We have measured stepwise-approximated sinewaves generated by a Programmable Josephson Voltage Standard with several different output configurations. These data are analyzed to characterize the dominant error mechanisms for rms applications, such as ac-dc difference measurements of a thermal voltage converter (TVC). We discuss the impact of the dominant systematic errors on the overall uncertainty and consider the implications of providing the load current required by low input impedance devices such as TVC's.