Weird Hardware problems

Posted on 2005-03-13
Medium Priority
Last Modified: 2013-11-08
For a while now I have been having some weird problems with my computer and seeing as I have formatted and reinstalled windows on a number of occasions I'm putting it down to the hardware.

I don't have any consistant problem but every so often when I restart or shutdown the computer, it freezes and I have to restart manually, when I shutdown the computer re-starts, whenever I play any game with high-end graphics after a while I get a serious error and the game crashes. sometimes my recycle bin shows that it's full so I open it and it's empty, then when I empty it it asks me if I'm sure I want to delete windows!!! normally I would put it down to a virus but as I have formatted and re-installed windows a number of times I don't think so...

I have replaced the processor fan, the graphics card, the memory and the harddrive and no joy...

I think it may be something to do with the graphics card over heating because if I open the case the graphics card is hot to touch.

Here is my spec:
CPU: Athlon AMD 2200
RAM: Corsair XMS 512mb
Hd: Seagate 80gg
Motherboard: Matsonic MS8137C
Graphics Card: ATI Radeon 9800 Pro

I would appreciate any suggestions...

Question by:liseperu
Welcome to Experts Exchange

Add your voice to the tech community where 5M+ people just like you are talking about what matters.

  • Help others & share knowledge
  • Earn cash & points
  • Learn & ask questions
  • 5
  • 4
  • 3
  • +4

Expert Comment

ID: 13529935
A couple of questions...

1.  When you formatted and reinstalled Windows did you do a Quick Format or Full Format?
2.  Prior to formatting the HDD did you zero the drive out with a disk utility?
3.  How old is your MB?
4.  Is your graphics card hot to the touch or is it the GPU that is hot?

A few suggestions...

1.  If you can get a temperature reading on the graphics card using a probe that may help to determine if this is the problem.  If you have onboard graphics you may want to remove your expansion graphics card and see if this helps solve your problem.
2.  If your HDD and MB support S.M.A.R.T. HDD monitoring - enable S.M.A.R.T. in BIOS and run HDTune from the following:


Author Comment

ID: 13530082
I used a full format.
I didn't zero the drive out. and what does that mean?
My motherboard is older than the flintstones!!! I bought it 5 years ago but I don't know when it came out.
it's my graphics card that is hot, the gpu has a fan over it so I can't touch it...
I ran everest and found that the cpu temp is around 41 and the tcpu temp is 55

personally I think that it is most likely to be the motherboard or the processor as they are the oldest things and also the only things I haven't changed. however, I don't have the money or time now to change them.

thanks for the suggestions, I'll try the hdtune thing.

Expert Comment

ID: 13530478
The temperatures you mention are nothing to worry about...

Have you tested your memory???  Virus, Temperature or memory problems would be the most likely culprits....

Formatting your hard drive will not neccessarily get rid of a virus though, it depends on the virus....   Yes if you have a virus, there is probably virus code on your hard drive, but it is also probably in memory as well...  If the virus is in memory, each time you format the disk it will simply copy itself back again and start to spread....

Different viruses act in different ways....  it is this variety that makes them so difficult to erradicate...

Reseting your computer will not clear your memory, it does not completely power down the computer....  It needs to be turned off to be sure that the memory is cleared...  But that does not necessarily mean it will get rid of the virus, it may already have re-infected your hard drive again anyway, lol....

If there is any chance that you have a virus, you need to get a reputable virus scanner.  Even if you dont, get one anyway....
Bringing Advanced Authentication to the SMB Market

WatchGuard announces the acquisition of advanced authentication provider, Datablink, with one mission – to bring secure authentication to SMB, mid-market, and distributed enterprises with a cloud-based solution, ideal for resale via their established channel & MSSP community.


Expert Comment

ID: 13530511
Hi liseperu

This sounds like a heat or power issue

Have good airflow through case ?

Psu with decent 12 volt performance, prefereably > 15 amps(sticker on PSU)

This could also be caused by bad capacitors on MOBO especially if 5 years old.
No bulging or oozing allowed, tops completely flat

LVL 13

Expert Comment

ID: 13531650

Test the memory, use Memtest86

This could very well be a power supply problem (or power supply inadequacy -- power supply too small).  The 9800 Pro uses a ton of power (which largely gets converted to heat).

