- Jun 30, 2004
- 16,632
- 2,027
- 126
I did several things with my computer this month.
I had been troubleshooting an infrequent, intermittent problem with my RAM -- first to identify it as a RAM problem, and then deal with it. Dealing with it meant mostly running the four modules of two kits at their spec. I finally decided to replace them with a two-stick set that ran spec at the settings that required overclocking of the previous modules.
Once I resolved my RAM problem (but before buying the replacement), I decided to "vet" or validate a higher CPU speed, so I went from 4.6Ghz to 4.7Ghz. This also increased my temperatures slightly, so I found one of my core sensors reporting 84C under load, with the average for all cores was around 78C.
The other thing I decided to do -- which I should've done at the beginning (but didn't) -- was to replace those limp Noctua (P12 and P14) fans that come with the D14. I was already sure that I could improve the cooler's performance with a single fan of higher top-end CFM and PWM for total thermal control off the mobo CPU_FAN plug.
There may be other fans on the market that fill the bill, but I had selected the Akasa Viper 140 R. "R" stands for "round," and the Viper is identical in size and fan-holes to the Noctua p14. In other words, I replaced both fans with a new fan installed in the center of the cooler -- between the two "towers." There is a slight complication with the springs that secure the Noctuas to the CPU-cooler fins. Of course you want to use the same springs, and there's nothing with my Akasa choice that would prohibit it, but you begin to wonder how Noctua was able to secure the springs to the fans in the first place. When you get to the point of using Noctua's little black-plastic pins to hold the springs in the fan-holes, make sure you have an assortment of the smaller nylon wire zip-ties handy. You shouldn't need further guidance here -- you'll figure it out. You end up using two of the four Noctua pins for the replacement fan, and two zip ties.
TO THE POINT: I had to do several runs of LinX testing to reach optimum settings for the increased clock-speed -- the most important of them was a marathon of 50 LinX iterations. So I can -- and did -- tell you about the temperatures -- all recorded at a room-ambient of 79F.
With the replacement fan, all temperatures have now dropped as much as 5C degrees with similar LinX iterations. The core average at load is now a little more than 73C. The "whacky" high temperature for Core #1, which had been 84C, now held steady around 78C.
And I haven't noticed much difference in noise -- no difference at all -- while I put the computer through its load-test paces.
ADDENDUM: Oh. Just to avoid needless questions. I said "I didn't notice any change in the noise level at full CPU load." This is true, but folks who lean toward "limp" fan and airflow would not have installed my Panaflo NMB-MAT 120x38mm case exhaust fan, which ramps up to 2,500 rpms for CPU temperatures over 60C (my choice of setting). Mostly -- the noise is just "Whoosh!" I am thinking I could replace this Panaflo with a second Akasa Viper and either a 120-to-140 fan adapter or a simple foam-board duct. And I'm thinking it would be more quiet at full load.
Then, there's the question about "kruft" and "before/after" measurement. The Kruft had been blown out of the D14 before I replaced the Noctua fans, and before I benched the peak-core and average temperatures.
So -- Yes!! YEs! Yes! Yes! -- you can improve the stellar performance of the NH-D14 cooler by replacing two fans with one better. My measurement shows 5C improvement -- YMMV.
This also has implications for comparison reviews by reputable reviewers -- Tom's Hardware, Anandtech, Bit-Tech and others. Those reviews only compare the D14 as it suits them, sometimes excluding the D14 if it overpowers everything else tested and their focus is "some other cooler." But take a review -- maybe at Tom's -- entitled something like "Socket-2011 air-cooling roundup." If you were at a crossroads between air and water, it might be incorrectly influenced by the Noctua default fans.
So just remember . . . . You can improve the D14's performance over those reputably-measured results.
I had been troubleshooting an infrequent, intermittent problem with my RAM -- first to identify it as a RAM problem, and then deal with it. Dealing with it meant mostly running the four modules of two kits at their spec. I finally decided to replace them with a two-stick set that ran spec at the settings that required overclocking of the previous modules.
Once I resolved my RAM problem (but before buying the replacement), I decided to "vet" or validate a higher CPU speed, so I went from 4.6Ghz to 4.7Ghz. This also increased my temperatures slightly, so I found one of my core sensors reporting 84C under load, with the average for all cores was around 78C.
The other thing I decided to do -- which I should've done at the beginning (but didn't) -- was to replace those limp Noctua (P12 and P14) fans that come with the D14. I was already sure that I could improve the cooler's performance with a single fan of higher top-end CFM and PWM for total thermal control off the mobo CPU_FAN plug.
There may be other fans on the market that fill the bill, but I had selected the Akasa Viper 140 R. "R" stands for "round," and the Viper is identical in size and fan-holes to the Noctua p14. In other words, I replaced both fans with a new fan installed in the center of the cooler -- between the two "towers." There is a slight complication with the springs that secure the Noctuas to the CPU-cooler fins. Of course you want to use the same springs, and there's nothing with my Akasa choice that would prohibit it, but you begin to wonder how Noctua was able to secure the springs to the fans in the first place. When you get to the point of using Noctua's little black-plastic pins to hold the springs in the fan-holes, make sure you have an assortment of the smaller nylon wire zip-ties handy. You shouldn't need further guidance here -- you'll figure it out. You end up using two of the four Noctua pins for the replacement fan, and two zip ties.
TO THE POINT: I had to do several runs of LinX testing to reach optimum settings for the increased clock-speed -- the most important of them was a marathon of 50 LinX iterations. So I can -- and did -- tell you about the temperatures -- all recorded at a room-ambient of 79F.
With the replacement fan, all temperatures have now dropped as much as 5C degrees with similar LinX iterations. The core average at load is now a little more than 73C. The "whacky" high temperature for Core #1, which had been 84C, now held steady around 78C.
And I haven't noticed much difference in noise -- no difference at all -- while I put the computer through its load-test paces.
ADDENDUM: Oh. Just to avoid needless questions. I said "I didn't notice any change in the noise level at full CPU load." This is true, but folks who lean toward "limp" fan and airflow would not have installed my Panaflo NMB-MAT 120x38mm case exhaust fan, which ramps up to 2,500 rpms for CPU temperatures over 60C (my choice of setting). Mostly -- the noise is just "Whoosh!" I am thinking I could replace this Panaflo with a second Akasa Viper and either a 120-to-140 fan adapter or a simple foam-board duct. And I'm thinking it would be more quiet at full load.
Then, there's the question about "kruft" and "before/after" measurement. The Kruft had been blown out of the D14 before I replaced the Noctua fans, and before I benched the peak-core and average temperatures.
So -- Yes!! YEs! Yes! Yes! -- you can improve the stellar performance of the NH-D14 cooler by replacing two fans with one better. My measurement shows 5C improvement -- YMMV.
This also has implications for comparison reviews by reputable reviewers -- Tom's Hardware, Anandtech, Bit-Tech and others. Those reviews only compare the D14 as it suits them, sometimes excluding the D14 if it overpowers everything else tested and their focus is "some other cooler." But take a review -- maybe at Tom's -- entitled something like "Socket-2011 air-cooling roundup." If you were at a crossroads between air and water, it might be incorrectly influenced by the Noctua default fans.
So just remember . . . . You can improve the D14's performance over those reputably-measured results.
Last edited:
Facebook
Twitter