Machinedesign 5833 Johnrinaldi042613 32 Web 0

What's my favorite PLC? The best I/O network? and other odd questions

April 29, 2013
Is there a best network? No. I once had a pesky lawyer trying to cajole me into analyzing 15 networks. He sent me a whole list. Essentially he was trying to figure out what the best one was — God knows why. I finally drove him away. You want to know how the most suitable network is picked? It’s simple.

I admit it — I’m a pretty opinionated guy. I read a lot of books with a goal of one every week. I read about everything from technology (as in OPC UA books, because they really help me sleep) to relationships (as in The 10-Second Kiss.) The stuff I read doesn't always appeal to my wife. She didn’t like me reading The Complete Guide to Picking Up Girls but I told her I just read Thou Shall Prosper (The Ten Commandments of Judaism) and I’m not wearing Yamulkes around the house.

From all that reading I've developed some pretty strong opinions on lots of things: Work (I’m for it), life (live it to the fullest), love and marriage (really good if you work at it), licorice (I would ban it), and lots more. If you ask my opinion on something I’ll not only give you an answer, but also launch into stories about something that happened to me one cold winter morning in Stockholm or that long day turned night in Bangkok.

But if you ask my opinion on some subjects, it’ll send me screaming into the server room — the only place in our building where my wailing can’t easily be heard. And no, the off-limits subjects don’t include dinner with one of the brother-in-laws or a loan to that work-challenged relative with appalling taste in boyfriends, husbands, and fathers for her children.

No, it’s when customers hit me with, "Can you recommend the best …” or “What’s your favorite ...” If these questions are about PLCs, I/Os, networks, ERP systems, or OPC driver vendors, you’ll make me stamp my feet, wince, and grip the phone like it’s my last lifeline keeping me from the cube farm at one of the corporate giants of the automation industry.

Why you might ask? Everything changes over time ... and there really isn’t an answer to these kinds of questions, anyway. Let me explain.

Consider the Mary Tyler Moore show in the 1970s. It started slow and a lot of people just hated it. People like me couldn’t figure out what happened to Rob. In the Dick Van Dyke show of the 60s, Mary was married to Rob Petrie and they were the perfect couple. Then in the '70s, Moore shows up sans Rob. Where’d he go? Did they divorce? Did that scumbag cheat on her? There were other problems: Dingbat Ted Baxter, snobbish Phyllis Lindstrom, and Rhoda — the abrasive, whiny, promiscuous neighbor. Rhoda was the antitheses of anybody ever on TV before. People hated her. So, the show started slowly, but became familiar over time. Viewers got more comfortable with it and eventually came to love it. The unfamiliar had become the familiar. In 1974, it was the number-one hit — and that was when everyone watched the same TV shows as everyone else. The critics started calling it one of the best comedies in the history of television. Then the bottom fell out: 1975, #17. 1976, #39. In 1977 the show was over. Mary walked out of the newsroom and turned out the lights. Screens across America went dark for three whole seconds. The familiar had become too familiar.

I’d argue that you see the same kind of process as new technologies roll out on the factory floor.

I remember the first reactions to DeviceNet. Users, vendors, lots of people hated it. You mean, you have to put one of those CAN chips in every device? Who’s gonna do that? Do you know how much that will cost? You have to buy odd and expensive connectors, Ts, and terminators? Lots of people thought this won’t fly. Unfamiliar it was.

But eventually, all these people came around. We all started using DeviceNet and once it became familiar we started to like it. I even came to love it. (I’m confident enough in my masculinity to talk about love.)

And where are we today? We haven’t turned the lights out on it like Mary did with her show, but who really wants to use DeviceNet? That’s yesterday’s technology. It’s too familiar. Now, I’m not arguing it’s technical merits. I am talking about the psychology of technology introduction on the factory floor.

Ethernet — so hot right now

Today, Ethernet’s the place to be. In my head I can hear my old systems-integrator buddy with his deep Texas drawl: "Why son, no one uses DeviceNet anymore." The familiar is now too familiar. It’s not fun anymore. It’s clear when you look at the number of DeviceNet network deployments vs. Ethernet network deployments.

