A minicomputer-controlled data acquisition system

Historic, archived document

Do not assume content reflects current
scientific knowledge, policies, or practices.

A S^)
f\7UU(j(^

A Minicomputer-controlled
Data Acquisition System

U.S. Department of Agriculture

Science and Education Administration

Advances in Agricultural Technology • AAT-W-1/August 1978

with this issue, the Western Region of the Science and Education Adminis-
tration (SEA) , U.S. Department of Agriculture , begins publication of a new
series of reports. Advances in Agricultural Technology. This series will con-
tain reports of technology that has been tested and is ready for application .
Examples are plans and specifications for a piece of equipment, new or improved
laboratory techniques, mathematical models, and studies resulting in practical
suggestions for improved operations.

Four U.S. Department of Agriculture agencies — Agricultural Research Service
(ARS) , Cooperative State Research Service (CSRS) , Extension Service (ES) , and
the National Agricultural Library (NAL) — have been merged to form a new orga-
nization, the Science and Education Administration (SEA) , U.S. Department of
Agriculture . The five regional series of ARS, the CSRS series, the ES Circular ,
and the NAL Library List have been discontinued . If you subscribed to any of
these discontinued series, you will continue to receive similar material from
SEA as new series are established . Subscribers wishing to modify or stop mail-
ings coming to them should write the office whose address appears below. Please
include the mailing label from this publication with your request.

Science and Education Administration, Advances in Agricultural Technology,
Western Series, No. 1, August 1978

Published by the Office of the Regional Administrator for Federal Research
(Western Region), Science and Education Administration, U.S. Department of
Agriculture, Berkeley, Calif. 94705.

ABSTRACT

A minicomputer-controlled data system was developed for measuring a wide
variety of plant and weather variables in the field. One to 26 sensors were
scanned at each of one to eight portable sensor stands placed in field plots.
Analog signals from thermocouples, a dewpoint meter, radiation sensing devices,
and a hot-wire anemometer were measured with a digital voltmeter. Digital data
were accumulated from a cup anemometer, an automatic evaporation pan, and a
beta-ray scaler. Data were processed by the computer, and the dimensioned val-
ues were recorded on punched paper tape and in a printed form. The software
was written in absolute assembler language. The program included routines that
would set the initial conditions and control the read, convert, and write oper-
ations; drive the peripheral equipment; provide mathematical functions; and
handle interrupt signals. A multipoint recorder furnished graphic data from
selected sensors.

KEYWORDS: Leaf thermometer, sensor stand, automatic evaporation pan,
stepping switch driver.

ACKNOWLEDGMENTS

The author received valuable assistance in setting up and calibrating the
beta absorption equipment from W. R. Ehrler, plant physiologist.

Trade names and the names of commercial companies are used
in this publication solely to provide specific information.
Mention of a trade name or manufacturer does not constitute
a guarantee or warranty of the product by the U.S. Depart-
ment of Agriculture nor an endorsement by the Department
over other products not mentioned.

CONTENTS

Page

System requirements 1

Computer control of operation.., 1

Sensor selection 2

Data processing 2

Data logging 2

System hardware 2

Computer and peripheral equipment 2

Multipoint recorder 4

Auxiliary power 4

Data reduction 4

Analog parameter measurements 5

Temperature measurements 5

Radiation measurements 7

Hot-wire anemometers 7

Digital parameter measurements 7

Cup anemometer 8

Automatic evaporation pan 8

Beta absorption 8

System operation 8

Discussion 9

Appendix A 9

Software program for data acquisition system 9

Key to subroutine labels in figure 4 10

Appendix B — single-step driver for stepping switch. . . 12

Appendix C 13

Crossbar switch 13

Interfacing peripheral equipment 13

A copy of this publication is available from the Western Cotton
Research Laboratory, 4135 East Broadway Road, Phoenix, Ariz. 85040.

