High-Efficiency, Regulated Charge Pumps forHigh-Current Applications

By Brigitte Kormann, Dipl.-Ing .

In Phase 2, also called transfer phase, theI. CHARGE PUMP TOPOLOGIES switches S 1 and S4 are closed and the flying

capacitor CF is placed in series to the inputvoltage. These two voltage sources charge theoutput capacitor COUT and supply the load .Phases 1 and 2 have a duty cycle of 50%, i .e .both have the same duration, At .

To transfer energy from the input to theoutput, the phases are periodically repeated witha frequency of several hundred kiloHertz. Acontrol circuitry and an oscillator control theoperation of the charge pump .

The output voltage ripple depends on the timeof the charge phase and on the capacitor size andits ESR. During the charge phase the outputcapacitor COUT supplies the load and is thereforebeing discharged. The longer the charge phase,the more charge is removed from the capacitor .The ESR of the output capacitor has an influencedue to the fact that current through the capacitoris reversed every cycle - during the chargecycle current flows out of the output capacitor,and during the transfer cycle charge flows intothe capacitor . The output voltage ripple can nowbe reduced by reducing either the charge time,i.e. increasing the switching frequency, or byreducing the ESR or by increasing the size of theoutput capacitor. All three possibilities have theirboundary: ESR is present in every capacitor,COUT is normally limited due to board space and

One of the best-known topologies is thevoltage doubler . Fig. 1 shows the basic switchconfiguration and the necessary capacitors for avoltage doubler charge pump .

VOUT

Fig. 1. Voltage doubler charge pump .

The operation of a charge pump can bedivided into two phases : In Phase l, also calledcharge phase, the switches S2 and S3 are closedand the flying capacitor CF is ideally charged toVIN . During this time, the output capacitor COUTsupplies the load and is therefore beingdischarged .

ABSTRACT

Charge pumps are well known as voltage doublers or inverters for very low power applications, e.g.below 50mA of output current. For applications where the input voltage could vary - as in batterypowered systems - the charge pump was not an acceptable solution, until now . The minimum additionalrequirement was a linear regulator for stabilizing the output voltage. Newer charge pumps areregulated internally and can deliver substantially higher output currents . This way charge pumps arebecoming more viable in applications that have been the domain of inductive DC/DC converter. Thistopic will give an overview on which topologies are used, how charge pumps are regulated and howexternal capacitors influence the performance of the system. Application examples show severalpossibilities to increase the performance by using additional external parts .

4-1

I

1 IN

I

Fig. 2. Push pull voltage doubler.

cost and the frequency is fixed by the design ofthe control circuit .

For further reduction of the output voltageripple, a new topology was invented: the push-pull charge pump. Fig. 2 shows the basiccircuitry of such a charge pump .

In this topology, two charge pumps are usedinstead of one. These two charge pumps operate180° phase shifted. While charge pump 1 is incharge phase, charge pump 2 is in transfer phase,and vice versa . The output is thereforecontinuously supplied from the input, thusreducing the ripple to a minimum (e.g . 5mV p-pfor TPS60100), resulting in only a small spikethat occurs during the turnover from onetransferring charge pump to the other .

In Fig. 3, the difference in output voltageripple of a single-ended charge pump to a push-pull version can be seen . The measurements aremade with the TPS60100 in both configurationswith the same ceramic capacitors :

CHARS PUM

11

S 2

S 3

S 4

0 2

4-2

OSC

CONTROL

V BOUT

LOUT LOAD

3 .4

3 .35 -

D 3 .30

3 .25

3 .20

3.40

3 .25

3 .20

180 0

0

2

2

CHARGE PUMP 2

S 4

3

4

4

2

OUT

(A)

6

8

10

12

14

16

t/ps

6

8

10

12

14

16

t/ps

S21

• C,N = 10µF• CF,. = 2.2µF Fig. 3. Output voltage ripple comparison• COUT = 22µF + 1µF between (A) single-ended and (B) push pull

charge pump .

In a voltage tripler, the two phases of acharge pump can also be observed . During Phase1 (charge phase) the switches S2, S5, S3 and S6are closed and the two flying capacitors CF, andCF2 are charged in parallel, ideally up to the inputvoltage VIN (the capacitors have to be the samevalue). The output capacitor COUT supplies the

I IN

I

Fig. 4. Voltage tripler .

