This website.

This is my first website. I have choosen for a simple design because the content is what counts. The purpose of this website is to share about my experience in building solar panels as much as possible, so others who might also want to do this have some information. The website will be regularly updated as my project progresses.

As a tool for this site I've used Kompozer which is a freeware (open source) tool. It is not the most simple tool for dummies, but after two nights I had the basic website (menu and colors) and I was ready to begin filling with text and that is very simple. For the menu and the colors I used Greg Tutor's examples and generate the menu I used I used ColorPic to "borrow" colors.

Due to severe time constraints and my preference to begin with the practical side this website will be updated after some time. I have a huge pile of fotos of which some should be placed here, but that will still take some time. So a page with only pictures which were only scaled down and higher compression so that the stored data is not too much ..... PHOTOS. More PHOTOS. Pictures of the first 72cells panel. Pictures of the following 72 cell panel, diodes, silica gel plug and tool plugwise.

Charger on solar energy.

One of the projects I have been thinking about is charging devices such as phones, rechargeable batteries etc. through a system that operates on solar energy. I still have a few cells so it is a start. But now the rest. Because the cells do not always have the same voltage and current supply and at night I certainly want to be able toDe verklarende tekst is in het engels, maar beschrijft alles wel. Ik weet nog niet welk optie werkt, dus heb ik voor beide mogelijkheden de componenten besteld. Beide circuits gebruiken een LDO om van de batterij spanning naar gewenste constante spanning te komen. Daarna wordt een weerstand gebruikt om de gewenste stuurstroom voor de led te genereren. Bovenstaand schema is dus de vervanger voor het "load" symbool van een schema eerder. Het geheel zit dus acter de mosfet in geval van de buitenverlichting die altijd aan gaat in het donker. Dus in de praktijk wordt de LDO niet rechtstreeks op de batterijen aangesloten, maar krijgt zijn input via de FET.
charge a device a system with extra batteries is needed. These batteries are charged during the day and once fully charged supply enough power for a phone or something else to charge.

The number of cells should be limited otherwise the whole size of the device would become to large. I might try to break the cells, but that seems to be very difficult. Broken cells still deliver 0.5V at only half the power (current). The problem is that I should have enough voltage. Even with a DC-DC step-up converter you need a minimum input. If the cells would have to supply 5.0V then you need a minimum of 10 cells and should it also work with less light than you need even more. This is a rather large surface of cells and not really desirable.

For this project, I intend to use as many items as I can which are fairly cheap to obtain. It should not be more expensive than a charger that you can buy in the shop, which is rather big challenge. Well it's a hobby so maybe it will cost a tiny bit more. The cost of the cells I do not count.

Basic idea.

Use solar cells to charge a number of batteries. These batteries provide 5V USB power. At first I thought to step-up voltage of the solar cells to for example 8V they charge 5 batteri which then create 5V (smoothed) supply. The disadvantage is that you there are two DC-DC steps. First to create 8V (if even possible) and then from anything between 7.5V and 6V (5 batteries) and it should also switch off once the batteries have a too low voltage (batteries should not be drained completely). Then I found something on the internet what is called MintyBoost. It is a charger using two batteries as a backup for your device (phone). The advantage is that it has everything to 5V USB power to deliver on a lower input voltage. To chargen two batteries, you need a minimum of about 3.3V. We need a diode between the batteries and the cells and you have at least 3 volts needed to charge two rechargeables. 3.3V is still at least 7 cells which is still a rather large surface. Actually, I would still like to see a bit smaller. So maybe half dc-dc converter. The advantage of a DC-DC on the cell battery is that you bring down the current and voltage up. A solar cell provides up to .5V at 3.5A. Which is a rather low voltage at a relatively high current. Hardly any device such as a phone can use 3.5A. Most of the charging current is below 1A at 5V. When a cell is not in direct sunlight (shade) than the voltage drops. That makes it even more difficult. A dc-dc converter behind the solar cell may in some cases convert a variable input voltage to a fixed output voltage. You could for example chooAfter some investigation it appears that there many more DC-DC converters which provide 5V output. For example the MAX1674/1676. These are for sale on ebay and have been working from .7 V. These can also handle high amperages. The only drawback is that the package of this IC is very small. So it is very difficult to solder. The advantage of this chip is the .7 V minimum. That would make two or three cells are sufficient to generate power at all times to provide enough capacity. For example, by 4 using half-cells it will use only a small area (15 x 15 cm) with max 1.5W generation, but lower voltages still convert to a possible charge current to the a 3V output and a fixed resistor between the batteries and dc-dc converter can then create a charge current which if the batteries are full results in such a low current that the internal resistance of batteries can absorb the power. You will have no additional logic needed to switch off the charging.


