Mike

Who is talking about rationale? This kind of test is a) nothing to do
with SIR: it is designed to provoke electrochemical migration phenomena
and b) next to useless. The real reason (?) behind it is that it assumes
users must do their measurements manually which is a laborious
procedure. The tests are done effectively once per day on weekdays. If
there is a significant drop from 96 to 168 hours, then it is assumed
something has gone wrong. However, a real knowledge of what is happening
from the real start of the test to the end can be gained only from very
frequent testing over the whole period and analysing the numbers. Every
15 minutes over the week would not be too frequent and four weeks would
be better than one. This implies automatic equiment.

To explain this further I quote after the signature from a recent
publication I wrote.

Best regards

Brian

These special qualification tests are necessarily very lengthy. They may
include actually running the finished equipment in climatic chambers
under the most extreme conditions of temperature, humidity and pressure,
sometimes with cycling. However, the most usual qualification tests are
Surface Insulation Resistance (SIR), Electrochemical Migration (ECM) and
Corrosion Testing. A great deal of confusion exists over these terms,
which have been incorrectly applied within some standards.

SIR testing consists of determining the insulating properties of the
substrate. A special interdigitated double comb pattern is etched onto a
test coupon, using the same materials as those used in practice. It is
then passed through the process, exactly as it would in production. It
is then placed in a climatic test chamber at fixed, usually severe,
conditions. Typically, these may vary from 35°C, 95% RH to 80°C, 80% RH
according to which standard is chosen. The resistance is measured
between the two combs at regular intervals of, say, 1 hour to 1 day,
with a low voltage applied for, say, 30 seconds before measurement over
a period of typically 28 to 56 days. The measured resistance is plotted
against time and the resultant curve analysed.  No voltage is applied
between the tests. This will detect deterioration of the substrate due
to the manufacturing processes (and that of the laminate itself).

ECM testing uses exactly the same test pattern and climatic conditions
as SIR testing, hence the cause for confusion. The only difference is
that a constant bias or polarisation voltage is applied to the test
pattern between the test. This voltage may lie between 5 V and 500 V and
may be of the same or the opposite polarity to the test voltage,
according to which specification is chosen. The function of this test is
to provoke electrochemical migration from residues or other causes. This
will manifest itself as a sudden and intermittent drop in resistance as
dendrites are formed and broken. Although some tests suggest a low
resistance measurement frequency of up to five tests per week over two
weeks, this would very likely miss the drop in resistance provoked by
dendrites. Current thinking is that the most reliable results would be
obtained by one measurement every 15 minutes over 10 – 15 days with a
minimum of 20 samples.

In a way, ECM testing is a hit-and-miss affair although the statistical
probability of missing at least a few events would be fairly low but not
entirely negligible. Corrosion testing was first developed by Dr David
Bono in France and further developed by others and does not depend on
detecting ephemeral phenomena. Basically, a very fine conductor is
etched on a printed circuit, between two other conductors. The circuit
passes through the process and placed in a climatic chamber, as in the
last two processes. A bias voltage of, say, 10 V is applied between the
fine conductor (anode) and the other two conductors (cathodes) to
provoke galvanic corrosion. The electrical resistance of the fine
conductor is measured at intervals: this will increase with the onset of
corrosion. A plot of resistance against time will give a measure of the
risk of electrical corrosion and the form of the curve will show whether
it is essentially “mousebite” or uniform corrosion or a mixture of both.
If corrosion becomes evident in the first ten days of a test (i.e. an
increase in resistance), then it would be preferable to continue the
test until the conductors are entirely broken through.


Michael Simms wrote:

>  Happy New Year to all, I've been reading IPC-9012, searching the
> technet archives, and looking high and low for the rationalebehind
> recording 24-hr SIR measurements and yet using only 96 and 168 hr
> values for pass-fail criteria.If someone out there was involved with
> drawuing up the IPC test protocol,perhaps you would be kind enough to
> elaborate on the reasons for recording/documenting the 24-hr values
> and yet using only 96 and 168 hr values for acceptability criteria.Any
> input here would be appreciated.Reagards,Mike SimmsTrace Laboratories
> - Central

##############################################################
TechNet Mail List provided as a free service by IPC using LISTSERV 1.8c
##############################################################
To subscribe/unsubscribe, send a message to [log in to unmask] with following text in
the body:
To subscribe:   SUBSCRIBE TECHNET <your full name>
To unsubscribe:   SIGNOFF TECHNET
##############################################################
Please visit IPC web site (http://www.ipc.org/html/forum.htm) for additional
information.
If you need assistance - contact Gayatri Sardeshpande at [log in to unmask] or
847-509-9700 ext.5365
##############################################################