Paul, this sounds very much like the same problem I and a few others
discussed on this group starting on March 14th. (See: "System failed
CPU test") You were also a participant in that thread.
http://groups-beta.google.com/group...d22c3475e44/87ef664ea81abfb2#87ef664ea81abfb2
If this is the same thing, there doesn't seem to be a definitive
answer. At least none that I've been able to find.
My "solution" was to turn the machine on, leave it on for 15 or 20
minutes while turning down the volume so that constant voice message
("System failed CPU test") didn't drive one nuts. After few minutes, I
shut down and rebooted -- this time successfully.
If you or Hello can find the prior thread (link above) it may provide
more detail including a link to the rather lengthy ABXZone forum
discussion.
BTW, though it wasn't mentioned in that previous thread, I occasionly
received an alternative voice message indicating my failure to POST
was due to overclocking. (I have not overclocked)
f
Did you look at his symptom list ? His board is shutting down,
if the processor is installed and no RAM is present. Instead,
there should be a Vocal POST saying failed RAM test or the like.
And I think it should stay powered. I tried the cardboard test,
just for fun, on my P4C800-E while building it ip, and it never shut
down, no matter what combo of components were installed.
His symptoms are unique, and I haven't seen this set before. I've
repeated them again here:
"-CPU Out of system: voice message is "no CPU installed"
-CPU installed in system: no voice message, the system
just shuts down. I repeated the start sequence three times
just to make sure.
-Addition of ram: voice message is " system failed due to
CPU overclocking". (Basically, back to where I was)"
No CPU installed, is done by the Vocal POST chip detecting
whether a pin on the CPU socket is grounded or not. All that
tells us, is the Vocal POST chip is working.
The shutdown in the second test is the puzzler. It could be
BIOS code shutting the system down (like if the BIOS detects
a thermal problem, or no fan is present), it could be the
Vcore circuit latching off due to a perceived overload, it
could be the PSU shutting down due to a perceived overload.
The third test, suggests the processor got to run enough
instructions, to at least write an event into the Vocal POST
chip. Now, if the third test is running BIOS code, that
increases the odds that the shutdown in the second test,
is being done intelligently, rather than some hardware
being at its limits.
"System Failed CPU Test" is the result of a timer timing out
on the Vocal POST chip. If the CPU doesn't manage to execute
some BIOS code, then the CPU doesn't have an opportunity to
disable the timer on the Vocal POST chip. Now, there are so
many possible root causes of a general inability to compute,
that of course you are going to see that error message show
up a lot in news groups or private forums.
There are some failure modes, that have distinct enough causes,
that they've been given names. The "monitor leakage" problem,
where a system refuses to start, until the monitor cable is
unplugged. Those are easy enough to understand, in terms of
the workaround that can be used, but no one has a verifiable
theory as to exactly what is happening. (I can think of some
tests, but I don't have the problem, so cannot debug it.)
But, in the case of the majority of "System failed CPU test"
problems, the root cause is a failure of the motherboard or
the power supply, to provide the essentials to allow the
CPU to boot.
The thing is, you just don't have an appreciation of the
3000 wires inside that board, and how if any one of them
goes open circuit or short, the system will stop. It is a
miracle that computer motherboards could ever work, and
yet Asus has a 99.6% success rate with getting the boards
to work on the factory floor. (That means 0.4% are thrown
away because they cannot be reworked to make them 100%
functional.)
One of the things I consider to be a mistake on recent
Asus boards, is the use of Vcore regulators that latch
in the off state. Older regulator circuits were designed
to attempt a restart, after encountering a problem, which
made those motherboards more resilient to transient
startup problems. As an example, there was the P4P800 Deluxe
motherboard, that would not start when an Antec Truepower
supply was used. My theory is, this is due to the Vcore
regulator shutting down and staying shut down. With some
of the older designs, the user would never have even
noticed that the Vcore regulator needed a couple of
tries until the Vcore rail popped up. In which case, the
board would start every time, and the user would be unaware
there was any kind of problem.
To use another example, say you jump in the car in the
morning, turn the key, the motor turns over, but the car
won't start. As an amateur mechanic, you can see that a
lack of spark, lack of gas, lack of air, could all contribute
to the car failing to start. The symptoms in that case
have so many root causes, that you simply cannot expect to
jump to a simple conclusion. "System failed CPU test" is
exactly like that - it could be a lack of spark, a lack
of gas etc
Paul
