Tin whisker problems????

[qrk]

Had a silly problem with a friends computer. Hard drive wasn't
recognized by the bios in two computers. After inspecting for bad
solder joints, I brushed all the IC leads with a toothbrush. Presto,
drive is recognized by the bios. Unfortunately, the boot partition FAT
was munched, but, was able to recover most of the valuable data. Drive
is still working 2 days later.

I want to say this is a tin whisker issue, but it could have been a
bum joint too. Was Hitachi (IBM) using lead-free in mid 2005?

Mark

 

[BobW]

Had a silly problem with a friends computer. Hard drive wasn't
recognized by the bios in two computers. After inspecting for bad
solder joints, I brushed all the IC leads with a toothbrush. Presto,
drive is recognized by the bios. Unfortunately, the boot partition FAT
was munched, but, was able to recover most of the valuable data. Drive
is still working 2 days later.

I want to say this is a tin whisker issue, but it could have been a
bum joint too. Was Hitachi (IBM) using lead-free in mid 2005?

Mark

It's very unlikely that they were lead-free in '05. It was probably just
some tartar or plaque.

Bob

 

[Joerg]

Had a silly problem with a friends computer. Hard drive wasn't
recognized by the bios in two computers. After inspecting for bad
solder joints, I brushed all the IC leads with a toothbrush. Presto,
drive is recognized by the bios. Unfortunately, the boot partition FAT
was munched, but, was able to recover most of the valuable data. Drive
is still working 2 days later.

I want to say this is a tin whisker issue, but it could have been a
bum joint too. Was Hitachi (IBM) using lead-free in mid 2005?

Could have been, hard to say for individual divisons:

http://www.my-esm.com/print/showArticle.jhtml?articleID=171201497

If you still have access to that drive: Are the solder joints looking a
bit dull? Does the solder have a hard time melting at normal tip temps?
That would point to lead-free. Then it might make sense to check it for
more whiskers under a microscope

 

[Martin Griffith]

Had a silly problem with a friends computer. Hard drive wasn't
recognized by the bios in two computers. After inspecting for bad
solder joints, I brushed all the IC leads with a toothbrush. Presto,
drive is recognized by the bios. Unfortunately, the boot partition FAT
was munched, but, was able to recover most of the valuable data. Drive
is still working 2 days later.

I want to say this is a tin whisker issue, but it could have been a
bum joint too. Was Hitachi (IBM) using lead-free in mid 2005?

Mark

I keep seeing the occasional, same, report about satellites failing
due to tin whiskers etc. ( nope, no url)

I was just wondering how "they" know, did they send a guy up to have a
look at the pcb's?


Martin

 

[Glen Walpert]

I keep seeing the occasional, same, report about satellites failing
due to tin whiskers etc. ( nope, no url)

I was just wondering how "they" know, did they send a guy up to have a
look at the pcb's?

Next time the occasional same report shows up read it. If it is the
same one I have occasionally read they duplicated the problem on the
ground, and merely ass-u-me-d that the causes were the same.

But of course the wiskers grew on tin plating, not lead free solder.
The issue is not and has never been with whiskers growing from lead
free solder. There have never been any documented failures from
whiskers growing on the SAC or better lead free solders. The issue is
and has always been whiskers from the tin plating on component leads,
sometimes used instead of the old tin-lead plating. Whiskers cannot
grow if the tin plating is completely wetted by solder; either
tin-lead or lead free. Using tin-lead instead of lead free solder
with components having tin plating offers absolutely no protection
against whiskers growing from the tin plating on component leads not
completely wetted by solder. If you want immunity from tin whiskers
either insure all tin plating is entirely wetted with solder or don't
use components with tin plated leads.

All of the major manufacturers figured this out long ago, which is why
we aren't seeing the massive failures some on this NG have predicted.
So I agree with the poster who speculated that the problem was more
likely tartar and plaque than tin whiskers .

 

[Jeff Liebermann]

Had a silly problem with a friends computer. Hard drive wasn't
recognized by the bios in two computers. After inspecting for bad
solder joints, I brushed all the IC leads with a toothbrush. Presto,
drive is recognized by the bios. Unfortunately, the boot partition FAT
was munched, but, was able to recover most of the valuable data. Drive
is still working 2 days later.

I want to say this is a tin whisker issue, but it could have been a
bum joint too. Was Hitachi (IBM) using lead-free in mid 2005?

