Just ordered a Noctua NH-D14

[996GT2]

Directron had it on sale for $69.99 so I picked one up yesterday. I'm currently running a Hyper 212+ with two 1200 RPM fans in push-pull, and my 2500K hits about 70C at max load @ 4.5 GHz, 1.34V. Could I realistically expect maybe 10C off of that with the NH-D14?

My case setup is a Lian-Li PC-A05N with Scythe S-Flex G for intake at the rear and Gentle Typhoon AP15 for exhaust at the front. Both are throttled down to about 1100 RPM with a fan controller.

EDIT: I was also considering the Thermalright Silver Arrow since it came with better, quieter fans, but this review from Overcock3D of both coolers in a closed case (instead of open bench) really put me off on the SA. Anyone know why it would suddenly fail at 4.4 GHz whereas the noctua performed fine?

Link: http://www.overclock3d.net/reviews/cases_cooling/thermalright_silver_arrow_vs_noctua_nh-d14/2

 

[deimos3428]

Either they actually hit the thermal limit of the Thermalright cooler at 4.4Ghz...or they did something wrong. I suspect the latter given how similar the two designs are. Also, the delta bar for the Thermaltake is incorrect..it should be about 79C (over ambient) if the chip hit 100C.

 

[stahlhart]

Here's the video for that review:

http://www.youtube.com/watch?v=07XTN0Qll2o

I do think that some of his criticisms of the mechanical aspects of the Silver Arrow are valid, especially for someone like him who is constantly tinkering with it, assembling/disassembling it for tests, etc. Much less so for someone who puts it in place and leaves it.

And note that the temperature problem for the Silver Arrow in the review is for an i7-950, not a 2500K or 2600K -- I've had my 2500K up to 4.7GHz and 1.5V with no thermal issues -- warmest any of the cores got was in the low 60s.

Edit: I lied earlier -- core #2 (the consistently hottest in my 2500K) does reach 70C at 4.6GHz and 1.4V with IBT. The other three cores average about 5-6 degrees cooler. I must not have been stressing it hard enough the first time.

 

[Rvenger]

Directron had it on sale for $69.99 so I picked one up yesterday. I'm currently running a Hyper 212+ with two 1200 RPM fans in push-pull, and my 2500K hits about 70C at max load @ 4.5 GHz, 1.34V. Could I realistically expect maybe 10C off of that with the NH-D14?

My 2500k barely hits 70c at max load at 5.0ghz @ 1.45v on a Hyper 212+. My house stays around 76 degrees. I am only using 1 fan too.

 

[996GT2]

My 2500k barely hits 70c at max load at 5.0ghz @ 1.45v on a Hyper 212+. My house stays around 76 degrees. I am only using 1 fan too.

I think my apartment must be hotter than 76-77 in the summer, since the outside temperatures frequently go into the high 80s and I don't run my AC that often. I've never actually measured the temperatures inside my apartment, so I'm not sure exactly what it is. I've tried remounting the 212+ several times and with different TIMs and the results are about the same.

I have a fan controller in my case so I can dial up both case fans to a high RPM (1850 for the Gentle Typhoon and 1900 for the S-Flex), and that only lowers temps by a few C. So I think interior temperatures must make the difference.

 

[HiTek21]

I bought the noctua NH-D14 a couple of weeks back, its a great heatsink. I was running a Zalman CNPS10X Extreme before. Stock speed for my 2600k, I was getting 70c at full load but the noctua dropped those temps about 14 degrees and its a lot quieter. The fans are running full speed and its no louder than the case fans in my Lian Li PC-P50

 

[Rvenger]

I think my apartment must be hotter than 76-77 in the summer, since the outside temperatures frequently go into the high 80s and I don't run my AC that often. I've never actually measured the temperatures inside my apartment, so I'm not sure exactly what it is. I've tried remounting the 212+ several times and with different TIMs and the results are about the same.

I have a fan controller in my case so I can dial up both case fans to a high RPM (1850 for the Gentle Typhoon and 1900 for the S-Flex), and that only lowers temps by a few C. So I think interior temperatures must make the difference.

