700 MHz to 1000 MHz RF Vector Modulator AD8340 FEATURES FUNCTIONAL BLOCK DIAGRAM VPRF QBBP QBBM VPS2 90 RFOP RFIP RFIM RFOM 0 CMOP IBBP IBBM DSOP 04699-001 Cartesian amplitude and phase modulation 700 MHz to 1.0 GHz frequency range Continuous magnitude control of -2 dB to -32 dB Continuous phase control of 0 to 360 Output third-order intercept 24 dBm Output 1 dB compression point 11 dBm Output noise floor -149 dBm/Hz @ full gain Adjustable modulation bandwidth up to 230 MHz Fast output power disable 4.75 V to 5.25 V single-supply voltage Figure 1. APPLICATIONS RF PA linearization/RF predistortion Amplitude and phase modulation Variable attenuators and phase shifters CDMA2000, GSM/EDGE linear power amplifiers Smart antennas GENERAL DESCRIPTION The AD8340 vector modulator performs arbitrary amplitude and phase modulation of an RF signal. Because the RF signal path is linear, the original modulation is preserved. This part can be used as a general-purpose RF modulator, a variable attenuator/phase shifter, or a remodulator. The amplitude can be controlled from a maximum of -2 dB to less than -32 dB, and the phase can be shifted continuously over the entire 360 range. For maximum gain, the AD8340 delivers an OP1dB of 11 dBm, an OIP3 of 24 dBm, and an output noise floor of -149 dBm/Hz, independent of phase. It operates over a frequency range of 700 MHz to 1.0 GHz. The baseband inputs in Cartesian I and Q format control the amplitude and phase modulation imposed on the RF input signal. Both I and Q inputs are dc-coupled with a 500 mV differential full-scale range. The maximum modulation bandwidth is 230 MHz, which can be reduced by adding external capacitors to limit the noise bandwidth on the control lines. Both the RF inputs and outputs can be used differentially or single-ended and must be ac-coupled. The RF input and output impedances are nominally 50 over the operating frequency range. The DSOP pin allows the output stage to be disabled quickly to protect subsequent stages from overdrive. The AD8340 operates off supply voltages from 4.75 V to 5.25 V while consuming approximately 130 mA. The AD8340 is fabricated using the Analog Devices, Inc. proprietary, high performance 25 GHz SOI complementary bipolar IC process. It is available in a 24-lead RoHS-compliant LFCSP package and operates over a -40C to +85C temperature range. Evaluation boards are available. Rev. B Information furnished by Analog Devices is believed to be accurate and reliable. However, no responsibility is assumed by Analog Devices for its use, nor for any infringements of patents or other rights of third parties that may result from its use. Specifications subject to change without notice. No license is granted by implication or otherwise under any patent or patent rights of Analog Devices. Trademarks and registered trademarks are the property of their respective owners. One Technology Way, P.O. Box 9106, Norwood, MA 02062-9106, U.S.A. Tel: 781.329.4700 www.analog.com Fax: 781.461.3113 (c)2004-2007 Analog Devices, Inc. All rights reserved. AD8340 TABLE OF CONTENTS Features .............................................................................................. 1 Noise and Distortion.................................................................. 11 Applications....................................................................................... 1 Gain and Phase Accuracy.......................................................... 11 Functional Block Diagram .............................................................. 1 RF Frequency Range .................................................................. 11 General Description ......................................................................... 1 Applications Information .............................................................. 12 Revision History ............................................................................... 2 Using the AD8340 ...................................................................... 12 Specifications..................................................................................... 3 RF Input and Matching ............................................................. 12 Absolute Maximum Ratings............................................................ 4 RF Output and Matching .......................................................... 13 ESD Caution.................................................................................. 4 Driving the I-Q Baseband Controls......................................... 13 Pin Configuration and Function Descriptions............................. 5 Interfacing to High Speed DACs.............................................. 14 Typical Performance Characteristics ............................................. 6 CDMA2000 Application ........................................................... 14 Theory of Operation ...................................................................... 10 Evaluation Board ............................................................................ 16 RF Quadrature Generator ......................................................... 10 Schematic and Artwork............................................................. 18 I-Q Attenuators and Baseband Amplifiers.............................. 11 Outline Dimensions ....................................................................... 