Titan Fenrir & CM Hyper 212 Plus: Direct Touch Revisited

Want to talk about one of the articles in SPCR? Here's the forum for you.
danimal
Posts: 734
Joined: Mon Jun 08, 2009 2:41 pm
Location: the ether

Post by danimal » Thu May 13, 2010 4:40 pm

DAve_M wrote: But being realistic, the nexus 120mm is the SPCR reference fan of choice. Since it is running at at full speed, 12v, 100%, and is also pushing 60°C with a top heatsink like the megahalems, then for sure a more high speed fan would be needed for overclocking. The results already show that at full speed it is at 16dBA which is pretty much at the limit while being 'silent'.
the nexus 120mm fan is inferior for overclocking purposes, it puts out less pressure than the stock fan that comes with the mugen2, for instance, which you can see if you are patient enough to wade thru the rather controversial spcr cpu review thread on that cooler :) so yes, it's quieter, because it's doing less cooling at full speed.

spcr has agreed that most people will not swap out a stock fan for the nexus, especially when the fan is as good as the mugen2 fan, so putting a nexus on every cpu cooler that comes down the pike isn't relevant info for the majority of the spcr readership.

likewise, there are several people in this thread who value performance above power useage.

danimal
Posts: 734
Joined: Mon Jun 08, 2009 2:41 pm
Location: the ether

Post by danimal » Thu May 13, 2010 4:50 pm

LEO454 wrote:
This is probably your worst test to date. Its ok to address a group of people who prioritize silence but to assume none of them overclocks is a mistake.
I thought this review was pretty straight forward and am shocked to see all these confused replies. Using the reference fan, the Hyper 212 outperformed the titan at low cfm both in temperature and noise. The review was not about how much cooling either heatsink can do regardless of noise.
actually, no cooler outperformed the other, because the only requirement was that they had to keep the cpu within factory temp specs, which both of 'em did well... why would a few degrees difference at stock frequencies matter? there is nothing to be gained one way or the other, it's a useless measurement.

the fan noise, on the other hand, is valuable information.

if you were an overclocker, you'd understand why temp-controlled fans matter; running a fan constantly at full speed is rather annoying :D

LEO454
Posts: 20
Joined: Sat May 08, 2010 12:38 am
Location: CA

Post by LEO454 » Thu May 13, 2010 11:25 pm

if you were an overclocker, you'd understand why temp-controlled fans matter; running a fan constantly at full speed is rather annoying
Sorry to take this review off topic since the hyper 212 performed better, admittedly marginal, using a constant (reference fan) instead of two variables (independent fans from both companies).

I am a watercooling proponent so I would be happy to discuss fan noise in a seperate thread for constructive purposes.

dev
Posts: 43
Joined: Sun May 10, 2009 12:47 am
Location: Here

Post by dev » Fri May 14, 2010 2:06 am

Let me ask you folks something: wouldn't you find it at least funny if someone found the single situation where a heatsink is better than others? This is the case with this review: the combination of the weak Nexus fan and the relatively large Fenrir brought the 212+ in front.

I dislike both products equally but I know for a fact that the slightest increase in vcore will leave the 212+ in the dust while the Fenrir is going to be going pretty strong for a HDT product.

Maybe testing denser heatsinks with other fans, the Slipstream 140 for example is not such a bad idea after all.

MikeC
Site Admin
Posts: 12285
Joined: Sun Aug 11, 2002 3:26 pm
Location: Vancouver, BC, Canada
Contact:

Post by MikeC » Fri May 14, 2010 8:41 am

dev,

On the one hand, you're probably right that with a higher airflow fan and a hotter CPU, the Fenrir might outperform the 212+. Assuredly, it will be noisier than the 212+ while doing so.

On the other hand, you don't seem to get it. We don't care what a heatsink does with a louder fan. Info on that is all over the web, you don't need to come here for it. What you don't get anywhere else is performance of coolers under carefully controlled, consistently quiet conditions. The whole point of controlled conditions is not to change them willy nilly for this sample and that, but to keep them the same for all tested items. Some will fare better under different conditions, we don't pretend our results have universal applicability. However, if you want low noise cooling performance, our reviews are extremely valuable.

Another thing which will come to light when the latest fan tests are posted: The Nexus 120 is an exceptionally capable fan that's extremely difficult to match for noise or cooling. This is why it has remained a reference at spcr all these years.

dev
Posts: 43
Joined: Sun May 10, 2009 12:47 am
Location: Here

Post by dev » Fri May 14, 2010 10:00 am

I'm not sure I'm with you on this one Mike. Things changed since your last fan roundup, new products arrived and were well received I don't know if the old Yate loon based Nexus is still a top dog. Sure it hums nicely at 12v but is pretty weak when it comes to putting some pressure behind the air it outputs.

