The transmitter design presented here has an unusual design technique. The technique is to construct it using only 2 materials, glass and aluminum. The construction method arose from the study of basic RF principles and a capacitor testing method. The limitation of materials is to allow better understanding of electrical energy conversions without the additional confusion of an amplifier or other complicating factors.
The circuit design is a multivibrator. It has input energy conditioning circuitry added to allow a wider range of input voltages to be useable. Because there are no active components (transistors, tubes, etc) a nonlinear component is needed to create multiple unstable states which is oscillation. The spark gap is chosen due to availability and durability. It is the 'switch' referred to in the RF origins section in Trail of Discovery. It is convenient to think of a switch as being either on or off. Unfortunately things must get more complicated to see what the switch is doing to the electricity.
The equipment shown above is mostly support equipment for the transmitter. The transmitter itself is fairly hard to photograph in ways to see it clearly. This is so because its design is so unusual. The side A and side B views are (roughly) mirror images of each other, having left and right sides switched. Each set is 90 degrees of viewing positions available. All other views are identical because it is symmetrical.
The 'Switch'---When the switch opens it is interrupting a current flow. That flow has an energy bearing field around it. The field will not simply stop at that point. The entire field involved has to collapse equally as the switch opens. When the field collapses it returns its stored energy as EMF which then causes current flow. The problem is that the switch is opening and those electrons may not continue their intended path. The resulting arc allows that final settling of the circuit to occur.
As the arc occurs current is flowing but the charge energy of the power source is not being used, the EMF from the magnetic field is. There is a full voltage drop between the points immediately after opening and the electrons continue to flow. As the magnetic field collapses it turns into static charge energy which is separate from the original EMF source. It is free RF energy. The energy is 'created' between the on and off conditions of the switch. It is a conversion of excess energy to another energy form.
When a switch or spark gap closes the same things happen in reverse. The EMF which waits on the edges of an open switch remains in effect until the final settling of the circuit is over. The entire magnetic field of the circuit that includes the switch begins to rise. The magnetic field displaces the static charge energy that was in the circuit. That happens at the edge of the gap where the EMF image is seen across the gap by the electrons causing static charge energy to split from the current flow of charge carrying electrons. The free static charge energy becomes free RF energy. The magnetic field draws electron flow to create itself of EMF energy from the power source. That electron flow is what displaces the previous static charge energy into becoming RF energy. When the field is fully setup and stabilized then the switch is finally on. RF radiation is over at that point.
A spark gap functions as a switch for obvious reason. The air breaks down at a specific voltage and allows current flow. A vacuum tube is easy to understand because it is a similar idea with control elements added. A transistor functions in a similar way also. Its 'switch' is the base to collector junction. That junction area has a varying size (from on to off) and is similar to the vacuum tube in the way that it works. The conductance of the output circuit is controlled by an amount of energy consumed at the input control element. The major difference is that electron activity occurs in a silicon crystal structure instead of flying thru vacuum tube 'space'.
An addition item is learned from the Tesla Coil Builders Club: The spark gap will work up to 6 times better if an air flow is added. The surprising thing is how little pressure is needed to achieve optimal results from this. At first thought it may seem that an arc across a high pressure nozzle would have to be set up. There are two similar examples of gas & arc combinations to contemplate. There is a audio speaker design that has a gas flame modulated by a high voltage arc. It is a high power, high frequency design. The second example is an RF torch. The RF torch is a tool used in industries for several purposes including the spraying of metallic coatings. There are many commercial models available.
To implement this addition a vacuum cleaner is used. Instead of using pressure delivery lines to each spark gap, the entire circuit is installed inside of a 5 gallon plastic bucket. The air is allowed to leak in through flexible plastic soda straws. The air then flows through the glass tubing and across the spark gap. The air is then discharged inside the bucket and removed by the vacuum cleaner.
The back side of the bucket has aluminum foil glued onto it with contact adhesive. It is sort of a reflector. It main purpose is to be there to cause capacitive loading on the antenna elements. As a reflector dish it will probably have little value. A better one could be made with a carefully curved piece of sheet metal. Because a true dish shape is difficult and expensive to create it is not done. This transmitter cannot be used for regular beam transmitter purposes and so the reflector is not to important. The reason the transmitter is restricted to academic studies only is because the output from if (if any) will be unstable in several ways and full of harmonics on many spectrums. A lot of angry people will come looking for you if you use it without proper precautions.
