Hello everyone, I am attaching herewith few time domain visualizer output got from simulation using optiSystem, and some voltage overshoot and undershoot is noticed, it is seen in most of the optical transmitter design. what may be the possible reason for this.?
I am using DP-16-QAM modulation. the doubt area is shown by the red circles and it is in optical domain.
I have also attached the X-polarization of the transmitter structure.The two graphs are taken from the red circled visualizer.
With regards,
Dhiman Kakati
What seems to disturb you: voltage drops vs. zero.
Seems you output is from simulation, thus we should look for problems resulting from 'ideality' that may or may not surface in real-worls implementations...
Id look for unmatched delays - meaning that signal delays may differ between signal pathes. Reflections could add to that.
If your simulation is capable of that: to find out where the porolem is, add oscilloscope probes at the respective location on parallel paths and look at the difference resp. sum of the signals measured.
In a real system signals are causal and you will find some inductive peaking to enhance the bandwidth. Since this is electrical it may provide some overshoot and undershoot in the detected waveform. In a simulation there may be timing errors (the symbol is shorter than the symbol time). If the symbol is measured at the centre of the period then this will not be an issue anyway. There does tot seem to be any evidence of bandwidth shaping in these results. Take a look at some of the RF communication notes as well as the optical communication notes as this will help you. Simulation is useful but it is only a model.
Dhiman,
It looks like you are simulating a system operating with symbol rate around 15 GBd (67 ps symbol period), with step transitions between symbols.
As David points out, some of the ringing is acausal, and unlikely to be caused by explicit bandwidth shaping.
Are you able to display the transients with better resolution? What time step is used in the simulation? Are the peaks always within one simulation time step of the ideal transition time?
My guess is the problem occurs in Coupler_1 where the in-phase and quadrature optical signals are combined. Does the 90 degree phase shift in the Q arm introduce a path delay, or is it a pure mathematical rotation of the carrier phase?
As U. Dreher suggests, examine the waveforms at other points in the system.
Do you have a clean step modulation of the optical field between the output of the Mach-Zehnder modulators and Coupler_1?
Are you able to examine the transition in the optical field, rather than just the optical power envelope?
Do the Mach-Zehnder modulators introduce any chirp during transition? Are they idealized devices, or do they include bandwidth limitations which you have set to a very high value.
If it is a problem with transition timing errors in the simulation, does it help to low pass filter the electrical drive to the M-Z modulators? This will be more representative of a real system, and even a 50 GHz low pass would give you a very open eye with moderate clock delay. You will benefit from a rather lower bandwidth in your receiver (perhaps 10 to 15 GHz?) if there is significant noise in your system.
Good luck.
Alan.
- Badges
- Science topic
- Similar topics
- Physical Sciences
- Optics