A Programmatic Method For Selecting Transistors For High-Frequency Class-E Amplifiers

Radiative wireless power transfer at high efficiency can be achieved with Class-D or Class-E amplifiers. These amplifiers have achievable efficiencies that can be very high but heavily depend on transistor selection. The objective of this work was to partially automate the selection process to quickly generate near-optimal Class-E amplifier designs entirely within MATLAB
^®
. The candidate transistor SPICE models were parsed and converted into a system of first-order ordinary differential equations. A non-stiff numerical solver was used to find steady-state solutions for candidate circuits. A simplex search algorithm was used to find the maximum-efficiency candidate circuit for the transistor. The results were also simulated in LTspice
^®
to confirm accuracy. Eighty-four transistors were compared and seven were found to have high efficiency capabilities at 2.45 GHz. The best performing device achieved 60.6% power-added efficiency in MATLAB
^®
, and the same device achieved 53.0% power-added efficiency in LTSpice
^®
. The incongruency between this work and LTSpice
^®
was small enough to justify its use for eliminating large numbers of candidate transistors.