P4 IHS removal question

[Turbo]

Hi guys. Has anyone in here tried to remove the Heat Spreader on the P4 cpu, and if so are you using the Intel boxed cooler? I'm going to remove the IHS when I get my Swiftech MCX478 (seems to be a bit off unfortunately :( ) but I'm wondering if the Intel heatsink will actually make contact with the cpu die if the IHS is removed, because of the way it is mounted. With the Swiftech it won't be a problem. I'd be happy to know if it works, so I can push my "little" 1.5 P4 even further. It's not 100% stable over 2.0 Ghz now, and I'm sure it's because of heat. It will load windows up to at least 2055 Mhz and run most applications fine. However it crashes in cpu stability test with a temperature arond 69°c. It also crashes to the windows desktop in 3dmark2001 over 2010 Mhz. In my experience a behavior consistent with overheating. Well if anybody succesfully removed the IHS and are using the boxed Intel heatsink/fan I would really love to know :) Thanks

[Miss winnie]

To my knoledge if you are already experiencing heat problems don't make it worse by removing IHS from it :)

[Turbo]

Originally posted by Miss winnie
To my knoledge if you are already experiencing heat problems don't make it worse by removing IHS from it :)

Well I've heard from a couple of guys, that after removing the IHS their temperatures dropped 6-8°c with the same cooler. I'm pretty convinced that removing the IHS will lower the temperature, granted of course that the contact between the IHS and the heatsink is intact.

[ol' man]

This may be of some use to you. I am pretty sure it will work the same for the PIV too.

http://forums.overclockers.ws/vb/showthread.php?threadid=45095

I have heard of one guy removing his IHS and he said it cools alot better. It would be best to have before and after temps though but what you can do is reapply ASII or whatever thermal paste and put it back on to test how much better it is without the IHS.

[batboy]

Are we sure this will help cooling? Intel put those IHS on for a reason and I'm sure their engineers did a lot of research prior to deciding to add the IHS. I thought the main reason was to allow better contact with the heatsink which would allow better cooling. I'd be interested in before and after temps measured in a consistant manner.

[Pinky]

Originally posted by batboy
Are we sure this will help cooling? Intel put those IHS on for a reason and I'm sure their engineers did a lot of research prior to deciding to add the IHS. I thought the main reason was to allow better contact with the heatsink which would allow better cooling. I'd be interested in before and after temps measured in a consistent manner.

In all actuality, the fewer layers between the CPU and it's primary cooling contact the better the dissipation of heat. I think Intel knew this, but because of the fragile and small nature of this core, they needed to 'protect' the cpu.

[ol' man]

Originally posted by Pinky


In all actuality, the fewer layers between the CPU and it's primary cooling contact the better the dissipation of heat. I think Intel knew this, but because of the fragile and small nature of this core, they needed to 'protect' the cpu. ]


You would think so but after a large amount of times of taking my HS on and off I don't think so anymore. I mean yeah it may be more fragile than the cumine but IMHO not by much. You will just have to be more careful that is all.

[lennytiger]

i removed ihs from my k62 and the temp dropped 4C!

[batboy]

Ok, that would make sense that Intel wanted to protect the CPU core to prevent the problems that AMD has with core crushing and corner chipping, plus it does add stability for the heatsink since the IHS has a bigger contact area. What Pinky said about fewer layers between the core and heatsink makes sense too. Call me cautious or conservative or whatever, but I'd still be interested in hearing if people consistantly find removing the IHS helps cooling. I've read some hardware reviews last summer that claim the IHS does helps cooling a little.

