Rebirth to 90s: Start from Scratch

Xiang Ding said: "The lithography machine should be the most difficult, but there are also vapor phase epitaxial furnaces, oxidation furnaces, wafer dicing machines, magnetron sputtering stations, chemical mechanical polishing machines, ion implanters, and plasma etching technology. , Wire bond stitching machine, reactive ion etching system, low pressure chemical vapor deposition system, plasma enhanced chemical vapor deposition system, wafer thinning machine, probe test bench, etc."

"Among them, the lithography machine was once close to the world's leading level. As long as we persist for about five years, we may become the world's top level and compete with the United States and Japan."

"At that time, we started a little later than the U.S., and we developed it at about the same time as the country of Japan, and it was about ten years earlier than the country of Bangzi and Baodao."

Zhang Qianxiu was surprised and asked, "Have we made chips?"

Tian Shuyuan sighed and said: "Did it! Our school did it. I was still one of the members and was responsible for the research of a small part of the tasks."

"In 1965, the Academy of Sciences developed a 65-type contact lithography machine. At that time, it was still tens of microns and it was not difficult."

"In 1980, our university developed the fourth-generation distributed projection lithography machine with an accuracy of 3 microns."

"At that time, the mainstream technology in the world was 3 micron and 1.5 micron, but the chip was not only a bottleneck of the process, but also other technologies."

"In 1979, Intel released the first personal computer CPU, the model number is 8088. At that time, our technology was very close to it, but we did not have a mature and complete production system."

"We have no way to produce wafers, and there are many types of chips, not just one type. Some are for personal use, for servers, for airplanes, for machines, and so on."

"Used in different fields, the internal structure is different, and there must be corresponding software to support the development."

"We had more than 40 integrated circuit factories in the 1970s, and more than 30 production lines in the 1980s. Our integrated circuit factories were fighting for their own, and we were doing our own. Although we all made integrated circuits, The directions are different."

"Wafer production line, because we cannot develop and manufacture independently, we started to introduce the world's advanced 3-inch wafer production line as early as 1973."

"At this time, we are two years ahead of Baodao and four years ahead of Bangzi Country."

"But for various reasons, we originally planned to introduce seven production lines, but we finally delayed the introduction in 1980, and there were only three."

"Seven years! Mr. Zhang, do you know Moore's Law?"

Zhang Qianxiu nodded: "I know, that is to say, the number of transistors on integrated circuits will double every 18 months."

Tian Shuyuan smiled bitterly and nodded: "Yes, in fact, we have never believed in this law, and it shouldn't be said to be a law, it's probably a law, and it will definitely be invalid after a certain time."

"But he said this sentence, so far, it is probably correct. In seven years, it has doubled by 4.6 times. The advanced wafer process we originally introduced has also fallen behind."

"From 1980 to this year, a total of ten years, the country of Japan has developed rapidly. In terms of semiconductors and integrated circuits, it has surpassed the country of the United States. In fact, it has already surpassed."

"The U.S. is developing at its best. This has led to the fact that these two countries have grown too quickly in the semiconductor industry, and we do not have a mature industry, and the input-output ratio is seriously inconsistent."

"In addition, in the field of integrated circuits, the technology is updated very quickly, and you must always invest in research and development, otherwise the products you produce will fall behind."

"Our products in this area have not formed strong competition, so we can't sell any profit."

"In the 1970s and 1980s, the Japanese government concentrated all the large companies in the semiconductor field, concentrated research and development, shared results, those basic areas that did not make money, the government paid for research and development, and the fields that were practical to specific products were developed by the following companies. ."

"By the mid-1980s, 50% of the world's semiconductor products were produced in Japan."

"The most powerful in the U.S. is Fairchild. They wanted to find loopholes in Waguo products. They were surprised to find that Waguo's products have completely surpassed them, and the production cost is far lower than them."

"As for the various disputes that followed are related to finance, we won't talk about it."

Zhang Qianxiu asked: "Professor Tian, ​​what preparations do we need to make if we continue to catch up with them and achieve full localization?"

Tian Shuyuan glanced at Zhang Qianxiu and said, "In the 1970s and 1980s, we were able to follow their footsteps, but by now, we have been left behind by him."

