Power problem...

[Travis King]

Asus A7M266 motherboard. (Pushing four years old) Last night when mother
went to shut down the computer, the LEDs shut off, but the computer
continued to run. Holding in the power switch would not shut down the
system. You had to unplug it. When plugged back on, it would not boot (or
even turn on). Pressed power button several times and it wouldn't boot.
Walked away from the computer and then, it turned on by itself when mother
started to walk away from the computer. It's had the boot problem here and
there for about two or three months where it takes several tries to get the
computer to turn on. Motherboard is also losing time, which we've replaced
the battery already. We were told on this NG that it could be a hardware
conflict. The only hardware we changed was the processor, and it was a week
or two later before the time problem showed up again. What's wrong with the
computer? Thanks in advance.

 

[Paul]

Asus A7M266 motherboard. (Pushing four years old) Last night when mother
went to shut down the computer, the LEDs shut off, but the computer
continued to run. Holding in the power switch would not shut down the
system. You had to unplug it. When plugged back on, it would not boot (or
even turn on). Pressed power button several times and it wouldn't boot.
Walked away from the computer and then, it turned on by itself when mother
started to walk away from the computer. It's had the boot problem here and
there for about two or three months where it takes several tries to get the
computer to turn on. Motherboard is also losing time, which we've replaced
the battery already. We were told on this NG that it could be a hardware
conflict. The only hardware we changed was the processor, and it was a week
or two later before the time problem showed up again. What's wrong with the
computer? Thanks in advance.

The debugging techniques a home builder can use are
pretty limited. It takes some amount of luck to find
even the simplest of problems. Motherboards are
very complicated, and there are hundreds of electrical
connections inside them. Each failure case can give
different symptoms, and you cannot always uniquely map
a given symptom to a root cause of that symptom.

I've had one experience like yours. I had a 440BX based
motherboard, and one of my IDE cables was loose. It was
only half plugged in. That caused the computer to switch
on immediately, without needing to press the front power
switch. I was also not able to turn off the computer with
the front power switch, and I just pulled the plug on the
thing. Once I discovered the loose IDE cable, it recovered
(so no permanent damage).

The "power up" logic flows through the Southbridge and the
Super I/O chip. It may go through a few other anonymous
transistors and driver chips, if all the required functions
are not inside the Southbridge or Super I/O chip. If a fault
on the motherboard affects the power supplied to either of
those chips, they might cause the PS_ON# signal to be asserted
when it is not supposed to be. (I am referring to onboard
regulator circuits, and not necessarily just the power supply.
Either of those sources could be at fault.)

You mention problems with your CMOS battery, and a cause of
those problems can be not following the "clear CMOS" procedure
correctly. The older Asus motherboards are designed in a
dangerous way, such that if you plug the CMOS jumper to the
clear position, while there is still +5VSB flowing to the
board, a small three lead diode device gets burned. This
could result in the CMOS circuitry in the Southbridge, running
off the battery all of the time (as the connection to +5VSB
is burnt out). And when the Southbridge is awake and running at
full speed, the current drawn from the battery is a lot larger
than the small sleep current the battery is normally called on
to provide.

A couple of posters here, have managed to fix their
motherboards when that problem has happened. One guy, to
his credit, used a couple of separate diodes, to replace
the three pin package, as he presumably didn't have a place
to buy the proper part. The parts are tiny, and difficult to
solder. First you have to find the damn diode package, and
using pictures of motherboards, I cannot find the part all
of the time. (The part number on the top of it is so small, I
use a small lense to magnify it.) Sometimes the part is very
close to the CMOS battery, and it is a matter of "guilt by
association". If the part is placed half way across the board,
I'd never find it. You then have to prove electrically, that
it is faulty (which is easy if it is physically burned, but
is more difficult if still intact). The part number on the
top of it is "K45", with a couple of other digits that might
be a date code. The part is a BAS40-05 or a BAS40W-05 (one
package is a little taller than the other).

http://www.diodes.com/tracker/pdftr...com/product_catalog/search.php&ds=ds11006.pdf

The first part of your debugging process, is to simplify the
setup. A "cardboard" test is a good way to proceed, as you
have better access to stuff on the board, and there is less
chance of an accident happening inside a cramped case.

