MAGNET CORE FOR LOW-FREQUENCY APPLICATIONS AND METHOD FOR PRODUCING A MAGNET CORE FOR LOW-FREQUENCY APPLICATIONS

Abstract

A magnet core for low-frequency applications and method for producing a magnet core for low-frequency applications is provided. The magnet core is made of a spiral-wound, soft-magnetic, nanocrystalline strip. The strip essentially has the alloy composition Fe.sub.RestCo.sub.aCu.sub.bNb.sub.cSi.sub.dB.sub.eC.sub.f, wherein a, b, c, d, e and f are stated in atomic percent and 0a1; 0.7b1.4; 2.5c3.5; 14.5d16.5; 5.5e8 and 0f1, and cobalt may wholly or partially be replaced by nickel. The magnet core has a saturation magnetostriction .sub.s of .sub.s<2 ppm, a starting permeability .sub.1 of .sub.1>100 000 and a maximum permeability .sub.max of .sub.max>400 000. In addition, a sealing metal oxide coating is provided on the surfaces of the strip.

Claims

1. A method for producing a magnet core for low-frequency applications from a spiral-wound, soft-magnetic, nanocrystalline strip, the strip essentially having the alloy composition Fe.sub.RestCo.sub.aCu.sub.bNb.sub.cSi.sub.dB.sub.eC.sub.f, wherein a, b, c, d, e and f are stated in atomic percent and 0a1; 0.7b1.4; 2.5c3.5; 14.5d16.5; 5.5e8 and 0f1, and cobalt may wholly or partially be replaced by nickel, wherein the strip is provided with a coating with a metal oxide solution and/or an acetyl-acetone-chelate complex with a metal, which coating forms a sealing metal oxide coating during a subsequent heat treatment for the nanocrystallisation of the strip, and wherein, in the heat treatment for the nanocrystallisation of the strip, a saturation magnetostriction .sub.s of |.sub.s|<2 ppm is set.

2. The method according to claim 1, wherein an element selected from the group of Mg, Zr, Be, Al, Ti, V, Nb, Ta, Ce, Nd, Gd, further elements of the 2.sup.nd and 3.sup.rd main groups and of the group of rare earth metals is used as a metal for the coating.

3. The method according to claim 1, wherein a saturation magnetostriction .sub.s of |.sub.s|<1 ppm, preferably |.sub.s|<0.5 ppm, is set in the heat treatment process.

4. The method according to claim 1, wherein the heat treatment is carried out field-free on non-stacked magnet cores in a continuous annealing process.

5. The method according to claim 4, wherein the non-stacked magnet cores are placed on a carrier having a good thermal conductivity in the continuous annealing process.

6. The method according to claim 4, wherein the magnet core passes through the following temperature zones in the heat treatment process: a first heating zone in which the magnet core is heated to a crystallization temperature; a constant or slightly rising decay zone with a temperature slightly above the crystallization temperature, the passage through the decay zone lasting at least 10 minutes; a second heating zone in which the magnet core is heated to a maturation temperature for setting the nanocrystalline structure; a maturation zone with a substantially constant maturation temperature T.sub.x between 540 C. and 600 C., the passage through the maturation zone lasting at least 15 minutes.

7. The method according to claim 4, wherein the heat treatment is carried out in an inert gas atmosphere of H.sub.2, N.sub.2 and/or Ar, the dew point T.sub.P being <25 C. or T.sub.P<49.5 C.

8. The method according to claim 1, wherein the strip is wound at a descending skew.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Embodiments of the invention are explained in greater detail below with reference to the accompanying figures.

[0056] FIG. 1 is a diagrammatic representation of an AC-sensitive RCD according to an embodiment of the invention;

[0057] FIG. 2 is a diagrammatic representation of a possible temperature curve of a heat treatment according to a method for producing a magnet core according to an embodiment of the invention;

[0058] FIG. 3 shows the surface of an uncoated strip after heat treatment;

[0059] FIG. 4 is a diagram illustrating the influence of crystallization temperature on the change of the coercitive field strength of a magnet core under radial deformation;

[0060] FIG. 5 is a diagram illustrating the influence of crystallization temperature and of a coating on the (H)-commutation curves of a magnet core;

[0061] FIG. 6 is a diagram illustrating the influence of crystallization temperature and of a coating on the on the hysteresis loop of a magnet core;

[0062] FIG. 7 is a view of the underside of an uncoated strip after heat treatment;

[0063] FIG. 8 is a view of the underside of a coated strip after heat treatment;

[0064] FIG. 9 shows an XPS depth profile of an uncoated strip after heat treatment;

[0065] FIG. 10 is a scanning electron microscopy shot of a coated strip underside;

[0066] FIG. 11 is a diagram illustrating the influence of a coating on the formation of SiO.sub.2 layers on the strip surface;

[0067] FIG. 12 is a diagram illustrating the influence of the dew point of the inert gas atmosphere during the heat treatment process on permeability;

[0068] FIG. 13 is a further diagram illustrating the influence of the dew point of the inert gas atmosphere during the heat treatment process on permeability; and

[0069] FIG. 14 is a diagram illustrating the influence of effective roughness on starting permeability.

DETAILED DESCRIPTION OF EXAMPLE EMBODIMENTS

[0070] FIG. 1 is a diagrammatic representation of an AC-sensitive RCD 1 which disconnects all poles of the monitored circuit from the rest of the network if a specified residual current is exceeded.

[0071] The currents flowing through the RCD 1 are compared in a core-balance transformer 2 which adds the currents flowing to the load with correct signs. If a current in the circuit is discharged to earth, the sum of inward and return current in the core-balance transformer is unequal to zero; the result is a current differential leading to the response of the residual current device 1 and to the disconnection of the power supply.

[0072] The core-balance transformer 2 has a magnet core 2 wound from a nanocrystalline, soft-magnetic strip. The RCD 1 further comprises a tripping relay 4, a preloaded latching mechanism 5 and a test button 6 for manually checking the RCD 1.

[0073] FIG. 2 is a diagrammatic representation of a possible temperature curve of a heat treatment according to a method for producing a magnet core according to an embodiment of the invention.

[0074] In this continuous heat treatment process, an initial heating of the magnet core is followed by a much slower increase or even by a temperature plateau (both alternatives are shown in FIG. 2), in order to let the exothermal crystallization heat decay before the higher temperature used for the maturation of the structure is established. In this way local overheating of the core is avoided. The subsequent maturation of the structure for setting the final magnetic values is then performed at the temperature T.sub.x in the downstream temperature plateau of the maturation zone.

[0075] Using a pre-sample, the temperature in the maturation zone is adapted to the composition of the respective batch in such a way that magnetostriction values become minimal. Of the strip batches to be used, pre-samples are first produced and subjected to different temperatures T.sub.x between 540 C. and 600 C. in the maturation zone. The magnetostriction is then determined either directly on a piece of strip or indirectly on an undamaged core. Direct measurement can for example be performed by means of the SAMR method. An indirect method is a pressure test in which the circumference of the annular strip core is deformed into an oval, for example by 2%. The change in coercitive field strength which occurs in this process is determined by measuring the quasi-static hysteresis loop by means of a Remagraph.

[0076] As FIG. 4 shows, the batch-specific optimum value for T.sub.x can be read at the point where the change H.sub.C is minimal or even tends towards zero.

[0077] On the basis of this method, magnetic values (at 50 Hz) can be obtained in an alloy such as Fe.sub.73.13Co.sub.0.17Cu.sub.1Nb.sub.3Si.sub.15.8B.sub.6.9 on a large scale which lie in the range of .sub.1=120 000-300 000 and .sub.10>450 000, as well as B.sub.r/B.sub.s>70% (measured quasistically). According to FIG. 4, the optimum temperature T.sub.x in this case is approximately 570 C. In an alloy composition Fe.sub.73.41Co.sub.0.21Cu.sub.0.98Nb.sub.2.9Si.sub.15.4B.sub.7.1, on the other hand, the zero cross-over of magnetostriction is only reached at T.sub.x=580 C. to 585 C. In the same way, the optimum temperature found for the alloy Fe.sub.73.38Co.sub.0.11Cu.sub.1.01Nb.sub.2.9Si.sub.16B.sub.6.6 was T.sub.x=564 C.

[0078] If a large quantity of cores is annealed at the same time in large-scale production, a large amount of moisture which adheres to the surface of the strip wound into cores is dragged into the furnace system. On the one hand, this results in direct local corrosive surface reactions on the strip, and on the other hand, some of the moisture is diffused into the inert gas atmosphere and there increases the dew point in an undesirable way. In these conditions, crystalline deposits form on the strip surfaces; as FIG. 3 shows, these largely accumulate in the air pockets. As a surface analysis showed, these crystallites consist of Fe.sub.2O.sub.3, Fe.sub.3O.sub.4 or Nb.sub.2O.sub.5 and are therefore due to oxide reactions during the heat treatment process.

[0079] A further undesirable surface effect supported by increased dew points, which is superimposed on the crystalline deposits, is the growth of a glassy SiO.sub.2 layer. This is rigid and has a considerably lower coefficient of thermal expansion of 0.45 to 1 ppm/K than the strip material (approx. 10 ppm/K). As the bulk material contracts by 1-2% during the generation and maturation of the nanocrystalline grains, mechanical stresses build up. These likewise result in strong anisotropies which affect the magnetic values in an undesirable way.

[0080] The surface sample shown in FIG. 3 was taken from an assembly of 5000 cores having dimensions of 10.5 mm7 mm6 mm, which were wound from a strip having the composition of Fe.sub.73.13Co.sub.0.17Cu.sub.1Nb.sub.3Si.sub.15.8B.sub.6.9. 100 each of these cores were placed endwise on square copper plates having dimensions of 300 mm300 mm6 mm and successively annealed in a continuous furnace at a temperature profile corresponding to FIG. 2. The formation and maturation of the nano-grain occurred at the temperature T.sub.x=575 C., which is the optimum temperature for the zero adjustment of magnetostriction.

[0081] The humidity drawn into the furnace was detected by measuring the dew point of the H.sub.2 inert gas by means of a device called PARAMETRICS MIS1. Before the entry of the annular strip cores into the heating zone, this was 42 C., reaching a comparably high value of 16 C. as the cores passed through the heating zone. Owing to the parasitic anisotropies of the two superimposed surface effects, the magnetic values of the annealed cores were not optimal. The average batch values measured at 50 Hz were in the range of <.sub.1>=47 873, <.sub.10>=222 356, <B.sub.R/B.sub.S>=52% and <H.sub.e>=28 mA/cm.

[0082] To avoid such parasitic effects, the sealing coating of the strip surfaces with an annealing-tolerant substance has proved useful. Suitable materials are dissolved substances the starting materials of which form a thermally stable oxide layer in the annealing process in an H.sub.2, N.sub.2 or Ar inert gas atmosphere or mixtures thereof at temperatures up to 650 C. without being reduced by the effect of the inert gases.

[0083] Examples for base materials for such coatings are Be, Mg, Al, Zr, Ti, V, Nb, Ta, Ce, Nd, Gd and other elements of the 2.sup.nd and 3.sup.rd main groups and the group of rare earth elements. These are applied to the strip surfaces in the form of metal alkoxide solutions in the corresponding alcohol or ether, e.g. methylate, ethylate, propylate or butylate solutions in the corresponding alcohol or ether, or alternatively as tri- or tetra-isopropyl alkoxides. Further alternatives are acetyl-acetone-chelate complexes with the above metals. Under the influence of atmospheric humidity, these are converted into the respective hydrated hydroxides in the subsequent drying process at 80 C. to 200 C. In the later heat treatment process, this releases further water and becomes the respective metal oxide, resulting in a dense protective layer which adheres firmly to the surface and seals it. Typical layer thicknesses lie in the range of 0.05 to 5 m, a layer thickness of 0.2 to 1 m having sufficiently good properties and therefore being preferred in one embodiment.

[0084] With the coating, the material properties can be stabilized against surface reactions at the high temperatures required for the zero adjustment of magnetostriction. The application-relevant characteristic values influenced by surface effects are in particular the (H) characteristic measured at 50 Hz, the quasi-static coercitive field strength and the remnant induction.

