3.3 V Dual-Loop, 50 Mbps to 3.3 Gbps Laser Diode Driver ADN2872 FEATURES GENERAL DESCRIPTION SFP/SFF and SFF-8472 MSA compliant SFP reference design available Any rate from 50 Mbps to 3.3 Gbps operation Dual-loop control of average power and extinction ratio Typical rise/fall time: 60 ps Bias current range: 2 mA to 100 mA Modulation current range: 5 mA to 90 mA Laser FAIL alarm and automatic laser shutdown (ALS) Bias and modulation current monitoring 3.3 V operation 4 mm x 4 mm LFCSP Voltage setpoint control Resistor setpoint control Like the ADN2870, the ADN28721 laser diode driver is designed for advanced SFP and SFF modules, using SFF-8472 digital diagnostics. The device features dual-loop control of the average power and extinction ratio, which automatically compensates for variations in laser characteristics over temperature and aging. The laser needs only be calibrated at 25C, eliminating the need for expensive and time consuming temperature calibration. The ADN2872 supports single-rate operation from 50 Mbps to 3.3 Gbps, or multirate operation from 155 Mbps to 3.3 Gbps. With a new alarm scheme, this device avoids the shutdown issue caused by the system transient generated from various lasers. The average power and extinction ratio can be set with a voltage provided by a microcontroller DAC or by a trimmable resistor. The part provides both bias and modulation current monitoring, as well as fail alarms and automatic laser shutdown. The ADN2872, a SFF-/SFP-compliant laser diode driver, can work with the Analog Devices, Inc., ADuC7019/ADuC702x MicroConverter(R) family and the ADN289x limiting amplifier family, to form a complete SFP/SFF transceiver solution. The ADN2872 is available in a space-saving 4 mm x 4 mm LFCSP specified over the -40C to +85C temperature range. APPLICATIONS Multirate OC3 to OC48-FEC SFP/SFF modules 1x/2x/4x Fibre Channel SFP/SFF modules LX-4 modules DWDM/CWDM SFP modules 1GE SFP/SFF transceiver modules Figure 1 shows an application diagram with a microcontroller interface. 1 Protected by U.S. Patent 6,414,974. APPLICATIONS DIAGRAM VCC VCC VCC VCC Tx_FAULT L Tx_FAIL FAIL VCC ALS IMODN LASER R IMODP MPD DATAP PAVSET ANALOG DEVICES MICROCONTROLLER DAC VCC IBIAS CONTROL ADC DAC DATAN 100 IMOD PAVREF RZ IBIAS RPAV CCBIAS 1k GND ERREF ADN2872 1k ERSET GND IBMON IMMON 1k 470 PAVCAP ERCAP VCC GND GND GND GND 08013-001 GND 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 (c)2009 Analog Devices, Inc. All rights reserved. ADN2872 TABLE OF CONTENTS Features .............................................................................................. 1 Voltage Setpoint Calibration ..................................................... 11 Applications ....................................................................................... 1 Resistor Setpoint Calibration .................................................... 13 General Description ......................................................................... 1 IMPD Monitoring .......................................................................... 13 Applications Diagram ...................................................................... 1 Loop Bandwidth Selection ........................................................ 14 Revision History ............................................................................... 2 Power Consumption .................................................................. 14 Specifications..................................................................................... 3 Automatic Laser Shutdown (Tx_Disable)............................... 14 SFP Timing Specifications........................................................... 4 Bias and Modulation Monitor Currents.................................. 14 Absolute Maximum Ratings............................................................ 5 IBIAS Pin ..................................................................................... 14 ESD Caution .................................................................................. 5 Data Inputs .................................................................................. 15 Pin Configuration and Function Descriptions ............................. 6 Laser Diode Interfacing ............................................................. 15 Typical Performance Characteristics ............................................. 7 Alarms.......................................................................................... 16 Optical Waveforms ........................................................................... 9 Outline Dimensions ....................................................................... 17 Theory of Operation ...................................................................... 10 Ordering Guide .......................................................................... 17 Dual-Loop Control..................................................................... 