Your SAMD51 M4 board bootloader may need to be updated to fix an intermittent bug that can erase parts of internal flash.

Updating Your Bootloader

To see if you need to update your bootloader, get the UF2 boot drive to appear as a mounted drive on your computer, in a file browser window. If you're running MakeCode, click the reset button once. If you're running CircuitPython or an Arduino program, double-click the reset button.

When you see the ...BOOT drive (FEATHERBOOT, METROM4BOOT, ITSYM4BOOT, PORTALBOOT, etc.) , click the drive in the file browser window and then double-click the INFO_UF2.TXT file to see what's inside.

The example screenshots below are for a PyGamer. What you see for your board will be largely the same except for the board name and the BOOT drive name.

The bootloader version is listed in INFO_UF2.TXT. In this example, the version is v3.6.0.

If the bootloader version you see is older than v3.9.0, you need to update. For instance, the bootloader above needs to be upgraded.

Download the latest version of the bootloader updater from the circuitpython.org Downloads page for your board.

• The bootloader updater will be named update-bootloader-name_of_your_board-v3.9.0.uf2 or some later version. Drag that file from your Downloads folder onto the BOOT drive:

After you drag the updater onto the boot drive, the red LED on the board will flicker and then blink slowly about five times. A few seconds later, the BOOT will appear in the Finder. After that, you can click on the BOOT drive and double-click INFO_UF2.TXT again to confirm you've updated the bootloader.

To update firmware on your MONSTER M4SK board with the latest eye animation software, start by downloading this .UF2 file:

For a HALLOWING M4 board, use this .UF2 instead:

To install different “looks,” download and unzip this collection of eyeball graphics:

Inside the “eyes” folder are several sub-folders, one for each of our ready-made eye designs. “hazel” is our standard human eye from prior projects…then we’re adding more as Halloween approaches.

Each folder contains a set of .bmp images for that eye, plus a file called config.eye. Copy one of these folders to the CIRCUITPY drive — not the individual files, but the whole folder. Then copy or move the file called config.eye out of the folder and into the drive’s root directory.

Copy these files to the CIRCUITPY drive.

Remember that config.eye goes in the root directory, along with the hazel (or other eye name) folder.

Press the board’s reset button…

There will be a delay of several seconds as the eye code initializes. Then…animation!

If you get a flat-colored white eyeball with blue irises and no eyelids, something’s wrong! Verify that you’ve copied over a whole folder of graphics (not the individual files) and moved or copied that folder’s config.eye to the root directory.

Aside from the stock hazel eyes, some of the alternate designs include:

SECRET FEATURE: You can select among four different eye configurations on startup. Hold down one of the three buttons on top when powering on (or pressing reset) to load files config1.eye (inner button), config2.eye (middle button) or config3.eye (outer button) instead of the default config.eye. All go in the root directory for now (but graphics & such can go in folders).

If one of the ready-made designs does what you need…fantastic, you’re all done! If you want to make changes, or create your own custom eyes, read on…

If our ready-made eye designs don’t meet your needs, you can personalize quite a bit by editing a text file and providing some graphics.

On startup, our eyeball code loads a file called config.eye from the device’s root directory. This is a plain-text file that any simple editor can handle…Notepad, TextEdit, vim and so forth.

If customizing your eyes, maintain a backup of the board’s contents on your computer. There is a small but nonzero chance of a glitch that could require a full reinstall. See the Troubleshooting page for help.

The syntax of this file is a format called JSON. There’s good and bad news…

The good: JSON has a standardized syntax and we can leverage existing code to read it…we’re not starting from scratch with a new file format.

The bad: JSON files are really meant to be read and written by machines, as a way to preserve program state…not edited by humans. It’s phenomenally picky about getting syntax just right and does not fall back gracefully.

The saving grace: if you’re just changing colors and textures, you might not need to edit this file at all! Quite often you can just substitute different graphics files with the same names.

Troubleshooting JSON

If even a single character is out of place in config.eye, the whole thing will fail to load and there’s no helpful indication of where the problem lies.

If this happens, the eyes will run in a default state: blue eyes with no eyelids, and everything is “flat” colors, no textures. This isn’t helpful in isolating the problem but at least tells you there is a problem.

One way to troubleshoot this is to start with a known valid eye configuration, then change one setting at a time and restart the code. If the eyes revert to the default state, the last change contains the syntax error. Quite often it’s just a missing double quote or extra comma.

Another option is to use a JSON validator such as this one at hjson.org. Copy-and-paste your eyeball configuration into the left pane…if there’s a problem, this will be reported on the right.

Here’s an simple config.eye file to start with:

Some things to notice:

• The whole thing is contained within curly braces — { and }
• Each item has a name (always in quotes), a colon separator ( : ) and a value. Values might be numbers, strings (in quotes) or arrays (in square brackets) depending on what’s being configured.
• This is a list and each line ends with a comma…except for the last item in the list. Watch out for missing or surplus commas!
• JSON is case-sensitive. "Foo", "foo" and "FOO" are all different things.
• Using extra spaces to line up columns really isn’t necessary, just something I do to assist with legibility. Some folks just find it annoying.
• Comments (starting with // ) are not standard JSON syntax, but the library we use allows them and I find them very helpful.

Each option will be explained in detail on the “Configurable Settings” page. But first let’s talk about graphics…

So we’re singing from the same page, let’s lay out some eye terminology…

This program’s eye graphics are stored flat and unrolled, like a map projection. The horizontal (X) axis works like the longitude, or angle around the eye, while the vertical (Y) axis is the latitude. The images are wrapped around the pupil in a clockwise direction.

There are two images (or texture maps) associated with the eyes…one for the iris, another for the sclera.

The iris is what we think of as the “color” of the eye and is most often what you’ll want to edit. Sometimes you just need to edit the hue & saturation in a program like Photoshop, or you can make something totally custom if you’re after a particular look.

The sclera is the “white” of the eye…which really isn’t that white at all. There’s veins and blotches and gross stuff!

Image Storage

The texture maps are stored as 24-bit BMP images…nobody’s favorite, but easy for microcontrollers to handle. (This is also sometimes called “Windows Bitmap” format, though plenty of non-Windows software can read and write these images…and not to be confused with “X BitMap” or “Portable Bitmap Format,” different animals.)

Textures are loaded from drive into RAM, downsampled to 16-bit color (what the displays natively use) and then written to the chip’s internal flash memory (different from the flash filesystem). Although the code places no strict limit on image dimensions, RAM and flash are both finite resources, and this limits to how large these textures can be…both individually and in total.

No single image can exceed available RAM, or about 160 kilobytes. The total of all images must fit within flash, or about 360 kilobytes. These figures might change a bit in the future, so try to leave yourself some overhead.

Use the following formula to determine the space needed for an image:

width in pixels × height in pixels × 2 bytes

For example, a 500 × 150 pixel texture would consume 500 × 150 × 2 = 150,000 bytes. This fits in RAM just fine, and takes up a bit less than half of the flash space.

You don’t have to texture-map both the iris and sclera if you don’t want to…on the Configurable Settings page we explain how to use solid colors for either or both. Also shown there…it’s possible to assign independent textures to the left and right eyes. When both eyes are sharing the same texture, the code will place only one instance in flash, saving space.

To make the re-use of textures less obvious, the “seam” where a texture map wraps around is normally at the 12 o’clock position for the right eye and 6 o’clock for the left eye. The Configurable Settings page shows how to change this.

Too-large images may exceed available space, while too-small images may exhibit visible jaggies (unless you’re aiming for that effect as described above). The ideal size can be calculated based on the circumference of the iris or the whole eye…

The iris and overall eye size are configurable (shown on following page)…but for example, let’s assume you’ve got an iris with a 60 pixel radius (120 pixel diameter) and the eyeball has a 125 pixel radius (250 pixel diameter)…both of these are the defaults.

Multiply the iris and/or eye diameters by Pi (3.14) to get the ideal width in pixels for the iris and sclera images.

For example: iris with 120 pixel diameter. 120 px × 3.14 = 377 pixels wide. You can use that, or round up or down a smidge to a round number if you like (e.g. 360 or 380 pixels).

Sclera image width for 250 pixel diameter eye: 250 × 3.14 = 785 pixels wide…but again, OK to round up or down a little…use 800 pixels wide if you like, unless really pressed for space.

The ideal image heights are a bit different. First, although the code can load any size image, it won’t actually benefit above 128 pixels on the vertical axis, it’s just wasted space. For the iris, use its radius (60 pixels in the case described above) or even a little less, since the pupil is always open a bit. For the sclera…try 200 minus the iris radius, keeping in mind the 128 pixel recommended maximum (e.g. with a 60 pixel iris, 140 is our target, then cap it at 128). But…with an 800 pixel wide sclera as described above…800 × 128 × 2 = 204 kilobytes…quite a bit over the 160K RAM limit! Whittle down one or both axes, whatever you think can best handle less resolution, until you find a size that fits. Once in motion, and at a reasonable viewing distance, minor “jaggies” aren’t that noticeable.

Photoshop users: the eye code maps textures differently than Photoshop’s “Polar Coordinates” filter. Ours wraps clockwise, with the top of the image becoming the outer circumference, while Photoshop wraps counterclockwise, with the bottom becoming the outer circumference. If using Photoshop to create or preview textures, simply rotate the rectangular image 180° before applying a rectangular-to-polar filter, or after a polar-to-rectangular filter.

Eyelids

The eyelids have stricter requirements. There are always two files (one each for the upper and lower eyelids) both 240 × 240 pixels exactly, both 1-bit BMP images (NOT 24-bit like the texture maps!).

The white area — which should span the entire 240 pixel width — represents the extent of vertical motion, the “open-most” and “closed-most” range. The eyes won’t hold the open-most position all the time when idle…the upper eyelid “tracks” the moving pupil, because that’s how eyes work.

A good set of eyelid images will overlap by a few pixels, so the eye makes a good solid blink.

This particular set of eyelids is slightly asymmetrical to approximate the eye’s caruncle* — that triangular bit by the tear duct. The shape is mirrored between the two eyes. But it is super 100% okay to make symmetrical eyelids if you prefer…a simple “football shape”…that usually looks better on a single-eye board like the HalloWing M4. For some of the eye designs we’ll offer both.

* Which itself is a vestigial remnant of the nictitating membrane, the “third eyelid” that reptiles have. How cool is that!?

The eyelid images as shown above are for the right eye. That is…the monster’s right eye, meaning the eye on the left when looking at the M4SK.

Let’s look at the configuration for our stock “hazel” human eye. It doesn’t reference every configurable setting, but shows the general format of these files…

This is a plain text file that any simple editor should be able to handle…Notepad or TextEdit or whatever comes bundled on your computer.

It’s worth reiterating these points from the “Customization Basics” page:

• The whole thing is contained within curly braces — { and }
• Each item has a name (always in quotes), a colon separator ( : ) and a value. Values might be numbers, strings (in quotes) or arrays (in square brackets) depending on what’s being configured.
• This is a list and each line ends with a comma…except for the last item in the list. Watch out for missing or surplus commas!
• JSON is case-sensitive. "Foo", "foo" and "FOO" are all different things.
• Using extra spaces to line up columns really isn’t necessary, just something I do to assist with legibility. Some folks just find it annoying.
• Comments (starting with // ) are not standard JSON syntax, but the library we use allows them and I find them very helpful.

The “Customization Basics” page also has some JSON troubleshooting tips. If you encounter trouble, review that page!

Two settings define the basic geometry of the eye…

eyeRadius establishes the size of the overall eyeball, in pixels. This is a radius — center to edge — so the overall eye size is twice this across. For example eyeRadius:125 configures the eye to be 250 pixels wide. This is the default if left unspecified.

The screens are only 240 pixels wide. Reason the eye is made a little bigger is because the code uses tricks to fake a rotating sphere…and that faking is more apparent as the pupil approaches an edge. So we push the edge out a few extra pixels, then cover it up with eyelids.

If designing an eye with no eyelids, you might want eyeRadius:120 instead, which provides a nice perfect circle on the screen.

Remember that JSON is case-sensitive. This must be spelled eyeRadius. Different capitalization will cause it to be ignored!

irisRadius establishes the size of the iris…again a radius, in pixels.

irisRadius:60 will make the iris 120 pixels across, or half the width of the screen. If you plan to use lenses over the displays, consider scaling down this number a bit to compensate.

Some creatures…cats and so forth…have very large irises and almost no visible sclera. In that case you can set irisRadius much larger, up to (but not exceeding) eyeRadius.

A third setting, slitPupilRadius, lets you make cat or dragon type eyes with a vertical slit pupil (only a vertical slit is available, no goat pupils, sorry). If set to 0 (the default), a normal round pupil is used. Larger numbers (up to irisRadius) make a taller/thinner pupil. This number sets the height. You’ll probably want an in-between value…maybe irisRadius:80 (160 pixels round) and slitPupilRadius:60 (120 pixels tall) to start.

Note that using slitPupilRadius makes the program a bit slower to initialize…you’ll just see blank screens for several seconds while it works. This is normal and just an unfortunate math thing.

The texture map images for iris and sclera are specified with irisTexture and scleraTexture:

If you prefer a solid color to an image, omit those lines and instead use irisColor and/or scleraColor, with color specified as three values [red, green,blue] in brackets:

Colors can either be three integers in the range 0 to 255 (like most folks are used to) or three floating-point values from 0.0 to 1.0 (same idea, just different scale).

Solid colors save a TON of space compared to texture maps, if your design can get away with it.

A few more items have configurable colors:

pupilColor is self-explanatory…like if you want glowing red or white pupils or something.

backColor covers the outermost/backmost part of the eye where the sclera texture map (or color) doesn’t reach. With a little planning, this and your sclera texture map can be designed to blend together…otherwise you’ll see a conspicuous crescent of backColor when the eye is looking off to a side.

The eyelid (and background) color is also configurable, but it’s not a normal RGB color. Instead, use eyelidIndex and an 8-bit value:

Notice the hexadecimal value must be quoted. The value corresponds to one of the colors in this palette:

eyelidIndex doesn’t use a normal RGB value because the code is using an optimization trick here, which limits the available colors in this one area. The default eyelidIndex if unspecified is “0x00” — black.

The eyelid graphics format is explained on the “Preparing Graphics” page. These are 1-bit BMPs, 240 pixels square. Use the upperEyelid and lowerEyelid settings to specify the filenames of these images.

Eyelids are one of those global (not configurable per-eye) things. The design of the eyelid graphics might be symmetrical or asymmetrical…some eye designs might provide two sets of eyelid images, one for a single-eye device (like HalloWing M4) and another for the Monster M4SK, where the left and right eyelid shapes are mirrored.

