Through-the-Earth (TE) propagation subjects a radio signal to considerable attenuation. The uplink signal must, however, be detected in the presence of full-strength atmospheric noise. Consequently, through-the-Earth communication systems are generally restricted to relatively low data rates. The electromagnetic characteristics of the TE signal differ from those of the surface-propagating noise. In addition, much of the energy in low-frequency noise is contained in discrete impulses. This paper describes an integrated combination of adaptive noise cancellation (ANC), maximum-likelihood detection (MLD), nonlinear processing (NLP), and decision-aided feedback (DAFB) that can be used to combat atmospheric noise in TEC systems. Tests with real VLF noise demonstrate a 10-to-24 dB improvement in SNR. The reduced effective noise level allows higher data rates, greater depth of operation, smaller transmitters, and/or smaller antennas.

The outphasing technique (LINC) combines two nonlinear RF power amplifiers into a linear RF power-amplifier (PA) system. The two PA's are driven with signals of different phases, and the phases are controlled so that the addition of the PA outputs produces a system output of the desired amplitude. However, the resultant time-varying impedances presented to the PA's alter their dc power consumption and efficiency. Power and efficiency characteristics are derived for both simple (transformer-coupler) and Chireix (transmission-line-coupler with shunt reactance) outphasing systems using saturated class-B PA's. Simple outphasing systems have the efficiency characteristic of a linear class-B PA. Through proper selection of the shunt reactance, the efficiency of a Chireix outphasing system can be maximized at a specific output amplitude. The average efficiency with various amplitude-modulated signals is determined as a function of shunt reactance. Selecting the shunt reactance to fit the signal can improve efficiency by as much as a factor of 2.

The Field-Lewinstein-Modestino model represents ELF noise as the sum of Gaussian and impulsive components. This model has proven to be accurate and is especially convenient for analysis and simulation. However, the parameters of the model are bandwidth dependent and have been measured for only a few specific bandwidths. No method of calculating the parameters for other bandwidths has been presented. This paper derives such a method by equating the second and fourth moments of the impulsive component of the ELF noise at the input and output of the filter.

The width of a train of square pulses can be varied to produce a modulated carrier at the pulse repetition frequency. When the pulse train is generated by switching (class D) transistors, highefficiency operation is possible. The efficiency of this type of amplifier can be significantly higher than that of conventional pulsewidth modulation amplifiers, since the switching rate is reduced. In addition, the spectrum of a bipolar pulse train so modulated has the highly desirable property of all spurious products being band limited near the odd harmonics of the carrier.