LITERATURE CITED
( 1 ) Bard, A. J., Solon E., ANAL.C H E x 34, 1181 (1962). ( 2 ) Barker, G. C., “Transactions of thc
Symposiuni on Ele8:trode Processes,” E. Yeager, ed., pp. 336-44, Wiley, New York! 1961. ( 3 ) Barnartt, S., J . Ekctrochem. Soc. 108, 102 (1961). (2) Bewick, A , , BewicE., A , , Fleischmann, M.,Liler, M., Elect,.ochim. ilcta 1, 53 (1959). (5) Booman, G . Holbrook, W. R., A X A L . CHEN. 31, 10 (19591. ( 6 ) Boonian, G. L., Holbrook, W. B., Rein, J. E., Ibid., 29, 219 (1957). ( 7 ) Booman, (2. L , Morgan, E., Crittenden, A. I,, J . .-Inz. [‘hem. SOC.78, 5533 (1956). (S) Brown, 11. H., U.S. Atomic Energy Conini. Rcpt. IDO-16852, hIay 1963. (9) Chestnut, H I h l a l e r , R. JT ., “Servo-
mechanisms and Iiegulating Systeni Ilesign,” 1-01, 1, 2nll ed., n-iley, Sew York, 1959.
(10) Cockbaine, 11. H . , Atoniic Energy Research Establishment (Gt. Brit.), AERE-EL/R-1528, ?vIay 5 , 1955. ( 1 1 ) DeFord, 11. D. 133rd Meeting,
ACS, San Francisco, Calif., April 15, 1958; and unpublished research. (12) Delahay, P., “New Instrumental hfethods in Electrochemistry,” pp. 391-3, Interscience, New York, 1954. (13) Dynatron Electronirs Corp., 178 Herricks Road, Mineola, K. Y., Bitilctin 126, 1962. (14) Geo. A. Philbrick Rese:trclies, Inc., 127 Clarendon St., Boston 16, Mass., “The Lightning Empiricist,” p. 3, Issue No. 6, October 1958. (15) Ibid., “GAP/R .4pplivation Brief,” I). 1, Ko. D2, April 1, 1960. (16) Ibid., S;. 6140’ AB-261-1, p. 3 , Feb. 1, 1961. (17) Gerischer, H., Staubach, IC., Elektmchern. 61, 789 (1957). (18) ,Grabbe, E. AI., Rarno, S., TVooldndge, D. E., eds., “Handbook of Automation. ComDiitation and Control,” Vol. 1, “Conhi1 Fundnntentals,” Wiley, S e w l-ork, 1958. 1191 Hager. C. I C . Elctlronzcs 32. 44 ‘ (KO.33, Septeniber 4, 1959). (20) Harrar, J. E . , University of California, Lawrence Radiation Laboratory, Livermore, Calif., private communication, 1962. (21) Kelley, 11. T., ,Jones, H. C., Fislier, I). J., ASAL. CHEM.31, 485, 956 (1959).
( 2 2 ) Kuo, S. C., “Automatic Control Systems,” Prentice-Hall, Englewood Cliffs, N. J., 1962. ( 2 3 ) Shults, \Y. D., Dunlap, L. B., ASAL. CHEM.35,921 (1963). (24) Spolin, P., Hedett-Packnrd Journcd 14, Ko. ,556, pp. 5-8, Jan.-Feb., 1963. (2.5) Tekt,ronis, Inc., Reaverton, Oregon,
“Introduction to Operational Ampli-
fiew,” Service Scope S o . 19,.~ iluril 196:i (26) T‘alley, G. E,, Jr., Wallninn, Ii;! eds., “J’acuuin Tube Amplifiers. Massachusetts Institute of Teclinolopy Radiation IAoratory Series, T:ol. 18, LLIcGraw-Hill, S e w York, 1948. ( 2 T ) JTeber, .T. H., Spnce/Aeronaulics, p. 134. Snvemher 1058. (28) IVill, F. G., Z. Elektrochent. 63, 184 (19.50). (29) Wise, E. S . , L l s a ~CHEN. . 34, l l S l
(1962).
( 3 0 ) Zittel, H. IC., I)unlap, L. 13., Ibid.,
35, 125 (1963).
RECEIVED for review I1eceniber 26, l%2. Accepted Akugust 29, 1963. Division of Analytical Chemistry, 144th >Ieeting, ACS, I m Angeles, Calif., April 1963. Work supported by U.S.Atomic; Energy Commission under contract no. AT( 10-1)205 through Idaho Operations Ofice.
