Unstable A7V8X-X

[ruflicks]

Hi all.

I am looking some help from you because my A7V8X-X motherboard has
weird stability issues. Don't get me wrong, I'm not here to trash on
asus products, I know they have a very good reputation and that a lot
of you people are happy with them (this is why I bought an asus mobo).

So here is my problem: a lot of time, randomly, my system will freeze
(no blue screen, no reboot). The system will just stop responding. I
still can see my windows desktop, but mouse won't work, ctrl-alt-del
won't work, sound will stop. Guess that is what you would call freeze.
It will freeze pretty often, sometimes only after a few minutes the
computer is on (that is pretty annoying when you'are installing an
os).
So why do i complain about this on an asus newsgroup ? Because i am
now convinced that this freeze problem is related to the motherboard.

A few things i have checked before posting here:
- software: system does freeze on windows 98, windows xp home, linux
mandrake, linux fedora
- power supply: +12V->stable 12.2, +5V->stable 4.8, +3.3->stable 3.2,
350W seems ok
- Vcore: 1.664V stable
- cpu temperature: 46 to 50°C seems ok
- motherboard temperature: 30 to 33°C seems ok
- fan speed: stable 2600 rpm
- ram: does freeze with 266 and 333 DDR (kingston or something, the
guy in the store told me it's good stuff)
- hard disk: freeze with various hard disks
- other disks: freeze even if all disks (cdrom, floppy) but master
hard-disk are unplugged
- graphic card: freeze with agp cards (ati rage) and pci card (S3
virge)
- cpu: freeze with an athlon xp 2000+ and an athlon xp 2400+ (both
working ok on my abit NF7). Never overclocked i swear.

All components working ok with my NF7!!! System always freezing with
my asus!!!
So that is why i am finally _convinced_ it is a mobo issue.

A few things i've done related to the A7V8X-X:
- read the ****ing manual
- bios update: v1013, still freezing
- cleared rtc ram: not working
- on-board sound and lan off: freezing
- tweaked agp, udma: still freezing

Finally, found that the only way to have my system a bit more stable
was to underclock my cpu (running my xp 2400+ at 1500MHz=1800+? now,
_should_ run at 2000MHz). Now my A7V8X-X is working a little better
(but still freezing sometimes). Found on the net that i was not the
only one that had the very same problem and had to underclock(!).

So can anyone explain me what i am doing wrong ? Is there a way to get
my problem solved ?
Hope someone will help.

Best regards.

 

[dino]

if it has an option for "Spread Spectrum" for the CPU..disable it..also
disable the same option for the AGP if that M/B has it....I know my A7N8X-X
and A7N8X-deluxe both have it...I have read where that can cause some
stability issues.

 

[radiobirdman]

- ram: does freeze with 266 and 333 DDR (kingston or something, the

guy in the store told me it's good stuff)

I would not trust this, if I don't know the brand

This is prolly the problem with your system.

Fan seems also spin slow.

And check the CPU settings.

 

[Paul]

Hi all.

I am looking some help from you because my A7V8X-X motherboard has
weird stability issues. Don't get me wrong, I'm not here to trash on
asus products, I know they have a very good reputation and that a lot
of you people are happy with them (this is why I bought an asus mobo).

The first thing I would do, is check that the heatsink is installed
correctly. It is easy to rotate the heatsink 180 degrees, and then
the contact square doesn't line up with the die. What happens then,
is part of the die is poorly cooled, and the rest is cooled. The
poorly cooled part will have its timing screwed up by the high temps.
So, I would verify that the contact patch on the heatsink lines up
properly, with the silicon die centered in the middle of the
contact zone on the CPU heatsink.

After you've checked than, work on the memory. You didn't say
how many sticks of memory you've got, but if you have one, put
the stick in the slot furthest from the processor. If using two
sticks, use slot 1 and slot 3 for the RAM. Test the memory with
memtest86 from memtest.org. It loads from a floppy or an ISO
CD. If you have more than one stick of memory, test it one stick
at a time, then try combinations of sticks after that. You may
find that just one stick is flaky, and the other works.

