Delidded my i7-3770K, loaded temperatures drop by 20°C at 4.7GHz

[Phynaz]

All IDC has done is show that poor thermal compound is an issue on his unit. There have been reports of some Ivy Bridge CPUs responding relatively poorly to a change in TIM. This definitely shows what a poor paste job can do, but it doesn't make thermal density any less of an issue.

He hasn't shown this at all. He changed two variables, TIM and heat spreader distance from the cpu. When he puts the Intel TIM on the cooler, then we will know how "poor" the TIM is.

 

[soccerballtux]

A ridged surface (/\/\/\/\) doesn't work well because the effective surface contact is actually diminished, even though the surface area is greater.

If you had a heatsink with a base that was actually cut with ridges to fit inside of the IHS ridges, then more surface area would be great.

hm, that's an interesting idea!

 

[Idontcare]

So I pulled off the H100 and unmounted the IHS to move on with the next test - using the Intel CPU TIM from under the IHS as a TIM on top of the IHS of my 2600k for a comparison with other TIMs. (i.e. the "is that CPU TIM crappy stuff, or good stuff?" test)

As alluded to in the OP, I did my best to attempt to salvage the CPU TIM that Intel uses on their Ivy Bridge cpus so that I could test it as a plain-ole TIM on my 2600K and compare it with other TIMs.

After the decidedly unfavorable results from my re-sandwiching attempt with the original Intel CPU TIM (detailed in this post) I took the lid back off the CPU:

^ lots of CPU TIM to scrape off and recover for this test.

^ all cleaned up, but what does the CPU TIM look like once it has been removed?

Answer: it looks a lot like pile of metal shavings in a spoon (fascinating though that the TIM shavings realize it is they who bend rather than the spoon)

Seriously though the shavings of CPU TIM are really indicating to you just how stiff and non-pliable the TIM itself really is. It is hard, like a thick sheet of plastic that has been cut into shavings with a razor blade. Nothing at all like the malleability of traditional TIMs (including that used on Intel's stock HSF's).

There was no way I could just plop this pile of shavings on top of my 2600K and expect any kind of results that would be representative of the thermal conductivity properties of the bulk TIM as applied to the 3770k cpu.

So I decided I would make an attempt to rejuvenate the TIM...and since we use IPA to remove TIM because it does such a wonderful job of dissolving the TIM (after all that is exactly how I cleaned up the 3770k silicon die and IHS as shown in the photo above) I figured I'd dissolve the TIM shavings in some IPA and then let the IPA evaporate, leaving behind a puddle of pliable gooey CPU TIM.

Ah, if only. You know what they say about theory versus practice? In theory there is no difference between theory and practice, but in practice there is!

The shavings weren't interested in dissolving in the IPA by any means. I even heated up the IPA with a simple water bath (double-boiler style) and chopped up the shavings with the razor blade to get them as small as possible...but there was not "mushing" them, way too hard and stiff. Here's the best I could do to recover the CPU TIM as re-appliable TIM...this is after soaking in hot 91% IPA for nearly 2 hrs:

I left the spoon sit out long enough for the bulk of the IPA to evaporate, trying to give the shavings as much of a chance as possible of absorbing the IPA in hopes of the CPU TIM becoming malleable from the polymer swelling effect alone.

What I had left over was pretty much what I started with, albeit the bits of CPU TIM were now much smaller because of all the chopping I had done in the effort to get them to dissolve in IPA.

I elected to make 5 piles of the CPU TIM on my 2600k IHS, versus just one big pile in the middle of the IHS, because I wasn't overly optimistic that the TIM was going to spread out to cover the whole IHS.

When I was done with my tests I removed the H100 and checked to confirm how well, or not as it were, the CPU TIM had spread out under the pressure of the H100 and the thermal cycling of the LinX testing:

^ yeah, not good at all. What you can't see in this picture is that the 5 spots of CPU TIM were like pancakes, probably 0.5-1mm thick (not thin by any stretch of the imagination), acting as pedestals keeping the H100 hoisted above the IHS so thermal contact was truly not happening.

And the results? About as disastrous as you would expect them to be given how poorly the mount was in this effort:

Ouch At the stock clock of 3.4GHz for my 2600k, the CPU TIM test was pushing temperatures up by a solid +15-20°C over that observed with standard TIMs.

