edfas.org 33 ELECTRONIC DEVICE FAILURE ANALYSIS | VOLUME 28 NO. 2 damage without significant heat transfer to the surrounding package. The industry relies on three primary component-level models (Fig. 2) to simulate these threats: 1. Charged-device model (CDM): Simulates discharges from automated equipment (like ATE handlers). It is the fastest event, with rise times around 250 ps and oscillatory waveforms up to 5 GHz. 2. Human-body model (HBM): Simulates a person discharging into a device through a 1.5 kΩ resistor and 100 pF capacitor, characterized by rise times of 2.0 to 10.0 ns. 3. Machine model (MM): Simulates abrupt discharges from tools, featuring oscillatory currents with sinusoidal peaks. However, component-level models often fail to capture the severity of field failures. System-level ESD testing (IEC 61000-4-2) evaluates a product in its poweredon state, simulating a discharge into the full system. Technically, this is far more severe than HBM; the peak current is approximately 3.75 A/kV (vs. 0.7 A/kV for HBM), with sharper rise times (0.6 to 1 ns). Because these tests occur while the device is booted, they account for onboard ESD protection components often neglected in component-level testing. CASE STUDY: USB2 SYSTEM-LEVEL FAILURE To illustrate the application of advanced FAFI methodologies, consider a case involving an Intel H310 chipset that suffered functional loss following a system-level ESD Fig. 1 Root causes leading to EOS.[1] Fig. 2 CDM, MM, and HBM ESD current vs. time.
RkJQdWJsaXNoZXIy MTYyMzk3NQ==