Subject: Re: UV light source for photoresist?
From: jmelson@artsci.wustl.edu (Jonathan M. Elson)
Date: 18 Mar 1995 06:05:26 GMT

Randy Gordon-Gilmore (rgordongilmore@haley.genetics.bio-rad.com) wrote:
: I am doing some photoetching of sheet brass, and am using dry-film aqueous 
: photoresist from Dynachem (blue polymer film).  It polymerizes using 
: long-wavelength ultraviolet light, and I had (last summer and fall) been 
: using noontime sunlight with a 1'15" to 1'30" exposure time.  However, 
: there has been no appropriate sunlight for the past couple of months. :(

I am currently using a 400 Watt Mercury-Vapor lamp with the outer glass
envelope broken away, leaving the inner quartz bulb.  I got the whole fixture
with bulb for $25.00 at a junkyard.

Several people who do similar stuff for printed circuit boards like I do
suggested the black-light bulbs for patio bug-zappers.  They said those
bulbs worked great.  The wavelength you need are the mercury lines from 405nm
and 357nm.

Subject: Re: chemical etching question
From: jmelson@artsci.wustl.edu (Jonathan M. Elson)
Date: 23 Mar 1995 06:40:53 GMT

Jon Lake (jlake@ix.netcom.com) wrote:
: I would like to make some small flat prototype parts and I think that
: chemical etching would give me the best quality.  Are there any
: "hobbiest" kits out there for chemical etching or should I try to
: avoid doing this myself?  I would like the material to be copper, but
: aluminum would work.  On occasion, I would like to make small, flat

This is very similar to the manufacture of printed circuit boards.  The metal
on a PC board is about .001 to .002 copper, laminated to a fiberglass
substrate.  What you want to do is exactly the same thing without the
substrate.  I have tried several processes, and had poor results.
The material I now use is "dry film resist", made by DuPont and others.
It is a standard material in the printed circuit industry.  It is hot-roll
laminated onto the copper, exposed to UV light through a photo-master,
and developed with Sodium Carbonate.  The copper is then etched with Ferric
Chloride, or similar acidic etchant.  (Alkaline etchants will dissolve
the resist.)  The resist is then stripped with Lye (I use Drano crystals,
making a new solution each time).  This process is discussed in newsgroup
alt.electronics.cad in much detail.  This process should work fine for you,
too.  Making the photo-master from a laser printer output is a little tricky.
I contact expose KODAK GLP-7 photoplotter film, and then may contact print
that again to get back to a positive.  Any high-contrast litho film will 
work.  It takes a lot of light to penetrate the laser printer paper, so
you might want to use laser transparencies.  You probably still need the
litho film, however, because the contrast ratio of laser printer output is
not as high as you might think.  Copying to the litho film greatly enhances
the contrast of the laser printed original.  If you have any questions,
E-Mail me.  Oh, yes, the dry film resist is available from suppliers to the
printed circuit board industry.  Kepro in St. Louis sells it, among others.

Subject: Re: Precision drilling
From: jmelson@artsci.wustl.edu (Jonathan M. Elson)
Date: 27 Dec 1995 19:44:47 GMT

Michael Covington (mcovingt@ai.uga.edu) wrote:
: I'm needing to drill holes in printed circuit boards accurately.
: Typically, the holes are 100 mils apart and 40 mils in diameter.
: As a result, a hole that is even 10 mils off its intended position
: will look noticeably out of place.

: So far I've been using a cheap ("Task Force," $100) drill press
: at 2600 rpm (highest speed available).

: The drill press has about 1/64" (15 mils) of horizontal play, as
: measured at the end of a relatively short drill bit.  Is this excessive?
: Or is it desirable to prevent bit breakage?

This is absolutely not desirable, except to save money. (theirs)
You need a drill spindle with slop less than .003" or so to do circuit
board drilling.  Use carbide drills, at about 40,000 RPM for your
.040" holes.  You need an x-y table which can be moved precisely,
or a very steady hand.

: I think the key to success is probably going to be center-punching
: all the holes, which shifts the problem slightly -- now I've got to
: figure out how to center-punch with sufficient precision.

No, center punching the holes makes the drill follow the center mark,
which you can never get absolutely centered under the drill, so the hole
is drilled at an angle, and will cause drill breakage.

: Also, I've found that the carbide-tipped bits that are usually used
: for printed circuit board drilling have considerably more of a tendency
: to wander than conventional twist drills do.  I don't know whether they
: last longer; presumably they stay sharp longer, but they're very brittle
: and I always lose them to breakage rather than dulling!

These drills have CONSIDERABLY LESS tendency to wander than a conventional
jobber's length steel drill.  That's why they're made that way.  Commercial
PC board drilling is done without any marks or centers on the blank board,
so the drill must SELF center.  Only a drill of the PC board type will
do that properly.

: Any thoughts, anyone?  I've been thinking about trying to build a
: precision horizontal boring machine around a Dremel Moto-Tool, possibly
: with a jig to produce holes 100 mils apart for IC pins.

A dremel also has poor spindle bearings, but at least better than you
have presently.  You will also note it has no Jacobs chuck, but uses
collets.  This may be needed on your drill press.  Jacobs chucks are 
notoriously poor for small, precision work.
See if you can find something with better spindle bearings than a Dremel,
and definitely use a precision collet instead of a chuck.
Also, make sure you can drill at 40,000 + RPM, or your drills will burn
their way through the board instead of drill it.