How important is mirror quality for deep sky imaging?

[SkyViking]

Hi All, I was wondering how accurate does a primary mirror need to be for deep sky photography?
I'm thinking that for long exposures the accuracy of the optical surface just need to be good enough to achieve a resolution equal to that limited by the average seeing? (Well actually more like 1/3 of seeing, to get good sampling)
This would typically still be less stringent than 'diffraction limited' I suppose.

The reason I ask is because my primary mirror really needs a resurfacing, and local options here in NZ are limited especially when I also want a good protective overcoating. The cost of packing and shipping the mirror overseas, plus recoating, would most likely be higher than a new standard 10" mirror (for example one of the Bintel ones). And I would be without a mirror for a while too (oh the horror... :lol:)

My current mirror is just a no-name one that I bought 17 years ago as part of a kitset, with sonotube, cheap focuser etc. I have no idea of how good it is, but doubt it is anything special. But it does produce images that are fine for my purpose. As you may know I have taken images of various high resolution features over the years so the mirror definitely delivers. I'm just thinking that a cheap standard mirror would be just as good, or maybe even better? But I really have nothing to compare with as I have never owned another mirror.:shrug:

So does anybody know if it really matters - is there perhaps a way to calculate the resolution expected from a mirror, given its properties such as strehl etc? In short, is a high end mirror really worth it for deep sky imaging?

Regards,
Rolf

[alocky]

In the days of film astrophotography that was certainly the wisdom: you don't need 1/8th wave figures to get good astrophotos. That will still apply where there is any sort of long exposure, but the advent of adaptive optics means that the optical performance will become more important. Planetary imaging with a high resolution video certaily shows up defects in the optics, the degradation due to slightly miscollimated optics certainly shows up, and I suspect so would a bad figure.
It all depends on the arc-second per pixel spec of your current imaging system, and your typical exposure length.
regards,
Andrew.

[SkyViking]

Hi Andrew, yes planetary imaging is a whole different story for sure. I'm specifically thinking of deep sky imaging with exposures on the order of minutes or more, where the seeing effectively limits the resolution.

My native resolution is 0.86"/pixel and with that I seem to get FWHM's in the range of 2"-3" depending on seeing conditions, which seems quite reasonable to me. So perhaps my question really is how accurate does a mirror have to be in order to still achieve the same FWHM with that image scale?
When I image Jupiter using high FPS I can easily record sub-arsecond details, so that's why I assume that the accuracy of my (probably very average) mirror is clearly higher than necessary for deep sky imaging.

[alistairsam]

Hi Rolf,
I had the same question as well and didn't really get an answer anywhere about practical differences.
I bought mine from Orion optics uk, and they have 4 grades from 1/4pv wavefront to 1/10pv. as for how much of a practical difference it makes with deep sky imaging, I'm very curious to know as well.
I asked them this question and they were more than confident that even their professional grade would yield far better results than standard GSO mirrors for DSO's. so maybe it does make a difference.
there is a good write up here http://www.orionoptics.co.uk/OPTICS/allaboutopticspa.html
I will be testing my 10" soon with the same camera that I used with the GSO mirror. will be interesting to see the difference. I'd be very disappointed if there isn't much.
But generally it makes a larger difference with planetary imaging.

would love to hear what others would have to say.

[gregbradley]

Diffraction limited at 1/4 wave. Beyond that and the atmosphere won't let you see the difference unless you are on top of a mountain (and no mountain in Oz either!). Planetary high mag though, not so sure.


Greg.

[Shiraz]

good question Rolf. Best quality optics is always worth having, but your question is - is top quality necessary when the atmosphere dominates?.

