Tualatin IHS removal. (Kony)

[~misfit~]

Hey Dave (and crew).

Well that Tualatin I mentioned that I'd glued the HS onto with Zalman epxoy?
I ran a razor blade around the IHS and eased it off, it came away no
trouble. One thing I found interesting, there was quite a gap between the
core and the IHS, filled with thermal compound. Probably 2mm or more. That
can't be good for heat-transfer.

I then put the HS into a hot oven for a half-hour and tapped the IHS with a
screwdriver and it came off the HS fairly easilly. Gouged the (aluminium) HS
a little, it's a lot softer when it's hot. Thankfully not in an area where
it makes contact with the core. I then 'lapped' the HS with some 600 grit
wet & dry sandpaper (wet) on glass to remove the remaining epoxy. It
actually came up a lot smoother than it was originally. Plus it's a lot
flatter now, it was quite bowed, a good half-hour of elbow-grease.

I don't have a board to run the Tui in at the moment but when I get one I'm
just gonna run it bare (no IHS). I really didn't like the way there was such
a huge gap between the IHS and the core, certainly can't be good for
heat-transfer.

Thanks for the URL and the idea, I now have another CPU I can use when I can
find a cheap enough board. It's a Celly 1.3Ghz.

Cheers,

 

[kony]

Hey Dave (and crew).

Well that Tualatin I mentioned that I'd glued the HS onto with Zalman epxoy?
I ran a razor blade around the IHS and eased it off, it came away no
trouble. One thing I found interesting, there was quite a gap between the
core and the IHS, filled with thermal compound. Probably 2mm or more. That
can't be good for heat-transfer.

Yep, after seeing quite a few chips sans their IHS, it's clear that a user
with attention to detail will have a better interface without the IHS,
providing of course that they have a halfway decent heatsink... but for a
Tualatin any old POS 'sink is usually plenty good after it's lapped.

I then put the HS into a hot oven for a half-hour and tapped the IHS with a
screwdriver and it came off the HS fairly easilly. Gouged the (aluminium) HS
a little, it's a lot softer when it's hot. Thankfully not in an area where
it makes contact with the core. I then 'lapped' the HS with some 600 grit
wet & dry sandpaper (wet) on glass to remove the remaining epoxy. It
actually came up a lot smoother than it was originally. Plus it's a lot
flatter now, it was quite bowed, a good half-hour of elbow-grease.

I'd be worried about the oven though, if the localized heat was too high
compared to the thermostat it might get hotter than intended, harm the
chip.

I don't have a board to run the Tui in at the moment but when I get one I'm
just gonna run it bare (no IHS). I really didn't like the way there was such
a huge gap between the IHS and the core, certainly can't be good for
heat-transfer.

Thanks for the URL and the idea, I now have another CPU I can use when I can
find a cheap enough board. It's a Celly 1.3Ghz.

It's good to hear that the IHS came off without any damage... I probably
would've taken a hacksaw to the 'sink and cut out a chunk of the bottom.

 

[Spajky]

I ran a razor blade around the IHS and eased it off, it came away no
trouble.

good ... no damage to the CPU base (PCB); my friend did not have that
luck!

One thing I found interesting, there was quite a gap between the
core and the IHS, filled with thermal compound. Probably 2mm or more. That
can't be good for heat-transfer.

Sure not! I did not wanna remove for myself that IHS but resolved a
problem my way .. , see my site ... check my mods too ..

I then put the HS into a hot oven for a half-hour and tapped the IHS ....

thats nothing; I put my Tuallie on the pot & than on the stove! see on
my site this updated under comp/..CPU Burn-in ... ))

I then 'lapped' the HS ... it's a lot
flatter now, it was quite bowed....

I did it also on IHS ...

If you have enough gutts, you can try lapping the core itself too ..
(some guys did it, but its a risky business)

I don't have a board to run the Tui in at the moment but when I get one I'm
just gonna run it bare (no IHS).

be carefull not to crack the edges of a CPU core; put some foamed gum
washers at the edge of a CPU (like AMDs have) to prevent that when
mounting a HS!

Be carefull that in any case the HS does not fall off during operation
/crack of a holding plastics on s370 socket!/, since IHS helps
delay-ing the thermal runaway of a chip to engage its thermal
protection at IMHO 125°C shutting it down ...

I really didn't like the way there was such
a huge gap between the IHS and the core, certainly can't be good for
heat-transfer.

sure not, but that gap is not equal on all samples, some have been
just assembled worse (depends on in which Intel´s factory was done &
boxed are generally made better)




-- Regards, SPAJKY ®
& visit my site @ http://www.spajky.vze.com
"Tualatin OC-ed / BX-Slot1 / inaudible setup!"
E-mail AntiSpam: remove ##

 

[~misfit~]

On Tue, 6 Apr 2004 19:37:22 +1200, "~misfit~"


I'd be worried about the oven though, if the localized heat was too
high compared to the thermostat it might get hotter than intended,
harm the chip.

Maybe I wasn't clear, I put the HS in the oven, still attached to the IHS
but *after* I'd removed the CPU.

It's good to hear that the IHS came off without any damage... I
probably would've taken a hacksaw to the 'sink and cut out a chunk of
the bottom.

Butcher. <g>.

Cheers,

 

[~misfit~]

good ... no damage to the CPU base (PCB); my friend did not have that
luck!

I think it was because the CPU was stuck to a HS. I could only get the blade
in dead straight, therefore it didn't dig into the CPU substrate.

I did it also on IHS ...

If you have enough gutts, you can try lapping the core itself too ..
(some guys did it, but its a risky business)

Yeah, I've seen a site that gives instructions on how to do it. I don't

be carefull not to crack the edges of a CPU core; put some foamed gum
washers at the edge of a CPU (like AMDs have) to prevent that when
mounting a HS!

The site Kony gave me mentioned this. I have some double-sided foam tape
that I'm going to try using. It is a little thicker than the die but I think
(hope) that it will compress enough to allow good contact between core and
HS. It is quite 'spongy'.

Be carefull that in any case the HS does not fall off during operation
/crack of a holding plastics on s370 socket!/, since IHS helps
delay-ing the thermal runaway of a chip to engage its thermal
protection at IMHO 125°C shutting it down ...

Thanks for the tips.

Cheers,

 

[kony]

Maybe I wasn't clear, I put the HS in the oven, still attached to the IHS
but *after* I'd removed the CPU.

Naw, you were clear... 'twas prior to the AM coffee, i was not clear

Butcher. <g>.

LOL, any excuse to cut up a heatsink... the scraps are infinitely cheaper
than ready-made mosfet 'sinks

 

[kony]

The site Kony gave me mentioned this. I have some double-sided foam tape
that I'm going to try using. It is a little thicker than the die but I think
(hope) that it will compress enough to allow good contact between core and
HS. It is quite 'spongy'.

Foam tape, provided it's open-cell, should compress plenty, though you
should be ok without any tape... Coppermines survived. It was mostly the
Athlons that were such a problem because of the extra tension AMD spec'd
for the heatsink clip, and I suspect it didn't help that the ceramic
carrier was completely inflexible.

 

[~misfit~]

Foam tape, provided it's open-cell, should compress plenty, though you
should be ok without any tape... Coppermines survived. It was mostly
the Athlons that were such a problem because of the extra tension AMD
spec'd for the heatsink clip, and I suspect it didn't help that the
ceramic carrier was completely inflexible.

Thanks. Yeah, I already tried fitting it in a socket and fitting a HSF, just
for the hell of it. It was Ok. Shame it wasn't a board that could run it.

Cheers,