Jeff,
In principle, this should work, but as usual, the devil is in the
details. The most obvious to me is that you are limited to small
boards with this set up. The glass has to be larger than your
original, and it also has to have a flat field, so as not to
introduce distortion. Large glass with good optical characteristic
tends to be expensive. The glass will also have to be of a type
that has a high transmission of UV if you are exposing
photoresist. You will have to provide a mechanism to position the
lamp/lens combination to optimize results, and also a mechanism
for focusing the image on the board. Focus could also be an issue
if going to photoresist. The focal point for UV light will be a
bit different from white light because of the shorter wavelength.
Somewhere in the setup, there has to be a structure to place the
lens the right distance from the board to produce the exact
magnification. This would be an awesome project...
I am not sure what problem this solves. The major problem with
laser printing that I see is that the narrower the trace, the
fewer dots there will be to generate the trace width. With a 600
dpi printer, a 10 mil trace that is aligned with the printer scan
direction will have 6 dots. This is relatively reasonable. A 5 mil
trace with 3 dots is not. Changing to a 1200 dpi printer solves
that problem, but at expense. Clearly, the transfer process is
limited in trace size by the resolution of the printer. Printing
on transparencies, and using them as photomasks with the
photoresist process has the same limitation because we are still
dealing with too few dots. To do this properly requires a high
quality laser plotter to make the photo mask, or in direct write
to the photoresist.
Harvey
A setup for a reducing lens - http://archive.cnx.org/resources/fb001372e360c7bddb6939364bfb77a57420e7dd/5.15.png
