Note: Descriptions are shown in the official language in which they were submitted.
CA 02220747 1997-11-10
DC-DC CONVERTERS
This invention relates to DC-DC converters, and is particularly concerned with
single ended forward converters providing reduced power losses.
Back~round of the Invention
Single ended forward converters are well known, and it is also well known that in
such a converter it is necessary to provide for resetting of the transformer to avoid
saturation of the transformer core. Various forms of reset circuit are known, including
active reset circuits in which magnetizing energy is conserved rather than being dissipated
in a resistor, in order to reduce power losses in the converter, and in which the duty cycle
10 of the main switching transistor can exceed 50%. For example, such converters and reset
circuits have been described in "High Power SMPS Require Intrinsic Reliability" by
Bruce Carsten, Proceedings of PCI '81, September 1981, pages 118-133 and in "Design
Techniques For Transformer Active Reset Circuits At High Frequencies And Power
Levels" by Bruce Carsten, High Frequency Power Conversion Proceedings, May 1990,15 pages 235-246.
Distributed power supplies, including low output voltage supplies, constituted by
such converters are increasingly required to operate with high efficiency, and hence low
power losses, and at a high switching frequency to minimi7e size and power dissipation.
However, increased switching frequencies result in increased power losses in the form of
switching losses of active devices such as switching transistors. Although single ended
forward converters with active reset circuits can be intended to operate with zero voltage
switching in order to reduce switching losses, this is not necessarily achieved especially
with varying load conditions and with a high switching frequency which makes very
precise timing necessary for controlling the switching transistors.
An object of this invention is to provide an improved single ended forward
converter.
Summary of the Invention
One aspect of this invention provides a method of operating a single ended
forward converter comprising a primary switch in series with a primary winding of a
transformer, comprising the steps of, for each switching cycle of the primary switch:
coupling a resonant circuit comprising a capacitance and a first inductor to the transformer
by closing an auxiliary switch whereby energy is transferred from the transformer and the
capacitance to the first inductor in a resonant manner whereby a voltage of the capacitance
is reduced to substantially zero; closing the primary switch when the voltage of the
capacitance is substantially zero; opening the auxiliary switch to recover energy stored in
the first inductor via a second inductor coupled to the first inductor; and opening the
primary switch.
CA 02220747 1997-11-10
Another aspect of the invention provides a method of operating a single ended
forward converter comprising the steps of, for each of successive switching cycles of the
converter: closing an auxiliary switch to resonate a capacitor with a first inductor, the
capacitor being coupled to a combination of a primary switch and a primary winding of a
S transformer coupled in series to a voltage source, whereby a voltage across the primary
switch is reduced to substantially zero; closing the primary switch when said voltage
across it is approximately zero; and opening the auxiliary switch and the primary switch.
The auxiliary switch can be opened at substantially the same time as or soon, e.g.
immediately, after closing of the primary switch, or it can be opened later after closing and
10 no later than opening of the primary switch.
The time when the voltage across the primary switch is reduced to substantially
zero can be modified by an inductive component coupled in series with a winding, for
example the secondary winding, of the transformer.
A further aspect of the invention provides a method of operating a DC-DC
15 converter to provide substantially zero voltage switching to close a primary switch in
series with a primary winding of a transformer, comprising the steps of first closing an
auxiliary switch to resonate an inductor with a capacitor coupled to a winding of the
transformer whereby a voltage across the primary switch is reduced, and closing the
primary switch when the voltage across it is substantially zero.
Another aspect of this invention provides a method of operating a single ended
forward converter including a primary switch in series with a primary winding of the
transformer, comprising the steps of: providing a capacitor coupled in parallel with the
primary switch or a winding of the transformer; and resonantly removing energy from the
capacitor to an inductor when the primary switch is open.
This method preferably includes the step of timing closing of the primary switch to
be when a voltage across the primary switch is reduced to substantially zero by the
resonance of the capacitor with the inductor, to achieve zero voltage turn-on of the primary
switch. This method preferable further includes the step of recovering energy from the
inductor by switching current through the inductor with an auxiliary switch, and coupling
30 energy from the inductor via a second inductor, inductively coupled to said inductor,
when the auxiliary switch is open.
