Saturday, August 22, 2009
The DIY solar heater has been working flawlessly, tracking the sun when it was visible and providing free heat to our large swimming pool which reached a peak of 82 degrees F/27.7 degrees C. The kids say that the pool has never been this warm.
I've continued doing time lapse videos and posting these to youtube. Because the array moves so slowly (the sun moves 15 degrees in an hour) it is difficult to see the motion without timelapse. Here is another example:
There are currently 11 of my videos about this project on youtube. Please check them out. I am "georgeplhak" on http://www.youtube.com/.
The plans are coming along - honestly!
Tuesday, July 28, 2009
What is interesting about this particular piece is that it was mostly overcast with a few bursts of sun.
I had thought that the sensor didn't wander when there was cloud cover, but apparently it does. This is impossible to see in real time unless you have the patience I guess? When the sky is a solid gray, the sensor does not seem to wander. When there are clouds that obscure the sun for a time, the sensor may be mislead by the "silver lining" of the cloud which is brighter than the portion of the cloud that covers the sun. The good news is that the sensor recovers promptly when the cloud moves and the sun is clear once again.
Here is another attempt made this morning. The duration is about one half hour and the sensor is swinging the array from the west to catch the morning sun. Unfortunately, the sky clouded over AGAIN.
Wednesday, July 22, 2009
Hopefully where you are has great solar potential.
Tuesday, July 14, 2009
This is about right, since the heat from the sun is greatest from about 10am all the way though to about 6pm, or about 50 degrees either way from vertical. Also, the Red Rock sensor will not reliably sense light coming from behind it so there isn't much point going for a large range of rotation than about 100 degrees.
You can clearly see the backs of the closest five reflectors are painted white. The rest of the array has the gray finish as it comes from the manufacturer of the acrylic mirrors. You can also see that the backs of some of the reflectors are lit up by the partially focussed light from the adjacent reflector.
As the sun rises just a bit more, there will be no back reflection and the sun's heat will be fully focussed onto the collector tubes.
The very good news is that none of the five mirrors in the foreground that have the white applied to their backs have experienced any degradation due to back heating. They have been in place for about a month and a half now. I've ordered new mirrors to replace the ones that are still gray and which have deteriorated somewhat.
The title of today's post is an opportunity for me to tell you about another strangeness of the system that is a preamble to another improvement that I intend to make.
I have purposefully tried to keep the system as simple as possible. There is no computer required, for example.
There are only two ON-OFF switches to make the system operate, one for the pool pump which starts the flow of water and another for the tracker which enables the tracking circuit to drive the small gear motor that swings the reflectors to follow the sun.
The pool pump is a very hungry motor - one horsepower, about 750 watts. I don't like to leave it running all the time. It does not need to run all the time. But it should be on when the tracker is on otherwise the water in the collectors is stagnant and will boil inside of about 20 minutes. This sometimes causes a blowout of one of the fittings and this is not a good thing.
The tracker does not need to operate all the time either, so I've been switching it off at night. This leaves the array facing west. When the array goes as far as it can, 50 degrees from vertical to the west, it will stop anyway due to a limit switch.
Recently, I've been forgetting to turn off the tracker, or I have been leaving it on purposefully to make sure that it senses the morning sun when it is heeled over west, 100 degrees away.
The good news is that it does sense the morning sun and begins to swing east as soon as the sky begins to lighten. In fact, by the time I am having my first cup of coffee, about 5:30, it has already reached the east limit. The sun has not yet broken above the horizon.
But here's the strangeness: as the sun breaks the horizon in the east, the tracker starts to head west again. It is about 6am and it is heading west. If I let it continue, it will, by about 6:30, be headed fully west again, at the west limit. What is going on?
As I mentioned earlier, there are trees to the south. Until about 9:45, when the sun rises above the tree tops, the Red Rock sensor is in the shade. The western sky looks to it to be brighter than the trees to the east, so it heads west. Although it correctly sees the early first light in the east when the west sky is still dark, it then "makes a mistake" when the sun is rising behind the trees.
Obviously this is a site issue. If I didn't have the trees to the east, the tracker would be working fine. But I'm not going to cut the trees down.
