Dual/Quad Input Network Clock Generator/Synchronizer AD9547 FEATURES APPLICATIONS Supports Stratum 2 stability in holdover mode Supports reference switchover with phase build-out Supports hitless reference switchover Automatic/manual holdover and reference switchover 2 pairs of reference input pins, with each pair configurable as a single differential input or as 2 independent singleended inputs Input reference frequencies from 1 kHz to 750 MHz Reference validation and frequency monitoring (1 ppm) Programmable input reference switchover priority 30-bit programmable input reference divider 2 pairs of clock output pins, with each pair configurable as a single differential LVDS/LVPECL output or as 2 singleended CMOS outputs Output frequencies up to 450 MHz 20-bit integer and 10-bit fractional programmable feedback divider Programmable digital loop filter covering loop bandwidths from 0.001 Hz to 100 kHz Optional low noise LC-VCO system clock multiplier Optional crystal resonator for system clock input On-chip EEPROM to store multiple power-up profiles Software controlled power-down 64-lead LFCSP package Network synchronization Cleanup of reference clock jitter SONET/SDH clocks up to OC-192, including FEC Stratum 2 holdover, jitter cleanup, and phase transient control Stratum 3E and Stratum 3 reference clocks Wireless base stations, controllers Cable infrastructure Data communications GENERAL DESCRIPTION The AD9547 provides synchronization for many systems, including synchronous optical networks (SONET/SDH). The AD9547 generates an output clock that is synchronized to one of two differential or four single-ended external input references. The digital PLL allows for reduction of input time jitter or phase noise associated with the external references. The AD9547 continuously generates a clean (low jitter), valid output clock, even when all references fail, by means of digitally controlled loop and holdover circuitry. The AD9547 operates over an industrial temperature range of −40°C to +85°C. FUNCTIONAL BLOCK DIAGRAM ANALOG FILTER STABLE SOURCE AD9547 CLOCK MULTIPLIER CLOCK DISTRIBUTION CHANNEL 0 DIVIDER DIGITAL PLL CHANNEL 1 DIVIDER DAC REFERENCE INPUTS AND MONITOR MUX SYNC EEPROM STATUS AND CONTROL PINS 08300-001 SERIAL CONTROL INTERFACE (SPI or I2C) Figure 1. Rev. 0 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 ©2009 Analog Devices, Inc. All rights reserved. AD9547 TABLE OF CONTENTS Features .............................................................................................. 1 Digital Phase-Locked Loop (DPLL) Core............................... 32 Applications ....................................................................................... 1 Direct Digital Synthesizer (DDS) ............................................. 34 General Description ......................................................................... 1 Tuning Word Processing ........................................................... 35 Functional Block Diagram .............................................................. 1 Loop Control State Machine ..................................................... 36 Revision History ............................................................................... 3 System Clock Inputs................................................................... 37 Specifications..................................................................................... 4 SYSCLK PLL Multiplier............................................................. 38 Supply Voltage ............................................................................... 4 Clock Distribution ..................................................................... 39 Supply Current .............................................................................. 4 Status and Control .......................................................................... 44 Power Dissipation ......................................................................... 4 Multifunction Pins (M0 to M7) ............................................... 44 Logic Inputs (M0 to M7, RESET)............................................... 5 IRQ Pin ........................................................................................ 45 Logic Outputs (M0 to M7, IRQ) ................................................ 5 Watchdog Timer ......................................................................... 45 System Clock Inputs (SYSCLKP, SYSCLKN)............................ 5 EEPROM ..................................................................................... 46 Distribution Clock Inputs (CLKINP, CLKINN) ...................... 6 Serial Control Port ......................................................................... 50 Reference Inputs (REFA/REFAA, REFB/REFBB) .................... 7 SPI/I2C Port Selection ................................................................ 50 Reference Monitors ...................................................................... 7 SPI Serial Port Operation .......................................................... 50 Reference Switchover Specifications .......................................... 8 I2C Serial Port Operation .......................................................... 54 Distribution Clock Outputs (OUT0, OUT1) ........................... 8 I/O Programming Registers .......................................................... 57 DAC Output Characteristics (DACOUTP, DACOUTN) ....... 9 Buffered/Active Registers .......................................................... 57 Time Duration of Digital Functions ........................................ 10 Autoclearing Registers ............................................................... 57 Digital PLL .................................................................................. 10 Register Access Restrictions...................................................... 57 Digital PLL Lock Detection ...................................................... 10 Register Map ................................................................................... 58 Holdover Specifications ............................................................. 10 Register Bit Descriptions ............................................................... 67 Serial Port Specifications—SPI Mode ...................................... 11 Serial Port Configuration and Part Identification (Register 0x0000 to Register 0x0005) ...................................... 67 Serial Port Specifications—I C Mode ...................................... 12 2 Jitter Generation ......................................................................... 13 Absolute Maximum Ratings.......................................................... 14 ESD Caution ................................................................................ 14 Pin Configuration and Function Descriptions ........................... 15 Typical Performance Characteristics ........................................... 18 Input/Output Termination Recommendations .......................... 23 Getting Started ................................................................................ 24 Power-On Reset .......................................................................... 24 Initial M0 to M7 Pin Programming ......................................... 24 Device Register Programming .................................................. 24 System Clock (SYSCLK) (Register 0x0100 to Register 0x0108) ......................................................................... 68 General Configuration (Register 0x0200 to Register 0x0214) ......................................................................... 69 DPLL Configuration (Register 0x0300 to Register 0x031B)......................................................................... 72 Clock Distribution Output Configuration (Register 0x0400 to Register 0x0419) ...................................... 74 Reference Input Configuration (Register 0x0500 to Register 0x0507) .......................................................................... 76 Profile Registers (Register 0x0600 to Register 0x07FF) ........ 78 Theory of Operation ...................................................................... 26 Operational Controls (Register 0x0A00 to Register 0x0A10) ........................................................................ 88 Overview...................................................................................... 26 Status Readback (Register 0x0D00 to Register 0x0D19) ...... 92 Reference Clock Inputs .............................................................. 27 Reference Monitors .................................................................... 27 Reference Profiles ....................................................................... 28 Reference Switchover ................................................................. 30 Nonvolatile Memory (EEPROM) Control (Register 0x0E00 to Register 0x0E03) ..................................... 94 EEPROM Storage Sequence (Register 0x0E10 to Register 0x0E3F) .......................................................................... 95 Rev. 0 | Page 2 of 104 AD9547 Applications Information ...............................................................98 Outline Dimensions ......................................................................102 Power Supply Partitions .............................................................98 Ordering Guide .........................................................................102 Thermal Performance.................................................................98 Calculating the Digital Filter Coefficients ...............................99 REVISION HISTORY 7/09—Revision 0: Initial Version Rev. 0 | Page 3 of 104 AD9547 SPECIFICATIONS Minimum and maximum values apply for the full range of supply voltage and operating temperature variation. Typical values apply for AVDD3 = DVDD3 = 3.3 V, AVDD = DVDD = 1.8 V, TA = 25°C, IDAC = 20 mA (full scale), unless otherwise noted. SUPPLY VOLTAGE Table 1. Parameter DVDD3 DVDD AVDD3 3.3 V Supply (Typical) 1.8 V Supply (Alternative) AVDD Min 3.135 1.71 3.135 3.135 1.71 1.71 Typ 3.30 1.80 3.30 3.30 1.80 1.80 Max 3.465 1.89 3.465 3.465 1.89 1.89 Unit V V V V V V Test Conditions/Comments Pin 7, Pin 58 Pin 1, Pin 6, Pin 8, Pin 10, Pin 11, Pin 53, Pin 59, Pin 64 Pin 16, Pin 33, Pin 43, Pin 49 Pin 25, Pin 31 Pin 25, Pin 31 Pin 17, Pin 18, Pin 23, Pin 28, Pin 32, Pin 36, Pin 39, Pin 42, Pin 46, Pin 50 SUPPLY CURRENT The test conditions for the maximum supply current are the same as the test conditions for the All Blocks Running section of Table 3. The test conditions for the typical supply current are the same as the test conditions for the Typical Configuration section of Table 3. Table 2. Parameter IDVDD3 IDVDD IAVDD3 3.3 V Supply (Typical) 1.8 V Supply (Alternative) IAVDD Min Typ 1.5 190 52 24 24 135 Max 3 215 70 55 55 150 Unit mA mA mA mA mA mA Test Conditions/Comments Pin 7, Pin 58 Pin 1, Pin 6, Pin 8, Pin 10, Pin 11, Pin 53, Pin 59, Pin 64 Pin 16, Pin 33, Pin 43, Pin 49 Pin 25, Pin 31 Pin 25, Pin 31 Pin 17, Pin 18, Pin 23, Pin 28, Pin 32, Pin 36, Pin 39, Pin 42, Pin 46, Pin 50 Min Typ 800 Max 1100 Unit mW ALL BLOCKS RUNNING 900 1250 mW FULL POWER-DOWN 13 mW −105 mW Test Conditions/Comments fSYSCLK = 20 MHz 1 ; fS = 1 GHz 2 ; fDDS = 122.88 MHz 3 ; one LVPECL clock distribution output running at 122.88 MHz (all others powered down); one input reference running at 100 MHz (all others powered down) fSYSCLK = 20 MHz1; fS = 1 GHz2; fDDS = 399 MHz3; all clock distribution outputs configured as LVPECL at 399 MHz; all input references configured as differential at 100 MHz; fractional-N active (R = 10, S = 39, U = 9, V = 10) Conditions = typical configuration; no external pull-up or pulldown resistors Conditions = typical configuration; table values show the change in power due to the indicated operation fSYSCLK = 1 GHz1; high frequency direct input mode 7 13 mW mW 70 75 65 mW mW mW POWER DISSIPATION Table 3. Parameter TYPICAL CONFIGURATION INCREMENTAL POWER DISSIPATION SYSCLK PLL Off Input Reference On Differential Single-Ended Output Distribution Driver On LVDS LVPECL CMOS 1 2 3 Single 3.3 V CMOS output with a 10 pF load fSYSCLK is the frequency at the SYSCLKP and SYSCLKN pins. fS is the sample rate of the output DAC. fDDS is the output frequency of the DDS. Rev. 0 | Page 4 of 104 AD9547 LOGIC INPUTS (M0 TO M7, RESET) Table 4. Parameter INPUT VOLTAGE Input High Voltage (VIH) Input Low Voltage (VIL) INPUT CURRENT (IINH, IINL) INPUT CAPACITANCE (CIN) Min Typ Max Unit 0.8 ±200 V V μA pF Max Unit Test Conditions/Comments 0.4 V V 1 1 μA μA IOH = 1 mA IOL = 1 mA Open-drain mode VOH = 3.3 V VOL = 0 V Max Unit Test Conditions/Comments 1000 MHz V/μs % V mV p-p 2.1 ±80 3 Test Conditions/Comments LOGIC OUTPUTS (M0 TO M7, IRQ) Table 5. Parameter OUTPUT VOLTAGE Output High Voltage (VOH) Output Low Voltage (VOL) IRQ LEAKAGE CURRENT Active Low Output Mode Active High Output Mode Min Typ 2.7 SYSTEM CLOCK INPUTS (SYSCLKP, SYSCLKN) Table 6. Parameter SYSTEM CLOCK PLL BYPASSED Input Frequency Range Minimum Input Slew Rate Duty Cycle Common-Mode Voltage Differential Input Voltage Sensitivity Input Capacitance Input Resistance SYSTEM CLOCK PLL ENABLED PLL Output Frequency Range Phase Frequency Detector (PFD) Rate Frequency Multiplication Range VCO Gain High Frequency Path Input Frequency Range Minimum Input Slew Rate Frequency Divider Range Common-Mode Voltage Differential Input Voltage Sensitivity Input Capacitance Input Resistance Min Typ 500 1000 40 60 1.2 100 2 2.5 900 pF kΩ 1000 150 255 6 70 100.1 200 1 Internally generated Minimum voltage across pins is required to ensure switching between logic states; the instantaneous voltage on either pin must not exceed the supply rails; ac ground the unused input to accommodate single-ended operation Single-ended, each pin MHz MHz Assumes valid system clock and PFD rates MHz/V 500 MHz V/μs 8 1 V mV pp 3 2.5 pF kΩ 100 Minimum limit imposed for jitter performance Rev. 0 | Page 5 of 104 Minimum limit imposed for jitter performance Binary steps (M = 1, 2, 4, 8) Internally generated This is the minimum voltage required across the pins to ensure switching between logic states; the instantaneous voltage on either pin must not exceed the supply rails; ac ground the unused input to accommodate single-ended operation Single-ended, each pin AD9547 Parameter Low Frequency Path Input Frequency Range Minimum Input Slew Rate Common-Mode Voltage Differential Input Voltage Sensitivity Input Capacitance Input Resistance Crystal Resonator Path Crystal Resonator Frequency Range Maximum Crystal Motional Resistance Min Typ Max Unit 100 1.2 MHz V/μs V mV pp 3 4 pF kΩ 3.5 50 100 10 50 100 MHz Ω Max 500 Test Conditions/Comments Minimum limit imposed for jitter performance Internally generated This is the minimum voltage required across the pins to ensure switching between logic states; the instantaneous voltage on either pin must not exceed the supply rails; ac ground the unused input to accommodate single-ended operation Single-ended, each pin Fundamental mode, AT cut See the System Clock Inputs section for recommendations DISTRIBUTION CLOCK INPUTS (CLKINP, CLKINN) Table 7. Parameter INPUT FREQUENCY RANGE MINIMUM SLEW RATE COMMON-MODE VOLTAGE DIFFERENTIAL INPUT VOLTAGE SENSITIVITY Min 62.5 75 100 Unit MHz V/μs mV mV p-p DIFFERENTIAL INPUT POWER SENSITIVITY −15 dBm INPUT CAPACITANCE INPUT RESISTANCE Typ 700 3 5 pF kΩ Rev. 0 | Page 6 of 104 Test Conditions/Comments Minimum limit imposed for jitter performance Internally generated Capacitive coupling required; ac ground the unused input to accommodate single-ended operation; the instantaneous voltage on either pin must not exceed the supply rails Same as voltage sensitivity but specified as power into a 50 Ω load Each pin has a 2.5 kΩ internal dc bias resistance AD9547 REFERENCE INPUTS (REFA/REFAA, REFB/REFBB) Table 8. Parameter DIFFERENTIAL OPERATION Frequency Range Sinusoidal Input LVPECL Input LVDS Input Minimum Input Slew Rate Common-Mode Input Voltage Differential Input Voltage Sensitivity Input Resistance Input Capacitance Minimum Pulse Width High Minimum Pulse Width Low SINGLE-ENDED OPERATION Frequency Range (CMOS) Minimum Input Slew Rate Input Voltage High (VIH) 1.2 V to 1.5 V Threshold Setting 1.8 V to 2.5 V Threshold Setting 3.0 V to 3.3 V Threshold Setting Input Voltage Low (VIL) 1.2 V to 1.5 V Threshold Setting 1.8 V to 2.5 V Threshold Setting 3.0 V to 3.3 V Threshold Setting Input Resistance Input Capacitance Minimum Pulse Width High Minimum Pulse Width Low Min Typ 10 0.001 0.001 40 Max Unit 750 750 750 MHz MHz MHz V/μs V mV 2 ±65 25 3 Minimum limit imposed for jitter performance Internally generated This is the minimum voltage required across the pins to ensure switching between logic states; the instantaneous voltage on either pin must not exceed the supply rails kΩ pF ps ps 620 620 10 40 Test Conditions/Comments 250 0.9 1.2 1.9 MHz V/μs Minimum limit imposed for jitter performance V V V 0.27 0.5 1.0 V V V kΩ pF ns ns Max Unit Test Conditions/Comments 9.54 × 10 1.2 0.1 NPDP Δf/fREF NPDP = nominal phase detector period (NPDP = fREF/R) 1 Programmable (lower bound subject to quality of SYSCLK) 0.001 0.001 65.535 65.535 sec sec Programmable in 1 ms increments Programmable in 1 ms increments 45 3 1.5 1.5 REFERENCE MONITORS Table 9. Parameter REFERENCE MONITOR Loss of Reference Detection Time Frequency Out-of-Range Limits TIMERS Validation Timer Redetect Timer 1 Min Typ −7 fREF is the frequency of the active reference; R is the frequency division factor determined by the R divider. Rev. 0 | Page 7 of 104 AD9547 REFERENCE SWITCHOVER SPECIFICATIONS Table 10. Parameter MAXIMUM OUTPUT PHASE PERTURBATION (PHASE BUILD-OUT SWITCHOVER) MAXIMUM TIME/TIME SLOPE (HITLESS SWITCHOVER) Min Max 200 Unit ps Test Conditions/Comments Assumes a jitter-free reference; satisfies Telcordia GR-1244-CORE requirements 65,535 ns/sec Minimum/maximum values are programmable upper bounds; a minimum value ensures <10% error; satisfies Telcordia GR-1244-CORE requirements 5 NPDP 3 NPDP NPDP = nominal phase detector period (NPDP = fREF/R) 1 NPDP = nominal phase detector period (NPDP = fREF/R)1 315 TIME REQUIRED TO SWITCH TO A NEW REFERENCE Hitless Switchover Phase Build-Out Switchover 1 Typ 40 fREF is the frequency of the active reference; R is the frequency division factor determined by the R divider. DISTRIBUTION CLOCK OUTPUTS (OUT0, OUT1) Table 11. Parameter LVPECL MODE Maximum Output Frequency Rise/Fall Time1 (20% to 80%) Duty Cycle Differential Output Voltage Swing Common-Mode Output Voltage LVDS MODE Maximum Output Frequency Rise/Fall Time 1 (20% to 80%) Duty Cycle Differential Output Voltage Swing Balanced (VOD) Min Typ 725 180 Short-Circuit Output Current CMOS MODE Maximum Output Frequency 3.3 V Supply Strong Drive Strength Setting Weak Drive Strength Setting 1.8 V Supply Unit MHz ps % mV Test Conditions/Comments Using internal current setting resistor (nominal 3.12 kΩ) 45 630 770 315 55 910 AVDD3 − 1.5 AVDD3 − 1.3 AVDD3 − 1.05 V Magnitude of voltage across pins; output driver static Output driver static Using internal current setting resistor (nominal 3.12 kΩ) MHz ps % 100 Ω termination across the output pins 40 350 60 247 454 mV 50 mV 1.375 50 V mV 24 mA 725 200 Unbalanced (ΔVOD) Offset Voltage Common Mode (VOS) Common-Mode Difference (ΔVOS) Max 1.125 13 100 Ω termination across output pins Voltage swing between output pins; output driver static Absolute difference between voltage swing of normal pin and inverted pin; output driver static Output driver static Voltage difference between pins; output driver static Output driver static Weak drive option not supported for operating the CMOS drivers using a 1.8 V supply 10 pF load 250 25 150 MHz MHz MHz Rev. 0 | Page 8 of 104 AD9547 Parameter Rise/Fall Time1 (20% to 80%) 3.3 V Supply Strong Drive Strength Setting Weak Drive Strength Setting 1.8 V Supply Duty Cycle Output Voltage High (VOH) AVDD3 = 3.3 V, IOH = 10 mA AVDD3 = 3.3 V, IOH = 1 mA AVDD3 = 1.8 V, IOH = 1 mA Output Voltage Low (VOL) Min Max Unit 0.5 8 1.5 2 14.5 2.5 60 ns ns ns % 40 2.6 2.9 1.5 Test Conditions/Comments 10 pF load 10 pF load Output driver static; strong drive strength setting V V V Output driver static; strong drive strength setting AVDD3 = 3.3 V, IOL = 10 mA AVDD3 = 3.3 V, IOL = 1 mA AVDD3 = 1.8 V, IOL = 1 mA OUTPUT TIMING SKEW Between LVPECL Outputs Between LVDS Outputs Between CMOS (3.3 V) Outputs Strong Drive Strength Setting Weak Drive Strength Setting Between CMOS (1.8 V) Outputs Between LVPECL Outputs and LVDS Outputs Between LVPECL Outputs and CMOS Outputs ZERO-DELAY TIMING SKEW 1 Typ 0.3 0.1 0.1 V V V 14 13 125 138 ps ps 23 24 40 14 240 ps ps ps ps 140 10 pF load Rising edge only; any divide value Rising edge only; any divide value Weak drive option not supported at 1.8 V 19 ps ±5 ns Output relative to active input reference; output distribution synchronization to active reference feature enabled; assumes manual phase offset compensation of deterministic latency Max 450 15 Unit MHz mV Test Conditions/Comments VSS + 0.5 V Ω The listed values are for the slower edge (rising or falling). DAC OUTPUT CHARACTERISTICS (DACOUTP, DACOUTN) Table 12. Parameter FREQUENCY RANGE OUTPUT OFFSET VOLTAGE Min 62.5 VOLTAGE COMPLIANCE RANGE OUTPUT RESISTANCE VSS − 0.5 OUTPUT CAPACITANCE FULL-SCALE OUTPUT CURRENT GAIN ERROR Typ 0.5 50 5 20 −12 pF mA +12 Rev. 0 | Page 9 of 104 % FS This is the single-ended voltage at either DAC output pin (no external load) when the internal DAC code is such that no current is delivered to that pin Single-ended; each pin has an internal 50 Ω termination to VSS Programmable (8 mA to 31 mA; see the DAC Output section) AD9547 TIME DURATION OF DIGITAL FUNCTIONS Table 13. Parameter EEPROM-TO-REGISTER DOWNLOAD TIME Min Typ 25 Max Unit ms REGISTER-TO-EEPROM UPLOAD TIME 200 ms MINIMUM POWER-DOWN EXIT TIME MAXIMUM TIME FROM ASSERTION OF THE RESET PIN TO THE M0 TO M7 PINS ENTERING HIGH IMPEDANCE STATE 10.5 45 μs ns Test Conditions/Comments Using default EEPROM storage sequence (see Register 0x0E10 to Register 0x0E3F) Using default EEPROM storage sequence (see Register 0x0E10 to Register 0x0E3F Dependent on loop filter bandwidth DIGITAL PLL Table 14. Parameter PHASE FREQUENCY DETECTOR (PFD) INPUT FREQUENCY RANGE LOOP BANDWIDTH PHASE MARGIN REFERENCE INPUT (R) DIVISION FACTOR INTEGER FEEDBACK (S) DIVISION FACTOR FRACTIONAL FEEDBACK DIVIDE RATIO 1 2 3 Min 0.001 Typ Max 10 Unit MHz Test Conditions/Comments Maximum fPFD = fS/100 1, 2 0.001 1 × 105 Hz 30 1 8 0 89 230 220 0.999 Degrees Programmable design parameter; maximum fLOOP = fREF/(20R) 3 Programmable design parameter 1, 2, …1,073,741,824 8, 9, …1,048,576 Maximum value = 1022/1023 fPFD is the frequency at the input to the phase-frequency detector. fS is the sample rate of the output DAC. fREF is the frequency of the active reference; R is the frequency division factor determined by the R divider. DIGITAL PLL LOCK DETECTION Table 15. Parameter PHASE LOCK DETECTOR Threshold Programming Range Threshold Resolution FREQUENCY LOCK DETECTOR Threshold Programming Range Threshold Resolution Min Typ 0.001 Max Unit 65.5 ns ps 16,700 ns ps Reference-to-feedback period difference Max Unit ppm Test Conditions/Comments Excludes frequency drift of SYSCLK source; excludes frequency drift of input reference prior to entering holdover 1 0.001 1 Test Conditions/Comments HOLDOVER SPECIFICATIONS Table 16. Parameter FREQUENCY ACCURACY Min Typ <0.01 Rev. 0 | Page 10 of 104 AD9547 SERIAL PORT SPECIFICATIONS—SPI MODE Table 17. Parameter CS Input Logic 1 Voltage Input Logic 0 Voltage Input Logic 1 Current Input Logic 0 Current Input Capacitance SCLK Input Logic 1 Voltage Input Logic 0 Voltage Input Logic 1 Current Input Logic 0 Current Input Capacitance SDIO As an Input Input Logic 1 Voltage Input Logic 0 Voltage Input Logic 1 Current Input Logic 0 Current Input Capacitance As an Output Output Logic 1 Voltage Output Logic 0 Voltage SDO Output Logic 1 Voltage Output Logic 0 Voltage TIMING SCLK Clock Rate, 1/tCLK Pulse Width High, tHIGH Pulse Width Low, tLOW SDIO to SCLK Setup, tDS SCLK to SDIO Hold, tDH SCLK to Valid SDIO and SDO, tDV CS to SCLK Setup, tS CS to SCLK Hold, tC CS Minimum Pulse Width High Min Typ Max Unit 2.0 0.8 30 110 2 V V μA μA pF 2.0 0.8 1 1 2 V V μA μA pF 2.0 0.8 1 1 2 V V μA μA pF Test Conditions/Comments Internal 30 kΩ pull-up resistor Internal 30 kΩ pull-down resistor 2.7 0.4 V V 1 mA load current 1 mA load current 0.4 V V 1 mA load current 1 mA load current 2.7 40 8 12 3 0 14 10 0 6 Rev. 0 | Page 11 of 104 MHz ns ns ns ns ns ns ns ns AD9547 SERIAL PORT SPECIFICATIONS—I2C MODE Table 18. Parameter SDA (AS INPUT), SCL Min Input Logic 1 Voltage Input Logic 0 Voltage Input Current Hysteresis of Schmitt Trigger Inputs Pulse Width of Spikes That Must Be Suppressed by the Input Filter, tSP SDA (AS OUTPUT) Output Logic 0 Voltage Output Fall Time from VIHmin to VILmax TIMING SCL Clock Rate Bus Free Time Between a Stop and Start Condition, tBUF Repeated Start Condition Setup Time, tSU;STA Repeated Hold Time Start Condition, tHD;STA 0.7 × DVDD3 Stop Condition Setup Time, tSU;STO Low Period of the SCL Clock, tLOW High Period of the SCL Clock, tHIGH SCL/SDA Rise Time, tR SCL/SDA Fall Time, tF Data Setup Time, tSU;DAT Data Hold Time, tHD;DAT Capacitive Load for Each Bus Line, Cb1 1 Typ Max Unit Test Conditions/Comments No internal pull-up/pull-down resistor For VIN = 10% to 90% of DVDD3 50 V V μA V ns 0.4 250 V ns IO = 3 mA 10 pF ≤ Cb ≤ 400 pF 400 1.3 kHz μs 0.6 0.6 μs μs 0.6 1.3 0.6 20 + 0.1 Cb1 20 + 0.1 Cb1 100 100 μs μs μs ns ns ns ns pF −10 0.015 × DVDD3 20 + 0.1 Cb 1 0.3 × DVDD3 +10 300 300 400 Cb is the capacitance (pF) of a single bus line. Rev. 0 | Page 12 of 104 After this period, the first clock pulse is generated AD9547 JITTER GENERATION Table 19. Parameter CONDITIONS: fREF = 8 kHz 1 , fDDS = 155.52 MHz 2 , fLOOP = 100 Hz 3 Min Typ Max Unit Bandwidth: 100 Hz to 77 MHz Bandwidth: 5 kHz to 20 MHz Bandwidth: 20 kHz to 80 MHz Bandwidth: 50 kHz to 80 MHz Bandwidth: 4 MHz to 80 MHz CONDITIONS: fREF = 19.44 MHz1, fDDS = 155.52 MHz2, fLOOP = 1 kHz3 0.71 0.34 0.43 0.43 0.31 ps rms ps rms ps rms ps rms ps rms Bandwidth: 100 Hz to 77 MHz Bandwidth: 5 kHz to 20 MHz Bandwidth: 20 kHz to 80 MHz Bandwidth: 50 kHz to 80 MHz Bandwidth: 4 MHz to 80 MHz CONDITIONS: fREF = 19.44 MHz1, fDDS = 311.04 MHz2, fLOOP = 1 kHz3 1.05 0.34 0.43 0.43 0.32 ps rms ps rms ps rms ps rms ps rms 0.67 0.31 0.33 0.33 0.16 ps rms ps rms ps rms ps rms ps rms Bandwidth: 100 Hz to 100 MHz Bandwidth: 5 kHz to 20 MHz Bandwidth: 20 kHz to 80 MHz Bandwidth: 50 kHz to 80 MHz Bandwidth: 4 MHz to 80 MHz 1 fREF is the frequency of the active reference. fDDS is the output frequency of the DDS. fLOOP is the DPLL digital loop filter bandwidth. 4 fSYSCLK is the frequency at the SYSCLKP and SYSCLKN pins. 5 fS is the sample rate of the output DAC. 2 3 Rev. 0 | Page 13 of 104 Test Conditions/Comments fSYSCLK = 50 MHz 4 crystal; fS = 1 GHz 5 ; Q-divider = 1; default SYSCLK PLL charge pump current; results valid for LVPECL, LVDS, and CMOS output logic types Random jitter Random jitter Random jitter Random jitter Random jitter fSYSCLK = 50 MHz4 crystal; fS = 1 GHz5; Q-divider = 1; default SYSCLK PLL charge pump current; results valid for LVPECL, LVDS, and CMOS output logic types Random jitter Random jitter Random jitter Random jitter Random jitter fSYSCLK = 50 MHz4 crystal; fS = 1 GHz5; Q-divider = 1; default SYSCLK PLL charge pump current; results valid for LVPECL, LVDS, and CMOS output logic types Random jitter Random jitter Random jitter Random jitter Random jitter AD9547 ABSOLUTE MAXIMUM RATINGS Table 20. Parameter Analog Supply Voltage (AVDD) Digital Supply Voltage (DVDD) Digital I/O Supply Voltage (DVDD3) DAC Supply Voltage (AVDD3) Maximum Digital Input Voltage Storage Temperature Range Operating Temperature Range Lead Temperature (Soldering 10 sec) Junction Temperature Rating 2V 2V 3.6 V 3.6 V −0.5 V to DVDD3 + 0.5 V −65°C to +150°C −40°C to +85°C 300°C 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. ESD CAUTION 150°C Rev. 0 | Page 14 of 104 AD9547 64 63 62 61 60 59 58 57 56 55 54 53 52 51 50 49 DVDD M7 M6 M5 M4 DVDD DVDD3 M3 M2 M1 M0 DVDD IRQ NC AVDD AVDD3 PIN CONFIGURATION AND FUNCTION DESCRIPTIONS 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 PIN 1 INDICATOR AD9547 TOP VIEW (Not to Scale) 48 47 46 45 44 43 42 41 40 39 38 37 36 35 34 33 REFBB REFB AVDD REFAA REFA AVDD3 AVDD TDC_VRT TDC_VRB AVDD SYSCLKP SYSCLKN AVDD SYSCLK_LF SYSCLK_VREG AVDD3 NOTES 1. NC = NO CONNECT. 2. THE EXPOSED PAD MUST BE CONNECTED TO GROUND (VSS). 08300-002 AVDD AVDD VSS CLKINN CLKINP VSS AVDD OUT_RSET AVDD3 OUT0P OUT0N AVDD OUT1P OUT1N AVDD3 AVDD 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 DVDD SCLK/SCL SDIO SDO CS/SDA DVDD DVDD3 DVDD RESET DVDD DVDD VSS DACOUTP DACOUTN VSS AVDD3 Figure 2. Pin Configuration Table 21. Pin Function Descriptions Input/ Output I Pin Type Power Mnemonic DVDD Description 1.8 V Digital Supply. I 3.3 V CMOS SCLK/SCL 3 I/O 3.3 V CMOS SDIO 4 O 3.3 V CMOS SDO 5 I 3.3 V CMOS CS/SDA 7, 58 9 I I Power 3.3 V CMOS DVDD3 RESET 10, 11 I Power DVDD 12, 15, 19, 22 13 O O VSS DACOUTP 14 O 16 17, 18 I I Ground Differential output Differential output Power Power Serial Programming Clock. Data clock for serial programming. SCLK is used for SPI mode, and SCL is used for I2C® mode. Serial Data Input/Output. When the device is in 4-wire mode, data is written via this pin. In 3-wire mode, both data reads and writes occur on this pin. There is no internal pull-up/pull-down resistor on this pin. Serial Data Output. Use this pin to read data in 4-wire mode (high impedance in 3-wire mode). There is no internal pull-up/pull-down resistor on this pin. SPI Mode. Chip Select (CS) Input. Active low. When programming a device, this pin must be held low. In systems where more than one AD9547 is present, this pin enables individual programming of each AD9547. In SPI mode, this pin has an internal 30 kΩ pull-up resistor. I2C Mode. Serial Data Line (SDA) Input/Output. In I2C Mode, this pin is an output during read operations and an input during write operations. There is no internal pull-up resistor in I2C mode. 3.3 V I/O Digital Supply. Chip Reset. Assertion of this pin (active high) resets the device. This pin has an internal 50 kΩ pull-down resistor. 1.8 V DAC Decode Digital Supply. Isolate the supply associated with these DVDD pins from the supply associated with the other DVDD pins. Analog Ground. Connect to ground. DAC Output. DACOUTP contains an internal 50 Ω pull-down resistor. Pin No. 1, 6, 8, 53, 59, 64 2 DACOUTN AVDD3 AVDD Complementary DAC Output. DACOUTN contains an internal 50 Ω pull-down resistor. 3.3 V Analog (DAC) Power Supply. 1.8 V Analog (DAC) Power Supply. Rev. 0 | Page 15 of 104 AD9547 Pin No. 20 Input/ Output I 21 I 23 24 I O 25, 31 I 26 O LVPECL, LVDS, or CMOS OUT0P 27 O OUT0N 28, 32 29 I O LVPECL, LVDS, or CMOS Power LVPECL, LVDS, or CMOS 30 O OUT1N 33 34 I I LVPECL, LVDS, or CMOS Power 35 O 36, 39 37 I I Power Differential input AVDD SYSCLKN 38 I Differential input SYSCLKP 40, 41 I 42 43, 49 44 I I I Power Power Differential input 45 I 46, 50 I Differential input Power Pin Type Differential input Mnemonic CLKINN Differential input Power Current set resistor Power CLKINP AVDD OUT_RSET AVDD3 AVDD OUT1P AVDD3 SYSCLK_VREG SYSCLK_LF TDC_VRB, TDC_VRT AVDD AVDD3 REFA REFAA AVDD Description Clock Distribution Input. In standard operating mode, this pin is connected to the filtered DACOUTN output. This internally biased input is typically ac-coupled, and, when configured as such, can accept any differential signal with a singleended swing of at least 400 mV. Clock Distribution Input. In standard operating mode, this pin is connected to the filtered DACOUTP output. 1.8 V Analog (Input Receiver) Power Supply. Connect an optional 3.12 kΩ resistor from this pin to ground (see the Output Current Control with an External Resistor section). Analog Supply for Output Driver. These pins are normally 3.3 V but can be 1.8 V. Pin 25 powers OUT0. Pin 31 powers OUT1. Apply power to these pins even if the corresponding outputs (OUT0P/OUT0N, OUT1P/OUT1N) are not used (see the Power Supply Partitions section). Output 0. This output can be configured as LVPECL, LVDS, or single-ended CMOS. LVPECL and LVDS operation require a 3.3 V output driver power supply. CMOS operation can be either 1.8 V or 3.3 V, depending on the output driver power supply. Complementary Output 0. This output can be configured as LVPECL, LVDS, or single-ended CMOS. 1.8 V Analog (Output Divider) Power Supply. Output 1. This output can be configured as LVPECL, LVDS, or single-ended CMOS. LVPECL and LVDS operation require a 3.3 V output driver power supply. CMOS operation can be either 1.8 V or 3.3 V, depending on the output driver power supply. Complementary Output 1. This output can be configured as LVPECL, LVDS, or single-ended CMOS. 3.3 V Analog (System Clock) Power Supply. System Clock Loop Filter Voltage Regulator. Connect a 0.1 μF capacitor from this pin to ground. This pin is also the ac ground reference for the integrated external loop filter of the SYSCLK PLL multiplier (see the SYSCLK PLL Multiplier section). System Clock Multiplier Loop Filter. When using the frequency multiplier to drive the system clock, an external loop filter can be attached to this pin. 1.8 V Analog (System Clock) Power Supply. Complementary System Clock Input. Complementary signal to SYSCLKP. SYSCLKN contains internal dc biasing and should be ac-coupled with a 0.01 μF capacitor, except when using a crystal. When using a crystal, connect it across SYSCLKP and SYSCLKN. System Clock Input. SYSCLKP contains internal dc biasing and should be accoupled with a 0.01 μF capacitor, except when using a crystal. When using a crystal, connect it directly across SYSCLKP and SYSCLKN. Single-ended 1.8 V CMOS is also an option but can introduce a spur if the duty cycle is not 50%. When using SYSCLKP as a single-ended input, connect a 0.01 μF capacitor from SYSCLKN to ground. Use capacitive decoupling on these pins (see Figure 36). 1.8 V Analog (Time-to-Digital Converter) Power Supply. 3.3 V Analog (Reference Input) Power Supply. Reference A Input. This internally biased input is typically ac-coupled and, when configured as such, can accept any differential signal with a singleended swing of up to 3.3 V. If dc-coupled, input can be LVPECL, CMOS, or LVDS. Complementary Reference A Input. Complementary signal to the input provided on Pin 44. The user can configure this pin as a separate single-ended input. 