"It takes five to ten years to catch up with their technology."

"Even in the 1970s and 1980s, we were only able to follow their footsteps in terms of production technology, and there were still no breakthroughs in many technologies."

"But to be fully localized, to be honest, I dare not say a time."

"There are too many fields involved. I will probably tell you about it, including programming, design, optics, mechanics, polymer physics, chemistry, materials science, fluid mechanics, microelectronics, electromagnetics, mathematics, automation, etc. ."

"R&D is not actually the most difficult. Just invest money. The most difficult thing is to realize it and scale it up after R&D."

"Because our early research and development must first catch up and then surpass. In the process of catching up, all of our technologies are lagging behind. We have no advantage in front of others."

"In the process of becoming, we can also solve the problems encountered in actual mass production and accumulate more technology."

Speaking of this, Tian Shuyuan sighed again: “We have lost at least ten years of time. From 1980 to 1990, our country’s investment in R&D became smaller and smaller. We just didn’t see opportunities for profit. With this money, It's better to buy a production line that others have eliminated and come back and produce it yourself."

"But these eliminated production lines are not profitable, at most they are fooling people in the country."

Zhang Qianxiu probably also understood the difficulties of the government. From the 1950s, he started to invest money in semiconductor research until 1980. For thirty years, a lot of research and development funds were invested. Although there were many technological breakthroughs, they did not bring much production. If part of the money is used to buy production lines in other countries, the rest can support the development of other scientific and technological fields, which is more cost-effective than in this field.

All in all, if you develop your own technology, it is better to buy a mature production line, and it is better to buy a ready-made product than to buy a mature production line.

If R&D technology costs one hundred yuan, the production line will only cost fifty yuan, and the finished product will only cost twenty or even ten yuan. In particular, the technological iteration in this area is very fast.

In this era, the integrated circuit market does not have a large scale, and belongs to high-end, small-scale products. Automation and informatization have not yet been popularized, and integrated circuits have not yet exerted their true power and have not yet been widely used.

In the future, almost most electronic products will be applied, but not now.

Especially in the business of personal mobile phones, mobile phones are still big brothers now, and no one in the world can expect much change in the next 30 years.

Zhang Qianxiu felt the difficulty of this road, and at the same time, he understood the reasons for not vigorously developing semiconductors in this era. Thirty years later, he will blame the decision in the 1980s and 1990s. That is not to think about the situation. It is not that the country is not developing. After 30 years of continuous investment, I gave up without seeing the expected results. With this money, other fields have been developed early.

After this discussion, everyone present had a heavy expression, because all the three people present except Zhang Qianxiu were real experts in this field. They knew the situation of domestic semiconductors and could be described as "horrible".

Zhang Qianxiu turned to ask: "Professor Tian, ​​to tell you, I want to make a chip."

The professors present were all surprised to see Zhang Qianxiu, a field that the nation has not developed for decades. Zhang Qianxiu is a teenager who wants to do it. This is simply a fantasy.

Tian Shuyuan smiled bitterly: "Mr. Zhang, I have talked with you for so long. Although you don't talk much, you mostly listen to us, but I believe that you are a patriotic and ambitious person."

"But let me say something that sounds bad, this field can not be developed by a single person, a single company."

"In order to develop semiconductors, the Japanese government officially unites top companies in the country to establish VLSI, which is a super-large-scale joint research institute."

"The U.S. also formed an alliance after being surpassed by the Japanese nation in the early to mid-1980s."

"Bangzi Country is also developing at a particularly fast pace, and they are also developing this field with the strength of the whole country."

"If we want to have a place in this field, it is almost necessary to contend with the research and development capabilities of the three countries. The road is long and difficult!"

Zhang Qianxiu said: "I have to try. I am not so arrogant. I have prepared myself. Even if I fail, I can guarantee that I will have nothing to worry about for the rest of my life. As long as I have a penny, I will Keep doing it."

"The main purpose of this visit is to understand the current situation with the professors. To be honest, I always thought that our National Semiconductor, especially the chip field, was a zero."

"But I heard from professors that our country still has a certain foundation, at least I don't have to start from scratch."

"Now the most advanced process is 1 micron. We have a 3 micron process. If we work hard, there is still hope in the end."