I see your goal is under 20dB by your standards and I have a suspicion some fans will be able to hit that. The Gentle Typhoon 1150 rpm is a definite candidate.

All I'm saying is give other fans a chance because some heatsinks might benefit.

MikeC
Site Admin
Posts: 12285
Joined: Sun Aug 11, 2002 3:26 pm
Location: Vancouver, BC, Canada
Contact:

Post by MikeC » Fri May 14, 2010 10:17 am

dev wrote:I'm not sure I'm with you on this one Mike.
Obviously not. Our goal is NOT best cooling at any cost.
The Gentle Typhoon 1150 rpm is a definite candidate.
It is gentle only by typhoon standards. The SPL may be low, but the sound quality isn't even close to the Nexus.
All I'm saying is give other fans a chance because some heatsinks might benefit.
You still don't get it -- all this will do is make the fan a variable when we don't want it to be one. Besides, weren't you arguing before to increase the CPU heat? We're not going to do that either.

dev
Posts: 43
Joined: Sun May 10, 2009 12:47 am
Location: Here

Post by dev » Fri May 14, 2010 10:38 am

Ok since this is going nowhere fast let us tackle the tdp issue. :)

My biggest gripe with your methodology atm is that is so specific it might be unique. By that I fear the results you are getting are covering a very narrow and almost unpopulated niche: people with high end cpus that buy a heatsink for that cpu use on that cpu at stock settings with a silent fan. I'm not sure those people really exist.

On the other hand it seems to me you are targeting mostly htpc users that will never use a high tdp cpu. For this reason alone your previous setup was better. Honestly I would split the testing in 2 different categories: high end products and consumerish ones. For the former I would use your current setup as for the latter I think something like an Athlon II X4 or an i3 should be more appropriate.

nyu3
Posts: 144
Joined: Wed Feb 18, 2004 6:07 pm
Location: Canada

Post by nyu3 » Fri May 14, 2010 11:48 am

What about people who buy mid end CPUs, overclock mildly, and want a quiet / silent computer?

Or more generally, people who want the fastest computer that remains silent?

mkk
Posts: 687
Joined: Sun Sep 05, 2004 1:51 pm
Location: Gefle, Sweden
Contact:

Post by mkk » Fri May 14, 2010 11:56 am

There is no TDP issue. Either the clearly laid out SPCR mission statement fits your bill, or it does not.

DAve_M
Posts: 198
Joined: Sun Sep 02, 2007 9:58 am
Location: UK

Post by DAve_M » Fri May 14, 2010 12:29 pm

dev, don't get so hung up on whether the CPU is running at stock or not. It's just a means to an end. It's just a source of heat. A way of producing a specific and exact amount of power to be dumped into the heatsink so that the temperature can then be taken.

You and a few others are right and/or realize that the Fenrir would indeed be 'better' than the Hyper 212+ when you turn up the heat. It's so simple really, a bigger heatsink with more tightly compact fins and a powerful fan will win out against a smaller heatsink with a weak fan as you put it. More surface area = better cooling capability. Bigger = better.

But that result would be totally irrelevant. This being SPCR, which is aiming to keep very quiet, all the bigger heatsink will do is increase the resistance to the airflow and not help with cooling at all. Which is why it doesn't do as well in the tests. SPCR is all about testing heatsinks in low airflow conditions to find a result that is relevant for SPCR. A test that showed the Fenrir to be better would be an invalid result, this being SPCR.

Also, apart from maybe a scythe s-flex with it's FDB bearings, the 120mm nexus is hard to beat. It incorrect to call it weak as it's static pressure should be decent enough. Call it a low RPM fan or a slow fan instead. Both the slipstream and the gentle typhoon would be worse than the nexus. The slipstream has very low static pressure so is only suitable for case exhaust fans, and the gentle typhoon has double ball bearings which if you have compared an old hard disk like a 40GB one to a modern one with FDB bearings, you'll realize how bad DBB can be.

nyu3
Posts: 144
Joined: Wed Feb 18, 2004 6:07 pm
Location: Canada

Post by nyu3 » Fri May 14, 2010 12:37 pm

Actually, Slipstream performs just fine on heatsinks.

Some relevant results from http://www.silentpcreview.com/article961-page5.html

Slipstream:
15 dba - 14 C rise
12 dba - 17 C rise

Nexus:
16 dba - 15 C rise
12 dba - 19 C rise

----------------------------------------------
Also, check this page for huge test of various fans on the Megahalems ( http://www.overclock.net/air-cooling/72 ... 112-a.html ).

