Solved Random crash / shutdown when sleeping

[JohnHa]

Random crash? shutdown? power off? when sleeping

My HP Pavilion W7 Home Edition 64 bit desktop suddenly started to power off at random when sleeping - it has happened about 20 times in the past two weeks. I am almost certain that it does not happen if I stop it from sleeping. I return to the PC and find it has powered off. When I power the PC on I get the "Windows did not shut down properly ..." screen.

I reseated the memory and system board plugs; cleaned the dust from the fans and heatsinks; used the HP drivers update utility to update my drivers (which wasn't a good idea as it gave me drivers for things I don't have); used Device Manager to check each and every driver and updated only those which required updating; but none of these fixed it. I then restored to a previous restore point before the shutdowns started, but that did not fix it either (and may have un-done some of the driver updates). I thought it best to ask for advice so I gathered the data requested - see attached debug-Lounge-PC-xxxx.ZIP file. The machine is standard without modifications and Windows is up to date. chkdsk /f and sfc /scannow ran OK.

All help will be gratefully appreciated.

 

[ICIT2LOL]

Hello and welcome John mat soiunds a tad like a power sipply problem. Try this

Using HW Info

You can test the volts on the PSU with HWInfo HWiNFO, HWiNFO32/64 -Download < download the right bit version andclose the right hand window select Sensors and scroll down to the power sectionwhere you will see what the volts are doing see my pic.

The original right hand window shows themachine running and is handy for that but for looking at the components in some detail close it and use themain left hand side panel

FOR OTHER COMPONENTS

Open each + down into the componentitself and then click on it - in theright hand side will appear all sorts of details including brands speeds andother essential info that device. See pic for example.

 

[derekimo]

Hi There,

It looks like your graphics driver,

Unable to load image \SystemRoot\system32\DRIVERS\igdkmd64.sys, Win32 error 0n2
*** WARNING: Unable to verify timestamp for igdkmd64.sys
*** ERROR: Module load completed but symbols could not be loaded for igdkmd64.sys
 igdkmd64+0x18c8c9

It's pretty old too,

 Loaded symbol image file: igdkmd64.sys
    Image path: \SystemRoot\system32\DRIVERS\igdkmd64.sys
    Image name: igdkmd64.sys
    Timestamp:       [COLOR="Red"] Fri Oct 15 22:28:37 2010[/COLOR] (4CB93805)
    CheckSum:         00A2BBB0
    ImageSize:        00A209A0
    Translations:     0000.04b0 0000.04e4 0409.04b0 0409.04e4

I don't see one from HP available so you can try this one,

https://downloadcenter.intel.com/De...sion=Windows 7 (64-bit)*&DownloadType=Drivers

It's newer than yours. That may fix it.

I notice you have Cobian backup too, the bugcheck for most all of your dumps has a possible cause of incompatible backup software.

If that driver doesn't fix it I would try uninstalling Cobian.

Let us know how things go.

 

[JohnHa]

Thank you for your responses - I am very grateful.

I first looked at the voltages, and ran HWInfo, getting the ambiguous results below. I couldn't see values for 5V or 12V, and I couldn't believe that what should be 3V was only 1.6V, so I opened up the case and measured the actual voltages at the PSU plug on the motherboard, getting:

Black - Orange (3.3V) 3.36V measured
Black - Red (5V) 5.03V measured
Black - Yellow (+12V) +11.80 measured
Black - Blue (-12V) -11.90V measured

I then thought that a faulty PSU must be likely as I am pretty certain that I only get the "power down" when the PC is sleeping, which presumably means it isn't running software. On second thoughts, I suppose it could be "powering off" when it is asked to come out of sleep mode, so a faulty driver could be my problem.

A new 300W ATX PSU isn't expensive so I will order one even though the measured voltages looked OK when the PC was running.

It seems strange to me that it started suddenly on 8 June so I checked what software and updates I installed around then. On 8 June I installed Audacity and updated Dropbox to 2.8.2. No W7 updates were installed around then. I have therefore uninstalled Audacity (with difficulty - I needed an uninstaller as W7 refused). I also checked Dropbox and see that 2.8.3 is now available so I installed that - it seems strange that it should come out so quickly afer my updating to 2.8.2.

