A7N8X Deluxe and power shutdown

[johnandpeter]

Please help. We have two identical PCs. Spec is A7N8X deluxe MB,
Athlon 3.2 CPU, Winfast A350FX5900 graphics, 1 gig dual ram, 40gig HD
with XP Home. One PC is perfect but the other, when running graphic
hungry software, shuts down with all fans stopped but front main
switch light remains on. Pushing the switch for 5 seconds switches the
PC off. With the case covers removed the PC takes a little longer to
shutdown. If the PC is immediately switched back on it will shutdown
again right away.

The Asus Probe temp at this point is CPU 49 and MB 28deg. The
interesting point is that the perfect second machine running the same
graphic hungry software operates fine at Asus Probe temps of CPU 58
and MB 35 with no power shutdown.

Could this be a fault with the temp sensor under the CPU on the faulty
machine and if so is there a way to set it to a higher temp before
cutoff.? Any help or advice would be greatly appreciated.

 

[Ben Pope]

Please help. We have two identical PCs. Spec is A7N8X deluxe MB,
Athlon 3.2 CPU, Winfast A350FX5900 graphics, 1 gig dual ram, 40gig HD
with XP Home. One PC is perfect but the other, when running graphic
hungry software, shuts down with all fans stopped but front main
switch light remains on. Pushing the switch for 5 seconds switches the
PC off. With the case covers removed the PC takes a little longer to
shutdown. If the PC is immediately switched back on it will shutdown
again right away.

Certainly sounds heat related.

The Asus Probe temp at this point is CPU 49 and MB 28deg. The
interesting point is that the perfect second machine running the same
graphic hungry software operates fine at Asus Probe temps of CPU 58
and MB 35 with no power shutdown.

Use MBM5 and get the CPU Diode temp... the socket temp isn't a great guide
when diagnosing temperature related problems.

www.ben.pope.name/a7n8x_faq.html#MBM5

What PSU are you using?

Ben

 

[AJ]

On Sun, 30 Nov 2003 19

Could this be a fault with the temp sensor under the CPU on the faulty
machine and if so is there a way to set it to a higher temp before
cutoff.? Any help or advice would be greatly appreciated.

I would set the AGP speed to 66Mhz in the BIOS. Since "auto" is flakey
and that one might be going sky high. I would shut down Probe since
it's not perfect and that might be the trouble. If any fans plugged
into the mobo drop below 2000rpm's that would cause what you describe.
Plug a high speed case fan in the CPU header and use an adapter to get
power for the HSF from a PSU lead. High speed but not high powered
since you don't want to overload the header on the board.

 

[The Old Man]

On Sun, 30 Nov 2003 19


I would set the AGP speed to 66Mhz in the BIOS. Since "auto" is flakey
and that one might be going sky high. I would shut down Probe since
it's not perfect and that might be the trouble. If any fans plugged
into the mobo drop below 2000rpm's that would cause what you describe.
Plug a high speed case fan in the CPU header and use an adapter to get
power for the HSF from a PSU lead. High speed but not high powered
since you don't want to overload the header on the board.

Hi ,
I get this problem too. My case is nice and cool.
It's a nuisance but happens probaby 1 out of 10 times. I can live with it.

Graham.

 

[johnandpeter]

Thank you for advice. Now using MBM5. Results of monitor are -Case
49C, CPU Diode 192C (alarm sounds saying temp is set at 70C) and CPU
Socket 33C.

The CPU Diode temp is the same on the computer that is working fine.
What does all this mean? The PSU is a Q-Tec 400W Dual Fan Gold and the
CPU fan is a Cooler Master Aero 7+.

Many thanks for any further help.

John & Peter.

 

[johnandpeter]

We tried setting AGP speed to 66 but same fault. The fans we are using
are high speed.

Many thanks for your help.

John & Peter.

 

[Ben Pope]

Thank you for advice. Now using MBM5. Results of monitor are -Case
49C, CPU Diode 192C (alarm sounds saying temp is set at 70C) and CPU
Socket 33C.

The CPU Diode temp is the same on the computer that is working fine.
What does all this mean? The PSU is a Q-Tec 400W Dual Fan Gold and the
CPU fan is a Cooler Master Aero 7+.

Looks like you need to configure your sensors, 'cos your cpu diode is not
192°C! My site links to the page that shows you how to configure your
sensors. Be sure to get your case and socket temps the right way round,
they're swapped for different Revisions of the board, and I think you have
yours reversed.

