Soldering in space holds surprises

rk wrote:

Soldering in space holds surprises


http://www.aiaa.org/aerospace/images...ookmarch05.pdf

Very interesting. It shows that we need to be sure that we really
understand even mundane (that's not the right word, I'm sure)
things in space. It also suggests something useful to do with
ISS in the years ahead.

Here's a couple of excerpts from the story:

Soldering in space holds surprises
AEROSPACE AMERICA/MARCH 2005
Edward D. Flinn

[EXCERPTS]


There is nothing routine about working in space, as
astronaut Mike Fincke found out when he did some
soldering onboard the International Space Station (ISS).

Richard Grugel, a materials scientist at NASA Marshall,
watched his video monitor in disbelief. A transmission from
the ISS was playing. The scene: Astronaut Fincke touches the
tip of a soldering iron to a wire wrapped with rosin-core
solder. The solder, heated, becomes a molten blob with a
droplet of rosin clinging tightly to the outside. As the
solder melts, it is the behavior of the rosin that amazes
Grugel. As the temperature increases, the droplet begins to
spin, round and round, faster and faster, like a miniature
carnival ride.

"What a surprise," says Grugel. "I have never seen anything
quite like it."

Grugel is the principal investigator of the In-Space Soldering
Investigation (ISSI). Fincke was conducting the ISSI program
at the time of the discovery. The program's purpose is to find
out how solder behaves in a weightless environment. This is
important information for astronauts: If something breaks
during a long trip to Mars, they will likely reach for a
soldering iron to repair it.

[snip]

Many of the methods used to build and repair equipment on Earth
must now be adapted for space. ISS crews have tools for making
small repairs, but little research has been conducted on the
best ways to manufacture and repair equipment in space.

In the low-gravity environment inside the orbiting space station,
surface tension influences materials and fluids more strongly
than it does on Earth. Fluids that would splatter and spill to
the ground on Earth form drops held together by surface tension.
Left uncontained, these drops float through the air in space.
Convection and surface tension are two forces that influence how
a fluid moves or flows. Both play roles in fabrication and repair
techniques such as soldering and welding. Soldering involves
melting a metal or metal alloy, usually lead or tin. The molten
material is applied and flows between surfaces or joints of
materials that are being held together. When it cools and
solidifies, the solder joins the materials together. On orbit,
however, gravity is balanced by the equal but opposite centrifugal
force or orbital rotation, thus eliminating convection and leaving
surface tension to be the dominant force influencing flow.

[snip]

http://www.aiaa.org/aerospace/images...ookmarch05.pdf

Hmm. I guess some day I'll need to try soldering upside down on
earth. My first reaction is that gravity plays little role even here,
but I guess I don't know that until I make sure that gravity is
working against me, rather than with me, eh?

The items in the article range from the subtle (gas bubbles in the
solder) to the (should be) obvious (need to blow/suck the smoke away
so you can see what you're doing). Interesting.

"Jim Kingdon" wrote in message
news

http://www.aiaa.org/aerospace/images...ookmarch05.pdf

Hmm. I guess some day I'll need to try soldering upside down on
earth. My first reaction is that gravity plays little role even here,
but I guess I don't know that until I make sure that gravity is
working against me, rather than with me, eh?

I've done this a bit in the past (generally solidering connections under the
dash of a car or something similar) and it's generally not fun. Any excess
solder, when it's liquid, tends to flow downward and drip off. If you try
this, I'd recommend gloves, and a full face shield (if you've got one).

The way the solder flows in zero gravity is certainly someting interesting
to study.

The items in the article range from the subtle (gas bubbles in the
solder) to the (should be) obvious (need to blow/suck the smoke away
so you can see what you're doing). Interesting.

But at least it's possible to solider in zero gravity. There don't appear
to be any show stoppers.

Jeff
--
Remove icky phrase from email address to get a valid address.

rk wrote:

Yeah, and it will perhaps get to be even more interesting, as the European and
Japanese requirements for lead-free soldering take affect; this will also
affect component availability. Along with making the soldering just
different, for long duration equipment, as is often found in many aerospace
applications, the tin whisker problem is of concern to some.

