Repost of March Sound Off from ControlGlobal.com

3/31/2005 Further to the computer dependence issue... I heard a great story the other day. It seems that the student pilot was in the cockpit, waiting to take off on his final check-ride before his solo. The instructor noticed some items missing from the student's kit. "Where is your ruler and chart book?" the instructor said. The student pilot pulled out a pocket GPS and said, "Don't need those anymore, I've got this." The instructor asked to see it, looked it over, and then removed the batteries and handed it back. "Okay," he said, "your GPS batteries just died. What are you going to do?" I had that same sinking feeling when I heard a few years ago that the Navy and Coast Guard were going to stop teaching celestial navigation to officers and petty officers. I don't even own a slide rule any more, and I haven't got a clue where I could lay my hands on a table of natural logarithms. Yet any decent engineer thirty years ago had memorized the first 10,000 logarithms. Machines and software are wonderful, yes. But what do we do when they break down and nobody remembers how to do it the old fashioned way? Comments? 3/30/2005 So who's an expert? Sometimes I think the definition of an expert really is "a person at least a day's travel from home." Here, we are trying to make the final tweaks on our Ask the Expert pages at ControlGlobal.com. The way the module works is that we have one expert on each thing, and they get all the questions about that topic. Oh, well. Battle plans seldom survive contact with the enemy. So it is here. You see, we got very lucky. I was able to persuade our Uber-Expert, Bela Liptak, to agree to "moderate" the Ask the Expert section of our website. This, as he says, will allow people to stop paying him consulting fees by asking questions on the website. He's gone and roped in most of the contributing authors (including yours truly) of the monumental Instrument Engineer's Handbook, 4th Edition, to assist in answering those questions. This is likely to be a significant new resource for the process automation professional, if we can only figure out how to jigger the software to allow it to happen. Stop by and ask him a question. Comments? --Walt Boyes 3/29/2005 Patent wars...oh how fun! It seems that progress in applying RFID to manufacturing may be delayed based on the outcome of the war between Symbol and Intermec. In a report in eWeek's enews http://ct.enews.eweek.com/rd/cts?d=186-1826-10-80-190089-205143-0-0-0-1, Intermec and Symbol are slanging each other over a wide variety of patents because Symbol acquired a major Intermec competitor. While in other RFID news, RSA has found new problems with security, and Forrester Research has published a new paper indicating that there are more problems than expected with the new technology. Watch this space. Comments? --Walt Boyes 3/28/2005 For several days now, I've been having severe laptop problems...and I don't use a desktop computer any more. My laptop is becoming increasingly unreliable, and right now, it is "Not Dead Yet" like the plague victim from Spamalot. This has concentrated me quite forcefully on the issue of computer dependence in manufacturing. No plant can be run without its control system any longer, and in many cases, nobody remembers how to do it the old fashioned way. This can be a real Achilles' Heel, especially if you have a system working at near capacity, and requiring 95% or better availability. Most redundant control systems are fairly expensive, and difficult to engineer and keep running, since they were intended for the enterprise IT world, which fairly often works "working hours" and can do maintenance and re-boot systems on the weekends. At the Iconics Summit last week, I saw a very interesting and relatively inexpensive take on redundancy for single computer control systems: Marathon Technologies. Essentially, they run two completely independent computers, with identical software and hard drives, in lockstep. You can even have the two computers separated physically so that if one building is demolished, the other will continue to operate. I wish I had a redundant laptop right now. Comments? --Walt Boyes 3/24/2005 Jim Pinto's ENews yesterday provided a really interesting and balanced view of Schneider Electric...almost too favorable, since he neglected to mention Schneider's central role in the Solaia patent mess. Schneider, as Steve Kuehn's article in the April CONTROL magazine points out, sold the '318 patent to Solaia (whoever they are) for $1 and a share in the licence fees Solaia was to collect. I suppose this is a valid way to compete in the automation marketplace. Pinto also made some excellent points about innovation and the outsourcing of R&D to India and China. He also made some real valid sense about corporations who think "making the numbers" is the be-all and end-all of corporate purpose. The fact is, Marx's analysis of capitalism, while being right on the money point for point (though not his predictions about whatever "socialism" was to become), are coming true just as destabilizing technology is also coming to fruition. The monolithic concentration of capital in the hands of a relatively few giant corporations with so much political control that they essentially become the government is happening. If you don't think so, just look around. Marx was right about that. But at the same time, there are tools