A MINICOMPUTER-CONTROLLED DATA ACQUISITION SYSTEM

By K. E. Fry^

SYSTEM REQUIREMENTS

Methods of monitoring plant and microclimate variables in field crops are
continuously needed.^ ^ Present solid-state systems offer reasonable flexibil-
ity for interfacing with a wide variety of electrical sensors; however, as the
demand for flexibility increases, so does the cost of the equipment. As a re-
sult of the recent advancements in microprocessor integrated circuits and rea-
sonably simple combinations with peripheral input-output devices, experimenters
can assemble highly flexible systems inexpensively.

Although the data acquisition system described here was designed around a
minicomputer (Hewlett-Packard Co., Model 2114A, 8K x 16 bit memory locations),
the principal organization of the hardware and software (computer programs) is
applicable to available microprocessors in combination with adequate random-
access-memories (RAM) and input-output peripheral equipment.

The follov7ing requirements were incorporated into the total system to col-
lect plant, microclimate, and weather data:

Computer Control of Operation

Operate the system with computer program (see Appendix A) , which is composed
of a control loop that selects program subroutines to scan sensors, convert
data, and record final data for further analysis.

Sound a buzzer to warn field help before scans are started.

Operate multiplexing and measuring devices under program control in the
field plots.

Place the system in a wait mode during periods of power failure.

^Plant physiologist. Western Cotton Research Laboratory, Science and Educa-
tion Administration, Phoenix, Ariz.

^Boving, P. A., and Winterfeld, R. G. A mobile recording and monitoring
weather tower. U.S. Dept. Agr., Agr. Res. Serv. ARS W-41, 13 pp. 1976.

^Smith, J. W. , Stadelbacher , E. A., and Gantt, C. W. A mobile unit for
automatic collection of computer-compatible macroclimatic and microclimatic field
data. U.S. Dept. Agr., Agr. Res. Serv. ARS S-136, 6 pp. 1976.

1

Sensor Selection

. Scan 1 to 26 analog sensors at each of one to eight portable sensor stands
placed in field plots.

. Record rapidly changing plant and weather parameters at predetermined in-
tervals (1 to 30 min) .

. Record slowly changing parameters (for example, soil temperatures) at
hourly intervals.

. Accumulate and record pulse counts for preset time periods.

. Be able to change the pattern of scan under program control or manually
without halting operations.

Data Processing

Convert the sensor voltage readings or pulse counts into the desired dimen-
sion by using linear, polynomial, or exponential regression equations.

Utilize multiple sensors and multiple regression equations to calculate the
final dimension, when it is influenced by more than one weather or plant
variable.

Data Logging

. Print Julian day, hour, and minute.

. Print channel code number and corresponding data in block format for immed-
iate observation.

. Record the above information of punched paper tape in an appropriate format
for further processing.

. Print statements indicating the channel and stand number when a malfunction
of equipment occurred.

. Print the time when line power was lost and when it returned.

SYSTEM HARDWARE

Computer and Peripheral Equipment

The components of the data acquisition system are shown in figure 1. Only
signal lines are shown connecting the components. The plant and microclimate
sensors were mounted on field sensor stands (fig. 2), which were usually located
near the center of the experimental plots. A 26-position stepping switch at
each stand performed the necessary multiplexing of the sensor signals to a com-
mon signal line. Appendix B describes a single-step driver, which was used to
operate a stepping switch under computer control. Signal, control, and direct
current power lines connect the stands to the signal processing equipment, which
was located in an air-conditioned trailer.

2

TC's

ANALOG SENSORS

AIR Solar

TEMPERATURE SOIL RADIATION NET

LEAVES P.A.R.

HUMIDITY
DEW POINT

HOT-WIRE

ANEMOMETER

Figure 1. — Relation of system hardware components to minicomputer.

Figure 2. — Sensor stand support for air-temperature manifold and
sensor tubes, stepping switch, and terminal board.

