LIBRARY

Michigan State
University

 

This is to certify that the

thesis entitled

CHARACTERIZATION AND OPTIMIZATION
OF A SILICON VIDICON SPECTROMETER
THROUGH REAL-TIME MICROPROCESSOR CONTROL

presented by

James Edward Hornshuh

has been accepted towards fulfillment
of the requirements for .

 

 

Ph .D. degree in Chemistry

Major professor

Date ’9 78/

0-7 639

 

CHARACTERIZATION AND OPTIMIZATION OF A SILICON VIDICON
SPECTROMETER THROUGH REAL-TIME MICROPROCESSOR CONTROL

BY

James Edward Hornshuh

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Chemistry

1978

(57d

63/ // g

1
'o-A

ABSTRACT

CHARACTERIZATION AND OPTIMIZATION OF A SILICON VIDICON
SPECTROMETER THROUGH REAL-TIME MICROPROCESSOR CONTROL

By

James Edward Hornshuh

A multiprocessor—controlled silicon vidicon spectrometer is
described. The spectrometer incorporates a silicon—target vidicon
with a fused-silica window to extend the usable instrument response
into the near ultraviolet region. The spectrometer can monitor a 230
nm window in the range of 250 nm to 1000 nm with 1.0 nm resolution.
Under computer control the wavelength window can be divided into
between 16 and 4096 wavelength channels. The tube readout beam can
be deflected to any channel at random or made to scan them sequen-
tially at frequencies between 10 kHz and 1 MHz. To increase the
target's integration time and enhance weak signals, the readout beam
can be inhibited for a user-specified time interval.

The real—time signal-to-noise ratio (S/N) of the system is 3086.
Signal averaging can be used to increase the S/N at the expense of
time resolution. For a fixed data acquisition time, a combination of
signal averaging and charge integration can be used to optimize the
output response. The detector responds linearly to the incident light
level over 4 orders of magnitude. The observed non-linearity of the
vidicon is 1% at 90% of saturation.

Real-time control of the vidicon spectrometer is provided by a

general purpose multiprocessor system based upon the Intel 8080A

James Edward Hornshuh
microprocessor. The development of this system not only illustrates
the increased system capabilities of a multiprocessor system over
those of a single processor system, but also provides design guidelines
for an efficient and modular multiprocessor system. Specific examples
of task allocation, efficient use of system resources, and utilization
of the inherent bandwidth of available microprocessor static memory are
provided in the documentation of the computer system used to control
the vidicon spectrometer.

The vidicon spectrometer was applied to three basic types of dual-
wavelength spectroscopy; dual-wavelength non-scanning spectroscopy,
dual-wavelength scanning spectroscopy, and derivative spectroscopy.
These dual-wavelength techniques are easily implemented with the
vidicon spectrometer which, because of its multi-wavelength capability,
requires only changes in the data processing software to achieve the
various techniques. Realized advantages of the dual-wavelength mode
of operation included continuous correction for scattering and sample
settling and the elimination of the influence of an interfering
component (Co+2) in the analysis of a highly overlapping two-component

mixture (Co+2 and Ni+2 in 2N HClOu).

To Jean, Mom, and Dad

ii

ACKNOWLEDGMENTS

I wish to offer my sincere appreciation to Professor C.G. Enke
for his guidance, encouragement, and friendship during the course of
my research at Michigan State University.

Thanks are given to Professor S.R. Crouch for serving as my
second reader and for many helpful discussions and comments.

I wish to thank the MSU Chemistry Department's Electronics Shop
and Machine Shop for the excellent support they offered. In addition,
I would like to thank Marty Rabb, MSU Chemistry Department Electronics
Designer, for helpful discussions on a wide range of electronic
problems.

I am very appreciative of financial support provided by Michigan
State University and the Office of Naval Research.

I thank my fellow group members in Professor Enke's research
group for their friendship, interaction, and assistance. I especially
thank Tom Last, Erik Carlson, and Spyros Hourdakis.

To my parents I am grateful for their interest in my studies and
their unfailing encouragement and moral support.

Mbst importantly, I thank my darling wife Jean for her love and
support. She has encouraged me throughout my graduate education,
worked full-time to support our family, typed the initial and final

draft of this thesis, and helped me to proofread the final copy.

iii

TABLE OF CONTENTS

LIST OF TABLES . . .
LIST OF FIGURES.
CHAPTER 1 - INTRODUCTION .

CHAPTER 2 - OVERVIEW OF THE SILICON VIDICON TUBE
IN ANALYTICAL SPECTROSCOPY .

A. Introduction.
B. Operation of the Silicon Vidicon Tube .
C. Recent Developments in Vidicon Tube Technology.
1. Silicon—Target Vidicon .
2. Focus-Projection-Scanning Vidicon.
3. Pyroelectric Vidicon .
4. X-Ray-Sensitive Vidicon.

5. Piezoelectric-Target Vidicon .

D. Spectroscopic Applications of the Silicon Vidicon Tube.

CHAPTER 3 - OPERATION OF THE SILICON VIDICON SPECTROMETER.
A. Introduction.
B. Real—Time Control and Data Acquisition.
C. Post—Processing of Spectral Information . . . . . .

CHAPTER 4 - CHARACTERIZATION AND OPTIMIZATION OF THE

SILICON VIDICON SPECTROMETER . . . . . . .
A. Introduction. . . . . . . . . . . . . . . . . . .
B. Multichannel Spectral Response. . . . . . .

iv

Page
viii

ix

12
12
13
14

14
15

16
22
22
22

33

41
41

42

C. Wavelength Linearity and Response . . . . . . . . . . . 47

1. Linearity of the Wavelength Deflection

Circuit Elements . . . . . . . . . . . . . . . . . 47
2. Limiting Resolution of the Monochromator . . . . . 43
3. Limiting Resolution of the Vidicon Detector. . . . 53

D. Analytical Signal . . . . . . . . . . . . . . . . . . . 56

1. Vidicon Detector Operating Voltages. . . . . . . . 58
2. Response Near Saturation . . . . . . . . . . . . . 60
3. Linearity and Range. . . . . . . . . . . . . . . . 62
4. Signal Enhancement with Charge Integration . . . . 64
E. Electronic Background and Random Noise. . . . . . . . . 65

CHAPTER 5 - DUAL-WAVELENGTH SPECTROSCOPY . . . . . . . . . . . . 76
A. Introduction to Dual-Wavelength Spectroscopy. . . . . . 76

B. Dual-Wavelength Applications with the Vidicon
Spectrometer. . . . . . . . . . . . . . . . . . . . . . 79

1. Introduction . . . . . . . . . . . . . . . . . . . 79

2. The Dual-Wavelength Non-Scanning Mode. . . . . . . 80

3. The Dual—Wavelength Scanning Mode. . . . . . . . . 88

4. The Derivative Mode. . . . . . . . . . . . . . . . 90

C. Conclusions . . . . . . . . . . . . . . . . . . . . . . 99
CHAPTER 6 - COMPUTER SYSTEM ARCHITECTURE . . . . . . . . . . . . 100
A. Introduction. . . . . . . . . . . . . . . . . . . . . . 100

B. Development of a Multiprocessor System. . . . . . . . . 101

1. Dedicated Minicomputer System. . . . . . . . . . . 101

2. Distributed System with Real-Time
Microprocessor Control . . . . . . . . . . . . . . 103

3.

4.

5.

Multiple Microprocessor System for
Real—Time contr01 o o o o o o o o o o o o

Interprocessor Communication .

Dedicated Multiprocessor Bus System. . . .

C. Multiprocessor Control of the Vidicon Spectrometer.

CHAPTER 7 - DOCUMENTATION OF SYSTEM DESIGN .

A. Introduction. . . . . . . . . . . . . . . . . . .
B. Scan Control Micro.
1. Introduction .
2. Slave Microprocessor Design.
3. Analog Scan Circuits . . . . . . . . .
4. Digital Control and Interface Circuits . . . .
5. Real-Time Clock and Interrupt System . . . . .
6. Dedicated Microprocessor for Number Processing .
7. Shutter Control Circuit. . . . . . . . . . .
C Data Acquisition System . . . . . . . . . .
1. Introduction . . . . . . . . . . . . . . . . .
2. Analog Conditioning and Sampling Circuits. . .
3. Digital Control Circuits . . . . . . . . . . .
4. Hardware Adder for Signal Averaging. . .

D. Optical Configuration . . . . . . . . . . . . . . .

APPENDIX A - Program Listings. . . . . . . . . . . . . . . .

A. 8080

CONVERS Routines . . . . . . . . . . . . . . .

B. Minicomputer Programs and Subroutines . . . . . . .

APPENDIX B - Supplementary Documentation of System Design. .

vi

Page

105
108
110
114
119
119
119
119
121
124
130
135
137
139
141
141
141
143
148
150
151
151
187

213

Page
APPENDIX C - Front Panel Documentation . . . . . . . . . . . . . 222

REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . 230

vii

LIST OF TABLES

Table Page
3-1. CONVERS Software for Operation of the Vidicon

Spectrometer. . . . . . . . . . . . . . . . . . . . . . . . 25
3-2. Operator Dialogue Mnemonics . . . . . . . . . . . . . . . . 29
4-1. Channel Response for a Line Source. . . . . . . . . . . . . 44

4—2. Wavelength Linearity. . . . . . . . . . . . . . . . . . . . 49
4-3. Vidicon Linearity and Range . . . . . . . . . . . . . . . . 64
4-4. Signal Enhancement with Charge Integration. . . . . . . . . 66
5-1. Dual-Wavelength Correction for Scattering . . . . . . . . . 82

5-2. Dual-Wavelength Correction for Scattering and
Sample Settling . . . . . . . . . . . . . . . . . . . . . . 83

5-3. Determination of Nickel in the Presence of Cobalt . . . . . 86
7-1. Analog Power Supply Requirements. . . . . . . . . . . . . . 126
B-1. Peripheral Addresses for Vidicon Micro (octal). . . . . . . 213
B-2. Data Acquisition Addresses for Vidicon Micro (octal). . . . 214
B-3. Vidicon Control Addresses for Scan Micro (octal). . . . . . 216
B-4. Backplane Connections for Vidicon Slave Micro . . . . . . . 218
B-S. Backplane Connections for Fast Adder Board. . . . . . . . . 219
B-6. Backplane Connections for Memory Board. . . . . . . . . . . 220
B-7. 3M Connector Pin Assignments. . . . . . . . . . . . . . . . 221

C-1. Keyboard Output Code. . . . . . . . . . . . . . . . . . . . 224

viii

Figure

2-5.
3-1.

3-2.

LIST OF FIGURES

Schematic of a Silicon-Target Vidicon. . . . . . . .

Silicon Vidicon Target Construction and Signal
Production . . . . . . . . . . . . . . . . . . . . .

Signal Output and Dark Current vs Target Voltage at
Constant Illumination. . . . . . . . . . . . . . . . .

Silicon Vidicon Tube Geometry. . . . . . . . . . . .
Beam Scan Pattern. . . . . . . . . . . . . . . .

CRT Format for Operator Dialogue . . . . . . . . . .
Program VEDO Terminal Interactions . . . . . . . . . .
Absorption Spectrum of PrCl3 (0.20M) in IN HCl . . .
Emission Spectrum of a Lead Hollow Cathode Lamp. . . .

Emission Spectrum of Neon Filler Gas in a Hollow .
Cathode Lamp . . . . . . . . . . . . . . . . . . . . .

Responsivity of RCA C23246 Vidicon vs Wavelength . . .
Quantum Efficiency of RCA C23246 Vidicon vs Wavelength
Channel ReSponse of Vidicon Target to Hg 436 nm Line .

Channel Response of Vidicon Target to a Continuous
Source 0 O O C I O O O O O O O O C O C C O O O 0

Emission Spectrum of a Lead Hollow Cathode Lamp. .

Emission Spectrum of a Mercury Lamp. . . . . . . . . .
Vidicon Response vs Number of Wavelength Channels. . .
Components of the Vidicon Spectrum . . . . . . . . . .
Vidicon Response vs Grid #1 Voltage. . . . . . . . . .

Vidicon Response Near Saturation . . . . . . . . . . .

ix

Page

11
11
28
34
38

39

40
43
43

45

46
51
52
55
57
59

61

Figure Page
4-11. Vidicon Transfer Function. . . . . . . . . . . . . . . . . 63
4-12. Vidicon Response vs Exposure Time. . . . . . . . . . . . . 67
4-13. Vidicon Dark Signal. . . . . . . . . . . . . . . . . . . . 69
4-14. S/N Enhancement with Background Subtraction. . . . . . . . 70
4-15. S/N Enhancement with Signal Averaging. . . . . . . . . . . 72

4-16. S/N Enhancement with Signal Averaging and Charge
Integration. . . . . . . . . . . . . . . . . . . . . . . . 74

5-1. Absorption Spectrum of a Holmium Oxide Filter. . . . . . . 81

5-2. Absorption Spectra of Nickel (0.10M) and Cobalt (0.17M)
in 2N HClOH. . . . . . . . . . . . . . . . . . . . . . . . 85

5-3. Calibration Curve for Nickel Determination . . . . . . . . 87

5-4. Effect of Sample Settling with Single-Wavelength
Scanning . . . . . . . . . . . . . . . . . . . . . . . . . 89

5-5. Dual-Wavelength Correction for Sample Settling . . . . . . 89
5-6. Absorption Spectrum of PrC13 (0.20M) in 1M HCl . . . . . . 93
5-7. First Derivative of PrC13 Absorption Spectrum. . . . . . . 94

5-8. Second Derivative of PrCl3 Absorption Spectrum . . . . . . 94

5-9. Mercury Emission Spectrum. . . . . . . . . . . . . . . . . 95
5-10. First Derivative of Mercury Emission Spectrum. . . . . . . 96
5-11. Second Derivative of Mercury Emission Spectrum . . . . . . 96

5-12. Intensity of First Derivative Signal vs Absorbance
of Neutral Density Filter. . . . . . . . . . . . . . . . . 98

5-13. Intensity of First Derivative Signal vs Time
During Sample Settling . . . . . . . . . . . . . . . . . . 98

6-1. Dedicated Minicomputer for Instrument Control. . . . . . . 102
6-2. Dedicated Microcomputer for Instrument Control . . . . . . 104

6-3. Distributed Multiprocessor System for Instrument
Control. . . . . . . . . . . . . . . . . . . . . . . . . . 106

Figure

6-4.

6—6.
6-7.
6-8.
6—9.
7-1.
7-2.

7-3.

Multiported System . . . . . . . . . . . . . . .
Single Time-Shared Bus System. . . . .

Dedicated Bus System .

Dedicated Bus with Bus Switch. . . . . . . . . .
Block Diagram of Multiprocessor Bus System .
Block Diagram of Vidicon Spectrometer System .
Block Diagram of Vidicon Control Circuits.
Vidicon Slave Micro.

Vertical Deflection Drive Circuit. . . . . . . .
Vertical Deflection Circuit Waveforms (25 kHz)
Horizontal Deflection Drive Circuit. . . . .
Horizontal Deflection Circuit Waveforms (25 kHz)
Vertical Deflection Circuit. . . . . . . . . . .
Horizontal Deflection Circuit. . . . . . . . .
Real-Time Clock and Interrupt System . .
Interface for 57109 Number-Oriented Processor. .
Shutter Control Circuit. . . . . . . . . . . . .
Analog Conditioning and Sampling Circuits. .
Fast Adder Controller Circuit. . . . . .

Fast Adder Timing Waveforms. . . . . . . . . .
Fast Adder Circuit . . . . . . . . . . . . .
Block Diagram of Front Panel Input/Output. . .
Keyboard/Address Display Logic . . . . . .
Keyboard Interface . . . . . . . . . . . . . . .

Front Panel Control Logic. . . . . . . . . . .

xi

Page
109
109
112
112
113
115
120
122
127
127
129
129
131
133
136
138
140
142
145
147
149
223
225
226

229

CHAPTER 1

INTRODUCTION

The dominant technique for the measurement of electromagnetic
radiation in the UV-VIS-IR region is the dispersive system based on
the diffraction grating. The typical dispersive system decodes inci-
dent frequency information into an optical spatial array in the focal
plane of the instrument, which is usually a monochromator or spectro-
graph. In the past, the two main detection systems used to measure
the spectral intensities in the image have been the photographic
emulsion (plate) and the photomultiplier tube in conjunction with an
exit slit. Although the photographic plate is a multichannel detector
capable of simultaneously recording thousands of lines during a single
exposure, it has several severe disadvantages, namely; a nonlinear
response, limited dynamic range, and tedious readout. Thus, even
though the photomultiplier tube—exit slit combination is limited to
measurement of one spectral resolution element at a time, it is the
detection system of choice for the majority of spectrochemical measure-
ments because of its wide dynamic linear range, sensitivity, and direct
conversion of light intensity to an electrical signal.

Fortunately, thhssingle-channel limitation is now being overcome
by the utilization of electronic image sensors as spectroscopic
detectors. These image sensors combine the desirable characteristics
of photomultiplier tubes with a simultaneous multichannel capability.
One type of electronic image sensor is the silicon—target vidicon. The

most attractive features of this image detector for spectroscopic

applications are its broad spectral coverage (250-1100 nm), high
quantum efficiency, linear response over a fairly wide dynamic range,
low geometrical distortion, and reliability.

The design, characterization, and optimization of a multiprocessor—
controlled silicon vidicon spectrometer are presented in this disserta-
tion. The versatility of the computer-controlled instrument lies in
the ability to change easily the scanning format by software changes
rather than by hardware modifications. In addition to instrument
control, the computer is used to implement several data processing
options for signal-to-noise (S/N) enhancement.

Chapter 2 presents a thorough discussion of the operation of the
silicon vidicon tube, as well as surveys of recent developments in
vidicon tube technology and recent spectroscopic applications of the
silicon vidicon tube. Chapters 3 and 7 discuss the operation and
construction details, respectively, of the silicon vidicon spectrometer.
The instrument can monitor a 230 nm window within a broad spectral
range of 250 nm to 1000 nm. Spectral information within this 230 nm
window is acquired simultaneously by the integrating diode array and
read out sequentially by a scanning electron beam under computer
control. Under software control, the user can specify the wavelength
resolution (number of channels/scan), the scan frequency, vertical
integration interval, detector integration time per scan, number of
scans to be averaged for S/N enhancement, and the magnitude of the
beam current.

A.major objective of this research was the characterization and

optimization of the performance of the silicon-target vidicon as a

spectroscopic detector. The experiments designed to obtain this infor—
mation and the corresponding experimental results are discussed in
Chapter 4. Characteristics examined included the multichannel spectral
response, the wavelength linearity and response, and the signal depen-
dence upon incident light level and integration time. Factors limiting
the S/N and methods for S/N enhancement were also investigated.

Chapter 5 describes the application of the vidicon spectrometer
to three basic types of dual-wavelength spectroscopy; dual-wavelength
non-scanning spectroscopy, dual-wavelength scanning spectroscopy, and
derivative spectroscopy. Because of the vidicon's multi-wavelength
capability, the implementation of these dual-wavelength techniques with
the vidicon spectrometer requires only changes in the data-processing
software. Realized advantages of the dual-wavelength mode included
continuous correction for scattering and sample settling and elimi-
nation of the influence ofznlinterfering component in the analysis of
a highly overlapping two-component mixture.

To provide real-time control of the instrument, a multiprocessor
system was designed and built in our laboratory. Chapter 6 describes
the development of the multiprocessor system and its advantages in real-
time instrument control. Multiprocessor control of the vidicon spec-

trometer is also discussed in this chapter.

CHAPTER 2

OVERVIEW OF THE SILICON VIDICON TUBE

IN ANALYTICAL SPECTROSCOPY

A. Introduction

The purpose of this chapter is to provide a clear perspective on
the operation and application of the silicon vidicon tube as a spectro—
scopic detector. The first section explains in detail the theory of
operation and signal generation of the silicon vidicon tube. Familiar—
ity with this material is necessary to understand fully the following
chapters on the design, characterization, and operation of the vidicon
spectrometer developed in this research. The second section reviews
recent developments in vidicon tube technology, including several
unique imaging devices based upon a conventional vidicon tube. The
last section surveys recent spectroscopic applications of the silicon
vidicon tube to illustrate its usefulness as a multichannel detector

in analytical spectrosc0py.

B. Operation of the Silicon Vidicon Tube

The RCA C23246 silicon-target vidicon used in this research is a
very high sensitivity photoconductive camera tube with extremely broad
spectral response, high resolution, low residual signal, and low dark
current characteristics. In electron-optics design, in mechanical con-
struction, and in physical appearance, it is virtually identical to the
standard vidicon photoconductive camera tube. However, whereas a

standard vidicon tube has an evaporated photoconducting film as the

image sensing target, the silicon-target vidicon has a planar array of
reverse-biased silicon photodiodes. The most obvious advantages of the
silicon target over more conventional targets are its sensitivity,
which results from an almost unity quantum yield throughout the entire
visible and near infrared region, and its resistance to image burn-in
or damage resulting from overexposure (1-9).

Figure 2—1 illustrates the silicon-target vidicon tube. The
Optical image is focused by a lens onto the substrate of the silicon
photodiode array. This diode array has two distinct functions: 1) to
convert the optical image into a corresponding electrical charge
pattern, and 2) to accumulate and store the charge pattern until the
electron beam and scanning mechanism read out the signal. The diode
side of the array is scanned by an electron beam that has passed
through the appropriate electron optics for focusing and deflection.
When the scanning electron beam recharges the diodes, a current exists
in the VT circuit (Figure 2-1). This recharging current constitutes
the video signal.

A silicon photodiode array typically consists of a matrix of
660 by 660 diodes; about 436,000 diodes within a 1.85 cm square. The
substrate is nominally 10 Q-cm n-type silicon with a diameter of
2.16 cm. The substrate in the area of the diode array is 5-50 pm
thick, while the perimeter of the wafer is usually 100 um thick to
ensure a self-supporting structure (2,3,7). The diodes are islands of
p-type dopant diffused through gaps in a silicon oxide insulating film
(Figure 2-2). The diode array has a typical center-to-center diode

spacing of 15 um and an oxide hole diameter of 8 um.

COwaw> umwumHICOUflme m mo owumEmzom .~IN muswwm

mo<So> s H. uqzoa .30on
Box: > lh. 39>

  

 

 

 

 

 

 

fl
\
msz .\\\\\
A «My supm>m
\\
L
\Y / 2<wm zo.hom4umo
zomhowqw oz< 025300“.
mun—0.0 00— x ¢.OA
/ ><mm< moo—o

med:—
Zocu bro:

PHOTO - GENERATED

LAYER I
CHARGE CARRIERS \ \

LIGHT INPUT

+

  
 

 
          
     
 

  
 

DEPLETION ,"x

.\\\\\\\\‘

1665

 

ELECTRON BE AM

7////
/\

 

k

\\\

    

’1'

 

 

\
BEAM %
LANDING PADS

 

(METALLIC)
MONOCRYSTALLINE
SILICON WAFER
“N" - TYPE
(THICKNESS Sioz
lO-ZO pm)

  

 

VIDEO DIODE CELLS FORMED

SIGNAL BY INTRODUCING P-TYPE
+ DOPANT (DIODE SPACING
VT 3 l5 pm)

Figure 2-2. Silicon Vidicon Target Construction and
Signal Production

 

 

SIGNAL OUTPUT

 

 

 

 

250? - so
f; 200- -40 <
C
I I
.—
g.
2 I50- - 30 z
t. I.“
3 K
° 5
- - o
i 100 20
2 3%
(3 <
H H
0 T I I I I I 0
O 4 8 I2 IS 20 24 28 32

TARGET VOLTAGE - V

Figure 2-3. Signal Output and Dark Current vs Target
Voltage at Constant Illumination

The silicon oxide film separates the diode matrix and prevents
the scanning beam from landing outside of the diode and thereby
finding a short—circuit path through the silicon. Unfortunately, if
the oxide is exposed to the beam, it will accept and store electrons
until it becomes negatively charged to the point where it will repel
any further incident electrons from the beam, i.e. the openings into
the diode centers will be closed off (1-3). In order to minimize
oxide exposure to the scanning beam, RCA adds a pattern of raised
conducting "beam-landing pads" to their vidicon targets (1) (see
Figure 2-2). Precision fabrication of these metallic pads gives
nearly complete shielding of the silicon dioxide from the scanning
beam, while still isolating each diode pad.

In normal operation, the substrate of the diode array is biased
positively with respect to the cathode of the electron gun. For an
RCA silicon-target vidicon tube, the optimum bias (target) voltage is
8 volts (see Figure 2-3) (1,3). At this voltage the peak target cur-
rent is approximately 750 nA. Higher target voltages result in
excessive dark current and do not increase the sensitivity of the
vidicon. Lower target voltages result in inefficient collection of
the generated charge carriers, inability of the tube to handle reason-
able levels of signal current, and lower resolution.

The electronic charge deposited by the electron beam places a
reverse bias of 8 volts on the photodiodes as it scans over the array.
This reverse bias creates a depletion width of approximately 5 pm with
a 10 Q-cm substrate and results in an effective charge storage capac-
itance for the target of approximately 2000 pF/cm2 (1,2,7). Thus the

diodes store electrons left by the beam as it charges them negatively

to the potential of the cathode of the electron gun. The insulating
effect of the reverse-biased junction prevents the charges on the
diodes from migrating within the target. Because the diodes have very
low leakage currents (less than 0.1 pA per diode), they remain in the
full reverse-biased condition throughout an entire frame period if
illumination is absent.

When light is incident on the target, photons are absorbed in the
n-type region, and an electron-hole pair is created for each absorbed
photon. For visible light the majority of photo-generated carriers
are created near the illuminated surface. As the creation of photo-
generated carriers increases the minority carrier (hole) density above
its thermal equilibrium value, illumination causes a net diffusion of
holes towards the reverse-biased diodes. If the lifetime of the holes
is sufficiently long and the illuminated surface has been treated to
reduce recombination effects, a large fraction of the photo—generated
holes will diffuse to the depletion region and contribute to the junc-
tion current. As long as the diodes remain reverse-biased, the light-
induced junction current will continue to discharge the diode capaci-
tance. In this manner a charge pattern develops on the diode array
that is an electronic image of the optical image focused on the target.

The video output from each diode is created when the scanning
electron beam returns to a diode and restores the original charge by
re-establishing the full value of the reverse bias. If a fixed voltage
of -8 V is applied to the cathode, the video output current from the

target can be directly coupled to a current-to-voltage converter to

10

produce a position—encoded video output voltage (10) (see Figure 2-4).
This arrangement maintains the necessary 8 V potential difference
between the cathode and the target, while it eliminates the need for
capacitive coupling between the target and the amplifier.

The electron beam is deflected across the target by two mutually
perpendicular magnetic deflection coils. Typically, the scan pattern
is a series of parallel lines that start in one corner of the target
and end in the opposite corner (see Figure 2-5). Normally the rates
of deflection for the two axes are 30-100 Hz and 15 kHz. In spectro-
scopic applications, the slow axis is usually positioned parallel to
the wavelength axis in the focal plane of the monochromator.

The video output current, I, produced by a silicon vidicon is

expressed by:

I = QeGA(N)YTe/Tr in A

where Q quantum efficiency of the photosurface, dimensionless

e = electron charge, C

G = target gain (G = 1 for a silicon vidicon)

A = area of target scanned per output channel, cm2

N = incident photon flux, quanta cm'2 sec"1

7 = slope of the current vs luminous flux curve (y = 1 for a

silicon vidicon)

Te = detector exposure time between readouts, s
Tr = active readout time per output channel, 5 (11).

If the photon flux is constant over the exposure time and the scan

pattern is constant and reproducible, this equation reduces to

11

SILICON TARGET

 

 

FUSEO
SILICA ELECTRON BEAM
WINQW E FOCUS, ALIGN, AND SCAN coILs CATHODE

., f * 1 (-av)

 

 

 

 

II \ I/II‘II‘F/
I: \\\ \

A \ x
I I \ GRID I HEATER
GRIoz

GRID 4 GRID 3

 

 

 

 

 

 

 

 

 

Figure 2—4. Silicon Vidicon Tube Geometry

 

 

 

 

 

 

 

 

 

‘ STOP
I I

2 II II I \ I\ I

4 \ \ \ \ \

,. I \ I I

If \ \ I I \

3 I I I I I

I3 \ \ I I \

=2: \ I I I I

2’ I I I I I

2 \ \ I I\ \

E I I I \ \

> I I II I I I
START

 

HORIZONTAL (WAVELENGTH) AXIS

Figure 2-5. Beam Scan Pattern

12

I = k(Te/Tr), where k equals the constant QeGAN. Thus, the average
output current of any given channel is directly proportional to the
ratio of exposure time to readout time for a constant incident photon
flux. This forms the basis for the charge integration technique for

signal enhancement (see Chapter 4).

C. Recent Developments in Vidicon Tube Technology

In the past several years a number of excellent review articles
have appeared in the scientific literature concerning the development
of camera and storage tubes and self-scanned solid—state image sensors
for pick up and reproduction of optical images (2,7,12-18). Because of
the depth and breadth of coverage of these review articles with respect
to currently available imaging devices, a review of imaging devices and
their applications in spectroscopy is not included in this chapter.
Thus, the following review covers only the more recent developments in

vidicon tube technology.

1. Silicon-Target Vidicon

The concept of using a large—scale array of silicon photodiodes
as a photosensitive target for a camera tube was first proposed at
Bell Laboratories in 1967 (19). Since then, rapid advancements in
silicon technology and large-scale integrated-circuit technology have
led to arrays with more than 600,000 individual diodes on a silicon
wafer (1). Continuing improvements in manufacturing efficiency and
techniques have reduced the number and prominence of blemishes from
faulty diodes and have resulted in lower-cost tubes. Also, the

mechanical strength of the tubes has been increased by mounting the

13

target as a part of the faceplate assembly. The introduction of fused—
silica and quartz windows has extended the response of the silicon
vidicon into the UV to less than 250 nm. These ongoing improvements
have helped to expand the usefulness of the silicon—target vidicon in
existing applications and to extend its use into new areas.

One of the most recent developments with this type of tube is a
reduced-blooming target that greatly reduces the highlight—blooming
characteristic of the silicon-target tubes (1). (Blooming is a lateral
spreading of a highlight image that occurs when the signal saturates a
localized area of the diode array.) These tubes will eventually rephxm
the standard silicon-target tubes and should offer extended dynamic

ranges of 105 or more (12).

2. Focus-Projection-Scanning Vidicon

An important advance in the electron optics for high resolution
cameras in recent years was the invention of the Focus—Projection—
Scanning (FPS) vidicon (22—24). This vidicon combines the advantages
of electrostatic deflection and magnetic focusing into one unique
electron Optics assembly. Inherent advantages of the FPS design
include: optimized beam landing and edge focus, lower power require-
ments, higher beam current densities at the photoconductor, and smaller
size and weight. Additionally, the FPS electron optics are compatible
with any photoconductive target. Because FPS vidicons have 60 to 70
percent fewer internal parts than conventional magnetic vidicons and a

correspondingly lower mass, they are extremely rugged devices.

14

3. Pyroelectric Vidicon

The search for improved cameras for thermal imaging has resulted
in the development of the pyroelectric vidicon (PEV) (20,21). The
pyroelectric vidicon is a low-cost, infrared-sensitive camera tube that
operates at room temperature. It can image in any wavelength band
within the 2—400 um range and is compatible with all conventional
television formats. Because the PEV responds only to changes in temper-
ature, it is insensitive to any constant radiation incident upon it.
However, a chopping and storage mode has recently been developed to
enable the integrated signal from several frames to be displayed as a
static image (20).

The pyroelectric vidicon is quite similar in appearance to the
normal photoconductive vidicon. It differs in that the front window is
usually germanium, which is opaque to visible radiation, but highly
transparent in the 8-14 um range. The most common target material for
the PEV is TriGlycine Sulfate (TGS). The PEV uses a vidicon gun with
a conventional thermionic cathode to provide a low velocity electron
beam to sample the pyroelectric target. For the 8-14 um region, the
PEV has a typical minimum resolvable temperature of less than 1°C at
250 television lines (TVL). Successful areas of application of the
PEV include infrared laser detection, forest fire mapping,

security/surveillance, and medical diagnostics.
4. X-Ray-Sensitive Vidicon

Recently a new type of imaging tube for x-rays and y-rays in the

80-keV and higher energy range was developed at Northwestern University

15

(25). The uniqueness of this x-ray—sensitive vidicon to the measure—
ment of radiation lies in the vidicon's inherent ability to image low
levels of radiation by charge integration in the target plate. The
target plate for this tube consists of a layer of lead deposited on an
aluminum substrate. The lead converts a portion of the incident radi-
ation into photoelectrons. These photoelectrons travel through a low-
density layer of cesium iodide deposited on the lead layer and create
secondary electrons. An electric field across the cesium iodide layer
causes the secondary electrons to be collected by the conductive back-
ing of the target plate, which creates a charge pattern on the surface.
A video output signal is created by scanning an electron beam, which
neutralizes the charged array. There is no lateral spreading of the
charge on the cesium iodide surface because of the high resistivity
(1017 0/cm) of the cesium iodide layer. In order to maximize the
signal current for a given radiation energy, it is necessary to
optimize the thickness of the lead layer and the thickness and density
of the cesium iodide layer. Anticipated areas of application of this
device include uptake scanning (visualization of the distribution of
radioactive material in an organ or gland) and high—kV radiographic

imaging.
5. Piezoelectric—Target Vidicon

Another interesting variant of the vidicon is the piezoelectric-
target (quartz) vidicon that is used for visualizing ultrasonic images
(26,27). The target of this vidicon consists of a mosaic of quartz

crystals, which provide an electronic image of the incident acoustic

16

field distribution. Piezoelectric voltages are generated in the
quartz target at all regions corresponding to the received acoustic
image. As in the conventional vidicon, a charge pattern corresponding
to the acoustic image is formed on the signal plate side of the target
and subsequently neutralized by a scanning electron beam. A promising
application for this imaging device is in the ultrasonic imaging of

human anatomy.

D. Spectroscopic Applications of the Silicon Vidicon Tube

Imaging devices are finding steadily increasing application as
multichannel detectors in analytical spectroscopy. During the past
several years multichannel detectors, which employ either silicon
vidicon or photodiode array detectors, have been used for atomic
absorption (28-34), emission (10,11,18,33,35-56), and fluorescence
(49,50,57,58) Spectroscopy, molecular UV-visible absorption (10,18,32,
55,59-82,91) and fluorescence spectroscopy (83-90), and Raman spec-
troscopy (92,93). This rapid expansion in the application of multi-
channel detectors has been aided by the availability of complete
detector systems from several commercial companies, such as Princeton
Applied Research, Nuclear Data, EMR, RKB, Quantex, Tracor-Northern,
and Tektronix. Additionally, the detector heads and electronic
components are also available on a modular basis from the following
companies; Reticon, GE, Hamamatsu, EMI, Amperex, ITT, RCA, Fairchild,
Teltron, TI, and Westinghouse. Nieman (18) has written an excellent
review on the historical development and application of imaging devices

in analytical spectroscopy. Consequently, this review covers only

17

recent spectroscopic applications of the silicon vidicon tube.

Nieman and Enke (61) and Pardue et. al. (10,59) have recently
developed and characterized computer-controlled vidicon spectrometers
for molecular absorption and atomic emission spectrometry. The vidicon
detector was found to be uniquely suited to multicomponent analysis and
preliminary investigations of the kinetics of chemical reactions,
including the detection of reaction intermediates with lifetimes as
short as 10 ms (59). Recent applications by these authors include
simultaneous multicomponent drug determinations (66,67), kinetic
methods for simultaneous enzyme determinations (66,74), and reaction-
rate methods of analysis (63,78-80).

More recently, Pardue and coworkers (60,81) have developed a
vidicon based derivative spectrometer that simultaneously generates
the normal spectrum and its first derivative. Electronic wavelength
modulation is employed to generate the derivative spectrum. It is
achieved by superimposing a low amplitude periodic waveform on the
horizontal (wavelength) deflection signal. A phase-sensitive lock-in
amplifier, referenced to the frequency and phase of the wavelength
modulation waveform, generates a signal proportional to the first
derivative of the optical spectrum incident on the vidicon detector.
This instrument is currently being applied to the determination of
bilirubin in biological fluids (60).

Ostertag (70) converted a commercial optical multichannel
analyzer (PAR) into a pulse-operated system that fully synchronizes
on an external event occurring even at low repetition rates. The
optical spectra measured with this instrument are those of the

luminescence of crystalline samples that have been excited by a

18

picosecond laser. Because of the low repetition rate of the measured
events, a simultaneous background subtraction feature was added to
provide automatic correction for long range drifts of dark current in
the vidicon detector.

Silicon vidicon tubes have been used in an integrating mode for
ground-based telescopic observation of celestial objects (77,94,95).
Silicon vidicons are preferred over the intensified versions (SIT)
because of the requirements for high photometric precision (<1%) and
IR sensitivity to relatively bright objects. The vidicons are mounted
in vacuum-insulated boxes and cooled with dry ice to minimize the dark
current and to extend the maximum usable integration time. At dry
ice temperature, the accumulation of dark current in 1 hr is only 4%
of full-scale saturation. To further minimize the accumulation of a
background signal during signal integration, the filament current is
turned off during integration, because backlighting of the IR-responsive
silicon target by the cathode heater contributes a nonspatially uniform
signal on the order of 1% of full-scale saturation in 1 min of warm
operation at 6.5 V. Consequently, the integration time is limited by
the intensity of the celestial object under observation rather than by
the electronic background. Images of extended sources through the
telescope agreed photometrically with photomultiplier measurements to
about 1%.

The silicon vidicon tube is also a viable multi-wavelength
detector for liquid chromatography (96). It offers the following
advantages over single wavelength detectors: fingerprint spectral

information to complement retention times, selection of wavelengths

19

which optimize sensitivity and resolution for each component, ability
to resolve quantitatively overlapping chromatographic peaks, and the
ability to detect impurities in chromatographic peaks. A preliminary
application of this configuration (96) involved the analysis of
pharmaceutical preparations (Sinutab II and Primatene tablets).
Similarly, a silicon intensified-target vidicon (SIT) was used as a
multichannel detector for the identification of fluorescent liquid
chromatographic petroleum fractions (58) and as a gas—phase fluores-
cence detector for gas chromatographic analysis of polyaromatic
hydrocarbons (89). The SIT gave limits of detection about a factor
of five times larger than those obtained with a photomultiplier tube.
To be useful in the clinical laboratory, multicomponent fluores-
cence analysis requires the rapid measurement of the fluorescence
intensity at a variety of excitation and emission wavelengths. Because
of the low light intensities often encountered in fluorescence spectro—
scopy, a SIT detector was used as the multichannel detector in a new
video fluorometer (84,97). A complete set of emission and excitation
spectra for perylene were obtained in less than 2 s at concentrations

of 10.10 M. Another recent application of the video fluorometer

involved a study of the fluorescence of the living brain and heart (57).
As a quantitative tool, the video fluorometer is potentially of enor—
mous value in measuring the damage in strokes and shock, as well as in
basic studies of circulation and metabolism in the heart and brain.
Pardue and coworkers (55) and Morrison and coworkers (37,40) have
used vidicon spectrometers for simultaneous flame—emission analysis of
electrolytes in serum. Morrison's group combined first and second

order diffraction lines to obtain analytical lines within the detector‘s

20

wavelength window, and used filters to partially absorb selected lines
to equalize emission intensities. Results for control sera and clin-
ical samples indicated an accuracy and precision of better than 2% (37).
Simultaneous emission determinations of up to eight metals have been
performed (10,11,36,43,45,58,49), but the S/N for the silicon—target
vidicon is an order of magnitude lower than for single channel detec-
tion (11). Caruso and coworkers (48) demonstrated that a microwave
induced plasma coupled to a vidicon detector has potential for simul-
taneous multielement trace analysis. Winefordner and coworkers (49)
compared single channel and multichannel detection limits for eleven
metals. The multichannel detection limit was an order of magnitude
higher than the single channel detection limit. Busch, Howell, and
Morrison (38) used spectral stripping to remove either molecular bands
or undesired concomitant interferences from analytical lines of
interest.

Simultaneous determinations of up to four elements (28,29,31,34)
have demonstrated that the silicon vidicon is also a viable multi-
channel detector for atomic absorption spectrometry. The observed
detection limits were an order of magnitude higher than for single
channel detection. Multichannel detection allowed summation of the
signal from several lines of the same elements for increased sensi-
tivity (31). A random access vidicon echelle spectrometer provided
both high resolution and wide spectral coverage for atomic absorption
measurements (34). By using the random access mode of target inter-
rogation to measure only the line intensities of interest, a reduction

in computer time and memory space was realized, as well as greatly

21

simplified data reduction. For atomic absorption, the use of a non—
integrating, electron-amplifier detector, such as an image dissector,
would allow rapid signal modulation to eliminate interference from
flame emission, and the electron amplification should improve S/N
ratios to a level compatible with that of a photomultiplier tube (29).
The sensitivity characteristics of the ultraviolet-sensitized
silicon vidicon and silicon intensified-target (SIT) vidicon tubes were
compared in the ultraviolet-visible wavelength region (52). Flame
emission detection limits for 23 elements were reported and compared
to the best reported photomultiplier tube values. The results
demonstrated that the silicon vidicon tubes are capable detectors in
the visible wavelength region, and that the SIT vidicon tube provides
detection power equivalent to commonly used photomultipliers in the
visible spectral region. The image tube detectors' poor responsi-
tivity characteristics in the ultraviolet region limited their applic-
ability to measurements in this spectral region. Winefordner and
coworkers (87) demonstrated that while the silicon vidicon is not
analytically useful for atomic fluorescence or molecular luminescence
spectrometry, intensified image devices possess considerable
analytical potential in molecular luminescence in the visible spectral

region.

CHAPTER 3

OPERATION OF THE SILICON VIDICON SPECTROMETER

A. Introduction

The purpose of this chapter is to describe the basic operation of
the microprocessor-controlled vidicon spectrometer as a general purpose,
single-beam scanning spectrophotometer. Since all instrument control
and data acquisition functions are under computer control, this chapter
focuses primarily on the operating software that was developed for the
vidicon spectrometer. The software and terminal interactions for
operation of the instrument are discussed in the following section.
Although the acquired spectral information can be processed locally by
the micro, the stored spectra are usually transmitted to the mini-
computer for final processing. Accordingly, details of the minicomputer
software developed for post-processing of the acquired spectra are
discussed in Section C. All programs relevant to this chapter are

listed in Appendix A.

B. Real-Time Control and Data Acquisition

Real-time control and data acquisition in the vidicon spectrometer
are provided by a multiprocessor system that was developed jointly with
Erik M. Carlson (98) and Spyros Hourdakis (99). The multiprocessor
system not only provides real—time control of the vidicon detector's
scan format and the data acquisition system, but also controls the

monochromator wavelength drive, optical shutter, and the positions of

22

23

the sample and reference cells in the cell module. Details Of the
computer system architecture are presented in Chapter 6. Documentation
Of the design details of the electronic, mechanical, and optical com—
ponents of the silicon vidicon spectrometer are presented in Chapter 7.

The vidicon detector is mounted on the faceplate of a GCA/McPheranl
EU-700 Czerny-Turner monochromator with the beam slow-scan axis paralkfl.
to the wavelength axis of the dispersed spectrum. The monochromator
was modified so that the grating image is focused on the target of the
vidicon tube. With the grating and optics used in this research, the
target monitors a wavelength window of 230 nm. Since the vidicon's
silicon target provides usable response from about 250 nm to 1000 nm,
the detector's 230 nm wavelength window can be positioned anywhere
within this broad spectral range. The wavelength control circuit
divides the target into a user—defined number of parallel electronic
channels. Information in each wavelength channel is read out by a scan-
ning electron beam under computer control and integrated over the
vertical trace by the data acquisition system to enhance the signal-
to—noise ratio. Through the software the user can vary the above scan
parameters to optimize the response of the vidicon spectrometer to the
chemical system under study.

A new software package, CONVERS (100), was modified to run on a
multiprocessor system (98) and was subsequently used as the software
Operating system for the vidicon spectrometer. This system is based
upon an interpretive compiler and offers the advantages of high speed
operation, excellent memory efficiency, stack manipulation, and a

variety of high level constructs, i.e. the programmer can develop his

24

own individual modifications and additions to the language itself.
Once software modules have been entered into the system software
library, they can be called up and executed in one of two ways: the
name of the software module can be entered at the terminal for immedi-
ate execution, Or the module name can be included as a line of code in
another software routine. Thus, once the assembly language routines
for instrument control and data acquisition have been written, the
programmer can develop his own instrument-specific high—level language
by designing software modules that contain the desired sequence Of
assembly language module names.

In the vidicon spectrometer, the coordinator micro controls and
coordinates all activities in the multiprocessor system, and all termi-
nal interaction is with this micro. The software that was developed
for the coordinator micro is outlined in Table 3-1. The order of these
routines is consistent with the order in the program listing in Appen-
dix A.

The power of the CONVERS language is evident in the following
sequence of CONVERS commands that was used for characterization Of the
vidicon spectrometer:

VINIT
DATA
PLT
EVAL
WFILE
Although this command sequence contains only five CONVERS calls, these

individual modules allow maximum flexibility in optimizing the response

25

 

 

.Hmo ecu Ou MOQEDG HmEHomv m museuso HDoomn
.aom ofiweeum on sumcfin BABIES amasz
.cowmw>fic wocwfimcs headwn ufiplw Sn kumcwn ufinlcfi >HQ
.%maamww canvascmzaam onu mmapmcw
cam >maamww muflsamuw mcu mmapmmwv coca .vumon%mx one Eoum o Houuaou m now wxomno oo
.Hmu one so mxumexuwu can moxm mzmuo me<
.wumOQSOx mcu Eouw Hones: HOEHOOO m musaaH zHomo
.zumcwn ufinlofi Ou mum ufiwwclm szoma
.Aouuwe coowcw>v N chow: Ou AHOODCHOHOOOV C One“: Eouw mumc muowmcmuk NOB
.HHOO mucoumwmu mnu wuowaom mmm
.HHOO OHmEmm ecu muooaom mz<m
.uouuazm ecu momoao mmoqo
.Amuusnm one mcmao zmmo
.AHOOOOMODEEOOV A chow: Ou AHOODDwvuOOOV o OMOHZ Eouw Dunc muommcmuH Hoe
:Ofiuocam mEmz

 

umumEOHuomem coowcw> ozu mo cowumuoeo pom wumBumom mmm>zoo

.aum Danae

26

 

 

.cofiuomaaoo meme neumfiuflcH <H<n
.Hmo ecu Du waoccmso Luwcmam>m3
wouumamm you mononuomnm paw .Oocmuqumcmuu .oocoumwou .OHaEmw mo mosam> mundane A<>m
.Hmu ecu co Esuuooam mocmpuomnm no .mocmuu«Emcmuu .moamuowmu .OHaEmm mo muwono muOHm Ham
.mafim omww haeoam m Ou xuoeoe mumv mo museucoo acouuso mmufiHS mAHmB
.umuoeouuooam Douawfi> onu mmnwamwuwcH HHZH>
.>HOEOE mumw Mm osu mumoao Mwmgu
.muwsoufiu woven ummm use mmufiamauHcH Hmmm
.muwsuuwo :mom coowwfi> Osu mmuwamwuwcH Hmm>
.mmaemfic ow»mE:cm£aHm mcu moanmwfiv cam Smammfiw wuwzamuw mnu moanmcm zoo
.cumonmmx osu Eoum
mosam> Oasmwmv 3m: musacw can emu ecu Ou muoumempma cmom wasmmmv acouuso mundane Haom
.Hmu\cumonzmx EOHM\OO :Owuwmoa MOOOEousuocoE muaauso\mu=a:H HDOZHS
.A.v.ucouv .film OHQOH

 

 

27

of the vidicon spectrometer. The CONVERS routines resident in the
other two micros are called as needed by the coordinator routines.
Consequently, the routines in the two slave micros are completely
transparent to the user.

The user initiates an experiment by entering VINIT from the key-
board. The coordinator then requests the current location of the mono-
chromator. After the wavelength has been entered, the coordinator
starts the operator dialogue routine (POUT). The CRT format for the
operator dialogue is illustrated in Figure 3-1. It consists of a
table of mnemonics, typical mnemonic default values, and an operator
prompt (>). Table 3-2 explains the mnemonics and their functions.
The user can change any or all of the mnemonic values by entering the
mnemonic and its new value (see underlined portion in Figure 3-1). The
processor checks the validity of each keyboard entry and updates the
CRT display after each valid user response. If an illegal mnemonic or
mnemonic value is entered, the processor outputs a question mark and
waits for a valid input. If an entered mnemonic value is within the
established boundary limits, but is not one of the discrete values
allowed for that mnemonic, the processor selects the nearest allowed
value and assigns it to the mnemonic. An entered mnemonic becomes the
new default value for that mnemonic.

Although the number of wavelength channels (PTS) can vary from 16
to 4096, only values of between 64 and 512 points/spectrum are useful.
Fewer than 64 points results in too little resolution for most appli-
cations, while more than 512 points requires excessive data storage

and actually degrades the signal (see Chapter 4). The number of

Figure 3-1.

WAV
PTS
D LY
B LK
B YT
AUG

ADR
OPT

>(MNEMDNICI

 

28

550
5|2

IOO

8|92

(VALUEIIC I

 

FRQ
BEM
ACQ
GAI
BIT
INT

CNT
SRD

CRT Format for Operator Dialogue

25
60

I2

I536

29

 

.AmH ou ~ I H<u pow H<U cw mwmouocw

 

Luwz mommouocfi :fimuv MOORMwODCw one mo Aawmwv ucmumcoo DEMO mHouucoo cfimo H<o
.AooH ou a n 2 How .z\nmz g u ommv NuHOLOHHx a“ Socoavoum cmom zocoocoum 0mm
.Aofi v sun + oo<v
coauwmwscom meme wcfiuumum ouowmn Swamp o» mHm>HOOCw Hmowuuo> mo Honesz Smaon wan
.Amem x arm n Hzov wouOum ma OO mouhn mo noses: HOuOH ucoou 920
.Am no .N .HV ucfioa some Hoe owOMOum mo wouzn mo monsoz mouam Hum
.chaxaman I a .wcaxcman o: I SO
coaumpwoucw Hmcwwm cam oomuuou weapon coon mo wCchmHn moanmcm wcwxcmam mam
.Am Lo .oH .NHV Deanne o\< ca soamfiumoa mo moan mo Amnesz Dam eHm
.Azmm x > mm.ou
u mwmuao> vapov He capo Ou mwmuao> wcwHHOHucoo an ucouuoo Emma mumm Emom 2mm
.Aeooq OO Hv cwwmuo>m on cu moemum snow mo poaasz owmuo>< o><
.owmu0um mumv MOM mmouvnm wafluumum mmonvv< mn<
.AoH coooxo uoc cmo >49 new Oo< mo 85m OLHV
.coOmquOCw on cu we Hmcwwm osu £OH£3 uo>o mam>uoOcH Hmowuuw> mo poneoz mufioco< oo<
cofiuocsm GOfimamaxm oa:oEo:z

 

mowcoaocz moonmfio Houmuoao

.NIm manmh

30

 

 

.mswofimfic MOOOHDQO mmumcflshoe pom ozm
.uOumeousoocoE mo Luwcoao>m3 acouuou LuwcmHm>m3 ><3
xuma
mucouomom
.AH n a .N u m .q n my OOOOOHHOO on 0u muuooem mo«maooam OHaEmm mam
.Aomm 0u H n z MOM .z\omoq u memv maoccmco :uwcoaw>m3 mo Monasz mucwom mam
.owcmupo>o cu umuuo>coo o\<
onu omsmo uoc moon umcu :Hmw ecumquOCH umowuma onu wCHuOoHom ha Emumhm
cowuwmwovom mumv onu mo :Hmw ecu monaEHuao HOODQEOO osu ..H. Ou uom MH muwefiuao Hmo
.Ammm Ou ov madam cmom Hoe mmEmum cowumuwoucw mo wonaoz :oHumuwoOcH HzH
.A.e.ocouv .Nnm Danae

31

wavelength channels also affects the time required to acquire a
spectrum at a given scan frequency. For example, at 25 kHz the time
required to acquire a spectrum ranges from 2.5 ms for a 64 point
spectrum to 20.5 ms for a 512 point spectrum.

The scan frequency, which is programmable from 10 kHz to 1 MHz,
also affects the data acquisition time. Since a high scan frequency
was not a prerequisite for this research, a scan frequency of 25 kHz
was used throughout this work as a satisfactory compromise between
sensitivity (integration time between successive readouts by the scan-
ning electron beam) and acquisition time.

For each spectrum acquired, the user can control three parameters
to enhance the S/N: the number of bits in the A/D output, the number
of spectral scans averaged, and the number of signal integrations per
scan. Although faster conversion speeds can be obtained by "short
cycling" (e.g. 4us for 8 bits vs 10us for 12 bits), the A/D quantiza-
tion error is minimized by using the full 12-bit output of the con-
verter. With a 12-bit converter output, the maximum S/N of the vidicon
detector without signal averaging is measured to be 3086. (Maximum

S/N is defined as the ratio of the maximum linear output level (~ 90%

of saturation, see Chapter 4) to the root-mean-square (RMS) value Of
the Observed background of a mercury lamp.) When time permits, signal
averaging can be used to enhance the S/N by reducing random noise and
converter quantizing noise (see Chapter 4).

The charge integration capability of the silicon vidicon target
can be used to enhance weak signals by increasing the exposure time.

The exposure time is increased by blanking the electron beam and

32

inhibiting the deflection circuits for an integer number of frames.
Signal intensity increases linearly with exposure time until the
accumulated signal approaches saturation (see Chapter 4). Since
charge integration integrates the signals in all wavelength channels
equally, this technique is limited by the most intense feature in a
spectrum. If integration continues after an intense region has sat-
urated, the signal from the saturated channels will spread into
adjacent channels. For a fixed data acquisition time, a combination
of charge integration and signal averaging will often maximize the
output response.

The SRD mnemonic specifies the spectra that are to be acquired.

A value of 7 results in the sequential acquisition of the sample, the
reference (for absorption), and the dark spectrum under computer
control. For a 512 point scan at a scan frequency of 25 kHz, all
three spectra can be acquired in less than 10 sec. The ADR and CNT
mnemonics are set by the computer and are not alterable by the user.
Optimum beam current is achieved with BEM set to a value Of 60. DLY
and ACQ may be changed as desired.

The END mnemonic terminates the operator dialogue routine. Upon
termination, the coordinator initializes the scan and data acquisition
control circuits and clears the data memory. Next, the DATA command
is entered to initiate data acquisition. Following completion of the
data acquisition cycle, the PLOT routine can be used to plot the sample,
reference, transmittance, or absorption spectrum on the CRT terminal.
Alternately, the EVAL routine can be used to output to the terminal the

sample, reference, transmittance, and absorbance values for specified

33

wavelength channels. Finally, WFILE outputs the sample, reference,
and dark spectral data to the floppy disc. Prior to storing the
spectrum on disc, WFILE allows the user to input the file name and
file header text from the keyboard.

Although the acquired spectral information can be processed
locally by the micro, stored spectra are usually transmitted to the
minicomputer for final processing. The communication link between
the microcomputer and the minicomputer was designed by S. Hourdakis
(99). The software written for this communication link (98,99) allows

bidirectional transter of programs between the two computers.
C. Post-Processing of Spectral Information

Post-processing of the spectral output from the vidicon spectrom-
eter is typically performed on the minicomputer to take advantage of
the computational ability of the mini and to use the mini's peripherals.
The data—processing software was written in FORTRAN IV and MACRO
(assembly language) to run on a PDP 11/40 (Digital Equipment Corpor-
ation) with 32K words of memory, RKOS hard disc, floppy disc, CRT
terminal, and line printer. Listings of the program and its sub-
routines are included in Appendix A.

An example of a typical terminal interaction with the program is
illustrated in Figure 3-2. The underlined portions are entered by the
user. The program first requests the file specification, then looks up
the specified file and outputs the file header text to the terminal.

If the file is to be processed, the user specifies the normalization

factor and indicates whether or not the background spectrum is to be

34

SRU VEDO
FILE SPECIFICATION - IFL18POL§2§oDAT

PCLS SPECTRUM .25M 122277
512 PTS D/A-3/6 25 KHZ
BEN 8 6C

PROCESS FILE 7:
NORMALIZATION FACTOR r A. 64) s.§
SUBTRACT DARK SIGNAL 7:

DATA OUTPUT CODES:
I NOTHING

DARK

SAMPLE
REFERENCE

3T
ABSORBANCE
WRITE FILE
LOCATE
EVALUATE
ANOTHER FILE
IO 8 EXIT

11 I SIN

OQQOUDUN—D
I I I I I I I I I

LIST :9
PLOT : i
THRESHOLD : 1g

MIN I OoOCOCOOO AT POINT 5
MAX 8 788.0000000 AT POINT 239

SCREEN BOTTOM - g;
SCREEN roe - 806-
LINE PLOT . a. POINT PLOT . I : CHOICE : g.

Figure 3-2. Program VEDO Terminal Interactions

35

subtracted from the sample and reference spectra. The upper limit on
the normalization factor is equal to the number of signal averages.

The lower limit represents the computer-selected prescaling factor that
was required to fit the 24-bit data word from the micro into a 16—bit
integer word at the mini. The normalization factor allows the user to
specify the number of extra bits from signal averaging that are to be
retained. Typically, half of the extra bits are retained for proces-
sing.

Next, the dark, sample, reference, transmittance, or absorbance
spectrum can be selected for plotting on the terminal. Prior to plot-
ting, the program searches out maximum and minimum intensity values in
the spectrum and allows the user to scale the data as required for
plotting. When plotting %T and absorbance, the user can input a threSh—
old value which helps to reduce the distortion that can result at low
intensities. At each data point the reference signal must exceed the
sample signal by the threshold; otherwise, the point is assigned a
value of 100% T or 0.0 A.

After plotting the desired spectrum, the S/N of the spectrum can
be further enhanced by smoothing the data using quadratic/cubic
Savitzky and Golay convolution functions (103). To limit the amount of
signal distortion, the width of the smoothing function should be no
wider than the full-width-at-half—maximum (FWHM) of the narrowest
spectral feature (101).

The "locate" option uses the graphical input capability of the
terminal to identify the channel number of a specific spectral feature.

The data point of interest is selected with a movable cursor. The

36

program then outputs to the terminal the channel number of the
selected data point and its corresponding values of sample, reference,
%T, and absorbance. To ensure that the cursor did in fact select the
desired point, the user can input the channel number to this routine
and it will draw a vertical line from the baseline to the specified
channel. Each time the "locate" option is called the spectrum is
first plotted on the terminal. The'ENaluate" option can be used to
output the values of sample, reference, ZT, and absorbance for spec-
ified channels without plotting the spectrum on the terminal.

The "write file" option records the spectrum on a user—specified
file-structured storage medium for later plotting with MULPLT (102),
a versatile and extremely powerful plotting package. The "S/N" option
calculates the S/N and background RMS noise level of a spectrum. The
"another file" option allows the user to specify another file for
analysis. "Exit" returns control to the system monitor.

Several input/output conventions were adopted in designing the
VEDO analysis software. All operator prompts that require a yes or
no answer end with a question mark (?). Secondly, all prompts that
require an integer response end with a colon (:). Thirdly, all prompts
that require a real number input, i.e. require a number with a decimal
point, end with an equal sign (=). By coding all of the prompts in
this manner, user interaction with the program is greatly simplified in
that the user is not required to remember the form of the response for
each prompt. Rather, the user only needs to remember three basic
programming conventions.

Examples of emission and absorption spectra recorded with the

37

vidicon spectrometer are given in Figures 3-3 thru 3-5. Figure 3-3 is
the absorption spectrum of PrCl3 in IN HCl. Figure 3-4 is the emission
spectrum of a lead hollow cathode lamp, and Figure 3-5 is the emission

spectrum of the neon filler gas in the hollow cathode lamp.

38

.01

 

 

 

 

 

 

 

 

 

0.8- 442 nm
to H
L)
2 0.6-
<1
00
m 467cm:
° n
(I)
(D 0.4- 482. nm
<
n 5860m
0.2-
O T
I j i _"1
400 450 500 550 600

WAVELENGTH (NM)

Figure 3-3. Absorption Spectrum of PrCl3 (0.20M) in 1M HCl

 

 

 

 

 

 

 

 

39
I600? Pb
406 nm
).
P -
0‘)
2!
m
g.
z
m
:>
_ Pb
'-
4 _ 368 nm
-’ 364 nm
In
t:
W A: WM
(3 l I I 1 I
250 300 350 400 450 500

WAVELENGTH (NM)

Figure 3-4. Emission Spectrum of a Lead Hollow Cathode Lamp

40

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

I000!

>_ .—

’—

U)

:5

LU

h _

2:

DJ

> _

l-

4

.J

.. f

“‘ - f

W b

0

 

1 l l l l
O IOO 200 300 400 500
CHANNEL. NUMBER

Figure 3-5. Emission Spectrum of Neon Filler Gas in a Hollow
Cathode Lamp

CHAPTER 4

CHARACTERIZATION AND OPTIMIZATION OF THE

SILICON VIDICON SPECTROMETER

A. Introduction

The microprocessor—controlled silicon vidicon spectrometer
developed in this research is a powerful investigative tool for the
analytical chemist. Microprocessor control of the instrument
increases its versatility by allowing the scan format to be changed by
software rather than by hardware modifications. Thus the computer can
easily alter the operating parameters to optimize the performance of
the instrument. This chapter describes the experiments used to charac-
terize and optimize the performance of the vidicon spectrometer.

The vidicon detector's multichannel spectral response was examined
to determine its uniformity of response within a selected wavelength
window. Wavelength linearity and resolution were measured and resolu-
tion-limiting factors were identified. The relationships among the
video output, scan rate, and beam current were studied and Optimum
values were established. The response of the detector to incident
light intensity was evaluated, and the linear response region was iden-
tified. Charge integration was studied to evaluate its potential for
signal-to-noise (S/N) enhancement. Next, the electronic background and
random noise components of the video output were studied in order to
determine their effect upon the maximum S/N of the system. Lastly, S/N

enhancement techniques were evaluated.

41

42

B. Multichannel Spectral Response

The RCA C23246 silicon vidicon tube used in this research is a
variant of the RCA 4532 family of silicon-target vidicons with a fused-
silica faceplate for increased ultraviolet response. The active target
area of the vidicon is 12.8 mm (horizontal) by 9.6 mm (vertical). The
reciprocal linear dispersion of the monochromator is 18 nm/mm in the
focal plane. Consequently, the vidicon detector can monitor a 230 nm
wavelength window in the present system. The spectral response curves
in Figure 4-1 and 4-2 (101) show that the vidicon's silicon target
provides usable response from about 250 nm to 1000 nm. Thus the
detector's 230 nm wavelength window can be positioned anywhere within
this broad spectral range. Spectral information within the 230 nm
window is acquired simultaneously by the integrating diode array and
read out sequentially by a scanning electron beam under computer
control.

As different areas of the active target area are illuminated by and
respond to different wavelength resolution elements simultaneously, it
is necessary to measure the uniformity of response across the active
target area. This uniformity was experimentally determined for two
different emission sources; a line source (436 nm emission line from a
mercury lamp), and a continuous source (tungsten lamp, 546 nm).

The channel response of the vidicon detector to a narrow emission
line was measured by monitoring the height of the 436 nm line from a
mercury lamp as it was moved to different positions across the vidicon

target. At each position data acquisition was delayed for several

43

mA/W

RESPONSIVITY
m
l

 

 

' l l l 1
200 400 600 800 I000
WAVELENGTH (NM)

Figure 4-1. Responsivity of RCA C23246 Vidicon vs Wavelength

 

 

 

IO -
a -
Es
z'i
E
94-
u.
u.
uI
2 _
:’ 2
l-
2!
'<
3 I
‘3 K) ‘
8 d
o\°
s _
I I l l
200 400 600 800 IOOO

WAVELENGTH (NM)

Figure 4-2. Quantum Efficiency of RCA C23246 Vidicon vs
Wavelength

44

seconds to ensure that the detector output had reached a steady-state
value. Target exposure and scanning were continuous throughout the
experiment. The resulting response curves are shown in Figure 4—3.
Peak heights, areas, and widths at half-height of the emission line at

selected channels from this curve are listed in Table 4—1. At either

Table 4-1. Channel Response for a Line Source

 

 

Channel # Relative Relative Width at

(512) Peak Height Peak Area Half-Height(nm)
46 542 675 1.62

126 744 783 1.35

199 952 877 1.23

294 974 907 1.14

384 977 924 1.33

480 646 983 2.03

 

end of the target the observed peak profile broadened and the peak
height decreased; however, the corresponding peak area increased as the
emission line was moved from left to right. Busch, Howell, and
Morrison (36) reported the same results using a commercial optical
multichannel analyzer with an RCA 4532 UV-sensitive silicon diode
vidicon. They suggested that part of the observed response could be
accounted for by a lack of focus at the edges of the target. This lack
of focus apparently results from placement of the flat vidicon target
in the curved focal plane of the monochromator.

The channel response of the vidicon detector to a tungsten source

is shown in Figure 4-4. The tungsten spectrum was moved across the

45

mafia E: omq w: 0n umwpmh coowvw> mo mmcoammm Hmccmco .mnq wuowwm

mmmZDZ 4mzz<1o

 

 

 

 

 

oom 00¢ oon CON oo. o
— b P p — p b p b F O
a
a
3
.I
v
:08 u
A
:85: 3
x<me I
I N
I.
3
o N
+ s
41.1 I l
3.5 coo. A
x<me
fee

46

00m

00¢

mousom msoscHucou m cu umwume coowow> mo mmcoammm Hmccmzo

mwmiaz szz<I0

00m CON 00.
p _ p _

 

.eue eeewee

I 000

BAILV-IBH

l 000.

AlmNBLNI

 

—l000_

47

vidicon target and the response at 536 nm was recorded as a function

of channel number. The observed response is linear across the entire
target area. The intensity of the signal increased slightly across the
target (6% higher at the right edge). This response curve suggests
that the response of the silicon diodes to incident light is relatively
uniform across the target.

The data obtained with the mercury and tungsten sources are appar-
ently inconsistent, particularly with respect to the change in peak
area with channel number. At this time it is not known whether the
observed response is due to optical considerations or to the vidicon

itself.

C. Wavelength Linearity and Resolution

1. Linearity of the Wavelength Deflection Circuit Elements

Wavelength scanning in the silicon vidicon spectrometer is accom—
plished by electronic deflection of the scanning electron beam rather
than by more conventional mechanical systems that rotate either a
mirror or grating. Consequently, the linearity of the wavelength axis
is determined by the linearity of the analog deflection circuit
elements. Wavelength linearity was measured by recording the positions
of selected lines from the emission spectrum of a mercury lamp and from
the absorption spectrum of a holmium oxide filter. A least squares
line through the data points had a slope of 0.5109 i 0.0019 nm/channel,
an intercept of 332.9 E 0.8 nm, a linear correlation coefficient of

0.99999, and a standard error of estimate of Y on X of 0.37 nm.

48

The least squares slope and intercept were then used to convert channel
numbers to wavelengths. The results are shown in Table 4—2. The
average wavelength error and percent deviation are 0.29 i 0.19 nm and
0.13 i 0.08 percent respectively. The limiting wavelength resolution
for a 512 line scan is approximately 0.5 nm, so that the observed wave-
length error is a result of wavelength quantization rather than any

nonlinearity in the horizontal scan circuit elements.

2. Limiting Resolution of the Monochromator

The resolution of the vidicon spectrometer reflects its ability
to distinguish adjacent adsorption bands or spectral lines as separate
entities. The resolution is determined by the size and dispersing
characteristics of the grating, the optical system of which the
grating is a part, the slit width of the monochromator, and the
horizontal (wavelength) resolution of the vidicon detector.

The fundamental resolution of a grating is

Rs = mN
where m is the order number and N is the total number of rulings (105).
Thus in the vidicon spectrometer the resolving power of the grating
(Bausch and Lomb certified precision 50 mm grating with 133.6
grooves/mm) is 6680 in the first order. At the mercury doublet wave—
length of 577/579 nm, the smallest interval that can be resoved is
AA = (578 nm)/6680 = 0.09 nm. Since the maximum resolution of the
vidicon detector is 0.5 nm for a 512 line scan, the resolving power of
the grating does not limit the instrumental resolution.

The slit width of the monochromator also affects the resolution of

49

 

 

mo.o- ~.ou w.wem o.eem awe mm
mz.o+ m.o+ m.eem o.enm wee mm
m~.on e.ou e.mem o.eem eae w:
oo.on o.on e.emm e.emm mom om
eo.o+ ~.o+ e.mme «.mme 0mm om
ea.o+ e.o+ e.eee o.eee NNN em
eo.ou a.o- e.mme m.mme HON w:
ma.o+ m.o+ w.wze m.wae mom on
ma.o+ m.o+ o.moe e.eoe Lea mm
om.ou e.ou e.mem m.eem so we
Aeev Aeev Aecv Amamv
Ea Omm\Amsuu<Iono<v N msuuxloamux uamux msuux uwaasz Hwacmzo Esuuomam

 

hufiummafia suwcmam>m3

.N-e eHeee

50

the instrument. The minimum spectral bandpass (width of a spectral
line at half of its full height) that can be obtained with the vidicon
spectrometer is a quantitative measurement of its resolution. This
spectral bandpass is linearly dependent upon the slit width. Although
resolution can be increased by decreasing the slit width, the light-
gathering power of the monochromator and the sensitivity of the
detector place a lower limit on the slit width. Sufficient light must
reach the detector to enable spectral features to be distinguished from
background noise.

Typically a monochromator has two slits; an entrance slit, which
provides a narrow source of light to the dispersion optics to minimize
overlap of monochromatic images, and an exit slit to select a narrow
band of the dispersed spectrum for observation by the detector. In
practice these slits are usually the same width so that the image of
the entrance slit is just passed by the exit slit. However, in the
vidicon spectrometer the exit slit has been replaced by the silicon
target of the vidicon detector, which monitors a 230 nm window. Thus
the limiting resolution will be a function of both the monochromator
entrance slit width and the resolution of the vidicon target.

The experimental resolution of the vidicon spectrometer was
evaluated by measuring the full width at half maximum (FWHM) of the
406 nm emission line from a lead hollow cathode lamp and the 436 nm
emission line from a mercury lamp at a slit width of 25 pm. The
resulting emission spectra are shown in Figures 4-5 and 4-6. The FWHM
for these two lines were 1.26 nm and 1.28 nm respectively. Similar

results have been reported by Ridder and Margerum (78). The

51

I600
.1 Pb

406 nm

I200-

INTENSITY

BOO-

 

Pb

RELATIVE

400-
364 nm 368 nm

 

 

 

L440...“

1 I I T I
250 300 350 400 450 500

WAVELENGTH (NM)

 

Figure 4-5. Emission Spectrum of a Lead Hollow Cathode Lamp

52

 

 

 

 

 

 

 

 

 

800-1
“9
*
507nm
" (254nm)
>"600"
I-
a)
2
m _
P Hg
2
'" 546 nm
400-
II.)
2
I-
‘4 ._
.1
ul
1
200-
H9 H9 *
626nm
- 577"" 579nm (313 nm)
PHJ L—J l___A_fvv_ A—J—~
0 , I l l l
450 500 550 600 650

WAVELENGTH (NM)

Figure 4-6. Emission Spectrum of a Mercury Lamp

(* - second order Hg line)

53

monochromator spectral bandpass at this slit width is 0.45 nm.
Consequently, the resolution of the spectrometer is limited by the
resolution of the vidicon detector itself.

The resolution of the vidicon spectrometer was increased by
decreasing the slit width to 6 um. The FWHM of the 436 nm mercury
line at this slit width was 0.97 nm. However, the peak intensity at
this slit width was approximately one-tenth of the intensity at 25 um.
Below 6 um the peak was not distinguishable from the background noise.
Therefore a slit width of 25 um represents the usable lower limit for
the optical system used in this research. Higher slit widths can be
used, but only with a corresponding decrease in resolution. Resolution
can be increased at the expense of wavelength coverage by using a
grating with a larger resolving power (smaller reciprocal linear
dispersion).

Another indication of an instrument's resolving power is the
degree of separation between two closely spaced spectral lines. Figure
4-6 shows that the vidicon spectrometer attains baseline resolution of
the 577/579 mercury doublet at a slit width of 25 um. This doublet
resolution further supports the observed resolution of 1.2 nm at a

25 um slit width.

3. Resolution of the Vidicon Detector

If the scanning electron beam used for channel readout in the
silicon vidicon had an infinitely narrow width, the resolution of the
vidicon detector would be limited by the packing density of the diode

array (approximately 900 discrete diodes per target width) (3).

54

Unfortunately, the electron beam has a finite width, and it is this
beam width that limits the resolution of the vidicon detector (8).
Since the reciprocal linear dispersion of the monochromator is 18 nmfimn,
the vidicon's limiting spectral resolution of 0.97 nm implies a beam
diameter of 0.054 nm. This value is in excellent agreement with
previously reported beam diameters of 0.062 nm (34) and 0.051 nm (107).
The width of the electron beam was also measured using the
approach of Nieman and Enke (61). The 456 nm emission line of a lead
hollow cathode tube was focused on the center of the vidicon target and
its height was recorded as the number of wavelength channels was varied
from 64 to 512 by powers of two. Since channel exposure time is pro-
portional to the number of channels in the spectrum, the observed
intensity should increase as the number of channels is increased. The
observed response is shown in Figure 4-7. The observed intensity
increased linearly with the number of channels up to 256 channels per
spectrum; however, the intensity decreased when the number of channels
was increased to 512. When the step size between adjacent channels is
smaller than the beam diameter, the beam overlaps a portion of the next
adjacent position so that when this position is sampled a smaller
signal is observed than for no overlap. Using this interpretation of
Nieman and Enke, the experimental beam diameter is less than 1/256 of
the target width (0.050 nm), but substantially greater than 1/512 of
the target width (0.025 nm). Since the vidicon monitors a wavelength
window of 230 nm, the detector's limiting spectral resolution accord-
ing to this approach is approximately 0.90 nm. This value is in
excellent agreement with the observed spectral bandpass of 0.97 nm at

a slit width of 6 um.

55

RESPONSE (VOLTS)
N
I

 

 

I I I
64 I28 256 5l2
NUMBER OF CHANNELS

Figure 4-7. Vidicon Response vs Number of Wavelength Channels

56

This limiting spectral resolution of 0.97 nm applies only to the
center portion of the target surface. The resolution of the vidicon
decreases near the edges of the active target (see Section B). Conse-
quently, the limiting spectral resolution near the edges would be

greater than 1 nm.
D. Analytical Signal

The video signal from the vidicon detector consists of three
parts; the analytical signal, electronic background, and random noise.
These parts are illustrated in Figure 4—8, which shows three overlaid
mercury emission spectra and their corresponding signal-to-noise ratios
(S/N) for the 436 nm mercury line. Spectrum 1 represents 1 scan of the
electron beam across the target, while spectrum 2 represents the
average of 512 individual scans. For spectrum 3, the average of 512
scans of the background signal (shutter closed) was subtracted from the
average of 512 scans of the observed signal (shutter open) to obtain a
background-corrected spectrum. Signal-to-noise ratios for the 436 nm
mercury line were calculated by dividing the peak intensity by the
root-mean-square (RMS) background, which was calculated using 50 signal-
free channels adjacent to the 436 nm emission line.

A comparison of spectrum 1 and spectrum 2 reveals that the effect
of random noise on the signal—to-noise ratio is negligible compared to
the effect of the electronic background, and that there is little
advantage to signal averaging without background subtraction. However,
when the background is subtracted from the video signal, a significant

improvement in S/N is realized through signal averaging. Thus the

57

 

 

 

  
  
  

 

 

 

 

 

 

 

 

40007
436 nm
.II

>-3000-
.—
(I)
Z -
m
.—
Z

2000-

S/N =8490
LLI
>
I— " A
q l
'J i
m S/N=3l8l
a: IOOO" I
* A."
d ; S/N = 2880
O I 1 I I l
O IOO 200 300 400 500

CHANNEL NUMBER

Figure 4-8. Components of the Vidicon Spectrum

58

electronic background is not only reproducible from scan to scan, but
is sucessfully removed from the video signal by subtraction. The
remainder of this section examines the factors affecting the analytical
signal and its optimization. The following section examines the elec—

tronic background and random noise components.
1. Vidicon Detector Operating Voltages

The data sheet for the C23246 vidicon tube specifies a fixed
voltage of +8 volts at the target; however, this requires that the
target be capacitively coupled to the current-to—voltage converter.
Instead, a fixed voltage of -8 volts was applied to the cathode and
the target was connected directly to the summing point of the amplifier.
This arrangement maintains the 8 volt potential between the cathode and
target and permits a direct coupling between the target and amplifier
(10). The remaining tube voltages were maintained at the recommended
levels (see Table 7-1) (104).

The beam current of the vidicon is controlled by a negative
voltage (0 V to -150 V) applied to Grid No. 1 from a high voltage
amplifier (Datel AM301A high voltage, high speed FET 0A with i 140 V
output swing, slew rate of 100 V/us). The high voltage amplifier is
interfaced to the microcomputer via an 8-bit digital-to-analog con—
verter (Datel DAG-29 8-bit DAC with 5 us voltage settling time).

Figure 4-9 shows the relationship between detector response to the 436
nm mercury emission line and Grid No. 1 voltage at a scan rate of 25
kHz. The observed response is a maximum at a grid voltage of -38 volts.

This grid voltage was used for all subsequent measurements at 25 kHz.

59

 

 

600-

500‘
>_ o
I-
” 400-
z
m 0
g.
E

300‘
III
2
I-
<
E: 200-
a:

.
IOO
O 1 1 I I
0 -30 --60 -90 -|20 -l50

GRID #l VOLTAGE (v)

Figure 4-9. Vidicon Response vs Grid #1 Voltage

60

The Optimum beam current was not observed to change with the scan

frequency.

2. Response Near Saturation

The reverse—biased photodiodes in the vidicon target operate in
the charge storage mode and hence are integrating type detectors.
Between scans the reverse bias charge stored on the diodes can be
discharged by either photo—generated charge carriers (light falling on
the diode) or by thermally generated charge carriers (dark current).
Thus the signal level necessary to re—establish reverse bias on the
diodes during the next scan is a measure of the total light intensity
and dark current integrated over the time between scans of the array.
Unfortunately, these diodes have a saturation level (the reverse bias~
charge). If the total charge accumulated on a diode between scans
reaches the reverse bias charge, the diode will saturate. Once satu-
rated, the diode can not acquire additional charge so that all signal
information between the time of saturation and recharge by the electron
beam is lost. As non—linear response was expected near saturation (61),
the response of the detector near saturation was measured to identify
the linear response region.

The response near saturation was studied by measuring the tranamfiw
tance of a neutral density filter (T = 0.54) as the intensity of the
tungsten source was varied. The magnitude of the reference signal (no
filter) was used as a measure of signal intensity. The results of this
experiment are shown in Figure 4-10. The observed non-linearity is 1%

at 90% of saturation. The detector's linear response thus extends to

61

000».

 

 

 

 

cowumusumm Hmwz mmcommmm coowvw> .ofilq muswwm
$5232. 329.... mozmmmemm
coon ooo~ ooo. o
- h p - b - - moo
_ I —
d I
_ $8 I ed
_
— I
S.
v.
m; I no
u_
v.
H.- I
on
uN_ 1 may
_
_ I
_
_ I ad
_
_ .
r o._

 

HBL'IIJ :JO BONVlllWSNVHJ. OBHRSVBW

62

approximately 90% of saturation. In all subsequent experiments the
amplifier gain was adjusted to keep all measurements within this linear

response region.

3. Linearity and Range

The response of the vidicon as a function of intensity should

follow the relationship
Response a IntensityY

where y is the gamma function (3). For the silicon diode vidicon,
gamma is expected to approach unity. The linearity of the silicon
vidicon was experimentally measured by varying the tungsten source
intensity with combinations of neutral density filters and recording
the corresponding A/D converter response. With no filter in the optical
path (100% T), the gain of the data acquisition system was adjusted so
that the converter received a full scale (+10 V) input when the vidicon
target was at approximately 40% of saturation. The value of the
observed converter response versus filter transmittance is given in
Table 4-3. At two points the gain of the data acquisition system was
increased to make use of more bits in the converter's response; all
data were scaled to the original gain setting. The integration time
was the same for all data points. A plot of log (converter response)
versus log (% T) (Figure 4-11) shows that the vidicon's response is
linear over nearly four decades. A linear least squares analysis of
the linear portion of this response curve yielded a slope (y) of
1.025 t 0.005, a linear correlation coefficient of 0.9997, and a

standard error of estimate of Y on X of 0.024. Similar results have

63

cowuocsm ummmcmue couwvw>

mwhdm no

u-
D
db

 

 

.aaue eeemee

HELHBANOO

BSNOdSBH

64

been reported by Milano et. a1. (10).

Table 4-3. Vidicon Linearity and Range

 

 

Transmittance Response Transmittance Response
1.0000 3032 0.0159 84
0.7079 2194 0.0100 60
0.5012 1584 0.0071 48*
0.3548 1175 0.0050 32
0.3162 1122 0.0031 24
0.2238 859 0.0010 8
0.1585 599 0.0007 7
0.1122 447* 0.0005 5
0.1000 387 0.0004 4
0.0708 303 0.0002 2
0.0501 203 0.0001 2
0.0316 159

 

* amplifier gain change

4. Signal Enhancement with Charge Integration

The charge integration capability of the silicon vidicon target
can be used to enhance weak signals by increasing the exposure time.
Exposure time was increased by blanking the electron beam and inhib-
iting the deflection circuits for an integer number of frames (time
required to record one spectrum). The scan circuit was then enabled,
and the accumulated signal read out during one readout frame. The
total number of exposure frames during which the charge accumulates is
equal to the number of integration frames (scan time during which beam
is inhibited) plus one (one readout frame). Intensity versus exposure

frame data for three mercury emission lines (366.3 nm, 404.7 nm, and

65

435.8 nm) are listed in Table 4-4. One exposure frame is equivalent to
an exposure time of 20 ms. A plot of log (signal intensity) versus log
(number of exposure frames) (Figure 4-12) shows that all three curves
have a linear region. The slopes and linear correlation coefficients
of these linear regions are given in Table 4-4. As the accumulated
signals approach saturation, non-linearity occurs. Thus signal
intensity increases linearly with exposure time until the accumulated
signal approaches saturation.

Charge integration integrates all wavelengths evenly. Conse-
quently, this technique is limited by the most intense feature in a
spectrum. Charge integration can be utilized only until the most
intense feature saturates. If integration continues after an intense
region has saturated, the signal from the saturated channels will
spread into adjacent channels, in contrast to linear diode arrays (17).

Since dark current also increases linearly with exposure time, it
places an upper limit on the number of integration frames that can be
used for signal enhancement. For the present system the accumulated
dark signal approaches saturation at an exposure time of approximately
3 seconds (150 frames). Cooling of the silicon diode array would
significantly lower the dark current (77) and permit longer integration

times for increased enhancement capability of weak spectral features.

E. Electronic Background and Random Noise

The two remaining components of the vidicon video output that need
to be characterized are the electronic background and random noise.

These two components are not related to the properties being measured

66

Table 4-4. Signal Enhancement with Charge Integration

 

Number of Signal Intensity (volts)

Exposure Frames

 

 

 

366.3 nm 404.7 nm 435.8 nm
1 .024 .092 0.231
2 .048 .184 0.470
3 .069 .284 0.718
5 .123 .482 1.241*
9 .202 .870* 1.702
17 .386* .354 1.920
Log-log Plot
Slope .978 .031 1.044
Oslope .014 .006 0.008
Linear correlation 0.9996 0.99996 0.99994
coefficient

 

* Last data point included in least-square fit

67

- ---------- SATURATION -----

 
 
 

435.8 nm

 

(D

F-

.J

o C
)>

_I -
<I 0.|
Z -
(D

a) ‘

 

 

(ICU '* I I I I ' I 1
l

NUMBER OF EXPOSURE FRAMES

Figure 4-12. Vidicon Response vs Exposure Time

68

and must be removed from the composite signal. The electronic back-
ground is an operational characteristic of diode arrays and actually
consists of two main components, dark current and fixed pattern noise
(17). The dark current is a relatively constant DC signal and is the
source of the overall level upon which the dynamic video signal is
superimposed. The observed noise across the electronic background
results from fixed pattern noise. Diode-to-diode dark current vari—
ations are the major source of fixed pattern noise and are reproducible
from scan to scan. Random noise includes both photon noise and elec-
tronic noise from the preamplifier and the signal conditioning and
sampling circuits.

Figure 4-13 illustrates the dark signals (shutter closed) recorded
after I scan and after 512 signal averages. (The large noise spike in
the first several channels is caused by the horizontal deflection coil
during retrace.) Any random noise in the system should be a maximum
without signal averaging (1 scan) and significantly reduced after 512
signal averages. A comparison of these two spectra shows that random
noise is quite small compared to the electronic background and that the
electronic background is a relatively constant and reproducible DC
signal. Consequently, the electronic background can be removed from
the composite signal by background subtraction.

A mercury emission spectrum was used to measure the S/N enhance-
ment resulting from background subtraction. The test results are
illustrated in Figure 4-14. For spectrum one, 512 scans were averaged.
For spectrum two, the average of 512 scans of the dark current was

subtracted from spectrum one to yield a background-corrected spectrum.

69

 

 

 

 

 

 

 

 

 

 

F70C>-
).
I-
m
z
w
I-
z
w
2
I-
<
_.I
III
a: 5l2 scans
_ __ - v J
I scan
F _ I"-.. - _fi~
0 " I I I I I
0 l00 200 300 400 500

CHANNEL NUMBER

Figure 4-13. Vidicon Dark Signal

70

 

 

 

 

 

 

 

 

 

 

 

 

 

 

4000-
). 3000‘“I
I'-
(D
Z
LIJ
I'-
Z S/ - 4 0
" 2000- "'8 9
2

. 11 1 1
>
I-
<
_.I
m I000-
m

S/N=2880

I
0 T I I I j
0 IOO 200 300 400 500

CHANNEL NUMBER

Figure 4-14. S/N Enhancement with Background Subtraction

71

Background subtraction not only removed the noise spike from the first

several channels, but also noticably improved the form of the baseline.
This baseline improvement resulted in a significant increase in the S/N
of the 436 nm mercury line.

When sufficient data acquisition time is available, signal aver-
aging can be used to improve the real-time S/N of 3086 (4792 with back-
ground subtraction) by removing random noise from the output of the
vidicon spectrometer. Figure 4-15 illustrates the improvement in
maximum S/N achieved with signal averaging. (Maximum S/N is defined
as the ratio of the maximum linear output level to the root—mean-square
(RMS) value of the observed background, which was calculated using 50
signal-free channels adjacent to the 436 nm mercury emission line.)

The lower trace shows the improvement in S/N without background sub-
traction; the upper trace the improvement in S/N with background sub—
traction. Signal averaging is observed to have limited utility in S/N
enhancement when the background is not subtracted. The electronic
background, which is not affected by signal averaging, limits the
maximum S/N to approximately 4,200. However, the maximum S/N increases
linearly with signal averaging when the background is subtracted. Thus,
when used with background subtraction, signal averaging does success-
fully improve the maximum S/N by removing random noise from the video
signal. Further, Figure 4-15 shows that good S/N data can be obtained
with the vidicon spectrometer without the need for extensive signal
averaging.

Since the data acquisition system is a digital system, quantizing

noise can also limit the S/N of the system (108-111). Quantizing

72

AEOHuomuunsm wcsouwxomn unocuw3 I m

.cowuompupsm ccsouaxomn :uwB I <0 wcwamum>< Hmcwwm Lufiz ucmEmocmscm z\m .mHIq muswwm

owo<mw>< mz<ow ..._0 mums—Dz

 

 

n0. «0. .0. OO—
Pib (L p . _ . P. . b _ . . L .
1
IInn:
o
m I
r
o IIeh:
4‘

 

MIFIVVI XCV'MI

IVI’S

73

noise is a result of the finite resolution of the A/D converter (1
part in 4096 for a 12-bit converter). The effects of quantization on
a signal are most serious when the converter is used to measure low
level signals or when two quantized signals are subtracted to yield a
small result. Consequently, quantizing noise is of concern in the
vidicon spectrometer whenever the background signal is subtracted from
non-signal containing portions of the video output. Fortunately,
quantization effects can be time averaged away if a small amount of
random noise is present on the signal (109). Such is the case in the
present instrument. The random noise randomizes the quantization
error and makes S/N enhancement by signal averaging possible. This
averaging of the quantization error increases the effective bit
resolution of the A/D converter. After 16 averages four extra bits
are accumulated; however, only half of them are significant (108).
Once the quantization error has been randomized, the precision of the
resulting data depends essentially upon the square root of the number
of scans averaged if white noise predominates (110,111). Thus in
calculating the S/N for Figure 4-15 only one half of the extra bits
accumulated through signal averaging were retained.

Finally, for weak signals charge integration (for signal enhance-
ment) and signal averaging (for noise reduction) can be combined to
increase the S/N for a fixed data acquisition time. Figure 4-16 shows
three spectra in which the total data acquisition times were identical,
but the relative times spent on charge integration and signal averag-
ing were varied. The optically attenuated output of a mercury lamp

served as the weak spectral source. Signal-to-noise ratios are given

74

 

 

anh n Z\m

2<um\mw¢:m0axw 0.
ou0<¢w>( mz<um v0

COHUWHdUur—H Odh—wLU mgr—F. HCHNQHO>< HQCde SUM»; UCQEwUCNSCM Z\m

10.0

I o0

 

I 0..

BSNOdSSU

(SlWOA)

 

 

mtn u 2\m

Z<QM\mw¢Dm0uxu '
ou0<¢w>< m2<om wow

31!...

1 «.0

I00

 

F 0.0

I 0.0

BSNOdSBH

(Si-IDA)

.o~I< mpsafim

 

# A

 

 

vmon .I. 2\m

<IIIIILQ

z<0m\w¢:m0&xu _
02.3.93 «:35 VNO.

fiod

I «.0

 

r00

BSNOdSBH

(Sl'IOA)

75

for the 436 nm mercury line. These spectra clearly demonstrate that a
combination of these two S/N enhancement techniques provides a method

for optimizing the output response for a fixed data acquisition time.

CHAPTER 5

DUAL-WAVELENGTH SPECTROSCOPY

A. Introduction to Dual-Wavelength Spectroscopy

Dual-wavelength spectroscopy was first proposed in 1951 by Chance
(112) as a method for measuring small absorbance changes in turbid
biochemical samples. In this method two monochromatic light beams of
different wavelengths are alternately passed through a single cuvette,
and the difference in absorbance at the two wavelengths is recorded as
a function of time. One of the light beams (sample) is set at a wave—
length that corresponds to an absorption peak of the substance under
study. The second (reference) light beam is set at a wavelength at
which the substance under study has little or no absorbance. Variations
in light intensity at the reference wavelength reflect only scattering
effects, if there is no absorbance change due to the substance under
study. Thus, if the relative changes in scattered light are the same
at the two wavelengths, the intensity at the reference wavelength
provides continuous correction for scattering and sample settling.
Consequently, this method allows the measurement of small concentration
differences of single components in mixtures or the monitoring of small
absorbance changes that would normally be hidden in a highly absorbing
or scattering background.

The use of only one cuvette in the dual-wavelength method minimizes
several errors that result from the use of two cells in the classical
(single—wavelength) method. The errors in the dual-cell method result

from differences in cell position, cell path lengths, cell reflectivity

76

77

losses, and from differences between the sample and reference
solutions, such as turbidity and concentration (113). Because these
errorsare minimized in the dual-wavelength method, improved selec-
tivity, sensitivity, and accuracy of analyte absorption measurement
in the presence of severe spectral overlap can be obtained by using
this technique (114,181).

There are two important restrictions that should be observed when
applying the dual-wavelength technique (115). The reference wavelength
should be chosen so as to minimize interferences from other absorbing
components in the sample. Secondly, the two wavelengths should be
closely spaced in order to minimize the effects of light-scattering
differences. The second restriction is a result of the strong inverse
fourth power wavelength dependence of Rayleigh scattering (116).

Of the various optical systems that have been designed for dual—
wavelength spectroscopy, the most common is the time-division multi-
plexed system. In this system the two monochromatic light beams are
alternately passed through the sample and onto a common photodetector.
The output of the photodetector is then amplified, demultiplexed, and
recorded. Vibrating mirrors (112,123,124), rotating chopper mirrors
(113,117,120,125), acousto-optic devices (122), and air-turbine-driven
mirrors (127) have been used to multiplex the two monochromatic light
beams. An obvious advantage to these systems is that only a single
detector is required, and it can be positioned quite close to the cell
to ensure good collection efficiency. The disadvantages include the
mechanical complexity of the chopper and a response time that is

limited by the chopping rate (multiplex frequency).

78

More recently, dual-detector systems have been developed to
provide faster response times for dual-wavelength applications. Chance
et. al. (127) and Stewart (126) combined the two monochromatic light
beams, passed them through the sample, isolated the two beams with a
beam splitter and two interference filters, and detected them simul-
taneously with two photomultipliers. Woodriff and Shrader (128) used
a Rowland circle to construct a dual-wavelength spectrometer with only
one fixed grating and mobile exit slits and photomultiplier light
sensors. Both of these approaches allow the two wavelength channels
to be monitored independently and simultaneously. The primary advan-
tage of this dual-detector approach is that the response speed is
limited by the desired S/N value and not by a multiplex frequency.

There are three basic types of dual-wavelength spectroscopy that
are commonly used; dual-wavelength non-scanning spectroscopy, dual-
wavelength scanning spectroscopy, and derivative spectroscopy (156).

In the dual-wavelength non-scanning mode, the sample and reference
beams are fixed at different wavelengths. This mode is used in those
analytical situations where an interfering substance overlaps the
analyte or when the effect of turbidity must be minimized. The dual-
wavelength scanning mode, in which the reference beam is fixed and the
sample beam is scanned, is used to measure the spectral characteristics
of a highly turbid sample when a suitable reference can not be prepared.
The derivative mode provides the derivative function by scanning the
two monochromators with a fixed small wavelength difference between
them. These three types of dual-wavelength spectroscopy are discussed

further in subsequent sections of this chapter.

79

Compensation of the effects of light scattering by turbid
suspensions is probably the most important use of the dual—wavelength
technique (112,124,125,127,129,130). This technique has also been used
effectively to cancel signal disturbances caused by bubbles and various
shock pulses in temperature-jump and flow experiments and to minimize
the effects of lamp flicker and ripple (131,134,136). Other appli—
cations include spectroelectrochemical studies (138), stopped-flow
measurements (131,134,137), sensitivity enhancement in inorganic
spectrophotometry (113,133), simultaneous determination of mixtures
or masking of diverse components (119-121,128), and multi-wavelength
detection of liquid chromatographic effluents (132) and microwave

atomic emission lines (139).

B. Dual—Wavelength Applications with the Vidicon Spectrometer

1. Introduction

Dual-wavelength techniques are easily implemented with the vidicon
spectrometer because of its multi-wavelength capability. The vidicon
spectrometer is similar to the dual-detector systems in that its wave—
length channels are monitored independently and simultaneously by the
photodiode array. However, as with the single-detector dual-wavelength
systems, the response time of the vidicon spectrometer to time—dependent
features is limited by a multiplex frequency (scan rate of the electron
beam). The normal multichannel output of the vidicon spectrometer
facilitates the selection of a suitable wavelength pair for the appli-
cation of dual-wavelength techniques. In the remainder of this chapter,

the performance of the instrument in three dual-wavelength modes (the

80

dual-wavelength non-scanning mode, the dual-wavelength scanning mode,

and the derivative mode) is evaluated.

2. The Dual—Wavelength Non-Scanning Mode

In this method the radiant power incident on the sample is set at
a wavelength that corresponds to an absorption peak of the substance
under study, and the radiant power incident on the reference is set at
a wavelength at which the substance under study has little or no
absorbance. To the extent that relative changes in scattered light
are the same at the two wavelengths, the intensity at the reference
wavelength provides continuous correction for scattering and sample
settling. This mode is implemented with the vidicon spectrometer by
selecting two suitable wavelengths from an absorption spectrum of the
sample and by using the difference in absorbance at these two wave—
lengths as the analytical signal.

The ability of this technique to compensate for scattering was
evaluated by using neutral density filters to simulate scattering
effects. In this case the turbidity change is the absorbance of the
filter, and, for a good neutral density filter, this absorbance will
be identical at all wavelengths. The spectral feature of interest was
the 536 nm absorption band of a holmium oxide filter (see Figure 5—1).
The sample wavelength was set at 536 nm and the reference wavelength
at 563 nm. The sample and reference cuvettes in the sample compartment
were filled with distilled water, and the holmium oxide filter was
placed in front of the sample cell. Several neutral density filters

were inserted into the light beam and the changes in absorbance were

81

0.5 -I

0.4 -

0.3 -I 536 nm

0.2 -*

ABSORBANCE

 

O.I -

563 nm

 

0.0

 

1
500 550 600
WAVELENGTH(NM)

Figure 5-1. Absorption Spectrum of a Holmium Oxide Filter

82

recorded and tabulated (see Table 5-1). The average AA for these data

is 0.298 t 0.003 (:10) absorbance unit.

Table 5-1. Dual-Wavelength Correction for Scattering

 

 

(Neutral A (536 nm) A (563 nm) AA
Density Filter)
0.00 0.348 0.051 0.297
0.15 0.478 0.178 0.300
0.30 0.611 0.317 0.294
0.50 0.774 0.480 0.294
1.0 1.248 0.948 0.300
2.0 2.065 1.764 0.301

 

The data in Table 5-1 show that insertion of the neutral density
filters into the beam caused changes of approximately 0.003 A (1%) in
AA. Thus, the vidicon spectrometer corrected for 99% of the simulated
"scatter" over a 2 absorbance unit range. The small variation in AA
observed with the different neutral density filters may have resulted
in part from non-uniformity of absorbance of the filters over the
wavelengthsof interest or from random errors in positioning the
filters perpendicular to the beam.

Next, barium sulfate powder was added to the sample cuvette to
simulate a sample with changing light-scattering properties. After
the powder was added, data were collected at five minute intervals
(see Table 5-2). The data for the 536 nm peak show that after a large
initial deflection, the peak began approaching the original absorbance

as the barium sulfate particles settled in the cuvette. In contrast,

 

83

m.o + Nflm.o qu.o oom.o me om

 

o.o A Ham.o eoe.o o-.o oe ma
o.a I wom.o eom.o Noe.o Ne oa
N.m I acm.o ewe.o wwo.~ mm m
o.m I emm.o weo.a ewm.a No a
0.0 A Ham.o o~o.o Hmm.o o o
00H x Ham.o\fimam.o1<<0
I eoeem N << Ase memv < Ase emmv < eeeeuueem Onwaa N see .eafle

 

wcfiHuumm mHaEmm mam wcfiumuumom pow cofiuumuuoo :uwcmaw>m3IHmsa .NIm memH

84

the dual-wavelength data show that the effect of the settling barium
sulfate particles was almost eliminated in this mode. The observed
dual-wavelength error, which increases with increased scattering,
results from the wavelength dependence of scattering (Is~1/A“) (135).
It is a function of both the wavelength separation of the sample and
reference beams and the concentration of the light-scattering compo-
nent. Thus, this error can be reduced by decreasing the sample
turbidity and/or by decreasing the wavelength separation of the two
analytical light beams.

The dual—wavelength non-scanning technique can also be used in the
analysis of highly overlapping two—component mixtures to eliminate the
influence of an interfering component. In this case the two wave-
lengths are chosen such that the absorbance of the interfering compo-
nent at both wavelengths is identical (AA = 0). Also, the sample
wavelength must be at or near an absorbance maximum of the analyte
to provide adequate sensitivity to the analyte. If these conditions
can be met, the analyte can be analyzed directly and independently
of the interfering component.

The determination of nickel in the presence of cobalt was used as
a test of this application of dual-wavelength spectroscopy. The
absorption spectra of the hexaquo Ni(II) ion (0.10 M) and the hexaquo
Co(II) ion (0.17 M) in an approximately 2N perchloric acid solution
are shown in Figure 5-2. These curves demonstrate that an analytical
determination of nickel in the presence of cobalt necessitates the
elimination (masking) of the absorption contribution from cobalt.

The sample wavelength was set to the absorbance maximum for the

85

 

 

 

0.8-I
CO+2
0.6-
u;
I:
:z
< -
a: (14
(r Ni+2
c:
U)
a:
< 1
01!-
01) I I I I I I I
380 460 540 620

WAVELENGTH (NM)

Figure 5-2. Absorption Spectra of Nickel (0.10M) and Cobalt (0.17M)
in 2N 11010.+

86

hexaquo Ni(II) ion, 397 nm (Figure 5-2). Since the absorbance of the
hexaquo Co(II) ion at 610 nm minimized changes in the difference of
absorbance with changes in cobalt concentration, the reference wave-
length was set at 610 nm.

Solutions were prepared by weighing out appropriate amounts of
the metal salts (cobalt chloride and nickel nitrate), by dissolving
them in 5 ml of concentrated (approximately 70%) perchloric acid, and
by diluting to 25 ml with distilled water. The absorbance of each
solution was then measured at 397 nm and 610 nm and the absorbance
difference (A397-A610) was used as the analytical signal. Figure 5-3
illustrates the linear calibration curve (extrapolated to zero
absorbance) that resulted from the measurement of the absorbance
difference at the two wavelengths for varying amounts of Ni+2. Table
5-3 lists the results obtained when 3 mg/ml of nickel were measured
in the presence of varying amounts of cobalt. The data clearly show
that over a cobalt concentration range of 0 to 15 mg/ml the dual—

wavelength technique can successfully minimize the interference of

Table 5—3. Determination of Nickel in the Presence of Cobalt

 

Ni taken Co added Ni found Absolute Error

(ma/ml) (ms/ml) (ms/ml) (mg/m1)

 

3.0 0 3.0 0.0
3.0 3 3.0 0.0
3.0 5 3.1 + 0.1
3.0 10 3.2 + 0.2

3.0 15 3.3 + 0.3

 

50wumcwEumqu meuwz ~00 m>p30 cowumunwamo .mlm wunwwm

2235 555.2 .10 zo_e<mezmozoo

87

P

v
.

m
.

N
.

006

Iwod

r000

INEO

12.0

 

WONO

(“JUOISIV - (WULGQIV

88

cobalt in the determination of nickel. The interference of cobalt
was not completely eliminated because of a small absorbance difference

between the two cobalt wavelengths.

3. The Dual-Wavelength Scanning Mode

The dual—wavelength scanning mode utilizes the same beam geometry
as the dual-wavelength non-scanning mode. In the dual-wavelength
scanning mode, however, the reference wavelength is stationary and the
sample wavelength is scanned over a short wavelength interval. This
mode permits the acquisition of spectral information, while retaining
the advantages of dual-wavelength Operation. Sellers et. al. (135)
used this technique to study the reduction of a microsomal suspension
in a turbid sample.

The dual-wavelength scanning mode can be implemented with the
vidicon spectrometer by acquiring an absorption spectrum and then
subtracting the absorbance at the reference wavelength from the absorp-
tion spectrum. As in the previous section, barium sulfate powder was
used to simulate a sample with changing light—scattering properties
caused by sample settling. A holmium oxide filter was placed in front
of the sample cuvette and barium sulfate was added to the water in the
cuvette. After the powder was added, the peak was scanned at five
minute intervals. The resulting curves are shown in Figure 5—4 for the
single-wavelength mode. After a large initial deflection, the peak
began approaching the original trace as the barium sulfate particles
settled in the cuvette.

The corresponding curves for the dual-wavelength scanning mode

89

l.5-Ir

‘
‘ I «IIIIOIO
qr ‘—
I.0 '—
" 5 mlnmu
‘h A ;.—‘

P IO mine!“
.5 d1- l5 mine!"
4
b

 

 

  
 
 

 

 

ABSORBANCE

 

20 IIIIIIOIOO

 

I D o IIIIIIIII

 

‘1.

O A A l L A A A
I I

A
v v I 1 W T I

500 '560 600
WAVELENGTH (NM)

Figure 5—4. Effect of Sample Settling with Single-
Wavelength Scanning

.4
“I ~3 I
o
2
d
a
C: .2
o»
OI
ll
Id

 

o ans===2g/ ‘~ .a—_aa—s

500 550 600
WAVELENGTH (NM)

Figure 5-5. Dual-Wavelength Correction for Sample
Settling

90

(Figure 5-5) show that the dual-wavelength scans are almost fully
superimposed, and the effect of the settling barium sulfate particles
was minimized. Unfortunately, this technique was not able to elimi-
nate totally the effect of the sample settling because of the wave—
length dependence of scattering (135). The observed correction error
increases both with an increase in sample scattering and with an
increase in the wavelength interval between the reference and sample

beams.

4. The Derivative Mode

In the normal derivative mode, the beam geometry is identical to
that of the dual-wavelength non-scanning mode, and a single cuvette is
used. With single channel detectors, two monochromators are mechani-
cally locked together at a fixed small wavelength difference. The
monochromators are then simultaneously scanned across the wavelength
interval of interest to produce a differential absorption curve.
Sellers et. al. (135) found that a wavelength difference of 2 nm with
a 2 nm bandpass results in differential absorption curves that closely
approach the theoretical first derivative of a spectrum.

In 1955 Giese and French (140) published the results of their
study of the behavior of the derivatives of hypothetical overlapping
absorption bands. This paper clearly demonstrated the advantages in
accuracy that could be gained by using derivative measurements in
resolving overlapping bands. Advantages of the first-derivative of
absorbance include the following: detection of absorption bands even

if they completely overlap, resolution of small absorption bands

91

which lie on the steep portion of an absorption curve, precise location
of very broad absorption bands, linear relationship of the differential
coefficients at specific wavelengths to the concentration of the
corresponding components, and the elimination of variations in turbid-
ity or background.

The experimental approaches used to obtain derivative spectra fall
into two general categories: those that operate on the light beam in
the optical part of the spectrometer, and those that operate on the
output signal of the spectrometer. The first category includes dual—
wavelength spectrometry (112-114,118) and wavelength modulation with
synchronous detection (142-145). The second category includes elec-
tronic differentiation (146-148,153), mechanical tachometers (149,150),
and numerical differentiation (66,82,103,151,152,154).

In recent years the increased availability of on-line computer
systems has made the use of numerical differentiation for producing
derivatives of spectroscopic data more common. Once the spectroscopic
data are in digital form, numerous algorithms are available for smooth-
ing, averaging, or differentiating the original data (see references
above). In order to use these techniques, the experimental curves
must be continuous and the data must be sampled at uniform wavelength
intervals (155). The digital output of the vidicon spectrometer meets
these requirements and is well suited to the various post-processing
techniques currently available.

The vidicon spectrometer simultaneously collects spectral
information over a 230 nm window and outputs a user-defined number of

uniformly spaced data points. These data can be subjected to any of

92

the various numerical differentiation techniques to calculate the

first derivative. Pardue et. al. have used both Savitzky and Golay
smoothing functions (66) and wavelength modulation of the horizontal
sweep signal of a vidicon spectrometer (60,81) to generate the first
derivative for atomic emission and molecular absorption processes.
However, since the first derivative represents the slope at each data
point, the first derivative of the vidicon output can be approximated
by taking the difference between data points equally distant from the
data point at which the slope is to be determined, e.g. the slope at
the ith data point is equal to (i+n) - (1-n), where n is a user-
specified integer. The value of n depends upon both the number of out-
put wavelength channels and the form of the experimental curve. The
first derivative of the acquired spectral data is obtained by repeating
this calculation for all of the data points. Higher derivatives can

be obtained by repeating this calculation the appropriate number of
times.

This subtraction technique was qualitatively evaluated by calcu-
lating the first and second derivatives of the absorption spectrum of
praseodymium chloride (Figures 5-6 thru 5-8) and the emission spectrum
from a mercury lamp (Figures 5-9 thru 5-11). For each spectrum the
number of wavelength channels was 512 and the integer increment for
the derivative calculation was 2, i.e. the slope was calculated over a
2 nm interval at each point. The absorption spectrum of PrCl3 (Figure
5-6) illustrates the high quality and resolution capability of the
vidicon spectrometer. Figures 5—7 and 5-8 show the increased resolu-

tion obtained from the first and second derivative curves. Notice in

93

 

 

 

 

 

 

 

 

 

 

 

0.8“ 442 nm
DJ
C)
2 0.6
<1
00
m 467nm
CD
0)
m 0'4- 482. nm
<1

4 586 nm
0.2- \J
0 I h T‘—
400 450 500 550 600

WAVELENGTH (NM)

Figure 5-6. Absorption Spectrum of PrCl3 (0.20M) in 1H HCl

94
ISO '-

l20 -

BO-

 

40'I

INTENSITY

 

 

 

 

-BO - L

RELATIVE

 

'IZO "

 

 

‘IGO

 

I I T 1
400 450 500 550 600
WAVELENGTH (NM)

Figure 5-7. First Derivative of PrCl3 Absorption Spectrum

 

 

 

 

 

 

 

 

 

IOO 3
50 "
)-
* I
a:
z 0 “
m
.—
z
-50 I
U
>
- 'IOO -
p.
<
4
w
1 ‘I50 -*
'200 T I 1 ‘1
400 450 500 550 600

WAVELENGTH (NM)

Figure 5-8. Second Derivative of PrCl3 Absorption Spectrum

95

 

 

 

 

 

 

 

 

 

4000 -
436 nm
>.
I'—
— — 507nm
(I) (254nm)
2!
DJ
[.—
3 546 mu
m -
>
'—
<1
_I 405nm
DJ
E ‘
366nm 577/5n7m9
() I I I I “I
350 400 450 500 550 600

WAVELENGTH (NM)

Figure 5-9. Mercury Emission Spectrum

96

 

 

 

 

 

 

 

 

330 I
>' u
I- |65
In
2
MI
.- I
E 0 q,_J I I J I A

I F T

U
>
.-
<
_, -I65 '-
IL!
I

'330 I T I l I

350 400 450 500 550 600

WAVELENGTH (NM)

Figure 5-10. First Derivative of Mercury Emission Spectrum

 

 

 

 

 

 

300 '1

I50 ‘ I
> . I II 1
.- 0 «I v ’—
5 1
2
U
I— 450 - ‘50-
z
N -300 “
Z
.—
‘
A - -
u 450
K

‘600 ‘

fl I f I —
350 400 450 500 550 600

WAVELENGTH (NM)

Figure 5-11. Second Derivative of Mercury Emission Spectrum

97

particular the resolution of the shoulder on the 586 nm absorption
peak in the derivative curves. The mercury emission spectrum of
Figure 5-9 further illustrates the excellent resolution obtained with
a line spectrum. Although the vidicon spectrometer can not resolve
the mercury 577/579 nm doublet when it is incident upon the target
edge, the second derivative (Figure 5-11) clearly indicates the
presence of two peaks in the doublet. Thus this numerical differen-
tiation technique can be used to significantly improve the resolution
of spectroscopic data.

One of the reported advantages of the first derivative method is
the elimination of variations in turbidity or background (155). Also,
because the wavelength channels used in evaluating the first derivative
are only a few nanometers apart, the previously observed error due to
the wavelength dependency of scattering should be essentially elimi-
nated. The ability of the derivative mode to eliminate variations in
turbidity and sample settling was evaluated by applying this technique
to the scattering data reported earlier in this chapter. Figure 5-12
illustrates the results obtained when the derivative mode was used to
correct for the "scatter" caused by insertion of several neutral
density filters into the sample light path. Figure 5—13 illustrates
the results obtained when the derivative mode was used to compensate
for the effect of the settling barium sulfate particles in the sample
cuvette. In both instances the derivative mode successfully eliminated
the effect of variations in scattering in the sample. The derivative
mode also eliminated the previously observed error that resulted from

the wavelength dependency of scattering.

98

 

 

 

ICC)?
).
I:
In
5
o
E A—r—v—* o
50.
III
2
I-
<
.J
III
a:
O l —1
O I 2

ABSORBANCE OF FflTER

Figure 5-12. Intensity of First Derivative Signal vs
Absorbance of Neutral Density Filter

 

 

I00
, l
I:
(D
2
DJ
.—
E
50 -
[LI
2
'2
.J
ILI
a:
0 | I
0 IO 20

TIME (MINUTES)

Figure 5-13. Intensity of First Derivative Signal vs
Time During Sample Settling

99

C. Conclusions

Because of the multi-wavelength capability of the vidicon
detector, successful implementation of the various dual-wavelength
techniques on the vidicon spectrometer required only changes in the
data processing software. No modifications were required in either
the data acquisition hardware or software. Thus, the dual—wavelength
techniques form powerful post-processing options that can be matched
to specific data analysis requirements. This result is another
excellent example of the power and versatility of the silicon vidicon

tube as a spectroscopic detector in the UV-visible spectral region.

CHAPTER 6

COMPUTER SYSTEM ARCHITECTURE

A. Introduction

As microprocessor costs continue to fall and system designers
become more familiar with them, microprocessors are being incorporated
into many types of multiple processor systems (157-166). Advantages of
multiple processor systems over single processors include higher
throughput, improved real-time response, better system reliability, and
modular expansion (158). At this point in time, however, the struc-
tures and organizations of multiprocessor systems have not been well
defined. This is particularly true with respect to the implementation
of multiple processor systems with the relatively new large-scale inte-
grated (LSI) microprocessors. Current areas of research and develop-
ment include task allocation, efficient use of system resources,
sequencing and interaction between processors, and overcoming the
physical and architectural limits of the microprocessors.

The first section of this chapter describes the logical develop-
ment of a general purpose multiprocessor system based upon the Intel
8080A Microprocessor. The development of this system not only
illustrates the increased system capabilities of a multiprocessor
system over those of a single processor system, but also provides
design guidelines for an efficient and modular multiprocessor system.
The second part of the chapter describes the application of the

multiprocessor system to the real—time control of the vidicon

100

101

spectrometer. In documenting the computer system used to control the
vidicon spectrometer, specific examples of task allocation, efficient
use of system resources, and utilization of the inherent bandwidth of

available microprocessor static memory are provided.

B. Development of a Multiprocessor System

At the start of this research project it was not immediately
obvious that a multiprocessor system was required for real-time control
of the vidicon spectrometer. Rather, the need for a multiprocessor
system slowly evolved during the design and specification of a suit-
able computer system for real-time instrument control. Although
several single processor system configurations were considered, each
one was subsequently rejected because of one or more performance
limitations. Only the multiprocessor system was able to meet satisfac-
torily the design goals of standardization of design, software and
hardware modularity, ease of expansion in both hardware and software,

efficient use of system resources, and simplified programming.

1. Dedicated Minicomputer System

The first computer implementation considered was a dedicated mini-
computer system (Figure 6—1). This system was based upon a Digital
Equipment Corporation PDP 11/40 minicomputer with 32K of core memory.
Standard peripherals include an extended instruction set (KEll-E),
floating instruction set (KEll-F),Printronix Model 300 Line Printer
(LP), Sykes Mbdel 7250 Dual Flexible Disc Drive (FL), RKOS Decpack

(DSC), and a Tektronix 4010-1 Computer Display Terminal (TRM). Also,

102

Houucou ucwssuumcH you Housmaooficwz wmumowwoo .~I© muowfim

 

thmzamkmz_

 

. 2a.. @

 

—bzm2:mkm2_

 

we. 99
e.

103

a digital plotter (PLT) has been designed and interfaced to the
minicomputer by workers in our laboratory (167). In this configuration
one instrument at a time is operated through the minicomputer via its
instrument-specific interface.

The dedicated minicomputer system does provide on-line real-time
control of the instrument to which it is interfaced. While it is
active, the instrument has the full attention of the minicomputer and
its peripherals. There are, however, several distinct disadvantages to
this system. Only one instrument can be active at one time. (Real
time needs of instruments would be too demanding for time-share opera-
tion of the 11/40.) While the available response time and data acqui-
sition speed capability of the mini may be used to service a specific
instrument, effective use would not be made of the mini's computa-
tional capacity. Clearly, such a configuration seriously under-
utilizes the power of the mini and gives access to the peripherals
only to the current user. A dedicated minicomputer system also
requires an expensive minimal system, i.e. the minicomputer and its
expensive peripherals, and different control logic for each subsystem.
It also does not satisfy the design goal of modularity in software and
hardware. As a result of these disadvantages, the dedicated mini-

computer system approach was rejected.

2. Distributed System with Real—Time Microprocessor Control

The second system configuration that was considered is shown in

Figure 6-2. In this system design a dedicated microcomputer is directly

104

 

h2m231hmz.

 

 

Hzmznmhmz

   

 

Houucou ucmaauumcH pow nousasooouowz wmumowvwa .mlo muswwm

 

Jomhzoo
.232200

 

 

 

105

interfaced to each instrument for localized real-time control. This
configuration relieves the microsecond response requirement from the
mini and, as a result, several instruments can be active at one time.
Also, the minicomputer's expensive peripherals can be shared by several
instruments in this configuration. A communication controller (99) is
required to control the parallel communication link between the mini-
computer and microcomputers and to resolve all bus conflicts.

Although this second system design is an improvement over the
dedicated minicomputer system from the viewpoint of effective mini-
computer utilization, it has several disadvantages with respect to
instrument support. In the computer-controlled systems we are devel-
oping the microprocessor has to assume a variety of different tasks.
This gives rise to two undesirable consequences; the interleaving of
tasks makes hardware and software modularity difficult, and the micro-
processor architecture (hardware and software) tends to mimic mini-
computer architecture (i.e. large memory, advanced operating system,
etc.) (159). Also, in those applications in which a fast minicomputer
can barely keep up with the real-time control of an instrument, a
single micro with its slower cycle time can not possibly maintain
real-time instrument control. Consequently, this system configuration

was also rejected due to the several obvious performance limitations.
3. Multiple Microprocessor System for Real-Time Control

The third and final system configuration that was considered is
shown in Figure 6-3. In this design a multiprocessor system is used
for real-time instrument control and data acquisition. This system is

H H
based on one "master" (coordinator) processor and two or more slave

106

Houucoo acmesuumcH you Eoumhm uommmooumwuasz wmuonwuumwm

_»zu2:m»mz__

 

 

coco

 

 

Jomhzoo
.232200

 

 

ooo

 

szwzanmZ_

 

 

.mno muswam

107

processors in a hierarchical relationship. The coordinator processor,
which contains the system executive, controls or supervises the opera-
tion of the slave processors, which perform dedicated functions as part
of the single partitioned system. This type of multiple processor
system, often called a distributed system (157), uses a static alloca-
tion of tasks. Although a static allocation of tasks allows very
simple, specialized program modules to be developed for each slave
micro, this type of allocation requires that the tasks and their
actions be known during system development so that the system process-
ing load can be balanced.

One of the most powerful advantages of this system is its ability
to execute logically divisible tasks simultaneously for subsequent
correlation. This simultaneous execution of tasks significantly
increases system speed and throughput. Other important advantages of
the distributed system include standardization of design, modularity
in both hardware and software, ease of expansion in both hardware and
software, simplified programming, and the fastest possible response to
multiple, asynchronous, real-time tasks (159).

The distributed system does have two major disadvantages. If the
tasks and their actions are not fully known during system development,
inefficient utilization of the processors and awkward segmentation of
the system software can result. A second disadvantage of the static
allocation of tasks is that the failure of any processor seriously
degrades or even terminates system performance. The dedicated system
can not dynamically shift tasks that have been assigned to a defective
processor.

The advantages of the distributed system for real-time instrument

108

control far outweigh the disadvantages of the system. Consequently,
this system was subsequently developed in our laboratory and applied to
the real-time control of the vidicon spectrometer. However, before
actually designing the system it was necessary to define the bus

structure that would be used for system communication.

4. Interprocessor Communication

The majority of multiple processor systems that have been
reported in the scientific literature have used a combination and/or
modification of the five basic interconnection methods. These basic
methods are the dedicated bus, multiporting, a single time-shared bus,
multiple time—shared busses, and a crossbar matrix (157). Great care
must be taken in selecting one of these bus structures for a partic-
ular application because the communication link used in a computer
system is one of the most crucial factors affecting system performance.

The crossbar matrix is a method of connecting N active elements
(processors and input/output elements) to M passive memory elements.
Multiple simultaneous connections are possible if they are mutually
exclusive. The main advantage of this structure is its high through-
put. Unfortunately, the switch matrix that controls the M x N cross-
points is necessarily complex, costly, and physically large. This
structure also has a limited expandability because of the complexity
of the switch matrix. The complexity and limited expandability of this
structure eliminated it as a feasible communication link for our
multiple processor system.

The multiported system uses multiple dedicated busses for

communication. As illustrated in Figure 6-4, each passive element

 

 

 

Figure 6—4. Multiported System

Q5 <9 (I? [323:2

Figure 6-5. Single Time-Shared Bus System

 

 

 

110

(memory) has several ports, each tied to a different processor.
Although the throughput of this arrangement is excellent, each passive
device requires an arbiter to acknowledge requests from the multiple
processing elements. In addition, the cabling and connector costs for
a multiported system are large. The memory interface circuitry
required also limits the expandability of this system. This system
was rejected because of its complexity and cost. The multiple time-
shared busses approach was also rejected for the same reasons.

In the single time-shared bus approach (Figure 6—5) the processors
use common memory accessible on a common bus for instructions and data.
Since the bus is a shared resource, an arbiter must be provided to
resolve bus conflicts. The interference between processors requesting
use of the bus is dependent upon the microprocessor instruction execu-
tion time, memory cycle time, and the number of processors sharing the
bus. Advantages of a single time-shared bus include minimal intercon—
nection costs and a high degree of modularity. Although it is very
easy to expand this interconnection system, the addition of devices to
the system bus increases bus contention. This increased bus contention
and subsequent allocation overhead cause a corresponding decrease in
individual processor throughput. Also, since the bus is the only data
path in the system, a failure of the bus will result in loss of the
total system. Consequently, this interconnection method was also

rejected.

5. Dedicated Multiprocessor Bus System

A dedicated bus is either a permanent link between two devices or

a bus used solely for one function (i.e. an address or data bus). It

111

is the least complex and easiest to configure of the five basic inter-
connection methods. Figure 6—6 illustrates an example of a dedicated
bus system in which the bus serves as a communication link between the
coordinator processor and the slave processors. In such a system each
processor has its own local memory and can operate independently.

Figure 6-7 illustrates the dedicated bus system that was adopted
for interprocessor communication in the multiprocessor system used to
control the vidicon spectrometer. This method is a modification of
the dedicated bus system in which the intercommunication bus is an
extension of the coordinator's bus. A "bus switch" (98) is used to
switch the local busses of the slave processors onto the coordinator's
bus.

A block diagram of the multiprocessor bus system (Figure 6-8)
shows the backplane bus (extension of the coordinator's bus) and three
modules connected to the backplane; the coordinator and two slaves.

The backplane bus includes a 16-line address bus, an 8-line data bus,

 

and four control lines (MEMR, MEMW, DBIN, and RDY). In this system the
slaves are fully functional microcomputers capable of independent
operation; however, a user-defined portion of each slave's memory and
task space resides in the address space of the coordinator. The
coordinator thus has the capability to start and stop the slave
processors, read and/or write status words to them, instruct them to
begin execution of different portions of their resident programs,
reprogram them, or even take complete control of a slave's memory and
task.

If the coordinator wishes to take over control of the memory or

task space of one of the slave processors, it simply sends the address

112
T
T o o T
COORD
P

T T

Figure 6-6. Dedicated Bus System

T T T

coono
11

Figure 6-7. Dedicated Bus with Bus Switch

113

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  

 

F a 1
8 COO” [MEMORY]
A
c
K
P A 05000:)- --
L
A
2 SLfiVE[ [-MEMORYI TASK
A 05000:} --
43 .
S lMEMORYj- .ASK
V
/ 28

 

Figure 6-8. Block Diagram of Multiprocessor Bus System

114

of that slave's memory or task space down the backplane. The address
decoding circuitry on the slave processor board recognizes its address
and informs the processor that the coordinator desires control of the
local bus. Upon completion of its current cycle, the slave processor
goes on hold, and the address decoding circuit transfers (switches)
control of the slave's bus to the coordinator. At this point the
coordinator has complete control over the slave's memory and task.
When the coordinator has finished with the slave's memory and task, it
releases that address space and the address decoding circuit returns
control of the memory and task to the slave.

This mode of cummunication works very well when communication
between the processors does not need to exceed a speed of approximately
50 kilobytes/second, which is the maximum memory transfer rate achiev-
able under processor (8080A) control. When faster communication is
desired in the system, direct memory transfer (DMA) circuits must be

used.

C. Multiprocessor Control of the Vidicon Spectrometer

As illustrated in Figure 6-9, the multiprocessor system has been
used for real-time control of the vidicon spectrometer. This system
requires three processors; a coordinator micro, and two slave micros.
The coordinator micro supervises the operation of the slave processors
and coordinates all activities in the multiprocessor system. The
communication microprocessor at the top of the figure controls communi-
cation with the system peripherals; CRT monitor, ASCII keyboard,

floppy disc, and communication link to the minicomputer. The vidicon

15

Emum%m umuoEouuooam coofivw> mo Emuwmwa xooam .olc muowwm

 

_uo<umm»zf
_ zoos;

 

 

 

Tl?

  

$ cudqomhzoo

    
   

 

 

 

 

 

 

 

 

 

 

 

mung .53..
- c I
@2752; oz< a? :85: 53 :35:
ez_zo_:ozoo ._<zo_m Iv >522 ES. 2332:. m><._m
8:8 “smegma <

 

 

 

.7235 _ zo_._< _

=. zoos; 7393 w 1
-_ $52.6 38%|. 752mg.— 268
u4a2<m _
x23 , T oma - F3335 , - a

_ 0.; .2_ T3:— :.._ Tea: .
I, \. - - T - I
I: .l 356m .
\_ /

V
H— b .... .0883 <

 

 

 

 

 

 

I

 

 

 

 

 

 

 

 

 

CD<0¥QJ<2LU

116

microprocessor at the bottom of the figure controls the spectrometer;
monochromator wavelength drive, sample cell module, Optical shutter,
vidicon scan circuits, and the data acquisition system.

The three microprocessors and their dedicated I/O interfaces are
contained in a standard BUD card rack (168). The card rack has 12
card slots with 58-pair connectors. A backplane was soldered to the
lower 40 pair of connectors to buss them across the width of the rack.
The upper 18 pair of connectors were not bussed, but instead were left
open for wire wrapping. A typical micro card is a 6 x 8 inch double-
sided wire wrap board with holes on 0.1 inch centers. A block diagram
of a microprocessor board, the vidicon spectrometer, is included in
Chapter 7. The rack power supply was built by the Michigan State
University Electronics Shop (169).

All of the micros have their own local memory for the storage of
data and task-specific program modules. In addition, a user-defined
portion of the memory and task space of the two slave micros resides in
the address space of the coordinator for control and communication. A
new software package, CONVERS (100), was modified to run on a multi-
processor system (98) and was subsequently used as the software
operating system. The executive for this system resides in the
coordinator micro.

The communication micro controls communication between the multi-
micro system and the operator (CRT/keyboard), local bulk storage
(floppy disc), and the 11/40. Memory mapping is used to interface
these peripherals to the communication micro. The primary advantage of

the memory mapping technique is that all instructions that operate on

117

memory locations can be used in I/O, i.e. the accumulator is not the
only transfer medium in memory mapped I/O.

An Electronic Warehouse Model EW-100 ASCII Encoded Keyboard (170)
provides operator input to the multiprocessor system. It is interfaced
to the bus of the communication micro via an Intel 8212 8—bit I/O Port
(174). The display system consists of a Ball 9 inch CRT Data Display
(172), a Matrox MTX-16328L 16 x 32 ASCII Character Video Random Access
Memory for alphanumeric capability, and a Matrox MTX-2562-1 Graphic
Display (171) for graphic capability. The two Matrox units interface
directly to the micro bus as memory units. They provide composite
video, TTL video, horizontal synchronization, and vertical synchroni-
zation outputs that can be used to drive a CRT display. The MTX-163ZSL
is a slave unit that uses the synchronization outputs from the MTX-
2562-1 unit to ensure proper synchronization between the alphanumeric
and graphic displays. The Persai Model 1070 Intelligent Diskette
Controller (173), which is based upon a dedicated 8080 microprocessor,
also interfaces directly to the communicator's bus. The diskette
controller has a factory interface for the Model 270 Dual Diskette
Drive (173) and a PROM-resident file management program that greatly
simplifies use of the floppy disc unit for information storage and
retrieval. Communication between the multimicro system and the PDP
11/40 is provided via a parallel link designed by S. Hourdakis (99).
An MDB-l710 General Purpose Interface Module (175) was used to inter-
face the communication link to the minicomputer.

The vidicon microprocessor at the bottom of Figure 6-9 provides

real-time control fo the vidicon spectrometer, including;

118

GCA/McPherson Alternating Cell Module Model EU-721-11 (176),
GCA/McPherson EU-700 Czerny-Turner Monochromator (176), optical
shutter, vidicon scan circuits, and data acquisition system.
Documentation on the design details of the spectrometer circuits
are presented in Chapter 7. Chapter 3 presents an overview on

the routine operation of the vidicon spectrometer.

CHAPTER 7

DOCUMENTATION OF SYSTEM DESIGN

A. Introduction

The purpose of this chapter is to provide documentation of the
design details of the electronic, mechanical, and optical components
of the silicon vidicon spectrometer. This documentation may be used
either to duplicate the instrument or to compare the results obtained
with it to the results obtained with similar instruments. Additional
documentation is included in Appendix B to aid in the repair or modi-
fication of the vidicon spectrometer.

An overview of the complete vidicon spectrometer system was
presented in Chapter 6. In contrast, this chapter discusses in detail
the vidicon slave micro, the electronic circuits used to control the
vidicon scan format, and the data acquisition system. A block diagram
of the vidicon circuits discussed in this chapter is shown in Figure

7-1.

B. Scan Control Micro

1. Introduction

The primary function of the scan control micro is to provide real-
time control of the vidicon detector's scan format. This section
describes the design of the scan micro and the scan circuits. Details
of the micro's real-time clock and interrupt circuit, a dedicated

microprocessor for number processing, and the shutter control circuit

119

120

muwsouwo Houucou coowww> mo Emuwmwo xooam .Hln muswfim

 

j

 

— uo<u¢mhz.

H zoo_o_>

 

 

 

 

H

 

$2.4114m 02¢

ox<
2:22:33 .3206

 

 

mint: _ 22.2 _
zoos; _ «noon.—

mtsoEo 255 T

 

 

 

CNJJOKFZOU
amend Pm<u

 

 

>¢02m2 <.—.<O >¢OZNE :
\muon: hm<u 2<¢oomm m><4m

 

 

 

TIT

 

 

x0040 mmzoznmo
moth -4dmx KNDZDZ

 

 

121

are also included in this section.

2. Slave Micro Design

A block diagram of the scan control micro is shown in Figure 7-2.
The basic micro consists of an Intel 8080A Microprocessor (single chip
8-bit N-channel microprocessor with 2 us instruction cycle time), 8224
Clock Generator (controlled by an 18.0000 MHz crystal), 8228 System
Controller, and 2102A—4 Static RAM (IX X 1) (174). A bus driver is
also required to interface the static memory's data out lines to the
microprocessor's 8-bit bi-directional data bus.

In order to utilize the full 64K memory address capability of the
8080 microprocessor, an Intel 8205 1-of-8 Decoder (C1) was used to
decode address lines A13-A15. When C1 is connected as shown in Figure
7-2, an active low signal that is exclusive in nature and that repre-
sents the value of the input address lines is available at one of the
C1 outputs. Thus C1 provides eight 8K "chip selects". CZ is used to
further decode the lowest 8K of memory space into eight 1K memory
"chip selects" for the local memory (2102A—4 1K static RAM).

The control, data, and address lines on the left side of Figure
7-2 provide communication between the scan control micro and the coor-
dinator micro. These signals originate from the coordinator micro and
permit it to reset the slave micro, start or stop program execution,
read or write status words to the slave, or even completely reprogram
the slave micro.

Communication between the coordinator and the scan control micro

involves the transfer of control of the slave's local bus to the

122

ouofiz w>me cooHvH> .Nln muswfim

n_<
E<
n_<
ado:
12w!

 

>o<wz .ozooo

Pomamm om<om
wo<umm._.z_

hounfim
>m0§m2 <h<o

 

32w!

 

N2 uo<

2.00
JOKPZOU hmmDowm 040:

  
 

salon

  

wwmmooa. <._.z_

._.z_

 

mam <53
zuwam
>n<uz

\ knit

hum“:
dzooo

humum
muhm<2

\ hmwmm

     

2m
Amman x NV II

«5210 mam
o

 

  

>

     

_o

 

A¢u<~o_wv
>m02w2 4(004

 

 

 

 

m_<lo< No-00 12w! 32w} 0.5.5 "I: m.

 

  

123

coordinator via the bus switch (98). The bus switch consists of one
8—bit bi-directional data bus driver (2 x 8216) and one 15-bit tri-
state line driver (3 x 8095). The transfer process is initiated when
the coordinator addresses the memory space in which the scan control
micro resides. When the proper memory address is placed on address
lines A13—A15, the hold request control circuit sets the slave micro's
hold input (HOLD) to a logic '1' and pulls the coordinator's ready
line (Coord. Ready) low (logic '0'). A low on the coordinator's ready
line halts the coordinator until control of the local bus has been
transferred to the coordinator. If the coordinator is not temporarily
halted, an incomplete data transfer will result because the coordina-
tor will start the data transfer before it actually has control of the
slave's memory. After the hold input to the slave micro goes high,
the slave micro completes its current cycle and then suspends oper-
ation, floats its data and address busses, and acknowledges the hold
request by generating hold acknowledge (HLDA). The HLDA output com-
pletes the transfer of control by enabling the bus switch (logic '0'
at US and CS) and returning the coordinator's ready line to a logic
'1'. At this point the coordinator has control of a user-defined
portion of the scan control micro's memory and task address space.

The coordinator returns control of the local bus to the slave
micro by releasing the address space of the slave. The hold input to
the micro then returns to a low level, HLDA returns to a low level,
the bus switch is disabled, and the micro resumes operation.

The data acquisition system (see Section C of this chapter) can

also generate a hold request to the scan control micro. Consequently,

124

the hold request control circuit was designed to accept only one hold
request at a time. Once this circuit has latched a hold request, it
will not recognize another request until the current one has been
terminated and HLDA has returned to a low level. FAA is the hold
acknowledge signal sent to the data acquisition system in response to
a hold request. The lockout circuit is necessary because of the inter-
nal timing sequence of the 8080 microprocessor. Without this circuit
HLDA from a previous hold request could pass control of the local bus
to the requesting device at the same time that the scan control micro
was resuming bus control. The result of such a conflict would be the
logical 'AND' of information placed on the bus by the two bus control-
lers. This would lead to very unpredictable and undesirable results.

The control, data, and address lines at the top of Figure 7-2

 

(m, DBIN, MEMW, MEMR, DO-D7, A0-A15) and the READY and BUSEN
control lines are used to interface the scan micro to the hardware
front panel (see Appendix C). When it is connected to the micro, the
hardware front panel provides the operator with complete hardware
control of the micro and its peripheral circuits. This hardware
control capability is important not only in testing the micro and its
peripheral circuits during the construction phase, but also for later

trouble-shooting of non-functional circuits.
3. Analog Scan Circuits

The analog scan circuits used in the vidicon spectrometer include
the vidicon tube and deflection assembly drive circuits and their

power supplies. The silicon vidicon tube (RCA C23246), deflection

125

assembly (Cleveland Electronics VYLFA—959), magnetic focus and align—
ment circuits, horizontal and vertical drive circuits, and the preamp-
lifier are contained in a small aluminum box mounted directly to the
front of the monochromator (Heath EU—700) with an XY translation

stage (Newport Research Corporation Model 400). All circuits within
this detector module are shielded from one another by 1/16 inch
aluminum panels connected to chassis ground. Power for these circuits
and for the vidicon tube is obtained through shielded cables from the
power supply. The destination, voltage, and current requirements for
the power supply are listed in Table 7-1. Design details of the power
supply and the magnetic focus and alignment circuits are given in
Appendix B.

The magnetic focus and alignment control circuits are adjustable
current sources. These control circuits are controlled by potenti-
ometers located inside the detector module. As the focus and alignment
were observed to be quite stable with time, there is no need to
readjust these controls once their settings have been optimized.

The vertical deflection drive circuit (Figure 7-3) is based on
conventional television scan circuitry (177). Typical waveforms for
this circuit are illustrated in Figure 7-4. The line sync input is
derived from a crystal-controlled programmable clock. The line sync is
buffered, differentiated, and applied to the base of the 2N3906 PNP
transistor. The transistor is normally held in saturation, but the
differentiated square wave momentarily brings it out of saturation.
Flyback action of the 22 mH choke raises the resulting voltage pulse to

approximately 50 volts peak-to-peak at the transistor collector. This

126

Table 7-1. Analog Power Supply Requirements

 

 

Destination Voltage Current
Tube pin 1 Heater +6.3 V +0.1 A

2 Grid 1, beam current -38 V

3 Grid 4, decelerator +340 V

5 Grid 2, accelerator +300 V

6 Grid 3, beam focus +290 V

7 Cathode -8 V

8 Heater 0 V

Tube target flange (target bias) 0 V
Magnetic focus +24 V +45 mA
-24 V -45 mA
Alignment +15 V +90 mA
-15 V —90 mA
Vertical deflection circuit +5 V +250 mA
-40 V -0.5 A
Horizontal deflection circuit +15 V +0.5 A
—15 V -0.5 A
Preamplifier +15 V +200 mA

-15 V -200 mA

 

127

I000 IK HEIGHT
-40 V

  
   
 
 

IOOuF
+5V
13 ZZmH
470i) IN4I48 "IF
4.
I .4: VERTICAL
LINE SYNC :1, SWEEP
I COIL

 

74HI ‘””F

2N3906 IOfl

.OO39pF

Figure 7-3. Vertical Deflection Drive Circuit

 

 

 

 

 

 

 

RESISTOR - I.OV

v
LINE 5
SYNC
ov .
I I
l I
I I
TRANaSTOR 2V
BASE ov
I I
: I
ov ' '
I
TRANSISTOR : I
COLLECTOR : I
I
-50v 1 '
I
I I
I
flLGSV ' '
RAMP : I
ACROSS I I
IOII : I
I I
I I

¢——————-44O us----’

Figure 7-4. Vertical Deflection Circuit Waveforms (25 kHz)

128

voltage pulse is applied to the vertical deflection coil through the

1 uF capacitor. The coil requires a pulse waveform of this type to
generate a current ramp through its windings because of its inductive
reactance. The 1N4148 diode serves as a conventional damper, while
the 0.0039 uF capacitor is used as a coil ringing capacitor. The 10 0
sense resistor is used to monitor the current ramp through the coil.
This current ramp does not need to be linear; however, it must be
consistent from line to line.

The horizontal deflection drive circuit (Figure 7-5) drives the
scanning electron beam to the selected wavelength channels of interest.
Consequently, it must be both linear and reproducible. In order to
obtain a linear current ramp through the deflection coil, the coil was
included in the feedback loop of a National Semiconductor LH0042
operational amplifier. A Burr-Brown 3553 buffer amplifier was used to
increase the coil drive capability of the circuit. Current sensing was
provided by placing a 10 Q resistor in series with the coil, while a
1 k0 resistor in parallel with the coil suppressed coil ringing. A 0.1
uF feedback capacitor was used to compensate for reactive loading of
the amplifier by the coil. This drive circuit produces a linear cur-
rent ramp through the coil at frequencies up to 1 MHz. Two potenti-
ometers are included for adjustment of the amplitude and centering of
the deflection waveform. Typical waveforms for this circuit are

shown in Figure 7-6.

129

 

200.0.

 

 

    
 

HORIZONTAL

SWEEP

LHOO42 3553 COIL

CENTERING

-lsv>—WV—<+|5V
IOK

Figure 7-5. Horizontal Deflection Drive Circuit

IOV
DAC

m M /

+ 2 V
LHOO42

OUTPUT
- 2 V

 

 

C--- C

+ .
RAMP 0“,

ACROSS
IOO.

RESISTOR -o_| v

+——ZO ms——0

Figure 7—6. Horizontal Deflection Circuit Waveforms (25 kHz)

130
4. Digital Control and Interface Circuits

The control and interface circuitry consists of the horizontal
(wavelength) and vertical (intensity) deflection circuits. These
circuits not only control the scan format of the vidicon detector, but
also provide timing signals to the data acquisition circuits.

The vertical deflection circuit (Figure 7-7) generates the line
sync signal, which is the basic timing and synchronization signal for
the vidicon detector. The line sync is derived from a 10 MHz crystal-
controlled oscillator and a digital prescaler. This prescaler consists
of a programmable +N counter (for N from 1 to 256) and a +10 counter.
The prescaler thus generates two output waveforms, line sync and x10
line sync. Line sync provides the clocking waveform to the vertical
deflection drive circuit (Figure 7-3) and the horizontal deflection
circuit (Figure 7-8). It also initiates the timing sequence that
controls the sample conditioning and sampling circuits (Figure 7-12).
The x10 line sync waveform provides 10 clock pulses per line sync pulse.

A new vertical scan trace is initiated by the rising edge of line
sync. This rising edge enables the count-down input to the %N delay
counter by setting the Q output of FFl to a logic '1'. When the counter
is enabled, it counts a preselected number of x10 line sync pulses and
then underflows. The borrow output (B) disables the counter input by
clearing FFl, reloads the counter, resets the integrator by clearing
FF2, and triggers delay monostable M81. The output of M81 sets flip-
flop FF3. FF3 enable the integrator and the count—down input to the
SN integrate counter. This counter counts a preselected number of x10

line sync pulses and underflows. The borrow output of the integrate

131

.253 E~>zoo 34 TIA

333.. “#2552. .IA

w...m<zw wum<Iom5

ufisouwo coauooamoa Hmowuum>

.AIR muswfia

 

 

.A a. an

Ancoowv
¢ m2

.3...

.283
n m:

 

3::

 

 

 

 

 

u :33:
>33 m

awkzzoo 2....

v

.mz

 

h. H...

 

 

_ 3&8 _

team?

H
.2. On.

 

mm:

 

u 322..

upqzem»z_m
muhzaoo zm .

¢

 

1:55.44 .Em I @—

 

 

oz>m m2...
haoo<wm w>_...ow._ww

.5912.

 

oz>w m2... nVIIllr

oz>m wz... 0.x

 

mmhzaoo CI:

8me

 

 

 

 

 

Io...<._ .EmIo

3.va

 

 

at: o.

 

d.omo

132

counter disables the counter input by clearing FF3, reloads the
counter, and triggers delay monostables M82 and M83. M82 generates a
50 ns pulse that initiates an A/D conversion. M83 provides a 200 ns
delay before M84 discharges the integrator by setting flip—flop FF2.
This circuit not only allows the computer to select the scan frequency,
but also allows it to specify the vertical interval that is to be
integrated during the data acquisition process.

The line sync signal is also gated with two other signals, line
sync inhibit and selective readout. These signals are used in the
random access mode of target interrogation. In this mode the computer
interrogates only those wavelength channels of immediate interest by
inhibiting the oscillator-based line sync, positioning the readout beam
at the desired wavelength channel, and generating a line sync pulse by
pulsing the selective readout line. Although the circuits have been
designed to permit random access scanning, this mode has not been
implemented because of the slow cycle time of the 8080A microprocessor.

The wavelength deflection circuit (Figure 7-8) generates an analog
output that is used by the horizontal deflection drive circuit to drive
the electron beam to the wavelength channel of interest. This circuit
consists of a hardware adder circuit with programmable increment and a
high speed D/A converter (Datel DAC-lZHYBC high speed 0 to +10 V
converter with a 300 ns current settling time). Although the D/A
converter has an internal amplifier for voltage output, an external
high speed operational amplifier (Datel AM-100 fast settling FET 0A
with 400 ns settling time) was used to achieve a faster settling time.

Either the 12—bit output of the adder circuit or a 12—bit random

133

 

 

[ 8 'BIT LATCH

4-Bn'LATCH
(82I2)

(74l74)

 

 

 

 

\HOICOII 1T
CARRY

     

MS (IZOns)

  

  

 

 
 

IZ’BIT ADDER
(3 x 7483)

E

 

 
 

 

   

 

l2

 

 

   

 

 

Q 0
men LATCH I2-BIT LATCH
(2x 74l74) ° .2 ° (2x 74l74)
L2 CLK CLK Ll

 

 

 

 

_ ‘1?
n I

 

 

  
 

LINE SYNC E>

 

      

A B
It BIT A/B SELE
35L (3:: 74I58)

STB Y

  

 
 

SELECTIVE
ENABLE

 
 

 

    
 
  
 

 

 

 

 

 

 

I2
DAC IZHYBC
SCAN
ENABLE
V0

SCAN .
ENABLED ¢-<} “"00“
FLAG

Figure 7—8. Horizontal DefleCtion Circuit

134

access address determines the wavelength channel to which the beam is

deflected. A logic '1' applied to the selective enable input enables

the random access mode of scanning. The number of wavelength channels
per scan is 4096/N, where N is the 8—bit programmable increment.

The hardware adder circuit is clocked by the line sync signal from
the vertical deflection circuit. The rising edge of line sync causes
latch L1 to latch the 12-bit output of the adder. The output of L1 is
connected to latch L2 and a 12-bit 2-to-1 multiplexer. The falling
edge of line sync causes latch L2 to latch the output of L1 and apply
it to the hardware adder. The adder adds the output of L2 to the
programmable increment and applies the sum to L1. Thus the wavelength
address is updated every line sync until the adder overflows. When the
adder overflows, the adder carry pulse clears latches L1 and L2 (resets
the wavelength), sends a scan completed signal (vidicon carry) to the
computer, and triggers monostable MS. This monostable generates a scan
synchronization pulse that clocks the scan enable flip-flop (FF). If
scan enable is a logic '1' when the flip-flop is clocked, the DAC input
is enabled. If scan enable is not a logic '1', then the DAC input is
disabled (set to 0 volts). This scan enable control is necessary to
synchronize data collection with wavelength scanning after the readout
beam has been disabled and blanked during charge integration. A scan
enabled flag signals the computer when the wavelength deflection

circuit is enabled.

135

5. Real-Time Clock and Interrupt System

A versatile real-time clock and interrupt system (Figure 7-9) has
been designed for the 8080A microprocessor to allow accurate interval
timing. This system features a flexible time base, hardware or soft-
ware triggering of the clock, and vectored interrupts.

The real-time clock consists of a Mostek 5009P Counter Time-Base
Circuit (178) and the Intel 8253 Programmable Interval Timer (179).

The counter time-base serves as a programmable prescaler with frequency
division ranges from 1 to 36 x 108. The interval timer has three
programmable 16-bit counters, one of which is used as the real-time
clock. The count-down sequence of this counter is initiated by either
a hardware or software trigger pulse. The clock input to the real—time
clock circuit is a 1 MHz square wave signal derived from the crystal—
controlled 2 MHz micro clock signal (¢2TTL)°

At the end of the timing interval, the borrow output of the inter-
val timer sends an interrupt request to the priority interrupt control—
ler by clocking flip-flop FFl. If the interrupt system has been
previously enabled, the controller will generate an interrupt if the
priority of the requesting signal is greater than the priority in a
software-controlled status register. The interrupt controller also
issues vector information along with each interrupt to identify the
service routine. Requesting level priorities are assigned by the
designer. 1R7 is the highest priority input to the 8214 and R0 the
lowest.

The vidicon carry output from the wavelength deflection circuit

also generates an interrupt request by clocking flip-flop FF2. This

136

MEMW -
A SELECT " .

m A SELECT
CLK A SELECT

 

 

 
 
 

 

  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

00-03 1:; ,14 on74o'74
n
2.4V "5'”
_,5V p 4 MEMR
I2I<
" VGG 2"
T A0
MKSOOSP ‘ A.
Von EXT
DATA
Bus
0 TRIGGER I
92711 - CLK TRIGGER 2
av
M. CLEAR
VIDICON
CARRY
CHEER
00-03 - 4
INTE -
0
— L 5 "07°“
MEMW - . '- 04 'fi—‘AOORESS
A—SEL—ECT- -0
_ 3 8
5v _
2 INT '——-0 INT
2K ' 0?] - INTA
J; 2
REQUEST VECTOR __
R3 040 CLR MD
R; 050
5v

Figure 7-9. Real—Time Clock and Interrupt System

137

interrupt request is generated at the completion of each frame. The
vidicon carry interrupt request has a lower priority than the interval
timer so that it may be disabled without affecting interrupts from the
timer. Thus the micro can use either vectored interrupts or software
polling of the scan completed flag (see Appendix B) to count the

number of scans completed during signal averaging.

6. Dedicated Microprocessor for Number Processing

The National Semiconductor MM57109 MOS/LSI Number4Oriented Micro-
processor (180) is used in this microprocessor system for number
processing applications in place of user-written number processing
software. Figure 7—10 shows the interface circuit used to interface
the 57109 to the 8080A bus as a peripheral device.

Although the 57109 is a MOS device, all but two of its signal
inputs (HOLD and FOR) will respond properly to TTL logic levels if the
processor is operated from supplies of +5 V and -4 V. The two non-TTL
compatible signal inputs are driven by RCA 3140 operational amplifiers.
These amplifiers are used as comparators to convert TTL logic levels to
MOS compatible levels. A logic '1' applied to the POR input resets the
processor. A logic '1' applied to the HOLD input halts the processor.
This input must be high when the ready line (RDY) goes high unless a
6—bit instruction is present in the instruction latch (L1).

Prior to using the 57109, the microprocessor first resets the
cruncher (microprocessor refers to the 8080A - cruncher refers to the
57109). When it is ready to input a 6—bit instruction the cruncher

sets the ready flag (INRDY). The microprocessor then writes an

138

xm

>¢|
rm
mommm >01

3}.

AmOCInv

muxozzmo
mmmizz

0401
owo mod

uomwmooum OOOEOAAOIuonfisz aofinm pow wommuwuaH .oHIn muamfim

EOE

net
52%.

mam

Mr
:40 zIw
a 2:35 o :32;
5:: red «mama mam

N4

21.3:

10.2.. .2me mam

oSn <0:

>0

mm?
0 m

>w..
OSn <01

 

GOOD
>

m<oxm4<zm

139

instruction into the 6-bit instruction latch (L1). This write cycle
also pulses the HOLD input low. When HOLD goes low, RDY also goes low
and the cruncher reads and performs the 6-bit instruction. Numbers are
entered one BCD digit at a time into the cruncher as instructions.

When INRDY goes high the next instruction is entered.

Although data can be entered either synchronously or nonsynchro-
nously, they are output synchronously. The actual number of output
digits is determined by the mantissa digit count (MDC). The MDC is
initially 8, but it can be set to any value from 1 to 8 by using the
SMDC instruction. When an output (OUT) instruction is executed, the
cruncher sequences through the specified number of digits. As each
digit is output on the data out (DO) lines the R/W line clocks the
digit into a 6—bit data output latch (L2) and sets the data out ready
flag (OUTRDY). The microprocessor must then detect the data out flag
and read the latched digit before the next digit is latched. Time
between digits is approximately 140 us. The error flag (ERROR) is set

high if an arithmetic or output error is detected.
7. Shutter Control Circuit

The shutter control circuit (Figure 7-11) allows the scan micro to
control the optical shutter. A logic '1' applied to the base of the
2N6043 NPN transistor turns on the transistor, energizes the solenoid,
and opens the shutter. A logic '0' reverses the process and closes the
shutter. The 400 O resistor eliminates bursts of high frequency noise
that had been Observed whenever the solenoid was energized or de-ener-
gized. The 40 V zener diode protects the transistor from harmful volt-

age transients. The solenoid is mounted inside the sample compartment

140

>Ov

 

 

nvomzm

uwsoufio Houucou Hmuusnm .HfiIn muswwm

>0

CONN ~31.

 

EOquOm

3 O. .C.OO¢

 

ca.

>¢N+

 

 

 

 

 

 

 

 

 

 

 

 

 

 

K~ zudo I...
\\ ammono I .o.
Laoomu
I l.
2.2:. 2.: I H.
55.53: M" o<
mo<So> + l" o:
> cu + 1
F zmzcoumz<¢»
I <~._ >¢~

141

of a Health Alternating Cell Module Model EU-721-11 and is attached

directly to the shutter.

C. Data Acquisition System

1. Introduction

The data acquisition system for the vidicon spectrometer consists
of the analog conditioning and sampling circuits, hardware adder for
signal averaging, and the digital control circuits. The design details
of these circuits and the timing of a data acquisition cycle are

discussed in this section.

2. Analog Conditioning and Sampling Circuits

The analog conditioning and sampling circuits are illustrated in
Figure 7-12. The preamplifier consists of a current-to-voltage
converter (0A1) with a measured 3dB point of 88 kHz and a x10 buffer
amplifier (0A2 and 0A3). The buffer amplifier produces a video signal
capable of driving the 100 O twisted-pair shielded cable connection to
buffer amplifier 0A4 on the signal conditioning and sampling board.

The preamplifier is located in a small aluminum box with the vidicon
tube, deflection assembly, and drive circuits (see Section B.3.). It
is connected to the vidicon target by a 4 cm length of 50 O coaxial
cable. The length Of this connection was minimized because of the high
speed and low level of the vidicon detector video output.

The signal conditioning and sampling board is located in the micro
rack (see Chapter 6). It consists of a buffer amplifier (0A4),

integrator (0A5), sample—and—hold amplifier (Datel Model SHM-IC-l

142

muwsouwu wOHHmEmm cam wcficowuwccoo woama< .NHIn muswfim

 

 

 

 

 

 

 

 

 

 

 

 

can

.3 oo. _.II|I0 2328
803m. «.5 conch. I 3
I\m o\< ~_ 5&8 ~78 84 I a? 302... I 25
REA. 72-25 I zxm omooozn I v.26
I 01.va «a 330:..- 33
4 L 825
>
_
_ . Emu
_ dc...
xo. H
_ .3
_ xoo_ £3
_ e r
_ M:
_ L A. Can.»
_

 
  

mmijmzaummm
mh<mOm .—.z_

x. . ATIAU mom<zom5

 

143

integrated circuit sample and hold with an acquisition time of 2 us
and a hold mode droop of 500 mV/s), and 12-bit analog-to-digital
converter (Burr—Brown ADC85-12 12-bit successive approximation A/D
converter with 10 us conversion speed). The integrate, discharge,
and convert commands are generated by the data acquisition control
circuit (see Section C.4.).

The 1 k9 resistor in series with the discharge analog switch (82)
protects the switch from damage by limiting the discharge current.
Diodes D1 and D2 limit the integrator output to +10.6 V. D1 is
included because of its very low leakage current when off (1 nA). It
also ensures that the diode loop is open under negative input condi—
tions, although this condition should never occur. The 1 k0 resistor
between the two diodes keeps the anode of the switching diode (D1) near
ground unless the zener diode (D2) is conducting. This resistor thus
maintains the low leakage current advantage of the switching diode.
Without this resistor, leakage current from the zener diode could turn
on the switching diode. The 3.9 k0 resistor at the inverting input of

the integrator (0A5) protects the amplifier input from discharge

transients.

3. Digital Control Circuits

This section describes the design details of the digital circuits
that control data acquisition and signal averaging in the vidicon
spectrometer. Because of its slow cycle time the 8080A microprocessor
is unable to directly control the transfer of data into memory during

a data acquisition cycle. Consequently, a direct memory access (DMA)

144

controller circuit was developed for data acquisition and signal
averaging. This controller circuit uses the Hold feature of the 8080
to acquire the system bus and generates the control signals required
to allow the hardware adder to access or deposit data directly from or
to memory. The controller uses the Oz (2 MHz TTL) clock of the 8080
system for timing so that individual memory read or write cycles
require only 500 us.

The fast adder (F.A.) controller circuit (Figure 7—13) is the
hardware sequencer for the DMA data acquisition system. Prior to the
start Of data acquisition, the microprocessor loads the 4-bit byte
count (# bytes of precision - 1) latch and the 16-bit starting memory
address latch. The micro then sets the start scan (1 bit in the scan
control register, see Appendix B) input to a logic '1'. Next, data
acquisition is synchronized with the scanning electron beam by the
vidicon carry signal, which clocks the start scan flip-flop (FFl).
Finally, the F.A. circuit is enabled by a line sync pulse, which sets
the F.A. enabled signal to a logic '1' by presetting flip-flop FF3 and
enables the F.A. request flip-flop (FF2).

Once the F.A. circuit has been enabled by the micro, data acquisi-
tion and signal averaging proceed under hardware control independent of
the microprocessor. A data acquisition cycle is initiated by the
falling edge of the A/D status line. This falling edge, which indicates
that the converter has completed a conversion, triggers monostable M81.
This monostable sends a hold request (F.A. request) to the micro. The
data acquisition cycle starts when the F.A. controller receives a hold

acknowledge (FAA) from the micro.

145

 

 

rumnwm
3
..II
10...<
mm
n.< w.
I O< mm>Eo
mmmmoo<
FIB o.
.. 10.5... ....m I O.
NO
Ion
32w! c
10.2.. .20 I t
«In

hmwaowm .<....
ow4m<zu .(d

uwsoufiu HOHHOHOGOO umpw< ummm

E .30

«UPZDOO
mmmmoo<

v.40 Bommom
mwkzaoo wk>m

ON

0.

Ha

xJON
.uu 140.

ON

a

mu m~

mm: H

CO.

 

.2. oo;

o

.m:

<

.mHIA Buswfim

ntms

>mm<o
200.0;

z<om
hx<hm

 

w4m<zm
m>.._.ow..ww

oz>m m2...

 

m3...<._.m O\<

146

Timing waveforms for a triple—precision (24—bit) data acquisition
cycle are shown in Figure 7-14. The FAA signal from the micro enables
the controller clock (C1), the address bus driver, and the MEMR and
‘MEMW drivers. This clock provides the timing for the data acquisition
cycle; alternate 500 ns memory read (MEMR) and memory write (MEMW)
pulses, corresponding data latch (LATCH) and memory select (SELECT)
pulses, and clocking of the byte and address counters.

A data acquisition cycle continues until the required number of
data bytes have been averaged, at which time the byte counter under-
flows. The borrow output from the byte counter triggers monostable
M82. This monostable disables the F.A. controller clock, disables the
memory read and write lines, reloads the byte counter, and releases the
F.A. request to the micro. This data acquisition cycle, which requires
less than 4 us for a triple-precision average, is repeated each time
the converter completes a conversion.

At the end of every wavelength scan a vidicon carry signal is
generated by the wavelength deflection circuit. This signal not only
clocks the start scan signal every frame (one wavelength scan) to
synchronize the enabling and disabling of the F.A. circuit, but also
reloads the address counter with the starting address contained in the
address latch. The address counter must be reset at the begining of
each frame for signal averaging. However, the address counter must not
be reset at the end of a frame until the last F.A. cycle is completed.
Circuit timing was adjusted to properly synchronize these signals.

The F.A. controller also has a selective enable input ( 1 bit in

the scan control register, see Appendix B) for random access interroga-

tion of the vidicon target. The rising edge of selective enable

147

.1

—N

ETA/
BOARD E

 

 

 

 

 

 

 

 

SELECT

 

 

,l'nlllj

.l'll|"'

 

AIO

 

AI

 

TI

A 2

 

READIWRITE
BYTE 3

 

READ IWRITE
BYTE 2

 

READ IWRITE
BYTE I

ISOOns

 

 

STOP
IMMA

 

START
EMMA

Fast Adder Timing Waveforms

Figure 7-14.

148

triggers monostable M83, which loads the address and byte counters.

A logic '1' at selective enable enables the line sync input to flip—
flop FF3 and enables gate G. Gate G is used in the random access mode
to load the address counter whenever the high byte of a new address is
loaded into the 16-bit address latch (58h pulsed low). After the
address has been loaded into the address latch, line sync is pulsed
high to initiate a data acquisition cycle. Thus a data acquisition
cycle in random access scanning requires the micro to load the address
latch with a 16-bit address and to pulse line sync. As mentioned
earlier, random access scanning has not been implemented because of the

slow cycle time of the 8080A microprocessor.
4. Hardware Adder for Signal Averaging

The fast adder circuit (Figure 7—15) performs the actual data
acquisition and signal averaging. The timing signals for the fast
adder originate on the F.A. controller board and are illustrated in
Figure 7-14. At the completion of an A/D conversion, the falling edge
of the A/D status line triggers monostable MS. This monostable clocks
the inverted output of the A/D converter (converter's output is
complementary straight binary) into a 12—bit latch and clears control
flip-flops FFl and FF2 to select the lower 8 bits for the first
average. At the same time data from the corresponding memory location
is read into an 8-bit data latch. These two 8-bit numbers are added
by an 8-bit hardware adder, the sum is written back into the same
memory location, the carry is latched by flip-flop FF3, and the memory
address is incremented. The next byte is then read from memory into

the 8-bit latch and added to the upper 4 bits from the A/D conversion

 

MS (200 m)

   

149

 

ADC 85

 

    

 

   
   

 

 

 

 

 

DBIN

MICRO DATA
BUS

 

SELECT
MEMW
SELECT
ADDRESS

Figure 7-15.

    

 

 

   
  

 

 
   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Fast Adder Circuit

CONVERT
VOLTSIN
I2-BIT INVERTER
(2 x 7404)
I2-BIT LATCH
(2 x 74l74)
11'
If CLK
SEL A a I
B-BIT A/B SELECT
(2 x 74I57)
fi STB Y
B
C4 A A
BUS DRIVER
B'BlT ADDER {I (2 x 8095)
(2 x 7483) ._
I B CD E"
_ 4' a
B- BIT LATCH ” 8
_ (le2)
D ,
l=--' 5H5; [K)h——-dililDI
BUS MEMORY
‘ DB DRIVER (48 x 2I02A-4)
8 (axeme)
C—S DIP-1 DO R/W C-S ADR
5
i
D 1
’l3

150
together with the carry from the previous addition. This sum is
written back into memory and the memory address is again incremented.
Finally, the triple—precision addition is completed by adding the
remaining byte in memory to the carry from the second addition and
storing the sum in memory.

At a typical scan frequency of 25 kHz a data point is acquired
every 40 us. Since a triple-precision data acquisition cycle requires
less than 4 us, the micro is only slowed down approximately 10% at a
scan rate of 25 kHz. Thus the micro can count the number of signal
averages completed, monitor the quality of the incoming data, or

perform any other desired function at about 90% of full speed.

D. Optical Configuration

The dispersion system used in the vidicon spectrometer was a
GOA/McPherson EU-700 Czerny-Turner Monochromator (176). The vidicon
detector was mounted in front of the monochromator with a Newport
Research Corporation Model 400 XY Translation Stage. The exit slit
was covered, the second folding mirror removed, and the rear focusing
mirror moved forward to focus the grating image onto the window of the
vidicon tube. The original monochromator grating was replaced with a
Bausch and Lomb precision certified grating with 133.6 groves/mm.

The reciprocal linear dispersion of this system is 18 nm/mm. The
width of the active vidicon target is 12.8 mm, so that the target

monitors a wavelength window of approximately 230 nm.

APPENDICES

APPENDIX A

Program Listings

151

8
: HHCROO.CON
I
I JIM HORNSHUH NOVEMBER 1, 1977
8
I CONVERSE LIBRARY FOR COORDINATOR MICRO
I
<< T01 :STACKIXFER FROMICOORD TO COMHUN
PUSH H
PUSH PSW
PUSH 3
OFF
HOV B.A
HOV C,A
:10 OFF
PUSH PSW
DCR B
JNZ .10
LHLD 20014
NOV B.C
MOV A.H
ORI 40
NOV H,A
:20 DCX. H
POP PSW
HOV M.A
DCR. B
JNZ .20
DCX ‘H
HOV MAC
HOV A.H
AN! 37
NOV H.A
SHLD 20014
POP B
POP PSW
POP H
>>
<< OPEN :COORD ROUTINE TO OPEN SHUTTER
SETUP 2 SOPEN
RUN 2
>>
<< CLOSE :COORD ROUTINE TO CLOSE SHUTTER
SETUP 2 SCLOSE
RUN 2
>>
<< SAMP :COORD ROUTINE TO SELECT SAMPLE CELL
SETUP 2 SSAMP
RUN 2
>>
<< REF :COORD SELECT REFERENCE CELL
SETUP 2 SREF
RUN 2
>>
<< 102 :STACKZXFER.FROMICOORD T0 SLAVE 2
PUSH H
PUSH PSW
PUSH B
OFF
HOV B.A
HOV C.A
:10 OFF
PUSH PSW
DCR B

JNZ .10

I 20 DCX

POP
) >

< < DECB IN
PUSH
PUSH
PUSH
PUSH
OFF
HOV
LXI

: 10 IN I
OFF
I‘IOV
CF I
J NC
DAD
J C
DCR
J Z

: 30 DCR

: 40 NOV
ON
POP
POP
POP
POP
STC
) >

< ( DEC IN
PUSH

a» Em

HBUOF0>3I
t

a? 9?? aku
bafibt

I
(0
0

H553 ?U
1‘. Bl

PSW

152

:DECIIIAL TO BINARY CONVERSION

I CHARACTERS

:INITIALIZE THE BUFFER

IGET A DIGIT

818 IT BCD?

IJUHP IF ERROR

IADD TEE DIGIT TO THE BUFFER
:TOO LARGE IF CARRY

ILAS'I‘ DIGIT?

IJUHP IF DONE
IIF NOT THEN BUFFERX 10

:ADD 10 TIMES FOR MULTIPLY
ITOO LARGE IF CARRY
IADDED 10 TIMES?

IBACK '10 GET A DIGITH
IPUT THE RESULT ON THE STACK

ICLEAR THE CARRY (NO ERROR)

I CHARACTERSIEFT

ISET THE CARRY (ERROR FLAG)
IINPUT DECII‘IAL NUMBER.

153

:10 STRING IINPUT ASCII STRING FROM TI'Y
XRA A '
ON IASCII => OCTAL MODE
ASCII :CONVERT To BCD
DECBIN :CONVERT TO BINARY
JC .20 :ERROR FLAG SET?
POP PSW
RET
: 20 CLEAN IREMOVE GARBAGE FROM THE STAGE
WHAT :SEND A ’7’ TO TI'Y
JMP . 10 :TRY AGAIN
)>
<< MOVIT :COORD ROUTINE TO CHANGE WAVELENGTH
PUSH H
PUSH B
MVI H. 1 1
LXI B. .25
I 10 LDAX B :GET CHARACTER
ON :AND PUSH ON STACK -
INX B :INCREMENT ADDRESS POINTER
DCR H :AND DECREMENT CHARACTER COUNT
JNZ .10 :JUMP IF MORE TO DO
MESSAG :OUTPUT MESSAGE
DECIN :INPOT NEW WAVELENGTH
POP2
SHLD 40024 :STORE IN SLAVE 2
SETUP 2 WAV
RON 2 IRON MONOCHROMATOR PROGRAM
: 20 FLAG 2 : AT NEW WAVELENGTH?
.12 .20 :JUMP IF NOT
LDA 40036 :ELSE GET STATUS
ANA A
JZ .30 IJUMP IF NO ERROR
S'I‘C :ELSE SET CARRY
POP B
POP H
RET IAND RETURN
:30 81C
cm I IF NO ERROR. CLEAR CARRY
POP B
POP H
RET IAND RETURN
:25 15.12.127.101.126,40,75,40.10 :MESSAGE
>>
<< AXIS IAXIS ROUTINE FOR COORD
PUSH H
PUSH PSW
LHLD 40113 IGET POSITION OF AXES
PUSH2 :ON STACK
MVI A.2
ON INOMBER OF BYTES ON STACK
101 ITRANSFER '1!) COMMON.
SETUP 1 AXIS
RUN I :DRAV AXES
XRA A IA=0
ON :0 TO STACK
LHLD 40115 IGET X Y INCREMENTS
PUSH2 ITO STACK
LHLD 41-0113 IAXES
POSH2
MVI A.5
ON :BYTE COUNT
: 10 FLAG 1 :CHECK COMMON.
.12 .10 :JUMP IF NOT READY
T01 :TRANSFER TO COMMON.

>>
<< CC

>>
<< DIV

>>

SETUP I TTCMHK

RUN I
FLAG 1
J2
POP
POP

PUSH
PUSH
HVI
MVI
PUSH2
T01
SETUP I
RUN l
FLAG I
JZ
HVI
STA
POP
POP

PUSH
PUSH
PUSH
PUSH
OFF
MOV
POP2
MVI
XRA
HOV
HOV
RAL
HOV
MOV
RAL
HOV
DCR
JZ
NOV
RAL
JNC
SUB
JMP
SUB
JNG
ADD
JMP
MOV
CMA
HOV
NOV
CMA
HOV
PUSH2
HOV
ON
POP
POP
POP
POP

.20
PSW
H
PSW
L.I
H.3
CC

.10
A,I7
36500

PSW

PSV

HES:

5'
I»
H

O. O. O

' a. U: >' OH bl" DUDO
-- -- I9 ' a
O 0 O U: a» m> t"> N

' U
u
0

A.H

H,A
.A,L

L.A
A.D

PSV

154

:DRAW'TICKMARKS

ICHECKZFOR.C

I BYTE COUNT

:CONTRL C

ISEND TO COMMON.

sRUN CHECK PROGRAM

ICHECK FLAG. JUMP IF NO CONTRL C
IENABLE ALPHA. DISABLE GRAPHICS

IIG-BIT BY 8-BIT UNSIGNED DIVISION
ISAVE STATUS AND.REGISTERS

IGET DIVISOR

IGET‘DIVIDEND
IBIT COUNTER

:ROTATE LOW BYTE

IROTATE HIGH BYTE
:DECREMENT'COUNTER

ICOMPLEMENT HIGH BYTE OF QUOTIENT

:COMPLEMENT DOW BYTE
IOUOTIENT ON STACK

IREMAINDER ON STAGE
IRESTORE STATUS AND REGISTERS

<< BINDEC

»

PUSH
PUSH
PUSH
MV I
MOV
IN I
ON

D I V
OFF
PUSH
DCR
J NZ
POP2
MOV
POP
CP I
J NZ
DCR
DCR
J NZ

I NR

I NR
J MP
POP
ON
DCR
J NZ
MOV
ON
POP
POP
POP

< ( DECOUT

>>

PUSH

BINDEC

MVI
ON
ASCII

MESSAG

POP

< < wmom

PUSH
PUSH
PUSH
OFF
MOV

H

B
PSW
0.3
B.C
A. I2

A. B
PSW

PSW
A. I

PSI!

9

Lilla EFF!" HUN!
0 ° In?
8»

~r>
u
Cl

400 I7

155

:IG-BIT BINARY TO 5 DIGIT BCD

IINITIALIZE BCD DIGIT COUNTER
IDIVISOR FOR 10

IDIVIDE BINARY NUMBER BY 10
IGET REMAINDER

IAND SAVE

IDECREMENT BCD COUNT

IJUMP IF MORE LEFT

:GET QUOTIENT

:REINITIALIZE BCD COUNTER
IRECOVER REMAINDER

IIS IT A ZERO?

INO. REST OF DIGI'IS ARE SIGNIFICANT
IYES. DECREMENT COUNT

:JUMP IF MORE 'IODO

IPUT REMAINDER ON STACK

:GET NUMBER OF NON-ZERO DIGITS
IAND PUT ON STACK

IOUTPUT DECIMAL NUMBER
ICONVERT BINARY NUMBER TO DECIHAL

ICONVERT TO ASCII CHARACTERS
IOUTPUT

I INPUT( 0) /OUTPUT( 1) END LOCATION

IGET MODE

IAND STORE IN L

:BYTE COUNT

:STARTING ADDRESS OF MESSAGE
IGET CHARACTER

IPUSH ON STACK

I INGREDIENT ADDRESS
IDECREMENT BYTE COUNT

IJOMP IF MORE

IOUTPUT MESSAGE

:A=0

:MODE=0?

:NO. JUMP

IYES. INPUT CURRENT WAVELENGTH

IAND STORE IN SLAVE 2
IRESTORE AND RETURN

POP
RET
LHLD
PUSH2
DECOUT
POP
POP
POP
RET

PSN
40017
‘B

H
PSW

156

IGET‘CURRENT'POSITTON

IOUTPUT’CURRENT WAVELENGTH
IAND RETURN

15.12.127.101.126.40.75.40.10

PUSH
PUSH
PUSH
PUSH
LXI
PUSH2
MESSAG
LXI
SHLD
LXI

H
B
D
PSW

H.15401

B.0
.50
D..40
B.0
0.3
.20
H.4401

A.B
20

0 lo
C.3
.20

A.B
20
.5

‘H.37001

D..40
U0

A.M
3
.13

.11
H.

B.3
B
D
361
D
A.B
.15
21
.12
.14
21
.16

IDIALOGUE ROUTINE

:PAGE CRT

IINITIALIZE BYTE COUNT’POINTER
:ADDRESS OF MNEMONIC TABLE
IINITIALIZE MNEMDNIC COUNTER
I3 BYTES/MNEMONIC

IOUTPUT MNEMONIC AND ITS VALUE

IOUTPUT TAB

IALL MNEMONICS OUTPUT?
:YES. JUMP

:NO. OUTPUT NDIT ONE

:MORE LEFT?
:YES. JUMP

:OUTPUT ’>‘

IINPUT MNEMONIC

:ADDRESS OF MNEMONIC TABLE
IADDRESS OF STACK

ICLEAN STACK

IGET BYTES
:83?
:YES. JUMP

:NO. OUTPUT ’7'
IAND INPUT AGAIN

:INCREMENT MNEMONIC COUNTER
ICALL SYSTEM COMPARE ROUTINE

IJUMP IF MATCH FOUND

ICOMPARE ALL MNEMONICS?

IYES. OUTPUT ’?' AND INPUT AGAIN
:NO. COMPARE NEXT IN TABLE
:COMPARE WITH 'END'

IJUMP IF NOT END ON INPUT

:32
:34

:20

POP
POP

POP

PSW

A.B
.50

H. .41
43'?

11.13
.30
.32

M. D
POU'I2

.34

H. 20040

A.L

157

:RETURN

:ADDRESS TABLE ADDRI'ES
ICALL JUMP TABLE ROUTINE
IINPUT DECIMAL NUMBER

IPUT IST BYTE IN MEMORY

II OR 2 BYTES?

IJUMP IF ONLY I BYTE

ITRANSFER 2ND BYTE

ICHECK TO SEE IF MICRO 2 IS FINISHED
ILOOP IF NOT

IGO DO AGAIN
IPUT TWO SPACES ON STACK

IAND ADD ANOTHER
IGET MNEMONIC

IJUMP IF MORE CHARACTERS
:BYTE COUNT TO STACK
IOUTPUT MNEMONIC

I OUTPUT SPACE

IADDRESS OF ADDRESS TABIE
ICALL JUMP TABLE ROUTINE
I GET BYTE

I1 0R2 BYTES?

IGET 2ND BYTE

IOUTPUT DECIMAL NUMBER

158

LXI H..51
XCHG
LHLD .50
DAD D
MOV A.M
CPI 2
POP D
POP H
RET
=40 127.101.126.106.122.121.120.124.123.102.105.115.I04.114

131.101.103.121.102.114.113.107.101.111.102.131.124
102.111,124,101.126.107.111.116.124,101.104.122.103
116.124.117,120,124.123.122.104.105.116.104

:41 40024.40117.40120.40104.40122.40123.40124.40035.40037
40125.40026.40034.40102.40100.40126.40127

:50 0.0

:51 2,lgzgl,l.‘.l,lglglg2glq2g2gl.1

<< CON :ENABLE GRAPHICS ~ DISABLE ALPHA
PUSH PSV
MVI A.21
STA 36500
POP PSW
>>

<< SUBTR :SUBTRACT UNSIGNED 16-BIT NUMBERS
PUSH ISAVE REGISTERS
PUSH
POP2 IGET SUBTRAHEND
XCHG
POP2 IGET MINUEND
MOV
SUB :SUBTRACT LON BYTES
MOV
MOV
SBB
MOV
PUSH2
POP
POP

OH!

O

b HID f‘

ISUBTRACT HIGH BYTES

:ANSWER,BACK.ON STACK
IRESTORE

EEG fltflbt‘fll’

))
<< VSET
:10 FLAG 2
J2 .10
SETUP 2 VSETUP IMODIFY VIDICON SCAN FORMAT
> RUN 2
>

<< FSET :INITIALIZE FAST ADDER.CIRCUIT
:10 FLAG 2

J2 .10

SETUP 2 FASET

RUN 2
>>

<< CLRBK :CLEAR. 8K DATA MEMORY
:10 FLAG 2

J2 .10

SETUP 2 CLRMEM

RUN 2
>>

<< VINIT :COORI) INITIALIZATION OF MICRO 2
PUSH H
PUSH PSW
SETUP 2 VINIT2
RON 2 IRON INITIALIZATIONPROGRAM ON 2
:10 FLAG 2 IDONE?
JZ .10 ;N0. JUMP

:20

POP
POP
RET
XCHG
MOV
ANI
MOV
SHLD

.40
40024

40017

.47
A.H

.55
.50

WAV

.60
PSW

H
B
D

WFILEI

.10
.40
D
D
A
.30

H.60000
.42
H.65252

.42

H.72524

.42
CFILI

.15

man:

A.H
37
B’A
20014

159

:ELSE 0 ON STACK
:REQUEST CURRENT MDNOCHROMATOR.POSITTON

:MONOCHROMATOR.BACK'UNDER MICRO CONTROL
:DONE?

:NO

:USER I/O

:SET UP NEW SCAN FORMAT

ISET UP NEW DATA ACQUISITION PARAMETERS

IGET NEW WAVELENGTH
:CURRENT WAVELENGTH
:ANY'DIFFERENCE?
IGET‘ANSWER
IANSWERSO?

ICLEAN STACK
I REPOS ITION MONOCHROMA’IOR

IOPEN FILE ON FLOPPY'- COORD

IGET ADDRESS OF I‘S STACK

IGET STATUS

IJUMP IF FILE NOT OPENED
:ADDRESS OF DARK
:TRANSFER BLOCK OF DATA
IADDRESS OF REFERENCE
:TRANSFER

:ADDRESS OF SAMPLE
ITRANSFER

IUPDATE COMMUN'S STACK

XCHG

MOV
SHLD

SETUP I

RUN 1

FLAG I

CALL
POP
PUSH2

ON

D I V
OFF
STA
POP2
MOV
STA
LDA
ANA

DCR
STA
MV 1
MOV
J MP
LDA
ANA
J NZ
POP
RET

. 20
200 14

40
H.A

III
GUHHIHQI
8 u-
»: 3

$55-05:
G
O
N
“I

0026

U>UU?GG‘PUO
t"

A.L
.56
.56

.56
B. 100
an
.52
.55

.51

160

ISET CARRY IF FILE NOT OPENED
IAND JUMP TO EXIT

IGET ADDRESS OF STACK
:ADJUST ADDRESS FOR COORD

:HL -> DE

ISAVE DATA ADDRESS

IGET TOTAL BYTE COUNT
:AND SAVE

IGET BYTES/ DATA WORD

:NUMBIHI OF SIGNAL AVERAG- - HIGH BYTE

:SIGNAL AVERAGES - LOW BYTE

ISEND BYTE COUNTTOMICRO I

I SEND OF BYTES

I RE-ADJ UST ADDRESS

IUPDATE MICRO I‘S STACK

IGET CURRENT ADDRESS OF 1 ’S STACK

IGET BYTE COUNT
I NUMBER OF BYTES/TRANSFER

IGET REMAINDER
IGET QUOTIENT

:JUMP ON ZERO
:ELSE DECREMENT COUNT

:JUMP TO TRANSFER ROUTINE
IGET REMAINDER

IJUMP IF NON-ZERO
IELSE CLEAR STACK
IAND RETURN

:52

:56
>>
<< FLT

210

:17

NOV
MOV

STA
POP
DCX
HOV
INX
STAX

DCR
JNZ
NOV
STAX
PUSH
XEHG
HOV
ANI
MOV
SHLD

SETUP I

RUN 1

FLAG 1

CALL
JHP
000
000

PUSH
PUSH
PUSH
PUSH
CRLF

HVI
HOV
LDAX
ON
INX

JNZ
MOV
ON

HESSAG
STRING

OFF
CPI
JZ

CLEAN
WHAT
JHP
OFF
CPI

LXI
JHP
CPI
JNZ
LXI

CPI
JNZ

a>99
Q >J'

uuuuyuu
:1

flfiflbzl

EDIE

.10

123
.16

H.72524

.100
122
.17

H.65252

.100
I24
.20

161

zCLEAR REMAINDER
{GET ADDRESS OF DATA

tGET DATA
8AND TRANSFER

SDECREMENT COUNT
8LOOP IF BORE

tTRANSFER BYTE COUNT
zSAVE DATA ADDRESS

:ADJUST ADDRESS
8UPDATE STACK ADDRESS

:GET CURRENT STACKIADDRESS
300 FOR MORE

xPLOT SAMP,REF,A.T ON CRT - COORD

sMESSAGE ADDRESS
zBYTE COUNT

8GET BYTE

:LOOP IF BORE BYTES

:COUNT TO STACK?
:OUTPUT MESSAGE
:INPUT MNEMONIC
:GET BYTE COUNT
:1 BYTE?

:YES

:NO. REPLACE
SCLEAN OFF STACK
:OUTPUT '?’

:00 GET ANOTHER
IGET CHARACTER

:8?

:NO

3YES. GET'S ADDRESS

:3?
sNO
cYES. GET R.ADDRRESS

3T?
:NO

:22

JNZ

LDAX

POP

.27
A.L
40133
“T

A.D
.37

.35
40134
A.H
100
H.A

R. 60000

.65

C.0
A.H

36202
A. C
3620 1

A. I
36200
.62
B,A

A.D
.50
.47

. 45
PSW

162

gYES. D=O

3A?

3N0

:YES, D= 1
zHESSAGE ADDRESS
8COUNT

: GET BYTE

:LOOP IF MORE

8 INPUT THRESHOLD

3H SHOULD BE 0

:NOPE. OUTPUT ’7'
zAND TRY AGAIN

:STORE THRESHOLD FOR 8T CALC

8PLOT 8T OR A ?
{JUMP IF XT
:ELSE CALCULATE A

8GET STARTING ADDRESS OF A DATA
8ADJUST ADDRESS FOR COORD
:STARTING ADDRESS OF ST DATA
:SAVE ADDRESS

SSTARTING ADDRESS

8GET DATA POINT

:ST‘ORE IN Y POSITION

:STORE C IN X POSITION

:T'URN ON DOT
zGET INCREHENT

:JUI‘IP TO EXIT IF FINISHED

800 GET ANOTHER
aEXIT

361

863
:64
:65

:100

3105

2101
:102

8103

POP
POP
POP
RET

D
B
H

163

120,114.117.124,40.133.I23,54.122.54.101
54.124,135,40.76.40

HOV

HOV
MOV
ORA
JNZ
HVI
NOV

u}
E313
n
at

I‘>E!U

B,A

ooonfl
QC

UUPOO>OU°'OUO

3L
e>+g > a» above

UPMEHUUH
83

>E§°C5>15'I3>” >1l'
'3 3;; h; 3
UF‘.‘ 6 N

t BYTES/DATA WORD
8XFPOSITION
tGET OF POINTS

15.124,110,122,105.123.110.117.114.104.40.75.40

3L=0?

:SAVE ADDRESS

zGET BYTES

sGET INCREMENT FOR.PLOTTIRG
:GET BYTES

8 POINTS TO SKIP

SADDRESS IN HL
t INITIAL‘IZE X POSITION

tGET INITIAL YMAX
3 DECREI‘IENT COUNT

$GET NEXT BYTE
3D>E?

3YES

8N0. GET NEW YI‘IAX

8LOOP ’TIL DONE

15

OFF

MOV

DECIN
DECIN
CRLF
LXI
MVI
MOV
CALL
CRLF
POP2
MOV
MOV
POP2
PUSH
XCHG
CALL
LXI
CALL
CALL
POP
PUSH
XCHG
CALL
LXI
CALL
CALL
POP
PUSH
DCX
XCHG

CALL
CALL
POP
PUSH
DCX
XEHG

MOV
ORI
MOV
CALL
LXI
PUSHB

ON

D,.100
n.13
L.B
.75

.15

.10

131
.31

D,.105
H.24

.75

H.60000

.65
H
H
B

40134
A.H
100
H.A
.55

H.20040

A.2

164

:HNEHDNIC TABLE ADDRESS
zBY'I'ES

IOUTPUT PROMPT
:INPUT RESPONSE

:1 BYTE?

8YES. JUMP
:NO

:GET BYTE
gY?
:NO.EXIT

:CHANNEL NUMBER
8 CHANNELS

zADJUST ADDRESS
:SAMPLE ADDRESS
zOUTPUT SAMP
:TAB '

:REFERENCE ADDRESS

sxT ADDRESS
gOUTPUT KT

8POINTER.T0 A ADDRESS
*ADJUST ADDRESS FOR.COORD
:OUTPUT A

:70

:72

MESSAG
POP
PUSH
PUSH2
DECOUT

DCX
MOV
ORA
JZ
LHLD
PUSH2
POP
PUSH
PUSH2
SUBTR
POP2
MOV
ORA
JZ
POP

PUSH
JMP
POP
POP
POP

B
A.B

.30
40120

guumnhuumbr?
o o u

UFHU

0:!

8,4401

>aypu>g
we 3
q

11
to

165

8OUTPUT'SPACES AND CHANNEL

8EXTT IF DONE

.8'IOTAL mm or POINTS

8TOTAL POINTS ‘ CURRENT'CHANNEL

8EXFT IF AT LAST CHANNEL

8CLEAN OFF CHANNEL

8RETURN

sTAB

8 BYTES /DATA WORD

ACOOOOOGflflfl

)>
((

>>
(<

:30
>>

( < TICK

JIM HORNSHUH
CONVERSE LIBRARY FOR COMMUNICATOR MICRO

PUSH2

POP2

AXIS

PUSH

' XCHG

LHLD
DCX
MOV
DCX
MOV
SHLD
XCHG
POP

PUSH
LHLD
MOV
I NX
MOV
I NX
SHLD
XCHG
POP

PUSH
PUSH
XRA
STA
STA
OFF
OFF
STA
IN]
IN I
STA
DCR
J Z
ORI
J MP
OFF
STA
XRA
STA
MVI
MVI
STA
DCR
J Z
ORI
J MP
POP
POP

PUSH
PUSH
LXI
MOV

D

14
H
M.D
H
PLE
14

D

D

14
E.M
H
D,”
H
14

D

PSW

1 6203
1 6202

16201
A. 1

B.377
16200

B
PSW

B. 16200

A.L

166

MICROLCON
NOVEMBER 1. 1977

8CONTENT‘S OF HGL T‘O STACK

82 BYTES FROM STACK TD HOL

8DRAW X AND Y AXES ON MONITOR

:CLEAR A

sCLEAR SCREEN

:AND Y AXIS

:S'mIP OFF NUMBER OF BYTE
zGET POSITION OF x AXIS
:AND WRITE

alNTENSITY IN A

;COUNTER

:WRITE DOT

aDECREnENT COUNT

:JUI‘IP IF DONE

:ELSE. SET UP FOR IIOVE DOWN
:AND LOOP

8Y AXIS

:ZEND x
sGET INTENSITY

3WRITE DOT

8ROUTINE TO DRAW TICK MARKS

ICURSOR MOVE COMMAND

>>

STAX
STAX
STAX
NOP
NOP
NOP
NOP
ORI
STAX
STAX
STAX
S TAX
STAX
STAX

<< TICMRK

:20
:25

MOV

muuuwug

167

IMOVE CURSOR AND WRITE DOT

g DELAY

8REVERSE DIRECTION
8MOVE BACK

: DELAY

tCLEAR UPPER BITS TO REVERSE
:I‘IOVE BACK

IDRAWTICKMARKSONXANDYAXES
ISAVE STATUS AND REGISTERS

8SET UP Y ADDRESS
8CODE TO WRITE DOTS IN HORIZONTAL
ICLEAR OFF BYTE COUNT
IGET X-CURSOR. POSITION
8WRITE

8AND SAVE

IGET Y-CURSOR POSITION
zWRITE

8AND SAVE

8GET Y INCREMENT

IAND SAVE

3)0?

8N0. JUMP TO NEXT AXIS
8GET CURSOR

ISUBTRACT INCREMENT
8JUMP ON BORROW
8POSITION CURSOR
gDRAW SINGLE TICK MARK
:LOOP

IGET CURSOR AGAIN

8AND REPOSITION

zADD INCREMENT THIS TIME
xJUMP ON CARRY
8POSITION CURSOR
zDRAW SINGLE TICK MARK
sLOOP

8REPOSITION
IGET X INCREMENT
:AND SAVE

' 8>0?

8N0. EXIT

8NEW CURSOR MOVE COMMAND
zDECREMENT ADDRESS POINTER
gGET Y CURSOR

>>
<< CC

ON

POP
POP
POP
RET

'Uflflauh DWI"
m poo.
a oath

”'3
t

.10
PSW

§QUUA§

D.33

.10
.30
B.A

15
.22
1123
A, 12
1123
B
020
A.15
1123
.15
A91
035
A.3
1123
A

D

B
PSW

168

[EXCHANGE CURSORS

8GET X CURSOR
snsTUNN
:mwnms

8COMMUN. CHECKIFOR.REOUESTED CHARACTER

8GET CHARACTER IN H.1 IN L
8INPUT CHARACTER

8COMPARE WITH DESIRED CHARACTER
8JUMP IF WRONG CHARACTER

8OPEN FILE FOR WRITING - MICRO I

80UTPUT ’S‘

IINPUT COMMAND STRJNG
zCLEAN STACK JUMP ON C
sSEND COMMAND STRING TO FLOPPY
83 -> FILE OPEN FOR WRITING
xELSE JUMP BACK

:GET ADDRESS OF MNEMONICS
;GET BYTE COUNT

8GET MNEMONIC

:INCREMENT ADDRESS

8OUTPUT TO CRT

8DECREMENT BYTE COUNT

8LOOP ’TIL DONE

8INPUT COMMAND STRING
8CLEAN STACK JUMP ON C
8GET BYTE COUNT

8GET ASCII CHARACTER
:IS IT A CR

8N0

:YES, OUTPUT

8GET CODE FOR.LF
gOUTPUT TO FLOPPY
sDECREMENT BYTE COUNT
sJUMP IF MORE ‘

8OUTPUT CR

8GET ANOTHER COMMAND STRING
81 8 ABORT FILE OPENING
8JUMP EXIT

8OUTPUT EXT TO FLOPPY

111.116.120.125.124.40.124.105.130.124.40.55.40
34.103.40.124,105,122.115,111.116.101.124.105.123.15

<< CFILI

>>
<< TSF

)>

OFF
MOV
MOV
OFF
PUSH

J NZ
POP
CALL
DCR
J N2

1171
1106

. 10
1106

1 106
.20

.10

B.A
1123
. lo

C.A
B,A

PSW
. lo
PSW
1 123

.20

8CI.OSE FILE - MICRO

169

sSEND EOT

3GET ACK OR NAK

gREAD ERROR MESSAGE

:EOT?

8JUMP IF NO

IELSE EXIT

8OUTPUT TO SCREEN
8JUMP FOR MORE

8TRANSFER FROM STACK TD FLOPPY
8GET BYTE COUNT

IGET BYTE
sTRANSFER A TO FLOPPY

:DECREMENT BYTE COUNT
8JUMP IF MORE

8OUTPUT DATA TO FLOPPY -‘ MICRO
8GET BYTE COUNT

:GET BYTE

8AND PUT IT ON STACK

1

1

Ame-00.000000.

>>
((

>>
(<

))
<<

>>
<<

)>
(<

ON

OFF

J 1!! HORNSHUH

170

MICRO2.CON

NOVEMBER 1 , 1977

CONVERSE LIBRARY FOR VIDICON MICRO

PUSH
LHLD
DCX
MOV
SHLD
POP

PUSH
LHLD
MOV
I NX
SHLD
POP

PUSH2

POP2

SREF

PUSH
XCHG
LHLD
DCX
MOV
DCX
MOV
SHLD
XC HG
POP

PUSH
LHLD
MOV
I NX
MOV
I NX
SHLD
XCHG
POP

POP2
XCHG
LX I
MOV

I NX
MOV
ANA
J NZ

PUSH
PUSH
PUSH
LDA
AN I
STA
MOV
MV I

8CONTENT‘S OF REG A ON STACK

8CONTENTS OF STACK TO REG A

8 CONTENTS OF HOL TO STACK

82 BYTES FROM STACK TO HOL

8ROUTINE'TO MOVE CELL
:GET COMMAND WORD
:AND PUT IN DE

H. 1200008ADDRESS OF OUTPUT LATCH

M.D zSEND OUT SELECT COMMAND

H

A.M zGET STATUS OF CELL COMPARTMENT

E

. 10 :1F STATUS NOT 0, WAIT FOR CELL
8SELECT REFERENCE CELL

PSW 8SAVE STATUS AND RMISTERS

H

D

21 8GET CURRENT STATUS

6 {SET REFERENCE BIT

21 :UPDATE STATUS

H.A 8SET UP COMMAND WORD

L.2

PUSH2

CELL

POP

POP

POP
>>

<< SSAMP
PUSH
PUSH
PUSH
LDA
ORI
STA
MOV
MVI
PUSH2
CELL
POP
POP
POP

>>

<< DELAY
:10 DCR
R2
:20 DCR
JZ
JMP
>>

< < SOPEN
PUSH
PUSH
PUSH
:10 LXI
MOV
ARI
JNZ
STA
MVI
:18 LXI
DELAY
DCR
JNZ
JHP
:20 . POP
POP
POP

<< SCLOSE
PUSH
PUSH
PUSH
:10 LXI
MOV
ARI
JZ
STA
NV]
:15 LXI
DELAY
DCR
JNZ
JHP
:20 POP
POP
POP

171

:AND PUT'ON STACK
ICALL ROUTINE TO MOVE CELL

D :RFSNRE STATUS AND REGISTERS
H
Psw
sSELECT SAMPLE CELL
PSW sSAVE STATUS AND REGISTERS
H
D
21 :GET CURRENT STATUS
10 ;SET SAMPLE BIT
21 1UPDATE STATUS
H,A :SET UP COMMAND WORD
L 1
' :AND PUT ON STACK
:CALL ROUTINE TO MOVE CELL
D 3RES'I‘0RE STATUS AND REGISTERS
H
Psw
:PROGRAIINADLE DELAY ROUTINE
C
B
. 10
.20
1110me To OPEN SHUTTER
H
B
PSW
H.120004
A.H IGET STATUS
2 I'I'FST 811‘ 1
.20 HF BIT 1:1, ALREADY OPEN
120007 IOPER SHUTTER
A.6
E. 177777
:DELAY POR SHUTTER
A
.15
. 10 :DOUBLE CHECK
PSW
B
H
n :ROUTINE TO CLOSE SHUTTER
D
PSW
$120004
A.H :GET STATUS
2
.20 :1F BIT 1:0. ALREADY CLOSED
120007 :CLOSE SHUTTER
A.6
B. 177777
:DELAY FOR SHUTTER
A
.15
. 10 :DOUBLE CHECK
Psw
B
H

>)
<< STEP

)>
<< DOWN

PUSH
PUSH
PUSH
POP2
XCHG
OFF
MOV
LXI
MOV

I NR
MOV
LXI
DELAY
DCR
MOV
LXI
DELAY
MOV

I NX

ANI
JZ

AN I
J NZ
DCX
MOV
ORA
J NZ
POP
POP
POP
POP

CMC

POP
POP
POP
POP
STC

PUSH
POP2
LDA
AN I
ON
PUS.
STEP
J C

ON
LXI
PUSH2
STEP
J C
ORI
STA
ON
LXI
PUS-
STEP

172

t RDUTI NE TU STEP MNOC'HROMATOR

PSW ISAVE STATUS AND REGISTERS
H
D
B
:GET COUNT
sPUT IN DE
IGET STATUS
B,A
H.12000010UTPUT LATCH ADDRESS
A
M.A

H.107403ISET UP BC FOR DELAY ROUTINE
ICALL PROGRAMMABLE DELAY

A
ILA
E. 107403
B,‘
H
A.H IGET STATUS
4- :TEST REVERSE LIHIT
.20 :JUIIP IF NOT AT LIMIT
B :EISE CHECK FOR FORWARD HOTION
4
.30 11? STILL REVERSE. EXIT 01! ERROR
D E zDECREMENT COUNT AND TEST IF 8 O
A.
D
.10 :NO, CONTINUE
: IYES. EXIT UITHOUI‘ ERROR
H
PSW
:SET CARRY
xCLEAR CARRY( ERROR FLAG)
8AND RETURN.
B
D
H
PSW
ISET CARRYt ERROR FLAG)
: DECREHEIT mROCHROHAT'OR WAVELENGTH
H
IGET COUNT
2 l IGET HONOCHROHA'IOR STATUS
12 :CLEAR FORWARD BIT
IAND PLACE STATUS ON STACK
:ALSO COUNT
IIIOVE HONOCHROHATOR
.IO :RETURN ON ERROR
ISTATUS BACK ON STACK
H.200
ITAKE ADDITIONAL 128 STEPS DOWN
. IO IRETURN ON ERROR
4 :SET FORWARD BIT
21 :UPDATE STATUS
8AND PUT ON STACK
H.200

:STEP UP 128 STEPS

810

>>
<< UP

)>

CMC
POP
RET
STC
POP

PUSH
PUSH
POP2
LDA
ORI
STA
ON

PUSH2

STEP
POP
POP

<< SUBTR

>>
<< WAV

320

PUSH
PUSH
POP2
XCHG
POP2
MOV
SUB
MOV
MOV
SBB
MOV

PUSIE

POP
POP

PUSH
PUSH
PUSH
PUSH

XCHG
SHLD

PUSH2

XCHG

PUSH2
SUBTR

JC

STA
POP
POP
POP
POP

POP2
MOV
CMA
ADI
MOV

CMA

PSW
H

21
21

PSW

Uh!

avid»?

O

MU HU>FH>

.20

36
PSW

E’
r.

3’?“
In»

173

8 SET CARRY
ICLEAR.CARRY

:SET CARRY(ERROR.FLAG)

IINCREMENT MDNOCHROMATOR WAVELENGTH
ISAVE STATUS AND REGISTERS

IGET COUNT

8GET HONOCHROHATOR.STATUS

ISET FORWARD/EXTERNAL ENABLE BITS
:UPDATE STATUS

8AND PLACE ON STACK

:PLACE COUNT ON STACK

:HDVE HONOCHRONATOR

IRESTORE

SSUBTRACT UNSIGNED IO-BIT NUMBERS
ISAVE REGISTERS

SGET SUBTRAHEND
IGET MINUEND
:SUBTRACT LOW BYTES

sSUBTRACT HIGH BYTES

1ANSWER.BACKION STACK
: RESTORE

:MOVE TO NEW WAVELENGTH
ISAVE STATUS AND REGISTERS

IGET OLD WAVELENGTH

IGET NEW WAVELENGTH
:UPDATE WAVELENGTH
8AND PLACE ON STACK

gOLD WAVELENGTH ON STACK
:SUBTRACT

HUMP IF CARRY DURING SUBTRACTION
:INCREMENT WAVELENGTH

:CLEAR A

zCLEAR ERROR LOCATION

8AND RETURN

IGET ANSWER
ICOMPLEMENT AND INCREMENT COUNT'

>>

<< AUTO

>>
(< MAN

>>

ACI
MOV
PUSNE
DOWN
JNC

STA
JMP

PUSH

ORI
STA
STA
POP

PUSH
LDA

STA
STA
POP

<< CLRMEM

>>
<< DIV

PUSH
PUSH
PUSH
LXI
SHLD
SHLD
LXI
MVI
INX
INX
MOV
ANI
JZ
POP
POP
POP

PUSH
PUSH
PUSH
PUSH
OFF
MOV
POP2
MVI

MOV
MOV

MOV
MOV

MOV
DCR
JZ

MOV

JNC

0
H.A

.10
A.‘
36

.10

PSW
21

2

21
120000
PSW

PSW
21

14

21
120000
PSW

PSW
H
D

H.20000

120011
102
D.0
M.0
H

D
A.D
100
.10
D

H
PSW

E

o
H

D'OF b? >$>O 5 BEN
:30 >' IUD t‘> NI D

174

zCOUNT'BACK ON STACK
IDECREMENT MONOCHROMATOR
sJUMP IF NO ERROR

88TORE ERROR CODE IN LOC 36

IMONOCHROMATOR‘UNDER.MICRO CONTROL
sGET STATUS

ISET EXTERNAL ENABLE BIT
sUPDATE STATUS

IPUT MDNOCHROMATOR.ON MANUAL
IGET STATUS

ICLEAR EXTERNAL ENABLE BIT
:UPDATE STATUS

SCLEAR.8K DATA MEMORY

IMEMORY STARTING ADRESS
IINITIALIZE MEMORY POINTER
:INITIALIZE MEMORY POINTER
ICLEAR MEMORY

IINCREMENT MEMORY ADDRESS
8AND COUNTER

ITEST“BIT 15
8JUMP IF STILL MORE MEMORY

316-BIT'BY'B-BIT'UNSIGNED DIVISION
ISAVE STATUS AND REGISTERS

IGET DIVISOR

8GET DIVIDEND
IBIT COUNTER

zROTATE LOW BYTE

:ROTATE HIGH BYTE
6 DECREMENT COUNTER

M

SUB
J MP
SUB
J NC
ADD
J MP
MOV
CMA
MOV
MOV
CMA
MOV

PUS-

MOV
ON

POP
POP
POP
POP

<< BINDEC

:20
z 25

PUSH
PUSH
PUSH

DIV

«nu-1—
0°

u>r>no

HUG b l" >H DOU'U°U

"=3

??9§”“
EGO

PSW

PSW

GNU

PSW
I 17

H. 1750

.40
.10

A.L

175

sCOMPLEMENT HIGH BYTE OF QUOTIENT

z COHPLEMENT LOW BYTE
:QUOTIENT ON STACK

zREMAINDER ON STACK
IRESTORE STATUS AND MISTERS

:16-BIT BINARY TO 5 DIGIT BCD

IINITIALIZE BCD DIGIT COUNTHI
IDIVISOR FOR 10

:DIVIDE BINARY NUMBER BY 10
IGET REMAINDER

IAND SAVE

zDECREMENT BCD COUNT

zJUMP IF MORE LEFT

:GET QUOTIENT

sREINITIALIZE BCD COUNTER
:RECOVER REMAINDER

118 IT A ZERO?

3N0. REST OF DIGITS ARE SIGNIFICANT
IYES. DECREMENT COUNT

IJUMP IF MORET‘ODO

sPUT REMAINDER ON STACK

IGET NUMBER OF NON-ZERO DIGIT‘S
8AND PUT ON STACK

IMICRO 2 I/O ROUTINE

ICALCULATE SCAN FREQUENCY
110001012

:DIVIDE 10001012 BY DESIRED FREQ
8JUMP IF BIT 7 NOT SET

IELSE BUMP COUNT

8AND SAVE

z 22

MOV
LXI
CALL
JP

I NR
MOV
STA
MV I
MOV
PUSH2
B I N DEC
POP2
MOV
XRA
CMP

J NZ
MV I
DCR
MOV
DCR

J 2
OFF
CP I

J NZ
MV I
DCR
RAL
RAL
RAL
RAL
AN I
ORA
STA
m
STA
LXI
PUSH2
LHLD
MOV
MOV
XCHG
PUS
XCHG
SUBTR
J C
LXI

I NR

POP2
LDA
CP 1
J NC
MV I
STA
LXI
CALL
JP

I NX
SHLD
MOV
MOV
DAD
LDA
SU I
J Z

B.A
H. 1730
.40
.15

A.L
117

L.B

“w?lpafl>?
0H8 t"

° 0
"I

A. 12
A

360

H

1 10

A

. 43

H. 10000

120
D.H
E.L

.24

H,.43

.43

.26

176

IDIVIDE THIS TIME BY COUNT

ISTORE ACTUAL FREQ FOR POUT

IGET BCD DIGITS
I DIGITS
IIST DIGIT A 0?
:NO. JUMP

zYES. MAKE IT A 10
IDECREMENT

:JUMP IF ONLY 1 DIGIT
IELSE GET NEXT DIGIT
3=0?

:NO. JUMP

IYES..MAKE IT A 10

IMASK OFF LOVER 4 010118
zADD IST DIGIT
:DEPOSIT

IINITIALIZE COUNTER
84096 -) BL

INUMBER OF POINTS REQUESTED

3DE->HL
I HL-> DE

1 BUMP COUNTER

IGO SUBTRACT AGAIN
ICLEAN STACK

85 IS MINIMUM INCREMENT

:MAKE IT A 5
IST‘ORE INCREMENT

IST‘ORE POINTS FOR POUT
8 CALCULATE TOTAL BYTE COUNT

8 DOUBLE WORD COUNT
:NUMBER OF BYTES

8JUMP IF 2

DAD
SHLD

DCR
DCR
MOV

CPI

DCR

CPI

POP
POP
POP
POP

PUSH2
ON
DIV
OFF
POP2
ANA

D
100

360

105
L

122
124
112
60
B.A
H.106
A.H
317
MrA
125
12
.41
.42
A

37
.46
A.H
L
21
L,20
B.0

26
PSW

A

177

:STORE WORD COUNT
tH’- A00 8 L - DLY

tACQ+DLYKll7
zYES, JUMP

tDECREMENT’AND'TRY AGAIN

:UPDATE FOR POUT
tADJUST BLANKING

IKEEP ONLY D4 D5

:CLEAR.D4 D5
3AND SET THEM.

tsET UP CONVERTER
:10 BITS

$GET NUMBER OF BYTES

:83?

:YES. EXIT

:GET NUMBER OF SIGNAL AVERAGE!

:NUMBER OF AVERAGES < I? ?
:YES. EXPT

178

:42 MVI A.26
J"? .45

:43 0

>

(< DIV3 :24-HIT XZIG-HIT UNSIGNHD INTEGEH.DIVISIOH
PUSH tSAVE STATUS AND REGISTERS
PUSH
PUSH
PUSH
POP2
MOV
MOV
POP2
MVI
STA

i

:GET DIVISOR

:LOW BYTE TO REG C

831GB BYTE TO REG B

‘ADDRESS OF 3-BYTE DIVIDEND
I zBIT COUNT

P “FD 03 IF

MOV

MOV :ROTAT'E DIVIDEND LEFT ONE BIT
MOV
INX
MOV

MOV
INX
MOV
RAL :MSB OF DIVIDEND INTO CARRY

MOV ILA

DCX B

DCX B

LDA .25

DCR A :DECREHENT BIT COUNT

JZ .30 :1F A=0. JUMP TO EXIT

STA .25 gELSE STORE COUNT AND CONTINUE

MOV A.E

RAL :ISB OF DIVIDEND INTO REMAINDER LOU BIT
MOV E.
MOV A.
MOV D
JNC
MOV
SUB
MOV
MOV
SBB

my ya; >5m>~> no can
> a

?
a

8JUMP IF BORROW ON SUBTRACTION POSSIBLE
stBTRACT LOWER BYTES

U?§IO>'

tsUBTRACT RICH BYTES
sLOOP

:SUBTRACT LON BYTES

a» m>o a» h; a» no» a»

SUB
MOV
I‘IOV
SBB
J NC
MOV
MOV
ADD
MOV
MOV

J MP 3 AND LOOP
: 25 000 : STORAGE LOCATION FOR BIT COUNT
: 30 MOV : GET QUOT I ENT AND COMPLEMENT

Ubl‘flb’

cSUBTRACT HIGH BYTES
8JUMP IF NO BORROW
8RESTORE D

:ELSE ADD DIVISOR BACK

3’ ' U>1§6>U'
0 ~ 0 o o o p
3 Q

179

CIIA
MOV ILA
INX H
MOV A.H
CMA
IIOV ILA
INX H
MOV A.H
CMA
IIOV ILA
INX H
P US 3 ADDRESS OF NEXT DATA WORD ON STACK
XCHG
PUSH2 8REIIAINDER ON STACK
POP B {RESTORE STATUS AND RMISTERS
POP D
POP H
POP PSW
>>
<< NORIB :NORMALIZE DATA - MICRO 2
PUSH H
PUSH D
PUSH B
PUSH PSW
LDA 127 gGET MODE
RAR :BITO -> CARRY
JNC .15 zSKIP DARK IF NO CARRY
PUSH PSW
LXI H.20000 :STARTING ADDRESS OF DARK
CALL .40 8NORIIALIZE
POP PSW gGET A BACK
: l5 RAR :BITl -> CARRY
JNC .20 :SKIP REFERENCE IF NO CARRY
PUSH PSW :SAVE A
LXI B. 25252 :STARTING ADDRESS OF REFERENCE
CALL .40 zNORIIALIZE
POP PSW :GET A
:20 RAR sBIT2 -> CARRY
JNC .25 :SKIP SAMPLE IF NO CARRY
LXI B. 32524 zSTARTING ADDRESS OF SAMPLE
CALL .40
:25 POP PSW
POP B
POP D
POP H
RET
: 40 LDA 37 zGET NUMBER OF BYTES
CP I 3 : =3?
JZ .43 8YES. JUMP
XCHG 3N0. l6-BIT DATA( ADDRESS-)DE)
LHLD 120 :GET NUMBER OF POINTS
MOV B,H
IIOV ILL
:41 LDAX D :GET LOWER DATA BYTE
MOV L,A 3 IN L
INX D :INCREIIENT MEMORY POINTER
LDAX D :GET UPPER DATA BYTE
IIOV H.A 3 IN H
PUSH2 zPUSH ON STACK
LDA 26 sGET NUMBER OF AVERAGES
ON :PUSH ON STACK
DIV :NORIIALIZE
OFF zcET REMAINDER
POP2 : GET NORMALIZED DATA

ANA A :BIT'? I l?

))

JP

I NX
IIOV
STAX
DCX
MOV
STAX
I NX

I NX
DCX
MOV
ORA
J NZ
RET
XC HG

MOV
MOV
LHLD
XCHG

PUSH2

XC HG

PUSH2

XCHG
D I V3
POP2
DCX
IIOV
ORA

RET

<< SUBTRD

PUSH
PUSH
PUSH
PUSH
LXI
CALL
LXI
CALL
POP
POP
POP
POP

LXI

PUSH
LDAX
MOV
I NX
LDAX
IIOV

PUSH2

LDAX
MOV
I NX
LDAX
MOV

PUSH2
SUBTR

POP2
J NC
LXI

do
uA

l"

OEDUUHUDUUD
U

.4!

”OH.-
0' ° N
FIG

In

L

awayu yuuru
b >

$

180

3N0. JUMP
tYES, INCREMENT ANSWER

SRETURN NORMALIZED DATA TO MEMORY

:SET MEMORY POINTER TO NEXT DATA WORD
:DECREHENT COUNT

:JUMP IF MORE TO DO

:ELSE RETURN

:HERE FOR 24-BIT DATA( HL-> DE)
zGET NUMBER OF POINTS

x NUMBER OF AVERAGES

:DATA ADDRFSS TO HL

:TO STACK

:NUMBER OF AVERAGES TO BL
zTO STACK

3NORIIALIZE

ICLEAR OFF REMAINDER
8DECREMENT COUNT

8JUMP IF MORE TO DO
SELSE RETURN
38UBTRACT DARK FROM SIGNAL - MICRO 2

8ADDRESS OF REFERENCE
% ADDRESS OF SAMPLE

8ADDRESS OF DARK
8JUMBR OF POINTS
8GET LOW BYTE

{HIGH BYTE

8 POP ANSWER

835

>>
<< IT

:12

I‘IOV
STAX
DCX
MOV
STAX
INX
INX
INX
LDA
CP I

J NZ

I NX
INX
POP
DCX
I‘IOV
ORA

PUSH
JI‘IP

PUSH
PUSH
PUSH
PUSH

PUSH

I‘IOV

PUS.
SUBTR
POP2
J NC

DCX
IN I
J I‘IP
LDAX
MOV
DCX

O
H

UUUU?UU>
l"

' GOOD
lb '4
O

H I“?IIIIIUU
H

a
N

:H§§UUI

9

> 33::

HUEF‘U FUUP‘UUE'
> D

5»

.10

252
524

181

8CORRECTED DATA BACK TO I‘IEIDRY

:NUMBER.OF BYTES/DATA WORD
:83?
3N0. JUMP

aYES, INCREI‘IENT COUNTERS
tGET COUNT

8RETURN IF DONE

8CALCULATE XT - MICRO 2

8NUHBER OF POINTS
; ADDRESS OF XT DATA

3 REFERENCE ADDRESS
zSAIIPLE ADDRESS

8 REFERENCE TO STACK

88AHPLE TO STACK
zJUl‘IP IF R>S
sCLEAN OFF STACK
:T I 255 IF S>R
:GET THRESHOLD

:SUBTRACT THRESHOLD
8AND ANSWER

gIF R-S<THR, THEN T 8 255

810

>>

LDAX
MOV
SHLD

STA
LXI
PUSH2
LDAX
MOV
DCX
LDAX
MOV
PUS
D I V3
POP2
POP2
LDA

INX

<< FASET

OD
gun:

1"?HIIUU
I

huh
(I

.7
I

182

:STORE SAMPLE IN SCRATCH PAD

:SET UP SAMPLE X 256
8ADDRESS OF SCRATCH PAD

REFERENCE TO STACK
:DIVIDE (SAMP X 256)/REF
sCLEAR OFF REMAINDER
:AND ADDRESS

sGET KT

3ADDRESS OF ANSWER

BNU'I‘IBER OF BYTES/DATA WORD
:33?

3N0. JUMP

zYES. INGREDIENT COUNTERS AGAIN

:NUMBER OF POINTS

8JUMP TO EXIT IF DONE
SELSE SAVE NUMBER OF POINTS
8AND LOOP

:SET UP VIDICON SCAN PARAMETEIE

C
H. 4000038'I‘ARTING ADDRESS OF INTERFACE BOAIH)

D. 104

M.A
H

D

C
.10
2!

120000

PSW
D
B
H

xSTARTING ADDRESS OF PARAMETER STORAGE
8GET SCAN PARAMETER BYTE

8JUMP IF MORE PARAMETERS
8UPDAT'E MONOCHROMATOR STATUS

:SET UP FAST ADDER CIRCUIT

)>

PUSH
PUSH
PUSH
LXI
LDA
DCR
STAX
INX
LHLD
XCHG
MOV
INX
MOV
INX
LDA
MOV
INX
MOV
LDA
STA
POP
POP
POP

<< VINITZ

>>
<< A

VSETUP
MAN
SCLOSE
SREF
FASET
CLRMEM

PUSH
PUSH
PUSH
PUSH

D
H
PSW

183

D.120010:STARTING ADDRESS OF F.A.

37
A
D
D
102

MgE
H
M,D
H

35
"9‘

H

MLA
21
120000
PSW

H

D

mafia—wean
b 23

530
00

B. 040

°~?-?~F9~=
8 >03 93

I‘9F?>flll
O l

=QBL
1:0

8 BYTES OF PRECISION

8ADDRESS OF NEXT AVAILABLE MEM LOCATION
3L0“ BYTE OF ADDRESS
:HIGH BYTE OF ADDRESS

:F.A. STATUS
8INITIALIZE BYTE AND ADDRESS COUNTERS
8UPDATE MONOCHROMATOR.STATUS

zINITIALIZE MICRO 2

tSET UP DEFAULT SCAN PARAMETERS
:MDNOCHROMATOR.ON MANUAL CONTROL
cCLOSE SHUTTER

:SELECT REFERENCE CELL
iINITIALIZE FAST ADDER.CIRCUTT
tCLEAR DATA MEMORY

8CALCULATE A - MICRO 2

8GET NUMBER OF POINTS

{ADDRESS OF KT DATA
$STARTING ADDRESS OF A.DATA IN HL

zSTORE

zSTARTING ADDRESS OF A.TAHLE
:GET KT

:GET ADDRESS OF A IN TABLE
:GET A

8NUMBER.OF POINTS

:EXIT IF DONE
8ELSE CONTINUE

184

66.
56.
479
41.
34.
27.
23.
17.
14.
11.

6.

3.

0.

66.
56.
47.
41.
34.
27.
23.
I7.
14.
lo.

5.

2.

0.

65.
55.
46.
419
33.
27.
23.
17.
13.
10.

5.

2.

0.

65
55
46
40
33
27
22
17
13
10

5

2

0

POP H
RET
:30 000.000
940 377
360.322.300.264.252.242.234.226.221.214.210.204.201.176.173
170,165,163.160.156.154.152.150.146.144.143.141.137.136.134
133.132.130.127.126.125.123.122.121.120.117.116.115.114.113
112.111.110.107.106.105.105.104.103.102.101.101.100. 77. 76
76. 75. 74. 74. 73. 72. 72. 71. 70. 70. 67.
64. 64. 63. 62. 62. 61. 61. 60. 60. 57. 57.
54. 54. 53. 53. 52. 52. 51. 51. 51. 50. 50.
46. 45. 45. 44. 44. 44. 43. 43. 42. 42. 42.
40. 40. 37. 37. 36. 36. 36. 35. 35. 35. 34.
33. 32. 32. 32. 32. 31. 31. 31. 30. 30. 30.
26. 26. 26. 25. 25. 25. 25. 24. 24. 24. 23.
22. 22. 22. 21. 21. 21. 21. 20. 20. 20. 20.
16. 16. 16. 16. 15. 15. 15. 15. 15. 14. 14.
13. 13. 12. 12. 12. 12. 12. 11. 11. 11. 11.
10. 10. 7. 7. 7. 7. 7. 6. 6. 6. 6.
5. 5. 4. 4. 4. 4. 4. 3. 3. 3. 3.
2. 2. 2’ l, 19 1. 1. A. l. 0. 0,
>>
<< DATCOL OSIGNAL INTEGRATTON AND AVERAGING
PUSH PSW
PUSH H
PUSH B
PUSH D
LHLD 34
MOV C L
MOV A C 3 OF INTEGRATIONS
STA .25
MOV B.H 3F.A. STATUS IN REG B
LHLD 26
XCHG 3 OF AVERAGES IN DE
LXI H.106 :ADDRESS OF VIDICON STATUS
CALL .10 :CHECK SCAN FLAG
CALL .10 :CHECK AGAIN FOR SYNC
:50 XRA A
CMP C :ANY INTEGRATIONS?
CNZ .20 :YES. CO INTEGRATE
MOV A.H :GET F.A. STATUS
ORI 20 xsET F.A. START SCAN BIT
MOV B.A
STA 120013 :SET UP F.A. SCAN ENABLE
CALL .10 :START SCAN ON FLAG
DCX D :DECREMENT OF AVERAGES
MOV A.E
ORA D
JNZ .50 :JUMP IF MORE TO DO
MOV A.H
ANI 17 :CLEAR.SCAN BIT OF ADDER
STA 120013
CALL .10 :ADDER STOPS ON FLAG
LDA 120016 :CHECK ERROR FLAGS
ANI 3 ‘CLEAR NON-ERROR BITS
ON :PUSH FLAGS ON STACK
STA 120017 :CLEAR ERROR FLAGS
POP D SRESTORE
POP B
POP H
POP PSW
RET :RETURN WITH DATA IN MEMORY
:10 PUSH H
LXI H.140011
:12 MOV A.H ccET FRAME,FLAG

ANI
JZ
DCX
MOV
POP
RET
:20 MOV
ANI
MOV
STA
MOV
ANI
MOV
STA
=22 CALL
DCR
JNZ
MOV
ORI
MOV
STA
LDA
MOV
RET
:25 000
>)'
<< DDAT2
PUSH
PUSH
LXI
SHLD
FASET
SCLOSE
DATCOL
OFF
.LXI
ORA
MOV
POP
POP
>>

<< RDATZ
‘ PUSH

PUSH
LXI
SHLD
FASET
SOPEN
SREF
DATCOL
OFF
LXI
ORA
MOV
POP
POP

)>

<< SDARZ

PUSH
PUSH
LXI
SHLD
FASET
SOPEN
SSAMP

140002
.25
C.A

H

PSW
H.20000
102

PSW

H.25252
102

H.36
M
MFA
PSW
H
PSW

H.32524
102

185

ITEST BIT'7
tLOOP TIL FLAG

:CLEAR.FLAG
:RETURN
:CLEAR F.A. SCAN BIT

:GET VIDICON STATUS
:CLEAR SCAN BIT

leTEGRATION STARTS ON FLAG
IDECREMENT INTEGRATTON COUNT
:LOOP IF MORE INTEGRATIONS
8ELSE. RETURN FOR READOUT
stT SCAN BIT

8ENABLE SCAN CIRCUIT
SRESTORE INTEGRATION COUNT
SINTECRATION COUNT STORAGE
8DARK DATA - MICRO 2

1INITIALIZE ADDER
ICHECK SHUTTER
:COLLECT DATA

SGET STATUS

zREFERENCE DATA - MICRO 2

88TORE STARTING ADDRESS
8INITIALIZE ADDER
zCHECK SHUTTER

SSELECT REFERENCE CELL
sCOLLECT DATA

:GET STATUS

:SAMPLE DATA - MICRO 2

88TORE STARTING ADDRESS
:INITIALIZE ADDER
:CHECK SHUTTER

:SELECT SAMPLE CELL

))

DATCOL
OFF
LXI
ORA
MOV
POP
POP

<< DATAZ

PUSH
PUSH
PUSH

STA
LDA

JNC
PUSH
DDAT2
POP

JNC
PUSH
RDAT2
POP

JNC
SDAT2
POP
POP
POP

PSW
.35
PSW

186

scOLLECT’DATA
8GET STATUS

10R. STATUS WITH MEMORY
cSTORE STATUS

:COLLECT DATA - MICRO 2

3CLEAR ERROR LOCATION
IGET MODE
sBITO -) CARRY

';SKIP DARK IF NO CARRY

:COLLECT DARK

zBITl -) CARRY
zSKIP REFERENCE IF NO CARRY

xCOLLECT REFERENCE
xBIT2 -> CARRY

:SKIP SAMPLE IF NO CARRY
:COLLECT SAMPLE

000000000000

10

000

6000 N000 N000

0000

187

VEDO.FT‘N
JIM HORNSHUH NOVEMBER 21 . I977

PROGRAM TO EVALUATE AND DISPLAY SPECTRAL DATA
COLLECTED BY THE CG ENKE VIDICON SPECTROMETER

*1: COMPILE WITH ION SWITCH *1:

CALL SUBROUTINFS: T4010.SMOOTH. FORPIP.U'NFM.
TEXT.RFILE. REPACK

BYTE IFILE( 24) . IBUFF(256) .ANS.Y. NN(2) .NM(2)
BYTE JWORKI2460)

DOUBLE PRECISION Sl.R1.D1

DIMENSION ISAMP( 819) . IREF(819) . IDARK(819)
DIMENSION IWORK(819)

EQUIVALENCE ( ICNT.NN) . (NBYT.NM)

DATA Y /'Y’/

WRITE (6.900)

READ (6.901) IFILE

[MODE-=0

CHECK TO SEE IF FILE EXITS ORERROR

CALL FORPIP (IFILE.IMODE)
IF (IMODE.EO.. 1) GOTO 20
WRITE(6.902)

GOTO 10

OPEN FILE FOR UNFORMAT'I'ED BINARY READ

IERR=0

CALL UNFMO( IERR. IFILE.0. 1)
IF( IERR.EO..0) GOTO 25
WRITE(6.933) IERR

GOTO 40

READ 256 BYTES FROM FILE

IBYTE=256

IERR=0

CALL UNFMRI“ IERR. IBUFF. IBYTE)
IF( IERR.NE. 1) GOTO 30
WRITE(6.904)

GOTO 40

OUTPUT ASCII TEXT '10 TERMINAL

ammo

INDEX=1

CALL mm IBUI-‘F. mm. INDEX. 1mm»
IF(NEND.EQ. 1) com 35
IF(IBYTE.EQ.0) com 25

ALL TEXT OUTPUT - SEE IF FILE IS TO BE PROCM

WRITE(6.905)
READ(6.906) ANS
IF(ANS.EQ.Y) GOTO 50
CALL UNFMC( IERR)
WRITE(6.907)
READ(6.906) ANS
IF(ANS.EQ.Y) GOTO 10
CALL EXIT

0000

60

g 000 g 000

#000

110

120

000

188

BEGIN PROCESSING DATA - DARK FIRST

DO 55 181.819

IREF( I) 80

ISAMP( I) =0

IDARK( I) 80

CONTINUE

ICODE= 1

GOTO 100

ISCALD= 1

NAVG-=0

CALL REPACK( IDARK. JWORK. ICNT. ISCALD.NAVG)

PROCESS REFERENCE DATA NEXT

ICODE= 2

GOTO 1 00

ISCALR= 1

CALL REPACK( IREF. JWORK. ICNT. ISCALR. NAVG)

PROCESS SAMPLE DATA LAST

ICODE=3

GOTO 100

ISCALS= 1

CALL REPACK( ISAMP.JWORK. ICNT. ISCALS. NAVG)
GOTO 120

DATA PROCESSING ROUTINE

IF( IBYTE.EQ.0) CALL RFILE( mm. mm. INDEXJBUFF)
NN< 1)=IBUFF( INDEX)

1NDEX=INDEX+1

IBYTE=IBYTE-I

1F(IBYTE.EQ.O) CALL anux mm. mm. INDEX. mum
NM 2) = IBUFFC INDEX)

INDEX=INDEX+1

13YTE=IBYTE-l

no no I=I.ICNT

IF( IBYTE.EQ.0) CALL RFILEt mm. 1m. INDEX. mum
mem I) = mum INDEX)

INDEX= INDEX-I- 1

IBY’I‘E=IBYTE-1

CONTINUE

com (60.70.80). 1Com:

PROCESSING COMPLETE - CLOSE INPUT FILE

CALL UNFMC( IERR)

IF( IERR.NE.0) WRITE (6.934) IERR
NBYT¥0

NM( 1)=JWORK( 3)

NPT=( ICNT-3)/NBYT

NORMALIZE DATA BUFFERS

MAX-‘- MAXO( ISCALS. ISCALD. ISCALR)
WRITE( 6 . 929) MAX. NAVG

READ( 6 .909) NORM

IF( NORM. LE. MAX) NORM=MAX

IF( NORM. GE . NAVG) NORM= NAVG

IF( NORM. E0. 0) NORM=1

SCALS= FLOAT( ISCALS) /FLOAT( NORM)
SCALR= FLOAT( ISCALR) /FLOAT( NORM)

130

400

410

301
402
200

0650

190
191

189

SCALD= FLOAT( ISCALD) IFLOA'K NORM)
DO 130 III.NPT
SI=FLOAT(ISAMP(I))+32767.
R1=FLOAT(IREF(I))+32767.
D1=FLOAT<IDARK(I))+32767.
Sl=Sl*SCALS+O.5

Rl= R1 *SCALR+O . 5
Dl=DleCALD+O.5
ISAMP(I)=INT(SI-32767.)
IREF(I)=1NT(RJ-32767.)
IDARK(I)=1NT(Dl-32767.)
CONTINUE

BEGIN OUTPUT SECTION

WRITE(6.932)
READ(6.906) ANS
IF(ANS.EO..Y) GOTO 111
ICODE=1

GOTO 402

SUBTRACT DARK SIGNAL

DO 993 131.819

IWORK(I)=0

CONTINUE

DAVG=0

DO 400 1'1.NPT
DAVG=DAVG+FLOAT(IDARK(1))
DAVG=DAVGIFLOAT(NPT)
FSPOS=32767.

FSNEG=-32767.

DO 410 I=1.NPT

S=ISAMP(I)

RFIREF(I)

D=IDARK(I)

S=S-D+DAVG

RFRPD+DAVG
IF((S-FSNEG).LT.0) S=FSNEG
IF((S-FSPOS).GT.0) S=FSPOS
IF((RrFSNEG).LT.0) RFFSNEG
IF((R?FSPOS).GT.0) REFSPOS
ISAMP(I)=INT(S+0.5)
IREF(I)=INT(R&0.5)
WRITE(6.908)

WRITE(6.917)

READ(6.909) IA

ICODE=1

IF( IA.LT.0) GOTO 402
IF(IA.EQ.0) GOTO 2

GOTO f190.201.202.203.204.500.600.700.800
1899.860.820.840).IA
WRITE(6.908)

GOTO (1.2).ICODE
WRITE(6.910) (IWORKII).I=1.NPTT
GOTO 1

WRITE(6.928)

READ(6.909)IA

ICODE=2

IF(IA)402.1.301

DARK

DO 191 I=I.NPT
IWORK( I): IDARK( I)

£8000
HI‘

19000 1010000
HO
”N

O
69

303
312

304
305
213

204

306
313

303
309

310
311
214

300

(0000

190

GOTO (200.300).ICODE
SAMPLE

DO 211 I8I.NPT
IWORKKI)=ISAMP(I)
GOTO (200.300).ICODE

REFERENCE

DO 212 I'I.NPT
IWORK( I)= IREF( I)
GOTO (200.300).ICODE

KT

WRITE(6.911)
READ(6.909) ITH
THR=ITH

DO 213 I=I.NPT
RFIREF(I)

.S=ISAMP(I)

IF((RrS)-THR) 304.312.312
T31000.*(S-DAVG)/(RPDAVG)
IWORK(I)=T

IF(1000-IWORK(I)) 304.304.305
IWORK(I)=1000

CONTINUE

CONTINUE

GOTO (200.300) ICODE

ABSORBANCE

WRITE(6.9II)

READ(6.909) ITH

THRFITH

DO 214 I81.NPT

RFIREF(I)

S=ISAMP(I)

IF((RPS)-THR) 310.313.313
T*(S-DAVG)/(RPDAVG)
IF(1.-T) 310,310,308
IF(T) 310.310.309
IWORK(I)=-1000.*AL0010(T9
GOTO 311

IWORK(I)I0

CONTINUE

CONTINUE

GOTO (200.300) ICODE
CONTINUE

M?0

M=M§1

N=NPT

FIND MINIMUM AND MAXIMUM AND SCALE FOR PLOTTING

PMIN= IWORK(M)
PMAX=IWORK(M)
JMIN=M

JMAX=M

DO 8 I=M.N
TSIWORK(I)
IF(T‘PMIN) 5.6.6
PMIN‘T

JMIN'I

“O

@000

0000

000a

H000

19

191

PMAX=T

JMAX=I

CONTINUE
WRITE(6.9I2)PMIN.JMIN
WRITE(6.9I3) PMAX.JMAX
WRITE(6.914)
READ(6.9I5) SMIN
WRITE(6.916)
READ(6.915) SMAX
YSCALE= 600 . /( SMAX-SMIN)
XSCALE= 1000 . /FLOAT( NPT')
WRITE(6.918)
READ(6.909) 1AM
NCODE=0

IF( IAM.NE.0) GOTO 17

LINE PLOT
IX= XSCA1.E*FLOAT( M)

IY= ( FLOAT( IWORK( I ) ) -SMI N) *YSCALE
IF( IY. GT.600) IY=600

- IF( IY.LT.0) IY=0

ERASE SCREEN. GO TO GRAPHICS MODE. DRAW AXIS.
TICK MARKS. AND DARK VECTOR TO FIRST POINT

CALL T4010(0)

CALL T4010(5.10. 10)
CALL T4010(1 . IX. IY)
L=M+1

DO 18 I=L.N

IX= )(SCALE*FLOAT( I)
YI=( FLOAT( IWORK( I) )-SMIN)*YSCALE
IF( YI.GT.600.) YI=600.
IF(YI.LT.0) YI=0.
IY=YI

CALL T4010(2. IX. IY)
CONTINUE

RETURN TO ALPHA MODE AND HOME

CALL T4010( 1.0.780)
CALL T4010(3)
GOTO 19

POINT PLOT

CONTINUE

CALL T4010(0)

CALL T4010(5. 10. 10)

DO 9 I=M.N

IX= XSCALFXFLOAT( I)

YI= ( FLOAT( IWORK( I) )-SMI N) *YSCALE
IF( YI.GT.600.) YI=600.
IF( YI.LT.0) YI=0.

IY=YI

CALL T401006. IX. IY)
CONTINUE

CALL T4010(1.0.780)
CALL T4010(3)

IF( NCODE.EQ. 1) GOTO 850
WRIT’E(6.919)
READ(6.909) ISIZE

IF( ISIZE.LE.0) GOTO 2

15

500

510
506

507
511

602

604

“I000

192

IF( ISIZE.LT.5) GOTO 15

IF( ISIZE.GT.25) GOTO 15

M=( ISIZE+1)/2

N=NPT—M

CALL SMOOTH( IW’ORK. NPT. ISIZE)
GOTO 3

WRITE(6.925)

GOTO 19

ROUTINE TO OUTPUT DATAFILE TO STORAGE DEVICE

WRITE(6.920)
READ(6.901) IFILE
CALL ASSIGN(4.IFILE.24.IERR)
WRITE(4.926) N

IF( IA.LE.3) GOTO 506
D0 510 I=M.N
X=FLOAT( I)

Y= FLOAT( I WORK( I))
WRIT‘E(4.927) X.Y
CONTINUE

GOTO 511

_ DO 507 I=M.N

X=FLOAT( I)

Y= FLOAT( IWORK( I) )+32767.
WRITE(4.927) X.Y
CONTINUE

END FILE 4

GOTO (1.2) . ICODE

LOCATE

WRITE(6.931)
READ(6.906) ANS
NCODE=1

IF( IA.NE.0) GOTO 17
GOTO 870

IF(ANS.NE.Y) GOTO 602
CALL T4010(4. IX. IY)
XI=FLOAT( IX) /XSCALE
I= INT(XI)
WRITE(6.930) I

LOC=I

GOTO 702

WRITE(6.921)
READ(6.909) DOC
IF(LOC.LE.0) GOTO 602
IF(LOC.GE.NPT) GOTO 602
I X= XSCALE*FLOAT( LOC)
IY=0

CALL T4010( 1. IX. IY)
YI=( FLOAT( IWORK( LOC) )-( SMIN+5 . ) )*YSCALE
IY= INT(YI)

CALL T4010(2. IX. IY)
CALL T4010(1.0.780)
CALL T4010(3)

GOTO 702

WRITE(6.922)
READ(6.906) ANS
IF(ANS.EQ.Y) GOTO 600
GOTO ( 1 .2) . ICODE

EVALUAT‘E
WRITE( 6 . 923)

702

o

820

865

899

193

READ(6.909) LOC

IF( LOC. LE.0) GOTO 700
IF(LOC.GT.NPT) GOTO 700
S=ISAMP(LOC)
RFIREF(LOC)

S=S-DAVG

R=BPDAVG

128/R

A=-ALOG10(T)
WRITE(6.924) S.R.T.A
IF(NCODE.EQ.1) GOTO 604
GOTO (1.2).ICODE

GOTO 10

CALCULATE DERIVATIVES

WRITE(6.941)
READ(6.909) ISTEP
ISTEP=ISTEP/2
ISTART=ISTEP+1
IEND=NPT-ISTEP
WRITE(6.943)
READ(6.909) IORD

DO 825 I=ISTART.IEND
IDARK(I)=IWORK(1+ISTEP)-IWORK(I-ISTEP)
CONTINUE

DO 827 I=ISTART.IEND
IWORK(I)=IDARK(I)
IORD=IORD-l
IF(IORD.GT;0) GOTO 824
GOTO (200.300). ICODE

DUAL WAVELENGTH SCANNING

WRITE(6.942)
READ(6.909) ICHAN
JCHAN=IWORK(ICHAN)

DO 844 I=1.NPT
IWORK(I)=IWORK(I)-JCHAN
GOTO (200.300).ICODE

CALCULATE S/N AND STANDARD’DEVIATION

WRITE(6.935)

READ(6.915) SATD

WRITE(6.936)

READ(6.909) IPEAK

WRI TE( 6 . 937)

READ(6.938) IST.IEN

SUM=0.

SUMSQ=0.

DO 865 I=IST.IEN
SUM=SUM+IWORK(I)
SUMSO=SUMSQ+IWORK(I)**2
XMEAN=SUM/FLOAT(IEN-IST+1)
DEV=(SUMSO/FLOAT(IEN-ISTW1))-(XMEAN**2)
DEV=SQRT(DEV)
XMEAN=XMEAN*(10000./4096.)
DEV=DEV*(10000./4096.)
SNM=SATD/DEV
SNP=IWORK(IPEAK)/DEV
WRITE(6.939) SNM.SNP.XMEAN.DEV
GOTO (1.2).ICODE

CALL EXIT
8*******$**************mn**u**n***tm*m

900
90 1
902
904
905
906
907
908

909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924

925
926
927
928
929
930
931
932
933
934
935
936
937
938
939

940
941
942
943

194

FORMATUOFILE SPECIFICATION - ‘)
FORMAT(24AI)

FORMATt ’ tam FILE ERROR xzx')

FORMATU READ ERROR ')

FORMATUSPROCESS FILE 1")

FORMAT( 1A1)

FORMAT( 'sANOTHER FILE 1")

FORMAT(’ DATA OUTPUT CODES: V’ O t NOTHING '/
1' 1 a DARK'I' 2 = SAMPLEV' 3 = REFERENCE'/
2’ 4 = ZT'/' 5 a ABSORBANCEV’ 6 a WRITE

3 FILE'/’ 7 = LOCATEv' 8= EVALUATE'I

4' 9 = ANOTHER FILE'I' IO = EXITV

5' II 8 SN'/’ 12 = DERIVATIVEV

6' 13 = DUAL WAVELENGTH SCANNING'I)

FORMAT( I5)

FORMATI /( IX.7( I7.4X)))

FORMATt 'CTHRESHOLD : ')

FORMATU' MIN = '.F.' AT POINT ’.I5)

FORMATU MAX = ’.F.' AT POINT '.I5./)

FORMA'N 'OSCREEN BOTTOM I ')

FORIIAT(FIO.4)

FORMAT( ’GSCREEN TOP 8 ')

FORMAT( 'CLIST : ’)

FORMATt'sLINE PLOT = o, POINT PLOT ‘4 I : CHOICE : ‘)
FORMAT( 'SLENGTH OF SMOOTH IS ')

FORMAT( 'BOUTPUT FILE NAME - ')

PORMAT( ’OPOINT TO LOCATE IS ')

FORMAT(/. 'GLOCATE ANOTHER POINT ?‘)

FORMAT( ‘SPOINT TO EVALUATE IS ')

FORMATUOX.’ S'.18x.' R'.18x.’ T'.18x.' A’./
12X.F,5X.F.5X,F.5X.F)

FORMATt 'ORANGE OF SMOOTH 155 TO 25 POIN'I‘SI')
FORMAT( 'NU' . 15)

FORMAT(’RD'.2(EI5.7))

FORIIAT( 'SPLOT : ')

FORMATt'sNORMALIZATION FACTOR [’.I5.'.'.I5'] : °)
FORMATUO POINT NUMBER IS '.I5./)

FORMAT< 'sGRAPHICAL INPUT 7')

FORMAT( 'sSUBTRACT DARK SIGNAL '2')

FORMAT( ' ERROR WHILE OPENING FILE - IERR 8 '.I5)
NORMAN ’ ERROR WHILE CLOSING FILE - IERR 8 ".15)
FORMAT( ‘SSATURATION LEVEL 8 ')

FORMATVOPEAK LOCATION IS ‘)

FORMAT( ’BINTERVAL FOR RIB NOISE IS ')

FORMAT( 2( 15))

FORMAT( 'SMAXIMUM S/N - ’.Flo.2.5x.

1’ PEAK S/N = ’.FIO.2./' MEAN I ‘.F10.2.5X
2' STANDARD DEVIATION 8 ‘.F12.4./)

FORMAT( 'SSAMPLE OR REFERENCE ARRAY [l/RI : ')
FORMATPSCHANNEL WIDTH Is [EVEN J : ')
FORMAT! 'BREFERENCE CHANNEL IS ')
FORMA'N'BORDER OF DERIVATIVE IS ')

END

0000000000000

10

0000 N000 N000 000

(@000
a

3

195

VCVEDO . FTN
J III HORNSHUH NOVEMBER 28. 1977

PROGRAM TO EVALUATE ANI) DISPLAY SPECTRAL DATA COLLECTED
BY THE CG ENKE VIDICON SPECTROMETER

FOR USE ONLY ON THE DEPARTMENTAL PDP 11 COMPUTER
*8: COHPILE WITH /ON SWITCH *1:
CALLS SUBROUTINFS: VECTOR.SI‘IOOTB,FORPIP,UNPI‘I.TEXT,RFILE.REPACK

BYTE IFILE( 24) . IBUFF( 256) ,ANS,YES.NN(2) .NI‘I( 2) .JWORK( 2460)
DOUBLE PRECISION SI,R1.D1

DIMENSION ISAI‘IP(819), IREF(819) . IDARK(819)
DIMENSION IWORK(819)

EQUIVALENCE ( ICNT.NN) ,(NBYT.NI'D

COMMON /VEC'IY)R/XSCALE. YSCALE. IDEV

COMMON /VTEK/XUPPER.YUPPER.NIBS

INTEGER PMODE

DATA YES /'Y’/

WRITE (6.900)

READ (6.901) IFILE

IMODE=0

CHECK 'fl) SEE [1" FILE EXITS 0R ERROR

CALL FORPIP (IFILE. IIIODE)
IF {IMODE.EQ. l) GOTO 20
WRITE(6,902) Il‘IODE

GOTO 10

OPEN FILE FOR UNFORI‘IA'I'I'ED BINARY READ

IERR=0

CALL UNFIIO( IERR, IFILE.0. 1)
IF( IERR.EQ.0) GOTO 25
WRITE(6,903) IERR

GOTO 40

READ 256 BYTES FROM FILE

IBYTE=256

IERR=0

CALL UNFPIRI“ IERR. IBUPF . IBYTE)
IF( [ERR.NE.1) GOTO 30
WRITE(6,904)

GOTO 40

OUTPUT ASCII TEXT TO TERMINAL

NEND=0

INDEX=1

CALL TEXT( IBUFF, IBYTE, INDEX. NERD)
IF(NEND.EQ.1) GOTO 35

IF( IBYTE.EQ.0) GOTO 25

ALL TEXT OUTPUT - SEE IF FILE IS TO BE PROCESSED

WRITE(6.905)
READ(6.906) ANS
IF(ANS.EQ.YES) GOTO 50
CALL UNFPIC( IERR)
WRITE(6.907)

(I000

60

8 000 g 000

H000

110

120

000

196

READ(6.906) ANS
IF( ANS. EQ. YES) G0“) 10
CALL EXIT

BEGIN PROCESSING DATA - DARK FIRST

DO 55 I3 1.819

IREF( 1) =0

ISAMP( I) =0

IDARK( I)=0

CONTINUE

ICODE3 1

GOTO 100

ISCALD=1

NAVG=0

CALL REPACKI IDARK.JWORK. ICNT. ISCALD.NAVG)

PROCESS REFERENCE DATA NEXT

ICODE=2

GOTO 100

ISCALR= 1

CALL REPACK( IREF .JII'ORK. ICNT. ISCALR. NAVG)

PROCESS SAMPLE DATA LAST

ICODE=3

GOTO 100

ISCALS= 1

CALL REPACK( ISAMP. JWORK. ICNT. ISCALS. NAVG)
com 120

DATA PROCESSING ROUTINE

IF( IBYTE.EQ.0) CALL RFILE( IBYTE. IERR. INDEX. lBUFF)
NN( l)=IBUFF( INDEX)

INDEX= INDEX+ 1

IBYTE=IBYTE-1

IF( lBYTE.EQ.0) CALL RFILE( IBYTE. IERR. INDEX. lBUFF)
NN( 2) = IBUFF( INDEX)

INDEX= INDEX+ 1

IBYTE=IBYTE-1

DO!101=1,ICNT

IF( lBYTE.EQ.0) CALL RFILE( 1BYTE.IERR. INDEX. IBUFF)
JWOR.K( I)= IBUFF( INDEX)

INDEX: INDEX+1

IBYTE= IBYTE-l

CONTINUE

GOTO (60.70.80) . ICODE

PROCESSING COMPLETE - CLOSE INPUT FILE

CALL UNFMC( IERR)
NBYT=0

NM( 1) =JNORK( 3)
NPT=( ICNT-3)/NBYT

NORMALIZE DATA BUFFERS

MAX=MAXO( ISCALS. ISCALD. ISCALR)
WRITE( 6 .929) MAX. NAVG

READ( 6 . 909) NORM

IF( NORM. LE. MAX) NORM=MAX

IF( NORM. GE. NAVG) NORM= NAVG

IF( NORM. E0. 0) NORM=1

130

-0C§0

993

409

410

301
402

200

197

SCALS=FLOAT(ISCALS)/FLOAT(NORM)
SCALR=FLOAT(ISCALR)/FLOAT(NORM)
SCALD=FLOAT(ISCALD)/FLOAT(NORM)
DO 130 I=1.NPT
SI=FLOAT(ISAMP(I))+32767.
R1=FLOAT(IREF(I))+32767.
D1=FLOAT(IDARK(I))+32767.
SI=Sl*SCALS+O.5
R1=RI*SCALR+0.5
D1=DI*SCALD+0.5
ISAMP(I)=INT(SI-32767.)
IREF(I)=INT(Rl-32767.)
IDARK(I)=INT(Dl-32767.)
CONTINUE

BEGIN OUTPUT SECTION

WRITE(6.912)
READ(6.906) ANS
IF(ANS.EQ.YES) GOTO 111
ICODE=1 .
GOTO 402

SUB'I'RACT DARK SIGNAL

DO 993 I=1.819

IWORK(I)=0

CONTINUE

DAVG=0

DO 409 I=1.NPT
DAVG=DAVG+FLOAT(IDARKIII)
DAVG=DAVG/FLOAT(NPT)
FSPOS=32767.

FSNEG=-32767.

DO 410 I=1.NPT

S=ISAMP(I)

R=IREF(I)

D= IDARK( I)

S=S-D+DAVG

R=RrD+DAVG
IF((S-FSNEG).LT.0) S=FSNEG
IF(IS-FSPOS).GT.0) S=FSPOS
IF(IRrFSNEG).LT.0) RFFSNEG
IF((RrFSPOS).GT.0) RFFSPOS
ISAMP(I)=INT(S+0.5)
IREF(I)=INT(R#0.5)
WRITE(6.908)

WRITE(6.917)

READ(6.909) IA

ICODE=1

IF( IA.EO..0) 0011) 4
IF(IA.LT.0) GOTO 402

GOTO (190.201.202.203.204.500.600.700.800.899,820.840).IA
WRITE(6.908)

GOTO (1.2.4).ICODE
WRITE(6.9IO) (INORK(I).I=1.NPT)
GOTO 1

WRITE(6.928)

READ(6.909)IA

ICODE=2

IF(IA)402.I.301
WRITE(6.932)

READC6.909) IA

ICODE=3

IF(IA) 402.2.301

400

190
191

201
211

202
212

203

303
312

304
305
213

204

306
313

308
309

310
311
214

300

@000

198

VRITE(5.910) IIWORK(I),I=1,NPT)
GOTO 4

DARK

DO 191 I=1.NPT'
IWORK( I)= IDARK( I)
GOTO (200.300.400),ICODE

SAMPLE

DO 211 I=I.NPT
IWORK(1)= ISAMP( I)
GOTO (200,300.400),ICODE

REFERENCE

DO 212 I=1,NPT
INORK( I)= IREF( I)
GOTO (200.300.400).ICODE

%T

WRITEI6.911)

READ(6.909) 1TH

TRR=ITR

DO 213 I=1,NPT

RFIREF(I)

S=ISAMP(I)

IF((RrS)-THR) 304.312.312
1*1000.*(S-DAYG)/(RrDAVG)
IWORK(I)=T
IF(1000-IWORK(I)) 304,304,305
INORK(I)=1000

CONTINUE

CONTINUE

GOTO (200.300.400) ICODE

ABSORBANCE

WRITE(6.911)

READ(6.909) 1TH

THR=ITR

DO 214 I=1.NPT

RFIREFII)

S=ISAMP(I)

IF(IRFS)—THR) 310,313,313
1*(S-DAVG)/(RrDAVG)
IF(l.-T) 310.310.308
IF(T) 310.310.309
IWORK(I)=-1000.03ALOG10I19
GOTO 311

IWORK(I)=0

CONTINUE

CONTINUE

GOTO (200.300.400) ICODE
CONTINUE

UFO

M=M+I

N=NPT

FIND MINIMUM AND MAXIMUM AND SCALE FOR PLOTTING

PMIN=IWORK(M)
PMAX=IHORK(M)

G) RIG a

0650

0€30

420
19

15

500

199

JMIN=M

JMAX=M

DO 8 I=M.N
TPIWORK(I)
IF(T—PMIN) 5,6,6
PMIN=T

JMIN=I

IF(PMAXPT) 7.8.8
PMAX=T

JMAX=I

CONTINUE
XSCALE=FLOAT(N)/12.0
YSCALE=ABS(PMAXFPMIN)/9.0
WRITE(6.918)
READ(6,909) PMDDE
PMODE=PMODE+2

INITIALIZE VECTOR GENERAL AND DRAW AXES

XUPPER=12.0
YUPPER=9.0

IDEV=1

CALL VECTOR(0..0..0)
Y=ABS<PMAX-PMIN)
CALL VECTOR(0..Y.1)
CALL VECNR(0. .0. ,2)
X=FLOAT(N)

CALL VECTOR( X.0. .2)
CALL VECTOR(0.,0..1)

PLOT DATA FILE

DO 420 I=M.N

XSFLOATII)

YSFLOAT(IWORKII))
YSABS(Y‘PMIN)

IF(PMODE.EQ.3) CALL VECTORIXAY.1)
CALL VECTOR(X.Y.2)

CONTINUE

CALL VECTOR( X. Y. 4)
WRITE(6,919)

READ(6.909) ISIZE
IF(ISIZE.LE.0) GOTO 2
IF(ISIZE.LT.5) GOTO 15
IF(ISIZE.GT.25) GOTO 15
M=(ISIZE+1)/2

N=NPT5M

CALL SMDOTR(IWORK.NPT.ISIZE)
GOTO 3

WRITE(6.925)

GOTO 19

ROUTINE TO OUTPUT DATAFILE TO STORAGE DEVICE

WRITE(6.920)
READI6.901) IFILE
CALL ASSIGN(4.IFILE.24.IERR)
WRITE(4.926) N
IF(IA.LE.3) GOTO 507
D0 510 I=M.N
X=FLOAT(I)
YSFLOAT(IWORK(I))
WRITE(4.927) X.Y
CONTINUE

GOTO 511

200

507 D0 506 I=M.N
X=FLOAT(I)
Y=FLOAT(IWORK(I))+32767.
WRITE(4.927) X.Y

506 CONTINUE

511 END FILE 4
GOTO (1.2.4).ICODE

C

C LOCATE

C

600 WRITE(6.931)
READ(6.906) ANS

850 IF(ANS.NE.YES) GOTO 602

CALL VECTOR( X. Y, - 1 )
I=INT(X%0.5)
NRITE(6.930) I
LOC=I
GOTO 702

602 WRITE(6.921)
READ(6,909) LOC
IF(LOC.LE.0) GOTO 602
IF(LOC.GE.NPT) GOTO 602

505 X=FLOAT(LOC)
Y=0.
CALL VECTOR( X. Y, 1)
YSFLOAT(IWORK(LOC))
Y=ABS(Y¥PMIN)
CALL VECTORI X. Y. 2)
CALL VECTOR( X. Y, 4)
GO'IY) 702

604 WRITE(6.922)
READ(6.906) ANS
IF(ANS.EQ.YES) GOTO 600
GOTO (1.2.4).ICODE

C

C EVALUATE

C

700 WRITE(6,923)
READ(6.909) LOC

702 IF(LOC.LE.0) GOTO 700

IF(LOC.GT.NPT) GOTO 700
S=ISAMP(LOC)
R=IREF(LOC)

S=S-DAVG

R=RrDAVG

1¥S/R

A=-ALOGIO(19
WRITE(6.924) S.R.T.A
IF(IA.EQ.6) GOTO 604
GOTO (1,2,4).ICODE

800 GOTO 10

C

C CALCULATE DERIVATIVES
C

820 WRITE(6.941)

READ(6.909) ISTEP

ISTEP=ISTEP/2

ISTART=ISTEP+I

IEND=NPT-ISTEP

WRITE(6.943)

READ(6.909) IORD
824 D0 825 I=ISTART.IEND

IDARK(I)=IWORK(I+ISTEP)-INORKII-ISTEP)
825 CONTINUE

DO 827 I=ISTART,IEND

902
903
904
905
906
907
908

909
910
911
912
915
917
918
919
920
921
922
923
924

925
926
927
928
929
930
931
932
941
942
943

201

IWORK( I)= IDARK( I)
IORD=IORD-1

IF( IORD.GT.0) GOTO 824
GOTO (200.300.400).ICODE

DUAL WAVELENGTH SCANNING

WRITE(6.942)

READ(6,909) ICHAN
IV=1WORK(ICHAN)

DO 845 I=1.NPT
INORK(I)=1WORK(I)-IV
GOTO (200,300.400).ICODE

CALL EXIT
atm-an:xxx:**************axxxxxntmxutmxxntxxm
FORMAT( 'SFILE SPECIFICATION - ')

F ORMAT( 24A1 )

FORMAT(' FORPIP FILE ERROR - MODE ‘ '.I4)

FORMAT(' ERROR WHILE OPENING FILE - IERR.= '.I5)
FORMAT(’ READ ERROR ')

FORMAT(‘SPROCESS FILE 7')

FORMAT(IAI)

FORMAT(‘SANOTHER FILE ?')

FORMAT(’ DATA OUTPUT CODES: 'l’ 0 = NOTHING'I' 1 8 DARK7/
1’ 2 3 SAMPLE’l’ 3 = REFERENCE'/‘ 4 = XT'/’ 5 = ABSORBANCE‘/
2' 6 8 WRITE FILE'/' 7 = LOCATE'I' 8 = EVALUATE'/

3’ 9 3 ANOTHER FILE’/’ 10 = EXIT’/’ 11 3 DERIVATIVE’I
4' 12 = DUAL WAVELENGTH SCANNING'I)

FORMAT(IS)

FORMAT(/(1X.7II7.4X)))

FORMAT(‘BTHRESHOLD : ’)

FORMAT(‘SSUBTRACT DARK SIGNAL 7')

FORMAT(F10.4)

FORMAT(‘SLIST [KB] : ')

FORMAT(‘BLINE PLOT = 0. POINT PLOT = I 3 CHOICE : ’)
FORMAT(‘SLENGTH OF SMOOTH IS ')

FORMAT(‘OOUTPUT FILE NAME - ')

FORMAT('8POINT TO LOCATE IS ')

FORMAT(/.’CLOCATE ANOTHER POINT 7')

FORMAT(‘OPOINT TO EVALUATE IS ’)

FORMAT(IOX,’ S’.18X.’ R?.18X.‘ T'.18X,"A'./
12X.F.5X.F,5X,F.5X,F)

FORMAT( 'ORANGE OF SMOOTH IS 5 'IO 25 POINTSI')
FORMAT(‘NU’,I5)

FORMAT(‘RD'.2IE15.7))

FORMAT(‘BPLOT : ')

FORMAT(‘SNORMALIZATION FACTOR ['.I5,’,’.I5’I 2 ')
FORMAT(‘GPOINT NUMBER IS '.I5./)

FORMATI'SGRAPHICAL INPUT 7')

FORMAT(‘OLIST [LP] : ')

FORMAT(‘SCHANNEL WIDTH I: [EVEN] 3')
FORMAT(‘SREFERENCE CHANNEL IS ')

FORMAT(‘SORDER.OF DERIVATIVE IS ')

END

REPACK:

LOOP:

EXIT:

SCALE:

LOOP]:

202

SUBROUTINE REPACK.(IDAT,IBUFF.ICNT.ISCALE.NAVG)

I
I
I IDAT 3 DATA ARRAY ADDRESS
I IBUFF 8 BUFFER ARRAY ADDRESS
I ICNT = BYTE COUNT
I ISCALE = SCALING FACTOR
NAVG 8 NUMBER OF SIGNAL AVERAGES

MACRO SUBROUTINE TO REPACK AND SCALE VIDICON DATA

I

I

I

I 1977

JIM HORNSHUH NOVEMBER 20,

1
.TITLE REPACK

.IDENT /

V1.2/

.PSECT REPACK
.GLOBL REPACK

I
. SB’I'I'L DEFINITIONS

R0 8 20

R1 = XI

R2 = 32

R3 8 X3

R4 = 84

R5 3 25

SP 8 86

PC = 87

I

.MCALL .F4DEF

.F4DEF 0

I

.PAGE

.SBTTL MAIN

I

MOV RO.-(SP) ISAVE REGISTERS

MOV R1.-(SP)

MOV R2.-(SP)

MOV R3.-(SP)

MOV R4.-(SP)

MOV 2(R5),R0 IDATA ARRAY ADDRESS

MOV 4(R5).R1 IBUFFER.ARRAY ADDRESS

MOV 06(R5).R2 IBYTE COUNT

MOVB (R1)+.N1 INUMBER.OF AVERAGES TO CALLING PROGRAM
MOVB (RI)+,N2

MOV N1.012(R5)

SUB 3.R2 IADJUST BYTE COUNT

CMPB 3.(R1)+ I3 BYTES/DATA WORD?

BEQ SCALE IBRANCH IF YES

MOVB (R1)+.N1 IELSE REPACK 2-BYTE DATA WORDS
MOVB (R1)+,N2

MOV N1.R3 IGET REPACKED DATA

ADD 100000.R3 ICONVERT TO 2’S COMPLEMENT
MOV R3,(RO)+ IAND PUT IN DATA ARRAY
SUB 2.R2 IDECREMENT BYTE COUNT

BNE LOOP ILOOP 'TIL DONE

MOV (SP)+,R4 IRESTORE REGISTERS AND EXFT
MOV (SP)+,R3

MOV (SP)+.R2

MOV (SP)+,R1

MOV (SP)+.RO

HIS R5 8** EX]T.FROM SUBROUTINE *3
ADD 2.R1 IMSB OF IST DATA WORD

MOV 0.N5

MOVB (R1).N5 IGET IST BYTE

CMPB (R1).N5 ILOOK FOR LARGER BYTE

LO0P2:

0N2:

AGAIN:

N1:
N2:
N3:
N4:

N6:

BYTCNTR

BLOS
MOVB
ADD

BNE
MOV
MOV
BEQ
MOV

DEC
ASR

MOV
MOV

MOV
SUB
MOV
MOV
MOVB
MOVB
MOVB
MOV
MOV
ASHC
ADD
MOV
SUB
BNE
JMP

.EVEN
.BYTE
.BYTE
.BYTE
.BYTE
.BYTE
.BYTE
.WORD

ON

(R1).N5

3,R1
3.32
LOOPI
1.R2
N5.R0
0N2
0.N5
R2
N5
RO
LO0P2

R2,010(R5)
4(R5).R4

3.R4

06(R5),BYTCNT
3.BYTCNT
2(R5),R3

“a.”

(R4)+,N1
(R4)+,N2
(R4)+,N3

N3,R0
N1.R1
R2,RO

100000,R1
R1.(R3)+
3.BYTCNT

AGAIN
EXIT

8
.END REPACK

203

IBRANCH IF N5>(R1)

IBUMP ADDRESS
IAND BYTE COUNT
ILOOP 'TIL DONE
IINITIALIZE R2

IBRANCH IF RO=O
I 8:2

I / 2

ILOOP 'TIL R030

ISCALE FACTOR TO CALLING PROGRAM
IBUFFER ADDRESS

:BYTE COUNT
IDATA ARRAY ADDRESS

IHIGH DATA WORD IN R0

ILOW DATA NORD IN R1

ISHIFT BY R2

ICONVERT TO 2’8 COMPLEMENT
IAND PUT IN DATA ARRAY

LOOP 1 :

WRITE:

OUT:

«moon-000.000.00.00

204

SUBROUTINE TEXT ( IBUFF. IBYTE. INDEX. REID)

IBUFF = BUFFER ARRAY ADDRESS
[BYTE = BYTE COUNT

INDEX 3 ARRAY INDEX

NEND = ETX FLAG

.TITLE TEXT
. [DENT /V1.0/
.PSECT TEXT
.GLOBL TEXT

I
. SB'I'I'L DEF INITIONS

SUBROUTINE TO OUTPUT HEADER TEXT TO TERMINAL
JIM HORNSHUH

NOVEMBER 20, 1977

8
.MCALL .F4DEF. . INIT, .WRITE. .WAIT. .RISE

R0 = 20

R1 = 81

R2 = 82

R3 = X3

R4 = 54

R5 = 35

SP = Z6

PC = 37

.F4DEF 0

8

.PAGE

.SB’I'I‘L MAIN

I

MOV R0.-(SP)
MOV R1,-(SP)
MOV R2.-(SP)
MOV R3.-(SP)
MOV R4.-(SP)
MOV 2(R5) ,R4
MOV 04(R5) ,R3
MOV BUFF.R2
MOV TBUFF.R1
MOV R1.6(R2)
CLR R0

DEC R3

CMPB 3.(R4)
BEQ WRITE
MOVB (R4) , ( R1)+
INC R0

CMPB 15.(R4)
BNE LOOP!
MOVB 12.(R1)+
INC R0

CMP R3. 0
BEQ OUT

ADD 1.R4

BR LOOP

MOV 1.010(R5)
CMP R0. 0
BEQ OVER

MOV R0,4(R2)
MOV 'I'I'Y.R0

. INIT R0

.WRI'I'E Room

ISAVE REGISTERS

IGET ADDRESS OF BYTE ARRAY
IAND BYTE COUNT

IR2->LINE BUFFER HEADER
:R1->TEXT BUFFER

:SET DUMP POINTER IN HEADER
IINITIALIZE COUNT

:CHECK FOR ETX
IBRANCH ON ETX

IBYTE TO TEXT BUFFER
IINCREIIENT BYTE COUNT
ICR?

INO

:BRANCH ON ZERO BYTE COUNT
IELSE CHECK NEXT BYTE

ISET RETURN ETX FLAG
IHEADER TEXT?

IBRANCH IF NOT

ISET ACTUAL BYTE COUNT
IRO->TI'Y LINK BLOCK

OVER:

.RLSE R0

MOV R3.04(R5)
COMB R3

INC R3

INC R3

MOV R3,e6(R5)
MOV (SP)+,RA
MOV (SP)+,R3
MOV (SP)+,R2
MOV (SP)+.RI
MOV (SP)+,R0
HHS R5

8

.WORD 0

.RAD50 IKB/
.BYTE 1.0

.RAD50 /KB/

3

.WORD 0

.BYTE 6,0

.WORD 0

.WORD 0

3

.BLKB 300

8
.END TEXT

205

zBYTE COUNT TO CALLING PROGRAM

:ARRAY INDEX TO CALLING PROGRAM
:RESTORE REGISTERS FOR.EXJT

: 3* EXIT FROM SUBROUTINE **

:KEYBOARD LINKBLOCK
:LOGICAL DEVICE NAME
3WORDS TO FOLLOWVUNIT NO.
:PHYSICAL DEVICE NAME

zMAX BYTE COUNT

:STATUS/MODE =DUMP UNFORMATTED.ASCII
8ACTUAL BYTE COUNT

:DUMP POINTER

:TEXT BUFFER

OOODOOOOOOOOO

206

SUBROUTINE RFILE ( IBYTE. IERR, INDEX. IBUFF)

IBYTE 3 BYTE COUNT

IERR 3 ERROR FLAG

INDEX 8 BYTE POINTER FOR BUFFER ARRAY
IBUFF = ADDRESS FOR BUFFER ARRAY

SUBROUTINE TO REFILL IBUFF WITH VIDICON DATA
JIM HORNSHUH NOVMBER 20, 1977
3* COMPILE WITH ION SWITCH *8

SUBROUTINE RFILE ( IBYTE. IERR. INDEX. IBUFF)
IERR=0

IBYTE=256

CALL UNFMRW ( IERR. IBUFF. IBYTE)

INDEX‘-' .1

RETURN

END

000000

with

207

SMOOTH.FTN
JIM.HORNSHUH NOVEMBER 1, 1977
DIGITRL SMDOTHING SUBROUTINE

SUBROUTINE SMOOTH(IRAY,NRAY.ISIZE’

DIMENSION NOR!“ II) . IWEICIH 13. II) . IRAY(BI9) . IBUF( 26)

DATA NORM/35.21.231.429.I43,I105,323,226I,3059.805,5175/
DATA IWEIGH/-3,12,17,IO*0,-2,3.6.7,9*0.-21,14.39.54.59,
18*0'-36'9944969’MQ8997*0.-11.099. 1692192402596*09_78’
2-13.42.87.122,147.162,167,5*0.‘21,-6.7,18.27.34.39,42,
343.4*0,'I36.'51,24.89,144.IB9,224,249.264.269,3*0,“I7I.
4-76.9.84,149.204,249,284.309.324.329,2*0,-42.-21,‘2,15,
530.43.54,63.70.75.78.79,9.-253.'I38.’33,62.147,222,287,
6343.387,422,447,462.467/

DO 3 I31.ISIZE
IBUF( I)= IRAY( I)
IB=(ISIZE-3)/2
IC=(ISIZE+I)/2
XNORM=NORM(IB)
IA=IC-I
LSTPNT=NRAYBIA

DO I I=IC.LSTPNT
IBUF(ISIZE)=IRAY(I+IA)
PNT=FLOAT(IWEIGH(IC,IB))*FLOAT(IBUF(IC))
DO 2 J=I.IA
PI=IBUF(IC+J)
P2=IBUF(IC-J)
WGT=IWEIGH(IC-J.IB)
PNT=PNT+WGT*(PI+P2)
PNT=PNT/XNORM

DO 4 J=I.ISIZE
IBUF(J)=IBUF(J+1)
IRAY(I)=PNT+.5
RETURN

END

o.cocoa-«o.no.u-~uu~uuumu«uauuaaaauua«

208

14010 WRITTEN BY TOM ATKINSON AND JIM HORNSHUH FOR.THE
C.G.ENKE RESEARCH GROUP

18-JUN-73

SUBROUTINE FOR.GRAPHIC I/O WITH TEKTRONICS 4010
- ERASE SCREEN.ALPHA MODE. HOME(0,780)

- GRAPHIC MODE. CONSTRUCT DARK VECTOR

- CONSTRUCT LIGHT VECTOR

RETURN TO ALPHA MODE

GRAPHIC INPUT WITH CROSS-HAIRS

PLOT AXIS AND TICK MARKS

GRAPHIC MODE. POINT PLOT

GRAPHIC INPUT’WITHOUT SCALE SHIFT

NGGfiG’NI-O

FOR MODE 6 OPERATION CALL T4010(IMODE.IXDIV.IYDIV)
IMODE 8 MODE OF OPERATION
IXDIV = NUMBER OF X DIVISIONS DESIRED
IYDIV 3 NUMBER OF Y DIVISIONS DESIRED

FOR ALL OTHER MODES CALL T0010CIMODE,IXCORD.IYCORD)
IXCORD 8 INTEGER VALUE OF X COORDINATE (0-1000)
IYCORD 8 INTEGER VALUE OF Y COORDINATE (0-600)

THIS SUBROUTINE SHIFTS THE SCREEN ORIGIN TO (10.10)
FOR ALL PLOTTING AND GRAPHIC INPUT EXCEPT FOR.MODE 7.

.TITLE T4010

. SB'I'I'L PRELIMINARIES
.IDENT /T4010/

.PSECT T4010.REL,RW,I
.GLOBL T4010

.PAGE

.NLIST 'I'I'M

R0=%0 : REGISTER DEFINITTON
R1331
R2=Z2
R3333
R4=X4
R5=35
SP=Xé
PC=X7

MASKR5=177740 3 MASKRU 8 .NOT3000037
MASKSB=I76037 MMSKSR 3 .NOT.001740

TXCSR=177564 T4010 TRANSMITTER.COMMAND. STATUS REG
TXBUF=177566 T4010 TRANSMITTER DATA REGISTER
RVCSR=177560_ T4010: RECEIVER COMMAND. STATUS REG
RVBUF=177562 T4010: RECEIVER DATA REGISTER

CSK=33 T0010: ERASE. GO TO ALPHA MODE, AND

CL=14 T4010: RETURN TO HOME (0.780)

CSO=37 T4010: GO TO ALPHA MODE

CSM=35 T4010: GO TO GRAPHICS MODE

CZ=32 T4010: STARTS CROSS-HAIR WHEN
PRECEEDED BY CSK

.PAGE

.SBTTL MACROS
.MCALL .F4DEF

DUMB] :

T4010:

. F4DEF 0

.MACRO INPUT.DUMMI .DUMMY
TSTB RVCSR

BPL DUMMI

MOVB RVBUF . DUMMY
. ENDM

. MACRO COORD.HI .LO. VALUE
CLR R0

MOVB HI . R0

BIC MASKR5 . R0
ROL R0

ROL R0

ROL R0

ROL R0

ROL R0

BIC MASKS5 . R0
CLR RI

MOVB LO. R1

B IO MASKR5 , RI
ADD R1 . R0

MOV R0 . VALUE

. ENDM

.NLIST ME

. PAGE

. SBT'I'L MAIN

MOV R0 . " ( SP)
MOV R1 . -( SP)
MOV R2 . -( SP)
MOV R4 . -( SP)
MOV 2( R5) , R4
MOV ( R4) , R0
TST R0

BGE L0

JMP ERR

CMP 7 . R0

BEG. ADD2

BEG. ADD2

BEQ ADDl

BGE L1

JMP ERR

ADD 2 . R0

BR L1

INC R0

ROL R0

ADD AJT. R0

JMP (R0)

BR ERASE

BR GRAPHI

BR VECTR2

BR ALPHA

BR GIN

JMP AXIS

JMP PTPLT

JMP GIN

MOV CSK. R0
JSR R5 . OUTPUT
MOV CL. R0

JSR R5 . OUTPUT
JMP EXIT

209

. ASSIGN as CONVENTION 'm micnos

:TEST‘ B117. IS DATA READY
:BRANCH IF NOT READY
:DATA TO CPU

:HI BYTE->R0
zMASK OFF RIGHT 5 BITS
:ROTATE R0 5 TIMES TO LEFT

:MASK OFF BITS5-9

:LO BYTE->R1
IMASK OFF BITS 0-0
:ADD HI/ID BYTES

8 SAVE REGISTERS

IMODE ADDRESS -) R4

IMODE VALUE -) R0

TEST IF >0

YES, CONTINUE

NO.RETURN

IMODE .GT. 7 IS ILLEGAL

ADD 2 TO GET OVER JUMP IN TABLE
BYPASS JUMP

00.000.000.000“

: MULTIPLY BY TWO
: CONSTRUCT JMP ADDR

GRAPH! 8

VECTR2:

ALPHA:

GIN:

GINNS:

RECSR:
AXIS :

MOV
MOV
ADD
MOV
MOV
ADD
J SR
J MP
MOV
MOV
ADD
MOV
MOV
ADD
J SR
J MP
MOV
J SR
J MP
B I C
MOV
J SR
MOV
J SR
I NPUT
INPUT
I NPUT
I NPUT
I N PUT
COORD
COORD
MOV
MOV
CMP
BEQ
SUB
SUB
MOV
MOV
MOV
MOV
B I S
J MP
MOV
MOV
J SR
MOV
J SR
MOV
J SR
CLR
MOV

HOV

BNE
CLR
BR

D I V
MOV
CLR
MOV
MOV
MOV
TST

4(R5).R4
(R4) .XNUM
12.XNUM
6(R5) .R4
(R4) .YNUM
12.YNUM
R5.GRAPH
EXIT
4(R5) ,R4
(R4) .XNUM
12.XNUM
6(R5) .R4
(R4) .YNUM
12.YNUM
R5.VECT'R
EXIT
CSO,R0
R5.0UTPUT
EXIT

100 . RVCSR

CSK.R0
R5,0UTPUT

CZ,R0
R5.0UTPUT
AGAIN.R0
AGN1.HIX
AGN2.LOX
AGN3,HIY
AGN4.LOY

HIX. LOX. XCORD
HIY. LOY. YCORD

2(R5) .R0
(R0) .R1
7,Rl
GINNS
12.XCORD
12.YCORD
4(R5) ,R0

XCORD. (R0)

6(R5) ,R0

YCORD. ( R0)
100 , RVCSR

EXIT

1762.XNUM

12.YNUM
R5.GRAPH
12.XNUM
R5.VECTR

1142.YNUM

R5.VECT'R
no .
1762,R1
4(R5).R2
(R2) .R3
R3
YES
XDIV
AXI
R3,R0
R0.XDIV
R0
1142,R1
6(R5) .R2
(R2) ,R3
R3

210

:SHIFT ORIGIN TO (10. 10)

8 SET TO ALPHA MODE

:NON-SCALED GRAPHIC INPUT
zSHIFT DATA BACK TO ORIGIN (0.0)

:RFSET INT ENBL BIT

tARE X DIVISIONS REQUIRED
:YES. CONTINUE
:NO. SET XDIV=0

:ARE Y DIVISIONS REQUIRED

CONT:

PTPLT:

ERR:
EXsz

8
8 SUBROUTINE GRAPH:

8
GRAPH:

VECT'R:

BNE AX2

CLR YDIV

TST XDIV

BNE XTICK

JMP EXIT

DIV R3.R0

MOV R0.YDIV

MOV 12.XNUM

TST XDIV

BEQ YTICK

ADD XDIV.XNUM

MOV 12,YNUM
1762.XNUM

BMI LASTX

JSR R5.GRAPH

MOV 20.YNUM

JSR R5.VECTR

JMP XNEXT

MOV I762.XNUM

JSR. R5.GRAPH

MOV 20.YNUM

JSR. R5.VECTR

MOV 12.YNUM

MOV 12.XNUM

TST YDIV

BNE GO

JMP EXIT

ADD YDIV,YNUM

CMP 1142.YNUM

BGE CONT

MOV 1142.YNUM

JSR. R5.GRAPH

MOV 20.XNUM

JSR R5,VECTR

MOV 12.XNUM

JMP EXIT

JSR R5,GRAPH

MOV 20.XNUM

JSR R5.VECTR

MOV 12.XNUM

JMP GO

MOV 4(R5).R§

MOV (R4).XNUM

ADD I2.XNUM

MOV 6(R5).R§

MOV (R4),YNUM

ADD 12.YNUM

JSR. R5.GRAPH

JSR R5.VECTR

JMP EXIT

MOV (SP)+.R§

MOV (SP)+.R3

MOV (SP)+.R2

MOV (SP)+,RI

MOV (SP)+,R0

RTS R5

.PAGE

.SBTTL SUBROUTINES

MOV
JSR
MOV

CSM.R0
R5,0UTPUT
YNUM.R0

211

8YES. CONTINUE

:ARE X DIVISIONS REQUIRED
8YES DRAW TICKS
3N0. EXIT

8DRA‘W FINAL TICK MARK AT 1000

:DRAW FINAL Y TICK MARK AT 600

8 ADD OFFSET BIAS

8 RESTORE REGISTERS

8 AND RETURN

DRAW VECTOR.ON SCREEN

8 DO UPPER Y

8
8 SUBROUTINE ROTH:

8
ROT5:

JSR
BIC
BIS
JSR
MOV
BIC
BIS
JSR
MOV
JSR
BIC
BIS
JSR
MOV
BIC
BIS
JSR
RTS

ROB
80R
BOB
ROR
BOB
HTS
:
3 SUBROUTINE
8
OUTPUT: TSTB
BPL
MOV
CLR
BIS
MOVE
OUT]: TSTB
BPL
MOV
HTS
ETX: .BYTE
Lox: .BYTE
HIY: .BYTE
LOY: .BYTE
.EVEN
XEORD: .WORD
YCORD: .WORD
XDIV: .WORD
YDIV: .WORD
XNUM: .WOBD
YNUM: .RORD
AJT: .WORD

.END

OUTPUT:

R5.ROT5
MASKR5.R0
40.R0

R5.0UTPUT

YNUM.R0
MASKR5.R0
140.R0

R5,0UTPUT

XNUM.R0

R5.ROT5
MASKR5.R0
40.R0

R5.0UTPUT

XNUM,R0
.MASKR5,R0
100,R0

R5,0UTPUT

R5

R0
R0
R0
R0
R0

RS : RETURN
OUTPUT l BYTE T0 TERMINAL

TXESR
OUTPUT

TXESR4-(SP)

TXESR
200.R0
R0.TXBUF
TXCSR

OUT!

