KR20100062444A - Mixer - Google Patents

Abstract

PURPOSE: A frequency mixer is provided to minimize parasitic capacitance by interlinking the ends of a first resonance inductor and a second resonance inductor between a transconductance unit and a switching unit, and also connecting another end thereof to a grounding terminal. CONSTITUTION: A transconductance unit(310) amplifies inputted RF(Radio Frequency) signals. A switching unit(320) generates BB(Baseband) signals by temporarily switching the amplified RF signals according to an inputted LO(Local Oscillator) signal. An inductor is electrically interlinked between the transconductance unit and the switching unit. The inductor is comprised of a bonding wire which has a different inductance value depending on the length. The inductor is a variable inductor which selectively adjusts the inductance value.

Description

Frequency Mixer {MIXER}

The present invention relates to a frequency mixer, and more particularly, to a frequency mixer used in a direct conversion RF receiver.

In general, the frequency mixer refers to a circuit for converting an input signal into a signal of a desired frequency band. Such frequency mixers are not only used in transmitters and receivers in communication systems, but also widely applied in other areas.

In a typical receiver, the frequency mixer downconverts the incoming signal. There are two types of downconversion methods: heterodyne and direct conversion. The heterodyne method converts an RF signal into an IF signal and then converts an IF signal into a baseband signal, and a direct conversion method converts an RF signal directly into a baseband signal. Although both methods have advantages and disadvantages, the heterodyne method requires the use of a filter in each process of converting an RF signal into an IF signal and then converting the IF signal into a baseband signal. Commonly used filters are surface acoustic wave (SAW) filters or ceramic filters. These SAW filters or ceramic filters are manufactured in off-chip form. Therefore, there are many constraints on integration of the receiver. Therefore, the direct conversion method is used more than the heterodyne method.

Then, the structure of the reception of the direct conversion method will be described with reference to FIG.

1 is a diagram illustrating a receiver of a general direct conversion method.

Referring to FIG. 1, when receiving one or more RF signals through the antenna 110, the received RF signals pass through the bandpass filter 120. The RF signals passing through the bandpass filter 120 are provided to the low noise amplifier 130, and the low noise amplifier 130 amplifies the RF signals. The amplified RF signals are provided to the frequency mixer 140, and the frequency mixer 140 mixes the LO signal provided by the local oscillator 150 and the RF signal passed through the low noise amplifier 130 to baseband (BB). Output the signal. The output baseband signal finally passes through the low pass filter 160. Here, the frequency mixer 130 is generally implemented as a complementary metal-oxide semiconductor (hereinafter referred to as "CMOS").

)는 아래의 <수학식 1>과 같이 예측할 수 있다.The direct conversion receiver can eliminate off-chip elements required by the heterodyne scheme, and thus, the receiver can be easily integrated. However, the direct conversion method has problems such as DC offset, second order distortion (IM2), self mixing, and flicker noise. Among these problems, flicker noise is inherent in active devices, especially in the low frequency region. This flicker noise is seen as a bigger problem for CMOS devices than for Bipolar Junction Transistor (BJT) devices. This is because the flicker noise is due to the physical properties of the semiconductor device and is a noise generated by immobilizing electrons moving through the channel formed between the drain and the source of the CMOS transistor. Since the flicker noise is randomly distributed in the time when the electron does not move in the channel, it is impossible to generalize and predict the flicker noise characteristic. However, current density due to flicker noise in the saturation region of CMOS transistor

) Can be predicted as in Equation 1 below.

는 CMOS 트랜지스터의 포화영역에서의 플리커 잡음에 의한 전류 밀도이고,

는 CMOS 트랜지스터의 옥사이드 트랩의 밀도이고,

는 동작 주파수이다. 여기서,

는 CMOS 트랜지스터의 게이트 단자에 인가되는 전압에 비례하므로,

는 CMOS 트랜지스터의 게이트 단자에 인가되는 전압에 비례한다.In Equation 1,

Is the current density due to flicker noise in the saturated region of the CMOS transistor,

Is the density of the oxide traps in the CMOS transistors,

Is the operating frequency. here,

Is proportional to the voltage applied to the gate terminal of the CMOS transistor,

Is proportional to the voltage applied to the gate terminal of the CMOS transistor.

