A7N8X-E "System Failed Due to CPU Overclocking" - what does it really mean?

[Erik Harris]

This seems like a silly question, but can anyone give me some insight into
what this Voice POST error _really_ means (on an A7N8X-E Deluxe,
specifically)? On a CPU without a locked multiplier or locked FSB, how does
the motherboard know it it's being overclocked? What kind of signal is it
queuing off of to return that error?

Generally, overclocking results in some significant instability long before
it reaches the point of not being able to boot up at all. I'm seeing just
the opposite, though. I can't get much of any extra juice out of my Athlon
XP-M 2600+ without getting this voice POST error. When it successfully
boots, it works just fine - no instability, no temperature issues, nothing.
That is no surprise, though, since I'm currently only running at 182x11, no
real overclocking at all (stock clock speed, faster FSB), on a chip that
should easily at 2.3GHz and beyond with no special cooling setup.

My system will sometimes not boot up, and not give any error message (voice
or beep). Sometimes, it'll give me the "System failed due to CPU
overclocking" voice POST error (no beep). Sometimes, it'll load up just
fine. Sometimes it'll give me the voice POST CPU overclock error and load up
fine anyhow, but Asus Voice POST has never been very reliable, so that
doesn't alarm me. I thought I had a bad motherboard, and spent the time and
money to have mine replaced by the vendor. The first three or four times I
booted up, it seemed like I'd solved my problem, but now I seem to be back to
where I started - a system that I need to reboot four or five times to get it
started.

Is anyone else successfully using one of the low-voltage XP-M chips on an
A7N8X-E Deluxe? I haven't found anything to indicate that they're
incompatible, but I'm having a hell of a time, after trying this chip on two
A7N8X-E Deluxe motherboards, I'm beginning to wonder.

I'm at a loss, and am almost tempted to just give up and build a new system
from scratch, without re-using any of my components. It'd be a colossal
waste of money, but I'm feeling like I'd have a better chance with it, after
trying this system with two CPU's, two motherboards, two sets of RAM, two
video cards, and two power supplies - i.e. I've swapped every single element
out at some point in time, and nothing seems to solve the problem. The only
thing I haven't swapped out is the CPU fan. This one is working fine,
though, and doesn't so much as stutter on the failed boot-ups. Just in case,
I'll try swapping that out, or at least making the system think I did by
connecting a chassis fan to the CPU fan connector for a test (I'm getting
really sick of wiping off and re-applying thermal compound). I'm not
optimistic that the fan could be the problem, though.

Any help would be appreciated. Thanks!

--
Erik Harris n$wsr$ader@$harrishom$.com
AIM: KngFuJoe http://www.eharrishome.com
Chinese-Indonesian MA Club http://cimac.eharrishome.com

The above email address is obfuscated to try to prevent SPAM.
Replace each dollar sign with an "e" for the correct address.

 

[As]

This seems like a silly question, but can anyone give me some insight into
what this Voice POST error _really_ means (on an A7N8X-E Deluxe,
specifically)? On a CPU without a locked multiplier or locked FSB, how
does
the motherboard know it it's being overclocked? What kind of signal is it
queuing off of to return that error?

Generally, overclocking results in some significant instability long
before
it reaches the point of not being able to boot up at all. I'm seeing just
the opposite, though. I can't get much of any extra juice out of my
Athlon
XP-M 2600+ without getting this voice POST error. When it successfully
boots, it works just fine - no instability, no temperature issues,
nothing.
That is no surprise, though, since I'm currently only running at 182x11,
no
real overclocking at all (stock clock speed, faster FSB), on a chip that
should easily at 2.3GHz and beyond with no special cooling setup.

My system will sometimes not boot up, and not give any error message
(voice
or beep). Sometimes, it'll give me the "System failed due to CPU
overclocking" voice POST error (no beep). Sometimes, it'll load up just
fine. Sometimes it'll give me the voice POST CPU overclock error and load
up
fine anyhow, but Asus Voice POST has never been very reliable, so that
doesn't alarm me. I thought I had a bad motherboard, and spent the time
and
money to have mine replaced by the vendor. The first three or four times
I
booted up, it seemed like I'd solved my problem, but now I seem to be back
to
where I started - a system that I need to reboot four or five times to get
it
started.

