Having set up the solar powered electrical system for my house and seen how successful it was, the next step for less dependence on outside services controlled by big business, was to have my water heated by solar.

Learning about water heaters.
Late 1996, when my 15 gallon Zip water heater tank had corroded to the point of brazing up the pinholes being a futile exercise, I was forced to replace it. This is when I learned that most water heaters for sale now are essentially junk. My original tank was made of cusilman bronze and it had lasted 30 odd years with full mains pressure applied to it. I'd also lived with copper and copper lined tanks which were also reliable. Much to my surprise neither cusilman bronze, or my other choice, copper lined steel (eg. Rheem Coppermatic) heaters were still being made. Something about water quality the manufacturers told me...but they could offer me any number of models with glass lined tanks. No way would I spend money on something which was actually destined to fail from the first day it held water. For those unfamiliar with the concept, glass lined tanks are a mild steel (just saying that makes one think of rust) with vitreous enamel (fancy name for molten glass) sprayed onto the inside surface, supposedly protecting it from corrosion. Which it would, except that the steel cylinder expands and contracts as it heats and cools. As glass is not flexible, minute cracks appear and the tank is now doomed. To slow down the rate of corrosion, a sacrificial anode is inserted into the tank. It's made of magnesium and is meant to corrode instead of the steel. However, most owners of glass lined tanks never check the anodes, or replace them once corroded away. Also, the corroding magnesium ends up in the tank and hence the hot water that issues forth. Such tanks often have words like "Vitreous", "Enamel", "Glass" or that abominable misspelling by Rheem, "Glas", in their model names. The inside of them, if you haven't ever seen a cutaway model, is just like an oven baking dish. If you're unfortunate enough to be stuck with this kind of heater, run it at the lowest temperature and water supply pressure you can, and check the sacrificial anode. Edwards and Beasley seem to be the predominate manufacturers of stainless steel water heaters, so in the meantime it's worth familiarising yourself with their models, ready for when the time comes.
I should point out here that most water heaters in Australia are of the mains pressure type, hence the requirement for a strong tank. To allow for pressure build up, a TPR (Temperature and Pressure Relief) valve is fitted. This opens thermostatically at 99 degrees (in case the heating element doesn't shut off) so that steam doesn't build up causing an explosion. In addition, should the thermal part of the valve be not functioning, the valve will open with extreme pressure alone (chosen to be above the pressure of the water supply).
Once popular until the 1960's when the Rheem Coppermatics came into vogue, were gravity fed copper tanks, usually installed inside the roof space. These have an exceptionally good life, and fed from an open cistern, do not require all the pressure release paraphernalia of mains pressure heaters.
As I didn't have the roof space to accommodate such a heater, I stayed with the mains pressure option.

The Edwards Stainless Steel Water Heater.
Further into my enquiries, I soon discovered the stainless steel water heater. Problem solved. It does surprise me that so few water heaters sold are of the stainless steel variety. Perhaps the extra cost puts people off and they think five or ten years life is satisfactory.
The model I chose was an Edwards Energy Systems DES125 which is a 125 litre mains pressure model. It is made from 316 stainless steel and had a ten year warranty. I installed the lowest pressure reducing valve I could get (350kpa) to further increase the life. Interestingly, the Edwards water heaters come with a cold water expansion valve. This is a reflection of their Western Australian origins. In that State and also South Australia, this is apparently a compulsory requirement. In those areas the water supplies are prone to cause scale build up on the TPR valves. This can cause them not to reseat properly after being activated, and thus the familiar failure with constantly dripping hot water. The ingenious way out of this is to have an expansion valve on the cold water inlet of the tank. Because the cold water does not cause scale build up, the valve is reliable. The cold water expansion valve therefore has a lower pressure limit than the TPR valve, so in effect the TPR valve is only activated if the water approaches boiling point. The additional advantage of this scheme is that only cold water is lost during the heating of the tank, not hot water as is the case when there's only a TPR valve. Thus the heating energy is lessened.

