Hybrid Solar Panel (photovoltaic and Thermal)




Some time ago I bought some PV cells (6"x3", from rebeccayi0904 on ebay, very nice seller!) with intention of building a PV panel (after reading a plaethora of instructables on solar panels here!).  While testing the individual cells out there in the sun I noticed that they got quite hot.  I then realized that PV panels convert to electricity only about 12% of the solar power that gets to them.  What about the rest?  It becomes heat (about 88%).  I figured that perhaps in the same surface one can harness both the electricity and the thermal energy of the panel....

Step 1: Overall Panel Characteristics, Items, Etc.

I am not going to repeat all the details on how to build a solar panel, there are plenty of other instructables for this (search tool is your friend!).  I will give some basics, though... and then focus more on the "hybrid" nature of my panel (PV + thermal).

General characteristics:

- about 0.5 m^2 area, at a maximum of 1 kW/m^2 of irradiation and 12% efficiency this should produce UP TO 60 W of electrical power.  (at the same time this means that about 440 W of thermal power could potentially be harnessed!).


- 36 cells, 3"x6".   Cost:  about $150 from rebeccayi0904 (ebay, nice seller!) for 80 cells (used 36 for this panel).
- aluminum backplate (26"x32", can't remember thickness):  about $10 in a sheetmetal store.
- small roll of Begquist sil-pad 400, cost about $50 from ebay (can't remember seller)
- glass front cover, about $15 at the hardware store
- aluminum rails for borders, about $12 at the hardware store
- about 25 feet of 1/4" copper pipe, about $20 at the hardware store
- some 2-3 tubs of silicone caulk
- aluminum flux paste from McMaster-Carr (about $30, but you can buy a smaller quantity, I only used about 1/20 of the tub)
- solder
- a 12 V water pump, search on ebay "12 V pump laser & cpu cooling", cost about $10.

Caution: Aluminum Flux Paste is a very nasty material.  It contains fluoride and if handled improperly it can cause serious harm to you.  Read all instructions and the material safety data sheet (MSDS), and if you are not 100% confident that you can work safely, do not.

Step 2: Building the Heat Exchange System

Photo 1: Soldering 1/4" (~0.5 cm) flexible copper pipe to the Aluminum backplate. The backplate is about 0.80 m x 0.65 m. I used about 7 m of pipe. Make sure you use the proper tool to bend the pipe to avoid pinching it off! The Al. plate cost about $10. The pipe about $20.

Photo 2: I thought it was impossible to solder to aluminum. Incorrect! You need the right FLUX. I bought this from McMaster-Carr for about $30. Notes: 1) buy a smaller pot, this was way more than I needed. 2) Be careful, this stuff is nasty: it has Fl and breathing its vapors or touching is NOT recommended!

Photo 3: Here my friend Martin is soldering the pipe and plate. Note that we worked outside, there was a good breeze and we used some protection!

Photo 4: End-result: will not win a soldering contest but was not bad for our first time soldering aluminum...

Photo 5:
Close-up of the solder job...  Could be better, eh?

Step 3: Gluing the PV Cells

The aluminum needs to be covered by some electrical insulator (I used Begquist sil-pad 400, cost about $50) because the backside of the PV cells is one of the elctrodes. Sil-pad is a decent thermal conductor too, but I presume that tar-paper would have worked OK too and would be much cheaper! The PV cells are 15 cm x 7.5 cm, and this pane will use 36 cells. I bought the cells on ebay from rebeccayi0904 (nice!) at $150 for 76 cells (used ~1/2 of those for this panel). The cells need to be connected in series (how are your soldering skills?) and then I glued them to the sil-pad using silicone caulk. BE CAREFUL: the PV cells are only 0.20 mm thick and break if you look at them too hard!

Step 4: Electrical Connections

Added electrical interconnection between rows (electrical bus). Notice I cracked some cells, this will reduce the efficiency of the panel, sigh! 

Step 5: Testing Back in Winter... (electrical System)

First test! Just put the panel out and connected a 12 V halogen lamp, the panel was pumping 16.2 V, the lamp was very very bright! Open circuit voltage was just over 20 V. Short circuit current about 2 A. This was in winter, so the sun was not too bright.

Step 6: Testing Water Pump and Pipes

Photo 1: Testing the water system. Note that pump (12 V, bought on ebay for ~ $10) is powered by one half of the solar panel (generating ~ 9 V).  The plastic pipes were recycled from an old medical condition!

Photo 2: Water is circulating! This day was cloudy and 1/2 of the PV panel was only generating 4 V (should be over 9 V on a sunny day!).

Photo 3: cool guy in the "mirror" :-)

Photo 4: water circulates between the panel and a small cooler, will do some calorimetry later to check the thermal output of the system. (not enough sunlight this day...)

