JP2003510939A - Narrowband tuned resonator filter topology with high selectivity, low insertion loss and improved out-of-band rejection in the extended frequency range - Google Patents

Abstract

Abstract: Narrowband bandpass filters with high angular load Q and optimal coupling (because of low insertion loss) utilizing parallel tuned resonator topologies in the 1 to 2 GHz or higher frequency band. Discloses a tuned resonator circuit topology that enables the realization of Because it consists of a parallel tuning circuit mirrored to the signal lines of a conventional parallel tuning circuit, its topology acts to substantially cancel all induced currents between the inductive elements of the resonator. . This reduction in the induced current reduces the electromagnetic coupling between the resonators, canceling out the increase in total coupling between the resonators at higher frequencies, and optimally coupling between the resonators even at higher operating frequencies. Act to hold. In addition, the mirror image topology allows for increased parallelism between the inductor elements of the resonator to reduce inductance values and increase capacitance values. The increase in the capacitance value of the resonator effectively offsets the decrease in Q during angular loading as the frequency increases. The topology is valid for any number of parallel resonators. As manufacturing process accuracy decreases (eg, from printed circuit boards to integrated circuit processes), the range of operating frequencies increases with increasing accuracy.

Description

Detailed Description of the Invention

[0001] Technical Field The present invention is a high frequency (HF), very high frequency (VHF), related to the topology of the tuned resonator filter of a narrow band-pass for the effective use in ultra high frequency (UHF), and microwave, More specifically, in the frequency range of interest, it retains the high loaded Q for increased selectivity and optimal coupling to minimize insertion loss for improved out-of-band rejection. Related to possible topologies. The present invention allows for relatively simple, inexpensive and highly reliable manufacturing.

BACKGROUND OF THE INVENTION The processing of wideband multi-carrier signals creates a particularly rigorous and stringent situation for signal processing circuits such as filters. For example, baseband television signals have a bandwidth of approximately 5-6 MHz and are typically mixed with RF (radio frequency) carrier signals (
Placed on the RF channel in the range of 50-1000 MHz or higher, and achieves frequency division multiplexing (FDM). Other applications, such as microwave communications, may require operating ranges of 1-2 GHz or higher. In applications requiring the processing of wideband signals, including multiple channels simultaneously, such as transmitting and receiving television broadcasts (via air, fiber optic cable, or coaxial cable), only a portion of the total bandwidth is allowed to pass (ie , Those frequencies fit in a narrow passband, usually one channel of interest), while needing a filter to stop the rest of the frequencies in the full bandwidth (ie, those frequencies fall in the stopband) The situation has arisen. This is typically implemented using a narrow band filter. Depending on the system design for a particular implementation, these filters may be at the same RF frequency, such as the RF channel of interest, or at some other frequency (intermediate frequency, i.e. IF) is required, which is usually a system of widely varying frequencies.

Noise and image signals and various unwanted spurious signals may be inserted or collected at various points in the process, thus depending on the selectivity of the application, the out-of-band signals may be of very low levels. Bandpass filters are often required to attenuate (ie, block) down to. For example, even a signal attenuated by 60 dB can still be seen in the received video transmission. Therefore, in many cases it is very important to significantly attenuate all signals other than the baseband signal modulated in the desired carrier. This often results in high selectivity (total loss) in the pass band with little or no energy loss (ie low insertion loss), but with the required degree of attenuation for all other frequencies in the stop band. Bandpass filter) that ideally passes only the portion of the bandwidth that includes the baseband signal of interest. Furthermore, since the fraction of the full bandwidth baseband signal in wideband applications is very small (about 1-2%), such filters produce the required frequency response with high reliability, And the response must be retained for a long time (ie the response must not change). Furthermore, they must be relatively insensitive to coupling between the elements themselves and RF noise from external power sources. Finally, it is always desirable for the filter to be inexpensive, easy and highly manufacturable.

There are several existing techniques for implementing bandpass filters. As mentioned above, the Q value of a filter represents its selectivity, which is defined by how fast the response of the filter changes from the pass band to the stop band. Filter
The roll-off from the passband frequencies to the stopband frequencies is steeper for higher Q values. A more useful and practical criterion is the "in-circuit" or Q at square load (ie, Q _L ), as the input and output loads of the filter affect the Q value. The Q _L of the filter is approximately equal to the reciprocal of the partial bandwidth of the frequency response, which is usually set between points in the response curve that are 3 dB below the response peak (ie, the half-power point of the response). . Therefore, the Q _L of the filter that passes with the partial bandwidth (that is, the ratio of the bandwidth) of 1% is approximately 100. Narrowband filters for wideband signal processing applications often require high Q _L values, while exhibiting low insertion loss (ie, the signal amplitude in the passband is greatly attenuated). Should not),
Furthermore, the attenuation of the signal in the stop band must meet the requirements of the application.

One known method for bandpass filter implementation requires the use of lumped LC elements to create a classical filter based on the technique of lowpass to bandpass conversion. To do. To produce the desired bandpass filter response,
Some modifications of the topology can be integrated. For the purpose of processing wideband signals in the VHF and UHF frequency ranges, there are a number of such filter deficiencies, the most significant of which is the lumped element (especially the coil inductor) which is parasitic at frequencies above 100 MHz. It is very susceptible to the effects. Moreover, some aspects of the circuit elements must be cascaded together to achieve the required transfer function complexity for high Q _L values. Therefore, such filters occupy valuable space and are relatively expensive to manufacture.

Another known method for implementing a filter utilizes a helical resonator. Filters utilizing spiral resonators are magnetically and / or capacitively coupled, which are capable of producing a response with low insertion loss and high Q _L required for many wideband signal processing applications. is there. However, they are unsuitable for frequencies below 150 MHz. This is because, below that frequency, a very large inductor value may be required for the resonator. It is impractical or impossible to construct such an inductor. Furthermore, even at higher frequencies,
They have a very large mechanical structure (they need shields to reduce their effects from RF noise and operate properly), which makes them costly to manufacture (even for large volumes). It will be relatively high. Also, they are very susceptible to environmental shocks and changes, and usually require numerical adjustments in the manufacturing process to make sure they resonate correctly at the appropriate frequency.

Another known method for constructing a bandpass filter is a magnetic, embodied as either a printed strip transmission line sandwiched between two ground plane shields or a cylindrical coaxial cable. A dielectric resonator that is mechanically coupled and / or capacitively coupled is used. These resonators are short-circuited transmission lines and their length with respect to the wavelength of the transmitted input signal (the length of the transmission line is usually the wavelength λ of the resonance frequency).
Λ / 4) to exploit their ability to resonate at a particular frequency. Such resonators are capable of producing high Q _L values and have partial bandwidth characteristics (ie, 1-2) required for many wideband signal processing applications.
%) Is achievable. However, since the trace length increases as the required resonant frequency decreases, such resonators are UHF (ie, about 40
It is not suitable for anything other than 0 MHz to several GHz). These resonators are very costly for HF and VHF, as the length of the transmission path increases to very costly sizes. Even at 1-2 GHz, there are about 2
Each trace length of ~ 1 inch is required, which is still very large and occupies a considerable area. Moreover, their length required to achieve a quarter of a wavelength is much larger for higher precision manufacturing techniques (eg integrated circuits) and well suited for them. I can't let you do it. Finally, such long quarter-wave resonators are very sensitive to transmit and receive noise.