If the memory is stable, then it is safe to boot back into Windows.
Then, you can run Prime95 from mersenne.org, and it will load
the processor and help test whether things are stable. If the
processor crashes real fast with Prime95, and yet the memory
tests good, it could be the processor is cooked or the
Northbridge is bad. Since you've already tested the processor
on another board, that would leave the Northbridge. In that
case, if there is any time left on the 3 year warranty, it is
time to RMA.

Sometimes the instability is video card related, and could be
caused by the AGP interface. There have been chipsets in the
past that were noteworthy for their bad AGP interfaces, and
in some cases, reducing the AGP rate from 4X to 2X would fix
them (i.e. bad design by Via/SIS/Ali etc). I like a 3D program
that can be left looping all night, to use as a stability test,
and the demo mode of 3DMark2001SE is my favorite tool for that.

Many of the AGP/chipset problems have been cured in the current
generations of boards. I hear way fewer complaints about Via
and SIS, so things have changed for the positive there. Via is
now a major player in making Athlon64 systems work, and SIS
components are used in all sorts of mass marketed computers,
and even some Asus motherboards. But, for many users who got
bad boards in the past, they aren't likely to forgive and forget.
(I certainly wouldn't buy another board with an Ali chipset on it.)

Post back what you find.

Paul

 

[ruflicks]

The first thing I would do, is check that the heatsink is installed

correctly. It is easy to rotate the heatsink 180 degrees, and then
the contact square doesn't line up with the die. What happens then,
is part of the die is poorly cooled, and the rest is cooled. The
poorly cooled part will have its timing screwed up by the high temps.
So, I would verify that the contact patch on the heatsink lines up
properly, with the silicon die centered in the middle of the
contact zone on the CPU heatsink.

Easy to rotate the heatsink 180 degrees ? Come on... Tried with the
standard amd fan, then with a Titan fan. Used lot of thermal grease.
Still freezing.

After you've checked than, work on the memory. You didn't say
how many sticks of memory you've got, but if you have one, put
the stick in the slot furthest from the processor. If using two
sticks, use slot 1 and slot 3 for the RAM. Test the memory with
memtest86 from memtest.org. It loads from a floppy or an ISO
CD. If you have more than one stick of memory, test it one stick
at a time, then try combinations of sticks after that. You may
find that just one stick is flaky, and the other works.

The two sticks i'm trying on my machine are:
256 mb MDT DDR 400 (www.mdt-usa.com)
256 mb Kingston DDR 333 (www.kingston.com)

No, I don't use both sticks together. Tried that memtest, booted from
a floppy. Both modules passed the standard test 2 times with no error.

Thinking of that, may it be I need a DDR 266 module ? But both are
working ok at 266 on my NF7.

If the memory is stable, then it is safe to boot back into Windows.
Then, you can run Prime95 from mersenne.org, and it will load
the processor and help test whether things are stable. If the
processor crashes real fast with Prime95, and yet the memory
tests good, it could be the processor is cooked

Tried this Prime95. 100% cpu usage for like an hour with both cpu
(tested them on my NF7)

or the
Northbridge is bad. Since you've already tested the processor
on another board, that would leave the Northbridge. In that
case, if there is any time left on the 3 year warranty, it is
time to RMA.

Didn't find anything about a 3 year warranty on the asus website. Can
someone tell me more about that?

Sometimes the instability is video card related, and could be
caused by the AGP interface. There have been chipsets in the
past that were noteworthy for their bad AGP interfaces, and
in some cases, reducing the AGP rate from 4X to 2X would fix
them (i.e. bad design by Via/SIS/Ali etc). I like a 3D program
that can be left looping all night, to use as a stability test,
and the demo mode of 3DMark2001SE is my favorite tool for that.

