Archive of the former Yahoo!Groups mailing list: Homebrew PCBs

I have not quite made the jump to SMD's. I am still toiling with through
hole parts. Using pcb in a box toner transfer method has worked wonders for
the quality and repeatability of my boards. Here is what I think is most
important about the toner transfer method:

1. I prefer to use really fine grit sand paper to clear initial oxidation,
then iso alcohol, then tarnix, then alcohol once again, all with a coffee
filter to reduce paper lint.
2. I have switched to the thinner 1/32" pcb's with 1oz copper. 1/2 oz is
easier to etch, but for some reason I could only get my hands on the 1oz
copper boards from Mouser. It still works fine. I really prefer the thicker
boards, but not sure if my laminator will handle them without modification.
I swear the thinner boards with break under the wrong screw torquing OR the
pcb traces with torque right off the board. Maybe through hole is not a good
idea with these thin boards?
3. I can't seem to etch lines that are really close to the edge, the toner
tends to flake really easily there.
4. I green TRF foils tend to remove some of my traces when removing the
foils back off. This is most likely due to lack of good toner curing. I need
to work on this

Here is my real question about through hole parts and boards. There are a
lot of parts the have flat through hole leads. Of course your drill bit
yields a round hole. Even further, some pads intended for those flat (or
square pins) are really thin (like 12 mils), such as the .1" pin headers
used a lot on dev and breakout boards. The problem I am having is getting a
good solid connection to the board with these parts. The pins always seem
very precariously attached to the pcb since the pins are attached by so
little copper on the board and the different shape between the hole and the
pin. Do you guys glue the plastic headers down? Is this a good idea? What
about the different shapes? Is the solder supposed to bridge the gap there
or is there another way? I know solder really is not a structural component,
but without some bonding agent, it really is, right?


thanks,
Chris


[Non-text portions of this message have been removed]