SANYO DENKI Technical Report No.6 Nov. 1998
General Theses
Floating Output DC-DC Converter Using Single
Winding Reactor and Its Applications
Hirohisa Yamazaki
1. Introduction
Networking based on a PC server is becoming increasingly popular and
uninterruptible power supplies (UPS) need to be made more compact, lightweight
and cheaper. The typical UPS now functions not only as a backup power supply
during a power failure, but also improves the power factor and cleans the
commercial power supply by absorbing the harmonics that are generated by
equipment acting as loads.
To achieve such targets, various rectifier circuits have been developed for UPSs.
Although Sanyo Denki has been marketing step-up chopper type and high frequency
insulation type products, various problems as discussed in Chapter 2 of this paper
make it difficult to use the products with 200 V input.
We report here on recent developments and a new product using a DC-DC
converter
(1)
that can generate an output that is floating (separated) from the input.
The new product is the 3 kVA UPS using the technology in the rectifier circuit. The
operating system of this DC-DC converter and UPS, and its characteristics are
described.
2. Requirements for Rectifier Circuit of Compact UPS
The typical load of a compact UPS is a personal computer. This means that the
compact UPS supplies AC power to the switching regulator of the power supply
circuit of the personal computer. Therefore, the output of the UPS includes various
harmonics having a high current crest value. At the same time, arresters and ceramic
varistors are used to protect the system from surges, and a high frequency filter is
inserted in the input circuit of the switching regulator to protect the system from
noise. The earth line of the system and the neutral line of the input/output AC
power line should have minimum potential difference so that the circuit components
do not operate on an abnormal potential with regard to the earth potential.
For that purpose, the input circuit and output circuit of a UPS should not be
separated, but the neutral lines of the input and output circuits should preferably be
connected directly.
The circuit in which full wave rectification and a full bridge inverter are combined
in the general purpose inverter is shown in
Fig. 1.
In this circuit, the DC voltage that exists in the middle of the output of the inverter
is super-imposed on the output voltage. This makes it difficult to connect the input
and output because of the potential difference between the neutral line of the input
circuit and output circuit. Therefore, the circuit operation becomes unstable with
regard to the earth potential and this can cause troubles such as burning of the
varistor of the load circuit or troubles due to noise.
In order to connect the neutral lines of the input and output circuits, the input
1 / 5
could be isolated from the output.
The circuit configuration using an insulation transformer in the output circuit is
shown in Fig. 2. The circuit configuration of connecting an insulation transformer
in the input circuit, or inserting a DC-DC converter between the input and output,
has the same result, but insulation transformers are large, heavy and expensive. Use
of the DC-DC converter makes the circuit configuration complicated, and
manufacturing an insulation transformer for a large capacity inverter is difficult and
limits application of this circuit configuration to small capacity inverters.
A circuit configuration using a voltage-doubler rectifier + half-bridge inverter is
shown in
Fig. 3.
The rectifier circuit that is used in Fig. 3, is a high power factor converter circuit.
The DC voltage in the middle of this circuit becomes as follows when taking
fluctuation of the input voltage into consideration:
100V
1.15 2 = 325V
When this circuit is used for 200 V, double the voltage of the above figure becomes
necessary. This means that the semiconductors used in the latter stage of the
inverter must have a high withstand voltage, which reduces the efficiency.
Furthermore, it requires a larger air clearance and a longer creepage distance, which
hinder the design of a compact unit.
In addition, the mid-point voltage of capacity fluctuates in the half-bridge inverter
when an unbalanced load such as a half-wave rectifier load is connected.
Fluctuation of the mid-point voltage of capacity can create deviations of magnetism
and insufficient withstand voltage of the capacitor.
Each circuit system has its own merits and demerits as we have examined. We
have developed a floating output DC-DC converter (FCON) in order to solve these
problems for a UPS that operates on 200 V input and delivers output voltages of 3
kVA or more.
3. FCON Circuit
3.1 Operation of FCON
The FCON circuit that provides a floating output with respect to the input is shown
in
Fig. 4 (A).
The input energy is stored in the reactor L through the two semiconductor switches
Q1 and Q2. The stored energy is discharged through the two diodes D1 and D2. This
circuit operation is the same as that of a flyback converter or of a polarity-inverting
chopper circuit except that an insulating transformer is not necessary. The spike
voltage that is created normally in the polarity-inverting chopper circuit due to its
leakage inductance, is absorbed by Cdc and is suppressed.
The principle of obtaining a floating output using FCON is described below.
Fig. 4 (B) shows the energy storage mode in the reactor L.
When Q1 and Q2 are turned on, current flows into reactor L from the DC power
supply, and stores electromagnetic energy in L.
The capacity of the reactor that is required here is about 4% of the input
impedance, and is fairly small compared with the transformer.
During storage mode, the diodes D1 and D2 are reverse-biased by the reactor L
and the capacitor Cdc voltages, so that any current does not flow into the output
side that is separated from the input side.
When Q1 and Q2 are turned off, the system enters energy discharge mode as
shown in Fig. 4 (C). Energy that was stored in L during the storage mode is now
discharged into capacitor Cdc and the load side circuit through D1 and D2.
Because Q1 and Q2 are turned off during the discharge mode, the output side is
separated from the input side. This separation of the output from the input is
obtained by repeating the two modes of (B) and (C).
2 / 5