A Kilowatt of Power

An anonymous reader writes "There is finally a review available of a kilowatt power supply. The PC Power and Cooling 1KW produces 1000W of power output with 1100W peak. The review points out how great this product did in the testing but was not afraid to admit how much of an overkill it is for the enthusiast market. From the article, 'In the current computing world, where more always equals "better than" the 1KW is king.'"

Comments

Re:whooboy

[headwes]

Well, not quite... That 1000 watts is only what the suppy is capable of. The actual draw will be the combined needs of the computer's individual components, plus any inefficiencies inherent to the power supply. Likley this will be well under 1000 watts. Even so, its silly.

Read the article

[User 956]

No way. If you check the specs, that thing runs continuously at 91% peak power. So, as I said, running a few apps, you're looking at a hundred bucks a month extra in your power bill (at 12.5cents per kilowatt/h).

Re:Read the article

[MechaStreisand]

No, sir, the power supply is a switcher, not a linear power supply. All computer power supplies are. They do not operate by continuously wasting power - instead they transform the line power into the desired voltage and current at around 88% efficiency (depending on the model) all the way from a minimal load to a full load. In fact, for a normal load, this power supply won't draw any more wall current than a 300 watt supply - but it will be able to draw a lot more if it needs to, without failing, if its claims are true.

Re:Read the article

[LoRdTAW]

I want everyone to realise what efficency means. Simply put it means that if a powersupply is 90% efficent and you are drawing 90 watts on its output, you would be drawing 100W on the input. This is because of the 10% loss during the conversion, hence 90% efficiency. I guess you dont know the difference between a switcher and a linear. A power supply does indeed waste power but thats a given. Nothing is 100% efficent. There is an idle current so to speak but usually very very small. A power supply only draws what the load demands plus its own internal loss.

The switcher immediatly converts the ac line current to dc and takes the ripple out using 2 or 3 big fat capacitors. Thats why it sparks when you plug a cold switcher in, the caps charge up. The dc then feeds a transistor which is controlled by an oscillator and produces a 20+ KHz square wave. This feeds some high frequency transformers that kick the power down to the necessary voltages like 3.3, 5, and 12 volts. Each voltage gets its own transformer, more caps to filter out the high frequency and then to regulators for each voltage. Thats the benefit of a switcher, light weight and compact for the power it converts(although more complex and only a 70-80% efficiency). On a side note: If you ever wonderd why aircraft use 400Hz power here is your answer, smaller and lighter power systems.

A linear would take the 120/240v line power at 60/50HZ and put it strait through a big fat transformer made for 50/60 hz and lower it to 3.3, 5 and 12 volts then each to a set of smoothing caps and regulators. A 1 kw linear would weigh oh id guess 20-30 pounds! The higher the frequency, the less turns of wire and smaller core a transformer requires. Plus it would be more efficent then a switcher.

Each of those power supplies have there respective efficiencies but in no way does the linear waste it as you imply. Please if you dont know what you are talking about then dont post misinformation.

Re:Pfffft

[dancallaghan]

But surely all (or nearly all?) 4200 watts are eventually converted to waste heat. So you could have a heater *and* a sh#tload of PoE gear.

Re:Hang on a minute...

[Lisandro]

The power supply does not deliver 1000w all the time. 1000w with 1100w peak means that the PSU is rated to deliver up to 1Kw constantly and up to 1,1Kw for brief periods; but the PSU will deliver only the power that it's requested from them.

    Switching PSUs waste some power, of course, but are among the most efficient types of electrical power supplies available - that's what make them so well suited for computers.

Re:Insanity

[arivanov]

Not with Pentium 4. There is no standard OEM MB out there to carry the number of Pentium 4s to get that far. And the special ones are actually designed together with a power supply and a case.

In fact there is just a single "standard" MB I can think of to use with this beast.

It is the Assus 8 Opteron MB which has 4 CPUs on board and 4 CPUs on a daughther card. If we assume normal Opteron and throw in some video, cooling and disks in you end up having a 700W+ maximum power consumption.

If someone can think of something else to generate that much power without coming with a dedicated power supply - post it. I can't.

Re:how about a good power supply instead?

[NerveGas]

Well... there's a lot of room to fudge, and like so man other people have said, it's about amperage. Just about a week ago, I showed a coworker a couple of power supplies, one rated at 400 watts, the other at 425 watts. As I recall, the 400-watt power supply (which was quite expensive) could deliver upwards of 30 amps on the +12V line, where the 425-watt unit maxed out at 18.

Even assuming that both of those supplies reached their published specs, if he's trying to run a high-draw, overclocked CPU and a high-draw video card (both of which draw their power from the +12V line, and even without overclocking can pull more than 18 amps), then it's fairly obvious that just reaching the rated specs - even at 425 watts - isn't *always* enough.

steve

Re:Pfffft

[Technician]

Not trying to be an ass but it hurts the eyes when so many posts including opening article have kW wrong. It is kW not KW. It has a small k because kilo isn't taken from somebody's name.


I found a list of metric prefixes.

I have no idea why Kilo is not uppercase as are most multiplyers of greater than unity. In common pratice is is common for greater then unity multiplyers to be uppercase to avoid confusion with less then unity multiplyers. That is why most street signs read KM to the next exit and transistors are measured in nm and leakage current is measured in uA.

