Heart failure significantly impacts morbidity, mortality, physical functionality, and quality of life, presenting a high financial burden to healthcare systems. Early detection through non-invasive methods like ballistocardiography (BCG) and seismocardiography (SCG) can help monitor disease progression. Within our research, we developed an accelerometer-based sensor system for recording BCG and SCG signals and conducted a feasibility study with ten chronic heart failure patients. For each patient, seven triaxial accelerometers and a reference ECG recorded data during rest, a six-minute walk test, and recovery phases. Signal preprocessing involved calibration and band-pass filtering to obtain interpretable data. The resulting signals show recognizable SCG and BCG patterns. Future work will test the applicability of existing heart failure assessment algorithms to these new signals and investigate differences according to sensor position, physical exertion and NYHA classification.