Industrial/Automation Forum

A new wave of MEMS sensors and sensor networks promises to reshape the industrial landscape

CONVENED AND MODERATED
BY RICHARD COMERFORD

In the wake of July’s Sensors Expo conference in Rosemont, IL, Electronic Products convened its annual Industrial/Automation Forum. Many participants found that there was a shift in the industry, with a marked interest in sensors and sensor networks for use in new industrial applications as well as old. The following is an edited transcript of that meeting.

Electronic Products: I want to welcome all of you and thank you for joining Electronic Products ’ 2008 Industrial/Automation Forum. I think we have a really interesting and widely mixed group — we cover Europe, North America, and I think we have some perspectives on the Far East here as well.

We’re here to talk about issues in industrial/automation today. I’d like to begin that by asking all the participants to introduce themselves and tell us some of the issues you see today.

Bob Scannell (Business Development Manager for iSensors, Analog Devices) : iSensors is a MEMS-based technology group focused on solutions for industrial customers. We see this as an incredibly interesting application space and customer space, because the opportunities are truly wide open here. It has not in any way been fully exploited yet by, say, inertial MEMS. There is a lot of opportunity to add different types of value to these systems in terms of safety, reliability, efficiency, and performance.

But the issue is that, while there are opportunities for, say, motion sensing, among our customer base, they tend to be extremely complex system designs, and the customer base is traditionally not so familiar with MEMS.

So the challenge for us is to get a much better understanding of these customers and their needs, and to translate that into products that are much more specific to this customer base, as opposed to the products available today for, say, consumer automotive applications. We need to look specifically at the needs within the industrial automation space and be very focused on that.

George Iwaki (Director of Product Management, Arch Rock) : I like the way Bob introduced the topic, raising the issue of the complexity of these networks. Arch Rock is a wireless sensor networking company focused on providing complete systems to the market. And I think one of the main considerations is how best to scale these networks. Our point of view is that a standards-based approach should be used to solve that problem.

To put a fine point on it, the Internet protocol can enable this scaling of sensor networks because the Internet Protocol allows many of the management functions to move out of the sensor network gateways and back into the servers. This is essentially the model of the Internet — a very familiar networking technology — and would allow interoperability between the new forms of sensing and control that’ll happen on factory floors as well as back in the data centers.

Jason Goerges (Application Engineer, ACS Motion Control) : ACS Motion Control’s primary niche is high-performance motion, and that’s been our traditional niche in several markets. But recently we’ve really been pushing products which cater to higher-volume systems. Specifically, for industrial automation we’ve included support for standards such as CANopen which allows great flexibility in the industrial automation networks, allowing sensors, I/Os, and additional axes of motion.

We’ve seen a trend to more and more complex systems requiring coordination of high-performance motion along with discrete I/O, analog I/O, sensors, and many other products. My company has been focusing a lot on the ability to accommodate these applications where it’s no longer just a high-performance motion machine, where several other factors are involved. So a lot of our software and hardware development has been focused on that over the past 12 months.

Vinny Acampora (Worldwide Product Manager, C&K Components) : C&K Components is a leader in quality switches and has a broad range of different types —precision switches, dip switches, pushbuttons, T-locks, and more — which we design and manufacture both here and abroad.

There are an increasing number of applications in industrial automation requiring switch manufacturers to design and develop product with a lot more functionality and performance, and to try to get the overall cost of our product lower. As a switch manufacturer, we really have to continue to innovate and develop variations to our existing design that have functions and improve performance especially with the environment getting more harsh, you know, like seal switches. The need to increase our performance at a lower cost and smaller size has become an increasing demand from our customers. And for instance OEMs are extending the warranty life of their final product. That means our component within the end product must follow suit.

For instance, life cycles which used to be acceptable at 10,000 operations went up to 100,000. Most customers are really looking for a million cycles which puts a lot of pressure on our designers and engineers. To meet this requirement the engineers have to work with a variety of materials and processes to try to double the life of our switches.

Andrew Glascott-Jones (Applications Engineer, e2v) : My particular business unit produces custom mixed-signal ASICs specializing in precision analog design. And it was very interesting to hear what Bob was saying about MEMS technology. We have a number of ASICs that interface to MEMS for MEMS clients.