Dark inside my laundry, but here resides the Edwards DES125 stainless steel water heater.
 

Going Solar.
Having lived in New Guinea for a lot of my younger years where solar water heaters were standard, the thought certainly did cross my mind of replacing my failed tank with a solar heater. The water heaters at the time in New Guinea were a 1960's design with two flat plate collectors and a vented, vertical copper tank with a small float valve fed header tank, on top. This was all mounted on the roof on a suitable galvanised steel frame. Just about every house used this design and an indoor electric heater was a rarity. When Solahart started promoting their horizontal tank models in the 1970's, it was a retrograde step. While some people might prefer them from an aesthetic view, the thermosiphon is far less efficient, and so is the isolation between the hot and cold layers of water inside the tank. However, the improvement in solar collectors over the years would probably mean that although innefficient, the horizontal tank models of today are probably at least as good as what I was used to in New Guinea.

Solahart.
I did do some research into solar water heaters around 1989 when I'd wanted to install one at our weekender at Woodford. At the time I was keen on the 180 litre Solahart. They had developed this "black chrome miracle" collector panel which supposedly transferred more of the sun's heat to the water. They also had developed a novel heat exchange system for use in frost prone areas. In such locations, the panels can be damaged overnight in sub zero temperatures. The water in the panels freezes, expands, and thus ruptures the pipes in the collector. Ways to solve this problem are to use frost valves which open in the cold, allowing hot water from the tank to flow through the panel. Apparently such valves are unreliable as well as wasting hot water. Another method is to circulate a small amount of water through the panels by means of a pump. While there is no cold water loss, some of the heater water stored in the tank is wasted warming the panels. The Solahart approach was to heat the tank by means of antifreeze fluid which was in turn heated by the solar panels. This neatly solves the problem with no heat or water wastage. However, the heating efficiency is less due to losses in the heat exchange process, and given their tanks are glass lined, I would think the glass would function as an insulating layer.
When it came to replacing my water heater I could have used the Edwards version with stainless steel tank, but as I'd just started my house loan, cost really was an issue so I took the easy way out and stayed electric only.

Evacuated Tube Collectors.
Towards the end of 2007 I'd become curious again about solar water heating. I'd now had the solar electric supply operating for two years, as well as the wind generator. A large proportion of the house water supply was coming from a 1000 gallon rain water tank, so solar water heating was next on the agenda.
The people I'd spoken to at Edwards 11 years ago told me my tank could be converted to solar and this had always stuck in my mind...so onto the internet to see what I could find. My initial idea was just to get a conventional flat collector panel and connect to my existing tank. I quickly discovered that there was a new kind of collector with many advantages. It was called the Evacuated Tube Collector. Not only was it very light in weight, but very easy to install and frost protection was not an issue. In my area, frost protection would be necessary with a conventional system. The thought of antifreeze fluid leaking and running down the roof into my tank was also another reason to avoid the well known commercially made systems.
The collector consists of an array of glass tubes. Inside the glass tubes are copper tubes filled with liquid. These copper tubes are heated by the sun (they have black fins attached), and the hot liquid rises to the top of the tube which is in contact with a heat exchanger. Through this heat exchanger flows the water to be heated. The key to this working so well is that the glass tubes are constructed like a thermos flask. There is a vacuum between the inner and outer tube which means that heat trapped inside cannot easily get out. Also being round, it means the angle of the sun is not critical and the tubes receive radiation over a much greater period of the day than would a flat panel. As there is no water flowing through the actual collectors, they are immune to frost damage. The whole array weighs only about 60 kg, with the weight being only that of the glass tubes, heat exchanger manifold, and stand.