Already tried that AGP rate tweaking. Freezing with an old pci video
card (S3 virge DX).

Many of the AGP/chipset problems have been cured in the current
generations of boards. I hear way fewer complaints about Via
and SIS, so things have changed for the positive there. Via is
now a major player in making Athlon64 systems work, and SIS
components are used in all sorts of mass marketed computers,
and even some Asus motherboards. But, for many users who got
bad boards in the past, they aren't likely to forgive and forget.
(I certainly wouldn't buy another board with an Ali chipset on it.)

There is a VIA KT400 chipset on my board. May it be i got a crappy
board ? How is it even possible ? Did asus ship boards they didn't
even test ? I must say I am really pissed right now. I may start
trashing on asus products for real.

Anyway, thanks for you answers.

 

[Paul]

The first thing I would do, is check that the heatsink is installed
correctly. It is easy to rotate the heatsink 180 degrees, and then
the contact square doesn't line up with the die. What happens then,
is part of the die is poorly cooled, and the rest is cooled. The
poorly cooled part will have its timing screwed up by the high temps.
So, I would verify that the contact patch on the heatsink lines up
properly, with the silicon die centered in the middle of the
contact zone on the CPU heatsink.

Easy to rotate the heatsink 180 degrees ? Come on... Tried with the
standard amd fan, then with a Titan fan. Used lot of thermal grease.
Still freezing.[/QUOTE]

Some people put the heatsink on wrong, the temperature in the BIOS
monitor reads normal, but the board crashes. When they rotate the
heatsink, it works for them. It was just a suggestion...

The two sticks i'm trying on my machine are:
256 mb MDT DDR 400 (www.mdt-usa.com)
256 mb Kingston DDR 333 (www.kingston.com)

No, I don't use both sticks together. Tried that memtest, booted from
a floppy. Both modules passed the standard test 2 times with no error.

Thinking of that, may it be I need a DDR 266 module ? But both are
working ok at 266 on my NF7.

No. Memory is backward compatible, so a DDR400 or DDR333 stick can
run at DDR266 without any problem.

Tried this Prime95. 100% cpu usage for like an hour with both cpu
(tested them on my NF7)

You tested Prime95 on the A7V8X-X with your normal processor ?
I was hoping that would show if the problem was with the Northbridge
or with the Vcore circuit. It is surprising when Prime95 will run
without a problem, and yet the motherboard freezes. Maybe it
is a driver or other software problem ? Like, maybe if some
power management feature (ACPI) is activating.

Have you looked in the Event Viewer in Windows, to see if there
are any error messages logged ? Perhaps there is some evidence of
what the problem might be in there.

Didn't find anything about a 3 year warranty on the asus website. Can
someone tell me more about that?

I'm not certain the details of the warranty are exactly the same in
all countries. But here is a page to examine anyway...

http://www.asus.com/support/english/techref/warranty/index.aspx
http://Helpdesk.asus.com/logon.asp?URL=/Default.asp

Already tried that AGP rate tweaking. Freezing with an old pci video
card (S3 virge DX).


There is a VIA KT400 chipset on my board. May it be i got a crappy
board ? How is it even possible ? Did asus ship boards they didn't
even test ? I must say I am really pissed right now. I may start
trashing on asus products for real.

Anyway, thanks for you answers.

Some other things you can investigate.

I really expected Prime95 to have a problem.

Sometimes, the Vcore circuit on a motherboard is a little weak, and
under load, or when unloaded (processor goes to idle), the voltage
can be out of spec. A workaround for one of these cases, is to
adjust the Vcore setting manually in the BIOS. Add as much as 0.1V
to the normal setting, and see if that makes any difference to
stability.

Along the same theme, the AMD boards draw a lot of +5V, because +5V
is used for the Vcore converter. While running in Windows, you
can use a program like Motherboard Monitor (MBM5) from
http://mbm.livewiredev.com/download.html - only use one monitoring
program at a time. If you already have Asus Probe loaded, that
is good enough.

