K8N and Seagate Sata Drive

Just got my motherboard back from RMA and picked up a seagate 320 Gig HD to
reinstall XP and I`m having major problems. Mainly, the drive is recognized
by the MB, both in the bios and on the boot up splash screen but hangs on
installing XP when Windows tries to go to the setup/install phase.I`ve tried
using the drivers provided on the nforce 3 MB disk but with the same result.
Any advice would be appreciated. Thanks
Jeff

JKBK wrote:
> Just got my motherboard back from RMA and picked up a seagate 320 Gig HD to
> reinstall XP and I`m having major problems. Mainly, the drive is recognized
> by the MB, both in the bios and on the boot up splash screen but hangs on
> installing XP when Windows tries to go to the setup/install phase.I`ve tried
> using the drivers provided on the nforce 3 MB disk but with the same result.
> Any advice would be appreciated. Thanks
> Jeff
>
>

Are you using a WinXP SP1 or SP2 install CD ? Or is your WinXP so old that
it has no Service Pack integrated ?

If you have an old WinXP disk, you should slipstream in at least the SP1
service pack, or preferably the SP2 service pack. As far as I know, MS
is only supporting SP2 now, which is why you'd want to be using it. When
I did this with my Win2K install disk eons ago, I think I used Autostreamer.
You'll need a CD burner, to burn the resulting ISO9660 file created by
the slipstreaming process, to a new blank CD.

http://en.wikipedia.org/wiki/Slipstream_%28computing%29

The purpose of SP1 or SP2, is to handle large disks, or to access any
additional native disk drivers from Microsoft. You also get a USB2 driver,
which is handy.

Paul

Tried with both xp home and pro both with service pack 2. No luck. I have a
160 Gig IDE drive that I was going to put on for storage with the Sata 320
but I might put the IDE as my boot drive now to try and solve the problem.

"Paul" <> wrote in message news:fbqr0d$hvk$...
> JKBK wrote:
>> Just got my motherboard back from RMA and picked up a seagate 320 Gig HD
>> to reinstall XP and I`m having major problems. Mainly, the drive is
>> recognized by the MB, both in the bios and on the boot up splash screen
>> but hangs on installing XP when Windows tries to go to the setup/install
>> phase.I`ve tried using the drivers provided on the nforce 3 MB disk but
>> with the same result.
>> Any advice would be appreciated. Thanks
>> Jeff
>
> Are you using a WinXP SP1 or SP2 install CD ? Or is your WinXP so old that
> it has no Service Pack integrated ?
>
> If you have an old WinXP disk, you should slipstream in at least the SP1
> service pack, or preferably the SP2 service pack. As far as I know, MS
> is only supporting SP2 now, which is why you'd want to be using it. When
> I did this with my Win2K install disk eons ago, I think I used
> Autostreamer.
> You'll need a CD burner, to burn the resulting ISO9660 file created by
> the slipstreaming process, to a new blank CD.
>
> http://en.wikipedia.org/wiki/Slipstream_%28computing%29
>
> The purpose of SP1 or SP2, is to handle large disks, or to access any
> additional native disk drivers from Microsoft. You also get a USB2 driver,
> which is handy.
>
> Paul

Narrowed down the problem to Ram but still trying to work around. The board
will boot, install XP and run fine with a no name stick of PC2700 256Mb but
will hang at various points with any of the other 3 sticks of PC3200 512Mb I
have. Ive tried raising the voltage and loosing the ram timings but no joy.
One of these sticks worked with its predecesor before it was RMA`ed so I`m
at a loss for an explanation. I`ll be checking the approved list for new ram
but if anyone can suggest a workaround so I can use my existing Ram I would
appreciate it before I sink any more cash into this white elephant of a
motherboard.
Jeff
"JKBK" <> wrote in message
news:RV3Ei.134704$fJ5.45157@pd7urf1no...
> Just got my motherboard back from RMA and picked up a seagate 320 Gig HD
> to reinstall XP and I`m having major problems. Mainly, the drive is
> recognized by the MB, both in the bios and on the boot up splash screen
> but hangs on installing XP when Windows tries to go to the setup/install
> phase.I`ve tried using the drivers provided on the nforce 3 MB disk but
> with the same result.
> Any advice would be appreciated. Thanks
> Jeff
>

