Files
meshtastic_firmware/variants/esp32s3/EBYTE_ESP32-S3/variant.h
T
+19 4c91beeda9 Develop to master (#9618)
* Don't ever define PIN_LED or BLE_LED_INVERTED (#9494)

* Don't ever define PIN_LED

* Deprecate BLE_LED_INVERTED

* Add StatusMessage module and config overrides (#9351)

* Add StatusMessage module and config overrides

* Trunk

* Don't reboot node simply for a StatusMessage config update

* Missed in reviews - fixing send bubble (#9505)

* Prefer EXT_PWR_DETECT pin over chargingVolt to detect power unplugged (#9511)

* Make sure we always return a value in NodeDB::restorePreferences() (#9516)

In case FScom is not defined there is no return statement. This
moves the return outside of the ifdef to make sure a defined
value is returned.

* Inkhud battery icon improvements. (#9513)

* Inkhud battery icon improvements.
Fixes the battery icon draining from the flat side towards the bump, which is backwards from general design language seen on most devices
By request of kr0n05_ on discord, adds the ability to mirror the battery icon which fixes that issue in another way, and is also a common design seen on other devices.

* Remove option for icon mirroring

* Add border + dither to battery to prevent font overlap

* Fix trunk format

* Code cleanup, courtesy of Xaositek.

* Add reply bot module with DM-only responses and rate limiting (#9456)

* Implement Meshtastic reply bot module with ping and status features

Adds a reply bot module that listens for /ping, /hello, and /test commands received via direct messages or broadcasts on the primary channel. The module always replies via direct message to the sender only, reporting hop count, RSSI, and SNR. Per-sender cooldowns are enforced to reduce network spam, and the module can be excluded at build time via a compile flag. Updates include the new module source files and required build configuration changes.

* Update ReplyBotModule.cpp

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

* Update src/modules/ReplyBotModule.h

Match the existing MESHTASTIC_EXCLUDE_* guard pattern so the module is excluded by default.

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

* Update src/modules/ReplyBotModule.cpp

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

* Tidying up

---------

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
Co-authored-by: Ben Meadors <benmmeadors@gmail.com>

* HotFix for ReplyBot - Modules.cpp included and moved configuration.h (#9532)

* Undefine LED_BUILTIN (#9531)

Keep variant in sync with
https://github.com/meshtastic/firmware/commit/df40085

Co-authored-by: Ben Meadors <benmmeadors@gmail.com>

* Add agc reset attempt (#8163)

* Add agc reset attempt

* Add radioLibInterface include

* Trunk

* AGC reset don't crash, don't naively call

* Update src/main.cpp

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

* Use Throttle function

---------

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

* Remove unused hmx variable (#9529)

The variable is not used at all in the function, remove it to
silence the compiler warning.

Co-authored-by: Ben Meadors <benmmeadors@gmail.com>

* Rename LED_PIN to LED_POWER, move handling out of main to dedicated module (#9512)

* Rename LED_PIN to LED_POWER, move handling out of main to dedicated module

* Misc

* Remove errant endif

* Fix hop_limit upgrade detection (#9550)

Co-authored-by: Ben Meadors <benmmeadors@gmail.com>

* meshtasticd: Fix install on Fedora 43 (#9556)

* RPM: Include meshtasticd-start.sh (#9561)

* Add Slash Key to VirtualKeyboard (#9563)

Addition of ? and / to the virtual Keyboard via short and long press

* Add support for CW2015 LiPo battery fuel gauge (#9564)

* Add support for CW2015 LiPo battery fuel gauge

* Address Copilot's concerns, minor fixups

* Make LED_POWER blip even in critical battery (#9545)

* Enable FORTIFY and SP for native builds (#9537)

* Enable FORITFY and NX for native builds

meshtasticd does have an executable stack and is not built with fortify, which makes exploitation of memory corruption bugs easier than it has to be. This enables fortify and a non-executable stack.

This gives the following improvements on Debian Trixie:

$ checksec --file=./.pio/build/native/meshtasticd
RELRO           STACK CANARY      NX            PIE             RPATH      RUNPATH	Symbols		FORTIFY	Fortified	Fortifiable	FILE
Partial RELRO   No canary found   NX enabled    PIE enabled     No RPATH   No RUNPATH   13516 Symbols	  No	0		17		./.pio/build/native/meshtasticd

$ checksec --file=./.pio/build/native/meshtasticd
RELRO           STACK CANARY      NX            PIE             RPATH      RUNPATH	Symbols		FORTIFY	Fortified	Fortifiable	FILE
Partial RELRO   Canary found      NX enabled    PIE enabled     No RPATH   No RUNPATH   13519 Symbols	  Yes	12		20		./.pio/build/native/meshtasticd

Tested with --sim mode I do not get any crashes or similar.

* Enable FORTIFY and NX for native builds

meshtasticd does have an executable stack and is not built with fortify, which makes exploitation of memory corruption bugs easier than it has to be. This enables fortify and a non-executable stack.

