OCZ GameXStream 850W PSU - PAGE 4
J. Micah Grunert - Wednesday, January 3rd, 2007

Related Articles

-- Article Quick Nav -- 1 - Introduction2 - First Impressons3 - Engineering Aspects4 - Testing Procedures5 - In Conclusion

Sound Level Testing

The rated sound output of a power supply is an important factor in power supply choice and purchase. Our computers are loud enough already with the spinning of hard drives and the whirring of fans. Adding yet another audible irritation can be, well, irritating. That's why PSU manufacturers work with due diligence to provide the PC market with ever quieter power supplies.

To determine the real world noise levels of the OCZ GameXStream 850W we perform our own in-house sound testing using an Omega HHSL1 sound meter. With this we measure our ambient reading (the most silent location we can find) and the sound levels of a given power supply at two distances. First at 10cm and then again at 1m away. It is also important to note that we take caution to avoid positioning the sound meter in the path of the supplies output fan. Furthermore, I have personally found that holding the sound meter whilst measuring can create false readings as well.  I always lay the sound meter on a solid surface when taking readings.

So now that our procedures have been set forth, it's time for me to start testing. I happen to have a bathroom at home that will dip down to about 28.0 dBA ambient in the dark of night. If all others are sound asleep in their beds, my dog isn't snoring at the top of the stairs and there's no pounding of rain of the roof, it is ideal for sound level testing of all sorts.

All right now, I'll just jumper this supply to life and see what's to be heard. Unfortunately, it's something bad. The moment the supply was plugged in and switched on, a faint yet noticeable buzzing sound started to come from the inner reaches of the GameXStream 850W supply. This sound was somewhat akin to the quick snaps of electrical arching between two points. Now I do have extensive experience with electronics, both that of the industrial and integrated circuit affairs. This is not electrical arching, but rather the rapid oscillation of one of the SCRs (Silicon Control Rectifiers). An SCR is a component with a power supply, a part mounted to a motherboard of soldered into an electrical device that perform the specific function of augmenting current flow. Typical house hold AC current appears in a wave form, like rolling hills. Computers require a steady diet of DC current. It's a power supplies job to shift this AC current into DC current. But that portion of shifting is accomplished by the electrical transformer inside the power supply. This new DC current still has those rolling peaks and valleys that a computer cannot use. There's more to detail, but simply put; computers require an electrical signal that appears more digital than analog. That's where SCRs come into the fold.

The SCRs take these wave forms and turn them into a square sign wave. Computers need this type of electricity to function properly. SCRs use ferrous rings wound in Copper wire to essentially chop off the tops and bottoms of the analog wave form. Further augmentation by the SCRs will take the sweeping rise of these wave forms and straighten them out. But if a single SCR were to be of substandard production quality, it would create a magnetic field and begin to vibrate at the of the rate of electricity passing through it. That's where this buzzing sound is coming from. The buzz will even shift downwards in tone when the fan is physically stalled out. Another good indication that a semi-faulty SCR is to blame.

Now before anyone completely denounces OCZ power supplies for this one small defect, it really isn't much to worry about. This will in no way have any affect upon the out flow of electricity. Nor will it shorten the life span of the power supply by any significant manner. Switching ATX power supply's are far more likely to expire from excessive heat than a single bad SCR. And the buzzing sound, it is really quite minuscule. Yes, you can hear it, but I was doing my testing in the most quiet environment you will ever encounter in your homes or offices.  Toss this power supply into a case with a few other fans and a couple of hard drives running and you won't even know it's there.

So we'll perform our sound measurements now, but with the knowledge that the buzzing noise I heard may skew our sound figures ever so slightly. But not to the point where it will taint or stain our results.

 

	OCZ GameXStream 850W	Cooler Master RS-850-EMBA 850W	ThermalTake Toughpower 750W cable management	Thermaltake Toughpower 750W	CoolerMaster RS-550-ACLY	OCZ GameXStream 700W PSU	AeroCool Zero dBA 620W
Ambient(dbA)	31.8	28.8	28.2	29.2	30.5	30.5	30.5
1 meter from PSU	34.4	31.1	34.0	33.2	41.5	34.3	31.0
10cm from PSU	38.5	37.2	42.0	38.5	51.4	38.2	32.6

Well, it looks like my ambient reading was a little higher than I had expected to have, but in the greater scheme of things I'd say that the OCZ GameXStream did all right. With a quick adjustment for the different ambient levels the GameXStream comes pretty close to first place. The AeroCool Zero dBA of course our most silent power supply of the lot. Our Cooler Master RS-550-ACLY is a tad loud for my tastes, and the lower output OCZ GameXStream 700W PSU appears to fall right in line with its kin. No wonder since they both use the same fan, same housing and the same basic reference design.