Note how the 1200rpm slipstream performs versus the 1200rpm S-Flex, 1300 rpm Yate Loon D12SL, and 1300rpm Noctua P12.
Image

Mr. Perfect
Posts: 74
Joined: Sat Nov 27, 2004 2:53 pm
Location: United States

Post by Mr. Perfect » Fri May 14, 2010 3:50 pm

DAve_M wrote:dev, don't get so hung up on whether the CPU is running at stock or not. It's just a means to an end. It's just a source of heat. A way of producing a specific and exact amount of power to be dumped into the heatsink so that the temperature can then be taken.
Makes sense to me. A 130 watt TDP CPU is going to push a heatsink like any other 130 watt TDP, whether it's a bone-stock i7 or it's a heavily overclocked i5. What SPCR is testing here is how to silently cool any CPU that's doing 130 watts.

Oh, and I'm one of the people who buy a fairly-high spec CPU and then run it at stock settings. :) I'd rather buy an i7-870 with guaranteed, crash free performance(and hyperthreading) then grab an i5-750 and go through the hassle of overclocking what may or may not be a good sample. That and it's pretty had to kill a CPU at stock settings, I haven't manged to do it yet. :)

dev
Posts: 43
Joined: Sun May 10, 2009 12:47 am
Location: Here

Post by dev » Fri May 14, 2010 10:10 pm

As it turns out I own most of the fans mentioned before and since the discussion is about subjective factors I can tell that for me the Gentle Typhoons at around 1000 rpm have the most decent harmonics. I'm not sure how one should judge the Yate loon based fans or the 120mm Slipstreams because sample variance comes into play. The Nexus I own is alright but nothing special as is the Slipstream. And for heatsink use the Slipstreams are the best choice hands down.

You see I've always felt a fixed rpm fan is best suited for the case while pairing a pwm one to a heatsink. The are also other fans with good acoustics that undervolt nicely, Noctua NF-P12, Noiseblocker Xl or PL series to name just a few.

As for cpus an i7's selling point is the multithreaded performance which an i5 can never match. As for cpus dying from overclocking on air at decent voltages thats more of a myth really especially these days. I would worry more about the accuracy of the sensors.

DAve_M
Posts: 198
Joined: Sun Sep 02, 2007 9:58 am
Location: UK

Post by DAve_M » Sun May 16, 2010 7:42 am

nyu3 wrote:Actually, Slipstream performs just fine on heatsinks.
Indeed. Although not so close to that S-Flex. I was sure that the slipstreams had a small issue with static pressure, so would therefore struggle when attached to a heatsink. Maybe I was thinking of the 120mm x 12mm slipstream. That is 1200 RPM also.

Getting off topic a bit but it's pretty hard to kill a CPU at overclocked settings too. All the way back to at least the athlon XP days, Intel CPUs could cope with complete heatsink failure. Intel sets conservative specs on max voltage. So just don't go over the limit for voltage and it will be fine.

Haych
Posts: 106
Joined: Sat Oct 17, 2009 11:16 am
Location: Lancashire, UK

Post by Haych » Sun May 16, 2010 9:28 am

I must say that the results for the Hyper 212+ has piqued my interest, as has the debate regarding the Nexus fan being used on that heatsink with a high TDP cpu. After a little research I've figured I could probably get the Hyper 212+ and two Nexus 120's for marginally more (once P+P is factored in) than the cost of a Scythe Mugen 2. I'm guessing that using two Nexus in a push/pull configuration will add maybe 2 or so dbA to the total noise produced, so would anyone like to throw some theoretical numbers at me for what sort of temps to expect with two Nexus on a 212+ cooling a Phenom II x4 965 at stock (3.4Ghz) and at, let's say, an overclock to 3.8Ghz.

No need for accuracy, I'm just throwing the question out there.

MikeC
Site Admin
Posts: 12285
Joined: Sun Aug 11, 2002 3:26 pm
Location: Vancouver, BC, Canada
Contact:

Post by MikeC » Sun May 16, 2010 11:12 am

Haych --

imo, a mugen 2 w/ 2nd stock fan will outperform the 212 w/ dual fans on a really hot cpu. The mugen has substantially bigger surface area.