I'm also very suspicious that the motherboard is too old to properly support the 9800 Pro.

I'm fairly confidennt that your problem lies with one of these three things.
LVL 12

Expert Comment

ID: 13531743
Silly question, but what game are you playing?  I've seen the same behavior, the one where it asks if you want to delete Windows, hard to remember what the cause was.

When is the last time you replaced the CMOS battery?

You said "I have replaced the processor fan, the graphics card, the memory and the harddrive and no joy..."

If you've replaced the graphics card, why would you still think it's that card?

If you have a tower, run with the covers off, maybe put a big fan blowing on it; if it's a temperature related problem, the cooling of a large fan will bring it down enough for it not to fail, usually.

What is this "Serious error" you get?  If you can't do anything else at that point, try ALT+PrintScreen and then paste to a picture viewer or word or something, save, then shut down.

When's the last time you did a manual scandisk and defrag?

What is your Virtual Memory set to?  Try 750 megs, instead of letting Windows determine it.  That should get optimum performance and cut out a lot of unnecessary disk reconfiguring by Windows.

You said "every so often when I restart or shutdown the computer, it freezes"

Usually, that's an indication that while saving the environment, Windows encountered some form of disk or memory error.  Reason for scandisk and defrag.

Did you partition the 80gig disk yet?  I see no sense in one giant disk of 80 gigabytes, not really.

The power supply could also be getting ready to go.

Check back after trying some of this stuff.

Author Comment

ID: 13533246
ok, I'll answer all of the above questions:

1. I checked the memory, and there is no problem, as I mentioned before I have tried different memory sticks and no change so I guess it isn't the memory.

2. I have used Panda, and bit defender to check for viruses, Panda also scans the memory, neither could find anything, I also have adaware and spybot which I run regularily

3. I checked the capacitors according to that website posted, but I can't find anything resembling the problems they described. everything looks normal.

4. I thought  and still suspect that it is a temperature problem so I did check the airflow, it seems ok, and I ran it with the case open and put a fan on it from outside as suggested, no difference...

5. I can't say I fully understand what is written on the PSU but it is 300W, and it seems to be saying that at 12v it uses 13A.

6. The serious errors depend on the game, but they are all exceptions.

7. The games I am playing are: Command and Conquer generals: Zero Hour, Diablo 2, Pro Evolution Soccer 3, FIFA 2004.

8. I have never replaced the CMOS Battery

9. I don't think it's the card itself, but I thibk the card could be a part of it, either that it needs more power than available, or better cooling.

10. The disk is partitioned into two 40gig partitions.

11. how would I test the power supply without buying a new one?

ok, I just remembered another problem, non of the dos commands work anymore, for example, ipconfig, chkdsk, they all say "" is not recognised as an internal command!! strange huh?

I tried to run scandisk but when I restarted windows, it said system performing scandisk, then "scandisk has benn cancelled, re-starting windows"!! I tried a couple times but always the same result. I checked and all the dos commands are still in the system32 folder.

I'll try changing the virtual memory setting but I don't think that it is a problem with lack of memory....

thanks again

Expert Comment

ID: 13533988
When you formatted and reinstalled windows (98SE?), did you boot up from the installation CD?  If not you may well have some kind of virus.  If you did then it is unlikely.
One of my kids has the same model AMD and had similar problems using win 98.  An upgrade to XP solved most of the problems.  Another problem was the CPU speed setting in the Bios.  When we set it down to 100 then the games stopped bombing.  Note that the Athlon chip is optimized for Win XP and as its gets older gets more quirky with the heat.  Also make sure that the AGP aperture seting is set to max for your video card.  This may help in the gaming problem if it is not set correctly.

Dr. Z

LVL 13

Expert Comment

ID: 13534396
Your power supply is inadequate.  You need AT LEAST 350 and preferably 400 watts and ideally about 20 amps for the 12 volt supply.  You have 300 watts and 13 amps.  No way that's enough for the 9800 pro.

