MAGNETIC CORE, METHODS OF DESIGNING AND CONSTRUCTING THEREOF

AND DEVICES INCLUDING SAME

FIELD OF THE INVENTION

This invention relates in general to the field of magnetic cores usable in transformers, chokes and other electromagnetic devices, and more particularly to multi- gap high efficiency, low loss magnetic cores.

BACKGROUND OF THE INVENTION

Transformers and chokes designed for high current applications typically have a ferromagnetic core with a closed magnetic flux path. Such cores are typically made from a ferrite material as is well known in the art. Frequently, air gaps (or gaps comprising spacers made from electrically insulating (non-electrically conducting), non- ferromagnetic materials, such as Mylar® or another suitable polymer based material, or the like) are introduced into the ferrite structure of the core in order to increase the core magnetic impedance in order to avoid magnetic saturation of the core which may lead to substantial undesirable power losses. Transformer and choke ferromagnetic cores with one air gap, two air gaps and even four air gaps are known in the art. However, the number and geometrical arrangement of such gaps have been typically limited to a maximum of four gaps due to mechanical, assembly and manufacturing considerations, such as, for example, the availability of commercially manufactured ferrite shapes (such as L shaped ferrite parts and U shaped ferrite parts), the difficulty and the excessive cost of precisely assembling and aligning the different parts of the core to close tolerances.

Moreover, in prior art magnetic cores having four gaps, the gap positions are dictated by the use of commercially available core parts resulting in magnetic core designs in which the positioning of the four gaps are far from optimal. Thus, to the best knowledge of the inventor, the closed magnetic path ferromagnetic cores of prior art chokes and transformers have a maximum of four gaps formed therein (typically, one, two, three or four gaps are used in prior art ferromagnetic core designs of the prior art). Furthermore, because of such practical considerations and the limited shapes of commercially available core component shapes, very little experimentation has been performed regarding methods for reducing power losses in such ferromagnetic cores through methods of modifying the magnetic field flux lines through premeditated gap positioning and gap design. Since chokes and transformers are ubiquitous in many devices and electrical applications, and since even small increases (as small as several percent or even fractions of a percent reduction in energy losses) in device efficiency may result in very substantial energy savings for the entire lifetime of the devices and may extend the practical lifetime of such more efficient magnetic devices, resulting in substantial savings in cost and less environmental heat pollution. There is therefore a long felt need in the industry for methods of reducing power losses in such chokes and transformers and other electrical/magnetic devices using a magnetic core.

BRIEF DESCRIPTION OF THE DRAWINGS

The invention is herein described, by way of example only, with reference to the accompanying drawings, in which like components are designated by like reference numerals, wherein:

Fig. 1 A is a schematic isometric view illustrating a prior art ferrite core of a choke or transformer having two gaps;

Fig. IB is a side view of the prior art ferrite core of Fig.lA;

Fig. 1C is a cross sectional view of the prior art ferrite core of Fig. IB taken along the lines I-I;

Fig. 2A is a schematic isometric view illustrating a choke or transformer having a ferrite core with four gaps and illustrating the choke coils;

Fig. 2B is a side view of the choke of Fig.2A;

Fig. 2C is a cross sectional view of the choke of Fig. 2B taken along the lines II- Π;

Fig. 2D is a cross sectional view of the prior art choke of Fig. 2B taken along the lines III-III;

Fig. 3A is a schematic isometric view illustrating a ferrite core of a choke or transformer having six gaps in accordance with an embodiment of the magnetic cores of the present application;

Fig. 3B is a side view of the magnetic core of Fig.3 A;

Fig. 3C is a cross sectional view of the magnetic core of Fig. 3B taken along the lines IV-IV;

Fig. 4A is a schematic isometric view illustrating a ferrite core of a choke or transformer having eight gaps in accordance with another embodiment of the magnetic cores of the present application;

Fig. 4B is a side view of the magnetic core of Fig.4A;

Fig. 4C is a cross sectional view of the magnetic core of Fig. 4B taken along the lines IV-IV;

Fig. 5A is a schematic isometric view illustrating a magnetic core of a choke or transformer having ten gaps in accordance with another embodiment of the ferrite cores of the present application;

Fig. 5B is a side view of the ferrite core of Fig.5A;

Fig. 5C is a cross sectional view of the ferrite core of Fig. 5B taken along the lines

V-V;

Fig. 6 is a schematic diagram illustrating the testing configuration used for testing the inductance decrease of different magnetic cores illustrated in Figs. 1 A-5A and 15;

Fig. 7 is a schematic cross sectional diagram illustrating the dimensions of the L- shaped members and straight bar-like members of the magnetic cores of the present application;

Fig. 8A is a cross-sectional diagram schematically illustrating the magnetic flux lines in the vicinity of a gap in part of a ferrite core;

Fig. 8B is a cross-sectional diagram schematically illustrating the magnetic flux lines in the vicinity of a gap in part of a ferrite core which has a gap length smaller than the gap length of the air gap illustrated in Fig. 8A; Fig. 9 is a schematic graph illustrating the dependence of the inductance on the current flowing through the coils of a choke having the two gap ferrite core of Fig.1 A;

Fig. 10 is a schematic graph illustrating the dependence of the inductance on the current flowing through the coils of a choke having the four air gap ferrite core of Fig.2A;

Fig. 11 is a schematic graph illustrating the dependence of the inductance on the current flowing through the coils of a choke having the six gap ferrite core of Fig. 3A, in accordance with an embodiment of the magnetic cores of the present application;

Fig. 12 is a schematic graph illustrating the dependence of the inductance on the current flowing through the coils of a choke having the eight gap ferrite core of Fig.4A, in accordance with another embodiment of the magnetic cores of the present application;

Fig. 13 is a schematic graph illustrating the dependence of the inductance on the current flowing through the coils of a choke having the ten gap ferrite core of Fig.5 A, in accordance with another embodiment of the magnetic cores of the present application;

Fig. 14 is a schematic graph illustrating the dependence of the % inductance loss values on the number of gaps included in the magnetic core for each of the tested chokes having the magnetic cores of Figs 1A, 2A, 3A, 4A and 5A, as measured at current values of 45, 40, 35, 30, 25, 20, 15 and 10 ampere for each choke;

Fig. 15 is a schematic isometric view illustrating a prior art generally rectangular magnetic core having four straight magnetic members and four gaps such that the total gap length of the core is 11 mm;

Fig. 16 is a schematic cross sectional view of the prior art magnetic core of Fig 15, taken along the lines XVI-XVI;

Fig. 17 is a schematic cross-sectional view illustrating a magnetic core including four L Shaped magnetic members and four gaps and having gap positions in accordance with an embodiment of the cores of the present application;

Figs. 18-21 are schematic cross-sectional views illustrating the approximate magnetic flux lines distribution in four different gaps of a magnetic core;

Fig. 22 is a schematic graph illustrating the dependence of the inductance on the current flowing through the coils of a choke having the four gap ferrite core of Fig.15;

Fig. 23 is a schematic graph illustrating the dependence of the inductance on the current flowing through the coils of a choke having the four gap ferrite core illustrated in the cross-sectional view of 17;

Fig. 24 is a schematic graph illustrating a comparison of the dependence of the inductance on the current flowing through the coils of a choke having the four gap ferrite core illustrated in the cross-sectional view of Fig. 17, when the Total gap length is decreased from 1 1 mm to 9 mm; and

Fig. 25 is a schematic block diagram illustrating steps of a method for designing the multi-gap magnetic cores of the present application in accordance with an embodiment of the methods of the present application.

DETAILED DESCRIPTION OF THE INVENTION

Notation Used Throughout

The following notation is used throughout the present application.

Term Definition

A Ampere

AC Alternating current

DC Direct current

Hz Hertz

KHz Kilohertz

niA Milliampere

niH MilliHenry

mm Millimeter

mm ² Square millimeter

mV Millivolt

PBT Polybutylene Terephtalate

It is noted that the term "magnetic core" as used throughout the present application refers to the cores of electromagnetic devices such as, but not limited to, electrical transformers, chokes or any other magnetic core based devices which include a yoke or similar magnetic or ferromagnetic structure or member having an inductance and a magnetic impedance. While the present application mainly discloses practical examples of chokes and transformers, the magnetic core designs disclosed herein and the methods for their design and implementations may be used in other types of electromagnetically operated devices which include ferromagnetic or magnetic cores.

The present application discloses novel designs for ferromagnetic cores for use in low power loss chokes and transformers as well as chokes and transformers including such novel ferromagnetic cores and methods for designing and constructing such ferromagnetic cores, chokes and transformers.