'IN

OSC

OSC

Fig. 5 . 1.5-times transfer charge pump.

CONTROL

CONTROL

4-3

load. In transfer phase, the switches S1, S4 andS7 are closed while all others are open and thetwo flying capacitors are placed in series to theinput voltage to supply the load and charge theoutput capacitor, ideally to the output voltageVOUT = 3 • VIN .

T IOUT

IBOUT LOAD

T TOUT

LOUT LOAD

VOUT

I

VOUT

Fig. 5 shows the topology of a charge pumpthat operates in 1 .5-times transfer . Two phaseswitching is also applicable to this example . Inthe charge phase, the switches S3, S4 and S5 are

closed and the flying capacitors are charged inseries to half the input voltage (the two capacitorshave to be the same value) . During this phase theload is supplied by the output capacitor COUT. Inthe transfer phase, the switches S1, S2, S6 and S7close and the two flying capacitors C FX are placedin parallel to each other and in series to the inputvoltage, therefore the output capacitor COUT isideally charged to VOUT = 1 .5 • VIN .

11 . REGULATION SCHEMES

The topologies shown up to now have beenunregulated and deliver a multiple of the inputvoltage. In most applications, this would lead to

the need for an additional regulator to regulatethe output voltage if the input voltage varies, aswith batteries. To overcome this disadvantagenewer charge pumps are regulated and deliver a

fixed output voltage. There are actually threedifferent ways to regulate a charge pump . Fig. 6and Table I. give an overview of the differentregulation schemes . In Fig. 6A, I represents theaverage current delivered to the output duringone cycle. For the LinSkip method, threedifferent output loads are mentioned to describethe operation of the regulation .

VOPP =

TABLE I .ADVANTAGES AND DISADVANTAGES OF THE THREE REGULATION SCHEME

2 . VIN - VOUT4

'LOAD

>,RSi

ESRCOUT +

1

2 - fCLK ' COUT,

4-4

A. Pulse-Skip RegulationIn pulse-skip regulation, the output voltage is

held constant by skipping unneeded pulses . Aslong as the output voltage is below the requiredvalue, the charge pump operates and charges theoutput capacitor. If the output voltage exceeds acertain internally set level, the charge pump stopsoperation. This reduces the power consumptionof the internal circuitry to a minimum . When theoutput voltage sinks below the minimum outputvoltage threshold, also set internally, the chargepump starts operation again . The advantage ofthis regulation is a very low control current,because the charge pump operates only if theoutput voltage is too low . The disadvantages arethe variable frequency and the higher output

voltage ripple compared to Constant-Frequencyregulation .

Equations 1 and 2 give approximately the

output voltage and the ripple of the outputvoltage for a voltage doubler in Pulse-Skip

Atw is the time duration in which the charge

pump is not operating while At is half of the timeperiod the charge pump is operating . The sum ofthe four Rsi is the resistance of the internalswitches. In Pulse-Skip regulation, the resistanceof the internal switches is set to the minimum andonly Atw is regulated .

(2)

Regulation Scheme Advantages DisadvantagesPulse-Skip Very low quiescent current, thus higher

efficiencyHigher output voltage ripple . No fixedfrequency .

Constant Frequency Very low voltage ripple . Fixed frequency Higher quiescent currentLinSkip Low output voltage ripple . Low quiescent

current .For very low output currents no fixedfrequency .

regulation .

i Atw \ 4VOUT =2 'VIN - ILoad' \ 2 +

AtJRSi ( 1)

/ i=1

CLK

VOUT

0

(A)

(B)

(C)

VARIABLEDEPENDENT ON LOAD CURRENT

h 11111110

VARIABLEDEPENDENT ON LOAD CURRENT

2 3

oil 01110FIXED

C. LinSkip RegulationThe third and newest regulation scheme is

called LinSkip regulation . This scheme regulateson current conducted to the output and not likethe previous ones on voltage . LinSkip regulationuses 3 phases, charge phase, transfer phase andwait phase . In Phase 3, also called wait state, thecharge pump does nothing and all four switchesare open. During this phase the output capacitorCOUT supplies the load and the charge on the

Fig. 6. Different regulation schemes, (A) Pulse-Skip, (B) Constant/Fixed-Frequency and (C)LinSkip .