1.8 V Analog (Reference Input) Power Supply. Rev. 0 | Page 16 of 104 AD9547 Pin No. 47 Input/ Output I 48 I 51 52 54, 55, 56, 57, 60, 61, 62, 63 EP I O I/O O Pin Type Differential input Mnemonic REFB Differential input REFBB Logic 3.3 V CMOS Exposed pad NC IRQ M0, M1, M2, M3, M4, M5, M6, M7 Exposed pad Description Reference B Input. This internally biased input is typically ac-coupled and, when configured as such, can accept any differential signal with a single-ended swing of up to 3.3 V. If dc-coupled, input can be LVPECL, CMOS, or LVDS. Complementary Reference B Input. Complementary signal to the input provided on Pin 47. The user can configure this pin as a separate single-ended input. No Connection. This pin should be left floating. Interrupt Request Line. Configurable I/O Pins. These pins are configured under program control. The exposed pad must be connected to ground (VSS). Rev. 0 | Page 17 of 104 AD9547 TYPICAL PERFORMANCE CHARACTERISTICS fREF = input reference clock frequency, fO = clock frequency, fSYSCLK = SYSCLK input frequency, fS =internal system clock frequency, LBW = DPLL loop bandwidth, PLL off = SYSCLK PLL bypassed, PLL on = SYSCLK PLL enabled, ICP = SYSCLK PLL charge pump current, LF = SYSCLK PLL loop filter. AVDD, AVDD3, and DVDD at nominal supply voltage, fS = 1 GHz, ICP = automatic mode, LF = internal, unless otherwise noted. –90 PHASE NOISE (dBc/Hz) –110 –120 –130 –100 –110 –120 –130 –140 –140 –150 –150 1k 10k 100k 1M 10M 100M FREQUENCY OFFSET (Hz) –160 100 08300-068 PHASE NOISE (dBc/Hz) –90 –100 –160 100 –80 –70 INTEGRATED RMS JITTER (PHASE NOISE): 5kHz TO 20MHz: 333fs (–69.8dBc) 20kHz TO 80MHz: 430fs (–67.6dBc) (EXTRAPOLATED) 100k 1M 10M 100M INTEGRATED RMS JITTER (PHASE NOISE): 5kHz TO 20MHz: 310fs (–64.4dBc) 20kHz TO 80MHz: 330fs (–63.9dBc) –80 –90 PHASE NOISE (dBc/Hz) –100 –110 –120 –130 –100 –110 –120 –130 –140 –140 –150 –150 1k 10k 100k 1M 10M 100M FREQUENCY OFFSET (Hz) 08300-056 PHASE NOISE (dBc/Hz) 10k Figure 5. Additive Phase Noise (Output Driver = LVPECL), fREF = 19.44 MHz, fO = 311.04 MHz, LBW = 1 kHz, fSYSCLK = 1 GHz, PLL Off –90 –160 100 1k FREQUENCY OFFSET (Hz) Figure 3. Additive Phase Noise (Output Driver = LVPECL), fREF = 19.44 MHz, fO = 155.52 MHz, LBW = 1 kHz, fSYSCLK = 1 GHz, PLL Off –70 INTEGRATED RMS JITTER (PHASE NOISE): 5kHz TO 20MHz: 103fs (–74.0dBc) 20kHz TO 80MHz: 160fs (–70.1dBc) –80 08300-066 –80 –70 INTEGRATED RMS JITTER (PHASE NOISE): 5kHz TO 20MHz: 173fs (–75.4dBc) 20kHz TO 80MHz: 315fs (–70.2dBc) (EXTRAPOLATED) Figure 4. Additive Phase Noise (Output Driver = LVPECL), fREF = 19.44 MHz, fO = 155.52 MHz, LBW = 1 kHz, fSYSCLK = 50 MHz (Crystal), PLL On –160 100 1k 10k 100k 1M 10M 100M FREQUENCY OFFSET (Hz) Figure 6. Additive Phase Noise (Output Driver = LVPECL), fREF = 19.44 MHz, fO = 311.04 MHz, LBW = 1 kHz, fSYSCLK = 50 MHz (Crystal), PLL On Rev. 0 | Page 18 of 104 08300-067 –70 AD9547 –70 –80 –70 INTEGRATED RMS JITTER (PHASE NOISE): 5kHz TO 20MHz: 361fs (–69.0dBc) 20kHz TO 80MHz: 441fs (–67.3dBc) (EXTRAPOLATED) –80 50MHz CRYSTAL –100 –110 –120 –130 –120 –130 –140 –150 –150 –160 100 –160 100 1k 10k 100k 1M 10M 100M ROHDE & SCHWARZ SMA100 (1GHz) 1k 10k 100k 1M 10M 100M FREQUENCY OFFSET (Hz) Figure 10. Additive Phase Noise Comparison of SYSCLK Input Options (Output Driver = LVPECL), fREF = 19.44 MHz, fO = 311.04 MHz, LBW = 1 kHz Figure 7. Additive Phase Noise (Output Driver = LVPECL), fREF = 19.44 MHz, fO = 155.52 MHz, LBW = 1 kHz, fSYSCLK = 50 MHz, PLL On –70 –70 INTEGRATED RMS JITTER (PHASE NOISE): 12kHz TO 20MHz: 245fs (–65.0dBc) 20kHz TO 80MHz: 300fs (–64.3dBc) –80 INTEGRATED RMS JITTER (PHASE NOISE): 5kHz TO 20MHz: 336fs (–69.7dBc) 20kHz TO 80MHz: 425fs (–67.6dBc) –80 –90 –100 –110 –120 –130 –100 –110 –120 –130 –140 –140 –150 –150 10k 100k 1M 10M 100M FREQUENCY OFFSET (Hz) –160 100 08300-051 1k 1k 10k 100k 1M 10M 100M FREQUENCY OFFSET (Hz) 08300-052 PHASE NOISE (dBc/Hz) –90 –160 100 ROHDE & SCHWARZ SMA100 (50MHz) –110 –140 FREQUENCY OFFSET (Hz) PHASE NOISE (dBc/Hz) –100 08300-058 PHASE NOISE (dBc/Hz) –90 08300-069 PHASE NOISE (dBc/Hz) –90 Figure 11. Additive Phase Noise (Output Driver = LVPECL), fREF = 8 kHz, fO = 155.52 MHz, LBW = 100 Hz, fSYSCLK = 50 MHz (Crystal), PLL On Figure 8. Additive Phase Noise (Output Driver = LVPECL), fREF = 19.44 MHz, fO = 155.52 MHz, LBW = 1 kHz, fSYSCLK = 50 MHz (Crystal), PLL On with 2× Frequency Multiplier, ICP = 375 μA, LF = External (350 kHz) –90 10 –100 0 –110 –10 CLOSED-LOOP GAIN (dB) –120 –130 ROHDE & SCHWARZ SMA100 (1GHz) –140 20MHz OCXO –150 –30 –40 –50 ROHDE & SCHWARZ SMA100 (50MHz) 1k 10k 100k 1M FREQUENCY OFFSET (Hz) 10M Figure 9. Phase Noise of SYSCLK Input Sources –70 10 100 1k 10k 100k FREQUENCY OFFSET (Hz) Figure 12. Jitter Transfer Bandwidth, Output Driver = LVPECL, fREF = 19.44 MHz, fO = 155.52 MHz, LBW = 100 Hz (Phase Margin = 88°), fSYSCLK = 1 GHz, PLL Off Rev. 0 | Page 19 of 104 08300-047 –170 100 –20 –60 –160 08300-053 PHASE NOISE (dBc/Hz) CLOSED-LOOP PEAKING: 0.04dB AD9547 2.0 1.0 5pF LOAD 0.8 AMPLITUDE (V) AMPLITUDE (V) LVPECL 0.6 0.4 LVDS 1.5 20pF LOAD 1.0 10pF LOAD 0.2 300 400 500 600 700 FREQUENCY (MHz) 0 50 100 200 250 50 200 FREQUENCY (MHz) Figure 13. Amplitude vs. Toggle Rate, LVPECL and LVDS Figure 16. Amplitude vs. Toggle Rate, 1.8 V CMOS 4.0 4.0 3.5 3.5 10pF LOAD 5pF LOAD 3.0 AMPLITUDE (V) AMPLITUDE (V) 150 08300-062 200 08300-063 100 08300-049 0 08300-061 0.5 0 2.5 20pF LOAD 2.0 3.0 10pF LOAD 2.5 2.0 1.5 1.5 1.0 1.0 100 200 300 400 500 FREQUENCY (MHz) 0 08300-055 0 10 20 30 40 FREQUENCY (MHz) Figure 17. Amplitude vs. Toggle Rate, 3.3 V CMOS (Weak Mode) Figure 14. Amplitude vs. Toggle Rate, 3.3 V CMOS (Strong Mode) 40 140 130 20pF LOAD 35 120 10pF LOAD POWER (mW) LVPECL 100 90 5pF LOAD 30 25 80 LVDS 70 20 60 15 50 0 100 200 300 400 FREQUENCY (MHz) 500 08300-064 POWER (mW) 110 0 50 100 150 FREQUENCY (MHz) Figure 18. Power Consumption vs. Frequency, 1.8 V CMOS Figure 15. Power Consumption vs. Frequency, LVPECL and LVDS (Single Channel) Rev. 0 | Page 20 of 104 AD9547 160 34 140 32 30 10pF LOAD 100 5pF LOAD 10pF LOAD 80 28 20pF LOAD 60 24 40 22 0 50 100 150 200 250 300 350 FREQUENCY (MHz) 20 10 08300-060 20 15 20 25 30 35 40 FREQUENCY (MHz) Figure 19. Power Consumption vs. Frequency, 3.3 V CMOS (Strong Mode) Figure 22. Power Consumption vs. Frequency, 3.3 V CMOS (Weak Mode) 1.0 0.5 0.8 0.4 0.6 0.3 DIFFERENTIAL AMPLITUDE (V) 0.4 0.2 0 –0.2 –0.4 –0.6 –0.8 0.2 0.1 0 –0.1 –0.2 –0.3 –0.4 –1.0 1 2 3 4 5 TIME (ns) –0.5 08300-050 0 0 1 2 3 4 5 TIME (ns) Figure 20. Output Waveform, LVPECL (400 MHz) 08300-048 DIFFERENTIAL AMPLITUDE (V) 5pF LOAD 26 08300-059 20pF LOAD POWER (mW) POWER (mW) 120 Figure 23. Output Waveform, LVDS (400 MHz) 3.5 3.5 10pF LOAD 5pF LOAD 3.0 3.0 20pF LOAD 2.5 AMPLITUDE (V) 2.0 1.5 1.0 2.0 20 pF LOAD 1.5 1.0 0.5 0.5 0 0 0 2 4 6 8 10 12 TIME (ns) 14 16 0 10 20 30 40 50 60 TIME (ns) Figure 24. Output Waveform, 3.3 V CMOS (20 MHz, Weak Mode) Figure 21. Output Waveform, 3.3 V CMOS (100 MHz, Strong Mode) Rev. 0 | Page 21 of 104 70 80 08300-046 –0.5 –0.5 08300-057 AMPLITUDE (V) 2.5 AD9547 2.0 10pF LOAD 1.5 1.0 0.5 0 –0.5 0 2 4 6 8 10 TIME (ns) 12 14 16 08300-065 AMPLITUDE (V) 20pF LOAD Figure 25. Output Waveform, 1.8 V CMOS (100 MHz) Rev. 0 | Page 22 of 104 AD9547 INPUT/OUTPUT TERMINATION RECOMMENDATIONS 0.1µF DOWNSTREAM DEVICE 100Ω 100Ω HIGH IMPEDANCE INPUT 0.1µF AD9547 SELF-BIASED SYSCLKx INPUT 08300-006 08300-003 0.1µF (OPTIONAL) AD9547 3.3V LVDS OUTPUT 0.1µF Figure 26. AC-Coupled LVDS or LVPECL Output Driver Figure 29. SYSCLKx Input AD9547 100Ω DOWNSTREAM DEVICE 100Ω 3.3V LVPECLCOMPATIBLE OUTPUT AD9547 SELF-BIASED CLKINx INPUT 08300-007 08300-004 0.1µF (OPTIONAL) 0.1µF Figure 27. DC-Coupled LVDS or LVPECL Output Driver Figure 30. CLKINx Input AD9547 SELF-BIASED REFERENCE INPUT 08300-005 0.1µF (OPTIONAL) 100Ω 0.1µF Figure 28. Reference Input Rev. 0 | Page 23 of 104 AD9547 GETTING STARTED POWER-ON RESET The AD9547 monitors the voltage on the power supplies at powerup. When DVDD3 is greater than 2.35 V ± 0.1 V and DVDD (Pin 1, Pin 6, Pin 8, Pin 53, Pin 59, and Pin 64) is greater than 1.4 V ± 0.05 V, the device generates a 75 ns reset pulse. The power-up reset pulse is internal and independent of the RESET pin. This internal power-up reset sequence eliminates the need for the user to provide external power supply sequencing. Within 45 ns after the leading edge of the internal reset pulse, the M0 to M7 multifunction pins function as high impedance digital inputs and continue to do so until programmed otherwise. INITIAL M0 TO M7 PIN PROGRAMMING During a device reset (either via the power-up reset pulse or the RESET pin), the multifunction pins (M0 to M7) function as high impedance inputs, but upon removal of the reset condition, level-sensitive latches capture the logic pattern present on the multifunction pins. The AD9547 requires that the user supply the desired logic state to the M0 to M7 pins by means of pull-up and/or pull-down resistors (nominally 10 kΩ to 30 kΩ). The initial state of the M0 to M7 pins following a reset is referred to as FncInit, Bits[7:0]. Bits[7:0] of FncInit map directly to the logic states of M[7:0], respectively. The three LSBs of FncInit (FncInit, Bits[2:0]) determine whether the serial port interface functions according to the SPI or the I2C protocol. Specifically, FncInit, Bits[2:0] = 000 selects the SPI interface. Any other value selects the I2C port, with the three LSBs of the I2C bus address set to the value of FncInit, Bits[2:0]. The five MSBs of FncInit (FncInit, Bits[7:3]) determine the operation of the EEPROM loader. On the falling edge of RESET, if FncInit, Bits[7:3] = 00000, then the EEPROM contents are not transferred to the control registers and the device registers assume their default values. However, if FncInit, Bits[7:3] ≠ 00000, then the EEPROM controller transfers the contents of the EEPROM to the control registers with CONDITION = FncInit, Bits[7:3] (see the EEPROM section). DEVICE REGISTER PROGRAMMING Initialize the system clock. After the system clock functionality is programmed, issue an I/O update using Register 0x0005, Bit 0 to invoke the system clock settings. 3. Calibrate the system clock (only if using SYSCLK PLL). Set the calibrate system clock bit in the cal/sync register (Address 0x0A02, Bit 0) and issue an I/O update. Then clear the calibrate system clock bit and issue another I/O update. This action allows time for the calibration to proceed while programming the remaining device registers. 4. Program the multifunction pins (optional). This step is required only if the user intends to use any of the multifunction pins for status or control. The multifunction pin parameters reside in the 0x0200 register address space. The default configuration of the multifunction pins is as undesignated high impedance input pins. 5. Program the IRQ functionality (optional). This step is required only if the user intends to use the IRQ feature. IRQ control resides in the 0x0200 register address space. It includes the following: • • IRQ pin mode control IRQ mask The IRQ mask default values prevent interrupts from being generated. The IRQ pin mode default is open-drain NMOS. 6. Program the watchdog timer (optional). This step is required only if the user intends to use the watchdog timer. Watchdog timer control resides in the 0x0200 register address space. The watchdog timer is disabled by default. Program the DAC full-scale current (optional). This step is required only if the user intends to use a full-scale current setting other than the default value. DAC full-scale current control resides in the 0x0200 register address space. 8. Program the digital phase-locked loop (DPLL). The DPLL parameters reside in the 0x0300 register address space. They include the following: Program the system clock functionality. The system clock parameters reside in the 0x100 register address space. They include the following: • • • 2. 7. The initial state of the M0 to M7 pins establishes the serial I/O port protocol (SPI or I2C). Using the appropriate serial port protocol, and assuming that an EEPROM download is not used, program the device according to the recommended sequence that follows. 1. mended that the system clock stability timer be programmed, as well. This is especially important when using the system clock PLL but also applies if using an external system clock source, especially if the external source is not expected to be completely stable when power is applied to the AD9547. System clock PLL controls System clock period System clock stability timer It is essential to program the system clock period because many of the AD9547 subsystems rely on this value. It is highly recom- • • • • • • Rev. 0 | Page 24 of 104 Free-run frequency (DDS frequency tuning word) DDS phase offset DPLL pull-in range limits DPLL closed-loop phase offset Phase slew control (for hitless reference switching) Tuning word history control (for holdover operation) AD9547 9. Program the clock distribution outputs. 12. Generate reference acquisition. The clock distribution parameters reside in the 0x0400 register address space. They include the following: • • • • • • Output power-down control Output enable (disabled by default) Output synchronization Output mode control Output divider functionality Program the reference inputs. 10. The reference input parameters reside in the 0x0500 register address space. They include the following: • • • • Reference power-down Reference logic family Reference profile assignment control Phase build-out control 11. Program the reference profiles. The reference profile parameters reside in the 0x0600 and 0x0700 register address spaces. They include the following: • • • • • • • • • Reference priority Reference period Reference period tolerance Reference validation timer Reference redetect timer Digital loop filter coefficients Reference prescaler (R divider) Feedback dividers (S, U, and V) Phase and frequency lock detector controls After the registers are programmed, issue an I/O update using Register 0x0005, Bit 0 to invoke all of the register settings that have been programmed up to this point. If the settings are programmed for manual profile assignment, the DPLL locks to the first available reference that has the highest priority. If the settings are programmed for automatic profile assignment, then write to the reference profile selection register (Address 0x0A0D) to select the state machines that require starting. Next, issue an I/O update (Address 0x0005, Bit 0) to start the selected state machines. Upon completion of the reference detection sequence, the DPLL locks to the first available reference with the highest priority. 13. Generate the output clock. If the registers are programmed for automatic clock distribution synchronization via DPLL phase or frequency lock, the synthesized output signal appears at the clock distribution outputs (assuming that the output is enabled and the DDS output signal has been routed to the CLKINx input pins). Otherwise, set and then clear the sync distribution bit (Address 0x0A02, Bit 1) or use a multifunction pin input (if programmed accordingly) to generate a clock distribution sync pulse, which causes the synthesized output signal to appear at the clock distribution outputs. Rev. 0 | Page 25 of 104 AD9547 THEORY OF OPERATION AD9547 OUT_RSET DIFFERENTIAL OR SINGLE-ENDED REFA POST DIV OUT0P OUT0N POST DIV OUT1P OUT1N REFAA REFB REFBB DIGITAL PLL CORE ÷S CLKINP TDC/PFD 2 OR 4 ÷R PROG. DIGITAL LOOP FILTER TW CLAMP AND HISTORY CLKINN DDS/DAC EXTERNAL ANALOG FILTER HOLDOVER LOGIC PHASE CONTROLLER INPUT REF MONITOR M0 TO M7 IRQ IRQ AND STATUS LOGIC CLOCK DISTRIBUTION LOW NOISE CLOCK MULTIPLIER CONTROL LOGIC AMP DIGITAL INTERFACE SYSCLKN SYSCLKP 08300-009 SYSCLK PORT Figure 31. Detailed Block Diagram OVERVIEW The AD9547 provides clocking outputs that are directly related in phase and frequency to the selected (active) reference but with jitter characteristics primarily governed by the system clock. The AD9547 supports up to four reference inputs and a wide range of reference frequencies. The core of this product is a digital phaselocked loop (DPLL). The DPLL has a programmable digital loop filter that greatly reduces jitter transferred from the active reference to the output. The AD9547 supports both manual and automatic holdover modes. While in holdover mode, the AD9547 continues to provide an output as long as the DAC sample clock is present. The holdover output frequency is a time average of the output frequency history just prior to the transition to the holdover condition. The device offers manual and automatic reference switchover capability if the active reference is degraded or fails completely. A direct digital synthesizer (DDS) and integrated DAC constitute a digitally controlled oscillator (DCO). The DCO output is a sinusoidal signal (450 MHz maximum) at a frequency that is determined by the active reference frequency and the programmed values of the reference prescaler (R) and feedback divider (S). Although not explicitly shown in Figure 31, the S divider has both an integer and fractional component, which is similar to a fractional-N synthesizer. The SYSCLKx input provides the sample clock for the DAC, which is either a directly applied high frequency source or a low frequency source coupled with the integrated PLL-based frequency multiplier. The low frequency option also allows for the use of a crystal resonator connected directly across the SYSCLKx inputs. The DAC output routes directly off chip where an external filter removes the sampling artifacts before returning the signal on chip at the CLKINx inputs. Once on chip, an integrated comparator converts the filtered sinusoidal signal to a clock signal (square wave) with very fast rise and fall times. The clock distribution section provides two output drivers. Each driver is programmable either as a single differential LVPECL/ LVDS output or as a dual single-ended CMOS output. Furthermore, a dedicated 30-bit programmable divider precedes each driver. The clock distribution section operates at up to 725 MHz. This enables use of a band-pass reconstruction filter (for example, a SAW filter) to extract a Nyquist image from the DAC output spectrum, thereby allowing output frequencies that exceed the typical 450 MHz limit at the DAC output. Rev. 0 | Page 26 of 104 AD9547 REFERENCE CLOCK INPUTS Two pairs of pins provide access to the reference clock receivers. Each pair is configurable either as a single differential receiver or as two independent single-ended receivers. To accommodate input signals with slow rising and falling edges, both the differential and single-ended input receivers employ hysteresis. Hysteresis also ensures that a disconnected or floating input does not cause the receiver to oscillate spontaneously. When configured for differential operation, the input receivers accommodate either ac- or dc-coupled input signals. The receiver is internally dc biased to handle ac-coupled operation. When configured for single-ended operation, the input receivers exhibit a pull-down load of 45 kΩ (typical). Three user-programmable threshold voltage ranges are available for each single-ended receiver. REFERENCE MONITORS The reference monitors depend on a known and accurate system clock period. Therefore, the functioning of the reference monitors is not reliable until the system clock is stable. To avoid an incorrect valid indication, the reference monitors indicate fault status until the system clock stability timer expires (see the System Clock Stability Timer section). Reference Period Monitor Each reference input has a dedicated monitor that repeatedly measures the reference period. The AD9547 uses the reference period measurements to determine the validity of the reference based on a set of user provided parameters in the profile register area of the register map (see the Profile Registers (Register 0x0600 to Register 0x07FF) section). The AD9547 also uses the reference period monitor to assign a particular reference to a profile when the user programs the device for automatic profile assignment. The monitor works by comparing the measured period of a particular reference input against the parameters stored in the profile register assigned to that same reference input. The parameters include the reference period, an inner tolerance, and an outer tolerance. A 40-bit number defines the reference period in units of femtoseconds (fs). The 40-bit range allows for a reference period entry of up to 1.1 ms (909 Hz). A 20-bit number defines the inner and outer tolerances. The value stored in the register is the reciprocal of the tolerance specification. For example, a tolerance specification of 50 ppm yields a register value of 1/(50 ppm) = 1/0.000050 = 20,000 (0x04E20). The use of two tolerance values provides hysteresis for the monitor decision logic. The inner tolerance applies to a previously faulted reference and specifies the largest period tolerance that a previously faulted reference can exhibit before it qualifies as nonfaulted. The outer tolerance applies to an already nonfaulted reference and specifies the largest period tolerance that a nonfaulted reference can exhibit before being faulted. To produce decision hysteresis, the inner tolerance must be less than the outer tolerance. That is, a faulted reference must meet tighter requirements to become nonfaulted than a nonfaulted reference must meet to become faulted. Reference Validation Timer Each reference input has a dedicated validation timer. The validation timer establishes the amount of time that a previously faulted reference must remain fault free before the AD9547 declares it to be nonfaulted. The timeout period of the validation timer is programmable via a 16-bit register (see the validation register contained within each of the eight profile registers in the register map, Address 0x0600 to Address 0x07FF). The 16-bit number stored in the validation register represents units of milliseconds, which yields a maximum timeout period of 65,535 ms. Note that a validation period of zero must be programmed to disable the validation timer. With the validation timer disabled, the user must validate a reference manually via the force validation timeout register (Address 0x0A0E). Reference Redetect Timer Each reference input has a dedicated redetect timer. The redetect timer is useful only when the device is programmed for automatic profile selection. The redetect timer establishes the amount of time that a reference must remain faulted before the AD9547 attempts to reassign it to a new profile. The timeout period of the redetect timer is programmable via a 16-bit register (see the redetect timer register contained within each of the eight profile registers in the register map, Address 0x0600 to Address 0x07FF). The 16-bit number stored in the redetect timer register represents units of milliseconds, which yields a maximum timeout period of 65,535 ms. Note that a timeout period of 0 must be programmed to disable the redetect timer. Rev. 0 | Page 27 of 104 AD9547 REGISTER CONTROL BITS REFERENCE VALIDATION LOGIC (4 COPIES, 1 PER REFERENCE INPUT) D Q VALID FORCE VALIDATION TIMEOUT VALIDATION TIMER REF MONITOR BYPASS REF MONITOR OVERRIDE R 1 EN R TIMEOUT FAULTED REFERENCE MONITOR 08300-010 0 REF FAULT Figure 32. Reference Validation Override Reference Validation Override Control Register 0x0A0E to Register 0x0A10 provide the user with the ability to override the reference validation logic, enabling a certain level of troubleshooting capability. Each of the four input references has a dedicated block of validation logic, as shown in Figure 32. The state of the valid signal at the output defines a particular reference as valid (1) or not valid (0), which includes the validation period (if activated) as prescribed by the validation timer. The override controls are the three control bits on the left side of the diagram. The main feature to note is that when faulted = 1, the output latch is reset, which forces valid = 0 (indicating an invalid reference), regardless of the state of any other signal. Under the default condition (that is, all three control bits are set to 0), the reference monitor is the primary source of the validation process. This is because, under the default condition, the ref fault signal from the reference monitor is identical to the faulted signal. The function of the faulted signal is fourfold. • • • • When faulted = 1, valid = 0, regardless of the state of any other control signal. Therefore, faulted = 1 indicates an invalid reference. When the faulted signal transitions from 0 to 1 (that is, from not faulted to faulted), the validation timer is momentarily reset, which means that, when it is enabled, it must exhaust its full counting sequence before it expires. When faulted = 0 (that is, the reference is not faulted), the validation timer is allowed to perform its timing sequence. When faulted = 1 (that is, the reference is faulted), the validation timer is reset and halted. The faulted signal passes through an inverter, which converts it to a not faulted signal that appears at the input of the valid latch. This allows the valid latch to capture the state of the not faulted signal when the validation timer expires. The reference monitor bypass control bit (Address 0x0A10) enables bypassing of the reference fault signal generated by the reference monitor. When the reference monitor bypass bit = 1, the state of the faulted signal is dictated by the reference monitor override control bit. This is useful when the user relies on an external reference monitor rather than the internal monitor resident in the device. The user programs the reference monitor override bit based on the status of the external monitor. On the other hand, when the reference monitor bypass bit = 0, the reference monitor override control bit (Address 0x0A0F) allows the user to manually test the operation of both the valid latch and the validation timer. In this case, the user relies on the signal generated by the internal reference monitor (reference fault) but uses the reference monitor override bit to emulate a faulted reference. That is, when the reference monitor override bit = 1, faulted = 1, but when the reference monitor override bit = 0, faulted = reference fault. In addition, the user can emulate a timeout of the validation timer via the force validation timeout control register at Address 0x0A0E. Writing a Logic 1 to this autoclearing bit triggers the valid latch, which is identically equivalent to a timeout of the validation timer. REFERENCE PROFILES The AD9547 has eight independent profile registers. A profile register contains 50 bytes that establish a particular set of device parameters. Each of the four input references can be assigned to any one of the eight profiles (that is, more than one reference can be assigned to the same profile). The profiles allow the user to prescribe the specific device functionality that should take effect when one of the input references assigned to a profile becomes the active reference. Each profile register has the same format and stores the following device parameters: • • • • • • • • • • • • • • • Rev. 0 | Page 28 of 104 Reference priority Reference period value (in femtoseconds (fs)) Inner tolerance value (1/tolerance) Outer tolerance value (1/tolerance) Validation timer value (milliseconds (ms)) Redetect timer value (milliseconds (ms)) Digital loop filter coefficients Reference prescaler setting (R divider) Feedback divider settings (S, U, and V) DPLL phase lock detector threshold level DPLL phase lock detector fill rate DPLL phase lock detector drain rate DPLL frequency lock detector threshold level DPLL frequency lock detector fill rate DPLL frequency lock detector drain rate AD9547 Reference-to-Profile Assignment Control The user can manually assign a reference to a profile or let the device make the assignment automatically. The manual reference profile selection register (Address 0x0503 and Address 0x0504) is used to program whether a reference-to-profile assignment is manual or automatic. The manual reference profile selection register is a 2-byte register partitioned into four half bytes (or nibbles). The four nibbles form a one-to-one correspondence with the four reference inputs: one nibble for REF A, the next for REF AA, and so on. For a reference configured as a differential input, however, the device ignores the nibble associated with the two-letter input. For example, if the B reference is differential, only the REF B nibble matters (the device ignores the REF BB nibble). The MSB of each nibble is the manual profile bit, whereas the three LSBs of each nibble identify one of the eight profiles (0 to 7). A Logic 1 for the manual profile bit assigns the associated reference to the profile identified by the three LSBs of the nibble. A Logic 0 for the manual profile bit configures the associated reference for automatic reference-to-profile assignment (the three LSBs are ignored in this case). Note that references configured for automatic reference-to-profile assignment require activation (see the Reference-to-Profile Assignment State Machine section). Reference-to-Profile Assignment State Machine The functional flexibility of the AD9547 resides in the way that it assigns a particular input reference to one of the eight reference profiles. The reference-to-profile assignment state machine effectively builds a reference-to-profile table that maps the index of each input reference to a profile (see Table 22). Table 22. Reference-to-Profile Table Reference Input A Reference Index 0 Profile Profile # (or null) AA 1 Profile # (or null) B 2 Profile # (or null) BB 3 Profile # (or null) Each entry in the profile column consists of a profile number (0 to 7) or a null value. A null value appears when a referenceto-profile assignment does not exist for a particular reference input (following a reset, for example). The information in Table 22 appears in the register map (Register 0x0D0C to Register 0x0D0F) so that the user has access to the reference-to-profile assignments on a real-time basis. Register 0x0D0C contains the information for REF A, Register 0x0D0D contains the information for REF AA, Register 0x0D0E for REF B, and Register 0x0D0F for REF BB. Bit 7 of each register is the null indicator for that particular reference. If Bit 7 = 0, the pro-file assignment for that particular reference is null. If Bit 7 = 1, that particular reference is assigned to the profile (0 to 7) identified by Bits[6:4]. Note that Bits[6:4] are meaningless unless Bit 7 = 1. Following a reset, the reference-to-profile assignment state machine is inactive to avoid improperly assigning a reference to a profile before the system clock stabilizes. The state machine relies on accurate information from the reference monitors, which, in turn, rely on a stable system clock. Because the reference-to-profile assignment state machine is inactive at power-up, the user must initiate it manually by writing to the reference profile selection register (Address 0x0A0D). The state machine activates immediately, unless the system clock is not stabilized. In that case, activation occurs upon expiration of the system clock stability timer. Note that initialization of the state machine is on a perreference basis. That is, each reference input is associated with an independent initialization control bit. When initialized for processing a reference, the state machine continuously monitors that reference until the occurrence of a device reset. This is true even when the user programs a reference for manual profile selection, in which case the state machine associated with that particular reference operates with its activity masked. The masked background activity allows for seamless operation if the user subsequently reprograms the reference for automatic profile selection. Reference-to-Profile Assignment When a reference is programmed for manual profile assignment (see Register 0x0503 to Register 0x0504), the reference-to-profile assignment state machine puts the programmed manual profile number into the profile column of the reference-to-profile table (see Table 22) in the row associated with the appropriate reference. However, when the user programs a reference for automatic profile assignment, the state machine must determine which profile to assign to the reference, as explained in the following paragraphs. If a null entry appears in the reference-to-profile table for a particular input reference, the validation logic for that reference enters a period estimation mode. Note that a null entry is the default state following a reset, but it also occurs when a reference redetect timer expires. The period estimation mode enables the validation logic to make a blind estimate of the period of the input reference with a tolerance of 0.1%. The validation logic remains in the period estimation mode until it successfully estimates the reference period. Upon a successful reference period measurement by the validation logic, the state machine compares the measured period to the nominal reference period programmed into each of the eight profiles. The state machine assigns the reference to the profile with the closest match to the measured period. If more than one profile exactly matches the reference period, the state machine chooses the profile with the lowest numeric index. For example, if the reference period in both Profile 3 and Profile 5 matches the measured period, Profile 3 is given the assignment. To safeguard against making a poor reference-to-profile assignment, the state machine ensures that the measured reference period is within 6.25% of the nominal reference period that appears in the closest match profile. Otherwise, the state machine does not make a profile assignment and leaves the null entry in the reference-to-profile table. Rev. 0 | Page 29 of 104 AD9547 As long as there are input references programmed for automatic profile assignment, and for which the profile assignment is null, the state machine continues to cycle through those references searching for a profile match. Furthermore, unless an input reference is assigned to a profile, it is considered invalid and excluded as a candidate for a reference switchover. REFERENCE SWITCHOVER An attractive feature of the AD9547 is its versatile reference switchover capability. The flexibility of the reference switchover functionality resides in a sophisticated prioritization algorithm coupled with register-based controls. This scheme provides the user with maximum control over the state machine that handles reference switchover. The main reference switchover control resides in the loop mode register (Address 0x0A01). The user selection mode bits (Bits[4:3]) allow the user to select one of the reference switchover state machine’s four operating modes, as follows: • • • • Automatic mode (Address 0x0A01, Bits[4:3] = 00) Fallback mode (Address 0x0A01, Bits[4:3] = 01) Holdover mode (Address 0x0A01, Bits[4:3] = 10) Manual mode (Address 0x0A01, Bits[4:3] = 11) In automatic mode, a fully automatic, priority-based algorithm selects the active reference. When programmed for automatic mode, the device ignores the user reference selection bits (Register 0x0A01, Bits[1:0]). However, when programmed for any of the other three modes, the device uses the user reference bits. They specify a particular input reference (00 = REF A, 01 = REF AA , 10 = REF B, 11 = REF BB). In fallback mode, the user reference is the active reference when it is valid. Otherwise, the device switches to a new reference using the automatic priority-based algorithm. In holdover mode, the user reference is the active reference when it is valid. Otherwise, the device switches to holdover mode. In manual mode, the user reference is the active reference whether it is valid or not. Note that, when using this mode, the user must program the reference-to-profile assignment (see Register 0x0503 and Register 0x0504) as manual for the particular reference that is declared as the user reference. The reason is that if the user reference fails and its redetect timer expires, its profile assignment (shown in Table 22) becomes null. This means that the active reference (user reference) does not have an assigned profile, which places the AD9547 into an undefined state. The user also has the option to force the device directly into holdover or free-run operation via the user holdover and user free-run bits (Register 0x0A01, Bits[6:5]). In free-run mode, the free-running frequency tuning word register (Address 0x0300 to Address 0x0305) defines the DDS output frequency. Automatic Priority-Based Reference Switchover The AD9547 has a two-tiered, automatic, priority-based algorithm that is in effect for both automatic and fallback reference switchover. The algorithm relies on the fact that each reference profile contains both a selection priority and a promoted priority. The selection and promoted priority values range from 0 (highest priority) to 7 (lowest priority). The selection priority determines the order in which references are chosen as the active reference. The promoted priority is a separate priority value given to a reference only after it becomes the active reference. An automatic reference switchover occurs on failure of the active reference or when a previously failed reference becomes valid and its selection priority is higher than the promoted priority of the currently active reference (assuming that the automatic or fallback reference switchover is in effect). When performing an automatic reference switchover, the AD9547 chooses a reference based on the priority settings within the profiles. That is, the device switches to the reference with the highest selection priority (lowest numeric priority value). It does so by using the reference-to-profile table (see Table 22) to determine the reference associated with