This gives the following improvements on Debian Trixie:

$ checksec --file=./.pio/build/native/meshtasticd
RELRO           STACK CANARY      NX            PIE             RPATH      RUNPATH	Symbols		FORTIFY	Fortified	Fortifiable	FILE
Partial RELRO   No canary found   NX enabled    PIE enabled     No RPATH   No RUNPATH   13516 Symbols	  No	0		17		./.pio/build/native/meshtasticd

$ checksec --file=./.pio/build/native/meshtasticd
RELRO           STACK CANARY      NX            PIE             RPATH      RUNPATH	Symbols		FORTIFY	Fortified	Fortifiable	FILE
Partial RELRO   Canary found      NX enabled    PIE enabled     No RPATH   No RUNPATH   13519 Symbols	  Yes	12		20		./.pio/build/native/meshtasticd

Tested with --sim mode I do not get any crashes or similar.

* Enable FORTIFY and SP for native builds

meshtasticd does have a stack canaries and is not built with fortify, which makes exploitation of memory corruption bugs easier than it has to be. This enables fortify and stack canaries.

This gives the following improvements on Debian Trixie:

$ checksec --file=./.pio/build/native/meshtasticd
RELRO           STACK CANARY      NX            PIE             RPATH      RUNPATH	Symbols		FORTIFY	Fortified	Fortifiable	FILE
Partial RELRO   No canary found   NX enabled    PIE enabled     No RPATH   No RUNPATH   13516 Symbols	  No	0		17		./.pio/build/native/meshtasticd

$ checksec --file=./.pio/build/native/meshtasticd
RELRO           STACK CANARY      NX            PIE             RPATH      RUNPATH	Symbols		FORTIFY	Fortified	Fortifiable	FILE
Partial RELRO   Canary found      NX enabled    PIE enabled     No RPATH   No RUNPATH   13519 Symbols	  Yes	12		20		./.pio/build/native/meshtasticd

Tested with --sim mode I do not get any crashes or similar.

---------

Co-authored-by: Ben Meadors <benmmeadors@gmail.com>

* Update built-in documentation for current method of implementation (#9592)

* Refactor logging in ProtobufModule to ensure message details are logged after successful decoding (#9536)

* Automated version bumps (#9604)

Co-authored-by: thebentern <9000580+thebentern@users.noreply.github.com>

* Add missing openocd_target to custom nrf52 boards (#9603)

This stops platformio complaining about `Missing target configuration for me25ls01-4y10td` etc when trying to flash with nrfutil.

* Add sdl libs for native builds (#9595)

* Add sdl libs for native builds

* Alpine try again

* fix some random compiler warnings (#9596)

* Modify the dependency library of v4-tft (#9507)

* BaseUI: Favorite Screen Signal Quality improvement (#9566)

* Favorite Signal Quality improvement

* Show Voltage if node shares it.

* Trunk Fix

* Change Favorite tittle to encase name with Asterisks

* Add Pluggin In condition for Battery Line

* Adjust getUptimeStr Prefixes

* Create isAPIConnected for SharedCommon usage

* Correct leftSideSpacing to account for isAPIConnected

---------

Co-authored-by: Jason P <applewiz@mac.com>

* ExternalNotification and StatusLED now call AmbientLighting to update… (#9554)

* ExternalNotification and StatusLED now call AmbientLighting to update RGB LEDs. Add optional heartbeat

* Don't overwrite RGB state if heartbeat is disabled.

* Use the right define

* Remove another .h and make rgb static

* move rgb objects into AmbientLighting class

* Straighten out AmbientLighting Thread object

* Use %f for floats

* Fixes on SCD4X admin comands (#9607)

* Fixes on SCD4X admin comands

* Minor fix in logs for SEN5X

---------

Co-authored-by: Ben Meadors <benmmeadors@gmail.com>

* feat/add sfa30 (#9372)

* Move PMSA003I to separate class and update AQ telemetry

* AirQualityTelemetry module not depend on PM sensor presence

* Remove commented line

* Fixes on PMS class

* Add missing warmup period to wakeUp function

* Fixes on compilation for different variants

* Add functions to check for I2C bus speed and set it

* Initial implementation for SFA30Sensor

* Move PMSA003I to separate class and update AQ telemetry

* AirQualityTelemetry module not depend on PM sensor presence

* Remove commented line

* Fixes on PMS class

* Add missing warmup period to wakeUp function

* Fixes on compilation for different variants

* Add functions to check for I2C bus speed and set it

* Add ScreenFonts.h

Co-authored-by: Hannes Fuchs <hannes.fuchs+git@0xef.de>

* PMSA003I 1st round test

* Fix I2C scan speed

* Fix minor issues and bring back I2C SPEED def

* Remove PMSA003I library as its no longer needed

* Add functional SCD4X

* Fix screen frame for CO2

* Add admin commands to SCD4X class

* Add further admin commands and fixes.

