GT wrote:
> "Mike Walsh" <spam_sucks@bellsouth.net> wrote in message
> news:48AAD88D.AF17147@bellsouth.net...
>> The heat from a power supply will increase as the load increases.
>
> That is what I thought. So a significantly 'over-the-top' PSU for a PC would
> tend to run cooler (due to larger cooling potential) than a
> 'pushed-to-the-limit' smaller PSU, which would have its little fan blasting
> like mad to keep up.
>
>

Take actual_load, and divide by the efficiency.

Say I have a 68% efficiency power supply and I draw 200W,

200W
---- = 294W , and 294W - 200W = 94W of heat
0.68

If I purchase an 85% efficiency power supply and I draw 200W,

200W
---- = 235W , and 235W - 200W = 35W of heat
0.85

In those examples, the actual load power is what determines the
heat. My load in both examples, was 200W at the outputs. If I
draw an actual 650W of power from each supply, then the
68% efficient one is going to get impressively hot. Heat is
proportional to output power. Draw more output power, and
get proportionally more waste heat.

A power supply that doesn't state the efficiency, could be as
low as 68%. The more expensive ones will be proud of their
efficiency, and will state they are "80+" or 85% or whatever.
Some reviews will attempt to measure the efficiency, to see
how close to the stated figure it gets.

In terms of operating temperature, you cannot make any
statement, because the fan can be slowed down to achieve
better noise performance. If I completely stop the fan on
the one dissipating 35W, I might be able to make it as
hot as the 94W one with the hurricane fan on it. But generally
speaking, you know the 85% efficient one is not going to be
allowed to get that hot. They'll leave the fan running to
some extent.

The efficiency varies a bit, with the load range. The
85% unit, might be 85% at a load of 200W to 300W. Perhaps
the efficiency drops to 80% at 100W. But it is likely
ahead of the unit that is 68% at 100W, so you're still
"winning".

The only time the 85% one is not a winner, is if the
design is marginal. For example, some early Seasonics,
had weak 3.3V or 5V rails, with excessive droop. On modern
motherboards, that would tend not to be an issue, as the
3.3V and 5V rails don't draw as much current as they used to.
But if you had a dual Athlon S462 motherboard, with all the
processor power coming from the +5V rail, and you combined
that motherboard with an early generation Seasonic, then
you may regret your purchase.

The best way to spot problems, is read the reviews on Newegg,
even if you aren't going to buy the supply from there.

Paul