 

Test Setup

Some time ago, ATI had begun a certification program for power supplies intending to run CrossFire systems. Since ATIs CrossFire platform can draw unprecedented and near obscene amounts of current, a 'Certified' PSU needs to be of a certain output rating to run those cards properly and efficiently when they are fully loaded. As fate would have it, there just happens to be a pair of CrossFire kicking around and our video test bench system sitting idly by. Here's a quick detailing of the parts we assembled for testing the OCZ GameXStream 850W PSU

• Intel Core 2 Duo E6600 Conroe @ 3.0GHz (333 MHz * 9) processor
• Asus P5B Deluxe / Wifi-Ap motherboard 
• 2*ATI Radeon X1800XT video cards running in CrossFire
• 2 * 1024 MB OCZ Gold PC2-8800 memory @ DDR2-1000 5-5-5-15/2T
• Noctua NH-U 12 CPU cooler
• Seagate Barracuda 80GB 7200.9 SATA hard drive
• Samsung WriteMaster DVD burner
• Windows XP Professional SP2

This updated system was assembled when we had decided to expand our repertoire of video card benchmarks. This configuration has served us well and is near on par with the average system that a typical home user might have today. But for one added note, the E6600 Conroe processor we use clocks at 2.4GHz (266 MHz*9) for stock speeds. We've pushed it just a little bit further in the hopes of keeping up to date with future processor speed stepping expectations. So that chip should pull a fair bit more juice that usual. Tack on those electron chugging X1800 XT CrossFire cards and we should be able to labour this OCZ supply to its near limits.

As for what application we use to load the OCZ power supply, there are two in particular that can bring even the best of rigs to its knees. The first is 3DMark 2006, specifically that of the HDR (High Dynamic Range) rendering feature test. HDR is one of those great eye popping features of video graphics and is somewhat akin to Anisotropic Filtering (AF) on steroids. The other app we use is CPUBurn-in, times two. This little program throws bits at a processor to bring it to full load. We run two instances to help us fill up both cores.

Once everything is up and running we use our trusted old Mastercraft 052-0060-2 volt meter. Once ready we set out and test our three primary voltage rails, of those being the 3.3V, the 5V and the 12V rail. We test first at idle to see the voltage levels being drawn, then with 3DMark 2006 and CPUBurn-in*2 loading up both CPU cores and both GPU cores. It should also be noted that CPUBurn-in has a check-box to enable/disable error checking, loading the processor even further when the error checking is removed. Here's the results of our testing and some resulting voltages for other notable power supplies.

Voltage testing with 3DMark 2006 & CPU Burn-in times 2 (error checking removed)  

OCZ GameXStream 850W
	3.3 Volt Rail	5 Volt Rail	12 Volt Rail
Idle	3.37 Volts	4.96 Volts	12.53 Volts
Load	3.36 Volts	4.985 Volts	12.475 Volts

As we can see, there are reportable fluctuations between the idle and load phase upon the respective voltage rails. The volts measured did rise and fall between idle and load, but this is normal. Furthermore, the resulting figures from the OCZ 850W GameXStream PSU are more than acceptable. There is that large dip in the 12V rail when the power supply began working extra hard, but that would seem to be normal for pretty much every power supply on the market today. The one obvious exception being that of the two Thermaltake power supplies included in this lot. One impressive standing goes to the CoolerMaster 850W RS850-EMBA power supply. Its voltage readings were so stable that it probably could have handled another pair of CrossFire cards at the same time. 

The OCZ GameXStream passed our CrossFire torture test with flying colours. Performance wise it should handle anything that could conceivably be thrown at it, CrossFire or whatever. That irritating buzzing sound that emanated from the supply was a disappointment, but nothing that would kill the GameXStream 850W anytime soon. All in all, the OCZ 850W is a good supply, though I do wish that I could have loaded it up to the near point of breaking. Oh well, perhaps just a simple conclusion will suffice.

So the OCZ GameXStrem 850W has a nice ring to it. Time for some voltage testing.

next: In Conclusion »

« Engineering Aspects

send article   hardware newsletter   article comments (3)

Submit our article to: Digg this! del.icio.us