Is it safe to build sealed tanks in a rectangular shape filled with the sort of water pressures used typically in homes? Can home-built rectangular tanks be shut tight by bolts, and yet prevent leakage? I suppose that, where there are enough bolts, it can be done. But it's a risk to even try. Still, it won't be a lost cause because you can use it as a separate system with less-than-household pressure.
A liquid under pressure is not as dangerous as a gas under pressure. This is due to the incompressibility of a liquid. Virtually as soon as the liquid under pressure finds an escape route, the explosive potential or elasticity of the water body is gone. Homemade tanks of the sort I'm talking about here are best not to be put under high pressure, but it's your call.
If you are using 40 psi or more at the cold taps from a regular pump-and-tank situation while wanting far less pressure in your roof tank, the latter will need to be a separate system, with water obtained elsewhere than from your pressurized pipe, unless there is a pump on the market that takes from a high pressure pipe but reduces the pressure to your setting. The alternative is to have a separate water system as an open tank, filled at a tap, with an automatic shut-off valve at the top of the tank. You then need a separate, low-pressure pump to take the water from this open tank to the roof tank, which will start the second you begin taking water from the roof tank.
Pumps are not always very quiet. Unless you have yet another pump to take roof-tank water to your hot water pipe (ouch, three pumps going at once for a shower), you will need to feed your tub and/or kitchen sink and/or laundry machine with separate pipes from the roof tank, under a gravity-fed situation. But at least there's a chance that the waters are at least warm. No one wants to live this way, but for trib purposes, I'm doing this chapter. It assumes that you'll have a little solar electricity to run at least the one pump to get 40-50 psi throughout.
The beauty of a home-built heater tank that can handle 40 psi is that you retain the one pump only and simply run water into the tank. You can still retain one pump only if you end up filling the heater tank to a non-pressurized situation (with the automatic shut-off valve). Make sure that there are automatic shut-off valves that work in a thin tank (a few inches tall) tank positioned on the angle of your roof or elsewhere.
My laundry machine will automatically "fail" if the water to it is too slow. It's programmed to think that there is a flood on the floor if the water supply is too slow. I've cursed Sears for this and other traps that it's programmed for. It's not trib friendly, in other words, and computerized machines may all come like that now. I'm just giving you a heads up in case you begin to hook the laundry machine to a low-pressure roof tank. My machine won't even allow a soak period for more than about 30 minutes before failing; I am then required to call the repairman to reset it, because Sears won't tell me how to reset it myself. It's a rip-off scheme, guaranteed. Computerized machines allow manufacturers to add programmed booby traps that cost us money, to their advantage. It's what hell was designed for.
Low pressure pumps may be less noisy and, hopefully, run on less power. It would be worth having one if those two items are true. Again, a low-pressure pump takes from an open tank and feeds the heater tank, giving you some pressure. For all I know at the moment, some regular water pumps may have a very-low cut-off point less than 20 psi. At 10 psi, it's still 1,440 pounds per square foot, requiring thick plexiglass, lots of bolts, and dependable washers / gaskets.
A large tank the size of three or four manageable ones would not be monstrous. Its predictable size would be 4' x 8' at a depth of a few inches (roughly 80 gallons or less than 700 pounds of water) so that a single sheet of plywood could be used for the back side. However, for high-pressure situations, a 4' x 4' tank is far less risky because the two 8-foot frame pieces may bend too much. Bending wants to "stretch" the plexiglass, but it's not stretchable, unfortunately.
So long as the plywood lies flat on a solid "floor," water pressure won't harm it, but will transfer instead to the plexiglass and the four pieces that are the frames. There is a question on how much bending pressure a 4 x 4 piece of frame, four feet long, will handle, while bolted to the plywood back and to the plexiglass. Both the back and the plexiglass will act to keep the frame piece from bending, to a point, and that's why you may opt for a double-plywood backing, but only if the first piece of plywood is weaker than the plexiglass (otherwise what's the point?).
If you are gunning for some pressure (call it fun), you can't tighten the frame's bolts merely through the plywood and plexiglass, unless you space the bolts very closely together. You won't know how close until you try it. The better alternative so far as reliability is concerned, is that both the front and back surfaces get angle iron encircling the entire frame. You would then tighten the bolts to the angle iron, and hope to keep the seal waterproof between bolts. If not, drill more holes, ouch. This is the toughest part of the job, unless your angle iron, strong enough, comes pre-drilled. Theoretically, the thicker the angle iron, the fewer the bolts needed. It's surprising how fast a new drill bit will go through 1/8" of metal, but it will still be a longish job to do the perimeter on both front and back. How many drill bits will you go through? Depends on the quality.