Because we produce custom ASICs, the difficulties that we see are the long design cycle that comes with an ASIC and being able to convince potential clients of our capabilities within that design cycle. So these are the issues for us. We’ve developed certain ways around these issues — in particular, using development kits and intellectual-property sample components — that allow potential clients to evaluate and try out their sensors and systems with our interface blocks, before and during the ASIC design process.

Electronic Products: What ASIC requirements are changing for you?

Andrew Glascott-Jones : Bob has talked about it already; MEMS technology is pushing the boundaries. For example, with the MEMS sensor we need to really measure very, very small capacitances and so the precision that’s required to interface demands components with very low noise. I’m sure Bob sees it as well. There is quite a shift in technology.

Bruce Strachan (Regional Marketing Manager for North America, Infineon Technologies) : I’m dealing with magnetic sensors and pressure sensors. And the interest of our business is to increase efficiency of systems through better control. We offer components in microprocessors, power electronics, and sensors. We endeavor to put those three types of components together to improve systems.

George Blazas (North American Product Manager for Cannon RF products, ITT Industries) : One of my product lines is a connector we call CmX which is geared toward the servo and linear motor industrial market. This unique connector addresses the need to integrate signal and power into a single accessible housing.

What I’ve learned from visiting many of the major servo and linear motor manufacturers in North America is that the industry is continuing to find ways to reduce the size of the motors while increasing their power. Motor manufacturers are constantly evaluating alternatives to reduce the number of signal wires required to operate and have integrated on some models Ethernet interconnects for monitoring and programming.

So on the motor interconnect side, we continue to explore connector solutions requiring lower profile, space saving, and simplified assemblies. Also, I see a trend for more color coding to speed up and simplify the cabling installations for servo motors. In addition, manufacturers require IP sealing to protect servo motor interconnects that are exposed to liquids like water, oils and cleaning fluids commonly used in CNC machines and food processing equipment.

IP stands for Ingress Protection. It is a rating system developed in Europe that specifies the amount of environmental protection an enclosure provides. Our CmX connectors have a rating of IP 65 (dust and low-pressure water jets) and IP 67 (dust and fluid immersion between 15 cm and 1 m). These are typical ratings required for sealed interconnects for the servo and linear motor applications.

Electronic Products: Now that our participants have introduced themselves, let me start things off by making an observation. I recently attended a show in Chicago that had to do with sensing. And the two things that I noticed were: there was a heavy introduction of products related to wireless sensor networks, and everywhere I looked I was seeing new MEMS sensors.

So let me ask first what your experiences are with wireless sensing networks. Are those actually getting out there and getting deployed? And, if so, is there any early feedback on how those networks are operating?

George Iwaki : I also attended Sensors Expo. What we find is that typically in an automation environment — a factory floor or otherwise — there will be some legacy dataloggers machines that are recording information, quality, process, safety, et cetera. But that information tends to be stranded.

There are these islands of stranded data. And they’re maintained or monitored by people in a typical way. Someone will walk up with a clipboard or a reader and take a measurement, but it is largely a manual process.

In many cases, the solution is a wireless sensor network. And why does it need to be wireless? Well, pulling wires is very expensive. We know those costs. Also a wireless solution tends to be a redundant solution on top of what is typically a production environment where you don’t want to stop the machine to reconfigure it. And that’s why I think the appeal of wireless sensor networks is so great.

Andrew Glascott-Jones : What sort of standards do you see being put on to the shop floor, George?

George Iwaki : That’s a great question and the standards I’m referring to are the Internet protocol standards. We would be able to bring an IP-address down to the node so that the information that’s collected on these machines can be networked and passed seamlessly back to the data center without any kind of proprietary protocol that needs to be translated at an interface or gateway level. You may have an application or a server that sits comfortably in the data center, either co-located with that factory or back at the headquarters or wherever the data center is located. And it’s able to collect that data.

Electronic Products: Have you seen applications iusing Internet/ Ethernet protocols along with, say, CANopen protocols in industrial environments?

George Iwaki : For CAN, no. We’ve seen Modbus of course. And some of the HVAC systems, Modbus, BacNet are typically employed.

Jason Goerges : Yeah. We actually have been working a lot developing our controllers giving them libraries and standard interfaces to CANopen devices from several different vendors.

