mjungo / vistas Goto Github PK

3.0 2.0 2.0 104 KB

License: MIT License

Python 100.00%

simulation vcsel optical-communications vectorized-computation ode waveguide

vistas's Introduction

VISTAS Python

VISTAS is a 2D, multimode, time-domain model aimed at simulating the static and dynamic behaviour of Vertical Cavity Surface Emitting Lasers (VCSEL).

VISTAS accounts dynamically for the spatial interactions between the optical field and carrier distributions in the active layer. Effects such as spatial hole burning, modal gain competition, or carrier diffusion losses can therefore be studied in details.

The original VISTAS algorithm was published in the early 2000s [1][2]. It is based on 2D spatially-dependent rate equations, mathematically transformed so as to remove any explicit spatial dependency. The resulting model consists therefore in a system of Ordinary Differential Equations (ODEs). It can be implemented in a fully vectorized fashion and exhibits several orders of magnitude improvement in terms of computational efficiency compared to the original spatially-dependent rate equations.

The model can also easily be extended with a series of important effects affecting the performance of VCSEL-based optical links, such as carrier transport into the quantum wells, noise, thermal effects, etc.

The aim of the project hosted in this repository is to develop a modern implementation of VISTAS in Python. Contributions are warmly welcomed!

References:

[1] M. Jungo, "Spatiotemporal VCSEL Model for Advanced Simulations of Optical Links," in Series in Quantum Electronics, vol. 30, edited by H. Baltes, P. Günter, U. Keller, F. K. Kneubühl, W. Lukosz, H. Mechior, and M. W. Sigrist, 1st ed. Konstanz: Hartung-Gorre Verlag, 2003, ISBN 3-89649-831-2

[2] M. Jungo, D. Erni and W. Baechtold, "VISTAS: a comprehensive system-oriented spatiotemporal VCSEL model," ” IEEE J. Sel. Top. Quantum Electron., vol. 9, no. 3, pp. 939–948, May/Jun. 2003

vistas's People

Contributors

Stargazers

Watchers

Forkers

bipinsubedi yaohuic

vistas's Issues

Implement time-domain logarithmic gain model

g0(t)=gln*ln((Na(t)+1)/Ntr)/(Na(t)-Ntr)
where Na(t) is the average carrier density over the active area delimited by the oxide aperture.

Develop and implement concept for simulation data structure

VCSEL parameters (currently a file with a single function that outputs the hard-coded VCSEL parameters single function file)
simulation parameters (features, signal, time, , hyperparameters, resolution algorithm, output visualization, etc)
simulation results

Validate parasitics model

Validate parasitics modelling approach consisting in

defining a simple parasitics circuit (Cp, Rm, Ca, Ra) (NB. the model could easily be extended with additional parasitics elements, such as bond wire resistance and inductance)
deriving the frequency response of that circuit Hp(f)
transforming the ideal drive current signal to the frequency domain through FFT
applying the parasitics filter to the signal by multiplication in the frequency domain
converting the filtered signal back to the time domain through IFFT
running a standard simulation with the filtered time-domain drive current signal as input
evaluating the frequency response of the system as the addition (in dB) of the intrinsic response and the parasitics response

VCSEL params
modes solver
models
main (incl. visualization)

formulation of Langevin terms
Langevin term for barrier term (Nb) really required?
when converting the scalar spontaneous emission factor (beta) into a vector for each mode, the scalar is divided by the number of modes to keep the "total amount" of spontaneous emission independent of the number of modes. The spontaneous is distributed homogeneously across the modes (not proportional to the power in each mode). Sensible?
Noise applied to the carrier density through the zeroth carrier term Nw only. Sensible?
- ...