Various OLPC related trips to Haiti brought home to me how difficult it is for non-technical people to use lead acid storage batteries properly. Quite often the people who purchase and install battery systems are not the same ones that use and maintain them day to day. So the idea that deep cycle batteries should only be used in the top half of their discharge-charge range is often not acted upon, or even part of the user’s awareness.
Even if keeping batteries charged to a level of 50% or greater is known, the means of achieving that are not easily at hand. When I had forgotten to bring my hydrometer on one trip, I went to buy one locally. The largest building supply store in Port Au Prince, where storage batteries were sold in high volume, did not sell hydrometers.
To determine the health of a battery array requires taking the specific gravity of each of the cells. In my experience the batteries are in small enclosed spaces, with no windows, and or the power is off. So the process involves a flashlight, pencil and paper, and a hydrometer, trying all the time to keep the battery acid off the clothes, paper, and anything else that might be damaged.
Another way to determine the state of charge is using a voltage reading. If the battery has been resting for at least 4 hours, the top curve (C/100) approximates the relationship between voltage and contained charge (but only at a specific temperature).
If a voltmeter ever existed at most of the schools and orphanages I visited, it had been misplaced or stolen long ago.
And from the slope of the curve between 70% and 100% it is really hard to get required information in exactly the portion of the curve where the engineer would like the charging system to be operating most of the time.
In order to learn about solar panels, deep cycle batteries, and charge controllers, I purchased these items and set them up in my bedroom. What I discovered was that it was very easy to see when a battery was fully charged. It may just have been the accident of the specific type of charge controller which I happened to buy. It is what is called a pulse width controller. As long as the battery terminal voltage is less than 14.5, the solar panel is connected to the battery. Once the terminal voltage rises to 14.5. the controller disconnects the battery, until the battery voltage falls to 13.75, at which time the solar panel is again connected. The duty cycle of this connect/disconnect activity is a really clear indication of almost 100% charge. Initially it might be charge for 20 seconds, disconnect for 5 seconds. At the end, the battery voltage rises very rapidly, and the connection only lasts for less than a second, and the disconnected state remains at about 5 seconds.
For a school server to be available 24/7, it needs enough stored energy to last through a stormy period — say 3 days. The most efficient school server, the cubox with disk, needs 5 watts. So to last for 3 days, the batteries need to provide 360 watt-hrs, or expressed as a battery related value 30Amp-hrs. The standard sized battery next larger than 60Amp-hrs (twice the required energy, if we only use half of total stored) is 100A-HR battery. 100A-Hr is the capacity of the battery . The load on the battery from the Cubox is certainly less that 1 Amp. So this tells us that we should be using the “C/100” curve on the figure above. The conclusion of all this is that our arduino battery monitor should disconnect the school server from the battery right at the point where the “C/100” line crosses the 50% axis. Reading from the left column, it should do the disconnect when the voltages falls to 12.5V.
This is the 3/13/2014 desulfator/shunt/low voltage disconnect: battery-nurse