IPM Internal Fault Diagnosis: From Overcurrent/Overheat to Predictive Maintenance

In modern power electronics, the Intelligent Power Module (IPM) has evolved far beyond a simple block of IGBTs. It has become the nerve center of systems like variable frequency drives (VFDs), servo drives, and EV inverters, integrating not just power switching but also critical control and protection functions. For an engineer, this integration simplifies design, but its true value lies in the sophisticated internal fault diagnosis capabilities. Understanding these diagnostic features is the key to building robust, reliable systems and transitioning from reactive repairs to proactive, predictive maintenance.

This article provides an in-depth engineering perspective on the internal fault diagnostics of an IPM, covering the fundamental protection mechanisms, how to interpret fault signals, and how to leverage this intelligence for next-generation system health management.

The Core of IPM Intelligence: How Faults Are Detected

An IPM’s “intelligence” stems from its integrated gate driver and control IC, which constantly monitors the state of the internal power devices (IGBTs). This IC acts as a vigilant supervisor, using a network of internal sensors to detect abnormal conditions before they lead to catastrophic failure. The primary mechanisms include:

• Current Sensing: Many IPMs incorporate either a shunt resistor in the emitter path of the low-side IGBTs or use sense-emitter IGBTs. These methods provide a scaled-down, real-time measurement of the current flowing through the power stage, allowing the control IC to detect overcurrent (OC) and short-circuit (SC) events with incredible speed.
• Temperature Sensing: A strategically placed Negative Temperature Coefficient (NTC) thermistor on the module’s substrate provides direct feedback on the thermal state near the IGBT chips. This is far more accurate than an external sensor on the heatsink for detecting rapid temperature spikes.
• Voltage Monitoring: The control IC monitors its own supply voltage (Vcc). If this voltage drops below a safe operational threshold, the Under-Voltage Lockout (UVLO) protection is triggered to prevent erratic or damaging behavior from inadequately driven IGBTs.

When any of these monitored parameters exceed predefined safety limits, the IPM’s internal logic immediately takes protective action and communicates the event to the master controller, typically via a single “FAULT” output pin. For more on the evolution of these features, see our guide on the evolution of IPM intelligence.

Decoding the Faults: A Practical Guide to IPM Protection Functions

A system’s reliability hinges on the master controller (MCU) correctly interpreting and acting upon the fault signals from the IPM. While many IPMs use a single fault pin for all errors, understanding the underlying cause is crucial for effective system-level response and troubleshooting.

1. Overcurrent (OC) and Short-Circuit (SC) Protection

Overcurrent and short-circuit are the most dangerous events for an IGBT. An IPM distinguishes between a sustained, moderate overload (OC) and an extreme, near-instantaneous short-circuit (SC).

• Detection Mechanism: The internal control IC monitors the IGBT collector-emitter saturation voltage (VCE(sat)) or uses an integrated current sensor. During normal operation, VCE(sat) is low. If a massive current event occurs, the IGBT comes out of saturation, and VCE(sat) rises sharply. This rapid rise is the trigger for SC protection. A less extreme, but sustained, high current triggers OC protection.
• IPM Response: Upon detecting an SC or critical OC event, the IPM initiates a “soft shutdown.” Instead of instantly cutting off the gate voltage, which would cause a massive and potentially destructive voltage spike (V = -L * di/dt) due to stray inductance, the gate is turned off in a controlled, two-stage manner. This minimizes overshoot while still protecting the device. Simultaneously, the FAULT pin is pulled low to alert the MCU.
• MCU Action: When the MCU detects a FAULT signal, it must immediately halt all PWM signals to the IPM and initiate a system-level shutdown. It should then log the fault and alert the user. Because both OC and SC faults can indicate serious systemic issues (e.g., motor winding failure, phase-to-ground short), the system should typically require a manual reset after such an event.

2. Over-Temperature (OT) Protection

Thermal stress is a primary driver of power module aging and failure. IPMs provide direct, fast-acting thermal protection.

• Detection Mechanism: An integrated NTC thermistor provides a voltage signal proportional to the module’s substrate temperature. The control IC compares this against two internal thresholds: a pre-alarm (warning) and a trip point (shutdown).
• IPM Response: If the temperature exceeds the OT trip point, the control IC blocks all gate drive signals, shutting down the IPM and asserting the FAULT pin. The shutdown is immediate, as there is no di/dt risk.
• MCU Action: An MCU can implement a two-stage thermal management strategy. By continuously monitoring the analog voltage from the NTC (if available on a separate pin) or a digital temperature output, it can trigger a “pre-alarm” response first. For example, at 100°C, it might reduce the switching frequency or limit motor torque to lower losses. If the temperature continues to rise and the IPM’s internal OT protection trips (e.g., at 125°C), the MCU should perform an emergency stop and diagnose the cooling system (e.g., check for fan failure, clogged air filters, or poor thermal paste application).

