WO1990013939A1 - Power conversion system - Google Patents

Abstract

A power conversion system uses flat matrix transformer technology for distributed physical placement and improved heat transfer. A picture frame matrix transformer includes a plurality of interdependent magnetic elements (101a to 101f) arranged end-to-end in a closed pattern configuration and interwired as a matrix transformer having at least one primary (102), and at least one secondary (103a to 103f) winding, wherein the end of one secondary winding(103b) begins at one end to-end position between adjacent magnetic elements (101a, 101b), and ends at another end-to-end position between adjacent magnetic elements (101b, 101c). Rectifiers (204a to 204f) are connected to the secondary windings (103a to 103f) for providing a DC output voltage.

Description

POWER CONVERSION SYSTEM

BACKGROUND OF THE INVENTION The present invention relates generally to power supplies and deals more particularly with a power conversion system employing flat matrix transformer and inductor technology.

Conventional power conversion systems, particularly power converters, suffer from a number of drawbacks and limitations which hinder the miniaturization of the power converter. Such limitations are generally associated with the transformers used in the power converters and include high leakage inductance, high profiles, and hot spot temperature at the center of the transformer cord. The problem of leakage inductance is further aggravated when fractional turns are used to fine tune the voltage outputs. Further, fine tuning a transformer with fractional turns is uneconomical in mass production. In addition, known power systems generally use a single turn secondary winding to provide a low voltage output, such as for example, 5 volts, which does not permit an accurate turns ratio to provide voltages other than 5 volts. However, it does permit the transformer to have a minimum number of turns for a given turns ratio. Such transformers usually require considerable duration of the flux density to reduce core loss at high operating frequencies. Additionally, the transformer, for a given output current, must have a window large enough to accommodate reasonably low loss windings to be efficient which greatly adds to the size of the transformer.

The above limitations and others are generally overcome with the power conversion system of the present invention which uses a flat matrix transformer technology described in U. S. Patent 4,665,357, issued May 12, 1987; entitled "Flat Matrix Transformer; U. S. Patent H. 4,845,606, issued July 4, 1989 entitled "High Frequency Matrix Transformer", and U. S. Patent Application Serial Number 351,944, filed May 12, 1989 entitled "Transformer Having Symmetrical Push-Pull Windings" and each assigned to the same assignee as the present invention. The matrix transformer comprises an array or matrix of interdependent magnetic elements interconnected by at least two windings, each made up of at least one current carrying conductor and interacting by magnetic induction with the magnetic element and with all other current carrying conductor paths which pass through the same magnetic element. The whole of the interdependent, interconnected magnetic elements cooperate to function as a transformer. Due to the unique characteristics of the flat matrix transformer, the magnetic elements tend to have single turn windings and are relatively small and distributed thereby providing a high ratio of surface area to volume which provides excellent thermal characteristics by spreading the heat dissipation over the surface. The matrix transformer has high current capability, can be very flat and almost planar, and can be built using printed circuit board techniques. A matrix transformer can also insure current sharing between parallel power sources, and/or between parallel loads. In addition, leakage inductance is minimized and the single hot spot of a conventional transformer is eliminated since heat transfer is shared by all the magnetic elements.

The power conversion system of the present invention also permits precision control of the voltage output through use of cascaded matrix transformers and binary matrix transformers which permit control of the turns ratio.

The power conversion system of the present invention further includes interstage power conditioning through pre- and post-regulation and automatic current limiting.

The power conversion system of the present invention also includes a voltage distribution scheme which improves heat disipation by distributing voltage on a bus system in a chassis which provides contact along the edge of a printed circuit card carrying the components of the power convertor and includes a locking edge wedge to promote thermal conductivity from the printed circuit board edge to the chassis.

SUMMARY OF THE INVENTION

In accordance with the present invention, a power conversion system includes a buck converter for preconditioning and providing pre-regulation of a DC voltage.

The pre-regulated DC voltage is then switched as a 100% duty cycle square wave in one embodiment with a symmetrically divided push-pull winding matrix transformer. The voltage produced across the secondary of the symmetrical push-pull matrix transformer may be used directly or may be used to drive additional circuitry of the power converter.

In accordance with a further aspect of the invention uses a novel physical arrangement of the matrix transformer or a matrix inductor having a plurality of elements (or other equivalent structure) placed end-to-end in a closed pattern wherein the windings of the transformer or inductor respectively leave and enter the structure at various points around the closed pattern.

In accordance with a further aspect of the invention, transformers and inductors well adapted to distributed physical placement for improved heat transfer, and in particular, in which the transformer and inductor and associated power semiconductors are distributed and peripherally located for improved heat transfer to the surrounding ambient.

In accordance with a further aspect of the invention, the power converter includes a binary matrix transformer which is controllable to provide a variable ratio to produce a variable voltage output which output may be used directly or to drive another binary matrix transformer or other matrix transformer to produce the desired voltage outputs.

In accordance with another further aspect of the invention, an interstage regulator may be used to condition the power as it is transferred between one matrix transformer and the subsequent stages which may include cascaded matrix transformers and which cascaded transformers may include post regulation.

In a yet further aspect of the present invention, an auxiliary output winding may pass through a small number of magnetic elements to produce a desired output voltage other than the power converter voltage and which auxiliary output causes the transformer to be conditionally valid wherein the current in the auxiliary winding will limit at a constant value to provide automatic current limiting.

In a still further aspect of the invention, power busses extend down the sides of a chassis to distribute the high current, low voltage power produced by the power converter and to sink heat produced by the power converter components mounted on the printed circuit board. Heat sinking is provided along the edge of the printed circuit board using a card guide and locking wedge.

BRIEF DESCRIPTION OF THE DRAWINGS

Other features and advantages of the present invention will be readily apparent from the following written description and the drawings wherein:

Figure 1 is a schematic functional block diagram of one embodiment of the power conversion system of the present invention.

Figure 2 shows a picture frame matrix transformer which may be used with the power conversion system of Figure 1 wherein the transformer has a symmetrical push-pull primary winding and push-pull secondary windings with push-pull full wave rectifiers connected to the secondary windings.

Figure 3 shows a picture frame matrix transformer with full wave bridge rectifiers connected to the secondary windings. Figures 4 and 5 show round picture frame matrix transformers.

Figure 6 is a cross-sectional view taken along the line 6-6 of Figure 5.

Figure 7 is a schematic of a picture frame matrix transformer having a non-integer turns ratio, 1 to 3.14.

Figure 8 is a pictoral diagram of the matrix transformer of Figure 7.

Figure 9 shows an "open diagram" of a picture frame matrix transformer using a modified pot core, with rectifiers and switching FET's.

Figure 10 is a cross-sectional view taken along the line 10-10 of Figure 9.

Figure 11 is a schematic of a picture frame matrix transformer having a symmetrical push-pull primary and secondary windings -in which each secondary segment is a push- pull winding which couples two elements on each side.

Figure 12A shows the parts for a coaxial secondary picture frame matrix transformer element assembly. Figure 12B shows a progressively assembled assemblyof Figure 12A, and Figure 12C shows a portion of a picture frame matrix transformer using the assembly of Figure 12B.

Figure 12D is a partial schematic of the assembly of Figure 12C wherein the restifiers are shown directly connected to the secondary winding leads for heat sinking and to minimize lead inductance.

Figure 13 shows a partial schematic of a transformer using the assembly of Figures 12A-C.

Figure 14 shows the partial transformer schematic of Figure 13 with a coaxial matrix inductor and capacitor forming an L-C filtered output.

Figure 15 shows a picture frame matrix transformer with a picture frame matrix inductor with its switching FET's and rectifiers assembled on a printed circuit card which could be used with the power conversion system of the present invention.

