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

[dqniel]

Well, that may speak to the low quality of OEM TIM as much as diamond thermal properties.

Oh definitely. But, I also noticed a 3-5c difference from AS Ceramique 2 when used on my 6950's GPU. But yes, I think you could replace the stock TIM in Dell laptops with chocolate sauce and it would work better...

 

[AsusGuy]

I de-lidded my CPU a few months ago and was getting temps in the low 80's MAX (Intel Burn Test high stress) with my 3570K running at 4.6 with a fixed voltage @ 1.25 and cooler master hyper 212 using Liquid Metal Ultra under the IHS and on the HSF and IHS. I just switched to the zerotherm zt-10d heatsink with the stock CM 212 fan and temps are now at around 72-75 in IBT. I am thinking about doing push/pull with this HSF and see what my temps get down to. I will post results.

 

[BonzaiDuck]

I de-lidded my CPU a few months ago and was getting temps in the low 80's MAX (Intel Burn Test high stress) with my 3570K running at 4.6 with a fixed voltage @ 1.25 and cooler master hyper 212 using Liquid Metal Ultra under the IHS and on the HSF and IHS. I just switched to the zerotherm zt-10d heatsink with the stock CM 212 fan and temps are now at around 72-75 in IBT. I am thinking about doing push/pull with this HSF and see what my temps get down to. I will post results.

So . . . you haven't noticed any "adverse effects" for using the Liquid Ultra? That's good . .

There are a lot of things that can be done to add to cumulative temperature improvements, including a fuss over your fans. If you plan to stick with your cooler choice (the hyper-212?) -- and if those are the temps from the core sensors, you're already doing better @ 4.6 than I am with my Sandy Bridge and an NH-D14 . . .

 

[dqniel]

I de-lidded my CPU a few months ago and was getting temps in the low 80's MAX (Intel Burn Test high stress) with my 3570K running at 4.6 with a fixed voltage @ 1.25 and cooler master hyper 212 using Liquid Metal Ultra under the IHS and on the HSF and IHS. I just switched to the zerotherm zt-10d heatsink with the stock CM 212 fan and temps are now at around 72-75 in IBT. I am thinking about doing push/pull with this HSF and see what my temps get down to. I will post results.

I just bought a ZT-10D yesterday for $17.02

Since I already have a plethora of $3 YL 120mm fans sitting around here, I can't wait to test it out with a push/pull configuration when I get home. I'll be upgrading from a CM 212+ with YL 120mm push/pull.

If the ZT-10D performs well, and I think it will after reading your post, then I might buy a few more at that price...

p.s. - I'm jealous of your 3570K. Mine absolutely will not get prime stable at 4.6Ghz, and I've tried up through 1.3v. I can do 4.5Ghz at 1.216v, though. It's one of the more frustrating chips I've ever overclocked in terms of diminishing returns.

 

[bl00tdi]

p.s. - I'm jealous of your 3570K. Mine absolutely will not get prime stable at 4.6Ghz, and I've tried up through 1.3v. I can do 4.5Ghz at 1.216v, though. It's one of the more frustrating chips I've ever overclocked in terms of diminishing returns.

Tell me about it. My 3570k requires 1.36 to be absolutely stable @ 4.5ghz and refuses to be stable @ 4.6ghz. I might be one of the few Ivy Bridge owners actually running into a voltage wall (as opposed to a temperature wall) without delidding. It's a shame too; I have a fresh tube of CL liquid ultra sitting on my desk purchased in anticipation of a delid but alas, I received a dud processor.

My next plan is to buy the Intel PTPP and kill.....ahem...take advantage of the it.

 

[Idontcare]

p.s. - I'm jealous of your 3570K. Mine absolutely will not get prime stable at 4.6Ghz, and I've tried up through 1.3v. I can do 4.5Ghz at 1.216v, though. It's one of the more frustrating chips I've ever overclocked in terms of diminishing returns.

