Omron Ethernet/IP comms probs via EIP21 card

[kaiser_will 65]

Has anyone used the new CS1W-EIP21 Ethernet/IP interface card? We are using the card to communicate, via Ethernet/IP, with all of the I/O devices. As of now, there is no communicating going on. I have used an Ethernet packet analyzer to capture communication events comparing how the Omron communicates compared to an Allen-Bradley CompactLogix PLC (our typical solution). It appears that the A-B PLC talks broadcast, and the Omron PLC has very little traffic (probably limited broadcast traffic). I have been working with Omron's tech support on a daily basis and they appear to be snowed on the issue. We are an integrator, our customer required Omron controls, and our hardware supplier said "Omron can talk Ethernet/IP to the I/O devices you have identified".

[PMCR 66]

kaiser_will I have used Omron CJ EIP21 module to talk to some non-Omron devices: Yaskawa inverter Wago 750 series IO P&F Ident RFID controller In each case, I needed to modify the .eds file. From what I have seen, Rockwell does not use the .eds file. You simply fill in a few parameters (size and instance of produced, consumed, configuration) connections. So, many of the 3rd party .eds file have not been properly tested. In 1 case, the .eds file was wrong (produced and consumed assembly numbers were backwards). Usually it is simply a case of getting the correct number of produced and consumed correct, based on the product configuration. I another case, the Ethernet/IP interface for the product needed to be enabled. There are only 2 cases that I know of the Omron EIP modules cannot do: 1. Odd number of produced / consumed bytes. Only 1 product that I have seen has used this. 2. Configuration supported in the Forward Open. PMCR

[Berti Baker 7]

Hi, I though it was launched for PLC to PLC comms, but which are the devices you are trying to talk to? Bertie.

[PMCR 66]

Bertie The target of the product is PLC to PLC comms, using the EIP Network configurator and Tag Data Links. This same function can be applied to other IO devices such as RFID controllers, vision systems, robotic controllers, inverters, IO blocks, hydraulic / pneumatic valve manfolds, etc. Most any product that conforms to the CIP / EIP specifications will work with the EIP21 or CJ2 w/EIP built in. PMCR Edited 23 Sep 2008 by PMCR

[René Heijma 1]

The target for this product is really PLC to PLC comms. Reason is that device vendors make a mess of their EDS-files and when connecting it to RA fall back to the generic device. So RA has no problem. But Omron keeps to the standard as defined by the ODVA. That others create a mess omron can't help. it is even so that the offical policy from Japan that there is no support for third-party as because of this mess. Otherwise they would be asked to clean up for each and every vendor that likes to connect to their Ethernet/IP scanner.

[scottmurphy 1]

So why don't Omron make a generic profile available within the configurator? That Rockwell have done it makes sense to follow, otherwise everyone that uses ethernet/IP products with an Omron master is going top have to modify the eds files to suit, if they are wrong that is.

[PMCR 66]

Scott I am confused by your statement about negating multicasting with a managed switch. Multicasting is accomplished through IGMP (Internet Group Management Protocol). An UNmanaged switch, which cannot perform IGMP snooping to determine which devices are part of a group, essentially functions like a hub for multicast packets. A managed switch that supports IGMP snooping continues to function like a switch even in the case of multicast packets. Thus EIP installations should be using managed switches. To answer your question about Omron, this is a double edged sword. Omron is following the ODVA specification. The fact that RA allows the generic mode, and many manufacturers do not test their eds files, is not really a compelling reason for others to follow this model. In my opinion, RA and other vendors should make more of an effort to have correct eds files to be ODVA / Ethernet/IP compliant devices. I understand from a user perspective generic mode is 'easy' but 'easy' is not always 'best'.

[scottmurphy 1]

Again, I agree, but it seems that this is not happening. The last lot of VSD's I did on Ethernet/IP use the Rockwell generic profile as an example in their configuration manual, when one comes to use another master, it does not become as easy.

[René Heijma 1]

About multicasting. Omron knows two settings for its consuming connection. Settable by the user. It can be of course Multicast or Point-to-Point. Multicast uses IGMP so to make good use of it you need a switch that can handle IGMP thus a managed switch. So if you have multiple devices that consume from one you set up the multicast, if you have a managed switch. No managed switch then use point-to-point. You will have multiple connections but the multicast will not be all over the placce. About the Generic device. I have no objections against it but an EDS-file makes life so much easier. It shows the connections that you can choose from and the proper sizes. Also it supplies all information of the identity object. This is needed to identify the device when aconnection is established. If you have no EDS-file you need to get all these parameters from documentation. Ever tried that. Been there, done that. No thanks Further you have all parameters at hand to configure the device itself. And to help those out that think a Generic EDS-file is it. Maybe I will have a look creating a Generic EDS-file where you can fill in the Assembly objects and the sizes. Not a big deal.

