Testing on Asus L1N64-SLI WS



When AMD released their FX-74 for Socket-F (1207) this was their marketing answer to Santa Clara's Kentsfield Quad and was essentially two FX-74 Dual Core processors sharing resources across the Asus L1N64-SLI WS motherboard. Unfortunately the system couldn't surpass a single Intel Quad core mounted on a NVIDIA 680 platform with similar ancillaries. Ironically neither system was considered a true Quadcore in the strict engineering sense. The difference being Intel's Dual Duo Cores were housed under the same IHS sharing the same socket while AMD's FX CPUs were linked by what must seem like miles of traces. If there was one area from which the indigenous AMD / Asus option suffered, it was Power Consumption. When HardOCP first reviewed the AMD Quad FX System they did a phenomenal job, especially the section on Power Consumption. Ultimately running a 8800GTX cards they ended up measuring a 510W draw under stress testing versus 262W on the Intel Quad Core system. To me this was the ideal system to test power supplies at or above the 1KW level. Hence this will comprise our AMD test rig as specified in the chart below.



Representing our Intel test rig will be a Q6600 SLCR on the Gigabyte GA-P35-DS3R (rev. 1.0). This system will be water-cooled via the Danger Den NVIDIA 4101. Other components can be found in the cart below.

Intel LGA-775 / AMD Socket-F Test Systems:
CPUs	1. Intel Q6600 SLCR Retail Socket-775 2. 2x S-1207 FX-74 Retail
Mainboards	1. Gigabyte P35-DS3R (BIOS F4) 2. Asus L1N64-SLI WS (BIOS 0505)
Memory	1. Mushkin XP8500 (2x1GB) 2. GEIL PC2-6400 (4x1GB)
Graphics	1. Leadtek 7950GX2 2. BFG 8800GTX
Power Supply	1. NZXT Precise 1200W 2. PCPower&Cooling 1KW-SR
Cooling	1. Danger Den 4101 NVIDIA 2. Stock AMD cooling
Storage	1. Seagate Barracuda 2x 80GB 7200.10 SATA 2. Maxtor DiamondMax 10 300GB SATA
Optical	1. Plextor PX755-SA DVD/RW 2. NEC DVD/RW
Operating Systems	Windows XP

Testing:

AMD stress testing will utilize the TestMem CPU/Memory Stress Tester S&M (Ver. 1.9.0b) on that page you can compare S&M to just about every other stress tester. For Intel stress-testing I chose Intel's Thermal Analysis Tool which can be downloaded at TechPowerUp under TAT. While each of these programs also feature logging utilities all voltages will be measured using an EXTECH MiniTec MN26 multi-meter testing directly from the connectors Live. While internal thermal diode testing has become more accurate thanks to utilities such as Core Temp, voltage logging still remains hit and miss.

Each system was overclocked to the highest stable setting without raising Vcore but instead setting the Vcore manually. This demands more of the 12VRail which supplies Vcore to the motherboard DC ~ DC power stages. The Q6600 was run at 9x400FSB while the FX-74's ran at 16x200FSB.



Temps:

In each circumstance the room temp was measured at 21C ~ 23C. Both systems were tested in an "open chassis" including the Intel system in the Thermaltake Mozart VE1000SNA with both doors removed, and the AMD system mounted "open air" fashion. This eliminates having to consider internal case temps from ambient. I recorded temps at the exhaust outlet fan on the power supply itself using the MiniTec multi-meter.

During the AMD tests temps measured at the TurboCool 1KW-SR exhaust fan read between 27.5°C at IDLE 38.1°C at LOAD.

During the Intel tests temps measured at the exhaust fan registered 25°C at IDLE and 40°C at LOAD.

Addendum:
I also measured PSU temps at the exhaust with the Mozart Tx chassis in its normal state with both panels in place, or "buttoned up."

During the AMD tests temps measured at the TurboCool 1KW-SR exhaust fan read between 37.9°C at IDLE 48.2°C at LOAD.

During the Intel tests temps measured at the exhaust fan registered 30°C at IDLE and 46.1°C at LOAD.

Noise:

Using the IntellSystems AR824 SPL meter my test-environment registers ~37dBA ambient with all major appliances off, recorded at midnight. In other words my apartment is rather noisey. From a distance of one meter the AMD rig measured 41dBA at IDLE and 45dBA at LOAD. The Intel system measured 40dBA at IDLE and 43dBA at LOAD. Others have found the TurboCool 1KW-SR to be very noisy; I have never had this experience. In all fairness this may be due to not being able to load the unit 100% although I don't know any systems that would anyway!

Conclusion:

It wasn't so long ago PC Power & Cooling was considered the undisputed leader in the power supply world. For many in the know this is still true, although as of late a plethora of new brands out there make things much more complicated. We must not forget since the introduction of the multi-rail PSU many of these new brands are the same design made by larger manufacturers such as Topower. Unfortunately almost all MR power supplies are still manufactured following the Intel 12V ATX Power Supply Guidelines mentioned in the intro. There are a few higher quality designs out there such as Dual Transformer models (again made by Topower) yet they fall short compared to a strict proprietary rail design. With several 1KW PSU Round-Up's including the 1KW-SR TurboCool little mention is given to its single-rail power, distributed from proprietary power stages. In this writer's opinion the final analysis on multi-rail designs has proven them to be inferior to the tried and true single-rail. If you take anything with you from this review read the PCPwer publication discussing single vs. multiple-rail under Power Supply Myths Exposed and of course consider other independent sources.

If I were to give this PSU a badge of honor it would read: Most Innovative Design

Recommended for



PRO
Able to power the infamous L1N64-SLI dual FX-7 system
Single Rail power distribution
Excellent Build Quality

CON
Noise

You can find the PCPower&Cooling TurboCool 1KW-SR for $429 at Newegg...

Thank you for reading, until next time!