Error flashing latest firmware to RAK4631

[cwlls]

Describe the Bug
When using the rnodeconf -a /dev/... command the flashing proceeds through each step, but fails in the EEPROM validation step, leaving the device in an unusable state. This is true with versions 1.82 and 1.81. Flashing with version 1.80 works as anticipated.

To Reproduce
Attempt to flash a RAK4631 device using firmware versions 1.82 or 1.81

Expected Behavior
The flashing process should succeed leaving the device in a functional, albeit "experimental" state (given the current status of the RAK4631 within RNode)

Logs & Screenshots

------------------------------------------------------------------------------
                               Installer Ready

Ok, that should be all the information we need. Please confirm the following
summary before proceeding. In the next step, the device will be flashed and
provisioned, so make sure that you are satisfied with your choices.

Serial port     : /dev/cu.usbmodem12101
Device type     : RAK4631 779 - 928 MHz
Platform        : NRF52
Device MCU      : Nordic Semiconductor nRF52840
Firmware file   : rnode_firmware_rak4631.zip

------------------------------------------------------------------------------

Is the above correct? [y/N] y
[11:40:22] Checking firmware file availability...
[11:40:22] Retrieving latest version info from https://unsigned.io/firmware/latest/?v=2.4.0&variant=rnode_firmware_rak4631.zip
[11:40:24] The selected firmware for this board is version 1.82
[11:40:24] Using existing firmware file: rnode_firmware_rak4631.zip for version 1.82
[11:40:24] Verifying firmware integrity...
[11:40:25] Decompressing firmware...
[11:40:25] Firmware decompressed
[11:40:25] Flashing RNode firmware to device on /dev/cu.usbmodem12101
Upgrading target on /dev/cu.usbmodem12101 with DFU package /Users/cwlls/.config/rnodeconf/update/1.82/rnode_firmware_rak4631.zip. Flow control is disabled, Dual bank, Touch 1200
########################################
########################################
########################################
########################################
########################################
########################################
########################################
########################################
########################################
########################################
########################################
##################################
Activating new firmware
Device programmed.
[11:40:53] Done flashing
[11:40:53] Waiting for NRF52 reset...
[11:40:59] Opening serial port /dev/cu.usbmodem12101...
[11:41:02] Device connected
[11:41:02] Current firmware version: 1.82
[11:41:02] Reading EEPROM...
[11:41:03] Clearing old EEPROM, this will take about 15 seconds...
[11:41:26] Waiting for NRF52 reset...
[11:41:44] Loading signing key...
[11:41:44] Bootstrapping device EEPROM...
[11:41:50] EEPROM written! Validating...
[11:41:50] Setting firmware checksum...
[11:41:57] Waiting for NRF52 reset...
[11:42:18] EEPROM was written, but validation failed. Check your settings.

System Information

• OS and version: macOS 15.3.2 (24D81)
• Python version: Python 3.13.2 (running within a venv)
• Program version: rnodeconf 2.4.0

Additional context
Meshtastic was previously installed on this device, but after a successful flash to version 1.80 of RNode, the issues persist.

[oherrala]

Same issue here with Heltec T114 with firmware version 1.83:

------------------------------------------------------------------------------
                               Installer Ready

Ok, that should be all the information we need. Please confirm the following
summary before proceeding. In the next step, the device will be flashed and
provisioned, so make sure that you are satisfied with your choices.

Serial port     : /dev/cu.usbmodem2101
Device type     : Heltec Mesh Node T114 850 - 950 MHz
Platform        : NRF52
Device MCU      : Nordic Semiconductor nRF52840
Firmware file   : rnode_firmware_heltec_t114.zip

