of obsessives....


 

#628 Whichwayisnorth

Whichwayisnorth

    Gemini

  • *****
  • Posts: 3,331
  • Joined: 04 Jul 2011
  • Loc: Southern California

Posted 15 December 2020 - 05:36 PM

And all of this is to go unguided as your image scale will allow.

 

If you spent these prices so you can be 100% unguided all the time, then I support your efforts to track down any and every reason for the model to be anything less than amazing.

 

But at some point, if you have room in your image train's back focus, you may want to toss in an OAG and call it a night.


 

#629 mccomiskey

mccomiskey

    Explorer 1

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

Posted 15 December 2020 - 06:16 PM

Capture1.JPG I will eventually add guiding into the tool kit, but I should not need to at this image scale and this sub duration, and I am stubborn about getting things to work...

Most recent model results below, though unfortunately everything has clouded over now.

 


 

#630 EFT

EFT

    Vendor - Deep Space Products

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

Posted 15 December 2020 - 06:34 PM

Well at least the trends are the same this time.  Below 10 RMS should usually get you good result, particularly with a fast system.  Do not obsess over this.  At this point, you need to be looking at your results and ignoring these numbers.  I never worry about anything less than 10 RMS with the systems I work on and some systems may never get that far down and still do fine.  Just like obsessing over polar alignment, obsessing over RMS is nothing but diminishing returns.  The program would be better if it didn't show this data unless you specifically needed it for trouble shooting.


 

#631 RazvanUnderStars

RazvanUnderStars

    Apollo

  • *****
  • Posts: 1,149
  • Joined: 15 Jul 2014
  • Loc: Toronto, Canada

Posted 15 December 2020 - 06:36 PM

Very consistent with the previous one (look at the position of the red and orange points). So what was the position of the system in the worst points?


 

#632 Tonk

Tonk

    Cosmos

  • *****
  • topic starter
  • Posts: 9,434
  • Joined: 19 Aug 2004
  • Loc: Leeds, UK, 54N

Posted 16 December 2020 - 07:32 AM

Your polar chart of pointing accuracy is pretty uniform in increasing accuracy with increasing altitude no matter what the azimuth is ...

... this tells me you should be checking the refraction data you are using (pressure and temperature inputs)

If the refraction correction is NOT used then the errors increase as you reach lower altitudes above the horizon. The refraction error is zero directly overhead.

This is consistent with your charted results! You have either very wrong temperature/pressure values - or no corrections applied at all.


 

#633 Tonk

Tonk

    Cosmos

  • *****
  • topic starter
  • Posts: 9,434
  • Joined: 19 Aug 2004
  • Loc: Leeds, UK, 54N

Posted 16 December 2020 - 07:38 AM

You can check the trend in increasing pointing error using the formulas in here

https://britastro.org/node/17066

If the trend is consistent with not applying refraction corrections - then you know where to look


Edited by Tonk, 16 December 2020 - 07:38 AM.

 

#634 mccomiskey

mccomiskey

    Explorer 1

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

Posted 16 December 2020 - 01:11 PM

Tonk,

Thank you for the insights.  A couple of questions:

1. How are getting to Alt/Az from the polar chart?  I thought it was simply showing the degree of error in arc second for the point at that give compass position (i.e. degree of error for a give Az position), but don't see a way to get the altitude data from this?

2. Thanks for the reference.  Once I figure out how to get the Alt data, I will build a spreadsheet and check this.

3. I have an MGBox v2 attached to the mount. It appears (based on checking local weather data) to be giving accurate information, and I believe I have my mount set correctly to integrate this information (screen shots below).  Have I unwittingly mis set something here?

 

Capture6.JPG  Weather Station.JPG

 

Capture4.JPG Capture3.JPG


 

#635 RazvanUnderStars

RazvanUnderStars

    Apollo

  • *****
  • Posts: 1,149
  • Joined: 15 Jul 2014
  • Loc: Toronto, Canada

Posted 16 December 2020 - 01:55 PM

I'm obviously not Tonk but will provide answers in the meantime as I've looked into this more.

 

The chart itself does not contain altitude information. To double check this, I had a brief email exchange with ModelCreator's author (who was very prompt to reply, hats off) and he confirmed the chart is a polar chart of the error magnitude, as calculated by the 10Micron controller (not MC)  and it's not a sky map. But read below for the meaning of the angle.

 

That said, the table (not the chart) contains the RA/dec coordinates of the alignment stars so you can use them in a planetarium software to derive the altitude, if you know the time you built the model. You should be able then to build the spreadsheet.

