The pilot project.

My first project will cover a full panel of 72 cells and a few small panels. The small panel I will use for things like battery charge (at least that's the intent). They are purely for experimentation and to get started.

What is a solar panel, a solar panel is actually relatively simple if you look at the necessary components. I found the idea of Chris van der Zwaal very interesting and used his concept. See chapter solar panel for extensive solar panel information.

First, a simple "construction description: Take the first sheet of glass. Place cells (soldered into strings) up side down on it. Place them exactly where they are to be glued to the glass. NOTE: dont forget about the plus-minus direction of the link. Put a dot of sealant on each cell (not too much). Check again whether the cells are in the right position. Slowly and carefully place the second sheet of glass on it. Make sure space between glass does not exceed what is ultimately required. My goal is 3 mm space between the glass. Allow it to dry at least 1 day (1 is enough for this). The following day (or later), turn it around and take the non glued glass away. Now the strings are to be connected with each other with the broad buswires. Connect electrical wire (more on this later). Where desired tile crosses (or something else as spacers) and glue them with superglue. Now comes the gluing of the glass plates together, so place the whole in a place where it can quietly dry for a couple of day. Put the sealan on the glass. Place the second glass on top of it. Check the distance between the two plates and let it all dry for at least 24 hours, but do yourself a favor, really wait for at least a 4 to 5 days before you continue (I have not done this with the two small test panels and they are certainly not waterproof). Apply now aluminum, and it's done. I can not write much about the alu because I have not yet done that :-)

To start I had the parts ordered. I did this and there are details on the costs page.

The first panel is 3 * 4 cells. This ratio was chosen because the full panel has glass of size 165 * 73 cm. That means there will be 18 * 4 = 72 cells, so 4 cells for the width. To test whether size of the glass is ok I will use3 * 4 so that all sizes can be practically tested. I have 110 cells, but there are probably a few that will break (currently there are 4 broken while only 12 are soldered). So a total panel of 72 cells and two panels of 12 cells which hopefully be possible with the number of cells available. 

Overzicht 3x4 cellen op glas

Above a photograph of the glass plates with the cells that will indicate the arrangement. In the middle is room because I want to put some sealant which hopefully makes the construction stronger. Not really relevant to this small panel but I think it might be for the large panel. The size of the glass I have not just chosen. The glass that I have in stock is 165 * 73 cm. Basically I want to use this glass without cutting it in case of a 72 cell panel. In case of the large panel there are 18 * 4 cells. 18 cells makes a total of 151 cm in length (don't forget about space between cells, I used 3 mm). That leaves a space of 14 cm in total. For the first test panel I have 3 * 84 mm = 25.2 cm + 14 cm taken. I cut  the glass to 40 cm length to see how that turns out and whether it is acceptable for the large panel.

Note: This was the first time I've cut glass and I got the manual from the internet. At first it didn't work at all (I hit the glass many times, but it didnt break, till I hit so hard that it just broke, but not on the cutting line). But a second website half gave decisive information. I had to tap on glass from under the glass where I had been making the incision. Then it went very easy. The intention is: Cutting (make an incision) along a bar or something like that with a glass cutter in a smooth motion in one time on the glass. Then tap with the glass cutter on the glass at the bottom exactly where the cut is. Hold the glass firmly while you hit it. BUT PLEASE NOTE: put safety glasses on. At the moment you hit the glass from the bottom it is likely that you loot at it from above the glass. During the tapping very small pieces of glass will come of the glass and that you do not want in your eye. Only one small detail: the glass that I bought was not really sharp, but the edges that I had cut were really razor sharp. I encountered that when I made the glass clean and in one beautiful flowing motion moved my finger along the glass edge. I will not place pictures.

First the soldering of the cells. In the cells section you can read more about the soldering. The cells are interconnected by tabwire with a length of 15 cm. Only at the ends of the string of cells use different lengths. I have 8 cm to the bottom and 9 cm for the top. You can also use longer (15) and later on cut it away.

Vlak voor kitten cellen op het glas.

