 |
Description  |
|
|
BACKGROUND OF THE INVENTION
The present invention is directed to power regulating circuits, also called
converters and power supplies, and more specifically to a constant-current
type of series-switching regulator.
Series-switching regulators have several advantages over other types of
non-switching regulators and shunt-type switching regulators. The
efficiency of power conversion is inherently better in a series-switching
arrangement. Efficiency is particularly important in moderate and high
power applications since internal overheating of the regulator can result
in inefficient circuits. The overall efficiency for series-switching
regulators is typically 80-85%.
There are two different techniques used in series switching regulators to
control the transfer of power to the load. One design uses a separate
multivibrator (oscillator) to control the switching rate of the
series-regulator switch. A feedback signal from the regulator output
circuit indirectly adjusts the waveform of the multivibrator, which in
turn sets the duty cycle of the switched input power. The switched power
signal is usually integrated by a lowpass filter thereby providing a
steady flow of power into the load.
The series-switching regultor switch using a separate multivibrator has the
disadvantage of being difficult to control since multivibrators generally
prefer to operate at a fixed duty cycle. Also, the stability of the entire
regulator circuit can be a problem under no load or full load conditions.
Consequently, this design tends to be complex and uneconomical,
considering the added circuitry for stabilization of the overall regulator
and control of the multivibrator.
Another control technique used in switching regulators is known as the
self-oscillating technique. In the self-oscillating design, the power
needs of the output automatically determine the switching rate of the
switching regulator. This technique has the advantage of not requiring
separate multivibrators, and regulators using this design have a faster
response to changes in the load termination. There is less chance of
damage to the regulator circuitry if it is able to respond rapidly to
changes in load. This feature is particularly important if extreme changes
in the load termination are possible. Usually, the self-oscillating
technique is used for step-down voltage applications for reasons of
efficiency. Further, the output voltage polarity is the same as the input
unless a separate inverter circuit is used at the output of the regulator
This additional circuitry tends to add expense, complexity, and
inefficiencies.
OBJECTS AND SUMMARY OF THE INVENTION
It is an object of this invention to provide an improved series-switching
regulator circuit using the self-oscillating type design technique.
It is another object of this invention to provide an efficient constant
current regulator which generates either a positive or negative voltage
over a range which exceeds the input voltage range.
In accordance with the above objects, there is disclosed a novel circuit
which when connected to a source of DC input power, delivers a constant
output current to load resistance. A series-switch, in response to a
feedback control signal, adjusts the flow of DC input power by switching
ON and OFF. The switching operation generates a gated power signal having
a variable duty cycle. During the period when the switch is ON, i.e.,
closed, the energy in the gated power signal is stored in the magnetic
field of a transformer. And, during the period when the switch is OFF, the
transformer releases the stored energy to capacitive storage devices and
to the output load. The capacitive storage devices also supply current to
the output load during the period when the transformer is storing energy.
This assures that a constant flow of current is always provided to the
load. Feedback signals from voltage and current sensors combine in a
feedback network to form a single control signal. The control signal
controls the switching operation of the series switch, and hence, controls
the duty cycle of the gated power signal. As the energy demands of the
load increase, the duty cycle increases proportionately to just match the
power requirements of the load.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 is a block diagram of the preferred embodiment of this invention;
and
FIG. 2 is a timing diagram associated with the circuit of FIG. 1.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
FIG. 1 illustrates a constant-current switching regulator with input
terminals 10 and 11 and output terminals 39 and 40. In normal operation,
the regulator input is connected to an external source of unregulated DC
voltage (not shown), and the regulator provides at its output a
predetermined direct current to a load termination (also not shown)
connected across terminals 39 and 40. Since the regulator maintains a
constant current, the voltage at the output terminals varies depending
upon the particular load resistance. With a large resistance or an open
circuit at the output, the output voltage is a maximum and with a short
circuit across the output, the output voltage is a minimum. The actual
voltage range at the output can be much greater than the nominal input
voltage due to the inherent operation of the regulator circuit.
Input terminals 10 and 11 are internally connected to power switch 12,
which is typically a compound transistor circuit. The function of switch
12 is to adjust the transfer of electrical energy from the power source to
the rest of the regulator circuit and ultimately to the load. Switch 12
performs this regulating function by opening and closing the electrical
path from its input, terminal 10, to its output, path 16. The interrupted
signal on path 16, called a gated power signal, is applied to the primary
windings of transformer 14. Transformer 14 is an iron core device which
does not saturate under the voltage swing of the gated power signal. As
switch 12 opens and closes, the transformer periodically charges and
discharges in response to the energy in the gated power signal.
During the charge period, the energy transferred through the switch is
stored in the transformer's magnetic field. When the switch opens, the
flow of energy into the transformer stops, and the field collapses
inducing a current flow in the secondary windings. During each discharge
period, a current, I.sub.s, flows through the secondary windings, through
diode 15, and divides into capacitors C1 and C2, and into the load
connected across terminals 30 and 40. During the charge period, there is
no current flow out of the secondary windings, and, therefore, capacitors
C1 and C2 supply current to a load across output terminals 39 and 40.
