Note: Descriptions are shown in the official language in which they were submitted.
CA 02458137 2004-02-17
TITLE OF INVENTION
ZERO VOLTAGE SWITCHED FULL BRIDGE DC/DC CONVERTER
s
TECHNICAL FIELD OF THE INVENTION
The present invention relates generally to isolated DC to DC
io converters and more particularly to buck derived Full Bridge converters,
which feature zero voltage switching of the four controlled power switches.
BACKGROUND OF THE INVENTION
is High frequency switching DC to DC converters are frequently used to
convert DC voltage, provide galvanic isolation of the output from the input
and to regulate the output voltage and current. DC to DC converters are
frequently used as a portion of AC to DC power supplies. For example,
such power supplies are employed in Telecommunication or Cell Site Power
2o Systems to provide isolated and regulated +24Volt or -48Volt power to
system batteries and the paralleled load. DC to DC converters are also
frequently used to provide a second isolated DC voltage from another. For
example, +24Volt Cell Site Transmitter power is converted to -48Volts for
co-located telecommunications equipment by using DC to DC converters.
2s
Of the many topologies that can be used for DC to DC converters,
those that are buck derived are often preferred for medium (+24V, -48V) or
low voltage (+5V, +3V) outputs. This is due to the non-pulsating output
current and ease of control as the output is directly proportional to the duty
3o cycle of the controlled switching devices. It is also preferred to use
i
CA 02458137 2004-02-17
topologies that achieve zero voltage switching (ZVS) of the primary
controlled switches and zero current switching (ZCS) of the secondary
(usually) non controlled switches to maximize converter efficiency and to
minimize Electro-Magnetic Interference (EMI) generation. For low to
s medium output power it is desired to use such a ZVS converter employing
two controlled switches, such as the inventor has previously disclosed in US
Patent 6208529. For high output power (above 2.5kW) a full bridge
converter employing four controlled switches is more desirable as current in
the switching circuit is reduced.
io
A commonly used full bridge ZVS buck derived converter is the
Phase Shifted Bridge (PSB) Zero Voltage Transition (ZVT) Converter such
as disclosed in US patent 4864479 by Robert Steigerwald. A similar
converter is described in Unitrode (Now part of Texas Instruments)
is application note U-136A 'Phase Shifted, Zero Voltage Transition design
considerations and the UC3875 PWM controller' published in May 1997.
Such a PSB Converter can achieve ZVS of the primary switches at full load
but has difficulty achieving ZVS at partial loads. This converter also has a
loss of output voltage due to the commutation time of the load current an
Zo effect, which is worsened if the series commutation inductor value is
increased to extend the ZVS load range. Also during the freewheeling time,
when one of the diagonally opposite switches are off, there is excess
current (which does not contribute to load current) flowing in the primary of
the power transformer and two of the primary switches causing extra power
Zs loss. The rate of turn off (dl/dT) of the output non-controlled switches
(rectifier diodes) is limited by the series commutation inductance but with
typically used inductor values the dl/dT is still very high causing high
reverse recovery current in the diodes which can generate high EMI.
Another method of modulating the ZVS Full bridge converter, disclosed by
3o Barry Blair et al. in US Patent 6483724, involves alternately modulating
the
2
CA 02458137 2004-02-17
top two switches on at nearly 50% duty cycle each and alternately Pulse
width modulating the bottom two switches on anywhere from zero to nearly
50% duty cycle, achieves essentially the same result as Phase Shift
Modulation for Full Bridge Buck derived converters.
s
As a result of the limitations of the commonly used 'Traditional' PSB
Converter many researchers have proposed variations to this converter
topology to improve it's performance. These variations either consist of
adding extra components to the primary side or to the secondary side of the
io PSB Converter.