The benefit of angle iron around the perimeter, versus flat steel (less money, I know), is that it prevents some bending of the wooden frames. You do not want to use the plexiglass to keep the four frame pieces from bending. Alas, for the angle iron to be best in keeping the frame pieces from bending, it needs more bolts. Or, put it this way, that bolts every six inches is better than ten. Metal brackets at the four corners makes sense.
You can actually test how much pressure your tank can take. It's even easy. You just fill the tank with the pressurized house water, but let the water exit through a valve, for example at your drain hole. Just before this valve, you install your pressure gauge. You then start to close the valve a little at a time until you see signs that the tank is about to explode. Kidding. You check for leaks. No leaks? Close your valve a little more, and check for leaks again. If ever you close your valve completely, the tank will immediately go to the pressure that your house water is at, so don't close your tap until you are sure the tank can take the pressure.
If it leaks, it won't be any more powerful than when you open a tap, or when you do a bad plumbing joint. The good news is that, after you've done all that work, and it leaks, the fix may be as easy as one or two bore bolts. The bad news is if the plexiglass cracks and needs to be changed at the risk of cracking another. I don't see heavy plexiglass cracking easily, but I'm not an engineer.
The pressure of a sealed tank goes up a lot when heated by 60 or 70 degrees. One page: "As a point of reference, in a residential hydronic heating boiler we actually observe an internal water pressure rise from 12 psi cold up to 28 psi hot as the boiler temperature increases from perhaps 60 °F up to 180 °F." It appears that we can expect a doubling of pressure in these solar-heated tanks. However, industry computations such as this one could have a large safety factor built in, which gives a final pressure far above the reality. I know that structural engineers have a large safety factor when dealing with winds on buildings.
In any case, a relief valve is your tank's salvation. Fear not.
Same article: "Typical pressure for a residential boiler serving a two story home would show 12 psi cold, and less than 30 psi hot. Over 30 psi boiler pressure will cause the pressure relief valve to open."
It seems that the boiler maintains, with a pump made for the purpose, a lower pressure in the tank than the household pressure. That's the type of motor we want for our solar tanks.
My understanding is that, in a household hot-water tank, a relief valve is needed because the burner can remain on continually in a defected system. In the sun, however, with a batch heater of this type, there is a maximum temperature achievable well below boiling point, and the heat added at any time is nothing like that from a flame under a hot water tank. With a relief valve, if it can be set to your choice of pressure, you protect the entire tank from breakage.
The valve will spit out water for the entire day that the sun gets the tank to your set pressure, but water on your roof is okay. Install two valves in case one decides to malfunction, and inquire on whether the valve will last long if it's used regularly. I don't see a problem, but it won't hurt to ask.
There is a way to keep the valve from spitting much, if you have the cockpit situation I talked about in the last update, which is in the next chapter too. It allows you to shut the cold water to the tank from your shower stall. Once that's shut off, and the valve spits out some water, it will be a longer time before it spits out again, and it might do so only three or four times all the rest of the day. It might not be a bad idea to shut the cold water off as a daily routine, until shower time. The cockpit allows one to drain as much or as little of the roof tank as desired, which instantly relieves the pressure problem.
The same article gives some numbers that can suggest to some that the pressure build up is not nearly as bad as the numbers above:
We are using the coefficient of thermal expansion of water = 0.00021 (1/oC)
1 deg C rise in the water tank = 1.0002 Unit Volume Rise
or if our water container is closed, such as a water heater tank, we should be able to say 1 deg C rise in the water tank = 1.0002 Unit Water Pressure Rise
I think that's wrong. The increase in pressure versus increase in volume is equal only in a gas, but not in a liquid. If you could stuff twice as much water in a tank, it would far more than double the pressure. However, the numbers above are useful for showing us how much extra water volume is provided for a temperature rise of 70 degrees. The coefficient number above times 70 degrees says that volume increase is 1.015 times, meaning that 80 gallons in a tank turns to 81.2. It means that the relief valve will spit out only a fraction of 1.2 gallons, if the cold water to the tank is shut off. You could arrange for a slow leak by opening the drain valve a touch. A gallon of water wasted each day could be your target.