I'd recommend that you get a quality power supply (Enermax or similar) rather than a cheap one.
LVL 12

Accepted Solution

GinEric earned 1000 total points
ID: 13535779
This "is not recognised as an internal command!! "

is bath path statement and/or non-executable path.

User the full path to see if commands work: C:\Windows\system32\

from a DOS prompt:


or CD to C:\Windows\system32\

and issue any DOS commands.  They should work, and if they do, it may be that one of the games has altered your DOS path, a common occurrence with some games.

Can your restart to a DOS prompt?  If so, you should be able to run scandisk from there.  Set to automatically fix problems.  This effectively builds a list of bad blocks/sectors and clears them from Windows usage.  Then reboot to Windows, no DOS prompt, and do the defrag.

If scandisk still cancels, you have disk block problems; not that there are actual problems, just that the OS and BIOS are having a hard time with such a large disk.  Again, common on a lot of Windows systems unless you use extended large blocking for disks, as in Western Digital and others EZ Bios and so on.  Problem here is you have to format the disk with WD or whatever floppy, install EZ Bios, then disable it as it conflicts with most Windows OS, but is needed initially to partition the disk.  Older games, and some new, run in DOS environment, taking you back to DOS disk size limit, 4 gig and under, Win 95 and Win98, I think, had a limit of 8 gig, and/or something like 32 gig, so even 40 gig would be a problem unless the games run outside of DOS and in XP

I can't seem to find which OS you have in this post, but unless it's XP, you have a disk size limit without some correcting software.  Scandisk cancelled is a good indication of disk size problem.

Back to the CMOS battery; can we assume that it has not been changed in 5 years?  If so, go buy a couple and change it, keep the second as a spare.

Power supply : you are on the edge of power consumption, it's true.  If you get another one, please get one with an ON/OFF Switch on it!  Depending on power management to turn the power supply on and off, and thus the computer, doesn't always work.  I return all power supplies that don't have an ON/OFF Switch as a sign to manufacturers that the switchless power supply is unacceptible.

Overclocking: never do this, it will burn out the microprocessor, regardless of what gamers and other say, it runs hot, and shortens the life of the rated processor.  This is very technical and most gamers and computer consultants have no idea why there is a rating for the speed, in simple terms, it's what the batch of silicon for that chip production run can handle before encountering heat problems.  Same for overclocking the video; it is actually possible to cause the monitor's CRT to implode by overclocking it.  Amateurs should never be allowed to adjust some settings!


"4. I thought  and still suspect that it is a temperature problem so I did check the airflow, it seems ok, and I ran it with the case open and put a fan on it from outside as suggested, no difference..."

Absolutely right on the money, good reasoning, good logic.  

Technical approach, not recommended unless you have a lot of confidence in what you're doing:
Now, way to find intermittent heat problems.  Get a can of aerosol'ed freeze spray, use with extreme caution.  When the computer is running, say for a game, you can spray various components.  When the bug goes away, you've found the component with the heat problem.  Care is needed though, such as, don't freeze the fan on top of the microprocessor, spray at bottom.  Video card, spray the main chips.  The technique here is to see which chips are hot, usually by a quick touch, spray them, touch again to see if they cool, and observe system; if no failures, last cooled chip is overheating.

Same can be done to power supply.  At lower temperatures, the power supply puts out better regulation, provides higher power without failures, and so on.  Usually, just spraying inside of one of the holes in the cage is enough.  No need to ever take apart a power supply.

Don't spray CMOS or other obviously alloyed products.  They shrink and break contact, and can react to the cooling in bad ways.

Don't spray the hard drives.  They don't like rapid changes in temperature.

Try this as the last resort only!

Intermittents are the hardest problems to find.  It is a trial and error process and can take a lot of time.  But there is a method.  In the end, if you find no hardware problem, it has to be a software related conflict.  You have to accept that conclusion, unless you have a spare computer to start swapping in things like the Motherboard, power supply, video and other cards, etc., to narrow it down to the specific failing device.  Sometimes that doesn't work because of combinations of power supply/motherboard power consumption, etc..

