Vista keeps shutting down after a few minutes

S

spencertaylor

Hi all;
I am hoping that you can help me.
I am having a problem w/ Vista locking up
It happens after about 7 - 10 minutes of being on - even at the
Welcome screen.
The monitor goes black and it's like the system is dead.
I have an nForce 3 mobo - MSI 7025, 3800 X2 939 chip, XFX 6800 Ultra

Here is what I tried already:
3 different Nvidia drivers
I ran the memory test from the Vista disc, no problems
I have an identical motherboard from MSI, because I thought the board
was bad - I rma'ed it.
The same thing happens in both boards.
I think the video card is ok - I have a 6800 Ultra

I have been having problems with this new CPU. I had a 3500 single
core and had no problems until I upgraded.
I am going to rma the 3800 X2 to NewEgg.

Any ideas? Very frustrated. I can't use the PC for more that 10
minutes.
 
A

Adam Albright

Hi all;
I am hoping that you can help me.
I am having a problem w/ Vista locking up
It happens after about 7 - 10 minutes of being on - even at the
Welcome screen.
The monitor goes black and it's like the system is dead.
I have an nForce 3 mobo - MSI 7025, 3800 X2 939 chip, XFX 6800 Ultra

Here is what I tried already:
3 different Nvidia drivers
I ran the memory test from the Vista disc, no problems
I have an identical motherboard from MSI, because I thought the board
was bad - I rma'ed it.
The same thing happens in both boards.
I think the video card is ok - I have a 6800 Ultra

I have been having problems with this new CPU. I had a 3500 single
core and had no problems until I upgraded.
I am going to rma the 3800 X2 to NewEgg.

Sounds like you build this system. If yes, and the computer dies
within 7-10 minutes EVERYTIME you turn it on or nearly every time,
that suggests it could be a overheating issue. If you installed the
CPU chip yourself, did you make sure it is FIRMLY attached to the
heatsink, that you used the PROPER amount of compound between the chip
and the heatsink and that the CPU fan is both big enough and running
fast enough? Many people building their own systems either forget to
put on any compound between the CPU chip and the heatsink or more
often they go way overboard and put far too much on, like icing on a
cake. That defeats the purpose.

If you suspect overheating, most BIOS these days do have a option to
show temp readings, if you just stay in BIOS and watch, then start
over if it confirms your CPU is getting too hot.

Carefully remove the fan/heatsink. Next remove the CPU chip. Next get
some rubbing alcohol and put just a small amount on some gauze or some
rag, then carefully remove the gunk. Let dry, then try again. Many web
sites show the right way to apply compound to a CPU chip. Operative
word is SMALL AMOUNT. The right way is closer to applying paint to a
wall, a thin even coat, not like putting motar between bricks.
 
S

spencertaylor

Sounds like you build this system. If yes, and the computer dies
within 7-10 minutes EVERYTIME you turn it on or nearly every time,
that suggests it could be a overheating issue. If you installed the
CPU chip yourself, did you make sure it is FIRMLY attached to the
heatsink, that you used the PROPER amount of compound between the chip
and the heatsink and that the CPU fan is both big enough and running
fast enough? Many people building their own systems either forget to
put on any compound between the CPU chip and the heatsink or more
often they go way overboard and put far too much on, like icing on a
cake. That defeats the purpose.

If you suspect overheating, most BIOS these days do have a option to
show temp readings, if you just stay in BIOS and watch, then start
over if it confirms your CPU is getting too hot.

Carefully remove the fan/heatsink. Next remove the CPU chip. Next get
some rubbing alcohol and put just a small amount on some gauze or some
rag, then carefully remove the gunk. Let dry, then try again. Many web
sites show the right way to apply compound to a CPU chip. Operative
word is SMALL AMOUNT. The right way is closer to applying paint to a
wall, a thin even coat, not like putting motar between bricks.

Hi;
It's not the CPU, it's staying under 110 F under load.
I built many systems and I know how to install a CPU. Over heating is
not the issue.
It seems as though either the CPU is bad, or my video card is bad.
When the PC locks up, it doesn't shutdown. I t stays on, but the video
is gone.
Also, the Nvidia drivers keep crashing when I boot the system up.
The only things left are the CPU and Mobo.
Thanks!
 
J

John Barnes

If you had no problems until you upgraded the CPU, that pretty much tells
you it's probably bad unless you damaged something else swapping it. 110
seems a little hot to me but is nowhere near the maximum I have seen listed.
Mine pretty much runs about 10F above ambient temperatures. I have a 5200+
X2.
 
C

cquirke (MVP Windows shell/user)

On 8 Apr 2007 07:46:37 -0700, "(e-mail address removed)"

First step is to check the hardware. Unplug the mains and disconnect
the hard drive, so the installation and file system don't get screwed
up any more than they have already...

I'd use MemTest86, rather than Vista's RAM check, so that I do NOT
have to put the HD contents at risk (as the Vista check forces you to
do) and so that I have cumulative rather than last-pass results on
screen. I'd leave it running for 24 hours with a different boot CDR
in drive, so that if the system reboots during testing, I'll know.

That would also imply re-contruction, and thus different mileage on
what both Adam and I reckon is your most likely issue ;-)

We'll get to that later, but otherwise I'd want to know about the
power chain, i.e.:
- house mains and wiring
- PSU rating
- all internal power points connected, esp. graphics card
- all fans connected, including graphics card
- all components well seated

Make sure the new CPU is supported by the motherboard in every
respect, i.e. BIOS, base speed, power rating, mech clearance etc.
Sounds like you build this system. If yes, and the computer dies
within 7-10 minutes EVERYTIME you turn it on or nearly every time,
that suggests it could be a overheating issue.

A CPU with no heatsink will usually overheat in seconds, possibly
destructively. A poorly-seated heat sink will usually overheat in
minutes, as you describe. A well-seated heat sink with an unpowered
fan will take more minutes to overheat; maybe as you describe.

Make sure some dumb-ass supplier hasn't stuck a paper serial number
sticker on the heat sink surface of the processor. Yep, we still see
such dumbo practice even into the era of P4 laptops.

Make sure the heat sink snaps down properly, and that the retention
lugs are not broken. Even if you build systems for a living as I do,
make a point of reading the instructionbs whenever a new physical
package comes out, as sometimes best-practice changes. For example,
new mounting mechanics can warp and damage a motherboard that isn't
already mounted to a chassis plate before applying the CPU and sink.
f you installed the CPU chip yourself, did you make sure it is FIRMLY
attached to the heatsink, that you used the PROPER amount of
compound between the chip and the heatsink and that the CPU
fan is both big enough and running fast enough?

He's rebuilt the PC to swap mobos, but if the heat sink's mounting
surfaces are bad, then the problem could persist - as could a
persistent method failure. Other possibilities are a
poorly-ventillated case or underpowered PSU, or he's using a processor
that isn't supported by the mobo, or there are jumpers to be set on
the mobo (should be rare, in 2007).

The modern equivalent to jumpers is NVRAM; maybe he needs to clear
this if swapping CPUs on the same motrherboard?

<rest of Adam's good advice snipped>


---------- ----- ---- --- -- - - - -
On the 'net, *everyone* can hear you scream
 
S

spencertaylor

First step is to check the hardware. Unplug the mains and disconnect
the hard drive, so the installation and file system don't get screwed
up any more than they have already...


I'd use MemTest86, rather than Vista's RAM check, so that I do NOT
have to put the HD contents at risk (as the Vista check forces you to
do) and so that I have cumulative rather than last-pass results on
screen. I'd leave it running for 24 hours with a different boot CDR
in drive, so that if the system reboots during testing, I'll know.


That would also imply re-contruction, and thus different mileage on
what both Adam and I reckon is your most likely issue ;-)

We'll get to that later, but otherwise I'd want to know about the
power chain, i.e.:
- house mains and wiring
- PSU rating
- all internal power points connected, esp. graphics card
- all fans connected, including graphics card
- all components well seated


Make sure the new CPU is supported by the motherboard in every
respect, i.e. BIOS, base speed, power rating, mech clearance etc.


A CPU with no heatsink will usually overheat in seconds, possibly
destructively. A poorly-seated heat sink will usually overheat in
minutes, as you describe. A well-seated heat sink with an unpowered
fan will take more minutes to overheat; maybe as you describe.

Make sure some dumb-ass supplier hasn't stuck a paper serial number
sticker on the heat sink surface of the processor. Yep, we still see
such dumbo practice even into the era of P4 laptops.

