[Raytrace] A real world application

[Peter Chen]

Hi Mike:

            Many thanks for checking out and critiquing my spherical mirror 
design.

 >Someone corrected me once for inadvertently anglicizing this name, so 
I'll >do the same. It's Pressmann-Camichel

       Any idea who those folks are?   I have never come across any 
reference to these names except in Rutten and van Venrooij.


 >Plugging those numbers into OSLO I still get lousy results: after 
refocusing I> get a P-V error of over 2.7 waves. What gives? Well, I solved 
to >eliminate 3rd order spherical aberration, but this system has a 
huge >amount of residual 5th and even 7th order spherical. In order to get 
good >performance you're going to have to make your secondary a higher 
order >asphere.

        I plugged your design into Zemax.  I noticed that the spot diagram 
is fairly big, consisting of a several micron spot with a faint outer 
ring.  This persists even after optimization.   I puzzled over the cause of 
the ring.  My guess is it must be caused by those high order spherical 
aberrations that you mentioned.

 >It's also probably outside the scope of ATM projects,
 >I don't think many ATMs would care to tackle a fairly steeply 
curved >convex oblate spheroid with an 8th order correction.

         I agree this is definitely not main stream ATM stuff at the 
moment.  However, I submit that:
1.  The ATM community is a huge pool of people and expertise.  I don't know 
that anything is outside the scope.   Two examples:
    a.  I understand that some ATMs are making 40" and 72" 
scopes.   Definitely not main stream either.
     b.  Raytracing is an activity normally done by highly trained 
professionals.  I suspect that giving a talk on raytracing at a standard 
star party -or even at Stellafane or Riverside- would likely lose (or 
frighten away) a large part of the audience.   However, you yourself and 
Peter Smith and others on this list are experts on the subject -- and doing 
it for fun!   I rest my case.   -:) smile.

2.   Spherical mirrors are a rarity at the moment.  However, most of the 
designs for very large telescopes of the future use spherical primary 
mirrors for reasons of affordability.   I think it quite possible that, 
once some of these humengous beasts have been built and proven in use, 
there will be ATMs who'll want to try out similar ideas.

         Thanks for the discussion.   Wishing you and everyone on the list 
a Happy New Year.

Regards,
P.C. Chen

At 01:50 PM 12/22/01 , you wrote:
>At 01:19 12/21/2001 -0500, Peter Chen wrote:
>>          The primary being spherical, the best that can be done in a two 
>> mirror system is to make a inverse Dall-Kirkham (or Pressman-Carmichael, 
>> Rutten's terminology).  The secondary is a convex spheroid.  This is a 
>> zero field system.  The image is a spot only on the optical axis.  I've 
>> added a diagonal flat for finding a target in the field.
>
>Someone corrected me once for inadvertently anglicizing this name, so I'll 
>do the same. It's Pressmann-Camichel. Wilson calls these spherical primary 
>telescopes; for some reason he decided not to give the designers credit 
>(or blame) for this one.
>
>>         The system requirements are as follows:
>>Primary - 20 in diameter, 76 in. ROC, sphere
>>Secondary - 4 in diameter (to reduce fabrication cost, actually any size 
>>up to 6 inch is ok)
>>Diagonal flat -  to be located 2 inch in front of primary, for ease of 
>>mounting
>>Focus  - 13 in. above optical axis (to accommodate focuser)
>
>This is actually a more difficult design (and probably fabrication) 
>problem than you may have thought. It's also probably outside the scope of 
>ATM projects, but since there doesn't seem to be much list traffic right 
>now any question has to be considered "frequently asked."
>
>>The problems are:
>>    1.  It doesn't focus to a spot, even on axis.   The secondary roc 
>> doesn't seem quite right.
>>     2.   I have no idea how difficult it would be to make a spheroid 
>> with a conic constant of 8.959129.  Is that realistic?  How does one 
>> test such a piece?   (note:  The piece will be concave.  We replicate 
>> with it and get the convex part).
>>
>>:  Does anyone have any suggestions?
>
>Personally I prefer to use a spreadsheet for preliminary cassegrain 
>design, then fiddle it in OSLO as needed. The version of my cassegrain 
>design spreadsheet on my web page doesn't do spherical primaries, but I 
>added a line to my local copy.
>
>I see two likely errors in your layout. First, it appears you made your 
>secondary a flat. I reverse engineered your probable intended 
>configuration and came up with a secondary radius of curvature of 
>-472.89mm, which produces a system focal ratio of f/10.35 or a secondary 
>magnification of -5.447. The second probable error I see is that my 
>spreadsheet returns a conic constant of 7.086, which is still highly oblate.
>
>Plugging those numbers into OSLO I still get lousy results: after 
>refocusing I get a P-V error of over 2.7 waves. What gives? Well, I solved 
>to eliminate 3rd order spherical aberration, but this system has a huge 
>amount of residual 5th and even 7th order spherical. In order to get good 
>performance you're going to have to make your secondary a higher order asphere.
>
>OSLO has several ways to set up a general high order aspheric surface. I 
>picked the simplest one, which it calls a "standard asphere." I don't know 
>how to solve for higher order aspheric coefficients by hand, so I used 
>OSLO's default optimization method (they call it GENII) and defined the 
>secondary conic constant and 6th and 8th order aspheric coefficients as 
>variables. Here's the system specification that OSLO returns (I've omitted 
>the tertiary flat):
>
>
>*LENS DATA
>No name
>  SRF      RADIUS      THICKNESS   APERTURE RADIUS       GLASS  SPE  NOTE
>  OBJ       --        1.0000e+20    1.0000e+14             AIR
>
>  AST  -1.9304e+03   -772.160000    254.000000 AS      REFLECT
>
>   2   -472.890000    772.160000 P   50.800772 S       REFLECT   *
>
>   3        --        279.408778 S   13.501962 S           AIR
>
>   4        --            --          0.005258 S           AIR
>
>  IMS       --            --          0.005258 S
>
>*CONIC AND POLYNOMIAL ASPHERIC DATA
>  SRF        CC          AD          AE          AF          AG
>   2       7.092349      --     -9.6589e-14 -1.2996e-17      --
>
>As specified OSLO tells me P-V error is about 0.004 wave. You could 
>actually, probably, get away with a 6th order asphere, but P-V would only 
>be about 0.11 wave.
>
>I don't think many ATMs would care to tackle a fairly steeply curved 
>convex oblate spheroid with an 8th order correction. Depending on what 
>kind of resources you have available at NASA you might be able to make 
>one, but I doubt the opticians who work on it would thank you.
>
>Mike Peck
>
>_________________
>
>Michael Peck
>email mpeck1@ix.netcom.com
>Wildlife photography page http://home.netcom.com/~mpeck1/index.html
>Amateur telescope making http://home.netcom.com/~mpeck1/astro/astro.html
>