If you want no eyelids at all, just leave out any upperEyelid or lowerEyelid filenames. You’ll then have a circular unblinking eye…looks good for skulls!

With eyelids enabled, normally the upper lid “tracks” the movement of the pupil (when the eye looks down, the eyelid follows with it). This is something that eyes do in real life…but some folks think it looks sleepy, or just want a particular caffeinated look. Use the tracking keyword with a value of false to disable the eyelid tracking and maintain wide-open eyes.

The pupils normally do some dilation movement on their own…but you can have them respond to light if you like! Use the lightSensor keyword along with a pin number where the light sensor is connected. If it’s on a Seesaw interface chip (e.g. on MONSTER M4SK), add 100 to the pin number. For example, on MONSTER M4SK the built-in light sensor is:

"lightSensor" : 102

And for HalloWing M4, use:

"lightSensor" : 21

(Add a trailing comma if this appears mid-file.)

The next couple of pages explain how to set up the Arduino software to work with the SAMD microcontroller boards like the HalloWing M4 or MONSTER M4SK. If you’ve previously worked with Adafruit SAMD boards with Arduino, that part’s already set up and you can skip ahead to the “Libraries and Settings” page.

In summary: use the latest Arduino IDE, install the latest Adafruit SAMD boards package, install Windows drivers if necessary, and install all the library dependencies for this project…

The first thing you will need to do is to download the latest release of the Arduino IDE. You will need to be using version 1.8 or higher for this guide

After you have downloaded and installed the latest version of Arduino IDE, you will need to start the IDE and navigate to the Preferences menu. You can access it from the File menu in Windows or Linux, or the Arduino menu on OS X.

A dialog will pop up just like the one shown below.

We will be adding a URL to the new Additional Boards Manager URLs option. The list of URLs is comma separated, and you will only have to add each URL once. New Adafruit boards and updates to existing boards will automatically be picked up by the Board Manager each time it is opened. The URLs point to index files that the Board Manager uses to build the list of available & installed boards.

To find the most up to date list of URLs you can add, you can visit the list of third party board URLs on the Arduino IDE wiki. We will only need to add one URL to the IDE in this example, but you can add multiple URLS by separating them with commas. Copy and paste the link below into the Additional Boards Manager URLs option in the Arduino IDE preferences.

https://adafruit.github.io/arduino-board-index/package_adafruit_index.json

Here's a short description of each of the Adafruit supplied packages that will be available in the Board Manager when you add the URL:

• Adafruit AVR Boards - Includes support for Flora, Gemma, Feather 32u4, ItsyBitsy 32u4, Trinket, & Trinket Pro.
• Adafruit SAMD Boards - Includes support for Feather M0 and M4, Metro M0 and M4, ItsyBitsy M0 and M4, Circuit Playground Express, Gemma M0 and Trinket M0
• Arduino Leonardo & Micro MIDI-USB - This adds MIDI over USB support for the Flora, Feather 32u4, Micro and Leonardo using the arcore project.

If you have multiple boards you want to support, say ESP8266 and Adafruit, have both URLs in the text box separated by a comma (,)

Once done click OK to save the new preference settings. Next we will look at installing boards with the Board Manager.

Now continue to the next step to actually install the board support package!

The Feather/Metro/Gemma/QTPy/Trinket M0 and M4 use an ATSAMD21 or ATSAMD51 chip, and you can pretty easily get it working with the Arduino IDE. Most libraries (including the popular ones like NeoPixels and display) will work with the M0 and M4, especially devices & sensors that use I2C or SPI.

Now that you have added the appropriate URLs to the Arduino IDE preferences in the previous page, you can open the Boards Manager by navigating to the Tools->Board menu.

Once the Board Manager opens, click on the category drop down menu on the top left hand side of the window and select All. You will then be able to select and install the boards supplied by the URLs added to the preferences.

Install SAMD Support

First up, install the latest Arduino SAMD Boards (version 1.6.11 or later)

You can type Arduino SAMD in the top search bar, then when you see the entry, click Install

Install Adafruit SAMD

Next you can install the Adafruit SAMD package to add the board file definitions

Make sure you have Type All selected to the left of the Filter your search... box

You can type Adafruit SAMD in the top search bar, then when you see the entry, click Install

Even though in theory you don't need to - I recommend rebooting the IDE

Quit and reopen the Arduino IDE to ensure that all of the boards are properly installed. You should now be able to select and upload to the new boards listed in the Tools->Board menu.

Select the matching board, the current options are:

• Feather M0 (for use with any Feather M0 other than the Express)
• Feather M0 Express
• Metro M0 Express
• Circuit Playground Express
  
• Gemma M0
• Trinket M0
  
• QT Py M0
  
• ItsyBitsy M0
• Hallowing M0
• Crickit M0 (this is for direct programming of the Crickit, which is probably not what you want! For advanced hacking only)
• Metro M4 Express
• Grand Central M4 Express
• ItsyBitsy M4 Express
• Feather M4 Express
• Trellis M4 Express
• PyPortal M4
• PyPortal M4 Titano
• PyBadge M4 Express
• Metro M4 Airlift Lite
• PyGamer M4 Express
• MONSTER M4SK
• Hallowing M4
• MatrixPortal M4
• BLM Badge

Install Drivers (Windows 7 & 8 Only)

When you plug in the board, you'll need to possibly install a driver

Click below to download our Driver Installer

Download and run the installer

Run the installer! Since we bundle the SiLabs and FTDI drivers as well, you'll need to click through the license

Select which drivers you want to install, the defaults will set you up with just about every Adafruit board!

Click Install to do the installin'

Blink

Now you can upload your first blink sketch!

Plug in the M0 or M4 board, and wait for it to be recognized by the OS (just takes a few seconds). It will create a serial/COM port, you can now select it from the drop-down, it'll even be 'indicated' as Trinket/Gemma/Metro/Feather/ItsyBitsy/Trellis!

Please note, the QT Py and Trellis M4 Express are two of our very few boards that does not have an onboard pin 13 LED so you can follow this section to practice uploading but you wont see an LED blink!

Now load up the Blink example

And click upload! That's it, you will be able to see the LED blink rate change as you adapt the delay() calls.

Successful Upload

If you have a successful upload, you'll get a bunch of red text that tells you that the device was found and it was programmed, verified & reset

After uploading, you may see a message saying "Disk Not Ejected Properly" about the ...BOOT drive. You can ignore that message: it's an artifact of how the bootloader and uploading work.

Compilation Issues

If you get an alert that looks like

Cannot run program "{runtime.tools.arm-none-eabi-gcc.path}\bin\arm-non-eabi-g++"

Make sure you have installed the Arduino SAMD boards package, you need both Arduino & Adafruit SAMD board packages

Manually bootloading

If you ever get in a 'weird' spot with the bootloader, or you have uploaded code that crashes and doesn't auto-reboot into the bootloader, click the RST button twice (like a double-click)to get back into the bootloader.

The red LED will pulse and/or RGB LED will be green, so you know that its in bootloader mode.

Once it is in bootloader mode, you can select the newly created COM/Serial port and re-try uploading.

You may need to go back and reselect the 'normal' USB serial port next time you want to use the normal upload.

Ubuntu & Linux Issue Fix

 Follow the steps for installing Adafruit's udev rules on this page.

The ATSAMD21 and 51 are very nice little chips, but fairly new as Arduino-compatible cores go. Most sketches & libraries will work but here’s a collection of things we noticed.

The notes below cover a range of Adafruit M0 and M4 boards, but not every rule will apply to every board (e.g. Trinket and Gemma M0 do not have ARef, so you can skip the Analog References note!).

Analog References

If you'd like to use the ARef pin for a non-3.3V analog reference, the code to use is analogReference(AR_EXTERNAL) (it's AR_EXTERNAL not EXTERNAL)

Pin Outputs & Pullups

The old-style way of turning on a pin as an input with a pullup is to use

pinMode(pin, INPUT)
digitalWrite(pin, HIGH)

This is because the pullup-selection register on 8-bit AVR chips is the same as the output-selection register.

For M0 & M4 boards, you can't do this anymore! Instead, use:

pinMode(pin, INPUT_PULLUP)

Code written this way still has the benefit of being backwards compatible with AVR. You don’t need separate versions for the different board types.

Serial vs SerialUSB

99.9% of your existing Arduino sketches use Serial.print to debug and give output. For the Official Arduino SAMD/M0 core, this goes to the Serial5 port, which isn't exposed on the Feather. The USB port for the Official Arduino M0 core is called SerialUSB instead.

In the Adafruit M0/M4 Core, we fixed it so that Serial goes to USB so it will automatically work just fine.

However, on the off chance you are using the official Arduino SAMD core and not the Adafruit version (which really, we recommend you use our version because it’s been tuned to our boards), and you want your Serial prints and reads to use the USB port, use SerialUSB instead of Serial in your sketch.

If you have existing sketches and code and you want them to work with the M0 without a huge find-replace, put

#if defined(ARDUINO_SAMD_ZERO) && defined(SERIAL_PORT_USBVIRTUAL)
  // Required for Serial on Zero based boards
  #define Serial SERIAL_PORT_USBVIRTUAL
#endif

right above the first function definition in your code. For example:

AnalogWrite / PWM on Feather/Metro M0

After looking through the SAMD21 datasheet, we've found that some of the options listed in the multiplexer table don't exist on the specific chip used in the Feather M0.

For all SAMD21 chips, there are two peripherals that can generate PWM signals: The Timer/Counter (TC) and Timer/Counter for Control Applications (TCC). Each SAMD21 has multiple copies of each, called 'instances'.

Each TC instance has one count register, one control register, and two output channels. Either channel can be enabled and disabled, and either channel can be inverted. The pins connected to a TC instance can output identical versions of the same PWM waveform, or complementary waveforms.

Each TCC instance has a single count register, but multiple compare registers and output channels. There are options for different kinds of waveform, interleaved switching, programmable dead time, and so on.

The biggest members of the SAMD21 family have five TC instances with two 'waveform output' (WO) channels, and three TCC instances with eight WO channels:

• TC[0-4],WO[0-1]
• TCC[0-2],WO[0-7]

And those are the ones shown in the datasheet's multiplexer tables.

The SAMD21G used in the Feather M0 only has three TC instances with two output channels, and three TCC instances with eight output channels:

• TC[3-5],WO[0-1]
• TCC[0-2],WO[0-7]

Tracing the signals to the pins broken out on the Feather M0, the following pins can't do PWM at all:

• Analog pin A5

The following pins can be configured for PWM without any signal conflicts as long as the SPI, I2C, and UART pins keep their protocol functions:

• Digital pins 5, 6, 9, 10, 11, 12, and 13
• Analog pins A3 and A4

If only the SPI pins keep their protocol functions, you can also do PWM on the following pins:

• TX and SDA (Digital pins 1 and 20)

analogWrite() PWM range

On AVR, if you set a pin's PWM with analogWrite(pin, 255) it will turn the pin fully HIGH. On the ARM cortex, it will set it to be 255/256 so there will be very slim but still-existing pulses-to-0V. If you need the pin to be fully on, add test code that checks if you are trying to analogWrite(pin, 255) and, instead, does a digitalWrite(pin, HIGH)

analogWrite() DAC on A0

If you are trying to use analogWrite() to control the DAC output on A0, make sure you do not have a line that sets the pin to output. Remove: pinMode(A0, OUTPUT).

Missing header files

There might be code that uses libraries that are not supported by the M0 core. For example if you have a line with

#include <util/delay.h>

you'll get an error that says

fatal error: util/delay.h: No such file or directory
  #include <util/delay.h>
                         ^
compilation terminated.
Error compiling.

In which case you can simply locate where the line is (the error will give you the file name and line number) and 'wrap it' with #ifdef's so it looks like:

The above will also make sure that header file isn't included for other architectures

If the #include is in the arduino sketch itself, you can try just removing the line.

Bootloader Launching

For most other AVRs, clicking reset while plugged into USB will launch the bootloader manually, the bootloader will time out after a few seconds. For the M0/M4, you'll need to double click the button. You will see a pulsing red LED to let you know you're in bootloader mode. Once in that mode, it wont time out! Click reset again if you want to go back to launching code.

Aligned Memory Access

This is a little less likely to happen to you but it happened to me! If you're used to 8-bit platforms, you can do this nice thing where you can typecast variables around. e.g.

uint8_t mybuffer[4];
float f = (float)mybuffer;

You can't be guaranteed that this will work on a 32-bit platform because mybuffer might not be aligned to a 2 or 4-byte boundary. The ARM Cortex-M0 can only directly access data on 16-bit boundaries (every 2 or 4 bytes). Trying to access an odd-boundary byte (on a 1 or 3 byte location) will cause a Hard Fault and stop the MCU. Thankfully, there's an easy work around ... just use memcpy!

uint8_t mybuffer[4];
float f;
memcpy(&f, mybuffer, 4)

Floating Point Conversion

Like the AVR Arduinos, the M0 library does not have full support for converting floating point numbers to ASCII strings. Functions like sprintf will not convert floating point.  Fortunately, the standard AVR-LIBC library includes the dtostrf function which can handle the conversion for you.

Unfortunately, the M0 run-time library does not have dtostrf.  You may see some references to using #include <avr/dtostrf.h> to get dtostrf in your code.  And while it will compile, it does not work.

Instead, check out this thread to find a working dtostrf function you can include in your code:

http://forum.arduino.cc/index.php?topic=368720.0

How Much RAM Available?

The ATSAMD21G18 has 32K of RAM, but you still might need to track it for some reason. You can do so with this handy function:

Thx to http://forum.arduino.cc/index.php?topic=365830.msg2542879#msg2542879 for the tip!

Storing data in FLASH

If you're used to AVR, you've probably used PROGMEM to let the compiler know you'd like to put a variable or string in flash memory to save on RAM. On the ARM, its a little easier, simply add const before the variable name:

const char str[] = "My very long string";

That string is now in FLASH. You can manipulate the string just like RAM data, the compiler will automatically read from FLASH so you dont need special progmem-knowledgeable functions.

You can verify where data is stored by printing out the address:
Serial.print("Address of str $"); Serial.println((int)&str, HEX);

If the address is $2000000 or larger, its in SRAM. If the address is between $0000 and $3FFFF Then it is in FLASH

Pretty-Printing out registers

There's a lot of registers on the SAMD21, and you often are going through ASF or another framework to get to them. So having a way to see exactly what's going on is handy. This library from drewfish will help a ton!

https://github.com/drewfish/arduino-ZeroRegs

M4 Performance Options

As of version 1.4.0 of the Adafruit SAMD Boards package in the Arduino Boards Manager, some options are available to wring extra performance out of M4-based devices. These are in the Tools menu.

