Luda

Ive been running this cooler for about 6 months now, and decided to post up a review of its performance, since its one of the cheapest ways to move to watercooling.


Xigmatek AIO S80DP:

• Brand XIGMATEK
• Model AIO-S80DP
• SpecType Liquid Cooling System
• Block Compatibility:
  Socket 478/754/939/940/LGA775/AM2
  Intel Pentium 4 Processor / Extreme Edition
  Intel Pentium D Processor
  Intel Celeron / Celeron D Processor
  AMD Athlon 64 / 64 FX / 64 X2 Dual-core
  AMD Sempron
  AMD AM2
• Block Dimensions 70 x 54 x 14mm
• Block Material Copper
• Pump Bearing Ceramics Shaft
• Pump Noise 19 dBA
• Radiator Dimensions 80 x 88.7 x 22mm
• Radiator Material Aluminum
• Radiator Fan Dimensions 80 x 80 x 25mm
• Radiator Fan RPM 1800 - 3200 ±10% RPM
• Radiator Fan Air Flow 27.68 - 52.72 CFM
• Radiator Fan Noise 20 - 32 dBA

All temperatures are pulled from Everest ultimate, using the on screen display on a secondary monitor during stress testing.

While the AIO-S80DP has gotten some bad reviews, my personal experience is very much the opposite. I ran stock cooling for about 4 months on this machine, with about a 50% overclock [200x10.5 stock, overclocked to 300x10] and was seeing temps as high as 75C full load, and idling around 45-47C. This was far from satisfactory. so i made a run up to my local microcenter to take a look at what they had. i was planning on going with one of the standard cooling from the big boys in the air cooling realm. but this S80DP caught my eye, all in one watercooling [which is the only way i could go watercooling in my case, since its cramped already] for under 50$ [current price is about 36$ from newegg] i decided then that i had to try this. so i picked it up and made a mad dash home to get it installed.

Installation: Installtion was no worse then any other aftermarket cooler, it comes with bracketry for most modern sockets. very straighforward, including removing the board, install the S80DP and reinstalling board it took me about 20 minutes. a quality thermal paste is always recomended, even for stock applications. i decided to go venture from the normal path of Artic Silver 5 and picked up some of xigmatek's own thermal paste [since Microcenter was oddly out of AS5].

First impression:
On powering the system on my first thought was loud. with the radiator fans at full tilt it gets obscenely loud [not 92mm Vantec tornado loud, but louder then the 8800GT fan at 100%]. This was solved by configuring the bios to run the fan @ 40% until the CPU hit 45C, then turn it up to 100%. after saving these settings i could not tell there was a fan in the assembly, at 40-50% the assembly is very close to silent.

Synthetic performance: I measured performance in 3 separate CPU clock settings. All load test were down with everest System stability test.

Ambient Room temp is 26-27C through all test.
test results are in idle/load format

here are the temps from the stock cooler for reference to the S80DP

Stock cooler:
stock speed: 38C/47C
300x10 OC: 45/75C

Xigmatek AIO S80DP:

Stock speed: 21C/41C
300x10-1.45V OC: 24C/44C
300x10-1.45C W/ CnQ: 21C/44C

as you can see from the numbers above the performance is phenominal in the system stabilty tests. Keep in mind that this the fan is still @ 40% during all the testing, since it doesnt kick up to 100% until 45C.

For fun i turned the fan up to 100% and enabled cool n quiet to cut the multi down to 4x and the voltage down to 1.0V, which enabled me to idle at a whopping 18C.

Real World Performance:

Synthetic test are nice for pushing maxing the heat output, but they miss a crucial aspect of the overall cooling scheme, system heat. some cases this is not important, as the overall system flow keeps the system heat from heatsoaking on the CPU. unfortunatly my case is not one of those, it is a HTPC case, and has little to no air flow. so i decided to fire up some very intensive applications and see what the numbers looked like.


Constants:
CPU running @ 300mhzx10 w/ 1.45V vcore
8800GT running @ 750/1800/2000
8400GS running @ stock
AIO S80DP - 40% fan speed until 45C, then 100%.

i decided on 2 test for this, 3dmark Vantage, and Crysis, both are very GPU and CPU intensive.


3dmark Vantage: Standard settings - Max temp: 45C Average: 43C

Crysis: DX10|High|1440x900 - Max temp: 47C Average: 45C

as you can see the heat output from the video card(s) affected the temps on the CPU very slightly, still well within what i consider acceptable, but something to think of.


Conclusion:
For the price i would recomend this to anyone looking for a solid aftermarket solution. performance is on par with some higher end coolers, but you also get the benefit of near silent operation if you set it up correctly in your bios.

My only complaint is that i had to switch to 2x1GB sticks of ram instead of my previous 4x1GB due to the size of the cooler:



Pictures:

install on board:


compared to the Accerlero S1: