Inception Top With Magnetic Levitation

As someone who has a keen interest in making props and costumes I always pay attention to specific iconic props which catch my eye when watching any form of media. One such prop that has stuck in my head ever since I first saw the film is the spinning top used as a totem by Cobb (Leonardo Dicaprio) in Christopher Nolan's film Inception.

However I also don't like to make props that are too simple or non functional, I prefer a prop that I can interact with in some way, it makes the creative process more engaging. A spinning top in itself is functional but it felt a little too easy and mundane. As a result, since I first saw the film I've been contemplating what feature to add to this prop and ultimately decided to try and get the top to levitate using magnetic levitation.

I have played with the 'Levitron' toy that functions on the same principle and feel as though the gravity defying and self righting nature of the project would be well suited to the themes of Inception, the function of the prop as a totem within the films context and the surrealism that comes with it. So I set about designing and 3D printing a top and base on which to spin it that would satisfy both replicating the iconic top from Inception, and getting a successful, stable and repeatable magnetic levitation.

This Instructable aims to serve two purposes, explaining how I managed to build my design using trial and error in case you wish to follow the same steps, as well as show the process of assembling and finishing the 3D files I have attached if you wish to replicate my design exactly. If you want to make your own, or adjust parameters follow the whole Instructable. If you just want to replicate my design, only follow the steps marked (My Design).

This Instructable requires varying tools depending on if you are directly following my design or taking inspiration to design your own. In each step I will show what, if any, tools and supplies I am using.

If you are directly following my design the tools and supplies recommended are as follows:

• 3D Printer (to print the attached STL's)
• 5 1/4x1/16 Inch neodymium disk magnets
• 1 Magnet source round base magnet as shown or loose ring magnet referenced in step 2
• 4 6-32x 3/4 Stainless steel machine screws and accompanying hex nuts
• Phillips head screwdriver matching the machine screws
• Cyanoacrylate glue (Super Glue)
• Sand paper/ sanding sponges
• Deburring tool
• Acetone
• Respirator

If needing to extract the Magnet:

• Glass container with lid
• Nitrile Gloves
• Flat head screwdriver and toothbrush

While researching how to design this project, I needed to have a better understanding of how the physics behind the levitation actually worked in order to ensure the accuracy of my design.

During this research I found a great article written about the original Levitron toy that very eloquently describes the science behind how this principle works: Physics Behind the Levitron (mit.edu) (http://web.mit.edu/viz/levitron/Physics.html

However I will try to summarise roughly how this project will come together and the considerations that need to be made, as well as, the adjustments that can be made to different variables to troubleshoot why your project may not be working.

There are two main forces that allow magnetic levitation to occur in this project, the magnetic repulsion between opposing poles of the top and base, and the gyroscopic stability of the top. These account for the lifting force for the levitation and the stability of the levitation respectively. The top will remain levitating for as long as the top is spinning fast enough to equalise the lifting forces and prevent the unbalanced forces pushing the top outside of the 'sweet spot' and dropping from the air or being flung outside of the field.

In the original Levitron it uses two ring magnets, one in the base and one in the top itself. When sourcing my materials I, where I live, was unable to procure a neodymium ring magnet small enough for the top so I opted to make a ring of disk magnets instead. You can see from the attached magnetic field diagrams I have drawn, roughly how the fields of a ring and disk magnet function. You will notice how the ring magnet with the hole in the middle functions essentially as two disk magnets next to each other from a cross section view, therefore using an array of disk magnets for either the top or base should function as an adequate substitute provided the spacing is correct.

When the top spins, in order for it to remain gyroscopically stable, the axis of rotation is aligned with the force of gravity so if a plane is cut through the top, the rotational forces will be on a perpendicular plane to the forces of gravity. In order to achieve magnetic levitation, the base ring magnet needs to sit level perpendicular to the force of gravity so the magnetic forces of the top and base are aligned and oppose one another. Therefore, the base ring magnet needs to have a means of being levelled so that the resultant lifting force on the top directly opposes gravity rather than forcing the top off in an odd direction.

