Gain control filter circuit, power module comprising a filter circuit and method of adjusting an RF filter circuit to provide a controllable gain

The present invention relates to impedance processing between an amplifier and an RF filter. In particular, the present invention refers to providing filter systems that provide a controllable gain for a spectrum of frequencies.

For example in wireless communication devices transmission signals are sent via an antenna and reception signals are received via the antenna. Generally, transmission signals should have a high power level while reception signals usually have a low power level. Duplexers are utilized to separate transmission signals from reception signals. The power of transmission signals propagating to the antenna is usually increased by a power amplifier. RF filters between power amplifiers and an antenna are needed in order to avoid a contamination of the communication device's environment with unwanted signals. However, conventional power amplifiers have a low output impedance and RF filters have an input impedance that may be adjusted to, e.g., 50 ohm. Thus, the signal path between the power amplifier and the RF filter needs an impedance-matching circuit to match the output impedance of the power amplifier to the input impedance of the filter. Such a configuration works fine for a

predetermined frequency, a predetermined power level and a predetermined amplifier gain. However, dependencies between the amplifier gain, the power level and the signal's

frequencies - in particular if further external influences such as other signal path' s influences and the spatial proximity to other signal paths - result in problematic electric and electromagnetic properties.

As a consequence thereof, a possible solution would be to provide one individual RF filter for each needed combination of power level, amplifier gain and RF frequency.

In particular due to the trend towards more and more

frequency bands and more and more signal paths within a single mobile communication device such a solution would not be desirable.

Thus, what is needed is an RF filter circuit that provides a constant amplifier gain for a plurality of different

frequencies or for a plurality of different power levels.

To that end, a gain control filter circuit, a power module comprising a filter circuit and a method of adjusting an RF filter circuit to provide a controllable gain according to the independent claims are provided. Dependent claims provide preferred embodiments.

A gain control filter circuit comprises an input port and an output port. Further, the filter circuit comprises an

impedance-matching circuit between the input port and the output port. Further, the filter circuit comprises an RF filter between the matching circuit and the output port. The impedance-matching circuit is provided for matching the input impedance of the input port to the input impedance of the RF filter. The input port is connected to the output port of a power amplifier. The input port is configured to receive a constant power level. The input port is configured to receive power at a predetermined impedance. The impedance matching circuit is configured to provide an output impedance the value of which is located on a circle of a predetermined radius around the predetermined conjugated impedance of the impedance at the input port in a Smith chart representation.

The filter circuit can be connected between a power amplifier and an antenna port, e.g. in a mobile communication device. The filter circuit can be provided for working with different frequencies or a spectrum of frequencies. It is to be noted that the impedance matching is different from adjustment: the impedance matching denotes a transformation of impedance from a first impedance level to a significantly different

impedance level. For example impedance matching can be performed from an impedance in the milliohm range to a standard impedance of, e.g., 50 ohm, 100 ohm or 200 ohm. For example, the adjustment can be made to compensate for an impedance change of a factor of 10.

Varying RF frequencies can be needed, if a certain frequency spectrum, e.g. typical RF frequency bands for mobile

communication and the like are used. The possibility of varying the power level may be useful, e.g. if no high power signal is needed and energy could be saved by reducing the power level or vice versa.

In particular, the number of needed RF filters can be

significantly reduced.

Correspondingly, it is possible that the input port of the filter circuit is connected to the output port of an

amplifier of a mobile communication device. Further, it is possible that the output port of the filter circuit is electrically connected to or coupled to antenna, e.g. an antenna of a mobile communication device.

Further, it is possible that the filter circuit is part of an RF module comprising one or more power amplifiers and one or more RF filters.

The trend towards miniaturization demands for small spatial dimensions. One possibility of complying with this trend is to increase the integration density of electrical components. Correspondingly, modules can be provided that have the power amplifier, the filter circuit, and optionally other circuit elements integrated and mechanically and electrically

connected within a small volume. Such a module can, for example, be a power amplifier module with integrated

duplexers to be used in a fronted module of a mobile

communication device.

Other signal paths can also comprise additional power

amplifiers and corresponding filter circuits. Switches or further duplexers can be used to combine parallel signal paths to a common antenna port. However, it is possible that different signal paths or combinations of different signal paths are connected to different antennas.

The capacitance value, the inductance value and the

resistance value, respectively, can be controlled by a controller. The controller can be a part of the filter circuit or a part of an external circuit environment, e.g. of the mobile communication device. A varicap diode is a type of diode designed to have a

voltage-dependent capacitance. Such diodes can be used to establish voltage-controlled capacitors where a bias voltage applied to the diode determines the capacitance of the diode for AC signals, e.g. RF signals.