All of these performance tweaks involve a degree of uncertainty. There’s no guarantee of improved performance in any given project, and some may even be detrimental, failing to work in part or in whole. If you encounter trouble, select the default performance settings and re-upload.

Here’s what you get and some issues you might encounter…

CPU Speed (overclocking)

This option lets you adjust the microcontroller core clock…the speed at which it processes instructions…beyond the official datasheet specifications.

Manufacturers often rate speeds conservatively because such devices are marketed for harsh industrial environments…if a system crashes, someone could lose a limb or worse. But most creative tasks are less critical and operate in more comfortable settings, and we can push things a bit if we want more speed.

There is a small but nonzero chance of code locking up or failing to run entirely. If this happens, try dialing back the speed by one notch and re-upload, see if it’s more stable.

Much more likely, some code or libraries may not play well with the nonstandard CPU speed. For example, currently the NeoPixel library assumes a 120 MHz CPU speed and won’t issue the correct data at other settings (this will be worked on). Other libraries may exhibit similar problems, usually anything that strictly depends on CPU timing…you might encounter problems with audio- or servo-related code depending how it’s written. If you encounter such code or libraries, set the CPU speed to the default 120 MHz and re-upload.

Optimize

There’s usually more than one way to solve a problem, some more resource-intensive than others. Since Arduino got its start on resource-limited AVR microcontrollers, the C++ compiler has always aimed for the smallest compiled program size. The “Optimize” menu gives some choices for the compiler to take different and often faster approaches, at the expense of slightly larger program size…with the huge flash memory capacity of M4 devices, that’s rarely a problem now.

The “Small” setting will compile your code like it always has in the past, aiming for the smallest compiled program size.

The “Fast” setting invokes various speed optimizations. The resulting program should produce the same results, is slightly larger, and usually (but not always) noticably faster. It’s worth a shot!

“Here be dragons” invokes some more intensive optimizations…code will be larger still, faster still, but there’s a possibility these optimizations could cause unexpected behaviors. Some code may not work the same as before. Hence the name. Maybe you’ll discover treasure here, or maybe you’ll sail right off the edge of the world.

Most code and libraries will continue to function regardless of the optimizer settings. If you do encounter problems, dial it back one notch and re-upload.

Cache

This option allows a small collection of instructions and data to be accessed more quickly than from flash memory, boosting performance. It’s enabled by default and should work fine with all code and libraries. But if you encounter some esoteric situation, the cache can be disabled, then recompile and upload.

Max SPI and Max QSPI

These should probably be left at their defaults. They’re present mostly for our own experiments and can cause serious headaches.

Max SPI determines the clock source for the M4’s SPI peripherals. Under normal circumstances this allows transfers up to 24 MHz, and should usually be left at that setting. But…if you’re using write-only SPI devices (such as TFT or OLED displays), this option lets you drive them faster (we’ve successfully used 60 MHz with some TFT screens). The caveat is, if using any read/write devices (such as an SD card), this will not work at all…SPI reads absolutely max out at the default 24 MHz setting, and anything else will fail. Write = OK. Read = FAIL. This is true even if your code is using a lower bitrate setting…just having the different clock source prevents SPI reads.

Max QSPI does similarly for the extra flash storage on M4 “Express” boards. Very few Arduino sketches access this storage at all, let alone in a bandwidth-constrained context, so this will benefit next to nobody. Additionally, due to the way clock dividers are selected, this will only provide some benefit when certain “CPU Speed” settings are active. Our PyPortal Animated GIF Display runs marginally better with it, if using the QSPI flash.

Enabling the Buck Converter on some M4 Boards

If you want to reduce power draw, some of our boards have an inductor so you can use the 1.8V buck converter instead of the built in linear regulator. If the board does have an inductor (see the schematic) you can add the line SUPC->VREG.bit.SEL = 1; to your code to switch to it. Note it will make ADC/DAC reads a bit noisier so we don't use it by default. You'll save ~4mA.

At this point you should be able to upload code (such as the “Blink” sketch) to the board, the basics are confirmed working. If not, work though the prior two pages.

Source Code

Source code for this project is found in the Adafruit_Learning_System_Guides repository on Github, specifically in the M4_Eyes subdirectory. Here’s a direct download link:

Libraries

The following libraries are used by the eye code (or are referenced indirectly by libraries used). These can be installed through the Arduino Library Manager (Sketch→Include Library→Manage Libraries…)

• Adafruit_Arcada
• Adafruit_GFX
• Adafruit_BusIO
• Adafruit_ImageReader
• Adafruit_LIS3DH
• Adafruit_MSA301
• Adafruit_NeoPixel
• Adafruit_Seesaw
• Adafruit_Sensor (Adafruit Unified Sensor in Library Manager)
• Adafruit_SPIFlash
• Adafruit_ST7789
• Adafruit_TinyUSB
• Adafruit_TouchScreen
• Adafruit_ZeroDMA (Adafruit Zero DMA in Library Manager)
• Adafruit_ZeroPDM (Adafruit Zero PDM in Library Manager)
• Adafruit_ZeroTimer
• ArduinoJson (not Arduino_JSON)
• SdFat - Adafruit fork (not the standard SdFat fork)

Recent versions of the Arduino IDE install dependent automatically, so installing Adafruit_Arcada should cover most if not all of this.

Additionally, depending on which of the “user” code tabs is active, you might need:

• Adafruit_AMG88xx (used by user_watch.cpp)

Project Settings

• Tools→CPU Speed→180 MHz (overclock) (200 MHz is a bit too much for some boards and may lock up, but you can give it a try. We use 180 MHz for our prepackaged .UF2 files since it’s likely to work on more boards in the wild.)
• Tools→Optimize→Faster (-O3) (do not use higher settings for this code, it may corrupt the flash filesystem requiring a complete wipe)
• Tools→USB Stack→TinyUSB (the code will not compile without this selected!)

One of the best features of the M0 express board is a small SPI flash memory chip built into the board.  This memory can be used for almost any purpose like storing data files, Python code, and more.  Think of it like a little SD card that is always connected to the board, and in fact with Arduino you can access the memory using a library that is very similar to the Arduino SD card library.  You can even read and write files that CircuitPython stores on the flash chip!

To use the flash memory with Arduino you'll need to install the Adafruit SPI Flash Memory library in the Arduino IDE.

Open up the Arduino library manager

Search for the Adafruit SPIFlash library and install it

Search for the SdFat - Adafruit Fork library and install it

We also have a great tutorial on Arduino library installation at:
http://learn.adafruit.com/adafruit-all-about-arduino-libraries-install-use

Once the library is installed look for the following examples in the library:

• fatfs_circuitpython
• fatfs_datalogging
• fatfs_format
• fatfs_full_usage
• fatfs_print_file
• flash_erase

These examples allow you to format the flash memory with a FAT filesystem (the same kind of filesystem used on SD cards) and read and write files to it just like a SD card.

Read & Write CircuitPython Files

The fatfs_circuitpython example shows how to read and write files on the flash chip so that they're accessible from CircuitPython.  This means you can run a CircuitPython program on your board and have it store data, then run an Arduino sketch that uses this library to interact with the same data.

Note that before you use the fatfs_circuitpython example you must have loaded CircuitPython on your board. Load the latest version of CircuitPython as explained in this guide first to ensure a CircuitPython filesystem is initialized and written to the flash chip.  Once you've loaded CircuitPython then you can run the fatfs_circuitpython example sketch.

To run the sketch load it in the Arduino IDE and upload it to the Feather/Metro/ItsyBitsy M0 board.  Then open the serial monitor at 115200 baud.  You should see the serial monitor display messages as it attempts to read files and write to a file on the flash chip.  Specifically the example will look for a boot.py and main.py file (like what CircuitPython runs when it starts) and print out their contents.  Then it will add a line to the end of a data.txt file on the board (creating it if it doesn't exist already).  After running the sketch you can reload CircuitPython on the board and open the data.txt file to read it from CircuitPython!

To understand how to read & write files that are compatible with CircuitPython let's examine the sketch code.  First notice an instance of the Adafruit_M0_Express_CircuitPython class is created and passed an instance of the flash chip class in the last line below:

By using this Adafruit_M0_Express_CircuitPython class you'll get a filesystem object that is compatible with reading and writing files on a CircuitPython-formatted flash chip.  This is very important for interoperability between CircuitPython and Arduino as CircuitPython has specialized partitioning and flash memory layout that isn't compatible with simpler uses of the library (shown in the other examples).

Once an instance of the Adafruit_M0_Express_CircuitPython class is created (called pythonfs in this sketch) you can go on to interact with it just like if it were the SD card library in Arduino.  You can open files for reading & writing, create directories, delete files and directories and more.  Here's how the sketch checks if a boot.py file exists and prints it out a character at a time:

Notice the exists function is called to check if the boot.py file is found, and then the open function is used to open it in read mode.  Once a file is opened you'll get a reference to a File class object which you can read and write from as if it were a Serial device (again just like the SD card library, all of the same File class functions are available).  In this case the available function will return the number of bytes left to read in the file, and the read function will read a character at a time to print it to the serial monitor.

Writing a file is just as easy, here's how the sketch writes to data.txt:

Again the open function is used but this time it's told to open the file for writing.  In this mode the file will be opened for appending (i.e. data added to the end of it) if it exists, or it will be created if it doesn't exist.  Once the file is open print functions like print and println can be used to write data to the file (just like writing to the serial monitor).  Be sure to close the file when finished writing!

That's all there is to basic file reading and writing.  Check out the fatfs_full_usage example for examples of even more functions like creating directories, deleting files & directories, checking the size of files, and more!  Remember though to interact with CircuitPython files you need to use the Adafruit_Feather_M0_CircuitPython class as shown in the fatfs_circuitpython example above!

Format Flash Memory

The fatfs_format example will format the SPI flash with a new blank filesystem.  Be warned this sketch will delete all data on the flash memory, including any Python code or other data you might have stored!  The format sketch is useful if you'd like to wipe everything away and start fresh, or to help get back in a good state if the memory should get corrupted for some reason.

Be aware too the fatfs_format and examples below are not compatible with a CircuitPython-formatted flash chip!  If you need to share data between Arduino & CircuitPython check out the fatfs_circuitpython example above.

To run the format sketch load it in the Arduino IDE and upload it to the M0 board.  Then open the serial monitor at 115200 baud.  You should see the serial monitor display a message asking you to confirm formatting the flash.  If you don't see this message then close the serial monitor, press the board's reset button, and open the serial monitor again.

Type OK and press enter in the serial monitor input to confirm that you'd like to format the flash memory.  You need to enter OK in all capital letters!  

Once confirmed the sketch will format the flash memory.  The format process takes about a minute so be patient as the data is erased and formatted.  You should see a message printed once the format process is complete.  At this point the flash chip will be ready to use with a brand new empty filesystem.

Datalogging Example

One handy use of the SPI flash is to store data, like datalogging sensor readings.  The fatfs_datalogging example shows basic file writing/datalogging.  Open the example in the Arduino IDE and upload it to your Feather M0 board.  Then open the serial monitor at 115200 baud.  You should see a message printed every minute as the sketch writes a new line of data to a file on the flash filesystem.

To understand how to write to a file look in the loop function of the sketch:

Just like using the Arduino SD card library you create a File object by calling an open function and pointing it at the name of the file and how you'd like to open it (FILE_WRITE mode, i.e. writing new data to the end of the file).  Notice however instead of calling open on a global SD card object you're calling it on a fatfs object created earlier in the sketch (look at the top after the #define configuration values).

Once the file is opened it's simply a matter of calling print and println functions on the file object to write data inside of it.  This is just like writing data to the serial monitor and you can print out text, numeric, and other types of data.  Be sure to close the file when you're done writing to ensure the data is stored correctly!

Reading and Printing Files

The fatfs_print_file example will open a file (by default the data.csv file created by running the fatfs_datalogging example above) and print all of its contents to the serial monitor.  Open the fatfs_print_file example and load it on your Feather M0 board, then open the serial monitor at 115200 baud.  You should see the sketch print out the contents of data.csv (if you don't have a file called data.csv on the flash look at running the datalogging example above first).

To understand how to read data from a file look in the setup function of the sketch:

Just like when writing data with the datalogging example you create a File object by calling the open function on a fatfs object.  This time however you pass a file mode of FILE_READ which tells the filesystem you want to read data.

After you open a file for reading you can easily check if data is available by calling the available function on the file, and then read a single character with the read function.  This makes it easy to loop through all of the data in a file by checking if it's available and reading a character at a time.  However there are more advanced read functions you can use too--see the fatfs_full_usage example or even the Arduino SD library documentation (the SPI flash library implements the same functions).

Full Usage Example

For a more complete demonstration of reading and writing files look at the fatfs_full_usage example.  This examples uses every function in the library and demonstrates things like checking for the existence of a file, creating directories, deleting files, deleting directories, and more.

Remember the SPI flash library is built to have the same functions and interface as the Arduino SD library so if you have code or examples that store data on a SD card they should be easy to adapt to use the SPI flash library, just create a fatfs object like in the examples above and use its open function instead of the global SD object's open function.  Once you have a reference to a file all of the functions and usage should be the same between the SPI flash and SD libraries!

Accessing SPI Flash

Arduino doesn't have the ability to show up as a 'mass storage' disk drive. So instead we must use CircuitPython to do that part for us. Here's the full technique:

• Start the bootloader on the Express board. Drag over the latest circuitpython uf2 file
• After a moment, you should see a CIRCUITPY drive appear on your hard drive with boot_out.txt on it
• Now go to Arduino and upload the fatfs_circuitpython example sketch from the Adafruit SPI library. Open the serial console. It will successfully mount the filesystem and write a new line to data.txt

• Back on your computer, re-start the Express board bootloader, and re-drag circuitpython.uf2 onto the BOOT drive to reinstall circuitpython
• Check the CIRCUITPY drive, you should now see data.txt which you can open to read!

Once you have your Arduino sketch working well, for datalogging, you can simplify this procedure by dragging CURRENT.UF2 off of the BOOT drive to make a backup of the current program before loading circuitpython on. Then once you've accessed the file you want, re-drag CURRENT.UF2 back onto the BOOT drive to re-install the Arduino sketch!

The Arduino IDE gives me "Device Descriptor Request Failed"

This can require "manual bootloading".

If you ever get in a 'weird' spot with the bootloader, or you have uploaded code that crashes and doesn't auto-reboot into the bootloader, double-click the RST button to get back into the bootloader. The red LED will pulse, so you know that its in bootloader mode. Do the reset button double-press right as the Arduino IDE says its attempting to upload the sketch, when you see the Yellow Arrow lit and the Uploading... text in the status bar.

Don't click the reset button before uploading, unlike other bootloaders you want this one to run at the time Arduino is trying to upload

CircuitPython is a programming language designed to simplify experimenting and learning to program on low-cost microcontroller boards. It makes getting started easier than ever with no upfront desktop downloads needed. Once you get your board set up, open any text editor, and get started editing code. It's that simple.

CircuitPython is based on Python

Python is the fastest growing programming language. It's taught in schools and universities. It's a high-level programming language which means it's designed to be easier to read, write and maintain. It supports modules and packages which means it's easy to reuse your code for other projects. It has a built in interpreter which means there are no extra steps, like compiling, to get your code to work. And of course, Python is Open Source Software which means it's free for anyone to use, modify or improve upon.

CircuitPython adds hardware support to all of these amazing features. If you already have Python knowledge, you can easily apply that to using CircuitPython. If you have no previous experience, it's really simple to get started!

Why would I use CircuitPython?

CircuitPython is designed to run on microcontroller boards. A microcontroller board is a board with a microcontroller chip that's essentially an itty-bitty all-in-one computer. The board you're holding is a microcontroller board! CircuitPython is easy to use because all you need is that little board, a USB cable, and a computer with a USB connection. But that's only the beginning.

Other reasons to use CircuitPython include:

• You want to get up and running quickly. Create a file, edit your code, save the file, and it runs immediately. There is no compiling, no downloading and no uploading needed.
• You're new to programming. CircuitPython is designed with education in mind. It's easy to start learning how to program and you get immediate feedback from the board.
• Easily update your code. Since your code lives on the disk drive, you can edit it whenever you like, you can also keep multiple files around for easy experimentation.
• The serial console and REPL. These allow for live feedback from your code and interactive programming.
• File storage. The internal storage for CircuitPython makes it great for data-logging, playing audio clips, and otherwise interacting with files.
• Strong hardware support. CircuitPython has builtin support for microcontroller hardware features like digital I/O pins, hardware buses (UART, I2C, SPI), audio I/O, and other capabilities. There are also many libraries and drivers for sensors, breakout boards and other external components.
• It's Python! Python is the fastest-growing programming language. It's taught in schools and universities. CircuitPython is almost-completely compatible with Python. It simply adds hardware support.

This is just the beginning. CircuitPython continues to evolve, and is constantly being updated. Adafruit welcomes and encourages feedback from the community, and incorporate it into the development of CircuitPython. That's the core of the open source concept. This makes CircuitPython better for you and everyone who uses it!

Mu is a simple code editor that works with the Adafruit CircuitPython boards. It's written in Python and works on Windows, MacOS, Linux and Raspberry Pi. The serial console is built right in so you get immediate feedback from your board's serial output!

Download and Install Mu

Starting Up Mu

Using Mu

You can now explore Mu! The three main sections of the window are labeled below; the button bar, the text editor, and the serial console / REPL.

Now you're ready to code! Let's keep going...

One of the best things about CircuitPython is how simple it is to get code up and running. This section covers how to create and edit your first CircuitPython program.

To create and edit code, all you'll need is an editor. There are many options. Adafruit strongly recommends using Mu! It's designed for CircuitPython, and it's really simple and easy to use, with a built in serial console!

If you don't or can't use Mu, there are a number of other editors that work quite well. The Recommended Editors page has more details. Otherwise, make sure you do "Eject" or "Safe Remove" on Windows or "sync" on Linux after writing a file if you aren't using Mu. (This is not a problem on MacOS.)

Creating Code

Copy and paste the following code into your editor:

If you're using a KB2040, QT Py or a Trinkey, please download the NeoPixel blink example.

The little LED should now be blinking. Once per half-second.

Congratulations, you've just run your first CircuitPython program!

Editing Code

Your code changes are run as soon as the file is done saving.

There's one warning before you continue...

The CircuitPython code on your board detects when the files are changed or written and will automatically re-start your code. This makes coding very fast because you save, and it re-runs. If you unplug or reset the board before your computer finishes writing the file to your board, you can corrupt the drive. If this happens, you may lose the code you've written, so it's important to backup your code to your computer regularly.

There are a couple of ways to avoid filesystem corruption.

1. Use an editor that writes out the file completely when you save it.

Check out the Recommended Editors page for details on different editing options.

2. Eject or Sync the Drive After Writing

If you are using one of our not-recommended-editors, not all is lost! You can still make it work.

On Windows, you can Eject or Safe Remove the CIRCUITPY drive. It won't actually eject, but it will force the operating system to save your file to disk. On Linux, use the sync command in a terminal to force the write to disk.

You also need to do this if you use Windows Explorer or a Linux graphical file manager to drag a file onto CIRCUITPY.

Back to Editing Code...

Now! Let's try editing the program you added to your board. Open your code.py file into your editor. You'll make a simple change. Change the first 0.5 to 0.1. The code should look like this:

Leave the rest of the code as-is. Save your file. See what happens to the LED on your board? Something changed! Do you know why?

You don't have to stop there! Let's keep going. Change the second 0.5 to 0.1 so it looks like this:

Now it blinks really fast! You decreased the both time that the code leaves the LED on and off!

Now try increasing both of the 0.1 to 1. Your LED will blink much more slowly because you've increased the amount of time that the LED is turned on and off.

Well done! You're doing great! You're ready to start into new examples and edit them to see what happens! These were simple changes, but major changes are done using the same process. Make your desired change, save it, and get the results. That's really all there is to it!

Naming Your Program File

CircuitPython looks for a code file on the board to run. There are four options: code.txt, code.py, main.txt and main.py. CircuitPython looks for those files, in that order, and then runs the first one it finds. While code.py is the recommended name for your code file, it is important to know that the other options exist. If your program doesn't seem to be updating as you work, make sure you haven't created another code file that's being read instead of the one you're working on.

One of the staples of CircuitPython (and programming in general!) is something called a "print statement". This is a line you include in your code that causes your code to output text. A print statement in CircuitPython (and Python) looks like this:

print("Hello, world!")

This line in your code.py would result in:

Hello, world!

However, these print statements need somewhere to display. That's where the serial console comes in!

The serial console receives output from your CircuitPython board sent over USB and displays it so you can see it. This is necessary when you've included a print statement in your code and you'd like to see what you printed. It is also helpful for troubleshooting errors, because your board will send errors and the serial console will display those too.

The serial console requires an editor that has a built in terminal, or a separate terminal program. A terminal is a program that gives you a text-based interface to perform various tasks.

Are you using Mu?

If so, good news! The serial console is built into Mu and will autodetect your board making using the serial console really really easy.

Once you've opened Mu with your board plugged in, look for the Serial button in the button bar and click it.

The Mu window will split in two, horizontally, and display the serial console at the bottom.

Serial Console Issues or Delays on Linux

If you're on Linux, and are seeing multi-second delays connecting to the serial console, or are seeing "AT" and other gibberish when you connect, then the modemmanager service might be interfering. Just remove it; it doesn't have much use unless you're still using dial-up modems.

To remove modemmanager, type the following command at a shell:

Setting Permissions on Linux

On Linux, if you see an error box something like the one below when you press the Serial button, you need to add yourself to a user group to have permission to connect to the serial console.

On Ubuntu and Debian, add yourself to the dialout group by doing:

After running the command above, reboot your machine to gain access to the group. On other Linux distributions, the group you need may be different. See the Advanced Serial Console on Linux for details on how to add yourself to the right group.

Using Something Else?

If you're not using Mu to edit, are using or if for some reason you are not a fan of its built in serial console, you can run the serial console from a separate program.

Windows requires you to download a terminal program. Check out the Advanced Serial Console on Windows page for more details.

MacOS has Terminal built in, though there are other options available for download. Check the Advanced Serial Console on Mac page for more details.

Linux has a terminal program built in, though other options are available for download. Check the Advanced Serial Console on Linux page for more details.

Once connected, you'll see something like the following.

Once you've successfully connected to the serial console, it's time to start using it.

The code you wrote earlier has no output to the serial console. So, you're going to edit it to create some output.

Open your code.py file into your editor, and include a print statement. You can print anything you like! Just include your phrase between the quotation marks inside the parentheses. For example:

Save your file.

Now, let's go take a look at the window with our connection to the serial console.

Excellent! Our print statement is showing up in our console! Try changing the printed text to something else.

Keep your serial console window where you can see it. Save your file. You'll see what the serial console displays when the board reboots. Then you'll see your new change!

The Traceback (most recent call last): is telling you the last thing your board was doing before you saved your file. This is normal behavior and will happen every time the board resets. This is really handy for troubleshooting. Let's introduce an error so you can see how it is used.

Delete the e at the end of True from the line led.value = True so that it says led.value = Tru

Save your file. You will notice that your red LED will stop blinking, and you may have a colored status LED blinking at you. This is because the code is no longer correct and can no longer run properly. You need to fix it!

Usually when you run into errors, it's not because you introduced them on purpose. You may have 200 lines of code, and have no idea where your error could be hiding. This is where the serial console can help. Let's take a look!

The Traceback (most recent call last): is telling you that the last thing it was able to run was line 10 in your code. The next line is your error: NameError: name 'Tru' is not defined. This error might not mean a lot to you, but combined with knowing the issue is on line 10, it gives you a great place to start!

Go back to your code, and take a look at line 10. Obviously, you know what the problem is already. But if you didn't, you'd want to look at line 10 and see if you could figure it out. If you're still unsure, try googling the error to get some help. In this case, you know what to look for. You spelled True wrong. Fix the typo and save your file.

Nice job fixing the error! Your serial console is streaming and your red LED Is blinking again.

The serial console will display any output generated by your code. Some sensors, such as a humidity sensor or a thermistor, receive data and you can use print statements to display that information. You can also use print statements for troubleshooting, which is called "print debugging". Essentially, if your code isn't working, and you want to know where it's failing, you can put print statements in various places to see where it stops printing.

The serial console has many uses, and is an amazing tool overall for learning and programming!

The other feature of the serial connection is the Read-Evaluate-Print-Loop, or REPL. The REPL allows you to enter individual lines of code and have them run immediately. It's really handy if you're running into trouble with a particular program and can't figure out why. It's interactive so it's great for testing new ideas.

Entering the REPL

To use the REPL, you first need to be connected to the serial console. Once that connection has been established, you'll want to press CTRL+C.

If there is code running, in this case code measuring distance, it will stop and you'll see Press any key to enter the REPL. Use CTRL-D to reload. Follow those instructions, and press any key on your keyboard.

The Traceback (most recent call last): is telling you the last thing your board was doing before you pressed Ctrl + C and interrupted it. The KeyboardInterrupt is you pressing CTRL+C. This information can be handy when troubleshooting, but for now, don't worry about it. Just note that it is expected behavior.

If your code.py file is empty or does not contain a loop, it will show an empty output and Code done running.. There is no information about what your board was doing before you interrupted it because there is no code running.

If you have no code.py on your CIRCUITPY drive, you will enter the REPL immediately after pressing CTRL+C. Again, there is no information about what your board was doing before you interrupted it because there is no code running.

Regardless, once you press a key you'll see a >>> prompt welcoming you to the REPL!

If you have trouble getting to the >>> prompt, try pressing Ctrl + C a few more times.

The first thing you get from the REPL is information about your board.

This line tells you the version of CircuitPython you're using and when it was released. Next, it gives you the type of board you're using and the type of microcontroller the board uses. Each part of this may be different for your board depending on the versions you're working with.

This is followed by the CircuitPython prompt.

Interacting with the REPL

From this prompt you can run all sorts of commands and code. The first thing you'll do is run help(). This will tell you where to start exploring the REPL. To run code in the REPL, type it in next to the REPL prompt.

Type help() next to the prompt in the REPL.

Then press enter. You should then see a message.

First part of the message is another reference to the version of CircuitPython you're using. Second, a URL for the CircuitPython related project guides. Then... wait. What's this? To list built-in modules type `help("modules")`. Remember the modules you learned about while going through creating code? That's exactly what this is talking about! This is a perfect place to start. Let's take a look!

Type help("modules") into the REPL next to the prompt, and press enter.

This is a list of all the core modules built into CircuitPython, including board. Remember, board contains all of the pins on the board that you can use in your code. From the REPL, you are able to see that list!

Type import board into the REPL and press enter. It'll go to a new prompt. It might look like nothing happened, but that's not the case! If you recall, the import statement simply tells the code to expect to do something with that module. In this case, it's telling the REPL that you plan to do something with that module.

Next, type dir(board) into the REPL and press enter.

This is a list of all of the pins on your board that are available for you to use in your code. Each board's list will differ slightly depending on the number of pins available. Do you see LED? That's the pin you used to blink the red LED!

The REPL can also be used to run code. Be aware that any code you enter into the REPL isn't saved anywhere. If you're testing something new that you'd like to keep, make sure you have it saved somewhere on your computer as well!

Every programmer in every programming language starts with a piece of code that says, "Hello, World." You're going to say hello to something else. Type into the REPL:

print("Hello, CircuitPython!")

Then press enter.

That's all there is to running code in the REPL! Nice job!

You can write single lines of code that run stand-alone. You can also write entire programs into the REPL to test them. Remember that nothing typed into the REPL is saved.

There's a lot the REPL can do for you. It's great for testing new ideas if you want to see if a few new lines of code will work. It's fantastic for troubleshooting code by entering it one line at a time and finding out where it fails. It lets you see what modules are available and explore those modules.

Try typing more into the REPL to see what happens!

Returning to the Serial Console

When you're ready to leave the REPL and return to the serial console, simply press CTRL+D. This will reload your board and reenter the serial console. You will restart the program you had running before entering the REPL. In the console window, you'll see any output from the program you had running. And if your program was affecting anything visual on the board, you'll see that start up again as well.

You can return to the REPL at any time!

Each CircuitPython program you run needs to have a lot of information to work. The reason CircuitPython is so simple to use is that most of that information is stored in other files and works in the background. These files are called libraries. Some of them are built into CircuitPython. Others are stored on your CIRCUITPY drive in a folder called lib. Part of what makes CircuitPython so great is its ability to store code separately from the firmware itself. Storing code separately from the firmware makes it easier to update both the code you write and the libraries you depend.

Your board may ship with a lib folder already, it's in the base directory of the drive. If not, simply create the folder yourself. When you first install CircuitPython, an empty lib directory will be created for you.

CircuitPython libraries work in the same way as regular Python modules so the Python docs are an excellent reference for how it all should work. In Python terms, you can place our library files in the lib directory because it's part of the Python path by default.

One downside of this approach of separate libraries is that they are not built in. To use them, one needs to copy them to the CIRCUITPY drive before they can be used. Fortunately, there is a library bundle.

The bundle and the library releases on GitHub also feature optimized versions of the libraries with the .mpy file extension. These files take less space on the drive and have a smaller memory footprint as they are loaded.