 

 

Regarding the values themselves, ModelCreator plots what is returned by the controller in the :getalpN# commands. I have just checked this myself, looking at the network traffic between the laptop and the mount while fetching an old model, still stored in the mount. The excerpt below is from the 10Micron command protocol documentation. Please note the red-highlighted part which provides the correct meaning of the polar angle:

 

:getalpN#

Gets the alignment information for star number N in the alignment model, where N goes from 1 to the number returned by :getalst#, including the polar angle of the measured star coordinates with respect to the modeled star coordinates.

Returns: the string "E#" if N is out of range, otherwise a string formatted as follows:

  • "HH:MM:SS.SS,+dd*mm:ss.s,eeee.e,ppp#" where HH:MM:SS.SS is the hour angle of the alignment star in hours, minutes, seconds and hundredths of second (from 0h to 23h59m59.99s),

  • +dd*mm:ss.s is the declination of the alignment star in degrees, arcminutes, arcseconds and tenths of arcsecond,

  • eeee.e is the error between the star and the alignment model in arcseconds,

  • ppp is the polar angle of the measured star with respect to the modeled star in the equatorial system in degrees from 0 to 359 (0 towards the north pole, 90 towards east).

 

So the angle is not the azimuth of the alignment star as it might be interpreted.

 

An example of the data returned is:

 

:getalp11#

20:56:24.07,+52*29:58.1,   2.8,208#

 

Razvan

 

 


 

#636 Tonk

Tonk

    Cosmos

  • *****
  • topic starter
  • Posts: 9,434
  • Joined: 19 Aug 2004
  • Loc: Leeds, UK, 54N

Posted 16 December 2020 - 03:15 PM

1. How are getting to Alt/Az from the polar chart?  I thought it was simply showing the degree of error in arc second for the point at that give compass position (i.e. degree of error for a give Az position), but don't see a way to get the altitude data from this?


Ah! Sorry I was reading this in the same way as you would read the equivalent  polar graph in MountWizzard 3 - SCRUB what I said then if thats not an all sky polar alt/az chart - its too small on my screen to see the legends

Mount Wizzard plots the true position in the sky of the sampled point and codes the pointing error in size of circle around the point and with colours - darned sight more logical and informative - you can immediately read the where and by how much that way


Edited by Tonk, 16 December 2020 - 03:20 PM.

 

#637 mccomiskey

mccomiskey

    Explorer 1

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

Posted 16 December 2020 - 03:41 PM

Razvan,

 

I asked Martin also, and he responded immediately, so a call out to a great supporter of the 10Micron community.  My interpretation of the polar graph is that the location of the point on the "compass" tells you in which direction the mount has mispointed, while its distance from the center tells you by how much.  It doesn't tell you anything at all about where in the sky that point was.

 

I could run a transform on the RA/DEC coordinates, but I think we can get the approximate alt/az coordinates simply by looking at the model points when displayed on the points tab of the model creator interface.  I'm going to give that a go this evening, since we have a snow storm here tonight.

 

Tonk - that sounds very useful in terms of diagnosis.  I assume MW4 must be capturing the sky coordinates itself and the correlating that with the size and direction of miss information that comes back from the mount?  I would love to give that a go, but am having trouble figuring out how to connect my camera (an SBIG 16803 which doesn't appear to have any ASCOM drives; with ModelCreator, I connect through SGP which in turn offers native support). 

 

If Morion is reading this, is there anyway to connect MW4 to an SBIG 16803?


 

#638 mccomiskey

mccomiskey

    Explorer 1

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

Posted 16 December 2020 - 03:48 PM

And finally, as Razvan suggested, and as promised, I reran the points from last time and took photos of the mount at the position of the 10 worst points in the model.  I can't discern a pattern, but am inviting suggestions.  The 4 worst points attached here, and for the truly curious, the link at the bottom will get you the 9 worst, named from worst to least bad (all being bad, of course).  Photos are all taken standing south of the mount, facing north (obvious to you all, no doubt).

 

A couple of observations:

 

1. Errors appear to cover the gambit from high up to relatively low in alt.  Also appear to be spread around the Az.

 

2. Most of the worst errors are away from the house, so I am provisionally going to assume that the primary problem is not hot air coming off of the house roof.

 

3. When I am running in modelling/imaging mode, the power cable is tied off to both the CW bar and to the mount itself (hanging from the CW bar in the photos).

