Ethernet-APL: Implications for the field devices of tomorrow

Welcome, Jonas. A real pleasure to have you join us here today.

Jonas Berge: Thank you, Keith. Great to be here.

Larson: Well, thanks for taking the time. Maybe just to start things off, perhaps you can talk a little bit more about just how much more bandwidth and power we're talking about with Ethernet-APL. I think people talk about the bandwidth, and it's going to be faster. But I think compared to analog 4-20mA HART or even fieldbus technologies, there's a lot more power that's reaching instruments as well.

Berge: Yeah. So let's start with the bandwidth numbers. So, Ethernet-APL runs at 10 megabits per second, while HART runs at 1.2 kilobits per second. So, APL really is more than 8000 times faster. And fieldbus runs at 31.25 kilobits per second, so APL is still more than 300 times faster than that. So, APL will be able to carry a lot more data, which enables more devices on a network and it also enables other types of devices with higher bandwidth requirements.

So for the power supply, I'm glad you brought this up, because it's important but often forgotten, as we get all excited about bandwidth, right? So, yeah. Ethernet-APL provides not only communication but also power of the same two wires, just like 4-20mA, HART loop powered and fieldbus bus power. So, for two-wired, loop-powered HART devices, their power is limited by the 4-20mA signals. They must be able to operate on less than four milliamps, which is not much, and it limits what type of devices can be two-wire, loop-powered.

Now, depending on the application, APL provides more than half a watt or more than 1 watt of power to a device, which is almost, like, 10 to 20 times that of 4-20mA devices. So, devices in the past, you know, those that had to be separately powered in the past, will, in the future, actually be powered over the same single pair of wires. No separate power cable needs to be pulled.

Larson: Yeah. So, what sort of possibilities does this open up for existing devices? Can current devices from flowmeters to DP or temperature transmitters? What kind of new things are possible with more power other than, you know, obviously, all the secondary variables that can come in over the APL signal? But are there other new kinds of capabilities that we can power with this new technology?

Berge: Sure, yeah. So, well, 4-milliamp is sufficient to loop-power manual, the familiar instrument types, like pressure, temperature level, and some flow transmitters like DP and vortex, as well as liquid analyzers like pH and conductivity. But 4-milliamp is not sufficient to loop-power certain other types of flowmeters like Coriolis mass flow and mag flow, for instance. So, those today need separate power, at least for the larger sizes.

But yeah, in the future, with APL, many of these device types, but not all, could be two-wire powered, thus, simplifying installation. But even devices which already are two-wire power today can benefit from more power provided by APL. More power means more powerful microprocessors, which enables more advanced signal processing, which benefits some measurement types such as those based on radar and ultrasound, which become more accurate and more robust with better signal processing.

So, yeah, more power also means faster measurement updates for these measurements. Yeah. And then also, you know, for the type of instrumentation such as transmitters and valve positioners we are familiar with, the greater bandwidth that you also get with APL means higher speed. So, you can put more devices on a network than with traditional fieldbus and still get faster response time.

So yeah, you don't need as many networks as you did with fieldbus. It also gives you a greater spare capacity and flexibility to later add more devices to a network without having to re-engineer, to add another network and junction box. You know, APL provides greater ability to accommodate changes late in greenfield projects, or even after startup, you know, without extensive re-engineering. And such flexibility is critical to initial project execution. And later for plants to stay evergreen. That's an important outcome.

Larson: Absolutely. Maybe we can talk a little bit, dig into a little bit more about the types of capabilities that represents, maybe. Emerson is obviously known for its Coriolis meter. More power and bandwidth, I think that would be more sensitive, right? Because you can actually have more power to energize the sensor to the tubes as they vibrate. Is that right?

Berge: Yeah. And so that's, you know, because the... I mean, the tubes need power, right? So, that's why today they are, at least the larger sizes need a lot of power, right? So, they have to be separately powered to get that sensitivity and accuracy, right? You know, but with APL, you can get that sufficient power over the same two wires. So, you can get two-wire devices, which are more accurate, and up to larger sizes, you know, and so on, thanks to APL. You know, yeah. So, this is a very good example of what you can do with APL. And the same thing for mag flow, right? They also need to energize the flow tube there.

Larson: Yeah, that makes sense. What about in terms of devices that tend to have some kind of a spectrum capability, like a radar tank gauge where you're dealing with a lot of data? I mean, that's the kind of data that stays stranded in the transmitter. Nowadays, you're not seeing the spectrum of how the radar signal or other technologies or vibration type of instrumentation. Does this open up better abilities to measure processes or do diagnostics for those kinds of instruments?

Berge: Absolutely. You know, so for a radar level transmitter, more power means that it can use, you know, the frequency modulated continuous wave, or FMCW, sensing technology instead of the simpler pulsed radar. So, FMCW has greater temperature stability and is better at handling disturbing echoes, particularly in tall tanks. So, FMCW, does require, you know, advanced signal processing, right? But with the power from APL, transmitters will be able to provide a faster updates rate than in the past.

And you also benefit from high bandwidth. You know, during the instrument commissioning, you know, the radar commissioning, you don't have to wait for the radio-echo curve to kind of slowly load in your setup app, right? It comes instantly, and it can actually stream live like live video. You can adjust the installation and see how the echo curve dynamically changes accordingly, right? And you can see the services from agitators inside the tank. So, easier to commission.

Larson: Yeah. Wow, that sounds like that would definitely cut some startup time off a project, for sure. Yeah.

Berge: And I think another good one is, like, gas analyzers, right? You know, that might also be possible. Imagine, for instance, installing the gas analyzers in the field, near the process sampling point. So, we'd man minimal sampling lines, instead of running APL back to the people that need the data, instead, you use APL to run the data back to the people, right? So, this is a whole lot easier to build and maintain than constructing analyzer shelters and lay and maintain long sampling lines, you know, the traditional way.

Larson: Yeah, that makes sense. We haven't really talked about some of the whole movement towards the industrial IoT or doing more diagnostics, process diagnostics as well instrument diagnostics. Do you see a new wave of being able to handle those type of data loads, much larger data loads more easily?

Berge: Yeah. Well, I mean, thanks to digital transformation, IoT and industry 4.0, and fourth industrial revolution, right, there's a huge interest in equipment condition monitoring and analytics, right? And not just for the large critical turbomachinery, but, you know, turbo mesh, turbines and compressors already have vibration monitoring, and in some cases, prediction, right? You know, which is very successful. But these online systems are very expensive, right? So, plants now want to do that kind of condition monitoring on their second-tier essential assets, like pumps and fans and conveyors, etc., as well. But at lower cost, since it's less-demanding applications.

So, some of this is done by wireless transmitters, as you know. But for some assets, you know, plants now deploy asset monitors, which are edge devices sitting in the field monitoring multiple assets. So, today, these use regular Ethernet with separate power, and they do help plants to move to predictive maintenance for a much greater portion of their equipment. But yeah, in the future, these could use APL to be even easier to deploy.

Larson: Yeah. Are there new types of instruments that maybe we're not used to envisioning in the process field that this opens up as well?

Berge: Well, let's think about, you know, bear with me here for a minute. Okay. So, let's think about ethernet in the office. It's not just connecting laptops and servers to the Internet. Ethernet also connects printers and scanners, and, in our office, you know, the phones are on the Ethernet, CCTV cameras are on Ethernet. The video wall display is on Ethernet, and I'm sure there's more, right? So, the Ethernet is...

Larson: This GoToMeeting session that we're doing right now is on it.

Berge: Exactly, you know. So, Ethernet is the single common infrastructure, right? And there are no separate phone lines anymore, no separate CCTV cabling, and so on. You know, a single Ethernet infrastructure, instead takes the place of maybe five separate infrastructures in the past. You know, so, Ethernet-APL can do the same for process area in the plant. So, APL has potential to not only take the place of 4-20 mA infrastructure, but also the RS45 infrastructure, the video cabling infrastructure, and regular Ethernet infrastructure in the plant area into a single common APL infrastructure instead of five dedicated infrastructures that we have today. You know, so a dramatic simplification.

