Introduction: PVC Water Pump

Picture of PVC Water Pump

Build a long reach pvc water pump, handy for pumping right down to the bottom, and easily transfer water between rain barrels without resorting to the small plastic 5 litre paint buckets that fit the opening.

I had hoped that this small scale pumping windmill copy would start siphoning once I pulled the plunger past the T piece, alas the reducing internal diameter of the T is smaller than the internal diameter of the 50mm pvc pipe which makes it difficult to pull the plunger past the T.

Experiments have shown that the standing water level needs to be halfway up the inlet pipe, roughly 500mm high for the syphoning action to occur.

A check valve, made out of a hollow plastic ball found in roll-on deodorants, right at the bottom of the inlet pipe, ensures that you dont have to pull on the plunger like a stark raving lunatic in order to raise water.

~Now updated with a better valve~

Step 1: Parts

Picture of Parts

Not a whole lot of tools needed, I used a hacksaw, pliers, soldering iron and a M5 die for threading the 5mm steel plunger rod.

In addition you will need,

  • 2 Metres of 50mm pvc pipe and suitable pvc solvent for the joints.
  • 1 x 50mm solvent T piece. grey colored pool pvc fittings preferred
  • 1 x 50mm 90 degree solvent bend. grey colored pool pvc fittings preferred
  • 1 x 50mm solvent end cap. cut a disk out of this for the plunger backing plate
  • 1x plastic ball from a roll-on deodorant. mine was a 35mm mitchum roll-on ball. Not needed if using updated butterfly valve
  • a short length of thin wire used as a ball valve retainer to keep the ball near the seat when not sealing.Not needed if using updated butterfly valve

  • The updated butterfly valve will need a bit of duct tape, a disc of stainless steel and a brass rod for the valve pivot.
  • leather, enough for 2 x 50mm dia discs. 1 for the plunger and 1 in the cap used as a rod shaft seal.

  • 1 x 50mm steel disc for the leather retainer in the top cap
  • 50cmm 5mm steel rod. piece of recycled trellis.
  • 2 x M5 nuts for plunger.
  • 1 x top cap, a 50mm solvent end cap can be used but dont glue it on. I happened to use a cap from a can of shaving cream which was a good friction fit.

I used 105cm for the inlet, 20cm for the outlet cross piece and 75cm for the outlet down piece, I had a 40mm scrap piece for the join between the top cap and T piece.

Step 2: Full Flow Butterfly Valve

Picture of Full Flow Butterfly Valve

The check valve at the bottom is key to ease of use otherwise you'll land up yo-yo-ing with a 1.5 kg column of water, there's more efficient ways of heating water, so its kinda pointless even on a slow day.

Some thought on the matter of valves led me to this design, a double flapper butterfly valve, with a visible difference in water volume discharge.

I used a section of a 2mm brass brazing rod with duct tape as both the hinge and valve seat seal. Taking into account the weight of the full column of water, roughly 1.9kg's, makes plastic or rubber impractical and so I settled on a disc of 0.9mm stainless steel which I then cut in half to act as the valve body.

My sequence of assembly

  1. drill the pvc pipe and insert the brass valve pivot.
  2. halve the ss disc and add clearance for the brass rod, roughly 5mm should work.
  3. use a slice of pipe to align the 2 disc halves, a bit of masking tape keeps them from shifting around while you lift off the aligment slice and apply duct tape.
  4. insert the ss duct tape flap from the inlet side of the pipe.
  5. lastly insert the pvc slice valve seat and glue with solvent.

If you got the order wrong just swop ends no great problem. Before glueing in the pvc seat ring, sand it flat removing saw maks for a better performing valve.

I intentionally didnt allow the valve flaps to open perpendicuarly to maximise the opening due to the possibility that the downward flow of water wouldn't close them back against the valve seat, or might delay in doing so.

The new valve flows better and seats quieter than the old roll-on ball version.

