analog dialogue Aforum for the exchange of circuits and systems for test, measurement, and control POM Te Le ear AA er ONCE) COMO ce CU ue A eS eo eee ca a Ce ea) eeenceeneet) Volume 16, Number 2, 1982 ANALOG DEVICESEditor’s Notes el SOFTWARE & COMPUTERS Analog Devices has always been noted for sofware as well as hacd= ware. But the software that un recently has been our forte is che human-to-human kind exempli- fied by our comprehensive data sheets, by this Journal, by our ‘Teebnieal Handbooks, and by the now 1500+ page Databook (see page 26). This steength in publi cations is now being augmented by a mastery of software in the more usual sense. Human-to-machine software first appeared in these pages in rudimentary fashion in Analog Dialogue 7-1 (1973 - SERDEX), and—more authoritatively—sin 13-1 (1979), where MACSYM 2 and MACBASIC made their debut. fc may interest you that, in this issue, articles discuss sophisticated compucer software genersted by chree different entities within ADI: At page 11, John Mills, of the Systems Components Division, Siscusses the new software drivers that make the uMAC-4000"s versatile and intelligent sensor and digital VO (sce back cover) available ro the explosively growing family of users of Apple IBM, and HP personal compacers, as well as users of DEC PDP-11 and LSL-I hardware. © At page 12, Bill Schweber, of the Measurement and Control Products Division, discusses the vietues of rmvlticasking, an inher cnt feature of systems that use MACBASIC, The world is finally beginning to “discover” mnultcasking in small computers, 2 uuable feature of MACSYM systems since before 1978 @ At page 19, Mike Slocombe and Af Finger, of Component Test Systems, discuss the new generation of compatible, easier-to-use sofeware developed for che LTS-2010 and LTS-2000 test syscems, Impressive as all this may seem, we're secing just the tip of a mas sive software iceberg (oftware iceberg?) In our wanderings through the facilities that house the various groups at ADI, soft ware-related activities are everywhere—and they relate to more than just developing yer better, newer, and augmented versions of the user-oriented electeonic systems activities mentioned above. Computers and home-grown software arc used every day in the de- sign, manufacture, crimming, and test of vierually all of our prod uicts, And in management, markecing, and administeacion. And, in particular, in our new two million-dollar computerized customer order system (COS}—that is already making life easier for many fof one readers (especially those purchasers of a-few-parts-at-a~ time who must welcome the razing of the “*minimum-orde barrier) ly, we acknowledge a digital debe of gratitude (o our word processor, the sine qua non of those (formerly known as “ink- stained wretches") whose principal pursuits include putting words cogether to communicate with others in “writing.” Oue word pro~ cessor and the laser-typesetter it communicates with via an RS; 232C link weren’ton the scene a mere year ago, but itisimpossible to picture any future scene without them. If you've noticed that this issue has 28 pages—the most ever—the answer to the unaskcd question is ifn, ie wouldn't have. Dan Sheingold > seem n164_2076 ® Anatan Neuirse Ine 1089. THE AUTHORS Donald W. Brockman tpage 14) 's Technical Communications Man ager at ADI's Computer Labs Divi sion {formerly Computer Labs Incorporated), which he joined 0 1968 from Bell Telephone Labora: tories, His formal education in cludes a B.A. degrec, and he has worked in engineering, writing, re Cruiting, and sales functions before his present job in Marketing, Terry L. Brown (page 14) joined ADI's Computer Cabs. Division after receiving his BSEE from North Carolina State University and has already made significant contributions (© the Division's high-speed data-acquisition prod- uct line. Tery was responsible for the design of the HTC-0300A and HIC-0500 (see page 22) Hs and the HDD-1206 DAC. He is currently spearheading CLD's monolithic-cireuit develapment program. Alon Finger ipage 19) isthe Senior 5 Software Engineer in the Compo: nent Tese Systems organization. Previously he was Sofeware Project Leader and Senior Applications Engineer with MACSYM, where he developed the software for MACSYM 20 and a variety of hardwarelsofeware packages for MACSYM 2. Before joining ADI, hhe was a Research Physicist at Eastman Kodak Research Labora- cories. He has a BSEE from Clarkson College and occupies his spare hours with photography, bicycling, and. personal computing. Pat Hickey (page 3), designer of the AD7S28 dual DAC, is Design Engineer at Analog Devices B.V., in Limerick, He has a B.S. in Ble’ tronic Systems from the National Institute of Higher Education, Limerick. Prior to becoming a De- sign Engineer, he served as a Test Engincerat A.D.B.V. (nore authors on page 26) analog dialogue 2 Technology Way, P.O. Box 280, Norwood, Mass. 02062 Published by Analog Devices, Inc., and available at no charge to engi nets and scientists who use or think about LC. of discrete atalog, con- version, data handling and display circuits. Correspondence is welcome and should be addressed to Editor, Analog Dialogue, P.O. Box 280, Norwood, Massachusetts, U.S.A. 02062. Analog Devices, Inc, has representatives and sales’ offices throughout the world. For informa tion regarding our products and their applications, you are invited to use the enclosed Business Reply card, write to the above address, oF phone 617-329-4700, TWX 710.394.6577 or 710.336.0562, Telex 524.491, or cable ANALOG NORWOODMASS. ‘Analen Rialanne 18.2 19?AD7528: A DUAL MONOLITHIC DAC FOR MANY REASONS World’s First Twin Multiplying 8-Bit DACs in a 20-Pin (0.3”) DIP Separate On-Chip Latches Provide Easy Microprocessor Interface AD7528" isa buffered dual 8-bit CMOS muliplying Wa con- verter on a single monolithic chip, housed in a small 20-pin 0.% DIP package. ts twin DACs, matched to within 1%, gain access to the common data bus via 2 pair of #-bit latches gated by TTL: ‘or CMOS-compatible logic. As Figure 1a shows, in addition to the usual chip-seleet (CS) and write (WR) logic inputs, there is also a steering input (DATAVDAC B) b. Simplified functional circuit of DACA. Figure 1. Functional diagrams of the AD7528. Besides the obvious applications where wo oF more 8-bit DACS arenceded and space isata premium, the AD7528 makes available a whole subelass of applications calling for paired DACs having essentially identical performance. In addition co conventional fixed-reference applications, they may be used 3s two: oF four- quadrant multiplying DACs wih variable de or ac reference sige nals. You can think of many ofthese latter applicatians as uses for a matched pair of digital pots (cither separate or ganged) to ma- nipulace the properties of analog signalsor ciceuits. The range of such applicatio is limited only by imagination; i¢ includes, for example, filters, oscillators, resolvers, function sgeneracars, X-Y ploccers and oscilloscope displays, audio gain con- wuators, interpolators, high-low attenuation-function-gain circuits, and all kinds of test and mea- surement circuits involving multiple-paramecee setting, *Usethe reply card fortechnical daca 1Useal references from Analog Devices include the fee Application Guide to CMOS Mulniplying DIA Coomertere and the book, Andlog Digital Conversion Nove, sealable 85.98, ‘Analoo Dialonue 16-2 1982 by John Wynne and Pat Hickey ‘THE AD7528 IN DETAIL “The AD7528 contains two 8-bit multiplying dia converters with {arches and conteol logic. Since both of the DACS are fabricated by a monolithie CMOS process at the same time on the same chip, precise matching and tracking between DAC A and DAC Bare in herent, Each DAC (Figure 1b) consists ofa highly stable thin-filn R-2R ladder, eight N-channel currentsteering switches, and a matched feedback resistor, As the simplified circuit show: verted ladder structure is used, ie, binary currents are switched between che DAC output and AGND, thus maintaining fixed cure rents ineach ladder leg, independent of switch state.t (continued on the following page) Figure 2. Applying the AD7528 as a unipolar binary DAC pair. IN THIS ISSUE Volume 16,Number2, 1982, 28 Pages Ealnor's Notes, Authors... - 7 AD7828: A Dual Monolithic DAG for All Reaions Applying Uligh-Performance Monolchic FET-Input Op Amps Use Your Personal Computer for Measurement & Control Moditasking: The Key to Effective Measurement and Control proved Vector-Scan Displays Are Coming «> « Deglteing a 16-Bit Monolithic BIA Converter. | Versatile wP-Based Meter for -K-T-E-R-S Thermocouples. Logarithmic Mubiplying CMOS DIA Converter, 0.3754BiDit New Software Package for LTS-2010 Linear IC Test Systeme High-Performance Hybrids for Data Acquisitc Super-Performance Hybrid Op-Amp Families. New-Produet Brits: Fas Teac ld wih 13s max Acuston Time or 20-V Sep. 22 12-Big, $-MHz AID Converter for MIL Applications 3 12-BiePlasSign CMOS Monolithic Integrating AD Convereer’ | 22. Low-Cost Precision 2.5-V IC Reference (ADI403 and ADIA03A) 23. '8-Bie Synchro-to-Digital Converters 10.Bie Monolithic ICTDAC (High-Performance AD DACIO0) | | 33 ACC & DODC Power: Naw Calg Ins 18New Sopplics: 24 Uleafast Op Amps: AD3S54 and HOS-0S0C.. 4 Test Subpicoamp Bias Currents Awomatcaly wit 25 Accurate Mut Chanel Thermocouple Meassremensfor “MACSYM, ae > 4 25 Analog and Digial 2s ‘Worth Reading: The Analog Devices 1982 Databook, ADL “Authors in the Trade Press, Featuted in Macs}micer 2% Potpout 7 Advenisments Industral Sense VO fat Any Compute 13 4In normal applications, each DAC is buffered by an external oper: ational amplifier, a shown in Figure 2. The DAC output terminal is maintained by the op amp at ground potential, and the internal Rynis connected as the feedback resistor. Nominally, Vour = ~D Vari» where Da function of the input code, isa feactional binary wale from 0 10 (12%) and Vey is an ac oF de voliage (absolute maximum rating is #25V), As the Figure indicates, Schottky diode protection is unnecessary. For four-quadrant multiplication, an additional amplifier cakes the difference between 2D Vay and the analog inpot (Vrs), €o provide a function chac swings linearly from + Ver to Veer as D varies from all-ts to all-0s (offset bi- nary = 2s complement with complemented MSB). Relative aceuraey is guaranteed over the device’s full temperature range: * 1 LSB maximum error for J, A, and S grades, and 1/2 LSB max for K, L, B,C, T, and U grades. All grades are guaranteed monotonic over the temperature range (0°C t0 + 70°C for J, K,and 1, = 25°C to +85°C for A, B, and C, ~S5°C to + 125°C for 5.7, and U}. Maximum initial gain error is guaranteed a +4 LSBs for JeAv& S grades, *2 LSBs for K, B, and TT grades, and = 1 LSB for L, C, and U grades. The trim resistors shown in Figure 2 are used to sec the gain to better than the specified value; they ean be replaced by direct wiring if the performance is satisfactory as specified. The AD7528 will operate witha single + SV to + 15V power sup ply, and it consumes a maximum of SmW at Von = +5V. Maximum feedthrough for a 20-vole peaksto-peak sine wave is —70dB at 100kHz, either channel, and ~77dB channel-to-chan: nel, Total harmonic distortion is ~85UB, and digital crosstalk is 30 mVes ‘The AD7528 is available in Cerdip (AQ. BQ, CQ suffixes) and hermetic ceramic (SD, TD, UD); and it will shortly be available in plastic JN, KN, LN). Ie is also available in leadless chip carriers (CE” suffix) and in versions processed to the requirements of MIL- ‘STD-88348, Prices startat $5.95 in 100s (AD7528)N). BRIEE PERFORMANCE CHARACTERISTICS Whore Varvw = 4 10V, OUT = OUI © OVE Vyyy => 15Veand't = 25°C, tule eherwae esi} KOT LOU Reselation is) Differconisl 5B, eas) Gain Error’ SB. em), Gain Tempco EPC, may) Input resatance Vin Ave Vgry Angutvesistanctrismareh (%, 93%) De Supply rection (8 gain/8 Ving) (Ce perio max) oor owt AC Feedthrough, cither channel, 20 pp sinewave TOOK (nas) 7-70 =70 Channebto-Chasnel alton, 20V pp reese i 100K (8) 7-7-7 Harmonic Distortion, éVemeG@ tkHy(dB) 85 8S AS APPLYING THE AD7528 ‘The AD7528 has ioad eycle similar to the write cycle of a random access memory and is compatible with most 8-bit microprocessors, including the 6800, 8080, $085, and Z80. Data is transferred into cither of the two DAC data latches via a common 8-bit TTL CMOS-comparible input port. Contro! input, DAC A/DAC B, de teemines which DAC is co be loaded, while separace reference in- puts allow each DAC co independently perform 2- or 4-quadrant analog-digital multiplication. AA simple example will show how easy microprocessor inteefacing is, Figure 3 is a scheme for interfacing the AD7528 to a 6800-type microprocessor, Oucputs A and A + 1 of the address-decoue logic 4 ED) Figure 3. Applying the AD7528 with 2 6800-type micropro: drive GS and the DAC selector. The logic is arranged so that if neither A nor A + 1 is addressed, CS remains high; if A is ad dressed, CS is brought low and DAC A is updated when WR goes lows if A +_1 is addressed, CS is brought low and DAC B is up- dated when WR goes low. For applications using 8085-type micro- processors, WR is generated directly and the gate is unnecessary Theanalog circuiery has been omitted for clarity Single Supply. In che AD7528, the termination resistors of the R- 2 ladder arc connected to AGND within the device. This arrange ‘ment is convenient for single-supply operation, because AGND ray be biased at any voltage between DGND and Vp. In Figure 4, the two DAC ouputs are offset by a precise +5V by the use of an ADSSA) reference to produce an offset AGND. The «wo DAC reference inputs are both grounded. Thus the reference vol- tage applied to the DAC is ~ 5V, the DAC ourput swing is 5 volts, and the DAC ouspu range (cither DAC) is from +5V to + 10V, for codes from 0000000010 1111 1111 Figure 4. One form of single-supply operation of the AD7528. APPLICATION EXAMPLE- STATE-VARIABLEFILTER ‘This application is an excellent example of the advantages of the unique properties of the AD7528 in performing useful functions that rely on DAC matching and benefit from the AD7528°s high packing density. Such circuits can be especially compact when em- ploying dval or quad op amps, such as the AD644 or the TLO74. ‘The circuit described