Additionally, the opposition forces of the magnetic field will provide lift but need to balance the downward forces applied by gravity in order to levitate in a stable manner. This problem is addressed in the Levitron by changing the top's weight by adding/ removing additional ring weights until a stable levitation can be achieved.

This presents a problem for the project as unlike the Levitron which has a flat top where weights can be freely stacked over the crown of the top, the top in inception I aim to replicate will require all of the magnets and potential weights to be enclosed within a pre-established form.

If I was working with the same sort of ring magnet as the Levitron, I could hollow out the middle of the top and add extra weight with a non magnetic material like plastic, brass rods or lead shot. Which I did also experiment with in my design utilising disk magnets.

However, as I aim to use disk magnets rather than a set diameter ring magnet, the spacing of the magnets within the tops can be adjusted which will change the way the magnetic fields interact. If you hold a magnet directly over the centre of the ring magnet you will notice the force is less than if you hold it over the main body of the ring itself. Therefore adjusting the spacing of the of the disk magnets in the top so they have a wider radius of the circular pattern will make the magnets experience more lift (functioning like a wider diameter ring magnet), ergo reducing the radius and having the disk magnets closer to the centre will reduce this force.

Where I live I was unable to buy an isolated ring magnet so I had to reclaim one from another product. This step is optional and if you can find a magnet on its own you can skip this and move to the next step.

If however you are unfortunate like me and do need to do this step I recommend doing it first as it is a lot easier if you let it sit overnight.

To remove the ring magnet from the metal housing it came in, I needed to first remove the glue/plastic that it was set in. I was not sure specifically what material this was but thought the solvent most likely to be effective would be acetone as it works on a variety of adhesive and plastic materials. I found a reasonably sized glass container with a lid and left the magnet submerged in acetone overnight. If you cannot find a glass container to use you can substitute a plastic one but do a patch test to ensure it is not reactive to the acetone before you use it. Please, when working with any solvents, be careful and use proper PPE. I was wearing nitrile gloves and a respirator (with appropriate filter) as you do not want to mess with the fumes coming off of acetone. Due to the low evaporation point of acetone and the fumes produced, you want to be sure to use a lidded container. If you do not have a lid on your container it will all evaporate off, not only giving off fumes you don't want to breathe, but also no longer working on the magnet, so use a lidded container or place it in a non reactive plastic bag inside the container, or both.

After sitting overnight, the material surrounding the ring magnet had softened, some had fully dissolved into the acetone solution but not all of it. I decided to use a flat head screwdriver to scrape along the side and remove the rest of the remaining material while it was in a soft clay like consistency. If you find the acetone hasn't fully penetrated down to the metal, let the magnet soak a few more hours and repeat until you can scrape out everything fixing the ring magnet in place.

Now that the magnet was only held in place by the magnetic attraction to the steel housing, I was able to use my screwdriver as a lever and pry the magnet free. If you aren't able to do this, you could always cut the housing with a Dremel tool and slide the magnet out sideways.

Now that the magnet was out, I could just use an old toothbrush to scrub off the remaining gunk and I had a perfect reclaimed ring magnet to use for this project.

If you are wanting to use my print files, the magnet on its own has the dimensions of an outer diameter of roughly 70mm and an inner diameter of roughly 32mm with a thickness of 10mm and a force quoted at 43KG (according to packaging).

Knowing the theory of how magnetic levitation worked, I knew I would need to experiment a bit to find the right balance of factors to get my top to work. I decided to then begin by modelling a top in my preferred CAD software of choice, Autodesk Fusion 360. I will not explain the whole CAD process for you here but will highlight some of the most useful tools and features for making or modifying this design yourself if you wish to edit this project to accept different magnets or if you can't get it to work due to environmental factors. Be advised environmental factors such as temperature or elevation will effect the forces of your magnetic fields.