Step 3: Plunger and Flap Valve

Picture of Plunger and Flap Valve

This draws the water column up the inlet pipe, the leather flap and disc holes makes it easier to push the plunger rod down for the next power stroke.

I used a rod 50cm long, shorter would also work but you would need more power strokes to do the same work.

Step 4: Top Cap and Rod Seal

Picture of Top Cap and Rod Seal

I wanted the shaft to be sealed but didnt want the drag of rubber so settled on leather squashed against the cap with a SS disc, no doubt thick plastic could also be used.

I usually start with short quick power strokes in the beginning to overcome the tendency of the water column to drop back down while there is still a lot of air in the inlet pipe. Once the water reaches the T piece then its relatively easy during a power stroke which displaces a fair amount of water.

Step 5: Free Flowing T-piece

Picture of Free Flowing T-piece

I have since modified the first version with a free flow pvc fitting from the pool hardware section, thinking that it was the reason why the pump didn't start syphoning, but once the standing water level is halfway up the inlet pipe then rapid power strokes to get the column of water moving results in the pump syphoning, a quick full length pull on the plunger and it seats in the top of the T piece out of the water flow allowing the syphoning action to continue.

It works much better and its easier for the plunger to travel to the top, Ive attached a pic with the 2 T pieces side by side, the grey T is the pool hardware version and the white one is a general outdoor waste plumbing fitting.

Usually the pool pvc fittings are more robust with thicker walls and better flow capabilities than the waste fittings so I prefer to use them, but its not essential if they arent available.

Step 6: Roll-on Ball Check Valve~NOW RETIRED~

Picture of Roll-on Ball Check Valve~NOW RETIRED~

I created the ball seat from sections of scrap pvc, small gaps are ok, and then inserted the ball from the top.

The plastic ball was drilled with a 3mm drill bit to fill with water to keep a neutral buoyancy and then I melted the hole closed using off cut scraps from the shaving cream can cap.


buck2217 (author)2015-07-01

Great idea so much better than trying to siphon using your mouth!!

Misac-kun (author)2014-03-10

Where do you get this blue barrel? I'm looking for one of those around my contry but i have no clue where to find em.

petercd (author)Misac-kun2014-03-10

I got it from a place called Plastics for Africa, they deal mostly in chinese plastic containers, this one originally came from India and was used to transport green peppercorns in brine.
One of my other ones had 13mm fines? from Argentina, both are +/- 220 litres capacity.

DrAmpuT (author)2014-02-24

"with thicker walls and better flow capabilities" usually would indicate lesser diameter and more friction loses. The Darcy-Weisbach Equation shows length divided by diameter as being a major portion of power loss, so anything with less diameter (or that filled with other non-open substance) only makes the total pressure drop. This is one of the prime reasons for using 1.5 in hose preferable for delivering water to fire nozzles instead of the booster hose style 0.75 or 1.0 in diameter. It seems the flow capabilities would actually be reduced given the same amount of energy expended for it. So most likely the pressure wouldn't build high enough so as to need the smaller internal diameter for a somewhat more secure construction.

astral_mage (author)DrAmpuT2014-03-05

way too much math here. just go with the flow.

vincent7520 (author)2014-02-23

Nice !

But are you sure you can siphon with such a large diameter pie. In theory yes, but then again ?…

MrE (author)vincent75202014-02-28

Umm large diameter pie, arrrllggrghhh.

vincent7520 (author)MrE2014-03-01

Sweet as a pie can be ! ;))

petercd (author)vincent75202014-02-23

I got a vacuum cleaner hose to syphon and a U pvc (with 2 x 90 deg bends) so I figured this should too, but Im now of the opinion that the little cul-de-sac at the top of the T piece prevents it.

