DIY solar panels using PVC pipes.

[mas985]

When it is getting primed, water will flow left to right in the feed line until the last panel. Once there is no more space, water will start to rise in the rightmost panel but soon enough pressure will also rise there. So water will start to rise in the before last panel and so on. Like a diagonal going right back to the left. Now on top, there will be more pressure on the right since all flow must go right. Which means that pressure on the top left will be less allowing water to flow out from the top left.

That ensures symmetry but not equality. The issue is that the head loss change in the headers between pipes is different on the supply header vs the return header because the flow rates are different depending on the position of the node. On the supply side header, the flow rate starts off high and ends up low while the return header is opposite. So the pressure difference between branch pipes goes from high to low to high again meaning the flow rate through the pipes goes from high to low to high. I performed a nodal analysis for an 8 pipe example to show this effect and the chart below shows the flow rate difference between the pipes. The blue is for a 1/2" header and the red is for a 2" header. The total flow rate is the same for both setups but you can see how the 2" header is much more even than the 1/2" header. If you reduce the branch tube size, it becomes even more even.

 

[AllenA]

Thanks mas for your help and explanation. I understand somewhat what you say. I wish I knew this stuff myself but it unfortunately takes quite a bit of underlying understanding to get to a certain decent point.

You mentioned the position of the node. What if each panel branches off at the same distance from the main feed line? Wouldn't that solve the problem? Say something like the drawing below which reduces incrementally down to 1/2 inch from 2? In the model I show only 4 panels but there would be 8 in total. Of course the return side would also look the same. Such a model would only work with exponents of 2 obviously. For me 8 would probably be more than enough.

 

[mas985]

The panel itself will still have the problem. It gets much worse with 24 pipes per panel. To have near equal flow, you have to use large headers or have a binary split like you have shown for every single pipe (i.e. 2 pipe panels).

Or you could just buy some of these:

Amazon.com : 2-2` x 20` Solar Pool Panel System : Swimming Pool Solar Heaters : Patio, Lawn & Garden

Amazon.com : 2-2` x 20` Solar Pool Panel System : Swimming Pool Solar Heaters : Patio, Lawn & Garden

www.amazon.com

You will probably end up spending less for more sq-ft that way.

 

[AllenA]

Ok, got it. Makes sense. Thanks a bunch mas and others. So it is not easy / worth it. And after the lising you showed, this is reasonable. I'd probably need a few of those to get decent heat in spring/fall but my back will thank me for 1 less project to build from scratch. I don't know why I didn't stumble on those before. I kept on seeing other much more expensive products which led me to try and DIY from scratch.

 

[AllenA]

Hey @mas985,
I need to thank you once more. Not only due to this thread but some older thread where you told someone else also living in AZ to try and find used panels. I can't find the thread now but you said something like "there are always people selling solar collectors there...". The moment I read that, I quickly checked craigslist and found a guy selling 8 Helicol panels 4 × 12.5 with 1.5 inch headers for $400. Some of the clamps and clips are missing but still much better than 5-6k. So a BIG thank you to you!
Now, to figure out the rest
Allen

 

[mas985]

You welcome! I think you will be much happier with the performance of those panels.

 

[AllenA]

That, I have no doubts.

 

[AllenA]

Hi Mark and other experts,
If I could pick your brain once more for 3 additional questions highlighted below.

As mentioned above, I purchased 8 used Heliocol panels that seem to be in good shape. I won't be installing too soon but rather around the end of the season. Yet, I want to start preparing from now...

I read that it is a good idea to flip used panels. Well noted.
My wife insists on the panels be installed on the back roof. I prefer the shown area (South) for optimal results. I may be able to fit all 8 there. If not, I could split the system front and back.


Using SunCalc.org, I noted the approximate ground angle of the sun to the North side during each month of the year below in this crude diagram. It is not precise by any means. Our ultimate goal is to extend our season. Pool is hot enough in summer. The roof pitch seems to be 20 degrees from ground by eyeing it with a speed square from ground (not perfect, I know).


I read this thread and noticed that it may actually be decent enough being so far South in AZ. If I understood correctly, the system will be operating at 60% effectiveness for @CaptainCannonball. One thing to note is that I am at an elevation that gets a LOT of wind.

