Noctua NH-U12F: Quiet Air Cooling from Austria
by Wesley Fink on April 25, 2007 4:00 AM EST- Posted in
- Cases/Cooling/PSUs
Scaling of Cooling Performance
Scaling charts plot the performance of a cooler as the overclock is increased. This begins to provide a picture of the areas of most efficient performance for each cooler. It also provides a graphic way to compare performance curves for the various tested coolers. The lower the line for a cooler (representing lower temperatures), the more efficient that cooler is at cooling. |
| Click to enlarge |
At 2.93GHz the retail HSF is running at 41C, compared to 30C with the stock Noctua and 29C with the push-pull fans. This is not the best performance seen in our tests at idle, but it is competitive with the top coolers tested so far. This is a delta of 11C. The delta becomes greater as the overclock increases. At 3.73GHz the idle with the retail fan is 56C compared to the Noctua at 42C/39C dual - a delta of 14C/17C dual. While the cooling performance of the Noctua is much better than the Intel retail cooler at idle, it must be said that the Noctua still does not reach the same cooling levels seen in the Thermalrights or Tuniq 120.
Cooling efficiency of the NH-U12F under load conditions was compared to the retail HSF and other recently tested CPU coolers. Load testing can be very revealing of a cooler's efficiency. A basically flat line, particularly from 3.73GHz upward, indicates the cooler is still in its best cooling range. A line that is increasing rapidly indicates a cooler nearing the end of its ability to cool efficiently. Lines which parallel the best coolers over a range of values are showing similar efficiency slopes.
 |
| Click to enlarge |
The Noctua is very efficient in cooling in the 2.93GHz to 3.73GHz overclock range, particularly with push-pull fans. It continues on to a highest overclock of 3.88GHz, which is just short of the 3.90GHz most of the top coolers in our testing have achieved. However, the sharp slope of the line from 3.73GHz to 3.88GHz indicates the Noctua is rapidly losing its ability to cool effectively at these higher overclocks. Compare this, for example, to the lines for the Thermalright Ultra 120 Extreme and the Monsoon II Lite hybrid air/TEC cooler. Both these coolers have very flat stress performance lines as the overclock rises above 3.73GHz, which suggests they are still efficiently cooling and the CPU ran out of overclock ability rather than the cooler reaching the limit of what it could effectively cool.
As stated many times, the overclocking abilities of the CPU will vary at the top, depending on the CPU. This particular CPU does higher FSB speeds than any X6800 we have tested, but the 3.9GHz top speed with the top tier heatpipe towers is pretty average among the X6800 processors we have tested with Tuniq cooling. A few of the other processors tested with the best air coolers reach just over 4 GHz, but the range has been 3.8 to 4.0GHz. Stock cooling generally tops out 200 to 400 MHz lower, depending on the CPU, on the processors tested in our lab.
Facebook
Twitter
RSS
30 Comments
View All Comments
HeatMiser - Thursday, April 26, 2007 - link
Wes -- how far does this Noctua cooler hang over the edge of the motherboard closest to the Power supply? Just trying to figure out if I have room. Thanks for the great reviews on these coolers.Bumtrinket - Friday, May 4, 2007 - link
It largely depends on the motherboard - if the CPU socket is close to the top edge, you could have problems if there's not much space between the mobo and PSU. This also applies to the Thermalright Ultra 120 and other similar designs. The design of the heatsink mounting lets you rotate it through 90 degrees, but that's not the preferred orientation.There's more info at Noctua's site, but they say this in the FAQs:
"How much space between mainboard and power supply is required when the NH-U12 is installed upright?"
"When installed upright (see picture below), the NH-U12 may jut out over the upper edge of the mainboard. In this case, the gap required between the upper edge of the mainboard and the power supply in the case is (depending on the position of the CPU socket on the mainboard) up to 25mm. Measured from the upper mounting holes of the CPU socket, the required distance to the power supply unit is 35mm."
n7 - Wednesday, April 25, 2007 - link
I am a big fan of Noctua, so nice to see their heatsink do as well as i expected.Their fans are very good @ providing a mix of performance & realtive quietness, & it appears this heatsink is no different.
I too would really like to see a TT Big Typhoon added to the reviews, mainly to see how it does compared to the new leaders, something the Typhoon used to be.
