Aerospace Science and Technology Second Hospital Guo Yanying
Why modern advanced weaponry is increasingly dependent on high-end chips
What is a high-end chip? There is no strict definition and unified statement in the world. Nowadays, the chip is generally referred to as a circuit integrated on a silicon block. Then the high-end chip refers to a qualitative leap on the basis of ordinary chips. It is a digital logic circuit or a dedicated circuit that is more integrated, faster, more powerful, and even field programmable. Modern weapons and equipment have widely used various chips since the 1990s. But they have entered the information war since the new century. In the era of networked warfare, the proportion of electronic devices in weapon systems has increased rapidly, and the demand for high-end chips has also increased rapidly. Taking the F-35, the fifth-generation fighter in the United States, for example, it is called the world’s first completely informatized. The combat design requires a fighter because it can be 'seamably compatible' with the current US military's combat system. The F-35 exists as a node of information warfare during combat. The integrated electronic warfare system designed on board is designed to be the largest. Limiting the pilot's ability to perceive the entire battlefield situation; its stand-alone electronic warfare capability can even reach the performance of professional electronic warfare aircraft. In the event of a war, the F-35 can share information with all US weapons and equipment in the air and sea in the US military's combat system. How does it do this? The foreign authoritative Jane's Defense Yearbook says it's loaded on the body. The chip '!
In the field of missiles, radars and aerospace defense weapons, the most used high-end chips, in addition to the commonly used high-speed high-precision ADC/DAC chips, there are roughly three aspects, one is the weapon computer itself chip. Especially the bomb, the on-board computer, The requirements for the chip are extremely strict. Second, the communication chip; is the backbone of many weapon communication control centers. For example, C2BMC in the US Sade is used to achieve global networking. The third is to be widely used in the US radar, the US Department of Commerce. Frequently named DSP (Digital Signal Processing) and FPGA (Field Programmable Gate Array) chips.
The main task of electronic equipment in modern weapon systems is to accept information, output and command after processing and fast calculation, to control the execution components. Taking modern radar as an example, its data rate is up to tens of times per second, that is It takes only a few tenths of seconds to complete the calculation. Therefore, the information processing chip is required to have real-time, fast, and large-capacity calculation functions. The most commonly used algorithms at present are convolution operations and rich-rich transformations. These two operations are a large number of multiplications. The iterative process of accumulation is the easiest to use modern signal processing chips. Since its introduction in 1982, modern signal processing chips have experienced five generations of products, which are quite different from high-speed microprocessors in the past: First, it can be used in one instruction cycle. Completing 32-bit multiplication and accumulating operations, the time is only 1-2 nanoseconds; second is multi-function; and can be processed in parallel; third is the so-called 'Harvard structure'. It is a memory that separates program instruction storage from data storage. Structure. It has a separate address bus and a separate data bus. The two buses are shared by program memory and data memory. This overcomes the data flow. The bottleneck of transmission greatly improves the operation speed.
As for the FPGA, it contains a large number of gates, making the chip more integrated, faster, and more reliable. Especially with the ability to be reprogrammable (reconfigurable) in the system, including programmable logic blocks, programmable I/ O and programmable internal wiring. As information processing, FPGA has a great tendency to replace DSP.
The vast majority of designers are chip users, of course, 'bringing'. For example, FPGA or DSP, famous American FPGA manufacturers, such as Xilinx (Chinese name Xilinx), Altera, etc., set up sales organizations in China. The goods are sufficient, the service is comprehensive, and even the development tools are ready for you. Every time you launch a new model, it will be free for you to train, which will greatly facilitate your design work, and even make a lot of 'innovation'. Of course, these companies earn a lot of money. Full, this is what some people call 'win-win'. A senior academician said well: Many of the results are now based on the use of foreign chips; and it is claimed that the world is number one, this is not ' Take someone else's yesterday to dress up for your own tomorrow!
We can see from the published information pictures of S-300 (such as the S-300-2 system sold by Russia to Greece), and its fire control radar signal processing also uses FFT technology. The radar channel is also realized by FFT operation. Doppler speed branch. But Russia is made up of small-scale integrated circuits (ICs). A distance channel has to be made into a cabinet, while the United States only needs one FPGA or DSP chip (and high-speed high-precision ADC/DAC chip). And a printed circuit board. So Russian electronic equipment often gives people a stupid, big, thick feeling, often looked down on by some domestic scholars. But Russia has his way, does not hinder his S-300 and S- 400 became a world-class air defense missile system.
Why can't we make chips like Americans?
