Introduction: GaussGun / Coil-Gun Electronics and Engineering Project
I decided to do the Gold Crest Award at my college, which is a great certification from the British Science Association that gives students an opportunity to expand their knowledge of science, technology, engineering and maths by completing a project by themselves or with an experienced mentor of their choosing totalling 90 hours or more of work. I decided to build a Gauss-gun/Coilgun and took over 110 hours excluding the many sleepless nights I had trying to figure out the specifics of the circuitry.
At the end of my project I had a model which looked the business and did what it was supposed to but unfortunately the power with which it fired with was disappointing. At this point I'd done well over 90 hours and didn't have the money to rebuild large sections of the GaussGun like I would need to, so unfortunately I had to end the project before It achieved its full potential. I wasn't going to write the instructable, but I thought it would be very sad to let all of my research go to waste without letting it be shared.
Although I consider the end result to be something of a flop, I know exactly why (I discuss this in depth later on and in my final report), and I still think that my work can be a useful resource. Someone inexperienced in electronics could use my work as a guide to learn a lot about electronics and to make a final product which works, but is (somewhat unnecessarily) safe. An experienced electrician could build upon my circuitry but use my other more engineering oriented research to build an altogether much more powerful coilgun.
In any case, I hope you learn something from my project because I learned a huge amount, and I would encourage anyone to look at my final report for a much more in depth look at electronics, magnetism, CAD design, prototyping methods, materials science, solenoids, a broad range of electronics, ballistics and engineering. The main file is the .pdf entitled "GaussGun".
Step 1: Research
The first stage of the project was to thoroughly research Coilguns, solenoids, magnetism and anything else that might be useful. The topics I covered were:
- Existing Coilguns
- Material Science
- The Projectile
- 2D CAD and Laser Cutting
- Rapid Prototyping and 3D CAD
- Maximising the Force
- Overcoming issues in the Circuit
All of this research is available for your viewing at first page. Normally I would just get going with the project, but this one had to be much more well planned out, not just for safety reasons but because I wanted to make the most efficient device possible and really push the limits of what Coilguns can look/perform like built at home. I went through a lot of different Ideas of how it would work, and came out with a lot of different estimations of just how powerful it would be. In order to find an approximation for the force, I used the equation:
Where F is force produced, a is cross sectional area of coil, n is number of turns, I is current and g is gap between core and coil. With my original design, this equation told me that a 0.5kg projectile would leave the barrel at around Mach 25, which may have been a slight overestimation. Using a much more reasonable estimation of current at 250A I calculated the force to be around 80kN which is still quite absurd. I had to accept the fact that this equation was ever so slightly optimistic, but it did give me a very good idea of what values I should alter to get the most force out of my Coilgun. Its easy to see what values need to be increased or decreased, so although it was a bit off in its estimations it allowed me to easily see how I would make a more efficient Coilgun. With this in mind I made it the priority to have the number of turns and current as high as possible, and the gap as small as possible. Its also important to consider that you can get more turns with thinner wire (a) but current must be less so as not to melt the wires. Also, more turns means higher resistance, which reduces current. Therefore I opted for very thick cable for the coil, with a comparatively small number of turns and quite high current. An alternate option would be to go for as many turns as possible with as thick wire as possible and not worry about current, but that could be very expensive.
Later in the instructable I used the equation with several different wire thicknesses and two different barrel diameters to find the best wire for my needs.
Step 2: Circuitry
I also went through a lot of different circuit diagrams. What I originally didn't realise was that supercapacitors had lower voltage ratings, so my first circuit could pump out about 37kA. I also wasn't aware of the fact that this value of current could probably melt any cable I could get my hands on and of course, the wire would have to be incredibly thick to have such a low resistance so as to allow this current to pass through it. Using Yenka I was able to work through various designs to come to a relatively simple analogue circuit that would allow me to charge and switch the capacitors, as well as a detection circuit that could switch the coil off the moment the projectile reached the end of the coil. This may have been easier to do with a micro-controller, since thats what a lot of people just getting into electronics are using, but I prefer to use analogue electronics where possible.
The major feature which I wanted to have, which I haven't seen many other Coilguns use is a system in which an infrared detector would shut off the coils and keep it off for around one second using a 555-timer. One point worth noting is that my detector was placed immediately after the coils so that it could turn off just as the projectile protruded through the coils, however it may have been more efficient to have the detector somehow within the coils, so that they could turn off before the projectile got to the end (preventing any work being done in the wrong direction). Another option would have been to have the coils on a timer circuit, but I couldn't find a definite figure for the optimum time it needed to be on for.
Another important point was that I wanted the capacitors to charge very quickly just to make using the Coilgun more enjoyable. For this reason I included in my circuit diagram an array of 8 high power resistors, each of resistance 1Ω, with combined resistance of around 1Ω also but critically they could collectively handle upwards of 250W, potentially even over 1000W in a short burst.
One major dilemma was how I could switch the very high current. I had the option of either using a large IGBT with a massive power rating, or stacking upwards of 20 MOSFETS in parallel. Ebay was my best friend here, as I was able to find a very powerful used IGBT, capable of switching 400A continuous or 600A in a 1ms burst, for under £20. MOSFETS were still a viable option but they could have been very expensive and the wiring would no doubt have been a gigantic pain.
Also on ebay I found some very reasonably priced 500F supercapacitors, and was able to make an offer of 5 for £40 which is cheap enough to worry me slightly. In my circuit I wire them in series because they are such low voltage, even though this limits the capacity of the bank. In fact, the entire bank amounts to 100F, but the 12V I charge the capacitors with means that the total energy stored is a substantial 7200J and has a charge of 1200C. If I wanted to fire with 250A of current, and each shot took .25 of a second, I would be able to fire the gun around 20 times in one charge, which takes 4 minutes.
In reality, since supercapacitors are so low voltage, I was unable to get any substantial power out of the gun because I couldn't get the resistance low enough to give me a nice high current (Remember Voltage=Resistance x Current).
12V is super safe when you think about electrocution (although fire hazards are a much greater issue with the supercapacitors) even if the stored energy is massive. If you have enough experience to stay safe with this project but not enough to be confident with high voltages, I would suggest using a low voltage and experimenting with your barrel/projectile size to get the most out of the solenoid.
However, if you're a very experienced electrical engineer I think that, using very high voltage capacitors, you could probably have it breaking re-entry speeds (hyperbole maybe). Of course I can't advise anyone to do that nor can I be held responsible for any harm you do with it.
I chose 12V because it is a really convenient voltage that will allow me to use car batteries, jump starters or power supplies with ease and not have to worry about damaging the capacitors, but moreover it is very safe and its almost impossible to do any real damage to yourself with dry hands. However, if I wired the capacitors in parallel and used 2.7V (their max voltage) my stored energy would be something like 10,000J, and charge around 7000C. This is a lot more but not necessarily a big issue within the scope of this project. It was a better option for me to use a less powerful capacitor configuration than have to worry about how I would charge it.
Step 3: Materials/Tools
- I spent a little more than I would have liked on this project, and one major reason was that I overestimated how many safety precautions I would have to put in place (overpowered fly-back diode, more resistors than necessary, coil too thick initially), so with this in mind I think it would be very plausible to do this project for under £100, depending of course on what materials are available to you. Most of the expensive components I bought second hand on eBay or ...borrowed from my college.
- Approx 1mx0.5m polycarbonate sheet - there are a lot of different options for what to construct the casing out of, but I had more than enough polycarbonate already on hand.
- Thin Aluminium tubing for barrel - Mark I barrel was 1in dia, which meant the projectile was far too heavy. Mark II was 15mm (12mm internal).
- Rod for projectile - The metal you use needs to have a high permeability, so steel with a low carbon content that has been annealed is best.
- Various screws and bolts
- MDF for stand (optional)
- Small steel brackets
- Chunk of hardwood for handle
- Wires of various gauges (but the thicker the better for most of the wiring)
- Enamelled copper wire - Initially I used 5mm but I found 3.15mm to be optimal for a barrel of 15mm Diameter
- Matrix board
- 5x 500F 2.7V Supercapacitors
- Various small capacitors
- Various resistors
- 8x 1Ω 25W resistors
- Large current capability Schottky diode - this was an area which I overspent in. A much smaller diode would have been able to handle the fly-back current given the very short period of time it happens over.
- Large IGBT - again, my current ended up much lower than this huge component could handle, especially after the barrel amendments, so there would be no need to get one as powerful as mine
- Digital display voltmeter - looks super cool.
- 555 timer
- Op-Amp - this may be an area in the circuit which can be optimised. The op-amp likely isn't necessary but it does the job anyway.
- 2x 9V battery
- High current switch - I bought one that could handle 300A, but once again this was much too high as it only handles 10A.
- Infra-red LED
- Infra-red Photo-transistor
- Something to cut sheets with - a band-saw would likely be ideal. I'm not sure how a table saw would cope with polycarbonate, but a jigsaw worked okay.
- 3D printer - optional of course, I just wanted to be able to say I used it to make my trigger. It was very useful though.
Step 4: Barrel
I made the barrel first so that I could plan the rest of the circuitry around what resistance the barrel ended up being. The actual construction was quite simple, but if you're making this for yourself I would implore you to find a different way to construct this section because I was under time constraints and there are a lot of improvements I'd like to make. However, the science of making it as efficient as possible was fun and interesting to research.
I've written a great deal about this in my final report which you can download in the introduction, but the basics are as follows. A thicker wire will always be more efficient (because it has less resistance) until either there is too greater gap between adjacent wires due to their roundness (can be solved by using square wire), the wire becomes to thick to wind or if there are other sources of resistance elsewhere in the circuit which makes having low resistance wire useless.
I used the equation for force of a solenoid here:
I've no doubt it's extremely inaccurate, but the point is that it shows which factors are most important and can tell me which factors are better to choose than others, even if the final value it spits out for force is extremely inaccurate.
In order to find the ideal diameter of wire I found out the length of wire in 1kg for each diameter of wire, and found out the resistance, current, number of turns for the barrel I am currently using. I also figured out how these factors would affect the force in the force from a solenoid equation and found 4mm to have the highest coefficient of F. I also calculated all of the same values for a smaller diameter barrel to see if the same trend would appear, and I found that a thicker wire was slightly more effective for a smaller diameter barrel.
The force was largely dependent on the resistance, which was difficult to calculate. I took the resistance of the coil of the specific diameter, and added 0.03 which was the approximate resistance of the rest of the circuit. When I removed 0.03 from the values of resistance, my graph resembled an exponential slope with the thicker wire being always the better option. This is not true however, because thicker wires become extremely hard to wind, especially around the planned smaller barrel, and they would also mean that the resultant coil is extremely wide. 5mm wire was very difficult to wind by hand around my 28mm barrel, I think that it would be almost impossible to wind it around a 16mm barrel, let alone using 6mm. Also, the further away from the core the coils are, the less effective they will be in adding to the force delivered, which is not accounted for in the equation. In both instances (25mm barrel and 16mm barrel) I found 4mm to be the most effective diameter of wire given my parameters. My current barrel has a calculated force of 1837N (which is of course an overestimate, and assuming g=2mm), and using 4mm and a 16mm barrel my projected force is 5600N. Although both figures may not be accurate, not accounting for voltage drop over the IGBT and various other areas of inefficiency, they still show me what will likely be most effective.
If all that science has put you off, the important point is that different thicknesses of wire will be more or less effective depending on the barrel width. I calculated that for a barrel with a diameter of around 15-25mm, 4mm diameter wire will likely be most useful from both a practicality and efficiency standpoint.
Step 5: Building the Base
In an ideal world, I'd love to have made the whole casing out of aluminium sheeting welded or brazed together just to try out something new, but this would be super time consuming and expensive, not to mention that aluminium is conductive and I didn't want a shock off it.
I had an excess of polycarbonate sheeting, which seemed like the ideal material because it is extremely strong and no force of recoil or anything else would be able to crack/shatter it. If time constraints had been more generous I would have liked to try all sorts of joining techniques such as finger joints or dovetails but the most time efficient way was simply to screw/bolt it all together, which ended up being useful because it meant I could easily take bits apart to make quick modifications.
I started with the front tray which I can hold with one hand, but also contains the capacitors and has the barrel sitting on top of it. This was simply a case of cutting rectangular shapes and screwing them together. This would be super easy using a bandsaw, but I am not fortunate enough to own one. Perhaps a table saw would also be useful, but I cannot speculate as to whether or not polycarbonate would cut well using a table saw. As a rule, if you think something might go wrong using a table saw then don't do it if you like having all your fingers intact. Always use pilot holes! Particularly when screwing through polycarbonate, because the heads of my screws were twisting off before they could penetrate.
Here I learned the true importance of measuring properly and using close tolerances where appropriate, because in my haste all of the little imperfections I left in the panels added up and it began to look very wonky. I would always recommend using a file or even some kind of fast sanding tool to get every piece to the exact right measurements because, as I learned with this project, rushing and not measuring properly will probably make your project take even longer than if you'd taken more time in the first place.