Jeff wrote:
> Narrowed down the problem to Ram but still trying to work around. The board
> will boot, install XP and run fine with a no name stick of PC2700 256Mb but
> will hang at various points with any of the other 3 sticks of PC3200 512Mb I
> have. Ive tried raising the voltage and loosing the ram timings but no joy.
> One of these sticks worked with its predecesor before it was RMA`ed so I`m
> at a loss for an explanation. I`ll be checking the approved list for new ram
> but if anyone can suggest a workaround so I can use my existing Ram I would
> appreciate it before I sink any more cash into this white elephant of a
> motherboard.
> Jeff

K8N is S754. The memory controller is in the processor, so the behavior
with respect to RAM, should follow the processor around. (In other words,
if a mobo+S754_proc behave crappy with regard to memory, moving the
processor to another board will likely not help. That is because the
memory controller interface is inside the processor.)

S754 is a single channel memory design. But there is a twist. The
data bus is shared by all sticks, so that part of the design is normal.
The processor has two address busses, and the addresses are pumped out of
phase. Only the DIMM being addressed at the time, needs the correct address,
and the DIMMs that aren't being accessed, get the 1's complement of
the address.

By doing that, the electrical switching noise is balanced. If a bit
rises on one address bus, the corresponding bit falls on the other
address bus. The purpose, is to try to reduce switching noise on the
processor die.

Now, how are the address busses wired ? The motherboard has three memory
slots. One slot gets its own address bus. The second address bus, goes to
the two remaining slots.

When populating memory, first you can occupy the slot with the private
bus. Next, you put the second stick on one of the second bus segment's
slots. Installing a third stick, tends to be the deal breaker, as then
there are two loads on that address bus. So the best configs would be
one or two sticks, and in the two stick case, I might try slot 1 and slot 3.

If you examine your motherboard manual, Table 1 and row 18 of the table,
shows the best configuration for two sticks of memory.

Notice how row 17 of the table (where slot 2 and slot 3 are occupied),
causes a big drop in the expected max speed possible. That is the bus
loading issue I was referring to. Double sided has twice the load of
single sided memory. And most people will buy double sided RAM, as it
is the most dense per stick.

You can try dropping the memory frequency, and/or change the "Command Rate"
to 2T. You can leave the timing alone, until you've tried those. I
think you'll find Command Rate to 2T is enough to fix it, as this is
primarily supposed to be an address loading issue. (On your board,
the setting in question is called "1T/2T timing". Change to 2T.
The auto setting should have worked, but maybe you can verify
what value was used while in Windows.)

You can try using CPUZ (www.cpuid.com) to verify settings. At least
in the picture shown here, it appears CPUZ will show the Command Rate
value. But there are no guarantees, as on some chipsets, the Command
Rate register may not be exposed for examination by CPUZ.

http://www.lostcircuits.com/motherbo...mvp/cpuz1t.gif

Paul

> "JKBK" <> wrote in message
> news:RV3Ei.134704$fJ5.45157@pd7urf1no...
>> Just got my motherboard back from RMA and picked up a seagate 320 Gig HD
>> to reinstall XP and I`m having major problems. Mainly, the drive is
>> recognized by the MB, both in the bios and on the boot up splash screen
>> but hangs on installing XP when Windows tries to go to the setup/install
>> phase.I`ve tried using the drivers provided on the nforce 3 MB disk but
>> with the same result.
>> Any advice would be appreciated. Thanks
>> Jeff
>>
>
>

"Paul" <> wrote

> Jeff wrote:
>> Narrowed down the problem to Ram but still trying to work around. The
>> board will boot, install XP and run fine with a no name stick of PC2700
>> 256Mb but will hang at various points with any of the other 3 sticks of
>> PC3200 512Mb I have. Ive tried raising the voltage and loosing the ram
>> timings but no joy. One of these sticks worked with its predecesor before
>> it was RMA`ed so I`m at a loss for an explanation. I`ll be checking the
>> approved list for new ram but if anyone can suggest a workaround so I can
>> use my existing Ram I would appreciate it before I sink any more cash
>> into this white elephant of a motherboard.
>> Jeff
>
> K8N is S754. The memory controller is in the processor, so the behavior
> with respect to RAM, should follow the processor around. (In other words,
> if a mobo+S754_proc behave crappy with regard to memory, moving the
> processor to another board will likely not help. That is because the
> memory controller interface is inside the processor.)
>