20 Output Amplifier........................................................................ 11 Ordering Guide .......................................................................... 20 REVISION HISTORY 8/07--Rev. A to Rev B Replaced Pin Configuration and Function Descriptions Section ........................................................................ 5 Changes to Figure 30...................................................................... 12 Changes to Figure 39...................................................................... 18 7/07--Rev. 0 to Rev. A Replaced Pin Configuration and Function Descriptions Section ........................................................................ 5 Changes to Ordering Guide .......................................................... 20 6/04--Revision 0: Initial Version Rev. B | Page 2 of 20 AD8340 SPECIFICATIONS VS = 5 V, TA = 25C, ZO = 50 , f = 880 MHz, single-ended, ac-coupled source drive to RFIP through 5.6 nH series inductor, RFIM ac-coupled through 5.6 nH series inductor to common, differential-to-single-ended conversion at output using 1:1 balun. Table 1. Parameter OVERALL FUNCTION Frequency Range Maximum Gain Minimum Gain Gain Control Range Phase Control Range Gain Flatness Group Delay Flatness RF INPUT STAGE Input Return Loss CARTESIAN CONTROL INTERFACE (I and Q) Gain Scaling Modulation Bandwidth Second Harmonic Distortion Third Harmonic Distortion Step Response RF OUTPUT STAGE Output Return Loss f = 880 MHz Gain Output Noise Floor Output IP3 ACPR Output 1 dB Compression Point POWER SUPPLY Positive Supply Voltage Total Supply Current OUTPUT DISABLE Disable Threshold Maximum Attenuation Enable Response Time Disable Response Time Conditions Min Typ Max Unit 1000 -2 -32 30 360 0.25 10 MHz dB dB dB Degrees dB ps 20 dB 2 230 47 45 45 1/V MHz dBc dBc ns 47 ns 7.5 dB -2 -149 -147 24 62 dB dBm/Hz dBm/Hz dBm dBc 11 dBm 700 Maximum gain setpoint for all phase setpoints VBBI = VBBQ = 0 V Relative to maximum gain Over 30 dB control range Over any 60 MHz bandwidth Over any 60 MHz bandwidth RFIM, RFIP (Pin 21 and Pin 22) From RFIP to CMRF (with 5.6 nH series inductors) IBBP, IBBM, QBBP, QBBM (Pin 16, Pin 15, Pin 3, Pin 4) 250 mV p-p sinusoidal baseband input single-ended 250 mV p-p, 1 MHz, sinusoidal baseband input differential 250 mV p-p, 1 MHz, sinusoidal baseband input differential For gain setpoint from 0.1 to 0.9 (VBBP = 0.5 V, VBBM = 0.55 V to 0.95 V) For gain setpoint from 0.9 to 0.1 (VBBP = 0.5 V, VBBM = 0.95 V to 0.55 V) RFOP, RFOM (Pin 9 and Pin 10) Measured through balun Maximum gain setpoint Maximum gain setpoint, no input PIN = 0 dBm, frequency offset = 20 MHz f1 = 880 MHz, f2 = 877.5 MHz, maximum gain setpoint IS-95, single carrier, POUT = 0 dBm, maximum gain, phase setpoint = 45 Maximum gain VPS2 (Pin 5, Pin 6, Pin 14); RFOP, RFOM (Pin 9 and Pin 10) Includes load current DSOP (Pin 13) DSOP = 5 V Delay following high-to-low transition until device meets full specifications Delay following low-to-high transition until device produces full attenuation Rev. B | Page 3 of 20 4.75 110 5 130 5.25 150 V mA 2.5 40 15 V dB ns 10 ns AD8340 ABSOLUTE MAXIMUM RATINGS Stresses above those listed under Absolute Maximum Ratings may cause permanent damage to the device. This is a stress rating only; functional operation of the device at these or any other conditions above those indicated in the operational section of this specification is not implied. Exposure to absolute maximum rating conditions for extended periods may affect device reliability. Table 2. Parameter Supply Voltage VPRF, VPS2 DSOP IBBP, IBBM, QBBP, QBBM RFOP, RFOM RF Input Power at Maximum Gain (50 ) (RFIP or RFIM, Single-Ended Drive) Equivalent Voltage Internal Power Dissipation JA (with Pad Soldered to Board) Maximum Junction Temperature Operating Temperature Range Storage Temperature Range Lead Temperature (Soldering, 60 sec) Rating 5.5 V 5.5 V 2.5 V 5.5 V 13 dBm 2.8 V p-p 825 mW 59C/W 125C -40C to +85C -65C to +150C 300C ESD CAUTION Rev. B | Page 4 of 20 AD8340 24 23 22 21 20 19 VPRF CMRF RFIP RFIM CMRF VPRF PIN CONFIGURATION AND FUNCTION DESCRIPTIONS 1 2 3 4 5 6 PIN 1 INDICATOR AD8340 TOP VIEW (Not to Scale) 18 17 16 15 14 13 IFLP IFLM IBBP IBBM VPS2 DSOP 04699-002 CMOP CMOP RFOP RFOM CMOP CMOP 7 8 9 10 11 12 QFLP QFLM QBBP QBBM VPS2 VPS2 Figure 2. Pin Configuration Table 3. Pin Function Descriptions Pin No. 1; 2 Mnemonic QFLP; QFLM 3; 4 5, 6, 14; 19, 24 7, 8, 11, 12; 20, 23 9; 10 13 15; 16 17; 18 QBBP; QBBM VPS2; VPRF CMOP; CMRF RFOP; RFOM DSOP IBBM; IBBP IFLM; IFLP 21; 22 RFIM; RFIP Description Q Baseband Input Filter Pins. Connect optional capacitor to reduce Q baseband channel low-pass corner frequency. Q Channel Differential Baseband Inputs. Positive Supply Voltage, 4.75 V to 5.25 V. Device Common. Connect via lowest possible impedance to external circuit common. Differential RF Outputs. Must be ac-coupled. Differential impedance 50 nominal. Output Disable. Pull high to disable output stage. I Channel Differential Baseband Inputs. I Baseband Input Filter Pins. Connect optional capacitor to reduce I baseband channel low-pass corner frequency. Differential RF Inputs. Must be ac-coupled. Differential impedance 50 nominal. Rev. B | Page 5 of 20 AD8340 TYPICAL PERFORMANCE CHARACTERISTICS 0 0.4 PHASE SETPOINT = 0 0.2 PHASE SETPOINT = 270 -20 -25 -35 PHASE SETPOINT = 180 -40 0 0.1 0.2 0.3 0.4 0.5 0.6 GAIN SETPOINT 0.7 0.8 0.9 -1.2 -1.4 -1.6 0 180 225 270 315 360 330 300 