4-5

B. Constant-Frequency RegulationThe Constant-Frequency regulating system

regulates the output voltage by changing theresistance of the internal switches . The chargepump operates all the time independent of theoutput voltage . Regulation is achieved byvarying the charge that is transferred perswitching cycle . The internal switches and theexternal capacitors form a RC constant thatvaries with output voltage . If the output voltageis too high, the control circuitry regulates theresistance of the internal switches in such a waythat the time constant increases and the amountof transferred energy per cycle decreases . Theadvantage of this method is a fixed frequencyoperation with a very low output voltage ripplethat makes filtering easier . The output voltageripple is lower than in Pulse-Skip regulation, butincreases a bit with increasing output current .The higher the output current is, the higher is thecharge transferred to the load during the chargephase . Therefore the output capacitor isdischarged more than with low output current .The disadvantage of Constant-Frequencyregulation is higher supply current . Since theinternal circuitry is active as long as the systemis active and the switches have to be turned onand off in each cycle, more operating current isnecessary than required for Pulse-Skipregulation .

Equations (3) and (4) (next page) show theoutput voltage and the ripple of the outputvoltage of a voltage doubler in Constant-Frequency regulation .

4

VoUT = 2 . VIN - ILOAD .2 . >, Rsi

(3)i=1

flying capacitor CF stays nearly constant (despiteleakage) . If the output current is high, the chargepump operates in linear mode, i .e . the wait stateis 0 and the current transferred per cycle isregulated dependent on the load current . Whenthe output current demand decreases, the currentis fixed and the wait state increases its durationwith decreasing load current. Therefore thecharge pump now operates in skip mode .

The phases 1 and 2 have equal time durationAt. Phase 3 has the duration Atw . In steady statethis means that the charge transferred in the twophases 1 and 2 is equal, and in Phase 3 there isno transfer at all . If the time constant in chargeand transfer phase is big enough, the voltagechange on the capacitors during charge anddischarge is nearly linear. Now, the twoequations (5) and (6) for the output voltage VOUTand voltage ripple Vopp for a single-endedvoltage doubler can be derived :

For more details on how these equations areobtained, see reference 1 .

At

4VOUT = 2 • VIN - I LOAD 2 + ~' J R Si (5)At ~ i=1

Equations 1 and 5 are equal . The differenceis that the sum of the switch resistances is heldconstant and as low as possible in Pulse-Skipregulation, and only the wait time is changed . Inthe LinSkip regulation mode, there are twopossibilities for regulating the output voltage to aconstant value . The input voltage is variable butcannot be influenced by the control circuitry ofthe IC in a running system ; the load current isalso given by the system . The duration for thecharge and transfer phase At is given by thefrequency of the charge pump. Therefore the twovariables Atw and Rsi are influenced by thecontrol circuitry dependent on the actual outputcurrent. For high output currents, the duration ofthe wait state is held constant to zero and

VOPP ='LOAD, 2 • ESRCOUT +1

2 'fCLK . LOUT

VOPP = 'LOAD / At

At ~e,2ESRCOUT •+ 1

At i COUT \

4-6

changing the resistance of the internal switchesRs;, like in Constant-Frequency regulation,regulates the output voltage . Below a certainoutput current limit, the resistance of the internalswitches is fixed and the output voltage isregulated with the duration of the wait state . Alower output current yields to a longer durationof the wait state AtW .

D. Efficiency-Optimized Charge PumpDespite all the different regulation schemes,

the efficiency of a simple charge pump isdependent on the input voltage . The lower theinput voltage, the higher the efficiency (as longas the input voltage is not too low) . To increasethe efficiency for higher input voltages, a newtopology was invented. This "efficiency-optimized" charge pump can change from onetopology to another. Fig. 7 shows the necessaryswitches and capacitors for implementing thistopology.

Nine switches are necessary to build such acharge pump. With these switches it is possibleto build a voltage doubler, tripler or operate in1 .5-times transfer. In the TPS6012x chargepumps the voltage tripler possibility is notavailable and therefore the description is reducedto 1 .5-times transfer and voltage doubler . Aslong as the input voltage is high enough (higherthan approximately 2/3 of the output voltage),the charge pump operates in 1 .5-times transfer .The flying capacitors are charged by closing theswitches C and discharged by closing theswitches B. When the input voltage decreasesfurther, the charge pump changes the topologyand operates now as a voltage doubler . Theflying capacitors are now charged by closing theswitches A and discharged as above by closingthe switches B .

