Search

A new alternative is described for the clock-distribution problem. The main feature of the new scheme is that a number of stations arbitrarily located on a single reference line independently derive a clock that is aligned to a common systemwide phase. The central principle is that the midtime between appearances of an isolated pulse traveling down and back on a reference line is the same, regardless of the point of observation. This principle is adapted for clock distribution by the periodic injection of a reference pulse onto a single conductor, reflection of this pulse at the end of the reference line, and generation of a local clock at all stations (such clock being phase-locked to the interval midtime by a special PLL circuit). The paper describes the method and gives a theoretical performance analysis and measured results from an experimental system. The experiments achieve a subnanosecond skew over 30 m on a coaxial cable that has an uncompensated delay of 147 ns. This method should be considered for applications from the millimeter scale on chip to the kilometer scale in radio astronomy antenna arrays. Possible advantages include reduced wiring, reduced tuning, reduced EMI, and/or power consumption and easier in-service growth, relative to conventional methods for the same performance objectives.

An adaptive equalization technique using forward error control (FEC) coding is described and evaluated. Computer simulation is used to compare the FEC-assisted adaptive equalization to conventional adaptive equalization. The system model employed is based on the proposed digital cellular communication standard for North America. The FEC-assisted adaptive equalization technique performs better than the conventional equalization method at low and moderate Doppler frequencies with the same level of redundancy in the transmitted data. For systems employing both FEC coding and adaptive equalization, the FEC-assisted adaptive equalization method successfully combines these two functions while using the redundancy required for only one of them. In other words, training required for adaptive equalization can be provided at no extra cost in terms of data redundancy in a system using FEC coding and vice versa.

Dual-mode cellular systems based on the EIA/TIA IS-54 standard offer the eventual prospect of carrying up to six digital calls in the same bandwidth as a single analog call [1]. During the transition from analog to digital service, however, the call carrying capacity of such systems will be limited by the presence of existing analog users. In this situation, it is reasonable to ask if there are call-handling strategies that could increase the total traffic carried by providing preferential treatment to digital users. We consider four such strategies for maximizing the total traffic carried by a dual-mode cellular system. For two of these strategies, including the baseline 'no-control' strategy, we develop closed-form solutions for carried traffic and other related service statistics. The closed-form solution for the no-control case is then extended to provide a tight upper bound on carried traffic for any control strategy. We also present a method for finding the optimal control strategy by applying linear programming (LP) techniques. The strategies are compared for various proportions of analog and digital users and offered traffic levels. The findings show that it is actually quite difficult to obtain gains using strategies that exploit the difference in spectral efficiency between analog and digital calls, even with formally optimal strategies. While this is an unexpected finding, we feel the conclusion has been well validated and is now understood and explained in the paper. Index Terms--Land mobile radio cellular system, linear programming, queuing analysis, traffic control (communication).