Title of Invention	AN INTEGRATED CIRCUIT,AN AMPLIFIER AND A RECEIVER FOR A WIRELESS DEVICE
Abstract	An amplifier, which has good linearity and noise performance, includes first, second, third, and fourth transistors and an inductor. The first and second transistors are coupled as a first cascode pair, and the third and fourth transistors are coupled as a second cascode pair. The third transistor has its gate coupled to the source of the second transistor, and the fourth transistor has its drain coupled to the drain of the second transistor. The first transistor provides signal amplification. The second transistor provides load isolation and generates an intermediate signal for the third transistor. The third transistor generates distortion components used to cancel third order distortion component generated by the first transistor. The inductor provides source degeneration for the first transistor and improves distortion cancellation. The sizes of the second and third transistors are selected to reduce gain loss and achieve good linearity for the amplifier.

Title of Invention

AN INTEGRATED CIRCUIT,AN AMPLIFIER AND A RECEIVER FOR A WIRELESS DEVICE

Abstract

An amplifier, which has good linearity and noise performance, includes first, second, third, and fourth transistors and an inductor. The first and second transistors are coupled as a first cascode pair, and the third and fourth transistors are coupled as a second cascode pair. The third transistor has its gate coupled to the source of the second transistor, and the fourth transistor has its drain coupled to the drain of the second transistor. The first transistor provides signal amplification. The second transistor provides load isolation and generates an intermediate signal for the third transistor. The third transistor generates distortion components used to cancel third order distortion component generated by the first transistor. The inductor provides source degeneration for the first transistor and improves distortion cancellation. The sizes of the second and third transistors are selected to reduce gain loss and achieve good linearity for the amplifier.

