PowerLabs Rail Gun Research!

Project Introduction:

This page is no longer being updated. The PowerLabs Rail Gun research is being continued at the new PowerLabs Rail Gun 2.0 page (click on link to be taken to that page).The PowerLabs Linear ElectroMagnetic Accelerator (“Rail Gun“, or “RailGun”) research project officially began on February 21st 2001 when Sam Parler, research Director of Cornell Dubilier Electronics (CDE), e-mailed me praising me for my experiments with electrolytic capacitors. The experiments he was referring to all involved discharging such capacitors at very high rates to produce intense magnetic fields for hyperplasic deformation in metal as in the PL Solid State Can Crusher Page and also accelerating metallic objects by means of either eddy current induction and subsequent repulsion as in the PL Disk Shooter, or by ferromagnetic attraction as in both the single stage gauss gun and the multi stage linear magnetic accelerator pages. Mr. Parler asked if I had any future projects in mind, offering to have his company sponsor me a new capacitor bank (at this time I had recently moved to college and either discarded or left in storage all of my parts and equipment). This was an offer I could not refuse: although over the years I had attempted just about every experiment I could think of with capacitor banks, my previous attempts at building Rail Guns were never very successful: they would either weld the projectile to the rails, or blow apart from plasma pressure; never producing enough force to shoot a projectile. Nonetheless, I still believed that a larger capacitor bank would allow me to produce a Lorenz force strong enough to accelerate a metal armature before it became welded onto the rails.
On his e-mail Mr. Parler mentioned some tests he performed for CDE where currents as high as 40Kiloamperes were obtained from a single electrolytic capacitor. I found this very impressive and quickly realized that a capacitor bank built from such capacitors could perform as well as any other pulse capacitor, and as such deliver enough power to successfully accelerate a metal armature in a linear accelerator design.
Such a capacitor bank would, however, be well outside of my University student budget. Mr. Parler not only helped me design a capacitor bank which would fit my request: 3.2kV, 16000Joules (this was actually more than twice the energy I was hoping for!), and the capability to deliver a current pulse of up to 100thousand amperes, but also had CDE pay for all the costs of the bank, hardware and shipping. I can not thank him and CDE enough for it!
With a preliminary design and a capacitor bank to power it I submitted a research proposal to my departmental advisor at the Mechanical Engineering/Engineering Mechanics department of Michigan Technological University. He arranged for me to work at the Advanced Propulsion Laboratory under Dr. Brad King’s supervision. The gun was built over the summer of 2002 (my first summer in college, still as a freshmen), in 150 hours of machine shop work and was first fired on October 19th 2002, at Ressonance Research during the Wisconsin Dells Teslathon. It has been through several upgrades and been fired numerous times since, including televised shootings for TV6 and The Discovery Channel, as well as having been featured on 2 newspaper articles in local Michigan papers.
In 2004 the gun, currently located at the Michigan Technological University Dynamics research laboratory, has been decommissioned and is currently being replaced by a new, vastly improved design; The Rail Gun 2.0…

 Project Description and Goals:

The primary objective of this project was to successfully design and construct a linear electromagnetic accelerator utilizing Lorenz Forces from a high magnitude electrical impulse to propel an armature down two parallel conducting rails. A simple graphic representation of the effect is seen on the image to the left; the current flows up on rail, through the armature which travels perpendicular to the rails, and down the other rail. The result is a magnetic field between the two rails (B = 2x(u0/2pi)x(I/r)) and an intercepting field by the armature. The rails repel one another (F = u0I1I2L/2piR) and they both repel the armature (F = ILB). Since rails are both fixed the net result is a propulsive force on the armature, which will be accelerated forward by electromagnetic means (A=ILB/M). This differs from conventional mass accelerators in that no gases are used, and it differs from conventional electromagnetic accelerators in that the field trails behind the projectile at all times; since no coils are used coupling occurs at a much greater degree and efficiency values would tend to be higher. The potential velocities achievable can also be much higher.