[[OC]_SR20DE]

Pinky, good explanation there.. yes, that is correct. Also, the reason why P4 & P3S/T-Celey use IHS is for the protection from overheating when the HSF department stops working properly. What I mean by that is that we are talking about those standard factory HSF.. those general users (not insane overclockers like us). When the Fan on the their factory HS stops working for whatever reason.. or if the HSF falls off!! :eek: , this IHS will come handy.. providing the temporary heat dissipation because having this On helps to spread out heat than as if there was no IHS. If there was no IHS and the HSF goes wrong, it would cause more danger. IHS is for the safety purpose, mainly headed for the general public users who don't know much about cooling even though system is not overclocked. So yes, having IHS is good. It is part of nice little features that these new chips come with. But for us overclockers, we would want that thing off. so we can apply more powerful HSF to get as much heat out of that core as possible. There were some discussions about this necessacity of IHS at Anandtech also. Hope this helps.. :)

[Yodums]

Me 3... I think Pinky has a good explanation on why they added the IHS there to reduce cores being crushed..

[[OC]_SR20DE]

Originally posted by Yodums
Me 3... I think Pinky has a good explanation on why they added the IHS there to reduce cores being crushed..

yes.. but the more of reason why they probably did this is to help "reduce" the temperature on the core. The major purpose of IHS is to spread out heat instead of holding the heat. Once again, having IHS On is good for the general users who don't overclock, who don't use high-end aftermarket HSF, besides, they don't even know how to upgrade to a high-end HSF anyway. They would say.. "Huh? HSF? what is that?" heeheehee.. :beer: :burn:

[ol' man]

If this IHS is so great why does intel leave it off on their new i845 chipset here

http://www.watch.impress.co.jp/akiba/hotline/20010825/image/nep4.gif

and replace it with a very large HS here.

http://www.watch.impress.co.jp/akiba/hotline/20010825/image/s81.gif

Seems to me if this IHS was so great at removing heat from a core they would have kept the IHS on the i845 chipset like they have been doing for how many years now. When is the last time you got to see the chipset core of any mobo. Never until this one that I can remember. If you were to peal the IHS off all your chipsets I bet you would find this lurking under neath.

[ol' man]

Originally posted by batboy
Are we sure this will help cooling? Intel put those IHS on for a reason and I'm sure their engineers did a lot of research prior to deciding to add the IHS. I thought the main reason was to allow better contact with the heatsink which would allow better cooling. I'd be interested in before and after temps measured in a consistant manner.

One other thing, those engineers knew what they were doing too when they multi locked all their cpu's too. DOesn;t mean it is good for OC'rs. Let me check the forum name again..... Yep says this is a OC'n forum. I could be wrong though. Anyone else here like to keep their chip as cool as possible and squeeze the max amount of hurtz out of it. I must be the only one;)

[[OC]_SR20DE]

Have you read my last two replies? If you read every single words of my last two posts, it explains "all" there.

I said, having IHS On is good for those Non-overclockers. Intel is looking toward the average users, not overclockers. How many overclockers are there in the globe compare to the general public users?

Having IHS Off is for the overclockers, that is if you're going to overclock those cores.

To be clear, I am not against removing the IHS. I am the "supporter" of removing the IHS. Hope u understand what I mean. :( YES I SUPPORT REMOVING IHS!!!!!!!!!!!!!!!!!!!!! Hope your last reply was refering to others, not me cauze im on your side.

[[OC]_SR20DE]

Originally posted by ol' man


One other thing, those engineers knew what they were doing too when they multi locked all their cpu's too. DOesn;t mean it is good for OC'rs. Let me check the forum name again..... Yep says this is a OC'n forum. I could be wrong though. Anyone else here like to keep their chip as cool as possible and squeeze the max amount of hurtz out of it. I must be the only one;)


Wrong. I am in it too.
And yes, this is overclockers forum and I agree with you.

[ol' man]

Hey all in fun:D

[[OC]_SR20DE]

Hell yes :D :o :p :) :beer: :burn:

[CrystalMethod]

Next question...
How the heck do you get the thing off without damaging the CPU?

[Yodums]

Originally posted by CrystalMethod
Next question...
How the heck do you get the thing off without damaging the CPU?

