Switch-mode power supplies are used to convert electrical power of a given voltage to a different desired voltage. Given the pervasive use of power supplies for a broad range of electronic devices, it is advantageous to minimize their cost and complexity, particularly, but not exclusively, for price competitive consumer electronic products.
In general, voltage outputs of switch-mode power supplies are controlled by periodically turning switches that are connected to the source of power on and off periodically, and, for a given input voltage, the duty cycle (relating the on time to the period) is directly related to the output voltage. Therefore, switch mode power supplies use one or more active devices as switches along with some kind of on-off control to turn the switch or switches on and off with a desired duty cycle to produce a desired output voltage. Such duty cycle controllers or control circuits, which generate the on-off signals to control the switch or switches are called pulse width modulation (PWM) controls or controllers. Ideally, these switches are digital in the sense that they switch almost instantaneously between full on and full off and vice versa. There are a host of canonical circuit topologies known to persons skilled in the art and available publicly to achieve various power supply characteristics. Good PWM control circuits can be quite complex due to the number of functions and details they entail, including, for example, triangle or other waveform generation, comparison, loop stability, cycle-by-cycle current limiting, slope compensation, lock-out, and other considerations. However, integrated circuit PWM controllers are readily available commercially that take care of these functions and complexities.
A less complicated PWM control technique called hysteretic control uses the output ripple (small periodic oscillations in the output voltage) to create the PWM drive to switch to control the duty cycle, thus output voltage. This hysteretic technique simplifies the PWM function by eliminating the triangle or other waveform generator and some other complexities, but it introduces other problems. For example, the output voltage necessarily has voltage ripple, which is a requirement of the feedback technique, and the switching frequency is greatly dependent on the load.
The foregoing examples of related art and limitations are intended to be illustrative, but not exclusive or exhaustive of the subject matter. Other aspects and limitations of the related art will become apparent to those skilled in the art upon a reading of the specification and a study of the drawings.
The accompanying drawings, which are incorporated herein and form a part of the specification, illustrate some, but not the only or exclusive, example embodiments and/or features. It is intended that the embodiments and figures disclosed herein are to be considered illustrative rather than limiting.
In the drawings:
The most common switch mode power supplies, converters, or regulators that use pulse width modulation (PWM) to control duty cycle of one or more active switches to control voltage output are switched inductor converters, including buck converters, boost converters, and buck-boost converters (also commonly called inverters). Buck converters reduce the input voltage in direct proportion to the duty cycle (the ratio of a switch conductive or “on” time to total switching period), whereas the output voltage of a boost converter is always greater than the input voltage. A buck-boost output voltage is inverted, but it can be greater than, equal to, or less than the magnitude of its input voltage. In all conventional buck, boost, and buck-boost converters, the input, the output, and the ground all come together at one point or node, but one of three (input, output, or ground) passes through an inductor on the way to the node where they all come together, and the other two pass through switches, at least one of which must be an active switch and the other switch can be a diode. The active switch is controlled (e.g., turned on and off) by pulse width modulation (PWM), which sets the duty cycle, and a feedback control loop is used to regulate the output voltage by modulating the pulse width to vary the duty cycle to compensate for variations in input voltage. In the examples described herein, the feedback is obtained from the node that connects the active switch to the inductor, is called the switch node.
