I have a computer with a socket 468 motherboard that I found in a dumpster a few years ago. The computer has always worked perfectly but close to the processor I found two bulging caps that I decided to replace. The caps were 3300uF 6.3V Nichicon HM(M) and since there were four of them in total I decided to replace all four of them, despite only two of them showing visible damage. I left all the other caps alone. There was also some burn marks around the pin of the ATX socket where an orange (3.3V) and a brown lead goes. The original PSU was of really low quality (ACE Raw Deal 420W) and probably contributed to the damage.
Anyway, I replaced the four Nichicon capacitors with Chemicon KZG of the same specifications. The new capacitors were unused, manufactured in 2007 and appeared genuine as far as the vent, bung and build quality goes. They measured a capacitance of about 3900uF (within 20% of their specified 3300uF) and an ESR of about 0.05. The crappy PSU was also replaced with a FSP PSU that had served me well before in another computer.
The capacitors I replaced tested about 5500uF for the two without visible damage. One of the bulging ones were 4000uF and the other one was completely dead and not recogniced as a capacitor by the meter. The ESR of the first three was about 0.2-0.3.
After the capacitor change was done I rebuilt the computer and fired it up. It works well until it loads the graphics driver of an operating system; then the screen goes blank and loses contact with the graphics card. I've tried both Windows XP and a Linux Live CD. Both DVI and VGA. Two different graphics cards (AGP by the way). Two different processors. A complete reinstall of Windows. The problem is always the same. If I start Windows with no graphics driver installed it will run on the generic drivers, but as soon as I install the dedicated Nvidia drivers the screen goes blank after the second "test flash" that occurs during the installation process.
I tested the RAM with Memtest86+ and the processor with Prime95 and neither give any errors.
Does anyone have any experienced this problem before and do you think it's related to the new capacitors or damage to the motherboard inflicted during the recap procedure?
There are empty silkscreens for capacitors on the motherboard that seem to be connected in parallell to the capacitors I replaced. Could I improve stability by filling them with more capacitors?
Anyway, I replaced the four Nichicon capacitors with Chemicon KZG of the same specifications. The new capacitors were unused, manufactured in 2007 and appeared genuine as far as the vent, bung and build quality goes. They measured a capacitance of about 3900uF (within 20% of their specified 3300uF) and an ESR of about 0.05. The crappy PSU was also replaced with a FSP PSU that had served me well before in another computer.
The capacitors I replaced tested about 5500uF for the two without visible damage. One of the bulging ones were 4000uF and the other one was completely dead and not recogniced as a capacitor by the meter. The ESR of the first three was about 0.2-0.3.
After the capacitor change was done I rebuilt the computer and fired it up. It works well until it loads the graphics driver of an operating system; then the screen goes blank and loses contact with the graphics card. I've tried both Windows XP and a Linux Live CD. Both DVI and VGA. Two different graphics cards (AGP by the way). Two different processors. A complete reinstall of Windows. The problem is always the same. If I start Windows with no graphics driver installed it will run on the generic drivers, but as soon as I install the dedicated Nvidia drivers the screen goes blank after the second "test flash" that occurs during the installation process.
I tested the RAM with Memtest86+ and the processor with Prime95 and neither give any errors.
Does anyone have any experienced this problem before and do you think it's related to the new capacitors or damage to the motherboard inflicted during the recap procedure?
There are empty silkscreens for capacitors on the motherboard that seem to be connected in parallell to the capacitors I replaced. Could I improve stability by filling them with more capacitors?
Comment