High Voltage

An Inertial Electrostatic Confinement (IEC) fusion reactor uses a high potential difference to accelerate the hydrogen ions together. This potential difference needs to be high enough to overcome the Coulomb Barrier. In practice, this happens around 27kV for amateur systems. The voltage of a system determines what percentage of available ions have the kinetic energy to fuse, whereas the current determines how many ions per second can fuse. We need a very high voltage, but the current can be very small, around 10 milliAmps. Finding a power supply that meets these requirements is very difficult. Any that are made are marketed towards research labs and as such, are extremely expensive. I chose to wind my own coils to meet these requirements. At 27kV, fusion is possible but it is slow. I designed my supply to operate at 35kV to make sure I could detect fusion.

High Voltage Safety

Before anything else, it is important to note the safety concerns of these high voltages. 27 kV and 10 mA is absolutely enough to kill or injure with just a small mistake. Very high voltages such as this will do anything to find its way to ground, whether that be through inches of insulation or through you. This power supply is a switcher type and operates at a high frequency of around 40 kiloHertz, which introduces another concern of capacitive coupling. Because the output is rectified, the supply rapidly switches from 35kV to 0V. Capacitive coupling at these voltages can create a Corona discharge even if there is no direct path. This Corona can build up and create conditions for a large direct discharge if not properly handled.

Proper grounding is the single most important thing in the entire system. If the system is not properly grounded, the extremely high voltage can find its way to ground through you. The shell of the fusor will always be connected directly to an Earth ground. The ground connection is brought to the chassis of the fusor and onto a bus bar called the star ground connection. This is where everything else is grounded to. I drilled ground connections onto all of the equipment and frames that sit on top of the workbench, as well as the bench itself. I used a multimeter to check the resistance between the main ground and every piece of metal on the fusor and associated equipment and nothing was above 0.7 Ohms. This means if anything catastrophic happened, all HV would find its way to ground and I would be protected. Because I am using analog meters for the voltage and current metering, there is a risk of the meters failing open and becoming energized. To ensure that the system would fail safe, a Zeener Diode pair was used between the two meter connections. Due to the way silicon diodes work, under normal operation, this pair would not conduct and all metering would work as intented, but if the voltage between the pins reaches ~3V, the diodes would conduct with very low resistance. I also installed a Gas Discharge Tube on the HV meter, so if the voltage rose above 100V, it would conduct to ground. This way if the meters failed, there would be some way for the electricity to reach ground. Past this, when I am operating the fusor I stand on a rubber work mat approximately 5 cm thick and I have plastic extensions for every control I need to interact with. These are extended at least 10 cm from the metal surfaces. If all else fails, I have an emergency “chicken stick” that is a long insulating pole with a wire attached to the far end of it, which connects directly to ground. If the system is charged, or if I want to ensure all HV capacitors have been discharged, I can use this stick to directly connect the transformer’s HV output to ground.

All wires used on the HV side are rated to 40kV and are run through vinyl tubing as an extra safeguard. Any screws used to connect these wires have a rounded nut on the end to prevent Corona from forming on the sharp edges. Two fans are used to decrease the buildup of Corona discharge and Ozone, which could cause a large discharge. In order to get the HV into the inner grid of the vacuum chamber, a commercial 40kV Conflat feedthrough is used. This is a delicate piece and must be kept dust free for it to hold up to its voltage rating.

There are several ways to stop HV in case of an emergency and interlocks so that the system cannot be energized unless the fusor is in the correct configuration. Because of these reasons, I am confident, but cautions, when operating the fusor.

The Power Supply

The power supply consists of a Zero-Volt Switching (ZVS) driver connected to a set of hand-wound flyback transformers wired in series. This is then put through a voltage multiplier to get the 35kV and 10mA needed for the fusor. The whole system is powered using a Variac, which allows it to be fully adjustable.

I purchased a cheap hand winding machine that keeps count of the turns as well as a set of transformer cores. Due to the high voltages involved, one of the most common failures is the arcing of the secondary coil to the core of the transformer. In order to prevent this, I designed space for a 2mm sheet of neoprene rubber to sit between the coil and the core. There can also be arcing between the layers of the secondary coil because of the high turn ratios, so flyback transformers are normally wound on special bobbins that have different sections. I placed this in a 3D printed mold and submerged it in epoxy for added arcing protection. The small secondary wires have air gaps between them, so when I poured the epoxy I placed them into an epoxy vacuum chamber to pull all the air out and replace it with epoxy. These coils have a turn ratio of 220:1, where the output voltage is around 4,440V. With the four transformers in series, I end up with a voltage of 17.6kV. This is then the input to a single stage voltage doubler which gives the full 35kV. This whole device is prone to arcing due to Corona buildup in air, so it is completely submerged in oil. This oil also helps with heat dissipation of the transformer cores. Care was taken in soldering the capacitors and diodes so that there were no points or sharp edges, which can cause fast oil flow.