Battery type and location on Asus P4C800-E Deluxe?

The clock on this motherboard is going whacky, even while AC power
stays on. Perhaps the battery is dying? Not sure if anything else
could cause that.

I took a look at the motherboard but didn't spot the battery. PDF docs
don't say. Anyone know what type, and where it's located?

On Tue, 13 Jan 2009 15:03:12 -0500, _DD wrote:

The clock on this motherboard is going whacky, even while AC power
stays on. Perhaps the battery is dying? Not sure if anything else
could cause that.

I took a look at the motherboard but didn't spot the battery. PDF docs
don't say. Anyone know what type, and where it's located?

In the corner by the memory modules.

The battery is in upper right corner of your motherboard. You couldn't see
it! Perhaps you need to see the eye doctor. It is regular motherboard type 3
Volts battery model CR2032.


"_DD" wrote in message
...
The clock on this motherboard is going whacky, even while AC power
stays on. Perhaps the battery is dying? Not sure if anything else
could cause that.

I took a look at the motherboard but didn't spot the battery. PDF docs
don't say. Anyone know what type, and where it's located?

_DD wrote:
The clock on this motherboard is going whacky, even while AC power
stays on. Perhaps the battery is dying? Not sure if anything else
could cause that.

I took a look at the motherboard but didn't spot the battery. PDF docs
don't say. Anyone know what type, and where it's located?

You can change the battery, but I doubt it would help.

When the system is running, Windows maintains time with devices other
than the RTC. The RTC is mainly intended for timekeeping when the
system is sleeping or disconnected entirely from AC power. The
RTC is always running, but Windows stops reading it after
startup.

One mechanism for timekeeping in the OS, is counting clock tick
interrupts. These would be the same events that give the scheduler
a chance to run, and put the next task in a runnable state.

If some code stays at interrupt level for longer than a clock
tick interrupt interval, that can lead to a lost tick and a
drift in the time. There have been cases, where there was a
bug in the interrupt logic in the chipset, which caused
spurious interrupts and upset time keeping in an unpredictable
way. On the Nforce2 chipset, that bug was so bad, that even
frequent NTP updates to the time, couldn't keep the
drift under control. In Linux, you might try "noapic"
to try to eliminate that particular bug. I'm not
aware of a problem with the P4C800-E chipset, so another
place to look, might be at any hardware plugged into the
computer. Even the separate RAID controller might be
responsible - if you're not dependent on the Promise
20378, disable it in the BIOS and see what happens to
the Windows maintained timekeeping.

Paul

On Tue, 13 Jan 2009 16:42:39 -0500, "007" wrote:

The battery is in upper right corner of your motherboard. You couldn't see
it! Perhaps you need to see the eye doctor. It is regular motherboard type 3
Volts battery model CR2032.

Thanks. Some cables and drive bays cover that area. I'll have to take
them out. Of course, Paul indicates that this is probably not the
problem, so maybe that isn't necessary.


"_DD" wrote in message
.. .
The clock on this motherboard is going whacky, even while AC power
stays on. Perhaps the battery is dying? Not sure if anything else
could cause that.

I took a look at the motherboard but didn't spot the battery. PDF docs
don't say. Anyone know what type, and where it's located?

On Tue, 13 Jan 2009 17:10:32 -0500, Paul wrote:

You can change the battery, but I doubt it would help.

When the system is running, Windows maintains time with devices other
than the RTC. The RTC is mainly intended for timekeeping when the
system is sleeping or disconnected entirely from AC power. The
RTC is always running, but Windows stops reading it after
startup.

....
If some code stays at interrupt level for longer than a clock
tick interrupt interval, that can lead to a lost tick and a
drift in the time. There have been cases, where there was a
bug in the interrupt logic in the chipset.... I'm not
aware of a problem with the P4C800-E chipset,

I've been having lots of trouble with this system. Usually does not
boot, but when it does, it stays up (mostly). I've seen it lose a
couple settings in the CMOS, and thought that the clock was
confirmation of battery problems. But you're right--it drifts even
while the system is running.