Of those that add extra components to the primary side, Dong-Yun
Lee, in the publication "An improved Full-Bridge Zero-Voltage-Transition
PWM DC/DC Converter with Zero-Voltage/Zero-Current Switching of the
is Auxiliary Switches", adds an extra leg of controlled switches whose center
point is connected to the center point of one of the original controlled
switch
legs through series connected inductors. Xinbo Ruan et al., in "An
Improved Phase-Shifted Zero-Voltage and Zero-Current Switching PWM
Converter" adds diodes in series with the controlled switches of one of the
Zo bridge legs. Praveen Jain et al. disclose in US Patent 6016258 using two
large valued inductors connecting between each original bridge leg center
point respectively and a center point of a leg of large valued capacitors.
Rajapandian Ayyanar et al. in US Patent 6310785 disclose adding a second
transformer between one of the original bridge leg center points and a
Zs center point of a leg of large valued capacitors. Both Yungtaek Jang et al.
in "A New ZVS-PWM Full Bridge Converter" and lonel Dan Jitaru in "High
Efficiency Converter Using Current Shaping and Synchronous
Rectification", use a coupled inductor connected in series with the
transformer primary.
3
CA 02458137 2004-02-17
Of those that add extra components to the secondary side,
Christopher Henze in US Patent 4953068 discloses changing the output
rectification circuit to a full bridge of controlled switches. Michael Waiters
in
US Patent 5157592 discloses adding controlled saturable reactors in series
s with the transformer secondaries. Both Ju Won Baek et al in US Patent
5886884 and Byeong-Ho Choo et al. in "A Novel Secondary Clamping
Circuit Topology for Soft Switching Full-Bridge PWM DC/DC Converter"
disclose adding two extra diodes, a capacitor and inductor to the output
circuit.
to
All of these previously disclosed circuits either add extra magnetic
elements, or non controlled switches in series with the primary or secondary
current paths or extra controlled switches to the original ZVS Full Bridge
Converter. This significantly increases the size, cost and / or reduces the
is efficiency of the resulting converter from that of the original ZVT Fulf
Bridge
Converter.
The output diodes in a ZVT Full bridge Converter typically have
significant reverse recovery time, which results in energy being stored in the
Zo series commutation inductor. This energy causes large voltage spikes to be
present across these diodes as they turn off and can generate substantial
EMI. William McLyman discloses in US Patent 4245288 a way to reduce
these diode recovery spikes in a Buck Converter by adding a tap-up winding
to the output inductor and connecting a commutation diode to the end of the
Zs tap, however with the tap-up ratio necessary to achieve acceptably fast
current commutation the output current of the Buck Converter will become
pulsating. Also it is desirable to recover the energy stored in the series
commutation inductor as a result of the reverse recovery current in the
output diodes and return it to the source or the Load. For the Traditional
4
CA 02458137 2004-02-17
PSB Converter, Richard Redl discloses in US Patent 5198969 a way to
return the stored energy to the source by adding two diodes.
It is accordingly an object of the invention to provide a new and
s improved isolated Full Bridge Zero Voltage Transition DC to DC Converter
suitable for High output Power and medium and low output Voltages.
An object of the invention is to provide a Full Bridge DC to DC
Converter that can achieve Zero Voltage Switching of the primary side
io controlled switches over a very wide load range, has minimal loss of output
voltage due to the commutation time of the load current, reduces the excess
current flowing in the transformer and primary controlled switches to zero
during the freewheel interval, minimizes the rate of turn off (dl/dT) of the
output diodes and provides a non-pulsating output current to the load. It is
is an object of the invention to achieve the above performance improvements
without adding extra magnetic components, any extra controlled switching
devices, or any additional non-controlled switches in series with the current
paths.
Zo An additional object of the invention is to provide a means to recover
all the stored energy in the commutation inductor as a result of the reverse
recovery current in the output diodes and some of the stored energy in the
leakage inductance of the transformer and return this energy to the source.
Zs An additional object of the invention is to provide a method to control
the energy stored in the inductor which is connected effectively in parallel
to
the transformer primary, to constant so that it can be just sufficient to
achieve Zero Voltage Switching of the controlled switches for different
output voltage adjustment of the DC/DC converter.
