Imprimir Republish

João Antônio Zuffo

João Antônio Zuffo: From electricity to the digital world

The first to produce a chip in Brazil, the electrical engineer had been actively involved in various projects when computer science was just getting off the ground in Brazil

Léo Ramos ChavesIn the 1950s, in the São Paulo Cambuci neighborhood, the neighbors already knew that when a bomb exploded it was little João who had set it off. At 10, João Antônio Zuffo liked chemistry and made his own bombs and rockets for fun and to amuse the other children in the neighborhood. Inspired by the space race and the hydrogen bomb, subjects in the news at that time, he carried out “experiments” until, one day, he had an accident in the “laboratory” built by his grandfather in the backyard. Wounded by shards of glass embedded in his chest, his mother forbade him from playing with chemistry and enrolled him in an electronics course. Zuffo ended up becoming one of the most respected researchers in the fields of microelectronics and computer science in Brazil.

When studying electrical engineering at the Polytechnic School of the University of São Paulo (Poli-USP), he developed the first Brazilian chip and participated in projects at a time when valves were being replaced by transistors and computer processors were becoming faster, with greater storage capacity. He graduated in 1963 and completed his doctorate from 1964 to 1968, both at Poli.

In 1974 he set up the Integrated Systems Laboratory (LSI) at Poli. Now employing more than 150 professionals, including undergraduate and graduate students, the LSI has participated in the development of several national computer projects, principally those involving parallel computing with microprocessors, in addition to digital TVs and virtual reality caves [special rooms with screens and projectors]. In 1976 he participated in the initial phase of the CPqD project, which was the Telebrás research center and resulted in the first digital telephone central office in Brazil, called Trópico (see report).

At the age of 77 and retired since 2009, João Zuffo continues to go to Poli every morning. Of his four children, two are engineers (Marcelo and Cristina) and two are businesspeople (Paulo and Patrícia). In addition to collaborating with his son Marcelo, also at Poli, he is involved in the LSITec Association, which develops projects in partnership with companies in the fields of microelectronics and digital processing.

The author of 20 engineering textbooks and technical books, in recent years he has written about technological evolution and its future economic and social consequences.

Microelectronics and computer science
Undergraduate and doctoral degree in electrical engineering from the University of São Paulo (USP) (1958-1968)
USP Polytechnic School (Poli)
Scientific production
8 scientific articles, 16 books, 5 book chapters, 36 master’s students and 13 doctoral students supervised

Como foi a produção em laboratório do primeiro chip no Brasil, em abril 1971?
What was it like to produce the first chip in a laboratory in Brazil, in April 1971?
Shortly after completing my doctorate, in 1968, a group of professors at Poli proposed setting up a microelectronics laboratory. Then Professor Alberto Carvalho da Silva of FAPESP [then scientific director], José Pelúcio Ferreira [who was president of the Brazilian Innovation Agency (FINEP)] and Manoel da Frota Moreira [at that time director of the National Council for Scientific and Technological Development (CNPq)] visited USP. They were enthusiastic about what we had proposed and we received total funding of $1 million. This allowed us to build the laboratory, which was inaugurated in April 1970. One year later we built the first integrated chip.

Yes. We had a well-structured laboratory and a very good technician, Jean Serrano. It worked and I showed it to my colleagues at Poli.

Before, had you studied the subject outside Brazil?
No, I had never gone abroad. I had a theoretical design book from Motorola, which was specifically on how to build a chip.  Since I taught a course on integrated circuits, I already knew how to design a chip, more or less.

How did you come to like electronics?
I had liked wires and electrical materials since I was a child. Then I became interested in chemistry. I had old books showing how to produce hydrogen and oxygen, and I dreamed of doing that. My grandfather built a shed in the backyard “so that I would not set fire to the house.” I bought materials at the drug store called Ao Veado D’Ouro [an old compounding pharmacy in downtown São Paulo]. I even bought potassium chlorate, manganese dioxide, which was a stabilizer for chemical reactions, and sulfuric acid, and made hydrogen. I also had glassware for chemistry experiments.

