ParaTNC is a hardware and a software which offers the functionality of multi function APRS controler.
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
Mateusz Lubecki fbaf9df8ed retrieving raw results from pt100 sensor 12 hours ago
.settings api: mac calculation 2 weeks ago
STM32F100_ParaTNC daily 1 week ago
STM32L476_ParaMETEO daily 1 week ago
doc timeout during RTC initialization amd low vbatt cutoff patch 3 months ago
hardware waiting for a while if more than one frame is transmitted from single packet_tx_handler call 2 years ago
include max31865 pooling and state machine. warning! it seems that JTAG hangs up after few seconds of debugging 2 days ago
ldscripts bufix stm32f100: do not read a content of DR register in IDLE state handler within irq handler, callback and pooling callback for KISS_GET_RUNNING_CONFIG 1 month ago
src max31865 pooling and state machine. warning! it seems that JTAG hangs up after few seconds of debugging 2 days ago
system retrieving raw results from pt100 sensor 12 hours ago
.cproject spi expiremnts almost working 3 days ago
.gitignore gitignore and backup copy of config file 1 year ago
.project project config 4 years ago
LICENSE Create LICENSE 2 years ago
README Update README 5 months ago
README-pl documentation 2 years ago
analog_windspeed_after_wakeup.png reinitializing analogue anemometer each wakeup event 1 year ago
makefile.init compilation fixes 5 - makefiles and new path to the toolchain 2 years ago
parameteo-test-raw-packets.txt configuration files reorganized 6 months ago
skalowania.ods Sparkfun SEN15901 aka PM1S analog/mechanic anemometer 2 years ago
windspeed_pulses.png reinitializing analogue anemometer each wakeup event 1 year ago


ParaTNC version DF15, April 4th 2021

// Please look into 'doc' directory for more documentation and user manuals
// WARNING: This readme could be sligtly outdated due to my lack of time :)

// Ongoing changes
// ---------------
// Currently intense refactoring is done, which will result in new main software version EA00.
// EA00 will remove the support for any Hardware Revision before HW-RevC, so it won't work on STM32VLDISCOVERY
// anymore. Conditional compilation will be removed, so EA00 and all consecutive version will be build as
// one binary with support to everything. The configuration will be kept in Flash memory and could be
// altered 'on the fly' by using some external (not yet started) software tool, without a need to recompile
// and reflash the controller.

ParaTNC is a hardware and a software which offers the functionality of multi function APRS controler.
The software itself can also run on cheap STM32VLDISCOVERY board with slightly reduced functionality, but
with costs about 10..12 euro this gives the cheapest possible standalone APRS controller.

ParaAPRS supports key elements of what good APRS device should have:
-> Two directional KISS TNC (no init strings required).
-> WIDE1-1 digipeater, with an option to limit digipeating only to SSIDs 7, 8 and 9.
-> Weather station with support for various meteo sensors. Full list od supported devices in point 5.
-> Extensive telemetry with an information about the count of receved, transmitted and digipeated frames plus status of weather sensors.
-> Support for VE.Direct serial protocol used in Victron PV charging controllers. The data about currents and voltages in the PV system are transmitted using APRS telemetry.
-> Support for UMB Binary porotocol as a mater.

ParaAPRS can be used as a standalone weather station controller w/o using the radio transmission. The sw
by default sends the measuremenets values

The ParaTNC software can run on STM32VLDISCOVERY board but it really shows it full potential od it's own
dedicated hardware. This not only rub out all bodge wires required with prototype board. It gives
few features which are hardly not possible without dedicated PCB
a) Two different grounds separated one from the another (HW Revision A and B)
-> Controller
-> DB9 connector used to interface with a radio and Wind & Temperature sensor
aa) ONLY IN HW REVISION C: Three different grounds separated one from the another
-> Controller
-> DB9 connector used to interface with a radio
-> Wind and temperature sensor including RS485 bus
b) Isolated DC-DC coverter to supply sensors with an option to connect another external power supply itself
c) A relay used as a watchdog to reset wind & temperature sensor in case that no communication will be received.
d) RS232 serial port avaliable on Cisco style RJ45 connector.
e) RS485 serial port interchable with RS232 (with HW Revision A) or working fully independly (HW Revision B)
f) PCB designed to fit into Mikrotik RB411/RB711 enclosure.

