The influence of pulse separation and instrument input impedance on pulsed IV measurement results

Pulsed current-voltage (IV) characterization has arisen as a method that can he used in many cases to obtain more accurate large-signal device characterization. The accuracy of pulsed IV measurements requires the use of a sufficiently short pulse length and a sufficiently large pulse separation. For a small class of devices, if the pulse separation used is too short, the steady-state condition at the quiescent bias point may not be reached between pulses. In this case, the measurement results depend on the actual steady-state condition achieved, which is not the steady-state condition of the desired quiescent bias point and is dependent on several factors, including instrument input impedance. It is shown in this work that a longer pulse separation must be used in measurements involving devices with significant trapping effects (such as a GaAs MESFET) for quiescent bias points near pinch-off, while a shorter pulse separation is often acceptable for Class A quiescent bias points and for devices with minimal trapping effects (such as a Si MOSFET). The normalized difference unit is used to allow numerical analysis of the IV results.