FIELD OF INVENTION This invention relates to voltage and current transformers. More specifically, this invention relates to a guided energy transformer which is a combination of a conventional transformer and a transmission line transformer.

BACKGROUND OF INVENTION Transformers are well known electrical devices which step voltages (or currents) up or down and are often used to electrically isolate input power sources from output loads. Transformers are used in a wide variety of electrical applications. Examples include switch mode power supplies, DC-DC converters, and telecommunication transformers.

Presently, conventional transformers are usually constructed of magnetic cores wound with primary and secondary wires. Some conventional transformers are

presently being designed to operate at high frequencies (e.g. 1 MHZ) resulting in a reduction of the size, weight, and cost of the transformer. The higher frequency, however, requires the use of more expensive, higher frequency components in circuits used with the transformers. Also, the skin effect limits the current that the conductors can carry at 1 MHZ.

Three of the problems with conventional transformers are their large size, weight, and cost. For example, a 1 kW, conventional transformer can occupy 245 in3, weigh 35 lbs, and have high cost. Another problem with conventional transformer is that it is too large to fit on a typical silicon chip and it is exceedingly difficult, if not impossible, to fabricate primary and secondary windings around a magnetic core on a silicon chip using present fabrication technology.

Another related technology is the transmission line transformer (TLT) . This type of transformer is constructed from transmission lines. Present TLT designs do not electrically isolate the input from the output, and/or do not allow a wide range of load impedances to be used, and/or require a magnetic core to inhibit unwanted modes of operation. These limitations have made the TLT useful in only very specialized applications such as an impedance transformer (e.g. impedance matching an amplifier to an antenna at a radio transmitter). For a description

of the theory and applications of TLTs see, for example, "Transmission Line Transformers," Jerry Swick, American Relay League, 1987.

A brief description of present transmission line transformers will first be given with reference to FIGS. 1 and 2 of the accompanying drawings. FIG. 1 shows a transmission line 50 having two conductors 52 and 54 separated by a material 56 having an electric permittivity E and magnetic permeability ij. There are many shapes conductors 52 and 54 can have; e.g. they can be coaxial cylinders or parallel wires, parallel planar stripes or even a combination of these shapes. The material 56 between the conductors 52 and 54 provides a medium for inductive and capacitive coupling between the two conductors 52 and 54. This coupling supports the propagation of transverse electromagnetic waves between the two conductors 52 and 54. When the instantaneous current in one conductor is equal and opposite to the instantaneous current in the other conductor at all points along the length, then the pair of conductors 52 and 54 are said to be operating in a transverse mode which may also be termed a transmission line mode. Thus, the transmission line 50 may be used to supply power from an AC voltage source 58 to a load impedance 60. The voltage source 58 is coupled to conductors 52 and 54 via a pair of input terminals 62 and

64. The load 60 is coupled to conductors 52 and 54 via a pair of output terminals 66 and 68.

The voltage source 58 applies an AC voltage to input terminals 62 and 64 creating an electromagnetic wave which propagates along the transmission line 50 with velocity vp and reaches the load impedance 60 at a time This This traveling harmonic wave is mathematically represented by A e-Y" , where A is the amplitude of the harmonic wave, e = a + j (a is the attenuation factor, is the phase factor, and j is the square root of -1), and z is the distance along the transmission line 50 measured from the input terminals 62 and 64. When the wave reaches the output terminals 66 and 68, part of the wave is reflected back towards the input terminals 62 and 64 and part of the wave is transmitted to the load impedance 60.

The reflected wave is mathematically represented by Be8Z , where B is the amplitude of the reflected wave. The fraction of the wave which reflects back towards terminals 62 and 64 depends on the characteristic impedance of the transmission line Z0 and on the load impedance ZL. In the special case where Zo = ZL there is no reflected wave.

The attenuation factor a, phase factor P (which is 2s divided by the wavelength 1), propagation velocity vp, and the transmission line characteristic impedance Z0, are determined from four characteristics of the transmission line: 1) inductance per unit length L; 2) capacitance per

unit length C; 3) conductance per unit length G; and 4) resistance per unit length R. These four quantities depend on the materials used in the construction of the transmission line and the shape and size of the transmission line. There are published, theoretical values of L, C, G, and R for several transmission line shapes and sizes which can be used. If theoretical values are not available, then the values of L, C, G, and R can be measured directly from the transmission line 50. The relationship between a and and these quantities is given by a + j = [(R + jCL)L) (G + j#c]½ , where # is 2# times the frequency of the traveling harmonic wave. The propagation velocity is given by Vp = / and the characteristic impedance is given by Z0 = [(R + juL)/(G + jCL)C)]. The input impedance, Zin, of the transmission line shown in FIG.1 is given by the following: Note that for low loss transmission lines a is approximately 0 which gives tanh (α + jP)l = j tanPl. This equation then becomes: Two transmission lines 72 and 74 can be wired together in the manner represented in FIG. 2 to produce a 2:1

step down transmission line transformer 70. The top transmission line 72 has two conductors 76 and 78, separated by a first material 80 with permittivity and permeability given by E1 and y1l respectively. The second transmission line 74 is comprised of two conductors, 82 and 84, separated by a second material 86 with permittivity and permeability given by E2 and ji, respectively.

The configuration of conductors 76, 78, 82 and 84 represented in FIG. 2 supports three modes for transmission of power from a voltage source 88 which is coupled to input terminals 90 and 96. The first mode is a transmission line mode. Traveling harmonic waves are generated in the two transmission lines 72 and 74 at the input terminals 90, 92, 94 and 96 and the waves travel along the lengths of the transmission lines 72 and 74. In this way, power is transmitted from the voltage source 88, to a load impedance 106 which is coupled to output terminals 98, 100, 102 and 104.

The second mode is not in general a transmission line mode. In this mode, current travels around the loop from input terminal 90 to output terminal 98, then from terminal 98 to output terminal 102, then from terminal 102 to input terminal 94, then from input terminal 92 to output terminal 100, then from terminal 100 to output terminal 104, and finally from terminal 104 to input terminal 96. This path has a very low impedance and effectively short circuits the voltage source 88. In the third mode, current travels around the loop from input terminal 90 to output terminal 98, then

from terminal 98 through the load impedance 106 to output terminal 104, and finally from terminal 104 to input terminal 96. This path effectively applies the full voltage V, across the load impedarce, ZL.

The second and third modes are not desirable for the operation of the 2:1 transmission line transformer 70 because the second mode effectively short circuits the voltage source, drawing a large current from the source and bypassing the load while the third mode allows the application of the full source voltage across the load, without a 2:1 step down. A common way to suppress the second and third modes is to use one or two magnetic (e.g. ferrite) cores 108 and 110. The cores 108 and 110 increase the impedance of the second and third modes to such an extent that they no longer play a role in the circuit. The presence of the cores, however, does not affect the operation of the first mode; therefore, this TLT will operate in the first mode.

The operation of the first mode gives a broadband, 2:1 transmission line transformer. The characteristic impedances of the two transmission lines 72 and 74 must be equal to each other, the load impedance 106 must equal 1A the characteristic impedance of one of the two transmission lines 72 and 74, and the transmission lines 72 and 74 must be constructed such that P111 - p313 = + 0 2r, 4ir, 6s, ... (i.e. the waves must be in phase at the output of the transmission lines). Two equal-voltage, traveling harmonic waves are created in the upper and lower transmission lines of the transformer 70 shown in FIG. 2. The voltage of the waves is

half of the voltage of the source voltage V, so that the voltage that is applied to the load impedance is V/2; i.e. the circuit functions as a 2:1 step down transformer.

Other ratio transformers such as 3:1, 4:1, and more complex transmission line transformers have been designed.

See, for example, "Transmission Line Transformers", Jerry Swick, American Radio Relay League, 1987. Four limitations in this transmission line transformer design are: 1) The load impedance must be matched to the equivalent characteristic impedance of the paired transmission lines at their output; 2) The output of the transmission line transformer is not electrically isolated from the input; 3) One or more relatively large magnetic cores 108 and 110 are needed to suppress unwanted modes which makes the TLT relatively large and heavy; and 4) High turns ratio TLTs require a relatively large number of transmission lines (e.g. 10:1 requires a minimum of 10 transmission lines) and non-integer turns ratios also require a relatively large number of transmission lines (e.g. 4.5:1 is produced by stepping the voltage down by 2:1 then stepping the voltage up by 9:1 using a total of 11 transmission lines).

SUMMARY OF THE INVENTION The structure and operation of the simplest guided energy transformer, GET, the 1:1 isolation GET, will be described in detail. As with TLTs, there is more than one mode of operation of the 1:1 GET. To enhance the most

favorable mode and to suppress the most deleterious mode, the GET must be operated at particular ranges of frequencies.

Several methods will be described for obtaining voltage transformation. The first method involves wiring together identical 1:1 GETs so that their inputs are in series and their outputs are in parallel. The next method involves wiring together two 1:1 GETs, each with a different characteristic impedance.

There are many shapes and sizes that GETs can have.

Essentially, any shape that a transmission line can have can also be used for a GET. These shapes are described. Three types of single stage n:1 GETs are then described.

A specific example is given on the construction of a GET using stripline transmission lines. This example describes a novel construction technique in which the dielectric and magnetic materials are layered in a unique way to obtain extremely high permittivity and permeability; thus, the operating frequency of the GET is reduced significantly.

A second mode of operation is described using only the middle transmission line. A third mode of operations using an outside transmission line will also be described.

Examples will be described which minimize the second and third modes of operations. These examples meet various initial conditions for a n:l GET.

A five transmission line GET will be described which performs the same function of two GETs wired together.

Various configurations of five transmission line GETs are described.

For the most part, GETs are used in circuits. The GETs perform the important functions of stepping voltages up or down and/or electrically isolating one part of a circuit from another part. Some of the more important applications include AC-AC converters, AC-DC converters, DC-AC converters, and DC-DC converters. Often, the AC voltage used is the common 60 Hz wall plug voltage. Since this frequency is too low for practical GET designs, the circuit in which the GET operates must supply a much higher frequency (approximately 10 kHz to as high as a few 10s of MHZ) to the input of the GET Then, at the output of the GET, the high frequency is either used as is or is converted to something else; e.g. 60 Hz or DC.

This invention contemplates other features and advantages which will become more fully apparent from the following detailed description taken in conjunction with the accompanying drawings.

DESCRIPTION OF THE DRAWINGS FIG. 1 is a circuit diagram of a prior art transmission line connected to a voltage source and a load impedance.

FIG. 2 is a circuit diagram of a prior art transmission line transformer.

FIG. 3A is a circuit schematic of a 1:1 isolated guided energy transformer (GET) coupled to a voltage source and a load according to the present invention.

FIG. 3B is a circuit schematic of the 1:1 isolated GET with a the voltage source and load coupled to different transmission lines according to the present invention.

FIG. 4 is a circuit schematic of two 1:1 GETs wired together to form a 2:1 GET according to the present invention.

FIG. 5 is a circuit schematic of two 1:1 GETs wired together to form an n:1 GET according to the present invention.

FIGS. 6A-6E are cross section views, perpendicular to the length, of different conductor arrangements for a GET according to the present invention.

FIG. 7A-7B are cross section views, perpendicular to the length, of mixed conductor shapes for a GET according to the present invention.

FIG. 8A shows sheets of ferrite bonded to form a 2" high stack and a slice from a stack of bonded ferrite sheets.

FIG. 8B is an exploded view of the layers of materials in a GET according to the present invention.

FIG. 8C shows an enlarged view of the cross section, parallel to the length, of a GET according to the present invention.

FIG. 9A is a circuit diagram of the lumped element model of the GET structure shown in FIG. 8C.

FIG. 9B is a circuit diagram of the ideal lumped element model of a transmission line.

FIG. l0A is a circuit of a two-stage 2:1 GET according to the present invention FIG. lOB shows a schematic of a two-stage 2:1 GET in FIG. l0A with different connections between the two stages.

FIG. 10C shows the cross section of two GET sections surrounded with magnetic material.

FIG. 11 shows waveforms of the voltage of the voltage source and the voltage profile on the single-stage 1:1 GET in FIG.

3A.

FIG. 12 is a circuit schematic of a single-stage n:1 GET according to the present invention with a load attached to some point along the GET.

FIG. 13 shows the waveform of the voltage of the voltage source and the voltage profile on the single-stage n:1 GET in FIG. 12.

FIG. 14 is a circuit schematic of a dual-output, single-stage n:1 GET according to the present invention.

FIG. 15 is a circuit schematic of a single-stage n:l GET with shortened upper and lower transmission lines according to the present invention.

FIG. 16 is a circuit schematic of a single-stage n:1 GET with shortened upper transmission line.

FIG. 17 shows the çaveform of the voltage of the voltage source and the voltage profile on the single-stage n:1 GET in FIG. 16.

FIG. 18 shows the waveform of the voltage of the voltage source and the voltage profile on a single-stage 1:1 GET after using a second frequency.

FIG. 19 shows the waveform of the voltage of the voltage source and the voltage profile on a single-stage 1:1 GET.

FIG. 20A is a perspective view of a disk-shaped GET and a pie- wedge-shaped piece cut from the disk according to the present invention.

FIG. 20B is a cross section view of the disk-shaped GET in FIG. 20A.

FIG. 21 is a circuit schematic of a single-stage n:l GET with differing material thicknesses according to the present invention.

FIG. 22 is a cross section view of a parallel wire GET according to the present invention.

FIG. 23A is a circuit schematic of a GET according to the present invention configured to be a 2:1 step up isolation voltage transformer.

FIG. 23B is a circuit schematic of a GET according to the present invention configured to be a 2:1 step down isolation voltage transformer.

FIGS. 24A-24B are circuit schematics of GETs according to the present invention configured to be 1:1 isolation voltage transformers.

FIGS. 25A-25B are circuit schematics of GETs according to the present invention configured to be 2:1 isolation voltage transformers.

FIGS. 26A-26B are circuit schematics of GETs according to the present invention or an alternate 1:1 isolation voltage transformer coniguration.

FIGS. 27A-27B are circuit schematics of GETs according to the present invention for another alternate 1:1 isolation voltage transformer configuration.

FIGS. 28A-28B are circuit schematics of GETs according to the present invention for yet another alternate 1:1 isolation voltage transformer configuration.

FIGS. 29A-29C are circuit schematics of five transmission line GETs according to the present invention in different load voltage source configurations.

FIGS. 30A-30C are circuit schematics of five transmission line GETs according to the present invention in different load voltage source conigurations.

FIGS. 31A-31B aLe circuit schematics of five transmission line GETs according to the present invention in different load voltage source configurations.

FIGS. 32A-32D are circuit schematics of five transmission line GETs according to the present invention configured to be n:l isolation voltage transformers.

FIGS. 33A-33C are circuit schematics of five transmission line GETs according to the present invention configured to be n:1 step down isolation voltage transformers.

FIGS. 34A-34C are circuit schematics of five transmission line GETs according to the present invention configured to be n:1 step down isolation voltage transformers.

FIGS. 35A-35C are circuit schematics of five transmission line GETs according to the present invention configured to be n:1 isolation voltage transformers.

FIGS. 36A-36D are circuit schematics of five transmission line GETs according to the present invention configured to be n:1 isolation voltage transformers.

FIG. 37 is a block diagram of a converter circuit using a GET according to the present invention.

FIG. 38 is the waveforms of the input voltage, rectified voltage and modulated voltage of the circuit in FIG. 37.

FIG. 39 is a circuit schematic of the modulator component of the circuit of FIG. 37.

FIG. 40 is a block diagram showing the addition of a phase lock loop to the circuit of FIG. 37.

FIG. 41 is a block diagram showing the addition of a demodulator to the circuit of FIG. 37 FIG. 42 is a waveform of the output voltage from the demodulator shown in FIG. 41.

FIG. 43 is a circuit schematic of a polarity switching circuit.

FIG. 44 is a circuit schematic of a baseline restorer.

FIG. 45 is the waveform of the output voltage from the polarity switching circuit of FIG. 43.

FIG. 46 is the waveform of the output voltage from the baseline restorer circuit of FIG. 44.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention combines features of conventional voltage transformers (i.e. primary and secondary wires wound around a magnetic core) with transmission line transformers (TLTs). Like conventional transformers, there is, essentially, a constant voltage at the output regardless of the load impedance and the input is electrically isolated from the output. Like TLTs, this invention is constructed of

transmission lines and, hence, can be made very small and lightweight. Unlike both conventional transformers and TLTs, there is no magnetic (ferrite) core. To differentiate this invention from both conventional transformers and TLTs, it is termed a guided energy transformer (GET).

GETs differ from TLTs in several respects. First, GETs electrically isolate the output from the input. Second, GETs can maintain a constant voltage output or a constant current output even while the load impedance changes. Third, GETs are not broadband devices; they must be operated in particular ranges of frequencies. And fourth, GETs do not utilize magnetic cores.

The advantages of GETs over conventional (including high frequency) transformers are: 1) GETs are smaller, 2) GETs are lighter, 3) GETs are less expensive to manufacture. The advantages of GETs over transmission line transformers are: 1) GETs do not use magnetic cores, 2) GETs electrically isolate the input from the output, 3) GETs maintain a constant voltage (or constant current) output even while the load impedance changes. There are no known TLT designs which offer, simultaneously, all three of these advantages.

FIG. 3A is a circuit schematic of a 1:1 isolation GET 120. The GET 120 is comprised of four conductors 122, 124, 126, and 128. Conductors 122 and 124, as well as conductors 126 and 128, are separated by materials 130 and 132. In this embodiment, the materials 130 and 132 each have permittivity and permeability equal to E1 and , respectively.

Conductors 124 and 126 are separated by a material 134 which

in this embodiment has permittivity and permeability equal to E2 and 82, respectively.

The GET 120 may be viewed as three transmission lines (upper, middle, and lower) 123, 125 and 127 fabricated in direct contact with each other. Conductors 122 and 124 in combination with material 130 form an upper transmission line 123. Conductors 126 and 128 in combination with material 132 from a lower transmission line 127. Conductors 124 and 126 in combination with material 134 form a middle transmission line 125. The upper and lower transmission lines 123 and 127 of this particular GET are designed to have approximately the same characteristic impedance and the time for a wave to propagate the lengths of these transmission lines is approximately the same so that needless complication of this description is avoided. There are other GET designs in which this is not the case. A design in which the upper and lower transmission lines 123 and 127 do not have the same characteristic impedance will have a different fraction of the source voltage 136 applied to each at their respective inputs 142 and 144 as well as 146 and 148. A design which does not have the same propagation velocity on the upper and lower transmission lines 123 and 127 must be designed in a manner to ensure that the waves are in phase at the outputs 150, 152, 154, and 156 of the transmission lines 123 and 127.

The GET 120 has two transmission line modes which operate simultaneously. The first mode has traveling harmonic waves propagating between conductors 122 and 124 as well as identical traveling harmonic waves traveling between

conductors 126 and 128. Thus, only the upper and lower transmission lines 123 and 127 are used in this mode. In this first mode there are no first order electric or magnetic fields in the material 134 between conductors 124 and 126.

The first mode may be viewed as a simple pair of transmission lines in which the AC voltage from a voltage source 136 is divided by two, propagated along the upper and lower transmission lines 123 and 127 shown in FIG. 3, recombined at the output, and applied to a load impedance 138.

This pair of transmission lines is equivalent to a single transmission line with a characteristic impedance equal to the sum of the characteristic impedances of the upper and lower transmission lines 123 and 127; i.e. the characteristic impedances of the upper and lower transmission lines 123 and 127 are both Z1 which means the characteristic impedance of the mode 1 equivalent transmission line is 2Z1. Using equation 1 the input impedance for the first mode, Zinl may be determined.

This is the impendance that the voltage source "sees," after the initial traiisients have settled, due to mode 1: where p1 is 2r divided by the wavelength of the waves propagating in the upper and lower transmission lines 123 and 127 length of the transmission lines 123 and 127.

There are two special cases which are important.

The first case is when p11 = 0, ir, 2ir, 3ir, ... (i.e. the frequency of the voltage source is selected such that the

wavelength, and hence p1, approximately satisfies one of these conditions) . In this first case, the input impedance due to the first mode reduces to Zin1 = ZL. Thus, the voltage applied to ZL equals the voltage of the source 136. The actual value of the characteristic impedance of the upper and lower transmission lines is not a factor. For this case, the first mode of the GET 120 is behaving as a 1:1 isolation transformer; the voltage applied to the load 138 equals the voltage of the source 136 yet the output of the GET 120 is electrically isolated from the input. Note that Zm can take on any value and that the first mode of the GET 120 behaves as a voltage source.

The second special case is when p11 = it/2, 3#/2, 5#/2, ... (i.e. the frequency of the voltage source is selected such that the wavelength, and hence p1 approximately satisfies one of these conditions) . In this second case, the input impedance due to the first mode reduces to Zinl = 4Zl2/ZL.

Thus, as the value of ZL increases, the input impedance decreases and more current flows. For this case, the first mode of the GET 120 behaves as a current source, with I = V/ (2Z1), yet the output of the GET 120 is electrically isolated from the input.

The second mode uses only the middle transmission line comprising conductors 124 and 126 and material 134. No first order electric or magnetic fields are generated outside the material 134. This mode can be simply understood as being the voltage source 136 connected to the conductors 124 and 126 of the middle transmission line at input terminals 144 and 146

while the other end of the conductors 124 and 126 of the middle transmission line are short circuited between output terminals 152 and 154. The characteristic impedance of the middle transmission line is Z2 Using equation 1 above (with Zt = 0) to determine the input impedance for the second mode; the impedance that the voltage source "sees," after the initial transients, have settled due to the second mode is as follows: Zin2 = j Z2 tan 2l, where p2 is 2r divided by the wavelength of the waves propagating in the middle transmission line and 1 is the length of the transmission line. It is clear that the input impedance for the second mode can take on a range of values by changing the value of 2l (which is done by changing the frequency of the voltage source, and/or by changing the values of E2 and/or 82 of the material used between conductors 124 and 126, and/or by changing the length 1 of the transmission line). When the input impedance for the second mode is low, a large current fowls through the wire connecting terminals 152 and 154. This is a waste of energy.

The effects of the second mode may be minimized by making its input impedance as high as possible. To do this p21 is made to be equivalent to #/2, 3#/2, 5#/2, ... (i.e. the frequency of the voltage source is selected such that the wavelength, and hence p2, approximately satisfies one of these conditions). The first embodiment relates to a 1:1 isolation GET with all the materials separating the conductors of equal

length and where the materials in the upper and lower transmission lines 123 and 127 as being equivalent.

A more complex 1:1 isolation GET may be obtained.

In this second embodiment, the materials 132, 130 and 134 separating the four conductors 122, 124, 126 and 128 can each have different electrical and/or magnetic properties (i.e. p2, and p3 respectively) Also, as represented in FIG. 3B, the voltage source 136 may be coupled to the input terminals 142 and 148 and the load impedance 138 may be coupled to output terminals 152 and 154. The two circuits represented in FIGS. 3A and 3B are completely equivalent.

In the second embodiment of the 1:1 isolated GET 120, the deleterious second mode of operation is minimized by operating the middle transmission line 125 at a frequency that maximizes its input impedance. There are no magnetic (e.g. ferrite) cores needed as is the case for TLTs. The 1:1 isolated GET 120 can be used as a building block to produce more complex transformers; i.e. transformers which step the voltage up or down.

Multi-Stage n:1 GETs FIG. 4 shows a more complex 2:1 GET 160 which is built from two ':1 isolation GETs 162 and 164. Both of the 1:1 isolation GETs 162 and 164 are similar to the GET shown in FIG. 3A. GET 162 has an upper transmission line 166, a middle transmission line 168, and a lower transmission line 170. GET 164 has an upper transmission line 172, a middle transmission

line 174, and a lower transmission line 176. The quantities Zn and n, for n = 1 to 6, are the characteristic impedances and phase factors (i.e. 2 divided by the wavelength) for each of the respective transmission lines 166, 168, 170, 172, 174 and 176. The middle transmission lines 168 and 172 of each 1:1 GET 162 and 164 is to be operated such that its input impedance is very high. This means that the following must be approximately satisfied: 2l1 = #/2 3#/2, 5#/2, ... and 5l2 = #/2, 3#/2, 5#/2, .

The input impedance of the 2:1 GET 160 is to be directly proportional to the load impedance (i.e. this is to be a constant voltage source); thus, the following conditions must be approximately satisfied: p111 = 0, it, 2#, 3#, . . . and p311 = 0, it, 2It, 3It, ... and 4l2 = 0, #, 2#, 3#, ... and 6l2 = 0, #, 2#, 3#, ...

Also, the waves leaving the upper and lower transmission lines 166, 170, 172 and 176 of each 1:1 GET 162 and 164 must be in phase: 1l1 - p311 = + ), 2It, 4It, .. and p412 - 6l2 = + 0, 2#, 4#, And, the waves leaving the 1:1 GETs 162 and 164 must be in phase with each other: 1l1 - p412 = + 0, 2#, 4#, In addition, the voltage V, from a voltage source 178 must divide equally between the two 1:1 GETs 162 and 164; i.e the differential voltage applied to each of the 1:1 GETs 162 and 164 must be V/2. This will occur when Z1 + Z3 = Z4 + Z6.

When all of these conditions are approximately satisfied, the 2:1 GET 160 will apply an AC voltage to a load

impedance 180 which equals 1/2 the voltage of the voltage source 178. The frequency of the voltage source 178 cannot be a broadband source, it must have a value which allows the above conditions to be approximately satisfied. However, the load impedance 180 can take on any value. It is a simple matter to step the voltage up by a factor of two with the 2:1 GET 160 by exchanging the positions of the voltage source 178 and the load impedance 180 to produce a voltage on the load 180 which is twice the voltage source 178. n:l GETs may be obtained in a similar fashion to the 2:1 GET 160 shown in FIG. 4. A 3:1 GET may be fabricated by adding another 1:1 GET to the 2:1 GET 160. The additional 1:1 GET is wired in series with the voltage source 178 and in parallel to the load impedance 180. The individual transmission lines in the 1:1 GETs that are added, must have outputs that are in phase with each other and the added GETs must have outputs that are in phase with the other GETs.

Also, the differential voltage applied to each GET must be the same. These conditions are described in detail for the 2:1 GET (see above for those conditions) . By adding more 1:1 GETs in a similar way it is possible to get 4:1, 5:1, 6:1, etc GETs. The voltage steps, however, occur only in integer numbers.

Two-Staqe n:1 GETs Another n:1 isolation GET 190 is represented in FIG.

5. Two 1:1 isolation GETs 192 and 194 are wired in series.

The GETs 192 and 194 are similar to the GET shown in FIG. 3A.

GET 192 has an upper transmission line 196, a middle transmission line 198, and a lower transmission line 200. GET 194 has an upper transmission line 202, a middle transmission line 204, and a lower transmission line 206. As with previous GETs, certain conditions must be met. First, the middle transmission lines 198 and 204 of each 1:1 GET 192 and 194 must be operated in a manner that produces a very high impedance. Thus, the following must be approximately satisfied: 2l1 = #/2, 3#/2, 5#/2, ... and 5l1 = #/2, 3#/2, 5#/2, In addition, the following must also be approximately satisfied: 1l1 = #/2, 3#/2, 5#/2, ...; 3l1 = #/2, 3#/2, 5#/2, ...; 4l2 = #/2, 3#/2, 5#/2, ...; and 6l2 = #/2, 3#/2, 5#/2, ...

Also, the waves leaving the upper and lower transmission lines 196, 200, 202 and 206 of each 1:1 GET 192 and 194 must be in phase: 1l1 - 3l1 = # ), 2#, 4#, ... and 4l2 - 6l2 = # 0, 2#, 4#, Each 1:1 GET 192 and 194 behaves as though it were a single transmission line with a characteristic impedance equal to the sum of the characteristic impedances of the upper and lower transmission lines 196, 200, 202 and 206. For the first 1:1 GET 192 the characteristic impedance is Z1 + Z3 and for the second 1:1 GET 194 it is Z4 + Z6. For this case, equation 1 above simplifies to Zin = Z0²/ZL. The input impedance as seen by a voltage source 208 is determined by applying equation 1

twice (once for each 1:1 GET) to obtain: Zin = ZL(Z1 + Z3)2/(Z4 + Z6)2. The value of the output voltage, Volt, as seen by a load impedance 210 may be derived by assuming that the electrical power going into the GET 190 is equal to the electrical power coming out; Pin = Pout, and using P = V2/R to get V2/Zin = Vout2/ZL, Zin is substituted and simplified to obtain Vout = V(Z4 + Z6)/(Z1 + Z3). Thus, the n:l isolation GET 190 functions as a voltage transformer with the value of n set by choosing appropriate values for Z1 Z3, Z4, and Z6.

GET Geometries FIGS. 3 to 5 show circuit schematic representations of GETs. There are many materials and geometries that can be used for the actual design of n:1 isolation GETs. For this discussion, only the second embodiment of the 1:1 isolation GET 120 shown in FIG. 3A will be described. Other more complex GETs can be constructed by using the 1:1 GET as a building block. As with any transmission line, the conductors used in the construction of GETs can have many shapes. The most common shapes are wires, tubes, and planar stripes.

These shapes are then separated by, or surrounded by, materials which have the properties necessary to support the propagation of traveling harmonic waves along the lengths of the conductors.

FIG. 6A shows the cross section of one arrangement of the conductors, perpendicular to the length, of the 1:1 GET 120 in FIG. 3A The identical elements have identical reference numerals. Two conductors 122 and 124 are imbedded

in the material 130 with permeability and permittivity given by E1 and p1/ respectively. The other two conductors 126 and 128 are imbedded in the material 132 having permeability and permittivity given by E3 and 83, respectively. In this configuration, the conductors 122, 124, 126 and 128 are wires.

These two pairs of conductors are imbedded in the third material 134 with permeability and permittivity given by E2 and 82, respectively. Although FIG. 6A shows the conductors 122, 124, 126 and 12P with circular cross sections, the conductor cross sections may have any shapes (e.g. square, rectangle. elliptical, etc.). All the conductors 122, 124, 126 and 128 may be parallel to each other, conductors 122 and 124 may also form a twisted pair while 126 and 128 are parallel.

Alternatively, conductors 122 and 124 may be parallel while conductors 126 and 128 form a twisted pair. Conductors 122 and 124 may also form a twisted pair and conductors 126 and 128 may form another twisted pair then the two twisted pairs may then be twisted around each other. There are many other configurations of these four conductors which can be used to fabricate a 1:1 GET. This configuration is relatively easy to manufacture because parallel pairs of wires and/or twisted pairs may be purchased commercially and embedded within material 134 to form the GET.

FIG. 6B shows the cross section, perpendicular to the length, of another arrangement of conductors which will function as the 11 GET 120 shown in FIG. 3. In this case, the conductor 122 is a wire or a tube surrounded by the material 130. The remaining three conductors, 124, 126 and

128 are tubes which are coaxially placed around the center conductor 122 and materials 132 and 134. Although FIG. 6B shows the conductors 122, 124, 126 and 128 with circular cross sections, the conductor cross sections may have any shape (e.g. square, rectangle. elliptical, etc.). Conductors 122 and 124 are separated by the material 130 with permeability and permittivity given by E1 and y1 respectively. Conductors 124 and 126 are separated by the material 134 with permeability and permittivity given by E2 and 82, respectively.

Conductors 126 and 128 are separated by the material 132 with permeability and permittivity given by E3 and 83, respectively.

The cross section, perpendicular to the length of another arrangemen of coaxial transmission lines is shown in FIG. 6C. In this case, the center conductors 122 and 128 are wires, tubes, or both while the remaining two conductors 124 and 126, are tubes which are coaxially placed around the center conductors 122 and 128. Although FIG. 6C shows the conductors 122, 124, 126 and 128 with circular cross sections, the conductor cross sections may have any shape (e.g. square, rectangle, elliptical, etc.). Conductors 122 and 124 are separated by the material 130 with permeability and permittivity given by E1 and µ1, respectively. Conductors 126 and 128 are separated by the material 132 with permeability and permittivity given by E3 and 83, respectively. These two coaxial transmission lines are then imbedded in the material 134 with permeability and permittivity given by E2 and 82 respectively. Conductors 122 and 124 may be parallel to

conductors 126 and 128 or conductors 122 and 124 may be twisted around conductors 126 and 128 as a twisted pair.

FIG. 6D shows the cross section, perpendicular to the length, of another arrangement of the conductors in the GET 120 shown in FIG. 3A. In this case the conductors 122, 124, 126 and 128 are planar stripes. Two conductors 122 and 124 are separated by a material 130 with permeability and permittivity given by E1 and ,i, respectively. Conductors 124 and 126 are separated by a material, 134 with permeability and permittivity given by E2 and 82/ respectively. Conductors 126 and 128 are separated by a material, 132 with permeability and permittivity given by E3 and 83, respectively.

The arrangement of conductors and materials in FIG.

6D may be free-standing or may be fabricated on a substrate.

One type of substrate fabrication is on a semiconductor wafer (e.g. silicon, gallium arsenide, etc.) using the standard methods for integrated circuit fabrication (e.g. deposition of material by evaporation, sputtering, chemical vapor deposition, etc., through a series of masks to build up the structure) . Another type of substrate fabrication is on a multi-layer circuit board with the layers between the conductors being made up of the appropriate materials.

The structure shown in FIG. 6D has an additional consideration which is not an issue with the structures shown in FIGS. 6A-6C. With the structure shown in FIG. 6D the magnetic flux lines will extend outside the structure as is shown by the dashed lines. The magnetic flux lines form closed loops. For part of the loop, the magnetic flux passes

through material which has a permeability of µ1, µ2, or 83 and for the remainder of the loop the flux passes through the material surrounding the GET (e.g. air). Whenever µ1, 2/ and µ3 have values that are comparable to the value of the material surrounding the GET (e.g. air), then the reluctance around the closed loop of magnetic flux is relatively low and the GET will operate properly. However, if µ1, 2/ or 83 have values that are significantly more than the value for the material surrounding the GET, then the surrounding material will be an area of high reluctance and will limit the value of the magnetic flux around its entire closed loop. This defeats the purpose of using high values of y µ2, and µ3 to obtain high magnetic fluxes within the GET.

To overcome this problem, the GET structure in FIG.

6D may be surrounded by material which will provide a low reluctance path for the magnetic flux lines. FIG. 6E shows how the GET may be surrounded by material 212 with permeability µ4 to provide a low reluctance path for the magnetic flux. The actual value of 84 and the optimal shape of the material surrounding the GET 120 depends, in part, on the values of µ1, 2/ and 3 and on the thicknesses of the corresponding materials.

FIGS. 7A and 7B show the cross sections, perpendicular to the lengths, of two examples of the 1:1 GET 120 which use a mixture of shapes for the conductors 122, 124, 126 and 128. FIG. 7A shows a mixture of wires, or tubes, and planar stripes while FIG. 7B shows a mixture of a coaxial

structure with a planar stripe structure. It is clear that numerous other configurations can be devised.

FIGS. 6 and 7 all show cross sections, perpendicular to the length, of various conductor configurations of the 1:1 GET 120 shown in FIG. 3. All the conductors in these figures will have a length. The lengths may be in the form of straight lines, spirals, or any other shape that is physically possible.

FABRICATION OF GETs Several different examples of fabrication techniques may be used in order to produce GETs according to the present invention.

Example 1 One method of the construction of a GET using stripline transmission lines will now be explained. The two quantities which need to be determined are the characteristic impedance, Z0, of the GET and the frequencies, n, at which this GET may be operated so as to produce the condition Pl = it/2, 3it/2, 5it/2, ... One of these conditions must be approximately met so that this GET can be used in a two stage, n:l GET as shown in FIG. 5.

The stripline transmission line is a common configuration for a transmission line and equations have been derived which allow the calculation of: resistance per unit length, R; capacitance per unit length C; conductance per unit length, G; and inductance per unit length, L (the distributed coefficients of the transmission line). For background

information on transmission lines see, for example, "Transmission Lines," Robert A. Chipman, Schaum's Outline Series in Engineering, McGraw-Hill Book Co., 1968 which is hereby incorporated by reference. If an uncommon transmission line geometry is used then the values of R, C, G, and L may be directly measured from the transmission line. The construction materials for these GETs are ferrite, high dielectric ceramic, and copper. These materials are bonded together with specialized epoxies. The procedure is as follows: 1) 30 mil thick ferrite sheets are cut and bonded together to form a 2" high stack 220 as shown in FIG 8A. The sheets are bonded together with highly conducting silver epoxy (p = 8 x 10-6 Q-rvi). The other two dimensions of the stack 220 are chosen so that subsequent cutting operations to be performed on the stack 220 can be done easily.

2) The bonded 2" stack 220 is sliced into 60 mil thick sheets 222. The cuts are made perpendicular to the plane of the ferrite sheets as is shown in Fig. 8A. This process forms the laminated ferrite sheets 222.

3) 2 mil thick dielectric sheets 224 of a high dielectric constant (E > 20,000) ceramic are fabricated as shown in FIG. 8B.

4) The dielectric sheets 224, the laminated ferrite sheet 222, and 30 mil thick copper sheets 226 are bonded together as shown in Fig. 8B. The copper sheets 226 are bonded on top of the dielectric sheets 224 which are in turn bonded on top of the ferrite sheets 222. Two types of

epoxies are used to bond the sheets 222, 224 and 226. Epoxy # 1 is a highly conducting silver epoxy (p = 8 x 10-6 Q-m) and epoxy # 2 is a partially conducting graphite epoxy (p = 10-3 Q- m). Starting from the top and moving down, the order of materials is: a) copper; b) epoxy #1; c) ceramic; d) epoxy #2; e) ferrite; f) epoxy #2; g) ceramic; h) epoxy #1; i) copper; j) epoxy #1; k) ceramic; 1) epoxy #2; m) ferrite; n) epoxy #2; o) ceramic; p) epoxy #1; q) copper; r) epoxy #1; s) ceramic; t) epoxy #2; u) ferrite; v) epoxy #2; w) ceramic; x) epoxy #1; and y) copper. The cross section of this bonded stack is shown in Fig. 8C. There is a 0.1" overhang of the copper sheets 226.

5) The bonded stack is sliced into a 0.1" thick piece and a 0.2" thick piece. The laminated ferrite and ceramic sheets 222 and 224 are 2 inches long while the copper sheets 226 are 2.2 inches long.

The structure of the transmission lines in this GET operates in the same fashion as ordinary stripline transmission lines. The GET structure shown in FIG. 8C can be modeled as the electrical circuit shown in Fig. 9A. The components of this model are as follows. C1 is the capacitance due to the high dielectric ceramic 224. C2 and R2 are due to the partially conducting, graphite epoxy (epoxy #2). R1 is due to the resistance of the sheet of highly conducting epoxy (epoxy #1) bonding the ferrite sheets together. Epoxy # 2 is made up of a polymer mixed with graphite powder to make it slightly conducting; its permittivity is approximately 3 and its resistivity is around 10-3 Ohm-m. Epoxy # 2 bonds the

ceramic material to the magnetic material while not reducing the value of C1. When the resistivity of epoxy #2 is low enough, then R2 will allow current to bypass C2; effectively removing C2 fron the circuit. Each individual ferrite sheet is 30 mils thick by 100 mils wide so that the area A, is given by A = (0.03") (0.1") (0.0254 m/inch)2 = 1.94 x 10-6 m2. The thickness of the ceramic is 2 mils; thus, de = 5.08 x 10-5 m.

The thicknesses of epoxy #1 and epoxy #2, t1 and t2 respectively, is about 1 mil; thus, t1 = t2 = 2.54 x 10-5 m.

C = #A(l, so C1 = EA/de = (20,000) (8.85 x 10-12)(1.94 x 10-6)/(5.08 x 10-5) = 6.76 nF, and C2 = EA/t2 = 3(8.85 x 10-12)(1.94 x 10-6)/(2.54 x 10-5) = 2.03 pF.

R1 and R2 are calculated as: R1 = pl/A =(8 x 10-6) (0.060") (0.0254 m/inch)/(0.001") (0.1") (0.0254)2 = 0.189 Ohms and R2 = (10-3) (0.015") (0.0254)/(0.001") (0.1") (0.0254)² = 5.91 Ohms.

Since this GET operates at below 50 kHz, R2 is low enough to cause C2 to be perpetually discharged; therefore, C2 can be dropped from the model. The total resistance, R1+ 2R2, is not high enough to prevent C1 from completely charging and discharging; therefore, R1 and R2 can be dropped from the model. Now the model shown in Fig. 9A matches the ideal transmission line model shown in Fig. 9B, with CAz =C1/2.

Since Az is the thickness of a ferrite sheet (30 mil), then Az = 7.62 x 10-4 m.

The capacitance per unit length is

C = C1/(2Az) = 6.76 x 10-9/(1.52 x 10-3) = 4.45 x 106F/m.

The distributed resistance, R is calculated by: R = 2/(wad) where w is the width of the transmission line (w = 0.1"), a is the conductivity of the copper conductors (a = 5.65 x 107 mhos/m), and d is the skin depth of the copper conductor (d = 0.0661/v1/2 meters).

This gives R = 1.39 x 10-5 v12 Ohms/m.

Since the operating frequency, v, has not been determined yet, R is left in this form.

The distributed conductance, G is determined by: G = 2###w(tan 6)/de where v is the frequency of operation, E is the electrical permittivity of the dielectric, w is the width, tan 6 is the loss factor of the dielectric (tan 6 = 0.001 for the PLZT that is used in this example; since the laminated ferrite conductivity is much greater than the dielectric conductivity it is not a factor in the determination of the conductivity between the copper conductors), and ds is the total thickness of the dielectric materials. This gives G = 5.56 x 10-8 v mhos/m.

Since the operating frequency v, has not been determined yet G is left in this form for now.

The distributed inductance L is determined by:

L = µ#µ/w where p is the magnetic permittivity of the ferrite (m = 10,000 v0 = 1.26 x 10-2 Henries/m), 6p is the thickness of the ferrite (p = 0.06"; note that it is appropriate to use only the thickness of the ferrite because, in this case, all the magnetic field will be in the ferrite with none in the dielectric), and w is the width. This gives L = 7.54 x 10-3 Henries/m.

The values of frequencies which are necessary for this GET to operate are determined and then the characteristic impedance of the GET is calculated. The propagation velocity of a traveling harmonic wave along a transmission line is given by Vp = /8 and a + j = [(R + j#L) (G + j#C)]½. When 2##L/R > 10 and 2##C/G > 10 (i.e. when, in this case, v > 8.61 X 10-6 Hz) then α + j = j#(LC)½; i.e. = #(LC)½ which gives Vp = 1/(LC)½. It's also true that vp = ## = 2##/ . Solving these two equations gives v = /[2#(LC)½] = 8.71 x 102 p Hz.

For Pl = it/2, 3it/2, 5#/2... (where 1 is the length of the transmission line, 1 = 5.08 x 10-2 m) the calculated operating frequencies are v = 26.9 kHz, 80.7 kHz, 135 kHz...

When 2##L/R > 10 and 2##C/G > 10 (i.e. when, in this case, v > 8.61 x 10-6 Hz) then the characteristic impedance, Z0 = [(R + juL)/(G + j#C)]½, simplifies to Z0 = (L/C)i. Thus, in this case , the characteristic impedance of each transmission line in the GET is Z0 = 41.3 Ohms.

Each stage 192 and 194 of the isolation GET 190 is surrounded with a ferromagnetic material to provide a low reluctance path for the closed loops of magnetic flux as shown

in FIG. 6D. The two stages of the 1:1 isolation GET 190 are wired together as shown in FIG. 5. Each GET 192 and 194 may be fabricated in the manner described above. The copper conductors of the middle transmission line 198 are coupled to the voltage source 208. The copper conductors of the middle transmission line 204 are coupled to the load impedance 210.

When the frequency of the voltage source 208 is near one of the values calculated above (i.e. n = 26.9 kHz, 80.7 kHz, 135 kHz...), this structure will operate as a 1:1 isolation GET.

To form a 2:1 step down isolation GET, a GET is fabricated identical to the one represented in FIG. 5 but the width of the second GET 194 is changed to 0.2." The first GET 192 with a width of 0.1" has a characteristic impedance which is equal to the sum of the characteristic impedances of the upper and lower transmission lines 196 and 200 in GET 192.

The characteristic impedance of one of the transmission lines was calculated above and is 41.3 Ohms and thus the characteristic impedance of the first GET 192 is twice this value, i.e. Zl = 82.6 Ohms. The second GET 194 has a width of 0.2" and a characteristic impedance, Z2, that is 1A the first GET 192 because its width is 2 times the width of the first GET 192, thus, Z2 = 41.3 Ohms. The output voltage across the load impedance 210 is given by the following: Vout = V(Z2/Z1) where V is the voltage of the voltage source 208. In this example Vout = V(41.3/82.6) = V/2. Thus the structure shown is FIG. 5 may behave as a 2:1 step down transformer. Of course different dimensions may be used for either GET 192 cr 194 to change the transformer 190 to a step

up or step down at different ratios. The dimensions of the GETs 192 and 194 may also be different resulting in different impedances.

Rectangular bar structures (i.e. stripline structures) such as that used in this example need to be surrounded by high permeability material so that there is a low reluctance path for the magnetic flux to form closed loops. Rather than enclosing each GET 192 and 194 in FIG. 5 with high permeability material separately, the two sections can be placed side by side and then enclosed with high permeability material.

There are two ways in which the GET sections 192 and 194, can be placed side by side. FIG. 10A shows a GET 190 which is the same as that shown in FIG. 5 with the addition of a dashed line between the two GET sections 192 and 194 of the GET 190. The two GET sections 192 and 194 are placed together by folding, along the dashed line, one section onto the other.

The two GET sections 192 and 194 do not touch each other. The other way to bring the two GET sections 192 and 194 together is shown in FIG. 10B. The two GET sections 192 and 194 are wired together slightly differently in FIG. 10B than in FIG.

10A. As in FIG. 10A the GET sections 192 and 194 in FIG. 10B are folded over along the dashed line, and brought together but do not touch. These folded structures are then surrounded by a high permeability material.

It is usually easier to fabricate the two GET sections 192 and 194 as a single unit from the very start.

The cross section, perpendicular to the length, of the final

unit will look something like the structure shown in FIG. 10C with the two GET sections 192 and 194 side by side. A magnetic material 228 surrounds both GET sections 192 and 194 of the GET 190 in FIG. 10C. When the two GET sections 192 and 194 are folded over as shown in FIG. 10A then the magnetic fields will tend to cancel each other resulting in a much lower fields within the GET sections. This is helpful when the GET 190 is operated at high power levels so that the magnetic field saturation limit is not reached for the magnetic materials within the GET 190. When the two GET sections 192 and 194 are folded over as shown in FIG. 10B then the magnetic fields will be in the same direction and there will be essentially no change in the value of the magnetic fields within the GET.

It is clear that an n:1 isolation transformer may be constructed by connecting together two GETs with different characteristic impedances in series as is shown in FIG. 5.

The impedances may be made different by changing the widths of the GETs, by changing the values of E and p of the materials used in the GETs, or by changing the thicknesses of the dielectric and/or the ferrite material used in the GETs so that any ratio, n:l, can be achieved.

Single Stage n:l GETs: Type 1 There are other ways to use GETs to obtain n:1 transformers. Returning to the GET 120 shown in FIG. 3A, the upper and lower transmission lines 123 and 127 both have a characteristic mpedance of Zl and both have a phase factor of

p1. For this GET design, it is necessary that the phase factors of the upper and lower transmission lines 123 and 127 be equal but it is not necessary that their characteristic impedances be equal. The characteristic impedances are chosen to be equal here so that the upcoming graphs will be simpler.

For this example, an operating frequency and the materials in the transmission lines 123, 125, 127 of the GET 120 are chosen in a way to make the following conditions true: p21 = s/2 and p11 approximately equal to zero. With these conditions, the voltage that appears across the load impedance 138 equals the input voltage. To make things simpler, the input voltage is a triangular wave as is shown in the top graph in FIG. 11. The GET shown in FIG. 3A operates with this triangular wave for a period of time so that the initial transients die out leaving the steady state operation. At some time tl, a "snapshot" is taken of the voltage gradients that appear on each of the four conductors in the GET. The top graph in FIG. 11 shows that at t1 the input voltage is at a maximum. The bottom graph in FIG. 11 shows the voltages as a function of distance along each of the four conductors 122, 124, 126 and 128 in the GET 120 at time t1 . The voltage on the conductor 124 that extends from terminal 144 to terminal 152 varies from V/2 to 0 and the voltage on the conductor 126 from terminal 146 to terminal 154 varies from -V/2 to 0. This is as expected because there is a standing wave within the middle transmission line 125 that is 1/4 of the wavelength long. The voltage on the conductor 122 that extends from terminal 142 to terminal 150 is, at each point along the

length, less than the voltage that appears on conductor 124 from terminal 144 to terminal 152 by an amount of V/2. Thus, the bottom graph of FIG. 11 shows line a-b, corresponding to conductor 122, to be parallel to line c-d, corresponding to conductor 124, yet below it by an amount of V/2. This is as expected because p1l is approximately zero (i.e. the upper transmission line 123 is much shorter than the wavelength) and whatever voltage that appears across input terminals 142 and 144 also appears, to a close approximation, across output terminals 150 and 152 and all points along the length of the transmission line 130. Using similar reasoning, the lines for the voltage gradients for conductors 126 (e-f) and 128 (g-h) are derived and included in the bottom graph of FIG. 11.

As shown in FIG. 12 the load impedance 138 may be attached at point 232 of conductor 122 and point 234 of conductor 128. The points 232 and 234 are 3/4 of the distance along the GET 120. If a triangular voltage is applied at the input, as before, then the voltage on the load impedance 138 will be 3V/4. This is seen in FIG. 13. The upper graph shows the input voltage and the lower graph shows that the voltage between points 232 and 234 (i and j) at time tl, is -3V/4. Any output voltage, less than the input voltage, may be achieved by choosing to attach the load impedance 138 at the appropriate distance along the length of the GET 120. By exchanging the attachment points of the voltage source 136 and the load impedance 138, the GET 120 becomes a step up transformer and an output voltage, larger than the input voltage, can be obtained.

It should be noted that some care needs to be taken when designing GETs like those described in the previous paragraph. The shape of the waveform of the input voltage must be known so that the proper attachment point of the load impedance can be determined. If, for example, a sinusoidal input voltage is applied to the GET shown in FIG. 12 then the voltage across the load impedance 138 will not be -0.75V.

This is because the lines shown in the bottom graph of FIG. 13 will no longer be straight lines but will be shaped like 1/4 the wavelength of a sine wave. Thus, the voltage from points i to j will be -V sin(3X/8) = - 0.92V.

Multiple voltage outputs are achieved with the GET 120 shown in FIG. 12 by attaching more than one load impedance at different points along the length of the GET 120. FIG. 14 shows an additional load impedance 236 which is coupled to point 238 on conductor 122 and point 240 on conductor 128.

Load impedances 138 and 236 have different voltages applied to them. Although the two load impedances 138 and 236 are both electrically isolated from the input, they are not electrically isolated from each other.

