Integrated magnetics for soft switching converter

Abstract

In an integrated magnetic component for a switched mode power converter, comprising two magnetic cores forming an 8-shaped core structure and at least two first electric winding wires, wherein at least one magnetic core is an E-core, at least one of the first electric winding wires is wound on a flange of the E-core.

Claims

1. Switching mode power converter including a transformer, a parallel resonant inductor, two output filter chokes, and an integrated magnetic component, where the switched mode power converter is a soft switching converter, the integrated magnetic component comprising: two magnetic cores forming an 8-shaped core structure; a first electric winding wire constituting a winding of a first output filter choke wound on a first flange of a first magnetic core, the first electric winding wire further constituting a first secondary winding of the transformer wound on the second flange of the first magnetic core, a supplementary first electric winding wire constituting windings of a second output filter choke wound on a first flange of the second magnetic core, the supplementary first electric winding wire further constituting a second secondary winding of the transformer wound on the second flange of the second magnetic core, and a second electric winding wire constituting a first primary winding of the transformer wound on the second flange of the first magnetic core, a supplementary second electric winding wire constituting a second primary winding of the transformer wound on a second flange of the second magnetic core, wherein, the first magnetic core is an E-core, and the second magnetic core is an E-core of an I-core, the first electric winding wires are connected to each other and the second electric winding wires are connected to each other, and all windings are wound around the flanges of the magnetic cores such as to lead to a cancellation of flux in a common center leg of the magnetic cores.

2. The switched mode power converter according to claim 1, wherein at least one of the electric winding wires is wound directly on one of the magnetic cores.

3. The switched mode power converter according to claim 1, wherein the first electric winding wires are connected to each other by a first soldering joint and/or the second electric winding wires are connected to each other by a second soldering joint.

4. The switched mode power converter according to claim 1, wherein the 8-shaped core structure comprises an air gap.

5. The switched mode power converter according to claim 4, wherein the air gap is centrally located in the 8-shaped core structure.

6. The switched mode power converter according to claim 1, wherein the 8-shaped core structure comprises a slanted edge.

7. The switched mode power converter according to claim 1, wherein three windings are wound on each magnetic core.

8. A switched mode power converter including a transformer, a parallel resonant inductor, two resonant inductors, and an integrated magnetic component where the switched mode power converter is a LLC converter, the integrated magnetic component comprising: two magnetic cores forming an 8-shaped core structure; and a first electric winding wire constituting a winding of a first resonant inductor wound on a first flange of a first magnetic core, the first electric winding wire further constituting a first primary winding of the transformer wound on a second flange of the first magnetic core, a supplementary first electric winding wire constituting windings of a second resonant inductor wound on a first flange of a second magnetic core, the supplementary electric winding wire further constituting a second primary winding of the transformer wound on the second flange of the second magnetic core, and a second electric winding wire constituting a first secondary winding of the transformer wound on the second flange of the first magnetic core, a supplementary second electric winding wire constituting a second secondary winding of the transformer wound on a second flange of the second magnetic core, the first electric winding wires are connected to each other and the second electric winding wires are connected to each other, wherein, the first magnetic core is an E-core, and the second magnetic core is an E-core or an I-core, and all windings are wound around the flanges of the magnetic cores such as to lead to a cancellation of flux in a common center leg of the magnetic cores.

9. The switched mode power converter according to claim 8, wherein at least one of the electric winding wires is wound directly on one of the magnetic cores.

10. The switched mode power converter according to claim 8, wherein the first electric winding wires are connected to each other by a first soldering joint and/or the second electric winding wires are connected to each other by a second soldering joint.

11. The switched mode power converter according to claim 8, wherein the 8-shaped core structure comprises an air gap.

12. The switched mode power converter according to claim 11, wherein the air gap is centrally located in the 8-shaped core structure.

13. The switched mode power converter according to claim 8, wherein the 8-shaped core structure comprises a slanted edge.