here isa state-vaciable 2-pote filter with prog rammable center frequency, selectivity (Q),and gain. The second-order state-variable filter (or universal filter, a8 it is often called) is a convenient building black of higher-ordce malt crs (Bessel, Butterworth, Chebyshev, ete.) Simultaneously 1g low-pass, high-pass, and bandpass responses, its out- ‘Analon Dialonue 16-2 1982puts can be further combined to provide band-reject and all-pass (biquad) properties. If available in programmable form, it makes possible a huge variety of filter characteristics, readily provided by ‘a microprocessor, using either computation or alookup table. Its parameters are readily adjustable by manipulation of resiscance ratios. Figure S shows a typical conventional stace-variable filter iccuit, with the expressions for center frequeney (f,), Q,and gain forthe bandpass output Figures, Block diagram of state-variable filter. In form, ic ean be recognized as an analog computing circuit chat solves the classical second-order differential equation, modelling a mass, spring, and dashpot, or inductance, capacitance, and resis: tance. Ifthe low-pass output is interpreted as position, the band= pass outpur—its derivative—can be viewed as velocity, and the high-pass ouxput, which depends on the second derivative, can be considered s acceleration. ‘The gains of the 1wo cascaded integrators (1/R yC. termine the resonant frequency: if Ry = Ryand Ry = Roy then che natural frequency is 1/(2 RC). And the gain in the loop around integrator A3 determines the damping, which isinverse wich Q If che filter isto be digitally programmed, the resistor ratios must be replaced by fixed resistors and digitally concrolled gains. How this may done is shown in Figure 6. DACs 2A and 2B, which serve as digitally controlled attenuators with fixed admittance, replace and VR4C) de- Figure 6. Programmable state-vasiable filter employing AD7528. Analog Dialogue 16-2 1982 R3 and R4, to concral frequency. DAC 1A replaces RI, and con- trols gain; and DAC IB replaces R2 and RS, cocontrol 1/Q. For the component values shown, the programmable ange of f, is 0 to 1SkHz, and the programmable range of Qis 0.3 (0 4.5:Q and f, are mutually independent. Figure 7 shows parametrie plots ‘of amplitude vs. frequency as functions of Q (a) and {,(b). APPLICATIONS IDEAS ‘That the AD752R is “only” an 8-bit double-DACis a problem for some conventional DAC applications. However, there are many more applications, where the primary fine structure is already analog, that simply consider the 256 codes available ro be a limita~ tion only on the number of available gains, but not on signal resol- ution. Many of the potential applications of the AD7528 rely on its excellent analog properties: low feedthrough and distortion, ‘wide bandwideh and dynamic range, and low gain tempco. The small assortment here is merely intended to stimulate thought.* I you can build a programmable filter, you can use itina feedback Joop to make a programmable oscillator. If you can build prog~ rammable oscillators and fileers, you can assemble programmable spectrum analyzers. With a pair of matched multiplying DACs, you can produce such vector composition/resolucion pairs as: V sin at sin 6 and V sin wteos 6, where the functions of 6 are established digitally. The inherent moltiplication can be used to advantage in program: ‘ming analog polynomial fonctions combining x, x”, ec.,forexam- pile, in constructing simple approximations to the sine, Since al as- peets—voltage or current amplitude, phase, frequency, and waveshape—are programmable, the technique can be ased t0 build a powerful ac or de function generator with multiple ousputs for estingatloweost *Some worked-out examples con be found in an AD7S28 Application Note, available upon request. DIRS Figure 7. Amplitude vs. frequency as functions of Q and fo. 5APPLYING HIGH-PERFORMANCE MONOLITHIC FET-INPUT OP AMPS Super Singles and Duals Use New Trimmed Bipolar-FET Technology The Ideal Op Amp Still Doesn’t Exist, But These Come Close Enough byDon Travers From the earliest days of the monolithic operational amplifier, neve amplifier chips have came to market with continually improv- ing performance. As refinements to existing technologies and man- uifscuuring processing techniques evolve, total operational: amplifier performance approaches thae of an ideal amplifier ever more closely, for mos practical purposes. Refinements to bipolar operational amplifier pracessing involving ion-implanted FETs (in some cases called “BieFET™ techniques), in particular, have nar owed the gap between peactical and ideal amplifiers in de perfor: mance without sacrificing bandwidth. For example, itis now possible o trim single and dual FET-input monolithic op amps for both aninimuny dift and minimum offset—automatially, a¢ the wafer stage. The improvement in overall performance is a boon to users, simplifying many designs and providing high aceueney at reduced cost. Here are a few examples of the kind of performance available in off-the-shelf devices: the deift-trimmed ADS47LH has maximum offsce and drife of 0.25mV and IMV/C, open-loop gain of 250,000V/V, and bias current of 25pA max, at operating tempers ine the wider-bandwidth ADS44UH has 2MHz bandwidth and 13Vips slewing rate, with maximum offset and drift of 0.5mV and SuVPC, Table I outlines che salient properties of the op amps inthisseries Applications requiring excellent de performance at low cost coll for the ADS42* and ADS47* monolithic FET operational oyplfier families ond the equivalent duals—AD642" and AD647°. For example, precision instrument front ends, requiring, accurate amplification of millivolt-level signals from megohm source impedances, will benef fram the devices’ combination of low offsee voltage and drift, low bias current, and low 1/f noise, ‘The dual versions of these amplifiers, since they are closely matched, are ideal for true instrumentation amplifiers and log- ratio amplifiers The ADS44* family of single op amps and the AD644* family of equivalent dual ap amps arc recommended for any application re ‘quiring excellent dynamie—in addition to de—performance, The wide bandwidth, low offset voltage, and high open-loop gain en: sure superior accuracy in high-impedance buffer and sample-and hold applications. The AD644, with matched amplifiers, can be tused for wide-bandwidth instrumentation amplifiers, low-de-drife active filters, and as output amplifiers for four-quadeant muliply- ing Wa converters, such as the AD7541A 12-bit CMOS DAC. tn these pages, we will show you some concrete examples of popu: lar op-amp applications thac benefic from the improvements in precision monolithic FET-input op amp technology and will offer *Uscthe reply cad for technical data {his aril ts condensed from material collected for an Application Note on| sonolthic FET-inpat op amps of the above sri, For 3 copy’ of the complete ‘AlNtwhen avaiable usethe teply ca {For information an the interaction between op amp and DAC, se “Analog Sighal Handing for High Speed and Accueacy,” Analog Dialogue 11-2, 92605 NSD. 6 suggestions for obvaining best resalts, The applications include op- erations with d/a converters and sensors, practical log amps, and high-inpat-impedance instrumentation amplifier ciceuits.t D/A CONVERTER APPLICATIONS CMOS muliplying DACs, from the eaeiest 10-bit AD7520 to the oder 16-bit AD7S46, rely on op amps for interfacing and but fering. For che best linearity and most stable dynamics, the choice ofopampsis by no means trivial CMOS DACs will provide improved overall performance when used with amplifiers in this series co perform both 2-quadranc and 4-quadrant operations. For example, the ouipuc impedance of a CMOS DAC varies with the bit composition of the digital word, affecting the noise gain (1 + Ry/R,, where R, is the resistance be- tween OUT! and common) of the amplifier circuic. The effect is to cause a nonlinearity, che magnitude of which isis dependent on the offset voltage of the amplifier and its drift over the operating temperature range. TABLE 1, Monolithic Bipolar Fet Op Amps (Typical specications at 25°C sad ried spptywnless noted bens) Asam mk 182 ® ApwoM wh ees soe mk tego For example, in the one-bit code ceansition from 0011 1111 1114 to 0010 0000 0000, the output offset voleage changes from 2Vax to (4/3)Vgy a change of (2/3)Vyq If Vig is not much smaller chan the voltage equivalent ofthe least-significane bit, ands of the right (ovrong) polarity, the transition will be nonmonotonie. This factor is especially imporcant in multiplying applications, where the re: ference input can be considerably less than full seale, and the bit voltage accordingly quite small. ‘These monolithic FETs, with theie trimmed offset and low drift, ‘Analog Dialogue 16-2 1982will minimize this effect. In addition, DAC output circuits using these monolithic FET op amps do not require the Schottky diodes that are recommended for protection of many older CMOS DAC «ypes from amplifier turn-on transients. Figure 1a shows the ADS47 op amp and AD7541 CMOS dla con- verter configured for unipolar binary (2-quadrant-mvultiplying) operation. With a dereference voltage or current, of either positive ‘or negative polarity, applied at pin 17 of the DAC, the circuit oper- ates asa unipolar fixed-reference converter. With an ac reference voltage oF current, the circuit provides two-quadrant multiplica- tion (digically controlled attenuation). we “8 vend I ee sir20pels lion gas T | sae your ment ar nacseyt K i re : soy Ty ¥ 2, ADS47 used as DAC output ampli b. Response to + 20V p-p reference square wave. Figure 1. Voltage-output multiplying DAC. The oscilloscope photos of Figure Ib show the output of che circuic of Figure 1a. The uppee tcace represents che ae reference input, a 2OV peak-to-peak square wave at 33kHz, and the lower trace shows the output volrage while che digical input to the DAC isall- Us Gain of 1-2"), representing a settling time 10 0.01% of 15s, with well-behaved dynamies. The 47pF capacitor across the fced- back resistor compensates for the DAC output eapacitanee, and the 150pF load capacitor serves co reduce signal feedthrough spikes atthe output ‘The diagram of Figure 2a illustrates the 10-bit digital-to-analog, ‘converter, AD7533, connected for bipolar operation. Since the di- gical input ean aecepe bipolar numbers and Vey can accepta bipo- lar analog input, the cireuit performs 4-quadrant multiplication, a. AD644 and AD7533in 4-quadrant application. Analan Nialamie 16-9 1982 ‘This is accomplished with an AD644 dual FEI maintain optimal performance at low cost wi The scope plots exhibit the response to. step input at Vers 2b is che large-signal response and 2c shows the small-signal re= sponse. Cl,as.a phase compensator (15 pF), is required for stability when using high speed amplifiers; it serves to cancel the pole formed by the DAC I feedback resistance and the oueput eapacitanee of the DAC, ¢. Small-signal response to reference square wave. Figure 2. Four-quadrant multiplying DAC configur Figure 3 shows the AD647 used with the AD7546 16-bit DAC, a tovorstage device which uses 16 segmented gain levels to represent 3 12+bit R-2R ladder and switches for the rest, and external op amps for impedance buffering. Since | USB = 8V/65536 = 122,V. in this application, amplifier perfor: mance i critical. the overall performance of the AD7546, Figure3. AD647 used with AD7546 16-bit DAC. ‘The paired amplifiers of Al, the AD647, are used as he high-preci- sion dual buffer. Here, the offset voltage match, the low offset vol- tage and high open-loop gain of the AD647 ensure monotonicity and high linearity over the entire operating temperature range. A2, also an AD6A7, used asa track-and-hold, serves a dual funetion: 1A buffers the switched capacitor from the output of the ladder, and amplifier B buffers the hold capacitor from the output. ‘The performance of the amplifices of A2 is erucial co the accuracy of the system. Their errors are added to the errors due strictly «0 DAC imperfections. For this reason, great care is needed in 7amplifier selection, The matching characteristics, low bias current, and low temperature coctficiens of the AD647 make ic ideal or this application Here's another DAC application where a high-performance op amp is desirable: Ifa CMOS DAC is connected asthe feedback ele- ment in an op-amp circuic (the &-bit 4P-compatible AD7524, in Figure 4), eather than as éhe more-usual input element, the cir- cuit—performing the inverse operation—will divide the input vo} tage by the fractional digital number, eather than multiplying by ic chus providing a repertoire of gains greater than 1. Since feed: back around the amplificr is reduced, for smaller digital numbers, the gains for both signal and noiselerrors are inereased. Also, since the incremental step changes of gain become inereasingly large for smaller values of D, in hyperbolic fashion, any errors in the gain steps, due ta either DAC errors or limited open-loop gain of the ‘opamp, are magnified correspondingly. Figure 4. Analog divider with digital denominator [As a simple analog “fix” co provide a limited gain in the extreme case of all zeros—in order ¢0 prevent the output from “taking off”—a large value of resistance (compared ro Ryy/2") may be used to shunt the feedback path, ¢.,22MO. There are any number of digital fixes that may be used, in hardware oF software, ranging {rom simply tying the LSB High to performing a logical tes for all 0s—or for a number less than 1 LSB—and setting the nvinimam value ofthe digital word at 1 LSB. INSTRUMENTATION AMPLIFIERS When op amps ate used as building blocks of high-performance instrumentation amplifiers, low drife and bias current, and high input impedance, are essential characteristics. For most in- stromentation-amplifier applications, the definitive performance and low cost of the monolithic ADS24 make it preferable toa user ‘ia Fane cy fengery AN yoy FigureS, Wide-bandwidth instrumentation amplifier, assembled version. However, applications do exist for which the ‘uniquely low bias euerent of a FET-inpatamplifcris essential Figure 5 shows an easily constructed (wo-amplifier FET-inpus in strunentation amplifier employing the _wider-bandwidth, amplificrs in the AD644, DC offset and common-mode error are trinmable, and the CMR-trimming capacitor on the negative input of AIA permics highly accurate signal conditioning a¢ the higher frequencies, with CMR of 804B over che range de (0 VOkHz, and bandwidth of 200kHz (~ 348) at 20V peak-to-peak output. Offset vokage diftis 1OnV?C. The circuit of Figure 5 can be configured for a practical range of gains from less than 5 to well over 100 with typical nonlincaricy (of 0.01% at gain of 10. Ifadjustment of gain by a single resistance is desirable, a resistor, Re may be connected between the negative summing points, with some sacrifice in high-frequency common mode rejection, The gain isthen equalto(} + RyRy + 2RYRh LOG-RATIO AMPLIFIERS Log amplifiers and log-ratio amplifiers are useful in a wide variety of analog computational applications, ranging from the simple linearization of exponential cransducer outputs (0 the use of logarithms in computations involving mult-cerm products or arbi trary exponents. Log amps also facilitate the compression of wide- dynamic-range analog signals into a range that can be handled withouc using esoteric circuit techniques. “The picoamp-level input current and low offset voltage of the AD647 make it suitable for wide-dynamie-range log amplifiers. The log-ratio circuit depicted in Figure 6, employing the AD647, can achieve less than 1% conformance ercor over 5 decades of eur rent input, INA to 100A. For voltage inputs, the dynamic range is typically SOmV «0 10V for 1% ertor, limited at the low end by the amplifiers’ input offset voltage, which can be reduced consider- ably with a pair of crim circuitslike that shown ia the inset. Iy with aps a Figure 6. Log-ratio amplifier. ‘The conversion between current (or voltage) input and log output is accomplished by the base-emitter junctions of the high-B (great- cer chan 100) dual transistor, QI. If Bis high, the base-emitter vol: age of QIAis, toa close approximation, wh wat \ a | i a Ve where kis Boltzmann's constant, Tis the absaluce (kelvin) cemper- ature, qs the charge on an electron, and Igq isthe reverse satura tion current of teansistor A. This circuits arranged to take the di ference of the Vpx's of QUA and QIB, two transistors on the same Analog Dialogue 16-2 1982chip having matching and cracking Js, thus producing an output voltage proportional co the logof the ratio of theinputs: Vour = G (Vara ~ Varn) 2 ‘The seating constant, G, is set by R1 and Ryc.t0 about 16V/V, 50 sto produce a change af 1V in the output voltage per decade (fac- cor-of-ten change in the atio of the inputs). Rucisa special resistor with a + 3500ppm/C temperature coefficient, Since G is inversely proportional ro Rc, it varies in inverse proportion 19 absolute comperature (T), thus compensating for changes in T. As a result, Vourris independent of temperature, “This circu configuration is free from the dynamic problems that plague many other logcircuits. The ~ 348 bandwidth is SOkHz for the topmost three decades, 100nA to 100HA, and decreases smoothly at lower input levels. No addicional compensation is needed for stable operation with input current sources—such as photodiodes—which may have up to 100pF of shunt expacitance. For greater valucs of inpuc capacicance, a 20-pF integrating capacitor around each amptifier will provide smoother frequency response. To calibeace this log-ratio circuit, use this procedure: First apply equal voltages (V\ = V2 = ~ 10V) and adjust “balance” for Vou = 0.00V. Then let V, = —10Vand V; = ~1V and adjust gain for Voury = + LOOV, Ikerate this procedure until gain and balance errorsare within 2mV of the ideal values. SENSOR INTERFACE In many instrumentation circuits, an operational amplifier is used 25 a current-to-woltage converter for che low-level output current fof a sensor—which may be connected to 2 high-voltage source. A typical example isa lame detectorin a gas chromatograph. In such applications, if there is a sensor fault condition, a very high poten- tial may be applied co the inpur terminal of the operational amplifie. Some form of input protcerion must be used to permit the amplifie to suevive the faulecandition Same electrometer-type devices, especially, those involving CMOS circuitry, may require claborate protection schemes employing Zener diodes. However, the protection scheme may well com= promise the overall perfarmance of the ciecuit, nullifying the ad- vantages for which ic was chosen in the fest place, In comparison, monolithic FET amplifiers of the typediscussed here do not usually require such protection, unless the source is not current-limited; in this respect, chey are similar to amplifier types that use discrete JFET devices. The failure mode in such cases is overheating from ‘excess current, rather thana voltage breakdown, The amplifiers in this series are guaranteed for a maximum safe input potential (either common-mode or normal-mode) equal t0 the rated power supply voleage. The input-stage design maintains a high practical level of inpur resistance for differential input vol- Coges of up to +t vols. This property is useful where the amplifier is used as an open-loop comparator of two input signals, either or both of which are directly connected to high-impedance sources. U the souree is not current-limited, adequate protection ean be achieved if a resistorisused, series with che inpuc side the lop, ‘Analog Dialogue 16-2 1982 terminal to limit the maximum overload current to LOmA (for ex ample, 100,000 ohms fora 100-volt overload). The simple scheme shown in Figure 7 will cause no signifi on in perfor- mance, except for bandwidth, and will provide complete de over- load protection. WZ nervneatey nas v0 Fc % ey vom moments ay co Figure 7. Input protection of |:V converter. Output voltage of current-to-voltage converters is proportional to the value of feedback resistance (Ry in Figure 8). For greater sen- sitivity. increased valucs of feedback resistance are used (for exam ple, $ megohms, in Figure 9). However, beyond 10° ohms, eesis- torscend to become expensive, large, noisy, and unstable. Figure 9. Photodiode amplifier ~ current-output photovol- taicmode. Under some ciccumstances (see Figure 10), it fs fruitful 10 use in stead a high value of resistance and a low-impedance resistive di vider that attenuates the portion of the output voltage used for {cedback (since the signal current must now produce a sinaller vol- tage, Re may now be smaller chan it would otheewise have had to be). For example, a 10-megohm resistor and a 100051 divider is ‘nominally equivalent co. 10,000-megohm resistor. The drawback is, of course, chat input voltage errors are magnified relative o the attenuated output voltage at the cap of the R1-R2 divider. The low voltage offse, drift, and noise of the ADS47 enhance the attrac tiveness of this eireust technique. F Vour= tgp Figure 10. Current-to-voltage conversion with gain. High-impedance transducers, such as proportional counters and some accelerometers, require an amplifier which converts aninere: ment of charge into a change of voltage, through the use of a feed- 9back capacitor. The increment of charge in circuits of the type shown in Figure 11 is obtained by variation of capacitance in a capacitive transducer siteing at constane voltage between the sum- ming point and a voltage source. (AQ = C- AV + VAC, and, since SV = 0, AQ = V- AC). The op amp transfers charge to the feedback capacitor and develops the output voltage change Figure 11. Charge amplifier. In this circuit, Ry provides a leak for bias current to permit he out pot voltage change, due to an ac variation of charge, co be observed withour excessive de drift. The time constant, RyCy, should be long, compared with the slowest change of charge (1/(2fyiq)ls but Ryliyas must be small compared tothe full-scale output range. Rin serves to protect the amplificr’s input and output circuit against transient and steady-state voltage breakdown in the capacitive transducer Ve is substantial. Applying a voltage bias to a grounded device is a common need in low-level current measurement, for example, in leakage current cescing, or in using low-level current transducers (¢.f Clark oxy: igen sensors), Figure 12 shows a cechnique chat applies a fixed or adjustable bias at the noninverting terminal of an op amp, thus using feedback ¢o force che inverting terminal to the same poten- tial. The current through the cransducer is sensed by Re, anda low- ost instrumentation amplifier converts che off-ground voltage ‘across Ry toa single-ended output. Figure 12. Current-to-voltage converter with grounded bias and sensor. APPLICATION HINTS The low input bias current (25pA max) and low noise characteris tics of che ADS47K make it suitable for electrometer applications, such as photodiode preamplifiers and picoampere-level corrent- to-voltage converters. In such circuits, iis essential to use guard: ing techniques in the design and construction of the printed circuie board layout to realize the ultimate low-leakage performance of which cheamplifieriscapable. ‘The input guarding scheme shown in Figuce 13 for single and dual ‘op amps will minimize leakage co the input cireuitry from the sup= ply terminals, the outputs, oF other portions of the board wiring. 10 The guard ring should be connected to a low-impedance potential at the same de level as che inputs. High-impedance signal lines should be kept as short as possible; they should be sucrounded by ‘guard lines and kept away from any sourecs of noise and leakage. Off-board, rigid shielded cables should be used for wis Duat Versions 2 9. Single Versions Figure 13. Board layout for guarding inputs. “The fast setting cime and low offset voltage of the the ADS44 make it an excellent choice as an output amplifier for high-accur recy curtent-output dia converters. The upper trace of the oscillos- cape photograph in Figure 14b shows the sttling characteristic of the ADS44 inthe circuit of Figure 14a; che lower trace isthe input Although the ADS44 by itself will setle fast to within 0.01% of final output value, feedback componenss, circuit layout, and cir cuit design must be carefully considered to provide minimum sect! ing ime for the overall circuit in which icis used sin Fee : ree a, Testcircuit b. Detailof settling characteristic. Figure 14. Settling-time tests on ADS44, If substantial capacitive loads must be driven, they will lise the ‘output slewing rate; In addition, the extra pole thac they establish in the amplifier transfer function can be destabilizing and provide distortion, even when the circuit is being used for a low-frequency application, Improved dynamic performance can be obtained by the use of the circuit of Figure 15. A 100-ohm isolation resistor re~ moves the capacitive load from the amplifier’s output, peemicting the 30pF feedback capacitor to provide phase lead in the loop. Low-frequency accuracy is maintained by closing the main loop around the 100-ohm resistor, as shown. Figure 16 shows typical transient response of this circuit b. Transient response Ry =2kM andC, = 500pF. 2. Circuit Figure 15. Driving a large capacitive load. Analog Dialogue 16-2 1982USE YOUR PERSONAL COMPUTER FOR MEASUREMENT & CONTROL Software Support Packages Translate Between .MAC-4000 and Apple, H-P, IBM, DEC Interface with Real World via »pMAC-4000; Use High-Level (e.g., BASIC) Programs The wMAC-4000° is am intelligent, expandable, multi-option single-board measurement and conteo system chat neceptsanilog and digital data, oucpots analog and digital control signals, ee ions, and selectively receives or transmis data under an RS-232 oF 20-mA corn the direetion of a hast computer, slink (Figure 1). The HMAC-4000 imerfaces dircetly, via serew cerminals, with sources of analog input signals, such as thermocouples, RTDs, strain gages, other millivolt-level sources, and 4-(0-20mA current oops. An on-board microcomputer unburdens the host by con- «rolling and locally performing signal anypli zation, channel selection, and conversion « engineering units. tt provides all this at a price of only $1488 for fou hannels of perature measurement vressune itac- 4000 Syste oN | [MEASUREMENT agente Cec Moto “morons “oy Figure) uMAC-4000 measurement-and-control concept. The computer furnishes instructions to~and eeecives information from—the wMAC-4000, in the form of standard serial ASCIL characters, via the standard RS232C/20mA data link, This makes. it possible for the zMAC-4000 to incerface with any host com. puter using this form of communications since most computers do embrace it, as either a standard feature or an option, the wMAC: 4000 is essentially a universal data-aquisicion peripheral The come mation (° wsfer of infor between the host and the sw MAC-4000— including checks on che validity of che daa—is the "C™ protocol, deseribed in the BMAC-4000 technical data. The user must pro= ication format, which embodies the andshake’ vide appropriate software, for che type of compucer being used, to translace between the language of the “application the BMAC-4000's command set. program and Recognizing thac this req for scientists and engineers who do not wish to beconse software mene can be a cime-wascing burden specialists, we have developed—and are continuing to develop— sofeware support packages for use with popalar computers. The first of these ace sofewace drivers using BASIC progeams for IBM. Analog Dialogue 16-2 1982 by John Mills and Hewlet-Packard personal computers, Applesoft BASIC for Applet computees,and MACRO-11 for DEC computers.t The als, and communication cables. An optionab RS-232C communi- cationsirealtime-clock card, which phugs into the backplane, is available for the Apple. It is ideal for data-logging applications. Prives are: Apple—$150 (S388 with Mountain option), IBM. $200, HP-S250, ‘kages inehide Driver Diskettes (Cassette for H-P), manu The ease of writing programs «sing these packages ean be seen in the display segment in Figure 2, which LIST the entire program to read a temperature at one location, and then displays dhe re sponse to 9 RUN command, Since the ¢MAC-4900 retains up- dated data in memory, the response appears immediately after the RUN is executed. A sean of a mumber of cemperatures is almost as simple—and specdy—but of course the display portion of the program will call for a few excea progeam steps if the results are wobelisted, EF List 80,190 BO INE = Ton TRE = "a" FO GISUB S108: KEM LOO INE = "1": REM 110 GOSUE 612602 1a (NCP LAL LZ VALUES VEG CHANNEL 1 PM CNVOKE CIMNNEL 1 READ A200 PRINT TEMPIs 8 4 Tat + EE Reet DISPLAY RESULTS 1300 END A Temrl: 4908.4 BLO Figure 2. Display of program LIST and results of RUN, using Applet Use the reply card for techaical data, The wMAC-AO00 was fist sede in these pages as the cover Featre of Valve 18, Nuno TACIS20 for Apple Hy ACINTS for DEC RSNo1 ACIUIY for HERG: aad ACI22 toe HBA Personal Co se the rep sat far echnical lt, WMULTITASKING The Key to Effective Measurement and Control What Is It? How Does It Help Me Accomplish Results with MACSYM? by Bill Schweber When you are controlling an experiment in a lab oF a process in the plant, yow often need a system that allows you to divide your ‘overall operation into smaller, easier-to-define casks, run them a¢ specified times or under specified circumstances, and pethaps ean several stations simultancously, but at differenc cates—all under the direction of a single mini- or microcomputer systen You can do all this—and more—with che “multitasking” that is a standard feature of MACSYM 2* and MACSYM 10° Measure- ment And Control SYsteMs, Besides making MACSYM easier to use, multitasking allows you to partition your overall program into independene sections and co interact with the system while it is eunning your test or process. ASIMPLE EXAMPLE Suppase you wane co cest a bactery’s charge/discharge characteris ties under camputer conteal (Figure 1), with che following test re quirements: , Secthe char 2. Read the cell voltage and temperature once per second, 3, Perforas calculations based on the last reading and several pre Jing current co the bawery. vious readings 4, Compare the results of the calculations co an established value and raise.an alarm fit isexceeded. 