ON Semiconductor Corporation (ON) Earnings Call Transcript & Summary

Everyone, and welcome to today's technology webinar on inductive position sensors for industrial and transportation markets brought to by onsemi. I'm Kyli Miller with onsemi, and I'll be your moderator today. This webinar will show how onsemi's NCS32100 and NCV77320 sensors bring higher speed and accuracy in harsh industrial and automotive environments. At the end of the webinar, we'll be holding a Q&A session to answer any questions you may have. [Operator Instructions] I will also share the slide deck of the webinar with you during the Q&A session. This webinar will be recorded and posted on onsemi.com. You will be notified via e-mail when the recording is available. Now let's meet today's presenters. Bob Card works at onsemi as an ASG Marketing Manager. Bob has over 30 years' experience in the semiconductor industry, along with 12 patents and multiple publications and blogs. Now let's get started with the webinar.

Thank you very much, Kyli. My name is Bob Card. As Kyli mentioned, ASG Marketing Manager for the Americas. And today, we're going to talk about our inductive position sensors for industrial and transportation markets. I'm going to start this off with a press release we did in collaboration with HELLA back in April -- late April, where we announced our 1 billionth inductive sensor with HELLA, over 2 decades, specifically for the automotive x-by-wire systems, specifically for things like accelerator pedal sensing, steering and torque sensors as well as actuators for pressure boost and turbos. Towards the end of the slide -- presentation, I mean, I'll talk a little bit more about the x-by-wire system applications. But the intent now is I just want to point out that back then, for the past 25 years, our strategy has been ASIC business with onsemi and HELLA, which is application-specific ICs. And what we're doing is we're unleashing that all that we've learned to the mass market today and tomorrow. So this is the beginning of these 2 first products which we'll be talking about, is the beginning of many products that we'll be bringing to the mass market for inductive position sensing encoders. So for the markets for transportation, markets and applications for encoders, there's many that includes automotive, agriculture, machinery, mining machinery, rail, construction machinery, forestry machinery, autonomous mobile robots or AMRs for warehousing, marine shipping, drones using for that gimbal to keep that camera steady for the high-resolution imaging and zooming in and out while the drone is moving. And one of the big applications is motor position feedback. So for your process DC motors, it's really good for the commutation algorithm to understand where the rotor is in its rotation and its spin. For trapezoidal commutations, it's fine to have a HALL sensor there, which measures the magnetic -- measures the permanent magnets as they spin. However, if you get into more complicated and elaborate or more sophisticated algorithms for commutation like FOC or field oriented control which is more of a vector type approach more accurate -- a higher accurate position sensor or a coder might be needed and an inductive position sensor might be a fit there. Also the steer-by-wire or x-by-wire that I mentioned earlier with HELLA. The idea here is for the steering, instead of having that mechanical linkage from the steering wheel all the way to the axle, you do it by wire. And the same for pedal position and also power steering, and I'll talk a little bit more about that later. Some other applications of the powertrain turbocharge, camshaft position or vehicle suspension level sensing. Some of the sportier cars when they're going at high speed into tight churns, you want to make sure that, that vehicle is stable in terms of its coplanarity versus the road. And then also the throttle position sensor, radar sensors for marine applications and also strain sensors are pretty interesting application where let's say, for marine shipping, we have these large cranes lifting very, very heavy cargoes, you can distribute these sensors across the crane to make sure that the crane is not bowing or bending or -- in any kind of way. So lots of interesting applications there. That's for transportation. Now for industrial markets and applications. For encoders, you've got manufacturing industrial robots, warehousing autonomous mobile robots or AMRs, pick-and-place cobot. So cobots are an interesting or collaborative robots are an interesting product. In contrast to the manufacturing industrial robots over the top left, where half the time they've got these robots fenced in with hazard signs because you could definitely get injured if you walk by these robots and you're not aware of their perimeter of movement. Cobots or collaborative robots that's -- they're actually designed to work with humans. I can't count the number of times, like, for myself when I'm in my -- in the lab, trying to do something, and I wish I had a third hand let ideas where the collaborative robots come in, where they work with humans, in unison with humans to automate and expedite a lot of different tasks. They also learn very quickly and they're very easy to program. So a popular product, and you're finding coders in those as well. So moving right along, motor feedback with the rotor, I've already mentioned that, CNC and 3D printing, metal forming and fabrication, cut-to-length packaging, textiles, material handling, conveying in many different levels or different types of conveying just doesn't have to be limited to foods. And then also continuing into aerospace and defense, wind turbine pitch control, it's always -- there's a big advantage to have that turbine aligned properly to the incoming wind as the wind changes and then antenna positioning. And also, medical and health care, it's one of my favorite applications for encoders. For example, I've got over here, the surgical robotics, which is an example of [ Intuitive ] their da Vinci product where the surgeon sits meters away from the patient when the patient is receiving all these robotic tools that are entering the body to do all the surgical. So it's a very, very accurate surgical procedures with scalpels, tweezers and whatnot with -- along with this 3D imaging with the camera. So it's a really cool application. And I think you're going to see more and more of that. And then finally, the diagnostic equipment. Again, encoders help align automated