There ar? many other variations on this embodiment.

The manufacturig costs of the GET 120 shown in FIG. 12 can be reduced by shortening the length of the upper and lower transmission lines of the GET 120 as is shown in FIG. 15. It is also possible to attach the load impedance 138 at points that are not equidistant from the end of the GET 120 as shown in FIG. 16. The voltage across the load 138 can be determined

from FIG. 17. When the input voltage is a triangular wave, as is shown in the upper graph of FIG. 17, the voltage gradients on the conductors of the GET, at time t1, will be as is shown in the lower graph. The voltage across terminals h and i is approximately V/2(1 + 3/4) = 0.875V. Of course there are many other variations which incorporate these principles.

Example 2 The GET 120 shown in FIG. 3A is to be constructed so that the material between the conductors 122, 124, 126 and 128 consists of two layers of dielectric material and one layer of magnetic (ferrite) material. The upper and lower transmission lines 123 and 127 of the GET 120 are constructed of identical materials while the middle transmission line 125 is constructed of a different material. The middle transmission line 125 will be constructed in a manner identical to that in the above example. From calculations in the above example, it is known that the frequencies that are needed for middle transmission line 125 of GET 120 to operate properly are approximately v = 26.9 kHz, 80.7 kHz, 135 kHz, ... At the lowest frequency, v = 26.9 kHz, the length of the middle transmission line 125 is 1/4 the wavelength.

The upper and lower transmission lines 123 and 127 of the GET 120 shown in FIG. 3A may be constructed in a manner identical the above example, except that the permeability of the ferrite layer 222 and the permittivity of the dielectric material 224 are each 100 times smaller in the upper and lower transmission liies 123 and 127 as compared to the middle

transmission line 125. The wave propagation velocity is 100 times faster in the upper and lower transmission lines as compared to the middle transmission line and also that p11 is approximately zero (as is required for this particular GET design to work properly). As with the GET in the above example, the GET 120 must be surrounded by a ferromagnetic material to provide a low reluctance path for the closed loops of magnetic flux.

When the input voltage is a triangular wave, as is shown in the upper graph of FIG. 11, the voltage gradients at time t1, will be as shown in the lower graph of FIG. 11. When the load impedance 138 is attached at terminals 150 and 156 the voltage across the load 138 will be the same as the input voltage. The voltage across the load 138 can be reduced to any fraction of the input voltage by attaching the load impedance 138 at the appropriate points on conductor 122 and conductor 128.

The changes that occur when instead of using the lowest frequency, v = 26.9 kHz, may be shown by using the next to lowest frequency, v = 80.7 kHz, with the GET 120 described in the previous paragraph. p11, for the upper and lower transmission lines 123 and 127, is still approximately zero while the length of the middle transmission line 125 of the GET 120 is now 3/4 the wavelength. The voltage gradients on the four conductors 122, 124, 126 and 128 of the GET 120 are different. If the input voltage is a triangular wave as shown in the upper graph of FIG. 18 then the voltage gradients on the conductors 122 and 128 at time t1, is as is shown on the

lower graph of FIG. 18. There is a standing wave with a length of 3/4 of the wavelength in the middle transmission line. The voltage gradients on conductors 124 and 126 will have the shape of 3/4 of a wavelength. The voltage on conductor 122 is less than the voltage on conductor 124 by V/2 and the voltage on conductor 128 is greater than the voltage on conductor 126 by V/2 at all points along the length. As with the previous example, the load impedance 138 can be attached at any points along the length of the GET 120 to obtain the desired voltage. The lower graph of FIG. 18 shows that any voltage may be obtained between 0 and 2V by choosing the appropriate points for attaching the load. Thus, this GET structure, operated at a frequency such that p21 = 3it/2 (i.e. n = 80.7 kHz), operates as an n:1 isolation transformer for n < 2. (Or, by exchanging the voltage source 136 with the load impedance 138, n > The GET 120 in this embodiment has different characteristics when p11 = it. The length of the upper and lower transmission lines 123 and 127 will be M a wavelength.

A frequency of v = 26.9 kHz is used so that the length of the middle transmission line 125 is 1/4 of a wavelength. When a triangular voltage, such as is shown in the upper graph of FIG. 19, is input then the voltage gradients on the four conductors 122, 124, 126 and 128, at time t1, will be as shown in the lower graph of FIG. 19. The voltage gradients on conductors 124 and 126 are as before, shown in FIG. 11. The voltage gradients on conductors 122 and 128, however, are different. The voltage on conductor 122 is no longer less

than the voltage on conductor 124 by V/2 for the entire length of the GET 120. Instead, the voltage between conductors 122 and 124 grades from -V/2 to V/2. There is a similar grading in the voltage between conductors 126 and 128. It is important to understand that since p11 is not approximately zero, it is not possible to attach a load impedance at just any point along the length. At points other than at the output terminals 148 and 158 there will be reflections produced which severely interfere with the ability of the GET 120 to maintain a constant voltage output.

When the GET 120 is designed with p11 = 2#, 3w, 4w, ... and/or p21 = 5it/2, 7it/2, 9it/2, ..., the voltage gradients on the conductors become even more complex.

Single Stage n:l GETs: Tvoe 2 A further GET design is shown in FIG. 20A. FIG. 20A shows a GET 250 which is circular disk shape with radius r2 having a circular hole 252 in the center with radius r1. The cross section of the GET 250 is shown in the bottom drawing of FIG. 20B. The cross section is one of the familiar GET cross sections discussed above. There are four conductors, 254, 256, 258 and 260, separated by dielectric and magnetic (ferrite) materials 262, 264 and 266 to allow the propagation of electromagnetic waves. The GET 250 has an inner diameter 268 and an outer diameter 270 and may be connected to a voltage source and a load impedance as shown in FIG. 3A. When the voltage source is connected to the inside diameter 268 and the load impedance is connected to the outside diameter 270,

waves will propagate from the inside diameter 268 to the outside diameter 270 much like the waves made by dropping a pebble into a calm pool of water. As the waves propagate outward their amplitudes decrease. When the waves reach the outside diameter, the voltage will have dropped to a fraction, (r1/r2)%, of its initial value. Waves that are reflected from the outside diameter 270 will increase in amplitude as they move toward the center so that after a round trip the amplitude will return to its original value (assuming no losses have occurred).

Materials are selected for the GET 250 so that p1(r2 - r1) = # and p2(r2 - rl) = it/2. This is analogous to the second example discussed above except that the voltage across the load impedance will not be the same as the input voltage, V, but it will have been reduced to (r1/r2)iV. As with the previous example, the GET 250 operates as a constant voltage source. Reversing the voltage source and the load impedance results in a GET. that steps up the voltage by a factor of (r2/r1). It is clear that any turns ratio n:1 can be achieved by constructing the GET 250 with the appropriate values for r and r2.

The voltage must be input along the entire inside diameter 268 at approximately the same time and the voltage at the outside diameter 270 must be "collected" and appear across the load impedance at approximately the same time. The four conductors 254, 256, 258 and 260, overhang at the inside diameter 268 and tie outside diameter 270 of the dielectric and magnetic material 262, 264 and 266 that are between them.

When the voltage source is input at a single point on the inside diameter 268, the voltage and current quickly spreads around the inside circumference of the GET 250 at nearly the speed of light. Since the propagation velocity along a radius is designed to be much less than the speed of light, the wave propagates around the inside diameter 268 first then propagates outward toward the outside diameter 270. This assures that the propagating wave fronts are, for all practical purposes, circular and centered. The same steps happen at the outside diameter 270, only reversed. The wave front hits the outer edge of the dielectric and magnetic material 262, 264 and 266, then quickly, at nearly the speed of light, circles around the outer edge to the points of attachment of the load impedance.

A piece of the whole disk can have the same turns ratio, n:1, as the whole disk. If two cuts are made along radii of the GET disk 250, a smaller piece 272 is obtained.

When a voltage source is applied to the end of the smaller piece 272 with radius r1, the wave will propagate toward the end with radius r2 and the voltage will drop by a factor of (r1/r2)i. In principle, the angle between the two cuts that are made to form the smaller piece 272, may be made very small.

In practice, there is a limit. When the piece is made too small for the current that is passing through the GET 250, there will be excessive heating. Thus, the size of the piece 272 is matched to the application for which it is used.

The curved input and output ends of the smaller piece 272 which are described by radii r1 and r2 may have

slightly different shapes. Straight edges, with widths w1 and w2 may be used on the small pieces instead of the curved edges r1 and r2. In this case, the output voltage will be approximately (w1/w2)V instead of (r1/ r2)iV. Finally, as with the stripeline configuration discussed earlier, the magnetic flux lines will extend outside the smaller piece 272. Thus, when high permeability materials are used in the construction of this geometry, there will be an improvement in the operation of the device if it is surrounded by ferromagnetic material so that the closed loops of the magnetic flux will always pass through a low reluctance path.

Single Stave n:1 GETs: Tvpe 3 Another GET design is shown in FIG. 21. A GET 280 has conductors 282, 284, 286 and 288 with dielectric and ferrite materials 290, 292 and 294 therebetween. The distance between the conductors 282, 284, 286 and 288 which form upper and lower transrlission lines 283 and 287 is graded from a larger value to a smaller value as the distance x, increases.

This means that the characteristic impedance Z1 of the upper and lower transmission lines is a function of x. There are several ways in which the grading can occur. The dielectric material between conductors 282 and 284 and between conductors 286 and 288 may be thicker at x=0 than at x=l, and/or the magnetic (ferrite) material between conductors 282 and 284 and between conductors 286 and 288 can be thicker at x=0 than at x=l, and/or the values of E and p can be graded so that E is low and/or p is nigh at x=0 and E is high and/or p is low at

x=l. In all these cases, the characteristic impedance of the upper and lower transmission lines in the GET 280 is higher at x=0 than at x=l. A middle transmission line 285 in the GET 280 can also be graded, in a similar manner.

If the upper and lower transmission lines 283 and 287 in the GET 280 have thicker dielectric material at x=0 than at x=l, then the capacitance per unit length C, is lower at x=0 than at x=l. The characteristic impedance at x=0 is also higher than at x=l. If this GET 280 is operated at a frequency such that p21 = s/2 and p11 = it, then it will behave like a step down isolation transformer. The waves generated by the voltage source will propagate toward the other end of the GET 280. The amplitude of the wave will change as the characteristic impedances of the transmission lines change.

Waves that are reflected from the other end will propagate back to the input and their amplitude will return to their original value (assuming no losses). The turns ratio, n:1, may be determined as follows: the power going into the GET 280 from the voltage source equals the power leaving the GET.

Thus, Pin = V²/Z0,in = V²out/Zo,out. The variables are rearranged to <BR> <BR> <BR> <BR> obtain: Vout = (ZO,out/ZO,in) ½ V; where V is the voltage of the<BR> <BR> <BR> <BR> <BR> source, Z0 in is the characteristic impedance at x=0, and To out is the characteristic impedance at x=l. This is a step down transformer with a turns ratio of (Z0,in/Z0,out)½:1.

Finally, as with the stripeline configuration discussed earlier, the magnetic flux lines will extend outside the device. Thus, when high permeability materials are used in the construction of this geometry, there will be an

improvement in the operation of the device if it is surrounded by ferromagnetic material so that the closed loops of the magnetic flux will always pass through a low reluctance path.

Mixture of Single Stage n:1 GETs Designs The GET designs described in the three types of single stage n:l GETs (FIGS. 3, 20 and 21) may be fabricated in any combination. For example, a single stage GET may be fabricated in the shape of a disk with a hole in the center.

The dielectric and ferrite materials that are between the four conductors may be graded in thickness so that they are thicker near the center of the disk and thinner near the outside of the disk. A section of the disk can be cut out by cutting along two radii. And finally, several load impedances can be attached at different points along the length of the GET so that they have different voltages applied to them.

As with the stripeline configuration discussed earlier, the magnetic flux lines will extend outside the cut section of the disk. Thus, when high permeability materials are used in the construction of this geometry, there will be an improvement in the operation of the device if it is surrounded by ferromagnetic material so that the closed loops of the magnetic flux will always pass through a low reluctance path.

Second Mode Of Operation An additional mode of operation is described here for GET 120 in EIG. 3A. There are three transmission line

modes which operate simultaneously in GET 120. The first two modes were identified and discussed previously: The first mode has traveling harmonic waves propagating between conductors 122 and 124 as well as identical traveling harmonic waves traveling between conductors 126 and 128. Thus, only the upper and lower transmission lines are used in this mode. In this first mode there are no first order electric or magnetic fields in the material 134 between conductors 124 and 126.

The first mode may be viewed as a simple pair of transmission lines in which the AC voltage from a voltage source 136 is divided by two, propagated along the upper and lower transmission lines shown in FIGs. 3A and 3B, recombined at the output, and applied to a load impedance 138.

The second mode uses only the middle transmission line comprising conductors 124 and 126 and material 134. No first order electric or magnetic fields are generated outside the material 134. This mode can be simply understood as being the voltage source 136 connected to the conductors 124 and 126 of the middle transmission line at input terminals 144 and 146 while the other end of the conductors 124 and 126 of the middle transmission line are short circuited between output terminals 152 and ',.54. The characteristic impedance of the middle transmission line is Z2 Using equation 1 above (with ZL = 0 in the equation) to determine the input impedance for the second mode; the impedance that the voltage source "sees," after the initial transients, have settled due to the second mode is as follows: 1n2 = 3 Z2 tanp2l,

where p2 is 2s divided by the wavelength of the waves propagating in the middle transmission line and 1 is the length of the transmission line. The input impedance for the second mode can take on a range of values by changing the value of p2l. This value may be changed by changing the frequency of the voltage source, and/or by changing the values of E2 and/or 82 of the material used between conductors 124 and 126, and/or by changing the length 1 of the transmission line.

When the input impedance for the second mode is low, a large current flows through the wire connecting terminals 152 and 154. The large current flow wastes energy.

As described previously, the effects of the second mode may be minimized by making its input impedance as high as possible. p21 is made to be equivalent to it/2, 3it/2, 5it/2, (i.e. the frequency of the voltage source is selected such that the wavelength and hence p2 approximately satisfies one of these conditions) . This embodiment relates to a 1:1 isolation GET with all the materials separating the conductors of equal length and where the materials in the upper and lower transmission lines 123 and 127 as being equivalent.

The effects of the second mode may also be minimized by causing the value of p21 to approach but not equal, 0, it, 2r, 3it, ... (i.e. the frequency of the voltage source is selected such that the wavelength, and hence p2, approximately satisfies one of these conditions) . This will be described in more detail in Example 3 below.

A third transmission line mode in GET 120 shown in FIG. 3A occurs because the top conductor 122 and the bottom

conductor 128 form a transmission line which will be termed the outside transmission line. The outside transmission line has a shorting wire connected at terminals 142 and 148 at one end and the load impedance 138 is connected at the other end at terminals 150 and 156. Both the characteristic impedance, ZOS, and the phase factor, os, depend on all three of the materials 130, 134, and 132. The input terminals for the outside transmission line are 150 and 156.

The third transmission line mode is the outside transmission line with its output terminals 142 and 148 shorted together. Using equation 1 above (with ZL = 0 in the equation) to determine the input impedance for the third mode; the impedance tat the voltage source "sees," after the initial transients have settled due to the third mode, is as follows: Zin3 = j Zos tanP05l, where os is 27 divided by the wavelength of the waves propagating in the outside transmission line and 1 is the length of the transmission line. The input impedance for the third mode can take on a range of values by changing the value of p051. This value may be altered by changing the frequency of the voltage source and/or by changing the values of E and/or p of the materials 130, 134, and 132 used between conductors 122 and 128, and/or by changing the length 1 of the transmission line. When the input impedance for the third mode is low, a large current flows through the wire connecting terminals 142 and 148. This is a waste of energy.

The effects of the third mode may be minimized by making its input impedance as high as possible. One way to minimize the third mode effects is to make osl equivalent to it/2, 3it/2, 5it/2, ... (i.e. the frequency of the voltage source is selected such that the wavelength, and hence os approximately satisfies one of these conditions) . The effects of the third mode may also be minimized by causing the value of p051 to approach, but not equal, 0, it, 2#, 3#, ... (i.e. the frequency of the voltage source is selected such that the wavelength, and hence os approximately satisfies one of these conditions) . The following example uses the latter method to minimize the effects of the third mode.

Example 3 A 1:1 isolation GET, as is shown schematically in FIG. 3A, will be used in this example to illustrate how the effects of the second and third modes can be minimized by causing the values of 2l and osl to approach, but not equal, 0, it, 2it, 3it, ... 21 and osl approach 0 in this example; however, it will be clear from the mathematics that it is equivalent to cause the values of 2l and osl to approach, but not equal, it, 2#, 3#, ... instead. The GET geometry to be used in this example is shown in cross section in FIG. 6A. In this case, #1 = E3 and µ1 = SO SO that FIG. 6A matches FIG. 3A.

Considering mode two first: As previously stated, the input impedance for the second mode is given by: Zin2 = j Z2 tan 2l.

As p21 approaches 0, then tan 2l approaches 2l. Therefore the following is approximately true: Zin2 = j Z2p2l.