The temporary solution is to switch the tracker off until the sun is above the trees.
The longer term solution is to add a bit of automation: to sense hard sunlight when the sun is above the trees and to turn on both the tracker and the pump.
I will be discussing my approach to this is another post.
Thank you for your interest in my project.
Monday, July 06, 2009
I had been experiencing deterioration of the plastic mirror over time in the form of linear marks that appeared in the mirror and caused it to mark and then warp.
I'd shown a picture (similar to the one above) of the backs of the reflectors "lighting up" with the partially focussed light as occurs before the reflectors are properly tracking the sun or when they are left stationary and the sun passes overhead. When they are properly tracking of course, virtually all of the sun's light is focussed on the collector and virtually none of the light spills. If they are parked horizontally, no light is reflected onto the adjacent mirror.
Here is another picture of a badly deformed section of a mirror. It is very difficult to take a picture of a mirror, but I hope you will see that there are linear marks that severely distort the surface of the mirror.
I mentioned that I had left the array over the winter with the reflectors pointing at the western sky, thinking that this would allow the accumulated snow to fall off the reflectors thus reducing the heavy load of snow. That turns out to have not been that effective. Better to leave them horizontal since they seem able to take the load.
Also, on a sunny winter day the reflected light of the sun will move across the back surface of each mirror that has one beside it. The end reflectors did not show this effect and that was really the clue that I needed.
I mentioned that I have installed five new mirrors in the array after painting their backs with bright white paint. It has been over a month and none of these new mirrors have shown this effect.
Today, as a further test of the theory, I focussed the light from a magnifying glass in a clear undeformed section on the back of one of the mirrors that I had removed. It was a large magnifying glass. I tried not to focus the light right down to the smallest concentration that was possible, but rather, I tried to get about a 15:1 concentration (a disc of light about 3/4 of an inch across from a magnifying glass about 3" in diameter) to approximate the back-lighting that I thought that the mirrors might be receiving.
It didn't take more than a few seconds to hear a sort of a crackling sound. When I pulled the magnifying glass away and took off my protective eye glasses, I could not see anything unusual on the back side of the mirror. But looking at the other side, there was a mark that looked very much like the problem I have been fighting!
Here's picture of the damage I caused. See what you think.
What looks like double exposure is the dust on the front surface of the acrylic and it's reflection from the back surface mirror.
My hand was not steady enough to make a track like the sun does as it slowly moves across the sky, but the appearance of the spots that I made looks to me very much like smaller versions of the marks that I have seen far too many of.
With a solution in hand (painting the backs white) I am confident of the durability of the acrylic mirror.
Once again, this is not the only reflective material that can be used in the DIY Solar Heater, but it is the cheapest and easiest to use for this application.
Saturday, July 04, 2009
Any electrically controlled machine needs limit switches. In the case of the DIY solar heater, allowing the motor to drive the mechanism too far could cause various sorts of damage.
The switches mounted last year did not handle the winter well. The rubber seals had already started to crack and one metal part had started to show rust. I'll describe the solution that I have come up with that hopefully will prove to be more long term. This is the limit switch which will be detailed in the plans.
I mounted a weatherproof standard electrical box on the drive arm to hold the switches. As the drive arm rotates, the box (and the whole array) tilts with it.
Inside the box, I mounted two mercury switches arranged at 50 degrees from the vertical, so that as the assembly tilts with the drive arm, eventually one of the two switches will cut off the motor. The motor current runs through both switches.
You can see here (click any picture to get a closer view) that the two mercury switches are fastened to a wooden plate with small screws so that their angle can be changed slightly. Two silicon diodes wired across the switches allow the current to flow in the other direction if either switch is open at the limit.
In other words, if the west limit is reached, the Red Rock solar sensor can tell the motor to go east, but no further west.
Mercury switches are interesting devices. They contain a small drop of the metal mercury which is enclosed in a sealed glass vial along with two (or sometimes three) contacts. Since the mercury is conductive, when it sits in the end of the vial with the contacts, electricity will flow between the two contacts. When the vial is tilted so that the mercury flows to the other end, electricity cannot flow.
