622 Mbps Clock and Data Recovery IC with Integrated Limiting Amplifier ADN2804 FEATURES GENERAL DESCRIPTION Exceeds SONET requirements for jitter transfer/ generation/tolerance Quantizer sensitivity: 3.3 mV typical Adjustable slice level: ±95 mV Patented clock recovery architecture Loss-of-signal (LOS) detect range: 2.6 mV to 18.4 mV Independent slice level adjust and LOS detector No reference clock required Loss-of-lock indicator I2C® interface to access optional features Single-supply operation: 3.3 V Low power: 423 mW typical 5 mm × 5 mm, 32-lead LFCSP, Pb free The ADN2804 provides the receiver functions of quantization, signal level detect, clock and data recovery, and data retiming for 622 Mbps NRZ data. The ADN2804 automatically locks to 622 Mbps data without the need for an external reference clock or programming. In the absence of input data, the output clock drifts no more than ±5%. All SONET jitter requirements are met, including jitter transfer, jitter generation, and jitter tolerance. All specifications are quoted for −40°C to +85°C ambient temperature, unless otherwise noted. This device, together with a PIN diode and a TIA preamplifier, can implement a highly integrated, low cost, low power fiber optic receiver. The receiver’s front-end loss-of-signal (LOS) detector circuit indicates when the input signal level falls below a user-adjustable threshold. The LOS detect circuit has hysteresis to prevent chatter at the output. APPLICATIONS BPON ONT SONET OC-12 WDM transponders Regenerators/repeaters Test equipment Broadband cross-connects and routers The ADN2804 is available in a compact 5 mm × 5 mm, 32-lead LFCSP. FUNCTIONAL BLOCK DIAGRAM REFCLKP/REFCLKN (OPTIONAL) SLICEP/SLICEN 2 LOL CF1 CF2 FREQUENCY DETECT LOOP FILTER PHASE DETECT LOOP FILTER VCC VEE PIN NIN QUANTIZER PHASE SHIFTER VCO VREF DATA RE-TIMING 2 2 THRADJ LOS DATAOUTP/ DATAOUTN CLKOUTP/ CLKOUTN ADN2804 05801-001 LOS DETECT Figure 1. Rev. 0 Information furnished by Analog Devices is believed to be accurate and reliable. However, no responsibility is assumed by Analog Devices for its use, nor for any infringements of patents or other rights of third parties that may result from its use. Specifications subject to change without notice. No license is granted by implication or otherwise under any patent or patent rights of Analog Devices. Trademarks and registered trademarks are the property of their respective owners. One Technology Way, P.O. Box 9106, Norwood, MA 02062-9106, U.S.A. Tel: 781.329.4700 www.analog.com Fax: 781.461.3113 ©2006 Analog Devices, Inc. All rights reserved. ADN2804 TABLE OF CONTENTS Features .............................................................................................. 1 Jitter Specifications......................................................................... 13 Applications....................................................................................... 1 Theory of Operation ...................................................................... 14 General Description ......................................................................... 1 Functional Description.................................................................. 16 Functional Block Diagram .............................................................. 1 Frequency Acquisition............................................................... 16 Revision History ............................................................................... 2 Limiting Amplifier ..................................................................... 16 Specifications..................................................................................... 3 Slice Adjust.................................................................................. 16 Jitter Specifications....................................................................... 4 Loss-of-Signal (LOS) Detector ................................................. 16 Output and Timing Specifications ............................................. 5 Lock Detector Operation .......................................................... 17 Absolute Maximum Ratings............................................................ 6 SQUELCH Modes ...................................................................... 17 Thermal Characteristics .............................................................. 6 I2C Interface ................................................................................ 17 ESD Caution.................................................................................. 6 Reference Clock (Optional) ...................................................... 19 Timing Characteristics..................................................................... 7 Applications Information .............................................................. 21 Pin Configuration and Function Descriptions............................. 8 PCB Design Guidelines ............................................................. 21 Typical Performance Characteristics ............................................. 9 DC-Coupled Application .......................................................... 23 I2C Interface Timing and Internal Register Description........... 10 Outline Dimensions ....................................................................... 24 Terminology .................................................................................... 12 Ordering Guide .......................................................................... 24 REVISION HISTORY 2/06—Revision 0: Initial Version Rev. 0 | Page 2 of 24 ADN2804 SPECIFICATIONS TA = TMIN to TMAX, VCC = VMIN to VMAX, VEE = 0 V, CF = 0.47 μF, SLICEP = SLICEN = VEE, input data pattern: PRBS 223 − 1, unless otherwise noted. Table 1. Parameter QUANTIZER—DC CHARACTERISTICS Input Voltage Range Peak-to-Peak Differential Input Input Common-Mode Level Differential Input Sensitivity Input Offset Input RMS Noise QUANTIZER—AC CHARACTERISTICS Data Rate Output Clock Range S11 Input Resistance Input Capacitance QUANTIZER—SLICE ADJUSTMENT Gain Differential Control Voltage Input Control Voltage Range Slice Threshold Offset LOSS-OF-SIGNAL (LOS) DETECT Loss-of-Signal Detect Range (see Figure 6) Hysteresis (Electrical) LOS Assert Time LOS Deassert Time LOSS-OF-LOCK (LOL) DETECT VCO Frequency Error for LOL Assert VCO Frequency Error for LOL Deassert LOL Response Time ACQUISITION TIME Lock to Data Mode Optional Lock to REFCLK Mode DATA RATE READBACK ACCURACY Fine Readback POWER SUPPLY VOLTAGE POWER SUPPLY CURRENT OPERATING TEMPERATURE RANGE 1 2 Conditions Min @ PIN or NIN, dc-coupled PIN − NIN DC-coupled (see Figure 27, Figure 28, and