RC Pyrotechnic Flamethrower

**Disclaimer. Fire is dangerous. Treat it with respect. I'm not responsible if you don't.**

Though I've been a maker for roughly a decade I've never put much effort into documentation. I've decided to change that. My first step is creating a new YouTube series I'm calling HyperIon QuickBuilds (inspired by Adam Savage's one day builds). The basic idea is a set of (not necessarily simple) engineering projects which I complete in less than 3 days start to finish. If you're interested in seeing more of my projects/research, such as where I turn this same system into a Samus style arm mounted flamethrower, consider subscribing to my YouTube channel.

I figured what better way to kickoff a new channel and series than a project dedicated to making some hardcore practical effects. My original intention for these flamethrowers was to use them in the creation of a channel icon and channel art, but there are many more possible applications. These could easily be made to sync up to music or add some practical flare to a homemade movie. I chose a long and narrow flame for my uses but by changing the dimensions of the nozzle allows you to modify the flame shape to your exact application!

Without further ado lets get started.

To complete this project you'll need:

3D printed components (Thingiverse)

Mechanical Components:

1. 4x M4 40MM Screws (nuts optional)
2. Copper Welding Nozzle
3. Piece of Copper Tubing (size depends on application i used 1/4x2.5")
4. Bernomantic Butane Canister

Electrical Components

1. Servo (this also works)
2. Radio RX/TX
3. DC-DC Converter
4. Battery
5. Battery Connector
6. Locking Toggle Switch
7. Servo Wire

In order to minimize print time I chose to separate the top frame into two pieces and use four M4 40MM screws in order to bridge the gap. If you would rather deal with the extra print time I included the editable files for this project in its Thingiverse page. I am a strong believer that all Thingiverse users should do the same, yet it seems that's not common practice.

You'll want to first screw each of the M4 screws almost completely into the top piece, stopping right at its rim. This is in order to leave enough space after the fact to screw into the bottom piece. You

Next position the assembly you just made onto of their relevant holes of the bottom frame piece. Screw in each piece just enough to get it started then alternate around the assembly screwing in all the way. It does not have to be exact as the tolerance I designed allow the Butane canister to move even if the angle is rather extreme however it looks better if they are all within a few millimeters of each other.

You can optionally include some M4 nuts to further secure the assembly but I found this to be completely unnecessary as I made the M4 holes just tight enough to securely hold the parts together without being too hard to screw in. This can usually be done by making the hole diameter .05mm smaller than the actual screw diameter, and relying on the fact that 3d prints naturally shrink on their own.

The electroncis for this project are extremely simple. The order of operation I chose is:

1. Solder the power wire of a battery connector to the input of the DCDC converter.
2. Solder the other wire and input to different terminals of a locking toggle switch
   1. I chose a locking toggle switch because while we want the ability to turn the machine off if it were to randomly lose power it could be stuck in "flame thrower" mode which would be bad for a host of reasons.
3. Solder the output of the converter to a PWM connector in order to power the reciever
4. Check to make sure the voltage of the dc-dc converter is within the limits of your servo and rx (i usually use around 6.5-6.8v)

After that you should be able to plug the PWM wire into the "bat" of your receiver and the input to your servo into whatever channel you want to control it on. Everything should move correctly at this point.

Additional Commentary:

I originally planned to make a set of PCBs for this project based on an ATTINY44 and a 443mhz rx/tx pair. However, after messing up the first set by misreading a pin and having to remake them. I realized that while its not particularly difficult for me to make a set of PCBs that's a capability not everyone has, particularly one with SMD soldering. I was choosing to complicate the project in a way it did not need to be complicated. Not only that, with my PCB you could only control one flamethrower at a time. While I'd never simplify a project of mine to the point it loses any functionality just so its easier for others to follow I realized I was actually giving up usability for the sake of complexity. So from there on I switched to using an off the shelf rx/tx pair and gained a lesson to watch out for unnecessary complexity for the sake of making something "custom".

In order to have the best possible fit (which isn't actually necessary for this project, but good practice) I have come up with the following method to install the pinion. Its worth mentioning that the pinion is slightly undersized, which allows for an extremely tight fit but makes it next to impossible to get on with any other method.

1. Screw the M3 10mm screw into the hole of the pinion until it bottoms out.
2. Manually screw the screw protruding out of the pinion into the servo's hole, stopping just as the edge of the pinion touches the top of the servo's shaft.
3. Using an Allen wrench twist the screw while keeping the pinion (and therefore servo) stationary. Since the action of screwing in the screw doesn't cause that much friction the servo should not turn and the pinion will begin to slide over the servo shaft.

Congratulations! You just used one of the size basic simple machines (a screw) to gain enough mechanical advantage to compress a 3d printed piece onto a servo shaft without risking the rest of the servo.

The method I originally used to install the piston plunger is as following:

1. Align the plunger with roughly where you want it to be at its maximum stroke.
2. Set the servo to its maximum position.
3. Add hot glue to the bottom of the servo.
4. Slide the pinion into position on the plunger's rack and press the servo into place, allowing the glue to secure it.

Don't do this. After my first attempt I realized there's a far better method.

1. Trim the servo to its absolute maximum position.
2. Glue the servo in place.
3. Set the plunger on top of the pinion.
4. Reset your trim to zero.

This not only gives you access to a wider ranger of possible positions but is far easier to do.

This step controls the size and shape of your flame. Depending on how you let air interact with the gas as it comes out of the nozzle you can tune the flame height, width, and even temperature. I chose a tall, skinny flame for my purposes but the sky is the limit. It is worth noting that some flame shapes aren't compatible with the pilot light method and may need a electric igniter. More details in the "extra commentary" section at the end.

This step can be done one of two ways.

Copper Soldering:

This method is fairly straightforward to those who know how to pipe solder

1. Clean and flux both copper pieces
2. Heat the joint with a blow torch
3. Apply the solder to the joint, using the metal to melt the solder rather than the blowtorch.

As you can see from the video i accidentally nicked the solder with the blowtorch, dripping a big blob onto the joint. This isn't a huge deal but isn't good form.

3D Printed Version:

The tip can also be made by 3D printing a replaceable tube that fits over the copper nozzle. This is for those who don't have access to a blowtorch/solder. However one has to remember the tube will melt overtime and have to be replaced fairly frequently. I would recommend not doing the pilot light method with a 3D printed tube as it is a low power flame that will heat up the tube, a high power flame will actually cryo-cool the tip since there's not enough oxygen to let the flame get close enough to melt it. One advantage of the 3D printing method is you can easily control the nozzle shape to decide exactly what flame shape you desire.

Whichever method you choose you can then proceed to use a pair of vice grips to screw it into the top of the 3D printed frame.

Now its time to hot glue all the electronics in place and close it up. This is fairly straightforward, just put stuff where it fits and avoid wires getting in the rack and pinion. You can also install the screw now. I originally had the switch's hole too big and required a washer however I have since updated the model so you shouldn't have to worry about that.

To hold the whole thing together I designed four 3D printable thumb screws. Most of these should go in by hand but occasionally one will need a little help,

If you plan to have multiple flame throwers controlled by one controller its as simple as plugging in said servo's wire into another channel of the receiver. If you want them to be controlled together just add a PWM y-splitter.

Now it's ready to start making fire! First step is to trim your controller to the point that you have a pilot light going at equilibrium. I prefer to have a flame right next to the nozzle as I do this so you don't get a burst of flame but its slightly easier to trim it till you hear a light "flow" out of the canister and then light it.

After you do that, you're done! Depending on how your controller is calibrated one direction will increase the amount of fire you have and the other extinguishes the flame. To turn the flame off simply extinguish it then trim back to zero before flipping the toggle switch.

Those who have seen other project involving butane canisters have probably noticed something unusual. This system doesn't have a igniter. That was actually a deliberate safety decision.

An advantage this system has over all other "flamethrowers" using butane canisters I have seen is it has the ability to precisely set the exact rate you want butane to leave the container. That means that we don't need to use an electric igniter. Instead we can keep the flame running with a pilot light.

There are multiple reasons using a pilot light is safer. Firstly, it acts as its own warning light. If you see fire, it is armed. There is no "I thought it was turned off" because its very obvious whether its on or not. Secondly, its far easier to disable the entire flame thrower in the event of an emergency. A brief flip up on the transmitter and the flamethrower needs to be restarted. It's also possible to disable the fire just by kicking the flame thrower fast enough to break the flame or covering it with a blanket. If it had an integrated igniter that wouldn't be possible and you would have made the situation far worse for yourself.

That being said, I realize a lot of people would want an integrated igniter, say for a flame throwing robot. As such I made the hole I use to connect the two flamethrowers together large enough to run the wire for one if you so desire. RCLifeOn does many tutorials that show exactly how to make one in a way that is directly compatible with this system.

I hope you all enjoyed this tutorial! If you're curious to see what I
plan to do with this technique in the future/other projects I have planned make sure you follow me!

Stay Awesome!

-HyperIon