Rainwater Harvesting

Isn't it a crime to live in one of the wettest countries in the world (the UK) and yet flush treated drinking water down the loo? I've thought so for quite some time, but the complexity and cost of a Rainwater Harvesting system has always put me off. Nevertheless I have now taken the plunge and built one, and hopefully the design decisions I have made (having taken a considerable amount of time to think them through) will be of use to someone.

Rainwater Collection Diagram

Underground or Overground?

Rainwater Storage Search for any commercial rainwater harvesting system and it will be an underground tank. There are two good reasons for this: frost and stagnation. An underground tank keeps all the component parts of the system safe from excessively low or high temperatures. Frost will burst your pipes and bust your pump. High temperatures could create the perfect breeding ground for bacteria and insect larvae, which you really don't want to have swimming around in your cistern.

Sadly such tanks start from £1500, and that doesn't include the considerable task of burying it and plumbing it in which will add thousands. Therefore an underground tank was simply never an option, and I would have to seriously question whether such a system can ever justify its existence. As a result my system is based around external water butts for storage, and I will just have to put up with the risk of freezing. I have brought what components I can inside (namely the pumps) but all the pipework is outside. To mitigate the risk of burst pipes from frost I have used flexible garden hosepipe (also the cheapest option!), and the pumps I have chosen actually allow the water to drain back into the butts when they are powered off. Empty pipes can't freeze!

There is another slightly more subtle reason for building the system this way: it is removable. Anyone who has ever sold a house will know that the Surveyors will go over your house from top to bottom, assessing every DIY task you have ever undertaken with a fine tooth comb, hunting for violations of the dreaded Building Regulations. Each esoteric violation will be listed on the report, and estimates sought for their rectification. You therefore end up paying through the nose for minor infringements you yourself have lived with quite happily for years. One solution to this is to make sure you can easily remove anything that might cause an infringement! The other benefit is that you can take the whole system with you.

Justification of Existence

This is my term to describe whether a system is worth building. Money-grubbing accountants would simply ask what the payback period of a system is, as if the financial impact of a system is all that matters. Sadly there is more to it than that, because even if the system saves you money, it could quite easily be the case that the environmental cost of the resources expended in its manufacture, maintainance and disposal outweigh the benefits. Perhaps a better term is Total Resource Cost of Ownership.

The more I have studied supposedly green ideas the less and less I have become convinced of their overall benefit. What is the point of an Air Source Heat pump if the electricity to power it is generated by burning coal? What is the point of a wood burning stove if the Nitrous Oxide emissions are off the scale? And what is the point of Rainwater Harvesting system if you use mains electricity? It's very easy to over egg the pudding and produce a design where the complexity, maintenance and lack of reliability and short lifespan outweigh the net gain. Quite how you are to know when you have overstepped that line is another matter. I think you'd need a degree in Sustainable Design. All advice welcome.

Header tank

Cistern with Rainwater in it Nearly every other system I have read about uses a header tank somewhere at the top of the house as a convenient way to generate pressure. All the system has to do is put the water into the header tank as slowly as it likes, and then gravity will do the rest. It also makes a convenient place to introduce the mains water backup, should you decide to go down this route. However, to me this just seemed like an extra tank, would require a powerful pump to get it there, and would require complex replumbing inside the house. I also couldn't think of any way to get the water to the top of the house except up the outside, where it is subject to frost.

Therefore I have chosen not to have a header tank, and instead to have multiple pumps. This might sound very wasteful, but the pumps are only £25 each, and I only need two of them. Perversely it actually costs more to buy a valve than it does a whole pump. The downside is that it takes 4 minutes to fill the toilet cistern with 7 litres of water. Not very fast, but then if people really do need to flush the toilet more often than that, they can just reach down and undo the mains water stop cock. Note in the picture that the black 7mm hosepipe and the wire are fed up the overflow of the cistern. This avoided making any extra holes in the bathroom, which was a major consideration because we have stone walls 2 feet thick. Sadly the float switch had to be fitted upside down with its wires trailing in the water, but that's the only polarity they sell.

Mains water backup?

Solenoid Valve I decided early on that I wanted an automatic mainswater backup for the system, but now having built it I'm not sure it was worth the effort. For filling the toilet, it's pretty obvious when the cistern is empty, and on such occasions it is easy to reach down and undo the stop cock, and so long as you remember to switch it off again when it next rains that is a perfectly adequate system. For watering the garden, I'm not convinced that I actually want the system to automatically switch to using expensive mains water for a task that might well not be necessary.