FWIW, my current thinking:
1. highest possible quality is required for visual use - we are very good at seeing anomalies and can pick up the effects of relatively minor aberrations very efficiently (particularly if we are unfortunate enough to know what to look for). Assume that is why people still pay big bucks for tiny but perfect refractors for visual use.
2. system requirements dominate for planetary - collecting lots of photons is a large part of the game and a big, but less-than-perfect optic may outperform a smaller perfect one, since it allows higher frame rates to help with seeing problems. Other issues such as thermal management and mirror cell design for equatorial mounting are also vitally important. I have only once found conditions good enough to test the limits of my GSO 12 inch optics over a full year of planetary imaging - but it would still be nice to have top quality optics though, just in case that mythical perfect night comes along.
3. optical quality is much less important for deep space imaging. The atmophere limits the resolution much more than the optical system and you can correct for many defects such as lower contrast etc in post processing - which explains why RC systems are popular, even though they start out with low Strehl simply because they have such large secondaries. The main optical problem for DSO imaging is astigmatism, since out-of-round stars will remain somewhat out-of-round even after convolution with the atmospheric psf - but astigmatism is generally caused by mirror mounting problems rather than the mirror, so you should be able to correct this type of fault. Minor diffraction can also influence saturated star shapes, since you see the skirts of the saturated PSF and any minor sprays of light from intrusions into the light column will distort the star shapes, even though unsaturated stars look OK.

My guess is that you will find that a GSO mirror is great for DSO imaging. Interested to hear what others think. Regards ray.

[RobF]

Finding this a fascinating discussion. I suppose theoretically you would need an occasional exceptional night of seeing/clarity to see the difference with a mirror better than 1/4 wave from what people are saying for DSO imaging.

I'd always wondered if my chinese mirror was a "goodie", but probably in reality it's 90% seeing limited as people say (and Brissie is never going to put Mauna Kea out of business) :)

[Moon]

Good question, but I don't know if there is an answer since it depends on your personal value judgement.
I use a GSO mirror and my images look ok to me, but I notice people with more expensive mirrors often times take much better images. Of course there are additional factors involved, including the skill of the operator!
I would say have a good look on the internets at all the images taken with GSO mirrors and set your expectation level accordingly.
If you want to go down the premium path, you could consider a Zambuto swap over. This is something I might consider one day too.
Just my 2 cents worth.
James

[gregbradley]

One way to answer this is what is the mirror accuracy of these GSO RCs?

I doubt its even 1/4 wave but it must be fairly close. There plenty of sensational images using them.

I imagine (I don't know for a fact) that a computer controlled polishing machine can most likely take a mirror close to 1/4 wave without too much trouble but to get it further requires more expensive, more precise machines and probably a fair bit of human labour and work to get that last little errors smoothed out.


Greg.

[Shiraz]

Hi again Rolf.

Looks like I was wrong in an earlier post to suggest that optical quality was not such a major issue for DSO imaging. I did a quick and dirty analysis using Aberrator to generate PSFs for a variety of scopes and then convolved these with a 2 arc sec Gaussian (in IRIS) that represents good seeing averaged over a fairly long integration time. The top row of the diagram represents the resulting combined PSFs (star shapes) for scopes from 80 to 300 mm, with brightness normalised and assuming perfect optics and with identical angular scale. The chosen scopes are a couple of refractors, three Newts and an RC. For scale, the 300mm Newt produces pretty close to a 2 arc sec combined PSF (half power).
The two PSF images below the 250 f5 entry are a 250 f5 with a mirror that only meets lambda/4 and the lower one includes some pinching and tube currrents as well.

the whole pattern is repeated lower down, but with the synthetic stars driven into saturation by a factor of 2x.

So:
1. for 2 arc sec seeing, you get no resolution increase above about 250mm aperture, but even the 80mm refractor does a pretty good job.
2. moderate spherical aberration results in larger stars from the 250mm Newt - the half power area is ~20% larger at lambda/4 than with perfect optics. Looks like optics with low SA are definitely worth having in this application as well, if you are imaging in above average seeing conditions.
3. tube currents and improper mirror mounting can really ruin the image

These results are a bit unexpected and I am now going to re-read the relevant parts of Suiter's book to make sure I have it somewhere near right with Aberrator - maybe someone else can try something similar as a check.
Regards ray

[gregbradley]

Higher end mirrored scopes often quote spot sizes in microns for centre and for a certain distance off centre.

That is another way to evaluate star sizes you will get.

Greg.

[SkyViking]

Thanks to everyone for your valued input on this. I've decided to get a GSO mirror - they sell on Astronz for NZ$375 for a 10" f/5, I just can't really go wrong with that price :D and then I won't be without a mirror either (Imagine if Eta Carinae went bang while I was waiting for my mirror to be recoated... :eyepop::lol:)

I've also just installed a new 70mm secondary from Bintel so that's taken care of too. I believe the Bintel mirrors are GSO as well?
It will be great to once again have nice shiny optics throughout, should have done this years ago... :rolleyes:

I'll let you know how it goes with the new mirror once I get my hands on it.

