Whose P4 cpu IHS is concave?

FIZZ3 · Dec 25, 2002

I was just curious because I suspect mine is. I took off my waterblock in order to move things around and I noticed that there is a distinct circular imprint on the thermal paste. I don't think it is the waterblock as it was somewhat off-base.

I guess I'll be doing some lapping soon...

Overclocker456 · Dec 25, 2002

carefull lapping made my temps worse by 4-5C..

Daemonfly · Dec 26, 2002

My 1.6A was slightly concave, as well as my old 366 Celerons. It's quite a common occurance

If anything, it would work better if it were slightly convex due tot he center of the core now in contact with the HS/WB.

I slightly lapped my 1.6A as one edge was nicked & stuck up a bit. Once i finish my waterblock, I'll fully lap the P4.

FIZZ3 · Dec 26, 2002

Overclocker456 said:
carefull lapping made my temps worse by 4-5C..

How would you explain that?

FIZZ3 · Dec 26, 2002

Can O' Beans said:
My 1.6A was slightly concave, as well as my old 366 Celerons. It's quite a common occurance If anything, it would work better if it were slightly convex due tot he center of the core now in contact with the HS/WB.

I slightly lapped my 1.6A as one edge was nicked & stuck up a bit. Once i finish my waterblock, I'll fully lap the P4.

How do you determine where the surface is imperfect? Do you take measurements?

SHODAN · Feb 8, 2003

I have a Pentium 4 that seems to exhibit this problem. Is lapping suggested or not? I'm hoping more people have tried lapping their P4 since this thread was started. Thanks in advance.

Rio71 · Feb 9, 2003

hi.
... or remove the ihs.

i run a p4m, this come without ihs "out of the box"
see also tread "P4 Mobile Whats the score? "

FIZZ3 · Feb 9, 2003

In the mean time I have heard bad things about handlapping. It seems it can often result in a surface that isn't flat.

As for removing the IHS- it can be done, but there is an increased risk of damaging the core. I have not tried this myself.

SHODAN · Feb 9, 2003

Rio71, I'm a bit too clumsy to be comfortable removing the IHS, so that pretty much out of the question.

FIZZ3, I realize that hand lapping is rarely excellent, but I belive the most common error is producing a convex surface, due to the edges of the surface being lapped plowing into the lapping block; if this is the case, conservative lapping should improve a concave IHS. What do you think?

FIZZ3 · Feb 10, 2003

SHODAN said:
FIZZ3, I realize that hand lapping is rarely excellent, but I belive the most common error is producing a convex surface, due to the edges of the surface being lapped plowing into the lapping block; if this is the case, conservative lapping should improve a concave IHS. What do you think?

Sounds good in theory, but tentative results lead me to think it's not so simple. Some of BillA's pictures show highs and lows at unpredictable spots on a heatsink's surface after hand lapping. I am not sure whether the sum total of surface characteristics and hand lapping error would produce a positive result.

looktall · Feb 10, 2003

the best method i've used for lapping, to determine the flatness of the core/IHS/hsf whatever, is to get a permanent marker and draw a star on the object to be lapped.
corner to corner, side to side.

as you start lapping the high spots will wear down first. by having a look every few minutes, you'll see the start slowly disappearing.
keep going until the star has totally gone.

to ensure a flat surface, use a glass surface. a mirror works very well.

Ralf Hutter · Feb 10, 2003

It's real easy to see how bad an Intel IHS is because they're copper underneath the plating. As you start to lap it you'll immediately see the copper start to show where there's high spots.

dbwillis · Feb 11, 2003

Ive lapped my 1.8A, 2.0A and 2.4B, all had high spots, and temps got around 2'C average lower.

I tested out a 1.7M and went back to the 2.4B without the IHS though, seems to be about the same temp as the lapped IHS, at least forme

FIZZ3 · Feb 11, 2003

Well BillA has equipment that can show unevenness in surfaces really accurately. The results were surprising and his advice is not to hand-lap anything! That is why I have become reluctant to do it myself. However, in the case of 'huge' malformation like this on the IHS, I am in doubt still. Surely it's better to do something about it and have small imperfection than this stock situation...

speedy4500 · Feb 11, 2003

Well, I'd reccommend against lapping the IHS. Intel says that the IHS is supposed to be concave at room temp, and as it heats up, it will expand to make a flatter surface. That makes excellent sense, and I'm sure they know what they're talking about.

SHODAN · Feb 11, 2003

Thanks to everyone who's posted.

FIZZ3 said:
Well BillA has equipment that can show unevenness in surfaces really accurately. The results were surprising and his advice is not to hand-lap anything!

speedy4500 said:
Well, I'd reccommend against lapping the IHS. Intel says that the IHS is supposed to be concave at room temp, and as it heats up, it will expand to make a flatter surface. That makes excellent sense, and I'm sure they know what they're talking about.

Could each of you please link to your source for this information?

Also, if water-cooling, don't you think the IHS may never heat up as much as Intel anticipates? In this case, some lapping may be good.

FIZZ3 · Feb 12, 2003

SHODAN said:
Thanks to everyone who's posted.

Could each of you please link to your source for this information?

Also, if water-cooling, don't you think the IHS may never heat up as much as Intel anticipates? In this case, some lapping may be good.

http://www.overclockers.com/articles692/index02.asp

Intel has made the IHS mainly to prevent damage or so I hear. I don't see why lapping would be better if there was less heat either, TBH.

SHODAN · Feb 13, 2003

Intel says that the IHS is supposed to be concave at room temp, and as it heats up, it will expand to make a flatter surface.

I don't see why lapping would be better if there was less heat either, TBH.

My thought is that if the IHS is made in such a way that if deforms (flattens) as it heats up (multiple layers with differential coefficient of thermal expansion?), it probably doesn't deform (flatten) as much if heated less than the designed operating temperature.

Terry · Feb 13, 2003

Awhile back I had a 2.26 that would hit 3ghz but had problems about 3 hours into Prime95. I lapped it, looked like a mirror when I finished. It did 3.1ghz stable at 1.6v after that till I sold a couple months ago. It was concave in the middle.

FIZZ3 · Feb 13, 2003

SHODAN said:
My thought is that if the IHS is made in such a way that if deforms (flattens) as it heats up (multiple layers with differential coefficient of thermal expansion?), it probably doesn't deform (flatten) as much if heated less than the designed operating temperature.

The first quote you gave is not mine.

In any case I see what you mean with less heat meaning less deformation, but why would lapping be particularly useful in this case?

Whose P4 cpu IHS is concave?

Member

Member

Member

Member

Member

Member

Member

Member

Member

Member

Member

Member

New Member

Member

Member

Member

Member

Member

Member

Member

Similar threads