FEATURES FUNCTIONAL BLOCK DIAGRAM REFSEL XT1 XT2 XTAL OSC CMOS REFOUT REFCLK DIVIDE 1 OR 2 VCO 2.15GHz TO 2.55GHz FEEDBACK DIVIDER THIRD ORDER LPF PLL2 LVPECL/LVDS OR 2 x CMOS LVPECL/LVDS OR 2 x CMOS LDO VCO 2.15GHz TO 2.55GHz FEEDBACK DIVIDER SCL SDA DIVIDERS DIVIDERS LDO I2C CONTROL DIVIDERS DIVIDERS THIRD ORDER LPF fPFD PLL1 PFD/CP Fully integrated dual PLL/VCO cores 1 integer-N and 1 fractional-N PLL Continuous frequency coverage from 11.2 MHz to 200 MHz Most frequencies from 200 MHz to 637.5 MHz available PLL1 phase jitter (12 kHz to 20 MHz): 460 fs rms typical PLL2 phase jitter (12 kHz to 20 MHz) Integer-N mode: 470 fs rms typical Fractional-N mode: 660 fs rms typical Input crystal or reference clock frequency Optional reference frequency divide-by-2 I2C programmable output frequencies Up to 4 LVDS/LVPECL or up to 8 LVCMOS output clocks 1 CMOS buffered reference clock output Spread spectrum: downspread [0, -0.5]% 2 pin-controlled frequency maps: margining Integrated loop filters Space saving, 6 mm x 6 mm, 40-lead LFCSP package 1.02 W power dissipation (LVDS operation) 1.235 W power dissipation (LVPECL operation) 3.3 V operation PFD/CP Data Sheet Clock Generator with Dual PLLs, Spread Spectrum, and Margining AD9577 LVPECL/LVDS OR 2 x CMOS LVPECL/LVDS OR 2 x CMOS Low jitter, low phase noise multioutput clock generator for data communications applications including Ethernet, Fibre Channel, SONET, SDH, PCI-e, SATA, PTN, OTN, ADC/DAC, and digital video Spread spectrum clocking GENERAL DESCRIPTION The AD9577 provides a multioutput clock generator function, along with two on-chip phase-locked loop cores, PLL1 and PLL2, optimized for network clocking applications. The PLL designs are based on the Analog Devices, Inc., proven portfolio of high performance, low jitter frequency synthesizers to maximize network performance. The PLLs have I2C programmable output frequencies and formats. The fractional-N PLL can support spread spectrum clocking for reduced EMI radiated peak power. Both PLLs can support frequency margining. Other applications with demanding phase noise and jitter requirements can benefit from this part. The first integer-N PLL section (PLL1) consists of a low noise phase frequency detector (PFD), a precision charge pump (CP), a low phase noise voltage controlled oscillator (VCO), a programmable SSCG MAX_BW SPREAD SPECTRUM, SDM AD9577 09284-001 MARGIN APPLICATIONS Figure 1. feedback divider, and two independently programmable output dividers. By connecting an external crystal or applying a reference clock to the REFCLK pin, frequencies of up to 637.5 MHz can be synchronized to the input reference. Each output divider and feedback divider ratio is I2C programmed for the required output rates. A second fractional-N PLL (PLL2) with a programmable modulus allows VCO frequencies that are fractional multiples of the reference frequency to be synthesized. Each output divider and feedback divider ratio can be programmed for the required output rates, up to 637.5 MHz. This fractional-N PLL can also operate in integer-N mode for the lowest jitter. Up to four differential output clock signals can be configured as either LVPECL or LVDS signaling formats. Alternatively, the outputs can be configured for up to eight CMOS outputs. Combinations of these formats are supported. No external loop filter components are required, thus conserving valuable design time and board space. The AD9577 is available in a 40-lead, 6 mm x 6 mm LFCSP package and can operate from a single 3.3 V supply. The operating temperature range is -40C to +85C. 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 (c)2011 Analog Devices, Inc. All rights reserved. AD9577 Data Sheet TABLE OF CONTENTS Features .............................................................................................. 1 Example Application.................................................................. 28 Applications....................................................................................... 1 Functional Description.................................................................. 29 General Description ......................................................................... 1 Reference Input and Reference Dividers................................. 29 Functional Block Diagram .............................................................. 1 Output Channel Dividers.......................................................... 30 Revision History ............................................................................... 2 Outputs ........................................................................................ 30 Specifications..................................................................................... 3 Reference Output Buffer ........................................................... 31 PLL1 Characteristics .................................................................... 3 PLL1 Integer-N PLL................................................................... 31 PLL1 Clock Output Jitter............................................................. 5 PLL1 Phase Frequency Detector (PFD) and Charge Pump . 32 PLL2 Fractional-N Mode Characteristics ................................. 6 PLL1 VCO ................................................................................... 32 PLL2 Integer-N Mode Characteristics....................................... 7 PLL1 Feedback Divider ............................................................. 32 PLL2 Clock Output Jitter............................................................. 9 Setting the Output Frequency of PLL1.................................... 32 CMOS Reference Clock Output Jitter...................................... 11 PLL2 Integer/Fractional-N PLL ............................................... 32 Timing Characteristics .............................................................. 12 PLL2 Phase Frequency Detector (PFD) and Charge Pump . 33 Clock Outputs ............................................................................. 13 PLL2 Loop Bandwidth............................................................... 33 Power............................................................................................ 14 PLL2 VCO ................................................................................... 33 Crystal Oscillator........................................................................ 15 PLL2 Feedback Divider ............................................................. 33 Reference Input........................................................................... 15 PLL2 - Modulator ................................................................. 33 Control Pins ................................................................................ 15 Spur Mechanisms ....................................................................... 33 Absolute Maximum Ratings.......................................................... 16 Optimizing PLL Performance .................................................. 34 Thermal Characteristics ............................................................ 16 Setting the Output Frequency of PLL2.................................... 34 ESD Caution................................................................................ 16 Margining .................................................................................... 35 Pin Configuration and Function Descriptions........................... 17 Spread Spectrum Clock Generation (SSCG).......................... 35 2 Typical Performance Characteristics ........................................... 19 I C Interface Timing and Internal Register Description........... 38 REFOUT and PLL1 Phase Noise Performance ...................... 19 Default Frequency Map and Output Formats ........................ 40 PLL2 Phase Noise Performance................................................ 20 I2C Interface Operation ............................................................. 40 Output Jitter ................................................................................ 21 Typical Application Circuits ..................................................... 42 Typical Output Signal ................................................................ 22 Typical Spread Spectrum Performance Characteristics ........ 24 Power and Grounding Considerations and Power Supply Rejection...................................................................................... 43 Terminology .................................................................................... 25 Outline Dimensions ....................................................................... 44 Detailed Block Diagram ................................................................ 27 Ordering Guide .......................................................................... 44 REVISION HISTORY 10/11--Revision 0: Initial Version Rev. 0 | Page 2 of 44 Data Sheet AD9577 SPECIFICATIONS Typical (typ) is given for VS = 3.3 V, TA = 25C, unless otherwise noted. Minimum (min) and maximum (max) values are given over full VS (3.0 V to 3.6 V) and TA (-40C to +85C) variation. AC coupling capacitors of 0.1 F used where appropriate. A Fox Electronics FX532A 25 MHz crystal is used throughout, unless otherwise stated. PLL1 CHARACTERISTICS Table 1. Parameter NOISE CHARACTERISTICS Phase Noise (106.25 MHz LVPECL Output) At 1 kHz At 10 kHz At 100 kHz At 1 MHz At 10 MHz At 30 MHz Phase Noise (156.25 MHz LVPECL Output) At 1 kHz At 10 kHz At 100 kHz At 1 MHz At 10 MHz At 30 MHz Phase Noise (625 MHz LVPECL Output) At 1 kHz At 10 kHz At 100 kHz At 1 MHz At 10 MHz At 30 MHz Phase Noise (106.25 MHz LVDS Output) At 1 kHz At 10 kHz At 100 kHz At 1 MHz At 10 MHz At 30 MHz Phase Noise (156.25 MHz LVDS Output) At 1 kHz At 10 kHz At 100 kHz At 1 MHz At 10 MHz At 30 MHz Phase Noise (625 MHz LVDS Output) At 1 kHz At 10 kHz At 100 kHz At 1 MHz At 10 MHz At 30 MHz Min Typ Max Unit Test Conditions/Comments 1 Na = 102, Vx = 4, Dx = 6, fPFD = 25 MHz -121 -127 -128 -150 -156 -158 dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz -117 -124 -124 -147 -156 -156 dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz -105 -112 -112 -135 -150 -150 dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz -119 -127 -128 -148 -156 -156 dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz Na = 100, Vx = 4, Dx = 4, fPFD = 25 MHz Na = 100, Vx = 2, Dx = 2, fPFD = 25 MHz Na = 102, Vx = 4, Dx = 6, fPFD = 25 MHz Na = 100, Vx = 4, Dx = 4, fPFD = 25 MHz -116 -124 -124 -145 -155 -155 dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz -104 -111 -112 -134 -149 -149 dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz Na = 100, Vx = 2, Dx = 2, fPFD = 25 MHz Rev. 0 | Page 3 of 44 AD9577 Parameter Phase Noise (106.25 MHz CMOS Output) At 1 kHz At 10 kHz At 100 kHz At 1 MHz At 10 MHz At 30 MHz Phase Noise (156.25 MHz CMOS Output) At 1 kHz At 10 kHz At 100 kHz At 1 MHz At 10 MHz At 30 MHz Phase Noise (155.52 MHz LVPECL Output) At 1 kHz At 10 kHz At 100 kHz At 1 MHz At 10 MHz At 30 MHz Phase Noise (622.08 MHz LVPECL Output) At 1 kHz At 10 kHz At 100 kHz At 1 MHz At 10 MHz At 30 MHz Phase Noise (155.52 MHz LVDS Output) At 1 kHz At 10 kHz At 100 kHz At 1 MHz At 10 MHz At 30 MHz Phase Noise (622.08 MHz LVDS Output) At 1 kHz At 10 kHz At 100 kHz At 1 MHz At 10 MHz At 30 MHz Phase Noise (155.52 MHz CMOS Output) At 1 kHz At 10 kHz At 100 kHz At 1 MHz At 10 MHz At 30 MHz 1 Data Sheet Min Typ Max Unit -118 -127 -127 -149 -156 -157 dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz -115 -124 -124 -146 -155 -155 dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz -117 -122 -123 -148 -156 -156 dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz -105 -110 -110 -136 -150 -150 dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz -117 -122 -123 -146 -155 -155 dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz -105 -110 -110 -134 -149 -150 dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz -117 -122 -123 -147 -155 -155 dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz Test Conditions/Comments 1 Na = 102, Vx = 4, Dx = 6, fPFD = 25 MHz Na = 100, Vx = 4, Dx = 4, fPFD = 25 MHz Na = 112, Vx = 2, Dx = 7, fPFD = 19.44 MHz Na = 128, Vx = 2, Dx = 2, fPFD = 19.44 MHz Na = 112, Vx = 2, Dx = 7, fPFD = 19.44 MHz Na = 128, Vx = 2, Dx = 2, fPFD = 19.44 MHz Na = 112, Vx = 2, Dx = 7, fPFD = 19.44 MHz x indicates either 0 or 1 for any given test condition. Rev. 0 | Page 4 of 44 Data Sheet AD9577 PLL1 CLOCK OUTPUT JITTER Table 2. Parameter 1 LVPECL INTEGRATED RANDOM PHASE JITTER RMS Jitter (625 MHz Output) Min RMS Jitter (156.25 MHz Output) RMS