Optical Quality for Imaging

[thehand]

I've seen in mentioned, and I think there was a thread here or over on Cloudy Nights, that discussed a study that was done to see if people could identify PV/4 optics vs. PV/10.  The conclusion was that it was very difficult.

That said, forget the human eye and what about a camera?  If I were to use a 16MegaPixel camera, or therebouts, with an 8" Newtonian would I be able to see the difference between PV/4 and PV/10 optics?  I would expect a maybe for full image usage but what if I want to zero in on 10% of the image?  Thanks.

[Demonperformer]

Interesting question to which, I'm afraid, I don't know the answer.  But hopefully someone will.

[jambouk]

No idea, but this thread may help you on your quest:

http://stargazerslounge.com/topic/70608-orion-optics-110-wave-thoughts/

James

[thehand]

Thanks for the link, unfortunately the discussion is focused on subjective evaluation vs. quantitative/qualitative like you would have with an imaging system.

[Stu]

I don't know about the imaging piece, but I'm sure the differences show up most at high power on planetary and lunar observing/imaging. Seeing, cooling and collimation obviously plays a big part too

[Big Jim Slade]

Thanks for the link, unfortunately the discussion is focused on subjective evaluation vs. quantitative/qualitative like you would have with an imaging system.

Do we have a definitive set of tests that we can call quantitative/qualitative for an imaging system? ADU for magnitude? FHWM for sharpness? Given that when two or three astro imagers are gathered together they shall disagree about something, is it even possible?

I'm the kind of guy who looks at Planewave spot diagrams and thinks: "I bet that would give me tiny stars..." so  I understand the motive, but is it realistic?

[thehand]

I'm new to this field but it is surprising how little quantitative and qualitative information there is for what amounts to precision instruments that command hundreds to thousands of dollars of investments.  What is even more amazing is how accepting people are of a largely subjective qualification criteria.  Just speaking of Newtonians:

For thermal, where are the 3D thermal analysis plots and thermal time constant information?  Scope vendors seem to be aware of thermal if they are investing in baffles and fans but nary a lick about how this improves the performance.  It costs about $8k - $10k and takes a couple of weeks for a contract Thermal Analyst to run the analysis and do the calculations.  Surely these companies are investing in analysis if they are going to throw baffles and fans at the problem.  Where is the data?  It's an opportunity for the high end manufacturers to really shine.

For focusers, where is the performance data, tolerances, etc...?   Not hard stuff but potential discriminators

For optics I have only found one company willing to sign up for optical quality and that's Orion Optics UK (no relation to Orion USA).  Not only that but they send you the test report which includes measurements of PV absolute, PV RMS, Strehl Ratio, and one other I can't recall as well as a topographical map of the mirrors showing where deviations in the surface are.  I've contacted other manufacturers to ask why they can't do the same.  In one case one manufacturer said they did the same test but did not include the reports (IE they push the envelope for sub standard mirrors and will replace if you complain).  Another I talked to tried to convince me the country of origin made it a good scope and that it's optical characteristics were top notch but they are also not in the habit of giving reports.  Frankly I don't care where it was made or by whom.  That has nothing to do with whether it's a good scope or not.

There are other aspects of the OTA that could probably be addressed as well such as ease of collimation and the susceptibility to loosing collimation on transport (a day on a contract shaker table in typical carry cases to MIL-STD-810 or USDOT transport vibration spectra, it's not hard or expensive).  How many people complain about some scopes being more difficult to collimate than others?

Seeing conditions are what they are but to learn you are limited by a sloppy telescope manufacturer on one of those few perfect days is heart breaking.

[Demonperformer]

Just some thoughts I have had, but I may be on the wrong track, so caveat emptor ...

Ignore the atmosphere for a moment (which for two scopes at the same place/time could be regarded as a constant)

One way of looking at PV/n is that it is a measure of the maximum imperfection of the optics and these imperfections cause a point source to appear dispersed. So it should be possible to produce a figure for the spread of a perfect point source for a given set of optics at PV/n [i suspect it would be different for different sets of optics (aperture, f/length, etc)]. Let's say PV/4 optics gives an image diameter of A um and PV/10 gives an image diameter of B um.