The accelerating force between the rails and armature depends on the magnetic field present (which in turn is a product of the rail separation distance and the current through the rails) and on the area this field acts upon. In order for acceleration to be maximized optimum parameters must be chosen for all these variables (and others which will be mentioned later). The rail separation distance was set at twice the electrical breakdown threshold of air at the peak power supply voltage assuming dry at air STP; 6mm. The 2x safety margin was chosen due to dielectric creepage considerations. As far as pulse current is concerned, it can be seen that in order for a high acceleration to occur, VERY high currents must be employed, which in turn requires a high voltage so that circuit impedance can be overcome and the required current can be achieved. The final design is a series of tradeoffs where higher voltages bring higher currents but at the cost of a higher rail separation distance. A typical design utilizes around 4 – 10kV, with higher voltages being used at higher energies. This particular design calls for a 100 kiloampere pulse which should be accomplished at 3.2kV. Good part of the many amateur Rail Gun attempts seen on the Internet failed because their power supplies were simply incapable of supplying the currents required; even “small” military and research designs employ currents in the 300KA+ range, with some of the larger guns going over 5 million amperes per pulse. Acceleration drops off quickly with lower currents and at a certain point drag becomes higher than accelerating force and the projectile becomes welded by the resistive heating that occurs. At the same time however, a very high current will cause dramatic rail erosion and resistive losses.
Once the goal of successfully accelerating the armature was achieved the device was then fine tuned for maximum efficiency so that it can be used as a platform for investigating the following aspects of electromagnetic acceleration:
1- Rail Erosion: This is currently the biggest issue holding back the implementation of linear electromagnetic accelerators: The very high currents employed to accelerate the armature must flow through very small contact patches, which often arc and dissipate a large portion of the energy available. The resistive heating that occurs quickly rises the surface temperature well above any metal’s vaporization point, causing extensive erosion. By studying the erosion caused under different conditions it is hoped that a solution will be found for maximizing rail life (plasma armatures, different armature compositions, conductive greases, coatings, lower current pulses, are some of the options being considered).
2- Capacitor bank Life expectancy: Should a capacitor failure occur it will probably be due to magnetic forces within the capacitor causing connections to physically tear due to mechanical motion; in this case it must be determined at what current this occurs so that a practical limit for the power supply performance can be determined (3 spare capacitors are available in case of individual failures).
Also, the capacitance of the capacitors is expected to increase due to cathode anodization (a phenomenon only common to electrolytic capacitors subjected to voltage reversals), and their ESR will consequently increase due to the dissolution of the resulting Hydrogen gas into the electrolyte. This will be analyzed through current waveforms.

Power Supply and pulse shaping:

The power supply consists of 32 Cornell-Dubilier Inverter Grade capacitors, each rated at 6300uF and 400V (450V Surge). Operating Temperature is -40C to +95C. These capacitors utilize the latest technology in electrolytic capacitor construction to store 640J each in a can measuring only 3″ dia x 5.63″ length and weighting 900grams each. To put in perspective, that is 40 times the amount of energy it takes to electrocute a human in a package the size of a cola can!
The capacitors are assembled in 8 sub banks wired in series, each bank containing 4 capacitors in parallel, for a total rating of 3200V nominal, 3.6kV peak charge and 3088.3uF (measured) capacitance. Stored energy (1/2CV^2) is thus 16kJ nominal, 20kJ at peak charge (see graph). Each individual capacitor has a 50KOhm 10W wire wound resistor for charge equalization and also to serve as a bleeder to prevent unwanted charge buildup when power is switched off. They are charged through a 900Ohm current limiting resistor and can be safely discharged through a 6.25kOhm resistor bank mounted inside the bank.
The capacitor bank has a measured internal resistance (ESR) of 14.7mOhms and an internal inductance (ESL) of approximately 1uH. The sub banks were designed using a Genetic Algorithm program that matches every individually measured capacitance value in such a fashion that the overall capacitance only varies by 0.02% from the total average capacitance between each sub bank. The design current is 100kA from the bank, meaning 25kA per capacitor, with a theoretical minimum pulse length calculated at 56uS, giving the pulse an equivalent frequency of approximately (1s/(56uSx2)) 9KHZ, which implies that the current will only flow on the outer (66/sqrt9000)= 0.7mm of the copper inter connecting bus bars due to skin effect. In order to counter skin effect related losses the capacitors are inter connected by very large surface area (30in^2) oxygen free copper strips each 0.064″ thick (1.6mm). The actual pulse length of the bank was measured at 63.4uS. The bank is currently fitted with a Fluke 80K6 6-kV probe for voltage monitoring.
To the left some of the inductors experimented with on the gun can be seen. By adding inductance to the circuit the pulse length can be increased, making power available for the projectile for a greater duration of its travel through the rails. At the same time, ohmic heating is decreased and the heat generated is dissipated over a larger area. It is expected that an optimum balance exists between the benefits of lower losses and acceleration over a longer period of time  and the drawback of lower accelerating forces due to the lower currents provided by the inductor. 

