Thanks for joining me today. My name is Jenn van Steig, and I am responsible for industrial Ethernet in TI's ethernet group. Today I would want to tell you a little bit about some advances in single-pair Ethernet I have some really exciting information to share with you during the course of the presentation. But it's not just about that. I'm going to share actual learning information with you as well.

So what we're going to do through the course of the conversation is talk about where Ethernet fits today in industrial communications applications. I'm going to talk about why single-pair Ethernet has been invented. Why it's been necessary to develop, and what its differences are with what I'll call standard Ethernet that exists already today. Then I'll talk about how it helps smart factories and most industrial communications applications, because it's able to move data much faster and much farther than has been possible before.

And at the end of the presentation, we have a slide available where there are several links you can visit afterwards. And you can go back and read more information about what you're most interested in. And that includes my previous webinar on industrial Ethernet, just in case any of the concepts in today's material leave any of those finer industrial Ethernet details out.

So again, thank you for joining me. And let's start talking about it. First of all, what is Ethernet? That might scare you a little bit. You might think this is going to be a very rudimentary presentation. But the fact that you might think that is exactly why it's such a great choice in today's industrial applications.

Ethernet is well established and understood in many communication networking protocols. It's used at multiple levels of enterprise communication, domestic communication, and all over the world today. With the increase in big data, the internet of things spreading into industrial applications, what we're seeing is that existing communications networks in industrial applications are being pushed to their limits. And so we need to make some improvements.

That data, as I mentioned a moment ago, needs to be able to move both faster and farther, not just one or the other. And you'll learn why that's important in a second. This has to happen without increasing complexity and, preferably, without increasing cost. In some cases, specific requirements need to be met.

An example of this might be intrinsic safety requirements in volatile environments. And then being able to do all of these things with a protocol that is as well understood and well established as Ethernet is extremely beneficial. Because in the long term, what it can do is make the extension to Ethernet across the entire range, all the way to the edge node-- it makes it easier to program. Because Ethernet does use TC PIP socket programming. Or you could look at that as internet programming from the get go.

And so making your systems more connected, the industrial Internet of Things becomes a lot simpler from a firmware perspective. The benefits in the long term include being able to monitor the system's status real time. And you can check its health. This will help provide more effective, predictive maintenance. And indeed, it could potentially provide the possibility for better throughput and, therefore, lower costs throughout the system. Lower cost per unit.

So let's level set a little bit, and take a look at how this communication is performed today. What you see here is a simplified version of many factory automation systems. We have the factory level, the control level, and the field level. And of course, at the top of the factory level, it's how and where most of the decisions are made.

You likely know this already, but ERP is enterprise resource planning system. And so that's where all the planning occurs. And MES is manufacturing execution system. And so that is what is actually executing all of the plans already made in ERP.

So in the factory level, as you know, most of the communications to the cloud and to storage are performed already using Ethernet. And then moving onto the control level, with recent improvements in the determinism of Ethernet, the communication, on to the actual field level, where things happen. That has also become industrial Ethernet.

Now what generally occurs is, at the field level, a lot of these different sensors, actuators, transducers require different levels of communication. And so, all of these really good, clever protocols have emerged over the years to be able to support those requirements. And they've done a very good job doing that.

Now the challenge is that the systems have become quite complex. And it requires a protocol conversion between the field level and the control level, whichever direction the data might be going in. And of course, that's exactly what it sounds like.

It may be a conversion between HART and Ethernet that occurs. And so adding this additional step adds complexity and another failure point of potential in the system. So this fragmentation has been necessary today. And in addition, the reason that Ethernet has not been used beyond these particular gateways in the past has largely been because--

Well, a couple of reasons. One is that Ethernet traditionally requires four-wire or either-wire cabling harnesses. So those are a little heavy, a little more expensive, than the two-wire harnesses available for the protocols that you see here. Another reason that it has not gone into these kinds of applications is that often these protocols are in place. They're working really well. And the budget isn't there necessarily to change from two-wire to something of larger count.

Also Ethernet traditionally has only reached up to 100 meters. And I'll tell you a little bit more in a sec about the distance and bandwidth trade-off for a lot of these various protocols. So what is often the case with these different industrial protocols, fieldbus protocols, is that they have to trade off between bandwidth and distance.

