George,
Since RS-232 is a point-to-point protocol, multiple TX drivers cannot
exist together.
Yes, tri-state RS-232 drivers can be constructed (and were likely used
in your particular arrangement), but only one driver can be allowed to
be active at one time. The fact that you say the drivers (transmitters)
were "wire or'd" means that the inactive drivers would have to present a
high impedance. That is not the normal case for RS-232 drivers. The
fact that you state it took 20 man-hours to develop the hardware
indicates that these are not normal RS-232 drivers.
I repeat - RS-232 is a point to point protocol, and the hardware
reflects that situation. One cannot have two active drivers on the same
line at a time. The voltages for RS-232 are up to -25 volts for the
'mark' level, and up to +25 volts for the 'space' level (the idle
condition or logic zero). Now think about connecting two active RS-232
drivers together, one trying to send a logic '1' (-25 volts), while the
other is idling (at +25 volts). The combination of the two drivers
equals zero volts at the receiving end, and zero volts is in the RS-232
undefined area, so the voltage level should be undefined and ignored.
Further complicating the situation is that some PC serial ports and many
"almost RS-232" devices have strayed from the normal spec, and will
respond to zero volts as a logic '1' and any positive voltage as a logic
'0'. Normal RS-232 levels would ignore anything between -3 volts and +3
volts as 'noise'. But recent psuedo RS-232 receivers will respond to
that zero volt level as a logic 1 (mark) and receive it as such.
The above situation works for very short 'RS-232' lines as usually
encountered in PC and ham equipment situations. but will not work well
with compliant RS-232 receivers that will ignore levels between -3 and
+3 volts.
In summary, multiple RS-232 drivers on a single line will conflict with
one another unless special enabling circuits are built into the driver
and a means of controlling which driver has control of the line are
present - not a trivial task. In other words, do not "wire OR" multiple
RS-232 drivers together unless those drivers are specifically designed
to allow only one at a time to be active. That means a total system
control must exist, and that can only happen in specific controlled
situations.
That is why George's example took 20 hours of hardware design, but
required 100 hours of writing and debugging software. This was a
specific system design and cannot be expanded to "wire or'd" RS-232
drivers in general - that will not work.
73,
Don W3FPR
On 1/5/2015 10:43 PM, George Danner wrote:
Harry,
Years ago we used 8 devices controlled from one PC com port on one RS-232
data link.
All of the devices bridged the receive line and were wire or'd for transmit.
The devices were polled by the PC and only allowed to transmit when polled.
This worked well with no conflicts for many years.
So it is possible to use one RS232 port for multiple devices -but- the
devices must be made to play nice in all circumstances and have special
hardware interfaces.
My best memory is we spent about 20 man hours developing the hardware
solution and another 100 hours writing & debugging software. BTW this was in
DOS days!
______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[email protected]
This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html
Message delivered to [email protected]
