Source: Associated Press

The series of interviews in the technology circle was led by Zhao Xiaoguang, vice president of Tianfeng Securities and director of the research institute. Through 1-3 interviews a week, the "global cutting-edge technology" was revealed to find "China's outstanding technology companies" to help investors track technology trends Tap into the future technology leader.

Interviewed guests: Sun Zhen, CEO of Silicon Core Technology, former COO and director of Shanghai Advanced Semiconductor Manufacturing Co., Ltd.

Guest views:

1.In the next 5-10 years, one market that will have the greatest impact on China's economy and GDP is the Internet of Things.

2. The Internet of Things market will generate a lot of customization requirements, which will make the sensors have more value.

3. The performance of the sensor involves two aspects, the first is the accuracy of a single sensor, and the second is whether it can achieve the fusion of multiple data, which can calculate some things through multiple data, and generate new performance and new data. .

4. The core of the sensor company's competition is the product technology itself.

5. In terms of next-generation sensors, I personally think that there are several key directions: the first is the form of combined sensors. Second, it needs to have certain performance and low power consumption. Of course, there are certain requirements for size and cost. Third, packaging technology packages devices with different functions together.

6. Regarding the development of China's semiconductors, how to combine our advantages with the advantages of development in various places and in various links, and come out of our own path of China's semiconductor development, I think this is the key. An important point in this process is that we cannot close ourselves and develop ourselves.

Regarding demand: In the next 5-10 years, one market that will have the greatest impact on China's economy and GDP is the Internet of Things.

Zhao Xiaoguang: We know that the biggest theme of the entire technology industry in the future is 5G. 5G is the entire communications infrastructure. After 5G, there are three main issues. The first is how to enter data, the second is how to handle data, and the third is Output conversion data. In the input data, the core is the Internet of Things, and the most important thing in the Internet of Things is the sensor. So the future space of the entire sensor industry is relatively large.

Let's first ask President Sun to introduce sensors, first, market demand, and demand space; second, introduce the technical route; third, introduce the current competitive landscape.

Sun Zhen: Since September 2012, the company has been regarded as a hard technology company. If you look at motion sensors, IMU (accelerometer and gyroscope) I just mentioned, in the consumer market, only three companies in the world supply a lot, one is Bosch, one is STMicroelectronics, and one is Inven Sense, 2017. The one acquired by TDK. Through these you can know that the sensor entry threshold is very high.

Zhao Xiaoguang: How big is the motion sensor market, and what are the technological improvements in the consumer market?

Sun Zhen: From another perspective, one of the markets that will have the biggest impact on China's economy and GDP in the next 5-10 years is the Internet of Things. There are three main parts in the Internet of Things. The first is data collection, and the second is algorithm. For example, we talk about artificial intelligence (AI) or edge computing (Edge Computing) and so on to improve computing power. The third is communication, which is 5G.

Data is collected mainly by sensors, which are the main converters that convert signals from the physical world into signals from the digital world. There are a wide range of sensors, from traditional air pressure, CIS images, microphones to motion sensors, and we mainly focus on motion sensors.

Motion sensors have a long history, beginning with military industries like rockets. During the Second World War in Germany, the V2 missile rocket first began to use an acceleration gyro to make an inertial navigation system. Later, in all sports equipment, such as cars, robots, and some devices in the Industrial Internet of Things, motion sensors were used to sense its actions .

In the past, these sensors were often made using non-semiconductor methods, which are characterized by high accuracy, but they are large in size, high in power consumption, and high in cost. Since 2000, MEMS has been used to make motion sensors. MEMS is called Micro-Electro Mechanical Systems. It is basically a miniature electromechanical system. The structural principle of the mechanical system is the same as some large non-semiconductor methods. It is a The mechanical device, the accelerometer, senses the parallel movement of the object, up and down, left and right, and the gyro measures the rotation of the object. After the power is applied, the gyro rotates at a high speed and stably. The angle change can be calculated by the change of the angle of the object.

After using MEMS, combined with semiconductor manufacturing capabilities, the entire motion sensor has developed very rapidly. Its core characteristics are that it can achieve small size, low power consumption, and very low cost, so it has been widely used in the consumer market.

Zhao Xiaoguang: After the smart phone, our analysts are particularly optimistic about the development of smart watches next year. What does Mr. Sun think?

Sun Zhen: After smart sensors, wearables is undoubtedly a big explosive market. Because a smart phone is superimposed with a lot of smart functions on the basic communication function starting from the phone, but wearable is completely different. It starts from the original needs of people and is more suitable for people's use of electronic products. The need for convenience and, for example, many wearable people now detect people's body and movement, including physical signs, so they are more people-oriented products.

Wearable products have actually been around for a long time. Some companies in the United States, like Fitbit, started doing them very early 7-8 years ago, but they didn't get up quickly. I believe there are a few big reasons. First, the hardware technology at the time was not good enough. Secondly, the communication technology has not been well developed. Now the wearable sensor technology on the 5G platform includes our motion sensors, and the power consumption has dropped a lot, including things like MCU base, Bluetooth and so on. The differences and progress eight years ago were very large.

Now wearable is not just wearable. Actually it has a very close relationship with mobile phones. A large amount of calculation can be done with mobile phones, but some very personal calculations related to people are done in wearables. Including your watch, the watch is combined with a TWS headset, which constitutes a closed loop. The detection of human signs, including some simple communications, can be done very well.
Regarding consumer applications: Location physical space information is critical, and motion sensors are actually the most important foundation.

Zhao Xiaoguang: In addition to smart watches, what other consumer-level applications are there?

Sun Zhen: In addition to the watches, headphones, etc. we talk about, there are many consumer applications. We define it as an industrial product, such as a sweeping robot, and also a pos machine. Now, more and more motion sensors are installed on the pos machine to track it. I believe that such a concept will soon be applied to high-value logistics on a large scale.

At the same time, all smart home related applications, including even smart door locks, are also increasingly equipped with some positioning sensor systems based on motion sensors.

Zhao Xiaoguang: In terms of consumer applications, competition has been fierce in the past and prices have fallen rapidly. How does Mr. Sun see future price trends and how does Silicon Core respond?

Sun Zhen: Price competition is of course an eternal theme in the consumer market, but I personally feel that because the sensor provides two concepts, one is function and one is performance. More and more applications are related to performance. Good performance requires more accurate sensors, or the following ASIC processing will be more complicated. In the past, mobile phone applications did not require sensors mainly at the functional level. Without high-performance requirements, everyone may simply pursue low prices. However, with the requirements for performance, future sensor trends will find one among performance and price. The right balance.

Zhao Xiaoguang: The performance that President Sun said mainly refers to precision or other?

Sun Zhen: The performance of the sensor involves two aspects, the first is the accuracy of a single sensor, and the second is whether it can combine multiple data. Through the multiple data, some things can be calculated to generate new performance, new data.

Zhao Xiaoguang: President Sun just mentioned the headset. Will the headset also become a smart product that can detect sports sleep?

Sun Zhen: Headphones will definitely be a big market for intelligent products, because basically, you can see that mobile phones, watches, and headphones are completely connected to human eyes, ears, and touch. Ears are very important. One of them, detecting the needs of sports sleep in this, is not a problem for sensor technology.

Of course, the current TWS headset is mainly focused on the functions of the headset itself. Just like the beginning of the mobile phone, everyone focused on the basic functions of the mobile phone and communication. However, more sensors will gradually be integrated on this headset, including a light sensor in it, an acceleration sensor, or now it becomes acceleration + gyroscope in it.

Through some algorithms, functions such as heart rate detection can be combined, which can be achieved completely. Collection of physical information such as exercise and sleep just mentioned.

Zhao Xiaoguang: Will the motion sensor be integrated with other sensors in the future and will it develop in a modular manner?

Sun Zhen: We believe that modular sensors are definitely a direction. With the Internet of Things, the basic configuration of each unit is the configuration that requires communication and the configuration that requires algorithms. This configuration is a foundation. On this If there are more sensor modules in it, more different signals can be received, of course, the amount of data and value generated will be more.

The difference between Silicon Core Technology and quite a few sensor companies is that other companies often make a sensor, and we make a combination sensor. There are several different sensors in one direction of the combined sensor. In the other direction, we also combine the ARM Base MCU on the sensor, including Bluetooth, to form such a module that generates data, calculates data, and transmits data. In terms of modularization, another aspect of your sensor is to extend one of the antennas of your sensor. For example, in the case of city navigation, in addition to motion sensing + gyroscope, magnetic sensors need to be put together as a fusion algorithm to enable better positioning.

About mobile trends: new sensor applications will be added

Zhao Xiaoguang: President Sun thinks that in addition to the current sensors on mobile phones, what types of sensors are most likely to be added to mobile phones in the future?

Sun Zhen: I think the new sensor is added to the mobile phone. First, it may take advantage of a large computing power platform; second, rely on a larger device such as a mobile phone to carry a new feature.