Retrofitting to an existing tank.
A typical vertical water heater tank has a cold water inlet right at the bottom and the hot water outlet right at the top. A third connection at the top is for the TPR valve, but some heaters combine this with the hot outlet. The heating element is located in the lower quarter of the tank and thus heats the water above it. A tank made for solar use has an extra connection about halfway up. This is where the warm water from the solar panel is fed into the tank. The cold water for the panel comes from the bottom of the tank. Thus when the panel is heated, thermosiphon occurs. How to use an existing tank? The cold supply for the panel is easy as there's already a connection at the bottom of the tank; i.e.. the cold water supply inlet. The hot output from the panel is the tricky part. If it was merely fed into the top connection of the tank, there could be conflict with the hot water return from the solar collector; eg. if the water from the collector is cooler it will mix with the hot water from the tank and thus reduce the outgoing water temperature. I eventually discovered the trick is to use something called a "five way valve", or "solar conversion valve".

This diagram came from the eBay seller of my kit and illustrates the five way valve.

What happens is that the cold water connection at the bottom of the tank performs two functions. The cold water is fed into here as well as to the solar panel. However, the hot water is also fed via this connection to a length of tubing inserted into the tank, the end of which is away from the cold inlet. So, cold water is drawn from the bottom of the tank, pumped through the solar collector, heated, and then injected into the bottom of the tank (away from the cold area) and rises to the top. Thus, no modifications to the tank are required. Obviously, the tube injecting the hot water must be of smaller diameter than the cold water inlet. Typically, the cold water inlet is 3/4" which allows a 1/2" injection tube. The remaining 1/4" gap is quite sufficient for cold water flow.

Five way valve made from ordinary plumbing fittings. The length of copper tube protruding into the tank is somewhat longer in my installation.

This type of conversion applies to a ground standing tank with solar panels on the roof. Unfortunately, for thermosiphon to occur, the tank needs to be above the collectors as hot water naturally rises. With this type of installation a circulating pump is therefore required, so unless this is powered from a photo voltaic panel the hot water is not entirely free. However, the pump is only for circulation instead of having to push a head of water so power requirements aren't that huge. My Wilo pump supposedly draws 46W. Somewhat cheaper to run than a 3.6KW element! The controller means the pump only runs when the water in the solar collector is warmer than that at the bottom of the tank, so it isn't running all the time. In practice, the pump cycles every few minutes on a sunny day; the stronger the sun, the longer the pump stays on.

The Wilo circulating pump. It has three speeds, the lowest of which is used for this application. It would have been nice if the proper unions were supplied with mine; instead a slightly bodgie adaptor came with the kit which fortunately hasn't leaked.

My Evacuated Tube Collector.
After looking around on the internet at these, it became obvious the best place was to get one was off eBay. I wasn't interested in paying someone else to install the system, so therefore wouldn't be eligible for the government rebates. That lessened the chances of me buying off a commercial supplier who would no doubt insist a licenced plumber do the installation etc.
I ended up with a kit of parts from a seller by the name of "pekinglook". It cost $1250, which is the typical cost of these units in Australia. Included was the roof stand, heat exchanger, 24 evacuated tubes, a Wilo circulating pump, the controller, and a few brass plumbing fittings meant for making your own five way valve. The seller promptly delivered the boxes containing said parts one evening, and I could see that his claims about it all being light enough to transport on the back of some yaks in Mongolia was quite true. So much less cumbersome than the usual flat plate set ups. You certainly need to be proficient with domestic plumbing, and understand how solar water heaters work to be able to do the installation. What I did find very useful was to look at installation instructions for commercially made water heaters. In particular, the Rheem Loline which is a solar retrofit. Although it uses a flat plate collector, the principles are the same and so is the tank retrofit and controller operation. Also of considerable use was this hot water installation course compiled for Australasian conditions.

Evacuated tube collector installed on my roof. Note the photovoltaic panels in the foreground.
 

