[Coco] Re: Any reason to put a 6309 in a Coco2?
jdaggett at gate.net
jdaggett at gate.net
Fri Jan 2 12:14:10 EST 2004
On 2 Jan 2004 at 9:44, Brad Grier wrote:
> Torsten Dittel wrote:
> > Hi Brad,
> > Would be better to know how the 6809 is "hardwired"... this would
> > allow an exact prognosis of what is happening with the illegal
> > opcodes...
> You would think some documentation had to exist somewhere for these
> opcodes to have ever been used in the first place.
There would have been a very good docment and a collection of weekly status
reports for the whole project. Probably now that exists is somewhere in an archive is
the final design report, MC qualification report, a synopsis of the changes and the
mask sets. Then some of that may now be ready to become nonexistant.
I had some free time and I just looked at the memory map and tried to do lagic
diagram of the Address mode decode section. Surprisingly from the Boolean
equations it only required a total of 40 gates ranging from 5 input totwo input gates.
The greatest propogation delay is four gates or about 20 to 30 nS max.
Also did the same thing for the logic for the postbyte decode. It takes less gates.
>From that you get three bytes, (opcode, mode and postbyte) that can control a very
complex and massive state machine. Unused opcodes were more than likely
directed to a common state and exit out to next instruction fetch. More likely the so
called "Undocumented Opcodes" are inputs to the instruction decode state machine
and were never removed once decided not to use them for fear of disrrupting or a
massive redesigning of the instruction decode state machine. Byte Magizine article
is a good article. The one picture of this massive blueprint hanging on the wall with
gates and wiring diagram was typical of time frame when I worked at Motorola.
Having worked for Motorola, I have a bit of insight as to the process of how the
6809 may have been designed. No I did not work on that project. I had projects that
I worked on that required interfacing with IC designers back in the early 80's. At one
time somewhare in Motorola was a very large breadboard with NMOS gates in DIP
packages wire wrapped to test the logic design.
Once convinced the design was solid, then they went for first metal. You have to
remember VHDL did not become a standard until 1987. The 6809 project started in
1977. This was all before the days of HDL and CAD were commonly excepted. The
6809 is LSI chip and by today's standards not a real dense chip.
> > Much better now with the "toggle buttons", any feedback from the
> > Apple users?
> No feedback on the newest version. The only question now is whether
> the labels fit all the controls on the Mac (font differences). In any
> case, Apple users should definitely have access to all the buttons
> > regards,
> > Torsten
> Coco mailing list
> Coco at maltedmedia.com
More information about the Coco