A Small Scale Solar Photovoltaic System
4 August 2007
This article describes our solar photovoltaic (PV) system, designed to provide us with a degree of independence from the national electricity grid. This system will not reduce our carbon dioxide emissions, since we already buy our electricity from Good Energy, a 100% renewable supplier. However, by reducing our load on the grid, it will help the UK share out its existing renewable energy capacity among more customers, thus increasing the proportion of the UK’s electricity needs that can be met by renewable energy.
A standalone PV system also provides us with security of energy supply, and helps insulate us from rising energy prices.
Choice of PV system
Conventional solar PV systems are grid-intertie. They generate AC (alternating current) electricity that feeds into the house mains electricity supply. If you generate more power than you consume, the meter spins backward, as that surplus electricity flows back into the grid for someone else to use. By returning surplus electricity to the grid, no battery is needed.
However, there’s a major drawback of grid intertie systems: when the grid fails, then to ensure safety the PV system is automatically disconnected. This means that your house will no longer be able to use the electricity generated by your PV array while the grid is down. Thus, contrary to expectations, you have no independence from the grid.
A lesser – but nonetheless significant – drawback of grid-intertie systems is their requirement for costly power inverters. The more complex system components and wiring needed increases maintenance costs and the likelihood of failure.
The system I chose to install is entirely independent from the grid-supplied mains electricity, and is purely DC (direct current), operating at 12 volts. It feeds into a battery, which ensures electricity is available at nighttime when the PV array is no longer generating power. It is very simple, and should therefore be very reliable.
I’ve chosen to start off with a low power system, but designed it with expansion in mind. By measuring the system performance I can then decide whether to add more PV panels to generate more power. In order to do this, I need to be able to monitor the energy the array puts into the battery.
My initial system will consist of a custom 12V lighting circuit run throughout the house, powering 5 low-energy lightbulbs, each consuming around 11 watts. I anticipate running these lights for up to 4 hours each evening.
Assuming that all the light bulbs are switched on at the same time (which is rather wasteful), the maximum power for 5 bulbs, each consuming 11 watts, is 55 watts.
In the worst case, should I leave all these lights blazing for 4 hours, they will consume 4 hours * 55 watts = 220 watt hours each evening. At 12 volts, this energy is equivalent to 220 volt.amp.hours/12volts = 18.3 AmpHours.
This power will come from a lead-acid battery. The model I have is an Elecsol Carbon Fibre 110 AHr battery which, when fully charged, can supply 110 Amp Hours over a long period (likely to be 100 hours, a standard rating since lead acid batteries give out their charge most efficiently at low currents). Unfortunately I don’t have the data to calculate this battery’s capacity when supplying 18.3 amp hours over 4 hours; but it is likely to be more than enough.
The battery is charged by 2 Unisolar triple junction panels. These amorphous panels have a good overall efficiency in the overcast and cloudy conditions we experience in the UK. The panel power ratings are 21 and 42 watts, making for a total of 63 watts.
At our latitude in the UK, the average sunshine hours in the summer is around 4 hours – ie, each day we get a total amount of sunlight that is equivalent to 4 hours of midday maximum intensity sun. Over those 4 hours, the panels will generate a total of 252 watt hours; this is just sufficient to recharge the battery for the worst case scenario.
During the winter there will be considerably less sunlight, and I expect not to be able to run all the lights all the time.
Of course these numbers are just guidelines. The real life system performance is what counts – hence the need to collect data, before deciding whether to uprate the capacity.
A pair of solar panels feeds into a charge controller, which then regulates the battery charging. When the battery is fully charged, the charge controller disconnects the solar panels to prevent the battery from becoming overcharged.
The battery runs a set of 12 volt lights. For simplicity, and to keep system losses down, no inverter or voltage converters are used.
Solar panel mount
There are a variety of ways of mounting solar panels, including roof mount, pole mount and ground mount. To keep costs down, this system uses a ground mount. I’ve built an aluminum rack using 25mm L-shaped struts available from DIY stores. This holds the two panels together, and is secured to a strong wood/concrete fence post.
The panel is oriented south, and tilted to maximise winter gain, when there’s the least solar power available.
Morningstar Tristar 60 charge controller
A charge controller is a key system component. The model I’ve selected is the Tristar 60, made by Morningstar. It can handle up to 60 amps at 12, 24 or 48 volts.
The Tristar has an RS232 computer interface, through which the controller can be programmed, and the firmware updated. Later this year Morningstar plan to update the MSView software (available from their website) to include a data-logging facility.
The Tristar is solidly built, and has a five year warranty. Installation is fairly straightforward.
Fuses, circuit breakers and fuse box
I decided to use standard electrical components for the circuit breakers, fuses and fuse box, purchased from Newey and Eyre, a UK-wide trade supplier of electrical parts.
The fuse box contains a ‘DIN’ rail – a metal strip whose cross-section resembles a top-hat, onto which fuses and switches snap on and off, making installation straightforward 10amp French industrial type fuses are held in Ferraz-Shawmut fuse holders (see Newey and Eyre’s catalogue for details).
10amp fuses are used to protect the battery, PV charge controller and load circuits individually.
I selected an Elecsol 110AmpHour carbon fibre wet lead acid battery. This can supply 110AmpHours if discharged over 20 hours (ie C20 rating) or just under 5 amps continuously for a day.
The manufacturer claims this battery will give over 1000 deep discharge cycles, and that the carbon fibre matrix prevents sulphation when the battery is left discharged for a long period, or overcharged. The matrix also prevents the battery plates from buckling under high discharge currents, so this battery could be used for very high current applications – perhaps useful in future if I were to fit a large inverter and run a 2kW electric kettle.
Wiring is a surprisingly complex area of solar PV installations, mainly because the type of cable needed is not easily available, except through specialist PV suppliers.
Most cable is designed for mains applications, which, because of its higher voltage, carries much lower current. In general, PV systems should use large diameter cables – eg at least 2.5 square millimetres, sized to the maximum current that may be carried, allowing for a safety factor (eg 25%). In general cables should be able to handle more current than the fuses are rated for, so that if an overload short-circuit occurs, the fuses blow before the cables overheat and melt.
A set of 12 volt Steca Solsum ESL 11 watt lamps are used to provide lighting. I also run a 12V fast battery charger, for keeping torches, digital camera etc. charged.
Results so far over the summer months are good; the charge controller does a good job keeping the battery topped up, and there’s plenty of power available to keep the lights running. However, during the winter I expect there will be less surplus capacity as the daylight hours reduce, and the demand for electric lighting increases. Watch this space for an update.