It’s simply a fact that we actually all behave pretty similarly. It’s true of our TV shows and our automation technologies. And I’m going to tell you a secret about that. Something that we in Industrial Automation tend to forget. And something that we need to be reminded of from time to time.

Here it is: We are all humans. Yup, your customer, the guys and gals that build you gizmos, even your boss and that guy in marketing (okay, not him) are all humans. We have human emotions. We fall in love with things and as engineers, we fall in love with technologies. We’re just like all the people that hated the Mary Tyler Moore show, came to love it, and then dumped it. The same thing goes on in this very logical, precise, and staid industry.

The second reason I have for my very odd behavior when asked these questions is that there is no answer to these questions. None at all.

In the old days I was an engineer, a Software Engineer. I did some time at one of the big paper companies and I worked at that big PLC company with the meatball logo. I never talked to customers in those jobs. There was some unwritten rule about engineers talking to customers. Something about the truth was what was whispered to me one time over an adult beverage.

Then when I started my own company, Real Time Automation. I still didn’t talk to customers a lot. In the early days we mostly concentrated on building our IP library with Profibus, DeviceNet, Modbus, Modbus TCP, EtherNet/IP and all the rest.

But now that I have more of a Marketing & Sales position (I was kicked out of Engineering) I talk to customers a lot more. And you wouldn’t believe how many people apparently think that I have the secret locked away in my credenza right under the picture of me and the Octopus in Hawaii (another good story). They seem to think I can give them the magic ingredient to build the perfect system. That I can tell them the perfect PLC to use in their application. That I can point them to the best sensor network, the best Ethernet network, and the best I/O devices.

Folks, they don’t exist. The truth is that most PLCs are pretty much like every other PLC. Everybody has the same architecture. There’s a Baby Bear PLC, a Mama Bear PLC, and a big hulking Papa Bear PLC. Sometimes there’s even a lot more. The differences between the Mama Bear PLC of one manufacturer and the Mama bear PLC of another manufacturer isn’t really that much. Yes, they vary — but not as much as those manufactures would like you to think. The truth about PLCs is that the decision to select a single vendor was made a long time ago and nobody is going to change it now unless that vendor goes out of business. There is too much momentum in a facility to make a change. They have the units in place, the backups in stores, the configuration files, the drawings, the training systems. So change? Can’t be done. If you’re a Rockwell, Modicon, or Siemens house it’s pretty likely you’ll stay that way.

Same thing with the networks. Is there a best network? No. You want to know how the network is picked? It’s simple. 99% of the time, if you’re using Siemens you are going to use Profibus or Profinet IO. If you’re using Modicon PLCs you’ll probably use Modbus or Modbus TCP. With Rockwell it’s EtherNet/IP and DeviceNet. The PLC really drives the network selection. After that, the requirements of the application dictate the specific flavor.

I once had a pesky lawyer trying to cajole me into doing an analysis for him of about 15 networks. He sent me a whole list. Essentially he was trying to figure out what the best one was — God knows why. I finally drove him away. Every time he emailed me I emailed him back asking for a credit card number.

It’s pretty much the same in the I/O game. There are a bunch of variables that dictate the requirements — do you need IP65 or RTD inputs or Open Collector and all that. But once you have that set of requirements, all the I/O modules that meet those requirements are going to be pretty much alike.

At that point, the purchasing nerds get involved and make the vendors crawl over broken glass, eat roaches, and swim through a crocodile infested lake to get the order.

I told you I had opinions. You were warned.

Editor’s note: John wanted the following added to his column. We hope you’ve enjoyed his humor. Stay tuned for more from John in coming months.

This is a column of well-reasoned, personal opinions that for some odd reason doesn’t necessarily reflect the opinions of Machine Design, the management, staff or the non-english speaking, nighttime cleaning crew at Real Time Automation — though it really should. And as for my staff: You shouldn’t be reading this anyway, as you’re lucky to be employed ... and shouldn’t be sitting around reading magazines when you could be washing my car or updating my facebook page. Before you contact John Rinaldi directly, think about finding something useful to do with your time and not using the office computers for such nonsense.

Voice your opinion!

To join the conversation, and become an exclusive member of Machine Design, create an account today!