Desktop pc freezes when using either Win XP or Win 7 OSs

[apairofpcs]

Hello Ladies and Gentlemen,

This is my first post on a pc related forum. The decision to write was long overdue. My problem has plagued me for the last two months. My desktop pc has been freezing without warning, disabling the mouse and keyboard and forcing me to restart by pressing the reset button on the front of the case. During the freeze, the hard drives and fans continue to spin.

It started two months ago in Windows XP Professional Edition, which has been my main boot OS since Apr. 2009. Then on Jan 27th, I started using Windows 7 Ultimate (32bit). The problem persisted even though I was using two physical hard drives and two distinct OSs. Everything points to an intermittent hardware failure, as it appears. I’m not so sure, though.

I feel that the more data I give you about my system’s components and my countless attempts at stopping the freezes, the better it will be for any of you to decide on the cause of the freezes. For starters, please bear with me as I list the vital components of my desktop pc. Keep in mind that I have had the desktop pc since Apr. 2009, and this is the first time it has given me a persistent problem.

This section is vintage Apr. 2009 equipment:
· ASUS P5Q SE/R socket 775 motherboard, with integrated Atheros L1E Gigabit PCI-E Ethernet Controller
· 2 x 2 GB G.Skill DDR2 1066 SDRAM memory cards
· Intel Core 2 Duo E8400 “Wolfdale” CPU
· Cooler Master RC-690-KKN1-GP BK case
· Sapphire HD4670 graphics card (ATI Radeon 4670)
· CORSAIR CMPSU-450VX 450W power supply
· 2 x Western Digital WD2500AAJS 250 GB SATA-2 hard drives (each with Windows XP installed)
· HAUPPAUGE WinTV HVR-1800 TV Tuner
· Logitech Performance Mouse MX cordless/optical laser mouse (USB receiver)
· Samsung SP2504C 250 GB SATA-2 hard drive (with Windows 7 32 bit installed)
· Acer P223w 22” LCD display
· Kensington Masterpiece Plus surge protector (vintage July 1992)
· Westell 6100 Wind River DSL Modem
· Linksys E2000 Wireless Router

From a Dell Dimension 4100 system (vintage Nov. 2000):
· Microsoft Internet keyboard with 104 keys (PS2 plug)
· Altec Lansing ACS340 speaker system with sub-woofer

From a Dell Dimension 2400 system (vintage 2004):
· Dell M-UVDEL1 optical/corded two button mouse with scroll wheel (USB plug)
· Dell Basic keyboard with 104 keys (PS2 plug)

The first freeze came about the end of Dec. 2011, when I was using the Windows XP OS. On Jan. 27, 2012, I started using the Windows 7 OS.
I have made countless attempts to stop the freezes in two months. But just when I think they have stopped due to an action I took, they creep up again. One thing is for sure. The freezes occur ONLY when I am connected to the internet in GUI. But when I am in safe mode with networking, there are no freezes. Neither Avast nor AVG load in safe mode, so I was playing a little “dangerously.” By the way, the cable routing is ethernet card, to wireless router, to modem. I have never had a freeze when my internet connection is disabled. It’s well known that an anti-virus program can create problems, because of how heavily it is involved with the OS. I was using Avast FREE for a few years, but changed to AVG FREE on Feb. 27th to see if Avast was the problem. The freezes continued.

I have had freezes when browsing with Firefox, writing emails in Thunderbird, checking device drivers in Device Manager, playing audio.video clips from the Internet or my hard drive and downloading files. In fact, I can have the pc resting at the desktop for hours, and when I move the mouse…..NOTHING! I have found installed the latest device drivers. I have rolled back device drivers to previous versions. I have disabled and enabled devices, such as the ethernet card…..which I suspect to be the culprit. I have swapped mice. I have swapped keyboards. I have replaced the yellow cable from the ethernet card to the wireless router. I have done defrags. in GUI and Safe Mode. I have done numerous chkdsk scans with the /R parameter, with no bad sectors found. I have done basic and detailed virus scans with Avast and AVG in GUI and safe mode, with few threats found and the unimpeded deletion of any that are found. I have been creating manual system restore points daily. I have been using Disk Cleanup daily. I used CPU Burn-In for six hours, with a successful termination. I used the HD Tune surface scan program to check for bad blocks on the two XP drives and the Win 7 drive. There were none. I have been manually checking Windows Update on a daily basis. I have reset the modem and router by pressing their recessed red buttons with a pen. I have disconnected the power and data cables from each device, waited a minute and reconnected everything, a tactic suggested by Verizon my ISP and the provider of the DSL modem.

Here is an usual way for a pc to conduct a restart!! There have been rare times when I do a case button reset and I lose my CMOS Setup settings. I have to manually configure them. There are times that I do the same reset, and the hard drives and fans stop spinning for about 5 seconds, before they start up again. I have selected “Load default settings”, although they don’t suit my style of pc use. I have unseated and reseated my two memory cards. I have unseated/reseated the SATA cables from the three hard drives. I have removed/reseated the graphics adapter. The CMOS battery is new. The pc clock keeps good time. I have disabled all interventions to the pc’s function as part of Power Options’ High Performance plan.

Something unusual happens with the mouse pointer when using any mouse. It will jump to another part of the screen and return back to where it was before. I installed the Logitech Setpoint program and have enabled it and disabled it within MSConfig’s Startup tab. I have disabled the screen saver function.

I can successfully place the pc in Sleep mode or Hibernate mode, and have it awaken without a problem.

Can anybody on this planet suggest something that I have not yet done, to stop my freezes?

Thank you all,
apairofpcs

 

[MaxBurn]

I guess you could start by checking the motherboard for bulging and leaking capacitors. With a multi meter I might check the 5v and 12v lines but that is a decent PS so unlikely to find something there. Proceed on to memtest86 and something like superpi or burn in tester. If you have another computer there are things you can do to check out the network card with point to point tests. Someone here in the business likely has a checklist. Myself I don't know how much time I would waste troubleshooting it before I called it time and did a core upgrade. Troubleshooting when you aren't set up for it can take a lot of time and money and you could potentially save both by just replacing.

