Two widely used electrical tomography systems, the Sheffield Mark I and the DAS-01P, were quantitatively evaluated and compared to the newly developed Goe-MF system. The performance was quantified using a hardware phantom which closely matches the real input and transfer impedances of the human thorax and allows measurements equivalent to different states of lung inflation. Our results demonstrate that adequate averaging is necessary for noise reduction for the Mark I and especially for the DAS-01P system to get meaningful results even in visualizing maximal respiratory manoeuvres. The Goe-MF system showed a notably improved signal-to-noise ratio which allows also dynamic measurements at low levels of lung volume changes, e.g., in intensive care lung injury patients.