USB Accelerator datasheet


  • Google Edge TPU ML accelerator coprocessor
  • USB 3.0 (USB 3.1 Gen 1) Type-C socket
  • Supports Linux, Mac, and Windows on host CPU


The Coral USB Accelerator adds an Edge TPU coprocessor to your system. It includes a USB socket you can connect to a host computer to perform accelerated ML inferencing.

The on-board Edge TPU is a small ASIC designed by Google that provides high performance ML inferencing with a low power cost. For example, it can execute state-of-the-art mobile vision models such as MobileNet v2 at almost 400 FPS, in a power-efficient manner.

System requirements

  • A computer with one of the following operating systems:
    • Linux Debian 10, or a derivative thereof (such as Ubuntu 18.04), and a system architecture of either x86-64, Armv7 (32-bit), or Armv8 (64-bit) (Raspberry Pi is supported, but we have only tested Raspberry Pi 3 Model B+ and Raspberry Pi 4)
    • macOS 10.15, with either MacPorts or Homebrew installed
    • Windows 10
  • One available USB port (for the best performance, use a USB 3.0 port)
  • Python 3.5, 3.6, or 3.7

Component overview

The Coral USB Accelerator is a small PCB that provides an Edge TPU coprocessor for your host CPU system.

Coral USB Accelerator components and features
Feature Details
ML accelerator
Edge TPU
  • ASIC designed by Google that provides high performance ML inferencing for TensorFlow Lite models
  • USB 3.0 (USB 3.1 Gen 1) port and cable (SuperSpeed, 5 Gbps)
  • Included cable is USB Type-C to Type-A, and 300 mm (12 in) in length

Product dimensions

Figure 1.USB Accelerator dimensions and mounting hole details

USB cable length: 300 mm ± 20 mm

Software and operation

The USB Accelerator does not require any software to be installed, but the host computer must have the Edge TPU runtime and API library. Instructions on how to install these are included in the guide for how to get started with the USB Accelerator.

For information about how to create models and run inferences on the Edge TPU, read the documentation for TensorFlow models on the Edge TPU.

Performance settings

When you first set up the USB Accelerator, you can select whether the device operates at the maximum clock frequency or the reduced clock frequency. The maximum frequency is twice the reduced setting, which increases the inferencing speed but also increases power consumption.

Caution: If you set up the device to operate at the maximum clock frequency, the USB Accelerator can become very hot to the touch during operation. This might cause burn injuries. To avoid injury, either keep the device out of reach when operating it at maximum frequency, or use the reduced clock frequency.

To change the clock frequency at which the device operates, simply install the alternative runtime, as described in the instructions for how to install the Edge TPU runtime.

LED behavior

The USB Accelerator includes an LED that indicates the current system state. The following table describes the behavior of the LED.

USB Accelerator LED behavior and meaning
System state LED behavior
Edge TPU initialized On (solid)
Edge TPU running Pulse (breathe)

Power specifications

The USB Accelerator is powered by 5 V from the USB interface through on-board power management chip (PMIC).

The device should be connected to a host computer with a USB cable that provides at least 500 mA at 5 V. Please note that when the device runs at maximum speed, it can draw a peak current of 900 mA.

Depending on the computation workloads and operating frequency, the USB Accelerator may heat up during normal operation. Contact with the metal part of the device may lead to discomfort and/or skin burns if the device has been operating for an extended period. We recommend the following operating temperature ranges, depending on the operating frequency you have selected.

USB Accelerator operating conditions
Operating frequency Maximum ambient temperature
Reduced 35° C
Maximum 25° C

Google does not accept any responsibility for loss or damage if the device is handled unsafely or operated outside of the recommended ambient temperature range.

Environmental and electrical reliability tests

Verified results for USB Accelerator environmental and electrical reliability tests
Test Conditions Verified
Heat soak Non-op: 60° C @ 90% RH 72 hrs
Temp cycling Non-op: -20° C @ 60° C 300 cycles
Short term survivability Non-op: 85° C @ 20% RH, 18 hrs 18 hours
Long-term operational stress (LTOS) Op: 40° C @ 90% RH 1000 hours
ESD - Air discharge 15~35° C, 30~60% RH 12 kV
ESD - Contact discharge 15~35° C, 30~60% RH 8 kV

Mechanical reliability tests

Verified results for USB Accelerator mechanical reliability tests
Test Conditions Verified
0.5 m controlled drop on wood 6 faces, 12 edges, 3/4" wood 18 drops
0.5 m tumble Impact surface: steel plate > 3 mm thick 50 drops
Repetitive drop 75 mm drop on plywood, six faces 50 drops
Vibration 3 axes (X, Y and Z), 15 minutes per axis, 10-500 Hz. Amplitude: 2.16 Grms 45 minutes
USB-C connector cycling Manual plug/ unplug 1000 cycles
USB-C connector side load 2.5 kg 2.5 kgf load, 15 mm from the plug, 2 orientations 50 cycles
Pressure & squeeze - 50 kg 50 kgf, 100 mm rubber pad 10 cycles
Pressure & squeeze - 5 kg 5 kgf, 10 mm rubber pad 10 cycles
USB cable 120° bend 60° CW + 60° CCW, 453 gf tension, 42 rpm 100 cycles


USB Accelerator certifications
Country Agency
European Union CE
Hong Kong CE
Japan VCCI
Korea KC

CAD files

USB Accelerator 3D CAD file
Coral-USB-Accelerator.STEP 3D CAD file in STEP format

Document revisions

History of changes to this document
Version Changes
1.3 (April 2020) Updated system architecture requirements (supports both ARMv7 and v8)
1.2 (January 2020) Updated system requirements for Mac and Windows support
1.1 (October 2019) Clarified power requirements
1.0 (September 2019) Updated information about performance settings
Beta (March 2019) Initial release