Magnetically Coupled Water Pump




Introduction: Magnetically Coupled Water Pump

About: I like Physics and electronics. Enthusiast of projects related to these topics. I also like aerial photography with the use of drones.

In this INSTRUCTABLE I will explain how I made a water pump with magnetic coupling.

In this water pump there is not a mechanical connection between the impeller and the axis of the electric motor which makes it work. But how is this achieved and what motivated me to give this solution? It was possible by applying the principle of attraction and repulsion that naturally occurs between magnets. I was motivated to carry out this project because I needed a Modular Water Pump, to which I could easily change some of its characteristics such as the shape of the impeller blades, its radius, material types etc , and check the results that derived from these changes, maintaining the same electric motor and voltage. At first, I started building traditional centrifugal pumps, but I faced multiple problems of water leaks (between the electric motor shaft and the impeller). Coincidentally these days the YouTuber GreatScott (great experimenter and whom I admire) has had similar problems as stated in this video.

If magnets are attached to the electric motor shaft and also in the impeller, perhaps it could be turned and propel water, even if there is no mechanical connection. This idea was what sparked my interest to carry out this project that I hope you find useful.

The experience I gained during the completion of this project has allowed me to conclude that there are many practical applications for these principles not only in the field of hydraulic pumps.


Plexiglass sheet of at least 200mm by 150mm, 6mm thick (used to make the impeller cavities and the electric motor coupler).

Two sheets of 80mm by 80mm plexiglass, 4.5mm thick (used to make the impeller and the DC motor magnet holder).

Plexiglass sheet 200mm by 150mm 4mm thick (for electric motor mounts).

Two M3 screws 8mm long and corresponding nuts (for the union of the electric motor with the coupler).

Six M4 screws 20mm long and 2 corresponding nuts (for the upper and lower union of the impeller cavities).

Two M4 spacer nuts 18mm long.

Two female banana type connectors for chassis

A power switch.

An electric motor 40mm in diameter and 55mm in length, 24V direct current (DC) with 5mm diameter shaft.

Instant glue, epoxy or similar.

Neodymium magnets 12mm long, 2mm thick and 4mm wide.

Electric soldering iron and cables for electrical connections.

Permanent black marker.

Phillips screwdriver.




CNC milling machine with work area of at least 300mm by 200mm.

Endmill 1.5mm cutter

Flexible water hose 8mm outside diameter and at least 250mm in length.

Water containers

Cable ties.

19V or 24v direct current source

Teacher Notes

Teachers! Did you use this instructable in your classroom?
Add a Teacher Note to share how you incorporated it into your lesson.


The magnets that were used in this project were extracted from a brushless DC motor. With the help of a flat screwdriver I put a little pressure on the base of the magnets and one by one I managed to take them off. At first I thought it would be very difficult, but the truth is that it was not. In the end you will get a set of magnets that have been accommodated according to the principle OPPOSITE POLES ARE ATTRACTED AND EQUAL REPELLED. With the help of the compass begin to mark the poles of each magnet separately. If you make an imaginary and horizontal cut to each magnet one face will be NORTH and another SOUTH in this type of magnets


The impeller with magnet holder was made from a single piece of 80mm by 80mm of Plexiglas. This required having to make two-sided cuts. In the cuts of ALL the pieces a milling cutter ENDMILL of 1.5mm of diameter was used. Plexiglass sheets are ALWAYS larger than the cuts to be made so that you can fix it correctly to your work table, leaving a margin for it.

The method I used was the following:

First the cavities for the magnets are made and a through hole located 5mm by 5mm from the origin of the coordinate axis of the plexiglass and the CNC machine.

Second, a 50mm by 50mm square cut is made to the entire depth of the material, thereby detaching the piece.

Third the piece is inverted and glued with instant glue in the same position it occupied in the first cut, but with the opposite side facing up (use possible marks left by the cutter in the scrap table. It is verified with the help of the reference hole that the part was stuck in the correct position (If the position X = 5mm, Y = 5mm and Z = 0 is executed in the control software of your CNC machine, it must match exactly with the beginning of the reference hole) .

Fourth, the cutting of the impellers fins is executed and the central and through hole of 5mm diameter is made.

Fifth the round cut is executed to the whole piece and gets detached from the rest of the Plexiglas material

Step 3: Glue the Magnets to the Impeller

Do you remember in step 1 when we identified the polarity of the magnets? Now it is the time to use this knowledge. Place a small amount of instant glue in the first cavity of the magnets and then the first magnet. Hold it in that position for a few seconds until the glue works. Depending on how you have placed the magnet you will have a NORTH or SOUTH face up, the next magnet will go with the opposite face up. PLEASE VERIFY THAT YOU PERFORM THIS PROPERLY, IT IS CRUCIAL FOR THE SUCCESSFUL DEVELOPMENT OF THIS PROJECT.

At the end and after repeating the previous step 6 times you should see something very similar to the photo I am
showing here.

Check again with the help of the compass if the magnets alternate their polarity. THERE Shouldn’t be TWO MAGNETS CONTIGUATED WITH THE SAME POLARITY.

It is important to clarify that the magnets should not exceed the surface of the plexiglass, so the amount of glue used should be moderate.

Step 4: Machining the DC Motor Magnet Holder

The DC motor magnet holder was created from a piece of 80mm by 80mm Plexiglas. The DC motor magnet holder is responsible for transmitting the torque to the impeller when it interacts magnetically with it. First the cuts of the cavities for the magnets and the central hollow are executed, then the external circular cut should also be made. In my case the motor shaft had a 0.5mm chamfering and was considered in the vector drawing. In the event that the electric motor you use does not have it, use the 5mm vector circle found in the final step.

