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

[tw33k]

So the guy on OCN ruined a chip by delidding and fiddling with it (which voids the warranty) and then sent it in to Intel? Not sure people realize, but this is the type of behavior that drives up costs. Not cool, in my opinion.

Intel wouldn't replace the chip under warranty if it had of been delidded

 

[Sheep221]

If you would not remove the heat shield you could return it and get a new one via RMA. The thing is warranty is voided on delidded CPUs.

On other hand you should wear the discharging wristband when working with electronics.

 

[SonDa5]

No Liquid Pro results? Supposed a be 2-3c lower than CL Liquid ultra.

CL Liquid Pro is so thin that when applied properly that you need to have as perfect as possible contact coverage between DIE and block for it to work right.

I used contact paper pressure test before doing bare die mount to ensure proper contact pressure between die and bare block mount. By doing this I found out the best orientation of block for best contact and was able to see contact pressure for mount.

With CL Liquid Pro TIM and bare die against block not alot mount pressure is needed. Just a few turns on the bolts for the water block mount and it is good to go. Springs on the mount bolts work well.

 

[BonzaiDuck]

No Liquid Pro results? Supposed a be 2-3c lower than CL Liquid ultra.

CL Liquid Pro is so thin that when applied properly that you need to have as perfect as possible contact coverage between DIE and block for it to work right.
. . . . . .

With CL Liquid Pro TIM and bare die against block not alot mount pressure is needed. Just a few turns on the bolts for the water block mount and it is good to go. Springs on the mount bolts work well.

Did IDontcare prove experimentally that there is a pumping-out problem with Liquid Pro and/or Ultra? Or was it only a logical deduction and inference that he made? Going back, I see that he said the pumping-out was "reversible" or that there wouldn't develop permanent voids -- or maybe any voids would occur over a longer period of time.

Earlier, he expressed a preference for IC Diamond or something similar (but not IC's Perehelion or whatever it was called.) He noted that it isn't conductive (which -- it isn't) and the lack of pumping out (probably due to the micronized diamond solids locked together under pressure of the HSF retention mechanism.)

For myself, any Ivy Bridge project has been placed on a back-burner for "budgetary" reasons. I had to replace the entire gas line between the meter and our house after the gas co discovered a leak. It wasn't cheap.

By the time I get around to a de-lidded IB system, Haswell may be shortly forthcoming . . . Yet I assume there is still great uncertainty about how Intel plans to "package" their Haswell die . . . . Or so I guess . . .

 

[Idontcare]

^^Maybe they have the 3570K instock? That was just under $200 for me the other day at the chicago store. No OC/delidding on that one though. Going into someone else's machine which needs to have a valid warranty.

In fact, the combo deal was pretty good. 3570K + p8z77-v lk for ~$280 after rebates.

No 3570k either The store looks like it either got completely cleaned out by a stampede of DIYers or the store intentionally limits their inventory to 2-3 units of any given SKU that is running a special :hmm:

If you would not remove the heat shield you could return it and get a new one via RMA. The thing is warranty is voided on delidded CPUs.

On other hand you should wear the discharging wristband when working with electronics.

Wearing a wristband is not practical in these situations where multiple worktables are involved spanning a working space that is some 15ftx25ft.

Typically I am very good about routinely discharging myself as I work with electronics by way grounding myself by touching metal in the workspace. In all my years of handling electronics, 20+, this would be the first time I've managed to kill something by ESD if in fact it is ESD that killed my 3770k and not something else.

Not to discourage others from using discharge straps, but relying on a discharge strap while working in an environment that is not discharge-strap friendly is not good advice IMO, you want people to develop ESD prevention practices that will actually be used (which means they need to be practical).

No Liquid Pro results? Supposed a be 2-3c lower than CL Liquid ultra.

CL Liquid Pro is so thin that when applied properly that you need to have as perfect as possible contact coverage between DIE and block for it to work right.

I used contact paper pressure test before doing bare die mount to ensure proper contact pressure between die and bare block mount. By doing this I found out the best orientation of block for best contact and was able to see contact pressure for mount.







With CL Liquid Pro TIM and bare die against block not alot mount pressure is needed. Just a few turns on the bolts for the water block mount and it is good to go. Springs on the mount bolts work well.

