Electroseismic (ES) exploration remotely identifies the presence of hydrocarbons using the conversion of electromagnetic energy to seismic energy. These conversions are relatively larger in a porous, permeable resistive body (such as an oil or gas reservoir) when compared with background conversions. Typical ES signals, however, may be several orders of magnitude below the expected ambient noise levels, thus requiring repetitions of long source sequences. In addition, the ES signals may vary linearly or nonlinearly with the input current. Two classes of coded waveforms are presented for the detection of these linear and nonlinear conversions. The linear conversions are detected using a Golay sequence pair with side lobes that cancel when the correlated pair is summed. A set of pseudo-random binary sequences (PRBSs) can be modified to detect a nonlinear squared response with minimal side lobes while removing undesired linear conversions. These source waveforms are implemented in the power waveform synthesizer (PWS) that is capable of switching three-phase power to create the specified waveforms. The PWS is an important part of the overall electroseismic exploration system that includes source electrodes, pickup-free receivers, and acquisition and processing features aimed at effective detection of small ES signals.