The characteristic impedance and the propagation velocity have been determined for the parallel wire transmission line used in this example (see, for example, "Transmission Lines," Robert A. Chipman, Schaum's Outline Series in Engineering, McGraw-Hill Book Co., 1968): Z0 = 120 (µ/#)3/2 ln(s/(2a) + ((s/(2a))2 - ) where p is the relative permeability (i.e. multiply p by Po to obtain the actual permeability), E is the relative dielectric constant (i.e. multiply E by E0 to obtain the actual dielectric constant), s is the separation distance between the two parallel wires, and a is the radius of a wire.

The propagation velocity is given by: vp = w/p = 2itv/P = (3 x 108 m/s)/(µ#)3/2 where v is the frequency of operation, p and E are the relative permeability and the relative dielectric constant respectively. Solving the propagation velocity equation for p gives: P = 2## (µ#)½/(3 x 108). The expressions for p and the characteristic impedance Zo (which equals Z2 in this example) are substituted into the equation for the input impedance. After simplifying, Z102 = 240itupl(ln(s/(2a) + ((s/(2a))2 - 1)½))/(3 x 108). The input impedance does not depend on the dielectric constant. The input impedance may be made as high as is needed by increasing the operating frequency v, and/or the relative permeability p, and/or the length 1, and/or the ratio s/a.

The input impedance for the third mode, the outer loop transmission line, is derived in the same way:

Zin3 = j 240Tvy1(1n(s/(2a) + ((s/(2a))2 - ))/(3 x 108). The parallel wire GET will now be described so that Zin2 and Zin3 can be determined.

FIG. 22 shows the cross section of the parallel wire GET in more detail. Materials 130 and 132 are a dielectric ceramic such as alumina. Two molybdenum wires, 122 and 124, each with a radius of a = 25 mils, are surrounded by the ceramic 130 while the ceramic is still in the "green" state; i.e. the ceramic has not yet been fired. The wires have a separation of s1 = 175 mils and the thickness of the ceramic lies between 2 and 4 mils at the thinnest point to provide electrical isolation. The pair of wires, 126 and 128, are surrounded in exactly the same way. The two pairs of wires, separated by s2 = 175 mils, are then surrounded by ferrite material 134 while it is still "green" so that the overall dimensions of the cross section are 0.1" x 0.7". The whole GET is then fired o finish the ceramic materials 130, 132 and 134. Material .34 is a Mn-Zn ferrite with a relative permeability p = 2,900 and a resistivity p = 1,000 Ohm-cm.

Because material 134 is conducting, all the electric fields will be within materials 130 and 132 and because the permeability of material 134 is so high, all the magnetic flux density (i.e. the B field) will be within material 134. The input impedance for mode two, Zin2 is then determined as follows.

The length of the GET is 1 = 1" and it is configured electrically as shown in FIG. 3A. For mode two, the separation distance s between wires 124 and 126 (see FIG. 22)

is s = s2 = 0.17". When the operating frequency is v = 300 kHz then the input impedance for mode two is calculated from the equation given above: Z102 = 107 Ohms. The input impedance for mode three, Zin3 is then determined.

The separation distance to be used to calculate Z103 is not the actual separation distance between wires 122 and 128 (see FIG. 22). Because the input impedance comes entirely from the magnetic material that separates the wires and not from the dielectric material, it is the width of the magnetic material that is used for the separation distance; i.e. s = 0.17". Therefore, Z103 = 107 Ohms.

The impedances of the second and third modes of this GET are comparable in magnitude to the impedances due to the primary and secondary windings of conventional transformers yet the GET is much smaller and lighter. This GET can directly replace conventional transformers operating at 300 kHz without the need for additional circuitry. Similar GETs, which can directly replace conventional transformers, can be designed to operate at frequencies from about 20 kHz to a few megahertz.

The GET described above, in Example 3, is a 1:1 GET.

To obtain other turns ratios GETs can be combined in the manner shown in FIG. 4 (i.e. the inputs of the GETs are wired in series and the outputs are wired in parallel) or a single stage GET can produce any turns ratio when constructed as shown in FIG. 12. There can be additional constraints on the design of the GETs so that they will operate properly when configured as shown in FIGs. 4 and 12. See the text that

describes these two figures to find the additional constraints that may be involved.

Figures 23A - 28B show schematic drawings of additional GET embodiments. Each is wired differently from the GETs that were described previously in FIGs. 3A and 3B.

As with the previously described GETs, the GETs schematically represented in FIG. 23A - 28B can be fabricated in any geometry that a transmission line can embody; e.g. see FIGs.

6A - 7B. The unique features of each GET will be described along with the unique design constraints (i.e. the unique materials and geometries). For each, methods will be described for obtaining different turns ratios. And finally, a specific example will be given describing how a GET can be constructed for each of the figures. Of course it is to be understood that these are merely examples for purposes of illustrating the present invention which is not limited to those examples.

FIG. 23A is the same as FIG. 3A except that the wire connecting terminals 152 and 154 in FIG. 3A is absent in FIG.

23A. There are four transmission lines in the structure shown in FIG. 23A that are of importance: The upper 123, the middle 125, the lower 127, and the outside transmission line which is comprised of the top conductor 122 and the bottom conductor 128, separated by the three materials 130, 134 and 132. The phase factors for these four transmission lines are p1, p2, p3, and p05 There are two sets of conditions for this GET to operate properly. Either one set or the other must be met for this GET to operate properly. The first set of conditions is:

1l = 0, #, 2#, 3#, ..., 2l = 0, #, 2#, 3#, ..., 3l = 0, #, 2#, 3#, ..., and osl can be either osl = 0, #, 2#, 3#, ..., or osl = #/2, 3#/2, 5#/2, ...

(Note the symbol "=" used above to designate that osl approaches, but does not equal 0, it, 2#, 3#, ... This symbol will be used throughout the remainder of this description.) The upper and lower transmission lines must be in phase: 1l - 3l = 0, # 2#, # 4#, ...

And the middle transmission line must be out of phase with the upper and lower transmission lines: 1l - 2l = # #, # 3#, # 5# When all of these conditions are met this GET is a 2:1 step up isolation voltage transformer. A 2:1 step down voltage transformer is obtained by exchanging the voltage source and the load impedance, as is shown in FIG. 23B. Other turns ratios are obtained by wiring GETs together in the fashion shown in FIG. 4 (i.e. the inputs of the GETs are wired in series and the outputs are wired in parallel); the more GETs that are wired together in this fashion the higher the turns ratio. Another way to change the turns ratio is to attach the load impedance at different points such as is shown in FIG.

12.

The second set of conditions for this GET is: 1l = #/2, 3#/2, 5#/2, ..., 2l = 0, #, 2#, 3#, ..., 3l = #/2, 3#/2, 5#/2, ..., and osl can be either osl # 0, #, 2#, 3#, ..., or osl = #/2, 3#/2, 5#/2, ...

The upper and lower transmission lines must be in phase: 1l - 3l = 0, # 2#, # 4#, ...

When all of these conditions are met this GET is a constant current source with current I = V/ (Z1 + Z2) where V is the voltage of the source 136 and Z1 and Z2 are the respective characteristic impedances of the upper and lower transmission lines 123 and 127 of the GET. When the voltage source and the load impedance are exchanged, as is shown in FIG. 23B, the GET behaves in exactly the same way. A constant voltage source, with any desired turns ratio, is obtained by wiring two of these GETs together as shown in FIG. 5 and described previously.

Examples 4-7 below use a rectangular bar geometry, however, any other geometry may be used. The dielectric and ferrite materials used in the construction of these GETs are assembled with epoxy as described above in Example 1. To simplify these examples, it is assumed that the thicknesses of the epoxies are so small that they can be ignored. This will produce an appropriate result that differs from the actual result by only a few percent. If a more accurate result is required then the procedure detailed in Example 1 may be followed.

Fig. 8C shows the cross section of these GET designs. Note that the dielectric materials 224 and the ferrite materials 222 have different thicknesses which are given by de and d", respectively. To meet the conditions needed for these examples, the design shown in FIG. 8C will be modified slightly in a manner to be described shortly. First the equation describing the phase factor for rectangular bar

(i.e. stripline or parallel plane) transmission lines is derived: The propagation velocity is vp = wlp = 1/(LC)i, where L is the inductance per unit length and C is the capacitance per unit length of the transmission line and # = 21rv (where v is the operating frequency) . The phase factor is solved to obtain p = 2##(LC)½. For the transmission line geometry used in these examples C and L have been published (see for example, "Transmission Lines," Robert A. Chipman, Schaum's Outline Series in Engineering, McGraw-Hill Book Co., 1968) and are given by C = #w/d# and L = µdµ/w where w is the width of the transmission line. Substituting these values and multiplying by the length, 1, results in: Pl = 2##l (µ#dµ/d#)½ Example 4 This is an example of a GET that satisfies the first set of conditions for the GETs shown in FIGs. 23A and 23B.

The GET in this example is to be operated at a frequency of v = 300 kHz and the length is 1.86 cm. The upper transmission line 123 and the lower transmission line 127 each have a single piece of dielectric material with E = 1.7#0 and thickness d# = 10 mils. Since there is no separate ferrite material d# = d, and µ = p. Calculating Pl for the upper and lower transmission lines: 1l = 3l = 2#(3 x 105) (0.0186) [(4# x 10-7) (1.7) (8.85 x 10-12)]½ P11 = P31 = 1.5 x 10-4 (this is approximately zero).

The middle transmission line 125 has two pieces of dielectric material, each with E = 20,000#0 and thickness d# = 2 mils, and a single piece of ferrite material with p = 2,900µ0 and thickness dp = 50 mils. Calculating l for the middle transmission line 125: 2l = 2#(3 x 105) (0.0186) [(2,900) (4# x 10-7) (20,000) (8.85 x 10-12) (50)/(2 + 2)]1/ p21 = 3.148 (this is approximately it).

The outside transmission line uses the materials that are in the upper, middle, and lower transmission lines 123, 125 and 127; therefore, the inductance per unit length is given by Los = L1 + L2 + L3 and the capacitance per unit length is given by 1(Cos = 1/C1 + 1/C2 + 1/C3. In this example Los = L2 and COS = C1/2. This gives osl = 2##l(L2C1/2½ = 2# (3 x 105) (0.0186) [(2,900) (4# x 10-7) (1.7) (8.85 x 10-12) (50)/2(10)]½ p051 = 1.30 x 10-2 (this approaches zero).

This example has shown that this GET design meets all of the first set of conditions for Pil, 2l, p31, and osl that are necessary for the GETs in FIGs. 23A and 23B to function properly.

Example 5 This example of a GET satisfies the second set of conditions for the GETs shown in FIGs. 23A and 23B. The GET in this example is to be operated at a frequency of v = 300 kHz and the length is 0.93 cm. The upper transmission line 123 and the lower transmission line 127 each have two pieces

of dielectric material, each with e = 20,000#0 and thickness d# = 2 mils, and a single piece of ferrite material with p = 2,900µ0 and thickness dH = 50 mils. Calculating Pl for the upper and lower transmission lines: 1l = 3l = 2#(3 x 105) (0.0093) [(2,900) (4# x 10-7) (20,000) (8.85 x 10-12) (50)/(2 + #1l = p31 = 1.574 (this is approximately #/2).

The middle transmission line 125 has a single piece of dielectric material with e = 1.7#0 and thickness d# = 10 mils.

Since there is no separate ferrite material dE = dµ and p = µ0.

Calculating l for the middle transmission line: 2l = 2#(3 x 105) (0.0093) [(4# x 10-7) (1.7) (8.85 x 10-12)]½ p21 = 7.62 x 10-1 (this is approximately zero).

The outside transmission line uses the materials that are in the upper, middle, and lower transmission lines 123, 125 and 127; therefore, the inductance per unit length is given by Los = L1 + L2 + L3 and the capacitance per unit length is given by 1/Cos = 1/C1 + 1/C2 + 1/C3. In this example Los = L1 + L3 = 2L1 and Cos = C2. This gives osl = 2##l(2L1C2)½ = 2#(3 x 105) (0.0093) [2(2,900) (4# x 10-7) (1.7) (8.85 x 10-12) (50)/(10)]½ osl = 1.3 x 10-2 (this approaches zero).

This GET design meets all of the second set of conditions for p11, p21, p31, and osl that are necessary for the GETs in FIGs.

23A and 23B to function properly.

Another GET embodiment is shown in FIG. 24A. The GET in FIG. 24A is the same as the GET in FIG. 3A except that the wire connecting terminals 142 and 148 in FIG. 3A is

replaced by the load impedance and terminals 150 and 156 are not connected in FIG. 24A. As described for FIG. 23A, there are four transmission lines in the structure shown in FIG. 24A that are of importance. The phase factors for these four transmission lines are 1, 2, 3, and os. There are two sets of conditions for this GET to operate properly. Either one set or the other must be met for this GET to operate properly.

The first set of conditions is: 1l = #/2, 3#/2, 5#/2, ..., 2l = 0, #, 2#, 3#, ..., 3l = #/2, 3#/2, 5#/2, ..., osl = 0, #, 2#, 3#, ..., The upper and lower transmission lines must be in phase: 1l - 3l = 0, # 2#, # 4#, When all of these conditions are met this GET is a 1:1 isolation voltage transformer. When the voltage source and the load impedance are interchanged, as is shown in FIG. 24B, the GET is also a 1:1 isolation voltage transformer. Other turns ratios are obtained by attaching the load impedance at different points such as is shown in FIG. 12.

The second set of conditions for this GET is: 1l = #/2, 3#/2, 5#/2, ..., 2l = #/2, 3#/2, 5#/2, ..., 3l = #/2, 3#/2, 5#/2, ..., osl = 0, #, 2#, 3#, ..., The upper and lower transmission lines must be in phase: 1l - 3l = 0, # 2#, # 4#, ...

When all of these conditions are met this GET is a 1:1 isolation voltage transformer. When the voltage source and the load impedance are interchanged, as is shown in FIG. 61B, the GET is also a 1:1 isolation voltage transformer.

The GET described in Example 5, when wired as is shown in FIGs. 24A and 24B, meets all the first set of conditions. The following example describes a GET which meets all the second set of conditions for FIGs. 24A and 24B.

Example 6 This GET satisfies the second set of conditions for the GETs shown in FIGs. 24A and 24B. The GET in this example is operated at a frequency of v = 300 kHz and the length is 2.79 cm. The upper transmission line 123 and the lower transmission line 127 each have two pieces of dielectric material, each with E = 20,000#0 and thickness d# = 3.33 mils, and a single piece of ferrite material with p = 2,900µ0 and thickness dU = 9.26 mils. Calculating Pl for the upper and lower transmission lines: 1l = 3l = 2# (3 x 105) (0.0279) [(2,900) (4# x 10-7) (20,000) (8.85 x 10-12) (9.26)/(3.33 + 3.33)]½ p11 = p31 = 1.575 (this is approximately #/2).

The middle transmission line 125 has two pieces of dielectric material, each with E = 20,000#0 and thickness dE = 4 mils, and a single piece of ferrite material with p = 2,900µ0 and thickness Q = 100 mils. Calculating Pl for the middle transmission line: p21 = 2#(3 x 105) (0.0279) [(2,900) (4# x 10-7) (20,000)(8.85 x 1012) (100)/(4 + 2l = 4.722 (this is approximately 3#/2).

The outside transmission line uses the materials that are in the upper, middle, and lower transmission lines 123, 125 and

127. In this example, since the dielectric and ferrite materials in all three transmission lines have the same respective permittivity and permeability, the sum of all three ferrite thicknesses and the sum of all six dielectric thicknesses are used in the equation: oSl = 2#(3 x 105) (0.0279) [(2,900) (4# x 10-7) (20,000)(8.85 x 10-12) (9.26 + 100 + 9.26)/(3.33 + 3.33 + 4 + 4 + 3.33 + 3.33)]½ osl = 3.149 (this is approximately #).

This example demonstrates that this GET design meets all of the second set of conditions for 1l, p21, p31, and osl that are necessary for the GETs in FIGs. 24A and 24B to function properly.

The GET embodiment shown in FIG. 25A is the same as FIG. 3A except that the load impedance is connected to terminals 142 and 148 and terminals 152 and 154 are open. As with FIGs. 23A and 23B, there are four transmission lines in the structure shown in FIG. 25A that are of importance. The phase factors for these four transmission lines are 1, P21 3, and P05. There are two sets of conditions for this GET to operate properly. Either one set or the other must be met for this GET to operate properly. The first set of conditions is: 1l = 0, #, 2#, 3#, ..., 2l = 0, #, 2#, 3#, ..., 3l = 0, #, 2#, 3#, ..., and osl can be either osl # 0, #, 2#, 3#, ..., or osl = #/2, 3#/2, 5#/2, ...

The upper and lower transmission lines must be in phase: 1l - 3l = 0, # 2#, # 4#, ...

And the middle transmission line must be out of phase with the upper and lower transmission lines:

1l - 2l = # #, # 3#, # 5# ...

When all of these onditions are met this GET is a 2:1 step up isolation voltage transformer. A 2:1 step down voltage transformer is obtained by exchanging the voltage source and the load impedance, as is shown in FIG. 25B. Other turns ratios are obtained by wiring GETs together in the fashion shown in FIG. 4 (i.e. the inputs of the GETs are wired in series and the outputs are wired in parallel); the more GETs that are wired together in this fashion the higher the turns ratio. Another way to change the turns ratio is to attach the load impedance at different points such as is shown in FIG.

