PowerLabs 3kJ Single Stage Coil Gun Page
The goal of this project is to develop a high power, highly efficient single stage linear electromagnetic mass accelerator capable of reaching over 50 joules of muzzle kinetic energy on a single stage. This will be used to research and demonstrate the principles of magnetic acceleration, and the accelerator is to be kept as a test bay for different coil/projectile combinations. The use of a single stage keeps design parameters simpler and leaves more room for experimentation.
10 Nippon-Chemi Com (brown) and 10 Powerlytic (blue) electrolytic capacitors inside a closed acrylic box measuring 70x15x15 cm and weighting a total of 13.5kilos (24pounds). All capacitors are rated for 450V max and store a 1500uF charge. This amounts to 150Joules each, or 3000Joules in total. The capacitors are interconnected using 2cm wide, 1mm thick copper buss bars (for low inductance) in a (2X10)/2 configuration (2 banks of 10 capacitors in parallel each, and the banks are in series), for 900V 7500uF. ESR for each capacitor is 130mOhms, and total bank ESR is 0,026Ohms.
Due to the very large currents encountered in this device I had to resort to a rather pricey solid state switch; A ceramic disk "Hockey Puck" SCR rated for 1200V, 1000A RMS. Peak current for this device is on the 14kA range. peak voltage is 1.5kV. This particular unit measures 8cm dia. and is 2cm tall. I paid $200 for it. Peak power rating for it is 21Megawatts for a half wave pulse. Its switching time is quite low however, since it was probably designed for a low frequency inverter or similar. Energy wise this is quite an overkill for my current application, but as energy and power levels are increased it will give me the ability to switch them without any risk of a junction meltdown. It will also allow me to perform "dry shots" (discharging the capacitor on the coil without any projectile in place) in order to evaluate the cause of the coil form crushing problems I am having with my other (multi stage) coilgun.
In order to conduct properly and dissipate all of the heat formed during high power duty this type of SCR has to be mounted on a special heat sink under pressures in the order of 1ton. To the right is the one I am using. It is made up of two parts (the SCR is sandwiched in the middle) and weights almost 4.5kilograms of solid anodized aluminium! With the SCR in place it measures 12 X 12 X 23cm. Although this heatsink is not required for pulsed operating since no significant amount of heat is produced as the SCR only dissipates 1 - 3% of the 3kJ of the shot, it was the only way I found of securely mounting it.
is the assembled SCR on the heat sink.
The SCR is so large that I had to make notches on its sides in order for the mounting bolts to come through!
In order to maximize its potential for conducting short, high magnitude electrical impulses the SCR gate must be given a sharp current pulse several orders of magnitude greater than the nominal junction trigger ratings. This rapid pulse allows the junction to go from non conducting to fully conductive state as quickly as possible, hence minimizing junction heating due to ohmic current absorption at the trigger phase. This circuit is still experimental and was designed and drawn by Steve Roadway. I will post results here as soon as I have tested it.
Here both the projectile and the glass coil form around which the coil will be
wound are shown. The projectile consists of a 9mm diameter, 5cm long ferrite
ceramic slug with its leading 2mm machined into a smooth curve. It has a density
of 5g/cc and hence weights 15,5 grams. It is a very large, very heavy
projectile; larger than any handgun bullet I am aware of (by comparison, a
typical 9mm pistol round is less than half its length). I do not expect it to go
very fast, and it will also probably shatter on any impact, but I am using it
nonetheless because its large weight allows me to slow down the pulse to a
manageable power value for my SCR and at the same time have a very high magnetic
field coupling between the projectile and the coil, which should translate into
a higher efficiency gun.
The coil form is a 11mm diameter glass tube with 1mm thick walls. It is 18cm long. I have had several bad experiences with glass coil forms crushing on firing in my other gun, but unfortunately this was the best I could obtain. The projectile is varnished with polyurethane varnish in order to minimize the impact between it and the tube's walls as it aligns itself to the magnetic field.
Here my first prototype coil is shown, with two projectiles above it (the top projectile is 1cm shorter than the one pictured above). The coil has a 10Cm winding length (hence a 1:1/2 coil/projectile ratio) and 3cm thickness. It consists of 200 turns of AWG 10 (2mm dia) soft copper wire with modified polyurethane insulation (Class "C", 180C duty).
Specifications are (measured with a digital LCR meter):
35mOhms resistance, 90uH inductance, 313uH inductance with the 5cm long projectile fully inserted.
