Application of microchip

The density of devices on microchips has reached the level of neuron density in human brain. Microchips of this level will promote the upgrading of computer and communication industries and greatly change the face of people's production and life. Scientists have been discussing the possibility of implanting microchip memory circuits into people's brains to treat Alzheimer's disease or improve people's memory ability. Portable supercomputers, electronic notebooks, micro translators and portable telephones made of microchips have appeared one after another.

Microchip is a computerized chip, which is encapsulated in a biochemical glass tube compatible with animal tissues and marked with an identification number. The whole microchip is only the size of a grain of rice (13mm ×2. 1mm). Aseptically packed in a needle, it can be implanted into the subcutaneous tissue of animals by subcutaneous injection, which can remain permanently in animals. It provides a safe, accurate, simple and economical permanent logo for animals, and can protect the logo of pets when they are lost or stolen. The microchip is full of a set of numbers that never repeat. When encountering the low-energy electromagnetic wave emitted by the reader, the reader will appear the numbers on the chip on the display screen. The whole induction process takes only a few millionths of a second to complete. Once the chip is implanted in the body, it will be wrapped in a protein bag produced in the animal, and the chip will be fixed at the implantation site for the life of the pet. The chip will not wear or lose, and there is no need to change energy, so even if it stays in animals for decades, there will be no problem, and its service life far exceeds that of most pets. Animals of any age can be implanted with chips. When puppies and kittens start vaccination programs, they can implant chips; Pets such as birds and horses can implant chips at any time.

25 microchips that shocked the world

Excerpted from the article "Special Report: 25 microchips that shocked the world" published by IEEE Spectrum magazine.

Microchip design is sometimes like our life, and sometimes small things will gradually accumulate and become big things. Design a clever microcircuit and then carve it on a silicon chip, and your little masterpiece may trigger a scientific and technological revolution. It all happened on the Intel 8088 processor. Mostek MK4096 4 kilobit dynamic random access memory (DRAM) and Texas Instruments TMS320 10 digital signal processor also triggered the scientific and technological revolution. In the half century when integrated circuits dominated, many excellent microchips were born out of people's disbelief, but only a few of them became the best among them. Their designs have proved to be so advanced, so avant-garde and so advanced that we have to create more technical terms to describe them. We can even say that they have brought us technologies to simplify our lives. Without these technologies, our lives would be boring.

We have compiled a list of 25 integrated circuits (ICs) here. In the buildings built by jack kilby and robert noyce, we think they should be regarded as outstanding people. Some of them have become immortal monuments in the history of chip development. For example, Signitics Ne555 timer of Signetic Company. Some of them have become design examples in textbooks, such as Fairchild Semiconductor's Fairchild 74 1 operational amplifier. Some have sold billions and are still selling, such as the PIC microcontroller of Chipscreen Technology Company. Very few of them have opened up new markets, such as Toshiba's flash memory. At least one of them has become a nuisance in popular culture. Question: What processor can Bander, a bad robot who is both an alcoholic and a smoker in the American animation series Futurama, carry? The answer is: MOS technology is the most 6502 microprocessor.

Of course, this list will definitely be controversial. Some people may accuse us of being too casual and lengthy (this is not the first time someone has accused us of this). Why is Intel's 8088 microprocessor not 4004 (the first microprocessor released by Intel) or 8080 (Intel's most famous microprocessor)? Where is the military RCA 1802 processor with ultra-low calorific value, which used to be the brains of many space shuttles?

You should understand one thing in the introduction, that is, the author, the source trusted by the author and several editors of IEEE Spectrum magazine argued for several weeks before compiling this list. We are not going to list every microchip that has achieved commercial success or made significant technological progress. In addition, the microchip with excellent design but only remembered by the five people who designed it is not on our list. We only focus on microchips that have proved to be unique, fascinating and amazing. All kinds of microchips on our income list are designed by large companies and small companies. Some of these companies have a long history, while others have only recently been established. Most importantly, we collect microchips that have affected many people's lives. These microchips are part of the shocking equipment, which symbolizes the development trend of science and technology and makes people feel happy.

For each microchip, we will use the introduction of engineers and executives who designed them to explain their design experience and the reasons for their innovation. Since we are not working in the IEEE Yearbook of Computing History, we will not sort these 25 microchips in chronological order, type or importance. In the article, we just randomly list these 25 microchips in the order that we think they are easy to read. Please note that the chronological order of the 25 microchips in the article is chaotic.

In addition, we also asked some excellent technical experts about their favorite microchips. We even want to know which kind of integrated circuit has special significance in the minds of Dr. Gordon Moore, founder of Intel Corporation, and Zhang Zhongmou, chairman of TSMC. ..

