Bare-die testing: A delidded 3770k, an H100, and 9 different TIMs

[D2ultima]

They should call themselves TIM refurbers then, delidding is baredie These noobshields should be optional (hint, hint Intel please offer a "mobile" K series).

A bit late, but there is a mobile "K" series; it's just called the "MX" chips. the i7-4930MX and 4940MX are Haswell's variants. I know someone who gets his up to 4.8GHz for benchmarks even.

 

[crashtech]

How do you get that mobile chip onto a desktop board?

 

[D2ultima]

How do you get that mobile chip onto a desktop board?

I don't think you do. There ARE boards that take desktop socket 2011 chips and put them with mobile RAM/GPUs/etc though.

Why did you think you could get one of those mobile chips onto a desktop board?

 

[crashtech]

I didn't think any mobile cooling solutions would be up to the task of cooling a Haswell at 4.8Ghz.

 

[D2ultima]

I didn't think any mobile cooling solutions would be up to the task of cooling a Haswell at 4.8Ghz.

With some liquid ultra and maybe some light cooling mods, it's possible. Some people just put it near an A/C vent for a while during tests or something even. But it can be done.

Here's an ivy bridge mobile quad in 3DMark 11. I can't find his haswell screenies right now, but he always says he doesn't let stuff go above 85 degrees and he uses 4.5-4.6GHz for his day-to-day usage. Not sure what TIM he was using here though.

 

[chuck1723]

I really need to delid my 3770k.

 

[DonWuLFF]

Any news for this Coollaboratory Liquid Ultra? Is it safe for die and IHS? Any user still using till today?

 

[Idontcare]

Any news for this Coollaboratory Liquid Ultra? Is it safe for die and IHS? Any user still using till today?

Works fine, no degradation in performance. Next best thing to being soldered on like the EP line.

 

[DonWuLFF]

Works fine, no degradation in performance. Next best thing to being soldered on like the EP line.

Ok. Planning to get one and delid my 4790K for the first time trying.
Read some review,some said its not good for the die,some not good for the IHS. Now I'm confused. Thank you IDC. :thumbsup:

 

[Idontcare]

Ok. Planning to get one and delid my 4790K for the first time trying.
Read some review,some said its not good for the die,some not good for the IHS. Now I'm confused. Thank you IDC. :thumbsup:

Good luck! Do be careful in your delidding, that is some cherry CPU you got there, would be a shame if it gets buggered up. I killed my first IB 3770K by delidding, dang that stung the wallet to replace, but my second and third deliddings were not a problem.

 

[soccerballtux]

what should we get if we want to avoid pumpout?

also, bump for all the black friday buyers upgrading.

 

[SlowSpyder]

what should we get if we want to avoid pumpout?

also, bump for all the black friday buyers upgrading.

Because of this thread, I went with NT-H1 for my FX. For whatever it's worth, I use it on my overclocked FX and have used the same application of NT-H1 since May without issue. Not sure how long, how many hot / cold cycles it is supposed to take. I've gotten my CPU alone to pull over 400 watts (according to my Kill-o-Watt), so I'm sure it got plenty toasty at times, but no issues yet.

 

[soccerballtux]

Because of this thread, I went with NT-H1 for my FX. For whatever it's worth, I use it on my overclocked FX and have used the same application of NT-H1 since May without issue. Not sure how long, how many hot / cold cycles it is supposed to take. I've gotten my CPU alone to pull over 400 watts (according to my Kill-o-Watt), so I'm sure it got plenty toasty at times, but no issues yet.

heat itself not the problem, it's the cycling. so, if you ran scripted linpack to heat the CPU up to 70C, turned it off for 30 seconds and let it cool back down to 45, and rinse/repeat for a day, you should see it in action. I wonder if pasting tons of the thermal compound would prevent this? IE, pump out, sucked back in? Hm.

