Reading Optics Reports

[Rob Sellent]

From time to time when discussing a scope's optics or performance, I often see folk posting up optic reports. The problem is being a bit dim witted, I just can't make head or tail of them.

I've picked out this comparable report - not entirely by random - and was wondering what it is I should be 'reading' and what it is those figures/colour images are telling me?

Below are the images simply cut and pasted from the link:

Vixen 102s:

 

Tak 128:

 

Edited November 27, 2019 by Rob Sellent

[John]

Personally I cant make much of the pictures so I tend to concentrate on the figures, particularly Peak to Valley, RMS and Strehl. The wavelength of light that these measurements were made under is also significant I believe. Some wavelengths are more significant for imaging and others for visual observation.

The other thing is that the tests just relate to a single example of course. You get some variation across a production run. Less probably for a premium product.

I doubt there are any poor Vixen 102 fluorites or Tak FS 128's around though !

(Canon Optron made the objectives for both of them BTW)

 

 

[jetstream]

38 minutes ago, Rob Sellent said:

and was wondering what it is I should be 'reading' and what it is those figures/colour images are telling me?

Mr Rohr is VG and is a trusted source.

First thing, you have selected test reports done in 2 different wavelengths, they all need to be checked as results can vary color to color- espc doublets IMHO -s there a blue test result?

Anyway, a top optician explained how to read an interferogram very simply- the lines should be straight. Also the best test report includes a sample of the many actual pictures, not just computer generated synthetic interferogram images.

PV is the least of my concern if the RMS and Strehl are good.

One thing I always ask for is a truly smooth mirror, something that is not on the forefront of discussion necessarily. These 2 top refractors will perform very good I would imagine- and Canon makes very smooth, top quality lenses IMHO.

[vlaiv]

1 hour ago, jetstream said:

Also the best test report includes a sample of the many actual pictures, not just computer generated synthetic interferogram images

This is debatable, and it depends on technique used to asses wavefront. If interferometer is used then yes, it should include image, but interferometer is not always needed to determine above.

2 hours ago, Rob Sellent said:

rom time to time when discussing a scope's optics or performance, I often see folk posting up optic reports. The problem is being a bit dim witted, I just can't make head or tail of them.

As pointed above - probably single interesting piece of data on such report is Strehl figure at particular wavelength.

I'll just do a brief explanation on what is going on for all interested (and partly because I'm actually rather involved with that topic at the moment ) :

All tests are done on optical axis, so this sort of report is only giving you idea of how the scope will perform on axis - it tells you nothing about off axis aberrations (which also depend on telescope design). Principal idea is that wavefront from a star (point source - single point along optical axis) is arriving at telescope aperture perfectly flat - uniform.

After telescope does it's thing and focuses light to a point - any optical "defects" of particular telescope act as if someone "twisted" that wavefront. Above colorful diagram titled : Wave front is displaying 3D image of wavefront phase (you can think of it as how much is rest of wavefront late with respect to point that first reaches aperture of scope).

This does not show surface of your mirror or lens or whatever! It relates to light wavefront.

Perfect flat wavefront will create perfect image that telescope aperture is capable of delivering. Any bending/ripple in that wavefront causes image degradation.

That is all you need to characterize performance of the telescope (on axis) - wavefront. It defines PSF (or image of the star with perfectly still atmosphere), MTF or diagram that shows how higher frequencies get attenuated and basically all other info on that report.

Any sort of deformation of wavefront can be characterized by set of functions (their sum) - called Zernike polynomials. Many of those polynomials correspond to a certain optical aberration or other effect on telescope. For example tilt and defocus (removed in this report because they are not inherent in optics) - represent angle of incident light (either star that is not on optical axis or tilted focuser / sensor) and actual defocus (focuser distance to ideal focus position).

Then there are astigmatism, coma, spherical aberration, etc ... All of these are represented by particular Zernike polynomial.

