fxfn; "WY:

4‘.

A #1
iv

a}:

'i '.

w, v.“ «. . ‘
E‘Mfl“ NEW w h,” 1
11L“ ‘7“; "'5'“ 991%?»
mkgsi‘wizgw‘ ‘

‘11.
L
i

L! y
if}. "v V

:3 a . v} , ‘1 ‘1' w‘.

9 l ' ‘1‘ ‘1‘ _ “:2" :1 ‘33.“, t‘x‘x‘u 21' 31:3“ ‘1“ -. 1‘

~ . “. '~ - :1.‘ “WWW ‘x‘x‘ ‘r: ‘1‘ "«

W akfi‘fiwfiggg \fisfiw‘fi‘g a.
V l.‘ ‘2‘ _ 5‘“ ‘7‘: ‘ a -.

try . ’1 HT?! ‘3‘." :fflbfi‘kkck 392%}:12‘1 - .~. Jaw‘xxx : ‘

.. x , .cv xv
i‘ _. f“. '..‘!‘Hc‘1

. - “‘1 . fix'i‘é‘r’ "111* L‘ * ‘L‘;

\V‘» ‘i‘xn‘f‘ ‘ .

1‘3. ' v u E:

‘

.‘I ~

..

,4 a .
' ‘ . I
1‘ “ g‘m W-

.

\J i

. {Emu
“12%

. win him ‘ga‘ a
' w‘g‘fl‘, Efi‘ék as}
' , . .

$5
. d
- 1
. 0- 1
.x‘u

\\

mm 1

N‘r
=. ‘9‘-

q:

1‘ -“1-:: w 1

“‘3“ ‘~ ~ "FM“ u xx 1

‘ ' ' ‘ -.1='*‘1‘13J‘IEC‘9’ ‘ '

fiééi ‘ ’ER “5133"“. ‘hvgv: ~ , ~ ‘ , V .

‘2. 3 " ‘fi 1:“? Et‘m . . ‘iwz’u‘ukwn". V}. "‘3“ “r“ ‘ EVE R
H. 4"“: .. ‘ ; . mm s ‘- Qfiiufing‘t

." WEE-3‘19-

" .

ii'zwm‘é

L“ t t‘. ,

iW‘Wé‘EQ‘HVéhm“ ‘1: :fi‘vhb“

3 i” Yi‘tHEH‘i'j 130‘ 3‘33”?!le
MW“ HM: 4‘ W'We-AFW‘
Li MW ‘8 ”*Wn'n‘i‘v

~in 3: Mm. 1 '3 ‘14:,“ v
‘ -'. 01%;"- ‘M gig!” ‘

am I. }‘
1

“N”
r"!

 

This is to certify that the

thesis entitled

THE DESIGN AND APPLICATION OF A
VIDICON RAPID SCANNING SPECTROPHOTOMETER
T0 STOPPED—FLOW KINETICS

presented by

Erik Mattis Carlson

has been accepted towards fulfillment
of the requirements for

Ph.D. Chemistry

degree in

Major professor

Date Se tember 5 1978

0-7639

   
     

LIBRARY

Michigan State , ‘

University >

  

 

 

 

 

THE DESIGN AND APPLICATION OF A
VIDICON RAPID SCANNING SPECTROPHOTOMETER
TO STOPPED—FLOW KINETICS

By

Erik Mattis Carlson

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Chemistry

1978

 

 

 

 

 

ABSTRACT

THE DESIGN AND APPLICATION OF A
VIDICON RAPID SCANNING SPECTROPHOTOMETER
TO STOPPED-FLOW KINETICS

BY

Erik Mattis Carlson

A rapid scanning stopped-flow spectrophotometer has been devel-
oped. The use of a silicon vidicon array detector enables reactions
to be simultaneously monitored over a spectral region of approximate-
ly 280 nm. Under computer control, this spectral region can be
divided into between 16 and 4096 wavelength channels, which can be
scanned at frequencies from 4 kHz to 100 kHz. This allows a spec—
trum with 100 channels to be acquired in as little as 1 msec.

The slow response time of the silicon vidicon detector was found
to limit its usefulness in the monitoring of fast reactions with large
changes in absorbance. Further experiments showed that the response
time of the detector can be improved by decreasing the frequency at
which the wavelength channels are scanned, increasing the number of
channels per spectrum and/or adjusting experimental conditions such
that the signal level from the detector is about 50% of saturation.

The instrument is controlled by a multi-microcomputer system.
One microcomputer controls the vidicon spectrophotometer. The
complete automation of the stopped-flow mixing system is handled by
a second microcomputer. Another microcomputer handles interactions
with the operator, mass data storage and communications with a

laboratory minicomputer. A fourth microcomputer coordinates the

 

 

 

Erik Mattis Carlson
activities of the other three. The use of a multi—microcomputer
system simplifies the software requirements for the individual
microcomputers, improves system modularity and allows a number of
operations to proceed simultaneously.

A new computer language called CONVERS-M was developed to
operate in the multi—microcomputer system. This language allows
new commands, which are geared directly to the instrument, to be
easily created and used. Commands have been created which dramat—
ically simplify the coordinations of activities among the micro-
computers. The use of CONVERS-M has greatly improved the versatility
of the instrument by allowing new modes of operation to be rapidly
developed and easily modified.

The instrument has been applied to an investigation of unusual
reaction rates encountered in a kinetic method for determining the
total protein content of human blood serum using the biuret reaction.
The unusual behavior was found to be primarily due to a shift in
the wavelength of maximum absorbance in the early part of the reac-
tion. Further studies indicated that the cupric tartrate complex
in the reagent was absorbing over the same wavelength region and

causing the shift.

 

 

 

To my devoted wife Stephanie and Mom and Dad

 

ii

 

 

 

ACKNOWLEDGMENTS

My sincere appreciation goes to Professor Chris Enke for the
help and encouragement he has shown me throughout the course of my
research at Michigan State University.

I want to especially thank Professor G. E. Leroi for serving as
second reader for this thesis and Professor Stan Crouch for many
valuable discussions.

I would like to thank Michigan State University and the Office
of Naval Research for the financial support they provided.

I want to thank the members of Professor Enke's research group
for their help and friendship, especially Spyros Hourdakis, Rick
Yost and Jim Hornshuh.

My special appreciation goes to my parents for their encourage-
ment and confidence which has helped so much.

To my wife, Stephanie, I offer my highest appreciation. Her love
and good humor have been a constant aid. The help she provided in
the preparation of this thesis, by typing the initial and final drafts,

proofreading and drawing nearly all of the figures, was tremendous.

iii

 

 

 

 

TABLE OF CONTENTS

LIST OF TABLES.
LIST OF FIGURES .
CHAPTER 1 — INTRODUCTION.
CHAPTER 2 — ROUTINE OPERATION OF THE INSTRUMENT .
A. Introduction .
B. An Overview of the Instrument.
C. Multi-Microcomputer System Software.
D. Data Analysis Software . . . . .

CHAPTER 3 — A GENERAL PURPOSE MULTI-MICROCOMPUTER
SYSTEM FOR RESEARCH INSTRUMENTATION .

A. Introduction . . . . . . . . . . . . .
B. The Evolution of the Multi-Microcomputer System.
C. The Structure of the Multi-Microcomputer System.

D. The External Communications System .

CHAPTER 4 - THE DESIGN OF A RAPID SCANNING
SPECTROPHOTOMETER FOR STOPPED-FLOW KINETICS .

A. Introduction .

B. Rapid Scanning Techniques in Stopped-Flow Kinetics .

1. Mechanical Scanning Techniques.
2. Array Detectors .

C. The Design of the Vidicon Rapid Scanning
Spectrophotometer. . . . . . .

CHAPTER 5 — THE DESIGN AND CONSTRUCTION OF AN AUTOMATED
STOPPED-FLOW MIXING SYSTEM. . .

iv

Page
. vii

. viii

16

21
21
22
3O

44

47
47
48
48

50

51

63

 

 

1.3.34

 

 

Page
A. Introduction . . . . . . . . . . . . . . . . . . . . 63
B. The Stopped-Flow Mixing System . . . . . . . . . . . 63

C. Stopping Syringe Velocity/Position Transducer. . . . 68

CHAPTER 6 - CHARACTERIZATION OF THE VIDICON, RAPID SCANNING,
STOPPED—FLOW SPECTROPHOTOMETER. . . . . . . . . . 73

A. Introduction . . . . . . . . . . . . . . . . . . . . 73 I

B. Characterization of the Vidicon Spectrophotometer. . 73

1. Introduction. . . . . . . . . . . . . . . . . . 73
2. Vertical Scan Frequency . . . . . . . . . . . . 74
3. Resolution. . . . . . . . . . . . . . . . . . . 80
4. Electron Beam Current . . . . . . . . . . . . . 87
5. Charge Integration. . . . . . . . . . . . . . . 87
6. Lag . . . . . . . . . . . . . . . . . . . . . . 91
C. Characteristics of the Stopped-Flow Mixing System. . lOl

D. The Monitoring of a Relatively Rapid Reaction. . . . 108

 

CHAPTER 7 — AN INVESTIGATION OF THE BIURET REACTION . . . . . 110
A. Background . . . . . . . . . . . . . . . . . . . . . 110
B. Experimental . . . . . . . . . . . . . . . . . . . . 112
C. Analysis of the Biuret Reaction Spectra. . . . . . . 120
CHAPTER 8 - A USER-DEFINABLE LANGUAGE

FOR COMPUTERIZED INSTRUMENTATION. . . . . . . . . 129

A. The Role of Software in Computerized
Instrumentation. . . . . . . . . . . . . . . . . . . 129
B. Programming with CONVERS—M . . . . . . . . . . . . . 136

C. The Structure of CONVERS-M . . . . . . . . . . . . . 141

APPENDIX A - CONVERS—M USER'S GUIDE . . . . . . . . . . . . . 147
A. Introduction . . . . . . . . . . . . . . . . . . . . 147
v

 

.45. l- .‘a' -"-

wars-'1- um:- "..-' .--..--.- "r1 '2-:.';.--r-:1_ -'-::..:1 - "*dTfiffi.)

 

Page
B. Special Characters . . . . . . . . . . . . . . . . . 147
C. Labels . . . . . . . . . . . . . . . . . . . . . . . 148
D. Parameter Passing. . . . . . . . . . . . . . . . . . 149

E. Mode Word. . . . . . . . . . . . . . . . . . . . . . 152

 

F. Using the Floppy Disk Drive. . . . . . . . . . . . . 152 v
i G. Assembly (Compilation of Stored Routines). . . . . . 154
H. Communication with the PDP ll/40 . . . . . . . . . . 155
I. Error Codes. . . . . . . . . . . . . . . . . . . . . 157
APPENDIX B - CONVERS—M BASIC COMMAND LIST . . . . . . . . . . 160
A. Coordinator's Basic Command List . . . . . . . . . . 160
B. Communicator's Basic Command List. . . . . . . . . . 168
APPENDIX C - CONVERS—M FLOW CHART . . . . . . . . . . . . . . 171

APPENDIX D - PROGRAM LISTINGS . . . . . . . . . . . . . . . . 174

 

A. Microcomputer Programs . . . . . . . . . . . . . . . 174
B. Minicomputer Programs. . . . . . . . . . . . . . . . 206

REFERENCES. . . . . . . . . . . . . . . . . . . . . . . . . . 214

vi

 

 

 

 

LIST OF TABLES

 

 

Table Page
2—1. Routine Instrumental Commands . . . . . . . . . . . . . 8
2—2. Instrumental Parameter Codes. . . . . . . . . . . . . . 11
3-1. Backplane System Bus Assignments. . . . . . . . . . . . 45
8—1. Comparison of Common Programming Languages. . . . . . . 135
A-l. Monitor Error Codes . . . . . . . . . . . . . . . . . . 158
A-2. 'PIP' Error Codes . . . . . . . . . . . . . . . . . . . 159
vii

 

I #7 '7 v
l“ .. 1

LIST OF FIGURES

Figure Page
2-1. Block Diagram of a Rapid Scanning

Stopped-Flow Spectrophotometer . . . . . . . . . . . . 5 '
2-2. Block Diagram of the Vidicon Rapid Scanning

Stopped—Flow Spectrophotometer . . . . . . . . . . . . 7
2-3. Absorption Spectra at Various Times During the

Biuret Reaction with Human Serum . . . . . . . . . . . 18
2—4. Absorbance (575 nm) Versus Time for the Biuret

Reaction with Human Serum. . . . . . . . . . . . . . . 19
2-5. Absorbance (839 nm) Versus Time for the Biuret

Reaction with Human Serum. . . . . . . . . . . . . . . 20
3-1. Distributed Single Minicomputer System.

LP = Line Printer; FL - Floppy Disk Drive;
DSC = Disk Drive; TRM Terminal . . . . . . . . . . . 23

Ill

3—2. Distributed Single Microcomputer System.
LP = Line Printer; FL = Floppy Disk Drive;
DSC = Disk Drive; TRM = Terminal; MC = Microcomputer . 24

 

3-3. Dedicated Bus System . . . . . . . . . . . . . . . . . 26
3-4. Time Shared Bus System . . . . . . . . . . . . . . . . 26
3—5. Cross-Bar Matrix Switch System . . . . . . . . . . . . 27
3—6. Master-Slave System. . . . . . . . . . . . . . . . . . 31
3-7. Distributed Multi—Microcomputer System. LP =

Line Printer; FL = Floppy Disk Drive; DSC = Disk
Drive; TRM = Terminal; MC = Microcomputer; TTY =

Teletype . . . . . . . . . . . . . . . . . . . . . . . 32
3-8. The Basic Bus Switch System. . . . . . . . . . . . . . 34
3-9. Memory Maps of Typical Slave and
Coordinator Microcomputers . . . . . . . . . . . . . . 35
3-10. Basic Slave Bus Switch . . . . . . . . . . . . . . . . 37
viii

 

 

 

Figure

3—11.
3-12.
3-13.
3-14.
4-1.
4-2.
4-3.
4—4.

4-5.

5-1.
5-2.
5-3.
5-4.

6-1.

6-3.

6—4.

6—5.
6—6.
6—7.
6—8.

6—9.

Addition of DMA Mass—Transfer . . . . . . . . . .
Basic Coordinator Bus Switch. . . . . .

Addition of the Register Bus.

Register Bus Switch . . . . . . . . .

The Silicon Vidicon Tube. . .

Raster Scan Pattern .

The Silicon Vidicon Target.

Equivalent Circuit for the Silicon Vidicon Target .

Detector Control System .

Data Aquisition System.

Block Diagram of the Stopped—Flow Mixing System .
Stopped-Flow System Mixer and Observation Cell.

Stopping Syringe Velocity/Position Transducer .

Velocity/Position Transducer Processing Circuitry .

Relative Intensity Versus Frequency .

Relative Intensity Corrected for Varying
Exposure Time Versus Frequency.

Relative Vertical Drive Amplitude Versus Frequency.

The Effect of Instantaneous Changes in Vertical
Scan Frequency. . . . . . . . . . .

Relative Intensity Versus Resolution.

Relative Intensity Versus Channel Separation.
Beam Width Versus Focus Coil Current.

The Effect of Instantaneous Change in Resolution.

Relative Intensity Versus Grid 1 Voltage
at Various Frequencies.

ix

Page

39

4O

42

43

53

53

56

57

58

61

65

67

69

71

75

76

77

79

81

83

84

86

88

 

   

  

vi...

' _ iii-a ' __ ti:- :’
I wan-'5 fishing.“ . ..
- I“ .3311
. __ ' -_ '-".I' ; .-:—:' . s . -
I -- ”$1.. -,.~‘:.“-1-:I-
l . - I

 

 

Figure Page

 

6-10. Effect of Rapid Changes in Charge Integration at
‘ Various Integration Levels . . . . . . . . . . . . . 90
1
j 6-11. Signal Decay After Removal of Illumination . . . . . 93
1 6-12. Relative Lag Versus Relative Intensity . . . . . . . 95

6-13. Relative Lag Versus Resolution (Frequency = 25 kHz). 96

 

6—14. Relative Lag Versus Vertical Scan Frequency '

(Resolution = 255 Channels/Frame). . . . . . . . . . 97
6—15. Relative Lag Versus Relative Vertical Target

Position . . . . . . . . . . . . . . . . . . . . . . 99
6-16. Relative Lag Versus Grid 1 Voltage . . . . . . . . . 100
6-17. Flow Velocity Output for 5 Separate Runs . . . . . . 103
6-18. A - Syringe Distance Traveled Versus Time

B - Syringe Velocity Versus Time . . . . . . . . . . 104
7—1. Absorption Spectra at Various Times During the

Biuret Reaction with Human Serum (2.2 g/L) . . . . . 115
7-2. Absorbance Versus Time for the Biuret Reaction

with Human Serum. (a) 575 nm; (b) 752 nm. . . . . . 116
7-3. Absorbance (839 nm) Versus Time for the Biuret

1 Reaction with Human Serum. . . . . . . . . . . . . . 117

7—4. Absorption Spectra at Various Times During the

Biuret Reaction with Bovine Serum Albumin (3.3 g/L). 118

7—5. Absorbance Versus Time for the Biuret Reaction
with Bovine Serum Albumin. (a) 586 nm; (b) 779 nm . 119

7-6. Absorption Spectra at Various Times During the
Biuret Reaction with Bovine Serum Globulin (3.3.g/L) 120

7-7. Absorbance Versus Time for the Biuret Reaction
with Bovine Serum Globulin. (a) 578 nm; (b) 775 nm. 122

7-8. Absorption Spectrum of the Copper-Protein Complex
from the Biuret Reaction with Human Serum After
About 25 Minutes . . . . . . . . . . . . . . . . . . 123

7-9. Absorption Spectrum of the Biuret Reagent. . . . . . 124

 

1 H.5- 31: -.- Luv-1min?) :r-J- : '.':-' : ' -. '- 1 "-'-" .‘J. Hi

 

 

 

 

Figure

7-10.