 

 

 

https://1drv.ms/u/s!...sLWC3g?e=hciNLh

Attached Thumbnails

  • 1.jpg
  • 2.jpg
  • 3.jpg
  • 4.jpg

Edited by mccomiskey, 16 December 2020 - 03:57 PM.

 

#639 RazvanUnderStars

RazvanUnderStars

    Apollo

  • *****
  • Posts: 1,149
  • Joined: 15 Jul 2014
  • Loc: Toronto, Canada

Posted 16 December 2020 - 04:16 PM

Regarding the interpretation, just be careful that the polar angle is relative to the alignment star, not to a fixed point in the sky such as NCP, so the same angle will mean different cardinal directions, depending on the star. So just by looking at the chart it's not possible to immediately derive if the errors are clustered towards particular directions.

 

I think 10Micron's model uses the data as it is because what it does, essentially, is to build a mapping function between the sphere of the sky and the sphere of the coordinates as returned by the encoders. I imagine these like two concentric spheres, with the sky outside. Ideally, they are perfectly aligned so from the perspective of the centre, where the telescope is, the stars align perfectly in all directions. In practice, they don't so one of them needs to be stretched a bit here and there - the model says where to stretch, how much and in which direction. That's why, I speculate, the model points have the data they have. 

 

There's a shortcut you can do to get the visualization: connect with MountWizzard to the mount and load the model stored in it (connect to the mount, go in Manage Model then Model Positions and Error Distribution). Since the model was not built with MW, the more sophisticated analysis is not available, but nonetheless, you'll get a quick visualization.

 

Now, you already wrote that there's no pattern to the errors - a mystery indeed. Perhaps contact the 10Micron support as well? On the other hand, the mount is within the specs (pointing accuracy <20") so I think real issue is whether tracking is within specs (<1" or 0.7" RMS for 15-min exposures). It should be possible, I think, to calculate the tracking error based on the eccentricity but what the latter doesn't tell is whether the eccentricity is due to the mount or to another factor. A scientific experiment to check this would involve using an identical equipment on the same mount, but alas, that's not practical.

 

 

 

Razvan,

 

I asked Martin also, and he responded immediately, so a call out to a great supporter of the 10Micron community.  My interpretation of the polar graph is that the location of the point on the "compass" tells you in which direction the mount has mispointed, while its distance from the center tells you by how much.  It doesn't tell you anything at all about where in the sky that point was.

 

I could run a transform on the RA/DEC coordinates, but I think we can get the approximate alt/az coordinates simply by looking at the model points when displayed on the points tab of the model creator interface.  I'm going to give that a go this evening, since we have a snow storm here tonight.

 

Tonk - that sounds very useful in terms of diagnosis.  I assume MW4 must be capturing the sky coordinates itself and the correlating that with the size and direction of miss information that comes back from the mount?  I would love to give that a go, but am having trouble figuring out how to connect my camera (an SBIG 16803 which doesn't appear to have any ASCOM drives; with ModelCreator, I connect through SGP which in turn offers native support). 

 

If Morion is reading this, is there anyway to connect MW4 to an SBIG 16803?


 

#640 Tonk

Tonk

    Cosmos

  • *****
  • topic starter
  • Posts: 9,434
  • Joined: 19 Aug 2004
  • Loc: Leeds, UK, 54N

Posted 16 December 2020 - 06:19 PM

It doesn't tell you anything at all about where in the sky that point was.


So its usefulness as a diagnostic tool is somewhat diminished :(
 

#641 nathanm

nathanm

    Vostok 1

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

Posted 16 December 2020 - 08:38 PM

It's dangerous for me to opine since I don't have the mount yet, but I am pretty sure that the RMS "pointing" error would be better explained as "modeling error".

 

The 10micron mount has a multi-parameter model which can cover a variety of factors that impact tracking.   So, that includes issues with the gears, with the alignment of the telescope, with some amount of flexing of the telescope, refraction through the atmosphere etc.

 

My understanding is that, as a previous post said, the mount deals with mapping two things - the sky and the internal encoders.

 

I think 10Micron's model uses the data as it is because what it does, essentially, is to build a mapping function between the sphere of the sky and the sphere of the coordinates as returned by the encoders. I imagine these like two concentric spheres, with the sky outside. Ideally, they are perfectly aligned so from the perspective of the centre, where the telescope is, the stars align perfectly in all directions. In practice, they don't so one of them needs to be stretched a bit here and there - the model says where to stretch, how much and in which direction. That's why, I speculate, the model points have the data they have.

