We received a trio of 1000 W power supplies priced between $200 and $300, so we ran them through our usual suite of tests to see if they really live up to their 80 PLUS Gold certifications. Surprisingly, all three hiccuped during efficiency testing.
It's easy enough to assume that 80 PLUS Gold-certified PSUs with power ratings in excess of 1000 W are not built with the broad masses in mind. It takes a serious configuration to require such high power delivery ceilings. Nevertheless, the findings in this roundup make it pretty clear how much importance the manufacturers attach to the quality of their products. Even the slightest deficiency is exposed immediately at such high loads, and the negative effects on energy efficiency are often quite severe. We received three PSUs for this piece, and we put them all through our gauntlet.
Compared to our recent roundup of gaming PSUs where the manufacturers almost buried us in test samples, the range of products is much more manageable in the high-end space. We're looking at two 1000 W PSUs from OCZ and Rosewill, along with a 1250 W PSU from Sparkle. Can the Sparkle PSU exploit its significant power rating advantage in any way? And how will these 80 PLUS Gold perform at the low loads typical of an idle PC?
Also Tested: Standby Power Consumption, EuP Standard
The European Union’s Eco-design Directive 2009/125/EC, also known as the EuP (short for Energy-using Product), contains increased stringency regarding the standby mode power consumption of PSUs from the year 2010. As more and more manufacturers are advertising the EuP certification on their PSUs, we are introducing the appropriate test methods in our reviews. Unlike previous standby measurements made with the 5 V-sb rail active, there are no loads on the rails in the EuP tests. The PSU must manage a standby power consumption of less than 1 W.
- Extreme Power For Extreme Users
- OCZ Z-Series Z1000M
- OCZ Z Series Z1000M: Measurements
- Rosewill Lightning-1000
- Rosewill Lightning-1000: Measurements
- Sparkle GW-EPS1250DA
- Sparkle GW-EPS1250DA: Measurements
- Test Configuration, Holdup-Time, Inrush Current, Peak, And Short Circuit Tests
- Efficiency, Temperature, And Standby Power Consumption
- Power Profile Efficiency Overview
- Conclusion And Comparison Table
You ignore when units fail ATX spec by having too much ripple on the 3.3V rail, you say the OCZ unit is the cheapest at $205, then 2 pages later say the rosewill is cheaper than the OCZ at $220, you ding the rosewill unit for having 47% efficiency at 2.5% load, no one designing a PSU even cares about a load percentage that low, and then you say the Sparkle cannot get up to the theoretical maximum of 120A between its 12V rails, its max is 1250W(104A) not 120A which would be 1440W, you made the noob mistake of adding the rails not looking at the combined power.
Go look at the jonnyguru review of it if you want to see a real review of the sparkle and not this farce.
http://www.jonnyguru.com/modules.php?name=NDReviews&file=print&reid=212
http://www.newegg.com/Product/Product.aspx?Item=N82E16817139014
takes any of the psu's in this review to school.
Seriously where is PC Power & Cooling,Corsair, or Antec!
OCZ and Rosewill thats it? OCZ is alright, but Rosewill is crap.
try theses next time:
http://www.pcpower.com/power-supply/turbo-cool-1200.html (not GOLD I know but still)
http://www.newegg.ca/Product/Product.aspx?Item=N82E16817139014
http://www.ncix.com/products/?sku=56794&vpn=HCP-1200&manufacture=Antec
Running Bionc on CPU and F@H on both GPUs, I draw 441 watts with ALL components under HEAVY load. That's 100% on all CPU threads and GPUs. 1000W is enough for a lot of components.
I also idle at 189 Watts with SLI enabled.
I just wish I could afford the kind of components that would require this kind of PSU.
can someone explain this to me, because i dont get it.
Its got to do with varying efficiency at different voltage levels.
Consider the following:
Power(P) = Voltage(V) X Current (I) X Power factor (Cos phi)
Now for the same amount of power transfer, at lower voltages, the current required is more (See the equation below):
P = V1 X I1 X Cos phi1 = V2 X I2 X Cos phi2
(Substitute for V1 = 230, V2 = 115, neglect the slight difference in Cos phi1 & Cos phi2)
The losses are given by :
H =I^2 X R X t (where R is the resistance of the current carrying conductor, t is time)
Thus losses increase in proportion to the current squared.
So you have higher losses (hence lower efficiency) at lower voltages.
where's the logic in this?
bigger temp difference means the cooling system is more efficient and it takes the heat away from the components. why is air temp so important to you? If you cannot measure temp from inside the psu case (the components) why bother with air temp?
You can't switch between the two. US & Japan & some S. American countries (IIRC) use 115V while the rest of the world uses 230V. PSUs meant for both markets often have a switch and you need to set it to the correct voltage for your country and plug it in. Many modern PSUs will often have a large input voltage range spanning both the voltage levels so you won't find the switch in them (My Corsair TX650 for instance).
Many computer mags plus other sites have recommended this brand for rock solid current/voltage tests including one of the best for low noise and very high current demands with no voltage drops.
They cost a bit more but in this case you get what you pay for. Their caps are HUGE compared to other power supplies. Though now that another company has bought(OCZ I think) them they may loose their industry rating they used to have and are not being as well built as before.
I've seen enough of this, like Fluke who used to make top quality meters now have most of their products made in China and I'd never give them any vote of confidence. I know this because of a friend of mine worked there as product engineer and saw how they were going downhill. When They were merged with the Danahur corp (not sure of their name spelling) their quality has gone into the gutter.
It's how it is in these times of high profits first, quality last.
But so far I've had very good results with my 750W PC Power and cooling supply.
Have seen enough people complaining about video card problems that can be traced back to the power supply they use. I've never had any prob. with my hardware. You need a PS that has very little or no surge drop of the current/voltage and low noise levels. PC Power has been tops in this area and is the only PS I would recommend to my friends.
Many computer mags plus other sites have recommended this brand for rock solid current/voltage tests including one of the best for low noise and very high current demands with no voltage drops.
They cost a bit more but in this case you get what you pay for. Their caps are HUGE compared to other power supplies. Though now that another company has bought(OCZ I think) them they may loose their industry rating they used to have and are not being as well built as before.
I've seen enough of this, like Fluke who used to make top quality meters now have most of their products made in China and I'd never give them any vote of confidence. I know this because of a friend of mine worked there as product engineer and saw how they were going downhill. When They were merged with the Danahur corp (not sure of their name spelling) their quality has gone into the gutter.
It's how it is in these times of high profits first, quality last.
But so far I've had very good results with my 750W PC Power and cooling supply.
Have seen enough people complaining about video card problems that can be traced back to the power supply they use. I've never had any prob. with my hardware. You need a PS that has very little or no surge drop of the current/voltage and low noise levels. PC Power has been tops in this area and is the only PS I would recommend to my friends.
You ignore when units fail ATX spec by having too much ripple on the 3.3V rail, you say the OCZ unit is the cheapest at $205, then 2 pages later say the rosewill is cheaper than the OCZ at $220, you ding the rosewill unit for having 47% efficiency at 2.5% load, no one designing a PSU even cares about a load percentage that low, and then you say the Sparkle cannot get up to the theoretical maximum of 120A between its 12V rails, its max is 1250W(104A) not 120A which would be 1440W, you made the noob mistake of adding the rails not looking at the combined power.
Go look at the jonnyguru review of it if you want to see a real review of the sparkle and not this farce.
http://www.jonnyguru.com/modules.php?name=NDReviews&file=print&reid=212
I would greatly enjoy that