A Digital Readout Device for Analog integrators E. CLIFFORD TOREN, Jr. and CHARLES P. DRISCOLL Department of Chemktry, Duke University, Durham,
b A circuit, capable of direct digital readout of analog integrators within a relative error and relative standard deviation of .tO.lc%, uses an operational amplifier to reset the integrator a t preset voltages. The unit may b e programmed to read equivalents directly. h A L O G I P U ’ T C G R A I ~ O Rl ~i a ~ e
iecently found extensir iipplication in as analytical instrumentation--e.g., c~mlonieters. Relatiw standard deviations of = t 0 . 0 5 ~ ofor integration of electrolysis curieiits have been reported ( 1 ) . -1 readout device as accurate as the integrator is necessaiy to achieve the maximum capabilities of the integrator. Rooman ( 1 ) has u-ed manual and recording potentioineters. Kelley, Jones, and Fisher (4, 5 ) have further utilized digital voltmeters. I n thi. work a Ion cost, simple, accurate, and direct readout device is developed to avoid th. inconvenience of manual potentiometers and the high cost of recorders and digital voltmeters. This curcuit is based cm the principles of a n instrument origin tted by Higinbot h a m and Rankowita (S), but is extended to use commercially available operational amplifiers 3
CIRCUIT DESCRIPTION AND OPERATION
Figure 1illustrates the complete circuit for the digital readout unit and block
N. C.
diagrams of aisociated modules necessary for circuit operation. The Integrator module is a conventional circuit used by DeFord ( 2 ) , but modified to switch to various values and types of capacitor, C. The integrators deslcribed by Booman (1) and by Kelley, Jones, and Fisher (4, 5 ) could also be used esentially without modification. The Voltmeter module used by DeFord ( 2 ) has a range of 10 mv. t o 10 volts fullscale which i i variable in steps by means of a panel snitch. d potentiometer in parallel with the ha+ IO-mv. movement and external .hunt is w e d to perform minor calibration adjustments of *lo% of the iiominal fullscale setting of all ranges. The Discriminator voltage is obtained from the Secondary Standard Voltage module designed by DeFord ( 2 ) . This unit, consisting of a Dekapot (Electro/ Scientific Industries, iLIodel DP-211) in series with O.lyoprecision resi>tors, divides the +300-volt calibrated power supply output (George A. Philbrick Researches, Inc., &Iode1 R-100B) with a linearity of 0.1%. Operational amplifier, OA, without feedback is a voltage crossing detector ( 7 ) used as the sensing element in the digital readout unit. 08 is balanced in the Zero position as shown in Figure 1 by adjusting the Discriminator voltage input such that zero voltage is obtained at pin 6. By applying E , volts, in addition to the bias 1 oltage necessary to balance O A , the output of 0-1 rapidly jumps from -50 to +50 .iolts for a K2-IT7 amplifier or from -100 to
+lo0 volt.. for a C/lOO/B, when the Integrator output attains a value of E , 1 0 003 volt. This jump is sufficient to bring T’, normally held below cutoff, into saturation to activate the relay. Ppoii closure of the relay, C is discharged through R4 to avoid burning the relay contact, and the mechanical register is advanced 1 count. The fraction of the last count is read on the x oltmeter, the fullscale value of rshich is adju-ted to E , volts. A convenient value for coulometric applications i* 9.65 volt> for direct readout in eyuivalcnt-1snitch, SIV> is provided for rmcr+ ing the input terminals of Orl when the polarity of the Integrator output is changed, and for balancing O24. The switch positions +Ill-T and --I?VT refer to the polarity of the signal applied to the Integrator input. voltage divider network of R2 and Ra is used to ensure that VI remains cutoff during balancing or warmup time. CI is placed in parallel with VI to hold the relay clobed for sufficient time t o activate the mechanical register. CI need not exceed 1 pf. a. determined by experiment. Thi. provides time of ea. 5 milliseconds to activate the couiiter and to discharge C almo.;t completely-Le., five or more time constants, R4 X C seconds--since C has a maximum value of 10 pf. The abqolute performance of the circuit i- checked wing the high quality, polystyrene dielectric, integrating capacitor, and the digital readout circuit The capacitors are atljuted in the InteVOL. 35, NO. 12, NOVEMBER 1963
1809
grator circuit such that the relative error of the output voltage does not exceed O.l%, as measured with a Leeds and h’orthrup Type K potentiometer in the integration of constant, known input voltages.
