This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Troubleshooting Coherent Optical Communication Systems
Created by:Michael KoenigsmannApplication SpecialistDigital & Photonic Test Division
• Requirements for a test instrument:• Flexibilty to address different modulation schemes• Clean signal to test your device and not your instrument• Bandwidth at least 20 GHz• 4 sychronized channels to support dual polarization (= 2 pairs of I/Q signal)
• Clean Signal generation (electrical & optical)• Clean Signal at defined test points• High datarates 32 GBd and beyond• Accurate and Repeatable Test Signals• Distortion Emulation• Flexible Modulation Formats
Page
Accurate and Repeatable Test Results
Out-of-the-box calibration ensures clean signal at the frontpanel connector
In-situ calibration – extend clean signal to the receiver test point
• S-Parameters of channel are embedded or de-embedded
• Frequency/Phase response is measured in-system andthen de-embedded
11
QPSK, 32 GBd PRBS 6 Gbit/s
Without correction With correction Without correction With correction
Measure frequency/phase response as determined by the coherent receiver End-to-end calibration of the whole transmission system This includes the characteristic of the receiver as well !
Typical test requirements: BER (EVM) vs. Power or OSNR Spectral behavior over 1-2 neighbor channels Signal integrity along the link (time domain modulation analysis) Nonlinear characteristics (PMD, PDL, CD,…)
Typical requirements: Availability of numerous modulation formats for research and development High quality PAMx, (D)QPSK and QAMx signal for manufacturing Arbitrary constellations for advanced research required (arbitrary stress) BER vs various stress parameters Some customers need RZ DQPSK Support of up to 56 GBd rates for >> 100G research PRBS test pattern, Pre-defined or User-defined Test of receiver algorithm robustness with distorted signals
Page 17
Test application: Rx stress test (electrical)
Typical requirements: Availability of numerous modulation formats for research and development High quality PAMx, (D)QPSK and QAMx signal for manufacturing Arbitrary constellations for advanced research required (arbitrary stress) BER vs various stress parameters Some customers need RZ DQPSK Support of up to 56 GBd rates for >> 100G research PRBS test pattern, Pre-defined or User-defined Test of receiver algorithm robustness with distorted signals
Multi-format Reference Transmitter oroptical AWG with or without stress generation
M8195A + OMFT + Tunable Laser+ Optical Modulation Gen. Tool
Typical test requirements: Cut-off frequency for each path In-application test to specify best EVM and other measures Polarization behavior Phase Response, Skew Modulator Bias Control
Error Vectorconnects the measuredVector and the expectedvector, Error VectorMagnitude or EVM is magnitude of this vector
EVM Q-factor1% ~ 100
2% ~ 50
5% ~ 20
10% ~ 10
The Q-Factor describes the signal-to-noise ratio at the decision points. It is calculated from the EVM. The formula is proportional to 1/EVMa and the result is converted into dB. It is calculated from the Eye-Diagram
EVM_pctlEVM_pctl defines the radius of a circle around a group of measured constellation points centered at the reference constellation point.In contrast to EVM_pctl, the classical EVM value from the VSA software includes I-Q Imbalance and Quadrature Error and is a RMS averaged value. The “Hit Ratio” multiplied with the number of points is equal to the number of points outside the circle. For Gaussian Noise dominated impairments, the classical EVM and the EVM_pctl are equal if a “Hit Ratio” of 0.3173 is selected. The "Hit Ratio" can be set in the EVM_percentile algorithm of the OMA Software.
IQ Gain Imbalancecompares the amplitudeof the I signal with the amplitude of the Q signal and shows the difference in dB.The effects of IQ gain imbalance are best viewed in constellation diagrams where the width of the constellation diagram is different than its height. Rectangular shape of constellation, different amplitudes in I- and Q-Eye
IQ OffsetDC offsets at the I- or the Q-Signals cause I/Q or origin offsets as shown. I/Q offsets can also resultin carrier feedthrough.It is a measure for the shift between the origin of the measured constellation with regard to the origin of the reference constellation (yellow).Without DC offsets, the carrier feed through as well as the IQ offset becomes zero (-infinity dB). Vertically shifted Eye-
error between the I and Q Quadrature-Phase. Ideally, I and Q should be orthogonal (90 degrees apart). In the screenshot to the right a quadrature error of 22.91 degrees means I and Q are 67,09degrees apart instead of 90 degrees.This could result from wrong bias point setting for the 90° phase shifter in the Mach-Zehnder-Modulator distorted IQ Plot,
Frequency Errorshows the carrier's frequency error relative to the VSA's center frequency displayed in Hertz. It is the amount of frequency shift from the VSA's center frequency that the VSA must perform to achieve carrier lock. The maximum allowable Frequency Error depends on the Modulation Format used. Here’s a list of formats:
2 GHz
Modulation Format Maximum frequency offsetQPSK 9.6% symbol rate
Symbol Rate ErrorIf the Digital Demodulator is only able to recover the clock phase but not the clock rate, a wrongsymbol rate shows up as typical „V“ shape whenlooking at EVM vs timeplot. Symbols start to spreadall over the constellation High EVM
Optical Signal Summary ScreenDisplay of- Timing skew betweenI- and Q-Signals for X-and Y-Polarization- Timing skew betweenX- and Y-Polarization. - Gain Imbalance between X and YPolarization- Symbol Rate Display- Frequency Error between Tx and Rx (in this case OMA)- EVM and Q-Factor
Effects of IQ PRBS Delay in bit (PRBS length 215-1)
0 bit
2 bit
4 bit
1 bit
3 bit
5 bit
Effects of Delay issues between I- & Q-Data Source
Effect of PRBS delayUsing the same PRBS sequence for I and Q and decorrelating it with a toosmall delay might lead to missing transitions and asymetric spectralcontent !Screenshots were takenbased on a PRBS 215-1with different delaysbetween I and Q showingresult in the correspondingconstellation and spectrum.
PhaseNoise Analysis of modulated signals.Display of Carrier Phase,Lorentzian Linewidth, Flicker & Random Noise.Graphical result of Phase Spectrum and corresponding Phase Spectrum Model
Stokes Space Analysis in OMA: Great Circle & Slicer
36
Troubleshooting Coherent Optical Comms Systems
Analysis of Stokes Space Trajectory
The OMA allows to analyzepolarization changes. Two patterns„Great Circle“ and „Slicer“ are generated with the AWG electrically. This electrical signal can be used for receiver stress testing.Here the electrical signal is directlyconnected to the OMA inputs.
The OMA allows to compensate and analyze the PMD.An electrical signal is generated with the AWG and analyzed by the OMA.This signal could beused for receiver stress testing.
Publication title Publication No.N4391A Optical Modulation Analyzer – Data Sheet 5990-3509ENN4392A Integrated Optical Modulation Analyzer – Data Sheet 5990-9863ENInfiniium Z-Series Oscilloscopes - Data Sheet 5991-3868ENM8195A 65 GSa/s Arbitrary Waveform Generator 5992-0014EN
Metrology of Advanced Optical Modulation Formats - White Paper 5990-3748ENKalman Filter Based Estimation and Demodulation of Complex Signals – White paper 5990-6409EN
Vector Signal Analysis Basics - Application Note 5989-1121ENDigital Modulation in Communications Systems - An Introduction – Application Note 5965-7160EEssentials of Coherent Optical Data Transmission - Application Note 5991-1809EN
More Information is available from these Jumpstations:Arbitrary Waveform Generator M8195A andOptical Modulation Generator Software: www.keysight.com/find/M8195AOptical Modulation Analyzer: www.keysight.com/find/OMALightwave Component Analyzer: www.keysight.com/find/LCA