[strongmanmike]

Sounds like a plan Rolf :thumbsup:

[bmitchell82]

You know for most of the time when your seeing is pretty average then I think the wisdom that has been told is spot on, but even here in Perth I can see where my mirror lacks in clarity.

On crystal clear, still winter nights my mirror isn't far off atmospherics. 0.89 arc sec per pixel.

My train of thought is the errors of the PV across the mirror don't "Hide" under the seeing and to some effect are additive, think of it this way if your Mirror was 100% flat, no peaks or vallys (extreme case) then the atmospherics would 100% be the cause of all errors, but if your adding in the light that comes though the aperature is already distorted then it hits the side of peak/vally it will not come back dead straight so its added to the error.

Why would you buy a RCOS telescope for its mirrors why would you get ION milled RCOS mirrors? when GSO sell the same?

Even looking at Mikes ETA and mine side by side, the Qhy9 has 5.4micron pixels Vs 9 micron pixels. The Arc second per pixel isn't too different. but there is a noticable change in resolution. I put it down to higher quality optics.

[clive milne]

A couple of thoughts:

Missing from the discussion so far is the limitations imposed by the (CCD) detector.

To extract all the information at the focal plane you will need to oversample the image by at least a factor of 2x
(basic nyquist sampling theorem)

However, the following caution also needs to be included:
Do not assume that a CCDs sampling rate is defined by its pixel size.
A real world example of what I am alluding to can be found at the bottom of page 11 on the following PDF:
http://www.sta-inc.net/wp-content/uploads/2010/08/Overwhelmingly-Large-CCDs-for-Astronomy.pdf
In this example, when a 2 micron star image was focussed on to an 9 micron array, 90% of the image was
registered by the target pixel, the surrounding pixels acquired the bulk of the remainder, rendering
it as a 27 x 27 micron cross. (13.5x the size of the original image) This is not due to the central pixel's well
depth being exceeded.

Implicitly therefore, for mirror quality to become a significant limitation in deep sky imaging, you need to
be over-sampling to a significant extent. The magnifications used by planetary imagers might serve as a data point here.

The bottom line is that if you put an ST10 at the prime focus of an 8" F5 GSO Newtonian equipped with a coma corrector,
the resulting spatial resolution in the image will be indistinguishable from what you would get with a 5" f8 Astrophysics
refractor, all else being equal (not withstanding the diffraction spikes & better image depth with the GSO).

best
~c

[clive milne]

A hypothetical question to perhaps lend some context....

What is the first thing that comes in to your mind when you hear the words; Riccardi Honders...
I'll bet London to a brick that it would be something along the lines of; 'ultimate imaging tool... want.. might consider trading internal organ to acquire'

Now consider the effect of its 50% central obstruction and what level of spherical aberration
constitutes an equivalent offence to the PSF. fwiw) It is waaaaaaay more than 1/4wl of SA.

ergo.. for prime focus deep sky imaging, the only purpose that 1/10th wave optics serve is to gratify your ego.

[alistairsam]

Clive,
How about two 8" reflectors, one with a 1/4 PV wavefront spec, and other a 1/10 PV wavefront spec with the same secondary?
if mirror quality doesn't lend discernible differences for DSO's, does ccd quality be it build or sensor quality make more of a difference?
eg,similar spec'd qhy, sbig, FLI or other high end cameras? or is it sensor size similar to aperture with telescopes? bigger sensor, more detail?
I'm referring to a raw capture without any processing?

[Shiraz]

Hi Rolf, will be very interested to find out how the new mirror performs.

Clive, thanks very much for the info - never seen specs for lateral charge diffusion (assume that is what it is) published for any astro sensors, except maybe Kodak's claim to produce "low smear" chips.
This could possibly explain why planetary imagers find that going well beyond Nyquist sampling requirements yields additional detail - maybe we are using oversampling to deal with sensor diffusion blur. Heavy handed oversampling may also be worth pursuing for DSO resolution improvement as well. Regards ray

[clive milne]

Hi Alistair,
To my understanding, the only real difference between the two OTA's you describe above is that the higher spec example will (in theory) be very, very slightly more tolerant of seeing conditions. I doubt however that you will ever be able to notice.... I would guess that it would be buried in the noise.

