Introduction: We introduce an extension of the PhysioNet Brno University of Technology Smartphone PPG Database (BUT PPG), which is suitable for health monitoring and diagnostics applications. This significant extension enables the efficient use of the database for the development of advanced algorithms including AI-based methods. Methods: BUT PPG is extended by PPG, ECG, and ACC signals recorded simultaneously in healthy subjects. Three-channel (RGB) PPG signals were measured in a telephoning position from the ear using the front smartphone camera and from the finger by the rear camera. One-lead ECG signals and 3-axis ACC data were acquired by the Bittium Faros 180 mobile device. The signals were measured either at rest or under various conditions causing artifacts: higher finger/ear pressure on the camera, motion of finger/ear on the camera, walking, coughing, laughing, changing of ambient light, and talking. PPG signals were annotated in terms of quality into two classes (good/poor) by three experts and their consensus was provided. In each ECG signal, QRS complexes were detected and manually verified. Results: The presented extension of the BUT PPG database includes 3,894 new 10s PPG signals from 38 subjects (18 male, 20 female, age 19-76). In the extension of the BUT PPG database, 24% of signals are of good quality. Signals from the ear (n=1,962) are of lower quality (11% of good quality signals) than signals from the finger (n=1,932) where 37% of signals are of good quality. The dataset includes 50,361 manually verified heartbeat positions. Conclusion: The complete BUT PPG database includes 3,942 signals from 50 subjects. The database is primarily intended for the development of algorithms for the assessment of smartphone PPG signal quality and estimation of heart rate from PPG.