PHASE SETPOINT = 0 270 PHASE SETPOINT = 315 PHASE SETPOINT = 270 -3 GAIN SETPOINT = 1.0 240 PHASE (Degrees) 0 210 180 GAIN SETPOINT = 0.1 150 GAIN SETPOINT = 0.5 120 90 PHASE SETPOINT = 180 -5 04699-004 PHASE SETPOINT = 225 -7 0 0.1 0.2 0.3 0.4 0.5 0.6 GAIN SETPOINT 0.7 0.8 0.9 04699-007 60 -6 30 0 1.0 Figure 4. Gain Conformance Error vs. Gain Setpoint at Different Phase Setpoints 0 30 60 90 120 150 180 210 240 270 300 330 360 PHASE SETPOINT (Degrees) Figure 7. Phase vs. Phase Setpoint at Different Gain Setpoints 6 0 -2 4 GAIN SETPOINT = 1.0 -4 PHASE ERROR (Degrees) -6 -8 GAIN (dB) 135 360 1 GAIN SETPOINT = 0.5 -10 -12 -14 -16 -18 -20 0 45 90 135 180 225 270 315 GAIN SETPOINT = 0.1 2 0 -2 GAIN SETPOINT = 1.0 -4 GAIN SETPOINT = 0.5 -6 -8 -10 04699-005 GAIN SETPOINT = 0.1 -22 -24 90 Figure 6. Gain Conformance Error vs. Phase Setpoint at Different Gain Setpoints PHASE SETPOINT = 90 -4 45 PHASE SETPOINT (Degrees) 2 -2 GAIN SETPOINT = 0.1 -2.0 1.0 PHASE SETPOINT = 45 -1 GAIN SETPOINT = 0.5 -1.0 PHASE SETPOINT = 135 3 GAIN CONFORMANCE ERROR (dB) -0.8 -1.8 Figure 3. Gain Magnitude vs. Gain Setpoint at Different Phase Setpoints, RF Frequency = 880 MHz 4 -0.6 04699-003 -30 -0.4 -12 360 PHASE SETPOINT (Degrees) Figure 5. Gain Magnitude vs. Phase Setpoint at Different Gain Setpoints Rev. B | Page 6 of 20 04699-008 GAIN (dB) -15 -0.2 04699-006 PHASE SETPOINT = 90 -10 GAIN SETPOINT = 1.0 0.0 GAIN CONFORMANCE ERROR (dB) -5 0 45 90 135 180 225 270 PHASE SETPOINT (Degrees) 315 360 Figure 8. Phase Error vs. Phase Setpoint at Different Gain Setpoints AD8340 -142 0 -143 -0.5 -145 GAIN FLATNESS (dB) RF PIN = +5dBm -146 -147 RF PIN = -5dBm -148 RF PIN = 0dBm -149 -1.0 -1.5 -150 -2.0 -151 04699-009 NO RF INPUT -152 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 04699-012 NOISE FLOOR (dBm/Hz) -144 -2.5 700 1.0 750 800 GAIN SETPOINT Figure 9. Output Noise Floor vs. Gain, Noise in dBm/Hz, No Carrier, With Carrier (20 MHz Offset) PIN = -5 dBm, 0 dBm, and +5 dBm -10 -12 -14 -16 -18 GAIN SETPOINT = 0.1 04699-010 -20 -22 750 800 850 900 950 -30 -40 -50 -70 -90 200 300 400 500 600 700 800 900 1000 DIFFERENTIAL BASEBAND INPUT LEVEL (mV p-p) (I OR Q CHANNEL DRIVEN AT 1MHz) Figure 10. Gain vs. Frequency at Different Gain Setpoints (700 MHz to 1000 MHz), Phase Setpoint = 0 Figure 13. Baseband Harmonic Distortion (I and Q Channel, RF Input = 0 dBm, Balun and Cable Losses of Approximately 2 dB Not Accounted for in Plot) 14 -146 12 -147 10 OP1dB (dBm) -145 -148 -149 2 04699-011 -151 0 700 1000 Figure 11. Output Noise Floor vs. Frequency, Maximum Gain, No RF Carrier, Phase Setpoint = 0 TEMP = -40C TEMP = +85C 6 4 950 TEMP = +25C 8 -150 800 850 900 RF FREQUENCY (MHz) THIRD BASEBAND HARMONIC PRODUC T, 877MHz, 883MHz -80 FREQUENCY (MHz) 750 SECOND BASEBAND HARMONIC PRODUC T, 878MHz, 882MHz -60 -100 100 1000 FUNDAMENTAL POWER, 879MHz, 881MHz 04699-013 GAIN SETPOINT = 0.5 -20 04699-014 GAIN (dB) -8 -10 RF OUTPUT AM SIDEBAND POWER (dBm) GAIN SETPOINT = 1.0 -6 NOISE (dBm/Hz) 1000 0 -4 -152 700 950 Figure 12. Gain Flatness vs. Frequency, Maximum Gain, Phase Setpoint = 0 0 -2 -24 700 850 900 FREQUENCY (MHz) 750 800 850 900 FREQUENCY (MHz) 950 1000 Figure 14. Output 1 dB Compression Point vs. Frequency and Temperature, Maximum Gain, Phase Setpoint = 0 Rev. B | Page 7 of 20 AD8340 30 30 28 TEMP = -40C 24 GAIN SETPOINT = 0.5 20 OIP3 (dBm) 22 TEMP = +85C 20 TEMP = +25C 18 15 10 16 14 GAIN SETPOINT = 0.1 04699-015 5 12 10 700 750 800 850 900 FREQUENCY (MHz) 950 0 1000 Figure 15. Output IP3 vs. Frequency and Temperature, Maximum Gain, I Only 04699-018 OIP3 (dBm) GAIN SETPOINT = 1.0 25 26 0 45 90 135 180 225 270 PHASE SETPOINT (Degrees) 315 360 Figure 18. Output IP3 vs. Gain and Phase Setpoints, 2.5 MHz Carrier Spacing 0 0 RBW 30kHz VBW 30kHz SWT 100ms REF LVL 0 dBm 20dB RF ATT MIXER -10dBm UNIT dBm A -5 -60 -70 -80 04699-016 -25 200mV p-p BB INPUT -30 0 50 100 150 200 250 FREQUENCY (MHz) 300 350 -90 -100 400 CENTER 880 MHz 500 kHz/DIV 1 RM 04699-019 -50 UNDESIRED SIDEBAND -20 -40 RF FEEDTHROUGH 500mV p-p BB INPUT -15 -30 DESIRED SIDEBAND OUTPUT POWER (dBm) -10 SECOND BASEBAND HARMONIC -20 1V p-p BB INPUT SECOND BASEBAND HARMONIC RF OUTPUT AM SIDEBAND POWER (dBm) -10 SPAN 5 MHz FREQUENCY (MHz) Figure 16. I/Q Modulation Bandwidth vs. Baseband Magnitude Figure 19. Single-Sideband Performance, 880 MHz, -10 dBm RF Input; 1 MHz, 500 mV p-p Differential BB Drive 90 60 120 14 GAIN SETPOINT = 1.0 12 30 150 10 8 GAIN SETPOINT = 0.5 180 0 1.5GHz 4 500MHz 2 1.5GHz 0 210 330 -2 500MHz -4 -6 240 GAIN SETPOINT = 0.1 04699-017 -8 -10 -12 0 45 90 225 180 270 135 PHASE SETPOINT (Degrees) 315 300 270 IMPEDANCE CIRCLE S11 RF PORT WITH 5.6nH INDUCTORS S11 RF PORT WITHOUT INDUCTORS 360 Figure 17. Output 1dB Compression Point vs. Gain and Phase Setpoints Rev. B | Page 8 of 20 Figure 20. Input and Output Impedance Smith Chart (with Frequency Markers) 04699-020 OP1dB (dBm) 6 AD8340 90 60 120 150 1.5GHz 0 30 500MHz -5 RF OUTPUT POWER (dBm) -10 0 180 500MHz 1.5GHz 210 330 300 -20 -25 -30 -35 -40 270 04699-024 240 -15 -45 -50 04699-021 IMPEDANCE CIRCLE S22 PORT WITH 1 TO 1 TRANSFORMER SDD22 PORT DIFFERENTIAL Figure 21. Output Impedance Smith Chart (with Frequency Markers) 0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 DSOP VOLTAGE (V) 4.0 4.5 5.0 Figure 24. Power Shutdown Attenuation TEK FAST ACQ SAMPLE 6 4 PHASE SETPOINT = 45 2V/DIV PHASE SETPOINT = 0 CHA1/CHA3 (V) 0 -2 -4 -6 DSOP 200mV/DIV 3 -8 PHASE SETPOINT = 90 -10 -14 0 0.1 0.2 0.3 0.4 0.5 0.6 GAIN SETPOINT 0.7 0.8 0.9 CH1 200mV CH3 2.0V 1.0 134 133 132 5.25V 131 130 5V 128 127 0 10 20 30 04699-023 4.75V 125 -40 -30 -20 -10 ET 200ps/pt 74.0ns TIME (10ns/DIV) 135 126 M 10.0ns 5.0GS/s A CH2 160mV Figure 25. Power Shutdown Response Time Figure 22. Phase Error vs. Gain Setpoint by Phase Setpoint, 5 V DC, 25C, 880 MHz 129 04699-025 04699-022 RF OUTPUT -12 SUPPLY CURRENT (mA) PHASE ERROR (Degrees) 2 40 50 60 70 80 TEMPERATURE (C) Figure 23. Supply Current vs. Temperature Rev. B | Page 9 of 20 AD8340 THEORY OF OPERATION VBBI I CHANNEL INPUT LINEAR ATTENUATOR V-I SINGLE-ENDED OR DIFFERENTIAL 50 INPUT Z (VBBI / VO ) ( ( + V BBQ / VO PhaseSP = arctan VBBQ /VBBI ) OUTPUT DISABLE LINEAR ATTENUATOR Q CHANNEL INPUT VBBQ Figure 26. Simplified Architecture of the AD8340 Vq MAX GAIN = 0dB +0.5 A |A| -0.5 +0.5 MIN GAIN < -30dB -0.5 The correspondence between the desired gain and phase setpoints, GainSP and PhaseSP, and the Cartesian inputs, VBBI and VBBQ, is given by simple trigonometric identities. Gain SP = SINGLE-ENDED OR DIFFERENTIAL 50 OUTPUT I-V V-I A change in sign of VBBI or VBBQ can be viewed as a change in sign of the gain or as a 180 phase change. The outermost circle represents the maximum gain magnitude of unity. The circle origin implies, in theory, a gain of 0. In practice, circuit mismatches and unavoidable signal feedthrough limit the minimum gain to approximately -40 dB. The phase angle between the resultant gain vector and the positive x-axis is defined as the phase shift. Note that there is a nominal, systematic insertion phase through the AD8340 to which the phase shift is added. In the following discussions, the systematic insertion phase is normalized to 0. 2 0/90 04699-026 By controlling the relative amounts of I and Q components that are summed, the AD8340 allows continuous magnitude and phase control of the gain. Consider the vector gain representation of the AD8340 expressed in polar form in Figure 27. The attenuation factors for the I and Q signal components are represented on the x- and y-axis, respectively, by the baseband inputs, VBBI and VBBQ. The resultant vector sum represents the vector gain, which can also be expressed as a magnitude and phase. By applying different combinations of baseband inputs, any vector gain within the unit circle can be programmed. Pure amplitude modulation is represented by radial movement of the gain vector tip at a fixed angle, while pure phase modulation is represented by rotation of the tip around the circle at a fixed radius. Unlike traditional I-Q modulators, the AD8340 is designed to have a linear RF signal path from input to output. Traditional I-Q modulators provide a limited LO carrier path through which any amplitude information is removed. Vi 04699-027 The AD8340 is a linear RF vector modulator with Cartesian baseband controls. In the simplified block diagram shown in Figure 26, the RF signal propagates from the left to the right while baseband controls are placed above and below. The RF input is first split into in-phase (I) and quadrature (Q) components. The variable attenuators independently scale the I and Q components of the RF input. The attenuator outputs are then summed and buffered to the output. Figure 27. Vector Gain Representation RF QUADRATURE GENERATOR 2 ) where: VO is the baseband scaling constant (500 mV). VBBI and VBBQ are the differential I and Q baseband voltages, respectively. Note that when evaluating the arctangent function, the proper phase quadrant must be selected. For example, if the principal value of the arctangent (known as the arctangent(x)) is used, Quadrant 2 and Quadrant 3 would be interpreted mistakenly as Quadrant 4 and Quadrant 1, respectively. In general, both VBBI and VBBQ are needed in concert to modulate the gain and the phase. The RF input is directly coupled differentially or single-endedly to the quadrature generator, which consists of a multistage RC polyphase network tuned over the operating frequency range of 700 MHz to 1000 MHz. The recycling nature of the polyphase network generates two replicas of the input signal, which are in precise quadrature, that is, 90, to each other. Because the passive network is perfectly linear, the amplitude and phase information contained in the RF input is transmitted faithfully to both channels. The quadrature outputs are then separately buffered to drive the respective attenuators. The characteristic impedance of the polyphase network is used to set the input impedance to the AD8340. Rev. B | Page 10 of 20 AD8340 I-Q ATTENUATORS AND BASEBAND AMPLIFIERS GAIN AND PHASE ACCURACY The proprietary linear-responding attenuator structure is an active solution with differential inputs and outputs that offer excellent linearity, low noise, and greater immunity from mismatches than other variable attenuator methods. The gain, in linear terms, of the I and Q channels is proportional to its control voltage with a scaling factor designed to be 2/V, that is, a full-scale gain setpoint of 1.0 (-2 dB) for VBBI (Q) of 500 mV. The control voltages can be driven differentially or single-endedly. The combination of the baseband amplifiers and attenuators allows for maximum modulation bandwidths in excess of 200 MHz. There are numerous ways to express the accuracy of the AD8340. Ideally, the gain and phase should precisely follow the setpoints. Figure 4 illustrates the gain error