computer starts then stops then starts again etc

bit of a weird one
computer powers on and shuts down almost imediately [all lights go
out] it restarts itself to the same point and repeats the shut down
it will do this 4 to 5 times prior to booting through to windows...
computer is less than a month old and started this caper 3 days ago
everytime it starts this happens
its running win xp + vista in dual boot scenario

computer spec:
Cases COOL MASTER ELITE 330
Motherboards Gigabyte P35-DS3L ATX
Processor INTEL CORE 2 DUO E6850
Memory 4 gigabytes ddr 667
Hard Disks SAM 500GB SATAII HDD x3 units
CD/DVD DVD WRITER DUAL x2 units
Sound SoundBlaster Xfi elite pro
Video Gigabyte 8800GTX 768 meg
Keyboard ms key mouse wireless desktop 6000
Monitor Samsung 24 wide 244T
Card reader Generic
OS XP HOME SP2+
usb port up to total 12 Generic - integrated
Power supply 500 watt twin rail-generic

any ideas??
scul
scul

scully wrote:
> bit of a weird one
> computer powers on and shuts down almost imediately [all lights go
> out] it restarts itself to the same point and repeats the shut down
> it will do this 4 to 5 times prior to booting through to windows...
> computer is less than a month old and started this caper 3 days ago
> everytime it starts this happens
> its running win xp + vista in dual boot scenario
>
> computer spec:
> Cases COOL MASTER ELITE 330
> Motherboards Gigabyte P35-DS3L ATX
> Processor INTEL CORE 2 DUO E6850
> Memory 4 gigabytes ddr 667
> Hard Disks SAM 500GB SATAII HDD x3 units
> CD/DVD DVD WRITER DUAL x2 units
> Sound SoundBlaster Xfi elite pro
> Video Gigabyte 8800GTX 768 meg
> Keyboard ms key mouse wireless desktop 6000
> Monitor Samsung 24 wide 244T
> Card reader Generic
> OS XP HOME SP2+
> usb port up to total 12 Generic - integrated
> Power supply 500 watt twin rail-generic
>
> any ideas??
> scul
> scul

My best guess, would be something is up with +5VSB. That is the
standby voltage, which powers the circuit that looks at the front
panel power switch. Without +5VSB, the motherboard has no juice to
drive the PS_ON# signal with a logic driver. And then the PSU goes
off.

The only thing that doesn't make sense, is why the supply comes
back on again. Unless you have the recovery on power fail, set to
"last state", I suppose that could make the logic try to turn on
the supply again.

If you don't have access to any instruments, I'd swap the supply.

If you have a multimeter, you can check the +5VSB wire on the power supply.
If you look in your user manual, the main connector pinout is shown,
and labels the pins. Pin 9 is +5VSB. Poking the meter into the back
of the nylon shell of the connector, allows making contact with
conductive metal in the connector. That way, you can measure what is
going on, with power applied.

It would be better if a LED and limiting resistor was tied to
+5VSB, because that is a more sensitive test of transient behavior.
A multimeter is not that good at detecting spikes on a rail, while
the light from a monitoring LED is a better indicator.

HTH,
Paul