Trends in switch-mode power supplies – how far we've come

Speaker 1 (00: 01):

Speaker 1 (00: Sound on. Power on. Your power electronics podcast, powered by PCIM Europe.

Prof. Marco Jung (00: 09):

Prof. Marco Jung (00: Hello everybody, and welcome to the fourth episode of Sound on. Power on. Your power electronics podcast, powered by PCIM Europe. Today from the exhibition and conference in Nuremberg. My name is Marco Jung. I'm a professor for E-mobility and Electrical Infrastructure at the Bonn-Rhein-Sieg University of Applied Science at Sankt Augustin, as well as Head of Department, Converter and Electrical Drives, at the Fraunhofer Institute for Energy Economics and Energy System Technology at Kassel. Both are located in Germany. Additionally, I am the chairman of the IEEE Joint IES/IAS/PELS German Chapter. And today, our technical theme is 40 years power electronics, a journey through time with a view into the future of switch-mode power supplies and this, I will discuss with the Senior Director, Research and Development, Dr. Peter Wallmeier from Delta Energy Systems. Hi, Peter, how are you?

Dr. Peter Wallmeier (01: 09):

Dr. Peter Wallmeier (01: I'm fine. Thank you.

Prof. Marco Jung (01: 10):

Prof. Marco Jung (01: So welcome to our podcast here, and we are now sitting at the exhibition and conference. And so, Peter for me and for our listeners, it is interesting to know a little bit more about you. And I saw in your CV that your first contact with power electronics was during your study at the University of Paderborn. What was the state of the art of power electronics, and what were presented at the lectures?

Dr. Peter Wallmeier (01: 39):

Dr. Peter Wallmeier (01: It was the beginning of power electronics at that time. We had a history of power electronics starting at that Institute since 1982, and obviously switched-mode power conversion was already in place, and the first dissertations have been written in that topic. And that was the time of Mr. Clemente, International Rectifier, EFP-450, large RDS on, large QRRs, lousy body diode. So yeah, and then we basically spent our time and our research on circumventing this disadvantages of this devices and developed hard-switched and soft-switched circuits, but that was basically my entrance point.

Prof. Marco Jung (02: 23):

Prof. Marco Jung (02: Mm-hmm. And this was during your PhD time?

Dr. Peter Wallmeier (02: 26):

No, no, no. It was kind of an accident, I would say. Surprisingly accident in 1991 friendly speaking. There was a lot of activities around our sponsor ABB Power Supply Systems. At that time I think the name was ABB CEAG in Soest. Bigger company developing custom design switch-mode power supplies for data center infrastructure. And at that time, a lot of research was done on super computers, so five volt, 1000 amps devices. And in that framework, the company spent quite a bit on research money and financed PhD students at the University of Paderborn. And I stepped into the office of such kind of a PhD student as a student. And he was asking me: "Hey, do you need a semester job?", and I said: "Yes, I need a semester job." And then I was catched by him and did finite-element simulations on resonant chokes. Yes.

Prof. Marco Jung (03: 22):

Prof. Marco Jung (03: Ah, great. So that was your task later in the PhD?

Dr. Peter Wallmeier (03: 27):

Dr. Peter Wallmeier (03: No, no, no. So yeah, it was starting point and the research topic was to circumvent switching losses of the basically lousy semiconductors at that time, and then develop resonant converters. And in these resonant converters, you have soft switching of the power devices, but you have AC currents in the windings of the magnetics. I mean, still the same problem as today.

Prof. Marco Jung (03: 48):

Prof. Marco Jung (03: Yeah.

Dr. Peter Wallmeier (03: 49):

Dr. Peter Wallmeier (03: And to understand skin and proximity effects in the winding, and took all the paralytic effects of hysteresis losses and eddy-current losses. That was basically the task at that time. And I started as magnetics design and analyst, I would say, in that institute and then devoted my student thesis, at that time I did a normal diploma, at that university in Paderborn for high frequency inductors, and then later on for high frequency transformers. And also, my PhD was along the automated optimization of these magnetics. So that was basically, as freindly speaking, everything was very related to magnetics, but then also extended them to circuit technology.