A block diagram of an example buck converter circuit 10 utilizing switch-node feedback for pulse width modulation (PWM) to control duty cycle, thus output voltage, is shown in
Referring now primarily to the example buck converter 10 in
Essentially, the buck converter operates by closing the active switch S1 to connect the input voltage Vin to the inductor L1 to store energy in the inductor L1 as current flows to the load Rload, and then, when the active switch S1 is opened, the energy stored in the inductor L1 discharges into the load Rload, where it performs work or creates heat. While the active switch S1 is closed, there is a voltage (Vin−Vout) across L1, the current rises linearly, a magnetic field builds in L1 (storing energy), and the diode D1 is reverse biased (turned off), thus prevents current flow from the input, through D1 to ground G so that the current flows through the load Rload. However, when the active switch S1 is opened, the applied voltage Vin is removed, the magnetic field of L1 collapses, reversing the voltage and releasing the energy to continue the current flow, which falls linearly as the energy depletes, and the now forward biased (thus turned on) diode D1 provides a current path (closed circuit) for the L1 induced current to flow through the load Rload. Of course, the more time the active switch is turned on during each cycle, the more energy will be stored in the inductor L1 during each cycle, and the higher the output voltage Vout will be. Therefore, as mentioned above, the higher the duty cycle (ratio of time the switch is one to the total time of a cycle period), the higher the output voltage Vout will be. Likewise, the lower the duty cycle, the lower the output voltage Vout will be. Also, if the input voltage Vin varies for a given duty cycle, the energy transferred into the inductor L1 will vary accordingly, so the energy discharged by the inductor L1 to the load Rload will also vary, causing a comparable variations in the output voltage Vout. Therefore, in order to maintain a constant output voltage Vout for a given load Rload when the input voltage Vin decreases, the duty cycle (ratio of time the active switch S1 is on in a cycle period) can be increased to maintain the same level of energy transferred into the inductor L1 in each cycle in spite of the lower input voltage Vin. Likewise, in order to maintain a constant output voltage Vout for a given load Rload when the input voltage Vin increases, the duty cycle can be decreased. Also, a variation in the load Rload will change the voltage drop across the load Rload and affect the current flow, thus energy transfer into and out of the inductor L1, causing variation in the output voltage Vout. Consequently, the duty cycle may have to be varied to maintain a constant output voltage Vout as variations in the load Rload occur.
In the example buck converter 10 shown in
As mentioned above, the node 12, which connects the active switch S1 to the energy transfer inductor L1 is called the switch node. The pulse width modulation (PWM) control of the active switch S1 to control the duty cycle, thus the output voltage Vout, is provided in this example buck converter 10 by an inverting comparator COMP, as indicated diagrammatically by the output 14 to the active switch S1. As also indicated by the connection 16 of the inverting input (−) of the comparator COMP to the switch node 12 via the resistor R1, the feedback for the comparator COMP to regulate the output voltage Vout is obtained from the switch node 12. When the inverting comparator COMP compares the voltage on the inverting input (−) to a reference voltage Vref connected to the non-inverting (+) input of the comparator COMP, and it outputs an on signal to the switch S1 when the inverting input (−) voltage is lower than Vref and outputs an off signal when the inverting input (−) voltage is lower than Vref, thereby completing a control loop. A capacitance, represented diagrammatically by the capacitor C2 in
The duty cycle of the drive signal 14 from the comparator COMP will be that required to create an average voltage equal to the reference voltage Vref across C1, i.e., between the switch node 12 and ground G. The average voltage across the capacitor C2 is the same as the average voltage at the switch node 12, because the average current in the capacitor C2 has to be zero, and ideally, assuming the comparator COMP does not require any current so that the only current that flows through the feedback resistor R1 charges and discharges the capacitor C2, the average voltage across the feedback resistor R1 is also zero. Further, ideally, the average voltage across the inductor L1 has to be zero, so the output voltage Vout is ideally equal to the average voltage at the switch node 14. Consequently, by these equivalencies, the output voltage Vout, like the voltage at the switch node 12, will also be equal to the reference voltage Vref.