Well from reading ol man's posts..

You would get the thinest razor blade you can find..

And go from the corners and push in but not slice...

Then turn the blade to the side push in and then turn to the corner and push in..

Continue.. And I think you can wiggle the thing off at the end..

[ol' man]

Originally posted by Yodums


Well from reading ol man's posts..

You would get the thinest razor blade you can find..

And go from the corners and push in but not slice...

Then turn the blade to the side push in and then turn to the corner and push in..

Continue.. And I think you can wiggle the thing off at the end..

Don't wiggle the IHS. Once you cut all the way around it it should lift right off no problem. Who ever asked this question should look at the link given above.

[Yodums]

Originally posted by ol' man


Don't wiggle the IHS. Once you cut all the way around it it should lift right off no problem. Who ever asked this question should look at the link given above.

Oh.. So your pushing the blade to the middle?

[Pinky]

Originally posted by Yodums


Oh.. So your pushing the blade to the middle?

No, just a 1/4" in around the edges... don't want the blade to come into contact with the core in the middle...

[ol' man]

Originally posted by Pinky


No, just a 1/4" in around the edges... don't want the blade to come into contact with the core in the middle...

Right! Only go in far enough to cut the glue. No further, as hitting the side of the core would be bad for its health possibly. LIke I said I did this procedure with out getting a scratch on my chip!

[Turbo]

Ol'man, thanks for your input. Also thanks to everybody else who shared their $0.02. I had allready read the other thread you reffered to ol'man, and I've seen almost exactly the same explanation on a different forum, so I'm very convinced about 2 things: 1) it will work and 2) it will lower temperatures. Now the only real concern I have is whether the die will have sufficient contact with the original Intel heatsink after the procedure. One guy over at the Abit forum didn't think he had (he couldn't boot), and the guy who performed the IHS procedure was using a Thermal Take cooler. When I get my Swiftech MCX478 it won't be a problem, but it may not get here anytime soon, unfortunately.

Now how thick is that IHS anyway? I may just have to try it out later today. I mean somebody has to try out these things right? I guess I may as well be the first then, if nobody else tried it on a P4 in here yet :)

[batboy]

Turbo, according to your signature, you have that P-4 running at 2.1 gig. That's a really great overclock. Personally, I think you're close to max for that chip, so you might be wise to wait and try the new CPU cooler first before considering anything extreme or radical. No shame in being cautious.

I have heard conflicting results about those IHS, some say it helps with cooling and a couple individuals here say it hinders cooling. One of those guys probably has never done this procedure. Sounds like Ol' Man had good luck removing it from a Tualatin (but not a P-4). Will everone have the same results? Who knows? I'm a scientist and those in my profession prefer lots of data. There simply is not much data yet to support the claim that removing the IHS on a P-4 helps cooling. Maybe it will and maybe it won't. Common sense and logic says that it might, if you're very careful and don't damage the core.

Hmmm, I looked at the name of this site again and it says overclockers.com not radical-overclockers.com or extreme-overclockers.com, so I think there is still a place here for us cautious and conservative overclockers. I've been overclocking for more than 4 years now and have never damaged or fried any components. I like that track record.

[ol' man]

Originally posted by batboy
Turbo, according to your signature, you have that P-4 running at 2.1 gig. That's a really great overclock. Personally, I think you're close to max for that chip, so you might be wise to wait and try the new CPU cooler first before considering anything extreme or radical. No shame in being cautious.

I have heard conflicting results about those IHS, some say it helps with cooling and a couple individuals here say it hinders cooling. One of those guys probably has never done this procedure. Sounds like Ol' Man had good luck removing it from a Tualatin (but not a P-4). Will everone have the same results? Who knows? I'm a scientist and those in my profession prefer lots of data. There simply is not much data yet to support the claim that removing the IHS on a P-4 helps cooling. Maybe it will and maybe it won't. Common sense and logic says that it might, if you're very careful and don't damage the core.

Hmmm, I looked at the name of this site again and it says overclockers.com not radical-overclockers.com or extreme-overclockers.com, so I think there is still a place here for us cautious and conservative overclockers. I've been overclocking for more than 4 years now and have never damaged or fried any components. I like that track record.

Well batboy you can always reattatch it if you are not satisfied with it. Dude common sense says it helps period. I could do the shi#in' math for ya but I hope you understand why it cools better with out it. I am not for sdomeone going and ruining their cpu but if it is done consistently the way I showed and explained it will no hurt the cpu. For **** and gigles I reapplied my IHS with rubber cement and ASII and the temps were the same as before I took off the IHS so if one want to reattach it like this it will work. I seriously doubt though after he takes it off he want to reattach it cause his temps will go down.


Turbo you are not the first to take it off, there is another guy at hardforum that has taken his off too and he noticed the same affect which was cooler temps. Good luck and believe what you want.

[[OC]_SR20DE]

**Post deleted by Moderator**

Dammit, you know better than that! You, being a Senior, should know we do not tolerate flaming and name calling like this!

Mr B

[Turbo]

Originally posted by batboy
Turbo, according to your signature, you have that P-4 running at 2.1 gig. That's a really great overclock. Personally, I think you're close to max for that chip, so you might be wise to wait and try the new CPU cooler first before considering anything extreme or radical. No shame in being cautious.

Yeah I need to change that signature now. The thing is I "won" this 1.5 Ghz P4 in the Abit P4 giveaway contest. So it didn't cost me anything. Besides I'm a very curious person by nature. As I discovered my 1.5@2.0 gives me pretty much the same benchmark results as my 1.7@2.1, I've decided to keep the 1.5 and sell the 1.7. I'm going to get a Northwood sometime next year anyway.

And as I couldn't wait any longer I did take off my IHS. My system is up and running again at the same speed as before. Here are my initial findings:

1) The IHS is really thick. It's ~2 mm on the sides, and in the middle it's at least 75% of that. The blue die almost doesn't "stick up" from the green plate. That means that without it, the stock Intel retail heatsink barely touches the die. I had to apply much more thermal grease than I usually do. This almost defeats the purpose of removing the IHS. This could indeed be an issue for a lot of people, and for this reason alone I would not recommend removing the IHS if you don't plan on getting an aftermarket HQ HS + fan

2) Temperature is indeed lower without the IHS. I suspect that if the heatsink was actually having better contact, the temperature drop would be even bigger. In my case the max. temperature under load has dropped 4°C/7°F. This is not measured scientifically in any way, just a readout of winbond software using cpu stability test.

3) Overclockingwise nothing has changed. I get errors in cpu stability test at exatly the same fsb speed, not even 1 Mhz more or less :(

4) CPU still needs the same Vcore as before. Lower it 0.05V and it immediately gives errors in Windows.

So, I think my CPU has reached it's limit here, and I can live with that. I did suspect my ram may have been at fault, but I've tested them in DOS using Memtest86 and DocMem (from simmtester) and they pass up to at least 3 Mhz more. I'm still going to get a better cooler anyway, but my conclusion for now is, that if you're using a retail Intel heatsink and fan, don't bother. I guess I just NEEDED to know for sure :D

[Yodums]

Whoa nice overclock..
Rather than sell the 1.7 give it to me hehe...

[lennytiger]

Thats a nice a oc....

Maybe you guys with fast cpu's can join us "folders"

please go to the Folding@Home board and look at the new member sticky...

We'd love to have more folders with cpu's like your on our team!

Thanx
lenny

[Yodums]

Originally posted by lennytiger
Thats a nice a oc....

Maybe you guys with fast cpu's can join us "folders"

please go to the Folding@Home board and look at the new member sticky...

We'd love to have more folders with cpu's like your on our team!

