Project Curacao2 – Part 9: Status / Shipped for Deployment

May 24, 2017 Arduino, Grove, Pi Camera, Project Curacao2, Raspberry Pi, SwitchDoc Products, Weather, WXLink

Project Curacao2 – Part 9: Status / Shipped for Deployment

The Project Curacao2 Family

We have packed the Project Curacao2 system for shipment.   Only one hardware issue remains and it can be easily fixed down in Curacao when received.

Solar Power
New Solar Panels on Top of the Project Curacao Box – WeatherRack in Background

Project Curacao2 is a redesign and rebuild of the Project Curacao environmental monitoring system that was running down on the island nation of Curacao in 2014 and 2015.   It finally died after a loose wire in the solar panel assembly (not in the panels, but in the internal wiring) finally came completely loose.   We are now using a LoRa WXLink wireless transmitters for data collection.

The project consists of two LoRa transmitters based on a Mini Pro LP and a Raspberry Pi Zero powered base unit.   Lots and lots of sensors!   All solar powered.

 

Project Status

The ProjectCuracao2 project is ready for deployment and is ready to start sending data as soon as it is taken out of the box and mounted.   One system that has been thoroughly tested and vetted is the remote update mode, so we can supply software updates easily to the Raspberry Pi Zero base unit.  It is very solid.   We are planning (just as they did on the Mars Curiosity rover mission to Mars) to supply updates remotely from 3500 miles away to complete the entire software package.

Hardware Status

The entire hardware system has been tested for functionality and it all checks out with the exception of part of the Solar Power system.    What is the trouble with the Solar Power system?   In very bright sun, the system can experience what we call “Solar Panel Voltage Collapse”.   When a solar panel is fully exposed to the direct sun the solar panel voltage can collapse during high current draw.    This shuts off the charger, causing the solar panel voltage to recover and then causing the collapse again.   It makes the charging process not stable.   It reduces the efficiency of the solar power system during the exact time you have the most solar power available to charge the battery.

This is not a problem in most cases, but we are occasionally seeing this during our full sun testing of the Solar Panel system.   We use the SunAirPlus board (measures battery voltage/current, solar panel voltage/current and load voltage/current) making the collapse problem easily detected.   We have a couple ways of addressing the collapse problem (including just leaving it alone and putting up with the recduced efficiency) and will report on this in the coming weeks.

 

Software Status

We used the GroveWeatherPi software as the starting point for ProjectCuracao2 software.   We used the apscheduler based software to build our additional tasks and added software for the new sensors included (the PiCamera, Air Quality Sensor, etc.).   It is basically working.  We put most of the effort before shipment into the remote management system (well, duh) and the solar power system, including reducing power using some innovative techniques by using Grove PowerSave boards to shut off unneeded devices.  Cut the current requirement from 420mA to 200mA.   We ran out of time to make the LoRa receiver into a system that we could shut off (see part 8 of the the ProjectCuracao2 series).   Software issues and updates follow:

 

  1. We occasionally get a bad read from an I2C device.   Looking at the problem, it is clear that occasionally we have two processes trying to read the I2C bus at the same time.  While this doesn’t crash the software, it should be fixed.  Either by changing the schedule, or more correctly, adding a semaphore to tasks that make I2C calls so we can make all of our processes work all the time.
  2. Adding the software to periodically take Air Quality Sensor readings.  We have thoroughly characterized this sensor and are comfortable adding it remotely.
  3. Adding temperature and humidity control inside the base unit using the 5V fan.   We have tested the fan and know it works.
  4. Updating the RasPiConnect interface from Project Curacao – again easily done remotely.
  5. Adding additional graphs for the two remote LoRa WXLink sensors.  We are storing all the data in MySQL that is received from the two external units, so this is easy to do using MatPlotLib
  6. Adding the code that will ftp a new picture from Curacao every hour during the day.

 

Deployment

It will take about 5 days to get the units to Curacao and so deployment will happen shortly after June 1st, 2017.   It is planned to set up all three units on the ground, run a multi-day test,  look at the solar power performance, test the range of the LoRa WXLink units with the Yagi attached to the base unit and finally place the sensing units and the base unit on their respective radio towers.   Stay tuned.