So, you have to really use your brains on these problems, logically eliminating suspects one at a time, and make an extrapolated, or best guess, as to what the real problem is.

You seem to have focused on the video card, yet you swapped it with no difference, this should tell you that it is most likely not the video card.  Perhaps the software, drivers, etc., but not the card itself.  Don't be afraid to draw the logical conclusion here.

If the power supply truly can't supply enough power, you may have what is called a "pinching" problem.  That is, at peak demand the current gets "pinched" off to protect the power supply itself.  This causes voltages to drop, and subsequently intermittent failures.  Cooling of the power supply or computer allows for power consumption to drop, even for the same demand, because of fact that as temperature goes up, current goes up, and vice versa, for the exact same set of circumstances.  So, cooling makes the current go down, which makes the power demand go down, so the computer works at cool temperatures.  This is also the basis of superconductivity and supercooled computers.

Your conclusions thus far:
1. - good
2. - good
3. - good, capacitor failure is rare unless machine off for a year or two.
4. - 330 Watts may be low for ATX.  Power: IE, generally, 12x13=260W, for 12v regulator, add +5V times current rating, and -5V times current rating, or, just look at the fuse and multiple by 110 or 120.  Example, 3 amp fuse times 110 Volts = 330 Watts, actual power used from AC line.  This is the rating of your Power Supply.

6. - vague, exact "exception" you get.  Should be in eventviewer.  GPF's mean memory access problems, etc., you need to state some of the "exceptions" they will point in right direction.
7. - good games!  I'm strictly a Doom player myself.
8. - then try replacing the CMOS battery.
9. - Very good conclusion.
10. - Acceptible partitioning, however, consider a third partition just for games.  Not sure, but I don't think you should need more than 7 gig partition for games, eliminating the FAT32/NT barrier, i.e., 8 gigs for FAT32, in general.
11 - That's a hard one to answer.  Usually, you have to eliminate some of the heavier loads, which is not always easy to determine.  Memory chips use a lot of current, but you can't remove them a lot because it defeats the purpose of seeing if it's a lack of memory problem.  You have plenty of memory.  The memory should be at least the size of the operating system.  Doubling it allows for all programs to have sufficient memory to make all calls local to RAM.  Since XP is about 200 meg, 400 meg is more than sufficient.  There will be no increase in performance at higher volumes of RAM.  1 gig won't make it run any faster, but 1 gig will raise your disk cache size, causing more disk caching, and thus slowing down performance.  Many argue about more memory vs less memory, drawing conclusions that more is better.  That is not how it works; it works on "optimum" memory.  Optimization is a technical term for where the rate of access with respect to the rate execution is at its peak value [a Differential Equation]; rates are in the realm of Calculus, not simple mathematics.  And the "more is better" argument does not hold up in that realm.

Don't worry about how long my response it, I'm writing technical manuals while I do this, it's benefitting me too.

Let's go back to the very beginning for a moment.

You've tried all kinds of things and you suspect a heat problem.

You admit eliminating the video card itself, even though it seems hot.

Problems show up running games.

Computer freezes up on shutdown.

These all point to a load problem, which generates heat, causing a self-feeding spiral of more heat, more current, until power demand exceeds supply.

Whether or not during that load a bad sector write is getting out to disk, as in Shutdown's attempts to save off memory, is uncertain, but seems more than likely.  As pointed out by Scandisk canceling itself.  The one exception to this is that the OS is not allowing you to run Scandisk from the OS, so you revert to a DOS prompt to run it [which I figure you have yet to do].

Games are load intensive.

Load lowers voltage, even to critical components such as CMOS and hard drives, interfaces, etc..

You're at the crossroads of either a heat problem or a power problem.  The simplest next step is to get a better power supply.  Also, CMOS batteries eventually do not recharge, a simple fix with another CMOS battery.  This too is a form of power problem.

Conclusion: unless you've got a real fluke of a problem, you're at about 90% indication that it is inadequate power.  Replace the Power Supply next.  400 Watts is good, 500 Watts would be better.