------------------------------------------------------------------------------

Is the above correct? [y/N] y
[20:46:12] Checking firmware file availability...
[20:46:12] Retrieving latest version info from https://unsigned.io/firmware/latest/?v=2.4.0&variant=rnode_firmware_heltec_t114.zip
[20:46:12] The selected firmware for this board is version 1.83
[20:46:12] Using existing firmware file: rnode_firmware_heltec_t114.zip for version 1.83
[20:46:12] Verifying firmware integrity...
[20:46:13] Decompressing firmware...
[20:46:13] Firmware decompressed
[20:46:13] Flashing RNode firmware to device on /dev/cu.usbmodem2101
Upgrading target on /dev/cu.usbmodem2101 with DFU package /Users/oherrala/.config/rnodeconf/update/1.83/rnode_firmware_heltec_t114.zip. Flow control is disabled, Dual bank, Touch 1200
########################################
########################################
########################################
########################################
########################################
########################################
########################################
########################################
########################################
########################################
########################################
########################################
#######
Activating new firmware
Device programmed.
[20:46:43] Done flashing
[20:46:43] Waiting for NRF52 reset...
[20:46:49] Opening serial port /dev/cu.usbmodem2101...
[20:46:52] Device connected
[20:46:52] Current firmware version: 1.83
[20:46:52] Reading EEPROM...
[20:46:52] Clearing old EEPROM, this will take about 15 seconds...
[20:47:05] Waiting for NRF52 reset...
[20:47:24] Loading signing key...
[20:47:24] Bootstrapping device EEPROM...
[20:47:29] EEPROM written! Validating...
[20:47:30] Setting firmware checksum...
[20:47:37] Waiting for NRF52 reset...
[20:47:58] EEPROM was written, but validation failed. Check your settings.

[jedahan]

Same with RAK4631

------------------------------------------------------------------------------
                               Installer Ready

Ok, that should be all the information we need. Please confirm the following
summary before proceeding. In the next step, the device will be flashed and
provisioned, so make sure that you are satisfied with your choices.

Serial port     : /dev/cu.usbmodem101
Device type     : RAK4631 779 - 928 MHz
Platform        : NRF52
Device MCU      : Nordic Semiconductor nRF52840
Firmware file   : rnode_firmware_rak4631.zip

------------------------------------------------------------------------------

Is the above correct? [y/N] y
[17:59:15] Checking firmware file availability...
[17:59:15] Retrieving latest version info from https://unsigned.io/firmware/latest/?v=2.4.0&variant=rnode_firmware_rak4631.zip
[17:59:15] The selected firmware for this board is version 1.83
[17:59:15] Using existing firmware file: rnode_firmware_rak4631.zip for version 1.83
[17:59:15] Verifying firmware integrity...
[17:59:16] Decompressing firmware...
[17:59:16] Firmware decompressed
[17:59:16] Flashing RNode firmware to device on /dev/cu.usbmodem101
Upgrading target on /dev/cu.usbmodem101 with DFU package /Users/micro/.config/rnodeconf/update/1.83/rnode_firmware_rak4631.zip. Flow control is disabled, Dual bank, Touch 1200
########################################
########################################
########################################
########################################
########################################
########################################
########################################
########################################
########################################
########################################
########################################
##################################
Activating new firmware
Device programmed.
[17:59:45] Done flashing
[17:59:45] Waiting for NRF52 reset...
[17:59:51] Opening serial port /dev/cu.usbmodem101...
[17:59:54] Device connected
[17:59:54] Current firmware version: 1.83
[17:59:54] Reading EEPROM...
[17:59:55] Clearing old EEPROM, this will take about 15 seconds...
[18:00:18] Waiting for NRF52 reset...
[18:00:36] Loading signing key...
[18:00:36] Bootstrapping device EEPROM...
[18:00:42] EEPROM written! Validating...
[18:00:42] Setting firmware checksum...
[18:00:50] Waiting for NRF52 reset...
[18:01:10] EEPROM was written, but validation failed. Check your settings.

[B08Z]

Has support for the RAK4631 been dropped.
I have 1.83 install but just noticed that there is no RAK 4631 firmware in the latest release.

Sorry if I missed something.

[markqvist]

No, not at all. I just had some toolchain problems when building the RAK / ARM versions of the firmware for v1.84, so haven't uploaded any for those yet. They are still on v1.83.

Either way, there's not actually any real changes for the RAK builds between 1.83 and 1.84 either way. Only ESP stuff was really updated in this version, which is why I figured I'd just get the ESP stuff out now and not let that hang on the other stuff.

[B08Z]

Amazing, thanks for the quick reply.

Saw there was no change for the RAK but wanted to know.

All working on the Heltec32v3
Battery state reporting correctly (at least while its on charge)

Thanks very much its a great project / tool

No, not at all. I just had some toolchain problems when building the RAK / ARM versions of the firmware for v1.84, so haven't uploaded any for those yet. They are still on v1.83.

