EP0715323A1

EP0715323A1 - Setting inductance value of magnetic components

Info

Publication number: EP0715323A1
Application number: EP95308655A
Authority: EP
Inventors: Patrizio Vinciarelli; Lance L. Andrus; Sean Crilly
Original assignee: VLT Corp
Current assignee: VLT Corp
Priority date: 1994-12-01
Filing date: 1995-11-30
Publication date: 1996-06-05
Also published as: JPH08250361A

Abstract

A method for setting a desired value of winding inductance in a magnetic component of the kind having a permeable core structure which defines a magnetic flux path and has a gap along the path. The method includes: a. measuring the actual value of the inductance of the winding; and b. automatically changing the gap length to cause the actual value to approach the desired value. Apparatus for setting the desired value of inductance includes: a. inductance measurement circuitry having a port for connection to the magnetic component; and b. a device having a movable support for the core structure and a controller configured to automatically change the gap length to cause the actual value to approach the desired value.

Description

This invention relates to setting inductance values of magnetic components.
An example of a gapped magnetic component 10 is shown in Figure 1. The component consists of two permeable core halves 20, 22 and one or more conductive windings (only one such conductive winding 24 is shown). The faces of the core halves 26a, 26b, 26c, 26d are not in contact with each other, but are separated by gaps 30, 32 which are small relative to the cross-sectional dimensions of the core faces. The core halves have a magnetic permeability, µ₂, which is greater than the permeability, µ₁, of the medium outside of the cores. A current, I, flowing in the winding 24, produces magnetic flux, φ, within the core halves and the gaps. For a non-ideal component, the inductance of the component 10 (e.g., the inductance measured at a winding 24 with any other winding open-circuited) and the maximum amount of magnetic energy which can be stored by the component (e.g., under conditions for which the total magnetic flux, φ, causes the flux density in the core halves to approach their saturation flux density, Bsat) are dependent upon the aggregate length of the gaps, GAP = g1+g2. As GAP increases, the energy storage capability increases and the inductance decreases. If we assume, for example, that all of parameters which characterize the component of Figure 1 are fixed (e.g., µ₂, µ₁, the dimensions of the core halves, the number of turns on the winding, and the saturation flux density of the core material) then the ungapped device (e.g., with GAP = 0) will exhibit some value of inductance, Lu, and will be capable of carrying a maximum current, I = Iu, corresponding the point at which core saturation occurs. This will correspond to a stored energy equal Eu = Lu*Iu²/2 for the ungapped component. If a small gap is introduced, of total length GAP, then the new values of inductance, Lg, maximum current at saturation, Ig, and energy storage capability, Eg, for the gapped component, will be approximately: Lg = Kg/Lu, Ig = Kg*Iu and Eg = Kg*Eu, where Kg = 1+ (µ₁*GAP/µ₂*L), where L is the total mean magnetic path length of the core halves, L = ml1+ml2.
One reason for using a gapped structure is that adjustment of the length(s) of the gap(s) provides a means for setting a specific value of inductance despite lot-to-lot variations in the permeability of the core halves. Another reason for using a gapped magnetic structure is that the amount of energy which can be stored in the structure, at a particular flux density (e.g., at the saturation flux density, Bmax, of the core halves), can be increased substantially with an essentially insignificant increase in the overall volume of the structure. Another reason for using a gapped structure is that the amount of variation in inductance due to variations in the permeability of the core halves (due, for example, to lot-to-lot variations in material or to temperature) is reduced if the gap is filled with non-permeable material (e.g., air; plastic) having an essentially invariant value of permeability.
One way to set a gap is to place non-permeable spacers of known thickness between the core faces of the core halves.
In general, in one aspect the invention features a method for setting a desired value of winding inductance in a magnetic component of the kind having a permeable core structure which defines a magnetic flux path and has a gap along the path. The method includes: a. measuring the actual value of the inductance of the winding; andb. automatically changing the gap length to cause the actual value to approach the desired value.
Implementations of the invention may include the following features. The method may be continued (e.g., iteratively) until the actual value is within a desired range of the desired value. The difference between the actual value and the desired value may never change signs, or may be allowed to assume both positive and negative values. An adhesive material, lying in the gap or outside the gap, may be used to fix the gap length. The adhesive material may be applied prior to, or at the end of, the iterative process. The core structure may include two core pieces separated by the gap and the step of changing the gap length may include moving the two core pieces toward or away from each other. The desired value may be set to take into account changes in the volume of the adhesive that will occur during setting of the adhesive. The core pieces may be held in position relative to one another using vacuum suction during the process. Electromagnetic shields may be mounted on the core pieces. The method may be repeated for setting a different desired value of a second magnetic component. The change of gap length between iterations may be larger when the difference between the actual value and the desired value is larger and conversely. The gap may include two sub-gaps at different positions along the path.
