Cyber-physical systems (CPS) are increasingly used in manufacturing, transportation, health, and other industries. To develop these complex interdisciplinary systems, highly qualified CPS engineers are required who possess sound engineering knowledge and excellent transferable skills. Academic institutions offer a range of modules and curricula to teach CPS engineering. However, the literature reports a gap between expectations of industry and competencies of CPS graduates. To close this gap, this paper introduces and describes a holistic educational framework (T-CHAT) for teaching CPS engineering at the module level. To evaluate this framework, two use cases were analysed by conducting self-perception surveys and semi-structured interviews with students. Descriptive statistics and "t"-tests were calculated for the survey data. Interviews were coded and analysed using a General Inductive Approach. The analysis results were discussed by the comparison of the T-CHAT implementations in these two use cases.