Weather Station Update

Weather Station v2

Not happy with the performance of the current external weather station. Whilst it functions as deployed reading reliably the wind, rain, temp etc, is let down by the power demands. This is a flaw with the original design as I can’t exploit low power modes with the inability to determine why the device was woken from sleep being the issue. The 2.5Ahr batteries run for several days but require that it is connected to power to charge weekly.

Version 2 will feature reprogrammed GPIOs and a RTC module. Staying with the SparkFun ESP32 Thing as breadboarding is a breeze. The ESP32 modules don’t fit so a nuisance. When you order PCB (once tested) you will typically get 5. The design has been is tweaked to make more general in potential uses including;

  • replacing the 7805 regulator with a buck converter
  • exposing 5v and 2 GPIO pins to support future relays
  • additional I2C port (support OLED unit for debugging)
Weather Station v2

So currently waiting for an additional SparkFun ESP32 Thing to be delivered. Aussie post not keeping up with the demand from all the online shopping.

Power Supply

Decided to break the power off completely. This unit and a Raspberry PI will be deployed out on the tower. Have to get the wind sensors higher as is too sheltered off the deck and want to deploy a WSPR transmitter.

The Sparkfun ESP32 also has a LiPo charger built in but can’t keep up with the demand. Ok yes reduced sunlight during winter may be be in play here but let’s fix it for good.

The RPi zero wants 120 mA @ 3.3v and there is an on board voltage regulator which creates this voltage from the 5V power input. The ESP32 wants ~240mA @ 3.3V. The ESP32 can demand up to 800mA in Active mode (WiFi, the Processing Cores and the Bluetooth module) ON at all times.

Decided to go with a commercial board, new 100W 18v solar cell and 9 Ahr lead acid battery. Yes, lead acid batteries are old school but will work a treat in a static solar low drain install. Have no desire to carry up a hill and have a serviceable unit already.

Solar Cells. Have an old school full sized one (85w) and couple of smaller ~5W units. The smaller units run the internal sensors (Heltec CubeCell ASR6501) happily. Solar cells have come a long way recently and compact 100W units are quite cheap so will source a new unit.

Anyway more to come as I test the Solar Power charger board looking for RF issues, demand and resiliency.

Update – 01012020 – v2 Layout being verified.

v2 Prototype

Started receiving “Receiving ‘A fatal error occurred: MD5 of file does not match data in flash!” error on upload. Tried test ‘Hello World’ code same result. The device was in sleep mode so had to push the boot ‘GPIO0’ button to get it to wait for download. Watching the serial monitor and noted boot issues. A search for ‘RTCWDT_RTC_RESET’ was fruitful. The discussions matched the use of GPIO12 for raising ints.

16:25:15.979 -> rst:0x10 (RTCWDT_RTC_RESET),boot:0x33 (SPI_FAST_FLASH_BOOT)
16:25:15.979 -> flash read err, 1000
16:25:15.979 -> ets_main.c 371
16:25:16.334 -> ets Jun 8 2016 00:22:57

It was the use of GPIO 12 for the AS3935 that was causing the device to fail boot. Moved to 15 and it boots and can be flashed ok.

So as it stands, the device is supporting;

  • SSD1306
  • AS3935
  • BMP280
  • DS3231
  • SX1278

So now to test sleep mode and the inuts from the weather sensors (Wind & rain)