Full Text	FORM 2 THE PATENTS ACT, 1970 (39 of 1970) & THE PATENTS RULES, 2003 COMPLETE SPECIFICATION (See section 10, rule 13) 'AMPLIFIER WITH ACTIVE POST-DISTORTION LINEARIZATION" QUALCOMM INCORPORATED, a company incorporated in the state of Delaware, U.S., of 5775 Morehouse Drive, San Diego, California 92121-1714, U.S.A The following specification particularly describes the invention and the manner in which it is to be performed. WO 2007/016552 PCT/US2006/029905 AMPLIFIER WITH ACTIVE POST-DISTORTION LINEARIZATION [0001] This application claims the benefit of provisional U.S. Application Serial No. 60/705,256, entitled "Linearity improvement technique for CMOS amplifiers from low frequency to high frequency by using vectorized post-distortion," filed August 2, 2005. BACKGROUND I. Field [0002] The present disclosure relates generally to circuits, and more specifically to an amplifier suitable for wireless communication and other applications. II. Background [0003] Amplifiers are commonly used in various electronics devices to provide signal amplification. Furthermore, different types of amplifiers are available for different uses. For example, a wireless device may include a transmitter and a receiver for bi-directional communication, and the transmitter may utilize a power amplifier (PA) and the receiver may utilize a low noise amplifier (LNA) and a variable gain amplifier (VGA). [0004] An LNA is commonly used in a receiver to amplify a low-amplitude signal received via a communication channel. The LNA is often the first active circuit encountered by the received signal and hence has a large impact on the performance of the receiver in several key areas. First, the LNA has a large influence on the overall noise figure of the receiver since the noise of the LNA is injected directly into the received signal and the noise of subsequent stages is effectively reduced by the gain of the LNA. Second, the linearity of the LNA has a large influence on both the design of subsequent stages in the receiver and the receiver performance. The LNA input signal typically includes various undesired signal components that may come from external interfering sources and leakage from a co-located transmitter. Nonlinearity in the LNA causes the undesired signal components to mix and generate cross modulation distortion (XMD) that may fall within the desired signal bandwidth. The amplitude of the cross WO 2007/016552 PCT/US2006/029905 modulation distortion is determined by the amount of nonlinearity in the LNA. The cross, modulation distortion component that falls within the desired signal bandwidth acts as noise that degrades the signal-to-noise ratio (SNR) of .the desired signal. The degradation in SNR caused by LNA nonlinearity impacts the design of (and often places more stringent requirements on) subsequent stages in order to meet the overall SNR specification for the receiver. Therefore, having a more linear LNA can alleviate the performance requirements for other stages, which may result in lower power consumption and smaller circuit area for the receiver. [0005] There is therefore a need in the art for an amplifier having good linearity and noise performance. SUMMARY [0006] Various embodiments of an amplifier linearized using active post-distortion (APD) are described herein. The amplifier is simple in design, has good linearity and noise performance, and is suitable for wireless communication and other high frequency applications. For example, the amplifier may be used as an LNA for a receiver in a wireless device. Active post-distortion may also be used to linearize other active circuits such as, e.g., a mixer. [0007] In an embodiment, an amplifier (e.g., an LNA) includes first, second, third, and fourth transistors (e.g., N-FETs) and an inductor. The first and second transistors are coupled as a first cascode pair, and the third and fourth transistors are coupled as a second cascode pair. The first transistor has its source coupled to the inductor and its gate receiving an input (voltage) signal. The second transistor has its source coupled to the drain of the first transistor and its drain providing an output (current) signal. The third transistor has its gate coupled to the source of the second transistor. The fourth transistor has its source coupled to the drain of the third transistor and its drain coupled to the drain of the second transistor. The first transistor provides signal amplification. The second transistor provides load isolation and further generates an intermediate signal for the third transistor. The third transistor receives the intermediate signal and generates distortion components used to cancel third order distortion component t generated by the first transistor. The fourth transistor provides load isolation. The inductor provides source degeneration for the first transistor and improves the cancellation of the third order distortion. In other embodiments, the fourth transistor WO 2007/016552 PCT/US2006/029905 may be omitted, and the drain of the third transistor may be coupled to the drain of either the first or second transistor. The sizes of the second and third transistors may be selected to reduce gain loss for the amplifier and to cancel as much third order distortion as possible. [0008] Various aspects and embodiments of the invention are described in further detail below. BRIEF DESCRIPTION OF THE DRAWINGS [0009] The features and nature of the present invention will become more apparent from the detailed description set forth below when taken in conjunction with the drawings in which like reference characters identify correspondingly throughout. [0010] FIG. 1 shows a radio frequency (RF) portion of a wireless device. [0011] FIGS. 2A, 2B and 2C show a received signal from an antenna, an LNA input signal, and an LNA output signal, respectively. [0012] FIG. 3 shows a schematic diagram of an LNA with active post-distortion linearization. [0013] FIGS. 4A and 4B show plots of IIP3 for the LNA for low and high frequencies, respectively. [0014] FIG. 5 shows an equivalent circuit for the LNA. [0015] FIG. 6 shows a vector diagram illustrating active post-distortion cancellation. [0016] FIGS. 7A and 7B show schematic diagrams of two additional embodiments of an LNA with active post-distortion linearization. [0017] FIG. 8 shows a schematic diagram of an LNA with active post-distortion linearization and multiple gain settings. [0018] FIG. 9 shows a schematic diagram of an LNA implemented with P-FETs. DETAILED DESCRIPTION [0019] The word "exemplary" is used herein to mean "serving as an example, instance, or illustration." Any embodiment or design described herein as "exemplary" is not necessarily to be construed as preferred or advantageous over other embodiments or designs. [0020] The amplifier and other linearized active circuits described herein may be used for various applications such as communication, networking, computing, consumer WO 2007/016552 PCT/US2006/029905 electronics, and so on. These linearized active circuits may be used in wireless communication systems such as a Code Division Multiple Access (CDMA) system, a Time Division Multiple Access (TDMA) system, a Global System for Mobile Communications (GSM) system, an Advanced Mobile Phone System (AMPS) system, Global Positioning System (GPS), a multiple-input multiple-output (MIMO) system, an orthogonal frequency division multiplexing (OFDM) system, an orthogonal frequency division multiple access (OFDMA) system, a single-carrier FDMA (SC-FDMA) system, a wireless local area network (WLAN), and so on. The amplifier may be used as an LNA, a VGA, a PA, and so on. For clarity, an LNA used in a receiver of a wireless device for a CDMA system is described below. The CDMA system may implement cdma2000, Wideband CDMA (W-CDMA), and/or other CDMA radio access technologies. [0021] FIG. 1 shows a block diagram of a radio frequency (RF) portion of a wireless device 100. Wireless device 100 may be a cellular phone, a personal digital assistant (PDA), a wireless modem card, or some other device used for wireless communication. Wireless device 100 includes a transmitter and a receiver that provide bi-directional communication. [0022] On the transmit path, a power amplifier (PA) 110 receives and amplifies a transmit (TX) modulated signal and provides a transmit signal. The transmit signal is routed through a duplexer 120 and transmitted via an antenna 130 to one or more serving base stations. A portion of the transmit signal also couples or leaks through duplexer 120 to the receive path. The amount of TX leakage is dependent on the isolation between the transmit and receive ports of duplexer 120, which may be approximately 50 decibels (dB) for a surface acoustic wave (SAW) duplexer at cellular band. A lower TX-RX isolation results in higher level of TX leakage. [0023] On the receive path, a received signal containing a desired signal and possibly a jammer is received via antenna 130, routed through duplexer 120, and provided to an LNA 140. LNA 140 also receives a TX leakage signal from the transmit path. The input signal at the input of LNA 140 may thus include the desired signal, the TX leakage signal, and the jammer. LNA amplifies the input signal and provides an amplified RF signal. A SAW filter 150 filters me amplified RF signal to remove out-of-band components (e.g., the TX leakage signal) and provides a filtered RF signal. A WO 2007/016552 PCT/US2006/029905 mixer 160 frequency downconverts the filtered RF signal with a local oscillator (LO) signal and provides a downconverted signal. [0024] FIG. 2A shows the received signal from antenna 130, which includes a desired signal 210 and a jammer 220. Jammer 220 is an undesired signal and may correspond to, for example, a signal transmitted by a nearby base station in an AMPS system. The jammer may be much higher in amplitude than the desired signal and may be located close in frequency to the desired signal. [0025] FIG. 2B shows the input signal at the input of LNA 140. The input signal contains desired signal 210 and jammer 220 in the received signal as well as a TX leakage signal 230 from the transmit path. The TX leakage signal may be large relative to the desired signal, especially if wireless device 100 is far from the serving base station(s) and needs to transmit at a high power level in order to reach the base station(s). [0026] FIG. 2C shows the signal at the output of LNA 140. Nonlinearity in LNA 140 can cause the modulation on TX leakage signal 230 to interact with narrowband jammer 220 and generate cross modulation distortion 240 around the jammer. A portion 250 of the cross modulation distortion, which is shown with shading, may fall within the desired signal band. Portion 250 acts as additional noise that degrades the performance of the receiver. This noise also degrades receiver sensitivity so that the smallest desired signal that can be reliably detected by the receiver needs to have a larger amplitude. [0027] FIG. 3 shows a schematic diagram of an embodiment of an LNA 140a with active post-distortion (APD) linearization. LNA 140a has good linearity and noise performance and may be used for LNA 140 in FIG. 1. LNA 140a includes four N-channel field effect transistors (N-FETs) 310, 320, 330 and 340, an inductor 350, and a capacitor 352. N-FET 310 has its source coupled to one end of inductor 350, its gate receiving an input voltage vi, and its drain coupled to the source of N-FET 320. The other end of inductor 350 couples to circuit ground. N-FET 320 has its gate receiving a bias voltage Vbias and its drain coupled to an output node. N-FET 330 has its source coupled to circuit ground, its gate coupled to one end of capacitor 352, and its drain coupled to the source of N-FET 340. The other end of capacitor 352 couples to the source of N-FET 320. N-FET 340 has its gate receiving the bias voltage vbias and its WO 2007/016552 PCT/US2006/029905 drain coupled to the output node. The output node provides an output current iout for LNA 140a. [0028] N-FETs 310 and 320 form a first cascode pair for a main signal path used for signal amplification. N-FET 310 provides signal amplification. N-FET 320 provides load isolation for N-FET 310 and further generates an intermediate voltage v2 for N-FET 330. N-FETs 330 and 340 form a second cascode pair for an auxiliary signal path that generates cross modulation distortion used for distortion cancellation. N-FET 330 generates the cross modulation distortion, and N-FET 340 provides load isolation for N-FET 330. Inductor 350 provides source degeneration and further provides a 50-ohm match looking into the gate of N-FET 310. Inductor 350 is also used for active post-distortion linearization and improves distortion cancellation. Capacitor 352 provides AC coupling. [0029] N-FET 310 has a small-signal transconductance of g1, which is determined by various factors such as the size (e.g., length and width) of N-FET 310, the bias current for N-FET 310, the gate-to-source voltage vgs of N-FET 310, and so on. N-FET 320 has a small-signal transconductance of g1/a , where a is the ratio of the transconductance of N-FET 310 to the transconductance of N-FET 320. The factor a is typically determined by the ratio of the width of N-FET 310 to the width of N-FET 320. N-FET 330 has a small-signal transconductance of gx Ib, where b is the ratio of the transconductance of N-FET 310 to the transconductance of N-FET 330. The factor a is typically determined by the ratio of the width of N-FET 310 to the width of N-FET 330. The factors a and /? may be selected as described below. [0030] Linearization of LNA 140a using active post-distortion may be achieved at low frequency as follows. At low frequency, inductor 350 does not come into play and is effectively shorted, and the input voltage vi is equal to the Vg, voltage for N-FET 310. The drain current i'i of N-FET 310 may be represented by a power series as: Eq(l) where g2 is a coefficient that defines the strength of second order nonlinearity; g3 is a coefficient that defines the strength of third order nonlinearity; and h(vg,) is the dram current of N-FET 310 as a function of VgS. WO 2007/016552 PCT/US2006/029905 For simplicity, nonlinearities higher than third order are ignored in equation (1). Coefficients g1, gz and g} are determined by the device size and the bias current for N-FET 310. Coefficient g% controls the third order intermodulation distortion (IMD3) at low signal level and hence determines the third order input intercept point (IIP3), which is a metric commonly used to specify the linearity of an amplifier. [0031] N-FET 320 may be assumed to be linear, hi this case, the drain voltage V2 of N-FET 310, which is also the vgs voltage for N-FET 330, may be expressed as: Eq(2) Equation (2) indicates that the V2 voltage generated by N-FET 320 is dependent on a. The drain current i3 of N-FET 330 may be represented by a power series as: Eq(3) Equation (3) indicates that the coefficients for N-FET 330 and the coefficients for N-FET 310 are related by p. [0032] Equation (2) may be substituted into equation (3) so that the drain current i3 of N-FET 330 can be expressed as a function of the drain current i\ of N-FET 310. Equation (1) may then be substituted into equation (3) so that the drain current 13 of N-FET 330 can be expressed as a function of the vgs voltage of N-FET 310. The expanded equation (3) includes multiple terms for each order of nonlinearity due to the interaction between the power series in equation (1) and the power series in equation (3). [0033] The drain currents of N-FETs 310 and 330 are combined to generate the output current i„a, as follows: Eq(4) where gxx and g3Z are the first and third order power series coefficients, respectively, for the output current iout and may be expressed as: WO 2007/016552 PCT/US2006/029905 The term g2J. in equation (4) may be ignored since only the fundamental frequency and the third order nonlinearity are of interest. [0034] Equation (5) represents an overall gain for LNA 140a and shows a gain loss resulting from the use of active post-distortion linearization. The overall gain for LNA 140a with distortion cancellation is giz whereas the gain for the LNA without distortion cancellation is g\. The gain loss of (1 - a/fi) is directly related to a and /3 and may be kept small by selecting /? to be large relative to a. A larger J3 leads to less gain loss but does not necessarily mean less distortion cancellation. Equation (6) represents the combined third order distortion in the output current iout. The first term in equation (6) represents the contribution from third order nonlinearity, and the second term in equation (6) represents the contribution from second order nonlinearity. [0035] FIG. 4A shows a plot 410 of HP3 for LNA 140a with distortion cancellation (with N-FETs 330 and 340 connected) and a plot 420 of IIP3 for LNA 140a without distortion cancellation (with N-FETs 330 and 340 omitted) at low frequency. For a given device width and power consumption, equation (6) may be solved such that the third order distortion component approaches zero. The value of J3 is selected to prevent excessive gain loss. For a specific exemplary design, fi is selected to be equal to 8, and a value of 1.35 for a provides good distortion cancellation. Because of the second order nonlinearity in equation (6), the distortion cancellation is dependent on bias voltage, which is the operating vgs voltage for N-FET 310. [0036] LNA 140a may be used for high frequency applications such as wireless communication. At high frequency, reactive elements such as capacitors and inductors affect linearity performance and further cause performance to be frequency dependent. [0037] FIG. 5 stiows a schematic diagram of a simplified equivalent circuit 500 for LNA 140a in FIG. 3. For the embodiment shown in FIG. 5, N-FETs 310, 320, 330 and 340 are modeled with ideal current sources 510, 520, 530 and 540, respectively, and parasitic gate-to-source capacitors 512, 522, 532 and 542, respectively. N-FETs 310, WO 2007/016552 PCT/US2006/029905 320, 330 and 340 have gate-to-source capacitances of Cgs1, Cgs2, Cgs3 and CgsA, respectively, and further have gate-to-source voltages of Vgs1, v gs2, V gs3 and V gs4, respectively. Inductor 350 is modeled with an ideal inductor 550. A circuit 508 models the input impedance Z\ of N-FET 310. [0038] For simplicity, the following assumptions are made for equivalent circuit 500: • All parasitic capacitances are negligible except for Cgs for each N-FET; • Parasitic resistances are zero; • The body effects of the N-FETs are negligible; and • LNA 140a operates in a weakly nonlinear region with a small input signal Vi. [0039] The drain currents for current sources 510, 520 and 530 may be expressed as: where v2 = vgs3 = -vgr2. For simplicity, only the nonlinearities of N-FETs 310 and 330 are considered, and N-FETs 320 and 340 are assumed to be linear, as indicated by equation (7b). [0040] The output current iout for equivalent circuit 500 in a weakly nonlinear region may be expressed as: Eq(8) where C„(.s1,,...,sn) is a Laplace transform of the n-th order Volterra kernel for iout, which is often called the n-th order nonlinear function; s = je> is the Laplace variable; sv..., sn are frequencies operated on by the n-th order Volterra kernel; and " o " denotes a complex multiply of each frequency component of v" by WO 2007/016552 PCT/US2006/029905 [0041] Equation (8) is for a Volterra series that is often used for nonlinear analysis. The Volterra series includes a Volterra kernel for each order of nonlinearity. The n-th order nonlinearity corresponds to the term v," and generates n frequency components. The n-th Volterra kernel is a set of n coefficients that operates on the n frequency components generated by the n-th order nonlinearity. The coefficients for each Volterra kernel may be determined by mathematical derivation or some other means. In equation (8), the third order Volterra kernel C3(s1s2s3) determines third order nonlinearity at high frequency, which is of interest. [0042] The gate-to-source voltage vgs1 of N-FET 310 may be expressed as a function of the input voltage v1, as follows: Eq(9) where An(sl,...,sn) is the Laplace transform of the n-th order Volterra kernel for vgs1 [0043] N-FET 310 generates a nonlinear current ids1 based on the input voltage vi, as shown in equations (7a) and (9). A portion of the ids1 current passes through N-FET 320 and generates the v2 voltage. The V2 voltage generates a nonlinear current ids3 through N-FET 330, as shown in equation (7c). The output current iout is equal to the sum of the ids1 current and the i^ current. [0044] Equation (8) may be evaluated to determine all distortion components. The distortion components of interest are those that affect HP3. The distortion components generated by third order nonlinearity of N-FET 310 are denoted as £m. The distortion components generated by nonlinearities of N-FET 330 may be categorized as follows: • 51: distortion components generated by second and third order nonlinearities of N-FET 310 and attenuated by a factor ofa/b; • 52 : distortion components generated by second order nonlinearity of N-FET 310 multiplied by second order nonlinearity of N-FET 330; and • 53 : distortion components generated by third order nonlinearity of N-FET 330. With active post-distortion linearization, the terms 51,, 52 and 53 are actively generated with N-FET 330 and are used to cancel the term 5m1 from N-FET 310. WO 2007/016552 PCT/US2006/029905 [0045] The term £", includes distortion components generated by second and third order nonlinearities of N-FET 310. For example, the second harmonic (2w) at the source of N-FET 310 can mix with the fundamental frequency (w) at the gate of N-FET 310 to generate third order intermodulation distortion. The second harmonic is due to second order nonlinearity of N-FET 310, which corresponds to the term g2'v2gs1 equation (7a). The fundamental frequency can also generate third order intermodulation distortion due to third order nonlinearity of N-FET 310, which corresponds to the term g3'v3gs1 in equation (7a). These distortion components from N-FET 310 are amplified by N-FET 330 through the g, g2'vgs3 term in equation (7c) and are attenuated by a factor of a/b by the combination of N-FETs 320 and 330. [0046] The term 52 includes distortion components generated by second order nonlinearities of N-FETs 310 and 330. For example, the second harmonic generated by the second order nonlinearity of N-FET 310 can mix with the fundamental frequency due to the second order nonlinearity of N-FET 330, which corresponds to the term g2'v2gs3 in equation (7c), to generate third order intermodulation distortion. [0047] The term £, includes distortion components generated by third order nonlinearity of N-FET 330. The fundamental frequency from N-FET 310 can generate third order intermodulation distortion due to third order nonlinearity of N-FET 330, which corresponds to the term g2'v2gs3 in equation (7c). [0048] The nonlinearity terms for N-FETs 310 and 330 may be expressed as: WO 2007/016552 PCT/US2006/029905 [0049] Equations (14), (15) and (16) indicate that the inductance Ls of inductor 350 is included in various intermediate terms that make up 51 52 and 53,. Inductor 350 improves cancellation of third order distortion generated by N-FET 310, at high frequency. [0050] In the above equations, s = jw , s1= jw1 , s2 = jw2 , and s3 = jw3 are different closely spaced signal frequencies, with w«w1x w a>2«= m2 -a>1 is much smaller than cax and a>2. Substituting equations (14), (15) and (16) into equations (10), (11), (12) and (13), and assuming conjugate match at co, the total third order distortion in the output current iout, IM3S., may be expressed as: IM3j. in equation (17) corresponds to the third order Volterra kernel C3(s1,s2,s3) in equation (8). [0051] In equation (17), the term in the first row represents third order nonlinearity, the term in the second row represents second order nonlinearity with second order harmonic, and the term in the third row represents second order nonlinearity. The values of a and fi may be selected such that these three distortion components cancel out as much as possible, the total third order distortion is minimized, and the highest possible IIP3 is achieved for LNA 140a. [0052J FIG. 6 shows a vector diagram that illustrates the distortion cancellation mechanism of active post-distortion. The terms 51, 52 and 53, are dependent on the signal frequency (S = Jw), the coefficients g1, g2 and g3 of the N-FETs, and the WO 2007/016552 PCT/US2006/029905 degeneration inductance Ls. The terms 51,52 and 53 can have different amplitudes and phases at a given frequency, as shown by the three vectors for these three terms. The sum of the three terms 51, 52 and 53 is shown by a dashed vector, which should be equal in amplitude but opposite in phase with the vector for 5m1, so that the total distortion is minimize. [0053] FIG. 4B shows a plot 430 of EDP3 for LNA 140a with distortion cancellation and a plot 440 of HP3 for LNA 140a without distortion cancellation at high frequency. For a given device width and power consumption, equation (17) may be solved to make the third order distortion component approach zero. The value of b is selected to prevent excessive gain loss. For a specific exemplary design, b is selected to be equal to 8, and a value of 1.77 for or provides good distortion cancellation. The value for a that minimizes distortion at high frequency may be different from the value for a at low frequency. The different a for high frequency is due to the distortion component generated by second order nonlinearity interacting with second harmonic, which corresponds to the second line in equation (17). [0054] The noise performance of LNA 140a is degraded slightly with active post-distortion linearization. The noise from N-FET 310 is approximately the same as the noise from a conventional inductively degenerated LNA. With active post-distortion linearization, additional noise is generated by N-FET 330 in the form of gate induced noise and drain noise. Both of these additional noise sources may be reduced by increasing /?, which results in less gain loss and smaller degradation in noise figure. [0055] FIG. 7A shows a schematic diagram of an embodiment of an LNA 140b with active post-distortion linearization. LNA 140b includes N-FETs 310, 320 and 330, inductor 350, and capacitor 352, which are coupled as described above for FIG. 3. However, the drain of N-FET 330 is coupled directly to the output node. N-FET 340 is omitted in LNA 140b. The linearity and noise performance of LNA 140b is similar to that of LNA 140a in FIG. 3. Omitting N-FET 340 mainly affects load isolation for N-FET 330. [0056] FIG. 7B shows a schematic diagram of an embodiment of an LNA 140c with active post-distortion linearization. LNA 140c includes N-FETs 310, 320 and 330, inductor 350, and capacitor 352, which are coupled as described above for FIG. 3. However, the drain of N-FET 330 is coupled directly to the source of N-FET 320. N- WO 2007/016552 PCT/US2006/029905 FET 340 is omitted in LNA 140c. The linearity and noise performance of LNA 140c is similar to that of LNA 140a in FIG. 3. [0057] FIG. 8 shows a schematic diagram of an embodiment of an LNA 140d with active post-distortion linearization and multiple gain settings. LNA 140d includes N-FETs 810, 820, 830 and 840, an inductor 850, and a capacitor 852 that are coupled in the same manner as N-FETs 310, 320, 330 and 340, inductor 350, and capacitor 352, respectively, in FIG, 3. LNA 140d further includes additional circuitry that provides biasing, gain control, and impedance matching. [0058] The bias circuitry for LNA 140d includes a current source 858, an N-FET 860, and resistors 862, 864, 866 and 868. Current source 858 has one end coupled to a power supply VDD and the other end coupled to the drain of N-FET 860. N-FET 860 is diode connected and has its source coupled to circuit ground and its gate coupled to its drain. Resistor 862 has one end coupled to the gate of N-FET 810 and the other end coupled to the gate of N-FET 860. Resistor 864 has one end coupled to the gate of N-FET 830 and the other end coupled to the gate of N-FET 860.. The bias current for N-FET 810 is determined by (1) the current provided by current source 858 and (2) the ratio of the width of N-FET 810 to the width of N-FET 860. Similarly, the bias current for N-FET 830 is determined by (1) the current provided by current source 858 and (2) the ratio of the width of N-FET 830 to the width of N-FET 860. Resistor 866 has one end coupled to the VDD supply and the other end coupled to the gates of N-FETs 820 and 840. Resistor 868 has one end coupled to circuit ground and the other end coupled to the gates of N-FETs 820 and 840. Resistors 866 and 868 determine the gate bias voltage for N-FETs 820 and 840, which does not need to be precisely set. [0059] The gain control circuitry for LNA 140d includes N-FETs 870 and 880, a capacitor 872, and resistors 882, 884 and 886. N-FETs 870 and 880 have their sources coupled to the gate of N-FET 810 and their gates receiving two gain control signals. Capacitor 872 has one end coupled to the drains of N-FETs 820 and 840 and the other end coupled to the drain of N-FET 870. Resistors 882 and 884 are coupled in series. Resistor 882 has one end coupled to the drain of N-FET 880 and the other end coupled to resistors 884 and 886. The other end of resistor 884 is coupled to the drains of N-FETs 820 and 840, and the other end of resistor 886 is coupled to circuit ground. [0060] N-FETs 810, 820, 830 and 840 form a gain signal path, N-FET 870 forms a pass-through signal path, and N-FET 880 forms an attenuation signal path. One of the WO 2007/016552 PCT/US2006/029905 three signal paths is selected as any given moment based on the two gain control signals. If N-FET 870 is turned on and the pass-through signal path is selected, then the input signal passes through N-FET 870 and AC coupling capacitor 872 to the LNA output. If N-FET 880 is turned on and the attenuation signal path is selected, then the input signal passes through N-FET 880 and is attenuated by the resistor network. [0061] An input impedance matching circuit 890 couples between an RF input and the gate of N-FET 810. An output impedance matching circuit 892 couples between an RF output and the VDD supply. Each impedance matching circuit may include one or more inductors, capacitors, strip lines, and so on. Matching circuit 892 also provides bias current for N-FETs 810, 820, 830, 840 and 880. [0062] FIG. 9 shows a schematic diagram of an embodiment of an LNA 140e with active post-distortion linearization. LNA 140e includes four P-channel FETs (P-FETs) 910, 920, 930 and 940, an inductor 950, and a capacitor 952. P-FET 910 has its source coupled to one end of inductor 950, its gate receiving the input voltage vi, and its drain coupled to the source of P-FET 920. The other end of inductor 950 couples to the VDD supply. P-FET 920 has its gate receiving the bias voltage vbias and its drain coupled to the output node. N-FET 930 has its source coupled to the VDD supply, its gate coupled to one end of capacitor 952, and its drain coupled to the source of P-FET 940. The other end of capacitor 952 couples to the source of P-FET 920. P-FET 940 has its gate receiving the bias voltage Vbias and its drain coupled to the output node. The output node provides the output current iout for LNA 140e. [0063] As noted above, the techniques for linearizing an active circuit using active post-distortion may be used for various types of active circuit such as amplifier, mixer, and so on. The main signal path for the active circuit generates distortion due to nonlinearity of the circuit elements in the main signal path. An auxiliary signal path actively generates distortion components used to cancel the distortion components generated by the main signal path. [0064] The amplifier and other linearized active circuits described herein may be used for various frequency ranges including baseband, intermediate frequency (IF), RF, and so on. For example, these linearized active circuits may be used for frequency bands commonly employed for wireless communication, such as: • Cellular band from 824 to 894 MHz, • Personal Communication System (PCS) band from 1850 to 1990 MHz, WO 2007/016552 PCT/US2006/029905 • Digital Cellular System (DCS) band from 1710 to 1880 MHz, • GSM900 band from 890 to 960 MHz, • International Mobile Telecommunications-2000 (IMT-2000) band from 1920 to 2170 MHz, and • Global Positioning System (GPS) band from 1574.4 to 1576.4 MHz. [0065] The amplifier and other linearized active circuits described herein may be implemented within an integrated circuit (IC), an RF integrated circuit (RFIC), an application specific integrated circuit (ASIC), a printed circuit board (PCB), an electronic device, and so on. These linearized active circuits may also be fabricated with various IC process technologies such as complementary metal oxide semiconductor (CMOS), N-channel MOS (N-MOS), P-channel MOS (P-MOS), bipolar junction transistor (BJT), bipolar-CMOS (BiCMOS), silicon germanium (SiGe), gallium arsenide (GaAs), and so on. [0066] The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein. WO 2007/016552 PCT/US2006/029905 WE CLAIM : 1. An integrated circuit comprising: a first transistor electrically coupled to an inductor and operative to receive and amplify an input signal; a second transistor electrically coupled to the first transistor and operative to generate an intermediate signal and to provide an output signal; and a third transistor electrically coupled to the second transistor and operative to receive the intermediate signal and generate distortion components used to cancel a distortion component generated by the first transistor. 2. The integrated circuit of claim 1, wherein the first and second transistors are coupled as a cascode pair. 3. The integrated circuit of claim 1, wherein the third transistor is operative to generate distortion components used to cancel third order distortion component generated by the first transistor. 4. The integrated circuit of claim 1, further comprising: a fourth transistor electrically coupled to the second and third transistors and operative to provide load isolation. 5. The integrated circuit of claim 1, wherein the third transistor has a drain electrically coupled to a drain of the second transistor. 6. The integrated circuit of claim 1, wherein the third transistor has a drain electrically coupled to a drain of the first transistor. 7. The integrated circuit of claim 1, wherein the first, second, and third transistors have first, second, and third gains, respectively, wherein the first and second gains are related by a first factor, and wherein the first and third gains are related by a second factor. WO 2007/016552 PCT/US2006/029905 8. The integrated circuit of claim 7, wherein the second factor is selected to reduce gain loss, and wherein the first factor is selected to cancel the distortion component generated by the first transistor. 9. The integrated circuit of claim 7, wherein the second factor is greater than one, and wherein the third gain is a fraction of the first gain. 10. The integrated circuit of claim 1, wherein the first, second, and third transistors are N-channel field effect transistors (N-FETs). 11. The integrated circuit of claim 1, wherein the first, second, and third transistors are P-channel field effect transistors (P-FETs). 12. The integrated circuit of claim 1, wherein the first, second, and third transistors are bipolar junction transistors (BJTs). 13. The integrated circuit of claim 1, wherein the first, second, and third transistors form a low noise amplifier (LNA). 14. The integrated circuit of claim 1, further comprising: a gain control circuit electrically coupled to the first and second transistors and operative to provide gain control for an amphfier formed by the first, second, and third transistors. 15. The integrated circuit of claim 14, wherein the gain control circuit comprises a fourth transistor electrically coupled to the first transistor; and a capacitor electrically coupled between the fourth transistor and the second transistor. 16. The integrated circuit of claim 14, wherein the gain control circuit comprises a fourth transistor electrically coupled to the first transistor; and WO 2007/016552 PCT/US2006/029905 at least one resistor electrically coupled between the fourth transistor and the second transistor. 17. The integrated circuit of claim 1, wherein the input signal is a Code Division Multiple Access (CDMA) signal. 18. An amplifier comprising: an inductor operative to provide source degeneration; a first transistor having a source electrically coupled to the inductor and a gate receiving an input signal, the first transistor operative to provide signal amplification; a second transistor having a drain providing an output signal and a source electrically coupled to a drain of the first transistor, the second transistor operative to generate an intermediate signal; and a third transistor having a gate electrically coupled to the source of the second transistor, the third transistor operative to receive the intermediate signal and generate distortion components used to cancel a distortion component generated by the first transistor. 19. The amplifier of claim 18, further comprising: a fourth transistor having a source electrically coupled to a drain of the third transistor and a drain electrically coupled to the drain of the second transistor. 20. An apparatus comprising: means for amplifying an input signal to generate a first signal having a desired component and a distortion component; means for generating a second signal having distortion components used to cancel the distortion component generated by the amplifying means; and means for combining the first and second signals to generate an output signal having the distortion component generated by the amplifying means canceled. 21. The apparatus of claim 20, wherein the distortion component generated by the amplifying means is a third order distortion component. WO 2007/016552 PCT/US2006/029905 22. The apparatus of claim 20, further comprising: means for controlling gain of the output signal. 23. A receiver for a wireless device, comprising: a low noise amplifier (LNA) comprising an inductor operative to provide source degeneration, a first transistor having a source electrically coupled to the inductor and operative to provide signal amplification, a second transistor having a source electrically coupled to a drain of the first transistor and operative to generate an intermediate signal, and a third transistor having a gate electrically coupled to the source of the second transistor, the third transistor operative to receive the intermediate signal and generate distortion components used to cancel a distortion component generated by the first transistor; an input impedance matching circuit electrically coupled to a gate of the first transistor and receiving an input signal for the LNA; and an output impedance matching circuit electrically coupled to a drain of the second transistor and providing an output signal for the LNA. 24. The receiver of claim 23, wherein the LNA further comprises a fourth transistor having a source electrically coupled to a drain of the third transistor and a drain electrically coupled to the drain of the second transistor. 25. The receiver of claim 23, wherein the LNA further comprises a gain control circuit electrically coupled to the first and second transistors and operative to provide gain control for the LNA. ABSTRACT AMPLIFIER WITH ACTIVE POST-DISTORTION LINEARIZATION An amplifier, which has good linearity and noise performance, includes first, second, third, and fourth transistors and an inductor. The first and second transistors are coupled as a first cascode pair, and the third and fourth transistors are coupled as a second cascode pair. The third transistor has its gate coupled to the source of the second transistor, and the fourth transistor has its drain coupled to the drain of the second transistor. The first transistor provides signal amplification. The second transistor provides load isolation and generates an intermediate signal for the third transistor. The third transistor generates distortion components used to cancel third order distortion component generated by the first transistor. The inductor provides source degeneration for the first transistor and improves distortion cancellation. The sizes of the second and third transistors are selected to reduce gain loss and achieve good linearity for the amplifier.

Full Text

FORM 2
THE PATENTS ACT, 1970
(39 of 1970)
&
THE PATENTS RULES, 2003
COMPLETE SPECIFICATION
(See section 10, rule 13)
'AMPLIFIER WITH ACTIVE POST-DISTORTION LINEARIZATION"
QUALCOMM INCORPORATED, a company incorporated in the state of Delaware, U.S., of 5775 Morehouse Drive, San Diego, California 92121-1714, U.S.A
The following specification particularly describes the invention and the manner in which it is to be performed.

WO 2007/016552

PCT/US2006/029905

AMPLIFIER WITH ACTIVE POST-DISTORTION LINEARIZATION
[0001] This application claims the benefit of provisional U.S. Application Serial No. 60/705,256, entitled "Linearity improvement technique for CMOS amplifiers from low frequency to high frequency by using vectorized post-distortion," filed August 2, 2005.
BACKGROUND
I. Field
[0002] The present disclosure relates generally to circuits, and more specifically to an amplifier suitable for wireless communication and other applications.
II. Background
[0003] Amplifiers are commonly used in various electronics devices to provide signal amplification. Furthermore, different types of amplifiers are available for different uses. For example, a wireless device may include a transmitter and a receiver for bi-directional communication, and the transmitter may utilize a power amplifier (PA) and the receiver may utilize a low noise amplifier (LNA) and a variable gain amplifier (VGA).
[0004] An LNA is commonly used in a receiver to amplify a low-amplitude signal received via a communication channel. The LNA is often the first active circuit encountered by the received signal and hence has a large impact on the performance of the receiver in several key areas. First, the LNA has a large influence on the overall noise figure of the receiver since the noise of the LNA is injected directly into the received signal and the noise of subsequent stages is effectively reduced by the gain of the LNA. Second, the linearity of the LNA has a large influence on both the design of subsequent stages in the receiver and the receiver performance. The LNA input signal typically includes various undesired signal components that may come from external interfering sources and leakage from a co-located transmitter. Nonlinearity in the LNA causes the undesired signal components to mix and generate cross modulation distortion (XMD) that may fall within the desired signal bandwidth. The amplitude of the cross

WO 2007/016552 PCT/US2006/029905
modulation distortion is determined by the amount of nonlinearity in the LNA. The cross, modulation distortion component that falls within the desired signal bandwidth acts as noise that degrades the signal-to-noise ratio (SNR) of .the desired signal. The degradation in SNR caused by LNA nonlinearity impacts the design of (and often places more stringent requirements on) subsequent stages in order to meet the overall SNR specification for the receiver. Therefore, having a more linear LNA can alleviate the performance requirements for other stages, which may result in lower power consumption and smaller circuit area for the receiver.
[0005] There is therefore a need in the art for an amplifier having good linearity and noise performance.
SUMMARY
[0006] Various embodiments of an amplifier linearized using active post-distortion (APD) are described herein. The amplifier is simple in design, has good linearity and noise performance, and is suitable for wireless communication and other high frequency applications. For example, the amplifier may be used as an LNA for a receiver in a wireless device. Active post-distortion may also be used to linearize other active circuits such as, e.g., a mixer.
[0007] In an embodiment, an amplifier (e.g., an LNA) includes first, second, third, and fourth transistors (e.g., N-FETs) and an inductor. The first and second transistors are coupled as a first cascode pair, and the third and fourth transistors are coupled as a second cascode pair. The first transistor has its source coupled to the inductor and its gate receiving an input (voltage) signal. The second transistor has its source coupled to the drain of the first transistor and its drain providing an output (current) signal. The third transistor has its gate coupled to the source of the second transistor. The fourth transistor has its source coupled to the drain of the third transistor and its drain coupled to the drain of the second transistor. The first transistor provides signal amplification. The second transistor provides load isolation and further generates an intermediate signal for the third transistor. The third transistor receives the intermediate signal and generates distortion components used to cancel third order distortion component
t
generated by the first transistor. The fourth transistor provides load isolation. The inductor provides source degeneration for the first transistor and improves the cancellation of the third order distortion. In other embodiments, the fourth transistor

WO 2007/016552 PCT/US2006/029905
may be omitted, and the drain of the third transistor may be coupled to the drain of either the first or second transistor. The sizes of the second and third transistors may be selected to reduce gain loss for the amplifier and to cancel as much third order distortion as possible.
[0008] Various aspects and embodiments of the invention are described in further detail below.
BRIEF DESCRIPTION OF THE DRAWINGS
[0009] The features and nature of the present invention will become more apparent
from the detailed description set forth below when taken in conjunction with the
drawings in which like reference characters identify correspondingly throughout.
[0010] FIG. 1 shows a radio frequency (RF) portion of a wireless device.
[0011] FIGS. 2A, 2B and 2C show a received signal from an antenna, an LNA input
signal, and an LNA output signal, respectively.
[0012] FIG. 3 shows a schematic diagram of an LNA with active post-distortion
linearization.
[0013] FIGS. 4A and 4B show plots of IIP3 for the LNA for low and high
frequencies, respectively.
[0014] FIG. 5 shows an equivalent circuit for the LNA.
[0015] FIG. 6 shows a vector diagram illustrating active post-distortion cancellation.
[0016] FIGS. 7A and 7B show schematic diagrams of two additional embodiments
of an LNA with active post-distortion linearization.
[0017] FIG. 8 shows a schematic diagram of an LNA with active post-distortion
linearization and multiple gain settings.
[0018] FIG. 9 shows a schematic diagram of an LNA implemented with P-FETs.
DETAILED DESCRIPTION
[0019] The word "exemplary" is used herein to mean "serving as an example, instance, or illustration." Any embodiment or design described herein as "exemplary" is not necessarily to be construed as preferred or advantageous over other embodiments or designs.
[0020] The amplifier and other linearized active circuits described herein may be used for various applications such as communication, networking, computing, consumer

WO 2007/016552 PCT/US2006/029905

electronics, and so on. These linearized active circuits may be used in wireless communication systems such as a Code Division Multiple Access (CDMA) system, a Time Division Multiple Access (TDMA) system, a Global System for Mobile Communications (GSM) system, an Advanced Mobile Phone System (AMPS) system, Global Positioning System (GPS), a multiple-input multiple-output (MIMO) system, an orthogonal frequency division multiplexing (OFDM) system, an orthogonal frequency division multiple access (OFDMA) system, a single-carrier FDMA (SC-FDMA) system, a wireless local area network (WLAN), and so on. The amplifier may be used as an LNA, a VGA, a PA, and so on. For clarity, an LNA used in a receiver of a wireless device for a CDMA system is described below. The CDMA system may implement cdma2000, Wideband CDMA (W-CDMA), and/or other CDMA radio access technologies.
[0021] FIG. 1 shows a block diagram of a radio frequency (RF) portion of a wireless device 100. Wireless device 100 may be a cellular phone, a personal digital assistant (PDA), a wireless modem card, or some other device used for wireless communication. Wireless device 100 includes a transmitter and a receiver that provide bi-directional communication.
[0022] On the transmit path, a power amplifier (PA) 110 receives and amplifies a transmit (TX) modulated signal and provides a transmit signal. The transmit signal is routed through a duplexer 120 and transmitted via an antenna 130 to one or more serving base stations. A portion of the transmit signal also couples or leaks through duplexer 120 to the receive path. The amount of TX leakage is dependent on the isolation between the transmit and receive ports of duplexer 120, which may be approximately 50 decibels (dB) for a surface acoustic wave (SAW) duplexer at cellular band. A lower TX-RX isolation results in higher level of TX leakage.
[0023] On the receive path, a received signal containing a desired signal and possibly a jammer is received via antenna 130, routed through duplexer 120, and provided to an LNA 140. LNA 140 also receives a TX leakage signal from the transmit path. The input signal at the input of LNA 140 may thus include the desired signal, the TX leakage signal, and the jammer. LNA amplifies the input signal and provides an amplified RF signal. A SAW filter 150 filters me amplified RF signal to remove out-of-band components (e.g., the TX leakage signal) and provides a filtered RF signal. A

WO 2007/016552

PCT/US2006/029905

mixer 160 frequency downconverts the filtered RF signal with a local oscillator (LO) signal and provides a downconverted signal.
[0024] FIG. 2A shows the received signal from antenna 130, which includes a desired signal 210 and a jammer 220. Jammer 220 is an undesired signal and may correspond to, for example, a signal transmitted by a nearby base station in an AMPS system. The jammer may be much higher in amplitude than the desired signal and may be located close in frequency to the desired signal.
[0025] FIG. 2B shows the input signal at the input of LNA 140. The input signal contains desired signal 210 and jammer 220 in the received signal as well as a TX leakage signal 230 from the transmit path. The TX leakage signal may be large relative to the desired signal, especially if wireless device 100 is far from the serving base station(s) and needs to transmit at a high power level in order to reach the base station(s).
[0026] FIG. 2C shows the signal at the output of LNA 140. Nonlinearity in LNA 140 can cause the modulation on TX leakage signal 230 to interact with narrowband jammer 220 and generate cross modulation distortion 240 around the jammer. A portion 250 of the cross modulation distortion, which is shown with shading, may fall within the desired signal band. Portion 250 acts as additional noise that degrades the performance of the receiver. This noise also degrades receiver sensitivity so that the smallest desired signal that can be reliably detected by the receiver needs to have a larger amplitude.
[0027] FIG. 3 shows a schematic diagram of an embodiment of an LNA 140a with active post-distortion (APD) linearization. LNA 140a has good linearity and noise performance and may be used for LNA 140 in FIG. 1. LNA 140a includes four N-channel field effect transistors (N-FETs) 310, 320, 330 and 340, an inductor 350, and a capacitor 352. N-FET 310 has its source coupled to one end of inductor 350, its gate receiving an input voltage vi, and its drain coupled to the source of N-FET 320. The other end of inductor 350 couples to circuit ground. N-FET 320 has its gate receiving a bias voltage Vbias and its drain coupled to an output node. N-FET 330 has its source coupled to circuit ground, its gate coupled to one end of capacitor 352, and its drain coupled to the source of N-FET 340. The other end of capacitor 352 couples to the source of N-FET 320. N-FET 340 has its gate receiving the bias voltage vbias and its

WO 2007/016552 PCT/US2006/029905
drain coupled to the output node. The output node provides an output current iout for LNA 140a.
[0028] N-FETs 310 and 320 form a first cascode pair for a main signal path used for signal amplification. N-FET 310 provides signal amplification. N-FET 320 provides load isolation for N-FET 310 and further generates an intermediate voltage v2 for N-FET 330. N-FETs 330 and 340 form a second cascode pair for an auxiliary signal path that generates cross modulation distortion used for distortion cancellation. N-FET 330 generates the cross modulation distortion, and N-FET 340 provides load isolation for N-FET 330. Inductor 350 provides source degeneration and further provides a 50-ohm match looking into the gate of N-FET 310. Inductor 350 is also used for active post-distortion linearization and improves distortion cancellation. Capacitor 352 provides AC coupling.
[0029] N-FET 310 has a small-signal transconductance of g1, which is determined by various factors such as the size (e.g., length and width) of N-FET 310, the bias current for N-FET 310, the gate-to-source voltage vgs of N-FET 310, and so on. N-FET 320 has a small-signal transconductance of g1/a , where a is the ratio of the
transconductance of N-FET 310 to the transconductance of N-FET 320. The factor a is typically determined by the ratio of the width of N-FET 310 to the width of N-FET 320. N-FET 330 has a small-signal transconductance of gx Ib, where b is the ratio of the
transconductance of N-FET 310 to the transconductance of N-FET 330. The factor a is typically determined by the ratio of the width of N-FET 310 to the width of N-FET 330. The factors a and /? may be selected as described below.
[0030] Linearization of LNA 140a using active post-distortion may be achieved at low frequency as follows. At low frequency, inductor 350 does not come into play and is effectively shorted, and the input voltage vi is equal to the Vg, voltage for N-FET 310. The drain current i'i of N-FET 310 may be represented by a power series as:
Eq(l)
where g2 is a coefficient that defines the strength of second order nonlinearity;
g3 is a coefficient that defines the strength of third order nonlinearity; and h(vg,) is the dram current of N-FET 310 as a function of VgS.

WO 2007/016552 PCT/US2006/029905
For simplicity, nonlinearities higher than third order are ignored in equation (1). Coefficients g1, gz and g} are determined by the device size and the bias current for N-FET 310. Coefficient g% controls the third order intermodulation distortion (IMD3) at low signal level and hence determines the third order input intercept point (IIP3), which is a metric commonly used to specify the linearity of an amplifier.
[0031] N-FET 320 may be assumed to be linear, hi this case, the drain voltage V2 of N-FET 310, which is also the vgs voltage for N-FET 330, may be expressed as:
Eq(2)

Equation (2) indicates that the V2 voltage generated by N-FET 320 is dependent on a. The drain current i3 of N-FET 330 may be represented by a power series as:
Eq(3)
Equation (3) indicates that the coefficients for N-FET 330 and the coefficients for N-FET 310 are related by p.
[0032] Equation (2) may be substituted into equation (3) so that the drain current i3 of N-FET 330 can be expressed as a function of the drain current i\ of N-FET 310. Equation (1) may then be substituted into equation (3) so that the drain current 13 of N-FET 330 can be expressed as a function of the vgs voltage of N-FET 310. The expanded equation (3) includes multiple terms for each order of nonlinearity due to the interaction between the power series in equation (1) and the power series in equation (3).
[0033] The drain currents of N-FETs 310 and 330 are combined to generate the output current i„a, as follows:

Eq(4)
where gxx and g3Z are the first and third order power series coefficients, respectively, for the output current iout and may be expressed as:

WO 2007/016552 PCT/US2006/029905

The term g2J. in equation (4) may be ignored since only the fundamental frequency and the third order nonlinearity are of interest.
[0034] Equation (5) represents an overall gain for LNA 140a and shows a gain loss resulting from the use of active post-distortion linearization. The overall gain for LNA 140a with distortion cancellation is giz whereas the gain for the LNA without distortion cancellation is g\. The gain loss of (1 - a/fi) is directly related to a and /3 and may be
kept small by selecting /? to be large relative to a. A larger J3 leads to less gain loss but does not necessarily mean less distortion cancellation. Equation (6) represents the combined third order distortion in the output current iout. The first term in equation (6) represents the contribution from third order nonlinearity, and the second term in equation (6) represents the contribution from second order nonlinearity.
[0035] FIG. 4A shows a plot 410 of HP3 for LNA 140a with distortion cancellation (with N-FETs 330 and 340 connected) and a plot 420 of IIP3 for LNA 140a without distortion cancellation (with N-FETs 330 and 340 omitted) at low frequency. For a given device width and power consumption, equation (6) may be solved such that the third order distortion component approaches zero. The value of J3 is selected to prevent excessive gain loss. For a specific exemplary design, fi is selected to be equal to 8, and a value of 1.35 for a provides good distortion cancellation. Because of the second order nonlinearity in equation (6), the distortion cancellation is dependent on bias voltage, which is the operating vgs voltage for N-FET 310.
[0036] LNA 140a may be used for high frequency applications such as wireless communication. At high frequency, reactive elements such as capacitors and inductors affect linearity performance and further cause performance to be frequency dependent. [0037] FIG. 5 stiows a schematic diagram of a simplified equivalent circuit 500 for LNA 140a in FIG. 3. For the embodiment shown in FIG. 5, N-FETs 310, 320, 330 and 340 are modeled with ideal current sources 510, 520, 530 and 540, respectively, and parasitic gate-to-source capacitors 512, 522, 532 and 542, respectively. N-FETs 310,

WO 2007/016552 PCT/US2006/029905

320, 330 and 340 have gate-to-source capacitances of Cgs1, Cgs2, Cgs3 and CgsA, respectively, and further have gate-to-source voltages of Vgs1, v gs2, V gs3 and V gs4, respectively. Inductor 350 is modeled with an ideal inductor 550. A circuit 508 models the input impedance Z\ of N-FET 310.
[0038] For simplicity, the following assumptions are made for equivalent circuit 500:
• All parasitic capacitances are negligible except for Cgs for each N-FET;
• Parasitic resistances are zero;
• The body effects of the N-FETs are negligible; and
• LNA 140a operates in a weakly nonlinear region with a small input signal Vi.
[0039] The drain currents for current sources 510, 520 and 530 may be expressed as:

where v2 = vgs3 = -vgr2. For simplicity, only the nonlinearities of N-FETs 310 and 330
are considered, and N-FETs 320 and 340 are assumed to be linear, as indicated by equation (7b).
[0040] The output current iout for equivalent circuit 500 in a weakly nonlinear region may be expressed as:
Eq(8)
where C„(.s1,,...,sn) is a Laplace transform of the n-th order Volterra kernel for iout, which is often called the n-th order nonlinear function; s = je> is the Laplace variable;
sv..., sn are frequencies operated on by the n-th order Volterra kernel; and " o " denotes a complex multiply of each frequency component of v" by

WO 2007/016552 PCT/US2006/029905

[0041] Equation (8) is for a Volterra series that is often used for nonlinear analysis. The Volterra series includes a Volterra kernel for each order of nonlinearity. The n-th order nonlinearity corresponds to the term v," and generates n frequency components. The n-th Volterra kernel is a set of n coefficients that operates on the n frequency components generated by the n-th order nonlinearity. The coefficients for each Volterra kernel may be determined by mathematical derivation or some other means. In equation (8), the third order Volterra kernel C3(s1s2s3) determines third order nonlinearity at high frequency, which is of interest.
[0042] The gate-to-source voltage vgs1 of N-FET 310 may be expressed as a function of the input voltage v1, as follows:
Eq(9)
where An(sl,...,sn) is the Laplace transform of the n-th order Volterra kernel for vgs1
[0043] N-FET 310 generates a nonlinear current ids1 based on the input voltage vi, as shown in equations (7a) and (9). A portion of the ids1 current passes through N-FET 320 and generates the v2 voltage. The V2 voltage generates a nonlinear current ids3 through N-FET 330, as shown in equation (7c). The output current iout is equal to the sum of the ids1 current and the i^ current.
[0044] Equation (8) may be evaluated to determine all distortion components. The distortion components of interest are those that affect HP3. The distortion components generated by third order nonlinearity of N-FET 310 are denoted as £m. The distortion components generated by nonlinearities of N-FET 330 may be categorized as follows:
• 51: distortion components generated by second and third order nonlinearities of N-FET 310 and attenuated by a factor ofa/b;
• 52 : distortion components generated by second order nonlinearity of N-FET 310 multiplied by second order nonlinearity of N-FET 330; and
• 53 : distortion components generated by third order nonlinearity of N-FET 330.
With active post-distortion linearization, the terms 51,, 52 and 53 are actively generated with N-FET 330 and are used to cancel the term 5m1 from N-FET 310.

WO 2007/016552 PCT/US2006/029905
[0045] The term £", includes distortion components generated by second and third order nonlinearities of N-FET 310. For example, the second harmonic (2w) at the source of N-FET 310 can mix with the fundamental frequency (w) at the gate of N-FET 310 to generate third order intermodulation distortion. The second harmonic is due to second order nonlinearity of N-FET 310, which corresponds to the term g2'v2gs1
equation (7a). The fundamental frequency can also generate third order intermodulation distortion due to third order nonlinearity of N-FET 310, which corresponds to the term g3'v3gs1 in equation (7a). These distortion components from N-FET 310 are amplified
by N-FET 330 through the g, g2'vgs3 term in equation (7c) and are attenuated by a factor
of a/b by the combination of N-FETs 320 and 330.
[0046] The term 52 includes distortion components generated by second order
nonlinearities of N-FETs 310 and 330. For example, the second harmonic generated by
the second order nonlinearity of N-FET 310 can mix with the fundamental frequency
due to the second order nonlinearity of N-FET 330, which corresponds to the term
g2'v2gs3 in equation (7c), to generate third order intermodulation distortion.
[0047] The term £, includes distortion components generated by third order
nonlinearity of N-FET 330. The fundamental frequency from N-FET 310 can generate third order intermodulation distortion due to third order nonlinearity of N-FET 330, which corresponds to the term g2'v2gs3 in equation (7c).
[0048] The nonlinearity terms for N-FETs 310 and 330 may be expressed as:

WO 2007/016552

PCT/US2006/029905

[0049] Equations (14), (15) and (16) indicate that the inductance Ls of inductor 350 is included in various intermediate terms that make up 51 52 and 53,. Inductor 350
improves cancellation of third order distortion generated by N-FET 310, at high
frequency.
[0050] In the above equations, s = jw , s1= jw1 , s2 = jw2 , and s3 = jw3 are
different closely spaced signal frequencies, with w«w1x w a>2«= m2 -a>1 is much smaller than cax and a>2. Substituting equations (14), (15) and (16) into equations (10), (11), (12) and (13), and assuming conjugate match at co, the total third order distortion in the output current iout, IM3S., may be expressed as:

IM3j. in equation (17) corresponds to the third order Volterra kernel C3(s1,s2,s3) in
equation (8).
[0051] In equation (17), the term in the first row represents third order nonlinearity,
the term in the second row represents second order nonlinearity with second order
harmonic, and the term in the third row represents second order nonlinearity. The
values of a and fi may be selected such that these three distortion components cancel
out as much as possible, the total third order distortion is minimized, and the highest
possible IIP3 is achieved for LNA 140a.
[0052J FIG. 6 shows a vector diagram that illustrates the distortion cancellation
mechanism of active post-distortion. The terms 51, 52 and 53, are dependent on the
signal frequency (S = Jw), the coefficients g1, g2 and g3 of the N-FETs, and the

WO 2007/016552 PCT/US2006/029905
degeneration inductance Ls. The terms 51,52 and 53 can have different amplitudes and phases at a given frequency, as shown by the three vectors for these three terms. The sum of the three terms 51, 52 and 53 is shown by a dashed vector, which should be
equal in amplitude but opposite in phase with the vector for 5m1, so that the total distortion is minimize.
[0053] FIG. 4B shows a plot 430 of EDP3 for LNA 140a with distortion cancellation and a plot 440 of HP3 for LNA 140a without distortion cancellation at high frequency. For a given device width and power consumption, equation (17) may be solved to make the third order distortion component approach zero. The value of b is selected to prevent excessive gain loss. For a specific exemplary design, b is selected to be equal to 8, and a value of 1.77 for or provides good distortion cancellation. The value for a that minimizes distortion at high frequency may be different from the value for a at low frequency. The different a for high frequency is due to the distortion component generated by second order nonlinearity interacting with second harmonic, which corresponds to the second line in equation (17).
[0054] The noise performance of LNA 140a is degraded slightly with active post-distortion linearization. The noise from N-FET 310 is approximately the same as the noise from a conventional inductively degenerated LNA. With active post-distortion linearization, additional noise is generated by N-FET 330 in the form of gate induced noise and drain noise. Both of these additional noise sources may be reduced by increasing /?, which results in less gain loss and smaller degradation in noise figure. [0055] FIG. 7A shows a schematic diagram of an embodiment of an LNA 140b with active post-distortion linearization. LNA 140b includes N-FETs 310, 320 and 330, inductor 350, and capacitor 352, which are coupled as described above for FIG. 3. However, the drain of N-FET 330 is coupled directly to the output node. N-FET 340 is omitted in LNA 140b. The linearity and noise performance of LNA 140b is similar to that of LNA 140a in FIG. 3. Omitting N-FET 340 mainly affects load isolation for N-FET 330.
[0056] FIG. 7B shows a schematic diagram of an embodiment of an LNA 140c with active post-distortion linearization. LNA 140c includes N-FETs 310, 320 and 330, inductor 350, and capacitor 352, which are coupled as described above for FIG. 3. However, the drain of N-FET 330 is coupled directly to the source of N-FET 320. N-

WO 2007/016552 PCT/US2006/029905
FET 340 is omitted in LNA 140c. The linearity and noise performance of LNA 140c is similar to that of LNA 140a in FIG. 3.
[0057] FIG. 8 shows a schematic diagram of an embodiment of an LNA 140d with active post-distortion linearization and multiple gain settings. LNA 140d includes N-FETs 810, 820, 830 and 840, an inductor 850, and a capacitor 852 that are coupled in the same manner as N-FETs 310, 320, 330 and 340, inductor 350, and capacitor 352, respectively, in FIG, 3. LNA 140d further includes additional circuitry that provides biasing, gain control, and impedance matching.
[0058] The bias circuitry for LNA 140d includes a current source 858, an N-FET 860, and resistors 862, 864, 866 and 868. Current source 858 has one end coupled to a power supply VDD and the other end coupled to the drain of N-FET 860. N-FET 860 is diode connected and has its source coupled to circuit ground and its gate coupled to its drain. Resistor 862 has one end coupled to the gate of N-FET 810 and the other end coupled to the gate of N-FET 860. Resistor 864 has one end coupled to the gate of N-FET 830 and the other end coupled to the gate of N-FET 860.. The bias current for N-FET 810 is determined by (1) the current provided by current source 858 and (2) the ratio of the width of N-FET 810 to the width of N-FET 860. Similarly, the bias current for N-FET 830 is determined by (1) the current provided by current source 858 and (2) the ratio of the width of N-FET 830 to the width of N-FET 860. Resistor 866 has one end coupled to the VDD supply and the other end coupled to the gates of N-FETs 820 and 840. Resistor 868 has one end coupled to circuit ground and the other end coupled to the gates of N-FETs 820 and 840. Resistors 866 and 868 determine the gate bias voltage for N-FETs 820 and 840, which does not need to be precisely set.
[0059] The gain control circuitry for LNA 140d includes N-FETs 870 and 880, a capacitor 872, and resistors 882, 884 and 886. N-FETs 870 and 880 have their sources coupled to the gate of N-FET 810 and their gates receiving two gain control signals. Capacitor 872 has one end coupled to the drains of N-FETs 820 and 840 and the other end coupled to the drain of N-FET 870. Resistors 882 and 884 are coupled in series. Resistor 882 has one end coupled to the drain of N-FET 880 and the other end coupled to resistors 884 and 886. The other end of resistor 884 is coupled to the drains of N-FETs 820 and 840, and the other end of resistor 886 is coupled to circuit ground.
[0060] N-FETs 810, 820, 830 and 840 form a gain signal path, N-FET 870 forms a pass-through signal path, and N-FET 880 forms an attenuation signal path. One of the

WO 2007/016552 PCT/US2006/029905
three signal paths is selected as any given moment based on the two gain control signals. If N-FET 870 is turned on and the pass-through signal path is selected, then the input signal passes through N-FET 870 and AC coupling capacitor 872 to the LNA output. If N-FET 880 is turned on and the attenuation signal path is selected, then the input signal passes through N-FET 880 and is attenuated by the resistor network.
[0061] An input impedance matching circuit 890 couples between an RF input and the gate of N-FET 810. An output impedance matching circuit 892 couples between an RF output and the VDD supply. Each impedance matching circuit may include one or more inductors, capacitors, strip lines, and so on. Matching circuit 892 also provides bias current for N-FETs 810, 820, 830, 840 and 880.
[0062] FIG. 9 shows a schematic diagram of an embodiment of an LNA 140e with active post-distortion linearization. LNA 140e includes four P-channel FETs (P-FETs) 910, 920, 930 and 940, an inductor 950, and a capacitor 952. P-FET 910 has its source coupled to one end of inductor 950, its gate receiving the input voltage vi, and its drain coupled to the source of P-FET 920. The other end of inductor 950 couples to the VDD supply. P-FET 920 has its gate receiving the bias voltage vbias and its drain coupled to the output node. N-FET 930 has its source coupled to the VDD supply, its gate coupled to one end of capacitor 952, and its drain coupled to the source of P-FET 940. The other end of capacitor 952 couples to the source of P-FET 920. P-FET 940 has its gate receiving the bias voltage Vbias and its drain coupled to the output node. The output node provides the output current iout for LNA 140e.
[0063] As noted above, the techniques for linearizing an active circuit using active post-distortion may be used for various types of active circuit such as amplifier, mixer, and so on. The main signal path for the active circuit generates distortion due to nonlinearity of the circuit elements in the main signal path. An auxiliary signal path actively generates distortion components used to cancel the distortion components generated by the main signal path.
[0064] The amplifier and other linearized active circuits described herein may be used for various frequency ranges including baseband, intermediate frequency (IF), RF, and so on. For example, these linearized active circuits may be used for frequency bands commonly employed for wireless communication, such as:
• Cellular band from 824 to 894 MHz,
• Personal Communication System (PCS) band from 1850 to 1990 MHz,

WO 2007/016552 PCT/US2006/029905
• Digital Cellular System (DCS) band from 1710 to 1880 MHz,
• GSM900 band from 890 to 960 MHz,
• International Mobile Telecommunications-2000 (IMT-2000) band from 1920 to 2170 MHz, and
• Global Positioning System (GPS) band from 1574.4 to 1576.4 MHz.
[0065] The amplifier and other linearized active circuits described herein may be implemented within an integrated circuit (IC), an RF integrated circuit (RFIC), an application specific integrated circuit (ASIC), a printed circuit board (PCB), an electronic device, and so on. These linearized active circuits may also be fabricated with various IC process technologies such as complementary metal oxide semiconductor (CMOS), N-channel MOS (N-MOS), P-channel MOS (P-MOS), bipolar junction transistor (BJT), bipolar-CMOS (BiCMOS), silicon germanium (SiGe), gallium arsenide (GaAs), and so on.
[0066] The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

WO 2007/016552 PCT/US2006/029905
WE CLAIM :
1. An integrated circuit comprising:
a first transistor electrically coupled to an inductor and operative to receive and amplify an input signal;
a second transistor electrically coupled to the first transistor and operative to generate an intermediate signal and to provide an output signal; and
a third transistor electrically coupled to the second transistor and operative to receive the intermediate signal and generate distortion components used to cancel a distortion component generated by the first transistor.
2. The integrated circuit of claim 1, wherein the first and second transistors are coupled as a cascode pair.
3. The integrated circuit of claim 1, wherein the third transistor is operative to generate distortion components used to cancel third order distortion component generated by the first transistor.
4. The integrated circuit of claim 1, further comprising:
a fourth transistor electrically coupled to the second and third transistors and operative to provide load isolation.
5. The integrated circuit of claim 1, wherein the third transistor has a drain electrically coupled to a drain of the second transistor.
6. The integrated circuit of claim 1, wherein the third transistor has a drain electrically coupled to a drain of the first transistor.
7. The integrated circuit of claim 1, wherein the first, second, and third transistors have first, second, and third gains, respectively, wherein the first and second gains are related by a first factor, and wherein the first and third gains are related by a second factor.

WO 2007/016552 PCT/US2006/029905
8. The integrated circuit of claim 7, wherein the second factor is selected to reduce gain loss, and wherein the first factor is selected to cancel the distortion component generated by the first transistor.
9. The integrated circuit of claim 7, wherein the second factor is greater than one, and wherein the third gain is a fraction of the first gain.
10. The integrated circuit of claim 1, wherein the first, second, and third transistors are N-channel field effect transistors (N-FETs).
11. The integrated circuit of claim 1, wherein the first, second, and third transistors are P-channel field effect transistors (P-FETs).
12. The integrated circuit of claim 1, wherein the first, second, and third transistors are bipolar junction transistors (BJTs).
13. The integrated circuit of claim 1, wherein the first, second, and third transistors form a low noise amplifier (LNA).
14. The integrated circuit of claim 1, further comprising:
a gain control circuit electrically coupled to the first and second transistors and operative to provide gain control for an amphfier formed by the first, second, and third transistors.
15. The integrated circuit of claim 14, wherein the gain control circuit
comprises
a fourth transistor electrically coupled to the first transistor; and a capacitor electrically coupled between the fourth transistor and the second transistor.
16. The integrated circuit of claim 14, wherein the gain control circuit
comprises
a fourth transistor electrically coupled to the first transistor; and

WO 2007/016552 PCT/US2006/029905
at least one resistor electrically coupled between the fourth transistor and the second transistor.
17. The integrated circuit of claim 1, wherein the input signal is a Code Division Multiple Access (CDMA) signal.
18. An amplifier comprising:
an inductor operative to provide source degeneration;
a first transistor having a source electrically coupled to the inductor and a gate receiving an input signal, the first transistor operative to provide signal amplification;
a second transistor having a drain providing an output signal and a source electrically coupled to a drain of the first transistor, the second transistor operative to generate an intermediate signal; and
a third transistor having a gate electrically coupled to the source of the second transistor, the third transistor operative to receive the intermediate signal and generate distortion components used to cancel a distortion component generated by the first transistor.
19. The amplifier of claim 18, further comprising:
a fourth transistor having a source electrically coupled to a drain of the third transistor and a drain electrically coupled to the drain of the second transistor.
20. An apparatus comprising:
means for amplifying an input signal to generate a first signal having a desired component and a distortion component;
means for generating a second signal having distortion components used to cancel the distortion component generated by the amplifying means; and
means for combining the first and second signals to generate an output signal having the distortion component generated by the amplifying means canceled.
21. The apparatus of claim 20, wherein the distortion component generated
by the amplifying means is a third order distortion component.

WO 2007/016552 PCT/US2006/029905
22. The apparatus of claim 20, further comprising: means for controlling gain of the output signal.
23. A receiver for a wireless device, comprising: a low noise amplifier (LNA) comprising
an inductor operative to provide source degeneration,
a first transistor having a source electrically coupled to the inductor and operative to provide signal amplification,
a second transistor having a source electrically coupled to a drain of the first transistor and operative to generate an intermediate signal, and
a third transistor having a gate electrically coupled to the source of the second transistor, the third transistor operative to receive the intermediate signal and generate distortion components used to cancel a distortion component generated by the first transistor;
an input impedance matching circuit electrically coupled to a gate of the first transistor and receiving an input signal for the LNA; and
an output impedance matching circuit electrically coupled to a drain of the second transistor and providing an output signal for the LNA.
24. The receiver of claim 23, wherein the LNA further comprises
a fourth transistor having a source electrically coupled to a drain of the third transistor and a drain electrically coupled to the drain of the second transistor.
25. The receiver of claim 23, wherein the LNA further comprises
a gain control circuit electrically coupled to the first and second transistors and operative to provide gain control for the LNA.

ABSTRACT
AMPLIFIER WITH ACTIVE POST-DISTORTION LINEARIZATION
An amplifier, which has good linearity and noise performance, includes first, second, third, and fourth transistors and an inductor. The first and second transistors are coupled as a first cascode pair, and the third and fourth transistors are coupled as a second cascode pair. The third transistor has its gate coupled to the source of the second transistor, and the fourth transistor has its drain coupled to the drain of the second transistor. The first transistor provides signal amplification. The second transistor provides load isolation and generates an intermediate signal for the third transistor. The third transistor generates distortion components used to cancel third order distortion component generated by the first transistor. The inductor provides source degeneration for the first transistor and improves distortion cancellation. The sizes of the second and third transistors are selected to reduce gain loss and achieve good linearity for the amplifier.

AN INTEGRATED CIRCUIT,AN AMPLIFIER AND A RECEIVER FOR A WIRELESS DEVICE

Documents:

Inventors:

PCT Conventions: