Recently, it was announced that Professor Willy Sansen, who was known for his research in electronics, circuits and analog design, has passed away. We also found a message from Caeleste for Professor Willy Sansen's condolence:
It is with deep sadness that we announce the passing of Willy Sansen, Professor Emeritus of Engineering Sciences at KU Leuven and member of the Board of Directors of Caeleste.
Willy has left an indelible mark on both academia and the corporate world.
Throughout his illustrious career, he was a beacon of knowledge and innovation in the field of engineering. For more than four decades, he has been a professor of engineering sciences at KU Leuven, dedicated to educating and inspiring countless students. His passion for research and teaching has earned him widespread respect and admiration from colleagues and students alike.
His influence extends far beyond academia. As a valued member of Caeleste's Board of Directors since 2017, he brings valuable expertise and guidance that has shaped the direction of the company with his visionary leadership and unwavering commitment to excellence. His keen intellect, combined with a deep understanding of engineering principles, makes him an indispensable asset to our team.
In addition to his professional accomplishments, Willy will be remembered for his continued dedication to colleagues, friends, and family.
He is a mentor to many, providing guidance and support with kindness and humility.
His contributions to the field of engineering and his impact on Caeleste will continue to inspire us for years to come.
We are with Willy's family and loved ones at this difficult time. May they find solace in the memory of an illustrious person.
安息吧,Willy Sansen。
We will miss you dearly, but we will never forget you.
According to Wikipedia, Professor Willy Sansen received his master's degree in electrical engineering from the Catholic University of Leuven in 1967 and his Ph.D. in 1967, and then his Ph.D. in electronics from the University of California, Berkeley in 1972.
维基百科指出，他撰写和合著了 650 多篇论文和 16 本书，包括《Analog Design Essentials》、《Symbolic Analysis for Automated Design of Analog Integrated Circuits》、《Low-Noise Wide-Band Amplifiers in Bipolar and CMOS Technologies》、《Biosensors: Microelectrochemical Devices, Design of Analog Integrated Circuits and Systems and Distortion analysis of analog integrated circuits》。 Willy Sanse获得了 2011 年IEEE Donald O. Pederson 固态电路奖 ，并且是IEEE的终身院士。
Professor Willy Sansen, go all the way!
附:Willy Sansen自传
I'm honored to be invited to talk about myself, but what I want to say is simple. That's what I've been trying to be a better analog designer. I'll explain what aspects of my life contributed to all this. Maybe this will help you become a better designer too. Analog design never stops. There will always be something that is invented or improved. Progress has always been possible. But in the struggle to achieve the optimal solution, designers can get frustrated. Although he knew he couldn't really find the best answer. It's a constant frustration. Perhaps he will resort to drunkenness, or listen to music, to forget about the unpleasantness. At the 1999 VLSI conference in Hawaii, I began my talk about the idea that analog design is both a science and an art. Art is full of inspiration, and science is full of insight. Analog design requires both. That's how amazing it is to build your amplifiers and filters based on current and voltage. This gives you a lot of satisfaction and frustration at the same time, because there is always the possibility of doing better. Analog design isn't just about current and voltage. In fact, he also married noise and distortion. They can't be seen in the circuit diagram, so it's more difficult to deal with. These are the factors mentioned earlier, but they are crucial in art and music. In this article, I'm going to try to find out where science stops and where art begins, or vice versa, in analog design, does art make you a better analog designer, does music make you a better analog designer, or do you just look for it all online? I'll answer this question based on my own experience. Many of them are related to PhD students whom I have had the pleasure of supervising. The reason I often cite literature is to give their names and the time they graduated. The full list is given in the PhD students supervised or co-supervised by Willy Sansen below. This article is therefore quite a personal checklist of what makes me a better analog designer.
My first electronic plaything
When I was a little boy, I was curious about not being able to see if the current was flowing. Looking at the colorful wires weaving between the boards and connectors full of components, I couldn't tell if it was all working. That's the beauty of a horn. No one could tell if anything was coming in or out, but the music was beautiful. I still think music is the most important reason why so many engineers are lured into the world of analog electronics. I was convinced of this when I started playing music when I was 7 years old. So I started to connect the different devices together: the VCR and the radio, the wideband recorder and the amplifier with the split speaker box, and so on. Not all combinations work, for reasons that I will only understand later. When I was 16 years old, I included a lesson in how to build a radio receiver in middle school. I learned the application of Ohm's law and Kirchhoff's law and built a radio receiver. It is a beautiful radio (see Figure 1). It has a vacuum tube Valvo 385 with a large silver dome and four pins. Next to the tube is an Amroh Mu-CORE superheterodyne coil. Gently flick the mica varactor diode on the front panel and you'll be able to hear the broadcast through your headphones. That's when I thought, maybe I should do more.