other than the armed struggle to permit the individual to win: the Internet, the Web, and the collaborative engineering and business process tools that have been produced in the past ten years. Why is this important to the process industries? The consolidation of capital since the late 1970s has caused a huge rearrangement of the way the process industries are controlled, and where new plants are located, and what older plants have to do to to remain open and competitive. Tools like PlantWeb, Iconics' BizViz, Experion, etc., etc., have made it possible for plants to work smarter, and be more productive than ever before. They have ALSO increased the individual worth of each knowledge worker...no longer proletarian cogs in the machine, process automation professionals have intrinsic added value, and are getting hard to find. This is good news. Comments? --Walt Boyes 3/23/2005 Through the magic of 802.11x I am blogging today from the Iconics meeting in Orlando, as I listen to a session on Fault Tolerant Servers. This makes me think a lot about connectivity. I am reminded of the story from the Shah Nameh (the "Arabian Nights") of the blind men and the elephant. Approaching the elephant from different sides, each blind man describes the same critter differently. To a large extent, this is going on in automation as well. Automation, whether process automation or machine automation, today, is all about connectivity, and the transparency of the process of data transfer. We are trying to move from essentially proprietary systems to systems that are open. Yet there is "open" and there is "truly open." Some vendors boasting openness mean that it is easy for them to integrate somebody else's legacy system into their own not all that open system. Others actually mean that their systems are truly open...that they have had OPC testing done, that they have had Fieldbus Foundation or HART Communications Foundation testing and approval done. The fastest way to even more productivity increases...and the way to save US manufacturing, if that matters to you...is to make it easy for the plant floor data to be visible throughout the enterprise. Simple, open, modular, object-oriented. I was absolutely floored last night at dinner to learn that Iconics has fewer than 20 technical support engineers worldwide. Most SCADA and HMI and MES software companies have many more support staff than that. This either means that it is hard to get service support from Iconics, or that Iconics software is really as simple to use as they say it is, take your pick. Comments? --Walt Boyes 3/22/2005 See this shiny flat spot? Right there in the middle of my forehead? Well, that's what you get from beating your head against the wall. For years, I've been preaching simplicity to automation vendors and field instrument manufacturers. Keep it simple! But this industry is seriously infected with tool envy. As Tim the Tool Man used to say, "More Power!" The problem with this is that end-users and integrators don't necessarily want to admire your tool. They want to solve problems and they want tools to do that with that are transparent and easy to use. Too few software packages for process automation, and very very few for MES integration and enterprise integration are designed for the ease of the user or integrator. SAP is proudly telling people that their new Netweaver platform will allow ERP and MES down to the $100 million company. Great. What about all the manufacturing and process facilities that are in the $10 to $90 million range? Especially consider that that range includes 90% of the vendors of process automation products and all of the integrators of those products. So what to do for the rest of us? I'm once again in Orlando (I ought to buy a frequent commuter pass or something) to attend the very first Iconics Worldwide Customer Summit. Considering it is the first one, the attendance (approximately 250, mostly end-users and integrators) is outstanding. And the theme of the meeting is that powerful software (by which Iconics means Genesis and BizViz, of course) doesn't have to be hard to use or hard to learn to use. There are impressive new features in Genesis and BizViz, and anybody who is looking for an integration platform that will do darn near everything anybody in process manufacturing wants should take a look at them. Comments? --Walt Boyes 3/21/2005 Somehow I got onto thinking about alarm management this morning. One of the most pitiful things I've ever seen is a cascade of alarms in a large control room. It was easier in the old days, when there was a huge analog annunciator where there was a painted description of the process train and the alarms displayed at the point in the train where the alarm was going on. In many modern HMI systems, alarms are displayed in log lists, as well as pictorially, trying to replicate the ease of use of the old analog annunciator panels. Why does the average operator prefer the old fashioned way? It ain't because they are opposed to change. In reality it has to do with the significant amount of the human brain that is hardwired for pattern recognition. Remember digital dashboards?? They went away because it is very hard to judge rate of change when looking at a number, while looking at a pointer is much easier for us. Evolutionarily, we are pattern recognition animals...my earliest ancestor was better able to discern the leopard hiding in the tree than his