In practice, of course, inductors and comparators are not ideal, so the output voltage Vout will not be exactly equal to the voltage across the capacitor C2. For example, leakage current in the comparator COMP will cause a slight voltage drop across the feedback resistor R1, thus biasing the equivalence. Also, resistance in the inductor L1 will create droop in the output voltage Vout, which does not affect the voltage at the switch node 12, thus will not be detected or cause duty cycle correction or compensation by the pulse width modulation of the COMP. Other non-ideal losses or leakages might also bias the equivalences described above. However, there are ways to mitigate and/or compensate for such affects due to non-idealities, which are known to persons skilled in the art to help achieve acceptable operation of this kind of power supply or converter for many applications and performance specifications. For example, but not for limitation, an inductor with a specified series resistance that will keep the resulting droop within tolerable specifications for a particular application can be selected and used for the energy transfer inductor L1. Also, for example, but not for limitation, a compensatory signal can be provided to the comparator COMP based on measured output current so that the inherent droop in the output voltage Vout from resistance in the induction L1 can be cancelled by a corresponding rise in the average voltage at the switch node 12. These and other mitigations and compensations can be used, if necessary, to take advantage of the benefits of switch node feedback principles and/or example implementation described herein.
As mentioned above, the example buck converter 10 with switch-node feedback shown in
Like the
Essentially, when the voltage across the capacitor C2 drops below the low threshold set by the voltage on the base of Q3, the COMP switches on, and current flows through Q3 to drive the base of the switch Q1 to turn Q1 on. When the switch Q1 is turned on, the switch-node 12 voltage goes high, substantially equal to the input voltage Vin, and current flows from the input, through the switch Q1 and through the energy transfer inductor L1 to the load Rload. The voltage across the capacitor C2 follows the switch-node 12 voltage and increases approximately linearly. Eventually, when the voltage across the capacitor C2 reaches the high threshold set by the voltage at the base of Q3, the COMP turns off, i.e., Q2 turns on and Q3 turns off, so current ceases following through Q3 to drive the base of the switch Q1, and flows through Q2 instead. With no current flowing through Q3 to drive the base of the switch Q1, the switch Q1 turns off, and the switch-node 12 voltage drops to a diode drop below ground G as the current induced by the inductor flows through the diode D1 and through the load Rload. With that voltage drop, there is a negative voltage across R1, which gives rise to a current that makes the voltage on C2 decrease. When the voltage of C2 falls again to the threshold voltage set by the voltage on the base of Q3, then Q3 turns on as a result of Vref on the base of Q3 then being higher than the voltage at the emitter of Q3, and Q2 turns off. The turned on Q3 again turns on the switch Q1, and the cycle repeats.
The resistor R2 is provided to drain the charge stored in the base-emitter junction of the switch Q1 when Q3 turns off, so that Q1 turns off quickly. The resistor R3 sets the operating current of the COMP, which will go through either Q2 or Q3, depending on which one is turned on.
Hysteresis, which defines how far the voltage on C2 ramps up and down chasing the switching threshold, is created by positive feedback from a capacitive coupling between the Q2 output and the zener diode D2, which provides the reference voltage Vref on the base 24 of Q3. In the example buck converter 20 circuit shown in
As mentioned above for the
Although the average output voltage Vout is theoretically equal to Vref, as explained above, it can vary instantaneously from Vref. Capacitor C1 and inductor L1 form an LC low-pass filter with damping factor being dependent on the load Rload. In some cases, the dynamic impedance of the load may be very high, for example, when the load has a large inductive component or drives a circuit that draws a constant current, which can allow excessive and/or undesired ringing at the output. Such undesired effects can be mitigated by the addition of capacitance coupling from the output to the summing junction (the inverting input of the comparator) as shown, for example, by the capacitor C4 in
In some case, the capacitor C2 can be eliminated, as illustrated, for example, in
The non-idealities of the example buck converters with switch-node feedback shown in
The hysteretic comparator for use in switch-mode power supplies using switch-node feedback can also be provided in other ways, for example, with a Schmidt trigger, as illustrated in an alternate embodiment in
As the voltage at the summing junction 32, which is derived as a predetermined fraction of the voltage of the switch-node, decreases below the lower threshold, the Schmidt trigger turns on, and the current flow through Q2 drives the switch transistor Q1 to turn on. Turning on switch Q1 drives the output of Q1, thus the switch-node 12, to Vin, and causes the voltage at the summing junction 32 to increase, which is caused by capacitor C4 charging through resistor R1. The resistors R6 and R7 facilitate setting the output voltage Vout at a desired level, which may be different than the reference voltage or hysteresis thresholds.