The power on/off function should still work, even if you
have just the motherboard and a PSU to work with. With the
PSU hooked up, the power supply fan should remain off until
the computer case power switch is pressed. (You can fake a
power switch, by momentarily touching a screwdriver tip to
the two pins, where normally you would install the power
switch. A momentary touch is latched by the motherboard and
is turned into a steady level for the PS_ON# signal to the
motherboard.)

As you have noticed a problem with your CMOS battery, as
part of my debugging procedure, the first thing I would do
is probe with a voltmeter, to the pin on the Southbridge
that is fed the sustaining voltage. Since you have a
VIA Southbridge, no pinout information is available, and
that means all you can do, is look at the flow of copper
wires, and try and figure out which copper wire carries
the sustaining voltage. If the problem was with your
Northbridge, AMD does have datasheets for the 761 and 762,
but knowing the pinout for them, is not going to help with
the debugging of the Southbridge.

As I say, the amount of debugging you can do is strictly
limited, unless you have a lot of resources (schematic,
datasheets, tech info) to aid in the effort.

If the motherboard behaves better by itself, plug stuff in
one component at a time, until the motherboard is again
misbehaving. You'd have to be pretty lucky to isolate the
problem to a single component, but you never know until
you give it a try.

This isn't likely to be a power supply problem, but you
can either measure the voltages being fed to the (empty)
motherboard, or swap a known good power supply in and
see if the symptoms change.

There are some people who do motherboard repair at the
component level, but expect the repair charge to start
at about $50. For that price, it is a lot simpler to try
to find a replacement motherboard. Finding a good one,
that will reuse your RAM, will be the tough part.

Paul

 

[Travis King]

The debugging techniques a home builder can use are
pretty limited. It takes some amount of luck to find
even the simplest of problems. Motherboards are
very complicated, and there are hundreds of electrical
connections inside them. Each failure case can give
different symptoms, and you cannot always uniquely map
a given symptom to a root cause of that symptom.

I've had one experience like yours. I had a 440BX based
motherboard, and one of my IDE cables was loose. It was
only half plugged in. That caused the computer to switch
on immediately, without needing to press the front power
switch. I was also not able to turn off the computer with
the front power switch, and I just pulled the plug on the
thing. Once I discovered the loose IDE cable, it recovered
(so no permanent damage).

The "power up" logic flows through the Southbridge and the
Super I/O chip. It may go through a few other anonymous
transistors and driver chips, if all the required functions
are not inside the Southbridge or Super I/O chip. If a fault
on the motherboard affects the power supplied to either of
those chips, they might cause the PS_ON# signal to be asserted
when it is not supposed to be. (I am referring to onboard
regulator circuits, and not necessarily just the power supply.
Either of those sources could be at fault.)

You mention problems with your CMOS battery, and a cause of
those problems can be not following the "clear CMOS" procedure
correctly. The older Asus motherboards are designed in a
dangerous way, such that if you plug the CMOS jumper to the
clear position, while there is still +5VSB flowing to the
board, a small three lead diode device gets burned. This
could result in the CMOS circuitry in the Southbridge, running
off the battery all of the time (as the connection to +5VSB
is burnt out). And when the Southbridge is awake and running at
full speed, the current drawn from the battery is a lot larger
than the small sleep current the battery is normally called on
to provide.

