Choosing the Right Components for Embedded System Development: Chips, RTOS, and Peripheral Devices

Source: DigiKey

Choosing the Microprocessor for Embedded Systems

Every electronic system requires a microprocessor (MPU) core; some systems may opt for microcontrollers (MCU), digital signal processors (DSP), field-programmable gate arrays (FPGA), or single-board computers (SBC) to handle the computation and control tasks. Below is an overview of microprocessors and related products.

The Microprocessor is the Core Component of a Computer

The microprocessor unit (MPU) is an integrated circuit (IC) whose main function is to execute the instruction set in a computer system, regarded as the brain of the system, responsible for processing and executing various calculations, controls, and operations, including program code execution, data retrieval/sending to memory, and applying mathematical logic to data. The microprocessor is also known as the central processing unit (CPU) and is one of the most important parts of a computer. The main functions of a microprocessor include computational capabilities to perform various mathematical and logical calculations, allowing it to handle large amounts of data and execute complex computations, which enables the computer to run various applications. Additionally, microprocessors include a control unit that interprets and executes instructions stored in memory. The speed of a microprocessor is typically measured by frequency, representing the number of clock cycles executed per second; higher frequencies indicate faster processing speeds, allowing more instructions to be processed simultaneously. Modern microprocessors often adopt multi-core technology, meaning they contain multiple processing cores, enabling them to handle multiple tasks simultaneously, improving overall performance and multitasking efficiency. Furthermore, microprocessors contain caches to temporarily store frequently used instructions and data to enhance access speed, reducing the need to read data from main memory and improving system performance. When designing with a microprocessor, one must typically choose the appropriate microprocessor based on the desired instruction set architecture (RISC-V, x86, and ARM). Microprocessors are primarily used in servers, workstations, computers, mobile devices, and gaming machines.

Microcontrollers (MCU) are Easier to Integrate into SystemsMicrocontrollers (MCU) are integrated circuits in embedded systems. Although microcontrollers offer lower performance than microprocessors, they are easier to integrate into systems and are also less expensive. The functions of a microcontroller typically include one or more processing cores responsible for executing embedded program code, these processors usually operate at lower frequency speeds but are sufficient for specific applications. Additionally, microcontrollers often include internal flash memory for storing program code and random access memory (RAM) for temporarily storing runtime data. Microcontrollers also feature input/output interfaces to connect with external devices, such as sensors, displays, communication interfaces, allowing data exchange with the outside world. Microcontrollers typically come with built-in timers and counters for executing timed tasks, counting pulses, or generating timing signals. They can communicate with other devices through different protocols such as UART, SPI, I²C, and include power management circuits to effectively manage energy consumption, extending battery life or saving energy. Moreover, some microcontrollers may have dedicated hardware for specific applications, such as analog-to-digital converters (ADC), digital-to-analog converters (DAC), and PWM control (Pulse Width Modulation), to support various application needs. The characteristics of microcontrollers include high integration and optimization for specific applications. This makes them an ideal choice for many embedded systems, providing sufficient computing and control capabilities while being compact, low-power, and cost-effective. Microcontrollers typically run real-time operating systems (RTOS) and are mainly used in vending machines, medical devices, household appliances, and robotics.

Digital Signal Processors (DSP) Meet Real-Time Processing Needs

Digital Signal Processors (DSP) are specially designed microprocessors for processing digital signals. Their main function is to efficiently execute tasks such as audio processing, image processing, and communication processing. Typically, DSP chips have high-resolution analog-to-digital converters (ADC) and digital-to-analog converters (DAC), as well as digital filtering capabilities. DSPs are optimized and designed to execute complex digital signal calculations quickly to meet real-time processing demands. DSPs focus on processing digital signals, with optimized hardware and instruction sets to enable efficient mathematical and signal processing calculations, allowing them to handle high-density data streams in real-time applications, such as audio or image data. DSPs often have multiple parallel processing units or channels, making it easier to process multiple signals simultaneously, which is useful for applications requiring synchronous processing of multiple data streams, such as multi-channel audio processing or multi-antenna communication systems. The instruction sets of DSPs typically include instructions optimized for digital signal processing tasks, such as fast Fourier transforms (FFT) and convolution calculations, making it more efficient to implement digital signal processing algorithms on DSPs. Many DSPs are designed for low-power applications, making them suitable for battery-powered or power-efficient applications, such as portable audio devices or wireless communication devices. Additionally, there are digital signal controllers (DSC), which can be seen as a specialized combination of DSP and MCU. DSCs typically include functions found in microcontrollers, such as watchdog timers, pulse width modulation channels, and the ability to program using low-level languages (such as C or assembly language).

Field-Programmable Gate Arrays (FPGA) with Reconfigurable Capabilities

Field-Programmable Gate Array (FPGA) is a semiconductor device that allows users to program and configure logic blocks (CLB) interconnections to adapt to application needs and add more functionality. One of the main characteristics of FPGAs is their reconfigurability, allowing users to reconfigure their internal digital logic devices without changing hardware. This allows FPGAs to adapt to different application requirements and be updated over time. FPGAs have configuration memory to store logic device configuration information, which can be generated by software or hardware design tools and then loaded into the FPGA to perform specific functions. Due to the presence of multiple programmable logic devices within FPGAs, they have parallel processing capabilities, enabling them to perform multiple tasks simultaneously, making FPGAs suitable for applications requiring high parallel processing. FPGAs are widely used in digital signal processing, embedded systems, high-performance computing, communications, image processing, networking acceleration, and prototyping. Because of their reconfigurability and high customization features, FPGAs are often used for developing and testing new hardware designs. While most FPGAs can be reprogrammed multiple times due to the use of static random-access memory (SRAM), there are also one-time programmable (OTP) options, and FPGAs typically have a higher unit price compared to other embedded options.

Single Board Computers (SBC) with Compact Size and High Integration

Single Board Computer (SBC) is a complete computer system with all major components integrated onto a single circuit board, providing computational power and processing speed in a compact form factor. SBCs include a central processing unit (CPU), memory, input/output (I/O) interfaces, storage devices, and other necessary components such as USB, HDMI, and network interfaces. Most SBCs are designed as low-power devices, making them suitable for energy-efficient applications such as embedded systems and portable devices. SBCs are widely used in embedded systems such as industrial control, automation, and medical devices that require a complete computer system to perform specific tasks. Due to the simple architecture and relatively low price of SBCs, they are also widely used in educational settings for students to learn about computer science, programming, and hardware design. SBCs are also ideal platforms for development and prototyping, allowing developers to quickly validate concepts, test software and hardware designs, and develop product prototypes. Many SBCs have a large community support base, including developers, manufacturers, and enthusiasts. These communities can provide technical support, development tools, tutorials, and third-party expansion modules, enabling users to better utilize these SBCs. Common SBCs include Raspberry Pi, Arduino, BeagleBone, and Odroid, each with its unique features and uses, allowing selection of the appropriate model based on specific needs.

Development of Embedded Systems and Selection of Real-Time Operating Systems

Once you have determined the instruction set architecture for the embedded system, you can start selecting evaluation boards and development boards corresponding to that instruction set architecture, as well as choosing the appropriate real-time operating system (RTOS) to begin system development. This article will introduce the characteristics of evaluation boards, development boards, and considerations for selecting real-time operating systems.

Evaluation Boards and Development Boards to Accelerate New Product DevelopmentWhen starting the development of any new product, engineers first consider the stages of prototyping, functionality improvements, and production design. To assist designers in completing this design process, a series of circuit boards are available for experimentation, development, and production to assist engineers in their development work. Evaluation boards and development boards are two common tools in electronic product design that help engineers evaluate and develop new electronic products. Additionally, expansion boards and System on Module (SoM) can be used to enhance development efficiency and expand functionality. In the process of electronic product design, evaluation boards and development boards can be used together; evaluation boards are used for rapid evaluation of hardware performance, while development boards are used for software development and integration. This combination helps accelerate the time to market for products while improving the overall reliability and effectiveness of the design.

Evaluation BoardEvaluation boards are circuit boards used for creating the first prototype or conducting experiments, providing a hardware platform for engineers to evaluate the performance of specific chips or devices. Evaluation boards include basic circuits of the key chips of interest and contain the minimum number of components needed to form an executable circuit. Engineers can use evaluation boards to quickly assess the performance of target devices, test their functionality, and understand how to integrate these devices into the final electronic product. Although their design is simple, they often have a larger footprint to provide space for separating each line of the chip, allowing designers to accurately select which circuits to enable or connect. Evaluation boards typically offer a user-friendly interface to facilitate experimentation and evaluation by engineers.

Sparkfun Development Board

Development BoardDevelopment boards are hardware platforms used for prototype development and software writing, typically disconnecting the chip being tested from all external communication lines to facilitate access to available interfaces and programming options. The power is controlled by onboard regulators or other power devices, allowing developers to focus on testing the core chip’s functionality rather than other supporting circuits. Development boards usually include the target processor, memory, ports, and other necessary hardware, enabling developers to conduct software development, testing, and debugging. The most popular series of development boards typically have some degree of interoperability between manufacturers or product lines, allowing expansion boards to be reused across multiple devices. The Arduino series is an example of this, with its plug/socket arrangement standard becoming the de facto standard for many development boards.

Arduino Development Board

Expansion BoardsIn most cases, to easily integrate sensors, displays, motor drivers, or other chips, many manufacturers have placed external chips and necessary devices on compatible plug-and-play boards called expansion boards. Expansion boards can be easily installed on top of development boards and provide downloadable development environment software libraries.

Microchip’s Module System

Module SystemIn addition to evaluation boards and development boards, there is also a System on Module (SoM) used for the final step of transferring projects from the workbench to the production line. SoM is the most critical part of the circuit board, compressed into the smallest possible space for direct integration into application circuit boards. Typically, SoMs occupy a very small footprint, with the largest overall area usually only a few square inches and are made of surface-mounted devices. Most SoMs have a development platform that places the SoM directly on a breakout board for mass use and experimentation.

Diverse Expansion Boards Available on DemandExpansion boards are used to expand system functionality. These expansion boards typically have a series name or thematic functionality to link them to the main board. Here is a small list of some of the most popular development board series and expansion boards.

Arduino Series

Arduino is an open-source electronic prototyping platform designed to simplify hardware and software prototyping. Its expansion boards come in the standard UNO R3 format or the larger MEGA R3 format. The Arduino R3 circuit board layout has become the de facto standard, with many companies and third-party manufacturers using the same standard format to produce circuit boards.

Beaglebone Series

Beaglebone is a low-power open-source single-board computer (SBC) produced in collaboration between Texas Instruments, DigiKey, and Newark element14. It is also a system built to run open-source software, with its mainboard designed with open-source hardware/software capabilities.

Raspberry Pi Series

Raspberry Pi is a low-cost, compact, high-performance single-board computer developed by the Raspberry Pi Foundation in the UK.

Its HAT (Hardware Attached on Top) physical expansion module may have an older 26-pin variant or a newer 40-pin variant. pHAT (partial HAT) is a variant designed to match the form factor of Raspberry Pi Zero and Zero W boards.

Feather Series

Feather is an open-source hardware platform developed by Adafruit Feather for creating lightweight, compact embedded systems and modular electronic prototypes. Its development board system’s add-on and expansion boards allow for prototyping, sensing, and motor control, wired and wireless connectivity options, audio and display, designed to provide flexible, scalable, and easy-to-use solutions.

STM32 Nucleo Series

STM32 Nucleo is a development board platform launched by STMicroelectronics, aimed at simplifying prototype development and application development for STM32 microcontrollers. These expansion boards allow for adding additional functions to the basic STM32 Nucleo boards, such as sensing, control, connectivity, power, and audio.

MikroE Series

MikroE’s Click board (mikroBUS™ Click™) is a series of expansion boards with over 1000 different options, providing unparalleled design choices and speed when adding sensors, communication, displays, and data storage. To save more space on the circuit board, MikroE Shuttle and Click Ribbon cable systems allow adding up to four different Click boards to the same data bus.

MicroMod Series

MicroMod is a modular embedded platform launched by SparkFun Electronics, designed to simplify and accelerate the prototyping of embedded systems, where the core module contains the main processor and basic peripherals, while functional expansion modules (Carrier Boards) provide additional peripherals and connectivity options that can easily connect processors, communications, and sensors.

Real-Time Operating Systems Meet the Needs of Resource-Limited SystemsReal-time operating systems (RTOS) are small, lightweight operating systems typically designed to run on resource-limited small embedded systems, specifically designed to meet the requirements of real-time systems. RTOS is very sensitive to task response times, requiring tasks to be processed within specified time limits, which is why RTOS is widely used in fields requiring high reliability and determinism, such as aerospace, medical devices, industrial control, and automotive electronic systems.

Real-Time Operating System (RTOS)

RTOS focuses on the real-time nature of tasks, ensuring that tasks are completed within specified time frames and providing predictability for task execution, meaning that the time to complete tasks can be predicted within a certain range. RTOS uses a special scheduling algorithm to ensure that high-priority tasks are executed within specified time frames. Since real-time systems require rapid responses to external events, RTOS typically features efficient interrupt handling mechanisms. RTOS provides effective resource management mechanisms to ensure that tasks can access system resources according to priority and timing requirements, including shared memory and device access. As real-time systems are often applied in embedded systems and resource-constrained environments, RTOS typically has a modular design, allowing it to be trimmed according to application needs. Some well-known RTOS include FreeRTOS, Zephyr, VxWorks, RTOS-32, and QNX. The choice of an appropriate RTOS depends on specific application requirements, such as real-time requirements, system resource limitations, development costs, system complexity, and available hardware resources.

Zephyr

Free and Open Source Zephyr and FreeRTOS Real-Time Operating SystemsZephyr and FreeRTOS are currently the most popular RTOS choices, both being completely open-source systems specifically designed for embedded systems and Internet of Things (IoT) devices. Zephyr is supported by the Linux Foundation and is licensed under the Apache License 2.0, while FreeRTOS is licensed under the MIT License, allowing developers to freely access, view, modify, and distribute the source code. Both Zephyr and FreeRTOS support multiple processor architectures, with Zephyr supporting ARM, x86, RISC-V, etc., and FreeRTOS supporting ARM, MIPS, RISC-V, etc., making them compatible with various hardware platforms and suitable for different types of embedded systems and IoT devices. In comparison, Zephyr RTOS supports a more complete network protocol stack, with a built-in full TCP/IP network protocol stack, making it convenient for embedded systems requiring network connectivity. Zephyr emphasizes security, providing security-related features such as TrustZone support and cryptographic libraries to meet the security needs of IoT devices, and offers rich peripheral and driver support, including sensors, communication buses, etc., making it easier to integrate with various hardware and peripheral devices. However, due to the rich functionality of Zephyr RTOS, it may present a certain learning curve for beginners, and compared to some lightweight RTOS, the kernel of Zephyr may be relatively large, potentially unsuitable for extremely resource-constrained systems. Related Zephyr real-time operating systems include SparkFun WRL-17354, NXP MIMXRT1060-EVKB, Nordic Semiconductor THINGY53, Raspberry Pi SC0915, and a complete list of Zephyr RTOS compatible boards can be found at https://docs.zephyrproject.org/latest/boards/index.html. FreeRTOS is designed to be lightweight, suitable for resource-constrained embedded systems, with a smaller memory footprint and memory usage. FreeRTOS has extensive hardware support and can run on various processor architectures, providing versatility across different hardware platforms. FreeRTOS has a large community support base, rich documentation, sample code, and forum resources, enabling developers to easily get started and troubleshoot. However, compared to Zephyr, FreeRTOS has relatively limited network support, which may require additional work to implement complex network connections, and for some applications with complex functional requirements, FreeRTOS may not offer as many advanced features as more comprehensive RTOS. Related FreeRTOS products include STMicroelectronics NUCLEO-L476RG, Microchip Technology DM320003-3, Infineon KITXMC13BOOT001TOBO1, and a complete list of FreeRTOS compatible product lines can be found at https://www.freertos.org/RTOS_ports.html. The choice between Zephyr and FreeRTOS depends on specific application requirements, hardware platforms, and the experience and preferences of developers. If there are high requirements for network support, security, and multi-architecture support, Zephyr may be the better choice. If lightweight, broad hardware support, and a simple learning curve are prioritized, FreeRTOS may be more suitable for resource-limited applications.