Seeing that our goal is season extension, I believe that the South side would be the ultimate setup. The big question is: would it help extend the season at all on the North side or just heat the pool warmer during summer? Even with a cover...
Mark, I also noticed your excel spreadsheets but could not figure out how to use them. Which one to use? I don't want to take up your time so if you have it already written up, I'm glad to read whatever you throw at me.


Additionally, since the headers of what I purchased are 1.5 inch (beggars can't be choosers), and my plumbing out of the filter is 2.5" which can feed a 3" to a waterfall, I do not want to introduce any restrictions to enjoy hot water out of the waterfall (and get bonus points from my better half). So I'd like to split the panels into sections from the 2.5". I noticed the information that @pooldv gave here and read the Vortex manual which indicates how they can be split such as shown in a double row. I am somewhat aware of all other information about gravity, position of inlet vs outlet, VRV etc... But the way they show it is pretty simple and does not consider all variables such as direction, height, run length, etc... I assume that the direction of the main feed line to the splits should be considered? For example, is my assumption correct that if the main feed line were to go straight into the first section of panels, it would not be a good design? Something line this below diagram would be bad? Meaning the angle at which the main feed line were to enter the split should optimize the furthest section? Or would it not make a difference?


Furthermore, since I would go from 2.5 to 1.5 and seeing that the area of a 2.5" is 4.91 sq in and the area of a 1.5" is 1.77 sq in, would a 2 way split be "good enough" or would a 4 way split be even better?

Thanks again in advance to all who chime in,
Allen

 

[mas985]

You want to use the Heat Transfer workbook and the "Heat Txfr" worksheet. Most of what you want to enter is in column B the yellow cells. The fields are commented for more explanation.

Also, the panels only require about 40 GPM so most of the water to the waterfall can be bypassed assuming the flow to the waterfall is much more than 40 GPM.

As to the geometry, what is important is to have the pipe connections at the diagonal to the panel sets. So if the supply connects at the closest point, the return would connect at the farthest point. Were you planning to take the 2.5" pipe to the roof? Given the flow rates required for the 8 panels, 2" is more than sufficient.

 

[AllenA]

Hey Mark,
Thanks again. Ok, yeah I knew about the diagonal but wondered about other factors and so it seems not too important. Got it.
I was planning on taking the 2.5" to the roof to keep the most flow for the waterfall. I think a good flow with highest possible temp would be really nice coming down the shoulders on a cooler day. Not sure what 40 vs 60 GPMs will look/feel like but my concern is getting the most when possible. I understand that there may be a bit less heat when they flow faster but it seems that it is possible (just found this. Sorry should have checked beforehand):


This would anyway only be so when the waterfall is in use and not to heat the pool per se. I don't even know if the pump will be able to deliver this much GPMs since my roof is quite high. We'll see...

I'm trying out the excel sheet now...
Allen

 

[mas985]

In terms of the bypass, it won't affect the water temperature at the waterfall much, only the flow rate. Slowing the water in the panels raises the water temperature out of the panel but you would be mixing it back with the original water temp so net net it is about the same with or without bypass. If you want to run at a full 60 GPM, that is ok too. But keep in mind that either way the temp rise is only a couple of degrees, it isn't very warm but that's point. For maximum efficiency, you don't want a large temp rise in the panels.

 

[AllenA]

Hmmm, good point. It's higher flow / feel (massage factor) vs higher temps.
Trying to fill out the excel... Am I correct that the position of the panels on North side vs South side of the roof would be controlled with row 80 where 0 is North and 180 is South?
And given all other variables remain the same on the same day, Which calculation would you say is the determining factor of yea / Nae? A row in section "Summary (Daily Averages)"? Or maybe row 184?

 

[mas985]

Trying to fill out the excel... Am I correct that the position of the panels on North side vs South side of the roof would be controlled with row 80 where 0 is North and 180 is South?

Correct. East is 90, West -90 or 270. Same as a compass.

And given all other variables remain the same on the same day, Which calculation would you say is the determining factor of yea / Nae? A row in section "Summary (Daily Averages)"? Or maybe row 184

Depends on what you are interested in. 182-184 show the extremes in temperature (max & min) vs 185-186 are the average daily and nightly temperatures which is basically the average of the max and min. If you swim mostly in the evening, then that is what you should use. If it is anytime during the day, the daily average might be more appropriate. They are highly correlated and only differ by a couple of degrees.