Design Of Low Voltage Low Power And Highly Efficient DC-DC Converters

2 Switched DC-DC Converters2.1 OverviewThe main purpose of a DC-DC converter is to supply a regulated DC outputvoltage to a variable-load resistance from an unstable DC input voltage. DC-DCconverters are commonly used in applications requiring regulated DC power,such as computers, medical instrumentation and communication devices. DCDC converters are also used to provide a stable variable DC voltage for DCmotor speed control applications.There are three types of DC-DC converters in use today, linear converters,switched capacitor converters (also known as charge pumps), and switchedconverters. Linear converters can only generate lower output voltage from thehigher input voltage [2]. Their conversion efficiency is never greater thanVout/Vin. In practice, most linear converters operate with typical conversionefficiencies of only 30% [18]. This is the major limitation which makes linearconverters not suitable for the task of this thesis. However they are commonlyused in analog circuits to ensure a constant (or nearly constant) power supplyvoltage.Switched capacitor converters implement switches and capacitors to performvoltage conversion. Since they do not use magnetic components like inductorsthe amount of EMI (Electro Magnetic Interference) is low which makes theseconverters suitable for applications which are sensitive to this phenomenon.However, in analyses presented in [4] it was shown that these converters are notappropriate for the applications which require the output voltage to be regulatedbetween several different values. Switched converters operate by passing energyin discrete packets over a switch. Hence, the output voltage can be higher, loweror inverted compared to the input voltage. They offer higher power efficiencythan their linear and switched capacitor counterparts [4], [18]. However,switched converters generate significant amounts of electrical noise caused bythe switching activity. Presence of this phenomenon known as voltage ripple istheir main drawback, fortunately this problem can be solved as it will be shownlater in this section.7

The output voltage in switched DC-DC converters is generally controlled usinga switching technique, as illustrated by the basic switched DC-DC convertershown in figure 2.Figure 2: Basic switched DC-DC converter.There are tree main topologies of switched DC-DC converters used today:1. Buck or step-down converters are used to produce an output voltage betweenground and the input voltage.2. Boost or step-up converters operate in the opposite manner compared to thestep-down converters generating higher voltage at the output than at theinput.3. Buck-Boost converters are used in applications where the output voltage isrequired to have levels both higher and lower than the input voltage.The most widely used method for controlling the output voltage through theswitch (see figure 2) is pulse-width modulation (PWM). The pulse-widthmodulation control technique maintains a constant switching frequency andvaries the ratio of the charge cycle (time when the switch is on) and thedischarge cycle (time when the switch is off) as the load varies. This techniqueaffords high power efficiency. In addition, because the switching frequency isfixed, the noise spectrum is relatively narrow, allowing simple low-pass filtertechniques to greatly reduce the peak-to-peak voltage ripple at the output. Thisis a reason why, PWM is popular in telecommunication applications where noiseinterference is of concern [2].2.2 Buck ConvertersAs mentioned earlier, step-down converters are used to convert an input voltageto a lower level at the output. Basic principle of a buck topology is shown infigure 3. When the switch is in position one, the output voltage is equal to theinput voltage and when the switch is in position two, the output voltage is equal8

to zero. The resulting average voltage level at the output is a function of the timewhen the switch is in position one and two respectively. This function is calledduty ratio and it is defined by the expression D Vo/Vin, where Vo donates theaverage output voltage and Vin is the DC voltage generated by the power source.Figure 3: Basic buck topology.The main problem of this basic circuit is the voltage ripple of the output signalof the converter. For this reason a LC-filter is used to decrease the voltageripple. This modified circuit is shown in figure 4.Figure 4: Buck topology, modified circuit.Since the average current through the load resistor R is approximately the sameas the average current of the inductor, the voltage Vo across the load resistorcontains less ripple. A diode is used when the switch is in position two. Thisallows the capacitor to be charged in both switching positions. When the switchis in position one the energy is transferred from the power source to thecapacitor and when the switch is in position two the capacitor is charged withthe energy stored in the inductor. This type of operation results in high powerefficiency for buck converters.9

2.3 Boost ConvertersA boost converter topology is obtained by rearranging the components of a buckconverter according to figure 5.Figure 5: Boost topology.During the time the switch is closed energy is transferred to the inductor whilethe diode is preventing the capacitor to discharge through the switch. When theswitch opens current through the inductor continues to flow in the samedirection as during the previous cycle. This forward-biases the diode and boththe input voltage source and the inductor are transferring energy to the load.Hence, a voltage boost occurs across the load, which causes the output voltageto be higher than the input voltage. The capacitor must be large enough to keepthe output voltage approximately constant.2.4 Buck-Boost ConvertersCertain applications require voltage levels to be both higher and lower than thesource voltage. A solution to this is a buck-boost converter. A simple buck-boostconverter is shown in figure 6.Figure 6: Buck-Boost topology.The basic operation of a buck-boost converter is the following. When the switchis closed energy from the source is transferred to the inductor and the diode isreversed-biased, thus, it is off. At the same time the capacitor is discharged intothe output load RL and the output voltage is falling.10

Next, the switch is open and the inductor maintains the current direction. Thisforward-biases the diode. During this period the inductor is transferring energyto the capacitor. In other words, the capacitor is being charged as the inductor isbeing discharged, and the output voltage is rising.Previous discussion implies, that by adjusting the on time of the switchcompared to the time of one switching period, the output voltage Vo can be set toeither lower or higher levels than the input voltage Vs. If the ratio of the on timeof the switch and the switching period approaches zero than the output voltagealso approaches zero. If the ratio approaches one, than the output voltage leveltheoretically has no upper limit [19].2.5 Converter Topology SelectionIt will be shown later that in order to obtain higher power efficiencies for lowvoltage low power converters the diodes in the converters described in previoussections should be replaced by an active switch Mn as shown in figure 7 (buckcase).Figure 7: Buck converter with active switch Mn instead of diode.The basic converter topologies described earlier are a small subset of many DCDC converter topologies that have been reported in the literature. Otherimportant classes of converter topologies include transformer-coupled circuitsand soft-switching topologies, such as resonant converters. Although many ofthese topologies have important advantages in some applications, transformercoupling is usually unsuitable in portable systems, and soft-switching can beobtained without the use of resonant techniques. Thus, the basic topologies areappropriate for most portable applications [4].In buck and boost converters, a part of the output energy is supplied directlyfrom the input source, reducing the energy storage requirement of the inductor,and thus, its physical size. In a buck-boost converter, because none of the energyis transferred directly, it is transferred from the input into the inductor, and thenin a separate portion of the cycle, from the inductor to the output, a larger11

inductor is typically required in this circuit. Thus, the buck and boost topologiesare generally preferred since the inductor is a significant part of the circuit area.Because of its more severe inductor requirements, a buck-boost topology shouldonly be used for voltage polarity inversion, or in applications which require bothup- and down-conversion of the input voltage source.Linear regulators and switched-capacitor converters, which have the advantagethat they require no magnetic components (inductors), are analyzed in [4].However, their power efficiency is fundamentally limited by the conversionratio. They should therefore be used judiciously in applications where physicalsize and thereby the production cost are of far greater concern than powerconsumption, or where the conversion ratio is within a range that allowsacceptable power efficiency.The discussion above implies that the most appropriate solution for the task ofthis thesis is the switched buck converter. Since, it offers high power efficiencyover a wide range of regulated output voltage levels.12

3 Control Schemes for DC-DC ConvertersNext step is to select a suitable control scheme for the buck circuit consideringthe high power efficiency requirements of this project. Three control schemeapproaches will be analyzed and described in the following sections.3.1 Pulse-Width Modulation (PWM)The PWM control technique has been briefly described in section 2.1. Itemploys switching at constant frequency, i.e., Ts ton toff where Ts is constanttime switching period and ton and toff represent the time the switch is on and off,respectively. By adjusting the ton/toff ratio the average output voltage can becontrolled. This operation can be represented by the following equationD t on Vout where D donates duty ratio of the switch.Ts VinA popular solution for generation of switch control signal is to compare Vcontrolwith a repetitive waveform as shown in figure 8a and 8b.13

Figure 8: Pulse-Width Modulation, a) system block schematics, b) control signals (source [5]).vcontrol is obtained by amplifying the difference between the actual outputvoltage from the converter and its desired value. The frequency of the repetitivewaveform, represented by the sawtooth voltage in figure 8b, establishes theswitching frequency. This frequency is kept constant in a PWM control. Whenthe amplified error signal, which varies slowly with time relative to theswitching frequency, is greater then the sawtooth waveform, the switch controlsignal becomes high, causing the switch to turn on. Otherwise, the switch is off.In terms of vcontrol and the maximum value of the sawtooth waveform V̂st infigure 8b, the switch ratio can be expressed asD t on v control TsVˆst(1)Lower power efficiency for small load is the main drawback of this controlscheme [4]. The main advantage is the use of single switching frequency whichmakes the level of output ripple highly controllable.14

