Unexpected Shutdown

[Shoe]

Several times in the last few days, my computer has shut down while in
sleep mode. Windows 7 Home Premium 64 bit. When I start after this
happens, I get the black screen saying that windows had an unexpected
shutdown and asking if I want to start windows normally, in safe mode,
or what. So far, this has not happened while I am using the computer.
This is a machine that I built myself and has been running for a
couple of years with no problem. I have no idea how to troubleshoot
this problem, whether it is a software issue or hardware. Any thoughts
on how to go about resolving this?

 

[BillW50]

Several times in the last few days, my computer has shut down while in
sleep mode. Windows 7 Home Premium 64 bit. When I start after this
happens, I get the black screen saying that windows had an unexpected
shutdown and asking if I want to start windows normally, in safe mode,
or what. So far, this has not happened while I am using the computer.
This is a machine that I built myself and has been running for a
couple of years with no problem. I have no idea how to troubleshoot
this problem, whether it is a software issue or hardware. Any thoughts
on how to go about resolving this?

Does the event log tell you anything? You know, my first thought is an
overheating problem. You might ask, overheating? It's in standby! Yes
standby almost always shuts down the fans. And the heatsinks on the RAM
or CPU isn't large enough. Although it could be something else of course.

 

[BillW50]

Does the event log tell you anything? You know, my first thought is an
overheating problem. You might ask, overheating? It's in standby! Yes
standby almost always shuts down the fans. And the heatsinks on the RAM
or CPU isn't large enough. Although it could be something else of course.

Or I should add dirty heatsinks.

 

[charlie]

Several times in the last few days, my computer has shut down while in
sleep mode. Windows 7 Home Premium 64 bit. When I start after this
happens, I get the black screen saying that windows had an unexpected
shutdown and asking if I want to start windows normally, in safe mode,
or what. So far, this has not happened while I am using the computer.
This is a machine that I built myself and has been running for a
couple of years with no problem. I have no idea how to troubleshoot
this problem, whether it is a software issue or hardware. Any thoughts
on how to go about resolving this?

Is the computer connected directly to A/C, or does it have a UPS between
it and the wall plug?
(Power glitches?)

A few computers can implement a full shutdown after being in standby or
hibernate after a user settable time.)
To get this sort of thing to work, BIOS and the windows power options
have to be set properly.

It sounds like your shutdown may be AC power related. (A more than
uncommon problem in some areas.)

 

[Paul]

Or I should add dirty heatsinks.

Where the heat would be, is inside the power supply.

The +5VSB isn't very efficient. And there's a bit of
heat involved, and the PSU fan does not spin.

It all depends, on whether the OP noticed the +5VSB glitching
or not running, or whether the power was stable, and
a crash was only observed on recovery from sleep (i.e.
bad RAM, unable to keep data while running in self-refresh
mode).

On Asus motherboards, the single LED on the motherboard
is likely tied directly to +5VSB. And can sometimes be
used to detect a problem with +5VSB. Like, if you see
the LED "oscillate" when the computer is sleeping or
hibernating, or when it's trying to boot or whatever.
The +5VSB should be stable whenever the PSU is switched
on at the back, and there shouldn't be any blinking
of such LEDs.

My Asrock motherboard, has no LED for that, and it's
a nuisance to come up with another monitoring solution
to take the place of that LED.

The sleep LED on some computer cases, on the front,
typically blinks at a 1 Hertz rate, and that too
isn't all that useful for detecting a +5VSB problem.
The LED that operates with steady light output,
on the motherboard surface, is the best monitoring
method, short of dragging out the multimeter or
oscilloscope or data logging instrument.

If the computer is used to recharge portable
devices (iPad, iPod, phone), then disconnect
all those loads and retest the sleep function
with a bit less DC loading.

Paul

 

[VanguardLH]

Several times in the last few days, my computer has shut down while in
sleep mode. Windows 7 Home Premium 64 bit. When I start after this
happens, I get the black screen saying that windows had an unexpected
shutdown and asking if I want to start windows normally, in safe mode,
or what. So far, this has not happened while I am using the computer.
This is a machine that I built myself and has been running for a
couple of years with no problem. I have no idea how to troubleshoot
this problem, whether it is a software issue or hardware. Any thoughts
on how to go about resolving this?

Have you tried configuring the power options so the computer always
remains powered up (i.e., NEVER goes into any low-power mode) to see if
when you leave it for long times if the awake problem crops up? If that
test passes then go back to your old power scheme and reboot Windows
with no startup items. Use msconfig.exe to do a selective startup.
That disables all the startup items (but not any services). Again test
for awhile to see if the problem crops up again or not. If not, you've
got software loading on startup causing the problem later when your
computer entered low-power mode.