8-1.

8-2.

8—3.

8-4.

C-1.

C—2.

C—3.

 

Page
Absorption Spectrum of the Copper-Protein Complex
from the Biuret Reaction with Bovine Serum
Albumin After About 25 Minutes . . . . . . . . . . . 126
Programming Levels of Various Computer Languages . . 134
The Structure of a Command Definition. . . . . . . . 142
CONVERS-M Library Searching Procedure. . . . . . . . 143
CONVERS-M Memory Allocation. . . . . . . . . . . . . 145
Flow Chart of the CONVERS—M Command
Decoding Software. . . . . . . . . . . . . . . . . . 171
Flow Chart of the CONVERS-M Assembler. . . . . . . . l72
Flow Chart of the Label Decoding Software. . . . . . 173

xi

 

 

 

 

 

r——’—'

CHAPTER 1

INTRODUCTION

Most traditional spectrophotometric detection systems used for
stopped—flow kinetics are capable of monitoring reactions at only a
single wavelength at a time. This is a severe limitation for many
applications. For example, shifts in the wavelength of maximum
absorbance due to interfering species and the presence of inter-
mediates in reactions are very difficult to detect with this type
of system. Recent developments in techniques for rapid wavelength
scanning have helped to overcome this limitation. Rapid scanning
systems allow spectra to be acquired rapidly enough to allow re—
actions to be monitored over an entire spectral region.

This dissertation describes the design and characterization of
a multi-microprocessor controlled vidicon rapid scanning stopped-flow
spectrophotometer and its application to stopped-flow kinetics.
Each chapter was written to be as independent of the other chapters
as possible. This allows the reader to study only those chapters in
which he has special interest. A better understanding of the total
instrument, however, can be obtained by first reading Chapter 2.
Chapter 2 contains an overview of the instrument and a description
of its normal operation.

A powerful new multi—microcomputer system for instrumental
control is described in Chapter 3. The first part of the chapter
reviews traditional multi-microcomputer architecture. The rest of

the chapter contains a discussion of the design details of the new

 

 

 

 

 

multi-microcomputer system. This system simplifies hardware and

software design and makes possible the acquisition of multiple types
of data, coupled with simultaneous data analysis and instrument
control. An understanding of this chapter is not necessary before
reading any of the other chapters; however it does help to show how
the various parts of the instrument are coupled together.

A new user—definable language which operates in the multi-
microcomputer system is described in Chapter 8. This language allows
the creation of commands which interact directly with the instrument.
Because of this, instrumental control routines can be rapidly
created and modified making the total instrument very versatile.

Chapters 4 and 5 provide construction details of the vidicon
spectrophotometer and the stopped-flow mixing system respectively.
The first part of Chapter 4 also includes a review of all of the
rapid scanning techniques which have been applied to stopped-flow
kinetics.

Before any instrument can be operated with confidence, its
operating characteristics must be well understood. Chapter 6
describes a variety of experiments which were designed to help
characterize both the vidicon rapid scanning spectrophotometer and
the stopped—flow mixing system. The vidicon detector's slow re-
sponse time is its most serious limitation to its use in stopped—flow
kinetics. Chapter 6 contains a detailed study of the conditions
under which the vidicon's response time can be optimized. A final
test of the system was its application to the determination of the

rate constant for the well known iron/thiocyanate reaction.

 

 

    

A,

”mmujflg'fi 5.“,va aims}? lmchi- «If? '3: 3.113 mutil-

m‘ 155: '1' 1-.-l.i|h"‘"5 ~_--.u J'-."u-.' --.'i. --"‘- " 235::- '- -~ ' -""

_;" r‘-_'.:.;..' B:i::_ -|; . if“-

 

The biuret reaction is the basis for the standard clinical chem—

istry technique for determining total serum protein. The traditional
procedure involves letting the biuret reaction go to completion, which
takes at least 20 to 30 minutes. Recently, kinetic methods utilizing
the biuret reaction have decreased the analysis time to only a few
seconds. Although these techniques appear promising, a number of
problems still exist. Chapter 7 describes the application of the
rapid scanning stopped-flow system to an investigation of these prob-

lems.

 

a». liking-2 man. .m-

 

' " 'fh _.‘gfi..:.o.1,._
‘ “.7 .

 

 

 

CHAPTER 2

ROUTINE OPERATION OF THE INSTRUMENT

A. Introduction

The purpose of this chapter is to introduce the reader to the
routine operation of the instrument. This is done by first pre-
senting an overview of the basic parts of the instrument and des-
cribing the function of each part. The next section describes the
instrument's basic operating software. The data evaluation rou—

tines for use on our minicomputer are presented in the last section.

B. An Overview of the Instrument

In traditional stopped-flow spectrophotometers the reaction
under investigation is monitored by recording the change in absorb—
ance with time of either a reactant, product or intermediate in the
reaction. Since only a single wavelength can be monitored at a time,
this process must be repeated for each wavelength at which the
reaction is to be studied. The use of a rapid scanning spectropho—
tometer adds another dimension to the spectrophotometric study of
stopped-flow reactions by enabling the reaction to be monitored over
an entire spectral region. A block diagram of a typical rapid
scanning stopped—flow instrument is shown in Figure 2-1. The upper
half of the figure shows the stopped-flow mixing system. This is
simply a solution handling system for rapidly mixing two solutions
(here labeled sample and reagent) and delivering them to an obser-

vation cell. To monitor any subsequent reaction, light is passed

 

 

m.) in {gig-xiii“.'o:.uI-r.¢:1-uzx_{-"n'v .1. rum-taf- Lil"? “.1- para-‘11.; u...

. ... -. _ .: .- - _ ' ,
'IIr-l‘..';l.1'. ;s_--- - ‘.r -. 5'-

 

    
   
    
    

SAMPLE

DETECTOR

ABSORBANCE

DATA
PROCESSING

 

Figure 2-1. Block Diagram of a Rapid Scanning Stopped-Flow
Spectrophotometer.

 

through the observation cell, dispersed by a monochromator and
focused onto the rapid scanning detection system. The result is

a three—dimensional portrait of the reaction, with time, absorbance
and wavelength forming the three dimensions.

, The instrument described in this dissertation is controlled by

four microcomputers. A block diagram of this system is shown in '

 

‘ Figure 2-2. Three of the microcomputers are used for controlling

the three major sections of the instrument, while the fourth micro-
computer coordinates the activities of the other three. Microcom—
puter number 1 of Figure 2—2 controls the instrument's communication
with the outside world (see Chapter 3). This includes interaction
with the operator and a floppy disk mass storage device, and commun-
ication with our PDP 11/40 minicomputer. Microcomputer number 2
controls the vidicon rapid scanning spectrophotometer (see Chapter 4).

The vidicon spectrophotometer is a very sophisticated television

 

camera which allows spectra to be acquired very rapidly. The oper-
ation of the stopped-flow mixing system is controlled by microcom-
puter number 3 (see Chapter 5). This microcomputer not only handles
the entire mixing sequence, it also determines when the solutions
have completed mixing. This is very important, for it is at that

time that the rapid scanning system must begin acquiring data.

C. Multi-Microcomputer System Software

A listing of all of the routines mentioned in this chapter can
be found in Appendix D. The most important instrumental routines

are summarized in Table 2—1. This table lists the routine's name,

 

 

 

 

 

N

 

III
VID

.
_‘_ ICON
uu" —’
.
COORD
AAICRC) '
C

Figure 2—2. Block Diagram of the Vidicon Rapid Scanning
Stopped-Flow Spectrophotometer.

 

 

 

[Iii-villi;

iI .mpon m30fl>ogm mcflmmho Houmm mo uommumc muon we
r .Houusgm may mcomo .
M .muon m50fl>ogm mcflmwho psocuwz we wommuwv muofim mo
.mwommpmw Haw mo>osom .
.muomuuzo 30: z moumosu 2
.mo wommumc mo Edaficfle can ezefixme esp mwcflm ma
.cmmgm ecu mo mow
ecu um mafia m mcflzwhc kn uofim >.x ecu :o x we COHuflmom 0:“ moumooq x
.moEHp z Eoumxm zofimlwommoum can mesmSHm z
.AHOEoE dump cooflwfl>
8 ecu :fl 2 camcofi mo muommumc mo gonad: ofinfiwmom Edeflxms ecu mouwoau 0.2
.m0 pomuuso
:fi monoum wed womwumw :ooflcfi> :owo mo 2 Hoccmzo on «web ecu moxmh z.mu
.xaoEoE mump :oofivfl> memofio :
.mo wommumw mnmofiu mo
.Hma :fl pfismoy mohoum can Hwo Scam Nmo wommumw muomypnsm NmQ.HmQ

Foam
2mmo
Foam:
HHZHE
H302

x<2

004

:wqu

ZMZQ

EDD
xwmqu
m<mqu

¥U<m

IIIIIIlIIlIlIlllIIIIIIIIII|IIIIIIIIITIIIIIIIIIIIIIIIIIIIlIlIIllIIIIIIII|llII||lIlIIllllllllllllllllillllll

:ofiumflhomoo whouoadgwa

camEEou

IIlIIlIIIIIIlIlIIIIIllllllllllllllllllllllllIlIlIIllIllIIIlIIIIIIIIIIIIIlIIIIIIlIIIIIIIIIIIIIIIIIIIIIIIII

.mvcmesou prcoESaumcH ocflusom

 

.Hum ofinmh

 

 

 

 

 

 

 

.wu uequse :u ecuwoum qu00ue> ew2uuxm m meuoum mo cm>
.xuuusouuo mcucceem :oouwu> ecu meNuHmuuucc I HH2H>
.Zm ecum ecu meEe: paw cmuc AQQOcm :o muuoemm mo segue euuuce ecu meuoum 2m mmOHmH
.Euom HeEuoew cu chusueu ecu ou ma uemeuew musmuso mo H209
.zc euue ecu mesa: new cmue smm0uu so we commune meuoum zc.mo macaw
.ueuuscm ecu memoHu I 932m

.mhuoemm cm mcummucm Heume Esuuuemm c meuuzvom may ucuom Hewwuuu
Heume :ouuumusvue euew mcumec MmeEuu z Eeumxm mcuxue ecu mecmSHm Mm.mh.z z<ummm
.peheuue ec ou mueueEmuem cmuceesuumCu mzocc< I Hmm
.mueueeeuem :eom ecu xc weumHoemm we coeuwu> ecu Eoum mceom meuusvo< I z<um
.mueueaeuem deuceESMumcu ecu mo mesce> uceuuso ecu musmuso I esom

.c.e.ucooc .HIN muses

 

. c159

10

any input parameters and the function of each routine. All of this
software is written in the new computer language CONVERS-M, which is
described in Chapter 8. The normal operation of the instrumental
software is presented in this section by describing the procedure
involved in a typical stopped-flow experiment.

Before the instrument can be used for an experiment, the de-
sired scan parameters for the vidicon spectrophotometer must be
established. The input and output of these scan parameters, as well
as a number of other instrumental parameters, are controlled by a
set of parameter handling routines. Each parameter is identified
by a 2 to 4 letter code. The important instrumental parameters,
their identification codes, and their default values are listed in
Table 2-2. When the operator wishes to examine the current value of
the parameters, he simple types POUT (Parameter OUTput) at the ter-
minal. This causes the terminal display to be cleared and the
identification code for each parameter along with its current value
to be displayed. The operator can alter any of the parameters by
using the SET command. Upon typing this command, the routine outputs
an asterisk (*) to the terminal. The operator then types the
identification code for the parameter he wishes to alter along with
its new value. After this parameter has been entered the routine
outputs another asterisk indicating it is ready to input another
parameter. Parameter input is completed by typing END. For example,
if the operator wished to change the number of scans which were
averaged together (AVGS) to 4 and the number of integration frames

per readout frame (INT) to 2, he would type:

 

 

 

 

 

 

.muemeuec mo Hecesz o at

.muequdo mo Hecesz o 0*

W .euHsvoe ou euuoemm mo Hecesz H mzum
.mqeom Hem Hecceco mo Recess ecu wecHEHeueoV .uceEeHUCH anew HeucoNHHo: mmHecceco NoHvov wen

.uon ecu mo ESEHcHE Ho uemeuec e mo esHe> EDEHcHz 0 2H:

.uon ecu mo ESEwaE Ho uemeuec e mo esHe> ESEme: o x<z

.eueuweucu ou mceem mo hecasz o HzH

.wucHom cmmum mo Recesz o memo

.cmms xocesceum deem HeoHuHe> fimzxmmvnm ommm

1 .uemeuew :ooHcH> come :H mayo: Ho Hecasz 0 meme
1 .xde :OHueumeucH Use meew oHcouuueHm mHH >cmo
.uequso coee :H meuos mo Hecesz o memo

.3528 52: mafia; z. 3.5 93-32 32%

.Hecuewou emeue>m ou euuoemm Ho hecssz H mo><

.uecuemou emeue>e ou muuoemm mo mxeuue mo Hecesz H <w><

COHumHuomeo esHe> uHsmmeo ewou

 

.mevou ueueaeuem HmucessuumcH .NIN eHcmH

 

 

 

 

12
SET
* AVGS 4
* INT 2
* END

Many of the data handling routines of the instrument operate on
an entire set of data at a time. The size and number of these sets
of data are variable and must be initialized by the operator. The
software recognizes two distinct types of data. The most common
type of data is the data generated by the vidicon spectrophotometer.
Sets of this type of data are simply referred to as "datasets".

Sets of the other type of data are referred to as "cut-sets". This
name comes from the type of data which is normally stored in these
sets, which involves "cuts” of intensity with time through the
stopped—flow spectra at a single wavelength. All vidicon datasets
must reside within the 8 k-byte vidicon data memory area. The
memory allocation for the datasets is established by the DMEM com-

mand. If the following command sequence is typed at the terminal:
DMEM 102 0

the DMEM routine will divide the data memory into datasets, each of
which contains 102 words of storage (2 bytes/word). The largest
possible number of datasets will be created (40 in this case) and
each will be initialized to contain zeros. The creation of cut—sets
is controlled by the command MCUT. The creation of two cut—sets

would be accomplished by typing:

 

 

1 1 1 .5 I - _I- " - - I
. u h 4 I
- 'r '-

 

 

 

 

as".

13
MCUT 2

The length of each cut-set in words would be equal to the current
number of datasets. When a reference to a specific set of data is
required by a data handling routine, the reference must consist of
two parts. The first part of the reference is a single letter which
specifies the type of data involved (C for cut-sets and D for vidicon
datasets). The second part of the reference specifies the decimal
number of the set of data (from 1 to a number equal to the total
number of sets). For example, if the operator wished to plot the

data contained in vidicon dataset number 23, he would type:

PLOT D 23

Normally it is desired that the reaction data from the instrument
be in the form of the absorbance versus time over a specified spectral
region. Since the only output from the vidicon spectrophotometer is
in the form of the signal intensity versus wavelength, it is necessary
to calculate the absorbance values. This is done by a data evaluation
routine on our laboratory minicomputer. The absorbance (A) at wave-
length I is derived from the signal intensity (S) of a reaction spec-

trum by the equation:

udnere RI is the signal intensity with a reference solution and DI is

the signal obtained when no light is falling on the detector (dark

 

    

.. . . ' .I
._r- __ . r,

liiL-lai Tar-.35! $11.35*. ‘- . in. one: 3' '! ui.’ ".. 31:; .!='.'I '5‘“; "'1 - --. | -

 

 

14

current signal). This requires that three different types of data be
acquired. First, the observation cell must be filled with the sel-
ected reference (100% transmittance) solution. In order to assure

that old solutions have been removed completely from the observation
cell and drive syringes of the stopped-flow mixing system, the system
must be flushed a number of times with the reference solution. This is
accomplished by the FLUSH command. This command acquires an octal
number from the operator which designates the number of times the
stopped—flow mixing system is to be flushed. After the flushing is
completed, the operator acquires and stores the reference spectrum by

typing the following set of commands:

OPEN

CLR8K

SCAN

STORE D 1

* FILEl

The OPEN command causes the shutter to the light source to open. The
next instruction (CLR8K) clears the dataset memory. The spectrum is
obtained by the SCAN command. The spectrum, which is placed in data—
set D l by the SCAN command, is then stored on a floppy disk mass
storage device by the command STORE. This command sends an asterisk
to the terminal and the operator then specifies the name for the file
which is to be created. The dark spectrum is obtained in a similar
Jnanner except that the command SHUT (which closes the shutter) is used

in place of the OPEN command.