You know, now, so to answer your question with that, for instance so many plants, they want to be able to see foaming or pebble-sized or stratifications in their process, right? So today, they send field operators out to check, peeping through an inspection window. In the future, they can just drop in an APL video camera. I'm not talking about your perimeter CCTV here, that will continue to use regular Ethernet.

But, you know, I'm talking about video playing a more important role in the process area in the future. So, you could easily drop in an APL HMI panel, for instance if that would help you in certain tasks. And perhaps we will see LiDAR in-process applications. So, some of these things are theoretically possible today, you just need to lay 100 meters or a few 100 meters of Ethernet fiber, right, and power for that to one device. You know, it's just too costly for a single-use case, so it doesn't get approved.

But once you have APL in every corner of the plant, dropping in these cameras, HMI panels, LiDAR, acid monitors, flow meters and gas analyzers, it becomes easy and low cost, paving the way for that continuous improvement.

Larson: That makes sense. I do think one of the neat things is that it accommodates multiple different protocols, since you're really only talking about the physical layer with APL. From a practical perspective, do you think we'll have, being a very conservative industry, will there be an instinct to say, "Okay, we want just control on these segments," but, you know, there may be a separately segmented network over here for say electrical equipment running, you know, maybe Profinet over there versus the HART-IP over for the instrumentation and control networks? What's your sense on that? Or is there sufficient bandwidth that we probably wouldn't even worry about? I'm just kind of curious what your take on that is.

Berge: Well, there's plenty of bandwidth, right? And Ethernet-APL supports IP. And with that, yeah, it enables multiple application protocols like you said. So with APL, it's no longer HART Modbus or Profibus. It's really HART-IP and Modbus TCP and Profinet, plus other protocols, such as HTTPS for web browsing, or RTSP for video, right? All on the same wire at the same time, and there will also be, obviously, a way to HMI panels and analyzers.

But yeah, having said that, it's good engineering practice to have subnets and security zones, right? So, there will be a subnet for the electrical gear in that room, right? And most of the traffic on that subnet might be Profinet, like you say. But it will not be purely a single protocol like in the bus days, right? You know, that there will also be HTTP and OPC-UA at the same time. And, and oh, by the way, this kind of electrical gear might continue to use regular Ethernet since MCC is usually indoor with short distances.

Larson: Oh, that's true too.

Berge: But yeah. Yeah, APL instrumentation will most likely use HART-IP. But yeah, the subnets for instrumentation will also not just be a single protocol like that. There will be HART-IP, sorry HTTP and OPC UA at the same time, right? So, the mix of devices and use cases, sharing the same infrastructure and the mix of software, tapping into their data, really necessitates a mix of protocols, you know. So, the benefit of Ethernet is really the shared infrastructure, and you can't do that without mixing protocol.

Larson: I think it'll be just kind of a mind shift required for people to take that in this environment, you know, that they've become used to in the office environment, and at home, obviously, as well to say, "Okay, yeah, it's going to work here, too." So, it might take a little time, but it makes sense.

Berge: Yeah. Yeah. In fact, you can apply the office analogy here, too, right? The beauty of IP is that one protocol doesn't have to do everything, right? You know, both in the office and at home, right, we use HTTP for web browsing, and FTP for file transfer, and IMAP for mail. Each protocol is specialized for a particular function, and yeah, APL will be just like that.

Larson: Yeah. Great. From an installed base perspective, we've been talking a little bit or touching on the HART-IP, which is the global community of instrumentation. Technicians, engineers are certainly more comfortable with the HART protocol when it comes to managing and communicating with smart instruments than other protocols. I mean, it seems like we're on the verge of some competing interests, I would say, at the protocol level for Ethernet-APL with people advocating for HART-IP, obviously, but also Ethernet as well as, you know, OPC, perhaps going straight to OPC field-level communications. And even ethernet IP, meaning the ODVA version of Ethernet IP.

What's your sense in how that's going to play out in terms of the landscape for protocols when it comes to, especially the process instrumentation side of things, which is our bread and butter in the controls community?

