Testing large number of processors

[Robert Redelmeier]

Generally speaking, a P4 processor will run without thermal paste for
about 5-10 minutes before it starts overheating (maybe less time if

I'd be very surprised it if could make it that long with a completely
dry (clean) joint. Any left over phase change pade would help alot.

On the other hand, if the heatsink isn't clamped down, the processor
will usually overheat and shut itself off in less than 15 seconds.

True if the joint is dry or there's a very stiff grease /
old pad being used. But I've run CPUs for hours unclamped
by using a fairly fluid grease (Radio Shack). Balancing the
HSF to be dead flat does take care.

Now I'm sure that these OEMs don't put a huge amount of
emphasis on testing CPUs, though I know that they at least
ensure the system boots to Windows.

This would suprise me. Any NIB OEM system behaves differently
on first boot than on later boots, so I presume has never
been booted on the disk image.

The disks are all pre-iamged in a RAID-type device.

An OEM might boot a nearly finished system from floppy,
network or USB for a final QC check. But a boot into
MS-Windows takes far too long.

-- Robert

 

[keith]

True. After all, these CPUs once _did_ pass.

Assuming that these aren't floor-sweepings, sold out the back door, that's
a good point.

I think his only problem is cooling. With Socket-7 processors
we'd just sit a heatsink (no goop) on 'em and let gravity do the work.
This strategy isn't going to cut it today though. I think either a
fansink with no goop, but latched down, or perhaps a fansink with goop and
no latch would work long enough to POST. At least this should work with
parts with an IHS. Bare chips are going to be more difficult.

As far as goop goes, I'm with Del. Anything with salt is a very bad idea
(particularly with the bare-die chips). Perhaps some zinc-oxide sun block?

Check the pinout. I consider any of the outside pins
likely candidates for ESD.

I'd also check the outside pins to make sure they're still there. If
there is one with a zapped I/O, oh well! Refund the money. The customer
knows they're not buying new parts. There has to be some risk assumed at
that end too.

 

[keith]

So let's say I go with Vaseline. Then there's no controversy about
whether the saline is a problem. Can I use Vaseline as a cheap
substitute for thermal paste that's adequate for a post or
even a boot to Windows?

As I said before, what about zinc-oxide sun-block? You should be able to
buy that stuff by the pound.

If so, this is starting to sound like an attractive shortcut:

1. I coat the processor with Vaseline.
2. I rest a clean heatsink carefully on top of the processor so it
rests flat, but I don't clamp it down.
3. I do a post or even a boot to Windows.
4. If it boots, I sell it. If the buyer has a problem, I refund
or replace.

Does anyone have a practical objection to this approach?

Maybe weight down the fansink, arranged such that it doesn't block the air
flow. Support the back of the board so it doesn't stress the board.
We had good luck supporting the back of the board with a sheet of
anti-static foam. The boards were just lined up on the bench with the
PSU's tucked out of the way.

I don't think I'd even bother booting Windows, unless you use it as an
advertising gimick; "Guaranteed to boot Windows". If it doesn't, refund
the money with no questions asked. Don't even ask for the onld one to be
returned.

 

[Robert Redelmeier]

As I said before, what about zinc-oxide sun-block?
You should be able to buy that stuff by the pound.

Zinc oxide sunblock is too good! Danm close to the RS stuff.
It is also somewhat hard to remove well.

-- Robert

 

[Rob Stow]

As I said before, what about zinc-oxide sun-block? You should be able to
buy that stuff by the pound.

Why not a flat-bottomed steel bottle full of liquid nitrogen ?

Set it on processor, test, lift it off. Repeat with next CPU.
You can|should insulate the walls of the bottle so that the
nitrogen lasts a little longer.

With something nice and cold like that a perfect fit between the
CPU & bottom of the bottle shouldn't be necessary for a simple
one or two minute test.

 

[GSV Three Minds in a Can]

Bitstring <[email protected]>, from the
wonderful person keith <[email protected]> said

Perhaps some zinc-oxide sun block?

Why would you bother when you can get Zinc Oxide 'generic' thermal paste
in litre pots for a few shillings? I mean you really don't have to use
the latest Antarctic-Palladium-Ceramic-Version 6b. In fact is doesn't
make a degree difference anyway, if the HS is actually flat.

 

[Rob Stow]

Why not a flat-bottomed steel bottle full of liquid nitrogen ?

Set it on processor, test, lift it off. Repeat with next CPU.
You can|should insulate the walls of the bottle so that the
nitrogen lasts a little longer.