Step 5: Glue the Magnets to the Magnet Holder

The same principles stated in step 3 apply here. Place a small amount of instant glue in the first cavity of the magnets and then the first magnet. Hold it in that position for a few seconds until the glue works. Depending on how you have placed the magnet you will have a NORTH or SOUTH face up, the next magnet will go with the opposite face up. FOLLOW THE RECOMMENDATIONS EXPOSED IN STEP 3


It is very likely that you have to transform the vector drawing of this piece depending on the characteristics of the electric motor you use. The function of this piece is to fasten the assembly of the impeller to the body of the electric motor, achieving a separation between them. In my case, I machined the piece from a 200mm by 150mm and 6mm thick plexiglass sheet from where I cut the Impeller cavities. The body of the electric motor used has two threads for M3 screws, so two of the holes in this piece are for M3 screws and two for M4.


The DC motor magnet holder must be securely attached to the electric motor shaft and completely perpendicular to it. In my case it was convenient for me to place it on the shaft, apply some instantaneous glue on the joint, wait 20sec and apply a voltage of 5V to the electric motor, making them turn at low revolutions and wait for the assembly to dry. With this I managed to make the magnet holder perpendicular to the axis. DO NOT OVERCOME WITH THE QUANTITY OF GLUE, WHEN THE SYSTEM BEGINS TO ROTATE THE GLUE WILL BEGIN TO SPREAD ON EVERY SIDE (TAKE CARE OF YOUR EYES)


The support system I designed is quite simple and only requires four cable ties to attach it to the electric motor. In one of the bases the cavities for the switch and the banana connectors were made. They were cut from a 200mm by 150mm and 4mm thick Plexiglas sheet.


The impeller cavities were obtained from a 200mm by 150mm Plexiglas sheet of 6mm thick. The FEED RATE was set at 200mm per minute. This is the process that consumes the most time (about 25min per face). If in any case you notice that the 1.5mm diameter endmill cutter begins to get stuck with plastic debris, try to lubricate the cutter with some type of oil for these purposes. At the beginning I joined the assembly with a gasket, but I found it more complicated to achieve a good tightness than if I joined the pieces directly. If you notice that, during operation, air is sucked through the joint, try to cover the leak with very little glue.


The electrical connections are very simple:

First identify the correct polarity where the DC motor rotates clockwise and mark them as Positive Cable and Negative Cable.

Second, establish an electrical connection with the soldering iron between the positive banana plug (red) and one of the power switch legs.

Third, solder a wire from the other leg of the switch to the positive wire of the electric motor.

Fourth solder the negative DC motor cable directly to the negative banana connector (Black).

Join the whole set with the corresponding screws and nuts. Insert the hose through the hole created for this purpose and place glue to hold it in position. Avoid causing clogging with the impeller.


If they are very close it will cause excessive frictional force between the impeller and one of its cavities. If they are very detached, the magnetic interaction may not be enough to transmit the necessary torque for the correct operation of the pump.

Something that I accidentally discovered is that when the system is pumping water the impeller seems to be "floating" inside the cavity and the friction is minimal with the cavities (something I will have to investigate further).

If you have completed these steps, you probably already have your own variant of this water pump. I hope you enjoyed it as much as I did.

Magnets Challenge

Runner Up in the
Magnets Challenge

Be the First to Share


    • Trash to Treasure Contest

      Trash to Treasure Contest
    • Raspberry Pi Contest 2020

      Raspberry Pi Contest 2020
    • Wearables Contest

      Wearables Contest

    6 Discussions


    14 days ago

    . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10.02.2020
    Nice work, I like the clear material for observing impeller operation.
    Interested to see your investigative results of the impeller float.
    Just some thoughts...
    :- Might be able to include some fines (maybe paint flake) in the liquid to observe turbulence etc & capture with high speed video.
    :- The change in impeller axial position "float" might be due to load, possibly repulsion as the impeller lags behind no load magnet alignment. Should be able to confirm this with strobe (android app?).
    :- Try putting a small low friction washer (or machine a small step) between the impeller & back housing, this might enable closer magnet coupling & enable easier impeller startup until float is attained.

    I would add a disk piece on the drive plate to cover the magnets & mitigate the hazard of a loose magnet flying off.


    Reply 13 days ago

    Thanks for all these suggestions. I appreciate that you spent some time giving me your ideas. Of course, I think it would be very feasible to use a camera at high fps and study the behavior of the fluid inside the impeller cavity. Can you imagine the impact it would have on technological development if a device similar to this could work safely without rollers attached to the impeller? As I have observed, the impeller begins its movement by touching one of the faces of the cavity, but when the water flows I think it takes off, apparently due to the appearance of some force on the back area of ​​the impeller that compensates the magnetic attraction between the magnets . In my results this happens at the distance that I recommend from 5 to 8mm approximately. Of course so far they are assumptions that must be verified at the empirical (experimental) level. Theme for a new INSTRUCTABLE, I think so. regards


    Tip 14 days ago on Step 10

    Due to the axial design of the magnets, there is a large axial force that the impeller bearings must bear. It is better to use a radial flux arrangement (like the original motor you got the magnets from, with impeller having magnets facing out, and the motor magnets facing in, as in the inside of a cup). This largely balances the magnetic forces. The wall between the 2 must have a projection for the motor magnets to fit over, and the impeller magnets to fit into. This is the way that commercial magnetically-coupled pumps are made.


    14 days ago

    I love it!!


    17 days ago

    Nicely done Instructable. This principle could be used as a drive clutch as well.


    Reply 16 days ago