Have you ever made the statement, “If I had it to do over again, I would’ve done it differently”? Well, here’s your chance. Many companies are migrating from DeviceNet to EtherNet/IP and now you can take this opportunity to do it differently. As a former Allen Bradley Instructor, I have seen enough bad DeviceNet installs to write a book. Here are six topics to consider in your new EtherNet/IP installations:
1. Topology – Star, Linear (Daisy Chain), Ring, which one is the best for your application? All three have their strengths and weaknesses. Star allows for quick troubleshooting where one cordset or device can easily be identified as faulty, but this topology utilizes many network cordsets, as well as, having a high usage of Ethernet switch ports. Linear is a great topology for applications with long network runs, like conveyors, but one faulty cordset or device will kill all devices down the remainder of the line. Ring topology brings the advantage of one fault tolerance. One cordset problem will not bring the network to a halt, but there are some added expenses to adding this feature, as well as product availability to support such a topology.
2. Document, Document, and Document – I wanted to put this as #1, but in the sequence of events, this comes after choosing a topology. How can you troubleshoot, add new devices, or do preventive maintenance without knowing what you have in your application? You need to know your cordset lengths, device locations, switch locations, cordset flexing locations, and known problems area. Once you have this, make sure it’s available – make multiple copies, keep one at the machine, save it on your server, and be able to access it from your HMI. Just do it, no one has ever complained about too much documentation.
3. Cable Routing – Remember these are network cordsets, they carry a lot of data at a high speed. Stay away from high noise locations. Use the right cable jacket for the environment. Do not use standard cordsets in flexing applications. Use the correct cordsets at the beginning because it’s always a pain to go back and fix it later. Don’t forget to document any cordset changes that may occur over time.
4. Diagnostics – Simply put, use what’s available. Many DeviceNet users ignored diagnostic data in their PLCs from either their communication cards and/or the individual devices in the field. Use what EtherNet/IP offers, what diagnostics are available to you via the communications cards, PLC, field devices and managed switches. Making this data available on your HMIs is also a big plus. Don’t forget that many devices, PLCs, and switches have their on built in web servers.
5. Establish Procedures – Define procedures early. How do I swap out a device? If the switch is dead, how do I replace it and reconfigure it? Can I use a longer cordset to replace a shorter bad one? All these questions will come up at some point, be proactive and have a procedure in place. Document any changes.
6. Bandwidth and Package Usage – “Just add another device it’ll be fine.” Sometimes this statement starts a network down the wrong path making it unreliable. Know what a new device will do to your network traffic. Yes, EtherNet/IP is faster, yes it can pass more packets, but it does have its limits. Know where your funnel point is in your network, for most industrial Ethernet networks it is at the communication card in the PLC chassis. Know its limitations; be able to calculate its bandwidth and packet usage. Document your calculations and have a procedure to update these values when anything changes.
I hope these six items can help put you on a path to a reliable and easyily maintainable industrial network. Please remember to document, document and document.
Click here to learn more about EtherNet/IP and devices.