Controlling the duty cycle of switch 12, is the function of feedback
signals from current and voltage sensors. In the primary circuit of
transformer 14 is resistor 31, which functions as a primary current
sensor. The voltage developed across the resistor is proportional to the
instantaneous primary current I.sub.p. A second current sensor 35,
generates a binary feedback signal on lead 36. When the current through
current sensor 35 is above a predetermined threshold, the binary signal on
path 36 exhibits a high state, and when this current is below the same
predetermined threshold, the binary signal on 36 exhibits a low state.
This current sensor could be implemented with a resistor connected between
path 41 and ground. Then the voltage developed on path 41, could be
applied to the input of a bistable voltage comparator having a fixed
voltage connected to the second input.
Voltage sensor 38 senses the output voltage on terminal 39 with respect to
ground 40. When the output voltage exceeds a predetermined maximum, the
binary output on lead 37 changes state to a high level. When the output
voltage drops below this maximum level, the binary output signal changes
state to its low level. The voltage sensor, coupled to other feedback
circuitry, serves as a voltage limiter protection device for open circuit
conditions. Another voltage sensor, 24, senses the polarity of the gated
power signal on path 16. This sensor provides a reference voltage to
comparator 13. A small negative voltage (-1V) is on output 25 when the
voltage on path 16 is negative, and a small positive voltage (+0.5V) is on
path 25 when the voltage on path 16 in positive. The binary voltage on
path 25 is applied to the "b" reference input of voltage comparator 13.
The function of the various current and voltage sensors is more easily
understood in relation to the waveforms as shown in FIG. 2. The particular
magnitudes of these waveforms are given in FIG. 2 only to illustrate the
operation of this particular embodiment, and are in no way unique to the
invention.
Referring now to both FIGS. 1 and 2, from an operational standpoint, it is
clear from FIG. 2 that when switch 12 closes, i.e., turns ON, the full
input voltage -50V, appears across the primary windings of transformer 14.
As long as switch 12 remains closed, I.sub.p, the current through the
primary windings, will increase linearly at a V/L rate. (L being the total
inductance from paths 16 to 32). One of the functions of current sensor 31
is to prevent I.sub.p from rising to too high a level. When I.sub.p
reaches a predetermined maximum current, switch 12 must be turned OFF.
Operationally, current sensor 31 senses the rising current and generates
an analog voltage on path 27 proportional to the instantaneous magnitude
of I.sub.p. The analog signal is fed back through resistor 22 to the "a"
input of voltage comparator 13. The operation of voltage comparator 13 is
such that as long as the "b" input voltage is more negative than the "a"
input voltage, the comparator will cause switch 12 to be in an ON state.
When the "b" input voltage is more positive than the "a" input voltage,
comparator 13 will cause switch 12 to be in an OFF state. Hence, when the
negative voltage fed back to "a" via path 27 causes "a" to exceed the
negative voltage on input "b", comparator 13 will turn switch 12 OFF.
When comparator 13 turns switch 12 OFF, the driving potential is removed
from the transformer input, and the transformer discharge period begins.
The energy stored in the magnetic field of the transformer, 1/2 LI.sup.2,
causes V.sub.p to reverse polarity as shown by the V.sub.p waveform in
FIG. 2. Since current cannot flow through the open primary circuit,
I.sub.p drops to zero, and an induced secondary voltage causes a secondary
current, I.sub.s, to flow. When I.sub.p drops to zero, an unwanted
oscillation in the input circuit could develop were it not for the
positive feedback action of voltage sensor 24. As long as V.sub.p is
positive with respect to ground, voltage sensor 24 will inhibit switch 12
from turning ON by placing a small positive voltage at the "b" input of
comparator 13. Were this "b" reference not changed, current sensor 31
could permit switch 12 to falsely turn ON when the primary current ceased,
and thus inducing an unwanted oscillation.
During the discharge period, an induced secondary voltage causes a current,
I.sub.s, to flow in the secondary. This current has three circuit loops in
which to flow. The first loop consists of conductor 34, diode 15,
capacitor C1, and back through conductor 33 to the secondary. The second
loop consists of conductor 34, diode 15, capacitor C2, current sensor 35,
and back through conductors 41 and 33 to the secondary. And the third loop
consists of diode 15, the load normally connected across terminals 39 and
40, current sensor 35, and back through conductors 41 and 33 to the
secondary. Since the capacitance of C2 is much less than C1, most of
I.sub.s flows in the first loop to charge capacitor C1. The rest of the
current flows in the other two loops. Current sensor 35 is in both the
second and third loops, and therefore, senses the sum of the currents in
these loops. Waveform 41 in FIG. 2 shows the time varying characteristics
of the current in conductor 41. As shown in FIG. 2, during the discharge
period the current waveform is triangular. The current through the load is
relatively constant so that waveform 41 reflects a fixed DC component. The
triangular nature of the waveform is caused by capacitor C2. At the start
of the discharge period, C2 provides current sensor 35 with a sudden surge
of current. This current surge pushes the sensor over its "high" current
threshold point, thereby causing the output control signal on conductor 36
to exhibit a high binary state, as shown by waveform 36 in FIG. 2. As the
current in path 41 decreases to the previous current level, the output of
current sensor 35 changes state to a low level as shown in waveform 36.