Another known circuit topology for producing a bandpass filter response is that of a magnetically coupled double-tuned resonant circuit. The bandpass filters of such embodiments are at least high manufacturing cost compared to various other prior art techniques described herein, which can each be manufactured in a few cents. Embodiments of such filters are not able to achieve the large Q _L values required to produce a response with a small sub-bandwidth, as well as the low insertion loss required for many applications such as wideband signal processing. It has previously been considered impossible (typically they could not achieve sub-bandwidths above about 15%). In the following description, the reasons for those drawbacks in such applications will be apparent to those skilled in the art.

FIG. 1 a shows the general topology of the series double-tuned circuit 10, and FIG. 2 b shows the general topology of the parallel double-tuned circuit 100. The series double-tuned circuit 10 has an input resonator circuit 12 magnetically coupled to an output resonator circuit 14. Similarly, the parallel double-tuned circuit 100 has an input resonator circuit 120 magnetically coupled to an output resonator circuit 140. The input resonator 12,120 is connected to the power supply V _S 18,1.
80 and the associated power source impedance R _S 16, 160 respectively connected to the input power source. The output resonators 14,140 are connected to the output load impedances produced by the resistors R _L 15,150, respectively.

The input and output resonators 12 and 14 of the series tuning circuit 10 are series lumped capacitors.
Formed as a connection between C _S1 11 and C _S2 13 and inductors L ₁ 17 and L ₂ 19, respectively. The two series tuned resonators 12, 14 as well as the two parallel tuned resonators 120, 140 are magnetically coupled as a function of the physically close distance between their inductors, so that the mutual inductance M21 is equal to their mutual inductance M21. Occurs in between. M = k (L ₁ · L ₂ ) ^1/2 , where k is a coupling coefficient having a value of the physical distance from each other and a function of the positional relationship of the inductive element. Therefore, the coupling coefficient k represents the percentage of all possible mutual coupling between the two resonators. Two inductors 17, 19
Alternatively, the closer 170, 190 are, the greater the value of k, and thus the greater the mutual inductance between the resonators. Similarly, if the inductances are more separated, a lower mutual inductance is reflected, reflecting a lower value of k.

The parallel double-tuned circuit 100 is theoretically dual with the series double-tuned circuit 10 and thus operates in exactly the same way. Resonators 120, 140 of parallel tuned circuit 100 are formed as a parallel connection between lumped capacitors C _P1 110 and C _P2 130 and inductors L ₁ 170 and L ₂ 190, respectively. In addition, the parallel tuning resonators 120 and 140 are
They are magnetically connected as a function of the physically close distance between their inductors, thus creating a mutual inductance M210 between them. The mutual inductance of the parallel tuned circuit is given by the same equation M = k (L ₁ · L ₂ ) ^1/2 , where the value of k is determined by the same consideration of the positional relationship as described above.

FIG. 2 shows three typical responses of a double-tuned resonator circuit (either series or parallel) for various values of coupling coefficient k. Response 22 was obtained when the two resonators of the circuit were critically coupled at the resonant frequency where the circuit exhibited optimal coupling with average selectivity and minimal insertion loss. The response 24 is the double tuned circuit 1 when each input and output resonator is under-coupled poorly.
Responses of 0 and 100 are shown. This occurs when the value of k approaches zero and can be caused by the circuit's resonators being more separated. If poorly coupled, the value of Q _L of the circuit will increase (decreasing the partial bandwidth) but also the insertion loss, which is undesirable. Response 26 occurs when the two inductors of the input and output resonators are close to each other and overcoupled (ie, the value of k approaches 1). Response 2
6 is characterized by two maxima on either side of the resonance frequency, the circuit exhibits the lowest Q _L value thereof (hence, they are the largest fractional bandwidth). From these results, it can be considered that there is a trade-off between the maximum achievable Q _L value and the insertion loss in the implementation of the double-tuned filter. For a given frequency, this trade-off comes as a function of the mutual inductance M between the resonators of such filter embodiments. Optimal coupling occurs clearly at or near the critical range, as it gives the best compromise between stopband performance and insertion loss.

It should be noted that with increasing frequency, the overall electromagnetic coupling between the resonators increases. This is because the overall electromagnetic coupling between the resonators is a direct function of frequency as well as a function of mutual inductance M (which is a function of distance and positional characteristics between the resonators). Function of (ie w · M)
For that reason. Therefore, as the frequency increases for a given value of M, the electromagnetic coupling between the resonators increases, eventually over-coupling the circuit. At some point
It is possible to compensate for the increase in this coupling by simply increasing the space between the inductors, and thus reducing k to a lower M. However, increasing space at frequencies of 1 GHz and above is not feasible.

The Q _L for series tuned circuits is generally determined as the reactance X of the tuned network at the resonant frequency (w _o L) divided by the load or source impedance connected to them. Therefore, the Q _L for the output resonator 14 is (w _o · L ₂ ) / R _L.
For a given resonance frequency w _o , Q _L can be increased by increasing the value of L ₂ (of course, to increase the overall Q _L for the series double-tuned resonator, likewise for L ₁ It can also be done for the input resonator 12 by increasing the value). The problem with this approach is that there is a practical limit to the size of the inductors L ₁ and L ₂ that can be manufactured and implemented at a reasonable cost. Moreover, as L ₁ and L ₂ increase, the parasitic shunt capacitance on the lumped value inductor (usually the coil) reduces the frequency response of the filter at frequencies above 200 MHz. Finally, the resonance frequency is expressed by the formula w _{o =} 1 / (L ₂ · C _S2 ) ^1/2 (output resonator 1
4), the value of C _S2 must decrease inversely to retain the value of w _o . There are also practical limits on how small C _S2 can be accurately formed.

FIG. 3 shows C _S1 11, C _S2 13, L ₁ 17 and C _S1 11, designed to increase the value of Q _L.
1 shows the values for L ₂ 19 while showing the series double tuned circuit 10 of FIG. 1 which retains the optimum coupling of the circuit at the resonant frequency of 400 MHz. 4a and 4b correspond to FIG.
3 shows a simulated response for a circuit 30 having element values as shown in FIG. The lower values in Figures 4a and 4b are the frequency (MHz) and attenuation (dB) values for points 1-4 shown in the response curve. The scaled response given in FIG. 4a shows that its filter performance is disqualified at high frequencies for applications processing television signals. At the smaller scale given in Figure 4b, the 3 dB sub-bandwidth is about 16% (thus an approximate value for Q _L is 6.25).
Is shown. As mentioned above, this is unsuitable for many wideband signal processing applications.

