Intel heat spreaders suck!

[rogue1979]

Bought a Q6600 here on the forums and it's running fine at 3.0GHz stock volts.

But when I first installed it, bios said 67C at idle, ouch!!!

Reinstalled the heatsink, same thing??

Finally pulled the chip out, physically mated it to the bottom of the heatsink in my hands. With a fresh smear of thermal compound, I saw the problem.

Concave city, I mean a real fish bowl. Contact only at the extreme edges of the heat spreader, it was unbelievable.

Sand, sand, and more sand. When I got done, idle temps dropped 30C, yeah that's right, down to 37C.

Lapped my other Q6600 and three other C2D chips of various speeds. None of them were close to flat, though not as bad as the first Q6600.

The other cpu's ran between 5C and 15C cooler after I was done.

From now on I won't install another Intel cpu without lapping it first, don't care about any warranty. (I'll check for DOA first)

 

[BTRY B 529th FA BN]

1/10

 

[Ben90]

Not all of them are bad, my d0 is near perfect.

My Pentium D 820...not so much, it really doesnt matter though because i finally found something that my little smithfield can accomplish better than even a core i7 overclocked to 8 ghz... I use it for scratching my arm cuz the chips on the bottom of the processor are way scratchier due to their sheer size over anything recent... They dont make em like they used to

 

[DrMrLordX]

AMD's IHSs aren't much better.

Originally posted by: Ben90

My Pentium D 820...not so much, it really doesnt matter though because i finally found something that my little smithfield can accomplish better than even a core i7 overclocked to 8 ghz... I use it for scratching my arm cuz the chips on the bottom of the processor are way scratchier due to their sheer size over anything recent... They dont make em like they used to

Bahahah, excellent. Put it on the end of a stick and use it as a backscratcher too! Keychains be damned.

 

[yh125d]

For the record, BIOS temps are *not* idle

 

[rogue1979]

For the record, BIOS temps are *not* idle


That's true with all the power saving features factored in.

But when I overclock I turn off all that stuff anyway.

If I wanted to save electricity I'd undervolt and then turn on that stuff. With proper cooling any cpu should be able to "idle" cool without it, usually meaning you have enough cooling for load too.

It would be easy to see real low idle temps with all the power saving features on, but under load that wouldn't mean anything.

 

[BD231]

Being that I've never gone so far as to lap the base of *anything*, I'd be upset if my heatspreader came to me that way. At least you were savy enough to do something about it, I on the other hand, would have been stuck.

I'm sure I'm not the only one. I'd actually be inclined to rip the heat spreader off, but, coolers aren't designed to cool without an IHS these days.

 

[yh125d]

Originally posted by: rogue1979
For the record, BIOS temps are *not* idle


That's true with all the power saving features factored in.

But when I overclock I turn off all that stuff anyway.

If I wanted to save electricity I'd undervolt and then turn on that stuff. With proper cooling any cpu should be able to "idle" cool without it, usually meaning you have enough cooling for load too.

It would be easy to see real low idle temps with all the power saving features on, but under load that wouldn't mean anything.

No what I'm saying is that the BIOS temp is a temp above any idle temp, whether power saving features are on or off. if it was 67c in BIOS by my experience it would probably be 70-75c in P95

 

[rogue1979]

Just got an E5200 in today, only paid $49 for it.

Tested it and it worked, then checked the heat spreader.

This one actually wasn't too bad, just a little concave. Lapped it anyway and lowered the temps about 5C.

Man, that chip is the crappiest overclocker I've seen in a long time. Successfully put a 266 fsb mod on it and then preceded to look over the Intel Data Sheets (motherboard has no vcore adjustment). After some trial an error, about the only feasible vid mod was to 1.3625v, maximum according to Intel.

This chip struggled to hit 3.2GHz at that voltage, stock it only hit 2950MHz.

 

[Pneumothorax]

Originally posted by: rogue1979
Just got an E5200 in today, only paid $49 for it.

Tested it and it worked, then checked the heat spreader.

This one actually wasn't too bad, just a little concave. Lapped it anyway and lowered the temps about 5C.

Man, that chip is the crappiest overclocker I've seen in a long time. Successfully put a 266 fsb mod on it and then preceded to look over the Intel Data Sheets (motherboard has no vcore adjustment). After some trial an error, about the only feasible vid mod was to 1.3625v, maximum according to Intel.

This chip struggled to hit 3.2GHz at that voltage, stock it only hit 2950MHz.

Is it the latest R0 stepping? I got a fresh one from Amazon and it's just as bad as an overclocker. Intel is getting more "efficient" at their binning.

 

[VirtualLarry]

That's disappointing to hear. I've got an M0 in a friend's machine that did 3.75Ghz @ 1.425v (BIOS), 1.4v (CPU-Z load). But then it became unstable after six months, so clocked it down to 290FSB, 3.625Ghz.

 

[rogue1979]

Originally posted by: VirtualLarry
That's disappointing to hear. I've got an M0 in a friend's machine that did 3.75Ghz @ 1.425v (BIOS), 1.4v (CPU-Z load). But then it became unstable after six months, so clocked it down to 290FSB, 3.625Ghz.

1.42v in pretty high for a Wolfdale. According to Intel 1.3625v is the max safe voltage.

I wonder if it's degrading and showing signs of an early demise?

 

[rogue1979]

Originally posted by: Pneumothorax
Is it the latest R0 stepping? I got a fresh one from Amazon and it's just as bad as an overclocker. Intel is getting more "efficient" at their binning.

I didn't check the stepping before I lapped it, but I got it from Amazon also.