My DIY rainwater harvesting project

Benster

I've been putting together a rainwater system to feed the toilets in our house and I thought I would share the how and the why, partly to pass on what I did and partly to get feedback and see if maybe there are better ways of accomplishing what I did.


Background

I'll keep this bit short as most folk will probably just want to know how it is put together. The original idea for this came from occasional but frustrating water outages due to our very low water pressure. We're at a height in Ashbourne and the pressure is just on the threshold of acceptable normally, but if there is a long
period of low rainfall then our pressure drops and the water doesn't even make it up to the tank in the attic. This has happened at least twice a year.
As well as that, when the water tax was confirmed as coming down the line, I thought this was a no-brainer and decided to pull the trigger on the project.
We have three toilets in all, and this project would convert two of those to use rainwater, the upstairs toilets. The downstairs toilet is too difficult to pumb in as
the supply pipe is hidden in the wall somewhere and it also supplies the hand basin. That would have been a major job involving pulling plasterboard off the wall and
that's somewhere I did not want to go. Plumbing in the upstairs toilets would (only!) consist of bringing a new pipe down from the attic and splitting that to supply
two adjacent rooms.


The Build

I have to start off by saying that I was not sure at the beginning if this was really going to work, so I didn't spend big on anything. Also, it was going to be a staged build as I wanted to see if one part worked properly before building the next.

The basic parts are:

- drainpipe diverter
- filters
- storage barrel(s)
- 2nd attic tank
- submersible pump
- pump controller
- mains backup mechanism
- associated plumbing parts


The collection starts with a square drainpipe diverter that links to a length of pond-pipe. Due to space restrictions, the only place the tanks could go was in the corner farthest from the drainpipe. That meant the pond-pipe was going to have to run the whole length of the back wall. Not pretty, but it was the only way.


[URL="[URL=http://pix.ie/yerman/3736305][IMG]http://photos3.pix.ie/9B/E1/9BE1AFC6AE1D43B89B975B60242D9606-0000314530-0003736305-00640L-9304794C45C94B96BA920435A8C958FC.jpg [/IMG][/URL]"][/url]


Within the diverter, I put the first stage of filtration, a wire mesh funnel. The idea is that lumps of moss etc would hit that and slide off down the drainpipe; water would seep through the mesh and enter the connected pipe. The next stage is filter no.2, one that is normally fitted to a pump, but I have it on the divert downpipe. This one is easily accessible and can be cleaned out when needed, all it needs is a scrub. I also have the filter wrapped in a black pastic bag to stop algae growth as that makes cleaning harder and needing done more regularly.


[URL="[URL=http://pix.ie/yerman/3757433][IMG]http://photos4.pix.ie/48/01/480116B0086D4F8E9F64A3489AB803A5-0000314530-0003757433-00695L-6084074AB05F49C8AE9E77DA9ED1D891.jpg[/IMG][/URL]"][/url]


Finally, the third filter is an idea I got from a Youtube video. It's a milk filter sock, the type used by dairy farmers. I have it stuck on the end of the downpipe. These fabric tubes are cheap and bought by the 100. I got these in an agri-suppliers for less than a tenner. I don't know the micron-level of filtration, but they catch the smallest bits of dirt that pass through the 2nd filter and the sock needs changed about once a month.



[URL="[URL=http://pix.ie/yerman/3736312][IMG]http://photos3.pix.ie/97/76/9776B29BD3B548299159BC6D2C63C231-0000314530-0003736312-00500L-C01D0B45B5D74D62B259DBF250E46B70.jpg [/IMG][/URL]"][/url]



The collectors are two 210-litre water butts I got in Aldi, but I see their design in garden centres all over the place. I have them connected at the bottom and also anchored to the wall. They are some weight when full and I don't want kids to get hurt it they somehow were to topple over. 420 litres sounds a lot, but for this purpose, it actually isn't. More on that later.


