Project Tornado

This project has a long history, spanning about a year. The idea was born when we were discussing ideas for a Caltech party; I didn't want to go with tesla coils because they're loud and obnoxious (and sort of expensive). I wanted to do something with fire, which has its own special kind of beauty. I then remembered a fog tornado on display at the Maryland Science Center, which was just a freestanding vortex you could wave your hand through and such... why not do the same thing, but with fire? Thus, the idea (nay, obsession) was born. This was during the winter break of 06-07, and it led to the following creation:


Tornado Mark 1

Made in my basement out of a shop vac, leaf blower, styrofoam, and old garden hose, this was the first attempt. The model imitated the one at the science center pretty closely, albeit on a much smaller scale. There are six vertical tubes with holes drilled in the sides that are arranged to blow towards the next clockwise tube. This creates the circulating current, while the shop vac sucks air out of a hole in the top to create the required updraft.
I at first wanted it to be completely freestanding, like the one at the science center; this refused to work, either because there was a strong air current in the room or because the circulating air wasn't being blown fast enough. There isn't much detailed information on tornado design and construction available on the internet, especially not for freestanding ones. So, I surrounded five of the six walls with plastic wrap; upon doing this, the tornado worked in an instant. It created a nice stable vortex using evaporated glycerin fog (which is pretty close to the commercial fog stuff) and managed to set off our fire alarm at 1 AM. Once my parents were out of the house, I set up a dish with some rubbing alcohol and fired it up to see what would happen. The results were very promising, though I didn't get pictures because I was too busy keeping the styrofoam and thus our pretty wooden house from catching fire. From this I learned the importance of having partial shielding around the tornado to keep the vortex stable; without this, your chances of success are slim.

Tornado Mark 2

The next step was to build the thing in a slightly larger form for the party at Caltech. I got to work, and had a basic frame about 7' high out of 2x4's up in no time. I found an old fan in one of the other student houses that I bought for a solid $30; it was a 36" belt-drive solid steel fan with five blades. Even with most of the housing removed, the thing weighed in at a good 40-50 pounds of steel, making it a rather unpleasant experience to mount and lift. I scrounged up a few window fans from nearby for the circulation and started testing. After putting up a 4'x8' sheet of corrugated metal, the vortex began to form very nicely, even with only one circulation fan. I usually just used sawdust on the ground to test the vortex; if the sawdust began to swirl in a tight loop, I knew there was a vortex reaching all the way up. I was also testing it with dry ice or liquid nitrogen fog. Here, positioning the wall and angling the side fan correctly seemed crucial, and for whatever reason it was set up really well right then. I only tried it with fire once, and the results can sort of be seen in the following pictures:

It was never very stable and the vortex refused to form properly, but it was getting pretty close... you can see the flame tapering off towards the top fan but not being stable enough to reach. Anyway, I had to move and raise the tornado for the actual party (where it was going to be with dry ice instead of fire), but once that happened it refused to work properly. While that was a disappointment, I learned enough to change the way I was going to do it for the next incarnation.


Tornado Mark 2.5

I was bored and slapped this one together out of some leftover plywood. There aren't any great pictures of it working, but it was just three small sheets of ~4' OSB plywood with a plastic box fan sitting on top and two across the sides blowing through. I used a sawdust/lighter fluid mix to provide the flame, later supplanted with some acetone. It worked very well - for most of the evening, there was a 4-6" wide vortex of fire at the center burning in a stable manner. Whenever I added more lighter fluid, especially, it would look great; a nice fat column of flame reaching through the box fan. Oh, and that box fan happened to be plastic. Honestly, I'm surprised it lasted as long as it did... the entire casing melted into a puddle before the motor gave out. Anyway, the success of that one gave me hope that a larger one, if properly constructed, could work well. Since I was going to be at Caltech over the summer, I figured I'd just put it off until then.

Tornado Mark 3

The only things I kept from the previous models was the large uplift fan; I got a new motor and a new belt (the old motor was irreversible, which forced me to mount it upside-down). I also designed this one from the start to run off of propane, since you can get a propane extension hose (read: flamethrower) for about $10. The propane source was a little metal pan at the bottom of the tornado, where the gas would be blown in from the side and swirl around. The design, as shown below, was still fairly big and lumbering, with a frame made of 2x4s and 4x4 legs. The whole thing was marked and bolted this time, though, to facilitate disassembly and storage. The center fan was mounted on metal rails going diagonally across the top (for structural support to prevent shear). The entire thing was extremely solid and stable and ran like a charm. My first tests involved trying to get it to work with fog, with no surrounding walls (the most difficult test). After a night or two of trying, I finally got it working using six circulation fans. After that, I decided to add metal walls as backing and use fewer fans; this turned out to be an excellent idea and the whole thing worked much better afterwards. Anyway, I ran it for two hours or so a night for about a week straight, after which it got taken down and has been in storage since. Most of the parts are now missing, so I'll have to rebuild it (probably lighter, faster, stronger, better) next time I want to use it. As for the propane, a single 5-gallon barbeque tank would let it run at the level shown in the video for about two hours. It's possible to invert the tank and let the liquid propane flow out directly; not only is this extremely dangerous, but it uses up the propane in about 15 minutes. The flame grows to about two feet wide, however, and makes the coolest roaring noise ever. Here are the schematics (I don't really have photos of the entire thing) and two short youtube videos:

...Tornado Mark 4?

Making the previous one once again taught me a couple of things. The two most important: walls are essential to create a circulating current unless you want to spend a bunch of time/money on directed fans, and that fire is hot enough and rises well enough that an uplift fan isn't really necessary (or at least, not a 50 pound metal piece of crap). With these in mind, at some point I'm going to design and build a much lighter and more compact model, maybe one that's completely enclosed to shield it from the wind; I haven't quite decided yet. Either way, there are definitely big improvements that could be made over the previous incarnation.