--
rk, Just an OldEngineer


Considering that most of the electronics will be built with surface
mount components it will be very difficult to repair, let alone
troubleshoot the complex equipment being used these days.

I worked on Telemetry receiving equipment and did my own rework on
chips up to 288 pins under a stereo microscope. It takes time to learn
how to probe these boards by hand, and requires a very steady hand to
resolder the leads without causing even more problems. Processes that
can be used on the ground would introduce unwanted contaminants in a
closed environment and you don't have the luxury of a cleaning room to
run repaired boards through to remove the flux after a repair.

--
Former professional electron wrangler.

Michael A. Terrell
Central Florida

Jim Kingdon wrote:

The items in the article range from the subtle (gas bubbles in the
solder) to the (should be) obvious (need to blow/suck the smoke away
so you can see what you're doing). Interesting.

I was fascinated to learn that in orbit, "gravity is balanced by the
equal but opposite centrifugal force".

--
Happy, sad, cross and concentrating.

rk wrote:

Michael A. Terrell wrote:

rk wrote:

Yeah, and it will perhaps get to be even more interesting, as the
European and Japanese requirements for lead-free soldering take affect;
this will also affect component availability. Along with making the
soldering just different, for long duration equipment, as is often found
in many aerospace applications, the tin whisker problem is of concern to
some.

--
rk, Just an OldEngineer


Considering that most of the electronics will be built with surface
mount components it will be very difficult to repair, let alone
troubleshoot the complex equipment being used these days.

I worked on Telemetry receiving equipment and did my own rework on
chips up to 288 pins under a stereo microscope. It takes time to learn
how to probe these boards by hand, and requires a very steady hand to
resolder the leads without causing even more problems. Processes that
can be used on the ground would introduce unwanted contaminants in a
closed environment and you don't have the luxury of a cleaning room to
run repaired boards through to remove the flux after a repair.

Well, yes and know. Some surface mount components aren't that hard, such as
resistors and capacitors and even some basic flat packs. Other large
integrated circuits, as you mention have a lot pins, some quite a bit more
than what you mention, but the key is that they are small. Professional
technicians do the work under microscopes and they now typically have a 0.5 mm
spacing between lead centers; that's itty bitty. But it gets more complicated
for space gear since you have additional elements such as the conformal coat
that seals it, which needs to be removed and reapplied, and perhaps structural
and/or thermal epoxy between the component and the board. The risk of damage
to a board for a remove and replace for a professional technician doing this
day in and day out can not be dismissed. So for components such as this, a
remove and replace job probably would not be the way to go at all, with the
replaceable unit being a higher level module. So soldering these sorts of
components probably isn't a key objectice.


Two and three terminal devices are easy: Apply a drop of liquid RMA
flux and use two soldering irons to heat and lift the components. On
small SMD chips you can use solder wick to remove all the old solder.
Then you use a small curved pic and the edge of the hold soldering iron
tip to melt the solder under the lead and gently lift it from the pad
and let the lead cool. If you're good it only takes 30 seconds to a
minute to remove a small chip. Replacement is almost as easy. Tack
solder two pins on opposite corners, check the alignment, then solder
all the pins. This took me under a minute for small SMD chips.

Module level repair is desirable, if you have good interchangability
testing at the module level and known good spares. Replacing some
components require recalibration or realignment following the repairs.

In four years I only damaged two boards while doing rework. The
first one I was replacing a MPU chip and an idiot walked up and slammed
his fist down on the microscope cart while I was soldering and ripped
half the pads off one edge. The other one was after the M.E.s replaced
the microscope carts and managed to arrange the AC outlets so that a
loop of .015" solder touched the AC line. When I touched the grounded
iron to the solder it vaporized in my hand and deposited metal vapor all
around the processor. Not a bad average if you consider that I did
about 100 assorted boards a week for four years.