[URL="[URL=http://pix.ie/yerman/3736309][IMG]http://photos3.pix.ie/85/40/8540A8C631E940ACB4510048B90FC3A2-0000314530-0003736309-00500L-307F9BCF3F784E04B50A0D2301723972.jpg [/IMG][/URL]"][/url]


So how to get all this clear water up to the tank that awaits it in the attic? I have a submersible pump (also purchased in Lidl) suspended in tank 2, connected to a looooong piece of qualpex pipe that goes directly to the tank, without any joints inbetween. This is a 240v pump, and the power cable also goes all the way up there to a socket.


[URL="[URL=http://pix.ie/yerman/3736301][IMG]http://photos3.pix.ie/D6/10/D610AB63C28B4528B4BE1472A0A9F6E4-0000314530-0003736301-00640L-F2D41FB5BA204E95B96A27BA9759D671.jpg [/IMG][/URL]"][/url]


The controller for the pump is worth a separate post on its own, but I'll summarise it here. Having a technical background, I wanted to build a controller myself, so I started with a Raspberry Pi board. On this board are connector pins that can link to sensors to detect water levels and relays that can turn on power. I bought float sensors and put them in the collector barrel and the attic tank and wired them to the Pi. With those in place, I could then detect if there was enough water in the barrels, and also when there was not enough water in the attic tank. Also attached to the board's pins was another board of relays that would connect the Pi's low voltage control to the pump's high voltage supply, enabling me to turn the pump on and off at will. Using a scripting language (Python), I then wrote the logic that would control when the pump had to turn on and, more importantly, when it should turn off. Because it's my own design and I know the programming language, there is a lot of flexibility in what I can do with the controller. At the moment I have it sending me emails of how each pump session went, and it also tracks how much water it has pumped.


[URL="[URL=http://pix.ie/yerman/3736307][IMG]http://photos3.pix.ie/1D/9E/1D9EED64FB3149AC9F3945730980EAF1-0000314530-0003736307-00640L-0FEFC498F4DC41269639A3F95E28B547.jpg [/IMG][/URL]"][/url]


And what if it doesn't rain for ages, like Summer 2014? What is going to supply the toilets then? That's where the mains backup comes in. There was no way we could rely solely on rainwater so a decent mains backup would make up the shortfall. Initially I wanted to have this as a mechanical backup, akin to a ballcock in a toilet cistern (techically called a part-fill valve). I wanted to make sure that in a power outage a non-electrical backup would still be functional. But the part to make that work came to almost the same as had been spent on the lot so far and I just couldn't justify it. (part-fill valve example) So I had to rethink that bit and reluctantly went with having the mains backup also controlled by the pump controller Raspberry Pi. I fitted an electrically-controlled water valve to a pipe Teed off the mains supply to the original attic tank and added the logic to the controller's program so it would open the valve if there was not enough in the barrels and the water level was low in the attic tank.


[URL="[URL=http://pix.ie/yerman/3736310][IMG]http://photos3.pix.ie/54/2C/542C02ED42E04BB1A509F62F7CD0E0ED-0000314530-0003736310-00800L-8B0F7BBC944B490B9A73F37288601404.jpg [/IMG][/URL]"][/url]



Results

---

So after all this, a project which on and off took most of last year, it has worked mostly as I planned, apart from a couple of occasions. I have yet to analyse the stats in detail but I would estimate perhaps about a third to a half of our toilet flushing needs have been met by rainwater. That includes having a few extended periods with no rainfall where we live too. Mains backup kicked in then and kept the attic tank at a minimum level. There have been two occasions when the tank ran dry though, causing an airlock in the cistern supply pipes. Once was early on when I had programmed just two tank-check times a day, when the controller would check the float sensor levels and work out if a rainwater or mains fill was required or not. I quicky learned that a morning and an evening check was just not enough for normal use, so I changed that to a check every two hours in daylight hours. Another time, we had visitors staying over and the tank was also on mains backup at that time. With higher demand from more people, it quickly went below the minimum level well before the next check. I had to do a forced manual mains-backup fill using another program similar to the main check script, running the mains backup for longer until it reached the float sensor at the top of the tank that would signal the cut-off.