The Installation.
First thing was to assemble the stand and mount it on the roof. The stand is stainless steel and the roof is zincalume. To those outside Australasia, this is a zinc/aluminium material with similar appearance to galvanised steel and is used for the same applications. Given the corrosion possibilities with dissimilar metals in contact with each other, I mounted the stand on neoprene pads before screwing it to the roof.
It was rather convenient that the roof area where the solar collector was to go, was right above the laundry where the hot water tank is. Thus there would be a short vertical run of pipe. There is approximately 10m of pipe in the circuit. I used green lagged copper tubing for the pipe run. All the fittings were designed to take 1/2" tubing.
Getting the pipe through the ceiling space was the most tedious part as there was a piece of framework that had to be cut away above the ceiling cornice. Once this was done, the pipes were installed and the roof sealed with a Rooftite ( a neoprene gland attached to the roof with stainless steel screws and sealed with silicone).
I was a bit concerned about mounting the pump due to its weight and the fact that it was meant to be supported only by the connecting pipes. As it turned out, 1/2" copper pipe was quite sufficient to bear the weight. I was rather doubtful about the connections to the pump as they were meant to be used with some kind of union. This wasn't supplied with my purchase, but instead were just simple 1/2" to 1" reducers with home made silicon washers to seal the whole thing. As it turned out this works perfectly with no leaks.
The plumbing was straightforward until I attempted to make the five way valve. The way the Edwards tank is constructed does not allow such a thing to work; not on the cold inlet anyway.

Pump installation and cold water plumbing. The cold water expansion valve is largely obscured by the pump terminal box. Pressure limiting valve is at the lower left.

Instead of the cold inlet simply being a port in the side of the tank, it's actually a connection to an internal pipe that runs underneath the tank to the centre. How did I find this out?
I tried making the five way valve and inserting the hot water copper tube but found it didn't go very far.
The design of this tank has the hot water outlet right in the centre of the top of the tank; not at the side. So, I could actually shine a torch down and see inside the tank to learn of its internal construction. Nice to see all that gleaming stainless steel inside with not a hint of corrosion. How many glass lined tanks would be like that after 11 years?
It was this top connection that solved my problem, for this was the logical place to put the five way valve. There's about 1.2m of copper tube going straight down into the centre of the tank for the hot water to be injected. The hot water is thus fed into the tank at the same location as if I'd put the five way valve at the cold inlet.

The diecast box at the bottom contains the relay for switching the electric element and socket for the pump. The additional switch on allows the element to be permanently turned off when required. Neon pilot lamps show pump and heating element operation. Green pipe at left is cold supply to collector; green pipe at right is the heated water return. Five way valve is largely obscured by the foam insulation, but hot water takeoff is visible to the left. T3 thermistor is inserted between insulation and brass T hotwater takeoff. Grey cable at left of green pipes is for the collector temperature sensor.

Once I'd finished the plumbing, turned on the water, and got the leaks out of the connections, it was time to install the glass tubes. The instructions recommend talcum powder to lubricate the neoprene glands in the heat exchanger, into which the tubes are inserted. I didn't have this so I tried CRC2-26. This worked perfectly and the tubes slid into position very easily. Next, the hose clamps at the base of the stand were tightened to secure the tubes. Finally, the stainless steel reflectors were attached between the tubes.
At this point I hadn't installed the control unit so the pump wasn't running. By the time I'd finished inserting the 24 tubes I found the hot outlet on the heat exchanger too hot to touch. Being a sceptical person I was quite delighted to see that at least the tubes were doing something.

Not my unit, but the Endless Solar version, this shows how the tubes are inserted into the heat exchanger.

