Debunking myths about removing the P4's IHS.

Printable View

Show 25 post(s) from this thread on one page

05-27-2003, 05:30 AM
FaTs

Debunking myths about removing the P4's IHS.

Okay I got kind of curious and decided to remove the IHS from one of my P4 2.4b’s, the first thing I want to ponit out is that, this is alot harder than the guides make it out to be. I severely bent serval pins, (even a couple fully flat), the glue is all the way round the IHS not just in the corners and is rubber based so it stretches which makes getting it off even harder.

As for the results, well, they are pretty disappointing. I’m sitting here typing this at wait for it....a massive overclock of 2411mhz. The thing refuses to go into windows at any sort of decent o/c, this chip did 3064 on 1.5v using the retail sink :(. My temps aren’t that great either, I’m going to try reseating the HSF and reapplying some thermal paste later...but I don’t think it will get me very far. I might try applying more clamping pressure, but I now have to be careful of crushing the core.

The thermal paste between the core and the heat spreader is indeed the same as that Intel ships with its Xeon CPU’s and 3.06/3.0ghz parts. Also the finish on the CPU die has to be seen to be believed...

Looks like my 2.4c will see action alittle earlier than planned...
05-27-2003, 05:36 AM
Colossus

WOW... I am shocked that you would even want to do that :( I wouldnt remove it for just a few degrees cooler... I know its always fun to experiment, but I do have a line that I will not cross and that was it :)

I may like expensive hardware but I dont want expensive dead hardware.. Like my 9800Pro :( Rest her little GPU soul!
05-27-2003, 05:41 AM
KaoTiK

Re: Debunking myths about removing the P4's IHS.

Quote:

Originally posted by FaTs
Okay I got kind of curious and decided to remove the IHS from one of my P4 2.4b’s, the first thing I want to ponit out is that, this is alot harder than the guides make it out to be. I severely bent serval pins, (even a couple fully flat), the glue is all the way round the IHS not just in the corners and is rubber based so it stretches which makes getting it off even harder.

As for the results, well, they are pretty disappointing. I’m sitting here typing this at wait for it....a massive overclock of 2411mhz. The thing refuses to go into windows at any sort of decent o/c, this chip did 3064 on 1.5v using the retail sink :(. My temps aren’t that great either, I’m going to try reseating the HSF and reapplying some thermal paste later...but I don’t think it will get me very far. I might try applying more clamping pressure, but I now have to be careful of crushing the core.

The thermal paste between the core and the heat spreader is indeed the same as that Intel ships with its Xeon CPU’s and 3.06/3.0ghz parts. Also the finish on the CPU die has to be seen to be believed...

Looks like my 2.4c will see action alittle earlier than planned...

Would you be able to put the heatspreader back on?

I mean, you might be able to get better temps if you put AS3 in between the core and the heatspreader instead of the generic crap that comes in there, if it isnt too good.

Worth a try, lol, 11MHz is pretty much useless.
05-27-2003, 05:59 AM
macutty

Re: Debunking myths about removing the P4's IHS.

Quote:

Originally posted by FaTs
Okay I got kind of curious and decided to remove the IHS from one of my P4 2.4b’s, the first thing I want to ponit out is that, this is alot harder than the guides make it out to be. I severely bent serval pins, (even a couple fully flat), the glue is all the way round the IHS not just in the corners and is rubber based so it stretches which makes getting it off even harder.

As for the results, well, they are pretty disappointing. I’m sitting here typing this at wait for it....a massive overclock of 2411mhz. The thing refuses to go into windows at any sort of decent o/c, this chip did 3064 on 1.5v using the retail sink :(. My temps aren’t that great either, I’m going to try reseating the HSF and reapplying some thermal paste later...but I don’t think it will get me very far. I might try applying more clamping pressure, but I now have to be careful of crushing the core.

The thermal paste between the core and the heat spreader is indeed the same as that Intel ships with its Xeon CPU’s and 3.06/3.0ghz parts. Also the finish on the CPU die has to be seen to be believed...

