Ultra Low Power (ULP) MCUs and IoT Security Solutions for Smart Homes

Artificial intelligence (AI) requires extreme computational horsepower, but Maxim is cutting the power cord from AI insights. The MAX78000 is a new breed of AI microcontroller built to enable neural networks to execute at ultra-low power and live at the edge of the IoT. This product combines the most energy-efficient AI processing with Maxim's proven ultra-low power microcontrollers. Our hardware-based convolutional neural network (CNN) accelerator enables battery-powered applications to execute AI inferences while spending only microjoules of energy.

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). The CNN weight memory is SRAM-based, so AI network updates can be made on the fly. The CNN engine also has 512KB of data memory. The CNN architecture is highly flexible, allowing networks to be trained in conventional toolsets like PyTorch^® and TensorFlow^®, then converted for execution on the MAX78000 using tools provided by Maxim.

In addition to the memory in the CNN engine, the MAX78000 has large on-chip system memory for the microcontroller core, with 512KB flash and up to 128KB SRAM. Multiple high-speed and low-power communications interfaces are supported, including I²S and a parallel camera interface (PCIF).

The device is available in a 81-pin CTBGA (8mm x 8mm, 0.8mm pitch) package.

Applications

• Audio Processing: Multi-Keyword Recognition, Sound Classification, Noise Cancellation
• Facial Recognition
• Object Detection and Classification
• Time-Series Data Processing: Heart Rate/Health Signal Analysis, Multi-Sensor Analysis, Predictive Maintenance

The MAX32655 microcontroller (MCU) is an advanced system-on-chip (SoC) featuring an Arm^® Cortex^®-M4F CPU for efficient computation of complex functions and algorithms that is qualified to operate at a temperature range of -40°C to +105°C. The SoC integrates power regulation and management with a single inductor multiple-output (SIMO) buck regulator system. The latest generation Bluetooth^® 5.2 Low Energy (LE) radio is on board, supporting long-range (coded) and high-throughput modes and medical body area network (MBAN).

The device offers large onboard memory with 512KB flash and 128KB SRAM, with optional error correction coding (ECC) on one 32KB SRAM bank. This 32KB bank can be optionally retained in BACKUP mode. An 8KB user OTP area is available.

The MAX32655 supports multiple high-speed peripherals, such as I²C, 50MHz SPI, and UART, plus one I²S port for connecting to an audio codec. An eight-input, 10-bit ADC is available to monitor analog input from external analog sources. In addition, a low-power UART (LPUART) is available for operation in the lowest power sleep modes to facilitate wake-up activity without any data loss. A total of six timers with I/O capability are provided, including two low-power timers to enable pulse counting, capture/compare, and pulse-width modulation (PWM) generation, even in the lowest power sleep modes.

The MAX32655 is available in two different packages:

• 81 CTBGA package (8mm × 8mm, 0.8mm pitch).
• 51 WLP (3.09mm × 3.09mm, 0.35mm pitch)

Applications

• Asset Tracking
• Fitness/Health and Medical Wearables
• Hearables
• Industrial Sensors
• Wireless Computer Peripherals and I/O Devices

DeepCover® embedded security solutions cloak sensitive data under multiple layers of advanced physical security to provide the most secure key storage possible.

The DeepCover secure microcontroller MAX32520 provides an interoperable, secure, and cost-effective solution to build new generations of trusted embedded systems and communication devices such as IoT, IoT gateways, and wireless access points.

The MAX32520 incorporates Maxim's patented ChipDNA™ PUF technology. ChipDNA technology involves a physically unclonable function (PUF) that enables cost-effective protection against invasive physical attacks. Using the random variation of semiconductor device characteristics that naturally occur during wafer fabrication, the ChipDNA circuit generates a unique output value that is repeatable over time, temperature, and operating voltage. Attempts to probe or observe ChipDNA operation modifies the underlying circuit characteristics, preventing discovery of the unique value used by the chip cryptographic functions. The MAX32520 utilizes the ChipDNA output as key content to cryptographically secure all device stored data including user firmware. User firmware encryption provides ultimate software IP protection. The ChipDNA can also generate a private key for the ECDSA signing operation.