10 Control ......................................................................................... 11 REVISION HISTORY 3/09--Revision 0: Initial Version Rev. 0 | Page 2 of 20 ADN2872 SPECIFICATIONS VCC = 3.0 V to 3.6 V. All specifications TMIN to TMAX, unless otherwise noted.1 Typical values as specified at 25C. Table 1. Parameter LASER BIAS CURRENT (IBIAS) Output Current, IBIAS Compliance Voltage IBIAS when ALS High CCBIAS Compliance Voltage MODULATION CURRENT (IMODP, IMODN)2 Output Current, IMOD Compliance Voltage IMOD when ALS High Rise Time2, 3 Fall Time2, 3 Random Jitter2, 3 Deterministic Jitter2, 3 Pulse Width Distortion2, 3 AVERAGE POWER SET (PAVSET) Pin Capacitance Voltage Photodiode Monitor Current (Average Current) EXTINCTION RATIO SET INPUT (ERSET) Resistance Range Voltage AVERAGE POWER REFERENCE VOLTAGE INPUT (PAVREF) Voltage Range Photodiode Monitor Current (Average Current) EXTINCTION RATIO REFERENCE VOLTAGE INPUT (ERREF) Voltage Range DATA INPUTS (DATAP, DATAN)4 Input Voltage Swing (Differential) Input Impedance (Single-Ended) LOGIC INPUTS (ALS) VIH VIL ALARM OUTPUT (FAIL)5 VOFF Min Max Unit 100 VCC 0.2 mA V mA V 90 VCC 0.05 104 96 1.1 35 30 mA V mA ps ps ps rms ps ps 20 mA < IMOD < 90 mA 20 mA < IMOD < 90 mA 80 1.35 1200 pF V A Resistor setpoint mode 25 1.35 k V Resistor setpoint mode Resistor setpoint mode 0.12 120 1 1000 V A Voltage setpoint mode (RPAV fixed at 1 k) Voltage setpoint mode (RPAV fixed at 1 k) 0.1 1 V Voltage setpoint mode (RERSET fixed at 1 k) 2.4 V p-p AC-coupled 2 1.2 1.2 5 1.5 60 60 0.8 1.1 50 1.2 1.1 1.2 1.2 0.4 50 2 0.8 VON IBMON, IMMON DIVISION RATIO IBIAS/IBMON3 IBIAS/IBMON3 IBIAS/IBMON Stability3, 6 IMOD/IMMON IBMON Compliance Voltage Typ 85 92 V <1.3 V 115 108 5 50 0 V V >1.8 100 100 1.3 Rev. 0 | Page 3 of 20 Conditions/Comments A/A A/A % A/A V Voltage required at FAIL for IBIAS and IMOD to turn off when FAIL asserted Voltage required at FAIL for IBIAS and IMOD to stay on when FAIL asserted 11 mA < IBIAS < 50 mA 50 mA < IBIAS < 100 mA 10 mA < IBIAS < 100 mA ADN2872 Parameter SUPPLY ICC7 VCC (with respect to GND)8 Min Typ Max Unit Conditions/Comments 30 3.3 3.6 mA V When IBIAS = IMOD = 0 3.0 1 Temperature range: -40C to +85C. Measured into a 15 load (22 resistor in parallel with digital scope 50 input) using a 11110000 pattern at 2.5 Gbps, shown in Figure 2. 3 Guaranteed by design and characterization. Not production tested. 4 When the voltage on DATAP is greater than the voltage on DATAN, the modulation current flows into the IMODP pin. 5 Guaranteed by design. Not production tested. 6 IBIAS/IBMON ratio stability is defined in SFF-8472 Revision 9 over temperature and supply variation. 7 See the Power Consumption section for ICC minimum for power calculation. 8 All VCC pins should be shorted together. 2 ADN2872 VCC R 22 L TO HIGH SPEED DIGITAL OSCILLOSCOPE 50 INPUT C IMODP BIAS TEE 80kHz 27GHz 08013-034 VCC Figure 2. High Speed Electrical Test Output Circuit SFP TIMING SPECIFICATIONS Table 2. Parameter ALS Assert Time Symbol t_off ALS Negate Time1 Time to Initialize, Including Reset of FAIL1 FAIL Assert Time ALS to Reset Time Typ 1 Max 5 Unit s t_on 0.83 0.95 ms t_init t_fault t_reset 25 275 100 5 ms s s Conditions/Comments Time for the rising edge of ALS (Tx_DISABLE) to when the bias current falls below 10% of nominal Time for the falling edge of ALS to when the modulation current rises above 90% of nominal From power-on or negation of FAIL using ALS Time to fault to FAIL on Time Tx_DISABLE must be held high to reset Tx_FAULT Guaranteed by design and characterization. Not production tested. VSE DATAP SFP MODULE DATAN 1H VCC_Tx 3.3V 0.1F 10F DATAP - DATAN 0V Vp-p, DIFF = 2 x VSE SFP HOST BOARD Figure 3. Signal Level Definition Rev. 0 | Page 4 of 20 Figure 4. Recommended SFP Supply 08013-003 0.1F 08013-002 1 Min ADN2872 ABSOLUTE MAXIMUM RATINGS TA = 25C, unless otherwise noted. Stresses above those listed under Absolute Maximum Ratings may cause permanent damage to the device. This is a stress rating only; functional operation of the device at these or any other conditions above those indicated in the operational section of this specification is not implied. Exposure to absolute maximum rating conditions for extended periods may affect device reliability. Table 3. Parameter VCC to GND IMODN, IMODP PAVCAP, ERCAP, PAVSET, PAVREF, ERREF, IBIAS, IBMON, IMMON, ALS, CCBIAS, RPAV, ERSET, FAIL DATAP, DATAN (Single-Ended Differential) Junction Temperature (TJ max) Operating Temperature Range, Industrial Storage Temperature Range Power Dissipation (W)1 JA Thermal Impedance2 JC Thermal Impedance2 JB Thermal Impedance2 Rating 4. 2 V -0.3 V to +4.8 V -0.3 V to +3.9 V 1.5 V 125C -40C to +85C -65C to +150C (TJ max - TA)/JA 48.6C/W 5.0C/W 28.4C/W ESD CAUTION 1 Power consumption equations are provided in the Power Consumption section. 