2 Sep
2006
2 Sep
'06
8:56 p.m.
Roy J. Tellason wrote:
People should also be aware of the misleading stuff in
the way the speed of
that part is presented. They called it "a 4 MHz Z80" but even though it was
indeed that chip the way they had it configured in there was so tightly
coupled with the rest of the system that it only ran some of the time,
rather than all of the time, giving an effective clock rate of somewhere
around 2.5 MHz.
Misleading it was, though we should assign blame to the Marketing folks.
When the C128 was being developed, Marketing wanted to hit the business
market (they had tried to position the Plus/4 there and failed).
Someone remember the old C64 CP/M cartridge and decided to play that up
(the C128 was C64 compatible, they reasoned, so the CP/M cart could be
used to offer CP/M support).
The problem was, the CP/M cart sucked, and would only work on the
earliest models of the C64. It was a direct copy of the Apple CP/M
cart, according to some reports, but didn;t take into account the
specifics of the C64 expansion port.
Brian Bagnall documents the specifics in his book, but in general, Bil
Herd got the CP/M cart to work by "accidentally" designing the Z80 into
the motherboard.
However, as the design-in was accidental, there was little time (or
ability) to do a proper design-in. If it had been a sanctioned design,
the full 4MHz would have been offered (and probably more), but CBM was
cheap, and Bil was trying to hit a mandated requirement, not a speedy
CP/M design.
Although the book does not document it, I think Bil had to fight tooth
and nail to keep the accidental work in place.
However, as Bil later notes, the Z80 ended up saving the CBM bacon, as
it was used to deal with badly misbehaving cartridges that did funky
things with the C64 MMU lines, which the C128 could not deal with on
machine startup. Thus, the Z80 actually starts the C128 up, and then
passes control to the 8502 (6502/10 compatible) CPU.
Jim