P4R800-V Deluxe and Crucial DDR400 RAM


K

Ken

I have an Asus P4R800-V Deluxe mobo with 512MB Crucial DDR400 (model
CT6464Z40B) RAM. I use the on-board video, sound, and lan. The
processor is a Pentium 4 2.4A GHz 533MHz FSB, Prescott Core, 1MB L2
Cache.

More than 50% of the time on boot-up the system will freeze at the
BIOS logo screen. If I press TAB to get out of the screen it will show
that it has froze right after the POST memory test. I think that there
is a problem with me RAM.

In the BIOS it shows that it gave my RAM these settings:
CAS: 3
TRCD: 4
TRP: 4
TRAS: 8

Do those settings look correct?

Does anyone have any other ideas as to what the problem could be?

Thanks!,
Ken
 
Ad

Advertisements

P

Paul

I have an Asus P4R800-V Deluxe mobo with 512MB Crucial DDR400 (model
CT6464Z40B) RAM. I use the on-board video, sound, and lan. The
processor is a Pentium 4 2.4A GHz 533MHz FSB, Prescott Core, 1MB L2
Cache.

More than 50% of the time on boot-up the system will freeze at the
BIOS logo screen. If I press TAB to get out of the screen it will show
that it has froze right after the POST memory test. I think that there
is a problem with me RAM.

In the BIOS it shows that it gave my RAM these settings:
CAS: 3
TRCD: 4
TRP: 4
TRAS: 8

Do those settings look correct?

Does anyone have any other ideas as to what the problem could be?

Thanks!,
Ken

The problem could be, that the memory speed and the processor
FSB are mismatched. Now, a properly designed chipset should handle
a wide range of combinations, but it is a tough thing to design for.
The Nforce2, for example, has a bug in it, that is triggered by
running memory and CPU bus at different clock rates, so it is
a possibility.

Your processor is FSB533. This is a basic clock of 133MHz*4,
as four chunks of data are passed per clock cycle. To match this,
two channels of memory, running at 133MHz*2 (DDR is two chunks
of data per clock) is a perfect fit. If you have set the memory
to some "Auto" mode or "By SPD" mode, then the BIOS will pick up
on the fact that the memory is capable of DDR400 and try to set
the memory to that speed. As I said, with well designed chips,
any rate mismatch should work, but it doesn't always turn out
that way.

First of all, you need to get yourself a Windows utility. This
will allow you to independently verify that the BIOS settings
you are using, are actually working. Sometimes, when a BIOS is
first released, the memory control code isn't finished, and
the memory is fed fixed values. Using a Windows program allows
you to snoop on what is really happening -

http://pcextreme.net/reviews/xms3700/cpuz-max3.jpg
http://www.cpuid.com/cpuz

Due to the fact your chipset is relatively new, not all
utilities support the ATI chipset yet. You may have to
"shop around", to find one that is updated. The former
AIDA32 (now Everest and still free) might also work:

http://www.lavalys.com/index.php?page=product&view=1&subpage=3

When the BIOS displays 3-4-4-8, those are "teaser" numbers, and
those particular numbers are the most relaxed numbers possible.
Those numbers are typically used for PC4000 overclocker memory,
to make the memory work.

Your Crucial memory is manufactured by Micron, as Micron owns
Crucial and Crucial is the retail arm. While Crucial might tell
you the memory is CAS3, Micron's datasheets give all the specs.
If you look at the DIMM, you might see both a Crucial and a
Micron sticky label. You can use the Micron label to look up the
datasheet. Here is a sample sheet for memory with -40B in the
part number - this is DDR400 5ns memory, and the datasheet says
it is 3-3-3 memory. With some luck, you might find the CPUZ
display reflecting those numbers, if the BIOS has used DDR400.

http://download.micron.com/pdf/datasheets/modules/ddr/DDA16C32_64_128x64AG.pdf

The 3-3-3 numbers, when converted to time, is 3*5ns, where the 5
nanoseconds comes from inverting the clock rate of 200MHz. If
you set the memory to DDR266, instead of DDR400, you can improve
the numbers on the memory. DDR266 is 133MHz, or 7.5ns. At DDR266,
your memory can be run at 2-2-2. (The CAS number is fractional,
as in 1.5, 2, 2.5, while the other two numbers are integers, and
in this case, the product of the number times the clock period
must be greater than or equal to that 15ns number.)