 

 

 

The mount model takes various things into account, and tries to model the discrepancy between the sky and the encoders.

 

If the model is able to compensate for the difference between encoders versus actual and does so perfectly for every point, that would have a "pointing error" of zero.   

 

The key reason I say this is that the mount returns the RMS pointing error immediately upon processing of the model.  It does not run an empirical test except insofar as is covered by the model points.   All of the work is done up front, so this is not a measured error it is a predicted modeling error

 

So this is really a curve fitting question.  The model fits the discrepancy between real and actual to its model.   If the curve fits, it has zero error.  But most curve fitting has some residual errors.  That is what the mount is calling "pointing error". 

 

So what does that mean in reality?   If one assumes that the mount is physically capable of perfectly executing the model, then the residual error is the average (RMS) error you could expect in pointing to a random point, if that point is statistically represented by the points in the model.

 

Of course the mount NOT be able to actually "walk the talk" and execute the model perfectly.  Things like inaccuracy in the encoders, wear and tear on the gears, or any other imperfections may mean it can't perfectly execute the model.  In that case the typical error in pointing will be larger.   I suspect that the 10micron hardware is well covered by their model, and that is actual practice imperfections in running the model are much smaller than modeling error,  but nothing or earth is absolutely perfect. 


 

#642 RazvanUnderStars

RazvanUnderStars

    Apollo

  • *****
  • Posts: 1,149
  • Joined: 15 Jul 2014
  • Loc: Toronto, Canada

Posted 16 December 2020 - 09:06 PM

I think you're right, the mount's manual reads "If four or more stars have been used, and the mount has been able to fit a flexure model to the alignment stars, the display will show the number of functions fitted and the expected RMS pointing accuracy (based on the alignment stars’ data)". 

 

In fact, there may be two models, according to Filippo Roccio from 10Micron who wrote the following in a post on their forum:

 

The number of "model terms" generally increases with the number of points/stars used, but not in a predictable way. Sometimes the controller is able to build a model with less terms, sometimes it uses more terms. With tens of points/stars, it doesn't really matter. If you find that the number of terms approximately doubles or halves with respect to another modelling run, it means that the controller has made a choice between fitting a single model to the whole sky, or fitting two separate models to the sky "west of the meridian" and "east of the meridian". This depends also on the zone you used for modelling, and you shouldn't worry about this.

 

 

I'm not sure how the returned RMS is calculated if there are two models since each will have its own RMS, perhaps it's the average.


 

#643 nathanm

nathanm

    Vostok 1

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

Posted 17 December 2020 - 01:30 PM

I looked on Google Scholar for scientific journal articles on "telescope pointing models" and found that there is a quite a literature. 

 

https://scholar.goog...ing model&btnG=

 

The term "pointing error" and the approach of building a model is used on most big telescopes used in astronomy. 

 

Many of the papers are only available if you have an institutional subscription, but this one by Patrick Wallace, one of the more prolific academic authors has a website that gives an overview  http://www.tpointsw.uk/pointing.htm and http://www.tpointsw.uk/gbt2016.pdf

 

Wallace is the author of a program called TPOINT used at many big observatories, and apparently has licensed it to Software Bisque as part of The Sky.  It appears to be the equivalent, for Software Bisque mounts, of the modeling software for 10micron mounts.

 

The various papers on making pointing models for telescopes cover many issues on the thread.  It seems quite likely that, armed with some of the knowledge from these papers, one could make a utility program that helps track down problems.


 

#644 RazvanUnderStars

RazvanUnderStars

    Apollo

  • *****
  • Posts: 1,149
  • Joined: 15 Jul 2014
  • Loc: Toronto, Canada

Posted 17 December 2020 - 05:29 PM

Another thought for the OP: is there any difference in star eccentricity in different orientations of the scope? I'm thinking about what if the filter wheel casing has a slight flexure. AFAIK, ModelCreator doesn't save the images but you can take them separately in various positions.

 

Things to look at are eccentricity both within a frame (suggesting tilt) and across frames.


 

#645 Tonk

Tonk

    Cosmos

  • *****
  • topic starter
  • Posts: 9,434
  • Joined: 19 Aug 2004
  • Loc: Leeds, UK, 54N

Posted 18 December 2020 - 05:53 AM

It seems quite likely that, armed with some of the knowledge from these papers, one could make a utility program that helps track down problems.