Within a few seconds of turning on the pump I had to let go of the hot water pipe from the exchanger going into the tank, as there was a burst of very hot water flowing through it.
It was at this point that I finally believed these tubes might actually work.
The final part of the installation was to install the controller. To this are connected two thermistors. One (designated T1) is inserted into the heat exchanger near the hot outlet and the other is located at the cold water connection at the bottom of the tank (designated T2). This thermistor was conveniently formed into a 1/2" threaded fitting so could be screwed into a T piece through which the cold water flows. The default setting of the controller is such that the pump turns on when the collector temperature is eight degrees warmer than the cold supply to the collector at the bottom of the tank. The pump then runs until the temperature difference is four degrees.
Two other functions which are not really necessary in my installation are frost and over temperature protection. Should the T1 sensor detect imminent frost (default setting is two degrees), the pump will circulate warm water until four degrees is reached. This prevents the pipes freezing and bursting. This would be more applicable to flat plate collectors, however. I have noticed that there is a degree of reverse thermosiphon at night anyway as the pipes to the solar collector remain warm even though the pump is off, so this feature is redundant (See postcript below for details on this). The other feature is to turn the pump off should T2 reach 70 degrees. This stops the hot water flow under extreme temperature conditions. I'll believe a solar water heater can get that hot when I see it. Still, these features are provided for and will come into action if they ever need to.

Diagram supplied for my solar heating setup. In addition I have made use of the T3 temperature sensor input and the H1 heating control for automatically switching the electric element. Unfortunately there's an important omission: There should be a non return valve either on the pump outlet or solar collector return to tank. Without this, reverse thermosyphon occurs at night. The pump wiring via tank thermostat should be ignored.

How does it perform?
I'd finished the installation late morning. By mid afternoon the controller was telling me that we'd reached 52 degrees in the tank. In reality the water available from it would be hotter given this temperature being taken from the coldest part of the tank.  This was late December 2007 and we'd had a few days of warm clear weather.
I have since done two further improvements. I wasn't impressed with the lagging on the copper pipe. Like most building products these days, there was obvious cost cutting. The green lagging of today is just a thin layer of plastic, unlike the foam of yesteryear. One could clearly feel the heat radiating from the pipes. I improved the insulation with the thick black foam stuff (sold by Bunnings as "Handitube"). What a surprise that was...up to 62 degrees and it wasn't a totally cloud free day. The weatherproof qualities of this insulation are questionable and I will probably have to cover it in duct tape.
The other improvement was to switch the electric element using the controller. Looking at the weather and deciding whether to go to the fuse box and turn the element on isn't convenient. The instructions with the kit imply that one simply reduces the thermostat setting on the heater to the minimal acceptable temperature.  As long as the increase in water temperature from the sun is above this, the element doesn't switch on. Undoubtedly this would certainly work, but I could see that there would be times when the water temperature could drop and I didn't need it heated; e.g.. overnight. Being in penny pinching mode, I decided to go one step further.

Water temperature would be about 38 degrees at the moment. This was in the early afternoon and it had been cloudy all day. Days like this require the electric element to be turned on later in the afternoon. Mine comes on at 17.30 until 18.30. This is quite sufficient to provide hot water for the evening and following morning.

The controller has a a feature where it can switch on the element up to three times a day for a specified time and until a preset temperature is reached. This temperature is sensed by a third thermistor at the top of the tank (T3), where the water is at its hottest. I simply pushed the thermistor in between the foam insulation and the brass T fitting at the hot water outlet. While it isn't in direct contact with the water, it works well enough as the brass is a good heat conductor. In practice, I found the thermistor shows about three degrees less than the water temperature. So, to have the element switch off at 50 degrees, I set the controller for 47 degrees. I have set the controller so that the heating element is switched on (if it needs to be) at 17:30 for an hour, which is just long enough to get the water temperature up to 50 degrees, from about 36 degrees, which is typical for a cloudy day. This works in with my hot water requirements being mainly at night, which is the best way to use a solar hot water heater.
The controller supposedly can switch 16A for the heating element, but I find this hard to believe given it's all on a printed circuit board and the relay doesn't look very substantial.
I decided to make a second box with a contactor inside, as well as some neon pilot lamps to indicate when the electric heating was activated, and when the pump was running.
The heating relay inside the controller simply supplies the low current of the contactor coil which then does the job of switching the 15A heating element. Unfortunately, the contactor having an AC coil buzzes loudly at 50 cycles. As the whole thing is mounted on a hollow fibro wall the sound is amplified to a large degree and is very unpleasant. Next step was to use a 12VDC relay with 250V 30A contacts. The coil is fed from a small transformer and rectifier. Being DC it is totally silent in operation.
I am very sceptical of claims that certain solar powered devices will work in cloudy conditions, and evacuated tube collectors seem to be no exception. The answer is simply that they will provide some heat but the truth is not to a usable degree. The electric element is essential after a two days of cloudy weather. Having said that, the use of solar collectors on a cloudy day will still save electric power because the element is only heating the tank from say 35 to 50 degrees, rather than starting off at say 18 degrees. It takes less energy to obtain a 15 degree rise in temperature than 32 degrees would require.
It is hard to believe that in Australia with its climate that there are so few solar water heaters. They clearly do work and save a substantial amount of electric power. I've had the electric element switched on only about five times in about two months. Yet, people continue to replace their failed (glass lined, of course) off peak or continuous hot water systems with the same. Don't get me started on the wasteful consumption of air conditioners, plasma televisions, and the greedy attitude of people that use them. Ideas of new power stations would simply be unnecessary if more people used solar energy. It works and people are using it right now.