3.2 Pulse-Frequency Modulation (PFM)One control scheme which obtains high power efficiency over a wide range ofloads is pulse-frequency modulation (PFM). In this scheme, the converter isoperated only in short bursts at small load as is conceptually illustrated in figure9.Figure 9: Pulse-Frequency Modulation, operation concept.Between bursts, both power switches, Mp and Mn in figure 7 are turned off, andthe circuit is idle with zero inductor current. During this period, the filteringcapacitor at the output sources the load current. When the output is discharged toa certain threshold V-, the converter is activated for another burst, charging Cf.Thus, the load-independent losses in the circuit are reduced [4]. Further, forsmaller load current the idle time increases and thereby decreases powerconsuption. Output is regulated when the charge delivered through the inductoris equal to the charge consumed by the load. This implies that the inductor mustbe designed to be able to deliver the maximum charge consumed by the loadduring system operation.The major drawback of PFM control is that the switching period (the timebetween charge bursts) is a function of the load. Thus, the converter appearsalmost chaotic and the switching noise is unpredictable. This is not well suitedfor wireless communications applications.15

3.3 Sliding ControlThe basic operation of the sliding control is shown in figure 10 where the outputvoltage, V, is the regulated output.Figure 10: Sliding control, system block schematics.The comparator switches the input to the buck converter based on the polarity ofthe compensator output,d(Vref-V)/dt (Vref-V)/τwhere Vref is the reference voltage. Unlike in PWM regulators, the switchingfrequency of the buck converter with sliding control is not fixed by an externalsource and is a function of Vref. The feedback is highly nonlinear due to thecomparator. However, this kind of system can be intuitively understood by itsphase portrait, as shown in figure 11.16

Figure 11: Sliding control, system phase portraits.The buck converter contains two poles, so the feedback loop is a second-ordersystem. The phase portrait in figure 11 describes the transient operation of thecircuit by the time trajectories of the state variable, (V, dV/dt), with the timevariable being implicit. The comparator introduces a boundary line,d(Vref-V)/dt (Vref-V)/τ 0that divides the state space into two regions. In the upper region, the input signalto the buck converter is low and the state follows the light trajectory curves. Inthe lower region, the input signal to the buck converter is high and the statefollows the dark curves. When certain, so called sliding condition on τ is met,the trajectories from both regions point towards the boundary line, and thus thestate is constrained on the line. Therefore, the system operates approximately asa first-order system with the time constant τ. This ideal sliding control lawforces the switching frequency to be infinitely high. Use of comparator withhysteresis like schmittrigger solves this problem. The comparator in figure 10drives the buck converter low when the compensator output is greater than ,and high when it is less than - . The larger the hysteresis, the lower theswitching frequency and the larger the voltage ripple. Sliding control offers highpower efficiency over a wide range of loads [9]. However, as in the case ofPFM, switching frequency is not constant making noise control difficult.17

Few actual implementations of sliding control in low voltage low powerapplications have been done so far. This makes it hard to fully evaluate potentialof this control technique. Detailed theoretical description of sliding control isgiven in [20].3.4 Control Scheme SelectionIn previous sections three control schemes have been presented. Theiradvantages and disadvantages compared to each other are summarized below PWMAdvantages:high power efficiency at large load, high noise and EMI control due toconstant switching frequency.Disadvantages:lower power efficiency at small loads. PFM:Advantages:high power efficiency at small loads.Disadvantages:variable switching frequency (noise and EMI problems in someapplications). Sliding ControlAdvantages:high power efficiency over a wide range of loadsDisadvantages:variable switching frequency (noise and EMI problems in someapplications).Since the solution (see section 1.2) of this thesis may as well be used in wirelesscommunications applications PFM and sliding control scheme can not beconsidered as an alternative due to noise and EMI issues. For the interestedreader, an example of application that uses PFM is given in [14]. Circuitimplementation of sliding control is summarized in [9]. Comparison of theprevious work where PWM have been used see [6-8] shows that PWM schemeis dominant in the field of low power, low voltage applications. The fact thatmost hand-held wireless communication applications are also the most commonlow voltage low power applications could be one reason for the popularity ofthis control technique.18

If the power efficiency is considered, since it is one of the main priorities indesign process of this solution, PWM provides solutions with power efficiencyover 90% including control circuitry.Good noise and EMI emissions control are other advantages which make PWMsuitable for implementation in this work. Furthermore, previous work done

procedures and techniques for design of variable low voltage, low power, and highly power efficient DC-DC converter with low levels of EMI (Electro-Magnetic Interference). Selection of the adequate control scheme for the DC-DC converter is the secondary objective. The result of this work can mainly be used for implementation in digital circuits