Bench Test from Hades...

I have a saying, that I can fix anyone's computer but my own. I personally can't believe how true this presently is, but I need help and even my own local peers are stumped. At least 4 different +A certified people including myself have looked at it. I have a P5PL2 Motherboard from ASUS, an NVidia 7600 PCI Express Graphics card from BFG, and an Intel 2.8 Dual core processor socket 775. I will not go into detail of the other items because I have multiples from different manufacturer's and I have narrowed the problem down to these three items. I have even replaced the motherboard which DOA'ed itself 5 months into the ongoing diagnostic, (yes I said 5 months....)

What is happening, started with the power supply. This computer spent a few months in a shed during march of this year where it ran, until i realized that the shed was getting too warm to support the system which does run warm on its own. After this, it started alternately failing. Now this failure did not occur during operation but if it was shut down and allowed to cool; it would refuse to start back up unless it was warmed up first. Now at the time I chose to simply keep the system running and assumed that a solder somewhere in the power supply had simply come loose and needed to be warm in order to make a proper connection and it was nostalgic to the old tube TV's which didn't like to run cold. But I did replace the power supply and it operated fine for 2 months. Now what's amusing is that the power supply I took out passed all bench testers for voltage out. Now I personally question what amperage stressing would do to change those results but that was something I haven't tried yet.

The new PS then simply started shorting out. It "would" randomly fail mid-operation and would restart and function without question or fail. But Average operating time was 30 seconds to 36 hours between failures. I switched back to the old power supply and started a heat gun treatment to get it started on occasion and simply left it running. Then the system started burping. It would fail just long enough to cause a reboot.

Now I am familiar with graphic card failures, and this mimicked one, so I started component replacement bench testing techniques with burn in testing while running hardware diagnostic software from Hiren's boot disk. Every single one of these passed, with flying colors... Even the original hardware configuration with no changes or replacements... I replaced the motherboard with a purchased replica of itself, the power supply 3 times, the ram twice, the graphics card once. the only thing I have not removed, and this is simply because it has passed every single processor check on every single burn in test I could perform, is the processor. I've never seen a random processor failure that could not be made to repeat itself. But one thing I have been able to repeat is that now 2 power supplies require multiple attempts to boot or must be "warmed up" before they will post properly. I've even upped the total wattage of the system which was running 380 and now 450.

I am running inside the system 1 graphics card, 2 SATA 3.5 inch hd's, a SATA raid controller PCI running RAID 0, a data IDE 3.5 inch HD, 1 extra fan that cools the hd rack, a 3.5 inch floppy, 1 internal CD-ROM, and an external self-powered DVD-RW, and 2 pci nic's. I have wiped Windows XP twice, running from both the IDE and the SATA's in normal mode and in RAID 0. The system was not originally but is now running in a power conditioned and back up controlled environment. It's internet access is firewalled by a separate computer running ipcop and even then I have netbios ports blocked as well as both registry and remote access services turned off. And these reboots will even occurr when it is not connected to the network. I have run out of things to test and my system is still randomly dos'ing itself.... Help? What should I be looking at now besides a priest or possibly napalm?

 

Well, since I don't believe in the power of a priest......I'll vote for the napalm.
(God yes, priests no)

Or maybe just a good tech.

That was a long post and possibly I missed a lot in my hurried scan, but have you thoroughly tested your ram? When a PC goes totally bonkers and there doesn't seem to be a logical answer, it's usually the ram.

I seldom replace ram. I usually just wash it.
Even in my own super system, my latest ram upgrade failed me, till I took it out and washed it in Denatured Alcohol. Now it's running like a scalded dog. lol
I never fuss with ram, either motherboard ram or Video ram. I just wash it.
Ram works on voltage levels, not current flow. Just a little dirt between the legs of a ram chip can cause it to fail. Dirt looks just like a copper wire to that ram chip.
Remove the short circuit and the ram stick will once again work like new.

Well, that's where I'd start anyway. I've solved weirder problems than yours, in my 29 years as a computer tech.

Good Luck,
The Shadow

 

I have three sets of ram and I have interchanged each of them into the system and continued my testing. The ram has nothing to do with the system PS needing to be warm before powering on, which is presently the only symptom I still have. Though if it is any consolation to your reply yes it was the second thing I replaced during my bench run, including a pair I bought from Frys only to return it when both it and my normal ram passed diagnostics the same, and the store bought even rebooted once.

