Chinese semiconductor thread II

[tokenanalyst]

Xingji Micro-Equipment's power device direct writing lithography equipment exported to Japan​

Xingji Micro-Equipment, a leading domestic manufacturer of direct write lithography equipment, exported its MLF series equipment to Japan for the first time.
The MLF series direct-write lithography equipment is designed for high-precision, high-efficiency pan-semiconductor packaging applications, especially for power semiconductors, such as the packaging process of insulated gate bipolar transistors (IGBTs). The equipment is equipped with an advanced equipment front-end module (EFEM), which can support the fully automatic operation process of 12-inch wafers, significantly improving production efficiency and process consistency. Core Micro will provide technical support and services to Japanese customers throughout the process to ensure efficient and stable operation of the equipment.
This export marks a solid step forward for Xingji Micro-Equipment in its globalization strategy and further consolidates its leading position in the field of direct writing lithography equipment.





The MLF series direct-write lithography machine adopts advanced digital lithography technology. It does not require a mask and can directly transfer layout information to a substrate coated with photosensitive materials. It is suitable for third-generation semiconductor silicon carbide process applications, power semiconductors ( IGBT ), ceramic substrates and other application fields. The MLF series products have a compact structure, a large depth of field, a fast speed, and good process adaptability for both dry film and photoresist. It is an economical and flexible mass production equipment.

Please, Log in or Register to view URLs content!

​

 

[ansy1968]

A good read especially coming from a Westerner POV, I hope he is referring to the 5 years gap between TSMC latest 3nm chips NOT the Intel 10nm.... Lol

Please, Log in or Register to view URLs content!

 

[antiterror13]

A good read especially coming from a Westerner POV, I hope he is referring to the 5 years gap between TSMC latest 3nm chips NOT the Intel 10nm.... Lol

Please, Log in or Register to view URLs content!

It does say 5 years behind global leader ... I think is about right. TSMC entered volume production of "5 nm" chips in 2020 and China will be in 2025. TSMC entered volume production 7nm in April 2018.

 

[gelgoog]

5 years is basically two generations.

I think the article is a hodgepodge of quotes and lacks structure and proper analysis. They smatter a bunch of charts in the article with bogus metrics, while more accurate metrics like manufacturing capacity, volume of chips sold are either hidden inside the text and only mentioned in a passing way, or not mentioned at all.

 

[natlou]

praying on Moore’s law being broken

 

[OptimusLion]

The next breakthrough direction of China's chips: making 7nm chips more powerful than 3nm chips

As we all know, the United States is currently carrying out the most severe crackdown on China's chip industry.

According to the US goal, logic chips are locked at 14nm process, DRAM memory is at 18nm, and NAND flash memory is at 128 layers. Any equipment that is more advanced than these technologies is restricted from entering the Chinese market.

Of course, before the US restrictions, some domestic manufacturers also bought some advanced equipment, but the output was not very high, so they also partially broke through the restrictions. For example, logic chips are definitely no longer 14nm.

However, judging from the current situation, the next direction of China's chip efforts may still be to make 7nm chips more powerful than 3nn chips.

Don't take this as a joke, it's the truth.

Previously, Huawei director Zhang Pingan said that if we cannot get 5nm or 3nm, it would be great if we can get 7nm. Our innovation direction cannot be on single-point chip technology, but on system architecture.

In fact, his words were very clear. At present, since we do not have EUV lithography machines, it is impossible to develop domestic EUV lithography machines in a short period of time, so it is basically impossible to enter 5nm and below.

What we can produce is actually 7nm chips, which is the best result at present.

But compared with 5nm, 3nm, and 2nm, the technology of 7nm chips is definitely behind. So we should not think about the volume technology, but innovate in the system architecture to make the 7nm chip comparable to 5nm, 3nm and others.

What is innovation in system architecture? First, adopt better transistor structures, such as 3D and stereoscopic ones, and second, adopt better packaging technologies, such as cascading, stacking, and stereoscopic packaging. Or adopt chiplet technology to combine many chips together. New materials can also be used, such as carbon-based chips and optoelectronic chips.