If you check the min/max page you can ensure your voltages are good too.

Ben

 

[Doug Ramage]

Thank you for advice. Now using MBM5. Results of monitor are -Case
49C, CPU Diode 192C (alarm sounds saying temp is set at 70C) and CPU
Socket 33C.

The CPU Diode temp is the same on the computer that is working fine.
What does all this mean? The PSU is a Q-Tec 400W Dual Fan Gold and the
CPU fan is a Cooler Master Aero 7+.

Many thanks for any further help.

John & Peter.

IIRC, your PSU is a bit light on the 12v rail (15 amps?). Something rated at
18 amps+ is likely to be a better option.

 

[johnandpeter]

We discovered that we incorrectly setup MBM to revision <2.0. It is
now set to revision >=2.0 which our MB is. Now the temp readings on
the computer that shuts down are Case 29C, CPU Socket 56C and CPU
Doide 44C. The good PC is Case 32C, CPU Socket 48C and CPU Diode 48C.
These temps are with the computer ticking over without stress working.

Could it be the CPU Socket temp causing the shutdown on the faulty PC?

Thanks again.
John & Peter.

 

[johnandpeter]

Further to he last post (sorry for confusion) we have installed
seti@home and here are the results when running it on faulty machine.
Case 32, Socket 49 and Diode 63. Results without seti@home running
dropped to Case 32, Socket 45 and Diode 55.

The same test on the good PC gave results of
Case 33, Socket 52 and Diode 63 and Case 33, Socket 51 and Diode 56.

The faulty machine did not close down during this experiement. The
temps look as good as the good PC. We'll continue tests.

Many thanks again.
John & Peter.

 

[Ben Pope]

Further to he last post (sorry for confusion) we have installed
seti@home and here are the results when running it on faulty machine.
Case 32, Socket 49 and Diode 63. Results without seti@home running
dropped to Case 32, Socket 45 and Diode 55.

The same test on the good PC gave results of
Case 33, Socket 52 and Diode 63 and Case 33, Socket 51 and Diode 56.

The faulty machine did not close down during this experiement. The
temps look as good as the good PC. We'll continue tests.

It's a bit strange, from your first post it still looks heat related but
your video card probably has a fairly good cooler on it. What I will say is
that your case temps are fairly high, about 8°C above my case temp, which
will affect the cooling ability of any fan/heatsink inside the case.

The other thing could be your voltages, after MBM5 has been running for a
while, what does it report for min voltages?

Ben

 

[johnandpeter]

It's a bit strange, from your first post it still looks heat related but

your video card probably has a fairly good cooler on it. What I will say is
that your case temps are fairly high, about 8°C above my case temp, which
will affect the cooling ability of any fan/heatsink inside the case.

The other thing could be your voltages, after MBM5 has been running for a
while, what does it report for min voltages?

Ben

It's now even more strange. We slowed the cooling fan on the CPU and
monitored the temps as they rose to 70C on the diode and 58C on the
socket and there was no shutdown. So it looks as though the shut down
ISN'T related to temp.

After the computer had been on for 15 mins with normal temps
displayed, we could force a shutdown by bringing in a long web page
and scrolling it down quickly with the mouse wheel and the pc shut
down exactly in the same way as it does when running graphic hungry
software. We swapped the graphics boards around between the good PC
and the faulty one and there was no change with the problem, which
seems to eliminate a fault on the graphics card. Any comments.

Sorry, we aren't sure how to setup MBM5 to get voltages. Any help on
this would be appreciated.

 

[Ben Pope]

It's now even more strange. We slowed the cooling fan on the CPU and
monitored the temps as they rose to 70C on the diode and 58C on the
socket and there was no shutdown. So it looks as though the shut down
ISN'T related to temp.

Isn't related to CPU temp...

After the computer had been on for 15 mins with normal temps
displayed, we could force a shutdown by bringing in a long web page
and scrolling it down quickly with the mouse wheel and the pc shut
down exactly in the same way as it does when running graphic hungry
software. We swapped the graphics boards around between the good PC
and the faulty one and there was no change with the problem, which
seems to eliminate a fault on the graphics card. Any comments.
Odd.

Sorry, we aren't sure how to setup MBM5 to get voltages. Any help on
this would be appreciated.

My FAQ page I linked to earlier has a link to an nForcersHQ article with
exactly what you need...

Ben