Abstract
Bioelectrical impedance in vivo measurement errors due to the connection between instrument and patient are analysed theoretically and experimentally. Special attention is paid to the patient cable capacitances and to the reduction of unwanted common-mode signals at the voltage amplifier input. Experiments with single-ended and symmetrical current sources, with different simulated body impedance values, with five types of cable connection and with three types of cable screen driving are carried out. The use of a balanced voltage-to-current transverter with a floating common point is recommended. Four separate patient cables should be used with separately driven screens by high-quality unity-gain buffers. Thus the impedance module measurement error can be reduced to below 1% in the range of 20 to and the phase measurement error to less than , for frequencies from 1 kHz to 300 kHz. Keywords: bioelectrical impedance, impedance spectroscopy, instrumentation, patient-instrument interface, measurement errors