Trinkey QT2040 - USB KEY RP2040 with StemmaQt / Qwiic conn. - I2C

Trinkey QT2040 : a RP2040 microcontroler on a USB Stick with Qwiic / StemmaQT connector

It's half USB Key, half Adafruit QT Py, and a lotta RP2040...

The Trinkey QT2040 is a half size USB Key fitted with the RP2040 microcontroler (same as Pico) with Stemma QT connector. The StemmaQT connector is also known as Qwiic (SparkFun). This connector ship an I2C bus together with power. Making electronic with microcontroler is really fun but sometime we just need to create a dead simple solution. The TrinKey is great to create plug and play solution Plugs it right into your computer's USB port and you are ready to program it OR to use it!

The PCB slip into any USB A port and fitted on RP2040 microcontroller with the required circuitry to make it running. The PCB also fit a RGB NeoPixel, a reset and bootloader button. The bootloader can also be used as user button. Finally, the PCB also brings a STEMMA QT/QWIIC port (for the I2C bus)

The board size is 25.4mm x 17.8mm as many of the QT Sensor boards. The TrinKey also have four M2.5 mounting holes for standoffs and screws. The QT connector at the end allows to use a QT cable to quickly wire QT compatible sensor board.

The TrinKey board comes with 8 MB QSPI Flash memory to store the firmware and filesystems!

Plug-and-play STEMMA QT

The STEMMA QT / Qwiic connector is the king of this board! It is a chainable I2C port that can be used with I2C components or StemmaQT sensors board and accessories. Thanks to this connector, no needs to do any soldering to get started.

The QT port allows to connect OLEDs, Inertial Measurment Units, air quality sensors and plenty of other board compatible with Qwiic (SparkFun) / Stemma Qt (Adafruit) boards. I2C is plug-and-play and chainable by design. Just plug in a compatible cable and attach it to your TrinKey and you’re ready to load up some software and measure some light.

Use any Seeed Grove I2C boards they will also work with this adapter cable.

Software Support

For use the Trinkey with Arduino IDE, check out the Philhower core. There is also great C/C++ Pico SDK from the foundation, an MicroPython and a CircuitPython port!
Adafruit recommend to use CircuitPython, supported out of the box so you can follow along with Adafruit's CircuitPython projects and tutorials.

The RP2040 has 264 KB of RAM (kind a lot) but doesn't integrates built-in FLASH memory. The RP2040 uses external QSPI flash chip and Adafruit fitted the board with 8MB Flash. Under CircuitPython or MicroPython, the Flash is used to store program and the file-system. When using C/C++ the whole flash memory is available for your program. When using Python script, about 7 MB of Flash is used for user script, files, images, fonts, etc.

Technical details

• Main body with four M2.5 mounting holes (25.4 x 15mm)
• USB Type A connector (extra-thick PCB to fit into a USB host port)
• RP2040 32-bit Cortex M0+ dual-core running at ~125 MHz @ 3.3V logic and power
• 264 KB RAM
• 8 MB SPI FLASH chip. No EEPROM
• Native USB supported by every OS (allow USB serial console, MIDI, Keyboard/Mouse HID, even a little disk drive).
• Can be used with Arduino, MicroPython, CircuitPython, C/C++
• Built-in RGB NeoPixel LED
• I2C connector (Qwiic / Stemma Qt)
• 3.3V regulator with 600mA peak output
• 12 MHz crystal
• Both Reset button and Bootloader button (allow quick restarts)
• Bootloader button can be used in 'user' code
• QT2040 Trinkey PinOut

About the RP2040

The RP2040 ROM contains an USB UF2 bootloader. This allows user to send a new firmware into the RP2040 Flash. To do so, the user press and hold down the BOOT button while plugging it into USB (or pressing and releasing the Reset button). When UF2 Bootloader get activen the microcontroler appears as an USB disk drive (and you can drag/drop the new firmware onto it).

The RP2040 is a powerful chip, which have two cores equivalent to SAMD M0 Cortex (SAMD21). RP2040 is an M0 chip, so it does not have a floating point unit nor DSP support which impact heavy floating point math calculation. On M0 core, the floatting calculation are performed by software (which is significantly slower than hardware implementation).