 

[BingBangBop]

Your mouse jumping could be as simple as a mouse setting: mouse properties->pointer options->Snap To (don't allow it to automatically move the pointer to the default button in a dialog box)

One of the things you have not done as a diagnostic test is to check for RAM errors. Memory errors will commonly cause very difficult to diagnose problems so if I see something particularly confusing Memtest86+ is where I go.

Another thing I would do is check the Windows system logs for hints at the cause of the crashing. Sometimes you can glean some pattern as to the cause.

The fact that you lose your CMOS settings upon a case button restart which is quite peculiar for there should cause any power interruption that could cause it. That would infer a voltage regulation issue.

To continue this theme is that you don't have any problems in safe mode. Normally, I associate that with malware or drivers but neither of those will survive a XP->Win7 install. But one of the characteristics of safe mode is the use of default non-accelerated video drivers. It is the video acceleration that will cause the video card to use lots of power.

 

[apairofpcs]

I guess you could start by checking the motherboard for bulging and leaking capacitors. With a multi meter I might check the 5v and 12v lines but that is a decent PS so unlikely to find something there. Proceed on to memtest86 and something like superpi or burn in tester. If you have another computer there are things you can do to check out the network card with point to point tests. Someone here in the business likely has a checklist. Myself I don't know how much time I would waste troubleshooting it before I called it time and did a core upgrade. Troubleshooting when you aren't set up for it can take a lot of time and money and you could potentially save both by just replacing.

Thank you, MaxBurn. I did a spot check for bulging/leaking capacitors, and my "untrained eyes" told me all is well. My buddy, jtr1962, had to replace some of his motherboard's capacitors when his pc would suddenly "shut down." He started a thread about it here. I will consult with him. My power supply has a good reputation, but I can check this as well.

I am quite sure that something in the "communication chain", between desktop, modem and router, is at fault. I ONLY get freezes when connected to the internet in GUI and never in safe mode with networking. I have a Toshiba laptop, so I can check the ethernet card on a point to point basis. How do I do this? The laptop is fine using the same DSL modem and wireless router to connect to the internet. This points to the desktop to modem to router link, as a place to deeply investigate.

I have plenty of time to troubleshoot. I have the money to bring back my desktop pc to a reliable state. I only have this desktop pc. But tell me what "core" components should I replace, when I don't know the one that is failing? I believe that "core" consists of CPU, RAM, power supply and motherboard. Didn't my successfully concluded six hour burn-in using CPU Burn-In, demonstrate the integrity of these components? Am I missing something as a hobbyist and not a trained professional in the world of computers?

 

[apairofpcs]

Your mouse jumping could be as simple as a mouse setting: mouse properties->pointer options->Snap To (don't allow it to automatically move the pointer to the default button in a dialog box)

One of the things you have not done as a diagnostic test is to check for RAM errors. Memory errors will commonly cause very difficult to diagnose problems so if I see something particularly confusing Memtest86+ is where I go.

Another thing I would do is check the Windows system logs for hints at the cause of the crashing. Sometimes you can glean some pattern as to the cause.

The fact that you lose your CMOS settings upon a case button restart which is quite peculiar for there should cause any power interruption that could cause it. That would infer a voltage regulation issue.

To continue this theme is that you don't have any problems in safe mode. Normally, I associate that with malware or drivers, but neither of those will survive a XP->Win7 install. But one of the characteristics of safe mode is the use of default non-accelerated video drivers. It is the video acceleration that will cause the video card to use lots of power.

Thank you as well, BingBangBop. I never used the "Snap To" mouse feature, but made sure the box was unchecked anyway. The mouse doesn't "Automatically move the pointer to the default button in a dialog box", it migrates to the place where it feels comfortable. It does this very quickly, sometimes startling me in the process.

I saw the free download of Memtest86+. I will make a bootable CD and proceed. I have yet to find the location of the Windows System Log files. How to check for poor voltage regulation? Are you implying a possible video card problem in GUI, with only the basic device drivers being loaded in safe mode? If so, how can I test the card's integrity in GUI?

 

[time]

You haven't said how often it freezes, or how consistent the behaviour is. If it freezes only once a day or less, how can you be sure that running in safe mode or with networking disabled is truly okay?

I know you're hoping it's just a driver issue or some other software glitch, but I'm struggling to remember any instances where that caused sudden loss of mouse and keyboard at random.

You haven't mentioned pulling the TV card - if not, that might be worth doing. Memtest is also useful for elimination, although to be honest I haven't seen either of these cause your problems exactly as described. Have you swapped the graphics card? That would fit the main symptoms, especially if it doesn't do it in safe mode (but doesn't explain your belief that disabling the network port fixes things).

Unfortunately, my diagnosis is a faulty motherboard, or a faulty CPU. The former is very considerably more likely, especially after a couple of years, but the latter is not completely impossible. Tweaks that might narrow it down include boosting the voltage to the CPU and underclocking the FSB. I assume the CPU is running cool, BTW? You could also check that the North and South bridge heatsinks are firm, and what their temperature sensors are saying.

The power supply is also a possible candidate (but not probable given the relatively high quality); if you want to pursue this then you'll have to try another one and see what happens.

Your motherboard has solid caps so there's no point inspecting them for leaks.

 

[apairofpcs]

Your mouse jumping could be as simple as a mouse setting: mouse properties->pointer options->Snap To (don't allow it to automatically move the pointer to the default button in a dialog box)

One of the things you have not done as a diagnostic test is to check for RAM errors. Memory errors will commonly cause very difficult to diagnose problems so if I see something particularly confusing Memtest86+ is where I go.