ParaTNC PCB could be manufactured used a set of generber files locates in ./hardware/ directory a long with schematics and
layout renders (top and bottom layer).

As usually every design has some flaws, less or more important. In case of ParaTNC they can be divided into two
parts, separately for software and hardware. Firstly in case of software an end user should USE ALWAYS THE SOURCE CODE
WITH THE NEWEST TAG marked as non-beta (like DE07). ParaTNC repository is used for daily development, so the last commit
can even sometimes don't compile at all, or consist some major bugs. 

Hardware workarounds in HW Revision A:
-> Missing resistor in series with base of transistor T2. External 1k must be added in series with TX line
-> Foortprint location of R7 (RX_LVL) and R8 (TX_LVL) helipots are 'not optimal'. If 'normal' pots are to be used the vertical
version must be chosen as flat one won't fit due to neighbouring components location.
-> J1 power-in jack has reversed polarity (negative center)
-> The chineese manufacturer dropped the TX20 anemometer from the production which means that a supoprt to it
became a little bit obsoloete for new installations. Support for Davis 6410 anemometer or any another with
pulse output for windspeed and potentiometer (resistance output) for direction is ongoing and should be
avaliable in HW-Revision B and further SW releases.

Hardware workarounds in HW Revision B:
-> Few missing traces in micro supply. Few bodge wired needs to be added. Please look at hardware-revision-b-errata.jpg in 'hardware'

ParaTNC software and hardware are licensed under terms included to the source code in 'LICENSE' file

a) Wind sensors:
- TX20 (not in manufacturing anymore)
- Davis 6410 or any another analogue anemometer with resisntance as a direction output and impulses as a windspeed
WARNING: Not avaliable in Hardware Revision A due to lack of circuitry. Fully working only in HW-Rev C
- Lufft V200A/Ventus or any other UMB compatible device.
- Any Modbus-RTU sensor
b) Pressure Sensors:
- MS5611
- Bosh BMP280
- Any UMB protocol compatible device
- Any Modbus-RTU sensor
c) Temperature sensors:
- Dallas One wire DS18B20
- Any UMB protocol compatible device
- Any Modbus-RTU sensor
d) Humidity Sensors:
- DHT22
- Bosh BMP280
- Any Modbus-RTU sensor (F&F MB-AHT1 is recommended)

Most of Victron devices have a support both for binary and text serial protocol. By default the text procol (VE.Direct) is
always enabled and a device will send from its own telegrams each couple of seconds. The communication via VE.Direct is
avaliable through dedicated socket on the charging controller which is just 3.3V TTL levels UART, so no external ICs is
required to connect the PV controller to an evaluation board. In the MPPT series the comm socket is located on the bottom
side of the charging controller below the fuse holder.

Exact pinout of the VE.Direct comm socket is as follows, assuming that terminal screws are facing down:
-> Ground
-> TX, data from host to the PV controller
-> RX, data from PV controller to the host

The controller sends a lot of different data which cannot be completely transmitted through APRS network due to radioprotocol
limitations. Only these parameters are transmitted:
-> Battery Current (charging as positive, discharging as negative) in third channel of telemetry.
-> Battery Voltage as fourth telemetry channel.
-> PV cell Voltage as fifth telemetry channel.
-> Charging Controller status (like current charging mode) and minimum/maximum battery current in last 10 minutes.
-> Error Codes if any (short circuit, overheat, overvoltage on PV or battery input)

If the VE.Direct protocol suppot is disabled the ParaTNC uses telemetry packets to send internal statistics and
general information about communication with sensors. Telemetry works in 10 minutes cycle, which means that packets are sent
every 10 minutes and they represents a state for the time between one and another.

Analog channels:
1st Channel: Number of packets received in 10 minutes.
2nd Channel: Number of packets transmitted in 10 minutes (digi-ed + generated internally by the controller)
3rd Channel: Number of digipeated packets in 10 minutes.
4th Channel: Number of packet received from Host by KISS protocol
5th Channel: Optional temperature from Dalls One Wire sensor.

