Re: ...thoughts on Retro-uC?

Den Thu, 13 Sep 2018 10:07:54 +0200 skrev Francesco Messineo
<francesco.messineo@gmail.com>:
> On Thu, Sep 13, 2018 at 9:50 AM Konrad B <konrad0x42@gmail.com> wrote:
> >
> > Well, fast 65c02 or maybe ce02 or c816 based MCUs with internal
> > Flash and RAM (and not these Mitsubishi MELPs devices, which you
> > could only buy in zillions and had 100 Flash rewrites guaranteed or
> > so) were something I dreamt about... last century ;)
> >
> > Now if we talk about 6530/6560-61/6581 and all these fancy CSG/MOS
> > chips - some time ago we learned about documents and backup tapes
> > being found in the abandoned GMT/MOS HQ. One gentelmen kindly
> > explained us that "there were no tape drives there so the tapes are
> > trash", but this is still something that concerns me - does anyone
> > know if all of these things (tapes, docs) were put in garbage bin ?
> 
> I refuse to believe there's no way to read those tapes. There're a few
> folks on vcfed.org forum that can surely find a way to read anything,
> it's just a matter of speaking with the right persons, imho.

Today you can also just look at an IC through a good enough microscope
and reverse engineer it from that.

> The problem is probably there's no chip farm with the right process
> active to reproduce the old designs. I was always curious about how
> would it costs to re-activate such an old process and start making
> these chips again.

I'm not sure that would be needed. By looking at how analogue IC's are
being manufactured, or rather what such IC's are available, we see that
it's not an unsolveable problem to make IC's with a modern process
having signal properties compatible with the old IC's.

For CIA's and the other similar IC's (6523, 6525 and maybe 6522 if the
current production is ended) it's good enough to make something
compatible, they don't need to be identical to the old ones.

P.S. don't forget that the 6522 and 6502, at least in CMOS versions, is
still produced today.

Not sure what it would take to change todays production of 6522's into
6526's, but it can't be totally impossible.

> Also, on the purely digital chips, one would just need a bonding
> machine to rewire a "modern logic implementation" to a compatible DIP
> chip package.

Aren't DIP-40 IC's still made in enough quantities to tag alone that
production process? (For example the microcontroller in SD2IEC is afaik
DIP-40 or some similar size).

> The mixed analog/digital chips on the other hand would need to be
> recreated with the old manufacturing process (SID, VIC and few
> others).

SID is for sure hard to make a clone of using some other manufacturing
process.

I doubt that VIC-II would be much of a problem though. Afaik it only
has 4 or 5 analogue levels on the luma signal, and the color signal
seems to only be on/off (and phase selects the actual color).

Worst case a new implementation of a VIC-II could use some surface
mount case and have 2/3 digital outputs for luminance, and you'd just
need a resistor network and some simple buffer on an adapter board from
the surface mount IC and the DIL 40 socket.

> I doubt however that there's enough market even to justify the
> "rebonding logic" approach which should be far cheaper than setting up
> a complete chip farm.

Yeah, the problem is that even though it might be possible to make new
IC's, they might be more expensive than known good used parts, and then
there is no market for the new IC's.


-- 
(\_/) Copy the bunny to your mails to help
(O.o) him achieve world domination.
(> <) Come join the dark side.
/_|_\ We have cookies.