EPA-600/4-77-025
April 1977
Environmental Monitoring Series
THE STATUS OF THE EPA LABORATORY
AUTOMATION PROJECT
Environmental Monitoring and Support Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL MONITORING series.
This series describes research conducted to develop new or improved methods
and instrumentation for the identification and quantification of environmental
pollutants at the lowest conceivably significant concentrations. It also includes
studies to determine the ambient concentrations of pollutants in the environment
and/or the variance of pollutants as a function of time or meteorological factors.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/4-77-025
April 1977
THE STATUS OF THE
EPA LABORATORY AUTOMATION PROJECT
BY
William L. Budde
Physical and Chemical Methods Branch
Environmental Monitoring and Support Laboratory
Cincinnati, Ohio ^5268
Bruce P. Almich
Computer Services and Systems Division
Office of Administration
Cincinnati, Ohio U5268
and
John M. Teuschler
Physical and Chemical Methods Branch
Environmental Monitoring and Support Laboratory
Cincinnati, Ohio ^5268
HWIRONMENTAL MONITORING AND SUPPORT LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO ^5268
-------
DISCLAIMER
This report has been reviewed by the Environmental Monitoring and Sup-
port Laboratory, U.S. Environmental Protection Agency, and approved for pub-
lication. Mention of trade names or commercial products does not constitute
endorsement or recommendation for use.
11
-------
FOREWORD
Environmental measurements are required to determine the quality of
ambient waters and the character of waste effluents. The Environmental
Monitoring and Support Laboratory-Cincinnati conducts research to:
o Develop and evaluate techniques to measure the presence and
concentration of physical, chemical, and radiological pollu-
tants in water, wastewater, bottom sediments, and solid waste.
o Investigate methods for the concentration, recovery, and
identification of viruses, bacteria and other microbiological
organisms in water; and to determine the responses of aquatic
organisms to water quality.
o Develop and operate an Agency-wide quality assurance program
to assure standardization and quality control of systems for
monitoring water and wastewater.
This report on the status of the EPA laboratory automation project was
developed by the staff of the Physical and Chemical Methods Branch, EMSL,
with the cooperation of personnel from the Cincinnati Computer Services and
Systems Division, Office of Administration. It describes the status of the
project as of March 1, 1977 and includes an outline of plans for future work
during fiscal years 1977 and 1978. This effort has been supported by the
Office of Research and Development and the Office of Planning and Management,
U.S. Environmental Protection Agency, Washington, DC 2Qi»60.
Dwight G. Ballinger, Director
Environmental Monitoring & Support Laboratory
Cincinnati
iii
-------
ABSTRACT
The status of the Environmental Protection Agency1s laboratory auto-
mation project is described in terms of currently installed systems, and
work in progress to develop and improve the system. The status report
includes a management review of the project goals, a management implementa-
tion plan, and a review of the quality control aspects of laboratory auto-
mation.
iv
-------
CONTENTS
Page
Foreword ill
Abstract iv
SECTION I - INTRODUCTION 1
SECTION II - A MANAGEMENT OVERVIEW OF THE GENERAL GOALS OF THE
PROJECT AND A MANAGEMENT IMPLEMENTATION PLAN 2
SECTION III - REVIEW OF THE QUALITY CONTROL ASPECTS OF LABORATORY
AUTOMATION 9
SECTION IV - SUMMARIES OF WORK IN PROGRESS DURING FISCAL YEAR 1977 I1*
SECTION V - AN OVERVIEW OF THE HARDWARE AND SOFTWARE OPERATIONS
OF THE INSTRUMENT AUTOMATION SYSTEM 21
SECTION VI - OVERALL PLAN FOR THE INTEGRATION OF INSTRUMENT
AUTOMATION AND LABORATORY DATA MANAGEMENT SYSTEMS 28
SECTION VII - REFERENCES TO REPORTS, SPECIFICATIONS, AND OTHER
INFORMATION 31
-------
SECTION I
INTRODUCTION
The purpose of this report is to describe the current status of the
Environmental Protection Agency's laboratory automation project. The first
phase of this project began in January of 1973 and was concluded in June of
197^- During this period feasibility studies were completed, detailed speci-
fications for instrument automation were written, an implementation design
was developed, and a contract was let for purchase of the commercial computer
hardware and the manufacturer's operating system and programming language
software. Some progress in refining designs was made during the subsequent
months, but the implementation phase actually began after delivery of the
computer hardware in December of 197^-
At the present time three systems are installed and several studies are
in progress to determine the feasibility of installing additional systems
based on similar hardware and software. Each of the following sections of
this report emphasizes and reviews different aspects of the project and each
was intended to stand alone for readers with different interests and view-
points. Readers' comments on any or all of these aspects are solicited and
will be considered as input to subsequent status reports.
-------
SECTION II
A MANAGEMENT OVERVIEW OF THE GENERAL GOALS OF THE PROJECT AND A
MANAGEMENT IMPLEMENTATION PLAN
Before 1965 it was unthinkable to put an electronic digital computer in
an analytical chemistry laboratory. Since that time the steady decreases in
the cost of computers and increases in their reliability have brought about
a revolution in analytical chemistry. Digital computers that were once in
the $100,000 to $1,000,000 class now sell for $500 to $100,000 and have far
more computational power. The analytical chemical instrumentation that has
all but completely replaced the buret and filter paper is rapidly becoming
computerized instrumentation.
Why is the digital computer invading the laboratory? As our technology
and society have become more complex the demand for more chemical analyses in
many fields has increased exponentially. Along with this increased demand
there are the requirements for more accuracy, better precision, higher sen-
sitivity, more timely results, greater selectivity, and of course all of
these at a lower cost per analysis. A good'example of this is the health
field. At one time the family physician's stethoscope was one of the few
routinely used diagnostic tools. Today a large clinical lab must do liter-
ally thousands of blood cholesterol and urine sugar analyses every day.
In the environmental field the measurement of specific air and water
pollutant chemicals is the basis for the whole environmental movement. Until
reliable measurements were made and correlated with undesirable health or
ecological effects, environmental concern was mostly limited to those con-
cerned with purely aesthetic values. Currently there is considerable emphasis
on setting standards for acceptable air and water, issuing permits for dis-
charge of wastes into rivers and oceans, monitoring these effluents to insure
compliance with permit limitations, and conducting enforcement actions when
violations occur. All of these activities are increasing the demand for
more and better chemical analyses. Better analyses embodies the ideas of
accuracy and precision and requires extensive use of analytical quality con-
trol techniques. Quality control is often deleted in analytical laboratories
because of its cost and time requirement. With this the validity of the
measurements decreases substantially. Another aspect of better analyses is
the desire for new kinds of measurements that are more revealing of the state
of environmental pollution than traditional measurements. These more reveal-
ing measurements are often more complex and simply cannot be accomplished
economically or at all without on-line computerization.
In response to these needs in EPA, a project was started in late 1972
to develop laboratory automation methods for analytical instruments that were
-------
owned "by EPA laboratories and widely used in many monitoring and research
applications. The project was a joint effort of the organizational predeces-
sor of the Environmental Monitoring and Support Laboratory (EMSL)-Cincinnati
and the Cincinnati Computer Services and Systems Division of the Office of
Planning and Management (0PM). The Lawrence Livermore Laboratory of the
Atomic Energy Commission (now Energy Research and Development Administration)
was retained as an expert consultant in systems design and development. The
goals to be achieved by laboratory automation were as follows:
(l) Improved throughput or productivity, i.e., increased sample
processing capacity and the production of more timely results
at a lower unit cost without the need for more personnel and
instruments;
(2) Improved analytical quality control;
(3) The cost-effective measurement of more meaningful parameters;
and
(M Improved management of information about samples in process
in the laboratory.
In addition to these results oriented goals, the system to be developed
was required to possess the following characteristics:
(l) Applicability. The EPA laboratory automation system was in-
tended to be applicable to a wide variety of multimedia envir-
onmental monitoring and assessment problems. Some previously
developed systems were intended for very-limited or specialized
application.
(2) Flexibility. The EPA laboratory automation system was intended
for easy modification by virtue of modular, easily duplicatable
hardware and major software elements written in readily under-
standable, self documenting high level languages. Some other
systems are literally hard wired in that it is either imposs-
ible or very costly to modify their hardware and software.
The flexible, open ended design permits the continuous up-
dating of the system with additional instrument's and allows
application in methods development research as well as the pro-
duction atmosphere. Automation in a methods research laboratory
makes possible careful testing of new procedures by allowing in-
dependent variation of a large number of method variables with
a statistically significant number of test samples.
(3) Transportability. The EPA laboratory automation system was in-
tended for easy duplication and installation in many EPA labor-
atories. In particular the designs for hardware interfaces
between the instruments and the computer and the EPA developed
software are in the public domain and may be used in any EPA
laboratory without further cost. Therefore the technology is
available to other EPA laboratories at a very significant cost
-------
and time savings. Some other systems either have proprietary
components or other aspects that preclude ready duplication.
The transportable items are described in more detail as follows:
(a) Instrument interface hardware designs; reproduction
of hardware requires funding, however, production
costs are only a fraction of design costs.
(b) A group of modular assembly language programs that
are written in Data General Nova assembly language.
These programs are assembled into relocatable binary
files that are loaded into foreground memory with
the relocatable binaries of the Basic language.
They are the real time data handlers, i.e., they
control the acquisition of data from instruments.
Since this programming is at the systems level,
it is relatively complex, and all installations will
use exactly the same programs. Selections of programs
from the group will depend on the particular instru-
ments automated in the laboratory.
(c) Several relatively modest modifications to the Basic
language program purchased from the Data General Corp-
oration. These modifications were necessary to incor-
porate the real time data handlers. These developments
are at the systems level and all installations will use
exactly the same program.
(d) A large group (over 25) of modular Basic language appli-
cations programs. These may be transferred directly
without modification, or where necessary, modules may
be modified easily to fit the operational procedures
of the particular laboratory. Again the selection of
programs from the group will depend on the particular
instruments automated in the laboratory.
(e) All documentation for all of the above. This consists
of user manuals, commented source listings, flow charts,
and engineering drawings.
(1).) Cost-effectiveness. The EPA laboratory automation system was
intended for true cost-effectiveness with special emphasis on
the above three characteristics. A major cost-effectiveness
factor is the relative ease of support, maintenance, and doc-
umentation of a number of similar systems. Some other systems
are relatively less cost-effective.
(5) Emphasis^ on Quality Control. The EPA laboratory automation
system was intended to emphasize the integration of analyti-
cal quality control concepts.
In the process of designing and implementing computer systems with the
k
-------
characteristics discussed above it has "become well established that the top-
down approach is required. This approach was employed in the EPA laboratory
automation project and consists of the following sequence of events:
(l) Development of system functional specifications.
(2) Rigorous definition of all data elements.
(3) Development of a system implementation design including
user interface specifications.
(U) Writing of program code concurrent with documentation.
(5) Installation, testing, and debugging of code followed
by user training.
(6) Post delivery maintenance.
Most individuals not experienced in computer systems are surprised to
discover that the majority of development costs for the systems that contain
the above characteristics are incurred in the rigorous specifications, data
definitions, and design phases of the project. With these designs in place,
program coding is a straightforward, non-experimental activity. Systems con-
scientiously designed in such a manner also tend to be easy to install, ac-
cept, and maintain.
MANAGEMENT IMPLEMENTATION PLAN
In order to facilitate the transfer of laboratory automation technology
to other EPA laboratories, a management plan was proposed. This plan is pre-
sented in terms of Functional Responsibilities and Definitions of Tasks.
Functional Responsibilities
The following functional responsibilities are consistent with Agency re-
sponsibilities:
The Office of Monitoring and Technical Support, Office of Research and
Development, has responsibility for budget preparation and review, coordin-
ation with 0PM and other Headquarters entities, and advocacy and support at
laboratory automation policy and planning meetings.
The Management Information and Data Systems Division (MIDSD) of the
Office of Planning and Management has responsibility for coordination, con-
trol, funding and approval of studies to determine the feasibility of trans-
porting the laboratory automation system to other EPA laboratories. In add-
ition MIDSD will approve procurements of commercially available computer
hardware and software.
The Environmental Monitoring and Support Laboratory-Cincinnati has over-
all responsibility for the implementation of laboratory automation systems,
the continued development of the system, and the long term support of the
-------
system. This includes the funding to assemble and rigorously test systems,
install systems, train EPA personnel, and provide complete documentation.
These functions will be accomplished in close cooperation with CSSD and MIDSD.
The Computer Services and Systems Division-Cincinnati (CSSD) has respon-
sibility for providing Agency-wide laboratory automation computer systems
software support. The CSSD will be the focal point for the distribution,
coordination, and installation of licensed, vendor's supplied systems soft-
ware and modifications. Included in this is software related technical
'assistance and training.
The laboratory being automated has the responsibility to provide sound,
realistic technical information to those who are directly responsible for the
system implementation. This information consists of current and projected
sample loads, operational procedures, personnel availabilities, and other
information necessary to define the functional specifications. The labora-
tory being automated has responsibility to pay for- all hardware and the com-
mercially available components of the software. The full costs of these will
be itemized in detail in the feasibility study along with the costs of poss-
ible alternatives.
After the purchase of a computer, the laboratory must provide for the
maintenance of the system. The feasibility study will also detail the annual
maintenance costs and include a description of personnel requirements to main-
tain the system and assist other personnel in its operation and use. Finally
the laboratory personnel have the responsibility to refrain from modifica-
tions to software at the systems level. Systems level modifications will be
supported from Cincinnati.
All laboratories interested in adopting the system will be considered
for feasibility studies. The results of the Office of Research and Develop-
ment laboratory evaluations will be considered if it is necessary to set
priorities for feasibility studies. No system will be installed without an
MIDSD approved feasibility study. Currently there exists an interagency
agreement between EPA and ERDA to provide the services of the staff of the
Lawrence Livermore Laboratory (LLL) for continued development and support of
the laboratory automation system. This agreement covers the period from
July 1, 1975 through June 30, 1980. The agreement is funded each fiscal year
at a level appropriate for the projects selected for work that year.
In general no instrument type will be included in an implementation until
the interface design and software have been developed, debugged, tested, and
documented by EMSL-Cineinnati, usually in cooperation with Lawrence Liver-
more Laboratory. EMSL-Cincinnati has a long term commitment to extend lab-
oratory automation methods to many other instruments.
An Agency-wide laboratory automation advisory committee will be formed
to provide periodic input of ideas and comment. This committee will meet at
least once a year to informally exchange views and technical information.
The chair of the committee will be the EPA project officer for the interagency
agreement. Headquarters coordination will be accomplished through the Office
of Research and Development's automated data processing coordinator.
-------
Definitions of Tasks
Each system implementation requires the completion of eleven distinct
tasks that are organized into six major milestones. This section contains a
description of these tasks.
Feasibility Study
1. A feasibility study and cost/benefit analysis is prepared
to determine whether technology developed in previous
years and currently installed in EPA monitoring laboratories
should be applied to an additional EPA laboratory at a diff-
erent site. Completion of this study is a major milestone,
allowing management to decide to commit funds for the next
phase.
Functional Specifications
2. A standard package of functional specifications and designs
is modified to fit the particular laboratory. After this
milestone it is possible to make valid estimates of the re-
maining cost to completion. Management can then, on the basis
of firm costs, decide to commit funds for the completion of
the project. Not until completion of the functional speci-
fications is it possible to define the cost of automation,
and management should keep open the option of discontinuing
or postponing the project.
Implement at i on
3. A specific implementation design is prepared for each project
to carry out the goals and objectives specified by Tasks 1
and 2. Designs will be based on prior designs, thereby obtain-
ing important engineering economies.
U. Purchase specifications are prepared for the items defined by
Task 3.
5. New programs are written as needed to carry out the automation,
data collection, data reduction, and sample file control. Again,
in the interest of standardization and economy, these will be
based on programs from prior implementations. Insofar as pos-
sible, programs will be directly transferred without modification.
6. Special interfaces and other hardware needed to carry out the
project are fabricated.
T. The hardware and software subsystems are assembled, tested, and
debugged. Completion of this group of tasks is the third mile-
stone .
-------
Installation
8. Equipment is shipped to the user site, reassembled, connected
to laboratory instruments, and proper operation of the com-
plete system is verified.
9- Selected users are trained to complete the debugging process.
Complete documentation of hardware and software, including
user's manuals and maintenance instructions, circuit descrip-
tions, etc. are provided. This will complete the fourth
milestone.
System Evaluation
10. The installed system is evaluated against the goals, objectives,
and specifications. This will include: a comparison of speci-
fied to achieved design objectives, an evaluation of user
reaction to the system and its documentation based on observation
and interviews, and evaluation of potential extensions to the
installed system, a cost breakdown of the actual implementation,
and a detailed evaluation of the benefits actually realized in
increased sample throughput, enhanced precision, improved quality
control, and savings in manpower.
User Assistance
11. Users are assisted in maintaining interfaces and programs and
in writing software for special needs. It is likely that pro-
prosals for new projects will be forthcoming from this work.
-------
SECTION III
REVIEW OF THE QUALITY CONTROL ASPECTS OF LABORATORY AUTOMATION
Most users of laboratory data tend, to assume that it is always of the
highest quality and therefore "beyond dispute. After all, analytical chemists
are thoroughly professional and invariably exhibit the highest possible in-
tegrity and competence. They are highly motivated, aggressive to adopt the
very latest methods, and can be trusted to put forward their very best at all
times. While these points are certainly well taken, the fact remains that
analytical chemists are also human beings, operating with often limited re-
sources, and frequently under pressure to get the results out. Therefore,
significant errors will and do occur and erroneous results could have a major
impact on important environmental or economic decisions. Clearly the problem
is one of quality control, that is, maintaining high standards for the lab-
oratory measurements.
A closely related problem is that no measurement is perfectly exact and
without some uncertainty. No matter how high the standards for the labora-
tory, this uncertainty will persist to a greater or lesser degree. Therefore
every environmental measurement recorded should be accompanied by a realistic
estimate of this error. The information required to develop this estimate
can be obtained from the quality control program of the laboratory.
In a manufacturing plant quality is rarely maintained by careful in-
spection and rigorous testing of each and every item that comes off the
assembly line. That would be too expensive! What is usually applied is a
periodic, regular or sometimes random, sampling of the product. This sample
is inspected and/or tested more completely than the balance of production,
and overall quality control is established on a statistical basis. Periodic
inspection also adds to the cost of a product, but most reputable manufac-
turers feel the cost is worth it.
The application of manufacturing quality control concepts to laboratory
analyses requires that the analytical method itself be tested and validated
periodically. Of course, this too will increase the average cost of a meas-
urement, perhaps strain the resources of the laboratory, and add to the sample
load. For many samples the investment in acquiring the sample and shipping
it to the laboratory may be several hundred dollars. It does seem worthwhile
to spend a few dollars per sample on analytical quality control (AQC). Never-
theless AQC is often deleted or reduced to a token level in the face cost and
time pressures.
One solution to this dilemma is a real time computer system that acquires
data from laboratory instruments, does all calculations including AQC statis-
tics, and has the capability of exercising some measure of control over the
-------
functioning of the instruments. Given such a system it should be possible to
generally improve the overall quality of data, make realistic estimates of the
uncertainty of measurements, and maintain reasonable costs and personnel
workloads.
Table 1 includes some of the kinds of quality control measurements that
are periodically made during a series of environmental measurements. EPA
guidelines for environmental monitoring laboratories call for 10-20% of their
effort to be devoted to quality control. The flow chart in Figure 1 shows
the sequence of events that occur during a series of laboratory measurements
with the EPA laboratory automation system. These events illustrate the inter-
active nature of the user-computer system, and explain the concept of real-
time quality control. The example is taken from the Technicon AutoAnalyzer
program.
The instrument operator starts the program and the program prints a ser-
ies of prompts at the user's terminal that request information about the an-
alysis to be performed. The operator responds to each question at the ter-
minal keyboard. Inputs include an analysis title, operator name, date, sample
numbers, standard concentrations, quality control pattern, and instrument
operating parameters. The quality control pattern is the sequence of samples,
check standards, duplicates, spikes, and blanks that will be run during the
analysis. At the conclusion of the input, a sample holder pattern is printed
to assist the operator in placing the appropriate sample, standard, etc. in
the proper place in the automatic sample changer. The operator enters a one
character command, and the analysis begins.
The first samples measured are standards to set the sample changer tim-
ing, blanks to measure the baseline, and concentration calibration standards.
After these the environmental samples and quality control samples are meas-
ured. At the end of the first quality control pattern—which may typically -
include 5-30 environmental samples, a duplicate, a spike, and a check stan-
dard—another blank is read to adjust for baseline drift. Within seconds all
computations are made, precision and accuracy are determined for the dupli-
cates and spikes, and the precision and accuracy is compared with previously
determined and stored values. The Shewhart and Cusum statistical methods are
applied and a summary quality control report is presented at the terminal for
the operator's inspection. If all is in control the analysis proceeds to the
next batch of samples already placed in the sample holder. If out of control
the operator is obliged to take appropriate action. The out of control sam-
ples will have to be rerun, but this information is known within seconds
while the samples are still in the sample holder and not several hours or
days later. If the system was in control and all samples were finished, the
program prints an operator's report that includes a complete digest of the
quality control information. If this is acceptable the operator orders, via
a terminal command, a customer report printed. This also contains quality
control information, but in a simpler and shorter format.
With the implementation of data management as described in another sec-
tion, the sample numbers may optionally be entered into the user program
from the backlog file. At the conclusion of the run, the operator will be
asked if the results are approved for release to sample file control. If the
10
-------
Procedure
Table 1
Some Analytical Quality Control Procedures
Application
Calibration Standard
Check Standard
Reagent Blank
Lab Duplicate
Field Duplicate
Lab Spike
Field Spike
Calibration standards are used to establish a
relationship between a measured phenomena and
concentration. Too few standards can lead to
erroneous conclusions about the shape of the
instrument response curve and inaccurate results.
The periodic measurement of a standard as an
unknown is a check on instrument drift, reagent
stability, and human errors.
A reagent blank contains all materials used in
an analytical measurement except the sample. A
periodic measurement of a blank is a check on
baseline drift and background signals from
reagent contamination.
A periodic lab duplicate is revealing of errors
due to instabilities in instruments, reagents,
or analytes. Lab duplicates are used to estab-
lish the precision of the method as expressed
by the standard deviation. A lab duplicate must
include a repetition of the complete analytical
procedure.
If the precision of the method is known, a
periodic field duplicate can provide information
about errors in sampling strategy or laboratory
errors in record keeping and sample dilution.
The periodic addition of a known amount of analyte
to the sample can give information about changes
in the analyte concentration due to settling,
the presence of interferences, or improper
preservation techniques.
Same as a lab spike except time begins in the
field and effects that occur during shipment and
storage are included.
11
-------
START
J
OPERATOR
INPUTS
MEASURE TIMING,
BASELINE, AND
CALIBRATION STANDARDS
MEASURE SAMPLES,
CHECK STANDARDS,
REPLICATES, SPIKES,
AND BLANKS
COMPUTE AND DISPLAY
INTERIM RESULTS AND
STATISTICS
ALERT
OPERATOR
TO
ACTION
Figure 1. Flow Chart of the Sequence of Events During a Controlled Series
of Laboratory Measurements.
12
-------
operator and the laboratory management enter the release password, the
environmental data and the quality control data will be transferred to the
data management system. The printed unconsolidated report should be signed
by the operator and retained as an extension of the laboratory notebook and
a permanent legal record of the validity of the results.
There appears to be two distinctly different types of quality control
derived from laboratory automation. One of these benefits is the ability to
process in a cost-effective way the quality control procedures indicated in
Table 1 and Figure 1. This quality control benefit is called active quality
control because it still requires some action on the part of laboratory per-
sonnel to prepare the quality control samples. These additional samples are
AQC overhead which is defined as the percentage of measurements made that do
not generate environmental data, but either provide control information to
the instrument operator or generate more realistic raw data for subsequent
calculations. These measurements are of reagent blanks, duplicates, spikes,
check standards, and calibration standards and all contribute to this over-
head:
_ Sum of Blanks + Duplicates + Spikes + Standards
— ~ ~—rrr— , A
Sum of All Measurements Made
The additional personnel effort required to process a hO% active quality
control overhead is estimated to be about ~L%. Additional instrument time is
required to process this overhead. The principal personnel time savings is
in the numerous calculations required to reduce the active quality control
data to meaningful results.
A different kind of quality control benefit from laboratory automation
is passive in the sense that no additional personnel effort is required to
receive the "benefit. It is purchased at the time of the original investment,
it is present at all times, and it does not require an additional price each
time it is used. Passive quality control is derived from the elimination of
errors that result from hand measurements of peak heights or areas, the elim-
ination of errors introduced by manual or desk calculator computations, and
the elimination of errors that result during the transfer of data from one
piece of paper to another piece of paper. Other errors eliminated are those
that result from baseline drift and assumptions of linear calibrations when
a non-linear equation is more accurate. Automation eliminates these problems
since the computer automation system acquires all peak heights or areas, does
all computations, corrects for baseline drift, calculates results based on
the best mathematical fit of the calibration data, and prints a report that
involves no hand copying of data.
13
-------
SECTION IV
SUMMARIES OF WORK IN PROGRESS DURING FISCAL YEAR 1977
SYSTEMS INSTALLED AND FEASIBILITY STUDIES
Currently EPA has three instrument automation systems installed. All
are based on the Data General model 8Uo minicomputer and use essentially the
same software:
Environmental Monitoring and Development and Test System
Support Laboratory-Cincinnati
(EMSL-Ci)
Cincinnati Computer Services Cincinnati Environmental Research
and Systems Division (CSSD) Center Production System (Supporting
MERL)
Region V Surveillance and Production System and Regional
Analysis Laboratory Test Site
Studies are in progress to determine the feasibility of installing Data
General based systems at several other sites. These studies are supported by
the Management Information and Data Systems Division (MIDSD), and monitored
by EMSL-Ci and CSSD. The status of these studies is as follows:
Region III Surveillance submitted for approval
and Analysis Laboratory
Region IV Surveillance in progress
and Analysis Laboratory
National Enforcement in progress
Investigation Center-Denver
Region VII Surveillance in progress
and Analysis Laboratory
Region II Surveillance planned for FY '77
and Analysis Laboratory
Open planned for FY '77
-------
SYSTEM SAMPLE PROCESSING CAPACITY
A number of benchmark tests of the capacity and quality control aspects
of the EPA laboratory automation system are in .progress. The Municipal En-
vironmental Research Laboratory's Waste Identification and Analysis Section is
one user of the Environmental Research Center's production system and recently
reported some capacity data that was of interest. The section has interfaced
four channels of Technicon AutoAnalyzer (TAA) and one flame atomic absorption
(AA) spectrometer with a computer controlled sample changer. The maximum num-
ber of measurements per day was determined to be 200 for the AA spectrometer
and 1600 for the four Technicon instruments. The Technicon equipment is
broken down into a dual channel TAA I usually dedicated to nitrite and nitrite
plus nitrate measurements, a single channel TAA II usually dedicated to phos-
phorus measurements, and a single channel TAA II usually dedicated to ammonia
or total Kjeldahl nitrogen measurements. All have large sample changers hold-
ing 200 samples and the second production run of the day concludes about
10:30-11:00 PM unattended.
To estimate the number of measurements possible per year in this one lab-
oratory it was assumed that maximum capacity runs would take place 50% of the
available time and that all other available time would be utilized for pre-
ventive maintenance, operator continuing education, personnel sick leave,
annual leave, and trouble shooting. The 50% capacity figures are as follows:
Instrument Measurement s/year
TAA I (2 channel) 10U,000
TAA II 52,000
TAA II 52,000
AA 26,000
23^,000
Presently four additional channels of Technicon and two AAs are being
interfaced to the computer system which is not yet at full instrument capac-
ity.
ELECTRONIC BALANCE AUTOMATION
An electronic balance was interfaced to the laboratory automation system
at the Region V laboratory, primarily for the purpose of weighing new and
loaded air filters for particulate matter determinations. This balance is
used for eight hours per day during certain periods and appears to be one of
the most successful implementations of the system. Since there are many
weighing operations involved in environmental laboratory operations, including
National Pollution Discharge Elimination System (NPDES) measurements of fil-
terable and non-filterable residues, it would appear reasonable to include an
electronic balance in a laboratory automation system. For the information of
those contemplating the purchase of an electronic balance, Region V has a
Mettler model HE20 with a model BE20 balance control and a model BA28 digital
readout. Interfacing these would be particularly cost-effective since the
interface design and software already exist.
15
-------
FLAMELESS ATOMIC ABSORPTION AUTOMATION
Recently the Perkin-Elmer model 503 atomic absorption spectrometer, which
is interfaced to the EMSL-Ci development system, was fitted with a model HGA
2100 graphite furnace and a model AS-1 automatic sample changer. A new set
of programs for application with the furnace and sample changer was developed
and these are available at no cost to anyone who has the hardware to use them.
We believe this combination is a superb example of instrument automation.
In one recent experiment 65 measurements of arsenic in the 2.5 - 100 yg/1
range were accomplished in a little over three hours, including all calibra-
tion and quality control calculations. The 65 measurements included all
standards and quality control samples. Actual operator intervention time was
only about 30 minutes for sampler and program set-up and retrieval of the
printed report. In our judgement the real advantage of these kinds of systems
is the quality of output that can be sustained over an extended period of time.
DATA GENERAL AND DIGITAL EQUIPMENT MINICOMPUTERS
The Environmental Protection Agency signed a contract during 1976 with
the Digital Equipment Corporation (DEC) for purchase of a number of PDP-11
model U5 and model TO minicomputer systems. These minicomputer systems were
intended for several applications including remote job entry to large Agency
batch processing computers, relatively small mathematical models, and small
regional databases. The terms of the DEC contract restrict direct customer
hardware interfacing with a PDP-11. The Data General minicomputer was
selected on the basis of quite different specifications including real time
data acquisition and control capabilities. However, the top of the line
PDP-11/TO will play a key role in laboratory automation according to current
studies for the implementation of the Laboratory Data Management. The role
of the PDP-11/TO in this area and the status of laboratory data management
are discussed in Section VI of this report.
ULTRAVIOLET-VISIBLE SPECTROMETER AUTOMATION
For those contemplating a laboratory automation system, one instrument
worth considering is a general purpose ultraviolet-visible spectrometer. At
EMSL-CI we chose an existing moderately priced Perkin-Elmer (Coleman) model
12it spectrometer for addition to the development system. This instrument has
about 1-2 nm resolving power and measurement accuracy sufficient for most
general purpose work. The following application programs have been developed
or are still under development, and all of these are available at no cost to
anyone who has the hardware to use them:
1. NBS performance test (completed).
2. Data acquisition and reduction for the chlorophylls a, b, c
and pheophytin a, and various chlorophyll and biomass ratios
and indexes (completed - see later note in this section).
3. A color analysis system compatible with 30U(g) requirements
(under development).
16
-------
U. A program for measurements of many NPDES analytes with manual
sample preparation. These are analytes often measured with Techni-
con systems except when only a few samples are on hand or the
samples require special treatment. All methods implemented will be
approved 30U(g) methods (under development).
EVALUATION OF TERMINALS
Keyboard/printers, keyboard/cathode ray tubes (CRT), and keyboard/CRT's
with optional hard copy are the devices (terminals) used for communication
between an instrument operator and a laboratory automation computer system.
The selection of terminals for a system is one of the more perplexing aspects
of the hardware design. The selection is a problem because there is a be-
wildering variety of terminals available and there needs to be a friendly
relationship between an instrument operator and his terminal. Since not all
operators are alike, a single terminal will not satisfy everyone. At EMSL-Ci
we have deliberately chosen to acquire and evaluate user's responses to a
variety of terminals. All terminals have a standard EIA plug and interface
and can be easily interchanged. Also all operate at a minimum of 30 char-
acters per second and therefore no ASR-33 teletypes are included. A list of
terminals in use is as follows and we will be happy to share our experiences
with anyone needing the information.
Terminal
Lear-Siegler
model ADM
Digital
Equipment
model VT-52
Texas Instru-
ments model
TOO
Hewlett-
Packard
model 26HOB
Digital
Equipment
model VT-55
Digital
model LA-36
Tektronix
model UOOO
series
Type
On EPA Standard
Contract
CRT No
CRT Yes
Printer Yes
(heat sensitive
paper)
CRT with memory No
for soft copy
of about 8 screens
CRT with optional No
hard copy (slow &
wet copy)
printer (ordinary No
paper)
CRT with optional Yes
hard copy (fast
copy)
Approximate
Price ($)
1500
2000
1700
2900
^000
1700
7000 - 13,000
Speed
char/sec
120
120
30
120
120
30
120
17
-------
TRANSFER OF DATA FROM INSTRUMENTS WITH DEDICATED DATASYSTEMS
There are many laboratory instruments available that have "built-in data
processing capability in the form of microprocessors or even minicomputers.
There is a general need to move data from these instruments to the laboratory
automation system so that it may be further reduced and/or consolidated with
other data for transfer to the sample file control system. Presently we have
under design consideration a general solution to this problem based on a
microprocessor controlled buffer concept.
A common feature within this class of instrumentation is the ability to
output the raw or partly reduced data to an output device, usually a tele-
typewriter, in a standardized code (American Standard Code for Information
Interchange, ASCII). Systems protocol inhibits the direct connection of the
dedicated system's output port to the laboratory automation system's input
port. However, if a microprocessor is placed between the two systems, the
system protocol of the dedicated and lab automation systems can be maintained.
The microprocessor buffer will be hardwired to the output device of the
dedicated system. This will allow the microprocessor to "catch" the ASCII
characters and store the outputted report in a buffer area of the micro-
processor 's memory. The buffer system will then log into the laboratory
automation system and transfer the ASCII data in its buffer memory to a disk
file on the lab system via Basic language commands. This design concept
should be applicable to many instrument types including:
1. Emission spectrometers with dedicated minicomputers.
2. GC/MS systems with dedicated minicomputers.
3. Gas chromatography data systems (small single purpose integrators
or large multi instrument systems).
k. Other systems such as scintillation counters, x-ray spectrometers,
etc.
GENERAL PURPOSE INTERFACE BUS (GPIB) AND EIA RS-232-C STANDARDS
The GPIB is a new Institute of Electrical and Electronics Engineers
standard for a digital interface for programmable instrumentation (IEEE
Standard W8-1975). The EIA RS-232-C is a widely used communications inter-
face protocol. If you are in a position to purchase any new instrument, and
the choices include one with a GPIB or EIA interface at moderate additional
cost, e.g., a few hundred dollars, it is worth serious consideration to choose
the GPIB or RS-232-C instrument if all other critical factors are equal. This
will save considerable time and expense at some later date if that instru-
ment is interfaced to a computer system.
POTENTIAL COST SAVINGS UNDERESTIMATED
The cost savings to the Agency that are possible with relatively stan-
dardized, but flexible, minicomputer systems are enormous. This is well
18
-------
documented in the instrument automation system feasibility studies and
the standard terminal procurement concept. One aspect of the cost savings
that appears understated is the cost of acceptable software documentation.
It appears in many cases that the cost of producing acceptable documentation
may actually exceed the cost of writing the software. Thus standardized
systems using similar general purpose software and the same documentation
will accrue another cost saving. Acceptable documentation is defined as:
1. Step-by-step operational instructions covering all possible
logical conditions, under both normal and most abnormal conditions.
2. System-level functional descriptions and operational flowcharts of
all programs.
3- Up to date program listings and indexed magnetic tape copies of all
programs with extensive commenting in the source code.
AUTOMATED CHLOROPHYLL ANALYSIS SYSTEM
The addition of the Perkin-Elmer (Coleman) model 12^ ultraviolet-
visible (UV-VIS) spectrometer to the EMSL-Ci pilot laboratory automation
system makes possible the automation of several analytical methods used in
the Environmental Protection Agency. The first method automated was the
chlorophyll (Chi) analysis system and several benchmark tests were conducted
to evaluate the productivity of the automated system. In this work five
replicate absorption curves were obtained from an unacidified and an acidi-
fied aliquot of a 90% aqueous acetone pigment solution containing chloro-
phylls &_, b_, c_ and pheophytin a^ using the unautomated Beckman ACTA V spectro-
meter and the automated Perkin-Elmer 12U. Records were maintained of the
time required to carry out the scans and to calculate the concentrations of
the pigments and biomass relationships in the sample. The mean, standard
deviation and relative standard deviation were calculated for the following
parameters: (l) Chi a_, Chi b_, Chi c_ and total chlorophyll (UNESCO trichro-
matic method); (2) Chi a/Chi b_ and Chi a/Chi c_ ratios; (3) (OD663 before
acidification)/(OD665 after acidification); (U) Biomass (AF¥)/Chl a_ (known
as the autotrophic index); (5) Pheophytin a_ and corrected chlorophyll a^
(monochromatic method); (6) Revised total chlorophyll (using corrected
chlorophyll ji value); (7) Revised autotrophic index (using corrected chloro-
phyll a. value); and (8) Revised Chi a/Chi b_ and Chi a/Chi c_ (using corrected
chlorophyll a. value).
The results obtained during the benchmark tests were as follows:
(l) Samples were scanned with the ACTA V; absorption spectra were recorded
on the strip chart recorder; all calculations were carried out manually on
the Wang 320 desk calculator. The time required was 288 minutes, and (2)
Samples were scanned on the automated Perkin-Elmer 12U; all calculations
were carried out on the laboratory computer. The time required was U3
minutes.
The mean chlorophyll £i values obtained from the analysis of 5 replicate
samples using the two instruments differed by less than 3%, and the relative
standard deviation of the data obtained with the automated system was less
19
-------
than 2%. We conclude that the saving in time and the expected quality assur-
ance benefits from the automated system are very significant and fully justify
the cost of the development work. Additional benchmarks are planned to
further evaluate the system with respect to the precision and accuracy of
the measurements.
20
-------
SECTION V
AM OVERVIEW OF THE HARDWARE AND SOFTWARE OPERATIONS
OF THE INSTRUMENT AUTOMATION SYSTEM
The EPA instrument automation system uses commercial computer hardware
and software. Extensions to the hardware and software were developed during
the project to complete the instrument automation. Flexibility and adapt-
ability in automating similar laboratories was incorporated into the basic
system design.
The Computer
The heart of the instrument automation system is a Data General Corp-
oration model 8UO Nova Computer or any compatible computer that uses the
same instruction set. Multiply-divide and floating point hardware are in-
cluded to enhance processing speed. A hardware memory mapping and protection
unit is used to allow two programming grounds to operate independently and
hardware protected from each other. The sixteen bit processor is equipped
with 6kK of core memory. This allows 31K of memory for the instrument
operators' multiuser Basic language programs executing in the foreground, and
approximately l8K of memory for utility programs executing in the background.
The remaining 15K of memory is dedicated to the Mapped Real Time Disk Oper-
ating System (MRDOS).
The mass storage devices are a high speed fixed head disk for system
overlays and swapping files, a slower moving head disk for fast access to
program and data files, and a 9 track magnetic tape for system dumps, off-
line storage, and interlaboratory distribution of programs. A medium speed
line printer is available for outputting final reports and program listings.
Figure 2 is a block diagram of the current EMSL-Ci pilot development
system. Figure 3 is a block diagram of the CSSD system that currently sup-
ports the Municipal Environmental Research Laboratory (MERL)'. The Region V
system has an essentially identical computer configuration, but a slightly
different mix of instruments.
User communication with the automation system is accomplished with any
type of terminal, i.e., printer or cathode ray tube (CRT) that has a stan-
dard EIA RS-232-C interface. The foreground and background terminals are
used for system type operations, while the instrument operators execute
applications programs from terminals at the instrument sites. Since the
system is a multiuser system, all users appear to have simultaneous access
to the Basic language capabilities through the terminal multiplexer.
21
-------
JARREL ASH
EMISSION
SPECTROMETER
PERKIN ELMER
A.A. WITH
SAMPLE CHANGER
f
\
VARIAN
A.A. WITH
SAMPLE CHANGER
f
*\
/
PERKIN ELMER
124 U.V./VIS
S\_
MA BELL LINK
MODEM
AUTO/ANS
CBS -20
PATCH
PANEL
TECHNICON
AUTOANALYZER
(2 channels)
TECHNICON
AUTOANALYZER
(2 channels)
DIGITAL
I/O
ro
COMPUTER LINK
FOREGROUND
TERMINAL
BACKGROUND
TERMINAL
PAPER TAPE
READER
SPORT
MUX
SLA
INTERFACE
REAL TIME
CLOCK
MEMORY
MANAGEMENT
UNIT
BECKMAN
TOC
A/D
CONVERTER
MBS
INTERFACE
A/D
INTERFACE
HARDWARE
MULTI/DIV
HARDWARE
FLOATING
POINT
64K
840
NOVA
CPU
512 K
FIXED HEAD
DISK
2.5 MBYTE
CARTRIDGE
DISKS
Figure 2. A Block Diagram or the EMSL-Cincinnati Pilot Laboratory Automation System
-------
ro
u>
,CSSD:
Waste Water
AUTOMATIC
SAMPLE
CHANGER
DATA GENERIAL NOVA
840 MINICOMPUTER
• 64K WORDS CORE
• FLOATING POINT
ARITHMETIC
• REAL TIME CLOCK
8-USER BASIC
TERMINAL MULTIPLEXOR
|_ 1
| OTY 0-7 |
f 'J 1 IMft 1 t IK.
V£s/ DRIVE J800 b.pj. DK
^ 1 PAPER TAPE 1
1 READER 1
— - TO LINE PRINTER
fsYCHRONOUS
COMMUNICATIONS
1 INTERFACE
1 500-9600 bps
1
5ITAL
/O
/
A
CALCOMP
PLOTTER
INTERFACE
DIGITAL
I/O
A/D1—-
I I
?£. 303
DUAL BEAM
FLAME
A. A.
REF
SWITCH
SAMPLE REF
| TERM
P.E. 303
DUAL BEAM
FURNACE
AJV.
SAMPLE
J J
ninrrAi i/n
OTYfl'-
Room 464 OTY3
iWaste
I Water:
INTERCOM
•0
- 1
FG/BG
TO' 1
TELECOMMUNICATIONS TO PLOTTER
NETWORK
ONS TO PLOTTER I
A/D
CONVERTER
lA/DO-N?
O- h
DIGITAL I/O
MASTER CONSOLE
Rom308
I OTY
, 2
LOGIN
A/D 3
OTY5
>upply: | SWITCH
P.E. 403
FLAME
A.A.
— A/D 9
— •- A/D 10
II II
P£. 403
FURNACE
A.A.
J
___
^Water Supply:
TECHNICON
AUTOANALYZER
NH3
A/D S
A/D 6
AUTOMATIC
SAMPLE
CHANGER
A/D 7
TERM
OTY 7-*.
I UII / -•-
n
DUAL
TECHNICON
AUTOANALYZER
NO2-NO3
II
z DUAL
TECHNICON
AUTOANALYZER
F- CN
TERM
Room 679
OTY6-
Room 671
Figure 3. A Block Diagram of the CSSD Production System.
-------
Instruments that generate analog signals are interfaced through a multi-
plexed analog to digital converter. Instruments with digital output or
instruments that are controlled with digital signals are interfaced through a
commercially available (MDB Systems) digital interface card. A more complete
description of the computer and interface hardware is in preparation and will
be published in a future report.
OPERATING SYSTEM AND SOFTWARE
The overall operation of the laboratory instrument automation system is
controlled by Data General's mapped real-time disk operating system (MRDOS).
This occupies about 15,000 words of core memory with overlays of the high
speed fixed head disk. The MRDOS handles all input and output from the
standard peripherals (disks, tapes, printer, etc.) and controls priorities
among the various tasks running in the system.
The high level interpreter language, extended Dartmouth Basic, is used
for the instrument data acquisition, data reduction, and data report appli-
cations (user) programs. This powerful but easy to learn language was speci-
fied for the user applications programs in order to assure that control of
computational methods and report formats would reside with the laboratory
science professionals. Basic has powerful computational capabilities, yet
it is very easy to learn and is widely taught in high schools and colleges.
It does not require detailed knowledge of computer systems and hardware, and
a professional programmer is not required for day to day laboratory operations.
The inevitable and frequent small program changes may be made in seconds by
laboratory personnel from the instrument terminals. The Basic language was
the only high level language, available on a mini-computer that permitted
some users to run data acquisition programs while other users, simultaneously,
modified existing programs or wrote new applications programs. This same
level of flexibility allows the convenient addition of additional applications
programs to the system.
The Basic language is somewhat less efficient in its use of computer
resources (memory, etc.) and it runs slower than a compiled language (Fortran)
or a machine language program. However, the advantages of Basic in increasing
overall laboratory personnel productivity far outweigh its disadvantages. The
execution speed of Data General's extended Basic is sufficient to allow rapid
response times at the user terminals. The vast majority of laboratory in-
struments have a sufficiently slow data rate and pose no serious problem.
The multiuser Basic language interpreter and all user programs run in
31,000 words of core memory called foreground. The remaining core memory
is assigned to the background. Foreground programs always have a higher
priority than background programs and input from instruments has the highest
priority of all tasks. Background programs run during intervals of time when
no foreground tasks require attention.
In the foreground, the Basic language interpreter and its extensions
occupy about 17,000 words of core memory with overlays on the high speed
disk. The extensions to Basic include a series of machine (assembly) lan-
guage subroutines that handle the critically timed acquisition of data from
2k
-------
instruments and the digital control functions. Applications programs for each
instrument in the Basic language make calls to the machine language subrou-
tines whenever data points are acquired from an instrument or a control
function is executed.
The Basic language applications programs vary in size and are stored on
the cartridge disk. When a user runs a program it is loaded into the re-
maining 13,000 words of foreground user memory. Several users may run pro-
grams simultaneously, and as long as this memory is adequate, all user pro-
grams are core resident. As soon as the combined size of the user programs
exceeds the 13,000 words of memory, swapping to the high speed fixed head
disk begins. Each user program is allocated a few milliseconds of run time,
and then the program is written to the swapping disk. At the end of another
few milliseconds the user program is returned to core memory for another
brief period of run time. This round robin swapping continues for as long as
the combined requirements of the users programs exceeds the 13,000 words of
foreground core memory. While the user program is swapped out of core mem-
ory, no instrument data is lost because the machine language data acquisition
and control subroutines are always core resident. The instrument user is
generally unaware of swapping as it is too fast to be noticeable.
LABORATORY INSTRUMENTS
The complement of instruments at each laboratory automation site was
intended to be flexible and allow for easy addition of new instrumentation.
Figures 2 and 3 show the current Cincinnati complements, and Table 2 gives
more information about these and the Region V system. Table 2 also lists
the maximum supported number of each type of instrument in a given system.
The maximum number of concurrent users is not exactly established, but prob-
ably is in the range of 12-16. For each instrument implemented, one or more
applications programs have been developed, tested, and largely debugged.
It should be noted that in a few selected applications of analytical
instrumentation, manufacturers have found it either necessary or profitable
to develop mini-computer systems dedicated to very specific instruments.
These applications include:
* Gas chromatography with conventional detectors
* Gas chromatography-mass spectrometry
* Nuclear magnetic resonance
* X-ray analysis
* Radiochemical analysis
The manufacturer's development costs (many millions in some cases) for
these are spread among the large number of units sold. Therefore, it was
more cost effective and technically feasible to use the commercial "stand
alone" systems than to interface the respective instruments to the laboratory
automation system. However, the laboratory automation system was designed
25
-------
Table 2
Maximum Supported Number of Laboratory Instruments of Each
Type on the EPA Laboratory Automation System
Types of Instruments
Technicon AutoAnalyzer I (single,
dual, triple, and 6 channel)
Technicon AutoAnalyzer II (single,
dual, triple, and 6 channel)
Maximum Supported
Number of Each
12 channels
12-n channels
(n = the number of
TAA I channels)
8
k
Partial Control for Technicon
Sample Changers
Perkin-Elmer, Instrumentation Lab,
etc. double beam flame atomic
absorption (AA) spectrometers
Varian single beam flame atomic h
absorption spectrometers
Lawrence Livermore Laboratory Computer 8
driven Sample Changers for flame
atomic absorption spectrometers
All types of flameless high temperature U
furnace atomic absorption spectrometers
Partial control for Perkin-Elmer flameless
AA high temperature furnace sample changers
Jarrel-Ash Analog or Digital Interfaced
Emission Spectrometer
Ultraviolet-Visible, Infrared, etc. wave-
length scanning spectrometers
Beckman type total organic carbon analyzers
Partial Control for total organic carbon 2
sample changers
Electronic Balances 2
Instruments with dedicated mini-computer 8
or micro-processors
Location Where
at Least One is
Implemented
MERL
Region V
EMSL-Ci
MERL
None
Region V
EMSL-Ci
MERL
EMSL-Ci
Region V
EMSL-Ci
MERL
Region V
EMSL-Ci
MERL
EMSL-Ci
EMSL-Ci
EMSL-Ci
EMSL-Ci
MERL
None
Region V
None
26
-------
for versatility, and plans for interfacing these systems were described in
Section IV of this report. This will allow an expansion of the capabilities
of the dedicated computers and permit output of comprehensive reports.
All instrument hardware interface designs are in the public domain and
are documented in the engineering drawings and parts lists. These inter-
faces may be readily reproduced on contract to organizations that specialize
in the manufacture of printed circuit board modules.
27
-------
SECTION VI
OVERALL PLAN FOR THE INTEGRATION OF INSTRUMENT
AUTOMATION AND LABORATORY DATA MANAGEMENT SYSTEMS
For the last three years there has "been a great deal of discussion and
study of a management information system for EPA laboratories. The major
factor that has impeded the development of this system has been the failure
to develop a set of sufficiently clear and detailed written functional spec-
ifications that are broadly acceptable within the Agency. A new set of
specifications is now available which integrated previous studies and added
much new information. If there is sufficient concurrence with the Agency,
a system based on these specifications could be implemented within the next
year.
In the meantime, several hardware environments have been considered for
the implementation of the laboratory data management capability:
(l) One of the large Agency-wide batch/time sharing computer systems.
The Data General instrument automation system and the large com-
puter system would communicate by remote job entry emulator soft-
ware operating in the background of the Data General minicomputer.
The system would be supported by laboratory and/or program (reg-
ional) office data processing staffs.
(2) The background of the Data General instrument automation computer
with support by the laboratory staff.
(3) A second Data General computer in the laboratory with communication
to the instrument system through a shared disk memory. The system
would be supported by the laboratory staff.
(U) A (regional) program office PDP-11/TO computer. The Data General
instrument automation system and the PDP-11/TO would communicate
by remote job entry emulator software and the system would be sup-
ported by laboratory and office data processing staffs.
After a lengthy period of study, including benchmark testing of several
alternatives, all but option k have been eliminated as viable implementation
possibilities. Option (l) was eliminated because of the slow turnaround and
high costs of the large Agency systems. Option (2) was eliminated because
the single Data General system does not have the capability to concurrently
and reliably process instrument inputs and laboratory data management.'
Option (3) was eliminated because of the unavailability of appropriate soft-
ware and utilities for laboratory data management. The INFOS software pack-
age and other general purpose software offered by Data General was judged to
28
-------
require a great deal of additional development in order to be applicable to
the EPA laboratory data management system. Also there vas a general lack of
capability to maintain data management type software in the EPA laboratories.
Option (h) was considered especially attractive because of the Agency
decision to adopt Digital Equipment Corporation PDP-11 computers for de-
centralized processing. This choice for laboratory data management would
provide functional independence of instrument automation and data management
systems. Reliability would be enhanced in the sense that hardware or software
malfunctions in one system would not affect the other system. Also program
or regional office data processing staffs have the general capability to
support data management software, and it would be easier to support if it
were integrated into the single data management package planned for EPA office
minicomputers. This alternative is also consistent with the fact that more
laboratories may require laboratory data management than instrument auto-
mation. Therefore, many laboratories would not have access to an instrumenta-
tion computer for data management, but could use the program office system
directly for this purpose.
Figure U shows a block diagram of a hierarchical distributed computer
system that includes on-line laboratory data acquisition, laboratory data
management, and links to large Agency batch processing computers with access
to national databases. The hierarchical network is the basis for current
planning of integrated instrument automation and laboratory data management
systems.
29
-------
LARGE AGENCY WIDE COMPUTERS
•BATCH JOB PROCESSING
"NATIONAL DATA BASES
MASSIVE ON/OFF LINE
STORAGE
PRINTER
1200 LPM
MAGNETIC
TAPE
n
DATA TRANSFER
HIGH SPEED HASP
PROGRAM (REGIONAL) OFFICE
PDP-11/70
* RJE EMULATORS
* LABORATORY
DATA MANAGEMENT
PRINTER
1300 LPM
MAGNETIC
TAPE
DATA TRANSFER
HIGH SPEED
HASP
TERMINAL
MULTIPLEXER
16 TERMINALS
USERS
LABORATORY
DATA GENERAL
NOVA INSTRUCTION SET
DIGITAL
INPUT/OUTPUT
(MICROPROCESSOR)
TURNKEY
MULTIUSER
MINICOMPUTER
SYSTEM
25-90
MBYTE
DISK
TERMINAL
MULTIPLEXER
8-16 TERMINALS
USERS
ANALOG TO
DIGITAL
CONVERTER
LABORATORY
ANALOG
INSTRUMENTS
TURNKEY
SINGLE
USER
MINICOMPUTER
SYSTEM
INTERNAL
MICRO-PROCESSOR
DIGITAL
OUTPUT
TTTTTII
INSTRUMENTS
I
INSTRUMENT
!
INSTRUMENT
r~
INSTRUMENT
Figure k. Hierarchial Computer Network for On-Line Laboratory Data
Acquisition Data Management.
30
-------
SECTION VII
REFERENCES TO REPORTS, SPECIFICATIONS, AND OTHER INFORMATION
An On-Line Real-Time Multi-User Laboratory Automation System, W. L.
Budde, E. J. Nime, and J. M. Teuschler, Proceedings of the first EPA Office
of Research and Development ADP Workshop, October 2-U, 197^, Bethany College,
Bethany, WV, p. 10**.
Improved Analytical Quality Assurance from Laboratory Automation,
W. L. Budde and J. M. Teuschler, Proceedings of the International Conference
on Environmental Sensing and Assessment, September lH-19, 1975, Las Vegas,
NV, Vol. 2, p. 36-6.
A Flexible Laboratory Automation System for an EPA Monitoring Laboratory,
B. P. Almich, Proceedings of the second EPA Office of Research and Develop-
ment ADP Workshop, November 11-13, 1975, Gulf Breeze, FL, p. 10.
Laboratory Data Management, W. L. Budde, Proceedings of the second EPA
Office of Research and Development ADP Workshop, November 11-13, 1975, Gulf
Breeze, FL, p. U2.
Technicon AutoAnalyzers: Functional Description, R. W. Crawford and
G. W. Barton, Lawrence Livermore Laboratory Report No. UCRL-520U6, April 1,
1976.
Total Organic Carbon Analyzer: Functional Description, R. W. Crawford
and L. P. Rigdon, Lawrence Livermore Laboratory Report No. UCRL-520U5,
April 1, 1976. -
Sample File Controller: Functional Description, R. W. Crawford and
H. S. Ames, Lawrence Livermore Laboratory Report No. UCRL-520U7, April 1,
1976.
Laboratory Data Management for the Environmental Protection Agency,
H. S. Ames, Lawrence Livermore Laboratory Report No. UCRL-5205^, April 15,
1976.
Atomic Absorption Instrument Functional Description, R. I. Bystroff
and W. G. Boyle, Lawrence Livermore Laboratory Report No. UCRL-52065,
April 26, 1976.
A Feasibility Study for the Computerized Automation of the Annapolis
Field Office of EPA Region III, H. S. Ames, G. W. Barton, R. I Bystroff,
R. W. Crawford, A. M. Kray, and M. D. Maples, Lawrence Livermore Laboratory
Report No. UCRL-52052, August 1976.
The Sample File Controller, Functional Description, Auerbach Associates,
Inc., March 7, 1977-
31
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/4-77-025
3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
The Status of the EPA Laboratory Automation
Project
5. REPORT DATE
April 1977 issuing date
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
William L. Budde, Bruce P. Almich
and John M. Teuschler
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Environmental Monitoring and Support Lab.
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 1*5268
- Gin., OH
10. PROGRAM ELEMENT NO.
1BD612A
11. CONTRACT/GRANT NO.
12. SPONSORING AGENCY NAME AND ADDRESS
Same as above
13. TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY CODE
EPA/600/06
15. SUPPLEMENTARY NOTES
16. ABSTRACT
The status of the Environmental Protection Agency's laboratory automation
project is described in terms of currently installed systems, and work in
progress to develop and improve the system. The status report includes a
management review of the project goals, a management implementation plan, and
a review of the quality control aspects of laboratory automation.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Computers
Scientific computers
Small scale computers
Laboratories
Research laboratories
Chemical laboratories
Quality Assurance
Laboratory automation
09-B
18. DISTRIBUTION STATEMENT
Release unlimited
19. SECURITY CLASS (ThisReport)
21. NO. OF PAGES
38
20. SECURITY CLASS (Thispage)
22. PRICE
EPA Form 2220-1 (9-73)
32 #U.S. GOVERNMENT PRINTING OFFICE: 1977-757-056/5559 Region No. 5-11
------- |