One or two fan for a CPU cooler: does it worth?

quest_for_silence · **Posted:** Tue Apr 19, 2011 4:31 am

Some data about CPU temp rise (°C), extrapolated from various SPCR reviews about some of the best coolers:

Code:

Noctua NH-C14 (NF-P14)        Noctua NH-D14 (NF-P14/12)   Prolimatech Genesis (NF-P14)      Thermalright Silver Arrow (NF-P14)
12V  29 dBA      35           12V   30 dBA      31       12V   30 dBA      34              12V   30~31 dBA   34
 9V   21 dBA      36          8.5V   21 dBA      33        9V   22 dBA      35               9V   22~23 dBA   35
 8V   18 dBA      37          7.3V   17 dBA      33        8V   19 dBA      36               8V   -- dBA      --
 7V   15 dBA      39          6.5V   14.5 dBA    34        7V   16 dBA      37               7V   16~17 dBA   39
 6V   12~13 dBA   41            5V   12 dBA      38        6V   13 dBA      39               6V   14 dBA      41

                                                          Prolimatech Genesis (Red Vortex)  Thermalright Silver Arrow (TY-140)    Thermalright U120E Rev. C (2)
                                                         12V   21~22 dBA   33              12V   26 dBA      34                   12V   19 dBA      36
                                                          9V   17 dBA      34               9V   19~20 dBA   35                    9V   14 dBA      38
                                                          8V   -- dBA      --               8V   -- dBA      --                    8V   -- dBA      --
                                                          7V   13 dBA      37               7V   14 dBA      38                    7V   12 dBA      40
                                                          6V   12 dBA      39               6V   12~13 dBA   40                    6V   -- dBA      --

Noctua NH-C14 (1 NF-P14)      Noctua NH-D14 (1 NF-P14)    Prolimatech Armageddon (NF-P14)   Thermalright Silver Arrow (1 NF-P14)  Thermalright U120E Rev. C (1)
12V  26~27 dBA   38           12V   30 dBA      32        ?V   -- dBA      --              12V   28~29 dBA   37                   12V   16 dBA      38
 9V   18~19 dBA   40          8.5V   20 dBA      34        8V   19 dBA      40               9V   21 dBA      39                    9V   13 dBA      40
 8V   16 dBA      41          7.3V   16 dBA      35        7V   15 dBA      42               8V   -- dBA      --                    8V   -- dBA      --
 7V   13 dBA      44          6.5V   14 dBA      38        6.5V 14 dBA      43               7V   15~16 dBA   45                    7V   12 dBA      43
 6V   11~12 dBA   49            5V   -- dBA      --        5V   12 dBA      48               6V   13 dBA      49                    6V   -- dBA      --

Noctua NH-C14 (1 Nexus)       Noctua NH-D14 (1 NF-P12)    Prolimatech Armageddon (SS-140)   Thermalright Silver Arrow (1 TY-140)
12V  16 dBA      39           12V   24 dBA      36        ?V   -- dBA      --              12V   23~24 dBA   36      
 9V   13 dBA      42          8.5V   17 dBA      39       10V   21 dBA      36               9V   18 dBA      38      
 8V   -- dBA      --          7.3V   14 dBA      41        9V   17 dBA      38               8V   -- dBA      --      
 7V   12 dBA      48          6.5V   12 dBA      44        8V   14 dBA      41               7V   13 dBA      42      
 6V   -- dBA      --            5V   11 dBA      51        7V   11~12 dBA   43               6V   12 dBA      46

After some testing with a Cooler Master Hyper 212+ and a couple of Enermax Cluster I have some doubts about what those number could mean for us, except for who does own a Nehalem i7 (now almost EOL).

Which are the real priorities?
What do you think about the subject, particularly with reference to the most modern (cool) CPUs (I hope the "code-table" above turn out to be enough readable/well formatted)?