Postscript:
I've just received my first power bill since installing the solar conversion. From $223 it has dropped to $161. This is a quarterly account and I should point out that the billing period in question commenced about a month before installation. So in actual fact the saving might be more. In any case, that is a very pleasing result.
As we are now into Autumn, something started to become very evident. My suspicions of reverse thermosyphon at night was confirmed. With cool night time temperatures outside, most of the heat accumulated during the day was simply dissipating into the night sky. I was beginning to think a non return valve was needed and as I looked at the installation manuals for the commercially made systems, it was now quite obvious; they all used one. To put it simply, I was rather annoyed that not only my kit supplier did not provide or mention the need for such a valve, none of the other eBay sellers mention it. I eventually decided on the RMC solar non return valve type SNR502. An ordinary non return valve should not be used as it may not be intended for the high temperatures in a solar hot water system. Luckily I'd seen the RMC price list for this valve before buying one. It was around $57 retail. So, when I rang up Reece to order one and was quoted $147, I looked elsewhere. Tradelink in Penrith had one ordered for me in a couple of days...and for the correct price.
Having installed the non return valve, the change in performance is like the difference between night and day. The water is still hot in the morning! The pipes running up the collector are cold at night as they should be, and now I can look at the T1 temperature reading to see what the outside night time temperature is.
One other annoying thing is leaking screwed fittings. I used compression fittings or soldered Yorkshire fittings where possible. Even with copious amounts of teflon tape, the screwed connections always leak. I've done them up to the point where the threads are about to be stripped, and yet I can almost count on seeing a drip when I turn the water on. So annoying did this get I just got the shits and ran solder into the threads. Problem fixed! I only did this where serviceability isn't goint to be affected of course. The real problem is the alot of fittings don't have a tapered thread and this is what's really required to prevent the problem.
I have also noticed with the onset of cooler weather how useless the green plastic insulation is. All it does is prevent you being burnt if you touch the pipe. So, I have put the black foam insulation on all the pipes in the hot water circuit. For outdoor use, this kind of insulation must be covered. It deteriorates fairly quickly under UV light.
This problem was solved by wrapping exposed sections in aluminium tape. I bought a roll from Jaycar. The catalog number is NM2860 and only cost $15 for a 50m roll.

The SNR502 installed on the pump outlet. No more loss of hot water at night.

Quite obviously, eBay sellers of evacuated tube retrofit kits do not understand the finer points of solar hot water. They are merely importers from China. You are on  your own if you buy one of these kits; do not rely on what they tell you. Having said that, I've learnt alot from this exercise, and hopefully my notes here will enable yours to be successful too.

home page

Thanks to eBay for providing some of the illustrations used. And no thanks to Australian eBay requiring all sellers to use paypal; a big company is getting greedy.

Email me: cablehack at yahoo dot com.