Coral Dev Board Micro - development board with NXP i.MX RT117, Edge TPU ML and Himax HM01B0 camera

Coral Dev Board Micro - board equipped with the NXP i.MX RT117 microcontroller (ARM Cortex-M7 @ 800 MHz and ARM Cortex-M4 @ 400 MHz), which offers support for TensorFlow Lite for Microcontrollers. The board also features the Edge TPU ML chip (designed by Google) for massive ML inference capabilities on Linux, Windows and macOS. It accelerates TensorFlow Lite models in a very energy-efficient way, thus enabling the implementation of low-power system-based applications that require fast in-device inference for vision and audio ML applications. It has the ability to perform as many as 4 trillion operations in just one second. This on-device ML processing reduces latency, increases data privacy, and removes the need for a constant internet connection. In addition, Coral Dev Board Micro is equipped with a Himax HM01B0 camera and a microphone. Its structure and available connectors allow you to add other peripheral devices to the project.

Coral Dev Board Micro - development board with NXP i.MX RT117, Edge TPU ML and Himax HM01B0 camera.

Possibility to program Arduino, FreeRTOS.

• NXP i.MX RT1176 microcontroller with support for TensorFlow Lite for Microcontrollers
  • ARM Cortex-M7 @ 800MHz
  • ARM Cortex-M4 @ 400MHz
• Edge TPU ML chip
  • peak performance 4 TOPS (int8)
  • provides high-quality ML inference
• 128 MiB NAND Flash memory
• 64MB of SDRAM
• Possibility to program Arduino, FreeRTOS
• Built-in Himax HM01B0 camera
  • Native resolution equal to 324 x 324 px or QVGA 320 x 240 px
• Built-in microphone
• 24 GPIO pins
• Additional connectors for connecting peripheral devices of the Coral Wireless series

1. Himax HM01B0 camera
2. Camera LED (indicates camera operation)
3. PDM microphone
4. Edge TPU LED (indicates Edge TPU operation)
5. User programmable button (also toggles power with 7 second hold)
6. LED - for use by the user
7. GPIO pins - digital, analog and power pins (right side)
8. GPIO pins - digital, analog and power pins (left side)
9. Status LED (indicates board operating status)
10. Reset button
11. USB Type-C connector
12. Additional connectors

Arrangement of components on the board.

NXP i.MX RT117:

• ARM Cortex-M7 @ 800MHz
  • 800Mhz clock speed
  • 32 kB L1 instruction cache
  • 32 kB L1 data cache
  • Single and double precision FPU, support for Armv7-M FPv5 architecture
  • TensorFlow Lite for Microcontrollers support
• ARM Cortex-M4 @ 400MHz
  • 400Mhz clock speed
  • 16 kB instruction cache
  • 16 kB data cache
  • Single precision FPU, support for Armv7-M FPv4-SP architecture
  • TensorFlow Lite for Microcontrollers support
• Memory
  • 2MB SRAM:
    • 512 kB of TCM memory - Cortex-M7
    • 256 kB of TCM memory - Cortex-M4
    • 1.25MB OCRAM
  • 4 kB of always-on RAM
  • 256KB boot ROM

• Authentication of cryptographic devices
• Secure sharing and storage of credentials

• Google-designed layout for high-performance machine learning inference for TensorFlow Lite models
• It can perform up to 4 trillion operations per second
• USB 2.0 interface with MCU RT1176

• Himax HM01B0 CMOS sensor
• Native resolution 324 x 324 px or QVGA resolution 320 x 240 px
• FOV 110º viewing angle
• Aperture ƒ/2.0
• Fixed Focus

• Built-in PDM microphone

• Random Access Memory (SDRAM): 64 MB
• Flash Memory (NAND): 128MB

• 2x12 pin layout
• Interfaces GPIO, PWM, UART, I2C, SPI, DAC, ADC, GND, 3V3, 1V8, VSYS

• 2x 100-pin connectors (Hirose DF40C-100DP-0.4V(51))
  • Provides various inputs/outputs from the MCU
• Supports Coral Wireless add-on board (sold separately) Support Coral PoE add-on board (sold separately)

• USB type C connector - power and data transfer

• Built-in 4 LEDs
  • Green LED to indicate camera activity
  • User programmable green LED
  • Orange LED indicating the card's operating status
  • White LED indicating the status of the Edge TPU

• Built-in 2 buttons
  • MCU reset button
  • User programmable button (for short presses); also responsible for the power supply - powers on the card when pressed for a long time

• User Guide
• Technical Documentation