In accordance with one embodiment of the magnetic cores of the present application, the novel ferromagnetic cores of the present application are closed magnetic path cores having six or more gaps. The gaps may be air gaps but are preferably (but not obligatorily) gaps having spacers disposed therein. The spacers may be made from made from any suitable electrically insulating (non-electrically conducting), non-ferromagnetic materials, such as, but not limited to, Mylar® or other suitable polymer based materials, cardboard, paper, or the like). The spacers may assist in keeping the gap dimensions at the desired gap length and may also assist in assembling and aligning the ferromagnetic core parts together to simplify construction and increase achievable precision of the gap dimensions and gap length.

Typically, when designing the prior art closed magnetic path ferrite cores for chokes and transformers, the designer first designs the dimensions of a gapless closed magnetic path ferromagnetic core (such as, for example, the ferromagnetic cores 10 and 20 of Figs. 1 A and 2A, respectively, but without any of the gaps in the cores such that the core is a continuous unbroken and non-segmented core). The magnetic impedance of such a gapless core is then computed. The designer then computes the length of a single air gap required to decrease the magnetic impedance of the core to a value which will sufficiently limit the magnetic flux within the core when the choke or transformer is operated at the nominal maximal current value such that the core would not magnetically saturate at the maximal nominal operating coil current. This calculation is performed taking into account, inter alia, the core's shape and magnetic and electrical properties, the device's operating temperature range, the maximal current levels, the electrical properties of air (or of other materials designed to be disposed as spacers within the gap if used in the design), as is known in the art. After the desired value of the gap length is computed as described above, the designer may determine the gap lengths of the two or four gaps of the actual ferromagnetic core design by either dividing the computed single-gap by the number of gaps to be formed in the core or by using different gap values such that the total gap length of all the gap lengths combined together equals the computed total gap length.

The inventor has unexpectedly found that increasing the number of gaps, preferably (but not obligatorily) by steps of two gaps ( such that six or more gaps are included in the core), resulted in a substantial reduction in the decrease of choke inductance with a concomitant substantial decrease in power losses. This effect is particularly noticeable at higher current levels which are close to the maximal nominally allowable choke current level (or maximal allowable transformer current level). The inventor believes that this surprising finding is based on the fact that while the total gap length of the core remains unchanged as the number of gaps is increased beyond four (such as, but not limited to, six gaps, seven gaps, eight gaps), the distortion of the magnetic field lines near the gap is smaller for shorter gap length and diminishes as the individual gap length is reduced. The decrease in the individual gap length resulting from using a large number of gaps (at least five gaps or more) contributes to improving the performance because it reduces the effective magnetic cross-sectional area in the gap, which in turn increases the inductance of the device back to the design goal inductance value.

It is noted that while all the experiments performed by the inventor and all the cores constructed and tested as described hereinabove had an even number of gaps, this is not obligatory, and the number of gaps usable in the novel cores described in the present application may also be an odd number such as but not limited to five gaps, seven gaps, nine gaps , eleven gaps, and the like.

The reasons that the number of gaps used in the experiments described herein was even is mainly for the sake of manufacturing convenience and commercial availability of "off the shelf ferrite members. Another reason why only cores with an even number of gaps were constructed and tested was because the use of an even number of gaps allows for a symmetrical core construction which may even further contribute to the reduction of magnetic flux losses and improve the device's performance. However, from theoretical consideration, a core having seven or nine gaps which are asymmetrically disposed within the core (or at least asymmetrically disposed on two opposing sides of a rectangular core), may still have a lower inductance loss at currents close to the nominal maximal current values than a core having four gaps each having a larger gap length than the gap length of the gaps in the cores having seven or nine gaps.

Reference is now briefly made to Figs. 8A and 8B. Fig. 8A is a cross-sectional diagram schematically illustrating the magnetic flux lines in the vicinity of an air gap in part of a ferrite core. Fig. 8B is a cross-sectional diagram schematically illustrating the magnetic flux lines in the vicinity of an air gap in part of a ferrite core. The gap length of the air gap of Fig. 8B is narrower than the gap length air gap illustrated in Fig. 8 A.

Turning to Fig. 8A, part of a ferrite core 1 is illustrated (the entire magnetic core 1 is not shown for the sake of clarity of illustration). An air gap 1C having a gap length GL1 which is the distance separating the parts 1A and IB of the magnetic core 1. The lines 5 A schematically represent the magnetic flux lines within the gap 1C and in the vicinity of the gap 1C. Turning to Fig. 8, part of a ferrite core 2 is illustrated (the entire magnetic core is not shown for the sake of clarity of illustration). An air gap 2C having a gap length GL2 which is the distance separating the parts 2A and 2B of the magnetic core 2. It is noted that the cross-sectional shape, material and other dimensions of the magnetic cores 1 and 2 are kept the same except for the gap length GL2 of the magnetic core 2 which is larger than the gap length GL1 of the magnetic core 1.

The lines 7A schematically represent the magnetic flux lines within the gap 2C and in the vicinity of the gap 2C. The gap length GL2 of the gap 2C is larger than the gap length GL1 of the gap 1C (GL2>GL1). It may be seen that while the magnetic flux lines within the middle of both gaps are very close to straight line perpendicular to the opposing faces of the core parts flanking the gaps, the magnetic flux lines get distorted as one moves closer to the lateral ends of the gaps. In the gap 2C, the distortion (curvature) of the magnetic flux lines at a predetermined lateral distance from the gap's end is larger than the distortion (curvature) of the magnetic flux lines of the gap 1 C at the same lateral distance from the gap's end. Increasing the gap length (while keeping the shape and structure of the magnetic core parts constant), results in increased distortion of the magnetic flux lines in the vicinity of the gap and increases the gap's leakage flux. Reducing the gap length reduces core losses due to reduced leakage flux in the gap. Typically, increasing the gap length reduces the core's magnetic permeability and increase the ability of the core to sustain DC bias fields.

Generally, large magnetic flux distortions (or large curvature of the magnetic field lines) out of the gap or in the vicinity of a gap are undesirable because they tend to increase power losses due to interference with current flow in the current bearing coils and other effects which decrease the efficiency of the device (such as, for example a transformer, choke or other magnetic devices using coils with magnetic cores) by increasing total energy losses. Another undesired effect of increasing the gap length is the increasing of the effective cross sectional area available for the magnetic flux to include regions in the vicinity of the gap. This effect of increasing the effective cross sectional area available for magnetic field lines which results in undesirable reduction of the magnetic inductance of the core, as explained in detail hereinafter with respect to Figs. 18-21.

As the distortion of the magnetic flux lines and the resulting leakage flux contributes to the total power losses during operation of the device (such as, the choke or the transformer or any other device including the core), when all other parameters are held equal, the energy losses (and core inductance losses due to the distortion will become smaller as the individual gap lengths becomes smaller.

To the best of knowledge of the inventor, prior art magnetic cores and magnetic core design methods ignored this phenomenon, because the standard techniques used in magnetic core design are based on using the limited number of existing standard shapes of commercially available ferrite core parts. Thus the total number of gaps in prior artmagnetic core designs never exceeds four gaps, as such numbers of gaps were never realized in actual core designs and the effect was not noticed or measured by those skilled in the art. Furthermore, one of the reasons for which prior art designs for magnetic cores do not teach more than four gaps in closed path magnetic cores having gaps is the fact that increasing the number of gaps beyond four gaps complicates the assembly and alignment of the larger number of core parts required to form a magnetic core having more than four gaps, which increases the total core cost and the achievable part alignment and assembly tolerances. In direct contrast, this reduction in operating energy losses due to smaller gap length resulting from increasing the number of gaps in the magnetic core was used by the inventor to even further reduce the total energy losses of multi-gap cores having more than four gaps. The Inventor of the present devices and methods has surprisingly noticed, that if the number of gaps in a magnetic core is increased while keeping the total gap length (the combined length of all the gaps in the core) of the magnetic core constant, the energy losses of each gap (and, therefore, of the entire device using the core) become smaller as the number of gaps increases. This is due to the reduction of the energy losses contributed by each gap as the individual gaps become narrower when more gaps are used but the total gap length is held constant. This results from the increased magnetic inductance of the narrower gaps.

Moreover, it was found that increasing the number of gaps in the core, allows the designer to even further reduce the total gap length while still keeping the device's specified inductance close to the specified value on which the design is initially based.