Is anyone else successfully using one of the low-voltage XP-M chips on an
A7N8X-E Deluxe? I haven't found anything to indicate that they're
incompatible, but I'm having a hell of a time, after trying this chip on
two
A7N8X-E Deluxe motherboards, I'm beginning to wonder.

I'm at a loss, and am almost tempted to just give up and build a new
system
from scratch, without re-using any of my components. It'd be a colossal
waste of money, but I'm feeling like I'd have a better chance with it,
after
trying this system with two CPU's, two motherboards, two sets of RAM, two
video cards, and two power supplies - i.e. I've swapped every single
element
out at some point in time, and nothing seems to solve the problem. The
only
thing I haven't swapped out is the CPU fan. This one is working fine,
though, and doesn't so much as stutter on the failed boot-ups. Just in
case,
I'll try swapping that out, or at least making the system think I did by
connecting a chassis fan to the CPU fan connector for a test (I'm getting
really sick of wiping off and re-applying thermal compound). I'm not
optimistic that the fan could be the problem, though.

Any help would be appreciated. Thanks!

--
Erik Harris n$wsr$ader@$harrishom$.com
AIM: KngFuJoe http://www.eharrishome.com
Chinese-Indonesian MA Club http://cimac.eharrishome.com

The above email address is obfuscated to try to prevent SPAM.
Replace each dollar sign with an "e" for the correct address.

Just a thought, check that the AGP bus speed is set to 66 MHz as when I 1st
overclocked it was set to 'AUTO' which meant it was being overclocked as
well and the system was very unstable. This should also lock the PCI bus to
33 MHz.

 

[Erik Harris]

Just a thought, check that the AGP bus speed is set to 66 MHz as when I 1st
overclocked it was set to 'AUTO' which meant it was being overclocked as
well and the system was very unstable. This should also lock the PCI bus to
33 MHz.

Ah ha! Thanks so much, I had forgotten about that. I was thinking that
"auto" stuck it at 66MHz by default, and didn't auto-scale with the bus
speed.

I'll give that a shot, and reply back with status.

--
Erik Harris n$wsr$ader@$harrishom$.com
AIM: KngFuJoe http://www.eharrishome.com
Chinese-Indonesian MA Club http://cimac.eharrishome.com

The above email address is obfuscated to try to prevent SPAM.
Replace each dollar sign with an "e" for the correct address.

 

[Erik Harris]

I'll give that a shot, and reply back with status.

No luck. I had forgotten to do this when I got the new motherboard. I know
I had it locked with the previous system (based on my old Thoroughbred
2400+), but I had forgotten that it wasn't locked in the auto position.

However, after setting the AGP bus to 66MHz, I rebooted the system five
times, and only got two successful POSTs out of it. One of the failures gave
me the "System Failed Due to CPU Overclocking" error, and the other two
failures gave me no beep or voice error at all.

--
Erik Harris n$wsr$ader@$harrishom$.com
AIM: KngFuJoe http://www.eharrishome.com
Chinese-Indonesian MA Club http://cimac.eharrishome.com

The above email address is obfuscated to try to prevent SPAM.
Replace each dollar sign with an "e" for the correct address.

 

[Erik Harris]

video cards, and two power supplies - i.e. I've swapped every single element
out at some point in time, and nothing seems to solve the problem. The only
thing I haven't swapped out is the CPU fan. This one is working fine,
though, and doesn't so much as stutter on the failed boot-ups. Just in case,
I'll try swapping that out, or at least making the system think I did by
connecting a chassis fan to the CPU fan connector for a test (I'm getting
really sick of wiping off and re-applying thermal compound). I'm not
optimistic that the fan could be the problem, though.

Amazing.. I just swapped fan headers - my motherboard now thinks that the
rear chassis fan is the CPU fan. I've booted the machine up five times now,
including once with a higher clock speed, and it has yet to have a problem.
My best guess is that _something_ about the startup motion of my CPU fan
triggers a fault on the motherboard, and it refuses to boot up. I wish I
knew that before I RMA'ed my motherboard with ZipZoomFly.

I'll run it this way for a day or so, and once I'm satisfied that this is the
solution, I'll go get a new HSF combination.

