Abstract-This paper will present the modelling, analysis and design of a load-independent Class EF inverter. This inverter is able to maintain zero-voltage switching (ZVS) operation and produce a constant output current for any load value without the need for tuning or replacement of components. The loadindependent feature of this inverter is beneficial when used as the primary coil driver in multi-megahertz high power inductive wireless power transfer (WPT) applications where the distance between the coils and the load are variable. The work here begins with the traditional load-dependent Class EF topology for inversion and then specifies the criteria that are required to be met in order achieve load-independence. The design and development of a 240 W load-independent Class EF inverter to drive the primary coil of a 6.78 MHz WPT system will be discussed and experimental results will be presented to show the load-independence feature when the distance between the coils of the WPT system changes.
A. IntroductionResonant soft-switching converters, such as Class E and Class EF 2 inverters, are commonly used in high power inductive WPT systems that operate at multi-megahertz frequencies due to their efficient operation and simple construction [1], [2]. However, they are only designed to operate at optimum switching conditions for a fixed load and therefore are highly dependent on the load value. They are less tolerant to load variations compared to other inverters, such as Class D inverters, which causes them to become less efficient as the load deviates from its optimum value. Consequently, this limits the WPT system to function efficiently only at a fixed coil separation distance and for a narrow load range.There have been attempts to tune the Class E inverter while in operation to compensate for any variations in the load by using saturable reactors and varactors [3]. While these methods can increase the load range that Class E inverter can tolerate, they require either the user to tune them manually by observing the MOSFET's drain waveform or require a control loop to be implemented. Additionally, the tuning elements can limit the Class E inverter to operate at higher power levels.It was shown in [4], [5] that the Class E and Class φ 2 inverters when used with a finite-DC inductor instead of the usual choke, can be designed such that the they achieve zerovoltage switching (ZVS) and produce a constant output voltage as the load resistance varies. These designs extends the load range at which the Class E and Class φ 2 inverters can operate efficiently from infinite load resistance (open circuit) to a certain minimum load resistance. Although these designs can be Fig. 1. The Class EF inverter applied to several applications, such as high frequency DC/DC converters, they cannot be applied efficiently in inductive WPT application where the distance between the coils changes. This is because the load will range from zero resistance (short circuit) when the coils are completely separated from each other to a certain ma...