[Raytrace] Optimization
Richard F.L.R. Snashall
rflrs@javanet.com
Thu, 31 Jan 2002 20:34:33 +0000
--------------030605040309010905030809
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
Peter John Smith wrote:
>Richard,
>
>I have run those files and get different results from yours.
>
>
>Step 1 was to optimize the design using the built-in wavelengths.
> The Polychromatic Strehl was measured using both the measures
> provided in Zemax and OSLO.
>
>
>Your file
>
>Attached (I hope). Step 1 is in file bk7_f4_doub_2.zmx
>
>
>Gives Strehl is close to 0.74
>
>
>The next file
>
> and Step 2a is
>in file bk7_f4_doub_3.zmx (I used up bk7_f4_doub_1.zmx with the original
>Baker design).
>
>
>Gives Strehl close to 0.755
>
>Then when I complete step 2b by replacing wl with the set
>
>>> wavelength weight
>>> 0.470 0.091
>>> 0.510 0.503
>>> 0.555 1.000
>>> 0.610 0.503
>>> 0.650 0.107
>>>
>
>I now get Strehl close to 0.675 which is more or less what one would expect
>to happen.
>
> These are different from your values of 0.84 and 0.86 which you reported.
>Just check these to make sure they are 0.8 + and not 0.7 +. Its not likely
>but double check it.
>
The discrepancy here is that we were looking at different Zemax outputs,
I think. For my measure, I was using the Stehl ratio given in the
Huygens PSF, while I believe you were using the RMS vs field Strehl
ratio.
Going back to the manual...
Zemax specifies that the RMS vs field measures use the primary wavelength
as a reference, but provide a wavelength weighted measure. I found that
the output seems to be independent of the primary wavelength.
The Huygens measure does not specify that it is wavelength weight sensitive,
but I tweaked a couple of weights and found it is, in fact, sensitive to
the weights.
I have, however also found that the Huygens PSF measure is sensitive to
the sampling. For step 1, I got:
32x32 0.840
64x64 0.826
128x128 0.820
256x256 0.816
The OSLO PSF for this is 0.812... looks like Zemax might be heading
there.
Even with the PSF value of 0.812, there is still some difference between
that and 0.74. Are these supposed to be the same value (or one an
approximation of the other)?
But also back to the question that brought all these meanderings:
When I originally tried this, I was attempting to match
the locations of the nulls to those of Baker's original design.
Why would Baker have chosen the solution closer (at least I think it
is) to step 2b, than to step 1?
Rick S.
>
>I do not know what is going on. My Zemax that I used here is Version 10.0 -
>Aprip 2001.
>
>As a matter of interest.
>
>I optimised after your step 1 - stays the same
>
>Then optimised step 2a - stays the same
>
>Then changed wl to the original and optimised and Strehl went back to 0.74
>so the optimiser seems identical.
>
>I really dont know.
>
>Peter.
>
>
>
>
>
>
>
--------------030605040309010905030809
Content-Type: text/html; charset=us-ascii
Content-Transfer-Encoding: 7bit
<html>
<head>
</head>
<body>
<br>
<br>
Peter John Smith wrote:<br>
<blockquote type="cite" cite="mid:02b601c1aa4d$773af200$644b8690@yourte6r9bjnhd">
<pre wrap="">Richard,<br><br>I have run those files and get different results from yours.<br><br><br></pre>
<pre wrap=""><!---->Step 1 was to optimize the design using the built-in wavelengths.<br> The Polychromatic Strehl was measured using both the measures<br> provided in Zemax and OSLO.<br></pre>
<pre wrap=""><!----><br>Your file<br><br></pre>
<pre wrap=""><!---->Attached (I hope). Step 1 is in file bk7_f4_doub_2.zmx<br></pre>
<pre wrap=""><!----><br>Gives Strehl is close to 0.74<br><br><br>The next file<br><br></pre>
<pre wrap=""><!----> and Step 2a is<br>in file bk7_f4_doub_3.zmx (I used up bk7_f4_doub_1.zmx with the original<br>Baker design).<br></pre>
<pre wrap=""><!----><br>Gives Strehl close to 0.755<br><br>Then when I complete step 2b by replacing wl with the set<br><br></pre>
<blockquote type="cite">
<blockquote type="cite">
<pre wrap=""> wavelength weight<br> 0.470 0.091<br> 0.510 0.503<br> 0.555 1.000<br> 0.610 0.503<br> 0.650 0.107<br></pre>
</blockquote>
</blockquote>
<pre wrap=""><!----><br>I now get Strehl close to 0.675 which is more or less what one would expect<br>to happen.<br><br> These are different from your values of 0.84 and 0.86 which you reported.<br>Just check these to make sure they are 0.8 + and not 0.7 +. Its not likely<br>but double check it.<br></pre>
</blockquote>
<tt>The discrepancy here is that we were looking at different Zemax
outputs,<br>
I think. For my measure, I was using the Stehl ratio given in the<br>
Huygens PSF, while I believe you were using the RMS vs field Strehl<br>
ratio.<br>
<br>
Going back to the manual...<br>
Zemax specifies that the RMS vs field measures use the primary wavelength<br>
as a reference, but provide a wavelength weighted measure. I found that<br>
the output seems to be independent of the primary wavelength.<br>
The Huygens measure does not specify that it is wavelength weight sensitive,<br>
but I tweaked a couple of weights and found it is, in fact, sensitive to<br>
the weights.<br>
<br>
I have, however also found that the Huygens PSF measure is sensitive to<br>
the sampling. For step 1, I got:<br>
<br>
32x32 0.840<br>
64x64 0.826<br>
128x128 0.820<br>
256x256 0.816<br>
<br>
The OSLO PSF for this is 0.812... looks like Zemax might be heading<br>
there.<br>
<br>
Even with the PSF value of 0.812, there is still some difference between<br>
that and 0.74. Are these supposed to be the same value (or one an<br>
approximation of the other)?<br>
<br>
But also back to the question that brought all these meanderings:<br>
</tt><tt> When I originally tried this, I was attempting to match<br>
the locations of the nulls to those of Baker's original design.</tt><br>
<tt>Why would Baker have chosen the solution closer (at least I think
it<br>
is) to step 2b, than to step 1?<br>
<br>
Rick S.<br>
</tt><br>
<blockquote type="cite" cite="mid:02b601c1aa4d$773af200$644b8690@yourte6r9bjnhd">
<pre wrap=""><br>I do not know what is going on. My Zemax that I used here is Version 10.0 -<br>Aprip 2001.<br><br>As a matter of interest.<br><br>I optimised after your step 1 - stays the same<br><br>Then optimised step 2a - stays the same<br><br>Then changed wl to the original and optimised and Strehl went back to 0.74<br>so the optimiser seems identical.<br><br>I really dont know.<br><br>Peter.<br><br><br><br><br><br><br><br></pre>
</blockquote>
<br>
</body>
</html>
--------------030605040309010905030809--