Figure 1 My first radioI started building amplifiers and radios with kits from MBLE, a company that does the same work as Heathkit. I renamed the BBO 845 to a 2*10-W ultra-linear hi-fi amplifier that can put out great sound with MBLE's 9710M speakers. I often use it to play music at my sister's parties. A few of my friends were very happy to be able to build an amplifier at such a low price. However, the most amazing radio receiver was my Galene unit. It uses only a crystal and a transformer container, and it doesn't even have a battery! The circuit diagram is shown in Figure 2. It's very simple. It also requires wires as antennas, as well as a 2kΩ headphone. When I was in middle school, I took it under my bed to listen to the radio in Brussels. The metal mattress mesh becomes the antenna. It's amazing that you don't need batteries or solar power to play sound.
Fig.2 Galene radio without power Receiver B is a variable capacitor and D is a crystal. The most important thing I learned in middle school was "mens sana in corpore sano", which means that if you feel good, you can do well. Secondary education in Belgium is the foundation for the rest of your life. I went to the Jesuit school in Alster. It is the true centre of the humanities. Most of his time is spent in the humanities and linguistics like Latin, Greek, French, English and moral education. With so much time spent in the humanities and languages, I had to spend a year in mathematics at university before studying engineering. Another thing I learned from Jesuit school was that even the best things weren't good enough. This thought never left my mind. I would explain to anyone who listens to me that you have to be the best in the world in your discipline. If it's not obvious in 10 years, you'd better try something else.
Of course, always striving for the best will inevitably bring you a lot of frustration and stress. To counteract this, I hung a photo of the Great Buddha in Kamakura in my office. It was the only photo I had in my office for many years. There were no ostentatious plaques, no pictures of my distracted wife and children, just the Kamakura Buddha (see Figure 3). Visitors think I like to travel, and I do. But it has nothing to do with tourism, just the pursuit of a mens sana in corpore sano. To explain: In 1981, I was invited to Japan as one of the first Ten Outstanding Young Persons (TOYP Program) selected by Osaka Jaycees. Over the course of two weeks, ten Europeans learned about Japanese culture, lifestyle, economy, and more. Another Belgian participant was the violinist Edith Volckaert, who sadly died a few years later. I was the only engineer among them. At one point, I spent 24 hours discussing religion and meditating at a Buddhist inn. Here I learned that the idiom of the four emptiness is a path to Buddhahood and more happiness in life. I found it to be an anti-anesthetic obsession with always wanting to be the best. Since then, I have hung the Great Buddha of Kamakura (Kamakura is a city near Yokohama) in my university office. For at least a decade, it was the only photo in my office. It reminds me that most missteps are not as important as they seem.
Fig.3 The most complex piece of equipment I built at the Kamakura Taifo was a high-frequency automatic gain control amplifier measurement device. This is my doctoral dissertation, completed in 1971 under the supervision of Bob Meyer at the University of California, Berkeley. (See Bob Meyer's review of Willy's research, Willy Sansen's Research: A Review). It's all about noise and distortion. I learned grounding and decoupling to get it to work. However, at the time I couldn't say it was all about hardware. We got a lot of help from earlier versions of SPICE when we figured out what we wanted. I even wrote the first Fourier analysis of SPICE to find out what harmonics were present in my circuit. The last electronic device I built myself was a 12-way public address system and a 10V/10A power supply for a digital gate demo board for a small company. I was 23 years old, had completed my university studies and military service, and was eager to enter the field of practical microelectronics design. By that time I had already developed an awareness of measuring equipment. Indeed, I was taught that there is no analog design without measurement, and there is no analog design without hardware.
The last time I worked with hardware was in an in-lab competition organized by my students. We needed to set up a trigger in as little time as possible, and it had to work!Wim Dehaene and I were the first to finish the race (see Figure 4).
Figure 4 Flip-flop with Wim Dehaene
Circles everywhere
So many electronic devices are flocking to our world. An ever-increasing number of tools control our lives. Where did it all start? Where are the simplest electronic devices? What was the first thing to come about? The first thing an electronics engineer encounters in practice is to measure impedance. Indeed, in the whole field of electronics, applying a current and then seeing what voltage is generated, or applying a voltage and then seeing what current is generated, is probably the most basic thing to do. Early examples include the measurement of the input and output impedances of amplifiers, the characterization of resistive sensors, the processing of bioimpedance, etc. More recent examples include the measurement of fruit freshness, the measurement of wall humidity, and the measurement of electrode contact resistance in living tissues.
Figure 5 (a) circuit (b) circles of the RC low-pass filter All similar measurements are subject to the same electrical model—a parallel RC circuit is connected in series with a small loss resistor, as shown in Figure 5. In a Bode plot, this impedance looks like a low-pass filter, but in a polar plot, it is a circle (see Figure 5). It's easy to know that all impedances are subordinate to the circle – the omnipresent circle.
