10 tiny chips for mobile devices

Mobile device and wearable designers face two big challenges: reducing board real estate and increasing power efficiency. Many component manufacturers address these challenges with new chips and solutions with higher integration and smaller packaging. The result is space savings or more room for a bigger battery, particularly for power-hungry devices with new features.

Power consumption continues to be a big challenge. Many of these new components implement new features like multiple power modes or advanced power management to extend battery life.

Designers also look for devices that are easy to implement to reduce development time, and they often need higher-performing devices to add new features. AI is also starting to drive a lot of the performance and functionality of these components.

Here are 10 tiny chips that shrink solution size and deliver higher performance for mobile device and wearable designs. These components include sensors, storage devices, power management ICs (PMICs), wireless microcontrollers (MCUs) and protection ICs.

Alif Semiconductor, Balletto wireless MCUs

Alif Semiconductor recently launched the Balletto family, claimed as the industry’s first Bluetooth Low Energy (LE) and Matter wireless MCU, featuring hardware optimization for AI/ML workloads. It leverages the company’s advanced MCU architecture with DSP acceleration and a dedicated neural processing unit (NPU) for fast and low-power execution.

The Balletto wireless MCUs target sophisticated AI/ML functions, such as speech recognition, adaptive noise cancellation, vocal targeting and beam forming in true wireless stereo earbuds, and sensor fusion in lifestyle wristbands and other space-constrained devices.

Key features include high-performance audio and sensor functionality based on DSP and AI/ML capabilities and a large on-die memory. An Arm Cortex-M55 core, with an EEMBC CoreMark score of 704 at 160 MHz, includes the Arm Helium M-profile vector extension, yielding a 500% improvement in DSP performance, according to the company.

The Balletto MCUs also integrate an Arm Ethos-U55 NPU, performing up to 46 GOPS, backed by a large 2-MB tightly coupled memory, for extended audio capabilities with hardware acceleration of AI/ML models. Neural network processing is up to 15× better than a Cortex-M4 processor can achieve, Alif said.

Other features include digital interfaces like I³C, USB-HS (480 Mbits/s) and twin CAN-FD channels; an analog front end with high-precision ADCs, including a 24-bit sigma-delta DAC; up to 2 MB of high-speed MRAM and 2 MB of zero-wait state SRAM; an Octal SPI interface; and up to 77 GPIOs, including I/O for audio input/output.

The Balletto family leverages the company’s advanced security platform with its own dedicated processor and memory, providing hardware root of trust and chain of custody. The MCU family is housed in an ultra-small wafer-level chip-scale package (WLCSP).

Bosch Sensortec, BMA530 and BMA580 MEMS accelerometers

Bosch Sensortec claims the industry’s smallest MEMS accelerometers with the introduction of the BMA530 and BMA580 MEMS accelerometers for wearable and hearable devices. Both devices measure 1.2 × 0.9 × 0.55 mm and include built-in features that make them easy to design into portable products. The BMA530 tracks activities with its step counter, targeting wearables, while the BMA580 is suited for hearables with voice activity detection.

Compared with Bosch’s current-generation accelerometer (BMA253), the BMA530 and BMA580 are housed in a 76% smaller footprint, with a reduction in height from 0.95 mm to 0.55 mm, thanks to their WLCSP packaging. These acceleration sensors offer multiple power modes and automatically switch from low-power to high-power modes to extend battery life.

The BMA530 acceleration sensor includes integrated functions like step counter and generic interrupts. Key specs include measurement ranges of ±2, ±4, ±8 and ±16 g; 1% sensitivity error; and 16-bit digital resolution. In addition to wearables, the BMA530 can be used for gesture recognition in toys, fall detection in laptops and other devices, and power management functions.

The BMA580 features voice activity detection, an advanced feature set for hearable devices and low power consumption. It enables user interaction with a hearable in response to the user tapping the device, for example, to answer or end a call. The software includes an algorithm that can distinguish between single, double and triple taps. Key specs include 16-bit digital resolution, 0.5% sensitivity error and measurement ranges of ±2, ±4, ±8 and ±16 g.