3

In the trailer, a crossbar switch (see Crossbar Switch, Appendix C) multi-
plexed the signals from the different stepping switches at the sensor stands
and also signals from individual sensors located elsewhere. The signals were
measured by a digital voltmeter (DVM, Hewlett-Packard Co., Model 2401C) , were
converted to the desired dimensions in the computer, and, by means of a tele-
typewriter (Teletype Corp., Model ASR-35) , were logged on punched paper tape
and as printed data. Each system component was operated by the computer under
program control (see Interfacing Peripheral Equipment, Appendix C) . A software
clock determined the preselected time to sound the warning buzzer and start a
scan of the sensors on line. The code numbers of the channels to be scanned
were read from paper tape by the photo tape-reader (Hewlett-Packard Co., Model
2753) and were stored in computer memory. Digital counters (four-decade) accum-
ulated the numbers of electrical pulses received from digital-output sensors
over preselected time periods. The counts were read and logged by the computer.

Multipoint Recorder

Graphic recordings of selected weather and plant parameters were obtained
with a multipoint recorder (Leeds and Northrup, Model H, 1 to 12 channels).
Under computer control, a 12-pole relay rerouted the signals from the selected
sensors between the scans.

Auxiliary Power

Regulated and uninterruptible alternating current power was supplied to
the computer and digital voltmeter through a charger-inverter system connected
to seven 12-volt automobile batteries (Elgar Corp., Model UPSlOOl) . In the event
of line power failure, the data acquisition system was held in a wait mode; the
software clock continued to run. Normal operations were resumed if the line
power returned within 45 min; if not, the total system was shut down orderly.

Data Reduction

The field data were printed on folded teletypewriter paper (8.5 by 11
inches) , which was conveniently bound in pressboard binders for a quick refer-
ence; however, studies of the weather and plant variables, including hourly
averages of air and leaf temperature, required further data reduction.

During 1976, a small office computer (Wang Laboratories, Inc., Model 2200S,
32K X 8-bit registers) became available. Computer programs in BASIC were written
to read the paper tapes (Photo tape-reader, Wang Model 2203), select specific
data, perform statistical treatments, and list and graph the reduced field pa-
rameters (Plotter, Wang Model 2212).

4

ANALOG PARAMETER MEASUREMENTS

Temperature Measurements

The choice of sensors for the air, soil, and leaf temperatures was copper-
constantan thermocouples for the following reasons: (1) Thermocouples were in-
expensive and easily constructed, (2) sensors for a wide number of uses could
be fabricated from commercially available wire with diameters from 1.4 to 0.012
mm (Omega Engineering Inc.), (3) only a single calibration was needed regardless
of sensor size for all thermocouples made from the same materials, and (4) inex-
pensive hand-held meters were available for use with thermocouples.

Air Temperature

At each sensor stand (fig. 2), air temperature was measured at the soil
surface, at the height in the densest canopy, and about 1 m higher. Sensors
were mounted inside a horizontal polyvinyl tube (1/2-inch diameter, 30 cm long) .
An outer tube (1-inch diameter) surrounded the sensor tube, and both were wrapped
with reflective aluminized polyester tape (Scotch No. 850). Holes were drilled
near the base of the inner tube to allow airflow between the tubes. All joints
were lubricated with silicone grease to prevent sticking during adjustments and
disassembly. The meter-long manifold (l^^-inch diameter) supported two sensor
tubes and a connection to the third sensor tube through a flexible hair -dryer
hose. Aspiration velocities were 508, 330, and 152 cm sec“^ for the upper, cen-
ter, and lower sensor tubes, respectively, when obtained with a small blower
(Dayton, No. 40012) attached to the manifold with its exhaust point upward.

Soil Temperature

Copper -constantan thermocouples were inserted to depths of 1, 5, 10, 20, 40,
and 100 cm below the soil surface of the cotton rows. The junctions were insu-
lated with several coats of vinyl glue.

