Re: Stellation Mill 6809 card clone / FPGA

[steve.howell@altium.com.au]

>
> Ok, question Steve.   If we were to use a real CPU, would we really
> need
> an FPGA for such a small amount of glue logic?  Seems like it would be
> extreme overkill.   Why would we not want to implement the 6809 in the
> FPGA
> itself?

Absolutley, putting the whole board including its CPU into the FPGA is
just touching on how most designs, if not all, will be in the future.
Which is why the FPGA and the PB04 ASIC adaptor board are in two
halves.
One of the most notable issues when implementing this stuff, is how to
hunt down bugs, and for first the first time, I found myself in awe at
the fact that Alex's IIe came to a stop when I tried to use Apple's
3.5" floppy disk controller, due to a bug in the actual processor code
itself (we are currently building a known good 65C02 core to fix this
- the freebies from the net are buggy). So what Im trying to say is,
by using a real 63C09 in PB04, we can design the support logic first
and confirm it all works correctly. Then we bring in the soft 63C09 to
replace the real one, and make sure it works correctly. Once its all
ok, the PB04 board is removed leaving the final design all inside one
weeny FPGA, hopefully with all out hair intact.

Is the FPGA we are using overkill for such a small amount of logic we
have to support? Yes. But with the 50AN costing only $9, and the 250E
only $18, its definatley worth it. It makes for a cheap and easy 2
layer board, simple to design in the FPGA space, as well as to
download, makes a nice simple bill of materials (one chip), and gives
us lots of logic to experiment with in the future - ie, you could add
Alex's Z80 soft card in there well. But more importantly, it gives us
the required logic to put all the soft tools, that once sat on the
bench, into the FPGA whilst we build the design. Tools such as logic
probes, logic analyzers, decoders, ROM images, etc.

And the FPGA's just keep getting cheaper.

Steve