You really aught to have a look around the 10Micron users forum as customers have been writing their own software analysis tools and sharing them there. You have to register to gain access but that's no problem as its free.
 

#646 Greg F

Greg F

    Vostok 1

  • -----
  • Posts: 144
  • Joined: 11 Jul 2017
  • Loc: Clover, SC

Posted 18 December 2020 - 11:27 AM

I am late to the party but have some thoughts that might help. 

 

My setup is virtually identical except I have a FSQ130 instead of the 106. I have used The SBIG 16803 the 16200 and currently the QHY600 on the back end. I also run the eagle on top as OP does. My RMS runs anywhere from 1.8 to 2.4 without removing points, all models are 98 points.

 

Things that may help.

 

Tripod -  as strange as this may sound my RMS went down when I fully extended the tripod to its highest point.

 

Rings - get rid of the PrimaLuce lab rings and buy the Takahashi rings. As Tonk has pointed out the felt on some rings can cause problems. My friend in the observatory next to mine had the Primaluce lab rings and kept getting worse and worse models. We finally saw that the felt in the rings was slipping out. The Tak rings are expensive but I get no slippage at all. 

 

Dovetail - yours is much too long and the scope is not centered over the mount. I suspect you have a large amount of flexure here. I realize with the heavy Moonlite and SBIG cameras you are back heavy if centered over the mount, probably 5 pounds or so back end heavy. I will cover this in the balance section. Center the rings over the mount and use a shorter dovetail.

 

Balance - The 10Micron has a built in balance routine. Disregard recommendations on what is acceptable and get both axis to zero, its very easy, it tells you where you are heavy and at what percentage. For RA put one counterweight fully at the top of shaft, you have them both sandwiched together, then move the second weight back and forth to reach 0.00%, mark the spot to help you get back to 0 faster the next time you set up. For Dec you are going to need some weight. I use an ADM accessories V series counterweights  https://www.admacces...counterweights/ . For finetuning I use stainless steel washers with stainless steel lockwashers and bolts to hold them. Make sure your mount can move in every direction and not hit the bolt holding the weights. Place the bracket as close to the mount as possible. there will be a lot of weight but it will be as close as possible to the mount. Currently you have it shoved far forward and hanging on a thin dovetail that is flexing, with the Moonlite and SBIG i'm not sure how you came to balance with the way it is currently setup. Both axis should read 0.00% and test it on both sides. if it is different from side to side or front and back then do not tie your cables to the CW bar but instead run it off the back and secure your cables to the tripod. To get your best RMS you must be properly balanced and it cannot change from side to side or North and South.

 

Camera Tilt - With the SBIG you should not have a problem but check it, if there is tilt then add a tilter into your train, you will have the room. I had no tilt issues with either SBIG but i do with the QHY600, was able to fix with the Gerd Neumann tilter, i also have the Moonlite sidewinder which works well too. I use the SKYX for plate solving with MC and so the SKYX uses cropped images for solving so tilt should not be too much of an issue but with the QHY and the smaller pixels my RMS jumped to 3 - 3.4 . After fixing my tilt issue I am back currently to 2.0RMS .

 

Polar alignment and model building - I can see by your picture that you mark your spots to help with setup, continue this. Your polar alignment even when using your marked spots should be fine. I have found with the 10Micron that super polar alignment is not necessary as long as you have addressed all the above points and also build your model correctly. I have a permanent setup and have experimented with my polar alignment taking it all the way down to 7 seconds. I gained nothing doing this, My current alignment error is 2 minutes 57 seconds RMS at 2 and image at 30 minutes unguided with no problems, the mount just adjusts for it, amazing machine. Of course being picky I will fix the polar alignment I just don't feel rushed to do it anymore like i used to now that I know the mount doesn't care and adjusts for the error. For building a model start with your  3 base points spread them apart as much as possible such as at 330 210 and 90. Place them as low on the horizon as possible. I only image from 35 degrees and up when at the observatory and 40 degrees and up when remote as I cannot lower my wall remotely. My model is built on points 35 degrees and up. I use a 5 second delay, I do think I could use a 2 second delay no problem, I use a Lum filter but also I will use a red filter for better contrast sometimes but I have never seen a difference in my final results. I expose for 5 seconds to help with seeing conditions but again I think 3 would be fine, i bin 2x2 for modeling. Oh make sure your refraction parameters are set which i believe you already have set. I never do the number of columns and rows but instead always use the evened spaced box which gives me 95 points plus the 3 original for 98. It may take 2 times as long as the 5x5 rows but 30 minutes more time is worth it plus no polar alignment adjustments to check and change and then redo model. Just do the 98 total and be done with it. Set your minimum altitude to the lowest you can with out hitting trees or homes.

