linear technology linear technology
Post on 21-Dec-2021
7 Views
Preview:
TRANSCRIPT
MAY 1995 VOLUME V NUMBER 2
Big Power for BigProcessors: The LTC1430Synchronous Regulator
IN THIS ISSUE . . .
COVER ARTICLEBig Power for BigProcessors: The LTC1430Synchronous Regulator... 1Dave Dwelley
Editor's Page ................... 2Richard Markell
LTC in the News .............. 2
DESIGN FEATURESPower Factor Correction,Part Two —Filling in the Boxes ......... 3Dale Eagar
The LT1319: A Light-to-Digital Converter forInfrared Communications........................................ 7George Feliz
The LTC1392: Temperatureand Voltage Measurementin a Single Chip ............ 10Ricky Chow and Dave Dwelley
LT1580 Low-DropoutRegulator Uses NewApproach to AchieveHigh Performance ......... 13Craig Varga
Quad Current-to-VoltageConverter is Ideal forOptical Disk Drives ....... 15William H. Gross
Design Ideas ............. 22–37(complete list on page 22)
New Device Cameos ....... 38
Design Tools.................. 40
Sales Offices ................. 40
by Dave Dwelley
IntroductionAs computer technology advances,
microprocessor designers pack moreand more transistors into less andless space with each new design. Inan effort to save power and reduceheat, many of the newest micropro-cessors run from nonstandard supplyvoltages well below the traditional5V, often with supply tolerance re-quirements tighter than the typical±5%. Those millions of transistors,all switching at the same time, re-quire prodigious amounts of currentfrom the low-voltage supply. Addi-tionally, many of the supporting chipsstill require 5V supplies; this forcessystem designers to generate mul-tiple high-power, nonstandard outputvoltages. Last minute power-supplyvoltage changes by microprocessormanufacturers and differing voltagerequirements for otherwise pin-com-patible processors add to theconfusion and risk in such designs. Apopular solution is to add a second-ary DC/DC converter on themotherboard to convert the 5V mainsupply to the lower supply the micro-processor requires. For this purpose,Linear Technology introduces theLTC1430 high-power switching-regu-lator controller, targeted specificallyat high-power, 5V step-down applica-tions where efficiency, output-voltageaccuracy, and board-space require-ments are critical.
Functional DescriptionThe LTC1430 is a new switching-
regulator controller designed to beconfigured as a synchronous buckconverter with a minimum of exter-nal components. It runs at a fixedswitching frequency (nominally200kHz) and provides all timing andcontrol functions, adjustable currentlimit and soft start, and level-shiftedoutput drivers designed to drive anall-N-channel synchronous buck con-verter architecture. The switch driveroutputs are capable of driving mul-tiple, paralleled power MOSFETs withsubmicrosecond slew rates, provid-ing high efficiency at very high currentlevels while eliminating the need for aheat sink in most designs. TheLTC1430 is usable in converter de-signs providing from a few amps toover 50A of output current, allowingit to supply 3.3V power to the mostcurrent-hungry arrays of micropro-cessors. A novel “safety belt” feedbackloop provides excellent large-signaltransient response with the simplic-ity of a voltage-feedback design. TheLTC1430 also includes a micropowershutdown mode that drops the quies-cent current to 1µA.
The LTC1430 is designed to beused in an all-N-channel synchro-nous buck architecture (Figure 1,page 19), allowing the use of cost-effective, high-power N-channel
continued on page 19
, LTC and LT are registered trademarks of Linear Technology Corporation.Burst Mode is a trademark of Linear Technology Corporation.Pentium is a trademark of Intel Corporation.PSpice is a trademark of MicroSim Corporation.
LINEAR TECHNOLOGY LINEAR TECHNOLOGY LINEAR TECHNOLOGY
2 Linear Technology Magazine • May 1995
DESIGN FEATURES
Hot Products for Hot ProcessorsHow many of us in today’s so-
called information age have lostinformation or had a computer sys-tem crash? Might this have been aresult of poor power-supply designand implementation? It’s really hardto say how many system crashes arethe result of power supplies that can-not output the required current, staywithin the voltage tolerances, orprovide the transient response de-manded. What is certain is that thesituation will get worse. FuturePentiumTM and P6 processors andprocessors from other vendors hop-ing to plug into these sockets willundoubtedly demand stricter power-management solutions than arecurrently required.
Linear Technology leads the in-dustry in supplying devices andcircuits to provide reliable, cost-effective solutions for powering alltypes of computer systems, from desk-top systems to small, handhelddevices. This issue provides insightinto several new devices from LTCfor powering the next generation ofcomputer products.
This issue’s lead article highlightsthe LTC1430 synchronous switch-ing-regulator controller. The LTC1430is specifically designed to provide highcurrents at the precise voltages re-quired by today’s and tomorrow’smicroprocessors. The LTC1430 con-verts the 5V main supply (in the silverbox) to the lower supply voltage re-quired by the microprocessor. It doesall this with efficiencies approaching95% and the excellent transientresponse required to meet the speci-fications set by most microprocessorvendors. This issue also introducesthe LTC1392, a single-chip data-ac-quisition system for temperature,voltage, and current measurement.The LTC1392 can be used as an envi-ronmental monitor inside a computer.No external components are requiredfor temperature or voltage measure-ments; current measurements canbe configured with a single resistor.
The device provides a 10-bit digitaloutput through a three- or four-wireserial interface that can talk to virtu-ally any microprocessor.
The LT1580 is a new, very low-dropout NPN regulator for poweringdesktop microprocessor-based com-puters and systems. The LT1580 cansupply all but the most extreme of themany voltages that today’s micropro-cessors require and that tomorrow’swill demand. The LT1580 requires asupply voltage higher than the mainpower source (for example, 12V) toprovide the power for the control cir-cuitry and to provide the drive for theNPN output stage. The LT1580 hasexcellent transient response, and canprovide currents up to 7A with a 0.8Vinput-to-output voltage differential.The LT1311 is a quad current-to-voltage converter useful forphoton-to-electron conversion andother I-to-V applications. The LT1311design is based on a new approach toI-to-V conversion, which providessuperior DC and AC performancewithout external DC trims or AC fre-quency compensation. The –3dBbandwidth of the LT1311 is 12MHzand its settling time is less than 175nsto 0.1% for a 2V output step. TheLT1311 is ideal for converting mul-tiple photodiode currents to voltagesand for general-purpose matched in-verting amplifier applications.
The LT1319 is LTC’s dedicatedlight-to-digital converter IC. TheLT1319 is a flexible, general-purposebuilding block that contains all thecircuitry necessary to convert modu-lated photodiode current into a digitalsignal. The LT1319 is flexible enoughto be configured for a variety of stan-dards, including IrDA-SIR, Sharp/Newton, FIR, and 4PPM. This issuealso features the second part of thePower Factor Correction article be-gun in the last issue.
Finally, we present many usefuldesign ideas, circuits that provideproven, tested designs for specificapplications.
EDITOR'S PAGE
by Richard Markell
LTC in the News...In the April 13, 1995 issue of
EDN Magazine, the results of vot-ing by EDN’s readership showednone other than Bob Dobkin, LTC’sV.P. of Engineering, as EDN’s“Innovator of the Year.” This isthe fifth year for EDN to havetheir readership nominate andvote for the Innovation and Inno-vator of the Year. Congratulationsto Bob and thanks to all of you whovoted for him.
LTC posted a new record for netsales at $68,135,000 for the quar-ter ended April 2, 1995, an increaseof 32% over the third quarter of theprevious year. Net income for thequarter also hit a new high at$21,805,000 representing an in-crease of 43% over the third quarterof last year. In addition, theCompany paid a cash dividend of$0.07 on May 17, to shareholdersof record on April 28,1995.
Financial World, April 1995,named LTC in its “IndependentAppraisals” section showing thetop 900 over-the-counter tradedcompanies in terms of market capi-talization. Within this list, LTCreceived an A+ (superior) ratingbased on five criteria emphasizingrecent growth and profitability.
In the April 1995 edition of Elec-tronic Business Buyer magazine,our President and C.E.O., RobertSwanson is interviewed by BobRistelhueber in the “Market Pulse”section which focuses on trends inthe analog market. In it, Mr.Swanson points out that despitedire predictions for the future ofanalog, the need for high perfor-mance analog has continued togrow. As Mr. Swanson put it, “....nodigital solution has eliminated theneed to interface to the analogworld, ....they've just created newopportunities for analog...”
Linear Technology Magazine • May 1995 3
DESIGN FEATURES
Power Factor Correction,Part Two — Filling in the Boxes
An Ideal Boost ConverterA simple boost converter is shown
in Figure 1. The boost converter hastwo modes of operation, each with itsown characteristics. The two modesare known as discontinuous mode andcontinuous mode. A boost converterfunctioning as a PFCC will operate inboth modes. The criterion for deter-mining in which of these two modes aswitcher is running at any given timeis whether the inductor is left un-loaded for any part of a switchingcycle (transitions don’t count). If theinductor is unloaded (SW1 and D1both off) during a switch cycle, theswitcher is operating in the discon-tinuous mode. With the circuit shownin Figure 1, operation becomes dis-continuous when the inductor currentdecays to zero. This happens duringthe part of the switching cycle whenSW1 is open.
To understand the workings of aswitching regulator, it is necessary tohave at least a mild immunity to“inductorphobia.”
The global outbreak of Inductorphobiaof the late 20th century threatened towipe out all analog circuit design. For-tunately, the requirement for powerfactor correction mandated the use ofthe inductor. This use is largely re-sponsible for the continuation of thepractice of the art through the blackage of digital design. [See “The BlackAge.” History of the Sol System, Vol-ume 17, p. 12,947]
Immunization against Induc-torphobia involves exposing oneselfto inductors, and is highlyrecommended.
When used in a boost-mode con-verter, the inductor is placed acrossthe input line and allowed to inter-cept and store some energy. Theinductor is then placed between theinput and the output to dump itsenergy, (along with some additionalintercepted line power) into the load.Figure 1. Simple boost converter
If, during a switch cycle, the induc-tor can successfully unload all of itsenergy, the operating mode is said tobe discontinuous. Otherwise, the in-ductor is forced to store some amountof energy through multiple switchcycles. The presence of such storedenergy indicates the continuous modeof operation.
When used to implement a DCvariac, a boost-mode switcher con-trols the duty cycle of the switch SW1so that the volt seconds across theinductor always add up to zero overany complete switching cycle.
The Continuous-ModeBoost Converter
In the steady state, the continuous-mode boost converter implements thefunction of the DC variac. The dutyfactor is set by the constraint on voltseconds, namely that the total volt
An Introductionto the Ideal Inductor
1. An ideal inductor will act toprevent DC voltage across itsterminals. The inductor willsteal energy from any sourcethat attempts to impose avoltage across its terminals.
2. An ideal inductor will store theminimum possible energy.The inductor will attempt todump any energy that it hasstolen at the first possiblemoment.
3. An ideal inductor, having aninductance of L, will stretchand store L volt seconds ofcharge for each ampere ofcurrent flowing through it. Theinductor will relax and return Lvolt seconds of charge to thecircuit upon withdrawal of eachampere of current flow.
In part one of this article, we inves-tigated power factor correction (PFC)by looking at its line frequency volt-age, current, and power waveforms.The device that performs PFC is calleda power factor correction conditioner(PFCC).
The waveforms shown in part oneare ideal, in that they reflect an aver-age of what is happening inside thethree boxes, ignoring higher frequencyeffects. We showed that PFC could beperformed if the appropriate compo-nents were implemented. We furtherdeveloped the concept of an instanta-neously adjustable DC variac as anequivalent circuit for the power han-dling part of the PFCC. In part two, wewill develop the circuitry for the imple-mentation of the DC variac byintroducing the boost converter.
Why the Boost Converter?Even though several circuit
configurations can perform PFC con-ditioning, the boost topology is by farthe most popular, because of thetopology’s inherently low input ripplecurrent. This ripple current is at theswitching frequency of the boost con-verter, and must be filtered by an EMIfilter at the input terminals of thePFCC. Unfiltered switching-frequencyripple may be conducted down thepower line as EMI.
by Dale Eagar
pfc2_1.eps
E1 C1 R1 E2SW1
L1 D1
I1 I2
I3
4 Linear Technology Magazine • May 1995
DESIGN FEATURES
pfc2_2b.eps
C1 E2
L1
E1
seconds imposed on the inductor willbe zero when looked at over a com-plete switch cycle.
The voltage transformation ratio ofthe DC variac is:
E2/E1 = 1+ DF/(1 − DF)where DF = Duty Factor
Some interesting properties of thisDC variac are:
DF = 0, E2 = E1
DF = 0.5, E2 = 2 × E1
DF = 1, E2 = Infinity
It’s easy to see that as duty factorapproaches unity, things get inter-esting, and can, in fact, become quitea problem.
The problem is not merely that youget infinite voltage, but that you getinfinite voltage and infinite currentat the same instant—that means in-finite power, which tends to rearrangegalaxies.
We know that this is a problem,because after a past catastrophic ga-lactic self destruction, we sent scoutships to the estimated center of thegalaxy, only to find a breadboard ofthe circuit shown in Figure 1 floatingin space with switch SW1 open. Evi-dently, ideal components don’tvaporize.
One interesting property of a boostconverter with a DF of unity is thatthe switch SW1 never opens, whichcan be good or bad. In the non-idealboost converter, the switch simplyblows up, limiting current. In theideal circuit, the power stored in L1increases with the square of the du-
pfc2_2a.eps
C1 E2SW1
(I-DF) SW2L1
I1
I2
I3
(DF)E1
Figure 2b. Simpler version of Figure 2a
Figure 2a. Continuous-mode boost converter
–400
–300
–200
–100
0
100
200
300
400
VOLT
S
0
1.0
2.0
3.0
4.0
5.0
AMPS
I1
0
1.0
2.0
3.0
4.0
5.0
AMPS
pfcb_3 a.eps
I1
I3
I2
I3
I2
–400
–300
–200
–100
0
100
200
300
400
VOLT
S
0
1.0
2.0
3.0
4.0
5.0
AMPS
pfcb_3b.eps
0
1.0
2.0
3.0
4.0
5.0
AMPS
I1
I1
I2
I3
I2
I3
0
20
40
60
80
100
DU
TY F
ACTO
R (%
)
pfcb_3aa.eps
0
20
40
60
80
100
DU
TY F
ACTO
R (%
)
pfcb_3bb.eps
ration of the time SW1 is closed. Thisleads us to one final statement aboutthe boost converter with a unity DF:
If, in your engineering adventuresyou happen across the circuit shownin Figure 1, implemented with ideal
components, and with the switch SW1closed, do not open SW1! Call 911,and, for added safety, take the firsthyper-light shuttle out of the galaxy!
Figure 3. Waveform gallery
Linear Technology Magazine • May 1995 5
DESIGN FEATURES
values of I1, I2, and I3. Thus, Kirch-hoff ’s laws apply to the averagedsteady-state values of I1, I2, and I3just as it applies to the instanta-neous values. Once past all of thatlinear system stuff, we can get tomuch more interesting things likecircuits and waveforms.
To implement a continuous-modePFCC with a boost converter, a slightmodification needs to be made to thecircuit in Figure 1. The modificationinvolves the substitution of a secondswitch SW2 for the output diode D1,as shown in Figure 2a. The openingand closing of the switch SW2 is outof phase with the opening and closingof SW1. Thus, the action of SW1 andSW2 constitutes a single-pole, double-throw switch, as shown in Figure 2b.This modification causes the boostconverter to always operate in thecontinuous mode.
Figure 2b details this continuous-mode boost converter implementa-tion of the PFCC. Figure 3 shows thewaveforms of the PFCC shown in Fig-ure 2b. One can see that the dutyfactor is changing continuously, andis directly related to the input volt-age. An interesting property of thecontinuous-mode boost converter isthat the duty factor does not changesignificantly with load current. (Thisis to be expected for a collection ofthings whose purpose is imitating aDC variac.)
The circuit of Figure 2b is not trivialto implement in the real world. It notonly requires a switch to be subsistedfor D1 (Figure 1), but also requiresfour additional switches to implementthe input rectifier bridge (so conve-niently missing from Figure 2b).
The Nonsteady-StateBoost Converter Operatingin Continuous Mode
One of the problems of an idealapproach to a problem like PFC isoversimplification. Here we have de-veloped a model for a DC variac thatworks wonderfully well in all steady-state conditions. If the load currentchanges, does the duty factor need tochange?
–400
–300
–200
–100
0
100
200
300
400
VOLT
S
0
1.0
2.0
3.0
4.0
5.0
AMPS
pfcb_3c.eps
I1
0
1.0
2.0
3.0
4.0
5.0
AMPS
I2
I3
I1
I2
I3
0
20
40
60
80
100
DU
TY F
ACTO
R (%
)
pfcb_3cc.eps
–400
–300
–200
–100
0
100
200
300
400
VOLT
S
0
1.0
2.0
3.0
4.0
5.0
AMPS
pfcb_3d.eps
0
1.0
2.0
3.0
4.0
5.0
AMPS
I1 I2
I3
I2
I3
I1
0
20
40
60
80
100
DU
TY F
ACTO
R (%
)
pfcb_3dd.eps
Figure 3. Waveform gallery (continued)
The Continuous BoostConverter as a PFCC
By setting the switching frequencyof the boost converter to many hun-dreds to several thousand times theline frequency, we get the freedom toanalyze the boost converter in terms
of average values over multiple cycles.Because the system detailed in Fig-ure 1 behaves as a linear system inboth states of SW1, the average val-ues of I1, I2, and I3 will obey the samelaws obeyed by the instantaneous
6 Linear Technology Magazine • May 1995
DESIGN FEATURES
Stretch — The measure of latticedeformation due to magnetostric-tion in a ferromagnetic material.Stretch is also used to describe thepresence of volt seconds in an in-ductor. This usage is not strictlycanonical in that it is used irre-spective of the medium that actuallycontains the magnetic lines of force.—Solclopedia, 2120 Volume V,p. 285.