Thanx
lenny

I prefer Seti ehehehe

I tried Folding.. Bah it doesn't even show your % to finish your WU..

Unless I didn't check an option to allow it..

[funnyperson1]

umm yodums why would you need a percentage when all Work Units are divided into 100Frames??? Right now i have finished 70 frames, so i am 70% done.....

[Turbo]

Originally posted by Yodums
Whoa nice overclock..
Rather than sell the 1.7 give it to me hehe...

I bet you would like that, wouldn't you :D I need the money for a new scanner and the Swiftech cooler, whenever it gets here :(

[ol' man]

Originally posted by Turbo


Yeah I need to change that signature now. The thing is I "won" this 1.5 Ghz P4 in the Abit P4 giveaway contest. So it didn't cost me anything. Besides I'm a very curious person by nature. As I discovered my 1.5@2.0 gives me pretty much the same benchmark results as my 1.7@2.1, I've decided to keep the 1.5 and sell the 1.7. I'm going to get a Northwood sometime next year anyway.

And as I couldn't wait any longer I did take off my IHS. My system is up and running again at the same speed as before. Here are my initial findings:

1) The IHS is really thick. It's ~2 mm on the sides, and in the middle it's at least 75% of that. The blue die almost doesn't "stick up" from the green plate. That means that without it, the stock Intel retail heatsink barely touches the die. I had to apply much more thermal grease than I usually do. This almost defeats the purpose of removing the IHS. This could indeed be an issue for a lot of people, and for this reason alone I would not recommend removing the IHS if you don't plan on getting an aftermarket HQ HS + fan

2) Temperature is indeed lower without the IHS. I suspect that if the heatsink was actually having better contact, the temperature drop would be even bigger. In my case the max. temperature under load has dropped 4°C/7°F. This is not measured scientifically in any way, just a readout of winbond software using cpu stability test.

3) Overclockingwise nothing has changed. I get errors in cpu stability test at exatly the same fsb speed, not even 1 Mhz more or less :(

4) CPU still needs the same Vcore as before. Lower it 0.05V and it immediately gives errors in Windows.

So, I think my CPU has reached it's limit here, and I can live with that. I did suspect my ram may have been at fault, but I've tested them in DOS using Memtest86 and DocMem (from simmtester) and they pass up to at least 3 Mhz more. I'm still going to get a better cooler anyway, but my conclusion for now is, that if you're using a retail Intel heatsink and fan, don't bother. I guess I just NEEDED to know for sure :D


Whoa dude, you are saying that your HS isn't flat against the core??????????

If that is true then that thing should be hotter than a sunsa. I know if my HS is only off by a little bit then it throughs my temps way out of wack. My pal was top heavy so the top of the HS wasn't touching the cpu core and it through off my temps by about 10 deg. C. So you have taken the HS off to see what kind of contact you are getting?????????? You are saying this is a thick layer????????????? If this is the way it is then you shouldn't have it like that. I don't know maybe simply put your IHS back on. I am saying with my experience that if my HS isn't flat against my core then it overheats and I mean right now. ASII has a thermal conductivity 1/20th of copper and about 1/10th of aluminum. I would put it on the order of playdoe if it is thicker than 1/32 inch. The thermal conductivity decreases hardcore with increasing thickness.. If what you are saying is true then the IHS should be raisd in the middle a little bit then. Mine was just the opposite and indented. You better check to make sure. You wouldn't want to be giving out wrong info about this.

[lennytiger]

You can just put a small flat piece of copper between thie core and the hsf this increases thermal conductivity and drops the temp

[ol' man]

I reread your post. You are saying that the die does stick up a little bit then right. Well that is all you need! There is no reason to use more ASII than before. I always put it on my HS any way and not on my core. I can get thinner layers that way. You want your thermal paste to be as thin as possible while still being there. The thinner the better. One nice thing about having the core so close to the base is that it sounds like you should never have to worry about chipping your core. You got almost a pefect built in shim then. If your temps dropped like you said they did then that is a good thing no matter what. I know if I could get a 4 deg. C drop in temps I could hit higher. That cause I am around 43 deg. C @ load. If I open the window and let the temps drop in my room about 5 deg. C I can hit higher so I know this to be true.

