Solved PC Wakes immediately after Sleep(S3), but doesn't respond after wake

[automagical]

Hey everyone,

I've got a self-built PC that has been running great since 2010 without issue. However, I recently found that the sleep function no longer performs as expected. The computer sleeps on command but wakes just 3-4 seconds later. Worse, the system is totally unresponsive after waking, with no video to the monitor. I own a backlit USB keyboard (Corsair K90) and the keys weirdly flicker when I try to type in this state. The only way to exit the state is by a hard shutdown. I can't seem to pinpoint the issue since "powercfg lastwake" returns an empty wake history by next boot. I also haven't installed any new hardware since having the problem. I found a troubleshooting guide for sleep issues and followed that, and the only device that is able to wake my computer is the keyboard, according to windows. I'm wondering what to do since I'd love to have my computer always available but leaving it on wastes a ton of power. Thanks for reading my question.


PS. I think this could be related to RAM. Under normal operation (after a fresh boot), the ram status led's are all green, meaning that there are no issues. However, when in the above unresponsive state, I fond that the same status led's are orange, a warning indicator. I am using S3 sleep, so the session is always saved to RAM. Just what I know, hopefully it's a clue to the solution. Thanks again.

 

[Carl Lawrence]

Have you tried system restore?

 

[Layback Bear]

How often do you do a normal proper shutdown or reboot?

 

[automagical]

I haven't done a system restore yet, but I will if nothing else works.
I just thought that maybe the problem could be resolved easily through software.


Also, I do shut off and power cycle my PC quite frequently, especially in the recent weeks at about once every other day.

Thanks for your quick replies.

 

[automagical]

I performed a system restore, but nothing happened. In fact, the system cannot even shutdown now (though I don't think that was the the fault of a restore). It simply turns back on and enters the same unresponsive state. I haven't got a clue what's causing it. Thanks for the help.

 

[F5ing]

You may want to check out Event Viewer to see what's going on around that time.

 

[Layback Bear]

Do a hold the power button down restart. May be after doing that you can do a normal start.
How much ram do you have?

 

[automagical]

I did check event viewer but I didn't find any explanation for the system turning back on. I'll keep checking it, though. Also, I tried a reset (via the reset button on my case) but the problem persisted. I have 6GB of Corsair RAM installed, and it hasn't thrown any errors during stress testing with IntelBurnTest. I'll try running the Windows memory diagnostic as well. Thanks.

 

[automagical]

Nothing at all seems to work. The only way to turn off the system after a failed sleep or shutdown is to either hold down the power button or open the PSU switch. Miraculously, Windows resumes as if nothing happened even after the switch has been cycled. Still, there seems to be no way to naturally turn off the system from Windows. This is becoming very confusing.

 

[F5ing]

Might try these:

http://www.sevenforums.com/tutorials/1538-sfc-scannow-command-system-file-checker.html

http://www.sevenforums.com/tutorials/105647-ram-test-memtest86.html (better mem test)

 

[Layback Bear]

When you do a forced shutdown go into bios and see if anything looks strange. Any thing that indicates auto restart. In Windows 7 stop all sleep and hibernate options. Do you have another power supply you can try?

 

[automagical]

/scannow recognized some corrupt files but I already knew that from a previous checkup, before the current problems. Still, do you think restoring or repairing windows could be the solution?

Also, I've checked the BIOS and even loaded the fail-safe defaults but nothing changes.

But if it is a windows issue, shouldn't system restore always be able to resolve it?

 

[F5ing]

/scannow recognized some corrupt files but I already knew that from a previous checkup, before the current problems. Still, do you think restoring or repairing windows could be the solution?


Also, I've checked the BIOS and even loaded the fail-safe defaults but nothing changes.


But if it is a windows issue, shouldn't system restore always be able to resolve it?

Maybe one of those corrupt files has something to do with it.

System Restore only takes you to the condition the system was in at the date/time the restore point was created. If it was broken at that time it'll be broken when restored.

You might try this instead: http://www.sevenforums.com/tutorials/3413-repair-install.html

At least it'll fix those corrupt files.