* Remove unused I2C speed functions and cleanup

* Cleanup of SEN5X specific code added from switching branches
* Remove SCAN_I2C_CLOCK_SPEED block as its not needed
* Remove associated functions for setting I2C speed

* Unify build epoch to add flag in platformio-custom.py (#7917)

* Unify build_epoch replacement logic in platformio-custom

* Missed one

* Fix build error in rak_wismesh_tap_v2 (#7905)

In the logs was:
"No screen resolution defined in build_flags. Please define DISPLAY_SIZE."

set according to similar devices.

* Put guards in place around debug heap operations (#7955)

* Put guards in place around debug heap operations

* Add macros to clean up code

* Add pointer as well

* Cleanup

* Fix memory leak in NextHopRouter: always free packet copy when removing from pending

* Formatting

* Only queue 2 client notification

* Merge pull request #7965 from compumike/compumike/fix-nrf52-bluetooth-memory-leak

Fix memory leak in `NRF52Bluetooth`: allocate `BluetoothStatus` on stack, not heap

* Merge pull request #7964 from compumike/compumike/fix-nimble-bluetooth-memory-leak

Fix memory leak in `NimbleBluetooth`: allocate `BluetoothStatus` on stack, not heap

* Update protobufs (#7973)

Co-authored-by: thebentern <9000580+thebentern@users.noreply.github.com>

* T-Lora Pager: Support LR1121 and SX1280 models (#7956)

* T-Lora Pager: Support LR1121 and SX1280 models

* Remove ifdefs

* Trunk

* Trunk

* Static memory pool allocation (#7966)

* Static memory pool

* Initializer

* T-Lora Pager: Support LR1121 and SX1280 models (#7956)

* T-Lora Pager: Support LR1121 and SX1280 models

* Remove ifdefs

---------

Co-authored-by: WillyJL <me@willyjl.dev>

* Portduino dynamic alloc

* Missed

* Drop the limit

* Update meshtastic-esp8266-oled-ssd1306 digest to 0cbc26b (#7977)

Co-authored-by: renovate[bot] <29139614+renovate[bot]@users.noreply.github.com>

* Fix json report crashes on esp32 (#7978)

* Tweak maximums

* Fix DRAM overflow on old esp32 targets

* Guard bad time warning logs using GPS_DEBUG (#7897)

In 2.7.7 / 2.7.8 we introduced some new checks for time accuracy.

In combination, these result in a spamming of the logs when a bad time is found

When the GPS is active, we're calling the GPS thread every 0.2secs.

So this log could be printed 4,500 times in a no-lock scenario :)

Reserve this experience for developers using GPS_DEBUG.

Fixes https://github.com/meshtastic/firmware/issues/7896

* Scale probe buffer size based on current baud rate (#7975)

* Scale probe buffer size based on current baud rate

* Throttle bad time validation logging and fix time comparison logic

* Remove comment

* Missed the other instances

* Copy pasta

* Fix GPS gm_mktime memory leak (#7981)

* Fix overflow of time value (#7984)

* Fix overflow of time value

* Revert "Fix overflow of time value"

This reverts commit 084796920179e80a7500d36c25fd4d82b3ef4214.

* That got boogered up

* Remove PMSA003 include from modules

* Add flag to exclude air quality module

* Rework PMSA003I to align with new I2C scanner

* Reworks AQ telemetry to match new dynamic allocation method
* Adds VBLE_I2C_CLOCK_SPEED build flag for sensors with different I2C speed requirements
* Reworks PMSA003I

* Move add sensor template to separate file

* Split telemetry on screen options

* Add variable I2C clock compile flag

* Added to Seeed Xiao S3 as demo

* Fix drawFrame in AQ module

* Module settings override to i2cScan module function

* Move to CAN_RECLOCK_I2C per architecture

* Add reclock function in TelemetrySensor.cpp
* Add flag in ESP32 common

* Minor fix

* Move I2C reclock function to src/detect

* Fix uninitMemberVar errors and compile issue

* Make sleep, wakeUp functions generic

* Fix STM32 builds

* Add exclude AQ sensor to builds that have environmental sensor excludes
* Add includes to AddI2CSensorTemplate.h

* SEN5X first pass

* WIP Sen5X functions

* Further (non-working) progress in SEN5X

* WIP Sen5X functions

* Changes on SEN5X library - removing pm_env as well

* Small cleanup of SEN5X sensors

* Minor change for SEN5X detection

* Remove dup code

* Enable PM sensor before sending telemetry.

This enables the PM sensor for a predefined period to allow for warmup.

Once telemetry is sent, the sensor shuts down again.

* Small cleanups in SEN5X sensor

* Add dynamic measurement interval for SEN5X

* Only disable SEN5X if enough time after reading.

