Industrial Inverter IGBT Modules: The Reliability Parameter Most Often Skipped in 34 mm and 62 mm Package Selection
Many engineers follow roughly this sequence when selecting IGBT modules for an inverter or servo drive: set the voltage class from the DC bus voltage, set the current class from motor rated current times a margin factor, compare VCE(sat), and then check price and lead time. The flow looks complete, but it omits a parameter that largely determines field lifetime—power cycling capability. The real operating profile of an industrial inverter on a production line is rarely “constant load, constant
1. Real Inverter Duty: Junction Temperature Is Constantly Cycling
Consider a typical case. A 22 kW inverter drives the oil-pump motor of an injection molding machine. Each molding cycle includes a low-speed standby phase, a high-pressure/high-speed phase, and a return/cooling phase. IGBT losses vary with load current, and chip junction temperature rises and falls within each cycle.
Industrial servo duty is more severe. A positioning move is often completed within hundreds of milliseconds to a few seconds. Current peaks during acceleration/deceleration, junction temperature rises quickly, and after positioning the current falls and junction temperature drops. An automated line running tens of thousands of cycles per day means the chips, bond wires, and solder layers inside the module experience tens of thousands of temperature excursions every day.
Mechanical stress accumulates with each excursion. An IGBT module is a multilayer composite: silicon chip, chip solder, DBC ceramic substrate, substrate solder, and copper baseplate. Each material has a different coefficient of thermal expansion. Every junction-temperature cycle imposes a shear stress on each interface. When stress accumulates beyond a threshold, failure occurs. Typical failure modes include:
- Bond wire lift-off: microcracks form at the joint between the aluminum bond wire and the chip surface metallization under repeated thermal expansion and contraction. As cracks propagate, the bond wire lifts. Current density in the remaining bond wires increases, creating a positive feedback loop that accelerates degradation.
- Solder fatigue: chip solder or substrate solder cracks under cyclic shear stress, and Rth(j-c) increases. For the same loss, junction temperature then drifts upward.
2. ΔTj: An Underestimated Lifetime Lever
The mainstream method for quantifying power-cycling lifetime in the industry is the Coffin-Manson power-law model and its extensions (the LESIT model and the Infineon CIPS2008 model). These models share a direct conclusion:
Cycle count Nf has an inverse power-law relationship with junction-temperature swing ΔTj, and ΔTj is the dominant lifetime variable. (Source: LESIT project and Infineon CIPS2008 model literature.)
This implies two things:
- A small increase in ΔTj is amplified exponentially into a large reduction in lifetime. Industry experience commonly cited is that for every approximately 10°C increase in ΔTj, cycling lifetime may be reduced by tens of percent. The exact percentage depends on module design, but the direction is consistent: larger ΔTj means shorter lifetime.
- Mean junction temperature Tj,m also matters. At the same ΔTj, a higher mean junction temperature accelerates bond-wire fatigue and solder creep (Arrhenius temperature acceleration term). This conclusion also comes from the LESIT model.
Therefore, whether a module can survive ten years of start/stop duty depends on two questions: How large is ΔTj under your load profile? And will Tj,max frequently approach its limit? The answers lie in two parameters that are most easily skipped—Tjmax and thermal resistance Rth(j-c).
This is why power-cycling lifetime is more decisive for field performance than “a sufficiently large rated current.”
3. 34 mm / 62 mm Packages: A Mainstream Platform for Medium-Power Inverters
In Hypersemi’s IGBT module portfolio, 34 mm and 62 mm packages together cover 19 models, making them the highest-share package family after Half-Bridge (data source: Hypersemi official product list). These two package systems have mature heatsink mounting holes and assembly processes in the medium-power inverter and servo drive market. For replacement or upgrade, users do not need to re-machine the heatsink or change the mounting structure.
Corresponding mainstream application classes:
- 62 mm package / 1200 V / 150 A / Six-Pack: covers the mainstream class for 30 kW–75 kW general-purpose inverters, benchmarked against Infineon FS150R12KT4 (Infineon’s website confirms this model as a 1200 V / 150 A six-pack in EconoPACK 3 package, 62 mm width).
- 34 mm package / 1200 V / 75 A / Six-Pack: suitable for lower-power servo drives or brake-unit branches.
On “whether the current rating is large enough” alone, 150 A is clearly superior to 75 A. Once power-cycling lifetime is brought into the picture, the trade-off becomes more nuanced.
4. Same Voltage Class, Two Tjmax Values: The Real Trade-off Between the 150 A and 75 A Options
The two models below are from Hypersemi product datasheets (data source: Hypersemi official product page for the corresponding model):
Parameter | 1200 V / 150 A Six-Pack | 1200 V / 75 A Six-Pack |
|---|---|---|
Package family | 62 mm package | 34 mm package |
VCE(sat) | 1.53 V | 1.9 V |
Eon / Eoff | 11.09 mJ / 10.18 mJ | 5.2 mJ / 3.6 mJ |
Tjmax | 150°C | 175°C |
Typical application | Continuous high-load general-purpose inverter | Frequent start/stop servo, branch circuits |
The key difference is Tjmax: the 150 A option allows a junction-temperature limit of 150°C, while the 75 A option reaches 175°C. In power-cycling terms, this 25°C gap means:
- Larger allowable junction-temperature swing space. Under the same ambient temperature and cooling conditions, a 175°C limit leaves more safety margin for ΔTj. Every reduction in ΔTj extends cycling life exponentially under the Coffin-Manson power law.
- Cycling margin remains after derating. For servo duty with frequent start/stop, high peak current but low average current, the 75 A option after appropriate derating often still has a lower ΔTj than the 150 A option running close to its limit under the same duty. Note: “often” here is a theoretical inference; the specific conclusion requires thermal simulation based on the actual load profile.
- Conversely, the 150 A option’s advantage lies in continuous high load. For pumps, fans, and compressors running at full load for long periods, ΔTj is small while average junction temperature is high. The main benefit then comes from lower loss (VCE(sat) 1.53 V) and lower current density.
In one sentence, the trade-off logic is: “a higher rated current” does not equal “longer field life.” In frequent thermal-cycling duty, a “smaller” module with higher Tjmax and lower ΔTj after derating may outlast a “larger” module with ample current margin but a lower junction-temperature limit. The right approach is to calculate ΔTj for the actual load profile and let Tjmax and the power-cycling curve decide.
Conclusion
Competition in the inverter and servo drive market ultimately comes down to equipment availability and maintenance cost. As the power core, when the IGBT module fails directly determines when the inverter stops. On the selection checklist, Tjmax, Rth(j-c), and power-cycling capability deserve to be placed before rated current—current determines whether the module can operate; power cycling determines how long it will stay in service.
Hypersemi’s 34 mm / 62 mm 1200 V portfolio of 19 models is now available for selection. Please contact us.