This dataset contains displacement measurements from an instrumented rotor obtained in experimental tests with controlled introduction of a shaft crack and discrete levels of unbalance. The dataset is intended to provide dynamic response time series under healthy and damaged conditions, enabling studies on diagnostics, system identification, data-driven modeling, and uncertainty quantification in rotating machinery. The data were first used in the doctoral thesis of Olympio Belli, entitled Reliable Data Driven Models Applied to Rotor Machines. Fault Configurations and Reference Conditions The experiment was designed with two main factors: Crack severity (depth): 0 mm, 1 mm, and 3 mm. The 0 mm condition corresponds to a healthy shaft. Unbalance (added mass): 0 g, 2 g, and 3 g. The 0 g condition corresponds to a balanced rotor. The combination of these levels yields 9 discrete operating conditions (a 3×3 grid), spanning from the healthy and balanced reference state to cases with higher crack and unbalance severities. The dataset corresponds to a sequence of run-ups with rotational frequency swept from 15 Hz to 45 Hz. The rotor uses a shaft with 12 mm diameter and 800 mm length. The system critical speed is experimentally observed at approximately 27 Hz, so the run-up protocol crosses the resonance region, which is essential to highlight dynamic changes associated with crack and unbalance. For each condition defined by the pair (crack depth, unbalance mass), the dataset provides: - Measured displacement series over the run-up (15 Hz to 45 Hz); - Multiple experimental realizations, as indicated by the “sequence of run-ups”, supporting statistical analyses and repeatability assessment (without assuming a specific number of runs). This dataset is well suited for fault classification, severity regression, anomaly detection, feature extraction around the critical speed, and validation of both physics-based and data-driven models under controlled parametric variation. Legend: p1 – 1° Bearing / p2 – Disc / p3 – 2° Bearing