I know all of this is confusing, so to start, change the memory
clock to DDR266, and see if it settles down. If the motherboard
has a Vdimm adjustment, it should be set to 2.6V for PC3200
memory, as PC3200 uses 2.6V while lesser memories use 2.5V.
Anything up to 2.75 or 2.8 volts or so is reasonable.

As for your choice of processors, I think you should have spent
a few more bucks and avoided the Prescotts as well. Prescotts
are power hogs, and will heat your room up.

A 2.4A 533 1MB Prescott is $124 at Newegg and uses 89Watts

A 2.4B 533 512KB Northwood is $148 at Newegg and uses 59.8Watts.

A 2.4C 800 512KB Northwood is $169 at Newegg and uses 66.2Watts
(supports hyperthreading if you have an OS that can use it -
disabling HT reduces the power closer to the level of the 2.4B)

The 2.4C will nicely match two DDR400 rams and allow you to
run them synchronously and get your money's worth from the ram.
You could try overclocking the other processors, because
especially in the case of the Prescott, there is no reason to
expect it won't do FSB800 (but, of course, the Prescott will
give you a collossal cooling problem). The FSB533 Northwood
might make it to FSB640, but I'm not sure whether a recent
stepping has the ability to make it all the way to FSB800 or
not. (http://www.cpudatabase.com/CPUdb shows FSB710 is possible
with air cooling and reasonable voltages - higher needs exotic
cooling.)

Tell us how it works out. I see the ATI chipset is being used
in a lot of small systems, and for all these companies to use
the chipset, it must be demonstrated to work for them.

BTW - if you get stuck after adjusting memory timing, if you
see any messages about corrupted BIOS and the like, DON'T
listen to the prompts that suggest reflashing the BIOS. The
BIOS is not really corrupted. The Asus overclocking recovery
code doesn't work worth a damn, and all you really need to do
to escape from messed up settings, is the "clear CMOS" procedure
listed in the manual. Always remember to unplug the computer
before doing it, and wait 30 seconds for the ATX power supply
to discharge, before moving the CMOS jumper. Failure to do so,
will damage a tiny dual diode on the motherboard.

HTH,
Paul
 
K

Ken

The problem could be, that the memory speed and the processor
FSB are mismatched. Now, a properly designed chipset should handle
a wide range of combinations, but it is a tough thing to design for.
The Nforce2, for example, has a bug in it, that is triggered by
running memory and CPU bus at different clock rates, so it is
a possibility.

Your processor is FSB533. This is a basic clock of 133MHz*4,
as four chunks of data are passed per clock cycle. To match this,
two channels of memory, running at 133MHz*2 (DDR is two chunks
of data per clock) is a perfect fit. If you have set the memory
to some "Auto" mode or "By SPD" mode, then the BIOS will pick up
on the fact that the memory is capable of DDR400 and try to set
the memory to that speed. As I said, with well designed chips,
any rate mismatch should work, but it doesn't always turn out
that way.

First of all, you need to get yourself a Windows utility. This
will allow you to independently verify that the BIOS settings
you are using, are actually working. Sometimes, when a BIOS is
first released, the memory control code isn't finished, and
the memory is fed fixed values. Using a Windows program allows
you to snoop on what is really happening -

http://pcextreme.net/reviews/xms3700/cpuz-max3.jpg
http://www.cpuid.com/cpuz

Due to the fact your chipset is relatively new, not all
utilities support the ATI chipset yet. You may have to
"shop around", to find one that is updated. The former
AIDA32 (now Everest and still free) might also work:

http://www.lavalys.com/index.php?page=product&view=1&subpage=3

When the BIOS displays 3-4-4-8, those are "teaser" numbers, and
those particular numbers are the most relaxed numbers possible.
Those numbers are typically used for PC4000 overclocker memory,
to make the memory work.

