10-Micron. Coping with OTA swaps

[ollypenrice]

No.

It should be, but hey-ho.

I think you would need to remodel with every change of scope but I dare say Per will be along.

Olly

[CCD Imager]

Hi Olly

Yes, you need a model for each OTA, but these can be saved, so recalled for a scope swap over. But there are potential complications like changed focuser slop, slightly different orthogonality etc. Probably best to do a fresh one, and Per will tell you it only takes 20 minutes whilst you have coffee, which I reckon he drinks gallons of

I've just taken delivery of a GM2000 HPSII and may even be bringing it down to you next year

Adrian

[Big Bang!]

Evening all,

If this helps...  I took delivery of a GM1000HPS from Ian King about two months ago. I cannot remember the exact figure but it was ballpark £5.5k - he gave me a better price than I was expecting, but then again, the total order was over £12k.

I upgraded from an LX90 12".  I wanted a GM2000 but just couldn't go the distance.  However, the GM1000 has given me everything I expected and more.  The unguided tracking accuracy is astonishing. I have an Altair RC250TT on it and on top of that I have a 70mm APO. Including the Atik 9 I am fairly close to the 25kg limit but the electronic balancing is great! It is permanently sited in an Obs, polar aligned to 31" and the 30 star model does the rest (after the first 10 stars very little centering was necessary). It goes where I point it and it stays there, first time, every time. As an absolute novice to AP that was very important to me because I just don't know enough to be able to solve problems easily - I needed a rig that was easy to use out of the box.  So, I have gone from nothing to a superbly accurate, remote operation in one go and this mount is the foundation of it.

I did initially consider the Paramount because of its capacity but the unguided accuracy of the 10 Micron is what swayed it.  The Paramount is no doubt excellent; all I can say is you will not be disappointed with a 10 Micron.

Gus

[CCD Imager]

My view is that both the Software Bisque mounts and 10 micron mounts are both excellent, but what really swayed me to get the 10 micron was the absolute encoders. They are not yet available from SB yet, although in the pipeline. But when they are, they will add considerable cost, making the 10 micron mounts look excellent value for money. Absolute encoders offer so much to the Astrophotographer that I am over the moon to now have them

No periodic error, the mount knows where it is, power on or off, long unguided imaging. There will be no excuse for poor images from now on...

Adrian

[Big Bang!]

I concur - an absolutely excellent feature.  

If anybody in the Fareham area wants to come and have a hands-on look/play, by all means PM me.  When you are spending this kind of money you really need to KNOW what you are getting before you part with it.

Gus

[ChrisLX200]

A lot of 10-Micron owners here! I've had mine for a week or so but have yet to set it up on my pier (or even power it up for that matter). The AP16000 also lists absolute encoders as an option, either factory fitted or as an end-user fitted add-on. Given the mechanical accuracy requirements I'm not sure how having the user fit them will work...

ChrisH

[perfrej]

The AP encoder solution is very different as the encoders are mounted in such a way that they lose contact with the axis when the clutches are disengaged. If you bump your encoded AP mount you have to sync the encoders, and every time you power it on you have to sync them as well. In the 10Micron mounts, the encoders are on the axis, so any de-clutching or power off has no effect on their knowing their position.

An encoder has to be mounted with micron-precision, so I, for one, do not believe in customer-mounted encoders. You may also note that encoders on AP and Bisque mounts set you back $3000+ per axis... The Bisque solution is yet to be released, and the AP solution really doesn't add anything to the AP mounts.

A "proper" encoder solution includes the encoders as an integral part of the actual motor control loop, something that very few mounts can brag about. Using them as "virtual guiders" really isn't the way to go.

As for models, yes, every scope needs its own model - no way around that. Under good conditions, a model previously used with a certain scope can be re-aligned when the scope once again is on the mount, but you cannot re-use a model if the mount positioning has changed.

These mounts are sub arc-second precision. If a tripod leg sinks one millimeter into the ground, the pointing is offset by more than four arc-minutes! It is, however, an offset, so the model can be re-aligned, but it also gives a sense of the precision involved.

/per

[ChrisLX200]

Thanks for the insight Per! I shall be making sure my GM2000HPS doesn't move at all - I just finished machining the pier adapter and it's... solid. :-)

ChrisH

[Zakalwe]

Thats going nowhere!