We also want to know what you think. Are you angry that the microchip you expect is not on our list? Then please take a few deep breaths, make a cup of chrysanthemum tea and join our discussion. It was the summer of 1970, and its designer Hans Camenzind could even recall one or two things about China restaurants at that time. In downtown Sunnyvale, California, the company has three offices, and Camenzind's office is sandwiched between the two offices, with a small area. At that time, Camenzind was a consultant of local semiconductor company Signitic. At that time, Camenzind was not rich, with an annual salary of no more than $65,438 +0.5 million, and a wife and four children at home. Camenzind really needed to invent something outstanding.

Of course, he did the same. In fact, this is the most outstanding microchip in history. 555 is a simple integrated circuit that can be used as a timer or an oscillator. This microchip became the best-selling product of its kind, and soon it was widely used in kitchen supplies, toys, spaceships and thousands of other products.

The 75-year-old Camenzind is still working on designing microchips, but his home is no longer near the China restaurant. Camenzind recalled: "At that time, this microchip was almost unsuccessful."

When the idea of Design 555 was born, Camenzind was designing a system called PLL. With some modifications, this circuit can work like a simple timer. You set the time, and it will run in a specific period of time. It sounds simple, but it's not the case.

First of all, the engineering department of Signitic rejected the idea. Because the company sells some parts, customers can use these parts as timers. This put an end to this idea, but Camenzind has always insisted on his own ideas. He found Art Fury, the marketing manager of Signitic. Frey appreciates the idea very much.

To this end, Camenzind spent nearly a year testing the prototype of the circuit test board, repeatedly scribing circuits on paper and cutting off the Rubylith shielding film. Camenzind said: "At that time, it was all made by hand, and there was no computer." His final design has 23 transistors, 16 resistors and 2 diodes.

555 was put into the market at 197 1, which caused a sensation in the market at that time. Signitic was acquired by Philips Semiconductors in 1975, which is now NXP Semiconductors. The sales of 555 reached billions. At present, engineers are still using 555 to design some useful modules and some useless things, such as designing lights in the style of the movie Knight Rider for the car air intake grille. Without TMC028 1, E.T. may never be able to "call home". This is because TMC028 1 is the first monolithic speech synthesizer, and it is also the "Speak &；; Learn the "heart" of toys (should I say "mouth"? )。 In Steven Allan Spielberg's movies, aliens use it to build their own interstellar transmitters (in the movie, E.T. also uses clothes hangers, coffee cans and circular saws).

TMC028 1 uses a technique called linear predictive coding to transmit sound, which sounds like a series of buzzing, hissing and bonding sounds. The four engineers of that year, Gene A. Frantz, are still at Texas Instruments. He said that this surprising solution is considered "impossible to complete through integrated circuits." The improved microchip was used in Atari arcade game and Chrysler K-cars. 200 1 Texas instruments sold the speech synthesis chip production line to Sensory, which stopped the production line at the end of 2007. If you need to make long-distance or long-distance calls, you can spend about $50 on Yi Bei and buy a "speaking and spelling" toy that is still in good condition to meet your needs. The whole universe was shocked when a juror put the microchip into the computer and started it. This weirdo is stephen wozniak, one of the founders of Apple. The computer is Apple I, and the processor is an 8-bit microprocessor 6502 developed by Mostek. This processor is also the brain of classic computers such as Apple II, Commodore PET, BBC Micro and game systems such as Nintendo and Atari. Chuck Peddle, one of the designers of the processor, recalled that they introduced the processor at a trade show in 1975. He said, "We filled two cups with potato chips. My wife and I sat there selling these potato chips. " The 6502 microprocessor finally stands out, because the speed of the 6502 is not much faster than that of its competitors, but its price is cheap, 25 dollars each, while the 8080 of Intel and the 6800 of Motorola cost about 200 dollars.

Bill Mensch, who designed 6502 with Peddle, said that the breakthrough lies in combining a minimum instruction set with the production program, and its output is 10 times that of other competitive products. 6502 forced the price of processors to drop, which contributed to the personal computer revolution. Today, some embedded systems still use these chips. The greater interest may be that 6502 is the brain of the degenerate robot Bander in Fly Out of a Future, and this information appears in the plot of 1999.

In the real image of Bender's brain, David X. Cohen, the film producer and main author of Fly to the Future, will explain why he chose 6502 as Bender's brain. Operational amplifiers are all silicon boards with similar designs. You always use some of them. You can use them in almost anything, and they can also accomplish some tasks beautifully. Designers use them to make audio and video preamplifiers, voltage comparators, precision correctors and many other systems, which are part of everyday electronic products.