The inventor of the present invention has performed a series of experiments to investigate the effect of increasing the number of gaps in the ferrite core of the transformer or choke to investigate the possibility of the overall effect of increasing the number of gaps on the total power losses of the transformer (or choke) at high currents. A series of experiments were conducted in which chokes were constructed with two, four, six, eight and ten gaps in their ferrite cores. Reference is now made to Figs. 1A-1C, 2A- 2D, 3A-3C and 4A-4C. Fig. 1A is a schematic isometric view illustrating a prior art ferrite core of a choke or transformer having two gaps. Fig. IB is a side view of the prior art ferrite core of Fig.1 A. Fig. 1C is a cross sectional view of the prior art ferrite core of Fig. IB taken along the lines I-I. Fig. 2A is a schematic isometric view illustrating a prior art choke or transformer having a ferrite core with four gaps and illustrating the choke coils. Fig. 2B is a side view of the prior art choke of Fig.2A. Fig. 2C is a cross sectional view of the prior art choke of Fig. 2B taken along the lines II-II. Fig. 2D is a cross sectional view of the prior art choke of Fig. 2B taken along the lines III-III. Fig. 3A is a schematic isometric view illustrating a ferrite core of a choke or transformer having six gaps in accordance with an embodiment of the ferrite cores of the present application. Fig. 3B is a side view of the ferrite core of Fig.3A. Fig. 3C is a cross sectional view of the ferrite core of Fig. 3B taken along the lines IV-IV. Fig. 4A is a schematic isometric view illustrating a ferrite core of a choke or transformer having eight gaps in accordance with another embodiment of the ferrite cores of the present application. Fig. 4B is a side view of the ferrite core of Fig.4A. Fig. 4C is a cross sectional view of the ferrite core of Fig. 4B taken along the lines V-V. Fig. 5A is a schematic isometric view illustrating a magnetic core of a choke or transformer having ten gaps in accordance with another embodiment of the ferrite cores of the present application. Fig. 4B is a side view of the ferrite core of Fig.4A. Fig. 4C is a cross sectional view of the ferrite core of Fig. 4B taken along the lines V-V.

It is noted that in Figs. 1A-1C, 3A-3C and 4A-4C, 5A-5C only the magnetic core of the choke or transformer or other magnetic devices is illustrated, while the electrical coils of the device are not shown for the sake of clarity of illustration. In Figs. 2A-2D, the coils of the device and their relation to the magnetic core of the device are shown.

Turning to Figs 1A-1C, the ferromagnetic core 10 includes two U-shaped magnetic members 3A and 3B. The members 3A and 3B are disposed facing each other in a spaced apart manner, such that two gaps 12A and 12B are formed between the members 3A and 3B. Spacer members 6A and 6B are disposed within the gaps 12A and 12B, respectively. The spacer members 6A and 6B may also be attached or glued to the U-shaped members 3 A and 3B, but this is not obligatory. The spacer members 6 A and 6B may be formed from any suitable electrically isolating materials such as, but not limited to Mylar®, an engineering plastic material, a polymer based material, paper, cardboard and the like, as is known in the art. The spacer members 6C and 6D are useful for keeping the gap length GL constant and may assist in keeping proper alignment of the magnetic members 3 A and 3B.

Alternatively, if the material between the U-shaped magnetic members 3A and 3B is air, the gaps 6C and 6D are called air gaps (such air gaps may be used if there is another structure or housing (not shown) which may holds the magnetic members fixed relative to each other. It is noted that while in all the magnetic structures illustrated in Figs. 1A-1C, 2A-2D, 3A-3C and 4A-4C, the gaps in the magnetic cores have spacers disposed therein, it is possible to implement the multi-gap cores of the present application by using air gaps by using suitable structures or housings to support and maintain the position of the various parts of the magnetic cores used. The core 10 (as well as all the other ferrite cores disclosed in the present application) is made from N87 soft ferrite (commercially available from Apcot, Siemens, Germany) or from P3 soft ferrite (commercially available from Changshu Xingda Electronics Ltd. Jiangsu, China). It is noted that the above described soft ferrite types gave similar results. It is further noted that the selection of the specific ferromagnetic material was done taking into account the operating AC frequency of the choke (as both of the above disclosed soft ferrite types are usable for operating at AC frequencies of approximately 20-30 KHz). However, other types of ferromagnetic material may be used as is known in the art depending, inter alia, on the device's operating AC range and other considerations.

The dimensions of the magnetic core 10 were as follows:

The total core length C = 138 ± 2 mm.

The core width F = 74 ± 2 mm.

The core height E = 75.5 ± 2 mm.

The core 10 has a rectangular cross-sectional shape having a width V and a height

E. V = 28.0mm and E = 75.5mm (with a manufacturing tolerance of ± 2 mm). The crosssectional area of the core 10 is 2114 mm .

It is noted that the dimensions C, F, E and V of all the assembled ferromagnetic cores disclosed in the present application and illustrated in drawing Figures - were kept identical (within the manufacturing tolerances of ± 2 millimeter) in order to compare the differences of the number of gaps (and of the gap distribution in the cores illustrated in Figs. ). However, the gap length GL was different in different cores.

The core 10 is a prior art closed path magnetic core. And was made for comparing its magnetic properties with the magnetic properties of the novel magnetic cores illustrated in Figs 2A-2D, 3A-3C, 4A-4C and 5A-5C.

In the magnetic core 10, the total gap length GLT = 11.5mm and the gap length of each of the two gaps 12A and 12B is approximately 5.75 mm (GLi = GLT/2).

Turning to Figs 2A-2D, the choke 30 includes a ferromagnetic core 20 including four Lshaped magnetic members 25A, 25B, 25C and 25D. The magnetic members 25A, 25B, 25C and 25D are disposed in a spaced apart manner, such that four gaps 22A, 22B, 22C and 22D are formed between the members 25A, 25B, 25C and 25D. Spacer members 26A, 26B, 26C and 26D are disposed within the gaps 22A, 22B, 22C and 22D, respectively. The spacer members 26A, 26B, 26C and 26D may also be attached or glued to the L-shaped members 25A, 25B, 25C and 25D, but this is not obligatory. The spacer members 26A, 26B, 26C and 26D are formed from Mylar® but may be made from any other suitable electrically insulating material having suitable thermal and mechanical properties, such as, but not limited to, glass-fiber reinforced PBT (Polybutylene Terephtalate) and the like. The spacer members 26A, 26B, 26C and 26D are useful for keeping the gap length of the gaps 22A, 22B, 22C and 22D constant and may assist in keeping proper alignment of the magnetic members 25A, 25B, 25C and 25D. It is noted that while the magnetic core 20 uses spacer members disposed within the gaps, air gaps may also be used as disclosed hereinabove.

The choke 30 also includes two coil holding members 27A and 27B made from glassfiber reinforced PBT (commercially available from Polyram Ltd, Israel), and two coil assemblies 29A and 29B disposed within the coil holding members 27A and 27B, respectively. Each of the coil assemblies 29 A and 29B includes three separate coils (individual coils are not shown for the sake of clarity of illustration) as the device is to be operated using a three-phasic power source. Each coil of the three separate coils on opposing core arms is electrically connected to a similar coil wound on the opposing core arm to form a contiguous coil pair as is known in the art of three-phasic choke design. Each of the coil pairs (not shown in detail in Figs.2A-2D) includes 2x25=50 windings of insulated copper wire. It is noted that the three-phasic coil design is a specific non- limiting example and all the core design principles of the examples given herein may be equally applied to core designs for use in mono-phasic current choke designs as well as to any type of transformer designs. Thus, many different types of coil designs may be used in devices having the magnetic cores of the present application. It is also noted that the type, structure and construction of the coils of chokes and transformers is well known in the art, is not the subject matter of the present application and is therefore not discussed in detail hereinafter.

The dimensions of the ferromagnetic core 20 are the same as the dimensions of the core 10 except that the total gap length GLT = 10.5mm and the gap length of each of the four gaps 22A, 22B, 22C and 22D is approximately 2.625 mm (GLi = GLT/4). The gaps 22B and 22D are positioned symmetrically with respect to a first axis of symmetry 21 of the magnetic core 20 and the gaps 22 A and 22C are positioned symmetrically with respect to a second axis of symmetry 23 of the magnetic core 20 (see Fig. 2C). The first axis of symmetry 21 and the second axis of symmetry 23 are orthogonal. This symmetrical gap arrangement is the preferred arrangement of gaps in magnetic cores including four gaps of the present application, as is explained in detail hereinafter with respect to Fig. 17.

