EP2643839B1 - Soft magnetic metallic ribbon for electromechanic element - Google Patents

Description

The application relates to a soft magnetic metal strip for electromechanical components, in particular AC residual current switch.

The EP 2 015 321 A1 discloses a magnetic core for a current transformer composed of an amorphous alloy, the composition being represented by the general formula Fe _100-xz Ni _x X ' _z , where X' is composed of Si and B.

A method for manufacturing a magnetic core for soft magnetic alloy AC fault current circuit breaker is disclosed in U.S.P. US 5,922,143 disclosed. An amorphous band of an iron-based alloy is produced by a rapid solidification technology, wound into a magnetic core and then heat-treated to produce a nanocrystalline structure.

This magnetic core is less affected by mechanical stresses so that the desired permeability value is more reliably achieved. However, there is still a need for further improvements.

The object of the application is to provide a soft magnetic metal strip for electromechanical components, which is particularly suitable for applications at 50 Hz, such as AC fault current switches, which can be produced reproducibly.

According to the invention, a soft magnetic metal strip is created for electromechanical components. The soft magnetic metal strip has a nanocrystalline or an amorphous structure and ratios of strip thickness to roughness d / Ra of ≦ 5 d / Ra ≦ 25, where Ra is the center roughness value. The metal band has a fish scale pattern with a structure arranged transversely and obliquely to the band longitudinal direction.

Magnetization properties of this soft magnetic metal strip are dependent on a strip thickness d and a roughness Ra. A ratio of belt thickness to roughness d / Ra of ≦ 5 d / Ra ≦ 25 enables the improvement of the permeability in AC applications as well as the reliable generation of this improved permeability.

The roughness Ra of this ratio is the measured roughness of the underside of the soft magnetic metal strip, the lower surface being the side of the soft magnetic metal strip which lies on a casting wheel during the solidification of a melt.

This soft magnetic metal strip has the advantage that electromechanical components with short response times such as fault current switch or speed sensors can be realized with a toroidal core, which can trigger a switching operation at low coercive force of a few 10 milliamps per centimeter or signal the Vorbeidrenhen a permanent magnet to speed measurements instead to enable Hall generators.

The soft magnetic metal strip has a nanocrystalline or amorphous structure. The soft magnetic metal strip is characterized by a nearly rectangular hysteresis loop and low eddy current losses, both of which can be used for fast-reacting electromechanical devices such as fault current switches used at 50 Hz.

In one embodiment, the soft magnetic metal strip contributes a maximum value of the magnetic induction values With an average roughness of approximately 1 μm, this means a strip thickness d of the soft magnetic metal strip between 10 μm ≦ d ≦ 20 μm.

The metal band may have a ratio Br / Bm> 80%, where Bm is measured at 200 mA / cm.

The metal band has a fish scale pattern with a structure arranged transversely and obliquely to the band longitudinal direction. Such a pattern can be selectively adjusted, for example, by reducing the casting pressure and / or increasing the casting wheel speed. Other options for selectively influencing the surface topology of the strip include, for example, surface structuring of the casting roll or subsequent laser scribing of the soft magnetic metal strip.

Comparative tests have shown that the influence of these geometric parameters is stronger than the influence of the alloy composition of the metal strip with alloy constituents of silicon, boron, niobium and copper in more than 73 atomic% iron.

Thus arise for a metal strip of an alloy comprising Fe _75.5 Cu ₁ Nb ₃ Si _12.5 B _8, and a metal belt having an alloy with Fe _73.5 Cu ₁ Nb ₃ Si _13.5 B _9, no serious differences, so that both alloys have maximum induction values in the roughness ratio thicknesses given above. Comparatively, strip qualities of thinner and rougher metal strip have better AC magnetic values, The alloying with Fe _73.5 Cu ₁ Nb ₃ Si _13.5 B ₉ shows no serious differences, so that both alloys have maximum induction values in the above-indicated thickness to roughness ratio. By comparison, strip qualities of thinner and rougher metal strip have better AC magnetic values than a smooth, zero-roughness, and more than 50 μm thick metal strip.

For this purpose, it is advantageous to heat the metal strip thermally at a temperature between 500 ° C and 600 ° C for a period of 0.5 hours to 2 hours in a longitudinal field of 5 A / cm to 15 A / cm. After the annealing treatment, the metal strip has a quasi-stationary coercive force independent of a strip quality in terms of strip thickness and roughness, and an AC-determined coercive force of the strip metal increases linearly with a strip thickness-to-roughness ratio d / Ra.