Step 7: Insulating the Backside

Photo 1: Insulated backside. I just glued some Styrofoam to the back using silicone caulk.

Photo 2: detail, I tried to keep it as weather tight as possible.  The electrical cables go through a hole in the aluminum frame and there is plenty of caulk to keep things in place.

Step 8: Two Days of Real Testing (May/5 and May/6)

Finally we have some sunny days!  I can do some testing.  Pity can't do it until 4:00 pm and later...  Sun is bright but not at its brightest.

Photo 1: First day of real testing (May/5). Half of the panel is powering some 12 V halogen lamps. Connecting 1 lamp this half panel was providing 9.3 V at 1.47 A (power = 14 W), connecting 2 lamps the voltage dropped to 7.9 V at 2.69 A (power = 21 W). Extrapolating to the whole panel this would be at least 42 W of electrical power (not too bad for doing this experiment at 4:30 pm, considering that the theoretical maximum for this panel is 60 W, assuming 1 kW/m^2 solar power and 12% efficiency of the PV system). The other half of the PV panel powers the water pum, which makes water circulate between the panel and the cooler. There was 3 kg of water in the system.

Photo 2 and Photo 3: Results from May/5.  Given water's heat capacity [4.2 kJ/(C kg)], one can conclude that the thermal power transferred was about 200 W maximum (again not bad considering that the panel received at most 500 W of solar light!). Note that I am neglecting possible other possible heat transfers (cooler, etc). See the second day of experiments below. The max. temperature reached was 52 C (126 F). Quite hot to the touch!

Photo 4 and Photo 5: Results from May/6.  Second day of experimentation. Again I used 3 kg of water in the cooler. Note that at the 94 minute I turned off the water pump and the water started to cool off (at an energy loss rate of ~ 56 W). See the next picture. I estimated that in the "more stable" region of temperature rise the thermal power was about 210 W. This day the solar flux was similar, notice that the electrical power for 2 lamps was almost the same as the previous day. Also notice that connecting 3 lamps in parallel reduces the net electrical power: this is an important factor when designing a complete PV system, you need to optimize the IV operational point!  In the plot, the very high rate of temperature rise in the first few minutes is probably an artifact of having used cold water rather (BTW: the air temperature was about 24 C). At the 94 minute I turned off the water pump and the water started to cool off (at an energy loss rate of ~ 56 W). Notice that as happened before the maximum temperature was about 52 C (126 F).

Step 9: Conclusions

Overall I am quite pleased with the results, but I did spend a lot more $$ than I had anticipated!!!  Whether this can be a cost-effective solution for people's homes remains to be seen.  I like the idea of using as much solar power as possible, and it makes for a most complete utilization of the surface area (e.g., a home's roof).  One extra advantage:  the water keeps the temperature of the PV cells lower, increasing their efficiency (or so I hope).

If you find this instructable interesting please comment.  If you have suggestions please comment.  What other types of heat exchangers could be used?  I still have enough PV cells to make a second panel, but I would like a cheaper and simpler way to harness the thermal energy.  Any suggestions are welcome.

Please enjoy and if you decide to make a panel like this let me know. THANK YOU.

Carlos Wexler

Note (added May/13/2010):  a google search reveals that there is apparently some commercial systems that use the same concepts (using both the electricity and the heat in a single package).  Please see, for example:  http://solarwall.com/en/products/solarwall-pvt.php, where they use air convection on the backside of the panels (instead of water).



    • Warm and Fuzzy Contest

      Warm and Fuzzy Contest
    • Build a Tool Contest

      Build a Tool Contest
    • Remix Contest

      Remix Contest

    137 Discussions


    2 years ago

    Can you tell me what size sil-pad did you use. I am unable to find a sil-pad of this dimension. Please REPLY..!!

    4 replies

    Reply 3 years ago on Introduction

    A semiconductor will not fry when exposed to reasonably high T's. You should read some more about mechanisms for heat transfer too:

    1) The insulation is not perfect. If you could make perfect insulation you'd be really rich.

    2) Then there is radiation. The cells are very close to a blackbody radiator. Even with the greenhouse effect of the glass the PV cells will radiate back IR radiation.

    3) There are shunt diodes dispersed in the circuit. This makes sure that one shady cell does not fry.

    4) Heating because of current is minimal. These are 2-3A maximum, at 0.5-0..7 V per junction this will be 1-2 W of heat. Not too significant for a cell this size.

    5) The panel has been working for 5 years without cooling...


    Reply 3 years ago on Introduction

    Again… be smart and read before you make comments that are not so smart. From 2010 to 2015 PV cells and panels have dropped in price. Surprised? Maybe I have a 286 computer that I can sell you (you know, these used to cost $1,000 or more back in the early 1990's).