1963, 26-year-old engineer Robert Widlar designed the first monolithic operational amplifier IC――μA702 in Fairchild Semiconductor Company. At that time, each set sold for $300. Subsequently, Widlar improved the design of μA709, and the cost was reduced to $70, which made the product a great commercial success. At this point in the story, Widlar, whose career is in full swing, asked for a promotion. Vedra resigned after his demands were not met. If the National Semiconductor Company won the treasure, it quickly hired Wedra. In the National Semiconductor Company, Widlar helped set up a similar IC design department. 1967, Widlar developed a better operational amplifier for National Semiconductor Corporation, LM 10 1.

Although the management of Fairchild Semiconductor Company has been overwhelmed by the sudden competition, David Fullagar, a new entrant, has made a careful study of LM 10 1 in the company's R&D laboratory. Soon, Fullagar found that although the design of LM 10 1 was very clever, there were still many shortcomings. In order to avoid specific frequency distortion, engineers must connect an external capacitor on the chip. In addition, due to the fluctuation of semiconductor quality, the input stage of IC, the so-called front end, makes some chips very sensitive to noise.

He said: "The front end looks a bit makeshift."

Fullagar started his own design. He expanded the limitations of semiconductor manufacturing procedures at that time and integrated a 30-picofarad capacitor in the chip. How to improve the front end now? The solution is very simple, adding an extra pair of transistors. "At that time, I didn't know how to solve this problem. I drove to Lake Tahoe. " The extra circuit makes the amplification smoother and becomes very stable from one chip to another.

Fullagar found Gordon Moore, director of R&D of Fairchild Semiconductor Company, with his own design. Moore then sent the design to the company's business department. The new chip μA74 1 has become the standard of operational amplifier. The integrated circuits and similar products developed by Fairchild's competitors have sold millions. Today, you can buy thousands of 74 1 chips at the price of $300 per 702 chips. Critics have always ridiculed ICL8038 for its limited performance and irregular operation. This kind of orthodox, right angle, triangle, sawtooth, pulse waveform generator must be a bit unreliable. But engineers soon found a reliable way to use this chip, and then 8038 was a great success. In the end, 8038 sold millions of copies, and found its own application method in countless applications, such as the famous Moog music synthesizer and the "blue box" of the telephone company defeated by the circuit thief in the 1980s. 8038 is so popular that Intersil published a material entitled "Everything you always wanted to know about ICL8038". There is a question: "Why can I get the best temperature performance by connecting pins 7 to 8?" Intersil abandoned 8038 in 2002. But enthusiasts are still collecting ICL8038 to make their own generators and Taylor Ming electronic organ.

At present, neither the public relations department of Intersil nor the engineers who have been in contact with ICL8038 in the company know the exact data of ICL8038. Do you know that?/You know what? Do you know that?/You know what? No chip can bring Intel into the Fortune 500 list? Intel will say yes, that's 8088. This is a 16 bit CPU. IBM used it as the CPU of its own PC production line at that time, and then 8088 dominated the desktop computer market.

In the whirlpool of fate, this processor based on the famous x86 architecture has no "86". 8088 is only slightly modified on the basis of Intel's first 16-bit CPU 8086. After the introduction of Intel engineer Stephen Morse, 8088 was called "castrated version 8086". Because the main innovation of the new chip is not the name, its innovation lies in that 8088 uses 16-bit words to process data, but it uses 8-bit external data bus.

When the 8086 design was nearing completion, the Intel manager kept the 8088 project strictly confidential. Peter A. Stoll, the chief engineer of the 8086 project, also participated in some design work of 8088. He said: "The management doesn't even want to delay 8086 for one day. They dare not tell us that they have modified 8088 in their minds, which will affect the completion time of 8086. One day's task forces us to solve microcode vulnerabilities that used to take three days to solve. "

After the first 8086 was launched, that is, after Intel shipped 8086 exhibits and documents to a design department in Israel, two engineers, Rafi Retter and Dany Star, decided to replace the processor with an 8-bit bus.

Robert noyce and Ted Hoff of Intel said in the article +098 1 for IEEE Micro magazine that this modification proved to be Intel's most successful decision. Comparatively speaking, 8088 integrates 29 000 transistors, which requires less transistor data and is cheaper than 8086. It provides faster processing speed, is fully compatible with 8-bit hardware, and can be smoothly converted into 16-bit processor.

