Low Phase Noise, Fast Settling PLL Frequency Synthesizer ADF4193 FEATURES GENERAL DESCRIPTION New, fast settling, fractional-N PLL architecture Single PLL replaces ping-pong synthesizers Frequency hop across GSM band in 5 μs with phase settled by 20 μs 0.5° rms phase error at 2 GHz RF output Digitally programmable output phase RF input range up to 3.5 GHz 3-wire serial interface On-chip, low noise differential amplifier Phase noise figure of merit: −216 dBc/Hz Loop filter design possible using ADI SimPLL The ADF4193 frequency synthesizer can be used to implement local oscillators in the upconversion and downconversion sections of wireless receivers and transmitters. Its architecture is specifically designed to meet the GSM/EDGE lock time requirements for base stations. It consists of a low noise, digital phase frequency detector (PFD), and a precision differential charge pump. There is also a differential amplifier to convert the differential charge pump output to a single-ended voltage for the external voltage-controlled oscillator (VCO). The Σ-Δ based fractional interpolator, working with the N divider, allows programmable modulus fractional-N division. Additionally, the 4-bit reference (R) counter and on-chip frequency doubler allow selectable reference signal (REFIN) frequencies at the PFD input. A complete phase-locked loop (PLL) can be implemented if the synthesizer is used with an external loop filter and a VCO. The switching architecture ensures that the PLL settles inside the GSM time slot guard period, removing the need for a second PLL and associated isolation switches. This decreases cost, complexity, PCB area, shielding, and characterization on previous ping-pong GSM PLL architectures. APPLICATIONS GSM/EDGE base stations PHS base stations Instrumentation and test equipment FUNCTIONAL BLOCK DIAGRAM SDVDD DVDD1 DVDD2 DVDD3 AVDD1 VP 1 VP2 VP3 RSET REFERENCE /2 DIVIDER + PHASE FREQUENCY DETECTOR – SW1 CPOUT+ CHARGE + PUMP – CPOUT– SW2 VDD HIGH Z MUXOUT 4-BIT R COUNTER ×2 DOUBLER REFIN DGND OUTPUT MUX CMR LOCK DETECT DIFFERENTIAL AMPLIFIER – AIN– + AIN+ RDIV NDIV AOUT N COUNTER SW3 FRACTIONAL INTERPOLATOR CLK DATA LE 24-BIT DATA REGISTER RFIN+ RFIN– FRACTION REG MODULUS REG INTEGER REG AGND1 AGND2 DGND1 DGND2 DGND3 SDGND SWGND 05328-001 ADF4193 Figure 1. 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 ©2006 Analog Devices, Inc. All rights reserved. ADF4193 TABLE OF CONTENTS Features .............................................................................................. 1 Function Register (R3) .............................................................. 18 Applications....................................................................................... 1 Charge Pump Register (R4) ...................................................... 19 General Description ......................................................................... 1 Power-Down Register (R5) ....................................................... 20 Functional Block Diagram .............................................................. 1 Mux Register (R6) ...................................................................... 21 Revision History ............................................................................... 2 Programming .................................................................................. 22 Specifications..................................................................................... 3 Worked Example ........................................................................ 22 Timing Characteristics ................................................................ 4 Spur Mechanisms ....................................................................... 22 Absolute Maximum Ratings............................................................ 5 Power-Up Initialization ............................................................. 23 ESD Caution.................................................................................. 5 Changing the Frequency of the PLL and the Phase Look-Up Table ............................................................................................. 23 Pin Configuration and Function Descriptions............................. 6 Typical Performance Characteristics ............................................. 8 Theory of Operation ...................................................................... 11 Reference Input Section............................................................. 11 RF Input Stage............................................................................. 11 Register Map.................................................................................... 14 FRAC/INT Register (R0)........................................................... 15 Applications..................................................................................... 25 Local Oscillator for A GSM Base Station ................................ 25 Interfacing ................................................................................... 27 PCB Design Guidelines for Chip Scale Package .................... 27 Outline Dimensions ....................................................................... 28 Ordering Guide .......................................................................... 28 MOD/R Register (R1) ................................................................ 16 Phase Register (R2) .................................................................... 17 REVISION HISTORY 6/06—Rev A. to Rev. B Changes to Table 1............................................................................ 3 Changes to Figure 32...................................................................... 18 Changes to Power-Up Initialization Section............................... 23 Changes to Timer Values for Tx Section and Timer Values for Rx Section ........................................................................................ 25 11/05—Rev 0. to Rev. A Updated Format..................................................................Universal Changes to Features Section............................................................ 1 Changes to Table 1............................................................................ 3 Changes to Reference Input Section ............................................ 11 Changes to RF N Divider Section ................................................ 11 Changes to the Lock Detect Section ............................................ 13 Changes to Figure 29...................................................................... 15 Changes to the 8-Bit INT Value Section ..................................... 15 Changes to Figure 33...................................................................... 19 Replaced Figure 35 ......................................................................... 21 Changes to the Σ-Δ and Lock Detect Modes Section................ 21 Changes to the Power-Up Initialization Section ........................ 23 Changes to Table 8.......................................................................... 23 Changes to the Local Oscillator for a GSM Base Station Section ....................................................................... 25 Changes to the Timer Values for Rx Section .............................. 25 Changes to Figure 36...................................................................... 26 Updates to the Outline Dimensions ............................................ 28 Changes to the Ordering Guide ................................................... 28 4/05—Revision 0: Initial Version Rev. B | Page 2 of 28 ADF4193 SPECIFICATIONS AVDD = DVDD = SDVDD = 3 V ± 10%, VP1, VP2 = 5 V ± 10%, VP3 = 5.35 V ± 5%, AGND = DGND = GND = 0 V, RSET = 2.4 kΩ, dBm referred to 50 Ω, TA = TMIN to TMAX, unless otherwise noted. Table 1. Parameter RF CHARACTERISTICS RF Input Frequency (RFIN) RF Input Sensitivity Maximum Allowable Prescaler Output Frequency 2 REFIN CHARACTERISTICS REFIN Input Frequency REFIN Edge Slew Rate REFIN Input Sensitivity REFIN Input Capacitance REFIN Input Current PHASE DETECTOR Phase Detector Frequency CHARGE PUMP ICP Up/Down High Value Low Value Absolute Accuracy RSET Range ICP Three-State Leakage ICP Up vs. Down Matching ICP vs. VCP ICP vs. Temperature DIFFERENTIAL AMPLIFIER Input Current Output Voltage Range VCO Tuning Range Output Noise LOGIC INPUTS VIH, Input High Voltage VIL, Input Low Voltage IINH, IINL, Input Current CIN, Input Capacitance LOGIC OUTPUTS VOH, Output High Voltage VOL, Output Low Voltage POWER SUPPLIES AVDD DVDD VP1, VP2 VP3 IDD (AVDD + DVDD + SDVDD) IDD (VP1 + VP2) IDD (VP3) IDD Power-Down B Version 1 Unit Test Conditions/Comments 0.4/3.5 –10/0 470 GHz min/max dBm min/max MHz max See Figure 21 for input circuit 300 300 0.7/VDD 0 to VDD 10 ±100 MHz max V/μs min V p-p min/max V max pF max μA max For f > 120 MHz, set REF/2 bit = 1 26 MHz max 6.6 104 5 1/4 1 0.1 1 1 mA typ μA typ % typ kΩ min/max nA typ % typ % typ % typ 1 1.4/(VP3 − 0.3) 1.8/(VP3 − 0.8) 7 nA typ V min/max V min/max nV/√Hz typ 1.4 0.7 ±1 10 V min V max μA max pF max VDD – 0.4 0.4 V min V max 2.7/3.3 AVDD 4.5/5.5 5.0/5.65 27 27 30 10 V min/V max V min/V max V min/V max mA max mA max mA max μA typ Rev. B | Page 3 of 28 AC-coupled CMOS-compatible With RSET = 2.4 kΩ With RSET = 2.4 kΩ Nominally RSET = 2.4 kΩ 0.75 V ≤ VCP ≤ VP – 1.5 V 0.75 V ≤ VCP ≤ VP – 1.5 V 0.75 V ≤ VCP ≤ VP – 1.5 V @ 20 kHz offset IOH = 500 μA IOL = 500 μA AVDD ≤ VP1, VP2 ≤ 5.5 V VP1, VP2 ≤ VP3 ≤ 5.65 V 22 mA typ 22 mA typ 24 mA typ ADF4193 Parameter SW1, SW2, and SW3 RON (SW1 and SW2) RON SW3 NOISE CHARACTERISTICS 900 MHz Output 3 1800 MHz Output 4 Phase Noise Figure of Merit 5 B Version 1 Unit 65 75 Ω typ Ω typ –108 –102 –216 dBc/Hz typ dBc/Hz typ dBc/Hz typ Test Conditions/Comments @ 5 kHz offset and 26 MHz PFD frequency @ 5 kHz offset and 13 MHz PFD frequency @ VCO output with dither off 1 Operating temperature range is from –40°C to +85°C. The prescaler value is chosen to ensure that the RF input is divided down to a frequency that is less than this value. 