Random Shutdowns

  • Thread starter Thread starter YankeeSanta
  • Start date Start date
Y

YankeeSanta

Hello all, just built me a new compy. I have an AMD Athlon 2200+
running at 1800 MHz. No name RAM that is pc2100 (512mb), a Sapphire
Radeon 9600 xt 256 mb RAM and an old 10 gig WD HDD. Also I'm running
an antec 400w power sup. I am having some problems with the compy
though, when I try to play a game (any game) it shuts down, and I have
to turn the master switch off until that little green light on the
mobo shuts off, then, and only then will the thing turn back on. It
runs fine at 1.49 GHz (50c, but that's besides the point). I'm sure
it's not cooling, because I've had and air conditioner right infront
of the thing and it's been idling around 46c. I'm running the vcore at
1.700 and the multiplier is 1.6 x 10. I HAVE NO IDEA WHY IT DOES THIS.
Could it be the RAM? The CPU is oem, could it be overated? Would a
defective hard drive do it? PLEASE SOMEONE I'M TORTURED HERE!
 
Hello all, just built me a new compy. I have an AMD Athlon 2200+
running at 1800 MHz. No name RAM that is pc2100 (512mb), a Sapphire
Radeon 9600 xt 256 mb RAM and an old 10 gig WD HDD. Also I'm running
an antec 400w power sup. I am having some problems with the compy
though, when I try to play a game (any game) it shuts down, and I have
to turn the master switch off until that little green light on the
mobo shuts off, then, and only then will the thing turn back on. It
runs fine at 1.49 GHz (50c, but that's besides the point). I'm sure
it's not cooling, because I've had and air conditioner right infront
of the thing and it's been idling around 46c. I'm running the vcore at
1.700 and the multiplier is 1.6 x 10. I HAVE NO IDEA WHY IT DOES THIS.
Could it be the RAM? The CPU is oem, could it be overated? Would a
defective hard drive do it? PLEASE SOMEONE I'M TORTURED HERE!
Click to expand...

Well, start out with a little test, to absolutely eliminate
processor overheating as the cause.

Install Asus Probe or Motherboard Monitor 5 (mbm.livewiredev.com)
(one or the other, not both, as they don't play nice together
when using the SMBUS).

Bring up the interface for either program, so you can see the
current CPU temperature.

Now, get a copy of Prime95 (mersenne.org). Install and run it.
Select "Torture Test" - the memory model doesn't matter, as
this is probably going to die pretty fast. Start the test
running, then watch the temperature display on the temp
monitor.

If the temp shoots up, or if the test shuts off the computer,
then it is time to pull the HSF. Have a look at the contact
patch of the heatsink - the square of material should be
centered with respect to the silicon die on the CPU. As
I understand it, it is possible to rotate a heatsink 180
degrees, and have the thing misaligned. If only part of the
CPU die is covered by the metal, the die will overheat.

You might also check that the heatsink is sitting flat.
On my first build a few years ago, the handle on a
S370 socket was lodged under my heatsink, and the heatsink
didn't sit flat. Look for stuff like that.

If it passes torture test for hours, and the temps stay
under 65C, then maybe it is time to look at the video card.
But I didn't think a video card could shut off a computer ?

As far as the Antec goes, using the same monitor program,
check the voltages. The power supply will have a label, and
it will say the voltages are regulated to within 5%. See if
any of your voltages are dipping by 10% (allows some room
for measurement error), in which case you'll need to investigate
whether a PSU swap is needed or not.

Paul
 
I ran the torture test on prime95, it didn't get 5 mins into it and
the compy shut down. I checked the history of the temp and voltage,
and the temp, at the time of shut down, was 57c and the voltage was in
it's normal operating perameter. I've had this running at a slower
clock rate, but during the summer when the temp was 64-67, and it
didn't shut down. The only thing that was changed was the clock speed.
Based on this information, I think it's safe to say that the mobo and
CPU overheating aren't the causes of this problem, correct me if i'm
wrong.

What's the next step?
 
I ran prime95's torture test, and it shut it down, but the temp was at
57c at the time. Now, i'm new at this, but is that ok? It's at the
high end, yes, but it's nowhere near the threshold. The voltage line
didn't move at all throughout the test. Thanks for the ideas.


Anymore? What's the next step?


Thank you
Matt
 
I ran prime95's torture test, and it shut it down, but the temp was at
57c at the time. Now, i'm new at this, but is that ok? It's at the
high end, yes, but it's nowhere near the threshold. The voltage line
didn't move at all throughout the test. Thanks for the ideas.


Anymore? What's the next step?