Or, what about putting a few kid-size, rubber footballs in the tank that compress under pressure? Sounds excellent to me, but won't do any good if the cold water is open to the tank. If they can take the heat (do a test), put the footballs in, shut the cold water after each final shower in the day, or each morning, and the pressure problem due to temperature increase can be obliterated...not at all meaning you're fully in the clear. The tank still needs to withstand the cold-water pressure.
The problem with balls in the tank is that they get equal pressure all around, which tends not to flatten them, but to keep their spherical shape. It seems to me that we want items in there that flatten out so that all their spaces are obliterated. Something to think about.
We can do the math another way if we begin with a tank four inches deep. We just multiply 4 x .015 to find that the volume increase wants the plexiglass or the plywood base to move over by .06 inches, or 1/16". That's 1/32" to each the plexiglass and the plywood, which isn't much, especially for the plywood. It doesn't matter how much the pressure goes up, the added water volume is the fact of life here. Any bowing of the plywood relieves the plexiglass of pressure, as well as all other parts. We can therefore welcome some plywood bowing, and even invite it. We might use "soft" washers of hard rubber or plastic, for example, between the plywood and the bolt head. In other words, don't use metal strips under the wood dividers down at the plywood bottom, as you would over the dividers at the plexiglass. Let the plywood bow.
Unfortunately, the plywood's memory might be as bad as mine. If the plywood doesn't bounce back to its original shape after it's been bowed, and new water comes into the tank while it's still partially bowed, isn't it going to want to bow more the second time around than on the first? Yes. But the bowing will reach a maximum at some point, and from there I don't know how much relief the plexiglass will get sunny day to sunny day.
There are flexible plastic sheets that can take this beast a lot better than acrylic, but, drats, they are not transparent. If you choose, you can do your pressure test with plywood on top, to check whether the bolts or gaskets leak on such a scale that you decide to scrap the project altogether. In fact, you can still get warm water if the top sheet is metal on a permanent basis, but I don't know the details on the respective performances between a metal lid versus plexiglass.
If you have success at a certain water pressure, leave it there all night. Give the tank a good look-over the next day to see how it seems to be handling it. If this is your small test tank, you can make necessary changes for your larger, permanent tank(s).
Aside from bolts that go clear through the plexiglass and plywood, you can lace the back of the plywood with screws into the frame. Every little bit helps. Aircraft skins have rivets every centimeter or so for a fantastic hold. The same theory seems helpful at the plexiglass. How badly do you really want warm water? It's best to ask that question in a freezing shower stall.
However, plexiglass expands a lot more under increased temperatures than wood. If the bolt / screw holes are made tight to the bolts / screws, there's no give for expansion, and I don't know whether the expansion process will forgive the bolts / screws by going around them. It's probably best to allow the frame pieces to bend slightly, by making the plexiglass' bolt and screw holes larger than the bolts / screws. This is not necessary or advisable for the bolt / screw holes through the plywood (they should be snug).
You can use plastic-wood (fully plastic but made to act as good as wood in every way) as an alternative to the wood frame and dividers, and even use a plastic sheet as the plywood alternative. It's expensive, but will solve all expansion problems, and allow tight bolt / screw holes. And it will look and keep better. But this chapter is for an on-the-cheap attempt to get at least some pressure in a heater tank.
How in tarnation can plastic lumber be several times the cost of wood lumber when no one needs to go cut plastic trees, or to drag plastic trunks of multiple thousand pounds to a saw mill for heavy-machine handling? Government-paid trucks just drive the plastics from our homes, to a slave house where workers separate plastics, and its then transferred quite easily to plastic manufacturers who in turn transfer it in a different form to those who melt it for casting into plastic lumber. Piece of cake. It doesn't require extreme temperatures to melt it, nor does anyone need to mine plastic rocks with huge machinery and explosives, and yet cast plastic is roughly the same price per pound as iron rods and tubes. Does someone smell a rip-off industry here? Governments are now practicing a secret tax on certain materials to pay for a global government, and to get us to use less materials, because that's what globalists want, to have us using less of everything. Raising prices is their attempt to have us use less, and it probably works, but it's painful on our end.
Forget the plastic sheet if you don't want long term. Use plywood covered with a waterproof material. What about a plastic film, the typical vapor barrier at building-supply stores, maybe a second layer as insurance against a leak. If water in a slow leak gets past a bolt to the underside of the plastic film, the water will run down the plywood, which is why you can drill a few small holes (1/8 inch) down at the bottom to let the water escape. If you want longer term, assure no leaks.
