No power?

[Ste...@pctamers.eu]

I have an old system which I use now and again for music. It has stopped booting. The Antec psu has a power switch on it (in addition to the power and reset switches on the front). When I switch it on a light comes on on the motherboard - the light is labelled "SB Power) (it may be S8, it's v.small text). No other noise, lights or action: the fan within the psu unit doesn't turn at all. If I switch off the psu with the switch on the back of the psu there is a faint whirring noise as if the machine is powering down.

If I use the power switch on the front nothing happens at all.

Any help/advice greatly appreciated.

Steve

[Paul]

So we know from your description, it's a desktop
computer, and it has an Asus motherboard in it.
Asus is pretty consistent about placing a green
monitoring LED on the +5VSB supply.

Your Antec is making +5VSB, but the step that is
failing, is PS_ON# from the motherboard, which
travels from the motherboard to the PSU, is not
making good logic levels. Either the driving end
(motherboard) is defective, or the Antec component
receiving the signal is defective.

A typical standalone test, is to remove the Antec
and drive the PS_ON# with a ground strap. Which
would cause the Antec to start and the cooling
fan on the Antec to spin. It's normally advised
to have a test electrical load connected to the
Antec, and with a test load connected, you can
use a multimeter to verify the voltages.

Doing such a test, verifies the Antec recognizes
a zero volt (logic "0") on the PS_ON# signal wire.
The supply remains powered, for as long as there
is a logic 0 on the PS_ON# wire. The "#" symbol in
the signal name, means "active low" and the signal
is active, when a logic "0" is present.

On the motherboard side, the drive is open collector
logic. A resistor (like maybe a 2.2K or a 10K) pulls
PS_ON# to approximately the same voltage as +5VSB
(5.0V). The open collector driver on the motherboard,
pulls the signal wire to around 0.4V or so when it
wants the PSU to come on. And normally, that would be
sufficient to pull down the PS_ON# on the power supply.

At one time, monster chips were used to drive PS_ON#.
Like maybe a 7407 or a 74F07. In modern times, they
don't like to waste money on additional components,
so perhaps one of the larger chips is driving it.
And in such a case, the drive capability might be
a bit less.

This interface doesn't have a good track record, and
users are called upon to debug this particular wire,
way more than they should have to. As that user,
you don't know if it is the motherboard driver that
is defective, or the Antec end which is defective.
Doing the ground strap on PS_ON# test, is one way
to verify the power supply is holding up its end.

Knowing it's an Antec, probably a Truepower series,
my guess is the Antec has given up. The Antec supplies
built by ChannelWell weren't the best, as I had two
fail here. The main transformer inside says "CWT" on
some of the printed text, which is one identifier.
But mine didn't fail on PS_ON#, and instead they
had leaking caps (sizzling noise at startup). There
were some later Antecs built by Delta, and each OEM
supplier will have their own failure characteristics.
(Ways they like to fail.) Antec doesn't build their
own supplies, but buys them from an OEM. This is
quite common in the power supply industry - although
several of the majors do actually make their own.

Since the wiring on PS_ON# is wired OR logic, you
can also ground the PS_ON# signal while the system
is completely assembled. But that's physically a
bit difficult to do. The system will start if you
do that, but it also overrides THERMTRIP and you
would have no protection against a CPU overheat
if run that way.

For more info, there are three power supply specs on
formfactors, and the one I'm including here is for
24 pin (modern) supplies. The second link gives an overview
on the connectors used on ATX, in a general way. The second
link shows you that the 20 and 24 pin cables share a
subset of pins, so work the same. It's just the pin
numbering which is a bit screwy. So I might use wire
color to identify my PS_ON# signal. The spec has
a wire color table in it.

http://www.formfactors.org/developer/specs/ATX12V_PSDG_2_2_public_br2.pdf

http://www.playtool.com/pages/psuconnectors/connectors.html

Paul

[Ste...@pctamers.eu]

Hi Paul

thanks for the great reply. I'm not a hardware person (the system was home built by a friend). I like the idea of the standalone test that you mention.