The Q _L for a parallel tuned circuit is generally determined as the admittance of the network at the resonant frequency multiplied by the load or power source impedance connected to it. Therefore, the Q _L for the parallel tuned output resonator 140 is w _o · C _P2 · R _L. Therefore, it is believed possible to increase the values of C _P2 and R _{L in} order to increase the Q _L for the parallel tuned output resonator. Since the signal tends to be shunted to ground via the parasitic shunt element, it is impossible for R _L to increase above 100Ω. To increase C _P2 , L ₂ needs to be very small. It is very difficult to manufacture lumped inductors of about 5nH within the permissible range using known techniques, and such inductors are susceptible to positional changes, especially longitudinal changes. Furthermore, obtaining and maintaining a proper coupling between such small coils with a given reproducibility is difficult. In order to maintain an optimal bond (usually in a critical or near critical bond), the small coil requires a small gap between them, and its coupling coefficient depends on the size of this small gap. Very susceptible to changes in. If a sub-bandwidth of about 1% is required, such element and dimension variations cannot be tolerated.

[0017]   Figure 5 shows the Q for a circuit with optimal coupling at a resonant frequency of 400MHz. _L The ratio of L to C designed to increase k, k, C_P1110, C_P2130, L₁ 170 and L₂2 shows the parallel double-tuned circuit 100 of FIG. 1 with a value for 190. Figure
6a and 6b show a circuit for a circuit 50 having the values of the elements shown in FIG.
Shows the simulated response. The lower values in Figures 6a and 6b are shown in the response curve.
The frequency (MHz) and attenuation (dB) values for points 1 to 4 are shown. Given in Figure 6a
The response shown on the scale is that it has a higher frequency compared to the series tuned circuit 30 of FIG.
Even though it operates more symmetrically in wavenumber, the filter performance in the stopband is
Indicates ineligibility. Limit the value of the prior art coil used in this example.
Even with the change to the field, the bandwidth of this filter is still in many applications.
On the other hand, it is not narrow enough. For the smaller scale given in Figure 6b, the 3 dB part
The minute bandwidth is about 15.5% (thus Q_LThe approximate value of is 6.45). Previous
As mentioned, it requires a sub-bandwidth of 1-2% (ie Q_LValue of 50-100
This is unsuitable for many wideband signal processing applications.

Accordingly, those skilled in the art will recognize the need for bandpass filter circuits with the properties required for many wideband signal processing applications in the bandwidth range of about 50-2000 MHz or higher. Ah These properties are to provide high selectivity, i.e. a high Q _L value, thus giving a small sub-bandwidth, high attenuation in the stop band, low insertion loss in the pass band, and more It can be manufactured at lower cost and with higher reliability than circuits.

SUMMARY OF THE INVENTION It is an object of a first embodiment of the invention to provide a bandpass filter that utilizes the topology of magnetically coupled parallel double-tuned resonators, which topology is achieved by the prior art. it is possible to realize a Q _L higher than possible Q _L.

Another object of the first preferred embodiment is to achieve higher values of Q _L for topologies that can be manufactured with low cost and high reliability.

Yet another object of the first preferred embodiment is to be relatively insensitive to RF noise in the environment.

The purpose of the second embodiment of the present invention is to add only one additional element to the first embodiment to reduce the low between the pass band and the stop band at lower and higher frequencies. Achieving higher Q _L values with insertion loss and steeper roll-off.

The purpose of the third embodiment of the invention is to add only one additional element to the known magnetically coupled series double-tuned resonator topology, (lower frequency and higher Achieving higher values of Q _L , with lower insertion loss (in frequency) between the pass band and stop band and a steeper roll-off.

The purpose of the fourth embodiment of the present invention is to utilize the circuit topology with novelty and inventive step which is low cost and easy to manufacture, and makes use of the UHF frequency band, ie about 500 MHz.
And to achieve the desired values of Q _L , absolute bandwidth and insertion loss, even in the higher range above 2 GHz.

The above and other objects will be apparent to those skilled in the art from the detailed description of the invention.

A first preferred embodiment of the bandpass filter of the present invention uses an electrically short (about 1% of the wavelength of the resonant frequency) transmission line as a very small inductive element to which the resonator is magnetically coupled. Utilizes a parallel double-tuned resonator topology that achieves higher Q _L values. The transmission line is fabricated as a metal trace with precisely controlled geometry that achieves the desired inductance value. The dielectric constant of the printed circuit board material is 4.65 at a thickness of 1.5 mm. The trace is formed of 0.018 mm thick copper. The microstrip inductor is physically arranged to obtain a coupling coefficient (k) of about 0.01-0.02, depending on the value required to maintain optimum coupling for a given frequency. One end of the transmission line trace is connected to the series capacitor and the other end is at ground. Inductor value is about ± 2% accuracy
It can be accurately generated up to 0.5 nH.

In a second preferred embodiment of the bandpass filter of the present invention, the parallel double-tuned resonator of the first preferred embodiment is modified by adding a coupling capacitor to each of the magnetically coupled resonators. At, the capacitor is coupled in series and has a very small value compared to the shunt capacitance in parallel with the magnetically coupled microstrip transmission line inductor.

In a third preferred embodiment of the invention, the topology of the prior art series double tuned resonator is modified by adding a shunt capacitance in each resonator, the shunt capacitance being two resonators. Is connected in parallel with the series element of and has a larger value than the value of the capacitance in series with the inductance. The inductance is preferably implemented using an air core coil or other known lumped inductance element.

The second and third embodiments are used as electrical tuners by simply substituting the veractor or other known controllable capacitance for either the series or shunt capacitors of the resonator. It is possible.

A fourth embodiment of the present invention discloses a topology for compensating for the decrease in Q and the increase in electromagnetic coupling which causes problems for the first three embodiments when the tuning frequency exceeds about 1 GHz. To do. The topology includes a mirror image structure of each resonator of the previously mentioned tuned parallel resonator topology that is symmetrical about their respective signal line. The mirror image structure of each resonator serves generally to cancel out the mutual inductance between the two resonators, and thus the offset which can otherwise be significantly increased in the electromagnetic coupling with increasing frequency. Further, the parallel characteristics of the mirror-image-related inductors reduce the effective inductance value for each resonator by 50% or more, and the value of C _P for each resonator is equal to Q of the square load of the circuit with increasing frequency. It can increase to offset the decrease.

The inductive element and its mirror image structure for each resonator may be embodied as a single strip of metal or further reduce the effective inductance for each resonator without a proportional increase in electromagnetic coupling. For this reason, it is preferably embodied as several strips in parallel. By embodying the inductive elements as parallel strips, adding strips gives an additional degree of freedom to adjust the value of the effective inductance for each inductance value, thus allowing the filter to be tuned during the test. Becomes Of course, the inductance can be reduced by adding the width of the strip, but the electromagnetic coupling is inversely proportional to the decrease in the value of L in order to tune the filter circuit by adjusting without increasing the electromagnetic coupling. Increase. Furthermore, by creating the inductance for each resonator and their enantiomer as a parallel structure, it is possible to create a short circuit by adding metal between the strips to tune the filter during testing. Become. Of course, it is also possible to use laser trimming to achieve a similar purpose.

