Hello,

     This message is going to everyone who's emailed me about my homebuilt
900 mm lens project that I discussed on rec.photo.advanced.  Thanks for 
the response.  I've heard from about 25 people.  That's a large enough
number to be gratifying, but also a small enough number that I'll be able
to send some prints to everyone who's asked.  (So far anyway.)

     By the way, at least two of you have asked for photos but didn't
include your mailing addresses.  If that's you, please save me some time
and send me another message.  Otherwise I'll have to contact each of
you individually.

     I thought I'd also answer some of the questions and comments that
I've gotten so far.  Here goes:


1.  How about sending some engineering drawings of the parts?

     I will be able to send some exploded isometric views, but I've decided
that there wouldn't be much value in sending detailed drawings of the
parts.  The parts were designed to specifically match the tools I have
available, the materials at hand, and my fabrication skills.  The chances
of my design details being the right solution for anyone else are pretty
small.  I think my sketches will give you enough detail that you could
make up your own design if you want to.


2.  What would it take to make the lens faster, say, f/8 or so rather
    than f/11?

     I did a little more ray tracing on this one.  At f/8, the maximum
spot size grows to about 34 microns (compared with 16 microns at f/11).
I'm a little unclear on just how spot size relates to MTF, but clearly
the lens would be less sharp at f/8.  It's easy to make another aperture
and try it out, so maybe someday I'll get a chance to find out.

     I do know that resolution is related to the distribution of the
rays inside the spot, so just knowing the maximum spot size is only an
approximation at understanding the resolution.  

     The other problem is that the edges of the picture will almost
certainly vignette at f/8, especially at close focus distances.  But I
estimate that at 50 yards or more, the effect won't be noticeable.


3.  How about scanning the photos and exchanging the electronically
    instead of using the U.S. Mail?

     I do have access to scanners  (My job at HP is mechanical design
of scanners.  I designed some of the internal parts of the ScanJet IIc
and ScanJet IIcx.),  so I can create good image files.  But I'd need some
instruction about where would be an appropriate place to post large
files.  The file would have to be quite large to even begin to represent
the detail you can see on a 3X5 print.  I'd estimate four megabytes or 
so.

     Also, for security reasons, my workstation isn't accessible on the
open net for ftp transfers.  I don't think it's practical to email such
large files.  Maybe someone else can offer to be the ftp server?

     Anyway, you can probably tell that even though I know how to make
the files, I'm hazy on the internet techniques and etiquette involved
in exchanging them.  Please feel free to educate me.


4.  What kind of ray-tracing software did you use?

     I used a program I wrote myself for work.  Had I known it would grow
to what it is, I would have bought a commercial program, but originally
I was just trying to model some rays bouncing off of flat mirrors.
Eventually, I worked out the formulas for reflection and refraction off
of and through flat surfaces, cylinders, spheres, toruses (tori?), and
even some ashperic shapes.  I can also do some automated optimization.

     As with most homebrew software, it's probably not usable by anybody
but me.  


5.  What special tools were needed to make the lens?

     Mostly I used standard machine-shop stuff.  All the parts were made
on either a lathe or milling machine.  I did use two different lathes for
the large and small parts.  Probably the trickiest thing was boring a
62mm diameter hole all the way through the six-inch pipe.  I had to do
it all from one side so that the pieces mounted on the two sides would
be in good alignment.  I had to ask several machinists before I found a
boring bar long enough.


6.  Is the performance of the lens limited by design or by the mirror
    quality?  

     All of the performance information I've given is for a perfect
mirror.  I haven't made any serious efforts to model mirror imperfections.
The Edmund catalog says the mirror is "1/4 wave", which sounds pretty
good to me.  

     The problem with modeling mirror imperfections is that the magnitude
of the imperfection isn't the important thing.  The important thing is 
the _slope_ of the deviation from perfect.  I suppose I could try to estimate
the curvature induced if the worst-case 1/4 wave deviation happened in
a one-inch traverse across the mirror, and then model a mirror with _that_
curvature.  I think it's also possible to characterize a surface on an
interferometer and input that data into some commercial ray-tracing
packages.  But that sounds more serious than I feel like getting.


7.  How can the image quality be any good with just a single-element
    system?  (That's paraphrased a little.)

     The key here is the small aperture.  Spherical aberration grows
dramatically as the aperture increases, and I understand that the other
aberrations are also very sensitive to it.  The narrow field angle due
to the long focal length also helps.


8.  Do you have any plans to market this lens or plans for it?

     I doubt that would be practical, given the labor-intensive nature
of it and the almost-certainly-small market.  I did it for fun.



     So there you have it.  My plan is to get some good shots of the
upcoming full moon, and get stuff mailed out within two or three weeks
of that.

     Thanks for your interest,

     Dave Boyd
     Hewlett-Packard, Greeley, Colorado
     daveb@hpgrla.gr.hp.com
     (303) 350-4566


P.S.  Some late-breaking news:  I may have found someone to make some
image files accessible to ftp retrieval.  Those of you who would prefer
that, I'll keep you posted.