The mintyboost is a project I've learned from. This mintyboost a USB device charger with only two batteries. It gave me insight into how voltage can be increased. I'm not going to mintyboost build one as i do not want one, but I wanted to mention it anyway because the project has been the springboard for me into the device that I'm building now.

Two versions are listed. The disadvantage of version 1.2 is the relatively low power. The 2.0 goes up to 400mA. For me this is still fairly low. It appears after some searching that many more DC-DC converters exist which provide 5V output. For example a the MAX1674/1676. These are for sale on ebay and have been working from .7 V and also can handle large amperages. The only drawback is that the package of this IC is very small, so it will be very difficult to solder. The advantage of this chip is the .7 V minimum. That would allow me to use two or three solar cells only which would be sufficient to use as charging capacity. For example, by using four half-cells which is small area (15 x 15 cm) a max of 1.5W can be generated, but lower voltages would still convert to a possible charge current for the battery.

A cell is 15 x 7.5 cm (3 x 6 inches) and contains two cell ribbon connectors. It is possible for a cell to break down the middle (in the length of cell ribbon) you get two parts where each part produces 0.5V at 1.8A (max). Or if you put you back in series with 1V 1.8A. The biggest problem is the breaking of the cells. Because I have the necessary waste cells, I try this.

Breaking the cells.

Never done this before, so lets take some already broken cells and just try ... It actually went surprisingly well. In any case better than expected. Somewhere on the Internet I had seen a movie of someone who broke cells but that were other (thicker) cells. Anyway I've done similar thing:

Break cell.

First I took a small glass plate and used a waterproof pen to draw half a cell. Then I put the cell on it. Then another layer of glass over it but two spacers on the end of the glass (far away from the breakpoint). The result is that the cell is jammed between the edges of the glass. Tidy up and hold them secure then use other piece of glass to break the cell:

Break cell using glass

En this is the result:

Result of breaking glass

The break is really not perfect but there are no cracks in the cell. So the cell will still function efficiently. A quick measurement showed 0.2V under lamplight. Incidentally, this werew all cells that were already broken on one side. And only the part below the glass remains intact. The part which is pressed down gets completely broke. So do not expect that you can make two smaller cells out of one.

To make the "device" look a bit nice I have a picture frame in mind. A picture frame with a thicker edge so the batteries can be inserted inside. Finally I found the following list who appears to be ideal. It costs a 7E:

Picture frame

This list is 24x18 cm (about 9.5 x 7 inches) and it would just fit in 6 half cells. After soldering the cellwire I have them placed on a cardboard plate and interconnected the cells. This was the easy part. I now have a frame containing six half-cells This list is 24x18 inches and it would just fit in 6 half cells. After the first cell of wire fitted, I have them on a cardboard plate and the cells themselves transferred. This was the easy part. I now have a frame containing six half-cells which in theory should be able to output a maximum of 3V at 1.5A.

Picture frame with cells.

Under a bright light I measured an open circuit voltage of 3V. Now something has to be made that stores the energy in batteries. In addition, there are a lot of possibilities to consider. I aim to charge devices that normally have a USB connector to charge. USB connector provides a standard 5V and 100mA at startup and a max 500mA after enumeration in a PC. However USB chargers now often provide higher amperages, up to 850ma. If you want to charge a phone, you often have a proper charging current needed otherwise it will simply not charge, so I can offer at least 500ma and actually even more. So ultimately a USB connector which provides at least 5V 500mA minimum is what is desired.


Note: None of the following schematics is built by me. Some are copied from other sites/documents, but some are my creation. But none of them were actually built by me or tried.

Charger project.

For the charger project I distinguish three parts. The first is the charger from the solar energy creating a charging current with which batteries (5) can be charged. The second is a voltage regulator which will via the USB connector provide +5 V and finally I want a battery charger for rechargeable batteries.

Solar powered battery charging.

