An Engineer’s Guide to Accurately Measuring IGBT Rth,j-c and Zth,j-c

In the world of power electronics, the numbers in a datasheet are the foundation of any design. Yet, experienced engineers know that relying solely on typical values can lead to systems that underperform or, worse, fail prematurely. This is especially true for thermal parameters. The junction temperature (Tj) of an IGBT module is the ultimate arbiter of its performance and lifespan. Miscalculating it can be a costly mistake. This is why accurately measuring and understanding junction-to-case thermal resistance (R_{th, j-c}) and transient thermal impedance (Z_{th, j-c}) is not just an academic exercise—it’s a critical engineering practice.

The Critical Role of Thermal Characterization in Power Electronics

Every watt of power lost in an IGBT module becomes heat. The primary job of a thermal management system is to efficiently move this heat from the silicon die to the ambient environment. The path from the junction to the case is the first and most critical leg of this journey. An inaccurate R_{th, j-c} or Z_{th, j-c} value can throw off all subsequent cooling calculations, leading to thermal runaway and catastrophic failure.

Why Datasheet Values Aren’t Enough

A datasheet provides a snapshot under specific, idealized conditions. R_{th, j-c}, for instance, represents the steady-state thermal resistance. It tells you the temperature rise per watt after the device has been operating under a constant load for a long time and has reached thermal equilibrium. However, many real-world applications, like motor drives or solar inverters, involve pulsed or variable loads. For these scenarios, the steady-state value is insufficient. This is where Z_{th, j-c} comes in. A detailed guide on using this curve for thermal design can be found in our article on mastering the Zth curve.

Differentiating Rth,j-c (Steady-State) vs. Zth,j-c (Transient)

It’s crucial to understand the distinction between these two parameters:

• R_{th, j-c} (Thermal Resistance): This is a single value, measured in °C/W or K/W. It describes the temperature difference between the junction and the case under steady-state DC power conditions. It’s useful for applications with very stable, long-duration loads.
• Z_{th, j-c} (Transient Thermal Impedance): This is a curve, not a single value, plotting thermal impedance against time. It describes the temperature rise in response to a power pulse of a specific duration. For any time (t) > 0, Z_{th, j-c}(t) will always be less than R_{th, j-c}. It is essential for calculating the peak junction temperature during short-duration events like motor startup, braking, or short-circuit conditions.

In essence, R_{th, j-c} is the final value on the Z_{th, j-c} curve as time approaches infinity.

Fundamental Principles of IGBT Thermal Measurement

Directly placing a thermocouple on a microscopic silicon junction operating at high voltage is impossible. Therefore, we rely on indirect measurement methods that use the chip’s own electrical properties to determine its temperature.

The Concept of the Thermal Path: From Junction to Case

Heat generated at the IGBT junction travels through several layers before reaching the module’s baseplate (case): the silicon chip itself, a solder or sinter layer, a Direct Bonded Copper (DBC) substrate, and another solder/sinter layer connecting to the baseplate. Each of these layers has its own thermal resistance and capacitance. The Z_{th, j-c} curve is a representation of this entire thermal stack. For a deeper dive into the physics, resources on thermal resistance provide excellent background information.

Temperature Sensitive Electrical Parameters (TSEPs) as the Key

The core principle of electrical thermal measurement is to find an electrical parameter that changes predictably with temperature. This is known as a Temperature Sensitive Electrical Parameter (TSEP). For an IGBT, several parameters can be used, but one is most common for its accuracy and stability.

The Role of VCE(sat) as a TSEP for Junction Temperature Sensing

The collector-emitter saturation voltage, V_CE(sat), is the most widely used TSEP for IGBTs. When a small, constant sense current is passed through the IGBT (typically around 1/1000th of the nominal current), the V_CE(sat) exhibits a nearly linear negative temperature coefficient. This means as the junction temperature increases, the voltage drop decreases.
The process involves two stages:

1. Calibration: The unpowered IGBT module is placed in a thermal chamber and its temperature is slowly varied. At each temperature step, the small sense current is applied, and the corresponding V_CE(sat) is recorded. This creates a highly accurate calibration curve (e.g., mV/°C).
2. Measurement: During an actual test, the V_CE(sat) is measured, and the calibration curve is used to determine the junction temperature with high precision.

The Continuous Power Method: Measuring Rth,j-c

Measuring the steady-state R_{th, j-c} is conceptually straightforward but difficult to perform accurately. The now-obsolete MIL-STD-883E standard was one approach, but it was plagued by reproducibility issues. The primary challenge is accurately measuring the case temperature (Tc).

Test Setup and Procedure

The traditional method involves:

1. Mounting the IGBT module onto a high-performance, liquid-cooled cold plate.
2. Drilling a small hole in the cold plate to embed a thermocouple directly beneath the chip to measure Tc.
3. Applying a constant heating power (P_H) to the IGBT until both the junction temperature (Tj) and case temperature (Tc) stabilize.
4. Tj is determined using the V_CE(sat) TSEP method just after the heating power is removed.
5. R_{th, j-c} is then calculated as (Tj – Tc) / P_H.

Calculation and Sources of Error

The biggest source of error in this method is the Tc measurement. A shift of just 1 mm in the thermocouple’s position can lead to a temperature reading difference of several degrees, potentially causing a 15-20% error in the final R_{th, j-c} value. This is why this method has poor reproducibility between different labs and manufacturers.

Transient Thermal Analysis: The Gold Standard for Zth,j-c Measurement

To overcome the limitations of steady-state methods, the industry has standardized on a transient approach defined by JEDEC. This method not only provides the complete Z_{th, j-c} curve but can also be used to derive a more accurate R_{th, j-c}.

The JEDEC JESD51-14 Standard Electrical Test Method

The JEDEC JESD51-14 standard, known as the Transient Dual Interface Method (TDIM), is the definitive modern procedure. Its key advantage is that it eliminates the need for direct case temperature measurement with a thermocouple, thus removing the largest source of error.

Step-by-Step: The Heating and Cooling Curve Analysis

The process, as defined by standards like IEC 60747-9 and JESD51-14, works as follows:

1. The device is heated with a constant power pulse until it reaches full thermal equilibrium.
2. The heating power is abruptly turned off, and a small, constant sense current is immediately applied.
3. The V_CE(sat) is measured and recorded over time as the device cools down. This is the “cooling curve.”
4. Using the prior TSEP calibration, this voltage curve is converted into a temperature decay curve, Tj(t).
5. This cooling curve is then mathematically processed to generate the Z_{th, j-c} curve.

The TDIM technique involves performing this measurement twice, each time with a different thermal interface material (TIM) between the module and the cold plate (e.g., one with and one without thermal grease). The point where the two resulting thermal impedance curves diverge indicates the boundary between the module’s case and the external environment. Everything before this point on the curve represents the true Z_{th, j-c}.

From Raw Data to a Zth,j-c Curve: Foster vs. Cauer Models

The Z_{th, j-c} curve in datasheets is typically represented by a series of RC pairs that can be used in simulation software. There are two common network models:

Model Type	Description	Pros	Cons
Foster Model	A mathematical chain of parallel RC elements. The values are curve-fitted to match the measured thermal response.	Simple to derive from measurement data.	The nodes in the network do not correspond to physical locations within the IGBT module. It cannot be used to determine temperatures at intermediate layers (e.g., the DBC).
Cauer Model	A ladder network of series thermal resistances and shunt thermal capacitances.	Represents a physical discretization of the heat flow path. Each node can be mapped to a physical layer (chip, solder, DBC, baseplate).	More complex to derive directly from measurement. Often, a Foster model is created first and then mathematically converted to a Cauer model.

For accurate system-level thermal simulation, converting the datasheet’s Foster model to a Cauer model is often a necessary step. Further reading on leveraging such models is available in our guide to accurate IGBT loss modeling.

Application Example: Calculating Peak Tj in a Motor Drive Overload

Imagine a motor drive experiences a 1-second overload, causing 200W of power loss in the IGBT. The case is maintained at 85°C.

1. Look up the Z_{th, j-c} value at t=1s from the datasheet curve (e.g., 0.08 K/W).
2. Calculate the transient temperature rise: ΔTj = Power × Z_{th, j-c} = 200W × 0.08 K/W = 16°C.
3. Calculate the peak junction temperature: Tj_peak = Tc + ΔTj = 85°C + 16°C = 101°C.

This quick calculation, impossible with R_th alone, confirms the device remains within its safe operating limits.

Practical Challenges and Best Practices in Thermal Testing

Achieving accurate thermal measurements requires meticulous attention to detail. For insights into the importance of thermal management, IEEE provides extensive research on the topic, such as the document found here: Thermal Management for Power Electronic Systems.

• Ensuring a Stable and Accurate Case Temperature (Tc): When a Tc reference is needed, use a high-performance cold plate and ensure minimal thermal interface resistance. The surface flatness of both the module and the heatsink is critical.
• Calibration of the TSEP (V_CE(sat) vs. Tj): The TSEP calibration is the foundation of the entire measurement. The device must be allowed to fully “soak” at each temperature point to ensure thermal equilibrium. Any temperature gradient within the module during calibration will skew the results. For a detailed view on how VCE(sat) varies with temperature, see this article from Infineon.
• Minimizing Electrical Noise During Measurement: The switch from high heating current to low sense current must be extremely fast to capture the initial rapid temperature drop. This transition can inject significant noise. Proper grounding, shielding, and sometimes the use of a gate resistor are necessary to get a clean signal.

Key Takeaways for Accurate IGBT Thermal Measurement

Understanding the thermal behavior of an IGBT module is non-negotiable for reliable power system design. Here is a summary of the critical points:

Parameter / Method	Key Characteristic	Primary Use Case	Best Practice
Rth, j-c	Single value for steady-state.	Constant DC load applications.	Use value derived from JEDEC-compliant Zth, j-c testing, not older, error-prone methods.
Zth, j-c	Time-dependent curve.	Pulsed or variable loads (inverters, drives, power supplies).	Use the full curve to calculate peak Tj under transient conditions (overload, short circuit).
JESD51-14 TDIM	Transient cooling curve analysis.	Industry standard for accurate Zth, j-c and Rth, j-c measurement.	The most reliable and repeatable method, as it eliminates thermocouple placement errors.
TSEP (VCE,sat)	Using device electricals to sense temperature.	The core sensing method for all modern thermal measurement.	Requires careful, precise calibration in a thermally stable environment.

By moving beyond a simple reliance on a single datasheet number and embracing the principles of transient thermal measurement, engineers can build more robust, reliable, and efficient power electronic systems that stand the test of time.