Prof. Marco Jung (04: 38):

Prof. Marco Jung (04: And this you have later done in different companies or what?

Dr. Peter Wallmeier (04: 42):

Dr. Peter Wallmeier (04: Yeah, yeah. That's correct. So when I left the university, I spent one year... It was very hard to get a job at that time, 1994, it was very difficult. I worked at a startup company doing finite-element and find a difference computing or field computing on parallel computers. And after one year... So it was the time also when the EMC Directive was set up in 1995, when the power factor correction was requested and you had the CE compliant of the information technology and in that framework, anechoic chambers have been used. And for that, you needed find a difference algorithms in order to calculate the damping of these absorbers. And that was basically my task. And I returned after one year to the university because I decided to do my PhD and continued the high frequency magnetics optimization, and also circuit optimization of switch mode power supplies, always in conjunction with ABB CEAG.

Prof. Marco Jung (05: 49):

Prof. Marco Jung (05: Great. So maybe for all our listeners, if we discuss about 1991. I was there in the elementary school. So it's a really different science.

Dr. Peter Wallmeier (06: 01):

Dr. Peter Wallmeier (06: No, I get depressions here.

Prof. Marco Jung (06: 01):

Prof. Marco Jung (06: Really funny.

Dr. Peter Wallmeier (06: 05):

Dr. Peter Wallmeier (06: I get depressive. How time flies, yes.

Prof. Marco Jung (06: 06):

Prof. Marco Jung (06: Yeah. It's true, it's true. Okay. But I think let's come to our theme for today and go a little bit more deeper in discussion and let us have a focus back to the other Sound on. Power on. podcast. And there we discussed future trends in power electronics, new semiconductor devices, and so on, in the context of renewable energies. Now we want to have a look into the direction of loads, especially power supplies. Can you show or describe us where we come from?

Dr. Peter Wallmeier (06: 38):

Dr. Peter Wallmeier (06: Yes, for sure. So the origin of switch mode power conversion is mostly in the information technology. So that means computers, in the beginning TV sets. So very simple stuff. And we started in information technology for computers for Siemens Nixdorf, if you maybe remember the name.

Prof. Marco Jung (07: 01):

Prof. Marco Jung (07: Yes, I know it.

Dr. Peter Wallmeier (07: 01):

Dr. Peter Wallmeier (07: So Nixdorf computers in the late '80s, beginning of the '90s, and then there was Amdahl, Cray, and, so basically the super computers requesting TTL technology. So five volts and then thousands of amps in a water cooled system, and that was basically our starting point. And you needed in this information technology applications, always high, efficient, high dense switch mode, power supplies, and conversions. Basically converting from AC to 12 volt, 5 volt, 3.3, to the processor voltages. This application was the technology driver for basically my career. So still today. We started with below 90, 85, in the 80s efficiency, and now we're at 98% efficiency of power supply. So that was the cause of development and obviously semiconductor technology, magnetics packaging, everything contributed to that development. Yes.

Prof. Marco Jung (08: 01):

Prof. Marco Jung (08: Mm-hmm. As I said, efficiency, I think efficiency is one of the key drivers for R and D in the present. What is the state of the art?

Dr. Peter Wallmeier (08: 10):

Like I said, the efficiency. So we have now the word ultra efficient it's high efficient. It is basically ending at 98% and they're ultra efficient converters. So the publications from the ETHZ from Professor Kolar and his team, they have 99% efficient. And now we are approaching the 99 and I said always: "The limit is a 100." Okay. So there's nothing beyond 100, but we are right now at 98% for our server power supplies. So the three kilowatt and five kilowatt units converting from single face to 48 volt or 12 volt are in that range. And, since it's always good, right? Because the losses is waste and waste means there's a chain of waste, right? Until the generators, until the coal fired power plant or the nuclear power plant it's an addition of waste and we should avoid this waste in order to save our climate.

Prof. Marco Jung (09: 04):

Prof. Marco Jung (09: True. And I think so for the people, the cost in this case.