I don't think so. IBM/Hitachi was shipping some really awful disk
drives from about 2000 to 2005. There was a class action against
Hitachi on the 75GXP series:
<http://www.berdonclaims.com/cases/details.asp?CaseID=173>
<http://en.wikipedia.org/wiki/Hitachi_Deskstar>
<http://www.astro.ufl.edu/~ken/crash/index.html> (autopsy)

They symptoms were somewhat similar to what you experienced. The boot
loader would fail to load from the hard disk because of read errors.
The drive would either not be recognized or not boot depending on
which point in the boot load process was failing. What was weird was
that removing the drive and retesting would sometimes induce a
miraculous recovery. I usually immediately made an image backup at
this point as it always failed later.

Incidentally, I just recycled about 40 IBM/Hitachi Deskstar drives
(mostly IDE) which is about a 2 year supply of failures. Disassembling
some of the drives showed the same failure as the aformentioned
autopsy.

 

[PhattyMo]

I don't think so. IBM/Hitachi was shipping some really awful disk
drives from about 2000 to 2005. There was a class action against
Hitachi on the 75GXP series:
<http://www.berdonclaims.com/cases/details.asp?CaseID=173>
<http://en.wikipedia.org/wiki/Hitachi_Deskstar>
<http://www.astro.ufl.edu/~ken/crash/index.html> (autopsy)

They symptoms were somewhat similar to what you experienced. The boot
loader would fail to load from the hard disk because of read errors.
The drive would either not be recognized or not boot depending on
which point in the boot load process was failing. What was weird was
that removing the drive and retesting would sometimes induce a
miraculous recovery. I usually immediately made an image backup at
this point as it always failed later.

Incidentally, I just recycled about 40 IBM/Hitachi Deskstar

You mean "DeathStar" drives...they've earned themselves a nickname
because of these issues.

 

[Robert Baer]

Could have been, hard to say for individual divisons:

http://www.my-esm.com/print/showArticle.jhtml?articleID=171201497

If you still have access to that drive: Are the solder joints looking a
bit dull? Does the solder have a hard time melting at normal tip temps?
That would point to lead-free. Then it might make sense to check it for
more whiskers under a microscope

"Dull"????
Lead free tin/silver solder is rather shiny and neet looking...

 

[Joerg]

"Dull"????
Lead free tin/silver solder is rather shiny and neet looking...

The hand-soldering at one client was shiny. However, all the industrial
lead-free stuff I've seen so far was dull and had failure rates several
times higher than non-RoHS. When it was for the lab I just re-soldered
the whole board with leaded and the failures went away.

 

[qrk]

I don't think so. IBM/Hitachi was shipping some really awful disk
drives from about 2000 to 2005. There was a class action against
Hitachi on the 75GXP series:
<http://www.berdonclaims.com/cases/details.asp?CaseID=173>
<http://en.wikipedia.org/wiki/Hitachi_Deskstar>
<http://www.astro.ufl.edu/~ken/crash/index.html> (autopsy)

They symptoms were somewhat similar to what you experienced. The boot
loader would fail to load from the hard disk because of read errors.
The drive would either not be recognized or not boot depending on
which point in the boot load process was failing. What was weird was
that removing the drive and retesting would sometimes induce a
miraculous recovery. I usually immediately made an image backup at
this point as it always failed later.

Incidentally, I just recycled about 40 IBM/Hitachi Deskstar drives
(mostly IDE) which is about a 2 year supply of failures. Disassembling
some of the drives showed the same failure as the aformentioned
autopsy.

I think I'll buy your explaination. Although, this particular drive
wasn't recognized by the BIOS. All the other failed IBM drives I've
came across were recognized by the BIOS, but, failed to boot due to
corruption or some other read problem. I've seen IBM drives die within
1 week in the early 2000s. Then I ran into Maxtor drives going bad
shortly thereafter. Now I'm using Samsung. See how long it takes for
them to go into early failure mode.

Mark

 

[JosephKK]

Martin Griffith mart_in_medina@ya___.es posted to
sci.electronics.design:

I keep seeing the occasional, same, report about satellites failing
due to tin whiskers etc. ( nope, no url)

I was just wondering how "they" know, did they send a guy up to have
a look at the pcb's?


Martin

I actually doubt that you will see this issue in satellites. The
reliability of expensive, place it and use it until it dies
technologies do require avoiding and preventing any failure mode
preventable or avoidable. The non-lead solder issues have been known
since the 1960's or earlier.