2 JA, JB, and JC are estimated when the part's exposed pad is soldered on a 4-layer JEDEC board at zero airflow. Rev. 0 | Page 5 of 20 ADN2872 GND VCC IMODP IMODN GND 24 IBIAS PIN CONFIGURATION AND FUNCTION DESCRIPTIONS 19 1 18 CCBIAS FAIL PAVSET IBMON GND ADN2872 VCC VCC TOP VIEW (Not to Scale) ERREF IMMON PAVREF ERSET RPAV 12 NOTES 1. THE LFCSP PACKAGE HAS AN EXPOSED PADDLE THAT MUST BE CONNECTED TO GROUND. 08013-004 ALS DATAN DATAP GND 7 ERCAP 13 PAVCAP 6 Figure 5. Pin Configuration Table 4. 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 (EPAD) Mnemonic CCBIAS PAVSET GND VCC PAVREF RPAV ERCAP PAVCAP GND DATAP DATAN ALS ERSET IMMON ERREF VCC IBMON FAIL GND VCC IMODP IMODN GND IBIAS Exposed Paddle (EPAD) Description Control Output Current. Average Optical Power Set Pin. Supply Ground. Supply Voltage. Reference Voltage Input for Average Optical Power Control. Average Power Resistor when Using PAVREF. Extinction Ratio Loop Capacitor. Average Power Loop Capacitor. Supply Ground. Data, Positive Differential Input. Data, Negative Differential Input. Automatic Laser Shutdown. Extinction Ratio Set Pin. Modulation Current Monitor Current Source. Reference Voltage Input for Extinction Ratio Control. Supply Voltage. Bias Current Monitor Current Source. FAIL Alarm Output. Supply Ground. Supply Voltage. Modulation Current Positive Output (Current Sink), Connect to Laser Diode. Modulation Current Negative Output (Current Sink). Supply Ground. Laser Diode Bias (Current Sink to Ground). The LFCSP package has an exposed paddle that must be connected to ground. Rev. 0 | Page 6 of 20 ADN2872 TYPICAL PERFORMANCE CHARACTERISTICS 1.2 90 1.0 0.8 JITTER (rms) RISE TIME (ps) 60 0.6 0.4 30 0 0 20 40 60 80 08013-037 08013-022 0.2 0 100 0 20 MODULATION CURRENT (mA) 40 60 80 100 MODULATION CURRENT (mA) Figure 6. Rise Time vs. Modulation Current, IBIAS = 20 mA Figure 9. Random Jitter vs. Modulation Current, IBIAS = 20 mA 250 80 TOTAL SUPPLY CURRENT (mA) 220 40 20 0 20 40 60 80 IBIAS = 40mA 160 130 IBIAS = 20mA 100 70 08013-025 0 IBIAS = 80mA 190 08013-038 FALL TIME (ps) 60 40 0 100 20 MODULATION CURRENT (mA) 45 60 40 55 SUPPLY CURRENT (mA) 35 30 25 20 15 10 80 100 45 40 35 30 25 20 -50 100 Figure 8. Deterministic Jitter vs. Modulation Current, IBIAS = 20 mA 50 08013-027 5 40 60 80 MODULATION CURRENT (mA) 60 Figure 10. Total Supply Current vs. Modulation Current, Total Supply Current = ICC + IBIAS + IMOD 08013-042 DETERMINISTIC JITTER (ps) Figure 7. Fall Time vs. Modulation Current, IBIAS = 20 mA 0 20 40 MODULATION CURRENT (mA) -30 -10 10 30 50 70 90 110 TEMPERATURE (C) Figure 11. Supply Current (ICC) vs. Temperature with ALS Asserted, IBIAS = 20 mA Rev. 0 | Page 7 of 20 ADN2872 60 120 58 115 56 110 IMOD/IMMON GAIN 100 95 52 50 48 46 90 08013-028 80 -50 -30 -10 10 30 50 70 90 08013-031 44 85 42 40 -50 110 -30 -10 10 30 50 70 90 110 TEMPERATURE (C) TEMPERATURE (C) Figure 15. IMOD/IMMON Gain vs. Temperature, IMOD = 30 mA Figure 12. IBIAS/IBMON Gain vs. Temperature, IBIAS = 20 mA OC48 PRBS31 DATA TRANSMISSION t_OFF LESS THAN 1s FAIL ASSERTED FAULT FORCED ON PAVSET 08013-029 08013-045 ALS Figure 16. FAIL Assert Time,1 s/DIV Figure 13. ALS Assert Time, 5 s/DIV OC48 PRBS31 DATA TRANSMISSION TRANSMISSION ON t_ON ALS 08013-046 POWER SUPPLY TURN ON 08013-032 IBIAS/IBMON GAIN 54 105 Figure 17. Time to Initialize, Including Reset, 40 ms/DIV Figure 14. ALS Negate Time, 200 s/DIV Rev. 0 | Page 8 of 20 ADN2872 OPTICAL WAVEFORMS VCC = 3.3 V and TA = 25C, unless otherwise noted. Note that there was no change to PAVCAP and ERCAP values when different data rates were tested. Figure 18, Figure 19, and Figure 20 show multirate performance using the low cost Fabry Perot TOSA NEC NX7315UA; Figure 21 and Figure 22 show dual-loop performance over temperature using the DFB TOSA Sumitomo SLT2486. 08013-047 (ACQ LIMIT TEST) WAVEFORMS 1001 08013-016 (ACQ LIMIT TEST) WAVEFORMS 1000 Figure 18. Optical Eye 2.488 Gbps, 65 ps/DIV, PRBS 231 - 1, PAV = -4.5 dBm, ER = 9 dB, Mask Margin 25% Figure 21. Optical Eye 2.488 Gbps, 65 ps/DIV, PRBS 231 - 1, PAV = 0 dBm, ER = 9 dB, Mask Margin 22%, TA = 25C 08013-048 (ACQ LIMIT TEST) WAVEFORMS 1001 08013-017 (ACQ LIMIT TEST) WAVEFORMS 1000 Figure 19. Optical Eye 622 Mbps, 264 ps/DIV, PRBS 231 - 1, PAV = -4.5 dBm, ER = 9 dB, Mask Margin 50% Figure 22. Optical Eye 2.488 Gbps, 65 ps/DIV, PRBS 231 - 1, PAV = -0.2 dBm, ER = 8.96 dB, Mask Margin 21%, TA = 85C 08013-020 (ACQ LIMIT TEST) WAVEFORMS 1000 Figure 20. Optical Eye 155 Mbps,1.078 ns/DIV, PRBS 231 - 1, PAV = -4.5 dBm, ER = 9 dB, Mask Margin 50% Rev. 0 | Page 9 of 20 ADN2872 THEORY OF OPERATION P1 P0 P1 + P0 PAV = 2 Gm 2 IEX BIAS SHA 1 BIAS CURRENT 1.2V VBGAP IPA VCC HIGH SPEED SWITCH ERSET PAVSET 2 MOD SHA P PAV 2 I LI = P0 ITH CURRENT P MOD CURRENT 100 2 Figure 24. Dual-Loop Control of Average Power and Extinction Ratio I 08013-005 OPTICAL POWER ER = P1 OPTICAL COUPLING MPD INPUT 08013-039 Laser diodes have a current-in to light-out transfer function, as shown in Figure 