practices. And so we'll start with the -- as Kyli mentioned, we're going to talk about 2 of our devices, the NCS32100, the industrial device, and then later, we'll move on to the NCV77320, which is automotive [indiscernible], both of which are inductive position sensors. So at a high level for encoders, you've got an MCU that's doing some sort of program control to some sort of actuator with some type of -- typically some kind of automation for many of these applications that I've described. And it's one thing for the MCU to say, I want something to move in the XYZ direction. It's another thing to say whether or not it actually got there. So that's where the encoder comes in. It provides this contactless sensing in terms of motion subdivided down into like position direction, speed and count, and that provides feedback to the encoder -- or I'm sorry, back to the MCU in the form of a digital or an analog signal. So that's its main role in the servo loop for actuator automation. Now for rotor encoders, there's many different types where the basics are, you have feedback, which is position, count, direction of speed. We mentioned that in the previous slide. And there's 2 major types, is the incremental and the absolute. So for position, the incremental basically informs the MCU that the position has changed. For absolute, it also informs that the position has changed, but it also informs the MCU of the exact position. So if you do a POR or power on reset of an absolute and an incremental encoder, that absolute encoder on the very first data reading, will tell you the exact position. For example, 36.128 degrees mechanical rotation. Whereas the incremental, it doesn't do that. It needs multiple readings to be able to tell you, okay, it's changed from the previous reading. And then both provide or can provide single and multiturn results. So rotary encoder -- there's many different technologies, but I'll summarize the top 3. The first of which is the optical encoder, it's probably the most popular. You've got a light source. And that light source shines through the slots of a disk, and that disk is attached to the rotating shaft, and then you've got a light detector. And that light detector, of course, as that shaft is spinning, will go high then low, then high, then low, then high then low. High every time the light goes through the slot, low every time the light is blocked. Now one downside to that is it really can't determine direction. So you can get a little bit fancy with this and do what's called a quadrature optical encoder where you have 2 light detectors really close to each other, such that when the light is going through the center of the slot, both light detectors, let's call them for now A and B, can both detect the light. So if you do that and you design your slots and the distance between the 2 lights detected properly, when that disk is spending clockwise, the -- for example, the A detector will see the light first. And when that disk runs counterclockwise, the B light detector will see the light first. And then an MCU reading that can determine the direction, whether it's clockwise or counterclockwise. Now you can get more fancy and have other types of slots that are longer or different shapes, moving closer towards -- in concentric circles moving closer towards the shaft, which can give you more information about position and also acceleration. So the optical encoder is a very popular device, simple to design, simple to use and can do a very, very high resolution. The next is a magnetic where you have a magnet attached to the shaft. And as the shaft spins, you have all elements or all effect products that basically capture the fact that the north pole has gone by them and then they can put out a digital one for that, and that tells an MCU, the basic spinning of the shaft. And then finally, there's the inductive encoder where you have 2 PCBs. You have a rotor PCB that's up above here that's attached to the shaft, and that freely spins with the shaft. And you also have a stator PCB that's stationary that mounts to the encoder housing. And that's where our encoder device resides. And you'll notice that there's printed patterns on both the stator PCB and the rotor PCB, and I'll go into great painful detail later in the next few slides how that works. But essentially, is these printed patterns are used to determine the rotor position. And this is our focus, the inductive encoder. Now some advantages of inductive encoding over optical and magnetic, number one, is the low sensitivity to vibration and contaminants. So when you have dust, water, oil, metal particles that can always cause an issue with the optical encoder common sense. You can see the -- you could block the slots, you could block the light source, you could block the light detector and so on. And then also metal particles could potentially degrade the performance of magnetic. Number two is you have a low component count, which I'll talk to in a slide further down and also no rare earth metals or no magnets. I'm a big fan of magnets. For those of you who've seen some of my other presentations, like, for example, permanent magnets and brushless DC motors. They're a wonderful thing, but they're not free and they are rare earth metals. Robustness, the inductive encoder is super repeatable, and also a very, very low tempco, or temperature coefficient. So it's very sensitive to changes in temperatures. And then finally, they're lightweight, right? You're just talking about the weight of 2 printed circuit boards. And by the way, the diameter of these -- the rotor and stator PCBs is on the order of 20 to 200 millimeters. We can work with any size. We typically, most of this presentation, will be referring to a 38-millimeter size or diameter just because that's kind of a sweet spot in the market for a lot of our applications. But we can -- our parts can be used from 20 millimeter out to 200 millimeter. And again, that's our focus, which is the inductive encoder. All right. So let's talk about how it works, which is my favorite part. And so I'm going to go back to the basics. So when you have -- so basically, what I'm just showing here on the bottom left is an inductor and we're putting an AC sign wave into that inductor. And so basically, an increase in inductor current causes that magnetic field to expand, okay? We typically store energy in that magnetic field. And then a decrease in inductive current here -- which I'm capturing here. That causes the magnetic