by reparator, published

Anti-Gravitator by reparator Jul 26, 2015

Verified Files

7 Share
Download All Files

Thing Apps Enabled

Order This Printed View All Apps


A part of these Groups

View All


Liked By

View All

Give a Shout Out

If you print this Thing and display it in public proudly give attribution by printing and displaying this tag.

Print Thing Tag

Thing Statistics

121342Views 15601Downloads Found in Physics & Astronomy


This is a model of an Anti-Gravitator, sometimes also called Levitator or Levitron. It lets a small magnet spere levitate in a magnetic field.
Video: https://youtu.be/0ftbobwkf2Y


Principle of the Anti-Gravitator:
Of course it is not possible to disable the gravity forces. The principle of the Anti-Gravitator is to compensate the gravity by a magnetic field generated in a copper coil. Referred to the Earnshaw´s Theorem (Samual Earnshaw, 1805-1888) it is not possible to get a stable magnetic levitation with permanet magnets. In this experiment the magnetic attraction force, generated by the electric coil, is controlled by a hall magnetic field sensor. If the floating magnet lifts up and approaches the sensor the current through the coil will be reduced leading to a lower magnetic force (and a drop down of the sphere). If the floating magnets descents, the hall sensor indicates a lower magnetic field to the OPAMP resulting in a higher coil current and a stronger magnetic force.
This sweep process takes place ~4.000.000 times a second resulting in a stable levitation state.

The case design of the levitron is adapted from the aluminium design of Andreas Titze
Thanks to the german Make Magazin for the inspiration to the topic.

What you need for this project:
1 x Hexagon socket head cap screw M6x40 mm as iron core
1x M6 nut and washer
100 g Cu wire Ø 0.28 mm
1 x electronic board as described in attachment
1x Power supply 15 V, 800 mA
1 x Hall sensor A1302 or simular.
2 x M4 Hexagon socket button head screws
4 x self-tapping screws ~2,5 mm
Nd-ball magnet (15 - 30 mm diameter) or similar Nd-magnet

First of all you have to print the carrier for the spool (Spule_Unterteil, Spule_Oberteil and Spule_Konus). These parts get sticked together with the M6 screw and fixed with the nut. The head of the screw fits in the part "Spule_Konus". Drill a 1 mm hole near the inner part of the spool so that the Cu-wire can be leaded through. After that you have to coil the copper wire on the spool. You need 120 m wire minimum so that the DC-resistance of the resulting coil is at least greater than 50 Ohms. If you need more wire to fill the spool this is not a problem. Fill the spool completely with wire. Solder a insulated copper wire on each end of the coil (approx. 20 cm is recommended).
Solder three isolated copper wires on the hall sensor legs, isolate the solder joints with heat shrink tubes.
After that glue the sensor with a big portion of hotmelt onto the top of the M6 screw head. If everything is solidified you can assemble the coil with the attached sensor into the "Cylinder"-part. Take care to lead the three wires from the sensor to one of the holes in the cylinder and the two wires from the coil to the other hole.

Print the other parts (in different colors would be nice) and build them together with hotmelt glue as you can see on the pictures. All parts are printed without support and with 20 % infill.
On the part "Sockel" four support sticks are included. They can easily be removed by breaking off. The inside of the "Sockel" needs some rework if you prefer a super finished surface. In principle it is not neccessary becaused it is covered by the electronic circuit board.
Assemble the two "Seitenteile" with the M4 screws on the "Sockel" part.
After that you mount the wires in the two small cable channels of the side parts. It is a good idea to secure the isolated wires with a small portion of glue. Solder the ends of the wires to the pads marked with "coil" and "Hallsensor" on the electronic board. Also add the cable for the power supply. I did not use the originally intended power supply chassis due to the oversize of that part, but soldered the wires for the supply directly onto the board. The pinhead connector called "Hall-control" is for future application and needs not to be soldered.
If everything is connected the electronic board can be assembled with the parts side downwards by use of four small self-tapping screws.
Glue the coil containing cylinder onto the side parts and cover the cable channels with the "Abdeckung_Kabelkanal" part.

Test and startup

Turn the precision potentiometer to the right end, hold the magnet sphere below the coil (approx. 15 mm), turn-on the power supply and turn the potentiometer screw slowly to the other end (as you can see in the video). If you reach the resonance point, the sphere will begin to levitate. If you turn further it will bounce to the coil cylinder above. Reduce the potentiometer than a small portion to get the system stable.

If you are not successful to get a stable levitation try at least to change the connections of the coil to the electronic board (exchanging the poles of the electromagnet).