You have been saying you have a 1.5@ 2GHz. You ever think maybe your other stuff is maxed out like PCI or something else? That is a pretty good OC dude.

I just did the math and I figured you are running a 33% OC which is real good but your PCI speeds are around 44.4MHZ which is the crapping out point usually for extreme boards, funny it didn't so earlier. Fact is at around 38MHz PCI my TUSL-2 would start crapping out but ABITs are known I guess to be able to handle the higher bus speeds than ASUS can but I think 44MHz PCI is the max dude for about any board.

[ol' man]

Originally posted by lennytiger
You can just put a small flat piece of copper between thie core and the hsf this increases thermal conductivity and drops the temp


Ummmmmmm, dude this is the reason he took away the IHS in the first place.It was so he would have only one thermal junction not two;) And if he is telling th truth then it did drop his temps but as I stated with a PCI of 44.4MHz you can't expect much higher than that no matter what, unless of course they are making new boards with a 1/4 divider now for these PIV's.

[Yodums]

Originally posted by funnyperson1
umm yodums why would you need a percentage when all Work Units are divided into 100Frames??? Right now i have finished 70 frames, so i am 70% done.....

Bah I guess I never figured that out..

Anyways I like the point of Seti better.

[ol' man]

Originally posted by Yodums


Bah I guess I never figured that out..

Anyways I like the point of Seti better.


BAH.........:D

I quit folding for now cause the way they made it sound at Stanford was that they wanted to learn how to simply make totally new polymeric prion proteins which to me is not finding a cure for alzheimers. I compare it too creating new viruses that cause cancer. And if you think viruses don't cause cancer think again. Instead of making new polymeric prions they should be concentrating on the 3 or 4 prions that cause alzheimers, mad cows disease and KURU. They have even found radical prions in the deer and elk around where I live. That is some scary shi#. Nope these folk want to figure out how to make totally new prions. Buncha BS if you ask me. Until someone shows me that they are simply trying to find a cure then I may fold again but I am not going to help create new and possibly more deadly prions.

I may do SETI though now.

[Turbo]

Originally posted by ol' man
SNIP
You have been saying you have a 1.5@ 2GHz. You ever think maybe your other stuff is maxed out like PCI or something else? That is a pretty good OC dude.

I just did the math and I figured you are running a 33% OC which is real good but your PCI speeds are around 44.4MHZ which is the crapping out point usually for extreme boards, funny it didn't so earlier. Fact is at around 38MHz PCI my TUSL-2 would start crapping out but ABITs are known I guess to be able to handle the higher bus speeds than ASUS can but I think 44MHz PCI is the max dude for about any board.

No worry's mate ;) The Abit TH7-II also has a agp/pci divider at 1/2 and 1/4 so at 133 fsb everything is right where it should be. Agp is at 66.5 and pci is at 33.25. That's why I'm keeping this cpu instead of the 1.7, because for some reason it would only do a cold boot with the agp/pci at 2/3 and 1/3 respectively, which left agp and pci at around 84 and 42 Mhz. Not good :(

How do you like this memory benchmark? I know I love it :D

[funnyperson1]

that is awesome, youre probably the only person i know with a P4 at 133...

[ol' man]

Wow those are nice but my 1.6 celery still with super low memory through put gives your PIV a run for its money in some apps. Memory through put isn't every thing. None the less it is impressive.

Well you still need to answer me if your HS really isn't making good contact with your die cause you should experience a temp increase then!

[Turbo]

Originally posted by ol' man
Wow those are nice but my 1.6 celery still with super low memory through put gives your PIV a run for its money in some apps. Memory through put isn't every thing. None the less it is impressive.