In general, in another aspect, the invention features apparatus for setting a desired value of inductance in a magnetic component having a core structure which defines a magnetically conductive path and has a gap along the path. The apparatus includes: a. inductance measurement circuitry having a port for connection to the magnetic component; and b. a device having a movable support for the core structure and a controller configured to automatically change the gap length to cause the actual value to approach the desired value.
Among the advantages of the invention are that the inductance can be set rapidly, accurately, and repeatably from unit to unit.
Other advantages and features will become apparent from the following description.
We first briefly describe the drawings:
Figure 1 is a side view of a gapped magnetic component.
Figure 2 is a cross-sectional view of a transformer. Figure 3 is a schematic side sectional view of apparatus for setting a winding inductance of a magnetic component.
Figure 4 is a flow diagram of an algorithm for automatically setting the inductance of a magnetic component.
Figure 5 is an exploded perspective view of a transformer.
Figure 6 is a schematic diagram of an inductance controller.
Figure 7 is a table of values associated with automatic setting of winding inductance of a transformer.
Figures 8A and 8B are side and sectional views showing how one of the stops of the apparatus of Fig. 3 is adapted to apply vacuum suction in the region of the face of the stop.
Figures 9A and 9B show two alternative ways of applying adhesive to a magnetic component.
Transformer 40 of Figure 2 has two symmetrical permeable core halves 42, 44 and two winding assemblies 50, 52. Each winding assembly has a hollow plastic bobbin 45, 46 overwound with a number of turns of wire to form a conductive winding 47, 48. The transformer is fabricated by selecting two wound bobbin assemblies each having a particular number of turns appropriate to a specific transformer and two core halves 42, 44, and inserting the cores into the ends of the winding assemblies 50, 52.
In some applications the core halves may be inserted with the faces 60a, 60b, 62a, 62b of the cores in contact (no gaps); in other cases, small gaps 64, 66 may be provided between the core faces to reduce the magnetizing inductance and increase the transformer magnetizing energy. Transformer 40 could be used in a wide variety of DC-DC power converters, e.g., for applications in which the nominal value of converter input voltage, Vin, may range from 5 to 300 Volts and in which the nominal value of converter output voltage, Vout, may range from 1 to 100 Volts. Different combinations of input and output voltages are accommodated by adjusting the numbers of turns on the primary and secondary windings 47, 48 (e.g., by selecting different winding assemblies 50, 52 with different numbers of turns). For each combination of winding assemblies and voltages there will be some desired, predefined, values of winding inductances (where the "winding inductance" is the inductance of a particular winding with all other windings left open-circuited). Such a transformer is one example of a more general class of magnetic energy storage components (e.g., transformers, inductors) which have a winding, or windings, which are linked by flux confined within a permeable core. In certain applications it is desirable to control accurately a value of winding inductance.
Putting a spacer of predefined thickness in between the faces of the cores will decrease the magnetizing inductance (relative to that for the ungapped structure) and increase the energy storage capability of the device, but the resulting value of winding inductance may not be accurate or repeatable from unit to unit. This is because the actual winding inductances of any particular transformer will depend upon the transformer configuration (e.g., size and geometric configuration of parts which comprise the transformer), the permeability of the core material (which may vary from unit to unit due to manufacturing tolerances), the number of turns in the winding, the combined length of the gaps 64, 66 and other factors, such as misalignment in the faces of the cores during assembly, and some of these factors will exhibit unit-to-unit variations owing to tolerances (e.g., permeability, size of parts, size of gap spacers) or assembly-related variations (e.g., alignment of core faces).
In Figure 3, an inductance-setting apparatus 80 accurately and repeatably sets the inductances of magnetic components despite unit-to-unit variations introduced by either the assembly process or variations in the characteristics or values of the specific set of components used in making the units. In Figure 3, transformer 40 is held between a fixed stop 62 and a movable stop 64. Before being placed between the stops, the transformer is pre-assembled as follows. Two core halves 42, 44 are selected from a supply of core halves (not shown) and a desired pair of winding assemblies 50, 52 (shown cutaway to reveal gaps between the faces of the core halves) is also selected from a supply of windings (not shown). The windings are assembled over one of the core halves (e.g., core half 44) and a drop of adhesive 90a, 90b is dispensed onto each face of the core half 44. A small amount of adhesive activator 91a, 91b is placed on the faces of the other core half 42. The resulting assembly is then positioned between stops 62, 64 without allowing the faces of the two cores halves 44, 42 to come together during positioning (which would cause the activator and adhesive to mix, thereby causing the activated adhesive to begin to set up).
The outer ends 43, 45 of the core halves are constrained to remain in contact with the faces 47, 49 of the stops 62, 64, e.g., by applying vacuum suction in the manner shown in Figures 8A and 8B. In Figure 8A the face 49 of one stop 62 has several orifices 94a - 94e. In Figure 8B (which shows a side view of stop 62 partially sectioned along the path labeled A-A in Figure 8A), vacuum suction is applied to fitting 95 and is conducted to the orifices via hollow region 92. A core placed in contact with the face 49 will be drawn into contact with the face by the vacuum suction. The same scheme may be applied to stop 64.
When the transformer is first installed between the stops 62, 64 the gap, G, is set to a relatively large value to ensure that the actual inductance of the winding 52, Lact, is initially much lower than the desired setpoint value, Lset. A stepper motor 70 is arranged to rotate a leadscrew 72. Rotation of the leadscrew causes the follower 74 to move along the length of the leadscrew in the directions indicated by the arrow "Y". Since the follower is connected to the movable stop 64 by the arm 76, and since the core half 42 is constrained to remain in contact with movable stop 64, rotation of the leadscrew will cause the gap G to vary.
The value of Lset is delivered to an inductance controller 84 which sends a motor control signal 85 to a stepper motor controller 82. Other information, described below, may also be delivered to the controller, as indicated by the controller input labeled "Data Input." An inductance measuring device 86, connected to winding 52 via conductive interconnections 78a, 78b
(which, in practice, might involve four wires in a "Kelvin" measurement scheme), measures the actual inductance, Lact, of the winding 52 and sends a measurement signal 83, indicative of the value of Lact, to the inductance controller 84. The motor control signal 85 both activates the stepper motor 70 and sets the direction of motor rotation. The control signal is adjusted by the inductance controller 84 to reduce the error, Lerr = Lset - Lact towards zero. This sets the gap G automatically to bring Lact equal to Lset. During the adjustment process the faces of the cores come close to each other causing the adhesive and activator to mix and fill the gaps between the core faces; after the inductance setting process is completed the assembly is allowed to sit undisturbed long enough for the adhesive to set up and fix the gap.
An initial measurement value of Lact will typically indicate a relatively large error, Lerr = Lset - Lact, due to the large initial value of gap. In one adjustment algorithm, the controller will respond to this large error by commanding the stepper motor controller 82 to take a relatively large number of steps (each rotational step of the motor 70 corresponding to a fixed variation in the total gap, ΔG) which the controller determines are sufficient to reduce, but not entirely eliminate, the error. A later measurement of Lset will therefore reflect a smaller error and the controller 84 will respond by commanding a smaller number of steps. This process is iterated a number of times until the value of Lact comes to a value which is within some predetermined tolerance band around Lset (e.g., +/- 1 %). This kind of "overdamped adjustment" algorithm, which converges on a final value of inductance without any overshoot (e.g., the controller never needs to correct for a negative value of error, Lerr), is useful when adhesive is placed in the gap prior to the adjustment process taking place. The adhesive is compressed as the gap G (and hence the volume between the core faces) is reduced, and excess adhesive is squeezed out from the gap into the space between the inner surface of the winding and the outer surface of the cores. If a negative error (as a result of overshoot) were allowed to occur, the subsequent compensatory increase in the gap size might leave insufficient activated adhesive in the volume between the faces to fill the increased volume and reliably hold the core halves together.
In Figure 4, an overdamped adjustment algorithm 100 begins with the receipt of data input values 102 indicative of the desired final value of inductance, Lset, and other data inputs used by the algorithm. It is assumed that different types of magnetic components (e.g., of different sizes, for example) may be constructed within the apparatus 80, and that each different type may be set to a variety of final inductance values depending, for example, upon the number of turns on the winding, or windings, being used. For the transformer of Figure 3, the other data inputs could include the values ΔG1, P1, P2, X1, X2, Y1 and Y2, as described below. The data input values may be manually entered by an operator or delivered electronically (e.g., digital values extracted from a database). An initial measurement of inductance, Lact = L1, is made 104, after which the gap is reduced by a predetermined fixed distance, ΔG1, and another measurement of inductance, Lact = L2, is made 106. For the apparatus 80 of Figure 3, the fixed distance would correspond to a fixed number of steps of the stepper motor 70. As an initial check, the value of L2 is compared to the value L1 108; if L2 is finite and greater than L1, then the apparatus is presumed to be operating correctly and the controller 84 proceeds to execute the balance of the algorithm; if, however, the value of L2 cannot be measured or is not greater than L1, then an error signal is delivered 110 indicating the possible existence of a problem (e.g., the apparatus might be jammed or the measurement interconnections 78a, 78b might be faulty or intermittent).