12.

The second set of conditions for this GET is: 1l = #/2, 3#/2, 5#/2, ..., 2l = 0, #, 2#, 3#, ..., 3l = #/2, 3#/2, 5#/2, ..., and osl can be either osl # 0, #, 2#, 3#, ..., or osl = #/2, 3#/2, 5#/2, ...

The upper and lower transmission lines must be in phase: 1l - 3l = 0, # 2#, # 4#, ...

When all of these conditions are met this GET is a 1:1 isolation voltage transformer. When the voltage source and the load impedance are interchanged, as is shown in FIG. 25B, the GET is also a 1:1 siolation voltage transformer.

The GET that is described in Example 4, when wired as is shown in FIGs. 25A or 25B, meets all the first set of conditions. The GET that is described in Example 5, when wired as is shown in FIGs. 25A or 25B, meets all the second set of conditions for FIGs. 24A and 24B.

The GET embodiment shown in FIG. 26A is the same as FIG. 3A except terminals 142 and 148 and terminals 152 and 154 are open. As with FIGs. 23A and 23B, there are four transmission lines in the structure shown in FIG. 26A that are of importance. The phase factors for these four transmission lines are 1, 2, 3, and P05. There is a set of conditions necessary for this GET to operate properly: 1l = #/2, 3#/2, 5#/2, ..., 2l = 0, #, 2#, 3#, ..., 3l = #/2, 3#/2, 5#/2, ..., osl = 0, #, 2#, 3#, ..., The upper and lower transmission lines must be in phase: 1l - 3l = 0, # 2#, # 4#, ...

When all of these conditions are met this GET is a 1:1 isolation voltage transformer. When the voltage source and the load impedance are interchanged, as is shown in FIG. 26B, the GET is also a 1:1 isolation voltage transformer.

The GET that is described in Example 5, when wired as is shown in FIG. 26A or 26B, meets the set of conditions necessary for the GETs to operate properly.

The GET embodiment shown in FIG. 27A is the same as FIG. 3A except that the load impedance is connected to terminals 142 and 148 and terminals 150, 152, 154 and 156 are left open. As with FIGs. 23A and 23B, there are four transmission lines in the structure shown in FIG. 27A that are of importance. The phase factors for these four transmission lines are 1, 2, 3, and os. There is a set of conditions necessary for this GET to operate properly: 1l = #/2, 3#/2, 5#/2, ..., 2l = 0, #, 2#, 3#, ..., 3l = #/2, 3#/2, 5#/2, ..., osl = 0, #, 2#, 3#, ...,

The upper and lower transmission lines must be in phase: 1l - 3l = 0, # 2#, # 4#, ...

When all of these conditions are met this GET is a 1:1 isolation voltage transformer. When the voltage source and the load impedance are interchanged, as is shown in FIG. 27B, the GET is also a 1:1 isolation voltage transformer.

The GET that is described in Example 5, when wired as is shown in FIGs. 27A or 27B, meets the set of conditions necessary for the GETs to operate properly.

The GET embodiment shown in FIG. 28A is the same as FIG. 3A except that the load impedance is connected to terminals 142 and 148 and terminals 150 and 156 are shorted out. As with FIGs. 23A and 23B, there are four transmission lines in the structure shown in FIG. 28A that are of importance. The phase factors for these four transmission lines are 1, 2, 3, and p05. There are two sets of conditions for this GET to operate properly. Either one set or the other must be met for this GET to operate properly. The first set of conditions is: 1l = 0, #, 2#, 3#, ..., 2l = 0, #, 2#, 3#, ..., P31 = 0, it, 27r, 3#, ..., and osl can be either osl = 0, #, 2#, 3#, ..., or osl = #/2, 3#/2, 5#/2, ...

The upper and lower transmission lines must be in phase: 1l - 3l = 0, # 2#, # 4#, ...

When all of these conditions are met this GET is a 1:1 isolation voltage transformer. When the voltage source and the load impedance are interchanged, as is shown in FIG. 28B, the GET is also a 1:1 isolation voltage transformer. Other

turns ratios are obtained by attaching the load impedance at different points such as is shown in FIG. 12.

The second set of conditions for this GET is: 1l = 0, #, 2#, 3#, ..., 2l = #/2, 3#/2, 5#/2, ... p31 = 0, it, 2#, 37r, ..., and osl can be either osl # 0, #, 2#, 3#, ..., or osl = #/2, 3#/2, 5#/2, ...

The upper and lower transmission lines must be in phase: 1l - p31 = 0, + 27r, + 4#, When all of these conditions are met this GET is a 1:1 isolation voltage transformer. When the voltage source and the load impedance are interchanged, as is shown in FIG. 28B, the GET is also a 1:1 isolation voltage transformer.

The GET that is described in Example 4, when wired as is shown in FIGs. 28A or 28B, meets all the first set of conditions. The following example, when wired as is shown in FIGs. 28A or 28B, meets all the second set of conditions.

Example 7 An example of a GET that satisfies the second set of conditions for the GETs is shown in FIGs. 28A and 28B. The GET in this example is to be operated at a frequency of v = 300 kHz and the length is 0.93 cm. The upper transmission line 123 and the lower transmission line 127 each have a single piece of dielectric material with E = 1.7E0 and thickness d# = 10 mils. Since there is no separate ferrite material d# = d" and p = µ0. Calculating Pl for the upper and lower transmission lines: 1l = 3l = 2#(3 x 105) (0.0093) [(4# x 10-7) (1.7) (8.85 x 10-12)]½

Pil = P31 = 7.6 x 10-5 (this is approximately zero).

The middle transmission line 125 has two pieces of dielectric material, each with E = 20,000#0 and thickness d# = 2 mils, and a single piece of ferrite material with µ = 2,900µ0 and thickness d, = 50 mils. Calculating Pl for the middle transmission line: P21 = 2#(3 x 105) (0.0093) [(2,900) (4# x 10-7) (20,000) (8.85 x 10-12) (50)/(2 + P21 = 1.57 (this is approximately #/2).

The outside transmission line uses the materials that are in the upper, middle, and lower transmission lines; therefore, the inductance per unit length is given by Los = L1 + L2 + L3 and the capacitance per unit length is given by 1/Cos = 1/C1 + 1/C2 + 1/C3. In this example Los = L2 and Cos = C1/2. This gives osl = 2##l (L2C1/2)½ = 2# (3 x 105) (0.0093) [(2,900) (4# x 10-7) (1.7) (8.85 x 10-12) (50)/2(10)]½ osl = 6.49 x 10-3 (this approaches zero).

This example has shown that this GET design meets all of the second set of conditions for P11, p2l, 3l, and osl that are necessary for the GETs in FIGs. 28A and 28B to function properly.

Five Transmission Line (5TL) GETs Figures 29A - 36D show additional embodiments of the GET design. While FIGs. 3A, 3B, and 23A - 28D are all GETs constructed of three transmission lines (3TL) in direct contact, the GET schematics in FIGs. 29A - 36D extend the concept to five transmission lines (5TL) in direct contact.

While the 3TL-GETs are mainly 1:1 isolation transformers (with a few 2:1 isolation transformers), the 5TL-GETs are all n:l isolation transformers with n > 2. Wherever two 3TL-GETs are to be wired together to obtain a higher turns ratio, a single 5TL-GET may be constructed which will perform the same functions yet with less material (making it lighter and less expensive to manufacture) and with higher efficiency (because there are fewer conductors there is a reduced I2R loss and because there is less ferrite material there is less "core" loss in the ferrite). 5TL-GETs themselves can be wired together in the manner shown in FIG. 4 (i.e. the inputs of the 5TL-GETs are wired in series and the outputs are wired in parallel) to obtain even higher turns ratios. The unique features of each 5TL-GET will be described along with the unique design constraints (i.e. the unique materials and geometries). While a specific example will not be given for any of the 5TL-GETs, their construction is similar to the construction of 3TL-GETs; therefore the construction of 5TL- GETs can be deduced from Examples 1 - 7. As with the previously described 3TL-GETs, the 5TL-GETs can be fabricated in any geometry that a transmission line can take on; e.g. parallel wire, stripline, coaxial, etc.

A 5TL-GET embodiment is shown in FIG. 29A. This GET is comprised of six conductors 400, 402, 404, 406, 408, and 410, that are separated by five dielectric/magnetic materials 412, 414, 416, 418, and 420, respectively. On one side of this GET are six terminals 422, 424, 426, 428, 430, and 432, and on the other side are six more terminals 434, 436, 438,

440, 442, and 444. This structure incorporates five transmission lines 446, 448, 450, 452, and 454 each of which has the respective characteristic impedance and phase factor Z and p1, Z2 and p2, Z3 and P3, Z4 and p4, and Z5 and p5. An additional transmission line, the outside transmission line, is formed as a consequence of the five transmission lines being in direct contact. The outside transmission line is comprised of two conductors 400 and 410 separated by the dielectric/magnetic materials 412, 414, 416, 418, and 420 with terminals 422 and 432 on one side and terminals 434 and 444 on the other side. The characteristic impedance and phase factor of the outside transmission line are Z05 and P,,. The lengths of the transmission lines are all 1. This structure forms the basic structure of the 5TL-GET. All the variations described in FIGs. 29A - 36D have this basic structure with the voltage source 456, load impedance 458, and shorting wires connected at different terminals for each figure. As such like numbers are used for like elements in these figures.

In the GET embodiment shown in FIG. 29A, the voltage source is connected to terminals 424 and 426, the load impedance is connected to terminals 434 and 444, and shorting wires are connected between terminals 422 and 432, 428 and 430, 436 and 438, 440 and 444, and 434 and 442. There are six transmission lines in the structure shown in FIG. 29A that are of importance: 446, 448, 450, 452, 454, and the outside transmission line (which is the top conductor 400 and the bottom conductor 410 separated by the dielectric/magnetic materials 446, 448, 450, 452, and 454 with input terminals at

434 and 444). The phase factors of these six transmission lines are 1, 2, 3, 4, 5, and os. The set of conditions for this GET to operate properly is: 1l = 0, #, 2#, 3#, ..., 2l can be either 2l # 0, #, 2#, 3#, ... or 2l = #/2, 3#/2, 5#/2, ..., 3l = 0, #, 2#, 3#, ..., 4l can be either 4l # 0, #, 2#, 3#, ..., or 4l = #/2, 3#/2, 5#/2, ..., 5l = 0, #, 2#, 3#, ..., and osl can be either osl # 0, #, 2#, 3#, ..., or osl = #/2, 3#/2, 5#/2, ..., Transmission lines 446, 450, and 454 must be in phase: 1l - P31 = 0, + 2#, + 4#, ... and P11 - P51 = 0, + 27r, + 47r, When all of these conditions are met this GET is a 2:1 isolation step-down voltage transformer.

In the GET embodiment shown in FIG. 29B, the voltage source is connected to terminals 424 and 426, the load impedance is connected to terminals 434 and 444, and shorting wires are connected between terminals 422 and 432, 436 and 438, 440 and 444, and 434 and 442.

The set of conditions for this GET to operate properly is: 1l = 0, #, 2#, 3#, ..., 2l can be either 2l # 0, #, 2#, 3#, ..., or 2l = #/2, 3#/2, 5#/2, ..., 3l # 0, #, 2#, 3#, ..., 4l = 0, #, 2#, 3#, ..., 5l = 0, #, 2#, 3#, ..., and osl can be either osl # 0, #, 2#, 3#, ..., or osl = #/2, 3#/2, 5#/2, ...

Transmission lines 446, 450, and 454 must be in phase:

1l - 3l = 0, # 2#, # 4#, ... and 1l - 5l = 0, # 2#, # 4#, Transmission line 452 must be out of phase with transmission lines 446, 450, and 454: P41 - P11 = + it, + 3#, + Sit ...

When all of these conditions are met this GET is a 3:1 isolation step-down voltage transformer.

In the GET embodiment shown in FIG. 29C, the voltage source is connected to terminals 424 and 426, the load impedance is connected to terminals 434 and 444, and shorting wires are connected between terminals 422 and 432, 440 and 444, and 434 and 442.

There are two sets of conditions for this GET to operate properly. The first set of conditions is: 1l = 0, #, 2#, 3#, ..., 2l = 0, #, 2#, 3#, ...

3l = 0, #, 2#, 3#, ..., 4l = 0, #, 2#, 3#, ..., 5l = 0, #, 2#, 3#, ..., and osl can be either osl # 0, #, 2#, 3#, ..., or osl = #/2, 3#/2, 5#/2, ...

Transmission lines 446, 450, 452, and 454 must be in phase: 1l - 3l = 0, # 2#, # 4#, ..., 1l - 4l = 0, # 2#, # 4#, ..., and 1l - 5l = 0, # 2#, # 4#, ...

Transmission line 448 must be out of phase with transmission lines 446, 450, 452, and 454: 2l - 1l = # #, # 3#, # 5# ...

When all of these conditions are met this GET is a 1.5:1 isolation step-down voltage transformer.

The second set of conditions for FIG. 29C is: 1l = 0, #, 2#, 3#, ..., 2l = 0, #, 2#, 3#, ...

3l = 0, #, 2#, 3#, ..., 4l = 0, #, 2#, 3#, ..., 5l = 0, #, 2#, 3#, ..., and osl can be either osl # 0, #, 2#, 3#, ..., or osl = #/2, 3#/2, 5#/2, ...

Transmission lines 446, 450, and 454 must be in phase: 1l - 3l = 0, # 2#, # 4#, ... and 1l - 5l = 0, # 2#, # 4#, Transmission lines 448 and 452 must be in phase: 2l - 4l = 0, # 2#, # 4#, Transmission lines 448 and 452 must be out of phase with transmission lines 446, 450, and 454: 2l - 1l = # #, # 3#, # 5# ...

When all of these conditions are met this GET is a 2:1 isolation step-up voltage transformer.

In the GET embodiment shown in FIG. 30A, the voltage source is connected to terminals 428 and 430, the load impedance is connected to terminals 434 and 444, and shorting wires are connected between terminals 422 and 432, 424 and 426, 440 and 442, 436 and 444, and 434 and 438.

The set of conditions for this GET to operate properly is: 1l = 0, #, 2#, 3#, ..., 2l can be either 2l # 0, #, 2#, 3#, ..., or 2l = #/2, 3#/2, 5#/2, ..., 3l = 0, #, 2#, 3#, ..., 4l can be either 4l # 0, #, 2#, 3#, ... or 4l = #/2, 3#/2, 5#/2, ..., 5l = 0, #, 2#, 3#, ..., and osl can be either osl # 0, #, 2#, 3#, ..., or osl = #/2, 3#/2, 5#/2, ...

Transmission liies 446, 450, and 454 must be in phase:

1l - 3l = 0, # 2#, # 4#, ... and 1l - 5l = 0, # 2#, # 4#, When all of these conditions are met this GET is a 2:1 isolation step-down voltage transformer.

In the GET embodiment shown in FIG. 30B, the voltage source is connected to terminals 428 and 430, the load impedance is connected to terminals 434 and 444, and shorting wires are connected between terminals 422 and 432, 440 and 442, 436 and 444, and 434 and 438.

The set of conditions for this GET to operate properly is: 1l = 0, #, 2#, 3#, ..., 2l = 0, #, 2#, 3#, ..., 3l = 0, #, 2#, 3#, ..., 4l can be either 4l # 0, #, 2#, 3#, ..., or 4l = #/2, 3#/2, 5#/2, ..., 5l = 0, #, 2#, 3#, ..., and osl can be either osl # 0, #, 2#, 3#, ..., or osl = #/2, 3#/2, 5#/2, ...

Transmission lines 446, 450, and 454 must be in phase: 1l - 3l = 0, # 2#, # 4#, ... and 1l - 5l = 0, # 2#, # 4#, Transmission line 448 must be out of phase with transmission lines 446, 450, anc 454: 2l - P11 = + it + 3#, + Sit ...

When all of these conditions are met this GET is a 3:1 isolation step-down voltage transformer.

In the GET embodiment shown in FIG. 30C, the voltage source is connected to terminals 428 and 430, the load impedance is connected to terminals 434 and 444, and shorting

wires are connected between terminals 422 and 432, 436 and 444, and 434 and 438.

There are two sets of conditions for this GET to operate properly. The first set of conditions is: 1l = 0, #, 2#, 3#, ..., 2l = 0, #, 2#, 3#, ...

3l = 0, #, 2#, 3#, ..., 4l = 0, #, 2#, 3#, ..., 5l = 0, #, 2#, 3#, ..., and osl can be either osl # 0, #, 2#, 3#, ... or osl = #/2, 3#/2, 5#/2, ...

Transmission lines 446, 450, and 454 must be in phase: 1l - 3l = 0, # 2#, # 4#, ..., 1l - 5l = 0, # 2#, # 4#, Transmission lines 448 and 452 must be out of phase with transmission lines 446, 450, and 454: 2l - 1l = # #, # 3#, # 5# ...

When all of these conditions are met this GET is a 1.5:1 isolation step-down voltage transformer.

The second set of conditions for FIG. 30C is: 1l = 0, #, 2#, 3#, ..., 2l = 0, #, 2#, 3#, ...