As the voltage at the summing junction 32 increases beyond the upper threshold, the Schmidt trigger turns off, stopping the flow of current through Q2, and thereby removing the drive from the switch transistor Q1. As a result, switch Q1 turns off (speeded by drain of energy through R5), and the voltage at the switch-node 12 goes to a diode drop below ground G. This voltage drop at the switch-node 12, in turn, causes the voltage at the summing junction (the charge in capacitor C4) to decrease. The cycle repeats itself, as explained for previous examples, with the frequency defined by the hysteresis voltage, the Thevenin charging resistance, the value of the capacitor C4, and the supply voltage Vin. Also, as in the previous examples, the frequency has little, if any, dependence on the load Rload or the load current.
A metal oxide semiconductor field effect transistor (MOSFET) or other kinds of transistors (not shown) can be used for the switch Q1, if desired, instead of the bipolar junction transistors (BJT) shown in the examples, including either the differential pair or other comparator versions or the Schmidt trigger versions described above. In general, modifications of the example circuits to accommodate various silicon switches would be obvious to persons skilled in the art, once they understand the principles of this invention, and such obvious variants are considered to be equivalents within the scope and intent of the invention as defined by the claims below.
A buffer (not shown) can also be placed between the drive signal and the gate of a MOSFET switching device in order to allow the input capacitance of the MOSFET to be charged more quickly than the collector resistor of the previously described examples allow. The increased switching speed reduces switching losses and improves overall converter efficiency.
As mentioned above, non-idealities, mostly resulting from losses in the inductor L1, can result in degradation in the output voltage Vout regulation against output load, and some, but not necessarily all, relatively simple techniques for reducing these effects and improving output regulation have been discussed above. A non-ideal inductor possesses a non-zero series resistance that introduces a voltage drop across the inductor, and such voltage drop biases the equivalence between output voltage and average voltage at the switch-node upon which this method and apparatus relies for output regulation. As a result, the output voltage will droop according to the product of the inductor series resistance and the load current.
One approach to counteract such droop in voltage output due to inductor series resistance includes the output current sense circuit for injecting a compensating signal to the comparator or Schmidt trigger summing junction, as shown, for example, in
When there is little or no output current, the voltage drop across the resistor R11 is small, and the current through resistor R12 is defined by the difference between the output voltage Vin and the voltage at the summing junction 34. As the current increases, the voltage drop across resistor R11 increases, lowering the current through R12. The lower current through R12 causes a higher duty cycle at the switch-node 12, thereby increasing the average voltage at the switch-node 12, which, in turn, counteracts the inherent voltage drop in the inductor L1.
The example circuit in
While a number of example aspects and implementations have been discussed above, those of skill in the art will recognize certain modifications, permutations, additions, and subcombinations thereof. It is therefore intended that the following appended claims and claims thereafter introduced are interpreted to include all such modifications, permutations, additions, and subcombinations as are within their true spirit and scope.
The words “comprise,” “comprises,” “comprising,” “composed,” “composes,” “composing,” “include,” “including,” and “includes” when used in this specification, including the claims, are intended to specify the presence of stated features, integers, components, or steps, but they do not preclude the presence or addition of one or more other features, integers, components, steps, or groups thereof. Also the words, “maximize” and “minimize” as used herein include increasing toward or approaching a maximum and reducing toward or approaching a minimum, respectively, even if not all the way to an absolute possible maximum or to an absolute possible minimum.
This invention claims the benefit of U.S. provisional application No. 61/108,784, filed on Oct. 27, 2008.
Number | Date | Country | |
---|---|---|---|
61108784 | Oct 2008 | US |