With something nice and cold like that a perfect fit between the
CPU & bottom of the bottle shouldn't be necessary for a simple
one or two minute test.

Took me an hour to find the nitrogen, but a friend of mine
already had a polished-bottomed copper cylinder left over from a
few crazy over-clocking experiments. I gave this idea a try with
an 2.4 GHz Athlon64.

With no attempt to improve the CPU/cylinder interface with things
like thermal paste, the system ran just fine so long as there was
still come nitrogen in the cylinder.

No idea what happens next: during two tests I shut down both
times as soon at the cylinder ran out of nitrogen. Relative
humidity here is 80% at the moment, so I quickly got frost on the
cylinder and when the nitro ran out I wanted to remove the
cylinder before the frost had a chance to melt.

 

[Robert Redelmeier]

Why not a flat-bottomed steel bottle full of liquid nitrogen ?

Dry ice in iso propyl alcohol is easier to obtain and handle.

With something nice and cold like that a perfect fit between
the CPU & bottom of the bottle shouldn't be necessary for
a simple one or two minute test.

The beaker will frost over, and if that frost is the right amount,
it will melt and act as thermal transfer medium. Water is actually
pretty good. Too much will make the chips wet.

-- Robert

 

[Bob Palmer]

I'd be very surprised it if could make it that long with a completely
dry (clean) joint. Any left over phase change pade would help alot.


True if the joint is dry or there's a very stiff grease /
old pad being used. But I've run CPUs for hours unclamped
by using a fairly fluid grease (Radio Shack). Balancing the
HSF to be dead flat does take care.


This would suprise me. Any NIB OEM system behaves differently
on first boot than on later boots, so I presume has never
been booted on the disk image.

The disks are all pre-iamged in a RAID-type device.

An OEM might boot a nearly finished system from floppy,
network or USB for a final QC check. But a boot into
MS-Windows takes far too long.

-- Robert

Just to add a little more confusion to the topic - why not just dampen the
chip - heat sink interface with distilled water and try a few tests? Good
thermal conductivity, no residue problem, trivial cost, easy to apply and
if it works temperature never is going to get above 100 degrees which would
be an issue. Sure it won't last but who cares? it only has to be there
for a few minutes. And as an added bonus, the low viscosity means easy to
obtain a good contact and easy removal. So long as the heat sink can sit
flat on the chip no hold down would be necessary. And as a last random
thought - roll your own heat sinks from heavy copper slab cut to fit and
keep several to allow the thermal mass to do the cooling and the weight to
do the hold down. Simply allow a warmed up sink to cool on its own and
use a fresh one for the next chip.

Bob

 

[keith]

Don't think about numbers -- calculate!

How much gel is left after wiping? Max 5um on the flat top
and 15um on the edges? That gives 800 nL or 800 ug. At 0.8%w
NaCl, that is 6400 ng. Call it low microgram if you wish.

Think if it a different way. You apply about one drop (50uL).
Good cleaning should remove 98+% leaving 1000 nL or 8000 ng.


I did and got no good hits on Google. Can you be
more specific? Certainly sodium can contaminate printed
semiconductors, but it's got to get there. How is it going
to migrate through 1+mm resin or 0.5 mm hi purity silicon?

A couple of atoms in the wrong place is 'nuff. Sodium is nasty stuff, as
anyone living in the North can attest to.

 

[keith]

Bitstring <[email protected]>, from the
wonderful person keith <[email protected]> said


Why would you bother when you can get Zinc Oxide 'generic' thermal paste
in litre pots for a few shillings? I mean you really don't have to use
the latest Antarctic-Palladium-Ceramic-Version 6b. In fact is doesn't
make a degree difference anyway, if the HS is actually flat.

It's really the same stuff. Sun-block is easier to find though. Look for
the lifeguards with the white noses. ;-)

 

[keith]

Why not a flat-bottomed steel bottle full of liquid nitrogen ?

It is cheap stuff, but it's a little hard to handle for such things.
Besides, there is such a thing as *too* cold. Yes, Liquid Nitrogen is
often used in testing, but water is easier unless testing at the low
extremes (0C, for commercial) is needed. For these purposes air should
work fine though.

Set it on processor, test, lift it off. Repeat with next CPU. You
can|should insulate the walls of the bottle so that the nitrogen lasts a
little longer.

Too cold could cause all sorts of problems.

With something nice and cold like that a perfect fit between the CPU &
bottom of the bottle shouldn't be necessary for a simple one or two
minute test.

Exactly. Why go to this extreme though.