Thank you
Matt
 
I ran prime95's torture test, and it shut it down, but the temp was at
57c at the time. Now, i'm new at this, but is that ok? It's at the
high end, yes, but it's nowhere near the threshold. The voltage line
didn't move at all throughout the test. Thanks for the ideas.


Anymore? What's the next step?


Thank you
Matt
 
I ran the torture test on prime95, it didn't get 5 mins into it and
the compy shut down. I checked the history of the temp and voltage,
and the temp, at the time of shut down, was 57c and the voltage was in
it's normal operating perameter. I've had this running at a slower
clock rate, but during the summer when the temp was 64-67, and it
didn't shut down. The only thing that was changed was the clock speed.
Based on this information, I think it's safe to say that the mobo and
CPU overheating aren't the causes of this problem, correct me if i'm
wrong.

What's the next step?
Click to expand...

To start with, you didn't identify the motherboard.
Is it an A7N8X ?

The AthlonXP 2200+ is 1800MHz at 133MHz clock. That is
133*13.5 multiplier. If the processor is locked, then your
multiplier setting is being ignored. If the processor is
an older one, then the multiplier should work.
http://www.qdi.nl/support/CPUQDISocketA.htm

Have you verified the settings with CPUZ ?
http://www.cpuid.com/cpuz.php

There is a feature called Asus COP (CPU Overheat Protection).
http://usa.asus.com/products/mb/cop.htm

Asus COP could be implemented with W83L785TS Winbond diode
monitoring chip. It is mentioned here. The second link
explains some of the details of the various A7N8X.

http://mbm.livewiredev.com/comp/asus.html
http://www.nforcershq.com/forum/viewtopic.php?t=16364 (A7N8X ?)

The Winbond 8 pin monitor is programmable and sits on the
SMBUS, just like the other hardware monitor chip. I believe
it can be programmed, as to what temperature it trips at.

http://winbond-usa.com/products/winbond_products/pdfs/PCIC/W86L785TS-S.pdf

In any case, the first step is to examine your heatsink install,
like I suggested in my first post. And make sure MBM is monitoring
what you think it should be. The Nforcershq post suggests testing
the dynamic response, so you know whether you are seeing the diode
temp or not. If the diode temp hits 85C, or if the Winbond chip
thinks it is, that might be enough to trip "temp_fault#".

It may not be overheat, but it could be the overheat detection
chip that is turning the computer off. That chip could malfunction,
but eliminate the HSF install as the problem first.

Why not post the values of the 3.3, 5.0, and 12.0 PSU voltages
from MBM just before shutdown ? So I can see if they are
within spec :-) After all, it could also be the PSU shutting
down on internal overload.

If the green motherboard LED winked out at shutdown, that would
suggest the power supply. (The green LED indicates the flow of
+5VSB, and if +5VSB is lost for even a second, the PSU will shut
down the main outputs and the PSU fan.)

HTH,
Paul
 
To start with, you didn't identify the motherboard.
Is it an A7N8X ?

The AthlonXP 2200+ is 1800MHz at 133MHz clock. That is
133*13.5 multiplier. If the processor is locked, then your
multiplier setting is being ignored. If the processor is
an older one, then the multiplier should work.
http://www.qdi.nl/support/CPUQDISocketA.htm

Have you verified the settings with CPUZ ?
http://www.cpuid.com/cpuz.php

There is a feature called Asus COP (CPU Overheat Protection).
http://usa.asus.com/products/mb/cop.htm

Asus COP could be implemented with W83L785TS Winbond diode
monitoring chip. It is mentioned here. The second link
explains some of the details of the various A7N8X.

http://mbm.livewiredev.com/comp/asus.html
http://www.nforcershq.com/forum/viewtopic.php?t=16364 (A7N8X ?)

The Winbond 8 pin monitor is programmable and sits on the
SMBUS, just like the other hardware monitor chip. I believe
it can be programmed, as to what temperature it trips at.

http://winbond-usa.com/products/winbond_products/pdfs/PCIC/W86L785TS-S.pdf

In any case, the first step is to examine your heatsink install,
like I suggested in my first post. And make sure MBM is monitoring
what you think it should be. The Nforcershq post suggests testing
the dynamic response, so you know whether you are seeing the diode
temp or not. If the diode temp hits 85C, or if the Winbond chip
thinks it is, that might be enough to trip "temp_fault#".

It may not be overheat, but it could be the overheat detection
chip that is turning the computer off. That chip could malfunction,
but eliminate the HSF install as the problem first.

Why not post the values of the 3.3, 5.0, and 12.0 PSU voltages
from MBM just before shutdown ? So I can see if they are
within spec :-) After all, it could also be the PSU shutting
down on internal overload.

