Noctua NH-D14

alienware

Banned
So, at last it is finally here. At last the rather fun squabbles and debates between Tom and I can come to a close and we can finally close the book on whether the NH-D14 actually does on AMD chips what it does on Intel chips.

Now a lot of people (myself included) would think that a £60+ish cooler would be overkill on an AMD cpu. Well, let me rephrase that, on a quad core 45nm AMD chip. This cooler would surely be worth it on a hex cored Thuban but is it worth spending half of what your CPU cost to put one of these on it?

Over the course of my banal bantering you should be able to make your own mind up. And, you'll also be able to get the word from some one who is not under NDA, doesn't have to be polite and therefore can be sickeningly blood thirstily brutal when talking about products.

As you guys and ladies have probably learned by now I am probably one of the most critical people on earth. And not just of products I do not own or I can not afford but with all of them . If I find a hole I will pick it, pull it and heave on it until it opens into a wound and then for good measure throw in some salt and vinegar. Basically what I am trying to say is I am an Ahole.

Now when Tom reviewed the NH-D14 on Intel I was skeptical. So skeptical that he and I engaged in forum fisticuffs numerous times. I had been through a good few coolers and no matter what one or how much they cost nothing really seemed to make much of a difference.

Let me explain that briefly. The first cooler I tried to tame this Phenom 2 with was the 92mm Noctua cooler running twin fans. It was far better than a stock cooler but with some Prime goodness it would always get worryingly close to what is considered the thermal limits of my CPU. And I don't like to push things.

So then I put on a Corsair H50. And the temps dropped to a level I had not seen from my CPU before. However those temps were a bit of a mixed blessing because the wee thing could just not keep up over time. Basically for the first hour or so the temps were great. Give it a few hours to warm up the coolant and it could not cope. Temps would start to increase and would not come back down until I switched off the PC and let the coolant go back to room temps.

This improved with dual fans on, but then it became a white elephant. Gone was the silence, temps were still OK-ish and it cost a fortune and made the same if not more noise (due to the rad being far more resistant with letting air go through) than an air cooler costing half the price.

However, being the sort of person who will never ever truly believe in something until he has actually seen it I was still not convinced by Tom's obvious excitment and idolisation of the Noctua.

So, it was time to get physical. Here is my CPU setting that was used for today's test proceedure.

cpusettings.jpg


So as you can see there is an overclock there and it is very real. It's not enormous however but it's there. It's over 10% and it causes more heat and needs more voltage than a Phenom 2 does out of the box. Now I can't remember what temps this CPU ran at when I got it because I never fitted the stock cooler. However, taking what has been said they usually idle at around 40-42c and max out at temps you don't want to see from one of these chips.

OK. Now, today's cooler comparisson comes by way of the NH-C12P cooler from Noctua. However, this is the one running their 140mm fan. I originally got this cooler because the NH-92 or whatever it is called was simply not good enough and the Corsair was beginning to get on my nerves. Temps were too unpredictable over lengthy runs and I was sick of constantly checking my temps.

Here is/was the NH-C12p fitted onto the CPU.

c12pfitted.jpg


So now it was time for the removal and refit.

Fitting the NH-D14.

Is not as easy as I thought it would be, but only down to my being a stupid idiot and making assumptions. See the NH-92 or whatever it's called used this mounting.

oldmounts.jpg


And because the NH-C12P did also I took for granted that the NH-D14 would use the same fittings. So after taking a look at the cooler (the D14) I saw some short sprung screws that looked like the shorter 775 Intel ones on my other coolers.

newscrews.jpg


So, I figured those will need to come off. What a bloody idiot I can be
rolleyes.gif


They're held on with circlips which I thought was a nice touch. Seeing as how on the other two I have the screws can fall out and annoy greatly. So I pulled the clips off (what an even more stupid idiot) and put the cooler onto the thermal paste. Then I realised the screws did not line up. So off it all came again and this time I read the manual. I then discovered that Noctua had redesigned the mounting system completely. And, for the greater good too.

So, the best had become even better.