Bill280 (author)2014-02-23

If you put the barrel up on a stand and make the outlet pipe longer so the end of the outlet is lower than the inlet pipe it will probably auto siphon. I am not sure any siphon will work if the outlet is higher than the inlet.

jack8559 (author)Bill2802014-02-23

Bill280, If you put the barrel on a stand, why wouldn't you just install a shutoff valve near the bottom of the barrel and eliminate the pump entirely?

altomic (author)jack85592014-02-24


actionjksn (author)2014-02-24

In the US our basements have sump pumps, which use a check valve which is made to go in between two pieces of "1.5" to "2" inch pipes, which should be around 35mm. This would be a good cheap source for a proper check valve. They are designed to keep the water that gets pumped out of a basement from flowing back into the sump. I'm sure they have them at the plumbing stores there, and I remember them being very inexpensive.

Also instead of using the short piece of pipe in the top of the fitting to reduce the size of the fitting to the size of the pipe. You should use what we call here in the US a street fitting. The purpose of a street T or elbow is to be able to glue a fitting directly to another fitting. This would eliminate a glue joint from the equation.

For anybody who has not worked with PVC pipe. Make sure you use a PVC cleaner of both the pipe and fitting. And before gluing it you must first use a PVC cleaner on the pipe and fitting, I prefer the clear cleaner because it doesn't stain. And after you glue the pipe and fitting you must work pretty fast, and push them together 90 degrees off from where it goes and twist the fitting a 1/4 of a turn to the proper position and hold them together and count to 30. If you don't hold them together for 30 seconds, they will push back apart. If you get something in the wrong place they will not come back apart. So make sure you have all your ducks in a row before applying any glue. A Sharpie marker can be handy. I do plumbing so I know about working with this stuff. Also read the directions on the cans of glue and cleaner.

mbeeby (author)2014-02-23

To syphon it needs the outlet to be lower than the inlet.

This is because gravity does the work once the water column in the pipe is below the lowest water level in the barrel.

You would also want a tap on the end to then cut off the flow. It probably won't remain primed unless it is oerfectly sealed.

The more height difference between outlet and inlet the faster the syphon flow will be.

petercd (author)mbeeby2014-02-24

@meeby & @Bill280, I get the whole height difference thing, as soon as the weight of the water column in the outlet exceeds that of the inlet then the system starts syphoning.

I have done the bucket on a chair and one on the ground and outlet extension as R&D experiments, but theres a whistle of air around the plunger shaft so that leak is most likely a large portion of the problem.

That said, I have devised a better flowing valve which now pumps like a demon, an 'ible update will follow shortly... and yet its still doesnt syphon with outlet extension and all.

Phimo (author)2014-02-24

It's a PUMP not a siphon. An excellent idea and one I'm certainly going to try!

vladivastok (author)2014-02-23


LP2 (author)vladivastok2014-02-23

Some people have vision problems partially blind and use caps it's easier for them to see. :-)

chuckyd (author)2014-02-23

The siphon effect only works when the outlet is below the inlet.

vladivastok (author)2014-02-23


kjsrocks (author)vladivastok2014-02-23

is your caps stuck or do you have a yelling disorder or are you just really exited about a realy simple project

petercd (author)kjsrocks2014-02-23

I think his elbow is leaning on the caps key, all his comments are in caps. :)

OldEd (author)2014-02-23

The reason that it won't "auto-siphon" (redundant: siphoning, by definition, is "automatic") is that the outlet of your system is above the inlet. Siphons work based upon the difference of pressure between the inlet and the outlet. They do have to be primed, and the tube connecting the input and output ends must be liquid and air tight, but, once primed, they will run for as long as there is fluid available at the input end.

Once the input runs dry, the siphon ceases to work. When supply is again available it will start up by itself ONLY IF THE HEAD ON THE SUPPLY SIDE IS HIGHER THAN THE HIGH POINT OF THE SIPHON. (Or if you "suck" on the output side.) If it isn't, as would be the case with your apparatus, it must be pumped to be primed.