 

Building the model is not the problem its all the factors before you build the model - the rings and dovetail, the balance cable management. I know many believe that tying to the CW bar is the way to go as I believe that is what Per did but I have found better results running everything off the back with the cable inside a cable tube and then tied to the mount pier base in the center using cable ties. You have one cable tube that maintains its shape that moves freely from side to side consistently instead of it laying over the CW bar at certain positions when you tie it to the bar.

 

I am sure I forgot a few things but this hopefully will help. Oh did I mention, get rid of those rings and get a shorter dovetail and balance properly over the center of the mount.


 

#647 SimonIRE

SimonIRE

    Apollo

  • -----
  • Posts: 1,283
  • Joined: 16 Dec 2018
  • Loc: South East England

Posted 18 December 2020 - 11:32 AM

I am late to the party but have some thoughts that might help. 

 

My setup is virtually identical except I have a FSQ130 instead of the 106. I have used The SBIG 16803 the 16200 and currently the QHY600 on the back end. I also run the eagle on top as OP does. My RMS runs anywhere from 1.8 to 2.4 without removing points, all models are 98 points.

 

Things that may help.

 

Tripod -  as strange as this may sound my RMS went down when I fully extended the tripod to its highest point.

 

Rings - get rid of the PrimaLuce lab rings and buy the Takahashi rings. As Tonk has pointed out the felt on some rings can cause problems. My friend in the observatory next to mine had the Primaluce lab rings and kept getting worse and worse models. We finally saw that the felt in the rings was slipping out. The Tak rings are expensive but I get no slippage at all. 

 

Dovetail - yours is much too long and the scope is not centered over the mount. I suspect you have a large amount of flexure here. I realize with the heavy Moonlite and SBIG cameras you are back heavy if centered over the mount, probably 5 pounds or so back end heavy. I will cover this in the balance section. Center the rings over the mount and use a shorter dovetail.

 

Balance - The 10Micron has a built in balance routine. Disregard recommendations on what is acceptable and get both axis to zero, its very easy, it tells you where you are heavy and at what percentage. For RA put one counterweight fully at the top of shaft, you have them both sandwiched together, then move the second weight back and forth to reach 0.00%, mark the spot to help you get back to 0 faster the next time you set up. For Dec you are going to need some weight. I use an ADM accessories V series counterweights  https://www.admacces...counterweights/ . For finetuning I use stainless steel washers with stainless steel lockwashers and bolts to hold them. Make sure your mount can move in every direction and not hit the bolt holding the weights. Place the bracket as close to the mount as possible. there will be a lot of weight but it will be as close as possible to the mount. Currently you have it shoved far forward and hanging on a thin dovetail that is flexing, with the Moonlite and SBIG i'm not sure how you came to balance with the way it is currently setup. Both axis should read 0.00% and test it on both sides. if it is different from side to side or front and back then do not tie your cables to the CW bar but instead run it off the back and secure your cables to the tripod. To get your best RMS you must be properly balanced and it cannot change from side to side or North and South.

 

Camera Tilt - With the SBIG you should not have a problem but check it, if there is tilt then add a tilter into your train, you will have the room. I had no tilt issues with either SBIG but i do with the QHY600, was able to fix with the Gerd Neumann tilter, i also have the Moonlite sidewinder which works well too. I use the SKYX for plate solving with MC and so the SKYX uses cropped images for solving so tilt should not be too much of an issue but with the QHY and the smaller pixels my RMS jumped to 3 - 3.4 . After fixing my tilt issue I am back currently to 2.0RMS .

 

