OCZ EliteXStream 800W and 1000W
by Christoph Katzer on April 7, 2008 5:00 AM EST- Posted in
- Cases/Cooling/PSUs
Internals, Continued
Going over to the secondary side, we can see the large sister PCB with the security features and fan control. There are two temperature diodes attached to the control unit, which are both attached to the secondary heatsink. The PS232S from Silicon Touch Technologies is also no stranger to us, with many of the companies using this or similar chips in their power supplies.
On the left side of the sister PCB, we see three potentiometers that the manufacturer can use to adjust the voltage output after production. The problem with this is that it's possible for the potentiometers to come loose after a while, and the output voltage would change if that occurs. To prevent this there is a little drop of blue glue you can see on the picture below. It's still a slight risk, though probably 90% of all power supplies in the retail market use this approach. It's an easy way to lower production failure rates, since the manufacturer can easily adjust the output voltages, potentially saving a unit from the scrap heap.
The two temperature diodes attach to the secondary heatsink. The first one is on the left side and the other is on the right side.
The PCB shows several markings, including 12V rails numbered 1 to 5. Indeed this power supply was designed for several separated 12V rails, but many manufacturers are ignoring the specifications related to maximum amperage on a single 12V rail. Instead, they do whatever they think is best to provide a problem-free product.
All secondary capacitors come from Teapo, which we find in many power supplies. We have never had any problems with these caps so far.
Facebook
Twitter
RSS
10 Comments
View All Comments
TechnoButt - Thursday, January 1, 2009 - link
I, too, bought this in May/08. Had to RMA within 2 months. Waited the 2 weeks or so to get it replaced (out another $20 for return shipping, of course). Got system built up with it again sometime in August, and here it's dead again. How can they possibly claim such high MTBF when they obviously didn't start testing these units so long ago (they're too new!).What's the point of a 5 year warranty if you have to return ship every 2-3 months to keep your product going?
Don't get me started on my other issues with OCZ from 2008 (Fata1ity memory kits and aggressive SPD programming that don't boot in dell boxes, Vista upgrade memory kits failing, the embarrassment of building corporate demos on the Core SSD drives only to have them corrupt data in front of a customer within 2 weeks of building, having my Vendetta2 hsf suddenly stop cooling my cpu properly (after about 6 weeks of great cooling.. no adjustments.. apparently a thin wall in the center heatpipe/pinprick leak), etc etc etc.
I'm starting to wonder if ocz even has a quality control department anymore, of if they just build for reviews and 'stand behind it' with a long warranty (way behind it.. I mean).
If you haven't already been suckered into buying this PSU by the great reviews here and elsewhere, you'd be better served by finding an older PCP&C unit from before the buyout days. My biggest fear is that PCP&C will succumb to the poor quality culture of their new masters.
StraightPipe - Saturday, December 20, 2008 - link
I took this recommendation and I've had terrible luck.The first PSU ran great, for 2 months. RMA took 2 weeks, and their Customer service would not even acknowledge my request for an advanced replacement.
Second PSU arrives. I'm told it's new, not a refurb. It runs for 5 days and dies. Same thing every time, red light in the back. The forums have a stickie recommending you unplug everything and short the PSU with a paperclip. That doesnt help either.
Now I'm about to buy another PSU while I wait for the OCZ RMA process...
TWO THUMBS DOWN
nubie - Monday, April 7, 2008 - link
I like it, the power output is re-freaking-diculous, you could run a pair of overclocked Dual-Core SLi systems from either of these!!The thing that most interests me is that huge area at the top of the curve in the efficiency charts, if only they will release a 400-600w model that would put that flat spot right where the majority of systems run, IE 150-300 watts.
Another thing of interest to me that I have been noticing lately in your (excellent) reviews, is that 230v is much more efficient. Since I live in the US, would it make sense to have a 230v outlet for my PC? The trade-off in terms of efficiency looks like it would be worth it.
Another thought is that these extreme power supplies that can theoretically draw more than a common 120v circuit breaker can deliver may be better off on their own 230v line. So if you are building or remodeling your house, or just running another power line to your PC, should you consider 220v? (obviously you should use a brightly colored cord for the 220v, and the proper receptacle for the wall outlet.)
vajm1234 - Monday, April 7, 2008 - link
do we really need that much power i have 1 corsair 620 and its enough i think....nubie - Monday, April 7, 2008 - link
Not me, I am still wondering whether this PCPower Silencer 470 is too much(26A on the 12v) for my single card overclocked 7900/Celeron 430 (650mhz, 3.17Ghz P5N-E SLi), but I realized that I might as well have it for when I choose to run more, or put a fan on my HR-05 chipset cooler and run for 4Ghz (It has booted stable at 3.6+, but then heat caused lockups after I got in the OS).See this testimonial http://www.clubit.com/product_detail.cfm?itemno=A6...">http://www.clubit.com/product_detail.cfm?itemno=A6... :
"I have an overclocked quad core, two overclocked 8800's, 3 HDDs, 8 big fans, and an X-Fi card and it's been totally stable."
The user above me is right though, with a Lian-Li v2000 and 12 HDD, overclocked Tri/Quad SLi setup, and/or a 12v refrigerant (phase change) compressor you would be glad for one of these PSU's.
Sunrise089 - Monday, April 7, 2008 - link
600 watt is great for most of us sure...but a few people do have SLI or Tri-SLI setups (or Crossfire) AND overclocked quad-cores. For those folks...I suppose 650w or so might be cutting it too close?petalsofpain - Monday, April 7, 2008 - link
Even though I am more of a biology person... I have been taught that the reason for the voltage being higher in the case of powerlines into orders of magnitude above that of what he have in household appliances has to do with the general fact that the higher the voltage the lower the conduction losses and also 'the less leakage'!So if it really is that way it stands to reason that would we be able to keep the voltages higher that we would be able to have less in the way of losses. You could try the idea out pretty easily but I think that you know the theory behind this... as I have forgotten XD! Fun stuff though. Have a comparison with some server ones that have an efficiency rating maxim of 93% for comparison or even direct explanations. I don!t remember the name of the model or whatever just know that it was here that I read about this model...
Excellent article as usual as Anandtech stands and remains one of the few and amongst the best hardware and software or even general tech sites period!
Let it be known... I AM AN ANANDTECH FANBOY! :)
... something to have those annoying vidia and amd and intel fanboyz be afraid of in the near future ...
homerdog - Monday, April 7, 2008 - link
The second graph on page 8 shows the 800W PSU to be more efficient with 90VAC than with 120VAC. Are the colors mixed up or is this really the case?Christoph Katzer - Monday, April 7, 2008 - link
Damn, yes right. I will fix it, thanks.piroroadkill - Tuesday, April 8, 2008 - link
Also, there's no colour legend on the PFC graph, for the record