As such, the noise characteristics of the frequency mixer using the CMOS transistor are mainly influenced by the flicker noise characteristics, which can be divided into the effects of the direct mechanism and the indirect mechanism. In order to specifically explain the effects of these direct and indirect mechanisms, we will look at commonly used frequency mixers.

2 is a circuit diagram of a general frequency mixer.

Referring to FIG. 2, a general frequency mixer includes a transconductance unit 210, a switching unit 220, and an output unit 230.

The transconductance unit 210 includes a pair of amplification elements M1 and M2, amplifies an RF + signal input to the gate terminal of M1, and amplifies an RF- signal input to the gate terminal of M2. Here, the RF + signal and the RF- signal have a phase difference of 180 degrees, and M1 and M2 are implemented as n-channel MOSFET devices.

The switching unit 220 includes a pair of M3 and M4 for switching the output current of M1, and a pair of M5 and M6 for switching the output current of M2. The LO + signal is input from the local oscillator to the terminals of the gates of M3 and M6, and the LO- signal is input to the common gate terminal of the M4 and M5. Here, the LO + signal and the LO- signal have a phase difference of 180 degrees. Meanwhile, the drain terminal of M4 and the drain terminal of M6 are connected to each other, and the drain terminal of M3 and the drain terminal of M5 are connected to each other. The switching unit 220 mixes the RF signal and the LO signal amplified by M1 and M2 and outputs a BB signal corresponding to the difference between the frequencies of the two signals. Here, M3, M4, M5, and M6 are implemented with n-channel MOSFET devices.

The output unit 230 outputs an IF signal output from the switching unit 220.

The frequency mixer amplifies the RF + signal and the RF- signal differentially input to the transconductance unit 210, and converts the amplified RF + signal and the RF- signal to the LO + signal and the LO- signal differentially input to the switching unit 220. Each of them is mixed to output the differential BB + and BB- signals to the output 230.

Next, the direct noise mechanism and the indirect noise mechanism of the flicker noise will be described with reference to the general frequency mixer shown in FIG. 2.

)는 <수학식 2>와 같다.The direct noise mechanism is due to the switching time of the transistors M3, M4, M5, M6 of the switching unit 220. Specifically, in the general frequency mixer shown in Fig. 2, the switching of the transistors M3, M4, M5, M6 should ideally take place in a very short time. However, since the LO signal input from the local oscillator to the frequency mixer is generally not an ideal square wave, the ratio between the voltage of the input LO signal and the output currents of the transistors M3, M4, M5, and M6 has a finite value. . If the ratio between the voltage of the LO signal and the output currents of the transistors M3, M4, M5, and M6 becomes infinity, the transistors M3, M4, M5, and M6 are turned on / off without switching time. It acts as an ideal switch to operate. Therefore, the faster the switching time, the more the direct noise mechanism can be minimized. Therefore, it is desirable to minimize the overdrive voltage (Vgs-Vth) for turning on the transistors M3, M4, M5, and M6. Do. The output noise current caused by this direct noise mechanism (

) Is the same as Equation 2.

는 LO 신호의 주기(

)이고,

는 스위칭 시간(

)동안의 LO 신호의 전압 기울기(

)이다.In Equation 2,

Is the period of the LO signal (

)ego,

Is the switching time (

Slope of the LO signal during

)to be.

Therefore, in order to improve flicker noise by the direct noise mechanism, the voltage of the LO signal may be applied to the switching unit 220 in the form of a perfect square wave. However, the flicker noise characteristic still appears in the signal output from the output unit 230. This is flicker noise caused by the indirect noise mechanism of the frequency mixer, and is caused by the parasitic capacitance of the switching unit 220.

Accordingly, the present invention provides a frequency mixer with improved noise characteristics.

In addition, the present invention provides a frequency mixer that can minimize the flicker noise by the indirect noise mechanism.

The apparatus according to the present invention, as a frequency mixer, is a transconductance unit for amplifying the input RF (Radio Frequency) signals, and amplified RF signals by switching in time according to the local oscillator (LO) signal is input to the baseband (BB) And a switching unit for generating signals and an output unit for outputting the generated BB signals, and an inductor electrically connected between the transconductance unit and the switching unit.

In addition, the inductor of the frequency mixer according to the present invention is preferably implemented by bonding wires having different inductance values depending on the length.

In addition, the inductor of the frequency mixer according to the present invention is preferably implemented as a variable inductor that can selectively adjust the inductance value.

The present invention has the advantage of providing a frequency mixer with improved noise characteristics. In addition, the present invention has the advantage of providing a frequency mixer that can minimize the flicker noise by the indirect noise mechanism.

Hereinafter, the present invention will be described with reference to the accompanying drawings. In the following description of the present invention, a part obvious to those skilled in the art will be omitted so as not to disturb the gist of the present invention. In addition, it is to be noted that each of the terms described below are only used to help the understanding of the present invention, and may be used in different terms despite the same purpose in each manufacturing company or research group.

3 is a circuit diagram of a frequency mixer according to a preferred embodiment of the present invention. The frequency mixer according to the preferred embodiment of the present invention shown in FIG. 3 is implemented using a double balance structure. The dual balance structure improves signal isolation between the LO signal and the RF signal input to the frequency mixer, and prevents the LO signal from flowing into the output unit 330. Here, although the frequency mixer of the present invention is implemented by the dual balanced structure, it should be noted that the idea of the present invention can be applied to the frequency mixer of the single balanced structure.

Referring to FIG. 3, a frequency mixer according to a preferred embodiment of the present invention includes a transconductance unit 310, a switching unit 320, and an output unit 330.

The transconductance unit 310 includes a pair of M1, M2, a first resonant inductor L1, and a second resonant inductor L2. Although two capacitors Cp1 and Cp2 are shown in FIG. 3, it should be noted that these capacitors are not parasitic capacitances actually implemented on the circuit, but parasitic capacitances on the circuit. The source terminals of M1 and M2 are electrically connected to each other, and the drain terminal of M1 is electrically connected to the common source terminal of M3 and M4 of the switching unit 320. The drain terminal of M2 is electrically connected to the common source terminal of M5 and M6 of the switching unit 320. One end of the first resonant inductor L1 is electrically connected between the drain terminal of M1 and the common source terminal of M3 and M4, and the other end is electrically connected to the ground terminal. One end of the second resonant inductor L2 is electrically connected between the drain terminal of M2 and the common source terminal of M5 and M6, and the other end is electrically connected to the ground terminal. Here, it is preferable that the first resonant inductor L1 and the second resonant inductor L2 are connected by a bonding wire technique.

The transconductance unit 310 amplifies the RF + signal input to the gate terminal of M1 and amplifies the RF- signal input to the gate terminal of M2. Here, the RF + signal and the RF- signal have a phase difference of 180 degrees, and M1 and M2 are implemented as n-channel MOSFET devices. The first resonant inductor L1 removes the first parasitic capacitance Cp1 existing between the common source terminal of M3 and M4 and the drain terminal of M1. Similarly, the second resonant inductor L2 removes the second parasitic capacitance Cp2 existing between the common source terminal of M5 and M6 and the drain terminal of M2.

The switching unit 320 includes a pair of M3 and M4 for switching the output current of M1, and a pair of M5 and M6 for switching the output current of M2. The LO + signal is input from the local oscillator to the terminals of the gates of M3 and M6, and the LO- signal is input to the common gate terminal of the M4 and M5. Here, the LO + signal and the LO- signal have a phase difference of 180 degrees. Here, M3, M4, M5, and M6 are implemented with n-channel MOSFET devices.

The switching unit 320 outputs a BB + signal corresponding to the difference between the frequencies of the two signals by mixing the RF + signal amplified by M1 and the LO + signal having the same frequency as the carrier frequency of the RF + signal. In addition, the RF-signal amplified by M2 and the LO-signal having the same frequency as the carrier frequency of the RF-signal are mixed to output a BB-signal corresponding to the difference between the frequencies of the two signals.

The output terminal 330 outputs the BB + signal and the BB- signal output from the switching unit 320.

As such, when the frequency mixer according to the preferred embodiment of the present invention shown in FIG. 3 is compared with the general Gilbert cell frequency mixer shown in FIG. 2, the frequency mixer according to the preferred embodiment of the present invention is used to remove parasitic capacitance. The apparatus further includes a first resonant inductor L1 and a second resonant inductor L2.

The frequency mixer according to the preferred embodiment of the present invention can effectively improve the flicker noise characteristic, and the reason thereof will be described below.

The direct noise mechanism of the flicker noise characteristic is solved by inputting the LO signal close to the ideal square wave to the switching unit 320. In this way, when an LO signal close to a perfect square wave is applied to each gate terminal of M3, M4, M5, and M6, M3, M4, M5, and M6 repeat on / off operations during the half period of the square wave.

However, since the indirect noise mechanism is due to parasitic capacitances Cp1 and Cp2 as described above, in order to eliminate such parasitic capacitances, the frequency mixer according to the preferred embodiment of the present invention has a first resonance inductor L1 and a second resonance. It further includes an inductor L2. To describe this in detail, reference is made to FIGS. 4 and 5.

4 and 5 are circuit diagrams for describing flicker noise generated in a switching unit of a general Gilbert cell frequency mixer.

As shown in FIGS. 4 and 5, the flicker noise of each of the transistors M3, M4, M5, and M6 of the switching unit 220 of the general Gilbert cell frequency mixer shown in FIG. 2 is modeled as Vn + and Vn−. can do. Here, for convenience of description, M3 and M6 of the switching unit 220 are represented by M + as shown in FIG. 4, and M4 and M5 are represented by M− as shown in FIG. 5. In accordance with the on / off operation of the LO + signal applied to the gate terminal of M +, Vn + of M + charges the parasitic capacitance Cp during the half cycle, and discharges the parasitic capacitance Cp for the next half cycle. This causes parasitic capacitance current (+ icp). Similarly, according to the on / off operation of the LO- signal applied to the gate terminal of M-, Vn- of M- charges the parasitic capacitance Cp for half a period, and the parasitic capacitance Cp for the next half period. Discharge. This causes parasitic capacitance current (-icp).

)를 수식화하면 <수학식 3>과 같다.Then, as shown in FIG. 6A, parasitic capacitance currents generated from flicker noises Vn + and Vn- may be represented by + icp and -icp as shown in FIG. 6B, and thus a general Gilbert cell. The differential output of the frequency mixer is shown in FIG. In this general Gilbert cell frequency mixer, the output current due to flicker noise according to the indirect noise mechanism of each of the switching units 220, M3, M4, M5, and M6 (

) Is formulated as shown in <Equation 3>.

Then, the output current due to the total flicker noise by the direct noise mechanism and the indirect noise mechanism of the general Gilbert cell frequency mixer is expressed by Equation 4.

는 스위칭부(220) M3, M4, M5, M5 각각의 트랜스컨덕턴스이고,

는 스위칭부(220) M3, M4, M5, M5 각각의 게이트 단자에 인가되는 LO 신호의 주파수이다.In <Equation 4>

Is the transconductance of each of the switching unit 220, M3, M4, M5, M5,

Is the frequency of the LO signal applied to the gate terminal of each of the switching units 220, M3, M4, M5, and M5.

)에 크게 영향을 받는다.As shown in Equation 3, the output current due to flicker noise of a typical Gilbert cell frequency mixer is determined by the parasitic capacitance (Cp) and the frequency of the LO signal.

) Is greatly affected.

Then, modeling the flicker noise (Vn) is shown in equation (5).

는 M3, M4, M5, M6 각각의 유효 채널 폭이고,

는 M3, M4, M5, M6 각각의 유효 길이이고,

는 옥사이드 커패시턴스이고,

는 주파수이고,

는 일반 상수이다.In Equation 5,

Is the effective channel width of each of M3, M4, M5, and M6,

Is the effective length of each of M3, M4, M5, M6,

Is the oxide capacitance,

Is the frequency,

Is a general constant.

According to Equation 5, the flicker noise Vn is inversely proportional to the area of the gate terminal of each of M3, M4, M5, and M6. Therefore, in order to minimize flicker noise, the area of the gate terminal of each of M3, M4, M5, and M6 should be increased. However, when the area of the gate terminal of each of the M3, M4, M5, and M6 increases, the parasitic capacitance that causes flicker noise increases. Since this amplifies the flicker noise by the indirect noise mechanism, the noise figure cannot be increased by adjusting the effective channel width and effective length of each of the M3, M4, M5, and M6.

Accordingly, the frequency mixer according to the preferred embodiment of the present invention shown in FIG. 3 connects one end of the first resonant inductor L1 and the second resonant inductor L2 between the transconductance unit 310 and the switching unit 320. In addition, by connecting the other end with the ground terminal, it is possible to minimize the parasitic capacitance.

Meanwhile, when the frequency mixer according to the preferred embodiment of the present invention illustrated in FIG. 3 is implemented as an integrated circuit, the first resonant inductor L1 and the second resonant inductor L2 may be implemented as chip inductors. Is difficult. This is because, in general, the chip inductor is fixed to a specific value during fabrication, and the parasitic capacitance value generated in the switching part of the frequency mixer is randomly generated. Accordingly, in order to implement the first resonant inductor L1 and the second resonant inductor L2 as chip inductors, the inductance value may be adjusted after fabrication of the integrated circuit of the frequency mixer, and the first inductor L1 may be adjusted for such adjustment. And the second inductor L2 are preferably manufactured by a bonding wire. When the first inductor L1 and the second inductor L2 are manufactured by the bonding wires, the bonding wires generally have an inductance value of 0.8 [nH] per 1 mm. The inductance values of the first inductor L1 and the second inductor L2 may be adjusted to minimize the capacitance.

In addition, when the first inductor L1 and the second inductor L2 are manufactured as chip inductors, the quality factor is 10 or less. However, when the first inductor L1 and the second inductor L2 are manufactured as bonding wires, the quality factor is used. Up to 20 and up to 80 can be guaranteed.

In addition, when the first resonant inductor L1 and the second resonant inductor L2 are implemented as chip inductors, the size of the chip inductor is the largest among the devices used in the integrated circuit. However, when the first inductor L1 and the second inductor L2 are manufactured using the bonding wires, the noise figure may be reduced at the same time without affecting the entire chip area.

The present invention described above is capable of various substitutions, modifications, and changes without departing from the technical spirit of the present invention for those skilled in the art to which the present invention pertains. It is not limited by the drawings.

1 is a simplified configuration diagram of a wireless communication system direct conversion receiver according to the present invention;

2 is a block diagram of a wireless communication system frequency mixer according to the present invention;

3 is a frequency mixer switching transistor noise voltage and parasitic capacitance according to the present invention,

4 and 5 are circuit diagrams for explaining flicker noise generated in a switching unit of a general Gilbert cell frequency mixer.

6 is a graph showing flicker noise (Vn +, Vn−), parasitic capacitance current and differential output of a typical Gilbert cell frequency mixer.

Claims (3)

In the frequency mixer, A transconductance unit for amplifying input RF (Radio Frequency) signals; A switching unit which generates baseband (BB) signals by temporally switching the amplified RF signals according to a local oscillator (LO) signal input thereto; Including an output unit for outputting the generated BB signals, And an inductor electrically connected between the transconductance section and the switching section. The method of claim 1, wherein the inductor, A frequency mixer implemented with bonding wires whose inductance values vary with length. The method of claim 1, wherein the inductor, Frequency mixer implemented as a variable inductor with optional inductance value adjustment.

Images