Polar alignment and model building - I can see by your picture that you mark your spots to help with setup, continue this. Your polar alignment even when using your marked spots should be fine. I have found with the 10Micron that super polar alignment is not necessary as long as you have addressed all the above points and also build your model correctly. I have a permanent setup and have experimented with my polar alignment taking it all the way down to 7 seconds. I gained nothing doing this, My current alignment error is 2 minutes 57 seconds RMS at 2 and image at 30 minutes unguided with no problems, the mount just adjusts for it, amazing machine. Of course being picky I will fix the polar alignment I just don't feel rushed to do it anymore like i used to now that I know the mount doesn't care and adjusts for the error. For building a model start with your  3 base points spread them apart as much as possible such as at 330 210 and 90. Place them as low on the horizon as possible. I only image from 35 degrees and up when at the observatory and 40 degrees and up when remote as I cannot lower my wall remotely. My model is built on points 35 degrees and up. I use a 5 second delay, I do think I could use a 2 second delay no problem, I use a Lum filter but also I will use a red filter for better contrast sometimes but I have never seen a difference in my final results. I expose for 5 seconds to help with seeing conditions but again I think 3 would be fine, i bin 2x2 for modeling. Oh make sure your refraction parameters are set which i believe you already have set. I never do the number of columns and rows but instead always use the evened spaced box which gives me 95 points plus the 3 original for 98. It may take 2 times as long as the 5x5 rows but 30 minutes more time is worth it plus no polar alignment adjustments to check and change and then redo model. Just do the 98 total and be done with it. Set your minimum altitude to the lowest you can with out hitting trees or homes.

 

Building the model is not the problem its all the factors before you build the model - the rings and dovetail, the balance cable management. I know many believe that tying to the CW bar is the way to go as I believe that is what Per did but I have found better results running everything off the back with the cable inside a cable tube and then tied to the mount pier base in the center using cable ties. You have one cable tube that maintains its shape that moves freely from side to side consistently instead of it laying over the CW bar at certain positions when you tie it to the bar.

 

I am sure I forgot a few things but this hopefully will help. Oh did I mention, get rid of those rings and get a shorter dovetail and balance properly over the center of the mount.

 

Very useful info. Thanks for putting this together. 


 

#648 mccomiskey

mccomiskey

    Explorer 1

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

Posted 18 December 2020 - 12:15 PM

I am late to the party but have some thoughts that might help. 

 

My setup is virtually identical except I have a FSQ130 instead of the 106. I have used The SBIG 16803 the 16200 and currently the QHY600 on the back end. I also run the eagle on top as OP does. My RMS runs anywhere from 1.8 to 2.4 without removing points, all models are 98 points.

 

Things that may help.

 

Tripod -  as strange as this may sound my RMS went down when I fully extended the tripod to its highest point.

 

Rings - get rid of the PrimaLuce lab rings and buy the Takahashi rings. As Tonk has pointed out the felt on some rings can cause problems. My friend in the observatory next to mine had the Primaluce lab rings and kept getting worse and worse models. We finally saw that the felt in the rings was slipping out. The Tak rings are expensive but I get no slippage at all. 

 

Dovetail - yours is much too long and the scope is not centered over the mount. I suspect you have a large amount of flexure here. I realize with the heavy Moonlite and SBIG cameras you are back heavy if centered over the mount, probably 5 pounds or so back end heavy. I will cover this in the balance section. Center the rings over the mount and use a shorter dovetail.

 

Balance - The 10Micron has a built in balance routine. Disregard recommendations on what is acceptable and get both axis to zero, its very easy, it tells you where you are heavy and at what percentage. For RA put one counterweight fully at the top of shaft, you have them both sandwiched together, then move the second weight back and forth to reach 0.00%, mark the spot to help you get back to 0 faster the next time you set up. For Dec you are going to need some weight. I use an ADM accessories V series counterweights  https://www.admacces...counterweights/ . For finetuning I use stainless steel washers with stainless steel lockwashers and bolts to hold them. Make sure your mount can move in every direction and not hit the bolt holding the weights. Place the bracket as close to the mount as possible. there will be a lot of weight but it will be as close as possible to the mount. Currently you have it shoved far forward and hanging on a thin dovetail that is flexing, with the Moonlite and SBIG i'm not sure how you came to balance with the way it is currently setup. Both axis should read 0.00% and test it on both sides. if it is different from side to side or front and back then do not tie your cables to the CW bar but instead run it off the back and secure your cables to the tripod. To get your best RMS you must be properly balanced and it cannot change from side to side or North and South.

 

Camera Tilt - With the SBIG you should not have a problem but check it, if there is tilt then add a tilter into your train, you will have the room. I had no tilt issues with either SBIG but i do with the QHY600, was able to fix with the Gerd Neumann tilter, i also have the Moonlite sidewinder which works well too. I use the SKYX for plate solving with MC and so the SKYX uses cropped images for solving so tilt should not be too much of an issue but with the QHY and the smaller pixels my RMS jumped to 3 - 3.4 . After fixing my tilt issue I am back currently to 2.0RMS .