(4)

At e,(6)

At i

I

1 IN

Fig. 7. Efficiency-optimized charge pump .

III. INFLUENCE OF CAPACITOR SELECTION ONOUTPUT VOLTAGE RIPPLE

There are two factors important for capacitorsused in charge pump applications : thecapacitance and the equivalent series resistance(ESR) . These two factors differ in theirimportance and influence by the topology andregulation scheme used in the charge pump, andwhether they are used as input, flying or outputcapacitors . This section discusses this influencein detail, in terms of the topology regulationscheme, starting with Pulse-Skip regulation in asingle-ended and push-pull charge pump . Thenthe capacitor selection in a Constant-Frequencysingle-ended and push-pull charge pump will beshown, followed by a discussion related to theLinSkip, push-pull charge pump method .

The ESR of a capacitor always influences thetransient response of a charge pump . The higherthe ESR of the capacitors, the longer is theresponse time to load changes . The ESR of theinput CIN and the flying capacitor(s) CF reducethe maximum charge that is delivered to theoutput during one cycle . This leads to morecycles being necessary until the load and theoutput capacitor voltages return to the nominalvalue .

4-7

V BOUT

LOUT

OSC

dCONTROL

LOAD VOUT

The ESR of the output capacitor COUTimpacts directly the voltage drop during a loadtransient . The voltage drop on the ESR of theoutput capacitor is linearly dependent on thecurrent drawn by the load. The larger the ESR ofthe output capacitor is, the larger is the outputvoltage drop for an increased load demand, andtherefore the more energy has to be transferred tothe output, resulting in an increased recoverytime .

I

A. Pulse-Skip RegulationIn Pulse-Skip regulation, regardless of

whether a single-ended or a push-pull chargepump is used, the influence of the value and theESR of different capacitor choices are nearly thesame. The size of the input capacitor CIN has noinfluence, as long as it is around 4 to 5 or moretimes bigger than the flying capacitor(s) .

The value of the flying capacitor(s) C F hasdirect influence on the output voltage ripple . InPulse-Skip mode, the minimum number of pulsesis one, and the duration of the pulse is fixed .Therefore the higher the capacitance of the flyingcapacitor(s), the greater the amount of energygiven to the output of the charge pump . The moreenergy delivered to the output, the higher thevoltage change on the output capacitor and so the

output voltage ripple amplitude . The frequency ofthe ripple is reduced with a higher value of C Fbecause the load needs more time to dischargethe output capacitor to the lower threshold .

In Pulse-Skip mode, toggles the outputvoltage between two levels, therefore the outputvoltage ripple has a minimum value given by thedifference between these two levels . The chargepump operates when the voltage is below theupper level and stops operation as long as theoutput voltage is above the lower level . Thereforethe minimum output voltage ripple is given bythe design of the charge pump because these twothresholds are internally defined . The ESR of theoutput and flying capacitor add additionalamount of ripple voltage to the inherentminimum ripple .

The size of the output capacitor COUT alsoinfluences the output voltage ripple, but thisinfluence is minimized since the charge pumpdesign has established minimum voltage ripple asdescribed above. Now as long as the outputcapacitor is large as compared to the flyingcapacitor, by approximately 8 to 10 times, theupper voltage level is normally reached with onecycle. The ripple is given dependent on the sizeof the flying capacitor compared to the size of theoutput capacitor . If the capacitance of the outputcapacitor increases, the voltage ripple becomeslower. The charge pump has to pump severaltimes until the upper level is reached andtherefore the ripple is now given by design.

Despite the increased time for a transientresponse, the ESR of the different capacitorsinfluences the operation of a charge pump inPulse-Skip regulation as follows .

The ESR of the input capacitor CIN and theflying capacitor(s) CF increase the number ofnecessary pulses to maintain the output voltage .This leads to additional pulses and lowers theefficiency of the charge pump . The ESR of theflying capacitor(s) has a bigger impact than theESR of the input capacitor . With low inputvoltages and high ESR flying capacitors, theoutput voltage can no longer be maintainedbecause not enough energy can be transferred tothe output . Therefore the ESR of the flyingcapacitors has to be minimized and only ceramic

4-8

capacitors with X7R or X5R material should beused for the flying capacitors .

