To solve such a problem, that when the primary and secondary windings of an output transformer are wound into an intricate multilayered and multi-division sandwich, in order to improve the characteristics of the transformer in high-frequency band, the thickness of the secondary winding cannot be increased due to the thickness relation of the window of a core, and consequently, the power efficiency of the trans former is deteriorated by an increased DC resistance on the secondary side and an increased internal loss.

In a winding method for audio-frequency transformer, the DC resistance of the output transformer is reduced and the high-frequency characteristics of the transformer are improved by using copper wires, having different thicknesses as the secondary winding of the transformer and putting the thinner secondary winding in between the thicker secondary winding and primary winding. When this method is used, the secondary winding having a lower DC resistance can be wound on the inside, outside, or both the inside and outside of the primary winding, resulting in a widened selection width. It is better, in addition, to wind the thinner secondary winding put between the turns of the primary winding in one or a plurality of stages, by taking the thickness of the core and the thickness, high-frequency characteristic, etc., of the primary winding into consideration.