14. The switched mode power converter according to claim 8, wherein three windings are wound on each magnetic core.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) The drawings used to explain the embodiments show:

(2) FIG. 1: Equivalent circuit of a soft switching converter circuit, one possible circuit in which an integrated magnetic component according to the invention can be used;

(3) FIG. 2: equivalent circuit of an LLC resonant converter circuit, another possible circuit in which an integrated magnetic component according to the invention can be used;

(4) FIG. 3: schematic view of an integrated magnetic component according to the invention (first embodiment);

(5) FIG. 4: reluctance model of the embodiment of FIG. 3;

(6) FIG. 5: runs of curves of voltage and current of secondary transformer winding as well as flux density (induction) in transformer core legs, choke core legs and centre legs;

(7) FIG. 6: schematic view of an integrated magnetic component according to the invention (second embodiment);

(8) FIG. 7: reluctance model of the embodiment of FIG. 6 and

(9) FIG. 8: runs of curves of series resonant choke current, parallel resonant choke current and primary winding current as well as flux density (induction) in transformer core legs, choke core legs and centre legs.

(10) In the figures, the same components are given the same reference symbols.

PREFERRED EMBODIMENTS

(11) FIG. 1 shows an equivalent circuit of a soft switching converter circuit, one possible circuit in which an integrated magnetic component according to the invention can be used. The soft switching converter circuit comprises an input circuit comprising four switching devices Q1, Q2, Q3, Q4 and an input capacitor C.sub.in, an output circuit comprising four diodes D1, D2, D3, D4 and an output capacitor C.sub.o, as well as the integrated magnetic component 1 according to the invention. The equivalent circuit of the integrated magnetic component 1 comprises two input resonant inductors L.sub.r1 and L.sub.r2, a parallel resonant inductor L.sub.m and a transformer T.

(12) FIG. 2 shows an equivalent circuit of an LLC resonant converter circuit, another possible circuit in which an integrated magnetic component according to the invention can be used. LLC resonant converter circuit comprises an input circuit comprising two switching devices Q1, Q2 and an input capacitor C.sub.in, an output circuit comprising two diodes D1, D2 and an output capacitor C.sub.o, as well as the integrated magnetic component 1a according to the invention. The equivalent circuit of the integrated magnetic component 1a comprises a parallel resonant inductor L.sub.m, a transformer T and two output filter chokes L.sub.r1 and L.sub.r2.

(13) FIG. 3 shows a schematic view of the integrated magnetic component 1 according to the invention. The integrated magnetic component 1 comprises an 8-shaped core structure 2 comprising a first magnetic core 3.1 and a second magnetic core 3.2. Both magnetic cores 3.1 and 3.2 are in the form of E-cores.

(14) A first electric winding wire 4.1 is wound around the first magnetic core 3.1. The first electric winding wire 4.1 comprises a first winding 5.1being wound around a first flange of the first magnetic core 3.1 and constituting a winding of the first output filter choke L.sub.r1and a second winding 6.1being wound around a second flange of the first magnetic core 3.1 and constituting a first secondary winding S1 of the transformer T. A supplementary first electric winding wire 4.2 is wound around the second magnetic core 3.2. The supplementary first electric winding wire 4.2 comprises a third winding 5.2being wound around a first flange of the second magnetic core 3.2 and constituting a winding of the second output filter choke L.sub.r2and a fourth winding 6.2being wound around a second flange of the second magnetic core 3.2 and constituting a second secondary winding S2 of the transformer T. The first electric winding wire 4.1 and the supplementary first electric winding wire 4.2 are connected to each other via a first soldering joint 7.1. The first electric winding wire 4.1 comprises a connecting section 8.1 which connects the first winding 5.1 to the second winding 6.1. Analogically, the supplementary first electric winding wire 4.2 comprises a connecting section 8.2 which connects the third winding 5.2 to the fourth winding 6.2. The connecting sections 8.1 and 8.2 are placed on opposite sides of the 8-shaped core structure 2. The first electric winding wire 4.1 comprises a wire end section 10.1 and the supplementary first electric winding wire 4.2 comprises a wire end section 10.2. The wire end sections 10.1 and 10.2 are placed on opposite sides of the 8-shaped core structure 2.

(15) A second electric winding wire 9.1 is wound around the second flange of the first magnetic core 3.1 thus creating a fifth winding 12.1, constituting a first primary winding P1 of the transformer T. A supplementary second electric winding wire 9.2 is wound around the second flange of the second magnetic core 3.2 thus creating a sixth winding 12.2, constituting a second primary winding P2 of the transformer T. The second electric winding wire 9.1 and the supplementary second electric winding wire 9.2 are connected to each other via a second soldering joint 7.2. The second electric winding wire 9.1 comprises a wire end section 10.3 and the supplementary second electric winding wire 9.2 comprises a wire end section 10.4. The wire end sections 10.3 and 10.4 are placed on opposite sides of the 8-shaped core structure 2.

(16) The wire end sections 10.1 and 10.3 as well as the connecting section 8.1 are placed on one side of the 8-shaped core structure 2 while the wire end sections 10.2 and 10.4 as well as the connecting section 8.1 are placed on the other side of the 8-shaped core structure 2.

(17) The 8-shaped core structure 2 comprises three air gaps 11.1, 11.2 and 11.3. Air gaps 11.1 and 11.3 separate the two lateral legs of each magnetic core 3.1, 3.2 from the respective lateral legs of the other magnetic core 3.1, 3.2. The centre legs of the magnetic cores 3.1, 3.2 are separated by air gap 11.2.

(18) The six windings 5.1, 5.2, 6.1, 6.2, 12.1 and 12.2 are wound directlyi.e. bobbin-lesson the four flanges of the 8-shaped magnetic core structure 2 supplied by the two magnetic cores 3.1 and 3.2. The magnetic core structures 3.1 and 3.2 each comprise several slanted edges 13. These slanted edges 13 are less prone to damaging the electric winding wires 4.1, 4.2, 9.1 and 9.2 than sharp 90-degree edges.

(19) The introduction of the air gap 11.1 in the flux path of the transformer T corresponds to an integration of the transformer T and the parallel resonant inductor L.sub.m (see FIG. 1). The thus created parallel inductance is adjustable through the configuration of the air gap 11.1 while the turn ratio of the transformer T is unchanged.

(20) The fact that the cores of the gapped transformer T and the output filter chokes L.sub.r1 and L.sub.r2 are put together in a common centre leg composed by the two centre legs of the magnetic cores 3.1 and 3.2 leads to a cancellation of flux in this common centre leg and thus to a reduction of core losses.

(21) The structure for the integration of the inductors L.sub.m, L.sub.r1 and L.sub.r2 and the transformer T (shown in FIG. 1) with parallel input inductance L.sub.m, transformer T and coupled output filter chokes L.sub.r1 and L.sub.r2 wound on bobbin-less E-cores as shown in FIG. 3 can be summarized as follows: The structure comprises two magnetic cores 3.1 and 3.2 formed as E-cores (use of ER-cores is also possible), three air gaps 11.1, 11.2 and 11.3 as well as six windings 5.1, 5.2, 6.1, 6.2, 12.1 and 12.2. Windings 12.1 and 12.2 constitute two primary windings P1 and P2 of the transformer T. Windings 6.1 and 6.2 constitute two secondary windings S1 and S2 of the transformer T. Windings 5.1 and 5.2 constitute the windings of the first and second output filter chokes L.sub.r1 and L.sub.r2. The first and second output filter chokes L.sub.r1 and L.sub.r2 are wound symmetrically on different flanges and are optimally coupled.

(22) The corresponding reluctance model of the embodiment of FIG. 3 is shown in FIG. 4. R.sub.L represents the reluctance of the inductance core (right lateral legs and transversal flanges of magnetic cores 3.1 and 3.2 as shown in FIG. 3) with consideration of its respective air gap 11.3, R.sub.T represents the reluctance of transformer core (left lateral legs and transversal flanges of magnetic cores 3.1 and 3.2 as shown in FIG. 3) with consideration of its respective air gap 11.1 and R.sub.c represents the reluctance of the centre core (constituted by the two centre legs of the magnetic cores 3.1 and 3.2) with consideration of its respective air gap 11.2.

(23) After mathematical description of the reluctance model and application of Faraday's Law on all windings, some equation manipulations yield the inductance matrix of the integrated component which is calculated to be:

(24) $\begin{array}{c}L=\left(\begin{array}{cc}{L}_{11}& M_{12}\\ M_{12}& L_{22}\end{array}\right)\\ =\left(\begin{array}{cc}\frac{\left(R_L+R_c\right).Math.N_p^2}{R_T.Math.\left(R_L+R_c\right)+R_L.Math.R_c}& \frac{\left(R_c.Math.N_L+\left(R_L+R_c\right).Math.N_s\right).Math.N_p^2}{R_T.Math.\left(R_L+R_c\right)+R_L.Math.R_c}\\ \frac{\left(R_c.Math.N_L+\left(R_L+R_c\right).Math.N_s\right).Math.N_p}{R_T.Math.\left(R_L+R_c\right)+R_L.Math.R_c}& \frac{\begin{array}{c}\left(R_L+R_c\right).Math.N_s^2+2R_c.Math.\\ N_s{N}_L+\left(R_T+R_c\right).Math.N_L^2\end{array}}{R_T.Math.\left(R_L+R_c\right)+R_L.Math.R_c}\end{array}\right)\end{array}$

(25) With
N.sub.p1=N.sub.p2=N.sub.p/2, N.sub.s1=N.sub.s2=N.sub.s/2 and N.sub.L1=N.sub.L2=N.sub.L/2.

(26) Using calculated elements of the inductance matrix, primary L.sub.11, secondary self inductances L.sub.22 and the mutual inductance M.sub.12, parameters of transformer model, the magnetizing inductance L.sub.m, the secondary leakage inductance L.sub.r and the equivalent secondary turn number N.sub.sn are respectively described as

(27) $L_m=L_{11}=\frac{N_p^2}{R_T+R_L//R_c},L_r=L_{22}-\frac{M_{12}^2}{L_{11}}=\frac{N_L^2}{R_L+R_c},N_{\mathrm{sn}}=N_p\frac{M_{12}}{L_{11}}=N_s+\frac{R_c}{R_L+R_c}{N}_L.$

(28) N.sub.s turns are wound but the transformer exhibits N.sub.sn turns. By introducing an air gap in the centre leg, the effective secondary number of turns N.sub.sn becomes higher than the factual number of turns N.sub.s which allows reducing secondary copper losses.

(29) For high permeability low saturation flux density material with no air gap in the centre core (g.sub.30), just R.sub.c<<R.sub.L,R.sub.T, the gapped transformer and output filter inductor are magnetically decoupled and the primary leakage inductance L.sub.r, the magnetizing inductance L.sub.m and the equivalent secondary turn number N.sub.sn are simplified to be respectively:

(30) $L_r\frac{N_L^2}{R_L},L_m\frac{N_p^2}{R_T}\mathrm{and}{N}_{\mathrm{sn}}{N}_s.$

(31) The fluxes and flux densities in transformer leg (.sub.T, B.sub.T), in choke leg (.sub.L, B.sub.L) and in centre leg (.sub.c, .sub.c) are respectively calculated as follows:

(32) ${}_T\left(t\right)=B_T\left(t\right).Math.A_T=\frac{R_T.Math.\left(\frac{N_p{I}_m\left(t\right)}{R_T}\right)-\left(R_L//R_c\right).Math.\left(\frac{N_L{I}_L\left(t\right)}{R_L}\right)}{R_T+R_L//R_c}$${}_L\left(t\right)=B_L\left(t\right).Math.A_L=\frac{R_L.Math.\left(\frac{N_L{I}_L\left(t\right)}{R_L}\right)-\left(R_T//R_c\right).Math.\left(\frac{N_p{I}_m\left(t\right)}{R_T}\right)}{R_L+R_T//R_c}$${}_c\left(t\right)=B_c\left(t\right).Math.A_c={}_T\left(t\right)-{}_L\left(t\right)$
where I.sub.m is the transformer magnetizing current.

(33) For high permeability low saturation flux density material with no air gap in the centre core (g.sub.30), just R.sub.c<<R.sub.L,R.sub.T, the gapped transformer and output filter inductor are magnetically decoupled and the transformer leg flux and the filter inductor leg flux are simplified to be respectively:

(34) ${}_{\mathrm{Td}}\left(t\right)=B_{\mathrm{Td}}\left(t\right).Math.A_T\frac{N_p{I}_m\left(t\right)}{R_T}$${}_{\mathrm{Ld}}\left(t\right)=B_{\mathrm{Ld}}\left(t\right).Math.A_L\frac{N_L{I}_L\left(t\right)}{R_L}$${}_{\mathrm{cd}}\left(t\right)=B_{\mathrm{cd}}\left(t\right).Math.A_c{}_{\mathrm{Td}}\left(t\right)-{}_{\mathrm{Ld}}\left(t\right)$

(35) FIG. 5 illustrates run of curve of voltage and current of secondary transformer winding as well as flux density (induction) in cores transformer, choke and centre legs. The flux density in centre core leg B.sub.c is reduced and therefore the core losses in there are minimized. When the transformer and inductor cores are fully separated the flux circulating in all transformer legs is .sub.Td and the flux flowing in all inductor legs is .sub.d.