Take in account that the power supply pulls a current of about 400 mA if the hall sensor does not detect a magnetic field (eg. because the floating sphere is falling down). This current leads to a slowly heatup of the coil. So don´t leave the instrument unattendet and turn off the power if no sphere is inside. During a stable levitation the current is much lower (< 100 mA).

NdFeB-magnets (Neodymium-magnets) are very strong and have to be handled carefully. If two of them collide they can break into small, sharp pieces. Also they are dangerous for people with heart peace-maker.
Keep children away from these magnets. Follow the instructions of the manufacturer.

Use this description at your own risk!

Have fun!

UPDATE 2016_01_20:
Added part V4_Sockel.stl in order to meet specific wishes :-)
This part has wider cylinders to give a better hold for the PCB (less fragile). I removed the support for the cylinders either. So you have to care for the support yourself by using the right parameter in your slicing software. Feel free to choose either V3 or V4 for the socket.

More from Physics & Astronomy

view more

File Name



All Apps

Auto-magically prepare your 3D models for 3D printing. A cloud based 3D models Preparing and Healing solution for 3D Printing, MakePrintable provides features for model repairing, wall thickness...

App Info Launch App

Kiri:Moto is an integrated cloud-based slicer and tool-path generator for 3D Printing, CAM / CNC and Laser cutting. *** 3D printing mode provides model slicing and GCode output using built-in...

App Info Launch App
KiriMoto Thing App

With 3D Slash, you can edit 3d models like a stonecutter. A unique interface: as fun as a building game! The perfect tool for non-designers and children to create in 3D.

App Info Launch App

Print through a distributed network of 3D printing enthusiasts from across the US, at a fraction of the cost of the competitors. We want to change the world for the better through technology, an...

App Info Launch App

Treatstock is an online platform that offers decentralized manufacturing services such as 3D printing and CNC machining for clients all over the world. We offer free and instant access to comparati...

App Info Launch App

3D print your favourite design with NinjaPrototype, a professional 3D manufacture with consistent quality and speed.

App Info Launch App

is it possible to move the ball up and down in a range of 100mm? but controlling the hall sensor distance somehow?

What is the length required for the button head M4 screws?

there is a gap in the "Sockel" (V3 & V4) below the hemisphere that will result in an empty layer if printing at 0.15mm.
already tried to repair the .stl without effect.
can you fix this?

I used a A 12v psu and the 100uf capacitor exploded does anyone know why that would happen. the polarity is correct

What is maximal voltage rating of your capacitor?

I used 100 µF/25 V. Are you really sure that your polarity at the power jack is right? The inner core of the jack must be VCC, the outer ground.

Hi guys, reparator.

I have received all the component except for the Hall sensor. I an wonder if I can replace the the one in the build A1302 with A1326 ? I compared the specs and I think the only notable difference is sensitivity.

A1302=1.3 mV/G
A1326=2.5 mV/G

Sensitivity is higher which I think can me compensated by the potentiometer rotation. Please correct me if I am wrong.
Both datasheets are here:


Another question , would 12V power work instead of 15V?


from datasheet I think the A1326 should work either. Another possibility is the SS495A.
As higher the supply voltage as higher the magnetic field strength as lower the levitation level of the magnetic ball. But 12 V power supply will work!

Thank you for offering such a cool file, and for keeping up with all the questions people have. I'm always grateful when a fellow maker takes the time to share their knowledge.

Aside from the magnetic ball providing a magnetic field for the hall effect sensor to read, what purpose does it serve? The reason I ask is this: would it be possible to levitate a similar steel ball - if the power adjustment to the electromagnetic coil was managed by an ir led and ir detecting photocell?

Thanks again.

If the "levitating part" would not be a magnet but only a ball of iron the electro magnet would be responsible completely for the levitation force. This should work either but I think you need a bigger coil / stronger electro magnet with iron core to get a levitation than. You can calculate the magnetic force you need very easy I think.
Of course you need a different control system than, as you said: an optical system should work.

Thank you for your response.

That's what I was thinking, but I'm just getting started with electronics, so I wanted to be sure. I really look forward to building the "Anti-Gravitator." Not only is it super cool, but I am hoping to learn a bit more about electronics and magnetism as I build it. I'll definitely post pictures when I finish.

Can anyone sell me the board? Awesome project, but I have no tools to make custom PCBs :((( Would love to build this, maybe in larger scale actually. Would be fun!

I have ordered 3 PCB boards from a remix model (http://www.thingiverse.com/thing:1982994) and can sell you one when it arrives (this week I guess) .