As far as variation between different cameras.. I have no feel for that, but intuitively, being that the manufacturer I quoted above is at the cutting edge of developing CCD's for professional astronomical research applications, I would be prepared to go out on a limb and suggest that their performance is close to 'best in class'.

~c

[clive milne]

Hi Ray...

There may be two factors at play here... electrical charge diffusion and/or light spill through the side walls of the pixels. Without some manufacturer specific 'secret sauce' the signal spill may be inherent to all CCD's made from any given substrate... I don't pretend to know the answer to that.

Incidentally, if anyone wants to simulate the effect of adding the equivalent of 1/4 wl of spherical aberration, just add a 35% central obstruction to your OTA.

Interesting that RC's have a 35% central obstruction basically by definition, and they seem to work fine more often than not. The fact that they routinely deliver higher resolution images than Newtonians of similar aperture might be just a function of the over sampling achieved by the longer focal length.

Also, (i'll re-iterate the point) the fact that optical tube assemblies with 50% central obstructions are practicable as prime focus telescopes should serve as a data point for how much spherical aberration can be tolerated in this application... at a guess, this is equivalent to 1/2 wave, give or take.


~c

[alistairsam]

Hi Clive,

So do you think it unlikely to yield any difference with details in areas with nebulosity, faint shadows and highlights or overall clarity?
you often read reports of visual observers noticing improved contrast, pinpoint stars etc with higher quality mirrors, I can't help but think it would be similar for imaging especially since the eye is far less sensitive to subtle variations in detail and if they're detectable by the eye, it might be more pronounced with a ccd? just guessing here.

I'm not sure of the details, but wasn't the initial flaw with the hubble mirror relatively minor but yielded dramatically inferior results? I guess the scale of that mirror is very different from what we're talking about here and could've been the contributing factor?

[Peter Ward]

Beg to differ with Clive's analysis. :)

With imaging it comes down to field correction, spot size and seeing. A well made Riccardi-Honders delivers 5 micron stars across around a 70mm flat field. Most other systems have spot sizes 2x larger on axis, and
degrade to *way more* off axis due to field curvature, astigmatism etc.

Other factors such as thermal and mechanical stability also come into play, and literally shift focus as an exposure is being taken. Exacting focusers and thermally stable materials usually come at a cost. My experience so far has been, you get what you pay for.

[Satchmo]

Peter, this kind of performance may be possible on a ray trace program but the reality of diffraction theory says its impossible. For a perfect F4 system of any design the diameter of the Airy disc to the first mimima of the airy pattern will be about 6 micron and then the 50% central obstruction will throw light into the first and second diffraction rings making them very bright , probably doubling the best blur spot diameter to twice that in the presence of seeing.

[Peter Ward]

Not really. Airy disk size and spot size are two different paramaters. I think you knew that :)

The physical size of the airy disk is purely a function of the F-ratio. 4.7 microns for a F3.8 system. With 305mm of aperture you'd be looking at 0.84 arc sec of sky.

That is not the same as the spot size.

Spot size (and shape) depends on all sorts of variables, but in a nutshell, mirror/lens quality & optical design.

You only need to look at the off axis "sea-gulls" delivered by most camera lenses to see what happens when a design is not corrected well.

The Honders design covers a 3 degree field with essentially perfect color correction from 400 to 1,000 nanometers (UV to IR). The telescope is fully corrected for spherical, coma, astigmatism, field curvature and distortion, longitudinal and lateral chromatic aberration.

[Satchmo]

Peter, Yes I know the difference, I was referring to the quoting of what I imagined were spot sizes from geometric ray tracing ( just talking about on axis to simplify things) . Assuming you had pixels to oversample , the true spot size ( allowing for diffraction effects) achieved in an image would be somewhat larger than 5 micron due to the extra energy thrown out of the airy disc by the 50% central obstruction.

See two simulations I have done with Aberrator. When you add in the blurring effects of seeing the true blur spot diameter of the real instrument is going to be somewhat larger than the one calculated by geometric ray tracing. If the first dark minima is at 4.8 microns, the blur spot will be closer to 10 micron in a time exposure.

Unless you are using very small pixels the obstruction probably makes no difference to the quality of the images.

[Peter Ward]

That's the rub. Agreed, On axis, most 'scopes can perform very well.