 

[Rob]

Have you tried configuring the power options so the computer always
remains powered up (i.e., NEVER goes into any low-power mode) to see if
when you leave it for long times if the awake problem crops up? If that
test passes then go back to your old power scheme and reboot Windows
with no startup items. Use msconfig.exe to do a selective startup.
That disables all the startup items (but not any services). Again test
for awhile to see if the problem crops up again or not. If not, you've
got software loading on startup causing the problem later when your
computer entered low-power mode.

Would not suprise me that the motherboard is buggered.

 

[Paul]

Would not suprise me that the motherboard is buggered.

It's possible, but in terms of probabilities, power supply
ranks pretty high up the chart, in terms of being the
faulty bit.

I've had two power supplies fail here at home, and not cheap ones.
While on motherboards, they've been perfect.

Bad RAM could do it, but then, you might have other signs
the RAM is bad, such as lots of crashes or errors while
the computer is running.

When a machine sleeps, the RAM sticks are put in self-refresh
as far as I know. The +5VSB from the power supply, continues
to be converted into low voltage (1.8V or 1.5V, DDR2 or DDR3)
to power the RAM. And I think it might still need a clock signal
as well. On recovery, the BIOS has to be smart enough not
to reset the RAM sticks by accident. As long as the controller
is not reprogrammed, the RAM contents stay intact.

More modern OSes, have various "hybrid" options, where not
only is stuff kept in RAM, but it can also be held on disk
just in case. In those cases, you can survive a RAM failure,
and still have your session preserved. The penalty, is slower
setup and recovery, depending on the scheme. Suspending to
RAM without any hybrid feature, should be pretty fast, but
a lot more dangerous.

I have one computer here, that about one in four times,
it doesn't survive coming out of S3 sleep. For that machine,
I use S4 Hibernate, and it's all good. I don't worry about it.
It's probably a design issue, and has been there since the
first day I got it. It's not worth fixing. And I've done
enough RAM swaps, to know the RAM is fine. The RAM has been
used in this machine before, and no issues.

Paul

 

[Rob]

It's possible, but in terms of probabilities, power supply
ranks pretty high up the chart, in terms of being the
faulty bit.

I've had two power supplies fail here at home, and not cheap ones.
While on motherboards, they've been perfect.

Bad RAM could do it, but then, you might have other signs
the RAM is bad, such as lots of crashes or errors while
the computer is running.

When a machine sleeps, the RAM sticks are put in self-refresh
as far as I know. The +5VSB from the power supply, continues
to be converted into low voltage (1.8V or 1.5V, DDR2 or DDR3)
to power the RAM. And I think it might still need a clock signal
as well. On recovery, the BIOS has to be smart enough not
to reset the RAM sticks by accident. As long as the controller
is not reprogrammed, the RAM contents stay intact.

More modern OSes, have various "hybrid" options, where not
only is stuff kept in RAM, but it can also be held on disk
just in case. In those cases, you can survive a RAM failure,
and still have your session preserved. The penalty, is slower
setup and recovery, depending on the scheme. Suspending to
RAM without any hybrid feature, should be pretty fast, but
a lot more dangerous.

I have one computer here, that about one in four times,
it doesn't survive coming out of S3 sleep. For that machine,
I use S4 Hibernate, and it's all good. I don't worry about it.
It's probably a design issue, and has been there since the
first day I got it. It's not worth fixing. And I've done
enough RAM swaps, to know the RAM is fine. The RAM has been
used in this machine before, and no issues.

Paul

There was a problem with this happening and a fixit from the MS site
could help. Do know what it is but its there.

I have had this problem of recent times the problem being a MB. Depends
on how it boots to see if there are problems with ram. Which you could
check with memtest 86.

 

[Paul]

It's possible, but in terms of probabilities, power supply
ranks pretty high up the chart, in terms of being the
faulty bit.

I've had two power supplies fail here at home, and not cheap ones.
While on motherboards, they've been perfect.

Bad RAM could do it, but then, you might have other signs
the RAM is bad, such as lots of crashes or errors while
the computer is running.

When a machine sleeps, the RAM sticks are put in self-refresh
as far as I know. The +5VSB from the power supply, continues
to be converted into low voltage (1.8V or 1.5V, DDR2 or DDR3)
to power the RAM. And I think it might still need a clock signal
as well. On recovery, the BIOS has to be smart enough not
to reset the RAM sticks by accident. As long as the controller
is not reprogrammed, the RAM contents stay intact.

More modern OSes, have various "hybrid" options, where not
only is stuff kept in RAM, but it can also be held on disk
just in case. In those cases, you can survive a RAM failure,
and still have your session preserved. The penalty, is slower
setup and recovery, depending on the scheme. Suspending to
RAM without any hybrid feature, should be pretty fast, but
a lot more dangerous.