So why didn’t you go into chemistry?
My penchant for chemistry ended in the first year of middle school [6th grade] because my laboratory exploded. I liked to make bombs for the neighborhood, mixing potassium chlorate with sulfur and other things. I did it for fun. That was in the 1950s, when the hydrogen bomb was developed, and the Americans and Soviets began launching satellites and making rockets. The other kids loved it. I hurt myself in the explosion, broken glass was embedded in my chest and my mother took me to the drug store to get bandaged up. I was 10. The work tables were made of crates and I put a test tube on one of them. It rolled over, fell on the ground and exploded. After that my mother forbade me to play with chemicals and enrolled me in Monitor’s radio and television course. That’s when I started to assemble electrical circuits. I still have one today, made in 1954.

At that time transistors were beginning to be introduced in electronic devices.
They were just starting. Transistor radios appeared in 1950, and it was very strange to see people walking along the street, listening to the radio. That was also when television began in Brazil.

You entered Poli in 1958. Did you do some important project as an undergraduate?
In my fourth year, Professor Jaime Gomes asked us to design a valve-based power amplifier. It came out really well and he took the process to Inbelsa, a company that manufactured transmitters for radio stations. I even published an article on the project in a journal that, at the time, was called Eléctron, published by IBRAPE [the Brazilian Institute for Teaching and Research]. Inbelsa used the process, but later folded.

And after graduating?
I went to work at Brasele, run by Rodolf Charles Thom, which manufactured a line for nuclear instrumentation. He hired me to replace valves with transistors in his component line. We had a lot of contact at the time with IPEN [Nuclear and Energy Research Institute] and the USP Physics Institute. At the time, I worked as a researcher at the company, on a FAPESP grant.

Was the instrumentation for reactors?
No, for tests, such as detecting cosmic rays.

Personal archive João Zuffo in front of his “laboratory,” where he made bombs and rockets for himself and the other children in the neighborhoodPersonal archive

How did you return to Poli as a professor?
I was invited by Professor Luiz de Queiroz Orsini. I wanted to pursue an academic career and completed my doctorate in 1968. At that time, USP did not require students to have an advisor. However, about three months before completing my dissertation the university instituted the requirement and Louis Richard Anderson became my advisor. But the person who helped me a lot with the mathematical aspects of the dissertation was professor Léo Borges Vieira.

After joining the microelectronics laboratory, what led you to set up the LSI in 1975?
Working too much, from 1968 to July 1974, led to a retinal effusion, a terrible scare. I was teaching here and at FEI [in São Bernardo do Campo]. I was also working on other projects, like a transistor-based elevator control panel for Átomo Elevadores with Professor Walter Del Picchia—that was the project that marked my entry into the digital era. After the problem with my eye, I had to convince myself to work less, and to not work on several projects at the same time. In late 1975 I submitted a proposal, IF 275, to FINEP and I founded LSI, which was a laboratory combining microelectronics and digital systems.

What what this proposal for?
It was to build a scientific-calculation machine with simple components, but still without integrated circuits. One of the projects we did was a printer to generate signals in Braille and produce books. For that, I used an IBM calculating machine.

And did this machine make it to the market?
We won a prize from the Dorina Nowill Foundation for the Blind, but the Braille printer was not manufactured. The sector evolved very rapidly and other companies with similar devices took the lead.

Why did you leave microelectronics? Were the chips you developed used?
Unfortunately no. The only thing that came of the microelectronics laboratory during that time was the design for building simple transistors, which was used by a company in Belo Horizonte, Transit, which later closed. We still do not have a microelectronics factory in Brazil. One is being built in Belo Horizonte and there is a center in Rio Grande do Sul, which is not yet operating.

LSI did not work with microelectronics?
Every microelectronics process is photolithographic and thermal and I had seen an article in 1966 on the use of refractory materials with multi-layer integrated circuits. As there was nothing further published on thermal processes for producing chips, I decided to learn about and enter this field. Then in 1980 I asked FINEP for funds to enlarge the laboratory. We needed special equipment for the simultaneous deposition of materials, which was done by Edwards, an English company. I went to them to help specify the equipment, because they did not yet have anything that did exactly what we wanted. In the meantime, inflation was soaring in Brazil in 1982 and 1983, and I had money stuck in cruzeiros [the currency at the time]. The Falklands War between Argentina and the United Kingdom also began, and the English no longer wanted to export strategic equipment. I was only able to obtain deposition equipment for LSI much later.

And were you able to specialize in this subfield of microelectronics?
No. In the early 1980s, supercomputers—they were enormously expensive, costing about $20 million to $30 million—and also the more advanced microcomputers started to appear. So then we came up with the idea of developing a supercomputer based on many microprocessors instead of a single super powerful processor. It would be the equivalent of replacing an elephant with lots of ants. It was also the era of parallel computing [in which several processors work simultaneously]. I had some brilliant students who developed several types of parallel computing schemes. In 1996, we presented these computers at the supercomputing conference in San Jose, California. We surprised them, as they did not expect this technology from Brazil. For example, even in the 1980s we had developed a supercomputer for Prológica, a Brazilian computer company, with a Unix operating system.

Did Prológica manufacture this computer?
It did produce and sell some. But it pirated the DOS operating system and was sued by Microsoft. It ended up closing. At the time, many companies pirated operating systems. Another interesting project in the 1990s was related to developing supercomputers and virtual reality, which is in fashion today. It was a medical-application project. The physician would see the virtual body of the patient, using these displays [showing a drawing]. The idea was that he could perform a virtual inspection and then undertake real surgery. The proposal approved in 1992 by FINEP generated a series of parallel computing models and a virtual reality system, but we were unable to develop the display to allow the physician to see inside the patient.

And what happened to the parallel computing project?
We didn’t have the processors ready. Since the set-up was successful, FINEP published an invitation to bid in 1995. Elebra won the competition and we developed a system for the company. Unfortunately, before Elebra produced it, the company went under. FINEP published a new invitation to bid and Itautec was chosen. We developed a custom computer for the company. Itautec produced several types of this equipment simultaneously. We also contributed to a large, mainframe computer using parallel computing.

After parallel computing, what other projects did you do?
With the parallel computing part ready, we began to work on virtual reality and built a digital cave—a projection room in which you can see 3D images using special glasses. That was in 2000. In order to build the cave, we bought special screens and projectors, and we needed to buy an oscillator. At the time, Silicon Graphics had this equipment. They wanted $1.5 million. I had only $150,000, and I said to my team: “We already did parallel computing for Itaú; let’s try to do parallel computing for computer graphics.” Then Marcelo [his son, already a professor at Poli] got involved. We worked on this, and we accomplished our goal. When the folks at Silicon Graphics saw that, they offered us their machine for $100,000, but by then we were no longer interested. Marcelo presented our work at a conference in the United States. Later, Silicon Graphics went under and now everyone uses parallel computing for computer graphics.

Personal archive Poli-USP electrical engineering class in 1963. João Zuffo circledPersonal archive

Did you participate in the discussion on the National Information Technology Policy, which protected the Brazilian computer market from foreign imports in 1984?
I started to participate earlier, in the conference funded by what was then called CAPRE [Coordinating Committee for Electronic Processing Activities], linked to the federal government, and which also included microelectronics specialists. We had annual meetings where we discussed whether policies were good for Brazil or not. There was an important meeting, in 1975, in Fortaleza, during which the foundation of what would become the government’s policy for developing the sector was established. But, in 1979, we met in Rio Grande do Sul and, suddenly, CAPRE was dissolved by the government. It was a shock for the researchers. Then they established SEI [Special IT Secretariat, under the National Security Council]. I was very involved in the discussion, but not in development of the policy. There was a microelectronics commission within SEI, established in 1979, to implement microelectronics in Brazil. They soon invited several professors: me, Carlos Inácio Mammana [professor at Unicamp], José Ellis Ripper Filho [professor at Unicamp], Carlos Morato de Andrade [Poli] and others. Later, SEI published a call for bids to select the companies that would receive government resources. Four companies were chosen: Elebra, Labo, Itautec and SID Microeletrônica. Each company had submitted a more or less simple proposal. The only good proposal was submitted by Itaú, which intended to build a microelectronics factory that would begin operating in 1990. According to them, there would be 2,000 PhDs working at the factory. But, to build that factory, the company would end up losing money for a decade. They wanted to deduct the loss from Itaú’s profits [Itaú is a large bank, part of the same group]. But Delfim Netto, Minister of Planning at the time, did not allow this. So, they set up a small factory in Jundiaí to assemble memory. Today we would have had a factory larger than Samsung’s.