I built the base in 3 sections, the front tray with the barrel on top, the rear with the handle and the top rear which contained most of the circuitry and control panel. All of them bolted together and could be easily removed. Unfortunately I didn't make any detailed plans because I was very time-constrained at this point, and I literally built all of this base by feel. Honestly this is quite stupid and I would advise that you make very detailed blueprints with measurements and actual time and effort since you're building something very dangerous. Not planning is my biggest flaw.
Step 6: Extras
The Trigger was 3D printed on my printrbot simple which, by the way, is extremely cheap and can perform just as well as printers several times its price if you put some time into calibrating it well. I've uploaded my 3D models if you want to just use my trigger (some assembly required!). Not the most efficient design, but by gluing a spring to the top of the central arm and attaching it above, you’ll have a fairly nice trigger which you can customise as much as you’d like. It should be fairly obvious from the 3D models, but pressing the trigger down forces the arm to press a push-to-make switch.
The Handle was carved from a chunk of pine. This was actually really fun and I wish I could have spent more time selecting a nice wood and getting it to a nice finish. I more or less (ab)used a jigsaw to carve the whole thing, in a very Jimmy Diresta like fashion. If only I could have used a band-saw instead!
I made a Control Panel to go on the upper rear section, unfortunately I didn't get to incorporate all of the features I’d have liked to, but it looks ace when its counting up the voltage across the capacitors. Again this was just very basic cutting and shaping of acrylic because I have so much of it, and the voltmeter was bolted in.
On the left side there is a Resistor Array, this is my favourite part as all of the anodised resistors look great lined up like they are, and their high power allows me to charge up 300 milliamp hours in 2 minutes (phone batteries have a similar charge (within the same power of 10 at least!) and they seem to take forever to charge). Once again I simply cut out a long bar of acrylic sheet and bolted it on to the side.
On the right there is my Fly-back Diode which, as discussed earlier is massively overpowered and doesn't need to be anything like as big as it is. Fortunately it happens to look quite cool. To secure it to the gun I bolted in one end and the other I had to attach using a steel bracket and bolting it on (see images).
Step 7: Projectile
I put a lot of thought into the design of the projectiles, and this was largely helped by the following instructable:
I strongly suggest reading through it, as it gave me a lot of points to consider and helped me to figure out the ideal length, diameter, material and shape. In my final report I've written extensively about the design of the projectile and you can download the report in the introduction, but the main points are that I've used an annealed, low carbon steel with a length approximately 5x its diameter, (the diameter being as close as possible to the inner diameter of the barrel) with as little material missing from the front (i.e. minimal grinding to a sharp point) to maximise the amount of magnetic flux used effectively during the gun's firing.
I'd love to have made it on a metal lathe but unfortunately I don't have that luxury, so I made it by simply filing down some round bar.
Step 8: Putting It All Together and Test Firing
As I'm sure you already gathered from the video at the beginning, the final product was quite disappointing. This is solely due to my descision to use supercapcitors. While safer and totally feasible to get very high currents using a very low resistance, 12V circuit, it just turned out to be too impractical. However, if I was more confident with high voltages and had the money to buy some massive electrolytic capacitors, I believe I could make this coilgun much more powerful with only minimal modifications.
Although less successsful than most of my other projects, I probably learned more than with any other project. Lessons learned are as follows:
- Never sacrifice accuracy or measurements for time constraints, because the likelihood is it will end up taking longer when nothing fits together.
- I've learned a great deal about electronics and gained a lot of confidence I didn't have before. I've always found that the quickest way to learn a lot about something quickly is to dive into a very ambitious project you definitely don't yet have the know-how to complete. Even if you fail you'll learn an incredible amount in a short space of time.
- While its a nice idea (and feasible) to try and get large currents from low voltages, its too impractical for an amateur like myself and you'd be better off working with high voltages (safety, disclaimer etc).
- Drill long pilot holes if you're putting screws into polycarbonate, because that stuff doesnt like to be screwed into.
We have a be nice policy.
Please be positive and constructive.
Gahhh people if you are going to build these you HAVE to calculate the ESR of your capacitors, and what your peak current of a discharge is going to be. You need to know how long the discharge takes and what the peak current is or you are just fumbling in the dark. The greencaps have such
Start there: https://www.coilgun.info/mark2/rlcsim.htm
The peak amperage those caps can supply is only marginally higher than a jump started Li-Po battery.
Additionally, it's snapping to the center of the coil. If you don't know the discharge time, you can be snapping to the center and just stopping (or at least majorly slowing down). An SCR isn't controllable after you trigger it until the voltage drops back off. So you HAVE to find the pulse time. This is the MOST important part.