I know a lot of people use the 'half of pea size' dab or two on the heatsink. I went ahead and applied ceramique 'old school athlon xp' style and spread it thinly across the entire CPU heat spreader. then I lightly filled in the crevices of the heatpipes on the 212+. Its about 80 degrees in my house right now and Realtemp is fluctuating between 27c -32c idle @ 4.2ghz. If I load the 5.0ghz profile then it idles around 37-41c.

 

[Piano Man]

I've got the NH-D14 coupled with Indigo Xtreme for my i7 920. 4.2GHz on air with no noise from a couple feet away. You'll be happy you made the purchase.

 

[Baasha]

great buy!

I have the Noctua NH-D14 and it's a great CPU cooler. It is probably the best air-cooler out there.

 

[Madcatatlas]

Very happy with mine aswell, BUT, im not sure why i cant controll both fans on it... Its probably gotto do with my mobo.. beats me

 

[996GT2]

Very happy with mine aswell, BUT, im not sure why i cant controll both fans on it... Its probably gotto do with my mobo.. beats me

Do you have both fans connected to one header via Y splitter? Otherwise only the fan connected to the CPU fan header will be speed controlled

 

[BonzaiDuck]

I'm attempting to "give" as well as "take" on this thread. Haven't logged in to Anandtech too much this last year, but I'm now starting to build my i7-2600K system with the Z68 chipset. So you can bet I'm . . . ah . . . coming back to . . . refresh my presence . . .

About three years ago, I felt "on top" of the air-cooling options, but I need to . . . come up to speed again.

First, I see that some company -- probably called the "NoFan" company -- has released a "NoFan" cooler -- a fanless cooler. But off the bat -- I see that the thermal resistance of the product is 0.48 C/W.

Friends! Friends! Let me remind us! Most of the "great" heatsinks -- albeit with fans -- for the last three years -- ThermalRight, Prolimatech, Noctua, etc. -- show thermal resistance specs closer to 0.1. Folks! The lower, the better!

Looks like the Noctua [what-izzit?] NH-D14 is probably a winner. I used an earlier Noctua model -- very similar -- which slightly outperformed a ThermalRight (TRUE -- I think we called them). Also, I'd say the Prolimatech Megahelem Rev. B is a safe bet. A good bet.

Supposedly, if the coolers are 'compatible" with socket-1156, they will fit a socket-1155 -- no problem.

Now . . . . I'm looking for a good cooler, y'all! So . . . . you can help the OP as first objective, but . . . . if you have any other suggestions, I'd like to hear them. Before I put together this monster. First time I put a heatsink on that sucker, it will also be the last time. . . .

EDIT: Oh. Yeah. I don't think this has changed, though I was surprised at the lackluster interest. For a 95W processor like a Sandy Bridge, you can probably expect a 5C drop in temperature under load if you sand off the nickel-plating on the processor (voiding your warranty, of course . . . ) You might be able to squeeze it down another 5C if you sand off any nickel-plating on the heatsink base. There had been plenty of detailed instructions for doing this with 320 and 400-grit wet-or-dry sandpaper -- back in the day. Go find 'em.

That leaves the paste. We still use thermal paste, don't we? Spend the money; get a tube of IC Diamond; go through the tedium of spreading it on. That's worth another few degrees in lowering load temps.

For these new cores, maybe it's not worth the trouble. But cooler . . . . is better. Like your CPU clock-speed, where Higher . . . is better.

 

[dmoney1980]

I'm attempting to "give" as well as "take" on this thread. Haven't logged in to Anandtech too much this last year, but I'm now starting to build my i7-2600K system with the Z68 chipset. So you can bet I'm . . . ah . . . coming back to . . . refresh my presence . . .

About three years ago, I felt "on top" of the air-cooling options, but I need to . . . come up to speed again.

First, I see that some company -- probably called the "NoFan" company -- has released a "NoFan" cooler -- a fanless cooler. But off the bat -- I see that the thermal resistance of the product is 0.48 C/W.

