Main project.

The main project started. The aim is to make a total of 16 panels and place them on the roof of the house. 16 because no more panels fit on the roof. First a little explanation about the "vague" of the panels. In the pilot project II chapter I have written that the first panel has become a little "blurry". Actually, all panels have this problem, but especially in the first panel was very visible. At first I thought it was mainly caused by improper cleaning the glass. The glass is second hand, but that is not the sole cause of the problem. The other two panels have a haze on the inside but it just is not so clearly visible from a distance. In the last two panels this blur was mainly concentrated around the cell wire so I suspect the cause to be the following: I put solder on the cell and then the cell wire is soldered to the cell. The soldertin is a tin with "no-clean" flux inside. What I hoped to achieve was to have the tin easily attach to the cell and nothing further needs to be cleaned. That goal is achieved very successful, however the cleaning is as it looks like still necessary. "No clean" tin does not mean that nothing is left behind after soldering, but it means that what remains is not aggressive for the parts. Often the flux is resin or something similar. The tin I use splashes quite a bit causing quite a bit of flux to remain on the cell. What I suspect (I do not know with 100% certainty) is that this resin is vaporized by the heat (one panel can be quite warm in the sun) and then attaches on to the glass. Since the first panel is made of second hand glass I cleaned it as best as possible. The bloom now visible is the zigzag pattern of wiping the glass. Because the latest two panels were made using new glass I suspect that therefor it is much less visible.

The panels output have as far as I can see not been affected by the haze. If you look at the panel from 90 degree angle then you hardle see it and when the sun shines nicely then it seems to disappear almost completely. However I don't like the looks of it and want to fix it. After some searching on the internet I discovered that the remains of resin during soldering is normal and that you always need to remove it unless it is no-clean resin. But in my case the resin caused other problems so I must still remove it, even though it is no-clean. The solution I chose is to cleanthe cells with IPA after the soldering. IPA is short for isopropyl alcohol or isopropanol. I'm not familiar in chemistry, but I believe it is an almost 100% (99.6 or something) alcohol which dissolves fat very good and evaporates afterwards. I read that IPA is also used during the production of solar cells so I hope it will make no further damage to the cell. I bought a bottle and experimented a bit with old cells and I see balls (of which I think it is resin) which are well solved with IPA.

First one panel but now with cells that have been cleaned. If that panel works as well as the already constructed panels then I will build more. The cleaning of the cells leaves a streak behind on the cells. I used cotton swabs with IPA to wipe away / resolve the resin. First I made a 6 cell panel. I put one of the first test panels on the shed and it is use to charge batteries. But lately, the batteries barely get charged. This is because the voltage is not high enough anymore. Therefore I have made the 6 cells panel which will be put in series to ensure that the batteries get charged in the autumn and winter as well. Moreover, production went down drastically. The panels are only a few hours per day in the sun and that is a bit unfortunate. Here is a screenshot of the week production measured with Plugwise:


The reason for this drop in production is ofcourse the lower strength of the sun, but mainly because the sun is less then one hour shining on the panels:


Lately I've been asked a few times how my panels are doing. If they are still waterproof. I also read on some Internet forums that some people have problems with moisture in their panels. This concerns usually panels made with a wooden back and I read the instruction on how to make such panels on diy websites, but never considered that as an option for me. I trust that my panels are waterproof and will last several years. The principle is that of double glazing. My panels are created in a similar way. Only for double glaze an aluminum spacer is used and other types of sealant. But that is primarily for production reasons. In any case, I can assure you that until now the panels are waterproof. However, there is water between the two sealings in the first panel. I have previously described that I expected that and that I was a little afraid of this to happen. Especially because this water will expand when freezing:

Vocht tussen kit randen

Yet it seems to work out fine sofar and everything is still waterproof. The small test panel is also still waterproof. And all this despite being quite damp lately. Of course, no guarantees if you want to try it yourself :-) I am convinced that my chosen method works but if I ever get moisture in it than I will definitely mention that.