35mm off axis is where many systems simply don't cope....as sensor sizes have grown, the merits, or shortcomings of various designs have become more obvious.

[gregbradley]

I asked RCOS once what the expected gain was from the Ion Milled optics which are to all practical purposes perfect. I was told the ion milling removed micro scratches from the polishing process and the gain was less light scatter.

Much like the gain from fluorite over FPL53. Minor but noticeable.
Less light scatter. This could mean slight improvements in contrast.

Greg.

[Satchmo]

My guess is that sky background darkness is a function baffling and F# alone and that smooth optics will mean that precious energy isn't scattered from diffuse objects . However I'm not sure what the relationship between spherical aberration and contrast of diffuse objects. I'd say it is much less important a consideration than optical smoothness for imaging faint diffuse objects. Users who report seeing `darker' sky background in premium reflecting optics I think are having themselves on. Perhaps the coatings are cleaner.

[bmitchell82]

Okay now im lost!

On one hand people are saying that commercially made optics are just fine as all this PV measurements are just a load of horse whooharr

On the other hand people are saying that the higher quality the mirror the better. Better images, Better magnification etc etc

If high quality mirrors don't mean anything then why would you have "Premium" mirror makers. Isn't it all just a crock?

I might as well just get the bog stock standard run of the mill mass produced optics at 1/4 of the cost!

Could somebody put it in terms that are quantifiable and not in high end optician language?

[Peter Ward]

Shades of gray.

The last 13% of any engineered system seeking "perfection" starts to cost serious $ with not a whole lot of performance return.

Hi-Fi's, cars, etc. all fall in that sort of category.

On a good night (i.e great seeing) average optics will satisfy most punters.

Excellent optics will return good images most nights (in short , it takes crappy seeing to unsettle them)

But, on a really good night (i.e superb seeing) excellent optics will also give you images as good as mother nature allows....

....A place you can never get to if the imaging system is already hamstrung by rough, poorly corrected or short-cut designed optics.

Question is: do you want to pay to capture those rare, albeit sublime, moments??

Proof?? the M104 image I took some years ago with great optics in superb seeing. The link is here (http://www.atscope.com.au/BRO/gallery20.html)

[bmitchell82]

That is what I thought all along Peter!

No i'm not just a happy snapper... :P Though I am majorly ham strung by uni! :P

I remember watching Top gear one night and they where testing the bugatti Veyron, the new one or the "suped" up version had close to 40hp at the rear wheels but only went a extra 4 or 5km a hour some piddly amount. At 300+ air becomes thick soup! Wind is my major source of loading where P=0.5*1.2kg/m3*wind speed^2*10^-3 = kN/m^2 :) so yes I understand that at low end it ramps up quickly but at the high end alot more effort is required.

I think presently I have almost reached the quality limits of a mass produced mirror. Its apparent when sitting side by side with other astrophotographers with similar styles of equipment but differing degrees of quality!

[RobF]

That's why you shouldn't be systematically ironing out the deficiencies in your mass-produced and self-tweaked system Brendan. The lure of additional quality images is about to COST big time ;) :lol:

[Bassnut]

Well put Peter, the proof is in the pudding, and that m104 is certainly proof, the detail in the edge is outstanding. I hope to also prove the point with my overpriced RCOS optics (albeit pissy 10") when I install my rig at siding spring shortly right amongst the big toys, that'll make for some interesting comparisements :P :lol:

[issdaol]

So along the lines of Hi-Fi audio perfectionists being called Audiophiles ,maybe there should be a term for Astronomers - Astrophiles :-)

[alocky]

I'm sorry I'm not familiar enough with your equipment list, Peter. What was the FL, aperture, f ratio and design of the system that produced that truly impressive image? How does the resolution in that image compare with the theroretical capabilities of the system? What was the sensor/camera used?
regards,
Andrew.

[Peter Ward]

Err...it's all there on the Web page. Just scroll down

What I don't know is: how the actual FWHM's compare to theory. The optical systems figure is however within 4% of perfection.

[Peter Ward]

Yep .....but you neglected to mention the fact: you got them for a song! (Fred actually only threatened to sing...so I gave him the scope to avoid any more Tinnitus than I already have :lol: )

[bmitchell82]

:) how you know it rob how you know it!