(36) FIG. 6 shows a second embodiment of the invention. The integrated magnetic component 1a comprises an 8-shaped core structure 2a composed of two magnetic cores 3.1a and 3.2a and suits an LLC resonant converter like the one shown in FIG. 2.

(37) Similarly to the embodiment shown in FIG. 3, the integrated magnetic component 1a comprises a first electric winding wire 4.1a and a supplementary first electric winding wire 4.2a. The first electric winding wire 4.1a comprises a wire end section 10.1a, a first winding 5.1a, a connecting section 8.1a and a second winding 6.1a. The supplementary first electric winding wire 4.2a comprises a wire end section 10.2a, a third winding 5.2a, a connecting section 8.2a and a fourth winding 6.2a. The first electric winding wire 4.1a and the supplementary first electric winding wire 4.2a are connected to each other via a soldering joint 7.1a.

(38) Also similarly to the embodiment shown in FIG. 3, the integrated magnetic component 1a comprises a second electric winding wire 9.1a and a supplementary second electric winding wire 9.2a. The second electric winding wire 9.1a comprises a wire end section 10.3a and a fifth winding 12.1a. The supplementary second electric winding wire 9.2a comprises a wire end section 10.4a and a sixth winding 12.2a. The second electric winding wire 9.1a and the supplementary second electric winding wire 9.2a are connected to each other via a soldering joint 7.2a.

(39) The 8-shaped core structure 2a comprises three air gaps 11.1a, 11.2a and 11.3a. Air gaps 11.1a and 11.3a separate the two lateral legs of each magnetic core 3.1a, 3.2a from the respective lateral legs of the other magnetic core 3.1a, 3.2a. The centre legs of the magnetic cores 3.1a, 3.2a are separated by air gap 11.2a.

(40) The six windings 5.1a, 5.2a, 6.1a, 6.2a, 12.1a and 12.2a are wound directlyi.e. bobbin-lesson the four flanges of the 8-shaped magnetic core structure 2a supplied by the two magnetic cores 3.1a and 3.2a. The magnetic core structures 3.1a and 3.2a each comprise several slanted edges 13a. These slanted edges 13a are less prone to damaging the electric winding wires 4.1a, 4.2a, 9.1a and 9.2a than sharp 90-degree edges.

(41) The integrated magnetic component 1a comprises an output wire 14, connected to the soldering joint 7.2a.

(42) In contrast to the embodiment shown in FIG. 3, all wire end sections 10.1a, 10.2a, 10.3a and 10.4a as well as the connecting sections 8.1a and 8.2a are placed on the same side of the 8-shaped structure 2a.

(43) With reference to FIG. 6 and FIG. 2, the first winding 5.1a is wound on a first flange of the first magnetic core 3.1a constituting the winding of the first resonant inductor L.sub.r1. The second winding 6.1a is wound on a second flange of the first magnetic core 3.1a constituting the first primary winding P1 of the transformer T. The third winding 5.2a is wound on a first flange of the second magnetic core 3.2a constituting the winding of the second resonant inductor L.sub.r2. The fourth winding 6.2a is wound on a second flange of the second magnetic core 3.2a constituting the second primary winding P2 of the transformer T. The fifth winding 12.1a is wound on the second flange of the first magnetic core 3.1a constituting the first secondary winding S1 of the transformer T. The sixth winding 12.2a is wound on the second flange of the second magnetic core 3.2a constituting the second secondary winding S2 of the transformer T.