However due to me being a complete idiot I now had to get down on my hands and knees and find the clips that I had yanked off in a moment of frenzy and put them back. Cue lots of swearing and belittling myself in ways that only I know how.

The new mounting system looks a little something like this.

newmounts.jpg


And the cooler itself then remains able to use the same locked on screws that can't fall off unless you're a bloody stupid idiot like me and you pull them off yourself. Put back on it was time to fit the thing in. But before that, and without further ado let's rewind shall we? It's time for some temps.

Now consider that the NH-C12P manages to cool pretty much on level terms with a Corsair H50 and then you can see the comparisson. You can also see why I was so happy with it because it is virtually silent.

Idle temp after the PC being on since 10am until I switched it off at 3pm.

idleC12p.jpg


Infact, on a single fan that beats the H50. My H50 idled at around 40 with a single fan on. And that would be the most fair comparisson.

So now let's kick this into gear and have some fun shall we? My weapon of choice was Prime 95. Please bear in mind that I only run this for about 30 minutes. Not because I don't want to test my overclock with all of the different tests it goes through, but because after half an hour my CPU never gets any hotter. I don't need to check for stability because I already know it's stable at these clocks because I've ran it all night before.

Load temps, NH-C12P after thirty minutes of Prime 95.

loadc12p-1.jpg


Which when you consider that AMD gurus reakon 62 is the limit is very respectable out of a low profile cooler. On the Corsair I was getting 52. After an hour on the Corsair (see coolant getting warmer) I was getting 55. Still nothing to worry about then.

So let's step back a moment.

Here is the NH-D14 fitted with the middle fan removed. You do not need to remove the one on the end and I haven't deliberately set it higher to clear my ram, this is out of the box here.

heightruler.jpg


Again I would like to pause for a moment and congratulate Noctua on their new fan holding system. It's bloody brilliant. You pull two levers and off comes the fan. No more yanking with pliers, no more it twanging and flying accross the room. I absolutely love them for sorting that out.

Why have I got a stupid Power Rangers ruler going accross the top? Because it fits. This cooler actually fits in my case and I can put the side on. Which I am indeed very happy about, and you're about to find out why.

nh14fitted.jpg


There it is. Lead time? five minutes. And three of those were removing the old brackets.

So, it's time to round this up.

NH-D14 idle. God I hate you Tom.

idlenh14.jpg


NH-D14 load Did I say that I hate you Tom? Oh yeah, I did. Well now I hate you ever more.

loadnh14.jpg


Forty degrees with a 11% or so overclock, at higher volts than stock, after thirty minutes of Prime.

Now what I could have done here was gone to the review of the NH-D14 by Tom, cut it, pasted it and then gone through it replacing the name Intel with AMD.

Conclusion

This is a sad day my friends. Firstly because it stops Tom and myself's funny antics and squabbles over the H50 vs the NH-D14. But secondly because I am wrong. And I absolutely !*!$*!! hate being wrong. Whether or not this will lead to bigger overclocks? Well, I just don't know. But that is no fault of the cooler and I suspect my chip is just from a crappy batch unless I have missed something very obvious.

Temps are disgusting. It looks disgusting. It feels disgusting and god dammit I hate Tom.

Positives.

It's big.

It' quiet.

It's cooler than a penguins bum.

It's gorgeous.

It fits my case.

It's cheap.

Negatives.

Tom was right.

I was wrong.

Tom sucks.

The Corsair H50 is a complete waste of money.

Global warming.

V.A.T

The final word.

Well, it was obvious that the lack of venom flying from my mouth meant something. It meant that for once in a blue moon I could not tear something limb from limb. And I'm kind of annoyed about that. You may be wondering why I said this cooler is cheap, so as a final word I am going to tell you why.

The Prolimatech Megahalems costs around £40. However, it comes with no fans. By the time you add those fans you are looking at £80. The Noctua costs (give or take a few quid) sixty pounds. If you deducted the price of the Noctuas from the cooler you would end up paying £20 for this cooler. The fans on the Noctua are twenty quid a pop and I know that only too well because I just bought four of them for my case. So, like for like I would put £20 fans on the Megahalems. And after spending £80 you would then realise it's not as good as the NH-D14.

So, to quote a very good chum of mine.

It's epic. It's BRILLIANT. It's gobsmackingly good.
 
Thats an epic read, not very often people have me laughing like this fella.

Final thought - OC3D Education rocks
biggrin.gif
 
LOL great read mate
cool.gif


I need a second one for my i5 750 still got the Coolit Domino on there, same story
biggrin.gif


Negatives.

Tom was right.

I was wrong.

Tom sucks.

The Corsair H50 is a complete waste of money.

Global warming.

V.A.T

laugh.gif
 
Epic read dude, especially love the Power Rangers ruler
biggrin.gif


Makes me sad I can't fit mine for a month as it's in Bristol whilst I'm stuck in Southampton until the holidays.
 
The Domino is basically the H50 which is basically the Asus one, the NorthQ Siberian Tiger - ETC until I turn blue. They all have a small pump and they all use a 120mm rad with a loop that simply isn't big enough to get the temps of the fluid down over the long term.

If I were to cheat and take my Corsair H50 out of the box and test it within 30 minutes then the results would be far better. However, I don't cheat and I don't try to cover up what was nothing but a mistake on my part. It says an awful lot when I have a H50 sat on a Phenom 9950 at idle. That's pretty much what I think about it and where it belongs.

The only excuse I really have is that the D14 did not exist when I got the H50. And at that time good AMD coolers were thin on the ground because pre Phenom 2 they were simply crap since the FX series (the rebranded Opterons where AMD were slaughtering Intel's P4 and Pentium D chips).

Since the Core 2 and since Intel's taking of the lead no one really gave a crap about cooling an AMD. Not until the Phenom 2 really.

The thing is none of the coolers I have used have been able to come close to this. On some things massive great lumps like this dont' make that much of a difference. As temps rise you can only do so much before a cooler craps out. Those ridiculous Delta fans could not cool this block and pipes any better than the stock Noctua fans and if they did? god it would be like living at Heathrow.

Right now I'm just having trouble at containing my glee that it actually fits in there. I have literally proven science wrong. If you take that awesome Power Rangers ruler and you use the measurements Noctua provide it does not fit. By about 2cm. However what's odd is when you get it on there it sits far lower than I thought which makes me really happy because unlike anything else in there it was free. Not only was it free but it's amazing to boot. And that doesn't happen very often.

Rune. I feel for you mate. However, this puppy is worth the wait. My mobo temps have come down too due to it being sideways on and not blasting warm air down onto the copper assembly. Which definitely deserves a mention as you will probably notice another one of my Doms got the snapping treatment today
biggrin.gif


But yeah. Snap your ram, take off the heat spreaders, do whatever you have to do but get this cooler on your CPU.
 
Whether or not this will lead to bigger overclocks? Well, I just don't know. But that is no fault of the cooler and I suspect my chip is just from a crappy batch unless I have missed something very obvious.

And that, my Noctua loving chums, will turn out to be the most prophetic thing I have ever said. Even more prophetic than a Nostradamus quatrain.

I had missed something very obvious. Whenever my system bombed, rebooted or garbled my CPU had hit 56c. Now I found that hard to accept because you can apparently push it to 62 before you should call it off. OK, so now I have let that slip maybe I should confess to what I just did.

oc.jpg


I was shaking as I got into Windows. Literally shaking. Not through excitement but with dread that any minute now I would get a buzz through my speakers and a reboot. So I ran Prime. I didn't want to as I found it rather scary, but I ran it. Max temp 50c after pass five. Now that may not sound too impressive but please consider than on anything over 3.4ghz on any other cooler it would either bomb as it switched to test two, or bomb as it switched to test three. And if it didn't bomb my windows dock bar would turn pink or purple and I would have visual artefacts.

Was I confident? no. Because the last time I thought I had cracked 3.6 and gotten it stable when I ran Mafia 2 it bombed.

mafia2.jpg


Not this time.

You know it's weird. I keep sitting here and forgetting to breathe because this just does not seem real. Those clocks are not possible on this CPU and I must be dreaming. I am idling with a fat surge of voltage into the CPU with an enormous overclock on air at close to 3.7ghz. What's even more remarkable is how stable it is.