Jitter (106.25 MHz Output) LVDS INTEGRATED RANDOM PHASE JITTER RMS Jitter (625 MHz Output) RMS Jitter (156.25 MHz Output) RMS Jitter (106.25 MHz Output) CMOS INTEGRATED RANDOM PHASE JITTER RMS Jitter (100 MHz Output) RMS Jitter (33.3 MHz Output) LVPECL INTEGRATED RANDOM PHASE JITTER RMS Jitter (622.08 MHz Output) RMS Jitter (155.52 MHz Output) LVDS INTEGRATED RANDOM PHASE JITTER RMS Jitter (622.08 MHz Output) RMS Jitter (155.52 MHz Output) CMOS INTEGRATED RANDOM PHASE JITTER RMS Jitter (155.52 MHz Output) RMS Jitter (38.88 MHz Output) LVPECL PERIOD AND CYCLE-TO-CYCLE JITTER (100 MHz OUTPUT) Output Peak-to-Peak Period Jitter Output RMS Period Jitter Output Peak-to-Peak, Cycle-to-Cycle Jitter Output RMS Cycle-to-Cycle Jitter LVDS PERIOD AND CYCLE-TO-CYCLE JITTER (100 MHz OUTPUT) Output Peak-to-Peak Period Jitter Output RMS Period Jitter Output Peak-to-Peak, Cycle-to-Cycle Jitter Output RMS Cycle-to-Cycle Jitter CMOS PERIOD AND CYCLE-TO-CYCLE JITTER (100 MHz OUTPUT) Output Peak-to-Peak Period Jitter Output RMS Period Jitter Output Peak-to-Peak Cycle-to-Cycle Jitter Output RMS Cycle-to-Cycle Jitter 1 2 Typ Max Unit 460 430 460 460 750 650 750 750 fs rms fs rms fs rms fs rms 470 450 470 470 820 790 790 790 fs rms fs rms fs rms fs rms 470 420 920 700 fS rms fS rms 500 460 480 680 590 680 fs rms fs rms fs rms 520 480 480 780 710 750 fs rms fs rms fs rms 470 440 700 650 fs rms fs rms 13 2 19 3 ps p-p ps rms ps p-p ps rms 17 2 25 4 ps p-p ps rms ps p-p ps rms 25 3 36 6 ps p-p ps rms ps p-p ps rms All period and cycle-to-cycle jitter measurements are made with a Tektronix DPO70604 oscilloscope. x indicates either 0 or 1 for any given test condition. Rev. 0 | Page 5 of 44 Test Conditions/Comments 2 25 MHz crystal used 12 kHz to 20 MHz, Na = 100, Vx = 2, Dx = 2 50 kHz to 80 MHz, Na = 100, Vx = 2, Dx = 2 12 kHz to 20 MHz, Na = 100, Vx = 4, Dx = 4 12 kHz to 20 MHz, Na = 102, Vx = 4, Dx = 6 25 MHz crystal used 12 kHz to 20 MHz, Na = 100, Vx = 2, Dx = 2 50 kHz to 80 MHz, Na = 100, Vx = 2, Dx = 2 12 kHz to 20 MHz, Na = 100, Vx = 4, Dx = 4 12 kHz to 20 MHz, Na = 102, Vx = 4, Dx = 6 25 MHz crystal used, 50 load 12 kHz to 20 MHz, Na = 96, Vx = 4, Dx = 6 12 kHz to 5 MHz, Na = 88, Vx = 6, Dx = 11 19.44 MHz crystal used 12 kHz to 20 MHz, Na = 128, Vx = 2, Dx = 2 50 kHz to 80 MHz, Na = 128, Vx = 2, Dx = 2 12 kHz to 20 MHz, Na = 112, Vx = 2, Dx = 7 19.44 MHz crystal used 12 kHz to 20 MHz, Na = 128, Vx = 2, Dx = 2 50 kHz to 80 MHz, Na = 128, Vx = 2, Dx = 2 12 kHz to 20 MHz, Na = 112, Vx = 2, Dx = 7 19.44 MHz crystal used, 50 load 12 kHz to 20 MHz, Na = 112, Vx = 2, Dx = 7 12 kHz to 5 MHz, Na = 112, Vx = 2, Dx = 28 25 MHz crystal used, Na = 96, Vx = 4, Dx = 6 10,000 cycles, average of 25 measurements 10,000 cycles, average of 25 measurements 1,000 cycles, average of 25 measurements 1,000 cycles, average of 25 measurements 25 MHz crystal used, Na = 96, Vx = 4, Dx = 6 10,000 cycles, average of 25 measurements 10,000 cycles, average of 25 measurements 1,000 cycles, average of 25 measurements 1,000 cycles, average of 25 measurements 25 MHz crystal used, 50 load, Na = 96, Vx = 4, Dx = 6 10,000 cycles, average of 25 measurements 10,000 cycles, average of 25 measurements 1,000 cycles, average of 25 measurements 1,000 cycles, average of 25 measurements AD9577 Data Sheet PLL2 FRACTIONAL-N MODE CHARACTERISTICS Table 3. Bleed = 1 Parameter NOISE CHARACTERISTICS Phase Noise (155.52 MHz LVPECL Output) @ 1 kHz @ 10 kHz @ 100 kHz @ 1 MHz @ 10 MHz @ 30 MHz Phase Noise (622.08 MHz LVPECL Output) @ 1 kHz @ 10 kHz @ 100 kHz @ 1 MHz @ 10 MHz @ 30 MHz Phase Noise (155.52 MHz LVDS Output) @ 1 kHz @ 10 kHz @ 100 kHz @ 1 MHz @ 10 MHz @ 30 MHz Phase Noise (622.08 MHz LVDS Output) @ 1 kHz @ 10 kHz @ 100 kHz @ 1 MHz @ 10 MHz @ 30 MHz Phase Noise (155.52 MHz CMOS Output) @ 1 kHz @ 10 kHz @ 100 kHz @ 1 MHz @ 10 MHz @ 30 MHz SPREAD SPECTRUM Modulation Range Modulation Frequency Peak Power Reduction 1 Min Typ Max Unit -107 -115 -122 -146 -153 -152 dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz -95 -103 -109 -133 -148 -150 dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz -107 -114 -122 -145 -154 -154 dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz -95 -103 -109 -132 -147 -149 dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz -107 -114 -122 -146 -154 -154 dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz Test Conditions/Comments 1 25 MHz crystal used Nb = 87, FRAC = 57, MOD = 625, Vx = 2, Dx = 7 Nb = 99, FRAC = 333, MOD = 625, Vx = 2, Dx = 2 Nb = 87, FRAC = 57, MOD = 625, Vx = 2, Dx = 7 Nb = 99, FRAC = 333, MOD = 625, Vx = 2, Dx = 2 Nb = 87, FRAC = 57, MOD = 625, Vx = 2, Dx = 7 +0.1 -0.5 31.25 10 % kHz dB x indicates either 2 or 3 for any given test condition. Rev. 0 | Page 6 of 44 Downspread, triangle modulation profile Programmable First harmonic of 100 MHz output, triangle modulation profile, spectrum analyzer resolution bandwidth = 20 kHz Data Sheet AD9577 PLL2 INTEGER-N MODE CHARACTERISTICS Table 4. Bleed = 0 Parameter NOISE CHARACTERISTICS Phase Noise (106.25 MHz LVPECL Output) @ 1 kHz @ 10 kHz @ 100 kHz @ 1 MHz @ 10 MHz @ 30 MHz Phase Noise (156.25 MHz LVPECL Output) @ 1 kHz @ 10 kHz @ 100 kHz @ 1 MHz @ 10 MHz @ 30 MHz Phase Noise (625 MHz LVPECL Output) @ 1 kHz @ 10 kHz @ 100 kHz @ 1 MHz @ 10 MHz @ 30 MHz Phase Noise (106.25 MHz LVDS Output) @ 1 kHz @ 10 kHz @ 100 kHz @ 1 MHz @ 10 MHz @ 30 MHz Phase Noise (156.25 MHz LVDS Output) @ 1 kHz @ 10 kHz @ 100 kHz @ 1 MHz @ 10 MHz @ 30 MHz Phase Noise (625 MHz LVDS Output) @ 1 kHz @ 10 kHz @ 100 kHz @ 1 MHz @ 10 MHz @ 30 MHz Phase Noise (106.25 MHz CMOS Output) @ 1 kHz @ 10 kHz @ 100 kHz @ 1 MHz @ 10 MHz @ 30 MHz Min Typ Max Unit Test Conditions/Comments 1 Nb = 102, Vx = 4, Dx = 6, fPFD = 25 MHz -116 -123 -127 -148 -156 -158 dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz -113 -120 -124 -146 -156 -156 dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz -101 -108 -112 -134 -149 -150 dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz -117 -123 -127 -147 -156 -156 dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz -113 -120 -124 -145 -155 -155 dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz -101 -108 -112 -133 -148 -149 dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz -117 -123 -127 -147 -156 -157 dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz Nb = 100, Vx = 4, Dx = 4, fPFD = 25 MHz Nb = 100, Vx = 2, Dx = 2, fPFD = 25 MHz Nb = 102, Vx = 4, Dx = 6, fPFD = 25 MHz Nb = 100, Vx = 4, Dx = 4, fPFD = 25 MHz Nb = 100, Vx = 2, Dx = 2, fPFD = 25 MHz Nb = 102, Vx = 4, Dx = 6, fPFD = 25 MHz Rev. 0 | Page 7 of 44 AD9577 Parameter Phase Noise (156.25 MHz CMOS Output) @ 1 kHz @ 10 kHz @ 100 kHz @ 1 MHz @ 10 MHz @ 30 MHz Phase Noise (155.52 MHz LVPECL Output) @ 1 kHz @ 10 kHz @ 100 kHz @ 1 MHz @ 10 MHz @ 30 MHz Phase Noise (622.08 MHz LVPECL Output) @ 1 kHz @ 10 kHz @ 100 kHz @ 1 MHz @ 10 MHz @ 30 MHz Phase Noise (155.52 MHz LVDS Output) @ 1 kHz @ 10 kHz @ 100 kHz @ 1 MHz @ 10 MHz @ 30 MHz Phase Noise (622.08 MHz LVDS Output) @ 1 kHz @ 10 kHz @ 100 kHz @ 1 MHz @ 10 MHz @ 30 MHz Phase Noise (155.52 MHz CMOS Output) @ 1 kHz @ 10 kHz @ 100 kHz @ 1 MHz @ 10 MHz @ 30 MHz 1 Data Sheet Min Typ Max Unit -113 -119 -123 -145 -154 -155 dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz -112 -118 -126 -147 -155 -156 dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz -100 -106 -112 -134 -149 -150 dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz -113 -118 -126 -145 -154 -155 dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz -101 -106 -112 -133 -148 -150 dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz -113 -118 -126 -146 -155 -155 dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz dBc/Hz Test Conditions/Comments 1 Nb = 100, Vx = 4, Dx = 4, fPFD = 25 MHz Nb = 112, Vx = 2, Dx = 7, fPFD = 19.44 MHz Nb = 128, Vx = 2, Dx = 2, fPFD = 19.44 MHz Nb = 112, Vx = 2, Dx = 7, fPFD = 19.44 MHz Nb = 128, Vx = 2, Dx = 2, fPFD = 19.44 MHz Nb = 112, Vx = 2, Dx = 7, fPFD = 19.44 MHz x indicates either 2 or 3 for any given test condition. Rev. 0 | Page 8 of 44 Data Sheet AD9577 PLL2 CLOCK OUTPUT JITTER Table 5. Bleed = 0 for Integer-N Mode, Bleed = 1 for Fractional-N Mode Parameter 1 LVPECL INTEGRATED RANDOM PHASE JITTER RMS Jitter (622.08 MHz Output) Typ Max Unit 660 1200 fs rms 500 900 fs rms 470 800 fs rms 380 650 fs rms RMS Jitter (155.52 MHz Output) 630 1100 fs rms RMS Jitter (156.25 MHz Output) 470 800 fs rms 660 1200 fs rms 510 900 fs rms 470 820 fs rms 380 650 fs rms RMS Jitter (155.52 MHz Output) 620 1100 fs rms RMS Jitter (156.25 MHz Output) 480 800 fs rms 630 1100 fs rms RMS Jitter (100 MHz Output) 490 800 fs rms RMS Jitter (33.33 MHz Output) 450 700 fs rms 510 800 fs rms 380 650 fs rms 470 800 fs rms 530 900 fs rms 390 700 fs rms 480 750 fs rms RMS Jitter (625 MHz Output) LVDS INTEGRATED RANDOM PHASE JITTER RMS Jitter (622.08 MHz Output) RMS Jitter (625 MHz Output) CMOS INTEGRATED RANDOM PHASE JITTER RMS Jitter (155.52 MHz Output) LVPECL INTEGRATED RANDOM PHASE JITTER RMS Jitter (622.08 MHz Output) RMS Jitter (155.52 MHz Output) LVDS INTEGRATED RANDOM PHASE JITTER RMS Jitter (622.08 MHz Output) RMS Jitter (155.52 MHz Output) Min Rev. 0 | Page 9 of 44 Test Conditions/Comments 2 25 MHz crystal used 12 kHz to 20 MHz, fractional-N operation, Nb = 99, FRAC = 333, MOD = 625, Vx = 2, Dx = 2 50 kHz to 80 MHz, fractional-N operation, Nb = 99, FRAC = 333, MOD = 625, Vx = 2, Dx = 2 12 kHz to 20 MHz, integer-N operation, Nb = 100, Vx = 2, Dx = 2 50 kHz to 80 MHz, integer-N operation, Nb = 100, Vx = 2, Dx = 2 12 kHz to 20 MHz, fractional-N operation, Nb = 87, FRAC = 57, MOD = 625, Vx = 2, Dx = 7 12 kHz to 20 MHz, integer-N operation, Nb = 100, Vx = 4, Dx = 4 25 MHz crystal used 12kHz to 20 MHz, fractional-N operation, Nb = 99, FRAC = 333, MOD = 625, Vx = 2, Dx = 2 50 kHz to 80 MHz, fractional-N operation, Nb = 99, FRAC = 333, MOD = 625, Vx = 2, Dx = 2 12 kHz to 20 MHz, integer-N operation, Nb = 100, Vx = 2, Dx = 2 50 kHz to 80 MHz, integer-N operation, Nb = 100, Vx = 2, Dx = 2 12 kHz to 20 MHz, fractional-N operation, Nb = 87, FRAC = 57, MOD = 625, Vx = 2, Dx = 7 12 kHz to 20 MHz, integer-N operation, Nb = 100, Vx = 4, Dx = 4 25 MHz crystal used, 50 load 12 kHz to 20 MHz, fractional-N operation, Nb = 87, FRAC = 57, MOD = 625, Vx = 2, Dx = 7 12 kHz to 20 MHz, integer-N operation, Nb = 96, Vx = 4, Dx = 6 12 kHz to 5 MHz, integer-N operation, Nb = 88, Vx = 6, Dx = 11 19.44 MHz crystal used 12 kHz to 20 MHz, integer-N operation, Nb = 128, Vx = 2, Dx = 2 50 kHz to 80 MHz, integer-N operation, Nb = 128, Vx = 2, Dx = 2 12 kHz to 20 MHz, integer-N operation, Nb = 112, Vx = 2, Dx = 7 19.44 MHz crystal used 12 kHz to 20 MHz, integer-N operation, Nb = 128, Vx = 2, Dx = 2 50 kHz to 80 MHz, integer-N operation, Nb = 128, Vx = 2, Dx = 2 12 kHz to 20 MHz, integer-N operation, Nb = 112, Vx = 2, Dx = 7 AD9577 Data Sheet Parameter 1 CMOS INTEGRATED RANDOM PHASE JITTER RMS Jitter (155.52 MHz Output) Min RMS Jitter (38.88 MHz Output) Typ Max Unit 470 700 fs rms 430 650 fs rms LVPECL PERIOD AND CYCLE-TO-CYCLE JITTER (100 MHz OUTPUT) Output Peak-to-Peak Period Jitter Output RMS Period Jitter Output Peak-to-Peak, Cycle-to-Cycle Jitter Output RMS Cycle-to-Cycle Jitter LVDS PERIOD AND CYCLE-TO-CYCLE JITTER (100 MHz OUTPUT) 13 2 19 3 ps p-p ps rms ps p-p ps rms Output Peak-to-Peak Period Jitter Output RMS Period Jitter Output Peak-to-Peak, Cycle-to-Cycle Jitter Output RMS Cycle-to-Cycle Jitter CMOS PERIOD AND CYCLE-TO-CYCLE JITTER (100 MHz OUTPUT) 17 2 26 4 ps p-p ps rms ps p-p ps rms Output Peak-to-Peak Period Jitter Output RMS Period Jitter Output Peak-to-Peak, Cycle-to-Cycle Output RMS Cycle-to-Cycle Jitter LVPECL PERIOD AND CYCLE-TO-CYCLE JITTER (100 MHz OUTPUT) 25 3 36 6 ps p-p ps rms ps p-p ps rms Output Peak-to-Peak Period Jitter Output RMS Period Jitter Output Peak-to-Peak, Cycle-to-Cycle Jitter Output RMS Cycle-to-Cycle Jitter LVDS PERIOD AND CYCLE-TO-CYCLE JITTER (100 MHz OUTPUT) 60 15 20 3 ps p-p ps rms ps p-p ps rms Output Peak-to-Peak Period Jitter Output RMS Period Jitter Output Peak-to-Peak, Cycle-to-Cycle Jitter Output RMS Cycle-to-Cycle Jitter CMOS PERIOD AND CYCLE-TO-CYCLE JITTER (100 MHz OUTPUT) 63 15 25 4 ps p-p ps rms ps p-p ps rms 70 15 36 6 ps p-p ps rms ps p-p ps rms Output Peak-to-Peak Period Jitter Output RMS Period Jitter Output Peak-to-Peak, Cycle-to-Cycle Jitter Output RMS Cycle-Cycle Jitter Rev. 0 | Page 10 of 44 Test Conditions/Comments 2 19.44 MHz crystal used, 50 load 12 kHz to 20 MHz, integer-N operation, Nb = 112, Vx = 2, Dx = 7 12 kHz to 5 MHz, integer-N operation, Nb = 112, Vx = 2, Dx = 28 25 MHz crystal used, integer-N operation, Nb = 96, Vx = 4, Dx = 6 10,000 cycles, average of 25 