It's a bit like decomposing a vector into unity vectors (3X + 7Y + 1Z sort of thing) - so does wavefront decompose into these Zernike polynomials. Those aberrations that are not symmetric (tilt, coma, astigmatism) have both angle and intensity, while symmetric only have intensity (defocus, spherical, ...)

For further info - check wiki page: https://en.wikipedia.org/wiki/Zernike_polynomials

These figures are reported on the left. If one had complete decomposition to Zernike terms, one could reconstruct wave front of that particular telescope, but you would need more and more terms to get more precise description of wavefront - here in these reports only first few terms are recorded so it is only coarse representation of recorded wavefront.

 

 

[davidc135]

As has been said above, as a wave front moves through a telescope parts of it are advanced or retarded by errors encountered in the optics. The overall range is given by the peak to valley term (pv). More useful is the RMS (root mean square) figure which averages the errors over the whole aperture. The Strehl figure refers to the amount of light that finds its way into the Airy disc compared to a perfect optic. As the Strehl goes down more light spills out of the disc into the surrounding rings reducing contrast. The PSF (point spread function) illustrate this.

The MTF (modulation transfer function) curve compares the image contrast of the instrument with the ideal for widely separated points at left through to very close, at the limit of resolution points, at right. These two objectives are both close to perfect, the curves being not far off the ideal and show why high quality refractors give such beautiful planetary images. With poorly figured objectives or with sizeable secondary obstructions the MTF curves fall well below the ideal especially in the middle section, reducing contrast.

Hope I got that right.  David

[Rob Sellent]

Thanks guys for the great explanations. In particular to @vlaiv and @davidc135for patiently going through this.

If I've understood correctly, for a appreciation of these type of graphs would it make sense to assume that the nearer to 0 better is the RMS and Peak to Valley (P-V) criterion and nearer to 1 better the Strehl ratio? Together indicating a flatter wavefront 3D image?

I also recall Dob guys - especially when considering buying a new mirror - suggesting that when evaluating potential optical systems it's a good idea to look for a pair of numbers, for example: 1/6 wave = .93, 1/7 wave = .94, 1/8 wave = .95, 1/9 wave = .96, 1/10 wave = .97, or something like that. In this case, I imagine that the wave fraction is related to the P-V and the decimal number to the Strehl ratio. Again, better the optic flatter the wavefront, lower the fraction towards an ever smaller portion and higher the decimal towards 1?

Edited November 27, 2019 by Rob Sellent

[vlaiv]

32 minutes ago, Rob Sellent said:

Thanks guys for the great explanations. In particular to @vlaiv and @davidc135for patiently going through this.

If I've understood correctly, for a appreciation of these type of graphs would it make sense to assume that the nearer to 0 better is the RMS and Peak to Valley (P-V) criterion and nearer to 1 better the Strehl ratio? Together indicating a flatter wavefront 3D image?

I also recall Dob guys - especially when considering buying a new mirror - suggesting that when evaluating potential optical systems it's a good idea to look for a pair of numbers, for example: 1/6 wave = .93, 1/7 wave = .94, 1/8 wave = .95, 1/9 wave = .96, 1/10 wave = .97, or something like that. In this case, I imagine that the wave fraction is related to the P-V and the decimal number to the Strehl ratio. Again, better the optic flatter the wavefront, lower the fraction towards an ever smaller portion and higher the decimal towards 1?

Indeed. You can think of P-V as difference between lowest point on a map (sea level) and highest mountain peak. It tells you what is the range of values on wavefront but it does not tell you anything about what the relief is (single mountain surrounded by ocean or something else). RMS is akin to "average terrain level", in the same sense guide RMS is average error of guiding (it is not average displacement as that would be 0 because any error in +RA would cancel any error in -RA, and similarly in DEC).

Strehl is very good indicator of optical performance as it deals with energy and that is what we see.

