Wednesday, December 30, 2009

Solar Hot Water Part V: Insulation

A much underappreciated  part of making a solar hot water system efficient is properly insulating it to reduce parasitic losses. There are two aspects to this:

  1. Properly insulating the pipes carrying the hot water and the heat transfer fluid from the solar collector.
  2. Ensuring that the solar tank and any backup storage (the gas fired tank in my case) are sufficiently insulated to reduce loss during water storage.

The contractor did a fine job on the pipes carrying the heat transfer fluid outside leading from the solar collector to the inside of the house. This is a particularly crucial area, because if the air temperature is low during the day the losses in this section can be really high. But for the pipes on the inside, if you take a look at my previous post on installation (which you'll find here), you can see that the contractor sort of insulated the pipes on the inside. The black rubber you see around the pipes in some of the pictures is pipe insulation. The contractor did an OK job with most of the pipes except around the gas fired hot water heater. The pressure relief valve pipe and the hot water pipe leading into the house were not insulated and radiating heat. Also, at places where the pipes had a bend, the black rubber pipe insulation stopped and left small areas of copper exposed that could radiate heat. Getting pipe insulation right takes a lot of time and effort, something most contractors don't want to do (otherwise they would have to charge you for it).

So I spent some time doing the pipe insulation myself. I installed the black rubber insulation around all the pipes which the contractor forgot, and sprayed some high temperature foam insulation around the pipe corners, and around the connections between the pumping station and the feed-in lines. Here's a picture of how the pipes look in the ground floor gas-fired hot water heater closet:

 The blobs of pink matter you see hanging down from the pipes where a corner is are the high temperature foam. Note that this is not the standard polyurethane foam that you can get for insulation at the hardware store, this is foam which is rated for installing, for example, between floors in apartment buildings to help retard the progress of a fire. It won't stop a fire, since it is organic and will ultimately burn, but it is rated to a much higher temperature than the standard, white polyurethane foam that's used for sealing around doors and windows. It's available at Home Depot, and the important distinguishing factor is, as you can see, that it is pink. I used high temperature foam because the temperatures around the pump station (which you can't see here) and around the furnace flue which leads up through the closet can get pretty high, maybe around 180F, which might cause the normal foam to get soft, if not melt. Finally, the white blobs you see are tags indicating what each of the pipes do (sorry, the picture  is a  little out of focus because the shutter was open for a long time and my camera doesn't have image stabilization).

The second aspect is the tank. The Superstor comes with 1.5" of closed cell foam insulation, and they claim that it loses around 0.5F per hour (as we will see in a later post, this is optimistic at best). To my mind, that is simply not enough insulation. My Hot Springs hot tub has around 6" of closed cell foam, and it loses maybe 2F overnight if I turn the temperature down to 80F after having it up around 102F, and that is outside where the ambient temperature in winter is a lot lower than inside the house. I think the reason they don't add more insulation to the Superstor is that it would make the tank's diameter bigger, and, of course it would make the tank even more expensive (the Superstor is pretty expensive as is).

My first plan was to spray the tank with closed cell foam before the contractor installed it, but a quick check with the city building inspector indicated that they would probably not approve the project if the tank was coated with closed cell foam. I then came up with the idea to take radiant barrier insulation made from aluminum foil and plastic bubble wrap and make a form the size of the tank, then spray it with closed cell foam to make a kind of jacket. So I tried that (sorry, no pictures) but the result was too rigid once it had dried and I couldn't get it into the narrow space between the tank and the wall.

Since I had some radiant barrier left, I decided to wrap both tanks in radiant barrier insulation as a start. I did this work in summer right after they installed the system, and the solar tank was radiating heat strongly, a bad sign. I did the wrapping in two phases, first the left side then the more complicated right side around all the connections. Here's a picture of what the Superstor looks like when it is completely wrapped in radiant barrier insulation:

One issue was what kind of tape to use. Standard duct tape probably won't work because the glue will melt from the high temperatures in summer. I first tried an aluminum tape from 3M that is specifically for furnace flues and other high temperature applications. It wasn't ideal, it was stiff and didn't adhere unless you really pressed it, also, it had sharp edges, but it sort of worked, although some places that weren't pressed hard enough came loose. I then found a similar tape from Nashua which sticks a lot better and doesn't require strong pressing to bind.

