Homebrew PCBs

It would be a lot easier to by a used process camera than to try and build
your own system for reduction. There are lots of business' closing doors in
this economy, and used equipement is really cheap.

If you did try to build your own, there are a multitude of things that can
go wrong. You are trying to accurately create a 2:1 reduction in image size
over a large area. Now, if you don't need fine resoultion over a large pcb,
eg 15//15 across a 4 x 6 pcb, a home brew unit may work fine, and its a
matter of experimentation to see what will work.

A single simple lens will not work mainly due to the fact that sin(x)=x for
only small angles which is the basis of the thin lense equation. Once you
start having running ray traces with angles greater than 10 degrees, you
need to start using series expansion techniques and the matrix of design
starts getting a little crazy. You also find out that in short order, you
either need to go with multiple lenses, and hope that you can build the
assembly to the tolerances needed, or you need to have an aspheric lense
made. Sorry to be a bit negative, but optical designs for high accuracy
reduction/enlargement are quite tricky.

If you do seriously want to get into optical design, Warren Smiths book on
Modern Optical Design is extremely valuable. I took a class from him when I
first started. There are a lot of very practical  hints in his book that are
lacking in a lot of the other more academic texts.

Steve Greenfield mentioned chromatic aberation, as most of the films used in
reproduction are monochromatic, it won't be an issue, but the other 4 common
aberations will be.

I consider myself at novice at optical desing too. I've taken probably 10
classes, designed hundreds of units, taken a number of designs through
production, and I still feel like I'm just beginning to understand the
concepts.
Thanks
Ron