    Reply 3 years ago on Introduction

    1) This was built in 2010. Read the article before making comments that are too smart for their own good. In 2010 this was a good price.

    2) It has lasted 5 seasons now and it is still generating electricity, and no… there is no cooling working because in the end this was a prototype and there was no need for the hot water (not worth on a panel this size to do all the piping for the house).

    Still charging a battery all these years (for a backyard night illumination using LEDs… I am quite sure you can buy the LEDs cheaper these days too).


    Reply 3 years ago on Introduction

    No safeguards. It is a prototype. However, I find it highly unlikely that the PV cells would "fry" (unless perhaps you are talking of the Sahara desert). What should happen is that the efficiency of the PV conversion goes down significantly with temperature.

    Russell McMahon

    6 years ago on Step 9

    Very nice.
    (1) Note that you REALLY don't want to use broken cells if only a few are nbroken as it affects the WHOLE panel output in proportion to the % area missing in the ONE cell. eg if a single cell has 10% area missing then the whole panel output will be 10% lower if all cells are otherwise the same. This is because current out is proportional to cell area while voltage is essentially nnaffected by area. So if you wire an -800- 800- 800- 600- 800- 800- mA cell string in series the 60- mA of the lowest cell "throttles the while string. In many cases if you have only one low output cell you may be better off shorting it out than leaving it in. eg if you have a 36 cell 18V nominal panel, if you short out one cell you get a 17.5V nominal panel. In practice it wioll usually be somewhat higher and in most cases when driving a 12V system with Vmax battery lead acid = 13.8V the 17.5V at full current is better than 18V at reduced current.
    (2) The cooling water will add 5% to 10% to panel output on a hot day. I have tried running a very thin film of water over a panel surface with excellent results. Needs a continuous water supply or a pump :-).


    6 years ago on Step 9

    A very good instructable. Is it possible to double the copper tubing to increase the efficiency of your heat exchanger?, Like two staggered "U" configurations both interconnected? Also what about adding (radiator) coolant to increase the heat transfer and use that in conjuction with a secondary heat exchanger to heat water or preheat water for the house?I will follow your work , you are doing research in a field and opening doors to many of us neophites. Thank you again for your article.


    7 years ago on Introduction

    Did you figure how much additional electricity you get from the PV with the 'cooled' vs 'not cooled' use?

    Cooling the PV is supposed to allow them to be more efficient.

    Thanks ... Jack

    2 replies

    Reply 6 years ago on Introduction

    That thought came to mind as well but in a different manner. Solar hot water and PV may not be compatable. The water temperature that would be good for PV may not be satisfactory for domestic hot water and vice versa.


    Reply 6 years ago on Introduction

    Respectfully aesthetics is a relative topic,relative to the individual. Personally I go for function first aesthetics last. All I can say is to look for collectors with the least amount of bulk & can be painted to match the surfaces the are mounted on. The frame and mounting hardware can be painted, not sure how to mask the actual collector while retaining the function. Was a time a TV antenna was a bit of a status symbol, but a lack of one indicated we can afford cable TV also a bit of status symbol :) Perhaps a good decorator could give advise how to mount a collector so it would blend in with other structural elements?


    8 years ago on Step 1

     It seems like mixing copper and aluminium is going to cause corrosion problems... probably on the aluminum. I would also assume that the aluminum backplane is electrically connected to one leg or the other of the cell, meaning you might be shorting all the cells together, wouldn't you?

    8 replies

    Reply 8 years ago on Step 1

    Oops, I just spotted the sil-pad... so that's taken care of, but I'm still concerned about corrosion.


    Reply 8 years ago on Introduction

    Probably would be an issue in the long run (years?), but for a prototype it is probably OK (months...).  I guess it will be important to keep the whole thing dry (at least where Cu and Al are in contact).  Hopefully my liberal use of silicone caulk will take care of it! 

    Does anyone know whether there is some kind of hollow aluminum panel of the right dimensions that could carry water inside?  Something with a profile like this:

    |        |         |          |          |         |

    Why not use stainless steel scrap sheets? Better heat transmission capability. Even though greater weight, can be thinner as its a much stronger material.


    Reply 6 years ago on Introduction

    Stainless is a horrible conductor of heat:

    Thermal conductivity is measured by the formula BTU*ft/hr*ftE2*F

    Copper: generally over 220
    Aluminum 73 to135, depending on the alloy
    SS 304 or 316 is 9.4

    Data source: Materials Selector, 1993

    Even the "worst" aluminum has 7 times the thermal conductivity of stainless.


    6 years ago on Introduction

    See size does matter. In size of panel and flow of water or current thru size of cable.