For example, in terms of vision, we may put a so-called non-intelligent infrared sensor on it, and you will have some new feelings, such as the image can be seen in a very dark place. In addition, there are some new sensor functions such as handheld ultrasound. Will it be added to this and the mobile phone will be used as the hardware base to truly achieve the effect of home use.

There are many sports that we have now. Generally, we go to the gym for coaches to teach you, but after returning, many of the exercises are done by ourselves, and I don't know whether it is accurate or not. If you put some sensors on your body and provide calculations, simulated movements, and corrective movements by your phone, it may be a great application.

Zhao Xiaoguang: This point made by President Sun is very good, in fact, it is a deeper combination of data input and application. For example, watches test our heartbeat and electrocardiogram, but what these data represent requires further data analysis and processing. Next question, how to look at the application of medical related sensors in the consumer level?

Sun Zhen: Just now I talked about handheld ultrasound. This is the consumer medical application we defined. I also defined an industry-like application. What is the industry? This product has similar sensor performance requirements to industry, but it does not need to provide 10, 20, 30 years of reliability as industrial products. It is a consumer industrial product, and the cost cannot tolerate as high as industrial products. So it is between industry and consumer. We define it as a kind of industry. The medical care and consumer medical care you just mentioned are similar concepts. Including the handheld B-ultrasound mentioned earlier, it is basically a probe made of MEMS technology combined with piezoelectric materials. It has a data line connected to the mobile phone. The mobile phone provides powerful computing capabilities, making the B-ultrasound The function is performed on a sensor-centric handheld device.

Zhao Xiaoguang: What is the main use of handheld ultrasound?

Sun Zhen: Just like a sphygmomanometer entering the home, hand-held ultrasound entering the home will also be a major breakthrough. Simply speaking, you can detect many items that you must go to the hospital before. In addition, the home-developed blood glucose measuring device is also handheld. Traditional blood glucose measurement is performed by stabbing bleeding, but now the sensor is attached to the human skin like an adhesive tape, and the body fluid of about 10 microns is measured by the epidermis Distribution, you can understand the situation of blood glucose.

In medical applications, the more difficult is the algorithm. Because the human body is a system, no matter which single sensor we use to measure only a part of a vital sign, it is not easy to accurately reflect the body based on such measurement information. Change, if there are different sensors, measure different data, and fuse through algorithms, there will be a chance to produce a more accurate description of the physical condition.

The traditional Chinese medicine pulse diagnostic instrument that we talked about some time ago is actually a good example. Traditional Chinese medicine understands a person's physical condition by diagnosing pulses and combining pulse theories of different genres.

We have also studied such a sensor system, using a matrix of pressure sensors to measure weak blood flow on the three positions of the inch / close / foot The characteristics of hydrodynamics, combined with motion sensors, turn physical phenomena into a set of data, but we do not know what kind of physical changes this person reflects. We need to restore this data to the original physical status, and combine pulse theory with Chinese medicine to reflect human vital signs. This process is the process of algorithms.

Zhao Xiaoguang: What do you think of applications in AR and VR?

Sun Zhen: AR and VR are products that have been proposed relatively early. They have been around for many years. Why haven't they started? There are two important reasons. One is that there is no 5G, so there is no way to transmit a large amount of data. Of course, these data cannot be of great value. Second, people are uncomfortable to use. Because AR and VR products are mainly related to human eyes and sports balance systems, people are very sensitive to this type of accuracy change and are very susceptible to uncomfortable reactions, such as dizziness. Motion sensors play a big role in this type of application, especially IMU. We have an IMU that was used in Google's DaydreamVR.

Zhao Xiaoguang: We think that Apple Watch will use blood glucose testing in the future. What does Sun think?

What is the principle of blood glucose detection?

Sun Zhen: There are actually many different ways to measure blood glucose. Traditional measurement usually requires blood collection. Wearable products such as watches generally use optical methods to measure the spectral changes of some elements in the blood to achieve blood glucose detection. However, the spectral signal that characterizes blood glucose is weak. What wavelength of light is used and how it can be done effectively is a considerable mechanism study. A device, whether it is a watch-based or bracelet-based, or a single blood glucose measurement product, if it can achieve no blood collection, I think it is a very good product.

Zhao Xiaoguang: The traditional Chinese medicine just mentioned is a good case. We are optimistic about the application of telemedicine after TV socialization.

