JPH0351133B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to an improvement in an electrode pattern to obtain a frequency response characteristic that is asymmetric with respect to a center frequency.

Conventionally, one method of obtaining asymmetric frequency response characteristics with a single interdigital transducer is
One known method is to change the distance between the centers of adjacent electrode fingers (hereinafter referred to as electrode pitch) and the crossing width of the electrodes along the surface acoustic wave propagation direction. This is a so-called variable pitch interdigital electrode, as described below. That is, when the frequency response characteristic is subjected to inverse Fourier transform, an impulse response as shown in FIG. 1, for example, is obtained. Since this impulse response has an asymmetric frequency response characteristic, it includes an imaginary part as a result of inverse Fourier transform, and the time intervals between peak points at which the imaginary part becomes zero are non-uniform. Then, by forming interdigital electrodes corresponding to the obtained impulse response, the desired frequency response characteristics can be achieved with these electrodes. The way to do this is to make the intersection width between adjacent electrode fingers (surface wave excitation/reception area) proportional to the size of each peak point (indicated by an arrow) in the impulse response, and to adjust the electrode pitch to the impulse response. It may be done in proportion to the time between the peak points in . However, since the time between peak points is non-uniform, the electrode pitch of the interdigital electrode is also non-uniform, and as a result, the interdigital electrode has a variable pitch type.

The conventional method described above can satisfy the desired characteristics, but since the electrodes are unevenly pitched, it is difficult to design the electrode pattern, and the electrodes are designed for high frequencies because they create thick and thin electrodes. It has the disadvantage of being easily short-circuited.

In order to solve the above-mentioned problems, attempts have been made to obtain asymmetric frequency response characteristics using equally spaced interdigital electrodes, and methods called the odd-even function method and the mirror method or reflection method, which will be described later, have been proposed. .

In the former odd-even function method, if H ₁ (ω) is a linear representation of the desired frequency response characteristic, then H ₁ (ω
This is a method that assumes H ₂ (ω) such that -ω ₀ )=H ₂ (ω ₀ -ω). The relationship between H ₁ (ω) and H ₂ (ω) is shown in Figure 2. Here, let the even component be H _R (ω) and the odd component H ₁ (ω), and define H _R (ω) and H ₁ (ω) as follows, then their functions are as shown in Figure 3. Become.

H _R (ω)=H ₁ (ω)+H ₂ (ω)/2 (1) H _I (ω)=H ₂ (ω)−H ₁ (ω)/2j (2) Also, H ₁ (ω) From equations (1) and (2), H ₁ (ω)=H _R (ω)−jH _I (ω) (3).

The impulse response is the Fourier transform of equation (3), h(t)=h _R (t)−jh _I (t) =∫H _R (ω)e ^j(2 〓 ^t+ 〓 ⁾ d+∫ −jH _I (ω)e ^j(2 〓 ^t+ 〓 ⁾ d (4).

When the impulse responses shown by h _R (t) and -jh _I (t) in equation (4) are determined at t=1/ ₂₀ , they become as shown by the solid line and broken line in FIG. 4, respectively. The two impulse response curves in the same figure are both uniform in that the time between their peak points is 1/2 ₀ (λ ₀ /2 when expressed in terms of wavelength), and the peak points of both curves are within the distance between the peak points on the other side. located in the middle of The interdigital electrodes corresponding to the impulse response shown by the solid line constitute the even component, and the impulse response represented by the dashed line constitute the odd component.

The electrode pattern shown in Fig. 6 consists of interdigital electrodes divided into two stages based on the two impulse responses shown in Fig. 4, and electrically connected in parallel.
This is disclosed in "A Design Method for Surface Acoustic Wave Filters" by Nakamura and Shimizu (issued September 28, 1972, materials from the 172nd Acoustic Engineering Research Meeting, Institute of Electrical Communication, Tohoku University). In FIG. 6, one interdigital electrode 1 is composed of two interdigital electrodes 2 and 3 arranged perpendicular to the propagation direction, with electrode 3 exciting and receiving the even component and electrode 2 exciting and receiving the odd component. (the reverse is also possible), and the other interdigital electrode 4 is formed so as to cover the propagation path of the two electrodes 2 and 3.

However, with the electrode 1 shown in Fig. 6 above, it is possible to realize an asymmetrical frequency response characteristic with equal pitch, but since two interdigital electrodes are arranged in a direction perpendicular to the propagation direction, the surface wave excitation and reception area is expanded, and the surface wave The disadvantage is that the wave board becomes wider. Further, the central portion where the surface wave excitation/reception intensity is high is divided into both sides, and the central portion of the electrode serves as a common electrode, which is not a preferable electrode pattern.

In order to eliminate the above-mentioned problem and realize an asymmetric frequency response characteristic with one equally pitched interdigital electrode, the two impulse responses in Fig. 4 are synthesized as shown in Fig. 5, and this synthesized impulse response is Based on this, an electrode pattern can be constructed as shown in FIGS. 7a and 7b. In the figure, one interdigital electrode 5 has main electrode fingers 6, 7, 8, and 9 with an electrode width of 1/8λo arranged at an electrode pitch of 1/4λo, and two adjacent main electrode fingers 6 and 9 have an electrode width of 1/8λo. 7, 8, and 9 are connected at common parts with different potentials, and the lengths of these two main electrode fingers are different, and each main electrode finger 6,
Auxiliary electrode fingers 10, 11, 12, ₁₃ with a width of 1/8λ 0 facing the free ends of the electrodes 7, 8, 9 and connected to a common part of different potentials are arranged with an electrode pitch of 1/4λ ₀ . Placed and formed. According to this interdigital electrode, an even component is excited and received in the area where adjacent main electrode fingers 7 and 8 of different potentials intersect (shaded area on the upper right), and the adjacent main electrode fingers 6 and 9 and auxiliary electrode fingers intersect. In the area where 11 and 12 intersect (the area with diagonal lines downward to the right), the odd component is excited and received at a distance of λ ₀ /4 from the even component. If such interdigital electrodes are used, the electrode width in the direction perpendicular to the surface wave propagation direction can be narrowed and the surface wave substrate can be made smaller, but the electrode fingers 6 and 8,
Since surface waves are also excited and excited in the area where points 7 and 9 intersect (cross diagonal lines), errors occur in the frequency response characteristics, and it is very troublesome to design with this error taken into consideration in advance. In addition, in order to reduce the influence of excitation and vibration between the electrode fingers 6 and 8 and between 7 and 9 to a negligible extent, the electrode fingers 7, 11, 8 and 1 located between them are
If the two fingers 7 and 11, 8 and 12 are brought closer together to reduce the cross-hatched area, there is a risk that the fingers 7 and 11, 8 and 12 will be short-circuited at the tips during pattern formation.

The latter reflection method or mirror method is
If the center frequency of a given frequency characteristic is ₀ ,
This method assumes a virtual image with a center frequency of ₃₀ that is line symmetric with respect to ₂₀ , and the impulse response obtained is the same as the odd-even function method described above, and the electrode pattern is also similar to that shown in Figures 6 and 7. It is determined in the same way as those for a and b, and has the same problems as described above.

The present inventor previously filed an application in Japanese Patent Application No. 57-104391 for a surface acoustic wave filter that eliminates the drawbacks of the prior art described above. The content is such that a common electrode is provided along the envelope of interdigital electrodes constituting even components, and interdigital electrodes having odd components are constructed on one or both sides of this common electrode.

The present invention is a further improvement of the above-mentioned earlier application,
In addition to the effects obtained in the previous application, it is possible to reduce the substrate size and suppress diffraction loss.

Embodiments of the present invention will be described in detail below with reference to the drawings.

In FIG. 8, input/output side interdigital electrodes 21 and 22 are formed at a predetermined distance apart on a surface wave substrate 20 made of LiNbO ₃ , PZT, a ZnO film on a glass substrate, or the like. One interdigital electrode 21 is composed of first and second interdigital electrodes 23 and 24. 1st
The electrode 23 is based on the impulse response (solid line) that defines the even component in FIG. , the intersection widths are weighted so that the envelopes are inclined (on the drawing) to the left and right of the vibration source 23C where the envelopes have the maximum intersection width. One 23 of the two common electrode parts 23a and 23b of the first electrode 23
b is formed so as to substantially follow the above-mentioned slanted envelope. The second electrode 24 is an interdigital electrode that defines the odd component in FIG.
The vibration source located near the vibration source 23C, which has the maximum intersecting width of the electrodes 23, is set in advance to be zero, and the surrounding electrode fingers are omitted, and the vibration source that is included in the remaining vibration sources is set to zero. By shifting and dispersing the zero vibration source in the length direction of the electrode finger, the first electrode 23
are formed on the propagation path of the electrode 23. With this configuration, each vibration source of the second electrode 24 can be concentrated at a position where the excitation intensity is maximum in the length direction of the electrode finger, for example, near the center.

Therefore, no other vibration source is disposed outside the vibration source 23C of the maximum intersecting width of the first electrode 23, and the substrate size can be reduced accordingly. Moreover, since the second electrode 24 is concentrated in the center where the excitation intensity is originally high, there is no influence of diffraction loss, etc., and calculation errors are reduced.

The second electrode 24 is a separate common electrode part 2 formed in close proximity to the common electrode part 23b of the first electrode 23 and the vibration source 23C of the maximum intersecting width of the first electrode 23.
The electrode fingers are made to protrude alternately from 4a. The common electrode part 24a of the second electrode 24 and the first
The common electrode portion 23a of the electrode 23 is connected to the common electrode portion 23a by a shield electrode 25. terminal electrode 26,
27, 28, 29 are formed at each corner of the substrate 20,
Each is connected to a predetermined common electrode section.

Next, a method for reducing the vibration source located near the vibration source 23C of the maximum width of the first electrode 23 in the second electrode 24 to zero will be briefly described.

As mentioned above, when the desired frequency characteristic is expressed by equation (3), the impulse response is h(t)=h _R (t)−jh _I (t) =∫H _R (ω)e ^j(2 〓 ^t+ 〓 ⁾ d+∫ −jH _I (ω)e ^j(2 〓 ^t+ 〓 ⁾ d (4), resulting in the characteristics shown in FIG. 1st
For convenience of explanation, Figure 0 does not necessarily match Figures 1, 4, and 5. According to a method such as the reflection method, data is sampled at a time interval of t=1/ ₄₀ in FIG. 10 (solid arrow), and even component interdigital electrodes 23 are formed based on data corresponding to even numbers. The parasitic component interdigital electrode 24 is configured based on data corresponding to odd numbers. The same electrode configuration is used for the odd-even function method. However, in this embodiment, by slightly changing the sampling time interval, the even component and the odd component are changed, and, for example, the source of the odd component near the maximum value of the even component is reduced. That is, when sampling is performed at a time interval of t' = 1/4 ( ₀ + △), the source of the odd component becomes smaller as shown by the broken line in Fig. 10. Furthermore, if the maximum value of the excitation intensity is set to 1 on a relative scale, an excitation source of, for example, 0.02 or less is forcibly set to zero. Of course, if the vibration source is forcibly set to zero, correction should be made using another vibration source. With this configuration, the conventional electrode pattern becomes the electrode configuration shown in FIG. 8, which is one embodiment described above.

FIG. 9 shows yet another embodiment, which differs from the above two embodiments in that the electrodes 23 and 24 are configured in the form of split electrodes in order to have the effect of removing TTE. According to this embodiment, the pair of electrode fingers of the split electrode can be configured to have the same length,
Calculation errors are reduced compared to the conventional mirror method (or reflection method). In the embodiment using this split electrode, when the number of zero vibration sources is small,
Alternatively, if it is desired to shift the electrodes by a large amount, the lengths of the paired split electrode fingers can also be changed to allow the electrodes to be shifted significantly, which is a very effective means in practice.

Although the electrodes in the above embodiments have very simple envelopes, the present invention is applicable to electrodes having any envelope. Moreover, the even component and odd component as used in this specification collectively refer to the even component and odd component in the odd-even function method, the symmetric component and the asymmetric component in the reflection method, and the like. moreover,
When the design is performed with the phase corrected, the maximum value of the crossing width may be on the odd function side, and in that case, the above-mentioned method may be applied on the even function side.

As explained above, according to the present invention, desired frequency characteristics can be reliably obtained without errors even with a substrate size comparable to or smaller than that of the mirror method, complicated calculations at the time of design can be reduced, and a single It can be configured with either an electrode or a split electrode, and furthermore, the influence of diffraction loss can be minimized.

[Brief explanation of drawings]

Fig. 1 is an impulse response characteristic diagram of a conventional variable pitch type electrode, Figs. 2 to 5 are diagrams used to explain the conventional example and the present invention, and Fig. 2 shows the characteristics of H ₁ (ω) and H ₂
(ω) frequency characteristic diagram, Figure 3 shows H _R (ω) and jH _I
(ω) frequency characteristic diagram, Figure 4 shows H _R (ω) and jH _I
(ω) impulse response characteristic diagram, Figure 5 shows H _R (ω)
Impulse response characteristic diagram that combines and jH _I (ω),
FIG. 6 is a diagram showing a conventional filter, FIG. 7a is a diagram showing another conventional filter, and FIG. 7b is a partially enlarged view.
8 and 9 are diagrams showing filters according to the present invention, respectively, and FIG. 10 is an impulse response characteristic diagram used for explaining the present invention.

Claims (1)

[Scope of Claims] 1. A surface acoustic wave filter having at least input and output side electrodes for obtaining a frequency response characteristic asymmetrical with respect to a center frequency, wherein at least one electrode is weighted in cross width. A first interdigital electrode that defines an even component (or an odd component) of the frequency response characteristic, and a frequency response that is mainly arranged in a non-intersecting region of the first interdigital electrode and that weights the intersection width. the second which specifies the odd component (or even component) of the characteristic;
and an interdigital electrode, and the second interdigital electrode is configured such that at least one vibration source near the vibration source, which is the maximum crossing width of the first interdigital electrode, becomes zero. 1. A surface acoustic wave filter, wherein the remaining oscillation sources are arranged such that the zero oscillation sources contained therein are dispersed toward the first interdigital electrode.