I then turned to the Graphics Driver. The HP site offers me

Released: 2009-10-14
File name: sp43162.exe [1/1, 27.14M]
Version: 8.15.10.1851
Compatibility: Microsoft Windows 7 (64-bit)

I used the Intel site to scan my PC and it was adamant that my driver was supplied by the vendor and I should only get an update from the vendor, and not Intel.

A customized computer manufacturer driver is installed on your computer. The Intel Driver Update Utility is not able to update the driver. Installing a generic Intel driver instead of the customized computer manufacturer driver may cause technical issues. Contact your computer manufacturer for the latest driver for your computer.

I therefore thought it best first to try restoring the previous version, so I did so, getting as below:
Manuf: Intel Corporation
Date: 15/10/2010
Ver: 8.15.10.2226.

I am confused because it looks the same as the one in the reports I uploaded, so perhaps it did not roll back. However it appears to be a later version (.2226 compared with .1851) that that offered by the HP site (see above).

I then downloaded the version derekimo suggested thinking I could always roll back if there were any problems, but it gave an error when I ran setup.exe saying it was not validated for this computer, and it exited.

NB I searched through the files I uploaded, including the dmp files with Blue Screen View, but I could not find the location of the error message derekimo quoted - in which file is it?

I will also temporarily disable Cobian, though I have used it for many years without problems. I use VSS with it so that I can backup "live, open" files.

I will keep posting with my experiences - it takes a bit of time to have the fault happen.

 

[Layback Bear]

JohnHa I noticed your vBAT 3V is way to low. It is the battery on your motherboard that allows the motherboard to remember the cmos when the computer is not running.
Replacing that battery might not solve your problem but it should be replaced and the cmos set to defauld. You will also have to set the time and date in the bios.

 

[ICIT2LOL]

I agree with Layback Bear and 300watts is not really that large you must take into account it is the +12v rail that is the probablt the most important and the 300 watts is made up of the total;s of the other rails. So just for example the +12v rail might only be delivering say 14 amps which equate to 140 watts - this is an example only.

I think if you can you could use this to check what wattage PSU you need and then it will give you the least and a recommneded size PSU. eXtreme Power Supply Calculator

Now to find the components I find this is the best thing to use . HWiNFO, HWiNFO32/64 - Download


Using HW Info

You can test the volts on the PSU with HWInfo HWiNFO, HWiNFO32/64 -Download < download the right bit version andclose the right hand window select Sensors and scroll down to the power sectionwhere you will see what the volts are doing see my pic.

The original right hand window shows themachine running and is handy for that but for looking at the components in some detail close it and use themain left hand side panel

FOR OTHER COMPONENTS

Open each + down into the componentitself and then click on it - in theright hand side will appear all sorts of details including brands speeds andother essential info that device. See pic for example.

 

[derekimo]

I then turned to the Graphics Driver. The HP site offers me

Released: 2009-10-14
File name: sp43162.exe [1/1, 27.14M]
Version: 8.15.10.1851
Compatibility: Microsoft Windows 7 (64-bit)

I used the Intel site to scan my PC and it was adamant that my driver was supplied by the vendor and I should only get an update from the vendor, and not Intel.

A customized computer manufacturer driver is installed on your computer. The Intel Driver Update Utility is not able to update the driver. Installing a generic Intel driver instead of the customized computer manufacturer driver may cause technical issues. Contact your computer manufacturer for the latest driver for your computer.

I therefore thought it best first to try restoring the previous version, so I did so, getting as below:
Manuf: Intel Corporation
Date: 15/10/2010
Ver: 8.15.10.2226.

I am confused because it looks the same as the one in the reports I uploaded, so perhaps it did not roll back. However it appears to be a later version (.2226 compared with .1851) that that offered by the HP site (see above).

I then downloaded the version derekimo suggested thinking I could always roll back if there were any problems, but it gave an error when I ran setup.exe saying it was not validated for this computer, and it exited.

