Time To Put A Fork In It

It may have been hard disk related. NTFS constantly hits the disk, and a marginal disk could cause the lockups you were seeing.

A good and inexpensive tool to deal with those kinds of disk issues is Spinrite by Gibson Research.

Dave

What - you didn’t even plug in a test power supply???

Hello,

If the machine is from the 2002-2003 era, then its motherboard may have been manufactured with faulty electrolytic capacictors that failed prematurely due to outgassing. This could cause the problem you noted.

If you were to examine the capacitors on the motherboard, particularly those around the ATX power supply connector or AGP video graphics card expansion slot, then some of them might have have a slight (or not-to-slight) bulge on the side or top of their cannister, along with a reddish brown or black oily residue at the base from the leakage. In many cases, removing these capacitors and soldering replacements in will solve the problem. For more information this problem, as was as the reason for the faulty capacitors, visit http://badcaps.net.

Other possibilities include a failing power supply or CPU cooling fan. These devices become less efficient over the years, resulting in improper voltage being delivered by the power supply or leading to an overheat situation with the fan. Replacing the devices in these cases can solve the problem.

Regards,

Aryeh Goretsky

Same thing happened to TVs and appliances. When I was a kid, there were TV and appliance repair shops everywhere. Now there are hardly any.

Computers have just become another appliance - they’re really not ‘technology’ any more because they can be cheaply reproduced.

I’ve seen lots of computer repair places pack it in (leaving most of the business to places like Best Buy / Geek Squad).

Matt,

So close! You ALMOST got it, then dropped the ball. How do I know? I have a pc with almost exactly the same problem. I don’t get the random freezes, but the rest is there.

Now to the root of the problem.

I believe the real culprit here is a really sneaky one- the system case, abd I’m not sure you can necessarily do anything about it.

The source of the real problem (I believe) is the geometry involved between the rear case wall and the motherboard. If you think about it for a second, it’s supposed to form a 90 degree angle where the two intersect. If it doesn’t, or the ‘bump-in’ for the backplane expansion card cutouts is too tall, then cards aren’t going to fit into the motherboard properly. I’ve experienced these problems for years and with different systems. Most times, it’s not critical and the system works reliably. Other times, a certain case-motherboard combo can result in cards that don’t seat fully and/or a warped motherboard. Let me tell you: it doesn’t take a lot of warp to make anything plugged into the motherboard totally unreliable as it heats up.

The big light that should have flashed on with this system was what you DIDN’T hear- the POST beeps. The specific item that we’re talking about here on the POST list is the video card- No card? No beep. You took care of reseating the RAM, so you’re next stop should have been that errant video card. Then you would have had your immediate reboot.

As you said, with a system that old, it may make more sense to go out and pay the big bucks to upgrade to a new system rather than buy a new case and probably motherboard. By now, the case he has will never allow total reliability and the motherboard, thanks to thermal action, is probably irreparably warped that tiny bit that is all that’s necessary.

At least you now know the missing piece of the puzzle.

I’ve heard good stuff about SpinRite… but the times that the machine won’t even POST and display a splash-screen, it’s not even trying to access the disk at that point and won’t boot from floppy or CD. That’s why I think it’s something else.

But I am going to check out SpinRite, as I think it could help me with other situations.

I agree an uncorrected bad sector on the hard drive can cause it to lock-up. But that does not explain the fact it would not boot again until it had plenty of time to cool down.

As of recent I have seen all manor of unusual seemingly hardware related problems get solved by simply replacing the little coin battery. If anything unusual happens, I pop the cover and check the battery with a DVM that has a battery test setting. If the coin battery shows anything less than 3 volts it is due for replacement. I used to believe that the battery was still okay but will need replacement soon if it shows anything more than 2.7 volts, but changing a battery that still showed 2.9~ volts solved a problem I was recently chasing.

Five years ago also puts the Power Supply and motherboard capacitors up as potential culprits.

Short of being in an actual repair shop with a selection of known good spares to swap with, narrowing the problem any further would have cost money for parts that may or may not solve the problem. Telling the customer to just buy new hardware may be a tough call, but maybe in the long run the better economical choice.

As for the old system, I still use systems that are more than 10 years old in applications where upgrading the hardware would require a rewrite of the software. These systems are in industrial control and telemetry collection systems where the hardware needs of the application does not justify the speed or extra power requirements a new system would represent.

Several states are about to pass laws that will disrupt the ability of companies that are in the business of recycling and reselling older computer and electronics equipment. One in particular has plans to limit the ability of a company to sell equipment within the state unless they first become “certified” by the state that they have proper procedures in place to accept and recycle systems returned by the customer. This particular law is already on the books and will go into effect later this year. It is one thing to pass laws prohibiting such equipment from being thrown away and going to land fills. It is another to impose requirements that all companies register with the state to be able to continue doing business. Until it changes, my company plans to simply stop selling to customers in that state. We have always had terms stating to our customers how they can contact us for returning their old equipment for recycling. Only thing we require is that the customer purchased it or what replaces it from us. This law will only serve to disrupt already existing recycling businesses, and cause more to wind up in land fills.

Exo

raymond — believe it or not, I didn’t have a test power supply available to me, although the idea did cross my mind. And I didn’t see the logic in going and purchasing one (or having the customer purchase one) because of the additional time and money involved (I know power supplies aren’t exactly expensive, but my time is).

Aryeh — you could very well be the issue, but again, it comes down to how much time needs to be invested in examining and replacing capacitors.

Exothermic — good call on the MoBo battery. That didn’t even enter in my mind, I guess probably because the system time was never incorrect.

Jeff — funny you should mention geometry. There was once instance where I couldn’t get the PC to pass POST, let alone boot, so I re-positioned the tower on its side, and it powered up (and froze up again like 45 minutes later, still on its side). But again, this kind of problem is so thorny to correct, it’s just not economically feasible to try to isolate and rectify it. I did check all the cards for proper seating, including the video card, and nothing would give me a result. Just letting the machine sit for different random periods of time.

Exo — good call about the MoBo battery. It never entered my mind, probably because the system was keeping time/date info correctly. It is interesting how more and more states and municipalities are regulating electronics disposal and recycling. It kind of reminds me of certain auto parts and materials (tires, batteries, etc).

Good comments everyone — I’m smarter for it!!!

Agree, no use trying to fix a machine that old.

Well done getting the guys data save. I’m sure he’ll be happy with a new machine.

You mean “loath” (reluctant), not “loathe” (despise). Otherwise a fine article. I think you made more than a sufficient effort to identify and remedy the problem before putting the fork in it.

Bad caps are easy to spot at a glance with a flashlight. There’s no troubleshooting needed. I’ve lost at least 4 mobos to bad caps :(

I’d at least check that all the cables in the case are fully seated. I once worked on a computer with what sounds like the same symptoms and it turned out to be one of the cables from the power supply to the motherboard was in it’s socket but not fully seated. At a glance everything looked okay, but this one cable was just resting in it’s socket, the way the cable was routed kept it from falling out completely.

My 7+ year old EMachine started locking up [fortunately I was able to copy my data on one of the near-dozen reboots that actually worked]. I used the 2 EMachine restore CDs but to no avail; I then installed a new hard drive and restored once again with the CDs but the cursor froze like before on the first couple of boots. At this point I was ready to trash the beast [I draw the line at hardware debugging] but lady luck was sitting on my shoulder and whispered, “go on! one more boot!”. Sure ’nuff, it’s been working like a champ ever since! Never say die unless lady luck tells you to.