Now obviously this result is not indicative of the actual performance of the CPU TIM that Intel is putting on their Ivy Bridge CPUs, these results are botched because of the inability of the TIM to be recovered and reapplied to subsequent surfaces like my 2600k's IHS.

All well, I guess for now the mystery of "how good, or crappy, is the stock Intel CPU TIM?" will have to go unanswered.

 

[T_Yamamoto]

As alluded to in the OP, I did my best to attempt to salvage the CPU TIM that Intel uses on their Ivy Bridge cpus so that I could test it as a plain-ole TIM on my 2600K and compare it with other TIMs.

After the decidedly unfavorable results from my re-sandwiching attempt with the original Intel CPU TIM (detailed in this post) I took the lid back off the CPU:

^ lots of CPU TIM to scrape off and recover for this test.

^ all cleaned up, but what does the CPU TIM look like once it has been removed?

Answer: it looks a lot like pile of metal shavings in a spoon (fascinating though that the TIM shavings realize it is they who bend rather than the spoon)

Seriously though the shavings of CPU TIM are really indicating to you just how stiff and non-pliable the TIM itself really is. It is hard, like a thick sheet of plastic that has been cut into shavings with a razor blade. Nothing at all like the malleability of traditional TIMs (including that used on Intel's stock HSF's).

There was no way I could just plop this pile of shavings on top of my 2600K and expect any kind of results that would be representative of the thermal conductivity properties of the bulk TIM as applied to the 3770k cpu.

So I decided I would make an attempt to rejuvenate the TIM...and since we use IPA to remove TIM because it does such a wonderful job of dissolving the TIM (after all that is exactly how I cleaned up the 3770k silicon die and IHS as shown in the photo above) I figured I'd dissolve the TIM shavings in some IPA and then let the IPA evaporate, leaving behind a puddle of pliable gooey CPU TIM.

Ah, if only. You know what they say about theory versus practice? In theory there is no difference between theory and practice, but in practice there is!

The shavings weren't interested in dissolving in the IPA by any means. I even heated up the IPA with a simple water bath (double-boiler style) and chopped up the shavings with the razor blade to get them as small as possible...but there was not "mushing" them, way too hard and stiff. Here's the best I could do to recover the CPU TIM as re-appliable TIM...this is after soaking in hot 91% IPA for nearly 2 hrs:

I left the spoon sit out long enough for the bulk of the IPA to evaporate, trying to give the shavings as much of a chance as possible of absorbing the IPA in hopes of the CPU TIM becoming malleable from the polymer swelling effect alone.

What I had left over was pretty much what I started with, albeit the bits of CPU TIM were now much smaller because of all the chopping I had done in the effort to get them to dissolve in IPA.

I elected to make 5 piles of the CPU TIM on my 2600k IHS, versus just one big pile in the middle of the IHS, because I wasn't overly optimistic that the TIM was going to spread out to cover the whole IHS.

When I was done with my tests I removed the H100 and checked to confirm how well, or not as it were, the CPU TIM had spread out under the pressure of the H100 and the thermal cycling of the LinX testing:

^ yeah, not good at all. What you can't see in this picture is that the 5 spots of CPU TIM were like pancakes, probably 0.5-1mm thick (not thin by any stretch of the imagination), acting as pedestals keeping the H100 hoisted above the IHS so thermal contact was truly not happening.

And the results? About as disastrous as you would expect them to be given how poorly the mount was in this effort:

Ouch At the stock clock of 3.4GHz for my 2600k, the CPU TIM test was pushing temperatures up by a solid +15-20°C over that observed with standard TIMs.

Now obviously this result is not indicative of the actual performance of the CPU TIM that Intel is putting on their Ivy Bridge CPUs, these results are botched because of the inability of the TIM to be recovered and reapplied to subsequent surfaces like my 2600k's IHS.

All well, I guess for now the mystery of "how good, or crappy, is the stock Intel CPU TIM?" will have to go unanswered.

Wow. You put in some hard work!
Your posts are always good read

 

[lehtv]

Nice quote.

 

[moonbogg]

It was abvious from the start that the poor thermal performance was BY DESIGN, period. Imagine how well it would do with solder? It would make buying a sandy-E a ludicrous idea.

