Validating the power loss model of a transformer by measurement: the price to pay

Two studies related to losses in ferrite core transformers are presented. They both aim to check the accuracy of transformer equivalent circuits. First, a 3 kW-two winding transformer is used in a converter which works at 25 kHz. In order to compare the losses forecasted using electronic circuit simulation together with the transformer electrical model, to those deduced from oscilloscopic electrical power measurements a lot of experimental precautions have been taken. Because the transformer have a power efficiency of 92%, knowing the lost part of the power within 10% needs input and output powers to be acquired within 0.4%. Second, to validate the copper losses electrical representation of a smaller transformer, calorimetric measurements have been carried out. Although in both cases high performance equipment is required, the most expensive is the time required to control it. As long as measurement is not considered as an unavoidable subject of investigation, no reliable result can be expected.