DirectNet RS-422 Multidrop Cabling

zmanvortex

Member
Join Date
Apr 2002
Location
Nebraska
Posts
216
Hi everyone. I am attempting to set up a DirectNet RS-422 multidrop network. I will be using an AutomationDirect DL-250 port 2 as the network master and up to 8 slaves consisting of a mixture of DL-06's and DL-240 processors depending upon the customer's specs. I understand how the master/slave network and instructions work. I also know how to make the cable. What I am looking for is a way for our installers to add another station (slave) and be able to just plug in another cable off of the last slave and move a terminator connector to the most recently added slave. I am looking for something similar to what Allen-Bradley did with DeviceNet cabling where you just screw on another cable from the device into a T or Y connector on the network cable.
 
It isn't quite that simple. You can multidrop an RS-485 network from PLC to PLC, but you need some RS-232/RS-485 converters. The serial ports on the 240 and DL-05 are all RS-232. One of the ports on the 250 and the DL-06 is both RS-232 and RS-485, so those you can just string together. The AutomationDirect.com manual has some diagrams.
 
Hi Tom, Thanks for the reply. I have discovered that this project will use all DL-250's and DL-06's. What I would like is for the installers to be able to add another station (slave) simply by adding another cable from a T connector near the previous station to another T on the new station. I will make the cables and send them to an installer to just plug in to T's and the new plc. The following is a simple drawing of what I am thinking.
[attachment]

multidrop.jpg
 
You can wire it that way, no problem. You can either use terminal strips (my preferred technique) to daisy chain them together or any kind of connector that is convenient. Remember, though, when you add stations the terminating resistors have to be at the end station.
 
Tee connections are not allowed with RS-422 networking. You tend to get reflection from the drops off of the main line. The proper way to wire RS-422 (or RS-485) is to daisy chain.

This is not to say that it can't or hasn't been done. There are people that do Tee networks and have them work fine. But it can be a problem.

The reason it can be a problem is that if the length of the Tee "drop" sections get over a certain length, the transmitters start seeing them as unterminated loads. Your impedence gets out of whack and then you start having faulty communications.

See this link for details: http://www.systech.com/support/nps1000/doc/nps1000/html/nps1kb-02.htm

First section, third paragraph, in bold type.
 
You're right, Icky, that indiscrimate tee connections of long length can cause problems with RS-485 (I know some ADC documentation calls it RS422, but that's not technically correct in my opinion.) However, we have done it with branches of several hundred feet successfully. Sometimes you have to have play with the terminating resistors, as you indicate. However, theoretically every connection to an RS-485 multidrop network is a tee - just very very short in some cases!
 
As I understand it, RS-422 has two pairs of wires. One pair comes from transmit at the master station and goes to receive at all the slave stations. The other pair comes from the transmits of all the slave stations to the receive of the master station. Of course only one slave is allowed to transmit at a time in response to a request from the master station. RS-485 is essentially the same chips with the transmit and receive pins tied together on one pair. This can be master-slave (like DirectNet) or multi master (like AB's DH-485 token-passing) depending on the protocol.
 
Thanks for the replies guys. I have set up a small network of 3 plc's on my desk successfully. It seems to be working fine, I have used terminal blocks next to each plc to create the network connections and I think I will use this theory in the real design. It seems simple enough to mount 4 or 5 terminal blocks on the DIN rail next to the plc in each station's panel.
 
Terminal blocks are another gotcha in the communications game. Anytime you use terminal blocks you are in essence adding small antenna to the line. Antennas transmit and receive noise. Noise causes poor line quality. Bottom line is, if you plan on using terminal blocks, put them as far away from other noise producers as possible and be prepared to remove them if necessary.

Like I said earlier, it has been done and can be done, but it can cause issues.

I get calls every day where the main culprit is noise or a poorly terminated line. The other main cause is using the wrong type of wire. IE. "It's shielded Belden, so it's good wire." Then I go into the explanation of impedance, reflection, and how radio signals work.

Tom, you are right. It's when you happen to get the right length of cabling that your impedance goes down the tubes. Most times you can make it work by playing with the termination resistors, sometimes it just won't.

Bernie, you are also correct. AD plc's do not support automatic token passing, but it can be done.
 

Similar Topics

I found problem in The PLC DL205-1 is using the protocol K-sequence and I'm try changing to DirectNet protocol that i can config to get success...
Replies
0
Views
3,093
We developed a DirectNet device driver for the the Koyo/Automation Direct PLCs, and we can access the PLC memory. Is there any information on how...
Replies
14
Views
9,288
Does anyone have any experience with the direct net protocol by direct logics. I would like to know if anyone has used the directNet protocol...
Replies
2
Views
5,155
MELSEC A RS422 (25pin) <> iQR C24 Serial RS232 (9pin) - Simple PLC Communication Communication from a MELSEC-A Series CPU (RS-422 – DB 25 Pin)...
Replies
4
Views
1,042
I have an application I'm considering where I need a device that's generating a master encoder output signal to drive two slave devices at the...
Replies
4
Views
2,533
Back
Top Bottom