[0085] At least three possible methods are available for applying the solution as starting product for the later formation of the sealing coating. The layer thicknesses referred to above can be obtained by adjusting concentration and by adapting the process parameters. If particularly thick layers are required, the process can be repeated.

[0086] In one possible method, the strip is continuously drawn via deflection rollers through the coating medium placed in a trough. Immediately before being wound to form a core, it passes through a drying section at a controlled temperature of 80-200 C. This Method results in a particularly uniform coating. Thicker layers can be obtained by a repeated passage.

[0087] In a further possible method, the strip, after being wound following its production, is dipped into the solution in a receiver in the form of a coil and evacuated. Owing to the effective capillary forces, which are sufficiently strong at a vacuum in the rough vacuum range of 10-300 mbar, the solution penetrates between the strip layers of the coil and wets the surfaces. The dried coils are then post-dried in a drying cabinet at 80-200 C. The coated strip is then wound to form magnet cores. This method is particularly economical.

[0088] In a further possible method, the cores wound from uncoated strip are dipped into the solution in a receiver. Following evacuation to the above vacuum, the solution penetrates between the strip layers and wets them. The dipped cores are then dried in a drying cabinet at 80-200 C. This method offers the advantage that the winding of the core cannot be affected by the coating medium on the strip surfaces.

[0089] Investigations have revealed that coatings with magnesium and zirconium are particularly easily processed, cost-effective and safe in processing.

[0090] The concentration of the dissolved metals was varied in the various organic solvents within a wide range between 0.1% and 5% by weight without causing any significant changes in the magnetic values. At very low concentrations, however, standard deviations were found to increase.

[0091] To check the effect of a surface coating, strips of the composition Fe.sub.73.6Co.sub.0.1Cu.sub.1Nb.sub.2.96Si.sub.15.45B.sub.6.84C.sub.0.05 produced in a melt spinning process and having a width of 10 mm were divided into three part-quantities of identical quality (fill factor =81.0-81.3%, R.sub.a(eff)=2.9%). The first and second part-quantities remained uncoated, while the third part-quantity was coated with a solution of 3.6% Mg-methylate in a receiver in a dipping process. The rough vacuum generated by means of a rotary slide-valve pump was approximately 110 mbar at the end of evacuation time. After a dwell time of 15 minutes, the saturated coils were dried at 110 C. for one hour, resulting in an adhesive layer of hydrated Mg(OH).sub.2 with a thickness of 0.8 m.

[0092] Both the coated and the uncoated strips were then wound at a descending skew to produce strain-free annular strip cores having dimensions of 32 mm16 mm10 mm. In preparation of heat treatment, 100 cores each were placed endwise in square copper plates having dimensions of 300 mm300 mm6 mm.

[0093] The subsequent heat treatment was carried out entirely field-free in a continuous process at a temperature profile similar to that shown in FIG. 2, the throughput speed through the heating zone being 0.16 m/min. Pure hydrogen with a dew point of 50 C. was used as an inert gas. Contrary to the presentation in FIG. 2, the temperature gradient in the first heating zone was increased such that the products reached a temperature of 480 C. after only 8 minutes. The temperature in the decay zone was not held constant, but increased to 505 C. along a 20 minute heating section. This was followed by a steep temperature gradient which the cores passed through within 3 minutes to reach the final maturation temperature T.sub.x. The passage through this temperature range was completed within 25 minutes. The cores were then cooled to room temperature at the same throughput speed in a cooling zone significantly longer than that shown in FIG. 2 in the presence of hydrogen of the same dew point. This greatly reduced cooling rate was chosen to avoid cooling-related strain effects.

[0094] To avoid overheating, which, together with atmospheric impurities, can result in increased surface reactions and thus in parasitic anisotropies, the maturation zone was for the first third of the cores made from uncoated strip adjusted as low as possible to T.sub.x=520 C. The (H) characteristic measured at 50 Hz and the quasi-statically (f=0.01 Hz) measured hysteresis loops shown in FIGS. 5 and 6 show by way of example that after a heat treatment at T.sub.x=520 C. high maximum permeability values of .sub.s=719 827 are reached, the starting permeability being .sub.1=105 238. The remanence ratio of B.sub.R/B.sub.S was approximately 77%.

