[Gllug] PSU voltages

[Christian Smith]

On Mon, 31 Jan 2005, Nix wrote:

>On Mon, 31 Jan 2005, Simon Wiehe suggested tentatively:
>>> VCore 1:   +2.83 V  (min =  +3.33 V, max =  +3.68 V)   ALARM
>>> VCore 2:   +2.83 V  (min =  +3.33 V, max =  +3.68 V)   ALARM
>>
>> It could be correct. I have an Abit Socket A motherboard and vcore was
>> very low. Investigation showed (via the Abit fora) that the
>> motherboard in question was known for its low vcore and that I should
>> raise it in the BIOS. This I did, but only once I managed to cool the
>> whole case and the CPU drastically to stop it overheating.
>
>Oh, right, so the CPU doesn't *mind* about the low VCore?


The CPU might be more unstable with lower than expected core voltage.
Higher voltages produce cleaner signals (or something like that) at the
expense of additional heat. Lower voltages may make the internal signals
dirty enough at the core speed to cause stability problems. Underclocking
can allow the reduction of core voltage.


>
>I mean, the last time I had lm-sensors working on that box (which has an
>asus motherboard), I don't recall any VCore alarms. But I'm not sure if
>it's the VCore or the lm-sensors alarm thresholds which have changed...
>
>How can we tell what the VCore is meant to be? It's CPU-dependent, isn't
>it?


It's CPU dependent. Modern CPUs generally have lower voltages than even
that, below 2V. As power is related to voltage by a square law, chips
manufacturers try to push core voltages down as low as possible. Pentium
Pro was probably the last desktop Intel CPUs with a 3.3V core voltage.
Pentium MMX and early PII both dropped to 2.8V core voltage, and it's been
going down ever since. I guess you have an early PII or Pentium MMX?  In
which case, the CPU core voltage is fine.

Christian

-- 
    /"\
    \ /    ASCII RIBBON CAMPAIGN - AGAINST HTML MAIL
     X                           - AGAINST MS ATTACHMENTS
    / \
-- 
Gllug mailing list  -  Gllug at gllug.org.uk
http://lists.gllug.org.uk/mailman/listinfo/gllug