8 Single-Event Transients (SET)

SET are defined as heavy-ion-induced transients upsets on the GATE (OUTH and OUTL tied together), REFCAP, V_LDO, and SS of the TPS7H502X-SP. When conducting testing on the device in flyback configuration the VOUT of the device (the DC signal after the transformer on the EVM) was also monitored.

Testing was performed at room temperature (no external temperature control applied). The heavy-ions species used for the SET testing was ¹⁶⁵Ho at 15 MeV/nucleon and ¹⁶⁹Tm at 19.5 MeV/nucleon (for more details refer to Ion LET_EFF and Range in Silicon). Flux of ≈10⁵ ions×cm²/s and a fluence of ≈10⁷ ions/cm², per run were used for the SET characterization discussed in this chapter.

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

Note¹: Only one Signal was used as a trigger source at a time, this table presents all possible sources for a given scope, the same is valid for the trigger type. All percentage specified on the trigger value are deviation from the nominal value.

Note²: The trigger signal V_OUT is only valid for and was only monitored during the flyback configuration testing on the flyback EVM.

Open-Loop Configuration

The primary focus of SETs were heavy-ion-induced transient upsets on output signal GATE (OUTH and OUTL tied together). SET testing was done at room temperature at ¹⁶⁵Ho which produced a LET_EFF of 75 MeV·cm²/mg. GATE was monitored using a NI PXIe-5110. During testing the scope was set to trigger if the signal exceeded |20%| from nominal using a pulse width trigger. During all SET testing, there was one type of transient recorded that was self-recoverable. The REFCAP, V_LDO, and SS signals monitored on the PXIe-5172 and PXIe-5162 scopes did not have any recorded transients. Because the V_LDO signal had a window trigger on it, being transient free shows that there is no overshoot on the device at 75 MeV·cm²/mg.

The SET results for 2 production and 7 pre-production devices are shown below in the following tables. The transient signature on GATE is shown and the number of transients across the runs, voltages, and frequencies is shown. Since only this transient signature occurred there is high confidence that the TPS7H502X-SP is SEFI free and the recorded transient signature does not show any overshoot indicating that the TPS7H502X-SP is safe for GaN operations. Note that for all testing V_LDO was programmed to be 5V except for the cases where V_IN=P_VIN=4.5V in which case it was programmed to be 4.5V as well.

The upper-bound cross-sections for all bias conditions are shown inFigure 8-1.

Figure 8-1 TPS7H5020-SP Silicon Mode GATE Pulse-Width Transient

Figure 8-2 TPS7H5020-SP Silicon Mode GATE Pulse-Width Deviation Histogram

Figure 8-3 TPS7H5020-SP GaN Mode GATE Pulse-Width Transient

Figure 8-4 TPS7H5020-SP GaN Mode GATE Pulse-Width Deviation Histogram

Figure 8-5 TPS7H5021-SP Silicon Mode GATE Pulse-Width Transient

Figure 8-6 TPS7H5021-SP Silicon Mode GATE Pulse-Width Deviation Histogram

Figure 8-7 TPS7H5021-SP GaN Mode GATE Pulse-Width Transient

Figure 8-8 TPS7H5021-SP GaN Mode GATE Pulse-Width Deviation Histogram

Flyback Configuration

To better understand functionality of the device closed-loop testing was conducted with the device in a flyback configuration. A flyback EVM was designed to allow the device to operate as a flyback converter with a power stage input voltage of 28V and an output voltage of 5V. During this testing TPS7H502X-SP was powered to operate in GaN mode with a nominal voltage of 12V at V_IN, P_VIN tied to V_LDO at 5V, and loaded to a 4A load on the output. The load was provided by a Chroma E-Load in constant resistance mode with a resistance of 1.25Ω. During the testing the GATE (OUTH and OUTL tied together before the transformer) signal and V_OUT (the DC output signal after the transformer) were monitored by a PXIe-5110 and PXIe-5172 respectively. Testing was done at room temperature at ¹⁶⁵Ho which produced a LET_EFF of 75MeV·cm²/mg. During testing the PXIe-5110 scope was set to trigger if the signal exceeded |20%| from nominal using a pulse width trigger. For V_OUT the PXIe-5172 was set to trigger on a |3%| window. During all SET testing, there were three types of transients recorded on GATE that were all self-recoverable. The first was a long transient only seen during the beginning of the run when the beam shutter opened. The second was a greater than +20% pulse-width transient. The third was a less than -20% pulse-width transient. There were no transients observed on V_OUT during this testing indicating that the DC output of the flyback is transient free at the specified operating conditions at 75MeV·cm²/mg.

Figure 8-9 GATE transient during beam shutter open

Figure 8-10 Typical Gate Transient >+20% Pulse-Width Deviation

Figure 8-11 Typical Gate Transient < -20% Pulse-Width Deviation