The Circuit---The usual method of building a spark gap oscillator is called a relaxation oscillator. As the name implies there is a capacitor which charges up to a point at which it then discharges to an output and is then relaxed. There is a theoretical maximum of energy that the circuit can produce. The output energy can only be 25 % of the input energy. The other 75 % turns to heat. Therefore, the multivibrator design is attempted. Because it has two sides that work against each other it may get a higher efficiency of 50 % as circuits using active components do. The maximum is 50 % because the output is, at best, a shadow or a copy of the input energy. That is for a push-pull amplifier. For a single transistor it is only 25 % because a single transistor can only push or pull but not both. For a spark gap it is only 25 % because the gap can only control the turn on voltage. The remaining circuit decides when the gap will turn off again.
If the source voltage is too great all of the spark gaps will go on and stay on until the voltage drops to a lower level. From this idea the input conditioning circuit was designed. It is basically a buffering capacitor. It will allow a higher voltage range to be used because when the multivibrator drains it, its voltage drops. The drop is brief but should allow the multivibrator to function without locking on as it otherwise would. To that extent it functions as a conventional large inductor or dropping resistor would. It allows a space or interruption of input power to occur caused by circuit loading.
The capacitors and inductors of this circuit are not as simple as they appear to be. The normal equations do not apply to either of them as will be explained. The capacitor testing method that inspired this design was very enlightening. We will examine it first and then study the inductors used later.
The Capacitor Test---The original purpose of the capacitor testing was to merely get some dielectric constants by direct measurement so that simple high voltage capacitors could be made of window glass. The experiment yielded far more than the desired results. If you're a beginner and you want to get a 'hands on' feel of what doing RF work is like you should duplicate this experiment. There are new discoveries to be explored if you do this.
The basic idea is to use a regular FM radio as a capacitance meter. The dial gets recalibrated in picofarads using known values of capacitors. The next step is to build several capacitors using glass or soda bottle plastic as the dielectric media. After measuring them the next step is to build a variable capacitor and test it also.
The data that comes from the experiment will not fit the equations as was intended. The exact shape and number of plates has a great influence on the final value. Because of this the equations may only be applied to specific categories of capacitor (ie. single plate, curved plate, variable) using values from similar types. There is not a single dielectric constant that may be used for a specific type of media material for all capacitor designs.
In addition to that, the variable capacitor shows another feature that we conveniently overlook. The unused plate area still counts as a capacitor even though we wish to ignore it. What this discovery shows you is that every thing the capacitor is connected to, or is even near to, will add heavily to the value of the capacitor. The variable capacitor helps to get an intuitive sense of just how much extra capacitance you get from nearby parts of the circuit. The final value of a capacitor will be what the carefully used equation tells you plus an amount that the builder must guess at.
The Construction Technique---After building and testing these things the idea to build an entire circuit using this construction technique came along. The idea became more challenging when it was realized that no connections could be made between pieces of aluminum. Each piece has to be made from a single aluminum can. That piece may include coils, capacitor plates and/or spark gap point holders. The maximum length allowed is cut spirally from the soda can. This will be the longest 1/4 wave antenna that can be produced in this way. Due to its short length the frequency required to use it will be above the FM band and getting close to microwaves. The UHF channels of a TV set may be used as test equipment at that frequency.
The technique appears simple and easy to do. It isn't. It will test your dexterity and patience to the limits. Here are several tips for those of you who want to try it. Do as much sanding in advance of cutting as possible. After the templates are stuck to the shiny side of the metal, scribe an outline of each part using the glass cutter. Turn it over and sand those areas before using scissors to cut the patterns. Then remove the patterns from the metal and carefully sand the shiny side as needed. Be prepared to replace any pieces that get crunched up or ruined during the sanding. Keep the metal as flat and smooth as possible at all times. Sanding the cut edges helps.
The coils are not wound on a spool. They are shaped into a spiral on a 3/16" metal rod. You will have a tendency to twist the coil around the rod. Don't do it. Doing so causes the windings to cut into each other. Shape the windings. Twisting may be used by guiding the windings to overlap each other later on. Don't try to twist the entire coil at once. When the coil is finished it will have a diameter of 1/4". If you cannot make a straight one put a piece of 1/4" plastic tube inside the finished coil to hold its shape. After 1 or 2 failures you will find it is much easier to do. Two of the coils are wound in one direction the other two coils are wound in opposite direction.
When producing the spark gap bases the most important thing to do is to have the points all be the same distance from the glass surface. They must align vertically. If they are off horizontally the error may be corrected when the metal is being stuck to the glass with double sided scotch tape. If point sets end up too far apart use a longer screw for the points.