Deucal · **Posted:** Tue Apr 19, 2011 5:16 am

For me it is about seeing the difference on that test rig and from that choose what is best for me (according to availability where I live).

Every noise component counts. And of course you have to able to see where you would benefit the most from.

Going for the top 3-4 will always net you the best results but at a price. It's for you to to decide.

Before I found this website I didn't even think about having a quiet computer. Now I'm semi obsessed. And all the work gone into all these tests have really helped me out to make my own systems like they should be

. Powerful and quiet.

Net result? It's reliable data which you can use for your own benefit, unlike the majority of the "reviews" out there.

ces · **Posted:** Tue Apr 19, 2011 6:36 am

Deucal wrote:

Net result? It's reliable data which you can use for your own benefit, unlike the majority of the "reviews" out there

Agreed.

I think the accompanying thoughtful commentary, assessments, conclusions and opinions are even more useful in comparison to that of other sites.

djkest · **Posted:** Tue Apr 19, 2011 9:43 am

It looks like on average we see a delta T of 3 degrees C between 1 fan and 2 fans. On the other hand, you have 5-6 dB increase in noise output. So essentially a fairly small decrease in temperature for a doubling in noise output.

Is it worth it? That depends on a number of factors, but #1 is what CPU are you using- how much cooling do you need? An overclocked 6-core intel is going to create much more heat than an undervolted i5-2500k.

I would say that on SPCR, the majority will answer no, not worth it. I guess the ultimate test would be- using 2 lower speed fans at the same noise output of 1 slightly higher speed fan, which is more effective? What provides better cooling for the same noise level?

A good cooler with a single fan at 7V or below will probably be the ideal configuration for most users on SPCR. For extreme overclockers or gamers running 2 video cards in SLI/crossfire, I think a push-pull configuration @12V will be the best because they aren't as concerned about noise as we are, and the other noise sources in their system are far louder.

ces · **Posted:** Tue Apr 19, 2011 10:43 am

djkest wrote:

It looks like on average we see a delta T of 3 degrees C between 1 fan and 2 fans. On the other hand, you have 5-6 dB increase in noise output. So essentially a fairly small decrease in temperature for a doubling in noise output.

I am not seeing that in the above data... but I am having some trouble sorting through it.

I have always assumed that two fans running slower should obtain the same amount of cooling as one fan running faster... and do so with less noise. For example two Nexus running at 500rpm compared to one running at 1000 rpm. It seems to me that the two slow fans would cool quieter and cool better than the one fast fan. But admittedly I have no data to support this.

My assumption may well be wrong. But to prove or disprove it I would like to see a comparison between (a) one Nexus and (b) dual Nexus run at various speeds and compared. I don't believe that kind of data is up in the table above, nor do I believe that data is available from SPCR testing.

If I am wrong I would be pleased if someone would be kind enough to politely correct me. I find it difficult to believe that MikeC has never made any subjective comments on this subject. Perhaps someone knows where he might have commented on it and provide a link to it. Or even where there is some clear data one way or the other (preferably from SPCR... I just don't trust sound related data from other sites).

Tephras · **Posted:** Tue Apr 19, 2011 3:10 pm

There was a comparison of 1 vs 2 Nexus fans in the Scythe Mugen 2 review, more specifically on page 6.

ces · **Posted:** Tue Apr 19, 2011 3:55 pm

Tephras wrote:

There was a comparison of 1 vs 2 Nexus fans in the Scythe Mugen 2 review, more specifically on page 6.

So what I am seeing is:

1 Fan at 12 volts (approximately 1080rpm): 16 dBA, 36°C
vs.
2 fans at 7 volts (approximately 680rpm): 12 dBA, 34°C

OR

1 Fan at 9 volts (approximately 850rpm): 13 dBA, 38°C
vs.
2 fans at 5 volts (approximately 490rpm): 11 dBA, 38°C

One or two fan for a CPU cooler: does it worth?

Who is online