Al Slagle said:

>Perhaps IPC is correct. Let's let the fab shops grab the designers by the hair
>and rub our noses in photo-resist! I may walk away a little red in the face,
>but at least I might learn something useful.
>
>I thought the process went like this:
>1. Fab shop produces film from gerbers.
>2. A chemical called a photo-resist is then deposited onto the board material.
>3. A light source exposes resist thru the film, be it a positive or negative.
>4. A chemical bath then removes the unwanted resist and unwanted copper.
>
>
>I may have something out of sequence or over simplified, but I must be close.

True. You have vaguely described wet lithography.

>Why can,t a process be created to expose the resist using the gerber files?
>And skip the photo-tool process? If a laser photoplotter can expose film at
>.00025 resolution? Why can't it expose the board w/resist with the same amount
>of accuracy? It would eliminate some of the problems inherent to the system.
>Registration would be greatly improved. Fine line technology can take great
>advantage of this improved resolution in trace to trace/pad capability.

It's too slow. You don't care about the time it takes to plot a set of
film, but the time it might take to expose each panel with a raster process
would be very significant.

Lithography as applied in the semiconductor world today can define lines
down below the micron line width. To do so, they apply resist by spinning a
coating on a wafer at thickness near the desired line width, then
b-staging, exposing, and developing the resist. This is merely in
preparation to etch, sputter or other wise apply/remove the pattern, then
you strip the resist.

Resolution is limited by how thinly and evenly you can apply the resist as
well as how accurately you can expose and develop.

Semiconductor lithography can expose by using a mask with either off
contact or projection exposure, or it can use beams of electrons, laser, or
x-ray to define  a pattern. Either way, it has much greater definition
capability than typical PWB fab lithography. As you go to
direct-step-on-wafer approaches (sort of like what you suggest) rather than
photomasks, it uses equipment orders of magnitude more expensive, and much
slower.

>It should make our boards cheaper! Right?

Doubt it. If PWB lithography follows the lead of semiconductor processes
(and it seems logical to me) it would not happen that way.

The process of using a projection mask to expose liquid resist can work
well down to a few microns in size (state of the art semiconductor fab
circa 1985) so I suggest the use of pattern masks to expose resist with
collimated light sources is _not_ a process hindrance in achieving fine
line definition, and it is a proven efficient approach.

regards,

Jerry Cupples
Interphase Corporation
Dallas, TX