China domestic sensors: from the supporting role to the manufacturing center

China Domestic sensors: from the supporting role to the manufacturing center

To open up the situation where sensors are weak and thin, and get stuck everywhere, Chinese sensors need a system layout in the industry chain.

1. The king of supporting roles has appeared

  In China's manufacturing strategy, the major players such as mainframe equipment play the leading role, while the sensors at the forefront of perception are insignificant supporting roles. This is not surprising. As a small component, sensors have always lacked a sense of existence; and because the scale of leading companies is also very small, if a company can achieve 100 million yuan in pure sensor revenue, it is already very powerful. Role. This kind of scale naturally lacks the right to speak. The positioning of the sensor in the manufacturing industry has become a "matching in a complete set, an edge in the edge".

  Sensors are often at the forefront of all industrial products. They provide the first checkpoint to perceive the physical world. Checkpoints never make mistakes, and factories operate well. This is the basic rule on which industrial operations depend. Until one day, a kind of pain spreads, and when you look closely, the checkpoint may be unreliable, but its value will be instantly magnified. For many years, FirstSensor, a German sensor company, has been providing pressure sensors for China's petrochemical instrument industry. However, in 2019, this German company was suddenly acquired by the American connector giant TE, so the entire petrochemical instrument industry naturally entered a shadow of fear and lingered. In recent years, the chip, as the most dazzling star of "Kneck", has acted as the super ambassador of Kneck. But in fact, many products, such as sensors, industrial software, and scientific instruments, all have the same function of sealing the throat, but they affect people's feelings in different ways. If the 7nm chip is gone, you will not be able to enjoy Huawei's P50 mobile phone at best; but if the pressure sensor is gone, a chemical plant with an output value of hundreds of billions will directly stop production.

  In fact, as the starting point of intelligence, sensors are the cornerstone of building smart industries and smart social buildings. Just like a high-rise building of several hundred meters, it cannot be separated from the foundation of several tens of meters deep underground. People are always attracted by the magnificent facades of skyscrapers, but few people care about the foundations of the ground. However, to build a smart manufacturing building, it is precisely by relying on a large number of incognito sensors to lay the foundation. If we go to dissect the foundation of manufacturing, the king of supporting roles is none other than sensors.

  However, it is indeed very difficult to recognize the true face of the sensor. There are so many types of sensors, I am afraid that it is difficult to say clearly, but more than 30,000 types, it should be a very certain number. To fully understand the belongings of the sensor, almost all manufacturing sectors have to be crossed, and it is as difficult as identifying a sky full of stars. To understand such an industry, it is inevitable to be daunting. This has also made that for many years, the sensor industry has rarely received direct and strong support from national policies: the king of supporting roles sits on the cold bench for a long time, and there is a reason.

2. In the three major areas, industry is the most urgent

  From the practical application point of view, the situation of sensors in military, civil, and industrial fields is also very different. The industrial sensor is an urgent need to break the game.

  As far as industrial-grade sensors are concerned, the gap between domestic products and foreign countries is the largest. On the one hand, it is limited by the level of domestic production and manufacturing, and the consistency and stability of the products are not enough; on the other hand, it is related to the domestic indifference to sensors. When it comes to technical research, the host is often taken the lead, and the sensor as a supporting role is basically doctrine. When importing foreign equipment, the whole machine is often used, and there are few opportunities for domestically-made parts. Industries such as nuclear energy, chemical industry, and electric power are generally directly input to complete sets of equipment, and these equipment basically come with their own meters and sensors. In many projects in the bidding, the construction party also clearly requires that high-end instruments and sensors must use foreign brand products.

  Since sensors account for a relatively low proportion of the overall equipment value, manufacturing companies would rather use higher-quality foreign brands. For example, for a machine tool of 300,000 yuan, the price of an acceleration sensor is only a few thousand yuan. Machine tool factories are unwilling to take risks to adopt domestic sensor brands, which makes domestic sensor manufacturers hungry for a long time, lacking industry applications and opportunities for trial and error growth.

