The next challenges of low power design

Article By : Maurizio di Paolo Emilio

Nowadays, the innovation of low power embedded systems is directly linked to battery innovation, which has created many interesting systems...

Nowadays, the innovation of low power embedded systems is  directly linked to battery innovation, which has created many interesting systems and applications.  Ravi Ambatipudi, vice president & general manager, mobile power business unit at Maxim Integrated, introduced his keynote for Embedded Electronics Forum at electronica 2020.

The emergence and exponential growth of breakthrough low-power IoT, wearable, hearable and edge devices has led to new system and IC design challenges where every nanowatt of power consumption or every picojoule of energy is drawn from the battery itself.

Lithium-ion batteries have created different types of systems, from low-power systems with less than one watt to high-power systems with high voltages. These systems have been created for different market segments, from consumer to healthcare. Lithium innovation continues improving  by offering different chemical combinations such as lithium cobalt oxide and lithium ferrous phosphate in different shapes and sizes, adapting to many applications. The advent of new battery management solutions will enable AI to be implemented even in ultra-low power devices.

Low power embedded systems
Every low power system that requires battery charging  uses a USB C charging port. All require fuel gauge devices to determine the charge status and, at the same time, protect the battery. These types of systems share the need for a sensor to detect information from the outside world, usually through a microcontroller that processes the information. Then, some kind of user interface and a communication transceiver are all managed by a power supply section that supplies power to all these blocks.

A designer has to guarantee a good charge shelf life, rapid recharging, and battery operating autonomy. “Now, these low-power devices are getting smaller and smaller. And there isn’t even room for a penny for the battery,” said Ambatipudi.

Some of these new generation low-power systems are in close contact with the skin. They can be worn inside the ear, for example, and they can’t get too hot. “It will be a very unpleasant experience if the earpiece or wearable device gets very hot; special attention must be paid to thermal performance. the integrity of the noise and signal and the quality of communication are also important aspects to consider,” said Ambatipudi.

Designers must also work to avoid providing a fully discharged product during first use. In other words, they have to work on the so-called “customer satisfaction”, as Ambatipudi pointed out. Manufacturers must meet their customers’ needs by providing a fully charged device at first use.This requires that the battery life is very long, avoiding unnecessary current losses. Current is a pivotal factor in achieving long battery life.  A value in the order of the quiescent current microampere provides a life of more than 50 months.

The energy density of a lithium-ion battery is increasing, not as fast as Moore’s parabolic law, but exponentially. The increase in density has many other consequences, especially on a safety basis. The battery separator becomes thinner, and safety becomes critical in some batteries. Small internal detections can actually develop with time due to harsh operating conditions, which can be aggravated by latent manufacturing defects, causing thermal leakage conditions. Therefore, fuel gauges are necessary (figure 1).

We all want to charge your luminaires faster. However, fast charging is not possible by simply increasing the power level of the charger as it also adds more power dissipation, which can actually heat the device. Efficiency is crucial to enable fast charging.

Within a power management, DC-DC plays an important role in efficiency. The designer must find a way to supply power to all the different sensors according to the subsystem form factor, keeping  in mind the small battery life and noise sensitivity. Each of these elements controls the audio amplifier, all the sensors and the LED display. All of them need a voltage current. But space is limited. Battery life is important, and at the same time low noise is crucial. So you need switching regulators, but with one inductor per switching regulator. By using the SIMO architecture, it is possible to generate multiple outputs with a single inductor.  By providing multiple outputs, the SIMO approach, together with the low standby current of the controller, extends the battery life of the wearable design. Regulators provide energy with minimal losses, and the architecture eliminates some duplicate components while saving on the bill of materials (Figure 2).Maxim low power design figure1

Figure 1. Energy density over the time (Source: Maxim Integrated)

Maxim low power design figure2 -simo
Figure 2: SIMO architecture (Source: Maxim Integrated)

AI for power
Bringing AI inferencing to the edge meant gathering data from sensors, cameras and microphones, sending that data to the cloud to execute an inference, then sending an answer back to the edge. This architecture works but is very challenging for edge applications due to poor latency and energy performance.

An alternative is given by low power microcontrollers that can be used to implement simple neural networks; however, the challenge is with latency and only simple tasks can be performed at the edge. The MAX78000 is designed to fill this gap. The MAX78000 is an advanced system-on-chip featuring an Arm Cortex-M4 with FPU CPU for efficient system control with an ultra-low-power deep neural network accelerator. The CNN engine has a weight storage memory of 442KB, and can support 1-, 2-, 4-, and 8-bit weights (supporting networks of up to 3.5 million weights).  This product combines the most energy-efficient AI processing with Maxim’s proven ultra-low power microcontrollers.

Artificial intelligence can be helpful if, with a more detailed analysis, it is able to determine whether there really is a walking person or whether it is a squirrel running outside. Other possible uses include better spatial awareness, or better vision, and deeper voice commands for small devices such as hearing aids or headphones.

Leave a comment