measurements 10,000 cycles, average of 25 measurements 1,000 cycles, average of 25 measurements 1,000 cycles, average of 25 measurements 25 MHz crystal used, integer-N operation, Nb = 96, Vx = 4, Dx = 6 10,000 cycles, average of 25 measurements 10,000 cycles, average of 25 measurements 1,000 cycles, average of 25 measurements 1,000 cycles, average of 25 measurements 25 MHz crystal used, 50 load, integer-N operation, Nb = 96, Vx = 4, Dx = 6 10,000 cycles, average of 25 measurements 10,000 cycles, average of 25 measurements 1,000 cycles, average of 25 measurements 1,000 cycles, average of 25 measurements 25 MHz crystal used, SSCG on, Nb = 100, FRAC = 0, MOD = 1000, Vx = 5, Dx = 5, CkDiv = 7, NumSteps = 59, FracStep = -8, fOUT = 100 MHz with -0.5% downspread at 30.2 kHz 10,000 cycles, average of 25 measurements 10,000 cycles, average of 25 measurements 1,000 cycles, average of 25 measurements 1,000 cycles, average of 25 measurements 25 MHz crystal used, SSCG on, Nb = 100, FRAC = 0, MOD = 1000, Vx = 5, Dx = 5, CkDiv = 7, NumSteps = 59, FracStep = -8, fOUT = 100 MHz with -0.5% downspread at 30.2 kHz 10,000 cycles, average of 25 measurements 10,000 cycles, average of 25 measurements 1,000 cycles, average of 25 measurements 1,000 cycles, average of 25 measurements 25 MHz crystal used, SSCG on, 50 load, Nb = 100, FRAC = 0, MOD = 1000, Vx = 5, Dx = 5, CkDiv = 7, NumSteps = 59, FracStep = -8, fOUT = 100 MHz with -0.5% downspread at 30.2 kHz 10,000 cycles, average of 25 measurements 10,000 cycles, average of 25 measurements 1,000 cycles, average of 25 measurements 1,000 cycles, average of 25 measurements Data Sheet AD9577 Parameter 1 LVPECL PERIOD AND CYCLE-TO-CYCLE JITTER (100.12 MHz OUTPUT) Min Typ Max Unit Output Peak-to-Peak Period Jitter Output RMS Period Jitter Output Peak-to-Peak, Cycle-to-Cycle Jitter Output RMS Cycle-to-Cycle Jitter LVDS PERIOD AND CYCLE-TO-CYCLE JITTER (100.12 MHz OUTPUT) 13 2 20 3 ps p-p ps rms ps p-p ps rms Output Peak-to-Peak Period Jitter Output RMS Period Jitter Output Peak-to-Peak, Cycle-to-Cycle Jitter Output RMS Cycle-to-Cycle Jitter CMOS PERIOD AND CYCLE-TO-CYCLE JITTER (100.12 MHz OUTPUT) 17 2 26 4 ps p-p ps rms ps p-p ps rms 25 3 36 6 ps p-p ps rms ps p-p ps rms Output Peak-to-Peak Period Jitter Output RMS Period Jitter Output Peak-to-Peak, Cycle-to-Cycle Jitter Output RMS Cycle-to-Cycle Jitter 1 2 Test Conditions/Comments 2 25 MHz crystal used, fractional-N operation, Nb = 100, FRAC = 15, MOD = 125, Vx = 5, Dx = 5 10,000 cycles, average of 25 measurements 10,000 cycles, average of 25 measurements 1,000 cycles, average of 25 measurements 1,000 cycles, average of 25 measurements 25 MHz crystal used, fractional-N operation, Nb = 100, FRAC = 15, MOD = 125, Vx = 5, Dx = 5 10,000 cycles, average of 25 measurements 10,000 cycles, average of 25 measurements 1,000 cycles, average of 25 measurements 1,000 cycles, average of 25 measurements 25 MHz crystal used, 50 load, fractional-N operation, Nb = 100, FRAC = 15, MOD = 125, Vx = 5, Dx = 5 10,000 cycles, average of 25 measurements 10,000 cycles, average of 25 measurements 1,000 cycles, average of 25 measurements 1,000 cycles, average of 25 measurements All period and cycle-to-cycle jitter measurements are made with a Tektronix DPO70604 oscilloscope. x indicates either 2 or 3 for any given test condition. CMOS REFERENCE CLOCK OUTPUT JITTER Table 6. Parameter JITTER INTEGRATION BANDWIDTH 12 kHz to 5 MHz 200 kHz to 5 MHz Min Typ Max Unit 680 670 1000 950 fs rms fs rms Rev. 0 | Page 11 of 44 Test Conditions/Comments Jitter measurement at 25 MHz is equipment limited 25 MHz AD9577 Data Sheet TIMING CHARACTERISTICS Table 7. Parameter LVPECL (see Figure 2) Output Rise Time, tRP Output Fall Time, tFP Skew LVDS (see Figure 3) Output Rise Time, tRL Output Fall Time, tFL Skew CMOS (see Figure 4) Output Rise Time, tRC Output Fall Time, tFC Min Typ Max Unit 170 170 225 230 20 300 310 ps ps ps 180 180 250 260 20 340 330 ps ps ps 250 680 950 ps 350 700 1000 ps Skew 20 Test Conditions/Comments Termination = 200 to 0 V, ac-coupled to 50 oscilloscope; CLOAD = 5 pF 20% to 80%, measured differentially 80% to 20%, measured differentially Between the outputs of the same PLL at the same frequency. SyncCh01/SyncCh23 set to 1 Termination = 100 differential; CLOAD = 5 pF 20% to 80%, measured differentially 80% to 20%, measured differentially Between the outputs of the same PLL at the same frequency; SyncCh01/SyncCh23 set to 1 Termination is high impedance active probe, total CLOAD = 5 pF, RLOAD = 20 k, 20% to 80% Termination is high impedance active probe, total CLOAD = 5 pF, RLOAD = 20 k, 80% to 20% Between the outputs of the same PLL at the same frequency; SyncCh01/SyncCh23 set to 1 ps Timing Diagrams SINGLE-ENDED DIFFERENTIAL 80% 80% VOD 50% CMOS tRP tFP 09284-002 LVPECL tRC Figure 4. CMOS Timing, Single-Ended, 5 pF Load Figure 2. LVPECL Timing, Differential DIFFERENTIAL 80% VOD 50% 20% tFL 09284-003 LVDS tRL tFC Figure 3. LVDS Timing, Differential Rev. 0 | Page 12 of 44 09284-004 20% 20% Data Sheet AD9577 CLOCK OUTPUTS AC coupling capacitors of 0.1 F used where appropriate. Table 8. Parameter LVPECL CLOCK OUTPUTS Output Frequency Min Typ 740 Max Unit Test Conditions/Comments 637.5 MHz 950 mV 55 % Load is 200 to GND at output pins, then ac-coupled to 50 terminated measurement equipment. Load is 200 to GND at output pins, then ac-coupled to 50 terminated measurement equipment. For differential amplitude, see Figure 2. Load is 200 to GND at output pins, then ac-coupled to 50 terminated measurement equipment. Load is 127 /83 potential divider across supply dc-coupled into 1 M terminated measurement equipment, outputs static. Load is 127 /83 potential divider across supply dc-coupled into 1 M terminated measurement equipment, outputs static. Output Voltage Swing, VOD 610 Duty Cycle 45 Output High Voltage, VOH VS - 1.24 VS - 0.94 VS - 0.83 V Output Low Voltage, VOL VS - 2.07 VS - 1.75 VS - 1.62 V 637.5 MHz 475 mV 25 mV 55 % 1.375 V 25 mV 24 mA 200 MHz V V % LVDS CLOCK OUTPUTS Output Frequency Differential Output Voltage, VOD 250 350 Delta VOD Duty Cycle 45 Output Offset Voltage, VOS 1.125 1.25 Delta VOS Short-Circuit Current, ISA, ISB CMOS CLOCK OUTPUTS Output Frequency Output High Voltage, VOH Output Low Voltage, VOL Duty Cycle 13 VS - 0.15 45 0.1 55 Rev. 0 | Page 13 of 44 Load is ac-coupled to measurement equipment that provides 100 differential input termination. Load is ac-coupled to measurement equipment that provides 100 differential input termination. For differential amplitude, see Figure 3. Load is ac-coupled to measurement equipment that provides 100 differential input termination. Load is ac-coupled to measurement equipment that provides 100 differential input termination. Load is dc-coupled to a 100 differential resistor into 1 M terminated measurement equipment, outputs static. Load is dc-coupled to a 100 differential resistor into 1 M terminated measurement equipment, outputs static. Load is dc-coupled to a 100 differential resistor into 1 M terminated measurement equipment, output shorted to GND. Sourcing 1.0 mA current, outputs static. Sinking 1.0 mA current, outputs static. Termination is high impedance active probe; total CLOAD = 5 pF, RLOAD = 20 k. AD9577 Data Sheet POWER Table 9. Parameter POWER SUPPLY LVPECL POWER DISSIPATION Min 3.0 LVDS POWER DISSIPATION CMOS POWER DISSIPATION Typ 3.3 1235 Max 3.6 1490 Unit V mW 1270 1530 mW 1020 1200 mW 1085 1290 mW 1065 1380 mW 1190 1510 mW POWER CHANGES Power-Down 1 LVPECL Channel 160 205 mW Power-Down 1 LVDS Channel 105 155 mW Power-Down 1 CMOS Channel 130 170 mW Test Conditions/Comments Typical part configuration, both PLLs enabled for integer-N operation, fOUT0 = 156.25 MHz, fOUT1 = 125 MHz, fOUT2 = 100 MHz, fOUT3 = 33.33 MHz, Na = 100, V0 = 4, D0 = 4, V1 = 4, D1 = 5, Nb = 96, V2 = 4, D2 = 6, V3 = 4, D3 = 18, 25 MHz crystal used, load is 200 to GND at output pins, then ac-coupled to 50 terminated measurement equipment Worst-case part configuration, PLL2 in fractional-N mode, with SSCG enabled, fOUT0 = 379.16 MHz, fOUT1 = 379.16 MHz, fOUT2 = 359.33 MHz, fOUT3 = 359.33 MHz, Na = 91, V0 = 3, D0 = 2, V1 = 3, D1 = 2, Nb = 86, V2 = 3, D2 = 2, V3 = 3, D3 = 2, FRAC = 300, MOD = 1250, CkDiv = 5, NumSteps = 77, FracStep = -7, -0.5% downspread at 32 kHz, 25 MHz crystal used, load is 200 to GND at output pins, then ac-coupled to 50 terminated measurement equipment Typical part configuration, both PLLs enabled for integer-N operation, fOUT0 = 156.25 MHz, fOUT1 = 125 MHz, fOUT2 = 100 MHz, fOUT3 = 33.33 MHz, Na = 100, V0 = 4, D0 = 4, V1 = 4, D1 = 5, Nb = 96, V2 = 4, D2 = 6, V3 = 4, D3 = 18, 25 MHz crystal used, load ac-coupled to measurement equipment that provides 100 differential input termination Worst-case part configuration, PLL2 in fractional-N mode, with SSCG enabled, fOUT0 = 379.16 MHz, fOUT1 = 379.16 MHz, fOUT2 = 359.33 MHz, fOUT3 = 359.33 MHz, Na = 91, V0 = 3, D0 = 2, V1 = 3, D1 = 2, Nb = 86, V2 = 3, D2 = 2, V3 = 3, D3 = 2, FRAC = 300, MOD = 1250, CkDiv = 5, NumSteps = 77, FracStep = -7, -0.5% downspread at 32 kHz, 25 MHz crystal used, load ac-coupled to measurement equipment that provides 100 differential input termination Typical part configuration, both PLLs enabled for integer-N operation, fOUT0 = 156.25 MHz, fOUT1 = 125 MHz, fOUT2 = 100 MHz, fOUT3 = 33.33 MHz, Na = 100, V0 = 4, D0 = 4, V1 = 4, D1 = 5, Nb = 96, V2 = 4, D2 = 6, V3 = 4, D3 = 18, 25 MHz crystal used, eight single-ended outputs active, CLOAD = 5 pF Worst-case part configuration, PLL2 in fractional-N mode, with SSCG enabled, fOUT0 = 189.58 MHz, fOUT1 = 189.58 MHz, fOUT2 = 179.66 MHz, fOUT3 = 179.66 MHz, Na = 91, V0 = 3, D0 = 4, V1 = 3, D1 = 4, Nb = 86, V2 = 3, D2 = 4, V3 = 3, D3 = 4, FRAC = 300, MOD = 1250, CkDiv = 5, NumSteps = 77, FracStep = -7, -0.5% downspread at 32 kHz, 25 MHz crystal used, eight single-ended outputs active, CLOAD = 5 pF Reduction in power due to turning off a channel of one VCO divider, one output divider, and one output buffer; data for Channel 1, with typical part configuration, both PLLs enabled for integer-N operation, fOUT0 = 156.25 MHz, fOUT1 = 125 MHz, fOUT2 = 100 MHz, fOUT3 = 33.33 MHz, Na = 100, V0 = 4, D0 = 4, V1 = 4, D1 = 5, Nb = 96, V2 = 4, D2 = 6, V3 = 4, D3 = 18, 25 MHz crystal used Load 200 to GND at output pins, and ac-coupled to 50 terminated measurement equipment Load ac-coupled to measurement equipment that provides 100 differential input termination Eight single-ended outputs active, CLOAD = 5 pF Rev. 0 | Page 14 of 44 Data Sheet AD9577 CRYSTAL OSCILLATOR Table 10. Parameter CRYSTAL SPECIFICATION Frequency ESR Load Capacitance Phase Noise Stability Min Typ Max Unit 19.44 25 27 50 +50 MHz pF dBc/Hz ppm 14 -135 -50 Test Conditions/Comments Fundamental mode Reference divider, R = 1, only 1 kHz offset REFERENCE INPUT Table 11. Parameter CLOCK INPUT (REFCLK) Input Frequency Input High Voltage Input Low Voltage Input Current Input Capacitance Min Typ Max Unit Test Conditions/Comments 19.44 38.88 2.0 25 50 27 54 MHz MHz V V A pF Reference divider, R = 1 Reference divider, R = 2 0.8 +1.0 -1.0 2 CONTROL PINS Table 12. Parameter INPUT CHARACTERISTICS SSCG, MAX_BW, and MARGIN Min Logic 1 Voltage Logic 0 Voltage Logic 1 Current Logic 0 Current REFSEL Logic 1 Voltage Logic 0 Voltage Logic 1 Current Logic 0 Current I2C DC CHARACTERISTICS Input Voltage High Input Voltage Low Input Current Output Low Voltage 2 I C TIMING CHARACTERISTICS SCL Clock Frequency SCL Pulse Width High High, tHIGH Low, tLOW Start Condition Hold Time, tHD; STA Setup Time, tSU; STA Data Setup Time, tSU; DAT Hold Time, tHD; DAT Stop Condition Setup Time, tSU; STO Bus Free Time Between a Stop and a Start, tBUF 2.0 Typ Max Unit Test Conditions/Comments SSCG, MAX_BW, and MARGIN have a 30 k internal pull-down resistor 0.8 240 40 V V A A 0.8 70 240 V V A A 0.3 Vcc +10 0.4 V V A V 400 kHz REFSEL has a 30 k internal pull-up resistor 2.0 LVCMOS; the SCL and SDA pins only, see Figure 48 0.7 Vcc -10 600 1300 ns ns 600 600 ns ns 100 300 600 1300 ns ns ns ns Rev. 0 | Page 15 of 44 VIN = 0.1 VCC or VIN = 0.9 VCC VOL with a load current of IOL = 3.0 mA LVCMOS; the SCL and SDA pins only, see Figure 48 AD9577 Data Sheet ABSOLUTE MAXIMUM RATINGS THERMAL CHARACTERISTICS Table 13. Parameter VS to GND REFCLK to GND LDO to GND XT1, XT2 to GND SSCG, MAX_BW, MARGIN, SCL, SDA, REFSEL to GND REFOUT, OUTxP, OUTxN to GND Junction Temperature1 Storage Temperature Lead Temperature (10 sec) 1 Rating -0.3 V to +3.6 V -0.3 V to VS + 0.3 V -0.3V to VS + 0.3 V -0.3 V to VS + 0.3 V -0.3 V to VS + 0.3 V -0.3 V to VS + 0.3 V 150C -65C to+150C 300C Thermal impedance measurements were taken on a 4-layer board in still air in accordance with EIA/JESD51-7. Table 14. Thermal Resistance Package Type 40-Lead LFCSP ESD CAUTION See the Thermal Characteristics section for JA. 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. Rev. 0 | Page 16 of 44 JA 27.5 Unit C/W Data Sheet AD9577 40 39 38 37 36 35 34 33 32 31 MAX_BW TST2A VSVA SCL GND GND OUT0P OUT0N VSOB0A SDA PIN CONFIGURATION AND FUNCTION DESCRIPTIONS PIN 1 INDICATOR AD9577 TOP VIEW (Not to Scale) 30 29 28 27 26 25 24 23 22 21 VSOB1A OUT1P OUT1N VSFA SSCG VSM VSFB OUT3P OUT3N VSOB3B NOTES 1. THE EXPOSED PADDLE ON THIS PACKAGE IS AN ELECTRICAL CONNECTION AS WELL AS A THERMAL ENHANCEMENT. FOR THE DEVICE TO FUNCTION PROPERLY, THE PADDLE MUST BE ATTACHED TO GROUND (GND). IT IS RECOMMENDED THAT A MINIMUM OF NINE VIAS BE USED TO CONNECT THE PADDLE TO THE PRINTED CIRCUIT BOARD (PCB) GROUND PLANE. 09284-005 TST1B TST2B LDO VSVB GND GND OUT2N OUT2P VSOB2B MARGIN 11 12 13 14 15 16 17 18 19 20 VSCA 1 VSI2C 2 REFOUT 3 VSREFOUT 4 VSX 5 REFCLK 6 XT2 7 XT1 8 REFSEL 9 VSCB 10 Figure 5. Pin Configuration Table 15. Pin Function Descriptions Pin No. 1 2 3 4 5 6 7, 8 9 10 11 12 13 14 15, 16, 35, 36 17 18 19 20 21 22 23 24 25 26 27 28 29 Mnemonic VSCA VSI2C REFOUT VSREFOUT VSX REFCLK XT2, XT1 REFSEL VSCB TST1B TST2B LDO VSVB GND OUT2N OUT2P VSOB2B MARGIN VSOB3B OUT3N OUT3P VSFB VSM SSCG VSFA OUT1N OUT1P Description PLL1 Power Supply. I2C Digital Power Supply. CMOS Reference Output. Reference Output Buffer Power Supply. Crystal Oscillator and Input Reference Power Supply. Reference Clock Input. Tie low when not in use. External 19.44 MHz to 27 MHz Crystal. Leave unconnected when not in use. Logic Input. Use this pin to select the reference source. Internal 30 k pull-up resistor. PLL2 Analog Power Supply. Test Pin. Connect this pin to Pin 13 (LDO). Test Pin. Connect this pin to Pin 13 (LDO). This pin is for bypassing the PLL2 LDO to ground with a 220 nF capacitor. PLL2 VCO Power Supply. Ground. LVPECL/LVDS/CMOS Clock Output. LVPECL/LVDS/CMOS Clock Output. Output Port OUT2 Power Supply. Logic 1 sets the margining frequency on the clock output pins. Internal 30 k pull-down resistor. Output Port OUT3 Power Supply. LVPECL/LVDS/CMOS Clock Output. LVPECL/LVDS/CMOS Clock Output. PLL2 Analog Power Supply. PLL2 Digital Power Supply. Logic 1 enables spread spectrum operation of PLL2. Internal 30 k pull-down resistor. PLL1 Analog Power Supply. LVPECL/LVDS/CMOS Clock Output. LVPECL/LVDS/CMOS Clock Output. Rev. 0 | Page 17 of 44 AD9577 Pin No. 30 31 32 33 34 37 38 39 40 Data Sheet Mnemonic VSOB1A SDA VSOB0A OUT0N OUT0P SCL VSVA TST2A MAX_BW EPAD Description Output Port OUT1 Power Supply. Serial Data Line for I2C. Output Port OUT0 Power Supply. LVPECL/LVDS/CMOS Clock Output. LVPECL/LVDS/CMOS Clock Output. Serial Clock for I2C. PLL1 VCO Power Supply. Test Pin. Connect this pin to the printed circuit board (PCB) ground plane. Logic 1 widens the loop bandwidth of the fractional-N PLL during spread spectrum. Internal 30 k pulldown resistor. The exposed paddle on this package is an electrical connection as well as a thermal enhancement. For the device to function properly, the paddle must be attached to ground (GND). It is recommended that a minimum of nine vias be used to connect the paddle to the printed circuit board (PCB) ground plane. Rev. 0 | Page 18 of 44 Data Sheet AD9577 TYPICAL PERFORMANCE CHARACTERISTICS -100 -100 -110 -110 PHASE NOISE (dBc/Hz) -120 -130 -140 -150 100k 1M 10M 100M Figure 6. Phase Noise, REFOUT Output, 25 MHz (fXTAL = 25 MHz) -160 1k -100 -110 -110 -120 -130 -140 -150 1M 10M 100M -120 -130 -140 100k 1M 10M 100M -160 1k 09284-007 10k Figure 7. Phase Noise, PLL1, OUT0 LVPECL, 106.25 MHz, Integer-N Mode (fXTAL = 25 MHz, Na = 102, V0 = 4, D0 = 6) -90 -100 -100 PHASE NOISE (dBc/Hz) -90 -130 -140 -130 -140 -160 -160 FREQUENCY (Hz) 10M 100M -170 1k 09284-008 1M Figure 8. Phase Noise, PLL1, OUT0 LVPECL, 156.25 MHz Integer-N Mode (fXTAL = 25 MHz, Na = 100, V0 = 4, D0 = 4) 100M -120 -150 100k 10M -110 -150 10k 1M Figure 10. Phase Noise, PLL1, OUT0 LVPECL, 125 MHz, Integer-N Mode (fXTAL = 25 MHz, Na = 100, V0 = 4, D0 = 5) -80 -120 100k FREQUENCY (Hz) -80 -110 10k 09284-010 -150 FREQUENCY (Hz) -170 1k 100k Figure 9. Phase Noise, PLL1, OUT0 LVPECL, 100 MHz, Integer-N Mode (fXTAL = 25 MHz, Na = 100, V0 = 5, D0 = 5) -100 -160 1k 10k FREQUENCY (Hz) PHASE NOISE (dBc/Hz) PHASE NOISE (dBc/Hz) -140 09284-009 10k FREQUENCY (Hz) PHASE NOISE (dBc/Hz) -130 -150 09284-006 -160 1k -120 10k 100k 1M FREQUENCY (Hz) 10M 100M 09284-011 PHASE NOISE (dBc/Hz) REFOUT AND PLL1 PHASE NOISE PERFORMANCE Figure 11. Phase Noise, PLL1, OUT0 LVPECL, 625 MHz, Integer-N Mode (fXTAL = 25 MHz, Na = 100, V0 = 2, D0 = 2) Rev. 0 | Page 19 of 44 AD9577 Data Sheet -110 -110 -120 -120 PHASE NOISE (dBc/Hz) -130 -140 1M 10M 100M -160 1k Figure 12. Phase Noise, PLL2, OUT2 LVPECL, 100 MHz, Integer-N Mode (fXTAL = 25 MHz, Nb = 100, V2 = 5, D2 = 5) 10M 100M -90 -100 PHASE NOISE (dBc/Hz) -120 -130 -140 -150 -110 -120 -130 -140 -150 100k 1M 10M 100M FREQUENCY (Hz) -160 1k 09284-013 10k Figure 13. Phase Noise, PLL2, OUT2 LVPECL, 156.25 MHz, Integer-N Mode (fXTAL = 25 MHz, Nb = 100, V2 = 4, D2 = 4) 10k 100k 1M 10M 100M FREQUENCY (Hz) 09284-016 PHASE NOISE (dBc/Hz) 1M Figure 15. Phase Noise, PLL2, OUT2 LVPECL, 106.25 MHz, Integer-N Mode (fXTAL = 25 MHz, Nb = 102, V2 = 4, D2 = 6) -110 Figure 16. Phase Noise, PLL2, OUT2 LVPECL, 625 MHz, Integer-N Mode (fXTAL = 25 MHz, Nb = 100, V2 = 2, D2 = 2) -100 -90 -100 PHASE NOISE (dBc/Hz) -110 PHASE NOISE (dBc/Hz) 100k FREQUENCY (Hz) -100 -120 -130 -140 -150 -110 -120 -130 -140 -150 10k 100k 1M FREQUENCY (Hz) 10M 100M -160 1k 09284-014 -160 1k 10k 09284-015 100k 09284-012 10k FREQUENCY (Hz) -160 1k -140 -150 -150 -160 1k -130 Figure 14. Phase Noise, PLL2, OUT2 LVPECL, 155.52 MHz, Fractional-N Mode (fXTAL = 25 MHz, Nb = 99, FRAC = 333, MOD = 625, V2 = 2, D2 = 8), Spurs Disabled 10k 100k 1M FREQUENCY (Hz) 10M 100M 09284-017 PHASE NOISE (dBc/Hz) PLL2 PHASE NOISE PERFORMANCE Figure 17. Phase Noise, PLL2, OUT2 LVPECL, 622.08 MHz, Fractional-N Mode (fXTAL = 25 MHz, Nb = 99, FRAC = 333, MOD = 625, V2 = 2, D2 = 2), Spurs Disabled Rev. 0 | Page 20 of 44 Data Sheet AD9577 OUTPUT JITTER 500 470 465 460 455 450 445 440 435 430 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 Na Figure 18. Typical Integrated Random Phase Jitter in fs rms for PLL1 and OUT0P LVPECL as Feedback Divider Value Na Swept (fXTAL = 25 MHz, V0 = 5, D0 = 5) 490 480 470 460 450 440 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 Nb 09284-021 INTEGRATED RMS PHASE JITTER (fs) 475 09284-018 INTEGRATED RMS PHASE JITTER (fs) 480 Figure 19. Typical Integrated Random Phase Jitter in fs rms for PLL2 and OUT2P LVPECL as Feedback Divider Value Nb Swept (fXTAL = 25 MHz, V2 = 5, D2 = 5, Integer-N Mode) Rev. 0 | Page 21 of 44 AD9577 Data Sheet 09284-027 5ns/DIV 09284-024 200mV/DIV 200mV/DIV TYPICAL OUTPUT SIGNAL 1ns/DIV Figure 23. Typical LVPECL Differential Output Trace, 625 MHz 1ns/DIV Figure 21. Typical LVDS Differential Output Trace, 156.25 MHz 09284-028 5ns/DIV 09284-025 100mV/DIV 100mV/DIV Figure 20 Typical LVPECL Differential Output Trace, 156.25 MHz 10ns/DIV Figure 25. Typical REFOUT Output Trace, 25 MHz Figure 22. Typical CMOS Output Trace, 200MHz Rev. 0 | Page 22 of 44 09284-029 1ns/DIV 09284-026 500mV/DIV 500mV/DIV Figure 24. Typical LVDS Differential Output Trace, 625 MHz Data Sheet AD9577 3.3 3.2 3.1 3.0 2.9 2.8 2.7 0.5pF LOAD 5.2pF LOAD 10.5pF LOAD 2.6 2.5 0 20 40 60 80 100 120 140 160 180 FREQUENCY (MHz) 200 09284-030 SINGLE-ENDED OUTPUT SWING (V) 3.4 1.6 1.5 1.4 1.3 100 200 300 400 FREQUENCY (MHz) 500 600 600 500 400 100 200 300 400 500 600 Figure 28. LVDS Differential, Peak-to-Peak Output Swing vs. Frequency 1.7 0 700 FREQUENCY (MHz) 1.8 1.2 800 0 09284-031 DIFFERENTIAL PEAK-TO-PEAK OUTPUT SWING (V) Figure 26. CMOS Single-Ended, Peak-to-Peak Output Swing vs. Frequency, for Loads of 0.5 pF, 5.2 pF, and 10.5 pF, Measured with a Tektronix P7313 Active Probe 900 09284-032 DIFFERENTIAL PEAK-TO-PEAK OUTPUT SWING (V) 3.5 Figure 27. LVPECL Differential, Peak-to-Peak Output Swing vs. Frequency Rev. 0 | Page 23 of 44 AD9577 Data Sheet TYPICAL SPREAD SPECTRUM PERFORMANCE CHARACTERISTICS 100.1 0 -10 100.0 POWER (dBm) FREQUENCY (MHz) -20 99.9 99.8 99.7 -30 -40 -50 -60 99.6 -70 99.5 TIME (10s/DIV) -90 695 Figure 29. Typical Spread Spectrum Frequency Modulation Profile OUT2, Nb = 96, FRAC = 0, MOD = 1000, CkDiv = 7, NumSteps = 59, FracStep = -8, fOUT = 100 MHz with -0.5% Downspread at 30.2 kHz, MAX_BW set to 0 10 0 POWER (dBm) -10 -30 -40 -50 -60 99.75 100.00 FREQUENCY (MHz) 100.25 100.50 09284-034 -70 99.50 701 Figure 31. Typical Nonspread and Spread Spectrum Power Spectra, OUT2, Nb = 96, FRAC = 0, MOD = 1000, CkDiv = 7, NumSteps = 59, FracStep = -8, fOUT = 100 MHz with -0.5% Downspread at 30.2 kHz, MAX_BW set to 0, Seventh Harmonic Shown, Spectrum Analyzer Resolution Bandwidth = 120 kHz, Maximum Hold On UNMODULATED CLOCK SPECTRUM MODULATED CLOCK 99.25 699 FREQUENCY (MHz) -20 -80 99.00 697 09284-035 99.4 09284-033 -80 Figure 30. Typical Nonspread and Spread Spectrum Power Spectra, OUT2, Nb = 96, FRAC = 0, MOD = 1000, CkDiv = 7, NumSteps = 59, FracStep = -8, fOUT = 100 MHz with -0.5% Downspread at 30.2 kHz, MAX_BW set to 0, First Harmonic Shown, Spectrum Analyzer Resolution Bandwidth =10 kHz, Maximum Hold On Rev. 0 | Page 24 of 44 Data Sheet AD9577 TERMINOLOGY Phase Jitter and Phase Noise An ideal sine wave can be thought of as having a continuous and even progression of phase with time from 0 to 360 for each cycle. Actual signals, however, display a certain amount of variation from ideal phase progression over time, which is called phase jitter. Although many causes can contribute to phase jitter, one major cause is random noise, which is characterized statistically as being Gaussian (normal) in distribution. This phase jitter leads to a spreading out of the energy of the sine wave in the frequency domain, producing a continuous power spectrum. This power spectrum is usually reported as a series of values whose units are dBc/Hz at a given offset in frequency from the sine wave (carrier). The value is a ratio (expressed in dB) of the power contained within a 1 Hz bandwidth with respect to the power at the carrier frequency. For each measurement, the offset from the carrier frequency is also given. It is meaningful to integrate the total power contained within some interval of offset frequencies (for example, 12 kHz to 20 MHz). This is called the integrated phase noise over that frequency offset interval and can be readily related to the time jitter due to the phase noise within that offset frequency interval. Phase noise has a detrimental effect on error rate performance by increasing eye closure at the transmitter output and reducing the jitter tolerance/sensitivity of the receiver. Time Jitter Phase noise is a frequency domain phenomenon. In the time domain, the same effect is exhibited as time jitter. When observing a sine wave, the time of successive zero crossings vary. In a square wave, the time jitter is seen as a displacement of the edges from their ideal (regular) times of occurrence. In both cases, the variations in timing from the ideal are the time jitter. Because these variations are random in nature, the time jitter is specified in units of seconds root mean square (rms) or 1 sigma of the Gaussian distribution. Additive Phase Noise It is the amount of phase noise that is attributable to the device or subsystem being measured. The phase noise of any external oscillators or clock sources has been subtracted. This makes it possible to predict the degree to which the device affects the total system phase noise when used in conjunction with the various oscillators and clock sources, each of which contributes its own phase noise to the total. In many cases, the phase noise of one element dominates the system phase noise. Additive Time Jitter It is the amount of time jitter that is attributable to the device or subsystem being measured. The time jitter of any external oscillators or clock sources has been subtracted. This makes it possible to predict the degree to which the device will affect the total system time jitter when used in conjunction with the various oscillators and clock sources, each of which contributes its own time jitter to the total. In many cases, the time jitter of the external oscillators and clock sources dominates the system time jitter. Random Jitter Measurement On the AD9577, the rms jitter measurements are made by integrating the phase noise, with spurs disabled. There are two reasons for this. First, because the part is highly configurable, any measured spurs