Eia 485?
- Thread starter MD88
- Start date
- Status
gar
Senior Member
- Location
- Ann Arbor, Michigan
- Occupation
- EE
090122-1808 EST
MD88:
A 485 system is a balanced line half-duplex type. Probably means you may require 3 wires. Two signal wires, preferably twisted, and a common reference path connection.
In a 485 system you can only send data in one direction at a time, and there has to be some means in the communication protocol to determine who is the sender. In a 422 system between two points you can have simultaneous transmission in both directions.
A twisted pair for the data lines is important to reduce inductive noise coupling. In addition a shielded cable would be used to reduce capacitively coupled noise. If the cable length is short anything may work. If it is long you want low capacitance cable. CAT-5E is twisted if you use the correct wires, and it is low capacitance.
A major problem that is usually not understood is the ground reference conductor, and difference voltages between different ports connected to the 485 or 422 cable. A system using electrical isolation in or external to the 485 (422) interface can greatly reduce problems from these ground voltage problems.
You need to tell us your maximum baud rate (bits per second), cable length, how many devices connected to the cable, the noise environment, ground voltage problems, whether any of the equipment includes isolation, etc.
.
MD88:
A 485 system is a balanced line half-duplex type. Probably means you may require 3 wires. Two signal wires, preferably twisted, and a common reference path connection.
In a 485 system you can only send data in one direction at a time, and there has to be some means in the communication protocol to determine who is the sender. In a 422 system between two points you can have simultaneous transmission in both directions.
A twisted pair for the data lines is important to reduce inductive noise coupling. In addition a shielded cable would be used to reduce capacitively coupled noise. If the cable length is short anything may work. If it is long you want low capacitance cable. CAT-5E is twisted if you use the correct wires, and it is low capacitance.
A major problem that is usually not understood is the ground reference conductor, and difference voltages between different ports connected to the 485 or 422 cable. A system using electrical isolation in or external to the 485 (422) interface can greatly reduce problems from these ground voltage problems.
You need to tell us your maximum baud rate (bits per second), cable length, how many devices connected to the cable, the noise environment, ground voltage problems, whether any of the equipment includes isolation, etc.
.
zbang
Senior Member
- Location
- Roughly 5346 miles from Earls Court
Picking nits
Picking nits
Just a minor nit, baud rate isn't necessarily the bits-per-second rate. Happens that they're the same for rs-232/422/485 and similar, but they are quite different when you're talking about things like modems or twister-pair ethernet. (For instance, a v.34 modem can run up to 33.8kbps but is operating at <3500 baud).
Just thought you'd like to know :grin:.
Picking nits
Just a minor nit, baud rate isn't necessarily the bits-per-second rate. Happens that they're the same for rs-232/422/485 and similar, but they are quite different when you're talking about things like modems or twister-pair ethernet. (For instance, a v.34 modem can run up to 33.8kbps but is operating at <3500 baud).
Just thought you'd like to know :grin:.
gar
Senior Member
- Location
- Ann Arbor, Michigan
- Occupation
- EE
090122-2047 EST
As far as the signal on the 485 cable is concerned baud rate and bits per second are identical. If you want to talk about the information transmitted as bits per second, then there can be a great disparity between the two. If there is a lot of redundancy in the information sent, or protocol garbage, then the information bit rate will be much lower than the baud rate.
If I use an asynchronous connection with 1 start bit, 8 data bits, 1 parity bit, and 1 stop bit followed with no delay by the next start bit, then the ratio of data bits to bits in the 485 channel is 8/11.
If I have an ensemble of four messages, then they can be represented by the four combinations of two bits. Thus, fundamentally 2 bits worth of information. Encode these messages in English as:
00 = This is message A.
01 = This is message B.
10 = This is message C.
11 = This is message D.
Each of the messages contains 18 characters. If 8 bits are used to define a character, then each message requires 18*11 bits in the 485 channel. Thus, my actual information data rate is (2/198)*baud rate. Note rate attached to baud is incorrect but is more natural and is commonly used. With 115.2 kbaud I can only send the above messages at the rate of 115.2/99 thousand messages per second. 1 % efficiency. The way these messages are composed does not do much for improving the reliability of signal transmission vs just sending XY where X is 0 or 1 and Y is 0 or 1. This was only to illustrate the difference between information rate and the signaling rate.
.
As far as the signal on the 485 cable is concerned baud rate and bits per second are identical. If you want to talk about the information transmitted as bits per second, then there can be a great disparity between the two. If there is a lot of redundancy in the information sent, or protocol garbage, then the information bit rate will be much lower than the baud rate.
If I use an asynchronous connection with 1 start bit, 8 data bits, 1 parity bit, and 1 stop bit followed with no delay by the next start bit, then the ratio of data bits to bits in the 485 channel is 8/11.
If I have an ensemble of four messages, then they can be represented by the four combinations of two bits. Thus, fundamentally 2 bits worth of information. Encode these messages in English as:
00 = This is message A.
01 = This is message B.
10 = This is message C.
11 = This is message D.
Each of the messages contains 18 characters. If 8 bits are used to define a character, then each message requires 18*11 bits in the 485 channel. Thus, my actual information data rate is (2/198)*baud rate. Note rate attached to baud is incorrect but is more natural and is commonly used. With 115.2 kbaud I can only send the above messages at the rate of 115.2/99 thousand messages per second. 1 % efficiency. The way these messages are composed does not do much for improving the reliability of signal transmission vs just sending XY where X is 0 or 1 and Y is 0 or 1. This was only to illustrate the difference between information rate and the signaling rate.
.
ELA
Senior Member
- Occupation
- Electrical Test Engineer
Gar,
The main backbone in a large piece of industrial equip. we designed was RS485 full duplex multi-drop with differential drivers. This used four wires. One set for transmit and one set for receive (with respect to the master).
I totally agree with the need for isolation to prevent common mode voltage issues. We learned that the hard way.
pbeasley
Senior Member
- Location
- Orange County, CA (for now)
There is alot of really good information posted here. As mentioned above, CAT5e will frequently work. 99% of the applications I have encountered use Belden 9729 or equivalent. It is 2 Individually Shielded 24 AWG twisted pairs with a 24AWG drain wire and a nominal impedance of 100ohms.
gar
Senior Member
- Location
- Ann Arbor, Michigan
- Occupation
- EE
090123-1300 EST
Your suggestion of 9729 is good.
Following is the data sheet:
https://edeskv2.belden.com/Products/#s=9729&r=0
then pick 9729.
http://www.farnell.com/datasheets/47776.pdf
this gets data sheet directly and faster.
.
Your suggestion of 9729 is good.
Following is the data sheet:
https://edeskv2.belden.com/Products/#s=9729&r=0
then pick 9729.
http://www.farnell.com/datasheets/47776.pdf
this gets data sheet directly and faster.
.
- Status