In other words, they might be used at high bandwidth, but they can't drive that signal very far. Or they might be used at lower bandwidth, and they can go a long distance, a very long distance indeed. Let me give you an example. CC Link has a long reach capability. And it can go up to 1.2 kilometers. And its fastest bandwidth is 10 megabits per second, so just over a kilometer and 10 megabits per second.

But the challenge is, you can't get both. You can either get a distance of 1.2 kilometers, but its data rate will be down to 156 kilobits per second. Or you get that higher data rate of 10 megabits per second, but it can only drive 100 meters of cable. So in summary, it's not able to do both. It can't drive 1.2 kilometers at 10 megabits per second.

So that too is becoming something that needs to be overcome in industrial systems. Because, of course, more and more data is needed to support the predictive maintenance that's going to be needed to support the industrial Internet of Things, or industry 4.0. So it's important that we came up with a way to be able to perform both high bandwidth and long distance. And even better, wouldn't it be great if we could keep these particular protocols, whatever we replace these with, on two wires instead of having to upgrade the wiring system. OK, I'm going to change the tone a little bit here. So single-pair Ethernet is going to be able to provide all the things I just mentioned. But let me make sure we all know what single-pair Ethernet is exactly. So regular Ethernet-- if you think about 10 100, or one gig-fi require multiple twisted pairs. So 10 100 requires four twisted pairs. Gig-fi's require-- sorry, two twisted pairs, I beg your pardon. Gig-fi's require four twisted pairs.

Now single-pair ethernet is full [? g-plex ?] communication. And it is capable of performing this with just a single twisted pair. So this provides the potential for significant savings. In both the amount of copper that has to be purchased in the first place, or if preferred, there's potential for reuse of existing wiring harnesses.

And just as another point here, of course, the specification, the i Tripoli 802 dot 3 CG spec does also allow for power in the form of power over data line to be coupled with-- onto the same pair of flyers with some external circuitry.

Let's go back to what we were talking about in terms of the trade-off between existing field buses with their reach and their data rates, as well. So here's the CC link example that I just gave you a moment ago. So this column shows you the service that each of these protocols is able to reach. And then this column here shows you the fastest this protocol is able to transmit and receive.

And down here, what you see, is the new single-pair Ethernet specifications. And so, what we have, of course, is the gig-fi level, the 100-meg-fi level, and then today, this is really the star of the show that IEEE802.3cg. In particular, this conversation is about the 10BASET1L, or long reach, version of this spec. And it has two specifications.

One is that the device, the T1L device must be able to transmit and receive data, full-duplex data, at 10 megabits per second over 200 meters. And that's at 1.0 volts. Then there's a second mode where maintaining that rate-- you can still transmit data at a higher voltage, 2.5 volts, over 1 kilometer.

So again, with 10BASET1L device, and this is IEEE spec, the device needs to be able to go all the way to a kilometer and still keep its speed of 10 megabits per second. So that, of course, is constant bandwidth with cable reach.

In addition to that, there are several additional benefits, most of which I've touched on before. But I just want to remind you once again that there are reduced failure points, reduced complexity, because of the lack of gateways, the lack of protocol conversion now needed. And the ability to potentially reuse cable infrastructure. And I'll give you an example of that really quickly. And of course, it's already in the form where your developing firmware for internet sockets.

So to give you examples of where the two-wire or the twisted pair could be reused. If you look where this red box is up here, you can see four examples of where there's really high potential for cabling reuse. And again, that could be helpful if you have a system, and it needs to have that greater bandwidth.

But perhaps you don't have the budget to replace all of the cabling. That's OK. You can do it gradually. I'll give you an example of that too in a moment. So that you aren't faced with a very large restructuring situation that takes a long time and a lot of money. So there are several different areas in which this can be used. Let's talk about that a little bit.

So today-- actually this has moved already. Most field buses are Ethernet now, except for down in the actual sensor and actuated level, where HART is about 50% of the use cases in factory automation. So-- sorry, in process automation, I beg your pardon. And lots of different kinds of protocols are used in building automation and in factory automation.

So to be able to provide this killer combination in the single-pair Ethernet device that we're talking about, we have developed a device-- I'm going to come back to this slide-- that is able to do some really amazing things. Now first of all, it has extremely low power consumption, 45 milliwatts.