The BIOS also reported a high CPU temp, around 62 C or so. And this is
a Northwood, not a Prescott. I cleaned the dust out of the fan and
heatsink, so maybe that will help. (Is there a runtime utility that
will report those numbers while XP is booted?)

You had recommended Knoppix for testing systems with boot problems. I
DL'd their latest image. It came up partway but locked before
completing the boot process. Tried many times. I'm not sure if that's
a compatibility issue, or if it was just turning up the same boot-lock
bug.

Memtest86 and Seatools say the RAM and Drives are good.

I'd love to have some kind of sophisticated hardware debugger for
this. It's been like taking shots in the dark.

_DD wrote:
On Tue, 13 Jan 2009 17:10:32 -0500, Paul wrote:

You can change the battery, but I doubt it would help.

When the system is running, Windows maintains time with devices other
than the RTC. The RTC is mainly intended for timekeeping when the
system is sleeping or disconnected entirely from AC power. The
RTC is always running, but Windows stops reading it after
startup.

...
If some code stays at interrupt level for longer than a clock
tick interrupt interval, that can lead to a lost tick and a
drift in the time. There have been cases, where there was a
bug in the interrupt logic in the chipset.... I'm not
aware of a problem with the P4C800-E chipset,

I've been having lots of trouble with this system. Usually does not
boot, but when it does, it stays up (mostly). I've seen it lose a
couple settings in the CMOS, and thought that the clock was
confirmation of battery problems. But you're right--it drifts even
while the system is running.

The BIOS also reported a high CPU temp, around 62 C or so. And this is
a Northwood, not a Prescott. I cleaned the dust out of the fan and
heatsink, so maybe that will help. (Is there a runtime utility that
will report those numbers while XP is booted?)

You had recommended Knoppix for testing systems with boot problems. I
DL'd their latest image. It came up partway but locked before
completing the boot process. Tried many times. I'm not sure if that's
a compatibility issue, or if it was just turning up the same boot-lock
bug.

Memtest86 and Seatools say the RAM and Drives are good.

I'd love to have some kind of sophisticated hardware debugger for
this. It's been like taking shots in the dark.


Speedfan from almico.com gives you the ability to measure
temps and voltages. I think my Northwood on the P4C800-E
ran around 43C or so. It was a 2.8C overclocked slightly
(10% or so).

If you boot from Knoppix, you try to note down which line
the thing is failing at. And not all the lines on the
screen are explicit enough to give good feedback. Still,
it is better than no feedback at all. I've booted several
Linux distros on the P4C800-E without a squeak or complaint.

The VBAT path on the computer, is the "diode OR"
of two sources, like this.

PSU---5VSB --- regulator --- 3VSB ---+
|
Dual
Diode ------ To Southbridge
| (Home of RTC and
CR2032 --- 1Kohm -----+ 256 byte CMOS)
resistor

The higher of the two voltages, is the one
that conducts ("wins"). When the computer is turned off
at the back, the 5VSB is gone and so the 3VSB
goes too. That leaves the CR2032, and at
that instant in time, the dual diode is conducting
through the lower arm.

If the computer is in S3 Suspend to RAM state, the 5VSB is
present. 5VSB is also present when the computer
is in the running state. In that condition,
the CR2032 is cut off, and virtually no current
flows from the battery. The battery path
is unidirectional for safety, so current
cannot flow backwards into the battery. The
battery is not rated for charging, and could
burst if that happened.

If is possible for the dual diode to fail. It could
be bad from the factory, or it could have been damaged
at some point after that. That can lead to premature
battery failure, or flaky CMOS/RTC. Note that when an
Asus motherboard detects "overclocking failure", it
can reset some settings - don't interpret every
occurrence of a settings change, as precipitated by
a CMOS problem. Sometimes, the BIOS is just resetting
to defaults, after the processor failed to finish POST.
If, on the other hand, the RTC clock value is reset,
then I might suspect a VBAT problem.

It sounds like more than that part of the circuit
is currently compromised, since the Knoppix boot
sequence is partially completing. It could be
that some other regulated motherboard voltage
(like the ones feeding the Northbridge or Southbridge)
might be out of sorts. But my previous attempts to
figure out how it works were unsuccessful. I know
the P4C800-E makes extensive usage of linear regulators
and op amps, for those voltages. The regulators
should be series pass type, but there seemed to be
more MOSFETs on the board than I could account for,
and I was getting all sorts of weird voltages from
the pins of some of those guys. So I never succeeded
in forming a mental picture of the "power train" on
my board.

Now, passing Memtest tells you something as well. Memtest
primarily focuses on the Northbridge and RAM. There isn't
much reason to go down the hub bus to the Southbridge. So
that narrows it down to Southbridge (or more "south" of that).
It is a wonder it finished booting from the memtest floppy
or CD, because that stage would be dependent on a working
Southbridge and/or SuperIO.

For a reference schematic of an 875P chipset board, you
can download this. This is one of the few documents I have
to go on, to analyze motherboard design. I wish Intel
continued the practice of releasing stuff like this, because
it is helpful in understanding board problems.

http://www.intel.com/design/chipsets...ics/252812.htm

The only problem with that one is, the Asus approach to building
boards is different than the Intel one. Also, the Intel schematic
lacks "real" parts in some sections of the schematic, like say
a real SuperI/O implementation. But I'll take anything I can
get, when searching for info.

Paul

On Tue, 13 Jan 2009 20:59:52 -0500, Paul wrote:

_DD wrote:
On Tue, 13 Jan 2009 17:10:32 -0500, Paul wrote:

You can change the battery, but I doubt it would help.


You had recommended Knoppix for testing systems with boot problems. I
DL'd their latest image. It came up partway but locked before
completing the boot process. Tried many times. I'm not sure if that's
a compatibility issue, or if it was just turning up the same boot-lock
bug.

Memtest86 and Seatools say the RAM and Drives are good.

I'd love to have some kind of sophisticated hardware debugger for
this. It's been like taking shots in the dark.


Speedfan from almico.com gives you the ability to measure
temps and voltages. I think my Northwood on the P4C800-E
ran around 43C or so. It was a 2.8C overclocked slightly
(10% or so).

That puts things into perspective. There was quite a bit of dust in
the fan path, and in front of the heat sink, so we'll see if that
helps.

This thing is really ****ing me off though. I wish that whatever is
flaking would just fail rather than working intermittently. It's been
impossible to get a handle on it, as it's been so inconsisten.

It just failed to boot again, with two beeps (error code for parity?
that's odd).

Next time, it went through most of POST with no error code, but tehn
stalled with a message "Reboot and select proper boot device". I
checked the BIOS settings, and boot priority looked normal.

Extreme delays in both boot cycles, blank screen before the POST
screen, then the POST screen hung on for a long while before
proceding.

Then of course, the third try, everything was back to normal speed and
it booted to XP as normal. No error messages in the event cue. And
when it does boot to XP, it tends to stay running.

If you boot from Knoppix, you try to note down which line
the thing is failing at. And not all the lines on the
screen are explicit enough to give good feedback.

I'll try that again. And thanks for your comments on the diode feeds
from the battery, etc. I do follow that.

If that diode module was failing, I could see some intermittent
operation. Not sure if it would result in something like the current
symptoms. And also, that module should be so simple that I'd think it
would either fail or not, rather than exhibit flakey intermittents.

I'm tempted to just scrap the motherboard at this point, but I'm not
positive that would fix things. And I'd like to keep the P4 in
service--I do get pretty good performance from the Northwood.

Are there better boards that would make use of the Northwood? I'll
have to buy more ram anyway, as one of the RAM modules failed, so I
don't care about compatible RAM sockets.

_DD wrote:
On Tue, 13 Jan 2009 20:59:52 -0500, Paul wrote:

_DD wrote:
On Tue, 13 Jan 2009 17:10:32 -0500, Paul wrote:

You can change the battery, but I doubt it would help.

You had recommended Knoppix for testing systems with boot problems. I
DL'd their latest image. It came up partway but locked before
completing the boot process. Tried many times. I'm not sure if that's
a compatibility issue, or if it was just turning up the same boot-lock
bug.

Memtest86 and Seatools say the RAM and Drives are good.

I'd love to have some kind of sophisticated hardware debugger for
this. It's been like taking shots in the dark.

Speedfan from almico.com gives you the ability to measure
temps and voltages. I think my Northwood on the P4C800-E
ran around 43C or so. It was a 2.8C overclocked slightly
(10% or so).

That puts things into perspective. There was quite a bit of dust in
the fan path, and in front of the heat sink, so we'll see if that
helps.

This thing is really ****ing me off though. I wish that whatever is
flaking would just fail rather than working intermittently. It's been
impossible to get a handle on it, as it's been so inconsisten.

It just failed to boot again, with two beeps (error code for parity?
that's odd).

Next time, it went through most of POST with no error code, but tehn
stalled with a message "Reboot and select proper boot device". I
checked the BIOS settings, and boot priority looked normal.

Extreme delays in both boot cycles, blank screen before the POST
screen, then the POST screen hung on for a long while before
proceding.

Then of course, the third try, everything was back to normal speed and
it booted to XP as normal. No error messages in the event cue. And
when it does boot to XP, it tends to stay running.

If you boot from Knoppix, you try to note down which line
the thing is failing at. And not all the lines on the
screen are explicit enough to give good feedback.

I'll try that again. And thanks for your comments on the diode feeds
from the battery, etc. I do follow that.

If that diode module was failing, I could see some intermittent
operation. Not sure if it would result in something like the current
symptoms. And also, that module should be so simple that I'd think it
would either fail or not, rather than exhibit flakey intermittents.

I'm tempted to just scrap the motherboard at this point, but I'm not
positive that would fix things. And I'd like to keep the P4 in
service--I do get pretty good performance from the Northwood.

Are there better boards that would make use of the Northwood? I'll
have to buy more ram anyway, as one of the RAM modules failed, so I
don't care about compatible RAM sockets.

It is odd that eventually it'll start.

Part of any testing I try, is to try to gather more symptoms, and get
some idea as to which part of the thing is most likely to be the cause
of the problem. Many of those will result in a "scrap motherboard"
result, except if there is something that is easy to fix.

For example, on those boards, a Southbridge failure has a higher than
normal probability, but then you'd see all the USB ports disappear at
the same point in time, or you'd see a burn mark on the top of the
Southbridge. Once the burn mark is there, the machine won't POST
after that. That is the latchup failure problem for ICH5/ICH5R.

Are you using the Promise controller and if not, is it disabled ?
I keep mine disabled, as that can speed up POST by a few seconds.

When you get some replacement RAM, perhaps you could test the
system a stick at a time, and see if the problem follows a
particular stick of RAM or not. Things like Memtest, cannot
test the "BIOS reserved" areas of RAM, so to test those areas,
you have to try comparing the behavior of single sticks of
RAM, and see if the problem disappears when another stick is
used.

Although my current system has a low end Core2 in it, the
P4 system remains for me a competent platform, and is my
current "backup system" if there are problems with the Core2.
So if you can keep it running, there is no particular reason
to scrap it. Anything which runs about 3GHz in the P4 line,
should make a good general purpose machine.

S478 has been out of production long enough, that you'll
have trouble finding retail boards. Some companies may
sell "pulls", which is a motherboard removed from a
working system. Some of those may come from "end of lease"
systems. Ebay would be another source.

The vip.asus.com forums are a place you can find problem reports
for various models of motherboards, so if you were looking for
another Asus board, you could check there to see what kind of
problems crop up.

http://vip.asus.com/forum/topic.aspx...Language=en-us

For a list of S478 boards, you can use this.

http://www.asus.com/products.aspx?l1=3&l2=-1

Or even traverse the ftp site:

ftp://ftp.asus.com/pub/ASUS/mb/sock478/

Paul

_DD wrote:
On Tue, 13 Jan 2009 20:59:52 -0500, Paul wrote:

_DD wrote:
On Tue, 13 Jan 2009 17:10:32 -0500, Paul wrote:

You can change the battery, but I doubt it would help.