23. Two key characteristics of this transfer function are the threshold current, ITH, and slope in the linear region beyond the threshold current, referred to as slope efficiency, LI. A dual loop is made up of an average power control loop (APCL) and the extinction ratio control loop (ERCL), which are separated into two time states. During Time 1, the APC loop is operating, and during Time 2, the ER loop is operating. Figure 23. Laser Transfer Function DUAL-LOOP CONTROL Typically, laser threshold current and slope efficiency are both functions of temperature. For FP and DFB type lasers, the threshold current increases and the slope efficiency decreases with increasing temperature. In addition, these parameters vary as the laser ages. To maintain a constant optical average power and a constant optical extinction ratio over temperature and laser lifetime, it is necessary to vary the applied electrical bias current and modulation current to compensate for the laser changing LI characteristics. Single-loop compensation schemes use the average monitor photodiode (MPD) current to measure and maintain the average optical output power over temperature and laser aging. The ADN2872 is a dual-loop device, implementing both this primary average power control loop and a secondary control loop, which maintains a constant optical extinction ratio. The dual-loop control of the average power and extinction ratio implemented in the ADN2872 can be used successfully with both lasers that maintain good linearity of LI transfer characteristics over temperature, and with those that exhibit increasing nonlinearity of the LI characteristics over temperature. Dual Loop The ADN2872 uses a proprietary patented method to control both average power and extinction ratio. The ADN2872 is constantly sending a test signal on the modulation current signal and reading the resulting change in the MPD current as a means of detecting the slope of the laser in real time. This information is used in a servo to control the ER of the laser, which is done in a time-multiplexed manner at a low frequency, typically 80 Hz. Figure 24 shows the dual-loop control implementation on the ADN2872. Average Power Control Loop The APCL compensates for changes in the laser diode (LD), ITH and LI, by varying IBIAS. APC control is performed by measuring the MPD current, IMPD. This current is bandwidth limited by the MPD. This is not a problem because the APCL must be low frequency and the APCL must respond to the average current from the MPD. The APCL compares IMPD x RPAVSET to the BGAP voltage, VBGAP. If IMPD falls, the bias current is increased until IMPD x RPAVSET equals VBGAP. Conversely, if the IMPD increases, IBIAS is decreased. Modulation Control Loop The ERCL measures the slope efficiency, LI, of the laser diode by monitoring the IMPD changes. During the ERCL, IMPD is temporarily increased by IMOD. The ratio between IMPD and IMOD is a fixed ratio of 50:1, but during startup, this ratio is increased to decrease settling time. During ERCL, switching in IMOD causes a temporary increase in average optical power, PAV. However, the APC loop is disabled during ERCL, and the increase is kept small enough so as not to disturb the optical eye. When IMOD is switched into the laser circuit, an equal current, IEX, is switched into the PAVSET resistor. The user sets the value of IEX; this is the ERSET setpoint. If IMPD is too small, the control loop knows that LI has decreased, and increases IMPD and, therefore, IMOD accordingly until IMPD is equal to IEX. The previous control cycle status of the IBIAS and IMOD settings are stored on the hold capacitors, PAVCAP and ERCAP. The ERCL is constantly measuring the actual LI curve; it compensates for the effects of temperature and for changes in the LI curve due to laser aging. Therefore, the laser can be calibrated once at 25C so that it can then automatically control the laser over temperature. This eliminates the expensive and time consuming temperature calibration of a laser. Rev. 0 | Page 10 of 20 ADN2872 Operation with Lasers with Temperature-Dependent Nonlinearity of Laser LI Curve The ADN2872 ERCL extracts information from the monitor photodiode signal relating to the slope of the LI characteristics at the Optical 1 level (P1). For lasers with good linearity over temperature, the slope measured by the ADN2872 at the Optical 1 level is representative of the slope anywhere on the LI curve. This slope information is used to set the required modulation current to achieve the required optical extinction ratio. 