Project Curacao Update – Loose Wire in the Caribbean

Project Curacao Update – Loose Wire in the Caribbean

Project Curacao has now been deployed for five weeks in the Caribbean heat.   The box has been operating extremely well but an issue has cropped up in the past week.

What is Project Curacao?

Project Curacao is a massive Raspberry Pi/Arduino solar powered project first published in MagPi magazine. It was first deployed in March 2014, lasted until a RFI (Radio Frequency Interference) incident in November 2014, came back to Washington state

Solar Power
Project Curacao Box – WeatherRack in Background

for a refit and is now back in the Caribbean.

Project Curacao uses the following SwitchDoc Labs products:

There are two independent power systems on board.  Two solar panels connected through SunAirPlus for the Battery WatchDog Arduino and three solar panels for the Raspberry Pi A+ in the box.   The new issue is with the solar panels and the Raspberry Pi A+.

Solar Panels and the Raspberry Pi

Project Curacao uses three VoltaicSystems 3.4W solar panels to charge up a 6000 mAh LiPo battery through a SunAirPlus solar power controller.  Everything was working very well for the first four weeks, then we saw the current coming in from the solar panels during mid day drop to less than 1/3 of the previous value.   It still was charging, but now the Raspberry Pi could only stay on for about 3 hours a day versus the 10 hours before.   The BatteryWatchDog was picking up the lower battery voltage and shutting down the Raspberry Pi (safely) and turning off the power supply because of the lower power available.   That part of the system was working perfectly.

So What Happened?

Take a look at this set of values from the RasPiConnect control panel:

IMG_0978

 

See how on July 13, 2015 the current from the solar panels dropped to ~100ma from a maximum of ~800ma the day before?  After looking at the real time values showing on the lower part of the screen for a few minutes it was obvious what had happened.   We had lost connection with the two solar panels connected to the Raspberry Pi on the front of the box and only the top panel was generating power.   We were sure that it was a loose connection, but from 3500 miles away, there was not much we could do.  The wire is clearly very, very close to making connection.  Every couple of days it reconnects (heat, wind, vibration, birds?) and we have good power for a few days as you can see from July 16, 2015 on the graph above.

That is what a loose wire looks like from 3500 miles away.   We have a person visiting the box in late September 2015 and we will ask them to open up the box and tighten the wires for those two solar panels.

From a system perspective, the box is doing exactly what it should be doing.  Changing the box behavior according to the power available.IMG_0980  In the log files, you can see the Battery WatchDog Arduino refusing to start up the Raspberry Pi in the morning because the battery is too low.  Very cool.   The chart below shows the Weather Rack data coming into the Raspberry Pi.  Why aren’t there gaps?  It is because the Battery WatchDog Arduino is gathering the weather data in a FRAM memory on the WeatherPiArduino board and relays it to the Raspberry Pi after the Raspberry Pi boots up.

A Picture of Paradise

We will end this post with a great picture of the Caribbean hot off the PiCamera in Project Curacao.

IMG_0977