Volt Seconds — the measure of thearea under the curve of voltagewhen plotted against time in theCartesian coordinate plane. Thevolt second, the unit of measure ofstretch, was popularized in the late20th century with the advent of theswitching power supply. Later inthe 21st century, the volt secondtook on its present sinister mean-ing when Sumlioux Midge wassentenced to twelve million voltseconds for making the absurdstatement “Digital electronics is amere subset of analog electronics.”Midge made the now infamousstatement in 2027 near the peak ofthe “Era of Digital Decadence.”[Solclopedia, 2120. Volume V,p. 324.]
pfcb_4.eps
+ –
*COULOMBS ARE THE SAME AS “AMP SECONDS”
CPU
5V
LT1073
WHAT ARE VOLT SECONDS ANYWAY?
BATTERY:
CHARGE MEASURED IN
AMP HOURS (AH) OR MILLIAMP HOURS (mAH)
ENERGY =E (JOULES) VOLTAGE =V E =V • AH •
IDEAL INDUCTOR:
“STRETCH” MEASURED IN
VOLT SECONDS (V • S)
ENERGY =E (JOULES) INDUCTANCE =L E =V S • L
V =L •
IDEAL CAPACITOR:
CHARGE MEASURED IN
COULOMBS* (Q)
ENERGY =E (JOULES) CAPACITANCE =C E =Q • C
I =C •
SEC HR
dI dt
dV dt
The answer is both yes and no. Themodel of the DC variac is not quiteaccurate for a continuous-mode boostconverter. The real model of the DCvariac needs to include inductanceL1 in series with the input. The neteffect of L1 in the model of the DCvariac can be seen when the systemresponds to steps in current at fixedinput and output voltages. To allow achange in current, L1 will captureand store L1 volt seconds per amp ofcurrent change. The duty factor willhave to change momentarily from thesteady-state value to allow the choketo capture the needed volt seconds.This is illustrated in Figure 4.
In the working PFCC, the effect of
capturing and releasing volt secondsis seen in a slight shift in phase of theduty factor waveform. The amount ofphase shift is determined by the valueof L1, and is negligible for all practicalpurposes.
ConclusionThe boost converter can be used to
implement the DC variac functionrequired to perform PFC, but thisrequires the duty factor to be wellcontrolled.
In part three of this series, we willinvestigate the discontinuous-modeboost converter and how it differsfrom the continuous-mode boostconverter.
pfc2_4.eps
∆DF
O
I1
Figure 4. Current in the ideal choke and duty factor
Linear Technology Magazine • May 1995 7
DESIGN FEATURES
The LT1319: A Light-to-DigitalConverter for InfraredCommunications
by George Feliz
Infrared communication will soonprovide a convenient, cableless, point-to-point connection between portablecomputers, PDAs (personal digitalassistants), desktop computers, andperipherals. Several communicationstandards exist, including IrDA-SIR(Infrared Development Association-Serial Infrared) and Sharp/NewtonASK (amplitude-shift keying), andmore standards are being developed.The LT1319 is a flexible, general-purpose building block that containsall the circuitry necessary to trans-form modulated photodiode currentto a digital signal. When coupled to anexternal photodiode, the LT1319 be-comes a light-to-digital converter thatcan be configured to receive multiplestandards. The LT1319’s flexibility isa key feature because of the vastdifferences between standards andbecause it can be reconfigured forfuture standards.
Operation of the LT1319Figure 1 is a block diagram of the
LT1319 with external filters for IrDA-SIR and Sharp/Newton. The preampis the secret of the part’s versatility.An external photodiode connected tothe preamp input produces a reversecurrent proportional to the incidentlight. The preamp is a low-noise(2pA/√Hz), high-bandwidth (7MHz)current-to-voltage converter thattransforms the photodiode current(IPD) into a voltage. The 7MHz band-width supports data rates up to4Mbaud. The low noise allows forlinks of two meters or more. When fullbandwidth is not required, sensitivitycan be increased by further reducingthe noise with a lowpass filter on thepreamp output. Encircling the preamp
is a loop formed by GM1, CF1, a buffer,and RL1. For low-frequency signals,the loop forces the output of thepreamp to VBIAS. High-frequency sig-nals are unaffected by the loop, so thepreamp output is effectively ACcoupled. The break frequency set bygm, CF1 and the ratio of RFB to RL1 iseasily modified, since CF1 is a singlecapacitor to ground. The loop rejectsunwanted ambient signals, includingsunlight and incandescent and fluo-rescent lights.
After the preamp stage, there aretwo separate channels, each contain-ing a high-input-impedance filterbuffer, two gain stages with lowpassloops, and a comparator. The onlydifference between the channels isthe response times of the compara-tors—25ns and 60ns, respectively.For modulation schemes with pulsewidths down to 125ns, the high-fre-quency comparator with its activepull-up output stage is ideal. The lowfrequency comparator, with its open-collector output and 5k internalpull-up resistor, is suitable for moremodest speeds, such as the 1.6µspulses seen with IrDA-SIR. BuffersA1 and A4 allow the use of a widerange of external filtering to optimizesensitivity and selectivity for specificmodulation methods. The externalcomponents shown are an 800kHzlowpass for IrDA-SIR, formed by RF2and CF3, and a 500kHz LC tank cir-cuit with a Q of 3 for Sharp/Newton,formed by RF1, CF2 and LF1. The loopscontaining GM2 and GM3 surroundthe gain stages and function similarlyto the preamp loop. They also provideaccurate threshold setting at the com-parator inputs by forcing the DC levelof the differential gain stages to zero.
The threshold is set by the currentinto pin 11, which is multiplied by 4in the VTH generator and then sunkthrough RC1 and RC3. With an RT1 of30k, the current into pin 11 is about130µA. The comparator thresholdsare 130µA × 4 × 500Ω = 260mV.Referred to the filter buffer inputs,the threshold is 260mV/400 or0.65mV.
Other features of the LT1319 in-clude a shutdown pin, which reducesthe supply current from a nominal14mA to 500µA. To reduce false out-put transitions due to power-supplynoise, the preamp and gain stageshave separate analog grounds andare operated off an internally regu-lated 4V supply bypassed at pin 16.The comparators, shutdown, andthreshold circuitry operate directlyoff the 5V supply and are returned todigital ground. To provide a low-noisebias point for the amplifiers, the partgenerates an internal 1.9V reference(V BIAS), which is bypassed exter-nally at pin 5.
FilteringOptimal filtering rejects interfer-
ence and improves sensitivity.Although the LT1319 data sheetshows filtering for several modula-tion standards, there are applicationsthat require custom filtering. Hereare some filter guidelines:
1. Limit the noise bandwidth witha lowpass filter that has a rise timeequal to half the pulse width. Forexample, for 1µs pulses, a 700kHzlowpass filter has a 10%–90% risetime of 500ns.
2. Limit the maximum highpass breakfrequency to 1/(4 × pulse width). For 1µspulses, the limit would be 250kHz.
8 Linear Technology Magazine • May 1995
DESIGN FEATURES
3. In setting the highpass filters,space the filter corners by a factor of5–10 to reduce overshoot due to filterinteraction. Overshoot becomes es-pecially important for high inputlevels, because it can cause falsepulses that may not be tolerated incertain modulation schemes.
4. As a general rule, place thelowest frequency highpass around thepreamp and the highest highpassaround the gain-of-400 stage, or be-tween the preamp and the filter-bufferinputs. The reason for this order isthat high light levels can have slowphotodiode-current tails that can in-crease the output pulse width. The
tail response can be filtered out by ahighpass of 200kHz–400kHz.
5. In all cases with custom filter-ing, or when modifying one of theapplications presented in the datasheet, evaluate the system at a vari-ety of distances and with data streamsthat exhibit the full duty-cycle range.
–
+13
DATAL
5V
11
VTH
12
DIG_GND
RC1 500Ω
RG2 1k
RL2 10k
VBIAS
VBIAS
gm 4k
RG1 1k
RFB 15k
DS1
RC2 500Ω
RSC 2k
LOW FREQUENCY COMPARATOR
COMP 1
–
+A3
AV = 20–
+A2
AV = 20–
+ A1 FILTER BUFFER
–
+PREAMP
VTH GEN
RL1 10k
RF1 1k
RF2 2k
RT1 30k
CT1 1µF
RH1 50k
RS3 20k
RS5 20k
RS6 1k
RS2 20k
RS1 20k
RS4 20k
RH2 50k
1
GM2
GM3
–
+10
DATA
RC3 500Ω
RG4 1k
RL3 10k
VBIAS
gm 4k
RG3 1k
RC4 500Ω
HIGH FREQUENCY COMPARATOR
COMP 2
–
+A6
AV = 20–
+A5
AV = 20–
+ A4 FILTER BUFFER
1
+–
+
+
+–
gm 4k
GM1
+–
+
+
+
CF1 10nF
CB4 1µF
CF4 2.2nF
CF2 1nF
+
+
+
CF3 100pF
15
5VVCC
16BYPASS
14
SHDN
CB2 10µF
+
+
+
CB3 10µF
+
CB1 0.1µF
CF5 1µF
9
FILT2
LT1319 • BD
Q1
Q2
Q4
VREG
Q3
DS2
3FILT1
6
FILTINL
7FILT2L
8
FILTIN
5
VBIAS
4
PREOUT
PHOTO- DIODE
1
AN_GND
1
2
INIPD
LF1 100µH
+
NOTE: EXTERNAL COMPONENTS ARE SHOWN FOR AN IRDA AND SHARP/NEWTON DATA RECEIVER.
Figure 1. LT1319 block diagram
Linear Technology Magazine • May 1995 9
DESIGN FEATURES
Overdrive at Short RangeAt short range there are two major
problems: huge photodiode currentsand slow photodiode-current decay.Typically, the maximum photodiodecurrent that the LT1319 can handleis 6mA. Beyond this level the preampinput voltage can sag and its recoverytime can cause wide pulse widths atthe output. The maximum input cur-rent can be increased to 20mA ormore by placing an NPN transistorwith its emitter tied to pin 2, its basetied to pin 4, and its collector tied tothe 5V supply. The choice of transis-tor depends on the bandwidthrequired for the preamp. The base-emitter capacitance of the transistor(CJE) is in parallel with the 15k feed-back resistor of the preamplifier andperforms a lowpass filtering function.For modest data rates, such as IrDA-SIR and Sharp/Newton, a 2N3904limits the bandwidth to 2MHz, whichis ample. For data rates with pulsesnarrower than 500ns, a transistor
with fT greater than 1GHz is needed,such as MMBR941LT1.
The second problem with large in-put signals is the photocurrent tail.This tail is proportional to the inputlevel and has a decay time constantgreater than 1µs. If the data has a lowduty cycle and the highpass filteringis below 200kHz, the output pulsewidth can become so wide that itextends into the next bit interval. Forthe case of IrDA-SIR, rejecting the1µs time constant can cause attenu-ation of the data pulses and reducedmaximum-link distance. An alterna-tive is shown in Figure 2: anapplication for IrDA-SIR and two ofits proposed higher-date-rate exten-sions—FIR and 4PPM. A clamp/squelch circuit consisting of Q1, Q2,and RC1–RC4 is added. Q1 is used asdescribed above to clamp the input,but the input-current level at whichthe clamp engages has been modifiedby RC1 and RC2. Without the resis-
tors, Q1 would turn on when thevoltage across the 15k resistor in thepreamp reached about 0.7V (an inputof 47µA). The drop across RC1 re-duces this voltage by about 350mV.At this new level of 23µA, Q1 turns onto clamp the preamp output. Thecollector current of Q1 provides basedrive for Q2, which saturates andpulls its collector close to 5V. TheFILT2 and FILT2L inputs are nowpulled positive by RC3 and RC4, whichforce an offset at the inputs to thefilter buffers that is high enough toreject the input tail.
ConclusionThe LT1319 is an ideal choice for an
infrared receiver because its high per-formance and flexibility allow it toimplement multiple modulationschemes. Point-to-point infrared linksbuilt around the LT1319 conform totoday’s standards and are easily modi-fied for the standards of tomorrow.
Figure 2. IrDA SIR/FIR/4PPM data receiver
AN_GND
IN
FILT1
PREOUT
VBIAS
FILTINL
FILT2L
FILTIN
LT1319
SHUTDOWN INPUT
SIR DATA
FIR/4PPM DATA
VCC
RT1 30k
CB1 0.1µF
1
2
3
4
5
6
7
8
16
15
14
13
12
11
10
9
CB3 10µF
CB2 10µF
CT1 1µF
CF4 3.3nF
CB4 1µF
CF5 10nF
AGNDDGND
LT1319_2.eps
CF1 10nF
CF3 100pF
CF2 22pF
CF6 65pF
RF2 2k
RF1 510ΩD1* RF3
5.1k
RC3 10k
RC4 10k
RC1 2k
**
RC2 11k
Q1 MMBR941LT1
BYPASS
VCC
SHDN
DATAL
DIG_GND
VTH
DATA
FILT2
* BPW34FA OR BPV22NF ** THESE COMPONENTS ARE ONLY REQUIRED FOR VERY LARGE INPUT CURRENTS THAT OCCUR WHEN THE PHOTODIODE IS LESS THAN 3cm AWAY. SEE TEXT
Q2 2N3906
10 Linear Technology Magazine • May 1995
DESIGN FEATURES
The LTC1392: Temperatureand Voltage Measurementin a Single ChipIntroduction
The LTC1392 is a new micropower,multifunction data-acquisition sys-tem designed to measure ambienttemperature, system power-supplyvoltage, and power-supply current ordifferential input voltage. It requiresno external components for tempera-ture or voltage measurements, andcurrent measurements can be madewith a single, low-value external re-sistor. An onboard 10-bit A/Dconverter provides a digital outputthrough a three- or four- wire serialinterface. Supply current is only350µA when performing a measure-ment; this automatically drops to lessthan 1µA when the chip is not con-verting. The LTC1392 is designed forPC board temperature, supply volt-age, and supply current monitoring,or as a remote temperature and volt-age sensor for monitoring almost anykind of system. It is available in SO8and DIP packages, allowing it to fitonto almost any circuit board.
TechnologyThe LTC1392 includes an
onboard, curvature-corrected temper-ature sensor, a 10-bit switched-capacitor ADC with sample and hold,an analog multiplexer, a trimmed in-ternal bandgap reference, and a three-or four-wire, half-duplex serial inter-face. The system is capable of makinga temperature measurement in aslittle as 142µs, or a voltage measure-ment in 72µs, with the maximum250kHz input clock frequency.Figure 1 shows an internal block dia-gram of the LTC1392. The inputanalog MUX passes the selectedinput signal to the sample-and-holdcapacitor CSAMPLE. The LTC1392adjusts the value of CSAMPLE auto-matically to provide the correct A/Dgain for each measurement mode.The 10-bit capacitive DAC, combinedwith the SAR register, converts thedifferential analog signal from themultiplexer inputs into a 10-bit digi-tal word. The resulting data is then
shifted out the serial interface to theexternal microprocessor.
The internal 2.42V bandgap refer-ence is stable over the operatingtemperature range. The temperatureand supply voltage measurementmodes use this 2.42V value directlyin the A/D conversion, providing 0.25°temperature resolution, or 4.7mVvoltage resolution in power-supplymode. The differential voltage inputmode uses an internal resistor di-vider on the reference output toprovide 1.0V and 0.5V unipolar full-scale ranges. The reduced voltageranges pay an accuracy penalty overthe full reference ranges: the 1V scaleprovides 8 bits of absolute accuracy,and the 0.5V range provides 7 bits.
The on-chip temperature sensorprovides a voltage output that is lin-early proportional to the LTC1392’sdie temperature. The low, 350µA ac-tive supply current, combined withthe fast temperature conversion rate
Figure 1. LTC1392 block diagram
1392_1.eps
+–+–+–
TEMPERATURE SENSOR
ANALOG INPUT MUX
CSAMPLE
VREF = 2.42VVREF = 1.0VVREF = 0.5V
INPUT SHIFT
REGISTERBANDGAP
SERIAL PORT
10 BITS
10-BIT SAR
CONTROL AND TIMING
COMP
10-BIT CAPACITIVE DAC
GND
+VIN
–VIN
VCC
VREF6
7
VCC
8GND
5
DIN1
CLK3
DOUT2
CS4
by Ricky Chow and Dave Dwelley
Linear Technology Magazine • May 1995 11
DESIGN FEATURES
operating temperature range. The10-bit A/D gives 0.25° resolutionover the 0°C–70°C (LTC1392C) or−40°C – 85°C range. To calculatetemperature from the LTC1392’s out-put, use the formula (ADC code/4) −130°. The theoretical maximum rangeis −130° to 125.75°, although theLTC1392 isn’t guaranteed to meetspec over this whole range. Figure 4shows the typical output tempera-ture error of the LTC1392 overtemperature.
In supply-voltage-monitor mode,the A/D makes a differential mea-surement between the 2.42V referenceand the actual power-supply voltage.Each LSB step is approximately4.727mV, giving a theoretical mea-surement range of 2.42V to 7.2V. TheLTC1392 has guaranteed accuracyover a voltage range of 4.5V–6V, witha total absolute error of ±25mV or±40mV over the commercial orindustrial temperature range, respec-tively. To calculate voltage, use theformula (ADC code × 4.727mV)+ 2.42V.
The differential voltage input modecan be configured to operate in either1V or 0.5V unipolar full-scale mode.Each mode converts the differentialvoltage between input pins +VIN and−VIN directly to bits, with the outputcode equal to [ADC code × (full scale/1024)]. The 1V mode is specified at 8bits accuracy, with the eighth bitaccurate to ±1/2 LSB or ±2mV, whereas
the 0.5V full scale mode is specified to7 bits accuracy ±1/2 LSB, giving thesame ±2mV accuracy. The differen-tial inputs include a common-modeinput range encompassing both powersupply rails, allowing them to be usedto measure the voltage across a senseresistor in either leg of the powersupply. They can also be used tomake a unipolar differential trans-ducer bridge measurement, or tomake a single-ended voltage mea-surement by grounding the −VIN pin.