Then again, pin counts of around two hundred is getting into the low category
and the use of BGA and CCGA packages, for example. The old hobby shopper with
the soldering iron can't do that job, needs special gear. and is often
inspected with X-Ray. Etc.

But testing isn't as bad as you make out. Electronic probing is getting to be
more and more standard. And military equipment often has a requirement to
isolate a certain percentage of all faults to a component. The IEEE 1149.1
JTAG boundary scan is often used, although one must be careful using this in
the space environment.


I worked as a production tech on one radio that's part of the space
station. I worked on over 75% of the boards and modules for that radio
because I had the reputation of making dam sure it was right, or
refusing to sign anything off. The radio doesn't have that much
digital, and the only part with JTAG was the MC68340 processor chip.
JTAG isn't much use in analog circuits and it would take a well trained
tech to repair something as complex as the telemetry receivers we built.

Another feature to consider is that components designed for military and space
are often required to have built-in test functions (e.g., MIL-STD-1553B
controllers, MIL-STD-1750A processors). Even older processors from the '60s
had test support in-flight (something I'm looking into now).

Anyways, I found it to be an interesting article. As is the topic of in-
flight fault detection, diagnosis, and repair.


I wish them well trying to repair equipment in flight. I know it
won't be me. I'm too old, I have too many health problems and I'm
almost blind.

--
rk, Just an OldEngineer


--
Former professional electron wrangler.

Michael A. Terrell
Central Florida

rk wrote:

Michael A. Terrell wrote:

rk wrote:

Michael A. Terrell wrote:

rk wrote:

Yeah, and it will perhaps get to be even more interesting, as the
European and Japanese requirements for lead-free soldering take
affect; this will also affect component availability. Along with
making the soldering just different, for long duration equipment, as
is often found in many aerospace applications, the tin whisker
problem is of concern to some.

--
rk, Just an OldEngineer


Considering that most of the electronics will be built with surface
mount components it will be very difficult to repair, let alone
troubleshoot the complex equipment being used these days.

I worked on Telemetry receiving equipment and did my own rework on
chips up to 288 pins under a stereo microscope. It takes time to
learn how to probe these boards by hand, and requires a very steady
hand to resolder the leads without causing even more problems.
Processes that can be used on the ground would introduce unwanted
contaminants in a closed environment and you don't have the luxury of
a cleaning room to run repaired boards through to remove the flux
after a repair.

Well, yes and know. Some surface mount components aren't that hard,
such as resistors and capacitors and even some basic flat packs. Other
large integrated circuits, as you mention have a lot pins, some quite a
bit more than what you mention, but the key is that they are small.
Professional technicians do the work under microscopes and they now
typically have a 0.5 mm spacing between lead centers; that's itty bitty.
But it gets more complicated for space gear since you have additional
elements such as the conformal coat that seals it, which needs to be
removed and reapplied, and perhaps structural and/or thermal epoxy
between the component and the board. The risk of damage to a board for
a remove and replace for a professional technician doing this day in and
day out can not be dismissed. So for components such as this, a remove
and replace job probably would not be the way to go at all, with the
replaceable unit being a higher level module. So soldering these sorts
of components probably isn't a key objectice.


Two and three terminal devices are easy: Apply a drop of liquid RMA
flux and use two soldering irons to heat and lift the components. On
small SMD chips you can use solder wick to remove all the old solder.
Then you use a small curved pic and the edge of the hold soldering iron
tip to melt the solder under the lead and gently lift it from the pad
and let the lead cool. If you're good it only takes 30 seconds to a
minute to remove a small chip. Replacement is almost as easy. Tack
solder two pins on opposite corners, check the alignment, then solder
all the pins. This took me under a minute for small SMD chips.

Me? Can't solder well at all. What we've seen over the years is that after
around 5 cycles or so with the fine lead pitch, pad lifts start to occur.
Depends on the particular technician, some having a lighter hand than others.
Chip caps and resistors are reasonably easy. It's made messier and staking
and coating and thermal goop. This is generally a job for someone who works
doing soldering and polymerics on a regular basis. Here's a picture of a
board I found on the internet, for those playing at home, fairly routine
technology (but no BGA or CCGA):

Before stake and coat:

http://klabs.org/images/messenger/rm...ttom020303.jpg


Some of the hand soldering looks a little rough, and I don't like the
wires used for R521 & R522 instead of zero ohm resistors. A number if
IC pins have solder almost to the body of the component, and gold plated
leads are kind of rare these days.


After stake and coat:

http://klabs.org/images/messenger/rm...erC&S22003.jpg


I've reworked some failed conformal coated boards in the past, but
they were industrial.

Module level repair is desirable, if you have good interchangability
testing at the module level and known good spares. Replacing some
components require recalibration or realignment following the repairs.

I think having that good level of module is the way to go to Mars for a crewed
mission. Reprogrammability of the hardware and software will also be a
factor. I'm talking mostly digital here, by the way. There has been some
work done in reprogrammable analog but not nearly as much.


In four years I only damaged two boards while doing rework. The
first one I was replacing a MPU chip and an idiot walked up and slammed
his fist down on the microscope cart while I was soldering and ripped
half the pads off one edge. The other one was after the M.E.s replaced
the microscope carts and managed to arrange the AC outlets so that a
loop of .015" solder touched the AC line. When I touched the grounded
iron to the solder it vaporized in my hand and deposited metal vapor all
around the processor. Not a bad average if you consider that I did
about 100 assorted boards a week for four years.

Quite excellent.

Then again, pin counts of around two hundred is getting into the low
category and the use of BGA and CCGA packages, for example. The old
hobby shopper with the soldering iron can't do that job, needs special
gear. and is often inspected with X-Ray. Etc.

But testing isn't as bad as you make out. Electronic probing is getting
to be more and more standard. And military equipment often has a
requirement to isolate a certain percentage of all faults to a
component. The IEEE 1149.1 JTAG boundary scan is often used, although
one must be careful using this in the space environment.


I worked as a production tech on one radio that's part of the space
station. I worked on over 75% of the boards and modules for that radio
because I had the reputation of making dam sure it was right, or
refusing to sign anything off. The radio doesn't have that much
digital, and the only part with JTAG was the MC68340 processor chip.
JTAG isn't much use in analog circuits and it would take a well trained
tech to repair something as complex as the telemetry receivers we built.

Again, referring mostly to digital components, seeing it on a lot of
processors, gate arrays, etc. Analog ... ewwwwww. :-)

Another feature to consider is that components designed for military and
space are often required to have built-in test functions (e.g.,
MIL-STD-1553B controllers, MIL-STD-1750A processors). Even older
processors from the '60s had test support in-flight (something I'm
looking into now).

Anyways, I found it to be an interesting article. As is the topic of
in- flight fault detection, diagnosis, and repair.


I wish them well trying to repair equipment in flight. I know it
won't be me. I'm too old, I have too many health problems and I'm
almost blind.

It's a tough problem. They took a crack at that back in the '60s when
electronics reliability wasn't as good as it is now. There was concern over
the two week lunar trips and this is discussed in Eldon Hall's book. For the
OAO mission, they went with quad redundant logic, if memory serves, for the
long-term one year mission! I think we'll be revisiting this for Mars trips,
which will be quite a bit different than lunar ones, in my opinion.

--
rk


--
Former professional electron wrangler.

Michael A. Terrell
Central Florida

rk wrote:

Michael A. Terrell wrote:

Some of the hand soldering looks a little rough, and I don't like the
wires used for R521 & R522 instead of zero ohm resistors.

Just curious, why?

Not bragging but I had to point out every solder joint I did to the
QC people because they couldn't tell my work from the solder flow done
by our Heller reflow oven until I pointed out the slight color
difference between the paste solder and the Ersin .015" rework solder we
used.