[0095] To protect against mechanical stresses caused by handling or processing steps, such as wire or conductor winding, these cores were bonded endwise into Ultramid troughs using silicone rubber as an adhesive. Owing to the magnetostriction of .sub.s measured by means of the SAMR method, the adhesive penetrating between the strip layers increased the quasi-static coercitive field strength from H.sub.c=3.9 mA/cm to 8.6 mA/cm, while the maximum permeability measured at 50 Hz was reduced to .sub.16=373 242 and B.sub.R/B.sub.S was reduced to 59%. Owing to their inadequate permeability, such cores were not suitable for use in RCDs.

[0096] The second third of the cores, uncoated like the first third, was annealed at a temperature T.sub.x=575 C., which in the pre-sample was found to be optimal for the zero adjustment of magnetostriction, to .sub.20 ppm.

[0097] In this case, however, the maximum permeability was reduced to 221 435, and the quasi-statically measured coercitive field strength of H.sub.c=13.2 mA/cm was found to be very high-see FIGS. 5 and 6. The remanence ratios were only around 51%.

[0098] To analyze the cause of these worse figures, the strip surfaces of the cores were checked by means of optical microscopy. As FIG. 7 shows, the air pockets on the underside of the strip were stratified with a dense layer of crystalline deposits which resulted in major parasitic anisotropies and a considerable degradation of the magnetic values. The surface analysis likewise performed on the underside of the strip by means of XPS (X-ray photoelectron spectroscopycf. Stefan Huffier, Photoelectron Spectroscopy Principles and Applications, Springer, 3.sup.rd edition, 1995/1996/2003) showed in the depth profile according to FIG. 9 in addition the existence of a highly straining SiO.sub.2 surface layer, which leads to major parasitic anisotropies. The structure of this layer is due to a segregation of Si atoms from the strip interior, followed by oxidation by residual atmospheric impurities.

[0099] The last third of the cores, which was coated with a 3.6% solution of Mg methylate, on the other hand, exhibited after annealing at T.sub.x=575 C., very good values as shown in FIGS. 5 and 6: H.sub.c was approximately 7 mA/cm, maximum permeability approximately .sub.8=692 163, B.sub.R/B.sub.S approximately 79%. At the same time, starting permeability .sub.1 rose to 243 562. Owing to the largely balanced magnetostriction of .sub.s0.1 ppm, a single-trough experiment using a silicone rubber adhesive resulted in a virtually unchanged permeability of .sub.8=679 322. Comparable results were obtained with cores which were not bonded into a trough, but loosely installed with a 2 mm thick rubber cushioning ring placed on their end faces.

[0100] As the scanning electron microscopy investigation of the strip surfaces as shown in FIG. 10 indicates, the strip surface of the last third of the cores was covered by a dense MgO sinter layer after annealing. As FIG. 8 shows clearly, this prevents the formation of surface crystallites in the air pockets. At the same time, the evaluation of XPS depth profiles recorded in individual sample states and shown in FIG. 11 indicates than an Mg coating suppresses the formation of a strain-inducing SiO.sub.2 surface layer. Similar results were obtained with coatings of 1.7% Zr-tetra-isopropyl alkoxide and 4% phenyl titanium tri-isopropyl alkoxide.

[0101] In the course of these investigations, the dew point of the H.sub.2 and N.sub.2 inert gas was discovered to be a further critical parameter in the production of maximum-permeability, magnetostriction-free magnet cores. This becomes more significant as the temperature required for the balance of magnetostriction increases. To investigate this effect, a large number of test annealing processes was performed in the continuous furnace on assemblies of 100 cores having dimensions of 26 mm10 mm6 mm produced from a strip of the composition Fe.sub.73.13Co.sub.0.17Cu.sub.1Nb.sub.3Si.sub.15.8B.sub.6.9. The strips used had an effective roughness R.sub.a(eff) of approximately 3% and a fill factor of about 81.5%. The cores were produced in the way described above. The whole strip was coated with a 2.4% solution of Mg methylate.

[0102] In the heat treatments, the dew point was varied between 20 C. and 55 C. by mixing humidified and dry H.sub.2 gas. A device PARAMETRICS MIS1 was used to measure the dew point.

[0103] In these atmospheres, the test cores were annealed on copper plates using the temperatures described with reference to FIG. 2. However, in a first passage the temperature in the maturation zone was adjusted to T.sub.x=540 C. without taking account of magnetostriction balance. From the averages of the permeability values measured at 50 Hz and =11.27 mA/cm as shown in FIG. 12, we can conclude that in these conditions a dew point of T.sub.p25 C. is required to obtain .sub.11.27(.sub.max400 000. As expected, all cores proved to be magnetostrictive in a deformation test and could therefore not be processed using the single-trough method commonly applied to magnetostriction-free cores. Special non-straining single-trough methods were required.

[0104] In a second run, the optimum temperature for magnetostriction balancing, T.sub.x=570 C., which had previously been determined in a pre-sample, was set. The average permeability values measured at 50 Hz and a field strength of 11.27 mA/cm are shown in FIG. 13. It can be seen that in these conditions a dew point of T.sub.p49.5 C. is required to obtain .sub.11.27(.sub.max)400 000.

[0105] In a further test series for limiting the influencing parameters, strip of the composition Fe.sub.73.13Co.sub.0.17Cu.sub.1Nb.sub.3Si.sub.15.8B.sub.6.9 and having a width of 6 mm was cast on the melt spinning line until the originally almost perfect surface of the casting roll exhibited considerable traces of wear. This wear resulted along the length of the strip in a continuous quality loss reflected in increased surface roughness. The cast strip was wound into coils of approximately the same size, and samples were taken from the beginning, the middle and the end of the coil. These samples were on both surfaces subjected to a measurement of their roughness R.sub.a in a tactile traverse scanning process, and the average thickness of the strip was calculated from the specific weight (as cast 7.07 g/cm.sup.3), the length, width and weight of the strip sample. Finally, the effective roughness R.sub.a(eff) of the strip samples were determined by dividing the sum of the R.sub.a values of the two surfaces by the strip thickness.

[0106] The completely wound coils were coated with three layers of a solution of 19% Zr-tetra-isopropyl alkoxide and then dried for one hour at 130 C. The whole strip was then wound into cores having dimensions of 26 mm10 mm6 mm in a strain-free process, maintaining the sequence of cores and their assignment to the original coils. This made it possible to assign to specific cores positions within the coils and therefore a value for R.sub.a(eff). After 50 cores each had been placed endwise on square copper plates having dimensions of 300 mm300 mm6 mm, a continuous annealing process was carried out using the temperature profile described above with a maturation temperature T.sub.x=570 C.

[0107] To determine the starting permeability, which depends on strip geometry, the .sub.1 values of the cores were measured at 50 Hz and plotted above the effective roughness in FIG. 14. As FIG. 14 shows, an effective roughness of R.sub.a(eff) 7% is required for obtaining .sub.1100 000. If .sub.1 is to be higher than 160 000, R.sub.a(eff) has to be less than 5%, and for .sub.1200 000 even less than 2.5%.

[0108] In the test series described above, the annealing process was carried out at a dew point of 53 C. and T.sub.x=570 C., which a SAMR magnetostriction measurement indicated to result in .sub.s=0.1 ppm. In view of this, the cores could be bonded into a plastic trough by means of silicone rubber or installed loosely into a plastic or metal protective trough by means of a mechanically damping foam rubber ring without changing their permeability in a significant way.

[0109] The results of the investigation are summarised in Table 1. The mark *) indicates fixing with silicone rubber and the mark **) indicates strain-free fixing with a high-viscosity acrylate adhesive.

