
Home Weather v3
The ‘Dashboard’ has received rework to improve existing and display additional information. Yes there are several home weather dashboards out there but by the time you install, register and integrate into existing systems it was more fun to build own. The interface may be blocky but I understand what is going on behind the scenes. This gives me a view as to what I’m recording, what is being recorded around me what is being forecast.
As it stands the dashboard supports;
- Recorded external conditions from Arduino device via LoRa to Pi gateway
- Pi Zero with BME680 conditions and AS3935 lightning
- Recorded internal conditions via Arduino and Pi devices using LoRa to Pi gateway
- Arduino with BME280
- Pi Zero with MS340 conditions and AS3935 lightning
- P4 based gateway using LAMP stack for storage and processing
- BOM via FTP for Radar and warnings
- Incidents by Vic Emergency RSS feed.
- Forecasts and Incidents via WillyWeather API
- Fronius Inverter status via API with JSON
- NBN throughput via speedtest cli
Device.

V3 included a voltage divider to monitor the input voltage, support for SPI RTC, Lightning detector, LoRa and SD card reader. I2C support for display and BM680. Analogue inputs for the wind, rain detectors. Multiple SPI devices provided challenges during dev however the addition of pull-up resistors on the SPI bus and CS lines fixed that. Currently the SD card has been disable in this code as the string class was consuming 32bytes each use and crashing the device. The discrete components are 805 SMD form factor on new PCBs.

define SERIAL_BAUD 115200
const bool TRUE = 1;
const bool FALSE = 0;
const int deviceID = 8;
const char* ssid = "Harvie-Home";
const char* password = "password";
const char* ntpServer = "192.168.1.1";
const char* cBanner = "Weather V4.0";
const char* outputFile = "/WX-v4.json";
define pinLED 5
// I2C
// I2C device found at address 0x3C SSD1306
// I2C device found at address 0x57 AT24C32 - eeprom
// I2C device found at address 0x68 DS3231
// I2C device found at address 0x76 BME280
define SDA 21
define SCL 22
// SSD1306 display connected to default I2C (SDA, SCL) pins.
define oledRST -1 // or use U8X8_PIN_NONE
define SCREEN_WIDTH 128 // OLED display width, in pixels
define SCREEN_HEIGHT 64 // OLED display height, in pixels
// VSPI SPI Defaults
// lora 1278x & AS3935
define SCK 18 // VSPI_CLK/SCK = GPIO18
define MISO 19 // VSPI_MISO = GPIO19
define MOSI 23 // VSPI_MOSI = GPIO23
// #define CS 5 // VSPI_CS/SS = GPIO5
// LORA
include
define loraBAND 433E6
define loraCS 4 // NSS
define loraRST -1 // connect the reset pin of the MCU 38
define pinLORA 2
// AS3935
define INDOOR 0x12
define OUTDOOR 0xE
define LIGHTNING_INT 0x08
define DISTURBER_INT 0x04
define NOISE_INT 0x01
define AS3935CS 33
define pinLIGHTNING 15
define pinRTC 16
define SDCardCS 32
define pinVOLTAGE 36
define SEALEVELPRESSURE_HPA (1013.25)
define pinVoltge 36
// Sensor variables
RTC_DATA_ATTR volatile unsigned long lastDataSent = 0; // Time since last data send
RTC_DATA_ATTR volatile int numRevsAnemometer = 0; // Incremented in the interrupt
RTC_DATA_ATTR volatile int numDropsRainGauge = 0; // Incremented in the interrupt
RTC_DATA_ATTR volatile int bootCount = 0; // Incremented on sleep wakeup
RTC_DATA_ATTR volatile int timerCount = 0; // Incremented on timer wakeup
RTC_DATA_ATTR volatile int intCount = 0; // Incremented on INT wakeup
RTC_DATA_ATTR volatile int numLightning = 0; // I
/*
Anemometer (Digital pin 2 - pin 2 of RJ11, pin 3 to ground),
wind vane (Analog pin 3 -> pin 1 of RJ45, pin 4 to ground)
rain gauge (Digital pin 3)
There are issues with using pins 32-39 to trigger wake up from deep sleep.
It will wake but you can’t determine which int its was (returns inf) whilst you can determine who called for the lower pins.
DIGITAL Anemometer - pinANEMOMETER 27 // 34
DIGITAL rain gauge - pinRAINGAUGE 26 // 35
ANALOGUE wind vane - pinWINDVANE 25 // 15 Dos not raise INAT so goes highrr
*/
define pinANEMOMETER 26
define pinRAINGAUGE 27
define pinWINDVANE 35
/*
Only RTC IO can be used as a source for external wake
source. They are pins: 0,2,4,12-15,25-27,32-39.
LIGHTNING 15 32768
RTC 16 65536
ANEMOMETER 26 67108864
RAINGAUGE 27 134217728
= 201424896
2^15 + 2^16 + 2^26 + 2^27 in hex
C018000
#define BUTTON_PIN_BITMASK 0xC018000
*/
define BUTTON_PIN_BITMASK 0xC018000
/*
Other pins. will have a header on board
pinTBA 35 Forth input on the weather sensors
pinRELAY01 34 Used to trigger relay ?
pinRELAY02 33 Used to trigget another relay ?
These wont be used to raise INTS
*#define pin037 37
*#define pin035 35
*#define pinTBA 36
The header file is above. Hopefully you will see the pin assignments align with the schematic. Its getting busy and the device is keeping up and I’m very happy with the end result.
Have two lightning sensors now deployed. Internally a DFRobot and externally a SparkFun unit. Both based on the AS3935 chip. The AS3935 Franklin chip is capable of detecting lightning up to 40km away with an accuracy of 1km to the storm front. The device is designed to pick up lightning events in the 500kHz band. Both will detect lighting events but not at the same time.
A RTC clock has been added to the external station to keep time. The time is currently set via NTP over wireless. Will be moving the station to the antenna tower. This is behind the house with the wireless behind colour bond walls. LoRa will be used so send readings however expecting wireless connectivity to be unavailable.

SD card unit has been added to support data logging. Current code is not deployed as using strings and they are fragmenting the heap. Both devices now read and send back the heap so I can monitor. Current code is solid as heap does not change.
The Hall sensor is recorded and sent back. Not sure what to make of these readings. They are indicating changes in the presence and magnitude of a magnetic field using the Hall effect. Currently the data looks random.
The voltage divider is using two 10k resistors from VUSB. Need to take an average to come up with a usable number.
voltage = ((float)sum / (float)NUM_SAMPLES * 5.015) / 1024.0;
Still seeing occasionally hangs for the external device . The BME680 sensor stops sending data. Suspect the unit currently deployed is marginal as a no name unit. Have a replacement unit here. This replacement unit is a SparkFun unit so expecting better quality but will need to update the code to use the SparkFun library before replacing.
Currently the external station is still on the edge of the deck. This spot is too protected to take usable wind speed readings. Not an accident as the house was positioned below the ‘ridge’ and we planted and grew a wind break (Koala trees) .
Happy to share code or KiCAD files on request. No promise they are of any value. There is also two spare boards available. Send me an email if you intent to use.
V4 updates will include;
- Multiple sources (APRS via RF for redundancy when iNet down) for alerts
- Automated control of battery charging when storms approaching (webiopi)
- Additional air quality and CO2 level reporting
- Update BME680 code with SparkFun libraries
- Calculations to make sense of the voltage readings
So yes could have purchased a commercial unit by now but where would the fun be in that ?