Subject: | |
From: | |
Date: | Thu, 27 Jun 1996 11:42:00 EDT |
Content-Type: | text/plain |
Parts/Attachments: |
|
|
\0
TO: I4235700 IBMMAIL new address for ipc technet 25.6.96
FROM: DSTEWART EX2 D.Stewart - Product Development Manager.
DATE: 27 June 1996
SUBJECT: re-odd number layers
REFERENCE: cons!
There is another issue with odd layer counts in volume fabrication,
which is logistical more than anything else:
We build odd layer count boards with foil on side 1 and laminate
for side n / n-1. We purchase copper foil as a stock item,
pretooled, which includes an offset tooling hole. As you can
imagine we end up with a large quantity of "left shoes".
More technically, the outside core is oxidised prior to lamination,
whereas the foil is clean copper. Resin sticks like the proverbial
to black oxide (as it is designed to) so there also tends to be an
issue on resin spots/short circuits on this side, compared to a
total foil build.
It has already been stated about warpage, so I need not add more,
except to say that we were once asked to make some PCMCIA type
cards which were blind via - sequentially bonded - drilled L1-2,
plated and bonded with foil, drilled L1-3 plated and bonded with
foil, and finally drill L1-4. The fact that it started as L1/2 on a
007" core, and the prepreg at each stage of bonding was 2 plies of
2125, combined with plated 25um of copper on each internal layer,
meant that by the time the panel came to be drilled finally, it had
to be taped down on all four sides to keep it flat enough to stay
on the tooling pins! At the final outer photomech stage we had to
revert to using double sided tape to hold the artwork to the panel
until the vacuum pulled it flat! SO please beware of assymetry.
Dougal Stewart
Product Development manager
Exacta Circuits
Scotland
|
|
|