3 fREFIN = 26 MHz; fSTEP = 200 kHz; fRF = 900 MHz; loop BW = 40 kHz. 4 fREFIN = 13 MHz; fSTEP = 200 kHz; fRF = 1850 MHz; loop BW = 60 kHz. 5 Calculated from the phase noise measured at 5 kHz with a 60 kHz loop BW. Increased noise contribution from the differential amplifier if the loop BW is reduced. 2 TIMING CHARACTERISTICS AVDD = DVDD = 3 V ± 10%, VP1, VP2 = 5 V ± 10%, VP3 = 5.35 V ± 5%, AGND = DGND = GND = 0 V, RSET = 2.4 kΩ, dBm referred to 50 Ω, TA = TMIN to TMAX, unless otherwise noted. Table 2. Limit (B Version) 1 10 10 10 15 15 10 15 Parameter t1 t2 t3 t4 t5 t6 t7 Test Conditions/Comments LE setup time DATA to CLOCK setup time DATA to CLOCK hold time CLOCK high duration CLOCK low duration CLOCK to LE setup time LE pulse width Operating temperature is from −40°C to +85°C. t4 t5 CLK t2 DATA DB23 (MSB) t3 DB22 DB2 DB1 (CONTROL BIT C2) DB0 (LSB) (CONTROL BIT C1) t7 LE t1 t6 LE 05238-002 1 Unit ns min ns min ns min ns min ns min ns min ns min Figure 2. Timing Diagram Rev. B | Page 4 of 28 ADF4193 ABSOLUTE MAXIMUM RATINGS TA = 25°C, unless otherwise noted. Table 3. Parameter AVDD to GND AVDD to DVDD, SDVDD VP to GND VP to AVDD Digital I/O Voltage to GND Analog I/O Voltage to GND REFIN, RFIN+, RFIN− to GND Operating Temperature Range Industrial (B Version) Storage Temperature Range Maximum Junction Temperature LFCSP θJA Thermal Impedance (Paddle Soldered) Reflow Soldering Peak Temperature Time at Peak Temperature Rating –0.3 V to +3.6 V –0.3 V to +0.3 V –0.3 V to +5.8 V –0.3 V to +5.8 V –0.3 V to VDD + 0.3 V –0.3 V to VP + 0.3 V –0.3 V to VDD + 0.3 V –40°C to +85°C –65°C to +125°C 150°C 27.3°C/W 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. This device is a high performance RF integrated circuit with an ESD rating of <2 kV, and it is ESD sensitive. Proper precautions need to be taken for handling and assembly. Transistor Count 75,800 (MOS), 545 (BJT) 260°C 40 sec ESD CAUTION ESD (electrostatic discharge) sensitive device. Electrostatic charges as high as 4000 V readily accumulate on the human body and test equipment and can discharge without detection. Although this product features proprietary ESD protection circuitry, permanent damage may occur on devices subjected to high energy electrostatic discharges. Therefore, proper ESD precautions are recommended to avoid performance degradation or loss of functionality. Rev. B | Page 5 of 28 ADF4193 32 VP3 31 AIN+ 30 CPOUT+ 29 SW1 28 SWGND 27 SW2 26 CPOUT– 25 AIN– PIN CONFIGURATION AND FUNCTION DESCRIPTIONS 1 2 3 4 5 6 PIN 1 INDICATOR ADF4193 TOP VIEW 24 VP2 23 RSET 22 AGND2 21 DGND3 20 VP1 19 LE 18 DATA 8 17 CLK 05328-003 7 DGND1 9 DVDD2 10 REFIN 11 DGND2 12 DVDD3 13 SDGND 14 SDVDD 15 MUXOUT 16 CMR AOUT SW3 AGND1 RFIN– RFIN+ AVDD1 DVDD1 Figure 3. Pin Configuration Table 4. Pin Function Descriptions Pin No. 1 Mnemonic CMR 2 3 4 5 AOUT SW3 AGND1 RFIN− 6 7 RFIN+ AVDD1 8 DVDD1 9 10 DGND1 DVDD2 11 REFIN 12 13 14 15 DGND2 DVDD3 SDGND SDVDD 16 MUXOUT 17 CLK 18 DATA 19 LE 20 VP1 21 22 DGND3 AGND2 Description Common-Mode Reference Voltage for the Differential Amplifier’s Output Voltage Swing. Internally biased to three-fifths of VP3. Requires a 0.1 μF capacitor to ground. Differential Amplifier Output to Tune the External VCO. Fast-Lock Switch 3. Closed while SW3 timeout counter is active. Analog Ground. This is the ground return pin for the differential amplifier and the RF section. Complementary Input to the RF Prescaler. This point must be decoupled to the ground plane with a small bypass capacitor, typically 100 pF. Input to the RF Prescaler. This small signal input is ac-coupled to the external VCO. Power Supply Pin for the RF Section. Nominally 3 V. A 100 pF decoupling capacitor to the ground plane should be placed as close as possible to this pin. Power Supply Pin for the N Divider. Should be the same voltage as AVDD1. A 0.1 μF decoupling capacitor to ground should be placed as close as possible to this pin. Ground Return Pin for DVDD1. Power Supply Pin for the REFIN Buffer and R Divider. Nominally 3 V. A 0.1 μF decoupling capacitor to ground should be placed as close as possible to this pin. Reference Input. This is a CMOS input with a nominal threshold of VDD/2 and a dc equivalent input resistance of 100 kΩ (see Figure 15). This input can be driven from a TTL or CMOS crystal oscillator or it can be ac-coupled. Ground Return Pin for DVDD2 and DVDD3. Power Supply Pin for the Serial Interface Logic. Nominally 3 V. Ground Return Pin for the Σ-Δ Modulator. Power Supply Pin for the Digital Σ-Δ Modulator. Nominally 3 V. A 0.1 μF decoupling capacitor to the ground plane should be placed as close as possible to this pin. Multiplexer Output. This multiplexer output allows either the lock detect, the scaled RF, or the scaled reference frequency to be accessed externally (see Figure 35). Serial Clock Input. Data is clocked into the 24-bit shift register on the CLK rising edge. This input is a high impedance CMOS input. Serial Data Input. The serial data is loaded MSB first with the three LSBs as the control bits. This input is a high impedance CMOS input. Load Enable, CMOS Input. When LE goes high, the data stored in the shift register is loaded into the register that is selected by the three LSBs. Power Supply Pin for the Phase Frequency Detector (PFD). Nominally 5 V, should be at the same voltage at VP2. A 0.1 μF decoupling capacitor to ground should be placed as close as possible to this pin. Ground Return Pin for VP1. Ground Return Pin for VP2. Rev. B | Page 6 of 28 ADF4193 Pin No. 23 Mnemonic RSET Description Connecting a resistor between this pin and GND sets the charge pump output current. The nominal voltage bias at the RSET pin is 0.55 V. The relationship between ICP and RSET is ICP = 0.25/RSET 24 VP2 25 26 27 28 29 30 31 32 AIN− CPOUT− SW2 SWGND SW1 CPOUT+ AIN+ VP3 So, with RSET = 2.4 kΩ, ICP = 104 μA. Power Supply Pin for the Charge Pump. Nominally 5 V, should be at the same voltage at VP1. A 0.1 μF decoupling capacitor to ground should be placed as close as possible to this pin. Differential Amplifier’s Negative Input Pin. Differential Charge Pump’s Negative Output Pin. Should be connected to AIN− and the loop filter. Fast Lock Switch 2. This switch is closed to SWGND while the SW1/2 timeout counter is active. Common for SW1 and SW2 Switches. Should be connected to the ground plane. Fast Lock Switch 1. This switch is closed to SWGND while the SW1/2 timeout counter is active. Differential Charge Pump’s Positive Output Pin. Should be connected to AIN+ and the loop filter. Differential Amplifier’s Positive Input Pin. Power Supply Pin for the Differential Amplifier. This can range from 5.0 V to 5.5 V. A 0.1 μF decoupling capacitor to ground should be placed as close as possible to this pin. Also requires a 10 μF decoupling capacitor to ground. Rev. B | Page 7 of 28 ADF4193 TYPICAL PERFORMANCE CHARACTERISTICS 0 FREQ. UNIT GHz KEYWORD R PARAM TYPE S IMPEDANCE 50 DATA FORMAT MA MAGS11 0.67107 0.66556 0.6564 0.6333 0.61406 0.5977 0.5655 0.5428 0.51733 0.49909 0.47309 0.45694 0.44698 0.43589 0.42472 0.41175 0.41055 0.40983 –5 ANGS11 –75.8206 –77.6851 –80.3101 –82.5082 –85.5623 –87.3513 –89.7605 –93.0239 –95.9754 –99.1291 –102.208 –106.794 –111.659 –117.986 –125.62 –133.291 –140.585 –147.97 4/5 PRESCALER –10 8/9 PRESCALER –15 –20 –25 –30 –35 05328-005 FREQ. 2.3 2.4 2.5 2.6 2.7 2.8 2.9 3.0 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 4.0 0 –30 –30 GSM900 Rx SETUP, 40kHz LOOP BW, DITHER OFF RF = 1092.8MHz, FREF = 26MHz, MOD = 130 N = 42 4/130 INTEGER BOUNDARY SPUR: –103dBc @ 800kHz –40 –50 –60 PHASE NOISE (dBc/Hz) –70 –80 –90 –100 –110 –120 –130 –90 –100 –110 –120 –130 –140 –150 –160 10k 100k 1M 10M –160 –170 1k 100M FREQUENCY (Hz) –60 DCS1800 Tx SETUP WITH DITHER OFF, 60kHz LOOP BW, 13MHz PFD. MEASURED ON EVAL-ADF4193-EB1 BOARD 1M 10M 100M DCS1800 Tx SETUP WITH DITHER OFF, 60kHz LOOP BW, 13MHz PFD. MEASURED ON EVAL-ADF4193-EB1 BOARD –70 400kHz SPURS @ 25°C –80 –90 –100 –80 600kHz SPURS @ 25°C –90 –100 –110 400kHz SPURS @ 85°C 1859 FREQUENCY (MHz) 05328-010 –110 –120 1846 100k Figure 8. SSB Phase Noise Plot at 1842.6 MHz (DCS1800 Tx Setup) SPUR LEVEL (dBc) SPUR LEVEL (dBc) –70 10k FREQUENCY (Hz) Figure 5. SSB Phase Noise Plot at 1092.8 MHz (GSM900 Rx Setup) vs. Free Running VCO Noise –60 DCS1800 Tx SETUP, 60kHz LOOP BW, DITHER OFF RF = 1842.6MHz, FREF = 13MHz, MOD = 65 DSB INTEGRATED PHASE ERROR = 0.46° RMS SIRENZA 1843T VCO –80 –150 –170 1k 5000 –70 –140 05328-006 PHASE NOISE (dBc/Hz) –60 4000 –120 1846 1872 Figure 6. 400 kHz Fractional Spur Levels Across All DCS1800 Tx Channels over Two-Integer Multiples of the PFD Reference 600kHz SPURS @ 85°C 1859 FREQUENCY (MHz) 05328-011 –50 2000 3000 RFIN FREQUENCY (MHz) Figure 7. RF Input Sensitivity Figure 4. S Parameter Data for the RF Input –40 1000 05328-007 ANGS11 –16.6691 –19.9279 –23.561 –26.9578 –30.8201 –34.9499 –39.0436 –42.3623 –46.322 –50.3484 –54.3545 –57.3785 –60.695 –63.9152 –66.4365 –68.4453 –70.7986 –73.7038 RFIN LEVEL (dBm) MAGS11 0.8897 0.87693 0.85834 0.85044 0.83494 0.81718 0.80229 0.78917 0.77598 0.75578 0.74437 0.73821 0.7253 0.71365 0.70699 0.7038 0.69284 0.67717 05328-038 FREQ. 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2.0 2.1 2.2 1872 Figure 9. 600 kHz Fractional Spur Levels Across All DCS1800 Tx Channels over Two-Integer Multiples of the PFD Reference Rev. B | Page 8 of 28 ADF4193 5 5 4 DCS1800 Tx SETUP, 60kHz LOOP BW. MEASURED ON EVAL-ADF4193-EB1 EVALUATION BOARD. TIMERS: ICP = 28, SW1/SW2, SW3 = 35. FREQUENCY LOCK IN WIDE BW MODE @ 5μs. 4 VTUNE 3 3 CPOUT– (V) (V) CPOUT+ 2 2 VTUNE 1 CPOUT+ 0 1 2 3 4 5 TIME (μs) 6 7 8 05328-040 0 –1 0 –1 9 Figure 10. VTUNE Settling Transient for a 75 MHz Jump from 1818 MHz to 1893 MHz with Sirenza 1843T VCO 50 30 10 0 –40°C –20 +85°C 3 4 5 TIME (μs) 6 7 8 9 20 10 0 –10 –40°C –20 +85°C –40 –50 –5 0 5 10 15 20 25 TIME (μs) 30 35 40 05328-009 –30 05328-008 –30 –40 –50 –5 45 Figure 11. Phase Settling Transient for a 75 MHz Jump from 1818 MHz to 1893 MHz (VTUNE 1.8 V to 3.7 V with Sirenza 1843T VCO) 0 5 10 15 20 25 TIME (μs) 30 35 40 45 Figure 14. Phase Settling Transient for a 75 MHz Jump from 1893 MHz to 1818 MHz (VTUNE = 3.7 V to 1.8 V with Sirenza 1843T VCO) 2.0 8 ICPOUT + P, ICPOUT – P IUP = | ICPOUT + P | + | ICPOUT – N | IDOWN = | ICPOUT – P | + | ICPOUT + N | 1.0 CHARGE PUMP MISMATCH (%) 0 0 –2 –0.5 NORMAL OPERATING RANGE –4 –1.0 –6 –1.5 (V) 0.5 2 AOUT (= VTUNE ) 4 MISMATCH (%) 4 VP1 = VP2 = 5V VP3 = 5.5V VCMR = 3.3V 5 1.5 6 ICP (mA) 2 DCS1800 Tx SETUP, 60kHz LOOP BW. MEASURED ON EVAL-ADF4193-EB1 EVALUATION BOARD WITH AD8302 PHASE DETECTOR. TIMERS: ICP = 28, SW1/SW2, SW3 = 35. +25°C PEAK PHASE ERROR < 5° @ 19.2μs 40 20 –10 1 50 PHASE ERROR (Degrees) PHASE ERROR (Degrees) 30 0 Figure 13. VTUNE Settling Transient for a 75 MHz Jump Down from 1893 MHz to 1818 MHz, the Bottom of the Allowed Tuning Range with the Sirenza 1843T VCO DCS1800 Tx SETUP, 60kHz LOOP BW. MEASURED ON EVAL-ADF4193-EB1 EVALUATION BOARD WITH AD8302 PHASE DETECTOR. TIMERS: ICP = 28, SW1/SW2, SW3 = 35. +25°C PEAK PHASE ERROR < 5° @ 17.8μs 40 05328-041 CPOUT– DCS1800 Tx SETUP, 60kHz LOOP BW. MEASURED ON EVAL-ADF4193-EB1 EVALUATION BOARD. TIMERS: ICP = 28, SW1/SW2, SW3 = 35. FREQUENCY LOCK IN WIDE BW MODE @ 4μs. 1 3 CPOUT+ (= AIN+) 2 ICPOUT + N, ICPOUT – N 0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 –2.0 5.0 CPOUT– (= AIN–) 0 1780 CPOUT + / CPOUT – VOLTAGE (V) Figure 12. Differential Charge Pump Output Compliance Range and Charge Pump Mismatch with VP1 = VP2 = 5 V 1800 1820 1840 1860 1880 1900 1920 05328-013 –8 05328-012 1 1940 FREQUENCY (MHz) Figure 15. Tuning Range with a Sirenza 1843T VCO and a 5.5 V Differential Amplifier Power Supply Voltage Rev. B | Page 9 of 28 ADF4193 100 7nV/ Hz @ 20kHz 10 10k 100k FREQUENCY (Hz) 1M 1.2 0.9 0.6 0.3 0 10M Figure 16. Voltage Noise Density Measured at the Differential Amplifier Output 1.5 0 13 26 +85°C SW1/ SW2 +85°C +25°C 70 REFIN 40 AGILENT HP8663A SIG. GEN. TUNING VOLTAGE RANGE 130 1880MHz 20 3 4 TEKTRONIX TDS714L OSCILLOSCOPE AD8302 EVB 1880MHz R&S SMT03 SIG. GEN. 05328-014 10 2 VPHS INPB 10MHz EXT REF 30 1 117 INPA –40°C 0 104 RFOUT 1805 +25°C 50 0 91 ADF4193 EVAL BOARD –40°C 60 RON (Ω) 104MHz 5dBm SW3 80 52 65 78 PHASE CODE Figure 18. Detected RF Output Phase for Phase Code Sweep from 0 to MOD 100 90 39 5 INTERVAL BETWEEN R0 WRITES SHOULD BE A MULTIPLE OF MOD REFERENCE CYCLES (5µs) FOR COHERENT PHASE MEASUREMENTS DRAIN VOLTAGE (V) Figure 19. Test Setup for Phase Lock Time Measurement Figure 17. On Resistance of Loop Filter Switches SW1/SW2 and SW3 Rev. B | Page 10 of 28 05328-045 1 1k MEASURED USING AD8302 PHASE DETECTOR Y-AXIS SCALE: 10mV/DEGREE RF = 1880MHz, PFD = 26MHz, MOD = 130 X-AXIS SCALE: 2.77°/STEP 05328-044 PHASE DETECTOR OUTPUT (V) 1.8 05328-042 NOISE (nV/ Hz) 1000 ADF4193 THEORY OF OPERATION The ADF4193 is targeted at GSM base station requirements, specifically to eliminate the need for ping-pong solutions. It works based on fast lock, using a wide loop bandwidth during a frequency change and narrowing the loop bandwidth once frequency lock is achieved. Widening the loop bandwidth is achieved by increasing the charge pump current. Switches are included to change the loop filter component values to maintain stability with the changing charge pump current. The narrow loop bandwidth ensures that phase noise and spur specifications are met. A differential charge pump and loop filter topology are used to ensure that the fast lock time benefit from widening the loop bandwidth is maintained when the loop is restored to narrow bandwidth mode for normal operation. RF INPUT STAGE The RF input stage is shown in Figure 21. It is followed by a 2-stage limiting amplifier to generate the CML clock levels needed for the prescaler. Two prescaler options are selectable: a 4/5 and an 8/9. The 8/9 prescaler is selected for N divider values greater than 80. BIAS GENERATOR 500Ω 100kΩ TO R COUNTER S3 NO 05328-016 BUFFER S1 Figure 21. RF Input Stage RF N Divider The RF N divider allows a fractional division ratio in the PLL feedback path. The integer and fractional parts of the division are programmed using separate registers, as shown in Figure 22 and described in the INT, FRAC, and MOD Relationship section. Integer division ratios from 26 to 255 are allowed and a third-order, Σ-Δ modulator interpolates the fractional value between the integer steps. POWER-DOWN CONTROL S2 AGND 05328-017 RFIN– The reference input stage is shown in Figure 20. Switches S1 and S2 are normally closed, and S3 is normally open. During powerdown, S3 is closed, and S1 and S2 are opened to ensure that there is no loading of the REFIN pin. The falling edge of REFIN is the active edge at the positive edge triggered PFD. REFIN NC 500Ω RFIN+ REFERENCE INPUT SECTION NC AVDD 1.6V Figure 20. Reference Input Stage RF N DIVIDER FROM RF INPUT STAGE The 4-bit R counter allows the input reference frequency to be divided down to produce the reference clock to the phase frequency detector (PFD). A toggle flip-flop can be optionally inserted after the R counter to give a further divide-by-2. Using this option has the additional advantage of ensuring that the PFD reference clock has a 50/50 mark-space ratio. This ratio gives the maximum separation between the fast lock timer clock, which is generated off the falling edge of the PFD reference, and the rising edge, which is the active edge in the PFD. It is recommended that this toggle flip-flop be enabled for all even R divide values greater than 2. It must be enabled if dividing down a REFIN frequency that is greater than 120 MHz. An optional doubler before the 4-bit R counter can be used for low REFIN frequencies, up to 20 MHz. With these programmable options, reference division ratios from 0.5 to 30 between REFIN and the PFD are possible. TO PFD N COUNTER THIRD-ORDER FRACTIONAL INTERPOLATOR INT REG MOD REG FRAC VALUE 05328-018 R Counter and Doubler N = INT + FRAC/MOD Figure 22. Fractional-N Divider INT, FRAC, and MOD Relationship The INT, FRAC, and MOD values, programmed through the serial interface, make it possible to generate RF output frequencies that are spaced by fractions of the PFD reference frequency. The N divider value, shown inside the brackets of the following equation for the RF VCO frequency (RFOUT), is made up of an integer part (INT) and a fractional part (FRAC/MOD): RFOUT = FPFD × [INT + (FRAC/MOD)] where: RFOUT is the output frequency of the external VCO. FPFD is the PFD reference frequency. Rev. B | Page 11 of 28 ADF4193 PFD and Charge Pump The PFD takes inputs from the R divider and N divider and produces up and down outputs with a pulse width difference proportional to the phase difference between the inputs. The charge pump outputs a net up or down current pulse of a width equal to this difference, to pump up or pump down the voltage that is integrated onto the loop filter, which in turn increases or decreases the VCO output frequency. If the N divider phase lags the R divider phase, a net up current pulse is produced that increases the VCO frequency (and thus the phase). If the N divider phase leads the R divider edge, then a net down pulse is produced to reduce the VCO frequency and phase. Figure 23 is a simplified schematic of the PFD and charge pump. The charge pump is made up of an array of 64 identical cells, each of which is fully differential. All 64 cells are active during fast lock, but only one is