 

 

15

The next step in the procedure is to fill the syringes with the
two reactant solutions. This is accomplished by again flushing the
system by using the FLUSH command. The following command sequence is
then used to acquire the intensity versus time profile for the re-

action:

OPEN
CLR8K

SFSCAN l,47,5

The first two commands open the shutter and clear the dataset memory.
The next command (SFSCAN) requests three parameters from the operator.
The first parameter indicates the number of times the mixing system is
to be flushed before the actual data run is taken. The next parameter
is the trigger point. The trigger point is used to specify the time
at which the reactant solutions are mixed and the spectra can begin

to be taken. Under normal operation the trigger point is always set
to 47. (For more information on the trigger point and its selection,
see Chapters 5 and 6.) The final parameter specifies the number of
spectra which are skipped between the acquisition and storage of
succeeding spectra. The array of spectra which is acquired by the
SFSCAN command can be stored on floppy disk by using the command TSTORE.
This command is similar to the STORE command described earlier, except
that TSTORE outputs the entire array of datasets instead of just a
single dataset. TSTORE also sends an asterisk to the terminal, in-
dicating that the operator should enter the name of the file that is

to be created.

 

 

 

16

The reference and dark spectra should be taken immediately before
the reaction spectra. This helps to eliminate any errors caused by
slow drifts in the light source, dark current, etc. in the vidicon
spectrophotometer.

The more powerful data analysis routines are all run on our PDP
11/40 minicomputer. The process by which the data files are trans-
ferred from the instrument to the minicomputer is described in Sec—
tion H of Appendix A. The operation of some of the data evaluation

routines is the subject of the next section.

D. Data Analysis Software

A great many routines were created on our PDP ll/40 minicomputer
for the analysis of various types of data from the vidicon stopped-
flow system. All but two of these were special purpose routines which
were used for only a short time (primarily during the characterization
of the instrument). The two routines which are described in this
section are general purpose routines which were used for most of the
standard data analysis. All operator interactions are handled through
the minicomputer's terminal.

When the data files are first transferred from the instrument to
the minicomputer, they are stored in binary format to conserve space.
Since files in binary format are more difficult to work with, almost
all of the data evaluation and plotting routines were designed to
work with standard ASCII file formats. The program CONV is used to
convert the binary files into ASCII files. The program first requests

the file names of the input and output files. The program assumes

 

 

 

 

 

17

that the body of the data in the input file consists of only a single
variable (the 'Y' value). The operator then supplies an initial
value and an increment for a second variable (the 'X' value). Upon
completion of the conversion the output file contains the entire set
of data in the form of X and Y data pairs.

The evaluation of the stopped—flow reaction data is performed
by the SF program. This program starts by requesting three binary
input files. The first 2 files must contain the reference and dark
spectra. The third file contains the array of stopped-flow reaction
spectra. At this point, the data can be thought of as a three—
dimensional array with absorbance, wavelength and time forming the
three dimensions. The SF program allows two—dimensional slices to
be made through this array. These slices can be in the form of either
the absorbance versus time at a specified wavelength of the absorbance
versus wavelength at a specified time.

Figure 2-3 shows typical data from the vidicon stopped—flow
spectrophotometer. This figure shows the absorbance versus wavelength
at various times during the biuret reaction with human serum protein
(see Chapter 7). Figures 2-4 and 2—5 show the absorbance versus time
for this reaction at wavelengths of 575 nm and 839 nm respectively.
The sensitivity of the vidicon system for monitoring reactions is

quite apparent in Figure 2-5, where the total absorbance change is

less than 0.01.

 

 

 

 

01$

ABSORBANCE

550

Figure 2-3.

18

g 700
WAVELENGTH (IIIII)

Absorption Spectra at Various Times During the
Biuret Reaction with Human Serum.

 

850

 

 

0.25

AISOR BANCE

Figure 2-4.

19

 

I 2 3 4 5
TIME (n

Absorbance (575 nm) Versus Time for the Biuret Reaction
with Human Serum.

 

 

 

 

0.02 3

ABSORBANCE

Figure 2-5.

20

 

I 2 3 4 5
TIME (s)

Absorbance (839 nm) Versus Time for the Biuret Reaction
with Human Serum.

 

 

 

CHAPTER 3

A GENERAL PURPOSE MULTI—MICROCOMPUTER

SYSTEM FOR RESEARCH INSTRUMENTATION

A. Introduction

For many years the automation of laboratory instrumentation was
accomplished almost exclusively through the use of minicomputers.
Although minicomputers were far less expensive than large computer
systems, they were still expensive enough that most research laborator-
ies could not afford to use a separate minicomputer for each automated
instrument. Normally, a single minicomputer was used to control a
number of instruments. This reduced the total cost of the computer
system; however, the finite speed of the minicomputer meant that only
a small number of these instruments could be operated simultaneously.
More recently, the falling costs of minicomputer systems and particu-
larly the introduction of small, inexpensive microprocessors has made
it economically feasible to use one or even more computers to control
each instrument. This has made it possible for laboratory computer
systems to be implemented in a variety of configurations (1-6). The
optimum configuration of a laboratory computer system depends on the
environment in which it is found. In our research group we have devel-
oped a powerful multi—microcomputer system for controlling sophisti-
cated instrumentation in research laboratories. The evolution of our
laboratory computer system from a single minicomputer is traced, and
details of its current hierarchical multi-microcomputer configuration

is presented in the following sections.

21

 

 

 

22

B. The Evolution of the Multi-Microcomputer System

Originally, all the computerized instrumentation developed in our
research laboratory was controlled by a single minicomputer. In its
most advanced form, this single minicomputer system contained a Digital
Equipment Corporation PDP 11/40 minicomputer and various peripheral
devices. This system is shown in Figure 3-1. This type of laboratory
computer system had a number of disadvantages, the major one being
that almost all of the instruments interfaced to the system required
the full-time attention of the minicomputer. This meant only a single
instrument could be operated at any one time. Also, relatively complex
interfaces were required to connect the instruments to the minicomputer.
In order to overcome these difficulties, we began to search for ways to
improve our laboratory computer system.

The first proposed improvement to the laboratory computer system
was to place a microcomputer on each instrument. The proposed system
is shown in Figure 3—2. The microcomputers would handle the high—speed
data acquisition and instrumental control for each instrument. Any
high—level data evaluations which were required would still be handled
by the minicomputer. With this system, routine operation of the
instrument could be performed without requiring intervention from the
minicomputer, thereby allowing any number of instruments to be operated
simultaneously. Also, due to the microcomputer's simpler bus structure,
interfacing to the microcomputer was much simpler than interfacing
directly to the minicomputer.

The use of a single microcomputer to control an entire instrument

would be perfectly adequate for a large variety of applications.

‘-

 

 

 

 

23

  

INSTRUMENT
INSTRUMENT

Figure 3-1. Distributed Single Minicomputer System. LP = Line Printer;
FL = Floppy Disk Drive; DSC = Disk Drive; TRM = Terminal.

INTERFACE

   
   

     
  

INTERFACE

   

 

 

 

 

24

    
 

INSTRUMENT
INSTRUMENT

   

COMMUN.
CONTROL

        

Figure 3-2. Distributed Single Microcomputer System. LP = Line Printer;
FL Floppy Disk Drive; DSC = Disk Drive; TRM = Terminal;
MC Microcomputer.

 

 

 

 

 

25

However, many of the instruments we planned on implementing required
the asynchronous acquisition of multiple types of data at extremely
high data rates, coupled with simultaneous data evaluation and instru—
mental control. These instruments were complex enough that it would
have been extremely difficult to control them even with a dedicated
minicomputer and obviously impossible with a slower microcomputer.

The solution to this problem was to use a number of microcomputers.

By careful planning in the initial design of an instrument, tasks
which must be performed simultaneously can be assigned to different
microcomputers. In this way complex problems of simultaneous data
acquisition and/or instrumental control can be reduced to a few vir-
tually independent, synchronous tasks. This makes the required hard-
ware and software for any single microcomputer in the system far easier
to create and modify.

The overall effectiveness of any multiple microcomputer system is
determined to a large extent by the structure of the communication
system within and between the various microcomputers in the system.

A great many different configurations of multi-microcomputer systems
have appeared in the literature (7—16). Almost all of these multi-
microcomputer systems can be described in terms of three basic con-
figurations: the dedicated bus system (Figure 3-3), and the time—
shared bus system (Figure 3-4), and the cross-bar matrix switch system
(Figure 3-5). More complex configurations consist of either expansions
of these basic systems (e.g. multiple time—shared buses) or combinations
of them.

The dedicated bus configuration (Figure 3—3) is the simplest form

 

 

 

26

 

 

Figure 3-3. Dedicated Bus System.

 

 

Figure 3-4. Time Shared Bus System.

 

 

 

 

27

 

 

I/O
I

I/O
2

Figure 3-5. Cross-Bar Matrix Switch System.

I

 

 

IIIIIIIIIIIIIIIIlIIIIIIIIIIIIII!!!!!!!!EEEEEEEEE============::::;__-—_g=z.s_-v ,

28

of multi-microcomputer system. It consists of a number of virtually
independent microcomputers interconnected by means of a communications
bus. The communications bus can have many different forms. The bus
can connect I/O ports on the microcomputers or even be a direct CPU to
CPU link. Both serial and parallel data transfer formats can be im-
plemented, depending on the speed and complexity requirements of the
system. Since the individual microcomputers are virtually independent,
timing conflicts between them are essentially eliminated. This results
in overall high system throughput with relatively simple hardware and
software required to maintain the communications link. The major
disadvantages of this type of system are difficulty in expansion and the
inability of the system to reassign tasks dynamically in case of a
single microcomputer failure.

In its simplest form, the time-shared bus system, (Figure 3—4)
consists of a single bus shared by all the CPU's, memory and I/O ports
in the system. A bus arbitrator is required to resolve timing con-
flicts on the bus. The major disadvantage of this type of system is
the reduction in system throughput caused by bus conflicts between CPUs.
This situation can be improved (but never eliminated) by removing the
program memory and some of the I/O ports from the time-shared bus and
dedicating them to individual CPUs. In this way bus conflicts can
occur only when the microcomputers are accessing common memory and I/O
ports found on the time-shared bus. Communication between microcom-
puters in this system is normally accomplished by storing status words
in ”mailbox addresses" located in the common memory. The major
advantages of this type of configuration are relatively low intercon-

nection costs, modularity and the ability to reassign tasks dynamically

 

 

29

on the time-shared bus.

The most powerful of the three basic systems is the cross-bar
matrix switch configuration (Figure 3-5). With this configuration,
any CPU can access any memory or I/O port in the system. Conflicts
only occur when one device is accessed simultaneously by more than
one CPU. This configuration has very high system throughput, while
retaining the ability for dynamic task reassignment. Unfortunately,
the extreme complexity and cost of the matrix switch severely limits
the size and expandability of this configuration.

In any multi-microcomputer configuration the individual micro-
computers can either all be of the same "importance" or one or more
can be selected to be more "important” than the others. In a system
utilizing dynamic task assignments, this can take the form of a micro—
computer which determines the tasks that will be assigned to the rest
of the microcomputers. Its most extreme form is a hierarchical master-
slave system in which the slave microcomputers are essentially intelli—
gent peripherals of the master microcomputer.

The multi-microcomputer configuration which we finally adopted can
best be described as a hierarchical master-slave system with a hybrid
dedicated-bus/time—shared bus configuration. The master-slave config-
uration was selected to simplify the coordination of the various tasks
in the system. The high throughput required in the complex instruments
we planned to computerize was best accomplished through the use of an
essentially dedicated bus configuration. Some characteristics of a
shared memory system were also utilized for a simple yet powerful inter—

processor communications system. An approximate block diagram of the

 

 

30

system is shown in Figure 3-6. A detailed description of this system
is presented in the next section.

The resulting laboratory computer system is shown in Figure 3—7.
Communications between the minicomputer and the multi—microcomputer
systems is accomplished through the use of a communications controller
developed in our laboratory (17). Simple instruments which require
only a single microcomputer could also be developed and could communi-

cate with the minicomputer through this communications controller.
C. The Structure of the Multi-Microcomputer System

The multi-microcomputer system which we developed is based on
Intel's 8080A microprocessor chip. This microprocessor was selected
because of its high speed, extensive and relatively powerful instruc-
tion set and the availability of a large variety of standard support
chips. Although the system hardware would allow the use of virtually
any microprocessor chip, this has never been attempted due to the
complexity of the software which would be required in a system having
a variety of instruction sets. Each individual microcomputer in the
system has a standard 8080A structure with memory-mapped I/O used
exclusively. A basic familiarity of this structure and the standard
8080A bus signals is necessary to fully understand the detailed
structure of the multi-microcomputer system. This information is
readily available in the literature (18).

During the actual execution of an assigned task, each micro-
computer is completely independent of the others. The sequencing and

coordination of the individual tasks is accomplished by a single

 

 

 

  

Figure 3-6.

_ -421“ kgrfi ,, _

31

SLAVE T SLAVE 2

 

MASTER

Master-Slave System.

 

 

 

32

INSTRUMENT

    

CONUAUN.
CONTROL

   

INSTRUMENT

Figure 3-7. Distributed Multi-Microcomputer System. LP = Line Printer;
FL Floppy Disk Drive; DSC = Disk Drive; TRM = Terminal;
MC Microcomputer; TTY = Teletype.

II II

 

 

33

microcomputer which we have appropriately labelled the coordinator
microcomputer. The coordinator microcomputer communicates with the
slave microcomputers in the system through the use of a dedicated
communications bus. When in use, this communications bus is a simple

extension of the coordinator's local bus. The communications bus

 

consists of 16 address lines, 8 data lines and 5 control lines (MEMW,

 

MEMR, DBIN, READY and MASTER—RESETU. A bus switching system is utilized
to insure that only one slave is accessed at a time. A basic block
diagram of this system is shown in Figure 3-8. The bus switch is
controlled by the upper address bits. We have found that in most

cases a total of three address bits (allowing a total of 7 slave micro-
computers) is optimum. The memory maps of typical slave and coordin-
ator microcomputers are shown in Figure 3-9. With this configuration,
an 8K block of each slave is accessible by the coordinator. This
address block is always arranged to coincide with the start of the
address space in the slave. Normally this block contains the slave's
random access memory (RAM) and any I/O ports which might be useful to
have accessible to the coordinator. The remaining 56K of the slave's
memory contains the rest of the I/O ports and the slave's read only
memory (ROM).

The basic communications system also includes a set of three lines
connecting the coordinator with the slaves. These lines consist of a
slave hold line, task completed flag and reset line for each slave in
the system. The hold line is used to allow the coordinator to control
When each slave in the system is running. The task completed flag

iJidicates when a slave has completed an assigned task. The coordinator

 

 

 

34

COORDINATOR

  

SLAVE I

   
 

COMMUNICATIONS BUS

Figure 3-8. Basic Bus Switch System.

 

 

35
64K COORDINATOR SLAVE 64 K
IUOT
SLAVE 7 ACCESSIBLE
BY
56K COORDINATOR

 

ROAA

SLAVE I

LOCAL

RAM, ROM, I/O

 

Figure 3—9. Memory Maps of Typical Slave and Coordinator Microcomputers.

 

 

 

can force any slave to start program execution at address zero by

using that slave's reset line.

The circuit diagram of a standard slave bus switch is shown in
Figure 3-10. When the coordinator attempts to access an address on
the slave, the address decoder causes a hold to be sent to the slave
CPU. The coordinator's READY line is also pulled down causing the
coordinator to go into a wait state. When the slave completes its
current instruction cycle the HLDA (HoLD Acknowledge) line goes high,
and the slave's system bus is tri—stated. This causes the bus switch
to be enabled and the coordinator to be released from its wait state
to complete the transfer. The DBIN signal from the coordinator‘s
control bus determines the direction of the data transfer through the
bus switch.

The bus switch and the three external control lines enable the
coordinator to exercise a high degree of control over the execution
of tasks by the slaves. The lowest address block in each slave
contains RAM program memory. The first few addresses always contain
a subroutine call instruction. When the coordinator requires a task
to be performed by a slave, the coordinator uses the bus switch to
place the address of the subroutine which accomplishes that task into
this subroutine call instruction. The coordinator then resets the
slave and releases its hold line. The slave will execute the sub—
routine call and perform the desired task. On returning from the task
subroutine the slave sends the task completed flag to the coordinator.
For simplicity, the task completed flag is usually generated by the

INTE (INTerrupt Enable) signal in the slave's CPU. This allows a

 

COMMUNICATIONS BUS

Figure 3-10.

