Choosing the best astrograph

[Stefan Buda]

A personal view on a rainy evening.

I deliberately used the word best in the title because it is important to understand that there is no such thing.
The range of AP is so wide that there is no astrograph that can be best at everything. What's best for wide field is not best for deep sky, and what is best for deep sky is not best for planetary.
One needs to divide up this range into several domains based on criteria such as field of view or resolution. After that choosing the "best" astrograph becomes easy.

I always urge people to start with the sensor they have, or intend to use. It sounds like a back to front approach but it make sense because sensors are evolving, while astrographs are not, and if one ignores that, one may waste money on buying the wrong instrument. It is likely to work out cheaper to match the scope to the sensor then vice versa, or end up with a mismatch.

For wide field imaging there are many possible combinations.
Once the sensor size is settled one needs to choose the desired field of view. With those two parameters one can work out the necessary focal length. From there it is a simple matter of choosing the largest aperture available for the calculated focal length.

For deep sky imaging things get more interesting.
Instead of starting with sensor size one needs to start with pixel size and work out the optimum focal length, based on empirically established best practice sampling rate of about one arc second per pixel. Again, it is just the matter of choosing the largest available aperture on the market, for the calculated focal length.
The interesting bit here is that you will end up with the best deep sky astrograph and yet it will not necessarily be the largest on the market. This calls aperture fever into question. And light gathering arguments too, unless you are going after Quasars.

I won't discuss planetary imaging as that is quite a different ball game and best done with telescopes rather than astrographs.

[gregbradley]

Interesting post Stefan.

I have approached it similarly but not exactly the same way.

I work backwards from the type of image I want to take. I look at lots of example images I like and then note what equipment they used and that starts as the base.

It does end up where you said though.

For widefield sensor pixel size seems less sensitive than at longer focal lengths. For example there are lots of FSQ106 images using large 9 micron pixelled cameras like the 16803 and 11002. See Mike's recent FSQ images for example.

I have found 1 arc sec/ pixel to work well also Roland Christen recommends that. On my CDK17 though I have found .43 arc sec/pixel with 6 micron pixels works even better but that is at a dark site and often with good seeing at highish altitude.

Widefield seems optimised around the 4 inch APO band like FSQ106 or other 4 inch true APOs. Hard to beat.

A 5 and 6 inch APO also works quite well but they get less widefield as you go up in aperture, become a lot more expensive (although an FSQ106 is not exactly cheap at about $7500).

Newtonians are popular as midrange focal length scopes like a 10 inch F4 type formula. They give very tight stars and sharp images. Fussy to collimate and prone to flexure are their weaknesses.

The trend though has been for astrographs to be quite wide aperture and fast F ratio.

Faster than about F3.8 though is asking for trouble. Everything mechanically and optically gets way harder.

I think I would be happy with a 105mm F6 CFF or a Tak FSQ106EDX4 and a long focal length CDK (oh wait, that is what I have hehehehe). Plus a large aperture fast scope for quick imaging in our limited time type imaging scene (for most anyway).Throw in a some nice lens imaging and its mostly covered except for planetary.

FSQ130 has hardly any example images and the ones I have seen are very good but not jaw dropping. Same with TV NP127i, very good, not the best around. So 4 inch APO seems to be a sweet spot so are 6 inch APOs (closest to an allround instrument with a reducer and a flattener). TEC140 is also another fairly good allrounder. Tak TOA seem popular but also are resold often (I think because they are so heavy and front heavy).

Greg.

[Stefan Buda]

My post did not consider any practical aspects of choosing the right instrument. It is simply my back to front way of going about it, from a theoretical point of view, so that one ends up with the right combination of sensor and scope. I guess the main conclusion of the exercise is that the size of the sensor should determine the focal length, when used for wide field (based on preferred field of view). And the size of the pixel should determine it, when used for deep sky. In both cases aperture should be maximised.
The surprising bit is that the sensor puts a limit on the size of the scope. You can waste money on buying a larger scope (longer focal length) without any tangible gains.
That is why one needs to be careful with aperture fewer. In other words aperture rules, but only if it is matched with the right sensor.

I agree with your comments on various combinations that you mention because the practical world is a bit more complicated than the outlined theory which contains a number of assumptions. For instance theory suggests that the FSQ106 should be best combined with smaller pixels than the 9 micron ones that people use so successfully. This particular instrument has been redesigned more than once and I've seen spot diagrams of an earlier version. I was surprised how large the spots were. It seems that they designed it like a photographic lens. They compromised the centre in order to be able to produce a uniformly corrected field. That may explain why it works so well with large pixels.

[Paul Haese]

My way of looking at this is that like screw drivers you need more than one type and size in your tool shed.

For planetary my choice was and still is a compound scope. At present and for the last 10 or so years is the C14 that I peltier cooled. It gives a great focal length that when combined with a barlow and nice sized pixels on a camera can mean I am imaging around 12m in focal length which in turn gives a really nice image size.

For DSO, well I have had a few telescopes but my main way of thinking is that you need to mix aperture/focal length with pixel size and image scale. I really like long focal length imaging, but you need big pixels to maintain a reasonable image scale. Something around .70-.90" per pixel I think is really good. Pixel size which is fairly close to f ratio in most circumstances produces the best results. For example I am using an f8 RC with 9um pixels. It makes for a nice image scale. Is it the best astrograph? No. It's slow photographically but does give nice sharp detail that does not require much sharpening. However it takes forever to gather very faint light from halo's around planetary nebula for example.

You can do something similar with focal length/image scale and a fast f ratio. At present I have another imaging system that has a fast f ratio of 4/3.8 with small pixels of 4.54um. This again gives a nice image scale of .82" per pixel. It's fast photographically, but the fine detail on some objects is not as nice as with a long focus scope. So almost the complete reverse of the above. Is this the best astrograph? Again the answer is no, although I can use this design to do both wide field and narrow field image, which is slightly better than the RC.

I have owned a variety of refractors but never really thought it worth looking at the image scale. I wanted these for imaging extended objects, so I was going for a field of view look. I know this does encompass a much larger image scale but it was not important to my decision. My decision was based on the size of the field and what I could afford to achieve that.

So again is there a best astrograph? I guess it all depends on what you are trying to achieve with your imaging. What you specifically want. Factors of focal length, f ratio, pixel size and resulting image scale are important.

[Peter Ward]

While we can ruminate on pixel size and focal length, I have found for "diffraction limited" optics...and Lord knows that is not a given :rolleyes:
that so many other aspect conspire to smudge your images, far more than an imperfect optic.

eg:
Environmental factors (seeing, local environmental temperature, transparency )

Mechanical factors: Tracking accuracy, guiding accuracy, wind buffeting and vibration suppression, OTA rigidity, collimation retention, focuser rigidity.

Thermal control: Suppression of tube currents, optical figure retention, focus retention.

.....and apologies to Stefan for the digression..

Best Astrograph? Simple.

*Anything* made by Astro-Physics. :D

[Stefan Buda]

I also said that wide field, deep sky and planetary, all require a different theoretical approach in optimizing the setup.

You use "image scale" and I use "sapling rate", a more general term, but we mean the same thing.

Your approach of associating focal ratio with pixel size is very interesting. I never looked at it from that angle. I will have to check the math.



That is precisely the problem with choosing a good setup. There are so many factor to consider that beginners may not know what are the most important.
My reverse method is only the start of the selection process and it has been simplified to the max. For a wide field setup, it is based on two parameters only - sensor size, and desired field of view. For deep sky it is based on a single parameter - the pixel size of the sensor.
Once the best match focal length has been calculated and set into concrete, the difficult process of choosing the largest aperture can begin. It is at this stage that all other factors must be considered. I must add that a bit of departure from the calculated optimum FL is not a disaster even if I just had it set in concrete.

Perhaps someone could set out some easy to follow rules on how to proceed with the aperture maximisation.
At this stage we no longer need to consider the sensor. We only have one parameter to consider and that is the focal length. Let's ignore pocket size as a parameter.
The next logical step would be to establish the astrograph type.
We do have a focal length but we don't want to restrict it to any particular "screw driver" size. I think, the only way forward, without complicating things too much, would be to establish several FL ranges, and treat them separately.
Maybe at the end we can come up with a flow chart that starts with a sensor, covers a lot of factors and ends up with several "best" astrographs. Would there be any made by Astro-Physics among them?:D

[Slawomir]

Interesting thread - thank you all for sharing your experiences and points of view.
From my limited experience, because theoretical parameters can be compromised during manufacturing process, mechanical quality has climbed up to the top of my priorities when choosing an astrograph within a given budget.

