Rolling element bearings are used at the heart of almost all machines. Recently, the electrical behavior of bearings has gained more interest of researchers. Especially bearing impedance models are of interest for multiple reasons. Firstly, measuring the bearing impedance allows one to draw conclusions about the load acting on the bearing. This enables a sensory utilization of the bearing, which is valuable for Industry 4.0. Secondly, the impedance provides insights into the lubrication film. It depends on the film thickness and changes its behavior in mixed friction. That enables the detection of poor lubrication conditions or early signs of fatigue, offering valuable insights for novel predictive maintenance approaches. Thirdly, bearings in electrical machines are subjected to potentially harmful bearing currents. In turn, the electrical impedance influences the voltage distribution in the system. In that way, all mentioned points drive the need for improved bearing impedance models. Multiple research institutes are involved in this area. Yet, efforts are barely streamlined because the existing models can hardly be compared to each other. Thus, this work aims to develop a reference capacitance calculation method for ball bearings that is easy to employ and widely applicable, while representing and expanding the current state of research. The model was developed in four steps. a) Investigating the capacitance of unloaded rolling areas. It is found that a semi-analytic approach is sufficient according to the requirements, provided it is based on the true geometry, including the rim outside the groove, rather than equivalent radii. b) Estimating the position of unloaded rolling elements. This is achieved by implementing a load distribution model that considers both hydrodynamic and centrifugal forces on the rolling elements, particularly those without load. c) An impedance measurement method was developed, allowing a sufficient sampling rate for the aforementioned experiments. It utilizes the Wheatstone measurement bridge principle with alternating current and reference capacitances. Results yield good performance even in high-noise environments and a linear frequency response within a sufficient bandwidth, meeting the requirements. d) Aggregation of all previous outcomes in a capacitance calculation model. It is capable of considering combined loads, deep groove and angular contact ball bearings of various bearing sizes, as well as arbitrary parameter sets of rotational speed and temperature. Half of the 8108 valid test points showed a relative deviation of less than 20%, surpassing most models in the state of research. A model adequacy assessment revealed that the model is suitable for serving as a reference model in the future, meeting the derived requirements. To improve the model's precision, various fields for future research are suggested.