 
   

 

37

   

SLAVE
CPU

III

I ADDRESS BUS
D

ATA BUS

 

 

 

 
 
 
 
 

   

SWITCH

CONTROL BUS

 
 

Basic Slave Bus Switch,

 

 

38

single, rapid one byte instruction to generate the flag. General
purpose routines have been created which greatly simplify the pro-
gramming required to enable the coordinator to force a slave to
execute a specific subroutine. Further details about these routines
are found in Chapter 8.

In most applications, the bus switch and external control lines
are adequate for all forms of communication between the various micro-
computers. Certain applications however, require more advanced forms
of communication. One such situation is the need to transfer large
blocks of data at very high data rates. The coordinator can be used
to transfer blocks of data, but its data transfer rate is limited to
a maximum of 50 kHz. By using direct memory access (DMA) circuits the
data transfer rate can be raised to at least 1 MHz. If these DMA cir—
cuits were located on the coordinator's bus the coordinator's through-
put would be seriously reduced. To alleviate this problem, the basic
system can be modified as shown in Figure 3-11. By placing a bus
switch between the coordinator's system bus and the communications bus,
the coordinator can be isolated from the communications bus when it is
being used for DMA data transfers between slaves. DMA transfers are
set up by the coordinator. Any conflicts on the communications bus
between the coordinator and the DMA circuitry are resolved by the
switch control. The coordinator is always given the highest priority
for communications bus access. Figure 3—12 shows the standard bus
switch for the coordinator. When the coordinator attempts to access
an address requiring the use of the communications bus, the communica-

tion bus request line (CBR) goes high. Whether or not the coordinator

 

 

 

39

       
   
 

COORDINATOR
CONTROL

l
o SLAVE I

CONTROL
I
0 SLAVE 2

 

Figure 3-11. Addition of DMA Mass-Transfer.

 

 

 

 

40

m
D
m
m
2
O
z
<
E
z
3
E
E
O
U

DATA BUS

CONTROL BUS

 

Figure 3-12. Basic Coordinator Bus Switch,

 

 

 

41

is granted immediate access to the bus is determined by the state

of the communication bus busy line (EBB). When this line is high, the
coordinator is granted immediate access. If the bus is currently busy,
this line will be low, and the coordinator will enter a wait state
until the current DMA transfer cycle is completed.

Another form of communication which is sometimes required is the
transfer of status words or a few bytes of data between slaves. This
can be accomplished by the coordinator, but the resulting transfer is
extremely slow and degrades the throughput of the coordinator and both
slaves. To overcome this limitation, an additional dedicated communi—
cations bus was designed. This new bus, named the register bus, is

shown in Figure 3-13. The register bus consists of 8 address lines,

 

8 data lines and 3 control lines (MEMR, MEMW and DBIN). All micro—
computers in the system have equal priority for access to the register
bus. By means of the register bus, it is possible to access special
registers which are located on the individual microcomputers. In

this way, slave 1 could write a status word into a special register
located on its bus and that status word could be read by any other
microcomputer in the system by using the register bus. Bus conflicts
on the register bus are resolved by the register bus switch circuit
(Figure 3-14) and a register bus controller circuit. Each micro—
computer requiring access to the register bus must contain the bus
switch circuit. When a microcomputer requires the use of the register
bus, it generates a chip select dfifi signal that causes a register bus
rcsquest (BBB) to be sent to the controller. The microcomputer will be

iri a wait state until the register bus busy line (RBB) goes high and

 

 

 

 
    
   
    
 

- m
I
I
I
I
OL
CONTR COORDINATOR CONTROL
I
I/O
0

2
g 9
m m

.4
U5 "'1
z CONTROL SLAVE] CONTROL a
Q a
p c
< I I V‘
g u
s - °
2
E l/O
O
u I
SLAVE 2
I ‘n
. 0
I/O
2

Figure 3-13. Addition of the Register Bus.

 

 

 

43

__ -, READY
Bl. .

ADDRE
DECOD

REGISTER BUS

CPU
SS
ER

  
  

3-STATE ADDRESS BUS
BUS DATA BUS

CONTROL BUS

  

SWITCH

DIR

Figure 3-14. Register Bus System.

 

 

 

 

 

44

it is granted access to the bus.

Even when the register bus and DMA transfer circuits are included,
the total circuitry required to implement the inter-processor communi-
cations system is surprisingly small. The standard use of wire wrapping
for all digital circuit connections allows the microcomputer and inter-
face chips to be packed tightly together. Often the entire slave
microcomputer system including the CPU, memory, task interface and all
the bus interfacing circuits can be easily contained on a single 6 x 8
inch circuit card. An entire multi-microcomputer system is normally
contained in a standard BUD card rack. This card rack contains
twelve, 58 pair printed circuit board edge connectors. The lower 39
pairs of the edge connectors are bused together. The pin assignments
of this backplane system bus are given in Table 3-1. The upper 19
pair of connectors are available for individual connections between

boards and to external circuits.

D. The External Communications System

One of the primary design goals of this laboratory computer
system was to make each instrument virtually independent of the mini—
computer during standard operation. In order to accomplish this goal
adequately, the operator must be able to run a variety of experiments
with the instrument, evaluate the results and store selected results
without any intervention from the minicomputer. The ability to design
and execute totally new experiments without using the minicomputer
would also greatly improve overall laboratory throughput.

The realization of these goals requires rather complex

 

 

Table 3-1.

Pin Component Side

20
21
22
23
24
25
26
27
28
29
30
31

33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58

 

45

Backplane System Bus Assignments.

Data

II
II
II
Address
II
II

II

DBIN

ONAOONAO‘

32 MASTER RESET

Address 14

II 10
II 8
Address 6
II 4
II 2
II 0

+ 24V
- 24V
- 15V
+ 15V

+ 5V

Wire Wrap Side

Data

II
II

II

Address

HWU‘IVI—‘LNUIV

El

Address

II

II

MEMW
MEMR
GND
+ 24V
— 24V
- 15V
+ 15V
GND

GND

Comments

Register Bus
II

II
II
II
II
II
II
II
II
II
Communications Bus
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
II
Power
II
II
II
II
II
II

II

 

-_,..

 

46

communications external to the multi-microcomputer system. The system
currently provides a variety of modes of external communication.
Typically, communication between the operator and the system is
accomplished through a video terminal. This terminal consists of a
Ball 9 inch CRT monitor, a Matrox MTX—16328L 16 x 32 character alpha-
numeric display circuit, a Matrox MTX-2562-1 graphics display circuit
and an Electronics Warehouse EW—lOO ASCII keyboard. The alphanumeric
capabilities of the terminal allow it to be used not only for operating
the instrument, but also for program development. The terminal's
graphics system allows preliminary data from the instrument to be
easily displayed. A Persci Model 270 Dual Diskette Drive with a Model
1070 Intelligent Diskette Controller provides local mass storage for
programs and data. Parallel, high speed data transfers between the
minicomputer and the multi-microcomputer systems are performed through
the use of a communications controller developed in our laboratory (17).
This controller also contains a link to a teletype for generating
printed copies of programs and data. All of these communications

tasks are performed by a single slave micro, the communicator slave
micro. Standard routines have been developed for all of these task;

details of these routines can be found in Chapter 8.

 

CHAPTER 4

THE DESIGN OF A RAPID SCANNING

SPECTROPHOTOMETER FOR STOPPED-FLOW KINETICS

A. Introduction

Most spectrophotometric detectors used for stopped—flow kinetics
monitor the course of a reaction at only one or two wavelengths. The
ability to monitor the course of a reaction simultaneously over a
range of wavelengths offers tremedous advantages in certain applica-
tions. One of the most important applications of rapid wavelength
scanning techniques in stopped—flow kinetics is in the determination
of the number and types of species present, particularly the detection
of reaction intermediates and the subsequent investigation of the
kinetics of the formation and decay of these intermediates (19, 20).
Rapid scanning techniques also allow two or more reactions to be
monitored at different wavelengths simultaneously (21, 22). Also,
many standard kinetics investigations could benefit by the initial
use of a rapid scanning system for establishing the optimum analysis
condition for later study with a more sensitive single wavelength
system (23).

The first part of this chapter contains a review of the types of
rapid wavelength scanning techniques which have been applied to
stopped-flow kinetics. The last section describes the design of the
silicon vidicon rapid scanning spectrophotometer which was used

throughout this research.

47

 

48

B. Rapid Scanning Techniques in Stopped-Flow Kinetics

A great many techniques have been developed which allow a range
of wavelengths to be rapidly scanned, and a number of excellent re—
views can be found in the literature (24—27). However, only a few
rapid scanning techniques have been applied to stopped-flow kinetics.
These techniques can be divided into two general categories according
to the method by which the wavelength is scanned. Mechanical scanning
techniques involve the mechanical movement of a mirror, grating, or
other element in the optical system, which causes the wavelength
region to be swept past a stationary detector. Array detectors com—
prise the other general category. In this method, the dispersed
spectrum falls directly on the array detector, which performs the

wavelength sorting electronically.

1. Mechanical Scanning Techniques

The first application of rapid scanning techniques to stopped-
flow kinetics was reported by Dye and Feldman (28). The system
utilized a Perkin—Elmer Model 108 rapid scanning monochromator (no
longer commercially available), and wavelength scanning was accom-
plished by a rapidly rotating mirror. The resulting spectra were
displayed on an oscilloscope. This system has since been modified
and interfaced to a minicomputer (29, 30). The scan rate can be
varied between 3 and 150 spectra per second. With a fixed sampling
rate of 20.4 kHz, the resolution at the above scan rates varies from
136 to 6800 sampled points per spectrum. The width of the wavelength

region which is scanned can be varied by the tilt of the rotating

 

 

 

49

mirror. Noise levels as low as 0.002 A.U. (absorbance units) have
been reported (31).

A mechanical scanning technique is also used in the system de—
veloped by Wightman, et a1. (32). This system is based on the
Harrick Scientific Corporation rapid scanning spectrometer originally
developed by Kuwana (33). Wavelength scanning is accomplished by
the mechanical movement of a small mirror mounted on a low inertia
galvanometer armature. The movement of the mirror is produced by
the application of a sawtooth or triangular waveform to the galva—
nometer. The width of the spectral region scanned, the number of
sampled points per scan, the scan rate and the repetition rate can
all be varied under computer control. At maximum, a 50—point 250 nm
spectrum can be scanned in 1.2 msec with a repetition rate of 500 Hz.
However, wavelength nonlinearities were observed at high scan rates.
Single scan noise levels as low as 0.01 A.U. were reported (32).

Hollaway and White have developed a rapid scanning stopped—flow
spectrophotometer for applications in biochemical studies (34).

This system utilizes a Norcon Model 501 scanning monochromator.
Wavelength scanning is accomplished by the rotation of a wheel con-
taining 24 mirror pairs. Variations in the mounting of the 24 mirror
pairs causes a different baseline for each spectrum. This problem
has been overcome by a circuit which switches in a correcting voltage
for each spectrum. The detection system is linear over a range of
1.5 absorbance units. The single spectrum noise was found to be
normally less than 0.02 A.U. A single spectrum of about 200 nm is

taken in l msec with a repetition rate of 800 spectra/sec. A total

 

 

 

50

of 32 spectra, each consisting of 256 8—bit words, can be stored
with the current system.

Another rapid scanning stopped—flow system for biochemical ap—
plications has been reported by Faini, et a1. (35). Mechanical
wavelength scanning is accomplished by the movement of a small grat—
ing attached to a galvanometer. A computer—generated triangular
waveform is used to move the galvanometer. A 128 point 400 nm
spectrum can be acquired in 6 msec. However, the repetition rate

is only 50 spectra/sec.

2. Array Detectors

The first application of an array detector to stopped—flow
kinetics was demonstrated by Milano, et a1. (36). This system
utilized a silicon vidicon TV camera tube (RCA No. 4532A) as a
detector. The original system could operate over the wavelength
region from 400 to 800 nm. Later, the use of a silicon vidicon tube
with a fused silica faceplate (RCA No. C23246) extended the useful
wavelength region down to about 200 nm (19). The wavelength region
scanned in single spectrum could be varied from 100 to 400 nm by
changing the grating. Scanning rates could be varied from less than
0.1 Hz up to 250 Hz for the standard 200 point spectrum. The de-
tector response was found to be linear with intensity over a range
of four orders of magnitude. The single scan noise of this system
was 0.002 A.U.

Nieman and Enke also developed a silicon vidicon rapid scanning

spectrophotometer which has been applied to stopped-flow kinetics (23,

 

 

51

37, 38). This system was shown to be linear over at least 3 1/2
orders of magnitude with a single scan signal-to-noise ratio of 220.
The number of electronic channels in the 230 nm wide wavelength
window could be varied from 32 to 4096 under computer control. Scan
repetition rates as high as 500 Hz could be achieved.

Dessy has reported the design of a rapid scanning stopped—flow
spectrophotometer which utilizes a self-scanning linear photodiode
array as a detector (39). However, no data have ever appeared on
the characterization or application of this system. The lack of
application of linear photodiode arrays as stopped-flow detectors
is probably due to the low sensitivity and S/N ratio which have been
reported for these devices. Yates and Kuwana evaluated the use of a
linear photodiode array as a rapid scanning detector and found that
it was linear from 0 to 1.5 A.U. and had a maximum signal-to-noise
ratio of only 100 (40). They believed that much of their noise was
due to their signal processing electronics. Other sources have

reported somewhat better results (41, 42).

C. The Design of the Vidicon Rapid Scanning Spectrophotometer

This section contains an overview of the design of the vidicon
spectrophotometer which was used throughout this research. The
actual construction of the scanning and detection circuitry was done
by Jim Hornshuh, and construction details can be found in his thesis
(43).

The silicon vidicon tube was selected as the detector in the

stopped—flow system for a number of reasons. Previous experience in

 

 

 

52

our group with the vidicon detector (38) had convinced us of the
power of this detector and suggested methods by which the overall
performance of the system could be improved. In general, vidicon
detectors tend to show lower noise levels than mechanically scanned
detector systems. This is due primarily to a combination of two
factors. The main source of noise in mechanical systems often results
from mechanical vibrations in the scanning element (32), which are
eliminated in the vidicon detector. Further noise improvements are
a result of the integrating character of the vidicon detector.
Finally, the vidicon detector allows sophisticated yet flexible
computer control of the electronic scanning circuitry.

The silicon vidicon tube which was selected in this work was
the RCA C23246. This tube has a fused silica faceplate to allow
operation in the UV region. Figure 4-1 shows the typical arrangement
of a silicon vidicon tube and associated magnetic deflection coils.
When the dispersed radiation is focused on the vidicon target an
equivalent charge pattern is generated. This charge pattern is read
out by scanning an electron beam across the target. Normally, scan—
ning is performed by a raster scan technique as shown in Figure 4—2.
With this type of pattern, the horizontal or wavelength axis is
generated by a staircase deflection of the electron beam by the
horizontal deflection coil. Each step in the staircase represents
a separate wavelength. Vertical scanning along the slit height axis
is accomplished by applying a sawtooth waveform to the vertical
deflection coil. The dashed lines in Figure 4—2 represent the ver-

tical and horizontal retrace paths followed by the electron beam.

 

 

53

ALIGNMENT COIL

\ FOCUSING cons
I: DEFLECIION YOKE
TARGETN

ELECTRON GUN
l

NELECTRON BEAM I

    
   
 

  

ANODE MESH—)1

VIDEO

SIGNAL

Figure 4-1. The Silicon Vidicon Tube.

 

WAVELENGTH

Figure 4—2. Raster Scan Pattern.

 

I . - ' ' I' I I a- ' I. - I" I it”
_-.-— _ . l- ..

_ . _ ..—.—— -_ .g._
I

 

 

 

,_
@‘m :._-=»-,.—_-, .3 z— .4

‘ 54

Any undesired read out of the target during retrace can be eliminated
by blanking (turning off) the electron beam. For a silicon vidicon,
the amount of current which exists during the scanning of the photo—
induced charge pattern, can be related to the incident photon flux

by the equation (44):

where: I = Video current, amps

Q = Quantum efficiency of the photosurface
e = Electron charge, c

N = Incident photon Density, quanta cm_2s_1
A = Target area scanned per channel, cm2

T = Detector exposure time, s

T = Time to scan area A, 5

Typically, the video output current varies from a few nanoamperes to
a maximum of about 700 - 800 nA. The integrating character of the
silicon vidicon detector is indicated by the fact that the video
current is directly proportional to the detector exposure time.

To understand fully the way in which the vidicon detector integrates
the incident photon flux, it is necessary to look at the vidicon
target in more detail. The target consists of an n—type silicon
wafer which is approximately 7/8 inch in diameter (45). The edges of
this wafer are made relatively thick for improved strength while the
center is extremely thin, normally about 5 - 50 pm thick. The

central section of the wafer contains a large array of silicon

 

 

55

photodiodes, which are simply islands of p-type silicon (see Figure
4—3). The packing density of the photodiode array is typically 1800
diodes per linear inch. The active video area of the wafer is
typically 5/8 inch in diameter. The p—n junctions form charge de-
pletion regions which essentially isolate the p-type islands from

the substrate and allow them to function as charge storage areas.

The p-type islands can be thought of as the combination of a capac-
itor (charge storage region) and a photodiode. Figure 4—4 shows an
equivalent circuit for the vidicon target based on this concept. The
electron beam is normally kept at a voltage of -8V with respect to
the substrate. As the electron beam moves across the target the p-
type islands are charged to —8V. When photons fall on the opposite
side of the wafer, they generate electron—hole pairs. The voltage
across the wafer tends to draw the holes into the p-type islands,
which thereby slightly discharge the stored charge. When the electron
beam rescans these diodes the current necessary to recharge the diodes
to -8V is equal to the amount of charge discharged by the photon—
generated electron—hole pairs. When no light is incident on the
target, charge depletion still occurs due to thermally generated
electron-hole pairs. This gives rise to a video output in the ab—
sence of illumination known as the dark current.

The silicon vidicon spectrophotometer control functions, which
are of primary importance to its use in rapid scanning applications,
are shown in Figure 4—5. All of these functions are completely
under computer control. The vertical scan frequency can be varied

from 3.91 kHz to 1 MHz, but the data acquisition system limits the

 

 

55

photodiodes, which are simply islands of p-type silicon (see Figure
4-3). The packing density of the photodiode array is typically 1800
diodes per linear inch. The active video area of the wafer is
typically 5/8 inch in diameter. The p-n junctions form charge de-
pletion regions which essentially isolate the p—type islands from

the substrate and allow them to function as charge storage areas.

The p—type islands can be thought of as the combination of a capac-
itor (charge storage region) and a photodiode. Figure 4-4 shows an
equivalent circuit for the vidicon target based on this concept. The
electron beam is normally kept at a voltage of —8V with respect to
the substrate. As the electron beam moves across the target the p—
type islands are charged to -8V. When photons fall on the opposite
side of the wafer, they generate electron—hole pairs. The voltage
across the wafer tends to draw the holes into the p-type islands,
which thereby slightly discharge the stored charge. When the electron
beam rescans these diodes the current necessary to recharge the diodes
to -8V is equal to the amount of charge discharged by the photon-
generated electron-hole pairs. When no light is incident on the
target, charge depletion still occurs due to thermally generated
electron—hole pairs. This gives rise to a video output in the ab-
sence of illumination known as the dark current.

The silicon vidicon spectrophotometer control functions, which
are of primary importance to its use in rapid scanning applications,
are shown in Figure 4-5. All of these functions are completely
under computer control. The vertical scan frequency can be varied

from 3.91 kHz to 1 MHz, but the data acquisition system limits the

 

 

56

Si02 INSULATION

   
    
      
  
    

    
    

(_..

G PHOTON
—->
’ DEFLECTION
\

\/ REGION
\

I
I

/
ELECTRON

BEAM ”I’m

-8V

 
     

P-TYPE
ISLAND

N-TYPE
SUBSTRATE

VIDEO
OUTPUT

Figure 4-3. The Silicon Vidicon Target.

 

 

 

57

__TARGET

E

E L ECTRON BEAM

    

VIDEO
OUTPUT

Figure 4-4. Equivalent Circuit for the Silicon Vidicon Target.

 

 

 

 

H Qqu

  

      

 

mmmmsm

 

mw<qu>uzuHI

mm>Hzo
onhumAme

4<HZONHmOI

mm>~mo
onHumqmmo
A<UHHmm>

 

   

 

moe<mmzmu
mm<umH<Hm
A<HZONHmOE

 

.Eopmxw Houpcou ~38an

          

moe<mmzmu
Hzmmmzu

z<mm

 

  

        

III III OMUHE
m><qm

    

moe<mmzmu
>uzm20mmm
4<uHemm>

 

.m-v ”Aswan

 

 

 

59

maximum practical vertical scan frequency to 100 MHz. The vertical
scan frequency is produced by taking the output of a 10 MHz crystal
oscillator through an 8-bit programmable divider circuit. The
vertical scan frequency is then produced by dividing the divider
circuit frequency by 10. The square wave created by the vertical
frequency generator is then sent to the vertical deflection driver,
which creates an approximately 200 mA peak—to—peak sawtooth waveform
in the vertical deflection coil. The final division by 10 in the
frequency generation system is done to create a control signal which
is always exactly 10 times higher in frequency than the vertical
scan frequency. This control signal is used in the video signal
processing system discussed later.

The horizontal deflection circuitry creates the programmable
staircase waveform used for electron beam deflections on the wave-
length axis. The staircase waveform is produced by the continuous
addition, into an accumulator, of a programmable 8-bit step width
value. The 12-bit accumulator is reset to zero on overflow. The
number of steps per staircase can be found by dividing 4096 by the
step width value. This technique allows the number of channels per
spectrum (steps per staircase) to be varied from 16 to 4096. The
addition is controlled by the vertical scan. The accumulator output
goes to a D/A converter (DAC) and finally to the horizontal deflec-
tion driver, which drives the 20 mA pp staircase through the hori-
zontal deflection coil.

The length of time required to read out a single spectrum de-

pends on both the vertical scan frequency and the number of channels

 

 

 

 

60

per spectrum. The single spectrum scan time in msec can be deter-
mined by dividing the number of channels per spectrum by the
vertical scan frequency in kHz. For example, a vertical scan rate
of 50 kHz with 100 channels per spectrum would result in a spectral
scan time of 2 msec.

The final control system of major importance in rapid scanning
regulates the electron beam current. The beam current is controlled
by the negative voltage applied to Grid 1 in the vidicon tube; the
more negative the voltage the smaller the beam current. This voltage
is controlled by an 8—bit DAC coupled with a high voltage buffer
amplifier. The voltage on Grid 1 can be varied from about zero V to
over -100 V.

Figure 4—6 shows the data acquisition system. Amplifiers Al —
A3 constitute the preamplification system for the vidicon detector.
They are mounted within a few inches of the vidicon tube. The
video output current from the vidicon target is converted to a
voltage signal by amplifier Al. This voltage level is then amplified
and driven back to the multi-microcomputer control system by A2 and
A3. Additional voltage gain is provided by A4, and the resulting
signal is electronically integrated by A5. The conversion of the
integrator output to a digital form is accomplished by a sample-
and-hold and a 12-bit A/D converter. A programmable, crystal-con-
trolled sequencer is used to control the data acquisition system.
The control of solid state switches 51 and $2 assures that A5 inte—
grates the signal only from a single vertical scan. These switches

can also be programmed to enable a limited amount of vertical scan

 

 

 

 

 

61

SLAVE
MEMORY M,CR° 2
I II II
"OGRAMMABH
FAST CRYSTAL-CONTROLLED

ADD‘R SEQUENCER
DMA

INTEGRATOR

VIDEO

INVERTER WITH GAIN

CURRENT TO VOLTAGE
CONVERTER

Figure 4-6. Data Acquisition System.

 

 

BUFFER WITH GAIN

    

BOOSTER

 

 

 

62

resolution to be obtained. The section of the vertical scan which
is integrated can be varied, under computer control, by varying
the timing of the closure of $1. In this way, up to 8 or 9 bands
(depending on the vertical scan frequency) can be resolved on the
vertical axis.

The data rates of up to 100 kHz of which the system is capable
are much too fast for the microcomputer to acquire through software.
To overcome this limitation, a specialized direct memory access (DMA)
circuit was designed. This DMA circuit can be programmed to store
spectra at any location within an 8K—byte data memory block. The
DMA circuit operates by adding together the contents of the data
memory and the incoming spectral data. This allows simple averaging
of multiple spectra. The number of bytes of precision used can be

varied under software control.

.. ,gfl4,>¥

 

 

CHAPTER 5

THE DESIGN AND CONSTRUCTION OF AN AUTOMATED

STOPPED-FLOW MIXING SYSTEM

A. Introduction

The stopped-flow mixing system described in this chapter is
simply a sample handling system which allows two solutions to be
rapidly mixed (in less than 0.01 s) and delivered to an observation
cell. Any subsequent reaction between the two solutions can then be
monitored by the vidicon spectrophotometer (see Chapter 4). The
first stopped-flow mixing system was constructed by Chance in 1940
(46—48). Since that time a large number of modifications and ap—
plications of stopped—flow mixing systems have appeared in the lit-
erature. Recently an excellent review of the stopped—flow technique
and existing stopped-flow instrumentation was published (49). The
level of automation in stopped-flow systems varies widely. In the
most highly automated systems (50-53), the instrument is nearly
completely automated from the initial sample preparation through
data acquisition and analysis. In the stopped-flow system described
in this chapter, a slightly more conservative approach to the auto-
mation of the instrument was taken. In this system the sample

preparation has been left to the operator.

B. The Stopped-flow Mixing System

The stopped-flow mixing system will be described in this section

by first describing the overall operation of the system. The function

63

 

 

 

64

and construction of each of the major components of the system will
then be presented in more detail.

A block diagram of the stopped-flow mixing system is shown in
Figure 5-1. In a normal stopped-flow run (or push), the first oper-
ation is the filling of the drive syringes with the sample solutions.
This is accomplished by the pneumatic drive piston shown at the top
of Figure 5—1. The drive piston pulls the syringe plungers back,
while the check valves insure that the syringes are filled from the
sample solutions only. Next, the direction of the air pressure in
the drive piston is switched, energizing the drive syringes. No
solution flows at this time since the check valves keep the solutions
from returning to the sample containers and the 3-way slider valve
shown at the right of Figure 5-1 is in the closed position. The
flow begins when the position of the pneumatically controlled slider
valve is switched, connecting the Observation cell to the stopping
syringe. The flow continues to force the plunger of the stopping
syringe to move back until the end of the plunger strikes the stopping
block. At this time the solution flow stops rapidly and the reaction
occurring in the observation cell can then be monitored. After the
reaction is completed, the slider valve changes position, which
connects the stopping syringe to the waste container. Springs mounted
on the stopping syringe push the plunger into the syringe, and force
the solution out into the waste container.

The various parts of the system are connected together by 1/8
inch Teflon tubing (1.5 mm I.D.) with Altex chemically inert tubing

end fittings (Altex Scientific Inc., part no. 200-15). The pneumatic

 

65

   

STOPPING BLOCK

DRIVE
HSTON

 

     
     
 

DRIVE
SYRINGES

STOPPING
SYRINGE

  
   
     
 
 

CHECK
‘VALvEs

SAMPLE
SOLUTIONS

3-WAY‘

SLIDER
VALVE

      

Figure 5-1. Block Diagram of the Stopped—Flow Mixing System.

 

 

 

66

drive piston (Schrader Gold Medallion miniature pneumatic cylinder;

1 1/8 inch bore, 1 inch stroke) is controlled by a Scoville pneumatic
solenoid air valve (part no. 41001—1115). The drive and stopping
syringes were made by the Michigan State University Glass Shop by
sealing together precision bore pyrex tubing (Wilmad Glass Co. Inc.,
with 0.1736 and 0.1751 inch I.D.) and Altex glass connectors (Altex

Scientific Inc., part no. 200—28). The syringe plungers are 1 ml

gas—tight syringe plungers (Glenco Scientific Inc., part no. 19925—P—l).

The check valves (The Lee Co., part no. TCKA6201050A) were modified
to fit the Altex connector system by drilling out the ends of the
check valves. The 3-way slider valve (Altex Scientific Inc., part no.
201—04, 201-12, 201-13) is controlled by a solenoid air valve (Altex
Scientific Inc., part no. 903-00). Both this solenoid air valve and
the one controlling the drive piston can be opened and closed under
computer control through the use of solid state relays (Teledyne
Relays, part no. 611—3).

A cut—away view of the stopped—flow mixer and observation cell
is shown in Figure 5-2. All of the components, except for the tubing
end fittings and the quartz rods, were constructed by the Chemistry
Department Machine Shop. The light path is connected to the stainless
steel observation cell by 2 mm quartz rods with ground and polished
ends (Wilmad Glass Co., Inc.). The quartz rods are sealed to the
ends of the observation cell by the compression of teflon washers
by aluminum compression fittings. The Notz—type, Kel F mixer (54)
is also sealed to the observation cell by compression from the'Teflon

mixer adapter.

 

67

02:»:

H
00¢

N:_<....O

20.35328 g
n.
n”

_ner
THU

.300 soapmcfiomno 28 ~23: Eoumxm onmIOOmmoum .Num 6.22m:

xuoam ._._mu

HH—
«mImsS

20::

fig
_HUHHD

B «32.2

C —W
\
\ EE<O<
82.2.3 $52
oz“

023:: I II I
20:: WK! I IVfl/N

 

 

68

C. Stopping Syringe Velocity/Position Transducer

The velocity/position transducer developed in this research is
based upon a design originally presented by Holler (55). The trans-
ducer enables the position of the stopping syringe plunger to be
determined at any time during a stopped-flow run. In routine opera-
tion, the transducer allows the computer system to determine the
time at which the solution flow stops (the trigger point). This is
accomplished by determining when the stopping syringe plunger reaches
the position at which it strikes the stopping block. The determina—
tion of the trigger point is very important, since it is at this time
that the vidicon spectrophotometer must begin data acquisition. The
transducer can also be used to monitor the stopping syringe velocity
during a stopped-flow run. This is useful in diagnosing any irregu—
larities in the solution flow and also in determining the dead time
of the mixing system (see Chapter 6).

The transducer is shown in Figure 5-3, and consists of two parts.
A precision ruling with 50 lines per inch is mounted on the stopping
syringe plunger. The ruling consists of a photographic contact neg—
ative made from a ronchi ruling (Edmund Scientific Co., part no. 30
511). The second part of the transducer is an opto-interrupter
module (General Electric H13A1). As shown at the bottom of Figure 5—3,
the opto-interrupter consists of a light emitting diode and a photo—
transistor separated by a narrow slot. The opto-interrupter is mounted
so that the ruling passes through the slot. As the dark lines of the
ruling pass through the slot, they block part of the light falling on

the phototransistor. This causes the resistance of the phototransistor

 

69

  
    

OPTO-

  

STOPPING
SYRINGE

IO k.n.
— fi

220.11
OPTO INTERRUTER

Figure 5-3. Stopping Syringe Velocity/Position Transducer.

 

 

.‘K4 ,5.-

70

to vary slightly, which produces a voltage waveform at the photo-
transistor's collector. This waveform resembles a sine wave of
varying frequency with an amplitude of a few tenths of a volt. The
period of the sine wave is equal to the length of time required for
the stopping syringe to move from one line to the next (0.0508 cm
with the 50 lines/inch ruling).

Figure 5-4 shows the processing circuitry which measures the
period of the transducer output. The output of the phototransistor
(Tl) has a 1 to 2 V d.c. offset which is removed by the high-pass
filter formed by C1, R2 and R3. R2 and R3 are used to set the d.c.
Offset of approximately 0.1 V. Amplifier Al (RCA CA3l60) amplifies
this signal and produces an output waveform of about 3 Vpp with a
d.c. offset of 1.5 V. A square wave with indentical frequency is
created by passing the signal through comparator A2 (National Semi-
conductor LM311). Resistor R7 prevents false triggering of the
comparator by providing about 0.5 V of hysteresis. The square wave
is then sent to the microcomputer rack. The period of the square
wave is measured by two 16 bit counters (Intel C8253) with a 1 MHz time
base. The square wave causes the output of the flip—flop (CI) to
change state after each period. Since the opposite outputs of the
flip-flop are connected to the enables of the two counters, each
counter is on for exactly one period of the square wave. The micro—
computer acquires the period measurement by watching the Q output of
the flip-flop. When the output changes state the microcomputer reads
the count acquired by the counter which was enabled during the

previous period of the square wave. The microcomputer then clears

 

 

71

IK 15K

+5

RI 10K Cl

 

  
 
    
  

    

   

COUNTER

 
 

      

STOPPED-FLOW EN

1
CLK
COUNTER CL
- 2

Figure 5-4. Velocity/Position Transducer Processing Circuitry.

     
   
   

MICROCOMPUTER

 

 

72

the counter and begins watching for the next transition of the
flip-flop output. The result is an array of data which represents
the length of time (in microseconds) required for each pair of lines
on the ruling to pass the opto-interrupter. The position of the
syringe at any time can be independently determined by counting the

number of transitions of the flip-flop output.