A couple of posters here, have managed to fix their
motherboards when that problem has happened. One guy, to
his credit, used a couple of separate diodes, to replace
the three pin package, as he presumably didn't have a place
to buy the proper part. The parts are tiny, and difficult to
solder. First you have to find the damn diode package, and
using pictures of motherboards, I cannot find the part all
of the time. (The part number on the top of it is so small, I
use a small lense to magnify it.) Sometimes the part is very
close to the CMOS battery, and it is a matter of "guilt by
association". If the part is placed half way across the board,
I'd never find it. You then have to prove electrically, that
it is faulty (which is easy if it is physically burned, but
is more difficult if still intact). The part number on the
top of it is "K45", with a couple of other digits that might
be a date code. The part is a BAS40-05 or a BAS40W-05 (one
package is a little taller than the other).

http://www.diodes.com/tracker/pdftr...com/product_catalog/search.php&ds=ds11006.pdf

The first part of your debugging process, is to simplify the
setup. A "cardboard" test is a good way to proceed, as you
have better access to stuff on the board, and there is less
chance of an accident happening inside a cramped case.

The power on/off function should still work, even if you
have just the motherboard and a PSU to work with. With the
PSU hooked up, the power supply fan should remain off until
the computer case power switch is pressed. (You can fake a
power switch, by momentarily touching a screwdriver tip to
the two pins, where normally you would install the power
switch. A momentary touch is latched by the motherboard and
is turned into a steady level for the PS_ON# signal to the
motherboard.)

As you have noticed a problem with your CMOS battery, as
part of my debugging procedure, the first thing I would do
is probe with a voltmeter, to the pin on the Southbridge
that is fed the sustaining voltage. Since you have a
VIA Southbridge, no pinout information is available, and
that means all you can do, is look at the flow of copper
wires, and try and figure out which copper wire carries
the sustaining voltage. If the problem was with your
Northbridge, AMD does have datasheets for the 761 and 762,
but knowing the pinout for them, is not going to help with
the debugging of the Southbridge.

As I say, the amount of debugging you can do is strictly
limited, unless you have a lot of resources (schematic,
datasheets, tech info) to aid in the effort.

If the motherboard behaves better by itself, plug stuff in
one component at a time, until the motherboard is again
misbehaving. You'd have to be pretty lucky to isolate the
problem to a single component, but you never know until
you give it a try.

This isn't likely to be a power supply problem, but you
can either measure the voltages being fed to the (empty)
motherboard, or swap a known good power supply in and
see if the symptoms change.

There are some people who do motherboard repair at the
component level, but expect the repair charge to start
at about $50. For that price, it is a lot simpler to try
to find a replacement motherboard. Finding a good one,
that will reuse your RAM, will be the tough part.

Paul

According to Asus Probe, all the voltages are ABOVE where they're supposed
to be. the 12 is at 12.5 and the others are just a little above. Does this
mean that the PSU is holding really strong or does this indicate a problem?
(It's 4% above 12.) Thanks.

 

[Travis King]

The debugging techniques a home builder can use are
pretty limited. It takes some amount of luck to find
even the simplest of problems. Motherboards are
very complicated, and there are hundreds of electrical
connections inside them. Each failure case can give
different symptoms, and you cannot always uniquely map
a given symptom to a root cause of that symptom.

I've had one experience like yours. I had a 440BX based
motherboard, and one of my IDE cables was loose. It was
only half plugged in. That caused the computer to switch
on immediately, without needing to press the front power
switch. I was also not able to turn off the computer with
the front power switch, and I just pulled the plug on the
thing. Once I discovered the loose IDE cable, it recovered
(so no permanent damage).

The "power up" logic flows through the Southbridge and the
Super I/O chip. It may go through a few other anonymous
transistors and driver chips, if all the required functions
are not inside the Southbridge or Super I/O chip. If a fault
on the motherboard affects the power supplied to either of
those chips, they might cause the PS_ON# signal to be asserted
when it is not supposed to be. (I am referring to onboard
regulator circuits, and not necessarily just the power supply.
Either of those sources could be at fault.)