Leaf Temperature

A clip-on type of leaf thermometer was developed to obtain maximum conven-
ience in attaching the thermometer to the leaf, minimum influence on the leaf’s
temperature and movement, and adequate reliability in monitoring leaf surface
temperature over extended periods. Figure 3 shows a thermometer clamped to a
cotton leaf. A copper-constantan junction bead (about 0.1 mm in diameter, welded
from 25 y wire) was epoxy-glued to the tip of a toothbrush bristle (0.38 mm in
diameter) and was held against the leaf surface with a force of about 1 g. The
coiled spring and wire arm that holds the bristle was formed from 0. 5-mm-diameter
phosphorbronze wire. Heavier spring wire (0.8 mm in diameter) was used for the
holding loops and arms that were soldered to the jaws of a miniature alligator
clip (Mueller, No. 30C) . The clamping force of the spring-loaded clip was suf-
ficient to hold the thermometer in place in winds of 670 cm sec”^ (15 miles per

5

Figure 3. — Thermocouple leaf thermometer on cotton leaf (underside).

hour). For use on cotton leaves, holding-loops were 2.7 cm in diameter and the
arms were 3 cm long. Flexible thermocouple lead wires (127 y in diameter, 40 cm
long. Teflon insulation. Omega Engineering Inc.) were threaded through polyvinyl
tubing (1.3 mm outside diameter) and connected to heavier lead wires (0.5 mm in
diameter) .

The leaf thermometers were tested for accuracy in measuring surface temper-
atures by clamping them to wet blotter paper. Their temperature readings were
compared with those from fine-wire thermocouples embedded in the blotter. The
leaf thermometers were in error about 0.16° Celsius above the blotter temperature
for every degree Celsius the ambient air temperature was higher than the blotter
temperature.

Possible leaf damage and its influence on the leaf temperature at the point
of measurement were tested in the following manner. Thermometers were allowed
to remain on field cotton leaves for 5 days. Then, during a period of relatively
stable leaf temperature and rapid transpiration, the leaf temperatures were re-
corded before and after moving the thermometer bead about 5 mm. Usually, no
significant mean difference in temperatures was observed. Indicating that the
1-g force of the bead against the leaf tissue did not greatly influence the
transpiration rate and the resulting leaf temperature.

6

Dewpoint Temperature

Air was sampled at the temperature sensor tubes and pumped at the sensor
stands (not shown in fig. 2) to a solenoid gas multiplexer (three channels) in
the trailer. Dewpoint temperatures were measured in a thermoelectric dewpoint
hygrometer (Cambridge Systems, Model 880) . During stormy weather, condensation
in the gas lines invalidated the readings.

Radiation Measurements

Solar Radiation

At an instrument shelter, the solar radiation was monitored with a Kipp and
Zonen solar imeter. Because the solar imeter was used at high field temperatures,
it was calibrated against a temperature-compensated Epply Black and White pyra-
nometer. Recorded body temperatures, measured on the north side of the Kipp and
Zonen solarimeter, were included in the multiple regression analysis along with
the radiation measurements. The resulting multiple linear equation was used for
the calibration.

Net Radiation

At each sensor stand, net radiation was measured 1 m above the crop and 1 m
south of the stand with Fr itschen-type net radiometers. Calibrations from the
manufacturer were used.

Photosynthetic Active Radiation (PAR)

At the instrument shelter, PAR was measured with a quantum sensor (Lambda
Instruments Co., Inc., Model LI-190S) . The manufacturer's calibration was used.

Hot-Wire Anemometers

Mu It iple-j unction, hot-wire anemometers (Hastings-Radist Inc., Probe Type
N-7B) were used to measure the horizontal air movement in the canopy. Unfortu-
nately, the alternating current in the lead wires often induced in the main sig-
nal wires a noise component. The level of this noise was reduced by cabling the
anemometer lead wires separately.

DIGITAL PARAMETER MEASUREMENTS

For each 15-min interval, the electrical pulses from sensors were counted
by means of solid-state, four-decade counters. At scan time, the count of each
unit was sampled in sequence and was logged by the computer. Then the count reg-
isters were set to zero and the counters were started again. Three devices were
operating currently.