The MAX32520 integrates an Arm® Cortex® -M4 processor, 2MB of Flash, 136KB of system RAM + 34KB ECC, 8KB of one-time-programmable (OTP) memory and 128KB of boot ROM.

The MAX32520 provides a FIPS/NIST compliant TRNG, environmental and tamper detection circuitry to facilitate system-level security.

Multiple high-speed interfaces are supported including SPI, UART, and an I²C. The four on-chip timers also support PWM output generation for direct control of external devices. One of the SPI ports has a serial flash emulation mode allowing direct code fetching enabling secure boot for a host microcontroller.

Applications

• Embedded Communication Equipment (Routers, Gateways, etc.)
• Embedded Connected Systems
• IoT Nodes and Gateways
• Secure Industrial Appliances, Sensors, and Controllers
• Set-Top Boxes

The MAX78000FTHR is a rapid development platform to help engineers quickly implement ultra low-power, artificial intelligence (AI) solutions using the MAX78000 Arm^® Cortex^® M4F processor with an integrated Convolutional Neural Network accelerator. The board also includes the MAX20303 PMIC for battery and power management. The form factor is 0.9in x 2.6in dual-row header footprint that is compatible with Adafruit Feather Wing peripheral expansion boards. The board includes a variety of peripherals, such as a CMOS VGA image sensor, digital microphone, low-power stereo audio CODEC, 1MB QSPI SRAM, micro SD card connector, RGB indicator LED, and pushbutton. The MAX78000FTHR provides a power-optimized flexible platform for quick proof-of-concepts and early software development to enhance time to market.

Go to https://www.maximintegrated.com/en/products/MAX78000FTHR to get started developing with this board.

Applications

• Audio Processing: Multi-Keyword Recognition, Sound Classification, Noise Cancellation
• Facial Recognition
• Object Detection and Classification
• Time-Series Data Processing: Heart Rate/Health Signal Analysis, Multi-Sensor Analysis, Predictive Maintenance

The MAX32655FTHR is a rapid development platform to help engineers quickly implement ultra low-power wireless solutions using MAX32655 Arm^© Cortex^®-M4F and Bluetooth^® 5.2 Low Energy (LE). The board also includes the MAX20303 PMIC for battery and power management. The form factor is a small 0.9in x 2.6in dual-row header footprint that is compatible with Adafruit Feather Wing peripheral expansion boards. The board includes a variety of peripherals, such as a digital microphone, low power stereo audio CODEC, 128MB QSPI Flash, micro SD card connector, RGB indicator LED, and pushbutton. The MAX32655FTHR provides a power-optimized flexible platform for quick proof-of-concepts and early software development to enhance time to market.

Applications

• Asset Tracking
• Fitness/Health and Medical Wearables
• Hearables
• Industrial Sensors
• Wireless Computer Peripherals and I/O Devices

The MAX32666FTHR board is a rapid development platform to help engineers quickly implement battery optimized Bluetooth® 5 solutions with the MAX32666 Arm^® Cortex^®-M4 processor with FPU. The board also includes the MAX1555 1-Cell Li+ battery charger for battery management. The form factor is a small 0.9in by 2.0in dual-row header footprint that is compatible with breadboards and off-the-shelf peripheral expansion boards. The board also includes a variety of peripherals, such as a micro SD card connector, 6-axis accelerometer/gyro, RGB indicator LED, and pushbutton. This platform provides power-optimized flexible for quick proof-of-concepts and early software development to enhance time to market.

