Dual-band mixer and its design flow

The present invention relates to a dual-band mixer for use in wireless communications, especially to a mixer that is able to work with two separate bandwidths with reduced number of components and die size.

Due to the wide application of the wireless communications nowadays, lower-cost and wider-bandwidth wireless communications devices have been a target of research and development in the industry. In order to satisfy the needs in the wireless applications, the industry has been driven to migrate to the 2-6 GHz frequency range. For wireless LAN applications, industrial standards target the carrier frequency to either the 5.2-GHz band, by the IEEE 802.11a standard, or the 2.4-GHz band, by the IEEE 802.11g standard. Dual-band transceivers are thus proposed by the industry to provide more flexible and convenient usages of the wireless LAN systems.

Among the dual-band transceivers as announced, most of them contain two different mixers in one transceiver chip. Under such a design, circuit of the transceiver chip is made even complicated and dimension of the chip would be made twice that of a single-band mixer, since all the circuit elements are duplicated. Among all the elements inductors at the load are the most space-consuming component, since 4 load inductors, 2 for each mixer, are required in the differential architecture of the circuit. This becomes a main obstacle in reducing the cost of dual-band applications.

In order to shrink the chip size, it is possible to merge two mixers into a single mixer. This may reduce number of inductors at load in the differential architecture from 4 to 2. However, the load of the mixer was designed to operate at only one specific frequency. As a result, such dual-band mixer can only be used in a dual-band receiver, not a dual-band transmitter. A load that can be shared by two bands was proposed by using a band pass filter with a notch at output to divide the pass band into two different bands. See S. Lavasani et al.: “Pseudo-concurrent 0.18/spl mu/m dual-band CMOS LNA,” Radio Frequency Integrated Circuits (RFIC) Symposium, pp. 8-10, June 2003. The circuit was simplified because 2 bands used shared elements. Unfortunately, under such a design an extra inductor was added to the load to introduce the notch. As a consequence, it still required 4 inductors at load to form a differential architecture.

It is thus necessary to provide a novel dual-band mixer wherein elements may be shared by two bands and no additional inductor will be needed.

It is also necessary to provide a dual-band mixer wherein working frequency bands may be easily selected.

The objective of this invention is to provide a novel dual-band mixer.

Another objective of this invention is to provide a dual-band mixer wherein elements may be shared by two bands without the need of additional inductor.

Another objective of this invention is to provide a dual-band mixer wherein working frequency bands may be easily selected.

Another objective of this invention is to provide a low-cost and small-size dual-band mixer.

Another objective of this invention is to provide a method to implement the design of dual-band mixers.

According to this invention, a dual-band mixer is disclosed. In the dual-band mixer of this invention, a current combined load is presented and is shared by two separate working frequency bands. In the invented dual-band mixer, a switch is provided to connect and disconnect an adjust capacitor series to the load inductors. By determining the capacitance of respective capacitors and the inductance of the load inductors, it is possible to generate resonance with related parasitic capacitances at particular frequencies, such that value of the load inductors may be changed. This enables the invented mixer to work with selected frequencies according to the operation of the switches. In addition, a systematic methodology is proposed to implement the design of the invented dual-band mixer. The efficient design method is approved by a 2.4/5.2-GHz CMOS mixer implementation.

These and other objectives and advantages of this invention may be clearly understood from the detailed description by referring to the following drawings.

FIG. 1 shows the circuit diagram of an embodiment of the dual-band mixer of this invention.

FIG. 2 shows the equivalent circuit of the load of the dual-band mixer of FIG. 1 , when working at the higher frequency band.

FIG. 3 shows the current combine operation of the dual band mixer at the higher working frequency.

FIG. 4 shows the equivalent circuit of the load of the dual-band mixer of FIG. 1 , when working at the lower frequency band.

FIG. 5 shows the flowchart of the method for designing a dual-band mixer of this invention.

FIG. 6 shows the return losses of one embodiment of the invented dual-band mixer under the frequencies of 5.2 GHz and 2.2 GHz.

FIG. 7 shows the waveform of output power of one embodiment of the invented dual-band mixer under a lower working frequency.

FIG. 8 shows the waveform of output power of one embodiment of the invented dual-band mixer under a higher working frequency.

It is known that most radio frequency CMOS mixers are implemented using the popular Gilbert-type double balanced active mixer. In the following description, the dual-band mixer of this invention will be described using the Gilbert-type mixer as an example. It should be understood that the dual-band mixer of this invention may be applied in mixers of other architectures or designs.

