Many digital delta-sigma radio frequency (RF) upconverters exhibit poor performance when the carrier frequency is tuned away from the centre of the band. The offsetting of signals from the nominal carrier frequency position generates unwanted distortion products which are dominated by the image and third harmonics. It is shown that these products can be moved out of the band by the correct choice of intermediate frequency. A novel technique for shifting both the signal and the noise null to the new frequency position is described.

No abstract

No abstract

This paper considers the realization of fiber-wireless (Fi-Wi) networks using the gigabits passive optical network (GPON) and the infrastructure-based wireless local area network (WLAN). The bottleneck of such hybrid system is the WLAN where interference limits the performance. As such, we focus on enhancing the WLAN performance by analytically deriving the optimum contention window (CW) sizes of access points (APs) and wireless users (WUs), respectively. An adjustable transmission priority factor is introduced to allow uplink-downlink transmission fairness. Further, an adaptive backoff technique using information from monitoring the GPON and WLAN networks is proposed. Simulations show the CW sizes of all WUs are maintained within a standard deviation of 1.5% at a cost of a 3% loss in throughput due to the effect of convergence and other estimates.

Idle Sense (IS) is a simple yet efficient access method to provide optimized throughput and fairness in wireless local area networks (WLAN). Each station dynamically controls its contention window (CW) size by monitoring the mean number of idle slots between transmission attempts. In general this quantity is subject to estimation error that is related to the number of transmissions, M , over which it is measured. Previous works in IS assumed all stations use similar CW sizes to contend the channel; an assumption that is not always valid in reality. Therefore, this paper studies the behaviour of the scheme when stations have different CW sizes. An analysis shows that bias caused by the additive increase multiplicative decrease (AIMD) algorithm used to control the CW sizes, combined with varying lengths of M can cause a fairness problem. Two classes of stations develop, with the first class gaining most of the channel bandwidth while the second class stays idle.