are a function of the current programmed state of the device. For example, there may be a small reference spur at the PFD frequency present on the output spectrum. If the PFD operates at 19.44 MHz (which is common for telecommunications applications), the resulting jitter falls within the normal 12 kHz to 20 MHz integration bandwidth. When the PFD operates above 20 MHz, the deterministic jitter is not included in the measurement. As another example, for PLL2, the value of the chosen FRAC and MOD values affects the amplitude and location of a spur, and therefore, it is not possible to configure the PLL to provide a general measurement that includes spurs. The second, and more significant reason, is due to the statistical nature of spurious components. The jitter performance information of the clock generator is required so that a jitter budget for the complete communications channel can be established. By knowing the jitter characteristics at the ultimate receiver, the data bit error rate (BER) can be estimated to ensure robust data transfer. The received jitter characteristic consists of random jitter (RJ), due to random perturbations such as thermal noise, and deterministic jitter (DJ), due to deterministic perturbations such as crosstalk spurs. To make an estimate of the BER, the total jitter peak-to-peak (TJ p-p) value must be known. It is the total jitter value that determines the amount of eye closure at the receiver and, consequently, the bit error rate. The TJ p-p value is specified for a given number of clock edges. For example, in networking applications, the TJ is specified for 112 clock edges. The equation for the total jitter peak-to-peak is TJ p-p = DJ p-p + 2 x Q x RJ rms (1) where the Q factor represents the ratio of the expected peak deviation to the standard deviation in a Gaussian process for a given population (of edge crossings). For 112 clock edges, Q is 7.03; therefore, for networking applications, the total jitter peakto-peak is estimated by TJ p-p = DJ p-p + 14.06 x RJ rms Rev. 0 | Page 25 of 44 (2) AD9577 Data Sheet Therefore, to accurately estimate the TJ p-p, separate measurements of the rms value of the random jitter (RJ rms) and the peak-to-peak value of the deterministic jitter (DJ p-p) must be taken. To measure the RJ rms of the clock signal, integrate the clock phase noise over the desired bandwidth, with spurs disabled (that is, removed) from the measurement. If the DJ spurs were included in the measurement, the DJ contribution would also be multiplied by 14.06 in Equation 2, leading to a grossly pessimistic estimate of the total jitter. This is why it is important to measure the integrated jitter with spurs disabled. Due to the 14.06 factor in Equation 2, the spurious DJ components on the clock output only have a small impact on the TJ p-p measurement and, consequently, the system BER performance. Therefore, it is clear that the DJ component (that is, the spur) should not be added to the rms value of the random jitter directly. However, if the phase noise jitter measurement was preformed with spurs enabled, this is exactly what the measurement would be reporting. For more background information, see Fibre Channel, Methodologies for Jitter and Signal Quality Specification-MJSQ, Rev. 14, June 9, 2004. Rev. 0 | Page 26 of 44 Data Sheet AD9577 DETAILED BLOCK DIAGRAM REFSEL XT1 AD9577 XTAL OSC XT2 CMOS REFOUT REFCLK DIVIDE 1 OR 2 LDO2 THIRD ORDER LPF PLL2 VCO 2.15GHz TO 2.55GHz DIVIDE BY 80 TO 131 FEEDBACK DIVIDER NB V0 D0 DIVIDE BY 2 TO 6 DIVIDE BY 1 TO 32 V1 D1 DIVIDE BY 2 TO 6 DIVIDE BY 1 TO 32 VCO DIVIDERS OUTPUT DIVIDERS V2 D2 DIVIDE BY 2 TO 6 DIVIDE BY 1 TO 32 V3 D3 DIVIDE BY 2 TO 6 DIVIDE BY 1 TO 32 0 FRAC_TRIWAVE 1 SDM MOD SSCG FORMAT1 LVPECL/LVDS OR 2 x CMOS LVPECL/LVDS OR 2 x CMOS OUTPUT BUFFERS FORMAT2 LVPECL/LVDS OR 2 x CMOS LVPECL/LVDS OR 2 x CMOS 4x GND + PADDLE LDO fPFD CKDIV FRAC FRACSTEP NUMSTEPS OUTPUT BUFFERS 14x VS 3-BIT FRAC OUTPUT DIVIDERS 220nF SSCG FRAC_TRIWAVE TRIWAVE GENERATOR 09284-036 SSCG MAX_BW VCO 2.15GHz TO 2.55GHz DIVIDE BY 80 TO 131 FEEDBACK DIVIDER NA THIRD ORDER LPF MARGIN I2C CONTROL fPFD SCL SDA VCO DIVIDERS LDO1 PLL1 PFD/CP 22pF PFD/CP 22pF Figure 32. Detailed Block Diagram Rev. 0 | Page 27 of 44 AD9577 Data Sheet EXAMPLE APPLICATION REFSEL 25MHz XTAL XTAL OSC CMOS 25MHz CMOS REFCLK DIVIDE 1 OR 2 THIRD ORDER LPF fPFD PFD/CP DIVIDERS DIVIDERS LDO PLL1 VCO 2.15GHz TO 2.55GHz FEEDBACK DIVIDER DIVIDERS DIVIDERS THIRD ORDER LPF PFD/CP 125MHz LVPECL LDO PLL2 VCO 2.15GHz TO 2.55GHz FEEDBACK DIVIDER SCL SDA 156.25MHz LVPECL I2 C CONTROL 100MHz LVDS 33.33MHz 2 x CMOS MARGIN SPREAD SPECTRUM, SDM AD9577 NOTE THAT ANY FREQUENCIES MAY BE PROGRAMMED. 09284-037 SSCG MAX_BW Figure 33. Example Application Achievable application frequencies include (but are not limited to) those listed in Table 16. Table 16. Typical Application Frequencies Applications Ethernet 10G Ethernet FB-DIMM Fibre Channel 10G Fibre Channel Inifiniband SAS, SATA Telecomm PCI Express PCI, PCI-X Video Wireless Infrastructure Frequency (MHz) 25, 62.5, 100, 125, 250 155.52, 156.25, 187.5, 161.1328125, 312.5, 622.08, 625 133.333, 166.666, 200 53.125, 106.25, 212.5, 318.75, 425 159.375 125 37.5, 75, 100, 120, 150; the AD9577 also meets the -0.5% downspread requirement 19.44, 38.88, 77.76, 155.52, 311.04, 622.08, 627.32962 100, 125, 250; the AD9577 also meets the -0.5% downspread requirement 33.33, 66.66, 100, 133.33, 200 13.5, 14.318, 17.7, 18, 27, 72, 74.25, 74.25/1.001, 148.5, 148.5/1.001 61.44, 122.88, 368.64 Rev. 0 | Page 28 of 44 Data Sheet AD9577 FUNCTIONAL DESCRIPTION The AD9577 contains two PLLs, PLL1 and PLL2, used for independent clock frequency generation, as shown in Figure 32. A shared crystal oscillator and reference clock input cell drive both PLLs. The reference clock of the PLLs can be selected as either the crystal oscillator output or the reference input clock. A reference divider precedes each PLL. When the crystal oscillator input is selected, these dividers must be set to divide by 1. When the reference input is selected, these dividers can be set to divide by 1 or divide by 2, provided that the resulting input frequency to the PLLs is within the permitted 19.44 MHz to 27 MHz range. Both reference dividers are set to divide by the same value. Each PLL drives two output channels, producing four output ports in total for the IC. Each output channel consists of a VCO divider block, followed by an output divider block. The output divider blocks each drive with an output buffer port. Each output buffer port can be configured as a differential LVDS output, a differential LVPECL output, or two LVCMOS outputs. Additionally, a CMOS-buffered version reference clock frequency is available. The upper PLL in Figure 32, PLL1, is an integer-N PLL. By setting the feedback divider value (Na), the VCO output frequency can tuned over the 2.15 GHz to 2.55 GHz range to integer multiples of the PFD input frequency. By setting each of the VCO divider (V0 and V1) and output divider (D0 and D1) values, the VCO frequency can be divided down to the required output frequency, independently, for each of the output ports, OUT0 and OUT1. The loop filter required for this PLL is integrated on chip. The lower PLL in Figure 32, PLL2, is a fractional-N PLL. This PLL can optionally operate as an integer-N PLL for optimum jitter performance. By setting the feedback divider value (Nb) and the - modulator fractional (FRAC) and modulus (MOD) values, the VCO output frequency can tune over the 2.15 GHz to 2.55 GHz range. The VCO frequency is a fractional multiple of the PFD input frequency. In this way, the VCO frequency can tune to obtain frequencies that are not constrained to integer multiples of the PFD frequency. By setting each of the VCO divider (V2 and V3) and output divider (D2 and D3) values, the VCO frequency can be divided down to the required output frequency, independently, for each of the output ports, OUT2 and OUT3. The loop filters required for this PLL are integrated on chip. The PLL2 can operate to modulate the output frequency between its nominal value and a value that is up to -0.5% lower. This provides spread spectrum modulation up to -0.5% downspread. Spread spectrum frequency modulation can reduce the peak power output of the clock source and any circuitry that it drives and lead to reduced EMI emissions. In the AD9577, the frequency modulation profile is triangular. The modulation frequency and modulation range parameters are all programmable. Both PLLs can be programmed to generate a second independent frequency map under the control of the MARGIN pin. This feature can be used to test the frequency robustness of a system. REFERENCE INPUT AND REFERENCE DIVIDERS The reference input section is shown in Figure 34. When the REFSEL pin is pulled high, the crystal oscillator circuit is enabled. The crystal oscillator circuit needs an external crystal cut to resonate in fundamental mode in the 19.44 MHz to 27 MHz range, with 25 MHz being used in most networking applications. The total load capacitance presented to the crystal should add up to 14 pF. In the example shown in Figure 34, parasitic trace capacitance of 1.5 pF and an AD9577 input pin capacitance of 1.5 pF are assumed, with the series combination of the two 22 pF capacitances providing an additional 11 pF. When the REFSEL pin is pulled low, the crystal oscillator powers down, and the REFCLK pin must provide a good quality reference clock instead. Either a dc-coupled LVCMOS level signal or an ac-coupled square wave can drive this single-ended input, provided that an external potential divider is used to bias the input at VS/2. The output of the crystal oscillator and reference input circuitry is routed to a reference divider circuit to further divide down the reference input frequency to the PLLs by 1 or 2. When the crystal oscillator circuit is used, the dividers must be set to divide by 1. The input frequency to the PLLs must be in the 19.44 MHz to 27 MHz range. The divide ratio is set to 1 by programming the value of R, Register G0[1], to 0. The divide ratio is set to 2 by programming the value of R to 1. REFSEL 22pF XTAL OSC DIVIDE 1 OR 2 22pF REFCLK TO PLLs 09284-038 On the AD9577, parameters can be programmed over an I2C bus to provide custom output frequencies, output formats, and feature selections. However, this programming must be repeated after every power cycle of the part. Figure 34. Reference Input Section and Reference Dividers Table 17. REFSEL (Pin 9) Definition REFSEL 0 1 Reference Source REFCLK input Crystal oscillator Table 18. Reference Divider Setting R, Register G0[1] 0 1 Rev. 0 | Page 29 of 44 Reference Divide Ratio Divide by 1 Divide by 2 AD9577 Data Sheet OUTPUT CHANNEL DIVIDERS OUTPUTS Between each VCO and its associated chip outputs, there are two divider stages: a VCO divider that has a divide ratio between 2 and 6 and an output divider that can be set to divide between 1 and 32. This cascade of dividers allows a minimum output channel divide ratio of 2 and a maximum of 192. With VCO frequencies ranging between 2.15 GHz and 2.55 GHz, the part can be programmed to spot frequencies over a continuous frequency range of from 11.2 MHz to 200 MHz, and it can be programmed to spot frequencies over a continuous frequency range of 200 MHz and 637.5 MHz, with only a few small gaps. Each output port can be individually configured as either differential LVPECL, differential LVDS, or two single-ended LVCMOS clock outputs. The simplified equivalent circuit of the LVDS outputs is shown in Figure 36. 3.5mA OUTxP OUTxN Divider Channel 0 VCO divider Channel 1 VCO divider Channel 2 VCO divider Channel 3 VCO divider Channel 0 output divider Channel 1 output divider Channel 2 output divider Channel 3 output divider I C Registers ADV0[7:5] ADV1[7:5] BDV0[7:5] BDV1[7:5] ADV0[4:0] ADV1[4:0] BDV0[4:0] BDV1[4:0] Parameter V0 V1 V2 V3 D0 D1 D2 D3 Divide Range 2 to 6 2 to 6 2 to 6 2 to 6 1 to 321 1 to 321 1 to 321 1 to 321 3.5mA Figure 36. LVDS Outputs Simplified Equivalent Circuit The simplified equivalent circuit of the LVPECL outputs is shown in Figure 37. 3.3V OUTxP Set to 00000 for divide by 32. OUTxN Asserting the SyncCh01 or SyncCh23 bits (Register ADV2[0] or Register BDV2[0]) allows each PLL output channel to use a common VCO divider. This feature allows the OUT0/OUT1 and OUT2/OUT3 output ports to have minimal skew when their relative output channel divide ratio is an integer multiple. Duty-cycle correction circuitry ensures that the output duty cycle remains at 50%. VCO VCO DIVIDER OUTPUT DIVIDER V0[2:0] D0[4:0] VCO DIVIDER OUTPUT DIVIDER V1[2:0] D1[4:0] VCO DIVIDER OUTPUT DIVIDER V2[2:0] D2[4:0] VCO DIVIDER OUTPUT DIVIDER V3[2:0] D3[4:0] OUT0 GND Figure 37. LVPECL Outputs Simplified Equivalent Circuit Output channels (consisting of a VCO divider, output divider, and an output buffer) can be individually powered down to save power. Setting PDCH0, PDCH1, PDCH2, and PDCH3 (Register BP0[1:0] and Register DR1[7:6]) powers down the appropriate channel. Output buffer combinations of LVDS, LVPECL, and CMOS can be selected by setting DR1[5:0] as is shown in Table 20 and Table 21. OUT1 Table 20. PLL1 Output Driver Format Control Bits, Register DR1[2:0] VCO OUT2 09284-039 OUT3 Figure 35. Output Channel Divider Signal Path 09284-041 1 2 09284-040 Table 19. Divider Ratio Setting Registers FORMAT1 (PLL1) Register DR1[2:0] 000 001 010 011 100 101 110 1111 1 OUT1P/OUT1N LVPECL LVDS 2 x CMOS 2 x CMOS 2 x CMOS LVPECL LVPECL 2 x CMOS OUT0P/OUT0N LVPECL LVDS LVPECL 2 x CMOS LVDS LVDS 2 x CMOS 2 x CMOS This indicates that the CMOS outputs are in phase; otherwise, they are in antiphase. Rev. 0 | Page 30 of 44 Data Sheet AD9577 FORMAT2 (PLL2) Register DR1[5:3] 000 001 010 011 100 101 110 1111 1 OUT3P/OUT3N LVPECL LVDS 2 x CMOS 2 x CMOS 2 x CMOS LVPECL LVPECL 2 x CMOS OUT2P/OUT2N LVPECL LVDS LVPECL 2 x CMOS LVDS LVDS 2 x CMOS 2 x CMOS In ac-coupled applications, the LVPECL output stage needs a pair of 200 pull-down resistors to GND to provide a dc path for the output stage emitter followers (see Figure 41). The receiver must provide an additional 50 single-ended input termination. VTERM 50 50 LVPECL 50 100 LVDS 09284-042 LVDS 50 200 REFERENCE OUTPUT BUFFER A CMOS buffered copy of the reference input circuit signal is available at the REFOUT pin. This buffer can be optionally powered down by setting Register DR2[0], PDRefOut to Logic 0. PLL1 INTEGER-N PLL The upper PLL in Figure 32, PLL1, is an integer-N PLL with a loop bandwidth of 140 kHz. The input frequency to the PLL from the reference circuit is fPFD. The VCO frequency, fVCO1, is programmed by setting the value for Na, according to fVCO1 = fPFD x Na See the AN-586 Application Note, LVDS Outputs for High Speed A/D Converters, for more information about LVDS. In a dc-coupled application, the LVPECL output buffer should be terminated via a pair of 50 resistors to a voltage of VCC - 2 V. This can be implemented by using potential dividers of 127 and 83 between the supplies, as shown in Figure 39. 3.3V 127 50 LVPECL 127 3.3V (3) where Na is programmable in the 80 to 131 range. The VCO output frequency can tune over the 2.15 GHz to 2.55 GHz range to integer multiples of the PFD input frequency only. By setting each of the VCO divider (V0 and V1) and output divider (D0 and D1) values, the VCO frequency can be divided down to the required output frequency, independently, for each of the output ports, OUT0 and OUT1. The fOUT0 frequency presented to OUT0 can be set according to f OUT 0 = f PFD x SINGLE-ENDED (NOT COUPLED) 200 Figure 41. LVPECL AC-Coupled Termination Figure 38. LVDS Output Termination 3.3V LVPECL 0.1F 50 This indicates that the CMOS outputs are in phase; otherwise, they are in antiphase. LVDS uses a current mode output stage. The normal value (default) for this current is 3.5 mA, which yields a 350 mV output swing across a 100 resistor. The LVDS outputs meet or exceed all ANSI/TIA/EIA-644 specifications. The LVDS output buffer should be terminated with a 100 differential resistor between the receiver input ports (see Figure 38). A recommended termination circuit for the LVDS outputs is shown in Figure 38. 50 0.1F 09284-045 Table 21.PLL2 Output Driver Format Control Bits, Register DR1[5:3] LVPECL Na V0 x D0 (4) The frequency fOUT1 presented to OUT1 can be set according to 50 83 83 f OUT 1 = f PFD x 09284-043 VT = VDD - 2V The loop filters required for this PLL are integrated on chip. Figure 39. LVPECL DC-Coupled Termination An alternative LVPECL termination scheme for dc-coupled applications is shown Figure 40. 50 LVPECL LVPECL 50 50 50 09284-044 50 Na V 1 x D1 Figure 40. LVPECL DC-Coupled Y-Termination Rev. 0 | Page 31 of 44 (5) AD9577 Data Sheet PLL1 PHASE FREQUENCY DETECTOR (PFD) AND CHARGE PUMP The PFD determines the phase difference error between the reference divider output and the feedback divider output clock edges. The outputs of this circuit are pulse-width modulated up and down signal pulses. These pulses drive the charge pump circuit. The amount of charge delivered from the charge pump to the loop filter is determined by the instantaneous phase error. The action of the closed loop is to drive the frequency and phase error at the input of the PFD toward zero. Figure 42 shows a block diagram of the PFD/CP circuitry. 3.3V HIGH REFCLK D1 Q1 UP LCM(156.25 MHz, 100 MHz) = 2.5 GHz (6) Therefore, set the VCO frequency to 2.5 GHz. With fPFD = 25 MHz, from Equation 3, Na must be set to 100. For 156.25 MHz on Port 0, set V0 x D0 = 16 (7) This can be achieved by setting V0 to 4 and D0 to 4. For 100 MHz on Port 1, set CHARGE PUMP V1 x D1 = 25 (8) This can be achieved by setting V1 to 5 and D1 to 5. With a reference frequency of 25 MHz, the reference divider value, R, must be set to 1 by setting Register G0[1] to 0. Table 22 summarizes the register settings for this configuration. CLR1 CP Table 22. Register Settings for Example PLL1 Configuration CLR2 DOWN D2 Q2 FEEDBACK DIVIDER GND 09284-047 HIGH To determine if both 156.25 MHz and 100 MHz can be derived from a common fVCO1 frequency in the 2.15 GHz to 2.55 GHz range, use the lowest common multiple (LCM) of 156.25 MHz and 100 MHz to determine the lowest VCO frequency that can be divided down to provide both of these frequencies. Figure 42. PFD Circuit Showing Simplified Charge Pump PLL1 VCO PLL1 incorporates a low phase noise LC-tank VCO. This VCO has 32 frequency bands spanning from 2.15 GHz to 2.55 GHz. At power-up, a VCO calibration cycle begins and the correct band is selected based on the feedback divider setting (Na). Whenever a new feedback divider setting is called for, the VCO calibration process must run by writing 1 followed by 0 to the NewAcq bit, Register X0[0]. Parameter Na V0 D0 V1 D1 R Divide Value 100 4 4 5 5 1 The lower PLL in Figure 32, PLL2, is a fractional-N PLL. The input frequency to the PLL from the reference circuit is fPFD. The VCO frequency, fVCO2, is programmed by setting the values for Nb, FRAC, and MOD according to f VCO 2 = f PFD x (Nb + SETTING THE OUTPUT FREQUENCY OF PLL1 For example, set the output frequency (fOUT0) on Port 0 to 156.25 MHz, the output frequency (fOUT1) on Port 1 to 100 MHz, and both the reference frequency (fREF) and the PFD frequency (fPFD) to 25 MHz. Register Value 010100 100 00100 101 00101 1 PLL2 INTEGER/FRACTIONAL-N PLL PLL1 FEEDBACK DIVIDER The feedback divider ratio, Na, is used to set the PLL1 VCO frequency according to Equation 3. Note that the Na value is set by adding the offset value of 80 to the value programmed to Register AF0[5:0], where 80 is the minimum divider Na value. The maximum Na value is 131. For example, to set Na to 85, the AF0[5:0] register is set to 5. I2C Register AF0[5:0] ADV0[7:5] ADV0[4:0] ADV1[7:5] ADV1[4:0] G0[1] FRAC ) MOD (9) where Nb is programmable in the 80 to 131 range. To provide the greatest flexibility and accuracy, both the FRAC and MOD values can be programmed to a resolution of 12 bits, where FRAC < MOD. The VCO output frequency can tune over the 2.15 GHz to 2.55 GHz range to fractional multiples of the PFD input frequency. By setting each of the VCO divider (V2 and V3) and output divider (D2 and D3) values, the VCO frequency can be divided down to the required output frequency, independently, for each of the output ports, OUT2 and OUT3. The fOUT2frequency presented to OUT2 can be set according to The frequency fOUT0 presented to OUT0 can be set according to Equation 4. The frequency fOUT1 presented to OUT1 can be set according to Equation 5. Rev. 0 | Page 32 of 44 f OUT2 = f PFD x FRAC ) MOD V2 x D2 (Nb + (10) Data Sheet AD9577 The fOUT3 frequency presented to OUT3 can be set according to fOUT3 = f PFD x FRAC ) MOD V3 x D3 (Nb + (11) The loop filters required for this PLL are integrated on chip. By setting the FRAC value to 0, powering down the SDM by setting Register ABF0[4] to 1, and turning the bleed current off by setting Register BP0[2] = 0, PLL2 can operate as an integer-N PLL. Equation 10 and Equation 11 are still used to set the output frequencies for fOUT2 and fOUT3. Operation in this mode provides improved performance in terms of phase noise, spurs, and jitter. PLL2 PHASE FREQUENCY DETECTOR (PFD) AND CHARGE PUMP The PLL2 PFD and charge pump is the same as that described in the PLL1 Phase Frequency Detector (PFD) and Charge Pump section. When operating in fractional-N mode, a charge pump bleed current should be enabled to linearize the PLL transfer function and, therefore, to minimize spurs due to the operation of the - modulator. Bleed is enabled by setting Register BP0[2]. PLL2 LOOP BANDWIDTH The normal PLL loop bandwidth is 50 kHz. When the SSCG input pin is asserted, the loop bandwidth switches from 50 kHz to 125 kHz, which prevents the triangle-wave modulation waveform from being overly filtered by the PLL. When the MAX_BW input pin is set high, it forces the PLL bandwidth to be 250 kHz instead of 125 kHz. PLL2 VCO PLL2 incorporates a low phase noise LC-tank VCO. This VCO has 32 frequency bands spanning from 2.15 GHz to 2.55 GHz. At power-up, a VCO calibration cycle begins and the correct band is selected based on the feedback divider setting (Nb). Whenever a new feedback divider setting is called for, the VCO calibration process must run by writing 1 followed by 0 to the NewAcq bit, Register X0[0]. PLL2 FEEDBACK DIVIDER The Nb feedback divider ratio is used to set the PLL2 VCO frequency according to Equation 9. Note that the Nb value is set by adding the decimal value programmed to Register BF3[5:0] to a decimal value of 80, where the minimum divider Nb value is 80. The maximum Nb value is 131. For example, to set Nb to 85, Register BF3[5:0] is set to 5. PLL2 - MODULATOR When operating in fractional-N mode only, PLL2 uses a thirdorder, multistage noise shaping (MASH) - modulator (SDM) to adjust the feedback divider ratio. The programmed Nb value can be adjusted over the -4 to +3 range on every rising clock edge from the feedback divider output (typically 25 MHz for networking applications). In this way, the average feedback divide ratio is adjusted to be a noninteger value, allowing for a VCO frequency that is a fractional multiple of the PFD frequency to be synthesized. By setting the FRAC and MOD values of the SDM, the PLL2 VCO frequency can be set according to Equation 9. The SDM must be turned on by setting PD_SDM to 0, Register ABF0[4]. 12-Bit Programmable Modulus (MOD) and Fractional (FRAC) Values Unlike most other fractional-N PLLs, the AD9577 allows users to program the modulus over a 12-bit range, which means they can set up the part in many different configurations. It also usually means that, in most applications, it is possible to design the PLL to achieve the desired output frequency multiplication with 0 ppm frequency error. The MOD value is set by setting Register BF1[3:0] and Register BF2[7:0]. The FRAC value is set by setting Register BF0[7:0] and Register BF1[7:4]. Bleed Current When the SDM is operational (Register ABF0[4] set to 0), bleed current should be enabled (Register BP0[2] set to 1), which increases the in-band phase noise but reduces the fractional spur amplitudes. All fractional-N jitter data is reported with bleed = 1. If bleed = 0 in fractional-N mode, the rms jitter decreases significantly; however, the fractional spur amplitudes increase. When PLL2 operates in integer-N mode, the bleed current should be disabled to improve the PLLs in-band phase noise. SPUR MECHANISMS This section describes the three different spur mechanisms that arise with a fractional-N PLL: fractional spurs, integer boundary spurs, and reference spurs. Fractional Spurs The fractional interpolator in the AD9577 is a third-order SDM with a modulus that is programmable to any integer value from 50 to 4095. The SDM is clocked at the PFD reference rate (fPFD) that allows PLL output frequencies to be synthesized at a channel step resolution of fPFD/MOD. The quantization noise from the - modulator appears as fractional spurs. The interval between spurs is fPFD/L, where L is the repeat length of the code sequence in the digital - modulator. For the third-order modulator used in the AD9577, the repeat length depends on the value of MOD, as listed in Table 23. Table 23. Fractional Spur Frequencies Condition If MOD is divisible by 2, but not 3 If MOD is divisible by 3, but not 2 If MOD is divisible by 6 Otherwise Rev. 0 | Page 33 of 44 Repeat Length 2 x MOD 3 x MOD 6 x MOD MOD Spur Interval fPFD/(2 x MOD) fPFD/(3 x MOD) fPFD/(6 x MOD) fPFD/MOD AD9577 Data Sheet Integer Boundary Spurs SETTING THE OUTPUT FREQUENCY OF PLL2 Another mechanism for fractional spur creation is the interactions between the RF VCO frequency and the reference frequency. When these frequencies are not integer related (the point of a fractional-N synthesizer), spur sidebands appear on the VCO output spectrum at an offset frequency that corresponds to the beat note or difference frequency, between an integer multiple of the reference and the VCO frequency. These spurs are attenuated by the loop filter and are more noticeable on channels close to integer multiples of the reference where the difference frequency can be inside the loop bandwidth; therefore, the name integer boundary spurs. For example, to set the output frequency (fOUT2) on Port 2 to 155.52 MHz and the output frequency (fOUT3) on Port 3 to 38.88 MHz using a reference frequency (fREF) and PFD frequency (fPFD) of 25 MHz, do the following. Reference Spurs Reference spurs occur for both integer-N and fractional-N operation. Reference spurs are generally not a problem in fractional-N synthesizers because the reference offset is far outside the loop bandwidth. However, any reference feedthrough mechanism that bypasses the loop may cause a problem. Feedthrough of low levels of on-chip reference switching noise, through the reference input or output pins back to the VCO, can result in noticeable reference spur levels. In addition, coupling of the reference frequency to the output clocks can result in beat note spurs. PCB layout needs to ensure adequate isolation between VCO/LDO supplies, the output traces, and the input or output reference to avoid a possible feedthrough path on the board. If the reference output clock (REFCLK) is not required, it should be powered down to minimize potential board coupling. The SDM digital circuitry is clocked by the reference clock. The SDM is enabled when PLL2 is in fractional-N mode. When PLL2 is in fractional-N mode, the switching noise at the reference frequency may result in increased spurs levels at the outputs. OPTIMIZING PLL PERFORMANCE Because the AD9577 can be configured in many ways, some guidelines should be followed to ensure that the high performance is maintained. For both PLLs, there can be a small advantage in choosing a lower VCO frequency because the VCO phase noise tends to be slightly better at lower frequencies. Both VCOs should not operate at the same frequency because this degrades jitter performance. The two VCO frequencies should differ by at least 2 MHz. The following guidelines apply to PLL2 operating in fractional-N mode only. If possible, denominators that have factors of 2, 3, or 6 should be avoided because they can produce slightly higher subfractional spur components. Avoid low and high fractions (that is, FRAC/MOD close to 1/MOD or (MOD - 1)/ MOD) because these are more susceptible to larger fractional spur components and integer boundary spurs. Avoid creating a low valued beat frequency between the output frequency and the PFD frequency to minimize the risk of low offset beat frequency spurs. For example, setting fPFD = 25 MHz, and fOUT = 100.01 MHz can create an output spur at 10 kHz offset to 100.01 MHz, depending on board layout. Choosing a smaller MOD value results in fractional spurs that are at a higher frequency and, consequently, are better filtered by the PLL loop filter bandwidth of 50 kHz. The frequency fOUT2 presented to OUT2 can be set according to Equation 10. The frequency fOUT3 presented to OUT3 can be set according to Equation 11. In this case, both 155.52 MHz and 38.88 MHz can be derived from the same VCO frequency because they are related by a factor of 4. The next step is to determine what the required values of fVCO2, V2, and D2 are to divide down to 155.52 MHz. Table 24 shows the available options. Table 24. Suitable Values of fVCO2 and V2 x D2, to Achieve fOUT2 = 155.52 MHz fOUT2 (MHz) 155.52 155.52 155.52 V2 x D2 14 15 16 fVCO2 (GHz) 2.17728 2.3328 2.48832 Choose a fVCO2 value of 2.48832 GHz. Next, determine that the multiplication ratio (Nb + FRAC/MOD) required to multiply a fPFD of 25 MHz up to 2.48832 GHz is 99.5328. Therefore, Nb must be set to 99 and (FRAC/MOD) = 0.5328. To convert 0.5328 to a fraction, 0.5328 can be the same as 5328/10000. This fraction can then be reduced to the lowest terms by dividing both the numerator and denominator by 16, where 16 is the greatest common divisor (GCD) of the 5328 and 10,000. This results in a solution for FRAC/MOD = 333/625. For 155.52 MHz on Port 2, set V2 x D2 = 16. This can be achieved by setting V2 to 4 and D2 to 4. For 38.88 MHz on Port 3, set V3 x D3 = 64. This can be achieved by setting V3 to 4 and D3 to 16. With a reference frequency of 25 MHz, the reference divider value, R, must be set to 1 by setting Register G0[1] to 0. Because both channels use VCO divide values of 4on V2 and V3, SyncCh23, Register BDV2[0], can be set to 1 to ensure that the clock edges on Port 2 and Port 3 are synchronized. Table 25 summarizes the register setting for this configuration. Table 25. Registers Setting for Example PLL2 Configuration Parameter Nb FRAC MOD V2 D2 V3 D3 R SyncCh23 Rev. 0 | Page 34 of 44 Value 99 333 625 4 4 4 16 1 1 I2C Register BF3[5:0] BF0[7:0], BF1[7:4] BF1[3:0], BF2[7:0] BDV0[7:5] BDV0[4:0] BDV1[7:5] BDV1[4:0] G0[1] BDV2[0] Register Value 010011 000101001101 001001110001 100 00100 100 10000 0000 1 Data Sheet AD9577 MARGINING fVCO VCO FREQUENCY fVCO - 0.5% DIVIDE BY 80 TO 131 FEEDBACK DIVIDER NB 3-BIT FRAC 0 FRAC_TRIWAVE 1 SPREAD SPECTRUM CLOCK GENERATION (SSCG) SDM MOD SSCG By asserting the SSCG (spread spectrum clock generator) pin, PLL2 operates in spread spectrum mode, and the output frequency modulates with a triangular profile. As the clock signal energy spreads out over a range of frequencies, it reduces the peak power at any one frequency when observed with a spectrum analyzer through a resolution bandwidth filter. This result improves the radiated emissions from the part and from the devices that receive its clock. The triangular-wave modulation is implemented by controlling the divide ratio of the feedback divider. This is achieved by ramping the fractional word to the SDM. Figure 43 shows an example implementation. The PFD frequency, fPFD, is 25 MHz. The starting VCO frequency, fVCO, is 25 MHz x (99 + 3072/4096), giving 2.49375 GHz. By continuously ramping the FRAC word down and up, this frequency is periodically reduced to 25 MHz x (99 + 1029/4096) = 2.481281 GHz. This results in a triangular frequency modulation profile, with a peak downspread (that is, peak percentage frequency reduction) of -0.5%. By controlling the step size, number of steps, and the step rate, the modulation frequency is adjusted. THIRD ORDER LPF fPFD TIME VCO 2.15GHz TO 2.55GHz FRAC_TRIWAVE = 3072 FRAC_TRIWAVE = 1029 TIME fPFD CKDIV FRAC FRACSTEP NUMSTEPS SSCG FRAC_TRIWAVE TRIWAVE GENERATOR 09284-048 When the MARGIN pin signal level is changed, a new frequency acquisition is performed. PFD/CP By asserting the MARGIN pin, a second full frequency map can be applied to the output ports. The values for the Na, V0, D0, V1, and D1 parameters, and the Nb, FRAC, MOD, V2, D2, V3, D3 parameters must be programmed over the I2C, although default values exist. There are some limitations: the output buffer signal formats cannot be changed, and the PLL2 fractional-N settings, such as power-down of the SDM, and bleed settings cannot be changed. The margining feature can be used to set higher than nominal frequencies on each of the ports to test system robustness. Figure 43. Spread Spectrum Clock Generator with Triangular Wave Modulation, fPFD = 25 MHz Basic Spread Spectrum Programming The SSCG is highly programmable; however, most applications require that the frequency modulation rate be between 30 kHz and 33 kHz and that the peak frequency deviation be -0.5% downspread. The AD9577 supports downspread only, with a maximum deviation of -0.5%. The key parameters (which are not themselves registers) that define the frequency modulation profile include the following: * * fMOD, which is the frequency of the modulation waveform. FracRange, which determines the peak frequency deviation by setting the maximum change in the FRAC value from the nominal. The following equations determine the value of these parameters: FracRange = FracStep x NumSteps f MOD = f PFD 2 x NumSteps x CkDiv (12) (13) where the following are programmable registers: * * * * Rev. 0 | Page 35 of 44 NumSteps is the number of fractional word steps in half the triwave period. FracStep is the value of the fractional word increment/ decrement, while traversing the tri-wave. CkDiv is the integer value by which the reference clock frequency is divided to determine the update rate of the triangular-wave generator, that is, the step update rate. fPFD is the PFD frequency. AD9577 Data Sheet Table 26 shows the relevant register names and programmable ranges. Table 27. CkDiv and FracStep Values Used in Worked Example Table 26. Registers Used to Program SSCG Operation Parameter NumSteps FracStep CkDiv Register Name BS2[7:0], BS3[7] BS1[7:0] BS3[6:0] Range +1 to +511 -128 to 0 +2 to +127 Because the register values need to be expressed as integers, there are no guaranteed exact solutions; therefore, some approximations and trade-offs must be made. The fact that neither FracRange nor fMOD needs to be exact is exploited. Note that the SSCG pin must be toggled every time the SSCG parameters are adjusted for the changes to take effect. Worked Example: Programming for fMOD = 31.25 kHz, Downspread = -0.5%, fPFD = 25 MHz Assume Nb = 100, MOD = 625, and FRAC = 198. In addition, a large number of frequency steps are desired to cover -0.5%. The objective is to find values for FracStep, NumSteps, and CkDiv that result in the required frequency modulation profile. CkDiv 2 3 4 5 6 7 8 9 10 FracStep = CkDiv x (-0.78375) (15) An approximate solution must be found to Equation 15 that produces an integer value for CkDiv, which gives a value that is very close to an integer for FracStep. In this case, considering CkDiv values in the range of 2 to 10 gives the FracStep values shown in Table 27. FracStep Error 21.6% 17.6% 4.5% 2.0% 6.0% 9.7% 4.5% 0.77% 2.0% FracRange = -7 x 45 = -315 The accuracy of this solution needs to be verified. Putting the derived values into Equation 13 gives f MOD = 25 MHz f PFD = = 30.86 kHz 2 x NumSteps x CkDiv 2 x 45 x 9 In addition, the percentage frequency deviation is obtained as FrequencyDeviation = By rearranging Equation 12 and Equation 13, it results in 2 x FracRange x f MOD (14) FracStep = CkDiv x f PFD Putting in the values for FracRange, fMOD, and fPFD from the previous information, the following results: Rounded FracStep -2 -2 -3 -4 -5 -5 -6 -7 -8 Both CkDiv and NumSteps must be integers. To minimize error, CkDiv = 9 and FracStep = -7 was chosen. With a target for FracRange = -313.5, Equation 12 is used to find the ideal value of NumSteps = 44.79, which is rounded to 45. From Equation 12, the actual used value for FracRange is The total feedback divider ratio is FRAC = 100 + 198/625 = 62,698/625 NTOT = Nb + MOD FracRange is set to -0.5% of 62,698, which results in an ideal value of -313.5. Ideal FracStep -1.5675 -2.35125 -3.135 -3.91875 -4.7025 -5.48625 -6.27 -7.05375 -7.8375 = 100 x FracRange MOD x N TOT 100 x -315 = -0.502% 62698 625 x 625 The fMOD and the percentage frequency deviation are very close to the target values. The register settings required for this example are detailed in Table 29. SSCG Register Summary Table 28 summarizes the programmable registers required to set up SSCG. Table 28. Register Values for SSCG Parameter NumSteps FracStep CkDiv FRAC MOD Nb Rev. 0 | Page 36 of 44 Register Names BS2[7:0], BS3[7] BS1[7:0] BS3[6:0] BF0[7:0], BF1[7:4] BF1[3:0], BF2[7:0] BF3[5:0] Range +1 to +511 -128 to 0 +2 to +127 0 to +4094 0 to +4095 0 to +51 Data Sheet AD9577 MAX_BW The normal bandwidth of PLL2 is 50 kHz. This low bandwidth is required to filter the SDM phase noise. When SSCG is activated, the bandwidth is increased to 125 kHz. There is a trade-off in setting the PLL bandwidth between allowing the triangular-wave modulation (that is, its higher order harmonics) to pass through the PLL unattenuated and passing more SDM phase noise through to the PLL output. Bringing the MAX_BW pin high changes the PLL bandwidth to 250 kHz from its default value of 125 kHz during SSCG operation. Increasing the PLL bandwidth results in more SDM phase noise being passed unfiltered through to the PLL output, but more of the triangular-wave harmonics are also passed through, improving the triangular-wave accuracy. Table 29. Register Values for SSCG Example Parameter NumSteps FracStep CkDiv FRAC MOD Nb Register Name BS2[7:0], BS3[7] BS1[7:0] BS3[6:0] BF0[7:0], BF1[7:4] BF1[3:0], BF2[7:0] BF3[5:0] Range +1 to +511 -128 to 0 +2 to +127 0 to +4094 0 to +4095 0 to +63 Value (Decimal) +45 -7 +9 +198 +625 80 + 20 = 100 Rev. 0 | Page 37 of 44 Value(Binary) 00101101 11111001 0001001 000011000110 001001110001 010100 AD9577 Data Sheet I2C INTERFACE TIMING AND INTERNAL REGISTER DESCRIPTION 1 0 0 0 0 0 0 X 0 = WR 1 = RD 09284-049 R/W CTRL SLAVE ADDRESS [6:0] S SLAVE ADDR, LSB = 0 (WR) A(S) SUB ADDR A(S) DATA A(S) DATA A(S) P 09284-050 Figure 44. Slave Address Configuration Figure 45. I2C Write Data Transfer SLAVE ADDR, LSB = 0 (WR) A(S) SUB ADDR S = START BIT A(S) = ACKNOWLEDGE BY SLAVE A(S) S SLAVE ADDR, LSB = 1 (RD) A(S) DATA A(M) DATA A(M) P P = STOP BIT A(M) = LACK OF ACKNOWLEDGE BY MASTER A(M) = ACKNOWLEDGE BY MASTER 09284-051 S Figure 46. I2C Read Data Transfer SDA SLAVE ADDRESS A6 SUB ADDRESS A5 A7 STOP BIT DATA A0 D7 D0 SCL S WR ACK ACK SLADDR[4:0] ACK SUB ADDR[6:1] DATA[6:1] Figure 47. I2C Data Transfer Timing tF tSU;DAT tHD;STA tBUF SDA tR tR tSU;STO tF tLOW tHIGH tHD;STA S tSU;STA tHD;DAT S Figure 48. I2C Port Timing Diagram Rev. 0 | Page 38 of 44 P S 09284-053 SCL P 09284-052 START BIT Data Sheet AD9577 Table 30. Internal Register Map Register Name C0 X0 BP0 AF0 BF3 BF0 BF1 BF2 ABF0 ADV0 ADV1 ADV2 BDV0 BDV1 BDV2 BS1 BS2 BS3 AM0 AM1 R/W W W W W W W W W W W W W W W W W W W W W Addr 0x40 0x1F 0x11 0x18 0x1C 0x19 0x1A 0x1B 0x1D 0x22 0x23 0x24 0x25 0x26 0x27 0x2A 0x2B 0x2C 0x30 0x31 AM2 W 0x32 BM0 BM1 BM2 BM3 BM4 W W W W W 0x33 0x34 0x35 0x36 0x37 BM5 W 0x38 DR1 W 0x3A DR2 G0 W W 0x3B 0x3D D7 0 0 0 0 0 D6 0 0 0 0 0 D5 0 0 0 D4 0 0 0 D3 0 0 0 D2 D1 D0 0 EnI2C 0 0 0 NewAcq Bleed PDCH1 PDCH0 Na[5:0], PLL1 feedback divider ratio Nb[5:0], PLL2 feedback divider ratio FRAC[11:4], SDM fractional word FRAC[3:0], SDM fractional word MOD[11:8], SDM modulus MOD[7:0], SDM modulus 1 1 0 PD_SDM 0 0 0 0 V0[2:0], Channel 0 VCO divider D0[4:0], Channel 0 output divider value V1[2:0], Channel 1 VCO divider D1[4:0], Channel 1 output divider value 0 0 0 0 0 0 0 SyncCh01 V2[2:0], Channel 2 VCO divider D2[4:0], Channel 2 output divider value V3[2:0], Channel 3 VCO divider D3[4:0], Channel 3 output divider value 0 0 0 0 0 0 0 SyncCh23 FracStep[7:0], SSCG fractional step size NumSteps[8:1], number of fractional word increments/decrements per half triangular-wave cycle NumSteps[0] CkDiv[6:0], reference divider output is divided by this integer to determine SSCG update rate 0 0 Na[5:0], PLL1 feedback divider ratio divider; MARGIN = 1 V0[2:0], Channel 0 VCO divider; D0[4:0], Channel 0 output divider value; MARGIN = 1 MARGIN = 1 V1[2:0], Channel 1 VCO divider; D1[4:0], Channel 1 output divider value; MARGIN = 1 MARGIN = 1 0 0 Nb[5:0], PLL2 feedback divider ratio divider; MARGIN = 1 FRAC[11:4], SDM fractional word; MARGIN = 1 FRAC[3:0], SDM fractional word; MARGIN = 1 MOD[11:8], SDM modulus; MARGIN = 1 MOD[7:0], SDM modulus; MARGIN = 1 V3[2:0], Channel 3 VCO divider; D3[4:0], Channel 3 output divider value; MARGIN = 1 MARGIN = 1 V2[2:0], Channel 2 VCO divider; D2[4:0], Channel 2 output divider value; MARGIN = 1 MARGIN = 1 PDCH3 PDCH2 FORMAT2[2:0], output format selection FORMAT1[2:0], output format selection for for PLL2 (see Table 21) PLL1 (see Table 20) 0 0 0 0 0 0 0 PDRefOut 0 0 0 0 PDPLL1, powerPDPLL2, powerR; 0 = 0 down PLL1 down PLL2 divide by 1 Rev. 0 | Page 39 of 44 AD9577 Data Sheet DEFAULT FREQUENCY MAP AND OUTPUT FORMATS Parameter Margining The power-up operation (without I2C programming) of the AD9577 is represented by a default frequency map and output formats (see Table 31). Table 31. Default Parameter Values, fPFD = 25 MHz Parameter PLL1 Value Na V0 D0 V1 D1 FORMAT1 SyncCh01 PLL2 80 + 20 = 100 4 4 4 5 000 0 Nb FRAC MOD PD_SDM Bleed V2 D2 V3 D3 FORMAT2 SyncCh23 SSCG FracStep NumSteps CkDiv Control EnI2C NewAcq PDCH0 PDCH1 PDCH2 PDCH3 PDRefOut PDPLL1 PDPLL2 R 80 + 16 = 96 0 0 1 0 4 6 4 18 000 0 Notes fOUT0 = 156.25 MHz, fOUT1 = 125 MHz OUT0/OUT1 are LVPECL fOUT2 = 100 MHz, fOUT3 = 33.333 MHz 0 0 0 0 0 0 0 0 0 0 0 0 0 Value PLL1 Na V0 D0 V1 D1 fOUT0 fOUT1 PLL2 Nb FRAC MOD V2 D2 V3 D3 Notes These parameters are applied only when the MARGIN pin = high fOUT0 = 156.25 MHz, fOUT1 = 125 MHz 80 + 20 = 100 4 4 4 5 156.25 MHz 125 MHz fOUT2 = 212.5 MHz, fOUT3 = 106.25 MHz 80 + 22 = 102 0 0 2 6 4 6 I2C INTERFACE OPERATION OUT2/OUT3 are LVPECL The AD9577 is programmed by a 2-wire, I2C-compatible serial bus driving multiple peripherals. Two inputs, serial data (SDA) and serial clock (SCL), carry information between any devices connected to the bus. Each slave device is recognized by a unique address. The slave address consists of the 7 MSBs of an 8-bit word. The 7-bit slave address of the AD9577 is 1000000. The LSB of the word sets either a read or write operation (see Figure 44). Logic 1 corresponds to a read operation, and Logic 0 corresponds to a write operation. To control the device on the bus, do the following protocol. First, the master initiates a data transfer by establishing a start condition, defined by a high-to-low transition on SDA while SCL remains high, which indicates that an address/data stream follows. All peripherals respond to the start condition and shift the next eight bits (the 7-bit address and the R/W bit). The bits are transferred from MSB to LSB. The peripheral that recognizes the transmitted address responds by pulling the data line low during the ninth clock pulse, which is known as the acknowledge bit. All other devices withdraw from the bus at this point and maintain an idle condition. The idle condition is where the device monitors the SDA and SCL lines waiting for the start condition and correct transmitted address. The R/W bit determines the direction of the data. Logic 0 on the LSB of the first byte means that the master writes information to the peripheral, and Logic 1 on the LSB of the first byte means that the master reads information from the peripheral. Rev. 0 | Page 40 of 44 Data Sheet AD9577 The AD9577 acts as a standard slave device on the bus. The data on the SDA pin is eight bits long supporting the 7-bit addresses plus the R/W bit. The AD9577 has 31 subaddresses to enable the user-accessible internal registers (see Table 30). Therefore, it interprets the first byte as the device address and the second byte as the starting subaddress. Auto-increment mode is supported, which allows data to be read from or written to the starting subaddress and each subsequent address without manually addressing the subsequent subaddress. A data transfer is always terminated by a stop condition. The user can also access any unique subaddress register on a one-by-one basis without updating all registers. Stop and start conditions can be detected at any stage of the data transfer. If these conditions are asserted out of sequence with normal read and write operations, they cause an immediate jump to the idle condition. During a given SCL high period, one start condition, one stop condition, or a single stop condition followed by a single start condition should be issued. If an invalid subaddress is issued, the AD9577 does not issue an acknowledge and returns to the idle condition. If the highest subaddress is exceeded while reading back in auto-increment mode, the highest subaddress register contents continue to be output until the master device issues a no acknowledge, which indicates the end of a read. In a no acknowledge condition, the SDA line is not pulled low on the ninth pulse. See Figure 45 and Figure 46 for sample read and write data transfers, and see Figure 47 for a more detailed timing diagram. To overwrite any of the default register values, complete the following steps: 1. 2. 3. Enable the overwriting of registers by setting EnI2C, Register C0[1]. Only write to registers that need modification from their default value. After all the registers have been set, a new acquisition is initiated by toggling NewAcq, Register X0[0] from low to high to low. An example set of I2C commands follows. These enable the I2C registers and program the output frequencies of both PLLs. fPFD is 25 MHz. A leading W represents a write command. Table 32. I2C Programming Example Register Writes Write/Read W W W W W W W W W W W W W W Register Name C0 AF0 ADV0 ADV1 BF3 BF0 BF1 BF2 ABF0 BP0 BDV0 BDV1 X0 X0 Data (Hex) 02 0A A6 CC 15 14 D2 71 C0 04 44 B0 01 00 Operation Enable I2C registers Na = 80 + 10 = 90; fVCO1 = 2.25 GHz Channel 0 divides by 5 x 6 = 30; fOUT0 = 75 MHz Channel 1 divides by 6 x 12 = 72; fOUT1 = 31.25 MHz Nb = 80 + 21 = 101; FVCO2 = 2.53832 GHz FRAC = 333 FRAC = 333, MOD = 625 MOD = 625 Power-up SDM, release SDM reset Turn on Bleed Channel 2 divides by 2 x 4 = 8; fOUT2 = 317.29 MHz Channel 3 divides by 5 x 16 = 80; fOUT3 = 31.729 MHz Force new acquisition by toggling NewAcq Rev. 0 | Page 41 of 44 AD9577 Data Sheet TYPICAL APPLICATION CIRCUITS VS CD 10k VS 100 DIFFERENTIAL TRANSMISSION LINE RT = 100 10k VS SDA OUT0N VSOB0A OUT0P GND GND SCL VSVA 30 VSI2C OUT1P REFOUT OUT1N VS VSREFOUT VS VSX 22pF VS VS 100 DIFFERENTIAL R = 100 TRANSMISSION LINE T VS VSFA SSCG AD9577 VSM VS XT2 VSFB VS XT1 OUT3P REFSEL OUT3N REFCLK 22pF CD VSOB1A VS VSOB3B MARGIN 21 CD 20 VSOB2B OUT2P OUT2N GND GND LDO VSVB 11 TST1B TST2B VSCB CD 10 CD 220nF CD VS 100 DIFFERENTIAL TRANSMISSION LINE DO NOT CONNECT OTHER TRACES TO PIN 15, PIN 16, PIN 35, AND PIN 36. 100 DIFFERENTIAL R = 100 TRANSMISSION LINE T CAPACITORS C D CONSIST OF 100nF IN PARALLEL WITH 10nF. VS RT = 100 Figure 49. Typical LVDS Application Circuit Rev. 0 | Page 42 of 44 09284-054 VSCA TST2A 40 MAX_BW CD 1 VS VS 31 CD Data Sheet AD9577 VS 127 83 127 83 VS 50 10k 50 VS CD 10k VS SDA VSOB0A OUT0N OUT0P GND GND SCL VSVA 30 VSOB1A VSI2C OUT1P 50 REFOUT OUT1N 50 VS VSREFOUT VS VSX 22pF VS VS VS VS VSFA VSM VS XT2 VSFB VS XT1 OUT3P 50 REFSEL OUT3N 50 21 127 83 83 VS 127 127 83 83 CD 20 MARGIN VSOB2B OUT2P OUT2N GND GND VSVB TST2B TST1B 11 CD 220nF 50 CD VS VS CAPACITORS C D CONSIST OF 100nF IN PARALLEL WITH 10nF. 50 DO NOT CONNECT OTHER TRACES TO PIN 15, PIN 16, PIN 35, AND PIN 36. VS VSOB3B VSCB CD 10 127 SSCG AD9577 REFCLK 22pF CD VS 127 83 127 83 09284-055 VSCA LDO VS TST2A 40 MAX_BW CD 1 VS 31 CD Figure 50. Typical LVPECL Application Circuit POWER AND GROUNDING CONSIDERATIONS AND POWER SUPPLY REJECTION Many applications seek high speed and performance under less than ideal operating conditions. In these application circuits, the implementation and construction of the PCB is as important as the circuit design. Proper RF techniques must be used for device selection, placement, and routing, as well as for power supply bypassing and grounding to ensure optimum performance. Each power supply pin should have independent decoupling and connections to the power supply plane. It is recommended that the device exposed paddle be directly connected to the ground plane by a grid of at least nine vias. Care should be taken to ensure that the output traces cannot couple onto the reference or crystal input circuitry. Rev. 0 | Page 43 of 44 AD9577 Data Sheet OUTLINE DIMENSIONS 6.10 6.00 SQ 5.90 31 40 30 0.50 BSC 1 0.80 0.75 0.70 0.45 0.40 0.35 4.60 SQ 4.50 10 11 20 BOTTOM VIEW 0.05 MAX 0.02 NOM COPLANARITY 0.08 0.20 REF SEATING PLANE *4.70 EXPOSED PAD 21 TOP VIEW PIN 1 INDICATOR 0.20 MIN FOR PROPER CONNECTION OF THE EXPOSED PAD, REFER TO THE PIN CONFIGURATION AND FUNCTION DESCRIPTIONS SECTION OF THIS DATA SHEET. 02-02-2010-A PIN 1 INDICATOR 0.30 0.25 0.18 COMPLIANT TO JEDEC STANDARDS MO-220-WJJD-5 WITH EXCEPTION TO EXPOSED PAD DIMENSION. Figure 51. 40-Lead Lead Frame Chip Scale Package [LFCSP_WQ] 6 mm x 6 mm Body, Very Very Thin Quad (CP-40-7) Dimensions shown in millimeters ORDERING GUIDE Model 1 AD9577BCPZ AD9577BCPZ-RL AD9577BCPZ-R7 AD9577-EVALZ 1 Temperature Range -40C to +85C -40C to +85C -40C to +85C Package Description 40-Lead LFCSP_WQ 40-Lead LFCSP_WQ, 13" Tape Reel 40-Lead LFCSP_WQ, 7" Tape Reel Evaluation Board Package Option CP-40-7 CP-40-7 CP-40-7 Z = RoHS Compliant Part. I2C refers to a communications protocol originally developed by Philips Semiconductors (now NXP Semiconductors). (c)2011 Analog Devices, Inc. All rights reserved. Trademarks and registered trademarks are the property of their respective owners. D09284-0-10/11(0) Rev. 0 | Page 44 of 44 Ordering Quantity 2,500 750