The upright dividers (from bottom to top of tank) can be 16-18" apart, more if better water pressure is the goal. Bolt the plexiglass to these dividers, with rubber washers inside and outside, but don't go bolt happy for a non-pressure situation, because each one is a leak possibility. The only thing stopping you from attempting a pressure situation is the cost of the plexiglass (hundreds of dollars), and the risk of using it in case it cracks due to expansion, or leaks without a solution. Otherwise, I know you'd try for a pressure situation.
For long term and for some water pressure, using dividers of 2 x 4 oak should be a lot better than spruce. The more dividers you add, the thinner your plexiglass can be. The dividers won't waste much solar energy if they are black to absorb light from the sides (morning and late-afternoon sun), but, the question is, how long will wood dividers hold the paint, and what happens to it when it peals? What about burning the wood black with a torch? Will that black last? Will it leach into the water and get on the plexiglass? There may be a black-rubber solution for these questions.
Instead of fiddling with connecting two or more 2' x 4' tanks with short pipes, we can make a 4' x 4' or 4' x 8' tank and just arrange holes lickity-split through the dividers, allowing water to pass from section to section. The downside is handling the larger tank, but you and a friend or two could raise the tank partially assembled to the roof (pull it up along two ladders), and finish it there.
Having a plywood back directly on a roof is asking for rot, yet it's a stable way to handle the weight of water on merely a sheet of plywood. If you plan on leaving the tank in one position, without changing its angle to the sun, just silicone at the top of the plywood to keep rain water from getting in. It should get you three or four years trouble-free.
If you want water pressure, the plywood will want to bow even though it's on a flat roof surface. With great-enough pressure, the force will overcome the overall weight of the tank, and start to bow the plywood into a pillow shape. To eliminate this, the frame will need to be bolted to the roof. The number of bolts will depend on the size of washers beneath the roof underlay.
There is a significant difference of 23 degrees between the sky of a March-21 sun verses a June-21 sun. From horizon to directly overhead, it's 90 degrees, meaning that 23 is significant. In the south, where solar-heated water is possible in the cold, the angle difference between January and June is even greater.
To change the angle of the tank (with water drained to make it light enough to raise one side), it's going to need a stable base such as a number of 2 x 8 spruce boards (flats facing roof) in the horizontal direction. Under the 2 x 8s, run and bolt to them at least seven 2 x 4s, on the flat, in the opposite direction. The 2 x 4s contact the roof but do not need to be bolted through the roof. The 2 x 8s not only allow you to lift the tank up for changing its angle, but allow water pressure against the plywood back.
It's not recommended that the tank's bottom end be lifted to change the angle, for the bottom needs to butt against a 2 x 6 bolted to the roof, to keep the tank from sliding off the roof. The "washers" on the inside of the attack, for the bolts of this 2 x 6, could be piece of 3/4" plywood roughly eight inches square (with the bolt holes drilled in, of course), not pieces of spruce that might split. You can screw the plywood to the roof's underlay to help keep the bolts from tilting under the weight, because that's what I imagine (especially in steep-roof cases) if the only thing holding the bolts is the 1/2-inch underlay. It won't hurt to include a large steel washer on each plywood "washer".
So, lift only the tank's upper end, with a man at each side. No hinges needed. Put something(s) stable under this top end to keep it from falling to the roof when filled with water. You'll need flexible hosing to a tiltable tank. No problem.
The metal heat collector can be in several pieces, each one between the dividers. I've read that it should be painted black, which is why you need the proper metal paint that won't peal very fast, and more than one coat. Maybe paint the back and sides of this metal in case rust is a problem. However, I'm wondering whether metal is necessary at all as the collector material.
Before you build the large tank, why not try a small one of 16 or even 12 inches wide (still four feet from top to bottom). Bang the case together quick only to see whether it can handle the water pressure you desire.
If you want longer term (a great idea), use a rubber backing instead of plastic film. Let it be one, seamless piece, folded up over the frames. If it's black rubber, are metal plates still needed? I don't know. My thinking is that any black color absorbs all light regardless of what the material is. So why do they use metal in solar heaters? Yes, metal absorbs heat well and readily, but we don't want the metal to absorb heat. We want it to give up it's heat to the water. Black color, as I understand it, is an indication of how little light is reflecting off a surface, failing to register as color in the eye. Therefore, both the black metal and the black rubber are absorbing most of the light. So why use metal? Black rubber in the sun gets darn hot too, doesn't it?
Whether you use rubber or plastic, it is best to keep it under water in summer, in case you decide not to use the tank for a few years. My experience with 6 mil plastic is that it's still good after five years, at least, in the sun, here in the far north. But the southern sun is much hotter longer.