After I finished the radiant barrier job, I discovered that a new line of high temperature duct tape had been introduced on the market. I used some that on the R-13 fiberglass blanket which I wrapped around both tanks on top of the radiant barrier. You can see that on the bottom of the picture above of the gas fired tank. Here's what the Superstor looks like nicely clothed in its new fiberglass blanket:

The gray stripes around the tank are where I taped together the R-13 batts.

One question that might come up is: why not used one of the premade vinyl insulation jackets that are commercially available? I thought about this, but there are a number of problems. First, I couldn't get a jacket big enough for the Superstor. It is 60" tall, much larger than a standard 40 gallon gas or electric domestic hot water heater. Second, they only provide something like R-7 of insulation. I now have R-13 with maybe R-7 or so of radiant barrier. Finally, vinyl is not a good material to have around the house, especially around hot areas where chemicals can outgas. I looked around for a jacket that didn't have a vinyl skin but could not find any.

In the next post, I'll talk a bit about how the system is performing, and provide some data about how well my insulating treatments have worked.

Wednesday, December 23, 2009

Solar Hot Water IV: The Installation Experience

Chris and Justin from Sunwater Solar came by and had a look at our house. We decided to put the collectors on the east side since the west side is occupied by the solar PV. The east side gets sun primarily in the morning, but since the solar PV is higher value in the afternoon (about 3x the tariff other times), having it on the west side is a better deal. So it made sense to keep the remaining space on the west side, about enough for another 3 PV panels, for solar PV if we decide to add some in the future.

Then came the question of where to put the tank. We initially considered putting it in a small house next to the sauna, or maybe next to the fireplace pod that sticks out from the house, but ultimately we decided to repurpose the cold air return for the forced air heating. We had the forced air heating taken out a couple years ago and put in hydronic floor heating, so we didn't need the space any more. It is on the second floor in the hallway. The space consisted of a plenium with a large grate on the front, which led into the downstairs furnace room. When we had the hydronic furnace installed in the furnace room, the plenium was blocked off from downstairs. As we discovered when we opened up the space and removed the plenium, the flue from the furnace and the gas fired hot water heater in the downstairs mechanical room ran up the middle. We had a bit of work to do on the space.

Point to note when planning a solar hot water system: be sure to include enough time and money for upgrading the space where you plan to put the tank and the pump station if you are not going to use your existing tank.

Sunwater doesn't do the kind of drywall and carpentry work needed to modify the space, so we asked a friend of ours who is a contractor, Tim Hmelar (check out his new business, Purple Coupon) to do it. Tim is the most reliable contractor I know, his work always comes in on time and at or under budget. Together with Tim's brother Frank, who sometimes helps him on jobs, we decided to put a door in the side of the space and a hatch where the grate for the cold air return was located. The pump station then could be installed in the upper space on the inside of the hatch, with the tank below it. Just under the hatchway, we decided to put a platform on which someone could sit while servicing the pump station.

Because the flue ran up the middle of the space, Tim and Frank had to move it to the back. You can see the new location before the platform was installed:

The black pipe is the vent from the bathroom, which is on the wall immediately to the right of the flue.

Here's a view of the bottom where a plywood floor was installed to support the tank:

One issue we ran up against was what to put on the inside walls. By code, if there was any loose wiring, we needed to have some kind of protection in place to keep someone from accidentally grabbing the wires. But the amount of space in the closet was pretty tight, especially since I wanted to wrap the tanks in additional insulation. The normal thickness of drywall is 1/2" but it turns out you can also get 1/4" drywall, which is what we ended up using. This gave us another 1/2" of space. In addition to that, Tim put 6" of closed cell foam insulation on the roof and at least 3" on the walls of the closet, to reduce the amount of heat transfer into the house in the summer, when the tank can get up to 175F.

