Re: 8520/21

On 2015-03-07 12:22, Gerrit Heitsch wrote:


> Might not be a 'we need it now' but more along the lines 'We're moving
> away from NMOS (The 8xxx-Chips started in 1984 after all)'. Keeping the
> NMOS line running besides the HMOS line would soon become too costly.
> Well, what do you do then? Do another process change and port the 6526
> to HMOS or take the 8520 which was already there and put the few extra
> transistors back in to make the counter behave as a clock again? The
> latter seems to be the more rational approach. Since the 8521R0 works it
> was the better way, chip revisions, meaning mask changes cost quite a
> bit of money, even at MOS/CSG.

That's more or less what I am about. What seems irrational is to go for 
the 8520 w/o the BCD-TOD and then make two chips rather than one, which 
at that time could serve all machines (Amiga was still being developed 
and the full software compatibility wasn't a factor). Of course those 
are pure speculations but to me it looks like a "wake up call" once the 
8520 was already in production. In this case the approach I meant and 
you described in more details looks like the most rational one, despite 
CBM/CSG being a corpo under heavy changes.

> Judging by the numbers, the 8565 in the C64C was started _before_ the
> 8566 for the C128, suggesting that the move to HMOS was an ongoing
> process for MOS.

I don't question it. Move to HMOS seems like being planned and executed 
properly. It was just the 8520/8521/6526A/B story that doesn't seem to 
make much sense. We might - obviously - be still missing something.

-- 
SD!

       Message was sent through the cbm-hackers mailing list