 

[Akantus]

The TIM looks just like the stuff that's used on notebook CPU's (also bare die).
I used to scrape that of with plastic pry, and the shavings looked the same.

BTW great thread IDC, I'm fan of scientific approach :thumbsup:

 

[pantsaregood]

It was abvious from the start that the poor thermal performance was BY DESIGN, period. Imagine how well it would do with solder? It would make buying a sandy-E a ludicrous idea.

Not any more than 2500K/2600K made the 980X/990X look ridiculous. Not to mention that the vast majority of people who buy Intel CPUs (likely >99.9%) have no interest in over locking. Intel would lose more money doing R&D finding some TIM that allows Ivy Bridge to run at stock but thermally limits overclocking than they do on SB-E sales lost to Ivy Bridge.

 

[Cookie Monster]

Hey IDC, is the CPU heatspreader glued back onto the CPU's PCB or just held in place by the socket retention brackets?

Your tempting me to buy an ivy bridge AND making me feel DIY..

 

[soccerballtux]

heroine is bad for you ande its bad for cpu!!!1

 

[Denithor]

I do and I will. I have AS5, Ceramique, TX-2, MX-1, MX-4, NT-H1, and Indigo Xtreme.

Looking forward to seeing these.

I assume you will be running each one in both spots (under and over IHS), right?

Ceramique is my paste of choice, interested to see how it stacks up against the rest.

 

[SunnyD]

Other than mounting tolerance issues, what's stopping you from directly mounting the heatsink to the die for even better cooling?

 

[BonzaiDuck]

Now obviously this result is not indicative of the actual performance of the CPU TIM that Intel is putting on their Ivy Bridge CPUs, these results are botched because of the inability of the TIM to be recovered and reapplied to subsequent surfaces like my 2600k's IHS.

All well, I guess for now the mystery of "how good, or crappy, is the stock Intel CPU TIM?" will have to go unanswered.

Too bad it couldn't just be submitted to a chemist for "qualitative analysis" to see what they used. Of course, that doesn't tell you about it's "performance."

I have a lot of "little projects" myself. I grow tomatoes; I make-a da sauce; I usa da sauce on-a my pizza. So when you get a great bargain on Crimini/Porta-bella mushrooms, what do you do to keep them from turning to slimy rot before you get to make-a-da-pizza?

You can blanch and freeze; you can pickle with vinegar; you can put them in the food-dryer. But dried shrooms just don't reconstitute well -- it's an irreversible process.

So much for trying to re-use TIM paste that's been dried out. At least, the diamond-stuff is re-usable. . .

 

[moonbogg]

Other than mounting tolerance issues, what's stopping you from directly mounting the heatsink to the die for even better cooling?

This has been asked a few times here with no response. I think its a great idea and might work really well. I would just be concerned about cracking the cpu die however (minor detail).

 

[Kristijonas]

Someone send all this to Intel please so they may learn how to set up CPUs right.

 

[PlasmaBomb]

The shavings weren't interested in dissolving in the IPA by any means

From cleaning off TIMs I found DMSO much more effective at dissolving TIM than IPA, which is great for cleaning as it doesn't leave a residue, and is much less toxic than DMSO (it being cheaper and easier to get hold of also helps...).

 

[HondaCop]

I would just be concerned about cracking the cpu die however (minor detail).

LOL Duct tape can fix that

- Grammar *censored*, this was sent from my phone.

 

[Avalon]

Nice work IDC. Some day I will scrape off the adhesive, as the results are much better than just replacing the Intel TIM alone.

Amusing to see that you used the hammer and blade method to de-lid, I did it fairly similarly with a blade and screwdriver handle. These really are tough chips.

 

[pantsaregood]

On this topic: would lapping the stock heatsink of a 3570K improve temperatures at all? A relative of mine is having me build a PC for them, and they opted to just use the stock cooler. I have some TIM to use on the heatsink, but don't want to bother lapping it if it won't improve temperatures.

I won't be delidding the CPU or lapping it. It isn't worth losing the warranty over, given that it isn't mine.

 

[graysky]

Don't lap one surface, lap both. In other words, do the HS + IHS or do neither.

 

[Idontcare]

Hey IDC, is the CPU heatspreader glued back onto the CPU's PCB or just held in place by the socket retention brackets?

Held in place by the socket's retention bracket. There is no need to physically attach the IHS, it works as intended when one simply reseats it over the CPU PCB and latches down the retention arm for the socket.

