The past few weeks have made it clear that out 2 solar panels weren’t enough at all for the Winter months. Short and overcast days weren’t enough to keep the battery charged with even very minimal use and we went dark a couple of days. I don’t have numbers yet but these 3 panels should do a lot better than the 2 I had. I also doubled the batteries’ amp-hours to have more buffer.
A solar powered blog
This blog is now powered by a Raspberry Pi using 100% solar energy. Nicole instrumented the Phidgets sensors so we would gain some visibility into our electricity production & consumption. This has already given us some great insights. We can see the effect that each device we use has on the system: how much the LED lamps take to charge, the hole that the inverter blasts through the battery when turned on. We can tell that not all sunny days are created equal in their ability to give a charge. We can even tell the increase in electricity consumption that rsyncing a whole bunch of data to the Pi has: 0.03A.
The sensors
- solar panels volts (a good indicator of sunlight)
- input amps (indicates when the charge controller uses produced electricity)
- output amps / load (what we consume with various devices)
- battery volts (whether this blog will make it through the night or not)
For now I’m only graphing using the Gnuplot one-liner from Hell. More to come…
It blows my mind way too hard that I have a system in which sunlight comes in and organized information comes out. And by organized information I mean lolcats.
Solar install tidied up & half instrumented
Charge controller (top), Raspberry Pi (top right), Phidgets interface kit (bellow the pi), Phidgets current sensors (bottom).
My new favorite project screws
The volt-meters (left) aren’t live yet but the amp-meters (bottom) are.
With the inverter hooked up and a properly fused distribution box
The Pi reading its sensors, Nicole is taking on the data aggregation.
The goal is to move Akrin to the Pi to have a solar powered server.