P4GD1 - Nothing but trouble

[Guest]

Archived from groups: alt.comp.periphs.mainboard.asus (More info?)

Hello,

to make a long story short I've so far experienced only Problems with my
new PC, based on the ASUS P4GD1 mainboard. The system will reboot/become
instable every once I a while (I cannot reliably reproduce it).

While other malfunctions have shown up in various components upon
assembly and first tests (defective DDR RAM-chip, defective Power Supply
shorting out and dying), I have now replaced all components except the
P4GD1 mainboard. I tried to send it back to the vendor, but they tested
it and claimed there are no errors, yet my system remains instable and
barely usable for any kind of serious work.

I think I have this tracked down to so called page faults, which cause
Windows XP to throw DRIVER_IRQL_not_less_or_equal blue screens, Linux to
oops refering to memory problems and FreeBSD to reboot with messages
like "panic: bad pte". I have thus tested, tested, swapped and tested
the RAM (Kingston PC400, double-sided) with tools like memtest86 which
showed no errors during an interval of 8+ hours.

Any input is greatly appreciated - I'm at a loss at what to do and how
to solve this problem. I'm hesitant to buy a new motherboard, since the
combination of the newly purchased Socket 478 CPU and the horribly
expensive PCI-Express Graphics card pretty much leaves the P4GD1 as only
option.

Again, any pointers/help/tips/comments are greatly appreciated.

Greets,
S.j.

 

[Paul]

Archived from groups: alt.comp.periphs.mainboard.asus (More info?)

In article <d4ajo1$8f3$1@online.de>, "S. j. Oth" <sjoh@xmail.net> wrote:

> Hello,
>
> to make a long story short I've so far experienced only Problems with my
> new PC, based on the ASUS P4GD1 mainboard. The system will reboot/become
> instable every once I a while (I cannot reliably reproduce it).
>
> While other malfunctions have shown up in various components upon
> assembly and first tests (defective DDR RAM-chip, defective Power Supply
> shorting out and dying), I have now replaced all components except the
> P4GD1 mainboard. I tried to send it back to the vendor, but they tested
> it and claimed there are no errors, yet my system remains instable and
> barely usable for any kind of serious work.
>
> I think I have this tracked down to so called page faults, which cause
> Windows XP to throw DRIVER_IRQL_not_less_or_equal blue screens, Linux to
> oops refering to memory problems and FreeBSD to reboot with messages
> like "panic: bad pte". I have thus tested, tested, swapped and tested
> the RAM (Kingston PC400, double-sided) with tools like memtest86 which
> showed no errors during an interval of 8+ hours.
>
> Any input is greatly appreciated - I'm at a loss at what to do and how
> to solve this problem. I'm hesitant to buy a new motherboard, since the
> combination of the newly purchased Socket 478 CPU and the horribly
> expensive PCI-Express Graphics card pretty much leaves the P4GD1 as only
> option.
>
> Again, any pointers/help/tips/comments are greatly appreciated.
>
> Greets,
> S.j.

I don't see too many postings in Google about this board, so it
is hard to know what percentage of boards have problems.

To start with, there is a beta BIOS released Apr14/2005, and
depending on what BIOS release you are currently using, you might
try flash upgrading the BIOS. If there are memory problems,
sometimes Asus makes changes to memory handling within the first
four or five BIOS releases. It can take Asus that long, to get
the basic bugs out of the BIOS, and actually finish writing the
BIOS code.

For a BIOS upgrade, I personally like to use a floppy based
method, booting from a DOS floppy. Depending on the size of the
BIOS file, it may take two floppy diskettes, so that you can
do a backup copy of the current BIOS image onto one floppy,
before actually flash upgrading to the new BIOS. I see there
is a copy of AFUDOS on the Asus download page, and your manual
should explain the proper procedure.