The first PC using 8088 was IBM's 5 150. The price of this personal computer at that time was 3000 dollars. Now all PCs with CPU in the world can regard 8088 as their ancestors. This is not bad for a castrated chip. Before the iPod, the diamond Rio PMP300 appeared. PMP300 was launched in 1998 and immediately became a hot-selling product. But it withers faster than Milevanilli. However, a notable feature of this player is the use of the MAS3507 MP3 decoding chip. This is a digital signal processor based on RISC, and its instruction set can optimize audio compression and decompression. The MAS3507 MP3 decoder chip developed by Weikai Semiconductor allows Rio to load several songs into its own flash memory, which seems a bit funny today, but it was enough to compete with portable CD players at that time. Hehe, isn't it interesting? Rio Tinto and its successors paved the way for the iPod. Now you can hold thousands of songs, and even you can put all Milli Vanilli albums and music videos in your pocket. Mostek MK4096 4 kilobit DRAM (1973)

Mostek is not the first company to launch DRAM. Intel also introduced DRAM. However, Mostek's 4,000-bit DRAM chip has an important innovation, which is called the circuit design of address multiplexing technology. This technology was designed by Bob Proebsting, co-founder of Mostek. Basically, through multiple addressing signals, the chip can access the rows and columns of the memory with the same pin. This makes the chip no longer need more pins after the memory density increases, which can reduce the cost. There will be a little compatibility problem here. The 4096 uses 16 pin, while the memory made by Texas Instruments, Intel and Motorola has 22 pins. What is the biggest confrontation in the history of DRAM?

Mostek put his future on the chip, and its manager began to lobby customers, partners, news media and their employees. At that time, Fred K. Beckhusen, who had just joined the company, was arranged to test the 4096 equipment. Baker Hu Sen recalled that Proebsting and CEO L.J. Sevin came to his night shift around 2 am to discuss with him. Baker Hu Sen said: "They boldly predicted that within six months, no one would hear about or pay attention to 22-pin DRAM." They are right. 4096 and its successors have gradually become the mainstream of DRAM. Xilinx XC2064 FPGA (field programmable gate array) In the early 1980s, chip designers have been trying to play the role of each transistor in the circuit. However, Ross Freeman has a rather radical idea about this. He designed a chip full of transistors, which were loosely organized into logic units. These logic units can be configured or reconfigured in turn by software. Sometimes, many transistors are not used. However, Freeman believes that Moore's Law will eventually make transistors really cheap. He made the right bet. In order to sell this chip called Field Programmable Gate Array (FPGA), Freeman co-founded Xilinx. The company's first product, XC2064, was launched in 1985. At that time, employees were assigned a task: they should draw a sample circuit by XC2064 logic unit like Xilinx's customers. Bill Carter, the former chief technology officer of Xilinx, recalled the task assigned by Bernie Vonderschmitt, the CEO, and said, "He had some minor difficulties in doing this job". Carter is happy to help his boss. He said, "We were all there, holding colored pencils and doing the tasks assigned by Bernie on paper." Today, FPGA sold by Xilinx and other companies are used in many products on this list. Federico Fagin clearly knows the energy and money spent on selling microprocessors. While working at Intel, he participated in the design of two classic products, the original 4004 and 8080. When he co-founded Ziglo with Ralph Unger Mann, a former Intel colleague, they decided to start designing a simpler chip: a monolithic microcontroller.

Fagin and Leon german rented an office in Los Altos, California, and began to draft a business plan to seek capital. Fagin recalled that they had lunch at a nearby supermarket called Safeway, and the lunch was "Camembert Cheese and Biscuits".

Engineers soon discovered that the microprocessor market was flooded with a large number of chips with excellent design. Even if their chips are better than other companies, they can only make a small profit and can only continue to eat cheese and biscuits. Ziglo had to focus on the higher level of the food chain, so it can be said that the Z80 microprocessor project was born.

Their goal is to do better than 8080 and be fully compatible with 8080 software, so as to attract customers to give up Intel. For several months, Fagin, Leon german and Masatoshi Shima, a former Intel engineer, worked 80 hours a week at their desks, drawing Z80 circuits. Fagin soon discovered that the smaller the microchip, the more beautiful it is, but it will hurt the eyes.

He said, "Finally, I had to wear glasses. I became nearsighted. "

The whole design team worked from 1975 to 1976. 1in March, 976, they completed a prototype chip. Mostek's Z80 and 6502 are products of the same era. Like 6502, their success lies not only in the design, but also in the low price (about $25). Bringing products to the market has brought them great confidence. "This is an exciting moment," said Fagin, who eventually suffered from gastric ulcer.