  As far as the military is concerned, sensors are the facial features of missiles, rockets, and various munitions, which are the benchmark competitiveness of all weapons. In the latest U.S. Department of Defense budget for the joint fighter F35, research on sensors and network warfare is still included. And the US Army Chief of Staff Milley (Milley) has always regarded sensors as the key core of future intelligent warfare. In an article in the United States at the beginning of this year, "Hunting the Opponent: The Battle of Sensors for 2035", he cited a "high-end war" statement put forward by Milli. In this concept, "the role of the ubiquitous sensor is to make the opponent die faster." In the war for 2035, the US military envisioned seven kinds of intelligence around sensors, including visual intelligence, communication intelligence, etc., and artificial intelligence is just one of them. From this point of view, talking about artificial intelligence can easily create a false illusion for manufacturing; only talking about intelligent manufacturing around sensors can it be easier to approach the truth of intelligence.

  The military industry has always attached great importance to the application of sensors, and it is even costly and expensive in design and manufacturing. As a military sensor, the pursuit of availability is the first goal, and the economy is the second. It often uses more advanced production lines and process methods, and it is inevitable to waste labor and materials. The manufacture of military sensors is in a completely different state from industrial sensors.

  Civilian sensors are quite different from military products. It must first consider economy, and strictly control production costs and defective product rates. At present, in the field of middle and high-end sensor manufacturing, my country has no advantages in intellectual property rights, equipment and materials, and the cost remains high. But in recent years, in some limited areas, this situation is gradually changing. For example, with the widespread application of high-fidelity headphones TWS driven by Apple, the development of acoustic sensors has also risen, and a number of star companies have emerged, such as GoerTek and AAC.

  In fact, the rapid development of civilian sensors has given industrial-grade sensors an important inspiration, that is, applications will drive the development of sensors, and a large number of applications will drive the development of sensors. In the past, acoustic sensors were monopolized under the American Lou's brand, but with the shipment of a large number of headsets from Weifang Goertek, it naturally added the most critical MEMS sensor (this kind of sensor uses a kind of sensor that is completely in line with traditional mechanical processing). Different semiconductor processes) initiate challenges. In spite of the active launch of a large number of patent litigation battles, Lou's is still retreating, and China's acoustic sensors are rising like a corps. Similarly, in emerging fields, China's sensors can often find confidence due to the huge domestic market. With the rise of life medicine and general health, biosensors for measuring blood pressure and blood sugar are also booming. This is also an opportunity for Chinese sensors.

  This is a lot of inspiration for industrial sensors that have been hard to return for many years. Of course, sensors in the three different fields have their own similarities. Foreign sensor manufacturers often take all three, because only in this way can they achieve the greatest scale effect. And this happens to be the shortcoming of Chinese sensors.

3. How to do it so small?

  Domestic sensor manufacturers have no scale. There are about 40 domestic listed companies that are involved in sensors, but sensors cannot become their main business. Zhengzhou Hanwei, which has made a sound and colorful sensor in the country, can be regarded as a bloody road, and its revenue in 2020 will be 2 billion. Less than 20% of these are pure sensors, and the others are IoT solutions formed around sensors. Moreover, such a high proportion has a lot to do with the sharp increase in demand for temperature sensors caused by the new crown epidemic. If looking at 2019, the revenue of sensors accounted for less than 10%.

  Hanwei is the lucky one of the sensor. Looking at the past, there are less than two thousand domestic industrial sensor manufacturers with a certain scale, and many of them are struggling with a revenue line of 30 million. Many sensor products are in small batch production, and the development cost of a single product remains high. Products made in China often give people a deep impression that they are cheap and have advantages over foreign countries. But in fact, on the contrary, the cost of many industrial sensors in China is often not competitive.

  This is related to the more complex and precise production process of the sensor. Taking the photoelectric sensor as an example, it is necessary to detect the weak signal to extract the signal, shaping and amplify, noise processing, etc., and the process flow for different sensors is the real test of the company's know-how. Sensor companies need to buy components such as chips, materials, and single-chip microcomputers, and then make their own optics, structures, light collection, light-emitting processing, and display. Many of the equipment, tooling, fixtures, calibration instruments, etc. required for this process need to be developed by the enterprise itself. As an intelligent starting product, the sensor is actually a nearly "self-sufficient" production method. This has left a frustrating impression that the sensor has hardly entered an "industrialized state."

