Dispersion wavelength Dependency

[GanbaruDM]

Dear contributors,

I am writing with regard to the Dispersion wavelength dependency.
At the moment I am simulating hundreds of channels (417) in the S+C+L bands along 4x80 km spans. In order to take into account the dispersion wavelength dependency I have tried to use the parameter dispersion_slope inserting in the equipment
"dispersion_slope": 90 s/m^3 ( obtained from [a] considering a G.652 D fiber: S0max < 0.092 ps/nm^2 km).

Unfortunately, eventhough I am extremizing the simulations increasing the input power to 5.00 dBm, I keep obtaining the same values of the channels SNR NLI (signal bw, dB) wheter inserting the dispersion_slope parameter or not,

In other words, it seems like that in GNPy's opinion the dispersion wavelength dependency phenomenon is negligible even while simulating multiband systems.

Thanks in advance

[a] https://www.itu.int/rec/T-REC-G.652-201611-I/en

[ojnas]

I'm not sure whether dispersion slope should have an impact on NLI but I think it may depend on the model used (GN vs GGN). @AndreaDAmico, any comment?

However, from looking at the code it seems that a dispersion_slope parameter specified in the equipment file is even completely ignored (since there is no default value in json_io.py), which clearly looks like a bug. @jktjkt, I'm not sure how we should address this but we probably need to include default values for all fiber types in the equipment file.

[AndreaDAmico]

Yes, the dispersion slope should have an impact on the NLI generation, but I am not sure about the intensity of this effect. In the current version this is considered only in the GGN.

[EstherLerouzic]

Chromatic dispersion, effective area, Raman Gain coefficient, and nonlinear coefficient can now be defined with a scaling
along frequency in v2.8 release.