If the green motherboard LED winked out at shutdown, that would
suggest the power supply. (The green LED indicates the flow of
+5VSB, and if +5VSB is lost for even a second, the PSU will shut
down the main outputs and the PSU fan.)

HTH,
Paul
Click to expand...

A7V8X-X
 
To start with, you didn't identify the motherboard.
Is it an A7N8X ?

The AthlonXP 2200+ is 1800MHz at 133MHz clock. That is
133*13.5 multiplier. If the processor is locked, then your
multiplier setting is being ignored. If the processor is
an older one, then the multiplier should work.
http://www.qdi.nl/support/CPUQDISocketA.htm

Have you verified the settings with CPUZ ?
http://www.cpuid.com/cpuz.php

There is a feature called Asus COP (CPU Overheat Protection).
http://usa.asus.com/products/mb/cop.htm

Asus COP could be implemented with W83L785TS Winbond diode
monitoring chip. It is mentioned here. The second link
explains some of the details of the various A7N8X.

http://mbm.livewiredev.com/comp/asus.html
http://www.nforcershq.com/forum/viewtopic.php?t=16364 (A7N8X ?)

The Winbond 8 pin monitor is programmable and sits on the
SMBUS, just like the other hardware monitor chip. I believe
it can be programmed, as to what temperature it trips at.

http://winbond-usa.com/products/winbond_products/pdfs/PCIC/W86L785TS-S.pdf

In any case, the first step is to examine your heatsink install,
like I suggested in my first post. And make sure MBM is monitoring
what you think it should be. The Nforcershq post suggests testing
the dynamic response, so you know whether you are seeing the diode
temp or not. If the diode temp hits 85C, or if the Winbond chip
thinks it is, that might be enough to trip "temp_fault#".

It may not be overheat, but it could be the overheat detection
chip that is turning the computer off. That chip could malfunction,
but eliminate the HSF install as the problem first.

Why not post the values of the 3.3, 5.0, and 12.0 PSU voltages
from MBM just before shutdown ? So I can see if they are
within spec :-) After all, it could also be the PSU shutting
down on internal overload.

If the green motherboard LED winked out at shutdown, that would
suggest the power supply. (The green LED indicates the flow of
+5VSB, and if +5VSB is lost for even a second, the PSU will shut
down the main outputs and the PSU fan.)

HTH,
Paul
Click to expand...

At the time of shut down the voltages were 12.224, 4.811, 3.312,
VCore is set at 1.600 and that was 1.648. I'm going to take the comp
apart now and take a look at the heatsink and fan. Thanks again.
 
To start with, you didn't identify the motherboard.
Is it an A7N8X ?

The AthlonXP 2200+ is 1800MHz at 133MHz clock. That is
133*13.5 multiplier. If the processor is locked, then your
multiplier setting is being ignored. If the processor is
an older one, then the multiplier should work.
http://www.qdi.nl/support/CPUQDISocketA.htm

Have you verified the settings with CPUZ ?
http://www.cpuid.com/cpuz.php

There is a feature called Asus COP (CPU Overheat Protection).
http://usa.asus.com/products/mb/cop.htm

Asus COP could be implemented with W83L785TS Winbond diode
monitoring chip. It is mentioned here. The second link
explains some of the details of the various A7N8X.

http://mbm.livewiredev.com/comp/asus.html
http://www.nforcershq.com/forum/viewtopic.php?t=16364 (A7N8X ?)

The Winbond 8 pin monitor is programmable and sits on the
SMBUS, just like the other hardware monitor chip. I believe
it can be programmed, as to what temperature it trips at.

http://winbond-usa.com/products/winbond_products/pdfs/PCIC/W86L785TS-S.pdf

In any case, the first step is to examine your heatsink install,
like I suggested in my first post. And make sure MBM is monitoring
what you think it should be. The Nforcershq post suggests testing
the dynamic response, so you know whether you are seeing the diode
temp or not. If the diode temp hits 85C, or if the Winbond chip
thinks it is, that might be enough to trip "temp_fault#".

It may not be overheat, but it could be the overheat detection
chip that is turning the computer off. That chip could malfunction,
but eliminate the HSF install as the problem first.

Why not post the values of the 3.3, 5.0, and 12.0 PSU voltages
from MBM just before shutdown ? So I can see if they are
within spec :-) After all, it could also be the PSU shutting
down on internal overload.

If the green motherboard LED winked out at shutdown, that would
suggest the power supply. (The green LED indicates the flow of
+5VSB, and if +5VSB is lost for even a second, the PSU will shut
down the main outputs and the PSU fan.)