The ESR of the output capacitor COUT can beseen on the output voltage waveform of thecharge pump. The voltage swings during eachswitching cycle increase proportional to a higherESR. This originates from charging anddischarging the output cap in each cycle . Thechange in current flow changes the direction ofthe voltage drop at the output capacitor .Therefore at the beginning of a charge cycle, apositive voltage change will happen and viceversa at the beginning of the discharge cycle .This reduces the efficiency of the charge pumpbecause of the need for more pulses per second .

Tantalum capacitors may be used as input andoutput capacitors if the influence of the higherESR - longer transient response and higheroutput voltage ripple - is not a concern .

B. Constant-Frequency RegulationA push-pull charge pump transfers charge

continuously from the input to the output .Therefore the value of the caps is negligible aslong as the minimum requirements given in thedatasheet are fulfilled. In a single-ended chargepump, the size of the flying capacitor CF and theoutput capacitor COUT determine the outputvoltage ripple. Increasing the flying capacitorincreases the ripple because more charge is givento the output per cycle . Selecting an outputcapacitor with a larger value reduces the outputvoltage ripple because during the charge cyclethe capacitor is less discharged . The inputcapacitor CIN only maintains the low impedanceof the source and therefore increasing thecapacitance of this capacitor has no influenceupon the output voltage .

The ESR of the input capacitor CIN has noinfluence on the functionality of the charge pumpas long as a low impedance input is maintained .The problem in this case is that the capacitor isunder current stress and will be heated as afunction of the ESR. Either ceramic or tantalumcapacitors should be used as input capacitors .

The ESR of the flying capacitor(s) C F has noinfluence as long as the input voltage is highenough . Since the output voltage level ismaintained by changing the resistance of the

internal switches, the efficiency does not changewith higher ESR. The power dissipation is nowpartly moved from the internal switches to theflying capacitors . If the input voltage is decreasedto a value near the minimum voltage, the chargepump is no longer able to maintain the outputvoltage because of the higher power dissipationin the flying capacitors. Therefore, as with chargepumps in Pulse-Skip regulation, only ceramiccapacitors are recommended for the flyingcapacitors .

The ESR of the output capacitor COUT willlead to voltage ripple on the output voltage .Fig. 8 shows an example with tantalumcapacitors and ceramic capacitors at the input andoutput. The tantalum capacitors are low ESRcapacitors that have only several tenth ofmilliohm of ESR, but even this is a lot higherthan in ceramic capacitors . It can be seen that

Single-Ended :3 .4,-- ..

7I

3 .35 -

D

3 .3 -0

3 .4

3 .25

3.20 2 4

tilts

6

8

10

t/lts

12 14 16

CIN = 1µF ceramicand 10µF tantalum

COUT = 1µF ceramicand 22µF tantalum

CIN = 1µF ceramicand 10µF ceramic

COOT = 1µF ceramicand 22µF ceramic

with tantalum capacitors the ripple is increased .In a single-ended charge pump the ESR leads to avoltage drop (positive or negative) during theturnover from one phase to the other . Thus thecharge and discharge of the output capacitorhappens with the RC time-constant given by theESR and the capacitance. In the push-pull chargepump the ESR leads to a voltage drop duringturn-over from the transfer phase of one chargepump to that of the other one . The ESR of theoutput capacitor increases the output voltagespikes and reduces the efficiency by somepercent. If the ESR is higher than several ohms,e.g. in some aluminum capacitors, the outputvoltage can no longer be maintained by thecharge pump. Therefore only ceramic or tantalumcapacitors are recommended for outputcapacitors .

Push-Pull :3 .40

I-0

3 .40

3 .35

3 .30

3 .25

3 .20

0

0

2

2

4

4

tilts

6

8

tilts

6

8

10

12

14

16

10

12 14 16

Fig. 8. Comparison of single-ended with push pull charge pump and influence of the ESR of the inputand output capacitors, CF., = 2.2,uF ceramic, measured on TPS60100.

C. LinSkip RegulationCharge pumps operating in LinSkip mode are

most of the time in a Constant-Frequency orlinear regulation . Only at low output currents thecharge pump switches to skip mode . Theregulation-scheme is based on current control . Inthis case the charge pump IC regulates thecurrent transferred to the output as long as thecharge pump is in the linear mode. Therefore thesize of the different capacitors has no influenceas long as the minimum requirements specified inthe datasheet are fulfilled .