Digital Channels:
DS_QF_FULL - Quality Factor for One Wire temperature sensor is FULL
DS_QF_DEGRAD - Quality Factor for One Wire temperature sensor is DEGRADATED
DS_QF_NAVBLE - Quality Factor for One Wire temperature sensor is NOT AVALIABLE
MS_QF_NAVBLE - Quality Factor for MS5611 pressure sensor ins NOT AVALIABLE
DHT_QF_NAVBLE - Quality Factor for DHT22 humidity & temperature is NOT AVALIABLE
WIND_QF_DEGR - LaCrosse TX20 anemometer driver dropped at least one measuremenet due to excesive slew rate.

Explantion of Quality Factors: Each measuremenets signal is kept along with the Quality Factor which represents the sensor condition
and value validity. If the QF is set to FULL it means that no communication problems happened between one telemetry cycle and another
DEGRADATED is set if at least once the communication problem happened (CRC calculation fail, timeout etc). NOT AVALIABLE is set
if all communication atempts failed and no valid measuremenet value is avaliable. By default each Quality Factor is set to
NOT AVALIABLE after the telemetry packets are sent.

The TX20 anemometer driver has a Slew Rate Limiter which keeps out all 'spikes' in values mostly due to RF interference. if the
difference between two consecutive measuremenets excedes hardcoded limit (9m/s) the software drops the value.

At this point ParaTNC is delivered in form of source code which needs to be manually compiled by the user. Most options are
configured by #define in './include/station_config.h' and then hard-coded by the C preprocessor during compilation. An example file
consist a lot of comments which explains what is what, but generally an user can choose there which mode should be enabled:

Hardware Revision A:

Hardware Revision B & C:

In Hardware Revision A there is no option to use a KISS modem and the VE.Direct protocol in the same time as only
one UART in the micro is used and handled by software. In hardware revisions B and C two UARTS are used and handled by
the software. The KISS modem is always enabled and avaliable on RJ45 Cisco style jack.

The KISS modem runs on default speed of 9600 bps. Telemetry is enabled by default and it will
trasmit channels values each 10 minutes and full channel descriptions each 70 minutes.

To build the ParaTNC software 'GNU ARM Embedded Toolchain' is required. This set contains gcc-arm-none-eabi compiler,
gdb debugger, linker, HEX generator and set of libraries. ParaTNC is developed in Xubuntu 16.04LTS and 20.04LTS
using toolchain in version 2018-q2. Please take note that You have to use 64-bit version of the operation system
as the 32-bit variant of the toolchain is not avaliable.

Exact GCC version is 7.3.1 and it present itself with a string like this below
gcc version 7.3.1 20180622 (release) [ARM/embedded-7-branch revision 261907] (GNU Tools for Arm Embedded Processors 7-2018-q2-update)

Self conatined TAR.BZ2 archive with all required tools can be found here:
in 'What's new in 7-2018-q2-update' section pulled down from the list located in the middle of this site.
Alternatively You can use this link:

After download the content of this archive has to be uncompressed into: /usr/local/bin/gcc-arm-none-eabi-7-2018-q2-update/
so the structure should looks like this

mateusz@mateusz-ThinkCentre-M720q:/usr/local/bin/gcc-arm-none-eabi-7-2018-q2-update$ ls -la
total 24
drwxrwxr-x 6 mateusz mateusz 4096 paź 10 08:35 .
drwxr-xr-x 7 root root 4096 paź 10 18:06 ..
drwxr-xr-x 6 mateusz mateusz 4096 cze 22 2018 arm-none-eabi
drwxr-xr-x 2 mateusz mateusz 4096 cze 22 2018 bin
drwxr-xr-x 3 mateusz mateusz 4096 cze 22 2018 lib
drwxr-xr-x 4 mateusz mateusz 4096 cze 22 2018 share

Both a makefile and an Eclipse project are configured to look for toolchain in this directory. In some cases to perform
a debugging in Elipse You will have to install libncurses5 library which is required to start GDB. To check if
this is a case try to start the debugger manually by issuing such command in the prompt

'/usr/local/bin/gcc-arm-none-eabi-7-2018-q2-update/bin/arm-none-eabi-gdb --version'

If such result will be printed in the console, libncurses5 must be installed, prefferably using system package
manager like aptitude in Debian/Ubuntu

'libraries: cannot open shared object file: No such file or directory'