Your Crucial memory is manufactured by Micron, as Micron owns
Crucial and Crucial is the retail arm. While Crucial might tell
you the memory is CAS3, Micron's datasheets give all the specs.
If you look at the DIMM, you might see both a Crucial and a
Micron sticky label. You can use the Micron label to look up the
datasheet. Here is a sample sheet for memory with -40B in the
part number - this is DDR400 5ns memory, and the datasheet says
it is 3-3-3 memory. With some luck, you might find the CPUZ
display reflecting those numbers, if the BIOS has used DDR400.

http://download.micron.com/pdf/datasheets/modules/ddr/DDA16C32_64_128x64AG.pdf

The 3-3-3 numbers, when converted to time, is 3*5ns, where the 5
nanoseconds comes from inverting the clock rate of 200MHz. If
you set the memory to DDR266, instead of DDR400, you can improve
the numbers on the memory. DDR266 is 133MHz, or 7.5ns. At DDR266,
your memory can be run at 2-2-2. (The CAS number is fractional,
as in 1.5, 2, 2.5, while the other two numbers are integers, and
in this case, the product of the number times the clock period
must be greater than or equal to that 15ns number.)

I know all of this is confusing, so to start, change the memory
clock to DDR266, and see if it settles down. If the motherboard
has a Vdimm adjustment, it should be set to 2.6V for PC3200
memory, as PC3200 uses 2.6V while lesser memories use 2.5V.
Anything up to 2.75 or 2.8 volts or so is reasonable.

As for your choice of processors, I think you should have spent
a few more bucks and avoided the Prescotts as well. Prescotts
are power hogs, and will heat your room up.

A 2.4A 533 1MB Prescott is $124 at Newegg and uses 89Watts

A 2.4B 533 512KB Northwood is $148 at Newegg and uses 59.8Watts.

A 2.4C 800 512KB Northwood is $169 at Newegg and uses 66.2Watts
(supports hyperthreading if you have an OS that can use it -
disabling HT reduces the power closer to the level of the 2.4B)

The 2.4C will nicely match two DDR400 rams and allow you to
run them synchronously and get your money's worth from the ram.
You could try overclocking the other processors, because
especially in the case of the Prescott, there is no reason to
expect it won't do FSB800 (but, of course, the Prescott will
give you a collossal cooling problem). The FSB533 Northwood
might make it to FSB640, but I'm not sure whether a recent
stepping has the ability to make it all the way to FSB800 or
not. (http://www.cpudatabase.com/CPUdb shows FSB710 is possible
with air cooling and reasonable voltages - higher needs exotic
cooling.)

Tell us how it works out. I see the ATI chipset is being used
in a lot of small systems, and for all these companies to use
the chipset, it must be demonstrated to work for them.

BTW - if you get stuck after adjusting memory timing, if you
see any messages about corrupted BIOS and the like, DON'T
listen to the prompts that suggest reflashing the BIOS. The
BIOS is not really corrupted. The Asus overclocking recovery
code doesn't work worth a damn, and all you really need to do
to escape from messed up settings, is the "clear CMOS" procedure
listed in the manual. Always remember to unplug the computer
before doing it, and wait 30 seconds for the ATX power supply
to discharge, before moving the CMOS jumper. Failure to do so,
will damage a tiny dual diode on the motherboard.

HTH,
Paul

OK, I flashed the BIOS to the most recent version and set the RAM to
run at 266 in the BIOS. I'm still having the problem though. There
seems to be no rhyme or reason to it. Sometimes it boots fine,
somtimes it stops right after the POST memory test. It's not every
other time, it's just seems random.

There is a rather important thing that I am not sure about though, and
that is the correct setting for the the clear CMOS jumper. In the
manual is shows that pins 2 and 3 are normal. However, it also came
with a mobo diagram that says pins 1 and 2 are normal. I currently
have it set to what the manual says. It actually boots up in either
setting which is strange. I'm afraid I'm going to mess something up if
it's wrong.

BTW, thanks for the info about the RAM settings Paul, if was very
helpful and informative.

Ken
 
P

Paul

OK, I flashed the BIOS to the most recent version and set the RAM to
run at 266 in the BIOS. I'm still having the problem though. There
seems to be no rhyme or reason to it. Sometimes it boots fine,
somtimes it stops right after the POST memory test. It's not every
other time, it's just seems random.