Mercury switches are capable of handling relatively large electrical currents, more than the one amp drawn by the motor used here. Since the vials are sealed and presumably in a vacuum, there is no oxidation or contact deterioration if there is a spark when the current starts to flow or when it is interrupted. Because they are sealed, they make ideal switches for outdoor use in this application.
Mercury switches are commonly used (or were commonly used) in room temperature thermostats and in auto applications such as turning on an under hood light when you raise the lid of the engine compartment. Mercury switches have a bit of a bad rap these days because the mercury is hazardous to the environment if it escapes. So this is in fact a good use for old mercury switches. I got mine from an old Honeywell commercial thermostat which actually contained four of these switches. I've since found a commercial source for the switches but they don't have the nice mounting bracket that the scavenged switches came with.
Here is the limit switch assembly being bench tested. I had clamped the motor drive vertically in a bench vise and I drove the motor with a bench power supply so that I could adjust the limit switches to cut off the motor in each direction at 50 degrees.
You can see that in the right switch the blob of mercury has moved to the left end of the vial, thus pulling away from the contacts at the right end of the vial. The motor drive was rotating the drive arm counterclockwise but now cannot go any further because the motor current which goes through the switch is now cut off.
There will be more detail and a circuit explanation in the plans, but for now, if you want to know more, you can visit the Red Rock website (www.redrok.com) to see how Duane has the circuit arranged.
Here is the finished limit switch assembly ready to be mounted on the drive arm.
It has been working well in the array for about two months now. After the initial bench checkout and a further test run on the array, I haven't had to worry about it or adjust it or even open it up. It has been very reliable.
As for all the components of the DIY solar heater of course, the long term performance in our Canadian climate will be the best test of all.
Friday, July 03, 2009
Experience with it has led me to improve it further still in several ways. In this photo (click any photo to enlarge it) you will see the current motor drive. It has been operating for about two months now and this will be the motor drive detailed in the plans.
I kept the same basic arrangement: the gear motor (inside the right most box) turns a lead screw (the horizontal shaft) on which is mounted a circular brass nut that is able to ride up and down the channel in the drive arm (the one with the smaller box attached to it). The turning of the lead screw pushes or pulls the drive arm so that it rotates around the pivot at it's top. This pivot is in the same plane as the collector tubes and it's length is the same as the length of the arms that hang down from each reflector. In this way, the push rod which runs down the entire length of the array and couples all the reflectors, moves up and down along the same arc as the motor drive. The push rod now stays perfectly straight along the whole range of motion of the array, 50 degrees to either side of vertical.
The first improvement was in the construction of the backing board. Previously I had laminated a number of 2x6 boards of pressure treated wood together to get a suitably sized slab. While it worked well for a time and helped to show that the principle was correct, it wasn't long before the whole slab started to warp from the sunlight which shines mostly on the backside of this slab. The whole thing started to cup so that the drive arm began to bind against it's surface.
I made a new motor drive using the same technique as for the ribs. I cut the profile from good quality plywood and treated it with the same marine paints as I had used on the ribs to seal it. After two months now in some fairly direct hot sunny days, it is still perfectly straight.
The second improvement was to stiffen the stainless steel push rod by changing the 3/16" rod to a larger 1/4" rod. It had had a tendency to bend when pushing the array to the west (west is to the right in this picture). When moving to the east (in my arrangement at least) the rod is pulling and didn't have this problem. With the various changes made, the reflectors are quite as well balanced as they were intended and with 13 reflectors, there is enough force required that the rod needed to be stiffer.
Here is a closer view.
Another improvement I made was the use of standard blank plastic electrical boxes to contain the motor drive (the right box) the battery (the lower box) and the limit switches (the small box on the drive arm). I will write about the new limit switches in another update shortly. The new boxes look better than my handmade wooden ones of course, but they will also help in the weather sealing of the whole system.
The wiring was also completely replaced and run within plastic flexible conduit, another standard electrical product available at most home centers. The previous three year old wiring had become unreliable and needed replacing. The new wiring includes terminal blocks and crimped terminals throughout so should be much more reliable. I am trying to achieve a 5-10 year life for all components.
I currently run the system from a small (6 amp-hour) gel cell that I replace with a freshly charged cell about once a week. I intend to install solar panels to run this system as well as a DC operated pool pump this summer, so this battery should disappear shortly.