The sale finally succeeded. Z80 is used in thousands of products, including Osborne I (the originator of portable computers) and Radio Shack TRS? 80 and MSX home computers. In addition, Z80 is also used in printers, fax machines, copiers, modems and satellites. Ziglo also uses Z80 in some embedded systems. In the basic configuration, the price of the Z80 today is $5.73, which is even cheaper than a cheese and biscuit lunch. Many years ago (about the early 1980s), when people were still watching Dallas in neon leg warmers, microprocessor designers could seek to increase the complexity of CPU instructions in order to get more calculations in each calculation cycle. The team at the University of California, Berkeley has always been a pioneer in anti-tradition. Their formulation is just the opposite. They put forward a simplified instruction set. They believe that processing instructions too fast will lead to less behavior in each cycle. The Berkeley team led by David Patterson proposed RISC, that is, reduced instruction set computer.

As a pure concept learning, RISC sounds very attractive. But is it feasible? Sun Microsystems is betting on this. 1984, a team of engineers from Sun began to develop a 32-bit RISC processor called SPARC (that is, an extensible processing architecture). Sun intends to use this chip in the new workstation product line. Patterson, consultant of SPARC project, recalled: "One day, Scott McNealy, then CEO, appeared in the R&D lab of SPARC again. He said that SPARC can increase the company's annual income from $500 million to billions of dollars. "

At that time, there was a lot of pressure, and many people outside expressed doubts about Sun's success. To make matters worse, Sun's marketing team has a terrible cognition: it's getting worse. To this end, the team had to swear not to disclose this news to others or even Sun insiders, so as not to leak this news to competitor MIPS Technologies. MIPS Technologies was also exploring the concept of RISC.

Robert Garner, who led the design of SPARC at that time and is now a researcher at IBM, recalled that the first minimum version of SPARC consisted of 20,000 gate array processors, and there were no multiplication/division instructions. There are 6.5438+million instructions per second, which is three times faster than the CISC processor at that time.

Sun decided to use SPARC in future high-profit workstations and servers. The first product based on SPARC was launched in 1987, which is a Sun-4 series workstation. This product quickly occupied the market and helped the company break through the revenue mark of $654.38 billion. All this is just as McNealy predicted. When Fujio Zenggang, manager of Toshiba factory, decided to reinvent semiconductor memory himself, the legend of flash memory invention began. We will think of it soon.

Before flash memory appeared, we had to use magnetic tape, floppy disk and hard disk to store a large amount of data. Many companies are trying to design a solid substitute. However, EPROM (which needs an ultraviolet eraser to erase data) and EEPROM cannot effectively store a large amount of data.

1980, Fujio Zenggang recruited four engineers and started a semi-secret project to develop a memory chip for storing a large amount of data and make it affordable for users. Their combat readiness is simple. "We know that as long as the size of the transistor is reduced, the cost of the chip will also be reduced," said Fujio Zenggang, who is currently the chief technology officer of Unisantis Electronics in Tokyo.

Fujio Zenggang's team introduced an improved EEPROM product, and the memory cell consists of a single transistor. At that time, the traditional EEPROM needed two transistors per memory cell. This slight difference has a great influence on the price.

In order to give this chip an easy-to-remember name, they called this chip "flash", which is also named because of its ultra-fast erasing ability. Now, you will think that Toshiba will put this invention into production soon, and look at the rolling wealth that this invention has brought to the company. You may not know the internal research and development of big companies here. When this discovery was successful, Fujio Zenggang's boss told him, well, forget this invention.

Of course, Fujio Masoka won't forget this invention. 1984, Fujio Zenggang went to San Francisco to attend the IEEE International Electronic Equipment Conference with his storage market drawings. This reminds Intel to start developing flash memory based on NOR logic gate. 1988, Intel introduced a 256K chip, which can be used in automobiles, computers and other devices. This has brought new business to Intel.

This prompted Toshiba to decide to market Fujio Masoka's invention. Fujio Zenggang's flash memory chip is based on NAND technology, which can provide higher storage capacity and is proved to be easier to manufacture. 1989 finally succeeded, when Toshiba's first NAND flash memory was put into the market. In fact, as predicted by Fujio Masoka, the price has dropped.

In the late 1990s, digital photography introduced the application of flash memory. Toshiba has thus become the biggest player in this multi-billion dollar market. At the same time, Fujio Zenggang's relationship with other Toshiba managers deteriorated. Finally, Fujio Zengzeng just resigned and left Toshiba.

Now NAND flash memory has become an important device in mobile phones, cameras and music players.