* Idle for SEN5X on communication error

* Cleanup of logs and remove unnecessary delays

* Small TODO

* Settle on uint16_t for SEN5X PM data

* Make AQTelemetry sensors non-exclusive

* Implementation of cleaning in FS prefs and cleanup

* Remove unnecessary LOGS
* Add cleaning date storage in FS
* Report non-cumulative PN

* Bring back detection code for SEN5X after branch rebase

* Add placeholder for admin message

* Add VOC measurements and persistence (WIP)

* Adds VOC measurements and state
* Still not working on VOC Index persistence
* Should it stay in continuous mode?

* Add one-shot mode config flag to SEN5X

* Add nan checks on sensor data from SEN5X

* Working implementation on VOCState

* Adds initial timer for SEN55 to not sleep if VOCstate is not stable (1h)
* Adds conditions for stability and sensor state

* Fixes on VOC state and mode swtiching

* Adds a new RHT/Gas only mode, with 3600s stabilization time
* Fixes the VOCState buffer mismatch
* Fixes SEN50/54/55 model mistake

* Adapt SEN5X to new sensor list structure. Improve reclock.

* Improve reClockI2C conditions for different variants
* Add sleep, wakeUp, pendingForReady, hasSleep functions to PM sensors to save battery
* Add SEN5X

* Fix merge errors

* Small reordering in PMS class for consistency

* If one sensor fails, AQ telemetry still reports data

* Small formatting fix

* Add SEN5X to AQI in ScanI2C

* SCD4X now part of AQ module with template list

* Fixes difference between idle and sleep
* In LowPower, sleep is disabled
* Requires testing for I2C clock comms for commands

* Remove unnecessary import

* Add co2 to serialized AQ metrics

* Add SFA30 with new sensor template in AQ module

* Update library dependencies in platformio.ini

* Fix unitialized variables in SEN5X constructor

* Fix missing import

* Fix uninitMemberVars

* Fix import error for SCD4X

* Fix I2CClock logic

* Fix not reclocking back to 700000Hz

* Fix multiple sensors being read simultaneously

*  The logic in AQ module is different to the one in EnvironmentTelemetryModule.  In Env module, if any sensor fails to getMetrics, no valid flag for the module. This prevents other sensors to report data.

* Fix pending clock change in PMSA003

* Cleanup of SEN5X class

* Exclude AQ sensor from wio-e5 due to flash limitations

* Fix I2C clock change logic

* Make sure clock is always set to needed value

* Fix returns

* Fix trunk

* Fix on condition in reclock

* Fix trunk

* Final SFA30 class implementation

* Add HCHO to screen and improve logs

* Add metrics to mesh packet serializer

* Minor fixes in logs

* OCD tidy up of logs

* Fix sleep function

* Remove old I2C_CLOCK_SPEED code

---------

Co-authored-by: nikl <nikl174@mailbox.org>
Co-authored-by: Hannes Fuchs <hannes.fuchs+git@0xef.de>
Co-authored-by: Nashui-Yan <yannashui10@gmail.com>
Co-authored-by: Ben Meadors <benmmeadors@gmail.com>
Co-authored-by: Tom Fifield <tom@tomfifield.net>
Co-authored-by: Mike Robbins <mrobbins@alum.mit.edu>
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
Co-authored-by: thebentern <9000580+thebentern@users.noreply.github.com>
Co-authored-by: WillyJL <me@willyjl.dev>
Co-authored-by: renovate[bot] <29139614+renovate[bot]@users.noreply.github.com>

* Log rxBad PacketHeaders with more info (id, relay_node) like printPacket, so we can try to match RX errors to other packets in the logs. (#9614)

* Exclude status message module

* fix: zero entire public key array instead of only first byte (#9619)

* Update protobufs (#9621)

* Xiao NRF - define suitable i2c pins for the sub-variants (#8866)

Co-authored-by: Christian Walther <cwalther@gmx.ch>

* Update src/detect/ScanI2C.cpp

Co-authored-by: Wessel <wessel@weebl.me>

* Update src/modules/Telemetry/Sensor/SFA30Sensor.cpp

Co-authored-by: Wessel <wessel@weebl.me>

* Update src/modules/Telemetry/Sensor/SFA30Sensor.cpp

Co-authored-by: Wessel <wessel@weebl.me>

* Update src/modules/Telemetry/Sensor/SFA30Sensor.cpp

Co-authored-by: Wessel <wessel@weebl.me>

* Update src/mesh/NodeDB.cpp

Co-authored-by: Wessel <wessel@weebl.me>

* convert GPS global and some new in gps.cpp to unique_ptr (#9628)

Trying this to see if anything bad happen if I were to replace most raw pointers with unique_ptr.

I didn't used std::make_unique since it is only supported in C++14 and onwards but until we update esp32 to arduino 3.x the ESP32 xtensa chips use a C++11 std.

* replace delete in RedirectablePrint.cpp with std::unique_ptr (#9642)

Is part of the unique_ptr modernization effort.