Another thing I would do is check the Windows system logs for hints at the cause of the crashing. Sometimes you can glean some pattern as to the cause.

The fact that you lose your CMOS settings upon a case button restart which is quite peculiar for there should cause any power interruption that could cause it. That would infer a voltage regulation issue.

To continue this theme is that you don't have any problems in safe mode. Normally, I associate that with malware or drivers but neither of those will survive a XP->Win7 install. But one of the characteristics of safe mode is the use of default non-accelerated video drivers. It is the video acceleration that will cause the video card to use lots of power.

Thank you as well, BingBangBop. I wrote a reply post, but it never showed up here. So here is my 2nd attempt.

I never used the mouse properties "Snap To" feature, but made sure the box was unchecked anyway. In fact, by definition, checking it will "Automatically move pointer to the default button in a dialog box." The pointer "jumps" to a distant part of the screen without my input, and returns to where it feels comfortable. This erratic behavior makes me "uncomfortable."

I found the free download of Memtest86+, will make a bootable CD and run it.

I looked for the location of the Windows System Logs, to no avail. Would you direct me?

Indeed, the reset button on the case is a true "reset button for a restart" and not a shut down button. The shut down button is above it. I am careful to only press the reset button. I have been doing case initiated restarts since I had my first pc in 1992 and I have never seen this behavior. How do I test the voltage regulation of the system?

Are you implying that the video card may be the culprit, because it has no chance to fail in safe mode when it is working off the basic device drivers loaded there? So how do I test the video card in GUI? If the display remains intact after a freeze, what does this tell you about the video card's function?

 

[BingBangBop]

If you need to, you can get yourself an inexpensive network card and then use that instead of the onboard one. If the freezing stops then you've ID'd the problem and also found at least a temporary solution.

Really, I discount your network activity hypothesis because of losing CMOS setup whenever you do a warm boot via the reset button. The CMOS problem really limits the flaw to power regulation or MB. Of course on your current setup you are using the MB's network card so there's the likely connection.

A CPU burn-in does not exercise all the capabilities of the components to their fullest. For example a gaming video card while gaming will can pull more power than the CPU+RAM+MB combined. So burning in the CPU won't test to see if the PS can supply proper voltage over its entire power range. A CPU burn-in will do a few CPU instructions in a long term loop but it does not write and read RAM at every location to see if one of your RAM sims has lost a bit. What I'm saying is a burn-in is not a universal test for all potential problems. A burn-in test will catch most of the obvious problems but won't work for lots of other more subtle issues.

If you can't diagnose the problem then what you are left with is swapping out individual components till the problem goes away. You just start out with the cheapest and move to the most expensive one replacement at a time.

P.S. I agree with Time that it is most likely the MB with a side possibility of the PS.

 

[time]

Our posts crossed. Jtr1962 has a very, very old PC; there really is very little in common with yours.

Communication between your modem and router cannot halt your PC so that even the mouse doesn't work. Similarly, I have never heard of a network problem causing a PC to actually halt. Why do you suspect the network comms? It's not clear from your information so far.

A 'trained professional' balances effort with reward, so they would probably run Memtest because they can be getting on with something else while it's running. There may be no payoff, but it's one less assumption for very little effort. I would probably then change the graphics card, hook up a different power supply and leave the PC running something for several hours to see what happened - again, minimal effort and something else hopefully eliminated.

If the freezes persist, I would be advising you to consider upgrading the motherboard, CPU and RAM because even the cheapest Sandy Bridge CPU (Celeron G530) is a match for your current CPU and newer motherboards have 6Gb/s SATA and USB 3.0. There's too much effort in changing the motherboard and/or CPU to justify similar vintage parts - but that's looking at it from the point of view of someone who needs to charge for their time. You may enjoy going through that process, as did Jtr1962.

A CPU burn-in test is useful as an initial test of overclocking stability or as a thermal stress-test, but otherwise isn't much use at all. Gaming benchmarks exercise your system far more broadly, but problems like yours might only manifest themselves when the CPU is relatively idle (or changing power states).

 

[apairofpcs]

You haven't said how often it freezes, or how consistent the behaviour is. If it freezes only once a day or less, how can you be sure that running in safe mode or with networking disabled is truly okay?

I know you're hoping it's just a driver issue or some other software glitch, but I'm struggling to remember any instances where that caused sudden loss of mouse and keyboard at random.

You haven't mentioned pulling the TV card - if not, that might be worth doing. Memtest is also useful for elimination, although to be honest I haven't seen either of these cause your problems exactly as described. Have you swapped the graphics card? That would fit the main symptoms, especially if it doesn't do it in safe mode (but doesn't explain your belief that disabling the network port fixes things).

Unfortunately, my diagnosis is a faulty motherboard, or a faulty CPU. The former is very considerably more likely, especially after a couple of years, but the latter is not completely impossible. Tweaks that might narrow it down include boosting the voltage to the CPU and underclocking the FSB. I assume the CPU is running cool, BTW? You could also check that the North and South bridge heatsinks are firm, and what their temperature sensors are saying.

The power supply is also a possible candidate (but not probable given the relatively high quality); if you want to pursue this then you'll have to try another one and see what happens.

Your motherboard has solid caps so there's no point inspecting them for leaks.

It freezes when it wants, many times a day if I remain online. When I uninstalled Avast AV and installed AVG AV, there were no freezes for about 20 hours. When I disabled/enabled the ethernet card from within Device Manager, there were no freezes for about 8 hours.

Knowing the risk of being on the internet without an AV program to protect me. I have used the pc in safe mode with networking for hours, with no freezes. I have also left the pc idle in safe mode with networking for many hours, with no freezes.

Yes Sir, device driver problem is my hope, but I am ready, willing and able to replace a "core" component ASAP.

The TV tuner card was active when using the XP OS, but I haven't installed the software on the Win 7 HDD.

I see your point about doing diagnostic testing in the background, and it couldn't hurt. I have not considered the graphics card to be the problem, so I have not swapped it.

I can handle replacing any component at all, even the MB or CPU. But my stubborn streak and relentless nature to solve a problem prevents me from arbitrarily swapping components with the hope of finding the one that was faulty. I would remain this way if I became filthy rich. Every pc problem starts a battle between my human intellect and a set of metal/plastic parts that are supposed to work for me, with me usually winning the battle.....in time!

I'll have to check the CPU temp in CMOS Setup, unless there is another loaction in GUI, and get back to you.

I've only seen a reference to the North/South bridge heat sinks in CMOS Setup, so where do I check for their temperature in GUI?

After jtr1962 saw the pictures of my motherboard and read it's specs., he also determined that I had solid caps. This means that if a cap was failing, the caps cannot be replaced and the MB has to be replaced.

Oh well, I have some work to do to implement the many diagnostic checks that are necessary to nail my problem.

 

[BingBangBop]

Install a temperature program like real temp and dynamically monitor via the systray. Checking the CPU temp in CMOS setup is totally useless unless you need an idle temp.

 

[apairofpcs]

If you need to, you can get yourself an inexpensive network card and then use that instead of the onboard one. If the freezing stops then you've ID'd the problem and also found at least a temporary solution.

Really, I discount your network activity hypothesis because of losing CMOS setup whenever you do a warm boot via the reset button. The CMOS problem really limits the flaw to power regulation or MB. Of course on your current setup you are using the MB's network card so there's the likely connection.

A CPU burn-in does not exercise all the capabilities of the components to their fullest. For example a gaming video card while gaming will can pull more power than the CPU+RAM+MB combined. So burning in the CPU won't test to see if the PS can supply proper voltage over its entire power range. A CPU burn-in will do a few CPU instructions in a long term loop but it does not write and read RAM at every location to see if one of your RAM sims has lost a bit. What I'm saying is a burn-in is not a universal test for all potential problems. A burn-in test will catch most of the obvious problems, but won't work for lots of other more subtle issues.

If you can't diagnose the problem then what you are left with is swapping out individual components till the problem goes away. You just start out with the cheapest and move to the most expensive, one replacement at a time.

P.S. I agree with Time that it is most likely the MB with a side possibility of the PS.

Despite all the suggestions that I have an MB, PS, CPU or RAM problem, replacing an integrated ethernet card with an independent one is a good first step. Actually, I dread replacing the MB the most, because of the amount of work required. Everything else is a piece of cake on labor and expense levels.

A correction is in order here. Out of 50 or so freezes in two months, with a case button reset starting a restart, I have had ONLY 10 or so hard drive/fan spin downs followed by their spin-ups. Of these 10, only 5 scratched my CMOS settings. By the way, after every reset button restart, the white text on a black background screen stops the activity, telling me that the pc was not shut down properly. I always select "Start your computer normally."

I understand how MY CPU Burn-In's results are inconclusive. I run it with nothing running in the background, meaning that the CPU is not stressed. The burn-in is supposed to stress an already stressed system and it has not done that in my case. I am not a gamer, so what can I do to stress the CPU before starting a CPU Burn-In test?

Thank you for laying out the list of possibly faulty components and explaining why you think they may be the culprit. If I didn't need expert opinions, then I wouldn't come to this forum. I am happy that I have a laptop as a backup pc.

 

[MaxBurn]

Yes core upgrade = motherboard, memory, cpu, video card and in your case I would likely do the PS as it is suspect. You get upgrades and end your troubleshooting, win-win. Throw in an SSD too.

Further testing for example would be looping 3d mark, that would answer the question of full system load, video card and just about everything down to PS. I don't know what you used for burn in testing but I was thinking of http://www.passmark.com/products/bit.htm. It does 3d testing and has a network loopback test as well.

Someone here recommended me a little dos utility to test actual network throughput that might put some stress on the networking components to see if that starts failing, I can't remember the name of it though. Need to run it on a source and destination.

 

[apairofpcs]

Our posts crossed. Jtr1962 has a very, very old PC; there really is very little in common with yours.

Communication between your modem and router cannot halt your PC, so that even the mouse doesn't work. Similarly, I have never heard of a network problem causing a PC to actually halt. Why do you suspect the network comms? It's not clear from your information so far.

A 'trained professional' balances effort with reward, so they would probably run Memtest because they can be getting on with something else while it's running. There may be no payoff, but it's one less assumption for very little effort. I would probably then change the graphics card, hook up a different power supply and leave the PC running something for several hours to see what happened - again, minimal effort and something else hopefully eliminated.

If the freezes persist, I would be advising you to consider upgrading the motherboard, CPU and RAM because even the cheapest Sandy Bridge CPU (Celeron G530) is a match for your current CPU and newer motherboards have 6Gb/s SATA and USB 3.0. There's too much effort in changing the motherboard and/or CPU to justify similar vintage parts - but that's looking at it from the point of view of someone who needs to charge for their time. You may enjoy going through that process, as did Jtr1962.

A CPU burn-in test is useful as an initial test of overclocking stability or as a thermal stress-test, but otherwise isn't much use at all. Gaming benchmarks exercise your system far more broadly, but problems like yours might only manifest themselves when the CPU is relatively idle (or changing power states).

I understand all the opinions about my communication path not having the means to cause freezes. But I can only state that I have never had a freeze when OFFLINE, although I have never been patient enough to stay OFFLINE for 6 hours or so. Maybe it needs that long to fail? I would be puzzled as you are, if I didn't see it happening for the last two months. It's quite illogical, by my standards.

Yes, jtr1962's pc has been around a loooong time. I admire him for not sending it to pc heaven before it's time, just because it misbehaved in a big way. He was patient, and it paid off with a fully functional MB.

No argument from me about the merits of using Memtest86. In fact, all of your ideas make sense and are not a problem to implement. I can live quite well with my laptop, in the meantime.