And please try to get a power supply with an ON/OFF Switch, to discourage this trend toward cheap and unsafe power supplies in the computer industry.

Author Comment

ID: 13537467
wow, that's really a lot of info GinEric, thanks a lot. I just ordered a new PSU. Hopefully that will solve the problem... I'll replace the battery too.

I'm running windows XP service pack 2. so I can't restart to a dos prompt. is there any other way I can run scandisk?

I did as you said and typed the full path to the dos commands, they do work that way, but on restart scandisk is still cancelled

The exact exceptions I get aren't very exact.

when Zero Hour crashes it says: "the system has encountered a serious error, this may be due to a hardware fault, overheating or a fault in the game" also if I am playing a skirmish with a lot of computer opponents set to hard difficulty, the game has periods of slowness.
when Battle for middle earth crashes I get the following exception: "EXCEPTION_ACCESS_VIOLATION The thread tried to read from or write to a virtual address for which it does not have appropriate access"

also in battle for middle earth, the film sequences have blotches in them, every so often...

In Diablo 2 I just get a windows error saying "the system has encountered a serious error" same in PES3 and FIFA

Also one other weird thing (though I think this is irrelevant, but you never know), the initial screens when the computer is turned on, the BIOS screen and the screen when windows is loading are displayed halfway down the screen!! howeverl, when I arrive at the screen to enter my password, everything is normal...

thanks again

Expert Comment

ID: 13538599
Since you are using SP2 then I agree that the problem is most likely temperature.  As I said, as the system gets older it gets more quirky and sensitive to high temperature.  Try and get a really good cooling system along with the power supply.  If you still have a problem then a new Microprocessor and board are called for.

Dr. Z
LVL 12

Expert Comment

ID: 13540818

"the initial screens when the computer is turned on, the BIOS screen and the screen when windows is loading are displayed halfway down the screen!!"

Indicates your Vertical Refresh and perhaps synch is set wrong.  It is being squeezed into the lower half of the CRT.  That's another problem for another time, maybe next time.

Actually, XP will boot to DOS, but under very specific circumstances and must be used with extreme care and caution.  You can set BIOS to boot first from floppy, then IDE 0 and use a DOS setup disk, simply use F3 twice to get to DOS prompt.  This is only recommended if you really know what you are doing!

Holding down Delete key during boot up will get you to BIOS.  Only use if you know exactly what you are doing!

Otherwise, holding down F8 key during boot up, until after three LED's on keyboard blink, will get you to the safe mode menu options.

SELECT Safe Mode with Command Prompt

You are now in DOS mode with Windows XP.

"cd .." will up directory one level

In first DOS, you should be in documents and settings\my documents

"cd .." twice will take you to C:\

cd Windows
cd system32

will take you to c:\Windows\system32


will run Check Disk.

chkdsk will issue a report at the end telling you if it found any bad sectors.

Sorry, I said scandisk above when I meant chkdsk

You can't run scandisk from DOS because it is incompatible with NTFS white space direcotory structures and will rename all files to their tilde values.

Scandisk on XP must be run from Disk Manager.

Start | All Programs | Administrative Tools | Computer Management
In Computer Management console:
right click over Disk Mangement, then right click over C drive, Properties, select tab Tools and click button Check Now
check radio buttons Automatically fix file system errors and Scan for and attempt recovery of bad sectors
then hit start.

There you will get the long message about how Windows can't do it because it needs some exclusive access, and the question "Do you want to schedule this disk check to occur the next time you restart the computer?

Click yes.
Click OK on properties.

In main windows of Desktop:
Hit Start , shutdown, restart

Computer will reboot.

After a short while, you should see a blue screen with "Checking file system on C:" blah, blah, blah. . .

You will also see the progress of the scandisk.

Be prepared for lunch or dinner!  Maybe a night out!

Anyway, when it gets done, you should have a clean disk drive with no bad sectors.  Note that chkdsk also runs here, so you don't need to do it from the other DOS method.