The beauty of a circle is that you know it's a circle. As a result, curve fitting allows for high-precision parameter extraction. A well-known example is the measurement of the impedance of a bipolar transistor, where the series resistance r is the basic resistance of the transistor that is considered to be an amplifier. The parallel RC circuit consists of rπ and Cπ. The lowest frequency is the gain cut-off frequency fT divided by the beta value of the transistor. This measurement fully depicts the characteristics of the transistor. Measurements like this can easily be transferred to MOS transistors with gate leakage currents. Now the problem is that when you insert two electrodes to measure the impedance of an apple, banana or living tissue, apply a continuous current, and you can't find such a circle. It is an ellipse or circle with a center that deviates from the coordinates, as shown in Figure 6. A new parameter has appeared called φ, which is not 90°. It is due to the distribution characteristics of resistance and capacitance in Figure 6. The resistor is no longer a discrete resistor, but an RC transmission line. This is also due to the cellular structure of apples. Fresh apples have a strong texture, resulting in a φ close to 90°. Squishy apples are subject to a lower F-number. That's why parametric φ can be used as a measure of the freshness of apples, bananas or kiwis.
Figure 6 Impedance of tissue or fruit
Now this impedance can be described in Figure 6, where the parameter n relates to the new parameter φ. As for the discrete circuit in Figure 5, the parameter n is no longer uniform, but is between 0.6 and 0.8, depending on the texture structure of the material (freshness of the fruit). For example, when n = 0.8, the φ is about 75° – a typical parameter for soft apples.
Electronics engineers have a problem here. Equations like RCs (where s=jw) have only exponents 1,2,3 depending on the order of the filter,...。 Decimals like 0.6 and 0.8 cannot be generated in electronic circuits. Drawing a Bode diagram is no longer so obvious. Some ubiquitous circles look like ellipses. We can no longer express impedance in terms of constant resistance and capacitance values. In some areas, the resistance value itself depends on the frequency, as does the capacitance. It's a nightmare. We all know that the impedance of a capacitor depends on the frequency, but the capacitor itself is not.
Similar results have been observed for bioimpedance. The impedance between the two electrodes in living tissue (or fresh meat) gives an elliptical curve as shown in Figure 6. In general, at least 3 electrodes are required to separate most of the impedance from the contact impedance. This is a zero-level discussion, however, only two electrodes are used here. This bioimpedance measurement is used to check how well the electrode is in contact with the tissue. Its application treasure trove of electrodes for pacemakers and electrodes for cochlear implants. After all, living tissue repels foreign materials like electrodes. It creates an isolation layer around the electrode that increases the values of R and C from time to time (see Figure 6). Surface impedance will have a similar result. Obviously, all the all-pervasive circles are transformed into ellipses.
The first time I measured bioimpedance was for a fracture stimulator. Fractures in ski accidents often do not heal like other fractures because these bones are frequently rotated and split. The addition of 50 μA of cathode current per electrode promotes healing. We built a four-channel telemetry system with a master oscillator and RF transmitter to input the current, measure the electrode impedance, and send the information out with the wireless telemetry system. This system originated from the relationship with the plastic surgeon J. Brown of the Pellenberg Institute of the University Hospital Leuven in Belgium. A collaboration between Mulier and M. Hoogmartens. It was presented at the 1980 Conference on Biological Telemetry in Japan. However, this system needs to be able to be implanted into living organisms. Many patients are opposed to receiving implantation on non-healing fractures and their interest is waned. All functions are integrated on a 1.3V bipolar technology chip (see Figure 7). This system was our first attempt into the biological realm, so we called it BIO 1. Among the myriad medical applications, there will be many next. It feels great to be able to do something that is helpful to medicine and humanity.
Fig.7 Fracture telemetry and stimulation chip
Our second foray into measuring the bioimpedance of cochlear implants. Similarly, an eight-channel electrical stimulation of the inner ear or cochlea is achieved with a single chip. An electrical current in the cochlea induces hearing restoration. In order to find out which electrode is still active, i.e. which electrode still has low ohmic contact with the tissue, this impedance must be measured. The electrodes that act must then be used to stimulate (Peeters, 1979; Van Paemel, 1990)。 This technology was developed by J. Murphy of the University of Antwerp, Belgium. Prof. Marquet and E. Developed by Professor Offeciers. It was the starting point for a long journey of cochlear implant technology, which is still developing today. We're still measuring bioimpedance, but today it's called a voltammetry sensor. A thin film that filters out specific ions is added to a three-electrode structure. It constitutes a chemical sensor. Provided with the right filter membrane that can be found and attached to a silicon substrate, it can be used to measure glucose or cholesterol (Lambrechts, 1989; Jacobs 1996)。 Due to the increased masking costs of smaller channel lengths twice a year, the design was integrated into multi-project chips (MPCs). We've been doing this since the spring of 1983. Part of the MPC 4 designed in August 1984 using a 3-μm CMOS process is shown in Figure 8. The ear stimulation chip can be clearly seen on the left side. With the preamplifier and the driver connected together, the chemical sensor is front-facing up in the middle.