But we’ve also seen a trend for high-performance networks towards Ether CAT, for example, an Ethernet-connected network. One of the problems with CANopen is that the data speed on the bus is not always guaranteed, meaning that it not always necessarily a real-time network for transferring data for CANopen devices. And as you can imagine, for high-performance devices if you want to record information such as position of a motor or very high sample rates of inputs, outputs, things like that, CANopen suffers from latencies.

The request for Ethernet-connected devices is becoming more and more popular as these networks become more complex and require faster data transfers.

Electronic Products: I see. What about some of the other wireless protocols? Protocols like ZigBee? Is anyone seeing that move into the marketplace with any steam?

Andrew Glascott-Jones : Yes. We are certainly looking at ZigBee as an option. Some of our ASICs are used in low-powered applications and so have essentially short-range requirements than other interfaces. So we’re certainly looking at ZigBee.

Electronic Products: So it seems that we’re faced with trying to integrate a large number of different types of networks into a consistent whole. Is that an accurate characterization of some of the work that’s going on today?

Bob Scannell : I think George did a good job of laying out the attractiveness of going wireless. I completely agree with that. And a question for us has been more of “when will this really happen” and to some extent, to understand “how it will happen.” Although, the “how,” in terms of which standard will be used, will eventually just shake out and be clear. The question is when is this really going to happen within the industrial automation sector?

Speaking from a sensor viewpoint, maybe there’s a first step here, where more integrated and more intelligent sensors become available to these applications and customers first. Currently, there’s quite a big gap between raw sensor technology and actually getting that implemented within these very complex designs.

The wireless feature can to some extent be integrated on top of the embedded sensors. For instance, vibration sensing for condition monitoring is an obvious, attractive application for wireless, and having fully embedded programmable sensor elements more accessible to these customers will then make the wireless step more natural.

One reason for this is that we need to have more embedded processing localized at the point we’re trying to sense, to first reduce the data rates and reduce the power requirements; and then perhaps do decision making remotely via the wireless link.

Electronic Products: You’re looking to see more intelligent, more capable sensors embedded to take on a lot of the reduction of data that needs to be transmitted wirelessly. Is that a fair characterization?

Bob Scannell : Yes, to reduce the data rates, as well as power, and provide a more accurate and autonomous sensing capability. It is important to have this technology available to be embedded in the first place and accepted by the industry.

Bruce Strachan : I’d like to comment on that. Our sensors contain a good deal of electronics that do necessary compensations and processing. They also do signal processing in order to determine speeds and the actual parameters that are ultimately transmitted. So there’s a good deal of intelligence in the sensor itself right at the sensing location, and I would echo Bob’s sentiments. There has to be a good deal of processing right at the sensor itself.

As for adding additional wireless capability to the sensor, one of the options you are going to run into is cost. In the automotive business, that is always an issue when people start bringing up different ways of networking sensors and other devices, and I expect that translates also into the industrial market.

One last comment, relative to the sensor show. From what I saw there, next year they are going to call it The Wireless Show. Everybody was absolutely flabbergasted that probably 40% of the booths there had something to do with wireless communications. So maybe we don’t know when or how all of this is going to take place, but there is an immense amount of interest and activity going on for wireless.

Electronic Products: I’d agree. But it also sounds like the problems are really very nicely separable. You can deal with the sensor and the embedded control technologies. And then separately deal with the issue of wirelessness. In other words, there is a very nice boundary between those two. And that you can therefore implement the networking in the way that you’d like to and implement the sensing in the way you’d like to. Is that a good characterization?

Andrew Glascott-Jones : You’re right in that there are blocks. I mean we have our sensor interface IC into which we’re embedding a microprocessor. And then we’re adding a block on what will be the RF interface. So, yeah, you’re right. You can look at things in different blocks.

Electronic Products: Okay. And that’s where you can implement the various standards you need — between the sensing and the network itself and how the interface takes place between those two for different rates.

George Iwaki : Right. I think an architecture that seems to work in what I’ve heard even on today’s call is having the intelligence at the sensor and, throwing that data onto the network.

In some cases, you need to have power autonomy, and a low-power, low-bandwidth IEEE-802.15.4 network is appropriate. And sometimes you need higher-bandwidth, real-time communications, which would call for another type maybe a Wi-Fi network. It’s finding the right fit for the right application, and not having to worry about protocol translation in between.

