Various Implemented Hardware
This section describes various hardware implemented not discussed in other parts of this chapter. These particular components characteristics and circuit functions are described with relevance to simulation results. Though they are significant in their contributions to the project, the amount of new work done in these areas for this project is limited.
Gate Drives & Optical Interface Board
The gate drive board has integrated isolated power supplies for driving the IGBTs. The gate signals were controlled using the MC33153 gate drive chip with integrated de-saturation protection which drives the IGBT off if a de-saturation condition is sensed. De-saturation detection and device auto turn off is a common form of IGBT short-circuit protection, with more advance versions being developed . The de-saturation trip point on the MC33153 is tunable to a particular IGBT and current level by varying a resistor in series with a high voltage clamp diode that feeds back to the de-saturation pin of the MC33153. During pulse testing, the gate drive for the auxiliary switch is modified by bypassing the resistor to maximize the allowable VCE de-saturation protection set point. This is necessary for the proper characterization and selection of the auxiliary IGBT and described further in section 3.2.
The gate driver boards are sent gate signals via an optical interface board. Gate drive generated fault signals are also sent back to the DSP through the optical interface board. The optical board Figure 7 Optical Interface Board [Figure 7] receives gate signals from the DSP generated at the 5V level.
In the final version, the gate driver is modified to operate on 15V CMOS logic using Agilent opto-couplers of the HCPL-0630 type. This is required for integration with the final application’s existing controller to meet the design drop in criterion.
The auxiliary diodes have few design requirements. They need to have larger voltage drop than the main diode to reduce circulation currents and they also need to be very fast with good reverse recovery characteristics. To minimize space and component count the IXYS VUE 130-12N07 6-pack diode module rated at 130A continuous with a 40ns reverse recovery time is selected. They work properly without incident in the implementation. A high peak current rating of a diode is very typical because of widespread use in 50 and 60 Hz applications. The particular selected diode has a forward current rating of 500A for a duration of 10ms which is adequate for the implemented ICZVT inverter. Since the diode is not a major cost item, no detailed analysis is performed to attempt to reduce the diode ratings. From simulations, the auxiliary diode peak and rms current ratings need to be at least 250A and 23A respectively.
The main IGBT used is the Eupec FF400R12KE3, rated at 1200V and an IC of 580A @ 25C. This device exhibits low conduction and switching losses and is capable of 5 kHz+ operation. The device is designed for traditional hard switching applications. Since the ICZVT almost completely eliminates switching losses, the packaging of the selected device is more substantial and costly than necessary. In a design for complete cost reduction, a device with packaging capable of dissipating the RMS conduction losses only is necessary. From simulations, the main switch peak and rms current ratings need to be at least 320A and 108A respectively.
DC Bus Capacitors
The main DC bus capacitors used are a parallel series combination of 8000uF capacitors. There are two 8000uF capacitors in parallel and two in series with the neutral point in the middle. These capacitors are connected to the main power stage via a low inductance/high current bus bar. Additionally, some 1uF, 1200V, Electronic Concepts “Lo Henry 88 Series” capacitors are used across the DC bus as snubber capacitors. Their low equivalent series inductance and equivalent series resistance design makes them suitable for this application.
The resonant capacitors are 1000V, 0.14uF, SBE 716P series polypropylene capacitors. This type of capacitor is needed due to the high current, low equivalent series resistance requirements of the ICZVT inverter. The resonant capacitor size is determined based on previous project work and device switching test results to give an optimal soft device turn off with minimal voltage overshoot.
Coupled Inductor and Saturable Inductor
The coupled inductor value is optimized to match the load to give the optimal inductor wave shape and soft-switching performance. The original coupled inductor value has an inductance of 1.2uH. After some testing, the inductance is changed to 1.6uH to help mitigate the direction of the negative current pulses. Additionally, the saturable inductor used to prevent circulation current is designed with two cores and various turns ratios. Three turns is the most common saturable inductor design tested. After several iterations and performance tests, the final design uses four saturable cores with two turns each.
Auxiliary IGBT Selection Methodology
Auxiliary IGBT Selection Foreword
For reasons of overall system cost it makes sense to attempt to find auxiliary IGBTs that will yield the lowest appreciable cost to the system but still match system goals for performance, cost and reliability. As switching characteristics, the device must have fast enough turn on and turn off characteristics to quickly turn on and off as to conduct the required auxiliary current through the coupled inductor for a short amount of time. Modern IGBTs are capable of turning on and off in this short period of time and have no problem operating at the 5 kHz inverter frequency. Since the auxiliary IGBT turns on under zero current and turns off under zero voltage conditions, the switching energy calculations are not of great importance either. Hence, for IGBT selection the switching characteristics are largely neglected.
The current carrying capabilities are the area of primary concern. IGBT cost is directly proportional to die area/current carrying capability and also highly dependent on packaging. In the ICZVT inverter, the auxiliary IGBTs conduct current for only a few microseconds resulting in a low RMS current carrying requirement, but have a high repetitive peak current requirement. From simulations, the auxiliary switch peak and rms current requirements are 280A and 43A respectively. Most IGBT datasheets describe device current carrying capabilities in terms of continuous collector current (IC), and some define a peak repetitive collector current (ICM or IRCM) at a duration on the level of milliseconds at a given junction temperature. However, the auxiliary switching device in the ICZVT is not being operated anywhere near continuously, and conduction time is three orders of magnitude shorter than the typical datasheet ICM rating. The datasheet current ratings are therefore inadequate for determining which IGBT will be right for the application. More information must be extrapolated from datasheet curves, and/or the IGBT must be tested to determine a suitable device.
Design for Reliability
To correctly select the auxiliary IGBTs for the system, the IGBT must be understood on a more fundamental level. There are two failure mechanisms of IGBT dies, over-voltage (avalanche energy breakdown) and over-current. There are additionally primary and secondary combinations of the two failures such as avalanche induced second breakdown limit (AISBL) and circuit oscillations  that can lead to device failure. In this application the 640V DC bus is very stiff and a device rating of 1200V prevents any voltage spikes from exceeding the device SOA. The device is operated within the SOA even under short circuit conditions so the avalanche characteristics/over voltage failure mechanisms of the IGBT are not considered. Of more importance is the high temperature behavior and over-current failure.
An over-current failure is actually an over-temperature failure caused by excessive conduction or de-saturation region power losses that the device packaging and external thermal system are unable to remove resulting in excessive increased heating of the device silicon. Once an IGBT’s junction temperature exceeds a limit defined by the manufacturer (a property of the die) the silicon may fail. This failure may not be immediate evident to the end user if the over-temperature condition is only seen for a short period of time and does not cause thermal runaway. Only measuring the case temperature to determine junction temperature is not a totally adequate technique to determine reliability. Additionally, though not failures seen early in the IGBT life, failures due to IGBT packaging such as, coefficients of thermal expansion and solder fatigue will limit lifetime and must be considered to ensure long IGBT life. Further analysis is required for robust design.
Transient Thermal Impedance and VCE
One IGBT characteristic investigated is the transient thermal impedance (ZthJC). This is the ability of the packaging material to absorb and transfer energy away from the die to the case in a short period of time. ZthJC is defined in terms of degrees K temperature rise seen per watt of power loss at the junction (K/W) and is defined on some datasheets for a given duty cycle (D) (square wave pulse) and pulse duration down to the microsecond level. This time duration is much closer to the conduction time of the auxiliary IGBT in the ICZVT inverter than the IC or ICM datasheet ratings enabling a better look into how to utilize the die size more effectively and optimize cost without sacrificing reliability by exceeding device peak junction temperature. An example transient thermal impedance chart is shown in Figure 8.
Chapter 1 – Introduction
1.2 Application Specifications
1.3 Soft-switching inverters
1.4 Thesis Objective and Overview.
Chapter 2 – The Fixed Timing Coupled Inductor Inverter.
2.1 Prior Art and Principles of Operation
2.2 Timing Parameter Calculations
Chapter 3 – Hardware & Software Design and Implementation
3.1 Various Implemented Hardware
3.2.1 Auxiliary IGBT Selection Foreword
3.3 Auxiliary IGBT “Single Shot” Testing
Chapter 4 – Single Phase Test Setup and Results
4.1 Single Phase Test Setup
4.2 Single Phase Test Results
4.3 Loss Comparison of Hard-Switched v. Soft-Switched Inverter
4.4 Simulation Waveform Comparison
Chapter 5 – Inductor Current Distortion.
5.1 Recreation of Distortion
5.2 Volt-Second Imbalance Root Cause Analysis
5.3 Summary of Simulation Results
Chapter 6 – Closing Remarks and Future Work
6.2 Future Work
GET THE COMPLETE PROJECT
IMPLEMENTATION OF A FIXED TIMING COUPLED INDUCTOR SOFT-SWITCHING INVERTER