* replace delete in EInkDynamicDisplay.{cpp,h} with std::unique_ptr (#9643)

Is part of the unique_ptr modernization effort.

* Undefine LED_BUILTIN for Heltec v2 variant (#9647)

* Undefine LED_BUILTIN for Heltec v2 variant

* Undefine LED_BUILTIN for Heltec v2.1 variant

---------

Co-authored-by: Jorropo <jorropo.pgm@gmail.com>

* replace delete in RadioInterface.cpp with std::unique_ptr (#9645)

Is part of the unique_ptr modernization effort.

* fix typo in PIN_GPS_SWITCH (#9648)

Wasn't caught by CI.

* replace delete in CryptoEngine.{cpp,h} with std::unique_ptr (#9649)

Is part of the unique_ptr modernization effort.

* workaround NCP5623 and LP5562 I2C builds (#9652)

Theses two appear to be buggy on r1-neo and nomadstar meteor pro, they rely on Wire.h being included previously to their import.

Idk why other platforms using the same smart LEDs are working while theses ones don't.

This should make CI green on the dev branch.

* replace delete in AudioThread.h with std::unique_ptr (#9651)

Is part of the unique_ptr modernization effort.

* Add USB_MODE=1 for Station G2 (#9660)

* InkHUD: Favorite Map Applet (#9654)

* fix a lot of low level cppcheck warnings (#9623)

* simplify the observer pattern, since all the called functions are const getters.
* use arduino macro over std: for numerical values and refactor local variables in drawScrollbar()
* oh, so Cppcheck actually complained about const pointers not being const.
* slowly getting out of ifdef hell
* fix inkHUD warnings as well
* last 2 check warnings
* git checks should fail on low defects from now on

* Feat/add scd30 (#9609)

* Merge develop into SCD30

* Add SCD30 class

* Fix logging and admin commands

* Minor cleanup and logging improvements

* Minor formatting issue

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

* Improvements on setTemperature

* Fix casting float-uint16_t
* Pass 100 for resetting temperature offset

* Fix issues pointed out by copilot

* Add quick reboot to set interval quicker on scd30

* Change saveState to only happen after boot and minor log changes

* Fix missing semicolon in one shot mode log

---------

Co-authored-by: Thomas Göttgens <tgoettgens@gmail.com>
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
Co-authored-by: Ben Meadors <benmmeadors@gmail.com>

* fix: respect DontMqttMeBro flag regardless of channel PSK (#9626)

The previous PSK check was broken from its introduction in #4643 —
memcmp was used in boolean context without comparing to 0, inverting
the condition. Since no one noticed for over a year, the PSK-based
filtering provided no practical value. Simplifying to always respect
the sender's preference is both more correct and easier to reason about.

* our firmware action is too clever

Update pio_target and add pio_opts for checks.

* fix detection of SCD30 by checking if the size of the return from a 2 byte register read is correct (#9664)

* fix detection of SCD30 by checking if thee size of the return from a 2 byte register read is correct
fix signedness warning in PMSA003 sensor code.

* Add alternate path for LPS22HB

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

* check EndTransmission for errors and compare returned length to expected value

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

---------

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

* #9623 resolved a local shadow of next_key by converting it to int. (#9665)

* zip a few gitrefs down (#9672)

* InkHUD: Allow non-system applets to subscribe to input events (#9514)

* Allow inkhud user applets to consume inputs with opt-in system
Adds a way for applets to subscribe to input events while keeping it off
by default to preserve compatibility and expected behaviours. Adds
example for use as well.
* Add check for nullptr on getActiveApplet uses
* Remove redundant includes
* Move subscribedInputs to protected
* More consistent naming scheme

* Fake IAQ values on Non-BSEC2 platforms like Platformio and the original ESP32 (#9663)

* BSEC2 Replacement
- add approximation for IAQ to non-BSEC2 platforms.
- Re-add this sensor to ESP32 targets, and refactor env_extra includes.
- Fix C++ 11 compatibility
* Check for gas resistance 0

* ULED_BUILTIN for 9m2ibr_aprs_lora_tracker (#9685)