The downstairs gas heater tank looks pretty normal as you can see from this picture:

The insulating blanket isn't doing much good since it is not tight against the tank, and, by the way, the outer skin is vinyl, which is quite unhealthy. But we were were going to change all that.

The guy Sunwater sent to do the installation was an older tradesman, quite pleasant to talk with, but I began to have some suspicions about his competence almost immediately. One day I came home from work after the tank had been installed, and found that he had connected the feed lines from the collector to the top heat exchanger instead of the bottom. Now, heat rises, so this configuration would have caused the tank to stratify with the incoming cold water line injecting cold water on the bottom and the hot water on the top, where the water outlet for the domestic hot water is. I called him up and made him change the plumbing.

In the picture below, you can see the bottom of the fully installed Superstor tank, where most of the interesting stuff is happening (there's some on the top too but not as much):

The two lines in the back with the black foam around them are the feed lines from the collector, going through the correct lower heat exchanger. The corrugated line is the pressure overflow from the tank, the line it is connected to is the pressure overflow coming down from the expansion tank and pump station in the upper chamber above the platform. The line entering the tank is the cold water input from the house and ultimately the main. The other line in the elbow on the right side is the hot water running to the gas hot water tank downstairs, you can't see the pipe running down from the top of the tank where the hot water outlet is because it is behind the door frame. There is also another pipe running across behind the cold water input, connected to the pan, that's also an overflow pipe to the drain. In the back is the furnace and gas hot water heater flue.

Below you can see the internals of the pump station in the upper chamber before the cover was put on:

The brass colored cylinder is the pump, three gauges on the front provide information about the temperature (to and from collector) and pressure in the feedlines. The feedlines coming from the collector through the outer wall of the house are on the top. The blue wire is the power cord (110v, 15 amp, we also had Tim install a plug for it). Here's what the pump station looks like when the cover is on:

And here you can see the white expansion tank:

The grey wire in the background is the sensor cable from the collector, the corrugated line is the pressure overflow from the expansion tank to the pipe going to the drain downstairs, and of course the ubiquitous flue.

Since the solar hot water acts as a preheater for the gas heated tank downstairs, it needed to be plumbed into the gas tank. Here's a picture of what that now looks like:

The grey knob you see in the middle is the mixing valve. It is where the hot water from the tank is mixed with cold water to keep the domestic hot water below 120F. Above that temperature, scalding is possible. As for the rest, the grey steel line on the left is the pressure overflow, the horizontal  line with the red lever on the  bottom is the hot line from upstairs, feeding into where the cold line would normally go (for preheating), the horizontal line on the top is the cold going to the tank upstairs. The grey knob is where the domestic hot water comes out of the gas tank. In the back on the right is a feed line for the hydronic boiler, from the cold line, and a cold line that goes into the tank, with the red lever now in the off position. Switching the front red lever to off and the back to on causes the solar tank to be cut out of the loop. Unfortunately, there is no way to switch the gas tank out of the loop (more on this in a later post).

Charging the collector feedlines with the heat transfer fluid requires a very specific pressure adjustment to match the height difference between the collector and the expansion tank, and also any air in the feedlines must be purged or the bubbles will cause a lock and prevent pumping. After the installer charged it up the first time, the tank didn't seem to be getting any heat. It turned out that the pump wasn't on because he had forgotten to turn on the power switch! He came and fixed that, but then the pumps were really loud and there still didn't seem to be any heat getting into the tank. So he came back again: he had forgotten to bleed the air out and so the lines were locked. When I checked the pressure after he finished, it turned out he had way overpressurized the lines, as if there were 10 vertical feet or so between the collector and the expansion tank instead of the 2-3 feet there is. By this time I had had enough of this guy. When I discovered that the unions on the tank and in the pumping station were leaking hot glycol (and copiously at that), I called Justin and told him I wanted somebody who knew what they were doing put on the project.

