GBPPR Electromagnetic Pulse Experiments - Part 1

Project Overview

This article will discuss the construction of a very simple Electromagnetic Pulse (EMP) generator.  This particular design won't be capable of destroying every computer in your neighborhood, but it will give the constructor a good overview of the concepts which make up electromagnetic pulse warfare.  It should be noted that you will be working with very dangerous voltages and currents on this project.  Muslims, Eurosavages, Digg kiddies - you will be immune to this.  People with brains may wish to exercise caution.

The EMP device described in this article will work as follows:

A surplus high-voltage DC power supply will be used to generate an output voltage between 3,000 and 4,000 volts.  This high-voltage will then charge an old 8 µF, 3,000 VDC General Electric Pyranol capacitor (the "C") via a current limiting resistor.  When the capacitor is fully charged, it will discharge via a spark gap.  The spark gap circuit will be part of an inductive circuit (the "L") which, along with the capacitor, sets up a natural resonant frequency.  To generate the actual emitted pulse, part of the LC-tank circuit will be made of fine wire, a lightbulb filament in this case.  Since the lightbulb filament can't handle the high-current discharge from the pulse capacitor, it will be instantly vaporized.  This "exploding wire" will essentially be turned into an electromagnetic pulse which is radiated from an impromptu parabolic dish antenna.  That's the idea at least.

Since the resonant frequency of this particular EMP device is very low (20 kHz or so), it will actually do very little damage to any electronic devices in its path.  Most "real world" EMP generators aim for the UHF or microwave RF bands by using tuned mechanical cavities.  The shorter wavelength of microwave RF energy is ideal for being coupled into the circuit board traces in the target electronic device.  The resonant frequency of this generator can be increased slighty by replacing the lightbulb with a long piece of small-gauge wire.  Experiment with the length, composition, and diameter of the wire used.

An optional ferrite core transformer will also be described.  When this ferrite core is placed over a power cord, or other similar exposed wire, the electromagnetic pulse can be directly injected into the target system.  This is a much more efficient method than the "exploding wire" idea.

The most critical component in an EMP generator is the high-voltage pulse capacitor.  The ideal capacitor will be non-polarized and with a low internal inductance and resistance.  The internal resistance inside the pulse capacitor will determine how fast, and to what final level, it can discharge.  Commerical pulse capacitors that are designed for this purpose are available, but their price is usually out of range for the hobbyist.  Search amateur radio swapfests for old Polychlorinated Biphenyls (PCB) high-voltage capacitors.  You can usually pick them up for free due to their high cost of disposal.  Just don't let the hippies know about them, or they'll try to tax amateur radio experimenters next.  Several capacitors can be banked together in parallel to increase the energy output.  You can also place low-voltage capacitors in series so they can handle higher voltages.

EMP Block Diagrams

For more detailed information on EMP generators and construction information on a similar design, refer the book Electronic Gadgets for the Evil Genius (ISBN 0-07-142609-4) by Bob Iannini, or see

Construction Notes & Pictures

High-voltage power supply parts overview.  The heart of the power supply is a General Electric 9T63Y2065G12 DC Power Supply.  Its maximum output voltage is approximately 12,000 VDC at around 1 mA.  It takes a standard 120 VAC input.  A variac will be used to control the power supply's final output voltage by controlling the input AC voltage.  If a variac is not available, it is possible to use the low-voltage secondary winding from a standard AC transformer to control the power supply's input AC voltage.

The other support components are an AC line filter, a Radio Shack metal-oxide varistor, a panel-mount SO-239 RF connector, a green neon light, two fuse holders (one panel-mount), an AC outlet, two binding posts (with rubber grommets), a solid-state relay, a surplus 0-120 VAC variac, a bunch of surplus ferrite cores, and an old military radio surplus voltage transformer.  This will be turned into an isolation transformer for feeding the variac.  Everything will be mounted inside an old ammo box.

And put it all together as so.  The AC line filter and dual fuses are probably overkill, but they're a good idea when working with EMP devices.  The isolation transformer is used to isolate the variac from the AC mains in case the "hot" and "neutral" lines are reversed.

Front panel rear view.  After the input AC line filter, the "hot" voltage line passes through a solid-state relay.  This relay will allow the high-voltage power supply to be remote controlled from a safe distance.  The solid-state relay's remote control is nothing more than a 9 volt battery and a switch.

The isolation transformer is made by tapping a 110 VAC secondary winding.  You can often find these transformers, or the military surplus radios they're inside, quite cheap at amateur radio swapfests.  Their secondary windings can only handle a few milliamps of current though.  There is a 100 ohm resistor and 0.1 µF AC-rated capacitor on the isolation transformer's primary winding to act as a "spike snubber" circuit.  The use of an isolation transformer before the variac is not required, but highly recommended.

The power supply's high-voltage output is on the left via the binding posts.  The binding posts are set inside rubber grommets to isolate them from the metal case.  The maximum voltage this entire setup can handle before arcing over is only around 6,000 volts.

Completed front panel view.  Main AC input is via the outlet shown.  The SO-239 connector is for the remote control.  The variac's knob and main input fuse are on the right.

Rear view showing the output high-voltage binding posts.

Parts for the remote control.  All you need is a metal outlet box, a cover plate, a switch (with guard), a panel-mount SO-239 connector, a 9 volt battery with a snap and holder, a panel-mount LED and 470 ohm resistor, a 0.01 µF capacitor, and assorted mounting hardware.