As a first step in an iterative portion of the algorithm a measurement of inductance 112, Lact = Ln, is made. The value Ln is compared to a fraction of the setpoint value P1*Lset 114, where 0<P1<1. For example, P1 might be 0.95 which would correspond to 95% of Lset. If Ln is less than P1*Lset, then a pair of constants, X and Y, are set to predetermined "coarse adjustment" values X = X1 and Y = Y1 118a. If, however, Ln is greater than P1*Lset, then the algorithm first determines if Ln has surpassed a value, P2*Lset, 0<P2<1, which is within a predetermined tolerance of Lset 116. For example, P2 might be 0.99, which would correspond to 99% of Lset. If this is not the case, the values of X and Y are set to predetermined "fine adjustment" values X2 and Y2. The values of Ln, Lset, X and Y are used to calculate an amount, ΔGn, by which the gap is to be reduced, using the exponential formula 120
where the function INT(Z) causes a number Z to be rounded off to the next lowest integer. In the case of the apparatus 80 of Figure 3, the value ΔGn would correspond to a number of steps taken by stepper motor 70. As Ln approaches Lset, ΔGn decreases toward a minimum decremental value ΔGnmin = INT(X*exp(-Y)).
After each reduction in the gap 120, another iteration of the process is begun by taking another measurement of Ln (112, Figure 4). After a number of iterations, the value of Ln will become greater than P2*Lset 116 and the inductance-setting process will be complete. The apparatus will be left idle for a fixed delay time 122, allowing the adhesive to set and completing the inductance-setting process. The stops 62, 64 may then be opened, the vacuum reduced and the completed magnetic component removed from the apparatus 80.
An apparatus of the kind shown in Figure 3 was used to set the winding inductance of a transformer of the kind shown in Figure 5 using an algorithm of the kind shown in Figure 4. The transformer, of the kind which has controlled leakage inductances (see, for example, our European Patent Application 92508315.8, now published as EP-A-0 532 360, incorporated by reference), has two ferrite cores 42, 44; two non-conductive plastic bobbins 45, 46; two windings 47, 48; four conductive winding terminators 202a, 202b, 204a, 204b; and two conductive shields 210, 212. As indicated by the arrows, the ferrite cores 42, 44 are inserted into the bobbins 45, 46 and the shields 210, 212 are placed over the ends of the cores. The ends of the two windings 47, 48 connect to pairs of conductive terminations 202a, 202b and 204a, 204b, respectively. The nominal mated dimensions of the pair of cores for one such transformer is approximately 1.4" x 0.9" x 0.25" (3.556cm x 2.286cm x 0.635cm). The cross-sectional area of the cores is approximately 0.08 inches square inches (0.516sq.cm). The cores are made of ferrite material having a nominal relative permeability value of 750. The manufactured tolerance variation in the value of permeability (at predetermined fixed values of temperature and flux density) is +/- 20%. With 20 turn windings 47, 48 the desired setpoint value of winding inductance, Lset, is 290 microhenry.
For the transformer of Figure 5, the data input values to the algorithm of Figure 4 are: Lset = 290 microhenry; P1 = 0.95; P2 = .995; X1 = 2000; X2 = 850; Y1 = Y2 = 5; ΔG1 = 1000 steps. In the inductance-setting apparatus 80 (Figure 3), the stepper motor 70, leadscrew 72 and follower 76 were embodied as a single axis positioning table, model number RM-200-SM, manufactured by New England Affiliated Technologies ("NEAT"), Lawrence, Ma., USA. The leadscrew and follower were both of 50 pitch, providing 0.02" (0.0508cm) of linear displacement for each revolution of the shaft 72. One full step of the stepper motor 70 produced 1.8 degrees of shaft rotation. The stepper motor controller 82 was a model 310M microstepping controller, also manufactured by NEAT, which provided for 10 microsteps per full step (e.g., each microstep provides one-tenth of a full step, or 0.18 degrees of shaft rotation). As a result, each microstep commanded by the motor control signal 85 resulted in 10 microinches (0.00001'' or 2.54x10^-7m) of linear displacement in the gap. The inductance measuring device was a model 4263A LCR metre manufactured by Hewlett-Packard, Test and Measurement Division, Santa Clara, Ca., USA. The meter was connected to a winding (e.g., winding 48, Fig. 6) using a standard 4-wire (e.g., two stimulus wires 178a, 178b and two Kelvin measurement wires 178c, 178d) measurement scheme.
The inductance controller 84 was implemented using an IBM-compatible 486-based personal computer (89, Figure 6) running the DOS operating system, version 5.0, manufactured by Microsoft Corporation, Redmond, Wa., USA. The algorithm was coded using Borland Pascal, version 7.0, manufactured by Borland International, Scotts Valley, Ca., USA. Any other computer, having roughly equivalent performance, and any other programming language could have been used. As shown in Figure 6, the personal computer incorporated a standard GPIB interface board 87a (e.g., a model PC<>488 GPIB interface controller, manufactured by Capital Equipment Corporation, Burlington, Ma., USA) for communicating with the GPIB port 87b included in the Hewlett-Packard 4263A LCR meter. This allowed inductance measurement values to be delivered to the personal computer from the LCR meter 82. As also shown in Figure 6, communication of motor control signal information 85 between the personal computer and the NEAT microstepping motor controller 82 was done via the standard RS-232 serial port interfaces 97a, 97b built into the computer and the stepper motor controller 82, respectively.
The adhesive 90a and activator 90b were, respectively, Loctite 325 adhesive and Loctite 792 activator, both manufactured by Loctite Corporation, Newington Connecticut, USA. For this type of adhesive a delay time (algorithm step 122, Fig. 4) of 20 seconds was used.
An actual sequence of iterative adjustments for the transformer and apparatus described above is shown in the table in Figure 7. Each line in the table shows results for a single pass through the iterative portion of the algorithm of Figure 4: an initial value of inductance (e.g., the value Ln at algorithm step 112, Figure 4), the resulting calculated value of ΔGn (algorithm step 120) and the elapsed time for the pass to be completed. Pass #1 begins with the measurement made at the time that algorithm step 112 is first entered (e.g., after the initial 1000 step (ΔG1) reduction in gap (algorithm step 106) has taken place). As indicated at the bottom of Figure 7, after eighteen passes the value of inductance was adjusted to be within 0.17% of the desired final value. The total elapsed time for the process was 2.33 seconds.
When adhesive is placed into the region of the gap, as described above, a small residual error in the final value of the inductance may result due to shrinkage of the adhesive following removal from the apparatus 80. This is because stresses build up in the adhesive as it sets up and these stresses will cause a small shrinkage in the gap when the unit is removed from the apparatus and the stresses are relieved. For the example cited, the error introduced amounted to about 3.5% of the setpoint value. Thus, the final measured value shown in Figure 7 (289.51 microhenry) increased to about 299.64 microhenry after removal from the apparatus.
One way to compensate for the effect of adhesive shrinkage is to measure the effect of shrinkage on inductance and use this information to derive an inductance setpoint, Lset, which is lower than the actual final desired value of inductance, Lfinal, in an amount which will be offset by the effect of the shrinkage. Thus, in the example cited above, the value Lset = 290 microhenry was selected by taking a desired final inductance value of Lfinal = 300 microhenries and adjusting it to compensate for a 3.5% shrinkage factor (e.g., Lset = 300 microhenry/1.035 = 290 microhenry).
The described overdamped adjustment algorithm, in which overshoot is avoided, is one example of a broader class of feedback control strategies which may be used to adjust the gap as a means of reducing Lerr = Lset = Lact. "Non-overdamped" algorithms, for example, which allow the value of Lact to oscillate about Lset (by allowing Lerr to go both positive and negative) may be faster to converge than overdamped algorithms. While non-overdamped algorithms may be unsuitable for use with constructions in which adhesive is placed within the gap, they can be used where the gap is set prior to application of the adhesive. For example, assume that the apparatus of Figure 3 includes a controller 84 which executes an ordinary proportional, proportional-plus-integral, or proportional-plus-integral-plus-derivative feedback control strategy. Also assume that adhesive 90a and activator 91a are not placed on the core faces prior to execution of the algorithm; the inductance value may be set rapidly by adjusting the gap without any adhesive present. Once the setpoint value of inductance is reached, adhesive may be applied to some appropriate location on the magnetic component. One way of doing this is shown in Figure 9A. In the Figure, a dispensing tip 250 is shown applying a fillet of adhesive 252 along the region at which the core enters the bobbin. Fillets of adhesive 254 are shown already in place at both ends of bobbin 46, whereas adhesive has not yet been placed at location 256. Other ways of applying adhesive after setting the gap include providing a conduit within the bobbin (e.g., by molding it into the part) for carrying dispensed adhesive down into the gap regions or arranging the cores so that the gaps are directly accessible to an adhesive dispenser after assembly. One such core pair 242, 244 is shown in Figure 9B. The gaps 251, 253 are located at the ends of legs 246, 248. If bobbins are placed over legs 246, 248 then adhesive may be placed directly over the outer periphery of the gaps by a dispenser (e.g., dispenser 250, Fig. 9A) after the inductance is set.
There are many ways of incorporating the invention into a high volume manufacturing line. An operator might manually retrieve components needed for assembly of a magnetic device (e.g., cores, bobbins and windings, adhesive and activator) and place them into the apparatus 80, Figure 3). Alternatively, the components needed to assemble particular devices might be automatically provided to an operator by a computer-controlled automatic storage and retrieval system. Likewise, the data inputs to the controller (84, Figure 3) can also be delivered from a database by a computer. As an alternative to a human operator, assembly of parts can be performed by a computer controlled robotic assembly station.