Electronic Products: I would expect that power consumption would become an issue if you don’t want to run wires to these sensors to embed them. So there must be quite an effort going on to reduce/minimize power.

Andrew Glascott-Jones : Yeah, we have a number of products that go into systems such as wristwatches and tiny portable systems. So we really have to work very hard to reduce power consumption, and at the same time have enough precision in the analog to perform the sort of difficult job of measuring the small changes in the capacitance or resistance of the sensor.

So yeah, power is one of the things we have to look at very carefully.

Electronic Products: Anyone else have power concerns?

Bruce Strachan : We have some products which implement such things; magnetic switches is one. The switches have extremely low power consumption as it is, but then what we also do is activate it only periodically and poll the device. We actually have a duty cycle for the measurement. For low-bandwidth measurements, we can push the power down even further so that the overall power consumption is extremely, extremely small.

Bob Scannell : I think Bruce makes a good point. Adding intelligence at the sensor in a lot of cases can improve the power. There are a lot of tradeoffs to be made there: the intelligent sensor can take on the duty cycle issues and everything else. It’s not just adding a bunch of processing power.

George Iwaki : We’re co-chairing a new IETF (Internet Engineering Task Force) working group called ROLL — which stands for Routing Over Low-Power, “Lossy” Networks — to further define standards and how to implement these protocols for inherently low-power networks.

Arch Rock’s CTO and co-founder David Culler and JP Vassuer of Cisco are the co-chairs. The process for IETF is generally an open invitation, and those who have the resources and the “passion” (if you will) to collaborate, get together and produce what are called RFCs, Requests for Comments. Typically within 12 to 15 months there’s an outcome where the RFC’s get solidified and blessed if you will. The first meeting was in March during what was called IETF 71 in Philadelphia.

Electronic Products: Let me switch gears here a little bit and talk about the need for greater precision on the sensing end of things. Let me put it in this context recently I was talking to some people about monitoring operations for preventive maintenance or predictive maintenance on systems in the industry.

And one of the things they were telling me was that actually most sensors of one type, or class, didn’t have the bandwidth necessary to actually detect some of the problems. I’m wondering if you’re seeing a request from industry to provide greater sensitivity, greater ranges in sensors to be able to do some of the things that people want to do.

Andrew Glascott-Jones : Hearing Bob talk about MEMS technology, I could really agree with what he said. MEMS technology is really growing, and we’re seeing more and more people coming to us with requests for an ASIC to interface with their MEMS.

But particularly, when I think about some of the capacitive type MEMS sensors, the change in capacitance is really, really small. So we really have to fight hard to get the resolution required to, say, detect a capacitance change of the order of hundreds of femtofarads, with noise levels as low as 20 attofarads. These are the sort of things that we have to interface to coming from the MEMS application arena.

Bob Scannell : That need for more precision is often related to the need to predict these errors early in predictive maintenance. So vibration monitoring is a good example. And instead of having relatively crude monitors external to the system, customers are wanting to mount sensors further down in the system; much closer to the point of interest.

And that’s why there is a need for wider-bandwidth sensors, that’s why we’ve moved in that direction. Wider bandwidth sensors allow you to detect error profiles that could not be seen before. They also allow designers to detect these issues much earlier and prevent downtime.

Electronic Products: So moving these sensors closer to, say, bearings or things like that could get you a much earlier indication that you need to do some work, like oiling or replacing gear that shows signs of wear.

Bob Scannell : Exactly. And you’re looking for a different profile at that point.

Electronic Products: We’ve mentioned MEMS a great deal. Until recently there had been some concern on MEMS reliability, lifetimes, and packaging MEMS devices for very hostile environments. A lot of these concerns seem to have been overcome. Can anyone address those issues? Bob?

Bob Scannell : I can’t directly speak for the Analog Devices division that serves the automotive industry with MEMS, but I believe that it has been overcome. It is an extremely robust technology that’s been shipping in very high volumes to that customer base for years now. And the automotive customer base is known for being quite stringent on quality and reliability. So from our standpoint anyway, that is an issue of the past.

Andrew Glascott-Jones : For sure we have product shipping to interface to MEMS sensors that go into automobiles. And so from our point of view that issue has been resolved.

