Index Home About Blog
From: REMOVE_THISdwilkins@means.net (Don Wilkins)
Newsgroups: rec.crafts.metalworking
Subject: Re: Photo etching
Date: Fri, 21 Aug 1998 00:34:32 GMT

On Thu, 20 Aug 1998 11:40:24 +0100, Leon Heller
<leon@lfheller.demon.co.uk> wrote:

>In article <35dfbc5d.367963635@news.newsguy.com>, Reckless
><reckless2x@home.com> writes
>>On 20 Aug 1998 01:19:19 -0400, dnichols@d-and-d.com (DoN. Nichols)
>>wrote:
>>
>>>      I've used a frame which had 4 foot UV fluorescents on both sides,
>>>with the PC board and negatives sandwitched between two flexible mylar
>>>films, with a vacuum pulled between them.  It was really nice for making
>>>double-sided PC boards.  We never tried for multi-layer, however, because
>>>that is a *lot* more work. :-)
>>
>>So how does one go about making the plated-through vias on a double
>>(or more) sided board in a home shop environment?  A lot of the PLCC
>>devices I would like to use need lots of traces and vias quickly
>>become necessary.  I'd rather not have to put in rivets or solder
>>wires through the board.
>
>It's virtually impossible to do this at home and get it to work
>properly. PCB manufacturers themselves have problems getting it right.

I wouldn't say impossible but if you don't do it right it ain't gonna
come out right either in the kitchen or in a commercial facility.

The key is cleanliness, proper sensitizing the BCB board (through the
hole), proper electroless plating followed by conventional
electrodeposition. Each part of the process must be done correctly
which leaves several opportunities for failure.

Murphy's law is always valid.



From: REMOVE_THISdwilkins@means.net (Don Wilkins)
Newsgroups: rec.crafts.metalworking
Subject: Re: Photo etching
Date: Fri, 21 Aug 1998 17:18:10 GMT

On 20 Aug 1998 21:48:41 -0400, dnichols@d-and-d.com (DoN. Nichols)
wrote:

>In article <35dfbc5d.367963635@news.newsguy.com>,
>Reckless <reckless2x@home.com> wrote:
>>On 20 Aug 1998 01:19:19 -0400, dnichols@d-and-d.com (DoN. Nichols)
>>wrote:
>>
>>>	I've used a frame which had 4 foot UV fluorescents on both sides,
>>>with the PC board and negatives sandwitched between two flexible mylar
>>>films, with a vacuum pulled between them.  It was really nice for making
>>>double-sided PC boards.  We never tried for multi-layer, however, because
>>>that is a *lot* more work. :-)
>>
>>So how does one go about making the plated-through vias on a double
>>(or more) sided board in a home shop environment?  A lot of the PLCC
>>devices I would like to use need lots of traces and vias quickly
>>become necessary.  I'd rather not have to put in rivets or solder
>>wires through the board.
>
>	I've done both of the last two.  The process which I have heard
>described involves first some mask on the boards, drilling them on some
>really repeatable setup (ideally put the registration pin holes in there at
>the same time), then wash with something conductive enough to get a start at
>electroplating.  (Perhaps graphite through the holes.)  Connect to the foil
>on both sides, and electroplate while flowing the solution through the holes
>(or you will get rapid depletion of the chemicals there).  Electroless tin
>is sometimes a starting point, I think.  Once you have that starting point,
>you can electroplate at will.  (This, of course, requires you to be able to
>do good durable electroplating. :-)

This is not quite how the process works. For a double sided board you
place registration marks on the artwork and have your artwork on
over-size sheets. You place the two artworks together so that the
registration marks line up and then staple or tape the two pieces of
artwork together to form a "sandwich". Place the double sided PCB
inside the sandwich, tape so the PCB can't move, and expose both
sides.

As an aside it is easy to get mirror images of your circuit on one or
both sides of the board. For that reason it is good practice to have
some written material on the artwork AND check to see that you can
read it correctly BEFORE you expose the resist. For better resolution
you want the side of the artwork with the printed stuff in contact
with the PCB.

If you are lucky and have access to a machine for drilling holes you
digitize first using the artwork and then sit back with your cup of
coffee and watch the little drill punch out the holes for you. If you
are only doing one board you probably don't waste time digitizing but
drill manually with a drill press or with a machine designed for the
purpose. It can take as long or longer to digitize the artwork as it
does to drill one board.

Plated through holes are done by first sensitizing the inside of the
hole using stannous chloride which leaves a tad of divalent tin on the
surface. This is followed by a treatment with a palladium solution.
The adsorbed divalent tin reduces some palladium in solution to
palladium metal which sticks to the PCB (in the hole). The PCB then
goes into an electroless copper plating bath where the palladium
catalyzes the reduction of copper in the electroless plating
solutions. The result is a thin conducting layer of electroless copper
through the hole. Too thin to use as is. (Of course this plating
through is done before any of the copper on the circuit board is
exposed. i.e. the mask has not been developed)

Finally the board goes into a conventional plating bath to produce  a
thicker coat and make contact with the copper on the circuit. This is
a rough outline without the gory details.

All of the steps are being observed closely by Murphy and his laws
frequently apply.

>	Once that is done, strip off the mask, put on photoresist, (well,
>the photoresist could be the mask, if you were willing to do all the plating
>steps under controlled lighting conditions.)

Yes this can be done. The resists (masks) are usually UV sensitive and
the room where work is done uses yellow light.

>	Using the registration holes for alignment, expose the photoresist
>(or for a large enough production run with no really high resolution
>demands, silk-screen the resist on.)
>
>	Now -- if you're going to do more than double-sided, etch each side
>which is going together, sandwich with an insulator with the holes also
>pre-drilled, and plate as before,

For more than one board you drill registration holes and use pins to
correctly align the boards.


>>Another question way out there: what glue is used to laminate
>>multi-layer boards?
>
>	Very thin uncoated material the same as the board's substrates (such
>as G-10 glass-epoxy).  A matching epoxy to that used to make the boards
>themselves.  Clamp and cure.  For this, you probably only want the
>registration holes until you have the various pairs of layers glued
>together, then drill to expose the edges of the foil where needed, coat
>through, and plate as before.
>
>	(Listen to me -- I sound as though I've done this before.  Don't
>believe it. Anybody who really *knows* this stuff is probably under a
>non-disclosure agreement. :-)
>
>>                     Any real good tricks to registration?  My first
>>thought is to make the board blank larger than necessary and use pins
>>to register everything, then trim the board.
>
>	I would say that you should trim to size just after the registration
>holes are placed in the board, perhaps after the etching of the two sides,
>and *before* any clamping and gluing.  You may want to (gently) sand the
>edges after the stack is finished, but I think that shearing would be very
>likely to break the boards apart near the point of shear, destroying your
>nice plated-thrugh holes.

We sheared the board and then sanded. You don't want to handle freshly
sheared boards containing fiberglass. At least you won't want to do it
a second time.

As a disclaimer this is not my field. I have built a lot of electronic
stuff including satellite receivers BUT we had a well equipped shop to
make PCBs and much of the above information (or misinformation) was
gleaned during the process of shepherding "government" jobs through
the plating shop. The experts did this work, I watched and listened.

Index Home About Blog