Bill Barthel
Electronic Assembly Corporation
[log in to unmask]  asked the following:

"What convention is used for calculating the area of a populated assembly 
when determining bulk ionic contamination?  TM-650 states the area of the 
components should be considered.  I've heard of people using length X width 
X 3.  Is there an industry standard?"

First, thanks to Doug Pauls & Bill Kenyon for presenting the two basic sides
of the issue.

Second, as a real life user of bulf contamination equipment for 15 years and
defending it against auditors of every type, I try to address the following
questions.

If you used a low value like length x width with no factors, would you pass
10 micrograms NaCl equivalent/sq.in. without the equivalency factor.
If yes, use that formula.  This keeps happy a group of people who feel that
only those surfaces directly contaminated should be used in the
calculations.  (This is where Mr. Pauls is coming from, as a contamination
expert)  This also makes me happy, because by cleaning process is robust
from a cleaning results angle and I have less to worry about.

If you pass using two times the length x width, this will keep all but the
most diehard fans happy.

If you still don't pass, it is legitimate to refer to the Navy studies of
1972 & 1978 and use actual measurement.  I have performed detailed studies
of various types of pin-in-hole designs, using cylindrical and rectangular
solids, as referred to by Mr. Kenyon.  Set up the basic formulas in a spread
sheet program, using a line for each line in your bill of materials, taking
into acount the body, leads, and quantities.  It isn't as hard as it once was.
>From this type of study, I have seen dense assemblies exceed a factor of
4.5....  This is an approach you can take to make even marginal equipment
look robust, in an attempt to justify less testing.

Having said all of that, please contact the EMPF for a copy of their report
(TR-0013??) on bulk ionic contamination testers so you can understand the
limitations of the test to tell you anything meaningful.  Please only use
the test as a simple process control to warn you when things have gone to
H*** in a hand basket, and spend more time making sure that your cleaner is
operating as it was when you did your SIR testing (you did do some to verify
your alternative cleaning process, didn't you?!).

Finally, please understand the concept of expectancies.
If my process is robust, and cleans IPC-B-36 boards to above average SIR
levels, and another board cleaned in the same process at the same time gives
another value, one of two things must be true.  Either, due to design or due
to the soldering process parameters, I have left a detrimental amount of
dirt on my board, or that number is the right number.  If I expect a certain
number from a particularly complex board, then that number is not high or
low compared to other boards, but is at 100% of the expected value.  The
tough part is to establish a sound protocal for testing assemblies with
"high" readings to determine if field failures are likely to occur.  I'll
leave that discussion for later...

Hope this helps
Steve Mikell, [log in to unmask]
SCI Systems, Government Division