 

[JosephKK]

qrk [email protected] posted to sci.electronics.design:

I think I'll buy your explaination. Although, this particular drive
wasn't recognized by the BIOS. All the other failed IBM drives I've
came across were recognized by the BIOS, but, failed to boot due to
corruption or some other read problem. I've seen IBM drives die
within 1 week in the early 2000s. Then I ran into Maxtor drives
going bad shortly thereafter. Now I'm using Samsung. See how long it
takes for them to go into early failure mode.

Mark

Speaking of disk drives and early failure, i bought a couple of
Maxtor "One-touch" fire-wire/USB drives a few months back. I was
using on as an on line drive for eventual migration use. It failed,
i decided to piss off the warranty and took apart the case, and found
a normal Seagate drive inside. I placed it in another case it the
drive proper was fine. I tested the mini-brick power supply and it
was fine. I checked the the internal PS to the external enclosure
and it is dead. There is no fan in the one touch enclosure. Make
any conclusions you like.

 

[Nico Coesel]

I think I'll buy your explaination. Although, this particular drive
wasn't recognized by the BIOS. All the other failed IBM drives I've
came across were recognized by the BIOS, but, failed to boot due to
corruption or some other read problem. I've seen IBM drives die within
1 week in the early 2000s. Then I ran into Maxtor drives going bad
shortly thereafter. Now I'm using Samsung. See how long it takes for
them to go into early failure mode.

They don't make 'm as good as they used to be. I had to replace the
Western Digital hard drive in one of my servers after 11 years due to
a power outage yesterday.

 

[Baron]

Speaking of disk drives and early failure, i bought a couple of
Maxtor "One-touch" fire-wire/USB drives a few months back. I was
using on as an on line drive for eventual migration use. It failed,
i decided to piss off the warranty and took apart the case, and found
a normal Seagate drive inside. I placed it in another case it the
drive proper was fine. I tested the mini-brick power supply and it
was fine. I checked the the internal PS to the external enclosure
and it is dead. There is no fan in the one touch enclosure. Make
any conclusions you like.

I had one of those die in exactly the same way. In my case I had to get
the data off it. I did similar to you stuck the drive in a machine
backed it up then wiped it. But I put it back in the case and RMA'd
it. The new one works just fine.

 

[Glen Walpert]

The hand-soldering at one client was shiny. However, all the industrial
lead-free stuff I've seen so far was dull and had failure rates several
times higher than non-RoHS. When it was for the lab I just re-soldered
the whole board with leaded and the failures went away.

That is the typical result everyone gets when they try to "drop-in"
SAC lead-free solder in a tin-lead process, merely setting the
temperature up for the higher melting point. But few manufacturers
actually ship anything like that. Look at the lead-free solder joints
in a new cell phone or a new PC motherboard. That is what a properly
done lead free solder joint should look and perform like.

In general, all new soldering equipment is required to handle the
tight profiles and nitrogen inerting required for good results with
the low cost SAC alloy; those stuck with older equipment can only get
good results with some of the relatively expensive 4 component
proprietary solders, and then only if it is good older equipment
capable of accurate time/temp control. Only some fluxes will produce
good results in air. A bit of a pisser for the small manufacturer,
but not an excuse for shipping boards with bad solder joints. The
procedures required to obtain good lead-free solder joints have been
published for anyone to read and follow. I can concieve of no valid
excuse for the incredibly poor workmanship you report, and completely
reject the notion that lead-free solder is in any way to blame. If it
were then all new cell phones and computers would be unreliable, and
they aren't.

 

[Jeff Liebermann]

I actually doubt that you will see this issue in satellites. The
reliability of expensive, place it and use it until it dies
technologies do require avoiding and preventing any failure mode
preventable or avoidable. The non-lead solder issues have been known
since the 1960's or earlier.

I happen to be catching up on my magazine reading today and found in
the Sept/Oct issue of Military Embedded Systems magazine:
"The costs of doing business in the new RoHS World"
<http://www.mil-embedded.com/pdfs/Aitech.Sep07.pdf>

The article notes that 3 satellites have probably failed due to tin
whiskers. Actually, the list in the article is incomplete. See:
<http://nepp.nasa.gov/whisker/failures/index.htm>
Seems to be a problem with the early Boeing HS601 series of birds:
<http://www.sat-index.com/failures/hs601.html>

More from the Mil Embedded Sys article:

Matte (dull) tin seems to be more effective at preventing whiskers
than shinny tin.

