[Raytrace] Spherical mirror telescopes

Peter Chen chen@stars.gsfc.nasa.gov
Fri, 07 Dec 2001 12:36:58 -0500


<html>
Hi:<br>
<x-tab>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;</x-tab>Just like
to express my thanks to John Upton for his post on raytrace
origins.&nbsp; The post included a message from Peter John Smith which
contained a pointer to the site
<a href="http://www.focussoftware.com/file_exchange/" eudora="autourl"><font color="#0000FF"><u>http://www.focussoftware.com/file_exchange/<br>
</a></font></u>which I did not know about.<br>
<br>
<x-tab>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;</x-tab>I went to
the site and downloaded the files for the Shafer telescope and Dan
Schroeder's  (author of Astronomical Optics) files.&nbsp; Dan's files
contained, as PJ Smith mentioned, examples of four mirror telescope
designs by RN Wilson.&nbsp;&nbsp; These are exactly the examples I was
looking for.<br>
&nbsp;<br>
<x-tab>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;</x-tab>For those
interested, the designs by RN Wilson are for the next generation of
extremely large (20-100m) ground telescopes.&nbsp; Telescope mirrors this
big cannot be made in one piece (for economical and practical
reasons).&nbsp; Hence they must be made of many segments.&nbsp;&nbsp; For
reasons of affordability and manufacturability, it is best to make all
the mirror segments identical.&nbsp; Hence the segments are all spherical
(spheres and flats are the only possible geometrical figures), resulting
in a giant spherical primary mirror.<br>
<br>
<x-tab>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;</x-tab>The
problem with a spherical primary mirror is that, unlike a parabola, it
does not focus parallel rays from infinity to a point.&nbsp; Therefore a
minimum of three mirrors are required to make a telescope that has good
resolution and an acceptable field of view.&nbsp; Three mirrors, however,
end up sending the light back towards the front of the scope.&nbsp;
That's why four is used in Wilson's design.&nbsp; Having an extra mirror
also permits some optimization, I'm told.<br>
<br>
<x-tab>&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;</x-tab>I loaded
the sample files into Zemax and looked at the 3d layout.&nbsp; They look
good.&nbsp; I need to learn more Zemax and raytracing to understand the
details.&nbsp;&nbsp; But thanks to this list, it has saved me a lot of
effort (as I have done in the past) trying to measure the diagrams with a
ruler and infer the optical parameters from the original published
articles.<br>
<br>
Regards,<br>
P.C. Chen<br>
</html>