Engineer’s Guide: Designing a 150kW+ DC Fast Charger with 1200V IGBTs

As a power electronics engineer tasked with developing the next generation of high-power DC fast chargers, you face a critical decision at the heart of your design: the main power switch. While new technologies like SiC are promising, the proven reliability and cost-effectiveness of silicon IGBT Modules make them the workhorse for today’s 150kW to 350kW systems. However, selecting the right one is far from simple. A common mistake is to assume that two modules with the same voltage and current ratings are interchangeable. This oversight can lead to thermal failures, unexpected EMI issues, and missed efficiency targets.

Consider two industry-standard 1200V, 400A half-bridge modules: the SKM400GB128D from SEMIKRON and the FF400R12KE3 from Infineon. At first glance, their datasheets show a nearly identical typical collector-emitter saturation voltage (VCE(sat)) of 1.70V at 400A and 125°C. An engineer focused solely on conduction losses might conclude they are equivalent. Yet, their underlying chip technologies—SEMIKRON’s SPT IGBT with a CAL diode versus Infineon’s TRENCHSTOP™ IGBT3 with an EmCon diode—result in different switching behaviors that are critical in a high-frequency EV charger topology. This guide will take you beyond the surface-level specs to provide an actionable framework for selecting, implementing, and protecting these high-power modules in a demanding DC fast charger application.

Decoding the Datasheet: Key Parameters for High-Frequency Charger Design

In a typical DC fast charger using a Phase-Shift Full-Bridge (PSFB) or LLC resonant topology, the IGBTs operate at switching frequencies between 16 kHz and 40 kHz. At these frequencies, switching losses can dominate, making a low VCE(sat) only part of the story. Here’s where to focus your analysis.

Conduction vs. Switching Losses: The Central Trade-Off

Every IGBT generation balances conduction losses (heat generated when the switch is on) and switching losses (heat generated during turn-on and turn-off). The key is to match the IGBT’s characteristics to your application’s dominant operating mode.

• Conduction Losses (P_cond): Determined primarily by VCE(sat) and the on-state collector current. A lower VCE(sat) is always better for reducing heat when the device is conducting.
• Switching Losses (P_sw): Determined by the turn-on (E_on) and turn-off (E_off) energies. These are heavily influenced by the speed of the IGBT and its integrated freewheeling diode. Faster switching reduces these losses but can increase voltage overshoots and EMI.

Let’s compare our two example modules based on their datasheet values:

Parameter (Typical Values @ 400A, 125°C, Vge=15V)	SEMIKRON SKM400GB128D	Infineon FF400R12KE3	Implication for EV Charger Design
Chip Technology	SPT (Soft Punch Through) IGBT + CAL 4 Diode	TRENCHSTOP™ IGBT3 + EmCon Diode	Different generations and designs lead to nuanced performance differences, especially in diode recovery.
VCE(sat)	1.70 V	1.70 V	Conduction losses will be very similar under identical load currents. No clear winner here.
Total Switching Energy (E_ts = E_on + E_off)	55 mJ + 55 mJ = 110 mJ	48 mJ + 50 mJ = 98 mJ	The FF400R12KE3 has ~11% lower total switching energy, making it a more efficient choice for higher frequency ( > 20 kHz) designs where switching losses are dominant.
Diode Forward Voltage (V_f)	1.80 V	1.85 V	The SKM400GB128D’s CAL diode has slightly lower conduction losses during freewheeling periods.
Short-Circuit Withstand Time (t_sc)	10 µs	10 µs	Both modules offer robust short-circuit protection, a critical safety requirement.

Engineer’s Takeaway: For a 150kW charger operating at 25 kHz, the ~12 mJ difference in switching energy per cycle on the FF400R12KE3 translates to a reduction of 300W of switching loss per device (12mJ * 25,000 Hz), significantly easing the thermal management burden.

Practical Thermal Design: From Datasheet to Heatsink Selection

Thermal failure is a leading cause of IGBT module failure. A robust thermal design is not optional. The goal is to ensure the maximum junction temperature (T_j,max), typically 150°C for these modules, is never exceeded, with a healthy safety margin (e.g., 20-25°C).

Step 1: Calculate Total Power Loss (P_total)

First, we must estimate the total power dissipated as heat. This is the sum of conduction and switching losses for both the IGBT and the freewheeling diode.

P_total = (P_cond_IGBT + P_sw_IGBT) + (P_cond_diode + P_rec_diode)

Let’s run a simplified calculation for one IGBT in a 150kW DC/DC converter stage (assuming 800V DC link, 25 kHz switching, and a 50% duty cycle for simplicity).

• Load Current (I_c): 150,000W / 800V ≈ 188A
• Conduction Loss (IGBT): P_cond ≈ VCE(sat) * I_c * DutyCycle = 1.70V * 188A * 0.5 = 159.8 W
• Switching Loss (IGBT): P_sw ≈ (E_on + E_off) * f_sw. Using the FF400R12KE3’s data (scaled for 188A, which is roughly linear for a first-pass estimate): (98 mJ * 188/400) * 25,000 Hz ≈ 115.2 W
• Total Loss per IGBT (approx.): 159.8 W + 115.2 W = 275 W

Note: This is a simplified calculation. For a real design, you must use the curves in the datasheet to find VCE(sat) and E_sw at your specific operating current and temperature.

Step 2: Use Thermal Resistance to Find the Required Heatsink

The concept of Thermal Resistance (R_th) is like Ohm’s Law for heat. The total thermal resistance determines the temperature rise for a given power loss.

ΔT = P_total * R_th_total, where R_th_total = R_th(j-c) + R_th(c-s) + R_th(s-a)

• R_th(j-c) (Junction-to-Case): An intrinsic property of the module. For both our modules, it’s 0.055 K/W per IGBT.
• R_th(c-s) (Case-to-Sink): Depends on the thermal interface material (TIM) or thermal grease. A typical value is ~0.015 K/W for a module this size.
• R_th(s-a) (Sink-to-Ambient): This is the value you are solving for—it represents the performance required from your heatsink (and fan).

Calculation Example:

Let’s set our maximum desired junction temperature T_j to 125°C for reliability, and assume an ambient temperature T_a inside the charger cabinet of 55°C.

Total allowed temperature rise (ΔT) = 125°C – 55°C = 70°C.

Total required R_th_total = ΔT / P_total = 70°C / 275W = 0.255 K/W.

Now, find the heatsink requirement:

R_th(s-a) = R_th_total – R_th(j-c) – R_th(c-s) = 0.255 – 0.055 – 0.015 = 0.185 K/W.

Actionable Decision: You must select a forced-air or liquid-cooled heatsink system that provides a thermal resistance of 0.185 K/W or lower to keep the IGBT junction safely below 125°C under these conditions. This is a demanding requirement that will likely necessitate a high-performance, fan-cooled aluminum heatsink or a liquid cooling loop. For a detailed guide on this topic, refer to our article on Why Rth Matters: Unlocking IGBT Thermal Performance.

Ensuring Reliability: Gate Drive and Protection Essentials

A perfect thermal design is useless if the IGBT is destroyed by electrical stress. The Gate Drive circuit is not just a signal buffer; it is the module’s first line of defense.

1. Short-Circuit Protection (Desaturation Detection)

The most catastrophic failure mode is a short circuit. Both the SKM400GB128D and FF400R12KE3 are rated for 10µs of short-circuit withstand time. The gate driver must detect this condition and turn off the IGBT within this window.

• How it works: The driver monitors the IGBT’s VCE voltage during the on-state. If VCE rises above a certain threshold (e.g., 7-9V) while the gate is commanded on, it signifies a massive current draw (desaturation).
• Action: The driver immediately initiates a “soft turn-off” to control the di/dt and prevent a destructive overvoltage spike, then flags a fault to the system controller.

2. Preventing Parasitic Turn-On with a Negative Gate Voltage

In a half-bridge configuration, the rapid turn-on of one IGBT can induce a voltage spike on the gate of the other (off-state) IGBT through the Miller capacitance (C_gc). If this spike exceeds the IGBT’s gate threshold voltage (V_GE(th)), the device can momentarily turn on, causing a shoot-through event that destroys the module.

• Solution: Use a gate driver that provides a negative turn-off voltage (e.g., -5V to -15V). This provides a much larger margin against parasitic turn-on. A gate driver IC like the SKHI 24 R is specifically designed for this purpose.
• Additional Protection: An active Miller Clamp can be used to short the gate to the emitter during the off-state to provide further protection. For more practical tips, see our guide on 5 Practical Tips for Robust IGBT Gate Drive Design.

3. The Importance of a Kelvin Emitter Connection

High-power modules like these feature a Kelvin Emitter (or auxiliary emitter) connection. This provides a separate, clean return path for the gate driver circuit, bypassing the main load current path. Using this connection is mandatory for clean, fast switching as it prevents the load current’s self-inductance from introducing noise into the gate drive loop.

Procurement and Alternative Solutions

While the SKM400GB128D and FF400R12KE3 are excellent choices, supply chain and design priorities may require alternatives. A suitable alternative in a similar performance class is the Fuji Electric 2MBI400TB-060. It offers comparable ratings (600V, 400A) and showcases Fuji’s chip technology, providing another option for engineers to evaluate based on dynamic performance and thermal impedance.

When designing your system, the choice between modules should be based on a holistic view of your system’s priorities:

• For Highest Efficiency at High Frequency (>25 kHz): The Infineon FF400R12KE3‘s lower switching losses provide a distinct advantage.
• For Robustness and Proven Field Reliability: The SEMIKRON SKM400GB128D, with its rugged SPT design and well-regarded CAL diode, is a benchmark for reliability in demanding applications.

Frequently Asked Questions (FAQ) for EV Charger IGBT Selection

1. Why not just use SiC MOSFETs for a new EV charger design?
While SiC offers superior switching speed and lower losses, mature silicon IGBTs like the ones discussed offer a significantly lower cost per amp, a more robust short-circuit withstand capability, and a proven supply chain. For many cost-sensitive 150kW designs, a well-implemented IGBT solution still provides the best balance of performance, reliability, and system cost.

2. What is the impact of stray inductance in my busbar design?
Stray inductance in the DC link is a major cause of voltage overshoot during IGBT turn-off. This can exceed the IGBT’s 1200V rating and cause immediate failure. It is critical to use a low-inductance laminated busbar design and high-quality DC link capacitors placed as close to the module terminals as possible.

3. Can I parallel these modules for higher power (e.g., a 300kW inverter)?
Yes, but IGBT Paralleling requires careful design. Mismatches in VCE(sat) can cause unequal current sharing, and asymmetrical PCB and busbar layouts can cause mismatched switching times, leading to one module taking excessive stress. It’s crucial to ensure symmetrical layouts and consider modules with a positive temperature coefficient for VCE(sat), which aids in balancing current.

Your power module is the heart of your EV charger. Choosing it wisely, designing its thermal environment meticulously, and protecting it with a robust gate driver are the keys to a successful, reliable, and efficient product. For help selecting the ideal module for your specific application or to get a quote, contact our team of application engineers.