Very cool results SonDa5 :thumbsup: Thanks for sharing them.

I had no intention of testing with Liquid Pro because my understanding of Liquid Pro was that it was intended to be a one-time application which more or less permanently binds the two surfaces together in similar fashion to that which occurs when using Indigo Xtreme. And since I was ultimately intending on using Indigo Xtreme, I didn't have a need to buy Liquid Pro (it would go unused).

So -- IDontCare -- I've caught up scanning through this latest thread.

And after all it has been through since mid-summer (? or earlier?) -- you finally killed your CPU.

I took notice where you said a page or two earlier in this thread that you'd pick the ICD for your final choice of a 24/7/365 TIM. Is that still your final choice?

The better results with Liquid Ultra versus ICD are about 2+C better than I would've imagined, but the data basis for my best guess were based on tests including the IHS.

Basically, it seems that you've proven a set of guidelines for an Ivy Bridge mod for use specifically with a water-block, or more specifically with the H100 cooler. I think some others here have used their heatpipe coolers in configurations that could vary from yours. If I were to do that, I think I'd avoid mounting fans on the cooler and avoid the extra torque. But the washer-standoff approach would seem to provide a stable approach, anyway.

Having read (or scanned) through all of this and the earlier thread, it has been an incredibly revealing and disciplined project.

Yeah I'd go with IC Diamond if I was going to do a bare-die mount for 24/7 application. My long-term plan was to actually put some Indigo Xtreme between the die and the IHS and resecure the IHS onto the CPU package.

Indigo Xtreme is a reflowable solder and it works great (used it with my 2600k).

But if I were to go with a true lidless operation 24/7 then I'd personally shy away from the liquid metal TIMs that remain a liquid metal. Liquid Pro might be an option if it is more pasty than liquid, but only if I didn't mind my H100 becoming fused to my CPU for the long term.

 

[Lepton87]

Do you still have that 2600K? If I remember correctly it did 5.1GHz without problems, so sandy at 5.1GHz would be just a tad bit slower than IVY at 4.9GHz, certainly not enough for most people to notice. If power consumption was of no importance, and no physical alterations involved, which one would you recommend sandy or ivy? I'm thinking about getting an 8 threaded CPU in my rig, as was the plan from the begging, only I expected IVY to be an improvement on all fronts like Kentsfield to Yorkfield, that's why I cheapened out and bought 2500k instead of 2600k.

 

[SlowSpyder]

Thanks for taking the time to do this, IDC. I have Arctic Silver 5 applied, and have used it for a while. I might pick up some NT-H1, for just a few bucks I may be able to drop 4-5C, if your results carry over to my PhII/Scythe Ninja copper. If nothing else, it'd be interesting to see the difference between well settled AS5 and the NT-H1. Again, thanks for posting all this great info.

 

[USER8000]

So would NT-H1 be fine with a non-modified SB or IB Core i5 for longterm use??

 

[Sunny129]

hey IDC, you pointed out that your AS5 seemed a bit watered down. i'm not bringing this back up to reiterate what others have already mentioned (that AS5 shouldn't be that runny), but rather to get your take on a similar scenario i had w/ Noctua NT-H1. my tube of NT-H1 seemed extremely thick and was a nightmare to spread evenly over my IHS (no de-lidded CPU here). i was wondering if my NT-H1 has possibly "gone bad" like the watered down portion of your AS5 tube, or if that's just the way NT-H1 is. and if so, should i be doing the "rice grain-sized drop" method on the IHS instead of trying to spread it first?

TIA,
Eric

 

[BonzaiDuck]

Yeah I'd go with IC Diamond if I was going to do a bare-die mount for 24/7 application. My long-term plan was to actually put some Indigo Xtreme between the die and the IHS and resecure the IHS onto the CPU package.

Indigo Xtreme is a reflowable solder and it works great (used it with my 2600k).

But if I were to go with a true lidless operation 24/7 then I'd personally shy away from the liquid metal TIMs that remain a liquid metal. Liquid Pro might be an option if it is more pasty than liquid, but only if I didn't mind my H100 becoming fused to my CPU for the long term.

