RS 485 network

[Ioannis]

check the data sheets of the device you will use. If the input of the RS-485 receiver puts high load on the lines then node number is limited (up to 32 for the RS-485 specs).

Newer devices like form analog devices (http://www.analog.com/en/interface/r...cts/index.html) present minimum load on the lines so up to 256 nodes can be implemented.

Ioannis

[ralfmayr]

Ah, sorry, was a misunderstanding from me,
thoght you have restrictions with the addresses...

[mychangl]

Thanks for the information. Get better understanding.
1. If i still use 75176, do you think the attach circuit will work ?

2. Do you think if should isolate the line as in the attachment if the wiring goes so long ? ( could have trouble with terminal resistance or isolation of fault )

3. Most of the articles i gone through does not mention the actual type of cable to be used except twisted pair. In order to determine terminal resistor value, i need to know the cable impedance. What if i do not have the information ? Will 120ohm resistor still helps. Or it is better to use 120 ohm with shorter distant cable but i can add isolator to extend the communication line ?

[mychangl]

Ooops, i missed out the diagram..

[Ioannis]

How long is "long"?

Do you really need isolation?

As for cable, I always use UTP cable. It is twisted and has characteristic impedance very close to 120 ohms.

If you don't have a copy of the Serial Port Complete by Jan Axelson, you must get it soon.

Lastly for short distances and low baud rate (less than 9600), termination is not really necessary.

Ioannis

[mychangl]

I will use baud rate of 9600. Total distance is estimated 500 meters since i cannot use star topology.

You use UTP, it is unshield - is it okay ? how about other noise ?

I am thinking that, may be isolation can localize the fault in case it happen.

Regards

[Ioannis]

OK. 500 meter is almost half the max distance of a RS-485 network. So termination is better to be used. The UTP is indeed unshielded. If there are no very strong interferences I see no reason to use other cable. The twisted pair has the property of self cancelate any noise presented on both the wires at the same time.

The opto isolation is really necessary if the two nodes are at different voltage referenced to (common) ground. If this is the case the do use isolation. Otherwise connect directly and the ground with 100 ohm resistor to be safe.

For the role of this resistor I direct you again to Axelson's book.

Ioannis