[SunRescue] CPU switching

James Lockwood lockwood at ISI.EDU
Fri Sep 17 19:17:25 CDT 1999


On Fri, 17 Sep 1999, Bjrn Ramqvist wrote:

> Luckily I have such a pocket-torch at home, which I usually need when I
> do some shrink-tubing on cables.
> Very handy piece.

Also great for bending plastics.  Very handy tool when you're building a
nitrogen laser.  :)

> ???
> And I was thinking it was just the jumpers a'la PeeCee...
> (although I must keep in mind I'm NOT dealing with a PeeCee)

Nope.  The SS5 derives CPU clock off of a single external oscillator.  The
jumpers only affect memory cycle timing.  This is the same in all
pre-Ultra systems (though the SS10 and SS20 feature selectable
oscillators), the Ultras have a software-settable CPU clock (they actually
start up at the lowest possible speed and then the PROM automatically
ramps up clock frequency at an early point).

> How could the CPU be fried? I mean... I've never ever experienced that a
> CPU actually can die.
> Is that common? Sounds very odd to me.

It's very odd that a SS5 is dead at all, they are quite robust.  The
motherboard contains more parts that can fail, but the CPU is more
complicated and more sensitive to damage internally.  The chances are
pretty even.

What's the failure mode?  Have you tried verifying that any other parts of
the system are functional?

> Are these hard to come by? Expensive? (Anyone knows where to find them?)

I saw one at a swap meet last week for $20, though they usually go for
closer to $200.  I believe that the boot PROM upgrade is needed, though it
may only be needed to run SunOS 4 (as it puts the CPU into microSPARC
emulation mode).

I agree with another poster on the idea of using a buffer socket to
prevent damage to your working motherboard.

-James







More information about the rescue mailing list