The LTC1392’s three- or four-wireserial interface allows it to fit into an8-pin SO or DIP package, and makesconnection to virtually any micropro-cessor easy. Four pins are dedicatedto the serial interface: active-low chipselect (CS), clock (CLK), data input(DIN), and data output (DOUT). The DINpin is used to configure the LTC1392
and the thermal mass of the package,keeps the die temperature to within0.1° of ambient temperature for singleconversions; continuous temperatureconversions with no idle periods inbetween will raise the die tempera-ture no more than 0.25° aboveambient temperature. The tempera-ture-sensor cell is based on a circuitdevised in 1979 by G. C. M. Meijer ofthe Delft University of Technology,The Netherlands. The circuit gener-ates a current proportional to absolutetemperature (IPTAT), and subtractsfrom it a current proportional to theVF of a P-N diode. Meijer showed thatthe resulting current is inherentlycalibrated when it is properly trimmedat any one temperature. The LTC1392temperature sensor (Figure 2) takesadvantage of this basic principle togenerate a voltage proportional to tem-perature by generating a VPTAT voltageand subtracting VF directly from it.Referring to Figure 3a, it can be shownthat the total output voltage is scaledproportionally to the Celsius tem-perature of the system. The circuitincludes a curvature-compensationcircuit to compensate for the inher-ent non-linearity of VF versustemperature (Figure 3b).
Measurement PerformanceWafer-level trimming allows the
LTC1392 to achieve a guaranteedaccuracy of ±2°C at room tempera-ture, and ±4°C over the entire
Figure 2. Block diagram of Celsius tempera-ture sensor
Figure 3a. Idealized output voltage, VOUT, andits components versus temperature
Figure 3b. Output voltage versus temperature,showing curvature and curvature-correctedoutput
1392_2.eps
CURVATURE COMPENSATOR
CIRCUIT
VOUT
VPTAT
A
R2
Q2Q1R1
RB
1392_3a.eps
VG0
0
–VG0
VOUT
TZ
VG0 = BANDGAP OF SILICON
VPTAT
T(K)
VBE
1392_3b.eps
VG0
0
–VG0
CURVATURE CORRECTION
OUTPUT WITH CURVATURE CORRECTION
OUTPUT WITHOUT CURVATURE CORRECTION
VBE IDEAL
VPTAT
T(K)
VBE ACTUAL
TEMPERATURE (°C)
–3.0
–2.5
–2.0
–1.5
–0.5
0
0.5
1.0
1.5
2.0
2.5
3.0
OU
TPU
T TE
MPE
RAT
UR
E ER
RO
R (°
C)
1392_4.eps
–40 –20 0 20 40 60 80 100 120
Figure 4. Sensor error versus temperature(temperature output)
12 Linear Technology Magazine • May 1995
DESIGN FEATURES
for the measurement, and the DOUTpin outputs the A/D conversion data.The DIN pin is disabled after a validconfiguration word is received, andthe DOUT pin is in three-state modeuntil a valid configuration word isrecognized, allowing the two pins tobe tied together in a three-wire sys-tem. The serial link allows severaldevices to be attached to a commonserial bus, with separate CS lines toselect the active chip. The small sizeand low pin count make the LTC1392useful in compact, remotely locatedsystems, or in isolated systems with alimited number of control wires. The350µA current consumption and 1µAshutdown mode make it usable inlow-power or battery-operated sys-tems, and single-supply operationeliminates the need for a negativesupply voltage.
Typical ApplicationFigure 5 shows a typical LTC1392
application. A single-point “star”ground is used along with a groundplane to minimize errors in the volt-
age measurements. The power sup-ply is bypassed directly to the groundplane with a 1µF tantalum capacitorin parallel with an 0.1µF ceramiccapacitor.
The conversion time is set by thefrequency of the signal applied to theCLK pin. The conversion starts whenthe CS pin goes low. The falling edgeof CS signals the LTC1392 to wake upfrom micropower shutdown mode.After the LTC1392 recognizes thewake-up signal, it requires an addi-tional 80µs delay for a temperaturemeasurement, or a 10µs delay for avoltage measurement, followed by a4-bit configuration word shifted intoDIN pin. This word configures theLTC1392 for the selected measure-ment and initiates the A/D conversioncycle. The DIN pin is then disabledand the DOUT pin switches from three-state mode to an active output. A nullbit is then shifted out of the DOUT pinon the falling edge of the CLK, fol-lowed by the result of the selectedconversion. The output data can beformatted as an MSB-first sequence
or as an MSB-first followed by anLSB-first sequence, providing easyinterface to either LSB-first or MSB-first serial ports. The minimumconversion time for the LTC1392 is142µs in temperature mode or 72µsin the voltage-conversion modes, bothat the maximum clock frequency of250kHz.
ConclusionThe LTC1392 provides a versatile
data acquisition and environmentalmonitoring system with an easy-to-use interface. Its low supply current,coupled with space-saving SO8 orDIP packaging, makes the LTC1392ideal for systems that require tem-perature, voltage, and currentmeasurement while minimizingspace, power consumption, andexternal component count. Thecombination of temperature- and volt-age-measurement capability on onechip makes the LTC1392 unique inthe market, providing the smallest,lowest power multifunction dataacquisition system available.
Figure 5. Typical LTC1392 application
1392_5.eps
DOUT
CS GND
+VIN
RSENSE
1µF 0.1µF 5V
ILOAD
–VIN
VCCDIN
LTC1392
CLK
P1.4
MPU (e.g. 68HC11)
P1.3
P1.2
Linear Technology Magazine • May 1995 13
DESIGN FEATURES
LT1580 Low-Dropout RegulatorUses New Approach to AchieveHigh PerformanceIntroduction
Low-dropout regulators have be-come more common in desktopcomputer systems as microprocessormanufacturers have moved away from5V-only CPUs. A wide range of supplyrequirements exist today, with newvoltages just over the horizon. In manycases, the input-output differentialis very small, effectively disqualifyingmany of the low-dropout regulatorson the market today. Several manu-facturers have chosen to achieve lowerdropout by using PNP-based regula-tors. The drawbacks of this approachinclude much larger die size, inferiorline rejection, and poor transient re-sponse.
Enter the LT1580The new LT1580 NPN regulator is
designed to make use of the highersupply voltages already present inmost systems. The higher voltagesource is used to provide power forthe control circuitry and supply thedrive current to the NPN output tran-sistor. This allows the NPN to be driven
Figure 1. LT1580 delivers 2.5V from 3.3V at up to 6A
into saturation, thereby reducing thedropout voltage by a VBE compared toa conventional design. Applicationsfor the LT1580 include 3.3V to 2.5Vconversion with a 5V control supply,5V to 4.2V conversion with a 12Vcontrol supply, or 5V to 3.6V conver-sion with a 12V control supply. It iseasy to obtain dropout voltages aslow as 0.4V at 4A, along with excel-lent static and dynamic specifications.
The LT1580 is capable of 7A maxi-mum, with approximately 0.8Vinput-to-output differential. The cur-rent requirement for the controlvoltage source is approximately 1/50of the output load current, or about140mA for a 7A load. The LT1580presents no supply-sequencing is-sues. If the control voltage comes upfirst, the regulator will not try to sup-ply the full load demand from thissource. The control voltage must beat least 1V greater than the output toobtain optimum performance. Foradjustable regulators, the adjust-pincurrent is approximately 60µA and
varies directly with absolute tempera-ture. In fixed regulators, the groundpin current is about 10mA and staysessentially constant as a function ofload. Transient response performanceis similar to that of the LT1584fast-transient-response regulator.Maximum input voltage from the mainpower source is 7V, and the absolutemaximum control voltage is 14V. Thepart is fully protected from over-current and over -temperatureconditions. Both fixed voltage andadjustable versions are available. Theadjustables are packaged in 5-pinTO-220s, whereas the fixed-voltageparts are 7-pin TO-220s.
The LT1580 BringsMany New Features
Why so many pins? The LT1580includes several innovative featuresthat require additional pins. Both thefixed and adjustable versions haveremote-sense pins, permitting veryaccurate regulation of output voltageat the load, where it counts, ratherthan at the regulator. As a result, thetypical load regulation over a range of100mA to 7A with a 2.5V output isapproximately 1mV. The sense pinand the control-voltage pin, plus theconventional three pins of an LDOregulator, give a pin count of five forthe adjustable design. The fixed-volt-age part adds a ground pin for thebottom of the internal feedback di-vider, bringing the pin count to six.The seventh pin is a no-connect.
Note that the adjust pin is broughtout even on the fixed-voltage parts.This allows the user to greatly im-prove the dynamic response of theregulator by bypassing the feedbackdivider with a capacitor. In the past,using a fixed regulator meant suffer-ing a loss of performance due to lackof such a bypass. A capacitor value of
1580_1.eps
+
+
+
+ C2 220µF 10V
VIN
1
3
2VCC
VSS
4VCONT
53.3V
5V
RTN
SENSE
VOUTADJ
R1 110Ω 1%
VOUT = 2.5V
U1 LT1580
C3 22µF 25V
C4 0.33µF
R2 110Ω 1%
C1 100µF
10V
100µF 10V × 2
1µF 25V × 10
MICROPROCESSOR SOCKET
by Craig Varga
14 Linear Technology Magazine • May 1995
DESIGN FEATURES
Circuit ExamplesFigure 1 shows a circuit designed
to deliver 2.5V from a 3.3V sourcewith 5V available for the control volt-age. Figure 2 shows the response to aload step of 200mA to 4.0A. The cir-cuit is configured with a 0.33µFadjust-pin bypass capacitor. The per-formance without this capacitor isshown in Figure 3. This difference inperformance is the reason for provid-ing the adjust pin on the fixed-voltagedevices. A substantial savings inexpensive output decoupling capaci-tance may be realized by adding asmall “1206-case” ceramic capacitorat this pin.
Figure 4 shows an example of acircuit with shutdown capability. Byswitching the control voltage ratherthan the main supply, the transistorproviding the switch function needsonly a small fraction of the currenthandling ability that it would need ifit was switching the main supply.Also, in most applications, it is notnecessary to hold the voltage dropacross the controlling switch to a verylow level to maintain low-dropoutperformance.
Figure 4. Small FET adds shutdown capability to LT1580 circuit
50µs/DIV
2A/DIV
50mV/DIV
2A/DIV
50mV/DIV
50µs/DIV
0.1µF to approximately 1µF will gen-erally provide optimum transientresponse. The value chosen dependson the amount of output capacitancein the system. Although the capacitor'sfinal value is empirically determined,it generally increases as the outputcapactance increases.
In addition to the enhancementsalready mentioned, the reference ac-curacy has been improved by a factorof two, with a guaranteed 0.5% toler-ance. Temperature drift is also verywell controlled. The part usesratiometrically accurate internal di-vider resistors. The part can easilyhold 1% output accuracy overtemperature, guaranteed, while
operating with an input/output dif-ferential of well under 1V.
In some cases, a higher supplyvoltage for the control voltage will notbe available. If the control pin is tiedto the main supply, the regulator willstill function as a conventional LDOand offer a dropout specification ap-proximately 70mV better thanconventional NPN-based LDOs. Thisis the result of eliminating the voltagedrop of the on-die connection to thecontrol circuit that exists in olderdesigns. This connection is now madeexternally, on the PC board, usingmuch larger conductors than are pos-sible on the die.
Figure 2. Transient response of Figure 1’scircuit with adjust-pin bypass capacitor. Loadstep is from 200mA to 4A
Figure 3. Transient response without adjust-pin bypass capacitor. Otherwise, conditionsare the same as in Figure 2
1580_4.eps
+
+
+ C2 220µF 10V
VIN
1
3
2
4VCONT
53.3V
5V
SHUTDOWN
RTN
SENSE
VOUTADJ
R1 110Ω 1%
VOUT = 2.5V
U1 LT1580
Q1 Si9407DY
C3 22µF 25V
C4 0.33µF
R2 110Ω 1%
R3 10k
LOAD
C1 100µF
10V
Linear Technology Magazine • May 1995 15
DESIGN FEATURES
Quad Current-to-Voltage Converter isIdeal for Optical Disk DrivesIntroduction
The LT1311 is a quad current-to-voltage converter designed for thedemanding requirements of photo-diode amplification. A new approachto current-to-voltage conversionprovides excellent DC and AC perfor-mance without external DC trims orAC frequency compensation. TheLT1311 is ideal for converting mul-tiple photodiode currents to voltages,and for general purpose, matchedinverting-amplifier applications.Figure 1 shows the LT1311 pin con-figuration and a typical photodiodeamplifier application.
The LT1311 has the excellent speedand power performance of a current-feedback amplifier with the DCaccuracy and low noise of a voltage-feedback amplifier. The internalfeedback resistor is 20k, resulting ina current-to-voltage gain of 20mV/µA. The −3dB bandwidth is 12MHzand settling time is less than 145nsto 0.1% of final value for a 2V outputstep. The four amplifiers draw only7mA of supply current while operat-ing on all supplies from ±2V (4V total)to ±15V (30V total). The input-re-ferred bias current is typically 75nAand drift is less than 0.5nA/°C. Theinput noise is only 5pA/√Hz. Table 1details the LT1311 performance.
Optical Disk DrivesThere are many types of optical
data storage. All have one thing incommon: the amount of lightreflected off the storage medium indi-cates whether a given data bit is a“one” or a “zero.” The detection of thereflected light requires a photo detec-tor, the most common of which is thephoto diode. The current that flowsthrough a photo diode is proportionalto the amount of light incident on thediode. Optical disk drives usually usea single-chip array of four to eightphoto diodes. These matched photo
diodes provide both position andintensity information to the servo sys-tems that keep the laser focused, ontrack, and at the correct output level.
In read-only optical drives, suchas audio CD players and CD ROMdrives, the laser level is constant andthe amount of reflected light is notcritical. This, combined with the ex-tensive data conditioning done beforerecording, allows automatic gaincontrol and AC coupling of the photo-diode signals. The amplifiers that
convert these photodiode currents touseful signals do not require good DCprecision.
Optical drives that record and play,such as magneto-optical and phase-change drives, require tight control ofthe laser output level. This is becausea high level of laser output is used towrite and a much lower level is usedto read the data. These drives alsoneed to record quickly, limitingthe amount of data conditioning thatcan be done before recording. The
Table 1. LT1311 electrical characteristics
VCC = 10V, VEE = Ground, Bias = 5V, TA = 0°C – 70°C unless otherwise stated.
Parameter Min Typ Max Units
Current to Voltage Gain 19.2 20 20.8 mV/µACurrent to Voltage Gain Drift –70 ppm/°CCurrent to Voltage Gain Mismatch 0.1 %Input Offset Voltage ±150 ±500 µVInput Offset Voltage Drift ±1 µV/°CInverting Input Current 75 250 nAInverting Input Current Drift 0.5 2.5 nA/°COutput Offset Voltage 1.5 5 mVOutput Offset Voltage Drift 10 50 µV/°COutput Noise Voltage Density, 1kHz 100 nV/√HzInput Noise Current Density, 1kHz 5 pA/√HzInput Noise Voltage Density, 1kHz 4.5 nV/√HzInput Resistance 0.2 2 ΩInput Impedance, 10MHz 300 ΩPower Supply Rejection Ratio, Vs = ±2 to ±15V 90 103 dB Bias = 0VMaximum Output Swing
Output High, No Load 8.8 9.0 VOutput High, 10mA Load 8.5 8.8 VOutput Low, No Load 1.0 1.2 VOutput Low, 10mA Load 1.2 1.5 V
Maximum Output Current ±30 ±55 mATotal Quiescent Supply Current 7 11 mASlew Rate 80 V/µsSmall Signal Bandwidth 12 MHzRise and Fall Time 35 nsSettling Time, 0.1% of 2V Step 145 ns
by William H. Gross
16 Linear Technology Magazine • May 1995
DESIGN FEATURES
photodiode signals must be DCcoupled into a wide dynamic rangesystem. The amplifiers that convertthese photodiode currents to usefulsignals require excellent DC and ACperformance.
For more details on optical diskdrives, please read the excellent ar-ticle by Praveen Asthana, entitled “ALong Road to Overnight Success,” inthe October, 1994 issue of IEEESpectrum.
PhotodiodeAmplifier Requirements
The read-write optical disk driverequires a fast photodiode current-to-voltage converter with very goodDC accuracy. The bandwidth of theconverter needs to be greater than10MHz and the output must settle to
within 0.5% of the final value in lessthan 200ns for a 100µA input step.The output current of the photo di-odes ranges from about 1µA to 100µA;a conversion gain of 20mV/µA re-sults in an output signal of 2Vpeak-to-peak, which is easy to handleon a single 5V or 10V supply. Theinitial offset errors of the photo di-odes and converters are easilytrimmed out at room temperature;however, the input-referred offset driftof the current-to-voltage convertermust not exceed 10% of the mini-mum input signal. For a 1µA inputand a 40°C maximum change in op-erating temperature, the convertermust have an input-referred offsetdrift of less than 2.5nA/°C. Hence, fora 20mV/µA conversion gain, the out-put offset voltage drift must be lessthan 50µV/°C. Additionally, there isa physical size constraint: fourcomplete converters in a small, sur-face-mount package would be idealfor mounting close to the diode array.
Traditional SolutionsMost photodiode current-to-volt-
age converters use the invertingamplifier circuit of Figure 2. The20mV/µA conversion gain implies a20k feedback resistor. Diode capaci-tance and/or stray capacitance ofjust 5pF combined with a 20k resistorresults in a pole at 1.6MHz. To movethe pole out to a higher frequency, asmaller resistor can be used, but thelost gain must be made up some-where. The additional voltage gainwould cause more output offset drift.Operating the amplifier at unity gaingives the best DC performance. Witha 20k feedback resistor, the pole dueto the diode capacitance must becanceled by the feedback capacitorin order to use a fast op amp.This cancellation must be quiteaccurate in order to get fast outputsettling. The diode and stray capaci-tance are not well controlled andthe worst case settling time is deter-mined by the mismatch in thepole-zero cancellation.