The metal strip according to any one of the preceding embodiments may be wound to indicate a magnetic core. This magnetic core can be used in various applications, for example in applications at frequencies of less than 1000Hz, such as an AC fault current switch, since the magnetic core has good permeability even at 50 Hz, or in distribution transformers.

It is envisaged that this soft magnetic metal strip is used for ac-sensitive electromechanical components with a soft-magnetic annular band core. As already mentioned above, the soft-magnetic metal strip can be used for residual current switches with a residual current limit value I _max ≦ 30 mA. Furthermore, a use of the soft magnetic metal strip for a speed sensor in cooperation possible with a segmented permanent magnet disc.

An AC-sensitive leakage circuit breaker is also disclosed, comprising a magnetic core made of a wound soft magnetic strip according to one of the preceding embodiments.

In one embodiment, at a frequency of less than 1000 Hz, in particular at 50 Hz, the magnetic core of the ac-sensitive leakage circuit breaker has a ratio Br / Bm> 80%.

A distribution transformer is also disclosed, comprising a magnetic core of a wound soft magnetic ribbon according to one of the preceding embodiments.

The magnetic core of the distribution transformer may have a ratio Br / Bm> 80% at a frequency of less than 1000 Hz, in particular at 50 Hz.

Fig. 1

zeigt ein Diagramm zur Definition einer dynamischen Koerzitivfeldstärke;

Fig. 2

zeigt ein Diagramm mit Werten von Banddicke und Rautiefe an verschiedenen Stellen eines untersuchten Metallbands;

Fig. 3

zeigt ein Diagramm mit Induktionsamplituden (Figur 3a) und Amplitudenpermeabilitäten (Figur 3b) als Funktion der erregenden Feldamplitude bei sinusförmiger Erregung mit 1 Hz für unterschiedliche Metallbandqualitäten;

Fig. 4

zeigt ein Diagramm mit Induktionsamplituden (Figur 4a) und Amplitudenpermeabilitäten (Figur 4b) als Funktion der erregenden Feldamplitude bei sinusförmiger Erregung mit 50 Hz für die Metallbandqualitäten aus Figur 3;

Fig. 5

zeigt ein Diagramm der dynamischen Koerzitivfeldstärke als Funktion des Verhältnisses von Banddicke und Rautiefe einer einzelnen Charge;

Fig. 6

zeigt ein Diagramm mit Induktionsamplituden bei einer erregenden Feldstärke von 10 mA/cm als Funktion des Verhältnisses von Banddicke und Rautiefe einer ersten Metallbandlegierung;

Fig. 7

zeigt ein Diagramm mit Werten von Banddicke und Rautiefe unterschiedlicher Chargen von Metallband;

Fig. 8

zeigt ein Diagramm mit Induktionsamplituden und Amplitudenpermeabilitäten als Funktion der erregenden Feldamplitude bei sinusförmiger Erregung mit 50 Hz für unterschiedliche Metallbandqualitäten im Vergleich einer nanokristallinen Legierung zu einer grobkristallinen NiFe-Legierung Ultraperm 10 mit der Kennlinie 020;

Fig. 9

zeigt ein Diagramm der dynamischen (50 Hz) und statischen (dc) Koerzitivfeldstärke als Funktion des Verhältnisses von Banddicke und Rautiefe unterschiedlicher Chargen;

Fig. 10

zeigt ein Diagramm mit Induktionsamplituden bei einer erregenden Feldstärke von 10 mA/cm als Funktion des Verhältnisses von Banddicke und Rautiefe einer zweiten Metallbandlegierung;

Fig. 11a

zeigt ein Diagramm mit Induktionsamplituden bei 50 Hz von drei nanokristallinen Metallbandqualitäten im Vergleich zu einem grobkristallinen Ultraperm 200 Metallband der NiFe-Legierung;

Fig. 11b

zeigt ein Diagramm der Änderungen der Induktionsamplituden nach einem Gleichfeldstoß und nach Klopfbelastung;

Fig. 12

zeigt ein Diagramm der dynamischen Koerzitivfeldstärke als Funktion des Verhältnisses von Banddicke und Rautiefe eines amorphen VITROVAC 6030 Z Bandmaterials;

Fig. 13

zeigt ein Diagramm mit Induktionsamplituden bei einer erregenden Feldstärke von 10 mA/cm als Funktion des Verhältnisses von Banddicke und Rautiefe eines amorphen VITROVAC 6030 Z Bandmaterials.

