In applications such as 220V servo drives and inverter air conditioners, Intelligent Power Modules (IPMs) serve as the core power conversion component. Their reliability directly affects the lifespan and stability of the entire system.
An IPM integrates not only IGBTs and gate drivers but also intelligent protection features such as undervoltage lockout (UVLO), overcurrent protection (often based on Vce(sat) detection), logic control, and overtemperature shutdown. However, under real operating conditions, electrical stress, thermal stress, and control parameter interactions may still lead to module failures.
The following sections analyze common IPM failure modes and provide practical solutions, drawing on semiconductor design principles and real-world application practices from companies such as SHYSEMI.
1. Unstable Gate Driver Power Supply
Failure Mechanism
The internal gate driver IC, level-shifting circuits, and fault detection logic within an IPM require a stable supply voltage, typically around 15V or 16V (VCC).
If the voltage falls below the undervoltage lockout (UVLO) threshold:
- The IGBT cannot fully saturate
- The device operates in the linear region
- Conduction losses and temperature rise increase rapidly
- False fault alarms or chip damage may occur
Solution
- Precision Voltage Regulation: Ensure the SMPS output remains within ±5% of the nominal 15V, with ripple below 100mV.
Many IPMs from SHYSEMI feature a UVLO threshold around 12.5V with hysteresis protection, but operation near the threshold should still be avoided.
- Local Decoupling and Monitoring: Add low-ESR capacitors close to the power pins (e.g., 10µF + 0.1µF MLCC) to suppress voltage dips. It is also recommended to monitor supply voltage through MCU ADC sampling and trigger soft shutdown if abnormal fluctuations occur.
2. PWM Signal Distortion and Bridge Shoot-Through
Failure Mechanism
If the PWM signal rise or fall time exceeds roughly 100ns, the IGBT may remain in the linear region for too long during switching transitions, increasing switching losses.
A more severe issue occurs when upper and lower bridge signals overlap. Even with configured dead time, Miller effect or parasitic inductance may trigger unintended conduction, resulting in catastrophic short circuits.
Solution
- Signal Conditioning and Driver Optimization: Optimize the interface between the MCU and IPM by using high-speed optocouplers or magnetic isolation drivers. Recommended signal edge times are below 50ns.
IPMs from SHYSEMI include optimized internal driver stages that help suppress input noise, but proper external circuit design remains essential.
- Adaptive Dead-Time Control: Adaptive dead-time algorithms can dynamically adjust switching timing based on load current direction. This improves switching safety while maintaining low total harmonic distortion (THD).
3. Continuous High-Current Operation
Failure Mechanism
During acceleration in servo systems, current may reach up to three times the rated value. Under such conditions:
- IGBT chips junction temperature rises rapidly due to conduction and switching losses
- Inductive loads generate voltage spikes during turn-off (V = L·di/dt)
- These spikes may exceed the IGBT blocking voltage rating (e.g., 600V or 1200V)
This can trigger avalanche breakdown.
Solution
- Transient Voltage Suppression: Reduce stray inductance through optimized DC bus layout. Add high-frequency snubber capacitors (such as CBB capacitors) or RCD snubber circuits to clamp switching spikes.
IPMs from SHYSEMI use Trench Field-Stop IGBT technology, which provides strong short-circuit capability and avalanche robustness. However, safe design margins should still be maintained, typically operating at about 80% of rated voltage.
- Thermal Management: Implement I²t protection algorithms based on thermal models to estimate junction temperature in real time and limit extreme operating durations. Proper heatsink contact and uniform thermal grease application are also critical.
4. Current Oscillation and Parameter Mismatch
Failure Mechanism
If current loop PID parameters are not properly tuned, particularly with excessive proportional gain (Kp), closed-loop current oscillation may occur.
These oscillations can produce peak currents far higher than the averaged ADC readings. In extreme cases, instantaneous current may exceed the Reverse Bias Safe Operating Area (RBSOA) of the IGBT, potentially causing bond wire failure or chip burnout.
Solution
- Automatic Parameter Tuning: Use servo system auto-tuning functions to ensure accurate motor parameters such as inductance (Ld/Lq). Hardware current filtering and higher PWM carrier frequencies can also improve sampling accuracy.
- Software Protection: Introduce current limiters and anti-windup protection in the control algorithm to prevent integrator saturation.
For systems using SHYSEMI IPMs, the integrated current detection function can provide fast hardware-level overcurrent protection as a backup to software safety mechanisms.
5. Overpower Operating Conditions
Failure Mechanism
When motors accelerate under heavy load at low speed, back electromotive force (BEMF) remains low. To maintain torque, the inverter must inject large current.
As speed increases:
- Output power rises rapidly
- Input power increases accordingly
With a fixed DC bus voltage, the bus current follows:
Ibus = Pout / (Vbus × η)
This may eventually exceed the IPM overcurrent protection threshold.
Solution
Power Monitoring and Control: Implement a power observer in the control algorithm to calculate real-time output power. Software protection thresholds should be set slightly lower than the hardware protection level to enable early intervention.
DC Bus Energy Buffering: Increasing DC bus capacitance can provide additional energy storage, reducing sudden current spikes during rapid power demand changes.
Failure Mechanism
During motor parameter identification or locked-rotor testing, applying large Vd or Vq voltage commands can cause extremely fast current rise.
Because the rotor is stationary and inductance is small:
V = L·di/dt + iR
Most of the applied voltage contributes to current change, causing current to reach hundreds of amperes within microseconds.
Traditional control loops sample current every tens of microseconds, which is too slow to respond. In these cases, only the IPM’s internal hardware protection—such as desaturation detection—can intervene.
Solution
Voltage Limiting and Soft Ramp: In voltage-mode operation, command voltage should be limited to the impedance drop corresponding to rated excitation current. Soft-start ramp functions should be used to prevent step voltage shocks.
High-Reliability Modules: For applications requiring frequent testing under these conditions, selecting modules with longer short-circuit withstand time (SCWT) and fast desaturation protection—such as certain models from SHYSEMI—can improve system safety.
Conclusion
IPM failures typically result from a combination of electrical stress, thermal stress, and control strategy interactions. By optimizing driver power design, switching timing, surge suppression, and control parameters at the semiconductor system level, engineers can significantly improve inverter reliability.
Combined with the advanced features of modern IPM solutions—such as low Vce(sat), wide safe operating area (SOA), and integrated protection mechanisms—high-performance modules from companies like SHYSEMI help move system design from reactive troubleshooting toward preventive engineering and precise control.
Would you like me to provide a comparison table of SHYSEMI IPM specifications to help you select the right module for your application?