This is what cutting celwrie for an evening gives:

Celwires cut

This is enough for the first 5 panels. The first panel is finished. In this panel are cells that have been cleaned with IPA. Cleaning I have done with a cotton swab where I used a new one every couple of cells. While cleaning, I found that thers is much more resin on the cell. I thought previously that bisters were produced by heat from soldering, but it appears to be resin. The first panel is finally tested and it certainly produces just as good as the existing panels. So we continue on this path and all the cells will be cleaned with IPA after soldering.

After a discussion on an internet forum I am in doubt about the used PU foam. As the photo above (water between sealants) shows the foam is very brown. That in itself is not bad, but PU foam breaks down under the influence of UV light. That means it will eventually fall apart. This may take a while before that happens and maybe it is still not very bad. I just hope that it drops to the bottom panel , but it is not desired. Therefore I started searching again and I found PVC weatherstrip. PVC handles UV light extremely well and is really well suited for what I want to use it for. Therefore I ordered it and I think it can work fine for me. The foam is slightly firmer but can be pushed far enough, it is only a little yellowish, but then again the PU foam is brown, so this is better.

The "mass" production is finally underway. I have been a little busy with other things but I made some time to build some panels. 5 Panels are now almost ready, except for the aluminum. Multiple panels at once to make it go much faster. Yet it still takes a lot of time.Winter 2010, snow on the panels.

It has been pretty cold and we have had several frost periods with snow. The panels are not affected by this winter. I once inspected the panels and everything still looked fine.

The "mass" production is finally underway. I have been a little busy with other things but I made some time to build some panels. 5 Panels are now almost ready, except for the aluminum. Multiple panels at once to make it go much faster. Yet it still takes a lot of time.

Mass production first 5 panels

The new PVC weatherstrip works perfectly. It is slightly thinner then the other one 9x4 mm, making the sealant go over it during the assembly, but it all still looks good and the color difference is minimal, and when looked at it from a distance you can hardly see anything of it.

New PCV foam with sealant

For now I'm still working on assembling the panels. The intention is that they are ready to be placed on the roof this spring. For the roof constructm I think I will use ClickFit. It is the cheapest I've found and is used quite often in the market so I trust it.

Attention !

Above, I described that I had made an additional panel to recharge the batteries. This involved a total of 4 NiMh batteries which used to be charged by a panel of 12cells through a diode. 12 cells provide 6.0V and the diode will make 5.7V of that. By directly connecting the panel to the batteries you have a fine (if there is sun) charger. For the autumn / winter  I created and connected an extra panel. This was an 8 cell panel and in the winter it worked fine. But the sun has started to gain strength and is shining again and now there was a problem. There are 20 cells in series (via two panels) and this results in a supply in bright sunlight of 10V 3A. If you connect this directly to a pair of batteries then this happens:

Melted batteries.

Do not do this !. I'm working on a solar charger on a small scale. See the miscellaneous section. I intend to buy all kinds of electronics including some material to recharge batteries in the barn which can supply LEDS. That must still be worked out, but in short, it means that I take a LDO which regulates the power from the panels to a fixed voltage, then a resistance that will set the current and a diode that ensures that the batteries won't be drained at night by the panels. Meanwhile, I have the 8 cell panel removed and the batteries charge just fine and are not heating up anymore and the light outside goes on at night again.


Because I had not enough cells to create all 15 panels I ordered on ebay cells. I thought I had found a bargain in the form of a box of cells with cell wire (cut) for $ 500 and this is what I got:

Box broken cells

At first glance it didn't look so good, but the left part of the box seems ok. When I sorted it out. this was the result:

Sorted box broken cells

It seems that half of the cells in this box are still good. What is much less than I expected, but the cells which appear to be whole in above picture are all broken. In total a little under 200 whole cell are usable, which is not really good quality. Furthermore I have approximately 3.5 kg of waste. Some of the cells are still capable of breaking in half and for the solar charger project I can still use them, but I already had enough waste cells.

After reporting this result to the seller, he responded with promising me refund. I ended up with 2/3 of the amount (including shipping) refunded, so the damage was limited. Now just a waste storage cells to get rid of :-)

After this little disaster, I tried it again and this time it went better. A little bit more expensive, but this time with more success. The cells were perfectly packaged and not damaged during transport:

Boxes good cells

And this is how one box looks like on the inside:

Box open with good cells

The cells from this order have been tabbed (on front and back). These cells I bought from the seller eco_iq on Since the handling went so well I would like to mention where I bought them.

