Hardware Problem Developing In An Old Computer

Computer Help and Discussions Hardware
1

At least I think it is a hardware problem. I can not see any pattern of circumstances that might be causing it. Maybe someone on here will recognize whats going on.

This is a ~ten year old win xp desktop which has performed without any issues for at least 6 or 7 years. That last issue was a malware infection, fixed by an OS reinstall. It has OEM drivers and hardware, except a newer hard drive, about 4 years old now.

What happens is that from time to time this computer just sort of locks up. Sometimes (rarely) once or twice a day, usually not for weeks at a time. Has been going on for several months. When it does, the windows and the system tray stop responding to the mouse or keyboard. The mouse and keyboard are OEM, PS2 type.

There is an odd circumstance, probably significant but I don't know what it means. If I have Process Explorer by Sysinternals running at the time of the lock up, it's window will still respond normally to the mouse and keyboard while the other windows will not.

Sometimes but not usually, after a restart check disk will run automatically, always on D: drive, the recovery partition. Check disk always reports that there are no bad sectors.

It doesn't matter how much load is created by Firefox or any other combination of softwares.

Overload can be created that will slow it down, but not lock it up. Also, sometimes it locks up while using very little resources.

Doesn't seem to matter which or how many website logins are active in tabs.

Doesn't seem to be a heat issue.

One more factor that may be important. I always have the C: drive running virtualized after updating Avast & Spyware Blaster. So any malware EXEs that may have landed there would be gone. Same for any other changes that may have been made.

Avast scans come back OK. Haven't done any other scans for a while, but there has been no indication of malware.

I'm pretty sure this isn't malware, so you can cut loose with any advice you please. :)

Thanks in advance for any suggestions.

2

do you have virtualized your cdrive the last 6-7 years too?

3

Just to add to Triums question, which "virtual" software are you running?

I know you've used PowerShadow for a long time and like me, with no issues, but I remember a little while ago you mentioned another one which was new to you. I can't even search the forum for it as I can't remember it's name.

Did you go down the road of that other program?

4

Damn login, where to start? :)

Talk about a needle in a haystack.

Cutting right to the chase, one would have to question the RAM.

If you have some lying around (who doesn't right) or can afford new ones or think the rig is worth the time/cost/effort than I'd be starting there.

I was going to mention heat, but you have gone down that rabbit hole by the sounds of it.

As to your 'no indication of malware' - well, if it was good malware, it shouldn't really be advertising itself should it ???!

So it'll be worth scanning the living bejeezus out of it.

Next as Trium as hinted, is it a VM related issue? Only using the PC in 'real mode' will identify that.

And therein lies your path - running the PC in it's most basic mode, and adding in one one thing at a time until it fails again.

Like starting in Safe Mode and running the rig for a while.

Replacing RAM and running it.

Not using VM software for a while.

etc, etc, etc.

On the good news front, it sounds like it is locking up frequently enough that you'll know in short order if a change hasn't helped.

Also remind yourself as to the age of that well-loved unit and the fact you have been a very good boy all year.

Don't you think Santa should bring you a new one this year? :D

5

I had the random lockup as you described Login123 in August 2016 (started a few months before though), the source of my problem was the antivirus software (Avira Free Antivirus) which they kept feeding program updates to on XP even though they ended support of it one year prior and even though the newest installer won't run on XP. Removing Avira solved the problem.

Your issue could be software-based or hardware-based. I'd run a memory testing program like Microsoft Windows Memory Diagnostic, or one that's included from the PC manufacturer. When I had faulty ram in Dec 2015 none of the memory testing programs found a problem, it was instead Window XP itself that detected bad RAM during boot which it promptly halted and displayed a warning during the boot process.

Edit:

Note that when testing RAM you need to test one stick at a time, that means remove any other sticks and test them individually one at a time - then try to boot/start Windows with them individually.

6

OK, well, it isn't gonna be easy.

I had hoped someone would jump in with "Hey, I have seen that and it is caused by (name of glitch here), and the fix is soooo easy (name of fix here). :P

As to the virtualization questions, (a very good place to start the investigation inho), it works like this. Each time I start the computer, I delete most history files, connect to the net, update Avast and SpywareBlaster, disconnect from the 'net, and start Powershadow. From there on, C: drive is "virtualized", meaning it runs in unused space on the HDD (I think that's how it works). Then I can get on the net, test most softwares, do pretty much anything I like, and when the computer is restarted all changes to C: drive are gone. DennisD uses Returnil for the same purpose. This setup has worked flawlessly for years, one reason why I tend to suspect hardware.

Dennis, The other virtualization software was Shadow Defender, used for win 7 since Powershadow doesn't work for anything later that win xp.

In the next day or two, I'll do this:

- recheck the heat issue using Speccy and a proprietary HP checkup software.

- Check the RAM ... is there a software for that?

- run a boot scan with Avast.

- run an ESET scan, and reinstall MBAM and scan with that.

A curiosity: why does the GUI for Process Explorer continue to work after the main desktop and systray have frozen?

7

A curiosity: why does the GUI for Process Explorer continue to work after the main desktop and systray have frozen?

I found that fact very strange - when a PC locks - it locks all.

Never seen one 'half lock' before. Which would suggest it's not really locked.

In such a circumstance can you Ctrl+Alt+Del and click File, Run New Task, and type in explorer and click OK and start a new explorer shell?

Has there been any software changes?

To follow your thought on the fact it's been running well, I'm wondering if the balance has been shifted due to a software update.

As to heat testing, I've found no need for software to show values, simply opening the case and in the worst scenario, redirecting a fan into it suffices for that theory elimination.

8

Andavari, I just totally missed your post...Duuhhh.. It answers the RAM question. Thanks. I'll get that software and test the RAM.

When the computer locks up, it does not respond to the keyboard or mouse unless the cursor is in the Process Explorer window.

The cursor will move around on the screen, but clicking will not make anything happen except within the Process Explorer window.

Don't remember if the shutdown option in Process Explorer will actually shut down the computer. Don't think so, will check next time.

One more clue maybe. Sometimes during a restart, the screen briefly displays multicolored confetti, just a quick flash. It might do that every time and I just don't see it because the monitor is not up yet, but I don't think so.

Maybe a failing graphics module on the motherboard ... seems like it would affect all windows.?.?

The idea that it's not really locked is a bit worrisome.

9

It has an integrated graphics card, not an add-on one?

Next time it 'locks' up, touch the heatsink on the grahpics chip and see how hot it is. (usually bottom right of motherboard)

Toasty warm is fine, "Ouch, I just got a burn blister" is not so fine.

What's the dust build-up within the chassis like?

10

One more clue maybe. Sometimes during a restart, the screen briefly displays multicolored confetti, just a quick flash. It might do that every time and I just don't see it because the monitor is not up yet, but I don't think so.

Display corruption during startup can in some cases mean a failing graphics display card, and if faulty with a still operational computer the display can look very discolored or multicolored as you've described it (I've had that happen with 3 graphics cards that were starting to fail) - that's if you're lucky enough for it to give you a very obvious clue like that because in some cases you won't have any display whatsoever and Windows may fully load but you can't see anything.

When my last graphics card died it took out a stick of RAM with it, or vice-versa perhaps the bad RAM took out the graphics card I don't know which way it happened.

11

This time when it locked up i took the side off and ran mta's highly sophisticated test. Touched the Graphics card. :P

The graphics card Is the hottest thing on the motherboard, feels hot to touch but not blister hot.. Everything else is about room temperature.

Dust buildup is minimal.

Right now, while all is working, Speccy and other system tools show temps normal, nothing shows more than 105 deg f, 40 deg C.

RAM shows at 4 gb, which is right.

Any suggestions about a graphics card for win xp? I would spend a bit to save old HAL here. Will try a graphics card first.

I'll go looking on the net but maybe some of you guys have experience?

Thanks a million for all the replies. You guys are the best.

12

I assume you are chasing an AGP graphics card for that mobo, or is it sexy enough to have PCI-x slots?

If AGP, can you still buy those??? (I can't from my retail parts supplier)

If PCI-x, is it a x1, or x4, or x8 etc type of slot? In other words, the mobo age may be a limiting factor.

Also, I'm concerned that even if you do go down the add-on graphics path, I'm wondering if the on-board card still may get hot.

I've never tested (felt up :) ) the heatsink on an on-board card when an add-on card is used.

As to the dust, even minimal can add a few degrees or even cause shorts when the dust gets moist enough.

Wouldn't hurt to take the rig to the shed and hit it with an air compressor. (substitute shed and air compressor with whatever works for you)

13

If AGP, can you still buy those??? (I can't from my retail parts supplier)

I looked all over the net from NewEgg to other popular sites, ended up finding what I needed on Amazon. Not the newest greatest technology, but it has heat sinks instead of a fan that will eventually stop working so I'm hoping with no moving part that being a fan to keep it cool that it may actually last longer that 1-2 years which was how often my system cooked graphics cards.

