You test a power supply by:
1) Removing it from the case.
2) Unplug the cables (making note of where everything goes for later).
The purpose of unplugging, is on the assumption the power supply is
nuts, and bad voltages are coming out. The less chance of burning something,
the better. Once you trust the supply again, you can put it all
back together.
3) Plug in the supply to the AC wall socket.
4) Turn it on at the back.
5) Now, with a multimeter, check that the +5VSB rail powers up.
The power supply has two sections. The +5VSB section, comes on as
soon as the switch at the back is ON. On your Asus motherboard, the
green LED on the motherboard, should light up if there is +5VSB
present. So if you were seeing a solid green LED, then that
would be proof you have voltage there. Or, use a multimeter, and
probe the appropriate pin on the main connector.
6) Next, connect PS_ON# on the main connector, to the adjacent COM
pin. Grounding PS_ON#, indicates to the supply it should come on.
That is the "soft power" feature. A typical method of doing that, is
using a paper clip, partially straightened out, as a jumper, to jump
the two appropriate pins on the main connector. Now, the -12V, +3.3V,
+5V, +12V1, +12V2 will come on. In fact, in step 6, all the rails
should be active, including the +5VSB which was ON in step 5.
(See Figure 7, appropriate for a 20 pin main connector)
(A 24 pin is similar, in terms of wire colors, short green to a black COM wire.
And then the fan should spin.)
http://www.directron.com/2powersupplies.html
You're probably failing to do step 6, and without step six, the
fan won't spin on the supply, and the main rails won't come up.
Strictly speaking, you should have some load on the supply. That is more
important on some supplies, that on others. Some of the older supplies
have a rather high "minimum load" spec. If you see such a thing printed
on the supply label (both "mins" and "maxes"), you might try to put
a dummy load on the supply. I bought some power resistors at the only
good electronics store in town, and that is what I use for loading. Without
the load, the supply could go out of regulation (when you use your
multimeter, the bad voltage readings would then not be conclusive
about the health of the supply). If it has OVP, the excess voltage
could even cause it to shut off. I've never had it happen to me, but
there are some pretty goofy supplies out there.
I have different resistor values for each rail. For example, I use a 50 ohm
power resistor for the -12V rail, which draws about 0.25 amps. 0.25 x 12 = 3 watts.
The resistors get warm, while I'm using them, and I keep a fan blowing on
them when testing. But to get those resistors, you won't find a good selection
at Radio Shack.
If the supply doesn't have a minimum load spec, you may get away with
doing step six, and noting the fan is spinning. That tells you the
supply can "switch on", but it doesn't give the supply a passing grade.
To give an example, the very first ATX supply I bought, I've still got.
It has a defect. Now, if I do step six, with no load, the +12V rail reads
12V. Great, you'd say, it's working. But the really funny part is, the
supply cannot take *any* load at all. If I connect a 12V 0.1 amp cooling fan
to the 12V rail, with a Molex connector, the voltage drops to 7 volts instead
of the 12V. This is well outside the allowed 5% tolerance. The supply is
a dud. So a load test, even one which doesn't apply the maximum allowed
load, can tell you something about the supply. And that's why I own a
set of resistors for that kind of testing. I test every new power supply
I buy, before it gets plugged to a motherboard. My old supply "looks nice",
but fails to perform, as soon as there is any load on the 12V rail. Even
the capacitors inside it, look nice. Very strange.
Paul
