31 Mar
2002
31 Mar
'02
9:35 a.m.
Andreas Freiherr wrote:
My "self-made 6809 computer" project
addresses the problem of complexity
vs. functionality by composing the machine from modules. This way, you
can add another low complexity module whenever you want additional
functionality.
The core is a board housing the processor, 60k of RAM and 4k of ROM,
with 256 bytes masked out of the address space in a decoder. I developed
this idea and later heard of the I/O space used in PDP-11 architecture.
It is 4k in the PDP-11, but it is the same idea. Why shouldn't it work
for me, too? The address space is layed out in a way that should make it
easy to eventually replace this board with separate boards with CPU and
MMU on one of them, plus one or more memory modules.
Better check to see how OS/9 level II is memory mapped. I think the
last 256 bytes
are common to all memory maps. Lets not forget the PDP-11 went from a 18
bit bus to I think a 22 bit bus.
A separate module has two 6850 ACIAs to provide two
serial interfaces,
complete with 1488/1489 drivers/receivers and COM8116 (or something like
that) programmable baud rate generators, the board even supports
split-speed operation for applications like BTX (50 baud send, 1200 baud
receive).
Nice. I like the maxim chip out that uses +5 volts and generates local
+-10 volts for RS232 buffers. This chip and 2 memory chips are wired to
FPGA prototype board to give me a minimal computer system. I am slowly
working on the PCB board with a limited front panel and a choice of
bootstraps: Front panel, serial port , EEPROM. The primary bootstrap at
the moment is the serial port as I download test programs from a PC to
test my CPU design. In time I will have a floppy disk and my system will
be stand alone.
To provide an easy way of configuring the hardware, I
created a board
with 8*8 DIP switches. The idea is to put these switches as 8 readable
bytes at a fixed address and use them to provide initialization
parameters to the ROM based monitor program. This board also came in
handy in initial testing of the processor board: I managed to squeeze a
9 byte loop into these 64 switches that made a LED on the processor
board (address decoder output) blink when running properly.
Neat.
A video board (with capabilities slightly above a
VT100) exists on paper
only, an EPROMmer on another board was partly built when time ran out.
All that stuff is still around, and I have a 19" rack mount frame that
contains the self-made power supply (with 24V ready to support 8" floppy
drives) and up to ten of my wrapped boards.
I would like to see that.
All this is based on standard components, but
intentionally incompatible
to existing buses: I wanted to force myself to build everything on my
own. Once I dig down to those pieces, I may be able to provide some
pictures. Stay tuned ;-)
Ok ... but the picture is a bit wavy :)
--
Ben Franchuk - Dawn * 12/24 bit cpu *
www.jetnet.ab.ca/users/bfranchuk/index.html