Did I damage my CPU while removing heatsink?

[Guest]

I had a tough time removing my GlobalWin heatsink from my AMD 1.0 GHz
CPU last night, and when I finally got it off, I noticed that the
raised square in the middle of the CPU (what actually comes in contact
with the heatsink) was damaged a bit. One edge of that raised square
had a lot of stuff flaking off of it. I'm sending my mobo in to get
bad caps fixed, so I didn't try to test the CPU. But does it sound
like I damaged the CPU?

 

[Collon]

The flakes were probably residue from that thermal stuff that is applied
when the chip is put in. Don't worry about it yet. You can't do anything
until you get your computer back anyway.

 

[George Pontis]

I had a tough time removing my GlobalWin heatsink from my AMD 1.0 GHz
CPU last night, and when I finally got it off, I noticed that the
raised square in the middle of the CPU (what actually comes in contact
with the heatsink) was damaged a bit. One edge of that raised square
had a lot of stuff flaking off of it. I'm sending my mobo in to get
bad caps fixed, so I didn't try to test the CPU. But does it sound
like I damaged the CPU?

That raised square in the middle is the backside of the CPU, the
silicon die itself. If you clean off the heat sink compound carefully
(Q-tip with solvent) and see that the silicon has actually been chipped
from the edge then you have damaged the CPU. A small chip off the
corner is not fatal and it has happened to the best of us. If there is
actually something that you can call a "flake" that spreads back toward
the center of the chip, then even if the chip is still functional it
won't last long since good thermal contact will be impossible. Both AMD
and Intel since moved to the newer packages that have a metal heat
spreader to avoid this problem.

--

 

[Mxsmanic]

I had a tough time removing my GlobalWin heatsink from my AMD 1.0 GHz
CPU last night, and when I finally got it off, I noticed that the
raised square in the middle of the CPU (what actually comes in contact
with the heatsink) was damaged a bit. One edge of that raised square
had a lot of stuff flaking off of it. I'm sending my mobo in to get
bad caps fixed, so I didn't try to test the CPU. But does it sound
like I damaged the CPU?

Yes. But you won't really know until you start it up.

However, even if the chip itself is intact, it may no longer be able
to shed heat to the exterior of the package well enough to stay cool.

 

[kony]

I had a tough time removing my GlobalWin heatsink from my AMD 1.0 GHz
CPU last night, and when I finally got it off, I noticed that the
raised square in the middle of the CPU (what actually comes in contact
with the heatsink) was damaged a bit. One edge of that raised square
had a lot of stuff flaking off of it. I'm sending my mobo in to get
bad caps fixed, so I didn't try to test the CPU. But does it sound
like I damaged the CPU?

Get a petroleum-based cleaner, like Goo-Gone or WD-40 or
gasoline or whatever-you-have-handy that says "contains
Petroleum solvents" on the label and clean off all the gunk.

Once it's clean, then assess the condition. If it has a
very, very minor chip off the corner of the raised area it's
probably fine still, but if there's a giant chunk gone it's
probably dead.

Why would you have such an old motherboard fixed? You can
get newer socket A boards for $30 now and memory $75 per
GB... hardly seems worthwhile to repair a board that old if
you have to pay someone to do it plus shipping.

 

[Guest]

Get a petroleum-based cleaner, like Goo-Gone or WD-40 or
gasoline or whatever-you-have-handy that says "contains
Petroleum solvents" on the label and clean off all the gunk.

Once it's clean, then assess the condition. If it has a
very, very minor chip off the corner of the raised area it's
probably fine still, but if there's a giant chunk gone it's
probably dead.

Looks bad... almost the entire right edge is crumbled (as much as
almost a millimeter's worth). I'm just going to assume it's dead, so I
can buy another one now off ebay and have it by the time i get my mobo
back.

Why would you have such an old motherboard fixed? You can
get newer socket A boards for $30 now and memory $75 per
GB... hardly seems worthwhile to repair a board that old if
you have to pay someone to do it plus shipping.

I've read that some components are still being manufactured today with
those bad caps, so it's no guarantee that a newer board wouldn't have
to have its caps replaced. I might as well just get my current board
fixed.

 

[Mxsmanic]

I've read that some components are still being manufactured today with
those bad caps, so it's no guarantee that a newer board wouldn't have
to have its caps replaced. I might as well just get my current board
fixed.

You must have some really inexpensive PC repair shops in your town if
you can get a board fixed for less than it would cost to replace it
(either that, or this is an extremely expensive board).

 

[kony]

Looks bad... almost the entire right edge is crumbled (as much as
almost a millimeter's worth). I'm just going to assume it's dead, so I
can buy another one now off ebay and have it by the time i get my mobo
back.

Tough call, "almost a millimeter" might be working but it
does seem a real potential there is damage.

I've read that some components are still being manufactured today with
those bad caps, so it's no guarantee that a newer board wouldn't have
to have its caps replaced. I might as well just get my current board
fixed.