Justin was very accommodating and worked with me to get the problems fixed (also, I didn't send him his last payment until he did, just to make sure). But it took another month before we could get the unions properly fixed. Justin sent one of his best guys over, and he first re-soldered the unions on the tank. But he couldn't get the pump station unions tight. It turns out they have a kind of pressure coupling which seals when tightened, but when overtightened, leaks. So he had to send his very best guy to come and cut through the pressure couplings. One slip and they would have had to reinstall the entire feedline set from the collector again (or, at least, a substantial part of it)! But thankfully he managed to cut just precisely right and they reinstalled the pressure couplings. Both sets of unions now looked solid.

Despite the fact that we had such problems with the installation, I don't hold it against Justin, and I would still recommend Sunwater. As it turns out, the installer had only been working for Sunwater for a few months, whereas the other two guys had been working with Justin for years. I've been in this situation myself: you hire someone with a great resume and they turn out to be a dud (or, even worse, they falsified the resume and don't know the basics of the job). So I didn't hold it against him. Justin really tried to get the system properly fixed, and ultimately did fix it to my satisfaction. However, this does point up a problem with the whole area of residential green energy and energy efficiency remodeling: many contractors and/or the people working for them don't have a clue. So they often do things, especially when it comes to insulation or weatherizing, that are ineffective or, in this case, downright harmful.

Point to note when planning a solar hot water system:be sure you understand how the system is supposed to work and what is needed in terms of insulation to make it efficient, and make sure the contractor does the work necessary to ensure it works properly.

Note to President Obama: "Cash for Caukers" won't work if the caukers don't know how to wield a cauking gun so they can properly insulate. Most people in construction in the US are in the business because they can't do any other work. Unlike Europe, where trades require a substantial amount of technical training, and where the trades have over the last 10 years upgraded their members' training through in service education to address energy conservation and green energy equipment installation, the assumption in the US seems to be that anybody can do this work without any training. If this kind of attitude goes into "Cash for Caukers" the resulting weatherization is not going to save energy, it will be a colossal waste of money. I'd recommend to the Congress if they want to do "Cash for Caukers" that they should budget enough that any contractor who plans to participate must take a 6 month course in energy efficiency building so they understand what they are doing. And - just to be on the safe side and avoid fraud - have people trained to inspect the job periodically the way building inspectors do, since most inspectors don't have the training to spot improperly installed green energy devices or ineffective insulation either.

Sunday, December 20, 2009

Solar Hot Water Part III: Our System Architecture

As mentioned in the previous post, the system architecture we decided on was to have the cold water from the city main feed into the solar tank, preheat with solar, and then feed into the gas-fired hot water tank. In addition, we decided to get a solar tank with a second heat exchanger coil. This would allow an external boiler, for example a geothermal heat pump or efficient closed combustion gas boiler, to be substituted for the gas-fired tank in the future if we wanted.

Here is a brief description of the system we had installed:

Collector: Schueco Slim Line II-80
Tank: Superstor SB 80 gallon w. backup heat exchanger
Balance of system: Pumping station PS 1.3 from Schueco, expansion tank, DeltaSol BS thermostatic control, etc. was part of the Schueco package

Normally, the Schueco system comes with a double walled steel tank from Rheem, but we dropped that and ordered the Superstor instead.

We decided on Schueco instead of another collector such as Heliodyne because the Schueco collectors are designed to force the heat transfer fluid out of the collector and into the expansion tank when the collector stagnates and begins to overheat. The water in the heat transfer fluid flashes to steam increasing the pressure and forcing the fluid into the expansion tank. This keeps the propylene glycol in the fluid from cooking, increasing the lifetime of the fluid. I've heard anecdotally that Schueco has tested their collectors for up to ten years without needing to change the fluid.