Friends! Friends! Let me remind us! Most of the "great" heatsinks -- albeit with fans -- for the last three years -- ThermalRight, Prolimatech, Noctua, etc. -- show thermal resistance specs closer to 0.1. Folks! The lower, the better!

Looks like the Noctua [what-izzit?] NH-D14 is probably a winner. I used an earlier Noctua model -- very similar -- which slightly outperformed a ThermalRight (TRUE -- I think we called them). Also, I'd say the Prolimatech Megahelem Rev. B is a safe bet. A good bet.

Supposedly, if the coolers are 'compatible" with socket-1156, they will fit a socket-1155 -- no problem.

Now . . . . I'm looking for a good cooler, y'all! So . . . . you can help the OP as first objective, but . . . . if you have any other suggestions, I'd like to hear them. Before I put together this monster. First time I put a heatsink on that sucker, it will also be the last time. . . .

EDIT: Oh. Yeah. I don't think this has changed, though I was surprised at the lackluster interest. For a 95W processor like a Sandy Bridge, you can probably expect a 5C drop in temperature under load if you sand off the nickel-plating on the processor (voiding your warranty, of course . . . ) You might be able to squeeze it down another 5C if you sand off any nickel-plating on the heatsink base. There had been plenty of detailed instructions for doing this with 320 and 400-grit wet-or-dry sandpaper -- back in the day. Go find 'em.

That leaves the paste. We still use thermal paste, don't we? Spend the money; get a tube of IC Diamond; go through the tedium of spreading it on. That's worth another few degrees in lowering load temps.

For these new cores, maybe it's not worth the trouble. But cooler . . . . is better. Like your CPU clock-speed, where Higher . . . is better.

Great info, looks like not much has changed over the years, other than the growing trend of closed water cooled setups.Noctua, Prolimitech, and Thermalright are still top names in cooling.

BTW- I would pay Christopher Walken $10 to read your post!

 

[BonzaiDuck]

Yeah . . . . Walken . . . . Did some good character roles . . . . can also be . . . funny . . .

Well . . . I'm torn at the moment. Got most of the essential parts; need to get a cooler; need to . . . adjust the July budget . . .

 

[Carfax83]

I recently bought this cooler myself, and I'm very impressed with it. With my Prolimatech megahalems with 2 Scythe fans, I could only hit 4.2ghz max stable clock on my 920.

With the NH-D14, I can do 4.3ghz stable, which isn't much more of course, but considering that 4.3ghz was unattainable with the Megahalems says something.

I'm just ticked that I missed the Directron deal. Directron is in Houston, and I live in Missouri City!

 

[BonzaiDuck]

I recently bought this cooler myself, and I'm very impressed with it. With my Prolimatech megahalems with 2 Scythe fans, I could only hit 4.2ghz max stable clock on my 920.

With the NH-D14, I can do 4.3ghz stable, which isn't much more of course, but considering that 4.3ghz was unattainable with the Megahalems says something.

I'm just ticked that I missed the Directron deal. Directron is in Houston, and I live in Missouri City!

Interesting that the OP found the Directron bargain. But I think it's still there. I was looking just a few minutes ago -- froogle.google, and all the listings therein. Maybe it went up from the OP's price, but the Directron deal was the lowest I could find out of 20, 30 or more.

Another point being -- the 0.1 Ghz improvement is worth choosing one cooler over another. It might be the case that -- as I'm currently thinking or planning -- someone would be perfectly happy with a 4.0 or 4.1 OC versus the extreme. And as we know, at the high end, the payoffs to voltage increases and even more extreme cooling have an exponentially diminishing effect, or conversely -- temperatures begin to rise exponentially.

But what's an extra $10 or $20 for a "better" cooler?

[Is Missouri City in Texas or is it in Missouri? I'll leave my thinking about this for you to figure out. See -- I'm a very scrupulous fellow, on the matter of my "public" obligations, and I split hairs over "total expense." But I don't split hairs for "reporting" less than 9% of a chump-change purchase less than $100. . . . . ]

EDIT: That "Spesh-ul Dee-ul" is "still on." Basically, it appears as a $79.99 purchase price, but it comes with an "Instant Rebate." That doesn't mean you have to submit a rebate claim-form. It means that they cut the price instantly by $10, charge you tax on the ~$80 amount, and ship it for less than $6. So -- all in all -- about $76 . . . plus tax . . . of course . . . .