Prefix Symbol Multiplier Exp
yotta- Y 1 000 000 000 000 000 000 000 000 10+24
zetta- Z 1 000 000 000 000 000 000 000 10+21
exa- E 1 000 000 000 000 000 000 10+18
peta- P 1 000 000 000 000 000 10+15
tera- T 1 000 000 000 000 10+12
giga- G 1 000 000 000 10+9
mega- M 1 000 000 10+6
kilo- k 1 000 10+3
hecto- h 100 10+2
deca- da 10 10+1
deci- d 0.1 10-1
centi- c 0.01 10-2
milli- m 0.001 10-3
micro- 0.000 001 10-6
nano- n 0.000 000 001 10-9
pico- p 0.000 000 000 001 10-12
femto- f 0.000 000 000 000 001 10-15
atto- a 0.000 000 000 000 000 001 10-18
zepto- z 0.000 000 000 000 000 000 001 10-21
yocto- y 0.000 000 000 000 000 000 000 001 10-24

Inefficient, compared to what?

[Firethorn]

The question is, compared to what? 75% efficient is great if the alternative is only 50%, which is what linear power supplies average. Linear supplies are simpler, but also bigger and heavier. The sheer mass of metal required can quickly make the simplicity moot, as the cost of the metal drives the price high.

Driving the frequency even higher can increase efficiency, but I believe it increases engineering complexities and strain on the components.

Re:Pfffft

[Anonymous Coward]

I never noticed that most units larger than unity are capitalized. Nevertheless, I imagine the lowercase k is because another SI unit uses uppercase K. (Kelvin)

1kW

[gauntlet420]

I seem to remember the same arguments from a few months back:

http://hardware.slashdot.org/article.pl?sid=05/08/ 22/2157244&tid=232&tid=126

And, since IAAPSD (I am a power supply designer), the same observations are applicable:

• - a 1 kW PSU and a 400W PSU will deliver equal amounts of power into a conventional system that only needs 250W - specifically, 250W
• - most commercial-grade PSUs are unlikely to survive long-term at their maximum rated power - that is to say, that 300W power supply that you bought for $10 from your local PC junk discount store is not the best thing to power your dual-core uber-gaming rig
• - most commercial-grade PSUs do have an efficiency sweet-spot because they do not use things like input power factor correction and are not engineered to exceed their requirements (i.e. ATX specifications) by even a fraction (to keep costs to a minimum)
• - a high-power PSU running at light load can be expected to last much longer than a lower-rated PSU at the same load as the component stress levels are lower - that is to say, the components in the 1 kW PSU are stronger than those in a 400W PSU, and industry standards dictate compoment stress levels to not exceed 50-70% of maximum (depending on the type of part) to ensure long MTBF (mean time between failures)
• - all power supplies fail eventually - if well-designed, the only things that will wear out are the electrolytic capacitors, which will eventually dry out (over a period of years) and cause the PSU to degrade and eventually quit

Re: Insanity

[Anonymous Coward]

I am constantly amazed that some people think that a couple of kilowatts is a lot of power. It is not. If it was a lot of power, then the breakers contacts for your in-house S/B wouldn't be about the size of a pencil eraser.

Let's put it in perspective (all units KWe unless noted):

A refrigerator when the compressor is running uses about 1.5 KW.
A microwave oven varies generally from 500W to 1.5 KW.
Your clothes dryer and oven may use 1.5 KW.
A hairdryer uses about 1.5 KW.
Your car will use about 150 KW thermal at full power (varies depending upon the power of the individual car).
I have run various pumps from fractions of a KW to 500 KW each.
A small silicon refiner outside my hometown uses 60 MW.
The residential areas of that same town use 30 MW (about 1 KW/person).
A locomotive uses about 2 MW.
A powerplant that puts out less than 100 MWe is considered 'tiny.'
Large nuclear plants exceed 3000 MW thermal per reactor (yielding about 1000 MWe).

A 1 KW power supply at 120V input will draw less than 10A from your wall outlet. I have drills that use more current.

Re:Who needs that?

[brontus3927]

Something tells me that panel isn't configured correctly. Check out this website: http://www.jscustompcs.com/power_supply/Power_Supp ly_Calculator.php?cmd=INTEL

Based on the limited information in your post, your computer would peak ~300W

Re:Who needs that?

[Kitsune78]

Umm.. Same PSU here, P4 505 (2.66Ghz/566Mhz FSB), 1 SATA Drive, a 9600, 2 120mm fans, DVD-R, and a water pump Is between 90-120 Watts depending on load, confirmed with my multimeter. 120-160 when overclocking. "Spikes" up near 200 or so when the drives spin up. Assuming I should shoot for 50% of rated capacity for reliability, that means even my crappy set up needs a "400W" PSU. Your processor alone should pull more than 50W..

Re:Hang on a minute...

[DoctorStarks]

Absolutely correct, and that's the important point.

This supply is not burning 1 kW all the time it sits there. It is providing exactly the same amount of power as your current power supply does. The only thing that it offers you is more headroom if you suddenly need to use your USB arc-welder.

And with efficiencies of close to 90%, these types of switching power supplies don't heat your room up when they run. Your processor and video card do, though.

Re:Pfffft

[DigitalRaptor]

Yeah, except that my Dell laptop pulls 80 watts when idle, and over 110 when doing something like unzipping a file. I tested it with my Kill-A-Watt P3 just a few weeks ago.

$16 a month to run your laptop is pretty steep if you ask me. My next one will definitely be a Centrino or other low power chipset. (I had already decided that, for a variety of reasons).

I know that no one is getting charged 25 center per KW/H, yet, but those same numbers would make a 75 watt lightbulb that is always on cost $15 a month to run.

Time for the whole country to switch to flourescent whereever possible. I'm not just talking about compact flourescents either. I just added a full spectrum flourescent 4' dual bulb fixture to my storage room and love it. It uses less juice than a single 75 watt bulb, puts out way, way more light, and it's perfectly white and bright light. And the bulbs last for 9 years.

I'm a convert...