3l = 0, #, 2#, 3#, ..., 4l = 0, #, 2#, 3#, ...

P51 = 0! it, 2#, 3#, ..., and osl can be either osl # 0, #, 2#, 3#, ..., or osl = #/2, 3#/2, 5#/2, ...

Transmission lines 446, 448, 450, and 454 must be in phase: 1l - 2l = 0, # 2#, # 4#, ..., 1l - 3l = 0, # 2#, # 4#, ...

1l - 5l = 0, # 2#, # 4#, ...

Transmission line 452 must be out of phase with transmission lines 446, 448, 450, and 454: P41 - P11 = + it, + 3#, + 5# When all of these conditions are met this GET is a 2:1 isolation step-up voltage transformer.

In the GET embodiment shown in FIG. 31A, the voltage source is connected to terminals 426 and 428, the load impedance is connected to terminals 434 and 444, and shorting wires are connected between terminals 422 and 432, 438 and 440, 436 and 444, and 434 and 442.

The set of conditions for this GET to operate properly is: 1l = 0, #, 2#, 3#, ..., 2l = 0, #, 2#, 3#, ..., 3l can be either 3l # 0, #, 2#, 3#, ... or 3l = #/2, 3#/2, 5#/2, ..., 4l = 0, #, 2#, 3#, ..., 5l = 0, #, 2#, 3#, ..., and osl can be either osl # 0, #, 2#, 3#, ..., or osl = #/2, 3#/2, 5#/2, ...

Transmission lines 446 and 454 must be in phase: 1l - 5l = 0, # 2#, # 4#, ...

Transmission lines 448 and 452 must be in phase: 2l - 4l = 0, # 2#, # 4#, Transmission lines 446 and 454 must be out of phase with transmission lines 448 and 452: P21 - P11 = + it, + 3#, + 5w ...

When all of these conditions are met this GET is a 3:1 isolation step-down voltage transformer.

In the GET embodiment shown in FIG. 31B, the voltage source is connected to terminals 426 and 428, the load impedance is connected to terminals 434 and 444, and shorting wires are connected between terminals 422 and 432, 436 and 444, and 434 and 442.

There are two sets of conditions for this GET to operate properly. The first set of conditions for FIG. 35B is: 1l = 0, #, 2#, 3#, ..., 2l = 0, #, 2#, 3#, ..., 3l = 0, #, 2#, 3#, ..., 4l = 0, #, 2#, 3#, ..., 5l = 0, #, 2#, 3#, ..., and osl can be either 5l # 0, #, 2#, 3#, ..., or osl = #/2, 3#/2, 5#/2, ...

Transmission lines 446, 450, and 454 must be in phase: 1l - 3l = 0, # 2#, # 4#, ..., 1l - 5l = 0, # 2#, # 4#, ...

Transmission lines 448 and 452 must be in phase: 2l - 4l = 0, # 2#, # 4#, ...

Transmission lines 446, 450, and 454 must be out of phase with transmission lines 448 and 452: 2l - P11 = + it + 3#, # l ...

When all of these conditions are met this GET is a 1.5:1 isolation step-down voltage transformer.

The scond set of conditions for FIG. 31B is: 1l = 0, #, 2#, 3#, ..., 2l = 0, #, 2#, 3#, ..., 3l = 0, #, 2#, 3#, ..., 4l = 0, #, 2#, 3#, ..., 5l = 0, #, 2#, 3#, ..., and osl can be either osl # 0, #, 2#, 3#, ..., or osl = #/2, 3#/2, 5#/2, ...

Transmission lines 446, 448, 452, and 454 must be in phase: 1l - 2l = 0, # 2#, # 4#, ..., 1l - 4l = 0, # 2#, # 4#, 1l - 5l = 0, # 2#, # 4#, ...

Transmission line 450 must be out of phase with transmission lines 446, 448, 452, and 454: 3l - 1l = # #, # 3#, # 5#

When all of these conditions are met this GET is a 2:1 isolation step-up voltage transformer.

In the GET embodiment shown in FIG. 32A, the voltage source is connected to terminals 422 and 432, the load impedance is connected to terminals 436 and 442, and shorting wires are connected between terminals 424 and 426, 428 and 430, 434 and 444, 436 and 440, and 438 and 442.

The set of conditions for this GET to operate properly is: 1l = 0, it, 2it, 3it, , . ., 2l can be either 2l # 0, #, 2#, 3#, ... or 2l = #/2, 3#/2, 5#/2, ..., 3l = 0, #, 2#, 3#, ..., 4l can be either 4l # 0, #, 2#, 3#, ... or 4l #/2, 3#/2, 5#/2, ..., 5l = 0, #, 2#, 3#, ..., and osl can be either osl # 0, #, 2#, 3#, ..., or osl = #/2, 3#/2, 5#/2, ...

Transmission lines 446, 450, and 454 must be in phase: 1l - P31 = 0, + 2#, + 4#, ... and P11 - P51 = 0, + 21r, + 4#, When all of these conditions are met this GET is a 2:1 isolation step-down voltage transformer.

In the GET embodiment shown in FIG. 32B, the voltage source is connected to terminals 422 and 432, the load impedance is connected to terminals 436 and 442, and shorting wires are connected between terminals 424 and 426, 434 and 444, 436 and 440, and 438 and 442.

The set of conditions for this GET to operate properly is: 1l = 0, #, 2#, 3#, ..., 2l can be either

2l # 0, #, 2#, 3#, ... or 2l = #/2, 3#/2, 5#/2, ..., 3l = 0, #, 2#, 3#, ..., 4l = 0, #, 2#, 3#, ..., 5l = 0, #, 2#, 3#, ..., and osl can be either osl # 0, #, 2#, 3#, ..., or osl = #/2, 3#/2, 5#/2, ...

Transmission lines 446, 450, and 454 must be in phase: 1l - 3l = 0, # 2#, # 4#, ... and 1l - 5l = 0, # 2#, # 4#, Transmission line 452 must be out of phase with transmission lines 446, 450, and 454: 4l - 1l = # #, # 3#, # 5# When all of these conditions are met this GET is a 3:1 isolation step-down voltage transformer.

In the GET embodiment shown in FIG. 32C, the voltage source is connected to terminals 422 and 432, the load impedance is connected to terminals 436 and 442, and shorting wires are connected between terminals 428 and 430, 434 and 444, 436 and 44C, and 438 and 442.

The set of conditions for this GET to operate properly is: 1l = 0, #, 2#, 3#, ..., 2l = 0, #, 2#, 3#, ..., 3l = 0, #, 2#, 3#, ..., 4l can be either 4l # 0, #, 2#, 3#, ... or 4l = #/2, 3#/2, 5#/2, ..., 5l = 0, #, 2#, 3#, ..., and osl can be either osl # 0, #, 2#, 3#, ..., or osl = #/2, 3#/2, 5#/2, ...

Transmission lines 446, 450, and 454 must be in phase: P11 - p31 = 0, + 2#, + 4#, ... and p11 - P51 = 0, + 2#, + 4#,

Transmission line 448 must be out of phase with transmission lines 446, 450, and 454: 2l - 1l = # #, # 3#, # 5# ...

When all of these conditions are met this GET is a 3:1 isolation step-down voltage transformer.

In the GET embodiment shown in FIG. 32D, the voltage source is connected to terminals 422 and 432, the load impedance is connected to terminals 436 and 442, and shorting wires are connected between terminals 434 and 444, 436 and 440, and 438 and 442.

There are three sets of conditions for this GET to operate properly. The first set of conditions is: 1l = 0, #, 2#, 3#, ..., 2l = 0, #, 2#, 3#, ..., 3l = 0, #, 2#, 3#, ..., 4l = 0, #, 2#, 3#, ..., 5l = 0, #, 2#, 3#, ..., and osl can be either osl # 0, #, 2#, 3#, ..., or osl = #/2, 3#/2, 5#/2, ...

Transmission lines 446, 450, and 454 must be in phase: 1l - 3l = 0, # 2#, # 4#, ... and 1l - 5l = 0, # 2#, # 4#, Transmission lines 448 and 452 must be in phase: 2l - 4l = 0, # 2#, # 4#, Transmission lines 448 and 452 must be out of phase with transmission lines 446, 450, and 454: P21 - Pil = + it, + 3#, + Sit ...

When all of these conditions are met this GET is a 4:1 isolation step-down voltage transformer.

The second set of conditions for FIG. 32D is: 1l = 0, #, 2#, 3#, ..., 2l = 0, #, 2#, 3#, ...,

3l = 0, #, 2#, 3#, ..., 4l = 0, #, 2#, 3#, ..., 5l = 0, #, 2#, 3#, ..., and osl can be either osl # 0, #, 2#, 3#, ..., or osl = #/2, 3#/2, 5#/2, ...

Transmission lines 446, 448, 452, and 454 must be in phase: 1l - 2l = 0, # 2#, # 4#, ... and 1l - 4l = 0, # 2#, # 4#, ...

1l - 5l = 0, # 2#, # 4#, ...

Transmission line 450 must be out of phase with transmission lines 446, 448, 452, and 454: 3l - 1l = # #, # 3#, # 5# When all of these conditions are met this GET is a 2:1 isolation step-down voltage transformer.

The third set of conditions for FIG. 32D is: 1l = 0, #, 2#, 3#, ..., 2l = 0, #, 2#, 3#, ..., 3l = 0, #, 2#, 3#, ..., 4l = 0, #, 2#, 3#, ..., 5l = 0, #, 2#, 3#, ..., and osl can be either osl # 0, #, 2#, 3#, ..., or osl = #/2, 3#/2, 5#/2, ...

Transmission lines 446 and 454 must be in phase: 1l - 5l = 0, # 2#, # 4#, ...

Transmission lines 448, 450, and 452 must be in phase: 2l - 3l = 0, # 2#, # 4#, ... and 2l - 4l = 0 , # 2#, # 4#, Transmission lines 446 and 454 must be out of phase with transmission lines 448, 450, 452: 2l - 1l = # #, # 3#, # 5# ...

When all of these conditions are met this GET is a 2:1 isolation step-down voltage transformer.

In the GET embodiment shown in FIG. 33A, the voltage source is connected to terminals 424 and 426, the load impedance is connected to terminals 422 and 432, and shorting wires are connected between terminals 434 and 444, 436 and 438, 440 and 442, 422 and 430, and 428 and 432.

The set of conditions for this GET to operate properly is: 1l = 0, #, 2#, 3#, ..., 2l can be either 2l # 0, #, 2#, 3#, ..., or 2l = #/2, 3#/2, 5#/2, ..., 3l = 0, #, 2#, 3#, ..., and 4l can be either 4l # 0, #, 2#, 3#, ..., or 4l = #/2, 3#/2, 5#/2, ..., 5l = 0, #, 2#, 3#, ..., and OSl can be either OSl # 0, #, 2#, 3#, ..., or 4l = #/2, 3#/2, 5#/2, ...

Transmission lines 446, 450, and 454 must be in phase: 1l - 3l = 0, # 2#, # 4#, ... and 1l - 5l = 0, # 2#, # 4#, When all of these conditions are met this GET is a 2:1 isolation step-down voltage transformer.

In the GET embodiment shown in FIG. 33B, the voltage source is connected to terminals 424 and 426, the load impedance is connected to terminals 422 and 432, and shorting wires are connected between terminals 434 and 444, 436 and 438, 422 and 430, and 428 and 432.

The set of conditions for this GET to operate properly is: 1l = 0, #, 2#, 3#, ..., 2l can be either 2l # 0, #, 2#, 3#, ..., or 2l = #/2, 3#/2, 5#/2, ..., 3l = 0, #, 2#, 3#, ..., 4l = 0, #, 2#, 3#, ...,

5l = 0, #, 2#, 3#, ..., and OSl can be either OSl # 0, #, 2#, 3#, ..., or 4l = #/2, 3#/2, 5#/2, ...

Transmission lines 446, 450, and 454 must be in phase: 1l - 3l = 0, # 2#, # 4#, ... and 1l - 5l = 0, # 2#, # 4#, Transmission line 452 must be out of phase with transmission lines 446, 450, 454: 4l - 1l = # #, # 3#, # 5# When all of these conditions are met this GET is a 3:1 isolation step-down voltage transformer.

In the GET embodiment shown in FIG. 33C, the voltage source is connected to terminals 424 and 426, the load impedance is connected to terminals 422 and 432, and shorting wires are connected between terminals 434 and 444, 422 and 430, and 428 and 432.

The set of conditions for this GET to operate properly is: 1l = 0, #, 2#, 3#, ..., 2l = 0, #, 2#, 3#, ..., 3l = 0, #, 2#, 3#, ..., 4l = 0, #, 2#, 3#, ..., 5l = 0, #, 2#, 3#, ..., and OSl can be either OSl # 0, #, 2#, 3#, ..., or 4l = #/2, 3#/2, 5#/2, ...

Transmission lines 446, 450, and 454 must be in phase: 1l - 3l = 0, # 2#, # 4#, ... and 1l - 5l = 0, # 2#, # 4#, Transmission lines 448 and 452 must be in phase: 2l - 4l = 0, # 2#, # 4#, ...

Transmission lines 448 and 452 must be out of phase with transmission lines 446, 450, 454:

P21 - p11 = + it, + 3#, + Sit ...

When all of these conditions are met this GET is a 1.5:1 isolation step-down voltage transformer.

In the GET embodiment shown in FIG. 34A, the voltage source is connected to terminals 428 and 430, the load impedance is connected to terminals 422 and 432, and shorting wires are connected between terminals 434 and 444, 436 and 438, 440 and 442, 422 and 426, and 424 and 432.

The set of conditions for this GET to operate properly is: 1l = 0, #, 2#, 3#, ..., 2l can be either 2l # 0, #, 2#, 3#, ..., or 2l = #/2, 3#/2, 5#/2, ..., 3l = 0, #, 2#, 3#, ..., and 4l can be either 4l # 0, #, 2#, 3#, ..., or 4l = #/2, 3#/2, 5#/2, ..., 5l = 0, #, 2#, 3#, ..., and osl can be either osl # 0, #, 2#, 3#, ..., or osl = #/2, 3#/2, 5#/2, ...

Transmission lines 446, 450, and 454 must be in phase: 1l - 3l = 0, # 2#, # 4#, ... and 1l - 5l = 0, # 2#, # 4#, When all of these conditions are met this GET is a 2:1 isolation step-dow voltage transformer.

In the GET embodiment shown in FIG. 34B, the voltage source is connected to terminals 428 and 430, the load impedance is connected to terminals 422 and 432, and shorting wires are connected between terminals 434 and 444, 440 and 442, 422 and 426, and 424 and 432.

The set of conditions for this GET to operate properly is:

1l = 0, #, 2#, 3#, ..., 2l = 0, #, 2#, 3#, ..., 3l = 0, #, 2#, 3#, ..., and 4l can be either 4l # 0, #, 2#, 3#, ..., or 4l = #/2, 3#/2, 5#/2, ..., 5l = 0, #, 2#, 3#, ..., and osl can be either osl # 0, #, 2#, 3#, ..., or osl = #/2, 3#/2, 5#/2, ...

Transmission lines 446, 450, and 454 must be in phase: 1l - 3l = 0, # 2#, # 4#, ... and 1l - 5l = 0, # 2#, # 4#, Transmission line 448 must be out of phase with transmission lines 446, 450, 454: 2l - 1l = # #, # 3#, # 5# ...

When all of these conditions are met this GET is a 3:1 isolation step-down voltage transformer.

In the GET embodiment shown in FIG. 34C, the voltage source is connected to terminals 428 and 430, the load impedance is connected to terminals 422 and 432, and shorting wires are connected between terminals 434 and 444, 422 and 426, and 424 and 432.

The set of conditions for this GET to operate properly is: 1l = 0, #, 2#, 3#, ..., 2l = 0, #, 2#, 3#, ..., 3l = 0, #, 2#, 3#, ..., 4l = 0, #, 2#, 3#, ..., p51 = 0, it, 2#, 3#, . . ..., and osl can be either osl # 0, #, 2#, 3#, ..., or osl = #/2, 3#/2, 5#/2, ...

Transmission lines 446, 450, and 454 must be in phase: 1l - 3l = 0, # 2#, # 4#, ... and 1l - 5l = 0, # 2#, # 4#, Transmission lines 448 and 452 must be in phase:

2l - 4l = 0, # 2#, # 4#, Transmission lines 448 and 452 must be out of phase with transmission lines 446, 450, 454: 2l - 1l = # #, # 3#, # 5# ...

When all of these conditions are met this GET is a 1.5:1 isolation step-down voltage transformer.

In the GET embodiment shown in FIG. 35A, the voltage source is connected to terminals 426 and 428, the load impedance is connected to terminals 422 and 432, and shorting wires are connected between terminals 434 and 444, 438 and 440, 422 and 430, and 424 and 432.

The set of conditions for this GET to operate properly is: 1l = 0, #, 2#, 3#, ..., 2l = 0, #, 2#, 3#, ..., 3l can be either 3l # 0, #, 2#, 3#, ... or 3l = #/2, 3#/2, 5#/2, ..., 4l = 0, #, 2#, 3#, ..., 5l = 0, #, 2#, 3#, ..., and osl can be either osl # 0, #, 2#, 3#, ..., or osl = #/2, 3#/2, 5#/2, ...