Rails, Rail Enclosure and Armature Design:

The rails consist of two 33.5cm lengths of 6mm thick, 3cm wide (12x.25×1.76″) silver plated oxygen free copper. This length was chosen to keep resistance and cost at a minimum whilst still allowing some flexibility in lengthening the electrical pulse.
Currently the Rail Gun is a “hot rail” design; I.E. the armature acts as the power switch when it meets the energized rails. For that, the first two inches of the rails is milled down 1/31th of an inch and covered by a glass-filled Teflon composite which insulates the rails from the projectile so as to ensure a stable magnetic field behind it once power switching takes place. The rails are held together one on top of the other (wide sides facing, so as to maximize contact area and magnetic field interaction with the armature whilst at the same time minimizing contact resistance) by a G-9 (Garolite/Melanite impregnated inter woven fiberglass) composite enclosure utilizing virgin grade Teflon spacers to keep them parallel to one another at a distance of 6mm. G-9 was chosen for its exceptional tensile strength (68KSI) and insulating properties. Teflon spacers were chosen due to the material’s high thermal resistance (one of the highest working temperatures of any polymer commercially available) and low coefficient of friction (the lowest known to man). The maximum expected tensile force between the rails can be estimated by the formula F = u0I^2L/2piR. Using half the rail thickness as the radius and 100kA as the maximum current through the rails, and assuming that all the current is carried by the entire length of the rails the repulsive force thus becomes F = ((4pi*10^-7)*100000*100000*0.3/2pi*0.002) = 100KN (1Ton). This can actually become a lot greater during plasma armature tests and with projectile injection. This force is equally distributed amongst 16 5/16″-24 Grade 8 ultra coated steel bolts, washers and pressure nuts so as to prevent buckling under the firing forces.
Two armatures were tested: Al1100 25x25x6mm and plasma-backed Teflon. Aluminum was chosen being used because it will melt before the rails do, and thus cut down somewhat on rail erosion. The length of the projectile was adjusted so that its effect on acceleration efficiency can be verified. On the Teflon projectile, the aluminum backing becomes a plasma during the discharge and recycles some of the efficiency losses in the form of propellant pressure. Unfortunately this propellant pressure was so great that it ultimately caused the failure of the rail gun enclosure in multiple places. Thus a new rail gun design was developed to be able to withstand plasma armature pressures, the Rail Gun 2.0 

 Projectile Injection:

If full power was to be applied to a static armature the rails and whatever was between them would instantaneously melt under the intense localized heat produced by Ohmic heating as 100thousand amperes tried to make it through the contact resistance. In order to prevent the Rail Gun from becoming a spot welder it is necessary that the armature be moving with some initial speed prior to electromagnetic acceleration. Most amateur designs fail because of lack of knowledge of this. In this design the armature is injected by a gas gun consisting of a 1000CC Schedule 80 PVC gas reservoir connected to a 30cm long barrel through a reducer that goes from 1/2″^2 to 1/4″ x 0.6″ through a 60 degree taper. A 1/2″ diaphragm pilot operated solenoid valve controls the gas flow and essentially serves as the trigger for the gun. Approximately 5% of reservoir capacity is used in one shot. The system is designed for 500PSI (35ATM), enough to consistently fire a 6 gram aluminum slug out at 150m/s, or a Teflon slug at 195m/s (634.5fps, 696km/h, 432.6mph). The barrel is an exact replica of the Rail Gun made from Polycarbonate with virgin grade Teflon rails for maximum efficiency and velocity. Currently the injector is being operated with Nitrogen gas, which, along with decreasing rail oxidation, also has 30% lower molecular weight than air, providing higher velocities. Ideally the injection velocity should be as high as possible, as it will allow the armature to travel the longest distance over powered rails and thus minimize localized rail erosion and kinetic friction. It would be desirable for a future Rail Gun design to employ a supersonic injector. 