You or someone else had mentioned that one (or the other) CoolLaboratory products remained in a liquid state -- good in one sense because you may be able to remove the processor from the water-block without damaging it, not so good even for a "reversible" pumping out. So -- yeah -- it probably makes more sense to reinstall the IHS with something like the Indigo: You get lower temperatures than with other solutions that include IHS replacement, but you can compromise those gains with the IHS and still be able to remove the water block or HSF.

So all the readers here have several options worth considering.

 

[Gillbot]

No 3570k either The store looks like it either got completely cleaned out by a stampede of DIYers or the store intentionally limits their inventory to 2-3 units of any given SKU that is running a special :hmm:

Intel Core i7 3770K 3.5GHz LGA 1155 Processor
10+ in stock at Northeast Ohio/Mayfield Heights Micro Center
$229

Intel Core i7 2700K 3.5GHz LGA 1155 Boxed Processor
10+ in stock at Northeast Ohio/Mayfield Heights Micro Center
$199

Core i5 3570K 3.4GHz LGA 1155 Processor
10 in stock at Northeast Ohio/Mayfield Heights Micro Center
$169

Core i5 2500K 3.3GHz LGA 1155 Boxed Processor
1 in stock at Northeast Ohio/Mayfield Heights Micro Center
$159

Need me to make a road trip? I kept talking myself out of getting a 3930k when they were at $399 and i'm still telling myself I don't need a 3820 also.

 

[SlowSpyder]

DOH! Sorry IDC, I just read through the thread, sorry to hear about your CPU.

 

[Idontcare]

Do you still have that 2600K? If I remember correctly it did 5.1GHz without problems, so sandy at 5.1GHz would be just a tad bit slower than IVY at 4.9GHz, certainly not enough for most people to notice. If power consumption was of no importance, and no physical alterations involved, which one would you recommend sandy or ivy? I'm thinking about getting an 8 threaded CPU in my rig, as was the plan from the begging, only I expected IVY to be an improvement on all fronts like Kentsfield to Yorkfield, that's why I cheapened out and bought 2500k instead of 2600k.

I do. The 2600k would do 5GHz without missing beat. For what I do, I need 3-4 desktop computers so the 2600k was already rotated out of my "fun hobby computer" and over into a "it works for a living" computer, but I can certainly rotate it back if need be.

As for your question - SB or IB - once power usage is removed from the equation, if the individual is not using the iGPU then I'd recommend a SB no question. It is a no-fuss no-muss situation where hitting good overclocks on economical air cooling is completely doable.

IB was fun for me because of all the fuss and all the muss, being able to delid opened the door to doing fun stuff as a hobbyist but the fact that you basically have no choice but to go through all that just to get the same upper-end OC's that you can get out of SB makes it a difficult choice to recommend under the constraints you gave.

Thanks for taking the time to do this, IDC. I have Arctic Silver 5 applied, and have used it for a while. I might pick up some NT-H1, for just a few bucks I may be able to drop 4-5C, if your results carry over to my PhII/Scythe Ninja copper. If nothing else, it'd be interesting to see the difference between well settled AS5 and the NT-H1. Again, thanks for posting all this great info.

I use to use AS5 religiously on all my rigs, I've have since completely switched over to NT-H1. I saw the same sort of improvement in temperatures on my Q6600, Q9505, 2600k and 3770k. I will test both AS5 and NT-H1 on my incoming FX8350, but it would be a damn odd thing if it suddenly didn't show the same kind of improvement

So would NT-H1 be fine with a non-modified SB or IB Core i5 for longterm use??

Absolutely fine. Noctua knows what they are doing and it shows whether you are looking at their NH-D14 coolers or their NT-H1 TIM. I even used NT-H1 on my GTX460 when I delidded it and remounted an Accelero XTREME Plus.

(I sound like a damned spokesperson for Noctua D: but I swears I'm just calling it like I sees it)

hey IDC, you pointed out that your AS5 seemed a bit watered down. i'm not bringing this back up to reiterate what others have already mentioned (that AS5 shouldn't be that runny), but rather to get your take on a similar scenario i had w/ Noctua NT-H1. my tube of NT-H1 seemed extremely thick and was a nightmare to spread evenly over my IHS (no de-lidded CPU here). i was wondering if my NT-H1 has possibly "gone bad" like the watered down portion of your AS5 tube, or if that's just the way NT-H1 is. and if so, should i be doing the "rice grain-sized drop" method on the IHS instead of trying to spread it first?