[ollypenrice]

Dare I wonder if I'm going to be rehabilitated into astronomical humanity given the large number of defectors from the EQ8, here and elsewhere?

Do I dare post this question? I don't scare easily so here goes!

lly

[Zakalwe]

Dare I wonder if I'm going to be rehabilitated into astronomical humanity given the large number of defectors from the EQ8, here and elsewhere?

Do I dare post this question? I don't scare easily so here goes!

lly

Peeps can't think about upgrading anymore without being called a "defector"?

[ollypenrice]

I'm only joking. 

The real problem with mounts, models and guiding is that in one revolution of the wheel there are 1296000 arcseconds so when we want to be down to half an arcsec error we are asking for one part in nearly 2.6 million. When you think about this it's a tall order.

Olly

[perfrej]

Modeling is an almost analog computing model and relies on many samples taken, from which rediculous accuracy can be obtained. The more samples, the better the accuracy. For a 10micron mount, a model of between 25 and 100 points is sufficient, and the number of points required is determined by the different factors of your rig behavior.

The internal result of a model is NOT a number of corrections to be applied on different parts of the sky, but rather a set a "model terms", usually 7 to 11. The terms are constants applied to a basic equation for each of them. One example is the gravitational ones which model the sag in your rig. They depend on the altitude that your scope is pointing at, so cosine gets involved, making the error go gradually worse as you approach the horizon. A simplified equation for that could be something like n * cos(alt), where n hence is the model term.

When you add a refine point to your model, the entire model is re-calculated by more or less zapping it and adding the point coordinates for all the previously entered points. After a model is calculated it will contain the term constants to apply to the equations. Look at the model as a best fit polynomial, like when you do a curve fit in Excel in order to find a mathematical representation of sampled data in a graph.

The mount model uses as many terms as are appropriate, up to the 11 ones it knows of. If it finds that the model terms fit on one side of the meridian but not the other it will split the model in two and say that it has up to 22 terms. The terms are also applied to negative values, which means that the model is valid even if points added have the weights up or if a model is used weights up (the mounts do up to 30 degrees of weights up).

When a model is ready, each entered point is compared to the mathematical results and given an error in respect to the model. All such errors are processed by a standard root mean square algorithm and the expected pointing error RMS is presented. Obviously, a low value is better, but it is also interesting to look at each point's individual error in respect to the model and find bad ones with high errors. Sometimes it is better to remove them, and sometimes it is not - it all depends and you have to try it out. This is no obstacle as you can save the model within the mount, delete the point in question and then recall it if the result is not to your liking.

So, in response to the actual sub arc-second precision discussed above; yes, it is possile!

A few points:

• A model consists of mathematical constants applied to a set of equations
• 10Micron uses up to 11 terms
• The model is split if the two halves of he sky behave differently
• Atmospheric refraction is de-calculated before a point is added and re-calculated as the mount applies the model
• The terms model consistent flexure, but not erratic or random
• The model is used both for pointing and tracking
• The mounts track in both Ra and Dec
• The encoders are truly absolute and have over 10 million ticks per rev which is software interpolated to even higher values
• Unguided imaging this way is no game and requires a steady rig

When you have a good telescope, without mirror flop, a good focuser, good adapters and a steady set of rings to hold it in place you can do all this. Do not expect to get good results with a stock cheapo focuser or a low cost RC with a primary rattling in its cage. Lock your Celestron mirror and add a separate high-quaity focuser at the cassegrain outlet, etc, etc.

Nothing like a morning rant, is there?

Al the best,

Per

p.s. The image below is from right now and geo-referenced...

[CCD Imager]

Hi Per

You are no doubt getting ready to introduce some new modelling terms for water ripple by the looks of things

The 10 micron mounts are perfectly set up for unguided imaging - no periodic error, no backlash, solid mount. You are just at the mercy of what telescope you put on the mount and its inherent flexural and flop problems. Basically, the shorter the FL and the more expensive (relatively speaking) your scope, the longer you can go unguided. A refractor is better as the optics are fixed (some cheap ones might wobble!). I think the two biggest issues are optic flop and focuser flop. Flexural issues are usually repeatable and can be modelled out.

