[rescue] sun 3/60 update

[Romain Dolbeau]

Le jeu. 22 mai 2025 à 08:28, Dan Moisa via rescue <rescue at sunhelp.org> a écrit :
> Putting the BOM together is a PITA (...)  It will get easier in a subsequent version if I sub out some of the rarer stuff.

Indeed, and using "modern" surface-mount components (when there's a
suitable part to substitute) would save a lot of real estate,
potentially. Figuring out what substitution would work is the hard
part :-(

> Video will be a weird one, I'd really like to get the schematics for the P4-based color CG6. Otherwise I'll implement the onboard BW framebuffer and modernize it to work on VGA instead of ECL, but it's like putting astroturf in the mansion :( At that point there's a question of whether this stays as a single-board-computer. If the P4 CG6 (which of course is different from the 3/80 version) was less of unobtanium, I'd just implement the P4 bus and be home free.

Depending on the level of documentation (and SW compatibility) of the
P4 cg6, it could be substituted. I implemented a cg3-compatible
framebuffer in the SBusFPGA [1], and enough cg6 compatibility as an
alternative to support NetBSD (but a *lot* of the acceleration is
missing, as almost only bit-blitting is actually used, nothing from
e.g. the 3d-specific stuff from the GX architecture). How different
are the 3/60 and 3/80 cg6 supposed to be? I was under the impression
from the pinouts (documented for the 3/60 in the schematics, partially
reverse engineered for the 3/80 from the motherboard) that they are
logically identical and only the backplate was different (my 3/80 blew
a capacitor and so far I haven't got around to trying to fix it). But
I don't own either devices (I only have an SBus and a VME in a 4/330).

I have also implemented enough of the '030 bus in the IIsiFPGA to have
a Mac framebuffer working from the PDS, Apple's process direct slot,
which is similar to the P4 (except for the '030 rather than the '020,
so you can be synchronous). The two shouldn't be too hard to combine
into a cg6 emulation on the '020/'030 bus.

A similar approach could be used to build a (cg3,cg6)-compatible
framebuffer on something similar/identical to the P4. The downside is,
it's a stupidly expensive solution due to the FPGA boards with enough
I/Os being expensive and the need for a fairly large carrier board.
It's around at least 300€ even using the cheaper options (such as
soldering the through-hole stuff by hand rather than using JLCPCB to
do everything).

Cordially,

Romain

[1] <https://github.com/MelkhiorVintageComputing/SBusFPGA>

-- 
Romain Dolbeau