RS485 on multiconductor with other stuff

[Joel Moore]

Is it a bad idea to run RS485 communications over a multiconductor cable
that has other signals on it?

For example, can I run a 12-conductor (twisted pair) cable that has RS485
communications on one pair while the other 5 pairs are connected to
miscellaneous other I/O?

If it is a problem, is it a problem for the RS485 communications or for the
other I/O (i.e. crosstalk from the RS485 triggering inputs)?

Thanks for any advice,

Joel Moore

 

[Rene Tschaggelar]

Joel Moore wrote:

Is it a bad idea to run RS485 communications over a multiconductor cable
that has other signals on it?

For example, can I run a 12-conductor (twisted pair) cable that has RS485
communications on one pair while the other 5 pairs are connected to
miscellaneous other I/O?

If it is a problem, is it a problem for the RS485 communications or for the
other I/O (i.e. crosstalk from the RS485 triggering inputs)?

RS485 as differential signal requires twisted pair.
What else would you offer ?
A standalone twisted pair ?

Rene
--
Ing.Buero R.Tschaggelar - http://www.ibrtses.com
& commercial newsgroups - http://www.talkto.net

 

[Joel Moore]

Rene Tschaggelar <none@none.net> wrote in
news:4075c1cc$0$709$5402220f@news.sunrise.ch:

Joel Moore wrote:

Is it a bad idea to run RS485 communications over a multiconductor
cable that has other signals on it?

For example, can I run a 12-conductor (twisted pair) cable that has
RS485 communications on one pair while the other 5 pairs are
connected to miscellaneous other I/O?

If it is a problem, is it a problem for the RS485 communications or
for the other I/O (i.e. crosstalk from the RS485 triggering inputs)?


RS485 as differential signal requires twisted pair.
What else would you offer ?
A standalone twisted pair ?

Rene

If I had the RS485 separate from the rest of the I/O one of the cables (or
both) would be shielded so they would be protected from interfering with
each other. By putting everything on one cable there is no shielding
between the RS485 and the other wires.

 

[Terry Given]

"Joel Moore" <asdadsasd@asdaadad.com> wrote in message
news:Xns94C5B91056FCAasdsadfgasdgadsa@130.81.64.196...

Rene Tschaggelar <none@none.net> wrote in
news:4075c1cc$0$709$5402220f@news.sunrise.ch:

Joel Moore wrote:

Is it a bad idea to run RS485 communications over a multiconductor
cable that has other signals on it?

For example, can I run a 12-conductor (twisted pair) cable that has
RS485 communications on one pair while the other 5 pairs are
connected to miscellaneous other I/O?

If it is a problem, is it a problem for the RS485 communications or
for the other I/O (i.e. crosstalk from the RS485 triggering inputs)?


RS485 as differential signal requires twisted pair.
What else would you offer ?
A standalone twisted pair ?

Rene

If I had the RS485 separate from the rest of the I/O one of the cables (or
both) would be shielded so they would be protected from interfering with
each other. By putting everything on one cable there is no shielding
between the RS485 and the other wires.

you'll suffer 2 types of "crosstalk" - magnetically coupled and capacitively
coupled. the twist pitch can be selected to minimise magnetic crosstalk
(tighter twisting is better - cat5 cable is about 1 twist per cm, is cheap
and readily available as stp and utp) and of course you will want to ensure
the RS485 line is correctly terminated (stp is about 105 Ohms IIRC) AT BOTH
ENDS to prevent reflections/ringing.

Make sure the RS485 transceiver has the slowest edge rise- and fall-times
that you can tolerate - dont use a 10Mbps part for 9600bps! This reduces
both magnetically and capacitively coupled noise at the source, which is the
best place. Also, the edge rate is pretty much what governs the electrical
length of the cable, not the baud rate. High edge rates mean multidrop
systems must be closer together to get away with sloppy termination (R at
either end of cable)

capacitive coupling then needs to be "screened". if the cable allows it, you
can use adjacent conductors (connected to 0V) as DIY shields; perhaps giving
each signal its own twisted pair, and an 0V wire. put the RS485 pair as far
away as possible from your most sensitive IO lines.

In the hardware at the other end of the cable, put as much capacitance as
you can allow on your IO lines (ie the biggest filter) - this will directly
attenuate capacitively coupled noise. common-mode chokes for individual
twisted pairs can also help. make sure your power supplies are well
filtered.