Due to the regular updates and space constraints, Adafruit does not ship boards with the entire bundle. Therefore, you will need to load the libraries you need when you begin working with your board. You can find example code in the guides for your board that depends on external libraries.

Either way, as you start to explore CircuitPython, you'll want to know how to get libraries on board.

The Adafruit Learn Guide Project Bundle

The quickest and easiest way to get going with a project from the Adafruit Learn System is by utilising the Project Bundle. Most guides now have a Download Project Bundle button available at the top of the full code example embed. This button downloads all the necessary files, including images, etc., to get the guide project up and running. Simply click, open the resulting zip, copy over the right files, and you're good to go!

The first step is to find the Download Project Bundle button in the guide you're working on.

The Download Project Bundle button downloads a zip file. This zip contains a series of directories, nested within which is the code.py, any applicable assets like images or audio, and the lib/ folder containing all the necessary libraries. The following zip was downloaded from the Piano in the Key of Lime guide.

When you open the zip, you'll find some nested directories. Navigate through them until you find what you need. You'll eventually find a directory for your CircuitPython version (in this case, 7.x). In the version directory, you'll find the file and directory you need: code.py and lib/. Once you find the content you need, you can copy it all over to your CIRCUITPY drive, replacing any files already on the drive with the files from the freshly downloaded zip.

Once you copy over all the relevant files, the project should begin running! If you find that the project is not running as expected, make sure you've copied ALL of the project files onto your microcontroller board.

That's all there is to using the Project Bundle!

The Adafruit CircuitPython Library Bundle

Adafruit provides CircuitPython libraries for much of the hardware they provide, including sensors, breakouts and more. To eliminate the need for searching for each library individually, the libraries are available together in the Adafruit CircuitPython Library Bundle. The bundle contains all the files needed to use each library.

Downloading the Adafruit CircuitPython Library Bundle

You can download the latest Adafruit CircuitPython Library Bundle release by clicking the button below. The libraries are being constantly updated and improved, so you'll always want to download the latest bundle. 

Match up the bundle version with the version of CircuitPython you are running. For example, you would download the 6.x library bundle if you're running any version of CircuitPython 6, or the 7.x library bundle if you're running any version of CircuitPython 7, etc. If you mix libraries with major CircuitPython versions, you will get incompatible mpy errors due to changes in library interfaces possible during major version changes.

Download the bundle version that matches your CircuitPython firmware version. If you don't know the version, check the version info in boot_out.txt file on the CIRCUITPY drive, or the initial prompt in the CircuitPython REPL. For example, if you're running v7.0.0, download the 7.x library bundle.

There's also a py bundle which contains the uncompressed python files, you probably don't want that unless you are doing advanced work on libraries.

The CircuitPython Community Library Bundle

The CircuitPython Community Library Bundle is made up of libraries written and provided by members of the CircuitPython community. These libraries are often written when community members encountered hardware not supported in the Adafruit Bundle, or to support a personal project. The authors all chose to submit these libraries to the Community Bundle make them available to the community.

These libraries are maintained by their authors and are not supported by Adafruit. As you would with any library, if you run into problems, feel free to file an issue on the GitHub repo for the library. Bear in mind, though, that most of these libraries are supported by a single person and you should be patient about receiving a response. Remember, these folks are not paid by Adafruit, and are volunteering their personal time when possible to provide support.

Downloading the CircuitPython Community Library Bundle

You can download the latest CircuitPython Community Library Bundle release by clicking the button below. The libraries are being constantly updated and improved, so you'll always want to download the latest bundle.

The link takes you to the latest release of the CircuitPython Community Library Bundle on GitHub. There are multiple versions of the bundle available. Download the bundle version that matches your CircuitPython firmware version. If you don't know the version, check the version info in boot_out.txt file on the CIRCUITPY drive, or the initial prompt in the CircuitPython REPL. For example, if you're running v7.0.0, download the 7.x library bundle.

Understanding the Bundle

After downloading the zip, extract its contents. This is usually done by double clicking on the zip. On Mac OSX, it places the file in the same directory as the zip.

Open the bundle folder. Inside you'll find two information files, and two folders. One folder is the lib bundle, and the other folder is the examples bundle.

Now open the lib folder. When you open the folder, you'll see a large number of .mpy files, and folders.

Example Files

All example files from each library are now included in the bundles in an examples directory (as seen above), as well as an examples-only bundle. These are included for two main reasons:

• Allow for quick testing of devices.
• Provide an example base of code, that is easily built upon for individualized purposes.

Copying Libraries to Your Board

First open the lib folder on your CIRCUITPY drive. Then, open the lib folder you extracted from the downloaded zip. Inside you'll find a number of folders and .mpy files. Find the library you'd like to use, and copy it to the lib folder on CIRCUITPY.

If the library is a directory with multiple .mpy files in it, be sure to copy the entire folder to CIRCUITPY/lib.

This also applies to example files. Open the examples folder you extracted from the downloaded zip, and copy the applicable file to your CIRCUITPY drive. Then, rename it to code.py to run it.

Understanding Which Libraries to Install

You now know how to load libraries on to your CircuitPython-compatible microcontroller board. You may now be wondering, how do you know which libraries you need to install? Unfortunately, it's not always straightforward. Fortunately, there is an obvious place to start, and a relatively simple way to figure out the rest. First up: the best place to start.

When you look at most CircuitPython examples, you'll see they begin with one or more import statements. These typically look like the following:

• import library_or_module

However, import statements can also sometimes look like the following:

• from library_or_module import name
• from library_or_module.subpackage import name
• from library_or_module import name as local_name

They can also have more complicated formats, such as including a try / except block, etc.

The important thing to know is that an import statement will always include the name of the module or library that you're importing.

Therefore, the best place to start is by reading through the import statements.

Here is an example import list for you to work with in this section. There is no setup or other code shown here, as the purpose of this section involves only the import list.

Keep in mind, not all imported items are libraries. Some of them are almost always built-in CircuitPython modules. How do you know the difference? Time to visit the REPL.

In the Interacting with the REPL section on The REPL page in this guide, the help("modules") command is discussed. This command provides a list of all of the built-in modules available in CircuitPython for your board. So, if you connect to the serial console on your board, and enter the REPL, you can run help("modules") to see what modules are available for your board. Then, as you read through the import statements, you can, for the purposes of figuring out which libraries to load, ignore the statement that import modules.

The following is the list of modules built into CircuitPython for the Feather RP2040. Your list may look similar or be anything down to a significant subset of this list for smaller boards.

Now that you know what you're looking for, it's time to read through the import statements. The first two, time and board, are on the modules list above, so they're built-in.

The next one, neopixel, is not on the module list. That means it's your first library! So, you would head over to the bundle zip you downloaded, and search for neopixel. There is a neopixel.mpy file in the bundle zip. Copy it over to the lib folder on your CIRCUITPY drive. The following one, adafruit_lis3dh, is also not on the module list. Follow the same process for adafruit_lis3dh, where you'll find adafruit_lis3dh.mpy, and copy that over.

The fifth one is usb_hid, and it is in the modules list, so it is built in. Often all of the built-in modules come first in the import list, but sometimes they don't! Don't assume that everything after the first library is also a library, and verify each import with the modules list to be sure. Otherwise, you'll search the bundle and come up empty!

The final two imports are not as clear. Remember, when import statements are formatted like this, the first thing after the from is the library name. In this case, the library name is adafruit_hid. A search of the bundle will find an adafruit_hid folder. When a library is a folder, you must copy the entire folder and its contents as it is in the bundle to the lib folder on your CIRCUITPY drive. In this case, you would copy the entire adafruit_hid folder to your CIRCUITPY/lib folder.

Notice that there are two imports that begin with adafruit_hid. Sometimes you will need to import more than one thing from the same library. Regardless of how many times you import the same library, you only need to load the library by copying over the adafruit_hid folder once.

That is how you can use your example code to figure out what libraries to load on your CircuitPython-compatible board!

There are cases, however, where libraries require other libraries internally. The internally required library is called a dependency. In the event of library dependencies, the easiest way to figure out what other libraries are required is to connect to the serial console and follow along with the ImportError printed there. The following is a very simple example of an ImportError, but the concept is the same for any missing library.

Example: ImportError Due to Missing Library

If you choose to load libraries as you need them, or you're starting fresh with an existing example, you may end up with code that tries to use a library you haven't yet loaded.  This section will demonstrate what happens when you try to utilise a library that you don't have loaded on your board, and cover the steps required to resolve the issue.

This demonstration will only return an error if you do not have the required library loaded into the lib folder on your CIRCUITPY drive.

Let's use a modified version of the Blink example.

Save this file. Nothing happens to your board. Let's check the serial console to see what's going on.

You have an ImportError. It says there is no module named 'simpleio'. That's the one you just included in your code!

Click the link above to download the correct bundle. Extract the lib folder from the downloaded bundle file. Scroll down to find simpleio.mpy. This is the library file you're looking for! Follow the steps above to load an individual library file.

The LED starts blinking again! Let's check the serial console.

No errors! Excellent. You've successfully resolved an ImportError!

If you run into this error in the future, follow along with the steps above and choose the library that matches the one you're missing.

Library Install on Non-Express Boards

If you have an M0 non-Express board such as Trinket M0, Gemma M0, QT Py M0, or one of the M0 Trinkeys, you'll want to follow the same steps in the example above to install libraries as you need them. Remember, you don't need to wait for an ImportError if you know what library you added to your code. Open the library bundle you downloaded, find the library you need, and drag it to the lib folder on your CIRCUITPY drive.

You can still end up running out of space on your M0 non-Express board even if you only load libraries as you need them. There are a number of steps you can use to try to resolve this issue. You'll find suggestions on the Troubleshooting page.

Updating CircuitPython Libraries and Examples

Libraries and examples are updated from time to time, and it's important to update the files you have on your CIRCUITPY drive.

To update a single library or example, follow the same steps above. When you drag the library file to your lib folder, it will ask if you want to replace it. Say yes. That's it!

A new library bundle is released every time there's an update to a library. Updates include things like bug fixes and new features. It's important to check in every so often to see if the libraries you're using have been updated.

CircUp CLI Tool

There is a command line interface (CLI) utility called CircUp that can be used to easily install and update libraries on your device. Follow the directions on the install page within the CircUp learn guide. Once you've got it installed you run the command circup update in a terminal to interactively update all libraries on the connected CircuitPython device. See the usage page in the CircUp guide for a full list of functionality

From time to time, you will run into issues when working with CircuitPython. Here are a few things you may encounter and how to resolve them.

Always Run the Latest Version of CircuitPython and Libraries

As CircuitPython development continues and there are new releases, Adafruit will stop supporting older releases. You need to update to the latest CircuitPython..

You need to download the CircuitPython Library Bundle that matches your version of CircuitPython. Please update CircuitPython and then download the latest bundle.

As new versions of CircuitPython are released, Adafruit will stop providing the previous bundles as automatically created downloads on the Adafruit CircuitPython Library Bundle repo. If you must continue to use an earlier version, you can still download the appropriate version of mpy-cross from the particular release of CircuitPython on the CircuitPython repo and create your own compatible .mpy library files. However, it is best to update to the latest for both CircuitPython and the library bundle.

I have to continue using CircuitPython 5.x or earlier. Where can I find compatible libraries?

Adafruit is no longer building or supporting the CircuitPython 5.x or earlier library bundles. You are highly encourged to update CircuitPython to the latest version and use the current version of the libraries. However, if for some reason you cannot update, links to the previous bundles are available in the FAQ.

Bootloader (boardnameBOOT) Drive Not Present

You may have a different board.

Only Adafruit Express boards and the SAMD21 non-Express boards ship with the UF2 bootloader installed. The Feather M0 Basic, Feather M0 Adalogger, and similar boards use a regular Arduino-compatible bootloader, which does not show a boardnameBOOT drive.

MakeCode

If you are running a MakeCode program on Circuit Playground Express, press the reset button just once to get the CPLAYBOOT drive to show up. Pressing it twice will not work.

MacOS

DriveDx and its accompanything SAT SMART Driver can interfere with seeing the BOOT drive. See this forum post for how to fix the problem.

Windows 10

Did you install the Adafruit Windows Drivers package by mistake, or did you upgrade to Windows 10 with the driver package installed? You don't need to install this package on Windows 10 for most Adafruit boards. The old version (v1.5) can interfere with recognizing your device. Go to Settings -> Apps and uninstall all the "Adafruit" driver programs.

Windows 7 or 8.1

To use a CircuitPython-compatible board with Windows 7 or 8.1, you must install a driver. Installation instructions are available here.

It is recommended that you upgrade to Windows 10 if possible; an upgrade is probably still free for you. Check here.

You should now be done! Test by unplugging and replugging the board. You should see the CIRCUITPY drive, and when you double-click the reset button (single click on Circuit Playground Express running MakeCode), you should see the appropriate boardnameBOOT drive.

Let us know in the Adafruit support forums or on the Adafruit Discord if this does not work for you!

Windows Explorer Locks Up When Accessing boardnameBOOT Drive

On Windows, several third-party programs that can cause issues. The symptom is that you try to access the boardnameBOOT drive, and Windows or Windows Explorer seems to lock up. These programs are known to cause trouble:

• AIDA64: to fix, stop the program. This problem has been reported to AIDA64. They acquired hardware to test, and released a beta version that fixes the problem. This may have been incorporated into the latest release. Please let us know in the forums if you test this.
• Hard Disk Sentinel
• Kaspersky anti-virus: To fix, you may need to disable Kaspersky completely. Disabling some aspects of Kaspersky does not always solve the problem. This problem has been reported to Kaspersky.
• ESET NOD32 anti-virus: There have been problems with at least version 9.0.386.0, solved by uninstallation.

Copying UF2 to boardnameBOOT Drive Hangs at 0% Copied

On Windows, a Western DIgital (WD) utility that comes with their external USB drives can interfere with copying UF2 files to the boardnameBOOT drive. Uninstall that utility to fix the problem.

CIRCUITPY Drive Does Not Appear or Disappears Quickly

Kaspersky anti-virus can block the appearance of the CIRCUITPY drive. There has not yet been settings change discovered that prevents this. Complete uninstallation of Kaspersky fixes the problem.

Norton anti-virus can interfere with CIRCUITPY. A user has reported this problem on Windows 7. The user turned off both Smart Firewall and Auto Protect, and CIRCUITPY then appeared.

Sophos Endpoint security software can cause CIRCUITPY to disappear and the BOOT drive to reappear. It is not clear what causes this behavior.

Device Errors or Problems on Windows

Windows can become confused about USB device installations. This is particularly true of Windows 7 and 8.1. It is recommended that you upgrade to Windows 10 if possible; an upgrade is probably still free for you: see this link.

