ECS MCP61M-M3(1.0A) shutting down

[Dennis]

I have a computer with the ECS MCP61M-M3(1.0A)Motherboard and an
AMD Phenom II X2 550 3.1Ghz Black Edition processor running Windows
XP Pro. This was built last November but recently started shutting
down at any time. It goes through the whole boot sequence when the
computer is turned back on. It will shut down while idle or while in
use. I added a case fan in case it was heat related but that did not
help.
Any help greatly appreciated...

[Paul]

http://www.ecsusa.com/ECSWebSite/Product/Product_Detail.aspx?DetailID=1006&CategoryID=1&MenuID=19&LanID=0

The CPU support table, shows processors up to 95W. The board has
a three phase regulator. I don't think your processor is too
much for it, unless it was overclocked a whole lot (driving power
over the 95W level). And you say it happens at idle. Is the
CPU VCore voltage consistent with operation of that processor ?
The BIOS could set the voltage too high, as an example of
a BIOS bug type.

It sounds like THERMTRIP, because THERMTRIP (a signal from the processor),
is used to turn off the PS_ON# signal sent to the power supply.

If VCore overcurrent protection was tripped, all the fans might
continue to run, but there'd be no CPU to POST, and a black screen.
To clear such a fault, might require turning off power at the back
of the computer.

I suppose it could be something related to the single chipset heatsink,
like it had come loose, but that would be more likely to cause it
to crash than to switch off. The only thing I know of, more or
less guaranteed because it's free in terms of protection, is THERMTRIP.
Everything else is optional, and could cost a few cents more for
components on the motherboard.

You could remove the CPU from its socket and re-seat it. In case
there is a bad contact in the socket or something. Then apply fresh
thermal paste and bolt the cooler back on. In some cases, AMD
aftermarket heatsinks, have been known to break off the plastic tab
on the socket retention, which causes the heatsink to fit loosely,
and perhaps even at idle the temperature of the CPU could end up too
high (because the cooler is no longer making good contact with the CPU).
Use something like Speedfan to monitor the temperature and see what it is,
just before shutdown.

http://www.almico.com/speedfan446.exe

http://4sysops.com/wp-content/uploads/2009/08/SpeedFanWindowsXP4_thumb.png

Paul

[Dennis]

The CPU fan never runs over 1200 RPM and as low as 500 RPM at times.
The fan spins free. So, is it the fan or is it not being told to run
faster by the motherboard?

[Paul]

My standard operating procedure is:

1) Enter the BIOS and disable fan speed control.
This is supposed to cause the fan header to have +12V on it,
and the fan should run full speed.

2) Verify the fan interface type. Three or four pin connector. Three pin
fans rely on modulating the +12V supply pin on the connector. On the
four pin headers, the fourth pin is PWM, and it's a control signal for
setting the fan speed. The desired fan speed, is communicated by the
pulse width of the square wave control signal (25KHz signal with variable
percentage pulse width).

I think if the PWM signal floats (isn't driven), it stays high 100% of
the time, and the fan runs full speed. If the signal is set to zero,
at least the Intel spec for fan behavior says the fan should continue
to spin at a minimum speed.

This is an example of the four pin (PWM controlled) CPU fans.

http://www.elma.it/IMAGES/fantt/4in4a.jpg

Motherboard makers like the four pin type, because they may get
the necessary control signal for it "for free". If they have to make
a three pin CPU fan control with speed control capability, they have to
fit a transistor (MOSFET) next to the header, and that costs money. So
they would prefer to use the four pin type. If you plug a three pin fan
into a modern four pin header, it should run full speed. If you plug in
a four pin fan, the speed should be adjustable by the motherboard.

3) Depending on the cooler design, you may be able to use a "case fan"
to replace the fan on the cooler. For example, the cooler on my CPU
takes a standard fan. And when I broke the fan, fitting a case fan was
easy. Not all coolers use standard fans. With the case fan fitted, the
fan runs full speed, all the time. I use a "quiet" fan, and the full speed
on that, doesn't make a lot of noise. No need for speed control at all.

The copper trace that powers fans on the motherboard, is not fuse protected.
It's possible to burn out that trace. But in that case, all your fans using
those headers, will stop turning. If the fan is moving slowly, either the fan
is bad, or the header is bad in some way (not putting out the voltage that
the fan needs, or the BIOS has done something to the hardware setting). You
can enter the BIOS, and disable fan speed control. And if that doesn't work,
try the fan speed adjustment in SpeedFan, and see if you can make the fan go
faster.

If those efforts fail, you can get an adapter cable, to power the
fan from a Molex (disk drive) connector. But those aren't exactly
easy to find when you need one. Since the CPU fan header needs an
RPM signal on it, you're still required to run the RPM signal over
to the CPU fan header. But that is a technique for "taking control"
of the setup.

This is an example of an adapter, for bypassing motherboard fan speed
control. Yellow/black carry a constant +12V for the fan, so it runs
full speed. You don't really want the twisted pair as shown. Just the
RPM wire should run to the right-most (monitoring) connector. They've
included the PWM wire as well, and I'd just pull that pin out of the
right-most connector. There's a little tab, over each pin, that releases
and allows the pin and wire to be backed out.

http://fanadaptors.com/wp-content/uploads/2011/04/4-pin_PSU_4-pin_PWM-RPM.jpg

In the diagram here, if you don't like the job the motherboard is doing,
of controlling the CPU fan speed, you have the option of pulling the PWM
(end-most) pin. You keep the "Sense" wire, a.k.a. RPM wire. You need that
signal, because some BIOS will turn off the computer if they detect no
fan RPM signal after a couple seconds of operation. A fan equipped with
RPM sensing, and with a working "Sense" wire is needed, on the CPU fan
header interface.

http://www.techpowerup.com/forums/showthread.php?t=145959

http://img.techpowerup.org/110517/pwm_diagram.jpg

HTH,
Paul

[Dennis]

I replaced the whole CPU cooler and the fan RPM went up to 3120 RPM.
The computer has not shut down since about noon Saturday. The fan on
it was supposed to run 4500 RPM and have a life over four years. I
found numerous complaints on the original fan only lasting 6-7 months.
So, don't use a Cooler Master DK9-7G52A-0L-GP CPU Cooler...