SOA models in Lumerical Interconnect

    • Badrul

      Dear Lumerical,

        I am trying to simulate a SOA with Interconnect, and I would like to have some directions.

      I have seen the TWLM section (, although it seems to be more centered around the model of a semiconductor laser. The example is more related to the electrical tuning of the laser, where the output of both optical ports is 9 dB optical signal; if I try to set a periodic square-wave optical in the optical input, the TWLM does not receive a gain from optical input to output.

      I have seen that Fraunhofer HHI have developed a SOA model (, and I would like to know which parameter I need to tune to obtain a similar effect.

      So, can you please tell me:

      1. what are the parameters required to obtain a gain from the optical port 1 (e.g. a square NRZ bit sequence) to the optical port 2?

      2. Also, I would like to ask you if it is possible to obtain a saturation effect of the SOA, which I show in the attached Figure.

      Sorry for the basic questions, but I am new to the Interconnect module, and overwhelmed by the many parameters.

    • Guanhui Wang
      Ansys Employee
      Hi Badrul
      Thank you for the post. When you use the TWLM as an SOA, we will need a input optical signal to the model then use the model as the amplifier. You may use an already modulated optical signal to get the NRZ waveform. The TWLM model in this case only uses as a gain medium without the reflections to form the cavity, so the parameters are basically the same as a normal TWLM laser model but without the relfections at the ends. You can also import and fit to your gain curve if you have a definition of it.
      For the saturation effect, we already have the self-heating thermal effect built into this model and it can model the power roll-off:
    • Badrul
      Dear Gwang, thank you very much for your answer. I have reduced the facet reflection, and the situation improved greatly. Apparently, even a small value of reflectivity greatly influences the distortion of the output signal. Still, the model is incomplete, and I will better explore the options and your link in these days, and update here for better reference.

      Can you also tell me if in the TWLM model the delay of the element can be related to the population (as happens in real SOAs)?

      Thank you again for any suggestion Best regards Badrul
Viewing 2 reply threads
  • You must be logged in to reply to this topic.