The diagram below shows the system in schematic form (note: the diagram is edited from the Schueco installation manual and is copyright Schueco, used here under the Fair Use exemption):

 The diagram shows the hot water coming into the solar tank, being heated by the solar heat exchanger and flowing from there into the gas hot water tank. The gas hot water tank heats it further if necessary then the water goes into the house. There is a temperature control valve on the house side of the gas hot water tank to keep domestic hot water temperature below 120F, since hotter temperatures are dangerous and can scald. The hot and cold heat transfer fluid flow to and from the collector is shown above the pump station.

The numbered items on the diagram are the following:
  1. The Schueco Slim Line II-80 collectors (2).
  2. The expansion tank and overflow hose. If the pressure gets too great and while filling the system, glycol can run out into the drain through the overflow hose.
  3. The Schueco PS 1.3 pump station.
  4. The Superstor tank. The bottom heat exchanger is connected to the pump station, the top exchanger is not connected to anything at this point.
  5. The DataSol BS thermostatic control. It compares the collector temperature at T1 and the tank temperature at T4 and only turns the pump on if the collector temperature is hotter than the tank, and turns the pump off if the tank temperature exceeds 180F.
  6. The existing hot water backup tank.
There are also two pressure release pipes coming out of the solar tank and the gas hot water tank that are not shown on the diagram, these are standard for hot water systems to avoid having the tank overpressurize  and rupture.

Friday, December 18, 2009

Spark Plugs

Last year I saw some information on the Internet about a new kind of spark plug that supposedly increased gas mileage, and also increased torque and power. Much as I love it, our 2002 Prius is no speed daemon, especially when you are trying to merge onto the freeway. In fact, I would say it is somewhat underpowered. The plugs are called Pulstar from Enerplus. They are pricey: $25 compared to maybe $1.50 for normal plugs. Being an early adopter at heart, I thought I would give them a try, so I ordered a set of four.

The plugs worked great initially. Torque and power was much better, and the gas mileage rose by another 2 mpg. I was really happy. Then, around 6 months later, the gas mileage started slowly dropping - from 46.6 to 44.3. Since that was the transition from winter to summer, I took it to be the normal Prius variation in mileage with temperature. Anybody who has a Prius notices this - you get much better mileage in summer than winter because the car warms up sooner and so the engine shuts off sooner.

On Wed., I was driving downtown when the engine suddenly began to chug. I thought it was an earthquake or something, Toyotas are usually reliable. Then the check engine light began flashing - whooo! whoo! red alert! That got me concerned. Fortunately, I was only a couple of blocks from my mechanic, so I slowly nursed the car to the garage and left it there for repair the next day.

The mechanic called me around noon the next day with the news. Apparently, the plugs had fouled and shorted out the ignition coil. He was really surprised, since the Prius comes with iridium plugs that supposedly last for 120,000 mi. I had the Pulstar plugs in for less than a year.

I asked him to give me the plugs, which he did, then today I called up Enerpulse and asked them to make good on their warranty, which is (quoting directly from their Web page):

We offer a 4-point guarantee on Pulstar™ Pulse Plugs. This means:
1) You can return the Pulstar™ pulse plugs for any reason within 30 days, no questions asked.
2) Pulstar™ pulse plugs will not harm your engine.

3) Pulstar™ pulse plugs will last 50,000 miles.

4) Pulstar™ pulse plugs will not void your warranty.

Near as I can tell, this incident violated at least 2) and possibly 3) if the plug that fouled shorted itself out too.

Enerpulse told me that I had to send the plugs and the ignition coil back to them, along with the bill from the mechanic and a letter explaining the situation. Fair enough. I guess they need to have the evidence before they are willing to cough up any money, since people do cheat.

I called the mechanic up again and asked if he had the coil. He did, so I picked it up today.

We will see whether they are as good as their word. Stay tuned.