Tell me about it. My 3570k requires 1.36 to be absolutely stable @ 4.5ghz and refuses to be stable @ 4.6ghz. I might be one of the few Ivy Bridge owners actually running into a voltage wall (as opposed to a temperature wall) without delidding. It's a shame too; I have a fresh tube of CL liquid ultra sitting on my desk purchased in anticipation of a delid but alas, I received a dud processor.

My next plan is to buy the Intel PTPP and kill.....ahem...take advantage of the it.

You both have enabled the "PLL overvoltage" in your BIOS, yes? (not to be confused with increasing the PLL voltage)

 

[AsusGuy]

I just bought a ZT-10D yesterday for $17.02

Since I already have a plethora of $3 YL 120mm fans sitting around here, I can't wait to test it out with a push/pull configuration when I get home. I'll be upgrading from a CM 212+ with YL 120mm push/pull.

If the ZT-10D performs well, and I think it will after reading your post, then I might buy a few more at that price...

p.s. - I'm jealous of your 3570K. Mine absolutely will not get prime stable at 4.6Ghz, and I've tried up through 1.3v. I can do 4.5Ghz at 1.216v, though. It's one of the more frustrating chips I've ever overclocked in terms of diminishing returns.

I am very happy with the ZT-10D so far. Got it at MC for $18.00 which I consider a great investment considering the quality of the heatsink. I will say that de-lidding and using Liquid Metal Ultra / Pro has had the most significant impact on temps. I guess I should consider myself fortunate with my CPU. I am going to try pushing it a bit further now that I seem to have temps under control and see if I can hit 4.7-4.8 without runaway temps and voltage. Running the CPU at fixed voltage does seem to help hit a lower stable voltage for the CPU. Evan though 1.55 is the "safe" limit. I can't see running it at that high a voltage in 24/7 setup without serious heat / throttling issues.

 

[WhoBeDaPlaya]

You both have enabled the "PLL overvoltage" in your BIOS, yes? (not to be confused with increasing the PLL voltage)

Should that be disabled? I'm running into Vcore walls with my 3570K as well (IBT temps ~89C). Managed to get it running at 4.5GHz w/1.24Vcore (this is the severely drooped value under IBT load, idle is ~1.33V with LLC @ 100% on an Asrock Pro4).

 

[BonzaiDuck]

You both have enabled the "PLL overvoltage" in your BIOS, yes? (not to be confused with increasing the PLL voltage)

Did ASUS thoroughly resolve that problem in BIOS with the wake/sleep-state malfunction per these settings and speeds over 4.5Ghz?

 

[bl00tdi]

You both have enabled the "PLL overvoltage" in your BIOS, yes? (not to be confused with increasing the PLL voltage)

Affirmative. Also, tried PLL voltages anywhere from 1.5 to 1.85 to no avail. VRM tweaking was similarly futile. Pretty upset about it since my temps @ 4.5ghz/1.36v is ~85c max during IBT. Aida64 stress test temps are 80c max with both @ 27c ambient. With even an AVERAGE 3570k there's a chance that delidding would not be necessary until 4.8-4.9ghz based on the temps I am seeing.

Z77 Sabertooth
Xigmatek Aegir push/pull rotated 90*
prolimatech PK-1

 

[Idontcare]

Should that be disabled? I'm running into Vcore walls with my 3570K as well (IBT temps ~89C). Managed to get it running at 4.5GHz w/1.24Vcore (this is the severely drooped value under IBT load, idle is ~1.33V with LLC @ 100% on an Asrock Pro4).

For whatever technical reasons, it must be "enabled" on SB and IB CPU's in order to get past 4.5GHz or so. It is not a specific threshold, it varies by chip. Some you need it to get above 4.3GHz, others you don't need it until you try to go above 4.8GHz.

Your Vdroop is crazy, is that normal for that mobo? Vdroop on mine is basically zero but I had to pay a pretty penny for that priviledge (MIVE-Z, cost more than my 2600k at the time).

Did ASUS thoroughly resolve that problem in BIOS with the wake/sleep-state malfunction per these settings and speeds over 4.5Ghz?

I don't know if they did, but knowing ASUS I would say they haven't.

 

[WhoBeDaPlaya]

Your Vdroop is crazy, is that normal for that mobo? Vdroop on mine is basically zero but I had to pay a pretty penny for that priviledge (MIVE-Z, cost more than my 2600k at the time).

No idea, it's an Asrock Z77 Pro4. Can't really complain for the price ($70 on clearance - $50 combo discount )
This is with 100% LLC enabled and the large range jump is a side-effect. Haven't tried it without LLC.

I just bought a ZT-10D yesterday for $17.02

Thanks for posting this! I just picked up the same thing at MC this evening

 

[BonzaiDuck]

Your Vdroop is crazy, is that normal for that mobo? Vdroop on mine is basically zero but I had to pay a pretty penny for that priviledge (MIVE-Z, cost more than my 2600k at the time).

Did the MIVE-Z board zero out at the level above "High"[50%]? P8Z68 . . Pro has "Regular," "Medium," "High," etc. If I go up one LLC-notch from "High," the lab-tests say I should precisely eliminate the droop.

It looks that way . . .

 

[Anth Seebel]

For whatever technical reasons, it must be "enabled" on SB and IB CPU's in order to get past 4.5GHz or so. It is not a specific threshold, it varies by chip. Some you need it to get above 4.3GHz, others you don't need it until you try to go above 4.8GHz.

Nice tip IDC , Im coming from X58 so didnt know about 'PLL overvoltage'. I incorrectly assumed it was only needed if you set PLL over 1.8v, but after doing some reading it is actually a setting that ramps up with clocks if enabled.

I managed to drop vcore because of it from 1.325v to 1.305 in bios.

 

[dqniel]

Soooo... an update for you all. Tonight I took off my 212+ push/pull and installed the ZT-10D. My biggest issue is that I have the "closed" Yate Loons and they won't work with the rubber mounting pins provided with the HSF. So alas, no push/pull configuration for now... until I get some "open" corner fans installed.

However, even with that disadvantage- HOLY CRAP! Big temperature drop at load.

Ambient temps about 24-25c in both tests.

CM 212+ with push/pull Yate Loons @ 2100RPM:
4.6Ghz @ ~1.27v Prime Blend ~12min 83c-91c-88c-86c max temp in RealTemp
then FREEZE

NT-10D with one CM fan (pos IMO) @ 1900RPM:
4.6Ghz @ ~1.28v Prime Blend ~22min 69c-80c-79c-78c max temp in RealTemp
and still going strong

So, using a worse fan configuration, the exact same TIM between the die and IHS (IC Diamond) and the IHS and HSF (ASC 2), and slightly increased voltage I see a ~10C drop at load?

Best $17.02 HSF ever.


I hope the temps drop even more once I get some CL Liquid Pro (or Ultra) and get some open corner fans to do push/pull action again.

Did the MIVE-Z board zero out at the level above "High"[50%]? P8Z68 . . Pro has "Regular," "Medium," "High," etc. If I go up one LLC-notch from "High," the lab-tests say I should precisely eliminate the droop.

It looks that way . . .

My vDroop with my (nearly free due to bundle savings) ASRock Pro4-M makes it so I go from 1.30v down to 1.28v. I'm OK with that, considering that I expected this cheapo board to vDroop badly.

 

[Idontcare]

Did the MIVE-Z board zero out at the level above "High"[50%]? P8Z68 . . Pro has "Regular," "Medium," "High," etc. If I go up one LLC-notch from "High," the lab-tests say I should precisely eliminate the droop.

It looks that way . . .

I don't even have to mess with that, but I could if I wanted too (the options are there).

For me, I just leave LLC set to "Auto" in the BIOS. It is magically delicious.

