P4G8X DLX beeping

  • Thread starter Thread starter Joachim Klein
  • Start date Start date
J

Joachim Klein

Hi folks,

ferengi rule of acquisition: if something seems to nice to be true it often
is so...
Well - I got a P4G8X DLX for very, very decent 40 bucks. After installing
CPU, AGP-card and so on I tried to start the machine but all I got was an
endless sequence of long beeps with a 2 second pause.
Any idea what this means ?
CPU defect ? Memory defect ?

Thanx a billion in advance for your input !

Cya -

Joachim
 
pull up the manual on the asus website and look up the error beep code

See section 2.5.2 of the manual. P4G8X uses unbuffered memory, and
if using one stick of memory, the manual says to use slot A2.

The Northbridge datasheet is here:
http://www.intel.com/design/chipsets/e7205/datashts/25193702.pdf

It says:

"Memory Operation
The MCH contains a dual channel DDR interface, with each channel
having 64 data bits and 8 ECC bits. The memory interface operates
in ³lock-step² with each other. The data is a double QWord
interleaved between the channels with the low DQWord on channel A
and the high DQWord on channel B. A burst of four data items, that
take two clocks, is required for one cache line (64 bytes). A
256-bit interface transfers the data at the core clock frequency
internally, matching the memory bandwidth. The memory must be
populated in identical DIMM configurations (i.e., Slot 0 of channel
A must contain the same configuration DIMM as Slot 0 of channel B).
The configuration consists of the same number of rows (1 or 2), the
same technology part (128 Mb, 256 Mb, 512 Mb), the same DRAM chip
width (x8, or x16), and the same speed."

so if using more than one stick, you would be advised to match the
memory on each channel. It is hard to say, but the controller on
this board may not be as flexible as an 865/875 board, which has
what is called virtual single channel mode, for use with non
matching memory DIMMs.

HTH,
Paul
 
Thanx a billion - but the strange thing is that my memory (2x 256 MB Crucial
PC-2100 CL 2.5) already ran flawlessly on my 1st P4G8X DLX - even in dual
mode and @ 166 MHz clock with manual CL2 timing. I replaced them by Kingston
PC-3200 when I upgraded from a P4 1600 (FSB 400) to a P4 2800 (FSB 800) -
the P4G8X handles FSB speeds up to 189 MHz rock stable, so the loss was not
too big.
But in the "new" P4G8X - nada. Beep---Beep---Beep---Beep..... :-((((


 
"Joachim Klein" said:
Thanx a billion - but the strange thing is that my memory (2x 256 MB Crucial
PC-2100 CL 2.5) already ran flawlessly on my 1st P4G8X DLX - even in dual
mode and @ 166 MHz clock with manual CL2 timing. I replaced them by Kingston
PC-3200 when I upgraded from a P4 1600 (FSB 400) to a P4 2800 (FSB 800) -
the P4G8X handles FSB speeds up to 189 MHz rock stable, so the loss was not
too big.
But in the "new" P4G8X - nada. Beep---Beep---Beep---Beep..... :-((((

Do you still have the Crucial RAM around ? Have you tried the new
board with that ? Does the new board use the same revision of BIOS as
the old board ? Maybe if you can get it to start, you can adjust the
timing on the ram, shut down, and then try the new ram, using manual
timings.

Paul
 
Sorry -seems my answer was a bit confusing.
In fact I have *two* systems with a P4G8X DLX now. The 1st one is in my
actual business machine. It formerly ran with the Crucial RAM and the P4
1600. These parts were replaced as I described it. The later (not
necessarily newer) P4G8X DLX replaces the aged TUSL2-C with a P3 Coppermine
750 in my download machine / MP3-server. Same type of board, same components
( P41600, Crucial RAM) I already used successfully - no go ! :-(

Paul said:
"Joachim Klein" said:
Thanx a billion - but the strange thing is that my memory (2x 256 MB Crucial
PC-2100 CL 2.5) already ran flawlessly on my 1st P4G8X DLX - even in dual
mode and @ 166 MHz clock with manual CL2 timing. I replaced them by Kingston
PC-3200 when I upgraded from a P4 1600 (FSB 400) to a P4 2800 (FSB 800) -
the P4G8X handles FSB speeds up to 189 MHz rock stable, so the loss was not
too big.
But in the "new" P4G8X - nada. Beep---Beep---Beep---Beep..... :-((((

Do you still have the Crucial RAM around ? Have you tried the new
board with that ? Does the new board use the same revision of BIOS as
the old board ? Maybe if you can get it to start, you can adjust the
timing on the ram, shut down, and then try the new ram, using manual
timings.

Paul
 
"John McClane" said:
Sorry -seems my answer was a bit confusing.
In fact I have *two* systems with a P4G8X DLX now. The 1st one is in my
actual business machine. It formerly ran with the Crucial RAM and the P4
1600. These parts were replaced as I described it. The later (not
necessarily newer) P4G8X DLX replaces the aged TUSL2-C with a P3 Coppermine
750 in my download machine / MP3-server. Same type of board, same components
( P41600, Crucial RAM) I already used successfully - no go ! :-(

Is the board in good enough shape to RMA ?

Maybe something you haven't tried is the Voice POST feature.
If you connect amplified speakers or your stereo line level
input to the motherboard lime colored lineout connector, there can be
Voice POST messages that play during POST. If an error is found,
there will be a message, and due to the heavy compression, it
pays to have the manual open to the list of messages. You have
to listen to the message and then match it to the text in the
manual, to figure out what the problem is.