I assume you will be running each one in both spots (under and over IHS), right?

Ceramique is my paste of choice, interested to see how it stacks up against the rest.

Yeah that's how I'll do it. Plus a few mix-and-match tests.

I was a big fan of ceramique when I was into vaporphase cooling as it is designed for the sub-zero temperatures whereas the other TIMs are not.

On this topic: would lapping the stock heatsink of a 3570K improve temperatures at all? A relative of mine is having me build a PC for them, and they opted to just use the stock cooler. I have some TIM to use on the heatsink, but don't want to bother lapping it if it won't improve temperatures.

I won't be delidding the CPU or lapping it. It isn't worth losing the warranty over, given that it isn't mine.

I wouldn't bother lapping a stock heatsink. Your time and money (cost of supplies) would be far better off spent on getting a 3rd party HSF to replace the stock HSF if you are looking to improve temps or noise.

 

[Idontcare]

Other than mounting tolerance issues, what's stopping you from directly mounting the heatsink to the die for even better cooling?

This has been asked a few times here with no response. I think its a great idea and might work really well. I would just be concerned about cracking the cpu die however (minor detail).

I've tried to answer this question every time I see it get asked, but I apologize if I missed some of you who have asked this question, I'm not intentionally ignoring folks but I guess I have missed this question a few times going by moonbogg's statement?

Is there a reason you haven't tried direct die contact? I guess it isn't irrational to worry about cracking the die, but we've all mounted a Coppermine or K7 back in the day. I don't see why it would be any more dangerous now.

No reason, I intend to make an attempt at it, I had a positive experience in doing that with my GTX460 when I delidded it last year. The trick then is the same as it would be here, dealing with adjusting the HSF standoffs from the socket to account for the thickness of the IHS.

Adjusting the standoffs on my video card was pretty straightforward, cutting down the standoffs for my CPU's cooler is going to be much more challenging (and irreversible).

AMD chips were exactly like that for the longest time. Of course my experience is based on years ago, but has it changed so much? -- I haven't had the opportunity to open up a newer AMD part; did either maker recently sell chips which were so very different from this ?

While I realize that this isn't my $300 piece of silicon to risk, I wonder at all this talk of sanding the IHS and spreading on double layers of TIM. Sure I seem to recall reading something about smaller processes having thinner, more fragile substrates, and thus dies... But damnit, is it so far out of the question to just stick a nice, massive copper heatsink (or waterblock) on top of a naked IB die? Didn't we all go years clumsily clamping fat aluminum heatsinks onto naked dies not that long ago?

I did just that about a year ago with my GTX460 for fun. The trick of course is building the right standoffs so the die doesn't take the brunt of the mounting pressure, particularly so the pressure is even across the entire die.

This can be done with IB of course, the retention bracket on the mobo will have to be removed (shouldn't be an issue, it is bolted on so it can be unbolted).

From my experience with delidding my GTX460, I'm definitely going to at least attempt to do the same with my 3770k at some point in the tests.

Once you're going to the trouble of delidding, why not skip the lid entirely and mount the cooler right on the die?

A bit riskier, but eliminates the problem of TIM and balancing the spreader etc. entirely.

I did that with my video card (see thread here) and have every intention of making a decent attempt of it with my 3770k. Just not right away, I want to get some other tests done first because there is always a chance of destroying my CPU with direct-contact testing and if that happens I want to at least have all the other data beforehand.

 

[BonzaiDuck]

Don't lap one surface, lap both. In other words, do the HS + IHS or do neither.

+1 :thumbsup: Abso-tively, Posi-lutely . . . Get the nickel plate ground off . . .

 

[Joseph F]

However the IHS on the 3770k was not going to go so easily. You can read a kinder and gentler version of how to get the lid off in fellow colleague graysky's 3770k delidding thread. My approach, however, was a little more energetic and involved a hammer ...but it was just a small hammer :sneaky:

 

[BonzaiDuck]

"Get me a hammer."

I'd seen another forum "DIY" thread -- another forum -- describing the gentle application of the razor blade, going around the lip of the IHS repeatedly and attempting to leave the bead of adhesive intact and the IHS clean of it.

No matter how the processor is supported to assure attenuated vibration, I'd rather suffer the tedium without the mallet.