Computer Won't Start Problem

C

Carl

I have a PC with an ASUS A8N-SLI Premium MB, in a Sonata II case (Antec
SmartPower 450 power supply).

The computer had been running for many days without interruption. This
morning I shut it down, using the Shutdown feature of Windows XP. It
turned off as expected.

But when I pushed the on button on the front of the case to turn it
back on, nothing happened.

I did some checking. My multimeter indicates the on button works, and
the lead to the motherboard power switch pins is correctly attached. I
checked the green wire running to the motherboard from the power
supply, and it shows 3.7 volts. But when I click the on button, the
voltage does not drop! (The on button is supposed to drop this voltage
to below a volt, signalling the power supply to turn on.)

For fun I shorted the green wire to ground, and the power supply turns
on and things start to happen. But when I remove the short the power
supply turns off immediately.

However, if after shorting the green wire as described above I then hit
the on button again, the computer starts up normally, and continues to
run OK.

Anyone have any idea about what is happening here?

TIA

--Carl
 
A

Andy

Is this a repeatable or a one time occurrence? Obviously there was
some kind of hiccup with the motherboard power control circuitry. If
it happens again, you can try turning off the standby power. Then turn
it back on after a minute or so. The power button should probably then
work.
 
C

Carl

Well, it happened for the first time today. I rarely turn the computer
off, but when I do I could always restart it with the on button. Not
today. It was off all day.

I am not sure what you mean by "standby power". I pulled the plug
several times during the day, without any apparenent effect. Maybe I
should have left it unplugged for an hour or so to see what that does.

I am wondering if some of the motherboard power control circuitry has a
bad resistor or capacitor. Or maybe the power supply is putting out
too much voltage on that green wire.

--Carl
 
W

w_tom

Input to the power supply controller include power switch and purple
wire (5 volt) wire from power supply. Using the multimeter and with
power supply connected to motherboard, that purple wire must measure at
least 4.87 volts.

Since green wire voltage does not drop below 0.8 volts, then measure
voltage where switch connects to motherboard (use wire or paper clip,
if necessary, to connect each meter probe to those motherboard pins).
When switch is not pressed, then voltage should be well above 2 volts.
When swtich is pressed, that voltage should drop to below 0.8 volts.

One final check. With power supply still connected to mothboard.
Put meter into current measurement with high amperage rating (maybe 2
amps). Connect meter from green wire to chassis. Power supply will
turn on. Lower multimeter current range until a non-zero reading is
measured. That current should be less than 5 milliamps.

Also report back numbers obtained since numbers may report other
useful facts.

If inputs to power supply controller are OK, if current (not voltage)
from green wire is sufficiently low, and if power supply powers during
that current reading, then power supply controller on motherboard is
defective. Inspect motherboard for obvious problems, or for a standoff
shorted to a motherboard printed circuit trace, or for an obvious
bulging electrolytic capacitor. Otherwise motherboard needs replacing.
 
C

Carl

w_tom said:
Input to the power supply controller include power switch and purple
wire (5 volt) wire from power supply. Using the multimeter and with
power supply connected to motherboard, that purple wire must measure at
least 4.87 volts.

I only get 3.74 volts on the purple wire.
Since green wire voltage does not drop below 0.8 volts, then measure
voltage where switch connects to motherboard (use wire or paper clip,
if necessary, to connect each meter probe to those motherboard pins).
When switch is not pressed, then voltage should be well above 2 volts.
When swtich is pressed, that voltage should drop to below 0.8 volts.

With the switch open the voltage reads 3.33 volts. It drops to zero
when the switch is closed. (That essentially grounds the pin.)
One final check. With power supply still connected to mothboard.
Put meter into current measurement with high amperage rating (maybe 2
amps). Connect meter from green wire to chassis. Power supply will
turn on. Lower multimeter current range until a non-zero reading is
measured. That current should be less than 5 milliamps.

I could not get a current reading on the green wire to ground.
Interestingly, when I attempted it the power supply did not turn on;
apparently the meter resistance was too high to drop the voltage on the
green wire. I had to actually short the green wire to ground to get
the power supply going; as soon as I removed the short it stopped. But
then I could press the computer's on button and it would start up (as
described previously).
Also report back numbers obtained since numbers may report other
useful facts.

If inputs to power supply controller are OK, if current (not voltage)
from green wire is sufficiently low, and if power supply powers during
that current reading, then power supply controller on motherboard is
defective. Inspect motherboard for obvious problems, or for a standoff
shorted to a motherboard printed circuit trace, or for an obvious
bulging electrolytic capacitor. Otherwise motherboard needs replacing.

Looks to me like a power supply problem. I obtained another power
supply, but installing it essentially involves re-building the
computer, including removing the motherboard. Meanwhile the thing is
running great; I just have this problem if I shut it down. Maybe this
weekend.

--Carl
 
W

w_tom

Carl said:
I only get 3.74 volts on the purple wire.

We need not go any farther. +5 volts digital electronics (number
varies slightly with different famillies) will not operate reliable
when voltage is below 4.8 volts. Only question remains is why that
voltage is below 4.87. Either the +5VSB (purple wire) voltage
regulator is defective, or motherboard power supply controller has a
defect - draws too much power.

Inspect for obvious problems. For example, a stray metal splinter
wedged somewhere on motherboard or an electrolytic capacitor with a
buldging top could account for this failure. However next paragraph
traces problem to power supply side or to motherboard side.

Disconnect power supply from motherboard (after disconnecting power
cord), then restore power and measure that 'purple to black wire'
voltage. If still low, then power supply is bad - case closed. If
voltage is good, we need do a little more. This $1.80 part from Radio
Shack is two one ohm resistors:
http://www.radioshack.com/product/index.jsp?productId=2062290&cp
Put both resistors in series and connect that resistor chain from
purple to black wire. Voltage across the resistor assembly (2 ohm
load) should remain above 4.87 volts. If not, then power supply +5VSB
power supply is defective. If loaded +5VSB (purple wire) voltage does
remain above 4.87 volts, then power supply controller circuit on
motherboard may be defective. Move on to next paragraph.

+5VSB is why some peripherals can turn on computer. For example,
modem can be enabled to power on computer when phone line rings.
Therefore modem could be a reason why the +5VSB load is too large.
Removing load (devices) that demand +5VSB (modem, keyboard, mouse,
ethernet card, etc) might then permit +5VSB to rise from 3.74 to above
4.87. Excessive loading may be due to that peripheral device. If not,
then excessive load would be due to motherboard. Again, we have traced
the +5VSB problem to specific devices. Note, suspect list was not
limited to power supply or motherboard.

+5VSB voltage is too low either due to power supply or due to
motherboard / peripherals failure. Since voltage is too low, then
power supply controller will not operate - cannot turn on power supply.
Above procedure identifies where problem is located - which side of a
purple wire connection is causing low voltage. 3.74 voltage high
enough to illuminate a light but too low for digital operation (of
power supply controller). Above traces the problem to left or right.
Good luck and report back.
 

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments. After that, you can post your question and our members will help you out.

Ask a Question

Top