The "exodus" of 16-bit MCUs

In 1971, the MCU was born. MCU Chinese name is microcontrol unit, which refers to the emergence and development of large-scale integrated circuits, the computer's CPU, RAM, ROM, timer and a variety of I/O interfaces are integrated on a chip to form a chip-level computer. Due to their flexibility, MCUs are widely used in consumer electronics, industrial control, automotive electronics, and communications.

With the continuous development of MCUs, from 1-bit, 4-bit, 8-bit, 16-bit to 32-bit or even 64-bit. Among the many types of MCUs, 8-bit and 32-bit MCUs have gradually become the current mainstream of the market, and the 16-bit MCUs sandwiched between them seem to be incomparable.

01

Excellent 8-bit and 32-bit MCUs

When discussing the changes in 8-bit, 16-bit, and 32-bit MCUs, let's first understand what "bits" are. The "bit" here refers to the "word length" of the MCU, that is, the length of the data involved in the operation in an operation. According to the general theory, the larger the number of bits, the higher the computing efficiency and the stronger the performance. But this is a theoretical result, and in the actual selection of MCU, factors that need to be considered include frequency, power consumption, price, etc.

The "stubbornness" of 8-bit MCUs

Forty years ago, Intel launched an 8-bit MCU with limited functions, the MCS-51, which is the ancestor of the current famous "51 microcontroller". Within a few short years of its introduction, the 8051 design led the embedded market, with several manufacturers launching their own versions of the 8051. MCS-51 series single-chip microcomputer is popular because of its reliable performance, simple and practical, and high cost performance, and is known as "the most classic single-chip microcomputer".

As more designs are added, the growth of this architecture is also adding numerous end products. The Atmel 8051 also became one of the best-selling products of the 8-bit Intel MCS-51. Later, the number of MCUs gradually increased, and 16-bit and 32-bit MCUs appeared in the market, so many people thought that 8-bit MCUs would gradually disappear over time.

Ten years ago, experts in the industry already said: "8-bit MCUs are dead". When Freescale launched the Kinetis L series, it also played the "8-bit MCU terminator", threatening to eat the 16-bit MCU market first, and then squeeze out the 8-bit MCU share.

However, ten years have passed, the reality seems to be not as expected, 8 MCUs still occupy a place in the market, from the official website of industry distributors, it can be seen that there are currently nearly 50,000 different MCUs available worldwide. Among them, the number of available 8-bit MCUs is the largest and accounts for the largest proportion.

What keeps an 8-bit MCU alive?

On the one hand, the most prominent feature of 8-bit MCUs is low power consumption and low cost. In terms of overall power consumption, although 32-bit MCUs can run up to four times faster, slower 8-bit MCUs always consume better power than faster 32-bit MCUs. In many embedded designs, 8-bit MCUs have the advantage of being easier to design than 32-bit, and software and hardware that follow 8-bit architecture are simpler than 32-bit. And in terms of price, with the increase in development cost pressure, embedded system brands will be more comfortable with the overall bill of materials (BOM) cost, and the low price of 8 digits still has a market advantage.

Therefore, in some cases an 8-bit MCU is sufficient. With this ultra-low cost and simple design, 8-bit MCUs are active in the market, especially in the Chinese market.

On the other hand, the development of 8-bit MCUs by manufacturers has not been reduced. To meet the demand for 8-bit microcontrollers (MCUs), Microchip introduced 5 new families, more than 60 devices, 8-bit PIC and AVR microcontrollers; BYD has further expanded its lineup of automotive-grade 8-bit general-purpose MCU series and launched the automotive-grade 8-bit MCU BS9000AMXX series.

That's why 8-bit MCUs are still fresh for many years.

32-bit MCU "show off their skills"

Excluding 8-bit MCUs, the current market share is 32-bit MCUs. According to IC Insights, the 32-bit share of the global MCU market will reach 67% in 2022.

In fact, it is reasonable that 32-bit MCUs can achieve such a proportion, although 8-bit MCUs are widely used and low cost, but overall performance is still limited. In high-end, complex scenarios, higher performance 32-bit MCUs are often required. For example, in automotive electronics, 32-bit MCUs are mainly used for vehicle control, intelligent meters, multimedia information systems, power systems, assisted driving and other high-end function control.

And automotive intelligence has greatly boosted the demand for 32-bit MCUs, according to the McClean report, more than three-quarters of automotive MCU sales in 2021 will come from 32-bit MCUs, which is expected to reach about $5.83 billion.

More and more MCU companies are planning to focus on the development of mid-to-high-end 32-bit products, MCU companies including Nuvoton Technology, Shengqun Semiconductor, Lingtong Technology and Weixian Electronics all believe that the proportion of mid-to-high-end 32-bit products must be increased to remain competitive.

In fact, at present, domestic enterprises represented by GigaDevice Innovation, Zhongying Electronics, Hangshun Chip, and Neusoft Carrier are also actively deploying the high-end MCU market. GigaDevice has become the largest supplier of ARMCortex-M based MCUs in China since 2013. In 2017, GigaDevice GD32MCU shipments reached 100 million pieces, and in 2021, it has risen to more than 1 billion units.