[alocky]

Thanks - I got a bit caught up in the image. Still - it's only really proof that you got a great image through your system: now someone with a similar aperture but inferior optics needs to step up to that mark.
regards,
Andrew.

[DavidU]

Here is one of Rohr's optical tests. This is a superb mirror in all respects , now check out the resolution it is capable of in the 3.3u artificial stars test.In there is also the star resolution test of a GSO mirror and a few others. It clearly shows fine resolution is better with an extremely good mirror.
http://translate.googleusercontent.com/translate_c?act=url&hl=en&ie=UTF8&prev=_t&rurl=translate.google.com.au&sl=de&tl=en&u=http://www.astro-foren.de/showthread.php%3F7977-traumhafter-Newtonspiegel%26p%3D30967&usg=ALkJrhigX-UHr3OCwL8MTKJAlwnVT-M0Kw#post30967

[Satchmo]

The issue of whether optical quality is justified for deep sky imaging is intimately bound up by the choice of camera and pixel size vs airy disc size. If you are using your eyeball, high quality optics will always win out as your eyes are capable of seeing the moments of clarity in even average seeing Planetary imagers who try to over sample the airy disc significantly with pixels will be fairly sensitive to mirror quality.

My attitude is that the lines of optical quality are probably more 'blurry' for prime focus imaging at faster f #'s but it helps to have good optics, but good optics will not help if you can't focus or guide properly.

[alocky]

I'm pretty sure the advantage of good optics isn't under question, my understanding of the original post was whether you could actually translate that superior resolution into a better image in real world conditions. Peter Ward has demonstrated what is possible with good optics, and I suspect more importantly, good seeing. It would be nice to know how close his image approaches the theoretical perfomance of the system, or would a c14 have done just as well that night? Would a 'cheapy' 20" have done even better (notwithstanding the challenges of mounting larger instruments)?
cheers,
Andrew.

[bmitchell82]

:) I have no issues of guiding for how ever long I want and regularly see less than .5 px deviation at 5.4 micron pixel size. Focus well... here (http://brendanmitchell.net/wp-content/uploads/2011/11/B33.jpg) you can see i know how to focus too. Hence why the original post had me very interested as soon I will be upgrading my old Synta 10"



Mid this year I am going to the Hallowed lands of unbelievable seeing! :) If i knew just how to make this measurement I would tell you.
Mark do you know how to preform this test?

[clive milne]

Peter,
I think are getting the wrong end of the stick here.
I used the Riccardi Honders as an example because it is without question as good as it gets for deep sky imaging, but let's be realistic about the factors that make it so. Even a perfectly constructed OTA (of any configuration) that has a 50% central obstruction is in reality no better than 1/2 wave once you include the effects of diffraction.

Ergo, optical quality really isn't the weakest link in most prime focus deep sky imaging chains. This is what the OP was asking. Field aberrations, mechanical construction, mount stability, seeing conditions, tracking error, operator skill, etc) are different issues and are really what separate the sheep from the goats.

fwiw) Roland does not suggest that a Riccardi-Honders delivers 5 micron images (which it cannot) What he says is that a 12" f3.8 produces an airy disk with a diameter of 5 microns, and that his OTA is diffraction limited over a 60mm field. It is left to the educated reader to fill in a couple of blanks: ie)
1) Only 50% of the encircled energy will actually fall within the airy disk, most of the rest being pushed into the first diffraction ring with a diameter of 10 - 20 microns depending on the wavelength... there is your limit (diffraction)
And 2) the point I was trying to establish.... That spec is entirely good enough for prime focus deep sky imaging at the highest level. The results speak for themselves.

best,
~c

[clive milne]

That is truly an excellent image Peter... I think the point that you have just established is that it is possible to take a world class image with an OTA that has the equivalent of 1/4 wave spherical aberration.

(An RC with a 32% central obstruction has exactly that)

114179

[clive milne]

My feeling is that good quality optics are worth the extra cost.
If you are using the scope as a visual instrument, then it's an absolute no brainer.
For imaging the merit function is a little different, but let me put it to you like this.... Just as an example, if you wish to get something around the 12" mark using a 16803 chip, the mount is going to cost you $10-20K, the camera a similar amount (you are up around the $20-40K mark already) The cost differential between OTA's is such that it isn't that much extra to go for something decent in the overall scheme of things.

The mechanical stability of the higher end products is probably worth more than anything else.

But you know... let's dispense with the idea that stratospheric strehl ratios quoted for instruments with large
central obstructions is anything other than an exercise in marketing hype.

[Moon]

As I said in my first post, this is a personal value judgement we each have to make. I guess that why this such an interesting discussion to follow.
In my case I put the mount & camera ahead of the optics. Hence my current combination is PME + STL + GSO 300mm f/4. Feel free to have a look at a fairly 'honest' example (http://deepspaceplace.com/show.php?id=ic5358) or here for the full list (http://deepspaceplace.com/images.php?sort=&filter=AT12IN) of this combination.
In my assessment the things holding me back from better images would be (in rough order) the weather, light pollution, seeing, my skills, focus and collimation. Better optics would be nice, but it's down the list. For now I'll just pretend its as good as the test report (http://translate.googleusercontent.com/translate_c?act=url&hl=en&ie=UTF8&prev=_t&rurl=translate.google.com.au&sl=de&tl=en&u=http://www.astro-foren.de/showthread.php%3F7977-traumhafter-Newtonspiegel%26p%3D30967&usg=ALkJrhigX-UHr3OCwL8MTKJAlwnVT-M0Kw#post30967) posted by Dave.:)
James

[clive milne]

Nice work James... Looks like you have got that 12" working well.

Your report on the OTA mirrors my experience with GSO. The optics in the smaller sizes (12" and below) are actually pretty reasonable, and are certainly good enough for imaging. It amazes me that they haven't lifted their game with the mechanical components though. I also wish that Andrew's would spare us the marketing B.S. I don't imagine that too many people would take it seriously.

btw) Absolutely nothing wrong with this image:
http://deepspaceplace.com/images/ic4628.jpg

Well done.

[Satchmo]

I wonder who made that mirror. Its certainly not a GSO being 2007.

[Peter Ward]

Well not quite, just because the airy disk energy is, well, not in the disk and has moved to the first ( or more ) diffraction ring doesn't mean the spot sizes will be bloated.

A modulation transfer function only tells part of the story....as does a ray trace which I think has more relevance to deep sky imaging.....

Differences in deep sky images from telescope with a figure error (eg 1/4 spherical) and one with none is obvious....with the latter putting all the energy into a small spot ( albeit larger than the "perfect" airy disk) rather than smearing it over tens of microns due the figure error.

One need look no further than Hubble images pre and post servicing missions to better understand my point....the secondary obstruction remained the same
but the figure and images were sooo much better :thumbsup:

[DavidU]

It is a hand made European mirror, possibly Italian or French.

[bratislav]

James,

no need to pretend or guess. That is why Roddier invented his test. Capturing extrafocal images of a star is rather trivial nowadays, and I'm quite good at running the test.
Next best thing to a proper interferogram, that's for sure. No guessing involved one bit.

(attached analysis of Stefan's 16" D-K in average seeing)

[Moon]

Thanks for reminding me about that Bratislav. I actually downloaded the software & joined the yahoo group (http://tech.groups.yahoo.com/group/roddier/) but never got round to it.
I must do it next time I have the camera off.
James

.

[Peter Ward]

I understand the modulation transfer function to only be a measure of frequency preservation....hence and not sure what you mean by saying a 50% obstructed telescope is like having a telescope with 1/2 wave of spherical error... hence.... I still don't agree with your analysis :)

That much spherical error looks bad visually and easily shows up in images. (I've had some clunkers in my time )

You are indeed correct about Roland's spec claims... but I did some extra research prior to getting the scope, and the Honders has spot diagrams well within the diffraction limit across a 60mm field (attached)

I'd be interested to see a ray trace of a 1/2 wave spherical error....but so far I've not manage to track any down.

[SkyViking]

Many valid points have been raised, and surely a Riccardi-Honders is great in practice with its large fully corrected field, but isn't that design mostly of benefit in wide field applications?

