I used four strings times 60 RGP10M diodes, and four HV capacitors made by Mate's recipe. Construction was pretty straight forward. First I made the capacitors with tin foil and overhead transparencies. I made sure to give at least 4cm of clearance from the sides so it wouldn't arc over. Basically the capacitors are made by paralleling 5-6 overhead transparency sheets, then putting one sheet of tin foil on top, another 5-6 overhead sheets, one more tin foil, then finally rolling together and taping. The voltage standoff is surprisingly good given the quick and cheap construction.

Then I moved on to the diode stack. 240 diodes, and they all had to be chained together. It took a few hours from I started until they were all linked together in strings. I chose RGP10M diodes because my electronics source had them in large quantity for a good price. Besides convenience, they are avalanche rated which means if the voltage exceeds their rating they can safely breakdown and dissipate a certain amount of energy. This is necessary when putting diodes in series as it prevents a diode from exploding if it is slightly mismatched and too much voltage is dropped across it. The use of 60 diodes per string also give a decent safely margin. RGP10Ms also have about 500ns of recovery time, which is acceptable when using an input source up to 100kHz.


Here the diodes are spiraled around a thin PVC pipe to give them some structure and voltage stand-off in a smaller package. Beside them is the pipe they will soon be fitted into. The diode tower was centered and suspended with twine within the main PVC pipe. Wires from the diode tower were pulled through small holes in the large PVC pipe. Some duct-tape really eases construction. Be sure to see the video as the pictures don't do justice, capturing a still photo of an arc at it's greatest is nearly impossible. Remember when drawing arcs from a high voltage CW like this, that care must be taken not to simply short the output leads together. The capacitors are charged to about 40kV, and will dump all of their energy into the spark and diodes at once, resulting in large peak currents flowing through the diodes. For one thing it's damaging to the diodes and will lead to failure. If the driver is too weak it can also prevent an arc from forming, as the capacitors will not have sufficient energy left to sustain an arc, resulting in sparks instead.

For driving the CW Tower I use my home-made HV transformers, driven by a multi-purpose inverter.