Mats Gustavsson
Ericsson
Network
Latest external collaboration on country level. Dive into details by clicking on the dots.
Publication
Featured researches published by Mats Gustavsson.
Journal of Lightwave Technology | 1990
Mats Gustavsson; Anders Karlsson; Lars Thylén
Near-traveling-wave semiconductor laser amplifiers for amplification and detection of optical signals are discussed. Measurements of gain, responsivity, and bandwidth are presented and compared with theory. The system performance of the laser amplifier detector is evaluated by a digital transmission experiment. The importance of using low-reflectivity amplifiers with high-responsivity and weakly wavelength-dependent devices is revealed by computer simulations. The various noise contributions of the laser amplifier detector are analyzed. Expected sensitivity values are given, and it is shown that there exists an optimum amplifier gain with respect to sensitivity. >
IEEE Photonics Technology Letters | 1995
W. van Berlo; Malgorzata Janson; L. Lundgren; A.-C. Morner; J. Terlecki; Mats Gustavsson; P. Granestrand; P. Svensson
Monolithically integrated 4/spl times/4 semiconductor laser amplifier gate switch arrays comprising twenty-four integrated laser amplifiers have been designed, fabricated, and evaluated. Low fiber-to-fiber loss, low polarization dependence, high extinction ratio, and low crosstalk are reported.<<ETX>>
Journal of Lightwave Technology | 1999
Lars Gillner; Claus Popp Larsen; Mats Gustavsson
The scalability of optical multiwavelength switching networks with respect to bit rate, number of wavelength channels per fiber, and number of inputs/outputs per optical node is examined. The limits considered here are those imposed by the combination of interferometric crosstalk, optical noise accumulation, and optical amplifier gain saturation. An analytical model that combines these effects is described. For the class of optical cross-connects examined, the number of first- and second-order crosstalk contributions per node for different node architectures and node sizes is derived. The results show that with a carefully designed optical cross-connect architecture, it should be possible to construct a fairly large meshed all-optical network (more than ten nodes in diameter with 32 inputs/outputs per node and 32 wavelengths per fiber) if the crosstalk of optical space switches, multiplexers, and demultiplexers is about -30 dB.
Journal of Lightwave Technology | 1996
E. Almstrom; C.P. Larsen; L. Gillner; W.H. van Berlo; Mats Gustavsson; E. Berglind
Packaged 4/spl times/4 semiconductor optical amplifier gate switch arrays have been studied experimentally and analytically. Experimentally, a dynamic range of the input power to the switch of more than 20 dB was obtained at 622 Mb/s with less than 2 dB power penalty. Wavelength division multiplexing transmission with four channels at 622 Mb/s each through the same path was successfully demonstrated. The switch was also evaluated in an installed fiber network, where a 2.5 Gb/s signal was transmitted through three cascaded switch routes and 160 km fiber. A numerical time-domain model incorporating an electrical filter in the receiver confirmed the experimental results. It furthermore indicated that residual reflections between the different waveguide sections within the switch and at the chip facets may degrade the performance of the switch.
Journal of Lightwave Technology | 1997
Mats Gustavsson; Lars Gillner; Claus Popp Larsen
General, statistical properties of accumulation of interferometric crosstalk due to a number of interfering signals are analyzed. The probability density function of the total crosstalk is derived as function of the sum of the incoherent crosstalk contributions associated with the interfering signals, Because the total crosstalk level can be expected to vary only slowly as compared to the time duration of one bit, the probability p that the total crosstalk exceeds a maximum tolerable value is introduced as a parameter in the crosstalk and bit error rate analysis of optical networks. General computations show that the crosstalk requirement on optical network elements may be significantly relaxed if it can be accepted that it is possible, although most unlikely, that the total crosstalk may exceed this maximum tolerable value. The statistical crosstalk model is then illustrated by investigation of optical network examples with respect to their crosstalk and transmission properties, and, finally, compared with previous crosstalk models.
Optical Amplifiers and Their Applications (1992), paper PD9 | 1992
Mats Gustavsson; B. Lagerstrom; Lars Thylén; M. Janson; Lena Lundgren; Ann-Caroline Mörner; Michael Rask; Björn Stoltz
Monolithically integrated 4×4 semiconductor laser amplifier gate switch arrays comprising twenty-four integrated laser amplifiers have been fabricated and evaluated. We report observation of net positive optical gain between fibres, high extinction ratio, and low crosstalk.
IEEE Journal of Quantum Electronics | 1989
L. Gillner; E. Goobar; Lars Thylén; Mats Gustavsson
Long-wavelength semiconductor laser amplifiers are investigated with respect to spectral gain properties such as peak gain wavelength shift and width of gain curve, employing different structural parameters such as thickness of the active layer and amplifier length. The model takes into account Auger recombination, thermal effects, and spontaneous emission. It is shown that there exists an optimum thickness of the active layer with respect to current density for a given gain and that increased length of the amplifier allows higher gains and reduced variation of peak gain wavelength with respect to variation of peak gain at the expense of increased saturation by amplified spontaneous emission and increased excess noise. An experimental verification of the theoretical model is reported. >
IEEE Journal of Quantum Electronics | 1993
Mats Gustavsson
Polarization rotation in periodically asymmetric passive and active waveguides is described theoretically. The analysis is based on the vector wave equation and analysis of the coupled mode equations. It is shown that these periodically asymmetric active waveguides can be employed as polarization-independent optical amplifiers, and that strictly polarization-independent amplifiers require a rotation of an integer multiplier of pi radians. An illustrative example of such an amplifier with a buried waveguide structure suitable for monolithic integration is analyzed. Polarization-insensitive passive Bragg filters employing, in addition to the ordinary grating, a periodically asymmetric perturbation are also considered, and polarization rotation in a distributed feedback laser filter is discussed. >
IEEE Journal on Selected Areas in Communications | 1996
Lars Gillner; Mats Gustavsson
The scalability of multiwavelength switching networks with respect to bit rate, number of wavelength channels per fiber, and number of inputs/outputs per optical node is examined. High bit rates will require use of dispersion compensation, a large number of channels per fiber will require low-loss wavelength demultiplexing and multiplexing, and many input/output fibers per node will require corresponding large space switching stages. These requirements implemented at the same time seem feasible from the power budget point of view to give networks with very high data capacity (3000 Tb/s/spl middot/km). The limits considered here are those imposed by optical noise accumulation, static and dynamic gain saturation in the optical amplifiers, and dispersion in the transmission fibers.
IEEE Journal of Quantum Electronics | 1991
Lars Thylén; Mats Gustavsson; T. K. Gustafson; I. Kim; Anders Karlsson
The noise properties of semiconductor laser amplifiers are calculated using the Heisenberg equations of motion. Photon noise as well as the recombination current noise are calculated with this formalism, and agreement with previously published results was found for the photon noise. The formalism also gave the conventional current shot noise in the appropriate limits, e.g., when using the amplifier reverse-biased as a detector. The method gives reasonable agreement with reported experimental results and the results for the current noise could be given a consistent physical explanation. Some concluding remarks concerning the relationship with the Langevin approach and the possibilities of minimizing the noise are given. >