Ethernet/IP news

Interesting and a huge confirmation that Rockwell had it right all along.

1. How long will Seimens hold out?

2. Do Rockwell get to embed Modbus in their TCP stacks?

3. Are we at long last within site of a genuine industry-wide connectivity standard?

Go Schneider!

Personally I can't understand why a company would not develop products that adhere to the Ethernet/IP standard. Most already have products that adhere to the Devicenet standard, and if you understand the Devicenet CIP structure the Ethernet/IP CIP structure comes very easy as it is similar. Seems this amount to less development time for PLC programmers and less training time since you only have to learn one protocol. Then you get to seamless bridging between the two networks and you can come up with some pretty intricate networks that very efficient.

The only argument I could see, which I don't know enough about to argue either way, is the 500 byte packet limitation on Ethernet/IP

Of course that's just my two cents

PhilipW said:

2. Do Rockwell get to embed Modbus in their TCP stacks?

Click to expand...

Why would Rockwell want to? Modbus/TCP is not a good protocol for sending I/O type data. There should have been a Modbus/UDP for that.

TWControls said:

if you understand the Devicenet CIP structure the Ethernet/IP CIP structure comes very easy as it is similar.

Click to expand...

Not really. Only the philosophy is the same.

Ot

Wow Peter is at 2500 post, keep up the good work.

PhilipW said:

1. How long will Seimens hold out?

Click to expand...

I think that Ethernet/IP is somewhat ahead of Profinet in market share.
But I also think that Siemens and Phoenix Contact are relying so much on Profinet that there is no option other than going ahead with it. And considering that the entire product line is up and and active, it would also be very stupid to do anything else.

PhilipW said:

2. Do Rockwell get to embed Modbus in their TCP stacks?

Click to expand...

Funny thing is that Siemens now support both Modbus TCP and Ethernet/IP in their panels. And they have a library for supporting Modbus TCP in their PLCs. I think that their strategy is to win as much as the PLC and HMI market share, and then the networking maket will follow. I think they are right, because there is much more invested in both PLC and HMI code than in networking. If it is technically available as an option, you can switch network type with a few mouse clicks (ok it is a simplification), but PLC and HMI code is locking you in.

PhilipW said:

3. Are we at long last within site of a genuine industry-wide connectivity standard?

Click to expand...

Ha ha ha ha ha, oh dear oh dear.

Nah, Profinet, Ethernet/IP, FF Ethernet, EtherCat, EPL etc. will live side by side. And we have not yet seen the last ethernet based field bus.

There may be a common protocol yet

http://ieee1588.nist.gov/
Go to the tutorials.
Chapter one is written by Rockwell.
Chapter two is written by Siemens.

IEEE 1588 is yet another time stamping and synchronization protocol. It just has bigger backers than the rest.

We use AutomationDirect Plc's exclusively. We do some networking with their proprietary protocols (K-Sequence, DirectNet), but recently started using their ethernet card for networking and it only supports Modbus/TCP. It seems to work well for us. I don't see AutomationDirect supporting the Ethernet/IP protocol in the near future.

Peter Nachtwey said:

Not really. Only the philosophy is the same.

Click to expand...

Perhaps I should rephrase. If I know how to assemble a CIP message is a Controllogix processor for Devicenet then I can quickly learn to do the same in Ethernet/IP

Your software gurus have already to all the internal "Stuff" that makes the magic work. When it comes to the magic part you pushing out of my league Peter

Nobody must have gone to www.odva.org

Omron and Eaton are on the list too.

Apparently Omron have Ethernet I/P on the way and they have CompoNet here now. For distributed I/O CompoNet is excellent - just not a lot around yet. It is controled by ODVA as are Ethernet I/P and Device Net.

Far more complex networks, faster speeds, longer distances, ordinary twisted pairs, repeaters etc. Bigger, better, faster, much cheaper Device Net???

Token ring also - no chance of crashes. Network update time at 30 metres over a twisted pair is 1000 I/O in 1 ms at 4mbaud!!!

I am using CompoNet on a job at the moment and have quoted into seven other jobs - just have to win them. With respect to cost, an AB 8 in, 8 out Device Net module costs me about $570 AU - an Omron CompoNet 16 input module costs me just over $300 AU - considerable savings with the safety of token ring and no crashes.

It also has provision for explicit messaging. I like it - on one job I quoted it reduced my costs for hardware (I/O cards) alone by just over $4000 AU, not to mention no very expensive Device Net cable. No configurater required for purely digital I/O either - just set an address on the master (rotary switch) and on each slave (DIP switches) and the network is set up. Auto baud rate detect controlled by the master and the slaves adjust to whatever is thrown at them. Uses EDS files for other brands where a configurator is required with mixed brands. The Omron configurator is a free download for Omron CX-One registered users.

Downside? Only one master.

Have heard some "how's your father" reports about Profinet here in Ozz on busy networks. Any comment/opinions from you guys in Europe that may have used it?

By the way, with Device Net, Ethernet I/P, CompoNet I would only use devices that have been fully tested by the ODVA labs for compliance with the standards. Not 100% guarantee of success but certainly reduces the odds of having a nasty conflict ocurr.

When I read about the ODVA confab in Ann Arbor, my first thought was "what is Schneider up to ?"

Remember, these are the guys who tried to use a proxy law firm to get royalties out of everybody (at least every Rockwell customer) who used OPC, an "Open Standard" that they had promoted along with everyone else.

Modbus is an open standard. Schneider's been supporting it, and Modbus.org, for years. Everybody uses it.

So what's keeping Schneider from suing over some patent they've got in their back pocket ?

Call me suspicious-minded, but I'm not going to forget being asked to sign an indemnity form saying I'd pay GM's legal bills resulting from installing a simple MMI application if Schneider's pinstriped hit-men sued them.

It is my understanding that they threatened to sue anyone who put a web server in PLCs as they claimed they had a patent on it. I think the only ones they frightened off were Omron.

Peter Nachtwey said:

Omron and Eaton are on the list too.

Click to expand...

Where, I know they are for Devicenet but neither actually have any conformance test products yet do they?
http://odva.org/default.aspx?tabid=155