active during normal operation. Because a single-ended control voltage is required to tune the VCO, an on-chip, differential-to-single-ended amplifier is provided for this purpose. In addition, because the phase-lock loop only controls the differential voltage generated across the charge pump outputs, an internal common-mode feedback (CMFB) loop biases the charge pump outputs at a common-mode voltage of approximately 2 V. D R DIVIDER CLR CHARGE PUMP ARRAY [64:1] CMFB CLR CPOUT– Q N DIVIDER EN[64:1] 05328-019 D P VBIAS P P UP DOWN CPOUT+ CPOUT– DOWN UP N VBIAS N N Figure 24. Differential Charge Pump Cell with External Loop Filter Components Fast Lock Timeout Counters CPOUT+ Q To pump up, the up switches are on and PMOS current is sourced out through CPOUT+; this increases the voltage on the external loop filter capacitors connected to CPOUT+. Similarly, the NMOS current sink on CPOUT− decreases the voltage on the external loop filter capacitors connected to CPOUT−. Therefore, the differential voltage between CPOUT+ and CPOUT− increases. To pump down, PMOS current sources out through CPOUT− and NMOS current sinks in through CPOUT+, which decreases the (CPOUT+, CPOUT−) differential voltage. The charge pump up/ down matching is improved by an order of magnitude over the conventional single-ended charge pump that depended on the matching of two different device types. The up/down matching in this structure depends on how a PMOS matches a PMOS and an NMOS matches an NMOS. 05328-035 The value of MOD is chosen to give the desired channel step with the available reference frequency. Thereafter, program the INT and FRAC words for the desired RF output frequency. See the Worked Example section for more information. Figure 23. PFD and Differential Charge Pump Simplified Schematic Differential Charge Pump The charge pump cell (see Figure 24) has a fully differential design for best up-to-down current matching. Good matching is essential to minimize the phase offset created when switching the charge pump current from its high value (in fast lock mode) to its nominal value (in normal mode). Timeout counters, clocked at one quarter the PFD reference frequency, are provided to precisely control the fast locking operation (see Figure 25). Whenever a new frequency is programmed, the fast lock timers start and the PLL locks into wide BW mode with the 64 identical 100 μA charge pump cells active (6.4 mA total). When the ICP counter times out, the charge pump current is reduced to 1× by deselecting cells in binary steps over the next six timer clock cycles, until just one 100 μA cell is active. The charge pump current switching from 6.4 mA to 100 μA equates to an 8-to-1 change in loop bandwidth. The loop filter must be changed to ensure stability when this happens. That is the job of the SW1, SW2, and SW3 switches. The application circuit (shown in Figure 36) shows how they can be used to reconfigure the loop filter time constants. The application circuits close to short out external loop filter resistors during fast lock and open when their counters time out to restore the filter time constants to their normal values for the 100 μA charge pump current. Because it takes six timer clock cycles to reduce the charge pump current to 1×, it is recommended that both switch timers be programmed to the value of the ICP timer + 7. Rev. B | Page 12 of 28 ADF4193 START MUXOUT and Lock Detect ICP TIMEOUT COUNTER FPFD SW1/SW2 TIMEOUT COUNTER The output multiplexer on the ADF4193 allows the user to access various internal points on the chip. The state of MUXOUT is controlled by M4 to M1 in the MUX register. Figure 35 shows the full truth table. Figure 27 shows the MUXOUT section in block diagram form. SW3 TIMEOUT COUNTER ÷4 SW3 CHARGE PUMP ENABLE LOGIC AOUT SW1 SW2 EN[64:1] SWGND DVDD 05328-036 WRITE TO R0 LOGIC LOW SERIAL DATA OUTPUT R DIVIDER OUTPUT N DIVIDER OUTPUT Figure 25. Fast Lock Timeout Counters MUX MUXOUT CONTROL THREE-STATE OUTPUT Differential Amplifier TIMER OUTPUTS VAOUT = (VAIN+ − VAIN−) + VCMR The CMR offset voltage is internally biased to three-fifths of VP3, the differential amplifier power supply voltage, as shown in Figure 26. Connect a 0.1 μF capacitor to ground to the CMR pin to roll off the thermal noise of the biasing resistors. As can be seen in Figure 15, the differential amplifier output voltage behaves according to the previous equation over a 4 V range from approximately 1.2 V minimum up to VP3 − 0.3 V. However, fast settling is guaranteed only over a tuning voltage range from 1.8 V up to VP3 − 0.8 V. This is to allow sufficient room for overshoot in the PLL frequency settling transient. Noise from the differential amplifier is suppressed inside the PLL bandwidth. For loop bandwidths >20 kHz, the 1/f noise has a negligible effect on the PLL output phase noise. Outside the loop bandwidth, the differential amplifier’s noise FM modulates the VCO. The passive filter network following the differential amplifier, shown in Figure 36, suppresses this noise contribution to below the VCO noise from offsets of 400 kHz and above. This network has a negligible effect on lock time because it is bypassed when SW3 is closed while the loop is locking. AIN– 500Ω 500Ω AOUT VP3 AIN+ 20kΩ CMR 500Ω 30kΩ Figure 26. Differential Amplifier Block Diagram C EXT = 0.1µF 05328-020 500Ω DIGITAL LOCK DETECT LOGIC HIGH DGND NOTE: NOT ALL MUXOUT MODES SHOWN REFER TO MUX REGISTER 05328-021 The internal, low noise, differential-to-single-ended amplifier is used to convert the differential charge pump output to a singleended control voltage for the tuning port of the VCO. Figure 26 shows a simplified schematic of the differential amplifier. The output voltage is equal to the differential voltage, offset by the voltage on the CMR pin, according to Figure 27. MUXOUT Circuit Lock Detect MUXOUT can be programmed to provide a digital lock detect signal. Digital lock detect is active high. Its output goes high if there are 40 successive PFD cycles with an input error of less than 3 ns. For reliable lock detect operation with RF frequencies <2 GHz, it is recommended that this threshold be increased to 10 ns by programming Register R6. The digital lock detect goes low again when a new channel is programmed or when the error at the PFD input exceeds 30 ns for one or more cycles. Input Shift Register The ADF4193 serial interface section includes a 24-bit input shift register. Data is clocked in MSB first on each rising edge of CLK. Data from the shift register is latched into one of eight control registers, R0 to R7, on the rising edge of latch enable (LE). The destination register is determined by the state of the three control bits (Control Bit C3, Control Bit C2, and Control Bit C1) in the shift register. The three LSBs are Bit DB2, Bit DB1, and Bit DB0, as shown in the timing diagram of Figure 2. The truth table for these bits is shown in Table 5. Figure 28 shows a summary of how the registers are programmed. Table 5. C3, C2, and C1 Truth Table C3 0 0 0 0 1 1 1 1 Rev. B | Page 13 of 28 Control Bits C2 C1 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 Name FRAC/INT MOD/R Phase Function Charge Pump Power-Down Mux Test Mode Register R0 R1 R2 R3 R4 R5 R6 R7 ADF4193 REGISTER MAP RESERVED FRAC/INT REGISTER (R0) 8-BIT RF INT VALUE CONTROL BITS 12-BIT RF FRAC VALUE DB23 DB22 DB21 DB20 DB19 DB18 DB17 DB16 DB15 DB14 DB13 DB12 DB11 DB10 DB9 DB8 DB7 DB6 DB5 DB4 DB3 0 N8 N7 N6 N5 N4 N3 N2 N1 F12 F11 F10 F9 F8 F7 F6 F5 F4 F3 F2 F1 DB2 DB1 DB0 C3 (0) C2 (0) C1 (0) MOD/R REGISTER (R1) DOUBLER ENABLE DBB PRESCALER DBB RESERVED DBB REF/2 DBB CP ADJ DBB DB23 DB22 DB21 DB20 DB19 DB18 DB17 DB16 DB15 DB14 DB13 DB12 DB11 DB10 DB9 DB8 DB7 DB6 DB5 DB4 DB3 F5 F4 0 F2 F1 R4 R3 R2 R1 M12 M11 M10 M9 M8 M7 M6 M5 M4 M3 M2 M1 4-BIT RF R COUNTER CONTROL BITS 12-BIT MODULUS DB2 DB1 DB0 C3 (0) C2 (0) C1 (1) PHASE REGISTER (R2) RESERVED DBB CONTROL BITS 12-BIT PHASE DB15 DB14 DB13 DB12 DB11 DB10 DB9 DB8 DB7 DB6 DB5 DB4 DB3 0 P12 P11 P10 P9 P8 P7 P6 P5 P4 P3 P2 P1 DB2 DB1 DB0 C3 (0) C2 (1) C1 (0) CPO GND RESERVED PFD POLARITY FUNCTION REGISTER (R3) DB15 DB14 DB13 DB12 DB11 DB10 DB9 DB8 DB7 DB6 DB5 DB4 DB3 0 0 0 0 0 0 0 0 0 1 F3 1 F1 RESERVED CONTROL BITS DB2 DB1 DB0 C3 (0) C2 (1) C1 (1) CHARGE PUMP REGISTER (R4) RESERVED TIMER SELECT 9-BIT TIMEOUT COUNTER DB23 DB22 DB21 DB20 DB19 DB18 DB17 DB16 DB15 DB14 DB13 DB12 DB11 DB10 DB9 DB8 DB7 DB6 DB5 DB4 DB3 0 0 0 0 0 0 0 0 0 1 C9 C8 C7 C6 C5 C4 C3 C2 C1 F2 F1 CONTROL BITS DB2 DB1 DB0 C3 (1) C2 (0) C1 (0) CP 3-STATE COUNTER RESET DB7 DB6 DB5 DB4 DB3 F5 F4 F3 F2 F1 PD DIFF AMP PD CHARGE PUMP POWER-DOWN REGISTER (R5) CONTROL BITS DB2 DB1 DB0 C3 (1) C2 (0) C1 (1) MUX REGISTER (R6) SIGMA-DELTA AND LOCK DETECT MODES RESERVED CONTROL BITS MUXOUT DB15 DB14 DB13 DB12 DB11 DB10 DB9 DB8 DB7 DB6 DB5 DB4 DB3 M13 M12 M11 M10 0 0 0 0 0 M4 M3 M2 M1 DB2 DB1 DB0 C3 (1) C2 (1) C1 (0) TEST MODE REGISTER (R7) DB15 DB14 DB13 DB12 DB11 DB10 DB9 DB8 DB7 DB6 DB5 DB4 DB3 0 0 0 0 0 0 0 0 0 0 0 0 0 DBB = DOUBLE BUFFERED BIT(S) Figure 28. Register Map Rev. B | Page 14 of 28 DB2 DB1 DB0 C3 (1) C2 (1) C1 (1) 05328-022 CONTROL BITS RESERVED ADF4193 RESERVED FRAC/INT REGISTER (R0) 8-BIT RF INT VALUE CONTROL BITS 12-BIT RF FRAC VALUE DB23 DB22 DB21 DB20 DB19 DB18 DB17 DB16 DB15 DB14 DB13 DB12 DB11 DB10 DB9 DB8 DB7 DB6 DB5 DB4 DB3 0 N8 N7 N6 N5 N4 N3 N2 N1 F12 F11 F10 F9 F8 F7 F6 F5 F4 F3 F2 F1 F12 0 0 0 0 . . . 