Turning to Figs. 3A-3C the magnetic core 50 includes four L-shaped magnetic member 55A, 55B, 55C and 55D and two straight bar-like members 51 A and 5 IB. The magnetic members 55A, 55B, 55C, 55D, 51A and 51B all have an identical generally rectangular cross-sectional shape. The magnetic members 55A, 55B, 55C, 55D, 51A and 5 IB are disposed in a spaced apart manner, such that six gaps 52A, 52B, 52C 22D, 52E and 52F are formed in the magnetic core 50. Spacer members 56A, 56B, 56C 56D, 56E and 56F are disposed within the gaps 52A, 52B, 52C 52D, 52E and 52F, respectively. The spacer members 56A, 56B, 56C 56D, 56E and 56F may also be attached or glued to the magnetic members contacting them, but this is not obligatory. The spacer members 55A, 55B, 55C, 55D, 51 A and 5 IB are formed from Mylar®. It is noted that while the magnetic core 50 uses spacer members disposed within the gaps, air gaps may also be used as disclosed hereinabove.

The dimensions of the ferromagnetic core 50 are the same as the dimensions of the cores 10 and 20 except that in the magnetic core 50, the total gap length GLT = 9.5mm and the gap length of each of the six gaps 52A, 52B, 52C 22D, 52E and 52F is approximately 1.583 mm (GLi = GLT/6).

The gaps 52A, 52B, 52C 22D, 52E and 52F are positioned symmetrically with respect to a first axis of symmetry 51 of the magnetic core 50 and with respect to a second axis of symmetry 53 of the magnetic core 50 (see Fig. 3C). The first axis of symmetry 51 and the second axis of symmetry 53 are orthogonal. This symmetrical gap arrangement is a preferred arrangement of gaps in a preferred arrangement of gaps in the magnetic cores including six gaps of the present application, as is explained in detail hereinafter.

Turning to Figs. 4A-4C the magnetic core 60 includes four L-shaped magnetic member 65A, 65B, 65C and 65D and four straight bar-like members 61A, 61B, 61C and 6 ID. The magnetic members 65A, 65B, 65C, 65D, 61 A, 6 IB, 61 C and 61D all have an identical generally rectangular cross-sectional shape. The magnetic members 65A, 65B, 65C, 65D, 61A, 61B, 61C and 61D are disposed in a spaced apart manner, such that eight gaps 62A, 62B, 62C 62D, 62E, 62F, 62G and 62H are formed in the magnetic core 60. Spacer members 66A, 66B, 66C, 66D, 66E, 66F, 66G and 66H are disposed within the gaps 62A, 62B, 62C 62D, 62E, 62F, 62G and 62H, respectively. The spacer members 66A, 66B, 66C, 66D, 66E, 66F, 66G and 66H may be attached or glued to the magnetic members contacting them, but this is not obligatory. The spacer members 66A, 66B, 66C 66D, 66E, 66F, 66G and 66H are formed from Mylar®. It is noted that while the magnetic core 60 uses spacer members disposed within the gaps, air gaps may also be used as disclosed hereinabove.

The dimensions of the ferromagnetic core 60 are the same as the dimensions of the cores 10, 20 and 50, except that in the magnetic core 60, the total gap length GLT = 9.0 mm and the gap length of each of the eight gaps 62A, 62B, 62C 62D, 62E, 62F, 62G and 62H is approximately 1.125 mm (GLi = GLT/8).

The gaps 62A, 62B, 62C 62D, 62E, 62F, 62G and 62H are positioned symmetrically with respect to a first axis of symmetry 61 of the magnetic core 60 and with respect to a second axis of symmetry 63 of the magnetic core 60 (see Fig. 4C). The first axis of symmetry 61 and the second axis of symmetry 63 are orthogonal. This symmetrical gap arrangement is a preferred arrangement of gaps in a preferred arrangement of gaps in the magnetic cores including eight gaps of the present application, as is explained in detail hereinafter.

Turning to Figs. 5A-5C the magnetic core 70 includes four L-shaped magnetic members 75A, 75B, 75C and 75D and six straight bar-like members 71A, 11B, 71C, 71D, 71E and 71F. The magnetic members 75A, 75B, 75C, 75D, 71A, 71B, 71C, 71D, 71E and 7 IF all have an identical generally rectangular cross-sectional shape. The magnetic members 75A, 75B, 75C, 75D, 71A, 71B, 71C, 71D, 71E and 71F are disposed in a spaced apart manner, such that ten gaps 72A, 72B, 72C, 72D, 72E, 72F, 72G and 72H are formed in the magnetic core 70. Spacer members 76A, 76B, 76C 76D, 76E, 76F, 76G and 76H are disposed within the gaps 72A, 72B, 72C 72D, 72E, 72F, 72G and 72H, respectively. The spacer members 76A, 76B, 76C 76D, 76E, 76F, 76G and 76H may be attached or glued to the magnetic members contacting them, but this is not obligatory. The spacer members 76A, 76B, 76C 76D, 76E, 76F, 76G and 76H are formed from Mylar®. It is noted that while the magnetic core 70 uses spacer members disposed within the gaps, air gaps may also be used as disclosed hereinabove.

The dimensions of the ferromagnetic core 70 are the same as the dimensions of the cores 10, 20, 50 and 60, except that in the magnetic core 70, the total gap length GLT= 9.0mm and the gap length of each of the ten gaps 72A, 72B, 72C 72D, 72E, 72F, 72G and 72H is approximately 0.90 mm (GLi = GLT/10).

The gaps 72 A, 72B, 72C 72D, 72E, 72F, 72G and 72H are positioned symmetrically with respect to a first axis of symmetry 71 of the magnetic core 10 and with respect to a second axis of symmetry 73 of the magnetic core 70 ( see Fig. 5C). The first axis of symmetry 71 and the second axis of symmetry 73 are orthogonal. This symmetrical gap arrangement is a preferred arrangement of gaps in a preferred arrangement of gaps in the magnetic cores including eight gaps of the present application, as is explained in detail hereinafter.

It is noted that the ferromagnetic cores 20, 50, 60 and 70 of Figs. 2A-5A all include four L-shaped ferromagnetic members as disclosed hereinabove and illustrated in Figs. 2A-5A. The ferromagnetic cores 50, 60 and 70 also include, in addition to the four L-shaped ferromagnetic members, two, four, and six straight bar-like ferromagnetic members, respectively. As the overall dimensions (C and F) of the entire cores are kept constant, while the individual gap length GLi and the total gap length GLT vary for the different cores 50, 60 and 70, the dimensions of the L-shaped ferromagnetic members and of the straight member vary for different cores 50, 60 and 70.

Reference is now made to Fig. 7 which is a schematic cross sectional diagram illustrating the dimensions of the L-shaped members and straight bar-like members of the magnetic cores of the present application. The bar-like member 91 (shown in cross- section has a length S and a width D. The L-shaped member 92 has a width D, a long arm length W and a short arm length K.

TABLE 1 includes the dimensions (with respect to the dimension notations illustrated in Fig. 7) of the L-shaped magnetic members and the straight magnetic members (if any) included in the magnetic cores 20, 50, 560 and 70 having 4, 6, 8 and 10 gaps, respectively.

TABLE 1

It is noted that in each of the cores 20, 50, 60 and 70 all of the L shaped members included in a core have identical dimensions (within the manufacturing tolerances of typically ± 2 mm). It is also noted that in each of the cores 50, 60 and 70, all of the straight bar-like members have identical dimensions (within the manufacturing tolerances of typically ± 2 mm). For example, it may be seen from TABLE 1 that the dimensions W, K and D of each of the four L-shaped members 55A, 55B, 55C and 55D of the core 50 (of Figs. 3A-3C) are : W=60mm, K=37mm and D=28mm, and that the dimensions S and D of each of the two straight bar-like members 51A and 5 IB are S=16mm and D= 28mm. Similarly, the dimensions W, K and D of each of the four L-shaped members 75A, 75B, 75C and 75D of the core 70 (of Figs. 5A-5C) are : W=52mm, K=29mm and D=28mm, and that the dimensions S and D of each of the six straight bar- like members 71 A, 7 IB, 71 C, 71D, 71E and 7 IF are S=16mm and D= 28mm. The height E of each of six straight barlike members 71A, 71B, 71C, 71D, 71E and 71F is E= 75.5mm ( it is noted that all the dimensions have a tolerance of ±2mm).

The differences in the dimensions W and K of the L-shaped members of Different cores are due to the fact that one or two of the L-shaped member's arms is shortened to accommodate the length of the straight bar-like member(s) used for providing the necessary number of gaps.