Figure 16 is a representative schematic diagram of the matrix transformer and matrix inductor of Figure 15.

Figure 17 is a partial schematic diagram of a conditionally valid matrix transformer which may be used with the power conversion system of the present invention wherein an additional voltage may be derived from several elements of the matrix transformer. Figure 18 is a partial schematic diagram of cascaded matrix transformers used to provide ratio multiplication of voltage and current.

Figure 19 is a schematic diagram illustrating a matrix transformer having an interstage winding driving two cascaded matrix transformers wherein a power conditioning device is located in the interstage winding.

Figure 20 is a schematic diagram of a voltage regulation device that may be used with the interstage winding of Figure 19.

Figure 21 is a schematic diagram of a switch mode voltage regulation device that may be used with the interstage winding of Figure 19.

Figure 22 is a schematic diagram of two variable ratio matrix transformers connected with an interstage matrix filter inductor.

Figure 23 is a schematic representation of a cascaded matrix transformer wherein the interstage windings are shown as separate windings.

Figure 24 illustrates one interstage winding of the cascaded matrix transformer of Figure 23.

Figure 25 is a fragmentary view of the cascaded matrix transformer of Figure 23 illustrating one output winding.

Figure 26 is a partial view of a flat matrix transformer power converter mounted on a circuit board with heat sinking castings.

Figure 27 is a perspective schematic view of a chassis for distributing power and heat sinking for the power conversion system of the present invention. Figure 28 is a fragmentary view showing the connection between a card edge and the chassis of Figure 27.

Figure 29 is a top plan view of a com operated spring biased wedging device to provide positive mechanical and thermal contact between a card edge and the chassis of Figure 27.

Figure 30 illustrates a simple binary matrix transformer.

Figure 31 illustrates another binary matrix transformer.

Figure 32 is a schematic representation of a generalized binary matrix transformer.

Figure 33 is a schematic representation of a binary matrix transformer having binary factors equal to zero.

Figure 34 is a schematic representation of a generalized binary M x 1 matrix transformer.

Figure 35 is a schematic representation of a generalized binary 1 x N matrix transformer.

Figure 36 is a schematic representation of a 3 x 4 binary matrix transformer.

DETAILED DESCRIPTION OF PREFERRED EMBODIMENTS

Turning now to the drawings and considering in further detail the power conversion system and its various components comprising the present invention, Figure 1 is a schematic functional block diagram of one general embodiment of the power conversion system and is generally designated 10. A DC voltage is connected to the input 12 of pre-regulation means contained within the lined box 14 to provide pre- regulation of the DC input voltage. The conditioned DC input voltage is outputted on line 16 and which DC output voltage is switched as a 100% duty cycle squarewave by the DC/AC switching means contained within the function block 18. Typically, the DC/AC switching means may comprise a symmetrical push-pull matrix transformer primary winding and associated controllers such as the one disclosed in the above- referenced co-pending patent application entitled Transformer Having Symmetrical Push-Pull Windings and which disclosure is incorporated herein by reference. The AC output on the lead 20 may be rectified to produce a DC voltage output or may act as the input for subsequent stages of the power converter. As illustrated in Figure l, the output 20 serves as the input to a matrix transformer contained within the function block 22 and which matrix transformer may be any of the type disclosed within the above-referenced patents and applications all of which are assigned to the same assignee as the present invention and all of which are incorporated herein by reference. The matrix transformer 22 may also be a picture matrix transformer or a binary matrix transformer to provide a controlled variable turns ratio and which picture frame and binary matrix transformers are described herein below.

The output 24 of the matrix transformer 22 may be rectified to provide a DC voltage output or the output 24 can be connected to the input for subsequent stages in the embodiment illustrated in Figure 1. The output 24 is provided with an interstage conditioning which may be either filtering or regulation as described below and which is contained within the function block 26. The conditioned voltage at the lead 28 may again be rectified to provide a DC voltage output or may serve as further voltage regulation to the input of a subsequent stage shown in Figure 1 as a matrix transformer within the function block 30. The output 32 of the matrix transformer 30 is shown rectified by the rectifier 34 to provide a DC output voltage at its output 36. The DC voltage at the output 36 is coupled to a power distribution means contained in the function block 38. As stated above, the output 24 may be directly rectified in which case the output

24 can be connected to the input of rectifier 34 to provide the D.C. output voltage as its output 36.

As further illustrated in Figure 1, matrix transformers may be cascaded to provide additional output voltages wherein the cascaded matrix transformer may be any of the type identified above. In Figure 1, the input to the cascaded matrix transformers is taken from the output 24 of the matrix transformer 22 and in the embodiment shown, serves as the input 40 to the cascaded matrix transformer contained within the function block 42 and also to a post-regulation means, shown within the function block 44, in series with the interstage winding and and with the input 46 of the cascaded matrix transformer shown within the function block 48. The output 50 of the cascaded transformer 42 is rectified to provide a DC output voltage as is the output 52 of the cascaded matrix transformer 48. Accordingly, Figure 1 illustrates an embodiment of a power converter operating at 100% duty cycle to produce 3 DC output voltages.

The pre-regulation circuit means shown within the function block 14 may be a time off-set matrix inductor buck converter. The operation of a buck converter is generally well known wherein a switching device pulses the input to an inductor and the output voltage is a function of the duty cycle times input voltage and reference may be made to UK patent application GB 2176951A for information on the principle of operation. A matrix inductor is disclosed and described herein below.

Considering now Figure 2, a basic picture frame matrix transformer is illustrated as having six cores løla-f, a primary winding 102 passing through and linking all of the cores løla-f and having a single series turn, and a secondary winding comprising six secondary winding sections 103a-f, each respective winding section linking one-for-one, one of the cores. Full wave rectifiers 204a-f and a load resistor 205, shown by way of example but not as a limitation are illustrated with the outputs of the full wave rectifiers 204a- g being connected in parallel with the proper polarity and connected to the load resistor 205. (For some applications it may be preferable to use a low-pass filter network with each of the full wave rectifiers 204a-f). The characteristics of matrix transformers insure that the current in each of the full wave rectifiers 204a-f will be equal and each rectifier current will equal the current in the primary_^winding 102

(neglecting magnetization currents). Preferably, long, slender cores løla-f as shown pictorially are used, and for good high frequency characteristics, it is preferred that the cores løla-f are placed close together end-to-end and the uncoupled wiring kept to a minimum.

Note in particular the interconnection of the secondary windings 103a-f to the full wave rectifiers 204a-f illustrating that the two ends of each secondary winding 103a- f are not brought to the same full wave rectifier. (Although bringing the two ends of a secondary winding to the same full wave rectifier would work schematically, the extra uncoupled wiring which would be needed to return along the length of the long, slender cores løla-f would be detrimental to high frequency operation). Instead, the two ends of wires exiting the transformer at each locationg where the elements (defined below) adjoin are connected to the same rectifier even though the wires are from different secondary windings. For example, the "end" 113 of a first secondary winding 103a and the

"start" 123 of a second secondary winding 103b are connected to the full wave rectifier 204b. This end of one and start of another wiring pattern is repeated around the transformer.

The end of one and start of another secondary winding wiring pattern is one of the distinctions between the matrix transformer of the present invention and prior art transformer circuits using multiple transformers with series primaries and parallel secondaries.

The nomenclature of matrix transformers is used with the picture frame matrix transformer of Figure 2 and through the specification and defines as an "element" a core with the secondary section and the segment of the primary winding 102 which couples it. In general usage, all of the secondary sections 103a-f of the matrix transformer are tied together in parallel with the proper polarity and doing so in the matrix transformer of Figure 2 produces an effective turns ratio of 1 to 6.