Figure 3. Basic current-feedback amplifier
Figure 2. Inverting op amp I-to-V converter
Current-FeedbackPhotodiode Amplifiers
If a current-feedback amplifier isused for the op amp in Figure 2, thefeedback capacitor becomes unnec-essary. To understand why, refer tothe simplified schematic of a current-feedback amplifier in Figure 3. Theinverting input of the amplifier is thejunction of the emitters of Q3 and Q4,and therefore a low impedance. Thepole formed by the capacitance at theinverting input is usually many timeshigher than the bandwidth of theamplifier, and therefore has almostno effect on settling time. Using acurrent-feedback amplifier eliminatesthe need to cancel the diode capaci-tance. For example, adding 50pF tothe input of the LT1311 only increasesthe settling time by a factor of two.This makes it feasible to locate the
1311_1.eps
–
+
–
+
–
+
–
+
10V
0.1µF
0.1µF
1
2
3
4
5
6
7
14OUT A
OUT B
OUT C
10V, VCC
OUT D
13
12
11
10
9
8
IN A
IN B
LT13115V BIAS
IN C
IN D
1311_2.eps
–
+OUT
RFEEDBACK
CFEEDBACK
CSTRAY
IDIODE
CDIODE
1311_3.eps
V+
V–
D1
D2
+IN
–IN
OUT
C
BUFFER
B I A S
RFEEDBACK
Q1
Q2
Q4
Q3
Figure 1. Photodiode current-to-voltageconverter
Linear Technology Magazine • May 1995 17
DESIGN FEATURES
Figure 4. LT1311 circuit concept
The LT1311 CircuitIn the basic LT1311 circuit of Fig-
ure 4, two current sources and Wilsonmirrors are used as the bias for theinput transistors Q3 and Q4. In thiscircuit the alpha errors are elimi-nated (first-order) and only the alphamatching between similar types oftransistors generates an invertinginput bias current. In a typical ICprocess, the beta matching of identi-cal transistors is better than 5%. IfQ3 and Q3B have a beta mismatch of5% and all the other transistors areperfectly matched, the inverting in-put bias current is 0.024% of thecollector currents. This small cur-rent is less than the other sources ofinput bias current. The current mir-rors are the largest source of DCerrors in a current-feedback ampli-fier and the LT1311 dramaticallyimproves the mirrors.
Figure 4 shows the basic idea ofreplacing the D1/Q1 current mirrorwith R1, R2, OA1, and Q5 and simi-larly replacing D2/Q2 with R3, R4,OA2, and Q6. There are three sourcesof input bias current due to this newcurrent mirror. The first results fromthe difference in the ratio of R1:R2and R3:R4. Note that the absolutevalue of this ratio does not generate
any input bias current. In standardIC processes with thin-film resistors,the resistor ratio matching is betterthan 0.1%.
The next source of input bias cur-rent in this new mirror is the differencein input offset voltage between OA1and OA2. The magnitude of this cur-rent is the difference in the two opamps’ offset voltages divided by R1.In a typical IC process, the op ampoffsets will match within 2mV, and atypical voltage drop across R1 wouldbe 200mV; therefore the input biascurrent due to OA1/OA2 mismatchwould be 1% of the collector current.This is ten times larger than the con-tribution due to the resistor-ratiomismatch.
The last source of input bias cur-rent in the mirrors is the alphamismatch of Q5 and Q6. The alphaerrors of Q5 and Q6 are canceled(first-order) by the base currents ofQ8 and Q9. Therefore, only the mis-match in current gain between twosimilar transistors causes input biascurrent. For a typical beta of 200,with a worst-case beta mismatch of5%, the input bias current would be0.02% of the collector current, so thiscontribution is very small.
amplifier a short distance away fromthe photo diodes.
The bandwidth of a current-feed-back amplifier is determined by thefeedback resistor and the internalcompensation capacitor. Hence, if thefeedback resistor that gives the de-sired gain also gives the desiredbandwidth, everything is OK. Unfor-tunately, most commercial CFAs areoptimized for feedback resistors of1k or less in order to make straycapacitance less of a problem. Forexample, if we use an LT1217 orLT1223 with a 20k feedback resistor,the bandwidth will be less than2MHz. In addition to the low band-width problem, there is a problemwith inverting-input bias-currentdrift. Even the low-current LT1217,with its guaranteed input bias cur-rent of less than 500nA, cannotguarantee less than 2.5nA/°C of biascurrent drift. In order to take advan-tage of the AC performance of currentfeedback, an improvement to theLT1217 circuit is required.
The Current Feedback CircuitReferring to the basic current-feed-
back amplifier schematic in Figure 3,we see that the error current thatflows in the feedback resistor is mir-rored by D1/Q1 and by D2/Q2 beforeit goes to the compensation capaci-tor. For a given bandwidth, increasingthe gain of the current mirrors D1/Q1 and D2/Q2 increases the size ofthe feedback resistor. In the 12MHzLT1311, the mirror has a gain ofthree, the compensation capacitor is2pF, and the feedback resistor is 20k.
Again referring to Figure 3, we canlook for the sources of DC error. Theinput offset voltage (and drift) of thisamplifier can be very low with theproper biasing; however, the invert-ing input bias current is hard tocontrol because it is the differencebetween the emitter currents of Q3and Q4. There are three things thatgenerate inverting input bias cur-rent: the mismatch in the alphas ofQ3 and Q4, the mismatch in thegains of current mirrors D1/Q1 andD2/Q2, and the input bias current ofthe output buffer.
1311_4.eps
–
+
I1
V+
R1 R2 R5
Q5Q7
Q8
–
+
I2
V–
R3
C
V–
V–
V+
OA1
OA2
RF
–IN
R4 R6
Q6
Q4A Q4B
Q4
Q3BQ3A
Q3
Q12
OUT+IN
PTAT
PTAT
Q11
Q10
Q9
18 Linear Technology Magazine • May 1995
DESIGN FEATURES
In summary, the new op amp-based mirror reduces the inverting-input bias current to about 1% of thecollector current, making it compa-rable to the input bias current of avoltage-feedback op amp. More im-portantly, the drift of the input biascurrent is very predictable. Mismatchin resistor ratios generates an inputbias current that has the same tem-perature coefficient as the collectorcurrents of Q3 and Q4. The op ampoffsets generate bias current with atemperature coefficient of exactly∆VBE divided by resistance. This isoften called “proportional-to-absolutetemperature,” or “PTAT.” If the collec-tor current in Q3 and Q4 is also PTAT,then the drift of the inverting inputbias current will be PTAT.
If the offset voltage of one of the opamps in one of the mirrors is adjusteduntil the inverting-input bias cur-rent is zero, the current will stay atzero because both op amps’ VOS havethe same drift: PTAT. This is very
Figure 5. Simplified schematic of one LT1311 amplifier
powerful, because it is not necessaryto actually know the offset of OA1 orOA2 in order to eliminate the driftthey cause; it is only necessary totrim their offsets so that the invert-ing-input bias current is zero.
In the simplified schematic of theLT1311 (Figure 5), we can see how allof this comes together. All four ampli-fiers on the chip are complete andidentical; only the supplies and thenoninverting inputs are common. Themain bias current source, I BIAS, whichis used to generate source and sinkcurrent sources for the rest of thecircuitry, is PTAT. Q23– Q27 and C1make up the signal portion of op ampOA1, and Q14, R13, Q15, R14, Q6,and R5 make up the bias portion ofOA1. Similarly Q35– Q39 and C2make up the signal portion of OA2whereas Q4, R3, Q5, R4, Q16, andR15 make up the bias portion of OA2.As mentioned earlier, the LT1311current mirrors have a gain of three,set by R6 and R9 (R16 and R19). This
allows a 2pF compensation capacitorto work with the 20k feedback resis-tor to set a 12MHz bandwidth.
The thin-film resistors in thebiasing circuitry of each amplifier arelaser trimmed at wafer sort. The off-set voltage of OA1 is trimmed in onedirection by R13 and in the otherdirection by R14. Similarly, R3 andR4 trim the offset of OA2. The ampli-fier input offset voltage is trimmed inone direction by R1 and in the otherdirection by R11. The 20k feedbackresistor is trimmed to set the gain.
The four current-feedback amplifi-ers are packaged in a 14-pin SOpackage with a nonstandard pinout.The four inverting inputs are on oneside of the package; the inputs areseparated by DC supply or bias pinsfor optimum channel separation. Thenoninverting inputs are tied to a com-mon bias point and the outputs areon the other side of the package forminimum output-to-input coupling.
continued on page 21
R1 1k
R2 2k
R3 2k
R4 2k
Q1 Q3 Q4 Q5
R5 1k
R6 4.5k
R11 1k
R12 2k
R13 2k
R14 2k
R15 1k
R16 4.5k
R9 1.5k
R10 1.5k
R22 10Ω
R23 10Ω
R20 1.5k
R19 1.5k
VCC
OUT
VEE
Q13 Q14 Q15
RF 20k
C1 5pF
CC 2pF
C2 5pF
INBIASIBIAS PTAT
Q28
Q33
Q44
Q41
Q39
Q40
Q17
1311_5.eps
Q12 Q16
Q38Q37
Q35 Q36
Q21
Q19
Q22
Q18
Q20
Q6Q23
Q26
Q24
Q25
Q29
Q32
Q42
Q27
Linear Technology Magazine • May 1995 19
DESIGN FEATURES
Figure 1. Functional block diagram, LTC1430circuit architecture
Figure 2a. Gate drive using 5V supply
Figure 3. One resistor sets current limit on the LTC1430
continued from page 1
MOSFETs. The on-chip outputdrivers feature separate power-supply inputs and internal levelshifters, allowing the MOSFET gatedrive to be tailored for logic-levelor standard threshold devices. Thestepped-up gate drive to M1 can begenerated with a simple charge-pumpscheme (Figure 2a), or it can be pro-vided by a low-power, higher-voltagesupply if one is available (Figure 2b).Low on-resistance MOSFETs can beused to minimize dissipation even athigh current levels; this maximizesefficiency in power-conscious designsand allows the elimination of the heatsink in many cases.
External component count in thehigh-current path is minimized byeliminating low-value current-senseresistors. Voltage feedback eliminatesthe need for current sensing undernormal operating conditions, andoutput current limit is sensed bymonitoring the voltage drop acrossthe RDS ON of M1 during its ON state.Current limit is set by specifying theRDS ON of M1 and setting the maxi-mum voltage allowed with a singleexternal resistor at the IMAX pin (Fig-ure 3). Current limit can also bedisabled if desired by tying the IMAXpin to ground. The current-limit cir-cuit is designed to engage slowly undermild transient overloads and to kickin more quickly to prevent componentdamage under severe overcurrent andshort-circuit conditions. Current-limit recovery time is set by theexternal soft start capacitor, provid-ing a controlled return to full outputvoltage after the fault is removed.
Performance FeaturesThe LTC1430 uses a voltage feed-
back loop to control output voltage. Itincludes two additional “safety belt”internal feedback loops to improvehigh-frequency transient response(Figure 4). The MAX loop respondswithin a single clock cycle when theoutput exceeds the set point by morethan 3%, forcing the duty cycle to 0%and holding M2 on continuously un-til the output drops back into theacceptable range. Similarly, the MINloop kicks in when the output sags3% below the set point, forcing theLTC1430 to 90% duty cycle until theoutput recovers. The 90% maximumensures that charge-pump drive con-tinues to be supplied to the topMOSFET driver, preventing the gatedrive to M1 from deteriorating duringextended transient loads. The MAXfeedback loop is always active,providing a measure of protection even
if the 5V input supply is accidentallyshorted to the lower microprocessorsupply. Under this condition, M2 willcrowbar the low supply to groundthrough the inductor until the mainsupply fuse blows or the higher sup-ply goes into current limit. The MINloop is disabled at start-up or duringcurrent limit to allow soft start tofunction and to prevent MIN fromtaking over when the current-limitcircuit is active.
The LTC1430 includes an onboardreference trimmed to 1.265V ±10mVand an onboard 0.1% resistor-dividerstring that provides a fixed 3.3V out-put. External resistors can be used togenerate other output voltages. Notethat a pair of 1% resistors will add2% to the output-error budget; 0.1%resistors are recommended forapplications that require very tight
Figure 2b. Gate drive using 5V and 12V supplies
1430_1.eps
3.3VOUT
COUTM2
M1
5V
LTC1430
L1
1430_2a.eps
VOUT = 3.3V
0.1µF
M2
M1
PVCC1
1N4148
PVCC2
G1
G2
PVCC = 5V
LTC14301430_2b.eps
VOUT = 3.3V
M2
M1
PVCC1PVCC2
G1
G2
PVCC = 5V12V
LTC1430
–+
CURRENT LIMIT
1430_3.eps
M2
M1
IFB
1k
10µA
PVCC = 5V
0.1µF
IMAX
RIMAX
LTC1430
20 Linear Technology Magazine • May 1995
DESIGN FEATURES
A Typical5V-to-3.3V Application
The typical application for theLTC1430 is a 5V-to-3.xV converteron a PC motherboard. The output isused to power a Pentium, P6, or simi-lar class processor, and the input istaken from the system 5V ±5% sup-ply. The LTC1430 provides theprecisely regulated output voltagerequired by the processor without theneed for an external precision refer-ence or trimming. Figure 5 shows atypical application with a 3.30V ±1%output voltage and a 12A output-current limit. The power MOSFETsare sized so as not to require a heatsink under ambient temperature con-ditions up to 50°C. Typical efficiencyis above 91% from 1A to 10A outputcurrent, and peaks at 95% at 5A(Figure 6).
The 12A current limit is set by the16k resistor R1 from PVCC to IMAX, andthe 0.035Ω on-resistance of theMTD20N03HL MOSFETs (M1a, M1b).
Figure 4. Two additional feedback loops improve the transient response of the LTC1430
Figure 5. Typical 5V-to-3.3V, 10A LTC1430 application
output tolerances. The LTC1430specifies load regulation of ±15mVand line regulation of ±3mV, result-ing in a total worst-case output errorof ±1.6% when used with the internaldivider or 0.1% external resistors.The internal reference will drift anadditional ±5mV over the 0°C–70°Ctemperature range, providing a ±2.0%total error budget over this tempera-ture range.
The LTC1430 includes a versatileinternal oscillator that can be set tofree run at any frequency between100kHz and 500kHz, or synchronizedto an external clock signal. The oscil-lator runs at a nominal 200kHzfrequency with the FREQ pin float-ing. An external resistor from FREQto ground will speed up the internaloscillator, up to a maximum operat-ing frequency of 500kHz; a resistor toVCC will slow the oscillator to below100kHz. The internal oscillator canbe synchronized to an external clocksignal by setting the free-running fre-quency to slightly slower than thesynchronizing clock frequency andapplying the clock signal to the SDpin. The LTC1430 will shut downonly if the SD pin is low continuouslyfor more than 50µs. In shutdownmode, the power-supply currentdrawn by the LTC1430 drops to below
1µA. When the shutdown pin isbrought high again, the LTC1430 willrun through a soft start cycle andresume normal operation.
–+
+
+
+
1430_4.eps
OSCPVCC1
90% DC
G1
PVCC2
G2
PGND
FB
60mV60mV
FB MIN MAX
PWM
COMP
1.24V
+
+
+
1430_5.eps
VOUT 3.3V
C3 0.1µF
M2 MTD20N03HL
CIN 220µF 10V × 4
L1 2.5µH/15A
COUT 330µF 6.3V × 6
M1B MTD20N03HL
M1A MTD20N03HL
PVCC1IMAX
FREQ
SGND
SHDN
COMP NC
SS
PVCC2SVCC
PGND
PGNDSGNDPGND AND SGND GROUND CONNECTED AT A SINGLE POINT
L1 = CIN =
COUT =
6 TURNS #16 WIRE ON MICROMETALS T50-52B CORE 4 EACH AVX TPSE 227M010R0100 6 EACH AVX TPSE 337M006R0100
*TRIM TO OPTIMIZE TRANSIENT REPONSE
SGND
R1 16k
RC* 33k
C1 0.1µF
CC* 3300pF
100pF*
CSS 0.01µF
C2 10µF
R2 100Ω
D1 1N4148
G1
IFB
G2
VTRIM
+SENSESHUTDOWN
NC
R3 1k
VIN 4.5V TO 5.5V
LTC1430
–SENSE
Linear Technology Magazine • May 1995 21
DESIGN FEATURES
tend to do), the inductor value willneed to be quite low, in the 1µH–10µHrange.
Loop compensation is critical forobtaining optimum transient re-sponse with a voltage-feedbacksystem like the LTC1430; the com-pensation components shown heregive good response when used withthe output capacitor values andbrands shown (Figure 7). The ESR ofthe output capacitor has a significanteffect on the transient response ofthe system. For best results, use thelargest value, lowest ESR capacitorsthat will fit the design budget andspace requirements. Several smallercapacitors wired in parallel can helpreduce total output capacitor ESR toacceptable levels. Input bypass ca-pacitor ESR is also important to keepinput supply variations to a mini-mum with 10AP–P square-wave currentpulses flowing into M1. AVX TPS-series surface-mount tantalumcapacitors and Sanyo OS-CON or-ganic electrolytic capacitors arerecommended for both input and out-put bypass duty. Low cost “computergrade” aluminum electrolytics typi-cally have much higher seriesresistance and will significantly de-grade performance. Don’t count onthat parallel 0.1µF ceramic cap tolower the ESR of a cheap electrolyticcap to acceptable levels.
Figure 7. Transient response: 0A-to-5A loadstep imposed on Figure 5’s output
1311, continued from page 18 ConclusionThe LT1311 is a new current-to-
voltage converter that solves theoptical disk drive photodiode ampli-fier problem with new circuit design,complementary bipolar processing,and laser trimming. The new circuitprovides current-feedback AC per-
formance and low power consump-tion with the DC precision ofvoltage-feedback amplifiers. Otherapplications that require matchedinverting amplifiers, such as colorscanners, will also benefit from theLT1311's performance.
The output-to-input stray capacitancemust be less than 0.5pF for propersettling performance; this pinout en-sures that the input and outputprinted circuit board traces are farapart.
The 0.1µF capacitor in parallel withR1 improves power-supply rejectionat IMAX, providing consistent current-limit performance when voltage spikesare present at PVCC. Soft start time isset by CSS; the 0.01µF value shownreacts with an internal 10mA pull-upto provide a 3ms start-up time. The2.5µH, 15A inductor is sized to allowthe peak current to rise to the fullcurrent-limit value without saturat-ing. This allows the circuit towithstand extended output short cir-cuits without saturating the inductorcore. The inductor value is chosen asa compromise between peak ripplecurrent and output-current slew rate,which affects large-signal transientresponse. If the output load is expectedto generate large output-currenttransients (as large microprocessors
ConclusionThe LTC1430 fits neatly into the
power-supply niche created by theadvent of new technology, power-sup-ply-critical microprocessors. Its tight,no-trims output-voltage tolerance,and simple, low external-parts-counthookup make it a good fit on high-end PC motherboards or plug-inmodules. Superior protection fea-tures, both for the power supply itselfand for the circuitry connected to it,help maximize system reliability, es-pecially in user-upgradable systemswhere unskilled screwdrivers arelikely to be roaming around. Highoverall efficiency reduces the heatgenerated by the power supply, mini-mizing cooling and heat sinkingrequirements and reducing the powerdrawn by “green” systems. The de-sign of the LTC1430, combined withLinear Technology’s unparalleled ap-plications support, simplifies the jobof powering today’s high performancemicrocomputers.
LOAD CURRENT (A)
40
70
100
90
80
50
60EFFI
CIEN
CY (%
)
10
1430_6.eps
0.1 1
VCC = 5V TA = 25°C VOUT = 3.3V
Figure 6. Efficiency plot for Figure 5’s circuit.Note that efficiency peaks at a respectable 95%
5A/div
20mV/div
22 Linear Technology Magazine • May 1995
DESIGN FEATURES
Humidity Sensor to Data AcquisitionSystem Interface
DESIGN IDEAS...Humidity Sensor toData Acquisition SystemInterface ....................... 22Richard Markell
Low-Power Signal Detectionin a Noisy Environment . 24Philip Karantzalis and Jimmylee Lawson
High Output-CurrentBoost Regulator ............ 26Dimitry Goder
LT1111 Isolated 5VSwitching Power Supply...................................... 27Kevin R. Hoskins
High-Efficiency ELDriver Circuit ............... 28Dave Bell
Adding Features tothe Boost Topology ....... 30Dimitry Goder
Bandpass Filter HasAdjustable Q ................. 31Frank Cox
Sallen and Key FiltersUse 5% Values............... 32Dale Eagar
Simple Battery ChargerRuns at 1MHz ............... 34Mitchell Lee
Lithium-IonBattery Charger ........... 35Dimitry Goder
Three-Cell to 3.3VBuck-Boost Converter ... 36Dimitry Goder
High Output-VoltageBuck Regulator ............. 37Dimitry Goder
IntroductionIt can be difficult to interface hu-
midity sensors to data acquisitionsystems because of the sensors’ driverequirements and their wide dynamicrange. By carefully selecting the de-
vices that comprise the analog frontend, users can customize the circuitto meet their humidity-sensing re-quirements and achieve reasonableaccuracy throughout the chosenrange. This Design Idea details theanalog front-end interface between aPhys-Chem Scientific Corp.1 modelEMD-2000 humidity sensor and auser selected (probably microproces-sor-based) data acquisition system.
Design ConsiderationsThe Phys-Chem humidity sensor
is a small, low-cost, accurate resis-tance-type relative humidity (RH)sensor. This sensor has a well de-fined, stable response curve and canbe replaced in circuit without systemrecalibration.
The design criteria call for a low-cost, high-precision analog front endthat requires few calibration “tweaks”and operates on a single 5 volt sup-ply. The sensor requires a squarewave or sine wave excitation with noDC component. The sensor reactancevaries over an extremely wide range(approximately 700Ω–20MΩ). Thewide dynamic range (approximately90dB) required to obtain the full RHrange of the sensor results in somechallenges for the designer.
The circuit shown in the schematicfeatures zero-drift operational ampli-fiers (LTC1250 and LTC1050) and aprecision instrumentation switched-capacitor block (LTC1043). Thisdesign will maintain excellent DCaccuracy down to microvolt levels.This method was chosen over the useof a true RMS-to-DC or log converterbecause of the expense and tempera-ture sensitivity of these parts.
Circuit DescriptionFigure 1 is a schematic diagram
of the circuit. Only a single 5 voltpower supply is required. Integratedcircuit U1, an LTC1046, converts the
5 volts supply to −5 volts to supplypower to U2, U3, and U4. U2A, part ofan LTC1043 switched-capacitorbuilding block, provides the excita-tion for the sensor, switching between5 volts and −5 volts at a rate ofapproximately 2.2kHz. This rate canbe varied, but we recommended thatit be kept below approximately2.4kHz, which is one-half the auto-zero rate of U3. We believe the de-viation from the Phys-Chem responsecurves taken at 5kHz is insignificant.
Variable resistor R2 sets the full-scale output. Since the sensorresistance is 700Ω at approximately90% humidity, setting R2 at 700Ωwill provide a 2:1 voltage divider that,when combined with the gain of U4(×2), results in an overall gain of one.U3 must be included in order for thecircuit to function properly; other-wise C4 and C7 form a voltage dividerthat is dependent on the resistance ofthe RH sensor. U3 is a precision auto-zero operational amplifier with anauto-zero frequency of approximately4.75kHz. U2B (the “lower” switch)samples the output of U3 and pro-vides this sample to the input of U4.U4 is set to provide a gain of two.
It is easy to digitize the output ofU4. Figure 2 is the schematic of a 12-bit converter that can be used for thispurpose. The range of humidity thatcan be sensed depends on the resolu-tion of the converter. The full-scaleoutput (which is equivalent toapproximately 90% humidity) is es-sentially independent of the numberof bits in the A/D converter, but thedry (low RH) end of the scale is depen-dent on the A/D resolution. As anexample, the above referenced 12-bitconverter will process humidity sig-nals that translate to approximately20% RH, since the voltage output atthis humidity is approximately 2.3millivolts, while 1/2 LSB is 1.2 milli-volts. Digitization down to 10% RH
DESIGN IDEAS
by Richard Markell
Linear Technology Magazine • May 1995 23
DESIGN FEATURES
requires the conversion of 350µV sig-nals or a 16-bit converter. From a coststandpoint this seems unwieldy. It ismuch more economical to use a two-channel 12-bit converter that changesranges somewhere in the humidityrange.
All of the above solutions measureoutput voltage from a voltage dividerconsisting of the RH sensor and afixed “calibration” resistor. The resis-tance of the sensor at a fixed output
voltage can be calculated from theformula
In this case, if R2 is set to 700ohms, and VFULL SCALE = 5.00V, then
be calculated from the quadratic ap-proximation in the Phys-Chemliterature:R2 VFULL SCALE
VOUT/2− R2R (Ohms) =
3500VOUT/2
− 700R (Ohms) =
Once R is calculated (probably bythe microprocessor), the humidity can
1. Phys-Chem Scientific Corp.36 West 20th StreetNew York, NY 10011(212) 924-2070—Phone(212) 243-7352—FAX
If a suitable humidity chamber isnot available, the sensor can be re-moved and fixed resistors substituted.The circuit should then be calibratedfrom the EMD-2000 “typical responsecurve.” This should provide approxi-mately 2% accuracy.
RH = LnR − 14.06 − √(14.06 − LnR)2 + 15.56−0.176
Figure 2. LTC1291 12 bit A/D converter interfaced to MC68HC11
DESIGN IDEAS
dIhumi_1.eps
BOOST V+
C+ OSC
GND
NOTES: UNLESS OTHERWISE SPECIFIED 1. ALL RESISTANCES ARE IN OHMS, 1/4 W 5% *C9 ADJUSTS OSC. FREQUENCY 2000pF YIELDS ~ 2.2kHz
LV
C–
1
2
2
3
7
6 13
17
16
11
8
14
12
72 7
6
4
3
4
4
3
4
8
7
6
5VO
U1 LTC1046
5V
C1 10µF C8
0.01µF
C4 1µF
R1 VOUTHUMIDITY SENSOR
C7 1µF
*C9 2000pF
–
+
–
+
R2 1kCALIBRATION
C3 0.1µF
C2 10µF
+
U2 LTC1043
U4 LTC1050
R3 10k
C8 1µF
C5 62pF
+
U3 LTC1250
R4 10k
Figure 1. Schematic diagram of humidity-sensor circuit
5V
CH0
GND DIN
DOUT
CLK
VCC(VREF)CS
LTC1291
1
2
3
4
8
7
6
5
2-CHANNEL MUX*
FROM VOUT LTC1050
CH1
+
0.1µF
22µF TANTALUM
*FOR OVERVOLTAGE PROTECTION LIMIT THE INPUT CURRENT TO 15mA PER PIN OR CLAMP THE INPUTS TO VCC AND GND WITH 1N4148 DIODES.
dIhumi_2.eps
DO
MC68HC11
SCK
MISO
MOSI
24 Linear Technology Magazine • May 1995
DESIGN FEATURES
Low-Power Signal Detectionin a Noisy EnvironmentIntroduction
In signal-detection applicationswhere a small narrowband signal isto be detected in the presence of wide-band noise, one can design anasynchronous (non-phase-sensitive)tone detector using an ultra-selectivebandpass filter, such as theLTC1164-8. The ultra-narrow pass-band of the LTC1164-8 filterband-limits any random noise andincreases the detector’s signalsensitivity.
The LTC1164-8 is an eighth-order,elliptic bandpass filter, with the fol-lowing features: the filter’s fCENTER(the center frequency of the filter’spassband) is clock tunable and isequal to the clock frequency dividedby 100; the filter’s passband is from0.995 × fCENTER to 1.005 × fCENTER(±0.5% from fCENTER). Figure 1 showsa typical LTC1164-8 passband re-sponse and the area of passband-gainvariation. Outside the filter’s pass-band, signal attenuation increases tomore than 50dB for frequencies be-tween 0.96 × fCENTER and 1.04 × fCENTER. Quiescent current is typi-cally 2.3mA with a single 5V powersupply.
An Ultra-Selective BandpassFilter and a Dual ComparatorBuild a High-PerformanceTone Detector
The LTC1164-8 has excellent se-lectivity, which limits the noise thatpasses from the input to the output ofthe filter. As a result, one can build atone detector that can extract smallsignals from the “mud.” Figure 2shows the block diagram of such atone detector. The detector’s input isan LTC1164-8 bandpass filter whoseoutput is AC coupled to a dual com-parator circuit. The first comparatorconverts the filter’s output to avariable pulsewidth signal. Thepulsewidth varies depending on thesignal amplitude. The average DCvalue of the pulse signal is extractedby a lowpass RC filter and applied tothe second comparator. The identifi-cation of a tone is indicated by a logichigh at the output of the second com-parator.
One of the key benefits of using ahigh-selectivity bandpass filter fortone detection is that when widebandnoise (white noise) appears at theinput of the filter, only a small amountof input noise will reach the filter’soutput. This results in a dramaticallyimproved signal to noise ratio at theoutput of the filter compared to thesignal-to-noise at the input of the
Figure 2. Tone detector block diagramFigure 1. Detail of LTC1164-8 passband
(S/N)OUT
(S/N)IN= 20 Log (BW)IN
(BW)f
DESIGN IDEAS
filter. If the output noise of theLTC1164-8 is neglected, the signal tonoise ratio at the output of the filterdivided by the signal to noise ratio atthe input of the filter is:
where: (BW) in = the noise bandwidthat the input of the filter and (BW)f =0.01 × (fCENTER) is the filter’s noiseequivalent bandwidth.
For example, a small 1kHz signal issent through a cable that is also con-ducting random noise with a 3.4kHzbandwidth. An LTC1164-8 is used todetect the 1kHz signal. The signal-to-noise ratio at the output of the filter is25.3db larger than the signal-to-noiseratio at the input of the filter:
Figure 3 shows the complete cir-cuit for a 1kHz tone detector operatingwith a single 5V supply. AnLTC1164-8 with a clock input set at100kHz sets the tone detector’s fre-quency at 1kHz (fCENTER = fCLK/100). Alow-frequency op amp (LT1013) andresistors RIN and RF set the filter’sgain. In order to minimize the filter’s
3.4kHz0.01 × 1kHz
= 20 Log(BW)IN
(BW)f= 25.3dB
–
+
–
+
LTC1164-8
fCLK
ULTRA NARROW BANDPASS FILTER
WITH GAINVIN
fIN = fCLK 100
REF 1
REF 2
COMPARATOR 2
VARIABLE PULSE WIDTH OUTPUT
PULSE AVERAGE
LOGIC HI WHEN SIGNAL PRESENT LOGIC LO WHEN NO SIGNAL PRESENT
dI1164_2.eps
COMPARATOR 1
AC BUFFER
LTC1164-8 PASSBAND (fCENTER = fCLK/100)
RF = 61.9k RIN = 340k
PERCENT DEVIATION FROM fCENTER
–18
–12
–15
–9
–3
–6
0
3
GAI
N (d
B)
dI1164_1.eps
–1.0 –0.75 –0.5 –0.25fCENTER
0.25 0.50 0.75 1.0
AREA OF PASSBAND GAIN VARIATION
by Philip Karantzalisand Jimmylee Lawson
Linear Technology Magazine • May 1995 25
DESIGN IDEAS
LTC1164-8 is from 0.5V to 3.5V, cen-tered at 2V. The divider also providesthe reference voltages for the LTC1040dual comparators (Ref. 1 = 1.9V andRef. 2 = 1V). Power supply variationsdo not affect the performance of thiscircuit because all DC reference volt-ages are derived from the same resistordivider and will track any changes inthe 5V power supply.
Theory of OperationThe tone detector works by looking
at the negative peaks at the output ofthe filter. Signals below 1.9V at theoutput of the filter trip the first com-parator. The second comparator hasa 1V reference and detects the aver-age value of the output of the firstcomparator. The R3–C2 time constantis set to allow detection only if theduty cycle of the first comparator’soutput exceeds 25%. Waveforms withduty cycles below 25% are arbitrarilyassumed to carry false information
The circuitry is designed so thattwo or more negative signal peaks of160mV at the filter’s output producea 25% duty-cycle pulse waveform atthe output of the first detector (the
1.9V and 1V references for compara-tors 1 and 2 respectively, set the160mVPEAK and the 25% duty cycle).The 25% duty-cycle requirement es-tablishes an operating point or“minimum detectable signal” for thedetector circuit. Thus, the circuitryoutputs a “tone-present” conditiononly when the duty cycle is greaterthan or equal to 25%. The 25% duty-cycle requirement sets two conditionsfor optimum tone detection at thedetector’s input.
The first input condition is themaximum-input-noise spectral den-sity that will not trigger the detector’soutput to indicate the presence of atone. When only noise is present atthe filter’s input, the maximum-input-noise spectral density isconservatively defined as the amountrequired to produce noise peaks atthe filter’s output of 160mV or loweramplitude. The 160mV maximumnoise-peak specification at the filter’soutput can be converted to outputnoise in mVRMS by using a crest factorof 5 (the crest factor of a signal is theratio of its peak value to its RMSvalue—a theoretical crest factor of 5
Figure 3. 1kHz tone detector with gain of 10
output noise and maintain optimumdynamic range, the output feedbackresistor RF should be 61.9k. Capaci-tor CF across resistor RF is added toreduce the clock feedthrough at thefilter’s output.
To set the gain for the LTC1164-8,RIN should be calculated by theequation:
RIN = 340k/gain
In Figure 3, the filter’s gain is 10(RIN = 34k). Capacitor C1 and a unity-gain op amp (LT1013) AC couple thesignal at the filter’s output to anLTC1040 dual low-power compara-tor. AC coupling is required toeliminate any DC offset caused by theLTC1164-8.
A resistive divider generates a 2Vbias for the LTC1164-8 “ground” (pins3 and 5) and the positive input of theLT1013 dual op amps. For single 5Voperation, the output swing of the
–
+
–
+–
–+
+
–
–+
+
COMP 1
5V
RIN, 34kR2, 10k
RF 61.9k
CF 200pF
C1 0.22µF
1
7
86
5
43
2
LTC1164-8
LTC1040
1/2 LT1013
VIN2 11
14
7
AGND3
5
5
1
6REF. 1 (1.9V)
REF. 2 (1V)
18
4
15VOUT
7
8
1
5V
6 8 10 12 13
4
0.1µF
5V
1/2 LT1013
0.1µF
C2 0.47µF
5V
0.1µF
STROBE
fCLK
1.0µF
0.1µF
30.1k
AGND (2V)
REF. 1 (1.9V)
0.1µF
1k
R1 10k
RIN = 340k/GAIN, fCENTER = fCLK/100 (1/(2 π RF CF) ≥ 10 • fCENTER) (1/(2 π R1 C1) ≤ fCENTER/10) (1/(2 π R3 C2) ≤ fCENTER/32)
R3 10k
REF. 2 (1V)
0.1µF
8.87k
10k
COMP 2
14
13
12
11
9 10
dI1164_3.eps
26 Linear Technology Magazine • May 1995
DESIGN IDEAS
predicts 99.3% of the maximum peaksof wideband noise with uniform spec-tral density). Therefore, the maximumallowable noise at the filter’s outputis 32mVRMS (160mVPEAK/5). The noiseat the filter’s output depends on thefilter’s gain and noise equivalentbandwidth and the spectral densityof the noise at the filter’s input.Therefore, the maximum input noisespectral density for Figure 3’scircuit is:
noise of the LTC1164-8 is indepen-dent of the chosen filter signal gain.
The second input condition is theminimum input signal required sothat a tone can be detected when it isburied by the maximum noise, asdefined by the first input condition.When a tone plus noise are present atthe filter’s input, the output of thefilter will be a tone whose amplitudeis modulated by the bandlimited noiseat the filter’s output. If a maximumnoise peak of 160mV modulates thetone’s amplitude, a 320mV tone peakat the filter’s output can be detectedbecause the product of the noise andthe tone crosses the (negative)160mVPEAK detection threshold andthe 25% duty cycle requirement isexceeded. Therefore, a conservativevalue for the minimum signal at the
filter’s output can be set to 320mVPEAKor 226mVRMS, but a value of200mVRMS was established experi-mentally. Therefore, the minimuminput signal for reliable tone detec-tion in the presence of the maximuminput-noise spectral density is:
VIN (MIN.) = 200mVRMS / GainFor optimum tone detection, the
signal’s frequency should be in thefilter’s passband, within ±0.1% offCENTER.
ConclusionA very selective bandpass filter, the
LTC1164-8, can be configured as anon-phase-sensitive tone detector.This allows signals to be detected inthe presence of comparatively largeamounts of noise or signal-to-noiseratios that are less than unity.
eIN ≤ 32mVRMS/(Gain × √(BW)f)VRMS
√Hzwhere: Gain is the filter’s gain at itscenter frequency and (BW)f is thefilter’s noise-equivalent bandwidth.
Note: Compared to 32mVRMS, the270µVRMS output noise of theLTC1164-8 is negligible. The output
High Output-Current Boost Regulator
Low-voltage switching regulatorsare often implemented with self-con-tained power integrated circuitsfeaturing a PWM controller and anonboard power switch. Maximumswitching currents of up to 10A areavailable, providing a convenientmeans for power conversion over wideinput- and output-voltage ranges.However, if higher switching currentsare required, a switching regulatorcontroller with an external powerMOSFET is a better choice.