  Sensor manufacturing also has software algorithm problems, which supports the intelligent requirements of sensors. The photoelectric switch looks simple, it can send and receive light, but in fact, a small photoelectric switch usually has more than 100 components, which have different requirements for different application scenarios, and it is not easy to make it well. Tianjin Yike has been engaged in sensor manufacturing since 2003, and has been looking for accumulation of various experiences in photoelectric sensor algorithms for many years. For example, in the inspection of biscuits on a food packaging line, there will be different biscuits on a packaging line, with different packaging quantities and packaging methods.The basic function of the photoelectric switch is counting, for example, counting how many biscuits there are. This sounds simple, but the actual situation is much more complicated. Some biscuits are mixed with sesame, chocolate beans, etc., which will cause interference; some biscuits are very thin and have weak signals; in mixed-line production, biscuits of different colors are mixed together. When numerous biscuits are connected together on the packaging line to move forward, it is not easy to count them one by one. In order to come up with an algorithm that can accurately count all kinds of biscuits, Ike’s R&D staff went to the supermarket to buy biscuits, bought hundreds of packages of various biscuits, and tested them one by one. Chocolate biscuits are black and white, and the algorithm should reduce the black-to-white ratio; when the biscuits with perforations cross the line, the light spot is not easy to be punched out, and the algorithm should be used to enlarge the light spot.

  When the remaining biscuits can finally be eaten, it means that the circuit board hardware, materials, and optics have passed the technical conditions. The next thing to consider is the stability of small batch production, which is also the most difficult part of sensor manufacturing. If there are large orders, the production line is likely to become more stable, and manufacturers are more willing to invest in the reliability of the quality control system and the stability of the supply chain. However, the batches of domestic sensor production are usually only a few thousand, and there is little demand for mass production. However, for small batch production, almost the same investment is needed to achieve incoming materials, process, quality inspection, etc., in this way, the production consistency is naturally difficult to control.

  For this kind of multi-variety, small-batch products, semi-manual and semi-automated products are inevitable, and it is difficult to find a suitable intelligent manufacturing production method at once. This is really a huge irony. When a sensor company's products are widely integrated into the intelligent manufacturing production line, its own production still uses a very backward method.

4.The whole chain moves, pulling and singing are wonderful

  In the 1960s and 1970s, the Japanese manufacturing industry, whether it was color TVs, automobiles, or semiconductors, was a model of "fighting on the battlefield and building its own arsenal" while developing industrial products while developing upstream equipment. This is a very effective way of nationwide deployment. At present, Japan's semiconductor manufacturing equipment is in a leading position in the world, not inferior to the United States, and it is driven by the rise of semiconductors in the past. Many semiconductor equipment in the United States, in the end, are strong first and then weak, and are squeezed out by the growing Japanese manufacturers. Although the current lithography machine ASML in the Netherlands is a great success, it has already overwhelmed Nikon and Canon. However, before the rise of ASML, Nikon and Canon's lithography machines relied on the synchronized polishing technology of their domestic memory to completely clear out the early lithography machine hegemons such as the United States.

  Similarly, this phenomenon also occurs in China's emerging industries. The rise of photovoltaics and lithium batteries has brought a golden opportunity to the development of China's equipment industry. They entered the track without any burden, or on the other hand, manufacturers of photovoltaics and lithium batteries, without any burden, chose upstream domestic equipment manufacturers, including laser equipment. There are at least four or five listed laser equipment companies in China, which are specifically designed for photovoltaic and power batteries. China's laser equipment industry has grown from an output value of 10 billion in 2010 to a current output value of 70 billion, becoming the most important market in the world. This is inseparable from the pull of China's advanced manufacturing industry. Advanced equipment and emerging industries simultaneously play and sing.

5.Small note: the spring of sensors

  China's sensor industry is estimated at a scale of 130 billion yuan, which is similar to the scale of artificial intelligence. But it can really play the role of leverage, and actually leverage the output value of more than 30 trillion made in China. In contrast, in terms of popularity and gold absorption, the sensor is too far behind. But the sensor has never been a raging fire. As far as the output value is concerned, its brightness is insignificant, which blurs people's attention to it. However, for Chinese manufacturing to move up a higher level, what is most needed is to re-recognize such a king of multiple, scattered, small, and weak supporting roles.

  Why do you want to light up the sky full of stars? The answer to this question is the response that the sensor industry is most willing to hear.