Because the windings of a matrix transformer are often single wires passing through elements, "turns" and "turns ratio" are misnomers, but they are accepted nomeclature in the art of transformers and its use is retained. Also "winding" is used generically to identify a conductive path, whatever its composition or physical embodiment, through one or more elements and includes wires, bus bars, tubes, coaxial sheaths, printed wiring board or flex print traces and the like, and can be any type of winding including, center-tapped, split, bi-filar, coaxial or multiple turn variations of those listed..

"Core" is also used generically to identify a structure or portion of a structure having a closed magnetic path. Included are toroids; sleeves; pot cores; cross cores or, depending on the context, one arm of a cross core; various lamination structures including, but not limited to U, U-I, D-

U, E-E, E-I, F-F, C or L laminations; special magnetic structures such as a block having grooves therein to receive windings, or a plate with a plurality of posts thereon, or a plate with a plurality of holes therein. Also included are gapped cores or cores having a special structure such as the cross cores used in constant current transformers or respective equivalents.

Figure 3 shows a picture frame transformer which is similar to the transformer of Figure 2, but includes push-pull windings in both the primary and secondary circuits. In general, any winding represented as a single wire can be replaced with a push-pull winding having two wires which are alternatively energized to provide the alternating excitation

(primary), or which have push-pull rectification or demodulation (secondary).

Four longer cores llla-d and four shorter cores

121a-d can be used. A push-pull primary 212 couples all of the cores llla-d and 121a-d, and is shown with switching FET's

216a,b. The primary 212 shown is a symmetrically divided push-pull winding and is also known as a "symmetrical push- pull winding". The switching FET's are alternatively switched between ON and OFF states to cause current to alternately flow in one direction and in the opposite direction through the cores to excite the transformer. For example, FET 216a may be turned ON to complete the primary circuit between the + and - terminals of the input D.C. voltage supply causing current to flow in one direction. When FET 216a turns OFF and FET 216b turns ON, the primary circuit between the + and - terminals is again completed, causing current to flow in the opposite direction.

For additional details related to the construction of matrix transformers having symmetrical push-pull windings, reference may be made to U. S. Patent application 07/220,532 entitled "Transformer having Symmetrical Push-Pull Windings", filed July 18, 1988 and assigned to the same assignee as the present invention, the disclosure of which is hereby incorporated by reference.

Full wave rectification is provided by six rectifier assemblies 214a-f, shown as common cathode dual rectifiers. A secondary winding 213 comprises six push-pull secondary segments 213a-f, and a bus structure paralleling the common cathode outputs of the rectifier assemblies 214a-f. A load is shown as a resistor 215. The common returns 217a-g are also part of the secondary 213, and are understood to be part of the bus structure.

Note that each of the push-pull secondary segments 213a-f has a common ground 217a-f located in close proximity with the common cathode output of each of the rectifying assemblies 214a-f and which proximity is preferred for high frequency operation to minimize the effects of lead inductance. Note that the two halves of the push-pull winding segments 213a-f pass in opposite directions through the adjacent cores llla-d and 121a-d to the next rectifying assembly 214a-f, and that the inputs to the rectifying assemblies 214a-f are windings 213a-f which have their respective ground connections 217a-f at the next rectifying assembly 214a-f in each direction around the picture frame matrix transformer.

For example, the start 223 of the secondary winding segment 213a is at the rectifying assembly 214f, and passes through core Ilia to the common portion 233 connected to ground 217a of the winding 213a which is located near the rectifying assembly 214a. The winding 213a then continues through core 111b to its end 243 at the rectifying assembly 214b. The winding pattern is similar for each segment 213a-f of the secondary 213.

To gain a better understanding of the invention, reference is made to Figure 4 and the following description wherein the principles of operation of the picture frame matrix transformer invention are explained without regard for optimum interconnection or configuration. A primary winding 602 makes a single pass through eleven core sections 601a-k. Multiple turns, center tapped or split windings and other variations are all usable). The flux capacity of the cores 601a-k taken together must be sufficient to support the volts- seconds of the primary voltage waveform, according to Faraday's law, as is true with any transformer. This relationship is well known to those skilled in the art. A first secondary 603 comprises seven parallel sections 6ø3a-g, and is wound through the cores 6øla-k with terminations at seven equally spaced places. A second secondary winding 604 comprises five parallel sections 604a-e, and is wound through the cores 601a-k with terminations at five equally spaced places. The seven outputs of the first secondary 603 are taken in parallel, observing polarity. The five outputs of the second secondary 604 are similarly taken in parallel, observing polarity.

To develop a step down transformer of ratio 1 to N, consider first an analytical model in which there is one long illustrative core with sufficient flux capacity to support the primary volts-seconds. A core sufficient to support the secondary volts seconds would be 1/N as long, the volts being 1/N, and the waveform being the same. Therefore it is sufficient simply to "cut" the illustrative primary core into N equal segments. The illustrative primary core comprises all of the core segments 6øla-k envisioned as being joined together. When "cut", the core segments 601a-k result. It is readily apparent that the techniques described above can be used to design a picture frame matrix transformer of any integer ratio.

Figure 5 shows a transformer which is equivalent to the transformer of Figure 4, however a pot core type structure is used rather than separated cores. The core structure 611 is circular and has a channel 616 to receive the windings. A cover 615 completes the magnetic circuit. The core structure 611 has a plurality of holes in it to allow the windings to enter and exit. Slots or notches would be equivalent and their location could be on the bottom as shown or on the inside, outside or top or some mix thereof. A primary winding 612 enters the core structure 611 at a first hole, makes one (or more) turns around the core structure through the channel and exits through the same hole.

A first secondary 613 comprises seven secondary segments 613a- g taken in parallel, observing polarity. The seven first secondary segments 613a-g enter and exit the core structure

611 through seven holes which are as being nominally equally spaced σircumferentially in the channel. Each of the first secondary segments 613a-g starts in one hole and ends in the next adjacent hole stitchwise around the channel, the last ending where the first started in a closed pattern. The second secondary 614 comprises five secondary segments 614a-e similarly wound in five holes. Although not essential, it is preferred to use separate holes for the different windings and to space the holes apart. There can be no net current difference in the wires passing through any hole. If there is a difference, the closed flux path around the hole will resist it. Thus, this implementation will be seen to be useful for attenuating common mode currents as might result from capacitively coupled noise.

Even if the voltages are equal, it is preferred to shift one winding relative to the other to eliminate any condition where more than one exit point of different windings are located together. To illustrate, consider a transformer having a five segment winding and a ten segment winding. This could be wound having five exit points at which both windings have terminations. It is preferred to shift the relative position of the windings so that no more than one exit point of different windings are located together. In the example, shifting the relative position of the windings results in none of the exit points of either winding being located with the exit points of the other.

Figure 7 is a partial schematic of a picture frame matrix transformer having non-integer turns ratios. A primary winding 1103 makes a single turn through four cores, lløla-c and 1102. A secondary winding 1104 comprises eight winding segments 1104a-h, each of which is half of a push-pull winding. Six segments 1104a-d,g,h make a single pass through their respective elements. The other two segments, 1104e,f each have 7 turns. Each winding segment is terminated at ground (-) at one end, and is coupled through one of eight rectifiers 1105a-h to a positive bus (+) on the other end. As drawn, the winding segments are suitably arranged and polarized to be full wave rectified push-pull windings. The special bracket like symbol 1106 above core 1102 indicates that this element is a side loop, and that core 1101b and 1101c are adjacent as explained below.