If not, try cleaning up your USB devices. Use Uwe Sieber's Device Cleanup Tool (on that page, scroll down to "Device Cleanup Tool"). Download and unzip the tool. Unplug all the boards and other USB devices you want to clean up. Run the tool as Administrator. You will see a listing like this, probably with many more devices. It is listing all the USB devices that are not currently attached.

Select all the devices you want to remove, and then press Delete. It is usually safe just to select everything. Any device that is removed will get a fresh install when you plug it in. Using the Device Cleanup Tool also discards all the COM port assignments for the unplugged boards. If you have used many Arduino and CircuitPython boards, you have probably seen higher and higher COM port numbers used, seemingly without end. This will fix that problem.

Serial Console in Mu Not Displaying Anything

There are times when the serial console will accurately not display anything, such as, when no code is currently running, or when code with no serial output is already running before you open the console. However, if you find yourself in a situation where you feel it should be displaying something like an error, consider the following.

Depending on the size of your screen or Mu window, when you open the serial console, the serial console panel may be very small. This can be a problem. A basic CircuitPython error takes 10 lines to display!

More complex errors take even more lines!

Therefore, if your serial console panel is five lines tall or less, you may only see blank lines or blank lines followed by Press any key to enter the REPL. Use CTRL-D to reload.. If this is the case, you need to either mouse over the top of the panel to utilise the option to resize the serial panel, or use the scrollbar on the right side to scroll up and find your message.

This applies to any kind of serial output whether it be error messages or print statements. So before you start trying to debug your problem on the hardware side, be sure to check that you haven't simply missed the serial messages due to serial output panel height.

code.py Restarts Constantly

CircuitPython will restart code.py if you or your computer writes to something on the CIRCUITPY drive. This feature is called auto-reload, and lets you test a change to your program immediately.

Some utility programs, such as backup, anti-virus, or disk-checking apps, will write to the CIRCUITPY as part of their operation. Sometimes they do this very frequently, causing constant restarts.

Acronis True Image and related Acronis programs on Windows are known to cause this problem. It is possible to prevent this by disabling the "Acronis Managed Machine Service Mini".

If you cannot stop whatever is causing the writes, you can disable auto-reload by putting this code in boot.py or code.py:

CircuitPython RGB Status Light

Nearly all CircuitPython-capable boards have a single NeoPixel or DotStar RGB LED on the board that indicates the status of CircuitPython. A few boards designed before CircuitPython existed, such as the Feather M0 Basic, do not.

Circuit Playground Express and Circuit Playground Bluefruit have multiple RGB LEDs, but do NOT have a status LED. The LEDs are all green when in the bootloader. In versions before 7.0.0, they do NOT indicate any status while running CircuitPython.

CircuitPython 7.0.0 and Later

The status LED blinks were changed in CircuitPython 7.0.0 in order to save battery power and simplify the blinks. These blink patterns will occur on single color LEDs when the board does not have any RGB LEDs. Speed and blink count also vary for this reason.

On start up, the LED will blink YELLOW multiple times for 1 second. Pressing the RESET button (or on Espressif, the BOOT button) during this time will restart the board and then enter safe mode. On Bluetooth capable boards, after the yellow blinks, there will be a set of faster blue blinks. Pressing reset during the BLUE blinks will clear Bluetooth information and start the device in discoverable mode, so it can be used with a BLE code editor.

Once started, CircuitPython will blink a pattern every 5 seconds when no user code is running to indicate why the code stopped:

• 1 GREEN blink: Code finished without error.
• 2 RED blinks: Code ended due to an exception. Check the serial console for details.
• 3 YELLOW blinks: CircuitPython is in safe mode. No user code was run. Check the serial console for safe mode reason.

When in the REPL, CircuitPython will set the status LED to WHITE. You can change the LED color from the REPL. The status indicator will not persist on non-NeoPixel or DotStar LEDs.

CircuitPython 6.3.0 and earlier

Here's what the colors and blinking mean:

• steady GREEN: code.py (or code.txt, main.py, or main.txt) is running
• pulsing GREEN: code.py (etc.) has finished or does not exist
• steady YELLOW at start up: (4.0.0-alpha.5 and newer) CircuitPython is waiting for a reset to indicate that it should start in safe mode
• pulsing YELLOW: Circuit Python is in safe mode: it crashed and restarted
• steady WHITE: REPL is running
• steady BLUE: boot.py is running

Colors with multiple flashes following indicate a Python exception and then indicate the line number of the error. The color of the first flash indicates the type of error:

• GREEN: IndentationError
• CYAN: SyntaxError
• WHITE: NameError
• ORANGE: OSError
• PURPLE: ValueError
• YELLOW: other error

These are followed by flashes indicating the line number, including place value. WHITE flashes are thousands' place, BLUE are hundreds' place, YELLOW are tens' place, and CYAN are one's place. So for example, an error on line 32 would flash YELLOW three times and then CYAN two times. Zeroes are indicated by an extra-long dark gap.

Serial console showing ValueError: Incompatible .mpy file

This error occurs when importing a module that is stored as a .mpy binary file that was generated by a different version of CircuitPython than the one its being loaded into. In particular, the mpy binary format changed between CircuitPython versions 6.x and 7.x, 2.x and 3.x, and 1.x and 2.x.

So, for instance, if you upgraded to CircuitPython 7.x from 6.x you’ll need to download a newer version of the library that triggered the error on import. All libraries are available in the Adafruit bundle.

CIRCUITPY Drive Issues

You may find that you can no longer save files to your CIRCUITPY drive. You may find that your CIRCUITPY stops showing up in your file explorer, or shows up as NO_NAME. These are indicators that your filesystem has issues. When the CIRCUITPY disk is not safely ejected before being reset by the button or being disconnected from USB, it may corrupt the flash drive. It can happen on Windows, Mac or Linux, though it is more common on Windows.

Be aware, if you have used Arduino to program your board, CircuitPython is no longer able to provide the USB services. You will need to reload CircuitPython to resolve this situation.

The easiest first step is to reload CircuitPython. Double-tap reset on the board so you get a boardnameBOOT drive rather than a CIRCUITPY drive, and copy the latest version of CircuitPython (.uf2) back to the board. This may restore CIRCUITPY functionality.

If reloading CircuitPython does not resolve your issue, the next step is to try putting the board into safe mode.

Safe Mode

Whether you've run into a situation where you can no longer edit your code.py on your CIRCUITPY drive, your board has gotten into a state where CIRCUITPY is read-only, or you have turned off the CIRCUITPY drive altogether, safe mode can help.

Safe mode in CircuitPython does not run any user code on startup, and disables auto-reload. This means a few things. First, safe mode bypasses any code in boot.py (where you can set CIRCUITPY read-only or turn it off completely). Second, it does not run the code in code.py. And finally, it does not automatically soft-reload when data is written to the CIRCUITPY drive.

Therefore, whatever you may have done to put your board in a non-interactive state, safe mode gives you the opportunity to correct it without losing all of the data on the CIRCUITPY drive.

Entering Safe Mode in CircuitPython 7.x and Later

To enter safe mode when using CircuitPython 7.x, plug in your board or hit reset (highlighted in red above). Immediately after the board starts up or resets, it waits 1000ms. On some boards, the onboard status LED will blink yellow during that time. If you press reset during that 1000ms, the board will start up in safe mode. It can be difficult to react to the yellow LED, so you may want to think of it simply as a "slow" double click of the reset button. (Remember, a fast double click of reset enters the bootloader.)

Entering Safe Mode in CircuitPython 6.x

To enter safe mode when using CircuitPython 6.x, plug in your board or hit reset (highlighted in red above). Immediately after the board starts up or resets, it waits 700ms. On some boards, the onboard status LED (highlighted in green above) will turn solid yellow during this time. If you press reset during that 700ms, the board will start up in safe mode. It can be difficult to react to the yellow LED, so you may want to think of it simply as a slow double click of the reset button. (Remember, a fast double click of reset enters the bootloader.)

In Safe Mode

Once you've entered safe mode successfully in CircuitPython 6.x, the LED will pulse yellow.

If you successfully enter safe mode on CircuitPython 7.x, the LED will intermittently blink yellow three times.

If you connect to the serial console, you'll find the following message.

You can now edit the contents of the CIRCUITPY drive. Remember, your code will not run until you press the reset button, or unplug and plug in your board, to get out of safe mode.

At this point, you'll want to remove any user code in code.py and, if present, the boot.py file from CIRCUITPY. Once removed, tap the reset button, or unplug and plug in your board, to restart CircuitPython. This will restart the board and may resolve your drive issues. If resolved, you can begin coding again as usual.

If safe mode does not resolve your issue, the board must be completely erased and CircuitPython must be reloaded onto the board.

To erase CIRCUITPY: storage.erase_filesystem()

CircuitPython includes a built-in function to erase and reformat the filesystem. If you have a version of CircuitPython older than 2.3.0 on your board, you can update to the newest version to do this.

1. Connect to the CircuitPython REPL using Mu or a terminal program.
2. Type the following into the REPL:

CIRCUITPY will be erased and reformatted, and your board will restart. That's it!

Erase CIRCUITPY Without Access to the REPL

If you can't access the REPL, or you're running a version of CircuitPython previous to 2.3.0 and you don't want to upgrade, there are options available for some specific boards.

The options listed below are considered to be the "old way" of erasing your board. The method shown above using the REPL is highly recommended as the best method for erasing your board.

For the specific boards listed below:

If the board you are trying to erase is listed below, follow the steps to use the file to erase your board.

       1.  Download the correct erase file:

       2.  Double-click the reset button on the board to bring up the boardnameBOOT drive.
       3.  Drag the erase .uf2 file to the boardnameBOOT drive.
       4.  The status LED will turn yellow or blue, indicating the erase has started.
       5.  After approximately 15 seconds, the status LED will light up green. On the NeoTrellis M4 this is the first NeoPixel on the grid
       6.  Double-click the reset button on the board to bring up the boardnameBOOT drive.
       7.  Drag the appropriate latest release of CircuitPython .uf2 file to the boardnameBOOT drive.

It should reboot automatically and you should see CIRCUITPY in your file explorer again.

If the LED flashes red during step 5, it means the erase has failed. Repeat the steps starting with 2.

If you haven't already downloaded the latest release of CircuitPython for your board, check out the installation page. You'll also need to load your code and reinstall your libraries!

For SAMD21 non-Express boards that have a UF2 bootloader:

Any SAMD21-based microcontroller that does not have external flash available is considered a SAMD21 non-Express board. Non-Express boards that have a UF2 bootloader include Trinket M0, GEMMA M0, QT Py M0, and the SAMD21-based Trinkey boards.

If you are trying to erase a SAMD21 non-Express board, follow these steps to erase your board.

       1.  Download the erase file:

       2.  Double-click the reset button on the board to bring up the boardnameBOOT drive.
       3.  Drag the erase .uf2 file to the boardnameBOOT drive.
       4.  The boot LED will start flashing again, and the boardnameBOOT drive will reappear.
       5.  Drag the appropriate latest release CircuitPython .uf2 file to the boardnameBOOT drive.

It should reboot automatically and you should see CIRCUITPY in your file explorer again.

If you haven't already downloaded the latest release of CircuitPython for your board, check out the installation page YYou'll also need to load your code and reinstall your libraries!

For SAMD21 non-Express boards that do not have a UF2 bootloader:

Any SAMD21-based microcontroller that does not have external flash available is considered a SAMD21 non-Express board. Non-Express boards that do not have a UF2 bootloader include the Feather M0 Basic Proto, Feather Adalogger, or the Arduino Zero.

If you are trying to erase a non-Express board that does not have a UF2 bootloader, follow these directions to reload CircuitPython using bossac, which will erase and re-create CIRCUITPY.

Running Out of File Space on SAMD21 Non-Express Boards

Any SAMD21-based microcontroller that does not have external flash available is considered a SAMD21 non-Express board. This includes boards like the Trinket M0, GEMMA M0, QT Py M0, and the SAMD21-based Trinkey boards.

The file system on the board is very tiny. (Smaller than an ancient floppy disk.) So, its likely you'll run out of space but don't panic! There are a number of ways to free up space.

Delete something!

The simplest way of freeing up space is to delete files from the drive. Perhaps there are libraries in the lib folder that you aren't using anymore or test code that isn't in use. Don't delete the lib folder completely, though, just remove what you don't need.

The board ships with the Windows 7 serial driver too! Feel free to delete that if you don't need it or have already installed it. It's ~12KiB or so.

Use tabs

One unique feature of Python is that the indentation of code matters. Usually the recommendation is to indent code with four spaces for every indent. In general, that is recommended too. However, one trick to storing more human-readable code is to use a single tab character for indentation. This approach uses 1/4 of the space for indentation and can be significant when you're counting bytes.

On MacOS?

MacOS loves to generate hidden files. Luckily you can disable some of the extra hidden files that macOS adds by running a few commands to disable search indexing and create zero byte placeholders. Follow the steps below to maximize the amount of space available on macOS.

Prevent & Remove MacOS Hidden Files

First find the volume name for your board.  With the board plugged in run this command in a terminal to list all the volumes:

Look for a volume with a name like CIRCUITPY (the default for CircuitPython).  The full path to the volume is the /Volumes/CIRCUITPY path.

Now follow the steps from this question to run these terminal commands that stop hidden files from being created on the board:

Replace /Volumes/CIRCUITPY in the commands above with the full path to your board's volume if it's different.  At this point all the hidden files should be cleared from the board and some hidden files will be prevented from being created.

Alternatively, with CircuitPython 4.x and above, the special files and folders mentioned above will be created automatically if you erase and reformat the filesystem. WARNING: Save your files first! Do this in the REPL:

>>> import storage
>>> storage.erase_filesystem()

However there are still some cases where hidden files will be created by MacOS.  In particular if you copy a file that was downloaded from the internet it will have special metadata that MacOS stores as a hidden file.  Luckily you can run a copy command from the terminal to copy files without this hidden metadata file.  See the steps below.

Copy Files on MacOS Without Creating Hidden Files

Once you've disabled and removed hidden files with the above commands on macOS you need to be careful to copy files to the board with a special command that prevents future hidden files from being created.  Unfortunately you cannot use drag and drop copy in Finder because it will still create these hidden extended attribute files in some cases (for files downloaded from the internet, like Adafruit's modules).

To copy a file or folder use the -X option for the cp command in a terminal.  For example to copy a file_name.mpy file to the board use a command like:

(Replace file_name.mpy with the name of the file you want to copy.)

Or to copy a folder and all of the files and folders contained within, use a command like:

If you are copying to the lib folder, or another folder, make sure it exists before copying.