Nevertheless, the system has a mains water backup, and it's worth discussing the various ways to do this. The simplest solution appears to be to use a ball valve with an extended arm, so that the ball is normally fully immersed, and only opens the valve when the water level falls low enough. My reservations about this are that I believe it violates UK Water Regulations, which require a Type AB Air-gap between mains water and, er, any other type of water. Also, in my system, the difference in water levels between full and empty is the best part of a metre, and no ball valve is going to cope with that.

Therefore I decided to use a Solenoid Valve instead. This proved to be a bit of a nightmare because it is almost impossible to buy a 12VDC Solenoid Valve. There are plenty of 24VDC ones (since 24VDC seems to be the standard voltage in the irrigation industry), but these can be very expensive, and would require an entire rethink of the electrics.

In the end I bought a 12VDC Solenoid Valve from Shanghai, and had to pay for the airmail, so it wasn't cheap. Neither did it come with 15mm connectors, only 1/4 inch, so I had to buy two rather uncommon adaptors, so all in all the valve came to about £35. Float switches aren't cheap either, and the one I bought was £15 from Maplin. Therefore the mainswater backup ended up adding over £50 to the cost of the system, and I'm not sure I need it. Now in use, the one good thing I would say about it is that it does help to avoid airlocks in the system, however you could achieve that equally well just by using a float switch to turn it all off.

Ideally the mainswater would be fed into Water Butt Number 1, so that during dry periods the fresh water will continue to flow through the whole system. It will also reduce the number of cycles of the solenoid valve to the absolute minimum. However, in practice you might need to feed it in closer to the float switch so that the action of the pumps does not lower the water level enough to cause an air lock. Only experimentation will reveal where the best place is. Feeding into the same Water Butt that contains the float switch can cause a lot of waves, which then cause the float to bounce up and down causing chaos for the electrics.

240VAC, 24VDC or 12VDC?

Solar Panel for Rainwater System I'm becoming a bit of a convert to solar power. In the past I've always regarded it as too weak and inefficient to be worthwhile, but I've realised that it's actually very good for applications that only require a bit of power now and again, and that would otherwise commit the cardinal sin of racking up a hefty standby power consumption. True, there are a lot of good 240VAC pumps on the market that seem purpose designed for pumping water out of external water butts. They are rugged and designed to cope with frost and dirty water. So why am I not using one?

The answer is that I didn't want to be put in the position of having to work out whether the electricity consumption of the system was undoing all the benefit of harvesting the rainwater in the first place. After all, this is only water we're talking about, and it's resource value is very low. The water companies are very efficient at what they do, and they happily produce millions of tonnes of fresh, pure, drinking water every year at a fraction of what it would cost you or me. The other benefit is that when it comes into your house it's already at mains pressure, so you don't have to pump it yourself. It can be very hard to do something better than a huge utility company.

My decision not to use 240VAC was helped by the fact that I had a spare 12V car battery that wasn't doing anything, and I didn't want to mess around with mains electricity near water. 24VDC would also have been a wiser choice, since it is the standard voltage of commercial irrigation equipment, but would necessitate either two batteries and two solar panels, or a cunning voltage doubling circuit. I decided not to create hard work for myself and just standardised on 12VDC. Even with the hassle of buying a 12VDC Solenoid Valve from China, it was still the right decision.


Two Whale Pumps Pumps can really be the weak point of a system, since they can be both expensive and unreliable. We shall see how long the ones I have bought will last, but on first impressions I am really pleased with them. I decided early on that the Whale Caravan In-Line Booster Pump GP1392 would be ideal for this system when I found a webpage listing their head versus flowrate capabilities (which are usually impossible to track down). It turns out that these pumps are used for groundwater testing, and can raise water over 10 metres, and boast some impressive flowrates. The first time I powered one up in a bucket of water I knew they were just right - even open-circuit they can squirt water halfway across the patio!

In a typical caravan arangement the pumps would be used with a pressure transducer, whereby the pipes are pressurised all the time. Opening a tap causes a drop in pressure, which is detected by the transducer, which switches on the pump. This ensures that water is available instantaneously, but leaves the system very vulnerable to frost, hence why caravans have to be drained down over winter. Since speed of operation is not an issue with irrigation or toilet filling, this arrangement is not required.

To Prime or not to Prime? The Whale GP1392's are distinctly non-self-priming, which means that if you get an airlock in the system they just whirr away uselessly. Worse still, each pump must not be used continuously for more than 15 minutes at a time, so if you aren't around to hear them whirring uselessly, they will get damaged. Although it's possible to create a timeout circuit to protect them, this would just be even more cost so we will just have to take the risk. Make sure that the pumps are located as low down as possible, so that they are always full of water. If you get a persistent air lock, it can be removed by using a hosepipe to force water through the pump from the water butt outlet. You can get a bit wet doing this, but it does work.

