"Absolute Maximum Ratings
Stresses greater than those listed in Table 13 may cause permanent
damage to the device and motherboard. Systems using this device must
be designed to ensure that these parameters are not violated.
Violation of these ratings will void the product warranty. Exposure
to absolute maximum rating conditions for extended periods may
affect device reliability.
VDD supply voltage relative to VSS 0.3 V to 1.65 V "
That is why the Asus BIOS doesn't allow more than 1.65V.
You'll just have to do a volt mod, if you want more.
It is my experience, that 50% of the voltage modding techniques
documented on the web, are just a bad idea. You must be
especially careful to do your own analysis of how each volt
mod technique proposes to work. Some of the ideas would pick up
noise and give a shaky voltage to the CPU - those are the ideas
I like to avoid. Modification via the VID signals, because they
are digital control signals, is a safer bet, but we don't know if
any GPIO signals driving these pins would be safe to ground or
not. (I have managed to trace the control signals on one of
my older motherboards, but it is a time consuming process. If
the GPIO use series limiting resistors, then forcing the VID
pins to ground would be safe.)
The SC2643 is especially flexible, and supports several VID codings.
The first step, is determining which coding is being used on
the chip. The VID5 pin is a multi-level logic input - the schematic
in the datasheet shows the three voltages that can be connected to
the pin. What I don't understand here, is how it would be possible
to use standard VID codings, and yet get more than the K8 coding
limit of 1.55V. Using the pin5 error offset adjustment terminal,
would boost all voltages, and I doubt that would be acceptable to
Asus (to reach 1.65V). Using the VRM9.X coding also does the
equivalent of boosting all voltages, so the only way this could
work, is if the processor VID signals are not directly
connected to the SC2643.
Perhaps you could use a multimeter, measure the voltage level
presented on all the VID pins, as a function of the voltage
set in the BIOS or via a Windows tweaking util. That might give
you some idea as to how the SC2643 is configured. (Take your
system out of the computer case, and do this work with the
system assembled on your work table - that is the only safe
way to be probing signals on a motherboard. It is too easy to
short stuff, if you are probing while the mobo is still in
the computer case.)
In any case, you'll have to do some sleuthing first, before jumping
into a volt mod. At least I hope, by presenting the information
from the AMD processor datasheet 31411, that you know you are
in suicide territory, in terms of voltage. I am all in favor of
overclocking, as long as it is done within the bounds of the
hardware - after all, the idea is to get something for nothing,
and a suicide overclock that lasts for only one month of operation
is not money well spent. If you wanted a "heroic overclock",
you would have started with a FX processor, right ?
Many of the volt mod techniques that use three terminal pots, will
kill the hardware if the pot is twisted all the way to one extreme
or the other. Also, if a wire breaks off, while the system is
powered, the instant overvolt can kill the motherboard, the processor
or both. The connections on the board are very tiny, and it is
difficult to get pots or resistors connected. I've modded my A7N8X-E
a while back, and was at the limits of my soldering skills (I need a
stereoscopic microscope to do better work, but cannot afford one of
those for home projects) - and I do that kind of stuff for a living.
If you've never soldered before, I don't recommend making your first
mod on a $175 motherboard.