A rubber sheet will be applied to the plywood before the dividers are bolted in, excellent for assuring no leaks in the bolt holes under them. But what happens at the four inside corners where the rubber needs to fold up around the rim (top of the frame). The rubber there needs to be as tight as possible to the frame, with as little gap behind it as possible, for a water-pressure situation. For the water pressure will stretch the rubber toward the gap, and could over stretch it with bad consequences. You need to leave enough rubber at each corner so that, under pressure, the rubber will go up snug against the frame and the floor of the tank. Is this doable? I think so, yes. But there's yet another hazard, with less reliability, as the rubber goes over the rims at the four corners. There's going to be too much material there, and you will need to cut some away, creating, oh-no, a seam.
You can't put too-few steel bolts and washers against the top of the plexiglass lid if you want significant water pressure, unless the lid is super-thick. There will be a lot of upward pressure concentrated at the small washers, making the plexiglass susceptible to breakage there. For this problem, one can add more bolts, but to reduce the risk of leaks from bolt holes, it's better to use few bolts, per divider, over a continuous washer that can be, at the least, a flat strip of metal from end-to-end of each divider. Rubber washers go under the plexiglass.
There is a question on whether these strips of metal (say 2 inches wide) should be provided thin enough to bend a little along both their edges when the plexiglass bows (between dividers) from pressure and expansion. A little bowing (probably not visible) is good to alleviate the expansion problem at the bolt holes. In that case, allowing more bow between bolts along each divider seems like a good idea, yet the metal strips won't allow any bowing unless the metal stretches. But even if it's thin enough and stretchable without snapping, will it still act well as a washer material, or will it, to the contrary, allow too much pressure to act under the bolts? Is there a perfect balance in its width and thickness? Yes, but only the structural engineer knows for sure what the balance is.
In other words, it's risky to guess at the best sort of metal strip for allowing some bowing along the dividers. For all I know, it may not be an issue at all. You may opt to sacrifice that bowing altogether, to allow bowing only between dividers. In that case, you can have fewer bolts while using angle iron, or square tubing (more expensive, double drilling) if you want it to look nicer, in place of the flat metal strips. I have the sense that flat metal strips can be sufficient.
Nuts should be on the top in case you need to remove the plexiglass for any reason. Leaks at the bolts could be solved merely by tightening the bolts, but if the bolt head turns at the bottom of the tank while you're tightening the nut at the top...that's why you want to arrange that the bolt heads don't turn. Hefty lock-nut washers should do the trick.
Wrapping the dividers with black rubber seems like a great idea, as it has the bonus of reducing the risk of leaks at the bolts. Slow leaks are tolerable and even desireable for relieveing pressure.
If you're not using water pressure, there's no need to bolt the upper end of the plexiglass (maybe use one central screw to keep out the flies or guard against wind). The automatic shut-off valve (in the tank) won't allow the tank to be filled to the very top.
You will drain it before the first freeze, you won't forget, will you? Do you relish putting a drain hole between every two dividers? I didn't think so. So cut a small notch at the base of each divider to let water flow from one section to the other, and let the tank be slightly off level so that water drains toward one corner. That's where the drain hole goes. Put a hose to the drain hole, and run the hose into the attic, then over to a gable wall if you have one, to let the hose go through that wall.
An insulated storage tank in the attic right below the outer tank? Sounds good to me. There are probably electrical gadgets that open a water valve when water reaches a certain temperature. This drains heated water to the storage tank automatically. The outer tank then gets cold water automatically in case there's plenty of sun left in the day. Why waste that energy? Well, maybe you do want to waste solar energy if you already have too much water for that day. That tank is best drained each night because adding heated water to cooled, the next sunny day, makes little sense. But this is trib-endurance stuff; you probably won't do it now. You may not want to drain a lot of water each night to the same place, and so you may not want automatic valves giving you more water than you need.
Can't a T joint in the drain hose be used to manually fill this storage tank? Yes. When you decide. Just turn the valve by hand when you want water in the storage tank. Yup, that means that you either go up into the attic, or send your son up grumbling. Which is worse? Be the man, you go up. The smaller the drain hose, the longer you will stand there with the cobwebs. You can be sure that every spider up there is eyeing you simultaneously, maybe wishing they were big enough to attack. Sure, you can have the storage tank on the main floor, but then you can't have gravity-fed water to the taps. Be the man, get up there. Or, find a water valve that opens at the press of a button on the main floor, and shuts by the same gadget. I knew you'd like that, no matter if it costs a hundred bucks.