Improvements to be made

- A bigger storage volume is required. I plan on moving up to an IBC tank holding 1000 litres next.

- A more intelligent control algorithm. There are a few things to change here, now that the system has bedded in. At the moment it is a bit on the paranoid side as I do not want to flood the house if something goes wrong. But it needs to have a better mains backup fill method than it has at the moment.

- Currently only the drainpipe at the rear has a diverter, so I'm missing the water that falls on the front part of the roof. That will be a bit of a challenge as the diverter will have to be even higher than the rear to keep the water moving by gravity.

- there's probably more and I'd appreciate your feedback. I know others have put their own systems together, so I'd love to swap tips too.



Cheers,
B.

Benster

Some more photos of the filter socks...

[URL="[URL=http://pix.ie/yerman/3736304][IMG]http://photos3.pix.ie/49/06/4906D8EC6B0D4509BE3BE77EAF273173-0000314530-0003736304-00800L-3CDD9AD639D44611BC9ECE039BD7FF2E.jpg[/IMG][/URL]"][/url]


A clean one fitted...

[URL="[URL=http://pix.ie/yerman/3757435][IMG]http://photos5.pix.ie/5D/0A/5D0A52BC454F4A20A9D8682252A92E2F-0000314530-0003757435-00640L-CFF64CE2689E4844A14F2C99857B838A.jpg[/IMG][/URL]"][/url]

phester28

will the pump not get damaged in the winter time if the water freezes?

Benster

phester - It's a possibility and I had intended to wrap the tanks in insulation initially. Over the last winter there was a small bit of freezing with about 1/3 cm of ice but it didn't go near the pump at that point.

As I alluded to in the post, I am ditching the tanks now and going with an IBC tank, so I'll have to look into a bit of insulation for that too maybe.

B.

phester28

just a thought but would not a ball cock set much lower in the tank provide a simple way to ensure that you never run dry while maintaining your priority on grey water?

Re the PI: I assume that you used it because you are more familiar with pearl. Would not a simple Ardurino (C complied) be sufficient as I assume you are only using digital i/p and o/p?

Great work and keep up the homebrew

Calahonda52

phester28 wrote: »

just a thought but would not a ball cock set much lower in the tank provide a simple way to ensure that you never run dry while maintaining your priority on grey water?

This idea is illegal as it runs the risk of grey water getting into the rising main

Benster

phester28 wrote: »

just a thought but would not a ball cock set much lower in the tank provide a simple way to ensure that you never run dry while maintaining your priority on grey water?

Re the PI: I assume that you used it because you are more familiar with pearl. Would not a simple Ardurino (C complied) be sufficient as I assume you are only using digital i/p and o/p?

Great work and keep up the homebrew

Yes, as Cal says, having a mains outlet submerged in what is essentially dirty water is a no-no from a building regulations point of view. I did try initially to use a ballcock up high at the same level as the RW inlet, but the arm length was too short to reach down to near the bottom of the tank. I thought of ways to work around that like extending it somehow (weld another on?) but I thought that it could easily end up a bodge job and I didn't want that for such an essential part of the system. Also, a ballcock set to that depth would mean the arm would be under a lot of stress when the tank would be full and that could also cause a problem at some point.
What I really would like is a reasonably priced part-fill valve like I linked to the original post. The only Irish place I found selling them had them at €150, too dear for my particular project.

As for the control system, if you're comfortable and handy with an Arduino that would do just as well. All it needs to do is read sensor inputs and open relays, well within its capabilities. Reuk.co.uk has even simpler boards that they make just for this exact purpose.

And thanks for the encouragement!

Calahonda52

I meant to add my good wishes to the project.
What follows may not be practical but given that you are relying on gravity to fill the toilets, then if you put a second small tank in at the same level as the RW tank and connect both to a three way Motorised valve which u can control with your electronic skills via float sensors.