On the off chance it is a dead memory bus or a dead socket, have
you tried a single DIMM in all available sockets, one socket at
a time ?

Finally, if you are bored, do the "cardboard test". Pull the
motherboard from the case and rest it on an insulator, like
cardboard. Start with the board stripped, and plug the power
supply to the board (both the 20pin ATX and the 2x2pin 12V
connector). The Voice POST should tell you the CPU is missing.
Install the CPU, and the Voice POST should tell you there is
no memory. And so on. (In fact, on the build I just finished,
I started with the cardboard test, and then I know the system
will work in the case when it is assembled.)

The advantage of assembly outside the case, is there is no chance
of an extra standoff shorting to some tracks on the bottom of
the board. And, you don't even need any other control wires
connected to the panel header, to make it work. Simply
identify the two pins where the power switch would normally
be connected. Drain the static off your screwdriver tip, then
touch the tip momentarily to the two pins where the power
switch would normally go. That will be enough to get the board
to switch on the power supply. A second touch can shut the board
down again, assuming it isn't in a state where it is looking for
a four second long contact (like after it has booted an OS).

Another diagnostic tool is the "POST card". This can be a combo
PCI/ISA card (card has two connectors, one fits ISA slots, the
other PCI slots). It has two hexidecimal LED displays, and as
the BIOS runs, the BIOS writes numbers into the two digit display.
Using lists of error code/routine names available on the web,
you can see what the BIOS is doing, and especially where it is
halted. Sometimes this can give a hint as to the problem as
well. POST cards retail for maybe $100, and regular jobbers
sell them on Ebay for much less, like less than $29.95.

HTH,
Paul
Paul said:
"Joachim Klein" said:
Thanx a billion - but the strange thing is that my memory (2x 256 MB Crucial
PC-2100 CL 2.5) already ran flawlessly on my 1st P4G8X DLX - even in dual
mode and @ 166 MHz clock with manual CL2 timing. I replaced them by Kingston
PC-3200 when I upgraded from a P4 1600 (FSB 400) to a P4 2800 (FSB 800) -
the P4G8X handles FSB speeds up to 189 MHz rock stable, so the loss was not
too big.
But in the "new" P4G8X - nada. Beep---Beep---Beep---Beep..... :-((((

Do you still have the Crucial RAM around ? Have you tried the new
board with that ? Does the new board use the same revision of BIOS as
the old board ? Maybe if you can get it to start, you can adjust the
timing on the ram, shut down, and then try the new ram, using manual
timings.

Paul
 
Woooow. Extremely detailed answer with many precious hints. I´m honored :-)
Well -I already tried some of the procedures you recommended, but not in
such a systematical way. I think I´ll pay extra attention to the voice post
option - never used this up to now. With the TUSL2-C I ran the I-panel and
its port 80 features, which is not as bad as its reputation (IMHO).
But the P4G8X does not support it...

Nevertheless thanx a trillion for spending so much precious time to help me
! :-)
Paul said:
"John McClane" said:
Sorry -seems my answer was a bit confusing.
In fact I have *two* systems with a P4G8X DLX now. The 1st one is in my
actual business machine. It formerly ran with the Crucial RAM and the P4
1600. These parts were replaced as I described it. The later (not
necessarily newer) P4G8X DLX replaces the aged TUSL2-C with a P3 Coppermine
750 in my download machine / MP3-server. Same type of board, same components
( P41600, Crucial RAM) I already used successfully - no go ! :-(

Is the board in good enough shape to RMA ?

Maybe something you haven't tried is the Voice POST feature.
If you connect amplified speakers or your stereo line level
input to the motherboard lime colored lineout connector, there can be
Voice POST messages that play during POST. If an error is found,
there will be a message, and due to the heavy compression, it
pays to have the manual open to the list of messages. You have
to listen to the message and then match it to the text in the
manual, to figure out what the problem is.

On the off chance it is a dead memory bus or a dead socket, have
you tried a single DIMM in all available sockets, one socket at
a time ?

Finally, if you are bored, do the "cardboard test". Pull the
motherboard from the case and rest it on an insulator, like
cardboard. Start with the board stripped, and plug the power
supply to the board (both the 20pin ATX and the 2x2pin 12V
connector). The Voice POST should tell you the CPU is missing.
Install the CPU, and the Voice POST should tell you there is
no memory. And so on. (In fact, on the build I just finished,
I started with the cardboard test, and then I know the system
will work in the case when it is assembled.)

The advantage of assembly outside the case, is there is no chance
of an extra standoff shorting to some tracks on the bottom of
the board. And, you don't even need any other control wires
connected to the panel header, to make it work. Simply
identify the two pins where the power switch would normally
be connected. Drain the static off your screwdriver tip, then
touch the tip momentarily to the two pins where the power
switch would normally go. That will be enough to get the board
to switch on the power supply. A second touch can shut the board
down again, assuming it isn't in a state where it is looking for
a four second long contact (like after it has booted an OS).

Another diagnostic tool is the "POST card". This can be a combo
PCI/ISA card (card has two connectors, one fits ISA slots, the
other PCI slots). It has two hexidecimal LED displays, and as
the BIOS runs, the BIOS writes numbers into the two digit display.
Using lists of error code/routine names available on the web,
you can see what the BIOS is doing, and especially where it is
halted. Sometimes this can give a hint as to the problem as
well. POST cards retail for maybe $100, and regular jobbers
sell them on Ebay for much less, like less than $29.95.

HTH,
Paul
 
Back
Top