POWERLABS Electromagnetic Solid-State Can Crusher!


 Over half a thousand US Dollars worth of equipment… The second largest semiconductor currently in production… 260 THOUSAND times the magnetic field strength of Earth… 22.7 MILLION Watts of electricity… Enough energy to kill a person 250 times over… Enough power during a single discharge to supply a large city… A 20 kilogram + device…Several days worth of work… The ability to smash an aluminium can with no physical contact, and to do it so fast and heat it so much in the process that it sounds as though the can is exploding as the drink inside it boils off…


(Sorry, couldn’t think of a better reason 🙂

The Device:

Electromagnetic Can Crushing is most probably a spin-off from the metal forming branch of pulsed power research. Several devices available in the market nowadays utilize the same underlying principle to remove dents from airplane fuselages, magnetically form metal objects into desired shapes, and unite metal parts by crimping. No one knows for sure who came up with the idea of using magnetic repulsion to smash aluminium cans first, but there are currently 4 can crusher projects on the Internet. All of them share one thing in common: The use of a large, bulky, high-voltage capacitor to deliver, through a spark gap, a massive electrical impulse to a coil of wire inside which the can is inserted. This enormous current induces an even larger current (by transformer effect: Suppose the coil has 3 turns of wire. The can, representing a single turn, will thus have a current that is 3 times as large induced on it) on the thin aluminium surface of a metal can. This rotating current (also known as an eddy current) has a magnetic field associated with it (proportional to the current involved), which, being the same polarity as the field on the coil, will cause the can walls to be repelled and hence to collapse in upon themselves.
Although the principle is simple, the actual design is tricky as it involves a fine balance of components in order to obtain as high a current rise time as possible (which dictates a low inductance) to produce as high a magnetic field as possible (which requires a large inductance) and as high a peak amperage as possible (which requires once again a low inductance and a low resistance). In practice, most projects make use of very powerful capacitors which, operating a high voltages, are able to push massive (100thousand amperes+) currents through short coils of wire. These currents sound like a shotgun blast when they arc over the device’s spark gap, and the power is so great that even at low design efficiencies it becomes possible to snap a can in two.
Living in an apartment, I am of course unable to operate such a device indoors. Furthermore, the price of shipping a multi hundred pound high voltage capacitor to Brazil is too great for my limited budget… Those two factors, allied to the challenge that such a device would present to design, made me opt for an entirely different approach to electromagnetic can crushing: I decided to make a device that would use no high voltages, no spark gaps, and no bulky, expensive high voltage capacitors…
The Solid State Electromagnetic Can Crusher!

SetupAt a basic level, it seems as simple as using a Silicone Controlled Rectifier (SCR) in place of the spark gap, and using lower voltage capacitors (since SCRs can’t cope with HV) on the device itself…
However, can crushing requires very high current rise times (DI/DT), which are very difficult to obtain at low voltages, and very high peak currents, which are impossible for silicon junctions to deal with. I pondered for quite a long time whether these problems could somehow be worked around… With an industrial 2800Volt, 2000Ampere ceramic disk (“Hockey Puck”) SCR, and a large enough capacitor bank, I though that perhaps a magnetic field could be produced that would be powerful enough to perform the task, without requiring so much current that the SCR was destroyed instantly. Fine tuning would of course be vastly more critical than on a conventional crusher design, but still, the mere possibility, allied to the fact that I already possessed a large enough capacitor bank that remained unused, in storage, made the decision for me an easy one.


– The Power supply:

For the power supply, a bank of 20 inverter grade electrolytic capacitors, as seen above, 10 are Nippon-Chemi Com (the brown ones) and 10 Powerlytic (the blue ones). They are all arranged inside a clear Plexiglas box measuring 70x15x15 cm and weighting a total of 13.5kilos (some 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) and, for this particular experiment, are connected as a 900V bank at 7500uF.
Above the capacitors is a Digital Multimeter reading the actual charge voltage (in this case 0volts). The multimeter is essential as it allows me to monitor the charging rate, the actual charge voltage, and any residual charge left in the capacitors after a discharge. It is also essential for safety. The box serves to insulate the capacitors from one another and prevent electrolyte from spilling out in the event of a capacitor failure. It is worth mentioning that such a capacitor bank is VERY LETHAL!

– Power Switching:

 SCR sitting on my handA Westcode 24870-708-01: A 10cm diameter, 4cm high, 350 gram ceramic disk SCR rated for an incredible 2000Amperes, 2800Volts RMS. These are the second largest SCRs currently in production and are not available on the common market. This particular unit, with a retail price of $1300, was obtained from a disassembled industrial induction oven. It is capable of switching 5,6megawatts continuously, and has a peak power rating of 90megawatts! The SCR is mounted on a 4,5kilogram aluminum heatsink with a force of 1000kilograms (2000pounds). I actually broke a wrench mounting it!!! The SCR/heatsink combo measures 12 X 13 X 23cm. Peak current rating for this SCR is in the 30000Ampere range.

