A Guide To Photovoltaic Panels
A Guide To Photovoltaic Panels
The following discussion is based around autonomous and semi-autonomous systems that use PV panels
to charge a bank of lead-acid batteries.
Semiconductor solar cells convert sunlight into electricity using the photovoltaic effect. Incident light falls
on the cells and creates mobile charged particles in the semiconductor which are then separated by the
device structure to produce electrical current.
The vast majority of solar cells are made from crystalline silicon. Single crystal cells are the most efficient
however, cheaper multicrystaline cells are also popular. Even cheaper amorphous silicon cells are also
available and used widely for small consumer products but rarely used for power systems.
A single PV cell will produce between 1 and 1.5W at a voltage of 0.5 to 0.6V under standard test
conditions. Standard test conditions are: an irradiance of 1kW/m2, standard reference AM1.5 spectrum*,
and a cell temperature of 25°C.
[* The AM1.5 spectrum is the spectrum (i.e. range of wavelengths) provided on a clear day by the sun
when its rays have passed through an average depth of atmosphere to the Earth's surface.]
Figure 1: A typical I-V curve for a silicon photovoltaic cell. Figure 2: A typical power output vrs. voltage curve for a
silicon photovoltaic cell.
In Practice PV cells do not operate under standard conditions. The two parameters that have the most
bearing on their performance are (i) temperature and (ii) irradiance.
(i) Temperature
Figure 3 shows the effects of temperature on the I-V curve of a PV cell. ISC increases slightly with
temperature by about 6μA per °C for 1cm2 of cell, this is so small that it is normally ignored. However, a
more significant effect is the temperature dependence of voltage which decreases with increasing
temperature. Typically the voltage will decrease by 2.3mV per °C per cell.
Figure 3: Temperature effects on the I-V curve of a PV cell. Figure 4: The effects of irradiance on the I-V curve of
PV cells.
(ii) Irradiance
Solar irradiance is a measure of the sun's energy, under standard conditions the amount of energy reaching
the Earth's surface on a clear day is taken to be 1kW/m2. The amount of irradiance reduces with the
slightest amount of haze and becomes quite small on over cast days. ISC is directly proportional to the
irradiance: so that if irradiance halves so does ISC. The voltage variation is very small and usually ignored
The power produced under different conditions, as a function of voltage, is shown in figure 5. Figures 4
and 5 clearly show that the voltage at which PMax occurs does not vary much with irradiance.
PV cells are connected in series to produce PC panels. These usually contain 36 or 72 cells to match 12
and 24V systems respectively. 36 cells in series will produce a panel rated at about 75W and 72 cells will
produce a panel that is rated at about 160W. Panels must be able to produce a voltage higher than that of
the battery bank (the nominal system voltage) otherwise the batteries will not charge, panels for a 12V
system normally have VOC in the region of about 17V. Values of ISC for panels will vary from make to
make but will be approximately the same for a single cell, 36 cells or 72 cells. The I-V curve for a panel
therefore looks the same as that for a single cell, only the voltages are larger (figure 6).
PMax is the preferred point of operation however, if the temperature is too high this may not be possible. If
a voltage below PMax, in the linear section of the I-V curve (figure 6), is acceptable the effect of
temperature can be eliminated and the output current is dependent only on irradiance. Some modern
charge controllers have maximum power point tracking that will alter the voltage across the panel to find
the maximum power output for any given conditions. Other charge controllers rely on a charging voltage
being set manually (e.g. 15V for a 12V battery bank) and you will have to take whatever current is
available at that voltage.
Manufactures provide data for ISC, VOC and PMax, also the characteristic I-V curve can usually be obtained.
These figures are quoted for standard conditions: and irradiance of 1kW/m2, spectral distribution of
AM1.5 and a cell temperature of 25°C. Panels are never used under perfect standard conditions and the
manufactures' data must be altered to find the true power output under relevant conditions. Figure 6
illustrates how a PV panel's output changes with temperature and irradiance, this curve if for a typical
panel from a 12V system.
Specification sheets may quote a value for the Temperature Coefficient of Voltage for particular makes,
for example for a BP 585 panel it is -80±10mV per °C. Note that this is almost exactly the same as
-2.3mV when multiplied by 36 cells.
The voltage at the maximum power point (VM) does not vary much with irradiance and can be estimated
as 80% of VOC under standard conditions.
Current
ISC is directly proportional to irradiance (G). Therefore the short circuit current at the given irradiance (ISC
(G)) is given by:
ISC does not vary much with temperature and this effect is normally ignored. However, manufactures'
specification sheets often provide a Temperature Coefficient of ISC, for example this is 0.064±0.015% per
°C for a BP 585 panel; this is about 3.25mA for a 36 cell panel.
Cell Temperature
The cell temperature will normally be higher than the air temperature because they are black and sitting in
the sun. The cell temperature under different conditions can be estimated using the Normal Operating
Cell Temperature (NOCT), which is defined as the cell temperature under the following conditions:
irradiance of 0.8kW/m2, spectral distribution AM1.5, ambient temperature 20°C and a wind speed < 1m/s.
NOTC is normally in the region of 42 to 46°C.
The following equation can be used to calculate the difference between the cell temperature TC and the
ambient temperature TA (ambient temperature is the air temperature measured in the shade):
NOTC 20
TC TA G(in kW/m 2 )
0.8
PMax VM I M
However, manufactures’ usually only provide us with ISC and VOC, also even if we did know IM and VM
under standard conditions they will change for different conditions. A scaling factor called the Fill Factor
(FF) is used, once calculated it can be used to scale the modified values of ISC and VOC to find PMax for the
true operating conditions (this makes the assumption that FF does not change with temperature or
irradiance).
Figure 7 shows an I-V curve, PMax can be found by maximising the area of the rectangle IMPMaxVM0.
The following equation is used to find FF from the manufacturers' data, which can then be used to find
PMax under non-standard conditions.
PMax VM I M FFVOC I SC
Example
Determine the parameters of a panel formed from 34 cells in series, under the operating conditions G =
700W/m and TA = 34°C. The manufacturer's values under standard conditions are: ISC = 3A, VOC = 20.4V,
PMax = 45.9W, NOCT = 43°C.
Other Reductions
The manufacturing process is not perfect and some PV panels will be rated slightly higher than others so
that the total power of an array of panels will be slightly less than the down rated power of a panel
multiplied by the number of panels. This phenomena is known as mismatch and the array output should be
down rated by the manufacturing tolerances; normally in the region of 2-4%. Another 2 to 6% can be
subtracted for a further mismatch caused by dirt and dust on the panels.
Most small PV systems have the panels fixed so that they do not track the sun across the sky throughout
the day. For fixed panels the maximum power output can be achieved when their surfaces are
Figure 8: Panel angles for latitudes from 70° to -70°, negative panel angles indicate that panel is inclined to face north.
perpendicular to the sun at solar noon. Note that due to the vagaries of national time keeping solar noon is
unlikely to be 12pm, rather it it the point in time at which the sun is at its daily zenith.
Panels should be fixed on the north-south axis since at all times of year, at solar noon, the sun will be
directly on this line. If your system is in the northern hemisphere and above the Tropic of Cancer (i.e. has
a latitude greater than +23.45°) your panels will always be inclined to face south because the sun's daily
zenith will always be in the southern skies. Similarly, if your system is in the southern hemisphere and
below the Tropic of Capricorn (i.e. has a latitude less than -23.45) your panels will always be inclined to
face north. Note that outside the tropics the sun is never directly overhead.
If your system is in the tropics (i.e. has a latitude between -23.45° and +23.45°) matters are not so simple.
At the equator the sun is directly overhead at solar noon on the equinoxes (21st-23rd March and 22nd-23rd
September), but reaches its daily zenith in the northern skies from March to September and the southern
skies from September to March. In the northern tropics (i.e. latitudes between 0° and +23.45°) as the
latitude increases the sun follows a similar pattern, although it will be directly overhead on days that
approach the summer solstice (21st-22nd June). If your site is on the Tropic of Cancer the sun's daily zenith
will always be in the southern skies and will be directly overhead on the summer solstice. The sun in the
southern tropics (i.e. latitudes between 0° and -23.45°) is in the southern skies for some of the year
between the autumnal equinox and the vernal equinox and will be directly overhead on dates approaching
the winter solstice as the latitude decreases, until at the Tropic of Capricorn the sun is always in the
northern skies and directly overhead on the winter solstice.
