Author Options:

How do I design an eductor-jet or venturi system to rapidly remove the sand from a pool filter? Answered

A friend of mine regularly changes pool filter sand for his clients and asked me to build a more effective way of doing this. So far I think an eductor would work best check this out: http://en.wikipedia.org/wiki/Eductor-jet_pump

1. be able to remove a large quantity of sand suspended in water - approximately 50kg (110 lbs)
2. use a pump to supply the motive force - 3/4 hp pump is readily available
3. the pump MAY NOT VACUUM OUT THE SAND DIRECTLY! that would ruin the pump's impellers instantly. It is essential that the pump only provides a motive force in the form of a fast-running jet of water which will create a venturi effect that sucks the sand/water slurry out of the filter.
4. the "vacuum hose" must be between 700mm and 1m long (27 - 40 inches) and about 1inch diameter.
5. The eductor/venturi must be strong enough to suck a sand/water mixture all the way up that pipe against gravity.

Check the attached diagram and tell me if my venturi design looks reasonable, also what diameters should I have at points 1 - 6?

I also attached a sketchup model of what I'm planning to build

Facilities at my disposal: Welder, drill press, cutoff saw, dremel, various grinders, jig saw.


OK, that's about it. Your commentary is appreciated!


see some sharp minds! I wish to dig in clay about 300 feet down maybe more. i was thinking injecting water in a larger pipe 4" in diameter that i call #7 down with a smaller pipe inside suck up material clear path for the pipe going down look at 4" # 7 inner pipe #6 be 2" with a 3" to 2" reducer at the bottom to increase water pressure at the bottom of pipe #7 be sucked up pipe #6.

I realize this is way old, but for those googling, a geyser pump might be a simpler solution. Sewage treatment plants use them to move sludge. I'm just saying.


6 years ago

The drawing is a good start for the project and there can be some details added to improve the eductor efficiency. Eductors use a high speed jet to mix with the fluid that needs to be pumped.

The suction point is where the jet enters the mixing region, indicated on the drawing at the 3 to 6 point. If there is a convergent nozzle (not shown in the drawing) the jet, the suction point, and the mixing region should be upstream of the convergent section.

If there is a diffuser (expanding or diverent as shown in 3 to 4), this should be downstream of any convergent section. The drawing indicates a convergence for the jet but not the mixed stream. The diffuser causes the local velocity to decrease and the local pressure to increase. The eductor in the drawing with the suction connected directly at the entrance to the diffuser (6 to 4) would be inefficient. It would be more efficient to have a concentric pipe-in-pipe arrangement or an actual mixing annulus around the jet with a convergent section (and then a diffuser) downstream.

Having built several over the years out of PVC and ABS, here is the simple answer. Buy a "gold dredge suction nozzle". 2" and smaller is about 100USD.

Or do a web search for the same to see what they look like. Forget about the proportional mixer type venturis ala carburetor type.

The curved section is the high pressure line coming from the pump. 3/4" garden hose will run a 2" nozzle as long as you have pressure, the more the better. On the end of the pressure line where it enters the 45 weld, braze or epoxy a washer for an orifice. Rule of thumb is 1/4 diameter of the pipe or in the case of 2" use a washer with a 1/2" hole.

ABS and PVC nozzles last for several hundred cubic yards of gravel while nozzles made from exhaust pipe will last a few thousand yards.

The tapered venturi is only for non-abrasive materials.

I like to use ABS due to the welding by the solvent (glue). A trip to the hardware store will set you up. The only reason I don't make them from steel is the hassle of jigging before welding and cutting the hole for the pressure line orifice. This needs to be centered. If I have the ABS on hand, I can make one in under 15 minutes.


6 years ago

The most important aspect of of an Eductor, as I learned in US Naval Pump and Compressor school, is the shape of the volute and the spray going into it. We had pumps providing 150psi/900gpm to our Eductors and we could de-water a compartment the size of a house in under an hour...the suction side was 10" pipe.
Scaling everything down to garden hose size shouldn't be that hard. Industry uses solid metal or cast Eductors that small. Using PVC pipe and your town water pressure should be enough. It's the shapes I spoke/typed of earlier that's important. Look at this pdf:


thanks set271, my friend I and have given up on that project for now but your information is interesting all the same and for that I thank you. The theory behind it is fascinating, it's one of those areas of physics that never fails to leave me in awe. Must say that clearing a house in an hour seems friggin' incredible! The military is definitely one place to go if you wanna see cool stuff.