One of the reasons we partition users big disks down to reasonable sizes is because of the time it would take to scandisk something like a 160 gig hard drive, etc.. That, and we can install alternate OS on other partitions and get their machine fixed a whole lot faster than with one humongous 160 gig partition.  Also, their data is stored and programs are usually run off another partition so that if they absolutely crash the first partition, none of the data is "ever" lost!  We can get it all back immediately.  These are just things for people, companies, and admins to think about.

Access violations are more common on NT than the GPF's were on Win95.  It means that the program tried to write into a system area; which is why a lof of games have a bad reputation for crashing systems, they think in DOS terms, not in NT terms.  Usually, a game loses it's pointers, and then tries to write to either address zero, or, to a location outside of its stack.  This gets all too very technical to explain, so just note that games are notorius for this behavior, check their compatibility with NT and XP.  Otherwise, run them in their own virtual environment.  Windows has an option for this using the "run" command from the Start menu of the Desktop, as in Start, run, and select the program by path, and the "run in virtual memory" or whatever.

After the scandisk completes and you're back up, then run defrag, from the click sequence stated above.

You can skip a lot of steps if you start Windows Explorer, right click on the C drive, select properties, and then either scandisk or defrag.

In the end, you've done everything you can.  If it turns out to be the games causing the errors, at least you now have a system you fully know.  The one thing is that game related errors do look an aweful lot like temperature problems, which is why most techs hate doing all this troubleshooting only to find out it's a gaming problem.

This problem should be solved either by all of this work, or when that new PSU gets there.



Author Comment

ID: 13542513
I finally managed to get chkdsk to work last night by typing its full path. But it didn't find any errors at all!!! I did a defraged the drive and at the end it told me that there were some areas that it couldn't defrag, otherwise everything normal!!!

It's ok you can get technical, I'm in my final year of a software engineering degree and I have just completed my own neural network in C++, which uses a lot of pointers!!!! I don't think it is the games causing the problems, unless there is some specific incompatibility issue with some of my hardware because I also have a laptop and they all run perfectly there, a lot slower but no errors at all... Thats a P4 2.8 with 768mb of ram...

<< Dr Z - Since you are using SP2 then I agree that the problem is most likely temperature

just out of curiosity, why does using SP2 affect the temperature?

also seeing as I tried the games even with an open case and a big fan blowing directly on the system and I got the same error, is it still possible that it is a temperature issue?

thanks again all

Expert Comment

ID: 13543517
This is unlikely to be a CMOS batery problem...  If the battery fails, it simply means that the bios will be unable to retain settings when you turn the power off...   A simple way to determine a battery dying is to look at the system clock in windows, if it keeps displaying the wrong date/time after you have turned off the computer, the battery is gone....   The bios being reset after turning off the computer is unlikely to cause the problems seen, as it would simply redetect your hardware and run in low performance mode to all intents and purposes (unless you have a real dinosaur on your hands)....

If I were you???

Download and install MadOnion, the trial is free.  This is a benchmarking tool: http://www.futuremark.com/
Make sure that DirectX is upto date on your machine....

I would then strip the computer right down to its core parts,:

Motherboard, CPU, 1 stick of Memory, HDD and Graphics card.  

This will remove a lot of strain from the PSU, testing that theory...
It will reduce the heat generated by the system....
It will remove a large number of possible hardware related issues....

Run the benchmark....  Does the error still occur???

If not, add other memory, still working ok????

Add your CDROM drive, try one of your games.....

If it still works fine, add the rest of your hardware back in, an item at a time, testing after each device....  Otherwise you will not be able to identify the culprit.....

If none of this has made any difference, you are left with a number of possible options....

You have a faulty hardware component, which is either Motherboard, CPU, PSU, Memory, HDD and Graphics card or PSU...   Or Drivers....

You can simply eliminate Memory, HDD, Graphics by simply switching these as you have already indicated that you have spares...   Do this whilst the system is still stripped down....

That leaves MBD, CPU, PSU....   Or Drivers....

You will be able to eliminate the PSU when your new one arrives, or you could borrow one from another computer case just for testing....

That will leave MBD, CPU, Drivers....

