The measurement of additional losses in induction motors: Discussion about the actually achievable uncertainty