Put it together like so.  An extra ferrite bead was slipped over the control's positive line.

Overview of the completed remote control.  Connect it to the high-voltage power supply via a good length of RG-58 coax with PL-259 connectors on each end.

The spark gap will be made from two drilled and tapped steel mouse balls.  Ideally, you'd want non-ferrous materials that are nickel or silver plated.

Flatten one side of the mouse ball with a grinder and drill an appropriate hole for the thread tap.  On this project, the mouse ball for the "hot side" (capacitor side) will have a #12-24 tap.  The other mouse ball will have a 1/4"-20 tap.

For the "cold side" of the spark gap, use 1/4" brass, bronze, or copper hardware.  It will be mounted on a small piece of wood which is then attached to the side of the pulse capacitor.  The "inductive" elements are mounted to the spark gap via a standard copper ground lug.  Use brass bolts with the head cut off for the threaded brass rod.  It's all kinda retarded, but it works.

Completed spark gap assembly.  The gap hasn't been set yet.  The air gap will be set at around 1 mm per 1,500 volts used.  You can use a spark plug feeler gauge to help set the initial gap width.  The split washers and nuts secure the mouse balls to their respective threads.  The gap can then be further adjusted by turning the brass rod in and out, then tightening the securing hardware.  The gap on this device was set to "spark" at around 3,500 VDC.  This is slightly over the voltage rating on the pulse capacitor, but it should handle it.  Be sure to fully discharge the pulse capacitor before adjusting the spark gap width.

The wooden spark gap holder is attached to the side of the pulse capacitor using some two-part epoxy putty.

The "exploding wire" holder will be made from an industrial heat lamp.  These have a nice porcelain lamp base and parabolic reflector.

Rear view of the porcelain lamp base.  This is what sets the maximum operating voltage on this EMP generator.  This particular model lamp base could only handle around 6,000 VDC before arcing over.

Solder two pieces of #6 solid copper wire to the porcelain lamp base like so.  You may wish to add a little "Q-Dope" to prevent high-voltage arcing between the two wires.

Next is the high-voltage input circuitry.  The current limiting resistor(s) and RF choke need to be mounted on little standoffs to prevent arcing.  These are secured to the side of the case with nylon hardware.  Note the extra ferrite beads slipped over the incoming power lines.  These, along with the 4700 pF bypass capacitor, help to suppress any "back-EMF" when the spark gap fires.  The RF choke shown (red cylinder thing) is from an old switching power supply.  Its value is around 8.5 µH, which is probably too low for this application.  Oh well.

The current limiting resistor(s) should have a high-voltage rating.  Lower resistance values will charge the pulse capacitor faster, but this may stress your high-voltage current source.  The time (in seconds) it takes for the capacitor to charge is approximately: t = C * V / I.  Where C is the capacitor's value (Farads), V is the capacitor's charging voltage (Volts), and I is the capacitor's charging current (Amps).

Make a securing bracket for the pulse capacitor from some 1-inch wide alumimum bar stock, two pieces of 5/16" allthread, threaded couplers, and other assorted hardware.  Drill two holes in the alumimum bar stock and place it so it can sandwich the capacitor to the bottom of the case.

Completed closeup picture.  A large hole is cut into the front of the case and the porcelain lamp base is epoxied in place.  The solid copper wires which make up the inductive elements of the circuit are clamped into the grounding lugs which attach to the "ground side" of the pulse capacitor and to the "cold side" of the spark gap.  These wires will need to handle several hundreds (or even thousands) of amps, so exercise solid construction practices when securing them.

Completed overview.

Parts for the direct coupling ferrite transformer core.  This is a total hack, but it does appear to work quite well.  The split ferrite (or powdered iron) cores are a swapfest grab, so start looking out for those!  You'll also need a 2-prong AC plug, some smaller sized grounding lugs, and an AC socket to lamp screw-in adapter thingy.

Wire it all up as shown.  You'll want to put a little hot glue on the AC plug to keep the prongs from moving.  Be sure the copper wire turns around the ferrite core are not shorted, or that they are so tightly wrapped they crush the brittle ferrite material.  Experiment with the number of turns needed, but you'll have a hard time getting more than three.  Slip some vinyl tubing over the copper wire for protection.

Screw the coupling ferrite transformer assembly into the lamp base as shown.  Secure the other half of the ferrite core with a plastic sliding-jaw clamp.

This connection method didn't work out too well, as it was too heavy for the small lamp socket.  You are better off just wiring the ferrite transformer directly off the spark gap.

To operate, just run your target's power, ground, Ethernet, etc. wire through the ferrite core and zap away!  Try to wrap the target cable multiple times through the ferrite core, if possible.

Be careful, as it is possible for the transformer's halves to shatter due to the induced current in the ferrite material.

Overview of the EMP generator and the high-voltage power supply connected together.  Note the metal-armored power cord on the high-voltage power supply and the coaxial cable for the remote control.  All the cables in the local area should be shielded to protect them from the electromagnetic pulse.

Various different lightbulbs were tried, and they didn't do too much except explode into little pieces.  It looks like you'll need to have a pulse capacitor output in the hundreds of Joules to have any really significant results.

Another possible EMP option is to connect metal Slinkys to each side of the spark gap to act like antennas.  This could help radiate the electromagnetic pulse a little bit more.

Schematics / Block Diagrams

Notes / Links

Return to Homebrew Military & Espionage Electronics Page