Sun Zhen: A solution such as telemedicine after the socialization of television will play a positive role in promoting the popularity of Chinese medicine. In fact, the biggest difference between traditional Chinese medicine and western medicine is that western medicine can quantify it. Everyone takes an X-ray picture or an electrocardiogram picture, which can be seen and judged in various places. Traditional Chinese medicine is more difficult because it has great relevance to people. The interpretation of the same pulse may be different for different Chinese medicine practitioners. The same person and the same Chinese medicine doctor will have different pulse veins in the morning, at noon, and at night. How to establish a quantitative system of data is the most critical. The sensor will definitely play a very, very important role.

Only after the sensor outputs data will there be a basis for quantification.

Because the human body is a complex system, I believe that this sensor is definitely a combination sensor, not a type of sensor. A type of sensor can only intercept its related data, and it is difficult to reflect changes in the surrounding environment and changes in human physical signs. The data has an impact, so the combination of sensors is needed to judge.

Zhao Xiaoguang: Medical sensors were discussed earlier. In fact, medical data also needs to be closely combined with movement position data to make sense?

Sun Zhen: A very simple example, measuring heart rate. People sitting there, walking slightly and violently, the measured results will be completely different. If you only rely on a heart rate monitor, there is no way to judge your state, you can only measure This change in exercise state, combined with two types of data, heart rate and exercise, has a more systematic understanding of the overall condition of the body.

Zhao Xiaoguang: Very reasonable. If combined, is it a motion-integrated organism or a bio-integrated motion sensor?

Sun Zhen: As for the movement-integrated organism or the bio-integrated movement, I think it depends on what is measured. Generally speaking, the biological data should be a relatively direct interception, but it knows what the organism is in by the movement, and can intercept the obtained interception Make some corrections to the data.

Zhao Xiaoguang: Basically speaking, on the consumer side, I think President Sun made it very clear. We can see that the first is that the entire movement, that is, the physical space information of the location is very important, because the most important thing to see on the Internet + now is Position information and motion sensors are actually the most important foundation. It will be better integrated with other sensors in the future to produce more accurate data and be combined with various applications.

Second, we can see that the trend on mobile phones is various new sensor applications. At the same time, I just mentioned that the application of smart phones, watches, AR, VR will be up.

Regarding the Internet of Things: "A lot of customization needs will make sensors more valuable."

Zhao Xiaoguang: In the second part, we discuss the Internet of Things in detail, part b. From the perspective of upstream sensors, what areas does Sun think the Internet of Things will come first?

Sun Zhen: The Internet of Things market is very interesting. On the one hand, I have not seen the so-called Killer Product, which is a killer product from the perspective of traditional consumer electronics. On the other hand, the Internet of Things is indeed all around us and is a very wide market. I personally think that it is difficult for us to see a so-called killer product like a mobile phone on the Internet of Things, but many applications of the Internet of Things are already quietly on the rise.

For a very simple example, the camera on the road is not a traditional consumer product, but if you look at the development of this industry in the past 5-8 years, it has actually pulled several companies into large companies. Now, including hardware companies and algorithm companies, if you go to see several unicorn companies of artificial intelligence, it is basically closely related to the development of this market.

Therefore, I think that we must look at the Internet of Things market aside from the concept that we used to look at the consumer market.

First, the IoT market is very broad and may require a lot of customization, not just general-purpose products. A lot of customization requirements will make the sensor more valuable.

Secondly, I feel that the Internet of Things market is like the concept of Internet ++. The Internet of Things is divided into two categories, one is the Industrial Internet of Things, and the other is the Consumer category. We used to talk more about the Industrial Internet of Things. In fact, the Chinese Consumer Internet of Things will develop more rapidly.

Zhao Xiaoguang: I think the important prerequisite for the development of the Internet of Things is the 5G foundation.

Sun Zhen: Of course there is no doubt that 5G is a big foundation, because the fundamental of the Internet of Things is data. You see the development of the entire society we talked about, from coal to oil to electricity, and finally to data, the platform for data flow is 5G. If there is no 5G data flow, it will not work. If there is no smooth mobile data, the value of the Internet of Things will be greatly reduced. I think this is a fundamental and basic thing.

Zhao Xiaoguang: In fact, the emergency industry and the energy Internet of Things are all typical big markets in the future. Mr. Sun, does the consumer Internet of Things refer to smart homes? Home Internet of Things?

Sun Zhen: The consumer Internet of Things may not only refer to smart homes. For example, the express delivery brother and POS machine I just mentioned also belong to the concept of the Consumer Internet of Things.