To start debugging session in Eclipse you must create new 'GDB OpenOCD Debugging' configuration and set paths
to OpenOCD and GDB in 'Debugger' tab. OpenOCD usually sits in '/usr/bin/openocd', the path to GDB is shown in
the paragraph before. Remember to set 'Config Options:' to tell the OpenOCD what JTAG adapter/debugger is used
in Your setup. If You're using ST-Link/V2 paste '-f interface/stlink-v2.cfg -f target/stm32f1x_stlink.cfg'


When everything is installed the reporistory can be cloned to local harddrive by using a command
'git clone'

The example config file is named 'station_config_example.h' and should be edited and then renamed to
'station_config.h'. When everything is configured it is a time to go to 'Debug' directory and invoke
command 'make' there. The source should be automatically compiled and new hex file 'ParaTNC.hex'
should appear in the same directory. Please do not use Release configuration!! It is screwed and produces an
application which doesn't works exactly as it should because of optimalization configuration.

Finished building target: ParaTNC.elf

Invoking: Cross ARM GNU Create Flash Image
arm-none-eabi-objcopy -O ihex "ParaTNC.elf" "ParaTNC.hex"
Finished building: ParaTNC.hex

Invoking: Cross ARM GNU Print Size
arm-none-eabi-size --format=berkeley "ParaTNC.elf"
text data bss dec hex filename
50542 576 5544 56662 dd56 ParaTNC.elf
Finished building: ParaTNC.siz

22:29:38 Build Finished (took 13s.231ms)

If You don't have a JTAG programmer/debugger or You just not want to or can't use it for any reason, You can choose
Internal Serial Bootloader provided by STMicroelectronics. It's code is stored in mask ROM within microcontroler
and can be used anytime, and in scope of ParaTNC it practically cannot be disabled or locked. Please remember that
this option is avaliable only if You're using ParaTNC PCB as STM32VLDISCOVERY board doesn't have TTL-RS232 converter.

To use serial bootloader You need:
- RJ45 to DB9 Cisco style RS232 cable
- FlashLoader Demonstrator software provided by STMicroelectronics for free

The required software can be downloaded from this link:
It is advised to read the user manual for this tool before proceding further. To jump to the bootloader You shall
disconect power to ParaTNC, then use something to short the jumper located at the left of JTAG connector (top-left
corner of PCB) and with jumper shorted reconnect the power supply. If procedure success You shall NOT hear
relay clicking and LEDs blinking for a short while. After the power supply is connected and micro stays in bootloader
you can disconnect the jumper and start the FlashLoader software to download the HEX file to micro. After process is done
you should do a cold reset without jumper shorter.

The STM32VLDISCOVERY board has an ST-Link/V1 programmer-debugger on board which can be used to load a HEX file.
This ST-Link appears normally as a mass storage device which makes in unusable to be used by HEX loadin software
in Linux (as the device will be 'blocked' by the mass-storage driver). To workaround this problem, a configuration
of modprobe daemon should be changed to ignore the ST-Link and not load any driver for it.

This is done by invoking a command:
'sudo echo "options usb-storage quirks=483:3744:i" >> /etc/modprobe.d/stlink_v1.conf'
What will create a new file called 'stlink_v1.conf' in modprobe directory. After the system reboot changes should
be applied and the programmer should be free to go. The kernel log should looks somehow like this below
[90639.895886] usb 2-1.1: new full-speed USB device number 13 using ehci-pci
[90639.990288] usb 2-1.1: New USB device found, idVendor=0483, idProduct=3744
[90639.990294] usb 2-1.1: New USB device strings: Mfr=1, Product=2, SerialNumber=3
[90639.990296] usb 2-1.1: Product: STM32 STLink
[90639.990298] usb 2-1.1: Manufacturer: STMicroelectronics
[90639.990300] usb 2-1.1: SerialNumber: Qÿr\xffffff86RV%X\xffffffc2\xffffff87
[90639.990796] usb-storage 2-1.1:1.0: USB Mass Storage device detected
[90639.992973] usb-storage 2-1.1:1.0: device ignored

The next step is to install texane-stlink which supports the ST-Link/V1 programmer and can be used to read an write
the flash memory. Installation is quite straight forward
'git clone git://'
'cd stlink.git'
'cd build/Relase'
'sudo cp st-* /usr/bin'

At the end the HEX file can be loaded into the microcontroler by typing a command
'sudo st-flash --format ihex write /dev/sr0 ParaTNC.hex'