A Solar Water Collector Made From CPVC 

This was a quick prototype and test of a solar flat panel collector using water as the heat transfer fluid.  The collector was constructed from CPVC plastic pipe.  Radiant floor heat transfer plates were used to absorb the solar energy and transfer it to the CPVC pipe (see pictures).  The collector is a small (5 sqft) size just to test the concept.  It is glazed with a single layer of PVC glazing from SunTuf (see note below on this).


As an alternative to using the (somewhat expensive) extruded alum fins, here are some homemade jigs to make your own from sheet alum.  Jigs.



The good news on this collector is that it is dead easy to construct.  It probably took less than an hour to put the whole thing together.   

It could not be simpler.



The CPVC pipe and fittings are cheap.  The radiant floor heat spreaders that I used were fairly expensive, but there may be cheaper ones out there, or it may be possible to fabricate them from sheet aluminum.



The performance graph and comments on the performance are provided below.



The life of the collector is somewhat doubtful.  The temperatures in the stagnated collector are at or perhaps a bit beyond the upper range of CPVC's rated capability. The fact that the collector loop need not be operated at a high pressure probably helps.    Mounting the collector on a vertical wall where it would get less radiation during the hotter summer period would also help.  Covering the collector with shade cloth in the summer would also help (this is actually recommended by some commercial flat plat collector makers who make copper and glass collectors).



Assembling the serpentine CPVC pipe run.   The alum extrusion radiant floor heat spreader.





Tapping the CPVC into spreader grooves.     The finished collector.


The performance test setup.  Tilt was set to give normal incidence around solar noon.


Performance on a sunny day.


The temperatures plotted are:

Water inlet temperature (blue)

Water outlet temperature (red)

Collector air temperature near the top (black)  -- located in shaded spot at top of collector box.

Ambient temperature (green) -- badly located.


Note that the ambient temperature sensor was badly located and are invalid.  The actual ambient temp was 62 F at 11:20 am, rising to 77 F at 2:53pm.


This plot shows the performance for the CPVC collector on a sunny day.  There were intermittent thin high clouds for part of the afternoon.


The collector was heating about 4.5 gallons of water from an uninsulated bucket.  The flow rate at which water was pumped through the collector was 0.5 gpm.


I also mounted thermocouples on the radiant heat spreader fin with the following typical results:


T fin at outer edge                    151 F           

T fin near CPVC pipe               147 F

T of outer wall of CPVC pipe   103 F

T fluid                                           100 F


There is a big temperature drop between the fin and the fluid (i.e. a high thermal resistance).  Most of this temperature drop appears to be taking place between the fin and the CPVC pipe.  Apparently there is not a good thermal bond between the CPVC and the radiant heat spreader.  This is in spite of the fact that the alum heat spreaders are a very snug fit on the CPVC and there is silicone in the groove.


I estimated the efficiency of the collector by calculating the energy out from the flow rate and the inlet and outlet temperature, and the energy in from the midday solar radiation at my latitude and tilt.   This came out to 60% -- I was surprised that it was this good.  And, 5 sqft of collector managed to heat about 4.5 gallons in an uninsulated bucket to about 125F -- not so bad -- I am sure it would have achieved a significantly higher temperature if the water storage bucket had been insulated (you can see from the performance graph that the collector continues to add heat to the bucket through the whole afternoon, the bucket is just losing more heat than it is gaining due to no insulation).


If the thermal bond between the pipe and fin were good, the fin temperature would drop down closer to the fluid temperature.  The fins would not run so hot, and would not lose so much heat out the glazing.  The efficiency and heat output would improve.  If you assume that for a good thermal bond that the temperature of the fins would drop by 40F, and that the glazing is R1, then the gain in efficiency comes out about +13% for a well bonded fin and tube.  So, it appears that you do take a pretty good hit for this construction.


Note on PVC glazing:  I would normally have use Polycarbonate glazing, but I had the PVC on hand and used it.  By the end of the test, there was an about 4 sqinch area near the top of the glazing that had been permanently deformed by the heat.  I would not recommend using PVC in this application.    Polycarbonate glazing is good to about 270F, and should work OK.

Note also that I used the "pink" foam board insulation because it was on hand.  Polyioscyanate insulation should be used because of its higher temperature capability.


Overall, I guess you have to weight the time and cost savings against the poorer performance and shorter life.  I can't rule out the possibility that the life might be a lot shorter.





Some Additional Notes On CPVC vs PEX:

It was pointed out to me by Burt M. that PEX has a greater thermal conductivity coefficient than does CPVC. He suggested that 1) most of the temperature drop may be in the CPVC wall (rather than the thermal bond between the fin and CPVC), and 2) that PEX might work better due to its higher thermal conductivity.


I am inclined to think that Burt is correct on both points -- here are some tentative calcs to support this:

I had a go at calculating the theoretical temperature drop across the
CPVC tube wall using the the thermal conductance values you provided
above.  If I did the math right (see below), then it takes a delta T
of about 26F to transfer the incoming heat from the fin (and sun)
through the wall of the tube.  So, this says that about 65% (26F/40F) of the 
thermal resistance seems to be in the CPVC wall. 

The PEX tubing I have has the same OD and ID as the CPVC, so the PEX
might be 2.1/0.9 = 2.3 times better, or about 11F to transfer the heat
through the PEX wall.

This says that PEX would do somewhat better than CPVC.

The wall heat transfer calc:

About (290 BTU/sqft-hr)(0.6 efficiency) = 174 BTU/sqft-hr gets to
each sqft of fin and is transferred into the fluid in the tubing.
Each sqft of fin (4 inch wide) is served by 3 ft of pipe.

CPVC half inch pipe:  r2 = 0.3875 inch, r1 = 0.3125 inch

Ucpvc = 0.9 BTU-in/ft2-hr-F = 0.075 BTU-ft/ft2-hr-F

The transfer through the CPVC wall is:
q = UA(Touter - Tinner)

UA = (2*Pi*L*K)/ln(r2/r1)
UA = (2)(3.14)(3ft)(0.075 BTU-ft/ft2-hr-F)/(ln(0.3875/0.3125) =    6.6 BTU/ ft2-hr-F

(Touter-Tinner) = q/UA = (174 BTU/sqft-hr)/(6.6 BTU-ft/ft2-hr-F) =   26F

Since the PEX has about twice the thermal conductivity of CPVC, and the OD and ID are the same, the temperature drop across the PEX wall would be less than half as great.

Source for thermal conductivities:

CPVC at http://www.boedeker.com/pvc_p.htm.   (and many other plastics -- a good ref)

This gives a thermal conductance of 0.9 BTU-in/ft2-hr-F.  Compare this with aluminum at about 1500, or copper at 2700!

PEX tubing thermal conductivity: 2.43 BTU in/hr-ft^2-F  per DIN52612

Apparently PEX has significantly higher conductivity than CPVC.  PEX-AL-PEX is even better.


09/11/05, April 4, 2008