Integrated circuits, or chips are all over the place. Without any noise, these small parts, which are based on silicon and are at the heart of computers, are multiplying. They are fitted into the most varied kinds of equipment, from high technology communications devices to children’s toys, from automobiles to domestic appliances and clocks. With such industrial importance, it should come as no surprise that mastering the manufacturing of chips causes a direct impact on the competitiveness of each country. In Brazil, there is still no consensus on the need for producing chips locally. High investments (about US$ 2 billion) and an excess supply of these devices in the world market are the two most important reasons alleged by those who are against it.
On one point, though, there is unanimity: the ability to design chips is fundamental, because this is the asset with the highest value in the electronics industry, and the one that can create qualified jobs. It was this proposition that drove FAPESP, back in 1994, to create the Special Program for Making Integrated Circuits Abroad, known as a multi-user (PMU in the Portuguese acronym), with the objective of offering postgraduate students in the universities or any other research institution in São Paulo the chance to see and to test the chips that they had designed. This was how the program contributed towards the formation of human resources to design electronic circuits, an important factor for the modernization of the electronic products industry, and also for the installation of a factory for these products in Brazil.
“Without the PMU, we would not have the funds to produce the chips, and it is the tests that determine the success of a project”, explains Professor Jacobus Swart, a director of the Center for Semiconductor Components (CCS), of the State University of Campinas (Unicamp), and a coordinator for the PMU. “The production of a chip is also a factor that arouses the interest of the students for a strategic area for the economic development of the country.”
Long time
The first PMU program was born during the 80’s, in the former Information Technology Center Foundation (CTI), today the Renato Archer Research Center (CenPRA), recalls Wilhelmus Van Noije, head of the Electronic Systems Engineering Department of USP’s Polytechnic School, and a member of the PMU’s advisory committee. This CenPRA program had national coverage and adopted a different technology from the one adopted nowadays. The projects were gathered together and sent abroad in large batches. In this method of work, the time it took to send the designs for the electronic circuits was very long, which resulted in a consequent delay in the chips being returned. In addition, the program was limited to a single manufacturing technique, called Complementary Metal Oxide Semiconductor (CMOS). In this way, FAPESP’s PMU came as an addition to the efforts made by CenPRA’s original venture, offering flexibility in the choice of technology, saving time and guaranteeing the necessary financial resources.
“At the beginning of the 90’s, we were living in a period of very high inflation, and the depreciation of the currency between the request for funds and their release made many projects impracticable”, Van Noije recalls. “With the PMU, things became far less bureaucratic”. For production, the universities that take part in the program use, basically, the services of two European companies – Circuits Multi-Projects (CMP), in Grenoble, France, and Europractice, in Belgium -, that attend to orders from various research centers, universities and companies spread over the world, in a period of from two to four months. These foundries offer several productive processes, for the most varied patterns of integrated circuits. Today, there is a stronger tendency forsystems on chip (SoCs), which is leading to the progressive replacement of the various components that used to be necessary to operate equipment by more developed and multifunctional integrated circuits.
The coordinators of the PMU assess all the projects before the parts are made, to analyze the merits of the work proposed and to prevent any errors from jeopardizing the prototypes, generating delays in the development of the work for a master’s or doctor’s degree. After all, the designs are almost always very complex. “Today, a single chip can concentrate a quantity of tracks (where the information is recorded) sufficient to portray a map of Brazil, with all the rivers, all the streets of the cities, and all the roads that crisscross the country”, Swart comments.
“Since its creation, the PMU has contributed to the formation of about a hundred designers, including both masters and doctors”, Swart recalls. Many of these are working at universities, research centers and companies, like the Motorola design house (read box). Much of the work carried our under the program has been published in specialized international periodicals. Prominent in this work was, for example, the development of a technique to increase the speed of the CMOS digital processors, by João Navarro Júnior, under supervision of Van Noije. At the beginning of June, the article was scheduled for that month’s issue of the prestigious magazine of the Institute of Electrical and Electronics Engineers (IEEE), which is headquartered in New Jersey, in the United States.
Fast processor
Starting from the traditional true-single phase clock technology, the study points to the possibility of attaining double the frequency of the clock (in speed), in processing data by means of the application of new structures formed by the connection of some of the tracks. By doing so, the computer’s consumption of electricity, for example, records a 30% fall. “There are other interesting projects under way, in particular relating to mobile communication”, Van Noije reveals. “The areas with the greatest concentration of work under way are telecommunications, instrumentation and medical equipment”, says Swart.
Another work with integrated circuits that is being carried out at Unicamp may be transformed into an innovation for the automobile industry, contributing towards the introduction of electronic controls and, for good measure, reducing production costs. The project, carried out by Professor Carlos Alberto dos Reis Filho, from Unicamp’s School of Electrical Engineering and Computing, and five students under his supervision between 1996 and 1999, in a partnership with Magnetti Marelli, a parts company from the Fiat Group, is based on the replacement of copper wires by a small wireless communication network; it makes a diagnosis of possible defects in, for example, the headlamp, and tips off the owner. In addition, with the help of specific software, it will be possible to recover the whole operational history of the part. “The solution was not adopted, because there were some internal problems at Magnetti Marelli”, Reis says. The company gave up its exclusive rights over the product, which may be the subjectof an agreement with another organization.
