8 Single-Event Transients (SET)

The primary focus of SETs were heavy-ion-induced transient upsets on the output signals HO and LO (with a 1nF capacitive load on the outputs as seen in block diagram). SET testing was done at room temperature across three ion species, ¹⁰⁹Ag, ¹⁴¹Pr, and ¹⁶⁵Ho which produced a range of LET_EFF of 48 to 75MeV × cm² / mg for more details, see Ion LET_EFF, Depth, and Range in Silicon. HO and LO were monitored by two different scopes, a NI PXIe-5110 and a MSO58B oscilloscope. During PWM and IIM_SW mode testing, each scope was configured to trigger based on an outside pulse width measurement, where the window for the output signal was 20% (±200ns). During the IIM_ST modes, the same two scopes were used, however, the trigger was a window which was 500mV above or 500mV below 0V with the signals AC-coupled. The signals in this mode were monitored to see if the signal ever went low when it should have been high, or high when it should have been low. During all SET testing, there was no cross-conduction in either PWM or IIM mode and in IIM mode the only transient that occurred was a high to low transient, no signals ever turned on when not required. For all recorded transients, there were never cases where the pulse deviated to be greater than 20%. The only captured SETs were missed pulses. During the IIM mode testing with LO high and during the PWM mode testing during a LO transient, the signature of the transient shows that there is some overshoot (approximately 410mV during static and approximately 480mV during switching) before leveling back out to 5V during the turn on. This is consistent across all transients, but all signals do recover back to nominal after the overshoot in the order of μs. Because of this BP5L was monitored on the MSO58B to show that this overshoot is from the internal LDO.

Waveform size, sample rate, trigger type, value, and signal for all scopes used is listed in Table 8-1.

Upper and lower bound cross-sections were calculated to 95% confidence. Weibull fit was done to calculate the onset value. The onset was only found for the TPS7H6003-SP device.

A Weibull fit was conducted to determine the best estimation of onset since there were transients at 65MeV, but none at 48MeV. Because of the gap between LET levels there is high probability that the true onset is somewhere between the two tested levels. Based on this fit the estimated true onset for the TPS7H6003-SP is 58MeV.