RS-422 source termination

[jeduffy]

I've been looking to use RS-422 for communication in a project, and I'm wondering if there is any reason to place resistors inline with the driver?
Some circuits ( https://www.pulseresearchlab.com/collections/rs422-output/products/prl-425rs?variant=29205881864 (the pair of 62 ohm resistors in this, not the 100 ohm resistor)) seem to do so, but I'm not sure if they refer to actual discrete resistors or just the natural source impedance.  The only thing I could think is limiting slew rate to reduce emissions, but I'm interested if that actually works, or if there's another reason?

[Doctorandus_P]

The "natural source impedance" you are refering to is a property of the cable.
To avoid reflections in the cable real resistors are placed at both ends of the cable.

With shared systems such as RS485 these restors are (should be) at the physical ends of the cable. With a single wire the output impedance of the transmitter should be equal to the cable impedance.
https://en.wikipedia.org/wiki/Electrical_termination

[apblog]

Read this and be enlightened:

http://www.ti.com/lit/an/snla034b/snla034b.pdf

The easiest and arguably the best way to terminate rs422 for a point to point connection is with a parallel resistor at the end that matches your cable impedance.  No source resistors required.

[T3sl4co1l]

The transmitter itself has a dynamic impedance in the 10s of ohms range; this is intended to drive the line to nearly the supply voltage, ensuring maximum signal at the far end.  The far end then gets terminated, and somewhat less than the supply voltage is available there (the loss being due only to line losses).

With source termination, the impedance of the line is the only load; this is good for saving power over short distances, but the risetime falls over more rapidly with distance, the line isn't suitable for multidrop connection (intermediate receivers will get a stepped edge), and the line also won't return to zero (which makes RS-485 problematic -- hence this is more of an RS-422 option, i.e., always-on transmitters).

Tim

[Cicero]

I have actually seen some implementations place low value series resistors in line to 'deal' with certain fault conditions, with the logic being that hopefully the resistors burn out instead of the driver. 

[T3sl4co1l]

Such is normal design for small devices, lamps and such.  A 4.7 or 10 ohm resistor limits inrush current without affecting efficiency much, and is fusible.

Another example, polyfuses usually at higher voltages (>30V?).  The Rmin spec is too low to be safe (i.e., Vin(max) / Rmin > Imax), so you may need to add external resistance to prevent from blowing it out (in which case it's a monofuse, not a polyfuse ).

Tim