Further, a capacitor bank as a possible element of variable capacitance can comprise two or more individual capacitors. Each capacitor can be individually switched to or separated from external connections of the variable capacitance

element. The individual capacitors of the capacitor bank can be electrically configured in a parallel configuration such that the overall capacity of the capacitor bank equals the sum of the individual capacities of the individual capacitors that are active, i.e. that are electrically connected via the corresponding switches to the capacity element's external contacts .

It is possible that the capacity of one individual capacitor has twice the capacity compared to the capacitor of the next smaller size. Thus, the capacitors have individual capacities equal to 2 ⁿ times a basic capacitance C _Q ·

With n individual capacitors a total number of 2 ⁿ equally spaced capacitance values can be obtained.

It is possible that the output impedance of the impedance matching circuit is independent from a frequency of an RF signal while the frequency is within a predetermined

frequency range.

The predetermined frequency range can be, for example, a transmission frequency band or a reception frequency band or a combined transmission/reception frequency band of a mobile communication system. Further, the width of the predetermined frequency range can be wide enough for two or more frequency bands of mobile communication systems to fit in.

Thus, for each needed frequency within a frequency band the electric behaviour of the filter circuit, especially with respect to gain and power level, is constant.

It is possible that the output impedance of the impedance matching circuit determines a conjugated impedance. The conjugated impedance determines a plurality of frequency- dependent output impedances of a single amplifier gain when the input port is connected to a power amplifier.

In this context the term "conjugated impedance" denotes an impedance that is characterized in that in a Smith chart diagram frequencies of a same amplifier gain lie on a circle of a certain diameter around the conjugated impedance such that a specific frequency refers to a specific point on the circle .

As the relevant frequencies can be limited to a finite frequency interval, it is possible that the corresponding impedance values do not lie on a whole circle but lie on a circle arc only.

It is possible that the RF filter is a bandpass filter.

In particular, it is possible that the RF filter is an electroacoustic filter working with electroacoustic waves. Thus, it is possible that the RF filter is selected from an SAW filter, a BAW filter, a GBAW filter and a TFSAW filter.

Electroacoustic filters work with acoustic wave propagating at a surface of a piezoelectric material or within a

piezoelectric material. SAW filters (SAW = surface acoustic wave) , GBAW filters (GBAW = guided bulk acoustic wave) , and TFSAW filters (TFSAW = thin film surface acoustic wave) employ acoustic surface waves. BAW filters (BAW = bulk acoustic wave) employ bulk acoustic waves.

In any case, electrode structures electrically and

mechanically connected to a piezoelectric material due to the piezoelectric effect convert between RF signals and acoustic waves. In filters working with surface waves comb like electrode structures are arranged at the surface of the piezoelectric material. In BAW filters a sandwich structure comprises a piezoelectric material between a bottom electrode and a top electrode and bulk acoustic waves propagate in a vertical direction.

Acoustic energy can be confined to a resonator structure due to acoustic reflectors such as structured finger-shaped elements or an acoustic mirror.

Such electroacoustic resonators can be combined, e.g. in a ladder-type configuration, to establish bandpass filters or band rejection filters. In a ladder-type configuration two or more electroacoustic resonators are arranged in series in a signal path while two or more shunt paths, each having at least one parallel resonator, electrically connect different nodes of the signal path to ground. It is correspondingly possible that a filter circuit as described is part of a power module. Thus, it is possible that a power module comprises a power amplifier and one or more filter circuits as described above.

It is possible that a method of controlling the gain of an RF filter circuit for a plurality of frequencies, comprises the steps of:

- determining an input power level based on a desired gain

- determining a conjugated impedance of an output impedance of an impedance matching circuit based on a desired gain,

- determining a radius of a circle - in a Smith chart

representation - around the determined conjugated impedance matching circuit,

- selecting an impedance on the circle as the input impedance of the filter circuit.

The filter is not limited to a duplexer. Quadplexers or multiplexers of a higher degree are also possible. Further improvements can be made to reduce impedance variation as much as possible. Then, variations in the transfer functions of the corresponding filters can be reduced. In particular with respect to reflections of power in signal paths that cause undesired ripple, the following is possible.

The impedance optimizations can be made with respect to a transmission filter so that the input impedance of the impedance matching circuit is as close as possible to the load light impedance, i.e. to the intrinsic impedance of the signal line. Thus, considering the specific properties of the signal line itself, it can lead to further optimizations of the filter's electrical properties. Compensation of

variations of the signal line's frequency dependence, power dependence or amplifier gain dependence can be performed at the input side of the corresponding RF filter.

Further, the input side of the filter can be provided such that its input impedance can be varied such that different gains caused by a frequency or power dependence of the circuit elements before the filter can be compensated. This can be obtained by making the filter impedance lie on a constant gain line (in a Smith chart) so that frequency variations do not alter the gain at the specific circuit node .

Another possibility to reduce passband ripple is to provide a small deviation from the circular line of a constant gain around a conjugated impedance to compensate for small errors in the filter transfer in the desired frequency band.

Filter structures can be optimized. An optimization of a filter structure can be to enhance the input impedance of the filter at those frequencies where the filter shows the greatest power dissipation.

A SAW filter (duplexer) has a maximum allowable power level for a cellular band. Defects in the duplexer usually occur with excessive power on the high side of the band generally in the smaller series elements. The maximum power depends strongly on the used power source and the duplexer

impedances. It is proposed to deviate from a desired duplexer impedance to achieve a different maximum power in the total system. The input impedance should be made at those frequency at which the respective filter receives too much power and exceeds the maximum power level. In some saw filters (duplexer) it is desirable to suppress a band close to the TX band. According to an embodiment a proper setting of the filter input impedance is used to increase the suppression in a neighbored channel. The goal is to get more gain in the desired frequency band with the aid of the whole system and to get less gain for the undesirable frequency band. This can be done by intentionally producing a mismatch of the power amplifier PA with the PA-matching circuit at the frequency of the band to be suppressed.

Thereby suppression of undesired bands can be maximized. In practice this means that the filter must be optimized for a given system. Enhancing the reflection Sll for the

undesirable band can be set as a new goal of the filter optimization routine.

In a PAMiD module the total gain for harmonics is determined by input impedance of the filter and output impedance of the PA-matching circuit. According to an embodiment it is not a goal to reduce this gain, but to shift the maximum gain from an undesirable place to a place where it is not important. By an adjustment of the PA matching network, or of the TX input impedance gives a different frequency at which maximum gain occurs. So it is possible to shift the frequency of maximal gain to location where it does not matter and where neither a neighbor channel nor a harmonics occur. By doing this less gain is produced in these channels and suppression of same can be improved.

There are four ways proposed to shift this gain peak to a point where the damage is most limited.

1) The output impedance of the PA-matching can be pushed a little bit by choosing the internal impedance a little bit different. 2) The line length of the interconnect between PA or PA matching and filter rotate the output impedance of the PA-matching. Thereby the gain peak can be shifted

3) The input impedance of a filter at high frequencies is capacitive, which means that the dimensions of the first filter element determines the input impedance of the TX filter to a large extent. Hence, by varying the dimension of the first filter element (preferably a series element) input impedance can be varied.

4) The first element of a filter element may be a series or a shunt element. A choice of a proper kind of first filter element can be used to determine the input impedance of the TX filter to a large extent.

In practice, this means that all four possibilities of shifting need to be suitably selected and weighted to achieve a proper balance towards the desired goal.

In mobile communication systems like cellular communications a system consisting of a PA, PA-matching and TX filter (e.g., a SAW duplexer) , the load line should be tuned for each frequency band to the correct impedance. For this purpose, parallel circuited capacities can be switched on or off. In a PAMiD fronted module in some places capacities are used while in other places too much capacity is already present.

According to an embodiment a method is disclosed to also use these additional capacities to make more insolation in the RX band. The additional input capacity is replaced by an

additional RX notch element with exactly the right capacity value in the TX band, then two problems have been solved. Instead of placing a capacity necessary for matching the power amplifier to the Tx filter in the matching circuit it is proposed to place the capacity at the input (towards PA) of the Tx filter parallel to the signal line. The capacitance value thereof can be selected to compensate for the frequency dependency of the matching circuit of the filter. At the same time, this capacitance is an additional notch to improve the suppression for an unwanted frequency.

The notch has not to be limited to its own RX band. The notch can be used for any frequency. In a carrier aggregation solution, the notch can be used for RX cross isolation.

Basic working principles and details of preferred embodiments are described in the schematic accompanying figures.

In the figures:

Fig. 1 illustrates a circuit configuration of a filter circuit .

Fig. 2 illustrates the use of a bandpass filter and the connection to an amplifier circuit.

Fig. 3 illustrates the integration in a power module.

Figs. 4 to 7 illustrate different possible power level and amplifier gain settings.

Fig. 8 illustrates the use of a variable capacitance element.

Fig. 9 illustrates the use of a capacitor bank.

Figure 1 shows a possible configuration of a filter circuit CGSC. The filter circuit has an input port PIN and an output port POUT. Between the input port PIN and the output port POUT a filter F is electrically connected. Between the input port PIN and the filter F an impedance matching circuit IMC is electrically connected. Between the input port PIN and the impedance matching circuit an adjustment circuit AC is electrically connected. The impedance matching circuit IMC converts between a substantially low impedance value seen at the input port PIN and an intermediate impedance at the input port of the filter F. Typically, the impedance at the input port of the F is several orders of magnitude higher than the impedance seen at the input port PIN.

However, it was observed that the seen input impedance is not constant but is susceptible to variations that have a width that is much smaller than what is to be handled by the impedance matching circuit. Thus, the impedance matching circuit IMC cannot appropriately handle the impedance

variations as stated above. To that end, the adjustment circuit AC is provided. In addition to the impedance matching circuit IMC a specific adjustment is performed by the

adjustment circuit AC.

It was observed that the output impedance of a power

amplifier that can be connected to the input port PIN depends on the power level and the frequency and determines the amplifier gain of the filter circuit. In order to avoid a plurality of individual filters F for a plurality of

different frequency and power level combinations, the

adjustment circuit AC compensates for the corresponding impedance variations such that a single RF filter F can provide the needed filter functionality.

Figure 2 illustrates the possibility of the filter F being a bandpass filter BPF. The bandpass filter can work with electroacoustic waves and comprise electrode structures electrically and mechanically connected to a piezoelectric material .

Further, Figure 2 illustrates the possibility of electrically connecting the adjustment circuit to an amplifier A, e.g. a power amplifier PA.

Figure 3 illustrates the relationship between the individual components: The filter circuit CGFC comprises the adjustment circuit AC, the impedance matching circuit IMC and the filter F and corresponding signal lines in between. The power module PM comprises the filter circuit CGFC and the amplifier, e.g. the power amplifier PA.

Figures 4 to 7 illustrate the working principle by referring to characteristic impedances: On the left-hand side of

Figures 4 to 7 a Smith chart is shown. On the right-hand side amplifier gain depending on input power level is presented for a plurality of frequencies f. The x shaped mark denotes a power level where the gain is constant for a plurality of frequencies .

For example as illustrated in Figure 1, the impedance

matching circuit IMC has a predetermined output impedance Z _m . For each output impedance Z _m of the impedance matching circuit IMC there is a conjugated impedance C(Z _m) that is characterized in that input impedances of the matching circuit are chosen to lie on a circle segment around the conjugated impedance.

In Figures 4 to 7 in the Smith charts, on the left-hand side the corresponding impedances are shown as dots lying on a circle segment having the conjugated impedance C(Z _m) as a center. The arrow accompanied by a "f" indicates the

direction in which the frequency increases in the left-hand side part and in the right-hand side part of the respective figures. On the right-hand side of the figures the input power is varied from -6 to 24 arbitrary units. The scale of the gain reaches from 7 to 11 arbitrary units.

For each of Figures 4 to 7 there is an input power level denoted by x on which the gain curves for different

frequencies f intersect each other.

Thus, for each wanted gain there is an input power level such that the gain is independent from the frequency if the corresponding interface impedances between the different stages of the filter circuit are chosen correspondingly.

Thus, a desired gain can be obtained by choosing a

corresponding input power level. Thus, the gain can be controlled such that the circuit provides a frequency

independent behaviour.

Further, it can be seen that for each wanted amplifier gain within a certain amplifier gain interval there is a

corresponding output power level that provides the wanted amplifier gain independent from the frequencies.

Thus, by choosing the output impedance of the impedance matching circuit corresponding to its conjugated impedance according to the circle segments shown in the left-hand parts of Figures 4 to 7 and by correspondingly utilizing the adjustment circuit a signal RF filter is sufficient to provide a frequency-independent behaviour for a given power level and a desired amplifier gain.

Figure 8 illustrates a basic embodiment of an adjustment circuit AC which comprises a variable capacitance element CE that electrically connects the signal path between the power amplifier and the impedance matching circuit to ground. The variable capacitance element can comprise a varicap diode (also known as varactor diode) , the capacity of which can be controlled by applying a bias voltage.

Further, Figure 9 illustrates another possibility of a variable capacitance element that bases on a capacitor bank with a plurality of individual capacitors C that can be coupled to or separated from the signal path via switches SW. In particular, for each individual capacitor C an individual switch SW can be provided.

Neither the filter circuit nor the power module nor the method of adjusting an RF filter circuit are limited to the shown technical details. Filter circuits can comprise further circuit elements and power modules can comprise further filter circuits. Consequently, methods for providing a controllable gain can also comprise further steps.

List of reference signs

A: amp1ifier

AC: adjustment circuit

BPF : bandpass filter

C (Z _m) conjugated impedance

C: capacitor

CE: capacitance element

CGFC : filter circuit

f : frequency

F: RF filter

IMC : impedance matching circuit

PA: power amplifier

PIN: input port of the filter circuit

POUT: output port of the filter circuit

SW: switch

VC: varicap diode

Zm : output impedance of the impedance matching circuit