A great tool when you are trying to replicate something in CAD is importing an image as a canvas. This allows you to have the reference image right there while you are sketching and modelling. You can see I imported a still frame of the top from the film and sketched my profile using a spline. Also note I scaled the canvas based on dimensions I found online for the top, and added a part line at its widest point to be able to print it in two halves. Printing in two halves allowed me to imbed the magnets and minimize any overhangs of support material. You can also see I offset the profile slightly to optimize it for 3D printing and not leave too many thin areas or towers. I then revolved the top in two separate parts, an upper and lower body and used the face that split them as my new sketch plane for the magnets. I sketched a line from the centre of the top and constrained it to be the radius of circular pattern I wanted to produce. I found 7mm to be perfect for my set up. I then sketched a circle the size of my magnets, offset them to account for tolerances of my printer and used a circular pattern around the centre point of the circle with 5 instances. This then allowed me to extrude the magnets as a cut, using both halves of the top as target bodies. This set up meant that with the constrained sketch and fusion's timeline feature, I could at any point go back in the timeline to edit the extrude feature to adjust tolerances, or edit the sketch and sketch dimension for the radius of the circular pattern and change the spacing of the magnets.

I used this principle to experiment by printing several tops through trial and error with different spacings till I found a sweet spot for my magnets and my specific climate and atmospheric conditions. I also played with some different sizes and numbers of magnets but found this size magnet with 5 instances and a 7mm radius for the circular pattern, to work the best.

I printed out a variety of tops with the different variables I was adjusting and then began experimenting with the levitation using the ring magnet so that I could then get an idea for the requirements and dimensions when designing a base for the top.

If you are not worried about your top looking like the one from Inception I would recommend replicating the original Levitron design and using something like nylon washers slipped over the top to adjust weight, it would be easier and require less printing when compared to working within a confined volume as I have done in this project.

If you do have any questions about the CAD process throughout the Instructable, please let me know in the comments.

Now that I had a prototype top printed out, I played around with it to try and determine the optimal offset to use and how profoundly the levelling of the magnet affected the levitation of the top.

A really easy way to do this is using a few decks of cards. I grabbed a flat non magnetic surface, a tupperware lid and used it as my lifting surface. You could use any relatively thin rigid non magnetic surface such as a notebook, cutting board, piece of wood or thick cardboard, so long as it will not disrupt the magnetic field (no ferrous or magnetic materials). then placing the ring magnet on a flat surface, I stacked playing cards to add a bit of space between the magnet and the top.

If the top is too close to the magnet, the fields will interact to try and pull the magnet through the centre of the ring; if it is too far away then force will push the magnet away and it will never be able to spin properly. There will be a sweet spot where those forces are transitioning and the top will spin stably. Keep adding and removing playing cards until you find the point that you can get the top to spin the most consistently.

Once the top will spin in a stable manner, begin to lift your lifting surface until the top begins to levitate. If the top doesn't ever fully levitate then the weight of the top is too much and needs to be reduced (or in my design, the magnets are too close together and need to be spread further apart). Once you get the top to begin levitating, it may begin to fall off to one side. This is likely due to the ring magnet not being level and therefore not perpendicular to the gravitational forces on the top so the magnets are not in parallel. Placing a card under the magnet in the direction the top falls off will shim up that side and begin to counter that inconsistency. Repeat this putting cards under the magnet until you get the top levitating in the centre and you can then fine tune the weight/position of the magnets if you are still not able to get a consistent levitation.

This is the most basic way of getting the top to levitate with house hold materials, and if that's all you want then feel free to stop here!

I however, wanted to make a base to hold the magnet that I could adjust without the use of playing cards and have it all in a self contained unit.

So it was time to go back to Fusion 360! But first, I took a measurement from this set up to figure out how much distance to put from the magnet to the lifting platform to utilise in my 3D model.

So I now know how far to offset my lifting surface from the ring magnet, but I don't yet know how I am going to deal with the adjustability of the ring magnet to make the levelling process easier. I decided the easiest and most consistent way would be to use some nuts and machine screws to hold the ring magnet in place, and have the possibility of tightening the screws to protrude from the bottom like legs of a table, therefore tightening or loosening the screws would raise or lower that side of the magnet.

