Brian S. Marks
University of Baltimore
Network
Latest external collaboration on country level. Dive into details by clicking on the dots.
Publication
Featured researches published by Brian S. Marks.
IEEE Photonics Technology Letters | 2002
Ivan T. Lima; Gino Biondini; Brian S. Marks; William L. Kath; Curtis R. Menyuk
In this letter, we use importance sampling to analyze polarization-mode dispersion compensators with a constant differential group delay (DGD) element. We optimize the value of the fixed DGD element of the compensator with respect to the outage probability. We show that the optimum value of the fixed DGD element of the compensator can reduce the outage probability by several orders of magnitude, even though it does not provide a substantial reduction of the average penalty due to polarization-mode dispersion in the cases that we studied. By contrast, choosing the fixed DGD element to maximally reduce the average penalty may lead to an outage probability that is orders of magnitude larger than the optimal choice.
optical fiber communication conference | 2003
Aurenice O. Lima; Ivan T. Lima; Brian S. Marks; Curtis R. Menyuk; William L. Kath
We compare the performance of single-section polarization mode dispersion (PMD) compensators with either a fixed- or variable-differential group delay (DGD) element as a function of both first- and second-order PMD using multiple importance sampling. We show that there is little benefit in using a variable-DGD element.
IEEE Photonics Technology Letters | 2007
Yu Sun; P. Hallemeier; H. Ereifej; Oleg V. Sinkin; Brian S. Marks; Curtis R. Menyuk
In this letter, we investigate the penalty after a 10-Gb/s Ethernet signal is detected and processed. We study the statistics of ideal electrical dispersion-compensation (EDC) performance, measured by the penalty of an ideal decision feedback equalizer (PIE-D), of 300-m multimode fiber links to determine the theoretically best possible EDC performance. We calculate the channel responses using two methods: the individual-mode method and the average-mode method. We find a strong dependence of the PIE-D value on the channel model that is used, especially when there is a connector with a large offset in the link. In system design, one should choose a suitable channel model, taking into account the length of the fiber, the magnitude of offset, and the percentage of coverage
Technical Digest: Symposium on Optical Fiber Measurements, 2004. | 2004
Hai Xu; Brian S. Marks; John Zweck; Li Yan; Curtis R. Menyuk; Gary M. Carter
It is well known that the distribution of differential group delay (DGD) in a straight-line optical fiber transmission system is a Maxwellian when the fiber realization drifts ergodically and the fiber is statistically homogeneous [I].
optical fiber communication conference | 2002
Yu Sun; Brian S. Marks; Ivan T. Lima; K. Allen; Gary M. Carter; Curtis R. Menyuk
The experimental and theoretical results show that polarisation dependent loss (PDL) plays a major role in the loop performance. In recirculating loop systems, the PDL will unrealistically improve the performance when the polarization evolution is optimized by using a polarization controller. The noise tends to polarize in the polarization state of the signal, and the noise orthogonal to the signal is reduced. We have characterized the behavior of the polarization state as a spiral evolution on the Poincare sphere and have associated different types of spirals with different system performance.
IEEE Photonics Technology Letters | 2008
Brian S. Marks; Curtis R. Menyuk; Anthony L. Campillo; Frank Bucholtz
It has previously been shown that dispersion compensation can dramatically affect the interchannel crosstalk in a multichannel analog fiber transmission link. In this work, we use a genetic algorithm to find five-segment dispersion maps that yield low crosstalk levels over two octaves of microwave-frequency bandwidth when amplitude modulation is used. The genetic algorithm suggests that optimal dispersion maps have low residual dispersion. Despite the genetic algorithms ability to optimize dispersion maps with many fibers, it is possible to obtain similar crosstalk levels from a simpler two-segment design whose dispersion is fully compensated.
optical fiber communication conference | 2002
Ivan T. Lima; Aurenice O. Lima; Yu Sun; John Zweck; Brian S. Marks; Gary M. Carter; Curtis R. Menyuk
A fundamental problem in the design of optical communication systems is to minimize channel outages due to the polarization effects. System designers commonly allocate a prescribed margin to polarization effects, such as 2 dB, with a certain probability that the margin will be exceeded, such as 10/sup -6/. When this margin is exceeded an outage is said to occur. Because outages are so rare, it has been difficult to obtain them from experiments or from standard Monte Carlo simulations.
lasers and electro-optics society meeting | 2003
J. Hu; Brian S. Marks; Qun Zhang; Curtis R. Menyuk
We describe a shooting algorithm for Raman amplifiers. Performance tests are done to compare the efficiency of two integration methods. Using a Jacobi weight and a continuation method yields more robust convergence of the algorithm.
quantum electronics and laser science conference | 2005
Jonathan Hu; Brian S. Marks; Jinchae Kim; Curtis R. Menyuk
We calculate the evolution of the loss, dispersion, and mode shape in a tapered microstructure optical fiber. We show that the primary limitation on mode compression is the loss due to mode leakage.
lasers and electro-optics society meeting | 2003
Gary M. Carter; Hai Xu; Hua Jiao; Yu Sun; Brian S. Marks; Li Yan
This study demonstrates detailed experiments and modeling on the effects of polarization dependent loss (PDL) and polarization mode dispersion (PMD) in recirculating loops. Experimental results and modeling show that the Q distribution is affected by noise which is partially polarized due to the PDL. Due to polarization dependence of the gain (PDG) in erbium doped finer amplifiers (EDFA), single channel experiments require input scrambling to eliminate this effect. Due to gain saturation in an EDFA chain the input scrambling rate can be chosen to eliminate unwanted amplitude modulation.