Richard said:
I have a P4P800se with a retail 2.4c P4 cpu. The vcore seems to be stuck
at 1.6v. No matter what the cpu voltage is set at the board reads 1.6v.
What gives?
Another question regarding 'vid cmos setting' in the bios. What is this?
Mine is set to '62'. Does this have any bearing on the vcore?
Thanks
I've been looking at your question since your first post, and
am no closer to coming up with a theory as to why it is doing that.
The "VID cmos" is a most curious setting. It appears next to the
multiplier setting section, implying the VID cmos thing has something
to do with ES (engineering sample) unlocked P4 processors. Since ES
processors don't have a spec that I've ever seen, it is hard to
guess why this setting is necessary, when there is a very nice
Vcore setting option elsewhere in the BIOS. There are only a
couple of posts on Abxzone about it, and that is where I
read about this speculation, that it is related to ES processors.
There are two power specs for Pentium processors. The ones I've got
are VRM 9 (5 bit VID code) and VRD 10.0 (6 bit VID code, socket 478).
There is a VRD 10.1 spec, but it is for the LGA775 socket.
http://www.intel.com/design/pentium4/guides/24920504.pdf (VRM9)
ftp://download.intel.com/design/Pentium4/guides/25288503.pdf (VRD10.0)
The reason I identified these two specs, is many Asus boards can take
two kinds of processors that correspond to those two specs. The first
spec handles voltages from 1.1 to 1.85V, and that was used for early
Pentium processors (1.75V being a typical value). Later processors
use the 0.8375 to 1.6V spec. Apparently, a pin called BOOTSELECT
on the later processors, can be used to tell whether the former
or latter spec is required. Page 22 of this Vreg spec shows an
example of a dual spec regulator circuit, using the BOOTSELECT pin.
http://www.analog.com/UploadedFiles/Data_Sheets/85022276ADP3181_prg.pdf
Now, the value "62" appears to be decimal, for the bit pattern 111110.
That is a 6 bit code, presumably corresponding to a VID code for a
VRD 10.0 processor. But the value 62, when you look it up in the spec,
does not correspond to a "useful" value. It would give 1.1V to a
processor, which would likely cause it to fail to boot. The value
42, a bit pattern of 101010, gives 1.6V.
What other things do we know ? Asus doesn't like to allow users
to undervolt processors. Generally, overvolting is prevented by
Asus, by limiting the hardware to a safe value. 1.6V is the highest
value that a VRD 10.1 processor can get.
The absolute max voltages for the 1.6 and 1.85 volt processors are
1.75 and 2.1 volts. So, your setting problem doesn't appear to be
related to overvolting, as that limit would more likely be 1.75V
at least.
About the only speculation I can offer, is part of your board
thinks it has got a VRM 9 processor, but the BIOS checks the
CPUID register in the processor and decides it is a VRM 10
processor, and sets the voltage as high as that spec will allow.
Maybe it has something to do with the BOOTSELECT pin malfunctioning
on the processor, or not making good contact ? Or some other
aspect of mis-identifying the processor ? Does your board have
any socket mods or volt mods ? Is it a new board or a used one
that someone could have modified ?
The 1.6V probably isn't going to hurt anything, as long as the
1.75 absolute max isn't being exceeded.
Another question for you - where is the 1.6V being measured ?
Are you seeing this in the BIOS hardware monitor page or
are you looking in Asus Probe/MBM5 ? Maybe the voltage you
are looking at isn't the Vcore, but is some other voltage ?
There has to be a logical explanation somewhere there, and
looking at Vcore related issues is making your symptoms
hard to explain. At least, I haven't been able to cook up
any half baked theories yet.
HTH,
Paul