In Brazil, says Swart, at least 40 people conclude the master’s course in microelectronics every year, and 20 the course for a doctorate. And, if the number is still small – in Switzerland, the example given by the researcher, 200 doctors are formed every year, there is no lack of good news in the area. In USP, Van Noije says, the demand on the part of undergraduate students of Electrical Engineering for themes from microelectronics grew a lot, after the optional subjects began to include knowledge in hardware, software and telecommunications. “In the last week for choosing optional subjects, 200 students sought information about these new options”, he says gladly.
Change of course
Another point in favor of microelectronics in the country was the launch of the National Program of Microelectronics (PNM), in August 2001, by the Ministry of Science and Technology (MCT), with the objective of reducing chips import. In 2000, integrated circuits generated a deficit of R$ 1,7 billion in the balance of trade of the Brazilian electronics sector, accounting for no less than 57% of the total negative balance of R$ 3 billion.
To try to turn round the trade deficit in this area, one of the targets of the PNM is to attract two design houses linked to big names in the semiconductor business, like Intel, AMD and Texas Instruments, for example, and to encourage the creation of small Brazilian companies specialized in designing circuits. Design houses are firms that can develop projects to order, aiming at just a few aspects of a large scale development venture, or they create innovative solutions to offer to the market. “It is sowing the seed to spur the intellectual capital in Brazil’s microelectronics industry”, explains Vanda Scartezini, the head of the MCT’s Secretariat for Information Technology Policy (Sepin). “Intellectual property accounts for no less than 66% of the United States’ export income, and Brazil has an enormous domestic market and ought to focus on the models adopted by the developed countries”.
With a forecast for investing about R$ 200 million up to 2005, the PNM selected two anchor states in its first ventures: São Paulo and Rio Grande do Sul. “It is important to add federal efforts to those that are being made by FAPESP”, Vanda reckons. To bring in the larger companies, the MCT is willing to give them support for their setting up, contributing towards the training of the project teams by hiring locally, and, possibly, grants for researchers, through the National Council for Scientific and Technological Development (CNPq), as well as offering funding. In compensation, the companies will have commit themselves to operate in the country for at least twice the period in which they are receiving benefits, to bring at least one integrated circuit project leader for each five project leaders recruited in Brazil, with the objective of disseminating knowledge of the processes. Another basic condition is respect for intellectual property rights in the negotiations.
Finished products
From Vanda’s point of view, attracting an integrated circuit factory is interesting for Brazil, but does not constitute a priority. “When we manage to transform chips designed in the design houses into products that have ready acceptance in the market, industry will come naturally, without being very demanding as to tax conditions”, she foresees. She believes that the best strategy to adjust the trade balance in electronics is not to sell chips, which, at least in the more common applications, have now become a commodity, but the export of finished products, for which the integrated circuit project is gaining more and more importance.
With all these prospects favoring the development of microelectronics in the country, FAPESP’s PMU has managed to form competent labor, and have at least another 15 months ahead of being in force, time enough to gather together new projects of those newly interested in designing chips.
Factory generates projects for chips
An international giant in the telecommunications sector, with world-wide sales of US$ 30 billion in 2001, Motorola started to pay attention to the ability of Brazilian designers in 1997. “That was the year when we started integrated circuit projects in Brazil, with a group of 14 persons”, recalls Antônio Calmon, a director of the company, who is responsible for the semiconductor area in Latin America.
In the beginning, the company took advantage the knowledge built up in Brazil in the period when the market reserve for information technology goods was in force. It was then that Vértice – the design house of the Machline Group, which wound up its activities in 1995 – and chip makers like Sid Informática and Itautec, went so far as to export know-how. And they took such good advantage that, today, Motorola’s Semiconductor Technology Center, installed since 2000 in it Jaguariúna, in the Campinas region, cannot accept orders: its capacity is fully taken up until the end of the year. The center employs more than 100 designers. Amongst them, Calmon says, there are various doctors – and almost 90% have a master’s degree.
This executive tells how dozens of chips designed in Brazil have already been built into Motorola’s products. “We develop integrated circuits for mobile communication, automobiles and networks, among many other applications”. His greatest pride, however, is the conquest by the Brazilian subsidiary of the design of a whole family of microprocessors, a large system that integrates several integrated circuits. “In the whole of Motorola’s history, never has the project for a central chip left the head office”, he reveals.
In April 2001, the company joined the National Microelectronic Program, taking on a commitment to donate equipment in the amount of roughly R$ 10 million to the Ibero-American Center of Excellence for Advanced Electronics (Ceitec), with the objective of making it feasible to manufacture chip prototypes in Porto Alegre (RS). Motorola also announced the donation of equipment to Unicamp, and it established a partnership with this university to set up a laboratory aimed at research into microelectronics.
The project
Special Program for Making Integrated Circuits Abroad Phase 4 (nº 01/04989-0); Modality Regular research benefit line; Coordinator
Jacobus Swart – Semiconductor Component Center – Unicamp; Investment R$ 161,687.50 and US$ 243,000.00
