K8N-E Deluxe - Wrong voltage for the CPU?

G

Guest

Guest
Archived from groups: alt.comp.periphs.mainboard.asus (More info?)

Hello all.

I am the not too proud owner of a K8N-E Deluxe and today I
have realized that my motherboard is supplying 1.55v to the CPU as
reported by the nvidia system utility (it is set to auto in the BIOS).
The CPU in question is an Athlon64 2800+ Clawhammer. Am I right to
assume that that voltage is wrong? Shouldn't it be supplying just
1.50v to it? Maybe the nvidia utility is wrong?

Kind regards.
--
Jose M. Arnesto
j[insert my surname here]@computer.org
 
G

Guest

Guest
Archived from groups: alt.comp.periphs.mainboard.asus (More info?)

Did you ever go to the website for ASUS and look at the voltage for
your CPU. You could probably have done that in the same amount of time
it took to complain on the NG.

On Tue, 23 Nov 2004 20:20:30 GMT, Jose M. Arnesto <j[insert my surname
here]@computer.org> wrote:

>
> Hello all.
>
> I am the not too proud owner of a K8N-E Deluxe and today I
>have realized that my motherboard is supplying 1.55v to the CPU as
>reported by the nvidia system utility (it is set to auto in the BIOS).
>The CPU in question is an Athlon64 2800+ Clawhammer. Am I right to
>assume that that voltage is wrong? Shouldn't it be supplying just
>1.50v to it? Maybe the nvidia utility is wrong?
>
> Kind regards.
 

Paul

Splendid
Mar 30, 2004
5,267
0
25,780
Archived from groups: alt.comp.periphs.mainboard.asus (More info?)

In article <mc67q0h9k2j9kd7293thn23ba372d6q4pa@4ax.com>, Jose M. Arnesto
<j[insert my surname here]@computer.org> wrote:

> Hello all.
>
> I am the not too proud owner of a K8N-E Deluxe and today I
> have realized that my motherboard is supplying 1.55v to the CPU as
> reported by the nvidia system utility (it is set to auto in the BIOS).
> The CPU in question is an Athlon64 2800+ Clawhammer. Am I right to
> assume that that voltage is wrong? Shouldn't it be supplying just
> 1.50v to it? Maybe the nvidia utility is wrong?
>
> Kind regards.

This doc has voltages, if you can figure out which processor
you've got. I think you are right, it should be 1.50V.

http://www.amd.com/us-en/assets/content_type/white_papers_and_tech_docs/30430.pdf

As I stated in a recent post, many different Asus motherboard
models are high by 60mV, as measured in the hardware monitor.
Now, many think the provided hardware monitor is a model of
hardware design, accurate to five digits. It is not. It is
worse than the cheapest digital multimeter you can buy at
the store. That means you cannot trust the readings, except
as an indication to dig out a meter and verify something
that doesn't look right.

To prove this point, I can look at the listed voltages for
my PSU, in the hardware monitor. They will be off by a bit.
If I dig out my multimeter, and measure the spare power
supply leads coming from the power supply, the voltages are
very close to being correct. Nowhere near the same
deviation from ideal values.

As a consequence, you cannot draw any useful conclusions from
the hardware monitor. Only a digital multimeter can give you
some idea of what the voltage is. My $100 meter is accurate
to maybe 1% or 1.5%, so it is not a model of accuracy either.
You need a serious lab quality multimeter, if you want to
prove or disprove the Asus design. And, you have to use the
proper measurement technique. On at least one processor
datasheet I've looked at, there is a monitor pin on the
processor socket, and that is the only place you can touch
with a multimeter, and get a number which indicates what the
processor is really getting. When currents of up to 60 amps
flow from the Vcore switcher, to the processor socket, there
will be voltage drop across the copper plane carrying the
60 amps, so randomly probing the copper plane is a waste of
time.

This document contains "Absolute Max" info for a Socket 754
processor. I got this a while back, so download the latest
one, to get more up to date info.

http://www.amd.com/us-en/assets/content_type/white_papers_and_tech_docs/31410.pdf

Absolute max is 1.65V for VDD (which is, I presume, Vcore).
You have a ways to go yet. The only consequence right now,
of the extra 0.05V, is a bit more heat. Also, for our
overvolting friends, note that the memory controller is
only rated for 2.9V max, so watch out when using winbond
BH-5 memory chips and 3.1V DIMM boosters! That 3.1V you
feed the memory, is also going to the processor's
memory interface power pins.

