The NCU-2000 has arrived. I couldn't wait to try this thing and I wasn't disappointed. A picture of the monster installed:



The first thing at hand was testing tempuratures. I like CPUburn for these tests but I'm using dual instances of SETI for the first test and then again at the end because it gave me a good look at the difference the NCU-2000 makes over the Shuttle's ICE cooler it is replacing. I've been running dual instances of SETI on my 3.0C @ 2.0 ghz with a .9500 Vcore under a passive ICE cooler with a temp of approximately 60C. According to this chart and this calculator the CPU should be dissipating a maximum of about 20 watts at these settings. Replacing the ICE cooler with the NCU-2000 and running dual instances of SETI with the same CPU settings nets a max CPU temp of 44C. Big improvement!!

Next I changed the CPU clock and Vcore to 2.49 and 1.1625, respectively, and ran dual instances of CPUburn. Using the same chart and calculator as above the CPU should be dissipating a maximum of close to 38 watts. This netted a max CPU temp of 53C.

Finally, the big test, running dual instances of CPUburn with the CPU at full clock and a Vcore of 1.4000, a maximum dissipation of about 66 watts. This setting netted a punishing 69C, not too practical as a little rise in ambient temps would send it overboard. But wait! What do I care about CPUburn temps? It's not like I'm going to be running CPUburn 24/7. I will, however, be running SETI 24/7. So what are the temps with SETI? A much more comfortable 62C, only slightly above the ICE cooler at 20 watts.

In an ideal open air 21C ambeint it looks like the practical limit of the NCU-2000 is close to 60 watts. Absolutely outstanding!

Not just a CPU cooler, it's also art!

Ahem.... where's my wallet

Thats nice...btw do you have any case fans for airflow? As from your pics it doesn't seem to have any; and that seems to be pretty good too

No case fans, not even a case. Just a wide open ghetto setup.



The two on the right are Shuttle Zen motherboards with the original Ice coolers. On the left is the NCU-2000.

I think maybe what would help your temps by at least a few degrees celcius if you had ONE 80mm L1A on top and created a duct attacking all three of those heatsinks to that one fan.

It wouldnt add i dont think ANYTHING to the noise, and for a few degrees, that's awesome.

The duct would have to be made of something stretchy (like the lycra/polyester one that i'm making right now. Sounds ridiculous? Wait till its done.)

I'm sure SPCR's upcoming review will give us an idea of what to expect from this thing when used near a fan, be it case fan or PSU fan. You see, SPCR is still on the practical side of things.

As for me, I've sunken into the bottomless quagmire of silent computing. A whole so wretched that even the sound of a single undervolted LIA is not to be tolerated. A place where managing 3 ghz passive, even at 62C, for a measly $50.00 is equal to Nirvana.

holy crap that's a good idea, thanks (even though i'll probably never use it)

you'd just need like... two metal rings or squares and a way to keep them the proper distance from each other

something similar

http://forums.silentpcreview.com/viewtopic.php?t=15450

The difference between my plan and that one is mainly the ridges. A cloth duct will not only be as flexible and usable, but will be smooth, not allowing the bouncing of soundwaves and impedence of airflow.

i think a stretchy cloth duct would be even easier than trying to fit dryer hose type stuff too, as long as it wasn't long or really curved - you'd basically just have to make a bracket for each end, then cut a little less fabric than you'd need to wrap around them.

I was thinking of putting this on my 2500xp-m using some extra zalman adapters until I can get an a64. How stable do you think it would be on an unprotected chip?