what is an integrated circuit?

source: /~lebok/gb/learn.htm

what is an integrated circuit? It's the small black blocks on the circuit board, which have many pins. It's not complicated inside, but just some triodes combined together, that's all.

top

------------. There is an electrode that controls the on-off of this channel. If the basic principle of the transistor is far-fetched by such a metaphor, it is very accurate when designing the layout of the transistor. We first draw a green line to represent the channel, and then draw a red line across the channel to represent the control grid, just like the green channel on the road and the red light mastered by the police. The on-off of the current in the green channel depends on the police's face. However, in an integrated circuit, a channel is not called a channel, but an active region, a strange name, but it is easy to remember. We usually call semiconductor devices active devices, resistors and capacitors passive devices, and triodes are active devices, so just remember that the region related to triodes is called active regions.

The source and the drain are composed of N-type or P-type semiconductor materials, and a layer of polysilicon is placed between the source and the drain as the gate, which forms a MOS transistor. Make more such transistors and connect them as required, which forms an integrated circuit. Making many triodes together is an integrated circuit.

Integrated circuits are really that simple. Please don't ask why semiconductors conduct electricity for the time being. What we are discussing here is how to learn how to design integrated circuits quickly, not semiconductor theory.

top

------------. In general, when the voltage between the gate and the source is V, it means off, and when the voltage above .7V is on the gate, it means on. It should be noted that the gate voltage is for the source.

We call the above MOS transistor N-type MOS transistor, and correspondingly, there is a P-type MOS transistor, which has exactly the opposite characteristics. When the current reaches the source from the drain, the MOS transistor turns on when the voltage on the gate is lower than .7V than the drain.

If it is stipulated that only one type of MOS transistor can be used, we can also design integrated circuits. We think that the original semiconductor technology was only suitable for making N-type MOS transistors. At that time, most of the integrated circuits were NMOS integrated circuits, and the early Z8 and 848, which we are familiar with, were all made by NMOS technology. Later, the process of making two different types of MOS transistors on the same chip was developed, which is called CMOS process, and now it has become the mainstream process in the semiconductor industry.

There is nothing different between the layout design of N-tube and P-tube. Just mark their types. This mark is used to inform people who make integrated circuits to make these tubes into a certain type of tubes. In the figure below, we enclose P-tube with a dashed frame as a mark.

top

------------. You can't always use electric soldering iron and solder wire, can you? In integrated circuits, this method is not used, but a method similar to double-sided circuit board is used. The vias on the double-sided circuit board connect the two sides of the circuit board, and vias are also used in integrated circuits. The two layers of conductive materials are aluminum and polysilicon, respectively. Aluminum can cross various areas and go anywhere without restrictions, but can polysilicon? It seems possible, but when polysilicon crosses the active region, the active region becomes a current path controlled by polysilicon: an extra transistor, which is not what we want, so we add a rule here: polysilicon cannot cross the active region. Connecting two transistors according to this rule, we design a simple integrated circuit with an inverter.

There are four solder joints around the chip, which are used to connect with external circuits. However, they are not called solder joints here, but are called pressure solder joints, which may be named after the process. If you call it a solder joint in the integrated circuit industry, it will make others stunned.

We might as well boldly understand the design technology and manufacturing technology of integrated circuits by printing. When we want to print business cards, we should first design the layout, think hard about a trademark pattern, draw on the draft paper with colored pens, and even pay artists to design this business card.

The designed business cards may contain several colors and fonts. When the design scheme is sent to the business card-introducing shop, it will be published according to the types of colors, then printed on paper with these plates, then cut a large piece of paper into small business cards, and finally put these business cards into small boxes, waiting for users to pick up the goods.

A small shop owner who prints business cards only needs to know how to play with such things as plate-making machines and stencils, and basically doesn't need to know many words, and only needs to have a primary school education. It is an unreasonable behavior to ask him to know the principles of art before he can be issued with an operation certificate.

The greatest requirement for the artist who designs business cards is that he can understand customers' requirements, and can express customers' intentions with layout, font selection and color matching. There is absolutely no need for us to require the artist to type in various fonts, read a large volume of Photographic Principles and be proficient in printing machinery.

accordingly, we should have only one requirement for the customer: simply and clearly put forward his intention.

The same is true for integrated circuits. The user puts forward the functions of his chip. The design process is to translate the functions into the layout, and the manufacturing process is to print the chip with the layout.

The striking similarity between printing and chips is that the artist only needs to use various fonts in the font library, and there is no need to make one font by himself. Chip designers actually only need to use the cells in the cell library.

