[ltp] Re: An "inverse" fan problem

[Michael Keyl]

On Friday 27 May 2005 04:19, Alejandro Bonilla wrote:

> Still, I see this going too hot. Whichever linux version or whatever app
> that you could be running, your lappy should not reach the 60 C, if it

"Intel=AE Pentium=AE M Processor Datasheet" at

ftp://download.intel.com/design/mobile/datashts/25261203.pdf

says to this point:

"The processor must remain within the minimum and maximum junction temperat=
ure=20
(Tj) specifications at the corresponding Thermal Design Power (TDP) value=20
listed in Table 23."

and table 23 says:

TJ Junction Temperature:  0C, min 100C max

Hence "not above 60C" seems to be a little bit too paranoid to mee ;-)=20
However, its probably not a good idea to come to close to the upper limit; =
in=20
particular because the thermal diode used for (user controllable) thermal=20
management does not provide a precise measure for this junction temperature=
=20
(but it should be close; cf. the discussion of these issues in Section 5.1 =
of=20
the above document). I have decided to trust the engineers who have provide=
d=20
the defaults for the trip points, which are

critical (S5):           94 C
passive:                 92 C: tc1=3D5 tc2=3D4 tsp=3D600 devices=3D0xdfffef=
e0

on my machine. At the passive temperature the clock is reduced from 1.7Ghz =
to=20
1.2Ghz (under ACPI, under APM it even goes down to 600Mhz) and the CPU=20
temperature drops below 80C within seconds. I guess this is sufficiently=20
secure.

Michael