---------

Co-authored-by: Jonathan Bennett <jbennett@incomsystems.biz>
Co-authored-by: Jason P <applewiz@mac.com>
Co-authored-by: Eric Sesterhenn <eric.sesterhenn@x41-dsec.de>
Co-authored-by: Vortetty <33466216+Vortetty@users.noreply.github.com>
Co-authored-by: Mattatat25 <108779801+mattatat25@users.noreply.github.com>
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
Co-authored-by: Max <rekin.m@gmail.com>
Co-authored-by: Colby Dillion <colby.dillion@pacshealth.com>
Co-authored-by: Austin <vidplace7@gmail.com>
Co-authored-by: Tom <116762865+NomDeTom@users.noreply.github.com>
Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>
Co-authored-by: Chloe Bethel <chloe@9net.org>
Co-authored-by: Thomas Göttgens <tgoettgens@gmail.com>
Co-authored-by: Quency-D <55523105+Quency-D@users.noreply.github.com>
Co-authored-by: HarukiToreda <116696711+HarukiToreda@users.noreply.github.com>
Co-authored-by: oscgonfer <oscgonfer@users.noreply.github.com>
Co-authored-by: nikl <nikl174@mailbox.org>
Co-authored-by: Hannes Fuchs <hannes.fuchs+git@0xef.de>
Co-authored-by: Nashui-Yan <yannashui10@gmail.com>
Co-authored-by: Tom Fifield <tom@tomfifield.net>
Co-authored-by: Mike Robbins <mrobbins@alum.mit.edu>
Co-authored-by: WillyJL <me@willyjl.dev>
Co-authored-by: renovate[bot] <29139614+renovate[bot]@users.noreply.github.com>
Co-authored-by: Wessel <wessel@weebl.me>
Co-authored-by: Christian Walther <cwalther@gmx.ch>
Co-authored-by: Jorropo <jorropo.pgm@gmail.com>
Co-authored-by: Eric Barch <ericb@ericbarch.com>
Co-authored-by: Clive Blackledge <clive@ansible.org>
2026-02-19 07:16:33 -06:00