NB I searched through the files I uploaded, including the dmp files with Blue Screen View, but I could not find the location of the error message derekimo quoted - in which file is it?

I will also temporarily disable Cobian, though I have used it for many years without problems. I use VSS with it so that I can backup "live, open" files.

I will keep posting with my experiences - it takes a bit of time to have the fault happen.

Yeah, that version from HP is right about the time Windows 7 was released, so that is an original driver.

Sometimes you can use the Intel site, obviously this wasn't one of them. Sorry about the run around.

As for why the driver rolled back to a newer version than what you found at HP, I'm guessing at some point it was updated since you got the computer or HP decided to just keep the bare minimum original drivers available for download.

For the detailed output I posted, which you couldn't find in Blue Screen View because it is a basic tool and only shows the Bug Check and a probably caused by, which is almost always not the actual cause.

I use Windbg to get those outputs with a few commands from the .dmp files.

You are getting some great advice from the others as well, if you do get any more BSOD's I'll be happy to look those over for you.

Very good feedback from you. Let us know how things go.

 

[JohnHa]

Thank you for your responses.

I removed the BIOS battery and measured it at 3.2V. For some strange reason, HWiNFO 64 (and the Beta) is only reporting my voltages as half what they really are.

I have since done the following and unfortunately it is still happening although it ran for several hours in sleep without a problem (I hate intermittent problems!):

1 Disabled Cobian and the VSS - the power downs still l happened. I have still got them disabled.

2 Using HWiNFO to tell me the motherboard, I downloaded the Evans BIOS firmware and installed it. I had taken the BIOS update from the HP Pavilion p6203uk site, but was not happy about it.

3 I noticed that the PC was in sleep mode with its sleep-light ON. I hit Enter to wake it up and it immediately powered down. So, sometimes it powers down when in sleep; but sometimes it powers down when I hit Enter to wake it.

4 I created a Ubuntu bootable USB stick using the instructions at How to create a bootable USB stick on Ubuntu | Ubuntu and ran it with Ubuntu in sleep mode for several hours. IT didn't fail. The idea was to separate out the hardware - if it is the PSU, it should fail under Ubuntu as well. I need to repeat this test leaving it longer.

5 I have got fed up waiting for it to reboot each time so I have bought a 240GB SSD SATA disk and I will install the W7 system on it when it arrives. I have always wanted a separate C: drive, with my data on D: and this will be ideal. I am looking forward to see if it is as good as everyone says.

6 PSU. My HP ATX PSU is a 300W PSU, but the label says +5V @ 25A and +12 @ 19A. I have an older 300W ATX, but it is rated at only 5v @ 20A and +12V @ 13A. I am tempted to try it because the eXtreme Power Supply Calculator Lite said I need only 155W and should use a 200W PSU. I also tried the PSU Watts And Rails Calculator at PSU Watts and Rails Calculator which said I needed 171W, and broke it down as 3.3V@2A, [email protected] and +12V@11A. So my old ATX PSU should be OK to try, especially if I don't power the CD/DVD drive and remove the Audio and Wireless PCI cards. I am also trying friends to borrow a PSU from a discarded machine.

6 Following one of the power downs I got the Restore page - I forget how - and I? it? chose to restore me to what it thought was the latest good system. It had obviously gone back because I had to un-install some software I had been cleaning out.

Proposed course of action from here:

1 Try my spare ATX PSU. It is current limited so it should not damage anything, nor itself.

2 Leave it in Ubuntu sleep mode overnight.

3 When the SSD disk arrives in 2-3 days time, re-install W7 on the SSD and see if that fixes it.

I will keep posting and if I find what is causing the problem, or find it goes away, I will report back.

 

[JohnHa]

derekimo

You may have be on the right track with your observation about the graphics driver failing to load.

The HP download site has the Intel G41 Express driver version 8.15.10.1851 as the one for my PC. For some reason I have 8.15.10.226 installed (I recollect upgrading it a year or two ago) which Free Driver Download for Intel(R) G41 Express Chipset - Pegatron Pegatron says is for an ASUS PC. I rolled back the driver finding yet another version (I did not record it) and installed the 8.15.10.1851 from the HP site. It is too early to say if it has fixed my problem.