Looks like my 2.4c will see action alittle earlier than planned...

OK, good to know. I was just thinking about try it myself.

One question though. I have seen a couple guys whoe have removed the IHS with glue only in the corners, I wonder if this chnged for certain CPU's?

The idea I had thought of (provided it was the corners only glued) was to use piano wire and "pull" rather than "push" to cut the glue. Do you think this owuld have helped?
05-27-2003, 06:26 AM
Pyro

Re: Debunking myths about removing the P4's IHS.

Quote:

Originally posted by FaTs
As for the results, well, they are pretty disappointing. I’m sitting here typing this at wait for it....a massive overclock of 2411mhz. The thing refuses to go into windows at any sort of decent o/c, this chip did 3064 on 1.5v using the retail sink :(.

That sucks. :o
05-27-2003, 06:30 AM
FaTs

Macutty, theres only about 2mm of heatspreader that isn't glued down on my cpu, I'd say it would be standard thing accross the northwoods.

I did just to do it, it didn't work, but you get that. I guess I have another bit of dead/useless hardware to add to my collection :|
05-27-2003, 06:53 AM
KaoTiK

Quote:

Originally posted by FaTs
Macutty, theres only about 2mm of heatspreader that isn't glued down on my cpu, I'd say it would be standard thing accross the northwoods.

I did just to do it, it didn't work, but you get that. I guess I have another bit of dead/useless hardware to add to my collection :|

If you dont want it, I will take it :)
05-27-2003, 07:00 AM
Colossus

Quote:

Originally posted by FaTs
Macutty, theres only about 2mm of heatspreader that isn't glued down on my cpu, I'd say it would be standard thing accross the northwoods.

I did just to do it, it didn't work, but you get that. I guess I have another bit of dead/useless hardware to add to my collection :|

Dont feel to bad.. I own a dead 9800Pro :|
05-27-2003, 07:03 AM
KaoTiK

Quote:

Originally posted by Colossus

Dont feel to bad.. I own a dead 9800Pro :|

I will take that too :)
05-27-2003, 07:05 AM
Colossus

hahaha, Why does Dora The Explorer come to mind!! God I hate children shows!!!

KaoTiK, no swiping!!
KaoTiK, no swiping!!
KaoTiK, no swiping!!

hehehe :) You have to watch it to understand!
05-27-2003, 07:06 AM
FaTs

Quote:

Originally posted by Colossus

Dont feel to bad.. I own a dead 9800Pro :|

I own a Dead and a semi dead GF2 Ultra, they cost more than your 9800pro :p
05-27-2003, 10:16 AM
pudad

Quote:

Originally posted by Colossus
WOW... I am shocked that you would even want to do that :( I wouldnt remove it for just a few degrees cooler... I know its always fun to experiment, but I do have a line that I will not cross and that was it :)

I may like expensive hardware but I dont want expensive dead hardware.. Like my 9800Pro :( Rest her little GPU soul!

Man bummer, I feel you pain man. But how did it happen, too much ocing? That sucks :( .
05-27-2003, 10:40 AM
Moridin

Re: Re: Debunking myths about removing the P4's IHS.

Quote:

Originally posted by KaoTiK

Would you be able to put the heatspreader back on?

There is no point. Without the machine applied downpressure the HIS would just hinder cooling.
05-27-2003, 08:24 PM
FaTs

Okay it's not looking too bad after all. These chips are stronger than I orginally thought, I tested an OLD 1.8a that had died of SNDS to destruction last and night and they are not easy to break, probably around the same sort of strength as a AMD tbred.

Moridin is right, my computer is running with the IHS on but it still is no performing correctly... that said the glue that holds the IHS is rubber based and seems to have some play in it :confused:

So I'm going to clamp down a waterblock on this thing as see how this goes. If it still wont play, I'll glue the IHS on and dig an SL6ZR out of one of the machines at work ;)
05-27-2003, 08:28 PM
Colossus

Quote:

Originally posted by FaTs
So I'm going to clamp down a waterblock on this thing as see how this goes. If it still wont play, I'll glue the IHS on and dig an SL6ZR out of one of the machines at work ;)

That is just so wrong in so many ways!!!
05-27-2003, 08:57 PM
FaTs

Quote:

Originally posted by Colossus

That is just so wrong in so many ways!!!