Any of the preferred embodiments may be arranged in various forms to offset any common mode noise that may be induced in the inductor from the environment by arranging the inductors so that the network current flows in opposite directions. It is possible to The preferred embodiment may be arranged in balanced-balanced and balanced-unbalanced configurations. Certain preferred embodiments may have the resonators physically located without having a particular positional relationship to each other. Other orientations (eg 45 degrees or equivalent angles) provide more topological flexibility and more freedom to adjust the coupling coefficient k, but the horizontal or vertical relative orientation of the resonators. The position (0 degree or 180 degree direction) is the main target. The values of the resonator elements of any of the preferred embodiments can be negotiated symmetrically or asymmetrically to adjust the impedance transformation or the frequency response of the filter. Finally, the multiple resonators in any of the preferred embodiments can be cascaded together to increase the complexity of the transfer function and increase Q _L as well as the passband to stopband slope or rolloff. It is possible.

Preferred Embodiments of the Invention Preferred embodiments of the invention are described in detail below. As mentioned above, the double-tuned resonators of FIGS. 3 and 5 do not reach the value of Q _L required for many broadband applications, even if the ratio of LC is increased so that Q _L is increased. In the parallel double-tuned resonator topology of FIGS. 1b and 5, the lower limit of the value of L is about 5 nH.

In the first embodiment of the invention shown in FIG. 7, the metal traces of copper formed on the printed circuit board include inductors L ₁ 72 and L ₂ 74 of the parallel double-tuned resonator 70. Used as. One end of each metal trace is connected to shunt capacitors C _P1 76 and C _P2 78, respectively, and the other end is grounded. With this method, the effective inductance value can be reduced to 0.5 nH with an accuracy of ± 2%.
Therefore, the inductance value is accurately lower than 5 nH, and C _P1 76 and C _P2 7
Since the value of 8 can be increased, the Q _L of the parallel double-tuned resonator can exceed the value obtained in the prior art.

Another advantage of the new method of using microstrip transmission lines as lumped inductive elements is that i ₁ and i ₂ (75) flowing at or near the resonator frequency is due to the extremely low impedance of the tiny inductor. It becomes extremely large. The increased current increases the energy available for transfer between the resonators. Therefore, the total electromagnetic coupling of the circuit will be greater than a given M, and the filter will be optimally coupled even if it is under coupled. Therefore, Q ₁ has a high tolerance for under-coupling, but the insertion loss is reduced because the current value is high. In addition, the physically small size reduces the inductance value, and the physical profile for PCBs is extremely low, so that the effects of PF noise (and interaction and electromagnetic waves) can reduce the effects of conventional lumped inductive elements. It is extremely small compared to. These are easy to manufacture, low cost, accurate and repeatable. Finally, this topology and its implementation are free to scale according to the precision of the manufacturing process used. The accuracy of the printed circuit board in the manufacturing process may limit the width of the inductor strip to a minimum of about 5 mm, but manufacturing these topologies on silicon with acceptable accuracy allows the formation of smaller inductors commensurate with it, resulting in smaller effective inductors. The value of becomes smaller accordingly.

FIG. 8a is a plan view of a portion of the PCB in which the inductive elements L ₁ 72 and L ₂ 74 (FIG. 7) are formed. In the preferred embodiment, each inductive element is formed on the surface 81 of the PCB 80 as copper microstrip races 82 and 84, respectively.
The microstrip is manufactured by the well-known metal welding method and etching technique. The size of the spacing 89 between the microstrips (ie height 86, width 87) determines the extent of the effective inductance of the inductive element and the mutual inductance M73 as a function of the coupling coefficient k. Trace thickness is preferably 0.018 mm
Is. The thickness or height 85 of the PCB is preferably 1.5 mm and the material used has an absolute dielectric constant of 4.65. One end of the microstrip is grounded to the ground plane 88 of the PCB 80 by the through hole 802. Through hole 802
Have their own inductance (depending on the diameter of the through hole, about 0.1 nH), which can each be varied depending on the application. If desired, multiple ground holes can be provided to reduce the total inductance of the holes. The ground plane 88 is typically formed on the back side of the PCB, but can also be formed on the front side of the PCB 80.

In the preferred embodiment, the microstrip can be divided into parallel microstrips by etching the metal portion 83 inside the microstrip as shown. As a result, the range of effective values of the inductance with respect to the coupling coefficient k is further expanded. For example, using three parallel microstrip lines each having a width of 2 mm and a length of 5.5 mm (see FIG. 8a), the inductive elements each having a larger inductance value are arranged in parallel, About 0.7
An effective inductance of 2 nH can be achieved. The effective inductance achieved by such a parallel structure is approximately equal to 1 / n × L. Here, n is the number of parallel microstrips each having an inductance value L. instead of one strip having a width equal to the sum of the widths of n parallel strips,
The advantage of using parallel microstrips is that the increase in coupling with increasing strip width is very small for parallel strips. However, there are manufacturing limitations on the number of microstrips that can be used in parallel strips. For one thing, as the return of each additional strip is reduced, and for another, as the overall width of the inductor strip increases, the impedances begin to spread rather than concentrate. The filter circuit using the inductive element consisting of three lines in FIG. 8a is illustrated in FIG. 8b as the value of each element.

For the prior art application of the topology with conventional lumped inductance elements (FIG. 5), the improved response of the double-tuned resonator topology with microstrip inductance elements is shown in FIGS. 9a and 9b. It becomes apparent by comparing the simulated output response of (invention) with FIGS. 6a and 6b (prior art). In a first embodiment of the invention, a Q ₁ of about 25 (about 4% at a resonant frequency of 400 MHz is
Can be achieved, and in the prior art, Q ₁ at the same frequency is about 6.5 (
(A partial bandwidth of about 15.5%). Attenuation other than bandwidth was also significantly improved.

The use of a microstrip transmission line as an effective inductive element of an electromagnetically coupled resonator is very characteristic and novel compared to the prior art which uses a microstrip transmission line as a resonator. One of ordinary skill in the art will understand that this is the case. The use of a microstrip transmission line as a resonator when the length of the microstrip transmission line is a suitable portion of the center frequency or resonance frequency (typically one quarter wavelength) is inherent to the transmission line. Due to resonance. The width of the microstrip of the present invention is approximately 0.5% to 10% of the wavelength of the desired resonance frequency.
Is. They can effectively act as lumped inductive elements rather than as distributed impedance of the transmission line resonator. As described above, when a transmission line is used as the resonator in the intended wide bandwidth application example, an extremely long transmission line is required at low frequencies.

A second preferred embodiment of the invention is shown in FIG. 10a. Additional capacitor (Cs ₁₄
31 and Cs ₂ 433), respectively, in the topology of the first preferred embodiment of the present invention (
It is arranged in series with the parallel tuned input resonator 432 and the output resonator 434 of FIG. 7). The values of Cs ₁ 431 and Cs ₂ 433 are very small when compared to the values of shunt capacitors C _P1 76 and C _{P 2} 78. Although adding capacitors in series in this way would be contrary to the idea of those skilled in the art, the addition of C _S1 431 and C _S2 434 significantly improves the response of the bandpass filter of the first preferred embodiment. It
This addition of two very cheap elements turns the bandpass filter from a fourth order filter to a sixth order filter. This can be confirmed by comparing the transfer functions obtained by implementing the parallel double-tuned topology of the present invention (FIG. 7) with the modified topology of FIG. 10a.