I too would tend to go for suction rather than pressure, unless the distance to be lifted is too great. A filter and/or settling system placed before your fan or pump really should be able to protect it adequately. (Or you could try the wet-vac trick -- sealed container, evacuate air to draw in the water, periodically dump container and repeat.)

But if you want to go with pressure... Hmm. You can buy cheap venturi submersion pumps intended to hook to garden hose, and believe it or not I have seen one used to empty a 20'x40'x8' swimming pool (slowly!). I don't think they'd hold up very long in this use, but they're essentially disposable...

How do the pros solve this problem? Your friend can't be the first to deal with it on a regular basis...

(My experience with pool filters is mostly the old-style low-velocity-big-tank sand filters, of the "bury it in the ground, fill it, and hope you never have to open it again" variety. Backwashing on a regular basis seemed to keep those reasonably happy...)

.  While it is very possible to build your own eductor, it not a job for the average DIYer. You have to maintain close tolerances to get maximum efficiency or you will end up wasting a LOT of water, fuel, time, &c (= money). For pumping abrasives, such as sand, you need to use very hard materials (ceramic would probably be a good choice for your application), which make working the eductor even more difficult for a DIYer. If this was just for cleaning your own filter once or twice a year, I'd say go for it, but, for your friend's uses, I recommend that you/he buy one - it will be much cheaper in the long run.
.  You can also buy pumps made specifically to move "trashy" streams. Try searching for "sand pump", "trash pump", and "mud pump".

Go for a SOFTER plastic, and you get less wear though sometimes.

.  True, but soft plastics, such as PTFE, can be as difficult to machine as harder stuff. No big deal for someone with experience and the proper tools, but not an easy job for the average DIYer.

Thanks for the replies NachoMahma and steveastrouk. I was thinking steel, I'm good at fabricating steel, I love steel. Close tolerances? I could potentially call in a favour or two and get a someone to machine the inner diameters of a cone that I have already welded up.

My questions are:

1. what tapers work for the motive and discharge areas?
2. What diameters? (although i suppose that depends on pump strength and the volume I need to move)
3. Is my diagram ok? would that produce any suction AT ALL? or is it more of a jet engine?

. All the eductors* I've seen had a different design than what you show. The ones ones I used were a venturi (often with a nozzle to boost velocity) with a port at the low pressure point (near the middle of the "throat").
.  I'm no engineer, but your design looks like it would work, but would be very inefficient.
.  As for dimensions, that's over my head. I've used a lot of them, but never had to design one. Somewhere around 1-1.5" (inlet/outlet/suction) should work if you use a store-bought eductor.

* Just an example, not a recommendation.

. PS: at the bottom of the eductor link are some links to pages on spec'ing eductors. You should be able to get some ballpark figures there.

Don' t think its a job for steel unfortunately, its going tp wear REALLY fast.

You need a high volumetric flow to get the pressure reduction, so your diameters can't be too big, I'd reckon on an inlet around 1".


I would suggest looking into purchasing a commercial eductor, like those used for certain toilets. While from an academic perspective it's interesting, it seems like a waste of time and money to design your own, since the work has already in large part been done, and designing a decent eductor/ejector is not childs' play (not saying you're a child!)

I would personnally recommend a shop-vac style air pump that just drives low pressure suction into a container (like an old oil drum, cheap).

At the end of the suction hose put a few water jets to soften/break up the sand. We recently did this to dig a hole down to our sewage line out of the house (instead of getting a traditional large hydrovac truck). Worked like a charm - just be sure to make it so the motor isn't spending a lot of time 100% bogged down with no airflow, or it'll overheat and you'll lose your vacuum.

Turned is probably the simplest approach.
I'd do it from solid PVC.