We would now have to consider the pixel size of the camera. And I am guessing that the criteria here is how probable is it that the entire star will fall on one pixel? (which would produce a point image for a point source). If we know A & B it should be possible to calculate these two probabilities.

So would the ratio of these two probabilities give you a measure of how much better the PV/10 is for imaging than the PV/4?

I don't have the math to take this much further (like putting in values or producing specific equations), but that would be my starting point.

[michael.h.f.wilkinson]

I'm new to this field but it is surprising how little quantitative and qualitative information there is for what amounts to precision instruments that command hundreds to thousands of dollars of investments.  What is even more amazing is how accepting people are of a largely subjective qualification criteria.  Just speaking of Newtonians:

For thermal, where are the 3D thermal analysis plots and thermal time constant information?  Scope vendors seem to be aware of thermal if they are investing in baffles and fans but nary a lick about how this improves the performance.  It costs about $8k - $10k and takes a couple of weeks for a contract Thermal Analyst to run the analysis and do the calculations.  Surely these companies are investing in analysis if they are going to throw baffles and fans at the problem.  Where is the data?  It's an opportunity for the high end manufacturers to really shine.

For focusers, where is the performance data, tolerances, etc...?   Not hard stuff but potential discriminators

For optics I have only found one company willing to sign up for optical quality and that's Orion Optics UK (no relation to Orion USA).  Not only that but they send you the test report which includes measurements of PV absolute, PV RMS, Strehl Ratio, and one other I can't recall as well as a topographical map of the mirrors showing where deviations in the surface are.  I've contacted other manufacturers to ask why they can't do the same.  In one case one manufacturer said they did the same test but did not include the reports (IE they push the envelope for sub standard mirrors and will replace if you complain).  Another I talked to tried to convince me the country of origin made it a good scope and that it's optical characteristics were top notch but they are also not in the habit of giving reports.  Frankly I don't care where it was made or by whom.  That has nothing to do with whether it's a good scope or not.

There are other aspects of the OTA that could probably be addressed as well such as ease of collimation and the susceptibility to loosing collimation on transport (a day on a contract shaker table in typical carry cases to MIL-STD-810 or USDOT transport vibration spectra, it's not hard or expensive).  How many people complain about some scopes being more difficult to collimate than others?

Seeing conditions are what they are but to learn you are limited by a sloppy telescope manufacturer on one of those few perfect days is heart breaking.

I do not fully agree with the above. There are quite a few scopes that give Strehl factor of the optics, which is a good quantitative indication of the performance of the optics. How accurate these numbers are is a moot point. The baffles you speak of are more for suppression of stray light than thermal. I have done thermal stability analysis of an optical instrument for the ESO 3.6m telescope as part of my MSc, and a single figure is not going to tell you very much, The effective cooling time of a scope depends on many factors, including the pose, wind condition (forced convection is the major cooling factor), etc. This means that even if you did the analysis, a signle figure would not just be meaningless, but probably even misleading.

Regarding collimation: there are many telescope reviews that will remark on ease of collimation, and several manufacturers advertise the ability of some Dobsons to have the primary cell collimated from the front of the scope.

[michael.h.f.wilkinson]

Just some thoughts I have had, but I may be on the wrong track, so caveat emptor ...

Ignore the atmosphere for a moment (which for two scopes at the same place/time could be regarded as a constant)

One way of looking at PV/n is that it is a measure of the maximum imperfection of the optics and these imperfections cause a point source to appear dispersed. So it should be possible to produce a figure for the spread of a perfect point source for a given set of optics at PV/n [i suspect it would be different for different sets of optics (aperture, f/length, etc)]. Let's say PV/4 optics gives an image diameter of A um and PV/10 gives an image diameter of B um.

We would now have to consider the pixel size of the camera. And I am guessing that the criteria here is how probable is it that the entire star will fall on one pixel? (which would produce a point image for a point source). If we know A & B it should be possible to calculate these two probabilities.