You mention problems with your CMOS battery, and a cause of
those problems can be not following the "clear CMOS" procedure
correctly. The older Asus motherboards are designed in a
dangerous way, such that if you plug the CMOS jumper to the
clear position, while there is still +5VSB flowing to the
board, a small three lead diode device gets burned. This
could result in the CMOS circuitry in the Southbridge, running
off the battery all of the time (as the connection to +5VSB
is burnt out). And when the Southbridge is awake and running at
full speed, the current drawn from the battery is a lot larger
than the small sleep current the battery is normally called on
to provide.

A couple of posters here, have managed to fix their
motherboards when that problem has happened. One guy, to
his credit, used a couple of separate diodes, to replace
the three pin package, as he presumably didn't have a place
to buy the proper part. The parts are tiny, and difficult to
solder. First you have to find the damn diode package, and
using pictures of motherboards, I cannot find the part all
of the time. (The part number on the top of it is so small, I
use a small lense to magnify it.) Sometimes the part is very
close to the CMOS battery, and it is a matter of "guilt by
association". If the part is placed half way across the board,
I'd never find it. You then have to prove electrically, that
it is faulty (which is easy if it is physically burned, but
is more difficult if still intact). The part number on the
top of it is "K45", with a couple of other digits that might
be a date code. The part is a BAS40-05 or a BAS40W-05 (one
package is a little taller than the other).

http://www.diodes.com/tracker/pdftr...com/product_catalog/search.php&ds=ds11006.pdf

The first part of your debugging process, is to simplify the
setup. A "cardboard" test is a good way to proceed, as you
have better access to stuff on the board, and there is less
chance of an accident happening inside a cramped case.

The power on/off function should still work, even if you
have just the motherboard and a PSU to work with. With the
PSU hooked up, the power supply fan should remain off until
the computer case power switch is pressed. (You can fake a
power switch, by momentarily touching a screwdriver tip to
the two pins, where normally you would install the power
switch. A momentary touch is latched by the motherboard and
is turned into a steady level for the PS_ON# signal to the
motherboard.)

As you have noticed a problem with your CMOS battery, as
part of my debugging procedure, the first thing I would do
is probe with a voltmeter, to the pin on the Southbridge
that is fed the sustaining voltage. Since you have a
VIA Southbridge, no pinout information is available, and
that means all you can do, is look at the flow of copper
wires, and try and figure out which copper wire carries
the sustaining voltage. If the problem was with your
Northbridge, AMD does have datasheets for the 761 and 762,
but knowing the pinout for them, is not going to help with
the debugging of the Southbridge.

As I say, the amount of debugging you can do is strictly
limited, unless you have a lot of resources (schematic,
datasheets, tech info) to aid in the effort.

If the motherboard behaves better by itself, plug stuff in
one component at a time, until the motherboard is again
misbehaving. You'd have to be pretty lucky to isolate the
problem to a single component, but you never know until
you give it a try.

This isn't likely to be a power supply problem, but you
can either measure the voltages being fed to the (empty)
motherboard, or swap a known good power supply in and
see if the symptoms change.

There are some people who do motherboard repair at the
component level, but expect the repair charge to start
at about $50. For that price, it is a lot simpler to try
to find a replacement motherboard. Finding a good one,
that will reuse your RAM, will be the tough part.

Paul

This also reminds me that the fan on the northbridge a few months ago when I
opened up the computer, I noticed it was not spinning and was froze. I
oiled the fan and it's fine now, but would that have damaged the NB? I
don't know exactly how long it went like that because I only open the case
once in a while when I noticed it wasn't spinning. Would this have anything
that would have made any of these problems long-term?

 

[Paul]

This also reminds me that the fan on the northbridge a few months ago when I
opened up the computer, I noticed it was not spinning and was froze. I
oiled the fan and it's fine now, but would that have damaged the NB? I
don't know exactly how long it went like that because I only open the case
once in a while when I noticed it wasn't spinning. Would this have anything
that would have made any of these problems long-term?