field to collapse. And finally, the induced voltage or counter EMF polarity opposes that change in inductor current. So the main point I'm trying to make here is really probably better described by Faraday's law of mutual induction. So you've got this on the blue on the left, you're putting an AC sign wave into this blue wire, okay, and you're creating that electromagnetic field, right? And that electromagnetic field is emanating around that left wire, and it's collapsing and expanding, collapsing and expanding. But some of this field, if it intersects with a parallel wire, in this case, I've got this red wire here, if it intersects with that wire, we will induce energy or current into that [ red ] wire as depicted by the timing diagram over on the right. So the point of this is parallel wires couple, right? So if you have energy going through one -- and that's a function of the distance between the blue and the red wire, okay. If that distance is close enough such that the magnetic field intersects with the red wire, you're going to couple energy from the blue wire to the red wire. This is the principle -- this is the governing principle of our inductive position sensors. Okay. So we make the NCS32100, that's our industrial device. That's a dual inductive position sensor. I'll explain what dual means in a moment. So we have the rotor attached to the shaft, and that's in blue. And we have the stator attached -- that's not attached to the shaft, that's stationary, and that's in red. And we have this contactless air gap. That's on the order of 0.1 millimeters to 0.5 millimeters kind of range. Now the stator PCB, it's a circular PCB, and it's -- as I mentioned earlier, we typically, for our evaluations, have been working with 38-millimeter, but it can range from 20 to 200. It has 3 different coil patterns on there, and I'll go into more detail about that later. Meanwhile, the rotor, it has 2 different coil patterns, and I'll go into that in more detail later. And our device sits at the bottom side -- the NCS32100 sits on the bottom side of the stator board. And on the top right, you can see the internal block diagram. And I'll draw your attention up to the top left. We've got LC1 and LC2 oscillator. That's a transmitter that leaves -- that comes out of our device. And then we have 3 -- we have 8 input receivers from REC 0 to REC 7, that's input to our device. We do demodulate those inputs, send them through 2 ADCs and then into a DSP, or a digital signal processor, to process that information. And then we have an MCU that takes that information, puts it to the outside world on a 2.5 megahertz UR. Okay. So now the very first thing that happens is, is that coming out of the LC1, LC2 oscillator is we generate a 4 megahertz sine wave into that stator excitation coil. So I've highlighted it here in red. And it's a coil that's relatively narrow. That's -- it's a circular configuration depicted -- highlighted in red here. And this acts like an antenna, okay? So that's the first step. The second step is the rotor PCB has, as I mentioned earlier, 2 coil patterns. On the perimeter of the pattern, is what we call the fine coil. And inside and closer to the shaft, is what we call the coarse coil. And what happens is, is that the energy from the excitation coil is coupled to the rotor PCB fine and coarse coils. Now I'll just point out -- you can actually probably see that the coarse [ file ], that coarse pattern, if I count them, it's a repeating pattern, it's 1, 2, 3, 4, 5. So there's 5 repeating coils -- sorry, 5 repeating patterns in the coarse pattern. I'll go into more detail about that later what that really means. So we've essentially, through Faraday's law of mutual induction, we've coupled energy into the rotor PCB. Meanwhile, we've also got 2 more patterns on the stator. We've got the stator PCB with fine and coarse receive coil patterns, right, on the perimeter and also through the center. And those are physically mapped or connected to up to 8 of our receivers in our block diagram above, right? So we can sense -- we're sensing with those on the stator with up to 8 receiver inputs. And when the rotor spins, the coupled energy in the rotor coils disturb the stator coils. That's basically what's happening. That's how we determine where we are in the rotor rotation spin. And so the NCS32100 measures the rotor position by processing these disturbances inside of our DSP, which is kind of where our special sauce is. So that's it in a nutshell, I'm going to dive a little bit deeper into the theory of operation. On the left, you'll see basically the rotor is superimposed on to the stator. The rotor coils are in blue and the stator coils are in red. I'm only showing one set of coils on the stator in red, and that's that 4 megahertz sign wave. That's an excitation coil that we stimulate or excite with the 4 megahertz sign wave coming out of LC1 and LC2. The rotor fine coil in blue is on the outside parameter and has 16 periods. You can count them. The rotor coarse coil in blue inside parameter near the shaft has 5 periods. And in green, we're capturing where the wires in parallel will be inductively coupled, right? So we've got the stator excitation coil. Those lines are in parallel with the blue fine rotor coils and the blue coarse rotor coils. So that's the deal, right? That's how we're -- with the law -- with Faraday's law of mutual induction, we're energizing the rotor, and the perpendicular wires do not inductively cover. So that's step one to energize the rotor. Now step two, I've got that same image. However, I've kept the 2 rotor coils, fine and coarse, 16 and 5 periods. I'm removing the stator excitation coil, and now I'm adding the fine and coarse coils to the stator. And those are in red and green, and all that means is that we're using 2 layers on the stator coil, I forgot to mention. The receiver coil was just 1 layer of metal on the layer that's facing the stator. But getting back to these 2 stator received coils, green is top metal and red is bottom metal. So we're -- and you can see [indiscernible] we're stitching back and forth. That's part of our design for the stator. And so in this case, I have on the stator the fine receive coils, I have 16 periods here in red and green. And what I do is I map that to receive 0 on my device, the 32100 I'm inputting the fine coil, stator coil and receive 0. And I'm going to input the receive coarse coil which has 5 periods and receive 7. Now that's a bit arbitrary on my side -- on my