Charging the battery is the trickiest part. In order to raise the voltage I will use a lm2623. The lm2623 is a dc-dc step-up converter. It has a "variable" input ans a fixed output voltage which is higher. From this voltage I will go through a resistor create a fixed current for charging the batteries. Because the capacity of solar cells have a maximum of about 3V at 1.5A, the amount of power available for charging is limited. So I'll assume a slightly higher output voltage and slightly lower power (compared to charging the batteries in the shed). Eventually I came at a voltage of 9.05V at 200mA. Which supplies about a maximum current of 400mA when the battery is empty and that is about the maximum the lm2623 can handle given the loss of lm2623 and the maximum load.


The schematic is more or less copied from the national documentation. I myself have a lot of difficulty to assess whether this will work and I guess I still have some component values to change, but I hope it will work. Here is my explanation for the chosen values of the different components. D1 is a schottky which I have a few because they are used in solar panels. Therefore this type. This is a SMD, but that's to solve. The LM2623 is a uSOIC8 (very small) and I'll buy an extra converter board. L1 is 4.7 uH because that is used in all sample schematics. R1 and R2 are values chosen based upon examples, but may need a bit of tuning. They affect the ripple of the output voltage. R3 and R4 provide the desired output voltage. This was put to 9.0V. Tantalum C2 would be to eliminate wrinkles. R5 provides a maximum current of about 200mA when the batteries are full. So that they would be able to handle. At the time the batteries are empty, the maximum charging current is about 400mA. And that is about the maximum the solar panel can provide after the dc-dc conversion. Vdd must be between 3 and 5 volts. Typically, the output voltage gets bootstrapped, but that's not possisble because I have chosen an output voltage of 9V. Therefore, the IC is powered by three batteries. By connecting the output of the solar panel to the EN pin I hope to shutdown the IC to reduce current consumption when the batteries are not being charged. I could have choosen an LM317 or similar for VDD to create a bootstrapped system but that requires additional components. That would have ensured that the batteries cant be drained, but turning off the IC when the solar panel is "down" will hopefully limit the current to an acceptable minimum.


The 5 NiMh batteries provide the power for the USB out. USB is always 5V, but the maximum current is different. A PC with a USB port has three situations in terms of maximum current.It begins at 100mA. This means that a device that connects to a PC initially can use a maximum of100mA. Once enumerated the device may indicate that it is a high power device and then it is allowed to use a maximum of 500ma. Ultimately, a device has a USB suspend / power down mode in which only a few milliamps maximum are allowed. Besides this definition, there is also the USB charger stadard. This is a standard which is made mandator for cell phone manufacturers by the EU and which defines to use a micro USB connector, and it also defines a number of other things in terms of power consumption. I have read the standard briefly and I am not sure I'm right, but this is what I've made of it; A USB charger had D+ and D- shortened out. This allows the device to detect that it concerns a USB charger. Then the device may choose to use 850ma max charging current, or it can see if the charger can deliver 1.8A. In the latter case, the device should first try to draw 1.8A. However, if the voltage drops below a certain level then the device should immediately switch back to 850ma. Most cell phone manufacturers that use USB as charge ports provide an adapter that can supply 850ma and this will therefore be the starting point.

The voltage of the 5 batteries must be converted to 5V. Additionally, there should be a protection for under voltage. When the voltage of the batteries go below a certain voltage then everything has to be disabled. Batteries should not be drained too far. There are ICs such as the LP2960 with a low voltage protection. They seem to be specially built for 5 cells and provide 5V. The only drawback is the limitation of 500mA. I really want to provide a higher current.

Therefore I made my own schematic. For the under voltage protection I used the LTC1440 because it has a very low power consumption. It is a voltage comparator with a hysteresis. The output goes through an inverter to the shutdown of the LM2941 which produces the 5V/1A:


Explanation for selected values:

See explanation in pdf LTC1440:
1) Vtrip is 5.40 V (5 * 1.0V)
2) = 0.2188 Ratio = 1.182V/5.40V
3) Hysteresis = 82mV * 0.2188 = 17.94
4) = 2.4m R4, R3 = 18k
5) R1 = 1.18M, R2 = R1 * ((5.40 / (1,182 +18.00 / 2)) -1) = 4,170 = 4.22M gives 5.45V (which is fine, shows hysteresis 83mv)