That's the convential wisdom, my experience (and that of quite a few posters
in MB forums and newsgroups when nf3 MBs first came out) suggests otherwise
and leads to the conclusion that the nf3 was a crap chipset that on far too
many motherboards couldn't handle double sided DDR400 RAM at its rated
speed. I've tried a variety of DS 512MB memory sticks on several nf3 boards
and a number of CPUs - none of them would pass memtest at 200. Yet the same
CPUs and memory on later S754 nf4x boards would work reliably over 200 to
the limits of the RAM. As an example, had some Kingmax tiny BGA RAM (can't
remember the name) that would work reliably at 250 MHz (1 or 2 sticks) on an
nf4x but still failed at 200 on nf3s.

> S754 is a single channel memory design. But there is a twist. The
> data bus is shared by all sticks, so that part of the design is normal.
> The processor has two address busses, and the addresses are pumped out of
> phase. Only the DIMM being addressed at the time, needs the correct
> address,
> and the DIMMs that aren't being accessed, get the 1's complement of
> the address.
>
> By doing that, the electrical switching noise is balanced. If a bit
> rises on one address bus, the corresponding bit falls on the other
> address bus. The purpose, is to try to reduce switching noise on the
> processor die.
>
> Now, how are the address busses wired ? The motherboard has three memory
> slots. One slot gets its own address bus. The second address bus, goes to
> the two remaining slots.
>
> When populating memory, first you can occupy the slot with the private
> bus. Next, you put the second stick on one of the second bus segment's
> slots. Installing a third stick, tends to be the deal breaker, as then
> there are two loads on that address bus. So the best configs would be
> one or two sticks, and in the two stick case, I might try slot 1 and slot
> 3.
>
> If you examine your motherboard manual, Table 1 and row 18 of the table,
> shows the best configuration for two sticks of memory.
>
> Notice how row 17 of the table (where slot 2 and slot 3 are occupied),
> causes a big drop in the expected max speed possible. That is the bus
> loading issue I was referring to. Double sided has twice the load of
> single sided memory. And most people will buy double sided RAM, as it
> is the most dense per stick.
>

The only way I could get DDR400 to work on my (MSI) nf3 was 1x512MB DS
combined with 2x256 SS. That configuration was stable up to around 225MHz -
roughly the limits of the Corsair (Winbond) RAM I was using at the time.

> You can try dropping the memory frequency, and/or change the "Command
> Rate"
> to 2T. You can leave the timing alone, until you've tried those. I
> think you'll find Command Rate to 2T is enough to fix it, as this is
> primarily supposed to be an address loading issue. (On your board,
> the setting in question is called "1T/2T timing". Change to 2T.
> The auto setting should have worked, but maybe you can verify
> what value was used while in Windows.)

2T instead of 1T didn't do anything for me. What is likely to work is single
sided 512MB - I've tested 2 sticks, don't know about 3. Or drop the memory
speed to 166. And if you overclock you start getting memory errors if the
memory speed rises into the 190-95 range.

>
> You can try using CPUZ (www.cpuid.com) to verify settings. At least
> in the picture shown here, it appears CPUZ will show the Command Rate
> value. But there are no guarantees, as on some chipsets, the Command
> Rate register may not be exposed for examination by CPUZ.
>
> http://www.lostcircuits.com/motherbo...mvp/cpuz1t.gif
>
> Paul