Monday, December 14, 2009

Solar Hot Water Part II: Background on Our System

I've been researching solar hot water systems for a number of years, waiting until the state of California instituted a subsidy. Unfortunately, a trial program in San Diego only netted double digit installations, indicating a general lack of interest in solar hot water (more on this in a later post). That and the sorry state of California's finances caused the state to indefinitely postpone any subsidy for solar hot water. For that, this year the Feds put in place a really nice 30% tax rebate (not a deduction, a rebate) on alternative energy. The rebate includes the cost of everything involved in installation, not just the equipment. So this spring we decided to take the plunge, and I began looking around for solar hot water installers.

One problem is that there are very few companies, even here in northern California, that specialize in solar hot water, in contrast to the case with solar PV electric installers which are really common. Most companies that do solar hot water include it as a sideline in addition to their mainstream solar PV business. The consequence of this is that there are few companies that have a lot of experience in installing solar hot water. As in any field, more experience leads to a more reliable outcome. The reason very few companies specialize in solar hot water is that very few people want it, and also, it is more difficult to install because it involves plumbing and tank installation. The amount of site specific customization that needs to be done for solar hot water is therefore much more than for solar PV, which leads to more costs and more possibility for errors.

After checking out another company that does solar hot water installations as a sideline to their main solar PV business, we decided on Sunwater Solar in Richmond. Sunwater only does solar hot water systems and only indirect systems.

Another question was what kind of system to go with (direct or indirect, thermosyphon or ICS or drainback or indirect). A friend had a thermosyphon-like system, which uses a non-water based heat transfer fluid, installed a couple of years ago for around $7K, but it involved putting a 40 gallon tank up on the roof. I favored an indirect system over a  thermosyphon system to reduce the potential aesthetic problems (looks less like a science project) and because of the lower efficiency in a thermosyphon system, and over a direct system because of the potential for freeze problems in a direct system (temperatures here in coastal northern California were around 24F at night a few days ago, and that's also the temperature my collector was at). Some research on the Internet indicated that flat plate collectors perform better in our climate, due to the overheating problem in summer. For that, they take up more roof space but all in all, flat plate seemed like the way to go. So we decided to go for an indirect system with flat plate collectors.

We also needed to decide what we were going to do for solar water storage. The cheapest alternative would have been to use our existing gas hot water tank for solar storage and install a heat exchanger between the cold water inlet and the tank. Silicon Solar makes such a heat exchange. The problem with this kind of system is that it is much less efficient than if the heat exchanger were in the tank itself because heat is lost to the air. In addition, the tank we have is 40 gallons which means we would be able to store less hot water for cloudy days. So that suggested instead installing a second tank for solar hot water storage.

I researched storage tanks a bit and found out that there are basically three types: glass (actually ceramic) lined, steel, and stainless steel. Glass lined tanks are much the same as the cheap electric and gas hot water tanks that one can find at Home Depot and elsewhere, and that most homes in the US have. Both glass lined tanks and steel tanks require a magnesium anode to prevent corrosion. A little known (at least, I didn't know it) fact about the cheap electric and gas hot water tanks is that they, too, need a magnesium anode. Over time, the anode dissolves until, after 5-10 years (depending on the ionic content of your water), the anode disappears. When it disappears, it needs to be replaced, something most homeowners probably never do. If it isn't replaced, the tank corrodes and starts leaking. The upshot of this was that I decided to go with a stainless steel tank, to reduce maintenance. Unfortunately, this added substantially to the cost of the system.

The final decision we needed to make about our system architecture was what we would use for backup. In the short term, it seemed to make sense to use our existing 40 gallon gas-fired hot water tank as backup, since it had originally been installed around 2001 and was therefore relatively new. In addition, I had vague plans to do some deeper energy remodeling in the future, and wanted to put off the ultimate decision about what to do for backup until then. I briefly considered getting a solar tank with an integrated gas heater, but rejected that because the cost was higher, and because I was hoping at some point to reduce our gas usage to a minimum. So we decided to keep the gas fired tank for now and use the solar tank as a preheater, at least in winter. In summer, we could turn off the gas tank (it has a pilot light, very wasteful) because there is enough sun and the tank gets hot enough to last for several days without sun, though we rarely get many cloudy days in summer.