These valves are not duck tight so you may get some flow from the mains tank when using RW but not a lot given the low head if both tanks are same level.
You can also throttle back the flow from the submersible RW pump to just exceed the toilet refill rate, this will mean that your RW header tank overflow should work easily.
I have a similar pump doing other stuff and I put a bypass on it so as the pump is not pumping against a reduced orifice in a gate valve but of course u could probably rustle up an VSD!!
Keep well

ps maybe 2 simple valves would work easier:)

What I m

North West

Hi Benster.
Your DIY rainwater system as described is a highly dangerous system to human health. The idea of saving rainwater is good. The system is flawed. Under no circumstances can you mix rainwater with mains water. Birds dropping alone from your roof contains about 50+ contaminates which when disolved in water from your roof is going straight to your tanks. These contaminates can seriously harm humans maybe even kill someone. Google bird droppings and rainwater.
It's not as simple as you think. It has to be designed properly. I recommend you do a lot more research on this topic and in the mean time seperate the rainwater system from the mains system for safety and health reasons. My apologies for taking the wind out of your sails. You must design a system safely.
NW

Benster

North West wrote: »

Hi Benster.
Your DIY rainwater system as described is a highly dangerous system to human health. The idea of saving rainwater is good. The system is flawed. Under no circumstances can you mix rainwater with mains water. Birds dropping alone from your roof contains about 50+ contaminates which when disolved in water from your roof is going straight to your tanks. These contaminates can seriously harm humans maybe even kill someone. Google bird droppings and rainwater.
It's not as simple as you think. It has to be designed properly. I recommend you do a lot more research on this topic and in the mean time seperate the rainwater system from the mains system for safety and health reasons. My apologies for taking the wind out of your sails. You must design a system safely.
NW

I'm not mixing the RW into the original mains tank:

Benster wrote: »

The basic parts are:

- drainpipe diverter
- filters
- storage barrel(s)
- 2nd attic tank
- submersible pump
- pump controller
- mains backup mechanism
- associated plumbing parts

Also, the pipe feeding the mains backup to the new attic tank has 2 non-return valves on it and the outlet at the end is 2 inches above the overflow outlet. I have thought quite a lot about how this system is put together as I do not want to harm either a person or the home I live in.

Calahonda52

North West wrote: »

Hi Benster.
Your DIY rainwater system as described is a highly dangerous system to human health. The idea of saving rainwater is good. The system is flawed. Under no circumstances can you mix rainwater with mains water. Birds dropping alone from your roof contains about 50+ contaminates which when disolved in water from your roof is going straight to your tanks. These contaminates can seriously harm humans maybe even kill someone. Google bird droppings and rainwater.
It's not as simple as you think. It has to be designed properly. I recommend you do a lot more research on this topic and in the mean time seperate the rainwater system from the mains system for safety and health reasons. My apologies for taking the wind out of your sails. You must design a system safely.
NW

You really do need to read, and understand the post before you pen such material which does a great disservice to the OP and his careful considered project.

daveville30

A float switch in the barrell would be a good idea to stop the barrell overfilling.have one in my boat when the water gets to a certain level it starts pumping.

Benster

daveville30 wrote: »

A float switch in the barrell would be a good idea to stop the barrell overfilling.have one in my boat when the water gets to a certain level it starts pumping.

Hi Dave, I have an overflow in one of the two joined barrels that empties into a drain for that exact purpose. The logic of when to pump from the barrel to the attic is a bit different than a boat though. In my case it's the attic tank level which is the driver; only if that is below the full level should the pump in the barrel fire up. In a boat I assumed you're just pumping bilge water overboard?

Counsellor Dublin

wow thanks I always had an interet in rainwater harvesting. great post

daveville30

Yea out through a hose overboard.I was thinking would it work putting a gutter on the gable end sloped part off roof and have a tank to collect water in the attic.

antodeco

Very impressive build! I hope it all works out in the end!

Benster

daveville30 wrote: »

Yea out through a hose overboard.I was thinking would it work putting a gutter on the gable end sloped part off roof and have a tank to collect water in the attic.