I have also discovered Sleeper from PassMark Sleeper - Hibernate and sleep state testing software which is a magic free utility which allows you to cycle the PC in and out of the various sleep modes while it records everything which happens.

 

[derekimo]

Well, when the dump files point out a driver failing to load, I can't really take the credit.

Sounds like you are doing some good troubleshooting, keep us posted with your findings.

We try not to recommend drivers from any place except the OEM or Manufacturer, sometimes there is no alternative but it is rare.

 

[JohnHa]

An update ...

Graphic driver: Windows update offered me a new Intel G41 Express driver. I couldn't see which level it was but decided not to take it.

PSU: I swapped out the PSU and replaced it with the spare PSU, which seems to power the PC OK. However, the problem persists with the swapped PSU. I removed the CD/DVD drive, the Soundblaster card and the Wireless card to minimise the PSU current load.

PassMark Sleeper: This seems to be very good. Help says:

Power State ... Description

Working ... The system is fully on. Some devices may independently conserve power if there usage falls below a certain threshold.

Sleeping ... The system seems as though it’s shut down. Power consumption is reduced to one of the three sleep levels (see below). The lower the level, the more power that is conserved, but the longer it takes for the system to wake. Mouse movement or key presses will usually wake the system.

Soft Off or Hibernate ... The system appears to be off. Power consumption is very low. The operating system context is saved out to disk and no longer exists in RAM. Wake up usually results in hardware boot, but not software boot.

Mechanical Off ... No power consumption. A full reboot will be required to start up the system.

ACPI (Advanced Configuration and Power Interface) is an industry wide standard that defines a power management and configuration mechanism for hardware and operating systems. ACPI defines six discrete power states. Lower states consume more power, but have a smaller latency on wake up.

S0 – Working
S1 – Sleeping
S2 – Sleeping
S3 – Sleeping
S4 – Hibernate
S5 – Off

Sleeper allows the user to put the system into sleep state S1, S2, S3 or S4. To turn off or reboot the computer, see the Rebooter application also available from PassMark.

Note that not all computer systems will support all power states. It is not unusual for just S1 and S4 to be supported or S1, S3 and S4. Attempting to put a system into an unsupported state results in the next lowest supported state being chosen.

It tells me that my PC supports S1, S3 and S4, so I set up a cycle of: sleep S1 for 30 sec > wake for 60 sec > sleep S3 for 30 sec > wake for 60 sec > sleep for S1 for 30 sec > ... etc. Sometimes it runs several times, but it always seems to fail while sleeping in the S3 (deepest) sleep state - see output below where it failed on the first S3 state.

I read elsewhere that the deepest sleep state leaves only the memory powered (lighter sleeps leave other stuff powered). I have 2 x 2GB memory cards = 4GB total.

EDIT. I now think that statement is wrong. My desktop supports Hybrid sleep as the default sleep mode, so I think these states are: (see http://www.tekrevue.com/tip/comparing-sleep-hibernation-and-hybrid-sleep-on-windows/)

S1 Sleep as in the old days. The memory is powered up (possibly at low voltage because it is not being changed, it just needs to retain the data), but everything else is powered down. The problem is that a power cut will lose what is in memory

S3 Hybrid sleep. User data in memory is first written to disk, and the PC is then put to sleep as in S1. If there is no power cut while sleeping, then it comes out of sleep exactly like S1.

But if there was a power cut, then user data and system in memory are lost, but user data is still available on disk and can be recovered.
​

This is the output I got from Sleeper for the last failure with the swapped PSU:

2014-Jun-25 12:11:11 **********************************************************************
2014-Jun-25 12:11:11 Sleeper V2.3 (Build: 1009), built on Oct 15 2012 10:38:25
2014-Jun-25 12:11:11 Attempting to load powrprof.dll
2014-Jun-25 12:11:11 **********************************************************************
2014-Jun-25 12:11:11 GlobalPowerPolicy.user.GlobalFlags = 0x12
2014-Jun-25 12:11:12 WAKE UP FROM S1, CYCLE 1, Sleeper restart, possible power cycle
2014-Jun-25 12:11:12 SLEEP TIME TOO LONG WARNING, Sleep time 1403694672sec

2014-Jun-25 12:11:26 Going to sleep, S1, Unforced, Duration 30sec
2014-Jun-25 12:11:26 Setting wake time at Wed Jun 25 12:11:56 2014
2014-Jun-25 12:12:14 Waitable timer triggered.
2014-Jun-25 12:12:14 Wake up from S1, Cycle 1
2014-Jun-25 12:12:14 Sleep time too long warning, Sleep time 48sec
2014-Jun-25 12:12:14 Successfully left sleep state S1...

2014-Jun-25 12:12:14 About to enter sleep state S3...
2014-Jun-25 12:13:16 Going to sleep, S3, Unforced, Duration 30sec
2014-Jun-25 12:13:16 Setting wake time at Wed Jun 25 12:13:46 2014

... and it powered itself off. I will continue checking to see if it powers off immediately, or when it gets the wake up signal - I expect the latter.

I have attached a new set of debug files - they are debug-Lounge-PC-0002.zip and I would greatly appreciate any hints they suggest. I left all the original dmp files in so it includes those I posted before. Hopefully the drivers bits might reveal something new.

 

[ICIT2LOL]

John may I ask what may be a daft question and that is why do you use sleep mode instead of powering off when not using the machine?.

One thing you did mention to was that one should not power up any PSU without and external load even if it is only a fan.

You could if you like test the POWER_GOOD signal at pin 8 on the 24 pin socket as this is the line that keeps the PSU going see Power Supply 101: A Reference Of Specifications - Power Supplies You will see the volts that it has to be and please use a digital multimeter a NOT and analogue one - they inject volts to work

This might be handy if I have not posted it before
http://www.sevenforums.com/tutorials/301799-psu-test-dc-output-voltage.html

 

[derekimo]

Sorry, I've been busy.

Of the 4 dumps from the 25th, two are pointing to the graphics driver,

Unable to load image \SystemRoot\system32\DRIVERS\igdkmd64.sys, Win32 error 0n2
*** WARNING: Unable to verify timestamp for igdkmd64.sys
*** ERROR: Module load completed but symbols could not be loaded for igdkmd64.sys
 igdkmd64+0x6250b

Loaded symbol image file: igdkmd64.sys
    Image path: \SystemRoot\system32\DRIVERS\igdkmd64.sys
    Image name: igdkmd64.sys
    Timestamp:        [COLOR="Red"]Tue Jul 21 15:32:45 2009[/COLOR] (4A66420D)
    CheckSum:         00710035
    ImageSize:        007014C0
    Translations:     0000.04b0 0000.04e4 0409.04b0 0409.04e4

This is dated about the same time as most original windows 7 system files so it is the original driver.

It is the 8.15.10.1851 version you mention earlier.

If windows update has a newer driver, it is WHQL certified so I would give that a try.

Graphic driver: Windows update offered me a new Intel G41 Express driver. I couldn't see which level it was but decided not to take it.

 

[JohnHa]

Thank you for your replies.

ICit2lol: I use sleep mode because W7 takes so long to boot that it drives me to distraction! I reboot every 10 days or so just to get a fresh build of W7. Thank you for the comment about the PSU, but I think I have now eliminated the PSU as a possibility, as I am pretty sure it is software.

My SSD has arrived and installing it it is like upgrading the PC. I will never again have a PC without an SSD for the OS. It just flies - reboot is very fast and applications snap open instantaneously. I now no longer have the long wait which drove me to use sleep.

I built a completely new W7 onto the SSD (I wanted to do so for other reasons as well) using the help on sevenforums (I have an HP OEM PC). During the installation of the (107) W7 updates, Resource Monitor reported the SSD was writing at 35 MBytes/sec and the updates went in very quickly. Normally, MS updates really slug the PC (2 x 2.6GHz, 4GB), but I could now start and use other applications as though update was not running. I think most PCs must be "disk I/O limited" because I noticed that my CPU was running at much higher levels than usual, suggesting that the CPU was able to run faster because it was not waiting for data.

