Originally posted by: Acanthus
While youre right, that operationally they are different, no one device draws enough power to make it an argument of any kind.
If your PC had a friggin stove element to draw off of one rail then yes it would make a difference.
You mean like a modern high-end video card?
16A+ will power anything on the market, combined or seperate rails.
Won't power Tech Report's
Radeon 2900 XT Crossfire rig under load. The 170 W difference between idle and load is almost all CPU and video cards - 12 V components; there goes 14 A right there. Unless the idle current draw was 2 A (we should be so lucky), that's just not going to work out. It's not even a certainty that 16 A would be enough for that system at idle.
I deliberately picked an extreme example, because it removes the wiggle room of "it'll work, the manufacturer just won't say so" etc.
If anything, having more than one rail makes a system less stable if its all drawing off of one or the other.
If the current draw is within the rail's limits, it's functionally no different from a single-rail supply under the same conditions.
FWIW I don't have any particular love of multi-rail supplies (at least as they are usually implemented), but they are a reality of the market.
But again... just adding them together haphazardly and watching brands will do more than buying PSUs for their "ripple" or overshooting your power requirements by 300w.
Adding ratings haphazardly usually works adequately until you run an SLI or Crossfire system, which may severely break the assumptions you are making. Otherwise it's a passable but poor first-order approximation.
Watching brands is a start, but not the whole story. I certainly would not equate the quality of all of Antec's products, some of which are excellent and some of which are not. Most Thermaltake power supplies are horrible, but several are very good.
Buying decisions should be based on more than ripple characteristics, but sometimes ripple is good to know about. Ripple is what keeps people from readily endorsing products based on Fortron's Epsilon platform.
And if your intention is to judge supplies based on brand, it is useful to have something concrete to base that on; a brand that consistently has supplies that produce in-spec voltages but wildly out of spec ripple is not very appealing. A brand that consistently uses cheap capacitors may have products that sail right through most "reviews" but die 6 months out of the gate. The problem with most power supply reviews is quite simply that they don't look at these things, often because they share your attitude that these are trivial concerns.
Over-shooting requirements (as has become so popular these days) is not an especially good solution, but frankly if you have no real information to go it isn't a terrible solution. Even a very low quality power supply will often function just fine if little is asked of it relative to its ratings. Is it better than making an informed choice? Of course not. But when all else fails, tossing in a factor of safety is better than nothing.
I understand the desire to reduce the complexity of dealing with this stuff, but sometimes there are reasons for that complexity. My hope is that technical advances, and maybe some revised standards, will simply the situation over the next few years. I'm not holding my breath.
) If you are 1/2 load on one #1 X/3 rail you've lost 15% of your total psu amps. If you're then at 1/3 load on #2 X/3 rail you've lost 25% of your total available psu amperage.
Facebook
Twitter