A transformer is one of the most important units in the power networks. Thus a breakdown of a transformer faults can cause costly repairs. To avoid this, its reliable operation has great significance and, therefore, the diagnosis system of the transformer is necessitated. The dissolved gas-in-oil analysis (DGA) is very popular diagnostic method, but is not a proper technique to identify mechanical faults. The transformer has a unique transfer function independent of the shape of the input waveform, which can be evaluated through sweep test. This transfer function changes by faults. In this paper, a new transformer diagnosis algorithm using proposed signal processing and comparison criteria is suggested.