Tony Duell wrote:
I think once
the 2901 "fell from grace" (?), this became a thing
of the past. I've designed two processors "from scratch" (TTL
Maybe, althoguh i never cared much for the 2901...
2901 really only makes sense for
a "classic" CPU design.
By the time you put all of the parts together to get a complete
CPU, it is quite a large solution. Some parts (29116, etc.)
could help a little but it still was big.
The 29116 never really seemd to be a classic bit-slice part. You couldn't
easily cascade thjem to an arbitrary data path width.
It was targeted at folks who were buying 4 2901's, 2911, etc.
It saved some board space and wiring...
[FPGA's]
I think it
depends on how big the item that you are designing
will end up -- and, if it stays a "one of a kind" design or if
you ever decide to make several of them.
Oh, exactly. If the thing is going into production, you want to use an
FPGA. And you want to use a PCB, not hand-wiring. This does not mean that
boards of TTL, hand-wired, are not more convenient for one-off prototypes.
I like prototyping in foil. It seems like I always end up needing
a second (or third) copy of a design for some reason. And, having
to hand wire a second (or third!) is really not fun.
And, for $WORK, I learned that no matter WHAT the client says,
he'll ALWAYS want a second prototype for some reason. :-(
This is an example of something that really annoys me.
Just because a
particular solution )here, the FPGA) is better in one case, it doesn't
mean it's batter in _all_ cases. Yet an awful lot of people use such
reasoning.
Of course! The first 50+ reading machines [1] that were built were
all wirewrapped. The design was still quite fluid (you kept records
of each individual machine since they were all slightly different)
so you wanted to be able to make a change on the fly -- as well as
IN THE FIELD! Sure, a nice 14" square (? Nova 3) PCB would LOOK
really professional. But, when you're stuck in Romney WV on a
Saturday night trying to hack a design change into it, you sure
as hell don't want to hang around until noon on Monday for something
to get air freighted in!
OTOH, I've watched hand built prototypes catch fire [2] during
testing and the realization that you've got to start all over
is more distressing than the fact that you've toasted a bunch
of expensive components :-/
[1] Kurzweil Reading Machine
[2] You quickly learn the value of "dead time" when switching
100A circuits! :-/