6) R8 = 1.0k * ((5.0/1.275) -1) = 2.922k => 2.94k

After "designing" the electronics for the barn, I decided to do things differently. The advantage of the LTC1440 is the low power consumption. The drawback is the price. Not only does the IC cost a few dollars, 1% high ohm resistors are not cheap either. Therefore a new scheme is much simpler and has a consumption of approximately 100u. I think that is acceptable. The disadvantage here is that you must do some calculation to get an appropriate value for R1 and R2:

5V USB versie 2

This is the schematic I'm going to use. The TC54 is a comparator with a fixed value. I chose the 4.3V version. Through R1 and R2, this can also be used for higher voltages. The disadvantage is that R1 and R2 are relatively low impedance and thus approximately 100u amp is drawn continuously frm the batteries.

USB powered battery charging.

The last part of the project is a battery charger, an USB powered charger. This is not really necessary and certainly not new, but something I would like to build.
And the goal for me was to create a solar powered battery charger. On the Internet, several loaders are available and there are a lot of ICs with which it is possible.
Finally I decided to build two. One based on an existing circuit by StefanV see:

I have redrawn his schematic, but it is exactly the same. Look for details on his website. He not only displays the schematic, but he explains exactly how it works.

Battery charger stefanv

This charger has a maximum charge current of 500mA and thus can be used on common USB ports (PC). Because I like to have a loader that loads faster and uses the max 1A from the solar charger USB connection, I decided to do a charger design. This charger uses the DS2712 chip. This is a special NiMH battery charger chip that is capable of DC-DC conversion boosting the power ultimately to 1.5A, using the 1A/5V from the solar charger. It might also be used for charging only one battery at a time. Incidentally, not that twice as fast. The schematic was derived from example schematics of maxim.

DS2712 charger

Outdoor light project.

At night I want to have a light shining, I made a few small solar panels and installed on the barn. These are the test panels that I built intially. These panels are charging batteries during the day and when the voltage of the panel collapses, as it is dim then the outdoor light automatically turns on. This lamp currently had a total of 15 small LEDs which currently use up to 300mA. To charge the battery I abuse the fact that the voltage of the battery rises as they get full. In addition to a simple diagram I found on internet which makes the LEDs turn on when it gets dark, but that schematic is limited (max power) and I do not understand it. I once made an extra panel in the winter, but because I did not disconnected it in the spring I melted down a few batteries. Because I am working with electronics I decided to build a better system. So here is what I want to make; A circuit so the batteries can be charged with the two panels even in the summer. This means an over-voltage and current which charges a battery without any problems. Furthermore, an under voltage protection of the batteries, so they do not get too far empty. In addition, new power LEDs, and finally a power source (battery powered) with constant current so that the LEDs have a constant light output regardless of the battery voltage.


For charging batteries in the shed I use the following method. A NiHm battery can handle up to 250mA continuous if it is full. If a NiMH is full and gets a load of 0.1C then the voltage of the battery is around 1.47V. To make sure that battery gets charged to the max a minimum charging voltage of 1.5V must be used. The following applies. The higher the voltage used, the lower the current (relative to lower voltage) when the battery is empty. Example: Suppose you load with 1.5V, then R = 0.03/0.250 = 0.12 ohms. This means that when the battery is empty than the charge current (battery is 1.2V) I = V / R = 0.3/0.12 = 2.5A. If the charging voltage is higher, for example, set to 2V, then R = 0.53/0.25 = 2.12 ohms. When battery is empty, then the current I = 0.8/2.12 = 0.38. Because the panel can produce plenty current (more than 3A when the sun shines), it is desirable to use as much power as possible. It is also true that when a low battery charging current is used then there is not enough time to charge the battery completely during daytime. Especially in winter the sun is too short for the battery to charge. It is important to make a good balance in the selected voltage and the resistor matching. 2.5A is quite high and really a bit overkill. A maximum charge current of 1.5A or 1.0A is enough. Moreover, the voltage of the battery increases very fast when being loaded. This means that after a short loading time the voltage of the battery will rise and hence the charge current will go down. In contrast, the voltage has to go over a top voltage when the battery is almost full. 1.55V gives 0.08/0.250 = 0.32 ohms. At 1.2V, this results in a charging current of 1.1A = 0.35/0.32. That seems like a good starting point.

For charging the following simple schematic will be used:

Schematic charger shed.