The primary-to-secondary ratio of the transformer of Figure 7 is 1 to 3.14 and this ratio can be seen by analyzing either the current or the voltage relationships. In each element, enz's law must apply so the primary ampere-turns equals the secondary ampere-turns. For three of the elements, there is a single primary and a single secondary turn (the push pull secondary has two wires, but they operate alternately as a single turn winding), so that the current in each of these three secondary winding segments equals the primary current. When summed, the three currents add to equal three times the primary current. The fourth element has a single turn primary and a seven turn secondary. Thus the secondary current of that segment is l/7th the primary current. Added to the other three, the total secondary current is 3 l/7th times the primary current, giving a 1 to

3.14 ratio.

Figure 8 shows a diagram of a possible physical arrangement for the transformer of Figure 7. Wherein the basic structure comprises three cores lløla-c arranged in a closed triangular shaped pattern. There is a dual rectifier pack, comprising rectifiers 1105a-c,g,h at each respective corner of the triangle. Thus the interconnections which are advantageous for high frequency operation are retained with balanced currents and very short lead lengths of the AC circuits. (Each output could also have an associated filter network). The last element is wound in a different core 1102 because it can have lower flux capacity and is optimized for high frequency operation by simply pulling a loop off the primary winding 1103. Smaller rectifiers 1105e,f are suitable for these secondary windings 1104e,f. The output of the rectifiers 1105e,f is paralleled with the outputs of the other rectifiers 1105a-c,g,h and since the four outputs are paralleled, they are constrained to have the same voltage.

Recalling that the element 1102 has seven turns, each has l/7th of the output voltage and Faraday's law shows that the primary voltage drop in this element will be l/7th the drop of the other elements (neglecting diode drops and losses).

Figures 9 and lø show a picture frame matrix transformer circuit assembled in a modified pot core 1701. As is usual with pot cores, there is center post and an outer return path. The outer return path of the pot core 1701 is divided, in this case into eight sections by slots 1711a-h.

These eight sections can be used in the manner of individual elements and wires as a picture frame matrix transformer. A primary winding 1702 comprising four straps 1702a-d is wired as a symmetrical push-pull primary. A secondary winding 1703 comprising sixteen straps 1703a-p is wired as eight parallel push-pull windings, each push-pull winding half coupling one element to give a picture frame matrix transformer having a one to eight ratio. Switching FET's 1704a,b and rectifiers

1705a-p are shown as an example of how external circuitry could be connected, the rectified outputs and the grounds, respectively, being connected in parallel.

The primary 1702 and the secondary 1703 of the picture frame matrix transformer of Figure 9 are preferably comprised of wide, insulated copper straps, for low inductance and high current capacity as illustrated in the cross-section in Figure 10. The ends of the straps could be made with pins extending downward for installation in a printed circuit board, and the exposed ends of the straps, fanned out somewhat, would be excellent heat sinking for the picture frame matrix transformer.

Figure 11 is a schematic of a picture frame matrix transformer comprising six cores 1901a-f, a primary winding 1902 which is a symmetrical push-pull winding and a secondary winding 1903 comprising six secondary winding segments 1903a-f wired as three parallel push-pull windings. Switches 1904a,b provide the alternating primary excitation for transformer action and rectifiers 1905a-c and 19ø6a-c provide a DC output. Each secondary segment makes a pass through two elements, so the ratio is 1 to 2/6, or 1 to 1/3, or 3 to 1. A unique feature of this circuit is that the rectifiers 1905a-c and 1906a-c are evenly spaced for easier layout and better heat spreading.

In the transformer of Figure 11, the three rectifiers 1905a-c could be replaced with one rectifier. The anode potential of these three rectifiers is the same magnitude and phase so the anode connections could be paralleled prior to rectification. Similarly, the three rectifiers 1906a-c could be replaced with one rectifier. Low inductance interconnecting methods should be used for circuits carrying AC currents.

Figure 12A shows the component parts of a picture frame matrix transformer element assembly with progressive assembly shown in Figure 12B and a portion of a picture frame matrix transformer using the same shown in Figure 12C.

Insulation between the conducting parts is necessary and is to be understood although not shown. The assembly comprises two cores 2301a,b, a center conductor 2302 with three terminals

2302a-c, two outer conductors 2303a,b with a joining clip

2303c, and two end terminals 2303d,e.

To assemble the element assembly, a first terminal

2302b is fixed to the center of the center conductor 2302, as by soldering or brazing. With suitable insulation (not shown) or spacing, two outer conductors 2303a,b are slipped over the center conductor 2302 with the first terminal 2302a.

Connecting clip 2303c is fixed to the outer conductors 2303a,b as by soldering to join them, and help locate them in place.

Two cores 2301a,b are slipped over the outer conductors. The end terminals 2303d,e of the outer conductors are then fixed to the outer conductors, as by rolling and soldering, and, with suitable insulation, (not shown) or spacing, the second and third terminals 2302a,c are fixed to the center conductor tube 2301, as by rolling and soldering.

The whole, when assembled and designated 2304, is suitable as a component part of a picture frame matrix transformer, such as in the portion of a picture frame matrix transformer shown in Figure 12C. Two element assemblies, 2304b,c, and parts of two other assemblies 2304a, d, partially cut away to show the primary 2305 and to indicate that the element assemblies 2304a-d are a portion of a larger assembly, are incorporated into a picture frame matrix transformer. The conducting parts of the element assemblies 2304a-d comprise part of the secondary circuit of the picture frame matrix transformer and are coaxial to the primary winding 2305, which primary winding itself is a coaxial winding, as shown, having an inner conductor 2305a and an outer conductor 2305b. A less expensive, but functionally equivalent part could be made of formed flat stock without much compromise. The coaxial configuration, and in particular the symmetrical push pull coaxial configuration, is preferred for higher frequency operation.

Figure 12D shows a portion of another assembly similar to the assembly shown in Figure 12C wherein rectifiers 2306a,b are shown directly connected to the secondary windings to minimize lead lengths to reduce lead inductance to a lowest possible amount. Heat sinking means 2308a,b surround the cores of the matrix transformer and are placed end-to-end in much the same manner as the cores. The secondary leads extend outward to the spacing between the heat sink and the rectifiers 2306a,b with proper polarity, are held in physical contact with the heat sink by wedges 2310a,b to improve heat transfer.

Figure 13 is a schematic of the portion of the picture frame matrix transformer shown in Figure 12C except rectifiers 2406a-f are added. Note that in each instance, the anode of the rectifier connects to the outer conductor and the cathode connects to the inner conductor. This symmetry could be exploited in developing a practical design for the power conversion system or other applications.

Figure 14 shows a schematic of a picture frame matrix transformer similar to the transformer of Figure 13 with the addition of output L-C filters comprising matrix inductors 2314a-c and capacitors 2608 a,b.

Figure 15 shows a possible finished assembly of a picture frame matrix transformer mounted on a printed circuit board 2800. The transformer comprises assemblies 2304a-d similar to the assembly of Figure 12B providing the cores and the secondary windings. Rectifiers 2807a-S provide a DC output. A connector 2808 is shown for example and terminates the assembly for external connection. The picture frame matrix inductor elements 2314a-j and the floating capacitors 3009a,b are also shown. Corner secondary elements 2324a-d are used. The primary winding 3002 is terminated at a corner and the symmetrical switching FET's 3ø06a,b are in the opposite corner from the primary winding terminating corner. This arrangement takes advantage of the openness of the corner and allows higher dielectric insulation and greater creepage distance. Figure 16 is the schematic for the matrix transformer assembly of Figure 15. It can be seen from the illustrative assembly of Figure 15 that the picture matrix transformer is advarrtageous in spreading circuit components over a relatively large area which promotes go.od heat dissipation and reduces the possibility of localized hot spots. In addition, the voltage output may be made in any number of ways well known to those skilled in the art. The power conversion system of the present invention also includes a distribution scheme described herein below.

