Reply by Paul Keinanen May 29, 20092009-05-29
On Thu, 28 May 2009 12:27:20 -0700 (PDT), Mike Silva
<snarflemike@yahoo.com> wrote:

>We're looking at a design that would have 1 RS485 master and 2 slaves, >with a maximum cable length of 50 meters (25m is more likely), and >we'd need to run a data rate of 460 kbaud. Since the RS485 spec talks >about data rates of 10 Mbaud and cable lengths of 1200 meters (not at >the same time) it would seem that we're well within safe-and-sane >territory, but I was hoping to hear from anybody who's run similar >data rates over similar distances, and how it went. Thanks.
Apart from avoiding slew rate limited transceivers (typically max 250 kbit/s) there should not be too much problems. Profibus-DP (which is basically RS-485 using asynchronous characters) runs without problems up to 1.5 Mbit/s on various hardware (some of which might be of dubious quality :-). However, running Profibus-DP at 12 Mbit/s (for up to 100 m) requires quite a lot attention to detail. For instance the stub length to each node must be as short as possible, i.e. the transceiver chip must be installed very close to the socket with very short PCB tracks and the passing through bus cable may require some series inductance at both sides of the plug, to compensate for the stub stray capacitance. Paul
Reply by Dombo May 28, 20092009-05-28
Mike Silva schreef:
> We're looking at a design that would have 1 RS485 master and 2 slaves, > with a maximum cable length of 50 meters (25m is more likely), and > we'd need to run a data rate of 460 kbaud. Since the RS485 spec talks > about data rates of 10 Mbaud and cable lengths of 1200 meters (not at > the same time) it would seem that we're well within safe-and-sane > territory, but I was hoping to hear from anybody who's run similar > data rates over similar distances, and how it went. Thanks.
Considering that the DMX protocol (used in stage lighting applications to control up to 512 lights, in use for over twenty years) runs at 250kbit/s over RS-485 under conditions that are typically more challenging than what you are describing, 460kbit/s doesn't appear to be too far fetched for your application.
Reply by John B May 28, 20092009-05-28
On 28/05/2009 Mike Silva wrote:

> We're looking at a design that would have 1 RS485 master and 2 slaves, > with a maximum cable length of 50 meters (25m is more likely), and > we'd need to run a data rate of 460 kbaud. Since the RS485 spec talks > about data rates of 10 Mbaud and cable lengths of 1200 meters (not at > the same time) it would seem that we're well within safe-and-sane > territory, but I was hoping to hear from anybody who's run similar > data rates over similar distances, and how it went. Thanks. > > Mike
DMX is RS485 running at 250kBaud and it's used (and abused) in the entertainments business all over the world. As long as you remember to terminate both ends of the run and don't have any stubs you'll be fine. If your comms is bidirectional add a pull-up and pull-down at one point only to force the space state when the line floats. This app note from National is very useful: http://www.national.com/apnotes/RS-485.html -- John B
Reply by Mike Silva May 28, 20092009-05-28
We're looking at a design that would have 1 RS485 master and 2 slaves,
with a maximum cable length of 50 meters (25m is more likely), and
we'd need to run a data rate of 460 kbaud.  Since the RS485 spec talks
about data rates of 10 Mbaud and cable lengths of 1200 meters (not at
the same time) it would seem that we're well within safe-and-sane
territory, but I was hoping to hear from anybody who's run similar
data rates over similar distances, and how it went.  Thanks.

Mike