The drivers, in windows, open the control panel, go to system, select the drivers and "uninstall" them, dont just simply remove the card and try with another, if the drivers are the problem and they are still on the machine??  See if windows will autodetect your card, try it with the drivers it installs....  They may not be as good or up to date as the ones from the manufacturers website but if they work it gets rid of the problem it will indicate that the problem is with drivers, if the problem still occurs, it may indicate that drivers are NOT the problem....

Is it working???  

By now you have erradicated all but the MBD & CPU...  Testing these is a bit more difficult unless you have another machine you can canibalise...

I would suggest taking out the CPU first by putting it into another machine that will support it....  If the CPU is damaged, putting it into another machine should not damage anything, you shoukl simply see the same symptoms appearing on this machine too...

I would be cautious about putting a CPU that you know works from another machine into your potentially damaged motherboard, as it could damage the good CPU....

If the CPU tested ok in the good machine and it functions without any signs of a problem, I personally would probably take the chance although using extreme caution, but you need to make that decision yourself....  

LVL 12

Expert Comment

ID: 13547204
This is a disk error:

"I did a defraged the drive and at the end it told me that there were some areas that it couldn't defrag, otherwise everything normal!!"

The only areas it can't normally defrag are the pagefile and system blocks.

You put a big fan on it, so let's consider the temperature eliminated for the time being.

Stripping down, as suggested by Kipski is fine, if it works though, it indicates a power supply problem.

CMOS is subject to trickle current, a dip in power can actually drain the CMOS, and the BIOS is constantly accessed at the hardware layer.  Vibrations also have an affect.  Example, if you thought the CMOS wasn't used during runtime, trying pulling it out during run and see what happens [just kidding, don't do that unless you have a machine you care whether or not you obliterate].

At this point, we know you have a clean system, except maybe for those undefragable areas.  Are you sure you didn't get a bad block report on them from defrag?  XP quarantines, or tries to quarantine, bad blocks and areas on disk, not always successfully.  Example, if a system code segment is in one of those areas, it can't replicate it because it's already corrupt, but it can't mark a system segment as bad either; result, XP dilemna and intermittent failure.

But it is more than likely that the games are doing this.  As I said before, games are "notorius" for this behavior.

You have to run most games "in their own memory space" which is to say a virtual memory area.  If the value of EAX Register is a pointer and, for whatever reason, that value is zero, then the reference is 0x0 plus the offset from maybe EBX or the operator's address couple, usually less than 0xFFFF, a system area.  65,536 is called a mod of code.  Access to the first mod is always restricted to the Operating System only.  This is where most games fail.  This is where viruses try to hijack a machine.  Knowing this, Microsoft and others have set the area to "protected storage" and it is off limits to all.  Games, which are designed more for arcade games, and then ported to Windows after a few years, do use this area, sometimes the developers forget to adjust for Windows sales after the arcades have made their money.

I'm belaboring the point a little because I've also worked on a few games and arcades in my time.

If after you strip down it works, consider power only if the games run in that mode.  If you still get the errors, even without the games, the error is pointing to power problems, RAM uses the most power/energy and this is where most upgraded computers fail, at the point they start getting around and above 512 meg of RAM.

I don't like removing RAM once it's installed myself because merely handling it can give similar errors thanks to static charges that are inevitable.  But seating of newly installed RAM is also sometimes a problem.  An example would be the Address Bus with a poor connection; same symptoms as you are getting, Address bus fails to load, points to 0X0000, and you get an exception.

Poor power regulation and not solid contact between lans of RAM strips and seat also affect dynamic refresh of the mostly planar memory.  These chips are very sensitive and often running on only a volt or so; bad contacts and oxide layer buildup between the lan and connector show up as intermittents with the same symptoms of intermittents because of the extra current required to "jump" the oxide layer or not solid contact.  These symptoms are exactly the same from mainframes down to PC's.

Some gamers don't even know what a long pointer is, or a virtual or global pointer, so the games are designed as if all memory is basically in RAM, which it may not be.  There is also a latency problem for games when the data is not local, some games don't allow for the wait state.  Gamers have a lot to learn about multitasking, multithreading, and parallel processing, Linked List Lookup and Vector Mode.  And lastly, that latency problem leads to a race problem; the pointer is assumed to have been loaded within a certain wait time, however, the disk takes longer than this period, the game proceeds anyway, and the data is not there, i.e., the pointer is zero.  The game crashes with an exception, or worse.