The Anti-Gravitator Thru-Hole Version

Awesome mate!. It's not only amazing, i real love that alien spaceship design.
Thanks for sharing!!

Does it have to be a ball? Or could I use a cube as well?

Check the video!

Does it have to be a ball? Or could I use a cube as well?

Does it have to be a ball? Or could I use a cube as well?

Does it have to be a ball? Or could I use a cube as well?

Does it have to be a ball? Or could I use a cube as well?

how does this work? (i am a little suspicious)

You know the film "The Black Hole" (1979)?
Works exactly the same way as Dr. Reinhardts!
(Read the instructions for more details....)

Hi Again,

I'm planning to make a dérivative of this thing where the controlled loop will be handled by an arduino.
The idea behing that, is to have several levitator side by side and have the controller animate the levitating magnets by moving them up and down.
Do you think it could be possible to move the levitating up and down by varing the controll loop? Will it be stable or will the magnet oscillate to much ?

Thanks for any feedback.

:-) Good luck!
I don´t think that it could work. The range, in which the magnet is levitating stable is too small.

Too bad !!! That's what I was afraid of.
Doing the controlled loop with an arduino should not be too difficult.
And if I can't move the ball up and down, well, I'll light it up with a RVB led :-)

I would not say that it is impossible at all. If you look at the following sides, there are arduino controls either.
In the second video you can see some moving up and down which seems to be controlled.

Shared the board on OSHPark


The Anti-Gravitator 2 layer board of 3.17x3.17 inches (80.44x80.44mm). $50.10 for three.


I bumped into your thing and I like it very much. Do you happend to have a BOM for the board?


Waht do you mean with "BOM for the board"? All parts and values are written in the schematic. You can take it from there.
The external parts are described in the instructions.


Yes, I've seen the schematic. But as an electronic novice, it is always quite difficult for me to go from the theoritical view (the schematic) to the real world (the PCB), because of the various choices (packaging, size, tolerances, capacitor types, etc, etc...) that can be made when choosing a part.
That's why I've asked if you happened to have a list of "suitable" parts for your Anti-Gravitator.


I dont know, if it really helps, but anyway:
Here is the BOM:
Name Type Value Form
C1 Capacitor 100nF 0805
C2 Pol-Capacitor 100µF / 25V Panasonic D
C3 Capacitor 100nF 0805
C4 Capacitor 100nF 0805
IC1 NE5532D8 SO08
IC3 LM1117MP-5.0 SOT223
Q1 IRF4905S D2Pack
D1 Diode 1N4148SMD Minimelf
R1 Potentiometer 1k RJ9W
R2 Resistor 680k 0805
R3 Resistor 4k7 0805
J2 Connector LUM1613
JP1 Pinhead 1x2
JP2 Pinhead 1x3
JP3 Pinhead 1x2


Will save me a lot of time :-)

What is the hall control header for? JP3 on the schematic provided. I can't equate this to anything on the original schematic. Thank you.

Good Question! You´re right, something should be written to that!
I used this header only to measure the hall response with an external multimeter. Therefore I assembled the pins on the downside. You don´t need this header for the circuits function. Regards!

A few questions, why use 15v and what's the purpose of using such an over capable mosfet? That one shows it'll take 170W max but the traces are so small. Seems that a much smaller mosfet would work, one rated for 20W max maybe?

You are right! This one I had in my restbox so I used it without calculation. Feel free to use a smaller one.

Is 15 volts still a requirement for the coil then? Or would the strength of the field be a function of current? I ask because these ICs will take 5 to 15v and the voltage regulator specified puts out 600mA to 800mA. Thanks.

I think you miss something: The voltage regulator is only for the hall sensor which has a supply range from 4.5 to 6 V. So the current through the coil is not regulated by the voltage regulator. The power consumption of the sensor is very low (as of the datasheet it should be below 11 mA) and even if you have an input voltage of 15 V which has to be broken to 5 V, the dissipation loss on the regulator is not that much (in the range of 100 mW). To achieve a sufficent field stregth to levitate the magnet you need some power. I can not calculate that exactly but I think the setup will not work if you reduce the input voltage near 5 V. Regards

I think I understand better now. Thank you.

can i use 20v battery source?

Hi Sir, after a year i manage to made it, but i have not understand where to place hall sensor. The small cap is the down side and the big cap is the upper side of the coil. I will place the hall sensor to the down side of the coil? the letters of the sensor will look down or up?If the sensor be placed there ,isn't it going to be affect form the magnetic field of the coil. Or i will place the sensor to the base of the construction
upwards of of the pcb?