So would the ratio of these two probabilities give you a measure of how much better the PV/10 is for imaging than the PV/4?

I don't have the math to take this much further (like putting in values or producing specific equations), but that would be my starting point.

You are right that increasing the errors in the optics blurs the image. If the errors are more-or-less randomly distributed, this would widen the diffraction pattern (Airy disk) in two ways: (i) widening the central peak, and (ii) increasing the wings (fringes) of the diffraction pattern. The two factors need not both occur: central obstruction tends to make the central peak a bit narrower, whilst widening the fringes.

Factoring seeing into the equation: The final appearance of a point source (point-spread function or PSF) is the diffraction pattern of the optical system convolved (blurred) with the seeing disk. Modelling the centre peaks of these disks by Gaussian distributions (bell-shaped curves), you wil find that the centre peak has a width which is the square root of the sum of the squared widths. If the seeing disk is 1" and the diffraction pattern 0.5", the PSF has a peak width of √(1²+0.5²) = 1.118". If my optics yield a peak of 1" the figure  becomes √2 = 1.412" . This is a rough approximation, of course.

Much depends on the kind of imaging you are doing. DSO imaging rarely samples the image anywhere near the Nyquist limit (unlike solar/lunar/planetary). Sampling at the Nyquist frequency means matching the PSF size to the pixel size optimally. A 5.85 micron pixel size requires a focal ratio somewhere near F/25 to achieve that. No DSO imager would consider that practical. This means that (fainter) stars should look like pinpoints in most cases, when using F/8 or faster.

In very bright stars the centre peak is NOT the main issue: the wings of the PSF become important. Poor optics distribute a lot of the energy into the wings of the distribution, causing "bloat" on bright stars. This is why top DSO imagers really want good control of aberrations. In planetary imaging, the wings mainly cause a reduction in contrast, not resolution. The reduction in contrast may be unwelcome, but can be dealt with to some degree by postprocessing. A fat centre peak is much more harmful, as it reduces resolution directly.

[thehand]

With regard to thermal: you're right, environmental conditions will have a big effect but so will the weather have an impact on seeing. It's a moot point. What is significant is the thermal equilibrium time in dead air with respect to thermal delta and humdity. With the thermal time constant you can calculate your longest cooldown time in steady state conditions. Or the manufacturer could give a few bounding points in a table. Rates will be dependant upon the thermal mass and transfer characteristics of the glass as well as the design of the mirror cell and tube.

It's a good metric for the manufacturer as a discriminator. We the consumers benefit from knowing prior to purchase whether the cooldown time will be acceptable. The only question is what constitutes thermal equilibrium and that 's something that can be measured and standardized.

[michael.h.f.wilkinson]

I doubt that will work. Thermal constants like this are fine indoors, and with objects with fixed pose, but outdoors they may mean little or nothing, and certainly if the pose of the object changes. Furthermore, thermal equilibrium itself is not the issue, it is tube currents. A little refractor like my 80mm will not suffer much because the tube is too small to support convection cells. The short, fat tube of the C8 is much more prone to this. What would make much more sense is to have specifications of effective cool-down times (i.e. until the tube currents no longer interfere significantly) for different initial temperature differences, and for some optimal pose of the scope.

[thehand]

Michael,

I have to disagree with you that environmental variables make any numerical approximation meaningless.  Measurements ratings for controlled tests are standard practice for everything from electrical components to salt spray testing for environmental enclosures.  Does it address all conditions, no, but it sets a standard by which things can be evaluated. 

I may not fully understand the mechanism of scope cool down, thank you by the way for clarifying, but I second the idea of effective cooldown times.  What is lacking is a standard to define the testing criteria.

The same goes for optics, and thank you by the way for clarifying the impact of mirror aberration. 

Optics is a mature field and I find it hard to believe that a telescope standard (IEC, ANSI, ect...) setting forth the methods of testing, measurement, and reporting is impossible.