A short guide on IoT Microcontroller chip selection

Date: 03/03/2017
IoT Microcontroller Chip Selection

The IoT structure is comprised of the edge nodes, the gateway nodes/hubs, and the cloud/datacenter. The interface between the virtual, digital world of the Internet or local network, and the real, analog world is provided by the edge nodes. An edge node is the "thing" in the IoT. Based on the type of application, the edge node can gather and/or receive sensed data, process and transmit it to the network. Hence, the functionality of the edge node can be described in terms of its integrated microcontroller and the associated transducers, interfaces, processing and communication to the network.

The IoT microcontrollers(MCU) with Ultra Low Power (ULP) are the computing engines that powers the edge of the Internet of Things (IoT). The ‘things’ living on the edge utilize Microcontrollers (MCUs) to handle user interfaces, collect and transmit sensor data, and provide security features. Hence, the application developers are challenged with a number of optimization challenges to ensure the very low power operation over periods of years and decades.

The MCU market has a strong relationship with the IoT trend, as the small nodes used for connectivity, and sensor hubs to collect and log data, are primarily based on MCU platforms. The IoT related MCU market is predicted to reach US $2.8 billion in 2019.

Security – Inarguably a prime factor to consider!

According to a study, it’s found that 70 percent of popular consumer IoT devices are easily hackable. It’s strongly advised to design and develop IoT solutions by adopting a multilayer approach that integrates device and data security into IoT solutions irrespective of the IoT device’s intended field of operation (industrial, commercial, or consumer systems). Effective device security would require efficient security processes, solid isolation of the security code and keys, and protection against both remote and physical attacks.

Being at the center of the IoT hardware, the selection of IoT microcontrollers is of great importance and indeed a tedious task itself. Given the IoT application, the foremost considerations to make while choosing the semiconductor chip (microcontroller or MCU) are given below.

Features / Functions neede

1. Integrated security (tamper resistance, optional JTAG lockout, CRC-16 and AES- 128 routines in ROM, reserved flash space for UID, true random number generator, etc.,).
2. Power-efficient at the edge (battery powered).
3. Wider software market (access to internal registers, debugging capabilities, etc.,).
4. Internet protocols supporting latest wired and wireless connectivity technologies.

Microcontroller-specific Chip Parameters / Specifications:

1. Very low power (Radio, Standby, Operating)
2. Data bus width (4, 8, 16, 32 bits)
3. Clock and CPU speed (MHz)
4. Device core (ARM, PIC, STM8, MSP430, RISC, etc.,)
5. Memory type and size – Data RAM and Program(Flash, EPROM, ROMLess, etc.,)
6. Number of I/O’s (Debug interface, DC-DC converter, Transceivers, etc.,)
7. Number of Timers (Independent RTCs, Periodic interrupt timers, Programmable delays, etc/.)
8. Interface Type (WiFi, Bluetooth, I2C, CAN, CMT, SPI, LPUART, TSI, NFC, etc.,)
9. ADC and DAC Channels
10. Supply or Input Voltage
11. Operating Temperature Range

Sample list of IoT MCUs from Top Manufacturers:

Tabble

Tabble2

Some more analytical points on IoT Microcontrollers


IoT applications ideally require a single chip at an acceptable form-factor and with ultra-low power consumption that can extend battery life. Low-power design techniques such as multi-threshold CMOS can be explored to optimize the design of power.

Combining GPS, accelerometer and gyro measurements or any such sensors in a PCB-mountable package is a better option than offering a single MCU chip.

The larger flash and RAM sizes allow designers to implement the entire networking stack and application code on the Microcontrollers without needing an additional processor in the system.

The higher processing capacity of MCUs enable them to complete processing sooner in order to enter sleep mode and preserve power.

An 8-bit MCU can be chosen for a simple end sensor node that senses and transmitts data a few times a day. However, for advanced end nodes or gateway devices that build in intelligence or need to run an RF protocol stack or other sophisticated algorithms, a 32-bit MCU is a better choice.

Some 32-bit MCUs such as those based on the ARM Cortex-M4 core also include a floating-point unit (FPU) that proves useful for implementing complex algorithms. It’s always better to choose a processor with more IoT-specific processing features inside the chip.

This article is in continuation to other main article

The hardware and software components eco of IoT – Internet of Things
http://www.eeherald.com/section/news/owns20170219001.html
Also read another related article:

Smart and connected IoT sensors for your IoT edge design
http://www.eeherald.com/section/news/owns20170326001-iot-sensors-sg.html

Author: Srinivasa Reddy N
Header ad