The Great Seal Hunt; Where the Rubber Meets the Plexiglas
I know, I know. You want water pressure in this tank, just a little, please. A little spit from the shower head could make the entire tribulation worthwhile.
Let's talk rubber type and rubber quality. Who in tarnation knows anything about rubber? What's there to know, you ask? Isn't it just that black stuff? I wish. What if the rubber you choose darkens the water after a period of no water use? What if the rubber stains the plexiglass because the latter loves rubber particles? Plastic has electrical properties that loves dirt, it seems. Wouldn't you just love to take all those bolts out to wipe the plexi? Should we use an imitation rubber or rubber alternative? Is that some other accident waiting to happen? What salesman can we trust anymore to care for what we care for?
The rubber website below once said, "Jesus is the way, the Truth and the Life." It must lose a lot of business by making that courageous claim. Each webpage has links to various other rubber and non-rubber products in rolls or sheets. See, for example, the SBR rubber page for a low-priced material. The recycled rubber rolls/sheets are very low cost and have operating temperatures of 150 F. If these temperatures are too borderline, enquire on nitrile rubber (good elasticity) with a service temperature of over 200 F. EPDM is rated 280 F, at a good price and including good elasticity.
Rubber is nice because it can act as its own washers at the bolt holes. I like. Here's a bit from an SBR page: "As with most higher tensile strength SBR rubber materials, the higher the tensile strength, the lower the elongation (or stretch)." "Tensile strength" refers to resistance from being pulled apart in opposite directions. We don't care about that. If elasticity is important for getting a nice cushion at the bolt holes, then that's more of what we want. Too much cushion may cause sharp edges to penetrate the rubber.
How do you plan on sealing the outer frame/perimeter under the plexiglass? A homemade gasket. So, we want appropriate gasket material: "This [black type] is a higher strength SBR Rubber material that is durable...SBR rubber in specification grade material is used in conveyor belts, chute liners and low pressure gaskets." We're probably on the low-pressure side. "Low pressure gasket" may be for metal parts unlike our wood and plastic combination. In other words, this SBR material (temperature rating is -20 to 180 F) may not be the best choice if it's too hard. Hard is okay gasket material with precision-straight metal parts, but our wood frame will not be precision. And we don't want a bolt every two inches in the plexiglass.
Most pages above do not offer rolls in 4' widths, meaning that if your tank is 4' x 8', you may need to piece together the tank's floor...meaning that you need to seal under each divider. I would suggest that each piece of rubber extend beyond a divider by an inch on both sides of it, and give it the appropriate glue. The only rolls I've seen in 4' widths was the general-purpose rubber page. The page (below) on large sheets of general purpose starts at 10' x 5' sheets (under $76 for 1/16" thick in January 2013), and shows a temperature rating of -40 to 225 F. It sounds good. If we decide not to use that rubber, then consider that the black SBR comes in 36" x 48" sheets, good where your dividers are spaced 18 inches or less.
See rubber flooring if that can help. The page below has some floor mats 4' wide and up to 60 feet long, some of it in 100%vinyl. Can we use vinyl? At their marbleized floor-runner page (100% rubber), their 4-footers are about $6 per square foot. They offer samples so that the texture on both sides can be checked out.
The best way to apply rubber is to let it cover the floor, then bend up the frames, then bend again over the lid to act as the gasket. If it bends again to cover the outer side of the wood frame, bonus. A few roofing nails to the bottom side. You know it'll be tricky at the corners, but your utility knife and your cool hand will get that just perfect...hopefully the first time.
Vinyl is cheaper. At the page below, Chinese-made black vinyl at 1.5 meters (almost 5') wide goes for under $20 per running meter. It's suggested for car bodies meaning that it can take heat. The thickness is 19c, but I don't know what that is:
...Service life: for outdoor 6 years, indoor 12 years...3D carbon fiber vinyl film is not just a cheap imitation vinyl, it is a high quality...3D carbon fiber vinyl film is also water resistant and UV stable, so applying it to exterior surfaces such as a hood, spoiler, or the whole car is not an issue.
The product is an adhesive back one, wherefore the service life above may be referring to the adhesive, not the waterproof performance of the material. I'm assuming it's fully waterproof, but rain is not the same as submerged. Yet, we don't need the glue. Ask.