Now remember the specification I showed you in IEEE802.3cg just a moment ago. And that is that the device can reach 1,000 meters when being used at 2.4 volts, or 200 meters at one volt. This device I'm introducing to you today, the DP83TD510 is capable of doing much, much more.

This device is capable of driving up to 1.7 meters at either 2.5 volts, but even more remarkably, at 1.0 volts as well. Let me repeat that, because it's very remarkable. The cable reach of DP83TD510 goes up to 1.7 kilometers, not just for 2.4 volt operation, but also for the lower 1.0 volt operation.

The additional applications possible because of this-- well, really they're just limited by an engineer's imagination. In addition to these very good performance characteristics, as always TI's Ethernet PHYs are known for their robustness and their long cable reach. This is another one of TI's very robust Ethernet PHY. It just happens to be a single-pair Ethernet PHY.

And of course because of the lengths that's possible to be driven with these options available through the device, the diagnostics start to become even more important. Here's what I mean by that. Imagine you have a cable that is going up a tower on a mountaintop somewhere. Now think about the cost involved in going up there and physically having to find where, perhaps a short or a break in the cable might be. It could even involve a helicopter.

Instead imagine being able to go to a base [INAUDIBLE] more than a kilometer away. And go in there and run some diagnostics and check to see where the issues are occurring. Then you suddenly have the information you need to be able to implement any physical maintenance, if that's necessary, down to a small area within the cabling or within the device. Or you can, of course, perform any kind of remote firmware, software upgrades that are needed.

So all in all, the 510 is very unusual with its exceptional cable reach. The diagnostics make it much, much easier to use. And of course, the very low power consumption provides either the ability to reduce overall the amount of power consumed in the end equipment, which is very convenient if you're trying to meet some kind of energy consumption standard. Or if you're trying to move some of that power budget elsewhere in the system, this lower power enables you to do that.

And of course, if you have very high-precision device somewhere else in the system, that's going to draw more power. Or sometimes more economical devices may not be so high performance, but they're cheaper. So you may be wanting to do that as well.

So now I'm going to go back to this example here. The elevator. Here's an example that I haven't really talked about just yet. An elevator today uses a traveling cable, often with can, and its length is generally up to about half a kilometer. Of course, with this the length and the bandwidths are somewhat limited.

So essential things can be transmitted along this moving cable. However, additional amenities, perhaps additional amenities, if the car breaks down-- let's hope that doesn't happen-- but if it does, it would be very nice to be able to still have some power in there, for example, which you can with PoDL.

And then, as I mentioned, there's a good potential for cable reuse. So if you don't want to go ahead and change absolutely everything in the system at greater expense in the beginning, it's perfectly possible to add a media converter at the elevator car. And then you can reuse the traveling cable that's in the system to be able to implement single-pair Ethernet until you get the benefit of more data over long distance, potentially longer than half a kilometer, as needed.

And then when you have that established, and you feel that you're ready to actually move on, you can replace the entire system where there's no longer a need for media converters. You can simply go from single-pair Ethernet to single-pair Ethernet.

I'd like to show you an overview of where this device fits into TI's portfolio of products. We do have what I called standard Ethernet at the beginning of the presentation, and single-pair Ethernet available in all flavors. So let me go through these at breakneck speed. But the overarching commonality between all of our Ethernet PHYs is our robustness, and often our long cable reach as well.

That is what you can think of when you think of TI's Ethernet PHYs. And we have them in '10 100 and gig levels, at the moment. And generally speaking, there should be a flavor that works for you. In single-pair Ethernet, we have just released to the web our 100BASE T1 device. So that's the gig-level single-pair Ethernet. And so it's possible to go there, and read more about it. It is kin to kin without existing 100-base device, the 811.

So what that means is, if you are already using the 811, it becomes easier to upgrade to a higher speed, with a 720. Because the [INAUDIBLE] is similar. And of course, my favorite-- you can probably tell that this is my baby, the DP83TD510, which is the 10BASET1L, or long-reach device, which also is available today on the web.

Let's go back again to a couple of the specifications of the device. I wanted to show you this, because I want you to know what the design team was targeting, and what they actually worked to attaining. So at the top here, you can see that there are three different configurations. The lowest power configuration is this one here. And the higher power but furthest to reach configuration is this one here.

