Russell McMahon wrote:

> BUT
> at 300 baud it will run with essentially 100% reliability over very
> large distances on wet string (well, maybe the water needs to be
> salty :-)).

  Sounds fine to me.  RS-232 *is* after all differential - it operates
on the difference between the TXD and GND wires.  If you use *real*
RS/ EIA-232 with *real* +/- 10V drive, it should work quite a long
distance at quite high baudrates.

  Accordingly, I see just three problems with it.

  1} The single-ended protocol is quite susceptible to bias from ground
returns, so it may well need an isolated receiver (just like that
recommended for RS-422/ 485!).

  2} You must *use* RS-232, not "imaginary RS-232" with 5V CMOS logic
levels, which is really "asking" for threshold problems.

  3} Many/ most receivers include termination because - that is
specified in the standard.  Paralleling these for "multi-drop" may well
result in out-of-spec termination, in which case again it would be
deceitful to say "RS-232 is not good enough".
--
  Cheers,
        Paul B.