July 26, 2003 - Soldered resistor chain into final form

I spent the day soldering the resistor chain together into its final form. The total length (including the 10Mohn resistor) is 1 meter and 8 cm with the gaps between resistors being anywhere form 5mm to 7mm.

Notice in the pictures below that I used two vises to hold the resistors while I soldered them. Originally I had the spacing between resistors at about 1cm so that I could attach the tips of needle nose pliers to either end to act as heat sinks, the purpose being to protect the resistors. However, in my first attempt the solder attached to the pliers so I gave up on that approach. It turned out that the vises themselves acted as perfectly good heat sinks allowing me to use a smaller spacing. I soldered the resistors into three chains of 7 and one chain of 4. The final assembly was done by taping the chains to the table surface and soldering the joints without heat sink. I guess the use of the vises was a bit of overkill. Despite the heat from all the soldering not one one resistor was ruined.

Resistors in the vises.

Close-up of the joint before soldering. The tips are touching.

Final soldered chain of resistors.

August 4, 2003 - Corona gap progress

I made a few attempts at building a corona ring but couldn't figure out any way to build it such that I could position it at location 5 along the chain and then reposition it at location 1 and still have the same distance between the ring and the resistor contact. Some pictures of what I tried are below.

What eventually worked out was to just have a single solder bulb instead of the ring. Also, rather than using the point at which I soldered the resistors together as the other side of the corona gap, I ended up using the much more consistent outer corner of the endcap on the resistor itself. Corona would therefore form between the solder bulb and the endcap. I confirmed this in the dark but unfortunately my camera is incapable of taking pictures at that light level.

Pictures of my attempts with a ring

Testing with a large ring to figure out the required gap. The ring was simply moved closer and farther from the resistor contact.

After measuring the required gap size (not very well) then I made a ring and supporting structure.

The idea was that I would mount a piece of cardboard cut to the gap size (inside the ring) and would then slide the ring over this and then hot-glue the ring's supporting structure into position. This was too difficult to do with any accuracy.

A solder bulb instead of a ring

There were a couple of advantages to using a solder bulb. The first is that I was able to make an adjustment screw that I could use to move the bulb closer out or farther in. The second is that the resulting structure could be pulled off the resistors at location 5 and put on the resistors at location 1 and be sure that the gap is the same at both locations.

As shown in the pictures below, the solder bulb assembly is made up of a nylon bolt with a smaller metal bolt screwed through the middle of it. The head of the smaller bolt is where the solder bulb is (solder was melted onto it). The other end of the smaller bolt is where the wire is attached. The clear glass-like piece is a 1"x1" piece of lucite. A 1/2" hole was drilled through the middle of the lucite piece and a nylon nut was epoxied into the hole. It is this nut that the nylon bolt is screwed into.

The solder bulb assembly and the parts that it is made up of.	Side view. The metal bolt is screwed through the middle of the nylon bolt (it is insde the nylon bolt).
The test setup.	The maximum voltage before corona prevents voltage from increasing higher is 49.6kV. Note that Cursor 1 is at 49.6V (the horizontal line at the top of the trace). The trace was set to linger for 5 seconds before vanishing - that's why you don't see a single trace line. The purpose for the linger was so that I could see if the voltage goes above 49.6V. The ripple is the 12kHz that was discovered previously.
The 49.6kV gap corresponding to the scope picture. Once this gap size was found, hot-glue was poured into the area between the nylon bolt head and the nylon nut to freeze it in that position. The idea is that I can now take the assembly off of the resistors at ths location and move it to the resistors at location 1.	If you look closely at where the resistors are going through the lucite supports you'll see that the lucite supports are actually made up of two pieces each. This is where I break it so that I can move it. Note that this means when I reassemble it at location 1 the gap size will remain the same.

August 11, 2003 - Corona gaps finished

As per Zoltan's advice I set the size of the second corona gap (the one further from the meter) such that it would leak before the first one.

Testing gap size for second corona gap.

Testing that the second gap arcs before the first gap.

Both gaps in their final positions.

August 11, 2003 - Stand-offs attached

In order to center the resistors in the final PVC tube I had to put lucite stand-offs around them.

Close-up of the stand-offs.

Because of the corona gaps the standoffs at that end had to be larger.

The complete chain with stand-offs.

August 11, 2003 - Probe tip made and attached

The probe tip was made to be fairly leakless and interchangable. The supporting part is a .5" diameter copper rod with a tapped hole in it. The actual tip is a bolt that I filed and then sandpapered down to a rounded tip. The threads on one end were left alone so that it could be screwed into the copper rod.

Probe tip made and attached.

August 23, 2003 - Waxing the connections

To prevent ion leakage at the connections between the resistors I kept the connections as short as possible, used large rounded bulbs of solder and applied paraffin wax. I first tried just dripping wax on but that created layers and flaked off fairly easily under stress. So then I made a small rectangular mold which I alternately put in place around each connection and poured wax into. This made a nice rigid, uniform block of wax that held on tightly. The pictures below show the results after a week of pouring (evenings).

Close-up of some of the blocks.

The resulting waxed chain.

rimstar.org