cousin or brother, and so he made it back to the safety of the cave, while his less percipient relative became leopard lunch. As we get to the point of having much larger display screens, we will be able to return to the easy recognition of alarm significance we used to have...and in the meantime, people like Honeywell, Matrikon, Emerson and others are working like bunnies on producing alarm management software that can triage alarms, display them in a hierarchy from "this alarm will blow up your plant" to "this alarm is an annoying pain but isn't critical." What is even more significant is that these alarm management tools can be used to link to expert systems that mine the knowledge base of the plant so that the alarm can follow the old adage, "If you bring me a problem, you better also bring me the solution." Comments? --Walt Boyes 3/18/2005 Moby RFID"¦ Steve Bender, Siemens' PR counsel, reports that Siemens is responding to the sudden issue of frequency glut by moving into the UHF range. Siemens, like Rockwell, has been in RFID since before it was cool, but unlike Rockwell, Siemens believes in vertical integration, and continues to make chips and readers under the Moby trade name, while Rockwell, you will recall, pushed chip manufacture into the Encompass Partners Program to concentrate on integration and consulting. Alex Stuebler, business manager of factory automation sensors for Siemens, said, "We expect to begin shipping the stationary readers as planned by mid-2005." Stuebler continued, "Our pilot tests have shown the devices to be extremely reliable and effective in providing industrial and retail customers with hardware components to design comprehensive RFID enabled solutions for improved productivity in industrial manufacturing environments and supply chain improvement." Stripped of the marketing speak, this means that they've made ruggedized RFID equipment designed specifically for the industrial and process environments. But all of this may be moot, since the exponential expansion of RFID technologies appears to be heading straight into a brick wall. Intermec (www.intermec.com) owns substantial intellectual property in the RFID space, and they've been confidently expecting to work out a deal with Symbol Technologies (www.symbol.com) over issues arising out of Symbol's acquisition of Matrics, which, Intermec says was infringing Intermec IP at the time of the acquisition. In an article on Line56.com, Todd Hewlin, Symbol senior vice president of global products was quoted as saying, "We believe Intermec's RFID IP position, if allowed to stand, could thwart the overall development and wide-scale implementation of RFID, which would be detrimental to the industry and all users of advanced data capture technology." Demir Barlas, the author of the Line56.com piece noted that, "In riposte, Intermec pointed out that it relaxed its licensing program in order to speed RFID Gen 2 progress and has not taken an exceptionally confrontational approach. These gestures distinguish Intermec from, say, SCO, which has been much more aggressive (qualifying as "terrorist," according to one analyst) in trying to enforce its IP around Unix." Tune in next week, or sometime soon I guess, for the next episode in the serial, "It's all about the money." Comments? --Walt Boyes 3/17/2005 Saint Patrick drove the snakes out of Ireland...at least so the legend says. News reports this morning put the number of botnets (computers controlled by malware) capable of initiating DOS and DDOS attacks at significantly higher than previously believed. Many of these botnets are capable of carrying out DOS (denial of service) attacks against most corporate networks. I wonder what it will take to drive the snakes out of the Internet? And while we're on the security subject, a couple of days ago, Bill Tarlinton, of Opal Software in Australia, posted this on the SCADA list, and it was important enough that I am choosing to post it in its entirety: There is another aspect of SCADA security which is consistently overlooked: How many SCADA systems will hold together following a big network disturbance where potentially thousands of points change value or state over a matter of a few minutes, resulting in large numbers of alarms and triggering a large amount of down-stream processing ? This is a particular problem for electricity distribution companies, especially in countries where industry reforms have led to the creation of "super sized" distribution companies with super sized SCADA systems monitoring several hundred thousand SCADA points. The problem is exacerbated by the increasing use of IED's, which are able to deliver up 100's more SCADA points per site than the traditional RTU. If the grid goes down for instance, virtually every analogue in the system is going to drop to zero and a very large number of digital changes of state will occur, not to mention failure of a large percentage of communications lines. SCADA/DMS/EMS systems are becoming more sophisticated: the amount of downstream processing performed following changes of SCADA points is increasing: many DMS systems for example now include functions to collate customer calls and predict outages, perform network tracing to show which areas of the network are dead and predict numbers of customer off supply (to name a few). The result of all this is that large scale SCADA events can trigger large scale demands on the resources of the