This acceleration sensor also uses bone conduction to detect the vibration of the user’s voice and then wakes the microphone from sleep. This delivers power savings compared with typical microphones in hearables that consume a lot of power because they need to be always-on to listen for voice activity, and it is the only sensor that combines bone conduction in this size, the company said.

GLF Integrated Power, GLF73610 battery-protection ICs

GLF Integrated Power’s IQSmart battery-protection IC family offers a small form factor, low power consumption and cost savings for wearable, hearable and IoT devices. Claimed as one of the smallest lithium-ion/-polymer battery-protection ICs, the GLF73610 product family is housed in a tiny, 0.97 × 0.97 × 0.55-mm CSP. These devices support 4.2-, 4.35-, 4.4- and 4.45-V lithium battery cells used in a variety of thin battery-powered devices, including Bluetooth LE wireless earphones, hearing aids and wearables.

GLF’s advanced IQSmart load switch topology, requiring minimal external components, virtually eliminates deep discharge of the battery during product shipment, storage and standby, the company said. A shipping-mode pin maintains capacity and provides discharge protection for devices with non-removable, pre-charged batteries during the delivery period.

The battery-protection IC’s “Zero Volt” feature allows even a deeply discharged battery to be recharged, and a self-lock feature prevents battery leakage during manufacturing, GLF said.

How it works: If a battery is charged past the overvoltage detection level, the battery-protection IC’s charging switch opens at a preset delay time. The short-circuit delay time prevents false triggers that might open the switch, the company said. When voltage decreases below the set overdischarge level, the switch cuts off the power rail, resulting in ultra-low battery consumption. When the load current reaches the short-circuit protection level, the IC switches and remains off to protect the system from damage.

Infineon Technologies AG, CYW43022 Wi-Fi + Bluetooth SoC

Infineon Technologies AG’s latest AIROC device, the ultra-low-power CYW43022 dual-band Wi-Fi 5 and Bluetooth combo chip, targets lower power consumption and size savings. Thanks to the ultra-low-power architecture, the CYW43022 Wi-Fi + Bluetooth combo offers up to a 65% reduction in power usage during deep sleep, extending battery life in applications like smart wearables and smart-home products using batteries. The Wi-Fi + Bluetooth combo chip is available in a 106-ball WLBGA (3.76 × 4.43 mm; 0.35-mm pitch) or 251-pin WLCSP (3.76 × 4.43 mm, 0.2-mm pitch).

The AIROC CYW43022 includes Wi-Fi network offloads and an embedded Bluetooth stack, which reduces power demands on host processors, Infineon said. It also features a Class 1 Bluetooth PA with 18-dBm transmit power to support designs with smaller antennas or designs that need a longer reach. For protection against hackers, the combo chip provides secure boot with firmware image authentication that requires signed Infineon firmware.

The AIROC wireless product family, including Wi-Fi, Bluetooth, Bluetooth LE and Wi-Fi and Bluetooth combos, uses a common software framework across Android, Linux and RTOS platforms and is pre-integrated with Infineon’s ModusToolbox software and tools. This allows designers to develop differentiated products that meet time-to-market and budget requirements, the company said.

Micron Technology Inc., UFS 4.0 mobile storage solution

Micron is sampling an enhanced version of its Universal Flash Storage (UFS) 4.0 mobile storage solution in the industry’s smallest UFS package, measuring 9 × 13 mm. The mobile storage chip features Micron’s new proprietary firmware upgrades that focus on delivering optimized and higher performance for flagship smartphones.

Aimed at AI-driven smartphones, the UFS 4.0 solution, built on Micron’s advanced 232-layer 3D NAND, offers up to 1-TB capacity and up to 4,300-MB/s sequential read and 4,000-MB/s sequential write speed, which is twice the performance of previous generations. The new solution also delivers improved data-intensive experiences in generative AI applications, with large language models loading 40% faster when initializing conversations with AI digital companions.