end. It could be receive 1, 2, 3, 4. It doesn't have to be receive 0 and 7. I was just pointing out that the number of coils, which is relatively low, 16 for the fine and 5 for the coarse, I can map that to 2 inputs. I can map it to 3 inputs. I can map it to 4 inputs. It's up to the application and what we're trying to do here, trying to get the accuracy. So when that rotor spins, right, which are the blue coils -- wires in parallel will definitely couple and you can see that the coils on the fine and the coarse stator will be disturbed because those lines are in parallel. So again, we're leveraging Faraday's law mutual induction. Now what we do is we take this and we kind of -- we then start adding more. Well, so this image, I've removed the rotor coils. So it's all the stator coils, all 3 stator coils just because of graphic traffic. So there's the excitation coil. There's the fine coil, stator coils, and there's the coarse stator coils. And the point here is I've added more periods to both, the fine and coarse. And so as the rotor spins an increase in higher density in the stator coils periods, fine and coarse, coupling to all 8 receiver inputs, you noticed now I've highlighted all lead inputs. I'm maximizing our input. I'm literally getting more information about how -- what's happening with the -- capturing more information there. This augments the NCS32100's, and this is my term, disturbance granularity information. So this dramatically improves resolution and accuracy for calculating the rotor's position, velocity and acceleration. And so you can see we're just adding more information, and this is one of the value adds for our device. Some parasitic coupling between the excitation coil of the stator and the stator-received coils can occur, but we've got a few tricks up our sleeve where we can cancel that out with the DSP. So for accuracy, you can say we've got the dual boards, so the 2 PCBs, coarse and fine, that's why we're dual. The dual means coarse and fine. That's where we come up with that. We have plus or minus 50 arcsecs accuracy for a 38-millimeter sensor. And I do want to point out that, that works for 20 to 200 millimeter. We support a wide range, but we just chose that as a reference design. And 50 arcsecs equals just under 14 millidegrees in mechanical rotation, which is very accurate. But I'll show in the next few slides, 50 arcsecs is actually a very conservative spec. We actually can do better than that. We also -- so our absolute encoder -- we're an absolute encoder. So we also -- we guarantee that 50 arcsecs accuracy up to 6,000 RPMs. However, we can run up to 45,000 RPMs, but our accuracy does reduce above 6,000 arcsecs. So it is a function of the RPMs. We support 20-bit single turn resolution output and 24-bit multiturn resolution outputs, so plenty of resolution, plenty of churns that we can count. Our output is a 2.5 megahertz UR and it can be pumped to a -- it could be input into an RS45 driver, for example. We also support what's called SSI. It's a very -- SSI is synchronous serial interface. It's a digital protocol that's very common with encoders. And then we also support [ SPI ] and direct GPIO control. Other features include 2.75 to 5.5-volt power supply, QFN-40 5 x 5 millimeter package. And we also support what's called open coil detect, which is cool. And what that means is if one of those 8 receiver coils for some reason in the system -- in the application, one of those traces that are printed on the metal top and bottom, somehow are cut or there's some abrasion or compromise, we detect that open and announce that as a fault to companion MCU. So that's nice. Now anything that someone like us makes some kind of silicon, that has any kind of parameter that's a high-value parameter, like this accuracy or whatever it is, let's just say you're making a serializer or deserializer and you have a 40 picosecond [indiscernible] spec you need to measure that with some kind of real-time sampling scope or time equivalent sampling scope, and its noise floor needs to be well below what you're trying to measure, right? You can't. If the best thing a scope can measure is 40 picoseconds and that's your spec, it's not going to be a very good thing. You're -- hopefully, you've got a scope that can go down to 100 [ femtoseconds ] to measure 40 picoseconds of jitter. Well, it's similar for an encoder, right? So we went searching for an accurate encoder and we came up with this Gurley [ reference ] encoder that advertises an accuracy at 10 arcsecs. So what we did is we attached that -- we attached our 32100 to the shaft of this Gurley and we ran the data, and you can see the results over here, the accuracy across 360 degrees which is the x-axis and the arcsec's accuracy is in the y-axis. And we're closer to plus or minus 10 arcsecs, a typical accuracy. So we're in the [ mud ] of this Gurley reference. So one of the things that you'll talk about when you're doing measurements is a thing called measurement system analysis, and we're done on the mud. So it's one of those things. But we realize that a lot of applications, the skew between the rotor and the stator, which is essentially any kind of wobble associated with the shaft, isn't going to be perfect. Like this Gurley has an excellent control. We call it eccentricity where the movement in the X and Y direction of the rotor is really well maintained. And we know that many applications that's not the case. So what we did is we did the second set of experiments where we varied the x-axis eccentricity, which is captured here on the left, meaning that's the relationship between the stator and the rotor moving in the X direction and also moving in the Y direction over on the right. We said, okay, we're going to move that plus or minus 140 micron. And everything is captured in green is our measured accuracy. So you can see we hit 50 up here in the top. We hit 50 over here on the bottom left. And so it's sort of like an envelope where our worst case accuracy is right around 49, 48, 50. However, if you look in the middle here, you can see that we're right around 10, 15, 20 arcsecs, and that correlates to when we hooked up to the Gurley. So we're very -- so all I'm trying to say is the plus or minus 50 arcsecs we spec, that's actually quite conservative. Now the airgap between the rotor to stator is typically 1 millimeter to 5 millimeter. It's captured over here. And the x-axis of this plot shows the relatively flat accuracy from 100 to 500 micron. And then you can see, as you increase that gap, our error. We start to incur some more error. Our tempco is flat as a pancake. We can go minus 40 -- minus 40 degrees to plus 85 degrees, which is the operating temperature of this device. And the -- there's no compromise for accuracy versus temperature, and that's what you want. We're also very easy to calibrate. You can do a self-calibration. It's a single command from a master companion MCU. It takes about 2 seconds for run time. You do not need a reference encoder for the first sort of calibration. It can be run in any time out in the field by the end customer. So it's easy to use. You just need -- the rotor just needs to be moving anywhere from 100 to 1,000 RPMs. Now we compared our NCS32100 dual inductive sensor on the far right with an optical device from another company which is optical, just to show the side-by-side comparison for the total cost of ownership for a module. That's everything. We just make the sensor. And so we found a device that has an accuracy of plus or minus 50 arcsecs. We have that accuracy up to 6,000 RPM. We can go to 45,000 RPMs, but this particular company, their max RPM is 6,000. And we said, well, that's our max. That's what we guarantee the 50 arcsecs or so. This is an apples-to-apples comparison for accuracy. However, there is a caveat that the optical device will be more sensitive to contaminants and temperature. We already covered that earlier. The main takeaway from this slide is that the optical needs 3 PCBs with over 100 components. And we only need 2 PCBs, and the rotor PCB is just a single layer with no components. And we need about 12 components on the back side of the stator. And so the total cost for building a module with a dual inductive is a lot less than [indiscernible]. This is just a slide on how to order our valuation board. I have the link further below, where you can go and order it on our website or work with our channel partners directly with onsemi. And basically, the rotor -- the PCB rotor and stator fixture is right here on the top right, and it's connected to this knob. So you move the knob with your hand, and it shows up on the bottom left here in degrees -- the physical rotation in degrees on the graphic user interface or GUI. And then we also capturing this time line from 0 to 360 degrees. We also captured the number of turns. And then over here in the bottom, a time plot, we capture the velocity of the [ trim ]. And then on the far left, we capture a number of different alarms that will go back to an MCU. So a very comprehensive evaluation board with the graphic user interface and really fun to use. This is a 1 slider on our NCS32100 inductive position sensor. I would have covered just about everything on here, a little bit more packaging information at the bottom, you've got your block diagram view, which I've touched on earlier, the accuracy. We do support what's called a 3-phase Y configuration receive system. I'll touch on that further down, which is a nice thing. Our max junction is 105 C, so our operating temperature -- ambient operating temperature is minus 40 to 85 degrees C. We do have nonvolatile memory to store calibration coefficients. We have an internal temp sensor, which is nice, and we run off -- we have integrated Arm Cortex-M0+ MCU, along with that DSP. So everything is built in. You don't really have to do anything which makes it very easy to use. This slide is more or less another summary slide of the device with a little bit more information on the application, how to capture the accuracy, and that's there for sort of a teaser. Now one of the big applications for inductors are robotics. Onsemi is a huge leader for industrial robotics as well as smart robotics, collaborative robotics, mobile robots for warehousing, delivery and drones. We have intelligent power products, intelligent sensing. So for the intelligent power, we've got BLDC MCU-based controls, our EcoSpin, ECS640, for example. We have a ton of MOSFETs, gate drivers, IGBTs and SIC. We've got the integrated power modules. We have a lot of the shrubbery, DLDOs, DC to DC, the ISENSE, eFuse, the ESD protection, all of these things are important for a full solution. We also have, for the switch mode power supply, we have power factor correction, quasi-resonant flyback, active clamp flybacks, synchronous rectifier controllers, LLCs, gate drivers, auxiliary power that includes voltage and current supervision. Really cool thing is 10BASE-T1S. It's sort of replaces CAN for Daisy Chain connectivity -- communication, but it's all Ethernet, A to D converters. We have visible light communication and power for Ethernet. And then under intelligent sensing, we've got the image sensors, the LiDAR and also the inductive position sensing. And then we have our RSL15, which is a really cool BLE 5.2 device, super low power consumption and supports angle of arrival and angle of departure. So we're building -- we've built an AMR demo that we take to trade shows and show people. It's a really cool device. And we're going to be building a next-generation device. We're in the process of doing that now. And that will have all -- it uses all of our products, right? So it's sort of a showcase for all of our comprehensive robotic solutions, including LED drivers, motor drive, our DC to DC, the EcoSpin ECS640A, the 10-BASE T1s communication and then finally, the inductive position sensor, which is a critical piece of all of that. And we also use the NVIDIA Jetson Orin to process all of that. There's also image sensors and LiDARs involved, battery packs and battery charging anyway. So that's my little plug for putting on my marketing hat. That's my little plug for our autonomous mobile robot. Now -- I'm going to now pivot to our automotive inductive position sensor. This is our NCV77320. And this is a one slider on this. Our NCS32100 that I've talked about earlier, that's a rotary inductive position sensor. The NCV77320 can be both a rotary or a linear inductive position sensor. And over on the far right, I've got a photograph of our linear version. And basically, you've got a long green PCB with the coil patterns on it, and you've got the smaller red PCB that slides up and down the green PCB. And that red PCB also has coils printed on it. And when you slide that up and down, the green PCB, our device, which is up here on the top, it determines where that red board is relative to the green board. So it does linear position. And that's very typical for pedal position or there's plenty