Документ взят из кэша поисковой машины. Адрес оригинального документа : http://www.stsci.edu/ops/swat/safing/01136/FailureAnalysis.txt Дата изменения: Tue May 29 18:56:03 2001 Дата индексирования: Sun Mar 2 04:33:01 2014 Кодировка: Поисковые слова: п п п п п п п п п п п п

From: "Rodger Doxsey"
Subject: FW: There is pure tin in STIS electronics

More re tin whiskers.


-----Original Message-----
From: Henning Leidecker [mailto:hleideck@pop500.gsfc.nasa.gov]
Sent: Friday, May 25, 2001 6:22 PM
To: Mitchell.L.Davis.1@gsfc.nasa.gov; Campbell, Dave; Sticka, Rob;
Fafaul, Bryan; Henning Leidecker; Randy Kimble; Phil Christon;
Larry.L.Dunham.1@gsfc.nasa.gov; Bacinski, John; Chiei, Roger; Lyle, Rob;
Paul Goudfrooij; Chris Long
Cc: pburch@hst.nasa.gov; dscheve@hst.nasa.gov; Mike Weiss; Leckrone, Dr.
Dave; Niedner, Dr. Mal; Carpenter, Dr. Ken; Cepollina, Frank;
mkienlen@hst.nasa.gov; Ruitberg, Ed; Wilkinson, Chris; doxsey@stsci.edu;
Rankin, Art; Vernacchio, Al; Walyus, Keith; Brian Vreeland; Kelley, Jim;
Clagett, Dick; Laura.J.Milam.1@gsfc.nasa.gov;
Bruce.E.Woodgate.1@gsfc.nasa.gov; Liqin.L.Wang.1@gsfc.nasa.gov
Subject: There is pure tin in STIS electronics


Gentles,

Conclusion:
==========
Based on observations made today, we now know that "tin whisker vacuum
arcing", caused by the growth of a tin whisker on one of the STIS relays
(or, possibly, on a nut or washer used on a filter), is a creditable
explanation for the present STIS adventure.

Consequence:
===========
IF the 20 A fuse was interrupted by a short circuit condition caused by a
tin whisker (growing across the bus) in the STIS, then it is highly likely
that the only residual effect is a blown 20 A fuse and a couple of
arc-scuffs at each end of the (former) whisker. We have to check for
possible latent damage to each current-carrying element in the cicuit. But
the fuse is designed to be the "weak link", and the evidence is that the 20
A fuse is indeed interrupted. It seems unlikely to me that this fuse would
have interrupted AND some other current-carrying element is on the verge of
interrupting and therefore has sustained damage. It seems more likely to
me that the fuse did its job and protected the other current-carrying
elements. Still, we must check each of these. It seems there is hope that
we can command C-bus on, and will then see the bus available at the output
of PDU2. And then command STIS on, and have it work.

Background:
==========
While he was at Ball Aerospace early this week, Richard Williams (6-8766)
requested a relay of the same lot date code (9234: 34th week of 1992) as
one of the "operate and hold" relays built into STIS. Phil Christon
provided him with one, and Richard Williams, Frederick Felt (6-9634), and I
examined it yesterday and this morning. A short examination using a
binocular microscope did not find any growths. However, the J-hooks had
the characteristic appearance of "bright pure tin".

SEM examination found many examples of "proper tin whiskers" growing from
the J-hooks. While these are all short (20 to 100 micrometers long), I can
report with confidence that they have precisely the appearance of whiskers
that have been seen in other studies to grow to as long as 9 millimeters.
A whisker as short as 1 mm could create a short of the type needed to
interrupt the 20 A fuse in the PDU2. That is the spacing between the
J-hook and the metal case of the relay.

Frederick Felt inspected the surface of the J-hook using energy dispersive
x-ray analysis of the surface of the J-hook, and found NO DETECTABLE LEAD.
A conservative estimate of the detection limit of the equipment he used is
1%.

Dr Liqin (Len) Wang (6-0962) has now completed examination of this relay.
He did not detect any lead at all in the J-hooks of this relay. Len's
equipment is more sensitive than that used by Frederick, and we can say
with confidence that the concentration of lead is substantially less than
1%. THE J-HOOKS OF THIS RELAY ARE PLATED WITH PURE TIN.

Pure tin surfaces, especially electroplated "bright" tin surfaces as these
apparently are, have been seen to grow tin whiskers of lengths up to 9 mm.
There is great variability among the tin whisker forests grown by tin
plated surfaces: it is not uncommon for some specimens from a given lot to
grow many long whiskers, while other specimens grow only a few short
whiskers. Therefore, I recommend that we concentrate on the finding that
the J-hooks are plated with pure tin, apparently with "brightener" added to
the plating bath, and the knowledge that such surfaces frequently grow long
tin whiskers, and the knowledge that these have caused spacecraft bus fuse
interruptions. I recommend that we do not attach importance to the finding
that this particular relay has not shown tin whiskers long enough to have
created the kind of short needed to interrupt the 20 A relay in the PDU2.

We know of nine cases in which tin whiskers growing from J-hooks plated
like the subject one have interrupted pairs of 10 A fuses within slightly
less than 2 ms, in a 50 V bus. Such whisker-based events would be able to
interrupt a 20 A fuse in a "28 V" (nominal) bus, but this would take longer
than 2 ms. I believe that a similar event would interrupt the 20 A fuse in
less than 100 ms, and is therefore potentially "invisible" to the telemetry
we have. I am working on a model for this time: my intent is to promote
this claim from a private belief to a public demonstration.

Therefore, "tin whisker vacuum arcing", caused by the growth of a tin
whisker on one of the STIS relays, appears to be a creditable explanation
for the present STIS adventure.

Additional item:
===============
Dr Liqin Wang has also inspected a feed-through filter also provided by
Ball to Richard Williams. This filter has the same lot date code as one
used in STIS. Len finds that this filter is plated with a leaded tin (2%
to 3% lead). Thus, it is improbable that any whiskers longer than some 4
to 5 mils (100 to 130 micrometers) long are present on the filter(s) in
STIS.

However, Len noted that the washer and the nut provided with this filter
are plated with pure tin: any lead was below the detection limit of Len's
equipment; therefore, its concentration is substantially less than 1%.
Stresses applied to nuts and to washers are well-known to launch tin
whiskers in abundance in pure tin, and have even produced some whisker-like
growths in leaded tin. Therefore, we can expect whiskers to form on the
nut and on the washer, and especially where either of these two items are
compressed against something. They can also form where the surfaces have
been compressed during the installation and final torquing of these two
items. I do not know the distance between the nut & washer and any exposed
conductor carrying a bus current (relative to the nut and washer). I do
not know if the nut and the washer are grounded. But if a whisker up to 9
mm long growing from the nut or the washer *could* short the bus, then we
must regard this too as a possible explanation for the present STIS
adventure.

Sincerely,
Henning Leidecker
301-286-9180