Check to see if the power supply has at least 4.75V on the 5V rail.
If the voltage is weaker than that, while Prime95 is running, then
a different power supply could help you.

It is normal for Vcore to "droop" under load. I'm just hoping that
a manual adjustment of Vcore can make enough of a difference for
you, that you won't have to RMA the board.

HTH,
Paul

 

[ruflicks]

Some people put the heatsink on wrong, the temperature in the BIOS

monitor reads normal, but the board crashes. When they rotate the
heatsink, it works for them. It was just a suggestion...

Sorry if I sounded rude. Checked on the net, heatsink is ok.

No. Memory is backward compatible, so a DDR400 or DDR333 stick can
run at DDR266 without any problem.

Didn't know that. That is pretty cool. Is it ok to use DDR333 and
DDR400 on the same board ?

You tested Prime95 on the A7V8X-X with your normal processor ?

No, last time i wrote, had just tested both my processor on my abit
NF7 with Prime95. Didn't even dare to test in on my A7V8X as i was
convinced it would crash right away. Just to be sure, tested my athlon
2000+ tonight with prime 95. Used the torture test. Would you believe
it, with two of the tests (small FFT and inplace FFT), the computer
didn't freeze for like an hour (very rare). Then i ran the third test
(blend). It took only 3 seconds to freeze. Found that weird, so tried
it again and again, got still the same results. The two first tests
use cpu a lot, but no ram at all. The third test uses both ram and
cpu. So i can say now (i guess) that my problem is ram related. But as
i said before, tested my ram with memtest and ram is ok on my abit
motherboard.

I was hoping that would show if the problem was with the Northbridge

Found on the net that the VIA KT400 Northbridge was handling memory.
Do you think i could have a chipset not working properly ?

I'm not certain the details of the warranty are exactly the same in
all countries. But here is a page to examine anyway...

http://www.asus.com/support/english/techref/warranty/index.aspx
http://Helpdesk.asus.com/logon.asp?URL=/Default.asp

3 years warranty for all motherboards! That is good news. Next thing
will be to convince my dealer that i never overclocked, that my
heatsink is ok, my power supply is ok, my ram is ok, my cpu is ok, my
hd is ok, my software is ok, and that i really need Prime95 running on
my system. Then maybe, he will send it back to asus. Knowing the guy,
it's gonna be hard.

Some other things you can investigate.

I really expected Prime95 to have a problem.

Sometimes, the Vcore circuit on a motherboard is a little weak, and
under load, or when unloaded (processor goes to idle), the voltage
can be out of spec. A workaround for one of these cases, is to
adjust the Vcore setting manually in the BIOS. Add as much as 0.1V
to the normal setting, and see if that makes any difference to
stability.

Tried that(1,8V). Took a few second to freeze with the Prime95 blend
torture test.

Along the same theme, the AMD boards draw a lot of +5V, because +5V
is used for the Vcore converter. While running in Windows, you
can use a program like Motherboard Monitor (MBM5) from
http://mbm.livewiredev.com/download.html - only use one monitoring
program at a time. If you already have Asus Probe loaded, that
is good enough.
Check to see if the power supply has at least 4.75V on the 5V rail.
If the voltage is weaker than that, while Prime95 is running, then
a different power supply could help you.

Asus Probe tells me my power supply has 4.8 to 4.9V on the 5V rail (at
least before starting the blend torture test, then i can't say cause
it freezes).

It is normal for Vcore to "droop" under load. I'm just hoping that
a manual adjustment of Vcore can make enough of a difference for
you, that you won't have to RMA the board.
HTH,
Paul

Thank you Paul for your help. I really appreciated it. Sorry if i
sounded rude. Was just a little angry about asus and via. Are new asus
card with the nvidia chipset more stable ?

Regards.