It would take a Northbridge failure, plus the Northbridge drawing
lots of current from a motherboard regulator used by both
the Northbridge and Southbridge, for that to cause a problem.
In other words, you'd see a failure, but it would be more along
the lines of the computer starts when the power button on the
front is pushed, but the computer wouldn't POST.

The Northbridge has the memory controller, and a path to get
to the Southbridge, so an overheated Northbridge might cause
random crashing due to memory errors. The AGP bus might also
have the odd error, causing screen freezes or VPU recover
being triggered.

On some Northbridges, the fan is decorative. A nice tall
Zalman passive heatsink might do as much for them, as a
short heatsink plus a fan. On the other hand, some of the
new chipsets with PCI Express, have theoretical maximum power
dissipation of over 20 watts, and practical power of 12 watts
or so, and some amount of forced air is a good idea for
those. For example, I personally wouldn't run an A8N-SLI with
a passive heatsink - unless the CPU fan blows lots of air over
the fins.

In terms of your debugging exercise, a good visual inspection
might turn up something. When you cannot afford million dollar
test equipment, your eyes, ears, and nose have to be your
substitute. That, and perhaps a cheap multimeter

With regard to your power supply observation, an out-of-spec
monitor chip could account for all readings being high, or
your idea of the PSU outputs being on the high side is also
possible. A lot of ATX PSUs have a single extra wire, which goes
to one of the ATX 3.3V pins. You'll see two wires that go to one
of the 3.3v pins, and the thinner of the wires is a "remote sense"
wire. That is supposed to help reduce the error on the 3.3V
output, so I would expect the tightest regulation on that
supply. There are some supplies, such as Antec Truepower, where
three monitor wires are used, one each for 3.3V, 5V, and 12V.
These remote sense wires allow the power supply to monitor
the voltage as seen right at the connector, and that means
the voltage drop in the cable is nulled out.

The monitor chip can warn you of a problem, but a multimeter
should be used to verify and cross-check what is going on.

In terms of your board's architecture, take a look at pg.17
This is the datasheet for your AMD761 Northbridge.

http://www.amd.com/us-en/assets/content_type/white_papers_and_tech_docs/24088.pdf

In order for your computer to POST and attempt to boot, you
must have a working PCI bus. (In other words, a bad PCI card
could hang the bus and prevent POST.) Newer motherboards wouldn't
have that restriction, as a higher bandwidth private bus runs
between the Northbridge and Southbridge. And, on the A7M266,
a Via Southbridge is used in place of the AMD766 shown in
the diagram.

Here is a datasheet for the Southbridge, or at least a 686A.
Your chip is a 686B, and has ATA100, while the 686A is ATA66.
The pin count is the same, at 352 balls (BGA package). This
chip also includes the Super I/O functions, so this chip might
handle virtually all the power-on functions.

http://web.archive.org/web/20030617203815/http://www.via.com.tw/pdf/productinfo/686a.pdf

The front panel switch should be driving PWRBTN# (see pg. 31 of
the datasheet). The output of the chip is called SUSC# (pg. 25)
and is supposed to connect to PS_ON# on the ATX PSU. SUSC# is
the opposite sense of PS_ON#, and there should be an inverting
logic gate with open collector output, to drive PS_ON#. Thus,
I believe SUSC# is a logic 1 (+3.3V), in order for PS_ON# to
be driven to logic 0 (0 volts). Some external chip would be used
for this function. (There are a number of 14 or 16 pin DIP chips
on your motherboard, but I cannot see what the part numbers are.
It could be that the signal from the Southbridge goes to one of
those smaller chips, before driving the power supply PS_ON#
signal.)