Other MacOS Space-Saving Tips

If you'd like to see the amount of space used on the drive and manually delete hidden files here's how to do so. First, move into the Volumes/ directory with cd /Volumes/, and then list the amount of space used on the CIRCUITPY drive with the df command.

That's not very much space left! The next step is to show a list of the files currently on the CIRCUITPY drive, including the hidden files, using the ls command. You cannot use Finder to do this, you must do it via command line!

There are a few of the hidden files that MacOS loves to generate, all of which begin with a ._ before the file name. Remove the ._ files using the rm command. You can remove them all once by running rm CIRCUITPY/._*. The * acts as a wildcard to apply the command to everything that begins with ._ at the same time.

Finally, you can run df again to see the current space used.

Nice! You have 12Ki more than before! This space can now be used for libraries and code!

Device Locked Up or Boot Looping

In rare cases, it may happen that something in your code.py or boot.py files causes the device to get locked up, or even go into a boot loop. A boot loop occurs when the board reboots repeatedly and never fully loads. These are not caused by your everyday Python exceptions, typically it's the result of a deeper problem within CircuitPython. In this situation, it can be difficult to recover your device if CIRCUITPY is not allowing you to modify the code.py or boot.py files. Safe mode is one recovery option. When the device boots up in safe mode it will not run the code.py or boot.py scripts, but will still connect the CIRCUITPY drive so that you can remove or modify those files as needed.

The method used to manually enter safe mode can be different for different devices. It is also very similar to the method used for getting into bootloader mode, which is a different thing. So it can take a few tries to get the timing right. If you end up in bootloader mode, no problem, you can try again without needing to do anything else.

For most devices:
Press the reset button, and then when the RGB status LED blinks yellow, press the reset button again. Since your reaction time may not be that fast, try a "slow" double click, to catch the yellow LED on the second click.

For ESP32-S2 based devices:
Press and release the reset button, then press and release the boot button about 3/4 of a second later.

Refer to the diagrams above for boot sequence details.

A lot of our boards can be used with multiple programming languages. For example, the Circuit Playground Express can be used with MakeCode, Code.org CS Discoveries, CircuitPython and Arduino.

Maybe you tried CircuitPython and want to go back to MakeCode or Arduino? Not a problem. You can always remove or reinstall CircuitPython whenever you want! Heck, you can change your mind every day!

There is nothing to uninstall. CircuitPython is "just another program" that is loaded onto your board. You simply load another program (Arduino or MakeCode) and it will overwrite CircuitPython.

Backup Your Code

Before replacing CircuitPython, don't forget to make a backup of the code you have on the CIRCUITPY drive. That means your code.py any other files, the lib folder etc. You may lose these files when you remove CircuitPython, so backups are key! Just drag the files to a folder on your laptop or desktop computer like you would with any USB drive.

Moving Circuit Playground Express to MakeCode

On the Circuit Playground Express (this currently does NOT apply to Circuit Playground Bluefruit), if you want to go back to using MakeCode, it's really easy. Visit makecode.adafruit.com and find the program you want to upload. Click Download to download the .uf2 file that is generated by MakeCode.

Now double-click your CircuitPython board until you see the onboard LED(s) turn green and the ...BOOT directory shows up.

Then find the downloaded MakeCode .uf2 file and drag it to the CPLAYBOOT drive.

Your MakeCode is now running and CircuitPython has been removed. Going forward you only have to single click the reset button to get to CPLAYBOOT. This is an idiosyncrasy of MakeCode.

Moving to Arduino

If you want to use Arduino instead, you just use the Arduino IDE to load an Arduino program. Here's an example of uploading a simple "Blink" Arduino program, but you don't have to use this particular program.

Start by plugging in your board, and double-clicking reset until you get the green onboard LED(s).

Within Arduino IDE, select the matching board, say Circuit Playground Express.

Select the correct matching Port:

Create a new simple Blink sketch example:

Make sure the LED(s) are still green, then click Upload to upload Blink. Once it has uploaded successfully, the serial Port will change so re-select the new Port!

Once Blink is uploaded you should no longer need to double-click to enter bootloader mode. Arduino will automatically reset when you upload.

CircuitPython is a programming language that's super simple to get started with and great for learning. It runs on microcontrollers and works out of the box. You can plug it in and get started with any text editor. The best part? CircuitPython comes with an amazing, supportive community.

Everyone is welcome! CircuitPython is Open Source. This means it's available for anyone to use, edit, copy and improve upon. This also means CircuitPython becomes better because of you being a part of it. Whether this is your first microcontroller board or you're a seasoned software engineer, you have something important to offer the Adafruit CircuitPython community. This page highlights some of the many ways you can be a part of it!

Adafruit Discord

The Adafruit Discord server is the best place to start. Discord is where the community comes together to volunteer and provide live support of all kinds. From general discussion to detailed problem solving, and everything in between, Discord is a digital maker space with makers from around the world.

There are many different channels so you can choose the one best suited to your needs. Each channel is shown on Discord as "#channelname". There's the #help-with-projects channel for assistance with your current project or help coming up with ideas for your next one. There's the #show-and-tell channel for showing off your newest creation. Don't be afraid to ask a question in any channel! If you're unsure, #general is a great place to start. If another channel is more likely to provide you with a better answer, someone will guide you.

The help with CircuitPython channel is where to go with your CircuitPython questions. #help-with-circuitpython is there for new users and developers alike so feel free to ask a question or post a comment! Everyone of any experience level is welcome to join in on the conversation. Your contributions are important! The #circuitpython-dev channel is available for development discussions as well.

The easiest way to contribute to the community is to assist others on Discord. Supporting others doesn't always mean answering questions. Join in celebrating successes! Celebrate your mistakes! Sometimes just hearing that someone else has gone through a similar struggle can be enough to keep a maker moving forward.

The Adafruit Discord is the 24x7x365 hackerspace that you can bring your granddaughter to.

Visit https://adafru.it/discord to sign up for Discord. Everyone is looking forward to meeting you!

CircuitPython.org

Beyond the Adafruit Learn System, which you are viewing right now, the best place to find information about CircuitPython is circuitpython.org. Everything you need to get started with your new microcontroller and beyond is available. You can do things like download CircuitPython for your microcontroller or download the latest CircuitPython Library bundle, or check out which single board computers support Blinka. You can also get to various other CircuitPython related things like Awesome CircuitPython or the Python for Microcontrollers newsletter. This is all incredibly useful, but it isn't necessarily community related. So why is it included here? The Contributing page.

CircuitPython itself is written in C. However, all of the Adafruit CircuitPython libraries are written in Python. If you're interested in contributing to CircuitPython on the Python side of things, check out circuitpython.org/contributing. You'll find information pertaining to every Adafruit CircuitPython library GitHub repository, giving you the opportunity to join the community by finding a contributing option that works for you.

Note the date on the page next to Current Status for:

If you submit any contributions to the libraries, and do not see them reflected on the Contributing page, it could be that the job that checks for new updates hasn't yet run for today. Simply check back tomorrow!

Now, a look at the different options.

Pull Requests

The first tab you'll find is a list of open pull requests.

GitHub pull requests, or PRs, are opened when folks have added something to an Adafruit CircuitPython library GitHub repo, and are asking for Adafruit to add, or merge, their changes into the main library code. For PRs to be merged, they must first be reviewed. Reviewing is a great way to contribute! Take a look at the list of open pull requests, and pick one that interests you. If you have the hardware, you can test code changes. If you don't, you can still check the code updates for syntax. In the case of documentation updates, you can verify the information, or check it for spelling and grammar. Once you've checked out the update, you can leave a comment letting us know that you took a look. Once you've done that for a while, and you're more comfortable with it, you can consider joining the CircuitPythonLibrarians review team. The more reviewers we have, the more authors we can support. Reviewing is a crucial part of an open source ecosystem, CircuitPython included.

Open Issues

The second tab you'll find is a list of open issues.

GitHub issues are filed for a number of reasons, including when there is a bug in the library or example code, or when someone wants to make a feature request. Issues are a great way to find an opportunity to contribute directly to the libraries by updating code or documentation. If you're interested in contributing code or documentation, take a look at the open issues and find one that interests you.

If you're not sure where to start, you can search the issues by label. Labels are applied to issues to make the goal easier to identify at a first glance, or to indicate the difficulty level of the issue. Click on the dropdown next to "Sort by issue labels" to see the list of available labels, and click on one to choose it.

If you're new to everything, new to contributing to open source, or new to contributing to the CircuitPython project, you can choose "Good first issue". Issues with that label are well defined, with a finite scope, and are intended to be easy for someone new to figure out.

If you're looking for something a little more complicated, consider "Bug" or "Enhancement". The Bug label is applied to issues that pertain to problems or failures found in the library. The Enhancement label is applied to feature requests.

Don't let the process intimidate you. If you're new to Git and GitHub, there is a guide to walk you through the entire process. As well, there are always folks available on Discord to answer questions.

Library Infrastructure Issues

The third tab you'll find is a list of library infrastructure issues.

This section is generated by a script that runs checks on the libraries, and then reports back where there may be issues. It is made up of a list of subsections each containing links to the repositories that are experiencing that particular issue. This page is available mostly for internal use, but you may find some opportunities to contribute on this page. If there's an issue listed that sounds like something you could help with, mention it on Discord, or file an issue on GitHub indicating you're working to resolve that issue. Others can reply either way to let you know what the scope of it might be, and help you resolve it if necessary.

CircuitPython Localization

The fourth tab you'll find is the CircuitPython Localization tab.

If you speak another language, you can help translate CircuitPython! The translations apply to informational and error messages that are within the CircuitPython core. It means that folks who do not speak English have the opportunity to have these messages shown to them in their own language when using CircuitPython. This is incredibly important to provide the best experience possible for all users. CircuitPython uses Weblate to translate, which makes it much simpler to contribute translations. You will still need to know some CircuitPython-specific practices and a few basics about coding strings, but as with any CircuitPython contributions, folks are there to help.

Regardless of your skill level, or how you want to contribute to the CircuitPython project, there is an opportunity available. The Contributing page is an excellent place to start!

Adafruit GitHub

Whether you're just beginning or are life-long programmer who would like to contribute, there are ways for everyone to be a part of the CircuitPython project. The CircuitPython core is written in C. The libraries are written in Python. GitHub is the best source of ways to contribute to the CircuitPython core, and the CircuitPython libraries. If you need an account, visit https://github.com/ and sign up.

If you're new to GitHub or programming in general, there are great opportunities for you. For the CircuitPython core, head over to the CircuitPython repository on GitHub, click on "Issues", and you'll find a list that includes issues labeled "good first issue". For the libraries, head over to the Contributing page Issues list, and use the drop down menu to search for "good first issue". These issues are things that have been identified as something that someone with any level of experience can help with. These issues include options like updating documentation, providing feedback, and fixing simple bugs. If you need help getting started with GitHub, there is an excellent guide on Contributing to CircuitPython with Git and GitHub.

Already experienced and looking for a challenge? Checkout the rest of either issues list and you'll find plenty of ways to contribute. You'll find all sorts of things, from new driver requests, to library bugs, to core module updates. There's plenty of opportunities for everyone at any level!

When working with or using CircuitPython or the CircuitPython libraries, you may find problems. If you find a bug, that's great! The team loves bugs! Posting a detailed issue to GitHub is an invaluable way to contribute to improving CircuitPython. For CircuitPython itself, file an issue here. For the libraries, file an issue on the specific library repository on GitHub. Be sure to include the steps to replicate the issue as well as any other information you think is relevant. The more detail, the better!

Testing new software is easy and incredibly helpful. Simply load the newest version of CircuitPython or a library onto your CircuitPython hardware, and use it. Let us know about any problems you find by posting a new issue to GitHub. Software testing on both stable and unstable releases is a very important part of contributing CircuitPython. The developers can't possibly find all the problems themselves! They need your help to make CircuitPython even better.

On GitHub, you can submit feature requests, provide feedback, report problems and much more. If you have questions, remember that Discord and the Forums are both there for help!

Adafruit Forums

The Adafruit Forums are the perfect place for support. Adafruit has wonderful paid support folks to answer any questions you may have. Whether your hardware is giving you issues or your code doesn't seem to be working, the forums are always there for you to ask. You need an Adafruit account to post to the forums. You can use the same account you use to order from Adafruit.

While Discord may provide you with quicker responses than the forums, the forums are a more reliable source of information. If you want to be certain you're getting an Adafruit-supported answer, the forums are the best place to be.

There are forum categories that cover all kinds of topics, including everything Adafruit. The Adafruit CircuitPython category under "Supported Products & Projects" is the best place to post your CircuitPython questions.

Be sure to include the steps you took to get to where you are. If it involves wiring, post a picture! If your code is giving you trouble, include your code in your post! These are great ways to make sure that there's enough information to help you with your issue.

You might think you're just getting started, but you definitely know something that someone else doesn't. The great thing about the forums is that you can help others too! Everyone is welcome and encouraged to provide constructive feedback to any of the posted questions. This is an excellent way to contribute to the community and share your knowledge!

Read the Docs

Read the Docs is a an excellent resource for a more detailed look at the CircuitPython core and the CircuitPython libraries. This is where you'll find things like API documentation and example code. For an in depth look at viewing and understanding Read the Docs, check out the CircuitPython Documentation page!

CircuitPython is designed to run on microcontrollers and allows you to interface with all kinds of sensors, inputs and other hardware peripherals. There are tons of guides showing how to wire up a circuit, and use CircuitPython to, for example, read data from a sensor, or detect a button press. Most CircuitPython code includes hardware setup which requires various modules, such as board or digitalio. You import these modules and then use them in your code. How does CircuitPython know to look for hardware in the specific place you connected it, and where do these modules come from?

This page explains both. You'll learn how CircuitPython finds the pins on your microcontroller board, including how to find the available pins for your board and what each pin is named. You'll also learn about the modules built into CircuitPython, including how to find all the modules available for your board.

CircuitPython Pins

When using hardware peripherals with a CircuitPython compatible microcontroller, you'll almost certainly be utilising pins. This section will cover how to access your board's pins using CircuitPython, how to discover what pins and board-specific objects are available in CircuitPython for your board, how to use the board-specific objects, and how to determine all available pin names for a given pin on your board.

import board

When you're using any kind of hardware peripherals wired up to your microcontroller board, the import list in your code will include import board. The board module is built into CircuitPython, and is used to provide access to a series of board-specific objects, including pins. Take a look at your microcontroller board. You'll notice that next to the pins are pin labels. You can always access a pin by its pin label. However, there are almost always multiple names for a given pin.

To see all the available board-specific objects and pins for your board, enter the REPL (>>>) and run the following commands:

Here is the output for the QT Py. You may have a different board, and this list will vary, based on the board.