Dr. Peter Wallmeier (09: 08):

Dr. Peter Wallmeier (09: Case, yeah. Clients, it's always adding infrastructure costs.

Prof. Marco Jung (09: 10):

Prof. Marco Jung (09: And if we have the discussion now about this, but what are the main applications and what are the power ranges for them?

Dr. Peter Wallmeier (09: 17):

Dr. Peter Wallmeier (09: Yeah, that is also interesting. I mean, when I refer a little bit also to my keynote is we started with IT equipment, right? This is still today key application, but the switch mode conversion has diverted to many, many fields, right? So the application field has, has tremendously increased and we have today medical applications. So medical power supplies are all switched modes. We have wireless charging, we have mobility. So onboard chargers on electrical vehicles, they are all switched mode and solar is switched mode. So it's everywhere. Switch mode, power conversion. I mean, your mobile telephone has an intelligent power management chip insight, which does all the chargers you use for your mobile equipment is driven by switch mode, conversion technology.

Prof. Marco Jung (10: 01):

Prof. Marco Jung (10: I think so. And I think for me and our listeners, it is not interesting only the state of the art, I think the future and the next steps are really interesting. In your presentation at the PCIM, you mentioned that the magnetics as a bottleneck for the next innovation steps. Please share us your thoughts about this.

Dr. Peter Wallmeier (10: 22):

Dr. Peter Wallmeier (10: Yeah. So, actually, as a magnetics engineer, it's very hard to, because I'm a magnetics engineer who would divert it in the switch mode technology. It's very hard to come to this conclusion, but when you look to this, to the development, which has been done over the last 40 years on power semiconductors, and when you go down here into the exhibition, you see ultra high efficient semiconductors. Very low RDS on almost zero charge with the GAN devices. But when you stop then at the magnetics manufacturer booth, you see that they're still fighting with the limits of materials. I mean, it's very hard to change copper and ferrite.

Prof. Marco Jung (10: 55):

Prof. Marco Jung (10: Yeah, I think so. That would be the next steps. And I think it's really necessary for next innovations.

Dr. Peter Wallmeier (11: 02):

Dr. Peter Wallmeier (11: Yeah. I think when you look back now, and so the last 10 years, so now we moved from silicon to silicon carbide to gallium nitride, it's all new material. But when you look to magnetics, the materials haven't changed. You still talk about ferrite, you don't talk about a new name, right? So I would expect a new name for a core material. We have amorphous and nanocrystalline, but you see they are still limited in increasing the switching frequency at reduced core losses. And for me it's... When I wrote this and when you reflect on yourself, it gets more and more evident that the investments have been done in billion-dollar factories. This Wolfspeed silicon carbide factory, close to New York. Everybody's investing in semiconductors. But what about the magnetics? I don't see that, friendly speaking.

Prof. Marco Jung (11: 50):

Prof. Marco Jung (11: I think so. Nobody wants to spend a lot of money in this area, but I think they must do it in the future and I think one main reason, and the problem is now they must speak two different cultures. Yeah. Chemistry and power electronic engineers. And maybe they have other thinking in this area.

Dr. Peter Wallmeier (12: 10):

Yeah, as we talk the power semiconductors are made by physicists. And when I worked for Infineon and when I talked to a diode developer or an IGBT developer, it has been different words because I have to explain to him: "Hey, I need less QRR." And he was telling: "I had to change this and this doping and I have to change the implantate and the structure of the diode." But in semiconductor business material science, physicists have spent really a lot of effort in order to make new materials and with much better capabilities and properties. And in magnetics, I don't see that, friendly speaking, I think that's an observation I make now more clearer than ever.

Prof. Marco Jung (12: 51):

Prof. Marco Jung (12: They must speak together.