This all would make your "handle" invalid, i.e., a value outside of allowable range, normally a zero offset with no global, virtual, or otherwise relative stack pointer or descriptor within the limits of the stack of the game.  Stack Underflow or Stack Overflow [not unrelated to or unlike a buffer overflow or buffer underflow].

SP2 increases the load by doing more processing over time interval t.

Increase in load ==> increase in current ==> increase in heat==> increase in probability of failure.

Engineering eh, do you recognize any Power Series or "as x==>m, y==>n" limits in what I'm writing?

In conclusion:  the big fan didn't make a difference, so no fan is going to make a difference.

You changed every component there is, no difference, EXCEPT the power supply; not always easy to just up and do.

Games or power supply, which do you think is more than likely the cause?

Expert Comment

ID: 13547424
Windows servicepack 2 ahem

Then you need to update drivers for Motherboard, and GFX in fact for all your components

13 amps on the 12 volt rail is a bit low, but should do ,if you are not loaded with other 12 volt consuming hardware - drives and fan's.

The 12 volt line solely powers your CPU And GFX.

here is a link to your GFX power need


so that means 5 amps to CPU and 5 amps to GFX, when gaming, and then some for fans and drives. So only if your PSU is very fit it will fulfill the needed power.

LVL 13

Expert Comment

ID: 13547633
Re: "The 12 volt line solely powers your CPU And GFX"

And disk drives and optical drives and fans and (if he has them) "lights" and RS-232 ports and ........

Expert Comment

ID: 13549065
If you read my first suggestion, I thought that perhaps the problem was with windows 98.  You did not say that you were using windows XP when you first asked your question.  So, since you are using XP then it is impossible that it is a win98 problem.  Ergo, more likely temperature.  If a better cooling system does not help, then perhaps (to repeat myself) the Microprocessor or the board are faulty.  In which case, you are out of luck.

Dr. Z

Expert Comment

ID: 13549290
Hi Watzman

Thanks for pointing out misunderstandable sentences

re: "13 amps on the 12 volt rail is a bit low, but should do ,if you are not loaded with other 12 volt consuming hardware - drives and fan's.

The 12 volt line solely powers your CPU And GFX."

The last line should be intrepreted as follows: the 12 volt line is the main source providing power to your CPU and GFX, each consuming "worst case" 60 watts ~ 5 amps at 12 volt.

Author Comment

ID: 13553328
I just got and installed my new power supply, command and conquer doesn't crash where it did before and I played diablo 2 for about 1 hour without any problems so i guess it was the power supply,

I'm going to accept GinEric's answer because he seems to have done the most work and his suggestions seem to have solved various other problems I was having too..

thanks a lot everyone.

Featured Post

Technology Partners: We Want Your Opinion!

We value your feedback.

Take our survey and automatically be enter to win anyone of the following:
Yeti Cooler, Amazon eGift Card, and Movie eGift Card!

Question has a verified solution.

If you are experiencing a similar issue, please ask a related question

On Beyond Tools A conversation I recently had with the DevOps manager of a major online retailer really made me think about DevOps monitoring tools (https://www.onpage.com/devops-incident-management-tool/). The manager and I discussed how sever…
Arrow Electronics was searching for a KVM  (Keyboard/Video/Mouse) switch that could display on one single monitor the current status of all units being tested on the rack.
Do you want to know how to make a graph with Microsoft Access? First, create a query with the data for the chart. Then make a blank form and add a chart control. This video also shows how to change what data is displayed on the graph as well as form…
In this video, Percona Director of Solution Engineering Jon Tobin discusses the function and features of Percona Server for MongoDB. How Percona can help Percona can help you determine if Percona Server for MongoDB is the right solution for …

777 members asked questions and received personalized solutions in the past 7 days.

Join the community of 500,000 technology professionals and ask your questions.

Join & Ask a Question