7

Cup Anemometer

A three-cup anemometer^ was used to integrate windspeed over 15-min inter-
vals and at 1 m above the crop. For every revolution of the cups, a pulse was
generated and counted. The calibration of the anemometer was based on the data
of Fritschen.^

Automatic Evaporation Pan

A standard (type A) evaporation pan was placed near the center of the field.
The water level was regulated at 17.8 cm by automatically adding water at 15-min
intervals. Chlorinated water was supplied from an elevated fiberglass tank (ca-
pacity 946 liters) at the edge of the field. The pan water level was controlled
at a small shelter 2 m east of the pan. A reservoir in the shelter was connected
to the pan with a garden hose so that pan water level could be sensed in the
shelter with a Styrofoam float. Attached to the float was permanent magnet,
which controlled a 6-v solenoid valve in the waterline by means of a reed switch.
A tipping -bucket mechanism, taken from a rain gage, triggered an electric pulse
to the four-decade counter for every 8 ml of water added to the pan. At 15-min
intervals, the count was logged and the counter was reset.

Beta Absorption

Relative leaf water content (RLWC) was estimated from beta absorption.^
Counts from a Geiger Muller detector were scaled for 10 min (Hamner Electronics,
Inc., Scaler Model NS-11) . The count was recorded by the computer, and the
scaler was reset at 15-min intervals. The high voltage to, and the signal from,
each detector were switched consecutively among four detectors by means of high-
voltage reed switches at the field plot. In this manner, the thickness of each
of four leaves was measured once every hour. The relation of RLWC to leaf
thickness was determined in experimental leaves at the end of a run.

SYSTEM OPERATION

The data acquisition system has been in operation since the summer of 1975,
primarily for field measurements. Modifications to the hardware and software
were made as needed to accomodate different kinds of sensors. The maximum number
of channels on line at one time was 125, which were distributed among six step-
ping switches.

^Fritschen, L. J. A sensitive cup-type anemometer. J. Appl. Meteorol.

6: 695-698. 1976.

^See footnote 4.

^Nakayama, F. S., and Ehrler, W. L. Beta-ray technique for measuring leaf
water content changes and moisture status of plants. Plant Physiol. 39: 95-98.

1964.

8

The system was used during winter months in the laboratory for logging leaf
weight losses. Ten balances were constructed in the laboratory, using miniature
strain gages. ^ For each of 10 detached leaves, 10 measurements were taken every
4 minutes for the first hour and every 10 minutes for the next 2 hours. In this
operation, a modified computer program was assembled from selected subroutines
(Appendix A). The subroutines, labeled LOOP, TBG, and PRINT, were modified to
suit the specific task.

DISCUSSION

Analog stripchart recorders log data that are immediately observable, yet
the number of channels is severely limited. Digital recording systems will han-
dle large numbers of data channels, but small changes in any parameter are diffi-
cult to observe unless the digitized data are graphed. Our data collection sys-
tem includes some of each advantage.

The greatest advantage of the system is its flexibility and ability to col-
lect data from a wide variety of sensors, to scan different sensors at different
intervals, to utilize multiple sensors and mathematical equations to obtain the
final dimensions, and to control both the field multiplexers and the sensing de-
vices during a measurement.

The main disadvantages of the system result from operation of the sophisti-
cated equipment, which is necessary for operation. The builder-operator must
have knowledge of digital electronics, assembler language programming, and sen-
sor characteristics. Changes made in the method of operation generally require
additional digital circuitry and computer program modifications; however, higher
levels of sophistication can be obtained when innovations of circuitry and pro-
gramming become available.

APPENDIX A

Software Program for Data Acquisition System

The computer program is written in assembler language for the Hewlett-Packard
2100 series minicomputers.® To reduce memory requirements for the total opera-
tional program, absolute (nonrelocatable) expressions were used. The program
occupied about 2,000 memory locations (16 bit). For the storage of data from
200 sensors, about 1,500 additional locations were needed. The subroutine sym-
bols and their linkage in the program are shown in figure 4, which is followed
by a key to the mission of each subroutine. Four more conversions subroutines
may be added without changing the format of the channel codes. In the program
listing (not shown) , the steps are well documented and are assembled in octal-
based integers. A current program listing is available from the author.

^Idle, D. B. An electrical weight transducer designed for the measurement
of changes of weight of detached leaves. Ann. Bot. 40: 473-477. 1976.

^Anonymous. A Pocket Guide to Hewlett-Packard Computers. Hewlett-Packard
Co., Cupertino, Calif., 626 pp., illus. 1972.