I have never worked in semiconductors and chips. I am not qualified to make irresponsible remarks about current chip issues. But I can reflect some usage and opinions from the user's point of view. On April 18, 2018, the United States announced the 'ban on sale' order. The next day, the Institute of Computing Technology of the Chinese Academy of Sciences held a roundtable meeting of experts in Beijing. The name of the meeting was 'Why can't we make chips like Americans?' But according to reports from the reporters of the Economic Observer, experts at the meeting After a whole day of discussion, there were different opinions, but I still couldn’t get a positive answer. However, the meeting still reflected some important information. For example, there are experts who take China’s supercomputer as an example. It is the fastest in international competitions. To the first. But according to the expert, the first six use of the US chip (author's note: should refer to the 'Tianhe II' supercomputer's six consecutive championships, using 80,000 Intel's Xeon). 2015 In April, the US Department of Commerce decided to ban the sale of related chips to four units in China. In 2016, the 'Shenwei·Taihu Light' supercomputer was equipped with a nationally produced chip' (author's note: Domestic CPU 'Shenwei 26010'. Does the so-called 'piggy' use a part?). Finally won the championship in the international competition with 930 billion floating point operations per second. Therefore, the expert said that 'banned sale' is also a The catalyst has made the researchers feel a sense of crisis, which has accelerated the development of China's independent chips to some extent.
As we all know, the chip industry is an industry that requires a lot of capital and talent investment in the early stage. It is an industry with a long return on investment cycle. It needs to be passed down from generation to generation. The generations of perseverance, hard work, and iterative innovation can only develop to today. At the meeting, experts pointed out that at present, most of the international semiconductor industry giants started in the 1970s and 1980s, and spent a long time and a huge amount of talents in exchange for today's technology accumulation. China's The semiconductor industry regrettably missed a golden age. After the reform and opening up, although we did not spend less effort, and achieved high-profile achievements in the middle and low-end chips, the development of high-end chips depends on market competition. It is difficult to catch up with the level of the United States.
In the 1990s, some forward-looking scientists such as Ni Guangnan and others advocated independent core technologies, and proposed to target the main core chips of the United States (such as Intel's Xeon) to overcome the difficulties of high-end chip technology; In the end, there will be some experts and economists who think that it is better to buy chips than chips, which is the so-called 'market-for-technology' theory; and that there is no need to worry that the United States will easily give up such a huge market and huge profits in China. Market-oriented leaders also worry about big investment and high risks. This makes China's chip industry once again regretted an opportunity.
Recently, a media in Taiwan (Zhongshi Electronics News) published a paper entitled "Revelation of Cross-Strait Semiconductors", which issued a straightforward view on this issue. The article is intended to be: The two sides attach great importance to the semiconductor industry, and the start is similar (simultaneously from Japan NEC and the US RCA introduced production lines. ) The mainland government's support for semiconductors is far more than that of Taiwan. Semiconductor technology is a study that requires hard work and long-term focus. Cutting-edge technology requires high-level professionalism. But now, in the process and In terms of high-end integrated circuits, the mainland is indeed not comparable to Taiwan (?). One of the reasons is that mainland young scientists and technicians are all state-owned enterprises, and there is no sense of crisis. Naturally, there is no sense of urgency, and they are afraid of hardships. I delve into the esoteric principles and key technologies. Some young people are too poorly dedicated and are keen to make a fortune as soon as possible. Some media have a habit of boasting, and they think they are already the first in the world. The article concludes with a 'humility to learn and ask and focus on long-term focus. It is the only way to succeed. The arguments put forward in this article are still objective and worthy of us. Test.
From the point of view of weapons and equipment, I always feel that some experts and scholars lack an important factor in the consideration of 'combat preparation.' Americans regard the chip as a kind of war preparations. Every time we impose sanctions on us, the ban on sales is an excuse. 'Out of national security'. The big strategy of the United States today is to contain China. Can he allow the Sino-US win-win situation to continue for a long time? This reminds us of the 1980s, whether China had to engage in its own Beidou navigation system. Some debates have been expressed. Some people have expressed their opposition. The reason is that the whole world is using the US GPS. It is costly to engage in Beidou. Even if it is built in the future, it will compete with the United States. Then GPS cannot be closed or encrypted, so the United States cannot use it. However, some local wars after the 1990s, as well as military exercises in some countries, the United States has targeted GPS-related functions and applied encryption interference, proving that the opinions of the comrades who advocated Beidou's success are correct.