The invention also provides a single ended forward converter comprising: a
primary switch; a transformer having a primary winding coupled in series with the
primary switch to supply voltage terminals and a secondary winding coupled via a35 rectifying and smoothing circuit to output voltage terminals; and a first inductor and an
auxiliary switch coupled in series, the f1rst inductor and auxiliary switch being connected
in parallel with the primary switch or a winding of the transformer; the converter including
an inductance comprising at least a leakage inductance of the primary and secondary
CA 02220747 1997-11-10
windings and a capacitance comprising at least a parasitic capacitance in parallel with the
primary switch or a winding of the transformer.
Conveniently said capacitance comprises, in addition to parasitic capacitance, acapacitor coupled in parallel with the primary switch or a winding of the transformer,
5 and/or said inductance comprises, in addition to the leakage inductance, an inductive
component coupled in series with a winding of the transformer.
The converter preferably includes a second inductor inductively coupled to the first
inductor for recovering energy from the first inductor when the auxiliary switch is open.
The second inductor can be coupled via a diode either to the supply voltage terminals or to
10 said capacitance.
The invention further provides a single ended forward converter comprising: a
transformer having a primary winding and a secondary winding; a primary switch
connected in series with the primary winding for coupling the primary winding to a
voltage source; an output circuit comprising forward and free-wheel diodes, an output
15 inductor, and a smoothing capacitor coupling the secondary winding to output voltage
terminals; and an auxiliary circuit comprising: a first inductor; and an auxiliary switch
controlled for each switching cycle of the converter to open before opening of the primary
switch and to close not before opening of the primary switch and not after opening of the
primary switch, the auxiliary switch when closed coupling the first inductor in parallel
20 with the primary switch or a winding of the transformer.
The converter preferably includes a capacitor coupled in parallel with the primary
switch or a winding of the transformer. The converter preferably further includes a
second inductor inductively coupled to the first inductor, and a diode in series with the
second inductor for coupling the second inductor to the capacitor or to the voltage source
25 for returning recovered energy thereto. The converter can advantageously further include
an inductive component coupled in series with the primary or secondary winding of the
transformer.
Thus different aspects of the invention provide for zero voltage switching of the
primary switch and/or recovery of magnetic energy in a single ended forward converter,
30 using resonant characteristics of the converter which involve leakage inductance of the
transformer and/or an inductive component in series with one of its windings, and
parasitic capacitance and/or a capacitor coupled in parallel with the primary switch or one
of the transformer windings.
The invention facilitates the provision of a single ended forward converter which
35 has the advantages of relatively lossless switching independent of supply and load
conditions, simple power and control circuitry, no increase in conduction losses between
switching times, and ability to operate in either current or voltage mode control.
CA 02220747 1997-11-10
Brief Description of the Drawings
The invention will be further understood from the following description with
reference to the accompanying drawings, in which:
Fig. 1 schematically illustrates a single ended forward converter in accordance
S with an embodiment of this invention;
Fig. 2 illustrates operating waveforms of the converter of Fig. 1; and
Fig. 3 schematically illustrates a modified single ended forward converter in
accordance with another embodiment of this invention.
Detailed Description
Referring to Fig. 1, a single ended forward converter is illustrated for providing
from an input or source voltage +V supplied at DC supply terminals 10 a desired output
voltage at DC output terminals 12 for supply to a load (not shown). The magnitude of the
source voltage +V is arbitrary; this may, for example, be in a range from 30 to 70 volts.
An input capacitor 14 connected between the terminals 10 ensures a low impedance source
15 for the converter. The magnitude of the output voltage is also arbitrary; this may, for
example, be S volts. A control circuit (not shown) serves to supply a pulsed control
signal G1 to the converter for m~int~ining this output voltage at its desired level in known
manner, for example using pulse width modulation in a current mode or voltage mode
feedback arrangement. The signal G1 typically has a high frequency, for example a fixed
20 frequency of 300 kHz, to permit the converter to be implemented using components of
relatively small size.
The converter includes a magnetic core transformer 16 having a primary winding
18 and a secondary winding 20 the senses of which are represented conventionally in
Fig. 1 by dots adjacent to the windings. The primary winding 18 is connected in series
25 with the drain-source path of an N-channel MOSFET 22 (each of the MOSFETs in Fig. 1
is illustrated as including its parasitic or body diode connected in parallel with the drain-
source path of the MOSFET) between the terminals 10. This series circuit may also
include a low impedance resistor (not shown) for current sensing. The MOSFET 22
constitutes a primary switch of the converter, and is controlled by the signal G1 supplied
30 to its gate.