[multiweb]

Interesting discussion. I tend to agree with Paul comments on image scale having used a wide variety of cameras, coupled with different apertures and focal lengths. My main concern at this point with my imaging journey is keeping collimation because I'm mobile and that becomes an issue for me with aperture.

[Startrek]

I’m a backyard astronomer and usually set up for a week at a time at both my Sydney residence and south coast holiday home. I check collimation on all my newts each time I use them , doesn’t takes long just a minute or so and forms part of my of procedure check list when either imaging or observing. Believe it or not my 6” tends to drift out of collimation more so than the 8” and 12”

[Stefan Buda]

Guys, please don't run ahead, or the thread will degenerate into the usual obscure tangle of opinions and we achieve nothing.
Let's see if it is possible to cut through the chaos and come up with a step by step procedure.
Remember that at this stage we are only have focal length to work with and we have to choose the astrograph type, not the make/brand. We can deal with those issues later, in a systematic way.

I propose that we cut out anything below 400mm FL, as that range would just generate a messy discussion of photographic lenses.

If we use exponential series, I think we can get away with three domains: 400 to 800; 800 to 1600; 1600 to 3200.

So we need to choose the best astrograph type for each range.
These are theoretical astrographs at this stage and are meant to be capable of producing near perfect fields that cover a full frame sensor.

Sorry Peter, your setup is out. People that use extreme gear like yours, don't need this kind of help.:) But your input would be appreciated.
Knowledge of optical systems is very helpful at this stage.

[Slawomir]

400-800mm domain.

I believe a 10" f/3 Newtonian would be one of strong potential candidates. Large enough aperture to be always seeing limited in average conditions, providing good resolution and also speed for extended objects. I believe TV makes coma correctors capable of correcting such fast Newtonians.

A potentially easier to manage (collimation etc) alternative could be a 130-140mm petzval. Possibly still large enough aperture to remain seeing limited in most typical backyard scenarios, when paired with right-sized pixels.

[gregbradley]

The only thing with aperture gets limited by pixel size is I haven't observed that to be true.

For example look at CDK20 images by M and T on this site and look at Russell Cromans STL11 and RC20 inch images. Of course there are benchmark images by Adam Block using a 24inch RC and 9 micron pixel CCDs. So 9 micron CCDs seem a bit of a jack of all trades for astro photography.

6 micron pixels as in the KAF16200 and 5.4 micron pixels as in the KAF8300 are also fairly universal in producing fine images but seeing starts to play a noticeable role.

I am finding I am getting sharper images using 6 micron KAF16200 and a CDK17 than with the 9 micron 16803 sensor. But that is under better seeing conditions.

I also once took the same image with both an 8300 sensor and a 16803 sensor on the CDK17 under reasonable usual seeing. I was surprised to see how much worse the 8300 image was. I didn't expect that. So that is showing what you mentioned in your original post. Seeing effects images more when poorly sampled.

Greg.

[gregbradley]

Takahashi Epsilon 180ED F2.8. I believe if the collimation goes out its a trip back to Japan.

Also looking at images from the 180 ED compared to same using FSQ106ED the FSQ images were quite a bit sharper with smaller stars.

The advent of these low read noise CMOS cameras seems to have changed the equation a lot as well.
Personally I would be interested in that approach once the sensors get larger.

Greg.

[Paul Haese]

1600-3200 range I like the corrected DK scopes and RC scopes. Both are great designs in my opinion.

800-1600 domain I like imaging Newtonians or Harmer Wynne scopes. More inclined to like the later though.

[Stefan Buda]

Your post made me realize that dividing the focal range into just a few domains is not a very good approach. In fact the more domains we set the more precise the outcome and probably more useful too.
A range of 400 to 800 is too large. It can accommodate the corrected Newtonian, as you suggested, but only at the upper end of the range and that would be a fuzzy outcome. Perhaps 400 to 600 would yield clearer results? It could include corrected prime focus catadioptric systems as well as Petzval or corrected APOs. But please don't discuss the merits of the various configuration at this stage. We need to have a list of candidates for each range before it becomes useful to rank them.