In fact, the chip industry itself is also exploring these technologies. Because the current chip technology has reached its physical limits and it is difficult to continue to shrink it, the next step is to use new materials, new architectures, etc. to achieve performance improvements without changing the chip.

Huawei’s previous Kirin 9000S and Kirin 9010 also proved that this is feasible. Although these two chips are not made with 5nm process, their performance is not inferior to the 5nm Qualcomm and MediaTek chips.

Therefore, the goal of domestic chips in the future is not to force a breakthrough of 5nm, 3nm, and 2nm, but to build on the already achieved 7nm and further explore it to make it comparable to 5nm and 3nm.

As for the advancement of technological miniaturization, we will not give up, but we will not fight to the death. Instead, we will move in multiple directions, combine with each other, and move forward together. What do you think?

 

[antiterror13]

The next breakthrough direction of China's chips: making 7nm chips more powerful than 3nm chips

As we all know, the United States is currently carrying out the most severe crackdown on China's chip industry.

According to the US goal, logic chips are locked at 14nm process, DRAM memory is at 18nm, and NAND flash memory is at 128 layers. Any equipment that is more advanced than these technologies is restricted from entering the Chinese market.

Of course, before the US restrictions, some domestic manufacturers also bought some advanced equipment, but the output was not very high, so they also partially broke through the restrictions. For example, logic chips are definitely no longer 14nm.

However, judging from the current situation, the next direction of China's chip efforts may still be to make 7nm chips more powerful than 3nn chips.

Don't take this as a joke, it's the truth.

Previously, Huawei director Zhang Pingan said that if we cannot get 5nm or 3nm, it would be great if we can get 7nm. Our innovation direction cannot be on single-point chip technology, but on system architecture.

In fact, his words were very clear. At present, since we do not have EUV lithography machines, it is impossible to develop domestic EUV lithography machines in a short period of time, so it is basically impossible to enter 5nm and below.

What we can produce is actually 7nm chips, which is the best result at present.

But compared with 5nm, 3nm, and 2nm, the technology of 7nm chips is definitely behind. So we should not think about the volume technology, but innovate in the system architecture to make the 7nm chip comparable to 5nm, 3nm and others.

What is innovation in system architecture? First, adopt better transistor structures, such as 3D and stereoscopic ones, and second, adopt better packaging technologies, such as cascading, stacking, and stereoscopic packaging. Or adopt chiplet technology to combine many chips together. New materials can also be used, such as carbon-based chips and optoelectronic chips.

In fact, the chip industry itself is also exploring these technologies. Because the current chip technology has reached its physical limits and it is difficult to continue to shrink it, the next step is to use new materials, new architectures, etc. to achieve performance improvements without changing the chip.

Huawei’s previous Kirin 9000S and Kirin 9010 also proved that this is feasible. Although these two chips are not made with 5nm process, their performance is not inferior to the 5nm Qualcomm and MediaTek chips.

Therefore, the goal of domestic chips in the future is not to force a breakthrough of 5nm, 3nm, and 2nm, but to build on the already achieved 7nm and further explore it to make it comparable to 5nm and 3nm.

As for the advancement of technological miniaturization, we will not give up, but we will not fight to the death. Instead, we will move in multiple directions, combine with each other, and move forward together. What do you think?

5nm can be manufactured with DUV, perhaps not that cost efficient with EUV, bit it can ..... I can see 5nm will be mass produced in China next year like 7nm currently

 

[tphuang]

Please, Log in or Register to view URLs content!

Huawei's 4-core MCU here is used in Astrolink S's smart watch. The interesting part is that the cores are RISC-V based. We've heard in the past about Huawei's efforts into RISC-V, but this is the first time in a while that I saw it using RISC-V cores in something.

And this Astrolink S watch also use Harmony OS

 

[gelgoog]

Huawei controls the whole stack. So they have no reason to continue paying royalties to ARM I think.

 

[Hyper]

5nm can be manufactured with DUV, perhaps not that cost efficient with EUV, bit it can ..... I can see 5nm will be mass produced in China next year like 7nm currently

I read on comments on semiwiki that CFET is easier for lithography than nanosheet transistor but requires BPR and BSPDN.