Either way, there's not actually any real changes for the RAK builds between 1.83 and 1.84 either way. Only ESP stuff was really updated in this version, which is why I figured I'd just get the ESP stuff out now and not let that hang on the other stuff.

[dannydelcaribe]

I'm having a somewhat related issue with a RAK4631. I have flashed firmware using Liam Cottle's RNode Flasher with success, but whenever I use rnodeconf -a /dev/ttyACM0 to flash it I get a Radio State Mismatch error when trying to run rnsd.

To clarify I am putting the RAK4631 in DFU mode by double-clicking the reset button on the Wisblock Base, then it pops up as a USB device, then I run rnodeconf -a...

After rnsd this is the output:

[2025-11-26 00:26:57] [Notice]   Opening serial port /dev/ttyACM1...
[2025-11-26 00:26:59] [Notice]   Serial port /dev/ttyACM1 is now open
[2025-11-26 00:26:59] [Verbose]  Configuring RNode interface...
[2025-11-26 00:26:59] [Verbose]  Waiting for radio configuration validation for RNodeInterface[Interface LoRa1]...
[2025-11-26 00:26:59] [Debug]    RNodeInterface[Interface LoRa1] Radio reporting frequency is 915.0 MHz
[2025-11-26 00:26:59] [Debug]    RNodeInterface[Interface LoRa1] Radio reporting bandwidth is 500.0 KHz
[2025-11-26 00:26:59] [Debug]    RNodeInterface[Interface LoRa1] Radio reporting TX power is 22 dBm
[2025-11-26 00:26:59] [Debug]    RNodeInterface[Interface LoRa1] Radio reporting spreading factor is 7
[2025-11-26 00:26:59] [Debug]    RNodeInterface[Interface LoRa1] Radio reporting coding rate is 5
[2025-11-26 00:26:59] [Verbose]  RNodeInterface[Interface LoRa1] On-air bitrate is now 21.88 kbps
[2025-11-26 00:26:59] [Debug]    RNodeInterface[Interface LoRa1] Radio reporting state is offline

[2025-11-26 00:26:59] [Error]    Radio state mismatch
[2025-11-26 00:26:59] [Error]    After configuring RNodeInterface[Interface LoRa1], the reported radio parameters did not match your configuration.
[2025-11-26 00:26:59] [Error]    Make sure that your hardware actually supports the parameters specified in the configuration
[2025-11-26 00:26:59] [Error]    Aborting RNode startup

Anyone experience this issue? It works fine whenever I use the web browser flasher, however I'd like to use rnodeconf via a Raspberry Pi instead.

Thanks in advance!

[markqvist]

[2025-11-26 00:26:59] [Debug] RNodeInterface[Interface LoRa1] Radio reporting state is offline

The radio refuses to turn on. This is most likely because some step of the boot checks failed, like the firmware hash being incorrect, or the device not having a correct configuration present in EEPROM.

If you put the device into DFU mode manually when flashing, the actual firmware flashing might proceed fine, but as soon as rnodeconf tries to restart the device and load the config, set the firmware hash, etc., the device might stay in DFU mode, and rnodeconf can't communicate with the firmware.

If there's no display on the device, you're not seeing the "Missing config" or "Firmware corrupt" error messages, which can make it less obvious, but the rnodeconf -a output has likely given you an error while flashing about some steps failing.

If you really have to put the device into DFU mode manually, you will also need to press the reset button manually once rnodeconf says something like "Waiting for device reset...".

But ideally, you shouldn't have to put it into DFU mode yourself, the flasher should handle it so it can properly reset the device when needed.

[dannydelcaribe]

[EDIT] I just noticed this was an issue back in DEC2024. It was fixed at 'Retrieve firmware image length from bootloader cfg, fix #95 #96' however I am now having the issue :( Not sure if there's something I need to do based on that post. I am new to all of this, sorry if I'm posting in the wrong section.

