Ivy Bridge Temperatures Could Be Linked To TIM Inside Integrated Heatspreader: Report

[STaRGaZeR]

so now we have tests concluding that removing the IHS works in lowering temps and removing the IHS does nothing at all and mounting on a naked core drops temps and mounting on a naked core does nothing.

conclusion: we are right where we started, nothing conclusive can be decided.

This.

[zalbard]

We had one test with people mounting their cooler directly onto the die (about the same temps, most likely poor contact), and two tests with people replacing TIM between IHS and CPU die with positive results. How is that not conclusive?

Shamino backing this up:

temps are not as good with IHS without exertion of force by ILM on the core but with good thermal paste and remounted IHS, temps are better and may bring the CPU to cold bug

[STaRGaZeR]

We had one test with people mounting their cooler directly onto the die (about the same temps, most likely poor contact), and two tests with people replacing TIM between IHS and CPU die with positive results. How is that not conclusive?

Shamino backing this up:

When you write "most likely" in your conclusions, well, that's non conclusive. There have been quite a few test on the subject and the results are all over the place. Most of them don't say how exactly they're doing it, like "remounting" and stuff. A post with links to all of them would be nice. Or a comparison table, etc.

[mibo]

So better check it yourself

The Intel TIM is not that bad - compared to other thermal grease. I reapplied AC MX4 paste 3 times and twice the temps were worse. In the third run I did achieve the same temps as the Ivy had before...
Then I used Liquid Pro and had more then 10° better temps.

I think that thermal grease is optimized to work on surfaces with a defined/high micro roughness. Heat spreader and cpu cooler are rough - in contrast to the die. Therefore the very different temperature results we see. It just strongly depends on the thermal grease.
And I would recommend to use liquid metal.

If you can read German: http://www.hardwareluxx.de/community...l#post18877078

[STEvil]

Should try with no TIM, or as small of an amount as you possibly can use. Any reduction in metal contact area or pressure reduces thermal transfer capabilities. TIM's are an insulator most of the time..

[WangChung]

These posts saying "No AS5 without the IHS!!" make me lol. You can tell who the youngins are who never had to roll the dice with their Socket A mobile chips.

[Hondacity]

TIM's are an insulator most of the time..

but you use thermal paste right?

[The Jesus]

http://www.engadget.com/2012/05/14/i...thermal-paste/

[kuroikenshi]

maybe this sounds ridiculous? but does anyone bother to also measure room temp at the time of the test? since different people carrying out different test in different regions and seasonal environments this can also affect results (only a little though unless your home is Antarctica and you're a penguin)

[Sparky]

These posts saying "No AS5 without the IHS!!" make me lol. You can tell who the youngins are who never had to roll the dice with their Socket A mobile chips.

Or any socket A chip in general. None of them had an IHS.

[RussC]

It looks to me that INTC is doing a poor job of applying the grease also. There putting way to much on. With thermal grease, less is more. Its really hard to find the right amount. Too much, too little is bad.

RussC

[Mats]

Or any socket A chip in general. None of them had an IHS.

You got me. I had forgotten about my AQYHA 2500+ I used with AS5 and never had any issues.

[Budwise]

I remember using AS5 on a delidded Opteron 170 with a Swiftech Storm

[M.Beier]

AS5... Ohh you new guys...
AS5 wasnt on the market when socket A was launched... I didnt have Arctic Silver 5 before my Pentium 4 Northwood 800QDR, the 400 I used AS3 !!!

[STEvil]

but you use thermal paste right?

Generally the smallest dab I can manage more so to identify fissures and/or mounting issues. Beyond a couple microns most TIM's out there offer no appreciable heat transfer capabilities.

and if you're still wondering: yes, I have run with no TIM at all, bare core or IHS.

[M.Beier]

Generally the smallest dab I can manage more so to identify fissures and/or mounting issues. Beyond a couple microns most TIM's out there offer no appreciable heat transfer capabilities.

and if you're still wondering: yes, I have run with no TIM at all, bare core or IHS.

Cant recommend that on DIE directly, unless you want the DIE to die
Killed a 3700+ because I was out of TIM and stores were closed,,, Patience didnt allow me to wait....

TIM is good, but... I would agree, it might be time to buy some indium.... cutting a tiny bit and put it on DIE and squeeze down heatspreader, should make a noticable difference....

[iboomalot]

wonder how good liquid metal would work vs AS5 ??

[STEvil]

Cant recommend that on DIE directly, unless you want the DIE to die
Killed a 3700+ because I was out of TIM and stores were closed,,, Patience didnt allow me to wait....

TIM is good, but... I would agree, it might be time to buy some indium.... cutting a tiny bit and put it on DIE and squeeze down heatspreader, should make a noticable difference....

Dont rock the boat (heatsink) and there should be no problem on die

[M.Beier]

Dont rock the boat (heatsink) and there should be no problem on die

Was with a Dangerden TDX s939, those 2 screw mounting did not make things easy, no.

[STEvil]

yeah, hoses put some serious torque on the water blocks making it twice as hard to keep them on the die flat..

[[XC] Oj101]

AS5... Ohh you new guys...
AS5 wasnt on the market when socket A was launched... I didnt have Arctic Silver 5 before my Pentium 4 Northwood 800QDR, the 400 I used AS3 !!!

Who said anything about overclocking when they were launched? I still do S462

[Mats]

Oh you new guys... only using AS5 with recently launched sockets.

[UrbanSmooth]

Then I used Liquid Pro and had more then 10° better temps.

This.

[solofly]

http://hardforum.com/showthread.php?t=1693361

[STEvil]

http://hardforum.com/showthread.php?t=1693361

I face-palmed at a couple of the guys posting there.. actually logged into [H] and replied lol. I feel dirty :p