Applications

• Connected Home
• Gaming Devices
• Hearables
• Industrial Sensors
• Payment/Fitness/Medical Wearables
• Telemedicine

The MAX78000 is an Arm Cortex M4F microcontroller with a Convolutional Neural Network (CNN) accelerator unit. The unique energy-efficient architecture of the MAX78000 enables battery-powered AI for at the edge applications such as face identification and keyword spotting. The MAXREFDES178 is a cube camera reference design based on the MAX78000 and MAX32666 microcontrollers to help AI at the edge device designers to accelerate their proof-of-concept to the market phase.

The MAXREFDES178 hardware consists of a connectivity board based on the MAX32666 and AI board based on two MAX78000 chips. A short 33-position flex cable connects the boards to each other.

The MAXREFDES9001 is a complete internet-of-things (IoT) security reference design featuring a LoRa radio based, low-power, temperature sensor node secured with a DS28S60 secure coprocessor, a LoRa gateway, and a cloud application implemented in AWS infrastructure. This reference design highlights a robust and easy to manage end-to-end security scheme with authentication and confidentiality capabilities independent of the transmission link in use—the LoRaWAN protocol in this case. The MAXREFDES9001 is designed to easily integrate into embedded systems enabling confidentiality, authentication, and integrity of information.

The sensor node is motioned by the tiny, low-power, Cortex-M4-based microcontroller MAX32660 which periodically measures the ambient temperature with the help of the DS7505, authenticates and encrypts the temperature value using AES-GCM with the DS28S60 secure coprocessor, and sends it to the AWS infrastructure over a LoRaWAN network, through a Raspberry Pi-powered gateway. To prevent rogue nodes from publishing data, joining the sensor nodes to the network requires a prior local verification using a convenient NFC-based strong authentication with the help of the MAX66242 Secure Authenticator and a dedicated Android application running on an NFC-enabled Android device.

Once the authentication is successful, proving that the sensor node is genuine, the Android device communicates with the cloud application through the Internet to provision the sensor node; that is, to generate a certificate for the sensor node and perform an AES-GCM key exchange between that sensor node and the AWS infrastructure. The Android device uses the MAX66242 as an NFC bridge to communicate with the sensor node device’s microcontroller application and store the certificate into the DS28S60 coprocessor, and to have the key exchange done between the DS28S60 and the cloud application using the Elliptic Curve Diffie-Hellman (ECDH) protocol. Once the key exchange is completed, the sensor node is ready to send its data to the cloud application using the negotiated AES-GCM key. Further sensor node authentication by the cloud application is possible using ECDSA since the sensor node now has a valid certificate with a matching key pair. Incidentally, the provisioning process also joins the end device to the LoRaWAN network implemented using the AWS IoT core, but this is not the main purpose of the reference design that shows a way to secure data without relying on the security of the various underlying communication links.

The DS2477 is a 1-Wire^® master that performs protocol conversion between the I²C master and any attached 1-Wire slaves. For 1-Wire line driving, the internal user-adjustable timers relieve the system host processor from generating time-critical 1-Wire waveforms, which supports both standard and overdrive communication speeds. The 1-Wire master has a selectable active or passive 1-Wire pullup. The strong pullup features support 1-Wire power delivery for 1-Wire devices that require additional current, such as electrically erasable programmable read-only memory (EEPROM) writes or cryptographic operations.

The DS2477 is also a SHA-3 coprocessor designed to operate with 1-Wire SHA3-256 secure authenticators, such as the DS28E50. The DS28E50 secure authenticator combines FIPS202-compliant secure hash algorithm (SHA-3) challenge and response authentication with Maxim’s patented ChipDNA^™ technology, a physically unclonable function (PUF), to provide a cost-effective solution with the ultimate protection against security attacks. The DS28E50 communicates over the single-contact 1-Wire bus at both standard and overdrive speeds.

The MAXREFDES9008 reference design shows how to develop a secure application using a DS28E50 and DS2477 authentication scheme. Featured is an Arm^® Cortex^®-M4 host microcontroller to be used as the system’s host processor. All hardware design files as well as C-code demonstration software are provided for this reference design in the Design Resources Section.