Saw filter comprising an additional pole

Abstract

In order to suppress an interference frequency in a ladder-type filter, an additional resonator (RZ1) that acts as a capacitance is connected in parallel to a series resonator (S1). The antiresonance of the additional resonator creates an additional pole in order for the interference frequency to be attenuated more effectively.

Claims

1. A SAW filter, comprising: a series branch connected between a filter input and a filter output, in which series resonators are arranged; n parallel branches, each including a different parallel resonator, connected to a fixed potential; an additional resonator, configured to act capacitively at a frequency, connected in parallel to a first series resonator of the series resonators, the first series resonator having a largest finger period of all the series resonators; wherein a pitch of the additional resonator is configured to have an anti-resonance frequency to form an additional pole at an interference frequency.

2. The SAW filter according to claim 1, wherein the pitch of the additional resonator is lower compared to a pitch of the first series resonator in order to achieve an improved attenuation at the interference frequency; wherein the interference frequency is higher than a passband of the SAW filter.

3. The SAW filter according to claim 2, wherein: the first series resonator is configured to have a lowest anti-resonance frequency of all the series resonators; a pole zero gap of the first series resonator compared to other ones of the series resonators is reduced so far that a parasitic shear mode of an acoustic wave is shifted out of the passband of the SAW filter; and the other ones of the series resonators have a pole zero gap and a finger period which are higher than in the first series resonator.

4. The SAW filter according to claim 3, wherein: the pole zero gap of the first series resonator is reduced by the additional resonator; and the additional resonator is dimensioned such that a parasitic shear mode of the acoustic wave is shifted out of the passband of the SAW filter.

5. The SAW filter according to claim 1, wherein the SAW filter is constructed on a lithium niobate substrate with a cutting angle between red 125 and red 130.

6. The SAW filter according to claim 5, wherein the SAW filter is configured with an SiO2 layer, arranged above an electrode metallization on the substrate.

7. The SAW filter according to claim 1, wherein the anti-resonance frequency of the additional resonator lies outside a passband of the filter.

8. The SAW filter according to claim 1, wherein the SAW filter is designed for operation in a band with a relative bandwidth greater than 3%.

9. The SAW filter according to claim 1, wherein any of the series resonators whose SH mode is in a range between a right passband edge and a foot of a flank are implemented with a reduced pole zero gap.

10. A method for generating an additional filter pole in a SAW filter, a) in which, in a first step, the SAW filter is designed from SAW resonators; b) wherein a series branch, connected between a filter input and a filter output, in which a plurality of series resonators of the SAW resonators are arranged, and n parallel branches connected to a fixed potential, in each of which one parallel resonator of the SAW resonators is arranged; c) a first series resonator of the plurality of series resonators, the first series resonator having a largest finger period of all the series resonators of the plurality of series resonators, is connected in parallel with an additional resonator acting as a capacitance at a frequency; d) wherein a pitch of the additional resonator is set at an anti-resonance frequency of the additional resonator that forms the additional filter pole at an interference frequency to be suppressed.

11. The method according to claim 10, wherein: the SAW filter is configured to have a passband for TX frequencies from 1,850 MHz to 1,910 MHz; and a frequency position of the additional resonator is set so that a pole at RX frequencies is obtained between 2,110 MHz and 2,155 MHz.

Description

BRIEF DESCRIPTION OF DRAWINGS

(1) FIG. 1 a first SAW filter according to the invention,

(2) FIG. 2 a circuit symbol and metallization of the resonators used according to the invention,

(3) FIG. 3 a detail of a filter according to the invention, in schematic cross-section,

(4) FIG. 4 the transmission curve determined by simulation for the filter shown in FIG. 1,

(5) FIG. 5 the impedance of a first resonator before and after variation according to the invention,

(6) FIG. 6 the impedances of the resonators used for a filter according to the invention, together with the resulting transmission curve,

(7) FIG. 7 the transmission curves determined by simulation for the filter shown in FIG. 6,

(8) FIG. 8 a filter known from the prior art,

(9) FIG. 9 the impedances of the resonators used for filters of FIG. 8, together with the resulting transmission curve,

(10) FIG. 10 the simulated passband of the filter according to FIGS. 8 and 9, with and without consideration of SH modes.

(11) FIG. 11 a plan view of an alternative usable series resonator, with removal weighting,

(12) FIG. 12 the relative space requirements of alternative solutions for reducing the pole zero gap.

DETAILED DESCRIPTION

(13) FIG. 1 shows a first exemplary embodiment according to the invention on the basis of a schematic block diagram. The filter consists of a series branch SZ, which is connected between two terminals TXIN and ANT. In series branch SZ, four series resonators S1, S3, S5, and S7 are arranged. Between each two series resonators, a parallel branch PZ branches off to ground GND from the series branch, in each of which a parallel resonator P2, P4, or P6 is arranged. Parallel to at least one of the series resonatorsin this case, the two series resonators S1 and S5an additional resonator RZ1, RZ5 is connected in each case. The additional resonator RZ1 connects a first circuit node N1 in the series branch in front of the series resonator S1 to a second circuit node N2 arranged in the series branch after the series resonator S1. The same applies to the additional resonator RZ2. All series and parallel resonators S, P are listed as SAW resonators.

(14) FIG. 2 shows, in the left part, the circuit symbol used for the resonators, while, in the right part, schematically, an exemplary metallization of a resonator usable for the invention is provided. Copper-based, multilayer systems are preferably used as metallizationsfor example, a layer sequence of chromium Cr, silver Ag, copper Cu and chromium Cr, or a layer sequence of titanium Ti, silver Ag, copper Cu and titanium Ti. However, other layer systems for the electrodes are possible, but preferably have at least one copper layer.

(15) The metallization is deposited on a piezoelectric substrate with high couplingin particular, on a lithium niobate crystal with a red-128 crystal cut.

(16) FIG. 3 shows a cross-section through a SAW filter, as it may be formed according to the invention. On the piezoelectric substrate SU, the component structures of the filterin particular, the electrode fingers and/or the fingers of the associated reflectorsare shown in the cross-sectional view transverse to the finger extension. Directly above the substrate SU coated with the component structures BS, a compensation layer KS is applied, by means of which the temperature coefficient of the frequency is reduced or even compensated for. For this purpose, usually, a SiO2 layer of sufficient thickness is used.

(17) In order to protect the moisture-sensitive compensation layer KS against environmental influences, a protective layer PL is applied as a final uppermost layer, e.g., a thin silicon nitride layer having a thickness of 10 nm to 200 nm.

(18) FIG. 4 shows the simulated transmission curve of the filter shown in FIG. 1. Compared to a known filter without additional resonators, the attenuation in the region of an interference frequency is improved. The improvement is circled in the figure. As an application example, a band 2 Tx filter is selected, which shows a clear pole, and thus the improved damping, in the Rx region of a band 4 duplexer.

(19) For a further explanation of the inventionin particular, for explaining the method according to the invention for shifting an interfering SH modethe design method of a filter according to the invention will be described in the following in extracts, and the measures required for this purpose will be explained.

(20) FIG. 8 shows a conventional SAW filter with the same basic structure as the SAW filter according to the invention shown in FIG. 1, in which only the additional resonators are missing. The frequencies of the series and parallel resonators S, P are chosen so that there is a desired passband. In that regard, the filter is optimized for the desired band.

(21) FIG. 9 shows the various resonators used for the known filter according to FIG. 8, with their impedances and the resulting transmission curve TF or transfer function.

(22) In order to realize the filter with the illustrated high bandwidth, the resonance frequencies fr(P) and fr(S) of the series resonators S and the parallel resonators P are offset from each other and, preferably, all selected so as to differ. In the region fr(P), the resonance frequencies of the parallel resonators occur, recognizable at the minimum of their impedance curves. At a constant pole zero gap, the anti-resonances of the parallel resonators can be found at a corresponding distance thereto in the range fa(P).

(23) In a same or similar frequency range fr(S), the resonant frequencies of the series resonators are found, which are preferably arranged symmetrically to the center of the passband.

(24) The anti-resonance frequencies of the series resonators are found in the frequency range fa(S) above the right passband edge of the transfer function TF.

(25) All curves shown in FIG. 9 are the result of a simulation that was determined without taking account of interfering shear wave modes (SH modes) that occur.

(26) FIG. 10: If, however, the occurrence of SH modes is allowed in the simulation calculation, the transmission behavior shown in FIG. 10 is obtained for the filter structure, shown in FIGS. 8 and 9, in the passband. Two passbands for the TX filter (left curves) and the RX filter (right curves) of a duplexer are shown. For better recognition of the effect of an interfering SH mode SHM, two curves are superimposed in the figure, corresponding to the simulation without consideration of the SH modes according to curve 1 and, once, taking into account SH modes occurring that correspond to curve 2. In the TX filter, an interfering resonance appears at the point marked with an arrow in the right passband edge, corresponding to the interfering SH mode SHM. However, such an SH mode in the region of the passband edge leads to a greater thermal load on the filter, which leads to an additional power load on the filter and, correspondingly, to a higher thermal load, which accelerates the aging of the filter and may damage the component structures BS.

(27) The calculation also shows that the interfering SH mode is generated by the series resonator with the lowest resonance frequency. The distance between SH mode and resonance frequency is only in the range of the passband edge when the resonance frequency of the use mode (Rayleigh wave) is arranged at a correspondingly low frequency, since the frequency distance of the SH mode to the Rayleigh mode in the resonators of the filter is almost constant. The impedance of this first resonator is designated by RSx in FIG. 9.

(28) The interfering SH-mode can be minimized by suitably matching the layer thicknesses of electrodes, compensation layer, and protective layer and by a suitable choice of the metallization ratio of the electrodes in their height. However, since the exact tuning has a tolerance due to production-related deviations, it cannot be effectively suppressed in a series production, or the proportion of filters with no or poorly suppressed SH mode is too high.

(29) A simple shifting of the resonant frequency of the series resonator with the interfering SH mode towards higher frequencies, and thus also a shifting of the interfering peak of the SH mode from the range of the passband edge, is not possible without adversely affecting the passband characteristic or the right passband edge. According to the invention, therefore, the pole zero gap of this resonator with the interfering SH mode is reduced by a frequency value f by connecting one or more additional resonators in parallel to this series resonator; see, for example, the resonators RZ1 and RZ2 in FIG. 1.

(30) Due to the capacitance of the additional resonatorcapacitive only at the center frequency of the filterthe anti-resonance frequency of the series resonator is shifted by the value f toward lower frequencies, and thus reduces the pole zero gap. However, in order to compensate for this effect and bring the anti-resonant frequency back to the right position important for the formation of the passband edge, the finger period is shortened in parallel, to move the resonance frequency by the value f to higher frequencies. This may also be necessary in all cases in which an additional pole is primarily intended to be generated by the additional resonator, so as to better suppress an interference frequency.

(31) FIG. 5 shows a simulation calculation of the impedance of the first series resonator, without an additional resonator (left curve), and with a parallel-connected resonator and extended finger period (right-hand curve). By virtue of the two measures, the position of the anti-resonance frequency remains practically unchanged according to the minimum of the two curves, while the resonance frequency shifts by the value f towards higher frequencies.

(32) FIG. 6 shows a representation corresponding to FIG. 9, but in which, now, the impedance of the series resonator, with the additional resonator connected in parallel in position and pole zero gap, is changed. The resonance frequency is now shifted upwards by a value f, while the position of the anti-resonance remains unchanged. It turns out that the passband TF remains almost unchanged, despite the additional resonator.

(33) FIG. 7 shows the passband of the corresponding filter in a simulation, with and without consideration of SH modes. The two curves are again shown above one another, corresponding to FIG. 10 already described. Compared to FIG. 10, FIG. 7 shows the effect produced by the invention that, in the region of the right-hand passband edge (see arrow), the SH mode (see SHM in FIG. 10) completely disappears in both calculations or in both curves, or no longer occurs. The remaining passband characteristics remain virtually unchanged, so that, with the measure according to the invention, no disadvantages in the range of the passband are to be accepted.

(34) In further embodiments, not shown, further series resonators can be connected in parallel with additional resonators beyond that shown in FIG. 1. In order to not reduce the bandwidth of the filter, only the frequency resonant series resonators, whose SH modes are within the passband or its edge, are connected in parallel with capacitances. Furthermore, the number of parallel branches and the number of series resonators can be further increased.

(35) In a further embodiment, not shown, individual, several, or all of the resonators are implemented as cascaded. Cascading here means that the individual SAW resonator as shown in FIG. 2 is replaced by a series connection of at least two partial resonators. Cascading lowers the voltage applied to the resonator, so that the area of the resonator is to be increased accordingly to compensate for this. At the same time, the power stability of the component structures is thereby increased, so that, in particular, those resonators are cascaded that are exposed to the strongest signal amplitudes. These areparticularly in the case of a TX filterthe series resonators arranged close to the input (TXin) and the parallel resonators with the highest resonance frequency. Also not shown are inductances, with which, in particular, the parallel branches can be connected in series with fixed potential.

(36) FIG. 11 shows a plan view of a resonator, with removal weighting. This is another alternative to using parallel capacitances for reducing the pole zero gap of a resonator.

(37) FIG. 12 shows, in the scale-related partial figures a, b, and c and thus, in a manner suitable for size comparison, the respective space requirement

(38) (a) of an unchanged series resonator SX,

(39) (b) of a serial resonator SX with scaled finger period, reduced aperture, and with additional resonator, and

(40) (c) of an equivalent series resonator SXW with removal weighting.

(41) From the figure, it is clear that the space requirement increases from a) to c), but the solution b) is the one that requires the least additional space.

(42) The invention has been described only with reference to a few embodiments and is therefore not limited to these. A filter according to the invention can therefore deviate from the illustrated structuresthat is to say, the block diagram shown. Furthermore, the filter can also differ in its layer structure of the known layer structure shown in FIG. 3.

(43) The filter according to the invention may be part of a duplexer, wherein the invention is used as a receiving filter, but also, in particular, as a TX filter of the duplexer.