Any IT professionals willing to offer advice?

sounds like you're on to something! keep following the trail

 

Well some research indicates some possible options.

Option 1: replace the CMOS battery

Option 2 is according to Dell:

The AB amber and CD green lights indicate memory modules are detected but a memory failure has occurred. The solution is to remove all but 1 module and restart, if it works add another module and restart. Continue until the system displays the problem and replace the last memory module installed.

That's what Mynorx suggested, too as I re-read his post. So, I think I'll try that route first.

Additional research also points to replacing CMOS battery and PSU.

I guess since my avatar is the famous detective Poirot, it's time to really put my little grey cells to work

Thanks to everyone for continued advice and for letting me bounce ideas off of y'all. I'll let you know what, if any of these solutions, works.

 

The ABCD lights on the back are diagnostic indicators, when you boot the system they will cycle through a sequence of codes and eventually go to all green when the bios post checks finish and the OS boots. The YYGY code you got indicates a potential processor failure. This could also be motherboard related and unfortunately the only way to know 100% would be to try an alternate processor to test.

Checkpoint errors will happen at random points and it's not uncommon if you see multiple checkpoints or even different ones on different boot attempts. These are often and indicator of a bad system board component or processor.

This is most likely not a thermal related issue. Dell motherboards are very thermal sensitive and will almost always shut off when they reach temperature thresholds. They will not restart. Most of the time the computer will power off and the power light will go to a flashing amber color indicating a voltage issue of some type.

I worked at Dell for 7 years and DCSE certified. I've seen this thousands of times, literally

 

oh boy, have I got some questions for you!!!

 

So I've got to tell this person the motherboard is on its way out??? Man, she'll blame me for sure since all this started after I defraged, diskclean, etc... You know, last one to touch it broke it.

Please tell me I have some options other than mobo replacement..please?

 

well, you're at a point where component swapping is the only sure way rule out parts which are good and parts that are bad...and that requires another Dimension 4600 (or very equivalent) that you can swap parts with.
Easy for me to say, with at least 10 of any given machine type we have, usually closer to 20.

one thing for sure to do...backup her hard drive!

 

Yep, backing up the hard drive I can do. Swapping out parts, not so much. I don't have any spare 4600's around.

Hmm...

 

just did a search on Dimension 4600's - it appears they were notorious from about a year out for failing power supplies.
However, it would be very odd for a bad PSU or CPU to reboot the machine and not completely power off.

Just to rule out software, do you have an Ubuntu LiveCD you could boot from and see if it reboots?

 

Yes, I do. That's a great idea. If I follow you, you're saying that if the Ubuntu LiveCD runs without the rebooting error, I may have a software issue (say a driver perhaps) but if it happens with the Ubuntu, then I can have some confidence in a hardware issue (CPU, motherboard, PSU). Is that right?

Since it's way too late to get started now, I'll try that tomorrow night.

 

If you are getting hardware diagnostic codes and machine check codes, this is all occuring outside the OS environment, before the OS even boots. I can tell you that it is very very unlikely that you are looking at a software related issue. There's nothing wrong with being thorough by any means, but what you are describing with the light codes is most likely pointing to hardware. I would use the time you have while the OS is booting and running for a period to back up the user's data and have the owner seriously decide if they want to invest in buying new hardware vs. just replacing the computer all together.

 

I have alot of experience working with old and new computers, most diagnostic software is a waste of time. your best bet is to swap out parts with known good ones. keep track of what parts you swap out. it seems like you have done alot of things and may be wasting your time. for future dont help people with computer problems, I stay away from that. I only touch them if im getting paid. because if something breaks people are easy to blame you for the problem. please try to swap out components one at a time. with known good ones. that will lead you to the problem. I used to buy pallets of old computers and sell them at swapmeets. so I do have some experience trouble shooting. memory test software doesnt detect all memory errors. set bios to default settings. good luck

 

Well, I gotta say this: The machine runs on Windows so nobody really knows why it reboots because there are no usable logfiles. Troubleshooting problems like these on Windows is a major PITA because of the fortress-like nature of the operating system (communication-wise, not security-wise ).

To give you at least some kind of advice: format and setup the machine from scratch, it's considerable less hassle and you'll get rid of all the junk that's built up since the last reinstall (for example Norton never *really* goes away which might be the source of the problem with it conflicting with AVG - seen that one personally).

//edit: One more thing for you to try: Download this CD image and burn it: http://www.phoronix-test-suite.com/?k=pts_desktop_live
Then boot from it and use it to benchmark the machine (put it under stress). It's Linux-based so if the machine randomly reboots there too, you're experiencing hardware failure. If it works like a charm, reinstall Windows (or use Linux instead )

 

Mike, I do plan to back up the data. I think that's critical. I don't have any hardware to swap out. So my options for them would be to replace the whole machine or replace the motherboard and CPU combo. Do I follow rightly?

 

Well, if I was doing this as a business, I agree I'd be wasting my time. However, I'm not so I consider it a good learning experience I may be able to use later on. My main limitation is that I don't have extra parts hanging around to swap.

 

I"ve seen Norton conflict with just about everything, too. And yes, Linux is my preferred OS but I have to work with what they like since it's not my machine.

 

Sigh, run prime 95's blend for 2+ hours, if you get zero errors, its not the board, processor or ram.
Amateurs... :happy102

 

I'll try not to take the perceived sarcasm personally, I hope. But since you mentioned this, do elaborate for the benefit of this amateur.

 

http://files.extremeoverclocking.com/file.php?f=103

Its used obviously for OC'ing but regardless, it can and will load your CPU to 100% and memory if you desire.

If you do this and it doesn't crash within 2 hours, its not hardware related.

Also a CPU-Z validation would be nice, so people can tell whats in the pos Dell.

I'm betting on it not being the board, however check the bios and make sure everything is detected, if a slot isn't being read, or the HD isn't being fully detected and you get random outage of the peripherals than yes its the board.

CPU failure is pretty rare, unless that rig has never been cleaned in its life *well probably it is a Dell user* its probably got a stock cooler on it, which is probably gunked up. If its an AMD CPU, and its running past 55c, and not a Thurban, then yes its the CPU.

Ram, if its bad ram, it probably wouldn't boot, or it would just BSOD however only having one gig of ram is not ideal either.

PSU, unplug it from the board, and short it with a paper clip, if the fans spin, its not the PSU.

Also I almost forgot, if the 12v rail drops below 11.80v at load or the rails fluctuate by more than 5%, then yes its PSU.

 

Prime 95 is a stress testing software that will thermally push the hardware to it's limits. I've never used it, but overclockers like to use it to find out when there 'clocks are stable and when it's too much. This would include the processor, memory north/south bridges.... in turn taxing the PSU as well. I don't know if upon rebooting if it will tell you which component caused the failure or not. If so that's a good place to start.

I'd re-seat the processor and re-apply thermal paste (Arctic Silver-5 is available at Radioshack) following msandoval858's direction about processor failure and check the 4 or so large caps right next to it for "popped hoods". Upon your findings on the CD-ROM, have you tried popping in a new one to see if opening/closing causes a reset?

 

It doesn't tell you what part is faulty, but going by temps, and the way the system responds you pretty much can get the idea of what is going on.