Note that you have to connect each pump directly to the water butt. If you use a tee-formation (as I did first time round) then all that one pump does is to suck water backwards through the other one, not out of the water butt. Note also that the hosepipe connects directly onto the pumps. Just soften the hose with a bit of hot water first and then push it on.

To Filter or not to Filter? The traditional way to protect unreliable pumps is to use unreliable filters. My suspicion was that I would simply end up clearing out blocked filters without really know whether the pumps needed them. It also occurred to me that a system with multiple tanks invites a rather more informal method of filtering - that of just letting the water travel slowly from one water butt to the next. Slow moving water clears itself, hopefully.

Another informal method of avoiding blockages is to encourage backflow when the pumps switch off, which is why this system does not feature any non-return valves. So far I haven't experienced any problems, but if I do then there is a plan B - I shall pour the contents of a bag of sand into one of the water butts. I imagine sand will make a very reliable filter, and it can be unblocked just by prodding it with a stick.


Rainwater Diverter There is no answer to this question, except that any capacity is better than no capacity. In a typical year how many times do we actually go a week without rain? Therefore so long as we can store about 250 litres, I'm happy. I can add more water butts if needed as time goes by. One point that is worth making is to make sure that the water butts are daisy-chained. This ensures that rainwater enters butt number 1, then proceeds to number two, and so on to the Nth butt, and then finally overflows from there down the drain.

This keeps the water in all the butts as fresh as possible. You will discover that the system will spend most of its time overflowing whenever it rains, so it is important to refresh all the water in all the butts whenever it overflows, not just the first one, or none! A typical downpipe diverter from a garden store will have an overflow built into it (which makes a very noise glug!). It is therefore important that the Nth butt overflows at a slightly lower level than the diverter into the 1st butt, so that the diverter overflow is never used.

Irrigation Timer

Electricals There are a vast number of irrigation timers on the market, but all of these are designed to work with a tap, i.e. with mains-pressure water. Since our system is unpressurised, none of these are an option. What we need is a way to switch 12VDC on and off to the irrigation pump at times that we can pre-programme. Sadly, you will quickly find that all the cheap programmable timers on the market (as sold in Tescos) only work with 240VAC. Stalemate.

Fortunately there is one company that sells a converted 240VAC programmable timer for use with DC. It is a bit Heath Robinson, but it retails for just £21.99 plus £1.95 postage and packing, which I assure you is only barely enough to cover the cost of the parts, let alone the time required to solder it together. I thoroughly recommend having an extended browse of REUK.co.uk's website, because it is crammed full of great ideas. The timer is dead easy to use, and comes with a 10A 30VDC relay that does the actual switching. This saved me so much time.


Some of the items listed below are not very competitive prices, e.g. the float switches. Others are definitely the best I could find, e.g. all the stuff from Wickes. Hunt around and see what you can get, and if you see a Solar Panel going cheap, nab it.

Item Vendor Code Unit Qty Shipping Total Notes
Water Butt Wickes £21.00 4 £84.00
Water Butt Stand Wickes £8.80 4 £35.20
Solar Panel Maplin L14BR £60.00 1 £60.00 (now up to £100 again)
Battery Free £0.00 1 £0.00 (find an old one)
Hosepipe Wickes £9.00 2 £18.00
Water Butt Taps Wickes £1.94 8 £15.52
Relays Maplin £2.70 2 £5.40
Solenoid Valve Virtual Village £14.99 1 £6.99 £21.98
Adaptors (¼ inch to 15mm) Partridges YP3LC £4.28 2 £2.75 £11.31
Irrigation kits B&Q £22.50 2 £45.00 (actually got them for £10 each)
Whale GP1392 pump Blazers Caravans £26.00 2 £52.00 (a bit cheaper on-line)
Float switch 1 Maplin N34FU £15.00 1 £15.00
Float switch 2 Maplin CL25C £10.00 1 £10.00
Converted 12V Timer REUK £21.99 1 £1.95 £23.94
Diverter B&Q £5.50 1 £5.50 (now discontinued)
External Tap HR Jones DIY £6.99 1 £6.99

Payback period. Water from Dŵr Cymru costs us about £3 per cubic metre. Our house uses 100 cubic metres per year. Estimate loo flushed 10 times per day at 7 litres per time, therefore 25 cubic metres per year saving. Payback period five years.