|
Mega Socket
A Heatsink Shootout The success of the AMD T/Bird/Duron has led to myriad socket A and compatible heatsinks in the market. Everyone is jumping on the T/Bird bandwagon hoping to cash in on the upward crest of frenzied buying. However, consumers are often left to fend for and ponder over each purchase of upgrade, especially the DIYers and overclockers. With so many of them, we are truly spoilt for choice.... which one is THE BEST? Well, it took all of 36 days to decide on the 43rd most powerful man in the world, President George W. Bush of the US of A. It will take a couple of pages to decide the most powerful air cooled socket A heatsink available on the market right now. Time to separate the boys from the men ................. Testbed setup...
The test is simulated in my "Typhoon" Casing. It consists of 5 x 120mm and 1 x 90mm high speed fans for 560 CFM worth of air movement, bringing the case temperature within 50C of ambient. In this situation, the respective heatsink will constantly receives fresh cool air to maximize heat dissipation to the atmosphere. The positioning of the fans is also optimized to eradicate dead zones and to prevent heat lock in the casing. Temperature will be taken using a bead probe bonded to the edge of the CPU core with Arctic Silver Thermal Adhesive. The surrounding of the probe is finally insulated with a layer industrial sealant. The probe is connected to a Fluke 54-2 thermometer for data logging over a period of 15-20 minutes. During this time, a barrage of applications e.g. Prime95, Quake III demo etc will be executed simultaneously to stress the CPU at 100% load.
Thermocouple bonded to the core using Arctic Silver adhesive.
PC-150 rams for the shootout. Both are
capable of 160 2-2-2. The professional Fluke 54-2 thermometer. IR port (top).
Creative Geforce 2 GTS Ultra
3Ware Escalade 6400 RAID Controller w/ Silver Cables.
Philips Acoustic Edge Sound Card. What I am trying to achieve is recreating the REAL WORLD scenarios. Many may debate the use of the above casing but bear in mind that there are just too many variables / possibilities. I am not going to give you 'lab-test' results. Telling you that a certain heatsink has the best efficiency but performs differently in your situation. Testing a component in a controlled environment (constant heat load in open air) is good for gathering data but it seldom hold true in reality. I am not claiming that my method of testing is the best but it should be sufficient.
O/C Typhoon Casing + Silver Series Rounded Cables
DIY fanbus + 650W PSU + rear & chimney exhaust fans.
Choice of KT133A Board. Left: Soltek SL75KAV-X Right: A7V133 Top: MSI K7T Turbo-R The Soltek was chosen for having the most clearance around the socket. NOTE: All default fans will be replaced by the Delta 38 CFM for the test. All thermal interface material will also be removed and replaced with a layer of Arctic Silver Thermal paste. How to read the chart?
|
|
Copyright 1998-2003
|