Figure 8 Ear stimulator and chemical sensor on the front of the MPC4 All of these early biological items are grouped under Bioelectronics. Is that correct?
Does bioelectronics really exist?
Bioelectronics is a hot topic right now, as it merges two of the world's biggest economies: microelectronics and medicine. In fact, the medical application of electronics is the first application area of electronics. In 1750, electrical stimuli were already being applied to all possible parts of the human body to enhance their function (see Figure 9). Due to the misuse of this method, most European countries have banned its use. Later, it resurfaced as a functional stimulus for nerves and many other organs.
Fig.9 Applications of electrical stimulation circa 1750 The first electronic stimulator we made for medical applications was used to help heal fractures, as described earlier. Is this bioelectronics? Obviously, integrated microelectronics with a lot of functions on a single chip. The transmission of data from the body to the outside is also a technical problem. For this purpose, a number of RF coupling systems have been developed. It may be that the effect of an electric current on a tissue or bone is bioelectronics, but it actually falls under the category of physiology. Indeed, our research continues in this direction: the effect of electrical stimulation on cell growth. This research is still ongoing, and it will always be. The reason is that cells are living materials. They are modified by stimuli, whether they are mechanical, chemical, electrical, or magnetic. Parameters of the stimulus, such as amplitude, pulse width, pulse train length, etc., must be modified in order to remain effective. Some of the early chips paved the way for fully programmed systems. Because the technical parameters are changing all the time, this kind of programmability is needed. Whether it's a doctor who changes his mind because of a meeting, or a workshop, or a patient change, or an application field, all have to be expanded. Only the largest markets – such as pacemakers, diabetics, the inconvenience of the elderly – can afford custom chips. Others require programmable chips. One such programmable system is called the Human Internal Regulation System (IHCS). IHCS is a starting point for chips to solve all biomedical problems. It contains three chips, a programmable preamplifier, a programmable stimulator, and a low-power microprocessor that operates at low frequencies. Between 1980 and 1992 I was part of COMAC in Brussels, where I learned a lot about biomedical problems and solutions. COMAC is a group of experts in the field of biomedical sciences (doctors and engineers) with two groups of experts from each of the participating countries. I'm an engineer from Brussels. Professor André Ermans from the Department of Nuclear Medicine at the Université Libre de Bruxelles is my Belgian colleague. COMAC advises on the financing of the European biomedical community. About 35 items were identified, ranging from ECG processing, to artificial skin, to biocompatibility. For each project, a project leader comes and organizes a workshop somewhere in Europe – usually on a small Greek island – and reports back to us in Brussels. Approximately 110 such seminars are organized each year. Of course, you can't attend them all. I picked out telemetry, cochlear implants, ECG processing, biochemical sensors, etc. This is really a great opportunity to learn about bioelectronics. At this stage of my life, I have learned two things. First, there is no such thing as bioelectronics. All aspects of electronics applications in medicine are well covered by other disciplines, such as microelectronic design, biochemistry, physiology and orthopaedic surgery. The person who claims to be an expert in bioelectronics must also be an expert in these areas. Second, there is no final answer, only progress. In a technical field such as electronics, there is a final product that works according to a limited number of technical parameters. This is not the case in the biomedical field. Things are always improving; So, a person's whole life can be dedicated to any of these "bioelectronics" problems.
More on the European level
In the early 70s of the 20th century, we were all a little more European. In 1976, I became a member of the Esprit Advisory Board in Brussels, the first framework project supported by the European Union for research. Most of the money disbursed goes to technology, as 12 companies – Philips, Siemens and STMicroelectronics are among them – say this is Europe's future. Now that Framework Project 7 is underway, the content is more application-driven, which is considered the future of Europe. Around the same time, I became the secretary general of the Semiconductor University Bulletin (SUB). SUB is an organization of European universities with semiconductor technology laboratories and semiconductor physics teaching. It was developed by O. of Twente. Prof. Memelink, W. Aachen Prof. Engl, P. Ravan Ranu Fu Prof. Jespers, R. Murphy of Leuven. Created by Professor Van Overstraeten. By the time I became the Secretary General of the organization, SUB had nearly 30 members from all over Europe and South Africa. Members are required to make presentations in the Bulletin and attend annual meetings organized by one of the members. I found it useful to discuss technical equipment such as furnaces, sputtering systems, ion implantation equipment, etc. For the first time, educational plans in the field of microelectronics have been discussed and compared by different universities. SUB played an important role in the establishment of ideas in the field of semiconductor technology in Europe. It became a gathering place for all semiconductor activity organizations. This European network is required to obtain research funding at the European level. However, SUB has recently disbanded, as the different pan-European projects have provided more than enough opportunities to meet and exchange experiences. An even more effective network to define the project is the European Simulation Research Network (NEAR), which is led by representatives from different institutions and universities. Samani, H. Casier, D. Logie, F. Dielacher, L. Moore, F. Maloberti, Y. Tsividis, J. da Franca, and J. L. Huertas, as well as I, as President, D. Murphy of the European Union Office. Initiated by Broster and M. Cecchini. The aim of NEAR is to produce more analog designers in Europe. NEAR organized workshops and defined projects. We also have a NEAR newsletter that lists current events related to mock design and impressions of the conference. It also has a corner for discussing issues. You can ask the Ph.D. Mock (see Figure 10) for answers to specific design questions. At one point, it had 220 members. Unfortunately, due to a lack of funding from the EU, NEAR disappeared four years later. Still, it's a great initiative. Regardless, there are still many analog designers who ask about it over the years.
