February_EDFA_Digital

edfas.org 25 ELECTRONIC DEVICE FAILURE ANALYSIS | VOLUME 23 NO. 1 damage to the component. This is especially impor- tant if the suspected component has a small package and a relatively large thermal resistance from junc- tion to case. 2. It is preferred to use a golden sample for com- parison to have confidence in identifying the faulty component(s). Having a golden sample eliminates the need to force potentially failed components to temperatures up to 60-70°C. 3. Some components generate very little heat, for example: transient voltage suppressors (TVS), metal- oxide varistors (MOV), capacitors, etc. The target 40°C case temperature while in a 25°C environment should give enough confidence to judge them as abnormal. 4. When injection voltage is very low but with big current, this could indicate a short with a small resistance. A small temperature rise on a component of 5 to 10°C could indicate that the component is abnormal. 5. When there’s high power injection but there are no component(s) with obvious elevated temperature, the PCBA may contain power components with heatsinks or large packages, which have good heat dissipation. In this situation, a temperature rise of 5 to 10°C on the component can indicate unusual behavior. CASE STUDY Described here is an example of a power board failure analysis with no schematic or layout drawing; however, the power rails and their voltages are identified on the PCBA (Fig. 3). The 24 V power input is converted to 12 V by an isolated SMPS, then 12 V is stepped down to 5 V, 3.3 V, 2.5 V, and 1.8 V. Visual inspectionof the PCBA showed no signs of component damage. After applying 24 V input power, all output power rails were measured to be 0 V. Amultimeter was used tomeasure the resistance of power rails and the 12 V power rail was found to be 16 Ω, which was not a short but could be considered to be leaky and abnormal. On the 12 V power rail, there were many components, for example, capacitors C137, C138, C140, C141, and C142. Wires were soldered on C142 to inject current. The injection process was followed to slowly increase voltage. When the voltage increased to 5 V, we found that C137 was hot, about 37°C (Fig. 4). This elevated temperature above lab ambient was considered abnormal for a multi- layer ceramic capacitor and it gave enough confidence to assume C137 has an issue. After C137 was removed from the circuit board, the 12 V power rail was measured again by multimeter to confirm the change. After applying a 24 V input power to the circuit board, all power rails werewithin specification. The IR-based direct current injection method found that C137 was the only damaged component. Fig. 3 Example PCBA and power supply topology. Fig. 4 Location to inject current and IR image result indicating C137 has elevated temperature. (continued on page 28)