The smaller footprint delivers space savings for foldable and ultra-slim smartphones or more space for larger batteries. In addition, the solution offers a 25% increase in power efficiency to extend battery life, even when running power-hungry features like AI, augmented reality, gaming and multimedia apps.

The new mobile flash storage delivers custom firmware updates, designed to significantly improve smartphone performance. Key features include High-Performance Mode (HPM), One Button Refresh (OBR) and Zoned UFS (ZUFS).

HPM is a proprietary feature that optimizes performance by prioritizing critical tasks over background tasks during intensive smartphone use, delivering over 25% improvement in speed, Micron said, while OBR automatically cleans and optimizes data, so smartphones continue operating in a “like-new state,” with 10% faster app launches.

With ZUFS, Micron UFS 4.0 allows the host to specify different zones where the data is stored. It “reduces write amplification to maximize the finite cycles of data that devices can program and erase without degrading device performance,” Micron said, “ultimately extending the smartphone lifespan while keeping devices feeling like new for longer.”

Nexperia, dual 500-mA RETs

Nexperia’s new series of 500-mA dual resistor-equipped transistors (RETs) offers double the space savings with integrated bipolar junction transistors (BJTs) and resistors in an ultra-compact DFN2020(D)-6 packaging, delivering significant board space savings. These devices are designed for load switching in wearables and smartphones as well as for use in digital circuits with higher power requirements, such as space-constrained computing, communications, industrial and automotive applications.

To reduce component count and simplify board design, the 12 new RETs combine dual BJTs with bias resistors integrated in the same package. They also include a second integrated resistor parallel to the base-emitter path to create a voltage divider that sets the base voltage, which offers finer tuning and better turn-off characteristic behavior, the company said.

Because these internal resistors have higher tolerances than external resistors, RETs are suitable for switching applications in which transistors operate in an on or off state, helping to overcome the temperature dependence of standard BJTs, the company added.

RET devices in this series are available in a choice of dual NPN/NPN, NPN/PNP and PNP/PNP options. Unlike similar competing devices, Nexperia’s RETs fully deliver their specified 500-mA output current in a tiny DFN2020(D)-6 package measuring only 2 × 2 × 0.65 mm, Nexperia said.

Nordic Semiconductor, nPM1300 PMIC

Nordic Semiconductor’s nPM1300 is a highly integrated PMIC. It integrates all necessary functions in a single package, comprised of two ultra-efficient buck converters, two load switches/low dropout voltage converters and integrated battery charging.

Targeting battery-operated applications, such as advanced wearables and portable medical devices, the PMIC reduces the bill of materials by combining circuitry that typically requires five or more discrete components into a single chip, Nordic said.

The PMIC claims unique system management features and accurate fuel gauging for low-power wireless applications. This includes an algorithm-based fuel-gauge functionality that uses voltage, current and temperature monitoring for higher accuracy than voltage-based fuel gauges and maintains the PMIC’s own power consumption significantly lower than coulomb-counter-type fuel gauges. Nordic said it is the first company to achieve this balance between high accuracy and low power consumption.

The nPM1300 operates from either a 4.0- to 5.5-V external power supply or a battery voltage down to 2.4 V. It features four controllable power rails.

The PMIC charges single-cell Li-ion, Li-polymer and LiFePO₄ batteries with a linear charging module that supports up to 800-mA charge current and programmable 3.5- to 4.45-V termination voltage. The battery charger features automatic thermal regulation with programmable maximum chip temperature during charging.

Other features include three LED drivers; five GPIOs; USB port detection with automatic current limits of 100 mA or 500 mA through standard USB or up to 1,500 mA through USB-C; dynamic power path management, which automatically switches to battery power if mains power is removed; and ultra-low-current ship and hibernate mode with a programmable wake-up timer.

The PMIC also features an I²C-compatible two-wire interface for configuration of advanced system management functions, including integrated hard reset functionality from one or two buttons, battery fuel gauging, system-level watchdog, power-loss warning and recovery from failed boot.