HTH,
Paul
Click to expand...

Here is something interesting: I dl MBM and got it up and
running, and it's telling me that the cpu is at 44c, and ASUS PC Probe
is telling me that it's at 50C, I don't know what this means, but it's
interesting :)
 
Are you running WindowsXP ?
If so go into Control Panel, System;
you will find an « error » tab. Open it and
disable everything.
 
To start with, you didn't identify the motherboard.
Is it an A7N8X ?

The AthlonXP 2200+ is 1800MHz at 133MHz clock. That is
133*13.5 multiplier. If the processor is locked, then your
multiplier setting is being ignored. If the processor is
an older one, then the multiplier should work.
http://www.qdi.nl/support/CPUQDISocketA.htm

Have you verified the settings with CPUZ ?
http://www.cpuid.com/cpuz.php

There is a feature called Asus COP (CPU Overheat Protection).
http://usa.asus.com/products/mb/cop.htm

Asus COP could be implemented with W83L785TS Winbond diode
monitoring chip. It is mentioned here. The second link
explains some of the details of the various A7N8X.

http://mbm.livewiredev.com/comp/asus.html
http://www.nforcershq.com/forum/viewtopic.php?t=16364 (A7N8X ?)

The Winbond 8 pin monitor is programmable and sits on the
SMBUS, just like the other hardware monitor chip. I believe
it can be programmed, as to what temperature it trips at.

http://winbond-usa.com/products/winbond_products/pdfs/PCIC/W86L785TS-S.pdf

In any case, the first step is to examine your heatsink install,
like I suggested in my first post. And make sure MBM is monitoring
what you think it should be. The Nforcershq post suggests testing
the dynamic response, so you know whether you are seeing the diode
temp or not. If the diode temp hits 85C, or if the Winbond chip
thinks it is, that might be enough to trip "temp_fault#".

It may not be overheat, but it could be the overheat detection
chip that is turning the computer off. That chip could malfunction,
but eliminate the HSF install as the problem first.

Why not post the values of the 3.3, 5.0, and 12.0 PSU voltages
from MBM just before shutdown ? So I can see if they are
within spec :-) After all, it could also be the PSU shutting
down on internal overload.

If the green motherboard LED winked out at shutdown, that would
suggest the power supply. (The green LED indicates the flow of
+5VSB, and if +5VSB is lost for even a second, the PSU will shut
down the main outputs and the PSU fan.)

HTH,
Paul
Click to expand...

With much thanks to you paul, I have gotten the temperature down
to 35c on both of the monitering programs, and the BIOS as well! The
computer works great now and hase been doing things that it never has
been! I took the fan and heat sink off and rotated it 180 degrees and
also did and very meticulas preperation and execution with the thermal
greese. Thank you so much, i'm going to go play some Far Cry now!
 
With much thanks to you paul, I have gotten the temperature down
to 35c on both of the monitering programs, and the BIOS as well! The
computer works great now and hase been doing things that it never has
been! I took the fan and heat sink off and rotated it 180 degrees and
also did and very meticulas preperation and execution with the thermal
greese. Thank you so much, i'm going to go play some Far Cry now!
Click to expand...

So does this mean you'll read the directions before putting one together
next time? ;p

Ed
 
Did you run memtest (www.memtest.org) If memtest give an error, believe me,
your memory is bad. . And did you check your log (xp), that can also be very
helpfull.

Driekes
 
Did you ckecked your mainboad on bad capacitors? If seen running (!) systems
with complete blowed out capacitors. If you have them, you could replace
them, otherwise, change your mobo.
 
Did you run memtest (www.memtest.org) If memtest give an error, believe me,
your memory is bad. . And did you check your log (xp), that can also be very
helpfull.

Driekes
Click to expand...
I have some Crucial PC2100 RAM installed on my 3-year-old
A7M266/Athlon XP 2100+ running Windows XP SP1. It has been rock
stable its entire life. I ran memtest in one of my geekier moments
just for the heck of it, and I DO get errors reported, but I've NEVER
had the slightest instability in this computer. My point is that an
error in memtest is not a 100%-reliable indicator that a computer's
problem is a bad DIMM. I'm not saying that memtest lies, just that it
can find errors that turn out to be trivial and of no practical
importance -- i.e., a memtest error does not always mean that the
memory is "bad" if "bad" means that it needs to be replaced.
(BTW, my memory tested perfect with Microsoft's free RAM-testing
program.)


Ron
 
I found out that if memtest gives only one error it is wrong. I would say,
change your mem and check out again.
 
Back
Top