In any case, P-V and RMS of 0 and Strehl of 1 represent perfect optic - flat wavefront.

1/6 wave means exactly that in P-V - meaning that max difference between two points on wavefront is less than one sixth of wavelength. It roughly corresponds to 0.93 Strehl. If low order spherical aberration was only one present and it had magnitude of 1/6 waves then Strehl would be 0.93. Similarly goes for other values.

In some cases aberrations can sort of cancel a bit out - look at first report. P-V for that scope is 1/5.6 but Strehl is 0.97. Now if scope had P-V of 1/5.6 and only aberration present was low order spherical aberration - strehl would be worse than this.

It is similar for relation of PV to RMS. In above two diagrams you have same P-V error, but different RMS errors as contribution of aberrations is different.

Mind you, if you get above report for newtonian mirror - that is not the whole story as you need to account for secondary as well. Secondary mirrors as well as diagonals have usually 1/10 or better surface, but since things compound that error can be actually beneficial or can make things worse - depending on respective surfaces.

Same goes for lens in refractor. Aberrations that are not symmetric can somewhat cancel out between elements and rotating lens elements can affect final wavefront. That is the reason why lens elements are marked for orientation. If you rotate them, you can actually make things worse as aberrations that canceled out start "reinforcing" instead.

If you are interested in doing such report yourself, there is a way to do it. It requires planetary type camera or guide camera and a piece of software.

Process is as follows: you take images of defocused star (both in and out focus), stack them in certain way and import into WinRoddier software for Roddier analysis. Software does it's magic and you get something like this:

Above is test for my RC8 scope. I'll probably repeat it at one point just to verify results and also make sure collimation is spot on this time (not sure about last measurement).

I'm currently fiddling around with Zernike polynomials and trying to figure out how I would implement above software (already have some ideas).

Both Aberrator and WinRoddier might be a bit outdated software and I think astro community could do with the same (or enhanced) functionality.

[johninderby]

R F Royce’s classic article on Strehl is well worth a read.

http://www.rfroyce.com/standards.htm

[Rob Sellent]

Incredible post @vlaiv. Expertly explained and wonderfully written for a dummy like me. I really appreciate your help and the time you've taken. Thank you 

[Captain Scarlet]

A read of Harold Suiter's "Star Testing Astronomical Telescopes" gives a very good guide through all of this, in some detail but not too text-book. And much much more besides, including a great section on Newtonian Collimation.

[jetstream]

18 hours ago, vlaiv said:

This is debatable, and it depends on technique used to asses wavefront. If interferometer is used then yes, it should include image, but interferometer is not always needed to determine above.

I like an actual picture as it is a quick, easily understood thing without the need for technical expertise. Crooked lines bad-straight ones good lol!

        

 

Anyway, the most important thing in all this is finding and using an honest trusted optician for the tests IMHO. Some of the best don't provide their test numbers ie Lockwood, but he will not sell a poor mirror only top ones.

If there is a dispute an interferogram picture is pretty hard to dispute compared to the computer generated representation of it and all the numbers including RMS and Strehl which uses the RMS to calculate.

All the numbers should be in line with each other (and picture) IMHO- except the PV... I'll take a 1/5 wave PV .92 Strehl with honest numbers over the many encountered 1/10 PV .98+ Strehl some makers put out any day.

A true "1/4 wave" system actually gives very good views IMHO.

I do appreciate your technical posts as I do gather more insight from them Vlaiv- thanks, Gerry

[jetstream]

5 hours ago, Captain Magenta said:

A read of Harold Suiter's "Star Testing Astronomical Telescopes" gives a very good guide through all of this, in some detail but not too text-book. And much much more besides, including a great section on Newtonian Collimation.

I like the MTF graphs in the book- they visually show the interesting effects of central obstruction, thermals, optical quality and seeing. Another  good source is Peach's article

http://www.damianpeach.com/simulation.htm

Edited November 28, 2019 by jetstream