What to Know About Semiconductor Failure and Replacement
Failures of power semiconductors in AC and DC drives can result in significant downtime for repair and recertification of affected equipment. These components are often subjected to substantial loads and stresses. The desire to avoid future issues as part of preventive maintenance has led to some myths and misconceptions about the accuracy of field tests.
Do Semiconductors Fail with Age?
Power semiconductor devices do age and can fail over time in service for a variety of reasons. As the devices age, degradation can include increased leakage current, lower threshold voltage for switching, dielectric breakdown, and loss of gate control. Device thermal cycles can also degrade in both bonding wires within the device and on the thermal junction in isolated base construction. However, conditions short of failure are not detectable by tests using a common multimeter.
Measuring resistance for anode-cathode and gate impedances offer no information because readings can vary widely, even among new electronic devices. When attempting to measure devices still connected in circuit, divergent paths within the circuit can contribute even more misinformation. Isolating the component and measuring resistance across the device is only accurate for readings of open circuit or short circuit, signifying device failure. Other indications may be useful to determine appropriate replacement intervals.
What Causes a Semiconductor to Fail?
Operation well within rated limits for voltage, current, and thermal characteristics allows consistent performance with less degradation over time. More commonly, specific conditions can hasten faults within the silicon until failure occurs. Here are some common conditions that affect semiconductor longevity:
Power cycling of the equipment can put additional and instantaneous stress that can exceed rated specifications. Unless mitigated by inrush effect suppression, these stresses can initiate cracks, voids, or separation of the silicon and heat-dissipating elements.
Negative bias temperature instability (NBTI)
NBTI occurs with constant electric field degradation of the dielectric. Some resulting issues can be a decrease of the threshold voltage and/or slower switching before failure occurs.
Hot carrier injection (HCI)
HCI or charge trapping occurs with electrons that become embedded into the gate material, causing damage and degraded performance.
Time-dependent dioxide breakdown (TDDB)
TDDB is a product of high electric fields applied over time, causing a breakdown of the gate. Catastrophic failure of the device follows.
Thermal transient induced damage
Thermal transient induced damage can occur more commonly with Isolated Base package semiconductors. In these cases, rapid heating and cooling of the device in normal operation can apply mechanical expansion/contraction fatigue and breakage in the device connecting wires. It can also cause the thermal impedance to increase due to material separation in the thermal cooling path to the heat sink and excessive heat in the active elements of the device. Damaged sections may exhibit increased heating of a device within the package as the load-sharing wires fail.
Nearby electronic component or system failure
Nearby electronic component or system failure can affect the operation conditions of the semiconductor. The result can be out of specification voltage, currents, and increased thermal stresses applied to the device. Many systems can affect semiconductor failure, including cooling system, high capacity capacitors, power supply, and control systems faults.
A failure analysis that includes checking for the degraded performance of semiconductors within its circuit and thermal imaging techniques can be more reliable indicators that one or more semiconductors should be replaced.
How to Replace a Semiconductor
Observe the following guidelines for successful installation of either the TO-200 AB (PUK) or Isolated Base module packages:
Keep it clean.
Make sure that the heat sink(s) and the device are clean and flat for installation. Clean the heatsink to a brilliant finish. Pitting in the heatsink surface should be corrected by machining before proceeding to the next step. When clean, act quickly to prevent micro oxidation from forming on either surface.
Apply Heatsink compound sparingly.
Make sure that the surface is covered lightly to present adequate conduction. Excess compound can prevent even surface contact, resulting in poor thermal and electrical conduction.
Align pins (PUK) or align base holes (isolated base modules) and apply to heatsink.
Note proper orientation for electrical connections.
Apply clamping force or fasteners according to the manufacturer’s guidelines.
There is no universal torque standard for this installation.
Make sure that the PUK heatsink faces are parallel by inserting the clamping fasteners the same amount to snug before tightening. The clamping force is determined by the displacement of the spring or a strain gauge indicator. Tighten ½ turn per side until indicators match manufacturer’s guidelines.
Isolated base modules require tightening of their fasteners in an order prescribed by the manufacturer.
Tighten in the prescribed order to the manufacturer’s specification.