Inductance Loss Measurements

Five different chokes were constructed from the magnetic cores 10, 20, 50, 60, 70 having two, four, six, eight and ten gaps , respectively ( as illustrated in Figs 1 A, 2A, 3A, 4A , 5A, respectively). The choke coils were identical to the coils 29A and 29B (of Fig. 2D) in all of the five constructed chokes except that for testing of inductance losses only one coil pair was wound on the magnetic cores 10, 20, 50, 60, 70 and a mono-phasic test setup was used.

The inductance of the constructed chokes was measured at various coil currents. For each constructed choke, the inductance was measured at several different current levels. Reference is now made to Fig. 6, which is a schematic diagram illustrating the testing configuration used for testing the inductance decrease as a function of coil current of the different magnetic cores illustrated in Figs. 1A-5A and 15. The test circuit 80 included a variable voltage source 81 (VARIAC, operated at 220V mains at 50Hz) connected to a stepdown transformer 82 (220V/24V) for providing high currents. A shunt resistor R (50ιηΩ, ± 5 %, rated for currents of 50A) is connected to one terminal of the step-down transformer 82 and to one terminal of the coil of the (serially connected) coils LI and L2 of the choke being tested (The ferromagnetic core of the tested coil is not shown in the schematic symbol notation of Fig. 6, for the sake of clarity of illustration). The test circuit 80 is closed by connecting the second terminal of the coil L2 of the tested choke's coil pair to the remaining terminal of the step-down transformer 82.

The voltage across the terminals of the resistor R is measured by a voltmeter 83 and the voltage across the first and second coils LI and L2 of the tested choke is measured by a voltmeter 84 as illustrated. The testing is done by measuring the voltage across the tested choke's terminals during brief current pulses at specific currents. The measurements are brief in order to avoid substantial heating of the shunt resistor R.

The voltage drop measured across the resistor R was 2.5 mV per 5A current, such that the current flowing through the coils could be adjusted by adjusting the voltage of the variac 81 to obtain a desired current level (for example, to obtain a current level of 10A, the variac 81 was adjusted such that the voltage measured by the voltmeter 83 was 5 mV. To obtain a current level of 20A, the variac 81 was adjusted such that the voltage measured by the voltmeter 83 was 10 mV, and so forth).

The inductance L at each specific current I was calculated from the following equation: L=V/(2nfI)

Wherein,

L is the inductance (in mH)

V is the voltage measured by the voltmeter 84

I is the test circuit current, and

f is the frequency of AC provided by the step-down transformer 82

(which was fixed at 50Hz)

Each of the mono-phasic chokes including the cores 10, 20, 50, 60, 70 and 150 was tested at currents of 5, 10, 15, 20, 25, 30, 35, 40 and 45 ampere and the inductance values were calculated as described above.

Reference is now made to Figs. 9-13 which are schematic graphs illustrating the results of measurements performed to determine the choke inductance values at various different current intensities flowing through the choke's coil for chokes with magnetic cores with two, four, six, eight and ten gaps, respectively.

In each of the graphs of Figs. 9 - 13, the vertical axis represents the Inductance

(in mH) and the horizontal axis represents the number of gaps in the magnetic core of the choke. The curves 100, 104, 108, 112 and 116 of Figs, 9, 10, 11, 12, and 13, respectively, represent the changes in Inductance (L) of the tested magnetic core as a function of the coil current. Each of the filled rhomboidal symbols represents the measured value of the inductance at a given coil current. For each measurement, the number below the rhomboidal symbol represents the measured inductance value and the number above the rhomboidal symbol represents the computed percent change of the inductance relative to the inductance measured for the magnetic core at a current level of 5 ampere. This number represents the percent inductance loss (the negative sign of the computed represents an inductance loss relative to the measured inductance measured at a 5 ampere current).

It should be noted that the percent change values are used to normalize the data against the inductance value measured at a current of 5 ampere because for each different core structure the inductance at a current 5 ampere is slightly different in different core structures.

Reference is now made to Fig. 14 which is a schematic graph illustrating the dependence of the % inductance loss values on the number of gaps included in the magnetic core for each of the tested chokes having the magnetic cores of Figs 1 A, 2A, 3A, 4A and 5 A, as measured at the current values of 45, 40, 35, 30, 25, 20, 15 and 10 ampere for each choke.

In the graph of Fig. 14, the vertical axis represents the absolute value of the percent inductance loss (as computed by normalizing the results as described hereinabove) and the horizontal axis represents the number of gaps in the cores used in the measurements.

The graph of Fig. 14 includes eight different curves 120, 122, 124, 126 128, 130 132 and 134. Each curve represents the dependence of the percent inductance loss on the number of gaps in the magnetic core at a specific coil current value. The curve 120 (having filled triangular symbols) represents the dependence of the percent inductance loss on the number of gaps in the magnetic core as measured at a coil current value of 10 ampere. The curve 122 (having hollow triangular symbols) represents the dependence of the percent inductance loss on the number of gaps in the magnetic core as measured at a coil current value of 15 ampere. The curve 124 (having filled circular symbols) represents the dependence of the percent inductance loss on the number of gaps in the magnetic core as measured at a coil current value of 20 ampere. The curve 126 (having hollow square symbols) represents the dependence of the percent inductance loss on the number of gaps in the magnetic core as measured at a coil current value of 25 ampere. The curve 128 (having filled triangular symbols) represents the dependence of the percent inductance loss on the number of gaps in the magnetic core as measured at a coil current value of 30 ampere. The curve 130 (having hollow circular symbols) represents the dependence of the percent inductance loss on the number of gaps in the magnetic core as measured at a coil current value of 35 ampere. The curve 132 (having hollow sqaure symbols) represents the dependence of the percent inductance loss on the number of gaps in the magnetic core as measured at a coil current value of 40 ampere. The curve 134 (having filled square symbols) represents the dependence of the percent inductance loss on the number of gaps in the magnetic core as measured at a coil current value of 45 ampere.

When comparing the curves it should be noted that the value of the total gap length systematically decreases as the gap number increases (except for the cores having 8 and 10 gaps for which the total gap length remains at 9 mm, but the single gap length decreases from 1.125 mm to 0.9 mm. The total gap length GLT (in millimeter) and the single gap length GLi (in millimeter) for magnetic cores having different gap numbers are given in TABLE 2.

TABLE 2

The results of the measurements indicate that for the current value of 45 ampere, increasing the gap number from two to six gaps monotonically decreases the normalized inductance loss by more than 4%, while further increasing the number of gaps to eight gaps increase the normalized inductance loss to a value higher than the normalized inductance loss at six gaps. Increasing the number of gaps to ten gaps, results in further decreasing of the normalized inductance loss to a value lower than the value at six gaps. The reason for the inflection of the curve and the minimum inductance loss at six gaps per core is not clear.

At lower currents the curves "flatten out" gradually and the differences in normalized inductance loss between cores with different numbers of gaps is not as large as those measured at high currents.

Thus increasing the number of gaps to a number larger than four gaps (such as for example, six gaps, eight gaps ten gaps, and the like), allows the designer of the core to decrease the individual gap lengths which reduces the core inductance by decreasing the magnetic impedance. This, in turn enables the designer to even further reduce the total gap length (without changing the number of gaps) by even further reducing the individual gap length to keep the core inductance closer to the initial design specification of the magnetic inductance value.

The inventor also noted that in magnetic cores having a generally rectangular shape with four corners, it is also possible to reduce inductance losses of magnetic cores having four gaps by using symmetrical gap positioning and by avoiding placement of the gaps at or near the corners of the magnetic core.

Reference is now made to Fig. 15-17. Fig. 15 is a schematic isometric view illustrating a prior art generally rectangular magnetic core having four gaps. Fig. 16 is a cross sectional view of the magnetic core of Fig. 15, taken along the lines XVI-XVI, schematically illustrating the approximate magnetic flux lines in the vicinity of the core's gaps. Fig. 17 is a cross-sectional view of a core 20A which is similar to the core 20 of Fig. 2A in having the same core length, core width and core height as core 20 of Fig. 2A, but has a total gap length of 9mm and an individual gap length of 2.25mm (GLT = 9mm and GLi = 2.25mm), schematically illustrating the approximate magnetic flux lines in the vicinity of the core's gaps.

Turning to Fig. 15, the magnetic core 150 includes four rectangular bar-like magnetic members 152A, 152B, 154A and 154B. The magnetic core 150 includes four gaps 153A, 153B, 153C and 153D. Spacer members 155A, 155B, 155C and 155D are disposed within the gaps 153A, 153B, 153C and 153D, respectively for mechanical stabilization of the spacing of the bar-like magnetic members, 152A, 152B, 154A and 154B. The gaps 153A, 153B, 153C and 153D are positioned close to the four corners of the core 150. The dimensions of the bar like members 152A and 152B are 74x75.5x28mm. The dimensions of the bar like members 154A and 154B are 80x75.5x28mm. The individual gap length of each of the four gaps 153A, 153B, 153C and 153D is GLi =2.75mm and the total gap length is GLT = 11mm. The cross-sectional area of each of the bars 152A 153B, 154A andl54B is 2114mm ² , which is identical (within manufacturing tolerances) to the cross sectional area of the magnetic cores 10, 20, 50, 60 and 70 of Fig. 1 A-5A, respectively. The overall length, width and height of the assembled core 150 are identical (within manufacturing tolerances) to the overall lengthy width and height (C, F and E, respectively) of the magnetic cores 10, 20, 50, 60 and 70 (of Fig. 1A-5A).

Turning to Fig. 16 and 17, the approximate distribution of the magnetic flux lines in the vicinity of the gaps 153 A, 153B, 153C and 153D of the magnetic core 150 is shown in Fig. 16 and the approximate distribution of the magnetic flux lines in the 5 vicinity of the gaps 22A, 22B, 22C and 22D is shown in Fig. 17. (It is noted, that all the representations of the distribution of the magnetic flux lines included in the drawing Figures of the present application are highly schematic and are intended to be used for rough qualitative comparison purposes only).

It may be seen that the distribution and curvature of the magnetic flux lines in the vicinity of the gaps 22A, 22B, 22C and 22D is different than the distribution and curvature of the magnetic flux lines in the vicinity of the gaps 153 A, 153B, 153C and 153D. In particular, the differences is quite substantial on the regions in the vicinity of the gaps 153 A, 153B, 153C and 153D facing the four internal corners of the core 150, where the magnetic flux lines are different ( the flux lines are asymmetrically distributed on the two sides of the gaps). This difference in flux lines distribution between the two cores increases the cross-sectional area available for flux lines flow near the gaps 153 A, 153B, 153C and 153D, as compared to the cross-sectional area available for flux lines flow near the gaps 22A, 22B, 22C and 22D, resulting in lowering of the core magnetic impedance in the core 150 as compared to the magnetic impedance of the core 20. The increased magnetic flux leakage near the gaps of the core 150 results in decreasing of the inductance relative to the inductance of the core 20 (at the same choke coils current level) with concomitant increase in power losses.

Reference is now made to Figs. 18-21. Figs. 18-21 are schematic cross-sectional views illustrating the approximate magnetic flux lines distribution in four different gaps disposed in a part of a magnetic core. In Fig. 18, the gap 102A in the magnetic core 102 (only part of which is shown in Fig. 18), is a relatively small gap with relatively low distortion of the magnetic flux lines 103 outside the gap 102A and relatively small magnetic flux loss, due to the small effective magnetic cross-sectional area available for magnetic flow. The effective magnetic cross-sectional area is schematically indicated by the bold lines 102E and 102F and is well approximated by the actual cross sectional area of the magnetic core 102.

It is noted that as a first approximation, the inductance L of a magnetic core having a gap is proportional to Nc ² , μο and A, and is inversely proportional to GL (wherein Nc is the number of coil turns, μο is a constant representing the permeability of the core, A is the core's effective magnetic cross-sectional area, and GL is the gap length). Therefore increasing GL will decrease the core inductance L and decreasing GL will increase the core inductance L. Increasing A will increase the cores inductance L and decreasing A will decrease the core inductance L.

In Fig. 19, the gap 106A in the magnetic core 106 (only part of which is shown in Fig.19), is a gap slightly larger than the gap 102A (of Fig 18). The distortion of the magnetic flux lines 105 outside the gap 102A is slightly higher than the distortion of the magnetic flux lines 103 and the magnetic flux loss of the gap 106A while still small is slightly larger than the magnetic flux loss in the gap 102A due to a slight increase in the effective magnetic cross-sectional area available for magnetic flux. The effective magnetic cross-sectional area is schematically indicated by the bold lines 106E and 106F. It is noted that in the gap 106A the effective magnetic cross-sectional area increases somewhat by spreading to the sides of the magnetic core which is schematically illustrated by the short bold lines orthogonal to the surfaces of the gap 106A which is larger than the actual cross sectional area of the magnetic core 106.

It is noted that the effective magnetic cross-sectional area of a gap is not a linear function of the gap length (GLi). For small gaps, the decrease in the effective magnetic cross sectional area as the gap length decreases is therefore smaller than the decrease in the effective magnetic cross-sectional area resulting from decreasing of a larger gap by the same length.

In Fig. 20, the gap 1 10A in the magnetic core 1 10 (only part of which is shown in Fig. 20), is a gap substantially larger than the gap 106A (of Fig. 19). The distortion of the magnetic flux lines 107 outside the gap 1 1 OA is substantially higher than the distortion of the magnetic flux lines 103 and the magnetic flux loss of the gap 1 1 OA is substantially larger than the magnetic flux loss in the gap 106A due to a significant increase in the effective magnetic cross-sectional area available for magnetic flow. The effective magnetic cross-sectional area is schematically indicated by the bold lines 110E and 1 10F. It is noted that in the gap 1 10A the effective magnetic cross-sectional area available for the magnetic flux lines 107 is substantially increases by spreading to the sides of the magnetic core which is schematically illustrated by the longer bold lines orthogonal to the surfaces of the gap 11 OA which is now substantially larger than the actual cross sectional area of the magnetic core 110.

In Fig. 21, the gap 114A in the magnetic core 114 (only part of which is shown in Fig.21), has a gap length equal to the gap length of the gap 106A (of Fig. 19).

However, the gap 114A is different than the gap 106 A because the gap 114A is disposed in a corner formed by the parts 114B and 114C of the magnetic core 114. Due to this corner position of the gap 114A, the distortion of the magnetic flux lines 109 A in one region in the vicinity of the gap 114A is different than the distortion of the magnetic flux lines 109B in a different region in the vicinity of the gap 114A as seen in Fig. 21. The distortion of the magnetic flux lines 109 A and 109B near the gap 114A is substantially higher than the distortion of the magnetic flux lines 105 near the gap 106A and the magnetic flux loss of the gap 114A is larger than the magnetic flux loss in the gap 106A due to a significant increase in the effective magnetic cross-sectional area available for magnetic flow. The effective magnetic cross-sectional area is schematically indicated by the bold lines 109E and 109F. It is noted that in the gap 114A the effective magnetic crosssectional area available for the magnetic flux lines 109 A and 109B substantially increases by spreading to the sides of the magnetic core which is schematically illustrated by the total length of the bold lines 109F and 109E which is substantially larger than the length of the bold lines 106F and 106E of the gap 106A.

The inventor has therefore serendipitously discovered that it is possible to improve the performance of a magnetic core as compared to the performance of prior art magnetic cores by increasing the number of individual gaps in the core above four gaps which results in each of the resulting individual gaps having a smaller effective magnetic cross-sectional area the individual gap length. Moreover, because as each of the resulting gaps has a smaller effective magnetic cross-sectional area (due to its having an individual gap length GLi which is substantially smaller than the total gap length GLT of the core), the total gap length may be even further decreased to compensate for the inductance decrease resulting from the reduction of the effective magnetic cross-sectional area due to the increase of the number of gaps. Thus, the individual gap length may be narrower than the initial value of GLT/N computed in step 232 of the design method illustrated in Fig. 25. Furthermore, The inventor has discovered that it is possible to further improve the performance and reduce the inductance losses of prior art magnetic cores having four gaps by carefully arranging the gaps in a special way such that they are positioned in the middle of each of the four sides of the core to minimize the effective magnetic cross sectional area of each of the gaps as compared to the prior art cores in which the four gaps are positioned close to the core's corners.

While the above effect is not large (when the percent inductance loss is calculated), the performance improvement in devices operating at the upper range of their nominal currents at high current levels (several ampere to several hundred ampere or greater are typical examples) may be quite substantial and is advantageous. It is concluded that the gap positioning within the magnetic cores of the present application is important and that it is generally desired to place as many gaps as possible away from corners of the core when rectangular cores are being designed. It is therefore clear from the above examples of cores having four gaps that in a magnetic core having a rectangular or square cross-section and four gaps it is preferred and advantageous to arrange each of the four gaps in the middle of the respective side of the rectangular or square core.

In accordance with one preferred embodiment of a rectangular magnetic core of the present application, the core is assembled from four identical L-shaped magnetic members (not Shown). The arms of each of the L-shaped members have equal length. The four identical L-Shaped members are symmetrically arranged with respect to each other and with respect to two orthogonal axes of symmetry of the magnetic core, such that the four gaps separating the ends of the L-shaped members are arranged in two pairs of opposing gaps and the sides of each gap are equidistant from the axis of symmetry passing through the gap pair.