the profile exhibiting the highest priority. If multiple references share the same profile, the device chooses the reference having the lowest index value. For example, if the A, B, and BB references (Index 0, Index 2, and Index 3, respectively) share the same profile, a switchover to Reference A occurs because Reference A has the lowest index value. Note, however, that only valid references are included in switchover of the selection process. The switchover control logic ignores any reference with a status indication of invalid. The promoted priority parameter allows the user to assign a higher priority to a reference after it becomes the active reference. For example, suppose that two references have a selection priority of 3 and a promoted priority of 1, and the remaining references have a selection priority of 2 and a promoted priority of 2. Now, assume that one of the Priority 3 references becomes active because all of the Priority 2 references have failed. Sometime later, however, a Priority 2 reference becomes valid. The switchover logic normally attempts to automatically switch over to the Priority 2 reference because it has higher priority than the presently active Priority 3 reference. However, because the Priority 3 reference is active, its promoted priority of 1 is in effect. This is a higher priority than the newly validated reference’s priority of 2, so the switchover does not occur. This mechanism enables the user to give references preferential treatment while they are selected as the active reference. An example of promoted vs. nonpromoted priority switching appears in state diagram form in Figure 33. Figure 34 shows a block diagram of the interrelationship between the reference inputs, monitors, validation logic, profile selection, and priority selection functionality. In holdover mode, the DDS output frequency depends on the holdover control settings (see the Holdover section). Rev. 0 | Page 30 of 104 AD9547 A ACTIVE A VALID PRIORITY TABLE INPUT PRIORITY PROMOTED A 0 0 AA 1 0 B 2 1 BB 3 2 A FAULTED AA ACTIVE AA VALID COMMON WITHOUT PROMOTION WITH PROMOTION A VALID AA FAULTED B ACTIVE AA VALID 08300-011 ALL VALID INITIAL STATE Figure 33. Example of Priority Promotion PROFILE SELECTION VALIDATION LOGIC PRIORITY SELECTION LOOP CONTROLLER … … MONITORS ÷R TDC 08300-012 … REF A/REF AA REF B/REF BB Figure 34. Reference Clock Block Diagram Phase Build-Out Reference Switching Phase build-out reference switching is the term given to a reference switchover that completely masks any phase difference between the previous reference and the new reference. That is, there is virtually no phase change that can be detected at the output when a phase build-out switchover occurs. The AD9547 handles phase build-out switching based on whether the new reference is a phase master. A phase master is any reference with a selection priority value that is less than the phase master threshold priority value (that is, higher priority). The phase master threshold priority value resides in the phase build-out switching register (Address 0x0507), and the selection priority resides in the profile registers (Address 0x0600 to Address 0x07FF). By default, the phase master threshold priority is 0; therefore, no references can be phase masters until the user changes the phase master threshold priority. When the AD9547 switches from one reference to another, it compares the selection priority value that is stored in the profile that is assigned to the new reference with the phase master threshold priority. The AD9547 performs a phase build-out switchover only if the new reference is not a phase master. Hitless Reference Switching (Phase Slew Control) Hitless reference switching is the term given to a reference switchover that limits the rate of change of the phase of the output clock while the PLL is in the process of acquiring phase lock. This prevents the output frequency offset from becoming excessive. The all-digital nature of the DPLL core (see the Digital PhaseLocked Loop (DPLL) Core section) gives the user numerical control of the rate at which phase changes occur at the DPLL output. When enabled, a phase slew controller monitors the phase difference between the feedback and reference inputs to the DPLL. The phase slew controller can place a user-specified limit on the rate of change of phase, thus providing a mechanism for hitless reference switching. The user sets a limit on the rate of change of phase by storing the appropriate value in the 16-bit phase slew rate limit register (Address 0x0316 and Address 0x0317). The 16-bit word, which represents units of ns/sec, puts an upper bound on the rate of change of the phase at the output of the DPLL during a reference switchover. A phase slew rate value of 0 (default) disables the phase slew controller. The accuracy of the phase slew controller depends on both the phase slew limit value and the system clock frequency. Generally, an increase in the phase slew rate limit value or a decrease in the system clock frequency tends to reduce the error. Therefore, the accuracy is best for the largest phase slew rate limit value and the lowest system clock frequency. For example, assuming the use of a 1 GHz system clock, a phase slew rate limit value of 315 ns/sec (or more) ensures an error of <10%, whereas a phase slew rate limit value above ~3100 ns/sec ensures an error of <1%. On the other hand, assuming the use of a 500 MHz system clock, the same phase slew rate limit values ensure an error of <5% or 0.5%, respectively. Rev. 0 | Page 31 of 104 AD9547 DPLL Overview A diagram of the digital PLL core of the AD9547 appears in Figure 35. The phase/frequency detector, feedback path, lock detectors, phase offset, and phase slew rate limiting that make up this second-generation DPLL are all digital implementations. LOCK DETECT PHASE SLEW LIMIT f DDS = DPPL CORE CLOSED-LOOP PHASE OFFSET REF A R+1 REF BB fTDC TDC AND PFD DIGITAL LOOP FILTER DDS/ fDDS DAC 2 S + 1 + U/V DACOUT 08300-013 fREF Figure 35. Digital PLL Core The start of the DPLL signal chain is the reference signal, fREF, which is the frequency of the reference input. A reference prescaler reduces the frequency of this signal by an integer factor, R + 1, where R is the 30-bit value stored in the profile register and 0 ≤ R ≤ 1,073,741,823. Therefore, the frequency at the output of the R divider (or the input to the time-to-digital converter (TDC)) is f TDC = f REF R +1 The TDC samples the output of the R divider. The TDC/PFD produces a time series of digital words and delivers them to the digital loop filter. The digital loop filter offers the following advantages: • • • • The DPLL includes a feedback divider that causes the DDS to operate at an integer-plus-fractional multiple (S + 1 + U/V) of fTDC. S is the 20-bit value stored in the profile register and has a range of 7 ≤ S ≤ 1,048,576. U and V are the 10-bit numerator and denominator values of the optional fractional divide component, also stored in the profile register. Together they establish the nominal DDS frequency (fDDS), given by Normally, fractional-N designs exhibit distinctive phase noise and spurious artifacts resulting from the modulation of the integer divider based on the fractional value. This is not the case for the AD9547 because it uses a purely digital means to determine phase errors. Because the phase errors incurred by modulating the feedback divider are deterministic, it is possible to compensate for them digitally. The result is a fractional-N PLL with no discernible modulation artifacts. Time-to-Digital Converter (TDC)/Phase Frequency Detector (PFD) The TDC is a highly integrated functional block that incorporates both analog and digital circuitry. There are two pins associated with the TDC that the user must connect to external components. Figure 36 shows the recommended component values and their connections. For best performance, place components as close as possible to the device pins. Components with low effective series resistance (ESR) and low parasitic inductance yield the best results. AD9547 Determination of the filter response by numeric coefficients rather than by discrete component values Absence of analog components (R/L/C), which eliminates tolerance variations due to aging Absence of thermal noise associated with analog components Absence of control node leakage current associated with analog components (a source of reference feed-through spurs in the output spectrum of a traditional analog PLL) The digital loop filter produces a time series of digital words at its output and delivers them to the frequency tuning input of a direct digital synthesizer (DDS), with the DDS replacing the function of the VCO in an analog PLL. The digital words from the loop filter tend to steer the DDS frequency toward frequency and phase lock with the input signal (fTDC). The DDS provides an analog output signal via an integrated DAC, effectively mimicking the operation of an analog VCO. f REF ⎛ U⎞ ⎜S +1+ ⎟ R +1 ⎝ V⎠ 41 40 TDC_VRB TDC_VRT 0.1µF 10µF 0.1µF 0.1µF 08300-014 DIGITAL PHASE-LOCKED LOOP (DPLL) CORE Figure 36. TDC Pin Connections The PFD is an all-digital block. It compares the digital output from the TDC (which relates to the active reference edge) with the digital word from the feedback block (which relates to the rollover edge of the DDS accumulator after division by the feedback divider). The PFD uses a digital code pump and digital integrator (rather than a conventional charge pump and capacitor) to generate the error signal that steers the DDS frequency toward phase lock. Rev. 0 | Page 32 of 104 AD9547 In addition, the user can adjust the closed-loop phase offset (positive or negative) in incremental fashion. To do so, program the desired step size in the incremental phase lock offset step size bits (Address 0x0314 and Address 0x0315). This is an unsigned number that represents units of picoseconds (ps). The programmed step size is added to the current closed-loop phase offset each time the user writes a Logic 1 to the increment phase offset bit (Register 0x0A0C, Bit 0). Conversely, the programmed step size is subtracted from the current closed-loop phase offset each time the user writes a Logic 1 to the decrement phase offset bit (Register 0x0A0C, Bit 1). The serial I/O port control logic clears both of these bits automatically. The user can remove the incrementally accumulated phase by writing a Logic 1 to the reset incremental phase offset bit (Register 0x0A0C, Bit 2), which is also cleared automatically. Alternatively, rather than using the serial I/O port, the multifunction pins can be set up to perform the increment, decrement, and clear functions. Note that the incremental phase offset is completely independent of the offset programmed into the fixed phase lock offset register. However, if the phase slew limiter is active (see the Hitless Reference Switching (Phase Slew Control) section), any instantaneous change in closed-loop phase offset (fixed or incremental) is subject to possible slew limitation by the action of the phase slew limiter. Programmable Digital Loop Filter The AD9547 loop filter is a third-order digital IIR filter that is analogous to the third-order analog loop shown in Figure 37. C1 R2 C2 C3 08300-015 R3 Thus, the β, γ, and δ coefficients always represent values less than unity. The α coefficient, however, has two additional exponential components, but the hardware interprets these as a positive exponent of two (that is, 2x). This allows the α coefficient to take on values that are greater than unity. To provide sufficient dynamic range, the positive exponent appears as two separate terms. α FRACTIONAL (16-BIT) α0 1/2x (6-BIT) α1 2x (3-BIT) α2 FRACTIONAL (17-BIT) α3 1/2x 2x (4-BIT) β (6-BIT) 51 β1 σ 0 1 FRACTIONAL (17-BIT) σ0 FRACTIONAL (15-BIT) 1/2x σ1 1/2x (5-BIT) (6-BIT) 48 LOOP FILTER (THIRD-ORDER IIR) IN OUT Figure 38. Third-Order Digital IIR Loop Filter DPLL Phase Lock Detector The DPLL contains an all-digital phase lock detector. The user controls the threshold sensitivity and hysteresis of the phase lock detector via the profile registers. The phase lock detector behaves in a manner that is analogous to water in a tub (see Figure 39). The total capacity of the tub is 4096 units with −2048 denoting empty, 0 denoting the 50% point, and +2048 denoting full. The tub also has a safeguard to prevent overflow. Furthermore, the tub has a low water mark at −1024 and a high water mark at +1024. To change the water level, the user adds water with a fill bucket or removes water with a drain bucket. The user specifies the size of the fill and drain buckets via the 8-bit fill rate and drain rate values in the profile registers. The phase lock detector uses the water level in the tub to determine the lock and unlock conditions. When the water level is below the low water mark (−1024), the detector indicates an unlock con-dition. Conversely, when the water level is above the high water mark (+1024), the detector indicates a lock condition. When the water level is between the marks, the detector holds its last condition. This concept appears graphically in Figure 39, with an overlay of an example of the instantaneous water level (vertical) vs. time (horizontal) and the resulting lock/unlock states. PREVIOUS STATE Figure 37. Third-Order Analog Loop Filter The filter requires four coefficients, as shown in Figure 38. The AD9547 evaluation board software automatically generates the required loop filter coefficient values based on user design criteria. The Calculating the Digital Filter Coefficients section contains the design equations for calculating the loop filter coefficients manually. β0 08300-016 The all-digital nature of the TDC/PFD provides for numerical control of the phase offset between the reference and feedback edges. This allows the user to adjust the relative timing of the distribution output edges relative to the reference input edges by programming the fixed phase lock offset bits (Address 0x030F to Address 0x0313). The 40-bit word is a signed (twos complement) number that represents units of picoseconds (ps). LOCKED UNLOCKED 2048 LOCK LEVEL 1024 Each coefficient has a fractional component representing a value from 0 up to, but not including, unity. Each also has an exponential component representing a power of 2 with a negative exponent. That is, the user enters a positive number (x) that the hardware interprets as a negative exponent of two (2−x). Rev. 0 | Page 33 of 104 0 –1024 FILL RATE DRAIN RATE UNLOCK LEVEL –2048 Figure 39. Phase Lock Detector Diagram 08300-017 Closed-Loop Phase Offset AD9547 During any given PFD phase error sample, the detector either adds water with the fill bucket or removes water with the drain bucket (one or the other but not both). The decision on whether to add or remove water depends on the threshold level specified by the user. The phase lock threshold value is a 16-bit number stored in the profile registers and carries units of picoseconds (ps). Thus, the phase lock threshold extends from 0 ns to ±65.535 ns and represents the magnitude of the phase error at the output of the PFD. It represents the magnitude of the difference in period between the reference and feedback signals at the input to the DPLL. For example, if the reference signal is 1.25 MHz and the feedback signal is 1.38 MHz, the period difference is approximately 75.36 ns (|1/1,250,000 − 1/1,380,000| ≈ 75.36 ns). DIRECT DIGITAL SYNTHESIZER (DDS) DDS Overview One of the primary building blocks of the digital PLL is a direct digital synthesizer (DDS). The DDS behaves like a sinusoidal signal generator. The frequency of the sinusoid generated by the DDS is determined by a frequency tuning word (FTW), which is a digital (that is, numeric) value. Unlike an analog sinusoidal generator, a DDS uses digital building blocks and operates as a sampled system. Thus, it requires a sampling clock (fS) that serves as the fundamental timing source of the DDS. The accumulator behaves as a modulo-248 counter with a programmable step size (FTW). A block diagram of the DDS appears in Figure 40. The phase lock detector compares each phase error sample at the output of the PFD to the programmed phase threshold value. If the absolute value of the phase error sample is less than or equal to the programmed phase threshold value, the detector control logic dumps one fill bucket into the tub. Otherwise, it removes one drain bucket from the tub. Note that it is not the polarity of the phase error sample but, rather, its magnitude relative to the phase threshold value that determines whether to fill or drain. If more filling is taking place than draining, the water level in the tub eventually rises above the high water mark (+1024), which causes the phase lock detector to indicate lock. If more draining is taking place than filling, the water level in the tub eventually falls below the low water mark (−1024), which causes the phase lock detector to indicate unlock. The ability to specify the threshold level, fill rate, and drain rate enables the user to tailor the operation of the phase lock detector to the statistics of the timing jitter associated with the input reference signal. The input to the DDS is the 48-bit FTW. The FTW serves as a step size value. On each cycle of fS, the accumulator adds the value of the FTW to the running total at its output. For example, given that FTW = 5, the accumulator counts by fives, incrementing on each fS cycle. Over time, the accumulator reaches the upper end of its capacity (248 in this case), at which point, it rolls over but retains the excess. The average rate at which the accumulator rolls over establishes the frequency of the output sinusoid. The average rollover rate of the accumulator establishes the output frequency (fDDS) of the DDS and is given by Note that when the AD9547 enters the free-run or holdover mode, the DPLL phase lock detector indicates unlocked. Also, when the AD9547 performs a reference switchover, the state of the lock detector prior to the switch is preserved during the transition period. ⎛ FTW ⎞ f DDS = ⎜ 48 ⎟ f S ⎝ 2 ⎠ DPLL Frequency Lock Detector Solving this equation for FTW yields The operation of the frequency lock detector is identical to that of the phase lock detector. The only difference is that the fill or drain decision is based on the period deviation between the reference and feedback signals of the DPLL instead of the phase error at the output of the PFD. ⎡ ⎛f FTW = round ⎢2 48 ⎜ DDS ⎢⎣ ⎜⎝ f S For example, given that fS = 1 GHz and fDDS = 155.52 MHz, then FTW = 43,774,988,378,041 (0x27D028A1DFB9). The frequency lock detector uses a 24-bit frequency threshold register specified in units of picoseconds (ps). Thus, the frequency threshold value extends from 0 μs to ±16.777215 μs. 48-BIT ACCUMULATOR 48 48 Note that the minimum DAC output frequency is 62.5 MHz; therefore, normal operation requires an FTW that yields an output frequency in excess of this lower bound. PHASE OFFSET 16 19 48 D Q 19 ANGLE TO AMPLITUDE CONVERSION 14 DAC+ DAC (14-BIT) DAC– fS Figure 40. DDS Block Diagram Rev. 0 | Page 34 of 104 08300-018 FREQUENCY TUNING WORD (FTW) ⎞⎤ ⎟⎥ ⎟⎥ ⎠⎦ AD9547 The relative phase of the sinusoid generated by the DDS is numerically controlled by adding a phase offset word to the output of the DDS accumulator. This is accomplished via the open loop phase offset register (Address 0x030D to Address 0x030E), which is a programmable 16-bit value (Δphase). The resulting phase offset, ΔΦ (radians), is given by ⎛ Δphase ⎞ ΔΦ = 2π⎜ ⎟ ⎝ 216 ⎠ The output of the digital core of the DDS is a time series of numbers representing a sinusoidal waveform. The DAC translates the numeric values to an analog signal. The DAC output signal appears at two pins that constitute a balanced current source architecture (see Figure 41). 16 CURRENT MIRROR CURRENT SWITCH ARRAY DACOUTP 13 IFS 1– UPPER TUNING WORD The user controls the frequency clamp boundaries via the pullin range limit registers (Address 0x0307 to Address 0x030C). These registers allow the user to fix the DDS output frequency between an upper and lower bound with a granularity of 24 bits. Note that these upper and lower bounds apply regardless of the frequency tuning word that appears at the input to the DDS. The register value relates to the absolute upper or lower frequency bound (fCLAMP) as CODE 214 – 1 14 DACOUTN SWITCH CONTROL 50Ω LOWER TUNING WORD Frequency Clamp IFS CODE IFS 214 – 1 FROM DIGITAL LOOP FILTER TO DDS However, regardless of the operating mode, the DDS output frequency is ultimately subject to the boundary conditions imposed by the frequency clamp logic as explained in the Frequency Clamp section. AVDD3 10 TUNING WORD CLAMP TUNING WORD ROUTING CONTROL TUNING WORD UPDATE When the DPLL is in free-run mode, the DDS tuning word is the value stored in the free-running frequency tuning word register (Address 0x0300 to Address 0x0305). When the DPLL is operating normally (closed loop), the DDS tuning word comes from the output of the digital loop filter, which changes dynamically to maintain phase lock with the input reference signal (assuming that the device has not performed an automatic switch to holdover mode). When the DPLL is in holdover mode, the DDS tuning word depends on a historical record of past tuning words during the time that the DPLL operated in closed-loop mode. DAC Output ISCALE FREE-RUN TUNING WORD Figure 42. Tuning Word Processing Phase offset and relative time offset are directly related. The time offset is (Δphase/216)/fDDS (seconds), where fDDS is the output frequency of the DDS (Hz). 14 50Ω fCLAMP = fS × (N/224) GND GND 08300-019 CODE where N is the value stored in the upper- or lower-limit register, and fS is the system sample rate. Figure 41. DAC Output Pins The value of IFS is programmable via the 10-bit DAC full-scale current word in the DAC current register (Address 0x0213 and Address 0x0214). The value of the 10-bit word (ISCALE) sets IFS according to the following formula: 3 IFS = 120 μA × (72 + ⎛⎜ ⎞⎟ × ISCALE) ⎝ 16 ⎠ The frequency tuning words that dictate the output frequency of the DDS come from one of three sources (see Figure 42). The free-running frequency tuning word register The output of the digital loop filter The output of the tuning word history processor Even though the frequency clamp limits put a bound on the DDS output frequency, the DPLL is still free to steer the DDS frequency within the clamp limits. The default register values set the clamp range from 0 Hz (dc) to fS, effectively eliminating the frequency clamp functionality until the user alters the register values. Frequency Tuning Word History TUNING WORD PROCESSING • • • TUNING WORD HISTORY PROCESSOR TUNING WORD HISTORY 08300-070 DDS Phase Offset The AD9547 has the ability to track the history of the tuning word samples generated by the DPLL digital loop filter output. It does so by periodically computing the average tuning word value over a user-specified interval. The user programs the interval via the 24-bit history accumulation timer register (Address 0x0318 to Address 0x031A). This 24-bit value represents a time interval (TAVG) in units of milliseconds (ms) that extends from 1 ms to a maximum of 4:39:37.215 (hr:min:sec). Rev. 0 | Page 35 of 104 AD9547 Note that history accumulation timer = 0 should not be programmed because it may cause improper device operation. The control logic performs a calculation of the average tuning word during the TAVG interval and stores the result in the holdover history register (Address 0x0D14 to Address 0x0D19). Computation of the average for each TAVG interval is independent of the previous interval (that is, the average is a memoryless average as opposed to a true moving average). In addition, at the end of each TAVG interval, the device generates an internal strobe pulse. The strobe pulse sets the history updated bit in the IRQ monitor register (assuming that the bit is enabled via the IRQ mask register). Furthermore, the strobe pulse is available as an output signal via the multifunction pins (see the Multifunction Pins (M0 to M7) section). History accumulation begins when the device switches to a new reference. By default, the device clears any previous history when it switches to a new reference. Furthermore, the user can clear the tuning word history under software control using Bit 2 of Register 0x0A03 or under hardware control via the multifunction pins (see the Multifunction Pins (M0 to M7) section). However, the user has the option of programming the device to retain (rather than clear) the old history by setting the persistent history bit (Register 0x031B, Bit 3). When the tuning word history is nonexistent (that is, after a power-up, reset, or switchover to a new reference with the persistent history bit cleared), the device waits for the history accumulation timer (TAVG) to expire before storing the first history value in the holdover history register. In cases where TAVG is quite large (4½ hours, for example), a problem arises in that the first averaged result does not become available until the full TAVG interval passes. Thus, it is possible that as much as 4½ hours can elapse before the first averaged result is available. If the device must switch to holdover during this time, a tuning word history is not available. To alleviate this problem, the user can access the incremental average bits in the history mode register (Register 0x031B, Bits[2:0]). If the history has been cleared, this 3-bit value, K (0 ≤ K ≤ 7), specifies the number of intermediate averages to take during the first, and only the first, TAVG interval. When K = 0, no intermediate averages are calculated; therefore, the first average occurs after Interval TAVG (the default operating mode). However, if K = 4, for example, four intermediate averages are taken during the first TAVG interval. These average computations occur at TAVG/16, TAVG/8, TAVG/4, TAVG/2, and TAVG (note that the denominator exhibits a sequence of powers of 2 beginning with TAVG/2K). The calculation of intermediate averages occurs only during the first TAVG interval. All subsequent average computations occur at evenly spaced intervals of TAVG. LOOP CONTROL STATE MACHINE The loop control state machine is responsible for monitoring, initiating, and sequencing changes to the DPLL loop. Generally, it automatically controls the transition between input references and the entry and exit of holdover mode. In controlling loop state changes, the state machine also arbitrates the application of new loop filter coefficients, divider settings, and phase detector offsets based on the profile settings. The user can manually force the device into holdover or free-run mode via the loop mode register (Address 0x0A01), as well as force the selection of a specific input reference. Switchover Switchover occurs when the loop controller switches directly from one input reference to another. Functionally, the AD9547 handles a reference switchover by briefly entering holdover mode then immediately recovering. During the switchover event, however, the AD9547 preserves the status of the lock detectors in order to avoid phantom unlock indications. Holdover The holdover state of the DPLL is an open-loop operating mode; that is, the device no longer operates as a closed-loop system. Instead, the output frequency remains constant and is dependent on the device programming and availability of the tuning word history as explained in the following paragraphs. If a tuning word history exists (see the Frequency Tuning Word History section), the holdover frequency is the average frequency just prior to entering the holdover state. If there is no tuning word history, the holdover frequency depends on the state of the single sample fallback bit in the history mode register (Register 0x031B, Bit 4). If the single sample fallback bit is Logic 0, the holdover frequency is the frequency defined in the free-running frequency tuning word register (Address 0x0300 to Address 0x0305). If the single sample fallback bit is Logic 1, the holdover frequency is the last instantaneous frequency output by the DDS just prior to the device entering holdover mode (note that this is not the average frequency prior to holdover). The initial holdover frequency accuracy depends on the loop bandwidth of the DPLL and the time elapsed to compute a tuning word history. The longer the historical average, the more accurate the initial holdover frequency (assuming a drift-free system clock). Furthermore, the stability of the system clock establishes the stability and long-term accuracy of the holdover output frequency. Another consideration is the 48-bit frequency tuning resolution of the DDS and its relationship to fractional frequency error, ΔfO/fO. f Δf O = 49 S fO 2 fO In this equation, fS is the sample rate of the output DAC and fO is the DDS output frequency. The worst-case scenario is maximum fS (1 GHz) and minimum fO (62.5 MHz), which yields ΔfO/fO = 2.8 × 10−14, which is less than one part in ten trillion. Rev. 0 | Page 36 of 104 AD9547 Recovery from Holdover System Clock Period When in holdover, if a valid reference becomes available, the device exits holdover operation. The loop state machine restores the DPLL to closed-loop operation, locks to the selected reference, and sequences the recovery of all the loop parameters based on the profile settings for the active reference. Many of the user-programmable parameters of the AD9547 have absolute time units. To make this possible, the AD9547 requires a priori knowledge of the period of the system clock. To accommodate this requirement, the user programs the 21-bit nominal system clock period in the nominal SYSCLK period register (Address 0x0103 to Address 0x0105). The contents of this register reflect the actual period of the system clock in units of femtoseconds (fs). The user must program this register properly to ensure proper operation of the device because many of its subsystems rely on this value. Note that if the user holdover bit (Register 0x0A01, Bit 6) is set, the device does not automatically exit holdover when a valid reference is available. However, automatic recovery can occur after clearing the user holdover bit. SYSTEM CLOCK INPUTS System Clock Details Functional Description The system clock circuit provides a low jitter, stable, high frequency clock for use by the rest of the chip. The user has the option of directly driving the SYSCLKx inputs with a high frequency clock source at the desired system clock rate. Alternatively, the SYSCLKx input can be configured to operate in conjunction with the internal SYSCLK PLL. The SYSCLK PLL can synthesize the system clock by means of a crystal resonator connected across the SYSCLKx input pins or by means of direct application of a low frequency clock source. The SYSCLKx inputs are internally biased to a dc level of ~1 V. Take care to ensure that any external connections do not disturb the dc bias because such a disturbance can significantly degrade performance. Generally, the SYSCLKx inputs should be ac-coupled to the signal source (except when using a crystal resonator). LF SYSCLKN 37 A block diagram of the system clock appears in Figure 43. The signal at the SYSCLKx input pins becomes the internally buffered DAC sampling clock (fS) via one of three paths. • • • High frequency direct (HF) Low frequency synthesized (LF) Crystal resonator synthesized (XTAL) Note that both the LF and XTAL paths require the use of the SYSCLK PLL (see the SYSCLK PLL Multiplier section). The main purpose of the HF path is to allow the direct use of a high frequency (500 MHz to 1 GHz) external clock source for clocking the AD9547. This path is optimized for high frequency and low noise floor. Note that the HF input also provides a path to SYSCLK PLL (see the SYSCLK PLL Multiplier section), which includes an input divider (M) programmable for divide-by −1, −2, −4, or −8. The purpose of the divider is to limit the frequency at the input to the PLL to less than 150 MHz, which is the maximum PFD rate. SYSCLK_VREG SYSCLK_LF 34 35 2× LOCK DETECT ÷M PFD AND CHARGE PUMP XTAL SYSCLKP 38 VCO CALIBRATION LOOP FILTER ÷N SYSTEM CLOCK 08300-020 HF Figure 43. System Clock Block Diagram Rev. 0 | Page 37 of 104 AD9547 The LF path permits the user to provide an LVPECL, LVDS, CMOS, or sinusoidal low frequency clock for multiplication by the integrated SYSCLK PLL. The LF path handles input frequencies from 3.5 MHz up to 100 MHz. However, when using a sinusoidal input signal, it is best to use a frequency in excess of 20 MHz. Otherwise, the resulting low slew rate can lead to substandard noise performance. Note that the LF path includes an optional 2× frequency multiplier to double the rate at the input to the SYSCLK PLL and potentially reduce the PLL in-band noise. However, to avoid exceeding the maximum PFD rate of 150 MHz, use of the 2× frequency multiplier is valid only for input frequencies below 125 MHz. The XTAL path enables the connection of a crystal resonator (typically 10 MHz to 50 MHz) across the SYSCLKx input pins. An internal amplifier provides the negative resistance required to induce oscillation. The internal amplifier expects a 3.2 mm × 2.5 mm AT cut, fundamental mode crystal with a maximum motional resistance of 100 Ω. The following crystals, listed in alphabetical order, may meet these criteria. Note that, although these crystals meet the preceding criteria according to their data sheets, Analog Devices, Inc., does not guarantee their opera-tion with the AD9547, nor does Analog Devices endorse one crystal manufacturer/supplier over another. AVX/Kyocera CX3225SB ECS ECX-32 Epson/Toyocom TSX-3225 NDK NX3225SA Siward SX-3225 SYSCLK PLL MULTIPLIER The SYSCLK PLL multiplier is an integer-N design and relies on an integrated LC tank and VCO. It provides a means to convert a low frequency clock input to the desired system clock frequency, fS (900 MHz to 1 GHz). The SYSCLK PLL multiplier accepts input signals between 3.5 MHz and 500 MHz, but frequencies in excess of 150 MHz require the M-divider to ensure compliance with the maximum PFD rate (150 MHz). The PLL contains a feedback divider (N) that is programmable for divide values between 6 and 255. The nominal VCO gain is 70 MHz/V. Lock Detector The SYSCLK PLL phase detector operates at the PFD rate, which is fVCO/N. Each PFD sample indicates whether the reference and feedback signals are phase aligned (within a certain threshold range). While the PLL is in the process of acquiring a lock condition, the PFD samples typically consist of an arbitrary sequence of in-phase and out-of-phase indications. As the PLL approaches complete phase lock, the number of consecutive in-phase PFD samples grows larger. Thus, one way of indicating a locked condition is to count the number of consecutive in-phase PFD samples and, if it exceeds a certain value, declare the PLL locked. This is exactly the role of the lock detect divider bits. When the lock detector is enabled (Register 0x0100, Bit 2 = 0), the lock detect divider bits determine the number of consecutive in-phase decisions that are required (128, 256, 512, or 1024) before the lock detector declares a locked condition. The default setting is 128. Charge Pump The charge pump operates in either automatic or manual mode, based on the charge pump mode bit (Register 0x0100, Bit 6). When Register 0x0100, Bit 6 = 0, the AD9547 automatically selects the appropriate charge pump current based on the N divider value. Note that the user does not have control of the charge pump current bits (Register 0x0100, Bits[5:3]) in automatic mode. When Register 0x0100, Bit 6 = 1, the user determines the charge pump current via the charge pump current bits (Register 0x0100, Bits[5:3]). The charge pump current varies from 125 μA to 1 mA in 125 μA steps. The default setting is 500 μA. SYSCLK PLL Loop Filter The AD9547 has an internal second-order loop filter that establishes the loop dynamics for input signals between 12.5 MHz and 100 MHz. By default, the device uses the internal loop filter. However, an external loop filter option is available by setting the external loop filter enable bit (Register 0x0100, Bit 7). This bit bypasses the internal loop filter and allows