S
CA 02458137 2004-02-17
SUMMMARY OF THE INVENTION
In one of it's aspects the invention consists of a switch mode DC to
DC converter comprising of a full bridge arrangement of controlled switches
at the input stage, a commutation inductance connected in series with
transformer primary winding and an inductance placed effectively in parallel
io with the primary winding, and a output stage including a center tapped
secondary winding of the transformer, an output diode connected to each
end of the center tapped winding, the other pole of the output diodes are
connected together and to a first pole of the output capacitor, first output
inductor winding connected between the secondary winding center tap and
is the other pole of the output capacitor, second output inductor winding
connected between the transformer center tap and a third output diode
whose other pole is connected to the common point of the other two output
diodes, third output inductor winding connected between the transformer
center tap and a output filter capacitor whose other pole is connected to the
Zo first pole of the output filter capacitor, and output terminals connected
to
each end of the output capacitor to which the load can be connected.
The primary stage is realized by a full bridge arrangement of
controlled switches. Two controlled switches are connected in series with
Zs another and in parallel with the input terminals and with a further two
controlled switches, which are also connected in series with each other.
Anti-parallel diode means are associated with each switch and are poled to
allow current to flow in a direction opposite to the normal direction of
current
flow in each switch. Capacitance means is also effectively connected in
3o parallel with each switch. An inductance means and optionally also a DC
6
CA 02458137 2004-02-17
blocking capacitor means is connected in series with the transformer
primary winding, the series combination of the inductance means, primary
winding and optional DC blocking capacitor is connected between the
common point of the first two controlled switches on the one hand and the
s common point of the further two controlled switches on the other hand.
In another aspect of the invention one or more additional windings are
added to the transformer, the windings if more than one are connected in
parallel with each other, a small valued capacitor is connected in series with
io these windings. The series combination of windings) and capacitor is
connected to a half bridge or full bridge diode rectification circuit
connected
in parallel with the input terminals of the converter.
Another aspect of the invention is a method and apparatus to control
is the peak current in the inductance connected effectively in parallel with
the
primary of the transformer. The output voltage of the converter is measured
directly or indirectly by means of an additional winding in the transformer,
whose voltage is rectified and averaged to provide a representation of the
output voltage. The operating frequency of the converter is then varied, by a
ao suitable circuit, to be inversely proportional to this measured voltage.
This
keeps the volt seconds integral across the inductance constant, thus the
peak current constant despite the output voltage being varied due to
adjustment or if the DC to DC converter enters into current limit condition.
CA 02458137 2004-02-17
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 is a circuit diagram of one embodiment of the invention
s FIG. 2 is a circuit diagram of the invention with the primary DC
blocking capacitor.
FIG. 3 is a circuit diagram of the invention with a half bridge energy
recovery circuit.
io
FIG. 4 is a circuit diagram of the invention with a full bridge energy
recovery circuit.
FIG. 5 is a circuit diagram of the invention with a output voltage
is simulation winding and with a full bridge energy recovery circuit.
FIG. 6 is a circuit diagram of a control circuit for the invention
including the invented method of controlling the peak current in the parallel
inductance Lp to be constant despite the output voltage on the converter
ao being changed.
FIG. 7 is a set of voltage waveforms for the embodiment of the
invention of FIG. 5.
is FIG. 8 is a set of current waveforms of the embodiment of the
invention of FIG. 5.
s
CA 02458137 2004-02-17
DETAILED DESCRIPTION OF THE PREFERRED
AND OTHER EMBODIMENTS
The basic schematic of one embodiment of the invention is shown in
s FIG. 1. The converter is connected to a voltage source Vs by input
terminals. Two sets of half bridges of controlled switches are connected to
the input terminals and thus are connected in parallel with the voltage
source Vs. The first half bridge consists of a series combination of
controlled switches S1 and S2, which are typically MOSFET's. These
io switches have diodes Ds1 and Ds2 placed across them in such a direction
that allows current to flow in the opposite direction to the switches. If
MOSFET's are used for S1 and S2 then integral diodes inherent in them
can be utilized as Ds1 and Ds2. Capacitors Cs1 and Cs2 are placed across
S1 and S2 respectively and similarly can be partly or completely realized by
is the MOSFET capacitance. The second half bridge similarly consists of
controlled switches S3 and S4, diodes Ds3 and Ds4 and capacitors Cs3
and Cs4. The switches S1, S2, S3, and S4 are turned on and off by drive
signals Drive A, Drive B, Drive C, and Drive D respectively. The drive
waveforms are organized so that only one switch of a series combination is
Zo on at any one time and a short dead time delay is imposed between one
switch turning off and the other turning on.
The series combination of a small valued inductor Lc and primary
winding, with turns Np, of transformer T1 is connected between the
Zs common point of the switches S1 and S2 and the common point of switches
S3 and S4. In this and the second embodiment described herein, inductor
Lr can be partly or mainly realized by the leakage inductance of the
transformer. A large valued inductor Lp is placed across the primary
winding of the transformer. Typically this inductance is realized by gapping
3o the core of the transformer to reduce the transformer primary magnetizing
9
CA 02458137 2004-02-17
inductance to the desired value, though it could also be realized as a
separate inductor.
The secondary of transformer T1 consists of two windings, of identical
s number of turns Ns, which are connected in series and phased as shown in
the figure. The two opposite ends of the windings are connected to the
anode poles of first and second diodes D1 and D2 respectively. The
common point of the secondary windings of T1 is connected to one end of
each of the three windings of inductor Lo, of N1, N2 and N3 turns, which are
io phased as shown in the figure. The other end of the second inductor
winding with N2 turns is connected to the anode pole of third diode D3. The
cathode poles of the diodes D1 D2 and D3 are connected together and to
one end of output capacitor Co and to one output terminal. The other end
of the main (first) winding of inductor Lo, of turns N1, is connected to a
is inductance LI. The other end of inductance LI is connected to the other end
of output capacitor Co and a second output terminal. The other end of the
third inductor winding of N3 turns is connected to one pole of capacitor Cf,
which can be of a value much smaller than Co. The other pole of Cf is
connected to the cathodes of D1 D2 and D3. The inductance LI can be
Zo realized by the leakage inductance between the main and third windings or
by a discrete inductance. This leakage inductance could be increased by
adding a second core to the inductor encompassing only the main winding,
however even a small value of inductance achieved by normal separation of
the windings works very well as long as the windings N2 and N3 are closely
Zs coupled. Typically the turns of the third winding (N3) are equal to the
turns
of the main winding (N1 ). The turns of the winding connected to third diode
D3 are typically half that of the other two windings.
The embodiment of FIG. 2 is identical to the first embodiment except
3o that a DC blocking capacitor Cb is added in series with the inductor Lc and
io
CA 02458137 2004-02-17
the parallel combination of the primary winding of T1 and the inductance Lp.
This series combination can be in any order. The capacitor Cb is useful to
block any residual DC voltage or DC current, which can otherwise result
from imbalances in the drive times, from being impressed across the
s primary of transformer T1 or the inductance Lp.
In the embodiment of FIG. 3 a forth winding of turns Nr is added to
the transformer T1 to allow clamping of the peak voltage across of D1, D2
or D3 when they turn off. A relatively small valued capacitor Cr is
io connected in series with this winding, the other end of which is connected
to
one of the input terminals. The other end of the winding is connected to
diodes Dr1 and Dr2 whose other poles are connected to the input terminals
in such a polarity to block the applied voltage Vs. This additional circuit
returns the energy stored in the commutation inductance Lc, due to the
is reverse recovery current of diodes D1, D2 and D3, back to the source
improving the converter efficiency and minimizing the excess voltage spikes
on diodes D1, D2 and D3 when they turn off. In this embodiment it is
preferred that the majority of the inductance Lc be realized by a discrete
inductor and that the leakage inductance of T1 is minimized. This can be
Zo achieved by realizing the primary or secondary windings) of the transformer
as two portions, connected in series or parallel, sandwiching the other
windings in between the two portions. The relatively small value of Cr
prevents flow of excess current during transient conditions and aids in
reducing the spike voltage on D1, D2 and D3 due to the unclamped leakage
Zs inductance of T1. This forth winding can be realized by two windings
connected in parallel if the transformer primary or secondary windings) is
actually in two portions to create a sandwich structure. In this embodiment
the number of turns Nr will typically be half that of the primary winding,
though a slight increase of turns from this value will improve clamping action
3o at the expense of more energy being returned to the source.