Bruce Strachan : There is a wide range of MEMS sensors. Pressure sensors, accelerometers, and maybe some gyro sensors were the most complex in the past. But now there are all sorts of devices like gas-analyzing devices and just amazing new things that have been miniaturized with MEMS.

Relative to pressure sensing at Infineon, one of the things we do is we avoid packaging electronics and MEMS devices separately. We build our MEMS Barometric Air Pressure sensors and also Side-Impact Air Pressure sensors on a single chip. The MEMS structure and the electronics are integrated on the silicon. This alleviates some of the problems with interfacing electronics such as capacitors. It also makes the package smaller and much more reliable. If you can do that, it is a step forward. Admittedly though it’s not possible in all applications.

Electronic Products: With sensors shrinking in size there must be some concerns about how you ensure the packaging for these new smaller devices actually protects them from harsh industrial environments. What developments are taking place in the sensor packaging to ensure reliability?

Bruce Strachan : For two of our product lines — Hall-effect and GMR magnetic sensors as well as our pressure-sensing products — half of the engineering effort goes into the package itself. This is especially true for pressure sensing. I think anybody who deals with sensors recognizes that a major issue is how you package the basic sensing element.

After you have the basic electronic device in a microelectronics package, like a surface-mount device, then that package has to be repackaged into something that could be put into a factory, a car, or whatever. That’s another step forward that has to be taken and there are certainly challenges.

Bob Scannell : I’d certainly echo that; the package effort is extreme — it’s an inherent part of the design and can’t be ignored. One piece is protecting the MEMS device; the other is understanding how the customers are going to use these and, in some cases, how the devices are assembled. Also, the type of system they go into can translate up into effects on the performance of the MEMS device.

We package our industrial-base sensors in a way that shields against these secondary effects, the assembly related effects. The performance stays intact, even when relatively little attention is paid to basic assembly practices.

Bruce Strachan : That’s absolutely true. We do two things to protect our Hall devices against environmental and electrical effects in the die itself. We package the product in such a way that amazingly enough the finished product can go to 180°C, which is absolutely amazing.

Again the repackaged part has to withstand the same environment.

Electronic Products: There are similar environmental requirements for new connectors, which also have to be small and yet also have to withstand a lot of different types of vibration, EMF, harsh chemicals. What are the requirements for the new connectors that are coming out like?

Vinny Acampora : Well, for instance, take our snap switch line or our switch line. We do have special materials we have to use for high-temperature plastics. Our packaging is antistatic bags, tape and reel, special plastics trays that can take high temperatures if you store them up to 180°C. The switch itself is sealed inside a tray. That the type of thing we’re seeing in the switch industry.

Electronic Products: And yet, to make your life even easier, while we’re putting more demands on the performance of the packaging, we’re also now starting to ban certain materials that have long been used in this area, because those material are being reclassified as dangerous to the environment as well. So now we’re experimenting with some new materials as well.

Vinny Acampora : Most of our switches or components now are RoHS compliant where we had to go to different plastics. We’re actually in the process of making many changes to our plastics that will allow us to meet the RoHS requirement.

Electronic Products: And the RoHS requirements are shifting, eliminating other materials.

Vinny Acampora : We have one gentleman, all he does all day is work on the shifting requirement to make sure we can meet them.

It’s become a challenge. You know, between the lower-current requirements of the sealing — and more and more companies are going to sealing — it becomes a challenge for the design engineer to start meeting these requirements. We get down to 20 mA and you have to meet 100,000 or a million operations, and it becomes a challenge.

Bruce Strachan : And after you solve the problem with the RoHS and the green product at your chip level, then you ship it off to the manufacturing plant and he’s got to deal with “how the heck do I solder this thing,” because he doesn’t want to use lead solder on it.

Vinny Acampora : Right. We have the lead free devices which we have to go to special plating.

Electronic Products: I really admire you for trying to keep up with this, because it must be nerve racking to have use these new materials and to assure that the quality will be there, to deliver the product performance that you need. It’s almost like working in a chem/physics lab at a research institute to be able to get these new materials into products.