193 lines
13 KiB
C

// Supporting information: https://github.com/S5NC/EBYTE_ESP32-S3/
// Originally developed for E22-900M30S with ESP32-S3-WROOM-1-N4
// NOTE: Uses ESP32-S3-WROOM-1-N4.json in boards folder (via platformio.ini board field), assumes 4 MB (quad SPI) flash, no PSRAM
// FIXME: implement SX12 module type autodetection and have setup for each case (add E32 support)
// E32 has same pinout except having extra pins. I assume that the GND on it is connected internally to other GNDs so it is not a
// problem to NC the extra GND pins.
// For each EBYTE module pin in this section, provide the pin number of the ESP32-S3 you connected it to
// The ESP32-S3 is great because YOU CAN USE PRACTICALLY ANY PINS for the connections, but avoid some pins (such as on the WROOM
// modules the following): strapping pins (except 0 as a user button input as it already has a pulldown resistor in typical
// application schematic) (0, 3, 45, 46), USB-reserved (19, 20), and pins which aren't present on the WROOM-2 module for
// compatiblity as it uses octal SPI, or are likely connected internally in either WROOM version (26-37), and avoid pins whose
// voltages are set by the SPI voltage (47, 48), and pins that don't exist (22-25) You can ALSO set the SPI pins (SX126X_CS,
// SX126X_SCK, SX126X_MISO, SX126X_MOSI) to any pin with the ESP32-S3 due to \ GPIO Matrix / IO MUX / RTC IO MUX \, and also the
// serial pins, but this isn't recommended for Serial0 as the WROOM modules have a 499 Ohm resistor on U0TXD (to reduce harmonics
// but also acting as a sort of protection)
// We have many free pins on the ESP32-S3-WROOM-X-Y module, perhaps it is best to use one of its pins to control TXEN, and use
// DIO2 as an extra interrupt, but right now Meshtastic does not benefit from having another interrupt pin available.
// Adding two 0-ohm links on your PCB design so that you can choose between the two modes for controlling the E22's TXEN would
// enable future software to make the most of an extra available interrupt pin
// Possible improvement: can add extremely low resistance MOSFET to physically toggle power to E22 module when in full sleep (not
// waiting for interrupt)?
// PA stands for Power Amplifier, used when transmitting to increase output power
// LNA stands for Low Noise Amplifier, used when \ listening for / receiving \ data to increase sensitivity
//////////////////////////////////////////////////////////////////////////////////
// //
// Have custom connections or functionality? Configure them in this section //
// //
//////////////////////////////////////////////////////////////////////////////////
#define SX126X_CS 14 // EBYTE module's NSS pin // FIXME: rename to SX126X_SS
#define LORA_SCK 21 // EBYTE module's SCK pin
#define LORA_MOSI 38 // EBYTE module's MOSI pin
#define LORA_MISO 39 // EBYTE module's MISO pin
#define SX126X_RESET 40 // EBYTE module's NRST pin
#define SX126X_BUSY 41 // EBYTE module's BUSY pin
#define SX126X_DIO1 42 // EBYTE module's DIO1 pin
// We don't define a pin for SX126X_DIO2 as Meshtastic doesn't use it as an interrupt output, so it is never connected to an MCU
// pin! Also E22 module datasheets say not to connect it to an MCU pin.
// We don't define a pin for SX126X_DIO3 as Meshtastic doesn't use it as an interrupt output, so it is never connected to an MCU
// pin! Also E22 module datasheets say to use it as the TCXO's reference voltage.
// E32 module (which uses SX1276) may not have ability to set TCXO voltage using a DIO pin.
// The radio module needs to be told whether to enable RX mode or TX mode. Each radio module takes different actions based on
// these values, but generally the path from the antenna to SX1262 is changed from signal output to signal input. Also, if there
// are LNAs (Low-Noise Amplifiers) or PAs (Power Amplifiers) in the output or input paths, their power is also controlled by
// these pins. You should never have both TXEN and RXEN set high, this can cause problems for some radio modules, and is
// commonly referred to as 'undefined behaviour' in datasheets. For the SX1262, you shouldn't connect DIO2 to the MCU. DIO2 is
// an output only, and can be controlled via SPI instructions, the use for this is to save an MCU pin by using the DIO2 pin to
// control the RF switching mode.
// Choose ONLY ONE option from below, comment in/out the '/*'s and '*/'s
// SX126X_TXEN is the E22's [SX1262's] TXEN pin, SX126X_RXEN is the E22's [SX1262's] RXEN pin
// Option 1: E22's TXEN pin connected to E22's DIO2 pin, E22's RXEN pin connected to NEGATED output of E22's DIO2 pin (more
// expensive option hardware-wise, is the 'most proper' way, removes need for routing one/two traces from MCU to RF switching
// pins), however you can't have E22 in low-power 'sleep' mode (TXEN and RXEN both low cannot be achieved this this option).
/*
#define SX126X_DIO2_AS_RF_SWITCH
#define SX126X_TXEN RADIOLIB_NC
#define SX126X_RXEN RADIOLIB_NC
*/
// Option 2: E22's TXEN pin connected to E22's DIO2 pin, E22's RXEN pin connected to MCU pin (cheaper option hardware-wise,
// removes need for routing another trace from MCU to an RF switching pin).
// /*
#define SX126X_DIO2_AS_RF_SWITCH
#define SX126X_TXEN RADIOLIB_NC
#define SX126X_RXEN 10
// */
// Option 3: E22's TXEN pin connected to MCU pin, E22's RXEN pin connected to MCU pin (cheaper option hardware-wise, allows for
// ramping up PA before transmission (add/expand on feature yourself in RadioLib) if PA takes a while to stabilise)
// Don't define DIO2_AS_RF_SWITCH because we only use DIO2 or an MCU pin mutually exclusively to connect to E22's TXEN (to prevent
// a short if they are both connected at the same time (suboptimal PCB design) and there's a slight non-neglibible delay and/or
// voltage difference between DIO2 and TXEN). Can use DIO2 as an IRQ (but not in Meshtastic at the moment).
/*
#define SX126X_TXEN 9
#define SX126X_RXEN 10
*/
// (NOT RECOMMENDED, if need to ramp up PA before transmission, better to use option 3)
// Option 4: E22's TXEN pin connected to MCU pin, E22's RXEN pin connected to NEGATED output of E22's DIO2 pin (more expensive
// option hardware-wise, allows for ramping up PA before transmission (add/expand on feature yourself in RadioLib) if PA takes