n
CA 02458137 2004-02-17
FIG. 3A shows essentially the same embodiment as FIG. 3 except that the
function of Capacitor Cr is realized by two Capacitors Cr1 and Cr2 both
connected to the first end of the forth transformer winding. The other pole of
s the capacitor Cr1 is connected to one input terminals and the other pole of
Cr2 is connected to the other input terminal. Capacitors Cr1 and Cr2 each
have half the value of Cr of FIG. 3 for the same effective operation. This
embodiment would be preferred over that of FIG.3 in that its clamping action
at the initial start up time of the converter will be balanced.
to
The embodiment of FIG. 4 is similar to the embodiment of FIG. 3
except that the other end of capacitor Cr is connected to the common point
of a series connection of diodes Dr3 and Dr4 whose other poles are also
connected to the input terminals in such a polarity to block the applied
is voltage Vs. In this case the turns of winding Nr will be equal or slightly
more
than the primary winding turns Np. This embodiment returns the energy
stored in the inductor Lc and clamps the spike voltage across D1, D2, and
D3 in much the same manner as that of that of FIG. 3.
2o Yet a further aspect of the invention is a method and apparatus to
control the peak current in the inductor Lp to a constant value despite the
output voltage of the converter being varied by adjustment of the output
voltage to different values or by the converter entering into a current
limiting
control region. This method entails measuring the output voltage of the
2s converter directly, (which is easily done if the control circuit is on the
secondary side of the converter) or indirectly using an isolated observer
circuit (needed if the control circuit is on the primary side) comparing it to
a
reference voltage with a error amplifier of fixed gain and generating a signal
which reduces the operating frequency of the converter proportional to the
30 output voltage for the normal output voltage range of the converter. Also
12
CA 02458137 2004-02-17
included in this method is a means to keep the loop gain and operating
pulse width or duty cycle of the converter constant despite the frequency
being varied. A current source controlled by the same error amplifiers
injects a current into the Phase Shift (or PWM) ramp capacitor of the control
s circuit to keep the peak ramp voltage constant despite it's ramp time being
increased due to lower operating frequency.
In the embodiment of FIG. 5 a fifth winding is added to allow
observation of the output voltage of the converter galvanically isolated from
to the actual output. Typically this winding only has one turn Nvs. The
winding is connected to a half wave rectification circuit (Though a full wave
rectification circuit could be used) consisting of Dvs1 and Dvs2 and a
voltage averaging circuit consisting of RC time constant Rvs and Cvs much
larger than one period of the switching circuit. The output voltage of this
is circuit portion Vosim is proportional to the output voltage of the
converter Vo
except for light loading of the converter whereupon the simulated output
voltage is reduced. This does not adversely effect the operation of the
converter as the resulting lower operating frequency extends the load range
of ZVS of the switches.
FIG. 6 shows a circuit, which can implement the method of keeping
the peak current in the inductor Lp to a constant value despite the output
voltage of the converter being varied. The circuit is designed to work with a
Phase Shift Modulation (PSM) Control IC such as Unitrode UCC2895, but a
2s similar circuit could be devised to work with other PSM or any Pulse Width
Modulation Control IC's. The output voltage of the converter is measured
directly using a resistive divider or indirectly using a circuit added to the
converter as shown in FIG. 5. This input signal Vosim is buffered by an
error Operational amplifier configured as a voltage follower. The input
3o signal is then compared to a reference voltage (Pin 8 of UCC2895) by an
13
CA 02458137 2004-02-17
Operational amplifier with fixed gain determined by input and feedback
resistors. A control signal, which varies the operating frequency of the PSM
control IC, is coupled to the frequency determining pin Rt (Pin 8 of
UCC2895) by a resistor and diode. A current mirror also controlled by the
s error Operational Amplifier injects a current into the ramp capacitor
connected to the Ramp pin (Pin 3 of UCC2895).