Saturday, December 12, 2009

Solar Hot Water Part I - Fundamentals

So this blog begins a series of posts about solar hot water systems. Solar hot water is one of my favorite alternative energy systems because it is (theoretically at least) relatively inexpensive and it seems quite natural - put water out in the hot sun and let it heat up. Yet, very few people have installed solar hot water, even though such systems were quite common in California prior to WWII. In this series, I will go through the basics of solar hot water systems and describe the system we had installed this summer. I'll also talk about financials and some of the pitfalls involved in our installation experience.

This first post is about the basics of solar hot water systems.

In principle, nothing could be simpler. Unlike solar PV systems for generating electricity, solar hot water systems don't depend on any fancy quantum physics, they just depend on good plumbing. The idea is to put something on your roof (or somewhere else that gets a lot of sun) which traps heat from the sun, converting light into heat by absorption on a black surface, then transfers that heat to a fluid. That fluid forms hot water for domestic use. Solar hot water heating can also be used for space heating in winter, by running the hot water through hydronic heating loop pipes embedded in the floor.

There are two basic kinds of solar hot water systems: direct and indirect. In direct systems, the fluid heated in the thing on the roof, the collector, is the actual water that you will use for your domestic hot water. In indirect systems, the fluid is typically propylene glycol, a nontoxic fluid that is the basic component of antifreeze in automobiles (antifreeze also has toxic components too for rust inhibition and other purposes but those are not in the fluid used for solar hot water systems). Propylene glycol is used because the range of temperatures over which it remains fluid is much higher than for water. For example, a 60%  propylene  glycol/water solution freezes at -55F rather than 32F. Also, the same solution of propylene glycol/water doesn't boil until around 225F instead of 212F. This means that the pipes and collector are protected both  against freezing temperatures at night in winter and boiling hot summer days when your hot water tank already has enough heat and the system shuts itself off. The water or propylene glycol  in the pipe loop between the collector and your hot water tank is called the heat transfer fluid because it transfers the heat from the hot collector to the water in the tank (either directly if water or indirectly - through a heat exchanger coil in the tank - if propylene glycol).

The cheapest way to obtain a direct solar hot water system is to build it and install it yourself. Take a used gas or electric hot water heater tank, paint it black, put it into a box with a glass top, and plumb it into the domestic hot water loop in your house. Another variation is a do-it-yourself collector made out of a glass box and copper tubing painted black with the hot water from the collector running down from the roof through a box set in the ground filled with rocks.  The rocks act to collect heat while the sun is shining, and radiate heat into the water loop when it's not. There are many sites on the Internet where you can find plans for systems you can build for under $1000. These systems are not very efficient but they certainly are cheap. I'll not recommend any here, because I've not tried them, but if you are handy with tools, don't have a lot of money, and don't particularly care about performance, you might want to give it a try.

If you have more money to spend, there are three kinds of commercially available direct systems - thermosyphon systems, integrated collector storage systems, and drain back or drain down systems. Thermosyphon systems have no active pumping or other means of forcing the water to circulate between the collector and the hot water tank. Instead, thermosyphon systems depend on the tendency of hot water to rise off the collector. They have a tank attached to the collector usually around 40 gallons where the hot water collects. You can see the tank at the top of the collector in the picture below:

Cold water enters the collector at the bottom and rises up as it is heated. The hot water is collected in the tank at the top of the collector and from there goes  to the domestic hot water system, usually into the backup heater tank.

Integrated collector storage (ICS) systems are similar to thermosyphon systems in that they have some water storage in the pipes making up the collector. Unlike thermosyphon systems, however, the water pressure from the city main or well drives hot water into an auxiliary storage tank in the house.