The following pins have labels on the physical QT Py board: A0, A1, A2, A3, SDA, SCL, TX, RX, SCK, MISO, and MOSI. You see that there are many more entries available in board than the labels on the QT Py.

You can use the pin names on the physical board, regardless of whether they seem to be specific to a certain protocol.

For example, you do not have to use the SDA pin for I2C - you can use it for a button or LED.

On the flip side, there may be multiple names for one pin. For example, on the QT Py, pin A0 is labeled on the physical board silkscreen, but it is available in CircuitPython as both A0 and D0. For more information on finding all the names for a given pin, see the What Are All the Available Pin Names? section below.

The results of dir(board) for CircuitPython compatible boards will look similar to the results for the QT Py in terms of the pin names, e.g. A0, D0, etc. However, some boards, for example, the Metro ESP32-S2, have different styled pin names. Here is the output for the Metro ESP32-S2.

Note that most of the pins are named in an IO# style, such as IO1 and IO2. Those pins on the physical board are labeled only with a number, so an easy way to know how to access them in CircuitPython, is to run those commands in the REPL and find the pin naming scheme.

I2C, SPI, and UART

You'll also see there are often (but not always!) three special board-specific objects included: I2C, SPI, and UART - each one is for the default pin-set used for each of the three common protocol busses they are named for. These are called singletons.

What's a singleton? When you create an object in CircuitPython, you are instantiating ('creating') it. Instantiating an object means you are creating an instance of the object with the unique values that are provided, or "passed", to it.

For example, When you instantiate an I2C object using the busio module, it expects two pins: clock and data, typically SCL and SDA. It often looks like this:

Then, you pass the I2C object to a driver for the hardware you're using. For example, if you were using the TSL2591 light sensor and its CircuitPython library, the next line of code would be:

However, CircuitPython makes this simpler by including the I2C singleton in the board module. Instead of the two lines of code above, you simply provide the singleton as the I2C object. So if you were using the TSL2591 and its CircuitPython library, the two above lines of code would be replaced with:

This eliminates the need for the busio module, and simplifies the code. Behind the scenes, the board.I2C()  object is instantiated when you call it, but not before, and on subsequent calls, it returns the same object. Basically, it does not create an object until you need it, and provides the same object every time you need it. You can call board.I2C() as many times as you like, and it will always return the same object.

What Are All the Available Names?

Many pins on CircuitPython compatible microcontroller boards have multiple names, however, typically, there's only one name labeled on the physical board. So how do you find out what the other available pin names are? Simple, with the following script! Each line printed out to the serial console contains the set of names for a particular pin.

On a microcontroller board running CircuitPython, first, connect to the serial console.

In the example below, click the Download Project Bundle button below to download the necessary libraries and the code.py file in a zip file. Extract the contents of the zip file, open the directory CircuitPython_Essentials/Pin_Map_Script/ and then click on the directory that matches the version of CircuitPython you're using and copy the contents of that directory to your CIRCUITPY drive.

Your CIRCUITPY drive should now look similar to the following image:

Here is the result when this script is run on QT Py:

Each line represents a single pin. Find the line containing the pin name that's labeled on the physical board, and you'll find the other names available for that pin. For example, the first pin on the board is labeled A0. The first line in the output is board.A0 board.D0. This means that you can access pin A0 with both board.A0 and board.D0.

You'll notice there are two "pins" that aren't labeled on the board but appear in the list: board.NEOPIXEL and board.NEOPIXEL_POWER. Many boards have several of these special pins that give you access to built-in board hardware, such as an LED or an on-board sensor. The Qt Py only has one on-board extra piece of hardware, a NeoPixel LED, so there's only the one available in the list. But you can also control whether or not power is applied to the NeoPixel, so there's a separate pin for that.

That's all there is to figuring out the available names for a pin on a compatible microcontroller board in CircuitPython!

Microcontroller Pin Names

The pin names available to you in the CircuitPython board module are not the same as the names of the pins on the microcontroller itself. The board pin names are aliases to the microcontroller pin names. If you look at the datasheet for your microcontroller, you'll likely find a pinout with a series of pin names, such as "PA18" or "GPIO5". If you want to get to the actual microcontroller pin name in CircuitPython, you'll need the microcontroller.pin module. As with board, you can run dir(microcontroller.pin) in the REPL to receive a list of the microcontroller pin names.

CircuitPython Built-In Modules

There is a set of modules used in most CircuitPython programs. One or more of these modules is always used in projects involving hardware. Often hardware requires installing a separate library from the Adafruit CircuitPython Bundle. But, if you try to find board or digitalio in the same bundle, you'll come up lacking. So, where do these modules come from? They're built into CircuitPython! You can find an comprehensive list of built-in CircuitPython modules and the technical details of their functionality from CircuitPython here and the Python-like modules included here. However, not every module is available for every board due to size constraints or hardware limitations. How do you find out what modules are available for your board?

There are two options for this. You can check the support matrix, and search for your board by name. Or, you can use the REPL.

Plug in your board, connect to the serial console and enter the REPL. Type the following command.

That's it! You now know two ways to find all of the modules built into CircuitPython for your compatible microcontroller board.

CircuitPython comes 'with the kitchen sink' - a lot of the things you know and love about classic Python 3 (sometimes called CPython) already work. There are a few things that don't but we'll try to keep this list updated as we add more capabilities!

Thing That Are Built In and Work

Flow Control

All the usual if, elif, else, for, while work just as expected.

Math

import math will give you a range of handy mathematical functions.

>>> dir(math)
['__name__', 'e', 'pi', 'sqrt', 'pow', 'exp', 'log', 'cos', 'sin', 'tan', 'acos', 'asin', 'atan', 'atan2', 'ceil', 'copysign', 'fabs', 'floor', 'fmod', 'frexp', 'ldexp', 'modf', 'isfinite', 'isinf', 'isnan', 'trunc', 'radians', 'degrees']

CircuitPython supports 30-bit wide floating point values so you can use int and float whenever you expect.

Tuples, Lists, Arrays, and Dictionaries

You can organize data in (),  [], and {} including strings, objects, floats, etc.

Classes, Objects and Functions

We use objects and functions extensively in our libraries so check out one of our many examples like this MCP9808 library for class examples.

Lambdas

Yep! You can create function-functions with lambda just the way you like em:

>>> g = lambda x: x**2
>>> g(8)
64

Random Numbers

To obtain random numbers:

import random

random.random() will give a floating point number from 0 to 1.0.

random.randint(min, max) will give you an integer number between min and max.

Adafruit SAMD21 (M0) and SAMD51 (M4) boards feature an improved bootloader that makes it easier than ever to flash different code onto the microcontroller. This bootloader makes it easy to switch between Microsoft MakeCode, CircuitPython and Arduino.

Instead of needing drivers or a separate program for flashing (say, bossac, jlink or avrdude), one can simply drag a file onto a removable drive.

The format of the file is a little special. Due to 'operating system woes' you cannot just drag a binary or hex file (trust us, we tried it, it isn't cross-platform compatible). Instead, the format of the file has extra information to help the bootloader know where the data goes. The format is called UF2 (USB Flashing Format). Microsoft MakeCode generates UF2s for flashing and CircuitPython releases are also available as UF2. You can also create your own UF2s from binary files using uf2tool, available here.

The bootloader is also BOSSA compatible, so it can be used with the Arduino IDE which expects a BOSSA bootloader on ATSAMD-based boards

For more information about UF2, you can read a bunch more at the MakeCode blog, then check out the UF2 file format specification. 

Visit Adafruit's fork of the Microsoft UF2-samd bootloader GitHub repository for source code and releases of pre-built bootloaders on CircuitPython.org.

Entering Bootloader Mode

If the bootloader couldn't start, you will get a red NeoPixel LED.

That could mean that your USB cable is no good, it isn't connected to a computer, or maybe the drivers could not enumerate. Try a new USB cable first. Then try another port on your computer!

Once the bootloader is running, check your computer. You should see a USB Disk drive...

Once the bootloader is successfully connected you can open the drive and browse the virtual filesystem. This isn't the same filesystem as you use with CircuitPython or Arduino. It should have three files:

•  CURRENT.UF2 - The current contents of the microcontroller flash.
•  INDEX.HTM - Links to Microsoft MakeCode.
•  INFO_UF2.TXT - Includes bootloader version info. Please include it on bug reports.

Using the Mass Storage Bootloader

You may get an alert from the OS that the file is being copied without it's properties. You can just click Yes

You may also get get a complaint that the drive was ejected without warning. Don't worry about this. The drive only ejects once the bootloader has verified and completed the process of writing the new code

Using the BOSSA Bootloader

As mentioned before, the bootloader is also compatible with BOSSA, which is the standard method of updating boards when in the Arduino IDE. It is a command-line tool that can be used in any operating system. We won't cover the full use of the bossac tool, suffice to say it can do quite a bit! More information is available at ShumaTech.

Windows 7 Drivers

If you are running Windows 7 (or, goodness, something earlier?) You will need a Serial Port driver file. Windows 10 users do not need this so skip this step.

You can download our full driver package here:

Download and run the installer. We recommend just selecting all the serial port drivers available (no harm to do so) and installing them.

Verifying Serial Port in Device Manager

If you're running Windows, its a good idea to verify the device showed up. Open your Device Manager from the control panel and look under Ports (COM & LPT) for a device called Feather M0 or Circuit Playground or whatever!

If you see something like this, it means you did not install the drivers. Go back and try again, then remove and re-plug the USB cable for your board

Running bossac on the command line

If you are using the Arduino IDE, this step is not required. But sometimes you want to read/write custom binary files, say for loading CircuitPython or your own code. We recommend using bossac v 1.7.0 (or greater), which has been tested. The Arduino branch is most recommended.

You can download the latest builds here. The mingw32 version is for Windows, apple-darwin for Mac OSX and various linux options for Linux. Once downloaded, extract the files from the zip and open the command line to the directory with bossac.

With bossac version 1.9 or later, you must give an --offset parameter on the command line to specify where to start writing the firmware in flash memory. This parameter was added in bossac 1.8.0 with a default of 0x2000, but starting in 1.9, the default offset was changed to 0x0000, which is not what you want in most cases. If you omit the argument for bossac 1.9 or later, you will probably see a "Verify Failed" error from bossac.  Remember to change the option for -p or --port to match the port on your Mac.

Replace the filename below with the name of your downloaded .bin: it will vary based on your board!

Using bossac Versions 1.7.0, 1.8

There is no --offset parameter available. Use a command line like this:

bossac -p=/dev/cu.usbmodem14301 -e -w -v -R adafruit-circuitpython-boardname-version.bin

For example,

bossac -p=/dev/cu.usbmodem14301 -e -w -v -R adafruit-circuitpython-feather_m0_express-3.0.0.bin

Using bossac Versions 1.9 or Later

For M0 boards, which have an 8kB bootloader, you must specify -offset=0x2000, for example:

bossac -p=/dev/cu.usbmodem14301 -e -w -v -R --offset=0x2000 adafruit-circuitpython-feather_m0_express-3.0.0.bin

For M4 boards, which have a 16kB bootloader, you must specify -offset=0x4000, for example:

bossac -p=/dev/cu.usbmodem14301 -e -w -v -R --offset=0x4000 adafruit-circuitpython-feather_m4_express-3.0.0.bin

This will erase the chip, write the given file, verify the write and Reset the board. On Linux or MacOS you may need to run this command with sudo ./bossac ..., or add yourself to the dialout group first.

Updating the bootloader

The UF2 bootloader is relatively new and while we've done a ton of testing, it may contain bugs. Usually these bugs effect reliability rather than fully preventing the bootloader from working. If the bootloader is flaky then you can try updating the bootloader itself to potentially improve reliability.

If you're using MakeCode on a Mac, you need to make sure to upload the bootloader to avoid a serious problem with newer versions of MacOS. See instructions and more details here.

In general, you shouldn't have to update the bootloader! If you do think you're having bootloader related issues, please post in the forums or discord.

Updating the bootloader is as easy as flashing CircuitPython, Arduino or MakeCode. Simply enter the bootloader as above and then drag the update bootloader uf2 file below. This uf2 contains a program which will unlock the bootloader section, update the bootloader, and re-lock it. It will overwrite your existing code such as CircuitPython or Arduino so make sure everything is backed up!

After the file is copied over, the bootloader will be updated and appear again. The INFO_UF2.TXT file should show the newer version number inside.

For example:

UF2 Bootloader v2.0.0-adafruit.5 SFHWRO
Model: Metro M0
Board-ID: SAMD21G18A-Metro-v0

Lastly, reload your code from Arduino or MakeCode or flash the latest CircuitPython core.

Below are the latest updaters for various boards. The latest versions can always be found here.  Look for the update-bootloader... files, not the bootloader... files.

Getting Rid of Windows Pop-ups

If you do a lot of development on Windows with the UF2 bootloader, you may get annoyed by the constant "Hey you inserted a drive what do you want to do" pop-ups.

Making your own UF2

Making your own UF2 is easy! All you need is a .bin file of a program you wish to flash and the Python conversion script. Make sure that your program was compiled to start at 0x2000 (8k) for M0 boards or 0x4000 (16kB) for M4 boards. The bootloader takes up the first 8kB (M0) or 16kB (M4). CircuitPython's linker script is an example on how to do that.

Once you have a .bin file, you simply need to run the Python conversion script over it. Here is an example from the directory with uf2conv.py. This command will produce a firmware.uf2 file in the same directory as the source firmware.bin. The uf2 can then be flashed in the same way as above.

Installing the bootloader on a fresh/bricked board

If you somehow damaged your bootloader or maybe you have a new board, you can use a JLink to re-install it.

Here's a Learn Guide explaining how to fix the bootloader on a variety of boards using Atmel Studio

Here's a short writeup by turbinenreiter on how to do it for the Feather M4 (but adaptable to other boards)

TFT Screen Adhesive

The TFT screen on the HalloWing can become un-adhered if it is bumped or wiggled too much (or left in a very hot Jeep near the windshield in the hot southern California sun, for a totally hypothetical example that didn't necessarily happen to the author). Fixing this is pretty easy -- you can use double-stick tape, E6000 glue, or Sugru to fix it back in place. Here are some action photos of these fixes.

E6000 Glue

This is strong glue and is even better than tape if you plan to leave the HalloWing screen exposed on a costume without the lens and cover.

Sugru

You can go all out and create a protective frame using Sugru moldable silicone rubber.

Diagnostics

Want to see some stats on your HalloWing? Double-click the board's reset button to get to bootloader mode, then drag the HallowWingM4_Diagnostics.UF2 onto the HALLOWBOOT drive!

This UF2 is for the Hallowing M4

If you have the Hallowing M0 use this UF2