Hi,
I hope this is an easy question to answer...
One of my older (dual-CPU) computers recently developed a problem. Under Windows, it was randomly giving all types of faults: IRQ_NOT_LESS_THAN_OR EQUAL, PFN_FAULT, and many others that I don't remember. Under Linux (which is what it mainly runs), it was freezing up randomly. Since both OS's were experiencing problems, I figured it had to be hardware...
After a day long troubleshooting session (using another computer to test all the hardware), I was able to rule out all the hardware: CPU's, memory, video card (AGP), PCI cards, etc..
I finally nailed it down: whenever either CPU is inserted into the CPU-1 slot, these errors happen. Using a flashlight (this is a big server board and case), I saw that one of the caps next to the CPU-1 slot is "dinked". Actually, I remember accidently dinking this cap many years ago when I inserted a CPU into the slot, but it didn't cause any problem then. From what I can tell, only the wrapper appears to be dinked. The metal casing under the wrapper seems to be fine.
The computer in question is running fine and stable using only one CPU in Slot-2, however I would like to get it running SMP again. I can definetly tell the difference between running 1 CPU and 2 CPU's under Linux. It runs much smoother usng both CPU's and it does even run "faster" when a lot of applications are running.
I can get to the area where the cap is on the underside of the motherboard without much difficulty. Well, I can reach it with multimeter probes, anyway. The only way that I would be able to de-solder and remove it would be by removing the entire motherboard from the case -- which is going to be a major PITA because there is a ton of hardware in the case (i.e., SCSI RAID).
Q: Can you test a capacitor without removing it? That is, can I just probe it with a multimeter and look for something? I seem to remember that you test caps by charging them up and then watching them drain with a multimeter. Does this work if they are connected still to other components? The cap in question is 10V and, (its at an angle that is difficul\t to read) I think 150 uF.
Strange that it worked fine when I dinked it way back then, but all of a sudden seems to be causing problems? Could there be a very small hole in the metal casing, which caused it to bleed away over the years...?
Thanks!