[0041]   The transfer function of the topology of FIG. 7 is calculated by the following formula.

[0042]

[Equation 1]

[0043]   The transfer function of the enhanced topology of FIG. 10a is given by:

[0044]

[Equation 2]

Where S is the composite frequency (ie, σ + jω), g ₀ and g _P are coefficients, and a ₁ , b ₁ , a ₂ , b ₂ , c ₁ , d ₁ , e ₁ , c The coefficients of the ₂ , ₂ , and c ₂ polynomials. A pole in the transfer function that determines the frequency response of the modified filter.
, The roll-off at high frequencies increases from the pass band to the rejection band by changing the slope from 1 / s to 1 / s ³ when S becomes infinite. Therefore, not only Q ₁ increases further, but the attenuation at high frequencies also increases. Therefore,
C _S1 431 and C _S2 433 improve the low frequency performance of the filter.

A band circuit having a center frequency of 70 MHz using the topology of FIG. 10a (including the microstrip transmission line of the first preferred embodiment) is shown in FIG. The simulated output response of the filter of Figure 11 is illustrated in Figures 12a and 12b. _{Q L} of the circuit is about 21, fractional bandwidth is about 4.8%.

An example of a band circuit having a center frequency of 400 MHz using the topology of FIG. 10a (including the microstrip transmission line of the first preferred embodiment) is shown in FIG. The simulated output response of the filter of Figure 13 is illustrated in Figures 12a and 12b. Q _L of the circuit is about 21, fractional bandwidth of about 4
． 8%.

An example of a band circuit having a center frequency of 800 MHz using the topology of FIG. 10a (including the microstrip transmission line of the first preferred embodiment) is shown in FIG. The simulated output response of the filter of Figure 15 is illustrated in Figures 16a and 16b. Q _L of the circuit is about 15, fractional bandwidth of about 6
． 6%.

An example of a band circuit having a center frequency of 400 MHz using the topology of FIG. 10a (including multiple parallel microstrip transmission lines of FIGS. 8a and 8b) is shown in FIG. The simulated output response of the filter of Figure 17 is illustrated in Figures 18a and 18b. _{Q L} of the circuit is about 34, fractional bandwidth is about 2.9%.

An example of a band circuit having a center frequency of 400 MHz using the topology of FIG. 10a (including the microstrip transmission line of the first preferred embodiment) is shown in FIG. The additional resonator 1900 includes an input resonator 432 and an output resonator 434.
Is connected between and. Resonator 1900 includes input and output resonators 432 and 4
It has a topology similar to 34 and includes a capacitor C _P3 1902 in parallel with a microstrip inductive element 1904. The simulated output response of the filter of Figure 19 is illustrated in Figures 20a and 20b. _{Q L} is about 19.5 parts bandwidth of this circuit is about 5%.

An example of a high frequency bandpass filter circuit having a center frequency of 400 MHz using the topology of FIG. 10a (including the microstrip transmission line of the first preferred embodiment) is shown in FIG. The circuit includes a balanced input of input resonator 432 and an unbalanced output of output resonator 434 (or vice versa). This circuit can be used as a signal combiner or a signal splitter in the band frequency range. The simulated output response of the filter of Figure 21 is illustrated in Figures 22a and 22b. _{Q L} of the circuit is about 2.4, fractional bandwidth is about 42%.

A third preferred embodiment of the present invention is illustrated in FIG. Additional capacitor (
C _P1 350 and C _P2 370) are respectively connected in parallel with the conventional topology series tuned input resonator 320 and output resonator 340 of FIG. C _P1 350
And C _P2 370 are much larger than the values of the series capacitors C _S1 11 and C _S2 13 of FIG. 1a. Although the addition of such a parallel capacitor seems contrary to the idea of one of ordinary skill in the art, the addition of C _P1 350 and C _P2 370 is shown in FIG.
a and the response of the conventional topology bandpass filter of FIG. 3 is significantly improved. By adding these two elements, which are very cheap, the bandpass filter can move from the 4th filter to the 6th filter in the same way as a series capacitor is added to the parallel tuned circuit of FIGS.
Change to place filter. The transfer function of the embodiment of the present invention having the modified topology of FIG. 23 is basically the same as the transfer function of the topology of FIG. 10a described above. This is because they are logical duels to each other.

An example of a band circuit with a center frequency of 70 MHz using the topology of FIG. 23 (using an air-core coil as an inductor to achieve the high inductive value required for high Q _L ) is shown in FIG. There is. The simulated output response of the FIG. 24 filter is illustrated in FIGS. 25a and 25b. _{Q L} of the circuit is about 46, fractional bandwidth is about 2.2%.

An example of a band circuit with a center frequency of 400 MHz using the topology of FIG. 23 (using an air-core coil as an inductor to achieve the high inductance value required for high Q _L ) is shown in FIG. There is. The simulated output response of the FIG. 26 filter is illustrated in FIGS. 27a and 27b. _{Q L} is about 33 of this circuit.
33, and the partial bandwidth is about 3%.

An example of a band circuit with a center frequency of 70 MHz using the topology of FIG. 23 (using an air-core coil as an inductor to achieve the high inductance value required for high Q _L ) is shown in FIG. There is. The simulated output response of the FIG. 28 filter is illustrated in FIGS. 29a and 29b. _{Q L} of this circuit is about 34.8
And the partial bandwidth is about 2.9%.

FIG. 30 is a table showing values for various examples of parallel double-tuned topologies using microstrip lines as inductive elements of the circuit, including dimensions and other relevant information.

Above a frequency of about 1 GHz, the electromagnetic coupling is beyond the point where the resonators are spaced apart and the decrease in mutual inductance M can be used to compensate for the increased electromagnetic coupling to maintain optimal coupling. Increase. Moreover, beyond the point where the value of the effective inductance M of each resonator (parallel tuned embodiment of FIG. 7 or FIG. 10a) can be reduced by simply shortening the length of the metal strip, the reduction in frequency is Q ₁ To reduce. In the preferred embodiment, the minimum width of a printed circuit board manufactured with standard manufacturing tolerances is typically 5 mm. If the length of the strip is influenced by manufacturing errors, the errors will be reflected in the response of the filter and it will not be possible to achieve the small partial bandwidth required for the intended application. Further, as noted above, there is a limit to the number of elements that can be placed directly in parallel to reduce the effective inductance of each resonator.

Therefore, a fourth embodiment of the invention is disclosed in FIG. 32a. Each resonator with its original topology (FIGS. 7 and 10a) is a mirror image connected to its signal line as shown. This topology provides two important features that allow it to be applied to frequencies in the range of about 500 MHz to over 2 GHz. First, it is possible to achieve an effective inductance value of each resonator that is lower than the value obtained by shortening the metal strip based on manufacturing accuracy. For inductive element _L 2a 510 and _{L 2} b 512 the inductive element _L 1a 508 and _L 1b 509 input resonator output resonator are parallel to each other, the effective inductance of the input resonator and the output resonator 50% It is reduced more than that.

