Optiwave Optisystem 'link' -
And that’s the difference between a simulation hobbyist and a system designer. Have you used OptiSystem for long-haul or data center interconnects? What’s the weirdest discrepancy you’ve seen between simulation and lab measurement? Let’s discuss below.
When simulating a dual-polarization 64-QAM system at 64 Gbaud, the difference between "works in simulation" and "works on the bench" comes down to three subtle settings: In OptiSystem, you can set linewidth independently for signal and LO lasers. But the hidden gem: phase correlation time . Many default simulations assume a Wiener process with infinite memory. Real lasers have finite correlation. To match hardware, enable the "Lorentzian with finite correlation time" model and set the correlation time to 1/(π * Δν) . This dramatically changes carrier recovery lock time. 2. Manakov vs. Scalar Nonlinearity For standard SSMF, the Manakov equation is the gold standard. But for few-mode fibers or strongly-coupled systems? Don't use it. Switch to the Coupled Nonlinear Schrödinger Equation (CNLSE) solver. It’s 10x slower but captures inter-modal nonlinear mixing. Watch four-wave mixing between modes—it will destroy your performance in ways Manakov hides. 3. Amplifier ASE with Polarization Effects EDFAs are not isotropic. Use the "Polarization-Dependent Gain (PDG)" model under the amplifier’s "Advanced" tab. Set PDG to ~0.05 dB (realistic for modern EDFAs). Then add a polarization scrambler before the amp. Without this, your simulation will show polarization hole burning artifacts that don't exist in a properly dithered system. Iterative Design Workflow: From Back-to-Back to 2000 km Here’s a professional workflow that saves hours of simulation time: optiwave optisystem
Discipline is key. Start with datasheet values from real components (Finisar, II-VI, Broadcom). Add realistic connector loss (0.5 dB per mated pair). Include filter penalties from ROADMs. Add a safety margin of 2 dB OSNR to your target. And that’s the difference between a simulation hobbyist
Add fiber with loss and CD only (disable SPM/XPM/FWM). Sweep launch power from -10 dBm to +10 dBm. The BER should improve with power (more OSNR) until you hit thermal noise. This curve is your linear baseline. Let’s discuss below