Fig.10 Mock Ph.D
In terms of cooperation in our field, the most important European development history, the Symposium on Advances in Analogue Design (AACD) was established. Han Huijsing of the University of Delft came up with the idea of a real workshop on analog design. Rudy van de Plassche and I were in favor of forming such a workshop, as long as three conditions were met: it had to consist of three days, a topic would be discussed by six invited speakers each day, and a panel of experts would be selected at the end of each day. It had to be located in a place where 100 designers could be found within a radius of 100 kilometers. Finally, it must be expensive enough to limit to 60 or so participants. (Otherwise, it's no longer called a seminar.) AACD was launched in 1992 in Scheveningen, the Netherlands (see Figure 11) and is still ongoing. The 20th AACD Symposium will be organized by Michiel Steyaert and will be held in Leuven from 5 to 7 April 2011. It is the best analog design workshop in the whole world because it allows for a lot of discussion. Seriously, this is a must for those who work in analog design.
图11 1992年AACD创始会员:(左至右)Rudy Van de Plassche, Willy Sansen, and Han Huijsing
The most important European event of my career was the European Conference on Solid State Circuits (ESSCIRC). The first meeting was held in Canterbury, England, in 1975, and I also attended. ESSCIRC is held in September in one location in Europe. The next conference will be held in Helsinki in 2011, followed by Bordeaux, Bucharest and Venice in the coming years. Figure 12 shows me in Munich with Professor Doris Schmitt-Landsiedel, President of the ESSCIRC2007, and Anne O'Neill, Executive Director of SSCS.
图12 Willy Sansen, Anne O'Neill, 和 Doris Schmitt-Landsiedel 在慕尼黑的ESSCIRC 2007会议
I attended all of the ESSCIRC meetings, except for two – both of which I was on vacation in the United States. It has become one of the most important gathering places in Europe for designers, managers, academics and others to discuss projects and collaborate of all kinds. That's why I find it so important to participate in all of the ESSCIRCs listed in IEEE Xplore, even for the first time in 1975. The International Conference on Solid State Circuits (ISSCC) is the big brother of ESSCIRC. It is the most important conference in the solid-state circuits community because it has the highest ratio of attendees to the number of papers presented (about 15, which is much higher than other conferences). In Leuven, this number is known as the Sansen factor, because I often use it to compare meetings and establish their quality factors. Having a paper published in ISSCC is enough for the whole world to know. Doing so greatly facilitates the company's project, as all companies participate in the ISSCC. I don't see how a simulation designer can skip this meeting and still call himself an expert. I was the first European to become the project chair of this conference. This was in 2002. A year ago, on September 11, 2001, I was on a plane to Washington, D.C., to attend a sequencing conference. I've never been like this. A few hours before the plane was supposed to land, it was diverted to St. John's, Newfoundland. No explanation was given. We were one of the five planes parked at the entrance of that little airfield, and we waited inside the plane. A few hours later, 17 planes were parked side by side. When I finally got on the phone with my wife on my triple-band phone, I tried to reassure her that I had encountered a small delay and would soon continue my flight to Washington. "Let me tell you what happened," she said. "She's seen it all on TV. I can't convince my neighbors on the plane what happened. It took us a few hours to get out of the plane and into a bus to make our way to St. John's Ice Hockey Stadium. There we watched the big screen hanging on the ice and the plane hitting the tower again and again and saw the numbness. Then we got on another bus and wandered around the city of St. John's to collect the people of the guesthouse, where too many people were dropped. Eventually, we arrived at a rubber mat at a school called the Sacred Heart. I was in room 412 with 15 other people. I remember the scene of us feeling extremely tired and grateful for the Sacred Heart School. Three days later, I managed to get back home. After that, there were no more sequencing conferences. Conference calls instead.
Interesting circuit
Conferences like ISSCC and ESSCIRC report on all aspects of analog design. Most circuits are constructed using basic circuit blocks. We all know that circuit blocks are built with differential pairs and mirror currents. However, there is a circuit module here that deserves more attention. It is a four-input differential current amplifier (see Figure 13) – one of the most interesting and versatile signal processing blocks of all analog circuits. It has up to 4 inputs and 2 outputs and can operate at a supply voltage of nearly 0.5V.