Adrian

[perfrej]

Adrian, you are right n all that. It is all up to what you put on the mount, simple as that. Now, you may be thrilled to hear that the one-hour subs that I have published as an example (M106), were shot with a Skywatcher 190MN telescope. I did, however, modify the primary mirror holder in order to stabilize it a bit... So, in essence, a simple and cheap scop may, if analyzed and treated properly, be OK.

/per

[Zakalwe]

Unguided imaging doesn't bother me to be honest. Guiding isn't an issue, and its certainly not a deciding factor.

[ChrisLX200]

Unguided imaging doesn't bother me to be honest. Guiding isn't an issue, and its certainly not a deciding factor.

Going 'Unguided' is just the bi-product of having extremely accurate tracking, and the more accurate the tracking the better the final image (and less subs lost to poor guiding). Accurate modelling provides the basis of accurate tracking, and another bi-product of that is not having to plate-solve in order to place your target perfectly on the sensor. I think you have to struggle with less than perfect tracking/pointing for a few years to really appreciate the benefits, and if you want to run remotely without issues then you need a mount like this.

ChrisH

[CCD Imager]

Being able to not use a guider is a reflection of the overall quality and capabilities of the mount, so I agree with Chris. BTW Chris, I think you need to change your name

Adrian

[ChrisLX200]

Being able to not use a guider is a reflection of the overall quality and capabilities of the mount, so I agree with Chris. BTW Chris, I think you need to change your name

Adrian

I think I'm stuck with it - I use it on many forums, and anyway I still have the LX200!

ChrisH

[Sp@ce_d]

Of course! I hadn't thought about plate solving too. That & trying to find a usable guide star, especially with an OAG at longer FL, are the most common time consuming & causes of failures I find running remotely with ACP!

Sent from my iPhone using Tapatalk

[JuergenB]

Coming back to the original question, I ask myself whether I would be able to exactly reproduce the position of my telescope tube when I have to fasten the felt-lined tube rings, clam shells or whatever after changing the scope. I hope that for example the Parallax rings for my Baby-Q are able to keep the arc second pointing accuracy over a couple of days with my semi-portable setup. Of course I believe that the telescope tube will be kept stable during the usual exposure times of 10-30 minutes, hopefully also during one night when temperatures are dropping. Any experiences with felt-lined tube rings here?

Juergen

[ChrisLX200]

The felt should compress to the point where movement is prevented, it is by no means a flexible 'padding' - but I suppose much also depends on the distance between the two rings. I don't anticipate problems here with my NP127is which is also mounted in this way with twin felt-lined rings. However, there will always be a tiny amount of thermal contraction along the length of the tube as temperature drops and the felt should allow for this degree of flexion (you wouldn't be able to prevent it anyway, and even trying would produce adverse effects).

ChrisH

[perfrej]

Even when going portable with a 10Micron mount, building a model from scratch is something that always fits in the time window between dark enough for plate solves and dark enough for imaging - with some time to spare. Everyone I know of who is portable always gets a full model run in that time window, usually while having a cup of coffee or something. It is all automatic and very simple. I wouldn't even bother reusing models...

I was asked to do a 10Micron presentation on the Astro Imaging Channel on Google Hangouts (it is on Youtube as well) and did a model run from my remote in Provence. There isn't much to it. https://www.youtube.com/watch?v=WfkQnZTFXzo, 56 minutes in approximately.

/per

[ollypenrice]

Guiding is very elegant in my view. A real-time feedback loop based on what the scope is actually doing. I would accept the idea that a perfectly constructed mount tracking at perfect sidereal rate would be more elegant if it were going to work - but it isn't. Once you have to correct for atmospheric refraction variable with altitude then you have to depend on a software sky model and encoders and, essentially, the 'elegance' of the system begins to evaporate. Is it more elegant to create a feeback loop based on a star actually being observed in real time or to build a feedback loop based on a model which is blind to real circumstances?

Nope, I think guiding is simple and elegant.

Olly

[swag72]

I agree with this Olly - I don't see the attraction of being able to gather 1 hour subs without guiding - I guide and my guiding works night after night. If I had such a mount I wouldn't for one minute think of not guiding. Perhaps I'm wrong, but I really don't see the issue with it nor the attraction of not doing it.