Thanks for the quick response!
I did some additional tests and am still having the same issue even after running rnodeconf -a without putting the RAK in DFU mode. In the LiamCottle RNode Flasher page I have to run the Set Firmware Hash step, which I don't really know how to do in rnodeconf. I wonder if this is what is causing it. Here is the last flash I made before running into the Radio State mismatch problem:

------------------------------------------------------------------------------
                               Installer Ready

Ok, that should be all the information we need. Please confirm the following
summary before proceeding. In the next step, the device will be flashed and
provisioned, so make sure that you are satisfied with your choices.

Serial port     : /dev/ttyACM0
Device type     : RAK4631 779 - 928 MHz
Platform        : NRF52
Device MCU      : Nordic Semiconductor nRF52840
Firmware file   : rnode_firmware_rak4631.zip

------------------------------------------------------------------------------

Is the above correct? [y/N] y
[18:25:30] Checking firmware file availability...
[18:25:30] Retrieving latest version info from https://unsigned.io/firmware/latest/?v=2.4.0&variant=rnode_firmware_rak4631.zip
[18:25:31] The selected firmware for this board is version 1.85
[18:25:31] Firmware /home/ubuntu/.config/rnodeconf/update/1.85/rnode_firmware_rak4631.zip not found.
[18:25:31] Downloading missing firmware file: rnode_firmware_rak4631.zip for version 1.85
[18:25:31] Firmware file downloaded
[18:25:31] Verifying firmware integrity...
[18:25:32] Generating a new device signing key...
[18:25:32] Device signing key written to /home/ubuntu/.config/rnodeconf/firmware/device.key
[18:25:32] Generating a new EEPROM signing key...
[18:25:32] Decompressing firmware...
[18:25:32] Firmware decompressed
[18:25:32] Flashing RNode firmware to device on /dev/ttyACM0
Upgrading target on /dev/ttyACM0 with DFU package /home/ubuntu/.config/rnodeconf/update/1.85/rnode_firmware_rak4631.zip. Flow control is disabled, Dual bank, Touch 1200
########################################
########################################
########################################
########################################
########################################
########################################
########################################
########################################
########################################
########################################
########################################
##################################
Activating new firmware
Device programmed.
[18:26:00] Done flashing
[18:26:00] Waiting for NRF52 reset...
[18:26:06] Opening serial port /dev/ttyACM0...
[18:26:09] Device connected
[18:26:09] Current firmware version: 1.85
[18:26:09] Reading EEPROM...
[18:26:10] EEPROM checksum correct
[18:26:10] Device signature validation failed
[18:26:10] Clearing old EEPROM, this will take about 15 seconds...
[18:26:33] Waiting for NRF52 reset...
[18:26:51] Loading signing key...
[18:26:51] Bootstrapping device EEPROM...
[18:27:22] EEPROM written! Validating...
[18:27:22] Could not calculate firmware partition hash. The contained exception was:
[18:27:22] module 'hashlib' has no attribute 'file_digest'
[18:27:24] Waiting for NRF52 reset...
[18:27:45] EEPROM checksum correct
[18:27:45] Device signature validated
[18:27:45] EEPROM Bootstrapping successful!

RNode Firmware autoinstallation complete!

I've tried using different RAK4631s, no success. Again, I am able to get them working successfully using the Liam Cottle RNode Flasher.
This line appears interesting:
[18:27:22] module 'hashlib' has no attribute 'file_digest'

After this flash, here is the current rnsd output:

[2025-11-28 18:33:25] [Debug]    Reticulum running in interpreted mode
[2025-11-28 18:33:25] [Debug]    Started shared instance interface: Shared Instance[rns/laptop]
[2025-11-28 18:33:25] [Debug]    Cleaning ratchets...
[2025-11-28 18:33:25] [Extra]    Cleaning resource and packet caches...
[2025-11-28 18:33:25] [Verbose]  Bringing up system interfaces...
[2025-11-28 18:33:25] [Debug]    Skipping disabled interface "Default Interface"
[2025-11-28 18:33:25] [Debug]    Skipping disabled interface "TCPServerInterface"
[2025-11-28 18:33:25] [Debug]    Skipping disabled interface "TCP_ClientInterface_VPS"
[2025-11-28 18:33:25] [Notice]   Opening serial port /dev/ttyACM1...
[2025-11-28 18:33:28] [Verbose]  Serial port /dev/ttyACM1 is now open
[2025-11-28 18:33:28] [Verbose]  Configuring RNode interface...
[2025-11-28 18:33:28] [Verbose]  Waiting for radio configuration validation for RNodeInterface[Interface LoRa1]...
[2025-11-28 18:33:28] [Debug]    RNodeInterface[Interface LoRa1] Radio reporting frequency is 915.0 MHz
[2025-11-28 18:33:28] [Debug]    RNodeInterface[Interface LoRa1] Radio reporting bandwidth is 500.0 KHz
[2025-11-28 18:33:28] [Debug]    RNodeInterface[Interface LoRa1] Radio reporting TX power is 15 dBm
[2025-11-28 18:33:28] [Debug]    RNodeInterface[Interface LoRa1] Radio reporting spreading factor is 7
[2025-11-28 18:33:28] [Debug]    RNodeInterface[Interface LoRa1] Radio reporting coding rate is 5
[2025-11-28 18:33:28] [Verbose]  RNodeInterface[Interface LoRa1] On-air bitrate is now 21.88 kbps
[2025-11-28 18:33:28] [Debug]    RNodeInterface[Interface LoRa1] Radio reporting state is offline
[2025-11-28 18:33:28] [Error]    Radio state mismatch
[2025-11-28 18:33:28] [Error]    After configuring RNodeInterface[Interface LoRa1], the reported radio parameters did not match your configuration.
[2025-11-28 18:33:28] [Error]    Make sure that your hardware actually supports the parameters specified in the configuration
[2025-11-28 18:33:28] [Error]    Aborting RNode startup
[2025-11-28 18:33:28] [Debug]    RNodeInterface[Interface LoRa1] hardware MTU set to 508
[2025-11-28 18:33:28] [Debug]    Skipping disabled interface "Interface LoRa2"
[2025-11-28 18:33:28] [Debug]    Skipping disabled interface "RNS Testnet BetweenTheBorders"
[2025-11-28 18:33:28] [Debug]    Skipping disabled interface "Sydney RNS"
[2025-11-28 18:33:28] [Verbose]  System interfaces are ready
[2025-11-28 18:33:28] [Debug]    Utilising cryptography backend "openssl, PyCA 3.4.8"
[2025-11-28 18:33:28] [Verbose]  Configuration loaded from /home/ubuntu/.reticulum/config
[2025-11-28 18:33:28] [Verbose]  Loaded 106 known destination from storage
[2025-11-28 18:33:28] [Verbose]  Loaded Transport Identity from storage
[2025-11-28 18:33:28] [Notice]   Started rnsd version 1.0.4

[markqvist]

module 'hashlib' has no attribute 'file_digest'

That's the problem. It can't calculate and set the firmware hash. What Python version and operating system are you on?

[dannydelcaribe]

I'm using:
Ubuntu 22.04.1
Python 3.10.12

I know my Ubuntu is older, but I need it this way. Not sure about Python. If thats the issue I will just continue flashing with the web page for now. Perhaps theres a way to do this manually with the -H option in rnodeconf? Thanks a lot. I love this project!

P.S. Can RNodes operate without a host as relays? As in, I want to put RNode transport nodes out in the wild, but not have to run a Raspberry Pi with it (due to size, power needs, etc.) It would be amazing to just be able to have multiple RNodes out acting as relays... I tried this by flashing Transport Firmware and putting them in TNC mode, but it didn't work.

[dannydelcaribe]

I found a solution to the rnodeconf / firmware-hash issues I was having.

The general steps were:

• Updated python3 to 3.11.4. [EDIT] Only install python3.11 necessary. Don't change the sym link.
• Updated pip to 25.3. [EDIT] Not necessary.
• Made python3 default to 3.11. [EDIT] This step broke my GUI. I recommend creating a virtual environment with python3.11 instead as such: python3.11 -m venv ~/.venv311 Then install inside this venv using pip and run rnodeconf from here. Update $PATH or give absolute path when running the program.
• Created venv using python3.11. python3.11 -m venv ~/.venv311
• Installed rns inside venv. ~/.venv311/bin/pip install rns
• Ran rnodeconf inside this venv. ~/.venv311/bin/rnodeconf -a...

[serialrf433]

It seems to be OS related like @dannydelcaribe have figured out. I used rnodeconf to flash a T-Echo that is also nRF52 chipset and i had no issues installing 1.85 firmware release.