Because, in 1979, Samsung’s management was in our laboratory, asking us about the advantages of making integrated circuits, semiconductors or not. It was a different Samsung, still very small, not the giant it is today. I remember that Philco manufactured transistors in 1966 in Brazil. The Asians came here to photograph the company’s clean rooms in Tatuapé [a neighborhood in the city of São Paulo]. So, we missed out on many microelectronics opportunities. The fact that I developed a chip during that period was not exceptional, because Brazil was on the same level as other countries. Even in industry, there were some integrated circuit factories in the 1970s. Phillips manufactured integrated circuits in Recife, and Texas Instruments in Campinas. But we missed our opportunity.

But wasn’t the idea of restraint of trade precisely to allow Brazil to develop its own industries?
Theoretically it was, but unfortunately the businessmen promised things they did not do. We bought a printer from a Brazilian company and when we lifted up the brand logo there was an original logo from a Japanese factory. I don’t know how SEI allowed this. It even forbade us from importing computers. The university itself was prohibited from importing them. The first PCs here were bought downtown, on Santa Ifigênia Street [where computers were assembled using contraband parts]. We wanted to experiment, but they wouldn’t let us import anything. We didn’t have the freedom to develop new projects. In my opinion, the problem with the IT law was execution, not the law itself.

You have published more than 20 technical books. Now that you have retired, are you still writing?
Yes, but I’m more worried today about the direction society will take as a result of all this technology. Since the 1970s, I have been writing books on digital topics, such as microelectronics, computers and semiconductors, because there was nothing in Portuguese.

And which of your books was most successful?
There are three of them: Subsistemas integrados e circuitos de pulso (Integrated subsystems and pulse circuits), from 1974; Dispositivos eletrônicos: Física e modelamento, [Electronic devices: physics and modeling] from 1976, e Circuitos integrados de média e larga escala, [Medium and large-scale integrated circuits] from 1977, in several editions. I have sold more than 200,000 copies total. But, starting in the 1990s, I began to worry about what will happen to society with everything changing at such a great speed. In 1998 I published the book A infoera: O imenso desafio do futuro [The infoera: the immense challenge of the future], which was pirated and is now on the Internet, for those who want to get it. In 2007 I also published a science fiction book, Flagrantes da vida da futuro [The flagrant life of the future], published by Saraiva.

What does the future hold in store for engineers?
The era of super-specialized engineering is over. An engineer has to be able to quickly delve into a field, and have a strong foundation in mathematics, physics, chemistry and humanities.  The world of tomorrow will not have room for purely technical professionals.

Technology is reaching saturation point. This is already happening with microelectronics. A modern cell phone is not very different from one from two, three or four years ago, it is reaching a threshold where there is no paradigm shift. This has already happened with other technological products. The discussions will focus more on human aspects and not on new technologies. Thus, those producing content, not just software, will have increased importance. Producing films, programs, documentaries, games, and interactive education is the great job market of the future. It is an opportunity for Brazil.

Also in a virtual sense?
Exactly, creating avatars, participating, discussing. This is the education I envision for five, 10 years from now. My dream, for example, would be to develop a low-cost virtual reality room. I think that, today, this could be done for about R$40,000, and install it in every secondary school in Brazil. It could be done immersively, where a person could enter the room, at an even lower cost.

Did you help set up the computer science departments in São Carlos?
It was at UFSCar [Federal University of São Carlos]. I was involved in setting up the undergraduate and graduate courses in the Electrical Engineering Department over a period of four years (1976-1980). I would travel there once a week with USP’s authorization.

Two of your children were your students. How did that go?
Paulo was admitted to Poli, did not like mechanical engineering much, and decided to switch to electrical engineering. He couldn’t stand electronics and failed my class twice. He only passed when I went on medical leave and another professor taught the course. Then he told me: “I hate electronics, but I am going to get my degree.” He graduated and went into business administration, did a master’s degree, and is now working on a doctorate at FGV [Getulio Vargas Foundation] where he is already a professor. My daughter Cristina also studied at Poli, but civil engineering.

What was it like to have your son Marcelo as a student and then as a colleague at Poli?
He was always a good student. He studied at Poli and became an intern at the LSI. The problem with being hired by Poli was a fear of being accused of nepotism. So, he first applied for a position to be a mathematics professor at USP, in the Institute of Mathematics and Statistics, and he was selected. He wanted to show that he was competent. Then a position opened up at Poli for an electrical engineering professor. He applied and was hired. It is great having him here; we have great conversations.