9

TIMIN

LEADR

TMOUT

BADSS

4 — » DVM.C

TBG

> INITIALIZE

START

READ

I

ssw

CODIN

CONTROL

LOOP

I

PROUT

WAIT

PR. ON

START

PRINT

^

1

f

BCD.V

MODE

j (LIBRARY)

4

VLT.

V.DEG

RADS

T.DIF

! i I.

1.7 L.

JUMP SUBROUTINE
UNCONDITIONAL JUMP

PTIME

PCODE

FPOUT

—f

INTGR

Figure 4. — Subroutines for computer program and their linkage.

Key to Subroutine Labels in Figure 4

INITIALIZE sets the initial conditions for the peripheral equipment.

START clears the storage arrays and resets the pointers.

TIMIN inputs the Julian day, hour, and minute from the keyboard.

CODIN inputs from paper tape the specific channels to be read, the correspond-
ing conversion subroutine code, and the DVM integration time (0.1 or 1.0
sec). The information is stored in array CODE.

LOOP tests for different operational commands that are manually or program con-
trolled. When present, they cause a jump to the corresponding subroutines.

LEADR punches 127 nulls to obtain leader and trailers on data tapes.

TMOUT is called manually to print the current Julian day and time.

READ inputs the channel code, sets the crossbar switch to the requested stepping
switch line, calls the stepping switch subroutine (SSW), and initiates the
DVM to start a reading.

SSW operates the requested stepping switch to find the requested switch level
(see Appendix B) .

10

BADSS prints the channel number at which a malfunction of the stepping switch
occurred .

BCD.V converts voltage data from array BCD (see DVM.C below) to a floating-
point binary (packed), checks for a DVM overload, and stores the data in
array VOLT.

MODE selects a conversion subroutine according to the channel code, gets the
data from array VOLT, and stores the converted data in array DATA.

VLT. transfers voltage data without conversion.

V.DEG converts voltage data to degrees Celsius using a quadratic regression
equation for copper-constantan thermocouples.

RADS converts voltage to langleys per minute. The temperature of the solari-

meter is recalled from the previous reading and is used in a multiple linear
regression equation. Coefficients for the equation were determined experi-
mentally.

T.DIF converts voltage data to degrees Celsius using a quadratic regression

equation for chromel-constantan-copper thermocouples, which were wired for
differential temperature measurements.

PRINT controls the output of day, time, code, and data. After an end statement
is printed, the program control is transferred to START.

PTIME outputs the Julian day, hour, and minute at which the scanning started.

PCODE outputs the channel code in decimal integers.

FPOUT outputs a floating-point number. When the number of integers exceeds 6,

$$ is printed.

INTGR outputs an integer (decimal) that was stored as an unpacked binary number
(manufacturer’s software).

LIBRARY includes selected mathematical subroutines (furnished by the manufac-
turer) that will: (1) convert a binary number of 16 bits to a packed
floating-point number of 32 bits; (2) double-store or double-load floating-
point numbers; and (3) add, subtract, multiply, and divide integers and
floating-point numbers.

BITIN (not shown in fig. 4) is called by READ. It inputs data from the counters
and scaler, and stores them in array BIT.

BTOUT (not shown in fig. 4) is called by PRINT. It outputs data from array BIT.

INTERRUPT SUBROUTINES An interrupt is a signal to the computer from an external
source, which immediately forces the control of the program to a specific
subroutine. At the end of the subroutine, the control is transferred back
to the next step in the program at which it was interrupted. The subroutines
are listed with decreasing priorities.

11

TBG A signal from the time-base-generator (every 10 sec) causes a jump to sub-
routine TBG, in which the software clock is updated. At preselected times,
signals are set to sound the warning buzzer, start a scan, scan the hourly
readings, punch a leader -trailer , and start a new page on the printer.

DVM.C When the DVM has finished a reading, it causes a jump to subroutine DVM.C,
in which the binary coded data (4-2 '-2-1) is input from the DVM and is
stored in array BCD.

PROUT When the auxiliary power source has signaled that the line power is off,
the control jumps to PROUT. The time of day is saved, and the control is
locked in a wait loop. Peripheral equipment is turned off, but the soft-
ware clock continues to operate.