Obviously I remove the Antec psu but I'm not sure how I "drive the PS_ON# with a ground strap." ? Physically, what do I do?

Cheers!

Steve

[Bob F]

There is one green wire on the large connector to the motherboard, with
black wire beside it. make an electrical connection between the
conductors to the green and black wires. You can do this with a rebent
paper clip or small piece of wire. The power supply should power up, and
it's fan should run. You can then measure, if you have a volt meter, the
voltages on the various connectors. Red should be about 5V, Yellow about
12V, and orange about 3.3V. Measure between the colored wire and any
black wire.

[Paul]

Being careful to use the colors in the wiring table
in the formfactors document, to make sure you have the
correct set of pins. In the 2.2 spec for 24 pin power
supplies, it's pin 15 and 16.

When the ATX supply is disconnected from the system,
you can insert a conductor into the round holes on the
very bottom of the main connector. Using the wire
color as your guide.

If you want to run the ATX supply (keep it on) using
a strap while the ATX supply is fully connected to
the computer, you can slide some wire (or a paper clip)
into the connector from the top. There is enough metal
available from the top side, to insert a meter probe
and take a reading. Or, to insert a paper clip and
have it make connections with pin 15 and 16.

Just be careful to short the right stuff :-)

*******

This one probably shorts 15 to 16 for you.

https://www.amazon.com/eForCity-24-pin-Power-Supply-Tester/dp/B005CTCD6S

What a unit like that uses, is window comparators to check
voltage ranges. The +12V monitor for example, would light the
green LED if the wire is between 11.40V and 12.6V (plus or minus
5%). A unit like that, should have a small dummy load for +5V
or similar. If the voltage is too high or too low, the
LED should go off.

*******

I have my own load tester (home made), and for pins, I use
a couple packages of WMF-3916MT. Those are crimp pins, which are
"male Molex" and the male end fits into the round holes
for pin 15 and 16. I have an electrical switch soldered
to two pins from the package of male pins, and that's how I
switch an ATX supply under electrical test On and Off.
It seems physically impossible today, to get a picture
of this particular product.

http://www.ba-electronics.com/molex.htm

39-00-0041 WMF-3916MT 15 pcs Contacts, male

In the ATX spec sheet, it mentions the 39-00-0041 number
as the model number of the male pin to be used. The WMF
number is a package of pins. Normally, you use a
proper crimping tool, so no solder is needed. But
I just mangle those with pliers and use a bit of solder :-)
I cover the body of the pin with polyolefin heat shrink
tubing. That's to prevent adjacent wires from touching.

You can also use a 24 pin extension cable for ATX, as
"raw materials" for shorting stuff, taking measurements
and so on. The deal with these, is getting one where
the wire colors are compliant with the ATX spec. That
makes it easier when you're trying to spot the green
and black wire, for a bit of "surgery". Cutting and
stripping wire ends, to make them touch and so on.
Normally, the extensions are a bit longer than this
cheesy example.

https://www.amazon.com/StarTech-8-Inch-Power-Extension-ATX24POWEXT/dp/B000FL60AI

When I bought one of those locally, the jokers who
made the cable, used the wrong colors for everything.
Like they didn't have a clue how important colors
are for the user population working with the
materials. That's why you need a picture of the
extension cable first, to see whether the colors
are in the right order.

So there are plenty of odds and ends floating
around. Just be careful (as they would say in
my electronics class).

As a kid, I got kicked across the basement
floor by 15kV, and the funny thing is, I was
careful after that. Professionally, the highest
voltage I've had to work with, is 800V. So, pretty
safe (unless you happen to touch it of course).
In another lab for the same company, when I asked
about the 800V thing, the guy says "oh, Jim gets
shocks all the time". Really building my
confidence in the stuff. Jim is, of course,
partially bald (from the abuse). The ATX power
supply is safe low voltage, so "burning something"
is about the worst you can do there. You could
get a small shock, but you're really have to
make an effort to do that (do something dumb).

Paul