It has been brought to the attention of POWERLABS that there is serious interest on the Internet on the topic of coil guns. However, all of the sites available describe small coil guns which are a pain to build and in the end yield unimpressive results. POWERLABS already leads the field with its 3Kilojoule coil gun (as of the time this page was written that is by far the largest on the web), but now wishes to produce a multi stage coil gun so as to further its research and perform efficiency comparisons. The objectives here are to study high power solid state switching and to produce a working multi segmented coil gun. The research will be conducted in modules, adding one segment at a time and optimizing it before adding the next module. The final design calls for 5 modules, each one storing and switching 1300Joules, which would make it capable of firing 7kJ with a very slight over voltage on the capacitors. If that energy is ever achieved, the projectile (which has its weight fixed at 4.5grams) may well break the sound barrier! (note: 7KJ amounts to the energy released when 8grams of gunpowder deflagrate, or in other words, its roughly 1/5th the propellant force of a small caliber naval gun, and 3 times that of a high powered sniper rifle!).
If we assume, for curiosity’s sake, a conservative 3% overall efficiency (all losses accounted for), the projectile velocity would be 310 meters per second, or 1100km/h (that’s mach 0.91). For now the goal is set at half of that (150m/s, 540km/h) and the prototype nearing completion (3 stages 4kJ total energy) is expected to reach that once it is finely tuned. Two more stages will than be added as money permits, and if the SCRs are able to handle the current.
Experiments are also being conducted with cryogenically refrigerating the firing coils with Liquid Nitrogen down to 196C below zero (-320.8F), which would cause their resistance to drop by approximately one third. Given the extreme heating experienced during firing (the smaller coils can actually reach 200C and burn up during firing sequences), it is assumed that resistive losses in the system are high and that decreasing them by that much will have a very significant effect on overall gun (if I can call it that. I don’t actually endorse the term since it gives the negative connotation that this is going to be used as a weapon, whilst the actual goal here is to prototype a research tool) efficiency, which is why I am hopeful for the unprecedented (in non commercial models) 3% efficiency. At 4% efficiency the gun will go supersonic (1296km/h). Breaking the sound barrier with this sort of linear electromagnetic accelerator would represent a true breakthrough in this field of research, and as far as I know has never been done before (it is no easy task though!).
Details on the final design: This drawing illustrates somewhat the 5 stages all arranged inside a box with their corresponding SCRs and coils on top. The box in the beginning is an inverter. Since the power consumption is moderate, it should be feasible to run this off a 12V battery. At 130 X 10 X 8 and weighting roughly 15kilos, it would be very easy to transport anywhere it was required. However this is not meant to be fired from a free standing position.
Looking for information as to how they work? Please check out The POWERLABS Gauss Gun Page
A note of warning on energy storage capacitors: Capacitors such as these, holding several thousands of joules at several hundreds of volts are absolutely lethal for anyone who is to come into contact with them. Aside from the severe electrocution hazard, they are also capable of exploding when overcharged and causing metallic objects to explode when shorted (see below). Only qualified people should attempt working with such dangerous devices.
Two brand new Philips Type A21509-532-01 Pulse Rated Electrolytic capacitors per stage. Each capacitor is rated at 450V surge at 6300uF (approximately 650Joules). They measure 22 X 7.6cm and weight about 1 kilogram. These differ from normal electrolytic capacitors by having screw-on terminals that are almost 2cm in diameter! They are meant to be used in switching power supplies, flash lamp power supplies, and other applications where high RMS and peak currents are encountered. They are, therefore, absolutely perfect for this application.
I obtain these capacitors brand new and unused from a special supplier. Apparently they were brought in to the country without paying taxes (Brazilian importation taxes are ridiculous), confiscated and subsequently sold at an auction. Because of these special arrangements I can get them for $80 a piece; still a lot of cash, but very good considering their market price of $300 a piece. I am yet to see anything nearly as good to power a coilgun and it was truly a shame that I only managed to buy 7 before their stocks ran out:(
The capacitors are arranged as 2 in series per stage (900V, inter connected with a 15mm wide, 3mm thick 95mm long (.6X.1X3.9″) copper buss bar. Each stage amounts to 3150uF and thus 1300J at max (.9kV) charge.
NEW: I now have seven capacitors. The second stage is complete, and the third is on the way. I can hardly wait to see it fire at 4kJ!
After experimenting with larger, heat sinked SCRs and finding them rather over-rated for this application, and blowing a 200A SCR in its 3rd firing, I decided to switch each stage using a brand new 1200V 300A bolt-type SCRs. Pictured here is a TENGEN model KP300A.12 I got straight off the box… The picture doesn’t do much justice to its size: That thread it has in the bottom is 1/2″ (1.5cm) in diameter! Specs are as follows (from what I could translate off the Chinese specs sheet:) :
Nominal current: 300A
Nominal voltage: 1200V
Trigger Current: 42mA
Trigger voltage: >20V
This is what it looks like open:
The nominal (RMS) power rating for these is 360Kilowatts. Since I never plan on running more than perhaps 500W nominal power through my coilgun (otherwise the coils would melt in seconds), a heat sink would be a complete waste of space. The peak power is what we are interested here. For this particular SCR it is in the order of 4.8MW (4Kiloamperes for 10mS with 1 second recovery time). Keeping the surge current below that critical value (the fact that the voltage through it is lower than its maximum rated voltage also helps since it keeps junction stresses low) is essential, as is the triggering which will allow the SCR to switch from non conducting to conducting in the shortest time possible so that not much energy will be dissipated in its semi-conductive junction. For that a special trigger circuit was devised which will apply a massive trigger pulse to each stage, timed by an optical sensor:
(note: having projectile sensing was chosen instead of open loop sequence as it will allow the firing energy to be varied whilst keeping its efficiency at a maximum)
<Circuit under testing, diagrams will be released upon completion>
Because of the extremely large currents encountered during the firing of the gun, one has to be very careful with fast cycling rates. The SCR junction requires 1 second recovery time from peak currents, and the capacitors have got enough thermal mass and a low enough resistance to fire several hundred shots every second. So one might think it possible to make this into a very fast shooter, perhaps as 60 shots per minute… However, as experience has thought me, firing as little as 10 shots in a row through even a large coil without allowing it to cool down rises its temperature so much that the 180C class insulation can actually burn right off! Improvements in coil design such as active cooling could, of course rise this figure, but it would make the entire device unnecessarily bulky and heavy. For mobility reasons a charger was chosen that would allow it to cycle quickly enough to keep an audience interested or to conduct experiments with ease, but still wouldn’t be so large and bulky to make transport of the device difficult. The original design called for a 500V at 1A filtered DC charger. A breakthrough in the design subsequently required high voltage charging, and a new, 900V @ 600mA charger was designed. Even with 4 stages the charging times should be kept below 30seconds, and even though an inverter would be a lot lighter, the custom wound transformer and the diode bridge still take up less space than a single stage of the gun, making it practical to move around. The transformer was wound with a 1:1,5 step up ratio, taking up to 250V input (through a variac) and outputting 650V at most. After rectification this translates into just a few volts over 900V when Vforward drop in the diodes is taken into account. So the charger can actually just be plugged in and it will charge the capacitors to the correct voltage. To protect the diodes and to prevent hazardous over voltages/currents from being injected into the mains supply a L-C-R filter network composed of two wire wound ferrite cored toroidal chokes, two 9kOhm 50W wire wound resistors, and a 20A 1100V MOV string is placed between the stages (which have individual protection networks themselves) in case of over volting (just hope it never reaches threshold: If it does, it will explode violently!).
Pictures and a schematic of the charger are coming soon.
The projectile is the silicon steel cylinder shown, measuring 2 X 6cm and weighting 4.5grams. So far 8 coils were tested. The coil forms used were acrylic, phenolite, and glass, which is what I am using right now. Coil length was varied from 10 to 20cm, and right now 15cm is being used. Wire gauge was varied from 1mm to 2mm, with 2mm being used now. The tube has 1mm thick walls and the wire is wound directly over it. The number of turns is varied so as to change the inductance and hence control the pulse length to fit any energy or projectile speed combination as is required. When lubricated with fine silicone oil this tube allows the projectile to pass through with virtually no friction (actually the lubrication is only important for the initial acceleration, as the magnetic field acts uniformly on the metal and makes it float down the middle of the tube).
Here is a small summary of the most recent tests (note: As new tests are performed the previous ones are deleted from the page to give space to new data): Summary: First 3 coils wound on plastic coil forms. The 2mm thick walls prevented the projectile from experiencing a very large magnetic field and hence introduced large losses. Plastic straw was than tested, but the coil is definitely moving during firings, since it *grabs* the projectile whereas before it would allow it to pass straight through. Reinforced cardboard snapped.
Finally glass coilforms were decided for.
AWG 14 wire coils: Became exceedingly hot during firings. One burned up. Obviously too much power is being wasted in resistive losses.
AWG 12 wire coils: These are a pain to wind. 3 layers proved optimal on a 14cm long coil. Projectile placement just before first turn.
AWG 10 wire coils: This is where I am right now. Requires a drill to wind. Results:First coil wound with 10AWG (2mm thick) annealed 180C class insulation wire. It required a solid coilform and a high torque cordless drill to turn the form. This coil was made 12cm long, which is shorter than the other coils, and once again 3 layers of wire were used (180 turns) with a shrink-wrap jacket around the coil to hold the turns in place. Initial tests showed that this is the most efficient coil so far, with its performance at 380V matching shots performed at 450V by the other coils. So far every single coil with a glass coil form has had the form destroyed. I am currently looking for a better coil form material so I can continue this research.
NEW! 900V, high “K” coil! (currently being designed).
