Magnus and the group,
in hi-rel projects and in our space and satellite projects we have
procedures to remove at least most of the gold cover eg. by
presoldering the contact areas and then sucking both, tin and gold,
away. Properly done the clean areas can then undergo the final soldering
process. So far all our satellites do work very well and years longer
than planned. I do not remember having had in 25 years ever real problems
with good prepared solder joints. Of course, the best is avoiding gold
contacts for solder joints at all and never use nickel plated surfaces for
solder joints!
The NASA does have similar special procedures as the ESA.
The best is using adequate crimping processes where possible.
regards,

Arnold

On Tue, 24 Jul 2007 04:36:40 +0200 (CEST), Magnus Danielson wrote:

Hi Magnus,

The 5372A which we bought new years ago reference polluted by the
switching power supplies. There were more of them with different
frequencies. It showed up in the fft of measurements and also seen at
the reference output. I do not have access to that instrument anymore
and do not know the serial number. How is this with your 5372A?

Henk

On Jul 23, 2007, at 2:36, Magnus Danielson wrote:

Gold mixes with solder to from an "intermetallic" that
suffers from "embrittlement".  I am not aware of silver
solder being a remedy.  I have heard of assembly lines
that do not allow any gold whatsoever in the building!
I heard our assembly line manager at work just
last week lecturing yet another engineer about NO GOLD.
Period!

Furthermore, gold plated brass connectors screwed into
aluminum (with corrosion treatment) (no soldering) will
corrode in a salt spray environment.  I guarantee it.
We always used stainless for military work.

Rick N6RK

Brooke Clarke wrote:

See:

http://journals.cambridge.org/download.php?file=%2FMAM%2FMAM11_S02%2FS1431927605504483a.pdf&code=7eb044ef7101eff8a8274c4fdfabca59

and

http://www.tkb-4u.com/articles/soldering/sgons/sgons.php

There is also at least one US patent deals with using pure tin
instead of gold for solderability. The problems are also well known
to anyone working with flip-chip bonding where gold bumps have been
used. Both tin and gold have been known to grow crystalline whiskers,
one microscopy journal carried a paper a couple of years ago from
Sandia Lab on the subject. Tin whiskering has been well known for many years.

John

At 11:21 AM 8/30/2007, Richard (Rick) Karlquist wrote:

That sounds totally unjustified. It's not like gold is going
to jump at the solder joint and contaminate it cold.

The effect of gold contamination is to weaken the joint (the
intermetallic compound of gold and tin is hard and brittle).
Over time, and particularly in high stress environments
(temperature cycling, vibrations), the solder joint
eventually breaks down at the interface and leaves you
with an intermittent connection.

The way we deal with it in the military business (where
temperature cycling and vibration is a way of life) is that
we first remove the gold plating from the solder pin side of
connector pins by tinning the pin a couple of times and each
time sucking the solder with fresh braid or other tool.
That dissolves the gold in solder that is then discarded,
leaving the solder end of the pin free of gold. At that point,
the pin can be soldered normally and the solder joint will show
good characteristics.

Alternately, we use a solder pot and dip the end of the part
that is to be cleaned in it. The problem with that is that
the concentration of gold in the pot increases over time until
the content of the pot has to be discarded and replaced, so
if you try to use the pot for too long, you may not get rid
of the gold completely.

Our manufacturing engineering people have developped rules
for how many times you can use the pot before replacement,
and that seems to work as long as it is controlled.

Didier KO4BB

From: "Didier Juges" didier@cox.net
Subject: Re: [time-nuts] HP 5370B low frequency modulation
Date: Thu, 30 Aug 2007 18:10:56 -0500
Message-ID: 061301c7eb5b$063275b0$6501a8c0@didierhp

Even these measures can be argued as overdoing it, even if I understand why
they run popular. The key point is that it is only when the gold concentration
in the solder joint is too high that you start to get these problems. For very
low levels of concentration there is no problem. The issues is really about
how much gold concentration do you end up with and do you have sufficient
control over amount of gold, amont of solder and amount of dissolvement time
for gold to distribute in the tin. These are process control variables.

Cheers,
Magnus

When you solder a connector pin that is gold plated, the concentration of
gold in the immediate vicinity of the interface is maximum, since you go
from pure gold to pure tin/lead and you have all the gradient in-between
because the gold diffuses in the tin, so you will always have an area where
the compound will have the right concentration to cause problem. I have seen
many solder joints contaminated this way and experienced the problem first
hand, it's not just a theoretical problem. Yet, it's quite preventable using
either of these simple methods. As soon as you remove the excess gold, the
concentration drops below the level where it can cause problems and you will
be OK.

I am not aware of a method that does not entail removing the excess gold?

Didier KO4BB

My experience is that thermal cycling and vibration exacerbate the problem
very greatly. We have been making both commercial and military products for
a long time, and we started seeing the problem in military products much
before we ever saw it in commercial products.

The intermetallic compound is conductive, so unless you have some mechanical
effect that pushes the conductors apart, the contact will not necessarily be
lost.

Look closely at a PWB that is gold plated (not just the plug-in contact
area, but some older PWBs were entirely gold plated). Look at the interface
where the solder joint meets the gold. On anything that is more than a few
years old, you are likely to see a darker grey ring where the two metals
meet. There you have it. You don't see that on a new board. Remove the
solder from one of these joints with solder wick or a desoldering iron,
clean real good with a small brush and redo the solder joint using a good
cleaning flux, and the solder joint will look normal.

Didier KO4BB