Well you still need to answer me if your HS really isn't making good contact with your die cause you should experience a temp increase then!

Well, bring it on ol'man :D :D I know very well how good the old coppermine core is. My son is using my old system, an Asus P3B-F with a PIII 1Ghz(eb)@1125 Mhz, 150 fsb, and it's still a great performer.

I'm pretty sure the HS is making sufficient contact. I do believe it could be better though. I may have led you to believe that I put in a load of thermal grease, and that's not really the case. It was pretty obvious that the HS wasn't 100% level, but it was pretty close. The cpu needs 2.2V to run at 2.0 Ghz 100% stable, and right now after several hours of typing, testing, benching etc. it's hovering around 42°C/107°F while writing this. Note that the bios Vcore setting doesn't respond to actual Vcore, this is the same with all i850 chipsets. The Vcore usually hoveres around 2.00-2.05V. These i850 chipsets are nothing like the BX/815's.

[ol' man]

Dude I can't believe you would run your chip with out the HS being flat on the core. I wish you would explain this a little better. I mean come on man we are not new to oc'n and we all now if the HS isn't touching perfectly flat then you will get very high temps and instability problems. Please don't BS us!

[ol' man]

Here is a pic of a naked PIV. Is this what your chip looks like? If so and you are not getting good contact I would definatly do something about it. Funny thing about it is this guy here is not having any problems with contact supposedly. He has a 1.7Ghz at 1.97GHz.




You know I noticed something when I removed my IHS, I found that my HS had to be turned around the right way where they milled it out for the socket like many HS's are. With the IHS you don't need to worry about that but with out the IHS you certainly do just like the older PIII's even. You couldn't have done that could you have?

[Turbo]

Originally posted by ol' man
Here is a pic of a naked PIV. Is this what your chip looks like? If so and you are not getting good contact I would definatly do something about it. Funny thing about it is this guy here is not having any problems with contact supposedly. He has a 1.7Ghz at 1.97GHz.


You know I noticed something when I removed my IHS, I found that my HS had to be turned around the right way where they milled it out for the socket like many HS's are. With the IHS you don't need to worry about that but with out the IHS you certainly do just like the older PIII's even. You couldn't have done that could you have?

Nah, the socket 478 Intel heatsink is flat. It isn't milled out like many socket A & socket 370 heatsinks are (like my Alpha PEP66), so you can't really get it in wrong. The thing is, that the edges of the retail heatsink rests on the mounting bracket (which is fixed on the MB) and thus "barely" touches the die, as it is much "lower" without the IHS. When I get my Swiftech it will no longer be a problem, and in all honesty it really isn't a problem now either, as the temperature consistently is lower than before, and the overclocking potential is identical. All I need now is my copper Swiftech MCX478 with a much better C/W value. I think it may well grant me at least 50-80 more Mhz.

Btw. here's what my P4 looks like.

[ol' man]

Sweet pic:D

Yeah dude I get what you are saying, that is a big problem and if it were making perfect contact it may improve your temps. In fact it would. So when you lift up your HS it looks like it isn' touching the core all the way like there is a gap or does it touch it all the way? If it isn't touching the core then you will have one side of the core where the thermal paste is like built up. You sure the swifty will touch the whole core too? Will you still be able to hit the core perfectly with that. If not I would mill it out so it will or similar. I must say you are nutz:D I mean you tell people it drops temps but that isn't good cause you can't get the HS to fit right???????? I can clearly see there is no long term reason why the HS should not sit perfectly flat on the core.

[Turbo]

LOL yes I am nuts :D Having tweaked hardware for as long as I can remember must have taken it's toll on my sanity.

With only a very small amount of paste on the HS, and losely setting the HS down, only about 2/3's of the die touched the HS. With a little more "grease" it seemed ok. If you have ever put on an Intel s. 478 HS you will notice how tight it fits. I mean the MB flexes, it's that tight. So I'm not REALLY nervous.