It is often necessary in a power conversion system to produce voltages other than the nominal power converter voltage and such voltages are needed to operate electronic circuit components associated with the power conversion system. In such instances, the matrix transformer may become conditionally valid. The concept of validity in a matrix transformer is based on Lenz's law and the necessity to keep the ampere turns at zero in each element. However, if the required current is very low and a small portion of the other currents, its effects may be ignored. For example, if one considers a 100 watt, 5 volt converter, the output current will be 20 amperes. If a plus (+) and minus (-) 15 volts at 10 milliamperes is required, there is no reason to be concerned about validity since a winding would be passed through any three of the magnetic elements of the matrix transformer.

Turning to Figure 17, a partial view of a conditionally valid picture frame matrix transformer illustrating two methods of driving a minus 5 volts from several elements of the matrix transformer is shown. It will also be seen that the dual cathode rectifier assemblies 85a-d are similar to and function the same as the dual cathode rectifier assemblies 214a-f of the picture frame matrix transformer shown in Figure 3. Consider that the portion of the matrix transformer illustrated in Figure 17 is part of a

15 to 1 matrix transformer used in a 5 volt power supply having a 150 watt output. In this case, the output current is

30 amperes and each of the 15 parallel elements of the matrix transformer provides 2 amperes. Further assume that a 500 milliampere minus 5 volt output is needed. It may be possible to put a second auxiliary winding on one element to derive the minus 5 volt output voltage however, the current balance in the secondary winding is upset with the difference being the current drawn by the auxiliary minus 5 volt secondary winding.

In order to maintain 0 net ampere turns, the current in the main secondary winding of the element will be reduced by the amount drawn by the auxiliary winding. In this case, the matrix transformer is conditionally valid and the condition is that the auxiliary winding current must always be less than the primary current. This means that if the main output voltage has a light load, the output capability of the auxiliary winding is reduced.

In Figure 17, the minus 5 volts appearing at the terminal 74 is derived using separate windings with one winding 76 passing through one element 78 and a second winding

80 passing through an element 82. The primary winding passing through all the elements is indicated as winding 84. In the second method illustrated, minus 5 volts appears at the terminal 86 which is connected to the anodes of diodes 88 and

90 and to the secondary windings 92 and 94, respectively. The secondary winding 92 passes through element 96 and the winding

94 passes through the element 98. It should be noted that the windings supplying the rectifiers carry extra current while the other windings in the elements have a lower current.

The conditional validity of the matrix transformer may be used as a current limit in the auxiliary winding. For example, in the case of an overload on the auxiliary output, the output current increases and more and more current is stolen from the main secondary winding output until there is none to steal which is the limit where the current in the auxiliary winding equals the primary current. If the load is further increased due to the low impedance being reduced, the current in the auxiliary winding cannot increase further and accordingly, will limit at a constant value which is determined by the current in the primary winding.

The conditional validity of the matrix transformer may be extended for multi-element auxiliary windings. Further considering the example as set forth above for a 15 to 1 matrix transformer, a 500 milliampere, 15 volt winding could steal from three elements and a 50 milliampere, 24 volt winding could steal from 5 others and so forth. This technique works very well if the main output element current is always larger than the auxiliary winding element current.

Another feature of the power conversion system is cascading matrix transformers so that the secondary of one matrix transformer ties the primary of one or more other matrix transformers. The principle of ratio multiplication using cascaded and matrix transformers may be understood by referring to Figure 18 wherein two matrix transformers are cascaded. The first transformer 101 comprises elements 100, 102, 104 and 106 wherein the primary winding 108 passes through elements løø, 102, 104 and 106. Each element has an associated secondary winding 110, 112, 114 and 116 respectively phased and connected in parallel to provide an output voltage across the leads 118, 120. The output of the first matrix transformer 101 serves to drive the primary winding 122 of the second matrix transformer generally designated 124. The matrix transformer 124 is a picture frame matrix transformer and, comprises elements 126, 128, 130, 132 and 134. Again, each of the elements has an associated secondary winding 136, 138, 140, 142 and 144 respectively phased and connected in parallel and produces an output voltage across the output leads 146 and 148. Assume the matrix transformer 101 has a ratio of Ml:l and the picture frame matrix transformer 124 has a ratio of M2:l, then the combination of the cascaded matrix transformers will have a ratio of Ml 21:1. In Figure 18, the matrix transformer 101 has a ratio of 4:1 and the picture frame matrix transformer

124 has a ratio of 5:1. Therefore, the ratio of the input to the output of the cascaded matrix transformers is 20:1 and there are 5 secondary output windings in parallel. The result is that the elements in the picture frame matrix transformer

124 have a lower voltage and carry a higher current. The concept of the cascaded matrix transformers can be extended so that one matrix transformer provides a secondary that can drive several additional matrix transformers each of which may have a different ratios and accordingly provide multiple voltage outputs. It will be seen that if the first matrix transformer is voltage regulated then all the cascaded stages will also have some degree of regulation.

Considering Figure 19, a matrix transformer generally designated 150 provides a 5 volt output voltage. An interstage secondary winding 152 is also wound on the elements of the matrix transformer 150 and its two ends 154, 156 and the voltage which appears across the ends 154, 156 form the input driving voltage to two other matrix transformers generally designated 158, 16ø, respectively. The secondary windings of the cascaded matrix transformers 158 and 160 are not illustrated to simplify the drawing. If it is assumed that the main .secondary winding of the matrix transformer 150 is regulated, then the regulation functions as a pre- regulation for the cascaded matrix transformers 158 and 160. All the conditions and criteria for transformer operation as stated in the foregoing also apply in this configuration.

The interstage winding may also be used to condition power as it is transferred between the first matrix transformer and the cascaded matrix transformers. The interstage of winding is usually carrying power at some intermediate level of voltage and current and it is isolated from both the input and output of the matrix transformer. The interstage power conditioning may be some type of voltage regulation or it might be a filter. As illustrated in Figure 19, the power conditioning device is generally designated within the function block 162 and is in series with the primary winding 164 of the cascaded matrix transformer 160.

A voltage regulation device that may be used with an interstage winding to provide interstage voltage regulation is illustrated in Figure 20 and generally designated 166. For illustrative purposes, a portion of an interstage winding 168 is shown passing through one element 170 and functions as a primary winding with the interstage current I flowing therein. A secondary winding of N turns provides a 1:N transformation ratio so that the current in the secondary is I/N. Any voltage drop appearing across the output of the secondary winding 172 is reflected as a voltage drop in the interstage winding. As illustrated in Figure 20, the output of the secondary winding 172 may be rectified by a rectifying device

174 and loaded by a linear regulator such as the FET 176. The voltage drop across the FET 176 is reflected back to the interstage winding to adjust the voltage output of the cascaded matrix transformer with which it is used.

The voltage drop reflected in the interstage winding could also be provided by a switched mode interstage regulator such as shown in Figure 21. As in case of the interstage voltage regulator of Figure 20, the interstage winding 168 passes through an element 170 and produces an output across the secondary winding 172 which is rectified by the rectifying device 174. The switched mode interstage regulator illustrated in Figure 21 is a regulated voltage sink. The interstage current I, transformed by the element 170 turns ratio 1:N must flow in the regulator secondary and will flow out of the regulator as (l-d)I/N where d is the regulator duty cycle. The switching device 176 is pulsed width a pulse with modulation device 178 such that the input current to the regulator is equal to the output current and the voltage at the terminal 180 must be able to sink the pulsing current produced as the switching device 176 is switched on and off by the pulse width modulator device 178.