Choosing Peripheral Devices for Embedded Systems

Besides the crucial choice of the processor, the associated peripheral devices are also an important component of embedded systems, including memory, clocks (oscillators), timers, communication interfaces, input/output, analog-to-digital conversion, etc., which can be selected based on the actual needs of the system. This article will introduce the types of peripheral devices mentioned above and considerations for their selection.

Types of Memory are Numerous and Have Different Characteristics

Memory

In embedded systems, memory is a critical device used to store program code, data, and other information required for system execution. In addition to the memory and storage space built into microcontrollers in IC packaging, external memory can also be added, and there are many types of memory, each with its unique purposes. First, the commonly seen flash memory in embedded systems is a type of non-volatile memory, meaning that it is not erased if the system is reset or powered off. It is typically used to store program code, firmware, and other resident application data. It has fast read speeds and relatively low power consumption. Flash memory comes in different types, including NOR and NAND, with different read/write characteristics and application scenarios. Another common type of memory is random access memory (RAM), which is a volatile memory used for temporarily storing data required for program execution. It has fast read/write speeds but loses stored data when the system is reset or powered off. Common RAM types in embedded systems include SRAM (Static Random Access Memory) and DRAM (Dynamic Random Access Memory). EEPROM (Electrically Erasable Programmable Read-Only Memory) is a type of memory that can be erased and is non-volatile, meaning that it is not erased when the system is reset or powered off, typically used to store configuration information, calibration data, and other non-volatile data that needs to be retained. Additionally, some systems may also connect external storage cards and embedded multimedia cards, including SD cards, MicroSD cards, etc., which are typically used to expand the storage capacity of embedded systems. Since different types of memory come at different prices, cost is an important consideration. Moreover, the read and write speeds of memory affect system performance, especially for applications requiring high efficiency, necessitating the selection of faster memory. The required memory capacity of the system should also meet application needs to ensure sufficient space for storing program code, data, and other necessary information. Additionally, different types of memory have different power consumption characteristics, particularly for mobile and battery-powered embedded systems, which typically require the use of lower-power memory. Embedded systems generally require durability, with some applications needing memory to have a long lifecycle and to withstand multiple write and erase cycles. On the other hand, some embedded systems may need to integrate multiple types of memory on a single chip, which will help save space and simplify design, and some applications may also require support for external storage devices, such as storage cards, to meet scalability requirements.

Clock Sources Ensure Synchronization of Various Operations in the System

Clock sources are critical devices used in embedded systems to ensure synchronization of various operations, including processor computations, external device communications, etc. Below are different types of clock sources with various characteristics. RC Oscillator is an oscillator made only with resistors and capacitors, temperature-dependent, with frequency signals that may vary by 1-5%. They can indeed meet some slower frequency timing requirements (such as low-frequency analog-to-digital conversion). Crystal Oscillator is commonly found in external oscillator circuits, characterized by high accuracy and good stability, typically measured in parts per million (PPM) rather than percentage (like RC oscillators), usually used in applications requiring high precision timing, often paired with microcontrollers, microprocessors, and other devices needing stable timing. Ceramic Resonator is less expensive but may have lower accuracy and stability than crystal oscillators, with tolerances in the tenths range, making them suitable for applications with less stringent timing requirements and cost-sensitive scenarios. MEMS Oscillator features small size, vibration resistance, and low power consumption, but generally has lower accuracy than crystal oscillators, suitable for size-constrained, low-power, vibration-resistant applications such as mobile devices and embedded sensors. Oscillator Module integrates oscillators and related circuit devices, providing a convenient external clock source, simplifying system design, commonly seen in highly integrated embedded systems. Another type is the GPS module, which provides high-precision clock synchronization by receiving global positioning system (GPS) signals, commonly used in applications requiring highly precise synchronization, such as communication systems and scientific instruments. Real-Time Clock (RTC) has low power characteristics, maintaining time counting even in power-off states, primarily used in applications requiring time counting to be maintained during power outages. When selecting clock sources in embedded systems, one should first consider the accuracy and stability of the clock source. Additionally, cost, power consumption, integration, and the application’s external environmental conditions (such as temperature and vibration) all affect the choice of clock source. In embedded systems, timers are common hardware modules used to generate accurate time baselines for executing timed and counting operations. Common timers are used to perform timing operations, such as generating precise time delays, calculating time intervals, etc., commonly found in applications requiring time control, such as communication protocols and sensor readings. Counters are used to count the occurrences of external events, such as pulse counting and frequency counting, commonly used in scenarios requiring the calculation of event occurrence frequency or counting, such as pedometers and measuring instruments. When selecting timers in embedded systems, one should first consider the precision of the timer, as different timers have different precisions, and the selection should be based on application requirements to determine the precision level. Additionally, the timing range (for timers and counters) should meet the time requirements of the application.

Communication Interfaces for Data Exchange Between Different HardwareIn embedded systems, communication interfaces are crucial components for enabling data exchange between different hardware modules, typically divided into parallel communication and serial communication types. Parallel communication sends multiple bits of data simultaneously, requiring data bus hardware, usually composed of multiple lines, allowing for faster data transfer while also using more I/O ports of the connected devices and requiring complex wiring configurations. Serial communication, on the other hand, sends one bit of data at a time over a single wire between paired devices, allowing device communication to use only one I/O port, reducing overall device complexity and cost. Serial communication can be further divided into two subgroups, depending on whether they use a frequency signal to control and synchronize data communication between linked devices (known as synchronous and asynchronous). Asynchronous serial means that data can be transmitted without requiring a frequency signal. Synchronous serial requires all devices to share a frequency signal to control data communication. While synchronous serial does require another timing signal across all devices, it indeed enables faster communication speeds. In terms of transmission modes, it can be divided into simplex, half-duplex, and full-duplex; simplex allows one-way data flow from source to destination, half-duplex allows bidirectional data transmission from paired devices, but only one direction can be transmitted at a time, while full-duplex allows two devices to transmit and receive simultaneously in both directions.

I²C Bus Protocol

The I²C bus protocol is a two-wire serial connection designed to allow multiple endpoints to communicate with one or more controllers. I²C is a half-duplex protocol that allows controller units and many endpoints to send and receive data. The protocol’s speed ranges from 0.1 to 5 Mbit/s (depending on the bus configuration), commonly used for connecting low-speed peripherals, such as temperature sensors and EEPROMs.

Serial Peripheral Interface (SPI) Protocol

Serial Peripheral Interface (SPI) is a full-duplex, synchronous serial connection requiring 3 or 4 wire connections. SPI connections require a shared synchronous frequency signal among all participants on the bus, enabling higher data rates. When two points are close to each other, speeds of up to 60 Mbps can be achieved. The downside of SPI is that it requires more I/O pins and connections, commonly used for connecting devices with SPI interfaces, such as memory, sensors, displays, etc.

Universal Asynchronous Receiver-Transmitter (UART) Protocol

Universal Asynchronous Receiver-Transmitter (UART) is a bidirectional asynchronous serial connection that can be set to simplex, half-duplex, or full-duplex. The data speed is slower and serves only one-to-one communication, with some devices possibly having multiple UART circuits, allowing multiple devices to communicate, commonly used for connecting embedded systems with external devices, such as sensors and GPS modules. Other common communication protocols include CAN (Controller Area Network), Ethernet, USB (Universal Serial Bus), and wireless communication interfaces. When designing communication interfaces, one can select based on the required data rate, transmission distance, power consumption, cost, integration, and real-time requirements, depending on specific application needs, such as embedded control systems, sensor networks, and communication devices. The input/output (I/O) interfaces in embedded systems are key components used to connect and control external devices, sensors, displays, etc. Common I/O interfaces such as GPIO (General Purpose Input/Output) offer versatility, configurable as input or output mode, for connecting various external devices, serving as general-purpose I/O connection interfaces, such as buttons, LEDs, switches, etc. Additionally, ADC (Analog-to-Digital Converter) is needed to convert analog signals into digital signals for reading sensor data, such as temperature and light. Conversely, DAC (Digital-to-Analog Converter) is used to convert digital signals into analog signals for producing analog outputs, such as audio output.

Source: DigiKey

Leave a Comment Cancel reply