Anyone have any suggestions for a good HSF combo for this A7N8X-E Deluxe
system that's both cool and quiet? I bought this one, the Vantec Aeroflow
TMD fan with heat sink based on the review on Tom's Hardware. On the
cooling/noise graph, it offered the best of both worlds for Cu/Al combo
heatsink setups.

--
Erik Harris n$wsr$ader@$harrishom$.com
AIM: KngFuJoe http://www.eharrishome.com
Chinese-Indonesian MA Club http://cimac.eharrishome.com

The above email address is obfuscated to try to prevent SPAM.
Replace each dollar sign with an "e" for the correct address.

 

[Paul]

This seems like a silly question, but can anyone give me some insight into
what this Voice POST error _really_ means (on an A7N8X-E Deluxe,
specifically)? On a CPU without a locked multiplier or locked FSB, how does
the motherboard know it it's being overclocked? What kind of signal is it
queuing off of to return that error?

Generally, overclocking results in some significant instability long before
it reaches the point of not being able to boot up at all. I'm seeing just
the opposite, though. I can't get much of any extra juice out of my Athlon
XP-M 2600+ without getting this voice POST error. When it successfully
boots, it works just fine - no instability, no temperature issues, nothing.
That is no surprise, though, since I'm currently only running at 182x11, no
real overclocking at all (stock clock speed, faster FSB), on a chip that
should easily at 2.3GHz and beyond with no special cooling setup.

My system will sometimes not boot up, and not give any error message (voice
or beep). Sometimes, it'll give me the "System failed due to CPU
overclocking" voice POST error (no beep). Sometimes, it'll load up just
fine. Sometimes it'll give me the voice POST CPU overclock error and load up
fine anyhow, but Asus Voice POST has never been very reliable, so that
doesn't alarm me. I thought I had a bad motherboard, and spent the time and
money to have mine replaced by the vendor. The first three or four times I
booted up, it seemed like I'd solved my problem, but now I seem to be back to
where I started - a system that I need to reboot four or five times to get it
started.

Is anyone else successfully using one of the low-voltage XP-M chips on an
A7N8X-E Deluxe? I haven't found anything to indicate that they're
incompatible, but I'm having a hell of a time, after trying this chip on two
A7N8X-E Deluxe motherboards, I'm beginning to wonder.

I'm at a loss, and am almost tempted to just give up and build a new system
from scratch, without re-using any of my components. It'd be a colossal
waste of money, but I'm feeling like I'd have a better chance with it, after
trying this system with two CPU's, two motherboards, two sets of RAM, two
video cards, and two power supplies - i.e. I've swapped every single element
out at some point in time, and nothing seems to solve the problem. The only
thing I haven't swapped out is the CPU fan. This one is working fine,
though, and doesn't so much as stutter on the failed boot-ups. Just in case,
I'll try swapping that out, or at least making the system think I did by
connecting a chassis fan to the CPU fan connector for a test (I'm getting
really sick of wiping off and re-applying thermal compound). I'm not
optimistic that the fan could be the problem, though.

Any help would be appreciated. Thanks!

I think you're getting just a little bit frustrated. If you
want to play with mobiles, you have to occasionally get your
hands dirty

Think about how a board like this starts up for a moment. There are
the bridges on the top of the processor, and they set limits or
initial conditions for the processor. Some of these signals go directly
to other hardware, like the clock generator for example. Other signals
are connected to GPIO (general purpose input output) pins, so that the
BIOS can program the value of the bridge.

When the board starts up, all the GPIO are tristated and aren't driving
anything. That means that, before the first instruction executes, the
Mobile will be selecting some weird values. According to the fab51 web
page, the multiplier defaults to 6X, and presumably the CPU clock for
your Mobile is 100MHz (this is based on the coding of the FSB values,
and as this is a desktop board, it interprets the FSB settings different
than a laptop would). This means the processor executes its first
instruction running at 600MHz (my guess, based on fab51 info and the
AthlonXP model10 datasheet).

The VID is higher than it needs to be, for that clock speed, because
it is a desktop board interpreting a mobile VID code. It is
possible the processor is not running smoothly enough before the BIOS
code starts to execute. As a result, if the BIOS doesn't get a chance
to softmod the multiplier, program the clockgen to your selected
frequency, then it is basically crashing.