Above is a photograph of the glass with 4 strings each of 3 cells. At the end of each string crosses about one cm from tabwire. The cells lie face down and are exactly where they should be. Now, I do drops of sealant on them and then the other glass is put on top of it. This I will leave 1 day to dry. Below are the cells with sealant and glass on top of it:

Cellen gekit op achterkant

Now the buswires have to be soldered on:

tabwire strings interconnect

Now the connection to the electric wire. But before showing I will first give a motivation for my choices. Before I started this pilot project I have long thought about a few things. And one of the problems I foresaw was in the connection with the electric wire. I came to the next trial. Many examples showed a box behind the panel where a buswire went in and inside the box the buswire is connected to the electric wire. I was not happy with some things. The "box" that you put at the back of the panel costs money. It has potential for additional problems (the more components the more pieces that can get broken). Furthermore I do not know how sustainable the design is and finally I foresee a problem with placing the aluminum profile. So I will try the following. Take an electric wire which is weather resistent etc. Strip it so the internal wires are exposed (these wires may be up to 3mm). The strip of the wire should be on the edge of the glass first. Then connect the electric wire to the tabwire. I think in this case a picture tells more ;-)

Connect tabwire with elektrawire
Here are the two glass plates that were on each with in between the electrical wires which are connected to the tabwires. I have a few problems / questions concerning this concept: The key is, can I do this so that the sealant is completely waterproof. Furthermore it can be seen that the "shell" of the electrical wire is against the glass. It is easy to imagine that if you pull the wire a bit then it will come loose. Because I want to put an aluminum profile (which has a  hole for the electrical wire) around it I would like if if the electric wire would be clamped into the hole in the aluminum. But if this is not far in it is likely that I can pull it out. Maybe not really bad, but not desirable. I think I found a solution, but more on that later. The inner wires of the electric wire I wanted to glue to glass with superglue, but that did not work. The material can not be bonded with superglue. That makes me wary when it comes to connection with sealant. So on this pane I have made two connections to experiment with this type of connection. On the second connection I have put the wires more across the plate to see which method can handle more traction after the sealant has dried:

Connection tabwire elektrawire 2 
The white rash that can be seen is from superglue. I couldn't get rid of this, but didn'r really try. The connections are made, and now lets put the sealant on and see how it comes out. But first I glue a few crosses on the glass (this time the superglue did work). These crosses are 2.5 mm high, but it will show that that is no problem. The panel is now ready to be sealed. Just before I put the sealant on I opened the bag of silica gel. Inside the aluminum foil was a bag as often found in packs. I have opened this bag and the beads I have put on the glass between the buswire and the cells so that they are not just rolling away. After applying the sealant it looks like this:

Paneel 1 na kitten

The first problem I had was that the opening of the sealant was very small. So it became a very thin "wire" of sealant and it took very long to apply. Eventually it showed that the sealant was inadequate in some places so I used a a straw to fill it up additionally.

With all the problems I faced, I decide to start with the second test panel right away. At first I thought I continue with a large panel but there are too many things that didn't go so well. So I decided to go for another test panel of 12 cells.

The soldering of the cells is quite a difficult and prolonged work and good tools are important. Some problems, such as the slipping of the cells and the difficulty of creating a nice straight string of cells has me decide to create a toolbar. The toolbar is of wood with tile crosses glued on it. I described the construction of the toolbar with pictures at the bottom of this page.

After letting the sealant dry for a day I can tell the following. Chris had already warned me, but, cocky: Let the sealant dry, leave the project alone and let it dry for a few days. After one day I pulled the electric wire and this was more or less coming loose. I first thought that this was because the electrical wire does not glue to the sealant, but after a few days drying it got stuck really well. I must say that the construction of the curved wire appeared to be stronger, but I went for a different method for the second panel.

The second panel I made is again twelve cells and the glass has the same size. Only now just one electric wire and another connection. This time I tried a number of changes. First I have the buswire go down further to bottom before soldering the wire to it. The idea is to have buswire in the sealant hopefully ensuring complete watertightness. The reason for this change is that the electric wire is against the glass en the sealand seems to not get around it completely. In this way I am afraid that things go wrong. I have made a notch in the glass with the glass cutter so that the wire remains better in its place when the aluminum is added. Finally I have increased the opening of the tube of the sealant. A picture of the connection of the electric wire:

Paneel 2 wire connection

Then the sealant. This time I wasn't really sparing:

paneel 2 gekit

Results. Maybe it was noticed from upper picture but probably not. The strings were not in the right way, so I have a vd-wire (black wire) used as a liaison between the group of cells at the left and cells at the right. Another good learning experience. Secondly, this means the sealant dries very, very, very slowly. Especially around the electric wire connection:

Paneel 2 wire connectie kit

After a week the sealant was still not dry, so the next time I will use a little less sealant. However the connection looks quite good and I like it. I am quite sure that this can work and I will try to make a 72 cells panel. But first wait for the aluminum. The problem with this second panel was that the sealant pushed the glass up. That is normally not really a problem because the little crosses create a distance which is not enough but when the glass is pressed and then comes back up the sealant is not having a good connection anymore. The sealant gets spread out. To prevent this, I will cut small pieces of the U profile and use them as clips. Lets not forget to put something underneath the lower glass plate so there will be some space available for these clips.

