barrowhill said:
SC Tom,
We've all made those mistakes...............Now get your head round
this....
Last thought was perhaps PSU is after all faulty. Thinks......Somehow
fires up fans and fan lights come on but other wise does nothing.
Fortunately have spare so replaced. Note: new PSU tested and voltage
measured off board by shorting pins14 (Power Supply On) to Ground
(COM). All Good.
Install main board molex connectors. Do not connect power to DVD's.
Press ON button ----- nothing ??? Old PSU starts fans/lights when ON
button pressed (?) . What's happening (?) Remove 20 pin molex, short pin
14 to ground fans start and fan lights come on. Now my understanding is
pin 14 (power supply on) is only active (low) when PSU receives
motherboard OK signal
Why should one PSU 'fire' up when ON is pressed and the other not??. Do
I have a faulty motherboard ?? I suspect I have but.......... some
rational behind what's happening with PSU's is welcome
I would start by checking whether you still had data cables connected
from the DVD. If you're going to disconnect a drive, it would be
better to disconnect both power and data. A data connector can be left
connected if an I/O pad is known as "failsafe", but we don't know whether
that is the case or not. So when the characteristics of a device are
unknown, it is best to disconnect both data and power to it at the same
time, just in case.
The motherboard has an open collector driver, to pull to ground the
PS_ON# signal (pin 14). The power supply end has a pullup, and it is weak and
easy for PS_ON# to overcome and pull to ground. A logic low level,
is what triggers the power supply into the ON state.
The front panel POWER button (soft power), gives a momentary low pulse
at logic levels. The front power button doesn't actually gate the
flow of power directly. It sends a logic signal to the motherboard.
The motherboard "latches" the front panel signal and converts in into
a steady level.
The motherboard logic has many inputs into that circuit.
1) The Southbridge and SuperI/O have a number of "Wake On" conditions,
that they accept and use to turn on the computer when you want to
wake it up. For example, if you enable "Wake On LAN" on your system,
when a LAN packet arrives, it asserts PME, which is then used to
(eventually) turn on PS_ON#.
2) Certain safety features on the motherboard, work in the opposite
sense. Modern motherboards (both AMD and Intel) have THERMTRIP.
If the processor asserts THERMTRIP during an overheat event,
the motherboard logic de-asserts PS_ON# immediately, and that
causes the power supply to go off.
Some of those safety features are "latching", and you must turn off
the main power switch on the back of the computer, before your next
attempt to turn on the computer again via the front (soft) power switch.
On some older Asus motherboards, the "AGP Warn" circuit did something
similar. It prevented power from coming on, when a 3.3V only video
card was inserted into a 1.5V only motherboard AGP slot. That would
not be a concern on a PCI Express motherboard.
When your system refuses to respond, you can see there are a couple kinds
of logic inputs, that are presented to the motherboard. There can also
be situations where "accidental" things happen. For example, on my old
440BX motherboard, if the IDE connector was installed crooked (only
half the pins touching), the resulting electrical stress on the
Southbridge would cause the power to come on, on its own, without
touching the button on the front of the computer.
On some motherboards, if the CMOS battery is flat, and reads zero volts
with a multimeter, that can prevent PS_ON# from working. A possible
theory is, it happens at the SuperI/O chip, as it can have a connection
directly to the CMOS battery. And in that case, perhaps the SuperI/O isn't
passing a logic signal from somewhere else on the motherboard, until
the battery is replaced.
The chip driving PS_ON# can become weak. Or, the power supply end
can do a poor job of dealing with the logic level on the PS_ON#
signal. The power supply doesn't use a logic chip, but can use a
discrete transistor circuit for the job.
The system is relatively complicated (and really deserves a better
treatment than I can give it).
The circuitry of this type, might be referred to as "Supervisory"
circuitry. On the motherboard, the +5VSB rail from the power supply,
powers the Supervisory circuits. If your power supply has a weak
or glitchy +5VSB, then that can cause problems turning on the computer.
On an Asus motherboard, they put a green LED on the motherboard wired
directly to +5VSB. On such a motherboard, you visually check that LED,
to make sure it remains lit as long as the switch on the back of the
computer is in the ON position. If that supply rail is overloaded
(+5VSB has a weak rating in amperes), the LED can go on and off rapidly,
or blink. And that would tell you that between the motherboard and
the power supply, something is happening to the ability to maintain
a steady 5V on the +5VSB (StandBy) rail. A weak supply can do that,
as could a short on +5VSB from say, a USB load or PS/2 load.
Paul