Charger:

Each individual capacitor in the 20000Joule capacitor bank is fitted with its own 50kOhm 10W resistor for charge equalization and also to serve as a bleeder which will drain the capacitor and prevent unwanted charge build up. These resistors cause the bank to dissipate 130Watts continuously when held at peak charge. That, and the 20kJ energy storage capability mean that the charger must output a lot of power to achieve the desired fast charge rates. The current charger consists in a variable autotransformer (Variac) and a microwave oven transformer (MOT) charging the capacitors through a half wave voltage doubler (0.86 capacitor and diode combo). Peak charging current is 0.8A and peak voltage is 3.5kV. Both are monitored from the power supply. This approach was chosen because of low cost, availability and the fact that microwave oven transformers are current limited, which simplifies the charge circuit significantly. The charging system is further protected by a fast blow 15Ampere fuse and a current limiting resistor within the Rail Gun. I hope to replace this with an inverter-based charger some time in the future. 

 Completed Device:

To the left is an electrical schematic of the gun. The projectile switching is assumed lossless and an arbitrary value of 1uH has been assigned to the rails temporarily. This PSpice 9.0 schematic can simulate the electrical pulse that takes place when the gun discharges and by comparing the simulation with actual oscilloscope waveforms values for rail inductance and resistance can be calculated. Then, by varying the inductance and resistance of the rails I can find the exact values once the expected and obtained waveforms match up (see Results below).
The gun has 4 major parts: The pneumatic injector consisting in an air tank/valve/teflon barrel assembly, the 20kJ capacitor bank, the Rail Gun per se (rails, enclosure and spacers), and a high voltage charger to charge up the 20kJ capacitor bank. Below parts of the gun can be seen (hover your mouse over the pictures for a description, or click on them for a full size image):

First the injector/air tank attaches to the Rail Gun, than the rail gun attaches to the capacitor bank, and finally the capacitor bank attaches to the charging supply and the tank attaches to the air compressor or Nitrogen tank at the charging supply. Below you can see the completed Rail Gun/Injector/Capacitor bank assembly: Please check the construction effort page for pictures and descriptions of how each individual part was made...

 Results!

The Rail Gun circuit has a measured shortest discharge time (full rail length) of 63.4uS. This is very close to the 56uS I originally designed the gun for, and means that the peak power of the discharge will be in the hundreds of Megawatts range (approx. 320MW). Notice the lack of oscillation in the discharge. Discharge current should be 80 – 90thousand amperes. This power brings with it a whole range of difficulties, with initial tests causing extensive vaporization of the projectile and rail damage. It thus became necessary to fit the gun with an inductor so that the discharge time could be lengthened so as to reduce the resistive losses and increase acceleration time.

The frame capture to the left shows the gun firing with 15kJ and no current limiting inductor; there was extensive rail and armature erosion, enough to produce a massive plasma cloud that was fired from the gun and traveled towards its target before extinguishing itself. The low velocities achieved are an indication that this metal vaporization is where most of the energy was spent.

The frame capture to the right shows the gun firing with a plasma armature at 6.5kJ without a current limiting inductor; the gun is very loud when fired this way. The plasma is a lot hotter but it lacks sparks; an indication of lower rail erosion.

