PPPS-2013: Electric field sensors effect on pulsed power measurements

Summary form only given. A coaxial pulse generator was designed and constructed for an experimental study on a sub-millimeter spark-gap, where the characteristic impedance of the Pulse Forming Line (PFL) and the Transmission Line (TL) is Zc = 100Ω the spark-gap is located into a pressurized chamber between PFL and TL, and the transmission line generator ends at a 100Ω resistance (LOAD).resistance (LOAD). D-dot sensors, used to register the waveforms in both the PFL and TL, are not located exactly on the spark-gap; instead, they are laid 40 mm from the pressurized chamber along both the PFL and TL. To determine the effect of sensor position on voltage measurements, simulations using the EMTP-ATP program were carried out. The PFL voltage (V1) and TL voltage (V2) recorded in different places along the lines are analyzed by using distributed parameter models of the corresponding coaxial transmission lines and also including the electric-arc nonlinear model in the spark-gap switch. Additionally, the distributed parameter circuit models representing the effect of the dielectric materials were included. A MODEL of the spark-gap channel resistance was included. In the MODEL, both the resistive phase and the inductive phase of the gas discharge channel proposedby Martin1 were implemented. The difference between the signal recorded just after the spark-gap (V2) and the one recorded some tens of millimeters forward (V2') is the time delay due to the displacement of the measurement point. Additionally, when comparing the voltage signal recorded just before the spark-gap (V1) with the signal recorded some tens of millimeters backward (V1'), two differences can be observed. The former is the delay in the onset of the two signals. The second difference is a variation in the waveform of the signals. A hypothesis was formulated claiming that this is because the sensor does not record all the charge moving along the transmission line. To confirm this hypothesis, both the charge stored up to the D-dot sensor location and also the charge stored up until the spark-gap were calculated on the basis of EMTP-ATP simulations. The difference between these charges (21.63 nC) was compared with the charge stored in the segment of the transmission line (21.99 nC). The results of these calculations show a difference of 1.6 %, thus the hypothesis was validated.