Often times it is desirable to connect successive variable ratio matrix transformers to provide a large number of voltage steps to increase voltage resolution. Because variable ratio matrix transformers have step changes in its ratio, a capacitor βannot be used on both its input and output. This is particularly true if the ratio steps are relatively large and modulation is used between the two closest ratio steps to provide infinite resolution of the output voltage. A solution to this problem is shown in Figure

22 wherein two variable ratio matrix transformers 182 and 184 are connected by an interstage filter matrix inductor 186.

The interstage filter inductor 186 is used to provide a suitable impedance allowing both transformers to have a variable ratio. In the illustration of Figure 22, the first transformer 182 functions as a pre-regulator responding with feed forward from the input and it is ratio modulated for infinite resolution. The second matrix transformer 184 is also a variable ratio transformer and responds to feedback from the output. The transformer 182 has a capacitor 188 at its input which prevents the input voltage from changing quickly. As the ratio of the transformer 182 is modulated between the closest steps, its output voltage Vi at the lead

190 will be V_p/Mι where the ratio is Mj_:l and the output voltage V will be V_p/M₂ when the ratio is M₂: 1• If the duty cycle d is defined as the time when the ratio is H_ >. 1 divided by the period then the average first stage output voltage V^ may be expressed as: V^ = V_p/ [dM^+( 1-d) M₂ ] . A similar expression relates the second stage input voltage V₂ at the lead 192 to the output voltage V₀ at the lead 194 and may be expressed as: V₂ = [d'M₃+( 1-d' )M ] V₀. The quantity d' is the second stage duty cycle M3: 1 and M4 : 1 are the ratios between which the transformer 184 is being modulated. On the average,

Vl must equal V and the interstage filter matrix inductor 186 provides the impedance to allow the instantaneous voltage differences. The voltage across the interstage inductor 186 on the average must be 0. The AC voltage will be small if both the transformers 182 and 184 have a sufficiently large number of steps so that the incremental steps between which modulation occurs are small. Although Figure 22 illustrates a variable ratio matrix transformer of the general type, it will be understood that a binary matrix transformer may be used to vary the turns ratio.

The current in an interstage winding is generally a high frequency AC current having high DI/DT. Indiscriminate mounting around the transformer structure would cause severe problems due to the leakage inductance in the wire and due to

EMI. If the cascaded stages are only for small, auxiliary output voltages with low current, then the problem is more manageable and the inductance of the interstage winding is a parallel impedance to the main secondary winding and will not degrade the operation of the switching of the primary winding.

However, where all the power passes through the interstage winding, it is important to keep the lead lenghts as short as possible and to minimize the inductance as much as possible.

This is difficult to accomplish if the outputs of the transformer are brought together, paralleled and then distributed as one winding. One solution to this problem is not to collect the current from the windings but rather to keep them as individual wires and to align them immediately with their destination. In other words, as the parallel wires leave the secondary windings of the first transformer they should enter immediately adjacent second transformer. Such an arrangement is illustrated in Figure 23 wherein a cascaded matrix transformer is shown in one potcore- like magnetic structure and generally designated 196. One of the interstage windings 198 of the transformer 196 of Figure 23 is illustrated in Figure 24. Figure 25 is a fragmentary view illustrating one of the output windings 200 of the cascaded matrix transformer 196 of Figure 23.

Referring to Figure 11, the primary winding 202 of the first stage matrix transformer makes a single turn around through the outer eight elements 204-218. Eight parallel secondary windings exit the first stage transformer in a similar manner as in a picture frame matrix transformer. Each secondary winding is taken to the second stage passing immediately into it and making a single turn through seven of the eight elements. This provides the optimum alignment so that each winding passes immediately from the first transformer to the second and back again. This provides eight overlapping primary windings for the second stage and each of the eight windings couple seven elements. Each has coupled the first primary in the first stage transformer so each of the eight windings carries a current equal to the primary winding. There are seven wires in any one core of the second stage transformer with each wire carrying a current equal to the primary winding. Accordingly, the first stage transformer ratio is 7:1 to the primary of the second stage transformer. As illustrated in Figure 23, the second stage secondary has a 4:1 ratio from the secondary stage primary. The total ratio of the two cascaded matrix transformers is therefore 28:1. It will be noted that the first stage cores must have seven times the flux capacity of the second stage cores and the current in the second stage is seven times the primary current. The current is carried in eight isolated wires in the second stage primary. The output windings of the second stage transformer also each carry seven times the primary current and the currents will be balanced. Again the above transformer principles apply.

The power conversion system utilizing flat matrix transformer technology such as described above lends itself to implementing and facilitating power distribution and heat sinking. Although printed circuit assemblies may use power and ground planes for thermal conductivity and EMI shielding, such planes are not possible with power converter systems due generally to the high currents that must be carried and which would necessitate the planes to be very thick. A further limitation is the printed circuit paths themselves must be carefully considered so as to not create hot spots. One approach to solving this problem is illustrated in Figure 26 wherein a partial view of a flat matrix transformer power converter mounted on a circuit board is illustrated and generally designated 220. The electrical components generating the heat are arranged on the circuit board 222 such that there is spacing around the edges and around the periphery of the circuit board itself. Two halves of castings 224 and 226 are fabricated to accommodate the circuit board topology and circuit component layout. The castings are then clamped on the circuit board to sandwich the circuit board and the inner surfaces of the castings 224 and 226 come into contact with the circuit board to transfer and dissipate heat. The castings 224 and 226 may further be used to provide electrical terminations for power cables and to distribute power to the converter circuit board. As illustrated in Figure 26, the casting 224 is arranged with a positive terminal 228 and the casting 226 is arranged with a negative terminal 230. The circuit board 222 has in this case, a bus arrangement such that the inner surface area 232 contacts the positive power bus of the circuit board 222. Likewise, the inner surface 234 contacts the negative power bus of the circuit 222. The assembly 220 of Figure 26 will be seen to be a combined power bus and heat sinking power converter.

The distribution of power in a power conversion system presents serious problems particularly as the voltages become lower and the currents become higher. Typically, the connector of a power supply is troublesome with plug-in assemblies since it is difficult design large bus bars or power and ground planes having sufficiently low impedance to carry power any significant distance. Improvements for solving the power distribution problem have included the use of distributed systems however, as power densities increase, heat sinking becomes a critical factor and accordingly miniaturization inner transaction of a power converter system is limited because of thermal conductivity considerations.

One method of heat sinking printed circuit cards is to use specialized card guides which have been optimized for heat conduction and even with such specialized card guides, heat sinking and thermal conductivity is limited. Heat sinking and power distribution apparatus of the present invention is illustrated schematically in Figure 27 and is generally designated 236. The assembly 236 includes a card 238 positioned within the chassis 240 of the assembly 236. The card 238 is shown without components for purposes of simplicity. The chassis 240 includes power buses extended longitudinally and distributed along the sidewalls 242, 244 to provide power connections to the edges of the card 236. In

Figure 27, the positive voltage buses 246, 248 are interleaved with the negative voltage buses 250, 252 and 254.

A fragmentary view showing the connection between a printed circuit card 238 and the bus structure of the assembly

236 of Figure 27 is illustrated in Figure 28. The card 238 is assumed to have a positive voltage plane on one side 256 and a negative voltage plane on its opposite side 258. The card guide assembly generally designated 260 includes a positive bus half 262 and a negative bus half 264. The card guide assembly 260 is connected further to the buses distributed along the sidewall of the chassis 266. As illustrated in

Figs. 27 and 28, the power buses extend along the wall of the chassis and are connected to the respective positive half 262 and negative half 264 of the card guide assembly 260. This provides a large area available at the card edge and presents a very low impedance connection. The respective positive and negative voltage buses are electrically isolated from the chassis 266 with an electrically insulating material 268.