Figure 13 (a) Four-input differential current amplifier (b) voltage input
This circuit is the result of the insertion of two co-source common gate stages that mirror the current, and is done by Rijns for the first time. Only input currents I3 and I4 are used. Two other input currents, I1 and I2, can be added to produce a dual-current differential amplifier. Usually not all input currents contain AC components, but they do. In this case, the output current is given in Figure 13(a). Part of the input power supply can be replaced with a voltage source. A clever way to introduce a voltage source is shown in Figure 13(b), which converts this amplifier into an AB class amplifier. Many designers claim that in order to save energy, all analog circuits will be Class AB in the future. However, distortion would prevent them from taking such a route. Perhaps the biggest challenge with this interesting circuit is that it can operate at very low supply voltages. Regardless of the path, it only needs a VGS+VDSsat voltage. For a 0.3V VT, the supply voltage can therefore be only 0.7V (much smaller if the transistor is operating in weak inverted phase). Much of the attention at the conference was on analog-to-digital converters (ADCs) and delta-sigma converters. I first heard about Δ-Σ signal processing in 1978 from a professor in the field of signal processing at the University of Pisa, Italy. He wanted to use the Δ-Σ technique to extract the basic parameters of the electrocardiogram (ECG) signal. Later, my first Ph.D. worked under the direction of Frank Op't Eynde to apply such a technique to the asymmetric digital subscriber line (ADSL) technique (1990) under the name Alcatel in Antwerp. The beauty of a delta-sigma converter is that you don't have to know everything about it to make it work. All technical parameters are related to distortion and noise, which are the basic limitations of analog design. In addition, it has a mix of analog and digital signals. That's why it's a must-have for anyone claiming to be a better analog designer. My students and I have designed so many of these converters since then! Delta-sigma converters struggle to compromise power consumption, and all ADCs, and all analog circuits face the same challenges. In fact, power consumption is used to increase speed and reduce noise (and distortion). All figures of merit therefore include power consumption, signal-to-noise ratio (SNR) or signal-to-noise distortion ratio (SNDR), and speed. Over the years it has been exciting to see how these figures of merit have changed, but they never look like they are going to be saturated. However, it is not clear whether this is due to the fact that the length of their channels has been decreasing due to technological advances, or because of the ingenuity of the designers.
Online simulation design
Some designers lose their joy in analog design. They can no longer accept the frustration of people who may not be the best designers. They look at the possible options. They look for analog IPs online. Of course, they'll never find what they're looking for – it doesn't exist. However, the search is good. It generates ideas for what to design next. Selling IPs is good business: they can become obsolete overnight, often due to technical upgrades or minor changes to technical specifications. It's selling like a bubble, so who's to say it's not good business to be able to sell in style? Analog design based on web content is a short-term answer. Designers want to skip the Insight phase and solve the problem no matter what. Needless to say, it's a risky design path. The probability that a particular module that has been designed has the same technical parameters as the designer must encounter is very low. Of course, the relative importance of technical parameters is not the same. The modules found on the web are therefore not optimal, in terms of power consumption corresponding to the required speed, noise, etc. Designers who are satisfied with suboptimal designs can indeed find everything online with the help of the search keyword IP module. However, they can never provide the best answers. In addition, picking modules from the web allows designers to skip the most important task in design, which is to build insight. As time goes on, let's not doubt insight. Without insight, a lot of time is wasted. If the design steps are correct, you can get a lot of insights in a short period of time. Analog designs are becoming more and more complex, while digital designs are used to solve simple problems. Only better insight can satisfy the design of eternally complex analog answers. More insight is a long-term answer, an investment in the designer's own future. Analog design is impossible without Insight. This is not the case with electronics in general (see Figure 14). Searching for encyclopedias on the web won't help much.
Figure 14 Analog design requires and obeys more than ordinary electronic designInsightInsight allows designers to "catch" the right compromise among a myriad of technical parameters. Analog design requires compromise. On the other hand, it's fun to make compromises. It gives a person a sense of ability. Analog design, and its compromises, lead to greater insight. It generates intuition, which is necessary to become more efficient for the next design. The question is this: if we agree that it takes 10,000 hours to create an expert, how do we change the web content so that we need less than 10,000 hours to do it? Clearly, the overlay of text with pictures is not the right format, as is the case with most published paper products. It is also clear that the use of the network is always broken. We never devote enough time to fully digest what we read. On the more technical side, the network feeds back short-term or working memory rather than long-term memory that has a growing effect on insight. Using the web to change our brains. Of course, positive changes outweigh negative changes. I just can't believe that having a lot of data can only have a negative effect. How do we turn having a lot of analog circuitry into an advantage? How do we reduce that 10,000 hours? I think there are two requirements. We need to select first, and then the data layer. Some peer reviewers need to decide what is worth reading. For example, a review like the IEEE Conference on Solid State Circuits is a must. We can even go a step further and the reviewers' comments and their names can be published alongside the content. In this way, it will be possible to make a more efficient choice. The second requirement is hierarchy. Technical papers are still written in the same way as texts created centuries ago. They contain continuous text and pictures. Obviously, there is another way to pass on the information behind an analog circuit. At the highest level, we need some indexes, the next level, we need a summary, then all the technical parameters, then a circuit, then all the sub-circuits, and finally, how and why it works. With such a level, it will be more efficient to browse online.
Why music produces better analog designs
Playing music in the early years led to better left-right brain connections. It certainly played an important role in my life. Music came to me when I was 7 years old, and it never left. As a child, I learned music in a local band. After my dad wanted to learn to read, reading music was the natural second step. A year later, I received a gift of clarinet to play. It had to be very small because I was only 7 years old. Here's how I started playing an E-key clarinet, which I kept until I was 16 years old. Then I switched to playing the clarinet in B, which is almost 50% longer and 50% heavier than the one in E, but it makes such a sweet tone that I'm still playing it.
The local band is not a lawn band. It includes clarinetists, who play the role of violinists in symphony orchestras. It mainly performs classical films, especially opera excerpts. I joined that band because my dad played the flute there, and my grandfather was also a member, and he played the bariton – a smaller version of the tuba.
I played in this band until I went to college. There I learned the oboe and played it in the symphony orchestra of the university. In my last year of university, I took a class at the Institute of Psychoacoustics and Electronic Music (IPEM) in Ghent. I found this very interesting because I ended up forming some insight into the relationship between the electrical signals and the sounds they produce. This has never left either.
Now I play the clarinet in a trio that plays classical chamber music. The photo shown in Figure 15 was taken in 2006 after a rehearsal of a Brahms clarinet trio.
Fig. 15 The trio of Raoul Vereecken (piano), Willy Sansen (clarinet), and Arthur Spaepen (cello) has many reasons why the music bears so much resemblance to the analog design. It requires compromise, and it's fun. You can never be perfect. There's always room for improvement. The more you do this, the more insight and intuition you build, which will make you perform better next time. In addition, musical poetry communicates at a higher level. It gives you intuition and the ability to think. It produces feelings and feelings that are shared with others, in fact, richer. It enhances the quality of life.
A better teacher
A good teacher meets two requirements: he is an expert, and he demonstrates his knowledge with enthusiasm. Maybe we all know this, but that's not all. Gardner has written an excellent book on education. The first time I experienced what I call the "dichotomy of education" was when I started going and explaining sheet music reading to other students. By the time I was 14 years old, I knew enough about music so I was involved in teaching music with five other classmates, a new band formed at my school in Poperinge, Belgium. It was a strange feeling. First, you are annoyed that students don't seem to understand what is obvious to you. You feel compelled to show that you do know. It's so easy to allow yourself to show off! On the other hand, you feel sorry for people who don't seem to understand such basic information. Where do you start to help him? I've come to understand that it's enough to help students learn 1% of what you know, and the other 1% next time. In a class, a student immediately understands that at least 1% of what is being taught really matters. He can't leave the classroom without mastering something. At least the same in all respects must be explained in great detail. I never forgot that. It has made me a better teacher. The second time I studied education was when I was 17 years old and in secondary school. Students were encouraged to sit next to their seats in my classroom to discuss answers to math and physics problems. I was amazed at how many people came, even though I didn't know all the answers. From here I learned discere docendi, which means "to teach for learning". If you're forced to explain something you don't know, then there's a crack in that question and the light comes in. My third time studying education was in the Solid State Physics Laboratory at the University of Leuven in Belgium. This lab is related to Professor Roger Van Overstraeten's course on solid state physics, which he started teaching a few years ago. I need to explain to my classmates how to make a countertop diode and how to predict its performance. Being ready means that you know everything about it. However, at this point in time, I wasn't sure if I would be able to express my insights with enthusiasm. My fourth involvement in education was at the University of California, Berkeley, as a research and teaching assistant for Bob Meyer and Paul Gray. The backgrounds of the students here are more diverse than those in Leuven. Sometimes it takes a bit of effort to get them to the bottom line. I've learned a lot from here. I've done a lot of teaching since then, and I'm still teaching analog design. I still want to know everything about it, and I try to present the material with knowledge and enthusiasm. I'm trying to suppress the tendency to show off, but I'm going to make sure that everyone in the class understands at least 1% of what I'm teaching. Every time I explain something I don't understand, more light comes in. That light is the beauty of teaching, especially on infinitely complex topics such as analog design. Now I teach with my slide textbook. These slides I also use to show everything about it. Each slide shows only one delta of Insight. In this way, it is no problem for students to follow the progress. He can read what I said again after that. I call it the Analog Design Essentials. These slides include a read-only CD, so other teachers can use it as well. Quite a few people use it. There is no teaching without content or enthusiasm. Content and industrial cooperation can be developed. Business guidance on what can be successful in the market. They know what to develop (or study) next. That's why I was involved in the design of four 10-μm PMOS for a dedicated automatic telephone exchange (PABX) system at GTE, when I was just establishing myself in Belgium in 1973, just after I got my PhD at Berkeley. I never gave up working with industry. Part of my career has been spent on projects to acquire industry. However, it is not easy at all. "You don't have to hope in order to get started, and you don't have to succeed in order to persevere. That's my motto, so well articulated by famous predecessors of the 15th century, such as Charles the Bold and William of Orange.
