On Wed, Feb 25, 2004 at 04:04:35PM +0100, Maciej W. Rozycki wrote:
> On Wed, 25 Feb 2004, Chris Tribo wrote:
> 
> > 	I believe that the psu fan has some kind of temperature or
> > rotation sensor system. If the fan is not spinning fast enough it will
> > kick off. I think there's also a thermal sensor in the CPU itself, but
> > NetBSD doesn't monitor it I don't believe. I'm probably spouting
> > conjecture more than actual fact, but hopefully that gets you on your way.
> 
>  There's a software-readable thermal sensor in the PSU of all DECstation
> 5000 machines.  It's available as a bit in the CSR in the /200 and in the
> SIR of the I/O ASIC in the others (therefore the latters can use it as an
> IRQ source).  The bit is active if an overheating condition happens -- I
> don't know what temperature it corresponds to (and I'm not that interested
> in checking it empirically).
> 
>  I haven't heard of any temperature sensor in any of the processors used
> for the DECstation.  There may be one outsidew the CPU on the KN04 and
> KN05 daughtercards as there's a bunch of new registers of littly known
> purpose in the combined MB/MT ASIC there.

Thanks for information guys.
I've since pulled the box to shreds and de-dusted it (actually very little
inside!)

On the overheating track, the PSU fan spins fine, but I noticed that the RZ25
SCSI drive gets rather hot (too hot to touch in fact).

Additionally, the Motorola branded chips on the CPU daughter board (I'm assuming
they're the cache?) get way too hot to even lay a finger on.
I'm wondering if this is normal and whether it might be a good idea to whack a
heatsink on them?

Guo-Rong