Your "back-flow preventer" (the deodorant dispenser ball) is call ed a check valve. The purpose of a check valve is to permit the pump to cycle completely - i.e., not "unpump" on the down stroke of the piston. In a siphon a check valve will block free flow once the system is primed. If the head of the fluid in the outlet side is greater than the head of the fluid above the check valve the fluid in the loop will vaporize and the prime will be lost.

In the case in which you wish to preserve the prime, so the system is auto-starting, a float-operated valve (a "float valve", such as is used, in a slightly different form in tank-type toilets) can close the OUTPUT side of the siphon and preserve the prime.

You should be aware that siphons, by their very nature, don't care which way they operate: i.e., they can siphon the "source" to the "sink", or, with equal alacrity, siphon the "sink" back to the "source". It is thus necessary to provide some sort of "back-flow preventer", or check valve, in such systems.

After reading this long-winded comment you may very well ask "howcum" I know so much about the subject. I have two reasons for you.

The first is four+ years of engineering education.

The second is that I experimented for several months as a young man - still in grade school - trying to build such a system so as to be able to avoid the onerous task of watering the live-stock (sheep, two goats, and a flock of chickens) before I went to school, when I got home from school, and before I went to bed.

I spent almost all of my free time from April through October attempting to devise such a system. In this I had the "help" of my younger-by-2-and-a-half-years brother, who anticipated more free time once "we" were successful.

I never made it work satisfactorily. Our parents and uncle who had watched all of our labors with ill-concealed humor, (and some financial help, once they were convinced that I was serious in my efforts) finally caved in - read "took pity on us" - and for my birthday I was presented with an electric-motor-driven pump, a suitable time clock (it could be set to several times during a 24-hour cycle), float-switches, electrically operated valves, and sufficient wire and flexible pipe and fittings to build the system.

Russell McMahon (author)2014-02-23

Looks useful.

Re auto-siphoning
(1) If you added a side piece at the bottom for the entry with the check valve in it and continued the tube down past it slightly you could (probably) enable auto siphoning. The same need for no restriction to the plunger applies as at the top BUT whereas the top T is structural, here the side pipe has almost no force on it (just the weight of a non return valve) and you could solvent cement a short pipe section at 90 degrees onto the main tube without using a joiner. (Cut a hole in the vertical tube 1st :-) ). Solvent joined pipes can be very strong.
(2) You could consider a tee at the top end that was oversized internally so the vertical tube could "just" push through - and would then need to work out [tm] how to seal the joins :-)

redmarkonthewall (author)2014-02-22

You could also use a bouncy ball for the check valve. It may make for a better seal.

petercd (author)redmarkonthewall2014-02-23

I gave this a try and it sealed very well with hardly a drip, unfortunately the down force of a 1.5kg column of water forces the ball into the seat making it much harder to pull the plunger on the upstroke.

Thanks for the input, its appreciated.

DDW_OR (author)petercd2014-02-23

I live in an area that freezes so i think i will use a solid ball instead of one tat may freeze.

petercd (author)DDW_OR2014-02-23

I didnt have much luck with a 35mm glass marble either, makes quite a noise when dropping down on the seat and its too heavy, which coupled with the weight of the water column inside the pvc pipe means you have to really haul on the plunger and a lot less water is drawn up the pipe as a result... so your options are diminishing somewhat.

bendingit (author)2014-02-22

Great design,thanks for posting!

Ironwave (author)2014-02-22

Nice Petercd!

rogue_saint_michael (author)2014-02-21

Looks great! I was looking for something pretty simple for this idea. I just couldn't put my finger on it. Great design and instructions. Tthis is so simple that even I can make it! :)

rimar2000 (author)2014-02-21

Good design, simple and effective.

About This Instructable




Bio: general bloke type of tinkering
More by petercd:Total Metal Hypercube 3D Printer (no Printed Parts)Hack a microswitch for better 3D prints. (ABL)Laser center finder. (drill press and mill)
Add instructable to: