InsaneTek

Test Setup and Results

Test Setup and Methods:

• Intel P4 2.4C
• Abit AI7 (GAT:  A-A-A-D-D)
• Radeon 9800 Pro (Omega drivers based on Catalyst 4.6)
• 2 x 512MB OCZ PC3200 Platinum Limited Edition
• Black Lite-On 16x DVD Rom
• Black Lite-On 52x CDRW
• Black Sony floppy
• DLink 802.11b NIC
• 2 x 120mm Aerocool Blue LED fans
• 1 x 80mm Aerocool Blue LED fan
• 1 x 80GB Western Digital 8MB 7200RPM
• 1 x 60GB Maxtor 2MB 5400RPM

Running Stock:

• Intel P4 2.4C -- 2.4GHZ @ 1.525v
• OCZ PC3200 Platinum Limited Edition -- 200FSB 1:1 200MHZ @ 2.6v
• Radeon 9800 Pro -- Gigacube BIOS @ 400/350
• VAGP @ 1.6v

Running Overclocked:

• Intel P4 2.4C -- 3.42GHZ @ 1.7v
• OCZ PC3200 Platinum Limited Edition -- 285FSB 5:4 228MHZ @ 3.2v
• Radeon 9800 Pro -- Gigacube BIOS @ 400/350
• VAGP @ 1.8v

Methods of testing:

Voltage readings will be from the ATX motherboard connector.  That is where the voltage lines are connected to the motherboard, and it is the most reliable source.  Readings off the Molex connectors can be off a bit due to wire efficiency, so it can vary from Molex to Molex.  For 12v load readings, Folding@Home will be used to test the stability of the 12v rail.  Folding puts a full load on the CPU, and because the CPU draws from the 12v rail, it will be the perfect test.  For 5v load readings, the video card will loop through the Nature test in 3DMark2001SE.  For 3.3v load readings, the memory will be run through the memory torture test in Prime95.

Results:

Testing the voltage readings is fairly quick and simple.  The voltages change immediately under full load.

The results were pretty impressive.  The rails over volt a little bit to allow the rails to remain stable at the specified voltages, even when there is some slight fluctuation.  The 5v and 3.3v rails are great when running at stock or at load.  The highest drop for 5v was a mere 0.02v, and it was still above 5v.  The 3.3v rail is excellent.  It already runs at 3.38v, so it's great for overclockers pumping VDimm to their memory.  The only thing I didn't like was the 12v rail.  It's true that it was still stable when powering an overclocked P4 CPU and running at full load, yet it still dropped a good 0.11v due to load.  That's a 1.7% dip in performance.  If the amps on the 12v rail were higher than 18, the 12v rail would have been much more stable.  Nonetheless, it was still kept within the 5% safe range for the 12v.