Figure 29) 223 − 1 PRBS, ac-coupled, 1 BER = 1 × 10–10 1.8 2.3 6 BER = 1 × 10–10 2.5 3.3 500 290 Max Unit 2.8 2.0 2.8 V V V mV p-p μV μV rms 622 622 ± 5% −15 100 0.65 Absence of input data @ 622 MHz Differential SLICEP − SLICEN = ±0.5 V SLICEP − SLICEN DC level @ SLICEP or SLICEN Typ 0.10 −0.95 VEE 0.11 Mbps MHz dB Ω pF 0.13 +0.95 0.95 V/V V V mV 1 RTHRESH = 0 Ω RTHRESH = 100 kΩ OC-12 RTHRESH = 0 Ω RTHRESH = 100 kΩ DC-coupled 2 DC-coupled2 14.9 2.6 16.7 3.5 18.4 4.4 mV mV 6.2 4.1 6.9 6.1 500 400 7.7 8.1 dB dB ns ns With respect to nominal With respect to nominal OC-12 1000 250 200 ppm ppm μs OC-12 2.0 20.0 ms ms In addition to REFCLK accuracy OC-12 3.0 Locked to 622.08 Mbps –40 100 3.3 128 3.6 +85 ppm V mA °C PIN and NIN should be differentially driven and ac-coupled for optimum sensitivity. When ac-coupled, the LOS assert and deassert times are dominated by the RC time constant of the ac coupling capacitor and the 50 Ω input termination of the ADN2804 input stage. Rev. 0 | Page 3 of 24 ADN2804 JITTER SPECIFICATIONS TA = TMIN to TMAX, VCC = VMIN to VMAX, VEE = 0 V, CF = 0.47 μF, SLICEP = SLICEN = VEE, input data pattern: PRBS 223 − 1, unless otherwise noted. Table 2. Parameter PHASE-LOCKED LOOP CHARACTERISTICS Jitter Transfer Bandwidth Jitter Peaking Jitter Generation Jitter Tolerance 1 Conditions Min OC-12 OC-12 OC-12, 12 kHz to 5 MHz OC-12, 223 − 1 PRBS 30 Hz 1 300 Hz1 25 kHz 250 kHz1 100 44 2.5 1.0 Jitter tolerance of the ADN2804 at these jitter frequencies is better than what the test equipment is able to measure. Rev. 0 | Page 4 of 24 Typ Max Unit 75 0 0.001 0.011 130 0.03 0.003 0.026 kHz dB UI rms UI p-p UI p-p UI p-p UI p-p UI p-p ADN2804 OUTPUT AND TIMING SPECIFICATIONS Table 3. Parameter LVDS OUTPUT CHARACTERISTICS (CLKOUTP/CLKOUTN, DATAOUTP/DATAOUTN) Output Voltage High Output Voltage Low Differential Output Swing Output Offset Voltage Output Impedance LVDS Outputs’ Timing Rise Time Fall Time Setup Time Hold Time I2C INTERFACE DC CHARACTERISTICS Input High Voltage Input Low Voltage Input Current Output Low Voltage I2C INTERFACE TIMING SCK Clock Frequency SCK Pulse Width High SCK Pulse Width Low Start Condition Hold Time Start Condition Setup Time Data Setup Time Data Hold Time SCK/SDA Rise/Fall Time Stop Condition Setup Time Bus Free Time Between a Stop and a Start REFCLK CHARACTERISTICS Input Voltage Range Minimum Differential Input Drive Reference Frequency Required Accuracy LVTTL DC INPUT CHARACTERISTICS Input High Voltage Input Low Voltage Input High Current Input Low Current LVTTL DC OUTPUT CHARACTERISTICS Output High Voltage Output Low Voltage 1 Conditions Min VOH (see Figure 3) VOL (see Figure 3) VOD (see Figure 3) VOS (see Figure 3) Differential 925 250 1125 20% to 80% 80% to 20% TS (see Figure 2), OC-12 TH (see Figure 2), OC-12 LVCMOS VIH VIL VIN = 0.1 VCC or VIN = 0.9 VCC VOL, IOL = 3.0 mA See Figure 11 760 760 Typ Max Unit 1475 mV mV mV mV Ω 320 1200 100 400 1275 115 115 800 800 220 220 840 840 ps ps ps ps 0.3 VCC +10.0 0.4 V V μA V 0.7 VCC −10.0 400 tHIGH tLOW tHD;STA tSU;STA tSU;DAT tHD;DAT TR/TF tSU;STO tBUF Optional lock to REFCLK mode @ REFCLKP or REFCLKN VIL VIH 600 1300 600 600 100 300 20 + 0.1 Cb 1 600 1300 300 0 VCC 100 10 160 100 VIH VIL IIH, VIN = 2.4 V IIL, VIN = 0.4 V 2.0 VOH, IOH = −2.0 mA VOL, IOL = +2.0 mA 2.4 Cb = total capacitance of one bus line in picofarads. If used with Hs-mode devices, faster fall times are allowed. Rev. 0 | Page 5 of 24 0.8 5 −5 0.4 kHz ns ns ns ns ns ns ns ns ns V V mV p-p MHz ppm V V μA μA V V ADN2804 ABSOLUTE MAXIMUM RATINGS TA = TMIN to TMAX, VCC = VMIN to VMAX, VEE = 0 V, CF = 0.47 μF, SLICEP = SLICEN = VEE, unless otherwise noted. Table 4. Parameter Supply Voltage (VCC) Minimum Input Voltage (All Inputs) Maximum Input Voltage (All Inputs) Maximum Junction Temperature Storage Temperature Range Rating 4.2 V VEE − 0.4 V VCC + 0.4 V 125°C −65°C to +150°C Stress 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 sections of this specification is not implied. Exposure to absolute maximum rating conditions for extended periods may affect device reliability. THERMAL CHARACTERISTICS Thermal Resistance 32-lead LFCSP, 4-layer board with exposed paddle soldered to VEE, θJA = 28°C/W. ESD CAUTION ESD (electrostatic discharge) sensitive device. Electrostatic charges as high as 4000 V readily accumulate on the human body and test equipment and can discharge without detection. Although this product features proprietary ESD protection circuitry, permanent damage may occur on devices subjected to high energy electrostatic discharges. Therefore, proper ESD precautions are recommended to avoid performance degradation or loss of functionality. Rev. 0 | Page 6 of 24 ADN2804 TIMING CHARACTERISTICS CLKOUTP TH 05801-002 TS DATAOUTP/ DATAOUTN Figure 2. Output Timing DIFFERENTIAL CLKOUTP/N, DATAOUTP/N VOH VOS 05801-032 |VOD| VOL Figure 3. Differential Output Specifications 5mA RLOAD 100Ω 100Ω VDIFF SIMPLIFIED LVDS OUTPUT STAGE Figure 4. Differential Output Stage Rev. 0 | Page 7 of 24 05801-033 5mA ADN2804 32 TEST2 31 VCC 30 VEE 29 DATAOUTP 28 DATAOUTN 27 SQUELCH 26 CLKOUTP 25 CLKOUTN PIN CONFIGURATION AND FUNCTION DESCRIPTIONS ADN2804* TOP VIEW (Not to Scale) 24 VCC 23 VEE 22 LOS 21 SDA 20 SCK 19 SADDR5 18 VCC 17 VEE * THERE IS AN EXPOSED PAD ON THE BOTTOM OF THE PACKAGE THAT MUST BE CONNECTED TO GND. 05801-004 PIN 1 INDIC ATOR THRADJ 9 REFCLKP 10 REFCLKN 11 VCC 12 VEE 13 CF2 14 CF1 15 LOL 16 TEST1 1 VCC 2 VREF 3 NIN 4 PIN 5 SLICEP 6 SLICEN 7 VEE 8 Figure 5. Pin Configuration Table 5. Pin Function Descriptions Pin No. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 Exposed Pad 1 Mnemonic TEST1 VCC VREF NIN PIN SLICEP SLICEN VEE THRADJ REFCLKP REFCLKN VCC VEE CF2 CF1 LOL VEE VCC SADDR5 SCK SDA LOS VEE VCC CLKOUTN CLKOUTP SQUELCH DATAOUTN DATAOUTP VEE VCC TEST2 Pad Type 1 P AO AI AI AI AI P AI DI DI P P AO AO DO P P DI DI DI DO P P DO DO DI DO DO P P P Description Connect to VCC. Power for Limiting Amplifier, LOS. Internal VREF Voltage. Decouple to GND with a 0.1 μF capacitor. Differential Data Input. CML. Differential Data Input. CML. Differential Slice Level Adjust Input. Differential Slice Level Adjust Input. GND for Limiting Amplifier, LOS. LOS Threshold Setting Resistor. Differential REFCLK Input. 10 MHz to 160 MHz. Differential REFCLK Input. 10 MHz to 160 MHz. VCO Power. VCO GND. Frequency Loop Capacitor. Frequency Loop Capacitor. Loss-of-Lock Indicator. LVTTL active high. FLL Detector GND. FLL Detector Power. Slave Address Bit 5. I2C Clock Input. I2C Data Input. Loss-of-Signal Detect Output. Active high. LVTTL. Output Buffer, I2C GND. Output Buffer, I2C Power. Differential Recovered Clock Output. LVDS. Differential Recovered Clock Output. LVDS. Disable Clock and Data Outputs. Active high. LVTTL. Differential Recovered Data Output. LVDS. Differential Recovered Data Output. LVDS. Phase Detector, Phase Shifter GND. Phase Detector, Phase Shifter Power. Connect to VCC. Connect to GND. Type: P = power, AI = analog input, AO = analog output, DI = digital input, DO = digital output. Rev. 0 | Page 8 of 24 ADN2804 TYPICAL PERFORMANCE CHARACTERISTICS 16 12 10 8 6 4 2 1 10 100 1k 10k Figure 6. LOS Comparator Trip Point Programming Rev. 0 | Page 9 of 24 100k 05801-005 TRIP POINT (mV p-p) 14 ADN2804 I2C INTERFACE TIMING AND INTERNAL REGISTER DESCRIPTION 1 A5 MSB = 1 SET BY PIN 19 0 0 0 0 0 X 0 = WR 1 = RD 05801-007 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 05801-008 Figure 7. Slave Address Configuration 2 Figure 8. I C Write Data Transfer SLAVE ADDR, LSB = 0 (WR) A(S) SUB ADDR 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 05801-009 S S = START BIT A(S) = ACKNOWLEDGE BY SLAVE Figure 9. I2C Read Data Transfer SDA SLAVE ADDRESS A6 SUB ADDRESS A5 A7 STOP BIT DATA A0 D7 D0 SCK S WR ACK ACK SLADDR[4...0] ACK SUB ADDR[6...1] DATA[6...1] Figure 10. 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 2 S Figure 11. I C Port Timing Diagram Rev. 0 | Page 10 of 24 P S 05801-011 SCK P 05801-010 START BIT ADN2804 Table 6. Internal Register Map 1 Reg Name FREQ0 FREQ1 FREQ2 MISC R/W R R R R Addr 0x0 0x1 0x2 0x4 D7 MSB MSB 0 x CTRLA CTRLB W W 0x8 0x9 CTRLC W 0x11 FREF range Config Reset LOL MISC[4] 0 0 1 D6 D5 D4 D3 MSB x LOS status Static LOL LOL status System reset 0 D2 D1 Data rate measurement complete Data rate/DIV_FREF ratio 0 0 Reset MISC[2] 0 0 Config LOS x D0 LSB LSB LSB x Measure data rate 0 Lock to reference 0 SQUELCH mode Output boost All writeable registers default to 0x00. Table 7. Miscellaneous Register, MISC D7 x D6 x LOS Status D5 0 = No loss of signal 1 = Loss of signal Static LOL D4 0 = Waiting for next LOL 1 = Static LOL until reset LOL Status D3 0 = Locked 1 = Acquiring Data Rate Measurement Complete D2 0 = Measuring data rate 1 = Measurement complete D1 x D0 x Table 8. Control Register, CTRLA 1 FREF Range D7 D6 0 0 0 1 1 0 1 1 1 Data Rate/Div_FREF Ratio D5 D4 D3 D2 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1 19.44 MHz 38.88 MHz 77.76 MHz 155.52 MHz 32 32 32 32 Measure Data Rate D1 Set to 1 to measure data rate Lock to Reference D0 0 = Lock to input data 1 = Lock to reference clock Where DIV_FREF is the divided down reference referred to the 10 MHz to 20 MHz band (see the Reference Clock (Optional) section). Table 9. Control Register, CTRLB Config LOL D7 0 = LOL pin normal operation 1 = LOL pin is static LOL Reset MISC[4] D6 Write a 1 followed by 0 to reset MISC[4] System Reset D5 Write a 1 followed by 0 to reset ADN2804 D4 Set to 0 Reset MISC[2] D3 Write a 1 followed by 0 to reset MISC[2] D2 Set to 0 D1 Set to 0 D0 Set to 0 Table 10. Control Register, CTRLC D7 Set to 0 D6 Set to 0 D5 Set to 0 D4 Set to 0 D3 Set to 0 Config LOS D2 0 = Active high LOS 1 = Active low LOS Rev. 0 | Page 11 of 24 SQUELCH Mode D1 0 = Squelch data outputs and clock outputs 1 = Squelch data outputs or clock outputs Output Boost D0 0 = Default output swing 1 = Boost output swing ADN2804 TERMINOLOGY OUTPUT NOISE 1 10mV p-p VREF SCOPE PROBE ADN2804 PIN + QUANTIZER – 50Ω Figure 13. Single-Ended Sensitivity Measurement When the ADN2804 is driven differentially (see Figure 14), sensitivity seems to improve if observing the quantizer input with an oscilloscope probe. This is an illusion caused by the use of a single-ended probe. A 5 mV p-p signal appears to drive the ADN2804 quantizer; however, the single-ended probe measures only half the signal. The true quantizer input signal is twice this value, because the other quantizer input is a complementary signal to the signal being observed. 5mV p-p 0 INPUT (V p-p) SCOPE PROBE PIN + OVERDRIVE QUANTIZER 05801-012 SENSITIVITY (2 × OVERDRIVE) 2.5V 3kΩ VREF OFFSET 50Ω VREF 05801-013 Input Sensitivity and Input Overdrive Sensitivity and overdrive specifications for the quantizer involve offset voltage, gain, and noise. The relationship between the logic output of the quantizer and the analog voltage input is shown in Figure 12. For sufficiently large positive input voltage, the output is always Logic 1; similarly, for negative inputs, the output is always Logic 0. However, the transitions between output Logic Level 1 and output Logic Level 0 are not at precisely defined input voltage levels, but occur over a range of input voltages. Within this range of input voltages, the output may be either 1 or 0, or it may even fail to attain a valid logic state. The width of this zone is determined by the input voltage noise of the quantizer. The center of the zone is the quantizer input offset voltage. Input overdrive is the magnitude of signal required to guarantee the correct logic level with 1 × 10−10 confidence level. NIN – Figure 12. Input Sensitivity and Input Overdrive 50Ω VREF 50Ω VREF 5mV p-p 3kΩ 2.5V 05801-014 Single-Ended vs. Differential AC coupling is typically used to drive the inputs to the quantizer. The inputs are internally dc biased to a commonmode potential of ~2.5 V. Driving the ADN2804 in a singleended fashion and observing the quantizer input with an oscilloscope probe at the point indicated in Figure 13 shows a binary signal with an average value equal to the common-mode potential and instantaneous values both above and below the average value. It is convenient to measure the peak-to-peak amplitude of this signal and call the minimum required value the quantizer sensitivity. Referring to Figure 13, the sensitivity is twice the overdrive because both positive and negative offsets need to be accommodated. The ADN2804 quantizer typically has 3.3 mV p-p sensitivity. Figure 14. Differential Sensitivity Measurement LOS Response Time LOS response time is the delay between removal of the input signal and indication of loss of signal (LOS) at the LOS output, Pin 22. When the inputs are dc-coupled, the LOS assert time of the AD2804 is 500 ns typical and the deassert time is 400 ns typical. In practice, the time constant produced by the ac coupling at the quantizer input and the 50 Ω on-chip input termination determines the LOS response time. Rev. 0 | Page 12 of 24 ADN2804 JITTER SPECIFICATIONS The ADN2804 CDR is designed to achieve the best biterror-rate (BER) performance and to exceed the jitter transfer, generation, and tolerance specifications proposed for SONET/SDH equipment defined in the Telcordia Technologies specification. Jitter Generation The jitter generation specification limits the amount of jitter that can be generated by the device with no jitter and wander applied at the input. For SONET devices, the jitter generated must be less than 0.01 UI rms and less than 0.1 UI p-p. Jitter Transfer The jitter transfer function is the ratio of the jitter on the output signal to the jitter applied on the input signal vs. the frequency. This parameter measures the amount of jitter on an input signal that can be transferred to the output signal (see Figure 15). This amount is limited. 05801-015 fC JITTER FREQUENCY (kHz) Figure 15. Jitter Transfer Curve Jitter Tolerance The jitter tolerance is defined as the peak-to-peak amplitude of the sinusoidal jitter applied on the input signal, which causes a 1 dB power penalty. This is a stress test intended to ensure that no additional penalty is incurred under the operating conditions (see Figure 16). 15.00 SLOPE = –20dB/DECADE 1.50 0.15 f0 f1 f2 f3 JITTER FREQUENCY (kHz) Figure 16. SONET Jitter Tolerance Mask Rev. 0 | Page 13 of 24 f4 05801-016 The following sections briefly summarize the specifications of jitter generation, transfer, and tolerance in accordance with the Telcordia document (GR-253-CORE, Issue 3, September 2000) for the optical interface at the equipment level and the ADN2804 performance with respect to those specifications. SLOPE = –20dB/DECADE ACCEPTABLE RANGE INPUT JITTER AMPLITUDE (UI p-p) Jitter is the dynamic displacement of digital signal edges from their long-term average positions, measured in unit intervals (UI), where 1 UI = 1 bit period. Jitter on the input data can cause dynamic phase errors on the recovered clock sampling edge. Jitter on the recovered clock causes jitter on the retimed data. JITTER GAIN (dB) 0.1 ADN2804 THEORY OF OPERATION Another view of the circuit is that the phase shifter implements the zero required for frequency compensation of a second-order phase-locked loop, and this zero is placed in the feedback path; therefore, it does not appear in the closed-loop transfer function. Jitter peaking in a conventional second-order phaselocked loop is caused by the presence of this zero in the closedloop transfer function. Because this circuit has no zero in the closed-loop transfer, jitter peaking is minimized. The delay and phase loops together simultaneously provide wideband jitter accommodation and narrow-band jitter filtering. The linearized block diagram in Figure 17 shows that the jitter transfer function, Z(s)/X(s), provides excellent secondorder low-pass filtering. Note that the jitter transfer has no zero, unlike an ordinary second-order phase-locked loop. This means that the main PLL loop has virtually no jitter peaking (see Figure 18), making this circuit ideal for signal regenerator applications, where jitter peaking in a cascade of regenerators can contribute to hazardous jitter accumulation. The error transfer, e(s)/X(s), has the same high-pass form as an ordinary phase-locked loop. This transfer function can be optimized to accommodate a significant amount of wideband jitter, because the jitter transfer function, Z(s)/X(s), provides the narrow-band jitter filtering. INPUT DATA X(s) e(s) d/sc o/s 1/n Z(s) RECOVERED CLOCK d = PHASE DETECTOR GAIN o = VCO GAIN c = LOOP INTEGRATOR psh = PHASE SHIFTER GAIN n = DIVIDE RATIO JITTER TRANSFER FUNCTION Z(s) 1 = n psh cn X(s) +1 s2 +s o do TRACKING ERROR TRANSFER FUNCTION 05801-017 e(s) s2 = d psh do X(s) s2 + s + c cn Figure 17. PLL/DLL Architecture JITTER PEAKING IN ORDINARY PLL ADN2804 Z(s) X(s) o n psh d psh c FREQUENCY (kHz) 05801-018 The delay and phase loops together track the phase of the input data signal. For example, when the clock lags the input data, the phase detector drives the VCO to a higher frequency and increases the delay through the phase shifter; both of these actions serve to reduce the phase error between the clock and the data. The faster clock picks up phase, whereas the delayed data loses phase. Because the loop filter is an integrator, the static phase error is driven to 0°. psh JITTER GAIN (dB) The ADN2804 is a delay- and phase-locked loop circuit for clock recovery and data retiming from an NRZ encoded data stream. The phase of the input data signal is tracked by two separate feedback loops, which share a common control voltage. A high speed delay-locked loop path uses a voltage controlled phase shifter to track the high frequency components of input jitter. A separate phase control loop, composed of the VCO, tracks the low frequency components of input jitter. The initial frequency of the VCO is set by yet a third loop that compares the VCO frequency with the input data frequency and sets the coarse tuning voltage. The jitter tracking phase-locked loop controls the VCO by the fine-tuning control. Figure 18. Jitter Response vs. Conventional PLL The delay and phase loops contribute to overall jitter accommodation. At low frequencies of input jitter on the data signal, the integrator in the loop filter provides high gain to track large jitter amplitudes with small phase error. In this case, the VCO is frequency modulated, and jitter is tracked as in an ordinary phase-locked loop. The amount of low frequency jitter that can be tracked is a function of the VCO tuning range. A wider tuning range gives larger accommodation of low frequency jitter. The internal loop control voltage remains small for small phase errors; therefore, the phase shifter remains close to the center of its range and thus contributes little to the low frequency jitter accommodation. Rev. 0 | Page 14 of 24 ADN2804 At medium jitter frequencies, the gain and tuning range of the VCO are not large enough to track input jitter. In this case, the VCO control voltage becomes large and saturates, and the VCO frequency dwells at one extreme of its tuning range. The size of the VCO tuning range, therefore, has only a small effect on the jitter accommodation. The delay-locked loop control voltage is now larger; therefore, the phase shifter takes on the burden of tracking the input jitter. The phase shifter range, in UI, can be seen as a broad plateau on the jitter tolerance curve. The phase shifter has a minimum range of 2 UI at all data rates. The gain of the loop integrator is small for high jitter frequencies; therefore, larger phase differences are needed to increase the loop control voltage enough to tune the range of the phase shifter. However, large phase errors at high jitter frequencies cannot be tolerated. In this region, the gain of the integrator determines the jitter accommodation. Because the gain of the loop integrator declines linearly with frequency, jitter accommodation is lower with higher jitter frequency. At the highest frequencies, the loop gain is very small, and little tuning of the phase shifter can be expected. In this case, jitter accommodation is determined by the eye opening of the input data, the static phase error, and the residual loop jitter generation. The jitter accommodation is roughly 0.5 UI in this region. The corner frequency between the declining slope and the flat region is the closed-loop bandwidth of the delay-locked loop, which is roughly 1.0 MHz at 622 Mbps. Rev. 0 | Page 15 of 24 ADN2804 FUNCTIONAL DESCRIPTION FREQUENCY ACQUISITION LOSS-OF-SIGNAL (LOS) DETECTOR The ADN2804 acquires frequency from the data. The lock detector circuit compares the frequency of the VCO and the frequency of the incoming data. When these frequencies differ by more than 1000 ppm, LOL is asserted. This initiates a frequency acquisition cycle. When the VCO frequency is within 250 ppm of the data frequency, LOL is deasserted. The receiver front-end LOS detector circuit detects when the input signal level falls below a user-adjustable threshold. The threshold is set with a single external resistor from Pin 9, THRADJ, to VEE. The LOS comparator trip point vs. the resistor value is shown in Figure 6. If the input level to the ADN2804 drops below the programmed LOS threshold, the output of the LOS detector, LOS (Pin 22), is asserted to Logic 1. The LOS detector’s response time is ~500 ns by design, but is dominated by the RC time constant in ac-coupled applications. The LOS pin defaults to active high. However, setting Bit CTRLC[2] to 1, configures the LOS pin as active low. Once LOL is deasserted, the frequency-locked loop is turned off. The PLL/DLL pulls the VCO frequency in the rest of the way until the VCO frequency equals the data frequency. The frequency loop requires a single external capacitor between CF1 and CF2, Pin 14 and Pin 15. A 0.47 μF ± 20%, X7R ceramic chip capacitor with <10 nA leakage current is recommended. Leakage current of the capacitor can be calculated by dividing the maximum voltage across the 0.47 μF capacitor, ~3 V, by the insulation resistance of the capacitor. The insulation resistance of the 0.47 μF capacitor should be greater than 300 MΩ. There is typically 6 dB of electrical hysteresis designed into the LOS detector to prevent chatter on the LOS pin. If the input level drops below the programmed LOS threshold causing the LOS pin to assert, the LOS pin deasserts after the input level increases to 6 dB (2×) above the LOS threshold (see Figure 19). LOS OUTPUT INPUT LEVEL HYSTERESIS LOS THRESHOLD t 05801-019 The limiting amplifier has differential inputs (PIN/NIN) that are internally terminated with 50 Ω to an on-chip voltage reference (VREF = 2.5 V typically). The inputs are typically ac-coupled externally, although dc coupling is possible as long as the input common-mode voltage remains above 2.5 V (see Figure 27 to Figure 29 in the Applications Information section). Input offset is factory trimmed to achieve better than 3.3 mV typical sensitivity with minimal drift. The limiting amplifier can be driven differentially or in a single-ended fashion. INPUT VOLTAGE (VDIFF) LIMITING AMPLIFIER Figure 19. LOS Detector Hysteresis SLICE ADJUST The quantizer slicing level can be offset by ±100 mV to mitigate the effect of amplified spontaneous emission (ASE) noise or duty cycle distortion by applying a differential voltage input of up to ±0.95 V to the SLICEP and SLICEN inputs. If no adjustment of the slice level is needed, SLICEP and SLICEN should be tied to VEE. The gain of the slice adjustment is ~0.11 V/V. The LOS detector and the SLICE level adjust can be used simultaneously on the ADN2804. This means that any offset added to the input signal by the SLICE adjust pins does not affect the LOS detector’s measurement of the absolute input level. Rev. 0 | Page 16 of 24 ADN2804 LOCK DETECTOR OPERATION Static LOL Mode The lock detector on the ADN2804 has three modes of operation: normal mode, REFCLK mode, and static LOL mode. The ADN2804 implements a static LOL feature that indicates if a loss-of-lock condition has ever occurred. This feature remains asserted, even if the ADN2804 regains lock, until the static LOL bit is manually reset. The I2C register bit, MISC[4], is the static LOL bit. If there is ever an occurrence of a loss-of-lock condition, this bit is internally asserted to logic high. The MISC[4] bit remains high even after the ADN2804 has reacquired lock to a new data rate. This bit can be reset by writing a 1 followed by 0 to I2C Register Bit CTRLB[6]. Once reset, the MISC[4] bit remains deasserted until another loss-of-lock condition occurs. Normal Mode In normal mode, the ADN2804 is a CDR that locks onto a 622 Mbps data rate without the use of a reference clock as an acquisition aid. In this mode, the lock detector monitors the frequency difference between the VCO and the input data frequency and deasserts the loss of lock signal, which appears on Pin 16, LOL, when the VCO is within 250 ppm of the data frequency. This enables the D/PLL, which pulls the VCO frequency in the remaining amount and acquires phase lock. Once locked, if the input frequency error exceeds 1000 ppm (0.1%), the loss-of-lock signal is reasserted and control returns to the frequency loop, which begins a new frequency acquisition. The LOL pin remains asserted until the VCO locks onto a valid input data stream to within 250 ppm frequency error. This hysteresis is shown in Figure 20. Writing a 1 to I2C Register Bit CTRLB[7] causes the LOL pin, Pin 16, to become a static LOL indicator. In this mode, the LOL pin mirrors the contents of the MISC[4] bit and has the functionality described in the previous paragraph. The CTRLB[7] bit defaults to 0. In this mode, the LOL pin operates in the normal operating mode, that is, it is asserted only when the ADN2804 is in acquisition mode and deasserts when the ADN2804 has reacquired lock. LOL SQUELCH MODES –1000 –250 0 250 1000 fVCO ERROR (ppm) 05801-020 1 Figure 20. Transfer Function of LOL LOL Detector Operation Using a Reference Clock In REFCLK mode, a reference clock is used as an acquisition aid to lock the ADN2804 VCO. Lock-to-reference mode is enabled by setting CTRLA[0] to 1. The user also needs to write to the CTRLA[7, 6] and CTRLA[5:2] bits to set the reference frequency range and the divide ratio of the data rate with respect to the reference frequency. For more details, see the Reference Clock (Optional) section. In this mode, the lock detector monitors the difference in frequency between the divided down VCO and the divided down reference clock. The loss-of-lock signal, which appears on Pin 16, LOL, is deasserted when the VCO is within 250 ppm of the desired frequency. This enables the D/PLL, which pulls the VCO frequency in the remaining amount with respect to the input data and acquires phase lock. Once locked, if the input frequency error exceeds 1000 ppm (0.1%), the loss-of-lock signal is reasserted and control returns to the frequency loop, which reacquires with respect to the reference clock. The LOL pin remains asserted until the VCO frequency is within 250 ppm of the desired frequency. This hysteresis is shown in Figure 20. Two modes for the SQUELCH pin are available with the ADN2804: squelch data outputs and clock outputs mode and squelch data outputs or clock outputs mode. Squelch data outputs and clock outputs mode is selected when CTRLC[1] is 0 (default mode). In this mode, when the SQUELCH input, Pin 27, is driven to a TTL high state, both the data outputs (DATAOUTN and DATAOUTP) and the clock outputs (CLKOUTN and CLKOUTP) are set to the zero state to suppress downstream processing. If the squelch function is not required, Pin 27 should be tied to VEE. Squelch data outputs or clock outputs mode is selected when CTRLC[1] is 1. In this mode, when the SQUELCH input is driven to a high state, the DATAOUTN and DATAOUTP pins are squelched. When the SQUELCH input is driven to a low state, the CLKOUTN and CLKOUTP pins are squelched. This is especially useful in repeater applications, where the recovered clock may not be needed. I2C INTERFACE The ADN2804 supports a 2-wire, I2C-compatible serial bus driving multiple peripherals. Two inputs, serial data (SDA) and serial clock (SCK), carry information to and from any device connected to the bus. Each slave device is recognized by a unique address. The ADN2804 has two possible 7-bit slave addresses for both read and write operations. The MSB of the 7-bit slave address is factory programmed to 1. B5 of the slave address is set by Pin 19, SADDR5. Slave Address Bits [4:0] are defaulted to all 0s. The slave address consists of the seven MSBs of an 8-bit word. The LSB of the word either sets a read or write operation (see Figure 7). Logic 1 corresponds to a read operation, while Logic 0 corresponds to a write operation. Rev. 0 | Page 17 of 24 ADN2804 To control the device on the bus, the following protocol must be followed. First, the master initiates a data transfer by establishing a start condition, defined by a high-to-low transition on SDA while SCK remains high. This 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. This is known as an 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 SCK 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. Logic 1 on the LSB of the first byte means that the master reads information from the peripheral. The ADN2804 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 ADN2804 has eight subaddresses to enable the user-accessible internal registers (see Table 6 through Table 10). It, therefore, interprets the first byte as the device address and the second byte as the starting subaddress. Auto-increment mode is supported, allowing 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 SCK high period, the user should issue one start condition, one stop condition, or a single stop condition followed by a single start condition. If an invalid subaddress is issued by the user, the ADN2804 does not issue an acknowledge and returns to the idle condition. If the user exceeds the highest subaddress while reading back in autoincrement mode, then the highest subaddress register contents continue to be output until the master device issues a no acknowledge. This indicates the end of a read. In a no-acknowledge condition, the SDATA line is not pulled low on the ninth pulse. See Figure 8 and Figure 9 for sample write and read data transfers and Figure 10 for a more detailed timing diagram. Additional Features Available via the I2C Interface LOS Configuration The LOS detector output, Pin 22, can be configured to be either active high or active low. If CTRLC[2] is set to Logic 0 (default), the LOS pin is active high when a loss-of-signal condition is detected. Writing a 1 to CTRLC[2] configures the LOS pin to be active low when a loss-of-signal condition is detected. System Reset A frequency acquisition can be initiated by writing a 1 followed by a 0 to the I2C Register Bit CTRLB[5]. This initiates a new frequency acquisition while keeping the ADN2804 in its previously programmed operating mode, as set in Registers CTRL[A], CTRL[B], and CTRL[C]. Rev. 0 | Page 18 of 24 ADN2804 REFERENCE CLOCK (OPTIONAL) A reference clock is not required to perform clock and data recovery with the ADN2804; however, support for an optional reference clock is provided. The reference clock can be driven differentially or in a single-ended fashion. If the reference clock is not being used, REFCLKP should be tied to VCC, and REFCLKN can be left floating or tied to VEE (the inputs are internally terminated to VCC/2). See Figure 21 through Figure 23 for sample configurations. The REFCLK input buffer accepts any differential signal with a peak-to-peak differential amplitude of greater than 100 mV (for example, LVPECL or LVDS) or a standard single-ended, low voltage TTL input, providing maximum system flexibility. Phase noise and duty cycle of the reference clock are not critical, and 100 ppm accuracy is sufficient. ADN2804 There are two mutually exclusive uses, or modes, of the reference clock. The reference clock can be used either to help the ADN2804 lock onto data or to measure the frequency of the incoming data to within 0.01%. The modes are mutually exclusive because in the first use the user knows exactly what the data rate is and wants to force the part to lock onto only that data rate, and in the second use the user does not know what the data rate is and wants to measure it. Lock-to-reference mode is enabled by writing a 1 to I2C Register Bit CTRLA[0]. Fine data rate readback mode is enabled by writing a 1 to I2C Register Bit CTRLA[1]. Writing a 1 to both of these bits at the same time causes an indeterminate state and is not supported. Using the Reference Clock to Lock onto Data In this mode, the ADN2804 locks onto a frequency derived from the reference clock according to REFCLKP Data Rate/2CTRLA[5:2] = REFCLK/2CTRLA[7, 6] 10 BUFFER The user must provide a reference clock that is a function of the data rate. By default, the ADN2804 expects a reference clock of 19.44 MHz. Other options are 38.88 MHz, 77.76 MHz, and 155.52 MHz, which are selected by programming CTRLA[7, 6]. CTRLA[5:2] should be programmed to [0101] for all cases. 11 100kΩ 100kΩ VCC/2 05801-021 REFCLKN Figure 21. Differential REFCLK Configuration VCC CLK OSC REFCLKP Table 11. CTRLA Settings CTRLA[7, 6] 00 01 10 11 ADN2804 OUT BUFFER 100kΩ 100kΩ VCC/2 Figure 22. Single-Ended REFCLK Configuration VCC Ratio 25 25 25 25 622.08 Mbps/19.44 MHz = 25 In this mode, if the ADN2804 loses lock for any reason, it relocks onto the reference clock and continues to output a stable clock. BUFFER 11 100kΩ VCC/2 05801-023 REFCLKN 100kΩ CTRLA[5:2] 0101 0101 0101 0101 For example, if the reference clock frequency is 38.88 MHz and the input data rate is 622.08 Mbps, CTRLA[7, 6] is set to [01] to produce a divided-down reference clock of 19.44 MHz, and CTRLA[5:2] is set to [0101], that is, 5, because ADN2804 10 REFCLKP NC 05801-022 REFCLKN Range (MHz) 19.44 38.88 77.76 155.52 While the ADN2804 is operating in lock-to-reference mode, a 0 to 1 transition should be written into the CTRLA[0] bit to initiate a lock-to-reference clock command. Figure 23. No REFCLK Configuration Rev. 0 | Page 19 of 24 ADN2804 Using the Reference Clock to Measure Data Frequency The user can also provide a reference clock to measure the recovered data frequency. In this case, the user provides a reference clock, and the ADN2804 compares the frequency of the incoming data to the incoming reference clock and returns a ratio of the two frequencies to within 0.01% (100 ppm) accuracy. The accuracy error of the reference clock is added to the accuracy of the ADN2804 data rate measurement. For example, if a 100 ppm accuracy reference clock is used, the total accuracy of the measurement is within 200 ppm. The reference clock can range from 10 MHz to 160 MHz. By default, the ADN2804 expects a reference clock between 10 MHz and 20 MHz. If the reference clock is between 20 MHz and 40 MHz, 40 MHz and 80 MHz, or 80 MHz and 160 MHz, the user must configure the ADN2804 for the correct reference frequency range by setting two bits of the CTRLA register, CTRLA[7, 6]. Using the reference clock to determine the frequency of the incoming data does not affect the manner in which the part locks onto data. In this mode, the reference clock is used only to determine the frequency of the data. Prior to reading back the data rate using the reference clock, the CTRLA[7, 6] bits must be set to the appropriate frequency range with respect to the reference clock being used. A fine data rate readback is then executed as follows: 1. Write a 1 to CTRLA[1]. This enables the fine data rate measurement capability of the ADN2804. This bit is level sensitive and can perform subsequent frequency measurements without being reset. 2. Reset MISC[2] by writing a 1 followed by a 0 to CTRLB[3]. This initiates a new data rate measurement. 3. Read back MISC[2]. If it is 0, the measurement is not complete. If it is 1, the measurement is complete and the data rate can be read back on FREQ[22:0]. The time for a data rate measurement is typically 80 ms. 4. Read back the data rate from FREQ2[6:0], FREQ1[7:0], and FREQ0[7:0]. The data rate can be determined by f DATARATE = (FREQ [22.0] × f REFCLK ) / 2 (14 + SEL _ RATE) where: FREQ[22:0] is the reading from FREQ2[6:0] (MSB byte, FREQ1[7:0], and FREQ0[7:0] (LSB byte). fDATARATE is the data rate (Mbps). fREFCLK is the REFCLK frequency (MHz). SEL_RATE is the setting from CTRLA[7, 6]. For example, if the reference clock frequency is 32 MHz, SEL_RATE = 1, because the reference frequency falls into the 20 MHz to 40 MHz range, setting CTRLA[7, 6] to [01],. Assume for this example that the input data rate is 622.08 Mb/s (OC12). After following Step 1 through Step 4, the value that is read back on FREQ[22:0] = 0x9B851, which is equal to 637 × 103. Plugging this value into the equation yields 637e3 × 32e6/2(14 + 1) = 622.08 Mbps If subsequent frequency measurements are required, CTRLA[1] should remain set to 1. It does not need to be reset. The measurement process is reset by writing a 1 followed by a 0 to CTRLB[3]. This initiates a new data rate measurement. Follow Step 2 through Step 4 to read back the new data rate. Note that a data rate readback is valid only if LOL is low. If LOL is high, the data rate readback is invalid. Table 12. D22 D21 ... D17 FREQ2[6:0] D16 D15 D14 ... D9 FREQ1[7:0] Rev. 0 | Page 20 of 24 D8 D7 D6 ... D1 FREQ0[7:0] D0 ADN2804 APPLICATIONS INFORMATION PCB DESIGN GUIDELINES If connections to the supply and ground are made through vias, the use of multiple vias in parallel helps to reduce series inductance, especially on Pin 24, which supplies power to the high speed CLKOUTP/CLKOUTN and DATAOUTP/ DATAOUTN output buffers. Refer to Figure 24 for the recommended connections. Proper RF PCB design techniques must be used for optimal performance. Power Supply Connections and Ground Planes Use of one low impedance ground plane is recommended. The VEE pins should be soldered directly to the ground plane to reduce series inductance. If the ground plane is an internal plane and connections to the ground plane are made through vias, multiple vias can be used in parallel to reduce the series inductance, especially on Pin 23, which is the ground return for the output buffers. The exposed pad should be connected to the GND plane using plugged vias so that solder does not leak through the vias during reflow. By placing the power supply and GND planes adjacent to each other and using close spacing between the planes, excellent high frequency decoupling can be realized. The capacitance is given by C PLANE = 0.88ε r A/d (pF) Use of a 22 μF electrolytic capacitor between VCC and VEE is recommended at the location where the 3.3 V supply enters the PCB. When using 0.1 μF and 1 nF ceramic chip capacitors, they should be placed between ADN2804 supply pins VCC and VEE, as close as possible to the ADN2804 VCC pins. where: εr is the dielectric constant of the PCB material. A is the area of the overlap of power and GND planes (cm2). d is the separation between planes (mm). For FR-4, εr = 4.4 and d = 0.25 mm; therefore, CPLANE ~ 15.5A (pF) 50Ω TRANSMISSION LINES VCC DATAOUTP + 22µF 0.1µF DATAOUTN 1nF CLKOUTP TEST2 VCC VEE DATAOUTP DATAOUTN SQUELCH CLKOUTP CLKOUTN CLKOUTN TIA 50Ω 1.6µF 50Ω 1.6µF 32 31 30 29 28 27 26 25 1nF VCC VEE LOS SDA SCK SADDR5 VCC VEE 1nF 0.1µF I2C CONTROLLER I2C CONTROLLER µC VCC 0.1µF µC RTH VCC 0.1µF 1nF 0.47µF ±20% >300MΩ INSULATION RESISTANCE Figure 24. Typical ADN2804 Applications Circuit Rev. 0 | Page 21 of 24 05801-031 0.1µF 24 EXPOSED PAD 23 TIED OFF TO 22 VEE PLANE 21 20 WITH VIAS 19 18 17 9 10 11 12 13 14 15 16 1nF 1 2 3 4 5 6 7 8 THRADJ REFCLKP REFCLKN NC VCC VEE CF2 CF1 LOL 0.1µF TEST1 VCC VREF NIN PIN SLICEP SLICEN VEE VCC ADN2804 Transmission Lines Choosing AC Coupling Capacitors Minimizing reflections in the ADN2804 requires use of 50 Ω transmission lines for all pins with high frequency input and output signals, including PIN, NIN, CLKOUTP, CLKOUTN, DATAOUTP, and DATAOUTN (also REFCLKP and REFCLKN, if a high frequency reference clock is used, such as 155 MHz). It is also necessary for the PIN/NIN input traces to be matched in length and for the CLKOUTP/CLKOUTN and DATAOUTP/DATAOUTN output traces to be matched in length to avoid skew between the differential traces. AC coupling capacitors at the input (PIN, NIN) and output (DATAOUTP, DATAOUTN) of the ADN2804 can be optimized for the application. When choosing the capacitors, the time constant formed with the two 50 Ω resistors in the signal path must be considered. When a large number of consecutive identical digits (CIDs) are applied, the capacitor voltage can droop due to baseline wander (see Figure 26), causing patterndependent jitter (PDJ). The high speed inputs, PIN and NIN, are internally terminated with 50 Ω to an internal reference voltage (see Figure 25). A 0.1 μF is recommended between VREF, Pin 3, and GND to provide an ac ground for the inputs. The user must determine how much droop is tolerable and choose an ac coupling capacitor based on that amount of droop. The amount of PDJ can then be approximated based on the capacitor selection. The actual capacitor value selection can require some trade-offs between droop and PDJ. As with any high speed, mixed-signal design, take care to keep all high speed digital traces away from sensitive analog nodes. For example, assuming that 2% droop can be tolerated, the maximum differential droop is 4%. Normalizing to V p-p: Droop = ΔV = 0.04 V = 0.5 V p-p (1 − e−t/τ); therefore, τ = 12t VCC ADN2804 CIN PIN 50Ω CIN NIN TIA 0.1µF VREF 3Ω where: τ is the RC time constant (C is the ac coupling capacitor, R = 100 Ω seen by C). t is the total discharge time, which is equal to nT, where n is the number of CIDs, and T is the bit period. 2.5V The capacitor value can then be calculated by combining the equations for τ and t: 05801-026 50Ω Figure 25. ADN2804 AC-Coupled Input Configuration C = 12 nT/R Soldering Guidelines for Lead Frame Chip Scale Package The lands on the 32-lead LFCSP are rectangular. The printed circuit board (PCB) pad for these should be 0.1 mm longer than the package land length and 0.05 mm wider than the package land width. The land should be centered on the pad. This ensures that the solder joint size is maximized. The bottom of the chip scale package has a central exposed pad. The pad on the PCB should be at least as large as this exposed pad. The user must connect the exposed pad to VEE using plugged vias so that solder does not leak through the vias during reflow. This ensures a solid connection from the exposed pad to VEE. Once the capacitor value is selected, the PDJ can be approximated as PDJpspp = 0.5 tr(1 − e(−nT/RC))/0.6 where: PDJpspp is the amount of pattern-dependent jitter allowed (<0.01 UI p-p typical). tr is the rise time, which is equal to 0.22/BW, where BW ~ 0.7 (bit rate). Note that this expression for tr is accurate only for the inputs. The output rise time for the ADN2804 is ~100 ps regardless of the data rate. Rev. 0 | Page 22 of 24 ADN2804 VCC V1 CIN V2 ADN2804 PIN TIA V1b CIN V2b 50Ω VREF 1 2 DATAOUTP CDR LIMAMP DATAOUTN COUT – NIN V1 COUT + 50Ω 3 4 V1b V2 VREF V2b VTH VDIFF VDIFF = V2–V2b VTH = ADN2804 QUANTIZER THRESHOLD 05801-027 NOTES: 1. DURING DATA PATTERNS WITH HIGH TRANSITION DENSITY, DIFFERENTIAL DC VOLTAGE AT V1 AND V2 IS ZERO. 2. WHEN THE OUTPUT OF THE TIA GOES TO CID, V1 AND V1b ARE DRIVEN TO DIFFERENT DC LEVELS. V2 AND V2b DISCHARGE TO THE VREF LEVEL, WHICH EFFECTIVELY INTRODUCES A DIFFERENTIAL DC OFFSET ACROSS THE AC COUPLING CAPACITORS. 3. WHEN THE BURST OF DATA STARTS AGAIN, THE DIFFERENTIAL DC OFFSET ACROSS THE AC COUPLING CAPACITORS IS APPLIED TO THE INPUT LEVELS CAUSING A DC SHIFT IN THE DIFFERENTIAL INPUT. THIS SHIFT IS LARGE ENOUGH SUCH THAT ONE OF THE STATES, EITHER HIGH OR LOW DEPENDING ON THE LEVELS OF V1AND V1b WHEN THE TIA WENT TO CID, IS CANCELED OUT. THE QUANTIZER DOES NOT RECOGNIZE THIS AS A VALID STATE. 4. THE DC OFFSET SLOWLY DISCHARGES UNTIL THE DIFFERENTIAL INPUT VOLTAGE EXCEEDS THE SENSITIVITY OF THE ADN2804. THE QUANTIZER CAN RECOGNIZE BOTH HIGH AND LOW STATES AT THIS POINT. Figure 26. Example of Baseline Wander PIN VSE = 5mV MIN NIN PIN PIN VCM = 2.3V MIN (DC-COUPLED) Figure 28. Minimum Allowed DC-Coupled Input Levels VCC 50Ω V p-p = PIN – NIN = 2 × VSE = 10mV AT SENSITIVITY 05801-029 The inputs to the ADN2804 can also be dc-coupled. This may be necessary in burst mode applications, where there are long periods of CIDs, and baseline wander cannot be tolerated. If the inputs to the ADN2804 are dc-coupled, care must be taken not to violate the input range and common-mode level requirements of the ADN2804 (see Figure 27 through Figure 29). If dc coupling is required and the output levels of the TIA do not adhere to the levels shown in Figure 28, level shifting must be performed and/or an attenuator must be placed between the TIA outputs and the ADN2804 inputs. INPUT (V) DC-COUPLED APPLICATION V p-p = PIN – NIN = 2 × VSE = 2.0V MAX ADN2804 INPUT (V) VSE = 1.0V MAX TIA NIN 0.1µF VREF 50Ω 3kΩ NIN 2.5V 05801-030 50Ω VCM = 2.3V (DC-COUPLED) 05801-028 50Ω Figure 27. DC-Coupled Application Figure 29. Maximum Allowed DC-Coupled Input Levels Rev. 0 | Page 23 of 24 ADN2804 OUTLINE DIMENSIONS 0.60 MAX 5.00 BSC SQ 0.60 MAX 25 24 PIN 1 INDICATOR TOP VIEW 0.50 BSC 4.75 BSC SQ 0.50 0.40 0.30 12° MAX 1.00 0.85 0.80 PIN 1 INDICATOR 32 1 EXPOSED PAD (BOTTOM VIEW) 17 16 3.45 3.30 SQ 3.15 9 8 0.25 MIN 3.50 REF 0.80 MAX 0.65 TYP 0.05 MAX 0.02 NOM SEATING PLANE 0.30 0.23 0.18 0.20 REF COPLANARITY 0.08 COMPLIANT TO JEDEC STANDARDS MO-220-VHHD-2 Figure 30. 32-Lead Frame Chip Scale Package [LFCSP_VQ] 5 mm × 5 mm Body, Very Thin Quad (CP-32-3) Dimensions shown in millimeters ORDERING GUIDE Model ADN2804ACPZ 1 ADN2804ACPZ-500RL71 ADN2804ACPZ-RL71 EVAL-ADN2804EB 1 Temperature Range −40°C to +85°C −40°C to +85°C −40°C to +85°C Package Description 32-Lead LFCSP_VQ 32-Lead LFCSP_VQ, Tape-Reel, 500 pieces 32-Lead LFCSP_VQ, Tape-Reel, 1500 pieces Evaluation Board Package Option CP-32-3 CP-32-3 CP-32-3 Z = Pb-free part. Purchase of licensed I2C components of Analog Devices or one of its sublicensed Associated Companies conveys a license for the purchaser under the Philips I2C Patent Rights to use these components in an I2C system, provided that the system conforms to the I2C Standard Specification as defined by Philips. ©2006 Analog Devices, Inc. All rights reserved. Trademarks and registered trademarks are the property of their respective owners. D05801–0–2/06(0) Rev. 0 | Page 24 of 24