Transmission lines 446 and 454 must be in phase: 1l - 5l = 0, # 2#, # 4#, ...

Transmission lines 448 and 452 must be in phase: 2l - 4l = 0, # 2#, # 4#, ..

Transmission lines 448 and 452 must be out of phase with transmission lines 446 and 454: 2l - 1l = # #, # 3#, # 5# ...

When all of these conditions are met this GET is a 3:1 isolation step-down voltage transformer.

In the GET embodiment shown in FIG. 35B, the voltage source is connected to terminals 426 and 428, the load impedance is connected to terminals 422 and 432, and shorting wires are connected between terminals 434 and 444, 422 and 430, and 424 and 432.

There are two sets of conditions for this GET to operate properly. The first set of conditions is: 1l = 0, #, 2#, 3#, ..., 2l = 0, #, 2#, 3#, ..., 3l = 0, #, 2#, 3#, ..., 4l = 0, #, 2#, 3#, ..., 5l = 0, #, 2#, 3#, ..., and osl can be either osl # 0, #, 2#, 3#, ..., or osl = #/2, 3#/2, 5#/2, ...

Transmission lines 446, 450, and 454 must be in phase: 1l - 3l = 0, # 2#, # 4#, ..., 1l - 5l = 0, # 2#, # 4#, ...

Transmission lines 448 and 452 must be in phase: 2l - 4l = 0, # 2#, # 4#, Transmission lines 448 and 452 must be out of phase with transmission lines 446, 450, and 454: 2l - 1l = # #, # 3#, # 5# ...

When all of these conditions are met this GET is a 1.5:1 isolation step-down voltage transformer.

The second set of conditions for FIG. 35B is: 1l = 0, #, 2#, 3#, ..., 2l = 0, #, 2#, 3#, ..., 3l = 0, #, 2#, 3#, ..., 4l = 0, #, 2#, 3#, ..., 5l = 0, #, 2#, 3#, ..., and osl can be either osl # 0, #, 2#, 3#, ..., or osl = #/2, 3#/2, 5#/2, ...

Transmission lines 446, 448, 452, and 454 must be in phase: 1l - 2l = 0, # 2#, # 4#, ..., 1l - 4l = 0, # 2#, # 4#, ..., and 1l - 5l = 0, # 2#, # 4#, ...

Transmission line 450 must be out of phase with transmission lines 446, 448, 452, and 454: 3l - 1l = # #, # 3#, # 5# ...

When all of these conditions are met this GET is a 2:1 isolation step-up voltage transformer.

In the GET embodiment shown in FIG. 35C, the voltage source is connected to terminals 424 and 430, the load impedance is connected to terminals 422 and 432, and shorting wires are connected between terminals 434 and 444, 422 and 428, and 426 and 432.

The first set of conditions is: 1l = 0, #, 2#, 3#, ..., 2l = 0, #, 2#, 3#, ..., 3l = 0, #, 2#, 3#, ..., 4l = 0, #, 2#, 3#, ..., 5l = 0, #, 2#, 3#, ..., and OSl can be either OSl # 0, #, 2#, 3#, ..., or 4l = #/2, 3#/2, 5#/2, ...

Transmission lines 446, 450, and 454 must be in phase: 1l - 3l = 0, # 2#, # 4#, ..., #1l - #5l = 0, # 2#, # 4#, ...

Transmission lines 448 and 452 must be in phase: #2 - #4l = 0, # 2#, # 4#, ..

Transmission lines 448 and 452 must be out of phase with transmission lines 446, 450, and 454: 2l - 1l = # #, # 3#, # 5# ...

When all of these conditions are met this GET is a 2:1 isolation step-up voltage transformer.

The second set of conditions for FIG. 35C is: 1l = 0, #, 2#, 3#, ..., 2l = 0, #, 2#, 3#, ..., 3l = 0, #, 2#, 3#, ..., 4l = 0, #, 2#, 3#, ..., 5l = 0, #, 2#, 3#, ..., and OSl can be either

OSl # 0, #, 2#, 3#, ..., or 4l = #/2, 3#/2, 5#/2, ...

Transmission lines 446 and 454 must be in phase: 1l - 5l = 0, # 2#, # 4#, ...

Transmission lines 448, 450, and 452 must be in phase: 2l - 3l = 0, # 2#, # 4#, ..., 2l - 4l = 0, # 2#, # 4#, .

Transmission lines 446 and 454 must be out of phase with transmission lines 448, 450, and 452: P21 - p11 = + it + 3#, + Sit ...

When all of these conditions are met this GET is a 2:1 isolation step-up voltage transformer.

In the GET embodiment shown in FIG. 36A, the voltage source is connected to terminals 422 and 432, the load impedance is connected to terminals 424 and 430, and shorting wires are connected between terminals 434 and 444, 436 and 438, 440 and 442, 424 and 428, and 426 and 430.

The set of conditions for this GET to operate properly is: 1l = 0, #, 2#, 3#, ..., 2l can be either 2l # 0, #, 2#, 3#, ... or 2l = #/2, 3#/2, 5#/2, ..., 3l = 0, #, 2#, 3#, ..., 4l can be either 4l # 0, #, 2#, 3#, ... or 2l = #/2, 3#/2, 5#/2, ..., 5l = 0, #, 2#, 3#, ..., and OSl can be either OSl # 0, #, 2#, 3#, ... or 2l = #/2, 3#/2, 5#/2, ...

Transmission lines 446, 450, and 454 must be in phase: 1l - P31 = 0, + 2#, + 4#, ... and p11 - p51 = 0, + 2#, + 4#, When all of these conditions are met this GET is a 2:1 isolation step-down voltage transformer.

In the GET embodiment shown in FIG. 36B, the voltage source is connected to terminals 422 and 432, the load impedance is connected to terminals 424 and 430, and shorting wires are connected between terminals 434 and 444, 436 and 438, 424 and 428, and 426 and 430.

The set of conditions for this GET to operate properly is: 1l = 0, #, 2#, 3#, ..., 2l can be either 2l # 0, #, 2#, 3#, ... or 2l = #/2, 3#/2, 5#/2, ..., 3l = 0, #, 2#, 3#, ..., 4l = #, #, 2#, 3#, ..., 5l = 0, #, 2#, 3#, ..., and OSl can be either OSl # 0, #, 2#, 3#, ... or 2l = #/2, 3#/2, 5#/2, ...

Transmission lines 446, 450, and 454 must be in phase: 1l - 3l = 0, # 2#, # 4#, ... and 1l - 5l = 0, # 2#, # 4#, Transmission line 452 must be out of phase with transmission lines 446, 450, and 454: P41 - P11 = + it + 3#, + 571. ...

When all of these conditions are met this GET is a 3:1 isolation step-down voltage transformer.

In the GET embodiment shown in FIG. 36C, the voltage source is connected to terminals 422 and 432, the load impedance is connected to terminals 424 and 430, and shorting wires are connected between terminals 434 and 444, 440 and 442, 424 and 428, and 426 and 430.

The set of conditions for this GET to operate properly is: 1l = 0, #, 2#, 3#, ..., 2l = 0, #, 2#, 3#, ...,

3l = 0, #, 2#, 3#, ..., 4l can be either 4l # 0, #, 2#, 3#, ... or 2l = #/2, 3#/2, 5#/2, ..., 5l = 0, #, 2#, 3#, ..., and OSl can be either OSl # 0, #, 2#, 3#, ... or 2l = #/2, 3#/2, 5#/2, ...

Transmission lines 446, 450, and 454 must be in phase: 1l - 3l = 0, # 2#, # 4#, ... and 1l - 5l = 0, # 2#, # 4#, Transmission line 448 must be out of phase with transmission lines 446, 450, and 454: 2l - 1l = # #, # 3#, # 5# ...

When all of these conditions are met this GET is a 3:1 isolation step-down voltage transformer.

In the GET embodiment shown in FIG. 36D, the voltage source is connected to terminals 422 and 432, the load impedance is connected to terminals 424 and 430, and shorting wires are connected between terminals 434 and 444, 424 and 428, and 426 and 430. The set of conditions for this GET to operate properly is: 1l = 0, #, 2#, 3#, ..., 2l = #, #, 2#, 3#, ..., 3l = 0, #, 2#, 3#, ..., 4l = #, #, 2#, 3#, ..., 5l = 0, #, 2#, 3#, ..., and OSl can be either OSl # 0, #, 2#, 3#, ... or 2l = #/2, 3#/2, 5#/2, ...

Transmission lines 446, 450, and 454 must be in phase: 1l - 3l = 0, # 2#, # 4#, ... and 1l - 5l = 0, # 2#, # 4#, Transmission lines 448 and 452 must be in phase: 2l - 4l = 0, # 2#, # 4#, ...

Transmission lines 448 and 452 must be out of phase with transmission lines 446, 450, and 454: P21 - Pil = + it, + 3it, + Sit ...

When all of these conditions are met this GET is a 4:1 isolation step-down voltage transformer.

The configurations for 3TL-GETs and 5TL-GETs have been described in detail here. These concepts can be extended to 7TL-GETs, 9TL-GETs, l1TL-GETs, ... As the number of transmission lines in the GET increases there is an increase in the turns ratio, an increase in the efficiency of the GET, and there is often a decrease in the size and weight of the GET (as compared to the size and weight of a group of GETs, with fewer tranmission lines, that are wired together to produce a given turns ratio).

GET Circuits For the most part, GETs are used in circuits. GETs perform the important functions of stepping voltages up or down and/or electrically isolating one part of a circuit from another part. Some of the more important applications include AC-AC converters, AC-DC converters, DC-AC converters, and DC- DC converters. Often, the AC voltage used is the common 60 Hz wall plug voltaye. Since this frequency is too low for practical GET designs, the circuit in which the GET operates must supply a much higher frequency (approximately 10 kHz to as high as the 10 MHz range) to the input of the GET. Then, at the output of the GET, the high frequency is either used as is or is converted to something else; e.g. 60 Hz or DC. The

following are descriptions of several GET circuits which do all of these things.

Circuits for the Conversion of DC and Low Frequency to High Frequency The circuit represented in FIG. 37 converts a voltage with a frequency that is too low for proper GET operation to a frequency that matches one of the frequency ranges that allows the GET to operate. Some examples of low frequency include 50 Hz and 60 Hz used in homes and businesses around the world and 400 Hz used in aircraft. In practice, any frequency that is too low for proper GET operation can be converted to a proper frequency. The circuit has an input voltage source 302 coupled to a rectifier 304 which is in turn coupled to a modulator 306. The modulator 306 modulates the voltage signal with the output of an oscillator 308. The modulator 306 is coupled to a GET 310.

If the input voltage 302, is a sine wave with a frequency of 60 Hz as shown in the top waveform in FIG. 383, after passing through the full wave rectifier 304, the voltage will look like the middle waveform in FIG. 38. Although it is not shown in the circuit, a filter can be added after the rectifier 304 to smooth the ripples and produce a DC voltage if desired. If the original voltage source 302 is DC, then it would be directly input into the modulator 306 and the rectifier 304 would be unnecessary. The voltage V1, must then be converted to a higher frequency by the modulator 306. FIG.

39 represents am H-bridge circuit 320 which may be used for

the modulator 306. The H-bridge circuit 320 is based on four MOSFETs 322, 324, 326 and 328 which act as switches. Zener diodes 330, 332, 334 and 336 are shown connected between the source and drain of the MOSFETs 322, 324, 326 and 328 and are intrinsic to the MOSFET. To compensate for the effect of the intrinsic Zener diodes 330, 332, 334, 336 diodes 338, 340, 342 and 344 are each connected to the drain of each of the MOSFETs 322, 324, 326, 328. The modulator 306 chops the input voltage V1, at a frequency set by the oscillator 308 to produce a voltage having the bottom waveform in FIG. 38. The actual chopping frequency is set to a value needed by the GET 310 and would usually be much higher than is shown in the bottom waveform. This voltage then passes through the GET 310. The voltage output from the GET Volt, also looks like the waveform except that the voltage may have been stepped up or stepped down. This voltage can be used as is, or it can be easily converted to DC by passing it through a full wave rectifier and filters.

The converter circuit is useable for many applications. There are other situations, however, which require that the oscillator frequency be varied. An example of this would be if the operating temperature of the GET is varied to such an extent that the operating frequency of the GET also varies. A circuit with a phase lock loop in it may be used to adjust the frequency of the oscillator 308 so that it is always matched to the operating frequency of the GET 310. FIG. 40 shows an example of a phase lock loop circuit which may be added to the converter circuit in FIG. 37. A

phase detector 352 takes an input from the GET 310 and passes its output to a low pass filter 354. A phase control pot 356 and the low pass filter 354 are coupled to the inputs of an operational amplifier 358.

A wire can be looped around the short circuit of the first stage of the GET 310 or, if there is only a single stage, the wire is similarly looped around its short circuit This wire measures the small AC current which passes through the short circuit and provides a signal to the phase detector 352. Since the phase difference between this current and the input current is known (e.g. if Pl = it/2 for the center transmission line of this GET then the phase must be approximately 90 degrees) the "phase control" pot, 356 is adjusted to set the phase, then the circuit will stay locked- on to this phase.

Circuits for the Reconstruction of Low Frequency or DC at the Output If the original low frequency is needed at the output, then a demodulator can be used after the GET 310.

FIG. 41 shows the circuit in FIG. 40 modified to add a demodulator 350 and a phase delay circuit 348. The demodulator 350 is positioned after the GET 310. The demodulator 350 is very similar to the modulator 306 and reverses the operation of the modulator 306. The demodulator 350 works at the same frequency that the modulator 306 operates at, but its phase is different. Since the phase shift introduced by the GET 310 is known, an equivalent phase

delay can be introduced in the signal from the voltage controlled oscillator 308 so that the demodulator 350 operates at the proper phase To maintain the electrical isolation of the output of the GET 310 from its input, a circuit to electrically isolate the oscillator signal going to the demodulator 350 must be incorporated. This may be accomplished with optical coupling somewhere along the path between the voltage controlled oscillator 308 and the demodulator 350. Alternatively, a second phase lock loop circuit can be dedicated to the demodulator with 350 all of its power and ground connections coming from the output side of the GET 310 so that electrical isolation is maintained.

FIG. 42 represents the output of the demodulator. The voltage spikes are due to the imperfect reconstruction of the rectified 60 Hz voltage.

Instead of the demodulator 350 there are two other methods to reconstruct the 60 Hz voltage. The first method is to pass the output of the GET 310 through a full wave rectifier to produce a waveform that is similar to that in FIG. 42. Then the voltage is passed through a polarity switching circuit to convert the rectified sine wave shown in FIG. 42 to a sine wave. The polarity switching circuit is shown in FIG. 43. There are four triacs 352, 354, 356 and 358 that are switched by a polarity and timing control 360. The polarity and timing control 360 receives an input from the original 60 Hz source and when the input voltage changes polarity; i.e. whenever the voltage goes from positive to negative or from negative to positive, a signal is produced

which toggles the triacs 352, 354, 356 and 358. A phase adjustment is made to this signal to compensate for the phase change produced by passing through the GET 310. Since there is an input from the original 60 Hz voltage source, there must be a means of electrically isolating the triacs so that there will be electrical isolation between the input and output of the GET circuit. Here, the isolation is achieved with an optical coupler 362. The result of the polarity switching circuit is that every other "hump" in FIG. 42 switches from positive to negative and a 60 Hz sine wave is produced as shown in FIG. 45.

The second method is to use a baseline restorer circuit such as is shown in FIG. 44, connected directly to the output of the GET 310 to produce a waveform that is like that shown in FIG. 46. The polarity and timing control 360 is similar to that used in the circuit in FIG. 43. Thus, the output voltage from the GET 310 is converted to the waveform shown in FIG. 46.

In the baseline restorer circuit shown in FIG. 44, one of the GET outputs is referenced to ground, while the other has bipolar voltage excursions, at the chopping frequency, following the original 60 Hz input waveform as is shown in the bottom figure of FIG. 38. The capacitor 364, on alternate half cycles of the original 60 Hz waveform, is connected respectively to the grounded cathode diode 366 and the grounded anode diode 368. During the negative half cycle of the original 60 Hz waveform the capacitor 364 is connected to the grounded anode diode 368. The polarity selecting

switch 370 may be simply a pair of optically coupled triacs connected in the single pole, double throw configuration shown in FIG. 43. Although the output of the baseline restorer shown in FIG. 46 has twice the peak to peak voltage of the input voltage shown in the bottom figure of FIG. 38, the peak to peak voltages will become equivalent after the output of the baseline restorer is passed through a low pass filter to completely restore a "pure" 60 Hz waveform. It is worth noting that the output of the baseline restorer circuit, shown in FIG. 46, may be phase synchronized or may be 180 degrees out of phase with the original 60 Hz waveform. Control circuitry can be provided to ensure one phase relationship or the other, or, for that matter, any phase relationship in between.

If needed, the final waveforms represented in FIGs.

42 and 45 may be "cleaned up" with a low pass filter. These filters are common and will not be discussed here.

It will be understood that modifications and variations may be effected without departing from the spirit and scope of the novel concepts of this invention.