WIP: p_ref and flexgrid

Flexgrid & Reference Power

At design time

• "spectrum"
  • spacing
  • freq range (might be later constrained by amps etc)
  • one reference channel:
    • baud rate
    • roll off
    • "reference power"
      • scalar value: "one power", input at the "reference span"
      • "launch power" but not at TXP egress, at first booster's egress instead
      • if I do a sweep, it doesn't affect this value directly; power sweep affects just the first span
        • so the actual input power at the 1st span is the "ref power" adjusted for span loss and adjusted by power sweep
        • a power sweep won't trigger a redesign today
          • power sweep was to verify that we're at good settings -- Jean-Luc's work
          • you cannot "sweep high enough" if the ref power was "too low" because of amplifier saturation (if amps were chosen based on that "too low" reference power)

The point of autodesign is to determine amplifier type and their operating points. We don't know the actual spectrum allocation (flexgrid), but for optimizing the amps, this is a good approximation. Autodesign is optional (you can feed the full set of parameters manually).

At propagation

User's input

• "spectrum" as created by transponders
  • a global reference channel
    • channel power
    • baud rate
      • for PSD, not required in every case
    • slot width
      • for OpenROADM-style equalization
  • a set of carriers
    • central freq
    • slot width
    • "whatever is in TXP mode"
    • something for equalization at the ROADM level:
      • "power difference" (a delta from the power of a reference carrier)
      • if not present, each ROADM has to determine how to equalize this carrier among the others

input to an element's propagate():

• "spectrum"
  • list of:
    • signal power
    • ase power
    • nli power
    • pdl
    • pmd
    • cd
• a result of a propagation of that virtual noiseless reference carrier
  • power of a noiseless carrier in dBm
  • today known as p_spani
  • (for output printing)
  • for something in EDFAs
    • if it was saturation, then it's incorrect
    • also target gain, not sure about the current status
    • TODO Esther: check this
• initial power at the beginning of each OMS for that virtual noiseless reference carrier
  • power of a noiseless carrier in dBm
  • today known as p_span0