Since the value of the inductance can be further lowered, increasing the frequency to cancel the small value of Q _L results in parallel capacitors C _P1a 504 and C _P1b 5
06 and C _P2a 514 and C _P2b 516 have large values. Furthermore, the effective inductance of each resonator is determined by the parallel combination of microstrips (606, 608, 610 and 612 of FIG. 10d) as described with reference to FIGS. 8a and 30.
Respectively) as L1a508, L1b509, L2a510 and L2b5
By implementing 12, the value of the effective inductance can be further reduced. As mentioned above, there is a limit to the number of microstrips that can be placed in parallel in this way. In the example illustrated in Figure 32b, a separate inductive element 606,
By placing microstrips such as 608, 610 and 612 in parallel, it is possible to obtain an inductance that is lower than the value of the inductance that is simply obtained.

A second important feature of this topology applicable to frequencies in the 1-2 GHz range is that it is substantially anti-parallel. Since the direction of the current flowing through the inductive element is in the opposite direction, the mutual coupling between the resonators tends to be canceled and the mutual inductance M between the resonators (and thus the overall electromagnetic coupling) is substantially reduced. . Therefore, it is easier to keep the electromagnetic coupling within the optimum range by varying M, which is a function of the distance between the resonators of the circuit, even if the frequency is in the range of 1-2 GHz or higher. Is possible.

The configuration of the antiparallel topology of the present invention in which the mutual inductance between the resonators is substantially canceled is described below with reference to FIGS. 33a-d. Analysis is performed at various stages, assuming that the width of the inductor is zero. In the first stage, the mutual inductance between the inductive elements L _1a 710 and L _1b 712 is first considered, as shown in FIG. 33a. The inductance between these two inductive elements can be calculated by the following formula.

[0062]

[Equation 3]

In the second stage, the mutual inductance between the inductive elements L _1a 710 and L _1b 714 is determined by the following equation.

[0064]

[Equation 4]

In a third step, the circuits of Figures 33a and 33b are combined to produce the circuit of Figure 33c. Calculate the total mutual inductance by the following formula.

[0066]

[Equation 5] Further, it can be expressed as the following equation,

[0067]

[Equation 6]

[0068] Further, the following equation is obtained.

[0069]

[Equation 7]

Therefore, the mutual inductance between L _1a and the dipole consisting of L _2a and L _2h is substantially independent of the spacing between the inductors. The last step of analyzing the mutual inductance between the mirror image resonator of the present invention, by determining the mutual inductance between the dipole consisting of the inductor _{L 1b} and _{L 2a} and _{L 2 b (M 1b, 2a} , 2b) and is there. This mutual inductance has the same formula as the mutual inductance of L _1a and L _2a and L _2b , but since the current flowing through L _1b is in the opposite direction to the current flowing through L _1a , the sign is opposite and is shown as follows. Be done.

[0071]

[Equation 8]

Superimposing L _1b and dipoles on top of the configuration of FIG. 33c results in the mirror image resonator topology of the invention illustrated in FIG. 33d. Therefore, the mutual inductance between the mirror image resonators is obtained by the following equation.

[0073]

[Equation 9]

Therefore, the mutual inductance between the mirror image resonators of the present invention is substantially zero because the length of the inductive element is longer than the spacing between the resonators.

As mentioned above, the above analysis was performed assuming a zero inductive element width. The preferred value of the mutual inductance of the mirror image resonator structure to achieve optimum coupling is:
It depends on the width of the inductive element. However, most of the induced currents, which increase with increasing frequency, cancel each other out and favor the circuit. Note that the mutual inductance between the resonators can be controlled to the extent that the inductive elements are in parallel with each other or to the extent that they are not. As one inductive element of the resonator rotates relative to the other inductive element, the cancellation of the inductive current decreases accordingly.

[0076]   Figure 32b illustrates a preferred embodiment of the mirror image resonator topology. Circuit of Figure 32b
32a has a higher order than that of the circuit of FIG. 32a due to the added third resonator 602.
Is. Resonator 602 is inverted with respect to resonators 600 and 604, but
The works are equivalent. Therefore, the resonators 600 and 604 correspond to the resonator 60 of FIG. 32c.
It may be inverted, as illustrated as 0i and 604i. Like this
By using the name, the degree of freedom in the physical layout of the circuit is expanded. Figure 32b
And the embodiment of FIG. 32c shows the inductive element L as three microstrips in parallel. _1a 606, L_1b608, L_2a610, L_2b612, L_3a614 and
L_3b616, each of which is a parallel microstrip inductor
It provides an effective inductance of about 1/3 of the drawer. Each of the three resonators
The total effective inductance is reduced by more than 50%, so individual microstrips
Is less than 1/6 of the inductance.

A shunt capacitor (eg, resonator 6) for each resonator of the embodiment of FIGS. 32a-32c.
00 C _P1a 618 and C _P12 620) are also in parallel, and their respective values add up to the total effective shunt capacitance of each resonator. Implementing each shunt capacitor as two or more capacitors in parallel has the additional advantage of creating parasitic resistances and inductances in parallel for each capacitor, which significantly reduce them and improve the performance of the filter circuit. Be improved.

An example of a mirror image resonator topology is shown in FIG. 34a. This is the same circuit as described using FIG. 32b. In these two figures, similar elements have been given similar numbers. The effective inductance of each of the resonators 600, 602 and 604 is 1.5 nH. The center frequency is 1015.75 MHz, which is a pass band of 30 MHz. Actual values of capacitors and inductive elements, length and width of inductive elements, gap dimensions, and gap G between parallel microstrips.
650, microstrip width W654, microstrip length L656
And the gap S652 between the resonators is disclosed. Using the mirror image resonator topology of the present invention, a resonator with an effective inductance well below 0.5 nH can be achieved. As the manufacturing process becomes more accurate, the minimum effective inductance of the resonator naturally decreases accordingly.

34b, 34c and 34d illustrate measured transfer functions for the circuit of FIG. 34a. The frequencies at the 3 dB point are 1000 MHz and 1030, respectively.
MHz, with a partial bandwidth of 3%, the circuit Q _L is 34. Figure 34e
Shows the measured return loss of the circuit of Figure 34a.

In addition to its application in filters, the invention can be used in a variety of applications where the low insertion loss and the inherent frequency agnostic properties are of great advantage. One of such application examples is application to the feedback path of the oscillator shown in FIG. By connecting the input / output ports of the coupled resonator filter 400 to the input / output ports of the RF amplifier 3100, a feedback path is provided from the output of the amplifier 3100 to the input, effectively looping around the amplifier 3100. To be closed. If the loop gain is greater than 1 (ie, the gain of the amplifier 3100 is greater than the insertion loss of the feedback path), then at a frequency where the phase shift around the loop is 0 degrees (ie, a multiple S of 360 degrees), Vibration occurs. The phase shift of the coupled resonator structure 400 is 180 degrees at the center frequency, and by using an inverting amplifier (internal phase shift is 180 degrees), the total phase shift is 3 degrees.
It is 60 degrees, which corresponds to the condition required for vibration. The coupled resonator with 0 degree phase shift (eg, rotating the microstrip at input 72 or output 74 180 degrees with respect to each other), and subsequently the non-inverting amplifier of amplifier 3100, is oscillated. Matches the condition of.