Haych
Posts: 106
Joined: Sat Oct 17, 2009 11:16 am
Location: Lancashire, UK

Post by Haych » Sun May 16, 2010 11:53 am

Thanks Mike :)

RoGuE
Posts: 540
Joined: Mon Jul 06, 2009 9:11 am
Location: Massachusetts, USA
Contact:

Post by RoGuE » Sun May 16, 2010 12:03 pm

dev wrote:
My biggest gripe with your methodology atm is that is so specific it might be unique. By that I fear the results you are getting are covering a very narrow and almost unpopulated niche: people with high end cpus that buy a heatsink for that cpu use on that cpu at stock settings with a silent fan. I'm not sure those people really exist.
I'm always amazed when I read posts like this..some people just dont get it. When I say "dont get it", I'm referring to scientific experimentation in general. In a way, SPCR reviews of heatsinks are like mini experiments. There must be a control, in this case, it's a quiet fan. Whether or not you chose to use that fan in your own home, is irrelevant. Allllllll that matters with SPCR reviews, is that they are consistant, and give us, as consumers, an accurate and comparable review on which to base a purchase decision! The accuracy of the cooling, extreme silence of the fan, etc, are completely irrelevant as long as all the reviews are done with a control and comparable.

...I have a looming feeling I'm wasting my time writing this, but I thought it needed to be said.

Thanks for the continued reviews SPCR!

Vibrator
Posts: 94
Joined: Sun May 31, 2009 9:46 am
Location: British Columbia

Post by Vibrator » Sun May 16, 2010 2:44 pm

danimal wrote:
DAve_M wrote: the nexus 120mm fan is inferior for overclocking purposes, it puts out less pressure than the stock fan that comes with the mugen2, for instance, which you can see if you are patient enough to wade thru the rather controversial spcr cpu review thread on that cooler
That is untrue.

Slipstreams actually have inferior static pressure to most other 120mm fans out there.

http://www.xtremesystems.org/forums/sho ... p?t=223391
http://www.youtube.com/watch?v=LE4Q2N3RL8E

The Nexus fan on the other hand, is a rebranded Yate Loon D12SL.
Yate Loons are known for their good static pressure and low cost.

MikeC
Site Admin
Posts: 12285
Joined: Sun Aug 11, 2002 3:26 pm
Location: Vancouver, BC, Canada
Contact:

Post by MikeC » Sun May 16, 2010 9:51 pm

You're welcome Rogue, and I'm glad you understand. 8)

W/regard to the Nexus 120 fan, I was under the impression for a long time that it was the same design as the standard sleeve bearing Yate Loons I obtained a bunch from NCIX a while ago. It turns out I was wrong, I didn't examine them closely enough. The Nexus 120's blades are more curved, and their edges are thinner. The blades look better formed, more aerodynamic, slightly wider. Sort of the way the old Panaflo 80 fan blades looked more finely formed than most other fans of its day.

The Nexus 120 also sounds and performs better than the Yate Loon. It's not a dramatic difference but it is there. More on this in the next fan roundup.

DAve_M
Posts: 198
Joined: Sun Sep 02, 2007 9:58 am
Location: UK

Post by DAve_M » Mon May 17, 2010 3:24 am

MikeC wrote:More on this in the next fan roundup.
Whoop! Can't wait lol.

PartEleven
Friend of SPCR
Posts: 279
Joined: Sun May 06, 2007 10:37 am

Post by PartEleven » Mon May 17, 2010 11:19 am

RoGuE wrote:
I'm always amazed when I read posts like this..some people just dont get it. When I say "dont get it", I'm referring to scientific experimentation in general.
+1

This is the feeling I get when reading a lot of user feedback comments on tech sites. I was thinking about posting about this, but you've worded it better are more concisely than I could have. SPCR is one of the few sites that regularly provides a control sample to compare to.

danimal
Posts: 734
Joined: Mon Jun 08, 2009 2:41 pm
Location: the ether

Post by danimal » Mon May 17, 2010 4:39 pm

Vibrator wrote:
danimal wrote: the nexus 120mm fan is inferior for overclocking purposes, it puts out less pressure than the stock fan that comes with the mugen2, for instance, which you can see if you are patient enough to wade thru the rather controversial spcr cpu review thread on that cooler
That is untrue.

Slipstreams actually have inferior static pressure to most other 120mm fans out there
not according to the spcr mugen2 review:

slipstream: 12V 24 dBA 33°C 12 0.15
nexus: 12V 16 dBA 36°C 15 0.19

cpu cooler testing isn't a controlled experiment, because every cpu cooler presents a different impedance to the fan.

RoGuE
Posts: 540
Joined: Mon Jul 06, 2009 9:11 am
Location: Massachusetts, USA
Contact:

Post by RoGuE » Mon May 17, 2010 6:49 pm

danimal wrote:
cpu cooler testing isn't a controlled experiment, because every cpu cooler presents a different impedance to the fan.
maybe so, but let's hope that if a heat sink has more impedance, it also has tighter fin spacing, and therefore more surface area. That said, if you use the same fan every time (the control), than you can get a fairly accurate comparison as to which heat sinks are more efficient at dissipating heat! Like I said before, the accuracy of the heat flux through the CPU isn't what we're after, but rather the accuracy of the comparison is what we all want.