Figure 1 shows an LTC1147-based5V-to-12V converter with 3.5A peakoutput-current capability. TheLTC1147 is a micropower controllerthat uses a constant off-time archi-tecture, eliminating the need forexternal slope compensation.Current-mode control allows fasttransient response and cycle-by-cyclecurrent limiting. A maximum voltageof only 150 millivolts across the cur-rent-sense resistor R7 optimizesperformance for low input voltages.
When Q2 turns on, current startsbuilding up in inductor L1. This pro-
vides a ramping voltage across R7.When this voltage reaches a thresh-old value set internally in theLTC1147, Q2 turns off and the en-ergy stored in L1 is transferred to theoutput capacitor C5. Timing capaci-tor C2 sets the operating frequency.The controller is powered from theoutput through R5, providing 10V of
gate drive for Q2. This reduces theMOSFET’s on-resistance and allowsefficiency to exceed 90% even at fullload. The feedback network compris-ing R2 and R8 sets the output voltage.Current sense resistor R7 sets themaximum output current; it can bechanged to meet different circuit re-quirements.
Figure 1. LTC1147-based 5V-to-12V converter
+
+
+
dI1147_1.eps
U1 LTC1147
C5,C6 SANYO 0S-CON EFFICIENCY AT 3A, 90%
VOUT 12V/3A
3.5A PEAK
VIN
CT
ITHC1 3300pF
C2 180pF
R1 510Ω
R7 0.01Ω 2%
R8 100k 1%
L1 15µH
C3 0.01µF
C4 100pF
R4 100Ω
R2 11.5k 1%
R3 100Ω
C7 3.3µF
D2 BAT54
C6 220µF 10V × 2
C5 150µF 16V × 2
VIN 5V
SENSE –
1
2
3
4
8
7
6
5
PDRIVE
R6 56k
Q1 VN2222LL
Q3 TP0610L
R5 100Ω
GND
VFB
SENSE +
Q2 IRL2203
D1 MBR735
by Dimitry Goder
Linear Technology Magazine • May 1995 27
DESIGN IDEAS
Figure 1. Circuit generates isolated, regulated 5V at 100mA
LT1111 Isolated 5VSwitching Power SupplyCircuit Description
Many applications require isolatedpower supplies. Examples includeremote sensing, measurement of sig-nals riding on high voltages, remotebattery-powered equipment, elim-ination of ground-loops, and dataacquisition systems where noiseelimination is vital. In each situation,the isolated circuitry needs a floatingpower source. In some cases, batter-ies or an AC line transformer can beused for power. Alternately, theDC–DC converter shown here createsan accurately regulated, isolated out-put from a 5V source. Moreover, iteliminates the opto-isolator feedbackarrangements normally associatedwith fully isolated converters.
Figure 1 shows a switching powersupply that generates an isolated andaccurately regulated 5V at 100mAoutput. The circuit consists of anLT1111, configured as a flyback con-verter, followed by an LT1121low-dropout, micropower linear regu-lator. An LTC1145 (winner of EDN’s
IC Innovation of the Year Award) pro-vides micropower isolated feedback.
The LT1111 is a micropower de-vice, which operates on only 400µA(max). This micropower operation isimportant for energy-consciousapplications. It works well with sur-face-mount inductors such as theCoiltronics Octa-pac shown in theschematic. Although the LT1111’sinternal power switch handles up to1A, a 100Ω resistor (R1) limits thepeak switch current to approximately650mA. This maximizes converter ef-ficiency. One side benefit of limitingthe peak switch current is that thecircuit becomes insensitive toinductance. The circuit operates sat-isfactorily with an inductance in therange of 20µH to 50µH.
It is important that the capacitor(C2 in Figure 1) have low effectiveseries resistance (ESR) and induc-tance (ESL) to minimize output ripplevoltage. Although aluminum capaci-tors are abundant and inexpensive,
they will perform poorly in thisswitcher application because of theirrelatively high ESR and ESL. Thetantalum capacitor shown (C2) haslow ESR and ESL and comes in asurface-mount package. Sanyo’s OS-CON series of capacitors are also goodchoices.
Circuit OperationThe LT1111 is configured to oper-
ate as a flyback converter. The voltageon the transformer’s secondary isrectified by D2, filtered by C2, andapplied to the LT1121’s input. As theLT1121’s input voltage continues torise, its output will regulate at 5V.The LT1121’s input voltage contin-ues increasing until the differentialbetween input and output equalsapproximately 600mV. At this pointQ1 begins conducting, turning onthe LTC1145 isolator. The output ofthe LTC1145 goes high, turning offthe converter. The feedback from theLTC1145 gates the LT1111’s oscilla-tor, controlling the energy transmittedto the transformer’s secondary andthe LT1121’s input voltage. The oscil-lator is gated on for longer periods asthe LT1121’s load current increases.Q1’s gain and the feedback throughthe LTC1145 force the converter loopto maintain the LT1121 just abovedropout, resulting in the best effi-ciency. The LT1121 provides currentlimiting, as well as a tightly regu-lated, low-noise output.
+ + +
C5 0.1µF
di1111_1.eps
IC3 LTC1145
IC2 LT1121CZ5
8
IC1 LT1111
7
6
5
18
1789NCOSC INGND2
VCC
DIN
GND1OSOUTDOUT
121110
1R1
100Ω
R2 30k
2
5V 5V
C1 10µF
C2 47µF
C3 10µF
D2 1N5818
500VRMS ISOLATION BARRIER
Z1 1N5355
D1 MUR120
Q1 2N3906
D3 1N4148
TR1*3
4
FB
SET
A0
GND
ILIM
VIN
SW1
SW2
*COILTRONICS CTX20-1Z
by Kevin R. Hoskins
28 Linear Technology Magazine • May 1995
DESIGN IDEAS
High-Efficiency EL Driver CircuitElectroluminescent (EL) lamps are
gaining popularity as sources of LCD-backlight illumination, especially insmall, handheld products. EL lampsresemble thin sheets of cardboardand are available in a variety of colors.Compared with other backlightingtechnologies, EL is attractive becausethe lamp is thin, lightweight, rugged,and can be illuminated with littlepower. Moreover, light is emitted uni-formly from the entire EL surface, sono diffuser is needed.
EL lamps are capacitive in nature,typically exhibiting around 3000pF/in2, and require a low frequency(50Hz–1kHz) 120VRMS AC drive volt-age. Heretofore, this has usually beengenerated by a low-frequency block-ing oscillator using a largetransformer. These large, inefficientpower modules have been suitable fortraditional EL applications, such asemergency exit signs and instrumentpanel backlights, but such space andpower inefficiencies are unacceptablein lightweight, battery-poweredproducts.
Figure 1 depicts a high-efficiencyEL driver that can drive a relativelylarge (12 in2) EL lamp using a smallhigh-frequency transformer. The cir-cuit is self oscillating, and delivers aregulated triangle wave to the at-tached lamp. Very high conversionefficiency may be obtained using thiscircuit, even matching state-of-the-art CCFL backlights at modestbrightness levels (10–20 foot-lamberts).
Since an EL lamp is basically alossy capacitor, the majority of theenergy delivered to the lamp duringthe charge half-cycle is stored as elec-trostatic energy (1/2CV2). Overallconversion efficiency can be improvedby almost 2:1 if this stored energy isreturned to the battery during thedischarge half-cycle. The circuit ofFigure 1 operates as a flyback con-verter during the charge half-cycle,taking energy from the battery andcharging the EL capacitance. During
the discharge half-cycle, the flybackconverter operates in the reverse di-rection, taking energy back out of theEL lamp and returning it to the bat-tery. Nearly 50% of the energy takenduring the charge half-cycle isreturned during the discharge half-cycle; hence the 2:1 efficiencyimprovement.
During the charge half-cycle, theLT1303 operates as a flyback con-verter at approximately 150kHz,ramping the current in T1’s 10µHprimary inductance to approximately1A on each switching pulse. Whenthe LT1303’s internal power switchturns off, the flyback energy storedin T1 is delivered to the EL lampthrough D3 and C5. Successive high-frequency flyback cycles progressivelycharge the EL capacitance until 300Vis reached on the “+” side of C5. Atthis point, the feedback voltagepresent at the LT1303’s LBI inputreaches 1.25V, causing the internalcomparator to change state.
When the LT1303’s internalcomparator changes state, the open-collector driver at the LBO output isreleased. This places the circuit intodischarge mode, and reverses theoperation of the flyback energy trans-fer. Q3 turns on, removing the gatedrive from Q2A, thereby disablingswitching action on the primary ofT1. Flip-flop U2A is also clocked,resulting in a high level on the Q-baroutput; this positive feedback actionkeeps LBI above 1.25V. Even thoughQ2A is turned off, the LT1303’s SWpin still switches into pull-up resistorR4. The resulting pulses at the SWpin are used to clock U2B and todrive a “poor man’s” current-modeflyback converter on the secondaryof T1.
Every clock pulse to flip-flop U2Bturns on Q2B and draws current fromthe EL lamp through C5, T1, D2, andQ4. (Q4 must be a 600V-rated MOS-FET to withstand the high peakvoltages present on its drain duringnormal operation.) Current ramps
up through T1’s 2.25mH secondaryinductance until the voltage acrosscurrent-sense resistor R12 reachesapproximately 0.6V. At this point Q5turns on, providing a direct clear toU2B and thereby terminating thepulse. Energy taken from the ELlamp and stored in T1’s inductance isthen transferred back to the batterythrough D1 and T1’s primary winding.This cycle repeats at approximately150kHz until the voltage on C5 ratch-ets down to approximately zero volts.Once C5 is fully discharged, the pre-set input on U2A will be pulled low,forcing the voltage on the LT1303’sLBI input to ground, and initiatinganother charge half-cycle.
This circuit produces a trianglevoltage waveform with a constantpeak-to-peak voltage of 300V, but thefrequency of the triangle wave de-pends on the capacitance of theattached EL lamp. A 12 in2 lamp hasapproximately 36nF of capacitance,which results in a triangle wave fre-quency of approximately 400Hz. Thisproduces approximately 17FL of lightoutput from a state-of-the-art ELlamp. Because of the “constant power”nature of the charging flyback con-verter, light output remains relativelyconstant with changes in the batteryvoltage. In addition, since EL lampcapacitance decreases with age, thecircuit tends to minimize brightnessreduction with lamp aging. C5, R9,and R10 maintain a zero average volt-age across the EL lamp terminals—anessential factor for reliable lampoperation.
Two options exist for EL lampswith different characteristics. Largerlamps can be supported by specifyingan LT1305 instead of the LT1303shown in Figure 1. The LT1305 willterminate switch cycles at 2A insteadof 1A, thereby delivering four times asmuch energy (energy stored in T1 isdefined by 1/2LI2). The value of R12must also be reduced to 7.5Ω toincrease the discharge flyback cur-rent by the same ratio. For smaller
by Dave Bell
Linear Technology Magazine • May 1995 29
DESIGN IDEAS
lamps, or for brightness adjustment,the circuit may be “throttled” by con-necting the LT1303/ LT1305’s FB pinto a small current-sense resistor inthe lower leg of the EL lamp. Contact
LTC for circuit details if your applica-tion calls for such brightness control.
Not only does the depicted circuitoperate very efficiently, it takes out-put fault conditions in stride. The
circuit, with C5 rated at 300V, toler-ates indefinite short-circuit andopen-circuit conditions across its ELlamp output pins.
+
5V
C2 10pF
VBATT
C5 4.7µF 160V
C4 47µF 16V
10µH1,2
4,5 10
T1
6
1:15
R6 10Ω
R1 18k
SHDN
D Q
R4 470Ω
LBI
FBQ3
2N7002
SW
LBO
6
81
3
5
7
2
4
SHDN
VIN
GND PGND
U1 LT1303U2A
HC74
C1 220pF
R3 2.2M
Q1 2N3906
R2 2.2M
R5 47k
D1 MBRS140T3
R7 4.7k
VBATT = T1 =
U2 =
5.4 TO 12V DALE LPE5047-A132 (605) 665-9301 POWERED FROM 5V
Q2A 1/2 Si9955
C6 0.022µF
+
R10 1M
EL LAMP (12IN2)
R9 1M
R8 2.2k
Q
D Q
Q
R11 10Ω
Q4 IRFRC20
D2 MURS160T3
D3 MURS160T3
VBATT
U2B HC74
R12 15Ω
Q2B 1/2 Si9955
R13 680Ω
C7 1000pF
Q5 2N3904
C3 0.1µF
dIEL_1.eps (V)
R14 10Ω
Figure 1. High-efficiency EL driver circuit
30 Linear Technology Magazine • May 1995
DESIGN IDEAS
Adding Features tothe Boost Topology
Figure 1. Q1 adds short-circuit limiting, true shutdown and regulation when there is a highinput voltage to the LT1301 in boost mode
Figure 2. Efficiency versus input for voltagesfor 5V/100mA output
A boost-topology switching regula-tor is the simplest solution forconverting a two- or three-cell inputto a 5V output. Unfortunately, boostregulators have some inherent disad-vantages, including no short-circuitprotection and no shutdown capa-bility. In some battery-operatedproducts, external chargers or adapt-ers can raise the battery voltage to apotential higher than the 5V output.Under this condition, a boostconverter cannot maintain regula-tion—the high input voltage feedsthrough the diode to the output.
The circuit shown in Figure 1 over-comes these problems. An LT1301 isused as a conventional boost con-verter, preserving simplicity and highefficiency in the boost mode. Transis-tor Q1 adds short-circuit limiting,true shutdown, and regulation whenthere is a high input voltage.
When the input voltage is lowerthan 4V and the regulator is enabled,Q1’s emitter is driven above its base,saturating the transistor. As a result,the voltages on C1 and C2 are roughlythe same, and the circuit operates asa conventional boost regulator.
If the input voltage increases above4V, the internal error amplifier, act-ing to keep the output at 5V, booststhe voltage on C1 to a level greaterthan 1V above the input. This voltagecontrols Q1 to provide the desiredoutput, with the transistor operatingas a linear pass element. The outputdoes not change abruptly during theswitch-over between step-up andstep-down modes, because it ismonitored in both modes by the sameerror amplifier.
Figure 2 shows efficiency versusinput voltage for 5V/100mA output.The break point at 4.25V is evidenceof Q1 beginning to operate in a linear
mode, with an attendant roll-off ofefficiency. Below 4.25V the circuitoperates as a boost regulator, andmaintains high efficiency across abroad range of input voltages.
The circuit can be shut down bypulling the LT1301’s shutdown pinhigh. The LT1301 ceases switchingand Q1 automatically turns off, fullydisconnecting the output. This staystrue over the entire input voltagerange.
Q1 also provides overload protec-tion. When the output is shorted, theLT1301 operates in a cycle-by-cyclecurrent limit. The short-circuit cur-rent depends on the maximum switchcurrent of the LT1301 and on theQ1’s gain, typically reaching 200mA.The transistor can withstand over-load for several seconds, beforeheating up. For sustained faults, thethermal effects on Q1 should be care-fully considered.
+ +
+
VOUT (5V/100mA)Q1
ZTX788B
MBR0520L
R1 1.5k
L1 22µH
dI1301_1.eps
6 7
2 4
3 5
8 1
VIN
SELECT
SHDN
PGND
SW
SENSE
ILIM
GND
SHUTDOWN
VIN (2V-9V)
C3 33µF
LT1301
C1 47µF
R2 3.3k
C2 100µF
INPUT VOLTAGE (V)
50
60
70
80
90
100
EFFI
CIEN
CY (%
)
9
dI1301_2.eps
2 3 4 5 6 7 8
BOOST RANGE
LINEAR STEP- DOWN RANGE
by Dimitry Goder
Linear Technology Magazine • May 1995 31
DESIGN IDEAS
H(s) = 1 + A(s) B(s)
A(s)
where A(s) is the forward gain andB(s) is the reverse gain. The forwardgain is the product of the transcon-ductance stage gain (gm) and the gainof the CFA (ACFA). For this circuit, gmis ten times the product of ISET andthe impedance of the tank circuit as afunction of frequency. This gives the
A(s) = 10 ISET ACFAsL
1 + s2 LC
The reverse gain is simply:
and ACFA =
B(s) = R7R6 + R7
R4R4 + R5
Setting B(s) =ACFA
RRATIO=1
and substituting these expressionsinto the first equation gives:
1 + 10 ISET sL
1 + s2 LC
10 ISET sL
1 + s2 LC
RRATIO
1H(s) =
The last equation can be rewritten as:
Comparing the last two equationsnote that
10 ISET √LC C
1√LC
ω° = and1Q
=
And therefore
S (ω°/Q)S2 + S (ω°/Q) + ω°
2H(s) = HBP
10 ISET √LC CQ =
The bandpass-filter circuit shownin Figure 1 features an electronicallycontrolled Q. Q for a bandpass filteris defined as the ratio of the 3dB passbandwidth to the stop bandwidth atsome specified attenuation. The cen-ter frequency of the bandpass filter inthis example is 3MHz, but this can beadjusted with appropriate LC-tankcomponents. The upper limit of theusable frequency range is about10MHz. The width of the passband isadjusted by the current into pin 5 (setcurrent or ISET) of the transconduc-tance amplifier segment of IC1, anLT1228. Figure 2 (page 33) is a net-work-analyzer plot of frequencyresponse verses set current. This plotshows the variation in Q while thecenter frequency and the passbandgain remain relatively constant.
The circuit’s operation is best un-derstood by analyzing the closed-looptransfer function. This can be writtenin the form of the classic negative-feedback equation:
complete expression for the forwardgain as a function of frequency:
It can be seen from the last equationthat the Q is inversely proportional tothe set current.
Many variations of the circuit arepossible. The center frequency of thefilter can be tuned over a small rangeby the addition of a varactor diode. Toincrease the maximum realizable Q,add a series LC network tuned to thesame frequency as the LC tank on pin1 of IC1. To lower the minimum ob-tainable Q, add a resistor in parallelwith the tank circuit. To create avariable-Q notch filter, connect theinductor and capacitor at pin 1 inseries rather than in parallel.