I used Double Driver to backup my "old disk" drivers - it is excellent as it finds them, names them and can backup and restore them. I then kept testing the sleep problem as I slowly rebuilt the PC and added drivers and installed software. So far I cannot force the sleep error.

When I have fully rebuilt the PC, I will either have found what causes the error during the rebuild; or I will compare the "new build" installed drivers with those on the "old disk" so I can identify precisely what is different.

derekimo: Please - no apology is required! I think it probably is the Intel/HP G41 Graphics driver but I need to do the confirmation tests.

During the rebuild of W7, one of the optional updates was "Intel Corporation - Display - Intel(R) G41 Express Chipset". More information gave Driver Information, but that site did not give me any details of the driver. Being cautious, I unticked it - it was optional - but it installed it in any case. Device Manager shows it is Intel 8.15.10.2202 from 25 Aug 2010. I note that this is later than the 8.15.10.1851 from 21 Jul 2009 driver which your report says is causing the crashes, so this may be the fix.

MS Update is still offering me an optional driver. It is Driver Information, dated Oct 2010, so again later than the one currently have installed. Now I know how to find, install, check, and roll back drivers (Double Driver is very good) I will see if I can prove positively that 8.15.10.2202 is causing the problem.

I had not realised that "If windows update has a newer driver, it is WHQL certified" . I was being lead (astray?) by the feeling I should stick to the manufacturer's driver, and HP presumably hasn't updated its site.

Thank you again for all your help - I will update as I find more.

 

[ICIT2LOL]

John no probs mate if you are having long boot up times take a read of this
http://www.sevenforums.com/tutorials/11728-optimize-windows-7-a.html one of the best tips in there is point 8. But they all really slash the boot up times

PS Meant to add I see in the specs you have free AVG of which I ma not a fan is there another suite you can use?? http://www.techsupportalert.com/best-free-anti-virus-software.htm
and what RAM have you got/ DDR2? or DDR3?

 

[JohnHa]

Thanks!

TaskManager and msconfig are two of the most heavily used programs on my PC as I watch startup programs and running processes like a hawk.

I am about to try Startup Delayer from Startup Delayer Trusted Freeware download and reviews - SnapFiles.com to sequence startup programs, and delay less essential ones. Cobian backup is a resource hog at startup as it scans the disk when started so that it can report the size of each of the backup jobs when you open it. There is no reason that cannot be delayed for 5 minutes to let me do what I wanted when I sat down at the PC.

AVG - it's just an old habit - I have thought of others but haven't got round to it. I have never had a virus so far - no doubt I'll get one tomorrow :-(

The RAM is DDR3.

 

[ICIT2LOL]

Yep John I only have the security ticked on the Startup tab and nothing in the Boot options.

A bit OT but here are three I keep as on call scanners and often refer to them in posts where malware is a might be (you might have them already)


http://www.superantispyware.com/

http://www.malwarebytes.org/products/malwarebytes_free/

http://www.bleepingcomputer.com/download/adwcleaner/

download from bleeping computer – delete any rubbishthese find.

If you are tempted to switch then there are plenty to choose from in that link I sent and I have tested most of them on the tester machine at home and I found Bitdefender, Avast, and Avira to be quite good.
But I have this thing about free stuff like that insomuch for about $0.10 a day I run the full Kaspersky 2014 and cannot really see that breaking the bank.
But I digress - sorry.

 

[derekimo]

derekimo: Please - no apology is required! I think it probably is the Intel/HP G41 Graphics driver but I need to do the confirmation tests.

During the rebuild of W7, one of the optional updates was "Intel Corporation - Display - Intel(R) G41 Express Chipset". More information gave Driver Information, but that site did not give me any details of the driver. Being cautious, I unticked it - it was optional - but it installed it in any case. Device Manager shows it is Intel 8.15.10.2202 from 25 Aug 2010. I note that this is later than the 8.15.10.1851 from 21 Jul 2009 driver which your report says is causing the crashes, so this may be the fix.