The Swiftech will use 4 screws, that can be independently adjusted. I'm sure it will work beautifully :p

And for the record, I don't think you guys are fools. That was a joke, right? :) I'm tired, I'll be back later....

[ol' man]

All right man you got me intrigued as to why your HS is not fitting right so I have included the bottom pic to help explain what I am speaking of. If you look at the socket 370 below you will notice I have labeled something shi# head. This is the problem I encountered when I removed my IHS. I then had to be more careful which way I put my HS on and to make sure the milled side was there which I hope is self explanatory.

It appears the sock 478 board has this and since it does is it higher than the actual core?

[ol' man]

Now here is another pic of the close up of the sock478 socket.

Once again is your cotton picken:D HS hitting the shi#head I mentioned above? If so you best do something about it. When My HS is at an angle like you say then it causes major problems. You say 2/3's of the HS is touching the core.If it is at an angle then only about .000001% is adhered to it the way it should. Well since your IHS is gone and you are only getting about .000001% of you HS adheared which is the corner what do you suppose you temps will be like once you get that taken care of???????

Like I said above I noticed when my HS is at an angle it increases temps 10 deg. C at least. I have seen higher but never made it since it craps out easlily around 50 deg. C.

Now lets get to work and fix this problem shall we:D

[ol' man]

And for the record, I don't think you guys are fools. That was a joke, right?

Yes that was a joke. Maybe I shouldn't have said it but it was a joke:D

[ol' man]

r

[ol' man]

UUUUUUUUUUHHHHHHHHHHGGGGG I hate that shi#. I right a long paragraph and then when you try to send it it makes you log in and you lose all you have written. Anyone else know what i am talking about here. ACKKKKK.

I was trying to say that your cpu is throttling down and that is most likely why you are seeing lower temps. The 2GHz PIV at default vcore pushes 75w. If your chip is at 2GHz but 2v too then you should be even higher around 80~90w. Thermal paste is not thermally conductive at all when it is thicker than 1/32" like I have told you above and will not offer a good path through which the heat should flow.

Here is a pic below showing you the thermal output of a PIV at 2GHz. You can goto this url...

http://support.intel.com/support/processors/sspec/p4p.htm

To get all the thermal specs of PIV's at a sock 478 status.

Now to breath in, post, and exhale..AHHhh much better...

[lennytiger]

I had an old clery 667@750 in the uk that i sold off and it had a core that really stuck out a lot, i wish i had a pic

[Turbo]

Darn ol'man you're all over the place. I took the HS off, and cleaned it of the residue of the black silicone/glue. The imprint of thermal grease on the HS indicated that the HS fully touches the die. I've now put on a little less thermal grease, than before, but that has increased the max temperature under load by 1°C/2°F, but that may very well be because of an increased ambient temperature. The die does stick up over the socket, ever so slightly. I've come to the conclusion, that the HS has reached it's thermal ability. With the Vcore at 2.2V and the CPU at 2.0 Ghz, according to the Intel thermal specs, this chip could well generate well over 100W of heat :eek: And I can tell you, that taking the HS out just after doing a cpu-stability test for 30 mins. it was indeed warm to the touch ;) I'm now confident that it is making sufficient contact with the standard Intel HS, and I'm going to keep it like this, until my Swiftech arrives. It will be really interesting to see how it will fare. Time will tell.

Thank ol'man for your involvement, you really dug in there, didn't you :D :D I'm not halfway shure I answered all your questions, but I'm confident you ask me again if I forgot something :p

[ol' man]

Well that is good if it fits right;)

[[OC]_SR20DE]

Originally posted by [OC]_SR20DE
**Post deleted by Moderator**

Dammit, you know better than that! You, being a Senior, should know we do not tolerate flaming and name calling like this!

Mr B

:confused: .... http://www.ytec.d2g.com/emoticons/cry2.gif