The secondary winding 20 of the transformer 16 has a first end connected via a
forward diode 24 and an output inductor 26 to one of the terminals 12, and a second end
connected via an inductance 32 to the other of the terminals 12, which conveniently may
be grounded as illustrated in Fig. 1. An output capacitor 28 is connected between the
35 output terminals 12, and a free-wheel or catch diode 30 is connected between the
grounded one of the terminals 12 and the junction between the diode 24 and inductor 26.
The inductance 32, which is discussed further below, can alternatively be connected in
series with the diode 24 on the first side of the winding 20, and/or can be partly or wholly
CA 02220747 1997-11-10
constituted by leakage inductance of the transformer 16 so that a separate inductive
component is not required. Thus the inductance 32 represents the leakage inductance
between the primary and secondary windings of the transformer 16, optionally with an
additional inductive component.
S Except in respect the inductance 32, the single ended forward converter as
described above has a known form and operates in known manner which need not be
further described here. As is well known, such a single ended forward converter also
requires some circuitry or mechanism for resetting the transformer core to avoid magnetic
saturation and to permit use of a physically small transformer. Typical of such circuitry in
10 the prior art is an active reset circuit as described in the Carsten articles referred to above.
The single ended forward converter illustrated in Fig. 1 includes an auxiliary
circuit 34, shown within a dashed-line box, which with the inductance 32 operates as
described below to reset the transformer on each cycle of the control signal and to provide
other advantages such as facilitating zero voltage switching of the primary switch 22 of the
15 converter, and hence low switching losses.
The circuit 34 comprises a capacitor 36, an auxiliary switch constituted by an
N-channel MOSFET 38, two coupled inductors 40 and 42, and a diode 46. The coupled
inductors 40 and 42 effectively constitute primary and secondary windings respectively of
a transformer 44 and have the relative senses shown in Fig. 1 by dots, and accordingly are
20 referred to below as primary and secondary inductors. The capacitor 36 is connected in
parallel with the primary switch constituted by the drain-source path of the MOSFET 22,
and hence in series with the primary winding 18 between the terminals 10. The drain-
source path of the MOSFET 38 is connected in series with the primary inductor 40 across,
i.e. in parallel with, the capacitor 36. An auxiliary control signal G2 is supplied from the
25 control circuit (not shown) to the gate of the MOSFET 38. The secondary inductor 42 is
connected in series with the diode 46 between the terminals 10, the diode being poled to
prevent a short circuit of the supply voltage by the inductor 42.
In order to avoid undesired oscillation under certain operating conditions, the
circuit 34 can also optionally include a diode 48, shown in dashed outline to indicate its
30 optional presence, connected in series with the drain-source path of the MOSFET 38 and
the primary inductor 40, poled oppositely to the body diode of the MOSFET 38.
The operation of the converter of Fig. 1 is described in detail below with additional
reference to the waveforms illustrated in Fig. 2. These waveforms comprise, for one
cycle at the switching frequency in normal operation of the converter, the control signal
35 Gl for the primary switch constituted by the MOSFET 22, the voltage Vl across the
drain-source path of this MOSFET 22 (and hence across the capacitor 36) and the current
Il through this path, the control signal G2 for the auxiliary switch constituted by the
MOSFET 38, the voltage V2 across the drain-source path of this MOSFET 38 and the
CA 02220747 1997-11-10
current I2 through this path (and hence also through the primary inductor 40), the current
I3 through the inductance 32, and the current I4 through the secondary inductor 42. The
locations of these voltages and currents are illustrated for convenience in Fig. 1. Each
switching cycle can be regarded as comprising sequential operating phases commencing at
times T 1 to T7 shown at the bottom of Fig. 2, and these phases are described below
starting for convenience at the time T1.