3. Under-Voltage Lockout (UVLO) Protection

Driving an IGBT with insufficient gate voltage is a recipe for disaster. It increases conduction losses (VCE(sat)) and can lead to thermal runaway.

• Detection Mechanism: The IPM’s internal driver IC monitors its own low-voltage control supply (typically +15V). If this voltage droops below a specified threshold (e.g., 12.5V), the UVLO function is activated.
• IPM Response: The IPM blocks all gate drive signals and pulls the FAULT pin low to prevent the IGBTs from operating in a high-loss, linear region. The IPM will typically remain in this state until the supply voltage recovers to a safe level, at which point it may automatically reset.
• MCU Action: A UVLO fault points directly to a problem with the control power supply. The MCU should immediately shut down the system and flag a “power supply fault.” Engineers should investigate the 15V rail for instability, poor regulation, or insufficient current capability, especially during high-frequency switching when the gate drive circuitry demands high peak currents.

Summary of IPM Fault Diagnostics

The following table summarizes the key fault types, their detection methods, and the expected system response.

Fault Type	Detection Method	Typical IPM Response	Recommended MCU Action
Short-Circuit (SC)	Desaturation detection (rapid VCE(sat) rise) or integrated current sensor.	Immediate soft-shutdown of IGBTs; FAULT pin asserted.	Halt all PWM signals instantly; require manual reset and system inspection.
Overcurrent (OC)	Integrated current sensor or VCE(sat) monitoring over a time window.	Soft-shutdown after a defined delay; FAULT pin asserted.	Halt all PWM signals; check for motor stall or load binding.
Over-Temperature (OT)	Internal NTC thermistor reading exceeds trip threshold.	Immediate shutdown of all IGBTs; FAULT pin asserted.	Initiate emergency stop; diagnose cooling system (fan, heatsink, airflow).
Under-Voltage Lockout (UVLO)	Control supply voltage (Vcc) drops below the minimum operating level.	Immediate shutdown of all IGBTs; FAULT pin asserted until Vcc recovers.	System shutdown; flag power supply error and investigate the 15V rail stability.

From Fault Diagnosis to Predictive Maintenance

The true power of an Intelligent Power Module is unlocked when we move beyond simple reactive protection and use its diagnostic data for predictive maintenance. This is a cornerstone of Industry 4.0, transforming maintenance from a fixed schedule to a condition-based, proactive strategy. By logging and analyzing data from the IPM over time, engineers can identify trends that signal impending failures before they occur.

Here are practical strategies to implement predictive maintenance using IPM data:

1. Thermal Trend Monitoring: Continuously log the NTC thermistor temperature under consistent load conditions. A gradual, steady increase in operating temperature over weeks or months can indicate:
   • Cooling Fan Degradation: A failing fan spins slower, reducing airflow and causing temperatures to rise.
   • Heatsink Contamination: Dust and debris accumulating on heatsink fins reduce their efficiency.
   • Thermal Interface Degradation: The thermal grease or pad between the IPM and the heatsink can “pump out” or dry over time, increasing thermal resistance.

   By setting a “warning” threshold for this long-term trend, maintenance can be scheduled to clean heatsinks or replace fans during planned downtime, avoiding a costly line-stop failure.

2. Fault Log Analysis: Modern MCUs have enough memory to log the type, frequency, and operating conditions (load, temp) of every fault. Analyzing this log can reveal hidden problems. For instance, intermittent OC faults that occur only under high torque could point to a mechanical issue in the drivetrain rather than an electrical fault. A rising frequency of UVLO faults might signal an aging auxiliary power supply.
3. Advanced Sensing and AI: The next frontier involves integrating more sophisticated data with AI algorithms. Some advanced DIPIPM™ and other smart modules provide real-time feedback on VCE(sat) and other internal parameters. By tracking subtle changes in these values over the module’s lifetime, machine learning models can predict the end-of-life of the IGBT itself, accounting for degradation from power cycling and thermal stress. This allows for a truly optimized asset lifecycle, replacing modules just before they are projected to fail. The accuracy of this depends heavily on quality data from on-chip sensors, a topic further explored in on-chip sensing for IGBTs.

Conclusion: The IPM as a System Health Monitor

The integrated diagnostic functions within an Intelligent Power Module are more than just safety features; they are a rich source of data about the health and performance of the entire power system. For design engineers and system integrators, mastering these functions is essential for building robust and fault-tolerant machinery. The initial step is to create a reliable fault-handling routine that correctly interprets the basic OC, SC, OT, and UVLO alerts. The ultimate goal, however, is to leverage this stream of data for intelligent, predictive maintenance. By logging and analyzing thermal trends and fault patterns, you can transition from fixing failures after they happen to preventing them before they start, maximizing uptime, reducing operational costs, and extending the life of your critical assets.