Comparison Of Different Types Of Ic Design Computer Science Essay

The topic of this assignment is to compare on different types of IC design. Therefore, we must first understand the meaning of IC. An IC, integrated circuit, is also known as microcircuit, microchip, silicon chip, or chip. It is a small electronic circuit that consists of semiconductor devices and other passive components. These components are manufactured on a ceramic or plastic container. Internal connections are welded from the chip to different number of external pins depending on the chip’s function.

IC design means that using logic and circuit design techniques to design and produce integrated circuits. It requires pathways so that information can flow properly and small electrical components are organized to maximize space so that computer systems can be as small as possible. Living in this modern age of data information, we need to design faster and smaller IC to increase our daily life quality and to catch up with the technology.

IC design is classified mainly into two categories of analog and digital IC design. Digital IC design is the most widely used in daily life such as microprocessor, FPGAs, memories (RAM, ROM, and flash) and digital ASICs. On the other hand, analog IC is used in the design of sensors, power management circuits, and operational amplifiers.

The figure of IC is shown below :

Figure 1(a) Figure 1(b)

(Adapted from introduction to ASIC, http://iroi.seu.edu.cn/books/asics/Book2/CH01/CH01.htm [1])

Figure 1(a) shows an IC chip where the pins will fit into holes in a printed-circuit board or breadboard. Figure 1(b), the silicon chip (more properly known as a die) is fitted in the cavity under the sealed lid.

A more detail classification of IC based on digital and analog is as follow :

Chart 1 : Hierarchical classification of IC

Referring to chart 1, as we are more interested in the branch of digital IC design, I expand more on its hierarchy compare to others. I will first compare between analog and digital IC design. Then I will explain on the advantages of each of them and mixed-signal IC design. Next, I will compare and explain on full custom and semi-custom IC follow by a simple explanation of silicon compilation as it is not discussed in lectures. After that, I will compare gate array, standard cell and PLDs.

Content

Analog, digital and mixed signal IC design

The comparison between analog and digital IC design is tabled as below :

Characteristics

Analog IC design

Digital IC design

Noise

More susceptible to noise

Less noise

Precision

Less precise

More precise

Design difficulty

Harder to design

Easier to design

Complexity

High

Low

Power consumption

Low

High

Heat dissipation

Low

High

Design work

Demands strong understanding of the principles, concepts and techniques

done by copying and reusing the same circuit functions or library

Life cycle

10years

1 to 2 years

Cost

Low

High

Information storage

Noise from aging degrades information

noise-immunity makes information not degraded

Table 1 : Comparison between analog IC design and digital IC design

Analog IC design

In analog IC design, the analog signals take any value from a given range, and each unique signal value represents different information. It is often found in op-amps, linear regulators, phase locked loops, oscillators and active filters. Therefore, a slight change in the signal may affect the design. For example, an analog signal is used to represent temperature, with one volt representing one degree Celsius. Therefore, 10 volts would produce 10 degrees, and 10.1 volts would produce 10.1 degrees. Analogue IC design produces noise, which is a random disturbance, variation or random thermal vibrations of atomic particles. Since any changes in an analogue signal is significant, any disturbance will change in the original signal and appears as noise. As the signal is copied and re-copied, or transmitted over long distances, these random variations become more significant and lead to signal degradation. Other sources of noise may include external electrical signals or poorly designed components. These disturbances are reduced by shielding, and using low-noise amplifiers.[2] However, some irreducible noise such as the shot noise in components will make an analog IC design imprecise. When designing an analog circuit, the choice of every single component, size, placement, and connection is crucial. Every small detail such as the resistance, placement and number of resistor, will affect the performance of final result. Therefore, designing an analog IC requires strong understanding of the principles, concepts and techniques. Hence, it is said that analog IC design is much more complex compare to digital IC design. It is harder to design because analogue circuit must be designed by hand, and the process is much less automated than digital IC. However, once an analog IC is designed successfully with high signal to noise ratio, low distortion, low power consumption, high reliability and stability, it can have a life cycle of more than 10 years. As a result of the long life cycle, the price of analog IC is low.

Digital IC design

A digital IC is designed to accept only input voltages of specific values and it uses only two states which are the binary quantities, “on” and “off” representing 1 and 0 or “true” and “false”. This is achieved by using the logic of Boolean algebra. The three basic logic functions in a digital IC are NOT, AND, and OR. A truth table is needed to design a digital IC. As discussed in the lectures, the design of digital systems is divided into combinational systems, which is a representation of a set of logic functions, and sequential systems, which are state machines. It is often found in microprocessors, FPGAs, memories (RAM, ROM, and flash) and digital ASIC.

Digital IC design produces less noise or even no noise. Digitally represented signals are transmitted using binary sequence of 1 and 0. It can be reconstructed, retransmitted or transmitted over long distance without any error provided the noise during transmission is unable to alter the 1s and 0s. Hence, digital IC is more precise compare to analog IC. Even in a compact disc of around 6 billion binary digits, the information or data can be presented precisely because each digit is handled by the same kind of hardware and there is no noise in the handling process. Digital IC is easier to design because it is controlled by software such as electronic design automation tools (EDA) so that functions can be altered without changing the hardware. If consumers detect error, they can simply upgrade the software to rectify the error. As digital IC is almost immune to noise, information can be stored and retrieved completely and precisely without any damage or degradation.

However, there are some disadvantages of digital IC. Because digital IC is very dense in circuitry, digital circuits use more energy than analog circuits to accomplish the same tasks, thus producing more heat. In portable or battery-powered systems this can limit use of digital systems.[3] Digital IC emphasizes on speed and cost ratio computing to achieve the lowest possible cost with the highest operating speed. Designers must use more efficient algorithms to process digital signals, or use new process to improve the integration cost. Therefore, the life cycle of digital IC is very short, about 1 year -2 years and the cost is higher than analog IC.[4] Since digital circuits involve millions of times as many components as analog circuits, much of the design work is done by copying and reusing the same circuit functions, especially by using digital design software that contains libraries of pre-structured circuit components. [5]

Mixed signal IC design

A mixed-signal integrated circuit is any integrated circuit that has both analog circuits and digital circuits on a single semiconductor die.[6] Mixed-signal can be found in ADC or DAC and digital radio chips. Since mixed-signal IC requires both analog and digital design, it is usually designed for a very specific purpose and because of that, their design requires a high level of expertise and careful use of computer aided design (CAD) tools. Therefore this type of design is very complicated andcostly.

Full-custom and semi-custom IC design

The comparison between analog and digital IC design is tabled as below :

Characteristics

Full-custom IC design

Semi-custom IC design

Circuits

Customized

Predesigned

Manufacturing time

Long

Short

Performance

Maximize

Moderate

Area of IC

Minimize

Moderate

Cost

High

Low

Table 2 : comparison between full-custom and semi-custom IC design

Full-custom IC design

Full-custom design is a methodology for designing integrated circuits by specifying the layout of each individual transistor, logic cells, mask layers and the interconnections between them.[7] Basically, the IC is designed from scratch and tailor-made to meet the requirement of a specific purpose. The main goal of having a full-custom design is to maximize the performance and minimize the area of an IC. Therefore, a lot of researches and studies are needed to produce a full-custom IC which results in very high production cost and long manufacturing period. Usually, full-custom IC is catered for large production so that the high production cost is fully utilized. Full-custom IC is produced from time to time when there is no suitable existing libraries available that can be used for the design. This is because existing libraries are outdated or consume too much power.

Semi-custom IC design

Semi-custom IC design can be partly customized to serve different functions within its general area of application.[8] It allows a certain extend of modification during the manufacturing process. It has the diffused layer fully defined but the libraries of pre-structured circuit components with the same circuit functions can be reused. This can save a lot of time and cost to in producing a semi-custom IC design. Therefore, the manufacturing cost if low and it is used widely in almost every IC design around the world. Semi-custom IC design is further classified into 3 groups which are gate array, standard cell and programmable logic devices circuits.

Silicon compilation IC design

Silicon compilation is to use a software system that takes a user’s specifications and automatically generates an integrated circuit (IC).[9] Generally, a designer is given a description of the system, by using a silicon compiler, mask and test information are produced which is either a simple combinational circuit or a finite state machine. The first step of silicon compilation is Convert a hardware-description language such as Verilog or VHDL or FpgaC into logic. Next, we shall place the logic gates on the IC followed by routing the standard cells together to form the desired logic.[9] A drawback of this method of IC design is that most of the silicon compilers do not utilize the area of silicon efficiently. Therefore, it is usually produced in small volume. It may used to generate simple cells to build up standard cell libraries.[10]

Gate array, standard cell and programmable logic devices (PLDs) IC

The comparison between gate array, standard cell and programmable logic devices circuits is tabled as below :

Characteristics

Gate array

Standard cell

PLDs (FPGA)

Flexibility

Flexible

Very flexible

Less flexible

Risk

Less risky

Less risky

Risky

Manufacturing Cost

Moderate

Cheap

Expensive

Manufacturing Difficulty

Moderate

Easy

Very complex

Manufacturing time

Shortest

Short

Moderate

Silicon size

Moderate

Small

Limited

Speed

Slow

Moderate

Fast

Heat dissipation

Low

Moderate

High

Table 3 : Comparison between gate array, standard cell and PLDs (FPGA) IC

Gate array IC

In a gate-array-based IC, the transistors, logic gates and other active devices are predefined on the silicon wafer. The only uncompleted part of the creation is the final surface layer, which defines the interconnect between the elements. Connecting these elements allows the function of the IC to be customized. Therefore, it is very flexible and less risky since it uses predefined elements. Furthermore, adding a surface layer of interconnects requires only a small cost and short time to complete. However, the chips designed using gate array techniques are a bit bigger in silicon area than standard cell IC, which makes them more expensive and harder to manufacture.[11]

Standard cell IC

In standard cell IC, different sizes of predesigned cells are used and a big combination of cells can be formed which is known as mega cells. Mega cells can be found in microcontroller or microprocessor. These cells, which consist of logic functions such as gates, latches, buffers and flip-flops, are known as standard cell library. Designer only needs to define only the placement of the standard cells and the interconnects in a standard cell IC.[12] Standard cell IC is flexible because it uses both digital and analog functions. The transistor sizes can be changed to improve speed and performance. It has a smaller silicon size and therefore a more compact are compare to gate array IC. In addition, it has faster speed which results in higher heat dissipation.

(Adapted from Standard-Cell-Based ASICs, http://iroi.seu.edu.cn/books/asics/Book2/CH01/CH01.1.htm#pgfId=1331)[1]

Figure 2 shows A cell-based IC die with a single standard-cell area (a flexible block) together with four fixed blocks.

Programmable logic devices IC

PLD is an electronic component used to build reconfigurable digital circuits an undefined function at the time of manufacture.[13] Before starting to implement PLD in a circuit, it must be configured or programmed to create a part customized to a specific application. This makes PLD a very flexible design to fulfill any custom specification. However, the limited size (<10000 equivalent gates) of PLD IC makes it not as flexible as array gate or standard cell IC. Since PLD can only be modified using software, it cannot have any last-minute design changes. The equipment to manufacture a PLD IC is cheap but large devices can be expensive to the extent of several hundred dollars per chip. Since PLD IC is very compact and operates in high speed, it produces a significant amount of heat at the same time.

(adapted from Programmable Logic Devices, http://iroi.seu.edu.cn/books/asics/Book2/CH01/CH01.1.htm)[1]

Figure 2 shows a PLD die. The macrocells typically consist of programmable array logic followed by a flip-flop or latch. The macrocells are connected using a large programmable interconnect block.