[Stefan Buda]

As I have just posted, I think we need more restrictive domains for more useful results. And perhaps work trough them in order so we can all keep focus on the same aspects.

[Stefan Buda]

Ok, this takes us back to the beginning. No problem. Let's find where the theory went wrong and make amends. If we can't agree on the theory then there is no point in proceeding any further.

[Paul Haese]

500-700 ED refractor (2 element) Apochromatic (3 element) or Petzval (4 element)
700-1200 Newtonian (imaging) or Harmer Wynne (fully corrected), Both around f4
1200 + RC (with flattener), corrected Dall Kirkham, traditional Cassegrain

Is this what you are thinking?

[Stefan Buda]

Yes, that is better but the 8" RASA would fall off range. I think wee need to start lower. Also let's agree on the theory first.

[Startrek]

I bought a new Bintel 8” f5 newt ($460 ) in March for imaging and I’m only touching the surface with its performance after using it for 3 months
Attached are some of my images captured in Sydney and down at my holiday home South Coast NSW
All images captured with a Canon 600D unmodded and all under 3 hours of data
I’m only a beginner,still a lot of things to learn !!!
To me it’s a great all round scope
Enough said !

[Slawomir]

Please forgive me for giving a specific example, but when looking at what is currently being used by astrophotographers, for the maximum wide-fields low resolution purposes, I believe it’s hard to improve on 106 FSQ + KAF16803 combo.

Therefore at the lower end of focal lengths range, petzval seems to be the most optimal astrograph.

[Stefan Buda]

I don't dispute that. It may be true, but as I said in an earlier thread, the FSQ106 does not play ball when it comes to theory - Greg will be pleased to hear that - and I have no idea how to deal with it, even though I indicated that I know the reason. I'm convinced that that the FSQ106 produces larger spots than an equivalent FL APO with flattener. But it does it in a nice uniform way, across a large field.

[gregbradley]

I see the usual tighter stars in the centre and slightly weaker further out in their spot diagrams for the FSQ106EDXiv (the various EDX models have identical optics and its only various mechanical changes to overcome focuser sag in some earlier models).

https://www.takahashiamerica.com/takahashi-fsq-106edx4-quadruplet-refractor-telesco.html

Greg.

[Slawomir]

From what I have seen a quality APO will have tighter stars that Petzval (like FSQ) but since one of the requirements for an outstanding astrograph is a larger corrected circle and the reason to use a short focal length is to take wide fields low resolution images, I still think petzval is the optimal solution for a short focal length wide field astrograph.

Having said that, if we are willing to give up on some of that wide field, then a well-corrected APO might be a better astrograph in terms of data quality, but not the speed and wide-field.

I really look forward to seeing what Greg will think about his new 105mm hand-aspherised oil spaced APO. It is promised to cover a full-frame at native f/6.

[gregbradley]

FSQ is a hard to beat widefield instrument. It has a few downsides though.
One is it is very sensitive to focus shift due to temperature change. The other is the flat colour it tends to have (EDX versions). The black paint used in the tube has a green bias. The coatings on the lens I am not 100% sure I like them. They look mustard in colour and no other scope I have had that.

The focuser has been a weak point in the EDX series hopefully finally fixed with the iv model although I would not be surprised to hear of the odd one here and there that flexes still. Replacing the focuser with a Moonlite or Feathertouch would be the safe option with a temperature compensation electronic focus option.

F5 and 106mm is a good imaging platform. The flat field is great as is the sharp to the corner stars. Its not wind affected on a decent mount which was a great advantage for me at my dark site. If it were too windy it was FSQ time!

A lot of the all time best images are from the various FSQ models. The EDX models optics as far as I know have not changed only the mechanicals like the infamous Captains Wheel and slight changes to the focuser.

The focus lock mechanism does not work and the micro focus adjustment is only 1:6 which is not fine enough for a scope like that with a small critical focus zone.

So its not all roses but its definitely the most famous and popular widefield imaging scopes of all time for a reason.

I can see another one in my future at some point. A TV127i also looks to a nice substitute for the $26,000 FSQ130 which seems outrageous for a 130mm APO scope but hey maybe its amazing (I doubt that its $26,000 amazing though).


Greg.

Greg.