 

[automagical]

I'll look for my install disc then.

However, I'm not certain that windows is still the problem. There's a ~5 second or so delay between the system powering down and automatically restarting. During this delay, I find that if I turn off the PSU, the system will still attempt to turn on; all fans and lights will activate for a brief moment. This is highly curious. I think this could be a fault of either the PSU or motherboard. Another thing I noticed was that the case speaker does not beep (normally indicating a successful POST, or potentially an error) at all when the system self-restarts. It's as if I'm running a headless system with power supplied to all devices without any result. The motherboard doesn't seem to recognize the boot sequence. This makes me lean closer to the faulty PSU theory. I'll be receiving a new one on Tuesday, so I guess I'll find out then. I apologize for posting hardware issues in a software forum, but I really could use some help in determining the source of the problem.

 

[F5ing]

Can you get into the BIOS?

Did you try running the RAM tests?

 

[automagical]

RAM tests were clean, I let memtest run for about 10 hours without issues. I can normally get into BIOS, its just that when I try to make the system sleep, it restarts but won't post at all.

 

[F5ing]

It does seem likely that it could be hardware. Until you get your new PSU, you might want to check a few things in your BIOS. Appears you have a pretty versatile motherboard.

I glanced at the manual for your board. It looks like you have the ability to check out quite a few system voltages. There are also controls/options that affect booting and POST:

GIGABYTE - Motherboard - Socket 1366 - GA-X58A-UD3R (rev. 2.0)

In your OP you state that you see the green LEDs for your memory and that you indicate that it's normal. But I found the following in the manual:

DDR Voltage
Off: Normal condition
L1: Level 1 (Slight, green)
L2: Level 2 (Moderate, yellow)
L3: Level 3 (High, red)

Not exactly sure what they mean by "slight" (green). Based on the other conditions stated I'm guessing it means "slightly above normal". They do state that at normal conditions the LEDs should be off.

If you are overclocking you probably should stop until this is resolved. It looks like you can save booting profiles and take your pick of which one it'll use. What I would do is save the current profile and restore the default to see if it solves the issue. If I wasn't sure how well that profile saving feature worked (and I'm not sure) I would take note of the non-default settings or take pictures of each screen before reverting to default settings.

 

[automagical]

Thanks for the reply. That's correct, I am using 1.64V for my memory as per manufacturer spec, even though Intel spec says 1.5V. That causes one of the 5 RAM LEDs to light green (this is all under normal operation.) However, when the system self-restarts, all the LEDs are lit, either orange or green. This persists even when I load Fail-Safe defaults, which pulls all voltages, settings and frequencies back within spec. For all I know, however, the lighting of the LEDs could just be a part of the normal startup sequence. Besides, a memory error should be represented by beeps from the case speaker.

Another point that may be worth a mention is that after the system restore, I noticed that my system clock had been reset. This is a sign of a depleted CMOS battery, right? I've also heard that a discharged battery can be the culprit of many other strange symptoms. I'll give it a try.

 

[F5ing]

Thanks for the reply. That's correct, I am using 1.64V for my memory as per manufacturer spec, even though Intel spec says 1.5V. That causes one of the 5 RAM LEDs to light green (this is all under normal operation.) However, when the system self-restarts, all the LEDs are lit, either orange or green. This persists even when I load Fail-Safe defaults, which pulls all voltages, settings and frequencies back within spec. For all I know, however, the lighting of the LEDs could just be a part of the normal startup sequence. Besides, a memory error should be represented by beeps from the case speaker.

Yeah, you would at least have to have made some previous observations of those LEDs during boot to have learned what was normal.

Another point that may be worth a mention is that after the system restore, I noticed that my system clock had been reset. This is a sign of a depleted CMOS battery, right? I've also heard that a discharged battery can be the culprit of many other strange symptoms. I'll give it a try.

A weak/dead CMOS battery can definitely cause some odd issues. Those little batteries seem to be so much more critical these days as compared to years ago.

 

[automagical]

Thanks for the help everyone. It turns out that it was just the CMOS battery, and replacing it fixed everything. I hope this helps someone else in the future.