As far as the wire jumpers, its quicker to use zero ohm resistors.
They are self aligning and you don't have to try to solder down a small
piece of wire then trim it. I can put on a dozen SMD resistors in the
time I can do one wire and they cost under a penny each. When you
factor in the time the SMD part is cheaper.


A number if
IC pins have solder almost to the body of the component, and gold plated
leads are kind of rare these days.

Actually, for that series of parts (the large 208-pin one), the MIL-SPECs only
call out that lead finish, note that it's the most readily available, and for
other finishes, go fish.

I've reworked some failed conformal coated boards in the past, but
they were industrial.

There's also quite a bit of staking used along with thermal goop under the
parts. More specialized skills for on-orbit repair, removal and replacement,
many of those substances needs to be kept quite cold, and has a limited
lifetime after shipment from the vendor. I'm not a polymerics guy, but
probably not long enough for a trip to Mars and back, if memory serves.


I'd be worried about the fumes from the coatings as they cured. You
can't just pump the contaminated air outside and replace it like you can
on the ground.

--
rk


--
Former professional electron wrangler.

Michael A. Terrell
Central Florida

rk wrote:

Michael A. Terrell wrote:

rk wrote:

Michael A. Terrell wrote:

Some of the hand soldering looks a little rough, and I don't like
the wires used for R521 & R522 instead of zero ohm resistors.

Just curious, why?

Not bragging but I had to point out every solder joint I did to the
QC people because they couldn't tell my work from the solder flow done
by our Heller reflow oven until I pointed out the slight color
difference between the paste solder and the Ersin .015" rework solder we
used.

Sounds like bragging. ;-)

I just consider it a skill developed over time, like logical
troubleshooting.

Actually, I got in trouble with the rework dept. for being too good.
One of the regular rework ladies would rant and scream that "None of the
techs can solder" then she couldn't find my rework without help. A
loudmouth tech that worked near my bench would go over to rework just to
stir up trouble by telling them I was better at it than any of the women
were. I learned to solder at eight and did a lot of repair work till I
was almost 50.

I was at a computer hardware show to see a new line of networking
hardware about five years ago. The manufactures rep laid out his sample
boards and was bragging about the quality. I found at least 20 reworked
solder joints on every board without a magnifier. I quietly pointed them
out to him and suggested that reworked products are not what you want on
display if you're trying to sell hundreds of boards at a time. He was
red faced as he told me thanks, and that he would have them send him
another set of samples that didn't have any rework. It looked like they
had given him engineering samples rather than production boards.

As far as the wire jumpers, its quicker to use zero ohm resistors.
They are self aligning and you don't have to try to solder down a small
piece of wire then trim it. I can put on a dozen SMD resistors in the
time I can do one wire and they cost under a penny each. When you
factor in the time the SMD part is cheaper.

Hmmm ... doesn't look like a piece of wire is that time intensive to prepare
and put down.

If it has to be done by hand VS Pick-N-Place it is.

Will the zero ohm SMD resistor have the same parasitics as the wire? I'll
look up the specification for it if you have the number handy, just curious.


Sorry, but I've been disabled for over three years now and no longer
have access to any of the data on the components. We used SMD 1208 and
0806 resistors at microwave frequencies, including where we needed a 0
dB pad. A zero ohm resistor from input to output and leave the other
positions open.


I don't know of any space grade part that costs under a penny each.


The 5% zero ohm was $26/5000, the 1% were a little higher.

--
rk, Just an OldEngineer

--
Former professional electron wrangler.

Michael A. Terrell
Central Florida

rk wrote:
Michael A. Terrell wrote:
The 5% zero ohm was $26/5000, the 1% were a little higher.

OK, that's a big more then under a penny.

I get 0.52 cents

And for boards such as the one
under discussion, 5,000 components is a lot, since there are
typically only a
few copies made.


Yes, but 5000 goes quick when spread across many products. Even if the
average product only uses 10 per board, and only 5 boards are made, the
next product is going to use 50 more, and the one after that 50
more....


/dps