Tian Shuyuan shook his head and said: "The chip needs an architecture and instruction set. Without this, even if we have the manufacturing process, it is useless to be able to make the chip, but this is another problem."

Zhang Qianxiu asked suspiciously: "What does this mean?"

Tian Shuyuan said: "Let's put it this way! A chip is an integrated circuit with many circuits, but in the final analysis, it is just an integrated circuit. It does not understand what you ask him to do, and it also needs a supporting instruction set."

"This instruction set is to connect the interface between human and machine so that the chip can understand the instructions you give. This is the simplest instruction."

"On the basis of this instruction set, everyone has developed software. The computer system we use is a piece of software."

"When you type the letter A, you give an instruction, and the system is converted into the language of the chip architecture. After the chip receives the command, it will operate on his complex circuit, and finally it will be displayed on your display."

"The more advanced the instruction set, the better the performance of the chip."

"The chip, instruction set, and system are a whole, and none of them are indispensable."

Zhang Qianxiu was stunned and asked, "What instruction set is there now?"

Tian Shuyuan said: "X86, MC68000, these are all, you may not know these."

"X86? On the computer?" Zhang Qianxiu asked.

"Yes!" Tian Shuyuan was a little surprised: "Unexpectedly, Mr. Zhang knew this."

Nowadays, computers are not popular, and they are not professionals and do not know this at all.

"Is ARM also an instruction set?" Zhang Qianxiu asked.

Tian Shuyuan froze for a moment, and said, "Yes, but it seems that there is no application on the market."

Zhang Qianxiu probably understands that in the chip wars, the only domestic manufacturer capable of designing chips used the ARM architecture. Almost all mobile phones use this architecture.

What Zhang Qianxiu didn't know was that there were two architectures in the world that monopolized the computer market and the mobile phone market, namely X86 and ARM.

For the X86 architecture, Intel only licensed it to AMD, and the win system was developed based on the X86 architecture.

This is why the company that can produce computer chips has its own Intel and AMD, and other companies cannot produce it even if they have this technology, because Intel does not authorize other companies, and other companies cannot use the X86 architecture. If they cannot use the X86 architecture, they cannot install WIN. System, win system cannot be installed, who buys your chip?

Later, all the instructions used in the architecture of the mobile phone were based on the ARM architecture. Qualcomm, Bangzi Country’s four-star, a certain company in my country, and MediaTek all used the ARM architecture. We often say that Kirin chips, Snapdragon and other chips are all Can't escape the ARM architecture, this company can make a lot of money by relying on instruction set authorization alone for one year.

Zhang Qianxiu said: "Then we still have a chance, we haven't developed it yet!"

Xiang Ding said: "Mr. Zhang, if you do semiconductors or integrated circuits, it's okay. There are other ways. If you want to make computer chips, you can't escape the X86 architecture."

"Intel will not authorize other companies. Without the X86 architecture, the WIN system cannot be installed."

Zhang Qianxiu said, "How feasible is it to develop a new instruction set and supporting system by ourselves?"

"Not less difficult than the CPU." Tian Shuyuan said.

Well, now I have another problem for myself. To make a chip, you must have an instruction set, and there must be a system on the instruction set.

If it is to be fully localized, it is not enough to solve a series of hardware problems, but also to solve the software problems.

Even if Zhang Qianxiu is a rebirth, he is not confident that he can develop a system to defeat Microsoft; then he developed an architecture to defeat the current Intel, and ARM that will dominate the mobile phone architecture in the future; and finally develop a chip manufacturing industry, giving all of the Bangzi, Japanese, and Chinese Dry up.

"It's up to people." Zhang Qianxiu said: "I can't let this life be a waste of life, I always have to do something."

Zhang Qianxiu said this to himself. After all, in the first rebirth, it is too simple to mix up food and clothing. It is too simple to do something that is good for the country and the people.

It’s not that Zhang Qianxiu is great, but that as a Chinese, he wants to make a little contribution to the land where his ancestors, descendants and grandchildren live on the premise of ensuring his living conditions. He himself doesn’t think this is much. great.

To put it simply, I have no worries about food and drink, and I have enough money to spend. It doesn't matter if I toss about it.