Changes in a certain parameter are often a few magnitudes smaller than the base value of the parameter, specifying significant requirements for the dynamic range and noise levels of the measurement system. In case of electrical bioimpedance acquisition, the variations can be 1000 times smaller than the entire measured value. Synchronous or lock-in measurement of these variations is discussed in the current paper, and novel measurement solutions are presented. Proposed methods are simple and robust when compared to other applicable solutions. A common feature shared by all members of the group of the proposed solutions is differentiation. It is achieved by calculating the differences between synchronously acquired consecutive samples, with lock-in integration and analog differentiation. All these methods enable inherent separation of variations from the static component of the signal. The variable component of the bioimpedance can, thus, be acquired using the full available dynamic range of the apparatus for its detection. Additive disturbing signals and omnipresent wideband noise are considered and the method for their reduction is proposed.

The paper investigates the impacts of the selected electrical equivalent circuit model, measurement setup, and surrounding environment on the trustworthiness of electrical bioimpedance measurement and obtained model data in the human body. The influence of these constitutive components of the system on finding the model parameters is analyzed and illustrated with examples. The results based on experimental measurements on a forearm near the wrist are provided by employing the model, measurement setup, and novel 16-bit compact wireless impedance analyzer (CIA) according to the outcome of the analysis. The area near the wrist is of interest because of attempts to get cardiac-activity-related impedance changes. It is concluded that a two-electrode system with voltage excitation suits better for determining bioimpedance model parameters in the β dispersion area. The results obtained with the CIA and two capacitive bracelet electrodes on a left forearm were used for the fitting model parameters. Despite the small dimensions of 60 × 60 × 25 mm of the CIA reducing stray capacitance to 8 pF, it provides relative impedance magnitude measurement error below 0.3% and phase error below 0.2 ° in the 10 MHz range. Analysis of the model parameters allowed separation of the electrodes, skin, and internal tissue spectra and revealed the relative significance of model components at different frequencies.