A variable-Q bandpass filter canbe used to make a variable-band-width IF or RF stage. Anotherapplication for this circuit is as avariable-loop filter in a phase-lock-loop phase demodulator. Thevariable-Q bandpass filter is set for awide bandwidth while the loop ac-quires the signal and is then adjustedto a narrow bandwidth for best noiseperformance after lock is achieved.
1. Thanks to Doug La Porte for this equation hack.
The transfer function of a secondorder bandpass filter can be expressedin the form1:
Figure 1. Circuit diagram: LT1228 bandpass filter
10 ISET √LC
RRATIO
1H(s) =
S2 + S
1√LC
1√LC
C
10 ISET √LC
1√LC C
+
S
Bandpass Filter Has Adjustable Q
–
+
–
+
dI1228_1.eps
R6 750Ω
R4 75Ω
R5 750Ω
gm (LT1228)
CFA (LT1228)
8
1
6
1
53
2 50Ω
R7 75Ω
R2 1k
R1 50Ω
R3 75Ω
ISET
5.3µH 536pF
by Frank Cox
32 Linear Technology Magazine • May 1995
DESIGN IDEAS
Sallen and Key Filters Use 5% Values
Table 2. Butterworth lowpass filter
Freq. R1 R2 R3 C1 C2 C31.0 0.36 3.3 3.3 0.47 0.10 0.0221.1 0.47 0.47 6.2 0.47 0.47 0.0101.2 0.36 0.62 1.0 0.47 0.47 0.0471.3 0.27 2.00 0.33 0.47 0.47 0.0471.5 0.24 1.60 0.3 0.47 0.47 0.0471.6 0.27 0.43 0.82 0.47 0.47 0.0471.8 0.43 1.20 0.13 0.22 1.00 0.0472.0 0.36 7.50 0.18 0.22 0.47 0.0102.2 0.24 0.24 3.00 0.47 0.47 0.0102.4 0.33 0.91 0.043 0.22 2.20 0.0472.7 0.27 5.60 0.062 0.22 1.00 0.0103.0 0.24 5.10 0.056 0.22 1.00 0.0103.3 0.22 1.60 0.30 0.22 0.22 0.0223.6 0.22 0.56 0.068 0.22 1.00 0.0473.9 0.24 0.39 0.68 0.22 0.22 0.0224.3 0.18 0.51 0.024 0.22 2.20 0.0474.7 0.16 1.30 0.039 0.22 1.00 0.0225.1 0.16 0.36 0.051 0.22 1.00 0.0475.6 0.13 1.10 0.033 0.22 1.00 0.0226.2 0.13 0.36 0.016 0.22 2.20 0.0476.8 0.24 1.60 0.33 0.10 0.10 0.0107.5 0.12 0.30 1.20 0.22 0.10 0.0108.2 0.12 0.11 0.024 0.22 2.20 0.0479.1 0.18 1.50 0.091 0.10 0.22 0.010
Table 1. Bessel lowpass filter
Freq. R1 R2 R3 C1 C2 C31.0 0.39 0.43 8.20 0.47 0.22 0.011.1 0.36 0.39 7.50 0.47 0.22 0.011.2 0.33 0.36 6.80 0.47 0.22 0.011.3 0.36 2.40 0.033 0.22 2.20 0.0471.5 0.33 4.70 0.012 0.22 4.70 0.0221.6 0.30 0.10 0.240 0.47 2.20 0.0471.8 0.30 3.30 5.10 0.22 0.022 0.0102.0 0.27 0.51 0.027 0.22 2.20 0.1002.2 0.24 2.70 0.43 0.22 0.10 0.0222.4 0.22 2.70 3.60 0.22 0.022 0.0102.7 0.27 0.43 1.30 0.22 0.10 0.0223.0 0.18 0.82 0.16 0.22 0.22 0.0473.3 0.15 0.056 1.00 0.47 1.00 0.0103.6 0.18 0.16 0.022 0.22 2.20 0.1003.9 0.15 1.50 2.20 0.22 0.022 0.0104.3 0.13 0.22 0.013 0.22 2.20 0.1004.7 0.20 0.12 1.20 0.22 0.22 0.0105.1 0.18 0.068 0.039 0.22 2.20 0.0475.6 0.20 1.10 0.036 0.10 0.47 0.0226.2 0.15 0.091 0.91 0.22 0.22 0.0106.8 0.16 0.91 0.03 0.10 0.47 0.0227.5 0.15 1.80 0.27 0.10 0.047 0.0108.2 0.10 0.12 1.00 0.22 0.10 0.0109.1 0.13 0.56 0.12 0.10 0.10 0.022
Figure 1. Sallen and Key lowpass filter
Lowpass filters designed afterSallen and Key usually take the formshown in Figure 1. In the classicSallen and Key circuit, resistors R1,R2, and R3 are set to the same valueto simplify the design equations.
When the three resistors are thesame value, the pole placement, andthus the filter characteristics, are setby the capacitor values (C1, C2, and
C3). This procedure, although greatfor the mathematician, can lead toproblems. The problem is that, in thereal world, the resistors, not the ca-pacitors, are available in a largeselection of values.
Taking advantage of the widerrange of resistor values is not alto-gether trivial; the mathematics canbe quite cumbersome and time con-suming.
This Design Idea includes tablesof resistor and capacitor values forthird-order Sallen and Key lowpassfilters. The resistor values are se-lected from the standard 5% valuepool, and the capacitor values areselected from the standard 10% valuepool. Frequencies are selected fromthe standard 5% value pool used forresistors. Frequencies are in Hertz,capacitance in Farads, and resistancein Ohms.
Figure 2 details the PSpiceTM simu-lation of a 1.6kHz Butterworth filterdesigned from these tables.
DESIGN IDEAS
dIs_k1.eps
–
+
C3 VOUTVIN
C1
R1 R3R2
C2
AV
0Ω
FREQUENCY
–80
–60
–40
–20
0
20
V OU
T
V IN
(dB)
100kHz
dIs_k2.eps
10Hz 100Hz 1.0kHz 10kHz
1.6kHz LPF
–
+
0.047µF
VOUT
0.47µF
VIN
270Ω 820Ω430Ω
0.47µF
LT1007
Figure 2. PSpice simulation of 1.6kHzButterworth filter
by Dale Eagar
Linear Technology Magazine • May 1995 33
DESIGN IDEAS
Table 3. Frequency multipliers0.1Ω 1Ω 10Ω 100Ω 1kΩ 10kΩ 100kΩ 1MΩ 10MΩ 100MΩ
1.0F 10 1.0 0.1 0.01 0.0010.1F 100 10 1.0 0.1 0.01 0.001
10,000µF 1k 100 10 1.0 0.1 0.01 0.0011,000µF 10k 1k 100 10 1.0 0.1 0.01 0.001
100µF 100k 10k 1k 100 10 1.0 0.1 0.01 0.00110µF 1M 100k 10k 1k 100 10 1.0 0.1 0.01 0.001
1µF 10M 1M 100k 10k 1k 100 10 1.0 0.1 0.010.1µF 100M 10M 1M 100k 10k 1k 100 10 1.0 0.1
0.01µF 1G 100M 10M 1M 100k 10k 1k 100 10 1.01,000pF 1G 100M 10M 1M 100k 10k 1k 100 10
100pF 1G 100M 10M 1M 100k 10k 1k 100
How to Designa Filter from the Tables:
Pick a cutoff frequency in Hertzas if it were a standard 5%resistor value in Ohms. (that is,if you want a cutoff frequency of1.7kHz, you must choosebetween 1.6k and 1.8k)
Select the component valuesfrom Table 1 or Table 2 as listedfor the frequency (think of the
first two color bands on aresistor).
Select a scale factor for theresistors and capacitors fromTable 3 by the following method:
1. Select a diagonal that representsthe frequency multiplier (think ofthe third color band on a 5%resistor).
2. Choose a particular diagonal boxby either choosing a capacitor
multiplier from the rows of thetable that give you a desiredcapacitor value or by choosing aresistor multiplier from thecolumns of the table that givesyou a desired resistance value.
Multiply the resistors andcapacitors by the scale factorsfor the rows and columns thatintersect at the chosen frequencymultiplier box. (for example, 0.68× 1µF = .68µF, 0.47 × 1kΩ =470Ω).
Figure 2. Network-analyzer plot of frequency response verses “set” current
Bandpass Filter, continued from page 31
–40
–30
–20
–10
0
REL
ATIV
E AM
PLIT
UD
E (d
B)
6MHz3MHz
dI1228_2.eps
300kHz
ISET =
250µA
200µA
150µA
100µA
50µA
25µA
34 Linear Technology Magazine • May 1995
DESIGN IDEAS
Simple Battery Charger Runs at 1MHz
Fast switching regulators have re-duced coil sizes to the point that theyare no longer the largest componentson the board. A case in point is theLT1377, which can operate at 1MHzwith inductances under 10µH.
The circuit shown in Figure 1 wasdesigned for a customer who wantedto charge a four-cell NiCd pack froma 5V logic supply. (This circuit willwork equally well with a 3.3V input.)Clearly the circuit needs an outputvoltage greater than 5V, which ishandled easily by the LT1377 boostregulator. The output current is lim-ited to approximately 50mA by a VBEcurrent-sensor (Q1/R1) controllingthe feedback pin (2) of the LT1377.This current is perfect for slow charg-ing or trickle charging AA NiCdbatteries.
Battery chargers are commonlysubject to a number of fault condi-tions, which must be addressed inthe design phase. First, what hap-
pens when the battery is discon-nected? In a boost regulator, theoutput voltage will increase withoutbound and blow up either the outputcapacitor or switch. Some voltage lim-iting is necessary, and in this designD2 serves the purpose. If the voltageon C3 rises to 11.25V, D2 takes overthe control loop at the feedback pin.
Another potential calamity is anoutput short circuit; a related faultresults from connecting a battery packcontaining one or more shorted cells,such that the terminal voltage is lessthan about 4V. Under either of thesecircumstances, unlimited currentflows from the 5V input supply,through D1 and Q1’s base-emitterjunction, frying at least Q1.
Q2 has been added to allow fullcurrent control even when the outputvoltage is less than the input voltage.In normal operation, where the out-put is boosted higher than 5V, Q2 isfully on. Its gate is held at 1.25V (pin
2 feedback voltage), and its source isgreater than 5V; hence it has no choicebut to be fully enhanced. Q2 becomesmore functional when the output volt-age drops to around 4V. First of all, at4V input the switching regulator stopsswitching because more than 50mAcurrent flows and the feedback pin ispulled up above 1.25V—Q1 makessure of that. But as Q1’s collectorcontinues to rise, Q2 is gradually cutoff, at least to the extent necessary tostarve the drain current back to about50mA. This action works right downto VOUT = 0. In a short-circuit, Q2dissipates about 200mW, not toomuch for a surface-mount MOSFET.
This circuit is useful for four to sixcells, and the output current can bemodified somewhat by changing senseresistor R1. A reasonable range isfrom very low currents (1mA or less)up to 100mA. The current will dimin-ish as Q1’s VBE drops about 0.3%/°Cwith temperature.
Figure 1. Battery charger schematic diagram
SHDN/SYNC FB
Q1 2N3906
Q2 Si9400DY
VIN = 5V C1
22µF 10V
C3 100µF
16V
D2 10V 400mW
C2 47nF
R2 2k
R3 1kΩ
R1 12Ω
4
1
2
6GND
7
LT1377
D1 MBR0520L
50mA (11V MAX)
L1 4.7µH
COILCRAFT DO-1608-472
VIN
VC
GND
VSW
85
C4 1nF
dI1377_1.eps
+
+
by Mitchell Lee
Linear Technology Magazine • May 1995 35
DESIGN IDEAS
Lithium-Ion Battery Chargerby Dimitry Goder
Lithium-ion (Li-Ion) rechargeablebatteries are quickly gaining popu-larity in a variety of applications. Themain reasons for the success of Li-Ion cells are higher power densityand higher terminal voltage comparedto other currently available batterytechnologies. The basic charging prin-ciple for a Li-Ion battery is quitesimple: apply a constant voltagesource with a built-in current limit. Adepleted battery is charged with aconstant current until it reaches aspecific voltage (usually 4.2V per cell),
then it floats at this voltage for anindefinite period. The main difficultywith charging Li-Ion cells is that thefloating-voltage accuracy needs to bearound 1%, with 5% current-limitaccuracy. These two targets are fairlydifficult to achieve. Figure 1 showsthe schematic of a full solution for aLi-Ion charger.
The battery charger is built aroundthe LTC1147, a high-efficiency step-down regulator controller. The IC’sconstant off-time architecture andcurrent-mode control ensure circuit
simplicity and fast transient response.At the beginning of the on-cycle, P-channel MOSFET Q1 turns on andthe current ramps up in the inductor.An internal current comparatorsenses the voltage, proportional tothe inductor current, across senseresistor R13. When this voltagereaches a preset value, the LTC1147turns Q1 off for a fixed period of timeset by C1. After the off-time, the cyclerepeats.
To provide an accurate currentlimit, U3A and Q2 are used to sense
Figure 1. Li-Ion battery charger schematic
+
R10 100Ω
R14 5.1k
100Ω
R13 0.1Ω
R3 51k 1%
D3 1N4148
D4 1N4148
C7 0.1µF
C6 0.1µF
R9 20k, 1%
R8 475k, 1%
1
+
–
–
U3A
2
3
LT1014LT1014
7
+
–
U3B
5
6
VIN
GND
8
1
7
5
2
3
4
CT
ITHU1
LTC1147
SENSE–
PDR
6VFB
SENSE+
R6 22k
8
4
11
+
U3C
10
9
LT1014
D1 MBRS130
Q1 Si9430
C1 270pF
R1 1k
R12 20k 1%
R4 22k
R5 100Ω
Q2 2N7002
Q3 2N7002
VREF
VIN
VREF
VIN
R11 20k 1%
C3 33µF 25V AVX TPS
C4 220µF 10V AVX TPS
C2 3300pF
+
0.1µF
D2 MURS320
VOUT 4.2V
1A MAX L1* 50µH
CTX50-4
R15 170k
0.25%
R16 249k
0.25%
dIbtcg1.eps
R2 24.9k
1%
VREF
C5 0.1µF
R7 20k, 1%
U2 LT1009-2.5
1000pF
VIN (6V TO 14V)
VREF
*L1 = CTX50-4 COILTRONICS (407) 241-7876
36 Linear Technology Magazine • May 1995
DESIGN IDEAS
the charging current separately fromthe LTC1147. U3A forces the voltageacross R11 to match the average dropacross the current sense resistor R13.This voltage sets Q2’s drain current,which flows unchanged to the source.As a result, the same voltage appearsacross R9, which is now referenced toground. Since C5 provides high-fre-quency filtering, U3A shifts theaverage value of the output current.N-channel MOSFET Q2 ensures cor-rect circuit operation even undershort-circuit conditions by allowingcurrent sensing at potentials close toground.
U3B monitors voltage across R9and acts to keep it constant by com-paring it to the reference voltage. DiodeD3 is connected in series with U3B’soutput, allowing the circuit to operateas a current limiter. The current-feedback circuit is not active if theoutput current limit has not beenreached.
U3C provides the voltage feedbackby comparing the output voltage tothe reference. The feedback resistorratio (R16/(R15 + R16)) sets the out-put at exactly 4.2V. U3C has a diode(D4) connected in series with its out-put. This diode ensures that the
voltage- and current-feedback circuitsdo not operate at the same time. Thereference voltage is supplied by theLT1009, with a guaranteed initial tol-erance of 0.2%. Together with the0.25% feedback resistors, the circuitprovides less than 1% output-voltageerror over temperature.
When the input voltage is notpresent, Q3 is automatically turnedoff and the feedback resistors do notdischarge the battery. Diode D2 isconnected in series with the output,preventing the battery from supply-ing reverse current to the charger.
Figure 1. Three-cell to 3.3V buck-boost converter
Obtaining 3.3V from three 1.2V(nominal) cells is not a straightfor-ward task. Since battery voltage canbe either below or above the output,common step-up or step-down con-verters are inadequate. Alternativesinclude using more complex switch-ing topologies, such as SEPIC, or aswitching boost regulator plus a se-ries, linear-pass element. Figure 1presents an elegant implementationof the latter approach.
The LT1303 is a Burst ModeTM
switching regulator that contains con-trol circuitry, an onboard power
transistor, and a gain block. Whenthe input voltage is below the output,U1 starts switching and boosts thevoltage across C2 and C3 to 3.3V. Thegain block turns on Q1, because thefeedback network R3–R5 biases thelow-battery comparator input (LBI)20mV below the reference. In thismode the circuit operates as a con-ventional boost converter, sensingoutput voltage at the FB pin.
When the input voltage increases,it eventually reaches a point wherethe regulator ceases switching andthe input voltage is passed unchanged
to capacitor C2. The output voltagerises until the LBI input reaches thereference voltage of 1.25V, at whichpoint Q1 starts operating as a series-pass element. In these conditions,the circuit functions as a linear regu-lator, with the attending efficiencyroll-off at higher input voltages.
For input voltages derived fromthree NiCd or NiMH cells, the circuitdescribed provides excellent efficiencyand the longest battery life. At 3.6V,where the battery spends most of itslife, efficiency exceeds 91%, leaving allalternative topologies far behind.
Three-Cell to 3.3VBuck-Boost Converter
VOUT 3.3V 300mA
D1 MBR0520L Q1
Si9433L1
20µH
+ + +
dI1303_1.eps
6 7
5 4
3 2
1 8
VIN
LBI
SHDN
GND
SW
FB
LBO
PGND
VIN 2.5V TO 8V
C1 33µF
C2 33µF
C3 330µF × 2
U1 LT1303
R2 100Ω
R1 100k
R3 200k 1%
R4 1.96k 1%
R5 121k 1%
C3: 330µF/6.3V AVX TPS C1, C2: 33µF/20V AVX TPS
by Dimitry Goder
Linear Technology Magazine • May 1995 37
DESIGN IDEAS
High Output-Voltage Buck Regulator
High-efficiency step-down conver-sion is easy to implement using theLTC1149 as a buck switching-regu-lator controller. The LTC1149 featuresconstant off-time, current-mode ar-chitecture and fully synchronousrectification. Current-mode operationwas selected for its well known ad-vantages of clean start-up, accuratecurrent limit, and excellent transientresponse.
Inductor current sensing is usu-ally implemented by placing a resistorin series with the coil, but the com-
mon-mode voltage at the LTC1149’ssense pins is limited to 10V. If ahigher output voltage is required, thecurrent-sense resistor can be placedin the circuit’s ground return to avoidcommon-mode problems. The circuitin Figure 1 can be used in applica-tions that do not lend themselves tothis approach.
Figure 1 shows a special level-shifting circuit (Q1 and U2) added toa typical LTC1149 application. TheLT1211, a high-speed precision am-plifier, forces the voltage across R5 to
equal the voltage across current-senseresistor R8. Q1’s drain current flowsto the source, creating a voltage acrossR6 proportional to the inductor cur-rent, which is now referenced toground. This voltage can be directlyapplied to the current-sense inputsof U1, the LTC1149. C12 and C4 areadded to improve high-frequencynoise immunity. Maximum input volt-age is now limited by the LT1211; itcan be increased if a zener diode isplaced in parallel with C12.
Figure 1. High output-voltage buck regulator schematic using LTC1149
+
+
–
+
dIbuck_1.eps
VIN 26V TO 35V
C13
C1
16
15
14
13
12
11
10
9
1
2
3
4
5
6
7
8
P-GATE Q2 RFD15P05
Q3 RFD14N05
Q1 VN2222LL
D1 MBRS140
D3 1N4148
L1 150µH
VIN
VCC
P-DRIVE
VCC
CT
ITH
SENSE–
CAP
SHDN
RGND
N-GATE
U1 LTC1149
U2A LT1211
PGND
SGND
VFB
SENSE+
C8 0.047µF
C7 1µF
C5 220pF
R4 510Ω
R13 12k 1%
R6 100Ω 1%
R5 100Ω 1%
R9 100Ω
24V 2A
R12 220k 1%
R10 100Ω
R8 0.05Ω
R9 100Ω
1
8 3
2
4C6 3300pF C11
100pF
C2 1000pF
C10 0.1µF
C12 0.1µF
C9 0.068µF
by Dimitry Goder
38 Linear Technology Magazine • May 1995
DESIGN IDEAS
New Device CameosLT1239: Backup BatteryManagement Circuit
The LT1239 is a micropower devicedesigned to be a complete manage-ment system for backup batteries inportable computers and other por-table devices. The device can provideboth charging and regulating func-tions for either lithium-ion or NiCdbackup batteries. The LT1239 pro-vides an uninterruptable powersource for the system’s backupmemory and power-management cir-cuitry. All circuitry is designed to runat micropower quiescent-currentlevels.
An adjustable linear regulatorsupplies a current-limited constant-voltage charge to the backup batteries.This regulator, normally powered fromthe system’s main battery pack, has aquiescent current of 15µA and ex-tremely low reverse-output current.The regulator acts like a switch, charg-ing the backup cells when the mainbattery pack is connected and dis-connecting the backup cells from thecharging circuitry when the mainbatteries are disconnected or dis-charged. The output voltage isadjustable from 3.75V to 20V. Be-cause of safety considerations relatedto the use of lithium-ion backup bat-teries, the regulator can operate withexternal current-limiting resistors inseries with its output. These current-limiting resistors can be placed in thefeedback loop of the regulator so thatthey will not affect output voltageregulation for normal operating con-ditions.
A second low-dropout linear regu-lator with a fixed output voltage of4.85V regulates the output of thebackup batteries. This second regu-lator also acts as a switch. When thesystem’s main battery is providingpower, the output of this regulator ispulled up to 5V and no power isdrained from the backup cells. If theoutput of the main power supply dropsbelow 4.85V, the second regulatorautomatically supplies power from
the backup cells to the backupmemory.
The LT1239 contains an error am-plifier to equalize the cell voltagesin two-cell lithium-ion systems. Inaddition, it includes a comparator forconnecting the main 5V system sup-ply to the backup circuitry.
A low-battery-detector circuit lim-its the discharge voltage of the backupbatteries to 5V. This circuit powersdown the 4.85V regulator and erroramplifier when the battery voltagedrops below 5V. In this shutdownmode the quiescent current drops to3µA.
Other features include independentshutdown pins for both regulators,and a current-monitor pin for eachregulator. The current monitor canbe used for gas gauging.
The LT1239 provides all the fea-tures needed to build a backup batterymanagement system. The LT1239 isavailable in a 16-pin, narrow-bodySO package.
LTC1334 Single 5VRS232/RS485 Transceiver
The LTC1334 is a single 5V supply,logic configurable, combinationRS232 and RS485 transceiver. Thisnew part is targeted at the software-configurable I/O port market.
Combining familiar functions inunfamiliar ways, the LTC1334 offersmultiple RS232 and RS485 ports inone package, along with the logic toallow various combinations of portconfiguration and an onboard chargepump to generate boosted voltagesfor RS232 levels. Inputs and outputsof both types are packaged togetherwith logic inputs that select whichwill be active for a given configura-tion. The LTC1334 features quadRS232 ports and dual RS485 ports,and is configurable as four RS232transceivers, two RS232 transceiv-ers, and one RS485 transceiver, ortwo RS485 transceivers. The configu-
ration of these transceivers is set byboth PC-trace routing and Select in-put logic states. For easy multiplexing,all drivers go into a high-impedancestate when deselected.
The LTC1334 features micropowershutdown mode, loopback mode forself-test, LTC’s usual high data rates(120kbaud for RS232 and 10Mbaudfor RS485) and 10kV ESD protectionat the driver outputs and receiverinputs.
The LTC1334 is ideal for comput-ers, multiplexers, networks, orperipherals that need to adapt to vari-ous I/O configuration requirementswithout any hardware adjustments.Remember the days of prying off theback cover and throwing DIP switcheswhen you set up your printer? Imag-ine the problems for a guy with a90-channel digital MUX —solved bythe LTC1334.
The LTC1334 is available in 24 pinSOIC packages.
LT1521: Micropower,Low-Dropout RegulatorHas 300mAOutput-Current Rating
The LT1521 is a 300mA, low-drop-out regulator with a quiescent currentof 10µA. Dropout voltage is 150mV at10mA, rising to 350mV at 300mA.Quiescent current is well controlledin dropout mode; it does not increasesignificantly as the device enters itsdropout region. The device can oper-ate with output capacitors as smallas 1µF.
The LT1521 has both reverse-battery and reverse-output protec-tion. Reverse output current isonly 6µA, making this device idealfor backup power applications.The LT1521 includes a shutdownfeature—quiescent current drops tojust 6µA in shutdown conditions.The LT1521 is available in fixed out-put voltages of 3.0V, 3.3V, and 5.0V.It is also available as an adjustabledevice with an output-voltage rangeof 3.75V to 20V. The LT1521 is avail-able in two surface-mount packages:the three-lead SOT-223 and the8-lead fused-leadframe SO-8.
NEW DEVICE CAMEOS
Linear Technology Magazine • May 1995 39
DESIGN IDEAS
The LT1529 has both reverse-bat-tery and reverse-output protection.Reverse output current is only 15µA,making this device ideal for backuppower applications. The LT1529 in-cludes a shutdown feature. Quiescentcurrent drops to just 15µA in shut-down mode. The LT1529 is availablein fixed output voltages of 3.3V and5.0V. It is also available as an adjust-able device with an output voltagerange of 3.75V to 20V. The LT1529is available in a 5-lead TO-220package.
The LTC1480:RS485 from 3.3V
RS485 transceivers enter the 3.3Vera with the introduction of the newLTC1480. Operating from a single3.3V supply, the LTC1480 is fullycompliant with all RS485 specifica-tions. The LTC1480 features amaximum quiescent current of 500µAin driver-disable mode and 600µA inthe driver-enable mode. It also pro-vides a shutdown feature, whichreduces the current consumption tobelow 1µA when the receiver anddriver are disabled at the same time.Its driver uses a proprietary CMOSoutput stage that connects twoSchottky diodes in series with theMOS output transistors. This allowsthe outputs to maintain high imped-ance when driven across the RS485common-mode range (12V to −7V), orwhen the power is off. The driver’soutputs also feature short-circuitprotection and thermal shutdown.
The LTC1480 features half-duplexoperation at up to 2.5Mbaud, withreceiver and driver propagation delayof 200ns (max) and 80ns (max) re-spectively. The LTC1480 is offered in8-pin DIP and SOIC packages, in bothcommercial and industrial tempera-ture grades.
LTC1487: Ultra-Low-Power5V RS485 Transceiverwith High Input Impedance
The LTC1487 is an improved sub-stitute for the LTC1483, designed witha high input impedance of 96kΩ (typi-cal) to allow up to 256 transceivers to
LT1528: 3-AmpPNP-Output Low-DropoutRegulator Optimized forMicroprocessor Applications
The LT1528 is a 3-amp low-drop-out regulator with a quiescent currentof 300µA. This device is optimized tohandle the large output current tran-sients associated with the currentgeneration of microprocessors. Thisdevice has the fastest transient re-sponse of all currently available PNPregulators and is very tolerant of varia-tions in capacitor ESR. Dropoutvoltage is 75mV at 10mA, rising to200mV at 1A and 500mV at 3A. Qui-escent current is well controlled indropout mode; it does not increasesignificantly as the device enters itsdropout region. The LT1528 can op-erate with output capacitors as smallas 3µF, although larger capacitorswill be needed to achieve theperformance required in most micro-processor applications. Although theLT1528 is available with a fixed out-put voltage of 3.3V, the external sensepin allows the user to adjust theoutput to voltages greater than 3.3Vwith a simple resistive divider. Thisallows the device to be adjusted easilyover a wide range of output voltages,including the 3.3V to 4.2V range re-quired by a variety of microprocessorsfrom Intel, IBM, and Cyrix.
The LT1528 has both reverse inputand reverse output protection. TheLT1528 includes a shutdown feature.Quiescent current drops to 150µA inshutdown mode. The LT1528 is avail-able in a 5-lead TO-220 package.
LT1529: 3-AmpPNP-Output, Low-DropoutRegulator Has MicropowerQuiescent Current
The LT1529 is a 3-amp low-drop-out regulator with a quiescent currentof only 30µA. Dropout voltage is100mV at 10mA and rises to 500mVat 3A. Quiescent current is well con-trolled in dropout mode; it does notincrease significantly as the deviceenters its dropout region. The devicecan operate with output capacitorsas small as 3µF.
share a single RS485 differential databus or line. With multiple transceiv-ers operating over the differential bus,the LTC1487 is fully compliant withall RS485 specifications. The LTC1487features remarkably low current, thelowest ever in the industry. It has amaximum quiescent current of 120µAin receiver-active mode and 200µA indriver-active mode under no-loadconditions. Significant power is savedby reducing quiescent current to be-low 1µA in the shutdown mode whenboth the receiver and the driver aredisabled. Like the other members ofLTC’s RS485 transceiver family, theLTC1487 uses a unique fabricationprocess and design that includesSchottky diodes in series with theMOS output transistors, allowing theoutput to maintain high impedancewhen driven across the full RS485common-mode range (12V to −7V) orwhen the power is off. The driveroutputs also feature short-circuitprotection and thermal shutdown.
The LTC1487 features half-duplexoperation at up to 250kbaud, withreceiver input propagation delay ofless than 250ns. Its driver slew rate isdeliberately limited to reduce EMIlevels in the transmitted signal. TheLTC1487 is available in 8-pin DIPand SOIC packages, in commercialtemperature grades.
NEW DEVICE CAMEOS
For further information on theabove or any of the other devicesmentioned in this issue of LinearTechnology, use the reader servicecard or call the LTC literature ser-vice number: 1-800-4-LINEAR. Askfor the pertinent data sheets andapplication notes.
Burst ModeTM is a trademark of LinearTechnology Corporation. , LTC and LT areregistered trademarks used only toidentify products of Linear Technology Corp.Other product names may be trademarksof the companies that manufacture theproducts.
Information furnished by TechnologyCorporation is believed to be accurate andreliable. However, Linear Technology makesno representation that the circuits describedherein will not infringe on existing patentrights.
40 Linear Technology Magazine • May 1995
DESIGN IDEAS
AppleTalk is a registered trademark of Apple Computer, Inc.
© 1995 Linear Technology Corporation/ Printed in U.S.A./27K
LINEAR TECHNOLOGY CORPORATION1630 McCarthy BoulevardMilpitas, CA 95035-7487
(408) 432-1900Literature Department 1-800-4-LINEAR
DESIGN TOOLSApplications on DiskNOISE DISKThis IBM-PC (or compatible) progam allows the user to calculate circuit noiseusing LTC op amps, determine the best LTC op amp for a low noise application,display the noise data for LTC op amps, calculate resistor noise, and calculatenoise using specs for any op amp. Available at no charge.
SPICE MACROMODEL DISKThis IBM-PC (or compatible) high density diskette contains the library of LTCop amp SPICE macromodels. The models can be used with any version ofSPICE for general analog circuit simulations. The diskette also containsworking circuit examples using the models, and a demonstration copy ofPSPICETM by MicroSim. Available at no charge.
Technical Books1990 Linear Databook, Volume I — This 1440 page collection of data sheetscovers op amps, voltage regulators, references, comparators, filters, PWMs,data conversion and interface products (bipolar and CMOS), in both commer-cial and military grades. The catalog features well over 300 devices. $10.00
1992 Linear Databook Supplement — This 1248 page supplement to the1990 Linear Databook is a collection of all products introduced since then.The catalog contains full data sheets for over 140 devices. The 1992 LinearDatabook Supplement is a companion to the 1990 Linear Databook, whichshould not be discarded. $10.00
1994 Linear Databook, Volume III — This 1826 page supplement to the 1990Linear Databook and 1992 Linear Databook Supplement is a collection ofall products introduced since 1992. A total of 152 product data sheets areincluded with updated selection guides. The 1994 Linear Databook Volume IIIis a supplement to the 1990 and 1992 Databooks, which should not bediscarded. $10.00
Linear Applications Handbook • Volume I — 928 pages full of applicationideas covered in depth by 40 Application Notes and 33 Design Notes.This catalog covers a broad range of “real world” linear circuitry. In addition todetailed, systems-oriented circuits, this handbook contains broad tutorialcontent together with liberal use of schematics and scope photography.A special feature in this edition includes a 22 page section on SPICEmacromodels. $20.00
1993 Linear Applications Handbook • Volume II — Continues the streamof “real world” linear circuitry initiated by the 1990 Handbook. Similar in scopeto the 1990 edition, the new book covers Application Notes 41 through 54 andDesign Notes 33 through 69. Additionally, references and articles from non-LTC publications that we have found useful are also included. $20.00
Interface Product Handbook — This 424 page handbook features LTC’scomplete line of line driver and receiver products for RS232, RS485,RS423, RS422, V.35 and AppleTalk applications. Linear’s particularexpertise in this area involves low power consumption, high numbers ofdrivers and receivers in one package, mixed RS232 and RS485 devices, 10kVESD protection of RS232 devices and surface mount packages.
Available at no charge.
SwitcherCAD Handbook — This 144 page manual, including disk, guidesthe user through SwitcherCAD—a powerful PC software tool which aids in thedesign and optimization of switching regulators. The program can cut days offthe design cycle by selecting topologies, calculating operating points andspecifying component values and manufacturer's part numbers. $20.00
1995 Power Solutions Brochure, First Edition — This 64 page collectionof circuits contains real-life solutions for common power supply designproblems. There are over 45 circuits, including descriptions, graphs andperformance specifications. Topics covered include PCMCIA power manage-ment, microprocessor power supplies, portable equipment power supplies,micropower DC/DC, step-up and step-down switching regulators, off-lineswitching regulators, linear regulators and switched capacitor conversion.
Available at no charge.
FRANCELinear Technology S.A.R.L.Immeuble “Le Quartz”58 Chemin de la Justice92290 Chatenay MalabryFrancePhone: 33-1-41079555FAX: 33-1-46314613
GERMANYLinear Techonolgy GmbHUntere Hauptstr. 9D-85386 EchingGermanyPhone: 49-89-3197410FAX: 49-89-3194821
JAPANLinear Technology KK5F NAO Bldg.1-14 Shin-Ogawa-Cho, Shinjuku-KuTokyo, 162 JapanPhone: 81-3-3267-7891FAX: 81-3-3267-8010
KOREALinear Technology Korea BranchNamsong Building, #505Itaewon-Dong 260-199Yongsan-Ku, SeoulKoreaPhone: 82-2-792-1617FAX: 82-2-792-1619
SINGAPORELinear Technology Pte. Ltd.507 Yishun Industrial Park ASingapore 2776Phone: 65-753-2692FAX: 65-541-4113
TAIWANLinear Technology CorporationRm. 602, No. 46, Sec. 2Chung Shan N. Rd.Taipei, Taiwan, R.O.C.Phone: 886-2-521-7575FAX: 886-2-562-2285
UNITED KINGDOMLinear Technology (UK) Ltd.The Coliseum, Riverside WayCamberley, Surrey GU15 3YLUnited KingdomPhone: 44-276-677676FAX: 44-276-64851
Linear Technology Corporation1630 McCarthy BoulevardMilpitas, CA 95035-7487Phone: (408) 432-1900FAX: (408) 434-0507
InternationalSales Offices
World Headquarters
U.S. AreaSales OfficesCENTRAL REGIONLinear Technology CorporationChesapeake Square229 Mitchell Court, Suite A-25Addison, IL 60101Phone: (708) 620-6910FAX: (708) 620-6977
NORTHEAST REGIONLinear Technology Corporation3220 Tillman Drive, Suite 120Bensalem, PA 19020Phone: (215) 638-9667FAX: (215) 638-9764
Linear Technology Corporation266 Lowell St., Suite B-8Wilmington, MA 01887Phone: (508) 658-3881FAX: (508) 658-2701
NORTHWEST REGIONLinear Technology Corporation782 Sycamore Dr.Milpitas, CA 95035Phone: (408) 428-2050FAX: (408) 432-6331
SOUTHEAST REGIONLinear Technology Corporation17000 Dallas ParkwaySuite 219Dallas, TX 75248Phone: (214) 733-3071FAX: (214) 380-5138
SOUTHWEST REGIONLinear Technology Corporation22141 Ventura Blvd.Suite 206Woodland Hills, CA 91364Phone: (818) 703-0835FAX: (818) 703-0517
top related