Another consideration I had noticed in my experimentation with the top and playing cards was that sometimes the top would wander and mess with the way the fields interacted, causing more wobbling and difficulties getting levitation going. I decided that it couldn't hurt to add a bit of a concavity to the spinning surface so that gravity would nudge the top to the centre as well as the magnetic fields themselves. This was probably unnecessary but I liked how it looked so decided to keep it.

With this brief in mind I began sketching out a profile I found aesthetically pleasing and complimented the shape of the top. I revolved those profiles in the same manner as the top to minimise supports and overhangs and to also keep the lifting surface separate to the rest of the base. Once I had that body I added a smaller piece to hold the magnet in place when it was assembled. I also added some notches to the base to allow the lifting plate to be more easily grabbed if necessary, and then cut away the spaces for the machine screw and nuts to the relevant parts to hold it together. I then used a circular pattern on the features to make the notched and hardware spaces, and repeated those around the base 4 times. I opted for four times in order to make adjustment of levelling easier and add more stability.

Finally I wanted to see if I would like the design when it was fully assembled so I moved the top to a rough levitating position and used Fusion 360's build in rendering engine to see how it could look.

Now that I had all of my parts modelled I just needed to print, post process and assemble them.

For all of these files I printed them using ABS as it is my preferred material of choice due to how well it sands and that acetone is its' solvent, allowing for opportunities for adhesion and surface finishing not as easily available with PLA. I also printed these files at 100% Infill and with a brim to increase adhesion, structural integrity of the parts and allow plenty of material for post processing. Other settings should be dialled in for your machine and material you are using.

Once the prints were all ready, I began sanding and assembling the base. This process starts by peeling off the brim and I like to use a deburring tool or sanding sponge to clean up along the edges of the print. Don't throw away that brim though, I will show you a useful way to reuse that extra material further down.

Once the edges are all cleaned up its time to start sanding surfaces.

Before doing any amount of sanding, a reminder to wear the proper PPE, I am using a respirator and wearing glasses as ABS will produce a lot of sanding dust and you do not want it in your lungs.

The first thing we are going to want to do is smooth out the spinning surface of the lifting plate as the ridges of the print lines will stop the top spinning smoothly. I started with some 100 grit sandpaper because that was the coarsest I had on hand, and took it to 150 grit, which is smooth enough for our purposes. If you have a coarser grit feel free to use that to remove material more quickly. I found the best way to sand this concave surface was using circular motions, alternating between circles concentrically around the entirety of the piece, and smaller circles about half the width of the piece to really knock down those larger edges.

As you sand, the sand paper will become less effective as its grit breaks free or it gets gummed up with plastic dust so save yourself the work and just use a clean area if it starts to become more difficult to knock down the edges.

This process can take a while to get it a nice smooth surface so I recommend taking some breaks and stretching your hand so its not locked in the same position or you may end up getting a bad hand cramp like me.

Keep sanding until its nice and smooth like you can see in the pictures. You can then sand the external edges of the lifting plate for cosmetic purposes until it looks the way you want, this will not change the function of the piece.

Then repeat this for the other two sections of the base if you'd like them to be a smoother final product.

Now its getting time to do a test fit and test our top. However, you can see when I put my lifting plate against its mating surface on the magnets housing, there is a gap towards the edges caused by some warping and edge peeling during the printing process. This is very common when printing ABS parts and the brim was intended to minimise this on the piece, but there has been some mild warping nonetheless.

In order to get the surfaces to mate together as flush as possible, I sanded the flat sides to improve the fit.

You can see how much dust was felt from this, that's exactly why a respirator is important. take a minute to clean up your surface and we can start assembling.

I noticed a slight gap when pressing the two halves of my top together so decided to give them a light sand, but this is not entirely necessary.

Starting with the bottom of the top, I slid the magnets into place by pushing them down into the hole then sliding the remaining magnets off from the side. I adjusted the tolerances of my design to be a snug press fit for my printer but if the magnets feel a little loose, add a bit of CA glue to hold them in place if necessary. It is important that all of the magnets are facing in the same direction, if you have them facing different directions the top will not levitate. You can help to ensure they are all facing the same way by marking the same side of each magnet with a sharpie, or just by checking that the face you are about to press into the hole is attracted to a magnet that is already affixed in place.