The circuit operation of the last shown embodiment (FIG 5) being a
preferred embodiment of the invention is described following with reference
to to FIG. 7 and FIG. 8, representing the voltage and current waveforms of the
converter for full load operation. In addition the phase shift control method
for the full bridge ZVS converter will be used in the description though other
known methods could be used.
is The gate drive voltage waveforms Drive A and Drive B for controlled
switches S1 and S2 are shown in FIGS. 7a and 7b respectively. They are
non-overlapping complementary rectangular waveforms of zero volts or
typically 12 Volts of an equal duty cycle of a little less than 50%. The gate
drive waveforms Drive C and Drive D for switches S3 and S4 are shown in
ao FIGS. 7c and 7d respectively. These are also non-overlapping
complementary waveforms of an equal duty cycle of a little less than 50%.
The two sets of gate drive waveforms are phase shifted with respect to each
other by a phase shift angle Phi 1.
Zs FIG. 7e shows the voltage at the common point of controlled switches
S1 and S2, Vab. When Drive A is high switch S1 is on causing Vab to be
equal to the voltage at the negative input terminal (zero volts). When Drive
B is high switch S2 is on causing Vab to be equal to the voltage at the
positive input terminal Vin. The gate drive dead time when both Drive A and
3o Drive B are low is selected to be slightly greater that the time for Vab to
14
CA 02458137 2004-02-17
slew naturally, due to the current in the inductor connected effectively in
parallel with the primary of the transformer, from one level to another so
that
both switches turn on only when the voltage across them is zero volts
achieving ZVS. In FIGS. 7 and 8 this time is exaggerated somewhat from
s what would typically occur for the sake of clarity.
FIG. 7f shows the voltage waveform at the common point of
controlled switches S3 and S4 Vcd, which is identical to Vab except that it is
phase shifted by angle Phi 1. FIG. 7g shows the voltage waveform Vab-Vcd
to that is applied to the series combination of inductor Lc and the
transformer
primary Np and paralleled inductor Lp. This is a bipolar pulse waveform
with pulse voltage of plus and minus Vs and duty cycle Phi 1/Period. FIG.
7h shows the voltage waveform applied to the primary winding of the
transformer. This waveform has the same amplitude as the waveform Vab-
is Vcd but has slightly lower duty cycle angle Phi 2 due to the commutation
time required to ramp the current up in the inductor Lc. It also has an
overshoot after the turn off time of diode D3 (t2-t3 and t9-t10). The clamp
winding Nr and associated components clamps this overshoot voltage. The
ramp of voltage during the clamp time is caused by the relatively small
Zo value of Cr.
The voltage across the secondary windings of the transformer is
essentially the same shape as the primary except that it's amplitude will be
dependent on the turns ratio of the transformer, Ns/Np. This secondary
Zs waveform is rectified by the main secondary rectifier diodes D1 and D2 to
produce a uni-polar pulse train voltages. The voltage waveform across
diode D1 is shown in FIG. 7i. It is ideally zero volts when diode D1 is
conducting and equal to -2Vin*Ns/Np during the phase angle Phi 2. During
the other times it is at an intermediate value determined by the windings on
3o the ouput inductor Lo. During normal operation there is a period when both
is
CA 02458137 2004-02-17
diodes D1 and D2 are not conducting and then both waveforms are at close
to zero volts at the same time. The voltage waveform across diode D2 is
shown in FIG. 7j. It is identical to the voltage waveform across D1 except
that it is phase shifted 180 degrees. The voltage waveform on the
s freewheeling diode D3 is shown in FIG. 7k. It is ideally zero volts when D3
is conducting and is Vin*Ns/Np*(N2+N1 )/N1 when the diode is non-
conducting. There is also an overshoot of this voltage waveform when
diode D3 stops conducting, the majority of which is clamped by the
clamping winding Nr of the transformer and it's associated components.
io The lower initial clamp voltage caused by the relatively small value of Cr
reduces a leading edge spike on this overshoot, resulting from the
unclamped portion of the transformer leakage inductance.