None of the pages speak on rubber leaching into water, however, which can either mean that they all do to a point but no one wants to tell, or that none do as it's a non-issue. Regardless of your choice, the plywood needs 100% waterproof material ongoing if you want to start using the tank now, and still have it in case the trib arrives.
Also, you can check out Wikipedia's article on synthetic rubbers for a list near the article's end. There's a link there for neoprene:
Neoprene rubber is an all purpose elastomer and an extremely versatile synthetic rubber used in thousands of applications. Neoprene rubber resists degradation from sun, ozone and weather...Neoprene rubber maintains its strength, flexing, twisting and elongation very well over a wide temperature range while having outstanding physical toughness...
Email us your drawings if you need rubber gaskets or rubber parts made.
There is also a page for rubber gaskets.
The most likely place to have a leak is where a gasket over a divider meets the gasket at the frames. How will you keep water from getting through the joint between those gaskets? That's why you need to cut the two pieces of rubber with precision, and why the rubber needs to compress to some degree, so that the edges of both squeeze toward one another hard. To cut them to precision, overlap the one gasket over the other, and make one slice through both with a straight, metal edge (such as a carpenter's square). You will be amazed at how nicely you get that fit, providing you have a firm hand that slices like a pro, not like a butcher going over and over the slice with choppiness. But even if you do a lousy job, the rubber's squeezability can save you.
Why is There's No Such Thing as Inexpensive Plexiglas?
It's perhaps best to make the tank top a multi-piece one rather than a full plexiglass sheet due to its expansion and contraction under tight bolts. Then again, if the full sheet is the final choice after much deliberation with all the wise men of the community, one may be fortunate where it simply bows a little between dividers, without cracking. I would rather use the entire sheet...in which case it would need to be drilled out while simultaneously putting the bolt holes into the 2 x 4 dividers. How else will you line up the holes in the plexiglass with the holes in the wood? One wrong shift, however small, without your knowing about it until you find that the bolts won't go in, could send you nutso. Thin plexiglass is not exactly a helpful partner when taking a drill, unless you get a special drill made for it. Thin plexiglass is prone to cracking, but this may not apply to sheets nearing 1/2" thick...which will give you lots of water pressure, providing the other parts of the tank hold the pressure.
If we don't drill the plexiglass at all, it's because we're using small plexiglass sections instead, with joints over dividers. True, we would then increase the locations for potential leakage. Having plexiglass joints above the dividers suggests that dividers should maybe be more than 1 1/2" thick. It's borderline. There are plastic-lumber lengths available in the dimensions of 2.5" from American Plastic Lumber. The edges of two plexiglass sections should be separated by a 1/2", allowing bolts to pass between them without need of drilling the plexiglass. I knew you'd like that.
How hot will the metal get that presses on the plexiglass? Will the plexiglass start to go soft? Whatever plexiglass we purchase, we should tell the salesperson what we're using it for, in full sun against metal. Or, put some cooler material between the metal and plexiglass, like some cheap and firm (and easily drilled) PVC plastic.
Another idea is cork-gasket strips. "Temperature and age stability: cork retains its properties at both high and low extremes of temperature and usually lasts 20 years without deterioration." Another webpage gives it ten years for wine bottles, while others flat out say that cork resists deterioration in water:
...cork is practically impermeable to liquids and gases. Its resistance to moisture enables it to age without deteriorating.
Compression, elasticity, resilience - cork can be compressed to around half its thickness without losing any flexibility, and recovers its shape and volume as soon as it is released. It is the only solid which, when compressed on one side, does not increase in volume on another; and as a result of its elasticity it is able to adapt, for example, to variations in temperature and pressure without suffering alterations. When it is compressed, the air inside the cell is squeezed to a smaller space...
It sounds great, but maybe it's a desperate sales pitch. Lowe's sells thinner cork sheets that can be cut into strips (I'm assuming cork is cork is cork). Here is a New York webpage with some cork sheets for cutting into strips:
The ultimate question may be how cork performs or deteriorates when saturated with water. There is "cork rubber" that you may want to look into to. The cork-rubber webpage below (California-based location with several locations worldwide) has a plexiglass link showing you pricing information.
Another gasket idea is a powder of some sort to be mixed with waterproof glue to form a paste. The idea here is to eliminate seams in the gasket (in case we can't get a seal using seams). For example, use cork powder or cork granules. With granules, it's not so much a paste as it's a mass of sticky granules. There are two options after applying the cork-paste to the frame: 1) tighten the plexiglass very loosely until paste is about 1/8" thick, or maybe 3/16", and let it harden, then tighten down hard after the glue has cured sufficiently; or, 2) tighten the plexiglass down hard from the start and let paste cure under pressure. With the last option, there could be some cork material falling into the tank that could plug your drain hole, for example. It would be an awful way to tear your hair out. Do some cork-and-glue tests before any attempts at the tank to assure success. If you chose option 1), give all bolts the same number of turns.