The binary output on path 36 is applied to an integrator 30 and to one
input of OR gate 29.
Current sensor 35 serves two functions. First, it causes the regulator
circuit to vary the output voltage for changes in the load resistance to
maintain output current constant. Second, it inhibits switch 12 from
closing prematurely until the stored energy is removed from the
transformer 14. Functionally V.sub.OUT is established for a given load by
regulating I.sub.p peak. The output voltage, V.sub.OUT, in steady state
operation depends not only on the input voltage, V.sub.IN, but also on the
ratio of charge period, t.sub.c, to discharge period, t.sub.d, or more
precisely Vout/Vin = tc/td. If V.sub.IN remains fixed, the ratio of
t.sub.c to t.sub.d is established to obtain the desired V.sub.OUT. And the
ratio of t.sub.c to t.sub.d is affected by regulating the peak input
current, I.sub.p. Operationally, the regulation control is performed by
the output of sensor 35 on path 36, via integrator 30 and resistor 21. By
integrating the feedback signal on path 36 in integrator 30 and applying
the result through a matching resistor 21, a control voltage is added to
the analog voltage from path 27. It is the sum of the voltages on paths 27
and 28 which determines the permitted value of I.sub.p peak. When input
"a" goes higher negatively than "b", comparator 13 turns switch 12 OFF
interrupting I.sub.p, and thus terminating the charge time. Correct choice
of component values for integrator 30 and resistor 21 will cause I.sub.p
to be the correct value so that energy input during the charge time just
balances the energy output for steady state operation. Once balanced the
system will tend to remain balanced for a wide range of input and output
voltages. Small errors in average output current will cause the time above
and below threshold to change incrementally. This in turn will
incrementally change the high to low level time of the feedback on path 36
which in turn will change the output voltage on 28 of integrator 30. This
voltage change modifies I.sub.p peak in a corrective way to return the
average output current to the desired magnitude.
As shown in FIG. 1 the control signal on path 36 is also applied to the
second input of OR gate 29. This feedback signal applied through OR gate
29, acts as an inhibit signal for switch 12 during the discharge period.
As long as the control signal on path 36 remains in a high state (-4V),
the effect of this feedback signal will be to keep switch 12 open.
Therefore, during the complete discharge cycle, the switch will be
prevented from closing. It may appear as though this feedback signal and
the feedback signal on path 23-25 perform the same function, i.e., they
both inhibit switch 12 during the discharge period, however, both inhibits
are desirable. Switch 12 must not come ON if V.sub.p is positive, even if
the desired feedback signal on 36 indicates output current is low. Also,
the switch 12 must not come ON if the control signal on path 36 indicates
I.sub.s is high, even if V.sub.p has dropped back to zero. At the start of
the discharge period, the immediate application of a digital inhibit via
OR gate 29 prevents any false and premature closing of switch 12. It is a
tenet of this regulator that the transformer be charged with an amount of
energy to just match the energy needs of the load for a full
charge/discharge cycle -- ignoring small internal circuit losses. This
requires that during the discharge period, all of the energy must be
removed from the transformer. Or, stated differently, the charge period
cannot be restarted until the transformer has been completely discharged.
Starting the charge period too soon, or too late, results in an increase
in output current ripple and an increase in internal power losses. As
previously described, several of the feedback control signals inhibit
switch 12 during the discharge cycle and theoretically, it would appear
that such controls are adequate to permit proper regulation. In practice,
however, it has been found that if the start of the charging period is
intentionally delayed momentarily, the circuit will operate in a more
stable manner. Also, a small time gap allows the energy storage devices
(other than C1 and C2) to discharge to a quiescent state. The transition
period in FIG. 2 between the discharge-to-charge periods, marked by
vertical dashed lines, indicates this intentional added delay period.
(This added delay period is shown somewhat exaggerated for purposes of
illustrating what occurs). Capacitor C3, connected across the primary
circuit of transformer 14, causes the existance of this delay. C3 is a
small damping capacitor (typically 5 .mu.F) which controls the rate
V.sub.p can reverse polarity, and hence it adds a very small delay between
the discharge-to-charge periods is negligible. The added period is a minor
wait period introduced deliberately to increase stability, and must not be
confused with excess energy wait periods common to other systems. If
I.sub.p peak is not controlled, excess energy pulses can cause an
excessive rise in output current, and a long dead period ensues until this
excess energy is dissipated. Under the normal operating load range of this
circuit, a dead period will not occur.
* * * * *
|
|
 |
|
 |
Description |
|
