Programming and Debugging

There are currently two ways of programming the nRF9160 Feather. You can use the built-in bootloader or use an external programmer.

Bootloader use

Currently the nRF9160 Feather uses the MCUBoot bootloader which comes standard with the nRF Connect SDK. It is the recommended way to load new firmware onto your nRF9160 Feather.

Using newtmgr

Note: v5 of the nRF9160 Feather no longer requires manual usage of the Mode + Reset pins. It's placed into bootloader mode automatically using the zephyr-tools CLI.

newtmgr comes with the VSCode extension. To learn more about configuring it see the setup page for your respective operating system:

Pre-check: MCUBoot needs to be enabled in your project before you can use it! Make sure that you have CONFIG_BOOTLOADER_MCUBOOT=y in your prj.conf

  1. Put your nRF9160 Feather into DFU mode.
    1. Hold the MODE button
    2. Then tap the RST button while holding mode
    3. Hold the MODE button until the Blue LED illuminates
  2. Building your application if you haven't already with west build. It will create a folder called build. The file we care about is build/zephyr/app_update.bin
  3. Load the file using the Zephyr Tools: Load via Bootloader task. Here's what it's doing behind the scenes:

Note:

  1. The transfer process is limited to 1M BAUD. In most cases it takes <15 seconds to transfer application code.
  2. The nRF9160 Feather does not respond to newtmgr commands unless it's in DFU mode. (See step 1 above to get it into DFU mode.)
  3. If you're having trouble completing a transfer you can also add these parameters to the end:
-r 3 -t .1

Requirements for external programming and debugging

You can also use external programmers with the nRF9160 Feather. Here are the current supported external programmers:

  1. nRF5340-DK
  2. nRF9160-DK

Note: Most commercial J-Link programmers *should* work with the nRF9160 Feather. In particular the J-Link EDU Mini is a great choice if you're building non-commercial products. (Supports Cortex M33 processors. Untested at this time.)

🚨Important: the nRF52 and nRF51 based development kits do not work with the nRF9160 Feather!

👉Additionally, you'll also need a Tag-Connect cable. The nRF9160 Feather was designed to use the TC2030-CTX-NL. For more information check out Tag-Connect's product page.

Installing programmer software

In this next section, I'll be focusing on using the nRF5340-DK as the programmer of choice. These steps should not be different from other J-Link supported boards.

  1. Download your version of the nRF Command Line Tools

Windows

  1. Then, run the .exe that was downloaded. It will do all the heavy lifting for you.
  2. Run nrfjprog in a cmd window to make sure your install is working.
  3. You may also have to add JLink.exe to your path. It's the exact same proceedure as adding newtmgr except the path you're adding is C:\Program Files (x86)\SEGGER\JLink Paths updated for Windows for JLink
  4. Close/restart VSCode and your Command Prompt
  5. Run jlink.exe and make sure it opens. 
    > jlink.exe
SEGGER J-Link Commander V6.86f (Compiled Oct 23 2020 18:01:48)
DLL version V6.86f, compiled Oct 23 2020 18:00:13

Connecting to J-Link via USB...O.K.
Firmware: J-Link OB-K22-NordicSemi compiled Jan 21 2020 17:33:01
Hardware version: V1.00
S/N: 960083363
License(s): RDI, FlashBP, FlashDL, JFlash, GDB
VTref=3.300V


Type "connect" to establish a target connection, '?' for help
J-Link>

Mac

  1. First run nRF-Command-Line-Tools_10_9_0_OSX.pkg
  2. Once that install is done, run JLink_MacOSX_V680a.pkg
  3. Open a fresh terminal window and run nrfjprog and jlinkexe to make sure your install is complete.

Setting up the nRF5340-DK

Here are a couple of close up shots of how to connect the nRF5340-DK:

nrf53-dk-connected

nRF9160 Feather w/ USB

nRF9160 Feather w/ Tag Connect

nRF9160 Feather w/ Tag Connect Connected

I highly recommend you jump SB47 on your nRF5340-DK with some solder. This forces the debugger to think an external devices is permanently connected. If you're only doing external debugging, this is very useful.

nRF5340-DK jumper highlighted

After hooking things up, It's time to do a quick smoke test. Running nrfjprog -r in a terminal should show this result:

$ nrfjprog -r
Applying system reset.
Run.