Preferably, the contact area with the chassis is large and relatively thin to promote good heat conduction. Wide flat conductors bonded with a good filled epoxy or one or more film dielectric layers could be included in addition to insulated mechanical fasteners to provide the insulating connection. In addition, the positive half 262 and negative half 264 of the card guide edge assembly 260 require their respective contact areas to be in substantially solid engagement with the card

256 to provide good heat and power conduction.

A wedging card guide assembly for providing positive mechanical and thermal contact with the edges of a printed circuit card is illustrated in Figure 29 showing the negative voltage contact lead to the printed circuit card 238 wherein the card is clamped in positive mechanical engagement with the card guide. The chassis 266 is electrically insulated from the negative bus 250 by an insulating layer 268. The wedging card guide assembly is generally designated 270 and includes a portion 272 in mechanical and electrical contact with the bus 250. The portion 272 includes an inclined surface 274 which extends the vertical length of the chassis wall from top to bottom and contacts each of the negative voltage buses. A complementary wedging portion 276 includes a flat surface contact area 278 which contacts the negative voltage portion 258 of the printed circuit card 238 when the wedging card guide assembly 270 is in the clamping position as illustrated in Figure 29. A similar wedging card guide assembly is used to contact the positive voltage bus and is designated generally as 280 in Figure 29 and operates similarly as the assembly 270 with all comments being applicable. The wedging card guide assembly 280 is shown in its released position removing mechanical and electrical contact from the surface 256 of the card 238. When the assembly 280 is operated to its clamped position, the card 236 is compressed between the two surface areas 278 and 282 to provide the positive mechanical and electrical connection for heat and power conduction. The wedge card guide assembly 270 includes an angular shaped spring 284 which has one end 286 in mechanical engagement with the portion 272 contacting the negative bus and has its opposite end 288 in cooperating engagement with a recess channel 290 in the moveable clamping portion 276. The portion 276 is urged by the action of this angular spring 284 in the clamping direction indicated by the arrow 292. The mechanism further includes a cam 294 extending lengthwise between the spring 284 and the contact portion 272. The cam

294 is rotated about its longitudinal axis causing the spring

284 to deform slightly from its shape urging it in a direction indicated by the arrow 296 with the result that the wedging portion 276 moves along the surface 274 and out of engagement with the surface 258 of the card 238. When the cam 294 is operated to release the wedging card guide assembly, the circuit card 238 may be removed.

Although the wedging card guide assembly is disclosed above as being used for both the positive and negative voltages, it is only necessary to use the card guide assembly for one voltage and utilize a fixed card guide for the other voltage. The single card guide assembly, when operated, will compress the circuit card between itself and the fixed card guide. It will also be recognized that a card guide assembly is used at the oppositely disposed edge of the printed circuit card thereby insuring that the card will remain connected during any vibration of the chassis.

As stated above in connection with Figure 1, the matrix transformer may also be a binary matrix transformer wherein the windings of the elements have a binary progression in one or both of the M or N dimensions. That is, for example, the primary windings of the elements of rows 1, 2, 3,

4, etc. can have 1, 2, 4, etc. turns respectively on each element in the row. The binary matrix transformer can be used for ratios that can be expressed more simply as the ratio of binary numbers and they are also particularly useful for variable ratio matrix transformers. Figure 30 illustrates a simple, two element binary matrix transformer wherein the primary winding 300 makes a binary progression from row 1 to row 2 (2⁰, 2*). The secondary has a single turn in each row and since the primary current passes through the first element once, the secondary current in the first element equals the primary current. However, the primary current passes through the second element twice so the secondary current in the second element is two times the primary current. Therefore, I

■ 3I_p and the ratio is 1:3, or expressed in binary form, 1:11.

Figure 31 illustrates another binary matrix transformer having dimensions 2 x 3. In the first column it is seen that I_sι = I_pι and I_s2 = 2 x I_pχ. In every column, the primary of row 2 has twice the turns of the primary of row

1 so the two to one relationship of I_s2 to I_sι holds throughout. Comparing the elements of row 1 in columns 1, 2, and 3 shows that the primary current I_pι, I_p2, and I_p3 are 1,

2 and 4 times the secondary current I_sι, respectively. In the second row, the primary of each element has two turns but I_s2 is known to be twice I_sι so again the primary current I_pι, I_p2 and I_p3 are 1, 2 and 4 times the secondary current I_sι- Using - ε l ^{= I}pl ⁱⁿ element 1, 1 is the common denominator, it can be seen that the primary current is seven times I_sl and the secondary current is three times I_pι so the ratio is 3:7. The ratio when expressed as a binary number is 11:111.

A generalized binary matrix transformer is illustrated in a schematic form in Figure 32. For a binary matrix transformer having dimensions M x N, the ratio is (2^M - 1):(2^N - 1). In binary form, this is a binary number with M binary factors (l's) to a binary number with N binary factors. A 4 x 5 binary transformer for instance, would have a ratio of 15:31 or binary 1111:11111. To make a matrix transformer having binary ratios which are not all 1's, the rows or columns corresponding to the binary factors where 0's appear in the binary numbers are left out. Such a binary matrix transformer is illustrated in Figure 33 wherein the resultant binary matrix transformer corresponds to a ratio of^'44:86 or a binary 10110:1010110.

Referring to Figs. 34 and 35, a generalized binary M x 1 matrix transformer and a generalized binary 1 x N matrix transformer is shown respectively. These are degenerate cases of the binary matrix transformer and are particularly useful in variable ratio matrix transformers. As in the case of a M x N binary matrix transformer, if the binary numbers contain

0's, the element corresponding to that factor is left out. By removing rows and columns from a binary matrix transformer, the ratio of the binary matrix transformer is changed and accordingly its binary representation is also changed. This feature is particularly useful in developing a variable ratio matrix transformer particularly for digital control applications. Referring to Figure 36, a 3 x 4 binary matrix transformer is illustrated as an example of removing a column and a row and which has the result of changing the corresponding binary factor to zero. The ratio of the binary matrix transformer of Figure 36 is changed from 7:15 to 5:11, or expressed as binary numbers, the ratio has changed from

111:1111 to 101:1011.

The degenerate binary matrix transformers illustrated in Figs. 34 and 35 may also be used as variable ratio matrix transformers due to their simplicity and ease of control which makes them particularly useful for digital control applications. If the M x 1 binary matrix transformer of Figure 34 is considered to have four rows and rows two and four are removed, the ratio is changed from 15:1 to 5:1

(binary 1111:1 to 0101:1). Likewise, considering the binary 1 x N matrix transformer of Figure 35 and assuming that there are four columns, and further that columns two and four are removed, the ratio is changed from 1:15 to 1:5 (binary 1:1111 to 1:0101) .

It will be appreciated from the foregoing that the binary matrix transformer may be configured as a fixed matrix transformer or when the appropriate rows and columns are removed may function as a variable ratio transformer.

A power conversion system utilizing flat matrix transformer technology has been described above in several preferred embodiments. It should be understood that numerous changes and modifications may be made by those skilled in the art and accordingly the invention has been described by way of illustration rather than limitation.