In terms of the device power, the datasheet has this to say:

"There are three power planes inside the VT82C686A. The
first power plane (VCCS) is always on unless turned off by
the mechanical switch. The second power plane (VCC) is
controlled by chip output SUSC# (also called ³PSON²). The
third plane (VCCRTC) is powered by the combination of the
VCCS and the external battery (VBAT) for the integrated real
time clock. Most of the circuitry inside the VT82C686A is
powered by VCC. The amount of logic powered by VCCS is
very small; its main function is to control the supply of
VCC and other power planes. VCCRTC is always on unless both
the mechanical switch and VBAT are removed."

I suppose a failure of one of those signals to make it to
the Southbridge would upset things. VCCRTC is the power source
that comes from the three pin diode ORing of the CMOS battery
and power derived from +5VSB. VCCS would also be derived from
+5VSB, and would likely be needed to make SUSC# work
properly. VCC is what powers the majority of the logic in the
chip. While the datasheet doesn't make it very obvious, the
chip uses mostly 3.3V level power signals.

So, there is a bit of info for your board, but it'll be tough
to guess at how it is all wired together.

Sounds like "cardboard test" time. Perhaps if you pick up a
replacement motherboard, you can do some probing of the A7M266
before giving up on it entirely.

Paul

 

[Travis King]

It would take a Northbridge failure, plus the Northbridge drawing
lots of current from a motherboard regulator used by both
the Northbridge and Southbridge, for that to cause a problem.
In other words, you'd see a failure, but it would be more along
the lines of the computer starts when the power button on the
front is pushed, but the computer wouldn't POST.

The Northbridge has the memory controller, and a path to get
to the Southbridge, so an overheated Northbridge might cause
random crashing due to memory errors. The AGP bus might also
have the odd error, causing screen freezes or VPU recover
being triggered.

On some Northbridges, the fan is decorative. A nice tall
Zalman passive heatsink might do as much for them, as a
short heatsink plus a fan. On the other hand, some of the
new chipsets with PCI Express, have theoretical maximum power
dissipation of over 20 watts, and practical power of 12 watts
or so, and some amount of forced air is a good idea for
those. For example, I personally wouldn't run an A8N-SLI with
a passive heatsink - unless the CPU fan blows lots of air over
the fins.

In terms of your debugging exercise, a good visual inspection
might turn up something. When you cannot afford million dollar
test equipment, your eyes, ears, and nose have to be your
substitute. That, and perhaps a cheap multimeter

With regard to your power supply observation, an out-of-spec
monitor chip could account for all readings being high, or
your idea of the PSU outputs being on the high side is also
possible. A lot of ATX PSUs have a single extra wire, which goes
to one of the ATX 3.3V pins. You'll see two wires that go to one
of the 3.3v pins, and the thinner of the wires is a "remote sense"
wire. That is supposed to help reduce the error on the 3.3V
output, so I would expect the tightest regulation on that
supply. There are some supplies, such as Antec Truepower, where
three monitor wires are used, one each for 3.3V, 5V, and 12V.
These remote sense wires allow the power supply to monitor
the voltage as seen right at the connector, and that means
the voltage drop in the cable is nulled out.

The monitor chip can warn you of a problem, but a multimeter
should be used to verify and cross-check what is going on.

In terms of your board's architecture, take a look at pg.17
This is the datasheet for your AMD761 Northbridge.

http://www.amd.com/us-en/assets/content_type/white_papers_and_tech_docs/24088.pdf

In order for your computer to POST and attempt to boot, you
must have a working PCI bus. (In other words, a bad PCI card
could hang the bus and prevent POST.) Newer motherboards wouldn't
have that restriction, as a higher bandwidth private bus runs
between the Northbridge and Southbridge. And, on the A7M266,
a Via Southbridge is used in place of the AMD766 shown in
the diagram.