servers, to the extent that the server's CPU load goes to 100% and stays there for a long period of time. Not a good thing to happen when you really need your system to maintain its real time performance and responsiveness. Moreover, large scale SCADA events can flush out bugs in the software, as the large scale event may be the first time the down stream processing algorithms get a chance to do their work against the whole network model. This can lead to complete failure of the system. Also not a good thing to happen when the grid has just collapsed. Lets face it, if a terrorist organisation wants to bring down the grid, there are a hundred things that can be done, relatively easily, to bring this about other than hacking into the SCADA system. Even without any help from the terrorists, natural events happen that result in large scale disturbances, the worst case being collapse of the grid. So one of the questions utility organisations should be asking themselves is.. "will my SCADA system be able to deal with the grid going down?" Some recent incidents in the UK and US where SCADA systems have failed under duress without any help from malicious attack suggest that this question can't be answered confidently in the affirmative. Bill Tarlinton Comments? --Walt Boyes 3/16/2005 In January, I talked about seeing Spamalot, the Monty Python-based musical by Eric Idle. If you are interested, visit www.montypythonsspamalot.com/HighBand/homepage.html For those of us of a certain age, and some of our children, Monty Python's unique brand of wacky humor epitomizes the travails of modern worklife. So it didn't surprise me when Bob Landman, the inventor of the fiber optic "modem," posted this on the SCADA email discussion list: Bob Landman wrote: You guys need to see this... > http://www.backuptrauma.com/video/default2.aspx Every bit as hysterical as an old-fashioned Python bit, John Cleese is the Director of the Institute for Backup Trauma. A word of advice, though. Whatever you do, don't press the third button. Comments? --Walt Boyes 3/15/2005 Ray Ozzie to the Dark Side of the Force The inventor of Lotus Notes and the founder of Groove Networks has joined Microsoft as Chief Technology Officer, as Microsoft acquires Groove. This is akin to Anakin Skywalker's metamorphosis into Darth Vader...except that Microsoft isn't the Evil Empire many people think it is. Microsoft has realized that its future is in facilitating collaborative work environments between people and enterprises. The Age of Computing is giving way to the Age of Information. In process automation, this translates to the Age of Process Control taking over from the age of Instrumentation. We're moving up a notch, and the tools are becoming much more sophisticated. Now we are concerned with the way processes are interconnected, instead of with how to measure them. We are concerned with optimizing information transfer between the plant floor and the business systems of the enterprise, instead of monitoring events. Things are going to start getting even more interesting. Comments? --Walt Boyes 3/15/2005 Walt wrote: "From the "Directory of Lost Companies" Dept.... Another icon bites the dust. I live in Aurora, IL and I drive past the manufacturing facility that makes Aurora Pumps..........." Sadly, this is the way we have evolved in today's global mergers and acquisitions mentality. I too live near the old Aurora plant (I live in Naperville). However, Aurora Pump is not your father's pump company. I've been in the W & WW engineering business for over 30 years and the Aurora Pump of my early years was a far different animal than today. Bean counters have killed the old products. What used to be a respected pump manufacturer is now marginal at best. I do agree that too many "brands" have vaporized. I wonder if Siemens will ever "get it" that most folks in the I&C field in this country find Milltronics to be more recognizable than the Siemens name, despite their attempt to eliminate the Milltronics moniker. Many of the favorites of my youth don't even exist anymore: Fischer and Porter Taylor Autocon Consolidated Electric Polysonics BIF (the real one) Geez, I feel old. Tim Greif, P.E. GREELEY AND HANSEN, LLC Comments? 3/14/2005 From the "Directory of Lost Companies" Dept.... Another icon bites the dust. I live in Aurora, IL and I drive past the manufacturing facility that makes Aurora Pumps, a truly long-lived icon which has provided huge numbers of sample transfer stream pumps, chlorine injection booster pumps, and other specialty centrifugal and turbine pumps to the process industries for close to a century. On my way to Orlando on March 3 I noted that the AP logo was on the sign. On my way back the next evening, it was gone. It was replaced by the Pentair Water logo. Pentair, of course, is the parent company of Aurora Pump. Pentair is also the parent company of another icon for process automation, Hoffman enclosures. Will we see "Pentair Enclosures" soon, too? It is amazing to me that companies with apparently normally intelligent executive management continue to simply throw away brands and brand equity like this. ABB lost enormous market share, as did Invensys, by doing the same things. Only a company like Emerson, whose "ur-brand" is strong enough to carry the message can devalue brands like Micro Motion, Rosemount, Brooks, and more without breaking stride. Honeywell should pay close attention to this as