 

 

 

CHAPTER 6

CHARACTERIZATION OF THE VIDICON, RAPID SCANNING,

STOPPED-FLOW SPECTROPHOTOMETER

A. Introduction

In order to realize the full potential of any instrument, its
operating characteristics must be well understood. Of primary in-
terest is the determination of the instrument's optimum operating
conditions and the identification of any features which limit its
effectiveness. This chapter examines some of the operating charac-
teristics of the vidicon, stopped-flow system which are important to
its application in stopped-flow kinetics. The first section presents
the characteristics which are specific to the vidicon spectrophotometer,
while the last section concentrates on the characteristics of the

stopped-flow solution handling system.

B. Characterization of the Vidicon Spectrophotometer

1. Introduction

The operating characteristics of the vidicon spectrophotometer
can be classified as either static or dynamic characteristics. Static
characteristics are observed when the vidicon detector is at steady
state (i.e. constant illumination and scan parameters). When the
system is perturbed from steady state by changing the illumination or
scan parameters, dynamic characteristics are observed. Dynamic

characteristics are often expressed as the time or number of scans

73

 

 

 

 

74

required to establish a new steady state condition.

Many of the basic static characteristics of the vidicon spectro—
photometer developed in our laboratory and of other similar spectro-
photometers have already been studied (36, 38, 43). The character—
ization studies presented in this section include new studies on
both static and dynamic characteristics of the vidicon spectropho—

tomet er .

2. Vertical Scan Frequency

The vertical scan frequency of the vidicon spectrophotometer,
which can be varied from 3.91 kHz to 100 kHz (see Chapter 4), has a
dramatic effect upon the intensities observed from the vidicon de-
tector. To study this effect, the vidicon target was illuminated with
a tungsten continuum and read out with 203 channels per frame at
varying frequencies. Figure 6—1 shows the relationship between the
relative intensity observed at the center of the target and the ver—
tical scan frequency. The rapid decrease in observed intensity at
higher frequencies is due primarily to two factors. Much of this
decrease is due to the decrease in exposure time at higher frequencies.
When the observed intensities are normalized to a constant exposure
time, a maximum response is observed at about 33 kHz as seen in Fig-
ure 6—2. The decrease in intensity at frequencies above and below
33 kHz is mainly due to incomplete scanning of the vidicon target.
Figure 6-3 shows the relative vertical drive amplitude at various
frequencies. The decrease in amplitude at low and high frequencies

results in incomplete scanning of the upper and lower edges of the

.‘-.‘. m eat-fur). - if

-&_'

 

 

>
.—
5
z
I“
p-
E
“I
>
Z
<
—I
III
8
0
Figure 6—1.

75

50
FREQUENCY (kHz)

Relative Intensity Versus Frequency.

 

IOO

 

 

>
y-
7:
z
I“
.—
E
O
IN
)—
S
8
8
O
U
0
Figure 6-2

76

 

so I00
FREQUENCY (ka)

Relative Intensity Corrected for Varying Exposure Time
Versus Frequency.

 

 

 

RELATIVE VERTICAL DRIVE AMPLITUDE

Figure 6-3.

77

so 100
FREQUENCY (kHz)

Relative Vertical Drive Amplitude Versus Frequency.

 

 

 

 

78

target and, therefore, a decrease in intensity.

In kinetics experiments it would often be very useful to be able
to change the vertical scan frequency during the course of a reaction.
For example, the early, rapid portion of a reaction could be observed
at a high scan frequency for good time resolution and then, as the
reaction slows down, the scan frequency could be decreased to maximize
the observed intensity. To study the effect of changes in frequency,
the vidicon target was exposed to constant illumination by placing a
red LED a few inches from the target. The horizontal resolution was
set at 255 channels per frame. The vidicon detector response was
observed at the center of the target for 10 frames. The frequency was
then changed and the response was observed for an additional 40 frames.
Figure 6—4 shows the results that were obtained. When the frequency
was increased from 25 kHz to 50 kHz the observed intensity decreased
as expected. Approximately 11 scans were required to bring the re—
sponse within 10% of the final 50 kHz value. When the frequency was
decreased from 50 kHz to 25 kHz, a tremendous spike was observed in
the vidicon response. This can be explained by remembering that at
50 kHz a significantly smaller portion of the target is being scanned.
The portions of the target which are not being scanned are in satura-
tion. When the frequency is reduced to 25 kHz, the portions which
were in saturation are again being scanned. This results in a sharp
increase in observed intensity which requires at least 30 scans to
bring it within 10% of the final 25 kHz value. The large number of
scans required to re-establish steady state severely limits the

usefulness of frequency changes during the course of a reaction.

3W2' Wm.

II 93%”;

4,-

 

>

p-

8

z

I“

o—

E

II

>

7.

<

-I

III

I

0

Figure 6-4.

79

IO 20 so 40
‘ SCAN NUMBER

The Effect of Instantaneous Changes in Vertical
Scan Frequency.

 

50

 

80

3. Resolution

It has already been established that the resolution of the
vidicon spectrophotometer is limited by the vidicon detector (43),
and it is determined by the number of channels per frame which are
used. The maximum resolution which can be obtained is limited by the
electron beam width of the vidicon detector. To study the resolution
of the vidicon detector, the target was illuminated with a tungsten
continuum. Figure 6-5 shows the relationship between the observed
intensity at a vertical scan frequency of 25 kHz and the number of
channels per frame. As the number of channels per frame is increased
the exposure time is also increased. Normally an increase in ex—
posure time results in an increase in observed intensity. However,
the observed intensity is relatively constant from about 65 channels/
frame to over 400 channels/frame. Previous studies (37, 38) have
interpreted this plateau region as an area in which the increase in
exposure time is balanced by the reduction in signal caused by the
partial overlap of adjacent channels. The extreme width of the
plateau region makes the ”balance” theory rather suspect.

A more reasonable explanation of the plateau region can be devel-
oped as follows. Since the exposure time is proportional to the
number of channels, the intensity is also proportional to the number
of channels. The number of channels per frame is also inversely
proportional to the effective target area scanned. Therefore, the
signal intensity increase due to the increased exposure time is
exactly canceled by the decrease in the effective target area per

channel. In order for this explanation to be valid, the charge

 

 

 

81

RELATIVE INTENSITY

 

o 500 IOOO
CHANN E LS/ FRAME

Figure 6-5. Relative Intensity Versus Resolution.

 

 

82

generated by photons falling upon the areas between the channels
must also be read out by the electron beam. This requires the
migration of charge from the saturated regions between the channels
into the channel areas themselves. Charge migration in regions of
saturation is a well known phenomemon which is normally referred to
as "blooming” or ”bleeding" (56). The implications can be more
easily understood if the data of Figure 6-5 are expressed in terms
of the observed intensity as a function of the channel separation.
The channel separation is determined by dividing the width of the
scanned portion of the vidicon target (12.8 mm wide) by the number
of channels. This relationship is shown in Figure 6—6. The rapid
decrease in intensity at small channel separations is due to the
increasing overlap of adjacent channels. A much less dramatic
decrease in intensity is also observed when the channel separation
becomes relatively large. This is due to the increasing probability
of electron—hole recombination when relatively extensive charge
migration is required. The intersection of the region of adjacent
channel overlap and the plateau region should give a good indication
of the electron beam width. This technique was used in an experiment
to determine the minimum electron beam width which can be obtained.
The results are shown in Figure 6-7. The optimum current through
the focus coil was determined to be approximately 39.2 mA. This is
very close to the recommended current of approximately 40 mA (57).
The minimum beam width, as determined by this technique, is approx-
imately 30 pm, which is significantly narrower than the 54 um beam

width previously reported for this instrument (43). It also differs

 

 

 

N

RELATIVE INTENSITY

Figure 6—6.

83

100
CHANNEL SEPARATION (pm)

Relative Intensity Versus Channel Separation.

 

200

 

 

 

3

40

BEAM WIDTH 00m)

30

20
37

Figure 6-7.

38 39 4O
FOCUS CURRENT (Inc)

Beam Width Versus Focus Coil Current.

4T

42

 

 

 

85

significantly from the 51 um (58) and 62 um (59) beam widths that
were reported for similar systems. This indicates that most of
the reading capability is concentrated in the center of the beam,
while the diffuse edges of the beam make the actual beam width much
larger. This new technique is nevertheless very important since it
provides a very reproducible indication of the relative beam width,
which is independent of the bandpass of the monochromator used in
the system.

At a given vertical scan frequency, the length of time required
to scan an entire frame depends on the number of channels per frame.
This means that a trade-off between resolution and frame scan rate
can be made. For example, it might be advantageous to select low
resolution, high frame scan rate conditions in the early, rapid part
of a reaction and then change to high resolution, low scan rate
conditions toward the end of the reaction. To investigate the feas—
ibility of this mode of operation, the vidicon target was illuminated
by placing a red LED a few inches from the target. The vertical scan
frequency was set at 25 kHz and the number of channels per frame was
changed from 127 to 511. The results of this experiment are shown
in Figure 6—8. 90% of the final steady state intensity was obtained
within 4 to 5 frames. Similar results were obtained when the number
of channels was changed from 511 to 127. The relatively rapid re—
establishment of equilibrium indicates that this mode of operation
would indeed be feasible, especially if the change in the resolution

were not as drastic as that shown in Figure 6—8.

 

30°

RELATIVE INTENSITY

2.5

Figure 6—8.

86

I27 CHANNELS

 

IO 20 30 4O 50
SCAN NUMBER

The Effect of Instantaneous Change in Resolution.

 

87

4. Electron Beam Current

The electron beam current is determined by the negative voltage
which is applied to Grid 1 in the vidicon tube. The beam current is
decreased by making the voltage on Grid 1 more negative. In the
1 current system the grid voltage can be varied from about +4 V to

-110 V. To determine the optimum beam current, the vidicon target
was illuminated by a tungsten continuum, and the resolution was set
at 511 channels per frame. The relationship between the intensity
observed at the center of the target and the Grid 1 voltage at var-
ious vertical scan frequencies is shown in Figure 6—9. The rapid
decrease in observed intensities at large grid voltages is due to

the beam current being reduced to a level where it no longer suffices
to read out all the charge stored on the target. As long as the
beam current is larger than the minimum required to read out the
charge, the observed intensity appears to be independent of beam
current. If care is taken in selecting a beam current which is

large enough to read out the largest signal expected, then the beam
current should never have to be changed. This is quite fortunate,
because even though the Datel AM301A high voltage amplifier used to
generate the grid voltage is supposed to have a slew rate of

100 V/usec, the actual observed slew rate of the grid voltage is only

15 V/sec.

5. Charge Integration

Charge integration is a signal enhancement technique which takes

advantage of the integrating character of the silicon vidicon. In

 

 

 

88

IO EH:

INTENSITY

RELATIVE

o W
O '50 - I00
GRID I VOLTAGE (V)

Figure 6-9. Relative Intensity Versus Grid 1 Voltage
at Various Frequencies.

 

 

89

charge integration the electron beam is turned off for a number of
scans and the photon-generated charge is allowed to accumulate. When
the accumulated charge is finally read out an increase in the signal
is observed which is linearly related to the integration time.

Since the S/N (signal—to-noise ratio) increases linearly with the
number of scans integrated, while the S/N increases only with the
square root of the number of scans averaged, charge integration can
be a very useful signal enhancement technique (38, 43, 60).

Signal enhancement by charge integration could be quite useful

in kinetics experiments. If a constant integration time was used
throughout a reaction, the signal level for the entire reaction could
be increased. Even greater benefit could be derived if the integra-
tion time could be varied during the course of a reaction. For
example, if the percent transmittance decreased dramatically during
the course of a reaction, the signal level could be increased by
using longer charge integration times. In order to determine if
integration times could be rapidly changed, the following experiment
was performed. The vidicon target was illuminated by a tungsten
continuum and the resulting charge was read out at a vertical scan
frequency of 25 kHz with a resolution of 101 channels per frame.
The number of frames during which the beam was turned off was varied
from 0 to 7. The results are shown in Figure 6—10. It is painfully
apparent that integration times cannot be rapidly changed. At least
15 to 20 frames were required to reach 90% of the final steady state
value. These results indicate that only constant integration times

can be used for kinetics applications.

 

 

 

90

RELATIVE INTENSITY

 

° 20 4o
SCAN NUMBER

Figure 6-10. Effect of Rapid Changes in Charge Integration at
Various Integration Levels.

 

 

 

 

91

6. Lag

As the electron beam travels across the vidicon target not all
of the charge stored there is read out. The percentage of the total
charge which is left behind is defined as the lag. Lag seriously
degrades the response time of the vidicon detector and, therefore,
it is the feature which most severely limits the applicability of
vidicon detectors to fast kinetics. This section contains the
results of a number of experiments which were designed to assess the
severity of the lag in the vidicon spectrophotometer and to determine
optimum conditions for limiting the effect of the lag.

Since it is virtually impossible to actually measure the amount
of charge left behind on the target, techniques must be developed
which provide an indication of the amount of charge left behind.

This is most commonly done by continously scanning the vidicon tar—
get under illumination, then rapidly removing the illumination and
observing the signal obtained after one or more complete frames have
been scanned. All of the lag experiments reported in this section
were performed using the same procedure. First the target was illum—
inated by placing a red LED a few inches from the target. A number
of frames were scanned with the LED turned on. The LED was then
turned off and the signal decay was observed until a constant signal

was obtained. The lag was then defined in the following manner:

 

 

 

92

where S1 is the signal measured one frame after the LED has been

turned off, S is the average signal obtained while the LED is on

ILL
and SD is the average dark signal. Figure 6-11 shows a sample lag
decay curve.

Comparison of the lag measured in different laboratories is
often extremely difficult due to the varying experimental conditions
under which the lag was measured (19, 27, 38). Often the conditions
are not mentioned at all. The most commonly published data are
measured at standard television scan rates. The lag decay curve
shown in Figure 6-11 was obtained with the scan parameters which most
closely approximate those used for standard television rates. The
vertical scan frequency was 15.87 kHz and the resolution was 272
channels per frame. All other parameters were optimized for min-
imum lag. The resulting lag was 12.0%. This is in good agreement
with one group (27) which reported a lag of 14% under similar con—
ditions. Another source (38) listed a lag of 7% after three read-
out frames, whereas our system shows only 4.9% lag after three read-
out frames.

The remainder of this section is devoted to studies of how the
various scan parameters affect the observed lag. The purpose of
these studies is to determine how the parameters can be varied to
minimize the lag. The reproducibility of the lag over periods of
up to 6 hours was found to be fairly good, with 1 to 5 percent
relative standard deviations being typical. When these same lag

measurements were made over a period of a few weeks, relative pre-

cisions as poor as 50% were observed. The poor reproducibility of

 

 

 

RELATIVE INTENSITY

Figure 6-11.

I0 20 30 4O
SCAN NUMBER

Signal Decay After Removal of Illumination.

   

 

 

94

the lag did not significantly affect the overall trends in the lag
which are reported in this section. Since the actual lag values are
not very reproducible, only the relative changes in the measured lag
are reported in each of the following experiments.

Lag is dependent upon the magnitude of the signal read out from
the vidicon target. Figure 6-12 shows the relationship between the
relative lag and the signal magnitude. This experiment was performed
at a vertical scan frequency of 25 kHz, a resolution of 255 channels
per frame and a Grid 1 voltage of -23 V. Normally, lag tends to
improve at higher signal levels. However, as the signal level ap-
proaches saturation, the lag increases rapidly. The increase in lag
at low signal levels is due primarily to the target potential being
discharged only slightly, resulting in a low potential difference
between the cathode and the target. This results in a lowered
beam landing efficiency and increased lag.

The resolution which is selected also appears to have a sig-
nificant effect upon the observed lag. Figure 6-13 shows the re-
lationship between relative lag and resolution at a vertical scan
frequency of 25 kHz and a Grid 1 voltage of -23 V. The decrease in
lag at higher resolutions is due primarily to the decrease in the
channel separation. Decreasing the channel separation decreases the
saturated area between the channels, therefore decreasing the rel-
ative lag.

The vertical scan frequency has a dramatic effect upon the ob-
served lag. Figure 6—14 shows the frequency dependence of the rel-

ative lag at a resolution of 255 channels per frame and a Grid 1

V

 

 

RELATIVE LAG

Figure 6-12.

95

    

SATURATION

102 103

RELATIVE INTENSITY

Relative Lag Versus Relative Intensity.

 

"‘I

 

 

0
<
-l
W
Z
)—
<
-J
“I
8
O
0
Figure 6-13.

96

300 600
CHANNELS/ FRAME

Relative Lag Versus Resolution (Frequency = 25 kHz).

 

97

RELATIVE LAG

 

0 so IOO
FREQUENCY (kHz)

Figure 6-14. Relative Lag Versus Vertical Scan Frequency (Resolu-
tion = 255 Channels/Frame).