Berge: Yeah. Well, I mean, the protocols can exist, right? So, but think about the mix of device signal pipes, right? So, let's right now, you're planning to build a plant based on APL, let's say. And we will see APL in DCS and PLC very soon. So, no problem, you know, for the core process control of the main plant to go APL. But what about the safety system, the firing gas system and the machinery protection? I think they will be slower to adopt APL. So for now, they have to use 4-20 mA power devices.

And I also believe packet unit vendors, like boiler manufacturers, will take a while to adopt APL. So, they will continue to use modules instrumented with 4-20 mA HART for now, right? So, and like we discussed, APL makes more sense for certain instrument types than others. So, it'll be a while before everything is available in APL. And let's not forget about wireless sensors, right? So, plants now use wireless sensors instead of portable testers for vibration and ultrasound leak test and ultrasound thickness for corrosion inspection, instead of temperature guns and so on.

And plants also put in wireless sensors instead of mechanical pressure and temperature gauges, side level glasses, and variable flowmeters, etc. So, these wireless sensors use wirelessHART, right? So, clearly it becomes a lot easier to manage this mix of instrumentation if APL devices use HART-IP, right? Because HART-IP, 4-20 mA HART, and wirelessHART, they're all HART. They all work and look and feel the same way, right? So, this makes the job of the instrument engineers and technicians so much easier.

So, because lots of functionality and terminology is common, right? So, and oh, by the way, you can run HART-IP all the way to the cloud. We run HART-IP from wireless gateways in the plant to analytics in the cloud without converting to intermediate protocol.

Larson: Yeah, that makes a lot of sense. It really does.

Berge: Maybe on another note there if we've got time?

Larson: Yeah.

Berge: You know, faster sampling time, also means faster control loops, which is better control. So, faster response time also means that approved APL devices could, in the future, be used for functional safety. So with digital, we can avoid covert failures associated with analog signals. So, high speed enables safety instrumented functions, fire and gas and machinery protection systems that currently are hardwired can go digital. That is, you know, so really all plant systems can go on APL, not just the basic process control.

And APL also means many devices will have embedded web servers, you know. So, you can configure, diagnose and otherwise manage these devices from a regular web browser.

Larson: Yeah. and I think it's also clear that all those 4-20 mA HART devices aren't going away anytime soon. I mean, you look at the rate of new plant construction here, certainly in the U.S., there's not a lot of new plants going up. So, HART is probably going to be around, or 4-20 mA HART, I should say, is going to be around for a matter of decades. So, having some commonality of a mixed environment is certainly a selling point for using HART-IP in those situations, as well as the wireless, as you were mentioning.

Berge: Yeah. So then, what users should do is, I mean, they are constantly upgrading their DCS, right? You do that every couple of years, you upgrade your DCS, you upgrade your device management software. Make sure that you upgrade your DCS to support HART-IP, so that you can benefit from HART-IP in your existing DCS, because HART IP, it's not just for the APL devices. You know, HART IP can be used in HART multiplexers also that some plants have, and it's also used in wirelessHART gateways. So, you don't wait until you're building your next plant to upgrade to HART-IP in your existing plant. And many users actually have HART-IP already, but they don't even know it. You know, it's just ticking away there in the background in all of their wirelessHART gateways, actually.

Larson: Yeah, it would make sense that that plays a role in many of the kind of configurable I/O solutions that are out there as well that are extracting our data on a regular basis from those 4-20 mA signals already. So, I would imagine in some DCSs, that's in the background functioning as well. Yeah.

Berge: Yeah, it is. And we're also seeing it in like, safety logic solvers, and so on, too. So now, HART-IP is out there. It's actually been around for more than 10 years, but it's a very well-kept secret, really.

Larson: Any other recommendations for users or folks at engineering firms out there listening about how they should prepare their current systems or their next designs for this coming revolution, I should say?

Berge: No. Well, no. But I think the main thing, okay, is to keep in mind that we are moving away from a single protocol. You know, that was a bus type of thinking. Now, we're moving into multiple protocols. Okay? So yeah, make sure to use the right application protocols for the right application. Okay? So that could mean like your motor controllers might be using Profitnet or Modbus TCP, more OPC UA, while your instrumentation uses HART-IP. You know, and with the modern DCS, that's not going to be any problem at all, because they support all of these protocols. You know, so no issue.