The Arctic Circle is at 66.5° and the Antarctic Circle is at -66.5°, beyond these latitudes the sun will be
completely absent for some of the year and ever present at other times. When the sun never sets it circles
in the sky, never being directly over head.
The above discussion illustrates that PV panels sited in the tropics will need to be inclined to face south
for some of the year and north at other times. Figure 8 shows a diagram that makes choosing panel angles
relatively simple. Find the latitude of your site on the y-axis on the right hand side, then follow the line
that corresponds most closely to your site and using the numbered days on the x-axis read the panel angle
off on the y-axis on the left hand side.
Figure 9: The equinox, summer and winter angles for panels at a latitude of 10°.
You could reset your panel angle every week however setting it four times a year will give good results.
Firstly in November set a winter angle, then in February set an equinox angle until April when a summer
angle is set and finally reset the equinox angle in August. Figure 9 shows how these angles can be
calculated, the correct angles for the beginning and end of each period should be bisected to find the
average angle for that period: for a latitude of 10° the winter angle is 28°, the summer angle is -7° and the
equinox angle is about 10°. A rule of thumb is that the equinox angle will be about equal to your latitude,
the summer angle will be about 15° less and the winter angle will be about 15° greater. You may wish to
change the panel angle every month around the equinoxes since the recommended angle is changing
rapidly from week to week at these times of year.
When the panel angles are connected together they are known as an array. The voltage of the array must
be matched to the voltage of the battery bank (if one is being used). Typically the bulk charging of a 12V
battery bank will be done at about 15V. It is clear from figures 10 and 11 that both the 85W and 160W
panels of 15V will deliver about 5A at 25°C therefore paying for 160W panels would be a waste of
money.
When panels are connected together in parallel, shown in figure 12a, they will operate at the same
voltage: if a parallel array of 85W panels are charging a twelve volt battery bank all of the panel will be
operating at the charging voltage (i.e. about 15V). The current from each panel in a parallel array are
added together so that two 85W panels in series will produce (2 x 5) 10A at 15V giving (10x15) 150W.
Note that the panels have not been down rated from the manufacturer's specifications.
Figure12: (a) two panels in parallel; (b) two panels in series; (c) four panels, two lots of two in series connected in parallel.
Panels connected in series, as depicted in figure 12b, work the other way round so that the voltage will be
the sum of the voltages across both panels but the same current will flow through both panels. Therefore,
two 85W panels in series can comfortably operate at 30V charging a 24V battery bank with 5A, once
again the total power is (5x30) 150W. However, if these two panels in series were connected to a 12V
battery bank they would operate at 15V and still produce about 5A, thus the power produced would only
be (15x5) 75W.
Panels can be combined in series and parallel to get the desired current at the battery bank voltage (figure
12c).
If batteries are not being used and the panels are connected to a electricity supply grid through an inverter
the panels are usually connected in a long series string. This has the advantage of keeping the current
small and thus losses can be kept to a minimum and thinner, cheaper wires can be used. This is not
possible with a battery system because the voltage across each panel is summed so that the operating of
voltage of a 20 panel series array may be about 300V.
A moderately sized 12V system will require about ten 85W panels in parallel, producing about 50A. This
is quite a large current therefore quite thick wires are needed to connect the panels together and to convert
the panels to a battery charger. Wires with a large diameter cause a smaller voltage drop and will not
burnout when substantial currents are fed through them.
Another consideration when wiring PV panels is that at night or when in deep shade the cells tend to draw
current from the batteries rather than sending current to them, this effect obviously causes the batteries to
lose charge. Most charge modern controllers contain diodes to prevent the flowing of a reverse current
Solar Electricity (2nd Edition), Edited by Tomas Makvart, John Wiley & Sons (2000).
Data for monthly mean solar irradiation on a horizontal plane. The database contains listings from
thousands of weather stations around the globe. http://energy.caeds.eng.uml..edu/fpdb/irrdata.asp