What wattage Power Supply?

[lehtv]

Just because this is an enthusiast forum doesn't change the general facts, and doesn't make what I said wrong. The vast majority of power supplies in the world are not that great. Lots of people reading this forum have bought pre-built computers, with no idea of what's in the power supply other than what the general specs said, if they even read that.

Notice how you added in a big IF there that I didn't say - IF you're buying a quality unit, and part of the definition of "quality unit" is that is splits up the output among the voltage rails intelligently.

1. The very fact that the rated wattage of low quality units does not represent their actual capability renders your argument irrelevant to low quality units. You can't meaningfully compare the difference in labeled wattage and +12v amperage of two low quality units when the labeled wattage is not to be trusted in the first place.

2. I interpret posts in the context of the thread they're posted in. I have no reason to think the OP is interested in buying a low quality unit.

Look again at a Huntkey PS. The previous comparison was a little confusing because they change the style of some models, but:
160: 3.3V - 7A, 5V - 8A, 12V - 7A
250: 3.3V - 20A, 5V - 21A, 12V - 17A

Are you actually looking at a 160w unit and a 250w unit here? What are the model numbers? I don't even. This is silly.

Let's just think about those specs for a minute. You've got a power supply supposedly giving 250 more watts than the other, and yet somehow it's providing 288 more watts just on the 12V line alone. You realize that doesn't add up, right? It's very difficult to compare apples to apples in these - I admit with my examples as well.

Good of you to admit that. You can equally well look at your Ultra LSP650 vs LSP450 comparison and say "only 168w more on the +12v - it doesn't add up". And that's exactly right.

 

[jeffnc]

This is a silly argument. I havent said anything controversial.

"The thing to keep in mind" is a figure of speech. In the context of this thread, it means when you calculate the amount of power you need, look at the power output of the voltage rail you'll be using. Big deal. Its a spec listed for all power supplies, and its at least as reliable as the total power rating.

If you want to tell the OP your opinion that it doesnt matter much in quality power supplies, then fine, say so.

But if you say my post is factually incorrect, then obviously I'm going to argue back. If its not important, I gotta wonder why youre spending so much time debating it.

 

[jeffnc]

Good of you to admit that. You can equally well look at your Ultra LSP650 vs LSP450 comparison and say "only 168w more on the +12v - it doesn't add up". And that's exactly right.

I compared apples to apples in context. I used the only specs the manufacturer made available, but I never misinterpreted them to come to a ridiculous conclusion like "the power increase on the 12v line alone is more than the total power increase of the whole PS".

 

[lehtv]

This is a silly argument. I havent said anything controversial.

"The thing to keep in mind" is a figure of speech. In the context of this thread, it means when you calculate the amount of power you need, look at the power output of the voltage rail you'll be using. Big deal. Its a spec listed for all power supplies, and its at least as reliable as the total power rating.

If you want to tell the OP your opinion that it doesnt matter much in quality power supplies, then fine, say so.

But if you say my post is factually incorrect, then obviously I'm going to argue back. If its not important, I gotta wonder why youre spending so much time debating it.

I'm concerned with what's useful and relevant. The distribution of power among rails is a non-issue. You don't have to look at the the output of the rail you'll be using. As long as you choose a quality unit (which is what everyone here recommends people to do), you can approximate available +12v power from the rated wattage, it's as simple as that. The end user doesn't need to understand rail distribution, he only needs to make sure the unit is a good quality one. Your posts only make sense in the context of comparing low quality to high quality units (which is not what this thread is about), not in the context of choosing a high quality unit correct for the user's needs (which this thread is about).

I compared apples to apples in context. I used the only specs the manufacturer made available, but I never misinterpreted them to come to a ridiculous conclusion like "the power increase on the 12v line alone is more than the total power increase of the whole PS".

You said: "a 750W power supply will never have 250 more watts available on the 12V line than a 500W power supply. The only question is how much less." I provided counterexamples to that claim, nothing was misinterpreted.

 

[TemjinGold]

jeffnc: I feel like there's 2 things you may not be in the know on that you should know. I apologize if you are already familiar with these concepts because I'm not trying to patronize you but your posts genuinely lead me to believe you are not aware of them.

1) You had a thing about "the numbers not adding up." Here's why: Trash PSUs provide what's know as a "Peak" rating (i.e. assuming they can even provide that much power, which most can't, that is supposed to be the MAXIMUM that PSU can provide for however brief an amount of time.) Quality PSUs on the other hand, provide what is known as a "Continuous" rating (i.e. they are rated to be able to safely deliver that level of power all day all the time for the warranty period). So how do quality manufacturers ensure their units can sustain such a performance level? Yep, they rate their units X percent lower than the maximum capable. What happens when the wattage goes up? You guessed it, the gap gets wider because the unit was rated conservatively percentage-wise. THAT's why the gaps in the examples provided by lehtv are bigger than their ratings difference would imply.

