Date : Tue, 04 Mar 2008 22:37:42 +0100
From : dw@... (David Warrington)
Subject: TUBE chip, accessing 'Parasite' side - 16Mhz 65C02S
PROJECT CO-PRO
There's been a lot of talk about building CoPros over TUBE in the last
days,esp. 6809. There seem to be a lot of challenges, which is, of course,
half the *fun*! But may I also suggest taking a look at a "modern 6502"
also? They only cost $8 one off, and matching VIAs are also available for a
similar price. I have been considering building a CoPro using a 65C02S ...
but no time to follow through at the moment.
The 65C02S is manufactured by WesternDesignCentre in both PIC and PLCC
packages and will happily clock up to 24Mhz (this data according to WDC
themselves, even though the "paper spec" is 14Mhz, they overclock easily).
www.westerndesigncentre.com
At 24Mhz, that's 12x faster than a standard beeb (at 2Mhz), or 8x faster
than standard 6502 second processor (at 3Mhz).
If this all sounds *too easy* and not enough *fun*, one could think about
paged memory on the CoPro rather than just a single memory map... and
therefore run a modified BASIC with one 64K bank for program, and another
64K bank for variables.
>> QUESTION 1: Does anyone have source code for BASIC4 or later? Note this
>> Q is for assembler source, not a hex ROM dump.
PROJECT HOST
I have a solidisk 4MEG board that replaces the original 6502 on the beeb
with a 65C02. This allows BASIC4 to be run on the host, and ALSO clocks it
at 4Mhz. The board was relatively simple, because EPROMS and static RAM
where capable of the 4Mhz speed. So the BASIC ROM and paged ROMS were
executed at 4Mhz. Data access to the mainboard was choked to 2Mhz. But
since 95% of the time, code was executed with a paged ROM, and loading and
saving to memory just 5%, the speedup was pretty much 100%.
I think ANYONE still using a beeb today should just "drop in a 65C02" or
"65C02S" into their beeb to allow them to use BASIC4 and the "65Cxx"
mnemonics in assembler, e.g. PHX, PHY, STZ, etc. This is a no brainer and a
5 minute exercise. Without the solidisk board, the CPU will still be running
at 2Mhz, but the availability of running 65Cxx code and BASIC4 is worth the
upgrade. (including faster Foating Point variables AND half a significant
digit more accurate, but still not IEEE754).
But to make a project, I think the solidisk 4MEG was very interesting. It
would be possible to build an 8 or 16Mhz host board using the 65C02S. The
major difficulty would be EPROMS. All existing 1980s UV-EPROMS are too slow.
So it would need an alternative solution to sideways ROM address space,
perhaps modern PLCC EEPROMS similar to a modern PC BIOS chip.
>> QUESTION 2: Does anyone else have a solidisk 4MEG board, or data/spec
>> for it?
With a 65C02S on board, BASIC4 is faster than BASIC2, as well as additional
machine code instructions. However, the OS remains the same "old" 6502 code.
If this could be upgraded to 65C02 code, it would be more efficient (less
memory space used) and marginally faster (use of PHX instead of TXA, PHA, or
STZ instead of LDA#0, STA, etc). Saving memory would be useful... some
addiitonal code or *new* OSBYTE commands could be put there for other
projects. Perhaps "stealing" from Master OS3.20 to create B32K OS1.30.
>> QUESTION 3: Does anyone have source code for B32K OS1.20? Or Master
>> OS3.20? Note this Q is for assembler source, not a hex ROM dump.
David