Fieldworthy ROFL/OFL Multimode Fiber Differential Mode Delay Measurement System

Transcription

1 Fieldworthy ROFL/OFL Multimode Fiber Differential Mode Delay Measurement System Lew Aronson and Lisa Buckman HP Labs Palo Alto February 2, 1998

2 Outline Measurement Goals and Issues Functional Block Diagram Source Characteristics ROFL and OFL Launch Implementations Detector Characteristics Final System System Impulse Response Typical Results Overview of Fibers Tested Future Issues

3 Acknowledgments Janis Valdmanis (Picometrix) - Detector Loan Richard Weiss (HP-San Jose) David Parkinson, Richard Wade, Stephen Gee (HP-Ipswich) Scott Corozine and Rick Schneider Jean Norman Dave Dolfi and Waguih Ishak (HP Labs Palo Alto) Steve Hinch and Doug Fullmer (HP -Santa Rosa)

4 Measurement Goals Perform Differential Mode Delay (DMD) Measurements on Greatest Possible Number of Fibers in Typical Installation Using Radially Overfilled Launch (ROFL) and Standard Overfilled Launch (OFL). Use Source Wavelengths Appropriate to 1000SX and 1000LX. Use Narrow Linewidth SX Source so Chromatic Dispersion is Negligible. Data to be used for: Developing a statistical base for understanding the distribution of DMD Measuring the details of the differential mode delay in these fibers Identifying particularly poor fibers on which to test conditioned launch solutions to the differential mode delay problem.

5 Issues related to Field Testing Previous ROFL DMD measurements conducted in Lab Environment In Field Test, Transmitter and Receiver Unit Separated Control Signals Not Available, Triggering A Special Problem Time Domain Method Selected Alignment Stages and Other Delicate Adjustments Must be Avoided Quick Connect ROFL Coupler Required Pigtailed Narrow Linewidth Single Mode Sources Needed Not Commercially Available for 850 nm Measurement Time at Site Limited, System must be Very Efficient.

6 Functional Block Diagram Transmitter System Laptop Data Acquisition Optical Pulse Source SX LX Ref RX Sampling Scope Sampling Scope SX 0.5 db ROFL SX 1.0 db ROFL SX 2.0 db ROFL Signal Trigger Launch Optics SX OFL LX 0.5 db ROFL LX 1.0 db ROFL LX 2.0 db ROFL Installed Fiber Under Test High Speed Detector LX OFL Receiver System Site 1 Site 2

7 Optical Pulse Source Gain Switching of Semiconductor Lasers Used for Generating Short Pulses 850 nm Source was Small Aperture VCSEL 1310 nm Source was MQW Fabry-Perot Laser Laser Sources Pigtailed into Appropriate Single Mode Fiber Corning Flexcore 780 for SX SMF28 for LX Narrow Linewidth Required at Short Wavelength

8 SMA Connector Optical Pulse Source Small Aperture VCSEL UV Cure Adhesive Flexcore 780 Pigtail Current Source Electrical Pulse Generator Pigtailed Laser Package 3.3 V 250 ps RF Switch Bias T Current Source Bias T 850 nm SM Source 1300 nm SM Source Flexcore 780 Pigtail SMF 28 Pigtail 1-5 mw < 30 ps FWHM 1-5 mw < 40 ps FWHM Picture of FP Package

9 Optical Source Characteristics - Short Wavelength Gain Switched Output Pulse (Measured with 25 GHz Receiver) 0-10 Spectrum When Gain Switched (Chromatic Dispersion Bandwidth > 10 GHz km) λ -3dB = 0.13 nm λ -10dB = 0.30 nm Power (au) ~ 30 ps Power (dbrel) Time (ps) Wavelength (nm)

10 Optical Source Characteristics - LX Gain Switched Output Pulse (Measured with 20 GHz Receiver) Spectrum When Gain Switched (Chromatic Dispersion Neglible at 1300 nm) λ -3dB ~ 10 nm Power (au) ~ 40 ps Power (dbrel) Time (ps) Wavelength (nm)

11 Radial Overfilled Launch Well Defined Launch to Selectively Generate Worst Case Modal Excitations Typical of Laser Launch Single Mode Fiber (for Wavelength under Test) Multimode Fiber Under Test (50 or 62 MMF) Gap Defined By Coupling Loss 12 ROFL Launch Types: (SX or LX) x (50 or 62 MMF) x (0.5 db, 1.0 db or 2.0 db Loss)

12 Measurement of ROFL Loss vs Gap Transmission (db) nm, SMF28 to 62MMF 1310nm, SMF28 to 50MMF 844nm, Flexcore780 to 62MMF 844nm, Flexcore780 to 50MMF Gap (um)

13 ROFL Spacer Mechanism Ceramic Split Sleeve From ST-ST Bulkhead Installed Ceramic Disk Spacer Defines Gap Between Connectors (Slides to Guarantee Contact with Ferrules) End View Ceramic Disk Spacer Diced from Ferrule (140 µm Aperture)

14 ROFL Spacer Mechanism ST Fiber Connector Connector Ferrule Before Making Contact with Spacer Disk

15 Log Magnitude (db) ROFL Spacer Disk Fabrication and Calibration 1) Disks Diced From Standard 2.5 mm Ferrule to Approximate Thickness 2) Disks Placed in ST/ST Bulkhead with Ceramic Split Sleeve 3) Gap Between Fiber Ends Measured with Precision Reflectometer to µm Accuracy. Measured Data Deltas (um): Measurement Span (microns) 4) Couplers Tested for Measured Losses. In Some Cases Predicted Gap Gave Wrong Loss: Variations in SMF Numerical Aperture (?) 5) Spacers Adjusted to Give Required Loss with Fibers Used in Final System

16 ROFL Spacer Disk Accuracy ROFL Coupler Measured Data Wavelength Fiber Type Nominal Loss (db) Disk Thickness Actual Loss (db) Error (db) SX 50 MMF SX 50 MMF SX 50 MMF LX 50 MMF LX 50 MMF LX 50 MMF SX 62.5 MMF SX 62.5 MMF SX 62.5 MMF LX 62.5 MMF LX 62.5 MMF LX 62.5 MMF

17 Overfilled Launch GI-SI Mode Scrambler from TIA Standard Used Input 62 MMF 100 um SI Fiber NA = MMF Front Panel Input Output MMF with 15 um Offset in Connector Spool Output Launch CPR Measured: 850nm/50MMF: 20.0 db 850nm/62MMF: 21.7 db 1300nm/50MMF: 16.1 db 1300nm/62MMF: 17.7 db 50MMF Class 1, 62 MMF Class 2 Try SI-GI-SI for Better Connector with 15 um Offset 100 um Step Index Fiber

18 Receiver Characteristics 70 um Pigtailed InGaAs Detector from Picometrix Used Required to Collect all Modes from 62 MMF Good Responsivity at SX (0.45 A/W) and LX (0.69 A/W) Fast Impulse Response (38 ps) 14 GHz and 12 GHz Picosecond Pulse Labs Amplifiers Splitter, Extra Amplification on Trigger Output

19 Transmitter Unit - Front Panel SX Reference Receiver Input SX Reference Receiver Output ROFL Output (to Fiber Under Test) Input to ROFL Generator Single Mode Source Outputs LX Reference Receiver Input LX Reference Receiver Output Laser Bias Control Input to OFL Generator OFL Output (to Fiber Under Test) Electrical Pulse Input Wavelength Selection

20 Transmitter Unit - Inside OFL Generator 20 GHz LX Ref Receiver 25 GHz SX Ref Receiver Microwave Switch Flexcore 780 ROFL Couplers 1300 nm Source SMF 28 to Front Panel Flexcore 780 to Front Panel 850 nm Source Bias Generator SMF 28

21 Transmitter Unit - Typical Configuration Flexcore 780 Jumper to Desired ROFL Unit MMF Patchcord to Fiber Under Test

22 Receiver Unit - Front Panel Recived Pulse Output Scope Trigger Output Optical Input from Fiber Under Test

23 Receiver Unit - Inside Signal Output 2 GHz - 20 db Amplifier (Trigger Only) Optical Input (62 MMF) Resistive Divider 14 GHz - 10 db Amplifier Picometrix Detector (~ 9 GHz, 62 MMF Input) 12 GHz - 10 db Amplifier

24 Laptop Data Acquisition Program

25 Complete Instrument Carts Laptop with GPIB for Data Acquisition RX Unit 20 GHz Sampling Scope 20 GHz Sampling Scope TX Unit Received Trace Electrical Pulse Generator

26 System Impulse Responses Short Wavelength (861 nm) Long Wavelength (1310 nm) Power (au) ~ 70 ps Power (au) ~ 80 ps Amplifier Related 40 Amplifier Related Time (ps) Time (ps) Bandwidth > 4 GHz Bandwidth > 3.5 GHz

27 Typical Fiber Measurement Recorded Data File Multimode Fiber Bandwith Test: 12/17/97 Filename: C:\Field Test Data\LG011116S0p.dat Repetition: p 25 Received Pulse Fiber ID: LG Fiber Info 1: LGUM Node 1 (B256) - B451 Fiber Info 2: ATT Lightguide Test Location: LLNL Fiber Length (m): 457 Fiber Type: 62 MMF Launch Type: 0.5 db ROFL Wavelength: 850 nm Estimated DMD (ps): 510 Comments: Power (au) Number of Averages: 64 Time (ps) Amplitude (mv) Time (ps). (Rough Corrected Bandwidth ~ 245 MHz km)

28 Overview of Fibers Tested Total of 78 Fibers Tested week of 12/15: MMF at Lawrence Berkeley National Labs MMF at Lawrence Livermore National Labs MMF at Sandia National Labs 8 or 9 Measurements per Fiber Last 3 Days Single Mode Launch at 1310 nm Also Conducted Tests on One Fiber Took 8-10 minutes Notes on Statistical Validity Most Fibers Selected by Ease of Access One Cable (12) at LBNL Selected for Previous Bad Performance (And the resulting Bandwidths (ROFL and OFL) were very Low) This Bad Cable Accounts for ~20% of 62 MMF Tested.

29 Histogram of Fiber Bandwidths Percent Below Bandwidth 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Cumulative Histogram MMF - SX Bandw idth (MHz - km) 0.5 ROFL 1.0 ROFL 2.0 ROFL OFL Percent Below Bandwidth 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Cumulative Histogram - 50 MMF - SX Bandw idth (MHz - km ) 0.5 ROFL 1.0 ROFL 2.0 ROFL OFL Percent Below Bandwidth 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Cumulative Histogram MMF - LX 0.5 ROFL 1.0 ROFL 2.0 ROFL OFL SM Bandw idth (MHz - km ) Percent Below Bandwidth 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Cumulative Histogram - 50 MMF - LX 0.5 ROFL 1.0 ROFL 2.0 ROFL OFL SM Bandw idth (MHz - km ) Represents Data corrected by full deconvolution with calibration data

30 ROFL vs OFL Bandwidth - 62 MMF SX ROFL and OFL Bandwidths - SX - 62 MMF Minimum of ROFLs OFL 640 Bandwidth (MHz km) Fiber ID Arrows Indicate Previously Known Bad Cable

31 ROFL vs OFL Bandwidth - 62 MMF SX -No 72X ROFL and OFL Bandwidths - SX - 62 MMF No 72X Fibers Minimum of ROFLs OFL 480 Bandwidth (MHz km) Fiber ID

32 ROFL vs OFL Bandwidth - 50 MMF SX ROFL and OFL Bandwidths - SX - 50 MMF Minimum of ROFLs OFL N05A1109 N05A1112 N05A1101 N05A1110 N05A1103 N05A1111 N05A1108 N05A1102 N05A1104 N05A1107 N05A1106 N05A1105 N04A1001 N03A0911 N03A0910 N03A0909 N04A1005 N04A1002 N03A0912 N04A1006 Fiber ID Bandwidth (MHz km)

33 ROFL vs OFL Bandwidth - 62 MMF LX ROFL and OFL Bandwidths - LX - 62 MMF Bandwidth (MHz km) Minimum of ROFLs OFL Fiber ID Arrows Indicate Previously Known Bad Cable

34 ROFL vs OFL Bandwidth - 62 MMF LX -No 72X ROFL and OFL Bandwidths - SX - 62 MMF No 72X Fibers Bandwidth (MHz km) Minimum of ROFLs OFL Fiber ID

35 ROFL vs OFL Bandwidth - 50 MMF LX Minimum of ROFLs OFL ROFL and OFL Bandwidths - LX - 50 MMF N03A0912 N05A1109 N05A1112 N03A0911 N04A1002 N04A1006 N03A0909 N05A1105 N03A0910 N04A1001 N04A1005 N05A1104 N05A1101 N05A1111 N05A1110 N05A1107 N05A1103 N05A1106 N05A1108 N05A1102 Fiber ID Bandwidth (MHz km)

36 Some Interesting Results - 62 MMF, SX Short Wavelength, 62 MMF [58 fibers, 12 fibers in Bad Cable ] 11 of 58 < 160 MHz km - Including Bad Cable 4 of 58 < 160 MHz km - Excluding Bad Cable All fibers below 160 MHz km had BW(OFL) < BW(ROFL) All fibers which pass OFL spec pass WCMB spec

37 Some Interesting Results - 62 MMF, LX Long Wavelength, 62 MMF [58 fibers, 12 fibers in Bad Cable ] 32 of 58 < 500 MHz km - Including Bad Cable 16 had BW(OFL) < BW(ROFL), 16 had BW(OFL) > BW(ROFL) 20 of 58 < 500 MHz km - Excluding Bad Cable 6 had BW(OFL) < BW(ROFL), 14 had BW(OFL) > BW(ROFL) 19 of 58 fibers failed 500 MHz km OFL spec 13 of 58 fibers passed 500 MHz km OFL spec but had ROFL < 500 MHz km All fibers which passed 500 MHz km OFL spec passed WCMB spec (250) 6 of 58 < 250 MHz km - Including Bad Cable 5 had BW(OFL) < BW(ROFL), 1 had BW(OFL) > BW(ROFL) 3 of 58 < 250 MHz km - Excluding Bad Cable 3 had BW(OFL) < BW(ROFL), 0 had BW(OFL) > BW(ROFL) 6 of 58 fibers failed 250 MHz km OFL spec All fiber which passed 250 MHz km OFL spec passed WCMB spec (250)

38 Some Interesting Results - 50 MMF, SX Short Wavelength, 50 MMF (20 fibers) 14 of 20 < 500 MHz km 1 had BW(OFL) < BW(ROFL), 13 had BW(OFL) > BW(ROFL) 11 of 20 < 400 MHz km 1 had BW(OFL) < BW(ROFL), 10 had BW(OFL) > BW(ROFL) 12 of 20 fibers failed 500 MHz km OFL spec 2 of 20 fibers passed 500 MHz km OFL spec but had ROFL < 500 MHz km 7 of 20 fibers failed 400 MHz km OFL spec 4 of 20 fibers passed 400 MHz km OFL spec but had ROFL < 400 MHz km

39 Some Interesting Results - 50 MMF, LX Long Wavelength, 50 MMF (20 fibers) 1 of 20 < 500 MHz km 0 had BW(OFL) < BW(ROFL), 1 had BW(OFL) > BW(ROFL) All fibers passed 500 MHz km OFL spec 1 of 20 fibers passed 500 MHz km OFL spec but had ROFL < 500 MHz km All fibers passed 500 MHz km OFL spec and had ROFL > 375 MHz km