(44) The corresponding reluctance model of the embodiment of FIG. 6 is shown in FIG. 7. R.sub.L represents the reluctance of the inductance core (left lateral legs and transversal flanges of magnetic cores of magnetic cores 3.1a and 3.2a as shown in FIG. 6) with consideration of its respective air gaps 11.3a, R.sub.T represents the reluctance of transformer core (right lateral legs and transversal flanges of magnetic cores legs of magnetic cores 3.1a and 3.2a as shown in FIG. 6) with consideration of its respective air gaps 11.1a and R.sub.c represents the reluctance of the centre core (constituted by the two centre legs of the magnetic cores 3.1a and 3.2a) with consideration of its respective air gap 11.2a. After mathematical description of the reluctance model and application of Faraday's Law on all windings, some equation manipulations yield the inductance matrix of the integrated component which is calculated to be

(45) $\begin{array}{c}L=\left(\begin{array}{cc}{L}_{11}& M_{12}\\ M_{12}& L_{22}\end{array}\right)\\ =\left(\begin{array}{cc}\frac{\begin{array}{c}\left(R_L+R_c\right).Math.N_p^2+2R_c.Math.\\ N_p{N}_L+\left(R_T+R_c\right).Math.N_L^2\end{array}}{R_T.Math.\left(R_L+R_c\right)+R_L.Math.R_c}& \frac{\left(R_c.Math.N_L+\left(R_L+R_c\right).Math.N_p\right).Math.N_s}{R_T.Math.\left(R_L+R_c\right)+R_L.Math.R_c}\\ \frac{\left(R_c.Math.N_L+\left(R_L+R_c\right).Math.N_p\right).Math.N_s}{R_T.Math.\left(R_L+R_c\right)+R_L.Math.R_c}& \frac{\left(R_L+R_c\right).Math.N_s^2}{R_T.Math.\left(R_L+R_c\right)+R_L.Math.R_c}\end{array}\right)\end{array}$

(46) With
N.sub.p1=N.sub.p2=N.sub.p/2, N.sub.s1=N.sub.s2=N.sub.s and N.sub.L1=N.sub.L2=N.sub.L/2.

(47) Using calculated elements of the inductance matrix, primary L.sub.11, secondary self inductances L.sub.22 and the mutual inductance M.sub.12, parameters of transformer model, the magnetizing inductance Lm, the secondary leakage inductance Lr and the equivalent primary turn number N.sub.pn are respectively described as

(48) $L_r=L_{11}-\frac{M_{12}^2}{L_{22}}=\frac{N_L^2}{R_L+R_c},L_m=L_{11}-L_r=\frac{R_L+R_c}{R_T.Math.\left(R_L+R_c\right)+R_L.Math.R_c}{N}_{\mathrm{pn}}^2=\frac{N_{\mathrm{pn}}^2}{R_T+R_L//R_c}$$\mathrm{and}$$N_{\mathrm{pn}}=N_s\frac{L_m}{M_{12}}=N_p+\frac{R_c}{R_L+R_c}{N}_L.$

(49) N.sub.p turns are wound but the transformer exhibits N.sub.pn turns. By introducing an air gap in the centre leg, the effective primary number of turns N.sub.pn becomes higher than the factual number of turns N.sub.p which allows reducing primary copper losses.

(50) For high permeability low saturation flux density material with no air gap in the centre core (g.sub.30), just R.sub.c<<R.sub.L,R.sub.T, the gapped transformer and resonant inductor are magnetically decoupled and the primary leakage inductance L.sub.r, the magnetizing inductance L.sub.m and the equivalent primary turn number N.sub.pn are simplified to be respectively:

(51) $L_r\frac{N_L^2}{R_L},L_m\frac{N_{\mathrm{pn}}^2}{R_T}\mathrm{and}{N}_{\mathrm{pn}}{N}_p.$

(52) The fluxes and flux densities of legs of this embodiment are calculated as with the integrated magnetics for the soft switching converter.

(53) FIG. 8 illustrates run of curve of series resonant choke current, parallel resonant choke current and primary winding current as well as flux density (induction) in cores transformer, choke and centre legs. The flux density in centre core leg B.sub.c is reduced and therefore the core losses in there are minimized.

(54) It is to be noted that the invention is not limited to the two embodiments described above. The scope of protection is rather defined by the patent claims.

(55) TABLE-US-00001 List of reference symbols 1 Integrated magnetic component 2 8-shaped core structure 3.1, 3.2 Magnetic cores 4.1, 4.2 First winding wires 5.1, 5.2 First winding, third winding 6.1, 6.2 Second winding, fourth winding 7.1, 7.2 Soldering joints 8.1, 8.2 Connecting sections 9.1, 9.2 Second winding wires 10.1, 10.2, 10.3, Wire end sections 10.4 11.1, 11.2, 11.3 Air gaps 12.1, 12.2 Fifth winding, sixth winding 13 Slanted edge 14 Output wire