Here is a datasheet for the Southbridge, or at least a 686A.
Your chip is a 686B, and has ATA100, while the 686A is ATA66.
The pin count is the same, at 352 balls (BGA package). This
chip also includes the Super I/O functions, so this chip might
handle virtually all the power-on functions.

http://web.archive.org/web/20030617203815/http://www.via.com.tw/pdf/productinfo/686a.pdf

The front panel switch should be driving PWRBTN# (see pg. 31 of
the datasheet). The output of the chip is called SUSC# (pg. 25)
and is supposed to connect to PS_ON# on the ATX PSU. SUSC# is
the opposite sense of PS_ON#, and there should be an inverting
logic gate with open collector output, to drive PS_ON#. Thus,
I believe SUSC# is a logic 1 (+3.3V), in order for PS_ON# to
be driven to logic 0 (0 volts). Some external chip would be used
for this function. (There are a number of 14 or 16 pin DIP chips
on your motherboard, but I cannot see what the part numbers are.
It could be that the signal from the Southbridge goes to one of
those smaller chips, before driving the power supply PS_ON#
signal.)

In terms of the device power, the datasheet has this to say:

"There are three power planes inside the VT82C686A. The
first power plane (VCCS) is always on unless turned off by
the mechanical switch. The second power plane (VCC) is
controlled by chip output SUSC# (also called ³PSON²). The
third plane (VCCRTC) is powered by the combination of the
VCCS and the external battery (VBAT) for the integrated real
time clock. Most of the circuitry inside the VT82C686A is
powered by VCC. The amount of logic powered by VCCS is
very small; its main function is to control the supply of
VCC and other power planes. VCCRTC is always on unless both
the mechanical switch and VBAT are removed."

I suppose a failure of one of those signals to make it to
the Southbridge would upset things. VCCRTC is the power source
that comes from the three pin diode ORing of the CMOS battery
and power derived from +5VSB. VCCS would also be derived from
+5VSB, and would likely be needed to make SUSC# work
properly. VCC is what powers the majority of the logic in the
chip. While the datasheet doesn't make it very obvious, the
chip uses mostly 3.3V level power signals.

So, there is a bit of info for your board, but it'll be tough
to guess at how it is all wired together.

Sounds like "cardboard test" time. Perhaps if you pick up a
replacement motherboard, you can do some probing of the A7M266
before giving up on it entirely.

Paul

My Mom's planning on getting a whole new computer when they get their income
taxes done in 3 or 4 months, so likely we'll probably just see if it will
hold for a few more months. Secondly, the power problem seems random as it
only does this once in a while - it doesn't have these power problems all
the time. It's random when it does it. Sometimes it takes weeks to do it,
other times it takes just a few days.

 

[Travis King]

My Mom's planning on getting a whole new computer when they get their
income taxes done in 3 or 4 months, so likely we'll probably just see if
it will hold for a few more months. Secondly, the power problem seems
random as it only does this once in a while - it doesn't have these power
problems all the time. It's random when it does it. Sometimes it takes
weeks to do it, other times it takes just a few days.

I've had experience with the cardboard test before. A person's computer I
built wouldn't turn on about four months after I built it. Swapped the PSU
thinking that would probably be it, nope. Still wouldn't turn on and the
motherboard wouldn't indicate it had power either. So, I put the
motherboard on cardboard, running only minimally and still, nothing. Turns
out it was the motherboard. We tried contacting the company and they would
either put you on hold and you would never get through or they wouldn't
answer the phones. Needless to say, that was the first and last Epox
motherboard that we would ever buy. (Since their support was lousy and
their webpage support was also terrible. They told us through e-mail to
contact them through phone (once we finally found out where to contact them
at), but we couldn't because they would never answer or put us on hold and
wouldn't come back to us.) By the way, we didn't see any bulging capacitors
either that apparently Epox is known for. Epox's website was slower than
molasis also. (especially when downloading anything off of it. It was like
they only had 56k or a low-end DSL.) So, went right online and bought an
Asus, that I should have bought in the first place. The computer's an
additional 5 months old now, and so far so good. (Knock on wood)