TIA,
Eric

Hi Eric,

Yeah the AS5 seemed odd to me but the performance wasn't any different than what I'm use to.

That said, I've been through 4 tubes of NT-H1 now and they were all very smooth and easy to apply. So it does sound like you either got a bad batch of NT-H1 or it had separated somehow inside the tube...either way you would be better off IMO tossing the tube and getting another one. $10 shipped is a pretty good price for any TIM IMO, not worth the hassle of wondering if your results are questionable because you think the tube might be questionable.

(FWIW I pitched my AS5 tube too, got two others I will use for future tests)

All the best,
Phil

You or someone else had mentioned that one (or the other) CoolLaboratory products remained in a liquid state -- good in one sense because you may be able to remove the processor from the water-block without damaging it, not so good even for a "reversible" pumping out. So -- yeah -- it probably makes more sense to reinstall the IHS with something like the Indigo: You get lower temperatures than with other solutions that include IHS replacement, but you can compromise those gains with the IHS and still be able to remove the water block or HSF.

So all the readers here have several options worth considering.

Having seen the way the Liquid Ultra sort of bonds and solidifies to my H100 (removable only by lapping, no different than the Indigo Xtreme in that sense) I actually suspect it will perform just fine against pump-out over time. We'll certainly know if it doesn't because it really is a favorite amongst delidders, so we'll be hearing reports of it in about 2-3 months time if that turns out to be the case.

Need me to make a road trip? I kept talking myself out of getting a 3930k when they were at $399 and i'm still telling myself I don't need a 3820 also.

Right now I've got my fingers crossed on snagging a 3770k for $110, so lets see if that pans out.

DOH! Sorry IDC, I just read through the thread, sorry to hear about your CPU.

Thanks for the condolences, I must say the part that sucks is sitting here thinking of all the additional tests that would have been fun to run but I can't because the chip is no longer available.

Would have been nice to make an attempt at that Goldilocks liquid ultra run, or test out the liquid pro, or re-test AS5 with a different tube, or do more tests with MX-4 to figure out how on earth it crapped out on me so badly

I was even wanting to run some tests with the Intel Stock HSF on the bare die to see if the necessity of the IHS as a heat-spreader for the low-cost stock HSF was fact or fiction. Alas, can't do that now.

 

[Gillbot]

Right now I've got my fingers crossed on snagging a 3770k for $110, so lets see if that pans out.

:thumbsup:

 

[Idontcare]

Remaining tests to perform:

1) Put IHS back on 3770k, attempt to boot to see if socket connections are the problem (the IHS pulls double duty as more uniformly apply the down-pressure on the PCB to properly get the LGA matched up to the pins in the socket)

2) Put 2600k into the MIVE-Z to confirm the MIVE-Z boots just fine.

3) Put 2600k into the Asrock Xtreme <insert random numer>+ to confirm the Asrock boots just fine

4) Put 3770k into the Asrock and see if it boots.

OK, ran the above tests and can confirm the 3770k has dieded.

The Mive-z boots fine with the 2600k, as does the asrock extreme6. Neither the mive-z nor the asrock will boot with the 3770k.

Just for extra measure I put the IHS back on the 3770k, and tried a different HSF. (yeah completely unlikely to matter but I had to check just to confirm)

I should have realized something was up with my 3770k when I noticed that it came wrapped in a red shirt

 

[SlowSpyder]

Too bad it is confirmed. RIP little guy, off to that silicon valley in the sky...

If testing would have continued, I would have had to of submitted a request for toothpaste, if you would have been willing to put that on your bare CPU. I have read tests in the past where it seems to hang with some of the lower end thermal compounds. Not sure what brand, Crest, Colegate, etc? But, it would have been interesting!

Incoming 8350, eh? Sounds like nerd paradise in your home.

 

[supremor]

I remember IDC tested tooth paste in his previous delidding thread, it did not fare well.

Tough luck killing your chip man, you think ESD was likely cause of death then?

 

[Avalon]

Thanks for the results IDC, fantastic as always!

You've got me damn near inspired enough to start tinkering with my 3570K again. I de-lidded it a while back and just replaced the stock TIM1 with MX2, then put it back on and applied MX2 for TIM2.

