On Feb 1, 8:55 pm, pauld <pauld...@hotmail.com> wrote:
> After running numerous tests I think that the memory controller on the
> original board is bad. The new board POSTs with any combination of RAM
> and with either the old or newpower supply, ...
A primary reason for fixing things is to learn. Described is how to
learn from your mistakes. First things were replaced only on
speculation. Second, after so much labor, unknown is what is and is
not 'definitively good'. Third, a computer can still boot and run
even with defective parts such as a power supply. Fourth, a good
power supply can appear defective in another system for numerous
reasons. Included in that list: it is a power supply 'system' where
power supply is only one 'system' component.
Doors in a house were sticking. So we plane down all doors? No.
First we locate a reason for all those sticking doors. Common factor
is a defective foundation. Instead of fixing doors (curing symptoms),
identify and fix the foundation - the actual problem.
Foundation of a computer is its power supply 'system'. Until that
foundation is confirmed, then anything can act unstable or fail. Start
an analysis using a two minute procedure "When your computer dies
without warning....." starting 6 Feb 2007 in the newsgroup
alt.windows-xp at:
http://tinyurl.com/yvf9vh
Connector chart to locate each color:
http://www.hardwarebook.net/connecto.../atxpower.html
You speculated a power supply (and other parts) was defective rather
than first identify the suspect. A failure of any other part in that
power 'system' also could have made the supply appear defective.
Supply and other parts are still unknown. Replacing a power supply on
speculation is called shotgunning. If a power supply was defective,
then that two minute procedure would have identified it 'definitively
bad'.
Same applies to motherboard and to memory. Dont' shotgun. As said
in CSI, "Follow the evidence". Before replacing anything, first
identify a suspect.
More responsible computer manufacturers provide comprehensive
hardware diagnostics just for this reason. If your manufacturer is
not so responsible, then download diagnostics from each component
manufacturer or from third parties.
For example, memory can report failures if a power supply system is
"unknown". When ready to test memory, also execute that diagnostic in
a 100 degree room or with semicondutors heated by a hair dryer on
high. Diagnostics - especially when executed with memory warm - takes
memory from 'unknown' to either 'definitively good' or 'definitively
bad'.
As noted up top, learn by fixing things. Shotgunning results in
wastefully buying components. What is 'defintively good'?
Demonstrated is how to convert speculation into suspects - before
buying anything. There is no reason to buy parts until the suspect is
first identified. That procedure also results in less time and less
labor as well as money saved. Mistake was to shotgun. Fix things to
learn from mistakes. Fixing things should not result in replacing
perfectly good parts.
A last point. Normal for a computer to boot when its power supply
is defective. Supply could have been defective when system was first
built. Using that same two minute procedure would have identified a
defective power 'system' when computer was first assembled. How to
eliminate a strange computer failure before it happens.