Claims

A method for setting a desired value of winding inductance in a magnetic component of the kind having a permeable core structure which defines a magnetic flux path and has a gap along the path, the method comprising:
a. measuring the actual value of the inductance of the winding, and

b. automatically changing the gap length to cause the actual value to approach the desired value.
The method of claim 1 further comprising
continuing the method until the actual value is within a desired range of the desired value.
The method of claim 1 wherein
the difference between the actual value and the desired value never changes sign.
The method of claim 1 wherein
the difference between the actual value and the desired value is allowed to assume both positive and negative values.
The method of claim 1 wherein said method for setting is performed iteratively.
The method of claim 1 further comprising
using an adhesive material to fix the gap length.
The method of claim 6 wherein the adhesive material lies in the gap.
The method of claim 6 wherein the adhesive material lies outside the gap.
The method of claim 6 wherein the adhesive material is applied prior to the iterative process.
The method of claim 6 wherein the adhesive material is provided at the end of the iterative process.
The method of claim 1 wherein the core structure includes two core pieces separated by the gap and the step of changing the gap length comprises moving the two core pieces toward or away from each other.
The method of claim 6 further comprising
setting the desired value to take into account changes in the volume of the adhesive that will occur during setting of the adhesive.
The method of claim 11 further comprising holding the core pieces in position relative to one another using vacuum suction during the process of setting.
The method of claim 1 further comprising
mounting electromagnetic shields on the core pieces.
The method of claim 1 further comprising
repeating the automatic process for setting a different desired value of a second magnetic component.
The method of claim 5 wherein the change of gap length between iterations is larger when the difference between the actual value and the desired value is larger and conversely.
The method of claim 1 wherein the gap comprises two sub-gaps at different positions along the path.
A method for setting a desired value of inductance in a magnetic component having a core structure which defines a magnetically conductive path and has two core pieces separated by a gap along the path, the method comprising:
a. measuring the actual value of the inductance, and

b. automatically changing the gap length to cause the actual value to approach the desired value by moving the two core pieces toward or away from each other,
the method being continued until the actual value is within a desired range of the desired value, and

using an adhesive material to fix the gap length.
The method of claim 18 wherein said method for setting is performed iteratively.
Apparatus for setting a desired value of inductance in a magnetic component having a core structure which defines a magnetically conductive path and has a gap along the path, the apparatus comprising:
a. inductance measurement circuitry having a port for connection to the magnetic component, and

b. a device having a movable support for the core structure and a controller configured to automatically change the gap length to cause the actual value to approach the desired value.