Testing
Since
I’ve never seen any power supply usage tests before, TargetPC is going to be the
guinea pig. While the proceeding certainly isn’t the only method, at least standardization
will help to eliminate various random factors. I have decided on two variables
to analyze. The first will be static or idling power consumption and the second
will be full tilt boogie or maximum power consumption. Idling power usage is easy
to define, just boot your machine and let it sit there. Peak consumption is harder
to define, so I’ve chosen the program WinTune98
to be the benchmark. After using many other programs, like 3Dmark2000 or any number
of 3D TimeDemos, I’ve found that WinTune98 has similar effects on current drain.
Before
you can blurt out, "my systems uses ___ watts," one formula must be
covered. The power formula is PIE. Mmmmmm, good…well, not actually the
eating kind, but the one that means power equals current times voltage.
Yeah, Bill, that spells PCV, not PIE. Current is also denoted by
I and PIE is easier to remember than PCV, which can be confused
for positive crankcase ventilation. Measure your line voltage
before hooking up your cheater cord. It should be in the range of 110 to 120 volts
AC if you live in the US or Canada. Other countries will have different readings,
such as 220-240 or 100 volts. Make a note of this number, as PIE requires it.
Observe
and note the idle current reading. Then run WinTune98 and note the higher peak
reading(s). It helps to have a "peak hold" feature to grab those pesky
10 mS surges that our eyes always seem to miss. Take both numbers and multiply
them by your line voltage. The result is in watts. For example, if the line voltage
is 120 and the idle current is 0.5 amps, 120x0.5=60 watts. This figure represents
the input power to your supply. It does not mean that your box is using 60 watts.
We have to introduce one more conversion to obtain a more accurate figure.
Efficiency
Power
supplies in general are very inefficient pieces indeed. Did you ever notice that,
at least for computer rated models, they have 5-7 amp fuses? For the 120VAC models,
this means that the total input power can be as high as 600-840 watts, even though
the output rating is only in the 250-300 watt range. What gives? Have you been
gypped? Computer switching supplies have a typical efficiency of around 65%. This
means that for every 100 watts of input power (i.e. 120 volts x 0.833 amps), about
65 watts can be used at the output. Nearly 35% of the input power is lost as heat
and conversion errors. To find your actual power usage, you must reduce your power
figures by 35% or simply multiply them by 0.65.