You had recommended Knoppix for testing systems with boot problems. I
DL'd their latest image. It came up partway but locked before
completing the boot process. Tried many times. I'm not sure if that's
a compatibility issue, or if it was just turning up the same boot-lock
bug.

Memtest86 and Seatools say the RAM and Drives are good.

I'd love to have some kind of sophisticated hardware debugger for
this. It's been like taking shots in the dark.

Speedfan from almico.com gives you the ability to measure
temps and voltages. I think my Northwood on the P4C800-E
ran around 43C or so. It was a 2.8C overclocked slightly
(10% or so).

That puts things into perspective. There was quite a bit of dust in
the fan path, and in front of the heat sink, so we'll see if that
helps.

This thing is really ****ing me off though. I wish that whatever is
flaking would just fail rather than working intermittently. It's been
impossible to get a handle on it, as it's been so inconsisten.

It just failed to boot again, with two beeps (error code for parity?
that's odd).

Next time, it went through most of POST with no error code, but tehn
stalled with a message "Reboot and select proper boot device". I
checked the BIOS settings, and boot priority looked normal.

Extreme delays in both boot cycles, blank screen before the POST
screen, then the POST screen hung on for a long while before
proceding.

Then of course, the third try, everything was back to normal speed and
it booted to XP as normal. No error messages in the event cue. And
when it does boot to XP, it tends to stay running.

If you boot from Knoppix, you try to note down which line
the thing is failing at. And not all the lines on the
screen are explicit enough to give good feedback.

I'll try that again. And thanks for your comments on the diode feeds
from the battery, etc. I do follow that.

If that diode module was failing, I could see some intermittent
operation. Not sure if it would result in something like the current
symptoms. And also, that module should be so simple that I'd think it
would either fail or not, rather than exhibit flakey intermittents.

I'm tempted to just scrap the motherboard at this point, but I'm not
positive that would fix things. And I'd like to keep the P4 in
service--I do get pretty good performance from the Northwood.

Are there better boards that would make use of the Northwood? I'll
have to buy more ram anyway, as one of the RAM modules failed, so I
don't care about compatible RAM sockets.

It is odd that eventually it'll start.

Part of any testing I try, is to try to gather more symptoms, and get
some idea as to which part of the thing is most likely to be the cause
of the problem. Many of those will result in a "scrap motherboard"
result, except if there is something that is easy to fix.

For example, on those boards, a Southbridge failure has a higher than
normal probability, but then you'd see all the USB ports disappear at
the same point in time, or you'd see a burn mark on the top of the
Southbridge. Once the burn mark is there, the machine won't POST
after that. That is the latchup failure problem for ICH5/ICH5R.

Are you using the Promise controller and if not, is it disabled ?
I keep mine disabled, as that can speed up POST by a few seconds.

When you get some replacement RAM, perhaps you could test the
system a stick at a time, and see if the problem follows a
particular stick of RAM or not. Things like Memtest, cannot
test the "BIOS reserved" areas of RAM, so to test those areas,
you have to try comparing the behavior of single sticks of
RAM, and see if the problem disappears when another stick is
used.

Although my current system has a low end Core2 in it, the
P4 system remains for me a competent platform, and is my
current "backup system" if there are problems with the Core2.
So if you can keep it running, there is no particular reason
to scrap it. Anything which runs about 3GHz in the P4 line,
should make a good general purpose machine.

S478 has been out of production long enough, that you'll
have trouble finding retail boards. Some companies may
sell "pulls", which is a motherboard removed from a
working system. Some of those may come from "end of lease"
systems. Ebay would be another source.

The vip.asus.com forums are a place you can find problem reports
for various models of motherboards, so if you were looking for
another Asus board, you could check there to see what kind of
problems crop up.

http://vip.asus.com/forum/topic.aspx...Language=en-us

For a list of S478 boards, you can use this.

http://www.asus.com/products.aspx?l1=3&l2=-1

Or even traverse the ftp site:

ftp://ftp.asus.com/pub/ASUS/mb/sock478/

Paul