[rescue] SunPCi2 woes

[Curious George]

Hi,

--- On Tue, 6/10/08, Marc Mirza <marcm at spri.levels.unisa.edu.au> wrote:

> Hi THere, While I've yet to see any caps fail on these

HAve you looked carefully at them?  THey tend not to be
very noticeable until you deliberately eamine them *or*
they outright *fail*  :>

> we've had all of 
> our SunPci CPU fans fail (around 15+). It was impossible to
> get a 
> replacement fan from Sun and I was unable to find any
> equivalent. I 
> think they are quite prone to failure as they rev. very
> high and the 
> slightest amount of dirt/ dust causes grief.

Yes.  Unfortunately, they *scream* when they fail.

> The fix I came up with was to:
> 1/ remove the heatsink/fan from the SUNPCI card (4 small
> screws with 
> springs hold these in).
> 2/ remove the plastic fan from the copper heatsink.

You can remove the fan without removing the heatsink
(some plastic "clips" hold the plastic shell together)

> 2/ using thermally conductive glue- glue a heatsink/fan
> combo from a 
> Pentium1 style CPU onto the existing copper heatsink.

Yikes!  Not just the fan??

> 3/ Power the new fan from a spare H/Disk power connector.
> 
> Hope this helps

Hmmm... I was hoping for something a bit more "elegant"
but I suspect any solution will have to be a kludge.  :<
I was trying to identify the socket and look for other
heatsink/fan assemblies designed to mate with that socket.

It *looks* like it might have similar mounting holes as
the "Slot one" (? I am not sure of correct terminology
here... intel has so many frigging package variants!)
package with "active heatsink".  I may sacrifice a P3
just to see if that fan assembly lines up with the mounts
in the SunPCi2 card.

(sigh)  What a wonderful design.  NOT!

THanks!

> If all that sounds too cryptic let me know and I'll
> send you a picture of what we've done..