The integration of MCU and sensor, the integration is unstoppable

Photocoupler

As the “sensing hub” of electronic products, sensors are becoming more and more widely used in the fields of consumer electronics, industry, medical care, and automobiles. In addition to basic functions, sensors are increasingly taking on auto-zero, self-calibration, and self-calibration functions. At the same time, it has the logic judgment and information processing capability, and can perform signal conditioning or signal processing on the measured signal. This requires that it has more and more intelligent processing capabilities, that is, toward the direction of intelligence. In order to enable customers to complete system development faster and more conveniently, some sensor manufacturers began to integrate MCUs and sensors to provide a modular development platform for MCU+ sensors, which has gradually become a product development trend.

However, whether it is scene, location and environment perception, or interaction functions based on motion, touch and gestures, it is often necessary to implement the sensor's “always on”, which is in line with the trend of low power consumption of the Internet of Things and mobile communication terminals. Related information. MCUs tend to have multiple modes of operation, requiring very low power consumption in sleep mode. But "always on" poses a huge challenge to the low power requirements of MCUs and sensors. How to solve this contradiction is related to the development path of MCU and sensor. To this end, the author interviewed mainstream manufacturers in the industry to explore the direction of technology development.

Geoff Lees, Senior Vice President and General Manager, Microcontroller, Freescale

MCU+ sensor will go to SoC

There is no doubt that the Internet of Things will become another emerging industry in the global information and communication industry. Security, scalability and energy efficiency are the three elements driving future IoT development. The basic requirement of the Internet of Things is that objects are connected, and each sensor that needs to be identified and managed needs to be equipped with a corresponding sensor. With the advancement of the Internet of Things technology, sensors are required to have not only basic information collection functions, but also intelligent information processing capabilities have become an important basis for judging their performance. Because it is impossible to put all the operations into the cloud, each node of the network has to complete its own computing tasks. Therefore, a single-chip integrated smart sensor with various functions combined with a sensor and a microprocessor has become a major development direction. Sensing technology and MCU are developing rapidly and integration is getting higher and higher. Although there are many differences between MCU and sensor technology, it is difficult to achieve integration. But SoC is a big trend in the industry. We are closely monitoring whether we can achieve the fusion of the two processes at the next process node, such as 28nm.

In addition, the growth rate of 32-bit MCUs has far exceeded that of 8-bit and 16-bit MCUs. 32-bit MCUs are growing at a rate of more than 15% globally, with 32-bit MCUs in the ARM architecture growing at twice this rate. In the past two years, 32-bit MCU shipments have doubled, and the previous rule was usually five years to double. The global growth in 2014 was mainly due to the contribution of IoT and smart internet growth. The Kinetis family based on the ARM architecture is one of the best performing MCU products on the market and has maintained strong growth momentum. We believe that the ARM ecosystem will enable us to better serve our customers, streamline their designs, and enable them to focus more on applications, software and services.

Ross Bannatyne, General Manager, NXP Microcontroller General Market Product Line

Solve sensor high power challenges with MCU "listening" function

Today's mobile devices introduce advanced features such as scene, location and context awareness, as well as motion, touch and gesture-based interactions, and even voice activation, through an ever-increasing "always-on" sensor. In the near future, other industries will also need product features similar to smartphones to enhance their customer experience, and thus make significant progress. Currently, many existing MCU sensor processing architectures consume too much power or cannot scale effectively as the number of sensors increases. The LPC54100 series requires only 3μA of very small current for continuous sensor monitoring, which is critical for applications that are always on. In addition, as a pioneering feature in sensor applications, the family's asymmetric dual-core architecture enables power consumption/performance at work, allowing developers to use the Cortex M0+ core (55μA/MHz) for sensor data acquisition. , integration and external communications to optimize energy efficiency, or use the Cortex M4F core (100μA/MHz) to perform complex mathematically intensive algorithms (eg, motion sensor fusion) faster while saving energy.

The architecture features a range of analog and digital interfaces designed for energy efficiency, including 12-bit, 4.8-Msps/s ADCs that support a wide range of performance over a wide voltage range (1.62V to 3.6V). Low-power serial interfaces that allow the LPC54100 Series to deliver lower power than other similar microcontrollers.

Russell Jordan, Senior Marketing Manager, Microelectronics, Austria

Integrating MCU and sensor SIP is the current mainstream

As an important part of the IoT perception layer, the role of smart sensors is becoming more and more obvious. The smart sensors formed by MCU+ sensors are increasingly becoming an option for microelectronics manufacturers to develop products. Integrating sensors and MCUs will be the future development trend, but the industry should go with the flow. For example, CO2 sensors, humidity sensors, brightness sensors, etc. are very difficult to integrate with MCUs. At present, there is not much disadvantage in the separation of the two. It is not necessary to force; but for some types of sensors that are relatively easy to integrate, such as touch screen controllers, accelerometers, and gyroscopes, after the market demand has expanded, some manufacturers have already made SoCs, and manufacturers should quickly seize the opportunity.

The MEMS process and MCU process used by the sensor are different, and the SoC integration of the two has certain challenges in semiconductor technology. In contrast, SIP solutions are more feasible. Many MCUs have now integrated with sensors in this way. We will see more of the MCU architecture designed for automatic acquisition of sensor data, enabling sensor data aggregation on IoT-optimized ultra-low power platforms.