 

Welcome to Major Geeks.
Not sure on this, but did a Google on your motherboard. Shows that it supports a P4 3.2GHz, 800MHz Bus speed CPU.
Don't know for certain, but don't believe that a Dual-core is a P4.

http://www.superwarehouse.com/Asus_P5PL2_Motherboard/P5PL2/p/728156#processors

 

http://usa.asus.com/product.aspx?P_ID=HPEy5xU6Uj31iLoX&content=specifications

Thanks for the welcome, btw. I've known about the site for just over a year now from another friend who thinks the site is a shrine to all good things IT. Got CCleaner from this site for the first time, as well as a few rogue registry, DLL, and services hunting tools.

The Motherboard supports, according to the ASUS website and from example of the fact that this is a 3 year old system, (should have mentioned that in the original part) that the board does indeed support 775 Pentium D's which my specific cpu is an Intel Pentium D 820 Smithfield LGA 775.

I also forgot to mention that I am running dual channel RAM, and when I say sets of ram, I mean 1GB per stick, 2 gig total sets of ram. I own 3 sets of them, two left over from other systems that I use for bench testing purposes.

 

Like my sig says"Continuing to learn"

 

All hail the God of Google, and it's prophet Wiki. Google knows all but only wiki will tell you the truth. (most of the time) And besides that does remind me to make sure that Windows does have the chipset drivers loaded. My old Soyo Dragon would randomly frag itself if it didn't have those loaded and resemble a graphics card failure.

 

The 2009/08/24 update suggests that the latest BIOS may fix your problem

 

Looks like a fun problem.

You say you have used different sets of RAM. Make sure you have the memory timings set properly for the set you are using. If that is set right, I would strip it down to just one stick of RAM and the GPU, see if it will POST and enter the BIOS. If it will, check the PSU voltage with a DMM. Perhaps add the CD and run memtest or some stress test to check it out. If it passes, add each component one at a time, and see when the problem starts.

Would not hurt to check the power cord and wall outlet as well.

 

BIOS flash on an unstable machine might not work out very well.

 

I just realized (while google hacking myself) that this thread was never finished, or closed.

End result. A few choice capacitors on the motherboard cooked themselves and created a randomized board failure. It would take 30 seconds or 30 minutes or it simply wouldn't work at all. the board was replaced. :cool

 

Thanks for being considerate and updating us.
Thoughtfulness is appreciated.

 

Excellent! LOL @ 'google hacking yourself'!roflmao

 

With all the millions of bad capacitors that the Chinese installed in our computers and power supplies over the past few years, that should be the very first thing looked for in any computer when it starts doing weird things.

A while back I scrapped out 14 computer power supplied that I had replaced over the past year and just dropped into my junk box. The worse ones were from Dell.....they seem to love buying the cheapest PSU's that they can get, from the lowest bidder. I saved the cooling fans from every power supply, but that was it and every power supply had at least one or more bulging and leaking capacitors. The term "There should be a law......." comes to mind.:cry
It should be a criminal offense to sell crapola like that to the American public.
But it's not.....Thank you Washington!:major

So if your computer is doing weird stuff, Look closely at the capacitors on the motherboard. If even one is bulging on the top, even a little, it's already failed and can seriously effect the operation of the computer.

The Shadow :cool
retired Computer Tech.

 

Sad, but true. Even sadder is the fact that the ruse also includes companies here at home: a few years back, I slipped out of a tour group at a TRW plant, and saw finished capacitors on a conveyor belt being funneled to separate bins - one marked 'Mil Spec', one not. They were the same components!

:confused

 

I had a nice response all typed, about a page of it, but I just lost it. Crap!

So here's the Condensed Version:
I had to replace ten bad 5v caps on a Dell motherboard, so to get guaranteed replacement parts, I ordered them from
http://www.badcaps.net/

The good, bad and ugly components are not made, but separated out of general run parts during final inspection and testing.
They may all look alike, because they come off of the same production line.
1% tolerance parts are directed into one bin, while the 5% parts to another, and 10% parts to another. Everything else is sold to Dell. ):-D:-D:-Drolleyes
(can't I joke, just a little?)

Celeron CPU chips were not made, but failed Cache-Ram testing, so were sold with the Cache disabled and bypassed.
Transistors, that failed testing because of one bad junction, were sold as diodes. It's nothing new. These little production tricks have been going on for a very long time.

But bad caps were not the result of production line mishaps or grading problems. It was the result of a faulty electrolyte formula.
Millions of them found their way into all sorts of electronic devices before the problem was discovered.
I scrapped out 14 PSU's (that had failed) and every one of them, from many different vendors, all had leaking or bulging capacitors.

Cheers Mates!
The Shadow :cool