4.0 The ER correction scheme, while using the average nonlinearity for the laser population, supplies a corrective measurement based on the actual performance of each laser as measured during operation. The ER correction scheme corrects for errors due to laser nonlinearity while the dual loop continues to adjust for changes in the Laser LI. For more details on maintaining average optical power and extinction ratio over temperature when working with lasers displaying a temperature-dependent nonlinearity of LI curve, contact sales at Analog Devices. RELATIVELY LINEAR LI CURVE AT 25C 3.5 OPTICAL POWER (mW) the laser and apply the feedback. This scheme is particularly suitable for circuits that already use a microcontroller for control and digital diagnostic monitoring. 3.0 2.5 CONTROL 2.0 The ADN2872 has two methods for setting the average power (PAV) and extinction ratio (ER). The average power and extinction ratio can be voltage set using the voltage DAC outputs of a microcontroller to provide controlled reference voltages to PAVREF and ERREF. Alternatively, the average power and extinction ratio can be resistor set using potentiometers at the PAVSET and ERSET pins, respectively. 1.5 1.0 NONLINEAR LI CURVE AT 80C 08013-008 0.5 0 0 20 40 60 80 100 CURRENT (mA) VOLTAGE SETPOINT CALIBRATION Figure 25. Measurement of a Laser LI Curve Showing Laser Nonlinearity at High Temperatures Some types of lasers have LI curves that become progressively more nonlinear with increasing temperature (see Figure 25). At temperatures where the LI curve shows significant nonlinearity, the LI curve slope measured by the ADN2872 at the Optical 1 level is no longer representative of the overall LI curve. It is evident that applying a modulation current based on this slope information cannot maintain a constant extinction ratio over temperature. However, the ADN2872 can be configured to maintain near constant optical bias and an extinction ratio with a laser exhibiting a monotonic temperature-dependent nonlinearity. To implement this correction, it is necessary to characterize a small sample of lasers for their typical nonlinearity by measuring them at two temperature points, typically 25C and 85C. The measured nonlinearity is used to determine the amount of feedback to apply. Typically, the user must characterize five to 10 lasers of a particular model to obtain a good number. The product can then be calibrated at 25C only, avoiding the expense of temperature calibration. Typically, the microcontroller is used to measure The ADN2872 allows an interface to a microcontroller for both control and monitoring (see Figure 26). The average power at the PAVSET pin and extinction ratio at the ERSET pin can be set using the DAC of the microcontroller to provide controlled reference voltages to PAVREF and ERREF. Note that during power-up, there is an internal sequence that allows 25 ms before enabling the alarms; therefore, the user must ensure that the voltage for PAVREF and ERREF are active within 20 ms. PAVREF = PAV x RSP x RPAV ERREF RERSET I MPD _ CW PCW (V) ER 1 PAV ER 1 (V) where: PAV (mW) is the average power required. ER is the desired extinction ratio (ER = P1/P0). RSP (A/W) is the monitor photodiode responsivity. IMPD_CW (mA) is the MPD current at that specified PCW. PCW (mW) is the dc optical power specified on the laser data sheet. In voltage setpoint, RPAV and RERSET must be 1 k resistors with a 1% tolerance and a temperature coefficient of 50 ppm/C. Rev. 0 | Page 11 of 20 ADN2872 VCC VCC VCC Tx_FAULT VCC L Tx_FAIL VCC FAIL ALS IMODN LASER R MPD IMODP DATAP PAVSET ANALOG DEVICES MICROCONTROLLER DATAN 100 IMOD VCC PAVREF DAC RZ IBIAS CONTROL ADC IBIAS RPAV 1k CCBIAS GND ERREF DAC ADN2872 1k ERSET GND IBMON IMMON PAVCAP ERCAP VCC GND GND 470 GND 08013-009 1k GND GND Figure 26. ADN2872 Using MicroConverter Calibration and Monitoring VCC VCC VCC VCC L FAIL VCC ALS IMODN LASER R VCC IMODP PAVREF DATAP MPD RPAV PAVSET DATAN 100 IMOD VCC IBIAS CONTROL RZ IBIAS GND CCBIAS ERSET GND ADN2872 VCC ERREF 1k GND IMMON 470 PAVCAP GND ERCAP GND GND Figure 27. ADN2872 Using Resistor Setpoint Calibration of Average Power and Energy Ratio Rev. 0 | Page 12 of 20 08013-010 IBMON VCC GND ADN2872 RESISTOR SETPOINT CALIBRATION Method 2: Measuring IMPD Across a Sense Resistor In resistor setpoint calibration, the PAVREF, ERREF, and RPAV pins must all be tied to VCC. Average power and extinction ratio can be set using the PAVSET and ERSET pins, respectively. A resistor is placed between the pin and GND to set the current flowing in each pin, as shown in Figure 27. The ADN2872 ensures that both PAVSET and ERSET are kept 1.2 V above GND. The PAVSET and ERSET resistors are given by: The second method has the advantage of providing a valid IMPD reading at all times but has the disadvantage of requiring a differential measurement across a sense resistor directly in series with the IMPD. As shown in Figure 29, a small resistor, Rx, is placed in series with the IMPD. If the laser used in the design has a pinout where the monitor photodiode cathode and the lasers anode are not connected, a sense resistor can be placed in series with the photodiode cathode and VCC, as shown in Figure 30. When choosing the value of the resistor, the user must take into account the expected IMPD value in normal operation. The resistor must be large enough to make a significant signal for the buffered ADC to read, but small enough not to cause a significant voltage reduction across the IMPD. The voltage across the sense resistor should not exceed 250 mV when the laser is in normal operation. It is recommended that a 10 pF capacitor be placed in parallel with the sense resistor. () 1.23 V I MPD _ CW ER 1 P P CW ER 1 AV () where: PAV (mW) is the average power required. RSP (A/W) is the monitor photodiode responsivity. PCW (mW) is the dc optical power specified on the laser data sheet. IMPD_CW (mA) is the MPD current at that specified PCW. ER is the desired extinction ratio (ER = P1/P0). VCC PHOTODIODE IMPD MONITORING MICROCONVERTER ADC DIFFERENTIAL INPUT IMPD monitoring can be implemented for voltage setpoint and resistor setpoint as follows. LD Rx 200 10pF PAVSET Voltage Setpoint 08013-011 RERSET 1.23 V PAV RSP ADN2872 In voltage setpoint calibration, the following methods can be used for IMPD monitoring. Figure 29. Differential Measurement of IMPD Across a Sense Resistor Method 1: Measuring Voltage at RPAV VCC The IMPD current is equal to the voltage at RPAV divided by the value of RPAV (see Figure 28) as long as the laser is on and is being controlled by the control loop. This method does not provide a valid IMPD reading when the laser is in shutdown or fail mode. A microconverter-buffered ADC input can be connected to RPAV to make this measurement. No decoupling or filter capacitors should be placed on the RPAV node because this can disturb the control loop. MICROCONVERTER ADC INPUT Rx 200 VCC LD PHOTODIODE PAVSET ADN2872 08013-012 RPAVSET Figure 30. Single Measurement of IMPD Across a Sense Resistor VCC Resistor Setpoint PHOTODIODE PAVSET ADN2872 MICROCONVERTER ADC INPUT R 1k 08013-043 RPAV Figure 28. Single Measurement of IMPD at RPAV in Voltage Setpoint Mode In resistor setpoint calibration, the current through the resistor from PAVSET to ground is the IMPD current. The recommended method for measuring the IMPD current is to place a small resistor in series with the PAVSET resistor (or potentiometer) and measure the voltage across this resistor, as shown in Figure 31. The IMPD current is then equal to this voltage divided by the value of the resistor used. In resistor setpoint, PAVSET is held to 1.2 V nominal; it is recommended that the sense resistor should be selected so that the voltage across the sense resistor does not exceed 250 mV. Rev. 0 | Page 13 of 20 ADN2872 VCC POWER CONSUMPTION PHOTODIODE The ADN2872 die temperature must be kept below 125C. The LFCSP package has an exposed paddle that should be connected such that it is at the same potential as the ADN2872 ground pins. Power consumption can be calculated as: PAVSET ADN2872 MICROCONVERTER ADC INPUT 08013-040 ICC = ICC min + 0.3 IMOD R P = VCC x ICC + (IBIAS x VBIAS_PIN) + IMOD (VMODP_PIN + VMODN_PIN)/2 Figure 31. Single Measurement of IMPD Across a Sense Resistor in Resistor Setpoint IMPD Monitoring TDIE = TAMBIENT + JA x P LOOP BANDWIDTH SELECTION To ensure that the ADN2872 control loops have sufficient bandwidth, the average power loop capacitor (PAVCAP) and the extinction ratio loop capacitor (ERCAP) are calculated using the laser slope efficiency and the average power required. For resistor setpoint control, PAVCAP 3.2 10 6 ERCAP LI PAV PAVCAP 2 (Farad) Thus, the maximum combination of IBIAS + IMOD must be calculated. (Farad) For voltage setpoint control, PAVCAP 1.28 10 6 ERCAP PAVCAP 2 LI PAV (Farad) (Farad) AUTOMATIC LASER SHUTDOWN (Tx_DISABLE) ALS (Tx_DISABLE) is an input that is used to shut down the transmitter optical output. The ALS pin is pulled up internally with a 6 k resistor and conforms to SFP MSA specifications. When ALS is logic high or open, both the bias and modulation currents are turned off. BIAS AND MODULATION MONITOR CURRENTS where: PAV (mW) is the average power required. LI (mW/mA) is the typical slope efficiency at 25C of a batch of lasers that are used in a design. The preceding capacitor estimation formulas are used to obtain a centered value for the particular type of laser that is used in a design and average power setting. Laser LI can vary by a factor of 7 between different physical lasers of the same type and across temperature without the need to recalculate the PAVCAP and ERCAP values. In the ac coupling configuration, LI can be calculated as P1 P0 LI I MOD where: ICC min is 30 mA, the typical value of ICC provided in Table 1 with IBIAS = IMOD = 0. TDIE is the die temperature. TAMBIENT is the ambient temperature. VBIAS_PIN is the voltage at the IBIAS pin. VMODP_PIN is the voltage at the IMODP pin. VMODN_PIN is the voltage at the IMODN pin. (mW/mA) where P1 is the optical power (mW) at the one level, and P0 is the optical power (mW) at the zero level. These capacitors are placed between the PAVCAP and ERCAP pins and ground. It is important that these capacitors are low leakage multilayer ceramics with an insulation resistance greater than 100 G or a time constant of 1000 sec, whichever is less. The capacitor tolerance can be 30% from the calculated value to the available off-the-shelf value, including the capacitor's own tolerance. IBMON and IMMON are current-controlled current sources that mirror a ratio of the bias and modulation current. The monitor bias current, IBMON, and the monitor modulation current, IMMON, should both be connected to ground through a resistor to provide a voltage proportional to the bias current and modulation current, respectively. When using a microcontroller, the voltage developed across these resistors can be connected to two of the ADC channels, making available a digital representation of the bias and modulation current. IBIAS PIN ADN2872 has one on-chip, 800 pull-up resistor. The current sink from this resistor is VIBIAS dependent. I UP VCC V IBIAS 0. 8 (mA) where VIBIAS is the voltage measured at the IBIAS pin after setup of one laser bias current, IBIAS. Usually, when set up, a maximum laser bias current of 100 mA results in a VIBIAS of about 1.2 V. In a worst-case scenario, VCC = 3.6 V, VIBIAS = 1.2 V, and IUP 3 mA. This on-chip resistor helps to damp out the low frequency oscillation observed from some inexpensive lasers. If the onchip resistance does not provide enough damping, one external RZ may be necessary (see Figure 32 and Figure 33). Rev. 0 | Page 14 of 20 ADN2872 The 30 transmission line used is a compromise between drive current required and total power consumed. Other transmission line values can be used, with some modification of the component values. The R and C snubber values in Figure 32, 24 and 2.2 pF, respectively, represent a starting point and must be tuned for the particular model of laser being used. RP, the pullup resistor, is in series with a very small (0.5 nH) inductor. In some cases, an inductor is not required or can be accommodated with deliberate parasitic inductance, such as a thin trace or a via placed on the PC board. DATA INPUTS Data inputs should be ac-coupled (10 nF capacitors are recommended) and are terminated via a 100 internal resistor between the DATAP and DATAN pins. A high impedance circuit sets the common-mode voltage and is designed to allow maximum input voltage headroom over temperature. It is necessary to use ac coupling to eliminate the need for matching between common-mode voltages. LASER DIODE INTERFACING The schematic in Figure 32 describes the recommended circuit for interfacing the ADN2872 to most TO-Can or coax lasers. These lasers typically have impedances of 5 to 7 and have axial leads. The circuit shown works over the full range of data rates from 155 Mbps to 3.3 Gbps including multirate operation (with no change to PAVCAP and ERCAP values); see Figure 18, Figure 19, and Figure 20 in the Typical Performance Characteristics section for multirate performance examples. Coax lasers have special characteristics that make them difficult to interface to. They tend to have higher inductance, and their impedance is not well controlled. The circuit in Figure 32 operates by deliberately misterminating the transmission line on the laser side, while providing a very high quality matching network on the driver side. The impedance of the driver side matching network is very flat vs. frequency and enables multirate operation. A series damping resistor should not be used. Take care to mount the laser as close as possible to the PC board, minimizing the exposed lead length between the laser can and the edge of the board. The axial lead of a coax laser is very inductive (approximately 1 nH per millimeter). Long exposed leads result in slower edge rates and reduced eye margin. Recommended component layouts and gerber files are available by contacting sales at Analog Devices. Note that the circuit in Figure 32 can supply up to 56 mA of modulation current to the laser, sufficient for most lasers available today. Higher currents can be accommodated by changing transmission lines and backmatch values. This interface circuit is not recommended for butterfly-style lasers or other lasers with 25 characteristic impedance. Instead, a 25 transmission line and inductive (instead of resistive) pull-up is recommended. Contact sales for recommendations on transmission lines and backmatch values. The ADN2872 also supports differential drive schemes. These can be particularly useful when driving VCSELs or other lasers with slow fall times. Differential drive can be implemented by adding a few extra components. A possible implementation is shown in Figure 33. VCC L (0.5nH) RP 24 VCC C1 100nF IMODP Tx LINE 30 RZ IBIAS CCBIAS Tx LINE 30 VCC R 24 In Figure 32 and Figure 33, Resistor RZ is required to achieve optimum eye quality. The recommended value is approximately 200 ~ 500 . C 2.2pF L 08013-014 ADN2872 BLM18HG601SN1D Figure 32. Recommended Interface for ADN2872 AC Coupling VCC L4 = BLM18HG601SN1D L1 = 0.5nH R1 = 15 L3 = 4.7nH C1 = C2 = 100nF TO-CAN/VCSEL IMODN 20 TRANMISSION LINES ADN2872 R3 C3 SNUBBER LIGHT IMODP CCBIAS IBIAS R2 = 15 (12 TO 24) L5 = 4.7nH L2 = 0.5nH L6 = BLM18HG601SN1D VCC RZ SNUBBER SETTINGS: 40 AND 1.5pF, NOT OPTIMIZED, OPTIMIZATION SHOULD CONSIDER THE PARASITIC OF THE INTERFACE CIRCUITRY. Figure 33. Recommended Differential Drive Circuit Rev. 0 | Page 15 of 20 08013-041 VCC ADN2872 ALARMS The ADN2872 has a latched active high monitoring alarm (FAIL). The FAIL alarm output is an open drain in conformance with SFP MSA specification requirements. The bias current alarm trip point is set by selecting the value of resistor on the IBMON pin to GND. The alarm is triggered when the voltage on the IBMON pin goes above 1.2 V. The ADN2872 has a three-fold alarm system that covers: FAIL is activated when the single-point faults in Table 5 occur. Use of a bias current higher than expected, likely as a result of laser aging. Out-of-bounds average voltage at the monitor photodiode (MPD) input, indicating an excessive amount of laser power or a broken loop. Undervoltage in the IBIAS pin (laser diode cathode) that increases the laser power. Table 5. ADN2872 Single-Point Alarms Alarm Type Bias Current MPD Current Crucial Nodes Mnemonic IBMON PAVSET ERREF (the ERRREF is designed tied to VCC in resistor setting mode) IBIAS Overvoltage or Short to VCC Condition Alarm if > 1.2 V Alarm if > 2.0 V Alarm if shorted to VCC (the alarm is valid for voltage setting mode only) Ignore Undervoltage or Short to GND Condition Ignore Alarm if < 0.4 V Alarm if shorted to GND Alarm if < 0.6 V Table 6. ADN2872 Response to Various Single-Point Faults in AC-Coupled Configuration, as Shown in Figure 32 Mnemonic CCBIAS PAVSET PAVREF Short to VCC Fault state occurs Fault state occurs Voltage mode: fault state occurs Resistor mode: tied to VCC Voltage mode: fault state occurs Resistor mode: tied to VCC Short to GND Fault state occurs Fault state occurs Fault state occurs Open Does not increase laser average power Fault state occurs Fault state occurs Fault state occurs ERCAP PAVCAP DATAP DATAN ALS ERSET IMMON ERREF Does not increase laser average power Fault state occurs Does not increase laser average power Does not increase laser average power Output currents shut off Does not increase laser average power Does not affect laser power Voltage mode: fault state occurs IBMON FAIL IMODP IMODN IBIAS Resistor mode: tied to VCC Fault state occurs Fault state occurs Does not increase laser average power Does not increase laser average power Fault state occurs Does not increase laser average power Fault state occurs Does not increase laser average power Does not increase laser average power Normal currents Does not increase laser average power Does not increase laser average power Voltage mode: does not increase average power Resistor mode: fault state occurs Does not increase laser average power Does not increase laser average power Does not increase laser average power Does not increase laser average power Fault state occurs Voltage mode: fault state occurs Resistor mode: does not increase average power Does not increase laser average power Fault state occurs Does not increase laser average power Does not increase laser average power Output currents shut off Does not increase laser average power Does not increase laser average power Does not increase laser average power RPAV Rev. 0 | Page 16 of 20 Does not increase laser average power Does not increase laser average power Does not increase laser average power Does not increase laser power Fault state occurs ADN2872 OUTLINE DIMENSIONS 0.60 MAX 4.00 BSC SQ TOP VIEW 0.50 BSC 3.75 BSC SQ 0.50 0.40 0.30 1.00 0.85 0.80 12 MAX PIN 1 INDICATOR 13 12 2.30 SQ 2.15 7 6 0.23 MIN 2.50 REF 0.05 MAX 0.02 NOM SEATING PLANE *2.45 EXPOSED PAD (BOTTOMVIEW) 0.80 MAX 0.65 TYP 0.30 0.23 0.18 24 1 19 18 0.20 REF COPLANARITY 0.08 FOR PROPER CONNECTION OF THE EXPOSED PAD, REFER TO THE PIN CONFIGURATION AND FUNCTION DESCRIPTIONS SECTION OF THIS DATA SHEET. 080808-A PIN 1 INDICATOR 0.60 MAX *COMPLIANT TO JEDEC STANDARDS MO-220-VGGD-2 EXCEPT FOR EXPOSED PAD DIMENSION Figure 34. 24-Lead Lead Frame Chip Scale Package [LFCSP_VQ] 4 mm x 4 mm Body, Very Thin Quad (CP-24-2) Dimensions shown in millimeters ORDERING GUIDE Model ADN2872ACPZ1 ADN2872ACPZ-RL1 ADN2872ACPZ-R71 1 Temperature Range -40C to +85C -40C to +85C -40C to +85C Package Description 24-Lead LFCSP_VQ 24-Lead LFCSP_VQ 24-Lead LFCSP_VQ Z = RoHS Compliant Part. Rev. 0 | Page 17 of 20 Package Option CP-24-2 CP-24-2 CP-24-2 Ordering Quantity 490 5,000 1,500 ADN2872 NOTES Rev. 0 | Page 18 of 20 ADN2872 NOTES Rev. 0 | Page 19 of 20 ADN2872 NOTES (c)2009 Analog Devices, Inc. All rights reserved. Trademarks and registered trademarks are the property of their respective owners. D08013-0-3/09(0) Rev. 0 | Page 20 of 20