Former Soviet Union and Russian microelectronics, lessons learned from chip technology
The former Soviet Union and Russia have poor levels of microelectronics technology and chips. The reasons are complicated. However, according to the Russian literature, there are three main reasons: First, the world’s transition from electron tubes to semiconductors in the 1960s. The former Soviet Union’s high-level listeners believe that semiconductor anti-nuclear radiation is far worse. The theory of electronic tubes is to advocate the development of small electronic tubes. This has caused the Soviet Union to miss the golden opportunity from the beginning. Second, the Soviet Union introduced a 'ternary' for a period of time (ie +1, 0, -1 instead of binary 1, 0) ), incompatible with the internationally popular binary, wasted a lot of time and energy. Third, the former Soviet Union distributed the microelectronics industry to the participating countries in order to balance the interests of the participating countries. The Soviet Union disintegrated and led to the fragmentation of the semiconductor industry. During the economic shock of ten years, the economic recovery was sluggish, unable to fully reinvigorate the past glory. Therefore, in the core technical problems, the directional error, the loss is often huge. The stone of other mountains can attack jade, these lessons make us avoid repeating the same mistakes.
Why are there no high-end chips in Russia that can still launch world-class new weapons and equipment?
Since the Ukrainian crisis in 2014, Russia has been subjected to rounds of sanctions in the West, but it is still able to launch world-class new weapons in batches. For example, in the air and space defense, there is a fifth-generation fighter Su-57. Production, the fourth-generation air defense system S-400 service, the fifth-generation air defense system S-500 target success, and the new generation of strategic anti-missile system A-235 combat duty, etc. Western reporter rendering: Russia 'frequently issued A chilling reminder! ' So some readers questioned that Russia's basic industry is poor, and microelectronics technology has always been a shortcoming of Russia. Russia has introduced these world-class new weapons, and lacks world-class microelectronics technology and high-end chips. Russia also emphasizes that self-reliance basically does not use foreign components (the West also bans high-end electronic devices). How do Russian scientists solve this contradiction?
I once analyzed how Russia solved this contradiction in the article 'Why does the Russian advanced weapon development have no chip crisis?'. In short, it relies on the idea of 'independent innovation'. More specifically, the first is to rely on the design of science and technology personnel to 'rise strengths and avoid weaknesses'. Thought. Make full use of Russia's basic technical advantages such as missile flight mechanics, inertial guidance technology and devices, engines, microwave electric vacuum technology, analog circuits, etc., relying on the wisdom and intelligence of Russian scientific and technical personnel to make up for the shortcomings of digital technology. This kind of compensation is not a simple substitute. However, it is an innovation. As the author cited in the article "Chip Crisis", Russia uses his crystal oscillator as the IF signal accumulation, which is even smaller than the integrated circuit. It is a typical successful example. Russian scientific and technical personnel On the one hand, it is out of helplessness, but on the other hand, it has promoted the development and progress of analog circuit technology.
The second is the design idea of the 'system first' of the scientific and technical personnel. The latter means that the level of the weapon components can be general, but the system level, including the weapon system performance indicators and weapon reliability must be first-class. For example, in the air and space defense weapons In the field, from the top leaders of Russia to the chief designer of the missile system, they are soberly aware that Russia's basic industry is far worse than the United States, and should seek truth from facts to develop their own air defense weapons. They have two unwritten design guidelines: one cannot be required The weapon system indicators are more comprehensive than the United States, or the item index is the world's first. Russia has no power to compete with the United States. Therefore, it is necessary to grasp the main indicators, solve the main contradictions, and make the main indicators world-leading. The second cannot require all equipment in the system and The components are first-class, but each equipment is required to innovate independently, try to achieve strengths and avoid weaknesses. The main pursuit of effectiveness and stability, high reliability, and not deliberately pursue the use of advanced components and advanced technology, do not care about the appearance of a little 'stupid, big, thick'. Then the general divisions combined the world-class air defense missile weapon system with outstanding performance and horizontal level. The grammar is hailed by some Western literature as the 'golden rule' of Russian research and development weapons.
The author believes that this has a very enlightening effect on us. In the previous stage, China's scientific and technological community showed some kind of impetuous emotion, that is, all walks of life must fully catch up with the United States, and claimed that many aspects have already surpassed the United States. Take high-end chips, to recognize that the gap between us and the United States is a long-term arduous struggle to catch up. Even then, we can not imagine that all high-end varieties of chips will catch up with the United States, in fact, as long as some major varieties catch up or Close to the world-class level is very good. Of course, the core technology must be in your own hands. Weapons R & D personnel pursue a world-class weapon level, he does not care whether the chip level used in the world is the first.
To be independent innovation, we must have the spirit of innovation. For example, the design of the Russian S-300 system, their spirit of innovation is still worth learning. The first is the band selection, the American patriots use the C band. And the Russian technology The personnel proposed to use the higher X-band to further improve the performance of the radar. However, it was very risky at the time. Not only the development and supply of components were difficult. The key debugging equipment required for phased array radar 'X-band large antenna The production of the near-field test system is also unsure (requires that when the sampling head on a test stand of several meters is used for three-dimensional movement at any point, the positioning accuracy is no more than one percent of the wavelength, ie 0.3 mm). But Russia dares to preempt Occupy this technological high point, and is the world's first near-field test system to develop a large X-band.
In addition, regarding the phased array radar antenna feeding method, American scholars and engineering circles have always advocated the use of branch feeding (through a large number of cables or waveguide branches) to negate the space feeding method. However, Russian scientists and technicians have made major breakthroughs in space feeding. And success. At the international radar annual meeting at the beginning of this century, US radar expert Baden listened to the report of Russian chief engineer Yevremov, knowing that space feeders have been officially adopted on the S-300, he praised Russia. The innovative spirit and perseverance of the scientific and technical personnel. And made such an interesting speech at the meeting:
Western phased array radar developers, who have been concentrating on branch feed design since 1960, tend to ignore and reject this kind of space feed.... By 1988, they thought they thought space feeds. Electrical technology is not mature enough. ....... In the detailed subject outline developed by the IEEEE phased array antenna in 1993-1994, there is not even a mention of space feed arrays. This attitude of the Western radar engineering community has developed the space feed array technology. All were left to Russian engineers, allowing them to enter this blank space energetically.
As the basic scientific research in Russia is very solid, everyone believes that as long as the Russian economy improves further, his microelectronics technology and high-end chip technology will certainly have new development and take-off. Otherwise, low-level microelectronics technology may become Russia's further development of informationization. The bottleneck of networked weapons. So we draw on Russia's experience in developing sophisticated weapons, mainly learning the design ideas and spirit of independent innovation and innovation, rather than emulating his specific practices.
The detours of core devices such as microwave electron tubes in China also show that the road of "changing technology with the market" cannot go
In fact, the lessons of this ZTE incident are not the first. In the 1980s and 1990s, when the United States and the Soviet Union were making great efforts to develop a new generation of microwave high-power tubes to equip a new generation of high-power radars, we Emphasizing the market economy, slamming the drums to shut down the non-profitable electronic tube factory, the university closed the electric vacuum professional that is not popular with students (it is said that the microwave is harmful, the electric vacuum material is harmful). Some experts are worried about the successor. A paradoxical concept is formed in the human brain: the electron tube is behind the transistor, the transistor is behind the integrated circuit, and the integrated circuit is behind the chip. If necessary, it can be spent on foreign markets. Until then, the West is strictly forbidden to sell high-power microwave tubes to China. At that time, I realized that. Russian scientists and technicians are smarter than us in this respect. Their microwave electric vacuum technology is very high and unique, and has played a major role in promoting the development of Russian advanced weapons.
'Patriotic scientists Qi Fali China's core will soon rise'
Every year, China's Ministry of Industry and Information Technology holds the 'China IC Industry Promotion Conference'. One of my students participated in the 2016 and 2017 sessions. After he returned, he conveyed to us the situation of the conference and the embarrassing situation of China's chip industry. The output value of the circuit industry is 500 billion yuan, accounting for 11% of the global share, while Taiwan in China accounts for 16%. According to the operating income, the world's top 20 semiconductor companies, the mainland one does not, and Taiwan has TSMC and other three finalists. China's annual imports of US semiconductor products, mainly high-end chips, amount to US$230 billion, comparable to the one-year funding of our military. Several chip giants in the United States have earned a lot of money in the Chinese market, but for many years On several occasions, on the grounds of national security, I waved intellectual property sticks. I will ban this sale for a while, and I will sanction the house for a while. The purpose is to kill our chip industry and keep the core technology in their hands forever.
This 'chip incident' is of course a bad thing, but it is also likely to become another golden opportunity for the development of high-end chips in China. According to my students, many scientific and technical personnel at the second meeting expressed their willingness to make a big effort. There are also a number of returnees, some of whom have been engaged in chip technology for many years abroad. They want to use their expertise to serve the motherland. They also call on colleagues who are still abroad to not miss the opportunity. They also proposed that 'patriotic scientists are working hard. The slogan of the rise will be the slogan of science and technology. In the final analysis, the dispute over talents is as long as the policy is right, the talents are attached, and all the forces that can be united are united. I believe that China's high-end chips will soon rise.