5. Cheek the position of an on/off switch mounted near che bate cory under test. 6. Printout logging report every 10 minutes. Figure 1. Functional diagram of batterytest setup. You would write an individual ¢ask (sub-progeam) for cach fune- tion, Task | would be set once (initialization) and be activated at startup. Task 2 would be periodically activated (once per second) Task 3 would be activated by Task 2, after the readings had been made. Task 4 would be activated by Task 3. Task 5 would be checkin but on 2 lower-priority basis, Task 6 would the switch beautivated every 1Ominutes While che tasks are being executed to run your test you ean choose whether the keyboaed is ca be active (live) or de-activaced. When use the keyboard odo the following Change the time period ofthe tasks itislive, youe, slngpect and print input values, switch-settings, and results of calculations -Change the alarm setpoint or the changing rate sListall or partofehe program ‘orto data ase the reply cae 12 fanage theron *Change outpot voltages, currents, or ondoff settings After you have made changes—or at any time-—you ean de-aeti vate the keyboaed to prevent unauthorized changes. Sample print ‘outs from typical test ransare shown in Figure 2. wgof the rasks—activate or suspend casks efit SEREAL NUMBER (4 praL Ts)? ee20 DESIRED EHARGE RATE CO-L ANE)? 3 DIetwED FINAL CELL VOLTAGE (0 40° YOLTSI™ 7.2 CEL ALARH TEMPERATURE (0-100 DEGREES C2? 98S eAeALARten® CELL OVERTENPENATURE: GEOL PUNSL VOLTAGES 6.00, FEL. FINAL TEMMERATIRES 96.1 Chu SORIA NUMBER (4 DIGITS)? 7104 TE SERED CHARSE RATE ("1 AMP)? 3 DESEREO FINAL CELL VOLTAGE. (020° VOLTS)? CELL ALARM TEMPERATURE (0-200 ULGRFES 0)? TEN FINAL TEHCERATIRES 7S: at 701/79 Gosn7st Figure 2. Printout of battery test results. WHAT IS MULTITASKING? WHAT ARE THE PROS & cons? Asyou have seen, multitasking is scheme that allows you to write the sections of your total program as individual casks and have them run with independent schedules, times, rates, oF occurrences, Vhe MACSYM keyboard can remain live, if you wish, while the program is running, MACSYM automatically shares its time and resources among the tasks, so it appears (0 you as if the processes arerunning simultaneously In che multitasking system, all variables are global, thatis, they are known and available co all casks. This means char data acquired in one task can be used by another task without any additional programmingeffore Since tasks ean be periodically activaced at known times, you don’t have to put timeswasting loops into your program in order co in- sure that a reading is taken at a specific time, or that a desired amount of time clapses between software setivities. In fact, while executing a WAIT stacement, the operating system will automati= cally go ¢o another task to keep the processor busy. The CPU re= source is therefore used much more efficiently and effectively chan inwould be without multitasking Another advantage of multitasking is chara cask may be normally “dorwant” but suddenly assume special importance and require immediate execution. An example fsthe procedure associated with the «ripping of a safety limi, A separate task ean be witten to test faedhe conditions dhat would invokethis rarely used task Yee another advantage exists under stactup conditions. When starting up a complex new program for the first time, and incegear- ing i¢ with rcal-cime Inpursfoutputs, ic helps if you ean being up. ‘Analog Dialogue 16-2 1982tasks one at a time, or in small groups at a rime, and selectively activate and de-activate them. This enables you to test che system Piece-by-piece (hardware and software). The nec effect of this ap~ proach is co reduce che time nceded to debug or certify that a com: plex large program with many inputs and outputs actually docs whatit is supposed to do. ‘The advantages of multitasking are not free—but the price that is paid is relatively small. The operating system for multitasking oc- cupies more memory space than for. system without multitasking, Also, it rans a little more slowly because it has more scheduling to do. However, the small penalty in speed—for most applic tions—is greatly outweighed by the increase in process utilization you can get with malticasking, WHAT CONTROL DOI HAVE WHILE THE PROGRAM IS. RUNNING? From the keyboard, or from within your program, you can activate a task (as a function of either time or circumstances) and assign i a priority, suspend the ewnning of a task, alert (or reactivate) the cask, starting with its next cimed event, or kill 9 task if itis no Tonger needed (to free up memory space and other resources). In dhe example above, you could SUSPEND Task 4, the alarm task, while you fine-tune your program, and then activate it when you are ready to use thealaem fanction. Note that tasks can control other tasks with the same task-control statements; this enables you to write flexible and powerful pro: grams, You can also assign relative priorities to casks so that ess ial tasks are completed before lesssimportant ones, and even change these priorities as needed within your programm. EXTENDING THE EXAMPLE Multicasking makes it easier co concrl several test stations ac he same time, yet have individual control over each one. Consider sition in which you want 10 control several basically sim auiery-test setups, bur with different setpoints and charging rates. MACSYM sets the charging race for each ba tery and monitorsthe cell voltage. With multitasking, you can write a task for each ba tery that would be identical in coneepe, buc might differ in details, stich as charging rate. The multitasking would allow you to change the chaeging raco—or constants in the calculacions—while the tests are running. When you want to change a battery under test, would SUSPEND the task associaced with that cell, then ACT VATE ha local on swvicch of at the system keyboard, The other cell test is unaffected and continites independently without intereuption. WHAT DOES MULTITASKING ACTUALLY DO? The multitasking capability is automatically available Co the user when the’ MACBASIC operacing system is loaded into MACSYM, This operating system cakes care of all the necessary work to make when the new cell is installed, cither ‘multicasking happen, The result is that all the tasks are interwoven automatically in a manner invisible co the user of the system. The molticasking operating system does these things: ‘Ic arbitrates among the casks and causes the execution of one Tine of each task on a “round-robin” basis ff te rasks have equal priorities). You can assign a higher prioricy to acask if you want. ic handlesthe time schedule of periodic tasks. ‘Xr goes co the next cask when che present cask is executing ac: tivities that have an inberont dead time and would otherwise te up the processor needlessly. Examples of such tasksinclide WAIT, fileinpurdourput, analog/digital input/output ‘Analog Dialogue 16-2 1982 ‘#1: keops track of what casks are running and what line of each «ask is being executed. ‘elt keeps the keyboard “live” (if not de-activated) so you can print, change variable values, control tasks, and list programs, while the programsare running, SAMPLE MULTITASKING PROGRAM Perhaps the quickest and most powerful way of demonstrating the power of multitasking is to show how much flexibility even a sinigle-task program can gain in the molttasking mode. The pro gram in Figure 3 measures a variable analog input voltage, X = AIN( 1,0), mltiplis ieby a programmed constant (K), and ourputs the result asan analog voltage, AOT(2,1). [FX exceedsa program= ‘med limit (L}, an alarm is sounded by PNT 7. This sequence is re~ peated at progeammed intervals (T). The initial values of the con: stants K,L,Taregiven as 1,5 volts, and0.5 seconds. 10 TASK 1450 zo activate 1 30 stor 30 Ket LoS 12.5 $0 1eAINt1 07 70 AOTC2 1} akex ec IF XL PHT? 90 ualt T 100 GOO 40 Figure 3. Multitasking: single-task program. The first three program lines (10, 20, 30) define the name and loca Gion of the cask (Task 1, ac line $0), activate Task 1, and indicae that there are no more tasks to be defined (STOP), therefore pro: ceding to execute the rest of the program. When the operator starts execution of the program by typing RUN (carriage retura), the response is “STOP AT LINE 30 (M)”. (M) indicates that MACSYM is runningin the multitasking mode. IE the program were an ordinary single-task progeam, the con: stants could not be changed from the keyboard without either a program change or including an INPUT statement in the progeam, ciher of which would require an interruption of the running ofthe program. With MACSYM's multitasking feavuee, on the other hand, new values of the multiplying constant, the limit threshold, ‘or the waiting period between repetitions can be keyed in at any time, by eyping in new definitions (K = 0.5, L = 8, T = 0.2, ete.) while the program is running, and they take effect immediately when the carriage recuen is keyed. The operator could also, at will, key in che request, PRINT X to print che current value of Ifa second task were added, to begin at line 140, the program preamble segment might look like Figure 4. 1.ine 30 identifies Task 2 as beginning at line 140, If ic were to be assigned priority 1, che Tine would read, TASK 2, 140, 1. Again, the keyboard could act with che program, including he aetivation or susp casks (SUSPEND 2 (carriage recurn)). ‘The simple examples given above should make it evident ¢hac mul n of sking is a powerlly iyused tool forcomputer-based measurementandeconteol [a to task 1,76 20 activate 1 30 TASK 2,140 4G activate 2 50 STOP ry Figure 4. Preamble to 2-task program 3IMPROVED VECTOR-SCAN DISPLAYS ARE COMING AKey is the Fast, 12-bit, Deglitched HDD-1206 D/A Converter ASelf-Contained Hybrid, It Updates at up to 6 Megapixels/Second by Don Brockman and Terry Brown Reams of material have been weitten in eecent years debacing the pros and cons of various cathode-ray-tube (CRT) display systems. ‘Aw easy conclusion is that theee isino “best” systems they must be compared in terms of cheir advantages and disadvantages for a specific applicxtion Even the methods of deflecting the clecteon beam (or beams, for multicolor displays) have been subject co discussion. Systems with yoke around che neck electromagnetic deflection use a magnet of the CRT to deflect che eleccrans; systems with electrostatic de- flection use voltages applied co sets of deflection plates built into che tube to steer the beam. Eleccromagnetic systems are slower and require more power chan electrostatic systems, but they are cheaper dare widely used in TV-type raster-scan displays. ‘The ewe principal forms of display are raster scan and veetor scan. “The TV-like raster scan variably iMluminates a large number of closely spaced clusters of docs along raster of closely spaced hori zontal lines every 1/30 ar 60 of a second. Yau can find a discus- sion of rasterscan techniques, and the application to them of che unique HDG family af compact hybrid deglirched DACs for com ation, in Analog Dialogue 15-2, pages 3-6." posite waveform gen Veetor-sean (alias calligraphic, stroke-writing, random-sean, or dirceted-bearn) display systems form XY plots of digitally deter- mnined points oF digitally programmed analog Vine-segments of randoms length and direction, in similar manner to oscilloscopes in the X-¥-plot mode. It forms a display on the sereen by veetoring the beam, i, by varying its X and Y coordinates, and curning it “on™ and “olf” or modulating ies intensity, as appropriate. Vector-scan systems move the beam only to those portions ol the sercew forming the illuminated portion of the pattern, while the bean chat illuminates the raster scans the entire screen, whether 6 not intensity information is present. In most applications, ras- ter-scan displays use clectromagnetic deflection (an inheritance {rou TV circuitry), while vector types use electrostatic deflection. ‘The former has the advantages of simplicity and low cost; the lat ter requires less power, provides better resolution, and can display Figure 1. Typical vector display system. 