I also look forward to seeing your testbed redone with a quieter PSU.
TA152H - Wednesday, April 25, 2007 - link
OK, I read this article with some interest, and as usual find the remarks about fanless power supplies completely off-base.I mean, if you're measuring overclocking a CPU, or some high end gaming system, or some high-end Peltier cooler, I could see the point in not using a fanless power supply because such a power supply wouldn't be useful in that configuration. But, I don't buy power supplies with fans anymore, and guess what type of heatsink I go after when I buy them? Yes, very quiet ones. So, when you're testing a heatsink/fan that is made to be especially quiet, or for example, fanless video cards, you SHOULD test with a fanless power supply.
I know most of the reviewers here think shooting space aliens is what everyone does, and they need 600 watt power supplies, but it's just not true. Most people in fact don't need anything that large, so try to think within the market of the product instead of personal preferences. I'm not knocking video games, by the way, they are a much better escape than drugs, or other bad things people do, or even TV. I'm just saying that you guys need to stop thinking everyone does this and needs equipment to handle these foul machines, especially when reviewing items that would never be put in one.
customcoms - Wednesday, April 25, 2007 - link
I agree that in order to get more accurate noise ratings a more quiet system based on a fanless psu would achieve higher accuracy. However, anandtech tests have generally catered to enthusiasts and overclockers, and imo, these tests do that well. For starters, they compare cooling performance with noise as a secondary consideration (mainly to see if the cooler advertised is quiet, but as it sates in the article, they are more curious about its cooling/overclocking ability than the noise). Yes, anandtech is using what could be considered a "noisy" test bed, but X6800's and 7900GTX's don't exactly come to mind when I think "silent" pc. Nor do the power, hungry, hot Nvidia 680i based boards. I suggest looking elsewhere for reviews on components that make up silent pc's, such as http://www.silentpcreview.com">http://www.silentpcreview.com. I'll stick with anandtech's review because I could care less about noise if it lets me gain an extra 300mhz out of my processor.TA152H - Wednesday, April 25, 2007 - link
Then why review a heatsink that is focused on being quiet? You completely missed the point. I don't have a problem with them reviewing noisy items, but you don't review a heatsink/fan that is made to be especially quiet, with a noisy power supply or video card. If the heatsink were a cooling beast that didn't pay much mind to noise problems, absolutely go with appropriate stuff. People that want it will not care about noise with the other components. But when someone is considering getting a unit because it is quiet, you don't use a noisy baseline because that same person is likely to look for that same quality in their other components.Ajax9000 - Wednesday, April 25, 2007 - link
Agreed.The "proper" (but time consuming :-) thing to do would be to have a "performance" testbed (which is also investigated for noise as a secondary test), and a "silent" testbed (which is also investigated for overclockability as a secondary test). I.e. people who want "quiet" systems don't automatically excude overclocking, they are just less concerned about it (and vice versa).
Paradox999 - Wednesday, April 25, 2007 - link
Great review and comparison.As an owner of a Scythe Ninja equipped with a silent fan, I can attest to it's effectiveness.....but your batch of reviewed HSF's is missing my personal favorite, the Thermaltake Big Typhoon. It's one of the best at cooling nearby capacitors since the fan blows DOWNWARDS onto the motherboard. I've replaced the original fans for some slower and silent Silenx fans since the core2duo isn't as hot as my old overclocked 820D.
Works awesome good.
Do us all a favor and please include the Big Typhoon (and not it's weaker cousin, the Mini Typhoon) in your review
Spanki - Wednesday, April 25, 2007 - link
I couldn't get my comments posted in here, so I added them as http://forums.anandtech.com/messageview.aspx?catid...">a new thread in the forums.Deusfaux - Wednesday, April 25, 2007 - link
"We also ran brief test results comparing the cooling efficiency and noise of the Scythe SFLEX SFF21F 64cfm/28 dB-A fan that was used in Thermalright reviews and the SilenX fan used in several of the most recent reviews. Cooling results were similar with either fan on the Noctua heatsink, with the SilenX providing better cooling on some overclocks. Both fans were quiet and came in below the noise floor of our test system, but the SilenX was subjectively quieter than the SFLEX. "REALLY. I have trouble believing this, but I guess we'll find out for sure in the upcoming roundup.