In terms of BIOS settings, I would start with the following.
This is not to cause overclocking, but to make more of the
BIOS settings visible to you:

AI Overclock [Manual]
Performance Mode [Standard]

To properly operate DDR400 memory, the JEDEC spec calls for
2.6V. (2.5V is appropriate for DDR333 memory or slower). Use
the following setting, to aid the operation of the memory. No
memory should be damaged by such a setting:

Memory Voltage [2.70V]

Since you are passing memtest86+ for eight hours, download a
copy of the free Prime95 from mersenne.org . There are Windows
and Linux versions. There is a testing option in Prime95,
called "Torture Test", which performs a calculation with a
known result. A chunk of memory will be used for testing,
which makes this test appropriate for testing processor,
Northbridge, and memory at the same time.

First, run the system at FSB800/DDR400. That should be the
default for the system. To verify the BIOS settings are
working properly, you can use www.cpuid.com/cpuz in
Windows, to see the various system clock frequencies and
memory settings.

Run Prime95 and see if there are any rounding errors or other
errors listed. There should be no errors shown, for the
system to be in good shape.

If you see errors, go back to the BIOS and try

DRAM Frequency [DDR333]

This should run the RAM slower than normal, and will help
the memory if it is not performing well.

Run Prime95 again, and see if the results are error free
or not. It can take several hours for Prime95 to do its
job, due to the amount of memory it uses.

By slowing the memory down, what I am trying to distinguish
between, is a defective processor, versus defective memory.
When the memory is run slower, the memory should not be
stressed, and continued failures could mean the processor
is bad.

Another important area for stability, is the power supply.
What brand and model number of power supply are you using ?
For a basic system, the label on the side of the supply should
be rated for at least 12V @ 15A, and there are some supplies
that provide less than this. Do you have a lot of power hungry
components in the system ?

HTH,
Paul

 

[Guest]

Archived from groups: alt.comp.periphs.mainboard.asus (More info?)

Hi,

First off, thank you very much for your suggestions!! It is much
appreciated.

Paul wrote:
> To start with, there is a beta BIOS released Apr14/2005, and
> depending on what BIOS release you are currently using, you might
> try flash upgrading the BIOS.

The Board shipped with the 1002 revision of the BIOS. I have since
upgraded to the 1003 and finally the beta 1005 rev, because I'm
desperate No difference in behaviour. Another thing I've noticed
just today is an increasing likelyhood of CRC-Errors packing and
unpacking files (.RAR, .CAB). Some applications crash randomly. This is
on WinXP SP2.

> AI Overclock [Manual]
> Performance Mode [Standard]
> Memory Voltage [2.70V]
> Since you are passing memtest86+ for eight hours, download a
> copy of the free Prime95 from mersenne.org . There are Windows
> [..]
> First, run the system at FSB800/DDR400.
> If you see errors, go back to the BIOS and try
> DRAM Frequency [DDR333]

I'll do this once I get home. Thank you for the suggestions!!

> Another important area for stability, is the power supply.

I have a rather high quality Tagan (TG380-U01, 380W) power supply. The
voltages it delivers are constant as far as I can verify. It powers: the
Board, 3.00ghz Northwood P4 CPU, 2x 512MiB RAM, a Leadtek Geforce 6600TD
graphics card, an Adaptec U160 SCSI Controller plus 10krpm SCSI drive by
Maxtor (Atlas 10k IV) and finally a Plextor DVD-RW. Connected via USB is
a Wacom Graphics tablet.

I have verfied the same behaviour using less power hungry devices
(slower CPU - 2.66ghz, Radeon X300 graphics card, Samsung hdd).

Greets,
S.j.

 

[Paul]

Archived from groups: alt.comp.periphs.mainboard.asus (More info?)

In article <d4aud9$pjp$1@online.de>, "S. j. Oth" <sjoh@xmail.net> wrote:

> Hi,
>
> First off, thank you very much for your suggestions!! It is much
> appreciated.
>
> Paul wrote:
> > To start with, there is a beta BIOS released Apr14/2005, and
> > depending on what BIOS release you are currently using, you might
> > try flash upgrading the BIOS.
>
> The Board shipped with the 1002 revision of the BIOS. I have since
> upgraded to the 1003 and finally the beta 1005 rev, because I'm
> desperate No difference in behaviour. Another thing I've noticed
> just today is an increasing likelyhood of CRC-Errors packing and
> unpacking files (.RAR, .CAB). Some applications crash randomly. This is
> on WinXP SP2.
>
> > AI Overclock [Manual]
> > Performance Mode [Standard]
> > Memory Voltage [2.70V]
> > Since you are passing memtest86+ for eight hours, download a
> > copy of the free Prime95 from mersenne.org . There are Windows
> > [..]
> > First, run the system at FSB800/DDR400.
> > If you see errors, go back to the BIOS and try
> > DRAM Frequency [DDR333]
>
> I'll do this once I get home. Thank you for the suggestions!!
>
> > Another important area for stability, is the power supply.
>
> I have a rather high quality Tagan (TG380-U01, 380W) power supply. The
> voltages it delivers are constant as far as I can verify. It powers: the
> Board, 3.00ghz Northwood P4 CPU, 2x 512MiB RAM, a Leadtek Geforce 6600TD
> graphics card, an Adaptec U160 SCSI Controller plus 10krpm SCSI drive by
> Maxtor (Atlas 10k IV) and finally a Plextor DVD-RW. Connected via USB is
> a Wacom Graphics tablet.
>
> I have verfied the same behaviour using less power hungry devices
> (slower CPU - 2.66ghz, Radeon X300 graphics card, Samsung hdd).
>
> Greets,
> S.j.

From your symptom description so far, there seem to be multiple
defective pieces of hardware. I'm having trouble suggesting a
single piece of hardware to focus on. I thought CRC errors are
detected by the drive hardware, and a root cause there could be
power related. Your original post seemed to show memory related
problems.

Right now, I see motherboard and PSU as candidates for swapping.
The TAGAN web page shows excellent specs for the power supply,
3.3@28A, 5@37A (220W max total)
12V@22A
and all I can suggest is borrow another PSU supply and test with
that.

Your motherboard might be drawing about 15 amps from +3.3V, and
not a lot of current from +5V. The 22 amp rating on +12V
should be plenty to do the job of powering the processor and disk.
That makes it hard to explain rationally, why the power is at
fault. (Check the voltages in the hardware monitor, and see
if any of them are on the low side of the nominal value. 5%
regulation is a normal expectation, so the 3.3 should not be
lower than about 3.1V for example. Check the 3.3, 5, and 12V
and see if any of them are low.)

It could be a power regulating device on the motherboard itself
that is defective. Not all the motherboard voltages are monitored,
and some can only be verified with a multimeter.

Now, when your vendor tested the board, did they use
their own power supply, processor and RAM ? Did you recommend
any specific tests to them ? The fact that your motherboard passed
at their facility, suggests that some other component is at fault.
The remaining common element is the power supply.

Very confused...
Paul

 

[Guest]

Archived from groups: alt.comp.periphs.mainboard.asus (More info?)

Paul schrieb:
> AI Overclock [Manual]
> Performance Mode [Standard]
> Memory Voltage [2.70V]

Whith the above set to default values, Prime95 gives errors on the first
test. Whenever I tinkered with those settings, the system would not even
boot but freeze at the WinXP "Welcome" screen.

I think I might have another clue as to what may be wrong: I bought two
512MiB DIMMS of Kingston Value-RAM. They're labeled as
"KVR400X64C3A/512". A datasheet is available here:

http://www.valueram.com/datasheets/KVR400X64C3A_512.pdf

Looking into the Quick-Start Guide it says "The board does not support
dual-sided x16 DIMMS". Is that the problem? The DIMMS have 8 chips on
each side - does that make them x16 DIMMS? I was careful to buy two
DDR400 of the same brand and type, when I should have bought a kit (dual
channel ready) of explicitly single-sided RAM?

Regards,
S.j.

 

[Guest]

Archived from groups: alt.comp.periphs.mainboard.asus (More info?)

Paul wrote:
> [..] regulation is a normal expectation, so the 3.3 should not be
> lower than about 3.1V for example. Check the 3.3, 5, and 12V
> and see if any of them are low.)

I have monitored those lines using the "Asus Probe" Software. All
voltages seem stable and do not fluctuate noticeably. At the moment I'm
writing this they're reading 12.302V, 5.171V, 3.344V and 1.56V Vcore.
The current power supply is a replacement itself. The original one was
specified at 330W (Tagan TG330-U01) which I replaced with the current
one to rule out the (possibly too weak) power supply as a cause of
instability.