14 data changes with precision. Rastee-scan systems are faster iflarge quantities af data are changing {com frame to frame, while vector scanis often the choice where precision isimportant, Although its growth doesn’t match that of the market for raster sean systems, which grows by leapsand bounds, the vector-Jisplay market is still growing at a modest rate and is expected to continue todoso. DAC REQUIREMENTS Figure 1 isa block diageam of a typical vector-sean display system that plots point-by-point, using 12-bit DACs forthe X and Y axes anda fast DAC with lower resolution to modulate the Z (intensity) axis, To obtain che perception of continuous lines when the display consists of diserete points, it is essential to use high-resolution Wa converters co deive the X- and Y-axis inputs. For example, a pair of 12-bie converters wll provide a display of 4096 * 4096 bits, or about 16.8 million piccute elements (pixels). Wich a 21° screen, this would provides resolution ofthe order of .1nom. The high dlensity of pixels makes it mandacory chat only high-re- solution, high-quality DACs be used for plotting the digital data stored in memory. Any aberratian in the DAC output, static or dy- namie, ean introduce an error in output voltage and erroneously position the plotted point. The HDD-1206+ fase deglitched hybrid 12-bic DACs shown in the figure are a new product af the Com puter Labs Division of Analog Devices. Figure 2 shows a simplified randomly progeammable vector-sean beam path. The X-coordinate DAC moves the beam left-and-right, the Y-coordinate DAC moves it up-and-down. The beam is al Figure 2. Randomly programmable vector-scan beam path. ‘Analog Dialoque 16-2 1982lowed to provide continuity between connected points, but it is blanked when it skips over a dark arca, Erravs in positioning in the X or ¥ direction produce lines with distorting lumps. D/A con- verter technology today is sufficiently advanced co ensure that problems with de linearicy and monotonicity are essentially non existent; so de characterstis are not a significant source of errors in vector scan displays, However, besides high de resolution, the dia converters need a few ‘other features. First and foremost is speed. The faster the DAG, che more points that can be illunsinated within a given sean (usually 113010 1/60 seconds, to avoid flicker). The 60-nanosecond settling time of the HDD: 1206 for one-bit changes (while plotting a line) is compatible with a 6MHz refresh rate, and only 2 microseconds are required for an accurate full-scale jump (less time for shorter jumps). Effective plowing strategies would call for many connected points and only shortjumps, when feasible Buc high speed and high resolution are useless if he transition from point 10 point isnot clean. A major problem in {ast high-cesolacion DAG is the discontinuity, or “glitch”, chat ean occur a¢ major transitions, for example, in the one-bit transition from OL1T11111111 co 100000000000. Ifthe less-significanr bits turn off faster than the MSB turns on, the output will rapidly swing to wards zero, then rapidly swing back towards the original level plus cone bit, causing che teace (0 seek to swing wildly aver the face of the CRT in the course of making that small change. Fven at less significant wransitions, highly visible perturbations ca Since the glitch is a code-sensitive nonlinear phenomenon, it can rnot be simply filtered by linear techniques. It must first be ‘minimized in duration and amplitude by designing the switching to be as symmetrical as possible; then the residual glitch ean be climinated by using a track-hold output amplifier circuit. When a new value of digital input is latched in, the output circuit switches 0 hold to retain the previous value until a time when the gliceh «ean be expected 10 have settled out, then switches (0 track (a ae- quire he new value. DOOM Tr rt Figure 3. HDD-1206 functional block diagram. ‘Thisis che essence of che job performed by the deglitching circuitry 10 the HDD-1206, designed for the nex« generation ‘mance point-by-point vector-graphic displays. THEHDD-1206 “This device represents atculy ccmarkable achievement in combin- ing many functions in a single thick-film 32-pin hybrid IC. In- cluded ace input registers, 12-bit current-ourput DAC network ‘Analog Dialogue 16-2 1982 and switches, a reference, the output amplifier with track-hold switching, ciming circuits, and associated electronics. Coding is negative-reference complementary binary for single-ended ouput (nd corplementary offset binary for bipolar outpud; that is, the single-ended output is 0 for all 15, ~5.119V for all Os (1 LSB = 1.25mV). Figure 3s block diagram of the device. The first step ing the glitch isan inpuc register 10 remove the effect of differences in the arrival vimes of the individual bits. All bits are simultane- ously latched and arrive at the input of the eurrent-outpur dia converter at the same time. Nevertheless, despite the best efforts ac balanced design for the converter, there still exists a difference between on and off times forthe switches, which produces. glisch.. The block labeled “timing generator” includes a one-shot multi- circuit provides a fixed hold vibracor and gating circuitry. TI period that exceeds the expected glitch duracion, starting at the Lime the strobe arrives and independent of stcobe frequency. Atthe conclusion ofthe bold interval, the ooxpot ampli is returned to the track mode and slews to the new valuc established by the digi- talinpor. “The role of che FET switches used in the raek-and-hotdis critica, because they affeet the dissipation and complexity of the drive cir ‘uit and the smoothness of che converter's output wavelorm. Cho sen for low input capacitance, low gate cutoff voltage, and tow rain-to-source “on” resistance, they make possible 6MHYz update rates with low power dissipation. Additional benefits are low charge transfer wichin the device—hence smaller residual spikes— and simplified drive ciruiery, resulting in more efliienc use of space and lowercost. WHAT IF NEED FASTER UPDATE RATES? Complete and cost-effective, the HOD-1206 isan ideal choice for vector-scan designs calling for 12-bie resolution and update rates through 6MH?.1f 12-bit resolution must be obtained at faster up: date rates, suitable circuits can be assembled, using excernal input registers, the HDS-1250t (12 bits, 50ns) or HDS-1240Et (ECI., 12 bits, 40ns) current-ourput DACS, and the HTS-0025+ crack: hold—to perform deglicched d/a conversion at update rates from 10 to 20MH Hps-1250. gore 4 isa block diagram of a scheme using che he Figure 4. Applying the HDS-1250 and HTS-0025 for fast vec: torscan. “Use che reply card for a copy ofthe article 4Usethe reply card for echnical dara {Prices feom $135 in 1008 5DEGLITCHING A 16-BIT MONOLITHIC D/A CONVERTER AD7546 Has an on-Board Sample-Hold Switch for Deglitching Simple Circuitry Provides Buffered Voltage Output byJohn Wynne The AD7546* 16-bie microprocessor compatible monolithic Wa converter, a voltage-output device designed to provide 16-bit re solution at high speed—with low power consumption and cost— ina 40-pin OIP package, wasintroduecd in Analog Dialogue 16-1, page 6. The plastic AD?S46KN and the ceramic AD7S46BD pro- vide 65,536 monotonic output levels, in response to 16-bit binary inputs, for respec to + 85°C semperature ranges of 0°C to 70° and = 25°C As Figure 1 shows, the four most-significant bits choose the appro- priate seginent of a {6-resistor strings the voltage across that resi tor is applied, via a pair of external buffer op amps, toa precision 12-bit DAC, cmployinga voltage-switching R-2R ladder. The out- part of dhe AD7546 is equal co the base level voltage, at the lower fend of the chosen resistor (OV, 1/16 Vagnty 2/16 Vir «» 15/16 Var), plus the fraction of the voltage across the resistor estab lished by the 12 less-significant bitsand the 12-bit DAC, the dynamic behavior of the analog output during code changes is nat clean, because the base level voltage can wary ducing sw ing, as intermediate cades appear and disappear; op-amp ciren dynamics also contributes. A typical major-carry “glitch” (8000, to JEFF) can be seen in thelower waveform of Figure 2a. The gliteh, being Large, must be suppressed. Since it is code-depen= dent and nonlinear, i¢ eannot be filtered out in a short time using linear Filtering. Ie is most successfolly dealt with by a track-hold ireuit at the output of the converter, switched t0 hold when data is put on the input, then (© track after che worst: case glitch will have sectled 0 . The glech occurs buts ignored. The AID7546 contains a single-pole, single-throw (SPST) switch, far use in raek-hotd circuits, The switch is synchronized with the two larch-loading signals; itis open when both CHIPSELECT and WRITE ace low. The internal lagic of the AD7S46 ensures thatthe switch opens before the data tothe latches can change, The switeh eb is normally connected as shown in Fi A3 buffers the ISk81 outpuc resistance of the 12-bit converter and drives the hold capacitor, Cy through the on-chip deglitcher s 1 Vout = Vy + Dus (Vs = Ve Figure 1. Functional diagram of the AD7546. switch, A3 should have wide bandwidth, be able to drive substan re load, and have low input bias current Mowing, imal contribution to offset tial capaci through the ladder resistance—for voltage. [AA butlers the voltage on Cy and supplies the deglitched ouput 2X low impedance. A4 should have low input bias current, (0 ‘mininsize droop of the hold-capacitor voltage. Offset voltageot AS and Ad isadded without amplification a the output ‘The degliccher swiceh isa specially designed CMOS switch with lows charge injection (typically 1S0/C, oF 15aV in 10nF) and on resistance of the order of 250 ohms. Cy should be large enough to keep the effects of chaege injection co less than }-LSB-worth of ‘output, yet be small enough for eapid charging through the de alitcher switch. The degltcher switch is brought out to pins 22 and 24; pin 23, a “norconncetion” pin, should be tied co the output ‘of A4asa guatd to minimize feedthrough and droop. The nexe-to-bottom trace of Figure 2 shows typical results, using the deglitcher switch in che configueacion of Figure 3, with AS = M4 = TLOZ1, Cu = 1OnF, and a hold period of 10s. The scale fof the ouput waveform is about 2 LSBs per division. The maximum gliech amplitude is reduced from 1.4V to about 2504V, a factor of about $600, or 75dB; the glitch area is reduced from 1.4RV-st0 1,6nV-s, about 875:1; and che delay is known and con: trollable. When the deglitcher switch is used, the period of the write pulse should be extended to give the ourpuc of A3 time to settle to the sta should remain stable while correct new value, Digital input the write pulse is low. When it returns high, the deglitcher switel closes, and Cychargescothe new value through the switch, [3 b. Circuit connections for deglitching Figure 2. Deglitching the AD7546. Fos techwicaldata,wsethe reply card ‘Analog Dialoque 16-2 1982VERSATILE j.P-BASED METERS FORJ-K-T-E-R-STHERMOCOUPLES AD2050/AD2051 Self-Calibrate, Cold-Junction Compensate, & Linearize Character-Serial ASCII &RS-232€ Digital, 1mV/°C Analog Outputs Available The AD2050and AD20511* are single-channel thermocouple t perature meters with a maximum range of —265°F to + 1999°F or = 165°C 10 +1760°C and resolution of 1%(F oF C). Micropror cessor-based circuitry provides automatic cald-unetion compen: sation, linarization, and self-calibration of gain and offset. The ‘AD2050 is programmed at the factory to interface directly with conc af the following six thermocouple types: (J, KT. E.R, $). The AD2051 is @ universal instrument; the user seleets appropriate conditioning for one of the six thermocouple types via switch prog ming. Typical applications include temperature monitoring and data logging in industrial and laboratory environments. Performance, comparable 10 that obtainable with meters costing 2 «0 3 times as much, is viewally driftfrees typical drift over the recommended 15-month re-calibration interval is less than + I°F (1 digic).Table I shows ranges and readout accuracies with various thermocouple ypc. Table 1. Readout accuracy (25°C ambient. Sensor Type Range Accuracy J =165°C10 760°C 207°C £12LSD J ~265°F 10 1400°F 13°F £12L8D K =50°Cw1250C 209°C + 12LSD K -38°F 0 1999F 6F =12LSD T =150°C 10 400°C 208°C £12LSD T -238°F 10752°F 4F +12LSD E -100°C10870°C oc =12LSD E -148°Ft01S98°F = 22.0°F 9 = 12LSD S,R +300°Cw0 176°C 15°C V2LSD 0°10299°C. 260°C = V2LSD S\R +572°Ft01999°F 23.0°F £1/2LSD +32°F10S70°F 2.0°F =U2LSD Temperature information is displayed on large 0.56°H (4.3mm) LEDs. It is also available digitally in character-serial ASCII format with race selection (25 o¢ 100 characters per second) for easy inter~ face to printers, cerminals, and other peripherals. An optional iso: laced 20mA bit-serial loop output—eapable of 4-witefullJuplex 300-oF 1200-baud operation over distances up to 10,000 f—and a non-isolated TTL-compatible interface are available for com poterinterfacin Linearized temperature information is also optionally available i analog form as a ImVidegree output voltage, for driving strip chart recorders and other analoginstruments, The operating ambient temperature range for specified operation of these meters is + 10°C to +40°C, with =25ppmu"C typical, =60ppmv"°C maximum tempeo. You can order cither meter for power supplies of 120V ac, 240V ac, or—at small extra cost +7.5V to + LSV de. The meters automatically indicate overrange and open-thermocouple conditions. They are protected for 300V. peak input (thermocouple to a¢ line) and up to 1400V peak com: mon-mode voltage (ac versions). Betore shipment, all devices ace bumed:in for 168 hours at 50°C, power-on-offeycled, and ealibrated (tseeable tothe National Bu- reau of Standards). They are supplied in rugged high-impact plas tic cases that mect DININEMA standard dimension requirements Base prices of the ADBOSO/AN2051 are $230/$290 (1-4) and $161/S203 in 100s. C3 * For wshwncaldatawsethe reply cae. is = ieee — a = ——_ Figure 1. Functional block diagram of the AD2050/AD2051. ‘Analna Nialanue 18.2 197 49LOGARITHMIC MULTIPLYING CMOS D/A CONVERTER AD7111 Controls a Wide Dynamic Gain Range of 88.5dB (26,600) with 8 Bits For Constant-Percentage-of-Reading Gain Adjustment, Audio Attenuators, etc. he AD7111° LOGDAC™ is a muti Finear response to unipolar or bipolar analog (reference) input sig nals and exponential eesponse to the digital input. In effect, che Alog input is attenuated by 0.375 dBbit ofthe 8-bitinpuc word. The AD7LL1's nominal transfer function in the cireuit of Figare tis ng dla converter with V Min = expiol = 0.375220) where Vand Viv are te or de voleages (© £25Vt and Nis the bbase-ten imeger value of the binaey umber. For example, if N= 128 (or 10900000,), ~ V,/Vis © eXP4ul ~ 4/20), 080.00398, Re- sistors RT and R2 are noc needed unless che gaits must be adjusted precisely tounity at OUD, Figure 1. Typical circuit configuration. al code and Table 1 shows the correspondence heeween en associated nominal attennation, in JB (= 20 lag, y(ratio)). Note at the more-significant nybble (D7-D4) has 16 steps of atten at 6AB per bit, while the lesssignificanc nyhble (DyDy) dic vides cach 6-dB interval into 16 equal 0.375-dB interval, Figure 2. Functional diagram of the AD7111 The block diageam of Figure 2 shows how the function is plished: The mieropracessor-compatible inpur-latches apply dhe input word ro a decoder, which feeds the code for the correct ratio toa 17-bit R-2R-ladder-type muhiplying DAC. Applications for the device include communications and modem test equipment, telephone and PCM test equipment, digitally eon: trolled automatic gain-comtrol (AGC) ciecvits, audio artenoators (and amplifiers, ifthe DACis used in the feedback path), and wide: dynamic-cange ald converters Accuracy and monotonicity depend on attenuation range, step size, device grade, and temperature range. For example, grades LICIU are accurate (0 £0.174B for attenuations from 0 to 36d and monotonic for 0 t0 $4dB, for 0.375-dB attenuation steps. I the attenoation sep-ste is increased to 0.754, accuracy #5 to within 0.27508 for 0 0 42dB and the device is monocot for 0 to 66dB—over the temperature range. For 3 steps, accuracy — oer temperacure—is 40 within 1.5 for 0 10 S4dB and the device is monotonic over the full eange—of both atienvation and The device is available packaged in plastic, ceramic andlor Ceri, and pricenstartat $16 in 1905(AD71 11KN). Ed Use he reply cand daetccbnical dats Depending theo senp's sealable ouupot range Table 1. Ideal attenuation (dB) vs. input code. eo 811d | too ease ders Nas 0 Sans 978 as as 1B Analen Nialnae 18.9 1087NEW SOFTWARE PACKAGE FOR LTS-2010 TEST SYSTEMS Includes Operating System, Utility Programs, and Family Test Packages Users Perform Tests More Easily; Test Designers Can Be More Creative ln just two short years since its incraduction, hundeeds of LTS: 2010° computer-based autoniatic benchtop test systems have come into use, testing op amps, Wa converters, afd converters, vil converters, and valtage references (employing standard off-the shelf Family Boards), plus an unknown number of devices whose nature is known only to users who have wired their wn test ci cuiery on User Protorype Family Boards. Daring this ime, the first-generation software has proven itself ef fective and quite 9 bit more flexible than that of existing alterna Lives, Asalways, though, there is plenty of room for improvement. Now, a major effort has resulted in new software that iseven more Mlexible, casier (© use, and—very important—eompatible with existing software and hardware. ‘The LTS-2010 package includes an Operating System , which fea- tures a File Manager program, an interactive Debug program, and several utility programs. A restructured, more versatile and easier che-blanks” Family Test progeam is $-2000 and the LTS-2010, BASIC may also be used with the LTS-2010 tooptimize system throughput. OPERATING SYSTEM With the File Manager program, the test programs stored on a single disk are readily accessed (read from, wricten co, deleted, ete.) without user concern for such details as physical location on the disk, block availability ete. There are ewo instewetions for saving programs: PSAVE allows a new program, currently in memory, to be saved on disk; DSAVE allows you to overwrite a program al- ready on disk with an edited version without having to first delete the disk version. You can also save portions af a program, so that later only che segments required need be loaded. ‘The two instructions for loading a program in memory feom disk are: NLOAD clears memory before loading so that the program is Toaded exactly s on the disk; PLOAD doesn't clear memory, al: lowing lines to be added to or replaced in the program existing in memory. he Debug program is an invaluable aid to troubleshooting both fhardware and software problems. When you issue the DEBUG in- struction and respond to the prompts, it causes the operating sys em to pause at ("trap") any selected point in che program (a spe~ cific line or test), allowing you to examine and modify hardware and software setups, The following may be displayed: Calculated output voltage and stace of all voltage sources ‘System reference DAC setup conditions All measurement system conditions Measurement eesult data ‘Codes last specified for the digital control and driver bits, Digital contro! readback and digital device readback bics. Utilities include a Handler Setup Program, a menu-type fill-in-the blanks progcam for automatic handlers; Convert to File, for pro grams written with previous eevisions of the operating system; Function Switch Module, for subroutines (0 automatically per Analog Dialogue 16-2 1982 by Mike Slocombe and Al Finger ar 1 form such functions as Datalog Stop on Fail, Yield, Statistical Analysis, ete; Statistical Print of files generated by the Funecion Switch Module: Disk Copy; and Remuntber, which ecads a BASIC program from a disk, modifies it, and returns it co the samme disk, but with a different file name. FAMILY TEST PROGRAMS, Family Test programs permit specific classes of devices (op amps, DACs, regulators, ete) to be comprehensively tested, by the use of menuctype “fillin-the-blanks" programming. Three executable programs areincluded: CREATE, used to generace andor edie the device test program; TEST, which eauses the tests to be performeds and CONVERT, which converts progeamscreated with previous software to the correct format for use with the new sofeware. In CREATE, there isa choice of four modes: Command, Edit, Set VO, and File. The last chree modes can also be reached from COM- MAND, which is also the jumping-off point to: Menu, List, and Define. MENU displays the list of available tests. LIST displays a listing of test programs currently in the test program buffer. DE= FINE causes the program co enter a question-and-answer dialogue to define the general characteristics and initial conditions of the de vice to be tested. When you have defined the device to be tested, the program will give you the opportunity to build a test program by stepping chrough each cest function, using these questions: NEXT TEST IS: test name, DEFINE? (Y,N) IFN", the testis skipped and it goes on to the next. If “Y" dialogue for that testis entered. After stepping through all the questions, ic will repeat co permit checking and revision—eseape is possible at any time by up-arrow and carriage return, leading co: REPEAT THISTEST? (Y/N) “Y" indicates that you wish the program co copeat this test; a new dialogue will be started. "N” takes you to the next test. EDIT mode allows you toeditexisting device test progeams. may entail listing, viewing, modifying, inserting, and deleting test fune= tions, as well as editing the Definition conditions. SET VO permits you to select or de-select output logging devices oF switch com- mand input over to the LTS console. FILE allows you to perform various disk maintenance functions and load a new progeam. “Fortcchnical data use the reply cae. The L-T5-2000 was described at length in Analog Dialogue 13-1 (1979), the LTS-2010 BASIC: programmable version waedeterbed in 14:2, pae 10. 19HIGH-PERFORMANCE HYBRIDS FOR DATA-ACQUISITION 12-Bit AD5210 Series ADCs for Accuracy Over Temperature AD5240 12-Bit ADC (51s Conversions) and AD346 S/H (2,1s Acquisition) for Speed ‘The ADS210* and ADS240* 12-bit a/d converter families AD346°" crack-hold amplifice are high-performance-and-r ity DIP-packaged devices with industry-standard pinouts designed for applications in military, industrial, and commercial data-ac- ‘quisition-syscem design. They have versions designed for high per- formance over the ~S5°C to + 125°C temperature range, are available with sercening to MIL-STD-883, and have existing sec- ‘ond sources. Because of their lower chip count and the use of a hermetically sealed SLAM ceramic DIP package, with its minimal number of bond wires, all theec devices are characterized by higher predicted reliability chan comparable devices now on che marker. ‘AD5210 HIGH-ACCURACY ADC FAMILY “The members of the ADS210 family are high-performance adjust~ ment-frce 12-bit a/d converters that guarantee + 1/2 LSB non- linearity and no missing codes over the operating cempeeature all versions) —and maximum gain error as smalls + 0.1% of full-scale-range (#4 LSBs) over cemperature (ADS214 and ADS215). Maximum conversion time for all versions is a fast 13, microseconds. The table indicates the principal distinecions be- tween members of the family = S810 =25%C10 Ste —10Vea Inscensl External Type suse gas esvynov "Re Rel ADsaiin x x x ost x x x x x x x x x x x x ApsMT x x x S22 x x x Apsast x. x x Other input voltage ranges are also available. Prices (100s) are $1$8/$257 for BDITD versions, $273/8385.50 for BO/TD/S838, with substantial discounts for higher quancities. itera SREY — Faia aeat o ae aes wet {B)ourrenour oe} fe sity wh) ny Fr omsonnt we we Pay msce scot P21 aroun i a = J Bint “ wernnee bse au mye S |||p=e svnome Gabe [RES FT tases Figure 1. Functional diagram of the AD5240. 20 AD5240 HIGH-SPEED ADC FAMILY The ADS240KP and ADS2405D are 12-bit a/d converters with Sus maximum conversion time and guaranteed no-missing-codles over their operating cemperature ranges. Besides being fully com patible wich requirements for $240-cype converters, its pin-com- patible with ADCBS sockets and can be used to upgrade conver- sion speed (from the 10s conversion time specified for ADC8S), for a substantial increase in chroughput rate, especially when used with the AD346 sample/hold, described below. (The AD ADC8S* is available for ADC8S applications where 10hs conversion time is adequate) An important feature of the ADS240 family is a“Z” version, chat can provide specified conversion speed over the full-MIL. tempera lure range while operating on + 12-vokt power supplies. Included in the ADS240 is an on-board buffer, which can be used to unload signal sources chat are “sofe” ac the high switching rates encoun: tered in fast successive-approximation ADCs. Prices of the ‘ADS240 family stare at $15 1in 100s (ADS240KD). AD346 FAST SAMPLE-HOLD The AD346 is a high-speed sample-hold suitable for use with 12. bit ald converters, such as the ADS210 and ADS240. Its guaran- teed maximum acquisition time is 2ys 6 0.01% of final value, for 1 10-V step, increasing to only 2.Sps for a 20-V step, ‘Thus, in a single-channel application with the ADS240, such as that shown in Figure 2, a theoughpot rate of 143kH2 is available, Price of the AD346 starts at $56in 100s. ED a I ot ve aot ene eT Figure 2. Applying the AD346 with the ADS240 in a 143kHz 12-bit conversion system. *Fortechnical data, ase the ply cae Analog Dialogue 16-2 1982SUPER-PERFORMANCE HYBRID OP-AMP FAMILIES Combine Low Offset Voltage, Low Drift, and Low Bias Current With Clean Fast Settling, Slewing, Wide Bandwidth, and Load Driving ‘The A380", AD381*, and AD382° FET-input op-amp families combine high speed and prevision, without external trim circuitry. For example, the AD381L and 382L have 0.25-mV offset and 700ns settling to 0.1%, the AD3BOL, with IyLV maximum offset, settles 60 0.01% in 250ns, and the AD380and AD382 families will Urive #10-V loads co £50mA. Their performance (typical at 425°C and nominal supply voltage, unless noted otherwise) is summarized in the following cable: AD3MOLAD3SIL_— AD3S2L Oper loop gain (with Ry, VV mie 25k(2000) 100k2kA) 40k(2000) Offset woltage, mV, max 19. 