>
> On Fri, 13 Aug 1999, Byron A Jeff wrote:
>
> > > On Thu, 12 Aug 1999, Byron A Jeff wrote:
> > >
> > > > >
> > > > > How about CSMA/CD?
> > > >
> > > > It's not clear if the physical layer of RS-485 allows for multiple trans
mitters
> > > > or what happens if multiple transmitters transmit different bits.
> > >
> > > If you are doing half-duplex RS-485, you just tie the UART tx line
> > > (inverted) to the RS-485 driver's R/W pin and tie the tx pin on the driver
> > > to ground. That way, the driver will only transmit when it wants to send a
> > > 0. A collision is defined when one transmitter is sending a 0 and another
> > > transmitter wants the line to be floating (i.e. it is sending a 1). There
> > > is an appnote on this somewhere (National Semi i think). Collisions can be
> > > detected in hardware with by: UART framing error, UART missing stop bit,
> > > UART parity error or in SW with CRC errors in the frames.
> >
> > Wire-OR. That's an interesting idea. I'd be interested in the latency betwee
n
> > my scheme and true collisions.
>
> That's what I (almost) did. In my case, the physical layer also performed
> arbitration (some of the datalink layer trickled down). After the last
> byte of a message was transmitted, the active master would enter an
> arbitration mode. This consisted of extending the stop bit for a
> pre-defined interval and then asserting an 'arbitration edge'. All
> potential masters monitored this edge with a hardware input-capture pin.
> After the edge was initiated, the active master would relinquish control
> by tri-stating its transmitter. The RS-485 termination resistors, which
> are necessary for reliable long-distance communications, were configured
> such that they would also passively maintain the DC level of the RS-485
> differential pair in the 'active low' state or the state to which the
> arbitration edge drove the bus.

>
> Now each master was preassigned an ID. This ID was used to determine an
> 'arbitration slot' or a time period that the masters would wait after the
> arbitration edge to attempt arbitration. An output compare module was
> programmed to generate the "I want the bus" edge. Finally, a master
> ascertained whether it had won arbitration by comparing when it sent the
> bus request to when the bus really went active. If the bus went active due
> to his request, then he was the winner. If the bus went active before his
> request then another master won.
>
> The comm rate was 38.4k baud, the arbitration slots were about 20uS, there
> were upto 32 nodes. This schem worked reliably for more than 300 meters of
> copper and 1500 meters of fiber. If it weren't for propogation delays
> through opto-couplers (the RS485 bus was optically isolated) the distances
> could be extended even more...

Scott, I'm unclear on one thing. This seems to be almost the same scheme I
came up with originally (which makes me happy because now I know it'll work).
But how do the UARTS react to channel activity? Also how did you prevent
starvation? It seems the lower ID'ed masters could swamp the channel. That's
why I went to a primary/secondary ID where after successfull transmission
a station went to its secondary ID until either it got the slot or until
no other primaries wanted to transmit.

BAJ
>