[rescue] Did i nuke my SS20?

Ray Brooks ray.userinterface at gmail.com
Mon Aug 2 09:50:44 CDT 2021


All,
  The good news is that the motherboard isn't nuked. I've managed to get it
to POST via serial by stripping it back to the absolute bare bones as
suggested by Romain. I need to do some more digging around here to fully
track down the problem but the good news is that nothing important seems to
be broken. Phew!

Thanks all for this invaluable and very timely information. Delighted to be
a part of the list! :)

Regards,

Ray

On Mon, 2 Aug 2021 at 15:25, Romain Dolbeau <romain at dolbeau.org> wrote:

> Le lun. 2 aoC;t 2021 C  15:37, Ray Brooks <ray.userinterface at gmail.com> a
> C)crit :
> > Can anyone offer any advice as to next steps, or to help me
> > identify the issue?
>
> Having just replaced a NVRAM chip in a SS20 (using a MT48T08-100PC1
> brand new from Mouser), the first POST can take a lot longer than
> expected.
>
> Suggestion:
>
> a) remove all DRAM but the one in the first slot, to minimize memory checks
> b) only keep the module in MBus slot 0 (bottom), to minimize mbus checks
> c) keep an eye on your keyboard leds
>
> When you power-up, all 4 LEDs should blink once, then the caps-lock
> one should start blinking to indicate POST.
> After a while (shortened by a) and b) above), the screen should sync
> and give you access to the PROM. Beware some screen will need nudging
> to detect the signal if there was none at power-up (my Iiyama 17" CRT
> defaults to the VGA port rather than the BNCs where the SS20 is, so I
> have to force the switch from VGA to BNC if the signal doesn't come up
> fast enough on the BNCs).
> All the details on the LEDs are available (and which slot/port is
> which) in the service manual 801-6189-12 (Which Oracle still has
> online, they are not all bad:
> <
> https://docs.oracle.com/cd/E19127-01/sparc20.ws/801-6189-12/801-6189-12.pdf
> >)
>
> If there's no blink at all (not even the power-up one), you may have
> troubles (or the keyboard is dead/undetected).
>
> If there's the initial all-blink but no follow-up on the caps-lock
> led, then there's a POST issue. Double-check the seating of the DIMM
> and MBus boards, and of the NVRAM chip; the NVRAM can easily be
> inserted off-by-one-pin!
>
> If the system reaches the caps-lock blink, then the final pattern will
> indicate an error if there' s one, IIRC; but you should get some form
> of display eventually.
>
> Cordially & good luck,
>
> --
> Romain Dolbeau


More information about the rescue mailing list