There were typically two kinds of capcitor problems- caps
that simply had instable electrolyte and when electricity
was applied, gasses built up over time, AND the more common
problem today- insufficient numbers of caps or lower quality
parts such that they run hot.

The former problem was mostly isolated to earlier
generations of boards- early socket A was about the last
effected with brands like GSC on some MSI and Epox boards,
G-Luxon on PCChips/ECS.

After that period, there were still some marginal designs
but overall it was more a matter of watching temperatures,
and it may matter less with an aging system since a lower
quality board may not last "as many" years but since the
other parts have already aged some, the board may not need
to last as many years before some other failure point is
encountered that is serious enough that whole system is
retired.

On the other hand, when a given technology is new then
typically only the lower-grade parts are cheap but now with
socket A being phased out, more median or better quality
boards are also at very low prices. Here's an nForce2
Ultra-400 refurb for $43 delivered... many refurbs now
aren't even used, they're just stock piled up by
manufacturers to cover warranty replacement but when
warranty is up they have the stock to get rid of. No
guarantees on any particular parts but the secondary
consideration was that with a newer board, there's also the
option of later upgrading the CPU.

 

[spodosaurus]

Tough call, "almost a millimeter" might be working but it
does seem a real potential there is damage.






There were typically two kinds of capcitor problems- caps
that simply had instable electrolyte and when electricity
was applied, gasses built up over time, AND the more common
problem today- insufficient numbers of caps or lower quality
parts such that they run hot.

The former problem was mostly isolated to earlier
generations of boards- early socket A was about the last
effected with brands like GSC on some MSI and Epox boards,
G-Luxon on PCChips/ECS.

Boards were affected a LOT more recently than that! I lost a GA-7VRXP
board, and I've seen more recent boards with bad caps as well.


--
spammage trappage: remove the underscores to reply

I'm going to die rather sooner than I'd like. I tried to protect my
neighbours from crime, and became the victim of it. Complications in
hospital following this resulted in a serious illness. I now need a bone
marrow transplant. Many people around the world are waiting for a marrow
transplant, too. Please volunteer to be a marrow donor:
http://www.abmdr.org.au/
http://www.marrow.org/

 

[spodosaurus]

I would avoid WD-40 unless you can remove the residue it leaves. It
penetrates a surface and leaves a residue (that's what makes it good
for many other tasks), and might prevent heatsink grease from doing
a good job since heatsink grease would need to displace the WD-40.

I just use acetone and then alcohol swabs (the kind you get for subQ
injections at the pharmacy). Cleans everything right up.

Ari

--
spammage trappage: remove the underscores to reply

I'm going to die rather sooner than I'd like. I tried to protect my
neighbours from crime, and became the victim of it. Complications in
hospital following this resulted in a serious illness. I now need a bone
marrow transplant. Many people around the world are waiting for a marrow
transplant, too. Please volunteer to be a marrow donor:
http://www.abmdr.org.au/
http://www.marrow.org/

 

[John Doe]

On 14 Jan 2006 11:34:38 -0800, (e-mail address removed)


Get a petroleum-based cleaner, like Goo-Gone or WD-40 or
gasoline or whatever-you-have-handy that says "contains
Petroleum solvents" on the label and clean off all the gunk.

I would avoid WD-40 unless you can remove the residue it leaves. It
penetrates a surface and leaves a residue (that's what makes it good
for many other tasks), and might prevent heatsink grease from doing
a good job since heatsink grease would need to displace the WD-40.

 

[kony]

Boards were affected a LOT more recently than that! I lost a GA-7VRXP
board, and I've seen more recent boards with bad caps as well.

What brand of caps?

Just having capacitor problems is not necessarily caused by
the often referenced taiwanese defective capacitors.

If one takes "most" motherboards that would run for several
years, puts them in a case with insufficient airflow,
especially with an overclocked or higher-heat CPU, that
alone may kill it.

Come to think of it, I've seen one of that model where
Gigabyte or some 3rd party (hired by Gigabyte) had replaced
the 3 caps near the audio jacks and it looked suspiciously
like one of the mosfets had heated up till it melted the
solder and slid down the board a few mm.

 

[kony]

I would avoid WD-40 unless you can remove the residue it leaves. It
penetrates a surface and leaves a residue (that's what makes it good
for many other tasks), and might prevent heatsink grease from doing
a good job since heatsink grease would need to displace the WD-40.

It does not penetrate a non-porous surface such as found on
a modern CPU or heatsink. It would not be good to use it on
an old ceramic carrier based CPU though, unless you know for
sure that the ceramic is sealed (it seems to be, but "seems"
is not definitive proof).

I've used WD-40 in a pinch though, it did fine. Just about
anything you use is going to "leave residue", including even
water, because the whole purpose was to clean off the CPU
and what you'd be cleaning off is residue until removed.

 

[kony]

I just use acetone and then alcohol swabs (the kind you get for subQ
injections at the pharmacy). Cleans everything right up.