We also talk about some areas of agriculture, food traceability, etc. Consumer Internet of Things concept. It is still quite wide. Even if you are talking about the sweeping robot today, we are just seeing the robot walking around. Will it produce so much data that it will eventually produce its value? Where can its value be reflected? If it can reflect another point of the Internet of Things.

The agricultural Internet of Things we just talked about is how to manage those animals, livestock, cattle, pigs, and chickens. This is an angle. Another angle is to go to the food chain of each of us. In this way, how can we trace the source or ensure the condition of the food, whether the pig's growth cycle, living conditions are sick, and usually exercise, if there is an animal wearable device, It also provides a lot of data to reflect the condition of these foods. Not only is it safe, but also whether these foods are healthier.

And once these data enter such an IoT based on the blockchain, of course, we can extract a lot of different information from these data.

Back to the Industrial Internet of Things, the Industrial Internet of Things in Europe, America, and Japan may speak of a vertical system. The Industrial Internet of Things we are talking about focuses more on the horizontal system, how different devices interconnect with each other, data transmission, and so on. I think these are two Different dimensions are described.

Zhao Xiaoguang: In addition to sports and medical treatments, what are the most promising sensors in other fields? What about optical sensors? President Sun mentioned the blockchain-based platform just now. I also think that the future of blockchain is an important part of the industry's in-depth information technology.

Sun Zhen: I think that in the past, due to technical limitations, the development of optical sensors was slow, and there were not many companies doing it. However, with the explosion of consumer electronics, the pull on this type of sensor is very large. Austrian semiconductor microelectronics AMS is a good example. It has transformed from an IDM company that was a pure semiconductor foundry company to an optical sensor in the past 15 years. Ten or two hundred million dollars in sales 10 years ago, and now almost a billion dollars.

Recently acquired Osram, very fierce development.

The main application that pulls it is the optical sensor on the mobile phone, including current facial recognition and so on. Combined with artificial intelligence, the application of facial recognition will have great development, far exceeding the consumer market.

A feature of facial recognition, like fingerprint recognition, is close. Different optical sensors can cover middle and long distances, such as lidar in the automotive field, etc., which are essentially different applications of optical ranging.

Zhao Xiaoguang: How do you think that global MEMS leaders are operating under the IDM model, and where are the growth opportunities for MEMS foundry companies?

Sun Zhen: Frankly speaking, is MEMS company IDM?

This is a false proposition. Looking at the entire stage of semiconductor development, everyone is IDM at the beginning. Why did it develop into Foundry? In fact, it is very simple. Because the investment in the factory is too large, this investment is not only hardware investment, but also research and development investment. In commercial form, Foundry is only responsible for manufacturing capacity investment and R & D of manufacturing capacity technology. Just like TSMC today, TSMC now accounts for 60% of the entire Foundry market and has almost achieved more than 40 billion US dollars in sales.

A good example is AMD. AMD was also IDM, and it has not been very good to compete with Intel on the CPU for decades. It sold its Fab 7-8 years ago (these factories are the predecessors of today's Global foundries), AMD became a design company, and it is now tied to TSMC.

TSMC ’s 5nm production makes its CPU all at once There has been a rapid improvement, and the degree of some fields has now competed with Intel, which was unimaginable.

The same is true for MEMS companies. If you look at the top 30 companies in the world, they are basically IDMs, all of which have evolved from system companies.

Most of them have evolved from system applications.

The reason is simple, because there was no Foundry to help you produce MEMS products. If you can make a small factory yourself, research and develop your own production, as the volume exceeds the capacity of your own factory, you may also find a foundry. Of course, it still has some differences, because performance is very closely related to manufacturing technology, and your own factory may still be retained for a long time. At present, the most successful design company is InvenSense in the United States, which was acquired by TDK 3 or 4 years ago. It is a very typical Fabless design company. The first-generation Apple phone ’s screen-turning function is implemented by its sensors. Do Foundry.

Therefore, I don't think that it is possible to talk about IDM and Fabless in a more general way, which is more conducive to the development of sensor companies. There are very successful companies.

If you target different sensor products and technologies required by different sensor products, analyze the contribution of its manufacturing to the performance of this product, and then look at which model of IDM or Fabless design company is better for this product and this technology. It would be more appropriate, maybe more reasonable.

We often have the idea that we hope that all resources are available, and of course we will not ask for help. But there is a price to owning resources. It is not a free lunch. The cost of owning a Fab is not just to have the money to build a factory. It is good to buy hardware and equipment. In fact, building a factory and buying equipment can not become productivity. It still requires a lot of technology investment, which is why TSMC can achieve a global market share of 60% today. The main global wafer production below 12 nanometers is TSMC.