The aluminum is 15x15x15x2.0 That means that there is 11 mm space between the legs of the "U". Since the glass 2 x 4.0 mm and then there is about 3.0 mm. Note: the glass is not evenly thick. It appears that it is somewhere between 3.5 and 4.2 mm.. I want to use the aluminum because it will look more beautiful, and the panel becomes better manageable because glass is very sharp.

The aluminum, however creates a problem. The idea is that the two glass plates become watertight / airtight to each other with adhesives. The sealant is not all placed on the edge. But if an aluminum frame is glue to it and water comes in there and it starts to freeze then the glass will creack. The crack is because the glass is trapped between the aluminum. So due to the aluminum water can stay between the plates and this can freeze which then has nowhere to go thanks to the aluminum. The sealant  has an elongation of 25% so that will not create a problem. To fix this, the edge has to be completely sealed. This seems to be possible just fine. If I seal (standard nozzle) than the sealant goes in for about 3 a 5 mm between the glass and it seems to create a perfect sealing. So this is what we are going to try.

When testing the aluminum, I broke the second test panel. This panel was in some places much thicker than 11 mm. I had the aluminum shoven over the glass and then subtracted, but at the time the aluminum is almost off the glass it creates so much pressure on such a small piece of glass that it starts to pulverize and sometimes it breaks. I still want to use this test panel, so I glued a piece of glass on top of  the crack with sealant. Furthermore I won't put aluminum around it so I'm just going to finish of the sides of the glass with sealant and nothing else and then see if it survives outside. 

Paneel 2 stuk en gekit.

The panel with 72 cells is in the making. I got a tip about the sealant to get it more decent on the plate and that really worked out great. On the glass I put a weatherstrip of pu foam.Along this strip, the sealant gets fitted. Then the glass is placed on top. The weatherstrip forces the sealant to one side.

Paneel 3 72 cells begin

Above three photos of the full 72 cells in panel construction. First, the 4 * 18 cells soldered to each other and put on the back of the glass to be glued. On the second photo there are many details about the changes I've made in this panel. First of all the band that is put on the glass. It is (difficult) to see but below the glass plate there are some bars of aluminum. This is to lift the glass for two mm of the wooden plate, so that later ont the clips can be placed. The clips which are also on this picture are remainders of the cutting of aluminum. The third picture is the complete panel ready to be sealed. The electrical wire is soldered and is already put through a pre-cut aluminum profile.

Details kitten paneel 3

Above a photo with details about the construction, the band still in there against the sealant is clearly visible. The connection of the electrical wire to the buswire is the same as for test panel 2 but this time the buswire a bit longer so it will stick into the outer edge sealant. That is only for increasing the strength. Pulling the electrical wire has no effect on the buswire, the buswire will not move. Clearly visible are the pieces of aluminum that are used as clips. In photo two the flat aluminium bar is clearly  visible (between cell and tour band) which is placed to ensure that the clips can be easily placed. Oops, I should have had the glass cleaned a bit better. Hopefully it does not affect attachment of the sealant. Just before sealing it, I had the glass cleaned with amonia on the places where the sealant is coming.

After leaving the panel several days to dry, (I managed to leave it alone for more then one day), I placed the aluminum. First, a photo of the window (to be honest I am pretty proud of it :-):

72 cells paneel without alu

So only the aluminium to put around and it is ready. I started with the piece that is more or less the alread connected to the panel because I have the electric wire going through the hole in the aluminum (I did not want to slide the aluminum over the entire length of the wire). In the above picture it is this piece of aluminum connected to the panel with clamps. For the alu I continued to use the same sealant as for the glass to bond. The sealant goes in as deeply as possible into the opening in the glass to create a second sealing. I don't want to leave much space to remain between the aluminum and the glass because water could freeze and that may break the glass:
 72 cells paneel met kit voor alu

I then pressed the aluminum around/on it, here the result:

72 cells eerste alu gekit.

