Electronic – How to go about powering appliances from solar energy

SolArduino v1

Solar Piston v1

Product Specification =
Peak Power (Pmax):100W;
Peak Circuit Voltage(Voc):21.61V;
Max Power Voltage(Vmp):17.43V;
Short Circuit Current(Isc):6.13A;
Max Power Current(Imp):5.73A;

Most useful values are Vmp, Imp, Wmp and Isc.
Wmp = wattage out in full sun at optimum load.
Vmp = loaded voltage in full sun at optimum load.
Imp = loaded current in full sun at optimum load.
Isc = current into short circuit in full sun at optimum load.

Isc is max current you will ever get.
In a simple system Vload is < Vmp or << Vmp and Iout approximates Isc x Zsun
Where Zsun is fraction of max solar output.

To properly determine energy delivered across a day you need to calculate Vload x Iload at regular intervals and sum these to get daily result. This is "not too hard" to do with eg an Arduino or any other microcontroller with 2 x ADC inputs.
Vload can be measured with a simple resistive divider.
Iload can be measrured by plaving a resistor in series with the panel feed and measuring the voltage drop across it. (There are a number of other ways but that is easiest. If you use a say 10 bit ADC with 1024 step resolution and max input of 5V then you get ~= 200 readings per Volt input. If you allow I_series_R to reach 0.5V max at Imax then you get about 1% steps. Current measurement accuracy will probably be in the +/0- 2.5% to 5% range (using a 1% series resistor).

You can allow higher voltage drop for more accuracy or use an opamp or instrumentation amplifier to increase the voltage into the ADC. A lower ADC Vref will allow greater accuracy and resolution

An "18V" panel such as you have is often used to charge a 12V lead acid battery. If a linear charging circuit is used the excess vopltage above battery voltage is wasted. In such a case, adding a 0.5V or even 1V drop in a sense resistor does not affect available charging energy.
If eg a buck converter is used to reduce panel voltage to battery voltage themn any extra drop represents a reduction in available energy.

Any of several simpler methods may be good enough:

You can measure and sum only Iload at regular intervals. This can be done with discrete ICs if desired. If charging a battery Vload may be approximately constant (depending on battery chemistry) and Iload will then approximate charge rate. In many cases the total mAh of charge is of more interest than the charge energy.

A VCO (Voltage Controlled Oscillator) can be used to produce a frequency output proportional to load current and this can be summed by an IC counter or by a microcontroller.

Iload can be used to charge a capacitor to a preset voltage at which point the capacitor is reset to zero and a count of the rests indicates mAh of charge. This is essentially a very simple Vco. If eg a comparator is used the rest threshold voltage can be low and available voltage is not significantly reduced.

Panel mean output if optimally loaded will be
Energy ~= (average equivalent sunshine hours per day) x (Panel max power out)
= Wmp x SSH ("sunshine hours")
Solar insolation is expressed in kWh/m^2 with "1 sun" = 1 kWh/m^2.
It happens that the typical midday solar output is approximately 1 sun.
This varies with site and may be higher or lower and depends on geographical factors, time of year, weather conditions and air purity.

An excellent site which gives mean monthly SSH plus many other related parameters is GAISMA.COM. Here is the page for Kabul in Afghanistan. A good way to access pages is to web search for gaisma city_name. This is usually faster than using the sites own search system.

This chart from the above page for Kabul hives "sunshine hours" or kWh/m^2/day for Kabul.
eg in June (month VI) there are an average of 7.4 equivalent full sun hours per day. This is an exceptionally high figure not achieved at most locations. In December (month XII) there are 2.17 SSH/day on average.

These SSH figures may be multiplied by panel Wmp rating to get the average maxi,mum output that can be expected. This needs to be reduced to allow for non-steering - both per day and across a year.

The sunpath chart on the same page shows sun angle across a day and across a year and sunrise and sunset times and is useful for predicting likely performance.

AVAILABLE ENERGY:

I plan to connect the panel to a 12V car battery and then keep an LED light turned on for about 14-16 hours per day.

Due to factors such as

• the voltage mismatch when charging and

• battery charge/discharge inefficiencies and

• panel pointing losses,

a rule of thumb is that you can expect the battery to output ABOUT one third to one half of the max possible panel energy per day.

• Available energy ~~~= SSH x Wmp / 3

eg if you have a 100 W panel and 2.7 SSH per day then you van notionally get 100 W x 2.7h = 270 Wh/day from the panel but after charging the battery and then using the battery at night the likely energy is 270/3 to 270/2 = 90 to 135 Wh/day.

• Load Watts ~~~= SSH x Wmp / 3 / Operating_hours

If you wish to operate a load for 16 hours then the available Wattage is ~~= 90/16 to 135/16 = 5.6 to 8.4 Watts of load.
If operating LEDs strips or similar that accept 12V directly then a 5 to 8 Watt load can be supported with 2.7 SSH of insolation. If the LEDS require constant current at about 3 to 4V then use of a buck converter or equivalent is required if efficient use of the battery energy is required.