 Hello everyone and welcome to We Meet at Digital Days. My name is Marie-Thérèse Kohl and I will moderate this presentation. We are very pleased that you took the time to participate in our virtual conference. The topic of this presentation is high-frequency three-phase PFC solutions for high-power charging stations from our partner ST Microelectronics. Our speaker is Francesco Genaro. He will hold the presentation and will answer your questions. But before we start, I would like to point out one thing. You will be muted during the presentation. This means that you cannot ask questions via microphone during the presentations. Nevertheless, you have the opportunity to ask questions during the presentation at any time via the chat function. You will find the chat function in the control panel. This presentation will be about 30 minutes long. The chat questions will be answered in a Q&A session following the webinar. There are 5 to 10 minutes in addition scheduled for this. If we are unable to answer all your questions within this time, we will answer them via email afterwards. If you still have any other questions left, just mail us at exhibition at we-online.com. We will try to answer all questions promptly. At the end of the webinar, you will be asked to participate in a feedback survey. You would be pleased if you take the time to fill out a survey and help us to improve our event. You will also receive the link to the presentation in the next few days. The recording will be available at our website shortly. But now I will hand over to our speaker Francesco, and I wish you an exciting presentation. Good afternoon. My name is Francesco Genaro, I am an application manager for Power Condition Development and a working system research and application group in SD Micropronics. Today we will discuss about high-frequency 3-phase BFC solution for high-power charge stations. With a very short agenda, we will give a look at the charge station as a general power factor correction, a short view of semiconductor technologies which are interested in this application, and a focus on a specific topology, which is the bi-directional TFC, named three-level T-type conductor. Then we will go to conclusions. Charge stations are the very companion of characterization, and they are really impacting on the grid infrastructure and management. In this presentation, I will go to a building block of a charge station, in particular a DC charge station, looking at the active front end, which is the interface of the charge station with the grid itself. In this case, due to the interface functionality, power quality performance are very important as well as efficiency and size of the charge station itself. General power converted topology for high power are available and will be considered in terms of efficiency and performance. With a specific look at the introduction of the new technologies and conductive technology which is based on silicon carbide, charge stations are basically energy-driven applications and we can see which is the trend towards 2030, and we are going up to 180 billion kilowatt-hours with a number of chargers which is increasing everywhere in the world, from US, Europe, and China as well. And the investment plan, as we can see from the source of this TASI, which is McKinsey, up to $50 billion by 2030. The charging station as a DC fast charger has basically two blocks. The first one is interface with the grid, which is the PFC or the power factor corrector circuit, and then there is a second stage, which is the isolated DC-DC converter, which is providing the regulated current to charge the battery of the car, for example. Giving a look at the specific or the three-phase PFC, of course, we have availability of power technologies for the switches, and we need also some signal conditioning and the control section, and basically for this application, just microcontrollers, standard microcontrollers are available and are able to drive the application. Shortly on power factor correction, as I said, the front-end converter is the power electronic stage which is in charge of controlling the input current, making the power converter behaving as a resistor, so basically having the input current in phase with the voltage and drawing a very sinusoidal current. This is important because this is impacting the quality of the grid and also the losses that we can feature on the grid. So two main factors, two main figures or merits, are important in terms of power factor correction and are the power factor itself, which is a number, but also the total harmonic distortion, which is a measure of the number of harmonics that are inside the current that we are drawing from the grid. In Europe, for example, there is a specific standard which is driving and regulating this harmonic distortion and all the behavior of the power factor correction. Giving a look at the history, this is the very first three-phase PFC, very simple, which is consist of input inductors plus a three-phase diode bridge and only one active switch, which is regulating the input current. Of course, it was not performing the best for this application, but it was just the first starting point for the new application. Total harmonic distortion was not the best or still quite high, but this was the thing, as I said, this was just the beginning of the history. And this was a second step still for the three-phase, where after several studies, we came up with a very simple topology, as we can see here, and it's called as a Vienna rectifier, where basically the three-phase converter has been optimized with bi-directional switches, and in addition, with a three-level modulation, which is able to stress the device with half of the output voltage. And this is more and more important as the output voltage is increasing. So in particular, for charging stations where a typical DC bus voltage is in the order of 800 volts, more or less, it is important to have reduced voltage across the switches, because this will save a lot of switching losses. There is also another approach to the three-phase solution, which is a modular approach starting from single phase. This solution is allowed to have three modules, three single phase modules, with a single output, which could be operated even in parallel, in order to improve the enlarge the available power. These topologies and these architecture are really suitable, where there is a possibility to supply the chargers station, even with a single phase or with a three-phase, directly three-phase voltage. So I would say that the three-phase module approach with a single-phase module is more suitable for onboard charging. But it's going also to introduce in order to reduce the industrial cost, also for the three-phase system for chargers station itself. Semiconductor technologies, which are very important in this field, because they are enabling the development of new structures increasing also the performance of the solution. So just to give a look at what is available in the market from semiconductor supplier and dividing the overall scenario in terms of frequency and output power, we can see that chargers stations are more or less positioned in a power range above 10 kW. So let's say starting from 20 to 22 kW up to some hundreds of kW. Of course, the 100 kW application is more and more modular, with some basic building blocks in the order of 30 or 50 kW. In the same area, in terms of power semiconductor, we can also position the silicon carbide, both for diodes and MOSFETs, since it's in these technologies providing a very high frequency and high power superior performance compared to other semiconductor technologies like silicon, in this case. There is also a smaller part for GAN, but of course this is limited with the lower power application and higher-reaching application itself. So for the chargers station, I would say that semiconductor of choice is the silicon carbide, and this is a short story of silicon carbide from 20 years ago, till now with the continuous shrinking of the size of the die and the superior performance, which are more and more improving in the last year. So also in terms of adoption in the market, there was in the last three or four years an increase of more or less 30 times from 2017 to today. The selection of this semiconductor technology is not only for a lower voltage class, which means 660 volts, but it's also available for 1200 volts devices, and this is where silicon carbide is providing the best performance and the best benefits for the overall application. Also, because silicon iGBTs is the counterpart in terms of the semiconductor technology, in terms of silicon, and of course there is no real substitution of silicon carbide of the silicon iGBTs, but it depends on what is the final aim of the application. It would like to have high power and high efficiency with no cost relationship, so then silicon carbide most that can be considered for sure. On the other side, if we need a balance between cost and performance, silicon iGBTs are the device of choice. I will show later which could be the benefit of using one or the other technology. Six benefits and advantages which are really summarizing in these slides. So higher performance and voltage operation, so extremely low power losses, high efficiency at low current, and there is also the intrinsic sick body diode, which is able to provide the forefront operation. We can increase operating frequency with the lowest reach losses and reaching performance of the diode as well. And there is one additional factor which is really important, which is the operating temperature that we can extend due to the wide-band gap material characteristics up to 200 degrees Celsius in terms of junction temperature. Of course, also the package has to sustain this high temperature, so there is not only semiconductor, but there is also package. So there is an overall efficiency regain in a typical application with higher, with lower, reaching losses, more or less one tenth, with one fifth of chip size, so also the size can be reduced and total losses reduced by 50% after switching frequency, which can be 10 times higher. So of course, there is no single answer to the question of the selection, but of course, as usual, is a trade-off between performance, cost, size, and overall design. As I said, there is a very strong link between technology of the semiconductor and the package. So in order to get the best benefit from this superior technology, we need also to have a suitable package in order to manage, basically, thermally, the final device. And also for silicon carbide, has been developed some specific package, which is a single dual-dye packages, like the HU3PAC and the STPAC, which are for very high current application, thanks to the superior thermal performance. When the discrete approach is not enough in terms of power, then the module, the low-cost and very simple module that we call ASPEC, or ASPEC one or two, or ASPEC three, can be used in order to have a multi-dye power module. Just to provide you a very, very simple comparison between silicon carbide and silicon. So basically, silicon carbide MOSFET and silicon GBT is inside the module. We can see that from a simple exercise, there will be a reduction of 10 times compared to silicon more, silicon IGBTs, thanks to the reduced size of the dial. And in terms of performance, if we give a look at the efficiency graph here for silicon carbide MOSFET, we can achieve more than 3% of efficiency gain compared to silicon IGBTs. It is important also to have a very good selection in terms of topologies in order to reduce the size of the overall converter. So the typical building block for the converter, like the alpha bridge, the full bridge, or the six-pack full bridge are available in the market, as well as three-level T-type leg, which is very common in this kind of application. Which is the real importance of the package technology. It's important basically for three steps. The first one is from an electrical point of view, because we can reduce the internal connection layout. So we can reduce the parasitics inductance of the connection. We can improve the overall layout of the board. So this will lead for sure in the reduction of the noise around the circuit and in improving of the switching losses, so the efficiency of the converter. There is also a thermal aspect. Since the module is much easier to dissipate the heat from the circuit and also to manage the overall thermal aspects of the application. It's quite easier since it's concentrated is one module. And going back to electrical aspects, the noise and the disturbance around the circuit, which are included inside the optimized layout of the module, will be reduced. So the overall EMI of the circuit will be improved using this package. Let's go to simple PFC topologies. So the three-level topologies are being summarized in two. So in the generative fire type one and type two, type two has a basic comparison between the two topologies. The main difference is that in type one, we can use only 650 volts devices. So we do not need any very high voltage device, even if the output voltage is up to 800 volts or 9900 volts. Since it is a three-level topology and the voltage across the device will not see higher than half of the output DC voltage. But there is a disadvantage which is related with the current path that we'll see always two switches in series, one diode and one arctic switch, with a lower, slightly lower efficiency, which is not the case of the type two generative fire, but there is the necessity to use 1200 volts devices. Or in this case, for example, D2 plus and minus 12 volts, and the volt diodes. Here we have a comparison between the performance of type one and type two converter and the spanning the frequency between 20 kilohertz and 80 kilohertz. We can see which is the difference between the three main semiconductor technologies. The first one is the silicon AGPTs in red, in green, the superjunction monster. So still based on silicon technology. And the blue one is based on silicon carbide MOSFETs. For the two topology, we don't see a very big difference when the switching frequency is quite low for a given device. But as soon as we increase the switching frequency, we approach 100 kilohertz, which can be something feasible for this kind of application. We can see a small difference in terms of efficiency in the order of 0.2, 0.3%, which is nothing for one kilowatt application. But if you consider 50 kilowatt application, this means a lot of energy that can be dissipated. For sick and able technology, what does it mean? It means that thanks to the introduction of silicon carbide technology, some topologies that are not able to be used with the superjunction MOSFETs or with general silicon-based power transistor are, again, valid also for very high power and high switching frequency. This is the case of the generative fire type II with input diodes using silicon carbide diodes and also for the bi-direction of the generative fire, which is an evolution of the type II generative fire, just replacing the input diode bridge with active devices. In this case, the switches from T1 to T6 has to be rated at 1200 volts. Why we are talking about bi-directional converters? This is very important. This is a new trend since there is a smart grid and there is a vehicle-to-grid capability of the converters more and more requested. So we need to reverse the energy from the battery to the grid in addition to charge the battery from the grid. So the bi-directional function is more and more requested and I would say is the selection choice for any new development for the range between 20 kilowatts and up to 350 kilowatts, for example. There is also the enabling of the standard fridge-boost converter, which is based on V6 technology. Thanks to silicon carbide MOSFETs, reduce switching losses and reduce size. So this is what's new. Thanks to the silicon carbide technology, the fridge-boost converter. So let's give a look at an example, an exercise that we did with the demo solution, a demo board. So we use a specification on the table, so for under volts VSE, in the input 800 volts DC on the output, 13 kilowatts up to RMS, up to 15 kilowatts peak power, the switching peak pressure of 70 kilowatts, and for the PFC design, we use a ripple current of 2.5 amps on the input with 10 volts peak to peak on the output as a ripple voltage. The question for the design are shown there and of course we can see that in order to select the design, the input inductor, we have to know the switching frequency as well as the ripple current that we can accept and the modulation index of the converter, which is defined as the ratio between the input voltage and half of the DC output voltage. The RMS current is defined as the DC power divided by three times the input voltage RMS, and this would be the RMS current of the select inductor with the peak value of 25 amps, more or less. This bidirectional structure demo board is able also to be operated also as a bridges three-phase or single boost three-phase solution when we disconnect the three-level T-type cell. So we can evaluate the board and the system not only as a three-level solution but also as a two-level. A very short summary of performance compared to the two-level and three-level converter in terms of semiconductor technology is shown here where silicon carbide is always performing better than IGBTs except a very, very small frequency span at a very low frequency, but once we increase the switching frequency above 50 kilohertz the difference between the two converter in terms of efficiency can reach up to 3-4%. So why selecting one solution against the other? For which are the pros and cons of the three-level versus the two-level? For the three-level we have losses distributed over more components so we can spread better the heat generated by the switches. We can have higher efficiency at higher switching frequency thanks to the three-level approach and also thanks to the reduced voltage we can reduce the volume of the input and output. Which are the cons of this solution? We need more components. We have higher complexity also in terms of control. We could need an output voltage balancing for the split capacitor output and also an expected higher cost because of more components. So in the table you can see how many PWM complementary we need for the two-level and two-level converter. It's more complex. We need more gate driver. We also more isolated the CDC converter to supply the gate driver. This is the basic schematic of the solution. It's based on a standard microcontroller and thus a microcontroller STM32G4 and is implemented in the bi-direction functionality thanks to the grid relay which are able to connect and disconnect from the grid. The system is based on a very simple control both for the AC, DC and DC to AC conversion based on DQ axis transformation. So we can decouple exactly the active and reactive power and then we can have a very accurate regulation of active and reactive power as well as the power factor of the system. So this is a summary of the specs of the board and the goal is to have a power factor higher than 0.90% starting from 20% of the load and the total harmonic distortion below the 5% which is basically according to the standard and an efficiency higher than 97% starting from 20% of the output or the output load. Key products are starting from the microcontroller, serial and carbide MOSFETs and of course there is one key part, key component which is not a semiconductor but is a magnetic so the input inductor which is not trivial to manage is not trivial to design and is not trivial to manufacture with the reasonable cost. To give you an idea of the complexity of the system the MCU is using high resolution timer to manage the high frequency operation and is also equipped with the math accelerator like the codec or the trigonometric function which is requested by the DQ axis transformation. Here is a picture of the system. We are the three input boost inductor given by VOOT. This is a system that has been developed in partnership with VOOT Electronics and we have developed and optimized this kind of components. Thanks to this we were able to reach very high efficiency so as I said above 97% after the 2kW and then up to 98% and in that range for the full range. In terms of power factor we are always in the range of 99%. So going back to the PFC inductor which are the main requirements of this part which is not semiconductor but is impacting the performance of the system not only in terms of efficiency but also in terms of total harmonic distortion and power factor controllability. We need in addition to have low thermal resistance because of thermal management we need low parasitic capacitance in order to have reduced losses on the components. It is able to work in offline application with very high voltage and has to be low leakage, low radiated magnetic field with a compact design and easy to be mounted on the board. These are the main specs of the inductor that we use and has been optimized and this is the final stage of a study that we did starting from a standard right core so the E65 and the EE80 which are able to manage the 27, 28 pic current grid and we finalize with the toroidal one and in order to give an explanation to this selection we of course gave a look also at the cost of these final components and if we took one the price and the cost of the E65 core but that was importantly able to handle only 17 amps we finished with 0.72% of the cost with the toroidal one against a 3.85 with the EE80 so we can understand which is the component of selection for low cost industrial solution. So let's go to the conclusion Chargers Station are an emerging application and DC Chargers are gaining a market up to 50 kW in terms of application. Of course this is a basic building block of the final application and we can stack several modules and interact them to have a very high power Chargers Station. The bidirectional functionality is more and more requested in order to allow this converter in general any battery charger with a smart grid equipped with the V2G functionality. The front-end converter represents the connection to the grid so it's a very very important part of the system so even if it's simple it has to manage the most difficult part of the game. Silicone carbide MOSFETs is as introduction is a very important factor in this application allowing a simple solution simple architecture to be used with a very high frequency solution and high performance in terms of efficiency and the switching frequency which is very very important in order to have a superior performance also in terms of power quality. Thank you for your time and I'm ready for your question. Thanks a lot Francesco for your interesting presentation. Now we would like to turn our attention to your questions and we will wait a second until some questions will come in and I already see one first question Francesco can the Vienna converters be realized for low voltage applications or they are confined to only high voltage applications? In general Vienna ratifier can be used even for low voltage applications it depends on the input voltage as well as the output voltage. The main benefit is when you have the high voltage on the output so you can split the output voltage you can divide it at least by two the output voltage so this will allow to have reduced switching losses. There is also let's say a decision to take consideration of the power that we have to manage so basically the specs has to be considered not only the voltage on the output but also the input voltage and the output power itself. Thanks maybe one next question is the board suitable only for silicon carbide MOSFETs? As I've shown there is also a possibility to use other semiconductor technology so super jump with MOSFET as well as IGBTs and for these different technologies some fine tuning for the switching frequency as being done for example for IGBTs in order to have the best performance the switching frequency should be reduced a little bit maybe around 30 kilohertz. Perfect thanks and maybe one last question is the board automotive grade qualified? The system is just a demo board and has been built using industrial grade components but in terms of concept we have all the components available to have the same solution based on automotive grade components. Thanks a lot. Thanks a lot for answering the questions I would say if there are any questions left we will answer them via email afterwards so thank you very much for your attention and I hope you enjoyed our presentation the next and last presentation for you today is case study how ESD could generate electromechanic issues and it will start at 5 p.m. See you there and enjoy our digital days. Goodbye. Thank you to all. Have a nice afternoon. Bye.