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

new boy · 13 May 2012 at 18:29

f5 f5 f5 f5

Jokester · 13 May 2012 at 18:36

Before:-

http://img441.imageshack.us/img441/555/ihson.jpg
Done at an ambient temperature of 17.5degC, peak temperature 79degC

After

http://img18.imageshack.us/img18/953/ihsoff.jpg
Done at an ambient temperature of 18.5degC, peak temperature 69degC

So an effective drop of 11degC removing the IHS, but I suspect I could possible get a bit more if I sort the mount out so that the core is taking all the pressure.

jeff6475 · 13 May 2012 at 18:42

good job but i would see what you get with the IHS back on b4 you mod the socket to make block fit better

Aurhinius · 13 May 2012 at 18:54

Jesus intel have some explaining to do. Article on guru3d from japanese website confirms your almost 20 degree drop. What were they doing? Trying to get to market too quick with large volume or purposefully limiting overclock by manipulating thermals?

Really great work mate

pastymuncher · 13 May 2012 at 18:55

Jokester, do you have the original mounting kit for the EK Supreme HF or did yours come with the EZ mount kit? I am pretty sure mine (original kit) does'nt have the pillars yours is bottoming out on.

-Mike- · 13 May 2012 at 18:55

Holy .... ! Can you try with decent paste ad the ihs back on?

jeff6475 · 13 May 2012 at 18:55

just cutting costs,they have no competition with AMD so can do what they want

Jokester · 13 May 2012 at 18:57

pastymuncher said:
Jokester, do you have the original mounting kit for the EK Supreme HF or did yours come with the EZ mount kit? I am pretty sure mine (original kit) does'nt have the pillars yours is bottoming out on.

Aye it's a new one with the EZ mount kit, my old one is the EZ as well.

Trying it with the IHS back on, but it's already upto 75degC.

Edit: Already the same temp as stock.

uk_unbeliever · 13 May 2012 at 18:58

I'm a bit out of the loop with Ivybridge, is it that temps shoot up after about 4.5ghz or is it that loads of extra volts are needed at this point (and then temps shoot up anyway)?

Jokester · 13 May 2012 at 19:11

It's actually hotter now than it was before, I think it's only going to be worth taking the IHS off if you're will to run without it, otherwise there's no real gain in doing so.

no_1_dave · 13 May 2012 at 19:14

I did this with my old opteron 64

moogle · 13 May 2012 at 21:52

The other problem with this IHS mount is that after a few years the paste will dry up and give really bad temperatures.

I had to get my AMD Athlon X2 RMA'ed just near the end of it's 3 years because of this, so I can imagine it might happen to IB users out of warranty.

Sasso · 13 May 2012 at 22:04

moogle said:
The other problem with this IHS mount is that after a few years the paste will dry up and give really bad temperatures.

or straight away because it's rubbish.. looking from the internal pics people have posted it already seems like it's turned to clay :\

Reaper123 · 13 May 2012 at 22:38

Jokester said:
It's actually hotter now than it was before, I think it's only going to be worth taking the IHS off if you're will to run without it, otherwise there's no real gain in doing so.

Did you also replace the TIM like they did in that japanese website?

pmc25 · 13 May 2012 at 23:04

Sasso said:
or straight away because it's rubbish.. looking from the internal pics people have posted it already seems like it's turned to clay :\

Exactly. You can be sure this will be a bigger headache for Intel in the channel than it will be with a few OC'ers. Are potentially big customers wanting to order a bunch of workstations going to be happy knowing that the TIM is about as effective as silly-putty immediately, and no doubt less than useless after 6 months to 2 years?

My guess is no, not at all.

They certainly won't pull this **** with the Ivy Bridge Xeons, as their potential clients would just laugh at them.

Sasso · 14 May 2012 at 00:03

Jokester... what thermal paste are you using? what application method did you use?

I've also noticed that you have removed the CPU "clamp" from the motherboard... was wondering how it would be possible to apply the block directly onto the CPU as this would be in the way xD

what is the CPU sitting like? I'm slightly worried that because you are applying the load directly onto the silicone die that it could damage it/warp the chip with too much pressure, also possibly cause it to curve so that the outer contacts wont be touching?

really thinking about doing this when i get my ivy chip (after running it as normal for a bit to make sure it all works before i butcher it) but wondering what are your, and others, opinions on these points.. I'll be getting the H80 so not sure what it's mounting is like.

Locky · 14 May 2012 at 00:25

What a joke from Intel, really is a poor show from them.

jallan · 14 May 2012 at 02:46

Surely if the paste becomes ineffective in the next few months then this would fall under the warranty that the chip is ineffective and manufacturing issue has reduced its life span? I mean we all know our paste begins to cook off after a few months so what can we expect Intel to do when a mass of Ivy bridge chips are beginning to sky rocket on temperatures no matter what you do?

pmc25 · 14 May 2012 at 03:12

jallan said:
Surely if the paste becomes ineffective in the next few months then this would fall under the warranty that the chip is ineffective and manufacturing issue has reduced its life span? I mean we all know our paste begins to cook off after a few months so what can we expect Intel to do when a mass of Ivy bridge chips are beginning to sky rocket on temperatures no matter what you do?

They can claim that the operational limit of the chip (as stated by them) is 105C. Of course, we all know the chip isn't going to last very long at all if it's consistently at 80C plus, at least not without massive degradation.

Jokester · 14 May 2012 at 04:50

Reaper123 said:
Did you also replace the TIM like they did in that japanese website?

I tried the thin layer method with AS 5 at first but wasn't convinced of the quality of the contact so I had to go with the blob method in the end. I think re adding the IHS isn't a great idea as you're just adding another TIM interface with variable results.