Making a Van de graaf

 The basic concept behind a Van de graff is creating high voltages, way beyond what any normal battery or generator could make. The average voltage of a Van de graff is around 100kV, which is 417 times more than the voltage that comes out of your plug socket. 

While your plug socket can kill you easily, the VDG cannot, due to the low current that is transmitted in every spark. Only around 0.35 joules of energy is released per spark, which isnt nearly enough to cause any sort of damage to your body, while 240V house lines can easily send 10 or more amps through your body in the right, or wrong, conditions, depending on how you look at it.

Van de graffs work in the exact same way as rubbing your socks on a carpet and getting shocked when you touch a doorknob or something metal. There is a rubber belt than rolls over two plastic rollers. As the belt spins around these rollers, it deposits its charge on one of them and takes in charge from another. These charges are picked up with 2 metal brushes. 

 

One metal brush is connected to the famous top ball, and other is connected to ground, so that the potential difference is between two things that you can choose - your body and the dome, another metal ball and the dome, etc. The analogy of the socks and carpet only work to a point, however. The van de graff uses contact charging, not frictional charging. In a van de graff, ideally, the friction is low, but it is the fact the belt rapidly comes in contact with a pulley and then leaves that creates high voltage.

 

This was the initial schematic of the VDG, a roller on the bottom that span due to an external motor, and a free wheeling roller on top. This, of course, was very simplified, so after confirming this would work, with the combs to collect charge being mounted near the belt, I started on 3D modelling it in CAD,

3D MODELLING (CAD) 

I began with modelling the roller, the part that would actually have the belt run over it, and came up with a tube with two nuts squeezing it in place.

 

Then I made the frame, simplyfing the base as there was no real need to model out the complex base I had found.

Then I modelled out the sphere that would go on top. This would have to have a hole cut in the bottom, as small as possible, to allow the top roller be hidden away.

 

Then I quickly modelled out what the comb would look like, that would collect charges from the belt.

 

Putting all this together with some extra models of bushings and bearings etc, we get this:

 

 

 

With this plan in mind, it was time to begin the build.

BUILDING

To make this Van de Graff, a frame must first be made. I used a long peice of scrap wood that i would cut in half to get two perfectly even wood planks as side supports

 

 

Next, I was walking around my area and found this beautiful speaker holder on someones wall, and they had no problem with me taking it as they were about to throw it away. After cleaning and revarnishing the wood, it is the perfect base for the van de graff, as it is solid and bottom heavy, with a low centre of gravity, making it very stable.

For the top roller, I was initially worried of how I would get the shape of it without a 3d printer. The roller must be taller on the sides and have side walls to stop the band from slipping off. 

Then i realised the hub of cheap kid's bikes are perfect for this, as they are slightly convex, are able to spin freely, have a shaft thats easy to work with, and are tapered. The only issue was the rest of the wheel connected to it, but this oppurtunity was too good to give up as my neighbours had their old kids bike in wretched condition in a skip.

 

To fix this issue, I spent the next hour of my life despoking this wheel, which is a painstanking process of removing the tyre, loosening the spoke nipples and then taking all 32 screws out of the rim, which is worse than it sounds because of the copius amounts of rust.

Eventually, all the spokes were removed. I unscrewed the bearing case and here my extensive knowledge of bikes helped, as I could remove all the ball bearings, the shaft and the housing so that they could be thoroughly cleaned and regreased.

This next step took almost an hour - the two wood planks had to be exactly vertical - I couldn't just eye ball it. They also had to be perfectly aligned. My solution to this took a long time to form, but I ended up using a 'level' app on my phone that could tell me if my phjone was horizontal to 0.1 degrees. 

I used this to level the side of the block so that it was horizontal, and then drilled a hole straight through the whole wooden block. I did this again so that there were two screw holes, as having only one screw would allow the planks to pivot.

 This led to a perfectly straight set of parallel wooden planks with 2 holes drilled in them for the bushings. These would be small sections of metal tube that would line the wood so that the wood wouldnt get eroded by the axle over time. 

These were made by drilling a hole slightly smaller than the diameter of the metal tube, and using my vice to press them extremely firmly into the hole. The wood clamps around the bushing, keeping it very securely in place.

The bushing came from a section of curtain rail, cut off. The rail was telescopic, so to find a peice of metal tubing just slightly thinner than the bushing that would fit perfectly, I simply used the tube that was fit in the bushing's tube in the original curtain rail. 

These came with stoppers which were screwed on, so I used these as axle limiters to stop the rod from being able to slip out.

 

To test the straightness of the system, I put this thinner axle tube through the bushings, and it span with very little friction and very little play. Most importantly, the axle was dead straight and horizontal.

After testing this design however, there were far too many issues. Since it was bare metal on metal, it was loud and had a lot of friction. There were a lot of issues with trying to make the roller stay centred - if the roller could slide laterally, so would the band, making the whole system derail. There were other issues with notches wearing into bushings and then a blindingly obvious issue came up.

For a Van de graff to function, the top roller and bottom roller should be made of different materials, ideally plastic PVC and a metal. The top roller by default must be metal, which is fine since the bicycle hub that I used for the top roller had a metal bearing case anyway.

The issue arises from using a metal rod as a bottom roller. I tried for almost a week to cover this tube with a PVC cover, but no progress could be made. The roller would not get centred, would rotate off centre introducing massive vibrations, and one prominent issue was that if the metal rod was spun and the plastic rod was held, the metal rod would easily work loose. I could not glue the plastic roller to the metal roller as this goes against my design philosophy of having everything able to be taken apart.

I settled on a new design using the other wheel's hub, from the same kids bike.

After painstakingly removing the tyre, loosening every spoke, removing the gears/cassette, soaking the assembly in WD-40 so it could be removed, despoking the hub and cleaning out the bearings, I had a clean bearing case. The plan was simple, to use the bearing case as a roller.

Firstly, i modelled out the roller in CAD using measurements from a ruler. This was just the raw casing.

 

The issue with this approach was how to make the casing spin with the axle - the whole point of a bicycle hub is to allow the hub itself to spin while keeping the axle still, so I needed to securely connect the hub to a shaft and not allow them to spin independately of eachother - when the axle spun, the hub must spin aswell.

To remedy this issue, the solution was simple - to use a threaded rod to squeeze the hub so that it was rotationally and laterally secured to the axle. This is hard to visualise, so to ensure this was possible, I modelled the whole axle assembly in CAD aswell.

First two washers are placed in the bearing cavity. I found that some M8 washers I had lying around fit perfectly, but a perfect fit wasnt even necessary, just a washer that wouldnt slip in the whole in the middle of the hub.

 

Then an M8 rod was cut to around 28cm and put through the hub.

 

Finally, two nuts would be screwed on opposite sides of the hub to squeeze the washes towards eachother. This in turn would squeeze the hub in place with the axle rod.

 

Incorparting this into the total, design, we get this:

 

The issue here is the axle itself will spin, so we cannot just have the axle going through a hole in the wood. For this reason, the hole in the wood is expanded to slightly less than the external diameter of a bearing, and the bearing is pressfit in with a vice.

 

Now, the axle can spin freely, but it can also move laterally freely. To solve this, we can simply tighten nuts on either side of the axle to squeeze the bearings inwards. This will prevent the bearing from ever slipping out and stop the hub from being able to move side to side.

 

Two nuts are used as they can be tightened against eachother to ensure a strong hold. Over time, if one nut is used, it would eventually get unscrewed as the bearing would slowly untighten it due to small amounts of friction.

The total assembly looks like this in CAD:

 

 

 

Here the issue was clear - it would not work. I could spin up the van de graff and record no potential difference between the top comb and the bottom comb. I tried everything, burning off the wax coating of the tin combs, making the belt thicker, remaking the whole system to the make the rollers more parallel and even changing the material of the rollers by wrapping tape around them