New bios (K8n 1011) and no more problems.Even using 2 different
manufacturers PC3200 512M single side.I would have thought that Asus (any
manufacturer really) would update their MB before shipping it as a
replacement (it had 2004 bios. Original I assume). Thanks for the advice
"ned ludd" <> wrote in message
news:0WwFi.34090$...
>
> "Paul" <> wrote
>
>> Jeff wrote:
>>> Narrowed down the problem to Ram but still trying to work around. The
>>> board will boot, install XP and run fine with a no name stick of PC2700
>>> 256Mb but will hang at various points with any of the other 3 sticks of
>>> PC3200 512Mb I have. Ive tried raising the voltage and loosing the ram
>>> timings but no joy. One of these sticks worked with its predecesor
>>> before it was RMA`ed so I`m at a loss for an explanation. I`ll be
>>> checking the approved list for new ram but if anyone can suggest a
>>> workaround so I can use my existing Ram I would appreciate it before I
>>> sink any more cash into this white elephant of a motherboard.
>>> Jeff
>>
>> K8N is S754. The memory controller is in the processor, so the behavior
>> with respect to RAM, should follow the processor around. (In other words,
>> if a mobo+S754_proc behave crappy with regard to memory, moving the
>> processor to another board will likely not help. That is because the
>> memory controller interface is inside the processor.)
>>
>
> That's the convential wisdom, my experience (and that of quite a few
> posters in MB forums and newsgroups when nf3 MBs first came out) suggests
> otherwise and leads to the conclusion that the nf3 was a crap chipset that
> on far too many motherboards couldn't handle double sided DDR400 RAM at
> its rated speed. I've tried a variety of DS 512MB memory sticks on several
> nf3 boards and a number of CPUs - none of them would pass memtest at 200.
> Yet the same CPUs and memory on later S754 nf4x boards would work reliably
> over 200 to the limits of the RAM. As an example, had some Kingmax tiny
> BGA RAM (can't remember the name) that would work reliably at 250 MHz (1
> or 2 sticks) on an nf4x but still failed at 200 on nf3s.
>
>> S754 is a single channel memory design. But there is a twist. The
>> data bus is shared by all sticks, so that part of the design is normal.
>> The processor has two address busses, and the addresses are pumped out of
>> phase. Only the DIMM being addressed at the time, needs the correct
>> address,
>> and the DIMMs that aren't being accessed, get the 1's complement of
>> the address.
>>
>> By doing that, the electrical switching noise is balanced. If a bit
>> rises on one address bus, the corresponding bit falls on the other
>> address bus. The purpose, is to try to reduce switching noise on the
>> processor die.
>>
>> Now, how are the address busses wired ? The motherboard has three memory
>> slots. One slot gets its own address bus. The second address bus, goes to
>> the two remaining slots.
>>
>> When populating memory, first you can occupy the slot with the private
>> bus. Next, you put the second stick on one of the second bus segment's
>> slots. Installing a third stick, tends to be the deal breaker, as then
>> there are two loads on that address bus. So the best configs would be
>> one or two sticks, and in the two stick case, I might try slot 1 and slot
>> 3.
>>
>> If you examine your motherboard manual, Table 1 and row 18 of the table,
>> shows the best configuration for two sticks of memory.
>>
>> Notice how row 17 of the table (where slot 2 and slot 3 are occupied),
>> causes a big drop in the expected max speed possible. That is the bus
>> loading issue I was referring to. Double sided has twice the load of
>> single sided memory. And most people will buy double sided RAM, as it
>> is the most dense per stick.
>>
>
> The only way I could get DDR400 to work on my (MSI) nf3 was 1x512MB DS
> combined with 2x256 SS. That configuration was stable up to around
> 225MHz - roughly the limits of the Corsair (Winbond) RAM I was using at
> the time.
>
>> You can try dropping the memory frequency, and/or change the "Command
>> Rate"
>> to 2T. You can leave the timing alone, until you've tried those. I
>> think you'll find Command Rate to 2T is enough to fix it, as this is
>> primarily supposed to be an address loading issue. (On your board,
>> the setting in question is called "1T/2T timing". Change to 2T.
>> The auto setting should have worked, but maybe you can verify
>> what value was used while in Windows.)
>
> 2T instead of 1T didn't do anything for me. What is likely to work is
> single sided 512MB - I've tested 2 sticks, don't know about 3. Or drop the
> memory speed to 166. And if you overclock you start getting memory errors
> if the memory speed rises into the 190-95 range.
>
>>
>> You can try using CPUZ (www.cpuid.com) to verify settings. At least
>> in the picture shown here, it appears CPUZ will show the Command Rate
>> value. But there are no guarantees, as on some chipsets, the Command
>> Rate register may not be exposed for examination by CPUZ.
>>
>> http://www.lostcircuits.com/motherbo...mvp/cpuz1t.gif
>>
>> Paul
>
>