Dual-Band Mixer

A dual-band mixer of the present invention will be described hereinafter. FIG. 1 shows the circuit diagram of an embodiment of the dual-band mixer of this invention. As shown in this figure, the mixer is a Gilbert-type double balanced active mixer, which is sometimes called “Gilbert cell” or “Gilbert multiplier) and comprises a first differential amplifier and two second differential amplifiers connected in series to the first differential amplifier. The two inputs of the first differential amplifier are received radio frequency signals and the two inputs of both second differential amplifiers are local oscillator signals. The received radio frequency signals are thus mixed with the local oscillator signals to generate intermediate frequency signals. As the Gilbert-type double balanced active mixer is known to those skilled in the art, detailed description is thus omitted.

Now refer to FIG. 1 . As shown in this figure the dual band mixer of this embodiment has at its output: 2 load inductors Lload, Lload, a main capacitor C 1 to connect the load inductors Lload, Lload, 2 adjust capacitors C 2 , C 3 connected in parallel with main capacitor C 1 and connected/disconnected by transistor 10 as switch. 2 transistors M 1 and M 2 controlled by bias circuit 11 provide the tail current. 2 transistors M 3 and M 4 as input transconductors to convert differential input base band signals into current. Transistors M 5 -M 8 work as the switch stage; their inputs are local oscillator signals LO. Transistors M 9 -M 10 provides resistive degeneration to suppress third-order distortion.

2 Load inductors Lload, Lload function as load stage to increase voltage headroom and to minimize process variation for the purpose of integration. The 2 inductors Lload, Lload are recommended to have about the same inductance. Of course, it is possible to use 2 inductors with different inductances and number of inductors is not limited to 2. It is also possible to use 1 or more than 2 load inductors to function as load. The main capacitor C 1 and the adjust capacitors C 2 and C 3 are used to perform current combination, so that differential output of the mixer may be converted into the single-ended form. Main capacitor C 1 and 2 adjust capacitors C 2 , C 3 may have about the same capacitance. Of course, this is not any limitation to this invention. To switch the working frequency band of the mixer, transistor 10 is used as switch to change the capacitance for current combination by connecting or disconnecting adjust capacitors C 2 , C 3 in parallel with the main capacitor C 1 . In the embodiment of this invention, the switch may be a large-dimension PMOS transistor. It is understood that other types of transistors or switches may also be used in this invention to determine connection of adjust capacitors C 2 , C 3 with main capacitor C 1 .

After having optimized the relation between linearity and gain of the mixer, form factors of the MOS transistors and parasitic capacitance at drain nodes of transistors M 5 , M 8 at the switch stage may be fixed. The parasitic capacitance at the drain nodes of the switch stage is herein defined as Cp, which shall be in a range to enable the parasitic capacitance to resonate with load inductors Lload, Lload.

Working with the Higher Frequency Band

For a mixer as shown in FIG. 1 to perform over dual frequency bands, some conditions shall be met. In the following description the first operating frequency, f 1 , is defined as the higher frequency band and the second operating frequency, f 2 , the lower frequency band, whereby, f 1 >f 2 .

When operating at the higher frequency band f 1 , since the current combining capacitance required at this band is much smaller in value than that of the lower frequency band, switch 10 is turned off and the adjust capacitors C 2 and C 3 are disconnected from the load. The load inductance Lload can be represented by an equivalent circuit consisted of two inductors L 1 and L 2 in parallel as:

The equivalent circuit of load of the dual-band mixer when working at band f 1 is shown in FIG. 2 . By selecting the value of the load inductor Lload, it is possible to make the device parasitic capacitance Cp to resonate with the equivalent inductor L 1 at the higher operating frequency band f 1 .

Both Cp and L 1 together act as short circuit when the mixer is operating at frequency band f 1 , if the following condition is met:

As shown in FIG. 2 , only the main capacitor C 1 at the load, along with the remaining equivalent inductor L 2 , performs the current combine operation. Due to the current combine operation, the main capacitor C 1 realigns the direction of the differential current to achieve differential to single transform, when the following condition is met, as show in FIG. 3 :

FIG. 3 illustrates steps of the current combine operation of one embodiment of this invention. As shown in this figure, at (a) one side of the current source and the parallel equivalent inductor L 2 can be converted to a voltage source and a series inductor with the same value. At (b), the main capacitor C 1 can be split into two capacitors with even capacitance, connected in series. In this embodiment, the main capacitor C 1 is chosen to meet equation (3). As a consequence, the series inductor and one of the series capacitors jointly act as a short circuit at the higher frequency band f 1 . The voltage source and series capacitor can be viewed back as a current source and a parallel capacitor. Inductor L 2 and series capacitor then act like an open at resonance. At the same time, the load network converts the differential output current of the mixer to a single-ended current.

Working with the Lower Frequency Band

On the other hand, while the mixer is set to the second operating frequency f 2 , the equivalent inductor pair L 2 shall resonate with device parasitic Cp. Thus, Cp and L 2 act jointly as a short circuit at frequency band f 2 , if the following condition is met:

Switch 10 is turned ON and adjust capacitors C 2 and C 3 are connected in parallel with main capacitor C 1 , to perform current combination at the lower frequency band f 2 . The simplified equivalent circuit of the load is shown in FIG. 4 . Together capacitors C 1 , C 2 and C 3 form greater capacitance at the load and realign the differential current to achieve differential to single transform with the equivalent inductor pair at the load in condition of:

The operation of current combiner at the lower frequency band is similar to the steps described in FIG. 3 .

The proposed mixer, therefore, can perform dual band operation by the same elements, especially the load. Only two load inductors are required in the proposed mixer to perform dual-band operation for a differential architecture. Compared to the conventional approach which has two mixers, one for each band, and totally four load inductors, two in each mixer, for a differential architecture, the number of load inductors of the proposed mixer is halved by sharing load elements at two bands. By simply setting the connection of the adjust capacitors, the load inductors can perform differential to single up-conversion and reduce parasitic effects at two different frequency bands.

Design Flow

Based on above design concepts, a design flow for dual band mixer is developed as shown in FIG. 5 . FIG. 5 shows the flowchart of the design of one embodiment of the invented dual-band mixer.

In designing the invented dual-band mixer, parameters for the circuitry as shown in FIG. 1 shall be determined first. The parameters include those for the load inductors Lload, the main capacitor C 1 , the adjust capacitors C 2 , C 3 and working frequencies f 1 and f 2 . As shown in FIG. 5 , when designing the dual band mixer of this invention, at 501 structure and parameters of the circuit device is first determined. Obviously, the structure of the dual band mixer is as that of FIG. 1 . At 502 modeling software kits are used to calculate the parasitic capacitance Cp at the drain node of the switch stage. At 503 the equivalent inductance L 1 required to resonate with Parasitic capacitance Cp may be obtained after optimizing the linearity and the gain. At 504 inductance of the other equivalent inductor L 2 is known from equations (3) and (5). At 505 value of the load inductors Lload, which is equivalent to L 1 and L 2 in parallel, is obtained.

At the higher frequency band f 1 , the adjust capacitors C 2 and C 3 are switched off and the main capacitor C 1 solely performs current combination. The required combiner capacitance to resonate with equivalent inductor L 2 is obtained by equation (4) at 506 .

All capacitors C 1 , C 2 and C 3 concurrently perform current combination at the lower frequency band f 2 . The required capacitance for current combination at frequency band f 2 is given by equation (6) at 507 :

Herein, parameter k=f 1 /f 2 , is the ratio of the two operating frequencies.

By following above procedure, the design of the load including parasitic effects for dual-band mixer is expeditiously accomplished. All the parameters of the load are presented to be in relation with parasitic capacitance and operating frequencies. At 508 compare the results with available cell library, and modify the components if required. An automatic net list generating program can be used to generate a circuit with obtained components, which can be easily integrated with commercial EDA tools to perform circuit simulation, layout, design value check etc, and post-layout simulation. At 509 fine tuning for switch effect is completed to assure design performances.

The proposed schematic design methodology is applied to 2.4/5.2-GHz CMOS dual-band mixer design. The dual-band mixer is designed to operate with a 1.8-V power supply by using 0.18 μm CMOS technology. Values of related parameters of the dual-band mixer are shown in Table I. The smallest feature size of the available capacitor library provided from the manufacturer, however, is 100 fF. The main capacitor size is thus modified to fit the cell library and the higher frequency band of the implemented mixer is therefore lowered. The measured return losses of the output port for operating at 5.2-GHz and 2.4-GHz are shown in FIG. 6 , wherein (a) represents return loss at 5.2 GHz working frequency band and (b) at 2.2 GHz working frequency band. FIG. 6 shows that the return loss is lower than −10 dB in both bands of interest.

The output power is measured with a 10-MHz baseband signal. As shown in FIG. 7 , the output power for the lower frequency band peaks at 2.46-GHz obtained by a 2.45-GHz LO signal and meets the design specifications. The output power for the higher frequency band as shown in FIG. 8 , however, peaks at 5.19-GHz due to the main capacitor being modified.

As the present invention has been shown and described with reference to preferred embodiments thereof, those skilled in the art will recognize that the above and other changes may be made therein without departing from the spirit and scope of the invention.