Future automotive engines have to comply with upcoming emission legislations (EURO 6, CARB 2) raising the requirements on onboard diagnostic systems (OBD). Faulty conditions of the engine leading to higher emissions must be detected with rising accuracy. Additionally, car manufacturers have a strong interest in improving the reliability of fault diagnosis in their workshops in the sense of being able to find the smallest changeable part. The legislation requirements can be reached using the present methodology, as has been shown in first series applications. But advanced methods of model-based fault detection and isolation can help to accomplish with future requirements as well as to extend the present OBD systems, especially with the ability of detecting small faults and the ability of a root cause isolation. This contribution describes a new approach to detect and isolate typical faults in the air path of common rail Diesel engines with high as well as low pressure exhaust gas recirculation (EGR). The presented fault detection system aims at bringing advanced model-based fault detection methodology to the ECU. The fault detection and isolation capabilities of the system are investigated with the help of measurement data acquired from a dynamic engine test bench.This paper presents results from a research corporation between the Institute of Automatic Control at TU Darmstadt and GM Powertrain Europe.