of applications where linear is preferred over vertical. This device does both. One of the differences, it has 3 inputs instead of 8 inputs. So it has less accuracy. It's on the order of 1,943 arcsecs, which is a little more than 0.5 degree mechanical rotation. Its max RPMs is 10,800. It does have what's called a cent output. That's an automotive protocol to our protocol, I believe. And that's with SPI or an analog communication. It's similar to the 32,100 and that's factored as Arm Cortex M0 plus MCU plus the DSP. It's got EEPROM inside. But one of the -- where it stands out is it's max junction temperature is 170 degrees C. So it can operate from minus 40 C to 150 C. So that's where it fits in for automotive. But there's plenty of applications in industrial where high temp is required. And so this finds a home in many industrial applications like motors, for example, where temperature is concerned. This is an example of some of our evaluation boards, and you can see rotary on the left, linear on the right. This is our main board here. And with this pivot from going from working just with a couple of customers to pivoting now to the mass market, we don't have this board available on our website yet. However, you can reach out to me if you're interested and I can help you get this, and we're working actively to get this on our website which I expect to see any week now. The NCS32100 industrial encoder, you can order that on our website. So this table captures compares and contrasts both inductive and automotive position sensor. The inductive is on the top. The automotive is on the bottom. They're both contactless, that air gap that I mentioned earlier. The industrial is rotary only, whereas the automotive is rotary and linear. The number of received coils is different. There's 3 received coils for the automotive. There's 8 received coils for the industrial. That's a big piece of why the industrial is a lot more accurate. But for some applications, 1,943, that's fine, and that equates to 0.539 degrees mechanical rotation. So not everybody needs [ 14 ] millidegrees of mechanical rotation. Here's the ambient operating range up to 150 C for the automotive. Here's the rotor speed max. And then finally, we're going left to right, at the far right is the difference in the package. So the automotive is TSSOP16. Okay. And then -- so one of the advantages for onsemi is to have -- to be able to do a 3-phase system. And so -- and we can do that because we've got 8 receiver inputs on our industrial device and 3 on our automotive device. And I've that drawn on the bottom left-hand corner. What this -- a 3-phase system is like a Wye ( Y) configuration. And so receiver 1 can come out and come -- it goes out through a coil and then it meets -- and it meets at a common point in the rotor. And then it comes back into receiver 3. And then we see receiver coil 2 goes out, and it also meets in that common point, and it comes in a receiver 2. That's an example of a Wye (Y) configuration. And so what you can do with that is when you're doing your measurements, you can do a differential measurement. So you've got those 3 coils, and you can subtract R1 from R3, you can subtract R1 from R2, and you can subtract R2 from R3, and that improves linearity. It reduces the harmonic distortion, and it gives you better signal noise ratio. In contrast, with a 2-phase system where you don't have that information, you can't do a differential measurement. Okay. So I'm returning to that original press release back in April 2023, where we announced we've shipped over 1 billion inductive sensors to HELLA, mostly for the automotive x-by-wire systems in the past 25 years. I mentioned before, it's kind of -- we know a thing or 2 about inductive position sensors and we're releasing that to the mass market. And so I just want to give you a few more examples of x-by-wire systems beyond the pedal position and the steer-by-wire. So it's sometimes called x-by-wire, and it's also sometimes called automotive wire. And so one example is the break pedal sensor, which stimulates the pedal field of the conventional brake system or accelerator pedal sensor, which determines the position of the accelerated pedal. It generates the corresponding pedal characteristics. There's a motor position sensor, and you can see these for both EV and hybrid powertrains. And it provides, as I mentioned before, motor precision with the rotor position for some applications that require that. And then there's also brake system sensors for accurate measurement of the brake pedal, vehicle-level sensor. And this can also extrapolate to the headlamp leveling or active suspension control for an accurate measurement of the car balancing load, which is important for your more sporty or high-performance vehicles. The transmission rain sensor, which measures the gear position of the transmission system. These are just some examples of where we fit into automotive. And it was surprising to me how many different instances there are. There's a throttle position sensor for the throttle body flap position. And this provides that information to the ECU, the electronic control unit. The exhaust gas recirculation sensor, or the EGR, used to measure the position and movement of exhaust gas valve. And you can use this for -- you can apply this to many industrial applications where you're mixing different gases. We're trying not to mix certain gases. And then you get the electric power sensing. And so at a high level, we're tapping into all these different megatrends in automotive. The idea is to move from mechanical to x-by-wire. These solutions include acceleration, steering, braking, suspension as well as regulation of actuators. And this is really an increased safety levels for enhanced autonomous driving. It's basically a part of electrification, so you can do e-motor for traction and auxiliary e-motors. And then you say, wait. So for BEVs or battery electric vehicles, one of the big hurdles for commercial adoption is the range anxiety. So if you can make your vehicle lighter for the same size battery, you can typically go further on charge size. That's the value add. So inherently, it's a higher reliability and lower cost of ownership. And so that's pretty much my presentation for today, my webinar. I hope all of you out there, if you're thinking encoders, think of onsemi and think of our inductive position encoders as a very high accurate and reliable source for determining your position in your applications.