15kJ power shot video (You may also view it online as a frame-by frame capture)

Clicking on the still frame to the left will download the latest RailGun test video; a 8.3kJ shot utilizing a pulse lengthening inductor and an aluminum backed Teflon projectile (plasma armature). Velocities are now estimated to be supersonic, although the muzzle flash of the gun does not allow a chronograph to be used to measure its velocity…

More on research objectives:

RailGuns are by far the most spectacular type of electromagnetic accelerators ever developed. They hold the record for fastest object accelerated of a significant mass, for the 16000m/s firing of a .1 gram object by Sandia National Research Laboratories’ 6mm Hypervelocity Launcher, and they can also propel objects of very sizeable masses to equally impressive velocities, such as in the picture to the left, where Maxwell Laboratories’ 32Megajoule gun fires a 1.6kilogram projectile at 3300m/s (that’s 9megajoules of kinetic energy!) at Green Farm research facility. Their ability to propel objects at speeds which are simply impossible for conventional (chemical or mechanical) means makes them extremely useful for a range of functions. The most obvious one being defense, where most of the research money in this area comes from nowadays, but NASA has also been funding RailGun research for hypervelocity impact simulations which will allow shields to be developed which will protect orbiting aircraft from high velocity debris surrounding the earth. NASA is also researching the possibility of a launcher which would deliver payloads into orbit at a fraction of the cost of a rocket launch. Similarly, other studies are under progress for the utilization of RailGuns in Fusion Fuel pellet Injectors for experimental nuclear fusion reactors, and also for metallurgical bonding; the University of Texas (UT) in Austin, identified that the Electromagnetic Powder Deposition (by a railgun) process is capable of achieving a coating of deposit material with bond strength equal to the base material while achieving less than 3% porosity. This should soon become a repair method for jet engine components, as similar processes are also being employed to produce extremely high shock pressures on collisions between dissimilar materials in an attempt to produce new materials.
Rail Gun technology also has the potential to revolutionize transportation: Sandia National Laboratories is working on a Segmented Rail Phased Induction Motor (SERAPHIM), a new type of linear induction motor offering unique capabilities for high-thrust, high-speed propulsion for urban maglev transit, advanced monorail, and other forms of high-speed ground transportation. Linear induction motors are already in use for applications such as airport transit systems, subway systems, theme park rides, and industrial material handling systems.

 Current Research Status:

After performing over 30 firings, the rail gun enclosure failed due to excessive in bore pressure from a plasma armature test. As such, it has been decommissioned and a new RailGun was developed. The PowerLabs RailGun research continues with this new, improved accelerator. Information on it is available at the RailGun 2.0 page.
Information on the tests performed with the first Rail Gun, along with videos, images, and detailed information on its failure are available through The RailGun Testing Page link.

 Future Research:

Ultimately I would desire to be employed in a professional research where my knowledge and skills displayed here could be put to use so as to further advance this promising technology. On a more immediate level, my plans are to take everything learned from this first prototype and build a second one, maximizing efficiency and minimizing rail erosion and maintenance costs. This second prototype is already under construction and can be seen on the Rail Gun 2.0 page. A third prototype will probably employ advanced power switching technology, probably solid state (SCR) as well as external magnetic field augmentation around the rails. It will use a supersonic injector and a higher energy storage capacitor bank. The exact specifications of PowerLabs Rail Gun 3.0 will depend a lot on what is learned from the second prototype, as well as what funding becomes available by the time that research begins. Any help -financial or otherwise- in designing it and acquiring the required components is of course welcome.
Below some of the components obtained can be seen: a 5500V, 3000uF, 46000Joule capacitor bank. Also seen is a photograph of 12 neodymium N45 grade supermagnets each 2×2″ base with a 2×1″ top and one inch high. These are the strongest magnets I have ever seen; in one word, they are Dangerous. Strong enough to crush fingers, drive metal objects through soft wood, or explode on impact when two are released together. I will be employing their fantastic field strength as external field augmentation for higher efficiencies in Rail Gun 2.0. They have been sponsored by Engineered Concepts; the cheapest and best magnet supplier on the ‘net!

 Construction, Plans, Schematics, How-To, FAQ:

Due to overwhelming demand for plans and schematics for the gun I may some day, time permitting, put together an amateur Rail Gun design and construction manual. This particular gun took 150 hours of design and construction work before it could finally be fired and although specific plans are not available at the time, a complete and detailed log of the construction effort, including pictures and videos, is available for free at the Rail Gun Construction Effort page. Information on the tests performed and cause of failure is available at the Rail Gun Testing Page.
For current information on PowerLabs’ RailGun research Project go to the PowerLabs Railgun 2.0 page.