368 IEEE TRANSACTIONS ON ELECTROMAGNETIC COMPATIBILITY, VOL. 46, NO. 3, AUGUST 2004 Predicting the Breakdown Behavior of Microcontrollers Under EMP/UWB Impact Using a Statistical Analysis Michael Camp, Student Member, IEEE, Hendrik Gerth, Heyno Garbe, Senior Member, IEEE, and Helmut Haase Abstract—Reproducible prediction of damaging effects is one of the main problems in intentional electromagnetic interference (IEMI). In this paper, the susceptibility of different types of single microcontrollers to unipolar fast rise time pulses [electromagnetic pulse (EMP), ultrawide band (UWB)] is determined. Therefore, pulses with rise times as fast as 100 ps and electric field amplitudes of up to 100 kV/m are applied to the devices. The results are gener- alized with a novel statistical procedure. Following discussion and rationale, the Weibull distribution is selected to describe the inter- ference behavior. The statistical analysis provides a new test pro- cedure for a confident determination of the interference behavior parameters. Index Terms—Electromagnetic field threat, intentional electro- magnetic interference (IEMI), susceptibility of electronics. I. INTRODUCTION T ODAY, an undisturbed operation of electronic equipment is of vital importance for the function of traffic systems (airplanes, traffic guidance), security systems, and communi- cations. A malfunction in one of these areas could cause personal, ecological, and economical disasters. Therefore, the susceptibility of electronics to pulsed electromagnetic fields like electromagnetic pulse (EMP) and ultrawide band (UWB) pulses is of great interest. The goal of this investigation is to measure the susceptibility of three different microcontroller types in CMOS technology (representative of current micro- controller designs) to a transient electromagnetic field threat. The influence of different circuit parameters on the breakdown effects is investigated and determined. Other parameters (e.g., angle of incidence) will be held constant in this investigation. Statistical analyses are performed and different options to de- scribe and predict the breakdown effects are discussed. Using a novel statistical test procedure, breakdown probabilities as a function of pulse amplitude are determined. Knowledge of this functional relationship allows the prediction of breakdown effects for different microcontroller circuits. II. GENERAL MEASUREMENT SETUP AND PARAMETERS The measurements are accomplished using TEM waveguides (IEC-61 000-4-20, [7]). The applied pulse shape is in general Manuscript received July 14, 2003; revised March 1, 2004. This work was supported in part by the German Federal Office of Defense Technology and Procurement under Contract E/E590/Y5140. The authors are with the Institute for Electrical Engineering and Measure- ment Science, University of Hannover, D-30167 Hannover, Germany (e-mail: camp@ieee.org). Digital Object Identifier 10.1109/TEMC.2004.831816 Fig. 1. Pulse shape and definitions. TABLE I PULSE SIGNAL CHARACTERISTICS a double exponential as shown in Fig. 1. Five different pulse- generating devices are available. 1 Table I shows the rise time and the full-width half-max time of the different pulse signals and their respective parameters. III. DEFINITIONS A. Failure Rates To describe the different failure effects, two quantities have been defined in earlier works [5]. The breakdown failure rate (BFR), (also known as upset rate or latchup rate) (1) is defined as the number of breakdowns of a system, divided by the number of pulses applied to the system. A breakdown means that the system is unable to perform the function for which it was designed, but without any physical damage. After a reset (self, external, or power-on reset) the system will return to its usual function. 1 Kentech PBG3 (1), Northstar (2), NP20 (3), WIS-Marx (4), and PAWS4000 (5). 0018-9375/04$20.00 © 2004 IEEE