If you want a way to control this overvoltage, turn on
Cool N' Quiet, then observe the voltage. Without CNQ,
you would see the normal load line (i.e. 1.55V at no
CPU load, maybe 1.4V at 100% computing load). When CNQ is
enabled, the OS turns down the Vcore and Fcore, when there
is no CPU load. They get turned up, when there is 100% load.
But, at full load, there will be the droop of the load line,
which in my example above, would be maybe 1.4V. So, with
CNQ enabled, there shouldn't be anywhere near the same
situation you have now. Note that, if you are doing a lot
of real time applications, CNQ may interfere with timely
delivery of streaming data, so test CNQ thoroughly with your
applications, before leaving it as your permanent setting.

HTH,
Paul
 
G

Guest

Guest
Archived from groups: alt.comp.periphs.mainboard.asus (More info?)

On Tue, 23 Nov 2004 21:31:47 -0400, nospam@needed.com (Paul) wrote:

>In article <mc67q0h9k2j9kd7293thn23ba372d6q4pa@4ax.com>, Jose M. Arnesto
><j[insert my surname here]@computer.org> wrote:
>
>> Hello all.
>>
>> I am the not too proud owner of a K8N-E Deluxe and today I
>> have realized that my motherboard is supplying 1.55v to the CPU as
>> reported by the nvidia system utility (it is set to auto in the BIOS).
>> The CPU in question is an Athlon64 2800+ Clawhammer. Am I right to
>> assume that that voltage is wrong? Shouldn't it be supplying just
>> 1.50v to it? Maybe the nvidia utility is wrong?
>>
>> Kind regards.
>
>This doc has voltages, if you can figure out which processor
>you've got. I think you are right, it should be 1.50V.
>
>http://www.amd.com/us-en/assets/content_type/white_papers_and_tech_docs/30430.pdf
>
>As I stated in a recent post, many different Asus motherboard
>models are high by 60mV, as measured in the hardware monitor.
>Now, many think the provided hardware monitor is a model of
>hardware design, accurate to five digits. It is not. It is
>worse than the cheapest digital multimeter you can buy at
>the store. That means you cannot trust the readings, except
>as an indication to dig out a meter and verify something
>that doesn't look right.
>[..]

Yes, I thought about the monitor not being too accurate, but
if I set the voltage by hand in the bios (to 1.50v), then the nVidia
system utility correctly reports 1.50, so that must not be the
problem.

>If you want a way to control this overvoltage, turn on
>Cool N' Quiet, then observe the voltage. Without CNQ,
>you would see the normal load line (i.e. 1.55V at no
>CPU load, maybe 1.4V at 100% computing load). When CNQ is
>enabled, the OS turns down the Vcore and Fcore, when there
>is no CPU load. They get turned up, when there is 100% load.
>But, at full load, there will be the droop of the load line,
>which in my example above, would be maybe 1.4V. So, with
>CNQ enabled, there shouldn't be anywhere near the same
>situation you have now. Note that, if you are doing a lot
>of real time applications, CNQ may interfere with timely
>delivery of streaming data, so test CNQ thoroughly with your
>applications, before leaving it as your permanent setting.

If I turn CnQ on, the voltage is correctly switched between
1.3v (at no load) and 1.5v (at full load), and those are the voltages
that the nvidia utility reports, so I think that this whole issue is
simply a bug in the BIOS.

Thank you very much for your long and informative post.

Kind regards.
--
Jose M. Arnesto
j[insert my surname here]@computer.org
 
G

Guest

Guest
Archived from groups: alt.comp.periphs.mainboard.asus (More info?)

On Tue, 23 Nov 2004 17:29:35 -0600, Bobbie Leslie <bleslie3@cox.net>
wrote:

>Did you ever go to the website for ASUS and look at the voltage for
>your CPU. You could probably have done that in the same amount of time
>it took to complain on the NG.
>
>On Tue, 23 Nov 2004 20:20:30 GMT, Jose M. Arnesto <j[insert my surname
>here]@computer.org> wrote:
>
>>
>> Hello all.
>>
>> I am the not too proud owner of a K8N-E Deluxe and today I
>>have realized that my motherboard is supplying 1.55v to the CPU as
>>reported by the nvidia system utility (it is set to auto in the BIOS).
>>The CPU in question is an Athlon64 2800+ Clawhammer. Am I right to
>>assume that that voltage is wrong? Shouldn't it be supplying just
>>1.50v to it? Maybe the nvidia utility is wrong?
>>
>> Kind regards.

Learn a bit of netiquette and then try again (or not).

Plonk!

--
Jose M. Arnesto
j[insert my surname here]@computer.org