The article is fundamentally flawed as the author measures the 110/220VAC side instead of the current on the 5V rail - the used power meter has 1.5% of tolerance, as well as that the switch power supply's RF noise is capable of corrupting the power meter's calibration!