I never thought I would live to see the day where Tom used the F word in a video, but now I know why.

This cooler is f*cking brilliant. It's not just brilliant, and it's not just bloody brilliant.

It's f*cking brilliant
 
laugh.gif
excellent read. I almost want to buy this cooler instead of going water cooled, but I have now convinced the wife it's something I need to do so no turning back.

Noctua should really take your post and use it in their sales material. Nice to see that Tom was right too. It's great to be part of an honest and unbiased review site.

Awesome mini review of that cooler. Nice job! I especially like this part:

As you guys and ladies have probably learned by now I am probably one of the most critical people on earth. And not just of products I do not own or I can not afford but with all of them . If I find a hole I will pick it, pull it and heave on it until it opens into a wound and then for good measure throw in some salt and vinegar. Basically what I am trying to say is I am an Ahole.

Had me literally lol.
 
Dude. Buy this cooler. But.this.f*cking.cooler.

It performs miracles. Watch this trick. It turns an AMD into an Intel
laugh.gif


Note the lack of any idle temp rise? Look at the f*cking overclock.
laugh.gif


oc2.jpg


Just for a laugh.

oc2load.jpg
 
And to conclude this experiment.

mafia22.jpg


Sadly it seems my PSU isn't going to cope with it. I got a red alarm warning on the 12v line from Probe. I shall see if it happens again but if it does I am going to have to wind it back. I tried dropping the volts from 1.4250 to 1.4 but it wouldn't boot
sad.gif
 
What an epic Read.

I'm going to recycle my H50 spray it with some air freshener and hang it from my rear view!
 
TBH that's about all it is good for tbh
sad.gif


Or a media PC you don't really care about.

I just can't believe the difference between this and all of the other coolers I have used. It is literally other worldly.

I realised that Asus Probe only allows something tiny on the 12v line as a margin, so I have upped it.
 
Hmmm... Looking at your TIM I think that I do not have enough on my cooler, I'll re seat everything tomorrow and I expect to see my temps drop a bit too.
 
This cooler is just godly.

3.9ghz ran fine with everything but a CPU torture. I've stopped using Prime95 now on account of it being poo. It's too inconsistent. Basically I ran it yesterday at the initial 3780 and it ran fine for 45 minutes. No issues at all.

I then ran Mafia 2, no issues. Ran Prime again later in the evening and temps were still the same, BSOD on test one. Reboot ran it again. BSOD on test 2.

So Stepy tells me about LinX. Ran 5 passes on 3780, CPU hit 45c passed with flying colours. However it wouldn't pass test 2 at 3.9ghz and I sort of respect that. I did push for 4ghz but even at 1.51v it only got as far as the Welcome screen. I have heard the chip can handle up to 1.55 but I'm not prepared to risk it. For the above clock I messed with the FSB (the 3900) and had to offset the ram to make it boot. So what I have done now is revert to 19x 200 (3800) with 1.42v. Idle is 35, max in LinX is 45, max in Mafia 2 bench is 42, max in Metro is 43.

So, like Furmark I won't bother using Prime again.

Any way, yeah, this is definitely the best air cooler money can buy. Thus far it's the best cooler I have used that money can buy.

If ever any one needed any proof of that at all? well, on three coolers (NH-92, NH-C12P, Corsair H50) I have never EVER been able to get past 3.4ghz even with 1.45v.

What's absolute genius about this cooler (and I have never seen any one mention it) is that it actually works. When I run LinX I see my CPU hit 45. Then after three seconds it drops to 43. Then it goes back to 46, then back to 42.

So what's happening is the heatpipes are doing the bizz. The heat is generated, goes up the pipes, gets hit by the huge wave of air and immediately drops the temps.

I can't see how this chip can go to 4ghz or beyond on any cooling system. At 3800mhz I get idle of 35, max (and the highest I saw it go) in LinX of 46. When I booted @ 3900mhz the temps idle were identical and hit 45 before I got a bsod. So it wasn't a temp issue.
 
Back
Top