Date   : Thu, 16 Apr 2009 12:43:37 +0100
From   : mfirth@... (Michael Firth)
Subject: ReCo6502 now available

> ----- Original Message ----- 
> From: "John Kortink" <kortink@...>
> >[...]
> >
> >You've all already mentioned the main points
> > * CPLDs don't have enough register bits versus going to 200+ pin 
> > packages
>
> Altera's Max II CPLDs could do, e.g. 570 cells (which I
> suppose is enough), available in a 100-pin flatpack. But
> only 3.3V tolerance, so needs level translators for most
> I/O.
>
>
> There are many copros that could make use of the same
> ULA replacement, ReCo6502 included. Which was an
> additional reason not to attempt integration of
> a ULA remake in ReCo6502 at this stage, and thereby
> sort of re-invent the wheel before its invention.
>
>
I guess there are several questions that need to be answered to determine 
whether a standalone ULA replacement, or a 6502 co-pro redesign 
incorporating a ULA replacement is the best road to go down.

For the standalone ULA replacement to be useful for other co-pros, the ULA 
would need to be a common first point of failure - does anyone have any feel 
as to whether that is the case?

I guess the next question is whether there is a demand for new 6502 co-pros 
beyond what can be easily resourced from eBay or ReCo6502s with ULAs taken 
out of other co-pros. Given that an external 6502 co-pro on eBay seems to go 
for at least ?50, usually a lot more, I would say that they are in demand.

Personally, I find the 6502 co-pro more interesting than most of the other 
BBC co-pros, as its the main one that is extending the BBC's capabilities, 
rather than trying to make the BBC emulate something else.

I think that if you want to run CP/M there are probably easier and better 
platforms than a BBC with a Z80 co-pro, and for running DOS there are 
certainly much easier platforms than the Master 80186 co-pro.

That only really leaves the 32016 scientific co-pro, and the ARM co-pro, as 
BBC extending alternatives and both of them are pretty rare anyway. Neither 
of them can run 6502 assembler either, which is what a lot of BBC 
enthusiasts are familiar with.

Another advantage of a new 6502 co-pro design incorporating a ULA 
replacement is that it would be possible to run the whole of the co-pro part 
at 3.3V, making the level translation task only relevant for the host 
processor facing pins of the ULA. It would possibly also make sourcing the 
parts for the co-pro side easier, as RAM chips seem to be more common as 
3.3V parts than 5V parts these days.

Regards

Michael