All right. Thank you for the excellent presentation. We have received a number of questions, so we'll jump right in. [Operator Instructions] All right. So the first question that we have is, where can I get an AMR robot?

As far as like one of our AMR robots, that's something that we don't really -- we don't necessarily sell as a full product. But if you want to reach out to me by e-mail, we might be able to work out something where I can share some more details of our AMR robot and schematics as well and how we build it. We make a couple just for trade shows and demonstrations. I hope that answers that question.

That's a fine answer. What is the maximum airgap where the NCV77320 will work reliably?

That's a good question. I don't actually have the answer, but I can get back to that person who's asking that question. I don't know. I know the 32100 max air gap is [ 0.5 millimeter ]. I don't know off the top of my head. I suspect it's similar, but I really don't -- I don't know. So I can -- if that person wants to e-mail me, my e-mail is, I think, on the website. It's [email protected]. I'd be happy to answer that question directly.

All right. Another question. Could you give us a ballpark price for the NCV77320, 500,000 pieces per year and/or minimal system with [ sensor ] PCBs?

I don't have pricing off the top of my head. It's -- we do have budgetary pricing on our website, so you can go to -- you can google NCV77320, go to our website, and we have budgetary pricing for low volume. So for 500,000, it would be less than that, but I apologize. I try to stay away from pricing because I never -- I'm more of a technical guy, and I always get thrown. So -- but I would reach out to one of our channel partners or one of our salespeople or again, reach out to me and I can get that information for you.

All right. And then another question, why 4 megahertz? This was earlier on question.

Yes, that's a fair question. And it's not exactly 4 megahertz. It depends. Like I do believe it's anywhere from 3 to 4.2. But it's -- you just have to pick a frequency that's out of the audio range that, based on the distance between the rotor and the stator, gives you the best mutual inductance return for what you're trying to do. And so I just think that, that's a good frequency that is out of band, so to speak, for many of the other frequencies in probably automotive applications, more specifically and especially out of the audio band. Otherwise, I'm not exactly sure why 4 megahertz, but that seems to be what we're doing.