> It could be a power regulating device on the motherboard itself
> that is defective. Not all the motherboard voltages are monitored,
> and some can only be verified with a multimeter.

This is a definite possibility. There's nothing I can do to check that,
though.

> Now, when your vendor tested the board, did they use
> their own power supply, processor and RAM ? Did you recommend
> any specific tests to them ? The fact that your motherboard passed
> at their facility, suggests that some other component is at fault.
> The remaining common element is the power supply.

They used their own CPU, RAM and power supply since I purchased those
components at a nearby store. The board was ordered by mail and a
different vendor.


Regards,
S.j.

 

[mercury]

Archived from groups: alt.comp.periphs.mainboard.asus (More info?)

I suggest you run HDD diagnostics - SCSI diags if you have any. I came
across a fellow with a similar repertoire of faults on a totally different
mobo and it was an IDEcontroller error.

So, since you have SCSI, check termination.

"S. j. Oth" <sjoh@xmail.net> wrote in message news:d4aud9$pjp$1@online.de...
> Hi,
>
> First off, thank you very much for your suggestions!! It is much
> appreciated.
>
> Paul wrote:
>> To start with, there is a beta BIOS released Apr14/2005, and
>> depending on what BIOS release you are currently using, you might
>> try flash upgrading the BIOS.
>
> The Board shipped with the 1002 revision of the BIOS. I have since
> upgraded to the 1003 and finally the beta 1005 rev, because I'm desperate
> No difference in behaviour. Another thing I've noticed just today is
> an increasing likelyhood of CRC-Errors packing and unpacking files (.RAR,
> .CAB). Some applications crash randomly. This is on WinXP SP2.
>
>> AI Overclock [Manual] Performance Mode [Standard]
>> Memory Voltage [2.70V]
>> Since you are passing memtest86+ for eight hours, download a
>> copy of the free Prime95 from mersenne.org . There are Windows
>> [..]
>> First, run the system at FSB800/DDR400.
>> If you see errors, go back to the BIOS and try DRAM Frequency [DDR333]
>
> I'll do this once I get home. Thank you for the suggestions!!
>
>> Another important area for stability, is the power supply.
>
> I have a rather high quality Tagan (TG380-U01, 380W) power supply. The
> voltages it delivers are constant as far as I can verify. It powers: the
> Board, 3.00ghz Northwood P4 CPU, 2x 512MiB RAM, a Leadtek Geforce 6600TD
> graphics card, an Adaptec U160 SCSI Controller plus 10krpm SCSI drive by
> Maxtor (Atlas 10k IV) and finally a Plextor DVD-RW. Connected via USB is a
> Wacom Graphics tablet.
>
> I have verfied the same behaviour using less power hungry devices (slower
> CPU - 2.66ghz, Radeon X300 graphics card, Samsung hdd).
>
> Greets,
> S.j.

 

[Guest]

Archived from groups: alt.comp.periphs.mainboard.asus (More info?)

Mercury schrieb:
> I suggest you run HDD diagnostics - SCSI diags if you have any. I came
> across a fellow with a similar repertoire of faults on a totally different
> mobo and it was an IDEcontroller error.

Do you have any recommendation as to which tools to use for diagnostics?
There might just be a problem with the IDE controller, since a memory
dump (after XP crashing) includes the following lines, and I've gone
through a lot of DVD-Rs trying to burn a few ISOs (mostly I/O errors
when the DVD was nearly finished).

Unloaded modules:
f13ca000 f13cd000 DirectIo.sys [..]
f7602000 f7605000 DirectIo.sys [..]
f1585000 f1599000 parport.sys [..]
f588b000 f5896000 imapi.sys [..]
f799f000 f79ae000 redbook.sys [..]
f7b37000 f7b3c000 Cdaudio.SYS [..]
f79bf000 f79cc000 cdrom.sys [..]
f7cc7000 f7cca000 Sfloppy.SYS [..]

> So, since you have SCSI, check termination.

I'd think termination (active LVD terminator is present) is alright,
since the PCI-card + Disk and cabling were inherited from my previous
system where this setup ran fine for two years.

Regards,
S.j.