Dr. Peter Wallmeier (12: 52):

Dr. Peter Wallmeier (12: Yeah. And why there is not so much investment into magnetics is a little bit a question to me because in the end for the semiconductor manufacturers, right now, also this becomes a blocking point. For the UPSs, definitely there is this classical rectifier and inverter conversion, which you have in the classical UPSs. They have now this eco mode. They basically bypass this two stage power conversion and have some clever ways to manage on the one hand, the power factor correction and the power security on the protected line. And that is a kind of a trick, yeah. To boost the efficiency in normal grid operation by around about 3%. But when it comes to online operation, then these converters are active. And then obviously the classical technology advantage advances like silicon carbite, gallium nitride devices, multilevel technologies, soft switching.

Dr. Peter Wallmeier (13: 51):

Dr. Peter Wallmeier (13: You also find this in the UPSs, right. And especially three level power conversion, Vienna rectifier is very famous, they are. So, and together with new components, they increase the efficiency up to 98% per stage. And then a normal UPS has 96, 97% online efficiency. When you bypass it, obviously it's 99 or plus, this eco mode which is promoted. But then you are directly connected to the grid and then your AC to DC power supplies, which we produce, are also connected to the grids, so directly. So the requirement for search and burst and abnormalities on the voltage, on the supply voltage gets higher for us due to this eco mode.

Prof. Marco Jung (14: 38):

Prof. Marco Jung (14: Okay. And what do you think typically UPS, they have the storage system is a battery, but maybe in the future, yeah, we have a look. I discussed in the past too, using hydrogen, or maybe using local integrated renewable energy. What do you think about, is it a possibility for the future to integrate it in one system, in a UPS system?

Dr. Peter Wallmeier (15: 00):

Dr. Peter Wallmeier (15: I think so. There's companies doing that already. So as demonstrators, there's one company Proton Motors doing this and they have fuel cells and integrate this into UPSs as an example. And we are also working on that new data center structures involving fuel cells and battery storage, for sure. So larger battery storage, replacing then the diesel gensets.

Prof. Marco Jung (15: 28):

Prof. Marco Jung (15: Yeah. So for fuel saving and so on.

Dr. Peter Wallmeier (15: 31):

Dr. Peter Wallmeier (15: Yeah and for long term storage, because normally a battery in a UPS just bridges the time when the diesel gensets turn on, right? And a few minutes, and this is now extended. So the trend is to leave out the genset and extend the size of the battery by using lithium ions.

Prof. Marco Jung (15: 49):

Prof. Marco Jung (15: And due to an expected future penetration of renewable energies, I think in the future power electronics, coupled loads, especially power supplies must provide more system services too. Is the industry already feelings the need there?

Dr. Peter Wallmeier (16: 05):

Dr. Peter Wallmeier (16: Yeah, sure. So we call it grid friendly power supplies, and so that you potentially, when the grid requires reactive power, that reactive power generation or consumption is provided by the power supply, by the load. So that basically the load supports the grid when the frequency starts to escape or the amplitude of the voltage starts to escape or fluctuate, I would say.

Prof. Marco Jung (16: 32):

Prof. Marco Jung (16: So you see as requirements for the future, but exist as products on the market right now.

Prof. Marco Jung (16: 38):

Prof. Marco Jung (16: So for loads.

Prof. Marco Jung (16: 38):

Prof. Marco Jung (16: Yeah. Not for generators. It's clear.

Dr. Peter Wallmeier (16: 40):

Dr. Peter Wallmeier (16: Yeah, yeah. Generators is clear. So there is the... When you call it the 61000-3-2 and 3-4 for power factor correction and low THDi. I would consider that already as a standard, deloading the grid from ugly reactive power and harmonics, and then having the intelligence too that the load looks to the voltage, which it's been supplied with and shifts the face between voltage and current. We see that already in the requirements for onboard chargers. So we have to comply with that so that we can shift current and voltage during loading and the load adaptation. I think we... I'm not really aware if there's a standard already in, but we see that for the dynamic and static load management in charging parks, that's kind of that thing, right? So you have a load and this load makes sure that's the point of common coupling, so where the feed in point is, is not overloaded. Either by dynamic tracking or static maximum power level is not exceeded. It goes into the direction.

Prof. Marco Jung (17: 54):

Prof. Marco Jung (17: We see it, I think for huge or big loads maybe in the future there, it's really needed. Let's say maybe for hydrogen production or it's on with big rectifiers for electrolyzers. Maybe there's a possibility that the requirements are still increasing. Now we see it in the past, in the photovoltaics, let's say I started with cosinus phi one, but then later step by step, they must have, or must provide more and more, and think so. If you decrease really heavy centralized power generation to renewable energies. So I think there's a must to do some like this direction.

Dr. Peter Wallmeier (18: 38):

Dr. Peter Wallmeier (18: Yeah. This electrolyzers, they should work when we have too much renewable energy in the grid, right? So when the grid voltage starts to increase and when the frequency starts to increase as well. Yeah? So that's kind of additional load and this overall integration of renewable energies in the grid, that is really a matter of a couple of research institutes. And actually we had a nice... On this 40 years anniversary, where I did this look back, we had a very good contribution from Professor Jansen from Rensselaer University in the US. And he was telling that when you want to integrate renewables, there is a problem of this missing inertia, right? So you have the synchronous generators. And actually the flying wheel could provide that. So, and what I want to say is for everything is a solution. When you have kind of engineers and engineering mindsets, analyzing the problem and then defining or searching for a solution, I think it's doable.

Prof. Marco Jung (19: 43):

Prof. Marco Jung (19: Yeah. I think there exists a lot of research at my institute too. So grid forming inverters that they have, or can provide a large virtual inertia too. So still working because in the past, let's say a lot of knowledge you got from eyelid grids. Now because everything is smaller and you have a lot of knowledge and now it's really important, yeah, to transport it to the interconnected power grids.

Dr. Peter Wallmeier (20: 09):

Dr. Peter Wallmeier (20: Yeah, I remember there was a demo project in Pellworm, I think, that was one where... Actually, when I worked for AT power solutions, we did investigated this power islands. So making sure that we have battery storage, diesel gensets, unfortunately, diesel gensets, so today it would be fuel sets and with gas feed, and yeah, and photo voltaic to make really a renewable and CO2 neutral supply. I think that's also encouraging for young engineers so...

Prof. Marco Jung (20: 38):

Prof. Marco Jung (20: But I think that is a really big topic for the next episodes of Sound on. Power on. because I think we can discuss more than 1, 2, 3, 4 hours.

Dr. Peter Wallmeier (20: 46):

Dr. Peter Wallmeier (20: Absolutely. Yeah.

Prof. Marco Jung (20: 47):

Prof. Marco Jung (20: But what we will see from Delta in the future, I think that is really important for our listeners too.

Dr. Peter Wallmeier (20: 51):

Okay. So I can recommend to go to our webpage at first, because then you can see the wide variety of Delta products and actually Delta has it's... I have to go to the last page of my presentation because then I get here. It's: "Smarter. Greener. Together" that's our slogan. And from the beginning of the foundation of Delta. So it was founded in 1971 by Bruce Cheng. It was always magnetics, switch mode conversion, and then diverting in all kinds of application areas. And today I think Delta is covering everything what you talked about. So we have energy storage solutions. We have fuel cell solutions. We have UPSs, data centers, switch mode power conversion, all kinds of thing, industrial. So go to our webpage and then you see the full variety of it.

Prof. Marco Jung (21: 42):

Prof. Marco Jung (21: Great. So, I hope we can work together in the future. Okay, Peter, thanks for your statements. It was really great and enjoy the conference and the exhibition. Thank you.

Dr. Peter Wallmeier (21: 52):

Dr. Peter Wallmeier (21: Thank you. I will.

Prof. Marco Jung (21: 54):

Prof. Marco Jung (21: To all the listeners, wherever you might be. Thank you very much for listening. We hope you have enjoyed today's episode and have gained some valuable insights. Make sure to subscribe to our podcasts, which is available on all major podcast platforms. If you would like to share your feedback with us, please do so via an email to podcast-pcim@mesago.com.

Speaker 1 (22: 21):

Speaker 1 (22: We hope you enjoyed this edition of Sound on. Power on. Powered by PCIM Europe. Do subscribe and share.