A non-intrusive fault detection technology for real time condition assessment of pipelines is highly desirable for pipeline operators and the use of artificially induced fluid transients for this purpose has been the topic of many studies in the past. Fluid transients are stress waves in the fluid that can propagate through pipelines at high speeds and can collect information on the pipe condition during its travel. Whilst many transient-based fault detection methods have been proposed in the literature, little work was done on the characteristics of the transient signal that is best suited for fault detection. Such a study is critical for the development of this technology as the generation of controlled pressure waves is a challenging key step of the procedure. This paper presents analytical, numerical and experimental results to illustrate the effect of signal bandwidth (frequency content) of the induced transient signals has on the spatial resolution, accuracy and range of transient based fault detection. The results demonstrate that higher bandwidth signals provide more accurate fault detection at the expense of detection range.