There is a rather important thing that I am not sure about though, and
that is the correct setting for the the clear CMOS jumper. In the
manual is shows that pins 2 and 3 are normal. However, it also came
with a mobo diagram that says pins 1 and 2 are normal. I currently
have it set to what the manual says. It actually boots up in either
setting which is strange. I'm afraid I'm going to mess something up if
it's wrong.

BTW, thanks for the info about the RAM settings Paul, if was very
helpful and informative.

Ken

Have you run memtest86 from memtest.org ? This is a program you
load via a floppy or a CDROM at boot time - the program is a
miniature boot loader and uses no OS. It will test the memory for
you, and help prove whether the memory is OK or not. You can
even use that program, while adjusting memory timings, and then
retesting with memtest86, to see if the setting is OK or not.
The download from the memtest site, contains a floppy formatter
program, and you feed it a blank floppy, and it installs the
test program on the floppy. Since the floppy doesn't contain a
normal file system, you cannot list the contents of the floppy
once it is prepped with memtest. Reformatting the floppy will
make the floppy available again for other uses, and erase
the memtest.

Since you say "stops right after the POST memory test",
it may have completed the memory test, and moved on to
enumerating some other hardware. So, maybe it is getting
stuck on something else. A "POST card', which has two hex
displays on it, can display the routine number currently
being executed by the BIOS, and that can sometimes give an
idea as to what routine the BIOS is hanging in. If the
amount of memory shown on the screen is correct, whether the
machine freezes or continues on, could mean it passed the
built-in BIOS memory test in either case.

As for the jumpering, my first source of info, would be the
info printed in the white paint of the motherboard silk screen.
Sometimes the jumper settings are explained right on the motherboard.
The designer of the board puts that info there, and that person
is the most knowledgeable about the design.

The jumpers are generally in the correct position, when the board
is shipped from the factory. So, that would give you another hint
as to which position is correct.

The pictures of the motherboard, in the manual, show the jumper
in the 2-3 position, implying that is normal. I used the downloadable
Japanese PDF manual, to see a clearer picture of the motherboard.

If you own a voltmeter, you could try the following.
The jumper shorts two pins together, and there is some copper
colored metal inside the plastic jumper. With a voltmeter set
on volts, touch one probe of the voltmeter to the copper
colored metal of the jumper, while the jumper is in place.
In the "normal" position, you should see a voltage of between
2 and 3 volts on that metal. If the jumper is in the "clear
CMOS" position, you'll read zero volts on the jumper metal.
The second lead of your probe can be connected to a screw on the
chassis of the computer, or the body of one of the I/O connectors
on the back of the computer (as that is GND).

1 2 3
X X<-->X
^
|
+---- Touch voltmeter to the jumper metal, or to
either of the two pins being joined by the
jumper. If the voltage reads 3 volts
(a full battery) with that jumper in place,
then that is the correct position. The other
lead of the voltmeter goes to GND.

An indirect test of correctness, is to leave the jumper in the
suspected good position, enter the BIOS and make some custom
settings, then save and exit. Shut down the computer and unplug
it. This removes +5VSB from the board, leaving the CMOS circuit
running off the battery. Plug in the computer the next morning,
allowing a number of hours to pass. If the custom settings
are replaced by the default values, then either the CMOS
battery is drained, or the jumper is in the wrong position.

Depending on the warranty/return policy with your vendor, you
may want to RMA the board, in case the symptoms indicate that
something is close to failing on the board. If you are past the
return period with your vendor, you still have plenty of time
on the Asus warranty, in case it completely dies.

HTH,
Paul
 
K

Ken

I have the same problem when I enable the Quiet Fan option.

Thank you so much Marc!! It was indeed the QFan option that was
screwing things up. Unfortunately, now that I disabled that my
computer is a lot louder. But I feel so much better now that I know
what the problem is :)

- Ken
 
Ad

Advertisements

P

Paul

(e-mail address removed) (Marc-Andr?) wrote in message

Thank you so much Marc!! It was indeed the QFan option that was
screwing things up. Unfortunately, now that I disabled that my
computer is a lot louder. But I feel so much better now that I know
what the problem is :)

- Ken

Another wierd one for the books :) Please send the problem to
Asus Tech support, so they'll fix this in their next BIOS
update. There is a support web page, where you can fill in
your details, and what you did to solve it. Maybe in a month
or three, they'll have a fix.

Paul
 

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