Side picture showing the mica insulators behind the bolt washersSCR mounted on heatsink with protection diode and resistor on top, bolted to capacitor bank on its rightThis heat sink was put together using two 10cm long, 8mm diameter steel bolts with 2 steel washers each and non slip nuts. Obviously having the metal bolts unite and pull together both parts of the sink presents a very serious problem since the two halves can be at as high a potential difference as 3kilovolts. My solution was to cover the bolts with two layers of plastic shrink wrap, for a 2mm all around cover, and than use 1mm thick mica squares behind each one the washers. This succeeded in keeping the entire device insulated up to the maximum rated voltage.

– Crushing Coil:

 3 turns of 10mm^2, 6 diameter (insulation included) polyurethane rubber insulated power cable cast in epoxy. Every single coil prior to the epoxy cast was destructed in the first firing; some crushed themselves, others ripped apart and shorted against the can. The epoxy cast allows the coil to be re-used a few times before it crushes itself.

– Circuit Layout:

 During the microsecond or so it takes for the SCR junction to go form non conducting into conducting stage, its resistance will vary from near infinite to near zero. During this time it will dissipate a substantial amount of the power going through it. When you are dealing with powers reaching peak values in the tens of Megawatts, dissipation can be a severe problem. The junction often develops very rapid, localized heating that can not possibly be sunk by the heat sink on time to avoid it from burning out completely. The solution is to minimize the junction turnover time by applying a short, high magnitude trigger pulse. Also, once the can has been crushed any residual magnetic field on the coil will be converted back into a reverse polarity current surge (Counter Electro Motive Force, or CEMF), which has the potential to destroy the SCR and to damage the capacitor bank. This pulse must be blocked before it can cause any damage. The circuit below, designed by Steve Roadway, accomplishes both those feats.

SCR trigger circuit and protection network

How Strong Is it?


Simulation of a full power can crusher discharge

Thus we can derive that the peak power is 22,5MW (25kA @ 900V).

 The magnetic field strength (B) at any point on a solenoid is given by the equation B = �0NI/L
�0 is the permeability of free space (4pix10-7TmA-1)
pi = 3,141592654…
N is the number of turns on the solenoid
I is the current (in amperes)
L is the length of the solenoid (in meters)
B is given in Teslas, where 1Tesla = 10 000Gauss

 With a 3 turn, 1,2cm long coil wound with 4mm diameter copper wire running the full 25Kiloamperes of the pulse provided by the SCR, the magnetic field would thus be 7,8Teslas, or 78THOUSAND Gauss… That is equivalent to exactly 260 000times the magnetic field of the Earth. During early tests of the device my computer monitor was magnetized from over a meter away. I now keep the device well away from all credit cards and sensitive electronic devices.

When a current I moves through a conductor of length L in the presence of a magnetic field B, the conductor experiences a force F according to the equation F=ILB
F=Force on conductor (Newtons)
I=Current through coil (Amperes-turn)
L=Conductor lenght (Meters)
B=Magnetic field strength (Teslas)

However, if this magnetic field were steady, nothing would happen to the can as there would be no current flow. What causes a force to be produced is the fact that the can, too, has a magnetic field. For that to happen there must be current induction, which is a product of the rate of change (DI/DT) of the field. Any idea how to calculate the force inflicted on the can as this field goes from zero to 7,8T and back to zero in 0,3microseconds? Mail me!


Sam removes the crushed can from his can crusher setupThe can crusher has been able to achieve a high enough peak power and magnetic field to allow the crushing of standard aluminum soft drink cans. However, since what matters is the rate of change of the magnetic field, as opposed to its final value, and knowing that SCR junctions take a certain time (approximately 2uS in my case) to change from nonconductive to fully conductive, the delay in junction turnover time is hurting performance: The can is reduced in diameter by 2cm or so per discharge, as opposed to being completely crushed, which is what would be expected from the energy delivered. I am looking into ways around this problem, but on the meanwhile you are welcome to observe these videos which give a glimpse of its performance. The first one was filmed during a presentation I have on electromagnetic induction to my IB HL-2 physics class, and the second one was filmed inside POWERLABS headquarters; my room.2 meter tall Liquid Nitrogel column resulting from the crushing of a liquid nitrogen filled can

 In the first video (328KB, .MPG) you can see a close up of the can as it crushes and appreciate just how loud the process is, without any interference from the sound of the switching device (SCRs are noiseless). In the second video (640KB .MPG) you can see a can full of Liquid Nitrogen (350ml) being crushed and shooting up a column of smoke. The 67% reduction in resistance on the can due to cryocooling should allow it to crush to a much greater extent. However, at -43 the polyurethane rubber insulation on the coil glassifies, and thus it becomes so brittle that it crushes itself. On the first try the coil shorted itself out against the can. On the second try it only broke in several places because the can was pre-cooled prior to the coil being inserted, and thus there was not enough time for it to completely glassify. Click on the pictures to download videos. The videos are high quality and can be played full screen.