In accordance with another preferred embodiment of a rectangular magnetic core of the present application, the core is assembled from four identical L-shaped magnetic members (not Shown). The arms of each of the L-shaped members have 5 different lengths. The four identical L-Shaped members are symmetrically arranged with respect to two orthogonal axes of symmetry of the magnetic core, such that the four gaps separating the ends of the Lshaped members are arranged in two pairs of opposing gaps and the sides of each gap are equidistant from the axis of symmetry passing through the gap pair.

Reference is now made to Figs. 22, 23 and 24. Fig 22 is a schematic graph illustrating the dependence of the inductance on the current flowing through the coils of a choke having the prior art four gap ferrite core of Fig.15. Fig 23 is a schematic graph illustrating the dependence of the inductance on the current flowing through the coils of a choke having the four gap ferrite core illustrated in the cross-sectional view of Fig.17. Fig. 24 is a schematic graph illustrating a comparison of the dependence of the inductance on the current flowing through the coils of a choke having the four gap ferrite core illustrated in the cross-sectional view of Fig. 17, when the Total gap length is decreased from 11 mm to 9 mm.

The horizontal and vertical axes of the graphs of Figs 22 -24 are as described in detail hereinabove for the graphs of Figs. 9-13 hereinabove. Similar to the notation of Figs. 9-13, the curves 11 1, 113, and 118 and 120 of Figs. 22, 23 and 24, respectively, represent the changes in Inductance (L) of the tested ferromagnetic core (in mH) as a function of the coil current (in Ampere). In Figs. 22 and 23, each of the filled rhomboidal symbols of the curves 11 1 and 113, represents the measured value of the inductance (L) at a given coil current. For each measurement, the number below the rhomboidal symbol represents the measured inductance value and the number above the rhomboidal symbol represents the computed percent change of the inductance relative to the inductance measured for the magnetic core at a current level of 5 ampere. This number represents the percent inductance loss at currents higher than 5A (the negative sign of the computed represents an inductance loss relative to the measured inductance measured at a 5 ampere current).

In Fig. 24 each of the filled triangular filled symbols of the curve 1 18 and the filled circular symbol of curve 118 represents the measured value of the inductance (L) at a given coil current. For each measurement, the number above the symbol represents the measured inductance value at the specific current value (of the horizontal axis of the graph).

As may be seen from Fig. 24, when the total gap length of the magnetic core 20A illustrated Fig. 17 is equal to the total gap length of the core 150 of Fig. 15 ( GLT = 1 1 mm in both cores), the inductance of the core of Fig. 17 measured at the reference current value of 5 ampere is L=0.818 mH which is much lower than the inductance value of the core 150 of Fig. 15 measured at the reference current value of 5 ampere, which is L=0.95. This occurs because the effective magnetic cross-sectional area of the gaps in the core 150 is significantly larger than the effective magnetic cross-sectional area of the gaps in the core shown in Fig. 17. This increase in the effective magnetic cross-sectional area of the gaps occurs due to the positioning of the gaps of the core 150 near the corners of the core 150, whereas the gaps of the core 20A of Fig. 17 are positioned in the middle of each of the four sides of the core 20A as seen in Fig. 17. The reasons for the difference in the effective magnetic cross-sectional area are explained in detail with respect to Figs. 18-21 hereinabove.

Therefore, the core 20A was reconstructed by slightly increasing the length of the arms of the L-shaped members 25A, 25B, 25C and 25D, to decrease the total gap length and the individual gap length by using the method illustrated in Fig. 25 hereinafter. In the end of the procedure, the total gap length was decreased to GLT = 9mm and the individual gap length was GLi = 2.25mm.

The curves 118 and 120 of Fig. 24 represent the actual resultsof inductance measurements performed at different currents in two different chokes constructed according to the structure illustrated in Fig 17. The ferromagnetic cores of the different chokes had the same total external dimensions identical to the external dimensions of the core 20 of Fig. 2A, except that the total gap length GLT of one core was 11mm (the inductance measurement of the choke including this core are represented by the curve 120) and the total gap length GLT of the other core was 9mm (the inductance measurement of the choke including this core are represented by the curve 118). The total external dimensions (length, width and height) of the core having a total gap length of 9 mm were kept identical to the total external dimensions (length, width and height) of the core having a total gap length of 1 1 mm by increasing the length of the arms of the L- shaped members 25 A, 25B, 25 C and 25D as explained in detail hereinabove.

It may be seen when comparing the curves 11 1 an 113 of Figs 22 and 23, respectively, that the percent inductance loss of the core 20 of Fig. 2A is significantly smaller than the percent inductance loss of the core 150 of Fig. 15 in the current range of 40A-45A which may result in performance improvement when the device is operating close to the nominal maximal current.

When comparing the inductance losses of the cores as illustrated in Figs 22 and 23, it should be noted that the initial inductance at 5 ampere of the core 150 is slightly higher at L=0.973 mH than the initial inductance at 5 ampere of the core 20A of Fig. 17 which was L=0.95 mH. As the percent inductance losses is calculated relative to the (arbitrarily chosen) reference inductance value at 5 ampere, it is reasonable to assume that had the core 20A been reconstructed to have a reference inductance equal to the reference inductance of the core 150, the results would have been a further improvement and reduction in the percent inductance losses over a wider current range as compared to the inductance percent losses of the core 150.

From the above tests and measurements, it became clear to the inventor that this decrease in the contribution of each gap to the total device energy losses as the individual gap narrows may be used to relax the total gap length of the designed magnetic core which is required to keep the core's magnetic impedance at a value sufficient for maintaining a large enough magnetic impedance to reduce energy losses due to core magnetic saturation.

The inventor realized that by increasing the number of gaps to be larger than four gaps per core, the resulting decrease in each gap's contribution to the energy losses due to the narrowing of the gap length and concomitant decrease in the core's inductance may also be used to further decrease the total gap length GLT of the core without decreasing the core's inductance and the associated total overall device operating energy losses because the decrease in energy losses resulting from the narrower gaps compensates or counteracts any increase in energy losses caused by the higher core magnetic impedance (and lower core inductance) resulting from making the total gap length GLT smaller than the initially computed value of the total gap. Therefore, increasing the number of gaps in the core enables decreasing not only the individual gap length but also the total 5 gap length of the core which is advantageous as it enables keeping the inductance of the device close to the originally specified device inductance as initially designed, while reducing power losses in the core.

Thus, in prior art core design methods, the method for designing a magnetic core was as follows: Receiving the desired set of device parameters including (but not limited to, the required device conductance and operating current range and the device dimension limitations. Computing the total gap length GLT which is required to provide a sufficient core magnetic impedance to prevent magnetic saturation of the core at the maximal nominal device current. Setting the number of gaps N of up to four gaps, and computing the individual gap length from the total gap length GLT and N.

In contrast, in accordance with an embodiment of the method of the present invention, the new design method of the present application is as disclosed in Fig. 25 which is a schematic block diagram illustrating the steps of a method for designing a magnetic core in accordance with an embodiment of the core design methods of the present application.

Reference is now made to Fig. 25 which is a schematic block diagram illustrating steps of a method for designing the multi-gap magnetic cores of the present application in accordance with an embodiment of the methods of the present application.

Turning to Fig. 25, the designer receives the desired set of device parameters including, but not limited to, and operating current range, the device dimensional limitations and the device inductance at the nominal maximal operating current LMAX (step 230) at this stage, the designer may decide on the material from which the core will be constructed and may determine the coils(s) structure the number of coil turns the coil wire parameters and the cross-sectional shape and dimensions of the magnetic core, as is well known in the art of core design . The designer then determines the total gap length GLT required for providing a core magnetic impedance value sufficient for preventing magnetic saturation of the core at the maximal nominal device current, as is known in the art (Step 232). The designer then selects and sets the number of gaps N and computes a starting value for the individual gap length GLi= GLT/N. (Step 234).

It is noted that in accordance with one preferred embodiment 5 of the design method of the present application, the number of gaps N is an even number and is equal to or greater than six (for example, N may be selected as 6 , 8, 10...., and the like). However, in accordance with another embodiment of the method of present application N=4. This embodiment may be useful in cases where a high number of gaps is not desired because of cost considerations but the method may still be useful for optimization of the gap length in order to achieve an inductance which is as close as possible (within the limitations of the value of AL as disclosed hereinafter) to the initially specified inductance design goal of the inductance at the nominal maximal operating current LMAX. When N=4 is used, it may be advantageous to use the novel core structure disclosed in Fig. 2C including four identical L shaped members arranged such that each of the gaps is positioned in the middle of the side of the core in which the gap is disposed, because as shown herein this novel arrangement further reduces the inductance losses and improves the efficiency of operation of the device by reducing flux loss.