I found some numbers to shed a bit more light on my original question:
Median FWHM of Paranal site: 0.75"
Median FWHM of Kitt Peak site: 1.10"
Airy disk size for 8000mm (VLT) aperture: 0.03"
Airy disk size for 250mm aperture: 1.03" = 6.2microns @ f/5

So the conclusion from that must be that atmospheric seeing is by far the limiting factor for a site like VLT, effectively reducing the theoretical resolution by a factor of 25.
For my location, which isn't on a remote mountain top with premium seeing conditions, I'd expect the average achievable FWHM to then be around maybe 2.00". I'm just guessing here, but it must surely be worse than the professional sites. So that means in my case the atmospheric seeing reduces my theoretical resolution by a factor of ~2. The question then is wether a GSO 250mm mirror is good enough to deliver star sizes no larger than the average seeing of 2.00"? and based on my experience I think it would be good enough, since with my old no-name mirror I can see much greater detail that this visually when looking at Jupiter for example.

So simply in the context of 'which f/5 mirror to use for long exposure deep sky inaging', I think a GSO would more than suffice based on what I can gather here.

In any case this is a very interesting discussion overall, keep it coming! :)

[Satchmo]

Peter, I think a 50% obstruction produces modification of the Airy pattern similar to a Strehl ratio equivalent to 0.5 wave or so spherical aberration in a perfect system. When you get down near the diffraction limit geometric ray trace diagrams are of intellectual interest only because of course the real image is subject to diffraction effects and the Strehl ratio tells us a lot more . I'll look into it for you.

Off the top of my head the geometric ray spot bundle at best focus for a 1/8 wave of pure spherical aberration optic ( wavefront ) is about half the theoretical airy disc diameter , and a 1/4 wave wavefront system has all it rays within 3 Airy Disc diameters.

Casual onlookers must keep in mind that in general discussions about wavefont quality vs utility for visual or imaging we are always talking about simple pure spherical aberration at those tolerances ( to keep things simple ) , and not other kinds of errors, which may or may not have similar effects on the Strehl ratio. The famous 1/4 wave `Rayleigh Criteria' which many mistakenly call `diffraction limited' referred only to pure spherical aberration, and even then was really a generally tolerable lower limit. Lord Rayleigh added that the effect on visibility of planetary detail was `already decidedly prejudicial ' at 1/4 wave.

I personally define the meaning of Diffraction Limited to mean that the only disturbance to the Airy Pattern attributable to the optics being the diffraction effects itself, and that won't happen until spherical aberration is reduced to 1/8 to 1/10 wave in the system . This caveat was first expressed by Danjon and Couder in the 1930's. Unfortunately few optics display SA in its pure form , but have generally more complex errors. In all the commercial optics I've looked at I find figure of revolution problems to be the greatest issue, and pure astigmatism itself is rare in favour of more complex irregular shapes.

[bratislav]

There is no need to simulate 1/2 wave of spherical, as all modern raytrace programs will be able to easily show the results of 50% of obstruction.

I don't have access to Riccardi's variant used in AP, and can't be bothered to modify my existing designs based on Honders' original ideas, but have another similarly capable design handy (Companar at 12" f/3.3, 50% obstruction). See attached spot diagrams for spot diagrams from zero to 2 degrees off axis (full 70mm circle) and corresponding diffraction affected PSF plots on axis and at 2 degrees off axis (apologies for low quality gifs, don't have access to all my tools at the moment). For PSF diagrams, square dimension is ~18 microns; those diffraction rings are bright enough to swell the calculated 4 micron RMS into 15 micron blur for anything but faintest stars.

I've exchanged many emails with Claas Honders in late 90's and finally managed to convince him to get design into the open, but unfortunately his untimely death put stop to that. Busack now claims to be the earliest inventor of such designs, but without Claas to tell us when did HE come with such ideas, we will never know.

[Peter Ward]

Guys...just so we are clear about the point I'm making....

An otherwise perfect optic with a 50% central obstruction does not degrade a telescope's resulting images in the same manner as a 1/2 wave spherical error....or turned edge or...or astigmatism etc. etc.... i.e the sort of errors you may find in an optic built to a price, rather than spec.

(download Aberrator...it's free...& play with the values )

One is a frequency domain filter (eg central obstruction), the other is a focus error.....which really will bugger up images, al la Hubble before it was fixed.

[SkyViking]

Hi Peter, the central obstruction is indeed a frequency filter but it does seem to affect a telescope's ability to resolve contrast.

There is a fairly comprehensive analysis here: http://www.telescope-optics.net/telescope_central_obstruction.htm

There is also Thierrys Legault's page here: http://legault.perso.sfr.fr/obstruction.html