02s 0.25 Drife,uWPC, max 10 5 5 IhyassPAmax 100 3s 25 Setcling ime 00.01% (10Vstep) 250nsN2as Ts Settling time 100.1%, ns 130 700 700 Slewing rate, Vins 30 30 30 Small-signal bandwidth, MHz 40 5 5 Gain-Bandwidth Product @10MH2, Miz 350 Joners and min, at = 10V =s0 =10=50 Package configuration Tos T0899 TOR Typical applications of these amplifiers are in high-resolution graphic displays, fast current-co-voltage translation for dia con- verter circuits, photodectector signal conditioning, fast sample- holds, pulse conditioning, precision buffering for driving coaxial ‘cables, and pin drivers for automatic testing. ‘AD380- FAST-SETTLING 50-mA DRIVER ‘The AD38O has a slewing rate of 330V/us and will output SOmA inco. 200-ohm load. When used without compensation, its 350MHe gain-bandwidth product permits wideband closed- loop response, for example, gain of 7OV/V with 3 SMH2 bandwideh. When compensated, it provides clean, fast setling, a8 noted in the table. At the same time, its low offse, drift, and bias ‘current make it suitable for general-purpose measurement circuits that muse respond stably and accurately co a wide range of high- and low-level de and ac signals, with a minimum of loading. Prices version) start at $34.50 in 100s, ange igen picira a) ourruts TUTSTSTSTET Figure 1. AD380in a butter application. ‘Analog Dialogue 16-2 1982 Figure 1 shows how i¢ might be used as a fast-sertling buffer be- tween a sensitive de measurement circuit, having slow response dynamic output impedance, and a successive-approxi- ration a/d converter, ta keep the fast bit-current pulses appearing at the ADS78 input from affecting the accuracy of the measure- mentcireuit. Figure 2 shows the large-signal output response of the AD380 and saring at a dummy summing point, in a unity-gain itemploying Skf input and feedback resistors. Figure 2. Unity-gain inverter, large-signal response of D380. AD381- LOW-DRIFT OP AMP The AD381 combines low offset and drift with compensated, moderately fast response. Its low bias current, as with all Analog, Devices FET-input op amps, is specified conservatively. Its out- ‘pot furnishes = 10V at # 10mA. Prices start a¢ $8.50 in 100s J version). AD382— LOW-DRIFT HIGH-OUTPUT OP AMP ‘The AD382 combines the de and ac characteristics of the AD381 ‘with the high outpuc drive of the AD380, a useful combination in line-driver and buffer applications that necd high gain-accuracy over the audio frequency range. Prices (J) scaccac $17.50 (1008). *Portechies| dats, vsethe reply card aNew-Product Briefs 12-BIT, 5-MHz ADC FOR MIL APPLICATIONS The MOD-1205MB: Complete, Compact, and Characterized A Documented Version of the Proven MOD-1205 A/D Converter FAST TRACK-HOLD 1.3.5 max Acquisition Time To 0.01% for 20-Volt Step The H¥C-0500° ic a fase track-hold ciecvie ina 14-pin hermetically sealed metal DID, is combination of speed, low power dissipa: tion, and low cost makes it attractive for ap- plications with a wide variety of monolithic and hybrid a/d converters; for example, «0 obtain a 250kH2 chroughput eae with the [2-bit, 28-ps HAS. 1202, [As Figure 1 shows, the HTC-0500 provides sign inversion, and the tracking loop is closed externally, permitcing the user to add external resistance in series with the input and feedback resistors tor adjust the gain around its nominal magnitude of 1.00. The HTC-0500's acquisition time to £0.1% is typically 700ns (8OOns max) for a 1Owolt step, and aperture time is 30ns (£5ns), with an uncertaincy of 100ps max. Its droop rate of 0.002mV/us permits rela tively long hold times. The HTC-0500 has modest quiescent power requirements: 810mW max. Besides serving asa low-aperture-time front end for afd converters, the HTC-0500 also hhas uses as a dleglitcher on the output of da converters, in analog peak detection, and for other analogapplications. Available for both the -25°C to +85°C and ~ 55°C co + 125°C temperature ranges, the HTC-0500 is also available processed to MIL-STD-883, Method 5008. Pricesstant at $79in 100s (HTC-0500AM) EI ox ween Figure 1. Functional diagram of HTC-0S00 *Usethe reply aed or technical bata 22 The MOD-1205MB* is 2 version of che MOD-1205, the popular 12-bit, 5-MH7, analog-to-digital converter (see Analog Dialogue 14-1, 1980), designed for use in radar/sonar digitizing, real-time spectral analysis, digical commonieations, and simmi- lar functions typically found in defense-re- lated applications. The following fearures may be of interest «0 users for these applications: © All specifications are characterized wich minimax limics, © All components are procured or sereencd in accordance with appropriate MILS'TDs. © Fach converter bears a serial number for traccability. © Each converter is shipped with serialized test data showing results of static and dynamic testing, Each device is complete; no external parts are required, And logic is TTL-compacible, climinating the need for external BCL-TTL- ECLeircviery, Constructed on a single 5°x5.43°%0.5 (127 138 x 12.7mm') _ printed-cireuit card, che MOD-120SMB includes a track hold amplifier with 25pm and output registers. "The singlepiece price is $4995, with substancial discounts for want. CA MONOLITHIC ADC—12BITS + SIGN CMOS AD7552 is Ratiometric, Microprocessor-Compatible Has Polarity, Overrange, Low Power Dissipation, Low Cost The AD7552* is a monolithic integeating aud converter with 12-bit-plus-sign resolu: ‘ion, plus overrange, housed in a 40-pin dal incline package, Drife is minimized by patented Quad-Slope conversion, which converts offsets 10 a digital number that is then used for correction, ‘The AD7552's parallel output daca lines hhave three-state logic and are micropro: cessor-compatible. ‘The lower cight LSBs and the five MSBs can be enabled separ- ately, for a choice of interfacing to 8- oF Vé-bit data buses. ‘Three flags—OVER- RANGE, BUSY, and BUSY, controlled by STATUS ENABLE, simplify interfacing comicroprocessors, ‘The AD7552 is essentially complete, requir ing an integrating R-C, a reference circuit, and clock capacitors for self-contained op: eration. The integrating amplifier, the com- parator, logie, and analog-switching circuits are on-board. With a 250kHz clock, che conversion time is approximately 160 mil- liseconds. Conversion may be either self starting or externally stimulated. A gated se- rial pulse crain is available for isolated cor, version using optocouplers and. off-chip ‘The AD7552KN is packaged in plastic and priced at only $9.95 in 100s. Ed ‘Analog Dialooue 14-2 1987New-Product Briefs LOW-COST PRECISION IC REFERENCE AD1403 & AD1403A: 4.5V to 40V In, 2.5V Out Improved Replacements for Standard Types The ADI403* and ADI403A* are three: cceminal devices thae accept input voltage fram 4.5V to 40V de and provide a stable, 25V at cureents up to 1ma Based on a patented bandgap circuit, these references. use Inser-wafer-trimmed. thin: film resistors to achieve initial tolerances co within 10mV and temperature stability 10 deter than +25ppm°C (AD L403). There are a number of applications for which they areideal: Their accuracy and sta~ bility ac low cost make them first choice for ia and a/d converters. Since they can oper ace with S-volt supplies, they are just right for applications where only a single 5-V logic supply is available. They will operate as precision current sources for currents from 1.SmA to ima. Last (but haedly least), they will serve interchangeably with economy and distinction as a superior first choice wherever generic 1403 or 1403A- type references have been specified. 18-BIT SYNCHRO-TO- The lowest-cost members of a distinguished family (inchiding MIL-temp and 883 ver sions) of monolithic precision bandgap re ferences including the ADS80 (2.5-V), the ADSB1 (10.0-V), the ADS84_ pin-proge rammable reference, and the ADS89 1.2-V ‘ovo-cerminal reference—the ADI403 and ADIAO3A are packaged in 8-pin miniature DIP packages and are specified for opera- tion from 0° 1 70°C. Prices for the ADIAD3N/ADI4O3AN (1003) are $1.65/ $475.8 DIGITAL CONVERTERS SDC/RDC1727 Accurate to Within + 19.8” Over Temperature Available for Systems with 400Hz or 1000Hz References Models SDC1727° and RDC1727*are Syn- chro- and Resolver-to-Digital Converters with resolution af 18 bits (4.9 are-seconds) and accuracy (o within + 19.8 arc-seconds max over the full temperature range (op- tionally OC w +70°C or ~SS°C 0 + 108°C). Packaged compactly in 3.5” x 2.5" x 0.875" (63.5 x 88.9 x 2.2mm) modules, they are self-contained, including transformers on both the signal and refer cence inputs. “They arc intended for vse with high-aceu~ racy synchros and resolvets as an alternative to two-speed (coarsefine) synchro and re- solver systems, Used with high-accuracy ror- ary devices, they ean provide an angular measurement system of comparable accu racy but having none of the wear problems associated with coarse/fine mechanical gear ing. Applications are to very-high-accuracy angular measurement systems, uch as test ‘Analog Dialogue 16-2 1982 equipment, gun sights, stabilized platforms, and fie-control systems. In addition to digital position information, these deviees also make available a de veloc nal for tachometric feedback. They are ccompatible with the industey-standaed jt-device pinout, but they have extra iy input pins for the increased resolution, Small-quantity prices are $1995 (0°C_t0 70°C) and $2295 (~$5*Cro + 105°C). EE sc the top ca for technical dat, 10-BIT IC DAC High-Performance AD DAC100 MIL-STD-883B Available The AD DACIO0® is 3 10-bit monolithic carcent-ourput digitalto-analog converter packaged in a hermetically sealed 16:pin ceramic DIP. The Analog Deviers AD DACIO0 is form, fnetion-, and_pin- compatible swith the industry DACIO0¢ itis available with equivalents © all he popular versions and options, includ ing processing to MIL-STD-883B. andard Using ten precision high-speed current- steering switches, control amplifier, a stable buried-Zener reference, trimmed Si-Cr resistor network, the device produces a fast, aecueate analog ontpat eur rent, Also incinded are laser-trimmed appli- cation eesistors for accurate, stable eurrent= and a laset cowvoltage conversion, using an external op amp, such as the ADS44 or the AD OP-07. Nonlinearity of dhe AD DACIOO is less chan = 1/2 LSB (0.05% FS) max for 10-bie con: version inthe K,L full-scale cempeo is = LS ppm?C in the Ly 230 ppm'Cin the K 8: T, and £60 ppmC in dhe J and S. Maximum settling time to 20.05% FS (1/2 LSB) is375ns, dropping co = 100ns for =0.8% FS (8 LSBs}. ‘The AD DACLOO is available in JD, KD, LD options (~ 25°C 19 +85°C), 8838 versions of chose options, and $D/883B and TDI BIB (~S5°C to + 125°C) options. S-vole- foll scale and chip versions are also avail- able. DIP prices starcat $14.10 JD=100s).E9 “Uscthe reply eaedfortcchnical data CELE 23New-Product Briefs POWER SUPPLIES 2ULTRAFAST OP AMPS: AD3554 & HOS-050C Modular AC/DC & DC/DC Types 18 New Models in New Catalog Ifyou build anything chat uses any of our ac: Live components, you need a power sup- ply—to obtain de power fromline voltage or tw convert de pawer at an existing supply voltage to 2 more-suitable level (or dual levels). Is not surprising that, recognizing this important requicement, we've always made low-cost, reliable power supplies of various eapaciies available with the eest of ‘our product line, The new 12-page eatalog shown above* de- scribes our peesent line of more than 40 modular acide and dele power supplies of various capacities, sizes, and numbers of ‘outputs for printed-eircuit board and chas- sis mounting. Included in the catalog are separate Selection Guides for acide and de! de power supplies and a set of outline draw. ings. Inaddition, coi power supplies, there is information on test conditions and procedures, power-distribu tion, grounding, decoupling, transients, and thermal deracing ACIDC: P-C-board-mounced supplies in this category provide single SV ouiputs at up (© 3 amperes, dual = 15V output at up to £350mA, and combinations of the ewo at upto 1A/= 1S0mA, Chassis supplies furnish 2 similar range, with up to *500mA at + 15V. Inaddition, here isthe unique 2835 transducer supply, which furnishes re sistorprogrammable +1V to +15V at 125mA for transducer excitation and 15V at £65mA forinstrumentation, users in applying the DC 10 DC: A wide range of supplies inthis clas, includinga new series of -wan 24-pin Dib-compatble unis, provides = 15V from SV, 12V, 24V, and 28V de primaries, at up to £410mA, and SV from SV at up to | ampere. *Use he reply catd wobiain copy. 4 AD3554 Is User-Compensated to Fitthe Application Slews @ 1000V/x5, Settles to 0.05% in 150ns Max ‘The AD3554* is a FET-input hybrid opera: tional amplifier, packaged in a T Jestyle hermetically sealed metal can, capable of de- i }OOmA output curren at + 1OV. Lishigh slewing rate (1Vins) and fase seucling ‘ime £0 0,01% (250nsmax) make it ideal for Wa and a/d conversion, sample-hold, and video instrumentation circuits. It is a pin compatible replacement in existing applica tions specifically calling for generic 3554- ype amplifiers ‘The illuscration shows the typical range of ‘gtins and bandwidths available as a func tion of the external compensation capaci tance, With its IL7GHz gain-bandwidth (A = 1000), ic is an ideal choice for small-sig- nal high-frequency amplifier applications. DC performance is also excellent: iis laser trimmed for ImV max offset (AD3S54B) and has bias current of omy SOpA max. leis available in 3 offsecdritt grades: A & B for 28°C w@ +85°C, and § for - 35°C 10 + 125°C. Prices start at $47.50in 1008, EA “Usethe reply card for technica dats Low-Cost HOS-050C Is Internally Compensated Settles to 0.1% in 80ns, 0.01% in 200ns The HOS-050C is the newest—lowest cost—member of the HOS-050 family of high-speed wideband hybrid operational amplifies, packaged in che heemetically sea- led TO-8 can. All members of the family ace characterized by a stable, essentially, 6-UB> per-octave rolloff co 100MH2 unity-gain bandwidth, slewing rate of 300V/gs, and settling time of 80n5100.1%. ‘All models have a rated output of = 10V at © 100mA. The “C” version differs only in temperature range (= 25°C to +85°%C) and offseudeift (65mV max, 200nV/°C max}. tis price is only $55 in 100s. APPLICATIONS With its high gain-bandwideh, fast settling, and high ourpuc drive, the HOS-050C is well-suited to applications with fast current ourpuc Wa converters. The Figure shows a circuit employing itin an inverting applica tion as an output (unipolar or bipolar) amplifier for the HDS® series of fast 8-10: 12-bit DACs (available in TTL and ECL ver~ sions, current-setting times of 10ns to 40ns). E “Use che reply caed for technical dataNew-Product Briefs TEST SUB-PICOAMP BIAS CURRENTS Automatically on the Computer-Based LTS-2010 Tester Using the LTS-0607 Low-Leakage Socket Assembly ‘The LTS-2010* can cese single, dual, or quad op amps for leakage current from Ina, to $0 (A, using the specially designed LTS. 0607 Low-Leakage Socket Assembly with the LTS-2100 Operational Aimplifier Family Board and the Low-Leakage Test sub: routine option in the Op-Amp Test Create program. Testing low bias currentst is difficult be- cause leakage in the test circuitey causes er rors, and the output voltage of the cest must be substantially larger chan che amplifice’s offset The LYS-0607% entite assembly is shielded and rounded to minimize ac pickup, and Tefton pins are used forthe socket-assemly inputs and the pins on the DUT socker ci cuit board. The socket itself is shielded by a hinged cover, which closes over the device. Conscientious care and cleaning on the part of the user ate key to continued satisfactory testing. Instead of unreasonable values of resistance, the LTS-0607 uses a stable, precise 1000pF capacitor. The bias current, lowing through the capacitor, develops a AV/AL = WC. ‘Thus, 1 pA of bias current would produce a AVIAUC of tmVis. The LTS sofware allows the eapacitor voltage to ramp for a suitable period of time, then computes 1 = C(V> — Vi)/{&0). An ADS90" on the fixture meas tures the ambient temperature to correct its effectonbiascurcent. EI + Fortcclical data use the eoply ted For detail, sce Eloctronic Design, Now. 12, 1981, p3ges225-230. ACCURATE THERMOCOUPLE MEASURING Reference, Cold-Junction Compensation, and Filtering MACSYM’s STBO3 14-Channel Isothermal Termination Panel The STBO3-01/STBO3-02" are 14/28-chan- nel isothermal