14

The chip that heats up is a bit over an inch square.

Printed on it is "ATI IXP 400".

It is just a chip, soldered to the MB, no fans, no heat sinks.

Is that chip the southbridge chipset or is it the ATI IXP SB400 Realtek chip, . . . or what?

I don't know enough about this stuff to interpret what I am reading.

Some links:

http://www.tomshardware.com/forum/id-1722405/ati-ixp400-sb400-chip.html
and
https://en.wikipedia.org/wiki/Xpress_200#SB400_.28IXP_400.29.2C_M1573_.28RS480.29
and
http://www.driverguide.com/driver/detail.php?driverid=1234626
a manual with a diagram (pdf)
http://www.ctechinfo.net/documentation/manuals/amberine_manual.pdf

Mta, the MB has some open PCI slots but i don't know what type.

That linked manual seems to say they are x1.

Will check further later, all the specs are stored on an external drive that I can't get to right now.

If I can figure out how, I will for sure install an after market replacement.

Possibly a fool's errand, but maybe a NEW CHALLENGE!!

What did you get to fix yours, Andavari?

Right now I have Puppy Linux running, will leave it running for the rest of the day, to see if the computer still locks up after it wakes up.

15

the integrated graphics card on that mobo is under the heatsink, it's the northbridge ATI RS-482 Radeon chip.

the slot labelled PCIEX1 is a x16 slot so almost any of the current PCI-ex graphics cards should work in it.

16

The chip that is heating up the most is the one marked ATI IXP 400. So apparently it's not the graphics card.

If I am reading all this right, that ATI IXP 400 chip is the Southbridge chip. It gets hot enough that you want to pull your finger away, but not blister hot.

The heat sink on the RS-482 chip gets warm also, not as much.

If it is the ATI IXP 400 getting too hot, does that mean that old HAL The Haunted Computer is doomed?

Edit: If so it's not the "end of life as we know it, Jim", :P but I'll fix it if possible.

17

weirdly, I searched Dr Google further and one site stated the ATI IXP400 as the video chip - so go figure.

but logic would dictate that the mobo manufacture would have put a heatsink on the video chip, so I'm quietly confident the RS-482 chip is your video controller.

you could get a little heatsink to put on that hot chip to cool it, but trying to anchor it would be problematic.

no anchor points and any sort of contact adhesive would interfere with the thermal properties.

although I do seem to remember once when looking for thermal paste, some site had thermal glue, so that may be a possibility, but it could also just as easily been the wrong use of terminology.

simply improving general airflow through the case may be enough.

getting that hot that you can't keep your finger on it is not a good sign. that's somewhere around 50oC so would be hitting the upper tolerances of the chip.

since you tried the highly scientific analytical method to test heat output, now it's time for an equally effective remedy and direct a fan onto the effected area. :)

18

What did you get to fix yours, Andavari?

This was the only one I could find at the time, an old design with old stable drivers that's XP compatible:

EVGA Nvidia GeForce 6200, 512MB DDR2, PCI

Manufacturer Links:

PCI: http://www.evga.com/Products/Product.aspx?pn=512-P1-N402-LR

AGP: http://www.evga.com/Products/Product.aspx?pn=512-a8-n403-lr

19

Andavari the Nvidia Geforce 6200 is what is in my XP machine also.

20

Thanks so far. Still working on it. :)

Temps shown by CPUID HWInfo have stayed below 51 deg C (124 deg F) for the hottest CPU core and 43 deg C (109 deg F) for any other reading (chips or HDD). I know 51 / 124 is hot, but it isn't oowwiiee blister hot, and should not be beyond the limit of the CPU. And anyway, that temp was for one of the CPU cores, not one of the chips. So, still digging.

One more clue. :ph34r: When it locked up this morning, I had been running it for 2 or 3 hours, doing all sorts of stuff, but when Avast updated it jumped to the top of the CPU usage list, and the screen froze. I could not detect a spike in temperature.

The guis for both CPUID and Process explorer continued to respond, but procexp froze when I tried to use it to shut the computer down. Also, to answer mta's earlier question, Ctrl+Alt+Del doesn't work at all, there is no response from the keyboard.

Also, on restart, the internet was connected. It should not have been. Since I had disconnected it before starting Powershadow it should have been disconnected after leaving shadow mode.