[rescue] voyager and pcmcia wifi with solaris

[Dan Moisa]

I noticed the thermals aren't that bad at the rated speed. The case itself
is pretty open and well laid out for passive cooling, and I wonder how
destructive it would be to remove the current stacked pancakes heat
spreader and put something bigger in. I also noticed that the CPU itself is
labeled -70, but reported by the system to be running at 60Mhz. Not sure if
that label is speed rating, it's a sticker and not printed on the heat
spreader like it is for the -110.

Romain you're likely right that a drop-in might not work out of the box
unless the clock multiplication is entirely internal. I'll look at what
clocks exist on the board.
https://datasheets.chipdb.org/Sun/microSparcII-STP1012.pdf seems to have
all the info I'd need. It's a bit weird that the system clock is pinned and
divided down for the bus, I'm used to bus speed being the guiding factor
and the multiplier determining the system clock.

For the power side there's a DC/DC plug-in/replaceable component that could
presumably be beefed up.

Dan.

On Sun, Nov 9, 2025 at 12:16 AM Romain Dolbeau <romain at dolbeau.org> wrote:

> Le dim. 9 nov. 2025 à 03:46, Dan Moisa via rescue <rescue at sunhelp.org> a
> écrit :
> > Speaking of Voyager, can I take a microsparc-II at 110Mhz out of a 5 or
> similar and substitute the @70 it comes with? So far I'm led to believe
> that yes.
>
> What you need to do depends on how the MicroSPARC II is wired.
> Presumably, it is using the internal PLL, so:
> * you need to change the clock supplied to the CPU to whatever you
> want to run it at;
> * you need to check what that clock is also feeding to make sure
> you're not overclocking unwanted components;
> * you need to change the input to the PLL divider (pins DIV_CNTL[1:0])
> so that the SBus is not overclocked. I assume it is 1/3 for a 70 MHz
> MicroSPARC II, you would need at least 1/4 up to 100 MHz (the voyage
> can probably take the SBus moving from 23.333 to 25 MHz), and 1/5 over
> 100 MHz.
> * whatever else that might be needed
>
> Both then as Andrew said, power/thermal are going to be the problem.
> They're unlikely to be dimensioned for +50% on the CPU.
>
> Cordially,
>
> --
> Romain Dolbeau
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://sunhelp.org/pipermail/rescue_sunhelp.org/attachments/20251109/35c59663/attachment.html>