Trying to set a full sheet neatly on the wet cork-paste of the entire frame could be a challenge. The glue mixed in will need to dry slowly so that one can apply the cork-paste to every rim section without any of it becoming too dry. Waterproof carpenter glue might not give that much time, especially in the heat. And dried carpenter glue is rigid, isn't it, not compressible. What happens to it when mixed with cork and then compressed as per option 1)? Are the seams between glue and cork compromised? Does the finished product go brittle like a cheap tortilla? Do a test. wood glue would adhere cork granules well and meanwhile will not, likely, stick well to plexiglass should we need to open the tank, or find another gasket method. But carpenter glue to the wood frame would make it hard to remover the cork, if that's the decision. A belt sander should do the trick.
I'm reading online: "I've seen wine cork instructables using wood glue, Gorilla (R) glue and rubber cement. I haven't worked with the first two." Another page: "Cut cork with scissors to fit size and shape of plywood. 2. Apply wood glue to the back of the cork and spread evenly..."
What about silicone as the glue? Fantastic, but it will stick mightily to the plexiglass lid. That's no good for opening the tank if need be. What about using cork strips throughout, aside from a little cork-paste only at the seams (where the strips would meet)? That sound easier and reliable. The paste is pressed while soft until the plexiglass is tight against the cork strips. But what if the small spots of paste shrink a little while drying out? I suppose that the lid can be loosened / removed with a little more paste added; then re-tighten the lid.
The thicker the plexiglass, the better it can handle the weight of accumulated snow without transferring pressure to the joints. For thin plexiglass, one should cover the tank with plywood in heavy-snow or long-winter regions. I've been praying that plexiglass manufactures be jailed for thievery. But if it's not the fault of manufacturers, then jail the distributors. Somebody deserves to be put away. Al Gore must have stock in glass and plexiglass.
To keep your solar heater hotter, make a thin-plexiglass cover for it that traps hot air in. Otherwise, water will lose heat when wind blows, and when the sun don't shine past clouds. Heavy snow on such a vulnerable cover would be an issue to resolve.
Beware the Expansion Phantom
If we don't bolt plexiglass to the frame, plexiglass expansion versus the rate of plastic-lumber expansion wouldn't matter near as much. The thermal-expansion coefficient of plastic lumber, according to a data sheet from American Plastic Lumber, is 0.000033 inches per inch per degree. At the data sheet below, cast acrylic has a value of .00005 inch / inch / degree, meaning that it expands larger than the plastic wood (see other plastic wood below with .000055).
If your plexiglass is 48" wide and 96" long, you need to multiply .000055 by 48 and 96 respectively, and then multiply each result by about 50 (degrees) to find the maximum expansion at 120 F if you bolt the plexiglass to the frame when the temperature is 70 F. The 96" length will grow by about .26" over that temperature difference, or an 1/8" on either side. It's not that much, but the problem may be that wood is said to expand a lot less. I'm reading online that the thermal expansion coefficient of wood is one to two millionths of an inch per degree. That is: .0000015. Hopefully, the expansion of the plexiglass will go around the bolts and not crack at the bolts. It's your call.
The thermal coefficient for steel is .00000645, about 8 times less than plexiglass, yet helpful as compared to no expansion. As the expansion distance for an eight-foot piece of steel (at 120 F) is therefore in the order of .03", the plastic is going to expand by about .2" more than the metal, suggesting that the bolt holes through the plastic lumber should at least be 3/16" larger than the bolt diameter.
I have this from online: "...the coefficient of thermal expansion of rubber is ten times that of steel." Is it an issue? I don't think so if the rubber stretches to begin with.
I want to say that these industry numbers may, or may not, come with safety factors built in, which is to say that the expansion may not be as much in reality. It gives me the jitters, anyway, and might cause me to forgo the single large sheet for smaller pieces. With an unsealed tank (not under water pressure), I would definitely go with a large sheet because it only needs to be fastened at the center, allowing it to buckle a little during expansion...though this doesn't answer for contraction. Loose bolts all around is advisable for that tank. I'd say to build it in the shade on a day at about 80 degrees, or midway between the coldest and hottest annual temperatures at the roof.