In addition, with the application of efficient and advanced process technology, the cost of 32-bit MCUs has been effectively reduced, and the price of 32-bit MCUs continues to approach the price of traditional 16-bit MCUs and high-end 8-bit MCUs by breaking through the $1 mark in 2015.

The combination of price reduction and performance improvement makes 32-bit MCUs irresistible appeal, and 32-bit MCUs can now be seen in many products that traditionally do not see 32-bit, such as watches or home appliances (such as refrigerators).

02

Why do 16-bit MCUs exist?

Under the continuous price reduction of 32-bit MCUs and the simple and durable attack of 8-bit MUC, the 16-bit MCU market has lost ground and retreated, becoming the product with the lowest shipment ratio, and even many companies have directly removed the product planning of 16-bit MCUs in their planning.

In fact, the 16 MCUs who are now "daddy doesn't hurt and don't love mother" also had a glorious moment. For a long time, the industry's long-term expectations for single-chip microcomputers have been low, and 8-bit computers are enough to use, believing that 16-bit and 32-bit are overperformance.

Later, 8-bit MUCs became more powerful, but due to limited processing power and addressable memory, performance limitations were inherent, so that 8-bit microcontrollers were often not used for high-performance applications. Systems based on 32-bit cores tend to go too far for these applications, especially if they result in unnecessary power consumption.

A 16-bit MCU located in the middle of an 8-bit and 32-bit core consumes as much power as a simple single-supply 8-bit MCU, while offering some of the performance and memory benefits of a 32-bit core. And, if the application does not require many threads to process simultaneously and memory requirements are not too high, a 16-bit MCU can provide the right level of performance and significant power savings.

As a result, for many low-power, mid-performance battery-powered embedded applications, developers at the time preferred to choose a suitable 16-bit MCU over a 32-bit core.

In 1996, Texas Instruments introduced a 16-bit ultra-low-power mixed-signal processor - MSP430 series single-chip microcomputer, mainly because of its practical application needs, many analog circuits, digital circuits and microprocessors integrated on a chip to provide a "single-chip" solution. Because of its low power consumption, fast speed, few instructions, assembly language flexibility and other characteristics, it is well received by the industry.

In 2005, Microchip announced that 16-bit MCUs will be one of the future development priorities, PIC24F, PIC24H in the PIC24 series, and the second dsPIC33F series with DSP control function, a total of 49 different models of 16-bit MCUs.

In 2011, 16-bit MCUs surpassed 8-bit MCUs for the first time, becoming the first product in shipments; Shipments increased by 23% in 2011 and continued to grow by 11% in 2012; By 2013, 16-bit MCU shipments continued to grow 6% to 6.7 billion units.

Until 2015, the global shipments of 32-bit MCUs exceeded the combined shipments of 4/8-bit MCUs and 16-bit MCUs, and 16-bit MCUs ushered in decline.

03

Will 16-bit MCUs die out?

From the perspective of current mainstream applications, 4-bit and 16-bit MCUs have almost disappeared, and 8-bit and 32-bit MCUs basically divide the entire application market. The 64-bit MCU market is still in its infancy, and most of the applications are in high-end workstations, multimedia interactive systems and advanced terminals.

Even in the current hot automotive field, 16-bit MCUs mainly used in the lower body, such as engine, electronic brake, suspension system and other power and transmission systems, are also losing armor in the growing performance and memory capacity of 8-bit MCUs and the growing number of 32-bit MCUs. However, in some critical applications of drivetrains, 16-bit MCUs still have a place to play.

However, 16-bit MCUs may not have to be so eliminated. It has higher performance than 8-bit and faster response time than 32-bit, and if it can compete with 8-bit products in terms of price and approach 32-bit in terms of performance, then 16-bit MCUs will also take a unique development path.

For example, Renesas introduced the RL78/G15, a 16-bit microcontroller with a main frequency of 16MHz, clearly stating that this MCU hopes to seize the market of 8-bit MCUs and use them to "replace" 8-bit MCUs.

The RL78/G23 is compatible with Renesas' current general-purpose RL78 MCUs, such as the RL78/G13, while improving power performance for battery-powered applications. In addition, to respond to market demand, on-chip flash memory capacity has been increased to 768KB and on-chip peripheral functions have been greatly expanded, thereby increasing functionality while improving security and reducing bill-of-materials (BOM) costs.

With these features, the new RL78/G23 is ideal for a wide range of applications that require both power and cost, including IoT end devices such as home electronics, remote controls, and sensors.

In addition, ROHM Semiconductor is also constantly upgrading 16-bit MCU technology to cope with the diversification of MCU demand in the home appliance market. ROHM's 16-bit MCU is based on the 8-bit MCU, including two models, ML620100 and ML620500, the former focusing on anti-interference and the latter focusing on low power consumption.

Perhaps under the diversified market demand, the 16-bit MCU that "left" can play its own light.