Powering up via Keyboard (only works after shutdown)

I've just assembled a 'media box' (Asus P4SP-MX, Antec Aria Case/PSU,
2GHz Celeron, Sapphire Radeon 9200SE, Win XP SP2) and run smack into
the brick wall that is my understanding of ACPI. I've got the BIOS set
to power up in response to a keyboard event; which works fine if the
computer was powered down using the 'shutdown' option in windows but
completely ignores me if I've just switched the power on at the mains
socket.

I *think* this must be because when windows shuts down the PSU
continues to draw a trickle charge (the 5VSB line?) which is then
enough to monitor the keyboard. For some reason though the PSU can't
respond to being switched on at the mains by going 'Oh good, power,
now I can turn on enough to wait for that nice man to fully start me
up with his keyboard'.

It is physically difficult to get to the power button on the machine
itself since it lives in a wardrobe with extension leads running to
monitor, speakers, network, KVM switch, etc. It runs cool enough (32C
under load) but I'd prefer to switch it off at the mains when I'm not
using it.

My questions, therefore, a

1. Am I right that a windows shutdown still leaves the box drawing
some power (which my switching off at the wall interupts)?

2. What options do I have for powering it up after turning it on at
the wall (i.e. will only the power switch on the box itself work)?

3. Does leaving the box switched on at the wall (after a windows
shutdown) represent more of a fire risk than say a DVD, Video or TV?

Regards,

Matt

Tomcat wrote:
I've just assembled a 'media box' (Asus P4SP-MX, Antec Aria Case/PSU,
2GHz Celeron, Sapphire Radeon 9200SE, Win XP SP2) and run smack into
the brick wall that is my understanding of ACPI. I've got the BIOS set
to power up in response to a keyboard event; which works fine if the
computer was powered down using the 'shutdown' option in windows but
completely ignores me if I've just switched the power on at the mains
socket.

I *think* this must be because when windows shuts down the PSU
continues to draw a trickle charge (the 5VSB line?) which is then
enough to monitor the keyboard. For some reason though the PSU can't
respond to being switched on at the mains by going 'Oh good, power,
now I can turn on enough to wait for that nice man to fully start me
up with his keyboard'.

It is physically difficult to get to the power button on the machine
itself since it lives in a wardrobe with extension leads running to
monitor, speakers, network, KVM switch, etc. It runs cool enough (32C
under load) but I'd prefer to switch it off at the mains when I'm not
using it.

My questions, therefore, a

1. Am I right that a windows shutdown still leaves the box drawing
some power (which my switching off at the wall interupts)?

2. What options do I have for powering it up after turning it on at
the wall (i.e. will only the power switch on the box itself work)?

3. Does leaving the box switched on at the wall (after a windows
shutdown) represent more of a fire risk than say a DVD, Video or TV?

Regards,

Matt

1. Yes. But the Southbridge has two sections to it. One is powered
by the "awake" voltages, the other half via +5VSB or the CMOS
battery. I would hope the ACPI related circuits are powered by
the CMOS battery, in which case the motherboard could be
awakened after a power failure. Maybe the state of the keyboard
has something to do with it ? Does a keyboard need to be configured,
or will it do the right thing when its power comes up ? Have you
tried waking the system after a power failure, using any other
method supported by your motherboard ?
2. See "AC Power Loss Restart". If you enable it, then every time
power is applied to the box, it will boot. The only question is,
whether it can be shut down properly again. I'm not sure "last
state" is always going to do the right thing. In any case,
play with that setting and maybe you'll get some functionality
you can use.
3. The +5VSB circuit inside the PSU is running, and the fan is not.
Something inside the PSU does get warm by doing this. As I've
never seen a circuit diagram for how they do +5VSB, it is hard
to guess at how it might fail. Now that I think about it,
it is possible the main caps on the PSU are charged by
line voltage, while +5VSB is running. That is why +5VSB can
run for 30 seconds after the switch on the back of the PSU is
switched off. The main caps are the only thing that could store
that much power. So, I would say there is still some nasty
stuff in the loop at that point (+5VSB being used). In a sense,
if +5VSB is still running, a portion of the main circuit
(rectifiers and caps) is still running too - it is just the
main switcher that is stopped. I guess that makes it a little
more reliable than when the whole thing is running, but not
by much. The main transistors still have potential across them.
But, no secondary side faults can get you - so no failed
components on the other side of the transformer can affect the
PSU when it is "sleeping".

HTH,
Paul

It is the same with Asus A7N8X-E.

"Paul" a écrit dans le message de news:
...
Tomcat wrote:
I've just assembled a 'media box' (Asus P4SP-MX, Antec Aria Case/PSU,
2GHz Celeron, Sapphire Radeon 9200SE, Win XP SP2) and run smack into
the brick wall that is my understanding of ACPI. I've got the BIOS set
to power up in response to a keyboard event; which works fine if the
computer was powered down using the 'shutdown' option in windows but
completely ignores me if I've just switched the power on at the mains
socket.

FG wrote:
It is the same with Asus A7N8X-E.

It has an AC Power Loss Restart setting also. I don't know
if it has the same wake up options, as they tend to differ
from product to product.

Paul