Hi Mendel_2, you have to put the sensor in the small cap of the downside of the coil-case. The sensor has two chamfered edges. Bend the pins of the sensor that way that the chamfered edges face downwards and the pins upwards. Solder the cables an the sensor pins, put a small portion of hotmeld glue in the small cap and fix the sensor on the bottom of the cap. After managing the cables inside the coil case fill the small cap with hotmelt glue and assemble it to the coil case.
You are right: The magnetic field of the coil influences the sensor. Therefore the sensor should be assembled as far as possible from the coil but on the same side of the magnet ball. Putting it on the buttom side just over the circuit will not work because the regulation circle is in the wrong direction: The coil is switched off if the ball is falling down and vice versa. Please take a look an my pictures again. My picture "Coil with screw-core" shows how to assemble the core-screw. The hexagon end should be on the down side and the open threaded end with the nut on the upper side. For me it seems that on your picture it is different.

Hi Sir, unfortunetly i have not find out where is the problem yet!
I fix hall sensor below coil http://www.alfadex.com/wp-content/uploads/2016/06/IMG_20160626_104855-Custom.jpg

i am using those magnets http://www.alfadex.com/wp-content/uploads/2016/06/IMG_20160626_120450-Custom.jpg

Now what i get: When i touch with my hand sensor pin of hall sensor , i feel magnet vibrating but canot be stabilized in the air,curent is around 150mA. I also see this to the output of the op.amp (pin1) http://www.alfadex.com/wp-content/uploads/2016/06/ADS00001-1.jpg
If i don't touch sensor pin, there is no vibration and current is only 15mA.
I replace op.apm but the same.I have check pcb

edit:or i have to touch the core of the coil for start vibrating magnet, but not staying on the air

As I wrote you have to glue the sensor with a portion of hotmelt to the head of the screw. See the new picture which I uploaded a few minutes ago (it is only a simular model, the original one is glueed together and I don´t want to open it again).
If you see the magnet-spool is switched on if the magnet sphere is approaching that´s good! So you see a hall response. So now you should fix everything together, turn the poti in one direction and approach the magnet from botton again. At one point you should feel that the sphere is attracted. Try now by turning the poti to the other direction to achieve a stable state. If you should not succeed start with poti from the other side as described in "Test and Startup" description.

Tnx, i thing i see, i will manage it

Would you be able to provide the files to have the board uploaded at https://oshpark.com/ or upload the files there and share the project?

Great project ! Could you explain the electronics a bit, esp. the operational amplifier part ? I have trouble understanding the comparison stage ... I understand this is a comparator, with integration (ie. the R2/C4 components) ? Is this right ? Thanks !

I really like this design and consider to make one. One thing: why are V3_Abdeckung_Kabelkanal and V3_Seitenteile not already one part? Why glue the parts afterward when it can be printed (including the cable hole) in one go? I think that will give a better looking end result. I am not into free cad, can you create this part, or do I have to dive into free cad myself?

I thought it would be difficult to insert the cables through the small hole over such a long distance. Therefore I designed the side parts open. Then it is easy to put the cables in the channel and cover everything with two additional parts and fix theses with superglue.
Anyway: You could try to print everything in one step. See the new optional parts V3_Seitenteile_mit_Abdeckung.stl for that!

Thanks for the fast reply and new optional part! I am going to print one and try if I can get the wires in.

Comments deleted.

Hi! Nice design, I have been looking to do something similar for some time now but I could not find an schematic on the web for it, maybe you can help me.
I want to make a ball floating in mid air, like those earth floating globes you can find on ebay, but the thing is that I want it without the side columns and the top part, this means that the top part should be on the bottom, something like this:


Do you think you have some schematic liying around or you can help me in the right direction on how to make this design possible? Or maybe let me know if its possible modifying your design somehow to make it work like that? Thanks!

You should buy one. Just $30.
We're about to design a Anti-gravity BT soundbox,Of course without the side columns and the top part, Will it be cool enough?

Unfortunatly I don´t have a design for that. I also think about building somewhat similar to that Floating Globe. The principle is slightly different. In my design the gravity helps to get the levitation stable (gravity pulls the object downward, the magnet pulls it upward to a point). To get a levitation as you want it, you have to push the object against gravity. To get a stable state with that is more difficult, because the object tries to break-out sidewards. So you have to stabilize it with at least three magnet spools minimum, each controlled by a hall sensor. Or the object must spin around his axis to stabilize the levitation. Search the web for "crealev" or "Repulsive electromagnetic levitation" and you will find some inspiration.

Hi, maybe I have a problem in the electronics. The circuit seems not to regulate the current thru the coil, I have a abrupt passage from oscillating to non-oscillating, while the hall sensor output is ok and proportional to the magnet distance. Seems that the opamp doesn't drive the mosfet properly, the output form the opamp is not proportional but a sort of on-off state change. Any hints?

OK, difficult to analyze:The hall sensor works, thats ok! But take in account that the coil on top will give you a magnet field either. So there is a hall voltage even if there is no magnet sphere nearby. First of all recheck your electronic board!
Than try the following:

  1. Turn the potentiometer at one end, approximate your magnet sphere (the one you want to levitate) to approximatly 15- 20 mm underneath the coil. Retain it from falling down but do not lever it too much so that the iron core will not pull it upwards. The sphere should not fly upwards if you let it. Than turn the potentiometer slowly to the other end by holding the sphere at the same high. At one point you may recognize that the sphere begins to vibrate because of the attraction force of the coil.This is the point where you have to fine tune the hall feedback by adjusting the potentiometer in each direction. If you have the right adjustment the sphere will begin to levitate if you take your hands off. If it falls down turn the poti to the direction where the nagnetic fields increases. The tuning process is a little bit tricky, so try to get a feeling for the right high where levitation starts.

  2. If this does not work, you should start the tuning process from the other side, means turn the potentiometer at the opposite end and start again.

  3. If all this trying fails you should switch the coil connections means swap the coil connections at JP1 and try again the tuning procedure.

Hope you succeed! Let us know!

Hi , how could i find 120 meters contentiously copper wire?? I found 0.035mm copper wire from a tv cathode choke, but it was sicked together and i could n't excuss 120 meter, because it was cutting off. I collect only around 15 ohms ,not 50. Does anybody has any other solution?

you can buy enamelled copper wire coil on a spool. This is used normally for winding mini-transformers. It is isolated with a thin film of polyurethane. You can buy it in different thicknesses in electronic specialized shop. The one I used had 0.28 mm thickness (100 gr were 175 m) but everything in simular range should work either. The other way is to cannibalize an old transformer. The resistance is not that critical but if it is too low the current through the windings is too high resulting in a heat build-up. The 50 Ohms is only a roughly value. If you want it exact:
Pouillet´s law: R = Rs l/A with Rs: specific resistance (0,0169 Ohm mm^2/m), l: length of the conductor, A: cross-section area).
With 170 m of copper wire (0.28 mm^2) you calculate the resistance to 47 Ohms. If you have only 120 m, your resistance will be 32 Ohms. Less than that you should not have.

Hi, I am a bit clumsy at soldering. Do you think it would be possible to send me a finished board (the electronic parts as discribed in the schemas) to Croatia. I will of coarse pay you for the work and postage.

Just asking!

Update: Exchanged file V3_Sockel.stl so that the 4 support pins have a correct starting high.

The feet inside the base (V3_Sockel) don't come all the way down to the surface; this makes a mess when you try to print because the printer is depositing plastic in air at that point. I solved this by lowering the model below my print bed by 0.2mm before slicing.

You are right! I saw that either when I started the print. Nevertheless the print was still ok. There was no warping in spite of it. I forgot to fix it in the stl-file.

Nice. I would like to build one !
some questions :

  • what is the .FCStd file ?
    Do I need it when I have all .stl files?
    (I use Rhinoceros for design)


The .FCStd file is the FreeCAD File. You don´t need it if you don´t want to modify the design. Just for printing you only need the .STL-Files.

i am starting printing the parts!!

Really cool, it would be nice to see how to setup the electronics part.

What do you mean with "setup the parts".
I soldered everything by hand .-)
You need good eye-classes!
You can use a strip grid board and conventional parts with bigger dimensions either.

During setup process you only have to turn the potentiometer screw to find the right hall response.

I was just wondering where to find a parts list and a schematic on how everything is setup on the electronics side of things. I wouldn't know to make a coil or make my own circuit board if my life depended on it;)

The schematic is included in the Thing Files (Levitator_SMD.sch.pdf and as EAGLE-File Levitator_SMD.sch). You can find the parts there.
To make the coil is easy: Just drill a small hole nearby the axis in the spool body, thread the copper wire through that hole to the outside, than wind the copper thread by turning the spool until it is completely full. Cut this end too and extend both ends by soldering an isolated copper wire at each end. That´s it!

Thank you, I totally overlooked the instructions section.