(SP)+,TXCSR

JTABLE
T4010

«.0000. 00.6.0000...“

212

ROTATE 5 RIGHT .
MASK.OFF RIGHT 5 BITS
INSERT UPPER Y TAG
SEND UPPER Y'BYTE

DO LOWER Y

MASK OFF RIGHT 5 BITS
INSERT LOWER.Y TAG

DO UPPER X

ROTATE 5 RIGHT

MASK OFF RIGHT 5 BITS
INSERT UPPER X‘TAG

DO LOWER X
MASK OFF RIGHT 5 BITS
INSERT LOWER.X TAG

ROTATE R0 RIGHT 5 TIMES

IS TERMINAL BUSY?

STATUS REGISTER POS MEANS BUSY
SAVE CURRENT TERM TRANS STATUS
BE SURE INTERRUPT IS OFF

SET BIT 8 OF BYTE

SEND IT OUT

TERMINAL BUSY?

no 111310113 STATUS
RETURN

8ADDR.OF JTRBLE

APPENDIX B

Supplementary Documentation of System Design

 

213

umuuszm mo mumum mmwcmzu mum: mmHsa < QonIQHHm Hmuusfim mocoNH

 

ammo I H .wmmOHu I o mew umuusnm HQ
mzumum pmuusam «ooomH
:mum mo wcm um unsuumuaH mmHnmcm I H mHnmcm uaaupmucw memum Hm
neumchuoou cu mem wcHuumm umumm
wumum vHOH cucH ow Cu ouoHE menmcm I H mHnmcm vHon on
umumwwmh unapumucH wcm cHom moooNH
m>Huum I H Heumckuooo cu mem NooomH
uHEHH suwamHm>m3 puma: um I H uHEHH vum3u0w ma
uHEHH suwcmHm>m3 umsoH um I H uHEHH mmum>mu No
:OHuHmoa cw tam wmuomem HHmu I o HHmo mucmumwmu HQ
coHuHmom cw can amuumHmm HHmu I o HHmo mHaEmm on
maumum neumEounuoaoz HooomH
mHaEmm I H .mucmuwmmu I o cowufimoa HHmu ma
vumzuow I H .mmum>mu I o uomem COHuomqu No
HOHucou ouoHE I H .Houuaoo Hmscme I o mHnmcm Hmcuwuxm HQ
mmHsa Mme 4 H.o wamum HoumEouzooco: A v use mmHsa on
uwumemu msumum neuMEousoocoz ooooNH
mummEEou coHuocsm ufim mwmuuv<

 

AHMuoov chow: :OUHvH> pow mmmmmuwv< HmumnaHumm .HIm mHan

 

214

HomHmm HHH w «a

 

 

w H o H H o uumHmm an oH mo
oH o H H o H Hoonm an.mw Na
NH H H o o o HumHmm m. HQ
muHHH m oH mm. m. m. Homem m. on
umumwwmu wsumum o\< mHooNH
mumucsou mmwuvvm vcm wuha mmNHHmHuHGH mNHHmwuficH «HoomH
:oHumuwwucH m>Huomem mmHHmcw I H mHnmcm m>HuumHmm ma
COHuHmHDUuw mumw mmHnmcm I H cmom uumum «a
m>Huum I H HoumHmmu chw «m ma
m>Huom I H HoumHmmu afimw mm mm
m>Huom I H Houmfimmu :Hmw NM HQ
m>wuom I H HeumHmmu :Hmw Hm on
ummwm co vmummHo umumwwmu Houucoo :mom mHoomH
mwmuoum mumv How
mmmuvcm wcHuumum mo muhn swam NHOONH
wmeOum wumc new
mmmuvwm wcHuumum mo muhn 30H HHooNH
ego; mume\mmuHH m no .N .H :onHumpa H0 mman H0 umpasz oHoomH
mucmEEou cowuocsm uHm mmmuvv¢

 

AHmuuov OHUHZ GOUHvH> pom wwwmmuv©< cowuflmfisvu< mama .NIm manna

215

 

 

mmem uouum Q\< mummHu ummHo NHOONH
m>Huom I H uouuw msumum Ho
m>wuom I H 30Hmum>o o\< on
msuwum Q\< oHoomH
.H.w.ucouV .NIm «Home

216

 

 

 

mem mEmum mummHo ummHo oHoqu
COHumuwmucw m>HuomHmm you mHamcm :mum HmoHuum> mmHsa :mom HmoHuum> mooqu
coHumuwmucH m>HummHmm hem muHa auwcmHm>m3 muuxm ucmEmausw auwcmHm>m3 oooqu
omm cu H u 2 wow .z\am0q u mHmccmao mo umaEsz ucmEmuucH auwcmHm>m3 mooqu
omm mu H u 2 wow .z\umz H u hucmsvmum uumHmm zuamavmum mooomH
muHsuuHu cmum momHvH> mummmm ummmm mooomH
. m uumHmm vaE Emma ma
Ao.HV wcwxcmHa .Ao.ov wmo Emma .AH.HV :0 Emma < ummHmm mvoE Emma «a
vaamcm :mum I H maOHum memamm ma
vaamcm cmom I o uumHmm mequn ma
nmuHHanH UEHm mcHH I o “HaHnaH ucxm mcHH Ha
vaamsm cmum I H mHamcm cmum on
umumemp maumum amom Nooqu
vmumuwmuafi mHm>umucH HmUHuum> mo umaEmz mumuwmucH nalco
vmaame mHm>umucH HmoHuum> mo pmaEsz mmHmn moIoa Hooqu
0mm cm H u z uoH .z x > mm.o u mwmuHo> H .02 cHuu Hamuuau emmm oooqu
mucmEEoo coauucsm mam mmmuvv<
AHmuuov Opuwz cmum How mmmmmuww< Houucou cooHcH> .MIm mHamH

217

 

mmmpwoua CH Emom I o .wmumHanu cmum I
cmHamcm ammm I

H
H

wmumHanu cmmm
mmHamcm cmom
mmew mmumum cmom

he
on

HHoqu

 

.H.v.u=ouv

.mIm mHamH

218

Table B-4. Backplane Connections for Vidicon Slave Micro

 

 

Pin Side A (wirewrap) Side B (component)
1 D0 D4
2 D1 D5
3 D2 D6
4 D3 D7
5 A0 A2
6 A1 A3
7 Board select Ifififii
8 TREK W
9 INT Integrate enable
IO Discharge enable A/D convert
11 DBIN Line sync

12 Vidicon carry out

 

219

Table B—5. Backplane Connections for Fast Adder Board

 

 

 

 

Pin Side A (wirewrap) Side B (component)
1 A select A0
2 B select A1
3 12 select A2
4 10 select A3
5 Line sync A4
6 A/D status flag A5
7 R1 A6
8 R2 A7
9 R3 A8
10 R4 A9
11 Integrate A10
12 Convert pulse A11
13 Adder reset A12
14 A/D overflow flag Memory board select
15 Clear error flag 8218 EATER
16 DIEN Average SEEEET
17 '03 2102 R/W (ME—MW)
18 8 select Vidicon carry out
19 MEMR A/D status error
20 Scan flag
33 D7 D6
34 D5 D4
35 D3 D2

36 D1 D0

 

220

Table B-6. Backplane Connections for Memory Board

 

 

Pin Side A (wirewrap) Side B (component)
1 A/D input - Bit 12 A0
2 11 A1
3 10 A2
4 9 A3
5 8 A4
6 7 AS
7 6 A6
8 5 A7
9 4 A8
10 3 A9
11 2 A10
12 1 A11
13 Adder reset A12
14 A/D overflow Memory board select
15 Clear error flag 8212 EATER
l6 'DTEN Average SEEECT
17 EE 2102 R/W (W)

18 MEMR A/D status error

 

221

Table B—7. 3M Connector Pin Assignments

 

 

Pin Signal Pin Signal
1 A0 18 D1

2 A1 19 D2

3 A2 20 D3

4 A3 21 D4

5 A4 22 D5

6 A5 23 D6

7 A6 24 D7

8 A7 25 DBIN
9 A8 26 m
10 A9 27 m
11 A10 28 ¢2TTL
12 A11 29 Ready
13 A12 30 m
14 A13 31 FAR
15 A14 32 Reset
16 A15 33 INTE

17 DO

 

APPENDIX C

Front Panel Documentation

APPENDIX C

Front Panel Documentation

The hardware front panel provides the operator with complete
hardware control of the micro and its peripheral circuits. This hard-
ware control capability is important not only in testing the micro and
its peripheral circuits during the construction phase, but also for
later trouble-shooting of non-functional circuits. Figure C-l illus-
trates the signal interaction between the front panel and the micro—

processor. The front panel generates all of the control signals (DBIN

 

 

GATE, MEMW, MEMR, READY, STSTB, BUSEN) needed to assume full control of
the micro. The logic that generates these signals will be discussed
later in this appendix. Data from the micro is buffered by bus driver
B1 and displayed with an 8-digit 7-segment LED display (D3). Status
information on the data lines is latched by latch Ll when STSTB is
pulsed low and displayed with individual LED's (D2). Keyboard input
is driven down the micro data bus by bus driver B2. A single 16—digit
7-segment display (D1) is used to alternately display the keyboard
output and the current memory address (see Figure C-2).

Data and address information are input from the front panel via a
surplus keyboard of unknown origin. Since this keyboard does not out-
put a standard keyboard code, it was necessary to design the front
panel around the keyboard output code. Table C-l illustrates the
keyboard output code for 16 keys and the value or function assigned
to each key. Bit M4 is used to distinguish a number from a control

function.

222

223

usauso\u=maH Hmcmm acoam mo Emuwmwa Hoon

.HIu mpstE

 

 

 

HI

 

 

 

 

 

 

  

 

 

 

 

 

 

 

 

 

 

     

 

  
       

Na.
><1Ema

m:._.<._.m
ow;

  

 

 

 

_o
2.363

mmumoo<
\a¢<om>ux
ou._

 

 

 

Au 5.33 a?

 

b >o<w¢

 

 

b 12w:

 

 

 

0 32m!

 

 

zumam A» “
mpmhm AI» .
mampm nvI H
_ £3.33 no
_ <p<n on;
_ o\
n \
mam <h<a o o
._<zo:.uu¢_o-_m I \\ u \N
- mupmaum
.r2:m
V . . o mm a 9:639.
zzao . m ~_~o
8.x. .
mam «3.33 nv \ .
n 5552.
ht.:m
- om<om>ux
.
>o<um A “
«gm: 4. _
32m: A .
~26 2.2. T .
.

om<om mommmoommomog

._mz<n_ #205 $4253:

3 230 2.0 o

224

 

 

 

Table C-l. Keyboard Output Code
Key M1 M2 M3 M4 N1 N2
None 1 1 1 1 1 1
0 1 O 1 1 1 O
1 1 1 O 1 1 1
2 1 O 1 1 1 1
3 1 O 0 1 1 1
4 O 1 1 1 1 l
5 O 1 0 1 1 1
6 0 O 1 1 1 1
7 O O 0 1 1 1
HLT O O O O 1 1
SS 0 O 1 0 1 O
EXM 0 1 O O 1 0
DEP O 1 l O 1 0
EXM NXT 1 0 0 O 1 0
DEP NXT l 1 0 0 1 0
CLR 1 1 1 O 1 O
RUN 1 1 1 0 0 1
* M ° 1 - number, 0 - control function

u.

225

KEYBOARD IADDRESS INPUT

  

, ...........
l/2-745O ' - ADDRESS BIT

    

 

KEYBOARD
7-SEGMENT . . ' ENTRY
LED 7 " <- M4

 

 

Figure C-2. Keyboard/Address Display Logic

The keyboard interface is illustrated in Figure C-3. When an
octal address or data word is entered at the keyboard, the correspon—
ding bit pattern for each entered digit is clocked into a 6-digit
shift register (3 x 74164). The shift register outputs are then
multiplexed onto the data bus at the appropriate time either to exam a
specified memory location or to deposit a data word in memory. To
deposit a word in memory, LOW BYTE ENABLE is set to a logic '1' to
enable the lower 8 bits from the shift register, bus driver B2 is
enabled (DST set to logic '0' and D82 to logic '1'), and BERN is
pulsed low to write the data into memory. To examine a desired memory
location, the front panel places a jump instruction on the data bus by
setting JMP ENABLE to a logic '1' and enabling bus driver B2. Next,
the micro's ready line is returned to a logic '1' to start the micro.
The micro then reads the jump instruction, which causes it to look for
a 16-bit memory address. Normally the micro would read the next two
memory locations to obtain the jump address; however, the front panel

disables the micro's memory input and places the user-specified jump

226

080.!
O...

.2051 0.5020 020 :0...
1.0-00 3 50 -0133: up». 300

U401! u
Garth»

£35

mummumucH numoammx .mIo mumwfim

.¢4u: («ado «at. o

2.8! tun: hxuz tubsuo an»; 5
83¢ 2»: O

.hxl Ixu. hxu! I<xu .0»: a

 

H5... 54: 3:. v
.2: an; 040...» 2:. 0
:38. :38 a»: u
:00. 20030 a»: .

227

address on the data bus by pulsing LOW BYTE ENABLE and HIGH BYTE
ENABLE at the appropriate times. After the micro has input the

address and completed a jump to the desired address, the front panel
returns the ready line to a logic '0' to halt the micro. At this point
the front panel displays the address and data of the specified memory
address.

The exam next (EXM NXT) function requires that the micro access
the next sequential memory location. To exam the next location, the
front panel disables the memory, places a NOP instruction on the data
bus, and raises the ready line to a logic '1' for one instruction
cycle. The micro then reads and executes the NOP, increments its
memory address, and halts at the next memory location. The deposit
next (DEP NXT) function combines the exam next and deposit functions.
For single stepping (SS) through programs, the front panel again lets
the micro run for only one instruction cycle at a time; however, in
this mode the micro reads and executes the code in its memory instead
of code supplied by the front panel. The clear (CLR) function writes
a zero byte (NOP) into the current memory location.

In order to latch the status bits output by the micro, the front
panel was interfaced directly to the micro bus. Consequently, the
system controller (8228) is between the front panel and memory. The
system controller is a bi-directional bus driver controlled by the
status output from the micro. When the micro halts, the controller is
in the input mode, i.e. driving data to the micro. Therefore, in order
to deposit data into memory from the front panel the direction of the

system controller must be reversed. The front panel does this by

228

placing a NOP instruction on the data bus and pulsing STETBI This
operation strobes a new status word into the system controller and
forces it to reverse direction. After depositing a byte of data into
memory, the direction of the system controller must again be reversed
to read the contents of the current memory location. The system
controller is returned to the input mode by enabling READ STATUS
ENABLE, enabling bus driver B2, and pulsing STSTEI This sequence
accompanies each deposit and deposit next instruction.

The function keys are decoded by the keyboard interface. When a
function key is depressed, the 74155 generates a 100 us pulse that
initiates the selected function. Figure C-4 illustrates the control
and timing circuit used to generate the necessary control signals for
front panel control of the micro and its peripherals as described

above.

229

 

w4m<zu
uh>m 304

mam<zm
wh>m 19.:

wam<zu
$21

wam<zm
m3k<hm
o<um

>o<wx

mhwhm

 

 

m
_N_vk

nN_¢h

ofiwoa

Houusoo Hmamm umoum

.qlo

mustm

 

REFERENCES

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

REFERENCES

R.G. Neuhauser, RCA Eng., 22, 40 (1976).

M.H. Crowell and E.F. Labuda, Bell System Tech. J., 42, 1481
(1969).

 

R.E. Johnson, Application of RCA Silicon Diode Array Target
Vidicons, Camera Tube Application Note AN—4623, RCA Electronic
Components, Harrison, N.J., 1971.

 

G.A. Robinson, RCA Eng., 22, 47 (1976).

A.H. Mengel, Optical Spectra, 29, 36 (1976).

 

E.C. Douglas, IEEE Trans. Electron Devices, ED—22, 224 (1975).

 

P.K. Weimer, Advan. Electron. Electron Phys., 22, 387, (1960).

 

R.G. Lubszynski, S. Taylor, and J. Wardley, Brit. Instn. Radio
Engrs., 29, 323 (1960).

 

D.J. Gibbons, Wireless World, 11, 1425 (1971).

 

M.J. Milano, H.L. Pardue, T.E. Cook, R.E. Santini, D.W. Margerum,
and J.M.T. Raycheba, Anal. Chem., 44, 374 (1974).

 

D.G. Mitchell, K.W. Jackson, and K.M. Aldous, Anal. Chem., 42,
1215A (1973).

 

Y. Talmi, Anal. Chem., 41, 658A (1975).

 

Y. Talmi, Anal. Chem., 42, 699A (1975).

 

 

P.K. Weimer, Adv. Electron. Electron Phys., 22, 181 (1975).
P.K. Weimer, RCA Rev., 24, 385 (1975).

C.I. Coleman and A. Boksenberg, Contemp. Phys., 22, 209 (1976).

 

G. Horlick, Appl. Spectrosc.,_20, 113 (1976).

 

T.A. Nieman, Ph.D. Dissertation, Michigan State University, 1975.

M.H. Crowell, T.M. Buck, E.E. Labunda, J.V. Dalton, and E.J.
Walsh, Bell Sys. Tech. J., 46, 491 (1967).

 

T. Conklin and E.E. Stupp, 02:. 13:13., is, 510 (1976).

R. Walton, C. Smith, B. Harper, and W.M. Wreathall, IEEE Trans.
Electron Devices, ED-21, 462 (1974).

 

 

230

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

231

D.B. Hunter, Optical Spectra, (November, 1975).

 

I.T. Saldi and K. Schlesinger, Optical Spectra, 2, (February,
1970).

 

K. Schlesinger and R.A. Wagner, IEEE Trans. Electron Devices,
ED-14, 163 (1967).

 

P. Hague and J.E. Jacobs, J. Appl. Phys., 42, 2655 (1977).

 

C.R. Hill, J. Phys. E, 2, 153 (1976).

 

P.H. Brown, R.P. Randall, R.F. Sivyer, and J. Wordley in
Ultrasonics International 1975, Conference Proceedings,
IPC Science and Technology Press, 1975, pp 73-79.

 

K.W. Jackson, K.M. Aldous, and D.G. Mitchell, Spectrosc. Lett.,
2, 315 (1973).

 

K.W. Jackson, K.M. Aldous, and D.G. Mitchell, Appl. Spectrosc.,
22, 569 (1974).

 

K.M. Aldous, D.G. Mitchell, and K.W. Jackson, Anal. Chem., 42,
1034 (1975).

 

G. Horlick and E.G. Codding, Appl. Spectrosc., 22, 167 (1975).

 

G. Horlick and E.G. Codding, Anal. Chem., 42, 133 (1974).

 

J.D. Winefordner, J.J. Fitzgerald, and N. Omenetto, Appl.
§pectrosc., 22, 369 (1975).

 

H.L. Felkel, Jr., and H.L. Pardue, Anal. Chem., 42, 1112 (1977).

 

H. van der Piepen, C. Claase, and D.B. de Villiers, Spectrochim.

 

Acta, 31B, 389 (1976).

K.W. Busch, N.G. Howell, and G.H. Morrison, Anal. Chem., 42,
575 (1974).

 

K.W. Busch, N.G. Howell, and G.H. Morrison, Anal. Chem., 42,
1231 (1974).

 

K.W. Busch, N.G. Howell, and G.H. Morrison, Anal. Chem., 42,
2074 (1974).

 

N.G. Howell, J.D. Ganjei, and G.H. Morrison, Anal. Chem., 42
319 (1976).

 

J.D. Ganjei, N.G. Howell, J.R. Roth, and G.H. Morrison, Anal.
Chem., 42, 505 (1976).

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

53.

54.

55.

56.

57.

58.

59.

60.

61.

232

A. Danielsson, P. Lindblom, and E. Soderman, Chem. Scr., 2, 5
(1974).

 

A. Danielsson and P. Lindblom, Phys. Scr., 2, 227 (1974).

 

D.L. Wood, A.B. Dargis, and D.L. Nash, Appl. Spectrosc., 22,
310 (1975).

 

G. Horlick and E.G. Codding, Anal. Chem., 42, 6490 (1973).

 

E.G. Codding and G. Horlick, Appl. Spectrosc., 22, 366 (1973).

 

G. Horlick, E.G. Codding, and S.T. Leung, Appl. §pectrosc., 22,
48 (1975).

 

J.A.C. Broekaert, Bull. Soc. Chim. Belg,, 24, 1159 (1975).

 

F.L. Fricke, 0. Rose, and J.A. Caruso, Anal. Chem., 42, 2018
(1975).

 

D.0. Knapp, N. Omenetto, F.W. Plankey, and J.D. Winefordner,
Anal. Chim. Acta, 22, 455 (1974).

 

T.L. Chester, H. Haraguchi, D.0. Knapp, J.D. Messman, and J.D.
Winefordner, Appl. Spectrosc., 22, 410 (1976).

 

A. Danielsson and P. Lindblom, Appl. Spectrosc., 22, 151 (1976).

 

N.G. Howell and G.H. Morrison, Anal. Chem., 42, 106 (1977).

 

M. Franklin, C. Baber, and S.R. Koirtyohann, Spectrochim. Acta,
313, 589 (1976).

 

T.E. Edmonds and G. Horlick, Appl. Spectrosc., 22, 536 (1977).

 

T.E. Cook, M.J. Milano, and H.L. Pardue, Clin. Chem., 22, 1422
(1974).

 

K.W. Busch and G.H. Morrison, Anal. Chem., 42, 713A (1973).

 

J.E. Callis, I.M. Warner, G.D. Christian, and E.R. Davidson,
Ind. Res., (November, 1976).

J.R. Jadamec, W.A. Saner, and Y. Talmi, Anal. Chem., 42, 1316
(1977).

 

M.J. Milano and H.L. Pardue, Anal. Chem., 42, 25 (1975).

 

T.E. Cook, H.L. Pardue, and R.E. Santini, Anal. Chem., 42, 452
(1976).

 

T.A. Nieman and C.G. Enke, Anal. Chem., 42, 619 (1976).

 

62.

63.

64.

65.

66.

67.

68.

69.

70.

71.

72.

73.

74.

75.

76.

77.

78.

79.

80.

81.

82.

83.

233

D.A. Yates, and T. Kuwana, Anal. Chem., 42, 510 (1976).

 

T.A. Nieman, F.J. Holler, and C.G. Enke, Anal. Chem., 42, 899
(1976).

 

R.E. Dessy, W.G. Nunn, C.A. Titus, and R. Reynolds,
J. Chromatogr. Sci., 24, 195 (1976).

 

R.E. Dessy, W.D. Reynolds, W.G. Nunn, C.A. Titus, and G.F.
Moler, Clin. Chem., 22, 1472 (1976).

 

H.L Pardue, A.E. McDowell, D.M. Fast, and M.J. Milano, Clin.
Chem., 22, 1192 (1975).

A.E. McDowell, R.S. Harner, and H.L. Pardue, Clin. Chem., 22,
1862 (1976).

 

A.E. McDowell and H.L. Pardue, Anal. Chem., 42, 1815 (1976).

 

R.S. Neiswander and 0.8. Plews, Appl. Opt., 24, 2720 (1975).

 

E. Ostertag, Rev. Sci. Instrum., 42, 18 (1977).

 

P. Mengers, Opt. Spect., 60, (August, 1977).

 

G. Zanella, M. Vascon, G. Novo, and A. Cavestro, J. Phys. E,
.22, 955 (1977).

 

J.T. Boyd and C.L. Chen, Agpl. Opt., 22, 1389 (1976).

 

M.J. Milano and H.L. Pardue, Clin. Chem., 22, 211 (1975).

 

P.H. Lloyd and M.P. Esnouf, Anal. Biochem., 22, 25 (1974).

 

J.B. Meyling and W.H. Hesselink, J. Phys. E, 22, 438 (1977).

 

T.B. McCord and M.J. Frankston, Appl. Opt., 24, 1437 (1975).

 

G.M. Ridder and D.W. Margerum, Anal. Chem., 42, 2098 (1977).

 

J.M. Marrs, American Lab., 57 (November, 1972).

 

R.E. Santini, M.J. Milano, and H.L. Pardue, Anal. Chem., 42,
915A (1973).

 

T.E. Cook, R.E. Santini, and H.L. Pardue, Anal. Chem., 42, 871
(1977).

 

M.J. Milano, S. Lam, and E. Grushka, J. Chromatogr., 125, 315
(1976).

 

I.M. Warner, J.B. Callis, E.R. Davidson, M. Gouterman, and G.D.
Christian, Anal. Lett., 2, 665 (1975).

 

84.

85.

86.

87.

88.

89.

90.

91.

92.

93.

94.

95.

96.

97.

98.

99.

100.

101.

102.

103.

234

I.M. Warner, J.B. Callis, E.R. Davidson, and G.D. Christian,
Clin. Chem., 22, 1483 (1976).

 

J.P. Fillard, M. de Murcia, J. Gasiot, and S. Chor, J. Phys E,
2, 993 (1975).

E. Kohen, C. Kohen, and J.M. Salmon, Mikrochim. Acta, 1976 II,
195 (1976).

 

R.P. Cooney, G.D. Boutillier, and J.D. Winefordner, Anal. Chem.,
42, 1048 (1977).

 

T.G. Curtis and W.R. Seitz, J. Chromatogr., 134, 513 (1977).

 

R.P. Cooney, T. Vo—Dinh, and J.D. Winefordner, Anal. Chim. Acta,
22, 9 (1977).

 

R.P. Cooney, T. Vo-Dinh, G. Walden, and J.D. Winefordner, Anal.
Chem., 42, 939 (1977).

R.E. Dessy, W.G. Nunn, C.A. Titus, and W.R. Reynolds,
J. Chromatogr. Sci., 24, 195 (1976).

 

W.H. Woodruff and G.H. Atkinson, Anal. Chem., 42, 186 (1976).

 

R. Wilbrandt, P. Pagberg, K.B. Hansen, and C.V. Weisberg, Chem.
Phys. Lett.,_22, 76 (1975).

 

J.M. Pasachoff, D.F. Muzyka, and J. P. Schierer, Jr., Appl. Opt.,
22, 2884 (1976).

 

T.B. McCord and J.A. Westphal, Appl. Opt., 22, 522 (1972).

 

A.E. McDowell and H.L Pardue, Anal. Chem., 42, 1171 (1977).

 

T. Vo-Dinh, D.J. Johnson, and J.D. Winefordner, Spectrochim. Acta,
33A, 341 (1977).

E.M. Carlson, Ph.D. Dissertation, Michigan State University, 1978.
S. Hourdakis, Ph.D. Dissertation, Michigan State University, 1978.
M.B. Denton and S.B. Tilden, paper no. 225 presented at the
Pittsburgh Conference on Analytical Chemistry and Applied
Spectroscopy, 1978.

C.G. Enke and T.A. Nieman, Anal. Chem., 42, 705A (1976).

 

Program written by T.V. Atkinson, Department of Chemistry,
Michigan State University.

A. Savitzky and M.J.E. Golay, Anal. Chem., 22, 1627 (1964).

 

104.

105.

106.

107.

108.

109.

110.

111.

112.

113.

114.

115.

116.

117.

118.

119.

120.

121.

122.

123.

124.

125.

126.

235

Data sheet for C23231 and C23246 Vidicons, RCA Electronic
Components, Harrison, N.J.

R.M. Barnes and R.F. Jarrell in Analytical Emission Spectroscogy,
Vol. 1, Part 1, E.L. Grove, Ed., Dekker, N.Y., 1971, pp 209-321.

 

OMA Catalgg, Princeton Applied Research, Princeton, N.J.

 

L.A. Selke, IEEE Trans. Electron Devices, ED-16, 625 (1969).

 

H.V Malmstadt, C.G. Enke, and S.R. Crouch, Electronic
Measurements for Scientists, W.A. Benjamin, Menlo Park, CA.,

 

 

1973, pp 743-750.

G. Horlick, Anal. Chem., 42, 352 (1975).

 

G.M. Hieftje, Anal. Chem., 44, 69A (1972).

 

G.M. Hieftje, Anal. Chem., 44, 81A (1972).

 

B. Chance, Rev. Sci. Instrum, 22, 634 (1951).

 

S. Shibata, Anal. Chim. Acta, 42, 271 (1969).

 

T.J. Porro, Anal. Chem., 44, 93A (1972).

 

J.C. Cowles, J. Opt. Soc., 22, 690 (1965).

 

J.E. Stewart, éppl. Spectrosc., 22, 287 (1973).

 

Manufacturer's literature on DW-2 UV-VIS Spectrophotometer,
American Instrument Company, Silver Spring, MD.

S. Shibata, Anal. Chim. Acta, 22, 49 (1973).

 

s. Shibata, Anal. Chim. Acta, _§_2_, 305 (1972).

 

T. Honkawa, Anal. Lett., 2, 839 (1976).

 

T. Honkawa, Anal. Lett., 8, 901 (1975).

 

C. Wade, Rev. Sci. Instrumu,42, 987 (1975).

 

R. Szentrimay and T. Kuwana, Anal. Chem., 42, 1349 (1977).

 

K. Ogdan, A. Essig, K.W. Beach, and S.R. Caplan, Rev. Sci.

Instrum., 42, 142 (1977).

R. Rikmenspoel, Rev. Sci. Instrum., 22, 497 (1965).

 

J.E. Stewart, American Lab., 51 (March, 1973).

 

127.

128.

129.

130.

131.

132.

133.

134.

135.

136.

137.

138.

139.

140.

141.

142.

143.

144.

145.

146.

147.

148.

149.

236

B. Chance, N. Graham, J. Sorge, and V. Legallais, Rev. Sci.

Instrum., 33, 62 (1972).

R. Woodriff and D. Shrader, Appl. Spectrosc., 22, 181 (1973).

 

B. Chance, Science, 120, 767 (1954).

 

B. Chance and N. Graham, Rev. Sci. Instrum., 42, 941 (1971).

 

R.M. Wightman, R.L. Scott, C.N. Reilly, R.W. Murray, and J.N.
Burnett, Anal. Chem., 42, 1492 (1974).

 

M.S. Denton, T.P. DeAngelis, A.M. Yacynych, W.R. Heineman, and
T.W. Gilbert, Anal. Chem., 42, 20 (1976).

 

 

S. Shibata, Anal. Chim. Acta, 22, 487 (1975).

J.A. Peterson and D.M. Mock, Anal. Biochem., 22, 545 (1975).

 

R.L. Sellers, G.W. Lowy, and R.W. Kane, American Lab.,
(March, 1973).

 

J.E. Stewart, American Lab., 51 (March, 1973).

 

R. DeSa and Q.H. Gibson, Rev. Sci. Instrum., 22, 900 (1966).

 

E.E. Wells, Jr., Anal. Chem., 42, 2022 (1973).

 

0. Rose, Jr., D.W. Mincey, A.M. Yacynych, W.R. Heineman, and
J.A. Caruso, Analyst (London), 101, 753 (1976).

 

A.T. Giese and C.S. French, Appl. Spectrosc., 2, 78 (1955).

 

J.P. Pemsler, Rev. Sci. Instrum., 22, 274 (1957).

 

M. Welkowsky and R. Braunstein, Rev. Sci. Instrum., 42, 399
(1972).

 

G. Baldini, E. Grilli, and M. Guzzi, Appl. Opt., 24, 2687 (1975).

 

G.M. Hieftje and R.J. Sydor, Appl. Spectrosc., 22, 624 (1972).

 

A. Gilgore, P.J. Stoller, and A. Fowler, Rev. Sci. Instrum., 22,
1535 (1967).

 

M.C. Shaw, Anal. Biochem., 44, 291 (1971).

 

W.L. Bulter, Arch. Biochem. Biophys., 22, 413 (1961).

 

J.P. Walters and H.V. Malmstadt, Appl. Spectrosc., 22, 193 (1966)

 

M.P. Klein and E.A. Dratz, Rev. Sci. Instrum., 22, 397 (1968).

 

150.

151.

152.

153.

154.

155.

156.

157.

158.

159.

160.

161.

162.

163.

164.

165.

166.

167.

168.

169.

170.

171.

172.

237

E.G. Olson and C.D Alway, Anal. Chem., 22, 370 (1960).

 

F. Grum, D. Paine, and L. Zoeller, Appl. Opt., 22, 93 (1972).

 

J.F. Brandts and L.J. Koplan, Biochem., 22, 2011 (1973).

 

R. Bird, J. Phys. E, 22, 953 (1977).

D. Lewis, P. Merkel, and W.J. Hamill, J. Chem. Phys.,_22,
2750 (1970).

 

T.C. O'Haver and C.L. Green, American Lab., 15 (March, 1975).

 

S. Shibata, Angew. Chem. Int. Ed. Engl., 22, 673 (1976).

 

B.C. Searle and D.B. Freberg, Computer, 2, 22 (1975).

A.J. Weissberger, Computer Design, 151 (June, 1977).

 

D.B. Chung, Electronic Design, 22, 133 (1976).

 

D.R. McGlynn, Microprocessors: Technology, Architecture, and
Applications, John Wiley and Sons, Inc., 1976, pp 182-187.

 

 

C.J. Sipple, Digital Design, 68 (August, 1977).

 

D.H. Chung, Electronics, 102 (January, 1977).

 

J. Pathak, Computer Design, 95 (July, 1977).

 

J. Pathak, Electronics, 108 (March, 1977).

 

D. Phillips and A. Goodman, Electronics, 109 (July, 1977).

 

G. Korn, Computer Design, 177 (June, 1977).

 

F.J. Holler, T.V. Atkinson, T.C. Kelly, and C.G. Enke,
American Lab., 9 (September, 1976).

 

Catalog No. MPS—77, BUD Industries, Inc., Willoughby, OH.

Design by M. Rabb, Department of Chemistry, Michigan State
University.

Data sheet on ASCII Encoded Keyboard Model EW-lOO, Electronics
Warehouse Inc., Redondo Beach, CA.

Data sheets on MTX-163ZSL and MIX-2562-1, Matrox Electronic
Systems, Montreal, Que.

Data sheet on Ball 9-inch CRT Data Display, Ball Brothers
Research Corporation, St. Paul, MN.

173.

174.

175.

176.

177.

178.

179.

180.

181.

238

Data sheets on Model 1070 Intelligent Diskette Controller and
Model 270 Dual Diskette Drive, Persai, Marina Del Ray, CA.

Intel Corporation, Intel 8080 Microcomputer Systems User's
Manual, Intel Corporation, Santa Clata, CA., 1976, Chp. 5.

 

Instruction manual on MDB-1710 General Purpose Interface Module,
MDB Systems, Inc., Orange, CA.

Product Bulletin 595-1208, Instrumentation for Spectroscopy
EU-700 Series, Health-Schlumberger, Benton Harbor, MI, 1969.

 

 

G. Davis, Radio Electronics, 23 (July, 1969).

 

Data sheet on MK5009P, Mostek Corporation, Carrollton, TX.

Intel Corporation, Intel Data Catalog, Intel Corporation, Santa
Clara, CA., 1977, Chp. 10.

 

Data sheet on MM57109, National Semiconductor Corporation, Santa
Clara, CA.

R.L. Ratzlaff and D.F.S. Natusch, Anal. Chem., 42, 2170 (1977).