On Apr 25, 6:29 am, "Orson Cart" <ex-pri...@parts.org> wrote:
> I was given a PC because it "crashed after a while".
> When I peeked in the BIOS settings, I noticed the CPU temperature
> was too low, e.g. 19 degrees. The fan seemed to be going pretty
> slow, so I guess it just overheated. I set the fan speed to full on,
> and replaced the CPU (in case it was toasted), and there has been no
> misbehaviour ever since. Although the new CPU still shows chilly
> temperatures. BTW the old CPU works in another mainboard - it
> was a Conroe 1860 MHz - probably not generating enough heat to
> cremate itself.
> Now I started wondering about how this is measured. I believe the
> BIOS used a thermistor under the CPU. I guess it is driven by a
> current source and the voltage measured. So if any of the circuit is
> defective, mainboard thinks the CPU is cool, and idles the fan.
> The CPU has a diode inside to monitor itself, and throttle if too
> hot (dropping multiplier). Ideally it should drop the voltage as
> well, then the CPU must tell the mainboard to lower the voltage
> when it cuts it's own multiplier.
> Are there datasheets for the chips that monitor the mainboard
> sensors, giving details how they work?
There's no correlation to a fan and an overheating processor. Fan
bearings wear out so they can squeak and squeal. A MPU is driven into
overheated conditions so a poor installer, or its overclockers, can
watch die pathways degrade into instability if not an irreversible
failure condition. Thankfully, due to an innate latter
predisposition, the condition occurs after they've lost interest and
the MP is figuratively thrown over a shoulder.
The Allendale, possibly updated to Wolfdale variant, operationally at
1860Mhz, once more and again in analogy to the fan is an independent
derivative of base operational frequency of failure skews. In other
words, whether a processor base speed can be said to exhibit a
permissible leeway for overclocking purposes, does not then reflect
upon a base frequency within which MPU is engineered. A 350Mhz,
because it is 350Mhz is no safer at 375Mhz, than can be said of
1860Mhz, rather than were it 1900 or 2000Mhz.
Temperature is a reasonable reflection of a wide range of cooling
provisions for the MPU, as well as the wattage draw within an
engineered framework for the processor placement. To be otherwise,
simply indicates a fault or compromised position due either to
mechanical components less apt to be within a tradition misconstrued
by an operator.
|
Similar Threads
