Test Setup / Conclusion



I ended up testing the HX620 over a period of several weeks and across several systems, including the AOpen i975X a-YDG / Yonah 2600ES. Due to its low power consumption I eschewed that system from test results. Most reviewers have a very limited amount of time with which to test products. While I certainly do not think it's feasible to test all products for extended periods, testing a product for a few days does not, in my opinion, give an accurate result. I learned a great deal about electronics and electronic circuitry from my Audiophile days. Of paramount importance was the critical "break-in period." This was especially true in high end audio since one would literally "hear" components breaking in.

As the expansion and contraction of molecules effects micro-circuits, so too does this transformation affect the sound produced by those components. Tonality, depth, high's, low's, midrange and perhaps most dramatic, the ability to image. This is especially true with solid state hardware such as Krell stereo amplifiers which may sound rather harsh out of the box until the circuits break in, at which time the sound has undergone a literal metamorphosis. This is the reason most high end amplifiers never really shut-down, instead they utilize stand-by switches which trickle low levels of current keeping transistors and devices "warm." For the same reasons it's better to leave all electronics, including your PC on if used daily. And for the same reason I feel testing products in less then 5-days just doesn't bode well, nor does it give an accurate picture of performance, not to mention longevity.

1. Intel / 2. AM2 Test Systems:
CPUs	1. Intel Conroe 6400 Retail Socket-775 2. AM2 3800X2 Retail
Mainboards	1. Asus P5W DH Deluxe (BIOS 1506) 2. Asus M2N32-SLI Deluxe/Wireless (BIOS 0706)
Memory	1. Crucial Ballistix PC2-8000 (2028MB) 2. Super Talent PC2-8000 T1000-UX2G4 (2x2048MB)
Graphics	1. AOpen Aeolus 7800GTX-DVD256 2. Leadtek PX7950GX2 TDH (2x) QUAD-SLI
Power Supply	1. Corsair CMPSU-620HX 2. PCPower&Cooling Silencer 750EPS12V 3. Silverstone Zeus ST85ZF
Cooling	1. Swiftech Apex Ultra 2. Danger Den NVIDIA 4101-kit
Storage	1. Maxtor 300GB SATA 2. 2x Seagate Barracuda 80GB SATA Perpendicular
Optical	1. Plextor PX-755SA DVD/RW 2. NEC DVD/RW
Operating System	Windows XP Home SP2+

Testing the HX620 consisted of powering two separate systems.

First in our Core Duo E6400 / Asus P5W DH Dlx and then an AM2 3800X2 / Asus M2N32-SLI Dlx. On the latter I ran two Leadtek PX7950GX2-TDH graphic cards in pseudo-QUAD SLI (pseudo since it's not four individual cards). On the Core Duo E6400 / P5W DH Dlx system a single AOpen Aeolus 7800GTX card was used. Measuring DC-voltages the Extech Minitech MN-26 leads were inserted into the reverse end of the 24-pin motherboard connector and 4-pin baseboard (CPU) connector as well as measuring Molex and 6-pin external video. Each system was run at default speed and then overclocked with all measurements recorded at full LOAD. On processor's this was produced using the program S&M v. 1.8.2. (160)

Noise measurements were taken using the Smart Sensor AR-824 SPL meter from a distance of 1-meter. For this test the following results were obtained. Ambient room noise = 31dB(A)
Test system 1. 3800X2 - Asus M2N32-SLI Dlx
Default speed - Idle = 39dB(A)
Default speed - LOAD = 42dB(A)
Overclocked - Idle = 39dB(A)
Overclocked - LOAD = 43dB(A)
Test system 2. Core Duo E6400 - Asus P5W DH Dlx
Default speed - Idle = 33dB(A)
Default speed - LOAD = 34dB(A)
Overclocked - Idle = 33dB(A)
Overclocked - LOAD = 36dB(A)

The large discrepancy in noise between the Core Duo and 3800X2 was due to the AM2 system running in QUAD SLI which constantly taxed the unit despite or regardless of, CPU speed. Onto voltages.



Conclusion:

To be honest I tend to cringe whenever I learn a specialty company in the technology sector decides on a quick product diversification when demand for said product happens to be up. It's often been the case a company is simply seeking to profit off their reputation which is not only a gamble, they often loose respect with more astute consumers who know they haven't built a better mousetrap, they've simply made one with a slightly better cheese holder.

In this case I have a one word - one acronym answer: QUAD SLI. The fact this PSU was able to power two Leadtek 7950GX2's in SLI (QUAD) while the AM2 3800X2 was overclocked to 300FSB says it all. Corsair isn't a PSU "maker" however I do applaud their effort to seek out a solid design, one which does belie its power rating to your advantage. The extra kick is most likely the result of "Load Sharing" which allows power from one rail to combine with another when taxed beyond its 18A limit. While I do see this as an ad hoc fix, it is a common work-around to design limitations inherent in the ATX/EPS12V Power Supply Guides. Corsair has incorporated an unique, user friendly modular cable system, with a fairly powerful multiple rail power supply. I must admit I was impressed.

PRO
Powerful, SLI no problem
Solid rails, minor fluctuation
User friendly, svelte modular design
Doesn't add much to overall system noise.

CON
Pricey

At the time of writing Corsair's CMPSU-620HX can be found at Newegg for $179

Questions/Comments: forum thread