Success!

Note: these pictures are with an early version of the nRF9160 Feather. The procedure is the same. The orientation of the Tag-Connect though is horizontal not vertical as seen in the pictures.

Securing the Bootloader

In order to deliver secure over the air updates we need to generate our own signing key. Here's the (simplified) process:

  • Change directories to /opt/nordic/ncs/v1.x.x/bootloader/mcuboot
  • Run the image generation script:
scripts/imgtool.py keygen -k your-key-rsa-2048.pem -t rsa-2048
  • Create a mcuboot.conf file in your app with the included contents:
CONFIG_BOOT_SIGNATURE_TYPE_RSA=y
CONFIG_BOOT_SIGNATURE_KEY_FILE="your-key-rsa-2048.pem"
  • In CMakeLists.txt add the following lines:
# MCUboot related
list(APPEND mcuboot_OVERLAY_CONFIG
  "${CMAKE_CURRENT_SOURCE_DIR}/mcuboot.conf"
  )

Then start a pristine build using west:

west build -b circuitdojo_feather_nrf9160ns -p

A pristine build  should happen before releasing that way the version data is up to date in the app_update.binIt's critical not to lose your-key-rsa-2048.pem as it's the key for (secure) OTA updates to succeed.

Programming with the nRF5340-DK

Programming with the nRF5340-DK is straight forward in Zephyr using west. Here's what it looks like:

west build -b circuitdojo_feather_nrf9160ns -p
west flash --runner nrfjprog
nrfjprog -r

In the above, i'm:

  1. Doing a pristine build of my application with the nRF9160 Feather as the target.
  2. Then flashing using the nrfjprog runner option. This is preferred for all J-Link boards.
  3. Resetting the board using nrfjprog -r. As of this writing, west does not reset the board after programming.

Getting a modem trace

Sometimes, you may be requested to get a modem trace of your device. This section will focus on helping you get one for your nRF9160 Feather.

In order to get a modem trace, the TX and RX pins on your board need to be free. You'll also need a UART to USB adapter of some type. I used an FTDI one that has each of the lines broken out.

  1. First, set your prj.conf to include the following lines:

    # Enable modem trace
CONFIG_BSD_LIBRARY_TRACE_ENABLED=y


# AT host library
CONFIG_UART_INTERRUPT_DRIVEN=y
CONFIG_AT_HOST_LIBRARY=y

    Note: version v1.5.x and newer uses this flag to enable modem tracing: CONFIG_NRF_MODEM_LIB_TRACE_ENABLED=y

  2. Then, create a folder in your project/sample called boards and add a new file called circuitdojo_feather_nrf9160ns.overlay We'll want to enable the UART1 interface on pins 23 and 24 like below:

    /*
* Copyright (c) 2020 Circuit Dojo LLC
*
* SPDX-License-Identifier: Apache-2.0
*/

&uart1 {
status = "okay";
current-speed = <115200>;
tx-pin = <24>;
rx-pin = <23>;
};

  3. Connect your USB to UART adatper. I've used clips from my logic analyzer to hold the wires in place. Connect the yellow wire to the TX on the board. Connect the orange wire to the RX on the board.

    img/programming-and-debugging/Screen_Shot_2020-09-16_at_4.42.58_PM.png

  4. Then, inside LTE link monitor enable the process with:

    AT%XMODEMTRACE=1,2
AT+CFUN=0

    This will perpetually save and keep modem traces on.

    Note: this step is no longer necessary.

  5. Then open the serial port in the modem trace App and click start.

    img/programming-and-debugging/Screen_Shot_2020-09-16_at_4.47.06_PM.png

  6. Then run your app as normal. You should see the Trace size go up little by little as connections are made, etc.

    img/programming-and-debugging/Screen_Shot_2020-09-16_at_5.12.48_PM.png

  7. Then grab the file according to the log output. For example: Tracefile created: /Users/jaredwolff/Library/Application Support/nrfconnect/pc-nrfconnect-tracecollector/trace-2020-09-16T20-47-19.741Z.bin

    For more information, check out Nordic's original article on the subject.

Debugging in Visual Code

Debugging your application is possible with Visual Code. You will need a J-Link programmer and a Tag Connect (TC2030-CTX-NL) cable for this process. Programmers include the nRF9160 DK, nRF532 DK, J-Link EDU (if your project is non-profit) and the standard commercial J-Link programmers.

Here's the process:

  1. Download and install your version of the nRF Command Line Tools

  2. Install the C/C++ Extension and the Cortex-Debug extensions. They're both very handy in development and debugging of Zephyr based projects. C/C++ Extension Cortex M Debug Extension

  3. If you don't have one already, create a .vscode folder in the root of your project. .vscode Fodler

  4. Create a file called launch.json. This is where we'll set up the configuration for debugging.

  5. Here's a real example of a config I was using to debug a project in OSX:

    {
  // Use IntelliSense to learn about possible attributes.
  // Hover to view descriptions of existing attributes.
  // For more information, visit: https://go.microsoft.com/fwlink/?linkid=830387
  "version": "0.2.0",
  "configurations": [
    {
      "name": "Cortex Debug",
      "cwd": "${workspaceRoot}",
      "executable": "${workspaceRoot}/pyrinas/applications/application/build/zephyr/zephyr.elf",
      "request": "launch",
      "type": "cortex-debug",
      "servertype": "jlink",
      "device": "nrf9160_xxAA",
      "interface": "swd",
      "armToolchainPath": "/Users/jaredwolff/gcc-arm-none-eabi-9-2019-q4-major/bin"
    }
  ]
}

    For folks on Windows you'll have to modify appropriately:

    {
 // Use IntelliSense to learn about possible attributes.
 // Hover to view descriptions of existing attributes.
 // For more information, visit: https://go.microsoft.com/fwlink/?linkid=830387
 "version": "0.2.0",
 "configurations": [
    {
        "name": "Cortex Debug",
        "cwd": "${workspaceRoot}",
        "executable": "${workspaceRoot}\\nrf9160-feather\\samples\\blinky\\build\\zephyr\\zephyr.elf",
        "request": "launch",
        "type": "cortex-debug",
        "servertype": "jlink",
        "device": "nrf9160_xxAA",
        "interface": "swd",
        "armToolchainPath": "C:\\Program Files (x86)\\GNU Tools Arm Embedded\\9 2019-q4-major\\bin"
    }
 ]
}

    Remember that workspaceRoot refers to the folder you have opened in VSCode. This will most likely be C:\Users\<your username>\ncs\v1.5.0\nfed\. (/opt/nordic/ncs/v1.5.0 for Mac) You will have to modify the "executable" entry to match the path of your zephyr.elf file.

  6. Change the executable path and the armToolchainPath to reflect your system. Make sure you point the executable option to the .elf file that gets produced during the compilation process.

  7. Next, go to your projects prj.conf and disable the bootloader by commenting out CONFIG_BOOTLOADER_MCUBOOT=y or changing the y to a n. As of this writing, disabling the bootloader is required as it prevents the debugging process from occuring.

  8. In prj.conf you'll also want to enable the CONFIG_DEBUG option. This disables compiler optimizations which makes the debug process hairy or impossible.

  9. Finally, program your project using west build && west flash

  10. At this point, if you've ever done any debugging in Visual Code, you should be able to follow the final steps to debug your application!

  11. Set some breakpoints in your code by pressing the line number you want. A red dot will appear where the breakpoint is set. Breakpoint

  12. Start debugging by clicking the debug icon on the left. Then click the play button in the top left. Debug menu Debug play button

  13. You can use the popup menu on the right to control traversal through your code. Popup debug control