Homebrew PCBs

photos in
http://tech.ph.groups.yahoo.com/group/Homebrew_PCBs/photos/browse/b888


My first attempt at a double sided board using smt and PLCC hardward.
Using a max resolution of .010" traces with .010" between them I
managed (using PressNPeel) to successfully create this board. The
technique I used to locate the top transfer to the bottom is outlined
below

1 I placed pads with a 1mm dia hole on the four corners of both the
top and bottom tranfer images.

2 Drill out the corner holes of the PnP tranfers both top and bottom

3 Ironed on the top transfer

4 drill out the four corner holes on the top PCP

5 Using Push Pins through the drilled holes on the pcb I Located the
bottom tranfer to the pcb

6 Tack ironed the bottom transfer to the pcb 

7 Remove the Push pins and finished the transfer to the bottom of the PCB

Using the technique I manages to align both bottom and top images by
with .020" not quite close enough but by making sure that anywhere I
was going thru hole I did not have any traces closer that .020" all of
the thru holes lined up close enough to solder.

In the process I discovered that as you tack the PNP it will slightly
stretch in the direction you are ironing and the PNP will appear
wrinkled so if you tack both the bottom and top PNP transfers in the
same directing trying to duplicate the exact method of transfer both
top and bottom you will achive better result than the .5mm error I had
on this example. Just remember to leave the push pins in for location
while tacking then remove the Push pins and finish the tonal transfer


One thing I did on the development board was to use 74ls04's to drive
leds on each channel of the PIC. This only places a .5ma load on each
channel of the pic while giving a visual indicator when any port pin
is turned on. It really lets you see whats going on on all ports at
any time! It also allows all the channels to drive other devices
without the addition load of driveind a led. I also included a LM34
temp probe, vref+, vfer- pots, and a 32.768khz osc that can be
disabled with jumpers. Also Soic pads for a 24C256 not in place yet.

Shown is a LCD displaying time and temperature using timer1 with the
32.768 crystal and the LM34Z probe

Regards
Dennis Waggoner

On Sat, 24 Nov 2007, Dennis Waggoner wrote:

> My first attempt at a double sided board using smt and PLCC hardward.
> Using a max resolution of .010" traces with .010" between them I
> managed (using PressNPeel) to successfully create this board. The
> technique I used to locate the top transfer to the bottom is outlined
> below

I've never used Press-n-Peel (always used inkjet photo paper for toner
transfer) so I don't know how much light it lets through - but I've found
that optically lining up the sides (i.e. a bright light behind the
transfer sheets, face in - using vias to make sure it's lined up), taping
the ends of the two sheets to form a hinge, then putting the board between
and just ironing the whole lot (turning over halfway through to iron the
other side - the heat transferring through the board gets the other side
to basically stick well enough for turning the board over) is extremely
accurate. I usually get it absolutely spot on with this method - most the
time, after etching and holding the board up to the light, there is no
detectable mismatch between the two sides.

It's really fast to do as well.

A light box is probably the easiest way to do it, but during daytime, I
just use sunlight and at night, a table lamp.