The narrow frequency band of the electromagnetically coupled resonator (ie, the large Q _L ) is associated with a steep slope near the center frequency. Due to the steep slope in this feedback loop, FIG.
1 oscillator phase noise performance is improved.

[Brief description of drawings]

FIG. 1A   The figure which shows the topology of the conventional series double tuning type electromagnetic coupling type resonator.

FIG. 1B   The figure which shows the topology of the conventional parallel double tuning type magnetic coupling type resonator.

FIG. 2 is a graph showing a typical response of the resonator of FIGS. 1A and 1B as the coupling coefficient k changes.

[3] has an ultimate value of a number of components in the embodiment of the existing resonator so as to achieve maximum Q _L, it illustrates an example of a series resonator of Figure 1A.

4A is a graph showing a simulation of the response of the conventional resonator of FIG. 3 on a broad scale for both frequency (40 MHz on one scale) and attenuation (10 dB on one scale).

FIG. 4B is a graph showing a simulation of the response of the conventional resonator of FIG. 3 on a smaller scale for both frequency (1 division 10 MHz) and attenuation (1 division 1 dB).

[5] has an ultimate value of a number of components in the embodiment of the existing resonator so as to achieve maximum Q _L, it illustrates an example of a parallel resonator in FIG 1B.

FIG. 6A is a graph showing a simulation of the response of the conventional resonator of FIG. 5 on a wide scale for both frequency (40 MHz on a scale) and attenuation (10 dB on a scale).

FIG. 6B is a graph showing a simulation of the response of the conventional resonator of FIG. 5 on a smaller scale for both frequency (1 division 10 MHz) and attenuation (1 division 1 dB).

FIG. 7 shows an example of a first preferred embodiment parallel resonator of the present invention using a small grounded microstrip transmission line that achieves a very small but precise effective inductance.

8A is a plan view showing the physical form of the microstrip effective inductance element of the present invention. FIG.

8B shows an example of the parallel resonator of FIG. 7 with the inductance element divided into three parallel microstrips as shown in FIG. 8A to achieve a low effective inductance in the resonator.

9A is a graph showing a simulation of the response of the resonator of FIG. 8B on a broad scale for both frequency (40 MHz on one scale) and attenuation (10 dB on one scale).

9B is a graph showing a simulation of the response of the resonator of FIG. 8B on a smaller scale for both frequency (10 MHz on a scale) and attenuation (1 dB on a scale).

FIG. 10A shows a parallel tuned resonator circuit using a microstrip transmission line as a bulk inductance element with an additional capacitive element connected in series between the resonator and the input and output signals.

10B illustrates the physical form of the parallel-tuned resonator of FIG. 10A using printed circuit board manufacturing techniques.

11 shows an example of the circuit of FIG. 10 giving numerical values of the components that achieve a 70 MHz narrow band filter.

FIG. 12A is a graph showing a simulation of the response of the resonator of FIG. 11 on a wide scale for both frequency (40 MHz per scale) and attenuation (10 dB per scale).

12B is a graph showing a simulation of the response of the resonator of FIG. 11 on a smaller scale for both frequency (10 MHz on a scale) and attenuation (1 dB on a scale).

FIG. 13 shows an example of the parallel tuned resonator of FIG. 10 with the numerical values of the components that achieve a 400 MHz narrow band filter.

FIG. 14A is a graph showing a simulation of the response of the resonator of FIG. 13 on a wide scale for both frequency (40 MHz per scale) and attenuation (10 dB per scale).

14B is a graph showing a simulation of the response of the resonator of FIG. 13 on a smaller scale for both frequency (10 MHz on a scale) and attenuation (1 dB on a scale).

FIG. 15 shows an example of the parallel-tuned resonator of FIG. 10A with the numerical values of the components that achieve the 800 MHz bandpass filter.

FIG. 16A is a graph showing a simulation of the response of the resonator of FIG. 15 on a broad scale for both frequency (40 MHz on one scale) and attenuation (10 dB on one scale).

16B is a graph showing a simulation of the response of the resonator of FIG. 15 on a smaller scale for both frequency (10 MHz on a scale) and attenuation (1 dB on a scale).

17 is a diagram showing an example in which the inductive element of each resonator is realized by three parallel microstrips in order to further reduce the inductance value of the resonator in the parallel-tuned resonator of FIG. 10A.

FIG. 18A is a graph showing a simulation of the response of the resonator of FIG. 17 on a wide scale for both frequency (40 MHz per scale) and attenuation (10 dB per scale).

FIG. 18B is a graph showing a simulation of the response of the resonator of FIG. 17, shown on a smaller scale for both frequency (10 MHz on a scale) and attenuation (1 dB on a scale).

19 has three parallel resonators to achieve a 400 MHz narrow band filter, FIG.
The figure which shows the Example of the parallel tuning type resonator of 0.

20A is a graph showing a simulation of the response of the resonator of FIG. 19 on a broad scale for both frequency (40 MHz on a scale) and attenuation (10 dB on a scale).

20B is a graph showing a simulation of the response of the resonator of FIG. 19 on a smaller scale for both frequency (10 MHz on a scale) and attenuation (1 dB on a scale).

FIG. 21: FIG. 1 with a balanced-unbalanced transformer to achieve a 400 MHz narrow band filter.
The figure which shows the Example of the parallel tuning type resonator of 0A.

22A is a graph showing a simulation of the response of the resonator of FIG. 21 on a broad scale for both frequency (40 MHz on one scale) and attenuation (10 dB on one scale).

22B is a graph showing a simulation of the response of the resonator of FIG. 21 on a smaller scale for both frequency (10 MHz on a scale) and attenuation (1 dB on a scale).

FIG. 23 shows a series tuned resonator with an additional capacitor connected in parallel between the input and output signals and the resonator and using an air-core coil as the inductive element.

FIG. 24 shows an example of the series tuned resonator of FIG. 23 with numerical values of the components that achieve a 70 MHz narrow band filter.

FIG. 25A is a graph showing a simulation of the response of the resonator of FIG. 24 on a wide scale for both frequency (40 MHz per scale) and attenuation (10 dB per scale).

FIG. 25B is a graph showing a simulation of the response of the resonator of FIG. 24 on a smaller scale for both frequency (10 MHz on a scale) and attenuation (1 dB on a scale).

FIG. 26 shows an example of the series tuned resonator of FIG. 23 with numerical values of components that achieve a 70 MHz narrow band filter.

FIG. 27A is a graph showing a simulation of the response of the resonator of FIG. 26 on a broad scale for both frequency (40 MHz per scale) and attenuation (10 dB per scale).

27B is a graph showing a simulation of the response of the resonator of FIG. 26 on a smaller scale for both frequency (10 MHz on a scale) and attenuation (1 dB on a scale).

28 shows an example of the series tuned resonator of FIG. 23 with the numerical values of the components that achieve the 800 MHz narrow band filter.

FIG. 29A is a graph showing a simulation of the response of the resonator of FIG. 28 on a broad scale for both frequency (40 MHz per scale) and attenuation (10 dB per scale).

29B is a graph showing a simulation of the response of the resonator of FIG. 28 on a smaller scale for both frequency (10 MHz on a scale) and attenuation (1 dB on a scale).

FIG. 30 is a diagram showing a table presenting equivalent bulk inductance values for the resonators of the examples shown in FIGS. 8B, 11, 13, 13, 15, 17, and 21.

FIG. 31 shows an example using the parallel tuned resonator circuit of FIG. 10A to achieve a 400 MHz oscillator.

32A illustrates an example of a mirror image topology of the present invention as applied to the parallel tuned resonator of FIG. 10A.

FIG. 32B illustrates an example of a mirror image topology of the present invention as applied to a parallel tuned resonator with three or more cascaded resonators with multiple strips in parallel for each inductance element of the resonator.

FIG. 32C   FIG. 32B shows the mirror image topology symmetry as applied to the cascade resonator of FIG. 32B.

FIG. 33 is a diagram including steps A to D showing steps of stepwise determination of an induced current in the mirror image topology of the present invention.

FIG. 34A is implemented using printed circuit board process technology and is 1015.75 MHz.
32B illustrates an example of a mirror image topology of the present invention applied to the cascade circuit of FIG. 32B having numerical values of the components that achieve the narrow band filter of FIG.

FIG. 34B is a graph showing a simulation of the response of the resonator of FIG. 34A on a broad scale for both frequency (100 MHz range) and attenuation (1 dB 10 dB).

FIG. 34C is a graph showing a simulation of the response of the resonator of FIG. 34A on a smaller scale for both frequency (6 MHz range) and attenuation (1 dB 0.1 dB).

34D is a graph showing a simulation of the response of the resonator of FIG. 34A on a very broad scale for both frequency (3 GHz range) and attenuation (1 division 10 dB).

FIG. 34E is a graph showing measured return loss of the resonator of FIG. 34A at a frequency range of 100 MHz and an attenuation scale 1 scale of 5 dB.

─────────────────────────────────────────────────── ─── Continued front page    (81) Designated countries EP (AT, BE, CH, CY, DE, DK, ES, FI, FR, GB, GR, IE, I T, LU, MC, NL, PT, SE), OA (BF, BJ , CF, CG, CI, CM, GA, GN, GW, ML, MR, NE, SN, TD, TG), AP (GH, GM, K E, LS, MW, SD, SL, SZ, TZ, UG, ZW ), EA (AM, AZ, BY, KG, KZ, MD, RU, TJ, TM), AE, AL, AM, AT, AU, AZ, BA, BB, BG, BR, BY, CA, CH, CN, C R, CU, CZ, DE, DK, DM, EE, ES, FI , GB, GD, GE, GH, GM, HR, HU, ID, IL, IN, IS, JP, KE, KG, KP, KR, K Z, LC, LK, LR, LS, LT, LU, LV, MA , MD, MG, MK, MN, MW, MX, NO, NZ, PL, PT, RO, RU, SD, SE, SG, SI, S K, SL, TJ, TM, TR, TT, TZ, UA, UG , UZ, VN, YU, ZA, ZW [Continued summary] The enclosure becomes wider with increasing accuracy.

Claims (15)

[Claims] 1. A predetermined circuit including a first resonator magnetically coupled to a second resonator, the first and second resonators having a first capacitance. A first capacitor having a first inductive element having a first inductance connected between a signal line and a first ground, and a second capacitor having a second capacitance and the first capacitor. A second inductive element having a second inductance connected between the signal line and a second ground such that currents through the inductive element and the second inductive element flow in generally opposite directions; And a product of the first capacitance and the first inductance is substantially equal to a product of the second capacitance and the second inductance. 2. The signal line of the first resonator is for transmitting an input signal to the first resonator, and the signal line of the second resonator is from the circuit to the load. For transmitting an output signal, the input signal being connected to the first resonator via a first coupling capacitor in series with the first resonator,
The circuit of claim 1, wherein the output signal is connected to the load via a second coupling capacitor in series with the second resonator. 3. The circuit according to claim 1, wherein the first and second inductances and the first and second capacitances have the same inductance value and capacitance value, respectively. 4. The optimum coupling between the first and second resonators is
Method according to claim 1, characterized in that it is maintained over a range of frequencies by changing the physical distance between the inductive elements of the first and second resonators. 5. One or more of said first and second inductive elements of each of said first and second resonators is a bulk inductance caused by a metal line residing on a generally non-conductive surface. The circuit according to claim 1, comprising: 6. The circuit of claim 5, wherein the one or more inductive elements are formed by two or more metal wires connected in parallel with each other. 7. The one or more first and second capacitors are formed by two or more parallel capacitors in order to reduce parasitic effects on the first and second capacitors. The circuit according to claim 1. 8. A predetermined circuit comprising two or more resonators magnetically coupled in series with each other, each of the two or more resonators having a first capacitor and a first capacitor having a first capacitance. Signal line and first
A first inductive element having a first inductance and a second capacitor having a second capacitance, the first inductive element and the second inductive element connected between the first inductive element and the second inductive element. A circuit further comprising a second inductive element having a second inductance connected between the signal line and a second ground such that the current passing therethrough flows generally in opposite directions. 9. The signal line of the two or more first resonators receives the input signal as the first signal.
For transmitting to the resonator, the signal line of the two or more second resonators is for transmitting an output signal from the circuit to the load, and the input signal is the first signal. Is connected to the first resonator via a first coupling capacitor in series with one resonator, and the output signal is coupled to the first resonator via a second coupling capacitor in series with the second resonator. The circuit according to claim 8, wherein the circuit is connected to a load. 10. The first and second inductances and the first and second capacitances of each of the two or more resonators have the same inductance value and capacitance value, respectively. The circuit according to 1. 11. Optimal coupling between the two or more resonators over a range of frequencies by changing the physical distance between the inductive elements of the first and second resonators. The method of claim 1, wherein the method is retained. 12. One or more of the first and second of each of the two or more resonators.
The circuit of claim 1 wherein the inductive element has a bulk inductance caused by a metal wire residing on a generally non-conductive surface. 13. The one or more inductive elements are connected in parallel with each other.
The circuit according to claim 5, wherein the circuit is formed by the above metal wire. 14. The one or more first and second capacitors are formed by two or more parallel capacitors to reduce parasitic effects on the first and second capacitors. The circuit according to claim 1. 15. A method for maintaining high angular load Q and optimum coupling of a parallel tuned series resonant circuit in an extended frequency range, said circuit being magnetically coupled in series with each other. Having the above tuning resonators, each of the resonators is
The inductive element comprising an inductive element having an inductance L connected between a signal line and a ground and a capacitor element having a capacitance C connected between the signal line and a ground. Is implemented as a bulk inductance generated by a metal wire on a substantially non-conductive surface, a step of canceling out almost all mutual induction currents between the two or more resonators, and Decreasing the value and increasing the value of C, and controlling the coupling between the two or more resonators by changing the physical distance between the two or more resonators. How to characterize.