I have one computer here, that about one in four times,
it doesn't survive coming out of S3 sleep. For that machine,
I use S4 Hibernate, and it's all good. I don't worry about it.
It's probably a design issue, and has been there since the
first day I got it. It's not worth fixing. And I've done
enough RAM swaps, to know the RAM is fine. The RAM has been
used in this machine before, and no issues.

Paul

I think I messed up that answer.

1) If a computer shuts down while sleeping, you have a
problem with +5VSB. Either the power supply is weak on
+5VSB, or there are too many loads on +5VSB. For example,
a person charging an excessive number of USB mobile devices,
could exceed the rating of +5VSB. Modern motherboards
run all the USB ports from +5VSB (for better or worse).

2) Other kinds of problems with sleep, would have to do
with the RAM contents getting corrupted. For example,
bad RAM might cause a crash when the computer comes
back. But that would not be the source of shutdowns.
Bad RAM or a bad motherboard design, could cause an
inability to successfully complete a sleep cycle.
I have a motherboard here, that cannot complete a sleep
cycle properly, but it doesn't happen every time.
Presumably the contents of RAM are corrupted.

HTH,
Paul

 

[Shoe]

I think I messed up that answer.

1) If a computer shuts down while sleeping, you have a
problem with +5VSB. Either the power supply is weak on
+5VSB, or there are too many loads on +5VSB. For example,
a person charging an excessive number of USB mobile devices,
could exceed the rating of +5VSB. Modern motherboards
run all the USB ports from +5VSB (for better or worse).

2) Other kinds of problems with sleep, would have to do
with the RAM contents getting corrupted. For example,
bad RAM might cause a crash when the computer comes
back. But that would not be the source of shutdowns.
Bad RAM or a bad motherboard design, could cause an
inability to successfully complete a sleep cycle.
I have a motherboard here, that cannot complete a sleep
cycle properly, but it doesn't happen every time.
Presumably the contents of RAM are corrupted.

HTH,
Paul

Thanks for all the responses - I will work through them to try to
resolve this. I do have a UPS between the computer and the AC line,
so power failure should not be a factor.
John

 

[Paul]

Thanks for all the responses - I will work through them to try to
resolve this. I do have a UPS between the computer and the AC line,
so power failure should not be a factor.
John

In (1), I refer to the behavior of the ATX power supply.

Check the label on the side, and it will say "5VSB @ 2 amps".
That tells you it supports 2 amps of current flow, 10 watts
total. That's all the power that is available during Sleep (S3).

Some +5VSB regulators inside power supplies, they're about 50%
efficient, so about 10 watts of waste heat is produced. The
heatsinks are big enough in the power supply, to handle this,
without the fans running.

The motherboard draws 5VSB @ 1 amp, leaving about 1 amp more
to charge Ipod or Ipad or whatever. Or run your USB desk lamp.

When a supply gets old, it is possible for the 5VSB to get
"tired" and trip out at lower current flow values. That would
lead to a shut off while in sleep.

The UPS, removed the power company as the potential culprit.

Paul

 

[Char Jackson]

Shoe wrote:

The UPS, removed the power company as the potential culprit.

A malfunctioning UPS might not remove the power company as a potential
culprit, and may even introduce a few unknowns of its own. As a
troubleshooting step, I sometimes remove a UPS from the equation.

 

[Rob]

In (1), I refer to the behavior of the ATX power supply.

Check the label on the side, and it will say "5VSB @ 2 amps".
That tells you it supports 2 amps of current flow, 10 watts
total. That's all the power that is available during Sleep (S3).

Some +5VSB regulators inside power supplies, they're about 50%
efficient, so about 10 watts of waste heat is produced. The
heatsinks are big enough in the power supply, to handle this,
without the fans running.

The motherboard draws 5VSB @ 1 amp, leaving about 1 amp more
to charge Ipod or Ipad or whatever. Or run your USB desk lamp.

When a supply gets old, it is possible for the 5VSB to get
"tired" and trip out at lower current flow values. That would
lead to a shut off while in sleep.

The UPS, removed the power company as the potential culprit.

Paul

Step one is to check to see if the fan is running to keep it cool.

Then you can pull the power supply out. Take the screws out and remove
the cover.

The biggest problem is the capacitors and the large ones, where all the
wires come out, should be nice and flat across the tops if not a new PSU
is required. Also make sure that its clean inside (and Fan) not full of
dust.

 

[Steve Hayes]

Thanks for all the responses - I will work through them to try to
resolve this. I do have a UPS between the computer and the AC line,
so power failure should not be a factor.
John

I just discussed a similar problem in the XP forum.

Eventually I found that the problem was a faulty multiplug adapter. See here:

http://ondermynende.wordpress.com/2012/12/02/wired/