the device to use an externally connected second-order loop filter, as shown in Figure 44. AD9547 The SYSCLK PLL has a built-in lock detector. Register 0x0100, Bit 2 determines whether the lock detector is active. When it is active (default), the user controls the sensitivity of the lock detector via the lock detect divider bits (Register 0x0100, Bits[1:0]). SYSCLK_VREG SYSCLK_LF 34 35 R1 C1 C2 08300-021 • • • • • Note that a value of zero must be written to the system clock stability timer (Register 0x0106 to Register 0x0108) whenever the lock detector is disabled (Register 0x0100, Bit 2 = 1). Figure 44. External Loop Filter Schematic Rev. 0 | Page 38 of 104 AD9547 To determine the external loop filter components, the user decides on the desired open loop bandwidth (fOL) and phase margin (φ). These parameters allow calculation of the loop filter components, as follows: C2 = Note that the monitors/detectors associated with the input references (REF A/REF AA and REF B/REF BB) are internally disabled until the SYSCLK PLL indicates that it is stable. I CP K VCO tan(φ) 2N (πf OL ) 2 I CP K VCO N (2πf OL )2 CLOCK DISTRIBUTION The clock distribution block of the AD9547 provides an integrated solution for generating multiple clock outputs based on frequency dividing the DPLL output. The distribution output consists of two channels (OUT0 and OUT1). Each channel has a dedicated divider and output driver section, as shown in Figure 45. ⎛ 1 − sin(φ) ⎞ ⎜⎜ ⎟⎟ ⎝ cos(φ) ⎠ where: KVCO = 7 × 107 V/ns (typical). ICP is the programmed charge pump current (amperes). N is the programmed feedback divider value. fOL is the desired open-loop bandwidth (Hz). Φ is the desired phase margin (radians). CLKINP CLKINN SYNC CONTROL ENABLE For example, assuming that N = 40, ICP = 0.5 mA, fOL = 400 kHz, and Φ = 50°, then the loop filter calculations yield R1 = 3.31 kΩ, C1 = 330 pF, and C2 = 50.4 pF. System Clock Stability Timer 4 The system clock stability timer, located in Register 0x0106 to Register 0x0108, is a 20-bit value programmed in units of milliseconds (ms). If the programmed timer value is 0, the timer immediately indicates that it has timed out. If the programmed timer value is nonzero and the SYSCLK PLL is enabled, the timer starts timing when the SYSCLK PLL lock detector indicates lock and times out after the prescribed period. However, when the user disables the SYSCLK PLL, the timer ignores the SYSCLK PLL lock detector and starts timing the instant that the SYSCLK PLL is disabled. The user can monitor the status of the stability timer at Register 0x0D01, Bit 4, via the multifunction pins or via the IRQ pin. Note that the system clock stability timer must be programmed before the SYSCLK PLL is either activated or disabled. SYSCLK PLL Calibration When using the SYSCLK PLL, it is necessary to calibrate the LC-VCO to ensure that the PLL can remain locked to the system clock input signal. Assuming the presence of either an external SYSCLK input signal or a crystal resonator, the calibration process executes after the user sets and then clears the calibrate system clock bit in the cal/sync register (Register 0x0A02, Bit 0). During the calibration process, the device calibrates the VCO amplitude and frequency. The status of the system clock calibration process is user accessible via the system clock status register (Register 0x0D01, Bit 1). It is also available via the IRQ monitor register (Bit 1 of Register 0x0D02), provided that the status bit is enabled via the IRQ mask register (Register 0x0209 and Register 0x0210). SYNC SOURCE ENABLEn/MODEn 4 4 Q0 OUT_RSET OUT0P OUT0N OUT0 OUT1 OUT1P OUT1N 08300-022 C1 = πNf OL ⎛ 1 ⎞ ⎜⎜1 + ⎟ sin(φ) ⎟⎠ I CP K VCO ⎝ RESET R1 = When the calibration sequence is complete, the SYSCLK PLL eventually attains a lock condition, at which point the system clock stability timer begins its countdown sequence. Expiration of the timer indicates that the SYSCLK PLL is stable, which is reflected in the system clock status register (Register 0x0D01, Bit 4). Figure 45. Clock Distribution Clock Input (CLKINx) The clock input handles input signals from a variety of logic families (assuming proper terminations and sufficient voltage swing). It also handles sine wave input signals such as those delivered by the DAC reconstruction filter. Its default operating frequency range is 62.5 MHz to 500 MHz. Super-Nyquist Operation Typically, the maximum usable frequency at the DAC output is about 45% of the system clock frequency. However, because it is a sampled DAC, its output spectrum contains Nyquist images. Of particular interest are the images appearing in the first Nyquist zone (50% to 100% of the system clock frequency). Super-Nyquist operation takes advantage of these higher frequencies, but this implies that the CLKINx input operates in excess of 500 MHz, which is outside its default operating limits. The CLKINx receiver actually consists of two separate receivers: the default receiver and an optional high frequency receiver, which handles input signals up to 800 MHz. To select the high frequency receiver, write a Logic 1 to Register 0x0400, Bit4. Rev. 0 | Page 39 of 104 AD9547 Super-Nyquist operation requires a band-pass filter at the DAC output instead of the usual low-pass reconstruction filter. SuperNyquist operation is viable as long as the image frequency does not exceed the 800 MHz input range of the receiver. Furthermore, to provide acceptable jitter performance, which is a consideration for image signals with low amplitude, the signal at the CLKINx inputs must meet the minimum slew rate requirements. Output Enable Clock Dividers Output Mode The output clock distribution dividers are referred to as Q0 and Q1, corresponding to the OUT0 and OUT1 output channels, respectively. Each divider is programmable with 30 bits of division depth. The actual divide ratio is one more than the programmed register value; therefore, a register value of 3, for example, results in a divide ratio of 4. Thus, each divider offers a range of divide ratios from 1 to 230 (1 to 1,073,741,824). The user has independent control of the operating mode of each of the two output channels via the distribution channel modes register (Address 0x0404 and Address 0x0405). The operating mode control includes With an even divide ratio, the output signal always exhibits a 50% duty cycle. When the clock divider is bypassed (a divide ratio of 1), the output duty cycle is the same as the input duty cycle. Odd output divide ratios (excluding 1) exhibit automatic duty cycle correction given by Output Duty Cycle = N + 2X −1 2N For example, if N = 5 and the input duty cycle is 20% (X = 0.2), then the output duty cycle is 44%. Note that, when the user programs an output as noninverting, then the device adjusts the falling edge timing to accomplish the duty cycle cor-rection. Conversely, the device adjusts the rising edge timing for an inverted output. Output Power-Down Each output channel offers independent control of power-down functionality via the distribution settings register (Address 0x0400). Each output channel has a dedicated power-down bit for powering down the output driver. However, if both channels are powered down, the entire distribution output enters a deep sleep mode. Even though each channel has a channel power-down control signal, it may sometimes be desirable to power down an output driver while maintaining the divider’s synchronization with the other channel dividers. This is accomplished by either of the following methods: • • • • Logic family and pin functionality Output drive strength Output polarity The three LSBs of both distribution channel mode registers comprise the mode bits. The mode value selects the desired logic family and pin functionality of an output channel, as listed in Table 23. Table 23. Output Channel Logic Family and Pin Functionality where: N (which must be an odd number) is the divide ratio. X is the normalized fraction of the high portion of the input period (that is, 0 < X < 1). • Each output channel offers independent control of enable/ disable functionality using the distribution enable register (Address 0x0401). The distribution outputs use synchronization logic to control enable/disable activity to avoid the production of runt pulses and to ensure that outputs with the same divide ratios become active/inactive in unison. In CMOS mode, use the divider output enable control bit to stall an output. This provides power savings while maintaining dc drive at the output. In LVDS/LVPECL mode, place the output in tristate mode (this works in CMOS mode as well). Mode Bits [2:0] 000 001 010 011 100 101 110 111 Logic Family and Pin Functionality CMOS (both pins) CMOS (positive pin), tristate (negative pin) Tristate (positive pin), CMOS (negative pin) Tristate (both pins) LVDS LVPECL Unused Unused Regardless of the selected logic family, each is capable of dc operation. However, the upper frequency is limited by the load conditions, drive strength, and impedance matching inherent in each logic family. Practical limitations set the maximum CMOS frequency to approximately 250 MHz, whereas LVPECL and LVDS are capable of 725 MHz. In addition to the three mode bits, both distribution channel mode registers include the following control bits: • • • Polarity invert CMOS phase invert Drive strength The polarity invert bit enables the user to choose between normal polarity and inverted polarity. Normal polarity is the default state. Inverted polarity reverses the representation of Logic 0 and Logic 1 regardless of the logic family. The CMOS phase invert bit applies only when the mode bits select the CMOS logic family. In CMOS mode, both output pins of the channel have a dedicated CMOS driver. By default, both drivers deliver identical signals. However, setting the CMOS phase invert bit causes the signal on an OUTxN pin to be the opposite of the signal appearing on the OUTxP pin. Rev. 0 | Page 40 of 104 AD9547 The drive strength bit allows the user to control whether the output uses weak (0) or strong (1) drive capability (applies to CMOS and LVDS but not LVPECL). For the CMOS family, the strong setting implies normal CMOS drive capability, whereas the weak setting implies low capacitive loading and allows for reduced EMI. For the LVDS family, the weak setting provides 3.5 mA drive current for standard LVDS operation, whereas the strong setting provides 7 mA for double terminated or double voltage LVDS operation. Note that 3.5 mA and 7 mA are the nominal drive current values when using the internal current setting resistor. Clock Distribution Synchronization A block diagram of the distribution synchronization functionality appears in Figure 46. The synchronization sequence begins with the primary synchronization signal, which ultimately results in delivery of a synchronization strobe to the clock distribution logic. As indicated, the primary synchronization signal originates from the following four possible sources: • • Output Current Control with an External Resistor By default, the output drivers have an internal current setting resistor (3.12 kΩ nominal) that establishes the nominal drive current for the LVDS and LVPECL operating modes. Instead of using the internal resistor, the user can elect to set the external distribution resistor bit (Register 0x0400, Bit 5) and connect an external resistor to the OUT_RSET pin. Note that this feature supports an external resistor value of 3.12 kΩ only, allowing for tighter control of the output current than is possible by using the internal current setting resistor. However, if the user elects to use a nonstandard external resistance, the following equations provide the output drive current as a function of the external resistance (R): I LVDS0 10.8325 = R I LVDS1 = 21.665 R I LVPECL = 24.76 R The numeric subscript associated with the LVDS output current corresponds to the logic state of the drive strength bit in the distribution channel modes registers (Address 0x0404, Bit 3 and Address 0x0405, Bit 3). For R = 3.12 kΩ, the equations yield ILVDS0 = 3.5 mA, ILVDS1 = 7.0 mA, and ILVPECL = 8.0 mA. Note that the device maintains a constant 1.238 V (nominal) across the external resistor. • • Direct synchronization source via the sync distribution bit (Register 0x0A02, Bit 1) Automatic synchronization source based on frequency or phase lock detection, as controlled via the automatic synchronization register (Address 0x0403) Multifunction pin synchronization source via one of the multifunction pins (M0 to M7) EEPROM synchronization source via the EEPROM All four sources of the primary synchronization signal are logic OR’d, so any one of them can synchronize the clock distribution output at any time. When using the multifunction pins, the synchronization event is the falling edge of the selected signal. When using the sync distribution bit, the user first sets then clears the bit. The synchronization event is the clearing operation; that is, the Logic 1 to Logic 0 transition of the bit. The primary synchronization signal can synchronize the distribution output directly, or it can enable a secondary synchronization signal. This functionality depends on the two sync source bits in the distribution synchronization register (Register 0x0402, Bits[5:4]). When sync source = 00 (direct), the falling edge of the primary synchronization signal directly synchronizes the distribution output. When sync source = 01, the rising edge of the primary synchronization signal triggers the circuitry that detects a rising edge of the active input reference. The detection of the rising edge synchronizes the distribution output. When sync source = 10, the rising edge of the primary synchronization signal triggers the circuitry that detects a rollover of the DDS accumulator (after processing by the DPLL feedback divider). This corresponds to the zero crossing of the output of the phase-toamplitude converter in the DDS (less the open-loop phase offset stored in Register 0x030D and Register 0x030E). The detection of the DPLL feedback edge synchronizes the distribution output. Rev. 0 | Page 41 of 104 AD9547 Active Reference Synchronization (Zero Delay) Active reference synchronization is the term applied to the case when sync source = 01 (Register 0x0402, Bits[5:4]). Referring to Figure 46, this means that the active reference sync path is active because Bit 4 = 1, enabling the lower AND gate and disabling the upper AND gate. The edge detector in the active reference sync block monitors the rising edges of the active reference (the mux selects the active reference automatically). The edge detector is armed via the primary synchronization signal, which is one of the four inputs to the OR gate (typically the direct sync source). As soon as the edge detector is armed, its output goes high, which stalls the output dividers in the clock distribution block. Furthermore, once armed, a rising edge from the active reference forces the output of the edge detector low. This restarts the output dividers, thereby synchronizing the clock distribution block. The term zero delay applies because it provides a means to edgealign the output signal with the active input reference signal. Typically, zero-delay architectures use the output signal in the feedback loop of a PLL to track input/output edge alignment. REGISTER 0x0402[5] PRIMARY SYNCHRONIZATION SIGNAL DIRECT SYNC AUTOMATIC SYNC SOURCE (REGISTER DX 0x0403) EEPROM SYNC SOURCE TO MULTIFUNCTION PIN STATUS LOGIC 0 TO CLOCK DISTRIBUTION SYNCHRONIZATION CONTROL 1 EDGE DETECT MULTIFUNCTION PIN SYNC SOURCE The fact that an active reference edge triggers the falling edge of the synchronization pulse means that the falling edge is asynchronous to the signal that clocks the distribution output dividers (CLKINx). Therefore, the output clock distribution logic reclocks the internal synchronization pulse to synchronize it with the CLKINx signal. This means that the output dividers restart after a deterministic delay associated with the reclocking circuitry. This deterministic delay has two components. The first deterministic delay component is four or five periods of the CLKINx signal. The one period uncertainty is due to the unknown position of the asynchronous reference clock edge relative to the CLKINx signal. The second deterministic delay component is one output period of the distribution divider. DPLL FEEDBACK EDGE ARM STALL DIVIDERS SYNC OUTPUT DISTRIBUTION EDGE DETECT SYSCLK/4 DPLL EDGE SYNC REGISTER 0x0402[4] RESET ARM REF A EDGE DETECT REF BB ACTIVE REFERENCE SYNC Figure 46. Output Synchronization Block Diagram Rev. 0 | Page 42 of 104 08300-023 DIRECT SYNC SOURCE (ADDRESS DX 0x0A02[1]) Active reference synchronization, however, operates open loop. That is, synchronization of the output via the distribution synchronization logic occurs on a single edge of the active reference. AD9547 The deterministic delay, expressed as tLATENCY in the following equation, is a function of the frequency division factor (Qn) of the channel divider associated with the zero-delay channel. tLATENCY = (Qn + 4) × tCLK_IN or tLATENCY = (Qn + 5) × tCLK_IN In addition to deterministic delay, there is random delay (tPROP) associated with the propagation of the reference signal through the input reference receiver, as well as the propagation of the clock signal through the clock distribution logic. The total delay is tDELAY = tLATENCY + tPROP The user can compensate for tDELAY by using the device’s phase offset controls to move the edge timing of the distribution output signal relative to the input reference edge. One method is to use the open-loop phase offset registers (Address 0x030D and Address 0x030E) for timing adjustment. However, be sure to use sufficiently small phase increments to make the adjustment. Too large a phase step can result in the clock distribution logic missing a CLKINx edge, thus disrupting the edge alignment process. The appropriate phase increment depends on the transient response of any external circuitry connected between the DACOUTx and CLKINx pins. The other method is to use the closed-loop phase offset registers (Address 0x030F to Address 0x0315) for timing adjustment. However, be sure to use a sufficiently small phase vs. time profile. Changing the phase too quickly can cause the DPLL to lose lock, thus ruining the edge alignment process. Note that the AD9547 phase slew limit register (Address 0x0316 and 0x0317) can be used to limit the rate of change of phase automatically, thereby mitigating the potential loss-of-lock problem. To guarantee synchronization of the output dividers, it is important to make any edge timing adjustments after the synchronization event. Furthermore, when making timing adjustments, the distribution outputs can be disabled and then renabled after the adjustment is complete. This prevents the device from generating output clock signals during the timing adjustment process. Note that the form of zero-delay synchronization described here does not track propagation time variations within the distribution clock input path or the reference input path (on or off chip) over temperature, supply, and so on. It is strictly a one-time synchronization event. Synchronization Mask Each output channel has a dedicated synchronization mask bit (Register 0x0402, Bits[1:0]). When the mask bit associated with a particular channel is set, that channel does not respond to the synchronization signal. This allows the device to operate with the masked channels active and the unmasked channels stalled while they wait for a synchronization pulse. Rev. 0 | Page 43 of 104 AD9547 STATUS AND CONTROL MULTIFUNCTION PINS (M0 TO M7) The AD9547 has eight digital CMOS I/O pins (M0 to M7) that are configurable for a variety of uses. The function of these pins is programmable via the register map. Each pin can control or monitor an assortment of internal functions, based on the contents of Register 0x0200 to Register 0x0207. To monitor an internal function with a multifunction pin, write a Logic 1 to the MSB of the register associated with the desired multifunction pin. The value of the seven LSBs of the register defines the control function, as shown in Table 24. Table 24. Multifunction Pin Output Functions (D7 = 1) D[6:0] Value 0 1 2 3 4 5 6 7 8 9 10 11 12 to 15 16 17 18 19 20 21 22 23 24 25 26 27 to 31 32 33 34 35 36 to 47 48 49 50 51 52 to 63 64 Output Function Static Logic 0 Static Logic 1 System clock divided by 32 Watchdog timer output EEPROM upload in progress EEPROM download in progress EEPROM fault detected SYSCLK PLL lock detected SYSCLK PLL calibration in progress Unused Unused SYSCLK PLL stable Unused DPLL free running DPLL active DPLL in holdover DPLL in reference switchover Active reference: phase master DPLL phase locked DPLL frequency locked DPLL phase slew limited DPLL frequency clamped Tuning word history available Tuning word history updated Unused Reference A fault Reference AA fault Reference B fault Reference BB fault Unused Reference A valid Reference AA valid Reference B valid Reference BB valid Unused Reference A active reference 65 66 Reference AA active reference Reference B active reference 67 68 to 79 Reference BB active reference Unused D[6:0] Value 80 81 to 127 Output Function Clock distribution sync pulse Unused Source Proxy Register 0x0D03, Bit 3 To control an internal function with a multifunction pin, write a Logic 0 to the most significant bit of the register associated with the desired multifunction pin. The monitored function depends on the value of the seven least significant bits of the register, as shown in Table 25. Note that the default setting is M0 through M7 configured as inputs and the input function set to unused (the first entry in Table 25). Table 25. Multifunction Pin Input Functions (D7 = 0) Source Proxy Register 0x0D00, Bit 0 Register 0x0D00, Bit 1 Register 0x0D00, Bit 2 Register 0x0D01, Bit 0 Register 0x0D01, Bit 1 Unused Unused Register 0x0D01, Bit 4 Unused Register 0x0D0A, Bit 0 Register 0x0D0A, Bit 1 Register 0x0D0A, Bit 2 Register 0x0D0A, Bit 3 Register 0x0D0A, Bit 6 Register 0x0D0A, Bit 4 Register 0x0D0A, Bit 5 Register 0x0D0A, Bit 7 Register 0x0D0B, Bit 7 Register 0x0D0B, Bit 6 Register 0x0D05, Bit 4 Unused Register 0x0D0C, Bit 2 Register 0x0D0D, Bit 2 Register 0x0D0E, Bit 2 Register 0x0D0F, Bit 2 Unused Register 0x0D0C, Bit 3 Register 0x0D0D, Bit 3 Register 0x0D0E, Bit 3 Register 0x0D0F, Bit 3 Unused Register 0x0D0B, Bits[1:0] Register 0x0D0B, Bits[1:0} Register 0x0D0B, Bits[1:0] Register 0x0D0B, Bits[1:0] Unused D[6:0] Value 0 1 2 3 4 5 6 to 15 16 17 18 19 20 21 to 31 32 33 34 35 36 to 47 48 49 50 51 52 to 63 64 65 66, 67 68 69 70 to 127 Input Function Unused (default) I/O update Full power-down Watchdog reset IRQ reset Tuning word history reset Unused Holdover Free run Reset incremental phase offset Increment incremental phase offset Decrement incremental phase offset Unused Override Reference Monitor A Override Reference Monitor AA Override Reference Monitor B Override Reference Monitor BB Unused Force Validation Timeout A Force Validation Timeout AA Force Validation Timeout B Force Validation Timeout BB Unused Enable OUT0 Enable OUT1 Unused Enable OUT0, OUT1 Sync clock distribution outputs Unused Destination Proxy Unused Register 0x0005, Bit 0 Register 0x0A00, Bit 0 Register 0x0A03, Bit 0 Register 0x0A03, Bit 1 Register 0x0A03, Bit 2 Unused Register 0x0A01, Bit 6 Register 0x0A01, Bit 5 Register 0x0A0C, Bit 2 Register 0x0A0C, Bit 0 Register 0x0A0C, Bit 1 Unused Register 0x0A0F, Bit 0 Register 0x0A0F, Bit 1 Register 0x0A0F, Bit 2 Register 0x0A0F, Bit 3 Unused Register 0x0A0E, Bit 0 Register 0x0A0E, Bit 1 Register 0x0A0E, Bit 2 Register 0x0A0E, Bit 3 Unused Register 0x0401, Bit 0 Register 0x0401, Bit 1 Unused Register 0x0401, Bits[1:0] Register 0x0A02, Bit 1 Unused If more than one multifunction pin operates on the same control signal, then internal priority logic ensures that only one multifunction pin serves as the signal source. The selected pin is the one with the lowest numeric suffix. For example, if both M3 and M7 operate on the same control signal, M3 is used as the signal source and the redundant pin is ignored. Rev. 0 | Page 44 of 104 AD9547 At power-up, the multifunction pins can be used to force the device into certain configurations as defined in the Initial M0 to M7 Pin Programming section. This functionality, however, is valid only during power-up or following a reset, after which the pins can be reconfigured via the serial programming port or via the EEPROM. IRQ PIN The AD9547 has a dedicated interrupt request (IRQ) pin. The IRQ pin output mode register (Register 0x0208, Bits[1:0]) controls how the IRQ pin asserts an interrupt based on the value of the two bits, as shown in Table 26. Table 26. IRQ Pin Control—Register 0x0208, Bits[1:0] Setting 00 01 10 11 Description The IRQ pin is high impedance when deasserted and active low when asserted and requires an external pull-up resistor (this is the default operating mode). The IRQ pin is high impedance when deasserted and active high when asserted and requires an external pull-down resistor. The IRQ pin is Logic 0 when deasserted and Logic 1 when asserted. The IRQ pin is Logic 1 when deasserted and Logic 0 when asserted. The AD9547 asserts the IRQ pin when any of the bits in the IRQ monitor registers (Address 0x0D02 to Address 0x0D09) are Logic 1. Each bit in these registers is associated with an internal function that is capable of producing an interrupt. Furthermore, each bit of the IRQ monitor register is the result of a logical AND of the associated internal interrupt signal and the corresponding bit in the IRQ mask register (Address 0x0209 to Address 0x0210). That is, the bits in the IRQ mask register have a one-to-one correspondence with the bits in the IRQ monitor register. When an internal function produces an interrupt signal and the associated IRQ mask bit is set, the corresponding bit in the IRQ monitor register is set. The user should be aware that clearing a bit in the IRQ mask register removes only the mask associated with the internal interrupt signal. It does not clear the corresponding bit in the IRQ monitor register. The IRQ pin is the result of a logical OR of all the IRQ monitor register bits. Thus, the AD9547 asserts the IRQ pin as long as any of the IRQ monitor register bits are Logic 1. Note that it is possible to have multiple bits set in the IRQ monitor register. Therefore, when the AD9547 asserts the IRQ pin, it may indicate an interrupt from several different internal functions. The IRQ monitor register provides the user with a means to interrogate the AD9547 to determine which internal function(s) produced the interrupt. Typically, when the AD9547 asserts the IRQ pin, the user interrogates the IRQ monitor register to identify the source of the interrupt request. After servicing an indicated interrupt, the user should clear the associated IRQ monitor register bit via the IRQ clearing registers (Address 0x0A04 to Address 0x0A0B). The bits in the IRQ clearing register have a one-to-one correspondence with the bits in the IRQ monitor register. Note that the IRQ clearing register is autoclearing. The IRQ pin remains asserted until the user clears all of the bits in the IRQ monitor register that indicate an interrupt. It is also possible to collectively clear all of the IRQ monitor register bits by setting the reset all IRQs bit in the reset functions register (Register 0x0A03, Bit 1). Note that this is an autoclearing bit. Setting this bit results in deassertion of the IRQ pin. Alternatively, the user can program any of the multifunction pins to clear all IRQs. This allows the user to clear all IRQs by means of a hardware pin rather than by a serial I/O port operation. WATCHDOG TIMER The watchdog timer is a general-purpose, programmable timer. To set its timeout period, the user writes to the 16-bit watchdog timer register (Address 0x0211 and Address 0x0212). A value of zero in this register disables the timer. A nonzero value sets the timeout period in units of milliseconds (ms), giving the watchdog timer a range of 1 ms to 65,535 ms. The relative accuracy of the timer is approximately 0.1% with an uncertainty of 0.5 ms. If enabled, the timer runs continuously and generates a timeout event whenever the timeout period expires. The user has access to the watchdog timer status via the IRQ mechanism and the multifunction pins (M0 to M7). In the case of the multifunction pins, the timeout event of the watchdog timer is a pulse that lasts 32 system clock periods. There are two ways to reset the watchdog timer (thereby preventing it from causing a timeout event). The first is by writing a Logic 1 to the autoclearing reset watchdog bit in the reset functions register (Register 0x0A03, Bit 0). Alternatively, the user can program any of the multifunction pins to reset the watchdog timer. This allows the user to reset the timer by means of a hardware pin rather than by a serial I/O port operation. Rev. 0 | Page 45 of 104 AD9547 EEPROM EEPROM Overview The AD9547 contains an integrated 2048-byte electrically erasable programmable read-only memory (EEPROM). The AD9547 can be configured to perform a download at power-up via the multifunction pins (M3 to M7), but uploads and downloads can also be done on demand via the EEPROM control registers (Address 0x0E00 to Address 0x0E03). The EEPROM provides the ability to upload and download configuration settings to and from the register map. Figure 47 shows a functional diagram of the EEPROM. Register 0x0E10 to Register 0x0E3F represent a 48-byte scratch pad that enables the user to store a sequence of instructions for transferring data to the EEPROM from the device settings portion of the register map. Note that the default values for these registers provide a sample sequence for saving/retrieving all of the AD9547 EEPROM-accessible registers. Figure 47 shows the connectivity between the EEPROM and the controller that manages data transfer between the EEPROM and the register map. The controller oversees the process of transferring EEPROM data to and from the register map. There are two modes of operation handled by the controller: saving data to the EEPROM (upload mode) or retrieving data from the EEPROM (download mode). In either case, the controller relies on a specific instruction set. DATA EEPROM ADDRESS POINTER DATA DEVICE SETTINGS (0x0100 TO 0x0A10) CONDITION (0x0E01 [4:0]) DEVICE SETTINGS ADDRESS POINTER DATA EEPROM CONTROLLER EEPROM (0x000 TO 0x7FF) SCRATCH PAD ADDRESS POINTER SCRATCH PAD (0x0E10 TO 0x0E3F) REGISTER MAP Note that, in the EEPROM scratch pad, the two registers that make up the address portion of a data instruction have the MSB of the address in the D7 position of the lower register address. The bit weight increases left to right, from the lower register address to the higher register address. Furthermore, the starting address always indicates the lowest numbered register map address in the range of bytes to transfer. That is, the controller always starts at the register map target address and counts upward, regardless of whether the serial I/O port is operating in I2C, SPI LSB first, or SPI MSB first mode. As part of the data transfer process during an EEPROM upload, the controller calculates a 1-byte checksum and stores it as the final byte of the data transfer. As part of the data transfer process during an EEPROM download, however, the controller again calculates a 1-byte checksum value but compares the newly calculated checksum with the one that was stored during the upload process. If an upload/download checksum pair does not match, the controller sets the EEPROM fault status bit (Register 0x0D03, Bit 1). If the upload/download checksums match for all data instructions encountered during a download sequence, the controller sets the EEPROM complete status bit (Register 0x0D03, Bit 0). Condition instructions are those that have a value from 0xB0 to 0xCF. Condition Instruction 0xB1 to Condition Instruction 0xCF represent Condition 1 to Condition 31, respectively. Condition Instruction 0xB0 is special because it represents the null condition (see the EEPROM Conditional Processing section). 08300-024 M7 M6 M5 M4 M3 The controller decodes the number of bytes to transfer directly from the data instruction itself by adding one to the value of the instruction. For example, the data instruction, 0x1A, has a decimal value of 26; therefore, the controller knows to transfer 27 bytes (one more than the value of the instruction). Whenever the controller encounters a data instruction, it knows to read the next two bytes in the scratch pad because these bytes contain the register map target address. SERIAL INPUT/OUTPUT PORT Figure 47. EEPROM Functional Diagram EEPROM Instructions Table 27 lists the EEPROM controller instruction set. The controller recognizes all instruction types, whether it is in upload or download mode, except for the pause instruction, which it recognizes only in upload mode. The I/O update, calibrate, distribution sync, and end instructions are mostly self-explanatory. The others, however, warrant further detail, as described in the following paragraphs. A pause instruction, like an end instruction, is stored at the end of a sequence of instructions in the scratch pad. When the controller encounters a pause instruction during an upload sequence, it keeps the EEPROM address pointer at its last value. This way, the user can store a new instruction sequence in the scratch pad and upload the new sequence to the EEPROM. The new sequence is stored in the EEPROM address locations immediately following the previously saved sequence. This process is repeatable until an upload sequence contains an end instruction. The pause instruction is also useful when used in conjunction with condition processing. It allows the EEPROM to contain multiple occurrences of the same register(s), with each occurrence linked to a set of conditions (see the EEPROM Conditional Processing section). Data instructions are those that have a value from 0x00 to 0x7F. A data instruction tells the controller to transfer data between the EEPROM and the register map. The controller needs the following two parameters to carry out the data transfer: • • The number of bytes to transfer The register map target address Rev. 0 | Page 46 of 104 AD9547 Table 27. EEPROM Controller Instruction Set Instruction Value (Hex) 0x00 to 0x7F Instruction Type Data Bytes Required 3 0x80 I/O update 1 0xA0 Calibrate 1 0xA1 Distribution sync 1 0xB0 to 0xCF Condition 1 0xFE Pause 1 0xFF End 1 Description A data instruction tells the controller to transfer data to or from the device settings part of the register map. A data instruction requires two additional bytes that, together, indicate a starting address in the register map. Encoded in the data instruction is the number of bytes to transfer, which is one more than the instruction value. When the controller encounters this instruction while downloading from the EEPROM, it issues a soft I/O update (see Register 0x0005 in Table 42). When the controller encounters this instruction while downloading from the EEPROM, it initiates a system clock calibration sequence (see Register 0x0A02 in Table 120). When the controller encounters this instruction while downloading from the EEPROM, it issues a sync pulse to the output distribution synchronization (see Register 0x0A02 in Table 120). 0xB1 to 0xCF are condition instructions and correspond to Condition 1 to Condition 31, respectively. 