With this new regulation scheme, theinfluence of the ESR is minimized . In linearoperation the ESR has a similar influence as in apush-pull charge pump in Constant-Frequencyregulation. And in skip operation are the sameproblems observed as described in Section III . A :Pulse-Skip Regulation . . Since Pulse-Skipoperation is only used for low load currents theproblems are therefore reduced. Thus the size ofthe capacitors is minimized in this regulationscheme, ceramic capacitors can be used for input,flying and output capacitors . Tantalum capacitorshave no longer a significant advantage in size orcost. For example, TPS60200 needs 2 .2µF inputand output capacitors and 1 µF flying capacitorsto supply an output current of 100mA .

In general the best capacitors for chargepumps are ceramic capacitors with X7R or X5Rmaterial . In some cases tantalum capacitors havea size or cost advantage but then thedisadvantages must also be kept in mind . SinceLinSkip regulation mode is a current regulation,this scheme will become more popular in thefuture. Its benefits are reduced capacitor size anda lower impact of the capacitor's ESR .

4- 1 0

IV. APPLICATION EXAMPLES

A. Reducing SpikesIf there is the need to further decrease the

switching spikes of a push pull charge pump,then an additional LC filter can be added to theoutput as shown in Fig. 9. Fig. 10 shows theoutput voltage of the TPS60100 without filter andFig. 11 with the implemented LC filter . Addingthe filter reduces the spikes from 18mV down to6mV. All measurements were taken with a loadresistance of RL = 16 .5Q to draw the maximumoutput current of 200mA .

FB is connected to the filter output to avoidthat the spikes before the LC filter enter the erroramplifier and the series resistance of the inductorinfluences the regulation of the output voltage .The filter corner frequency of 2 .3MHz waschosen well above the 300kHz switchingfrequency to avoid loop stability issues .

Fig. 11 shows the same amount of AC outputripple, however the spikes have been reduceddown to 6mV.

Be aware that the energy dissipated in theseries resistance of the inductor has to bedelivered by the charge pump . Therefore withlow input voltages and high output currents, theoutput voltage may exceed the limits given in thedatasheet, either in output voltage or intemperature .

The part numbers for the capacitors and theinductor are given in the Table II.

B. Additional Negative OutputSome applications like biasing a LC-display

require in addition to the main positive supplyvoltage also a negative voltage at low current .The power supply circuit should be able to easilyobtain this additional negative supply voltage .Fig. 12 shows how to derive a negative voltagefrom a charge pump by adding a few externalparts. The negative voltage is stabilized with ashunt regulator . The regulated negative outputcan be adjusted from -VREF to -3V, where VREF is1 .24V typical .

INPUT1 .8 V TO 3 .6 V

ON

OFF

C IN10 µF

Fig. 10. Output ripple without LC filter .

d I T

SKIP

COM

3V8 II

IIN

FB

I

OUT

TPS60100 OUT

C1+

C2+C FI

2.2 µF -~-C1-

C2

ENABLE

SYNCII PGND

GNDI

I

4- 1 1

I C F2

2 .2 µF

LFIL

47 nH

Co

CFIL22µF

100 nF

1C27C 9 VLFIL 0 CFIL

Fig. 9. Output LC filter reduces spikes and output ripples to an absolute minimum .

TABLE II .PART NUMBERS FOR CHARGE PUMP TPS60100 WITH LC FILTER

OUTPUT3.3 V

RL200 mA

16.5 Q

LJJ 10 .0mVtiBW

M 1 .OOµs

Fig. 11. Output ripple with LC filter .

Part Value Part Number ManufacturerCiN 10µF, 16V EMK325F106ZF (F/Y5V) Taiyo YudenCIF, C2F 2.2µF, 16V LMK212BJ225MG-T (BJ/X7R) Taiyo YudenC o 22µF, 10V LMK316F226ZL (F/Y5V) Taiyo YudenLFIL 47nH, 0.0750 1008G470GTE StetcoCFIL 100nF, 16V EMK107BJ104AA (BJ/X7R) Taiyo Yuden

Inductive DC/DC converters can easilygenerate several voltages by additional windingsto the inductor . Charge pumps do not allow thisapproach. The proposed solution uses the voltageripple at the flying capacitor CF, to generate anegative voltage at low current . Of course, thevoltage ripple is very low as well as stronglydependent on input voltage and output current ofthe main output . Therefore it is necessary to firstmultiply this ripple voltage with a cascade(voltage inverter plus doubler) and then stabilizeit with the TLV431 shunt regulator . Anadvantage of this solution is that the negativevoltage can easily be adjusted to another voltagelevel. The circuit has been tested with allTPS601xx devices. The TPS6010x andTPS601lx-devices has to run in skip mode toensure an appropriate voltage ripple at CFX, henceSKIP is connected to VIN .