TABLE-US-00001 TABLE 1 Strip thickness T.sub.x .sub.1 .sub.max .sub.max Alloy Core dimensions [m] Coating [ C.] unfixed unfixed fixed Fe.sub.73.13Co.sub.0.17Cu.sub.1 26.3 10.5 6.2 19.5 None 520 97 566 719 827 373 242 *) Nb.sub.3Si.sub.15.8B.sub.6.9 687 688 **) Fe.sub.73.13Co.sub.0.17Cu.sub.1 26.3 10.5 6.2 19.5 None 575 105 311 221 435 209 432 Nb.sub.3Si.sub.15.8B.sub.6.9 Fe.sub.73.13Co.sub.0.17Cu.sub.1 26.3 10.5 6.2 19.5 Mg methylate (3%) 575 244 562 692 163 677 322 Nb.sub.3Si.sub.15.8B.sub.6.9 Fe.sub.73.13Co.sub.0.17Cu.sub.1 26.3 10.5 6.2 21.0 Mg methylate (3%) 575 178 364 618 215 607 224 Nb.sub.3Si.sub.15.8B.sub.6.9 Fe.sub.73.13Co.sub.0.17Cu.sub.1 26.3 10.5 6.2 24.0 Mg methylate (3%) 575 63 078 188 474 Nb.sub.3Si.sub.15.8B.sub.6.9 Fe.sub.73.13Co.sub.0.17Cu.sub.1 26.3 10.5 6.2 19.5 Mg methylate (0.3%) 575 229 528 642 999 639 623 Nb.sub.3Si.sub.15.8B.sub.6.9 Fe.sub.73.13Co.sub.0.17Cu.sub.1 26.3 10.5 6.2 19.5 Ti propylate (1%), 575 198 466 621 523 615 872 Nb.sub.3Si.sub.15.8B.sub.6.9 3 layers Fe.sub.73.13Co.sub.0.17Cu.sub.1 26.3 10.5 6.2 19.5 Ti butylate (4%), 550 132 321 588 478 368 662 *) Nb.sub.3Si.sub.15.8B.sub.6.9 4 layers 581 014 **) Fe.sub.73.13Co.sub.0.17Cu.sub.1 26.3 10.5 6.2 19.5 Zr propylate (2%), 575 192 833 647 174 642 445 Nb.sub.3Si.sub.15.8B.sub.6.9 3 layers Fe.sub.73.13Co.sub.0.17Cu.sub.1 26.3 10.5 6.2 19.5 K methylate (3%) 575 47 642 68 540 Nb.sub.3Si.sub.15.8B.sub.6.9 Fe.sub.72.13Co.sub.0.17Cu.sub.1 26.3 10.5 6.2 19.5 K propylate (0.3%) 575 51 684 86 262 Nb.sub.3Si.sub.15.8B.sub.6.9 Fe.sub.72.53Co.sub.0.11Cu.sub.1.1 26.3 10.5 6.2 19.5 Mg methylate (4%) 585 173 354 662 551 392 444 *) Nb.sub.3Si.sub.16.5B.sub.6.75 658 676 **) Co.sub.0.11 Fe.sub.72.53Co.sub.0.11Cu.sub.1.1 26.3 10.5 6.2 19.5 Mg methylate (4%) 562 209 471 708 422 706 843 Nb.sub.3Si.sub.16.5B.sub.6.75 Co.sub.0.11 Fe.sub.72.43Co.sub.0.08Cu.sub.0.98 26.3 10.5 6.2 19.5 Mg methylate (4%) 562 126 927 565 618 382 464 *) Nb.sub.2.9Si.sub.15.45B.sub.6.95 529 930 **) Co.sub.0.21 Fe.sub.72.43Co.sub.0.08Cu.sub.0.98 26.3 10.5 6.2 19.5 Mg methylate (4%) 585 231 738 712 486 709 686 Nb.sub.2.9Si.sub.15.45B.sub.6.95 Co.sub.0.21 Fe.sub.73.13Co.sub.0.17Cu.sub.1 10.5 7 4.5 19.5 Mg methylate (3%) 575 188 431 629 644 632 381 Nb.sub.3Si.sub.15.8B.sub.6.9 Fe.sub.73.13Co.sub.0.17Cu.sub.1 180 140 20 19.5 Mg methylate (3%) 575 172 524 646 813 631 117 Nb.sub.3Si.sub.15.8B.sub.6.9