 

Polar alignment and model building - I can see by your picture that you mark your spots to help with setup, continue this. Your polar alignment even when using your marked spots should be fine. I have found with the 10Micron that super polar alignment is not necessary as long as you have addressed all the above points and also build your model correctly. I have a permanent setup and have experimented with my polar alignment taking it all the way down to 7 seconds. I gained nothing doing this, My current alignment error is 2 minutes 57 seconds RMS at 2 and image at 30 minutes unguided with no problems, the mount just adjusts for it, amazing machine. Of course being picky I will fix the polar alignment I just don't feel rushed to do it anymore like i used to now that I know the mount doesn't care and adjusts for the error. For building a model start with your  3 base points spread them apart as much as possible such as at 330 210 and 90. Place them as low on the horizon as possible. I only image from 35 degrees and up when at the observatory and 40 degrees and up when remote as I cannot lower my wall remotely. My model is built on points 35 degrees and up. I use a 5 second delay, I do think I could use a 2 second delay no problem, I use a Lum filter but also I will use a red filter for better contrast sometimes but I have never seen a difference in my final results. I expose for 5 seconds to help with seeing conditions but again I think 3 would be fine, i bin 2x2 for modeling. Oh make sure your refraction parameters are set which i believe you already have set. I never do the number of columns and rows but instead always use the evened spaced box which gives me 95 points plus the 3 original for 98. It may take 2 times as long as the 5x5 rows but 30 minutes more time is worth it plus no polar alignment adjustments to check and change and then redo model. Just do the 98 total and be done with it. Set your minimum altitude to the lowest you can with out hitting trees or homes.

 

Building the model is not the problem its all the factors before you build the model - the rings and dovetail, the balance cable management. I know many believe that tying to the CW bar is the way to go as I believe that is what Per did but I have found better results running everything off the back with the cable inside a cable tube and then tied to the mount pier base in the center using cable ties. You have one cable tube that maintains its shape that moves freely from side to side consistently instead of it laying over the CW bar at certain positions when you tie it to the bar.

 

I am sure I forgot a few things but this hopefully will help. Oh did I mention, get rid of those rings and get a shorter dovetail and balance properly over the center of the mount.

Greg,

 

Extremely helpful and informative.  Thanks much.  I am jealous of the FSQ130.  By the time I started to focus (pun intended) on astrophotography, those were no longer available.  Any chance you would be willing to share a photo of your setup?  I am always interested in seeing how other people do things, as I continue to search for best practices and solutions to adopt.

 

A few questions/comments:

  • I don't believe there are Tak rings for the FSQ106.  I think that they just make clamshell, which I assume would be inferior for these purposes.  I do have a set of Moonlite rings, which I believe are well regarded for this scope, so can move to those.
  • Balance - in its current position, the OTA is balance on both axis to 0.1%.  I had read somewhere that it was better to leave a slight imbalance (which never made sense to me for a mount with absolute encoders, but equally, I had assumed 0.1% out was not material, so why not).  I will get it to 0.0% on my next try.
  • I assume the rationale for putting the rings over the center of the dovetail is that the ring/dovetail interface is where the flexure is being initiated, and by positioning them at the center, this is minimized?  If so, does adding weight on to the front of the dovetail recreate this issue?  Are there other reasons to seek out this positioning?
  • In the FWIW category, I, like you, use 2x2 binning, 2 second settle, 4 second exposures with the Lum filter and a 60% crop for model building.  I have wondered whether an effective 18micron pixel with a sampling of 7" to a side might be part of the issue with my higher RMSs, and whether i shouldn't move to 1x1 binning.  Thoughts on that front?
  • I will give the cable off the back approach a try, too.

 

Many thanks again. 

 

Unfortunately, I will be away from my scope for the next two weeks, but this will give me much to work on in January.


 

#649 Tonk

Tonk

    Cosmos

  • *****
  • topic starter
  • Posts: 9,434
  • Joined: 19 Aug 2004
  • Loc: Leeds, UK, 54N

Posted 18 December 2020 - 12:18 PM

I have found better results running everything off the back with the cable inside a cable tube and then tied to the mount pier base in the center using cable ties.


Ditto - I only have 3 cables in my cable mesh - 12v power, network and 10Micron remote on/off.

 

All the USB stuff is on the rig with the NUC computer and USB hubs. That's how to solve cable management - otherwise I'd have 19 USB cables dragging about


 

#650 mccomiskey

mccomiskey

    Explorer 1

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

Posted 18 December 2020 - 12:30 PM

I stand corrected.  Tak does make rings.  Tracking down those now.


 


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