2 AA +CELLS

OUT

C3100 nF

∎

∎

∎

∎

∎

-------------TPS60100

1II

1

I

I

GND1

GND2

SYNC

ENABLE

COM

FB

SKIP

OUT

OUT

C1+

C2+

3V8

IN

IN2

C1-

C2-

PGND1

PGND4

PGND2 PGND5

I

I-----------

Fig. 12. Additional negative output with TPS601xx .

II

4- 1 2

Fig. 12 shows the application, which consistsof the TPS601xx charge pump, a cascade and the3-terminal adjustable shunt regulator TLV43 1 .Due to the low ripple and in order to minimizelosses, Schottky diodes are used for the cascadethat inverts the input voltage and doubles it . Toincrease the ripple voltage amplitude at the inputof the cascade, small 1 Q resistors have beenadded in series with the flying capacitors .

V I is the voltage after the cascade in front ofthe 47Q resistor. The negative output voltage Vois determined by the resistor network RI / R2 asfollows :

~

R 1~

V0 = - 1 +

VREF -R1 . II(REF)

(7)RZi

where VREF is typically 1 .24V . II(REF) is thereference input current, which is typically about0.15µA .

The resistor R should provide a cathodecurrent III ,,,, ;,, 0.08mA to the TLV431 atminimum V I .

D1 D2 47 0 OBAT54S BAT54S V0

-2V/1 mA

OUT3.3V/200mA1+

0

C5 C4R2 22 µF 100 nF

¶10

C 22 .2 µF

C6

C8

2.2

1 µF R1µF

TLV431 150 kOC7

C91µF

1µF R2250 kO

H0

Fig. 13. Additional transistor reduces supply current .

1. Application hints•

For best performance connect the input of thecascade with Cl- (PIN 8) when using theTPS60140

• Use low voltage Schottky diodes for thecascade, - e .g. BAT54S (double diode),LL103A, MBRM120LT3 (for highestperformance)

•

R in series with CFX must be at maximumresistance at (worst case)•

Maximum input voltage of the chargepump

•

Minimum output current at the mainoutput

•

Maximum output current at the negativeoutput

•

In some applications using the TPS6012x,TPS6013x and TPS6014x the series resistorwith CF, may not be necessary

•

The resistor in series with CF, will limit themaximum current at the main output of thecharge pump at low input voltages, but willincrease the performance of the negativevoltage. Use lowest value sufficient for theapplication

Since the charge pump generates a regulatedoutput voltage, the resistors in series with CFXwill not degrade the efficiency of the inverterwithin a wide operating range .

II

TPS60200

II

I

9

I

GNDI

4- 1 3

EN

C2+

2

C. Reducing Quiescent CurrentIn many battery-powered systems, the

quiescent current of the system is very important .Such systems are most of the time in a wait orshutdown mode and operate only during a shorttime. Afterwards the system returns into theshutdown mode for a longer period of time . Areduction of several microAmps of quiescentcurrent could increase battery lifetimesignificantly .

Another requirement in a battery drivensystem is battery voltage supervision to tell theuser when it's time to change or recharge thebattery. Additionally when using rechargeablebatteries, the supervision is also necessary toprevent the batteries from over-discharge . For themeasurement of the battery voltage, often aresistive divider in parallel to the battery is used .This resistive divider has to be high impedance inorder to minimize power dissipation . On theother hand, it needs to be low enough to preventnoise interference . Commonly a resistive dividerwith some megaOhms is used. If, for example, aresistive divider with a resistance of 1MQ isused, a constant current of l µA per volt batteryvoltage flows through the resistive divider .