 

[ElFenix]

[Is Missouri City in Texas or is it in Missouri? I'll leave my thinking about this for you to figure out.

MO city is a suburb of houston, tx. directron allows local pick up.

 

[Madcatatlas]

Do you have both fans connected to one header via Y splitter? Otherwise only the fan connected to the CPU fan header will be speed controlled

hmm, cant look as it is operating now. but i figur i have not used a Y splitter...as i dont think that was included...or should it have been..?

 

[BoomerD]

I went with the Noctua NH-U12P SE2 cooler for my i5-760 build this past winter. I like the unit. Dead quiet, does a great job at temperature control, although the tan/brown fans are butt-ugly. However, since I don't spend my day looking inside the case...I don't really care about those...as long as they work.

The kit I got came with a supplied "Y" cable so that I can run both fans on the CPU fan header.
According to Noctua, the NH-D14 also comes with that "Y split cable" as well.
http://noctua.at/main.php?show=productview&products_id=34&lng=en&set=1

 

[Madcatatlas]

Hey and cheers for the confirmation. I went trough the package again and found ..a Y type cable.

Forgive my understanding of electronics, but why is one of the heads only 2 pin, while the other is 3 (yellow in addition to black and red).

The i part of the Y itself has a 3 pin "socket" though, so i think ill try it out.




Edit: nope still not letting me control the cpu fan speed trough Asus Fan Xpert, but i can control chassis fans etc

 

[Dark Shroud]

I'm running the NH-U12P SE2 and it's great. The fans are no louder than my system fans or video card.

Lastly I would advice against lapping the CPU.

 

[996GT2]

Hey and cheers for the confirmation. I went trough the package again and found ..a Y type cable.

Forgive my understanding of electronics, but why is one of the heads only 2 pin, while the other is 3 (yellow in addition to black and red).

The i part of the Y itself has a 3 pin "socket" though, so i think ill try it out.




Edit: nope still not letting me control the cpu fan speed trough Asus Fan Xpert, but i can control chassis fans etc

The Y cable doesn't need two 3 pin headers because it can only read the RPM on one of the fans anyway.

Have you tried connecting the Y cable to one of the fan headers for the chassis fans? That should allow you to control the heatsink fans since you can control chassis fans connected to that header.

 

[996GT2]

I'm running the NH-U12P SE2 and it's great. The fans are no louder than my system fans or video card.

Lastly I would advice against lapping the CPU.

Yeah I'm definitely not planning to lap my 2500K for a few degrees gain.

I lapped an Athlon X2 a while back and the gains were minimal. I eventually removed the IHS (it was only a $60 CPU) but damaged the chip while putting the heatsink back on.

Lesson learned...

 

[BoomerD]

Hey and cheers for the confirmation. I went trough the package again and found ..a Y type cable.

Forgive my understanding of electronics, but why is one of the heads only 2 pin, while the other is 3 (yellow in addition to black and red).

The i part of the Y itself has a 3 pin "socket" though, so i think ill try it out.




Edit: nope still not letting me control the cpu fan speed trough Asus Fan Xpert, but i can control chassis fans etc

Here's what I get:

As you see, there's not much difference in the CPU fan speeds, but quite a bit in most of the others.

Running the Intel Burn Test on maximum, my CPU temp maxes out in the low to mid 60's.
I'm comfortable with that. (O/C'd to 3900)

 

[996GT2]

Just received it and installed it. The installation was a difficult, since I had to reverse the fans to match up with my Lian Li PC-A05's reversed airflow. As a result I had to remove the metal clips from the fans and re-attach them on the other side of the fans, and that definitely took some time to do.

Size comparison with my Hyper 212+:

Load temps are about 8-10C lower compared to Hyper 212+ with 2x1200 RPM fans in push-pull.