Furthermore, while it was shown hereinabove that using an increasingly larger even number of gaps in constructing the core reduces the inductance loss at high currents and results in improvement of device performance at higher operating currents, it is also possible to use an odd number of gaps N>5 and to construct magnetic cores with an odd number of gaps which is equal to or higher than 5. This may be useful in cases where the net reduction in flux losses achieved by using a number of gaps higher than four and the resulting reduction in the Individual gap length, exceeds the increase in flux losses resulting from any asymmetrical gap positions due to the odd gap number.

Thus, the method disclosed herein is applicable for being used for designing any cores having a number of gaps N>4. While in some cases, the selection of odd values for N may be less efficient than selection of an even number of gaps, it is still possible to improve the core efficiency by increasing the number of gaps to a larger even number. For Example, as explained hereinabove, from theoretical considerations, a core having seven or nine gaps which are asymmetrically disposed within the core (or at least asymmetrically disposed on two opposing sides of a rectangular core) which is designed by the method described herein, may still have a lower inductance loss at currents close to the nominal maximal current values than a core having only four gaps each having a larger gap length than the gap length of the gaps in the cores having seven or nine gaps.

Thus, the higher number of gaps in the seven gap core may allow decreasing in the individual gap length and in the resulting total gap length relative to the possible individual gap length and the total gap length allowable by a core design of the core having only four gaps, which may result in the core with seven gaps having a reduced magnetic flux losses and increased efficiency than the core with four gaps despite the reduced symmetry of gap distribution of the seven gap core as compared to the symmetry of the four gap core.

The designer then sets the value of the currently computed gap length GLINEW = GLi (step 236). The designer than computes a new (reduced) value for the currently computed individual gap length GLINEW = GLINEW - AGL , wherein AGL is a predetermined gap length reduction value which is a small fraction of the value of the starting individual gap length GLi , constructs a device with a core having N gaps with an individual gap length of GLINEW and experimentally measures the inductance LEM which is the device's inductance at the nominal maximal operating current (step 238). The designer then computes the absolute value of the difference between LEM and LMAX and compares it's value to a predetermined inductance tolerance value AL (step 240). If the absolute value of the difference between LEM and LMAX is smaller than the predetermined inductance tolerance value ( | LEM - LMAX | < AL), the design procedure is terminated (step 242) and the current values of GLINEW is the individual gap length to be implemented in the final core design. If the absolute value of the difference between LEM and LMAX is not smaller than the predetermined inductance tolerance value, the designer returns to step 238 and recalculates a new value of GLINEW (by subtracting AGL from the current value of GLINEW as explained hereinabove).

The loop continues until | LEM - LMAX | < AL.

It is noted that while the values of AGL and AL are predetermined values which are kept constant within the design procedure described above, their specific values may be different for different devices having different initial design parameters. Typically, but not obligatorily, for an initial value of GLi in the range of 1-3 mm, a value of AGL= 0.25mm may be practically useful in the design method described above. However, this value is not obligatory and other different values of AGL outside this range may be used, depending, inter alia on the cores dimensions, the required final inductance and other device parameters. Decreasing the value of AGL may increase the number of iterations required to satisfy the criterion of Step 240 but may improve the accuracy of the final value of the gap length. Increasing the value of AGL may reduce the number of iterations required to satisfy the criterion of step 240 but may reduce the ability to fine tune the gap.

Similarly, Typically, but not obligatorily, for an initial value of AL in the range of

AL ~LMAX/10 (AL equals approximately a tenth of the value of the target value of the inductance at the nominal maximal current value) may be practically useful in the design method described above. However, smaller or larger values of AL may be used in the method described hereinabove, depending on the user defined tolerances of the device's inductance at the nominal maximal current value.

It is noted that typically, the devices physical parameters, such as for example core's overall dimensions and the coil structure are not changed in the iterations and the only design parameter which are changed in the iterations are the individual gap length and the dimensions of the ferrite members which are changed in order to keep the overall core's dimensions constant. It is further noted that the shape of cross-sectional profile of the ferrite members of which the core is comprised is also kept constant and does not change during the design iterations.

At the end of the design procedure described, the core inductance (at nominal maximal current) LEM of the final design (after the procedure is completed at step 242) is close to the initial design parameter of LMAX within the defined inductance tolerance of ±AL. Using the core design method of the present application, the final total gap length GLTF = N X GLINEW is always smaller than the initially computed value of the total gap length GLT such that GLTF < GLT.

The inventor noticed that it is possible to further improve magnetic cores with four gaps by practicing the core structure illustrated in Fig. 2A in which the gap placement used in prior art magnetic cores is in the middle of each side of the rectangular core (as illustrated in detail for the core 20 of Fig. 2A) instead of the prior art type of gap placement in which the gaps are positioned in the corners of the magnetic core (as illustrated in detail for the prior art core 150 of Fig. 15). The preferred gap arrangement illustrated for the core 20 advantageously reduces magnetic flux losses in the core 20 as compared to the prior art core 150 as explained in detail hereinabove with respect to Figs.18-21 , hereinabove. Thus, the specific arrangement of the gaps in the core 20 is a novel and preferred arrangement in a rectangular magnetic core having four gaps, in accordance with another embodiment of the improved (more efficient) magnetic cores of the present application.

It is noted that the magnetic cores disclosed hereinabove and illustrated in Figs 1-

4 may be made from any type of suitable ferromagnetic material. For example, many types of ferromagnetic materials may be used, including, but not limited to, a ferrite, a soft ferrite, a NiZn Ferrite, a NiCuZn ferrite, a MgZn ferrite, a MnMgZn ferrite, and a ferrite containing, as an additive or substituting oxide, at least one oxide selected from oxide(s) of Ti, Cr, Al, Sn, Li, Co, Pb, Bi, V, Si and Ca. Other useable magnetic materials may include iron, soft iron, iron powder, composite magnetic materials such as, but not limited to silicon-metal alloys, various iron alloys or other magnetic metals or metal alloys, amorphous iron, or other amorphous metals or amorphous metal alloys may also be used. However, the above indicated magnetic materials are given by way of example only and are not meant to limit the type and composition of magnetic materials usable in constructing the magnetic cores of the devices described in the present application. Rather, any type of suitable magnetic material known in the art may be used for constructing the cores described herein. The type of magnetic or ferromagnetic material selected may depend, inter alia, on the particular application, the dimensions of the core, the maximal current allowable, the permissible power losses and other design considerations.

It will be appreciated by those skilled in the art that while the magnetic cores disclosed herein have magnetic members having a rectangular cross-sectional shape, this is not obligatory to practicing the cores of the application. Thus, the cross-sectional shape of the multi-gap magnetic cores of the present application may also have other, different shapes, including but not limited to, a circular cross-sectional shape, a square cross- sectional shape, an elliptical cross-sectional shape, a polygonal cross-sectional shape, a regular polygonal cross-sectional shape and any other suitable cross-sectional shape known in the art of magnetic core design.

Moreover, the specific magnetic and non-magnetic materials disclosed herein are given by way of example only, and the material composition of the ferromagnetic, and/or magnetic, and/or ferrimagnetic, and/or non-magnetic parts of the cores and/or devices using the cores disclosed herein may be different, as is known in the art, depending on the specific design and application being considered. For example, the magnetic material of the cores taught in the present application may vary from soft ferromagnetic material to hard ferromagnetic material (of the ferrite type or any other type of ferromagnetic core material known in the art), depending on the AC frequency at which the device including the core may operate. Similarly, the core may be made from soft iron, or from amorphous iron or other amorphous magnetizable metals, or iron powder, or other Fe/ Si type of magnetic material, depending, among others on the specific design.

Moreover, while the type of rectangular core used in the Examples disclosed is based on a rectangular shaped closed magnetic path cores, this is not obligatory and the concepts of increasing the number of gaps while decreasing the individual gap lengths (and decreasing the total gap length) may be similarly applied to various different types of cores, such as but not limited to, toroidal magnetic cores, curved magnetic cores, and the like.

It is also noted that while the specific dimensions of the cores used in the experiments described above are given for the sake of completeness of information of the exemplary core designs used to perform the measurements of the core inductance, they are by no means obligatory to practicing the cores of the present application, and cores having other different dimensions and shapes and cross-sectional shapes may be used, as may be easily understood and implemented by those skilled in the art.