Ah, I had thought these were (or could be) on the same PCB.

Since there is a cable, and only one cable connecting only two pc boards-
Each end has five wires
End one connects the three needed signals (TX, RX, GND) and leaves the
fourth and fifth positions unconnected
End two shorts the fourth position to TX, the fifth position to RX.
The fourth and fifth position on each PCB is a simple pull up and pull
down, respectively.

This is probably the easiest to build cable.  If you need to keep the
cable to three wires, then a single diode going from TX to RX with a
resister all built into the cable and some code would enable you to have
three wires and still determine who is master.

The key here is that you have two (or more) identical devices which
connect.  But they connect via one cable.  That one cable can be built
exactly the same for each pair, and yet be unsymmetrical so as to
provide a deterministic decision for the other parts of the system.

-Adam

John Nall wrote:

> At 12:22 PM 6/9/2003 -0400, you wrote:
>
>> Design the PCB so that one line (TX to RX) has a pull up resister, the
>> other (RX to TX) having a pull down resister.
>
>
> But....perhaps I am missing something...but it seems to me that the two
> PCB's would have a conflict in that case.    Call them A and B.  Then
> Tx(A)
> connects to Rc(B).  So the line that A sees as Tx to Rc is the exact same
> line that B sees as Rc to Tx.  A pull up which affects Tx(A) also affects
> Rc(B) because they are connected.   Ditto with a pull down.  No??
>
> BTE, I've had some very original and creative ideas come back, but
> some of
> them seem to ignore the two primary prerequisites:  (1)  The two PCB's
> are
> identical, and are turned on with a common power supply, and (2) The
> software (OK, firmware) in them is also identical.  It does not matter
> which is master and which is slave, just so long as it is done.   It may
> not be the same each time the system is powered on (although I suspect it
> will be the same).
>
> The suggestions that say "Have A check so-and-so and then do thus and
> thus..." seem to forget that  between the time A does the check and the
> time A does something, B can also make  the check, and make the exact
> same
> assumption.  Unfortunately, there is no single instruction that says
> "check
> and change". :-)
>
> I don't want to seem ungrateful, though, and am not.  Some of the
> ideas are
> really good ones and I am very appreciative of the responses.
>
> John
>
> --
> http://www.piclist.com#nomail Going offline? Don't AutoReply us!
> email listserv@mitvma.mit.edu with SET PICList DIGEST in the body
>
>
>

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email listserv@mitvma.mit.edu with SET PICList DIGEST in the body