Anecdote: For the second test, whose picture you see underneath, the camera was placed right over the left side of the coilgun. Since it is an USB port camera, the gun itself had to be placed roughly 20cm (8″) from the computer monitor so the cable would reach it. As soon as the trigger was pressed and 650J of energy flew through the coil in a single 3thousand ampere pulse, a magnetic field was created which was so intense that it COMPLETELY magnetized the monitor. The image swayed sideways as though the monitor had been dropped, and than all the colors were messed up and the image was severely distorted. Luckily, several degaussing sections later and a few reverse magnetizations with a neodymium bar and the monitor was functional again. Fortunately the field decays by the inverse square of the distance, which means that it doesn’t stay that powerful for very long, but still, at 1300J, I have to be very careful placing all watches, credit cards, and magnetic media at least one meter from the coil. If a coil was to be developed specifically for the projection of this enormous field, it might make it possible to completely run all magnetic media within a large range… Food for thought for all you EMP enthusiasts out there… I’ve done my experiments with EMP and it cost me a phone, so I am rather weary of repeating them in an equipment packed laboratory…
TEST NUMBER 2:
With the purpose of calculating the gun’s muzzle velocity, a test was set up in which the projectile’s predicted target was marked out using a laser beam through the center of the solenoid, and the slug was fired at maximum energy and the difference between the height of the predicted target and the actual target were inputted into the equation V = DX x SqRt(G / 2DY). Several shots were fired and the projectile placement was changed until maximum velocity was obtained.
For these shots the measured velocity was 66.667meters per second, or 240kilometers per hour. This means a muzzle energy of 10Joules, and an efficiency of 1.54% (650J input) Click here or in the picture of the punctured can or the test above to watch a video (98KB) of it firing. Notice how the gun does not make any sound… I absolutely love that feature!
NEW! 270km/h muzzle velocity reached with the new (thicker wire) coil. Now projectile kinetic energy is at 12.5Joules (with 650J input)! The goal has been reached and the next stage is to double the firing energy. A new prototype firing 1300J was built:
Click on the picture to the left to watch the NEW (1300J) prototype fire! As of now there are no velocity measurements, but this video gives a pretty good idea of the energies we are dealing with now. Note how the slug comes out so fast that it cuts right through the can and still has enough energy to knock the plastic shield down! (128KB)
NEW PROTOTYPE (Photos courtesy of Michell Zappa):
The capacitors are now placed side by side with 2cm wide, 3mm thick copper buss bars connecting them (see picture above, under “POWER SWITCHING”. Also notice how a filter network (2 fast (15nS) 300Vdiodes and a 550V 20A MOV) has been placed in parallel with the coil. This has proven to be an excellent addition to the project, as it blocks out all of the CEMF produced by the coil right after the shot (some significant energy, I might add: The resistors become quite warm), and in that way protects the SCR and capacitors from damaging reverse currents. It also should help efficiency somewhat by cutting off the reverse polarity magnetic field that develops during the “ringing” produced by the L-C circuit, which if allowed to ring down by itself would pull the projectile back in somewhat while at the same time stressing the capacitor dielectric.
Yeah, still using tape… think I might build this in an acrylic box, with the capacitors underneath the box and all coils on top… But that’s going to cost some serious cash, which is not worth investing until the project is completed and out of the prototype stage.
New: The coil continues to mysteriously crush the glass coil form every time it is fired. I have begun wondering if this is truly due to coil contraction (if so the problem could be fixed by having the coil on an epoxy cast) or because of the projectile bouncing around it. In order to test this, I decided to attempt a “dry” shot (discharging the capacitors on the coil without any projectile inside). Obviously, once the capacitor has fully discharged on the solenoid a massive magnetic field will form. And without any ferrous material to dissipate this field, it will collapse and come back full strength as back EMF into the SCR. Because of that I installed the 2X 500V 35nS fast diode and the 550V 20A MOV protection network to block out that EMF before it could get to the SCR, where it could cause damage. I slowly increased the variac setting, watching the needle on the voltmeter gradually climb up to 380V. Than I switched off the charger and pushed the button on the trigger, watching the coil to see if it would jerk or if the glass coil form would break. Having done this hundreds of times before, you can imagine my surprise when a blinding flash of light and a large explosion took place! The blast was so strong that I could feel the pressure wave hit me in the chest, and fragments cut through the tape that was holding the network in place. On close examination, the cause for this powerful explosion showed up: The fast diodes, upon blocking the counter EMF, were completely vaporized (see picture)!!! I was already aware that they were blocking a significant amount of energy since they heat up considerably with every shot, and this was the first test with the 10AWG coil. My thoughts were that the coil would dissipate the energy once it was prevented to ring down with the capacitors. However, since this coil has much lower resistance than the other coils, all that energy had to go somewhere, and it went to the diodes…
But it gets worst… Apparently the energy was enough to both vaporize the diodes AND ruin my expensive, brand new 200A SCR, as well as the whole bank of capacitors :(a 300 dollar loss).
Testing will continue once new SCRs and capacitors are acquired.
NEW: Work on multiple stages is stopped until a new version, firing at 900Volts can be perfected.
Don’t forget to check my single stage Gauss Gun page too. 3kJ in one single stage… Watch that coil smoke! 🙂