A talented person tells his experience of learning analog circuit technology

fish001

A talented person tells his experience of learning analog circuit technology [Copy link]

I have been studying analog chip design for microelectronics at Fudan University for five years, and I have been working for five years now. During this period, I have listened to the advice of many experts at home and abroad. Recently, at the invitation of a friend, I wrote some thoughts and shared them with you. I remember that when I just graduated from college, I planned to study sensors, but later I entered the State Key Laboratory of ASIC and System in the Yifu Building of Fudan University as a graduate student. Now I think the name of this laboratory has a deep meaning, but I was at a loss at that time. Circuits and systems seem to be two concepts and two levels. Some of my classmates studied electronics and information systems as a graduate student. At that time, I knew that they were "systems", while we were doing analog "circuit" design, so we naturally tended to circuits. Beginners of analog chip design always admire the circuits with strange ideas, especially the most authoritative magazine in this field, JSSC (IEEE Journal of solid state circuits). I used to like reading it very much. At that time, I was determined to read the articles of nearly 20 years and open up the eight meridians. I always looked forward to posting an article. At that time, there were very few articles published in this magazine in China. Even if you were a doctoral student abroad, you would be excellent if you could publish an article in it. When I was studying for a master's degree, my supervisor was Professor Zheng Zengyu. Teacher Li Lian had retired at that time. Yifu Building invited Teacher Li to come every week for guidance. Teacher Zheng is rigorous in academic research and a heroine. Teacher Li is a pioneer in analog circuits in China and is now employed as an expert or consultant in many companies. The book written by Teacher Li in 1987 (Operational Amplifier Design); even now it is a classic. Teacher Li and Teacher Zheng were classmates, so they were very good friends. Naturally, I was lucky to get the guidance of Teacher Li compared to my classmates. The training plan given to me by Teacher Li and Teacher Zheng was: Start with operational amplifiers. So I remember that I started designing from small current sources. At that time, I felt that design was just about adjusting parameters through simulation. But I will always remember Mr. Li's earnest words: op amp is the foundation. Once the op amp design is done, the rest will be easy. I didn't quite understand it at the time. My classmates' topics were all "high-end" things such as AD/DA and phase-locked loops, but Mr. Li and Mr. Zheng asked me to make "primitive" modules. The only paper I published in (Solid State Electronics) (a domestic garbage magazine) was a rail-to-rail amplifier. I was very depressed during the process and envied my classmates' projects very much. But I felt that what Mr. Li and Mr. Zheng said always made sense, so I specifically read JSSC op amp articles, basically all of them in the past 20 years. At that time, I thought I understood it very well, but later I found out that I didn't understand it. The so-called understanding is to truly integrate it, otherwise no matter how much knowledge you put in your head, it will be dead. But operational amplifiers are the cornerstone of analog circuits. Only with a solid foundation can you flourish. I only understood the good intentions of the two teachers after working. In general, what impressed me most at Fudan was Professor Zheng's rigorous style of scholarship and Professor Li's words. After graduating with a master's degree, I went to look for a job and received several offers. My senior brother Sun Liping, Professor Li's last disciple, recommended me to Xintao Technology. He said there was a very capable Chang Zhongyuan, who had a doctorate from the Catholic University of Louvain. I followed my senior brother's advice and went. Xintao had been acquired by IDT for 85 million US dollars at that time, becoming the first successful chip company in China. The person who interviewed me was Howard. C. Yang, the general manager and one of the founders of the company. Howard was a doctor from Oregon State University and an expert in phase-locked loops. During the interview, he asked me to draw a two-stage amplifier with Miller compensation. I was very skilled. He said there was a zero point on the surface. I was very surprised because I had never heard of it. I was confused. Later I learned that this was the first one proposed by Howard in the world. There was a resistor in the equivalent model, which he named Yang's resistor. At that time, I kept nodding my head out of politeness. But they were still very satisfied, so I got in anyway. As for me, the only regret during the interview was that I didn't meet Chang Zhongyuan, probably because he was on a business trip. After entering Xintao, I made up my mind to specialize in a certain field. Because I liked physics, mathematics and philosophy during my undergraduate and graduate studies, I spent some time on these. After work, I had to work hard. Every day, after work simulation and after get off work, I read English original books crazily. The first book was the popular book by Razavi. I read it three times. I felt that I gained a lot. At that time in Xintao, I was a young and brave man who was not afraid of tigers. I should say that I did a very good job, so I was appreciated by Mr. Chang and was evaluated by him as the most potential person in the company. Occasionally, Mr. Chang would come to give me some advice, and others were very envious. In fact, I only remember the experience that Mr. Chang told me during a chat. He said that there are three levels in analog circuit design: the first is to be able to do manual calculations, which means pencil-to-paper. The circuit should actually be calculated by hand, and simulation is just to prove the result of manual calculation. The second is to think after calculation and turn the circuit into something intuitive. The third is to create circuits. I generally followed these three steps. I carefully calculated the exercises at the back of Razavi's book. In the company's projects, I also tried to do manual calculations first. The parameters of the amplifier were calculated first and then compared with the simulation results. Over time, my ability to do manual calculations has greatly improved, and I feel very comfortable with some small signal analysis and calculations. Here is a small episode. Once in a project, the AC simulation of a protection circuit was always unstable. I adjusted it again and again, but it didn't work. I added capacitors here and resistors there. After trying several times, it didn't work, so I asked Mr. Chang for help. Because this circuit is very large, it feels like a blind man touching an elephant. Mr. Chang came and solved it in a few seconds. He looked at it carefully, and then derived a formula to find the main pole and bandwidth expression. Through this incident, I admired Mr. Chang very much, and I also knew the power of intuition. So when I read books later, I would carefully derive the formulas in the book, and then intuitively think about the signal flow. I would not stop until I was intuitive. After more than a year, I finally had a thorough understanding of amplifiers. I felt that I had learned it thoroughly, and after understanding it, I found that I could understand everything. In conclusion: There are two difficulties in amplifiers, one is frequency response and the other is feedback. In fact, the so-called circuit intuition is to think about the circuit from the perspective of feedback. Every time I analyze some "weird" circuits in books or JSSC, I will sigh: feedback, feedback! Then write the analysis experience on the paper. After mastering one field, learning other related fields will have a certain "acceleration" effect. Mr. Chang's way is to let his subordinates study it first every time he does a new project. Before I left Xintao, I made a phase-locked loop. I had never done it before, so I got my classmate's master's thesis, books and many papers to study. After studying for a month and a half, Mr. Chang came to ask me: Have you understood the 3dB bandwidth of the phase-locked loop? I smiled and answered: I have understood it a long time ago. My strong knowledge of the frequency response of op amps is a piece of cake for phase-locked loops. At this time, I have already studied the advanced phase noise and jitter. Not long after, a 30-page English research report was issued, and Mr. Chang praised it highly! . Later, when I was in the COMMIT process, there was a project to modify an RF transceiver chip to convert it from WCDMA to TD-SCDMA. There was a baseband analog filter in it. I had never been exposed to filters before, so I spent two months reading three original English books, the first of which had more than 900 pages, and many papers. I suddenly understood the entire filter field, including switch capacitors, GmC, and Active RC. When proposing a modification plan, since I had a solid foundation in op amps and it was easy for me to understand the filter signal flow when reading articles, I was able to propose a chip circuit principle analysis and modification plan in a short time. Finally, the report was written (also another proud work of mine) and sent to TI. TI was in awe of us. During the conference call, they first said that this report was "Great job!", I didn't understand the English, Julian gave me a thumbs up and said "They think very highly of you". Later, when I went to Dallas, TI respected us very much. When I gave a report, many people came to listen. In short, now I know that the foundation is very important for everything. With a solid foundation, it is easy to learn other things, and the more you learn, the faster you learn. I went to COMMIT in November 2002. My current boss Julian interviewed me at that time. Julian asked me: Where do you think the SOC (system on chip) design link is? I said: It should be the analog circuit, which is more difficult. Julian was wrong, it was the system. I didn't agree with it at the time, and I thought that analog circuit engineers should spend their energy on analyzing and designing circuits. Julian later ran his current company On-Bright and brought me with him. At the same time, he also pulled two people from TI, one of whom was Dr. Fang. I recommended Dr. Zhu to Julian. In the past one or two years, Dr. Zhu and I have admired Dr. Fang. Dr. Fang is the top expert among the Chinese in TI, Strong product development capabilities. On-Bright is currently making power chips. Dr. Zhu and I have been working on it for nearly two years and have learned the importance of systems. Chip design must eventually move towards systems, which is the fourth level of chip design. Circuits are like bricks and tiles, and systems are like buildings. Chip design engineers must consider problems from a system perspective, otherwise they will only see the trees and not the forest. In power chips, amplifiers and comparators are the most common, and the difficulty lies in a thorough understanding of the system. At On- Bright, I truly saw how to make a product, from definition to design, to debug, chip testing and system testing, and finally to RTP (release to production). Julian introduced TI's advanced product development process and project management methods to On-Bright, which opened Dr. Zhu and I's eyes and made us realize the hardships of making products.

秋天的琴湖

Awesome!