Continuous technology research and development investment, competitive production operations, these are all It is not available to ordinary design companies. This is also a big misunderstanding of semiconductor factory construction in China now.

Zhao Xiaoguang: What is the situation of foundry companies?

Sun Zhen: I mentioned the previous MEMS companies, such as Bosch, ST, etc. Many of them produced their own. There are several major foundries such as TSMC, as well as Silex in Sweden, Dalsa in Canada, etc., as well as Sony, Sony doing IDM, and also foundry.

Zhao Xiaoguang: Please introduce General Sun's situation and the company's core competitiveness. How can it rise from the breakout of foreign capital?

Sun Zhen: I personally think that the core of the sensor company's competition is the product technology itself, because sensors are similar to analog devices. Analog devices are very different from digital devices. Digital devices often depend on the size of the chip. The sensor is like an analog device, and its core is performance. Often, the performance is increased by 10%, and the price may be increased by 10% -20%. The core of performance improvement lies in product design and production process research and development.

First, we hope that our research and development will enable the technology of the entire product to reach the global level of 1.5, and achieve world-class performance in 2-3 years. It may not be a very realistic goal for a start-up company, because our first-class company They have been doing it for 10, 20 or even longer. The iteration of product R & D investment in the past 10 or 20 years far exceeds our current investment. A possible opportunity for us is the advantage of late development, because many trial and error in the past have been seen, and there is no need to repeat the same mistakes. But why our goal is not second-class, it is very simple. More and more sensor applications, especially sensor-related sensor applications, the performance / reliability of second-class products cannot be achieved.

Second, for now, I believe that the next major window for sensors is the Internet of Things. In IoT applications, China is without a doubt the place with the most application scenarios in the world. The Internet of Things requires a lot of customization, localized services, localized solutions, and even device design customization, which has become very critical. We believe that as a local company, in the competition with foreign companies, the customization of localized services and products will have more advantages.

Zhao Xiaoguang: The localization service spoken by Mr. Sun very much agrees that the future of chip integration with applications is getting deeper and deeper. Localization is very important. This is our advantage.

The development of China's sensor design and manufacturing enterprises is still very important. If we are under such control in the future, we must be sure that future development is dangerous.

Sun Zhen: The semiconductor industry is very global.

We can see from a few examples, such as the friction between Japan and South Korea some time ago, and finally the semiconductor industry as the winner. Japan ’s semiconductor chemical suppliers, very small The suspension of several chemicals has caused huge semiconductor companies Samsung and Hynix to have trouble in production! These companies are actually not large and are not among the semiconductor companies we generally pay attention to, but they have actually played a role.

Everyone knows that ASML is in the Netherlands. In the past, there were several lithography companies. The United States also has Japan, including Nikon and Canon. The lithography machine that competed to the highest end is basically ASML. Then it is a company in the Netherlands, in Europe.

In addition, for example, we talk about manufacturing capabilities. Now the world's largest manufacturing capacity is in Taiwan. TSMC accounts for 60% of Foundry's global volume. In the most advanced manufacturing process, 12nm, 7nm and even 5nm are basically the only ones capable of mass production. . So this is a very global, divisional and collaborative industry.

Not being controlled by others does not mean that we are self-closed and closed for development. It may be more important how to more closely cooperate with the semiconductor industry in our country and the semiconductors in other parts of the world to become an integral part.

Zhao Xiaoguang: President Sun can introduce the company's team, customers, and production capacity.

Sun Zhen: Si Rui Technology is a completely technology-oriented company, so the company team comes from all over the world. We have domestic and overseas returnees, we also have R & D centers in the United States, partners in Europe, and we have R & D centers in Taiwan. Into. Our market has developed from consumer to industrial to automotive electronics market.

We have shipped a considerable amount of motion sensors or magnetic sensors in consumer categories such as mobile wearables. Many major domestic motor manufacturers are users of our industrial magnetic sensors. Automotive sensors mainly include ABS and throttle switch sensors, which have entered customers such as Chery and Great Wall. We mainly adopt the model of design company. In automotive electronics, we have our own module production capabilities.

Zhao Xiaoguang: In addition to us, which sensors do better in China?

Sun Zhen: Compared with the products of other semiconductor companies in the global ranking of our domestic sensors, our ranking is relatively backward.

Today, I went to see the top 30 sensor companies in the world. There are only two in China. These two companies are not completely traditional sensor chip companies. They are more in the packaging and application of sensors. In the past, domestic sensors were mainly used in military research. There are many products at the beginning. The products are mainly targeted at the military industry or some industries, and now some are gradually moving towards unmanned driving, but basically they are still in a multi-variety, small-volume, high-unit-price market. Although in recent years, our domestic sensor companies have made great progress in some consumer markets, but I personally feel that we have a long way to go.

Zhao Xiaoguang: What do you think of Navy Technology? How competitive?

Sun Zhen: Naiwei Technology is a company that started with navigation systems. Its products were mainly used in the military industry in the past. After listing in 2015, Silex was acquired and entered the MEMS industry-related business. Sweden Silex is one of the top three MEMS foundries in the world. Our IMU is mainly produced in Silex.

Silex is a very typical Nordic traditional MEMS manufacturing company. Before 2016, Silex development was not very fast, and a large part of its revenue came from research and development revenue with customers. At that time, the production capacity was not large.
If you take a look at the Swedish factory, its factory is not very large, and its main competitiveness is its technical capabilities. MEMS sensors need to have many different structures. Different structures require many modules, chips, process technologies, The ability of process modules, they are relatively strong in this regard.

In fact, many sensor companies have come from Northern Europe, and there are such sensor companies in Finland, Norway, Sweden, and Switzerland. After the new Chinese investors entered, they made some changes to the development strategy of the entire company, making them gradually move from a more complete R & D orientation to a market orientation and to a model of a foundry.

In the past two or three years, the volume of the entire sensor market has also risen, so they have a pretty good development. Neville Technology also planned to build such a MEMS factory in Yizhuang. Now the factory has been completed. For Sweden, the problem of insufficient production capacity may also have a solution.

The expansion cost / production cost in Sweden is quite high. At present, the Chinese semiconductor industry is hot, especially the local government is very supportive of the factory, so this is actually a great opportunity for its development, especially in solving The top of the capacity provides a good solution.

Zhao Xiaoguang: Please ask President Sun to talk about a giant company like ON Semiconductor.

Sun Zhen: ON Semiconductor has traditionally been a company with discrete power chips. It ranks second in the world in power chips. After acquiring Fairchild, it is second only to Infineon. Power chips currently have a large layout in the automotive electronics market. With the rise of electric vehicles, this is a large market. In the automotive electronics market for power chips, ON Semiconductor has occupied a very good position.

Regarding the layout of automotive electronic sensors, ON Semiconductor is mainly realized through mergers and acquisitions. Like CIS, ON Semiconductor acquired Aptina, and on this basis, he acquired other related companies. The competition in this field mainly needs to provide a complete solution, just like a combination sensor. Fundamentally, it needs to provide a complete solution to integrate all different data. This is also a strategy of ON Semiconductor.

Domestic companies are limited to scale, R & D capabilities, market brands, etc., and basically can only do it individually. However, European and American companies all adopt another strategy, which is to provide customers with a complete set of solutions.
Regarding the technical route: "A lot of customization needs will make the sensor more valuable."

Zhao Xiaoguang: What do you think of the sensor technology line?

Sun Zhen: Talking about the technical route of sensors is actually a very broad question. It's not like Moore's law that semiconductors talk about. The technology of sensors is based on non-Moore's law. We call it Transcend Moore's Law. In fact, there are many different ways to implement the sensor. For example, the gyro we are talking about, some people use fiber optic gyro, but also use quartz. The semiconductor method is called MEMS. So I want to say that each product has its own development path for the sensor. I personally feel that there are different development paths in different application areas.

If we take the Internet of Things as the main direction, I personally think that there are several key directions in the next generation of sensors. The first is the form of combined sensors. Second, it needs to have certain performance and low power consumption. Of course, there are certain requirements for size and cost. Third, packaging technology packages devices with different functions together. The sensor just digs out a piece of data. If it can be processed and transmitted quickly at the port, such as edge computing, for example, ARM is used in our inertial navigation sensor to perform a localized calculation. In addition, For networking applications, sensor packaging technology will evolve from just traditional sensor packaging to more focused on packaging devices with different functions. As we are now trying to package sensors, MCUs, memories, Bluetooth and other communication chips together. Of course, once the scale is larger, different large chips are combined through packaging, which is also a farther format.

Zhao Xiaoguang: Are there major sensor design companies with similar routes? What are the main considerations when measuring sensor performance?

Sun Zhen: The route of major sensor design companies will be quite different. For example, there are more microphone products in China. Some of them are moving towards a low cost and others are moving towards high performance. Product line. Some are combined with different applications, such as BOE, which hopes to use the screen as a port to cut into markets like the Internet of Things. The carrier of this screen is not just a screen. It needs to put a lot of other devices in it to become a real carrier, which is applied to the Internet of Things.

Zhao Xiaoguang: Talk about the changes brought by 5G to sensor technology and industry.

Sun Zhen: In fact, 5G is very important for sensor technology and industry. Why do you say that? We are all talking about the Internet of Things as a very big outlet for sensors. The concept of the Internet of Things has been proposed in China for a long time. I remember starting to mention it in 2008 and 2009. At that time, the country established a sensor center in Wuxi, so We have mentioned it for at least ten years, and have done a lot of applications and made a lot of scenarios in it. But neither is very successful, and there is not a lot of volume, then it will completely change after 5G. On the surface, sensors are different scenarios, but the core of application development is big data. Port words on the cloud. Without 5G communication, its data granularity and the value of multiple data will be difficult to reflect. If the Internet of Things market is not as it is now, then for the sensors, the arrival of its tuyeres is greatly restricted.

Data is fundamental, sensors are big data "excavators"

Zhao Xiaoguang: See what else Sun has to add?

Sun Zhen: In conclusion, sensors are big data “diggers”. When it comes to the Internet of Things, many people think of 5G, not sensors, but sensors are important. The reason for this is that in Industry 4.0, data is fundamental.

The sensor is a big data "excavator", which is a very important device.

In terms of new trends in the application of the Internet of Things market, the Internet of Things is characterized by fragmentation, diversification, cross-domain, cross-platform and timeliness. At present, the “fragmentation” feature of IoT applications is weakening, and there are some single application products with market demand exceeding tens of millions of sets per year. Better integration and cost performance become necessary and possible.

At the same time, in the past, sensor modules commonly used in industrial and agricultural equipment and systems have increased interconnectivity and made them a type of terminal node in the Internet of Things.

The support of 5G standards for IoT / Internet of Vehicle applications overcomes the limitations of wireless interconnection standards such as Bluetooth BLE, Zigbee, LoRa / Sigfox, and provides a platform foundation for the popularization of IoT terminals.

In terms of the evolution trend of sensor products, the first stage is mainly in the industrial field, with low market usage, high unit price, low yield, and mainly performance pursuit.

The second stage is mainly in the consumer category. The market usage has increased dramatically, and MEMS sensors have entered a large-scale industrialization stage. The price has dropped significantly to meet basic functions.

The third stage is mainly in the category of industry, which puts more requirements on the functions and performance of the product, and the product line has more updated products based on performance upgrades.

The fourth stage is focused on the Internet of Things market. The rapid growth of the Internet of Things application market has made the functions and uses of sensors, taking data collection and effective use as consideration points, and it is becoming more and more common for sensor nodes to have connectivity functions.

"The key to the development of China's semiconductor industry is that it cannot be closed and closed."

Zhao Xiaoguang: President Sun is also a veteran of the semiconductor industry. What are your opinions and suggestions on the development of the Chinese semiconductor industry?

Sun Zhen: For the development of China's semiconductors, especially in such a large environment, I personally feel that we must adhere to the idea that it should not be self-closing, and should not be closed behind. The chain is very long. From the raw materials just mentioned, such as many chemicals, to the end market, this chain is very long.

In terms of materials, Japan and Europe may have developed earlier; in terms of equipment, Europe, the United States, Japan, and South Korea are relatively advanced; in terms of manufacturing, the entire chip manufacturing and packaging capabilities, production capacity, and technology are now the best in Taiwan; in the application market, There is no doubt that China is in a very good position. How to combine our advantages with the advantages of development in various places and links, and come out of our own development path of Chinese semiconductors, I think this is the key. An important point in this process is that we cannot close ourselves and develop ourselves.

Zhao Xiaoguang: The communication with President Sun is very sufficient today, which gives us a deep understanding of the sensor industry.

On the one hand, from the perspective of demand, sensors will be integrated together to provide users with better data sensing and application services on mobile phones and other consumer electronics.

On the other hand, we can also see that the very important part of the Internet of Things is the sensor, and the very important information in the sensor is the movement and location information. This is very important for the development of the entire Chinese Internet of Things and the entire industry in the future. step.

From the technical route and the development of the entire industry, Chinese companies and China still have a lot of room for development from design to manufacturing on the Internet of Things. Silicon Core is also one of the domestic leading companies in this regard. I hope that Silicon Core will The field of sensors is promising.

Sun Zhen: Thank you so much for your time. I also have a different perspective on our industry.

@ 雪球-虎妈牛小妞