The solar panel is connected to an LDO. This produces a fixed voltage, the exact voltage level is determined by R2 and R3. The schematic might not be exactly like above. It is depending on chosen LDO. The output of the LDO goes through a diode to a resistor. The resistance determines the height of the current and the diode ensures that the battery wont be drained when the solar panel is not producing. The diode that I use is the same that I use in the panel. This diode has a voltage which is dependent on the current that passes through. The key is to make sure the batteries are not too heavily "charged" when they are full. So 250mA maximum at that time. The diode that I use is the MBRD1045. It has a voltage drop of 0.35V at 250mA. That is 25 at degrees. If the temperature rises, the voltage drop goes down, so does the current will also go down because of R1.

Because I want to charge 4 batteries in series I come to the following: The charging voltage is 1.55 * 4 6.2V and output voltage of the LDO must be at 6.55V. The resistance R1 is therefore 6.2V - 5.88V = 0.32/0.25 = 1.28 ohms. Note: R1 has to withstand a little over 1A for his reason it means that the resistance should be at least 2W. No 1.28 ohm resistors exist, but 1.30 do (5% max deviation) and that is satisfying. For the LDO I chose the NJM2397, simply because it was the cheapest with reasonable specs. It has a 4% deviation on the vref which is annoying but can be solved by adjusting the resistance division (using a potentiometer). The NLM2937 has no shutdown pin. The formula for the output voltage is Vo = VREF * (1 + R2/R1). Applying this to the above shema then the formula becomes: Vo = VREF * (1 + R2/R3). The suggested value for R3 is 1k. Vo = 6.55, VREF = 1.29. That gives 6.55/1.29 - 1 = R2 (in kOhms) is 4.078k. However 1.0k 1% is quite expensive. So we choose a more standard value of 1.07k and it therefore requires an R2 of 4.799. The closest value to that is 4.75k, a small deviation of about 1%, but that is acceptable. Ultimately, the output voltage under load can be measured and with a small high impedance resistor parallel on R2 or R3 the output voltage can be adjusted.

Automatic night light and under voltage protection.

The second part of the outdoor lighting project is twofold. There must be something which ensures that the batteries won't be drained too far and something to make the LED(s) go on in the dark. Batteries should never be completely discharged. That is not good for lifetime. Therefore there is a need to protect against under voltage. To do so there is a component available that is specifically meant for this and as long as the voltage is above 4.3V the output is high and if the battery voltage goes below it then the output becomes low. This component is the TC54 which you can get in different voltages. The output of this IC is not suitable for a load (1W LED) and should be used to drive a FET. Also a schematic has to be made that ensures that the LED(s) turn on when it gets dark. The simplest way is to use the output voltage of the solar panel. Because this is the first time that I make electronics it is probably not the simplest circuit, but I think it will work:

Auto switch led on schematic.

In this schematic it is already shown how the output of TC54 is processed here. The transistor T2 is used to build a hysteresis. At the time that Q1 goes "on" T3 will be opened and thus the input voltage at T1 will be pulled down. On coming out of Q1 (ie when the voltage of the solar panel rises), the opposite happens. Load represents the LEDs and the electronics for the LEDs to send.


The last phase are the LEDs. I plan to use CREE LED to go up to around 5W. For night lighting that always (automatisc) is to want this LED at 350mA (1W). At this moment I have white LEDs. These are driven by a simple resistor. The result is that they become less bright light as the batteries drain, and thus the voltage collapses. This following schematic should prevent this:

The explanatory text describes everything. I do not know which option works, so I ordered the components for both possibilities. Both use a LDO circuit to bring down the battery voltage to the desired constant voltage. Then a resistor to set the exact current for the LED. The above schematic is the replacement for the "load" symbol of the earlier schematic. So the LEDS and drivers are behind the mosfet in the case of exterior lighting which will be automatically turned on in the dark. So in practice, the LDO is not directly connected to the batteries, but gets its input from the FET.

Led driver

The explanatory text describes everything. I do not know which option works, so I ordered the components for both possibilities. Both use a LDO circuit to bring down the battery voltage to the desired constant voltage. Then a resistor to set the exact current for the LED. The above schematic is the replacement for the "load" symbol of the earlier schematic. So the LEDS and drivers are behind the mosfet in the case of exterior lighting which will be automatically turned on in the dark. So in practice, the LDO is not directly connected to the batteries, but gets its input from the FET.