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

Jokester · 14 May 2012 at 04:55

Sasso said:
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?

It was sitting really nice, other than my suspicion that it was the mounting pillars of my water lock mounting kit that was taking the load. Any cooler with a backplate will stop the board warping.

I will add that there is a real risk of breaking your CPU doing this and will definitely invalid the warranty and wreck its resale value.

eddyr · 14 May 2012 at 08:31

Ballsy stuff. When do you plan on seeing how much futher this raises the oc headroom? Lower temps is great but it may still react similarly when pushed a litte higher and of course temps may not the only factor limiting clocks. Cant wait for the next chapter!

Zeed · 14 May 2012 at 11:39

Now this is what Overclocking is all about.
Best overclock i seen on OCUK forums in months. Ppl that think changing stuff in bios makes them hardcore overclockers are so wrong

. If i would went for ivy i would do it to.
Good Job mate.

muon · 14 May 2012 at 12:20

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.

What type of TIM is it? ALso does the IHS go firmly up against the chip? If there's a significant gap, some TIM under high heat could become runny and be poor conductors.

edit:

Just seen AS5, yeh that becomes very runny under high heat doesn't it?

Tombo · 14 May 2012 at 12:49

This is putting me off buying

-Mike- · 14 May 2012 at 12:51

Tombo said:
This is putting me off buying

Buy sandybridge

Tombo · 14 May 2012 at 12:53

I am considering it but I like shiny new things!

dreigns · 14 May 2012 at 13:00

So looking at all these problems i'll be better off getting a 2700K as i'll be able to over clock it more then Ivybridge of the same clock
All I am going to be using it for is Video converting and Games

pastymuncher · 14 May 2012 at 13:01

As a retail 2500k can be had for significantly cheaper than the 3570k elsewhere it would be where i would put my money. There is no telling how this problem will affect Ivy a year or two down the line if the TIM dries out.

eddyr · 14 May 2012 at 15:15

I dont see where the idea that as manufactured TIM drying out is coming from? ie unaltered with the IHS left on and sealed.
Is there a historical basis for it? My earliest cpu fiddling was 2006 so not too aware of before then.

PGdude · 14 May 2012 at 15:39

eddyr said:
I dont see where the idea that as manufactured TIM drying out is coming from? ie unaltered with the IHS left on and sealed.
Is there a historical basis for it? My earliest cpu fiddling was 2006 so not too aware of before then.

I don't think the IHS is "sealed". Refer tot he image on the last page and you'll see that the adhesive doesn't go all the way around the IHS. If it were sealed I'm sure there would be pressure problems leading to damage after some time.

WingZero30 · 14 May 2012 at 16:03

I wonder why intel decided to go with this TIM idea?

Surely they have way more technical data on this issue and would know which method is best for IHS integration onto cpu in the long run.

There seems to be a lack of care from Intel but still doesn't explain that it would look badly on them if this method of IHS integration was afterthought.

covenantuk · 14 May 2012 at 16:04

No idea where this drying out scare rumour is from, but if it's an issue don't you think Intel might have thought about it?

Look at the facts, clock speeds are massively conservative. They're rated at 3.4ghz but the majortiy of CPUs will run at 4.2ghz with no changes other than the multiplier.
At stock speeds the CPU barely breaks a sweat and even if (and it's an unproven "if") TIM degrades in short timeframes causing a 10C increase in CPU temperature over 3 years then for most users if won't be an issue at all as they're running at stock with max temps in the 40-50C range.

If you don't want an IB for overclocking then don't get one, but it's just hysterical to post unfounded and unproven rumours about them.

Bezzaboy1234 · 14 May 2012 at 17:12

covenantuk said:
No idea where this drying out scare rumour is from, but if it's an issue don't you think Intel might have thought about it?

Look at the facts, clock speeds are massively conservative. They're rated at 3.4ghz but the majortiy of CPUs will run at 4.2ghz with no changes other than the multiplier.
At stock speeds the CPU barely breaks a sweat and even if (and it's an unproven "if") TIM degrades in short timeframes causing a 10C increase in CPU temperature over 3 years then for most users if won't be an issue at all as they're running at stock with max temps in the 40-50C range.

If you don't want an IB for overclocking then don't get one, but it's just hysterical to post unfounded and unproven rumours about them.

http://www.youtube.com/watch?feature=player_embedded&v=yCW508yl_Eg

Skip to 5:35,

If that isn't dry and flaky then I must be blind 0.o, Might want to consider sending back your i5 Ivy.

Sasso · 14 May 2012 at 17:14

Think these rumours of seeing out were from my posts

sorry if seemed like I was stating fact. My ideas about TIM degradation are mostly speculation. This is just based on assuming it will behave like that cheap nasty white stuff we used before all this fancy metal compound stuff. I'm sure you remember trying to remove an old hsf and finding the stuff had gone completely dry and brittle

. So as I said, this is just my assumption that this white stuff they used is going to not perform/last as expected.

Plus from what I've seen from jokester and that Japanese guy it looks pretty dry already.

Nevakonaza · 14 May 2012 at 17:18

Have Intel actually commented on this yet?...ive not been following this at all.

moogle · 14 May 2012 at 17:22

Sasso said:
This is just based on assuming it will behave like that cheap nasty white stuff we used before all this fancy metal compound stuff.

We call it solder

Sasso · 14 May 2012 at 17:26

moogle said:
We call it solder

I knowwww.... :rolleyes:

.... Maybe

more that it's fancy compared xD the stuff I have has diamond in it now...... You can't say that having diamond in it isn't fancy

Bezzaboy1234 · 14 May 2012 at 17:45

Nevakonaza said:
Have Intel actually commented on this yet?...ive not been following this at all.

majnu said:
The Tech Report was able to get a comment from Intel. It's not exactly conclusive, but does shed a little light.
http://kingpincooling.com/forum/showthread.php?t=1730&page=2

http://techreport.com/discussions.x/22859

covenantuk · 14 May 2012 at 17:56

The change in TIM was because the old (and better thermically) solder had an issue where repeated heat cycling would cause it to degrade and create an air gap. That scenario is FAR worse - air is a terrible heat conductor. So people are advocating they keep using the old way? :rolleyes:

I've just removed my HSF from my 5 year old Conroe and the Artic Silver was dry and flaky. So why is that seen as in some way faulty?
In all the time since I installed that CPU I haven't noticed any appreciable difference in CPU temperatures. Until someone proves in a scientific way why the new TIM is faulty rather than just less efficient (as it probably is) I think I'm quite happy.