In my teaching, most of the technical content is derived from the doctoral dissertation of my university team. Turning the content of your doctoral dissertation into a dozen PowerPoint slides that can be used for teaching is a rewarding exercise. To be honest, sometimes I don't fully understand everything a paper is about until I do that.
Most of my teaching has already been implemented at the University of Leuven in Belgium. The last time I took a class was in May 2008 (see Figure 16).
Figure 16 In 2008, the author's last class in Leuven coincided with a lecture on MEAD education in Lausanne. I was happy to be with many famous speakers in Lausanne (see Figure 17). It is one of the very few institutions that distributes comments from attendees to all speakers. This feedback helped me a lot.
Fig.17 Lausanne speaker: (left to right) G. Temes, W. Sansen, H. Casier, E. Vittoz, and B. Gilbert. My last time on vacation in the U.S. was in Philadelphia in 1984, and Jan Van der Spiegel and Ken Laker (see Figure 18) organized a trip. This stay resulted in a complete textbook, which unfortunately was out of print.
Fig.18 (left to right) Sansen, J. Van der Spiegel, and K. Laker teach teachers in Philadelphia more in return because they have to get the message across. Most of the courses I have recently taught teachers were organized by IMEC last year. These teachers are all from different universities in China (see Figure 19). Their enthusiasm and dating are truly overwhelming.
Figure 19 Teaching in China in 2009 at KU Leuven in Belgium Our ESAT-MICAS organization works with companies from all over the world, including Europe, the United States, Japan and many other regions. Now I've started teaching the same business – an ideal exchange. We also support many of our ESAT-MICAS branches. The first was Silvar-Lisco, later Ansem and ICSense, and most recently zenso, with more to come. In this way, a closed loop is formed.
Advice for students:
Because I've been a teacher for so long, I have some advice for students: 1) Become an expert. There must be at least one little thing in life that you know everything about it. It can be expanded later, but nourishes that "world first" thing. Don't be afraid of depth, read and consult an IEEE journal regularly, and regularly update your knowledge in the field you are particularly aware of. 2) Be known as an expert. It's important for others to know that you're an expert in a small field. Have them guide at a conference or workshop, or demonstrate your knowledge at a local workshop. 3) Play on an international level and be an international communicator. Let's not forget that the world is becoming more and more globalized. Known all over the world. 4) Make a showcase. Sharing knowledge and experience is just as important as communicating. Technically functioning for the benefit of all. 5) Become an IEEE member. This will lead to all of the above recommendations.
Conclusion and acknowledgments
I love analog design because it deals with compromise, like life itself. It pushes you to do your good job. I've tried to explain how one other activity can lead you to better simulation design. The first time a person is exposed to technology, of course, plays a role, but a person's experience with humanity and anything else that shapes your life also plays a role. Finally, connecting with colleagues and friends can illuminate the answer in analog design (life). That's why I want to respect those who are superior to me and who help me choose the right path. They included Roger Van Overstraeten, who was the thesis advisor for my master's students in Leuven and convinced me to return to Leuven after my PhD at Berkeley. I studied distortion and noise from Bob Meyer, who was my doctoral dissertation advisor at Berkeley. I would also like to mention Hugo Deman, who is always a few steps ahead. He worked as a research assistant in the late 60s in the same office as me in the Department of Electrical Engineering at KU Leuven. He was with me at Berkeley in the early '70s, and he worked in Leuven for as long as I did. My PhD and I shared most of my adventures in analog design. I've listed them all on the sidebar and I thank them for all the beautiful and rewarding time spent together in college. I also get a lot of joy from interacting with my colleagues in the ESAT-MICAS department at the university. The first to join me there were Bob Puers and Michiel Steyaert, then Georges Gielen, Wim Dehaene, and most recently Patrick Reynaert. I believe they all came with the aura of better analog designers, and they want to keep it that way. I hope I didn't get in the way of them setting their own future, and that in college they enjoyed as much freedom as I did. They are all pictured in Figure 20 with their wives, which was taken after our annual town hall meeting.
Figure 20 ESAT-MICAS staff and their wives
Source | Semiconductor Industry Watch (ID: icbank) is synthesized from EETOP and others
☞ Business Cooperation: ☏ Please call 010-82306118 / ✐ or to [email protected]