12 May
2004
12 May
'04
9:31 p.m.
On Thu, 13 May 2004, Jochen Kunz wrote:
Hmmm. Replicating the RK611 in an FPGA is an
interresting idea. - If you
can get complete shematics and there are no funky things like an AMD bit
slice CPU in it.
Nope, just logic.
There isn't one if I use the original RK611
controler. That is an other
issue: If you emulate the entire controler, you have to make the
emulation good enough for all operating systems. It may work for a
RK611, but what about a UDA50 with its AMD bit slice CPU or a controler
that has a T11 with some non-trivial firmware on it. You can't stick
this in a (cheap, small) FPGA.
Well, since I wrote the SimH driver for RK611 (the "hk" driver),
I can safely say that emulating this is simple, if the design is
microcontroller based.
Since we have been bithin' and such about this since forever, why not
define a board spec?
The Unibus transceivers is indeed an issue- we dont have heaps of
those laying around. On the other hand, their exact function and
specs are known (see the Peripherals handbook, 1972-ish or so),
and we should be able to re-implement them in discrete logic,
or bite the bullet and do a pin-compatible replacement, and
have 1,000 made.
The board itself (not a funny thing, just large) could be a
general microcontroller design, with the basic stuff (uNibus
interfacing, CSR, BR and such logic) being standard, including
a general flash-programmable microcrontroller setup (CPU, RAM,
FLASHROM, I/O). The rest is "up to the function", for example
an IDE controller with 2.5" laptop drive, CF sockets, or what
not.
--f