1 1 1 1 F11 0 0 0 0 . . . 1 1 1 1 F10 0 0 0 0 . . . 1 1 1 1 F3 0 0 0 0 . . . 1 1 1 1 .......... .......... .......... .......... .......... .......... .......... .......... .......... .......... .......... F2 0 0 1 1 . . . 0 0 1 1 F1 0 1 0 1 . . . 0 1 0 1 DB2 DB1 DB0 C3 (0) C2 (0) C1 (0) FRACTIONAL VALUE (FRAC) 0 1 2 3 . . . 4092 4093 4094 4095 N8 N7 N6 N5 N4 N3 N2 N1 INTEGER VALUE (INT) 0 . . . 1 0 . . . 1 0 . . . 1 1 . . . 1 1 . . . 1 0 . . . 1 1 . . . 1 0 . . . 1 26 . . . 255 05328-023 0 = < FRAC < MOD Figure 29. FRAC/INT Register (R0) R0, the FRAC/INT register, is used to program the synthesizer output frequency. On the next PFD cycle following a write to R0, the N divider section is updated with the new INT and FRAC values. At the same time, the PLL automatically enters fast lock mode and the charge pump current is increased to its maximum value and stays at this value until the ICP timeout counter times out, and switches SW1, SW2, and SW3 closed and remains closed until the SW1, SW2, and SW3 timeout counters time out. Once all registers are programmed during the initialization sequence (see Table 8), all that is required thereafter to program a new channel is a write to R0. However, as described in the Programming section, it can also be desirable to program R1 and R2 register settings on a channel-by-channel basis. These settings are double buffered by the write to R0. This means that while the data is loaded through the serial interface on the respective R1 and R2 write cycles, the synthesizer is not updated with their data until the next write to Register R0. Control Bits The three LSBs, Control Bit C3, Control Bit C2, and Control Bit C1, should be set to 0, 0, 0, respectively, to select R0, the FRAC/INT register. Reserved Bit Bit DB23 is reserved and must be set to 0. 8-Bit INT Value These eight bits set the INT value, which determines the integer part of the feedback division factor. All integer values from 26 to 255 are allowed. See the Worked Example section. 12-Bit FRAC Value The 12 FRAC bits set the numerator of the fraction that is input to the Σ-Δ modulator. This, along with INT, specifies the new frequency channel that the synthesizer locks to, as shown in the Worked Example section. FRAC values from 0 to MOD − 1 cover channels over a frequency range equal to the PFD reference frequency. Rev. B | Page 15 of 28 ADF4193 CP ADJ REF/2 RESERVED PRESCALER DOUBLER ENABLE MOD/R REGISTER (R1) DB23 DB22 DB21 DB20 DB19 DB18 DB17 DB16 DB15 DB14 DB13 DB12 DB11 DB10 DB9 DB8 DB7 DB6 DB5 DB4 DB3 F5 F4 0 F2 F1 R4 R3 R2 R1 M12 M11 M10 M9 M8 M7 M6 M5 M4 M3 M2 M1 4-BIT RF R COUNTER CONTROL BITS 12-BIT MODULUS DB2 DB1 DB0 C3 (0) C2 (0) C1 (1) F4 REF/2 DISABLE ENABLE F2 PRESCALER F1 DOUBLER ENABLE 0 1 0 1 4/5 8/9 DOUBLER DISABLED DOUBLER ENABLED F5 CP ADJ 0 1 NOMINAL ADJUSTED R4 0 0 0 0 . . . 1 1 1 1 R3 0 0 0 1 . . . 1 1 1 1 M12 0 0 0 . . . 1 1 1 1 M11 0 0 0 . . . 1 1 1 1 R2 0 1 1 0 . . . 0 0 1 1 R1 1 0 1 0 . . . 0 1 0 1 M10 0 0 0 . . . 1 1 1 1 .......... .......... .......... .......... .......... .......... .......... .......... .......... .......... M3 1 1 1 . . . 1 1 1 1 M2 0 1 1 . . . 0 0 1 1 M1 1 0 1 . . . 0 1 0 1 INTERPOLATOR MODULUS VALUE (MOD) 13 14 15 . . . 4092 4093 4094 4095 RF R COUNTER DIVIDE RATIO 1 2 3 4 . . . 12 13 14 15 05328-024 0 1 Figure 30. MOD/R Register (R1) This register is used to set the PFD reference frequency and the channel step size, which is determined by the PFD frequency divided by the fractional modulus. Note that the MOD, R counter, REF/2, CP ADJ, and doubler enable bits are double buffered. They do not take effect until the next write to R0 (FRAC/INT register) is complete. Control Bits Reserved Bit Reserved Bit DB21 must be set to 0. Doubler Enable Setting this bit to 1 inserts a frequency doubler between REFIN and the 4-bit R counter. Setting this bit to 0 bypasses the doubler. 4-Bit RF R Counter With C3, C2, and C1 set to 0, 0, 1, respectively, the MOD/R register (R1) is programmed. It allows the REFIN frequency to be divided down to produce the reference clock to the PFD. All integer values from 1 to 15 are allowed. See the Worked Example section. CP ADJ When this bit is set to 1, the charge pump current is scaled up 25% from its nominal value on the next write to R0. When this bit is set to 0, the charge pump current stays at its nominal value on the next write to R0. See the Programming section for more information on how this feature can be used. REF/2 12-Bit Interpolator Modulus For a given PFD reference frequency, the fractional denominator or modulus sets the channel step resolution at the RF output. All integer values from 13 to 4095 are allowed. See the Programming section for additional information and guidelines for selecting the value of MOD. Setting this bit to 1 inserts a divide-by-2, toggle flip-flop between the R counter and PFD, which extends the maximum REFIN input rate. Rev. B | Page 16 of 28 ADF4193 CONTROL BITS 12-BIT PHASE DB15 DB14 DB13 DB12 DB11 DB10 DB9 DB8 DB7 DB6 DB5 DB4 DB3 0 P12 P11 P10 P9 P8 P7 P6 P5 P4 P3 P2 P1 .......... .......... .......... .......... .......... .......... .......... .......... .......... .......... P3 0 0 0 . . . 1 1 1 1 P12 0 0 0 . . . 1 1 1 1 P11 0 0 0 . . . 1 1 1 1 P10 0 0 0 . . . 1 1 1 1 P2 0 0 1 . . . 0 0 1 1 P1 0 1 0 . . . 0 1 0 1 10 DB2 DB1 DB0 C3 (0) C2 (1) C1 (0) PHASE VALUE1 0 1 2 . . . 4092 4093 4094 4095 = < PHASE VALUE < MOD 05328-025 RESERVED PHASE REGISTER (R2) Figure 31. Phase Register (R2) 12-Bit Phase The phase word sets the seed value of the Σ-Δ modulator. It can be programmed to any integer value from 0 to MOD. As the phase word is swept from 0 to MOD, the phase of the VCO output sweeps over a 360° range in steps of 360°/MOD. Note that the phase bits are double buffered. They do not take effect until the LE of the next write to R0 (FRAC/INT register). Therefore, if it is desired to change the phase of the VCO output frequency, it is necessary to rewrite the INT and FRAC values to R0, following the write to R2. The output of a fractional-N PLL can settle to any one of the MOD possible phase offsets with respect to the reference, where MOD is the fractional modulus. If it is desired to keep the output at the same phase offset with respect to the reference, each time that particular output frequency is programmed, then the interval between writes to R0 must be an integer multiple of MOD reference cycles. If it is desired to keep the outputs of two ADF4193-based synthesizers phase coherent with each other, but not necessarily with their common reference, then it is only required to ensure that the write to R0 on both chips is performed during the same reference cycle. The interval between R0 writes in this case does not have to be an integer multiple of the MOD cycles. Reserved Bit The reserved bit, Bit DB15, should be set to 0. Rev. B | Page 17 of 28 ADF4193 CPO GND RESERVED PFD POLARITY FUNCTION REGISTER (R3) DB15 DB14 DB13 DB12 DB11 DB10 DB9 DB8 DB7 DB6 DB5 DB4 DB3 0 0 0 0 0 0 0 0 0 1 F3 1 F1 C3 (0) C2 (1) C1 (1) F1 PFD POLARITY 0 1 NEGATIVE POSITIVE F3 CPO GND 0 1 CPO/CPO GND NORMAL CONTROL BITS DB2 DB1 DB0 05328-026 RESERVED Figure 32. Function Register (R3) R3, the function register (C3, C2, C1 set to 0, 1, 1, respectively), only needs to be programmed during the initialization sequence (see Table 8). PFD Polarity CPO GND Reserved Bits When the CPO GND bit is low, the charge pump outputs are internally pulled to ground. This is invoked during the initialization sequence to discharge the loop filter capacitors. For normal operation, this bit should be high. The Bit DB15 to Bit DB6 are reserved bits and should be programmed to hex code 001, and Reserved Bit DB4 should be set to 1. This bit should be set to 1 for positive polarity and set to 0 for negative polarity. Rev. B | Page 18 of 28 ADF4193 CHARGE PUMP REGISTER (R4) TIMER SELECT 9-BIT TIMEOUT COUNTER CONTROL BITS DB23 DB22 DB21 DB20 DB19 DB18 DB17 DB16 DB15 DB14 DB13 DB12 DB11 DB10 DB9 DB8 DB7 DB6 DB5 DB4 DB3 0 0 0 0 0 0 0 0 0 1 C9 C8 C7 C6 C5 C4 C3 C2 C1 F2 F1 C9 C8 C7 0 0 0 0 . . . 1 1 1 1 0 0 0 0 . . . 1 1 1 1 0 0 0 0 . . . 1 1 1 1 .......... .......... .......... .......... .......... .......... .......... .......... .......... .......... .......... F2 F1 TIMER SELECT 0 0 1 1 0 1 0 1 DB2 DB1 DB0 C3 (1) C2 (0) C1 (0) SW1/SW2 SW3 ICP NOT USED C3 C2 C1 TIMEOUT COUNTER xPFD CYCLES 0 0 0 0 . . . 1 1 1 1 0 0 1 1 . . . 0 0 1 1 0 1 0 1 . . . 0 1 0 1 0 1 2 3 . . . 508 509 510 511 0 4 8 12 . . . 2032 2036 2040 2044 1DELAY DELAY µs1 0 0.15 0.30 0.46 . . . 78.15 78.30 78.46 78.61 WITH 26MHz PFD 05328-027 RESERVED Figure 33. Charge Pump Register (R4) Table 6. Recommended Values for a GSM Tx LO Reserved Bits Bit DB23 to Bit DB14 are reserved and should be set to hex code 001 for normal operation. 9-Bit Timeout Counter These bits are used to program the fast lock timeout counters. The counters are clocked at one-quarter the PFD reference frequency, therefore, their time delay scales with the PFD frequency according to Delay(s) = (Timeout Counter Value × 4)/(PFD Frequency) For example, if 35 were loaded with timer select (00) with a 13 MHz PFD, then SW1/SW2 would be switched after (35 × 4)/13 MHz = 10.8 μs Timer Select 10 01 00 Timeout Counter ICP SW1/2 SW3 Value 28 35 35 Time (μs) with PFD = 13 MHz 8.6 10.8 10.8 On each write to R0, the timeout counters start. Switch SW3 closes until the SW3 counter times out. Similarly, switches SW1/SW2 close until the SW1/SW2 counter times out. When the ICP counter times out, the charge pump current is ramped down from 64× to 1× in six binary steps. It is recommended that the SW1, SW2, and SW3 timeout counter values are set equal to the ICP timeout counter value plus 7, as in the example of Table 6. Timer Select These two address bits select the timeout counter to be programmed. Note that to set up the ADF4193 correctly requires setup of these three timeout counters; therefore, three writes to this register are required in the initialization sequence. Table 6 shows example values for a GSM Tx synthesizer with a 60 kHz final loop BW. See the Applications section for more information. Rev. B | Page 19 of 28 ADF4193 PD CHARGE PUMP CP 3-STATE COUNTER RESET POWER-DOWN REGISTER (R5) DB7 DB6 DB5 DB4 DB3 F5 F4 F3 F2 F1 PD DIFF AMP NORMAL OPERATION COUNTER RESET F2 CHARGE PUMP 3-STATE 0 1 NORMAL OPERATION 3-STATE ENABLED DISABLED ENABLED 05328-028 DISABLED ENABLED DB0 COUNTER RESET 0 1 0 0 1 1 DB1 0 1 CHARGE PUMP POWER-DOWN DIFF AMP POWER-DOWN DB2 C3 (1) C2 (0) C1 (1) F1 F3 F5 F4 CONTROL BITS Figure 34. Power-Down Register (R5) R5, the power-down register (C3, C2, C1 set to 1, 0, 1, respectively) can be used to software power down the PLL and differential amplifier sections. After power is initially applied, there must be writes to R5 to clear the power-down bits and to R2, R1, and R0 before the ADF4193 comes out of power-down. Power-Down Differential Amplifier When Bit DB6 and Bit DB7 are set high, the differential amplifier is put into power-down. When Bit DB6 and Bit DB7 are set low, normal operation is resumed. Power-Down Charge Pump For normal operation, Bit DB5 should be set to 0, followed by a write to R0. CP Three-State When this bit is set high, the charge pump outputs are put into three-state. With the bit set low, the charge pump outputs are enabled. Counter Reset When this bit is set to 1, the counters are held in reset. For normal operation, this bit should be 0, followed by a write to R0. Setting Bit DB5 high activates a charge pump power-down and the following events occur: • All active dc current paths are removed, except for the differential amplifier. • The R and N divider counters are forced to their load state conditions. • The charge pump is powered down with its outputs in threestate mode. • The digital lock detect circuitry is reset. • The RFIN input is debiased. • The reference input buffer circuitry is disabled. • The serial interface remains active and capable of loading and latching data. Rev. B | Page 20 of 28 ADF4193 MUX REGISTER (R6) RESERVED CONTROL BITS MUXOUT DB15 DB14 DB13 DB12 DB11 DB10 DB9 DB8 DB7 DB6 DB5 DB4 DB3 M13 M12 M11 M10 0 0 0 0 0 M4 M3 M2 M1 DB2 DB1 DB0 C3 (1) C2 (1) C1 (0) M13 M12 M11 M10 SIGMA-DELTA MODES M4 M3 M2 M1 MUXOUT 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 0 0 0 0 1 1 1 1 0 0 0 0 1 1 1 1 0 0 1 1 0 0 1 1 0 0 1 1 0 0 1 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 3-STATE DIGITAL LOCK DETECT N DIVIDER OUTPUT LOGIC HIGH R COUNTER RESERVED SERIAL DATA OUT LOGIC LOW R DIVIDER/2 OUTPUT N DIVIDER/2 OUTPUT RESERVED RESERVED ICP TIMEOUT SIGNAL SW1/2 TIMEOUT SIGNAL SW3 TIMEOUT SIGNAL RESERVED 0 1 0 1 1 0 0 1 ALL OTHER STATES INIT STATE, DITHER OFF, 3ns LOCK DETECT THRESHOLD DITHER ON 10ns LOCK DETECT THRESHOLD RESERVED 05328-029 SIGMA-DELTA AND LOCK DETECT MODES Figure 35. MUX Register (R6) MUXOUT Modes With C3, C2, and C1 set to 1, 1, 0, respectively, the MUX register is programmed. Σ-Δ and Lock Detect Modes Bit DB15 to Bit DB12 are used to reconfigure certain PLL operating modes. In the initialization sequence after power is applied to the chip, the four bits must first be programmed to all zeros. This initializes the PLL to a known state with dither off in the Σ-Δ modulator and a 3 ns PFD error threshold in the lock detect circuit. To turn on dither in the Σ-Δ modulator, an additional write should be made to Register R6 to program bits [DB15:DB12] = [0011]. However, for lowest noise operation, it is best to leave dither off. These bits control the on-chip multiplexer. See Figure 35 for the truth table. This pin is useful for diagnosis because it allows the user to look at various internal points of the chip, such as the R divider and INT divider outputs. In addition, it is possible to monitor the programmed timeout counter intervals on MUXOUT. For example, if the ICP timeout counter was programmed to 65 (with a 26 MHz PFD), then following the next write to R0, a pulse width of 10 μs would be observed on the MUXOUT pin. Digital lock detect is available via the MUXOUT pin. To change the lock detect threshold from 3 ns to 10 ns, a separate write to R6 should be performed to program bits [DB15:DB12] = [1001]. This should be done for reliable lock detect operation when the RF frequency is <2 GHz. A write to R6 that programs bits [DB15:DB12] = [0000] returns operation to the default state with both dither off and a 3 ns lock detect threshold. Reserved Bits The reserved bits must all be set to 0 for normal operation. Rev. B | Page 21 of 28 ADF4193 PROGRAMMING The ADF4193 can synthesize output frequencies with a channel step or resolution that is a fraction of the input reference frequency. For a given input reference frequency and a desired output frequency step, the first choice to make is the PFD reference frequency and the MOD. Once these are chosen, the desired output frequency channels are set by programming the INT and FRAC values. WORKED EXAMPLE In this example of a GSM900 RX system, it is required to generate RF output frequencies with channel steps of 200 kHz. A 104 MHz reference frequency input (REFIN) is available. The R divider setting that set the PFD reference is shown in Equation 1. FPFD = REFIN × [(1 + D)/(R × (1 + T))] (1) where: REFIN is the input reference frequency. D is the doubler enable bit (0 or 1). SPUR MECHANISMS The Fractional Spurs, Integer Boundary Spurs, and Reference Spurs sections describe the three different spur mechanisms that arise with a fractional-N synthesizer and how the ADF4193 can be programmed to minimize them. Fractional Spurs The fractional interpolator in the ADF4193 is a third-order, Σ-Δ modulator (SDM) with a modulus (MOD) that is programmable to any integer value from 13 to 4095. If dither is enabled, then the minimum allowed value of MOD is 50. The SDM is clocked at the PFD reference rate (fPFD) that allows PLL output frequencies to be synthesized at a channel step resolution of fPFD/MOD. With dither turned off, the quantization noise from the Σ-Δ modulator appears as fractional spurs. The interval between spurs is fPFD/L, where L is the repeat length of the code sequence in the digital Σ-Δ modulator. For the third-order modulator used in the ADF4193, the repeat length depends on the value of MOD, as shown in Table 7. R is the 4-bit R counter code (0…15). Table 7. Fractional Spurs with Dither Off T is the REF/2 bit (0 or 1). The maximum PFD reference frequency of 26 MHz is chosen and the following settings are programmed to give an R divider value of 4: Doubler enable = 0 R=2 Condition (Dither Off) If MOD is divisible by 2, but not 3 If MOD is divisible by 3, but not 2 If MOD is divisible by 6 Otherwise Repeat Length 2 × MOD Spur Interval Channel step/2 3 × MOD Channel step/3 6 × MOD MOD Channel step/6 Channel step REF/2 = 1 RFOUT is the desired RF output frequency. With dither enabled, the repeat length is extended to 221 cycles, regardless of the value of MOD, which makes the quantization error spectrum look like broadband noise. This can degrade the in-band phase noise at the PLL output by as much as 10 dB. Therefore, for the lowest noise, dither off is a better choice, particularly when the final loop BW is low enough to attenuate even the lowest frequency fractional spur. The wide loop bandwidth range available with the ADF4193 makes this possible in most applications. INT is the integer part of the division. Integer Boundary Spurs FRAC is the numerator part of the fractional division. Another mechanism for fractional spur creation involves interactions between the RF VCO frequency and the reference frequency. When these frequencies are not integer related, spur sidebands appear on the VCO output spectrum at an offset frequency that corresponds to the beat note or difference frequency between an integer multiple of the reference and the VCO frequency. Next, the modulus is chosen to allow fractional steps of 200 kHz. MOD = 26 MHz/200 kHz = 130 (2) Once the channel step is defined, the following equation shows how output frequency channels are programmed: RFOUT = [INT + (FRAC/MOD] × [FPFD] (3) where: MOD is the modulus or denominator part of the fractional division. For example, the frequency channel at 962.4 MHz is synthesized by programming the following values: INT = 37 FRAC = 2 Rev. B | Page 22 of 28 ADF4193 These spurs are attenuated by the loop filter and are more noticeable on channels close to integer multiples of the reference where the difference frequency can be inside the loop bandwidth, thus the name integer boundary spurs. The 8:1 loop bandwidth switching ratio of the ADF4193 makes it possible to attenuate all spurs to sufficiently low levels for most applications. The final loop BW can be chosen to ensure that all spurs are far enough out of band while meeting the lock time requirements with the 8× bandwidth boost. The ADF4193’s programmable modulus and R divider can also be used to avoid integer boundary channels. This option is described in the Avoiding Integer Boundary Channels section. Reference Spurs Reference spurs are generally not a problem in fractional-N synthesizers as the reference offset is far outside the loop bandwidth. However, any reference feedthrough mechanism that bypasses the loop can cause a problem. One such mechanism is feedthrough of low levels of on-chip reference switching noise out through the RFIN pin back to the VCO, resulting in reference spur levels as high as –90 dBc. These spurs can be suppressed below –110 dBc by inserting sufficient reverse isolation, for example, through an RF buffer between the VCO and RFIN pin. In addition, care should be taken in the PCB layout to ensure that the VCO is well separated from the input reference to avoid a possible feedthrough path on the board. POWER-UP INITIALIZATION Table 8. Power-Up Initialization Sequence Step 1 2 Register Bits R5 [7:0] R3 [15:0] Hex Codes FD 005B Wait 10 ms 3 4 R7 [15:0] R6 [15:0] 0007 000E 5 R6 [15:0] 900E 6 7 8 9 10 R4 [23:0] R4 [23:0] R4 [23:0] R2 [15:0] R1 [23:0] 004464 00446C 004394 00D2 520209 11 R0 [23:0] 480140 12 R3 [15:0] 007B 13 14 R5 [7:0] R0 [23:0] 05 480140 Description Set all power-down bits. PD polarity = 1, ground CPOUT+/ CPOUT–. Allow time for loop filter capacitors to discharge. Clear test modes. Initialize PLL modes, digital lock detect on MUXOUT. 10 ns lock detect threshold, digital lock detect on MUXOUT. SW1/SW2 timer = 10.8 μs. SW3 timer = 10.8 μs. ICP timer = 8.6 μs. Phase = 26. 8/9 prescaler, doubler disabled, R = 4, toggle FF on, MOD = 65. INT = 144, FRAC = 40 for 1880 MHz output frequency. PD polarity = 1, release CPOUT+/ CPOUT–. Clear all power-down bits. INT = 144, FRAC = 40 for 1880 MHz output frequency. The ADF4193 powers up after Step 13. It locks to the programmed channel frequency after Step 14. CHANGING THE FREQUENCY OF THE PLL AND THE PHASE LOOK-UP TABLE After applying power to the ADF4193, a 14-step sequence is recommended, as described in Table 8. The divider and timer setting used in the example in Table 8 is for a DCS1800 Tx synthesizer with a 104 MHz REFIN frequency. Once the ADF4193 is initialized, a write to Register R0 is all that is required to program a new output frequency. The N divider is updated with the values of INT and FRAC on the next PFD cycle following the LE edge that latches in the R0 word. However, the settling time and spurious performance of the synthesizer can be further optimized by modifying R1 and R2 register settings on a channel-by-channel basis. These settings are double buffered by the write to R0. This means that while the data is loaded in through the serial interface on the respective R1 and R2 write cycles, the synthesizer is not updated with their data until the next write to Register R0. The R2 register can be used to digitally adjust the phase of the VCO output relative to the reference edge. The phase can be adjusted over the full 360° range at RF with a resolution of 360°/MOD. In most frequency synthesizer applications, the actual phase offset of the VCO output with respect to the reference is unknown and does not matter. In such applications, the phase adjustment capability of the R2 register can instead be used to optimize the settling time performance, as described in the Phase Look-Up Table section. Rev. B | Page 23 of 28 ADF4193 Phase Look-Up Table Avoiding Integer Boundary Channels The ADF4193’s fast lock sequence is initiated following the write to Register R0. The fast lock timers are programmed so that after the PLL has settled in wide BW mode, the charge pump current is reduced and loop filter resistor switches are opened to reduce the loop BW. The reference cycle on which these events occur is determined by the values preprogrammed into the timeout counters. A further option when programming a new frequency involves a write to Register R1 to avoid integer boundary spurs. If it is found that the integer boundary spur level is too high, an option is to move the integer boundary away from the desired channel by reprogramming the R divider to select a different PFD frequency. For example, if REFIN = 104 MHz and R = 4 for a 26 MHz PFD reference and MOD = 130 for 200 kHz steps, the frequency channel at 910.2 MHz has a 200 kHz integer boundary spur because it is 200 kHz offset from 35 × 26 MHz. An alternative way to synthesize this channel is to set R = 5 for a 20.8 MHz PFD reference and MOD = 104 for 200 kHz steps. The 910.2 MHz channel is now 5 MHz offset from the nearest integer multiple of 20.8 MHz and the 5 MHz beat note spurs are well attenuated by the loop. Setting double buffered Bit R1 [23] = 1 (CP ADJ bit) increases the charge pump current by 25%, which compensates for the 25% increase in N with the change to the 20.8 MHz PFD frequency. This maintains constant loop dynamics and settling time performance for jumps between the two PFD frequencies. The CP ADJ bit should be cleared again when jumping back to 26 MHz-based channels. Figure 10 and Figure 13 show that the lock time to final phase is dominated by the phase swing that occurs when the BW is reduced. Once the PLL has settled to final frequency and phase, in wide BW mode, this phase swing is the same, regardless of the size of the synthesizer’s frequency jump. The amplitude of the phase swing is related to the current flowing through the loop filter zero resistors on the PFD reference cycle that the SW1/SW2 switches are opened. In an integer-N PLL, this current is zero once the PLL has settled. In a fractional-N PLL, the current is zero on average but varies from one reference cycle to the next, depending on the quantization error sequence output from the digital Σ-Δ modulator. Because the Σ-Δ modulator is all digital logic, clocked at the PFD reference rate, for a given value of MOD, the actual quantization error on any given reference cycle is determined by the value of FRAC and the PHASE word that the modulator is seeded with, following the write to R0. By choosing an appropriate value of PHASE, corresponding to the value of FRAC, that is programmed on the next write to R0, the size of the error current on the PFD reference cycle the SW1/SW2 switches opened, and thus the phase swing that occurs when the BW is reduced can be minimized. With dither off, the fractional spur pattern due to the SDM’s quantization noise also depends on the phase word the modulator is seeded with. Tables of optimized FRAC and phase values for popular SW1/SW2 and ICP timer settings can be down-loaded from the ADF4193 product page. If making use of a phase table, first write phase to double buffered Register R2, then write the INT and FRAC to R0. The Register R1 settings necessary for integer boundary spur avoidance are all double buffered and do not become active on the chip until the next write to Register R0. Register R0 should always be the last register written to when programming a new frequency. Serial Interface Activity The serial interface activity when programming the R2 or R1 registers causes no noticeable disturbance to the synthesizers settled phase or degradation in its frequency spectrum. Therefore, in a GSM application, it can be performed during the active part of the data burst. Because it takes just 10.2 μs to program the three registers, R2, R1, and R0, with the 6.5 MHz serial interface clock rate typically used, this programming can also be performed during the previous guard period with the LE edge to latch in the R0 data delayed until it’s time to switch frequency. Rev. B | Page 24 of 28 ADF4193 APPLICATIONS LOCAL OSCILLATOR FOR A GSM BASE STATION Timer Values for Tx Figure 36 shows the ADF4193 being used with a VCO to produce the LO for a GSM1800 base station. For GSM, the REFIN signal can be any integer multiple of 13 MHz, but the main requirement is that the slew rate is at least 300 V/μs. The 5 dBm, 104 MHz input sine wave shown satisfies this requirement. To comply with the GSM spectrum due to switching requirements, the Tx synthesizer should not switch frequency until the PA output power has ramped down by at least 50 dB. If it takes 10 μs to ramp down to this level, then only the last 20 μs of the 30 μs guard period is available for the Tx synthesizer to lock to final frequency and phase. Recommended parameters for the various GSM/PCS/DCS synthesizers are given in Table 9. In fast lock mode, the Tx loop BW is widened by a factor-of-8 to 480 kHz, and therefore, the PLL achieves frequency lock for a jump across the entire band in <6 μs. After this, the PA power can start to ramp up again, and the loop BW can be restored to the final value. With the ICP timer = 28, the charge pump current reduction begins at ~8.6 μs. When SW1, SW2, and SW3 timers = 35, the current reaches its final value before the loop filter switches open at ~10.8 μs. Table 9. Recommended Setup Parameters Parameter Loop BW PFD (MHz) MOD Dither Prescaler ICP Timer SW1, SW2, SW3 Timers VCO KV Tx 60 kHz 13 65 Off 4/5 28 35 GSM900 Rx 40 kHz 26 130 Off 4/5 78 85 18 MHz/V 18 MHz/V DCS1800/PCS1900 Tx Rx 60 kHz 40 kHz 13 13 65 65 Off Off 8/9 8/9 28 38 35 45 38 MHz/V 38 MHz/V With these timer values, the phase disturbance created when the bandwidth is reduced settles back to its final value by 20 μs, in time for the start of the active part of the GSM burst. If faster phase settling is desired with the 60 kHz BW setting, then the timer values can be reduced further but should not be brought less than the 6 μs it takes to achieve frequency lock in wide BW mode. Loop BW and PFD Frequency Timer Values for Rx A 60 kHz loop BW is narrow enough to attenuate the PLL phase noise and spurs to the required level for a Tx low. A 40 kHz BW is necessary to meet the GSM900 Rx synthesizer’s particularly tough phase noise and spur requirements at ±800 kHz offsets. To get the lowest spur levels at ±800 kHz offsets for Rx, the Σ-Δ modulator should be run at the highest oversampling rate possible. Therefore, for GSM900 Rx, a 26 MHz PFD frequency is chosen and MOD = 130 is required for 200 kHz steps. Because this value of MOD is divisible by two, certain FRAC channels have a 100 kHz fractional spur. This is attenuated by the 40 kHz loop filter and therefore is not a concern. However, the 60 kHz loop filter recommended for Tx has a closed-loop response that peaks close to 100 kHz. Therefore, a 13 MHz PFD with MOD = 65, which avoids the 100 kHz spur, is the best choice for a Tx synthesizer. The 40 kHz Rx loop BW is increased by a factor-of-8 to approximately 320 kHz during fast lock. With the Rx timer values shown, the BW is reduced after ~12 μs, which allows sufficient time for the phase disturbance to settle back before the start of the active part of the Rx time slot at 30 μs. As in the Tx case, faster Rx settling can be achieved by reducing these timer values, their lower limit being determined by the time it takes to achieve frequency lock in wide BW mode. In addition, the PCS and DCS Rx synthesizers have relaxed 800 kHz blocker specifications and thus can tolerate a wider loop BW, which allows correspondingly faster settling. Dither Dither off should be selected for the lowest rms phase error. Prescaler The 8/9 prescaler should be selected for the PCS and DCS bands. The 4/5 prescaler allows an N divider range low enough to cover the GSM900 Tx and Rx bands with either a 13 MHz or 26 MHz PFD frequency. VCO KV In general, the VCO gain, KV, should be set as low as possible to minimize the reference and integer boundary spur levels that arise due to feedthrough mechanisms. When deciding on the optimum VCO KV, a good choice is to allow 2 V to tune across the desired band, centered on the available tuning range. With VP3 regulated to 5.5 V ± 100 mV, the tuning range available is 2.8 V. Loop Filter Components It is important for good settling performance that capacitors with low dielectric absorption are used in the loop filter. Ceramic NPO COG capacitors are a good choice for this application. A 2% tolerance is recommended for loop filter capacitors and 1% for resistors. A 10% tolerance is adequate for the inductor, L1. Rev. B | Page 25 of 28 ADF4193 ADI SimPLL SUPPORT Also available is a technical note (ADF4193-TN-001) that outlines a loop filter design procedure that takes full advantage of the new degree of freedom in the filter design that the differential amplifier and loop filter switches provide. The ADF4193 loop filter design is supported on ADI SimPLL v2.7 or later. Example files for popular applications are available for download from the applications section of the ADF4193 product page. 10pF 18Ω 18Ω 100pF RF OUT 5V 10µF 18Ω + 3V 5.5V + 100nF 100nF 100nF 15 8, 10, 13 SVD VDD 100pF DVDD RFIN+ SW1 RFIN– 100pF REFERENCE 104MHz, +5dBm 1nF 51Ω 11 19 RSET 2.40kΩ 29 REFIN SW2 27 18 DATA 17 CLK CPOUT– 26 23 RSET MUXOUT 16 14 AGND 4, 22 C2A 1.20nF R1A2 820Ω 28 LE SDGND 100nF 7 VP2 32 AVDD VP3 10µF R1A2 6.20kΩ ADF4193 SWGND 1nF 24 30 51Ω 5 100pF VP 1 CPOUT+ 6 100nF 20 C1A 120pF 31 INTEGRATED DIFFERENTIAL AMPLIFIER AIN+ R3 62Ω 2 R2 1.80kΩ SW3 AIN– R1B2 6.20kΩ 3 AOUT 25 + 100nF L1 2.2mH C3 470pF Ct 30pF CMR R1B1 820Ω C2B 1.20nF 1 100nF SIRENZA VCO190-1843T 38MHz/V C1B 120pF DGND 9, 12, 21 LOCK DETECT OUT Figure 36. Local Oscillator for DCS1800 Tx Using the ADF4193 Rev. B | Page 26 of 28 05328-037 10µF ADF4193 INTERFACING ADSP-21xx Interface The ADF4193 has a simple SPI®-compatible serial interface for writing to the device. CLK, DATA, and LE control the data transfer. When LE goes high, the 24 bits that have been clocked into the input register on each rising edge of CLK are latched into the appropriate register. See Figure 2 for the timing diagram and Table 5 for the register address table. Figure 38 shows the interface between the ADF4193 and the ADSP-21xx digital signal processor. The ADF4193 needs a 24-bit serial word for some writes. The easiest way to accomplish this using the ADSP-21xx family is to use the autobuffered transmit mode of operation with alternate framing. This provides a means for transmitting an entire block of serial data before an interrupt is generated. Set up the word length for eight bits and use three memory locations for each 24-bit word. To program each 24-bit word, store the three 8-bit bytes, enable the autobuffered mode, and then write to the transmit register of the DSP. This last operation initiates the autobuffer transfer. ADuC812 Interface Figure 37 shows the interface between the ADF4193 and the ADuC812 MicroConverter®. Because the ADuC812 is based on an 8051 core, this interface can be used with any 8051-based microcontroller. The MicroConverter is set up for SPI master mode with CPHA = 0. To initiate the operation, the I/O port driving LE is brought low. Some registers of the ADF4193 require a 24-bit programming word. This is accomplished by writing three 8-bit bytes from the MicroConverter to the device. When the third byte is written, the LE input should be brought high to complete the transfer. An I/O port line on the ADuC812 can also be used to detect lock (MUXOUT configured as lock detect and polled by the port input). ADuC812 SCLOCK MOSI ADF4193 DATA LE 05328-033 I/O PORTS Figure 37. ADuC812 to ADF4193 Interface ADF4193 SCLK DT TFS I/O FLAGS CLK DATA LE MUXOUT (LOCK DETECT) Figure 38. ADSP-21xx to ADF4193 Interface PCB DESIGN GUIDELINES FOR CHIP SCALE PACKAGE The lands on the chip scale package (CP-32-3) are rectangular. The printed circuit board (PCB) pad for these should be 0.1 mm longer than the package land length and 0.05 mm wider than the package land width. The land should be centered on the pad. This ensures that the solder joint size is maximized. The bottom of the chip scale package has a central thermal pad. CLK MUXOUT (LOCK DETECT) ADSP-21xx 05328-034 The maximum allowable serial clock rate is 33 MHz. The thermal pad on the PCB should be at least as large as the exposed pad. On the PCB, there should be a clearance of at least 0.25 mm between the thermal pad and the inner edges of the pad pattern. This ensures that shorting is avoided. Thermal vias can be used on the PCB thermal pad to improve the thermal performance of the package. If vias are used, they should be incorporated in the thermal pad at 1.2 mm pitch grid. The via diameter should be between 0.3 mm and 0.33 mm, and the via barrel should be plated with one ounce copper to plug the via. The user should connect the PCB thermal pad to AGND. Rev. B | Page 27 of 28 ADF4193 OUTLINE DIMENSIONS 0.60 MAX 5.00 BSC SQ 0.60 MAX PIN 1 INDICATOR TOP VIEW 0.50 BSC 4.75 BSC SQ 0.50 0.40 0.30 32 1 EXPOSED PAD (BOTTOM VIEW) 17 16 3.45 3.30 SQ 3.15 9 8 0.25 MIN 3.50 REF 0.80 MAX 0.65 TYP 12° MAX 1.00 0.85 0.80 PIN 1 INDICATOR 25 24 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-VHHD-2 Figure 39. 32-Lead Lead Frame Chip Scale Package [LFCSP_VQ] 5 mm × 5 mm Body, Very Thin Quad (CP-32-3) Dimensions shown in millimeters ORDERING GUIDE Model ADF4193BCPZ 1 ADF4193BCPZ-RL1 ADF4193BCPZ-RL71 EVAL-ADF4193EB1 EVAL-ADF4193EB2 1 Temperature Range –40°C to +85°C –40°C to +85°C –40°C to +85°C Package Description 32-Lead Lead Frame Chip Scale Package [LFCSP_VQ] 32-Lead Lead Frame Chip Scale Package [LFCSP_VQ] 32-Lead Lead Frame Chip Scale Package [LFCSP_VQ] Evaluation Board (GSM 1800) Evaluation Board (No VCO or Loop Filter) Package Option CP-32-3 CP-32-3 CP-32-3 Z = Pb-free part. ©2006 Analog Devices, Inc. All rights reserved. Trademarks and registered trademarks are the property of their respective owners. D05328-0-6/06(B) T T Rev. B | Page 28 of 28