If you are using a material other than ABS you can then glue the halves together with the CA glue ensure the magnets are fitting in place on both sides.

If you are using ABS I would recommend adding a bit of acetone to a jar and putting in any excess support material, failed prints or the brims we removed earlier to create an ABS slurry, this slurry makes a great adhesive as the acetone will melt the surfaces of the ABS you are joining together and create a strong bond like they are one piece of material. The ABS dissolved in the solution will also help to fill in any voids or gaps in the fit and can be sanded back flush. Don't use a nice paint brush for applying it though as it will be very difficult to get fully clean again, I have one brush dedicated for specifically this purpose.

This ABS slurry is also great if you have poor bed adhesion, brushing a small amount on your print bed will really improve the bond of your print to the build plate.

Now that the top is firmly joined with the magnets contained within it, you can sand the edges flush and smooth out the surfaces till you are happy with the finish.

Pay close attention to the bottom of the top at the point on which it will spin, you want as little friction as possible, and any inconsistencies in the shape, print lines or blobbing will disrupt its spin.

The assembly of the base is pretty straight forward, however you want to make sure that the magnet is facing the right direction. You want the magnets in the top and base to have the same poles facing each other so that they have a repulsion force.

When you place the top next to the magnet you may think that you want to feel it pushing away from the middle of the ring magnet, however this is not correct, due to the magnetic field that is produced by a ring magnet. You want the top to feel as though it is going to be sucked through the hole in the ring magnet. If it feels like there is an invisible bump when you try to pass the top through the hole you've got it upside down.

Begin by pressing the nuts into the bottom plate, again use a little super glue if you feel you need to.

Then place the magnet into the base with the correct side facing up where the spinning surface will sit. Then holding the bottom plate against the base with the magnet, screw the machine screws in from the obverse side until they are flush with the flat surface. I utilised stainless steel hardware as it was not inherently magnetic so hopefully any interaction with the magnetic field is negligible. You could also utilise brass hardware or another non ferrous metal if you can find it. If you use a different kind of steel hardware check it is non magnetic or it may make the top do wonky things. Once you have all of the screws in, the base is fully assembled.

The screws are currently flush with the top flat surface, this is their starting position. Tightening the screws further will begin to make them proud on the bottom and raise the base on that side. This is how the level is adjusted, in the same manner as the playing cards were used in step 4, if the top flies off to one side, adjust the screws corresponding to raising the side to which the top flew. It may take several tries to level the surface just right.

Now that we have it assembled its time to test it out and see if you can get it to levitate! Reminder to keep your screws flush to start with and slowly adjust them as you go. The screws are a much more finite method of adjusting than using the playing cards as shims so it will take a little bit of getting used to to feel how much each turn raises each side. If you are having issues with the screw feel free to just keep using the playing cards! I was able to get a decent levitation using both methods without too much difficulty, but it definitely is a skill you have to learn through trial and error.

Now the project is essentially complete! If you want to you can paint and finish it or leave it with the raw finish of the print or smooth it out further with some acetone vapour smoothing.

I opted to paint it and finish it using some graphite powder as I wanted to mimic the metallic look of the top in the film itself.

There are some still images of the top levitating but they really don't do it justice until you see it floating in real time.

I have uploaded a short clip of the levitation as an unlisted Youtube video that you will be able to access from this link:

Inception Top Levitating Video - YouTube (https://youtu.be/g1sG6nNOg0E)

In this video you will be able to see how the top wobbles and then rights itself as it adjusts to the imperfections in its spinning cycle and the influences of the magnetic field. It is just as much a skill to know how quickly to move the base up and if you need to lower it a bit then resume raising it, as it is to level the magnet and spin the top itself. There are a lot of variables that will effect the spin and levitation of the top so it has the potential for countless hours of tinkering and practice. I am sure many of you will be able to improve upon my prototypes and get tops that levitate for much longer.

With my magnets and current set up I have been able to get sustained levitation for an average of about 30 seconds. By the standard of the Levitron this is still pretty poor and can definitely use some more fine tuning but for my amateur project and first time attempting this I don't think its too bad!