Speaking of the most common scenario that precedes a freeze, it is when it is sitting idle for a while.

 

[apairofpcs]

Install a temperature program like real temp and dynamically monitor via the systray. Checking the CPU temp in CMOS setup is totally useless unless you need an idle temp.

Before I started having freezes, I was about to download a temp. program to play pc doctor a little. Is the one you recommend above called "Real Temp, or is this a general term?

By the way, I had a freeze while writing this post the 1st time....and I was naturally ONLINE when it happened.

 

[apairofpcs]

Yes core upgrade = motherboard, memory, cpu, video card and in your case I would likely do the PS as it is suspect. You get upgrades and end your troubleshooting, win-win. Throw in an SSD, too.

Further testing for example would be looping 3d mark, that would answer the question of full system load, video card and just about everything down to PS. I don't know what you used for burn in testing but I was thinking of http://www.passmark.com/products/bit.htm. It does 3d testing and has a network loopback test as well.

Someone here recommended me a little dos utility to test actual network throughput that might put some stress on the networking components to see if that starts failing, I can't remember the name of it though. Need to run it on a source and destination.

Oh boy, you really know how to UPGRADE like a champ!!! Of all the words you used to describe pc parts, the only one that put a smile on my face is "SSD." How about we chew the fat about this "luxury item", after my freezes stop? My mind is a little "overloaded" right now. Haa haa!

I used "CPU Burn-In" for the stress tests. I will use YOUR recommendation, since the program I used seemed very bare bones, and that bothers me a little. If there is no way to configure a program, I don't respect it that much.

I will seek that DOS program, since I would like to run that before anything else. I am still thinking "communication" as the problem.

Listen guys and gals. If my observations had even one inconsistency, a freeze when OFFLINE being that one, I would not be so fixated on a communication problem. It's hard for me to get this out of my head, when the symptoms point in one direction. Nevertheless, I respect the opinions of all of you as people who do more than fool around with pcs.

 

[apairofpcs]

Install a temperature program like real temp and dynamically monitor via the systray. Checking the CPU temp in CMOS setup is totally useless unless you need an idle temp.

Using Real Temp, core 0 = 41 C and core 1 = 39 C Cool as a cucumber!

 

[MaxBurn]

The network tool I was thinking of is iperf.exe discussed here http://www.storageforum.net/forum/s...e-Server-Home-Server-OS/page3?highlight=iperf

You could just shove a bunch of data back and forth and see what happens.

 

[BingBangBop]

Despite all the suggestions that I have an MB, PS, CPU or RAM problem, replacing an integrated ethernet card with an independent one is a good first step. Actually, I dread replacing the MB the most, because of the amount of work required. Everything else is a piece of cake on labor and expense levels.

A correction is in order here. Out of 50 or so freezes in two months, with a case button reset starting a restart, I have had ONLY 10 or so hard drive/fan spin downs followed by their spin-ups. Of these 10, only 5 scratched my CMOS settings. By the way, after every reset button restart, the white text on a black background screen stops the activity, telling me that the pc was not shut down properly. I always select "Start your computer normally."

I understand how MY CPU Burn-In's results are inconclusive. I run it with nothing running in the background, meaning that the CPU is not stressed. The burn-in is supposed to stress an already stressed system and it has not done that in my case. I am not a gamer, so what can I do to stress the CPU before starting a CPU Burn-In test?

Thank you for laying out the list of possibly faulty components and explaining why you think they may be the culprit. If I didn't need expert opinions, then I wouldn't come to this forum. I am happy that I have a laptop as a backup pc.

First, I'll start out with that the network card isn't that good of an idea. Unlike you, I really don't think of this as a network problem. I suggested it really to satisfy your obsession with the network association. If it is a network problem (non-motherboard related) then the odds-on favorite is that you actually have a driver issue or you re-caught some malware from your previous XP install. perhaps you went to the same site and caught the same malware or you copied the malware when you restored your old data. However, it is much cheaper and easier to do than replacing the MB so it potentially has some value

Next, the reason I'm obsessed with the loss of CMOS as a symptom is because on a working computer that should never happen upon pressing the reset button. I understand that it is something that does not bother much and it is not what you are complaining about but It is a guaranteed HW problem that limits the choices and solutions. I'm also assuming that you have only one related problem and not several independent issues so that fixing that one, will potentially fix the problem you have an issue with.

Windows should always tell you that the PC was improperly shut down upon reset button restart because it was. When you shut down normally, Windows modifies a file saying that it is shutting down normally. When Windows starts back up, that file is read (and the modification removed) and if it did not see that modification it will give that warning that it was improperly shut down and often it will want to check your HD too. The reason for that is caching HD data. If properly shut down all cached data is written out but if the power goes out or the reset button is pushed the shutdown bypasses Windows shutdown process so any cached data will not be written producing a potential file structure issue. Most people don't write cache their HD's and most caches auto-flush within a second so shutting down is not likely to cause a problem but it could so Windows always warns.

A good burn-in does actually stress the CPU to some extent. Not all parts of it because they normally only do a small subset of the instructions possible but it should run the CPU at 100% and it should at least thermally stress the CPU. It may not really test everything well, it does show that the CPU, MB, PS, RAM are minimally functional. The real purpose of burn-in tests were to detect initial massive failures before sending a new machine to the new owner for it is really embarrassing to ship out a computer that does not run at all. Most component failures occur either very early in ownership or after the warranty period. So the theory was to run the machine for a few hours to a few days so as to try to catch those early total failures. Burn-in testing was never designed to diagnosis or to detect subtle issues that can creep in over time. For your problem, it is not the type of test I would give much weight as to the reliability of your components for there are too many things that just aren't tested by just letting your CPU run.

 

[CougTek]

First off, I wish "apairofpcs" would be a pair of something else, although I wouldn't care unless you'd be half your age.

Secondly, you can probably forget the driver route, since it freezes in both WinXP and Win7 and the drivers for both are generally quite different.

Third, I believe you're stressing the CPU with softwares for nothing since your PC freezes even when ilding. Ditto for the temperature monitoring programs. Your problem is not there.

I'd test the RAM and hard drive just to be sure. If you haven't done a few passes of Memtest, then start there, as others have pointed out. Regarding the hard drive, I've seen hard drives deactivating themselves and freezing computers some times. This behavior doesn't always show when a hard drive testing program runs, but it wouldn't be a bad thing to check. Swapping the network card should be tried as well. It goes without saying that you should de-activate the old one in the BIOS and plug the network cable in the newer card as well.

The PSU swapping can be tried too, although my money (pennies, I'm quite poor) isn't on it. I'd suggest you to lower the frequency of your RAM. 1066MHz DDR2 were often unstable in the long term. Using your 1066MHz sticks at 800MHz and at the voltage specified in the DDR2 standard could improve your situation.

If none of the above fixes your problem, you have a very probable motherboard issue. Organize the funerals and after a solemn ceremony, replace it by a newer model filled with promesses (along the CPU and RAM at least).

 

[BingBangBop]

Using Real Temp, core 0 = 41 C and core 1 = 39 C Cool as a cucumber!

Now do your burn-in and see where those temps end up. It shouldn't take more than a couple of minutes to get to your maximum temperature. Make sure that the burn-in is running both cores at 100% using the task manager->performance tab.

Now, I don't think your problem is thermal for you said it occurs during network activity and that does not normally stress the CPU at all. Rather, this is just a quick test you might find interesting -- Your maximum operating temperature.

 

[apairofpcs]

Hello all,

I took a break yesterday, after a long day slaving over my desktop. This morning I removed the yellow cable connecting the ethernet port to the wireless router.....and I got a freeze a short while later. Most of you were right that my problem is not network related. This is the proof.

Today is another day and I'm on a new kick, spending all day in safe mode without networking and working the pc as I always do, but without an internet connection and AV protection. A few of my non-internet based tasks will not work, but that's OK. I'm on a mission.

If the day goes by and I don't get a freeze, what will this mean as far as a hardware problem? I thought that safe mode was devised to eliminate the complex GUI software that works with hardware, while loading simple device drivers to have the hardware function. If I get a freeze, we all know what THAT means!!! But if I don't, does it mean that none of my hardware is faulty? How can a pc's components work reliably in safe mode, and not work in GUI?

I hope I'm making sense.

 

[apairofpcs]

How could I forget that for the first time, safe mode has offered me about 8 different screen resolutions from which to choose. I have never seen this before. It's always opened at 800 x 640 and I didn't think I had a choice. I wish I did all along. A low resolution is terrible when there are many icons sharing space. The resolution that was applied when desktop arrived was 1,300 x XXX. But since my monitor is a wide aspect type, none of the offered resolutions would fit the monitor. I picked the resolution that most appealed to me.

Although I have never had a safe mode request end in anything other than 800 x 640, were the other resolutions always there when right clicking on the desktop and I never looked for them? The big surprise was the 1,300 x xxx screen that appeared. Any ideas why this happened?

 

[Mercutio]

Having read everything you and others have written, I'd probably just tell you to replace your motherboard and be done with it.

 

[apairofpcs]

First, I'll start out with that the network card isn't that good of an idea. Unlike you, I really don't think of this as a network problem. I suggested it really to satisfy your obsession with the network association. If it is a network problem (non-motherboard related) then the odds-on favorite is that you actually have a driver issue or you re-caught some malware from your previous XP install. perhaps you went to the same site and caught the same malware or you copied the malware when you restored your old data. However, it is much cheaper and easier to do than replacing the MB so it potentially has some value

Next, the reason I'm obsessed with the loss of CMOS as a symptom is because on a working computer that should never happen upon pressing the reset button. I understand that it is something that does not bother much and it is not what you are complaining about but It is a guaranteed HW problem that limits the choices and solutions. I'm also assuming that you have only one related problem and not several independent issues so that fixing that one, will potentially fix the problem you have an issue with.

Windows should always tell you that the PC was improperly shut down upon reset button restart because it was. When you shut down normally, Windows modifies a file saying that it is shutting down normally. When Windows starts back up, that file is read (and the modification removed) and if it did not see that modification it will give that warning that it was improperly shut down and often it will want to check your HD too. The reason for that is caching HD data. If properly shut down all cached data is written out but if the power goes out or the reset button is pushed the shutdown bypasses Windows shutdown process so any cached data will not be written producing a potential file structure issue. Most people don't write cache their HD's and most caches auto-flush within a second so shutting down is not likely to cause a problem but it could so Windows always warns.

A good burn-in does actually stress the CPU to some extent. Not all parts of it because they normally only do a small subset of the instructions possible but it should run the CPU at 100% and it should at least thermally stress the CPU. It may not really test everything well, it does show that the CPU, MB, PS, RAM are minimally functional. The real purpose of burn-in tests were to detect initial massive failures before sending a new machine to the new owner for it is really embarrassing to ship out a computer that does not run at all. Most component failures occur either very early in ownership or after the warranty period. So the theory was to run the machine for a few hours to a few days so as to try to catch those early total failures. Burn-in testing was never designed to diagnosis or to detect subtle issues that can creep in over time. For your problem, it is not the type of test I would give much weight as to the reliability of your components for there are too many things that just aren't tested by just letting your CPU run.

Based on my post #22, the network, or should I say ethenet section of the MB, is no longer suspect. Thank you for coddling me, though. Like you, I have never seen CMOS settings reset and hard drives and fans stop spinning for any reason, except sleep mode, hibernate or shut down.

I have done endless scans with AVG, in GUI and in safe mode. No viruses. No malware. Let me elaborate on what is on each of my 3 case resident hard drives. One is my primary XP boot drive. Two is my secondary XP boot drive. Third is my Win 7 boot drive. Neither drive has had any OS other than what I have described. The Win 7 drive is my most recent one. A month after the first XP drive started having freezes, I started using the Win 7 drive. The freezes continued immediately. Correct assumption, no other pc problems. Three drives working in perfect harmony, until the end of Dec. 2011.

Windows has been doing it's job, warning me after each improper shut down.

 

[apairofpcs]

First off, I wish "apairofpcs" would be a pair of something else, although I wouldn't care unless you'd be half your age.

Secondly, you can probably forget the driver route, since it freezes in both WinXP and Win7 and the drivers for both are generally quite different.

Third, I believe you're stressing the CPU with software for nothing, since your PC freezes even when idling. Ditto for the temperature monitoring programs. Your problem is not there.

I'd test the RAM and hard drive just to be sure. If you haven't done a few passes of Memtest, then start there, as others have pointed out. Regarding the hard drive, I've seen hard drives deactivating themselves and freezing computers some times. This behavior doesn't always show when a hard drive testing program runs, but it wouldn't be a bad thing to check. Swapping the network card should be tried as well. It goes without saying that you should de-activate the old one in the BIOS and plug the network cable in the newer card as well.

The PSU swapping can be tried too, although my money (pennies, I'm quite poor) isn't on it. I'd suggest you to lower the frequency of your RAM. 1066MHz DDR2 were often unstable in the long term. Using your 1066MHz sticks at 800MHz and at the voltage specified in the DDR2 standard could improve your situation.

If none of the above fixes your problem, you have a very probable motherboard issue. Organize the funerals and after a solemn ceremony, replace it by a newer model filled with promesses (along the CPU and RAM at least).

I also wish I was half my age, but not for the same reason as you propose!

"Probably" forget the driver route?

The CPU temp has always been within range.

The ethernet section is integrated with the MB. I could buy a card that plugs into an available slot. Thanks for the CMOS setting reminder.

Freezes occur when the machine is stressed and when at rest. This prevents me from putting some plan into action, when it would be a burden to have a freeze. I had one destructive freeze that occurred while my Thunderbird email program was logging all of my emails. I lost all contents of a buddy's sent folder. I used the equivalent files on the laptop, to replenish the folder.

A RAM test is a good start. As for the hard drives, how can 2 physical drives with different OSs fail at the same time, especially when the Samsung has only been a year in service? Other than HD Tune's surface testing and data collection program, how can I "fully" evaluate any hard drive? The secondary XP drive is still in the case. Although I have not booted to it in about nine months and it needs all kinds of updates, I could boot to it now and see if the freezes occur. It is an identical drive to the primary XP drive. Actually, it was part of a Raid mirror array that I had created when the pc was assembled. The RAID system failed and I started using it as a backup drive. All 3 drives are active during a startup, and I use all 3 as redundant backup locations.

 

[apairofpcs]

Now do your burn-in and see where those temps end up. It shouldn't take more than a couple of minutes to get to your maximum temperature. Make sure that the burn-in is running both cores at 100% using the task manager->performance tab.

Now, I don't think your problem is thermal for you said it occurs during network activity and that does not normally stress the CPU at all. Rather, this is just a quick test you might find interesting -- Your maximum operating temperature.

Based on your post, it sounds like my idea of using CPU Burn-In while doing a full AVG scan and a defrag. simultaneously would be enough to stress the CPU and raise it's temp. I have to open the program twice, one for each core. I always keep an eye on the Task Manager's Performance tab.

 

[apairofpcs]

Having read everything you and others have written, I'd probably just tell you to replace your motherboard and be done with it.

If the others had expressed your belief that the ONLY culprit could be the MB, then Newegg would have a replacement MB on it's way to me as we speak.

 

[ddrueding]

I'm with Merc on this as a likely MB issue. The only other thing I'd point to would be the PSU, but 90% motherboard / 10% PSU sounds about right.

 

[jtr1962]

If the others had expressed your belief that the ONLY culprit could be the MB, then Newegg would have a replacement MB on it's way to me as we speak.

In case it is only the MB, a cheaper route might be to look for used motherboards which support your processor on eBay, or perhaps some of the members here have a spare MB of similar vintage which they would be willing to sell. If you're going to buy a brand new motherboard, might as well spend the money on something current, and upgrade the CPU and RAM at the same time.

 

[apairofpcs]

Is anybody impressed by the fact that I have NOT had a freeze in almost 12 hours, while using my pc in safe mode the same way as I use it in GUI, with the exception being lack of internet access? Maybe I am not fully aware of the differences between the way all hardware in a system is being utilized in GUI vs. safe mode? Would somebody provide me with some clarity?

 

[BingBangBop]

I'm with Merc, replace the MB. I can't guarantee that it is the source but its quite likely.

It is not just that I've never seen a CMOS reset caused by a soft boot, rather it isn't supposed to happen. The CMOS is supposed to be continuously powered, in all circumstances. If the CMOS reset occurs, it is because the power went out or at least it went so low that the CMOS failed to retain it data. Even if you power the machine down, as long as the cord is plugged in the MB should still be supplying the CMOS with power. If you totally unplug the computer and thereby remove all power, the battery is supposed to automatically take over to maintain power. With the reset button, the power was never interrupted for the machine is still plugged into a power line. All the reset button does is short across two pins on the CPU and with that signal the CPU then knows to go to a set address in the bios to execute the code there i.e. reboot. It should never ever affect the CMOS.

When you reboot, the power draw on your computer goes up significantly. If the power supply can't maintain voltage or if the motherboard voltage regulation is problematic then the voltage to CMOS could drop to the point that it can't maintain memory i.e. a CMOS settings reset but even then the battery is supposed to take over. Those same problems would also show up as lockups or crashes if they occur while normal running. Really, I place the odds on that the MB is going bad.

It is all nice and fine to repeatedly scan using one AV product but if it doesn't catch it the first time it isn't going to improve with repeated attempts. If you want to do repeated scans, use different products between each because what one misses, another might catch.

 

[BingBangBop]

Is anybody impressed by the fact that I have NOT had a freeze in almost 12 hours, while using my pc in safe mode the same way as I use it in GUI, with the exception being lack of internet access? Maybe I am not fully aware of the differences between the way all hardware in a system is being utilized in GUI vs. safe mode? Would somebody provide me with some clarity?

Primarily what safe mode does is use default Microsoft drivers and not load any autostart items. As to how this affects hardware is that it won't exercise all the capabilities of custom drivers because the default drivers use the minimum common denominator. The best example I can think of are video cards for with the default drivers there is no hardware acceleration (everything is emulated in software and done at the CPU level) and with that comes a lot less power draw. Also, if there is no default driver that will work then Windows will operate as if the HW wasn't there

 

[Mercutio]

If the others had expressed your belief that the ONLY culprit could be the MB, then Newegg would have a replacement MB on it's way to me as we speak.

Well I just looked at this thread for the first time like 8 hours ago, but my intuition is that it's a motherboard issue just based on the variety and frequency of issues. It's relatively simple to slap in a different power supply and see if the issues go away.

It's actually relatively simple to test output voltage even if you assume the numbers you're seeing the motherboard BIOS aren't accurate, but full testing requires an oscilloscope and that's not something people keep sitting around.

Anyway, life's too short for motherboards that could be replaced and RMA'd.

 

[apairofpcs]

Now do your burn-in and see where those temps end up. It shouldn't take more than a couple of minutes to get to your maximum temperature. Make sure that the burn-in is running both cores at 100% using the task manager->performance tab.

Now, I don't think your problem is thermal for you said it occurs during network activity and that does not normally stress the CPU at all. Rather, this is just a quick test you might find interesting -- Your maximum operating temperature.

OK, more feedback for you. I just completed a 30 min. CPU Burn-In session, with an open window for each core. I confirmed 100% CPU usage in Task manager's Performance tab. Using Real Temp, the CPU cores before the burn-in were 40 C each. Near the 30 min. mark for the burn-in, the CPU cores were 50 C each. TJ Max. is 50 C each. The burn-in test completed successfully.

Along with the CPU Burn-In program, I ran an Ultimate Defrag. (Disktrix), an HD Tune surface test, a full system AVG scan (at the fastest speed using the most resources) and a few drive to drive copy operations.

Bottom line - does this convince you that the CPU is working properly?

 

[apairofpcs]

I'm with Merc on this as a likely MB issue. The only other thing I'd point to would be the PSU, but 90% motherboard / 10% PSU sounds about right.

Due to the amount of work involved in a MB replacement, I'm going to hold off on this item and pursue the diagnostic tests suggested by the others.

 

[apairofpcs]

In case it is only the MB, a cheaper route might be to look for used motherboards which support your processor on eBay, or perhaps some of the members here have a spare MB of similar vintage which they would be willing to sell. If you're going to buy a brand new motherboard, might as well spend the money on something current, and upgrade the CPU and RAM at the same time.

You know that I won't buy ANY high tech. device in used condition. But you made a good point about upgrading the CPU and memory along with a new MB. This makes sense in order to keep up with the fast pace of technology. These 3 items are the most important of a pc's core components. You'll be hearing from me when I decide the path I'm going to take.

 

[Mercutio]

Due to the amount of work involved in a MB replacement, I'm going to hold off on this item and pursue the diagnostic tests suggested by the others.

I strongly suspect that the amount of work and aggravation involved in replacing your motherboard will be lower than continuing to troubleshoot an issue that will in all likelihood lead to you replacing your motherboard.

 

[apairofpcs]

Primarily what safe mode does is use default Microsoft drivers and not load any autostart items. As to how this affects hardware is that it won't exercise all the capabilities of custom drivers because the default drivers use the minimum common denominator. The best example I can think of are video cards for with the default drivers there is no hardware acceleration (everything is emulated in software and done at the CPU level) and with that comes a lot less power draw. Also, if there is no default driver that will work then Windows will operate as if the HW wasn't there

So did my 12 hours of general pc use in safe mode without a freeze, not provide evidence that I don't have a hardware problem?

 

[apairofpcs]

I'm with Merc, replace the MB. I can't guarantee that it is the source but its quite likely.

It is not just that I've never seen a CMOS reset caused by a soft boot, rather it isn't supposed to happen. The CMOS is supposed to be continuously powered, in all circumstances. If the CMOS reset occurs, it is because the power went out or at least it went so low that the CMOS failed to retain it data. Even if you power the machine down, as long as the cord is plugged in the MB should still be supplying the CMOS with power. If you totally unplug the computer and thereby remove all power, the battery is supposed to automatically take over to maintain power. With the reset button, the power was never interrupted for the machine is still plugged into a power line. All the reset button does is short across two pins on the CPU and with that signal the CPU then knows to go to a set address in the bios to execute the code there i.e. reboot. It should never ever affect the CMOS.

When you reboot, the power draw on your computer goes up significantly. If the power supply can't maintain voltage or if the motherboard voltage regulation is problematic then the voltage to CMOS could drop to the point that it can't maintain memory i.e. a CMOS settings reset but even then the battery is supposed to take over. Those same problems would also show up as lockups or crashes if they occur while normal running. Really, I place the odds on that the MB is going bad.

It is all nice and fine to repeatedly scan using one AV product but if it doesn't catch it the first time it isn't going to improve with repeated attempts. If you want to do repeated scans, use different products between each because what one misses, another might catch.

I agree that under no circumstance in the running of a pc, should the CMOS lose it's settings, unless the CMOS backup battery is missing or dead. My battery is only two months old. The glitchy power failure even bypassed the positive role of the battery. How did it do that?