Okay. Another question, NCV77320 in linear mode, how fast can the rotor slide move in seconds or like inches per second or feet per second to provide maximum accuracy?

Well, when you're in a linear mode, you can only -- you can be in either linear mode or a rotor mode. You can have a rotary encoder or a linear encoder. So I'm a little bit thrown by the question because if you're in linear mode, it says how fast can the rotor slide. I guess, what you're saying is that red piece that's sliding up and down. How fast can it move? I don't know, but I do know Dale, and so I can find out and get back to Dale about that.

Perfect. Another question. What's the main purpose to use dual coil, fine and coarse, instead of 1 receiver coil to increase accuracy?

Yes. And that's my friend and colleague, Adam [ Tang ]. Yes, Adam, that's a good question, and this is kind of funny because I believe it has something to do with -- okay. So the real answer is I don't know, but I believe it has something to do with reducing the harmonics for the processing because you'll notice that the -- and I look at the patterns on the -- and I can go right to it, maybe that would help show you. Sorry, so if I look at these patterns between the coarse and the fine of the rotor, where the lines go vertical, they're not in sync. There's a little bit of an offset, right? They're not aligned where they go vertical, a little bit offset. And I believe that, that's done to help reduce and cancel harmonics which helps -- it's a signal integrity thing, I believe.

Okay. And this is just for...

There's some things -- sorry to interrupt you. There are some things that my internal business unit won't tell me because I'll blab it to everybody. So there's -- it may be part of our secret sauce.

This is just for clarification because there was a question about this, but is it course as an C-O-U-R-S-E or coarse as in C-O-A-R-S-E?

That's -- well, it's the coarse where you have fewer periods versus a higher number of periods. So it's -- gee I guess that's semantics. I believe it's C-O-R, C-O-A-R.

C-O-A-R?

I believe so. I'd have to look that up. You got me there. It's -- I think it's -- well, I've been given C-O-U-R-S-E by the business unit manufacturer that makes this. So that's what I've gone by.

Okay. No worries. All right. Next question, can the device cause EMI or EMC issues due to the excitation frequency of 4 megahertz into PCBs?

Well, that's something I've asked the business unit, and they've come back and said that's something that they work with, with different customers. I mean obviously, there's going to be some EMI, right? And what's nice is we've stayed out of the audio range of 20 hertz to 20 kilohertz. But that's kind of application specific. Some customers are fine with it. Other customers will build -- will shield it. And that's really dependent on the application, but it's a fair thing to talk about. I have asked my business unit that makes these devices to provide some EMI information, and they said they haven't done that yet, and that's something that's on their to-do list. But it's never been like a barrier -- it's never been a barrier. It's been something you work through. That's, I think, the best way to describe it.

All right. Next question. In terms of physics, what is the governing principle driving how NCS32100 and NCV77320 inductive position sensors works?

It's basically Faraday's laws of mutual induction. So if you've got 2 wires and you're injecting sine wave into one, it will couple to the other depending on how far away they are.

Perfect. Next, what is the accuracy of the NCS32100 inductive position sensor?

Yes. It's plus or minus 50 arcsecs or just under 14 millidegrees, but that assumes eccentricity or skew between the rotor and stator of plus or minus 140 microns. So if you tighten that up and you don't have any skew, it's closer to plus or minus 15 arcsecs.

Next question. Referring to the NCS32100, what does dual inductive position signify?

It signifies using the coarse and fine coil sets versus just using one set of coils on the rotor and stator.

All right. Then what are the advantages of using 3-phase sensor system over a 2-phase?

Well, the -- I guess I can go to that slide. It's really an active signal integrity. And so it helps reduce the linearity -- it improves the linearity, reduces harmonic distortion and gives you better signal noise ratio.

Okay. All right. And then I believe this is our final question. How do you see the trend of vehicle electrification affecting the x-by-wire system?

Well, the x-by-wire system will be -- I think will be growing in popularity as we have more BEVs, or battery electric vehicles, simply because you're replacing mechanical pieces by -- with wires. And they're always going to be lighter. And you can add redundancy, you can have 2 sets of wires. So our NCV77340 (sic) [ NCV77320 ] -- I'll just show you here. It actually can be -- it comes as ASIL B, but if you put 2 in parallel, you can achieve ASIL D safety because you can put 2 in parallel. So there's a lot of advantages. It's lighter, and you can augment the safety for a number of different applications, pedal and steering are 2 of them. And it's less susceptible to vibration, and also the whole thing about contaminants. These applications are factory floors in automotive or dirty environments. It's ugly, right? So anything you have and you're using, that's insensitive to those dirty environments is always going to be advantage.

All right. Well, thank you very much, Bob. These are all the questions for today. So on behalf of onsemi, I would like to thank everyone for attending, and I wish you a nice rest of your day.

Thank you.