I'm measuring Vcc with a voltmeter too, and it is rock-solid at whatever I set it at in AI Suite.

 

[BonzaiDuck]

I don't even have to mess with that, but I could if I wanted too (the options are there).

For me, I just leave LLC set to "Auto" in the BIOS. It is magically delicious.

I'm measuring Vcc with a voltmeter too, and it is rock-solid at whatever I set it at in AI Suite.

I can't remember if my board even has an "Auto" setting, but I didn't think that it did . . . ASUS once had some flagships that weren't all that stellar, but these Maximus boards appear to be knock-down, stand-up winners . . .

 

[AsusGuy]

Soooo... an update for you all. Tonight I took off my 212+ push/pull and installed the ZT-10D. My biggest issue is that I have the "closed" Yate Loons and they won't work with the rubber mounting pins provided with the HSF. So alas, no push/pull configuration for now... until I get some "open" corner fans installed.

However, even with that disadvantage- HOLY CRAP! Big temperature drop at load.

Ambient temps about 24-25c in both tests.

CM 212+ with push/pull Yate Loons @ 2100RPM:
4.6Ghz @ ~1.27v Prime Blend ~12min 83c-91c-88c-86c max temp in RealTemp
then FREEZE

NT-10D with one CM fan (pos IMO) @ 1900RPM:
4.6Ghz @ ~1.28v Prime Blend ~22min 69c-80c-79c-78c max temp in RealTemp
and still going strong

So, using a worse fan configuration, the exact same TIM between the die and IHS (IC Diamond) and the IHS and HSF (ASC 2), and slightly increased voltage I see a ~10C drop at load?

Best $17.02 HSF ever.


I hope the temps drop even more once I get some CL Liquid Pro (or Ultra) and get some open corner fans to do push/pull action again.



My vDroop with my (nearly free due to bundle savings) ASRock Pro4-M makes it so I go from 1.30v down to 1.28v. I'm OK with that, considering that I expected this cheapo board to vDroop badly.

I had this same problem. I got an NZXT 120mm fan at MC and was unable to mount it. I was a bit disappointed. So for now I am using the CM 212+ fan and like you noted my min temps seem to stay the same but under load temps have dropped a lot.

 

[WhoBeDaPlaya]

I had this same problem. I got an NZXT 120mm fan at MC and was unable to mount it. I was a bit disappointed. So for now I am using the CM 212+ fan and like you noted my min temps seem to stay the same but under load temps have dropped a lot.

I actually got worse load temps than a Hyper 212+. Most likely a case of PEBCAK and will take apart the system again after work (installing the ZT-10D is a giant PITA).

 

[dqniel]

I actually got worse load temps than a Hyper 212+. Most likely a case of PEBCAK and will take apart the system again after work (installing the ZT-10D is a giant PITA).

Yeah, the mounting instructions were a bit... odd to me? It took me a while to get it mounted. Hopefully you get a better mount the second time around and reach the full potential, because I'm certain after switching from the 212+ to this that the ZT-10D is much better under load.

 

[AsusGuy]

You both have enabled the "PLL overvoltage" in your BIOS, yes? (not to be confused with increasing the PLL voltage)

I have not checked this setting yet. I will take a look when I get home from work.

 

[Ferzerp]

Just so you know, I hate all you guys whose 3770k will go over 4.4Ghz with reasonable voltages.

 

[WhoBeDaPlaya]

Just so you know, I hate all you guys whose 3770k will go over 4.4Ghz with reasonable voltages.

Define reasonable? 1.25V under IBT load?

 

[Ferzerp]

Mine does 4.4 with 1.224V load, so sure.

 

[BonzaiDuck]

Mine does 4.4 with 1.224V load, so sure.

Do you think that voltages in the 1.23 to 1.30V range are unreasonable? IDC posted some tables in this thread that pairs over-clock speeds and voltages. I think 4.7 Ghz required something around 1.29+V.