Partial discharge (PD) is a micro discharge that occurs in defected regions within the insulating media. As these discharges are the main culprits that cause dielectric material aging, PD measurements have been used for assessing insulating materials, including solids, liquids, and gases for power applications. There are various methods and sensors available for measuring PD sensitive to specific characteristics and operable over a wide range of frequencies. Most PD measurement techniques provide patterns that enable PD interpretation more comfortable for users. For example, in AC applications, the phase-resolved partial discharge (PRPD) technique provides identifiable patterns for distinguishing various types of PDs. However, the establishment of meaningful patterns to multiple types of PD in DC systems requires more sensitive and accurate measurements of individual PD pulses with noise rejection functionality due to the lack of phase-resolved information. Investigating of the transient phenomena such as individual PD pulses requires well-designed circuits with sufficiently large bandwidths. Waveshapes can be easily disturbed by background noise and deformed by the frequency response of measuring circuits and data acquisition systems (DAQ). Noises are unwanted disturbances that could be suppressed by suitable filters or mathematical methods. Measurement circuits and DAQ systems consist of transmission lines, sensors, cables, connectors, DAQ hardware, and oscilloscopes. Therefore, matching the impedance of all components guarantees a reflectionree path for traveling signals and addresses most of the challenges relevant to transient measurements. In this dissertation, we proposed and designed an appropriate testbed equipped with high bandwidth transmission line and electromagnetic field sensors suitable for investigating PD under DC electrical stresses. We comprehensively used finite element analysis simulations through the COMSOL Multiphysics software to design the dimensions and evaluate the frequency response of the testbed, transmission line, and electromagnetic sensors. Furthermore, based on the new testbed, DC PD measurements were performed using conventional and non-conventional sensors. Finally, various types of DC PD were statistically classified based on the proposed testbed.