Silicon Labs crafts wireless MCUs for harvested power

Update: April 23, 2024 Tags:25080aecoelicltmcutransmitter

“As Silicon Labs’ most energy-efficient SoCs to date, all three will enable IoT device makers to build wireless devices for battery-optimised and battery-free devices that can harvest energy from external sources in their environments like indoor or outdoor ambient light, ambient radio waves, and kinetic motion,” said the company.

Each has an Arm Cortex-M33 application processor, and a 2.4GHz radio with its own Cortex-M0+ processor.

Features include: 150μJ 8ms wake-up. Sleep options include RFSense, GPIO-wale and real-time clock wake from the deepest (EM4) sleep mode.

The three family members are:

  • EFR32BG22E for Bluetooth LE and Bluetooth Mesh with a 76.8MHz Cortex-M33, up to 512kbyte of flash and 32kbyte of ram, as well as Bluetooth 5 angle-of-arrival, angle-of-departure and sub-metre location.
  • EFR32MG22E for Zigbee also gets a 76.8 MHz Cortex-M33
  • EFR32FG22E for proprietary protocols has a half-speed (38.4MHz) Cortex-M33

Each is available with up to 512kbyte of flash and 32kbyte of ram and a 6dBm 2.4GHz transmitter.

Security comes from hardware crypto acceleration (AES128/256, SHA-1, SHA-2 (up to 256-bit), ECC (up to 256-bit), ECDSA and ECDH), a true random number generator (NIST SP800-90 and AIS-31 compliant) and Arm’s TrustZone.

To this, FG parts add secure boot with a root of trust and secure loader, and secure debug with lock/unlock.

Focussing on the FG22E (proprietary radio version), the company claims consumption of 3.6mA 1Mbit/s GFSK Rx, 3.9mA 250kbit/s O-QPSK DSSS Rx, 4.1mA dBm Tx, 8.2mA 6dBm Tx, 26μA/MHz active (38.4MHz EM0 mode), 1.2μA deep sleep (EM2, 8kbyte ram retention, RTC (32.768kHz RC oscillator) or 170nA EM4 ‘shut-off’.

For all the devices, operation is from a single 1.71 to 3.8V rail and over -40 to +85°C.

Packaging is 5 x 5mm QFN40 or 4 x 4mm QFN32.

Teaming up with harvesting IC company e-peas, SiLabs has developed two energy harvesting extension boards for the xG22E Explorer Kit, each tuned for different energy sources and energy storage. One uses e-peas AEM13920 dual-harvester for twin energy sources, the other is based on AEM00300 and collects power from random pulsed energy sources.

Find the FG22E family here