MS Update is still offering me an optional driver. It is Driver Information, dated Oct 2010, so again later than the one currently have installed. Now I know how to find, install, check, and roll back drivers (Double Driver is very good) I will see if I can prove positively that 8.15.10.2202 is causing the problem.

I had not realised that "If windows update has a newer driver, it is WHQL certified" . I was being lead (astray?) by the feeling I should stick to the manufacturer's driver, and HP presumably hasn't updated its site.

Thank you again for all your help - I will update as I find more.

You're welcome John, it's a pleasure working with someone willing to research and troubleshoot like yourself.

Sounds like you're doing some good things and keeping us posted so that helps everyone involved.

Regarding this statement,

"If windows update has a newer driver, it is WHQL certified"

Here is a little more info on that,

WHQL Release Signature (Windows Drivers)

If your driver package is digitally-signed by WHQL, it can be distributed through the Windows Update program or other Microsoft-supported distribution mechanisms.

Thanks for keeping us updated, hoping to hear good results from you.

 

[JohnHa]

My apologies for the delay in this update, especially as I have to say that while it is now fixed (I rebuilt Windows 7) I have been unable to track down what was causing it. I had thought it was probably the Intel G41 Express Graphics driver but after rebuilding W7, I re-installed the graphics driver I had been using previously - 8.15.10.1851- when the "power off while sleeping" happened and despite sleeping 50 times, it never powered down. So something else fixed it and I have no idea what it was.

I am however very pleased that I rebuilt Windows 7, especially as it was much easier than I had expected it to be; and I am very pleased about the new SSD disk. Apparently the speed improvement you get with SSD has little to do with the fast read / write speed - rather it is because

a) the latency for accessing many smaller than 4kB files used during W7 startup is much lower, so the whole startup is much faster. I can go form clicking on RESTART to rebooting and using it in less than 50 seconds (2 x 2.6GHz, 4GB).

b) a typical HDD does about 10,000 I/O operations/second but my SSD does 70,000 I/O ops/second. This means that the PC is MUCH more responsive when running multiple applications. It doesn't behave sluggishly while installing Windows updates (and I applied 150+ updates in about 35-40 minutes!) and, as soon as the desktop appears, I can use applications.

Windows 7 recognised I had a SSD and automatically switched off defrag for it (defrag just wastes writes while doing nothing to improve performance), and also enabled TRIM (before you can write to a a deleted file's blocks marked "not in use", you must erase the data on thos blocks - TRIM goes round in the background and erases all the data on "not in use" blocks so you don't have to wait when writing).

Thank you everyone for all your help.

Added information. I have just discovered that the driver shipped with the PC was 8.15.10.1851. I have Factory Image partition on my disk which holds the system information for rebuilding the image if you use the HP Recovery. I was never able to look inside this partition but now it is not a partition on my boot disk I can see inside. F:\hp\Drv\APP00300\IntelGfxELCG_15.15.2.1851.WHQL_ALL_DRV_W6_WW.txt is a TXT file describing the driver and it says:

Block: IntelGfxELCG_15.15.2.1851.WHQL_ALL_DRV_W6_WW
Modified by: Minh Phu
Created by: Soyoung Ahn
Creation Date: Sat. July 25, 2009

Description:
------------

Intel Graphics Driver - WHQL'd

DriverVer=07/21/2009,8.15.10.1851

I don't know why Windows Update updated this driver to 8.15.10.2202 but it did. I therefore expected that once I re-installed the ...1851 driver, Update would promptly update it to ... 2202. But Update hasn't and when I check if I have the latest driver, it says I have.

It used to be thought a sign of madness to do the same thing while expecting a different outcome but Microsoft seems to have has established that as a norm! I shall take consolation in the Windows Haikku:
Yesterday it worked.
Today it is not working.
Windows is like that.

 

[derekimo]

I have to admit John, it's been a pleasure, very interesting following your progress.

I'm glad to hear you have things back, even if the cause wasn't found we have all learned something in the journey.

Thanks for posting back with your results.