Claims

I CLAIM :

1. A high frequency power conversion system, said system comprising: means for producing a regulated DC output voltage from a DC input voltage wherein said output voltage is at a high current; said DC producing means further including; means for pre- regulating the DC input voltage; switching means coupled to said pre-regulation means for converting said additional DC voltage to an AC voltage; picture frame matrix transformer means coupled to said switching means for transforming the magnitude of said AC voltage to a desired magnitude AC voltage; interstage power conditioning means coupled to said matrix transformer for regulating the magnitude of the AC voltage at the output of said matrix transformer; second matrix transformer means coupled to said interstage power conditioning means for further transforming said conditioned AC voltage to a desired AC voltage output; rectifying circuit means coupled to the output of said second matrix transformer for converting said AC voltage signal to a DC rectified voltage signal, and power distribution means coupled to said rectifying circuit means for providing an electrical and mechanical connection from said circuit rectifying means to an output terminal.

2. Power conversion system as defined in claim 1 further including cascaded matrix transformer means coupled to said interstage power conditioning means for producing a DC output voltage having a magnitude other than the magnitude of the DC output voltage available at said power distribution means.

3. Power conversion system as defined in claim 2 wherein said interstage power conditioning means further includes a post regulation means and a second cascaded matrix transformer coupled to said interstage power conditioning means whereby a third DC output voltage is produced and regulated by said post regulation means.

4. A power conversion system as defined in claim 1 wherein said power distribution means includes a chassis for receiving printed circuit boards and further including edge card connectors for distributing voltage produced by said power converter system.

5. A power conversion system as defined in claim 4 wherein said chassis includes a edging card guide for providing positive mechanical and electrical contact with power busses on said printed circuit card.

6. A power conversion system as defined in claim 5 wherein said wedging card guide includes a moveable portion including biasing means for urging said moveable portion into contact with the surface of a printed circuit card inserted in said chassis and further including a cam for urging said biasing means in an opposite direction to release a card inserted in said wedging card guide.

7. A high frequency power conversion system as defined in claim 1 further characterized in that said picture frame matrix transformer includes: a plurality of interdependent magnetic elements; at least one primary winding comprising an electrical conductor, each having first and second ends; at least one secondary winding comprising an electrical conductor, each having first and second ends; each of said plurality of interdependent magnetic elements being arranged end-to-end with an immediately adjacent magnetic element in a closed pattern configuration; one of said at least one primary winding and one secondary winding passing at least once through each of said plurality of interdependent magnetic elements, one of said first and second ends ending at an end-to-end position between adjacent magnetic elements where the other of said first and second ends begins; each of said plurality of magnetic elements further having at least one turn of the other of at least one primary and secondary windings with the winding being associated one- for-one with said plurality of magnetic elements, one of said first and second ends of said winding associated with one magnetic element ending at the end-to-end position between magnetic elements where one of said first and second ends of a winding associated with an adjacent magnetic elements begins, said ending of one winding and the beginning of an adjacent winding at an end-to-end position between magnetic elements being repeated for each of said end-to-end positions.

8. A high frequency power conversion system as defined in claim 7 further characterized by said interdependent magnetic elements being axially elongated cores having a cross sectional dimension substantially smaller than its longitudinal dimension.

9. A high frequency power conversion system as defined in claim 8 characterized in that each of said cores are further characterized by a number of core segments wherein the total flux capacity of said number of core segments is sufficient to support the volts-seconds of a primary winding voltage signal.

10. A high frequency power conversion system as defined in claim 9 further characterized in that said picture frame matrix transformer has N core segments, a one turn primary winding, and a plurality of secondary windings, each of said plurality of secondary windings passing through one of said N core segments and each having their respective first and second ends connected together such that said secondary windings are in parallel whereby a matrix transformer having a transformation ratio of 1 to N is made.

11. A high frequency power conversion system as defined in claim 10 characterized by said N core segments are of unequal longitudinal lengths, the length of the shortest core segment being no shorter than that length required for a core segment to have sufficient flux capacity to support 1/N volts-seconds of a primary winding voltage signal.

12. A high frequency power conversion system as defined in claim 10 characterized in that N is an integer.

13. A high frequency power conversion system as defined in claim 10 further characterized by at least one of said plurality of secondary windings having a first number of turns (X) and forming a first secondary circuit and at least one of other of said plurality of secondary windings having a second number of turn (Y) and forming a second secondary circuit, said first and second secondary circuits having different numbers of turns whereby the magnitude of current flowing in said first secondary circuit is 1/X of the current Ip flowing in said primary circuit and the magnitude of current flowing in said secondary circuit is 1/Y of the current Ip flowing in said primary circuit.

14. A high frequency power conversion system as defined in claim 7 further characterized by said picture frame matrix transformer having a first number (A) of first secondary circuits and a second number (B) of second secondary circuits wherein the sum of A+B=N, said first and second secondary circuits being connected in parallel and the total secondary current is Is * A(l/X Ip)+B(l/Y Ip) whereby the transformation ratio of the matrix transformer is 1 to [A(l/X)+B(l/Y) ].

15. A high frequency power conversion system as defined in claim 9 characterized in that the picture frame matrix transformer has N core segments, a plurality of primary windings and a plurality of secondary windings, each of said plurality of primary windings passing through P core segments and each having their respective first and second ends connected together such that said primary windings are in parallel, each of said plurality of secondary windings passing through S core segments and each having their respective first and second ends connected together such that said secondary windings are in parallel whereby a matrix transformer having a transformation ratio of P to S is made.

16. A high frequency power conversion system as defined in claim 9 characterized by said at least one secondary winding being a push-pull winding.

17. A high frequency power conversion system as defined in claim 16 further characterized by said picture frame matrix transformer having a plurality of push-pull devices being connected to said push-pull secondary windings, said push-pull devices comprising dual rectifiers having a common output terminal and being located at an end-to-end_. position, one of said rectifiers being connected to and terminating one end of said push-pull winding wound in one direction through said cores and the other of said rectifiers being connected to and terminating one end of another of said push-pull windings wound through said cores opposite to said one direction, each half of one of said plurality of push-pull windings passing in opposite directions through adjacent cores.

18. A high frequency power conversion system as defined in claim 9 characterized in that said at least one primary winding is a push-pull winding.

19. A high frequency power conversion system as defined in claim 7 characterized by said plurality of interdependent magnetic elements of said picture frame matrix transformer forming a closed structure having an interior channel for carrying said primary and secondary windings, said structure further including a number of spaced apart openings in registration with said channel and through which the ends of said windings pass.

20. A picture frame matrix transformer for use in a power conversion system for supply a D.C. voltage characterized by: a plurality of axially elongated magnetic cores having a cross sectional dimension substantially smaller than its longitudinal dimension, each of said plurality of cores being arranged end-to-end with an immediately preceding and an immediately following core in a closed pattern configuration; at least one primary winding and at least one secondary winding, each winding forming an electrical conductor having first and second ends and passing at least once through each of said cores and one of said first and second ends of one conductor ending at an end-to-end position between adjacent cores at the location where the other of sai first and second ends of another conductor begins, each of said plurality of cores further having at least one turn of the other of at least one primary and secondary windings with the winding being,.associated one-for-one with said plurality of cores, one of said first and second ends of said winding associated with one core ending at the end-to-end position between cores where one of said first and second ends of a winding associated with an adjacent core begins, said ending of one winding and the beginning of an adjacent winding at an end-to-end position between magnetic elements being repeated for each of said end-to-end positions; said ending of one winding and the beginning of an adjacent winding being connected to rectifying means one-for- one at each of said end-to-end positions, said rectifying means having an output, each of said outputs being connected to one another and to a load.

21. A picture frame matrix transformer as defined in claim 20 further characterized in that said rectifying means is a bridge rectifying device and providing a D.C. output voltage.

22. A picture frame matrix transformer as defined in claim 20 further characterized by said secondary winding being a symmetrical push-pull winding and said rectifying means being a dual rectifier having a common output terminal and providing a D.C. output voltage.

23. A picture frame matrix transformer as defined in claim 20 further characterized by said primary winding being a push-pull winding.