There is a mod called the L12 wire mod, that changes the default
FSB at startup. Maybe you'll have some success by doing that mod.
Basically, what the mod does, is set FSB1 to logic 0, by grounding
FSB1 to a VSS signal. If you believe the fab51 web site, your mobile
has the FSB signals set to 1 0 (reserved), which the clock generator
could interpret as 100MHz. If you do the L12 wire mode, this
should result in the CPU clock starting at 200MHz before the first
instruction is executed, meaning the core frequency will be at least
1200MHz. The high VID voltage is going to be more consistent with
this frequency than it would be with 600MHz.

Table 26. Front-Side Bus Sense Truth Table (model10 datasheet)
FSB_Sense[1] FSB_Sense[0] Bus Frequency
1 0 RESERVED
1 1 133 M Hz
0 1 166 M Hz
0 0 200 MHz

Fab51 L12 table
L12 [ 3 ] : Closed ( default )
L12 [ 2 ] : FSB_Sense [ 1 ]
L12 [ 1 ] : Closed ( default )
L12 [ 0 ] : FSB_Sense [ 0 ]
L12[3:0]
C:CC 133 MHz Barton - Mobile
^ ^
| |
Open, Closed = 1 0, the RESERVED value, but also obviously 100MHz.
Closing L12[2] will set the FSB to 200MHz, and the startup frequency
will be 1200MHz instead of 600MHz. One way to close L12[2] is
to connect the FSB1 signal to VSS, via a wire jammed into the socket.
(You can also do the mod using a soldering iron, between the pins on
the back of the motherboard. Warranty voiding, etc...)

There is a picture here, and they are connecting a signal called FSB1
to VSS (which is GND or logic 0). I recommend a strand of nickel plated
copper wire, which in years past was quite common, but is harder to
find now. Radio Shack generally has nothing but pure copper, and the
nickel plate prevents oxidation and guarantees the mod won't open on
you and put you back in "flake city". ( I _hate_ bare copper )
http://www.logan.eclipse.co.uk/abit_nf7_(s)_v2_0.htm#TBred_L12_Mod

You can compare this info to the Athlon Model10 datasheet (Barton).
I wish AMD would put a mobile datasheet on the web site, so there
would be a better reference to look at. But, like Intel, mobile info
seems to be "hidden".

(Model10 - see Pg.66 topside view = socket view = the "logan" pic)
http://www.amd.com/us-en/assets/content_type/white_papers_and_tech_docs/26237.PDF

I think you will still have the ability to program the CPU clock
in the BIOS. I believe the clockgen has a serial I2C interface, and
override values can be programmed through there, to override the
200MHz setting that the L12 wire trick creates. The L12 wire mod
is only required for the short interval needed for the BIOS to
finish programming the hardware according to your settings. After
that, the BIOS settings will be used for the rest of the time.

HTH,
Paul

 

[Sam]

I'm so glad I found this post guys. My XP2800+ is doing the same as
yours Erik... saying the bloomin' CPU is overclocked when everythnig is
at stock settings and speeds. And if i try to change settings, notthat I
need to, save to CMOS and reboot there is no display. I have yet to have
a single boot and I too have RMA'd with the supplier. Tis is the 2nd
board I've had. I;m eager to see if your solution has worked in the
long term, and if it would work for me, but i have no chassis fan
any suggestions as to what I can do? Can I buy one perhaps... I have an
old fan but it uses molex power connectors. Please reply soon.

Jonny

Update your BIOS. I think the latest version (1013) fixes that false
message you are getting. You may want to look here:
http://www.asus.com.tw/support/down...me=A7N8X-E Deluxe&Type=Latest&SLanguage=en-us

Sam

 

[Sam]

Oh good lord no! the last A7N8X-E DELUXE i had flipped out on me cos of
a corrupt bios. i am not touching it... invalidate my warranty i don't
think so :| looks like its back to supplier see if they can sort it.
cheers though.

I've upgraded BIOS's dozens of times without incident. Maybe I'm just
lucky. Since the updates come from the Asus site, I hardly think it will
invalidate your warranty. The most you'd have to pay for is a replacement
BIOS chip, which is relatively inexpensive. The main thing I do is used a
script to flash the BIOS after booting on a DOS floppy disk. I always make
sure the machine is on a UPS so that power problems won't interrupt the
process.

Sam