Articles | Volume 2
https://doi.org/10.5194/ars-2-71-2004
https://doi.org/10.5194/ars-2-71-2004
27 May 2005
27 May 2005

Measuring the upset of CMOS and TTL due to HPM-signals

N. Esser and B. Smailus

Abstract. To measure the performance of electronic components when stressed by High Power Microwave signals a setup was designed and tested which allows a well-defined voltage signal to enter the component during normal operation, and to discriminate its effect on the component.

The microwave signal is fed to the outside conductor of a coaxial cable and couples into the inner signal line connected to the device under test (DUT). The disturbing HF-signal is transferred almost independent from frequency to maintain the pulse shape in the time domain. The configuration designed to perform a TEM-coupling within a 50 Ohm system prevents the secondary system from feeding back to the primary system and, due to the geometrical parameters chosen, the coupling efficiency is as high as 50–90%. Linear dimensions and terminations applied allow for pulses up to a width of 12ns and up to a voltage level of 4–5 kV on the outside conductor. These pulse parameters proved to be sufficient to upset the DUTs tested so far.

In more than 400 measurements a rectangular pulse of increasing voltage level was applied to different types of CMOS and TTL until the individual DUT was damaged. As well the pulse width (3, 6 or 12 ns) and its polarity were varied in single-shot or repetitive-shot experiments (500 shots per voltage at a repetition rate of 3 Hz). The state of the DUT was continuously monitored by measuring both the current of the DUT circuit and that of the oscillator providing the operating signal for the DUT.

The results show a very good reproducibility within a set of identical samples, remarkable differences between manufacturers and lower thresholds for repetitive testing, which indicates a memory effect of the DUT to exist for voltage levels significantly below the single-shot threshold.