So that means NO, right? (see the AND?) If no such thing exists, why worry about it? Now, if you have access to Spartan-II, which I like very much, then you have options.



Back in '78 or so, a HAM friend of mine produced a circuit that generated Morse code from a bipolar (therefore very small) PROM and a counter, to generate the periodically recurring signature for his club's repeater. It produced very precise code at a very reasonable rate. Not being interested in Morse code, myself, I didn't make any other observation.

For outside help with the problem you're working, I'd be looking at the various RLL encoder/decoder SPLD's published out on the www, for a bit of "spiritual guidance" about how one might encode/decode a code with variable character length. I'm not, in any way, criticizing your approach, particularly as the objective is to learn, and I'm sure you're doing that, but I'd certainly recommend examining "prior art" related to what you're doing.

The tristate bus within an FPGA is apparently a thing of the past. It means that whenever the notion of a common bus pops up, there will be a huge timing penalty due to the required concatenation of dozens of logic cells in order to form a single mux wide enough and deep enough to support the requirement. I've seen some people deal with this with an ultra-high-speed serial bus, common to all the parties. That would be fine if you didn't mind having to divide the costly ultra-high bandwidth by the bus width. On devices with lots of pins, the tristate buffers can be external, i.e. each module talking on the common bus can be externally connected. That, of course, has its own price. You know what they say ... "where there's a will, there's mourners ..."

RE