they become a major system integrator with a big group of branded co-conspirators. Comments? --Walt Boyes 3/11/2005 National Manufacturing Week was a bust. Well, for us magazine folks it wasn't...all our advertising clients were sitting in their booths without anything to do except talk to us about buying more advertising to pay back their marketing budgets for a lousy trade show...Pre-announcing that they were moving to the (sure it is) more economical Rosemount Convention Center (the old Horizon) next year may be the kiss of death. Exhibitors were giving even money that next year's NMW doesn't even happen. In other news, GE Energy (formerly Bently-Nevada) announced equipment health monitoring systems to compete with the newly announced Emerson CSI system we saw in Orlando. It is a more traditional system, and costs about what Emerson's system does... My problem, again, with this whole thing is that I cannot imagine a successful end to the scenario that starts with a maintenance engineer trying to explain to the cost accounting department why he will get enough ROI from instrumenting ONE motor train to be able to pay for the necessary stand-alone capital purchase requisition. When the cost to do this gets down below the capital purchase line, people will line up to do rotating machinery monitoring and then predictive maintenance will really take off. AVG's EZautomation.net and AutomationDirect.com had booths close to each other at the entrance to the automation pavilion at NMW. Both companies indicated that, as I predicted earlier this year, the existence of a new direct sales company in automation has increased sales for both companies. Hooray for competition and choice. What this means to me is that people don't necessarily want One Big Vendor, they want inexpensive access to quality goods. What they want is value add! Comments? --Walt Boyes 3/7/2005 Too busy to update the blog the last few days...so, to recap: Emerson announced the ability to integrate AMS data into PlantWeb via a Foundation Fieldbus enabled field device that not only monitors rotating machinery for vibration, temperature, etc., but runs expert system software (of the proprietary type, of course) to determine and forward to the control system Machinery Health status, rather than raw data from sensors. The price point for a single machinery train will be between $5000 to $7000 depending on configuration. For people who have fieldbus systems throughout their facilities, or who have industrial ethernet throughout (system works with IE and Modbus, too) this is a great solution for absolutely essential rotating machinery trains, especially the predictive maintenance software embedded in the new product. The real issue that Emerson will have to overcome, though, is that each of these systems is over the line for capital purchase items in almost every brownfield plant in North America and Europe. I'd like to be a fly on the wall listening to the justifications presented to management on the ROI possible by instrumenting one machinery train. It seems to me that the real course Emerson and other vendors should explore is to marry these types of devices with wireless mesh networks, and provide them at a price point under $1000 per point. This then makes it a maintenance operating budget decision, not a capital purchase item... Comments? --Walt Boyes 3/2/2005 More on RFID... Is RFID the panacea, the next great thing that will save manufacturing. NO. Not only "No," but resoundingly so. "Yet it is already heading down the wrong (technology) path as people spend more time talking about how the RFID chip can be used, rather than the business issues it will resolve," said a report by AT Kearney Inc. Joe Owen used this quote on Monday to illustrate why Rockwell isn't getting back into the sensors and readers business. Sure, Alien and the other companies are going to make a lot of money, hopefully, but Rockwell is quite sensibly correct that it is the real changes in business processes and the solutions to business problems that currently exist that will be the long term gain and the long term business model for RFID. Rockwell, therefore, is sticking to consulting and systems integration. Comments? --Walt Boyes 3/1/2005 RFID reconsidered.... continued from yesterday... Alfonso Gutierrez, from the University of Wisconsin eBusiness Consortium, joined us for lunch. He commented, "RFID, and other AIDC technologies will have the effect of moving manufacturing from a transaction model to a placement model. That is, the fact that a package is located in a specific location, such as receiving, or factory service, will cause decisions to be made, based solely on where the package is. This is the most significant advance in manufacturing theory since we learned to debit all the components of the BOM from inventory automatically when a finished product shipped." The fact that it is becoming obvious that there is little or no ROI coming from the "slap and ship" model that is intended to just meet Wal-Mart's directive isn't terribly exciting news. Pretty much everybody looking at RFID could tell that was coming. Joe Owen, who is Director of Strategic Marketing (read "Rockwell's Corporate Fireman") pointed out that the ROI will come from people who say, "Well, we have to do this to satisfy Wal-Mart, or the DOD, or General Motors...what else can we do with this technology?" He's certainly right. Comments? --Walt Boyes