The invention is based on diagrams of the following figures, which show in detail:

Fig. 1

shows a diagram for defining a dynamic coercive field strength;

Fig. 2

shows a diagram with values of strip thickness and surface roughness at different points of a metal strip under investigation;

Fig. 3

shows a diagram with induction amplitudes ( FIG. 3a ) and amplitude permeabilities ( FIG. 3b ) when Function of the exciting field amplitude with sinusoidal excitation at 1 Hz for different metal strip qualities;

Fig. 4

shows a diagram with induction amplitudes ( FIG. 4a ) and amplitude permeabilities ( FIG. 4b ) as a function of the exciting field amplitude for sinusoidal excitation at 50 Hz for the metal strip qualities FIG. 3 ;

Fig. 5

Figure 12 is a plot of dynamic coercivity vs. function of the ratio of strip thickness and surface roughness of a single lot;

Fig. 6

Figure 10 shows a graph of induction amplitudes at an excitation field strength of 10 mA / cm as a function of the ratio of strip thickness and roughness of a first metal strip alloy;

Fig. 7

shows a diagram with values of strip thickness and surface roughness of different batches of metal strip;

Fig. 8

Figure 10 shows a graph of inductance amplitudes and amplitude permeabilities as a function of exciting field amplitude at 50 Hz sinusoidal excitation for different metal strip grades compared to a nanocrystalline alloy to a Coarse-Cr NiFe alloy Ultraperm 10 having the characteristic 020;

Fig. 9

shows a plot of dynamic (50 Hz) and static (dc) coercive field strength as a function of Ratio of strip thickness and surface roughness of different batches;

Fig. 10

Figure 10 shows a graph of induction amplitudes at an excitation field strength of 10 mA / cm as a function of the ratio of strip thickness and roughness of a second metal strip alloy;

Fig. 11a

Figure 4 shows a 50 Hz induction amplitudes plot of three nanocrystalline metal strip grades as compared to a Co-crystalline Ultraperm 200 metal strip of NiFe alloy;

Fig. 11b

Fig. 10 is a graph showing the changes of the induction amplitudes after a DC field impact and a knocking load;

Fig. 12

FIG. 12 is a graph of dynamic coercive force vs. function of the ratio of strip thickness and surface roughness of an amorphous VITROVAC 6030 Z strip material; FIG.

Fig. 13

Figure 10 shows a graph of induction amplitudes at an excitation field strength of 10 mA / cm as a function of the ratio of strip thickness and surface roughness of an amorphous VITROVAC 6030 Z strip material.

As part of the production transition of the nanocrystalline alloy Fe _75.5 Cu ₁ Nb ₃ Si _12.5 B ₈ , a puncture test was carried out to investigate the influence of strip thickness and band roughness on the magnet quality achievable after heat treatment. For this purpose, in a batch of strip (KA 1283, use 7 kg, belt width 15 mm) by lowering the roller speed during the Schussverlaufes the band thickness of about 16 microns (microns) varies to about 35 microns. By lowering the casting pressure towards the end of the weft, it was finally attempted to increase the tape roughness with a constant thickness (by 35 μm). From this batch band samples of varying thickness and roughness were taken over the course of the shot and wound therefrom ring band cores (RBK) measuring 22 mm × 16 mm × bandwidth for further magnetic examinations.

Included in the investigations were also a number of different KA batches of the alloy Fe _73.5 Cu ₁ Nb ₃ Si _13.5 B ₉ , in which the production parameters were varied with a view to improving the ductility in the production state. From these batches ring wire cores measuring 22 mm x 16 mm x bandwidth were also produced for magnetic investigations.

The toroidal cores were tempered longitudinally in a production furnace under hydrogen atmosphere. The exact starting conditions were: 1h holding time at 540 ° C in the longitudinal field (alternating field of about 10 A / cm), cooling at 1 K / min.

Measured were the quasistatic hysteresis loops, as well as the 50Hz commutation curves. The 50 Hz characteristics were measured in the "decreasing excitation field", which corresponds to a measurement of the demagnetized core. From the commutation curves was accordingly FIG. 1 the dynamic coercive field strength H _{c is} evaluated, ie in the following diagrams of the figures, H _c (dyn) designates that exciting field strength at which the saturation or maximum permeability is just about reached.

To characterize the band geometry, the mean band thickness was determined from the weight of the meter and the roughness at the band underside (transverse to the band direction) was measured.

example 1

Example 1 is based on investigations with respect to roughness and strip thickness with the alloy Fe _75.5 Cu ₁ Nb ₃ Si _12.5 B ₈ as batch KA 1283

The ratio of strip thickness to surface roughness d / R _a ranges from about 10 to 60. The FIG. 2 shows the associated variation of strip thickness d and roughness Ra. It is in FIG. 2 recognizable that the thinner tapes usually have an absolutely greater roughness.