Conformal coating doesn't help.

TI estimates that their conversion to RoHS will save the equivalent of
about 10 automobile batteries in lead.

Ford had some automobile computahs fail with trace corrosion when the
tin traces came in contact with modelling clay, which had a high
sulfur content.

There have been nuclear and conventional power plant shutdowns due to
tin whiskers.

 

[Jeff Liebermann]

The hand-soldering at one client was shiny. However, all the industrial
lead-free stuff I've seen so far was dull and had failure rates several
times higher than non-RoHS. When it was for the lab I just re-soldered
the whole board with leaded and the failures went away.

Umm... It's suppose to be dull to reduce whisker growth. See:
<http://www.aciusa.org/leadfree/leadfree_verdi-11-5-04.htm>
"It is well documented that internal compressive stress is
a driving force for the growth of tin whiskers. This stress
can come from the naturally occurring compressive stress
that occurs when tin is electrodeposited from a plating bath.
The more organic brightening additions in the plating bath,
the higher the internal stress in the tin deposit will be.
The highest stressed deposits are obtained from baths that
yield bright or specular (mirror-like) deposits, often used
for decorative electroplates. Somewhat less prone to whiskers
is the matte or dull finished tin, which is usually lower
stressed, but still will grow (usually fewer and/or shorter)
whiskers."

 

[JosephKK]

Jeff Liebermann [email protected] posted to sci.electronics.design:

I happen to be catching up on my magazine reading today and found in
the Sept/Oct issue of Military Embedded Systems magazine:
"The costs of doing business in the new RoHS World"
<http://www.mil-embedded.com/pdfs/Aitech.Sep07.pdf>

The article notes that 3 satellites have probably failed due to tin
whiskers. Actually, the list in the article is incomplete. See:
<http://nepp.nasa.gov/whisker/failures/index.htm>
Seems to be a problem with the early Boeing HS601 series of birds:
<http://www.sat-index.com/failures/hs601.html>

More from the Mil Embedded Sys article:

Matte (dull) tin seems to be more effective at preventing whiskers
than shinny tin.

Conformal coating doesn't help.

TI estimates that their conversion to RoHS will save the equivalent
of about 10 automobile batteries in lead.

Ford had some automobile computahs fail with trace corrosion when
the tin traces came in contact with modelling clay, which had a high
sulfur content.

There have been nuclear and conventional power plant shutdowns due
to tin whiskers.

Thank you for your very informative post. I never knew that HSC and
apparently most of Hughes S&CG seems to have been bought by Boeing.

 

[Joerg]

Umm... It's suppose to be dull to reduce whisker growth. See:
<http://www.aciusa.org/leadfree/leadfree_verdi-11-5-04.htm>
"It is well documented that internal compressive stress is
a driving force for the growth of tin whiskers. This stress
can come from the naturally occurring compressive stress
that occurs when tin is electrodeposited from a plating bath.
The more organic brightening additions in the plating bath,
the higher the internal stress in the tin deposit will be.
The highest stressed deposits are obtained from baths that
yield bright or specular (mirror-like) deposits, often used
for decorative electroplates. Somewhat less prone to whiskers
is the matte or dull finished tin, which is usually lower
stressed, but still will grow (usually fewer and/or shorter)
whiskers."

Yes, it's a new field. Yet the Eurocrats took it upon themselves to
declare that "it's going to be ok". We'll see. I have yet to see a
longterm study about lead-free, like some experiment that was started,
say, in the 60's. Anything else is playing with fire.

 

[Jeff Liebermann]

Jeff Liebermann [email protected] posted to sci.electronics.design:

Thank you for your very informative post. I never knew that HSC and
apparently most of Hughes S&CG seems to have been bought by Boeing.

What I found interesting is that the failure ocurred in relays.
<http://www.sat-index.com/failures/hs601.html>
which are apparently used for on-off switching. I would have expected
whisker bridging in ciruit board traces and tin plated connectors, but
not a mechanical relay. The article mumbles that the tin whisker
shorts the relay to the case, causing a fuse to blow. That means that
either the tin whisker is huge, or the fuse is really tiny. I would
imagine that the very thin tin whisker would blow long before any fuse
would blow. Methinks something is fishy here.

Boeing bought the Hughes Satellite biz in Jan 2000 for $3.75 billion
in cash. Much of the cash went to prop up PanAmSat and pay for
DirecTV aquisitions.