And in gray, you can see what we tasked the design team with hitting, in terms of the target specification. So low, lowest, and furthest reach over here. What they were able to do is actually beat all of these requirements. And most notably, the very low power consumption, that we were able to beat the target by 30, which is exceptional.

And as I mentioned before, we were also able to beat the target of meeting the IEEE802.3cg distance spec significantly by being able, in both modes-- again, I know I'm harping on this, but the one volt, being able to drive 1.7 kilometers is a very significant advancement in single-pair Ethernet technology.

So what I'd like to do is quickly show you a video of the 510 actually driving a kilometer of cabling here, so you get a sense of how it's possible to do so. So if you'll just stand by one moment, I will go and open that up.

Today we're introducing the [INAUDIBLE] Based on the IEEE802.3cg standard. This is a 10BASETIL5. On one side of our EDF, we have an external termination for intrinsic safety. On the other side, we have a DP83822 for media conversions to standard Ethernet.

Our test today is a smart-phase link to the DB 83822 to the 510, through 1,000 meters of single-twisted pair Ethernet. Turning it into a second 510 EDF through the external termination to the DP83822 standard Ethernet back to [INAUDIBLE]. Reflected over 17 million packets error free.

So if you turn your head a little bit, you can actually see this. This is what Justin, our applications engineer who filmed this, is talking about. You can see the 17 million units going through error free, which is a great demonstration of the performance of this device. So let me review real quick what I've gone over today with you.

So the 510, the DP83TD510 is going to give you more than you originally would have expected from an 802.3cg long-reach PHY. So it will provide both you and your end customer with a better experience that-- it will be better both immediately during design and testing, and then, of course, later in terms of maintaining the end product by the end user, your end customer.

Again, this device is able-- it's really able to take data much faster and way farther. It does ease upgrades, because of its ability to reuse existing cable wherever possible. And if there's a system where fiber has been used for the distance in the past, then it's actually possible to move over to more economical copper, a two-wire twisted pair, to save costs should that fiber need to be replaced at any time.

In addition it simplifies the entire system. Because you're doing everything in Ethernet. So if you think about, with so many of us working from home at the moment, being able to simply connect to the most remote node in your industrial communication network, without having to go through any protocol conversions, being able to perform all the things that you can via Ethernet remotely, is an extremely beneficial feature of single-pair Ethernet.

So once again, , single-pair Ethernet is enabling big data into industrial systems to improve the end customer's experience, to make it much easier to both develop, maintain, and have throughput of your particular system. And again, really the use cases for single-pair Ethernet are limited only by your imagination.

Here you see an evaluation module available to you. This is our media converter version. And what it does is, it allows you to connect through just simple two-wire here. It's not a specific connector. You can use whatever kind of cable you want.

And then over here, you connect these two boards together. This is the 510 board. You connect these two boards together, and the DP83822 10 100 PHY is present to convert the signal to standard Ethernet. And here's the jack. You can just plug right in here and perform your evaluation, without needing additional infrastructure to evaluate single-pair Ethernet.

It can be powered onboard or externally. Even though we do have external terminations for the 510 to enable things like PoDL or other implementations, we've actually included them on this board here just to simplify and enable you to test and evaluate the actual data transmission of the system.

With all that, there are a lot of different resources for you on TI.com for a single-pair Ethernet. Again we now have a 10-megabit device, a 100-megabit device, and 1,000-megabit device available to you on TI.com. You can go there. Samples of the device are available. EDNs are available for all of the devices. And then more information in each case is available through all of these different technical documents.

We do have reference designs already, that use single-pair Ethernet. You can see this up here. And then to be able to find-- if you get an EVN-- to be able to find the GUI that goes with it, you can just click on this thing. And then, of course, we do like to provide Linux drivers with all of our Ethernet PHYs. You can click here to go directly to that site. So with that, everyone, I thank you so much for your time today. That concludes my presentation on 10BASETIL, what it does, and what TI has to offer in single-pair Ethernet.

Thank you so much. Everybody, just a reminder, you can find this archived recording on TI.com/npu by the end of the day today. And you can also go to that site and see the topics for the next few weeks that we have coming up. And hopefully we see you next Thursday. Thank you so much.