For convenience below, the primary switch is also designated by the reference 22pertaining to the MOSFET constituting this switch, and the auxiliary switch is also
designated by the reference 38 pertaining to the MOSFET constituting this switch. In the
10 description below, n represents the reciprocal of the (primary to secondary) turns ratio of
the transformer 16, Ls represents the magnitude of the inductance 32, combining the
leakage inductance with the inductance of any additional inductive component, and it is
assumed that the inductance Lo of the output inductor 26 is relatively very large. The
magnitude of the inductance 32 reflected back to the primary winding 18 is Ls/n2, and
15 (when the diodes 24 and 30 are both forward biased) this appears in parallel with the
magnetizing inductance Lm of the transformer 16. It is also assumed that Lm is much
greater than the inductance Lp of the primary inductor 40; for exarnple, the inductances Lp
and Ls can be related by the equation Ls = n2Lp.
The control circuit determines in known manner as indicated above a duty cycle or
20 on period of the primary switch 22 of the converter, which starts at the time T3 and ends
at the time T5, as shown by the control signal G1 in Fig. 2. The control circuit also
determines the control signal G2 for the auxiliary switch 38 to turn on this switch starting
at the time T1, which is in advance of the time T3 by an amount which is determined by
characteristics of the converter as described below to be such that substantially zero
25 voltage switching of the primary switch 22 takes place at the time T3. Fig. 2 illustrates the
control signal G1 as turning off the auxiliary switch 38 at substantially the same time T3.
In practice, it is desirable to ensure that the auxiliary switch 38 is turned off only when the
primary switch 22 is substantially fully turned on. For this reason, the end of the pulse of
the control signal G2 may be delayed until after the time T3, as indicated by a broken line
30 50 in Fig. 2. The timing of the end of this G2 pulse is not particularly critical, so that this
delay can be extended potentially up to the time T5. However, during such delay period
the auxiliary switch 38 is on so that there is a conduction loss, and it is preferable to keep
the delay period short to reduce conduction losses and enhance the efficiency of the
converter.
Immediately prior to the time T1, the transformer has been reset and both switches
are off, with the currents I1, I2, I3, and I4 all being substantially zero. Consequently the
voltages V1 and V2 are the same and, as can be seen from the description below, are
CA 02220747 1997-11-10
greater than the supply voltage +V. Commencing at the time T1, the auxiliary switch 38 is
turned on by the control signal G2.
Between the times T1 and T3, a resonant tank circuit is formed by the capacitor 36
in parallel with the primary inductor 40, both in series with the magnetizing inductance Lm
5 of the transformer 16. The voltage V2 across the auxiliary switch 38 falls rapidly to
substantially zero, and the current I2 through it rises in a resonant fashion. This current
comprises two parts: discharge current of the capacitor 36, and current flowing via the
primary winding 18. The voltage V1 across the primary switch 22 and the capacitor 36
consequently falls also in a resonant fashion, until it reaches zero at the time T3.
At the time T2, when the voltage V1 reaches the supply voltage +V, current in the
output circuit starts to flow via the diode 24 (as well as flowing via the diode 30), so the
current I3 increases in a resonant fashion from zero until the time T3. Consequently, the
inductance 32 is reflected at the primary winding 18 to appear as the reflected inductance
Ls/n2 in parallel with the magnetizing inductance Lm of the transformer. The
lS consequently decreased inductance reduces the time constant for the resonant waveforms
of the voltage V1 and the current I2 between the times T2 and T3, in comparison to the
time constant between the times T 1 and T2.
At the time T3, the voltage V1 across the primary switch 22 is substantially zero,
and this switch 22 is turned on by the control signal G 1. Thus there is substantially zero
20 voltage switching of the primary switch 22, avoiding any significant switching loss, at
this time. Starting at the time T3, the supply voltage +V applied by the primary switch 22
across the primary winding 18 causes the current I3 to rise linearly, until at the time T4 it
reaches a current Io which flows through the output inductor 26 to the load connected to
the terminals 12. At this time T4 the free-wheel diode 30 ceases conducting, and the full
25 output inductor current Io continues to be supplied via the diode 24 for the remainder of
the on period of the primary switch 22 until the time TS, in the known manner of a
conventional single ended forward converter. Between the times T3 and TS the voltage
V1 across the primary switch 22 is substantially zero, and the current I1 through the
primary switch 22 similarly rises linearly between the times T3 and T4. Between the
30 times T4 and TS, the currents I1 and I3 ramp up slightly.
At the time T3, or subsequently as mentioned above and indicated by the broken
line S0 in Fig. 2, when the pulse of the control signal G2 ends the current I2 through the
auxiliary switch 38 is interrupted. The abrupt interruption in current through the primary
inductor 40 reverses the voltage polarity of the coupled inductors, so that the diode 46 is
35 forward biased and conducts, feeding the stored energy of the coupled inductors back to
the supply voltage line (terminals 10 and input capacitor 14). With the diode 46conducting, the secondary inductor 42 is coupled to the constant voltage +V, so that the
current I4 flowing through it decreases linearly from an initial value to zero, as shown at
CA 02220747 1997-11-10
52 in Fig. 2. The voltage V2 across the auxiliary switch 38 between the times T3 and T5
is the supply voltage +V multiplied by the turns ratio of the coupled inductors 40 and 42.
When the current I4 reaches zero, the diode 46 is no longer forward biased and the voltage
V2 drops to zero, tracking the voltage V1.
At the time T5, the pulse of the control signal G 1 ends to turn off the primaryswitch 22, so that its current I1 is interrupted. The current I3=Io in the output circuit is
maintained by the inductance of the output inductor (the diode 24 can be considered as a
short circuit, and the diode 30 an open circuit, at this time so that the output inductance is
connected in series with the secondary winding 20 and the inductance 32) to cause the
10 capacitor 36 to be charged with a substantially constant current. Consequently the voltage
V1 across the primary switch 22 and the capacitor 36 rises linearly from substantially
zero, until it reaches the supply voltage +V at the time T6. Thus there is substantially zero
voltage switching to turn off the primary transistor 22 at the time T5. The voltage V2
across the auxiliary transistor 38 is pulled up with the rising voltage V1.
At the time T6, the voltage across the primary winding 18 is reversing through
zero due to the rising voltage V 1, so that the diode 30 begins to conduct (the diodes 24
and 30 are both forward biased and hence can be considered to be short circuits,connecting the inductance 32 across the secondary winding 20). The inductance 32 is
reflected via the transformer 16 to appear in parallel with the magnetizing inductance of the
transformer, forming a resonant circuit with the capacitance 36 so that the energy stored in
the inductance 32 is quickly transferred to the capacitor 36 between the times T6 and T7.
Between these times the current I3 through the inductance 32 falls to zero, so that at the
time T7 the diode 24 is no longer forward biased. The voltage V1 continues to rise above
the supply voltage +V until the time T7, and is followed by the voltage V2.
For the remainder of the switching cycle, until a time T 1' for the start of the next
cycle, the capacitor 36 resonates with the magnetizing inductance Lm of the transformer
16, this resonant combination having a time constant that is much greater than the period
of the switching cycle so that there is relatively little change in the voltage V1, and in the
voltage V2 following it, between the times T7 and T1'. Starting at the time T1', the
switching cycle is repeated.
It can be appreciated that the transformer is reset during the interval from the time
T6 in each switching cycle until the time T2 in the following switching cycle, when the
polarity of voltage applied to the primary winding is reversed, the volt-seconds applied
during this interval being equal and opposite to the volt-seconds applied in the switching
cycle from the time T2 to the time T6.
As already mentioned, the inductance 32 can be constituted partly or entirely by the
leakage inductance of the transformer 16. Similarly, the capacitance of the capacitor 36
can be constituted partly or entirely by parasitic capacitances of the MOSFET 22 and of the
CA 02220747 1997-11-10
primary winding 18 of the transformer 16. In any event, these parasitic values can be
taken into account in determining desired physical component values of the inductance 32
and capacitor 36. Generally, it may be desirable to provide both of these components in
order to provide predictable and stable, and not too small, time constants for the resonant
5 operating phases of the converter as described above. This is particularly the case because
these determine for example the time T3 at which the control circuit is desired to provide
the start of the pulse of the control signal Gl to achieve zero voltage switching for turning
on the primary switch 22. The provision of an inductive component 32 additional to the
leakage inductance of the primary and secondary windings of the transformer 16, and/or a
10 capacitor 36 in addition to the parasitic capacitances of the circuit, enables time constants
of the resonant circuits to be increased so that switching times in operation of the converter
can be easily and accurately determined.
The following parameters are provided purely by way of example for a 100 watt
converter providing an output of 5 volts from a supply voltage +V of 40 to 60 volts:
Magnetizing inductance Lm 150 IlH
Reciprocal of turns ratio n 1/3
Capacitor 36 10 nF
Inductance32 Ls 0.33 IlH
Primaryinductor40 Lp 3 ,uH
Secondary inductor 42 100 ,uH
Fig. 3 illustrates a modification of the single ended forward converter of Fig. 1.
In the converter of Fig. 3, the blocking diode 46 is moved to the opposite, or OV, side of
the secondary inductor 42 and is still in series with it, and the other end of this inductor 42
is connected to the junction of the primary inductor 40, capacitor 36, primary switch 22
and primary winding 18. The coupled inductors 40 and 42 can thus be simplified to
provide only three connections. The optional diode 48, if provided in the converter of
Fig. 3, is connected in series with the primary inductor 40 for example between this
inductor 40 and the auxiliary switch 48. The converter of Fig. 3 operates in a similar
manner to that of Fig. 1 as described above, except for the recovery of energy from the
coupled inductors. In this case, when the auxiliary switch 38 is opened at the end of the
pulse of the control signal G2, at or after the time T3, current is initially circulated via the
closed primary switch 22 and the diode 46 with relatively little loss. When the primary
switch is opened at the time T5, this current instead serves to charge the capacitor 36,
whereby the energy stored in the coupled inductors is in this case recovered in the
capacitor 36 instead of in the capacitor 14.
In the converter of Fig. 3 the recovery of energy from the coupled inductors does
not start until the time T5, it can be seen that it is possible to delay until the time T5 the end
of the pulse of the control signal G2 as discussed above. The particular time for ending of
CA 02220747 1997-11-10
the pulse of the control signal G2 to turn off the auxiliary switch 38 can depend on
particular circumstances, such as relative conduction losses for the switches 22 and 38.
The single ended forward converter topologies described above provide particularadvantages in that substantially zero voltage switching of the primary switch 22 can be
5 achieved, for turning the switch both on and off, under diverse supply voltage and load
conditions, so that switching losses can be reduced and efficiency of the converter can be
increased. The auxiliary switch 38 is not switched under zero voltage conditions, but it
switches a lower current than the primary switch 22 so that it can be a smaller MOSFET
having less capacitance and, therefore, its switching losses are much lower than would be
10 the case for the primary switch. Consequently, higher frequency operation of the
converter, and hence smaller size and costs, are possible. In addition, there is no need for
a reset winding on the transformer 16, resetting of the transformer core being achieved as
described above by the capacitor 36 and inductance 32 (including leakage in(lllct~nce of
the transformer). As can be seen from the waveforms V 1 and V2 in Fig. 2, voltage
stresses on the switches 22 and 38 are not high and are limited by the capacitor 36.
Energy stored in each switching cycle in the inductance 32 (including leakage inductance
of the transformer) is transferred to the capacitor 36 and then recovered via the coupled
inductors, thereby facilitating efficient operation of the converter. Furthermore, the
primary switch duty cycle of the converter is not limited to 50~o, as is often the case for
20 single ended forward converters, but can be considerably greater than this.
Although as described above the capacitor 36 is connected in parallel with the
primary switch 22, it can be appreciated that a capacitor having an equivalent function can
alternatively be connected in parallel with the primary winding 18 of the transformer 16.
In addition, connections of the auxiliary circuit 34 can be re-arranged to suit particular
25 desires, for example to use different polarities of MOSFETs for the switches and/or to
interchange connections to the +V and 0V lines connected to the terminals 10.
Furthermore, although the auxiliary circuit 34, and specifically the components 36 to 40
(and 48 if present) of this circuit, are described above as being coupled to the primary
winding 18, these or their equivalent could alternatively be coupled on the secondary
30 winding 20 side of the transformer 16, or to an auxiliary winding (not shown) of the
transformer 16. Thus for example the capacitor 36 or its equivalent can be connected in
parallel with the secondary winding or in parallel with an auxiliary winding of the
transformer.
It can also be appreciated that any inductive component additional to the leakage
35 inductance between the primary and secondary windings of the transformer constituting
the inductance 32 can alternatively be provided in series with the primary winding 18 of
the transformer.
CA 02220747 1997-11-10
Thus although particular embodiments of the invention have been described in
detail, it can be appreciated that these and numerous other changes, variations, and
adaptations may be made without departing from the scope of the invention as defined in
the claims.