In the meantime I made a panel of the "bad" cells and that was not easy. One of the cells appeared to be broken, but I saw this when I put glue on the back of the cells. The result was that the cell had to be replaced when the string had been glued on the glass:

Broken cell on glass

Fortunately, after a little patiently fiddling it all turned out ok and the eighth panel is ready. The cells were of very poor quality. In total I made almost three panels with them and I had to replace cells twice. Hopefully they produce something, but I'm afraid these are not so good panels.


After more than half a year doing very little with my solar hobby I’m finally going to continue. The panels, a total of 14 units must be mounted on the roof. The past year I had to find another job and was not sure whether we would continue living in our house. I have therefore postponed the installation. In total I made ​​14 panels to be placed on the roof. I have immersed myself in the installation and decided to use a standard construction. I did that for two reasons. First, the system has to be on the roof for at least 10 years, and secondly I can’t think of anything to make myself that I thought it was safe enough and would also cheaper. I chose click-fit since this looks like a fairly easy to use system and it can be used in combination with my homemade panels. There is still a need for some own invention.


The click-fit system works as follows. I have a sloped roof with tiles. The system starts with hooks that go around the tiles. On these hooks is a horizontal bar. On this bar, the panels are "laid" and are then fastened with clamps on the bars. For each row of panels there are two bars.

My own work.

Part of click-fit system are "wheels" that you connect to the panel rotates. This is useful during installation because you can lay the panel on the bars without having the panel sliding down. A factory panel has an aluminum factory rim in which can be drilled. This edge is quite high, usually  between 30-50mm. But my panels do not have this, so I can not connect those wheels. However I do want something to support the panels when they lay on the bars. Therefore I do the following. I take a strip of aluminum 30x2 mm. That I bend in a hook that slides over the bottom of the panel, the other end of trip is a hook  which slides in the aluminum. If I made ​​one I will post a photo, because I understand that this is not so clear :-). Furthermore, I find a way to connect the first and last panel of each row. Normally you can use dedicated end pieces which can be attached to the ba, but they are only for panels of 30-50mm thick. Therefore I do the following. At each end of the row of panels I place a box profile of 30x15x2 mm thick aluminum. In this profile I drill a hole to connect it to the bar and then put a clamp between the profile and the panel. As I said, my panels are thin and so I also need shorter bolts for the clamps to secure them. So I contacted a supplier of the system. He gave me some good advise and told me that shorter bolts were no problem. I ordered with him.

First I made ​​a picture of the roof to determine how the panels can be placed and to calculate how much I had to order:

Roof showing how to place panels.

On this basis, I have determined how much to order and I did. The delivery was very fast and the items have already been received except for some bolts. The bars, I ordered on the length as indicated above, but I'm not quite sure if I let them stick out so far. It is maybe nicer to shorten them and not letting them stick out. But I'm not sure if that is helpful. Now the hardest part, placing the construction on the roof. Hopefully this will be done in the next couple of weeks. Making the picture was really helpful. Not only did I know exactly how much to order, but I also know now exactly how I’m going to install it. It makes me more confident that I can do this.


Unfortunately it has happened. The panels on my shed contain moisture. The last day I was at about 10watts lower production than expected. Of course this is partly due to the heat. The panels were covered with dust (from flowers) so I decided today to wipe them off and see if that would make much difference. Then I discovered the moisture, I first saw it only on my first panel. I had expected it, but then I saw it also with my other two panels. In all three is moisture. The panels are still functioning, but I'm afraid that it will be over soon. The humidity will be disastrous for the cells and celwire and sooner or later they will fail. This really sucks.

My biggest problem is that I still have 14 panels. I have purchased everything (inverters, Roof connection, harness) to install them, except the scaffolding. That I have not ordered. My first thought is to install them anyway. Even if they work only for two years, then they still produce something. Besides, I can replace them "real" panels which by then might be a little cheaper. The only downside is that I need to go two times on the roof, but hey, that is also manageable. Besides I do not know where it went wrong. It was very cold last winter and that may have been the cause, but then again this kind of weather can always come. So I should not assume that they last longer than two years. But even with a conservative estimate, these panels will still produce 450E of electricity in two years.