PR. ON A return of line power causes a jump to PR. ON. Peripheral equipment is
turned on, the times of day that the power went off and came back on are
printed, a leader and new page are obtained, and the program control is
transferred to START.

GRND. TRUE
STEPPING

PULSE 24 VDC

Figure 5. — Schematic diagram of stepping-switch (Type 45), single-step driver.

APPENDIX B

Single-Step Driver for Stepping Switch

The circuit shown in figure 5 causes the stepping switch (GTE Automatic
Electric Co., Type 45, 26 switch positions, 6 to 12 banks) to advance one step

12

for every voltage drop at the pulse input terminal. The quad two-input nand
gate (SN7400 or MC846) generates a positive pulse that triggers the silicon
controlled rectifier (SCR) to energize the magnet motor. Commutation of the
SCR is obtained at the interrupter switch on the stepper. The pulse timing is
controlled at the computer, which was programmed to send out 8 pulses per second.

The 26 positions of the stepping switch are wired (five banks) to send to
the computer the binary equivalent for each position. In operation, the computer
tests the binary signal for a requested position. Stepping is commanded until
the requested position is reached. When more than 76 stepping commands (three
rotations of the wiper assembly) are sent and the requested position is not
reached, a malfunction statement is printed.

For three summers, these drivers have operated satisfactorily without atten-
tion except for an annual cleaning of the interrupter contacts on the stepping
switch. A single powerline carrying 24 volts of direct current (vd-c) was sup-
plied to the drivers. The 5 vd-c power required by the nand gate was furnished
through a 7805 voltage regulator, which was mounted on a small heat sink (not
shown in fig. 5). An Insulated metal dust cover was placed over the switch (not
shown in fig. 2).

APPENDIX C
Crossbar Switch

The crossbar switch was a relay switching matrix containing 20 gold-plated
switches, each having six levels. It was an experimental model (Cunningham
Corp.) acquired from government surplus. The relay driving circuits were de-
signed and wired by us so that the switches would operate under computer control
through an interface card.

Interfacing Peripheral Equipment

The peripheral equipment was connected to the computer through interface
circuit cards, which were supplied by the manufacturer. These cards permitted
the exchange of commands and data signals. Fifty channels were made available
for control of the crossbar and stepping switches and to read digital data from
the digital counters. One card (not an interface card) furnished crystal-con-
trolled time pulses that were counted in the software clock. We wrote the soft-
ware to drive all the peripheral equipment.

13

GPO 795-867/396

g^l;'

PHn>.:» i'JWt'Ui*' If. "vW't aig|.&.t%:|^' 'li- } . '

'■v..*a*|i;;, ’ '.tf^^' " •rtHf-' f, • ;■ V ' .ft .•• ''•*»' •'. »

'A

J/iC-

.M' ^

'■ ; :r, '.' '"'JJ 'jy's

;--4t '».v . ■*■ j\ ,f'i)^«-»}

f i{'\ - ^ ;, .i.**l ■ V?:*)

’. -I * ' v"

> \ ■' 2 ..

••-■' , ^ ■ , 1'^ ^ . ■'■

c 3»0h- ^wifeWM ■ : > ' * *

i>'"' ■ . ' “'L'^)^'^- ■■ '^'' '' a ti if

"i/ s ■ '■* •S‘2^:i 'isi',. Ii4'. M fe'!''’’- «^^." U'tn '*•■- ^ Ujiii(|l¥».> ''7

i'ti' ^j\'Wi''-'- «^^." u'l'n ^

fi»:t^'^ii},i <•»,,, f Ka • 'HWiS^iBl

I

U. S. DEPARTMENT OF AGRICULTURE
SCIENCE and EDUCATION ADMINISTRATION
WESTERN REGION
2850 TELEGRAPH AVENUE
BERKELEY, CALIFORNIA 94705

OFFICIAL BUSINESS

PENALTY FOR PRIVATE USE, $300

POSTAGE AND FEES PAID
U. S. DEPARTMENT OF
AGRICULTURE

AGR 101