Another striking similarity between printing and chips is that when there is no specific symbol in the font, such as an unusual trademark, the artist has to make the symbol himself. Chip designers sometimes encounter units that are not in the library, and he also needs to draw the layout of this unit himself.

top

------------. It is the layout of an oscillator. You connect it with a resistor, a capacitor and a light-emitting diode, and then connect it with a power supply, and you can see that the light-emitting diode flashes.

that's about it for the layout design of integrated circuits, and there are no tricks.

You may find that the rightmost inverter is a little larger, which is to have greater driving ability to make the LED have higher brightness. Another problem that should be paid attention to is to make a relatively large current flow in the aluminum conductive layer as much as possible. The conductivity of aluminum is much better than that of polysilicon, which is a little different from that of double-sided circuit boards, and the conductivity of both sides of double-sided circuit boards is as good.

top

------------.

Now, everyone has his own ideas. If you want to play some tricks on this jigsaw puzzle, you can. There are many friends here who are willing to play with you. Or if you remember the original intention of coming here and want to make your circuit into an integrated circuit, you should stop playing jigsaw puzzle for a while and try to finish your chip design as soon as possible.

If you used many standard integrated circuits in the circuit designed with discrete components, such as inverters, NAND gates, D flip-flops, counters, and even a LCD driver module, it is reasonable to say that it is not an easy job for you to design the layout of these units. Day after day, month after month, you have to be stupid and stupefied. You originally designed a high-performance circuit.

After you worked hard for three months, you found that the farmer's uncle who had started planting the second season rice was not so hard. From this calculation, the workload of making a version of the library is about equal to the workload of planting two-season rice. Being a chip designer is not as good as farming.

But you don't have to be afraid even if you are asked to draw a map. The most wonderful feature of high technology on the computer is that its labor can be reused. It takes two months to make a unit gallery for the first time. When you make a gallery for the second time, you can COPY the first gallery, modify it, and it only takes two weeks. This is the beauty of copy in the computer. It's strange why we don't extend this computer technology to agricultural science and technology. If the farmer's uncle also adopts these technologies, he just needs to concentrate on planting an acre of land, and then copy it 18, acres step by step, so he stands at the head of an endless golden field and watches the magnificent wheat waves roll in the morning light, so it's hard to control the feeling of "being in the head of the field and keeping the world in mind".

Fortunately, the farmer's uncle hasn't mastered this technology, otherwise we chip designers will be ashamed, but we should still hurry up and make more contributions than the farmer's uncle under the advantage that chips are temporarily ahead of farming.

But the situation is different now. Here, there are ready-made unit layouts for you to use. These unit layouts are put in a library, which contains a wide variety, almost all the varieties you may use. Don't think this is a "demo version". This is a working library that many chip designers are using. What you have to do now is not layout design, but something similar to what you used to do: wiring double-sided circuit boards.

With the cell library, you can carry out normal design work. You draw circuit diagrams in the usual way, and then connect them with corresponding components. In the past, an inexperienced circuit designer often designed circuits according to outdated textbooks. As a result, buyers often complained that they could not purchase components, and even had to go to Intel's company history showroom to see samples. Nowadays, circuit designers are generally concerned about the supply of the original devices. This experience is also applicable here. You should be familiar with the situation of the unit library before designing, so as to avoid drawing a circuit diagram without any units available.

after all, the content in the cell library used in chip design is a little different from that in the CMOS 4 series or TTL 74 series on the market. For example, the commonly used JK flip-flop can't be found in the cell library (maybe there will be enthusiasts to make up this cell in the future), because it is unnecessary to use JK flip-flop in ASIC design, and it will cause confusion when analyzing the circuit. You can adapt to these differences quickly, even after you read through the introduction of the unit library, you will feel confident.

I remind you again that you are here to design integrated circuits, not to study layout design technology. If you are a beginner, the layout you designed will probably not be better than the existing layout in the library, right? So just know a little bit about layout design for the time being, don't delve into it, and use the ready-made unit library and double-sided wiring technology to complete your integrated circuit as soon as possible. If you encounter a unit that is not in the gallery, don't panic. Please come here and ask. Maybe an old hand will draw one for you. However, it should be some common units. Don't ask about the 16-bit A/D converter and the control circuit of the camera. In that case, you might as well sleep and let others design it for you for free.

To be honest, the great and mysterious layout design of integrated circuits is not much more complicated than the jigsaw puzzle played by children. The layout of the flip-flop below is almost the most complicated layout in the layout library, but it's just like this. It's no big deal. I don't think you can't master it. I just don't think you need to master it for the time being.

Don't think that this is a demonstration layout, but it is a real layout in use. I call it DF6A, and it took a designer about a month to make a similar layout. I drew this layout on his basis, so it only took about half a month, and some delay parameters have not been calculated. With this layout, almost one-third of the workload of the entire standard cell library (more than 1 cells) has been completed, because this is the most difficult layout in the library.

top

------------.