Then I turned the entire panel and left it a day to dry. After a day the alu was really stuck. I do not think the sealant was completely dried out (still not), but in any case it was really fixed. Now I could measure the size and cut the next pieces. I had two parts cut and glued on it and left it another day to dry. Then the last piece to the top. But before I did that I took the panel outside to try it out. I could not wait ...

Paneel 3 klaar.

And then .............. then nothing. I purchased 3 second hand converters (OK4E) on market place (  for a reasonable price, guaranteed to work. So the question naturally began with how to connect. Normally when chraging a battery you connect plus to plus and minus to minus and generally is black is minus and red is plus, duhhhhh. So I connect the panel like this plugged in a meter and put it in the wall socket. The meter worked, but I did not know whether it worked in both directions, so I connected two plugs together and use a multiple socket for the female-female connection. That would have to work. But it didn't. I tried all three converters but nothing. Plus and minus reversed, nothing. So then I started to measure. The panel delivered 40V open, so that was fine. Connected to the OK4E the voltage dropped  a little, but still 36V. When plus and minus were then there was only 1 Volt left, so that was probably not the intention, hopefully the OK4E is still in one piece. Switched to other OK4E and connected it normally and 36V, so that is good. Then the OK4E could be broken, but why the (what seemed to me) right voltage. It seems to track and would therefore have to work. Then just checked the 220 (Europe has 220, not 110) connection. On the back of the OK4E is a connection that I do not know but that seems to be waterproof and could possibly be a bad connection. So first of all I measured the 220 cable (you never know) and yes. There is a switch on the plug and it appears to be broken. I have never been so quick in replacing a plug and yes:
 Paneel 3, eindelijk productie.

What a fantastic feeling, even when I write this down it gives a good feeling. It worked, and suddenly I (well actually the solar panel) was producing power. Immediately after that I went to the the meter in the house to see if is going backwards. Then I learned that there seem to be many devices in our house which use quite a lot in standby, so the meter went merrily forward, but slightly less fast ;-)

Eventually I built a standard of garden wood with an angle of 30 degrees and placed that on my shed and now just wait and see how it goes:

Paneel 3 op het dak.

This I'll leave to work for the coming time and see how it goes.

The panel works well, but I had expected better. This expectation was probably not good. Here is some information that I have measured. A number of peaks that I have measured: 100, 99, 97, 90 and 89 watts, over 90 only with the gridfit. The OK4E does not get higher than 92 watts even if you think of the name (OK4E-100) that it is to give 100W max. These peaks are not bad I think. The panel produces better when it is cold. The peak is reached when the panel is cooled in the shade and suddenly the sun comes from behind the cloudes. Depending on the outside temperature the constant power supply is somewhere between 70 and 80W. If it is cold it gets to almost 80W continuous, but the average is low in the 70W continuous. The temperature has a significant influence and the design of the panel will have an influence on that. In three weeks' time, the panel produces 7.04 kWh. I can not estimate whether or not this is much, but it doesn't seem so much, but there were quite a lot of bad days, so perhaps it is not so bad after all. The OK4E seemed to work as well as the gridfit and I have a feeling that at low power it has a slightly better performance. And last but not least: This project went very well and I will certainly continue ......

Almost forgot, I determined the the price of one kWh. I switched to another provider via a website (some electricity companies give money back (in my case, 390E), they didn't know that they would subsidize solar panels :-)). On that site you can fill in exactly how much you use and then it calculates what it would cost. I used two realistic values, and come to 20 Eurocents per kWh, so I've already earned 1.40E, another 3 weeks and the sealant is paid for :-)

After this first project I decided to do a second pilot project. This is because I have a few things that I'm not quite sure of, so therefor pilot project II.

The toolbar.

Made of:


Saw twice length of 155 cm. Saw one of the remaining pieces in 4 * 17 cm. Keep it! Screw the outer two beams of length 155 to the 4 short pieces. Outer surface to surface. Place remaining piece of 115 in the middle and screw it together (should be screwed on three places if all goes well). Shorten the last remaining piece in order to cooperate with the 115 cm long piece to create a total 155. If everything went well there should be one piece left of about 7cm. Use it as part of interconnect between 40 cm and 115cm in the middle. Now glue the crosses on. Just start at the most outer. Search for largest solar cell that you have and start. I gave it an additional half millimeter space to top and left side.

I spent a whole whole evening on this tool, but soldering is now much easier and faster. Especially easier since the cell is now free at the place where it is to be soldered so the bolt is never put higher than 350 degrees celcius.