The output inductor Lo and output capacitor Co average the voltage
is waveform across D3 to realize the output voltage of the converter Vo that
is
present across the output terminals and connected load. The voltage
waveform present at the common point of the two secondary windings is
shown in FIG. 71. It is similar in shape to the voltage across D3 but is
reduced in amplitude by the ratio of the inductor windings N1/(N1+N2). The
Zo average of this waveform is also the output voltage Vo and the voltage
during the freewheel period is VO*N2/(N1+N2) which is the voltage that
reverse polarizes Lc commutating the Diodes D1 and D2 off. The voltage
across capacitor Cf is equal to the DC output voltage with a much higher
ripple voltage component.
2s
The current waveforms for the converter are shown in FIG. 8. The
currents flow as a result of the applied load and the voltage waveforms
impressed across the various components thus they will be explained in the
reverse order. FIG. 8m shows the output current I(RI) that flows through the
30 load is inversely proportional to the applied load resistance. Also shown
is
is
CA 02458137 2004-02-17
the current in the output inductor main winding I(N1 ) which is equal to the
output current plus a small ripple current component that is typically less
than 5% of the load current. This ripple current flows through the output
capacitor Co. The current in the filter winding I(N3) is also shown in FIG.
s 8m. This is an AC current, of RMS value of typically 10 to 15% of the output
current of the converter, which flows through filter capacitor Cf.
FIG. 8i shows the current which flows in output inductor winding of N2
turns and the freewheeling diode D3. This current is of a trapezoidal uni-
io polar wave shape which flows when the main output diodes are off or are
turning on or ofF. Its peak value is typically two-thirds of the peak value of
the current in the main output diodes. Figures 8j and 8k show the
waveforms of current in the two main output diodes D1 and D2 respectively.
These currents also flow in the secondary windings Ns1 and Ns2 of the
Is transformer respectively. These are two uni-polar trapezoidal waveforms
identical except for being phase shifted 180 degrees with respect to each
other. These currents start to flow when diagonally opposite switches S1
and S4 or S3 and S2 are on. The currents ramp up at a rate determined by
the value of inductor Lc and reach a peak values essentially equal to the
Zo output of the converter. The current ramps down at a rate also controlled
by
the ratio of windings N1 and N2 of the ouput inductor when one of the
diagonally opposite switches turns off. A small reverse recovery current of
the output diodes is also shown to flow for a short time before the diodes
turn off. It is important to note that in normal operation there is a time
during
Zs which both main output diodes are off and both the transformer secondary
currents are zero.
FIG 81 shows the waveform of the current in the primary winding Np
of the Transformer. This is an AC current equal to the difference of the two
30 output diode currents times the transformer turns ratio Ns/Np plus small
m
CA 02458137 2004-02-17
current pulses which result due to the action of the clamp winding Nr and
associated components.
FIG. 8h shows the current in the primary inductor Lp. This AC current
s is of a flat topped triangular shape, tamping up or down when diagonally
opposite switches are on and remaining at a constant value when one of the
switches turns off. This current is used to provide zero voltage switching of
the switches and can be controlled, by the method described herein, to be
the optimum value to do so despite the output voltage of the converter being
io varied.
FIG. 8g shows the current that flows in the clamp winding Nr, through
capacitor Cr and the diode bridge consisting of Dr1 to Dr4. This current is
a series of small pulses that returns the energy stored in Lc, as a result of
is the reverse recovery current of the main output diodes, to the source Vs.