Vinny Acampora : One of the challenges is, of course, is to keep the cost down. I mean, you get these materials, now you have to try to find a process to keep the costs the same or even lower. And the new materials are actually costing us more money, with the oil prices and metal cost rising. So the challenge is not only to get the material to meet the RoHS, but also keep getting a good manufacturing process where you keep your costs.

Electronic Products: You raise a very interesting issue. Everybody thinks about the oil crisis in terms of their automobiles and how much gas costs at the pump. But for many people in the electronics industry there is a real impact on the cost of devices from plastic materials derived from petroleum. Can any of you talk to that? What impact is this oil crisis having on your company operations and on your electronic design operations.

Vinny Acampora : In the switch industry, most of our product is a lot of plastic. Between the plastic costs, and gold and silver costs which have been skyrocketing too, it becomes a real big challenge to try to reduce some of the plastics, gold, and silver in our components, so we don’t have to raise our costs to the end customer, because they don’t want to hear raising prices.

Electronic Products: No. The demand is for costs to go down.

Vinny Acampora : That’s right. The whole challenge is the process: to reduce gold and silver, and still keep the high quality and long life that we put in our catalog.

Bruce Strachan : I would say that is the problem side of the issue — how do we manage the rising costs of raw materials in our products? At Infineon, we’re looking at this oil situation and everybody’s trying to envision how to make lemonade from the lemons.

We supply a wide range of microprocessors. We have an extremely wide breadth and capability in power electronics and a wide variety of magnetic sensors. Just to give you an example, I think they say 40% of all the electricity used in the United States is used in electric motors. If we can make those motors more efficient and cut the consumption of power there, we can make a major contribution. It’s good for our business to foresee these kinds of opportunities.

Electronic Products: But electric motors are actually pretty efficient, are they not, in using power?

Bruce Strachan : True. But it depends on what types of motors you’re talking about and what kind of applications they are going into. Also, there’s a growing number of motor-driven applications.

In automotive, for example a lot of motors just run constantly at a certain speed. So if you can modulate the speed that they are running and modulate the time their running based on needs, then you can cut the overall power consumption of the system. That’s the objective.

Electronic Products: So it’s actually the way in which a motor operates and is used, as opposed to just to flat-out efficiency.

Bruce Strachan : Yes. You’re trying to improve the efficiency and you’re trying also to control how it’s being run.

Electronic Products: Right. Is anyone else seeing an impact like in Vinny’s case, where the impact goes right down into the product itself?

George Blazas : It does for us also. We have factories located all around the world that receive raw materials and ship finished goods globally, and it has impacted our costs. Besides fuel, we’ve seen significant cost increases in copper and gold used in the manufacture of our products.

We are constantly trying to find ways to control costs. Can we use less gold in our products while maintaining the same quality standards our customers expect from us, or perhaps redesign components to run on more efficient CNC machines? We’re exactly in the same boat as Vinny describes.

Bob Scannell : I’d like to address the topic in a slightly different way. I think there have already been a lot of comments about how it’s affecting us today. One way we’re thinking about this is looking ahead — what does this mean in terms of product development and directions we want to take in terms of solutions for our customers. And I think it’s getting clearer and clearer that the whole idea of energy resource efficiency is key, resources being fuel, or water, or people, or whatever.

This is an area that sensors can obviously play a critical role in. Within the factory, as has already been mentioned, optimizing on/off time in motors and equipment is one area; using sensors to detect and optimize for predictive maintenance is another.

We’ve already talked about this within the factory, but it can be seen outside the factory as well. In the agricultural setting, as an example, equipment used for seeding, fertilizing, and irrigation, can be improved with sensors to provide better control of resources. These are instruments that can be used — inclinometers, gyros and other sensors — to detect the environment and the exact positioning of equipment; to optimize the amount of water that’s used and avoid runoff, for instance. This is a really interesting area to look at and I think it should influence everyone’s thinking as we look ahead. ■

Electronic Products: George Iwaki referred earlier to the islands of data that exist within factories. What experiences are you having trying to take existing equipment or older equipment and bring that into a new world to modernize and change it?

George Iwaki : I think some of the challenges are that those existing systems have legacy protocols that they use. A lot of the sensors that are used are analog. We’ve introduced a wireless endpoint device that we call the IPserial node; it is a serial node that has an RS-232 connector so that we can pick up some of the industrial protocols and start collecting and networking that data.