Fig. 13 shows how to reduce this current toless than 1OnA, when using a charge pump likeTPS60200. Here the system voltage is regulatedto 3 .3V. If the system enters the shutdown mode

10

8

6

CF21 µF

and the TPS60200 is disabled (EN = 0), thetransistor BSS138 disconnects the resistivedivider from the battery . The TPS60200consumes in shutdown mode only a fewnanoAmps. This means that the supply current inshutdown mode is reduced from 2 .4µA to somenanoAmps when 2 battery cells in series are used(nominal cell voltage 1 .2V) .

The transistor BSS138 was used because thistransistor can also be enabled with the lowestbattery voltage . Since the resistive divider has aresistance of around 1MQ, the measurement inactive mode is not influenced by the on-resistance of the transistor. With 1 .8V to enablethe system, the on-resistance of the transistor islow enough to be neglected. Differenttemperatures also didn't impact the accuracy ofthe voltage measurement of the battery voltage .

The influence of the quiescent currentincreases with the standby time of the system andwith decreasing current consumption duringoperation .

[l]

[3]

REFERENCES

Erich Bayer, Hans Schmeller, "ChargePump with ACTIVE CYCLE Regulation -Closing the Gap Between Linear- and SkipModes ", in PCIM 2000, Dublin, Ireland

[2] Brigitte Kormann, Jim Pelfrey, "SimpleDesign of an Ultra-Low-Ripple DC/DCBoost Converter with TPS60100 ChargePump", Analog Applications Journal, May2000, Texas Instruments Literature No .SLYTO15, pp . 15-18Brigitte Kormann, "TPS6010x/TPS6011xCharge Pump", Application Report, 1999,Texas Instruments Literature No.SLVA070A

[4] Thomas Schaffner, "Additional NegativeOutput with TPS601xx ", Design Idea Page,2000, Texas Instruments Literature No .SLVA098

IMPORTANT NOTICE

Texas Instruments Incorporated and its subsidiaries (TI) reserve the right to make corrections, modifications,enhancements, improvements, and other changes to its products and services at any time and to discontinueany product or service without notice. Customers should obtain the latest relevant information before placingorders and should verify that such information is current and complete. All products are sold subject to TI’s termsand conditions of sale supplied at the time of order acknowledgment.

TI warrants performance of its hardware products to the specifications applicable at the time of sale inaccordance with TI’s standard warranty. Testing and other quality control techniques are used to the extent TIdeems necessary to support this warranty. Except where mandated by government requirements, testing of allparameters of each product is not necessarily performed.

TI assumes no liability for applications assistance or customer product design. Customers are responsible fortheir products and applications using TI components. To minimize the risks associated with customer productsand applications, customers should provide adequate design and operating safeguards.

TI does not warrant or represent that any license, either express or implied, is granted under any TI patent right,copyright, mask work right, or other TI intellectual property right relating to any combination, machine, or processin which TI products or services are used. Information published by TI regarding third-party products or servicesdoes not constitute a license from TI to use such products or services or a warranty or endorsement thereof.Use of such information may require a license from a third party under the patents or other intellectual propertyof the third party, or a license from TI under the patents or other intellectual property of TI.

Reproduction of information in TI data books or data sheets is permissible only if reproduction is withoutalteration and is accompanied by all associated warranties, conditions, limitations, and notices. Reproductionof this information with alteration is an unfair and deceptive business practice. TI is not responsible or liable forsuch altered documentation.

Resale of TI products or services with statements different from or beyond the parameters stated by TI for thatproduct or service voids all express and any implied warranties for the associated TI product or service andis an unfair and deceptive business practice. TI is not responsible or liable for any such statements.

Following are URLs where you can obtain information on other Texas Instruments products and applicationsolutions:

Products Applications

Amplifiers amplifier.ti.com Audio www.ti.com/audio

Data Converters dataconverter.ti.com Automotive www.ti.com/automotive

DSP dsp.ti.com Broadband www.ti.com/broadband

Interface interface.ti.com Digital Control www.ti.com/digitalcontrol

Logic logic.ti.com Military www.ti.com/military

Power Mgmt power.ti.com Optical Networking www.ti.com/opticalnetwork

Microcontrollers microcontroller.ti.com Security www.ti.com/security

Telephony www.ti.com/telephony

Video & Imaging www.ti.com/video

Wireless www.ti.com/wireless

Mailing Address: Texas Instruments

Post Office Box 655303 Dallas, Texas 75265

Copyright 2003, Texas Instruments Incorporated