termination panels that inter face thermocouples to MACSYM systems ‘The STBO3 accepts any combination of popular thermocouple types Jy K Ty Ey Ry §).28 channels require only 2.7!" af vertical spacein a standard 19° mounting rack. The $TBO3, compatible with the AIMO3* solid-state analog input card, provides pas sive 2-pole filtering, open-thermocouple de- ection, an ADS90 temperature sensor, and input protection against 120V acems. The high-density package design minimizes rack space and insures accurate measure- ments by maximizing thermal conduction within the cerminal acea and insulating it from external thermal gradient. Dato aequired via $1303 and AIMO3 ean be subjected to the software AVG mode—co average many readings of a channel at specified cime intervals (at 60H for line-fre- quency rejeetion), up t© 4000 samples per second—and the SCAN mod. smany channels atup co 20008. ES “Usethe rely cad forteshnica data to sample ‘Analog Dialogue 18-2 1982 peMAC EXPANSION Analog Output & Digital 1/0 Increase System Capacity Ourpur and Digital VO Expansion Boards for use with the wMAC-4000 single-board incelligent measurement-and-contral sys- tem. Both boards plug directly into the AMAC-4000 card cage and are cconomi cally priced from $790 & $545, ‘The wMAC-4030 contains cight separate 12-bit dia converters and provides eight uscr-programmable channels of cither vol: fage ourpur (several ranges available) oF optionally —+ 1000-volt-peak isolated finpurto-ourpue or channelto-ehannel) 4-t0-20mA or O-t0-20mA _currentloop comput. Its 130-Verms-proteeted voltage and iso- lated current-loop ourputs may be mixed on the same board in blocks of four and are in tended for use in controlling motors, valve positioners, and servomechanisms, or for displays, using analog indicators or revord- ers. The high common-moue-voltage isola tion ensures reliable operation of the cur rent loop in the presence of transient and, fault voltage—a_ncecssity in industrial ‘The wMAC-4030's on-board micraproces- sor provides key functions, such as incre menuideerement for backup control, data readback for bumpless transfer from) nian: ual 10 automatic control, programmable slew eate to conteal the rate of advance from one sewing to another, and output scaling in cngincering units, viaa simple command set. ‘The AMAC-4040 isa low-cost 64-channel digital inpovourpat (VO) board with 32 channels of digital oucpat and 32 channels of optionally isolated digital input. Ic incee- faces contact closures and parallel digital in puts (0 the #MAC-4000 for load-control and direct digital conteol. EX *Use the eply card foe teaimeal da, oFWorth Reading 2-VOLUME 1982 DATABOOK The Analog Devices 1982 Databook is a 1536-page com prehensive wo-volume collection of technical information on daca acquisition components and sub: systems for measurement and con «rol. I's free, upon written request According to Corporate Director of Marketing, George Adams, “With the introduc than 60 new products since our on of more last catalog, Analog Devices con- tinucs to offer you the broadest selection of data-acquisition solu- tions, So many, in fact, that rwo volumes are now required to describe them, Our broad and rapidly growing line of integrated circuit products is described in Volume f. Volume 1 includes our wide range of modules and subsystems, from signal-conditioning components to complete intelligent data-acquisition systems. “The volumes arc organized functionally and are fully erosssin- dexed, so that you can easily find the solution you're looking for, tno matter which volume you hold in your hand. The identical Index in cach valume and the Selection Guides provide page loca~ tions of all products “Whether you're long-time user of our products or encountering, them in the Databook for the first time, we're convineed that you'll nd the Databook a bighelp.” Myouhaven't received your fee enpy yet, write now ADI AUTHORS IN THE TRADE PRESS Get a Fast, Accurate Handle on FET-Amplificr Bias Currene,” by Allan Ryan. Flectrowie Design, November 12, 1981, pp. 225-230 “Low-Level Analog Signal Conditioning,” by Bob Butler. Flec- tronic Products, November 30, 1981, pp. 65-67 “Low-Cost RMSIDC ICs Upgrade AC Measurements,” by Lew Counts, Chuck Kitchin, and Walt Jung, EDN, January 20, 1982, pp. 101-112 “Fixture, Tester Whip Through Instrumentation Amplifier Tests,” by D. Perkins, Flectronic Design, February 4, 1982, pp. 131-136 “VF Converters Demand Accurate Linearity Testing,” by Larry DeVito, Electronic Design, March 4, 1982, pp. 105-110 “Computerized Measurement & Control,” by Bill Schweber, Quality, March, 1982, pp. 38-41 FEATURED IN MACSYMIZER Volume t, Number 4 Special New Product Issue: 8 New Products, Plus: Communications Channels (Part 1), MACSYM 2/10 Error Code Cross Reference, Installing/Removing Boards, Ther- mocouple Table, MACSYM 2 Workshop announcement, Traist ing Course Update. Volume 2, Number 1: “MACSYM 2 Talks t0 PDP-11/34 at Di mond Shamrock,” plus these Features: Digital Signal-Handling, Techniques, MACSYM 2 as an IEEE Controller, MACSYM 2 (Disk) to MACSYM 20 Programs, Task State Diagram, plus Sys teni Notebook, New Items and Field Support. On AUTHORS (continned from page 2) Jobn Mills (page 11)is Senior Mar- ting Engineer for Subsystem Products at Analog Devices, John earned his BSEE degree at North- castern University. He fiest joined Analog Devices as an Application Engineer, then became a regional sales manager for Datel-Incersi and subsequently rejoined ADI in his present capacity. He enjoys softball, golf, cross-country skiing, and is a excquetball enthusiast Bill Schweber (page 12) is a Sys tems Application Engineer with ADI's MACSYM group. Before this, he designed microprocessor based controls for materials test ing equipment. Bill has a BSEE from Columbia University, an MSEE from the University of Massachusetts, and is a Registered Professional Enginecr. His hobbies include bicycling, photogeaphy, model railroading, and “just cak ing care of things around the house, which often has a priority over the first three items.” Mike Slacombe (page 19) joined ADI's Component ‘Test Systems organization as System Software Designer after 5 years in the ADI Semiconductor test-engincering ‘group and a brief stine at LTX. AC ADS, he designed computer-based inchouse test equipment and im- plemented cesting of many of the monolithic products on commer cial «esters. He formerly worked at Raytheon and has auended Northeastern University and the Berklee College of Music. Dow Travers (page 6), Senior Ap- plications Engineer at Analog De- vices Somiconductor, Wilmington, Mass, has BSEE from Northeast- crn University. Before joining Analog Devices, he supervise the converter test department at Hyb- rid Systems. Ac ADI, he assisted in the development of data-acquisi- tion products at our Microelee: tronics Division. After a bricf stint at Parlex Corp., where he de signed custom thick-film hybrids, he rejoined ADS in 1978 in his present capacity. John Wynne (pages 3and 16) isan Applications Engineer at Analog Devices B.V. (Limerick, leeland) Previously, he was Telecommoni- cations Project Engineer in the Gity of Dublin Devin Street Tech- nical College, and later a Design Enginecr at Sperry Gyroscope Av- ionies Division U.K.Potpourri ‘An Eclectic Collection of Miscellaneous Items of Timely and Topical Interest. Further Information on Products Mentioned Here May Be Obtained Via the Reply Card. IN THE LAST ISSUE (Volume 16, Number 1, 1982 - 24 pages) . . - Hybrid Isolation Amplifiers with a New Twist (AD293, AD294) . .'. World's First Monolichie 16-Ble D/A Converter (AD/546)~ - + HonoTithte 12-Bit DAC Has 6-Word FIFO Memory (AD7S44) .. . Monolithic Systems DAC (AD/527) . . . First Monolithic FET Op Amp with luv/"C Drift (AD547) . . . Low-Cost 10- and 12-Bit Interfaces for STD Bus (RTI-1225 & 1226, RTI-1260 & 1261) . . - Testing A/D Converters Automatically . . . MACSYM 10 for Indust rial-Plant-Floor Environments . . . Double—Buffered Complete Monolithic 12-Bit D/A Converter (AD367) . . + High-Resolution Data~Acquisition Family. . . and these New-Product Briefs: fvene-Counter aed for MAGS (EVCOI) + + Pulee=Teein Output Card for MACSYM (POCOL) . . « Loop-Fowered Isolator for Industrial Applications (2B24)- « 12-Bit uP-Compatible Monolithic Multiplying DAC (AD7545) . + . Improved 12-Bit Complete IC A/D Converter: High-Speed Interfacing and FID Operation (ADSTUA) ..”. Fast 10-Bie ADO Converts Tn T,dus wax (ADS79) « « . Software for ‘uMac-4000 Compatible with DEC RSX-11, HP~85 and Apple II Computers . . . 4-Channel Signal Conditioner and Wule{plexer (2834). .- 10-Bie Tracking Inductosyn/Resolver-to-Digital Converter . » . Ivar Wold Named First Analog es Corporate Fellow » . - plus Editor's Notes, Authors, Worth Reading, Potpourri, ete- ERRATA . . . DATABOOK, General Note: While assembling datasheet material to appear in the Databook, we seized the opportunity to incorporate corrections of all known errors in previous editions of the \eets. In cases of conflicts between the Databook and individual data sheets, give the Databook the benefit of the doubt for all data sheets published before December, 1981. If the data sheet is wore recent, it is the more likely to be correct. You can find the month and year in which a data sheet was published on the last page at the upper right, following a set of vertically oriented characters, e- g-, CO92-9-2/82 . . . Some errata corrected in the Databook: DACI106/1108 Absolute Accuracy Offset is +20nA (correct on page 10-12 of Volume IT). Logic foupattbiliey of the 158/460 -v/t converters (“Voltage to Frequency Operation") is cotreetly stated in the Databook (page 12-19 of Volume II). All AD2040 errata found in previous editions of the data eet have been corrected in the Databook (pages 16-53 and 16-54 of Volume II); they relate to: (1.) the notes below Figures 1 & 3, (2.) the steps in the calibration procedure, (3.) the mechanical drawing for the AC version, and (4.) the orfentation of fhe Span and Adjust pots in the AC version. ‘THC-0300 Track-and-Hold Amplifier input impedance is 10’) ohms, correctly given on page 14-16, Volume TT... In Analog Dialogue 14-2, page 7, Figure 3, we show a scheme for increasing the resolution of an audio attenuator (using an AD7110) from 1.5 to .75 dB/tit, by adding a sub-ISR. To make st vork properly, an inverter should be used in the new LSE line (either furnished externally or derived using the other half of the ADG200). DIGITAL PANEL-INSTRUMENT CATALOG . . . The Digital Panel-Instrument Catalog is now updated and integrated with the DATABOOK: Volume II, Section 16. It includes Selection Guides, General Information and Definitions of Specifications, and technical data on Logic-Powered and AC-Powered Digital Panel Meters and Temperature-Measurement and Signal-Conditioning Meters PRODUCT NOTES . . . Here's an interesting angle on the ADS79 fast 10-bit a/d converter: When shortcycled for 8-bit conversion, it becomes a high-linearity converter with 1.5us max conversion time for all grades above J. If you compare its cost with that of other fast 8-bit hybrid ADCs on the market, it turns out to be extremely competitive, especially for the MIL-STD-6638 vereions. And a +12-volt-supply "2" grade version is avaflable . - . Benchnarks: MACSYM, a minicomputer-based Beasurement-and-control system, with such important features as integral process I/O statements and nultitasking, is beyond comparison with other BASIC-oriented computers simply on the strength of conventional benchmark tests (e.g-, empty do-loops, division, subroutine jumps, etc.) Nevertheless, Te acquits itself rather well in many cuch contests when compared to such computers as Apple, IBM, ‘and TRS-80. If you're interested in details, get in touch with our nearby Systems sales office « yMAC-4000 has successfully passed humidity tests according to MIL~STD-202, Method 103, Steady-State Humidity Specifications, and has successfully completed more than 11,350 hours of life testing. We also have a report from one customer of flawless operation with no failures over an 8-month period including last summer's record heat spell, for 12 systems after installation in a remote oil-field ‘application - the equivalent of 8 device-yearst PATENTS . . . The following U. S. Patents have recently been issued: 4,309,693, to Bob Craven, Solid-State Digital-to-Analog Converter . . . 4,313,083, to Barrie Gilbert, Temperature-Compensated IC Voltage Reference . . . 4,323,/95, to Peter Holloway and Doug Mercer, Bias Current Network for IC Digital-to~Analog Converters and the Like (AD558). ‘Analog Dialogue 16-2 1982 2»MAC-4000 is the software packages [including DEC's LOG uMAG-6000 isthe se pes ia ANA makes implementation easy and user measurement and software development minimal DEVICES EMAC gives you features unequaled channel clusters, yet with ts family of tain precise contol in computer or economical anaiog and digital !/O manual modes boards. is versatile enough to expand for any growing needs. ‘HMAC is optimized for the highest performance in harsh enwronments, bring- ng crucial measurement and contol to exactly where i's Name control system that's i re tro inlow-cost measurement and convo pp 7 optimized for Allthe signal conditioning is done for For the complete details on what harsh industrial you. andwithmnx and matencapaoii. | Foumecempereceralson way | environments. onion wide vanetyot sensor inpuns | measurement and contro! solution, Afulyincegiated, pre calivated intel: nctuangthnermocoupies RTDs sain | Spl lout ard mailihi coupon | gente gata wdesgnestomter gugesara’ 20m Mighrosevgecion | {Analog Recess Box 720, face eal world signals to any computer. and + 1000V isolation assure signal OE Or Or Itucique sngie Bosra design mekesit_—integity andwwatwtal umplesscrans- | a the most costetfecve soon for sma fer and ncrement/dectement your | | a Tae _ CompanyName needed-—locally or miles. Gy from your host computer. Standard RS-232C/20MA ine ‘communication makes inter facing to any computer fast and simple. And with on board intelligence, uMAC not only unburdens the host CPU, but through a powerlu! command set and available .232C/20mA For any computer, this industrial sensor I/O is better across the board.