Larson: Alright, great. Well, it's going to be exciting to see how this new technology transforms what's possible. We've been through a lot over these years where we both date back to fieldbus days or before fieldbus days. So, it's going to be interesting to see how things continue to change, and there's certainly no lack of change in this, what used to be a somewhat sleepy industry. But, boy, things are changing fast now, aren't they?

Berge: Yeah. So, on that note, APL actually has two important things going for it. Okay? The first one being that, with fieldbus, everything was really connected very much in parallel, right? So, and some users might have experienced that. You know, if noise makes its way in on the bus, you can't really know at which device it entered, so it made troubleshooting quite challenging. Okay? The beauty with APL is that the switch, is really, like, well, it's a switch, right? So, each device, it actually sits, it's pretty much on its own network, or spur, if you will, right? And what happens is that if noise makes its way into one device, it only affects that device. It doesn't propagate to the other ones. Okay?

So first of all, the network becomes very robust, and secondly, it becomes very evident where the noise is getting in. So, it becomes easier to troubleshoot. So, that's one great advantage of APL.

The second one being is that the FDI package, you know, that device description file, that was always a big headache to get that, right? With APL devices that can be stored on board devices and automatically uploaded into the system, so you don't have to search the web to extract the right file for your particular device and model and version and all of that. You know, that's dramatically simplified.

And, oh, by the way, similarly, you can just upload the instruction manual from the device. That can also be on board with the device. So, perhaps the two biggest points there, we learned a lot from fieldbus, and we made an APL better in that sense.

Larson: So, instruments on Ethernet really hold the promise to have a higher level of self-configuration, and things we take for granted now with, say, a printer at home or in the office. Instruments will be a lot more like that in terms of having a layer of self-recognition. And, you know, you don't have to go dig out the CD ROM with the device driver, and that sort of thing, on these kind of things. So, you really see that happening in the instrument realm, as well.

Berge: Yeah, because that can now, it's shipped on the device, and thanks to the abundant bandwidth, the system uploads it. And so yeah, it solves that type of problem.

Larson: Yeah. One of the things you were mentioning about diagnostics for the noise levels on the different spurs, I understand that some of the chip level manufacturers, they can actually tell, like, if there's a fault in a wire. You know, somewhere along this kilometer, 1000-meter length, they can actually pinpoint exactly where that fault is within a few meters of it. So, if you're having problems with that wiring, not only is it just a bad wire, but they can actually really have good diagnostics around where that fault might be, which is another layer of diagnostic information that you may not have in traditional schemes.

Berge: Yeah. I mean, the silicon has come so far. You know, and yeah, and thanks to, I mean, even like network switches for APL, I mean, they will have embedded web servers just like they do for regular Ethernet and Wi-Fi, right? So yeah, so that's how you can access that type of diagnostics, right? So in the future, you will be troubleshooting with a web browser and not the test pen or a multimeter.

Larson: Yeah, yeah, that makes a ton of sense. Any other last comments, Jonas? I really appreciate you taking the time. Any other things that our listeners should remember, keep in mind?

Berge: No. So, the main thing is, you know, design for multiple protocols. And, yeah, make life easy for the people that are going to the plant, you know by using HART-IP.

Larson: That sounds great. Well, thanks so much, Jonas, for sharing your perspective with us today. It's really been thought-provoking and a real pleasure as always. Again, our guest today has been Jonas Berge, senior director of applied technology for Emerson. Thank you, all listeners, for tuning in. And thanks also to Emerson for sponsoring this episode. Once again, Jonas, thank you for joining us.

Berge: Thank you.

Larson: I'm Keith Larson. You've been listening to a Control Amplified podcast. And if you've enjoyed this episode, you can subscribe at the iTunes Store and at Google Podcasts. Plus, you can find the full archive of past episodes online at controlglobal.com. Signing off until next time.

For more, tune into Control Amplified: The Process Automation Podcast.