I
i
RESULTS AND DISCUSSION
The performance of the circuit is illustrated in Table I for Ej = 10.000 volts. The constant, K , depends only on the fixed circuit parameters of input voltage, E,; number of counts (including fraction of last count), n; and the total integration time, T . For the intryration of a constant voltage, the Integrator output voltagc, e,, is eo =
-E,t/RC
T cI I DISC.
-3OOV
IN
INT. OUT IN
.-
r--------l
(1)
A count is regislerrd when e, = El, and occurs a t t = t,, the time of a single count. Therefore Equation 1 becomes Ef
=
-E,L/RC
INTE B RATOR L_________--_______
(2)
and for n counts during the total time, T , without regard to sign, Equation 2 becomes _n_ ~- __ - K
(3)
EiT - EjRC
K , therefore, indicateb the accuracy of t h e unit as well as the linearity with respect to Ei. Table I compares tlie calculated values, K = l/EjRC. and the observed values, K&. = n/E:,T. Each value of Kobe, measured in quadruplicate, is found to have a relative standard deviation of =tO.l%. The values of Kob. agree within =t0.101, relative standard deviation in a given column for a given polarity of E,. This indicates a linearity of 0.1% for a given set of circuit parameters. The discrepancy between the positive and negative input values is explained by the h3-mv. input transition interval mentionrd above. Switching from + I X T t o -1NT reverses the input of OA to change from one end to the other of the 6 m v . interval. By reducing the discriminator voltage hy 6 mv. in the f I N T position. this error is eliminated, and Kab. becomes independent of polarity. At count rates less thnn 0.02 count/
Figure 1.
-
Wiring diagram for digital readout unit
C: Polystyrene adjustable capacitors, Southern Electronics types PC105D2E or PC106D2E 1 and 10 pf., 2 0 0 wvdc. or 1 pf., Mylor CI: 1 hf.,ZOO wvdc. L,: Relay coil, Guardian Electric Mfg. Co. type 121 5-05 OA: Amplifier, Embree type C/1 OO/B or Philbrick type K 2 - W Mech. register: General Controls type CE40BS402 R: Precision resistor selector switch ( 2 ) R1: lOOK 1 2 W Rz: 10 M,’l/2 W Ra: 1 M , 1 / 2 W RI: 100 ohm, 5 W Relay: See 11 above SW: Rotary switch, 3-pole, 3-position, nonshorting VI: 12AU7-A
second, the dispersion of the results increases owing t o inherent integrator drifts (6). Cnder these conditions Ei is too small for tlie corresponding value of RC, hence t h e integrator is not used efficiently. At count rates greater than 0.2 count/second, the relay closure time becomes an appreciable fraction of the time for one count-Le., 5 milliseconds as compared to 5 seconrlswhich causes errors greater than 0.1%. The use of relatively incyensive Mylar capacitors in the integrator circuit does not affect, the linearity or reproducibility of the results. However, discrepancies between K and Kohswithin the *IO% tolerance of the
capacitors are observed. This presents no practical difficulties since E, can be adjusted to calibrate the integratorreadout combination. ACKNOWLEDGMENT
hluch of the construction was done by John It7.Yarbrough. LITERATURE CITED
1) Rooman, G. L., ANAL. CHEAI.29,
’
D. J., ANAL.CHEM.31,488 (1959). (5) Ibid., p. 956. (6) Malmstadt, H. V., Enke, C. G., Toren, E. C., “Electronics for Scientists,” Ch. 8, W. A. Benjamin, Inc., New York, 1962. ( 7 ) George A. Philbrick Researches, Inc., Boston, Mass. “Applications Manual for Philbrick &tal Plug-in Computing Amplifiers,” 1956. ’
Table 1.
RC = 0.1000, K = 1.000 Ei Kobi +o. 1000 1.005 +O.O5OO I ,006 +0.0250 -0.1000 0.0500 0.0250
-
1 8 10
a
1.009
0.999
1.000
1.004
Results for Digital Readout Unit
nc
K Ei
= 1.000, = 0.1000
++ O1.0000 . 5000
+o. 2500
- 1.0000 - 0.5000 -0.2500
ANALYTICAL CHEMISTRY
Kotm
RC = 10.00, K = 0.01000 Ei Kobs 10.000 0.01008
0.1007
++5.000
0,1009 0.1000 0.1000 0.1001
+2.,500 - 10.000 -5.000
0,1008
-2.500
0.01009
0.10008 0.10000 0.10001 0.10001
RECEIVEDfor review June 17, 1963. Accepted August 8, 1963. Division of Analytical Chemistry, 144th Meeting, ACS, Los Angeles, Calif., April 1963. This work was supported in part by a grant from the Dreyfus Foundation and the Duke University Research Council.
