EEVblog Electronics Community Forum
Electronics => Beginners => Topic started by: gogoman on July 26, 2018, 12:24:55 pm
-
Hi the RS485 i'm looking at does not use a bias network, when the BUS is idle the differentials (A/B) have the identical voltage levels.
is this normal?
The RS232 is converted to RS485, when the bus is idle the signals A and B are at 0v and 5v.
the voltages swing rail to rail, I was expecting the signals to be ~1/2 of VCC(5v) with a different of 200mv.
Is this a exceptionable?
In both cases the bus has a single driver and multiple receivers
thanks
-
Yes, this is "normal". The RS485 driver has two diferencial outputs that go from 0 to 5 volts. Reciever on the other hand has to be able to opperate with 200mV difference. This way, you can make even very long bus work - f.e. 3km line of UTP is still ok.
-
If an RS485 wire pair is not driven because all nodes are in "receive" mode then the termination resistors will pull the wires to each other.
In this case noise picked up by long wires can easily be more than the threshold and the RS485 receivers may receive false data.
This is why there is normally a bias network on the wires.
Any RS485 network in which the direction is switched over a single wire pair should have these bias resistors.
RS485 should also have a 3rd GND wire to make sure the common mode voltage does not esceed the limits of -7V to +12V on the signal lines.
"10 ways to bullet proof RS485" is an (old) National Semiconductor AN and well worth reading.
https://duckduckgo.com/html?q=10+ways+to+bullet+proof+RS485 (https://duckduckgo.com/html?q=10+ways+to+bullet+proof+RS485)
Note:
A RS485 driver should be able to get reliable data from a 200mV signal, but it could be triggered by a much lower signal. I would not be surpized if a 50mV noise spike generates a false bit.