Integrated Magnetic Transformer for ZVS Phase
Shift Full Bridge Converter
A Dissertation submitted to the Faculty of the Graduate School
of the Gyeongsang National University
By Xinlan Li
In partial fulfillment of the requirements
for the degree
of Master of Engineering
July, 2008
Abstract
Integrated Magnetic Transformer for ZVS Phase
Shift Full Bridge Converter
Xinlan Li
Department of Electrical Engineering
Graduate School
Gyeongsang National University
Supervised by Professor
Hwi-Beom Shin
Integrated magnetic (IM) has been widely used in power electronics system.
The design of IM transformer is considered in a zero voltage switching (ZVS)
phase shifted full bridge converter. The purpose of this design is to improve the
performances, like power density, packaging, size and cost of IM transformer
that can be applied to high frequency converter.
In this dissertation a new IM transformer is proposed. The transformer is
located on the center leg; the output inductor is located on the outer legs with
air gaps. The design equations and equivalent magnetic circuit are presented.
The analysis of each timing interval is included for ZVS. For reducing the core
size, EE core is redesigned. The applications of these circuit are verified on a
1.2kW prototype converter through experimental results. The performance of
the proposed IM transformer is compared with conventional IM transformer.
The efficiency of proposed IM is similar to the one of conventional IM but the
size of proposed IM is reduced up to 20%. Therefore, with the proposed IM
transformer, the power density can be largely increased.
I. Introduction
The IM transformers begin to be used in the low output voltage and high
current power electronics system [1]. With the IM technique, two or more
magnetic components are integrated in one magnetic structure. The EE or EI
core is commonly utilized. The magnetic parts in the converter may be reduced,
so that the size and power density can be improved. In some case, overall
efficiency may be increased and an EMI problem may be reduced.
For a PSFB converter, several IM transformers are proposed [2]. The
transformer winding is typically located on the two outer legs in series
connection and the inductor is wound on the center leg. The AC flux of
transformer and a half of the DC flux of inductor flows in outer legs but the
inductor flux only flows in the center leg. The flux density in outer legs is
much higher than one in center leg because the center leg has nearly two time
cross-sectional area of the outer leg. This IM transformer has a high
magnetizing inductance. Since the transformer coupling is low, EMI may be
high.
A new IM transformer is proposed for ZVS PSFB converter in this paper.
The transformer winding is located on the center leg for close magnetic
coupling and the inductor is wound on two outer legs in series connection. The
proposed IM transformer is analyzed electrically and magnetically. An E-core
is redesigned and implemented.
The proposed IM transformer is
experimentally compared with the conventional one through a 1.2kW prototype
ZVS PSFB converter and the performance is discussed.
II. ZVS PSFB converter and IM transformer
ZVS PSFB converter is shown in Fig. 1. The center-tapped transformer
and the output inductor are usually integrated in one EE or EI core. Fig. 2(a)
shows the proposed IM transformer and the circuit connection in the output
stage. The transformer and output inductor are located on the center leg and
two outer legs, respectively. The conventional IM transformer given in [2] is
shown in Fig. 2(b).
2.1 Operation Principle
When the PSFB converter operates in ZVS, there are 8 operating modes during
each cycle as shown in Fig. 3. The ZVS transition intervals are magnified in
Fig. 3 and these intervals can be neglected when analyzing the magnetic circuit.
Hence, 4 modes only are considered for analysis. The duty ratio D is limited
below 0.5 because of the alternating current.
The magnetic states within each core leg are modeled by using the capacitive
modeling method [3] ~ [5]. For simplicity, all the devices are assumed to be
ideal. The permeance of core is assumed to be infinite. The operation is
explained as follows.