Nordic also offers the nPM1300 evaluation kit and nPM PowerUP PC app to make it easier to evaluate, configure and implement the nPM1300 without having to write any code.

Onsemi, Hyperlux LP image sensors

Onsemi’s Hyperlux LP image sensor family, comprised of three products, targets a range of applications from AR/VR/XR headsets, webcams and bodycams to smart doorbells, security cameras and machine vision. It features a stacked architecture design that minimizes size, with the smallest device nearly the size of a grain of rice. The new devices include the 20-MP AR2020, the 8-MP AR0830 and the 5-MP AR0544.

The 1.4-µm pixel sensors can extend battery life by up to 40% with low power consumption and wake-on-motion features, and they deliver high image quality even in challenging lighting conditions.

The Hyperlux LP family offers several features and proprietary technologies that improve performance and resolution. These include wake-on-motion that enables sensors to operate in low-power mode until it detects movement, a smart region of interest that provides a context view of the scene at reduced bandwidth and a separate region of interest in the original detail, and near-infrared performance for higher image quality.

Onsemi also said the Hyperlux LP family’s low power consumption reduces thermal noise that can negatively impact image quality and eliminates the need for heat sinks.

Prophesee SA, GenX320 Metavision sensor

The EP Product of the Year Award winner for sensors in 2023, Prophesee SA’s latest event-based Metavision sensor, the GenX320, delivers ultra-low power, low latency and high flexibility for applications like AR/VR headsets, wearables, security and monitoring systems, touch-free interfaces and always-on IoT. Claiming the smallest, most power-efficient and the first event-based vision sensor for consumer edge AI vision devices, the fifth-generation Metavision sensor, available in a tiny, 3 × 4-mm die size, has been designed for event sensing in energy-, compute- and size-constrained embedded-at-the-edge vision systems.

The 320 × 320 6.3-μm-pixel BSI stacked-event-based vision sensor offers a small, 1/5-inch optical format. It is AI-ready, with an on-chip histogram output compatible with multiple AI accelerators. MIPI or CPI data output interfaces offer low-latency connectivity to embedded processing platforms, including low-power MCUs and neuromorphic processor architectures.

Key features include low-latency microsecond-resolution timestamping of events with flexible data formatting and on-chip intelligent power management modes that reduce power consumption to as low as 36 µW and enable smart wake-on-events. It also offers deep-sleep and standby modes.

Other features include easy interfacing with standard SoCs with multiple integrated event data pre-processing, filtering and formatting functions to minimize external processing overhead and native compatibility with Prophesee Metavision Intelligence, a free event-based vision software suite. An evaluation kit is available, along with a range of adapter kits that provide connectivity to a variety of embedded platforms.

Silicon Labs, xG27 Bluetooth SoCs

Silicon Labs’ xG27 family of Bluetooth SoCs are designed for the smallest-form-factor IoT devices. The family is comprised of the BG27 for Bluetooth connectivity and the MG27 for Zigbee and other proprietary protocols, in packages as small as a 2.3 × 2.6 mm WLCSP. This makes them suitable for applications like medical patches, continuous-glucose monitors, wearable electrocardiograms and asset tags. They also deliver energy efficiency, high performance and trusted security.

The xG27 SoC family, built on the Arm Cortex M33 processor, is suited for tiny, battery-optimized devices. Shared features include an integrated DC/DC boost that allows devices to operate on batteries as low as 0.8 V, resulting in smaller device sizes and cost; an integrated coulomb counter for battery-level monitoring; and advanced security, including Silicon Labs’ Secure Vault with Virtual Security Engine for secure boot and debug hardened against glitch attacks and tamper protection.

The wireless SoCs also offer shelf mode that reduces energy use to less than 20 nA so that devices can be stocked on the shelves and still maintain nearly full battery life.

Silicon Labs said the xG27 SoCs extend the range of possible applications and use cases. Lura Health, for example, is using the BG27 SoC to develop a sensor that is smaller than a tooth to collect salivary data, which is used to test for over 1,000 health conditions. The health monitor is so small that it can be glued to a tooth.