The mean, quasi-statically measured coercive force is $${\mathrm{H}}_{\mathrm{c}}=7.5\pm 1\mathrm{}\mathrm{mA}/\mathrm{cm}\mathrm{}({\mathrm{at\; H}}_{\mathrm{Max}}=50\mathrm{}\mathrm{mA}/\mathrm{cm}$$

The mean, quasi-statically measured remanence ratio is $${\mathrm{B}}_{\mathrm{r}}/{\mathrm{B}}_{\mathrm{m}}=0.97\pm \mathrm{12:01}\mathrm{}\left({\mathrm{at\; H}}_{\mathrm{Max}}=200\mathrm{}\mathrm{mA}/\mathrm{cm}\right),$$

FIG. 3 shows the commutation curve measured at 1 Hz (sinusoidal field strength) (B _max over H) and the associated amplitude permeability μ.

With the thicker and smoother bands suggestively better magnet values can be seen. Overall, the influence of strip thickness and band roughness determined here is on However, the quasi-static and 1Hz measured values are low and are almost lost in the measurement accuracy. In contrast, in comparative studies on amorphous materials such as VC 6150 Z and VC 6030 Z, a significant deterioration of H _c and B _{r was} observed with increasing ratio of R _a / d.

At higher frequencies the situation looks very different. Even at 50 Hz, there are clear differences between the different band qualities. FIG. 4 shows the measured at 50 Hz (sinusoidal field strength) commutation curve (B _max over H), and the associated amplitude permeability μ. The comparison of the 1 Hz and 50 Hz characteristic curves shows that the 50 Hz characteristic is almost completely determined by anomalous eddy currents. As already explained above, the classical eddy current contributions are negligible. It is noticeable that in particular the thin and rough bands deliver by far better 50Hz magnet values than the thicker smooth bands.

FIG. 5 shows the dynamic coercive force H _c (as defined in FIG. 1 ) As a function of the ratio d / R _a. The values for f = 1 Hz coincide with the quasi-statically measured values and show, as already mentioned, virtually no dependence on d / R _a . The swirl current-determined 50 Hz values, on the other hand, increase linearly with d / R _a , which corresponds to the theoretical expectations set out above. The extrapolation of the 50 Hz values for d / R _a ≥ 0 leads to a value that roughly corresponds to the quasistatic coercive field strength.

For the possible application in 30mA residual current circuit breakers, the induction amplitude B _{10 is} relevant for an exciting field amplitude of 10 mA / cm. The consequences of the relationship d / R _a for this size are in FIG. 6 shown. B ₁₀ increases significantly with decreasing strip thickness and increasing roughness. The maximum value of B ₁₀ , ie B ₁₀ approximately equal to B _s , is expected when the dynamic coercive force H _c becomes less than 10 mA / cm. This is realized eg for smaller frequencies.

Example 2

Example 2 is based on investigations with regard to roughness and strip thickness with the alloy Fe _73.5 Cu ₁ Nb ₃ Si _13.5 B ₉ from various batches, as well as investigations of further influencing parameters.

The results obtained on different batches of the alloy Fe _73.5 Cu ₁ Nb ₃ Si _13.5 B ₉ are given in the Figures 7-11 shown. Here too, the just discussed dependence of the magnet values on the band roughness and band thickness arises. In contrast to the case discussed above, here the influence of the roughness is expressed somewhat more explicitly, since the mean strip thickness does not vary as strongly as in Example 1 ( FIG. 7 ).

Static coercivity values are about 3.5 mA / cm. This significantly lower value compared to the alloy of Example 1 Fe _75.5 Cu ₁ Nb ₃ Si _12.5 B ₈ is partly due to the fact that the static loop was only controlled at approximately 20 mA / cm. With a comparable modulation of H _max = 50 mA / cm, slightly higher H _c values result around H _c = 5 mA / cm.

The 50 Hz magnet values in the present case (Fe _73.5 Cu ₁ Nb ₃ Si _13.5 B ₉ ) are somewhat more favorable than those of the alloy Fe _75.5 Cu ₁ Nb ₃ Si _12.5 B ₈ . This is most clearly expressed when comparing the B ₁₀ values at the same d / R _a ratio (cf. FIG. 8 and 10 ). It should be noted that the alloy contains 13.5 atomic% Si and λ _s = 2x10 ^-6 and K _u = 20 J / m ³ , both a smaller magnetostriction constant λ _s by a factor of two, and a smaller magnetic field-induced anisotropy K _u (the corresponding values for the 12.5 atomic weight% Si alloy are: λ _s = 3.5x10 ^-6 and K _u = 40 J / m ³ ). Thus, the alloy with Si 13.5 atomic% has more favorable basic requirements both for the pure soft magnetic properties and for the dynamic properties.

Upon closer inspection of the Figures 9 and 10 an outlier is noticed (marked by an exclamation mark); this is a band from a batch KA 1114. Noticeable are, despite relatively small d / R _a ratio, relatively poor dynamic properties, as well as more than twice as high compared to the other batches static coercive field strength. Furthermore, this band also had a hysteresis loop greatly shifted with respect to the field zero point. A striking feature of the band geometry was a longitudinal pitch more than 5 μm deep, which was not included in the Ra value, ie it was not measured over this score during the measurement. This groove is due to slag particles in the nozzle, which eventually led to splitting of the belt and ultimately premature firing.

1. 1. Eine Reduzierung der Induktionswerte um bis zu 50-100 mT nach Gleichfeldvorbelastung (im vorliegenden Fall ca. 1A/cm).
2. 2. Eine Anhebung der Induktionswerte bis zu 100-200 mT nach Klopfen des Kernes.

For materials with a rectangular hysteresis loop, the magnetization characteristic is relatively sensitive to the exact measurement conditions and the state of the core before the measurement. The characteristics discussed so far were measured in a decreasing magnetic field (corresponds to a measurement of the demagnetized core) on cores in the state "as annealed, a few days after". In comparison to such a reference curve, it follows:

1. 1. A reduction of the induction values by up to 50-100 mT after DC field stress (in the present case approx. 1A / cm).
2. 2. Increasing the induction values up to 100-200 mT after knocking the core.

FIG. 11 shows these changes as a function of the exciting field together with the BH commutation curve in comparison with Ultraperm 200. It can be seen that the changes of the soft magnetic nanocrystalline metal strip are significantly larger than in the case of the ultracrystalline Ultraperm 200 metal strip of a NiFe alloy.

Noteworthy are the changes in the magnet values after a knock of the metal strip. While the alloy compositions Fe _73.5 Cu ₁ Nb ₃ Si _13.5 B ₉ examined here showed a significant change in the magnet values, the alloy Fe _75.5 Cu ₁ Nb ₃ Si _12.5 B ₈ (despite higher magnetostriction) was also a Ultraperm 200 virtually no change to watch. Currently, the phenomenon can only be understood as an indication that the handling of the cores can influence the magnet values on a case-by-case basis.

The influence of a DC field impact, however, was much more reproducible and observed in all cores studied. The phenomenon reflects the irreversible nature of the underlying magnetization processes (pinning of domain walls, after-effects). This effect may be uncontrolled when measuring, namely when switching the current on and off in a specific operating point, depending on the ohmic load in the primary and secondary circuits. This is due to inductive retroactivity-related surges that can temporarily drive the material into saturation.

In practical terms, these effects mean that the induction values for nanocrystalline materials with a rectangular hysteresis loop at a given operating point for one and the same core can currently only be specified with a tolerance of about ± 100 mT.

Example 3

Example 3 is based on tests on an amorphous comparison material VITROVAC 6030 Z (Z stands for material with a rectangular hysteresis loop) with regard to roughness and strip thickness.

For comparison with the nanocrystalline material show the Figures 12 and 13 a corresponding evaluation at 50 Hz on VITROVAC 6030 Z measured data. For d / R _a > 15 an analogue picture results. For d / R _a <15 {d = 15 μm, R _a = 1.5 μm to 3 μm), however, a deterioration of the magnet values is observed here in contrast to the linear increase in nanocrystalline metal bands. This is essentially due to the influence of the roughness on the quasi-static coercive field strength and the remanence magnetization.

The results for VITROVAC 6030 Z make it clear that the improvement of the dynamic properties by reducing the strip thickness and in the case of amorphous metal strips with rectangular hysteresis loops are set by raising the band roughness limits.

The present investigations show that the 50 Hz properties of amorphous and nanocrystalline materials with a rectangular hysteresis loop are decisively determined by the ratio of strip thickness d to surface roughness R _a . The influence of d / R _{a is} at least as important as the influence of the uniaxial anisotropy K _u induced in the direction of the ribbon. With regard to the production safety d / _Ra, gained nearly a crucial role because this parameter is much more difficult to master than the well-defined by the alloy composition and annealing treatment induced anisotropy K _u.

The best magnetic properties were previously (15 to 20 microns) to thin and rough bands (R _a = 1 to 1.5 microns) with d / R ratios observed _a 10 to 20 However, earlier studies on VITROVAC 6030 Z also indicate that the magnet values are deteriorated again due to increasing hysteresis losses if the d / R _a ratios are too small. The critical limit for nanocrystalline alloys has not yet been reached.

The physical cause for the improvement in magnet values with increasing roughness and decreasing ribbon thickness could be due to dynamic domain refinement in the nanocrystalline metal ribbon. The latter presumably results from the closure domain structures present around surface defects (such as air pockets on the underside of the belt, coarser crystalline precipitates, etc.) and comparable magnetization inhomogeneities.

For good magnet values, the band should be as thin as possible (if possible below 20 μm average band thickness) with "defined" roughness (around or above R _a = 1 μm).

The demand for rougher tape leads to an improvement in the dynamic magnet values, if the magnetization process is significantly supported by wall displacements. This is true for materials with rectangular hysteresis loop and conditionally for round hysteresis loops.

One possible magnetization process for rectangular hysteresis loops is that the magnetization proceeds across the movement of 180 ° domain walls transversely to the tape longitudinal direction. Due to the associated strong spatial localization of the magnetization changes, excessive, so-called anomalous eddy current losses result. In this case, the greater the increase in losses, the fewer domains involved in the magnetization process.

Dabei bezeichnen

• ρ_el den spezifischen elektrischen Widerstand,
• n _o die Zahl der Domänen pro Flächeneinheit beim quasi - statischen Durchlaufen der Hystereseschleife,
• V _o einen Mindestwert, um den das äußere Feld erhöht werden muss, um eine neue Domäne zu bilden, bzw. eine gepinnte Wand in Bewegung zu setzen, V_o ist damit letztlich eng verknüpft mit der statischen Koerzitivfeldstärke,
• f die Ummagnetisierungsfrequenz und
• B̂ die Induktionsamplitude.

The investigation of the frequency dependence shows that the magnetization characteristic of nanocrystalline ferromagnetic materials with a rectangular hysteresis loop is very well within the already successful theory of amorphous alloys Bertotti from Bertolli, J. Magn. Mat., 1556, 1986, pages 54 to 57 , can be described. After this, the contribution of the anomalous eddy current losses results as $${\mathbf{P}}_{\mathrm{Fe}}^{\mathrm{nano}}\alpha \sqrt{\frac{\mathrm{Vo}}{{\mathrm{\rho }}_{\mathrm{el}}\cdot {\mathbf{<figure-callout\; id="0"\; label="n"\; filenames="imgf0005.png,imgf0007.png"\; state="\{\{state\}\}">n</figure-callout>}}_0}}\cdot {\left(\mathbf{f}\cdot \widehat{\mathbf{B}}\right)}^{3/2}$$

Designate

• ρ _el the specific electrical resistance,
• n _o the number of domains per unit area during the quasi-static traversal of the hysteresis loop,
• V _{o is} a minimum value by which the outer field must be increased in order to form a new domain, or to set a pinned wall in motion, V _o is thus ultimately closely linked to the static coercivity,
• f the Ummagnetisierungsfrequenz and
• B is the induction amplitude.

The power dissipation is generally given by H dB / dt, where the rate of magnetization reversal dB / dt is proportional to f · B. H is the external field needed to compensate for locally generated eddy current fields. From equation (1) follows for this: $$H-H_c^{\mathit{stat}}\left(B\right)\alpha \sqrt{\frac{\mathit{Vo}\cdot f\cdot \widehat{B}}{{\rho }_{\mathit{el}}\cdot {\mathrm{<figure-callout\; id="0"\; label="n"\; filenames="imgf0005.png,imgf0007.png"\; state="\{\{state\}\}">n</figure-callout>}}_0}}$$

Here H _c ^stat (B) denotes the course of the quasistatic hysteresis loop, which is mainly determined by coercive field mechanisms. The contribution of the so-called classical eddy currents is neglected, which is justified for not too high frequencies (f <1000 Hz). For example, an estimate for f = 50 Hz and B = 1 T provides only fractions of mA / cm for the eddy current field resulting from classical eddy currents.

The ultimate critical parameters in Equations (1) and (2) are the domain density n _o , as well as the nucleation field strength V _o . To clarify below is how both sizes are related to the surface roughness and the strip thickness.

verbunden, wobei $$N_{\mathit{eff}}\approx \frac{{\mathrm{\pi }}^3}{8}\cdot \frac{R_a^2}{\mathrm{\lambda }\cdot d}$$

ein mittlerer durch die Oberflächenrauigkeit bedingter Entmagnetisierungsfaktor ist. Es bezeichnen ferner d die Banddicke, K die Anisotropie in Bandlängsrichtung und λ eine effektive Wellenlänge, die ein Maß für die Ausdehnung und den Abstand der Oberflächendefekte darstellt.In band-level magnetization, magnetic surface poles form on geometrical deviations from an ideal flat surface, resulting in local stray fields. In order to reduce the associated stray field energy, magnetization inhomogeneities are formed at the surface defects, in extreme cases these are jagged termination domains, for example at the air pockets of the underside of the band. Thus, the consequence of a slower entry into the ferromagnetic saturation and in particular a reduction of the remanence ratio according to $$J_r/J_s\approx 1-\frac{1}{2}{N}_{\mathit{eff}}\sqrt{\frac{J_s^2}{2{\mu }_{0}K}}$$

connected, where $$N_{\mathit{eff}}\approx \frac{{\mathrm{<figure-callout\; id="3"\; label="\pi "\; filenames="imgf0002.png"\; state="\{\{state\}\}">\pi </figure-callout>}}^3}{\mathrm{8th}}\cdot \frac{R_a^2}{\mathrm{\lambda }\cdot d}$$

is a mean degaussing factor due to the surface roughness. Furthermore d denotes the band thickness, K the anisotropy in the band longitudinal direction and λ an effective wavelength, which represents a measure of the extent and the distance of the surface defects.

In this case, increased band roughness should also entail an increased nucleation probability for new domains. Investigations of the dynamic domain structure FeSi sheet metal and amorphous metals are also indications that new domains are formed preferentially to surface irregularities with increasing frequency.

To estimate the influence of R _a on the domain density n _o , for example, the following simplified model can be used:
Domains which initiate the remagnetization preferably form on the surface irregularities and thus have approximately their effective extent λ. For N _o domains thus results in a cross-sectional portion N · λ ₀ / b = ηο · λ d (b = bandwidth, n ₀ = N ₀ / (b · d)). On the other hand, this cross-sectional proportion reduces the remanence magnetization and is thus proportional to 1-J _r / J _s . With that follows: $${\mathrm{<figure-callout\; id="0"\; label="n"\; filenames="imgf0005.png,imgf0007.png"\; state="\{\{state\}\}">n</figure-callout>}}_0\approx \frac{1-J_r/J_s}{\lambda \cdot d}\alpha \frac{R_a^2\cdot J_s}{{\left(\lambda \cdot \mathrm{<figure-callout\; id="2"\; label="d"\; filenames="imgf0001.png,imgf0003.png"\; state="\{\{state\}\}">d</figure-callout>}\right)}^2\cdot \sqrt{K}}$$

An expression similar to the R _a, d, and K dependence is obtained by assuming the average stray field energy 1/2 N _e ff J _s ² and setting it equal to the domain wall energy necessary to give n _o domains per cross section to create.

In addition, the relevant coercitive field mechanisms still have to be considered for the influence on eddy current losses. Here, one can distinguish between two cases:

a) H _c is determined by pinning to surface defects

gegeben. Dann folgt aus Gl. (2) für das Wirbelstromfeld $$H-H_c^{\mathit{stat}}\left(B\right)\propto \sqrt{\frac{\lambda \cdot d}{R_a}\cdot K_u\cdot f\cdot \widehat{B}}$$

The coercitive field strength is essentially determined by pinning on the surface defects. The minimum field strength V _o for involving a new, previously pinned domain in the magnetization process is then analogous to H _c $$V_0\alpha H_c\alpha \frac{R_a}{\lambda \cdot d}\frac{\sqrt{A\cdot K}}{J_s}$$

given. Then follows from Eq. (2) for the eddy current field $$H-H_c^{\mathit{stat}}\left(B\right)\alpha \sqrt{\frac{\lambda \cdot d}{R_a}\cdot K_u\cdot f\cdot \widehat{B}}$$

It should be noted that H _c ^{stat in} accordance with equation (5a) is proportional to R _a / d, that is ultimately more dependent on R _a / d than the anomalous eddy current losses. This results in inferior magnet values again due to the increasing hysteresis losses for bands that are too rough.

b) H _c is independent of surface defects

womit das Wirbelstromfeld als $$H-H_c^{\mathit{stat}}\left(B\right)\propto \frac{\lambda \cdot d}{R_a}\sqrt{K^{3/2}\cdot f\cdot \widehat{B}}$$

folgt. Dieser Fall ist insbesondere für nanokristalline Werkstoffe wichtig, bei denen auf Grund der endlichen Korngröße die mittlere Kristallanisotropie (K₁) einen entscheidenden Beitrag zu H_c liefern kann, im Gegensatz etwa zu amorphen Systemen. Bei konsequenter Rechnung wäre hier entsprechend der Term K^3/2 durch K_u ^1/2 <K₁> zu ersetzen.In this case, H _{c is} determined by nucleation or by pinning to intrinsic Anisotropiefluktuationen. Then $${\mathrm{<figure-callout\; id="0"\; label="V"\; filenames="imgf0005.png,imgf0007.png"\; state="\{\{state\}\}">V</figure-callout>}}_0\alpha H_c\alpha K/J_s$$

with which the eddy current field as $$H-H_c^{\mathit{stat}}\left(B\right)\alpha \frac{\lambda \cdot d}{R_a}\sqrt{{\mathrm{<figure-callout\; id="3"\; label="K"\; filenames="imgf0002.png"\; state="\{\{state\}\}">K</figure-callout>}}^{3/2}\cdot f\cdot \widehat{B}}$$

follows. This case is particularly important for nanocrystalline materials in which the finite grain size, the average crystal anisotropy (K ₁ ) can make a decisive contribution to H _c , as opposed to amorphous systems. In the case of a consistent calculation, the term K ^3/2 would have to be replaced by K _u ^1/2 <K ₁ >.

In this limiting case, the coercive field strength and thus the hysteresis losses are independent of the roughness. Accordingly, there is a more efficient reduction in overall losses with increasing roughness than in the amorphous materials case discussed above.

Overall, the general statements of the above model calculation seem to agree with the following experiment. However, one must be aware of the simplistic assumptions that are made in the above formulas.

In summary, it can be stated that the 50 Hz characteristic of nanocrystalline (as well as amorphous) materials with a rectangular hysteresis loop is decisively characterized by anomalous eddy current losses. The above investigations indicate that the ratio of the strip thickness d to the surface roughness R _a forms a significant influencing parameter, which can go far beyond the influence of the alloy composition. The best magnet values were found for ratios d / R _a <20, ie for thin bands (around 20 μm and smaller) with a relatively rough surface (R _a around 1 μm or slightly larger). With the alloy composition Fe _73.5 Cu ₁ Nb ₃ Si _13.5 B ₉ , in the operating point relevant for ac-current-sensitive 30 mA fault current switch at Hc = 10 mA / cm, induction values approximately twice as high as with coarsely crystalline NiFe alloys (such as Ultraperm 10, 200 ) reached become. The possibility of achieving such peak values and creating decisive starting points for their reproducible production can thus be achieved.

Claims (14)

1. Soft magnetic metal strip, wherein the soft magnetic metal strip has a nanocrystalline or amorphous structure, wherein the metal strip has strip thickness-to-roughness ratios d/Ra of 5 ≤ d/Ra ≤ 25, preferably 10 ≤ d/Ra ≤ 20, characterised in that
   the metal strip has a fish scale pattern with a structure arranged across and at an angle to the strip longitudinal direction.
2. Metal strip according to claim 1 or claim 2, wherein the metal strip contains alloy constituents of silicon, boron, niobium and copper in more than 73 % iron by atomic weight.
3. Metal strip according to any of claims 1 to 3, wherein the metal strip contains an alloy with FE_75.5Cu₁NB₃Si_12.5B₈ or FE_73.5Cu₁NB₃Si_13.5B₉.
4. Metal strip according to any of the preceding claims, wherein the strip thickness lies between 5 µm < d < 20 µm and/or the roughness Ra lies between 0.6 µm < d < 2.5 µm, preferably between 1 µm < d < 2 µm.
5. Metal strip according to any of the preceding claims, wherein Br/Bm is > 80%.
6. Metal strip according to any of the preceding claims, wherein one surface of the metal strip has a surface topology of a surface structuring of a casting roller.
7. Magnet core comprising a wound soft magnetic strip according to any of claims 1 to 6.
8. Use of the soft magnetic strip according to any of claims 1 to 6 for earth leakage circuit breakers with a leakage current limit value Imax ≤ 30 mA.
9. Use of the soft magnetic strip according to any of claims 1 to 6 for a speed sensor acting together with a segmented permanent magnet disc.
10. Use of the soft magnetic strip according to any of claims 1 to 6 for AC-sensitive electromechanical components with a soft magnetic annular strip core.
11. AC-sensitive earth leakage circuit breaker having a magnet core according to claim 7.
12. AC-sensitive earth leakage circuit breaker according to claim 11, wherein the magnet core has a Br/Bm ratio > 80% at a frequency of less than 1000 Hz.
13. Distribution transformer having a magnet core according to claim 7.
14. Distribution transformer according to claim 13, wherein the magnet core has a Br/Bm ratio > 80% at a frequency of less than 1000 Hz.

Images