2) You CANNOT simply add up all the amps across all rails to arrive at the amount of power a PSU can provide. No PSU out there, quality or otherwise, can deliver the sum of the listed amps across all rails. That's not how a PSU works. What's more, if say a multi-rail PSU says "12V1 30A, 12V2 30A, 12V3 30A," that does NOT mean you can get 90A off the 12V.

 

[jeffnc]

You said: "a 750W power supply will never have 250 more watts available on the 12V line than a 500W power supply. The only question is how much less." I provided counterexamples to that claim, nothing was misinterpreted.

In other words, you're concerned with what's factually correct, not what's useful or relevant? Your counterexamples were both incorrect and irrelevant, and if anything just showed an inability to interpret the specs. You're comparing the overall nominal rating of the power supply, rounded off by the manufacturer as they see fit, with an allegedly measured maximum power on a single rail, not 2 things that should be compared.

If a power supply increases overall power with all else being equal (for example within a manufacturer series in most cases), then to have a bigger power increase on the 12V line than the overall supply power increase, it would have to significantly decrease the power available on the 3.3 and 5V lines on the more powerful supply. This never happens, and it would be completely ridiculous to do so. Even the worst power supply designs don't do that. See if you can find one counterexample where that happens.

 

[jeffnc]

1) You had a thing about "the numbers not adding up." Here's why: Trash PSUs provide what's know as a "Peak" rating (i.e. assuming they can even provide that much power, which most can't, that is supposed to be the MAXIMUM that PSU can provide for however brief an amount of time.) Quality PSUs on the other hand, provide what is known as a "Continuous" rating

True, but that's not what I meant by the numbers not adding up. It does make comparisons difficult though.

2) You CANNOT simply add up all the amps across all rails to arrive at the amount of power a PSU can provide. No PSU out there, quality or otherwise, can deliver the sum of the listed amps across all rails. That's not how a PSU works. What's more, if say a multi-rail PSU says "12V1 30A, 12V2 30A, 12V3 30A," that does NOT mean you can get 90A off the 12V.

True, and that's the reason I said the numbers don't add up.

The difference is, I realize the numbers are off, but I interpreted them rationally given the context of my statement. Someone else came along and twisted the results to "prove" something what was actually completely incorrect and absurd.

The problem is, without access to rigorous testing and the resultant data, the manufacturer isn't giving us the real specs we need. We have a starting point of a "750W power supply", whatever that means exactly, and then we're giving maximum current ratings for each of the voltage rails. The overall rating of the power supply is more or less up to the category the manufacturer wants to put it in, not what it truly measures at. If the manufacturer wants a power supply to fit in a 750W category, then they call it model number xx750, whether it be 747W peak, or 794W continuous, or whatever.

The bottom line is what I originally said, and it's not really controversial at all. Given 2 PSs with all else being equal, if PS B is actually 200 watts more powerful than PS A, then something less than a 200 watt increase is going to be available to power your 12V devices with PS B. How much less depends on the design, and thus usually the quality, of the PS. Usually the cheaper the power supply, the smaller the power increase is going to be. But it is physically impossible for it ever to be more, without doing something retarded that to my knowledge no PS manufacturer ever does (i.e. significantly decrease the power available to the lower voltage lines in a model higher up in the series).

So people doing a calculation on how much power their computer needs should take that into account. This is basically common sense and nothing controversial at all. This is what all standard calculators do, and I quote "When selecting a proper power supply unit, pay attention to the +12V rail power ratings. Modern computers use the 12V rails to generate most of the voltages in the system."

http://powersupplycalculator.net/
http://www.extreme.outervision.com/psucalculatorlite.jsp
http://us.msi.com/power-supply-calculator/

 

[TemjinGold]

Buddy, I give up. There's really no helping you at this point. Let's make it simple: The OP came here with a system specced out and asked for advice. We gave it to him simply. There is zero need for any of the crap you spewed. I'm not going to waste my time on this thread because you clearly don't get it.

 

[jeffnc]

Buddy, I give up. There's really no helping you at this point.

Helping me? lol

It's really simple. If your system needs more 12V power, then the 12V output is the spec you should be looking at to provide it. If that simple advice is too far over the OP's head, he can ignore it. But it's something that is obviously true, and anyone who spends 10 minutes studying this stuff should know. It's how power requirement calculations are done when putting systems together.

And if you're trying to imply that the above is wrong just to try to win some argument you dreamed up with me, then shame on you.