Electronic Products: Are you seeing more in the way of robotics in factories today? And what are the challenges for new robotic factories?

Jason Goerges : I can comment a little bit on that. We’re seeing with the ability of motion controllers to have higher and higher data rates and speeds. We’ve really seen an increase in performance of robotic applications requiring complex mathematical calculations. And that’s something again our company has been working on a lot is increasing cycle times — or reducing cycle times and increasing frequencies at which we process our data.

For example, we have systems where we control hexapod robots, delta robots, and we’ve made a conscious effort in the company to increase our data processing capabilities so that we can increase servo band widths and increase the precision that we control robotic systems.

In the past I think the ability to precisely control robotics systems has been limited by processing power. But now for example in our motion controllers we use, X-86 Pentium processors and we often use them to their fullest capability to perform calculations in real time to control robotic systems like that.

Bob Scannell : Robotics is, obviously, is a key area of interest in terms of our motion sensors. There are two segments to look at — motion control and navigation. So I tend to have a broad definition of a robot or a robotic application; it can be basically any kind of remotely operated vehicle for instance; not just in a factory setting.

Electronic Products: And it could certainly be human-assisted robotics as well.

Bob Scannell : Right. So when you look at it from that perspective, motion sensing is obviously key. The other thing to think about here is the types of customers that are doing development in this area and their level of sophistication when working with MEMS. These are things that influence the type of product we manufacture, what interface is chosen and, how system-ready it is.

Electronic Products: I would think that the simpler the interface and the easier it is to use in the application, the less you have to know. And the more that you as a supplier of the sensor can do for the customer the better it is in general, and more rapidly it is adapted to the application. Is that correct?

Bob Scannell : From our point of view, absolutely, with a little bit of caution about taking that too far. But going back to your question about retrofitting to existing industrial systems, that’s a key area. If we can make these relatively easy to use, clients will add additional value onto their existing systems for safety, reliability reasons, or what have you.

And we see that this customer base tends to prefer a much easier to use device. This is particularly true when it comes to MEMS which, let’s say, is relatively unknown to a lot of people. But you also can take that too far, in terms of integration. We have to be very careful about how we partition; we also have to be conscious of where our customer’s IP strengths are and leave room for that.

Electronic Products: So in many instances it’s more like joint design: you can take your interface to their application only so far, to allow them to use their expertise to tailor it to the real needs of their customers in turn.

Bob Scannell : And it all comes down to cost effectiveness. So the more you do that, the more effective.

Electronic Products: You know one of the things in this area that I hear a lot about is the field of mechatronics the idea of a very integrated design of mechanical and electronics. Are you applying mechatronics more in your operations, in your designs, than in the past? Is it an issue for you, are people asking more questions about mechatronics?

Bob Scannell : Yes., we put a lot of product-development effort into the calibration of these devices to make them easier to use. And, calibration, in this case, is no longer just an electrical problem, it’s a motion plus electrical calibration problem.

So, automatically, mechatronics is part of our work here. And we’re certainly seeing that in terms of applications as well. Some are quite interesting, such as the use of gyroscopes to improve the operation of a simple torque wrench.

Electronic Products: That’s very interesting. How does a gyroscope improve the torque wrench?

Bob Scannell : A torque wrench can be used within automotive manufacturing to check fasteners along the assembly line; it’s a safety check to prevent overtorquing.. Placing a gyro at the head of this torque wrench eliminates a couple of layers of uncertainty in this measurement, and basically solves the overtorque problem.

Electronic Products: Does anyone else see mechatronics coming into their area?

Jason Goerges : Well maybe in a different flavor than what we’ve just been talking about. Obviously with the systems we work on, especially as these systems become larger and larger, we work with stage [positioner] manufacturers and we have to get very high performance with our electronic motion control. We see the differences between stage manufacturers in terms of allowed performance in motion control.

Take, for example, flat panel display inspection systems. As pixel sizes are decreasing and platform sizes are increasing, the need for precise mechanics is more and more stressed, because we see increases in requirements for move-and-settle times; we need to be able to move faster and larger distances, while at the same time we need systems that are mechanically sound so that we can settle on a very fine position, down to the nanometer level.