0xB0 is the null condition instruction. See the EEPROM Conditional Processing section for details. When the controller encounters this instruction in the scratch pad while uploading to the EEPROM, it resets the scratch pad address pointer and holds the EEPROM address pointer at its last value. This allows storage of more than one instruction sequence in the EEPROM. Note that the controller does not copy this instruction to the EEPROM during upload. When the controller encounters this instruction in the scratch pad while uploading to the EEPROM, it resets both the scratch pad address pointer and the EEPROM address pointer and then enters an idle state. When the controller encounters this instruction while downloading from the EEPROM, it resets the EEPROM address pointer and then enters an idle state. EEPROM Upload To upload data to the EEPROM, the user must first ensure that the write enable bit (Register 0x0E00, Bit 0) is set. Then, on setting the autoclearing save to EEPROM bit (Register 0x0E02, Bit 0), the controller initiates the EEPROM data storage process. Uploading EEPROM data requires that the user first write an instruction sequence into the scratch pad registers. During the upload process, the controller reads the scratch pad data byte by byte, starting at Register 0x0E10 and incrementing the scratch pad address pointer as it goes, until it reaches a pause or end instruction. As the controller reads the scratch pad data, it transfers the data from the scratch pad to the EEPROM (byte by byte) and increments the EEPROM address pointer accordingly, unless it encounters a data instruction. A data instruction tells the controller to transfer data from the device settings portion of the register map to the EEPROM. The number of bytes to transfer is encoded within the data instruction, and the starting address for the transfer appears in the next two bytes in the scratch pad. When the controller encounters a data instruction, it stores the instruction in the EEPROM, increments the EEPROM address pointer, decodes the number of bytes to be transferred, and increments the scratch pad address pointer. Then it retrieves the next two bytes from the scratch pad (the target address) and increments the scratch pad address pointer by 2. Next, the controller transfers the specified number of bytes from the register map (beginning at the target address) to the EEPROM. When it completes the data transfer, the controller stores an extra byte in the EEPROM to serve as a checksum for the transferred block of data. To account for the checksum byte, the controller increments the EEPROM address pointer by one more than the number of bytes transferred. Note that, when the controller transfers data associated with an active register, it actually transfers the buffered contents of the register (see the Buffered/Active Registers section for details on the difference between buffered and active registers). This allows for the transfer of nonzero autoclearing register contents. Note that conditional processing does not occur during an upload sequence (see the EEPROM Conditional Processing section). EEPROM Download An EEPROM download results in a transfer of data from the EEPROM to the device register map. To download data, the user sets the autoclearing load from EEPROM bit (Register 0x0E03, Bit 1). This commands the controller to initiate the EEPROM download process. During download, the controller reads the EEPROM data byte by byte, incrementing the EEPROM address pointer as it goes, until it reaches an end instruction. As the controller reads the EEPROM data, it executes the stored instructions, which includes transferring stored data to the device settings portion of the register map whenever it encounters a data instruction. Note that conditional processing is applicable only when downloading (see the EEPROM Conditional Processing section). Rev. 0 | Page 47 of 104 AD9547 Automatic EEPROM Download The condition originates from one of two sources (see Figure 48), as follows: Following power-up, assertion of the RESET pin, or a soft reset (Register 0x0000, Bit 5 = 1), if FncInit[7:3] ≠ 0 (see the Initial M0 to M7 Pin Programming section), the instruction sequence stored in the EEPROM executes automatically with condition = FncInit[7:3]. In this way, a previously stored set of register values downloads automatically on power-up or with a hard or soft reset. See the EEPROM Conditional Processing section for details regarding conditional processing and the way that it modifies the download process. • FncInit, Bits[7:3], which is the state of multifunction pins M3 to M7 at power-up (see the Initial M0 to M7 Pin Programming section) Register 0x0E01, Bits[4:0] • If Register 0x0E01, Bits[4:0] ≠ 0, then the condition is the value stored in Register 0x0E01, Bits[4:0]; otherwise, the condition is FncInit, Bits[7:3]. Note that a nonzero condition that is present in Register 0x0E01, Bits[4:0] takes precedence over FncInit, Bits[7:3]. EEPROM Conditional Processing The condition instructions allow conditional execution of EEPROM instructions during a download sequence. During an upload sequence, however, they are stored as is and have no effect on the upload process. The condition tag board is a table that is maintained by the EEPROM controller. When the controller encounters a condition instruction, it decodes Condition Instruction 0xB1 through Condition Instruction 0xCF as condition = 1 through condition = 31, respectively, and tags that particular condition in the condition tag board. However, Condition Instruction 0xB0 decodes as the null condition, for which the controller clears the condition tag board; subsequent download instructions execute unconditionally (until the controller encounters a new condition instruction). Note that, during EEPROM downloads, the condition instructions themselves and the end instruction always execute unconditionally. Conditional processing makes use of two elements: the condition (from Condition 1 to Condition 31) and the condition tag board. The relationships among the condition, the condition tag board, and the EEPROM controller appear schematically in Figure 48. During download, the EEPROM controller executes or skips instructions, depending on the value of the condition and the contents of the condition tag board. Note, however, that condition instructions and the end instruction always execute unconditionally during download. If condition = 0, all instructions during download execute unconditionally. If condition ≠ 0 and there are any tagged conditions in the condition tag board, the controller executes instructions only if the condition is tagged. Condition is a 5-bit value with 32 possibilities. Condition = 0 is the null condition. When the null condition is in effect, the EEPROM controller executes all instructions unconditionally. The remaining 31 possibilities, condition = 1 through condition = 31, modify the EEPROM controller’s handling of a download sequence. CONDITION TAG BOARD EXAMPLE CONDITION 3 AND CONDITION 13 ARE TAGGED M7 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 REGISTER 0x0E01, BITS[4:0] FncInit, BITS[7:3] 5 IF B1 ≤ INSTRUCTION ≤ CF, THEN TAG DECODED CONDITION IF INSTRUCTION = B0, THEN CLEAR ALL TAGS EEPROM STORE CONDITION INSTRUCTIONS AS THEY ARE READ FROM THE SCRATCH PAD. WATCH FOR OCCURRENCE OF CONDITION INSTRUCTIONS DURING DOWNLOAD. UPLOAD PROCEDURE 5 IF {0x0E01, BITS[4:0] ≠ 0} CONDITION = 0x0E01, BITS[4:0] ELSE CONDITION = FncInit, BITS[7:3] ENDIF 5 CONDITION CONDITION HANDLER SCRATCH PAD M3 EXECUTE/SKIP INSTRUCTION(S) DOWNLOAD PROCEDURE EEPROM CONTROLLER Figure 48. EEPROM Conditional Processing Rev. 0 | Page 48 of 104 IF {NO TAGS} OR {CONDITION = 0} EXECUTE INSTRUCTIONS ELSE IF {CONDITION IS TAGGED} EXECUTE INSTRUCTIONS ELSE SKIP INSTRUCTIONS ENDIF ENDIF AD9547 If the condition is not tagged, the controller skips instructions until it encounters a condition instruction that decodes as a tagged condition. Note that the condition tag board allows for multiple conditions to be tagged at any given moment. This conditional processing mechanism enables the user to have one download instruction sequence with many possible outcomes, depending on the value of the condition and the order in which the controller encounters the condition instructions. With the upload sequence written to the scratch pad, perform an EEPROM upload (Register 0x0E02, Bit 0). Table 28 lists a sample EEPROM download instruction sequence. It illustrates the use of condition instructions and how they alter the download sequence. The table begins with the assumption that no conditions are in effect. That is, the most recently executed condition instruction is 0xB0, or no conditional instructions have been processed. 3. Table 28. EEPROM Conditional Processing Example Instruction 0x08 0x01 0x00 0xB1 0x19 0x04 0x00 0xB2 0xB3 0x07 0x05 0x00 0x0A 0xB0 0x80 0x0A Action Transfer the system clock register contents regardless of the current condition Tag Condition 1 Transfer the clock distribution register contents only if condition = 1 Tag Condition 2 Tag Condition 3 Transfer the reference input register contents only if condition = 1, 2, or 3 Calibrate the system clock only if condition = 1, 2, or 3 Clear the condition tag board Execute an I/O update, regardless of the value of the condition Calibrate the system clock, regardless of the value of the condition Storing Multiple Device Setups in EEPROM Conditional processing makes it possible to create a number of different device setups, store them in EEPROM, and download a specific setup on demand. To do so, first program the device control registers for a specific setup. Then, store an upload sequence in the EEPROM scratch pad with the following general form: 1. 2. 3. Now reprogram the device control registers for the next desired setup. Then store a new upload sequence in the EEPROM scratch pad with the following general form: 1. 2. 4. Condition Instruction 0xB0 The next desired condition instruction (0xB1 to 0xCF, but different from the one used during the previous upload to identify a new setup) Data instructions (to save the register contents) along with any required calibrate and/or I/O update instructions Pause instruction (FE) With the upload sequence written to the scratch pad, perform an EEPROM upload (Register 0x0E02, Bit 0). Repeat the process of programming the device control registers for a new setup, storing a new upload sequence in the EEPROM scratch pad (Step 1 through Step 4) and executing an EEPROM upload (Register 0x0E02, Bit 0) until all of the desired setups are uploaded to the EEPROM. Note that, on the final upload sequence stored in the scratch pad, the pause instruction (FE) must be replaced with an end instruction (FF). To download a specific setup on demand, first store the condition associated with the desired setup in Register 0x0E01, Bits[4:0]. Then perform an EEPROM download (Register 0x0E03, Bit 1). Alternatively, to download a specific setup at power-up, apply the required logic levels necessary to encode the desired condition on the M3 to M7 multifunction pins. Then power up the device, and an automatic EEPROM download occurs. The condition (as established by the M3 to M7 multifunction pins) guides the download sequence and results in a specific setup. Keep in mind that the number of setups that can be stored in the EEPROM is limited. The EEPROM can hold a total of 2048 bytes. Each nondata instruction requires one byte of storage. Each data instruction, however, requires N + 4 bytes of storage, where N is the number of transferred register bytes. The other four bytes include the data instruction itself (one byte), the target address (two bytes), and the checksum calculated by the EEPROM controller during the upload sequence (one byte). Condition instruction (0xB1 to 0xCF) to identify the setup with a specific condition (1 to 31) Data instructions (to save the register contents) along with any required calibrate and/or I/O update instructions Pause instruction (0xFE) Rev. 0 | Page 49 of 104 AD9547 SERIAL CONTROL PORT SERIAL CONTROL ARBITER 13-BIT ADDRESS SPACE SPI READ-ONLY REGION I2C EEPROM EEPROM CONTROLLER 400kHz M7 M6 M5 M4 M3 M2 M1 M0 MULTIFUNCTION PIN CONTROL LOGIC READ/WRITE REGION ANALOG BLOCKS AND DIGITAL CORE 08300-026 SCLK/SCL CS/SDA SDIO SDO POWER-ON RESET Figure 49. Serial Port Functional Diagram The AD9547 serial control port is a flexible, synchronous serial communications port that provides a convenient interface to many industry-standard microcontrollers and microprocessors. The AD9547 serial control port is compatible with most synchronous transfer formats, including Philips I2C, Motorola® SPI, and Intel® SSR protocols. The serial control port allows read/write access to the AD9547 register map. In SPI mode, single or multiple byte transfers are supported. The SPI port configuration is programmable via Register 0x0000. This register is integrated into the SPI control logic rather than the register map and is distinct from the I2C Register 0x0000. It is also inaccessible to the EEPROM controller. A functional diagram of the serial control port, including its relationship to the EEPROM, appears in Figure 49. Although the AD9547 supports both the SPI and I2C serial port protocols, only one is active following power-up (as determined by the multifunction pins, M0 to M2, during the startup sequence). That is, the only way to change the serial port protocol is to reset the device (or cycle the device power supply). Both protocols use a common set of control pins as shown in Figure 50. 2 AD9547 SDIO 3 SDO 4 SERIAL CONTROL PORT CS/SDA 5 Figure 50. Serial Control Port Because the AD9547 supports both the SPI and I2C protocols, the active serial port protocol depends on the logic state of the three multifunction pins, M0 to M2, at startup. If all three pins are set to Logic 0 at startup, the SPI protocol is active. Otherwise, the I2C protocol is active with seven different I2C slave address settings that are based on the startup logic pattern on the M0 to M2 pins (see Table 29). Note that the four MSBs of the slave address are hardware coded as 1001. Table 29. Serial Port Mode Selection M2 0 0 0 0 1 1 1 1 M1 0 0 1 1 0 0 1 1 M0 0 1 0 1 0 1 0 1 Serial Port Mode SPI I²C (address = 1001001) I²C (address = 1001010) I²C (address = 1001011) I²C (address = 1001100) I²C (address = 1001101) I²C (address = 1001110) I²C (address = 1001111) SPI SERIAL PORT OPERATION Pin Descriptions 08300-027 SCLK/SCL SPI/I2C PORT SELECTION The SCLK (serial clock) pin (SCLK/SCL) serves as the serial shift clock. This pin is an input. SCLK synchronizes serial control port read and write operations. The rising edge SCLK registers write data bits, and the falling edge registers read data bits. The SCLK pin supports a maximum clock rate of 40 MHz. The SDIO (serial data input/output) pin is a dual-purpose pin and acts either as an input only (unidirectional mode) or as both an input and an output (bidirectional mode). The AD9547 default SPI mode is bidirectional. The SDO (serial data out) pin is useful only in unidirectional I/O mode. It serves as the data output pin for read operations. The CS (chip select) pin (CS/SDA) is an active low control that gates read and write operations. This pin is internally connected to a 30 kΩ pull-up resistor. When CS is high, the SDO and SDIO pins go into a high impedance state. Rev. 0 | Page 50 of 104 AD9547 SPI Mode Operation Write The SPI port supports both 3-wire (bidirectional) and 4-wire (unidirectional) hardware configurations and both MSB-first and LSB-first data formats. Both the hardware configuration and data format features are programmable. By default, the AD9547 uses the bidirectional MSB-first mode. The bidirectional mode is the default mode so that the user can still write to the device to switch to unidirectional mode, if it is wired for unidirectional operation. If the instruction word indicates a write operation, the payload is written into the serial control port buffer of the AD9547. Data bits are registered on the rising edge of SCLK. The length of the transfer (1, 2, or 3 bytes or streaming mode) depends on the W0 and W1 bits (see Table 30) in the instruction byte. When not streaming, CS can be deasserted after each sequence of eight bits to stall the bus (except after the last byte, where it ends the cycle). When the bus is stalled, the serial transfer resumes when CS is asserted. Deasserting the CS pin on a nonbyte boundary resets the serial control port. Reserved or blank registers are not skipped over automatically during a write sequence. Therefore, the user must know what bit pattern to write to the reserved registers to preserve proper operation of the part. Generally, it does not matter what data is written to blank registers, but it is customary to write 0s. Assertion (active low) of the CS pin initiates a write or read operation to the AD9547 SPI port. For data transfers of three bytes or fewer (excluding the instruction word), the device supports the CS stalled high mode (see Table 30). In this mode, the CS pin can be temporarily deasserted on any byte boundary, allowing time for the system controller to process the next byte. CS can be deasserted only on byte boundaries, however. This applies to both the instruction and data portions of the transfer. Table 30. Byte Transfer Count W1 0 0 1 1 W0 0 1 0 1 Bytes to Transfer 1 2 3 Streaming mode During stall high periods, the serial control port state machine enters a wait state until all data is sent. If the system controller decides to abort a transfer midstream, the state machine must be reset either by completing the transfer or by asserting the CS pin for at least one complete SCLK cycle (but less than eight SCLK cycles). Deasserting the CS pin on a nonbyte boundary terminates the serial transfer and flushes the buffer. In streaming mode (see Table 30), any number of data bytes can be transferred in a continuous stream. The register address is automatically incremented or decremented. CS must be deasserted at the end of the last byte transferred, thereby ending the stream mode. Communication Cycle—Instruction Plus Data The SPI protocol consists of a two-part communication cycle. The first part is a 16-bit instruction word that is coincident with the first 16 SCLK rising edges and a payload. The instruction word provides the AD9547 serial control port with information regarding the payload. The instruction word includes the R/W bit that indicates the direction of the payload transfer (that is, a read or write operation). The instruction word also indicates the number of bytes in the payload and the starting register address of the first payload byte. Most of the serial port registers are buffered. Refer to the Buffered/Active Registers section for details on the difference between buffered and active registers. Therefore, data written into buffered registers does not immediately take effect. An additional operation is needed to transfer buffered serial control port contents to the registers that actually control the device. This is accomplished with an I/O update operation that is performed in one of two ways: by writing a Logic 1 to Register 0x0005, Bit 0 (this bit is self-clearing) or by using an external signal via an appropriately programmed multifunction pin. The user can change as many register bits as desired before executing an I/O update. The I/O update operation transfers the buffer register contents to their active register counterparts. Read The AD9547 supports the long instruction mode only. If the instruction word indicates a read operation, the next N × 8 SCLK cycles clock out the data from the address specified in the instruction word. N is the number of data bytes read and depends on the W0 and W1 bits of the instruction word. The readback data is valid on the falling edge of SCLK. Blank registers are not skipped over during readback. A readback operation takes data from either the serial control port buffer registers or the active registers, as determined by Register 0x0004, Bit 0. Rev. 0 | Page 51 of 104 AD9547 SPI Instruction Word (16 Bits) When Register 0x0000, Bit 6 = 1 (LSB first), the instruction and data bytes must be written from LSB to MSB. Multibyte data transfers in LSB-first format start with an instruction byte that includes the register address of the least significant payload byte, followed by multiple data bytes. The serial control port internal byte address generator increments for each byte of the multibyte transfer cycle. The MSB of the 16-bit instruction word is R/W, which indicates whether the instruction is a read or a write. The next two bits, W1 and W0, indicate the number of bytes in the transfer (see Table 30). The final 13 bits are the register address (A12 to A0), which indicates the starting register address of the read/write operation (see Table 32). For multibyte MSB-first (default) I/O operations, the serial control port register address decrements from the specified starting address toward Address 0x0000. For multibyte LSB-first I/O operations, the serial control port register address increments from the starting address toward Address 0x1FFF. Unused addresses are not skipped during multibyte I/O operations; therefore, the user should write the default value to a reserved register and 0s to unmapped registers. Note that it is more efficient to issue a new write command than to write the default value to more than two consecutive reserved (or unmapped) registers. SPI MSB/LSB First Transfers The AD9547 instruction word and payload can be MSB first or LSB first. The default for the AD9547 is MSB first. The LSB-first mode can be set by writing a 1 to Register 0x0000, Bit 6. Immediately after the LSB-first bit is set, subsequent serial control port operations are LSB first. When MSB-first mode is active, the instruction and data bytes must be written from MSB to LSB. Multibyte data transfers in MSB-first format start with an instruction byte that includes the register address of the most significant payload byte. Subsequent data bytes must follow in order from high address to low address. In MSB-first mode, the serial control port internal address generator decrements for each data byte of the multibyte transfer cycle. Table 31. Streaming Mode (No Addresses Are Skipped) Write Mode LSB First MSB First Address Direction Increment Decrement Stop Sequence 0x0000 ... 0x1FFF 0x1FFF ... 0x0000 Table 32. Serial Control Port, 16-Bit Instruction Word, MSB First MSB I15 I14 I13 I12 I11 I10 I9 I8 I7 I6 I5 I4 I3 I2 I1 LSB I0 R/W W1 W0 A12 A11 A10 A9 A8 A7 A6 A5 A4 A3 A2 A1 A0 CS SCLK DON'T CARE SDIO DON'T CARE R/W W1 W0 A12 A11 A10 A9 A8 A7 A6 A5 A4 A3 A2 A1 A0 D7 D6 D5 16-BIT INSTRUCTION HEADER D4 D3 D2 D1 D0 REGISTER (N) DATA D7 D6 D5 D4 D3 D2 D1 D0 DON'T CARE REGISTER (N – 1) DATA 08300-029 DON'T CARE Figure 51. Serial Control Port Write—MSB First, 16-Bit Instruction, Two Bytes of Data CS SCLK DON'T CARE SDIO DON'T CARE R/W W1 W0 A12 A11 A10 A9 A8 A7 A6 A5 A4 A3 A2 A1 A0 SDO DON'T CARE REGISTER (N) DATA REGISTER (N – 1) DATA REGISTER (N – 2) DATA REGISTER (N – 3) DATA DON'T CARE Figure 52. Serial Control Port Read—MSB First, 16-Bit Instruction, Four Bytes of Data tDS tS CS DON'T CARE SDIO DON'T CARE tC tCLK tLOW DON'T CARE R/W W1 W0 A12 A11 A10 A9 A8 A7 A6 A5 D4 D3 D2 D1 Figure 53. Serial Control Port Write—MSB First, 16-Bit Instruction, Timing Measurements Rev. 0 | Page 52 of 104 D0 DON'T CARE 08300-031 SCLK tHIGH tDH 08300-030 D7 D6 D5 D4 D3 D2 D1 D0 D7 D6 D5 D4 D3 D2 D1 D0 D7 D6 D5 D4 D3 D2 D1 D0 D7 D6 D5 D4 D3 D2 D1 D0 16-BIT INSTRUCTION HEADER AD9547 CS SCLK DATA BIT N 08300-032 tDV SDIO SDO DATA BIT N – 1 Figure 54. Serial Control Port Timing—Read CS SCLK DON'T CARE A0 A1 A2 A3 A4 A5 A6 A7 A8 A9 A10 A11 A12 W0 W1 R/W D0 D1 D2 D3 D4 16-BIT INSTRUCTION HEADER D5 D6 REGISTER (N) DATA D7 D0 D1 D2 D3 D4 D5 D6 D7 REGISTER (N + 1) DATA Figure 55. Serial Control Port Write—LSB First, 16-Bit Instruction, Two Bytes of Data CS tS tC tCLK tHIGH tLOW tDS SCLK BIT N BIT N + 1 Figure 56. Serial Control Port Timing—Write Table 33. Serial Control Port Timing Parameter tDS tDH tCLK tS tC tHIGH tLOW tDV Description Setup time between data and the rising edge of SCLK Hold time between data and the rising edge of SCLK Period of the clock Setup time between the CS falling edge and SCLK rising edge (start of the communication cycle) Setup time between the SCLK rising edge and CS rising edge (end of the communication cycle) Minimum period that SCLK should be in a logic high state Minimum period that SCLK should be in a logic low state SCLK to valid SDIO and SDO (see Figure 54) Rev. 0 | Page 53 of 104 08300-034 tDH SDIO DON'T CARE 08300-033 SDIO DON'T CARE DON'T CARE AD9547 The transfer of data appears graphically in Figure 57. One clock pulse is generated for each data bit transferred. The data on the SDA line must be stable during the high period of the clock. The high or low state of the data line can change only when the clock signal on the SCL line is low. I²C SERIAL PORT OPERATION The I2C interface has the advantage of requiring only two control pins and is a de facto standard throughout the I2C industry. However, its disadvantage is programming speed, which is 400 kbps maximum. The AD9547 I2C port design is based on the I2C fast mode standard from Philips, so it supports both the 100 kHz standard mode and the 400 kHz fast mode. Fast mode imposes a glitch tolerance requirement on the control signals; that is, the input receivers ignore pulses of less than 50 ns duration. SDA SCL DATA LINE STABLE; DATA VALID CHANGE OF DATA ALLOWED 08300-035 The AD9547 I2C port consists of a serial data line (SDA) and a serial clock line (SCL). In an I2C bus system, the AD9547 is connected to the serial bus (data bus SDA and clock bus SCL) as a slave device; that is, no clock is generated by the AD9547. The AD9547 uses direct 16-bit memory addressing instead of traditional 8-bit memory addressing. Figure 57. Valid Bit Transfer Start/stop functionality appears graphically in Figure 58. The start condition is characterized by a high-to-low transition on the SDA line while SCL is high. The start condition is always generated by the master to initialize data transfer. The stop condition is characterized by a low-to-high transition on the SDA line while SCL is high. The stop condition is always generated by the master to terminate data transfer. The AD9547 allows for up to seven unique slave devices to occupy the I2C bus. These are accessed via a 7-bit slave address that is transmitted as part of an I2C packet. Only the device with a matching slave address responds to subsequent I2C commands. The device slave address is 1001xxx (the last three bits are determined by the M0 to M2 pins). The four MSBs (1001) are hardwired, whereas the three LSBs (xxx, determined by the M0 to M2 pins) are programmable via the power-up state of the multifunction pins (see the Initial M0 to M7 Pin Programming section). SCL S P START CONDITION I2C Bus Characteristics STOP CONDITION Figure 58. Start and Stop Condition A summary of the various I2C protocols appears in Table 34. Every byte on the SDA line must be eight bits long. Each byte must be followed by an acknowledge bit. Bytes are sent MSB first. Table 34. I2C Bus Abbreviation Definitions Abbreviation Definition S Start The acknowledge bit (A) is the ninth bit attached to any 8-bit data byte. An acknowledge bit is always generated by the receiving device (receiver) to inform the transmitter that the byte has been received. It is done by pulling the SDA line low during the ninth clock pulse after each 8-bit data byte. Sr Repeated start P Stop A Acknowledge A No acknowledge W Write R Read The no acknowledge bit (A) is the ninth bit attached to any 8-bit data byte. A no acknowledge bit is always generated by the receiving device (receiver) to inform the transmitter that the byte has not been received. It is done by leaving the SDA line high during the ninth clock pulse after each 8-bit data byte. MSB ACK FROM SLAVE RECEIVER 1 SCL 2 3 TO 7 8 9 ACK FROM SLAVE RECEIVER 1 S Figure 59. Acknowledge Bit Rev. 0 | Page 54 of 104 2 3 TO 7 8 9 10 P 08300-037 SDA 08300-036 SDA AD9547 bytes immediately after the slave address byte are the internal memory (control registers) address bytes with the high address byte first. This addressing scheme gives a memory address up to 216 − 1 = 65,535. The data bytes after these two memory address bytes are register data written into or read from the control registers. In read mode, the data bytes after the slave address byte are register data written into or read from the control registers. Data Transfer Process The master initiates data transfer by asserting a start condition. This indicates that a data stream follows. All I2C slave devices connected to the serial bus respond to the start condition. The master then sends an 8-bit address byte over the SDA line, consisting of a 7-bit slave address (MSB first) plus an R/W bit. This bit determines the direction of the data transfer, that is, whether data is written to or read from the slave device (0 = write, 1 = read). When all data bytes are read or written, stop conditions are established. In write mode, the master (transmitter) asserts a stop condition to end data transfer during the 10th clock pulse following the acknowledge bit for the last data byte from the slave device (receiver). In read mode, the master device (receiver) receives the last data byte from the slave device (transmitter) but does not pull SDA low during the ninth clock pulse. This is known as a no acknowledge bit. When receiving the no acknowledge bit, the slave device knows the data transfer is finished and enters idle mode. The master then takes the data line low during the low period before the 10th clock pulse, and high during the 10th clock pulse to assert a stop condition. The peripheral whose address corresponds to the transmitted address responds by sending an acknowledge bit. All other devices on the bus remain idle while the selected device waits for data to be read from or written to it. If the R/W bit = 0, the master (transmitter) writes to the slave device (receiver). If the R/W bit = 1, the master (receiver) reads from the slave device (transmitter). See the Data Transfer Format section for the command format. Data is then sent over the serial bus in the format of nine clock pulses, one data byte (eight bits) from either master (write mode) or slave (read mode) followed by an acknowledge bit from the receiving device. The number of bytes that can be transmitted per transfer is unrestricted. In write mode, the first two data A start condition can be used in place of a stop condition. Furthermore, a start or stop condition can occur at any time, and partially transferred bytes are discarded. MSB ACK FROM SLAVE RECEIVER 1 SCL 3 TO 7 2 8 9 ACK FROM SLAVE RECEIVER 1 3 TO 7 2 8 9 10 S P 08300-038 SDA Figure 60. Data Transfer Process (Master Write Mode, 2-Byte Transfer Used for Illustration) SDA ACK FROM MASTER RECEIVER 1 3 TO 7 2 8 9 1 3 TO 7 2 8 9 10 S P 08300-039 SCL NO ACK FROM MASTER RECEIVER Figure 61. Data Transfer Process (Master Read Mode, 2-Byte Transfer Used for Illustration) Data Transfer Format In write byte format, the write byte protocol is used to write a register address to the RAM starting from the specified RAM address. S Slave Address A W RAM Address High Byte A RAM Address Low Byte A RAM Data 0 A RAM Data 1 A RAM Data 2 A P In send byte format, the send byte protocol is used to set up the register address for subsequent reads. S Slave Address W A RAM Address High Byte A RAM Address Low Byte A P In receive byte format, the receive byte protocol is used to read the data bytes from RAM starting from the current address. S Slave Address R A RAM Data 0 A RAM Data 1 A RAM Data 2 P A Read byte format combines the format of the send byte and the receive byte formats. S Slave Address W A RAM Address High Byte A RAM Address Low Byte A Sr Slave Address Rev. 0 | Page 55 of 104 R A RAM Data 0 A RAM Data 1 A RAM Data 2 A P AD9547 I²C Serial Port Timing SDA tLOW tF tR tSU;DAT tHD;STA tSP tBUF tR tF tHD;STA S tHD;DAT tHIGH tSU;STO tSU;STA Sr Figure 62. I²C Serial Port Timing Table 35. I2C Timing Definitions Parameter fSCL tBUF tHD;STA tSU;STA tSU;STO tHD;DAT tSU;DAT tLOW tHIGH tR tF tSP Description Serial clock Bus free time between stop and start conditions Repeated hold time start condition Repeated start condition setup time Stop condition setup time Data hold time Data setup time SCK clock low period SCK clock high period Minimum/maximum receive SCL and SDA rise time Minimum/maximum receive SCL and SDA fall time Pulse width of voltage spikes that must be suppressed by the input filter Rev. 0 | Page 56 of 104 P S 08300-040 SCL AD9547 BUFFERED/ACTIVE REGISTERS There are two broad categories of registers on the AD9547: buffered and active (see Figure 63). Buffered registers are those that can be written to directly from the serial port. They do not need an I/O update to apply their contents to the internal device functions. In contrast, active registers require an I/O update to transfer data between the buffered registers and the internal device functions. In operation, the user programs as many buffered registers as desired and then issues an I/O update. The I/O update is performed by writing to Register 0x0005, Bit 0 = 1 (or by the external application of the necessary logic level to one of the multifunction pins previously programmed as an I/O update input). The contents of the buffered registers that are connected directly to the internal device functions affect those functions immediately. The contents of buffered registers that connect to active registers do not affect the internal device functions until the I/O update event occurs. An S or C in the Opt column of the register map identifies an active register (otherwise, it is a buffered register). An S entry means that the I/O update signal to the active register is synchronized with the serial port clock or with an input signal driving one of the multifunction pins. On the other hand, a C entry means that the I/O update signal to the active register is synchronized with a clock signal derived from the internal system clock (fS/32), as shown in Figure 63. When reading back a register that has both buffered and active contents, Register 0x0004, Bit 0 can be used to select whether to read back the buffered or active contents. Readback of the active contents occurs when Register 0x0004, Bit 0 = 0, whereas readback of the buffered contents occurs when Register 0x0004, Bit 0 = 1. Note that a read-only active register requires an I/O update before its contents can be read. EDGE DETECT 5 SDIO 3 SDO 4 SCLK/SCL 2 08300-041 CS/SDA TO INTERNAL DEVICE FUNCTIONS SERIAL CONTROL PORT ACTIVE C REGISTERS fS/32 ACTIVE S REGISTERS In general, when a group of registers defines a control parameter, the LSB of the value resides in the D0 position of the register with the lowest address. The bit weight increases from right to left, from the lowest register address to the highest register address. For example, the default value of the incremental phase lock offset step size register (Address 0x0314 to Address 0x0315) is the 16-bit hexadecimal number, 0x03E8 (not 0xE803). Note that the EEPROM storage sequence registers (Address 0x0E10 to Address 0x0E3F) are an exception to this convention (see the EEPROM Instructions section). FROM MULTIFUNCTION PIN LOGIC BUFFERED REGISTERS The register map spans an address range from 0x0000 through 0x0E3F (0 to 3647, decimal). Each address provides access to one byte (eight bits) of data. Each individual register is identified by its four-digit hexadecimal address (for example, Register 0x0A10). In some cases, a group of addresses collectively define a register (for example, the IRQ mask register consists of Register 0x0209, Register 0x020A, Register 0x020B, Register 0x020C, Register 0x020D, Register 0x020E, Register 0x020F, and Register 0x0210). I/O UPDATE I/O PROGRAMMING REGISTERS Figure 63. Buffered and Active Registers AUTOCLEARING REGISTERS An A in the Opt column of the register map identifies an autoclearing register. Typically, the active value for an autoclearing register takes effect following an I/O update. The bit is cleared by the internal device logic upon completion of the prescribed action. REGISTER ACCESS RESTRICTIONS Read and write access to the registers may be restricted, depending on the register in question, the source and direction of access, and the current state of the device. Each register can be classified into one or more access types. When more than one type applies, the most restrictive condition that applies at that time is used. When access is denied to a register, all attempts to read the register return a 0 byte, and all attempts to write to the register are ignored. Access to nonexistent registers is handled in the same way as for a denied register. Regular Access Registers with regular access do not fall into any other category. Both read and write access to registers of this type can be from the serial port or the EEPROM controller. However, only one of these sources can have access to a register at any given time (access is mutually exclusive). When the EEPROM controller is active, either in load or store mode, it has exclusive access to the registers. Read-Only Access An R in the Opt column of the register map identifies read-only registers. Access is available at all times, including when the EEPROM controller is active. Exclusion from EEPROM Access An E in the Opt column of the register map identifies a register with contents that are inaccessible to the EEPROM. That is, the contents of this type of register cannot be transferred directly to the EEPROM or vice versa. Note that read-only registers (R) are inaccessible to the EEPROM, as well. Rev. 0 | Page 57 of 104 AD9547 REGISTER MAP The register addresses and defaults are hexadecimal values. Use the default value when writing to registers and/or bits marked as unused. Table 36. Addr Opt 1 Name D7 D6 Serial Port Configuration and Part Identification 0x0000 E SPI control UnidirecLSB first/ tional IncAddr 0x0000 Dup I2C control Unused 0x0001 E Reserved 0x0002 R Silicon revision level 0x0003 R Device ID 0x0004 E Register readback 0x0005 A, E I/O update System Clock (SYSCLK) 0x0100 S Charge pump/ External Charge lock detect loop filter pump control enable mode (auto/man) 0x0101 S N divider 0x0102 S Unused System clock M divider input options reset 0x0103 C Nominal system clock 0x0104 C period 0x0105 C 0x0106 C System clock stability 0x0107 C period 0x0108 C General Configuration 0x0200 S M0 control 0x0201 S M1 control 0x0202 S M2 control 0x0203 S M3 control 0x0204 S M4 control 0x0205 S M5 control 0x0206 S M6 control 0x0207 S M7 control 0x0208 C IRQ pin output mode 0x0209 C IRQ mask D5 D4 Soft reset Long instruction D3 D2 D1 Unused Soft reset Unused 0x00 0x01 Device ID Unused Read buffer register I/O update Unused Charge pump current[2:0] Lock detect timer disable N divider[7:0] PLL enable 2× frequency multiplier enable Nominal SYSCLK period[15:0] (in fs) (1 ns at 1 ppm accuracy) M divider[1:0] SYSCLK source[1:0] 0x28 0x45 M0 function[6:0] M1 function[6:0] M2 function[6:0] M3 function[6:0] M4 function[6:0] M5 function[6:0] M6 function[6:0] M7 function[6:0] Unused Unused SYSCLK unlocked Unused 0x020A C 0x020B C 0x020C C 0x020D C Ref AA new profile Ref AA validated Ref AA fault cleared 0x020E C Ref BB new profile Ref BB validated Ref BB fault cleared 0x020F 0x0210 0x0211 0x0212 0x0213 0x0214 C C C C S S Closed Free run Unused SYSCLK locked Holdover History updated Ref AA fault Ref BB fault IRQ pin output mode[1:0] Unused Distribution sync Frequency unlocked Frequency unclamped Ref A new profile Watchdog timer Frequency locked Frequency clamped Ref A validated Ref B new profile Ref B validated SYSCLK cal complete EEPROM fault Phase unlocked Phase slew unlimited Ref A fault cleared Ref B fault cleared DAC current DAC shutdown DAC full-scale current[7:0] Unused Rev. 0 | Page 58 of 104 0x40 0x42 0x0F 0x01 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 SYSCLK cal started EEPROM complete Phase locked Phase slew limited Ref A fault 0x00 Ref B fault 0x00 Unused Unused Watchdog timer [15:0] (in ms up to 65,535 ms) Watchdog timer 0x00 0x18 System clock stability period[19:16] (in ms) M0 in/out M1 in/out M2 in/out M3 in/out M4 in/out M5 in/out M6 in/out M7 in/out 0x31 0x00 Lock detect timer[1:0] Nominal SYSCLK period[20:16] SYSCLK stability period[15:0] (in ms) Unused Def 0x10 Unused Silicon revision number Unused Switching D0 DAC full-scale current[9:8] 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0xFF 0x01 AD9547 Addr DPLL 0x0300 0x0301 0x0302 0x0303 0x0304 0x0305 0x0306 0x0307 0x0308 0x0309 0x030A 0x030B 0x030C 0x030D 0x030E 0x030F 0x0310 0x0311 0x0312 0x0313 0x0314 0x0315 Opt 1 Name C C C C C C A, C C C C C C C C C C C C C C C C Free-running frequency tuning word Free-running frequency tuning word[47:0] Update TW Pull-in range lower limit Unused Pull-in range lower limit[23:0] Pull-in range upper limit Pull-in range upper limit[23:0] Open-loop phase offset Open-loop phase offset word[15:0] Fixed closedloop phase lock offset Fixed phase lock offset[39:0] (in ps; signed) Incremental phase lock offset step size[15:0] (in ps/step) 0x0316 0x0317 0x0318 0x0319 0x031A 0x031B C C C C C C Incremental closed-loop phase lock offset step size Phase slew rate limit History accumulation timer History accumulation timer[23:0] (in ms) D7 History mode Clock Distribution Output Configuration 0x0400 S Distribution settings 0x0401 S 0x0402 S 0x0403 C 0x0404 S 0x0405 S 0x0406 0x0407 0x0408 0x0409 0x040A 0x040B 0x040C 0x040D 0x040E 0x040F S S S S S S S S S S Distribution enable Distribution synchronization Automatic synchronization Distribution channel modes D6 D5 D4 D3 D2 D1 D0 Update TW Phase slew rate limit[15:0] (in ns/sec) Unused Unused Unused Single sample fallback Persistent history Unused External Receiver distribution mode resistor Unused Sync source[1:0] Unused Unused Unused Unused Unused OUT0 CMOS phase invert OUT1 CMOS phase invert OUT0 polarity invert OUT1 polarity invert OUT0 drive strength OUT1 drive strength Unused Q0[23:0] Distribution channel divider, Q0 Unused Q0[29:24] Q1[23:0] Distribution channel divider, Q1 Unused Q1[29:24] Rev. 0 | Page 59 of 104 Incremental average[2:0] Def 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0xFF 0xFF 0xFF 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0xE8 0x03 0x00 0x00 0x30 0x75 0x00 0x00 0x0C OUT1 OUT0 powerpowerdown down OUT1 OUT0 enable enable OUT1 OUT0 sync mask sync mask Automatic sync mode[1:0] 0x00 OUT0 mode 0x03 OUT1 mode 0x03 0x00 0x00 0x03 0x03 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 AD9547 Addr Opt 1 Name 0x0410 S Reserved 0x0411 S 0x0412 S 0x0413 S 0x0414 S 0x0415 S 0x0416 S 0x0417 S Reference Input Configuration 0x0500 S Reference power-down 0x0501 0x0502 0x0503 S S C 0x0504 C 0x0505 0x0506 0x0507 C C C Reference logic family Manual reference profile selection Phase buildout switching Profile Registers—Profile 0 0x0600 Priorities 0x0601 Reference period 0x0602 0x0603 0x0604 0x0605 0x0606 0x0607 0x0608 Tolerance 0x0609 0x060A 0x060B 0x060C 0x060D 0x060E Validation timer 0x060F 0x0610 Redetect timer 0x0611 0x0612 0x0613 0x0614 0x0615 0x0616 0x0617 0x0618 0x0619 0x061A 0x061B 0x061C 0x061D D7 D6 D5 D4 D3 Unused D2 D1 D0 Unused Ref BB Ref B Ref AA Ref A powerpowerpowerpowerdown down down down Ref BB logic family[1:0] Ref B logic family[1:0] Ref AA logic family[1:0] Ref A logic family[1:0] Unused Ref AA manual profile[2:0] Ref A manual profile[2:0] Enable Enable Ref A Ref AA manual manual profile profile Ref BB manual profile[2:0] Ref B manual profile[2:0] Enable Enable Ref B Ref BB manual manual profile profile Unused Unused Unused Phase master threshold priority[2:0] Unused Promoted priority[2:0] Nominal reference period[39:0] (in fs up to 1.1 ms) Selection priority[2:0] Unused (write 0s to these bits) Inner tolerance[15:0] (1/tolerance) (removes fault status; 10% down to 1 ppm) Unused Inner tolerance[19:16] Outer tolerance[15:0] (1/tolerance) (indicates fault status; 10% down to 1 ppm) Unused Outer tolerance[19:16] Validation timer[15:0] (in ms up to 65,535 ms) Redetect timer[15:0] (in ms up to 65,535 ms) Alpha-0 linear[15:0] Digital loop filter coefficients Alpha-2 exponent[1:0] Alpha-1 exponent[5:0] Beta-0 linear[6:0] Alpha-2 exponent[2] Beta-0 linear[14:7] Beta-1 exponent[4:0] Gamma-0 linear[15:0] Unused Unused Beta-0 linear[16:15] Gamma-1 exponent[4:0] Delta-0 linear[7:0] Delta-0 linear[14:8] Delta-1 exponent[0] Alpha-3 exponent[3:0] Rev. 0 | Page 60 of 104 Gamma-0 linear[16] Def 0xF0 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 Delta-1 exponent[4:1] 0x00 AD9547 Addr Opt 1 Name 0x061E R divider 0x061F 0x0620 0x0621 0x0622 S divider 0x0623 0x0624 0x0625 0x0626 Fractional feedback 0x0627 divider 0x0628 0x0629 Lock detectors 0x062A 0x062B 0x062C 0x062D 0x062E 0x062F 0x0630 0x0631 Profile Registers—Profile 1 0x0632 Priorities 0x0633 Reference period 0x0634 0x0635 0x0636 0x0637 0x0638 0x0639 0x063A 0x063B 0x063C 0x06CD 0x063E 0x063F 0x0640 0x0641 0x0642 0x0643 0x0644 0x0645 0x0646 0x0647 0x0648 0x0649 0x064A 0x064B 0x064C 0x064D 0x064E 0x064F D7 D6 D5 D4 D3 R[23:0] Unused D2 D1 D0 R[29:24] S[15:0] Unused S[19:16] Unused V[7:0] U[3:0] Unused U[9:4] Phase lock threshold[15:0] (in ps) Unused V[9:8] Phase lock fill rate[7:0] Phase lock drain rate[7:0] Frequency lock threshold[23:0] (in ps) Frequency lock fill rate[7:0] Frequency lock drain rate[7:0] Unused Promoted priority[2:0] Selection priority[2:0] Nominal reference period[39:0] (in units of fs up to 1.1 ms) Unused (write 0s to these bits) Tolerance Inner tolerance[15:0] (1/tolerance) (removes fault status; 10% down to 1 ppm) Unused Inner tolerance[19:16] Outer tolerance[15:0] (1/tolerance) (indicates fault status; 10% down to 1 ppm) Unused Validation timer Redetect timer[15:0] (in units of ms up to 65.5 sec) Redetect timer Digital loop filter coefficients Outer tolerance[19:16] Validation timer[15:0] (in units of ms up to 65.5 sec) Alpha-0 linear[15:0] Alpha-2 exponent[1:0] Alpha-1 exponent[5:0] Beta-0 linear[6:0] Beta-0 linear[14:7] Beta-1 exponent[4:0] Gamma-0 linear[15:0] Unused Unused Alpha-2 exponent[2] Beta-0 linear[16:15] Gamma-1 exponent[4:0] Gamma-0 linear[16] Delta-0 linear[7:0] Delta-0 linear[14:8] Delta-1 exponent[0] Alpha-3 exponent[3:0] Rev. 0 | Page 61 of 104 Def 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 Delta-1 exponent[4:1] 0x00 AD9547 Addr Opt 1 Name 0x0650 R divider 0x0651 0x0652 0x0653 0x0654 S divider 0x0655 0x0656 0x0657 0x0658 Fractional feedback 0x0659 divider 0x065A 0x065B Lock detectors 0x065C 0x065D 0x065E 0x065F 0x0660 0x0661 0x0662 0x0663 Unused 0x0664 to 0x067F Profile Registers—Profile 2 0x0680 Priorities 0x0681 Reference period 0x0682 0x0683 0x0684 0x0685 0x0686 0x0687 0x0688 0x0689 0x068A 0x068B 0x068C 0x068D 0x068E 0x068F 0x0690 0x0691 0x0692 0x0693 0x0694 0x0695 0x0696 0x0697 0x0698 0x0699 0x069A 0x069B 0x069C 0x069D D7 D6 D5 D4 D3 R[23:0] Unused D2 D1 D0 R[29:24] S[15:0] Unused S[19:16] Unused V[7:0] U[3:0] Unused Unused U[9:4] Phase lock threshold[15:0] (in units of ps) V[9:8] Phase lock fill rate[7:0] Phase lock drain rate[7:0] Frequency lock threshold[23:0] (in ps) Frequency lock fill rate[7:0] Frequency lock drain rate[7:0] Unused Unused Promoted priority[2:0] Nominal period[39:0] (in units of fs up to 1.1 ms) Selection priority[2:0] Unused (write 0s to these bits) Tolerance Inner tolerance[15:0] (1/tolerance) (removes fault status; 10% down to 1 ppm) Unused Inner tolerance[19:16] Outer tolerance[15:0] (1/tolerance) (indicates fault status; 10% down to 1 ppm) Unused Outer tolerance [19:16] Validation timer[15:0] (ms) (up to 65,535 ms) Validation timer Redetect timer[15:0] (ms) (up to 65,535 ms) Redetect timer Digital loop filter coefficients Alpha-0 linear[15:0] Alpha-2 exponent[1:0] Alpha-1 exponent[5:0] Beta-0 linear[6:0] Beta-0 linear[14:7] Beta-1 exponent[4:0] Gamma-0 linear[15:0] Unused Unused Alpha-2 exponent[2] Beta-0 linear[16:15] Gamma-1 exponent[4:0] Gamma-0 linear[16] Delta-0 linear[7:0] Delta-0 linear[14:8] Delta-1 exponent[0] Alpha-3 exponent[3:0] Rev. 0 | Page 62 of 104 Def 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 Delta-1 exponent[4:1] 0x00 AD9547 Addr Opt 1 Name 0x069E R divider 0x069F 0x06A0 0x06A1 0x06A2 S divider 0x06A3 0x06A4 0x06A5 0x06A6 Fractional feedback 0x06A7 divider 0x06A8 0x06A9 Lock detectors 0x06AA 0x06AB 0x06AC 0x06AD 0x06AE 0x06AF 0x06B0 0x06B1 Profile Registers—Profile 3 0x06B2 Priorities 0x06B3 Reference period 0x06B4 0x06B5 0x06B6 0x06B7 0x06B8 0x06B9 0x06BA 0x06BB 0x06BC 0x06BD 0x06BE 0x06BF 0x06C0 0x06C1 0x06C2 0x06C3 0x06C4 0x06C5 0x06C6 0x06C7 D6 D5 D4 D3 R[23:0] Unused D2 D1 D0 R[29:24] S[15:0] Unused S[19:16] Unused V[7:0] U[3:0] Unused U[9:4] Phase lock threshold[15:0] (in ps) Unused V[9:8] Phase lock fill rate[7:0] Phase lock drain rate[7:0] Frequency lock threshold[23:0] (in ps) Frequency lock fill rate[7:0] Frequency lock drain rate[7:0] Unused Promoted priority[2:0] Nominal period[39:0] (in fs up to 1.1 ms) Selection priority[2:0] Unused (write 0s to these bits) Tolerance Inner tolerance[15:0] (1/tolerance) (removes fault status; 10% down to 1 ppm) Unused Inner tolerance[19:16] Outer tolerance[15:0] (1/tolerance) (indicates fault status; 10% down to 1 ppm) Unused Validation timer Digital loop filter coefficients Outer tolerance[19:16] Validation timer[15:0] (in ms up to 65,535 ms) Redetect timer[15:0] (in ms up to 65,535 ms) Redetect timer Alpha-0 linear[15:0] Alpha-2 exponent[1:0] Alpha-1 exponent[5:0] Beta-0 linear[6:0] 0x06C8 0x06C9 0x06CA 0x06CB 0x06CC Alpha-2 exponent[2] Beta-0 linear[14:7] Beta-1 exponent[4:0] Gamma-0 linear[15:0] Unused Unused 0x06CD 0x06CE 0x06CF 0x06D0 0x06D1 0x06D2 0x06D3 D7 Beta-0 linear [16:15] Gamma-1 exponent[4:0] Delta-0 linear[7:0] Delta-0 linear[14:8] Delta-1 exponent[0] Alpha-3 exponent[3:0] R divider Gamma-0 linear[16] Unused R[29:24] Rev. 0 | Page 63 of 104 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 Delta-1 exponent[4:1] R[23:0] Def 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 AD9547 Addr Opt 1 Name D7 D6 D5 D4 D3 D2 D1 D0 0x06D4 S divider S[15:0] 0x06D5 0x06D6 Unused S[19:16] 0x06D7 Unused 0x06D8 V[7:0] Fractional feedback 0x06D9 U[3:0] Unused V[9:8] divider 0x06DA Unused U[9:4] 0x06DB Phase lock threshold[15:0] (in ps) Lock detectors 0x06DC Phase lock fill rate[7:0] 0x06D D 0x06DE Phase lock drain rate[7:0] 0x06DF Frequency lock threshold[23:0] (in ps) 0x06E0 0x06E1 0x06E2 Frequency lock fill rate[7:0] 0x06E3 Frequency lock drain rate[7:0] Unused 0x06E4 to 0x06FF Profile Registers—Profile 4 through Profile 7 0x0700 Profile 4 The functionality of the Profile 4 through Profile 7 address locations (Address 0x0700 to Address 0x07FF) is identical to to through that of the Profile 0 through Profile 3 address locations (Address 0x0600 to Address 0x06FF) 0x07FF Profile 7 Operational Controls 0x0A00 S Unused General Reset sans SYSCLK Reference TDC DAC Dist Full power-down regmap powerpowerpowerpowerpowerpowerdown down down down down down 0x0A01 C Loop mode Unused User selection mode[1:0] User User Unused User reference holdover free run (write a 0 selection[1:0] to this bit) 0x0A02 S Cal/sync Unused Sync Calibrate distribution SYSCLK 0x0A03 A, C Unused Clear LF Clear CCI Reset Clear phase Reset auto Reset Reset Reset functions accumulator sync TW history all IRQs watchdog 0x0A04 A, C IRQ clearing Unused Unused SYSCLK SYSCLK SYSCLK cal SYSCLK cal unlocked locked complete started 0x0A05 A, C Unused DistribuWatchdog EEPROM EEPROM tion sync timer fault complete 0x0A06 A, C Switching Closed Free run Holdover Frequency Frequency Phase Phase unlocked locked unlocked locked 0x0A07 A, C Unused History Frequency Frequency Phase slew Phase slew updated unclamped clamped unlimited limited 0x0A08 A, C Ref AA Ref AA Ref AA fault Ref AA fault Ref A Ref A Ref A Ref A new profile validated cleared new profile validated fault cleared fault 0x0A09 A, C Ref BB Ref BB Ref BB Ref BB Ref B Ref B Ref B Ref B new profile validated fault cleared fault new profile validated fault cleared fault 0x0A0A A, C Unused 0x0A0B A, C Unused 0x0A0C A, C Unused Incremental Reset Decrement Increment phase offset phase offset phase offset phase offset 0x0A0D A, C Unused Detect BB Detect B Detect AA Detect A Reference profile selection state machine startup 0x0A0E A, C Unused Force Force Force Force Force validation Timeout BB Timeout B Timeout AA Timeout A timeout 0x0A0F C Unused Reference Ref Mon Ref Mon Ref Mon Ref Mon monitor Override BB Override B Override AA Override A override 0x0A10 C Unused Reference Ref Mon Ref Mon Ref Mon Ref Mon monitor Bypass BB Bypass B Bypass AA Bypass A bypass Rev. 0 | Page 64 of 104 Def 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 AD9547 Addr Opt 1 Name D7 D6 D5 D4 Status Readback (These registers are read only and are accessible during EEPROM transactions.) 0x0D00 R EEPROM Unused 0x0D01 R SYSCLK 0x0D02 R IRQ monitor 0x0D03 R 0x0D04 R 0x0D05 R 0x0D06 R 0x0D07 R 0x0D08 0x0D09 0x0D0A R R R, C 0x0D0B R, C 0x0D0C R, C 0x0D0 D R, C 0x0D0E R, C 0x0D0F R, C Unused Unused Switching SYSCLK unlocked Unused Closed Ref A input reference status Ref AA input reference status Ref B input reference status Ref BB input reference status Unused E A, E Condition Save 0x0E03 A, E Load EEPROM Storage Sequence 0x0E10 E SYSCLK settings 0x0E11 E 0x0E12 E 0x0E13 E I/O update 0x0E14 E SYSCLK calibration 0x0E15 E General configuration 0x0E16 E settings 0x0E17 E 0x0E18 E DPLL settings Distribution sync Holdover Frequency unlocked History Frequency updated unclamped Ref AA fault Ref A new profile Ref BB Ref B fault new profile Unused Ref AA new profile Ref BB new profile Ref AA validated Ref BB validated Ref AA fault cleared Ref BB fault cleared Offset slew limiting Frequency clamped Profile selected Phase build-out History available Phase lock Frequency Loop lock switching Active reference priority[2:0] D0 Watchdog timer Frequency locked Frequency clamped Ref A validated Ref B validated Load in progress Cal in progress SYSCLK cal complete EEPROM fault Phase unlocked Phase slew unlimited Ref A fault cleared Ref B fault cleared Save in progress Lock detected SYSCLK cal started EEPROM complete Phase locked Phase slew limited Ref A fault Ref B fault Holdover Active Unused Def Free running Active reference[1:0] Selected profile[2:0] Valid Fault Fast Slow Profile selected Selected profile[2:0] Valid Fault Fast Slow Profile selected Selected profile[2:0] Valid Fault Fast Slow Profile selected Selected profile[2:0] Valid Fault Fast Slow 0x0D10 R, C 0x0D11 R, C 0x0D12 R, C 0x0D13 R, C 0x0D14 R, C Holdover history 0x0D15 R, C 0x0D16 R, C 0x0D17 R, C 0x0D18 R, C 0x0D19 R, C Nonvolatile Memory (EEPROM) Control 0x0E00 Write protect 0x0E01 0x0E02 D1 Unused SYSCLK locked Free run D2 Fault detected Unused Stable Unused DPLL status D3 Unused Tuning word history[47:0] Unused Half rate mode Condition value[4:0] Unused Unused Unused Load from EEPROM Data: 9 bytes Address: 0x0100 Action: I/O update Action: calibrate system clock Data: 21 bytes Address: 0x0200 Data: 28 bytes Rev. 0 | Page 65 of 104 Write enable Save to EEPROM Unused 0x00 0x00 0x00 0x00 0x08 0x01 0x00 0x80 0xA0 0x14 0x02 0x00 0x1B AD9547 Addr 0x0E19 0x0E1A 0x0E1B 0x0E1C 0x0E1D 0x0E1E 0x0E1F 0x0E20 0x0E21 0x0E22 0x0E23 0x0E24 0x0E25 0x0E26 0x0E27 0x0E28 0x0E29 0x0E2A 0x0E2B 0x0E2C 0x0E2D 0x0E2E 0x0E2F 0x0E30 0x0E31 0x0E32 0x0E33 0x0E34 to 0x0E3F 1 Opt 1 E E E E E E E E E E E E E E E E E E E E E E E E E E E E Name D7 D6 D5 D4 D3 Address: 0x0300 Data: 26 bytes Address: 0x0400 Clock distribution settings I/O update Reference input settings Action: I/O update Data: 8 bytes Address: 0x0500 Profile 0 and Profile 1 settings Data: 100 bytes Address: 0x0600 Profile2 and Profile 3 settings Data: 100 bytes Address: 0x0680 Profile 4 and Profile 5 settings Data: 100 bytes Address: 0x0700 Profile 6 and Profile 7 settings Data: 100 bytes Address: 0x0780 I/O update Operational control settings Action: I/O update Data: 17 bytes Address: 0x0A00 I/O update End of data Action: I/O update Action: end of data Continuation of scratch pad area See the I/O Programming Registers section for an explanation of the Opt column. Rev. 0 | Page 66 of 104 D2 D1 D0 Def 0x03 0x00 0x19 0x04 0x00 0x80 0x07 0x05 0x00 0x63 0x06 0x00 0x63 0x06 0x80 0x63 0x07 0x00 0x63 0x07 0x80 0x80 0x10 0x0A 0x00 0x80 0xFF AD9547 REGISTER BIT DESCRIPTIONS SERIAL PORT CONFIGURATION AND PART IDENTIFICATION (REGISTER 0x0000 TO REGISTER 0x0005) Table 37. SPI Control/I2C Control Address 0x0000 Bit 7 Bit Name Unidirectional 6 LSB first/IncAddr 5 Soft reset 4 Long instruction [3:0] Unused Description Select SPI port SDO pin operating mode. 0 (default) = 3-wire. 1 = 4-wire (SDO pin enabled). Bit order for SPI port. 0 (default) = most significant bit and byte first (multibyte transfers use incrementing address). 1 = least significant bit and byte first (multibyte transfers use decrementing address). Device reset (invokes an EEPROM download if M[7:3] ≠ 0). 0 (default) = normal operation. 1 = reset. 16-bit mode (the only mode supported by the device). This bit is read only and reads back as Logic 1. Unused. Table 38. Reserved Register Address 0x0001 Bit [7:0] Bit Name Unused Description Unused. Table 39. Silicon Revision Level (Read Only) Address 0x0002 Bit [7:0] Bit Name Silicon revision number Description Default = 0x31 = 0b00110001. Table 40. Device ID (Read Only) Address 0x0003 Bit [7:0] Bit Name Device ID Description Default = 0x48 = 0b01001000. Table 41. Register Readback Control Address 0x0004 Bit [7:1] 0 Bit Name Unused Read buffered register Description Unused. For buffered registers, serial port readback reads from actual (active) registers instead of from the buffer. 0 (default) = reads values currently applied to the device’s internal logic. 1 = reads buffered values that take effect on the next assertion of the I/O update. Table 42. Soft I/O Update Address 0x0005 Bit [7:1] 0 Bit Name Unused I/O update Description Unused. Writing a 1 to this bit transfers the data in the serial I/O buffer registers to the device’s internal control registers. This is an autoclearing bit. Rev. 0 | Page 67 of 104 AD9547 SYSTEM CLOCK (SYSCLK) (REGISTER 0x0100 TO REGISTER 0x0108) Table 43. Charge Pump and Lock Detect Control Address 0x0100 Bit 7 Bit Name External loop filter enable 6 Charge pump mode [5:3] Charge pump current 2 Lock detect timer disable [1:0] Lock detect timer Description Enables use of an external SYSCLK PLL loop filter. 0 (default) = internal loop filter. 1 = external loop filter. Charge pump current control. 0 (default) = automatic. 1 = manual. Selects charge pump current when Bit 6 = 1. 000 = 125 μA. 001 = 250 μA. 010 = 375 μA. 011 (default) = 500 μA. 100 = 625 μA. 101 = 750 μA. 110 = 875 μA. 111 = 1000 μA. Enable the SYSCLK PLL lock detect timer. 0 (default) = enable. 1 = disable. Select lock detect timer depth. 00 (default) = 128. 01 = 256. 10 = 512. 11 = 1024. Table 44. N Divider Address 0x0101 Bit [7:0] Bit Name N divider Description System clock PLL feedback divider value: 6 ≤ N ≤ 255 (default = 0x28 = 40). Table 45. System Clock Input Options Address 0x0102 Bit 7 6 Bit Name Unused M divider reset [5:4] M divider 3 2× frequency multiplier enable 2 PLL enable [1:0] SYSCLK source Description Unused. Reset the M divider. 0 = normal operation. 1 (default) = reset. When not using the M divider, program this bit to Logic 1. System clock input divider. 00 (default) = 1. 01 = 2. 10 = 4. 11 = 8. Enable the 2× frequency multiplier. 0 (default) = disable. 1 = enable. Enable the SYSCLK PLL. 0 = disable. 1 (default) = enable. Input mode select for SYSCLKx pins. 00 = crystal resonator. 01 (default) = low frequency clock source. 10 = high frequency (direct) clock source. 11 = input receiver power-down. Rev. 0 | Page 68 of 104 AD9547 Table 46. Nominal System Clock (SYSCLK) Period1 Address 0x0103 0x0104 0x0105 1 Bit [7:0] [7:0] [7:5] [4:0] Bit Name Nominal SYSCLK period (expressed in fs) Unused Nominal SYSCLK period Description System clock period, Bits[7:0]. System clock period, Bits[15:8]. Unused. System clock period, Bits[20:16]. Units are femtoseconds (fs). The default value is 0x0F424 = 1,000,000 (1 ns) and implies a system clock frequency of 1 GHz. Table 47. System Clock Stability Period1 Address 0x0106 0x0107 0x0108 1 Bit [7:0] [7:0] [7:4] [3:0] Bit Name SYSCLK stability period (expressed in ms) Unused SYSCLK stability period Description System clock stability period, Bits[7:0] (default = 0x01). System clock stability period, Bits[15:8] (default = 0x00). Unused. System clock stability period, Bits[19:16] (default = 0x0) (default period = 0x00001, or 1 ms). Units are milliseconds (ms). The default value is 0x00001 = 1 (1 ms). GENERAL CONFIGURATION (REGISTER 0x0200 TO REGISTER 0x0214) Register 0x0200 to Register 0x0207—Multifunction Pin Control (M0 to M7) Table 48. Multifunction Pin (M0 to M7) Control1 Address 0x0200 0x0201 0x0202 0x0203 0x0204 0x0205 0x0206 0x0207 1 Bit 7 Bit Name M0 in/out [6:0] 7 [6:0] 7 [6:0] 7 [6:0] 7 [6:0] 7 [6:0] 7 [6:0] 7 [6:0] M0 function M1 in/out M1 function M2 in/out M2 function M3 in/out M3 function M4 in/out M4 function M5 in/out M5 function M6 in/out M6 function M7 in/out M7 function Description In/out control for the M0 pin. 0 (default) = input (control pin). 1 = output (status pin). See Table 24 and Table 25 (default = 0xb0000000). In/out control for the M1 pin (same as M0). See Table 24 and Table 25 (default = 0xb0000000). In/out control for the M2 pin (same as M0). See Table 24 and Table 25 (default = 0xb0000000). In/out control for the M3 pin (same as M0). See Table 24 and Table 25 (default = 0xb0000000). In/out control for the M4 pin (same as M0). See Table 24 and Table 25 (default = 0xb0000000). In/out control for the M5 pin (same as M0). See Table 24 and Table 25 (default = 0xb0000000). In/out control for the M6 pin (same as M0). See Table 24 and Table 25 (default = 0xb0000000). In/out control for the M7 pin (same as M0). See Table 24 and Table 25 (default = 0xb0000000). The default setting for all the multifunction pins is as an unused control input pin. Table 49. IRQ Pin Output Mode Address 0x0208 Bit [7:2] [1:0] Bit Name Unused IRQ pin output mode Description Unused. Select the output mode of the IRQ pin. 00 (default) = NMOS, open drain (requires an external pull-up resistor). 01 = PMOS, open drain (requires an external pull-down resistor). 10 = CMOS, active high. 11 = CMOS, active low. Rev. 0 | Page 69 of 104 AD9547 Register 0x0209 to Register 0x0210—IRQ Mask The IRQ mask register bits form a one-to-one correspondence with the bits of the IRQ monitor register (Address 0x0D02 to Address 0x0D09). When set to Logic 1, the IRQ mask bits enable the corresponding IRQ monitor bits to indicate an IRQ event. The default for all IRQ mask bits is Logic 0, which prevents the IRQ monitor from detecting any internal interrupts. Table 50. IRQ Mask for SYSCLK Address 0x0209 Bit [7:6] 5 4 [3:2] 1 0 Bit Name Unused SYSCLK unlocked SYSCLK locked Unused SYSCLK cal complete SYSCLK cal started Description Unused. Enables IRQ for indicating a SYSCLK PLL state transition from locked to unlocked. Enables IRQ for indicating a SYSCLK PLL state transition from unlocked to locked. Unused. Enables IRQ for indicating that SYSCLK calibration is complete. Enables IRQ for indicating that SYSCLK calibration has begun. Table 51. IRQ Mask for Distribution Sync, Watchdog Timer, and EEPROM Address 0x020A Bit [7:4] 3 2 1 0 Bit Name Unused Distribution sync Watchdog timer EEPROM fault EEPROM complete Description Unused. Enables IRQ for indicating a distribution sync event. Enables IRQ for indicating expiration of the watchdog timer. Enables IRQ for indicating a fault during an EEPROM load or save operation. Enables IRQ for indicating successful completion of an EEPROM load or save operation. Table 52. IRQ Mask for the Digital PLL Address 0x020B Bit 7 6 5 4 3 2 1 0 Bit Name Switching Closed Free run Holdover Frequency unlocked Frequency locked Phase unlocked Phase locked Description Enables IRQ for indicating that the DPLL is switching to a new reference. Enables IRQ for indicating that the DPLL has entered closed-loop operation. Enables IRQ for indicating that the DPLL has entered free-run mode. Enables IRQ for indicating that the DPLL has entered holdover mode. Enables IRQ for indicating that the DPLL lost frequency lock. Enables IRQ for indicating that the DPLL has acquired frequency lock. Enables IRQ for indicating that the DPLL lost phase lock. Enables IRQ for indicating that the DPLL has acquired phase lock. Table 53. IRQ Mask for History Update, Frequency Limit, and Phase Slew Limit Address 0x020C Bit [7:5] 4 3 Bit Name Unused History updated Frequency unclamped 2 Frequency clamped 1 Phase slew unlimited 0 Phase slew limited Description Unused. Enables IRQ for indicating the occurrence of a tuning word history update. Enables IRQ for indicating a state transition of the frequency limiter from clamped to unclamped. Enables IRQ for indicating a state transition of the frequency limiter from unclamped to clamped. Enables IRQ for indicating a state transition of the phase slew limiter from slew limiting to not slew limiting. Enables IRQ for indicating a state transition of the phase slew limiter from not slew limiting to slew limiting. Rev. 0 | Page 70 of 104 AD9547 Table 54. IRQ Mask for Reference Inputs Address 0x020D 0x020E 0x020F 0x0210 Bit 7 6 5 4 3 2 1 0 7 6 5 4 3 2 1 0 [7:0] [7:0] Bit Name Ref AA new profile Ref AA validated Ref AA fault cleared Ref AA fault Ref A new profile Ref A validated Ref A fault cleared Ref A fault Ref BB new profile Ref BB validated Ref BB fault cleared Ref BB fault Ref B new profile Ref B validated Ref B fault cleared Ref B fault Unused Unused Description Enables IRQ for indicating that Ref AA has switched to a new profile. Enables IRQ for indicating that Ref AA has been validated. Enables IRQ for indicating that Ref AA has been cleared of a previous fault. Enables IRQ for indicating that Ref AA has been faulted. Enables IRQ for indicating that Ref A has switched to a new profile. Enables IRQ for indicating that Ref A has been validated. Enables IRQ for indicating that Ref A has been cleared of a previous fault. Enables IRQ for indicating that Ref A has been faulted. Enables IRQ for indicating that Ref BB has switched to a new profile. Enables IRQ for indicating that Ref BB has been validated. Enables IRQ for indicating that Ref BB has been cleared of a previous fault. Enables IRQ for indicating that Ref BB has been faulted. Enables IRQ for indicating that Ref B has switched to a new profile. Enables IRQ for indicating that Ref B has been validated. Enables IRQ for indicating that Ref B has been cleared of a previous fault. Enables IRQ for indicating that Ref B has been faulted. Unused. Unused. Table 55. Watchdog Timer1 Address 0x0211 0x0212 1 Bit [7:0] [7:0] Bit Name Watchdog timer (expressed in ms) Description Watchdog timer, Bits[7:0] (default = 0x00). Watchdog timer, Bits[15:8] (default = 0x00). The watchdog timer is expressed in units of milliseconds (ms). The default value is 0 (disabled). Table 56. DAC Current1 Address 0x0213 0x0214 1 Bit [7:0] 7 Bit Name DAC full-scale current DAC shutdown [6:2] [1:0] Unused DAC full-scale current Description Full-scale current, Bits[7:0] (default = 0xFF). Shut down the DAC current sources. 0 (default) = normal operation. 1 = shut down. Unused. Full-scale current, Bits[9:8] (default = 0b01). (default current = 0x1FF, or 20.1 mA). The default DAC full-scale current value is 0x01FF = 511, which equates to 20.1375 mA. Rev. 0 | Page 71 of 104 AD9547 DPLL CONFIGURATION (REGISTER 0x0300 TO REGISTER 0x031B) Table 57. Free-Running Frequency Tuning Word1 Address 0x0300 0x0301 0x0302 0x0303 0x0304 0x0305 1 Bit [7:0] [7:0] [7:0] [7:0] [7:0] [7:0] Bit Name Free-running frequency tuning word (expressed as a 48-bit frequency tuning word) Description Free-running frequency tuning word, Bits[7:0]. Free-running frequency tuning word, Bits[15:8]. Free-running frequency tuning word, Bits[23:16]. Free-running frequency tuning word, Bits[31:24]. Free-running frequency tuning word, Bits[39:32]. Free-running frequency tuning word, Bits[47:40]. The default free-running tuning word is 0x000000 = 0, which equates to 0 Hz. Table 58. Update TW Address 0x0306 Bit [7:1] 0 Bit Name Unused Update TW Description Unused. A Logic 1 written to this bit transfers the free-running frequency tuning word (Register 0x0300 to Register 0x0305) to the register embedded in the tuning word processing logic. Note that it is not necessary to write the update TW bit when the device is in free-run mode. This is an autoclearing bit. Table 59. Pull-in Range Lower and Upper Limit1 Address 0x0307 0x0308 0x0309 0x030A 0x030B 0x030C 1 Bit [7:0] [7:0] [7:0] [7:0] [7:0] [7:0] Bit Name Pull-in range lower limit (expressed as a 24-bit frequency tuning word) Pull-in range upper limit (expressed as a 24-bit frequency tuning word) Description Lower limit pull-in range, Bits[7:0]. Lower limit pull-in range, Bits[15:8]. Lower limit pull-in range, Bits[23:16]. Upper limit pull-in range, Bits[7:0]. Upper limit pull-in range, Bits[15:8]. Upper limit pull-in range, Bits[23:16]. The default pull-in range lower limit is 0 and the upper range limit is 0xFFFFFF, which effectively spans the full output frequency range of the DDS. Table 60. Open-Loop Phase Offset1 Address 0x030D 0x030E 1 Bit [7:0] [7:0] Bit Name Open-loop phase offset (expressed in units of π/215 radians) Description DDS phase offset, Bits[7:0]. DDS phase offset, Bits[15:8]. The default DDS phase offset is 0. Table 61. Fixed Closed-Loop Phase Lock Offset1 Address 0x030F 0x0310 0x0311 0x0312 0x0313 1 Bit [7:0] [7:0] [7:0] [7:0] [7:0] Bit Name Fixed phase lock offset (expressed in ps) Description Fixed phase lock offset, Bits[7:0]. Fixed phase lock offset, Bits[15:8]. Fixed phase lock offset, Bits[23:16]. Fixed phase lock offset, Bits[31:24]. Fixed phase lock offset, Bits[39:32]. The default fixed closed loop phase lock offset is 0. Rev. 0 | Page 72 of 104 AD9547 Table 62. Incremental Closed-Loop Phase Lock Offset Step Size1 Address 0x0314 0x0315 1 Bit [7:0] [7:0] Bit Name Incremental phase lock offset step size (expressed in ps/step) Description Incremental phase lock offset step size, Bits[7:0]. Incremental phase lock offset step size, Bits[15:8]. The default incremental closed loop phase lock offset step size value is 0x03E8 = 1000 (1 ns). Table 63. Phase Slew Rate Limit1 Address 0x0316 0x0317 1 Bit [7:0] [7:0] Bit Name Phase slew rate limit (expressed in ns/sec) Description Phase slew rate limit, Bits[7:0]. Phase slew rate limit, Bits[15:8]. The default phase slew rate limit is 0 (or disabled). Table 64. History Accumulation Timer1 Address 0x0318 0x0319 0x031A 1 Bit [7:0] [7:0] [7:0] Bit Name History accumulation timer (expressed in ms) Description History accumulation timer, Bits[7:0]. History accumulation timer, Bits[15:8]. History accumulation timer, Bits[23:16]. Do not program a timer value of 0. The history accumulation timer default value is 0x007530 = 30,000 (30 sec). Table 65. History Mode Address 0x031B Bit [7:5] 4 Bit Name Unused Single sample fallback 3 Persistent history [2:0] Incremental average Description Unused. Controls the holdover history. If tuning word history is not available for the reference that was active just prior to holdover, then 0 (default) = use the free-running frequency tuning word register value. 1 = use the last tuning word from the DPLL. Controls the holdover history initialization. When switching to a new reference 0 (default) = clear the tuning word history. 1 = retain the previous tuning word history. History mode value from 0 to 7 (default = 0). Rev. 0 | Page 73 of 104 AD9547 CLOCK DISTRIBUTION OUTPUT CONFIGURATION (REGISTER 0x0400 TO REGISTER 0x0417) Table 66. Distribution Settings1 Address 0x0400 1 Bit [7:6] 5 Bit Name Unused External distribution resistor 4 Receiver mode [3:2] 1 Unused OUT1 power-down 0 OUT0 power-down Description Unused. Output current control for the clock distribution outputs. 0 (default) = internal current setting resistor. 1 = external current setting resistor. Clock distribution receiver mode. 0 (default) = normal operation. 1 = high frequency mode (super-Nyquist). Write a 1 to these bits. Power down clock distribution output OUT1. 0 (default) = normal operation. 1 = power down. Power down clock distribution output OUT0. 0 (default) = normal operation. 1 = power-down. When Bits [1:0] = 11, the clock distribution output enters a deep sleep mode. Table 67. Distribution Enable Address 0x0401 Bit [7:2] 1 Bit Name Unused OUT1 enable 0 OUT0 enable Description Unused. Enable the OUT1 driver. 0 (default) = disable. 1 = enable. Enable the OUT0 driver. 0 (default) = disable. 1 = enable. Table 68. Distribution Synchronization Address 0x0402 Bit [7:6] [5:4] Bit Name Unused Sync source [3:2] 1 Unused OUT1 sync mask 0 OUT0 sync mask Description Unused. Select the sync source for the clock distribution output channels. 00 (default) = direct. 01 = active reference. 10 = DPLL feedback edge. 11 = reserved. Unused. Mask the synchronous reset to the OUT1 divider. 0 (default) = unmasked. 1 = masked. Mask the synchronous reset to the OUT0 divider. 0 (default) = unmasked. 1 = masked. Table 69. Automatic Synchronization Address 0x0403 Bit [7:2] [1:0] Bit Name Unused Automatic sync mode Description Unused. Autosync mode. 00 (default) = disabled. 01 = sync on DPLL frequency lock. 10 = sync on DPLL phase lock. 11 = reserved. Rev. 0 | Page 74 of 104 AD9547 Table 70. Distribution Channel Modes Address 0x0404 0x0405 0x0406 0x0407 Bit [7:6] 5 Bit Name Unused OUT0 CMOS phase invert 4 OUT0 polarity invert 3 OUT0 drive strength [2:0] OUT0 mode [7:6] 5 Unused OUT1 CMOS phase invert 4 OUT1 polarity invert 3 OUT1 drive strength [2:0] OUT1 mode [7:0] [7:0] Unused Description Unused. When the output mode is CMOS, the bit inverts the relative phase between the two CMOS output pins. Otherwise, this bit is nonfunctional. 0 (default) = not inverted. 1 = inverted. Invert the polarity of OUT0. 0 (default) = not inverted. 1 = inverted. OUT0 output drive capability control. 0 (default) = CMOS: low drive strength; LVDS: 3.5 mA nominal. 1 = CMOS: normal drive strength; LVDS: 7 mA nominal. OUT0 operating mode select. 000 = CMOS (both pins). 001 = CMOS (positive pin), tristate (negative pin). 010 = tristate (positive pin), CMOS (negative pin). 011 (default) = tristate (both pins). 100 = LVDS. 101 = LVPECL. 110 = reserved. 111 = reserved. Unused. When the output mode is CMOS, the bit inverts the relative phase between the two CMOS output pins. Otherwise, this bit is nonfunctional. 0 (default) = not inverted. 1 = inverted. Invert the polarity of OUT1. 0 (default) = not inverted. 1 = inverted. OUT1 output drive capability control. 0 (default) = CMOS: low drive strength; LVDS: 3.5 mA nominal. 1 = CMOS: normal drive strength; LVDS: 7 mA nominal. OUT1 operating mode select. 000 = CMOS (both pins). 001 = CMOS (positive pin), tristate (negative pin). 010 = tristate (positive pin), CMOS (negative pin). 011 (default) = tristate (both pins). 100 = LVDS. 101 = LVPECL. 110 = reserved. 111 = reserved. Unused. Rev. 0 | Page 75 of 104 AD9547 Register 0x0408 to Register 0x0417—Distribution Channel Dividers Table 71. Q0 Divider1 Address 0x0408 0x0409 0x040A 0x040B 1 Bit [7:0] [7:0] [7:0] [7:6] [5:0] Bit Name Q0 Unused Q0 Description Q0 divider, Bits[7:0]. Q0 divider, Bits[15:8]. Q0 divider, Bits[23:16]. Unused. Q0 divider, Bits[29:24]. The default value is 0 (or divide by 1). Table 72. Q1 Divider1 Address 0x040C 0x040D 0x040E 0x040F 1 Bit [7:0] [7:0] [7:0] [7:6] [5:0] Bit Name Q1 Unused Q1 Description Q1 divider, Bits[7:0]. Q1 divider, Bits[15:8]. Q1 divider, Bits[23:16]. Unused. Q1 divider, Bits[29:24]. The default value is 0 (or divide by 1). Table 73. Reserved Address 0x0410 to 0x0417 Bit [7:0] Bit Name Unused Description Unused. REFERENCE INPUT CONFIGURATION (REGISTER 0x0500 TO REGISTER 0x0507) When all bits are set, the reference receiver section enters a deep sleep mode. Table 74. Reference Power-Down Address 0x0500 Bit [7:4] 3 Bit Name Unused Ref BB power-down 2 Ref B power-down 1 Ref AA power-down 0 Ref A power-down Description Write a 1 to these bits. REF BB input receiver power-down. 0 (default) = normal operation. 1 = power-down. REF B input receiver power-down. 0 (default) = normal operation. 1 = power-down. REF AA input receiver power-down. 0 (default) = normal operation. 1 = power-down. REF A input receiver power-down. 0 (default) = normal operation. 1 = power-down. Rev. 0 | Page 76 of 104 AD9547 Table 75. Reference Logic Family Address 0x0501 0x0502 Bit [7:6] Bit Name Ref BB logic family [5:4] Ref B logic family [3:2] [1:0] [7:0] Ref AA logic family Ref A logic family Unused Description Select the logic family for the REF BB input receiver (ignored if Bits[5:4] = 00). 00 (default) = disabled. 01 = 1.2 V to 1.5 V CMOS. 10 = 1.8 V to 2.5 V CMOS. 11 = 3.0 V to 3.3 V CMOS. Select logic family for REF B input receiver. 00 (default) = differential (REFB/BB is positive/negative input). 01 = 1.2 V to 1.5 V CMOS. 10 = 1.8 V to 2.5 V CMOS. 11 = 3.0 V to 3.3 V CMOS. The same as Register 0x0501, Bits[7:6] but for REF AA. The same as Register 0x0501, Bits[5:4] but for REF A. Unused. Table 76. Manual Reference Profile Selection Address 0x0503 0x0504 0x0505 0x0506 Bit 7 Bit Name Enable Ref AA manual profile [6:4] Ref AA manual profile 3 [2:0] 7 [6:4] 3 [2:0] [7:0] [7:0] Enable Ref A manual profile Ref A manual profile Enable Ref BB manual profile Ref BB manual profile Enable Ref B manual profile Ref B manual profile Unused Description Select manual or automatic reference profile assignment for REF AA. 0 (default) = automatic. 1 = manual. Manual profile assignment. 000 (default) = Profile 0. 001 = Profile 1. 010 = Profile 2. 011 = Profile 3. 100 = Profile 4. 101 = Profile 5. 110 = Profile 6. 111 = Profile 7. Same as Register 0x0503, Bit 7 but for REF A. Same as Register 0x0503, Bits[6:4] but for REF A. Same as Register 0x0503, Bit 7 but for REF BB. Same as Register 0x0503, Bits[6:4] but for REF BB. Same as Register 0x0503, Bit 7 but for REF B. Same as Register 0x0503, Bits[6:4] but for REF B. Unused. Table 77. Phase Build-Out Switching Address 0x0507 Bit [7:3] [2:0] Bit Name Unused Phase master threshold priority Description Unused. Threshold priority level (a value of 0 to 7, with 0 (default) being the highest priority level). References with a selection priority value lower than this value are treated as phase masters (see the profile registers for selection priority value). Rev. 0 | Page 77 of 104 AD9547 PROFILE REGISTERS (REGISTER 0x0600 TO REGISTER 0x07FF) Note that the default value of every bit is 0 for Profile 0 to Profile 7. Register 0x0600 to Register 0x0631—Profile 0 Table 78. Priorities—Profile 0 Address 0x0600 Bit [7:6] [5:3] Bit Name Unused Promoted priority [2:0] Selection priority Description Unused. User-assigned priority level (0 to 7) of the reference associated with Profile 0 while that reference is the active reference. The numeric value of the promoted priority must be less than or equal to the numeric value of the selection priority. User-assigned priority level (0 to 7) of the reference associated with Profile 0, which ranks that reference relative to the others. Table 79. Reference Period—Profile 0 Address 0x0601 0x0602 0x0603 0x0604 0x0605 0x0606 0x0607 Bit [7:0] [7:0] [7:0] [7:0] [7:0] [7:0] [7:0] Bit Name Reference period (expressed in units of fs) Unused. Description Nominal reference period, Bits[7:0]. Nominal reference period, Bits[15:8]. Nominal reference period, Bits[23:16]. Nominal reference period, Bits[31:24]. Nominal reference period, Bits[39:32]. Unused. Write 0s to these bits. Table 80. Tolerance—Profile 0 Address 0x0608 0x0609 0x060A 0x060B 0x060C 0x060D Bit [7:0] [7:0] [7:4] [3:0] [7:0] [7:0] [7:4] [3:0] Bit Name Inner tolerance Unused Inner tolerance Outer tolerance Unused Outer tolerance Description Inner tolerance, Bits[7:0]. Inner tolerance, Bits[15:8]. Unused. Inner tolerance, Bits[19:16]. Outer tolerance, Bits[7:0]. Outer tolerance, Bits[15:8]. Unused. Outer tolerance, Bits[19:16]. Table 81. Validation Timer—Profile 0 Address 0x060E 0x060F Bit [7:0] [7:0] Bit Name Validation timer (expressed in units of ms) Description Validation timer, Bits[7:0]. Validation timer, Bits[15:8]. Table 82. Redetect Timer—Profile 0 Address 0x0610 0x0611 Bit [7:0] [7:0] Bit Name Redetect timer (expressed in units of ms) Description Redetect timer, Bits[7:0]. Redetect timer, Bits[15:8]. Rev. 0 | Page 78 of 104 AD9547 Table 83. Digital Loop Filter Coefficients—Profile 01 Address 0x0612 0x0613 0x0614 0x0615 0x0616 0x0617 0x0618 0x0619 0x061A 0x061B 0x061C 0x061D 1 Bit [7:0] [7:0] [7:6] [5:0] [7:1] 0 [7:0] 7 [6:2] [1:0] [7:0] [7:0] [7:6] [5:1] 0 [7:0] 7 [6:0] [7:4] [3:0] Bit Name Alpha-0 linear Alpha-2 exponent Alpha-1 exponent Beta-0 linear Alpha-2 exponent Beta-0 linear Unused Beta-1 exponent Beta-0 linear Gamma-0 linear Unused Gamma-1 exponent Gamma-0 linear Delta-0 linear Delta-1 exponent Delta-0 linear Alpha-3 exponent Delta-1 exponent Description Alpha-0 coefficient linear, Bits[7:0]. Alpha-0 coefficient linear, Bits[15:8]. Alpha-2 coefficient exponent, Bits[1:0]. Alpha-1 coefficient exponent, Bits[5:0]. Beta-0 coefficient linear, Bits[6:0]. Alpha-2 coefficient exponent, Bit 2. Beta-0 coefficient linear, Bits[14:7]. Unused. Beta-1 coefficient exponent, Bits[4:0]. Beta-0 coefficient linear, Bits[16:15]. Gamma-0 coefficient linear, Bits[7:0]. Gamma-0 coefficient linear, Bits[15:8]. Unused. Gamma-1 coefficient exponent, Bits[4:0]. Gamma-0 coefficient linear, Bit 16. Delta-0 coefficient linear, Bits[7:0]. Delta-1 coefficient exponent, Bit 0. Delta-0 coefficient linear, Bits[14:8]. Alpha-3 coefficient exponent, Bits[3:0]. Delta-1 coefficient exponent, Bits[4:1]. The digital loop filter coefficients (α, β, γ, and δ) have the general form: x(2y), where x is the linear component and y is the exponential component of the coefficient. The value of the linear component (x) constitutes a fraction, where 0 ≤ x < 1. The exponential component (y) is an integer. See the Calculating the Digital Filter Coefficients section for details. Register 0x061E to Register 0x0628—Profile 0 Frequency Multiplication Table 84. R Divider—Profile 01 Address 0x061E 0x061F 0x0620 0x0621 1 Bit [7:0] [7:0] [7:0] [7:6] [5:0] Bit Name R Unused R Description R, Bits[7:0]. R, Bits[15:8]. R, Bits[23:16]. Unused. R, Bits[29:24]. The value stored in the R divider register yields an actual divide ratio of one more than the programmed value. Table 85. S Divider—Profile 01 Address 0x0622 0x0623 0x0624 0x0625 1 Bit [7:0] [7:0] [7:4] [3:0] [7:0] Bit Name S S Unused S Unused Description S, Bits[7:0]. S, Bits[15:8]. Unused. S, Bits[19:16]. Unused. The value stored in the S divider register yields an actual divide ratio of one more than the programmed value. Furthermore, the value of S must be at least 7. Rev. 0 | Page 79 of 104 AD9547 Table 86. Fractional Feedback Divider—Profile 0 Address 0x0626 0x0627 0x0628 Bit [7:0] [7:4] [3:2] [1:0] [7:6] [5:0] Bit Name V U Unused V Unused U Description V, Bits[7:0]. U, Bits[3:0]. Unused. V, Bits[9:8]. Unused. U, Bits[9:4]. Table 87. Lock Detectors—Profile 0 Address 0x0629 0x062A 0x062B 0x062C 0x062D 0x062E 0x062F 0x0630 0x0631 Bit [7:0] [7:0] [7:0] [7:0] [7:0] [7:0] [7:0] [7:0] [7:0] Bit Name Phase lock threshold (expressed in units of ps) Phase lock fill rate Phase lock drain rate Frequency lock threshold (expressed in units of ps) Frequency lock fill rate Frequency lock drain rate Description Phase lock threshold, Bits[7:0]. Phase lock threshold, Bits[15:8]. Phase lock fill rate, Bits[7:0]. Phase lock drain rate, Bits[7:0]. Frequency lock threshold, Bits[7:0]. Frequency lock threshold, Bits[15:8]. Frequency lock threshold, Bits[23:16]. Frequency lock fill rate, Bits[7:0]. Frequency lock drain rate, Bits[7:0]. Register 0x0632 to Register 0x067F—Profile 1 Table 88. Priorities—Profile 1 Address 0x0632 Bit [7:6] [5:3] Bit Name Unused Promoted priority [2:0] Selection priority Description Unused. User-assigned priority level (0 to 7) of the reference associated with Profile 1 while that reference is the active reference. The numeric value of the promoted priority must be less than or equal to the numeric value of the selection priority. User-assigned priority level (0 to 7) of the reference associated with Profile 1, which ranks that reference relative to the others. Table 89. Reference Period—Profile 1 Address 0x0633 0x0634 0x0635 0x0636 0x0637 0x0638 0x0639 Bit [7:0] [7:0] [7:0] [7:0] [7:0] [7:0] [7:0] Bit Name Reference period (expressed in units of fs) Unused Description Nominal reference period, Bits[7:0]. Nominal reference period, Bits[15:8]. Nominal reference period, Bits[23:16]. Nominal reference period, Bits[31:24]. Nominal reference period, Bits[39:32]. Unused. Write 0s to these bits. Table 90. Tolerance—Profile 1 Address 0x063A 0x063B 0x063C 0x063D 0x063E 0x063F Bit [7:0] [7:0] [7:4] [3:0] [7:0] [7:0] [7:4] [3:0] Bit Name Inner tolerance Unused Inner tolerance Outer tolerance Unused Outer tolerance Description Inner tolerance, Bits[7:0]. Inner tolerance, Bits[15:8]. Unused. Inner tolerance, Bits[19:16]. Outer tolerance, Bits[7:0]. Outer tolerance, Bits[15:8]. Unused. Outer tolerance, Bits[19:16]. Rev. 0 | Page 80 of 104 AD9547 Table 91. Validation Timer—Profile 1 Address 0x0640 0x0641 Bit [7:0] [7:0] Bit Name Validation timer (expressed in units of ms) Description Validation timer, Bits[7:0]. Validation timer, Bits[15:8]. Table 92. Redetect Timer—Profile 1 Address 0x0642 0x0643 Bit [7:0] [7:0] Bit Name Redetect timer (expressed in units of ms) Description Redetect timer, Bits[7:0]. Redetect timer, Bits[15:8]. Table 93. Digital Loop Filter Coefficients—Profile 11 Address 0x0644 0x0645 0x0646 0x0647 0x0648 0x0649 0x064A 0x064B 0x064C 0x064D 0x064E 0x064F 1 Bit [7:0] [7:0] [7:6] [5:0] [7:1] 0 [7:0] 7 [6:2] [1:0] [7:0] [7:0] [7:6] [5:1] 0 [7:0] 7 [6:0] [7:4] [3:0] Bit Name Alpha-0 linear Alpha-2 exponent Alpha-1 exponent Beta-0 linear Alpha-2 exponent Beta-0 linear Unused Beta-1 exponent Beta-0 linear Gamma-0 linear Unused Gamma-1 exponent Gamma-0 linear Delta-0 linear Delta-1 exponent Delta-0 linear Alpha-3 exponent Delta-1 exponent Description Alpha-0 coefficient linear, Bits[7:0]. Alpha-0 coefficient linear, Bits[15:8]. Alpha-2 coefficient exponent, Bits[1:0]. Alpha-1 coefficient exponent, Bits[5:0]. Beta-0 coefficient linear, Bits[6:0]. Alpha-2 coefficient exponent, Bit 2. Beta-0 coefficient linear, Bits[14:7]. Unused. Beta-1 coefficient exponent, Bits[4:0]. Beta-0 coefficient linear, Bits[16:15]. Gamma-0 coefficient linear, Bits[7:0]. Gamma-0 coefficient linear, Bits[15:8]. Unused. Gamma-1 coefficient exponent, Bits[4:0]. Gamma-0 coefficient linear, Bit 16. Delta-0 coefficient linear, Bits[7:0]. Delta-1 coefficient exponent, Bit 0. Delta-0 coefficient linear, Bits[14:8]. Alpha-3 coefficient exponent, Bits[3:0]. Delta-1 coefficient exponent, Bits[4:1]. The digital loop filter coefficients (α, β, γ, and δ) have the general form: x(2y), where x is the linear component and y is the exponential component of the coefficient. The value of the linear component (x) constitutes a fraction, where 0 ≤ x < 1. The exponential component (y) is an integer. See the Calculating the Digital Filter Coefficients section for details. Register 0x0650 to Register 0x065A—Profile 1 Frequency Multiplication Table 94. R Divider—Profile 11 Address 0x0650 0x0651 0x0652 0x0653 1 Bit [7:0] [7:0] [7:0] [7:6] [5:0] Bit Name R Unused R Description R, Bits[7:0]. R, Bits[15:8]. R, Bits[23:16]. Unused. R, Bits[29:24]. The value stored in the R divider register yields an actual divide ratio of one more than the programmed value. Rev. 0 | Page 81 of 104 AD9547 Table 95. S Divider—Profile 1 1 Address 0x0654 0x0655 0x0656 0x0657 1 Bit [7:0] [7:0] [7:4] [3:0] [7:0] Bit Name S S Unused S Unused Description S, Bits[7:0]. S, Bits[15:8]. Unused. S, Bits[19:16]. Unused. The value stored in the S divider register yields an actual divide ratio of one more than the programmed value. Furthermore, the value of S must be at least 7. Table 96. Fractional Feedback Divider—Profile 1 Address 0x0658 0x0659 0x065A Bit [7:0] [7:4] [3:2] [1:0] [7:6] [5:0] Bit Name V U Unused V Unused U Description V, Bits[7:0]. U, Bits[3:0]. Unused. V, Bits[9:8]. Unused. U, Bits[9:4]. Table 97. Lock Detectors—Profile 1 Address 0x065B 0x065C 0x065D 0x065E 0x065F 0x0660 0x0661 0x0662 0x0663 0x0664 to 0x067F Bit [7:0] [7:0] [7:0] [7:0] [7:0] [7:0] [7:0] [7:0] [7:0] [7:0] Bit Name Phase lock threshold (expressed in units of ps) Phase lock fill rate Phase lock drain rate Frequency lock threshold (expressed in units of ps) Frequency lock fill rate Frequency lock drain rate Unused Description Phase lock threshold, Bits[7:0]. Phase lock threshold, Bits[15:8]. Phase lock fill rate, Bits[7:0]. Phase lock drain rate, Bits[7:0]. Frequency lock threshold, Bits[7:0]. Frequency lock threshold, Bits[15:8]. Frequency lock threshold, Bits[23:16]. Frequency lock fill rate, Bits[7:0]. Frequency lock drain rate, Bits[7:0]. Unused. Register 0x0680 to Register 0x06B1—Profile 2 Table 98. Priorities—Profile 2 Address 0x0680 Bit [7:6] [5:3] Bit Name Unused Promoted priority [2:0] Selection priority Description Unused. User-assigned priority level (0 to 7) of the reference associated with Profile 2 while that reference is the active reference. The numeric value of the promoted priority must be less than or equal to the numeric value of the selection priority. User-assigned priority level (0 to 7) of the reference associated with Profile 2, which ranks that reference relative to the others. Table 99. Reference Period—Profile 2 Address 0x0681 0x0682 0x0683 0x0684 0x0685 0x0686 0x0687 Bit [7:0] [7:0] [7:0] [7:0] [7:0] [7:0] [7:0] Bit Name Reference period (expressed in units of fs) Unused Description Nominal reference period, Bits[7:0]. Nominal reference period, Bits[15:8]. Nominal reference period, Bits[23:16]. Nominal reference period, Bits[31:24]. Nominal reference period, Bits[39:32]. Unused. Write 0s to these bits. Rev. 0 | Page 82 of 104 AD9547 Table 100. Tolerance—Profile 2 Address 0x0688 0x0689 0x068A 0x068B 0x068C 0x068D Bit [7:0] [7:0] [7:4] [3:0] [7:0] [7:0] [7:4] [3:0] Bit Name Inner tolerance Unused Inner tolerance Outer tolerance Unused Outer tolerance Description Inner tolerance, Bits[7:0]. Inner tolerance, Bits[15:8]. Unused. Inner tolerance, Bits[19:16]. Outer tolerance, Bits[7:0]. Outer tolerance, Bits[15:8]. Unused. Outer tolerance, Bits[19:16]. Table 101. Validation Timer—Profile 2 Address 0x068E 0x068F Bit [7:0] [7:0] Bit Name Validation timer (expressed in units of ms) Description Validation timer, Bits[7:0]. Validation timer, Bits[15:8]. Table 102. Redetect Timer—Profile 2 Address 0x0690 0x0691 Bit [7:0] [7:0] Bit Name Redetect timer (expressed in units of ms) Description Redetect timer, Bits[7:0]. Redetect timer, Bits[15:8]. Table 103. Digital Loop Filter Coefficients—Profile 21 Address 0x0692 0x0693 0x0694 0x0695 0x0696 0x0697 0x0698 0x0699 0x069A 0x069B 0x069C 0x069D 1 Bit [7:0] [7:0] [7:6] [5:0] [7:1] 0 [7:0] 7 [6:2] [1:0] [7:0] [7:0] [7:6] [5:1] 0 [7:0] 7 [6:0] [7:4] [3:0] Bit Name Alpha-0 linear Alpha-2 exponent Alpha-1 exponent Beta-0 linear Alpha-2 exponent Beta-0 linear Unused Beta-1 exponent Beta-0 linear Gamma-0 linear Unused Gamma-1 exponent Gamma-0 linear Delta-0 linear Delta-1 exponent Delta-0 linear Alpha-3 exponent Delta-1 exponent Description Alpha-0 coefficient linear, Bits[7:0]. Alpha-0 coefficient linear, Bits[15:8]. Alpha-2 coefficient exponent, Bits[1:0]. Alpha-1 coefficient exponent, Bits[5:0]. Beta-0 coefficient linear, Bits[6:0]. Alpha-2 coefficient exponent, Bit 2. Beta-0 coefficient linear, Bits[14:7]. Unused. Beta-1 coefficient exponent, Bits[4:0]. Beta-0 coefficient linear, Bits[16:15]. Gamma-0 coefficient linear, Bits[7:0]. Gamma-0 coefficient linear, Bits[15:8]. Unused. Gamma-1 coefficient exponent, Bits[4:0]. Gamma-0 coefficient linear, Bit 16. Delta-0 coefficient linear, Bits[7:0]. Delta-1 coefficient exponent, Bit 0. Delta-0 coefficient linear, Bits[14:8]. Alpha-3 coefficient exponent, Bits[3:0]. Delta-1 coefficient exponent, Bits[4:1]. The digital loop filter coefficients (α, β, γ, and δ) have the general form: x(2y), where x is the linear component and y is the exponential component of the coefficient. The value of the linear component (x) constitutes a fraction, where 0 ≤ x < 1. The exponential component (y) is an integer (see the Calculating the Digital Filter Coefficients section for details). Rev. 0 | Page 83 of 104 AD9547 Register 0x069E to Register 0x06A8—Profile 2 Frequency Multiplication Table 104. R Divider—Profile 2 1 Address 0x069E 0x069F 0x06A0 0x06A1 1 Bit [7:0] [7:0] [7:0] [7:6] [5:0] Bit Name R Unused R Description R, Bits[7:0]. R, Bits[15:8]. R, Bits[23:16]. Unused. R, Bits[29:24]. The value stored in the R divider register yields an actual divide ratio of one more than the programmed value. Table 105. S Divider—Profile 2 1 Address 0x06A2 0x06A3 0x06A4 0x06A5 1 Bit [7:0] [7:0] [7:4] [3:0] [7:0] Bit Name S S Unused S Unused Description S, Bits[7:0]. S, Bits[15:8]. Unused. S, Bits[19:16]. Unused. The value stored in the S divider register yields an actual divide ratio of one more than the programmed value. Furthermore, the value of S must be at least 7. Table 106. Fractional Feedback Divider—Profile 2 Address 0x06A6 0x06A7 0x06A8 Bit [7:0] [7:4] [3:2] [1:0] [7:6] [5:0] Bit Name V U Unused V Unused U Description V, Bits[7:0]. U, Bits[3:0]. Unused. V, Bits[9:8]. Unused. U, Bits[9:4]. Table 107. Lock Detectors—Profile 2 Address 0x06A9 0x06AA 0x06AB 0x06AC 0x06AD 0x06AE 0x06AF 0x06B0 0x06B1 Bit [7:0] [7:0] [7:0] [7:0] [7:0] [7:0] [7:0] [7:0] [7:0] Bit Name Phase lock threshold (expressed in units of ps) Phase lock fill rate Phase lock drain rate Frequency lock threshold (expressed in units of ps) Frequency lock fill rate Frequency lock drain rate Description Phase lock threshold, Bits[7:0]. Phase lock threshold, Bits[15:8]. Phase lock fill rate, Bits[7:0]. Phase lock drain rate, Bits[7:0]. Frequency lock threshold, Bits[7:0]. Frequency lock threshold, Bits[15:8]. Frequency lock threshold, Bits[23:16]. Frequency lock fill rate, Bits[7:0]. Frequency lock drain rate, Bits[7:0]. Rev. 0 | Page 84 of 104 AD9547 Register 0x06B2 to Register 0x06FF—Profile 3 Table 108. Priorities—Profile 3 Address 0x06B2 Bit [7:6] [5:3] Bit Name Unused Promoted priority [2:0] Selection priority Description Unused. User-assigned priority level (0 to 7) of the reference associated with Profile 3 while that reference is the active reference. The numeric value of the promoted priority must be less than or equal to the numeric value of the selection priority. User-assigned priority level (0 to 7) of the reference associated with Profile 3, which ranks that reference relative to the others. Table 109. Reference Period—Profile 3 Address 0x06B3 0x06B4 0x06B5 0x06B6 0x06B7 0x06B8 0x06B9 Bit [7:0] [7:0] [7:0] [7:0] [7:0] [7:0] [7:0] Bit Name Reference period (expressed in units of fs) Unused Description Nominal reference period, Bits[7:0]. Nominal reference period, Bits[15:8]. Nominal reference period, Bits[23:16]. Nominal reference period, Bits[31:24]. Nominal reference period, Bits[39:32]. Unused. Write 0s to these bits. Table 110. Tolerance—Profile 3 Address 0x06BA 0x06BB 0x06BC 0x06BD 0x06BE 0x06BF Bit [7:0] [7:0] [7:4] [3:0] [7:0] [7:0] [7:4] [3:0] Bit Name Inner tolerance Unused Inner tolerance Outer tolerance Unused Outer tolerance Description Inner tolerance, Bits[7:0]. Inner tolerance, Bits[15:8]. Unused. Inner tolerance, Bits[19:16]. Outer tolerance, Bits[7:0]. Outer tolerance, Bits[15:8]. Unused. Outer tolerance, Bits[19:16]. Table 111. Validation Timer—Profile 3 Address 0x06C0 0x06C1 Bit [7:0] [7:0] Bit Name Validation timer (expressed in units of ms) Description Validation timer, Bits[7:0]. Validation timer, Bits[15:8]. Table 112. Redetect Timer—Profile 3 Address 0x06C2 0x06C3 Bit [7:0] [7:0] Bit Name Redetect timer (expressed in units of ms) Description Redetect timer, Bits[7:0]. Redetect timer, Bits[15:8]. Table 113. Digital Loop Filter Coefficients—Profile 31 Address 0x06C4 0x06C5 0x06C6 0x06C7 0x06C8 Bit [7:0] [7:0] [7:6] [5:0] [7:1] 0 [7:0] Bit Name Alpha-0 linear Alpha-2 exponent Alpha-1 exponent Beta-0 linear Alpha-2 exponent Beta-0 linear Description Alpha-0 coefficient linear, Bits[7:0]. Alpha-0 coefficient linear, Bits[15:8]. Alpha-2 coefficient exponent, Bits[1:0]. Alpha-1 coefficient exponent, Bits[5:0]. Beta-0 coefficient linear, Bits[6:0]. Alpha-2 coefficient exponent, Bit 2. Beta-0 coefficient linear, Bits[14:7]. Rev. 0 | Page 85 of 104 AD9547 Address 0x06C9 0x06CA 0x06CB 0x06CC 0x06CD 0x06CE 0x06CF 1 Bit 7 [6:2] [1:0] [7:0] [7:0] [7:6] [5:1] 0 [7:0] 7 [6:0] [7:4] [3:0] Bit Name Unused Beta-1 exponent Beta-0 linear Gamma-0 linear Unused Gamma-1 exponent Gamma-0 linear Delta-0 linear Delta-1 exponent Delta-0 linear Alpha-3 exponent Delta-1 exponent Description Unused. Beta-1 coefficient exponent, Bits[4:0]. Beta-0 coefficient linear, Bits[16:15]. Gamma-0 coefficient linear, Bits[7:0]. Gamma-0 coefficient linear, Bits[15:8]. Unused. Gamma-1 coefficient exponent, Bits[4:0]. Gamma-0 coefficient linear, Bit 16. Delta-0 coefficient linear, Bits[7:0]. Delta-1 coefficient exponent, Bit 0. Delta-0 coefficient linear, Bits[14:8]. Alpha-3 coefficient exponent, Bits[3:0]. Delta-1 coefficient exponent, Bits[4:1]. The digital loop filter coefficients (α, β, γ, and δ) have the general form: x(2y), where x is the linear component and y is the exponential component of the coefficient. The value of the linear component (x) constitutes a fraction, where 0 ≤ x < 1. The exponential component (y) is an integer (see the Calculating the Digital Filter Coefficients section for details). Register 0x06D0 to Register 0x06DA—Profile 3 Frequency Multiplication Table 114. R Divider—Profile 3 1 Address 0x06D0 0x06D1 0x06D2 0x06D3 1 Bit [7:0] [7:0] [7:0] [7:6] [5:0] Bit Name R Unused R Description R, Bits[7:0]. R, Bits[15:8]. R, Bits[23:16]. Unused. R, Bits[29:24]. The value stored in the R divider register yields an actual divide ratio of one more than the programmed value. Table 115. S Divider—Profile 3 1 Address 0x06D4 0x06D5 0x06D6 0x06D7 1 Bit [7:0] [7:0] [7:4] [3:0] [7:0] Bit Name S S Unused S Unused Description S, Bits[7:0]. S, Bits[15:8]. Unused. S, Bits[19:16]. Unused. The value stored in the S divider register yields an actual divide ratio of one more than the programmed value. Furthermore, the value of S must be at least 7. Table 116. Fractional Feedback Divider—Profile 3 Address 0x06D8 0x06D9 0x06DA Bit [7:0] [7:4] [3:2] [1:0] [7:6] [5:0] Bit Name V U Unused V Unused U Description V, Bits[7:0]. U, Bits[3:0]. Unused. V, Bits[9:8]. Unused. U, Bits[9:4]. Rev. 0 | Page 86 of 104 AD9547 Table 117. Lock Detectors—Profile 3 Address 0x06DB 0x06DC 0x06DD 0x06DE 0x06DF 0x06E0 0x06E1 0x06E2 0x06E3 0x06E4 to 0x06FF Bit [7:0] [7:0] [7:0] [7:0] [7:0] [7:0] [7:0] [7:0] [7:0] [7:0] Bit Name Phase lock threshold (expressed in units of ps) Phase lock fill rate Phase lock drain rate Frequency lock threshold (expressed in units of ps) Frequency lock fill rate Frequency lock drain rate Unused Description Phase lock threshold, Bits[7:0]. Phase lock threshold, Bits[15:8]. Phase lock fill rate, Bits[7:0]. Phase lock drain rate, Bits[7:0]. Frequency lock threshold, Bits[7:0]. Frequency lock threshold, Bits[15:8]. Frequency lock threshold, Bits[23:16]. Frequency lock fill rate, Bits[7:0]. Frequency lock drain rate, Bits[7:0]. Unused. Register 0x0700 to Register 0x07FF—Profile 4 to Profile 7 Profile 4 (Register 0x0700 to Register 0x0731) is identical to Profile 0 (Register 0x0600 to Register 0x0631). Profile 5 (Register 0x0732 to Register 0x077F) is identical to Profile 1 (Register 0x0632 to Register 0x067F). Profile 6 (Register 0x0780 to Register 0x07B1) is identical to Profile 2 (Register 0x0680 to Register 0x06B1). Profile 7 (Register 0x07B2 to Register 0x07FF) is identical to Profile 3 (Register 0x06B2 to Register 0x06FF). Rev. 0 | Page 87 of 104 AD9547 OPERATIONAL CONTROLS (REGISTER 0x0A00 TO REGISTER 0x0A10) Table 118. General Power-Down Address 0x0A00 Bit 7 Bit Name Reset sans regmap 6 5 Unused SYSCLK power-down 4 Reference power-down 3 TDC power-down 2 DAC power-down 1 Dist power-down 0 Full power-down Description Reset internal hardware but retain programmed register values. 0 (default) = normal operation. 1 = reset. Unused. Place SYSCLK input and PLL in deep sleep mode. 0 (default) = normal operation. 1 = power-down. Place reference clock inputs in deep sleep mode. 0 (default) = normal operation. 1 = power-down. Place the time-to-digital converter in deep sleep mode. 0 (default) = normal operation. 1 = power-down. Place the DAC in deep sleep mode. 0 (default) = normal operation. 1 = power-down. Place the clock distribution outputs in deep sleep mode. 0 (default) = normal operation. 1 = power-down. Place the entire device in deep sleep mode. 0 (default) = normal operation. 1 = power-down. Table 119. Loop Mode Address 0x0A01 Bit 7 6 Bit Name Unused User holdover 5 User free run [4:3] User selection mode 2 [1:0] Unused User reference selection Description Unused. Force the device into holdover mode. 0 (default) = normal operation. 1 = force device into holdover mode. The device functions as though all input references are faulted. Force the device into free-run mode. 0 (default) = normal operation. 1 = force device into free-run mode. The free-running frequency tuning word register (Address 0x0300 to Address 0x0305) specifies the DDS output frequency. Note that, when user free run is set, it overrides user holdover. Select the operating mode of the reference switching state machine. 00 (default) = automatic mode. The fully automatic priority-based algorithm selects the active reference (Bits[1:0] are ignored). 01 = fallback mode. The active reference is the user reference (Bits[1:0]) as long as it is valid. Otherwise, use the fully automatic priority-based algorithm to select the active reference. 10 = holdover mode. The active reference is the user reference (Bits[1:0]) as long as it is valid. Otherwise, enter holdover mode. 11 = manual mode. The active reference is always the user reference (Bits[1:0]). When using manual mode, be sure that the reference declared as the user reference (Bits[1:0]) is programmed for manual reference-to-profile assignment in the appropriate manual reference profile selection register (Address 0x0503 and Address 0x0506). Unused. Write a 0 to this bit. Input reference when user selection mode = 01, 10, or 11. 00 (default) = Input Reference A. 01 = Input Reference AA. 10 = Input Reference B. 11 = Input Reference BB. Rev. 0 | Page 88 of 104 AD9547 Table 120. Cal/Sync Address 0x0A02 Bit [7:2] 1 Bit Name Unused Sync distribution 0 Calibrate SYSCLK Description Unused. Setting this bit (default = 0) initiates synchronization of the clock distribution output. When this bit = 1, the clock distribution output stalls. Synchronization occurs on the 1 to 0 transition of this bit. Setting this bit (default = 0) initiates an internal calibration of the SYSCLK PLL (assuming it is enabled). The calibration routine automatically selects the proper VCO frequency band and signal amplitude. The internal system clock stalls during the calibration procedure, disabling the device until calibration is complete (a few ms). Register 0x0A03—Reset Functions Table 121. Reset Functions 1 Address 0x0A03 1 Bit 7 6 5 4 3 2 Bit Name Unused Clear LF Clear CCI Clear phase accumulator Reset auto sync Reset TW history 1 Reset all IRQs 0 Reset watchdog Description Unused. Setting this bit (default = 0) clears the digital loop filter (intended as a debug tool). Setting this bit (default = 0) clears the CCI filter (intended as a debug tool). Setting this bit (default = 0) clears DDS phase accumulator (not a recommended action). Setting this bit (default = 0) resets the automatic synchronization logic (see Register 0x0403). Setting this bit (default = 0) resets the tuning word history logic (part of holdover functionality). Setting this bit (default = 0) clears the entire IRQ monitor register (Register 0x0D02 to Register 0x0D09). It is the equivalent of setting all the bits of the IRQ clearing register (Register 0x0A04 to Register 0x0A0B). Setting this bit (default = 0) resets the watchdog timer (see Register 0x0211 to Register 0x0212). If the timer times out, it simply starts a new timing cycle. If the timer has not yet timed out, it restarts at Time 0 without causing a timeout event. Continuously resetting the watchdog timer at intervals less than its timeout period prevents the watchdog timer from generating a timeout event. All bits in this register are autoclearing. Register 0x0A04 to Register 0x0A0B—IRQ Clearing The IRQ clearing registers are identical in format to the IRQ monitor registers (Address 0x0D02 to Address 0x0D09). When set to Logic 1, an IRQ clearing bit resets the corresponding IRQ monitor bit, thereby canceling the interrupt request for the indicated event. The IRQ clearing register is an autoclearing register. Table 122. IRQ Clearing for SYSCLK Address 0x0A04 Bit [7:6] 5 4 [3:2] 1 0 Bit Name Unused SYSCLK unlocked SYSCLK locked Unused SYSCLK cal complete SYSCLK cal started Description Unused. Clears SYSCLK unlocked IRQ. Clears SYSCLK locked IRQ. Unused. Clears SYSCLK calibration complete IRQ. Clears SYSCLK calibration started IRQ. Table 123. IRQ Clearing for Distribution Sync, Watchdog Timer, and EEPROM Address 0x0A05 Bit [7:4] 3 2 1 0 Bit Name Unused Distribution sync Watchdog timer EEPROM fault EEPROM complete Description Unused. Clears distribution sync IRQ. Clears watchdog timer IRQ. Clears EEPROM fault IRQ. Clears EEPROM complete IRQ. Rev. 0 | Page 89 of 104 AD9547 Table 124. IRQ Clearing for the Digital PLL Address 0x0A06 Bit 7 6 5 4 3 2 1 0 Bit Name Switching Closed Free run Holdover Frequency unlocked Frequency locked Phase unlocked Phase locked Description Clears switching IRQ. Clears closed IRQ. Clears free-run IRQ. Clears holdover IRQ. Clears frequency unlocked IRQ. Clears frequency locked IRQ. Clears phase unlocked IRQ. Clears phase locked IRQ. Table 125. IRQ Clearing for History Update, Frequency Limit, and Phase Slew Limit Address 0x0A07 Bit [7:5] 4 3 2 1 0 Bit Name Unused History updated Frequency unclamped Frequency clamped Phase slew unlimited Phase slew limited Description Unused. Clears history updated IRQ. Clears frequency unclamped IRQ. Clears frequency clamped IRQ. Clears phase slew unlimited IRQ. Clears phase slew limited IRQ. Table 126. IRQ Clearing for Reference Inputs Address 0x0A08 0x0A09 0x0A0A 0x0A0B Bit 7 6 5 4 3 2 1 0 7 6 5 4 3 2 1 0 [7:0] [7:0] Bit Name Ref AA new profile Ref AA validated Ref AA fault cleared Ref AA fault Ref A new profile Ref A validated Ref A fault cleared Ref A fault Ref BB new profile Ref BB validated Ref BB fault cleared Ref BB fault Ref B new profile Ref B validated Ref B fault cleared Ref B fault Unused Description Clears Ref AA new profile IRQ. Clears Ref AA validated IRQ. Clears Ref AA fault cleared IRQ. Clears Ref AA fault IRQ. Clears Ref A new profile IRQ. Clears Ref A validated IRQ. Clears Ref A fault cleared IRQ. Clears Ref A fault IRQ. Clears Ref BB new profile IRQ. Clears Ref BB validated IRQ. Clears Ref BB fault cleared IRQ. Clears Ref BB fault IRQ. Clears Ref B new profile IRQ. Clears Ref B validated IRQ. Clears Ref B fault cleared IRQ. Clears Ref B fault IRQ. Unused. Rev. 0 | Page 90 of 104 AD9547 Table 127. Incremental Phase Offset Control Address 0x0A0C Bit [7:3] 2 1 Bit Name Unused Reset phase offset Decrement phase offset 0 Increment phase offset Description Unused. Resets the incremental phase offset to 0. This is an autoclearing bit. Decrements the incremental phase offset by the amount specified in the incremental phase lock offset step size register (Register 0x0314 to Register 0x0315). This is an autoclearing bit. Increments the incremental phase offset by the amount specified in the incremental phase lock offset step size register (Register 0x0314 to Register 0x0315). This is an autoclearing bit. Table 128. Reference Profile Selection State Machine Startup 1 Address 0x0A0D 1 Bit [7:4] 3 2 1 0 Bit Name Unused Detect BB Detect B Detect AA Detect A Description Unused. Setting this bit starts the profile selection state machine for Input Reference BB. Setting this bit starts the profile selection state machine for Input Reference B. Setting this bit starts the profile selection state machine for Input Reference AA. Setting this bit starts the profile selection state machine for Input Reference A. All bits in this register are autoclearing. Register 0x0A0E to Register 0x0A10—Reference Validation Override Controls Table 129. Force Validation Timeout 1 Address 0x0A0E 1 Bit [7:4] 3 Bit Name Unused Force Timeout BB 2 Force Timeout B 1 Force Timeout AA 0 Force Timeout A Description Unused. Setting this bit emulates timeout of the validation timer for Reference BB. This is an autoclearing bit. Setting this bit emulates timeout of the validation timer for Reference B. This is an autoclearing bit. Setting this bit emulates timeout of the validation timer for Reference AA. This is an autoclearing bit. Setting this bit emulates timeout of the validation timer for Reference A. This is an autoclearing bit. All bits in this register are autoclearing. Table 130. Reference Monitor Override 1 Address 0x0A0F 1 Bit [7:4] 3 2 1 0 Bit Name Unused Ref Mon Override BB Ref Mon Override B Ref Mon Override AA Ref Mon Override A Description Unused. Overrides the reference monitor REF fault signal for Reference BB (default = 0, not overridden). Overrides the reference monitor REF fault signal for Reference B (default = 0, not overridden). Overrides the reference monitor REF fault signal for Reference AA (default = 0, not overridden). Overrides the reference monitor REF fault signal for Reference A (default = 0, not overridden). All bits in this register are autoclearing. Table 131. Reference Monitor Bypass 1 Address 0x0A10 1 Bit [7:4] 3 2 1 0 Bit Name Unused Ref Mon Bypass BB Ref Mon Bypass B Ref Mon Bypass AA Ref Mon Bypass A Description Unused. Bypasses the reference monitor for Reference BB (default = 0, not bypassed). Bypasses the reference monitor for Reference B (default = 0, not bypassed). Bypasses the reference monitor for Reference AA (default = 0, not bypassed). Bypasses the reference monitor for Reference A (default = 0, not bypassed). All bits in this register are autoclearing. Rev. 0 | Page 91 of 104 AD9547 STATUS READBACK (REGISTER 0x0D00 TO REGISTER 0x0D19) All bits in Register 0x0D00 to Register 0x0D19 are read only. These registers are accessible during EEPROM transactions. Table 132. EEPROM Status Address 0x0D00 Bit [7:3] 2 1 0 Bit Name Unused Fault detected Load in progress Save in progress Description Unused. An error occurred while saving data to or loading data from the EEPROM. The control logic sets this bit while data is being read from the EEPROM. The control logic sets this bit while data is being written to the EEPROM. Table 133. SYSCLK Status Address 0x0D01 Bit [7:5] 4 Bit Name Unused Stable [3:2] 1 0 Unused Cal in progress Lock detected Description Unused. The control logic sets this bit when the device considers the system clock to be stable (see the System Clock Stability Timer section). Unused. The control logic holds this bit set while the system clock calibration is in progress. Indicates the status of the system clock PLL. 0 = unlocked. 1 = locked (or the PLL is disabled). Register 0x0D02 to Register 0x0D09—IRQ Monitor If not masked via the IRQ mask register (Address 0x0209 to Address 0x0210), the appropriate IRQ monitor bit is set to a Logic 1 when the indicated event occurs. These bits can only be cleared via the IRQ clearing register (Address 0x0A04 to Address 0x0A0B), the reset all IRQs bit (Register 0x0A03, Bit 1), or a device reset. Table 134. IRQ Monitor for SYSCLK Address 0x0D02 Bit [7:6] 5 4 [3:2] 1 0 Bit Name Unused SYSCLK unlocked SYSCLK locked Unused SYSCLK cal complete SYSCLK cal started Description Unused. Indicates a SYSCLK PLL state transition from locked to unlocked. Indicates a SYSCLK PLL state transition from unlocked to locked. Unused. Indicates that SYSCLK calibration is complete. Indicates that SYSCLK calibration has begun. Table 135. IRQ Monitor for Distribution Sync, Watchdog Timer, and EEPROM Address 0x0D03 Bit [7:4] 3 2 1 0 Bit Name Unused Distribution sync Watchdog timer EEPROM fault EEPROM complete Description Unused. Indicates a distribution sync event. Indicates expiration of the watchdog timer. Indicates a fault during an EEPROM load or save operation. Indicates successful completion of an EEPROM load or save operation. Table 136. IRQ Monitor for the Digital PLL Address 0x0D04 Bit 7 6 5 4 3 2 1 0 Bit Name Switching Closed Free run Holdover Frequency unlocked Frequency locked Phase unlocked Phase locked Description Indicates that the DPLL is switching to a new reference. Indicates that the DPLL has entered closed-loop operation. Indicates that the DPLL has entered free-run mode. Indicates that the DPLL has entered holdover mode. Indicates that the DPLL lost frequency lock. Indicates that the DPLL has acquired frequency lock. Indicates that the DPLL lost phase lock. Indicates that the DPLL has acquired phase lock. Rev. 0 | Page 92 of 104 AD9547 Table 137. IRQ Monitor for History Update, Frequency Limit, and Phase Slew Limit Address 0x0D05 Bit [7:5] 4 3 2 1 0 Bit Name Unused History updated Frequency unclamped Frequency clamped Phase slew unlimited Phase slew limited Description Unused. Indicates the occurrence of a tuning word history update. Indicates a frequency limiter state transition from clamped to unclamped. Indicates a frequency limiter state transition from unclamped to clamped. Indicates a phase slew limiter state transition from slew limiting to not slew limiting. Indicates a phase slew limiter state transition from not slew limiting to slew limiting. Table 138. IRQ Monitor for Reference Inputs Address 0x0D06 0x0D07 0x0D08 0x0D09 Bit 7 6 5 4 3 2 1 0 7 6 5 4 3 2 1 0 [7:0] [7:0] Bit Name Ref AA new profile Ref AA validated Ref AA fault cleared Ref AA fault Ref A new profile Ref A validated Ref A fault cleared Ref A fault Ref BB new profile Ref BB validated Ref BB fault cleared Ref BB fault Ref B new profile Ref B validated Ref B fault cleared Ref B fault Unused Description Indicates that Ref AA has switched to a new profile. Indicates that Ref AA has been validated. Indicates that Ref AA has been cleared of a previous fault. Indicates that Ref AA has been faulted. Indicates that Ref A has switched to a new profile. Indicates that Ref A has been validated. Indicates that Ref A has been cleared of a previous fault. Indicates that Ref A has been faulted. Indicates that Ref BB has switched to a new profile. Indicates that Ref BB has been validated. Indicates that Ref BB has been cleared of a previous fault. Indicates that Ref BB has been faulted. Indicates that Ref B has switched to a new profile. Indicates that Ref B has been validated. Indicates that Ref B has been cleared of a previous fault. Indicates that Ref B has been faulted. Unused. Table 139. DPLL Status Address 0x0D0A 0x0D0B Bit 7 6 5 4 3 2 1 0 7 6 [5:3] Bit Name Offset slew limiting Phase build-out Frequency lock Phase lock Loop switching Holdover Active Free running Frequency clamped History available Active reference priority 2 [1:0] Unused Active reference Description The current closed-loop phase offset is rate limited. A phase build-out transition was made to the currently active reference. The DPLL has achieved frequency lock. The DPLL has achieved phase lock. The DPLL is in the process of a reference switchover. The DPLL is in holdover mode. The DPLL is active (that is, operating in a closed-loop condition). The DPLL is free running (that is, operating in an open-loop condition). The upper or lower frequency tuning word clamp is in effect. There is sufficient tuning word history available for holdover operation. Priority value of the currently active reference. 000 = highest priority. 111 = lowest priority. Unused. Index of the currently active reference. 00 = Reference A. 01 = Reference AA. 10 = Reference B. 11 = Reference BB. Rev. 0 | Page 93 of 104 AD9547 Table 140. Input Reference Status Address 0x0D0C 0x0D0D 0x0D0E 0x0D0F 0x0D10 0x0D11 0x0D12 0x0D13 Bit 7 [6:4] Bit Name Profile selected Selected profile 3 2 1 Valid Fault Fast 0 Slow [7:0] [7:0] [7:0] [7:0] [7:0] [7:0] [7:0] Unused Description The control logic sets this bit when it assigns Ref A to one of the eight profiles. The index (0 to 7) of the profile assigned to Ref A. Note that these bits are meaningless unless Bit 7 = 1. Ref A is valid for use (that is, it is unfaulted, and its validation timer has expired). Ref A is not valid for use. If Bit 7 = 1, this bit indicates that the frequency of Ref A is higher than allowed by its profile settings. If Bit 7 = 0, this bit indicates that the frequency of Ref A is above the maximum input reference frequency supported by the device. If Bit 7 = 1, this bit indicates that the frequency of Ref A is lower than allowed by its profile settings. If Bit 7 = 0, this bit indicates that the frequency of Ref A is below the minimum input reference frequency supported by the device. Same as 0x0D0C but for REF AA instead of REF A. Same as 0x0D0C but for REF B instead of REF A. Same as 0x0D0C but for REF BB instead of REF A. Unused. Table 141. Holdover History 1 Address 0x0D14 0x0D15 0x0D16 0x0D17 0x0D18 0x0D19 1 Bit [7:0] [7:0] [7:0] [7:0] [7:0] [7:0] Bit Name Tuning word history Description Tuning word readback, Bits[7:0]. Tuning word readback, Bits[15:8]. Tuning word readack, Bits[23:16]. Tuning word readback, Bits[31:24]. Tuning word readback, Bits[39:32]. Tuning word readback, Bits[47:40]. These registers contain the current 48-bit DDS frequency tuning word generated by the tuning word history logic. NONVOLATILE MEMORY (EEPROM) CONTROL (REGISTER 0x0E00 TO REGISTER 0x0E03) Table 142. Write Protect Address 0x0E00 Bit [7:2] 1 Bit Name Unused Half rate mode 0 Write enable Description Unused. EEPROM serial communication rate. 0 (default) = 400 kHz (normal). 1 = 200 kHz. EEPROM write enable/protect. 0 (default) = EEPROM is write protected. 1 = EEPROM is write enabled. Table 143. Condition Address 0x0E01 Bit [7:5] [4:0] Bit Name Unused Condition value Description Unused. When set to a nonzero value (default = 0), these bits establish the condition for EEPROM downloads. Bit Name Unused Save to EEPROM Description Unused. Upload data to the EEPROM based on in the EEPROM storage sequence. This is an autoclearing bit. Table 144. Save Address 0x0E02 Bit [7:1] 0 Rev. 0 | Page 94 of 104 AD9547 Table 145. Load Address 0x0E03 Bit [7:2] 1 0 Bit Name Unused Load from EEPROM Unused Description Unused. Download data from the EEPROM. This is an autoclearing bit. Unused. EEPROM STORAGE SEQUENCE (REGISTER 0x0E10 TO REGISTER 0x0E3F) The default settings of Register 0x0E10 to Register 0x0E33 embody a sample scratch pad instruction sequence. The following is a description of the register defaults under the assumption that the controller has been instructed to carry out an EEPROM storage sequence. Table 146. EEPROM Storage Sequence for System Clock Settings Address 0x0E10 Bit [7:0] Bit Name System clock 0x0E11 0x0E12 [7:0] [7:0] System clock 0x0E13 [7:0] I/O update Description The default value of this register is 0x08, which the controller interprets as a data instruction. Its decimal value is 8, which tells the controller to transfer nine bytes of data (8 + 1) beginning at the address specified by the next two bytes. The controller stores 0x08 in the EEPROM and increments the EEPROM address pointer. The default value of these two registers is 0x0100. Note that Register 0x0E11 and Register 0x0E12 are the most significant and least significant bytes of the target address, respectively. Because the previous register contains a data instruction, these two registers define a starting address (in this case, 0x0100). The controller stores 0x0100 in the EEPROM and increments the EEPROM pointer by 2. It then transfers nine bytes from the register map (beginning at Address 0x0100) to the EEPROM and increments the EEPROM address pointer by 10 (nine data bytes and one checksum byte). The nine bytes transferred correspond to the system clock parameters in the register map. The default value of this register is 0x80, which the controller interprets as an I/O update instruction. The controller stores 0x80 in the EEPROM and increments the EEPROM address pointer. Table 147. EEPROM Storage Sequence for System Clock Calibration Address 0x0E14 Bit [7:0] Bit Name SYSCLK calibrate Description The default value of this register is 0xA0, which the controller interprets as a calibrate instruction. The controller stores 0xA0 in the EEPROM and increments the EEPROM address pointer. Table 148. EEPROM Storage Sequence for General Configuration Settings Address 0x0E15 Bit [7:0] Bit Name General Description The default value of this register is 0x14, which the controller interprets as a data instruction. Its decimal value is 20, which tells the controller to transfer 21 bytes of data (20 + 1) beginning at the address specified by the next two bytes. The controller stores 0x14 in the EEPROM and increments the EEPROM address pointer. 0x0E16 0x0E17 [7:0] [7:0] General The default value of these two registers is 0x0200. Note that Register 0x0E16 and Register 0x0E17 are the most significant and least significant bytes of the target address, respectively. Because the previous register contains a data instruction, these two registers define a starting address (in this case, 0x0200). The controller stores 0x0200 in the EEPROM and increments the EEPROM pointer by 2. It then transfers 21 bytes from the register map (beginning at Address 0x0200) to the EEPROM and increments the EEPROM address pointer by 22 (21 data bytes and one checksum byte). The 21 bytes transferred correspond to the general configuration parameters in the register map. Table 149. EEPROM Storage Sequence for DPLL Settings Address 0x0E18 Bit [7:0] Bit Name DPLL 0x0E19 0x0E1A [7:0] [7:0] DPLL Description The default value of this register is 0x1B, which the controller interprets as a data instruction. Its decimal value is 27, which tells the controller to transfer 28 bytes of data (27 + 1) beginning at the address specified by the next two bytes. The controller stores 0x1B in the EEPROM and increments the EEPROM address pointer. The default value of these two registers is 0x0300. Note that Register 0x0E19 and Register 0x0E1A are the most significant and least significant bytes of the target address, respectively. Because the previous register contains a data instruction, these two registers define a starting address (in this case, 0x0300). The controller stores 0x0300 in the EEPROM and increments the EEPROM pointer by 2. It then transfers 28 bytes from the register map (beginning at Address 0x0300) to the EEPROM and increments the EEPROM address pointer by 29 (28 data bytes and one checksum byte). The 28 bytes transferred correspond to the DPLL parameters in the register map. Rev. 0 | Page 95 of 104 AD9547 Table 150. EEPROM Storage Sequence for Clock Distribution Settings Address 0x0E1B Bit [7:0] Bit Name Clock distribution 0x0E1C 0x0E1D [7:0] [7:0] Clock distribution 0x0E1E [7:0] I/O update Description The default value of this register is 0x19, which the controller interprets as a data instruction. Its decimal value is 25; this tells the controller to transfer 26 bytes of data (25 + 1) beginning at the address specified by the next two bytes. The controller stores 0x19 in the EEPROM and increments the EEPROM address pointer. The default value of these two registers is 0x0400. Note that Register 0x0E1C and Register 0x0E1D are the most significant and least significant bytes of the target address, respectively. Because the previous register contains a data instruction, these two registers define a starting address (in this case, 0x0400). The controller stores 0x0400 in the EEPROM and increments the EEPROM pointer by 2. It then transfers 26 bytes from the register map (beginning at Address 0x0400) to the EEPROM and increments the EEPROM address pointer by 27 (26 data bytes and one checksum byte). The 26 bytes transferred correspond to the clock distribution parameters in the register map. The default value of this register is 0x80, which the controller interprets as an I/O update instruction. The controller stores 0x80 in the EEPROM and increments the EEPROM address pointer. Table 151. EEPROM Storage Sequence for Reference Input Settings Address 0x0E1F Bit [7:0] Bit Name Reference inputs 0x0E20 0x0E21 [7:0] [7:0] Reference inputs Description The default value of this register is 0x07, which the controller interprets as a data instruction. Its decimal value is 7, which tells the controller to transfer eight bytes of data (7 + 1), beginning at the address specified by the next two bytes. The controller stores 0x07 in the EEPROM and increments the EEPROM address pointer. The default value of these two registers is 0x0500. Note that Register 0x0E20 and Register 0x0E21 are the most significant and least significant bytes of the target address, respectively. Because the previous register contains a data instruction, these two registers define a starting address (in this case, 0x0500). The controller stores 0x0500 in the EEPROM and increments the EEPROM pointer by 2. It then transfers eight bytes from the register map (beginning at Address 0x0500) to the EEPROM and increments the EEPROM address pointer by nine (eight data bytes and one checksum byte). The eight bytes transferred correspond to the reference inputs parameters in the register map. Table 152. EEPROM Storage Sequence for Profile 0 and Profile 1 Settings Address 0x0E22 Bit [7:0] Bit Name Profile 0 and Profile 1 0x0E23 0x0E24 [7:0] [7:0] Profile 0 and Profile 1 Description The default value of this register is 0x63, which the controller interprets as a data instruction. Its decimal value is 99, which tells the controller to transfer 100 bytes of data (99 + 1), beginning at the address specified by the next two bytes. The controller stores 0x63 in the EEPROM and increments the EEPROM address pointer. The default value of these two registers is 0x0600. Note that Register 0x0E23 and Register 0x0E24 are the most significant and least significant bytes of the target address, respectively. Because the previous register contains a data instruction, these two registers define a starting address (in this case, 0x0600). The controller stores 0x0600 in the EEPROM and increments the EEPROM pointer by 2. It then transfers 100 bytes from the register map (beginning at Address 0x0600) to the EEPROM and increments the EEPROM address pointer by 101 (100 data bytes and one checksum byte). The 99 bytes transferred correspond to the Profile 0 and Profile 1 parameters in the register map. Table 153. EEPROM Storage Sequence for Profile 2 and Profile 3 Settings Address 0x0E25 Bit [7:0] Bit Name Profile 2 and Profile 3 0x0E26 0x0E27 [7:0] [7:0] Profile 2 and Profile 3 Description The default value of this register is 0x63, which the controller interprets as a data instruction. Its decimal value is 99, which tells the controller to transfer 100 bytes of data (99 + 1), beginning at the address specified by the next two bytes. The controller stores 0x63 in the EEPROM and increments the EEPROM address pointer. The default value of these two registers is 0x0680. Note that Register (0x0E26 and Register 0x0E27 are the most significant and least significant bytes of the target address, respectively. Because the previous register contains a data instruction, these two registers define a starting address (in this case, 0x0680). The controller stores 0x0680 in the EEPROM and increments the EEPROM pointer by 2. It then transfers 100 bytes from the register map (beginning at Address 0x0680) to the EEPROM and increments the EEPROM address pointer by 101 (100 data bytes and one checksum byte). The 99 bytes transferred correspond to the Profile 2 and Profile 3 parameters in the register map. Rev. 0 | Page 96 of 104 AD9547 Table 154. EEPROM Storage Sequence for Profile 4 and Profile 5 Settings Address 0x0E28 Bit [7:0] Bit Name Profile 4 and Profile 5 Description The default value of this register is 0x63, which the controller interprets as a data instruction. Its decimal value is 99, which tells the controller to transfer 100 bytes of data (99 + 1) beginning at the address specified by the next two bytes. The controller stores 0x63 in the EEPROM and increments the EEPROM address pointer. 0x0E29 0x0E2A [7:0] [7:0] Profile 4 and Profile 5 The default value of these two registers is 0x0700. Note that Register 0x0E29 and Register 0x0E2A are the most significant and least significant bytes of the target address, respectively. Because the previous register contains a data instruction, these two registers define a starting address (in this case, 0x0700). The controller stores 0x0700 in the EEPROM and increments the EEPROM pointer by 2. It then transfers 100 bytes from the register map (beginning at Address 0x0700) to the EEPROM and increments the EEPROM address pointer by 101 (100 data bytes and one checksum byte). The 99 bytes transferred correspond to the Profile 4 and Profile 5 parameters in the register map. Table 155. EEPROM Storage Sequence for Profile 6 and Profile 7 Settings Address 0x0E2B Bit [7:0] Bit Name Profile 6 and Profile 7 0x0E2C 0x0E2D [7:0] [7:0] Profile 6 and Profile 7 0x0E2E [7:0] I/O update Description The default value of this register is 0x63, which the controller interprets as a data instruction. Its decimal value is 99, this tells the controller to transfer 100 bytes of data (99 + 1) beginning at the address specified by the next two bytes. The controller stores 0x63 in the EEPROM and increments the EEPROM address pointer. The default value of these two registers is 0x0780. Note that Register 0x0E2C and Register 0x0E2D are the most significant and least significant bytes of the target address, respectively. Because the previous register contains a data instruction, these two registers define a starting address (in this case, 0x0780). The controller stores 0x0780 in the EEPROM and increments the EEPROM pointer by 2. It then transfers 100 bytes from the register map (beginning at Address 0x0780) to the EEPROM and increments the EEPROM address pointer by 101 (100 data bytes and one checksum byte). The 99 bytes transferred correspond to the Profile 6 and Profile 7 parameters in the register map. The default value of this register is 0x80, which the controller interprets as an I/O update instruction. The controller stores 0x80 in the EEPROM and increments the EEPROM address pointer. Table 156. EEPROM Storage Sequence for Operational Control Settings Address 0x0E2F Bit [7:0] Bit Name Operational controls 0x0E30 0x0E31 [7:0] [7:0] Operational controls 0x0E32 [7:0] I/O update Description The default value of this register is 0x10, which the controller interprets as a data instruction. Its decimal value is 16, this tells the controller to transfer 17 bytes of data (16 + 1) beginning at the address specified by the next two bytes. The controller stores 0x10 in the EEPROM and increments the EEPROM address pointer. The default value of these two registers is 0x0A00. Note that Register 0x0E30 and Register 0x0E31 are the most significant and least significant bytes of the target address, respectively. Because the previous register contains a data instruction, these two registers define a starting address (in this case, 0x0A00). The controller stores 0x0A00 in the EEPROM and increments the EEPROM pointer by 2. It then transfers 17 bytes from the register map (beginning at Address 0x0A00) to the EEPROM and increments the EEPROM address pointer by 18 (17 data bytes and one checksum byte). The 17 bytes transferred correspond to the operational controls parameters in the register map. The default value of this register is 0x80, which the controller interprets as an I/O update instruction. The controller stores 0x80 in the EEPROM and increments the EEPROM address pointer. Table 157. EEPROM Storage Sequence for End of Data Address 0x0E33 Bit [7:0] Bit Name End of data Description The default value of this register is 0xFF, which the controller interprets as an end instruction. The controller stores this instruction in the EEPROM, resets the EEPROM address pointer, and enters an idle state. Note that, if this were a pause rather than an end instruction, the controller actions would be the same except that the controller would not reset the EEPROM address pointer. Rev. 0 | Page 97 of 104 AD9547 APPLICATIONS INFORMATION POWER SUPPLY PARTITIONS THERMAL PERFORMANCE The AD9547 features multiple power supplies, and their power consumption varies with the AD9547 configuration. This section provides information about which power supplies can be grouped together and how the power consumption of each block varies with frequency. The AD9547 is specified for a case temperature (TCASE). To ensure that TCASE is not exceeded, an airflow source can be used. Use the following equation to determine the junction temperature on the application PCB: The numbers quoted in this section are for comparison only. Refer to the Specifications section for exact numbers. With each group, bypass capacitors of 1 μF in parallel with 10 μF should be used. where: TJ is the junction temperature in degrees Celsius (°C). TCASE is the case temperature in degrees Celsius (°C) measured by the customer at the top center of the package. ΨJT is the value that is indicated in Table 159. PD is the power dissipation (see the Power Dissipation section). Upon applying power to the device, internal circuitry monitors the 1.8 V digital core supply and the 3.3 V digital I/O supply. When these supplies cross the desired threshold level, the device generates an internal 10 μs reset pulse. This pulse does not appear on the RESET pin. 3.3 V Supplies The 3.3 V supply domain consists of two main partitions: digital (DVDD3) and analog (AVDD3). These two supply domains must be kept separate. Furthermore, the AVDD3 consists of two subdomains: the clock distribution output domain (Pin 25, Pin 31) and the rest of the AVDD3 supply connections. Generally, the ADD3 supply domains can be joined together. However, if an application requires 1.8 V CMOS driver operation in the clock distribution output block, provide one 1.8 V supply domain to power the clock distribution output block. Each output driver has a dedicated supply pin, as shown in Table 158. TJ = TCASE + (ΨJT × PD) Values of θJA are provided for package comparison and PCB design considerations. θJA can be used for a first-order approximation of TJ using the following equation: TJ = TA + (θJA × PD) where TA is the ambient temperature in degrees Celsius (°C). Values of θJC are provided for package comparison and PCB design considerations when an external heat sink is required. Values of θJB are provided for package comparison and PCB design considerations. Table 158. Output Driver Supply Pins Output Driver OUT0 OUT1 Supply Pin 25 31 1.8 V Supplies The 1.8 V supply domain consists of two main partitions: digital (DVDD) and analog (AVDD). These two supply domains must be kept separate. Table 159. Thermal Parameters for the AD9547 64-Lead LFCSP Package Symbol θJA θJMA θJMA θJB θJC ΨJT 1 2 Thermal Characteristic Using a JEDEC51-7 Plus JEDEC51-5 2S2P Test Board 1 Junction-to-ambient thermal resistance, 0.0 m/sec air flow per JEDEC JESD51-2 (still air) Junction-to-ambient thermal resistance, 1.0 m/sec air flow per JEDEC JESD51-6 (moving air) Junction-to-ambient thermal resistance, 2.5 m/sec air flow per JEDEC JESD51-6 (moving air) Junction-to-board thermal resistance, per JEDEC JESD51-8 (still air) Junction-to-case thermal resistance Junction-to-top-of-package characterization parameter, 0 m/sec airflow per JEDEC JESD51-2 (still air) Value 2 21.7 18.9 16.9 11.3 1.2 0.1 The exposed pad on the bottom of the package must be soldered to ground to achieve the specified thermal performance. Results are from simulations. The PCB is a JEDEC multilayer type. Thermal performance for actual applications requires careful inspection of the conditions in the application to determine whether they are similar to those assumed in these calculations. Rev. 0 | Page 98 of 104 Unit °C/W °C/W °C/W °C/W °C/W °C/W AD9547 CALCULATING THE DIGITAL FILTER COEFFICIENTS The digital loop filter coefficients (α, β, γ, and δ, as shown in Figure 38) relate to the time constants (T1, T2, and T3) that are associated with the equivalent analog circuit for a third-order loop filter (see Figure 64). FROM CHARGE PUMP R2 C1 C3 C2 08300-042 30,517,578,125 fS 233 D=S+ The design process begins by deciding on two design parameters related to the second-order loop filter shown in Figure 65: the desired open-loop bandwidth (fP) and the phase margin (θ). U +1 V where S, U, and V are the integer and fractional feedback divider values that reside in the profile registers. Keep in mind that the desired integer feedback divide ratio is one more than the stored value of S (hence, the +1 term in the equation for D in this equation). This leads to the digital filter coefficients given by R2 TO VCO C1 2 K= Figure 64. Third-Order Analog Loop Filter FROM CHARGE PUMP 1 ωC (T1 + T3 ) T2 = Calculation of the digital loop filter coefficients requires a scaling constant, K (related to the system clock frequency, fS), and the PLL feedback divide ratio, D. TO VCO R3 It can also be shown that the adjusted open-loop bandwidth leads to T2 (the secondary time constant of the second-order loop filter), which is expressed as 08300-043 C2 ωC 2T2 D T1K β= − 32 ⎛ 1 1 ⎞ ⎜ − ⎟ f S ⎜⎝ T1 T2 ⎟⎠ γ= −32 f ST1 δ= 32 f ST3 Figure 65. Second-Order Analog Loop Filter An analysis of the second-order loop filter leads to its primary time constant, T1. It can be shown that T1 is expressible in terms of fP and θ as 1 − sin(θ) T1 = where ωp = 2πfP ωP cos(θ) An analysis of the third-order loop filter leads to the definition of another time constant, T3. It can be shown that T3 is expressible in terms of the desired amount of additional attenuation introduced by R3 and C3 at some specified frequency offset (fOFFSET) from the PLL output frequency. T3 = ATTEN 10 10 ωOFFSET 1 + (ωCT2 )2 Calculation of the coefficient register values requires the application of some special functions, which are described as follows: The if() function y = if (test_statement, true_value, false_value) −1 where ωOFFSET = 2πfOFFSET Note that ATTEN is the desired excess attenuation in decibels (dB). Furthermore, ATTEN and ωOFFSET should be chosen so that 1 T3 ≤ 5 fP (T1 + T3 )tan(θ) ⎡⎢ T1T3 + (T1 + T3 )2 ⎢⎣ where: test_statement is a conditional expression (for example, x < 3). true_value is what y equals if the conditional expression is true. false_value is what y equals if the conditional expression is false. The round() function y = round(x) With an expression for T1 and T3, it is possible to define an adjusted open-loop bandwidth (fC) that is slightly less than fP. It can be shown that ωC (fC expressed as a radian frequency) is expressible in terms of T1, T3, and θ (phase margin) as follows: ωC = (1 + (ωCT1 )2 )(1 + (ωCT3 )2 ) α= 1+ ⎤ T1T3 + (T1 + T3 )2 − 1⎥ 2 [(T1 + T3 )tan(θ)] ⎥⎦ If x is an integer, then y = x. Otherwise, y is the nearest integer to x. For example, round(2.1) = 2, round(2.5) = 3, and round(−3.1) = −3. The ceil() function y = ceil(x) If x is an integer, then y = x. Otherwise, y is the next integer to the right on the number line. For example, ceil(2.8) = 3, whereas ceil(−2.8) = −2. Rev. 0 | Page 99 of 104 AD9547 Calculation of the α Register Values The min() function The quantized α coefficient consists of four components: α0, α1, α2, and α3, according to y = min(x0, x1, ... xn) where: x0 through xn is a list of real numbers. y is the number in the list that is the farthest to the left on the number line. α ≈ αquantized = α0 × 216 − α 1 The max() function y = max(x0, x1, ... xn) where: x0 through xn is a list of real numbers. y is the number in the list that is the farthest to the right on the number line. where: α0, α1, α2, and α3 are the register values. α2 provides front-end gain. α3 provides back-end gain. α1 shifts the binary decimal point of α0 to the left to accommodate small values of α. Calculation of α1 is a two-step process, as follows: w = if(α <1, −ceil(log2(α)), 0) α1 = if(α <1, min[63, max(0, w)], 0) The log2() function log2(x) = + α2 + α3 If gain is necessary (that is, α > 1), then it is beneficial to apply most or all of it to the front-end gain (α2) implying that the calculation of α2 is to be done before that of α3. Calculation of α2 is a three-step process that leads directly to the calculation of α3. ln(x ) ln(2) where: ln() is the natural log function. x is a positive, nonzero number. Assume that the coefficient calculations for α, β, γ, and δ above yield the following results: x = if(α > 1, ceil(log2(α)), 0) y = if(α > 1, min[22, max(0, x)], 0) α = 0.012735446 α2 = if(y ≥ 8, 7, y) β = −6.98672 × 10−5 α3 = if(y ≥ 8, y – 7, 0) Calculation of α0 is a two-step process, as follows: γ = −7.50373 × 10−5 z = round(α × 216 + α δ = 0.002015399 1 These values are floating point numbers that must be quantized according to the bit widths of the linear and exponential components of the coefficients as they appear in the register map. Note that the calculations that follow indicate a positive value for the register entries of β and γ. The reason is that β and γ, which are supposed to be negative values, are stored in the AD9547 registers as positive values. The AD9547 converts the stored values to negative numbers within its signal processing core. A detailed description of the register value computations for α, β, γ, and δ follows. − α2 − α3 ) α0 = min[65535, max(1, z) Using the example value of α = 0.012735446 yields w = 6, so α1 = 6 x = 0 and y = 0, so α2 = 0 and α3 = 0 z = 53416.332099584, so α0 = 53416 This leads to the following quantized value, which is very close to the desired value of 0.012735446: αquantized = 53416 × 2−22 ≈ 0.01273566821 Rev. 0 | Page 100 of 104 AD9547 Calculation of the β Register Values Using the example value of −γ = 7.50373 × 10−5 yields The quantized β coefficient consists of two components, β0 and β1, according to −β ≈ βquantized = β0 × 2−(17 + β ) x = 13, so γ1 = 13 y = 80570.6873700352, so γ0 = 80571 1 This leads to the following quantized value, which is very close to the desired value of 7.50373 × 10−5: where β0 and β1 are the register values. Calculation of β1 is a two-step process that leads to the calculation of β0, which is also a two-step process. γquantized = 80571 × 2−30 ≈ 7.503759116 × 10−5 Calculation of the δ Register Values x = −ceil(log2( β )) The quantized δ coefficient consists of two components, δ0 and δ1, according to β1 = min[31, max(0, x)] δ ≈ δquantized = δ0 × 2−(15 + δ ) y = round( β × 217 + β ) 1 1 where δ0 and δ1 are the register values. β0 = min[131071, max(1, y)] Calculation of δ1 is a two-step process that leads to the calculation of δ0, which is also a two-step process. Using the example value of −β = 6.98672 × 10−5 yields x = 13, so β1 = 13 x = −ceil(log2(δ)) y = 75019.3347657728, so β0 = 75019 δ1 = min[31, max(0, x)] This leads to the following quantized value, which is very close to the desired value of 6.98672 × 10−5: Using the example value of δ = 0.002015399, the preceding formulas yield Calculation of the γ Register Values The quantized γ coefficient consists of two components, γ0 and γ1, according to 1 where γ0 and γ1 are the register values. Calculation of γ1 is a two-step process that leads to the calculation of γ0, which is also a two-step process. 1 δ0 = min[32767, max(1, y)] βquantized = 75019 × 2−30 ≈ 6.986688823 × 10−5 −γ ≈ γquantized = γ0 × 2−(17 + γ ) y = round(δ × 215 + δ ) x = 8, so δ1 = 8 y = 16906.392174592, so δ0 = 16906 This leads to the following quantized value, which is very close to the desired value of 0.002015399: δquantized = 16906 × 2−23 ≈ 0.002015352249 x = −ceil(log2( γ )) y1 = min[31, max(0, x)] y = round( γ × 217 + γ ) 1 γ0 = min[131071, max(1, y)] Rev. 0 | Page 101 of 104 AD9547 OUTLINE DIMENSIONS 0.60 MAX 9.00 BSC SQ 0.60 MAX 64 49 48 PIN 1 INDICATOR 1 PIN 1 INDICATOR 8.75 BSC SQ 0.50 BSC 0.50 0.40 0.30 1.00 0.85 0.80 33 32 16 17 0.25 MIN 7.50 REF 0.80 MAX 0.65 TYP 12° MAX 0.05 MAX 0.02 NOM SEATING PLANE 0.30 0.23 0.18 6.35 6.20 SQ 6.05 EXPOSED PAD (BOTTOM VIEW) 0.20 REF FOR PROPER CONNECTION OF THE EXPOSED PAD, REFER TO THE PIN CONFIGURATION AND FUNCTION DESCRIPTIONS SECTION OF THIS DATA SHEET. COMPLIANT TO JEDEC STANDARDS MO-220-VMMD-4 091707-C TOP VIEW Figure 66. 64-Lead Lead Frame Chip Scale Package [LFCSP_VQ] 9 mm × 9 mm Body, Very Thin Quad (CP-64-4) Dimensions shown in millimeters ORDERING GUIDE Model AD9547BCPZ 1 AD9547BCPZ-REEL71 AD9547/PCBZ1 1 Temperature Range −40°C to +85°C −40°C to +85°C Package Description 64-Lead Lead Frame Chip Scale Package [LFCSP_VQ] 64-Lead Lead Frame Chip Scale Package [LFCSP_VQ] Evaluation Board Z = RoHS Compliant Part. Rev. 0 | Page 102 of 104 Package Option CP-64-4 CP-64-4 AD9547 NOTES Rev. 0 | Page 103 of 104 AD9547 NOTES ©2009 Analog Devices, Inc. All rights reserved. Trademarks and registered trademarks are the property of their respective owners. D08300-0-7/09(0) Rev. 0 | Page 104 of 104