Silverstone Zeus 1200W: Designed for the Power Hungry
by Christoph Katzer on July 7, 2008 2:00 AM EST- Posted in
- Cases/Cooling/PSUs
Conclusion
Talking to Silverstone reps always brings the same message about the company philosophy. They have no need for the mass market; Silverstone tries to make their mark with special products. Given that philosophy, it's no shock that the prices are higher than the competition, but there are clearly users that buy these products since the company continues to exist. Silverstone demonstrates with the Zeus 1200W that they are able to produce products with a few interesting features. The quality is good as always, and we saw a very well built PCB layout that was actually fun to take apart.
The length of 210mm (roughly 8 inches) means the power supply won't fit in many smaller cases, and users will need to pay attention to spacing as many cases have either hard drives or optical drives that can get in the way. Cases with a separate thermal area for the PSU (i.e. the popular Antec P182, some Lian Li offerings, and similar units) will have trouble mounting this unit. It's important to note that besides the casing length, the cable connectors at the front add another inch to the total required length.
 |
From the connector side we have few complaints. There are connectors for all kinds of graphics card arrays and in our opinion plenty of connectors for a large number of peripheral components. Six Molex and SATA connectors are available, so you will have to determine if that's enough for users with large disk arrays like on a server farms; for gamers, though, it should definitely be enough. The length of the cables is also good since we have a distance of 55cm to the power supply from ATX and PEG connectors. A nice feature to fight cable clutter is the two PEG connectors on one cable harness. This allows users with three graphics cards to run three harnesses instead of six through the case. One of the highest power consuming graphics cards, the NVIDIA 8800 Ultra, needs up to 160W, so it will pull a around 3A on each of the pins within the connector (with some power coming from the PEG slot on the motherboard).
Performance-wise, we saw some flaws today that are severe from the view of an enthusiast but not from someone who builds servers or workstations. The DC output is well regulated, with up to 2.5% around the nominal number of each rail. To archive this with a load up to 1320W is something many competitors can only dream of accomplishing. The output quality comes from the well designed PCB and the choice of components. Silverstone unfortunately had to make choices that hurt efficiency somewhat, so the Zeus only manages 84% with 230VAC. Up to 84% efficiency is nothing to be ashamed of, but in comparison to other high output PSUs we have seen better. For users wanting to save a few dollars more on electricity, we would recommend other power supplies. The acoustic noise is also poor in comparison to some other solutions. It is very good for a power supply to stay cool, but it should have been possible to reduce noise somewhat without temperatures getting too high. The fan rotates faster with every small increase in load, and it reaches its top speed of 3200RPM at only a 60% load. The noise certainly won't be a problem in a datacenter, but it does limit the appeal for the product.
Our only serious concerns relate to the acoustics and the length of this power supply; in terms of providing a stable output with huge amounts of power, the Zeus 1200W is top notch. As for the price, the Zeus 1200W is available in for around $350 to $400 in the US. We haven't actually seen availability in the European market yet, and that may not occur unless Silverstone is willing to bring this unit to the masses. The price is quite high, but then we really didn't expect a unit targeting the elite users to rate anywhere near affordable. If you have a need for 1200W of power, the Zeus may fit the bill; the rest of us with mere mortal budgets and PCs will have to be content with something else.
Facebook
Twitter
RSS
20 Comments
View All Comments
thebackwash - Tuesday, July 8, 2008 - link
At what point does the house's wiring really begin to be a factor in sustaining the power draw? I know you're measuring watts vs. volts, but if someone could give a practical range and a better explanation as to the practical problems with running appliances with a high current draw. Volts being constant generally in a house, is it amperage that cannot be increased past a certain point?What I'm trying to ask, to anyone who can elucidate, is at what point does the house become the limiting factor as compared with the computer PSU?
I know that if I run the air conditioner in a room and someone sends a job to the (laser) printer plugged in in the same room, it trips the circuit breaker downstairs, and everybody gets an 'oh gosh, that was silly' out of it. When does one have to get new wiring run in the house to run their über gaming rig/cluster running department of defense simulations?
Carnildo - Tuesday, July 8, 2008 - link
> At what point does the house's wiring really begin to be a factor in sustaining the power draw?Right about here. 1200 watts at 75% efficiency means that, at full load, this PSU is drawing 1600 watts from the outlet. Most 120v house wiring is limited to 1800 watts (15 amps) per circuit, so if this thing is sharing an outlet with almost anything (say, a laser printer), you'll be blowing fuses on a regular basis.
If your power is only 110v (common enough), then 1200 watts at ~73% efficiency is 1650 watts at the outlet, exactly the limit of what a 15-amp circuit can provide.
thebackwash - Tuesday, July 8, 2008 - link
"Is it amperage…"I should say, "Is it amperage and the correlated wattage that can't be increased beyond a certain rating?" What's generally the bottleneck or are the two tightly linked phenomena when it comes to encountering real world engineering limitations?
JarredWalton - Tuesday, July 8, 2008 - link
If you're running 115VAC (i.e. in the US), then you need to look at the circuit that's tripping. It's probably a 15A circuit, which means that you can only run around 1700W worth of equipment on that circuit before you have the problems you describe. (115V * 15A = 1725W) The question is then how much power the various devices use.I wouldn't be at all surprised if your AC unit can pull upwards of 750W... and if it's a powerful model it could easily reach the 1250W and higher range. (Yup, AC is expensive!) A laser printer might use anywhere from 100W to 300W I suppose. I'd suggest getting something like a cheap Kill-A-Watt device and plugging the various power users into it.
Also, don't forget that lights use power as well. That 60W light bulb uses 60W, so if you have a light fixture with three bulbs, there's another 180W (or 225W if you use 75W bulbs). I highly recommend the florescent bulbs as a power efficient alternative.
gameman733 - Tuesday, July 8, 2008 - link
I think theres an error in this graph. http://images.anandtech.com/reviews/psu/2008/silve...">http://images.anandtech.com/reviews/psu/2008/silve... (DC output at 12V, look at the left side, 12.12, 12.00, 11.88, 11.94, 11.40, out of order)JonnyDough - Monday, July 7, 2008 - link
The real problem with PSUs like this is that sometimes people that are well off and on their first build who want "the best" run out and buy something like this and absolutely do not need it. It just ends up wasting electricity, which we all know is largely derived from strip-mining/coal burning which is horrible for the ozone and natural habitats.serchaing - Monday, July 7, 2008 - link
This is actually a myth, one that I also thought to be true until recently. For example, a PC that requires 340W to operate will use 340W whether the power supply is a 450W or 600W. PC Power and Cooling's web site dispels this several other PSU myths here:http://www.pcpower.com/technology/myths/#m1">http://www.pcpower.com/technology/myths/#m1
C'DaleRider - Tuesday, July 8, 2008 - link
Not the PCP&C myths again. While one or two are actually correct, the modular cable "myth" they dispel has been proven, time and again by independent testing, to be just marketing fluff by PCP&C. Add to that the "single rail is better" myth PCP&C pushes....only taken up, by the way after PCP&C absolutely failed at their design of the multi-railed Turbo Cool 1000W unit (it was horribly under powered on the rails supplied and caused problems....and their solution, instead of fixing the rails and supplying proper voltage/amperage per rail was to dump it for an easier to design single rail.)But, outside of efficiency, you are correct in that a power supply will only draw what is needed from the wall to run whatever is connected to it....no more, no less. So, a 1kW ps will only draw XA or X volts from the wall to supply what's required from it, be it 200W or 900W. It's no more expensive to run a 1200W unit, again leaving efficiency out of the equation, than a 500W unit.
And if you really look at power supplies and their construction, you'd notice that the high power units tend to be built better with better quality internals than lower wattage units.
JarredWalton - Tuesday, July 8, 2008 - link
While that is technically true, efficiency comes into play. If a PSU reaches maximum efficiency with a load of 30-70% of the rated output, then a system that requires 350W should have a 500W PSU minimum, and for optimal efficiency you almost certainly wouldn't want anything larger than 1150W (*cough*).Personally, I try to shoot for around 30 to 50% load, but even my most powerful system only draws a rather piddly 400W at peak. With a roughly 80% efficient power supply, that means the system is only using in the vicinity of 320W. Idle power draw drops to under 200W (160W or less power used by the system). This is with a quad-core Q6600 G0 stepping running at 3.40GHz, 2x2GB DDR2-800 RAM, two HDDs, and dual HD 3870 cards. It's been running quite happily with a 650W power supply for over six months.
mattclary - Monday, July 7, 2008 - link
Can anyone explain to me, or point me to the info on how it is a power supply that will be plugged into a 20 amp circuit can provide more than 20 amps?