Drain back or drain down systems also use the main water pressure to push water through the collector, but they use a different sort of freeze protection. Thermosyphon and ICS systems have a temperature sensor on the collector. When the temperature reaches the freezing point, the sensor causes hot water to be circulated into the collector to prevent freezing. This decreased the efficiency because in winter the hot water may come from the backup electric or gas hot water heater. In drain back or drain down systems,  when the temperature nears the freezing point a valve on the main side  of the collector closes. A valve on the house side closes too and a valve emptying the system opens. All the water in the collector and collector loop plumbing drains out. When the temperature rises sufficiently above freezing, the drain valve closes and the main and house valves open and the loop repressurizes.

The most expensive type of system is the indirect system. Indirect systems don't need to have freeze protection since the heat transfer fluid won't freeze in winter. An electrically driven pump drives the heat transfer fluid between the hot collector and a heat exchanger coil in a water storage tank. The heat from the heat transfer fluid is absorbed by the water in the tank, and the water then is used for the domestic hot water supply. When the temperature of the collector drops below the temperature of the tank, the pump shuts off. Similarly, if the temperature in the tank reaches a maximum (usually no more than 180F) the pump shuts off and the collector stagnates. At stagnation, the collector can reach very high temperatures, usually near the boiling point of water. Indirect systems need to have some kind of protection against stagnation, otherwise the heat transfer fluid can cook and become acidic. If that happens, the fluid must be replaced.

There are two kinds of collectors available for indirect and drainback systems: flat plate and evacuated tube. Flat plate collectors look like big flat glass skylights, except they are black. Here's a picture:

The simplest flat plate collector is just a piece of glass on top of a metal manifold or metal tubing coil through which the heat transfer fluid circulates. The manifold or tubing is painted black. Usually, insulation separates the manifold from the back of the collector to prevent heat from being transferred onto the roof.

 An evacuated tube collector consists of a collection of glass tubes from which all the air has been removed. Here's a picture:

The collector tubes extend vertically along the roof between a manifold on the top of the collector through which the heat transfer fluid circulates. A black heat pipe extends through the middle of the tube and contains another heat transfer fluid sealed in the heat pipe. When the sunlight strikes the collector, it heats up the heat pipe causing the fluid to rise to the manifold. Heat doesn't escape from the collector because there is no air inside to transfer heat, just as in a vacuum bottle. The heat exchanger fluid running through the manifold absorbs the heat from the heat pipe and transfers it into the hot water tank in the house.

Evacuated tube collectors put out more heat during the winter than flat plate collectors, and consequently are better for hydronic space heating systems, which circulate water through pipes in the floor. On the other hand, evacuated tube collectors tend to become too hot in the summer. Ironically, one of the major problems with solar heating systems is that they tend to get too hot in the summer. For domestic hot water systems, this isn't such a problem because you can size the system for maximum summer temperatures and limit the number of times the system stagnates. For solar space heating systems, however, you need somewhere to dump the heat, like a hot tub or swimming pool. In moderate climates, such as here in coastal California, flat plate collectors perform better overall because winter temperatures usually aren't that cold for extended periods.

Generally speaking, direct systems are not recommended in areas where there is any tendency to freeze during the winter. This includes areas such as coastal California, since we do get freezing temperatures at night. The thermosyphon and ICS systems are less vulnerable to freezing because they keep more hot water in the collector or near it where it can be recirculated to inhibit freezing, but in a really hard freeze there may not be enough water to recirculate. The drainback systems can freeze if the power goes out. If the direct system is only used during summer and completely drained in autumn before the first freeze, they won't freeze, but in general, an indirect system is better if freezing temperatures occur.

Solar hot water systems require some kind of backup gas or electric water heating for when the sun doesn't shine and in winter when there may not be enough sun to fufill hot water needs. One common way to install such systems is to make the solar hot water system a preheater that heats water prior to entering the backup heater tank. Another way is to build an electric or gas heater into the solar hot water storage tank, or have a separate heat exchanger coil in the solar hot water storage tank through which a heat exchanger fluid from an external boiler, like a gas boiler or a geothermal heat pump, can be circulated. A final way that doesn't seem to be very common but seems sensible is to have an on-demand electric or gas heater provide additional heat to the solar heated water if it is not hot enough.