Acetone would certainly do it brute-force, but be very
careful with it as it will melt far more types of plastics
than other options... and the outer coating on a CPU is
plastic, as well as sleeves on capacitors, the socket
itself, the fan, wires, etc.

 

[John Doe]

John Doe wrote:

I just use acetone and then alcohol swabs (the kind you get for
subQ injections at the pharmacy). Cleans everything right up.

If acetone isn't too harsh.

The WD-40 will clean it very well too, but the point is to avoid
leaving residue that you cannot see.

 

[John Doe]

kony wrote:

Boards were affected a LOT more recently than that! I lost a
GA-7VRXP board, and I've seen more recent boards with bad caps as
well.

Maybe you should buy from a better merchant.

 

[John Doe]

It does not penetrate a non-porous surface such as found on
a modern CPU or heatsink.

What you think is a nonporous surface has lots of pores.

If you look very very closely when you apply WD-40 to metal, you can
watch it flow across the surface. And I'm not talking about it
flowing like water, it's much more subtle.

I've used WD-40 in a pinch though, it did fine.

Yes it cleans very well, but it leaves a residue.

Just about anything you use is going to "leave residue",

My use of "residue" is fitting for this context.

Leaving a protective residue/coating is one of WD-40's advertised
qualities.

The objective is to provide the best possible contact between your
CPU and heatsink. Therefore you want as little residue as possible.
It's likely to be counterproductive.

because the whole purpose was to clean off the CPU
and what you'd be cleaning off is residue until removed.

It works great for cleaning, but it might degrade the performance of
your heatsink grease unless you remove the residue.

After being told how heatsink grease works, I won't use Dubya D40
around heatsinks.

 

[kony]

What you think is a nonporous surface has lots of pores.

Not significant ones... which is the whole reason there are
the differing terms porous and non-porous.

If you look very very closely when you apply WD-40 to metal, you can
watch it flow across the surface. And I'm not talking about it
flowing like water, it's much more subtle.

Did you ever consider wiping off the cpu after spraying
(anything, doesn't matter what) all over it?

Yes it cleans very well, but it leaves a residue.

Are you assuming the CPU surface as clean to laboratory
standards in the first place? There's clean at a molecular
level and then there's relatively clean. After wiping off
the core or spreader it is clean enough to make good thermal
conduction. If you want to be extra (overly) cautious, you
might then clean it off with another solvent, but ultimately
once you no longer see any on it, it is clean enough to work
fine.

My use of "residue" is fitting for this context.

Leaving a protective residue/coating is one of WD-40's advertised
qualities.

Yes, when you spray it on and leave it alone... not when you
put a bit on a towel and wipe it off again.

The objective is to provide the best possible contact between your
CPU and heatsink.

Actually, no.
The objective is to provide a "good enough" contact, to keep
the CPU cool enough.

If we go down the hypothetical road of "best", suddenly you
can't use conventional heatsinks at all, and that choice is
far more significant than what you're claiming.

Therefore you want as little residue as possible.
It's likely to be counterproductive.

You want as little residue as _reasonably_ possible.
If you have a non-residue cleaner nearby, it's defintely
reasonable, even prudent to use it instead. If you have
such a cleaner but it doesn't dissolve the original TIM, but
will clean off the WD-40 residue, again it is a good idea to
use it as that is reasonable.

On the other hand, if what you have is WD-40, and your
heatsink is not rough, and only wd-40 is applied, then wiped
off thorougly with a dry paper towel, it WILL do fine. I've
done it, watched others do it, and have no suspersticions
against doing it again because it does fine. If you are
claiming it will make a fraction of a degree temp
difference, why would anyone care? It definitely does not
make a several degree temp difference if it has been wiped
off with a dry towel, let alone with another solvent later.

In short, no. WD-40 works fine and you're just guessing it
wouldn't. Do some experiements, otherwise it's just another
myth.

It works great for cleaning, but it might degrade the performance of
your heatsink grease unless you remove the residue.

A guess.
"Residue" has to be a relative term, the remaining residue
after thoroughly wiping the surface is inconsequential.

 

[John Doe]

Mostly semantics and falsehoods from someone who is a little too
wrapped up in his own advice.

 

[kony]

Mostly semantics and falsehoods from someone who is a little too
wrapped up in his own advice.

No, just common sense from someone who knows it's not
necessary to get obsessive onto a microscopic level of
cleaning. Your theory about residue only holds true when it
wasn't a reasonable attempt to wipe it off.

Perhaps more to the point is- I've already mentioned wiping
till you see no more of it, so if you suggest using some
other cleaning agent and wiping till you see no more of it,
how do you have any assurance that there isn't the same
level of residue remaining? You don't, unless you get out
the high powered microscope to examine it... which I don't
recall many assemblers doing. Anything you use to clean off
the CPU will, as explicitly desired, be all over the top to
clean it so some idea about porous or spreading is
completely wrong, anything will.

After you wipe off the CPU, if it looks clean enough, it is.
It is really that simple, maybe you should just try it
sometime.