Also, I think it needs to be said...a fan is a fan. Yes, some are better than others in terms of silence, but I get the feeling some people here think some fans are much much better performers than others! In reality, this simply isn't true. SPCR's reviews with a control fan gives us a semi-accurate, and comparable review. What more could you ask for? Test it with every fan on the market? come on...


*edited for grammar

KayDat
Posts: 222
Joined: Wed Aug 15, 2007 7:37 pm
Location: Melbourne, Victoria, Australia

Post by KayDat » Mon May 17, 2010 9:08 pm

RoGuE wrote:Also, I think it needs to be said...a fan is a fan. Yes, some are better than others in terms of silence, but I get the feeling some people here think some fans are much much better performers than others! In reality, this simply isn't true. SPCR's reviews with a control fan gives us a semi-accurate, and comparable review. What more could you ask for? Test it with every fan on the market? come on...
Well yes...and no. A fan is a fan, but then you have amount of noise produced for amount of air moved, and also static pressure. Also noise profile of the fan, which depends on the motor, fin profile and other various factors.

DAve_M
Posts: 198
Joined: Sun Sep 02, 2007 9:58 am
Location: UK

Post by DAve_M » Tue May 18, 2010 6:31 am

No no no... 1st of all, of course it is a controlled experiment! The CPU coolers are the things that are being tested! Everything else remains the same.

2nd, a fan is not just a fan. As if any old fan will do. SPCR is the place where the specific fan in use matters more than any other place I've come across. Every aspect of the fan in use as well as their number and even the case are all important factors and can make a significant difference. The difference between inaudible and audible is very slight indeed.

ilovejedd
Posts: 676
Joined: Tue Oct 21, 2008 4:14 pm
Location: in the depths of hell

Post by ilovejedd » Tue May 18, 2010 7:39 am

danimal wrote:
Vibrator wrote:Slipstreams actually have inferior static pressure to most other 120mm fans out there
not according to the spcr mugen2 review:

slipstream: 12V 24 dBA 33°C 12 0.15
nexus: 12V 16 dBA 36°C 15 0.19

cpu cooler testing isn't a controlled experiment, because every cpu cooler presents a different impedance to the fan.
I think the SPCR CPU cooler reviews are pretty controlled. Note that the cooling performance between various coolers are being compared at a given sound level which is what SPCR is after. Granted it would be nice if SPCR had a similar set-up to FrostyTech (mock processor providing a constant synthetic load instead of using an actual CPU).

MikeC
Site Admin
Posts: 12285
Joined: Sun Aug 11, 2002 3:26 pm
Location: Vancouver, BC, Canada
Contact:

Post by MikeC » Tue May 18, 2010 8:11 am

ilovejedd wrote: Granted it would be nice if SPCR had a similar set-up to FrostyTech (mock processor providing a constant synthetic load instead of using an actual CPU).
The trouble with this is that it is not CPU, and there is no CPU which behaves just like the synthetic load. Lots of detailed discussion on umpteen tech web sources, but the gist of the limitations for SPCR are:

1) mounting does not always replicate real app on a motherboard, because the die and "board" don't have the exact same physicality as a motherboard. This means the tension is never quite the same as in real use -- we've always insisted that the mounting efficicy is part and parcel of the heatsink.

2) different secondary heat paths -- ie, heat loss through the back, sides, and other ways are different, most likely not quite the same in direction or quantity.

3) the die itself is also quite different. In a CPU, the heat source is a small portion of the heat spreader. The heater coil (or simular) under the copper die simulator usually covers most of the whole die. 100W through that coil doesn't give the same thermal density as the CPU dissipating the same power, and therefore does not get as hot.

In the end, the variances from a real CPU are too many. With our testing platform, even though it is just one unique CPU on one unique motherboard, it is still a real CPU on a real motherboard.

*All this is specific to testing heatsinks.*

dev
Posts: 43
Joined: Sun May 10, 2009 12:47 am
Location: Here

Post by dev » Tue May 18, 2010 9:55 am

Mike I couldn't have said it better myself. :)

On the Nexus vs Slipstream vs Yate loons the issue is pretty simple: the Nexus is a round edged Yate loon modified by the dutch company. Slipstreams have surprising qualities, the 1200 rpm version for example outperforms the famous Noctua NF-P12 in every respect in spite of the fact that Noctua qualifies their fan as a high pressure designed for heatsinks product.

Post Reply