That sort if plan would very much depend on the shape of your roof. My gutters are just below the eaves so a tank in the attic is higher up than that, thus gravity wouldn't allow the collected water to flow in. But more important than that, you need some sort of large primary storage area for what you collect, to enable you to have water to use even when it doesn't rain for a while. Realistically you're talking 1000's of liters and that volume weighs over a ton, so keeping that in the attic is not possible unless you have designed it specifically to hold that weight. I added extra support to the existing joists just for the RW tank and that only holds 270-odd liters.
Also (II), I wanted the attic tank to be sealed and not need any maintenance; having to open pipes to clean filters is not something I think should be done in an attic.

Again, thanks for the interest and encouragement, folks. More details to come soon on the pump control system.

thejohner

I am thinking of installing a similar system but intend using a ballcock for mains supply. The arm could be bent down bringing the ball to a depth of ca. 6 " . this will also decrease the lever pressure on the valve . If the pressure is still too high then a smaller ball will work. A careful sizing of the tank would ensure minimum use of fresh water. Great post.

cdaly_

Benster wrote: »

I did try initially to use a ballcock up high at the same level as the RW inlet, but the arm length was too short to reach down to near the bottom of the tank. I thought of ways to work around that like extending it somehow (weld another on?) but I thought that it could easily end up a bodge job and I didn't want that for such an essential part of the system. Also, a ballcock set to that depth would mean the arm would be under a lot of stress when the tank would be full and that could also cause a problem at some point.

thejohner wrote: »

I am thinking of installing a similar system but intend using a ballcock for mains supply. The arm could be bent down bringing the ball to a depth of ca. 6 " . this will also decrease the lever pressure on the valve . If the pressure is still too high then a smaller ball will work. A careful sizing of the tank would ensure minimum use of fresh water. Great post.

This is quite a late response but you could do a fairly simple low-level float linked to a high-level ballcock thus:

• Fit a ballcock valve higher than overflow outlet.
• Instead of the ball, fit some kind of pivot to the end of the float arm.
• Mount the float to a length of brass rod and fit that to the pivot on float arm.
• Secure a length of wavin pipe vertically in the tank and put the float inside that.
• Fit some kind of stopper across the top of the wavin pipe to limit the upward movement of the float arm. This prevents excess pressure on the float valve.

In this arrangement, the float will move vertically in the wavin pipe based on a very low water level in the tank. The pivot will transfer that movement to the float valve. The stopper at the top of the pipe will prevent excess pressure on the float valve. In normal operation, the float valve remains closed. In water shortage or power failure conditions, the valve will open to prevent the tank from emptying.

John T Carroll

cdaly_ wrote: »

This is quite a late response but you could do a fairly simple low-level float linked to a high-level ballcock thus:

A neighbour enlisted my help for a much less ambitious project not too long ago,(I,m always in demand for this sort of work). There are 6 people in his house and he has upstairs+downstairs toilets, he says the downstairs one is used by everyone 90% of the time so we just looked at a gravity fed system to the downstairs toilet fed from a 1000 Litre IBC which he already has. His downstairs toilet is fitted with a combined cistern/toilet with the cistern top 30 ins from the ground, the cistern has a bottom entry filling device with no ball cock, these are quite common now. We went away and bought a bottom entry Opella Delcham ball cock which comes fitted with a 1/8 ins nozzle so we asked for and got a 1/4 ins nozzle and fitted this in his cistern. We decided that the minimum head required from the bottom of the IBC to the cistern top as 12 ins and then carried out a few flow tests at this head with the ballcock fully down (open) and the nozzle flowed 1 litre in 28 secs so the cistern should fill in 5/6 mins max, as the level in the IBC will normally be well above this, in practice it means that the cistern should fill in a few minutes ready for its next client!. He has plenty of room at the side of his house right adjacent to this toilet so if he supports the IBC (bottom) 3.5 feet from the ground and securely straps it to the wall to stop it toppling over then he should be in business. Thanks for all the ballcock filling designs, I'll pass them on to him.