This coil is quite heavy and should be capable of dissipating the full 3kJ of each shot without overheating.
Update: As was the case with the coils used on my multi stage coilgun, the glass coilform used here was pulverized during the first discharge. It appears as though the coil contracts during firing, and this gives rise to a crushing action on the form. I have temporarily substituted the glass coilform for one made from an aluminium gas tank for a pen torch. The new coilform is 12cm long and has 2mm thick walls (too thick). I cut a 1mm gap along its entire length to eliminate eddy current losses.
Here my first solid state 3KJ coilgun prototype can be seen. The SCR/Heat sink/protection network are all integrated into one block and that block is bolted onto the capacitor bank. The bank itself is closed off with an acrylic top (for safety) and there is a 9cm wide, 1mm thick copper plate exiting 10cm ahead of the SCR buss bar which serves as the other terminal. The Coil's terminals are 10 and 15cm long respectively, and the coil itself is taped onto the side of the capacitor bank (this is temporary). Ahead of the bank is the charger (notice how it has a voltmeter and an ammeter for monitoring charging rate and power) and next to it is the variac. In the beginning of the bank is the voltmeter and the white wire coming off the SCR leads into the trigger of the device. The entire assembly is 40cm wide, 20cm high and 1 meter long, and weights close to 30 kilograms. It looks awfully mean just sitting here on my workbench ;)
In order to better understand what goes on during the discharge, I have modeled my entire coilgun on MicroSim PSpice and ran a few simulations on it in, evaluating the current and voltage during the discharge (in order to ensure that the SCRs ratings can not be exceeded even during a "dry shot" event) and the pulse length (I hope this will allow me to eliminate experimentation and guesswork from coil sizing and make it possible for a coil that is just long enough to carry the pulse with maximum efficiency to be designed).
first simulation was a dry shot with the objective of determining the
shortest pulse length possible and maximum current achievable with the current
setup. The values obtained are 5,2KA peak current and 2,8mS pulse length. Power
of the discharge is hence 4,7MW. The values obtained indicate that in order to
stand that pulse the SCR would have to be rated for at least 900V, 1kA, which
means that my SCR is
well below its operational values with this coil and should even sustain
repeated firing at high rates, but the 300A 1200V bolt-type SCRs I am using on
the multi stage coilgun would be instantly destroyed on the pulse. The dry shot
simulation can be found here.
The second simulation can be found on the picture next to this text. It represents the closest match to what I believe happens during a discharge. All values inputted were calculated and/or measured to the highest degree of accuracy possible. The coil inductance used here is the average between the minimum (no projectile) and maximum (projectile in the middle) inductances the coil can have. Of course, in the real world the inductance is a dynamically changing variable, but since it goes through those two values I think an average of them should suffice for a simulation.
Peak power here equals to 3600kW (4kA x .9kV). There is a voltage reversal at 2,5ms, and anything after that will be blocked by the SCR protection diode (half the wave). If we integrate the area under the curve we should get the energy available for the projectile (I^2R losses included), and if we divided that (1440J) by how much energy we can couple into it (average coupling coefficient = (coil/projectile ratio)/2 = 1/4), we get the maximum theoretical value for electrical energy conversion into kinetic energy, all losses included. Doing this on my TI-83 Graphing calculator I got a value of 12%, which implies a projectile kinetic energy off (3000X0,12) 360J, which on a 15,5 gram projectile translates into a muzzle kinetic energy of (KE = 1/2MV^2) 215m/s, or 774km/h!
However, when I tried repeating the calculations I came up with a different value. I am not sure what I did wrong yet, but as soon as I have it figured out I will include all calculations here.
Click on the image to the left to watch a 158kb .mpg video of the assembled prototype firing at full power (the meter on the picture reads 948V: The capacitors were in fact overcharged for this experiment) on an aluminium soft drink can. The slug hits the bottom of the can and sends it spinning towards the 1" thick acrylic shield, which is than knocked down as the can bounces back onto the floor. A lot of work still needs to be done as far as efficiency goes, but I have discontinued experiments with new coils until I can obtain a new coil form... Enjoy the video!
Also be sure to check out my MULTI STAGE coil gun page!
Questions? Comments? Mail me!
People have been here since 01/04/00
Last updated 11/02/10
� 2000 - 2002 by Sam Barros. All rights reserved.
Removing any material from this site for display without consent from its author consists in an infringement of international copyright laws and can result in fines up to $50000 per infringement, plus legal costs. So ASK ME before you remove anything from here.