Make sure the heat sink snaps down properly, and that the retention
lugs are not broken. Even if you build systems for a living as I do,
make a point of reading the instructionbs whenever a new physical
package comes out, as sometimes best-practice changes. For example,
new mounting mechanics can warp and damage a motherboard that isn't
already mounted to a chassis plate before applying the CPU and sink.


He's rebuilt the PC to swap mobos, but if the heat sink's mounting
surfaces are bad, then the problem could persist - as could a
persistent method failure. Other possibilities are a
poorly-ventillated case or underpowered PSU, or he's using a processor
that isn't supported by the mobo, or there are jumpers to be set on
the mobo (should be rare, in 2007).

The modern equivalent to jumpers is NVRAM; maybe he needs to clear
this if swapping CPUs on the same motrherboard?



On the 'net, *everyone* can hear you scream

Thank you for all the advice. I am certain that the CPU is installed
properly, and that the HSF is installed right as well. I left the PC
running in the BIOS for a while and the temp never ran above 85 F.
110 F was wrong, it's never been above 95K, I was thinking of the 3500
I had installed.

According to MSI, the motherboard supports my CPU 100%. I RMA'ed the
board and the other board behaves exactly like the original.

I tried another video card, and the PC was stable, but the 3800 was
still running 30% lower points in 3DMark 03, which is the same that I
observed with the 6800 Ultra, this one is a 5900 XT that I had given
to a friend, and I have my old scores and was able to make a
comparison.

I am now installing XP on a separate drive. I am going to install SP2
and then run 3DMark 03 and 06. If the results are the same, then the
CPU is either partially incompatible or just busted. I will send it
back to NewEgg.

I appreciate all the help guys!
 
J

John Barnes

You can rma all the cpu's you want, the message is the same as I posted for
you on the previous post. If you want faster timings you have to buy a CPU
that has a core speed higher than your 3500+. There is also a small amount
of extra overhead but that shouldn't affect your tests. The 3800+ is about
10% slower core speed.
 
C

cquirke (MVP Windows shell/user)

On 8 Apr 2007 16:03:26 -0700, "(e-mail address removed)"
Does it do this if you "park" it in CMOS setup?
Thank you for all the advice. I am certain that the CPU is installed
properly, and that the HSF is installed right as well. I left the PC
running in the BIOS for a while and the temp never ran above 85 F.
110 F was wrong, it's never been above 95K, I was thinking of the 3500
I had installed.

I don't know Fareinheit and Kelvin, but can convert... OK, 95K is
unlikely and way too cold at well below -100C, whereas 85F and 110F
are OK. But, look at other device temperatures...
- 70C or less is OK for CPU
- 40C or less is best for motherboard
- 40C or less is best for HD

If you have a well-cooled CPU in a tight, poorly-ventillated case, you
will tend to cook the motherboard and have lockups on that basis.

Similarly, SVGA cards can cook themselves, or be cooked by a hot case
raising the ambient towards which the card's cooling has to tend.

Check the heat sink on the SVGA card is well-seated. Often there are
corner sprung posts, and pressing on a corner of the sink can lift
most of the sink off the chip, and it can sometimes stay like that.
According to MSI, the motherboard supports my CPU 100%. I RMA'ed the
board and the other board behaves exactly like the original.

OK, points away from mobo, unless both are being cooked in the case or
mis-fed by a lame PSU, etc.
I tried another video card, and the PC was stable, but the 3800 was
still running 30% lower points in 3DMark 03, which is the same that I
observed with the 6800 Ultra, this one is a 5900 XT that I had given
to a friend, and I have my old scores and was able to make a
comparison.

OK, whatever.

Fast is no good if it falls on its ass every 10 minutes :-/
I am now installing XP on a separate drive. I am going to install SP2
and then run 3DMark 03 and 06. If the results are the same, then the
CPU is either partially incompatible or just busted. I will send it
back to NewEgg.

Strange case, this. Speaking of cases; does yours have...
- front and rear fans in appropriate directions?
- horizontal rather than vertical PSU?
- passive CPU air duct?
- HD in fan path?


---------- ----- ---- --- -- - - - -
Proverbs Unscrolled #37
"Build it and they will come and break it"
 

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments. After that, you can post your question and our members will help you out.

Ask a Question

Top