After you've built a couple of these it doesn't seem so hard at all. If you think you're really good at it try to build a miniature one using aluminum foil, microscope cover slips and super glue. For the rest of us, build the first one using paper instead of aluminum. Complete the second one even though it has major flaws. Do it just for practice.
You will find that the aluminum breaks very easily at sharp bends. Make a pair of bending pliers. Get a cheap pair of needlenose pliers and lightly grind the edges off the sides of the jaws. When used to make the bends it will produce a tiny radius instead of a crease that breaks the grain of the metal. Use this even on the folded over flaps before crushing them flat. The flaps will break off when crushed flat if you don't.
The double sided tape makes construction easy. It also adds an unknown media material to the capacitors. An increase of capacitor value is expected because the adhesive will function as electrolytic media. If large amounts of energy are used the tape will probably blister and bubble and possibly fail. Super glue may be a better choice if you want to go to the trouble of using it for this. When the circuit is fully assembled put Goop adhesive around the edges of all parts. The tape will hold together long enough for the Goop to dry. The Goop will get soft at about the boiling point of water. It may fail due to heat buildup during operation because of this. Place it around the edges of the input capacitor to prevent arcing around the edges of the glass. Also place it at other places where arcing might occur. Do not place it on the sides of the glass spring bar near the antenna socket.
The Coils---From the capacitor tests it is known that it is very easy to produce capacitors of much greater size than needed. All capacitor values needed are in the picofarad range. The addition of stray capacitances make accurate prediction very difficult. That prediction becomes even harder when you consider the coils and how close they are to the capacitors. The coils have to be of a very small value. Normally a fine wire and a high quality core are used for their construction. Such coils have a very small diameter and only a few turns. These cannot be duplicated using aluminum cans for building material. A study of the coil winding equation shows that a long skinny coil allows the lowest possible value to be obtained.
What the equations do not allow for is to use sheet metal ribbon instead of round wire for the windings. The solution is to view the ribbon winding as several coils in parallel. That means that the coil will have a much lower value than the equation says it will have. If it does it becomes low enough to use for the design. Fortunately it allows coil value prediction to remain a mathematical exercise without the guesswork addition needed for capacitor prediction. To do so compare equal cross sectional areas of wire (from the coil equation) to cross sectional areas of square cut strips of ribbon (from the coil). Add all strip inductance values in parallel to get final value of spiral ribbon coil.
The circuit has not been tested at this time. It may need to have a 2 to 3 turn coil added in parallel to its power input to act as a 'base loading coil'. Because this coil is very close to being a short circuit it causes a large amperage to come from the power source. The leading edge of that amperage gets used by the circuit. The remainder of that amperage turns to heat. The circuit should not need this but it is one of the many experimenters options available. If large inductors are substituted for the capacitive coax feed lines the circuit should work as a relaxation oscillator instead. Motor field cores may be useful as feed inductors.
The Power Supply--The power supply is a motor/generator (alternator) unit with a step up transformer. The support equipment to operate it was built as modules. The rotor excitation source and the output control relay & breakers are multi purpose modules of test equipment. The rotor excitation power comes from a car battery. That power is controlled by the regulator module. The output from the alternator is controlled by a power relay and is also limited by circuit breakers. The 'BoomBox' is useful for testing salvage electrical & electronic equipment. It allows the service person to be several steps away from a test object when it gets powered up for the first time. It is also useful for the alternator testing experiments. The alternator has to be tested at many voltages and under many loads before it is ready to use. Some of the loads are very heavy or just plain accidental.
As fancy as it appears it is little more than a high voltage ac source based on a pair of automotive ignition coils. It has the ability
to deliver constant ac power at a low power level or brief pulses of high power ac. It also has the ability to fry the experimenter like
a bug on a bug light. You can't be too safe. Many other experimenters have died due to their activities. Even professional
engineers make simple fatal mistakes. 
This safety warning is put here because this is the
part that will kill you. If you experiment with microwave oven transformers or street light transformers you're getting in to
extremely dangerous areas. 
One advantage of modular construction is that the components can be used for other experiments. The alternator is useful in alternate energy experiments. An alternator may be used as the generator for a windmill or waterwheel. The voltage from it may be brought from the generation site using 3 skinny wires. The motor/alternator then converts the 3 phase high voltage to single phase line voltage. The inductive motor does the phase alignment automatically for the line voltage output.