[TheDr 0]

We are currently testing the CJ1W-EIP21 with the Wago 750-841 and are having some problems. The PLC can turn on outputs on the wago unit but we do not see any inputs on the wago from the PLC. Any ideas? Would it be possible to get a copy of the 750 eds file you used to compare it to the one we had to change? Thanks

[PMCR 66]

see attached. Wago_with_Parameters_JCH.zip

[René Heijma 1]

By any change a logfile of the ethernet traffic available? Then we can have a peek to see what is going on. Is the NS-led on the Wago turning on RED once in a while?

[TheDr 0]

Thanks for your post. We still can not get the unit to work. What IO modules did you have on the wago unit? we can get the PLC to turn on wago outputs but not the other way round. The NS light on the Wago unit is flashing red and sometimes the CJ1W-EIP module has a d9 error. Thanks

[René Heijma 1]

Is it possible to create a logfile with wireshark. Then we can have a look why it doesn't work.

[TheDr 0]

We did finally get the wago 750-841 unit to work with the Omron CJ1W-EIP21. What we had to do was to get an older version of firmware from Wago. Once we loaded that older version of firmware into the 750-841 then everything started working. Not sure what that means. Thank you everyone for your help.

[kaiser_will 65]

Sorry I did not see this until after you had the problem resolved. Your resolution (Wago firmware backdate) was very similar to our resolve. First I found the SMC Ethernet/IP combination I/O block/pneumatic output rack would not communicate with the Omron CS1W-EIP21 card due to the SMC EDS file not being correct. I experimented and found that the Omron Network Configurator software can "listen" to everything on the Ethernet network (it is connected to) and pull in basic EDS information. What the Omron card read was not the SMC EDS info but the AnyBus Ethernet/IP chip (that SMC uses). I modified the master SMC EDS file and was able to get the network up and running. Even though SMC said they had 1500+ installs of their Ethernet/IP units (typically on Allen-Bradley controller-based systems), they were not aware of any EDS file issues. It appears that Omron follows more closely to the ODVA rules compared to other hardware manufacturers. I found the WireShark tool to be very helpful in documenting what was going on on my Ethernet network. In the future, I will use this tool regularly for verifying proper communication amongst multiple devices. The other factor here is that Japanese controls manufacturers are not coming to market with Ethernet/IP options, compared to European or American controls manufacturers, or have less of an install base to reference. As more applications of their products occur, there will be more instances to reflect upon for troubleshooting options. Until then, it is quite possible that some of us may be in the "first on your block" situation to pave the way for others.

[PMCR 66]

kaiser_will has hit the nail on the head! Omron does follow the rules for the .eds files. Or more accurately, Omron uses them. In most cases, Rockwell does not use .eds files at all. The user simply types in the assembly number and size for the device that they are using, and away you go. Most 3rd party vendors have tested with Rockwell, determined what is necessary to make it work, and have written their manual around this configuration. Since this process usually does not use the .eds file, there are plenty of .eds files out there that have not been well tested. The WAGO firmware issue is still under investigation to determine the problem.

[René Heijma 1]

Had a issue with a Wago 750-341. Outputs where set but inputs where not coming into the PLC-memory. Every 10 seconds the NS-led of the Wago blinked shortly. Wireshark revealed that there seemed to be normal communication but every 10 seconds the EIP21 re-connected. I seemed that the EIP21 did not accepted the input messages from the Wago. Further investigation showed that Wago had the Don't fragment bit set in the EtherNet frame. An earlier firmware version of the Wago did not do this and the EIP21 had no problems communicating. After of course some tweaking of the EDS-file. Checking with the Wago firmware developer learned was what they did was according to spec. And after informing Omron they came with a fix in the firmware of the EIP21. Fix is still in Beta however. It is not in the current Version 2 of the EIP21. Bottom line. Yes Omron is following the EtherNet/IP spec closely but also they slip. (Only saw it once). All the other devices I saw had or EDS-files or firmware problems. I wonder if the firmware engineers ever did conformance testing???

[kaiser_will 65]

Rene, Omron's coming to market for Ethernet/IP interfacing products is slow and with reluctance. Japanese manufacturers, ironically, are dragging their feet to get on the bandwagon with Ethernet/IP, thus their base controls suppliers (Mitsubishi, Omron, Keyence) are moving slow to market with Ethernet/IP solutions. Japan may not be interested in an open controls Fieldbus communication standard like Ethernet/IP, but the rest of the world sure is (at least the United States is). You found a critical flaw in your application of the new EIP21 card. After this module has a broader install base, more of its firmware or application issues will be flushed out.