A high voltage full wave rectifier

Earlier I posted about building a full wave rectifier from an LED diode bridge to use with our foxhole radio.  The transformer we've been playing with is alternating current, which means the electrons in our partial vacuum are first accelerated in one direction, then reversed and accelerated in the other direction many times a second.  To live up to their full potential what we really need is direct current.  However, normal diodes won't work, the voltage will easily spark across them.  So I got a hold of four microwave diodes (W10163432) designed for handling high voltage.  Also, I didn't want circuitry with this much power laying around bare on a table top, so we planned to embed them in paraffin wax (pure paraffin wax is a better insulator than almost anything except some types of plastic) in an old brie container (M likes to snack on brie).  Perhaps this is overkill, but it also has the benefit of making the bridge rectifier a single portable unit. 

Above, gathering materials and diagramming our strategy. 

First we needed to set up a double boiler to melt the paraffin.  I was planning to use two of V's pots but she forbade me from getting melted wax all over her new soup pot, so I picked up a stack of recycled aluminum trays for a couple dollars and used them as our crucible to melt the wax. 

We punched holes and ran some copper wire through for handles to lift the crucibles out of the boiling water.  If you try this, do not use a single wire across the middle for a handle (like the one on the left).  We tested this with water first and it tilts to one side and dumps whatever is in it.  A much better arrangement is to have three connections like the one in the middle, this gives much more control and allows the contents to be tilted and poured out. 

Above, a block of paraffin is completely melted. 

M pouring wax into the base of the brie container. 

Then we soldered the diodes together with wires to access each corner of the bridge.  It was a horrible soldering job, the thick (AWG 12) wire and large diodes sunk all the heat away from our cheap soldering iron repeatedly giving cold joints and having to retry; it was ugly in the end with solder blobs but the circuit worked.  (Pay attention to the cathode stripes on the diodes.  They have to run from one corner to the other.) 

We mashed the circuit down part way into the first layer of wax while it was still warm and started melting the second block of paraffin to pour over it. 

Above, the now transparent upper layer of wax is setting over the diode circuit.  Below is with the lid added and holes punched for the wires.  The circuit takes AC at two of its corners and converts it to DC at the other two corners.  Later I marked the + and - ends so I wouldn't forget. 

It was a hot day, while we waited for the wax to set the kids went outside to play and cool off for a while. 

Then we hooked it up and tried it out.  The most obvious difference was the DC plasma beam was much more responsive to a magnetic field (on the end of the screwdriver). 

The path near the magnet is bent toward a curve perpendicular to the magnetic field from the magnet. 

I was curious if we could detect any bremsstrahlung (braking radiation) from the setup.  This is seen with high voltage particles that undergo changes in acceleration (like running into the wire or glass on the sides).  The change causes the electron (or ion) to shed energy in the form of a photon, like the light emitted when electrons are captured by positive ions in the tube, but the photon can be even higher in energy than visible light, in the ionizing X-ray range. Long story short, we got some very spurious results. 

I placed our homemade geiger counter on a remaining piece of the shoepad (that we used to insulate the quad gyros from vibrations) because the vacuum pump was vibrating the table top.  In this position you can see a reading of over 20,000 counts per minute, or 165 µSv/hr!  30 minutes of this is about the dose received by people living within 16 km of Three Mile Island during its nuclear accident, but it would take about 3 hours to equal the exposure of a mammogram.  Fortunately however, this is a completely false reading.  The electric field caused the counter to go haywire, some times it would read zero for long periods of time, once symbols went all over the screen and I had to re-upload the program, and at other time it gave transiently high spike readings like this. 

Above is a plot of the recorded counts per minute on a log scale.  Counts around 10 are normal background levels around here.  Counts above 100 and long stretches of zero are not normal and presumably complete artifacts.