Previously, I think, stage manufacturers have been less in tune to the whole process, whereas now we work very intricately with the mechanical stage designers on large projects. So I guess you could say that there is more of an integration of the mechanics along with our electronic control.

Electronic Products: What kind of tools do you have for co-design in those areas?

Jason Goerges : What we do primarily is electronic simulations and all mechanical stage manufacturers have multiple ways to measure, say, resonances in their stages. When we start a project with a customer, often the specs for mechanical issues are very tight — for example dampening of specific resonances and the granite or other frame bases.

In addition we have a software tool called the Frequency Response Function Analyzer, which simply injects a small sinusoidal current into the motors in a system, and measures the response at the encoders. This measurement provides information about the electronic servo control settings, while at the same time giving a mechanical signature of the machine — revealing resonances and other dynamic behaviors.

And so really we try to accommodate these mechanical manufacturers as they try and reduce imperfections in the system. We receive the specs that they are trying to meet and then we can simulate how well we can control motion with those mechanical systems.

Electronic Products: At this point, let me just throw the floor open for any final observations that any of you gentlemen would like to make.

George Iwaki : I think energy efficiency is going to have a big impact. It’s relevant. It’s immediate. And I think that sensors — to the extent that they can apply more precise control, monitoring, visibility and awareness of energy usage and efficiency on the factory floor and elsewhere — are going to be part of the solution. To that extent I’m definitely excited that there will be an even greater focus on instrumentation and finer, more granular monitoring and control.

Bruce Strachan : One of the things that I wrestled with for many years in working with customers and their systems is location of processors along with sensors and actuators. The question is, do you work with local control or do you work with the centralized control? Is the control distributed or is it centralized? There are tradeoffs between both of those. There are positives and negatives for each one of those approaches. Quite often it has to be solved on a case by case basis.

George Blazas : So between local and of course remote applications I agree it really depends on the specifics. I heard a lot of discussion today about the greater need for more processing, more logic at the point of sensing or control. Certainly that would dictate the kind of network or communications that you would need to send the information to some kind of centralized monitoring function that isn’t necessarily co-located.

So for instance on a — let’s call it a fat node or fat sensor — that is doing quite a bit of processing and sending out data signals which may or may not need to be sent real time. They could be filtered and a histogram sent instead of all the raw data. And that would dictate a certain kind of network. Of course in other applications you may want real time or you may need to get every point of data. And that would perhaps be in a burst.

I guess the additional level of analysis — be it remote or central but also depending on the bandwidth you would need would dictate the kind of network wireless or otherwise — that you would architect to bring that data back to be managed.

Jason Goerges : In applications that we’re starting now and that we’ve been starting recently, we see the need more and more for quicker integration. And with the amount of different vendors and devices that are available, you know, different nodes on a distributed network, we see a lot of companies that will pay a premium for the ability to more easily and more effortlessly integrate these networks.

My company has been focusing on that a lot — the ability to integrate with standards such as CANopen and upcoming versions of EtherCAT. We’re really seeing a trend of companies wanting to have one platform that they can become familiar with. And it becomes much easier when a new prototype comes out to have a simple model for it to be based on.

Vinny Acampora : In the switch industry, we’re seeing a lot of smaller size, logic level, high-temperature materials that can go through the solder process with the lead free solders, a lot of illumination for automation so they can see the switch to determine what position it’s in from a distance. Mostly that’s what we’re seeing in our industry.

We do a lot of work for automation equipment, and industrial automation companies do use a lot of our switches. They’re getting smaller and smaller. And again more logic level instead of the high current like they used to use in the older days. Now they are getting down to logic level.

Bob Scannell : From a sensor standpoint, what I find significant here is just the basic fact that we are having this detailed discussion about the application of sensors within industrial and automation. For years, certain sensors — including MEMS motion sensors — have been well adopted in various other markets. But there have been barriers to adoption in the industrial market .

So the idea of offering devices that are quicker to integrate , easier to use, and less expensive to implement — basically dropping some of these barriers and focusing on these fundamentals — is real key. It’s good to hear all the discussion about it.

Electronic Products: Yes, it is. In a certain sense, the consumer and automotive marketplaces are becoming the proving grounds for various technologies, determining how rugged they are, how good their performance is. And with experience in those markets, some of the things coming out of them are finding their way into real industrial applications, which I think is a new direction for product development. ■