 

 

98

voltage of -23 V. The increase in the lag at higher vertical scan
frequencies is due to a number of factors. Since the observed
intensity decreases rapidly at higher frequencies, the intensity
dependence of the lag results in an increase in the observed lag.
The beam residency time decreases with increasing frequency, also
causing an increase in the lag. Another cause of the increased
lag at high frequencies is the incomplete scanning of the upper and
lower edges of the target. As shown in Figure 6-3 the amplitude of
the vertical scan varies significantly with frequency. The edges of
the target are not scanned because it would be impractical to
increase the electron beam; rectangular scanning pattern to cover
the entire circular, active region of the target. Since the edges
are in saturation, the observed lag is increased in the scanned
portions of the target near the upper and lower edges. A limited
amount of vertical resolution is possible in the vidicon spectropho-
tometer by varying the electronic delay and integration times in the
amplifier chain (43). Figure 6—15 shows the dependence of the lag
on the vertical target position at a resolution of 255 channels per
frame. The vertical scan frequency was set to 25 kHz at which the
vertical drive amplitude is near maximum. The edges of the target
do show a significantly higher lag than the center of the target.
(The very high lag at the left edge of the graph is due to the
inclusion of the vertical retrace.)

The last parameter which was found to significantly affect lag
is the beam current. Figure 6-16 shows the relationship between the

relative lag and the voltage applied to Grid 1 (the higher the

 

99

RELATIVE LAG

0 RELATIVE VERTICAL POSITION I

Figure 6—15. Relative Lag Versus Relative Vertical Target Position.

 

 

RELATIVE LAG

Figure 6-16.

100

-40
GRID I VOLTAGE (V)

Relative Lag Versus Grid 1 Voltage.

 

'80

 

 

101

voltage, the lower the beam current). This experiment was performed
at a vertical scan frequency of 25 kHz and a resolution of 101
channels per frame. The lag is seen to reach a broad minimum at a
Grid 1 voltage between about -10 V and —30 V. The decrease in lag
at Grid 1 voltages above about ~60 V is quite unexpected. Above
-60 V the beam current is no longer large enough to completely read
out all the stored charge and the observed intensity decreases
rapidly (see Figure 6—9). Both of these factors should result in a
dramatic increase in the lag. No matter what the reason is for the
observed decrease in lag, the signal degradation at these voltages
makes operation under these conditions impractical.

In this section it has been demonstrated that although lag is
a serious problem in the vidicon spectrophotometer, its effects can
be minimized by the careful selection of the scanning conditions.
In general, the lag can be decreased by setting the Grid 1 voltage
to about —18 V, decreasing the vertical scan frequency and/or
increasing the number of channels per scan. Adjustment of the ex—
perimental conditions so that the signal intensity is about 50%

of saturation will also decrease the observed lag.

C. Characteristics of the Stopped-Flow Mixing System

The purpose of the stopped—flow mixing system is to mix rapidly
together two solutions and deliver them to an observation cell where
any subsequent reactions can be monitored by the vidicon spectro—
photometer. In order to be able to investigate relatively rapid

reactions the mixing system must be capable of mixing the two

   

51:33:9quan 531m: :52 E“ 03- .moflr. am; 2:11» SEVILLE

  
 

Was. {he '21-?! ['1' 2: :1g.xen~ mantel 145-nu! n- = ---:-:n 4mm win V"

 

 

102

solutions both rapidly and reproducibly. The mixing characteristics
and other important characteristics of the stopped-flow mixing system
developed in this work will be presented in this section.

The stopped—flow mixing characteristics are highly dependent
upon the solution flow velocity. In the current system, the solution
flow velocity is monitored by a transducer which consists of a
transparent ruling (with 50 lines/inch) mounted on the stopping
syringe, and an opto—interrupter module. (A complete description of
this transducer can be found in Chapter 5.) The transducer allows
the measurement of the length of time in usec, between the lines as
they pass the opto-interrupter. Figure 6-17 shows the flow velocity
output for 5 separate stopped—flow pushes. Line number 47 represents
the point at which the flow comes to a stop. The erratic output
observed after this line, is due to mechanical vibrations in the flow
transducer. (Further experiments showed that no vibration or cavi—
tation effects occured in the flow cell.) Curve A of Figure 6—18
shows the distance traveled by the stopping syringe during a typical
run. The entire distance is covered in about 0.03 s. Curve B of
Figure 6—18 shows the syringe velocity profile. The velocity pro-
file was calculated from the lower curve by performing a linear
least squares fit to 4 points at a time. The slope of this tangent
line was then taken as the velocity at the mid-point.

The single most important characteristic of a stopped-flow
mixing system is its dead time. The dead time can be defined as the
length of time between the theoretical start of a reaction and the

earliest time at which the reaction can be monitored. In a properly

 

TIME hnuc)

Figure 6-17.

103

30
LINE NUMBER

Flow Velocity Output for 5 Separate Runs.

 

60

 

 

I00

0
O

O
O

70

VELOCITY (cm/soc)

60

2.0

I.5

I.0

DHIANCE km)

I5

Figure 6-18.

104

25 40
TIME (mm)

A - Syringe Distance Traveled Versus Time.
B - Syringe Velocity Versus Time.

50

 

 

105

operating stopped—flow, dead time is determined primarily by the
length of time required for the solution to flow from the point of
first mixing to the center of the Observation cell. The actual dead
time can be longer than this due to slow mixing, cavitation and
other problems; however these problems should be minimal in a prop-
erly operating system.

For the dead time to be reproducible, the solution flow rate must
be reproducible during the final portion of the flow. Figure 6—17
shows that the flow profile is quite reproducible from about line
number 26 to the end of the flow. This corresponds to about the last
12 msec of the solution flow. The average syringe velocity over this
section is 87.8 i 0.9 cm/s, which yields a solution flow rate of
13.6 i 0.1 ml/s with the current stopping syringe. This means that
0.163 t 0.002 ml of solution flow through the cell during this time.
Since the combined volume of the mixing chamber and the observation
cell is 0.134 i 0.0007 ml, the entire observation cell should be
completely flushed during the most reproducible portion of the flow.
This should result in a very reproducible dead time.

The dead time of the system was investigated by two methods.
The length of time required for the solution to flow from the point
of first mixing to the center of the observation cell can be deter-

mined by the following equation:

 

where t is the dead time, V is the dead volume (the volume between

d d

   
 

-_—.- mushy-IRE

...-. -l..- . .--'-""-
:.'I..I.I: ...'. :a ..' I f - ... .' . . . I . . II' ""I if)",

 

 

106

the point of first mixing and the center of the cell), vS is the
syringe velocity and As is the cross sectional area of the syringe.
The value of vs used should be the average syringe velocity just
prior to the stopping of the flow. It should be measured over a
time approximately equal to the dead time. The average syringe
velocity over this period was 85.4 i 1.1 cm/s. Calculations based
on the cell specifications yielded a value of 0.103 t 0.005 ml for
Vd and 0.155 r 0.002 cm2 for As' This results in a dead time of
7.7 i 0.4 msec. The uncertainty in the dead time is primarily due
to the uncertainty in the dead volume. This uncertainty could be
improved by determining a more precise value for Vd'

The above method for calculating the dead time does not include
any effects from mixing, cavitation and other problems which can
affect the performance of the stopped-flow mixing system. The
traditional method for determining dead times is the extrapolation
technique (61), whereby the total dead time including mixing and
other problems is measured. In this technique the absorbance of a
chemical reaction in the stopped—flow is recorded with time, normally
on an oscilloscope. The experimental curve is then extrapolated back
to zero absorbance. The dead time is the time between the extrapol-
ated point of zero absorbance and the time at which the extrapolated
and experimental curves coincide. This technique, with a photomulti-
plier tube employed as a detector was also used to determine the
dead time. The dead time was found to be 8 i l msec. The large

uncertainty is due to the difficulty of performing the extrapolations.

The difference between the two dead time values is not significant.

 

107

Two solutions were used to determine whether equal volumes are
delivered from the two drive syringes. One syringe was filled with
0.0017 M K2Cr207 in 0.2 M H2804 and the other syringe was filled with
0.2 M H2804. The absorbance of this mixture at 435 nm was determined,
the solutions switched and the new absorbance determined. The ab-
sorbance which was found when the KZCrZO7 solution was in the left
syringe was found to be larger than the absorbance when it was in
the right syringe. The ratio of the absorbances was found to be
1.03 i 0.02. The ratio of the volumes of the syringes is also equal
to this value. Whenever quantitative measurements are performed
with this stopped-flow apparatus, the above difference in syringe
volumes must be compensated for in the calculations.

The pathlength of the stopped-flow observation cell was designed
to be 2.0 cm. The actual pathlength was determined by comparing the
absorbance of a solution in a cuvette having a calibrated pathlength
and the same solution in the stopped—flow cell. The pathlength of
the cell was found to be 1.96 i 0.04 cm.

When solutions to be used in stopped-flow experiments are
changed, the drive syringes must be flushed a number of times. The
number of times that the syringes must be flushed to ensure complete
removal of the old solutions varies with the type of solutions used.
In most cases at least 30 to 40 flushes of the drive syringes are

required. Since about 0.42 ml of solution is required for each

flush, the total amount of solution required is about 12 to 17 ml.

           

- - '4.

:4

. “a

h: :_~_
. .' _

 

108

D. The Monitoring of a Relatively Rapid Reaction

The capabilities of the entire system were determined by
monitoring a relatively rapid reaction which has a large change in
absorbance. The large change in absorbance causes a drastic change
in the amount of light falling on the vidicon detector. If the lag
of the vidicon is going to cause any errors when fast reactions are
followed, these errors should show up in this type of application.
The reaction which was selected was the iron—thiocyanate reaction
(62), carried out with an excess of iron. Under these conditions

the equilibrium equation involved is:

k

i 2-
<— 4
k2

FeSO + + SCN' Fe(SCN)2+ + so

4

and the rate law is (62):

d [Fe(SCN)2+] = k1 [FeSO4+] [SCN’] — k2 [Fe(SCN)2+].
““EF‘"‘"

A solution of 0.0800 M Fe3+ was mixed with a 0.0102 M SCN_

solution. The reaction was carried out in 0.02 M H2504. The ver-
tical scan frequency for the vidicon spectrophotometer was set to

25 kHz, 102 vertical channels per scan were acquired, and the Grid 1
voltage was set to —21.5 V. The resulting kinetics data were evalua-
ted by using a curve-fitting routine called KINFIT-4 (63) to deter-
mine the rate constant. This was done at two different wavelengths.
At 583 nm the average forward rate constant (k) was found to be

96 M-1 sec-l, while at 658 nm a value of 196 M.1 sec_1 was determined.

 

 

109

Previously reported values for this rate constant are 208 M-1 sec-1
(62), 228 M_1 sec"1 (54) and 212 M-1 sec_1 (36). The rate constant
1 —l

of 196 M_ sec is in about the right range; however, the value
determined at 583 nm is far too low. The reason for this is lag
problems in the vidicon detector. At 583 nm the light intensity from
the tungsten source is much lower than at 658 nm. Also, under the
conditions of this experiment the absorbance.change at 583 nm is
greater (total absorbance change of 1.68) than at 658 nm (total
absorbance change of 0.31). Both of these factors result in a
dramatic increase in the detector lag at 583 nm.

The results of this study indicate that although the stopped-
flow system has a far slower response time than a standard single
wavelength instrument, it can still be used to monitor fairly rapid
reactions provided care is taken in selecting the conditions under
which the reactions are studied. The light intensity falling on the
vidicon detector should be kept as high as possible. Also the

absorbance changes which are studied must be kept small.

 

CHAPTER 7
AN INVESTIGATION OF THE BIURET REACTION
A. Background

The biuret reaction is the basis for the standard clinical chem-
istry technique for determining the total protein content of blood
serum (64), which can be a very useful aid in the diagnosis of a
number of diseases. The normal total serum protein level in a
healthy adult is between 6.0 and 8.5 g/100 ml (64). A number of
diseases show symptoms of hypoproteinemia, in which the total serum
protein level drops below 6.0 g/100 ml. An elevation in total serum
protein level is encountered in multiple myeloma and in any disease
which results in dehydration.

The biuret reaction is a complexation reaction between the cupric
ion (Cu2+) and the peptide bonds of the protein which occurs in
alkaline solution. Although the exact structure of the complex is

unknown, the Cu2+ is believed to interact with the carbonyl (-C=O)

and imine (=N-H) groups of the peptide bond. The Cu2+ is normally
linked to between four and six peptide bonds (64).

When the biuret reaction was first introduced, the procedure
was relatively complex and the reagents were unstable, resulting
in poor reproducibility. Kingsley (65) simplified the technique and
made it reproducible enough for routine clinical use. Since that time
a number of variations and improvements on the basic technique have

appeared (66, 67). In the standard biuret technique, the reaction is

allowed to go to completion and then the absorbance of the complex is

110

 

 

111

measured at 550 nm. This technique tends to be rather slow since
the reaction, though very rapid over the first few seconds, does not
go to completion for at least 20 to 30 minutes.

Koch, et al. (68) reported a kinetic method for determining
total serum protein using the biuret reaction. In this study a
centrifugal analyzer, the CentrifiChem Fast Analyzer (Union Carbide
Corp.), was used. They found that reasonable results could be ob—
tained by this method in as little as 18 5. They reported a precision
of 2.9% for replicate runs and a 3.4% precision for day-to-day results.
One limitation of this technique is that assayed human blood serum
has to be used for standardizations. In the standard biuret technique,
assayed bovine serum albumin can be used as a standard. Bovine
albumin was found to react much more rapidly in the early part of the
reaction than human serum. They also found that bovine albumin could
be used as a standard if the analysis time was lengthened to at least
93 s. Savory, et a1. (69) tried to use this kinetic method, but
found it to be relatively imprecise and subject to analysis errors in
patients with certain diseases.

Recently, Wai-Tak Law and S.R. Crouch of the Michigan State Un-
iversity Department of Chemistry have been investigating a more rapid
rate method for total serum protein using the biuret reaction in a
stopped-flow spectrophotometer (70). The reaction is monitored by
following the growth of the absorbance of the copper—protein complex
at 550 nm. Good results have been Obtained in as little as 0.1 s.

The average relative standard deviation for replicate runs was 3.1%;

when the analysis time was increased to 10 s the precision improved

112

to 0.83%. The experimental rate curve for this reaction is very
complex. After an induction period of about 20 msec, the reaction
rate increases rapidly until it reaches a peak at just over 100 msec.
After this peak the rate falls Off irregularly over a long time per-
iod. Most of the reaction occurs within the first 30 s, but the
reaction doesn't go to completion until after 20 to 30 minutes.
Just as Koch had discovered earlier (68), bovine protein samples
react more rapidly than human protein samples and therefore cannot
be used as standards.

To gain a better understanding of the biuret reaction, Wai-Tak
Law and I decided to investigate the reaction by using the vidicon
rapid scanning stopped-flow spectrophotometer. The first goal of
this investigation was to study the very unusual rate behavior of the
biuret reaction with human serum protein. The second goal was to
study the variations in the reaction rates observed in the biuret
reaction with bovine serum proteins and human serum protein. We
felt that these studies would help to characterize the new rate method
and also serve as a basis for determining future directions in

characterizing the new rate method.

B. Experimental

The biuret reagent used in this work was prepared by dissolving
24.0 g of sodium hydroxide, 4.0 g of sodium potassium tartrate and
0.5 g of potassium iodide in approximately 800 ml of water. When
the reagents were completely dissolved, 1.2 g of cupric sulfate was

added and the solution diluted to 1 liter. After standing for a few

    

L80 ., ‘. .>.j':_._ _‘
nfi-Mgn ”Mural—n, [1.0 2153:1111: :-

 

. '- -: I ' 1' "'T III‘EII’I- I.‘

 

113

hours, the solution was filtered. The sodium potassium tartrate is
used to complex the copper and keep it in solution. The biuret blank
reagent was prepared in the same way as the biuret reagent except that
no cupric sulfate was added.

The various serum protein samples were prepared by making an
approximately 1 to 30 dilution of the serum samples with 1.9% saline
solution. The human serum samples were Moni—Trol I (Dade Division
American Hospital Supply Corp.). The bovine serum protein samples
were bovine serum albumin and bovine serum globulin (Sigma Chemical
Co.).

The following vidicon scan parameters were used throughout this
study: the vertical scan frequency was set at 25 kHz with 102 chan-
nels per scan frame. This results in a scan time of 4.08 msec/frame.
The beam current was set to about —23 V. No charge integration was
performed.

The following general procedure was used for all of the stopped-
flow data presented in this chapter: the stopped—flow system was
first flushed approximately 30 times with the appropriate serum sample
and the biuret blank reagent. With the shutter open, 16 scans from
the vidicon detector were averaged to yield the reference (100%
transmittance) spectrum. The shutter was then closed and another 16
scans were averaged to form the dark current spectrum. Next, the
blank reagent was replaced by the biuret reagent and the stopped-flow
system flushed another 30 times. The system was always flushed l
or 2 more times immediately before each stopped—flow run. The reaction

was monitored by collecting successive spectra, each of which was the

 

114

average of 4 frames. A total of 25 frames were skipped between the
acquisition of successive spectra. This results in the acquisition

of one spectrum every 118.32 msec. A total of 40 spectra are acquired
during each push. To gain better precision, four separate pushes

were always averaged together.