in decibels (dB) from a best fit line, normalized to the gain measured at the gain setpoint = 1.0, for the different phase setpoints. Figure 6 shows the gain error in a different form; the phase setpoint is swept from 0 to 360 for different gain setpoints. Figure 8 and Figure 22 show analogous errors for the phase error as a function of gain and phase setpoints. The accuracy clearly depends on the region of operation within the vector gain unit circle. Operation very close to the origin generally results in larger errors as the relative accuracy of the I and Q vectors degrades. OUTPUT AMPLIFIER The output amplifier accepts the sum of the attenuator outputs and delivers a differential output signal into the external load. The output pins must be pulled up to an external supply, preferably through RF chokes. When the 50 load is taken differentially, an OP1dB of 11 dBm and OIP3 of 24 dBm are achieved at 880 MHz. The output can be taken in single-ended fashion, albeit at lower performance levels. NOISE AND DISTORTION The output noise floor and distortion levels vary with the gain magnitude but do not vary significantly with the phase. At the higher gain magnitude setpoints, the OIP3 and the noise floor vary in direct proportion with the gain. At lower gain magnitude setpoints, the noise floor levels off while the OIP3 continues to vary with the gain. RF FREQUENCY RANGE The frequency range on the RF input is limited by the internal polyphase quadrature phase-splitter. The phase-splitter splits the incoming RF input into two signals, 90 out of phase, as previously described in the RF Quadrature Generator section. This polyphase network has been designed to ensure robust quadrature accuracy over standard fabrication process parameter variations for the 700 MHz to 1 GHz specified RF frequency range. Using the AD8340 as a single-sideband modulator and measuring the resulting sideband suppression is a good gauge of how the quadrature accuracy is maintained over RF frequency. A typical plot of sideband suppression from 500 MHz to 1.5 GHz is shown in Figure 28. The level of sideband suppression degradation outside the 700 MHz to 1 GHz specified range is subject to manufacturing process variations. 0 SB SUPPRESSION (dBc) -5 -10 -15 -20 -25 -35 500 04699-028 -30 600 700 800 900 1000 1100 1200 1300 1400 1500 FREQUENCY (MHz) Figure 28. Sideband Suppression vs. Frequency Rev. B | Page 11 of 20 AD8340 APPLICATIONS INFORMATION RFIM, a 50 match is achieved with a return loss of >10 dB over the operating frequency range. Different matching inductors can improve matching over a narrower frequency range. The single-ended and differential input impedances are exactly the same. The AD8340 is designed to operate in a 50 impedance system. Figure 30 illustrates where the RF input is driven in a single-ended fashion while the differential RF output is converted to a single-ended output with an RF balun. The baseband controls for the I and Q channels are typically driven from differential DAC outputs. The power supplies, VPRF and VPS2, should be bypassed appropriately with 0.1 F and 100 pF capacitors. Low inductance grounding of the CMOP and CMRF common pins is essential to prevent unintentional peaking of the gain. 100pF 5.6nH 100pF 5.6nH RF RFIM ~1VDC RC PHASE RFIP 04699-029 USING THE AD8340 50 Figure 29. RF Input Interface to the AD8340 Showing Coupling Capacitors and Matching Inductors RF INPUT AND MATCHING The RFIP and RFIM should be ac-coupled through low loss series capacitors as shown in Figure 29. The internal dc levels are at approximately 1 V. For single-ended operation, one input is driven by the RF signal and the other input is ac grounded. The input impedance of the AD8340 is defined by the characteristics of the polyphase network. The capacitive component of the network causes its impedance to roll off with frequency, albeit at a slower rate than 6 dB/octave. With matching inductors on the order of 5.6 nH in series with each of the RF inputs, RFIP and VP C2 100pF C1 0.1F IBBM VP IBBP C12 (SEE TEXT) VPS2 RFIM RFOM RFOP CMRF CMOP VPRF C4 100pF QFLP C11 (SEE TEXT) VPS2 RFIP QBBM L4 5.6nH QBBP C5 100pF C3 0.1F QBBM CMOP C17 100pF AD8340 VP QBBP DSOP CMOP CMRF QFLM RF INPUT L3 5.6nH OUTPUT DISABLE L1 120nH CMOP VPS2 C10 0.1F C9 100pF Figure 30. Basic Connections Rev. B | Page 12 of 20 ETC1-1-13 RF OUTPUT C18 L2 100pF 120nH C14 0.1F VP 04699-0-030 C6 100pF IFLP VPRF IBBM VP IBBP C7 100pF B IFLM C8 0.1F A AD8340 RF OUTPUT AND MATCHING The RF outputs of the AD8340, RFOP, and RFOM, are open collectors of a transimpedance amplifier that needs to be pulled up to the positive supply, preferably with RF chokes, as shown in Figure 31. The nominal output impedance looking into each individual output pin is 25 . Consequently, the differential output impedance is 50 . VP RFOM ISIG GM 100pF 1:1 100pF RF OUTPUT 50 DIFFERENTIAL 04699-031 RFOP RT DRIVING THE I-Q BASEBAND CONTROLS The I and Q inputs to the AD8340 set the gain and phase between input and output. These inputs are differential and should normally have a common-mode level of 0.5 V. However, when differentially driven, the common mode can vary from 250 mV to 750 mV while still allowing full gain control. Each input pair has a nominal input swing of 0.5 V differential around the common-mode level. The maximum gain of unity is achieved if the differential voltage is equal to +500 mV or -500 mV. Therefore, with a common-mode level of 500 mV, IBBP and IBBM each swings between 250 mV and 750 mV. 120nH RT The RF output signal can be disabled by raising the DSOP pin to the positive supply. The shutdown function provides >40 dB attenuation of the input signal even at full gain. The interface to DSOP is high impedance, and the shutdown and turn-on response times are <100 ns. If the disable function is not needed, the DSOP should be tied to ground. Figure 31. RF Output Interface to the AD8340 Showing Coupling Capacitors, Pull-Up RF Chokes, and Balun Because the output dc levels are at the positive supply, ac coupling capacitors are usually needed between the AD8340 outputs and the next stage in the system. A 1:1 RF broadband output balun, such as the ETC1-1-13 (M/A-COM), converts the differential output of the AD8340 into a single-ended signal. Note that the loss and balance of the balun directly impact the apparent output power, noise floor, and gain/phase errors of the AD8340. In critical applications, narrow-band baluns with low loss and superior balance are recommended. If the output is taken in a single-ended fashion directly into a 50 load through a coupling capacitor, there is an impedance mismatch. This can be resolved with a 1:2 balun to convert the single-ended 25 output impedance to 50 . If loss-of-signal swing is not critical, a 25 back termination in series with the output pin can also be used. The unused output pin must still be pulled up to the positive supply. The user can load it through a coupling capacitor with a dummy load to preserve balance. The gain of the AD8340 when the output is single-ended varies slightly with the dummy load value, as shown in Figure 32. The I and Q inputs can also be driven with a single-ended signal. In this case, one side of each input should be tied to a low noise 0.5 V voltage source (a 0.1 F decoupling capacitor located close to the pin is recommended), while the other input swings from 0 V to 1 V. Differential drive generally offers superior even-order distortion and lower noise than singleended drive. The bandwidth of the baseband controls exceeds 200 MHz even at full-scale baseband drive. This allows for very fast gain and phase modulation of the RF input signal. In cases where lower modulation bandwidths are acceptable or desired, external filter capacitors can be connected across Pin IFLP to Pin IFLM, and across Pin QFLP to Pin QFLM, to reduce the ingress of baseband noise and spurious signal into the control path. The 3 dB bandwidth is set by choosing CFLT according to the following equation: f 3 dB RL2 = SHORT -1.5 -2.0 GAIN (dB) -2.5 -3.0 RL2 = 50 -3.5 -4.0 -4.5 -5.5 RL2 = OPEN 800 900 FREQUENCY (MHz) 04699-032 -5.0 -6.0 700 C EXTERNAL + 0.5 pF This equation has been verified for values of CFLT from 10 pF to 0.1 F (bandwidth settings of approximately 4.5 kHz to 43 MHz). -0.5 -1.0 45 kHz x 10 nF 1000 Figure 32. Gain of the AD8340 Using a Single-Ended Output with Different Dummy Loads, RL2 on the Unused Output Rev. B | Page 13 of 20 AD8340 The AD977x family of dual DACs is well suited for driving the I and Q vector controls of the AD8340. While these inputs can in general be driven by any DAC, the differential outputs and bias level of the Analog Devices TxDAC(R) family allow for a direct connection between DAC and modulator. The AD977x family of dual DACs has differential current outputs. The full-scale current is user programmable and is usually set to 20 mA, that is, each output swings from 0 mA to 20 mA. The basic interface between the AD9777 DAC outputs and the AD8340 I and Q inputs is shown in Figure 33. The resistors R1 and R2 set the dc bias level according to the equation: 1.15 1.13 1.10 1.08 1.05 1.02 1.00 0.97 0.95 0.92 0.90 0.88 0.85 0.82 0.80 0.77 0.75 0.72 0.70 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120 125 130 R3 04699-034 DIFFERENTIAL PEAK-TO-PEAK SWING (R3) INTERFACING TO HIGH SPEED DACs Figure 34. Peak-to-Peak DAC Output Swing vs. Swing Scaling Resistor R3 (R1 = R2 = 50 ) Bias Level = Average Output Current x R1 For example, if the full-scale current from each output is 20 mA, each output will have an average current of 10 mA. Therefore, to set the bias level to the recommended 0.5 V, R1 and R2 should be set to 50 each. R1 and R2 should always be equal. Figure 34 shows the relationship between the value of R3 and the peak baseband voltage with R1 and R2 equal to 50 . Figure 34 shows that a value of 100 for R3 provides a peak-to-peak swing of 1 V p-p differential into the AD8340 I and Q inputs. If R3 is omitted, this results in an available swing from the DAC of 2 V p-p differential, which is twice the maximum voltage range required by the AD8340. DAC resolution can be maximized by adding R3, which scales down this voltage according to the following equation: When using a DAC, low-pass image reject filters are typically used to eliminate the Nyquist images produced by the DAC. They also provide the added benefit of eliminating broadband noise that might feed into the modulator from the DAC. Full - Scale Swing = 2 x I MAX ( To test the compliance to the CDMA2000 base station standard, a single-carrier CDMA2000 test model signal (forward pilot, sync, paging, and six traffic as per 3GPP2 C.S0010-B, Table 6.5.2.1) was applied to the AD8340. A cavity-tuned filter was used to reduce noise from the signal source being applied to the device. The 4.6 MHz pass band of this filter is apparent in the subsequent spectral plots (see Figure 35 to Figure 38). ) R2 R1 (R2 + R3) x 1 - R2 + R3 AD9777 AD8340 IBBP IOUTA1 R1 R2 OPTIONAL LOW-PASS FILTER R3 IOUTB1 Figure 35 shows a plot of the spectrum of the output signal under nominal conditions. POUT is equal to -5 dBm and VI = VQ = 0.353 V, that is, VIBBP - VIBBM = VQBBP - VQBBM = 0.353 V. Adjacent channel power is measured in 30 kHz resolution bandwidth at 750 kHz and 1.98 MHz carrier offset. Noise floor is measured at 4 MHz carrier offset. IBBM IOUTA2 QBBP R1 R2 OPTIONAL LOW-PASS FILTER R3 IOUTB2 CDMA2000 APPLICATION 04699-033 QBBM Figure 33. Basic AD9777 to AD8340 Interface Rev. B | Page 14 of 20 *RBW 30kHz *VBW 30kHz *SWT 100ms OFFSET 0.5 dB CH PWR ACP LOW ACP UP ALT1 LOW ALT1 UP -20 -30 10 MARKER 2 [T1 NOI] -148.76dBm/Hz 876.009615385MHz 0 -5.17dBm -60.94dB -60.08dB -86.40dB -86.80dB LVL MARKER 1 [T1 NOI] -148.89dBm/Hz 884.006410256MHz -40 -50 NOR -60 -70 SWP 50 OF 50 -80 04699-035 1 2 -110 CENTER 880MHz 1MHz/DIV -30 -40 -40 -50 -50 -60 -60 ACP - 750kHz OFFSET, 30kHz RBW -70 -70 ACP - 1.98MHz OFFSET, 30kHz RBW -80 -80 -90 -90 NOISE - 4MHz OFFSET, 1MHz RBW -25 -20 -15 -10 -5 OUTPUT POWER (dBm) 0 5 -100 04699-036 -30 NOISE - dBm @ 4MHz CARRIER OFFSET (1MHz RBW) ACP - dBc (30kHz RBW) ACP - 750kHz OFFSET, 30kHz RBW -40 -75 ACP - 1.98MHz OFFSET, 30kHz RBW -80 -50 NOISE - 4MHz OFFSET, 1MHz RBW 0 50 100 150 350 400 450 -90 500 In contrast to Figure 36, Figure 37 shows that for a fixed input power, ACP remains fairly constant as gain and phase are changed (this is not true for very high input powers). The noise floor still drops with decreasing gain, but it never reaches the -90 dBm level shown in Figure 37. Figure 38 shows the output spectrum for a 3-carrier CDMA2000 spectrum. Again, the signal being applied to the AD8340 is filtered by a cavity-tuned filter with a -3 dB bandwidth of 4.6 MHz. To reduce distortion, the total output carrier power was reduced to approximately -8 dBm (per-carrier power = -12.6 dBm). Adjacent channel power ratios of -61 dBc (2 MHz from center of spectrum) and -82 dBc (3.23 MHz from center of spectrum) were measured. The noise floor, measured at 5.25 MHz carrier offset, is approximately -149 dBm/Hz (-89 dBm in a 1 MHz bandwidth). While some dynamic range is lost due to output power back-off, ACP stays approximately equal and noise floor improves slightly. REF -15 dBm -20 With a fixed input power of 2.4 dBm, the output power was again swept by exercising the I and Q inputs. VI and VQ were kept equal and were swept from 10 mV to 500 mV. The resulting output power, ACP, and noise floor are shown in Figure 37. 200 250 300 VI = VQ =VIN (mV) -85 Figure 37. Output Power, Noise, and ACP vs. I and Q Control Voltages, CDMA2000 Test Model, VI = VQ, ACP Measured in 30 kHz RBW at 750 kHz and 1.98 MHz Carrier Offset, Noise Measured at 4 MHz Carrier Offset Figure 36. Noise and ACP vs. Output Power, Single-Carrier CDMA2000 Test Model, VI = VQ = 0.353, ACP Measured in 30 kHz RBW at 750 kHz and 1.98 MHz Carrier Offset, Noise Measured at 4 MHz Carrier Offset The results show that at an output power of 3 dBm, ACP is still in compliance with the standard (<-45 dBc @ 750 MHz and <-60 dBc @ 1.98 MHz). At low output power levels, ACP at 1.98 MHz, carrier offset degrades as the noise floor of the AD8340 becomes the dominant contributor to measured ACP. Measured noise at 4 MHz carrier offset begins to increase sharply above 0 dBm output power. This increase is not due to noise. but results from increased carrier-induced distortion. As output power drops below 0 dBm, the noise floor drops towards -90 dBm. -65 -70 -30 SPAN 10MHz Holding the I and Q control voltages steady at 0.353 V, input power was swept. Figure 36 shows the resulting output power, noise floor, and adjacent channel power ratio. Noise floor is presented as noise in a 1 MHz bandwidth as defined by the 3GPP2 specification. -100 -30 -60 -20 -70 Figure 35. Output Spectrum, Single-Carrier CDMA2000 Test Model at -5 dBm VI = VQ = 0.353 V, ACP Measured at 750 kHz and 1.98 MHz Carrier Offset, Noise Measured at 4 MHz Carrier Offset, Input Signal Filtered Using a Cavity Tuned Filter (Pass Band = 4.6 MHz) -55 -10 -60 -90 -100 -50 POUT vs. VIN 04699-037 *ATT 5dB *ATT 5dB *RBW 30kHz *VBW 300kHz *SWT 5s OFFSET 0.5dB MARKER 1 [T1 NOI] -148.83dBm/Hz 885.252403846MHz CH1 CH2 CH3 TOTAL ACP LOW ACP UP ALT1 LOW ALT1 UP -30 -40 -50 -12.65dBm -12.58dB A -12.87dB SOL -7.93dB -61.41dB -61.87dB LVL -82.36dB -81.92dB NOR -60 -70 -80 -90 1 -100 -110 CENTER 880MHz 1.5MHz/ 04699-038 REF -12dBm OUTPUT POWER (dBm) BS, 1X, C0 : ADJ CHANNE L ACP - dBc (30kHz RBW) NOISE - 4MHz CARRIER OFFSET - dBm (1MHz RBW) AD8340 SPAN 15MHz Figure 38. Output Spectrum, 3-Carrier CDMA2000 Test Model at -12.5 dBm/Carrier, VI = VQ = 0.353 V, ACP Measured at 2 MHz and 3.23 kHz Offset from Center of Spectrum, Noise Measured at 5.25 MHz Carrier Offset, Input Signal Filtered Using a Cavity-Tuned Filter (Pass Band = 4.6 MHz) Rev. B | Page 15 of 20 AD8340 EVALUATION BOARD The evaluation board circuit schematic for the AD8340 is shown in Figure 39. The evaluation board is configured to be driven from a single-ended 50 source. Although the input of the AD8340 is differential, it may be driven single-ended with no loss of performance. The low-pass corner frequency of the baseband I and Q channels can be reduced by installing capacitors in the C11 and C12 positions. The low-pass corner frequency for either channel is approximated by f 3 dB 45 kHz x 10 nF C EXTERNAL + 0.5 pF On the evaluation board, the I and Q baseband circuits are identical, so the following description applies equally to each. The connections and circuit configuration for the Q baseband inputs are described in Table 4. The baseband input of the AD8340 requires a differential voltage drive. The evaluation board is set up to allow such a drive by connecting the differential voltage source to QBBP and QBBM. The common-mode voltage should be maintained at approximately 0.5 V. For this configuration, Jumper W1 to Jumper W4 should be removed. The baseband input of the evaluation board can also be driven with a single-ended voltage. In this case, a bias level is provided to the unused input from Potentiometer R10 by installing either W1 or W2. Setting SW1 in Position B disables the AD8340 output amplifier. With SW1 set to Position A, the output amplifier is enabled and an external voltage signal, such as a pulse, can be applied to the DSOP SMA connector to exercise the output amplifier enable/disable function. Rev. B | Page 16 of 20 AD8340 Table 4. Evaluation Board Configuration Options Components R7, R9, R11, R14, R15, R19, R20, R21, C15, C19, W3, W4 Description I Channel Baseband Interface. Resistor R7 and Resistor R9 can be installed to accommodate a baseband source that requires a specific terminating impedance. C15 and C19 are bypass capacitors. For single-ended baseband drive, Potentiometer R11 can be used to provide a bias level to the unused input (install either W3 or W4). R1, R3, R10, R12, R13, R16, R17, R18, C16, C20, W1, W2 Q Channel Baseband Interface. See the I Channel Baseband Interface description. C11, C12 Baseband Low-Pass Filtering. By adding Capacitor C11 between QFLP and QFLM, and Capacitor C12 between IFLP and IFLM, the 3 dB low-pass corner frequency of the baseband interface can be reduced from 230 MHz (nominal). See the equation in the Evaluation Board section. Output Interface. The 1:1 balun transformer, T1, converts the 50 differential output to 50 single-ended. C17 and C18 are dc blocks. L1 and L2 provide dc bias for the output. Input Interface. The input impedance of the AD8340 requires 5.6 nH inductors in series with RFIP and RFIM for optimum return loss when driven by a single-ended 50 line. C5 and C6 are dc blocks. Supply Decoupling. T1, C17, C18, L1, L2 L3, L4, C5, C6 C2, C4, C7, C9, C14, C1, C3, C8, C10, R2, R4, R5, R6 R8, SW1 Output Disable Interface. The output stage of the AD8340 is disabled by applying a high voltage to the DSOP pin by moving SW1 to Position B. The output stage is enabled by moving SW1 to Position A. The output disable function can also be exercised by applying an external high or low voltage to the DSOP SMA connector with SW1 in Position A. Rev. B | Page 17 of 20 Default Conditions R7, R9 = open R11 = potentiometer, 2 k, 10 turns (Bourns) R14 = 4 k (Size 0603) R15 = 44 k (Size 0603) R19, R20, R21 = 0 (Size 0603) C15, C19 = 0.1 F (Size 0603) W3 = jumper (installed) W4 = jumper (open) R1, R3 = open R10 = potentiometer, 2 k, 10 turns (Bourns) R12 = 4 k (Size 0603) R13 = 44 k (Size 0603) R16, R17, R18 = 0 (Size 0603) C16, C20 = 0.1 F (Size 0603) W1 = jumper (installed) W2 = jumper (open) C11, C12 = open C17, C18 = 100 pF (Size 0603) T1 = ETC1-1-13 (M/A-COM) L1, L2 = 120 nH (Size 0603) L3, L4 = 5.6 nH (Size 0402) C5, C6 = 100 pF (Size 0603) C2, C4, C7, C9 = open (Size 0603) C1, C3, C8, C10, C14 = 0.1 F (Size 0603) R2, R4, R5, R6 = 0 (Size 0603) R8 = 10 k (Size 0603) SW1 = SPDT (Position A, output enabled) AD8340 SCHEMATIC AND ARTWORK IBBP IBBM C19 R7 0.1F (OPEN) R9 (OPEN) C2 (OPEN) R20 0 C15 0.1F R14 4k GND TEST POINT R19 0 W3 W4 R21 0 VP TEST POINT R11 2k R2 0 VS R15 44k C1 0.1F C12 (OPEN) C6 100pF VPS2 IBBM IFLP VPRF IBBP VP C8 0.1F R5 0 IFLM C7 (OPEN) SW1 CMOP CMRF L3 5.6nH R8 10k DSOP CMOP B A DSOP C17 100pF RFOM RFIM T1 ETC1-1-13 M/A-COM RFOP AD8340 CMRF CMOP VPRF C3 0.1F QFLP C11 (OPEN) L1 120nH CMOP C14 0.1F VPS2 R6 0 C10 0.1F R12 4k R10 2k R13 44k C18 L2 100pF 120nH C9 (OPEN) VP VS C16 0.1F W2 R17 0 R1 OPEN QBBP W1 R16 0 R18 0 R3 C20 OPEN 0.1F QBBM Figure 39. Evaluation Board Schematic Rev. B | Page 18 of 20 04699-039 R4 0 RFOP VPS2 C4 (OPEN) RFIP QBBM VP L4 5.6nH QBBP C5 100pF QFLM RFIN 04699-040 04699-0-041 AD8340 Figure 41. Component Side Silkscreen Figure 40. Component Side Layout Rev. B | Page 19 of 20 AD8340 OUTLINE DIMENSIONS 0.60 MAX 4.00 BSC SQ PIN 1 INDICATOR 0.60 MAX TOP VIEW 3.75 BSC SQ 0.50 BSC 0.50 0.40 0.30 1.00 0.85 0.80 12 MAX PIN 1 INDICATOR 19 18 24 1 2.25 2.10 SQ 1.95 BOTTOM VIEW 13 12 7 6 0.25 MIN 2.50 REF 0.80 MAX 0.65 TYP 0.05 MAX 0.02 NOM SEATING PLANE 0.30 0.23 0.18 0.20 REF COPLANARITY 0.08 COMPLIANT TO JEDEC STANDARDS MO-220-VGGD-2 Figure 42. 24-Lead Lead Frame Chip Scale Package [LFCSP_VQ] 4 mm x 4 mm Body, Very Thin Quad (CP-24-1) Dimensions shown in millimeters ORDERING GUIDE Model AD8340ACPZ-WP1, 2 AD8340ACPZ-REEL71 AD8340-EVALZ1 1 2 Temperature Range -40C to +85C -40C to +85C Package Description 24-Lead Lead Frame Chip Scale Package [LFCSP_VQ] 24-Lead Lead Frame Chip Scale Package [LFCSP_VQ] Evaluation Board Z = RoHS Compliant Part. WP = Waffle pack. (c)2004-2007 Analog Devices, Inc. All rights reserved. Trademarks and registered trademarks are the property of their respective owners. D04699-0-8/07(B) Rev. B | Page 20 of 20 Package Option CP-24-1 CP-24-1 Ordering Quantity 64 1,500 1