 

[Idontcare]

heat itself not the problem, it's the cycling. so, if you ran scripted linpack to heat the CPU up to 70C, turned it off for 30 seconds and let it cool back down to 45, and rinse/repeat for a day, you should see it in action. I wonder if pasting tons of the thermal compound would prevent this? IE, pump out, sucked back in? Hm.

The only TIM that purports to be specifically engineered and tested to prevent pump-out is ICD7 (IC Diamond).

The next best thing appears to be CL Liquid Ultra.

Both of those will stain/mar the surfaces of both your IHS and HSF, so if that is a concern for you then your fall-back option is to go with NT-H1.

Many people have used NT-H1 and not observed any thermal creep issues related to pumping out effects.

But if you are looking for bullet-proof guaranteed, then you have no option other than ICD7.

 

[Compddd]

What is pump out?

 

[BonzaiDuck]

The only TIM that purports to be specifically engineered and tested to prevent pump-out is ICD7 (IC Diamond).

The next best thing appears to be CL Liquid Ultra.

Both of those will stain/mar the surfaces of both your IHS and HSF, so if that is a concern for you then your fall-back option is to go with NT-H1.

Many people have used NT-H1 and not observed any thermal creep issues related to pumping out effects.

But if you are looking for bullet-proof guaranteed, then you have no option other than ICD7.

IDC!! About the ICD!! Did we ever get that resolved to a point of certainty? That the abrasive nature of nano-diamond would surely damage the processor? See -- I finally got my hands on a pristine, year-old, unabused i5-3570K. I'm thinking to de-lid. But after de-lidding -- what? Reinstalled IHS or bare-die? Either way -- which TIM to use between the die and anything else?

 

[Idontcare]

IDC!! About the ICD!! Did we ever get that resolved to a point of certainty? That the abrasive nature of nano-diamond would surely damage the processor? See -- I finally got my hands on a pristine, year-old, unabused i5-3570K. I'm thinking to de-lid. But after de-lidding -- what? Reinstalled IHS or bare-die? Either way -- which TIM to use between the die and anything else?

For bare die application: (either under the IHS or with direct HSF mount)

CLU > ICD > NT-H1

 

[soccerballtux]

IDC!! About the ICD!! Did we ever get that resolved to a point of certainty? That the abrasive nature of nano-diamond would surely damage the processor? See -- I finally got my hands on a pristine, year-old, unabused i5-3570K. I'm thinking to de-lid. But after de-lidding -- what? Reinstalled IHS or bare-die? Either way -- which TIM to use between the die and anything else?

painting in frame protected by glass, placed on top of sandpaper pedestal. the sandpaper etches the glass but the painting is safe. You'd have to sand for quite some time to get through the glass....

 

[Idontcare]

painting in frame protected by glass, placed on top of sandpaper pedestal. the sandpaper etches the glass but the painting is safe. You'd have to sand for quite some time to get through the glass....

In this case the grit of the sand paper is very small (the diamond bits in the TIM are very large) compared the thickness of the glass they are sanding. The thickness of the bits of diamond are going to be roughly 3-5x the thickness of the hermetic sealing layer (SiN) that covers the backside of the die.

Think about rubbing your iphone face down across a sidewalk or pavement. Now clamp it down with the force of your HSF, and then cycle it through a few hundred expansion/contraction therma cycles. Won't be pretty. (but I still wouldn't worry about doing it to my CPU)

 

[DrMrLordX]

Thought I'd throw in some minor data from an AMD delid of an A10-7700k. Lots of people are looking at delidding Kaveri chips since they have a notorious clockspeed wall which may or may not be related to thermals.

So I did mine with an x-acto knife, which worked out fairly well. I nicked it a few times on the outer edge of the pcb which has caused no known CPU malfunction. Also bent a few pins with my hands (oops) which I bent back. I lapped the ihs to 400 grit only which got all the nickel off and put a blurry shine on most of it. Then I relidded the CPU using CLU all around.

Temps got much worse. Thermal throttling @ 4.5 ghz, etc.

I double-checked my work muchly much later (was very busy) and discovered that the TIM contact pattern was obviously quite poor between the underside of the IHS and the die. Only the dead center of the die seemed to have any contact. This was painting an incredibly thin layer on the die and another on the underside of the IHS, which is my usual application technique with CLU. I added a wee bit more to the die + underside of the IHS (I ran out of CLU in the process, boo), and temps are better but still slightly worse than they were before the relid.

So, I am thinking that Idontcare's observations about CLU usage may also apply to relid operations.

I have more CLU coming and it should get here in a week or two. Will post further results as I get them.

 

[DrMrLordX]

update: I added a much, much larger amount of CLU to the die and now my temps are much better than before the delid. At stock (3.4 ghz) the hottest I can get the socket temperature is 30C and the hottest I can get the the thermal margin is 6C.

The chip now goes to 4.7 ghz easily, and might go higher. Big thanks to Idontcare for doing the extra work on CLU application. Adapting that work to my delid/relid operation saved my butt from bad temps.

 

[Idontcare]

update: I added a much, much larger amount of CLU to the die and now my temps are much better than before the delid. At stock (3.4 ghz) the hottest I can get the socket temperature is 30C and the hottest I can get the the thermal margin is 6C.

The chip now goes to 4.7 ghz easily, and might go higher. Big thanks to Idontcare for doing the extra work on CLU application. Adapting that work to my delid/relid operation saved my butt from bad temps.

:thumbsup: very happy to hear your endeavors led you to such great results!

The viscosity of CLU is so low that you really have nothing to fear from adding too much, outside of the obvious "don't add so much that it drips outside the CPU and causes an electrical short somewhere" common sense concern 😉 But it is easy to add too little, found that out the hard way.

 

[TurtleCrusher]

Solid write ups. Pretty much as far as anyone could want.

Was using Arctic Silver Ceramique on my P8700 and Radeon 6870 a bad idea? I cant recall the 6870 having an exposed die or not (think so with the springs) but the P8700 is. Should I swap it out?

Ceramique on my FX-8350 is a solid call though, right? It's got an Arctic Freezer Pro 7 V2 and temps aren't an issue. I used the stock TIM but may go to ceramique to get an extra 200Mhz if temps don't go past 55.

 

[Idontcare]

Solid write ups. Pretty much as far as anyone could want.

Was using Arctic Silver Ceramique on my P8700 and Radeon 6870 a bad idea? I cant recall the 6870 having an exposed die or not (think so with the springs) but the P8700 is. Should I swap it out?

Ceramique on my FX-8350 is a solid call though, right? It's got an Arctic Freezer Pro 7 V2 and temps aren't an issue. I used the stock TIM but may go to ceramique to get an extra 200Mhz if temps don't go past 55.

Ceramique is every bit as good as NT-H1 as far as I can tell.

Personally I prefer NT-H1 though just because it seems to clean up easier with less effort and in my experience it has a broader sweet-spot when it comes to achieving an optimal mount. (the too-little/too-much window for Ceramique is narrower than that of NT-H1, much bigger Goldilocks zone for NT-H1)

I think both points (ease of cleanup and ease of achieving an optimal mount) are due to NT-H1's lower viscosity relative to that of Ceramique, but that is just an opinion and not a data-driven fact.

 

[flexy]

One of the most interesting things re: delidding is what I read not too long ago:

That one reason for high temps might not be so much the TIM or how it is applied to the die, but rather how the IHS is glued (back) on. It's all about the gap between the IHS and the die. The thickness of the glue can prevent that the IHS touches the die evenly as it should.

This explains how some people did otherwise perfect CLU jobs but discovered afterwards it didn't work, exactly how you described where you did a perfect "flat" CLU job but needed to apply LOTS more for reasonable temps. (Because otherwise there was poor contact between die and IHS)

This would also mean that sanding down the IHS itself (at the base) would be interesting to achieve a nice, flat CLU application that actually touches the IHS. (As opposed to having to use TIM as a "filler")