As a part of the general re-wiring, I changed the support and packaging of the sensor housing slightly. Previously, the Leviton "while in use" cover that I had removed the lettering from and polished (described in a previous article) had been bolted to the support platform (the white thing) and there wasn't quite enough room for the wiring and another improvement that I had wanted. I wanted to be able to swing the Red Rock sensor slightly on a pivot to adjust it's alignment.
I added a standard plastic electrical box under the sensor cover to give more space for the wiring and for the sensor swing mechanism. This will be described in another update shortly. The improvements have worked out very well. The wiring is now more easily accommodated and it is now very easy to swing the sensor slightly to change alignment and have it stay in place reliably.
Previously I had been bending the sensor on its wiring posts to change its position and this was not a good arrangement at all.
Here is a view of the new motor drive from the back.
One thing that was tricky was getting the pivot for the drive arm in the same plane as the collector rods. You can see the crossed wood struts at the back of the motor drive used to mount it to the post. These were used with wood clamps temporarily to get things into the right position with a bit of trial and error and then screws were driven in to hold everything permanently.
So far, everything has been working as planned.
Thursday, July 02, 2009
When I started using acrylic plastic mirrors in 2005, they seem to be a perfect candidate: readily available, low cost, lightweight, UV stable, strong and with excellent reflectivity. The mirror surface (a sputter coated aluminum film) is actually on the back surface so that it is well protected from scratches that might result from cleaning. The mirror material that I use is 0.060" thick and it bends nicely into a parabola shape formed by the ribs.
It wasn't long however, before I saw the first of these "marks" you can see in the picture (click to enlarge). The first mark was a curious thing. I just noticed it one morning streaking across a few inches of the otherwise perfect mirror, a bit in from the edge. It seemed to be a discoloration of the plastic or the mirror, I could not tell which and it seemed as if something hot had hit the mirror at an angle. Looking from the bottom, it seemed that the grey material covering the aluminum coating was cracked or crazed in a fine pattern.
I remember writing to a friend at NASA with a picture enclosed asking him if he thought a meteorite had bounced off the mirror? That's what it looked like.
Soon more marks started to appear. Many more.
In 2006 I had only four reflectors in place. Two of them were metal sheets, one was aluminum and the other was galvanized steel. The metal sheets did not show the effect.
This is what the array looked like this spring. Notice that the marks seem to have a common orientation. Interestingly, most of them were on the west side of the reflectors. A few had appeared on the east sides, but only a few.
By themselves the marks were a curiosity but might not have been that bad except that as they got worse, the acrylic sheet started to buckle in the area of the marks, pulling the sheet out of contact with the parabolic shaped ribs. It did not take much of this for the focus to deteriorate and for the heat captured by the collector to fall.
Here is what one of the sheets looked like along the edge and you will see clearly what was starting to happen.
By this time I was on the phone to the vendor Laird Plastics (www.lairdplastics.com) sending them pictures and asking what could be going on?
As a footnote, most of the components of the DIY Solar Heater are available from building material stores (like Lowes or Home Depot). When I have used special items, like the plastic mirror or the gear motor, or the solar sensor, I have chosen items from national distributors so that these should be readily available, at least in North America.
The local Laird reps were great. They showed concern and gave me reassurance that: 1) the acrylic was UV stable and was regularly used in glazing (windows) in commercial buildings, so the mirror should be stable also and 2) it was good for outdoor use as it was fully capable of the temperature range. They offered some helpful suggestions also like was the acrylic constrained by the mounting so that there was no room for expansion? I assured them that the metal strip along the edge (furring strip) was springy and that it had some "give" if the sheet was not exactly the right size. They also asking if there was some cushioning against the ribs where they bear on the sheets. I had used foam strips on about half of the reflectors on the edge of the ribs which would otherwise touch the backs of the sheets.
Since most of the reflectors had the problem, that was not the cause. Most of them, that is, except for two, the ones at each end of the array. These two DID NOT have the problem.
The Laird people were great, but they were as puzzled as I was. In 2007 I replaced all the sheets with new ones.
But by this spring however, all of those (except for the end ones) had shown the curious marks and deformation again. I was pretty desperate by this time and this was one of the reasons the project was dragging a bit.
One morning about two months ago I was out waiting for the array to swing east to catch the morning sun. This is what I saw.
The array does take a while to swing the entire 100 degrees from west to east, about 1/2 hour, since I have the RED ROCK sensor turned to a very low duty cycle to conserve battery power. At the time of the picture, the reflectors have not yet pointed at the sun.
Notice in the picture that each reflector throws light on the reflector in front of it, except that is, for the one at the very end! The light is not focussed well since the reflector ahead is far out of the focus, but it is a good blast of concentrated light nonetheless. It also has hot spots if the reflector causing the beam is already somewhat distorted.
I put my hand on the back of one of the reflectors at a hot spot. It felt uncomfortably warm. Could it be that this was the cause of the streaking marks and distortion deterioration?
I'd imagined a few different solutions to this problem. I decided first to replace a number of the mirrors (I had five spares left over from last year's batch) but this time before mounting I painted the entire backs of these new mirrors with bright white acrylic paint. Three coats in all.
I'm hoping that the white paint will dissipate most of the unfocused light (heat) from the adjacent reflector. I've installed these new mirrors with the painted backs about a month ago at the west end of the array and I am watching them closely.
This picture is of one of the new, white-coated mirrors receiving its daily blast of light from the one beside it as the array tracks from west to east in the morning.
So far, none of the new mirrors have shown the marks or deformation.
As a test of my theory, I blasted the back of one of the mirrors that I had removed with a heat gun for quite a few minutes, long enough for it to get too hot to touch. No deformation or marks formed. That was a bit disappointing. Perhaps the effect comes about because of radiant heat as opposed to convection heat? Another experiment to try is to blast the back of a mirror with a magnifying glass aimed at the sun. I will do this shortly.
I'd mentioned other possible solutions. One is to use another type of reflective sheet . Another is to speed up the solar sensor so that the array sweeps from west to east more quickly. Another is to "park" the array in the vertical position, rather than leaving it at the west. In this position, no light would hit the back of any of the reflectors.
However, if the white paint on the backs solves the problem, then none of these will be necessary.
Incidentally, I had left the array parked facing east this winter, thinking that the snow load would be shed more easily if it was off to one side. The east direction was really arbitrary. This may explain the huge predominance of the marks on the west sides of the reflectors. These ends were sticking up in the air receiving the periodic blasts of light on the few days that it was sunny. The array had been parked this way for about six months. This winter I will park them horizontally, for sure.
Thanks for your continued interest in my project.
Wednesday, July 01, 2009
Performance has been somewhat of a moving target since I have been making various changes and improvements but I have enough experience now to provide some detail on the measurements that I have made, to comment on some of the factors that influence the results and to show the calculations that I make to determine the heat capture of the system and to estimate it's efficiency.
With apologies to my metric readers, I will be using Imperial measures.
My solar pool heater project currently consists of 13 parabolic trough reflectors faced with acrylic mirrors which rotate automatically to track the sun over 50 degrees east or west from vertical. At the focus and mass center of each parabola is a matte black coated copper collector tube through which the water flows and receives the sun's heat which is concentrated by the reflector. The reflectors hang down from the collector tubes and simply pivot on them. The reflectors are pushed or pulled into position by means of a slender steel rod which couples them all to a small gear motor drive.
The collector tubes are connected in series so that the pool water flows through each collector one after the other.
By measuring the flow rate and the inlet and outlet temperatures of the water, I can calculate the amount of heat which is being added to the water by the solar array.
Inside the pump shed, I have installed a pitot tube acrylic flowmeter from Blue-White Industries Ltd (www.blue-white.com) model F-300 with a range of 20-100 gpm as well as a dual differential thermometer from Fluke Electronics (www.fluke.com) model 52 II, shown here (click on any picture to enlarge it).
The Fluke thermometer is at the top, the Blue-White flowmeter is to the left.
The temperature sensors for the thermometer are located on the white ABS pipes under the green bands. The bottom pipe is the water feed to the array, the top pipe is the return line from the array.
The Fluke thermometer measures temperature to one tenth of a degree for each of two sensors and will conveniently display the temperature difference between the two, also to one tenth of a degree. Fluke claims an accuracy of plus or minus 0.2% when used with K type thermocouples (the sensors that I am using) when under 100 degrees C.
The Blue-White flowmeter is inserted in a hole that I've drilled in the ABS pipe and is clamped to the pipe with the band clamps. It protrudes slightly into interior of the pipe and draws a small sample of the flow which is diverted up through a channel in which rides a small steel puck. By reading the height of the top of the puck against the scale, I can read the flow rate, in this case 37 gpm.
Blue-White claims a full scale accuracy of plus or minus 10% for this meter. I would have preferred to use a more accurate type of flowmeter but the approx $75 cost of the F-300 fits the project budget.
The water flow rate is fairly constant at about 37 gallons per minute. It will fall slightly to about 35 gpm when the sand filter needs back flushing but after that operation is done the flow rate will return to 37 gpm. Because of the +/- 10% tolerance, my actual flow rate could be as low as 33.3 gpm or as high as 40.7 gpm. I've used 37 gpm for the calculations.
The temperature difference readings I have taken (between the inlet and outlet) over the past two years vary considerably from a low of about 1.8 degrees F to a high of 2.6 degrees F.
Obviously, a bright midsummer sun high in the sky (ideally at solar noon when it is directly overhead) against a crystal clear blue sky gives the highest reading. Passing clouds impair the result as does a hazy sky.
Dirty reflectors covered with dust or pollen give lower readings. My best results are obtained after the reflectors have been washed by a recent rainfall. I seldom wash the reflectors myself since I've found that once the spring pollen season is over, the reflectors pretty much keep themselves clean.
Poorly focussed reflectors (because the reflector drive linkage or the sensor aim has not been properly adjusted) will give lower results.
Reflectors that do not properly conform to the parabolic shape of the ribs due to deformation will give lower results. I have been experiencing some deformation of the acrylic mirrors due to backside heating which I have recently solved by painting the backs of the mirrors with white paint. Only five of the 13 mirrors have been treated this way and the results so far are that problem has been solved for those five reflectors. The other eight show various amounts of deformation so that I cannot achieve the best results now until those eight are replaced.
I will write more about this problem and it's solution shortly.
The high reading of 2.6 degrees F was reached last year when all mirrors were clean and not deformed in any way and the array and sensor focus had been carefully done. That reading was taken in late August. Had it been done on a clear day around the summer solstice (June 21) it might have been higher still. Nevertheless, I have used 2.6 degrees F for the calculations even though I have not recently seen a value that high for the reasons mentioned.
So, using a temperature rise of 2.6 degrees F at a flow rate of 37 gpm:
One BTU (British thermal Unit) is the amount of heat required to raise the temperature of one pound of water by one degree Fahrenheit. Heating is usually expressed in BTU/hr.
At a flow rate of 37 gpm, in one hour (37 x 60) = 2220 gph.
The weight of one gallon of water is 8.345 pounds, so the weight of water which flows through the array in one hour is (2220 x 8.345) = 18,526 lbs.
If the water is heated by 2.6 degrees F, then the heating power of the array is (18,526 x 2.6) = 48,168 BTU/hr.
One BTU/hr is approximately 0.0002931 kWh so the array generates approximately (48,168 x 0.0002931) = 14.118 kWh.
There are 13 sections in the array, so each contributes:
(48,168/13) = 3705 BTU/hr
(14.118/13) = 1.086 kWhr
I estimate the efficiency as follows:
The sun energy (insolation) in my area (43.7N -80.0W) that falls in one square meter (1550 square inches) is approximately 1150 watts or (1000/1550) = 0.645 watts/sq.in. The optical size of each section of my array (the aperture) is 19.375" x 96" or 1860 sq. inches and I am capturing 1086 watts per section or (1086/1860) = 0.584 watts/sq.in. So the capture efficiency is approximately (0.584/0.645x100) = 90%.
This result does seem to be a bit high and probably not right. I will continue to work on this.
As always, your interest and comments are welcome. I have much more to tell you about and you can watch this blog for updates.
The plans are coming along and should be available shortly. Thank you for your patience.
Update: My plan book is available here.