No :p They're my computers ;) I like to keep a stock of o/c ready cpu's at work :D
05-27-2003, 09:00 PM
Colossus

uh huh!!! Sure, yeah I buy that :D
05-27-2003, 09:08 PM
Moridin

I think you will still get better results by discarding the IHS at this point. If you get a good enough die-heatsink junction and lots of down pressure you should be able to get the chip near where it was before, but only if you don’t use the HIS.

Another thing. What temperatures are you running? Perhaps the bent pins are affecting your FSB signal integrity, and the reduced OC has nothing to do with cooling…
05-27-2003, 09:14 PM
FaTs

Quote:

Originally posted by Moridin
I think you will still get better results by discarding the IHS at this point. If you get a good enough die-heatsink junction and lots of down pressure you should be able to get the chip near where it was before, but only if you don’t use the HIS.

Another thing. What temperatures are you running? Perhaps the bent pins are affecting your FSB signal integrity, and the reduced OC has nothing to do with cooling…

Actully i was thinking pins, but hoping it wasn't because they are nearly impossiable to straighten, so I wanted to explore the pressure route first ;) .

Temps at 2.4 are 41c - case temp is 35c without the IHS, 44c with the IHS just chucked on. I think it was getting roughly 42-43c with an unmodded IHS.
05-27-2003, 10:48 PM
macutty

Quote:

Originally posted by FaTs

Actully i was thinking pins, but hoping it wasn't because they are nearly impossiable to straighten, so I wanted to explore the pressure route first ;) .

Temps at 2.4 are 41c - case temp is 35c without the IHS, 44c with the IHS just chucked on. I think it was getting roughly 42-43c with an unmodded IHS.

Wekk if you rtemps are lower I can't see removing the IHS as having an adverse afect on OCing. SOunds more like then pins than anyting else IMO.
05-27-2003, 11:18 PM
Moridin

Quote:

Originally posted by FaTs

Temps at 2.4 are 41c - case temp is 35c without the IHS, 44c with the IHS just chucked on. I think it was getting roughly 42-43c with an unmodded IHS.

Do you mean 44 now, with minimal overclock, and 43 before with the overclock? How about with the IHS out of the way?
05-27-2003, 11:32 PM
FaTs

Quote:

Originally posted by Moridin

Do you mean 44 now, with minimal overclock, and 43 before with the overclock? How about with the IHS out of the way?

Temps at 2.4 are 41c now.

I dont know what they where before, because I never ran the thing aircooled or at less than 3.06ghz :( 42-43 is an educated guess, based measuring the temps of two other P4 2.4's in my mobo.
05-27-2003, 11:33 PM
Ramuman

Quote:

Originally posted by FaTs

I own a Dead and a semi dead GF2 Ultra, they cost more than your 9800pro :p

I remember when that happened...you lost 32 megs on the semi dead one didn't you? I felt bad cause for you that was right around when I'd dropped $400 or so for my GF2 Ultra. I guess the sadest part of it is that I still use that GF2 Ultra in my computer now!
05-27-2003, 11:43 PM
FaTs

Quote:

Originally posted by Ramuman

I remember when that happened...you lost 32 megs on the semi dead one didn't you? I felt bad cause for you that was right around when I'd dropped $400 or so for my GF2 Ultra. I guess the sadest part of it is that I still use that GF2 Ultra in my computer now!

Good to see your still lurking :p

Yep one is now making a great door stop and the other is still serving in my celeron, with only 32mb of ram :)
05-29-2003, 09:16 AM
Sinnersis

The same thing was tried here:
http://www.overclockers.com/tips1087/
and they had some success with it. Althrough it wasn't neccessarily conclusive.
Bad Luck there :(

Show 25 post(s) from this thread on one page