FIG. 8e shows the current that flows through the capacitances Cs1
and Cs2 (half in Cs1 and half in Cs2) during the time that switches S1 and
S2 are off. These currents are small pulses start at the value of current in
2o Lc and decay to near zero when the voltage Vab has stewed to the other
rail. FIG. 8f shows the current that flows in capacitances Cs3 and Cs4
during the time that both switches S3 and S4 are off . This current is of
much larger pulse size decaying only slightly during the stewing time of the
Vcd voltage waveform.
FIG's. 8c and 8d show the current waveforms in the parallel
combination of S4 and Ds4 and the parallel combination of S3 and Ds3
respectively. These waveforms are identical except they are phase shifted
180 degrees. The current is initially at a large negative value flowing
3o through the diode when the voltage across the switch becomes zero, then
is
CA 02458137 2004-02-17
decays to the value of the peak current in Lp as the respective main output
diode turns off, then stays at this value until the other diagonally opposite
switch is turned off. It then ramps positive as the current ramps up in Lc
and becomes equal to the sum of the transformer primary current I(Np) and
s the current in Lp as the other main output diode takes over the load. It
drops sharply to zero when the respective switch S4 or S3 is turned off.
FIG's. 8a and 8b shows the waveform of the current in the parallel
combination of S1 and Ds1 and the parallel combination of S2 and Ds2
to respectively. These waveforms are identical except they are phase shifted
180 degrees. The current is initially at a small negative value when the
voltage across the switch reaches zero then ramps positive and becomes
equal to the sum of the transformer primary current I(Np) and the current in
Lp, as the current ramps up in Lc and the respective main output diode
is takes over the load. It ramps down to the value of current in Lp only, as a
result of the action of winding N2 of the output inductor when the diagonally
opposite switch turns off. It then drops sharply to zero when the respective
switch S1 or S2 turns off.
Zo As can be seen by the voltage and current waveforms that the
described DC/DC converter achieves zero voltage switching of the primary
side controlled switches at full load. As the main output diodes are
commutated off before the diagonally opposite switches are both turned off
and that the Bridge leg voltages are stewed by the current in Lp (which is
Zs independent of the load) it can also achieve zero voltage switching over a
very wide load range, typically down to 10% load. The converter has
minimal loss of output voltage due to the commutation time of the load
current as the current to commutate is reduced as the main output diodes
are commutated off already and loss of output voltage only occurs as a
3o result of the time to commutate the diodes on. The output voltage is infact
19
CA 02458137 2004-02-17
increased, by the action of the winding N2 of the output inductor, over that
of the 'Traditional' Phase Shifted Bridge Converter. The converter
minimizes the rate of turn off (dl/dT) of the output diodes achieving Zero
current switching of all output diodes It provides a non-pulsating current
s due to the action of filter winding N3 of the output inductor and capacitor
Cf .
ft can be also seen that the converter has no additional controlled switches,
and no extra non-controlled switches in series with the current path, extra
diode D3 is in parallel with the current path so no loss in efficiency
results.
The described converter also has no additional magnetic components, only
io two extra windings in the output inductor. Additionally a means is
provided,
by transformer winding Nr and associated components, to recover the
energy stored in the commutation inductance and part of the transformer
leakage inductance caused by the reverse recovery of the output diodes
and return this energy to the source. This means also clamps the reverse
is spike voltage on the main output diodes. Additionally a method is described
herein which allows control of the peak current in the parallel inductance Lp
to be at a constant value, optimum to achieve ZVS of the controlled
switches despite the output voltage of the converter being varied by
adjustment.
It will be appreciated by those skilled in the art that modifications to
the preferred embodiments described herein, including electrical
equivalents, may be made without departing from the principles of the
invention or the scope of the claims. For example the polarity of the output
2s diodes and l or controlled switches could be reversed thereby allowing
operation of the circuits with negative outputs or inputs. Also additional
circuit elements, well known by those skilled in the art, such as RC dampers
on certain elements may be desired to achieve optimum operation of the
invention. In addition a saturable reactor or ferrite beads may be added in
CA 02458137 2004-02-17
series with the output diodes to reduce their reverse recovery current
without departing from the scope of the invention or the claims.
21