// a while to stabilise, removes need for routing another trace from MCU to an RF switching pin, however may mean if in
// RadioLib you don't tell DIO2 to go high to indicate transmission (so the negated output goes to RXEN to turn the LNA off)
// then you may end up enabling E22's TXEN and RXEN pins at the same time whilst you ramp up the PA which is not ideal,
// changing DIO2's switching advance in RadioLib may not even be possible, may be baked into the SX126x).
/*
#define SX126X_DIO2_AS_RF_SWITCH
#define SX126X_TXEN 9
#define SX126X_RXEN RADIOLIB_NC
*/
// Status
#define LED_POWER 1
#define LED_STATE_ON 1 // State when LED is lit
// External notification
// FIXME: Check if EXT_NOTIFY_OUT actualy has any effect and removes the need for setting the external notication pin in the
// app/preferences
#define EXT_NOTIFY_OUT 2 // The GPIO pin that acts as the external notification output (here we connect an LED to it)
// Buzzer
#define PIN_BUZZER 11
// Buttons
#define BUTTON_PIN 0 // Use the BOOT button as the user button
// I2C
#define I2C_SCL 18
#define I2C_SDA 8
// UART
#define UART_TX 43
#define UART_RX 44
// Power
// Outputting 22dBm from SX1262 results in ~30dBm E22-900M30S output (module only uses last stage of the YP2233W PA)
// Respect local regulations! If your E22-900M30S outputs the advertised 30 dBm and you use a 6 dBi antenna, you are at the
// equivalent of 36 EIRP (Effective Isotropic Radiated Power), which in this case is the limit for non-HAM users in the US (4W
// EIRP, at SPECIFIC frequencies).
// In the EU (and UK), as of now, you are allowed 27 dBm ERP which is 29.15 EIRP.
// https://eur-lex.europa.eu/legal-content/EN/TXT/?uri=CELEX:32022D0180
// https://www.legislation.gov.uk/uksi/1999/930/schedule/6/made
// To respect the 29.15 dBm EIRP (at SPECIFIC frequencies, others are lower) EU limit with a 2.5 dBi gain antenna, consulting
// https://github.com/S5NC/EBYTE_ESP32-S3/blob/main/power%20testing.txt, assuming 0.1 dBm insertion loss, output 20 dBm from the
// E22-900M30S's SX1262. It is worth noting that if you are in this situation and don't have a HAM license, you may be better off
// with a lower gain antenna, and output the difference as a higher total power input into the antenna, as your EIRP would be the
// same, but you would get a wider angle of coverage. Also take insertion loss and possibly VSWR into account
// (https://www.everythingrf.com/tech-resources/vswr). Please check regulations yourself and check airtime, usage (for example
// whether you are airborne), frequency, and power laws.
#define SX126X_MAX_POWER 22 // SX126xInterface.cpp defaults to 22 if not defined, but here we define it for good practice
// Display
// FIXME: change behavior in src to default to not having screen if is undefined
// FIXME: remove 0/1 option for HAS_SCREEN in src, change to being defined or not
// FIXME: check if it actually causes a crash when not specifiying that a display isn't present
#define HAS_SCREEN 0 // Assume no screen present by default to prevent crash...
// GPS
// FIXME: unsure what to define HAS_GPS as if GPS isn't always present
#define HAS_GPS 1 // Don't need to set this to 0 to prevent a crash as it doesn't crash if GPS not found, will probe by default
#define PIN_GPS_EN 15
#define GPS_EN_ACTIVE 1
#define GPS_TX_PIN 16
#define GPS_RX_PIN 17
/////////////////////////////////////////////////////////////////////////////////
// //
// You should have no need to modify the code below, nor in pins_arduino.h //
// //
/////////////////////////////////////////////////////////////////////////////////
#define USE_SX1262 // E22-900M30S, E22-900M22S, and E22-900MM22S (not E220!) use SX1262
#define USE_SX1268 // E22-400M30S, E22-400M33S, E22-400M22S, and E22-400MM22S use SX1268
// The below isn't needed as we directly define SX126X_TXEN and SX126X_RXEN instead of using proxies E22_TXEN and E22_RXEN
/*
// FALLBACK: If somehow E22_TXEN isn't defined or clearly isn't a valid pin number, set it to RADIOLIB_NC to avoid SX126X_TXEN
being defined but having no value #if (!defined(E22_TXEN) || !(0 <= E22_TXEN && E22_TXEN <= 48)) #define E22_TXEN RADIOLIB_NC
#endif
// FALLBACK: If somehow E22_RXEN isn't defined or clearly isn't a valid pin number, set it to RADIOLIB_NC to avoid SX126X_RXEN
being defined but having no value #if (!defined(E22_RXEN) || !(0 <= E22_RXEN && E22_RXEN <= 48)) #define E22_RXEN RADIOLIB_NC
#endif
#define SX126X_TXEN E22_TXEN
#define SX126X_RXEN E22_RXEN
*/
// E22 series TCXO voltage is 1.8V per https://www.ebyte.com/en/pdf-down.aspx?id=781 (source
// https://github.com/jgromes/RadioLib/issues/12#issuecomment-520695575), so set it as such
#define SX126X_DIO3_TCXO_VOLTAGE 1.8
#define LORA_CS SX126X_CS // FIXME: for some reason both are used in /src
// Many of the below values would only be used if USE_RF95 was defined, but it's not as we aren't actually using an RF95, just
// that the 4 pins above are named like it If they aren't used they don't need to be defined and doing so cause confusion to those
// adapting this file LORA_RESET value is never used in src (as we are not using RF95), so no need to define LORA_DIO0 is not used
// in src (as we are not using RF95) as SX1262 does not have it per SX1262 datasheet, so no need to define
// FIXME: confirm that the linked lines below are actually only called when using the SX126x or SX128x and no other modules
// then use SX126X_DIO1 and SX128X_DIO1 respectively for that purpose, removing the need for RF95-style LORA_* definitions when
// the RF95 isn't used
#define LORA_DIO1 \
SX126X_DIO1 // The old name is used in
// https://github.com/meshtastic/firmware/blob/7eff5e7bcb2084499b723c5e3846c15ee089e36d/src/sleep.cpp#L298, so
// must also define the old name
// LORA_DIO2 value is never used in src (as we are not using RF95), so no need to define, and if DIO2_AS_RF_SWITCH is set then it
// cannot serve any extra function even if requested to LORA_DIO3 value is never used in src (as we are not using RF95), so no
// need to define, and DIO3_AS_TCXO_AT_1V8 is set so it cannot serve any extra function even if requested to (from 13.3.2.1
// DioxMask in SX1262 datasheet: Note that if DIO2 or DIO3 are used to control the RF Switch or the TCXO, the IRQ will not be
// generated even if it is mapped to the pins.)