Looks like if I both get a tube of NT-H1 and decrease the gap between IHS and die by shaving off the rest of the glue that previously held down the IHS to the chip, I'd see some nice gains.

 

[sm625]

Wow, this can be an expensive hobby. At least we have solid data that tells us the old and venerable AS-5 is "good enough".

 

[Revolution 11]

Wow, this can be an expensive hobby. At least we have solid data that tells us the old and venerable AS-5 is "good enough".

It's good enough in the sense that if you still have some AS-5 lying around, go ahead and use it. But don't buy any new AS-5 because other TIM evaluations (skinneelabs) don't rate AS-5 that well. Even in IDC's review, Ceramique did better than AS-5 here and its cheaper and non-capacitive.

 

[Idontcare]

It's good enough in the sense that if you still have some AS-5 lying around, go ahead and use it. But don't buy any new AS-5 because other TIM evaluations (skinneelabs) don't rate AS-5 that well. Even in IDC's review, Ceramique did better than AS-5 here and its cheaper and non-capacitive.

^ I agree, that is the take home message.

If you already have AS5 installed, don't remove it just to remount with NT-H1 for a few degrees improvement. A few degrees isn't going to change your max OC or lower your power usage by enough to matter.

If you happen to have an existing tube of AS5 lying around, don't throw it away, go ahead and use it.

If you happen to be in the market to buy a new TIM, don't short-change yourself when a tube of the good stuff only costs you $10 shipped, just buy some NT-H1 and use it going forward.

 

[Sunny129]

Hi Eric,

Yeah the AS5 seemed odd to me but the performance wasn't any different than what I'm use to.

That said, I've been through 4 tubes of NT-H1 now and they were all very smooth and easy to apply. So it does sound like you either got a bad batch of NT-H1 or it had separated somehow inside the tube...either way you would be better off IMO tossing the tube and getting another one. $10 shipped is a pretty good price for any TIM IMO, not worth the hassle of wondering if your results are questionable because you think the tube might be questionable.

(FWIW I pitched my AS5 tube too, got two others I will use for future tests)

All the best,
Phil

thanks for the advice...i had a sneaking suspicion that something wasn't right w/ my NT-H1...i'll grab another tube then.

If you already have AS5 installed, don't remove it just to remount with NT-H1 for a few degrees improvement. A few degrees isn't going to change your max OC or lower your power usage by enough to matter.

i'm not sure i follow...how would switching to a better TIM lower power consumption? just b/c a CPU runs cooler w/ NT-H1 than it does w/ AS5 doesn't mean its consuming any less power. a CPU is going to generate the same amount of heat at a specific power draw regardless of what TIM and/or fan/heatsink assembly is being used. its just that some TIMs, fans/heatsinks, or combinations thereof, will do a better job of dissipating that heat than others...at least that was the impression i was under. perhaps i'm overlooking something?

TIA,
Eric

 

[Idontcare]

i'm not sure i follow...how would switching to a better TIM lower power consumption? just b/c a CPU runs cooler w/ NT-H1 than it does w/ AS5 doesn't mean its consuming any less power. a CPU is going to generate the same amount of heat at a specific power draw regardless of what TIM and/or fan/heatsink assembly is being used. its just that some TIMs, fans/heatsinks, or combinations thereof, will do a better job of dissipating that heat than others...at least that was the impression i was under. perhaps i'm overlooking something?

TIA,
Eric

It turns out that power consumption is also temperature dependent, even if you have the same clockspeed and same voltage.

See Effect of Temperature on Power-Consumption with the i7-2600K

and i7-3770K vs. i7-2600K: Temperature, Voltage, GHz and Power-Consumption Analysis

 

[Sunny129]

^ i see i have some reading to do...thanks for the additional information - it seems like i learn something new everyday on the AT boards .

 

[Gillbot]

^ I agree, that is the take home message.

If you already have AS5 installed, don't remove it just to remount with NT-H1 for a few degrees improvement. A few degrees isn't going to change your max OC or lower your power usage by enough to matter.

If you happen to have an existing tube of AS5 lying around, don't throw it away, go ahead and use it.

If you happen to be in the market to buy a new TIM, don't short-change yourself when a tube of the good stuff only costs you $10 shipped, just buy some NT-H1 and use it going forward.

I'll stick to Ceramique. :thumbsup: