System does not completely power off

[jmf4150]

After system shutdown there is clicking which seems to come from the
hard drive and flashing of lights sometimes for a minute or so and
sometimes more persistently). When I disconnect the power source the
problems stops.

The clicking corresponds to the lights "on the tower" and seems to
slow down but never completelt shut off. I am trying to determine
whether it is a hardware or software problem.

Has anyone ever experienced this?

 

[JS]

As a starting point, check your Power Options settings in the control panel.

JS

 

[jmf4150]

Power Options are correct. I opened the tower and now I think it's the
fans and USB fan controller which seem to slow down after shutdown and
then the fan blade starts to flip back and forth rather than spinning,
almost like they are still receiving minimal power. Very strange. Any
thoughts?

 

[Paul]

Power Options are correct. I opened the tower and now I think it's the
fans and USB fan controller which seem to slow down after shutdown and
then the fan blade starts to flip back and forth rather than spinning,
almost like they are still receiving minimal power. Very strange. Any
thoughts?

Oscillating power supply. Either a supply problem, or the motherboard
drive on PS_ON# is not deasserting properly.

PS_ON# controls the power supply. It is open collector logic, and
a pullup resistor on the power supply, pulls it up to +5V.

When the power supply is sitting on a table, with no motherboard
connected, and you turn it on at the back, nothing happens. That
is because the internal pullup resistor on the power supply,
causes the PS_ON# signal to rest at positive 5 volts (+5V).
When at 5V, it is in the "off" position (it is inverted logic).

When a motherboard is connected, the motherboard controls the power
supply via PS_ON#. When the motherboard want the supply on, it
grounds PS_ON# (makes it zero volts). To do that, a driver chip
on the motherboard "sinks" or absorbs a small amount of current
from the pin.

Now, imagine the following scenario. The user just selected "shutdown"
in Windows from the menu. Windows is all done, and the OS flips
PS_ON# to the non-driven (resting at 5V) state. But the signal
doesn't go cleanly to the full 5V value. The power supply can be
in a "half on" state, where the voltages are no longer full value
and correct. Or the supply could flip itself on and off rapidly,
leading to twitching fan blades, and a very confused computer.

Now, in this situation, it could be the fault of the power supply,
or the motherboard may not be healthy. While there are probably
ways to identify the problem, with instruments, it is faster
to just try another PSU, and see if the behavior is the same.

(I think if I had to determine for sure, which side was at fault,
I'd have to insert a series resistor in the PS_ON# line, and
observe the voltage on either end of the resistor. Maybe
around 100 ohms or so. By looking at the voltage levels on either
end, it gives an idea if the driver is driving, and whether the
pullup inside the power supply is working or not. The voltage
across the resistor, also tells you the amount of current flow
in the line. To do this kind of test, I'd purchase an ATX main
power extender cable, so I could chop the PS_ON# line in two,
and insert the resistor. That is what I mean by a test. Swapping
the supply, is less work than all that nonsense.)

Paul

 

[jmf4150]

Oscillating power supply. Either a supply problem, or the motherboard
drive on PS_ON# is not deasserting properly.

PS_ON# controls the power supply. It is open collector logic, and
a pullup resistor on the power supply, pulls it up to +5V.

When the power supply is sitting on a table, with no motherboard
connected, and you turn it on at the back, nothing happens. That
is because the internal pullup resistor on the power supply,
causes the PS_ON# signal to rest at positive 5 volts (+5V).
When at 5V, it is in the "off" position (it is inverted logic).

When a motherboard is connected, the motherboard controls the power
supply via PS_ON#. When the motherboard want the supply on, it
grounds PS_ON# (makes it zero volts). To do that, a driver chip
on the motherboard "sinks" or absorbs a small amount of current
from the pin.

Now, imagine the following scenario. The user just selected "shutdown"
in Windows from the menu. Windows is all done, and the OS flips
PS_ON# to the non-driven (resting at 5V) state. But the signal
doesn't go cleanly to the full 5V value. The power supply can be
in a "half on" state, where the voltages are no longer full value
and correct. Or the supply could flip itself on and off rapidly,
leading to twitching fan blades, and a very confused computer.

Now, in this situation, it could be the fault of the power supply,
or the motherboard may not be healthy. While there are probably
ways to identify the problem, with instruments, it is faster
to just try another PSU, and see if the behavior is the same.

(I think if I had to determine for sure, which side was at fault,
I'd have to insert a series resistor in the PS_ON# line, and
observe the voltage on either end of the resistor. Maybe
around 100 ohms or so. By looking at the voltage levels on either
end, it gives an idea if the driver is driving, and whether the
pullup inside the power supply is working or not. The voltage
across the resistor, also tells you the amount of current flow
in the line. To do this kind of test, I'd purchase an ATX main
power extender cable, so I could chop the PS_ON# line in two,
and insert the resistor. That is what I mean by a test. Swapping
the supply, is less work than all that nonsense.)

Paul

Thanks that makes sense. Looks like I should replace the power supply.
I installed the one that in the tower sometime ago. On the otherhand,
any danger if I just turn off the surge protecter after shutting down
and continue to use it this way?

 

[JS]

I had a somewhat similar situation with a set of speakers that if I powered
on before turning on the PC caused a problem like yours.
My solution was simple, turn on the PC first then after booting into Windows
turn on the speakers.

JS

 

[Paul]

Thanks that makes sense. Looks like I should replace the power supply.
I installed the one that in the tower sometime ago. On the otherhand,
any danger if I just turn off the surge protecter after shutting down
and continue to use it this way?

I'll take a pass on answering that one.

You could always leave it running, and not use the shutdown, until the
replacement PSU has been purchased. Just switch off the monitor between
sessions. I've worked in places, where all the computers are left running
all the time, and the only thing we had that used to smoke, was the
occasional CRT monitor.

With my home computers, if one of them is ill, I just switch to another
until it is fixed. If I need the files, I can always move the disk from
one machine to another. All my disks are IDE, so no problem moving them
from machine to machine.

Paul

 

[jmf4150]

I'll take a pass on answering that one.

You could always leave it running, and not use the shutdown, until the
replacement PSU has been purchased. Just switch off the monitor between
sessions. I've worked in places, where all the computers are left running
all the time, and the only thing we had that used to smoke, was the
occasional CRT monitor.

With my home computers, if one of them is ill, I just switch to another
until it is fixed. If I need the files, I can always move the disk from
one machine to another. All my disks are IDE, so no problem moving them
from machine to machine.

Paul- Hide quoted text -

- Show quoted text -

Thanks again