The reaction was initially investigated by centering the wave—
length window of the vidicon detector at the absorbance maximum of
the copper-protein complex (550 nm). It was soon discovered that an
isobestic point existed at about 690 nm. The wavelength window was
then moved to cover the region from about 550 nm to 850 nm.

Figure 7-1 shows selected spectra from the biuret reaction with human
serum protein. Two isobestic points are observed, at about 690 nm
and 802 nm. The sharp irregularities that occur periodically in

the spectra (e.g. 640 nm, 730 nm, 810 nm, etc.) are due to slight
shifts in the vidicon's electron beam retrace pattern which are not
completely corrected for when the absorbances are calculated.

Figure 7—2 and 7—3 show the absorbance change with time which occurs
at wavelengths of 575 nm, 752 nm and 839 nm (corresponding to the
vertical lines in Figure 7-1).

Figure 7-4 shows selected spectra from the biuret reaction with
bovine serum albumin. In this case only a single isobestic point was
found, at 690 nm. For this reason, the reaction profile has a
dramatically different appearance from the human serum. Figure 7—5
shows the absorbance changes which occur at 586 nm (a) and 779 nm (b).

The final serum sample investigated with the biuret reaction was

bovine serum globulin; the results are shown in Figure 7—6. The

     

 

..- ’_ ’. 4‘ , z .: ‘. . ‘ — ,. . “1
.tnaEthrA}; iM.2;-m,_m-mnnsim suntan my in . ("a
, _ , . may .5 -- -

115

015

ABSORBANCE

 

550 700 350
WAVELENGTH (nm)

Figure 7-1. Absorption Spectra at Various Times During the
Biuret Reaction with Human Serum (2.2 g/L).

 

 

Figure 7-2.

    

ABSORIANCI

 

0.050

ABSORIANCE

 

TIME (I)

Absorbance Versus Time for the Biuret Reaction with
Human Serum. (a) 575 nm; (b) 752 nm.

 

 

 

 

ABSORBANCE

Figure 7-3.

  

I 2 a 4 5
TIME (s)

Absorbance (839 nm) Versus Time for the Biuret Reaction
with Human Serum.

 

 

0.4

ABSOR BANCE

550

Figure 7-4.

118

 

700 350
wAvnENGIH(nm)

Absorption Spectra at Various Times During the
Biuret Reaction with Bovine Serum Albumin

(3.3 g/L).

 

 

 

Figure 7-5.

 

In
9
z
<
I
I
O
a
I
d

a
I
0
o I 2 3 4 5
TIME (II
0.05
b
3
z
<
2
o
In
G
<
O
o I 2 3 4 5
TIME (I)

Absorbance Versus Time for the Biuret Reaction with
Bovine Serum Albumin. (a) 586 nm; (b) 779 nm.

 

 

   
     

 

.c.......

in! .3.
’1 . l.

 

 

 

1.67l

 

0.04 s

0.4“

0.25 s

ABSORBANCE

0.13 I

550

Figure 7-6.

 

700 ISO
WAVELENGTH (nm)

Absorption-Spectra at Various Times During the
Biuret Reaction with Bovine Serum Globulin

(3.3 g/L).

 

 

 

 

121

reaction profile is quite similar to the bovine serum albumin reaction.
The single isobestic point is shifted slightly to 676 nm. The ab-
sorbance changes which occur at 578 nm (a) and 775 nm (b) are shown

in Figure 7—7.
C. Analysis of the Biuret Reaction Spectra

The absorbance of the copper—protein complex should increase
during the course of the reaction. The existence of isobestic points
and wavelength regions over which the absorbance decreases with time
indicates the presence of at least one interfering species which
absorbs over this wavelength region. No interference had been pre-
viously reported, and this discovery is the most important result
of the current investigation.

To determine the extent of this interference the identity of
the new species and the spectral regions over which all of the species
present during the reaction absorb, must be determined. Figure 7-8
shows the absorption spectrum of the copper complex with human serum
protein. It was obtained by letting the biuret reaction go to com—
pletion (about 25 min.). A comparison of this spectrum and the
spectra in Figure 7-1 indicates that there is only a single inter-
fering species which absorbs over this spectral region. The increase
in absorbance with time at wavelengths below 690 nm and above 800 nm
is due to the copper-protein complex.

The cupric tartrate complex present in the biuret reagent was
identified as the interfering species. Figure 7-9 shows the

absorption spectrum of the cupric tartrate complex, with the biuret

 

 

 

Figure 7-7.

 

ABSORIANCE

AISORIANCE

 

"M! (II

Absorbance Versus Time for the Biuret Reaction with
Bovine Serum Globulin. (a) 578 nm; (b) 775 nm.

 

 

123

ABSORBANCE

 

550 700 :50
WAVELENGTH (nm)

Figure 7-8. Absorption Spectrum of the Copper-Protein Complex from

the Biuret Reaction with Human Serum After About
25 Minutes.

 

 

 

124

0.1

AB SORBANCE

 

550 700 850
WAVE LENGTH (nm)

Figure 7-9. Absorption Spectrum of the Biuret Reagent.

 

 

 

125

blank used as a reference. The cupric tartrate complex shows an
absorption maximum at about 660 nm and it still has significant
absorbance at 550 nm. The interference from the cupric tartrate
complex causes a decrease in the observed rate of absorbance change
for the copper-protein complex, and accounts for the wavelength shift
of the absorption maximum with time.

Since the cupric tartrate complex does not absorb above 800 nm,
it does not explain the relatively large absorbance changes observed
for both bovine serum samples between 800 nm and 850 nm (see Figure
7-4 and 7-6). Figure 7-10 shows the absorbance spectrum of the
copper—protein complex from bovine serum albumin after about 25 min-
utes of reaction. Again, there is essentially no absorbance above
800 nm. Both of the bovine serum samples also showed slight shifts
in the absorption maxima with time; however the shifts were far less
dramatic than those found in the human serum samples. These obser-
vations indicate that the absorbance from interfering species in the
biuret reaction with the bovine serum samples has a maximum at a
larger wavelength than the cupric tartrate complex. One possible
explanation for this is that a new copper complex of unknown composi-
tion is formed almost immediately after the bovine samples and the
biuret reagent are mixed. The decrease in the concentration of the
cupric tartrate complex reduces its interference at 550 nm. Also,
since the absorption maximum of the new interfering species occurs
at longer wavelengths that the cupric tartrate complex, it is likely
that its absorbance at 550 nm is significantly less than the

absorbance of the cupric tartrate complex. This lessened interference

 

 

126

(L3

03

ABSORBANCE

0.1

 

550 70° :50
WAVE LENGTH (nm)

Figure 7—10. Absorption Spectrum of the Copper-Protein Complex from
the Biuret Reaction with Bovine Serum Albumin After
About 25 Minutes.

 

 

127

in the bovine samples at 550 nm could account for the higher
rate of absorbance change observed in these samples.

Two important discoveries have resulted from this investigation
of the biuret reaction. The first major result was the discovery of
a chemical species in the reaction which absorbed over the same
wavelength region as the copper—protein complex. This discovery
enabled the new species to be identified as the cupric tartrate com-
plex and the extent of its interference on the observed rate of ab-
sorbance change for the copper—protein complex could be determined.
The second important finding was the discovery that the interfering
species in the biuret reaction with bovine serum samples are differ—
ent, and that the absorption maximum of the interference lies at
longer wavelengths than that of the cupric tartrate complex. On the
basis of these results it is possible to suggest further studies
which could help characterize the biuret reaction. The rate curve
should be investigated at wavelengths below 550 nm. In this region
the absorbance of the cupric tartrate complex should be relatively
small. This study would help to quantify the effect of the cupric
tartrate complex on the reaction rate at 550 nm. Also, comparisons
with the bovine serum samples would indicate if the differences in
the rates were due to the cupric tartrate complex interference. (At
the present time this study could not be performed on the vidicon
system, since a problem with the monochromator dramatically reduced
the sensitivity of the system below 550 nm.)

Another important study would be to identify the interfering

copper complex in the biuret reaction with the bovine serum samples.

 

128

Once identified, the spectral region over which it absorbs could be
determined and its effect on the rate curve at 550 nm could be more
precisely determined.

One critical question which has not been addressed at all is
whether the many different proteins found in human serum samples
react with the biuret reagent at different rates. The irregular
decrease in the rate curve for the biuret reaction could be due to
these varying rates. This is a very important problem, since cer—
tain diseases cause increases or decreases in only a single type
of protein (64), and kinetic analysis may be able to differentiate

the different types.

 

CHAPTER 8

CONVERS-M
A USER-DEFINABLE LANGUAGE

FOR COMPUTERIZED INSTRUMENTATION

A. The Role of Software in Computerized Instrumentation

As the use of computers in the sciences becomes more and more
common, their role is rapidly shifting. Throughout most of their
history computers have been large and expensive, forcing them to be
used primarily as remote, off—line data processors. In this type of
application the principal task of the computer was to perform data
file manipulations and complex mathematical processing of already
existing data. As computers became smaller and less expensive, ex—
amples of computers also being used for data acquisition and real—
time instrumental control became more common. With the recent intro-
duction of small and inexpensive microprocessors, computers are now
being found in a rapidly expanding variety of scientific instrumenta—
tion. This has caused a dramatic shift in the types of tasks that a
computer is commonly required to perform. Sequencing, formatting,
self—checking, operator interaction, monitoring, communications, and
many other modes of automation are increasingly common in dedicated
systems.

Since most of the common computer languages were written long
before the rapid increase in computerized instrumentation, a re—
evaluation in terms of current applicability is in order. It is

necessary to set up some criteria by which to compare the various

129

--. wm'mi‘m am

'Ié-ié‘awfi’mi“ I I."

fi:-"."0 IiJi'.‘ '3‘. -l'

 

 

130

computer languages. A large percentage of the most important criteria
can be summed up with the conclusion that the computer language must
enable the programmer to create the required software as rapidly as
possible. The speed with which the programmer completes the necessary
software is primarily determined by the number of lines of code that
he is required to write. Therefore, the language must be versatile
enough and of a high enough level that most tasks can be accomplished
with only a few lines of code. Also, it is important that there be a
library of commonly used routines (e.g. least squares fit, etc.) so
that the programmer need not create them all himself. Obviously, if
the completed software does not execute rapidly enough to accomplish
the desired task, than it does not matter how few lines of code were
necessary to create the software. Therefore, the execution speed of
the language is very important. It is always important that the
completed software take up as little computer memory as possible

and this is especially critical when the computerized instrument is

to be mass produced. The above criteria (program level, routine
libraries, execution time and memory requirements) will be applied to
the evaluation of the languages most commonly used in computerized
instrumentation.

The most basic type of computer language is machine code, which
is the numeric code used directly by the computer and therefore is
the only code upon which the computer actually operates. All other
computer languages must first have their commands translated into
machine code before they can be executed. Routine computer pro-

gramming is seldom done in machine code for many reasons. Since the

 

131

language is entirely numeric, programs in machine code are extremely
lengthy and very tedious to write. Also, each type of computer has
its own specific machine code, which makes the development of stan—
dard routine libraries impossible.

The programming level of assembly language is only slightly
higher than that of machine code. The real advantage of assembly
language lies in its use of English mnemonics to replace the numeric
codes of machine language. Before any assembly language program can
be used, all of the English mnemonics must be converted to their
machine code equivalents. This is most commonly done through the use
of a program called an ”assembler”. Assembly language is extensively
used in computerized instrumentation because it is the most common
computer language now in general use which can be applied to instru-
mental control and data acquisition. Well written assembly language
programs require less memory and execute faster than programs written
in any other type of code; however, it shares with machine code a
number of distinct disadvantages. Since it is a low level language,
even simple tasks require many lines of assembly code. Each type of
computer has its own assembly language, resulting in the establish-
ment of very few common routine libraries. The mathematical cap—
abilities of different assembly languages vary with the type of
computer which is used, but at best they include only integer arith—
metic, with multiply and divide capabilities being rather uncommon.

FORTRAN and BASIC are the two high level computer languages
which are most Often used in computerized instrumentation. Their

command structures enable them to handle file manipulations and

 

132

complex mathematics that would be extremely difficult and lengthy if
done with assembly language. As with any high level language, a
single line of FORTRAN or BASIC code can perform operations which
would require many lines of code if written in assembly language.
Before any FORTRAN or BASIC program can be executed, it must be
converted into machine code. All FORTRAN languages and a few types
of BASIC accomplish this conversion through the use of a "compiler”
program. The compiler creates an executable machine language program
which is equivalent to the original FORTRAN or BASIC program. Most
types of BASIC use an "interpreter" program instead of a compiler.
The lines of BASIC code are interpreted and executed one line at a
time. Since most computers have compiler and/or interpreter programs
available, FORTRAN and BASIC programs can be run on almost any type
of computer. BASIC, and especially FORTRAN, have been used long enough
that extensive libraries or common routines are in existence. Un-
fortunately, FORTRAN, and especially BASIC, routines tend to be quite
slow during actual execution and their memory requirements are high.
In addition, standard FORTRAN and BASIC contain no commands which
could be used for data acquisition or instrumental control. The way
in which FORTRAN and BASIC are adapted for such use is by programming
in assembly language. Assembly language subroutines are written for
the various modes of instrumental control and data acquisition.
These routines are then linked together with the FORTRAN or BASIC
routines which handle the data processing.

It is obvious that there is no single computer language currently

in general use which can easily solve all the problems normally found

 

0

0

79:61 7" '_

2.

.11... 4-2.,ril.M <s§§.. h3?—-.§gf, .7

    

133

in computerized instrumentation. There are very few languages which
have any high level "instrumental" commands for instrumental control
or data acquisition. Furthermore, it would probably be impossible
to create a computer language with a set of "instrumental" commands
versatile enough to handle all the extremely varied tasks found in
the different types of computerized instrumentation.

There is a possible solution to this dilemma. Recently a new
class of computer language has begun to develop. CONVERS-M, a lan-
guage developed at Michigan State University, is one version of this
new type of computer language. The basic concepts of CONVERS—M were
based on the language CONVERS developed at the University of Arizona
(71). CONVERS itself was based on FORTH (72), the most widely dis-
tributed language of this type. CONVERS-M does not have the rigid
command structure characteristic of most other languages. In fact,
its major advantage lies in the ease with which new commands can be
created and used within its structure. The result is a language which
is not restricted to a small range of programming levels but, as shown
in Figure 8-1, can be written on any level from machine code to
commands more powerful than FORTRAN commands. This makes it possible
to create ”instrumental” commands tailored to the specific application.
The execution speed of CONVERS-M depends upon the level at which it
is written. For most instrumental control applications the speed is
the same as an equivalent machine code routine. Its major disadvan-
tage is that it is a new language and therefore software libraries
have not yet been established.

Table 8-1 summarizes the advantages and disadvantages of the

      
 

. A

nl'amuoit was: is:.-=-I-~--~I:'£a:m!.-..:- nah TIE album 01' Mini-yd .
_m_-I-.-, Hum-"T 'I- --- '.' '.‘-".’ *ITMOI’IEB

'.."‘-: I ' -: .u-- ' "I "- - _ ..n'.‘ .‘... -.,._-r.-,

 

134

D MACHINE CODE

D ASSEMBLY LANGUAGE
BASIC

FORTRAN
CONVERS-M

 

SIMPLE COMPLEX
PROGRAMMING LEVEL

Figure 8-1. Programming Levels of Various Computer Languages.

 

 

 

Table 8—1.

Language

Machine

Assembly
BASIC

FORTRAN

CONVERS—M

Programming
Level
Lowest
Low
High
High

Any Level

135

Standard

Libraries
Available

No

Few

Yes

Yes

No

Comparison of Common Programming Languages.

Execution
Time
Fast
Fast

Very Slow
Slow

Fairly Fast

Core
Requirements
Small
Small
Large
Large

Medium

 

 

 

 

136

various languages examined. It shows that although CONVERS-M is not

the ultimate solution, it is at least a start in the right direction.

B. Programming with CONVERS-M

CONVERS-M represents a style of programming which is totally
different from any other programming technique. Whereas most pro-
gramming languages are based on a single, predefined set of commands
which can only be altered or expanded with difficulty, programming
with CONVERS-M consists almost entirely of creating new commands.
This type of dynamic command structure is best illustrated through

the use of examples.

One task of the computer system in the vidicon rapid scanning
stopped-flow spectrophotometer is to open a valve, thereby enabling
two chemicals to rapidly mix and flow into an observation cell.

Since this is a fairly common operation in the system, it was decided
to create a command called FLOW which would cause this to happen.
The simplest way to create this command is to type, at the terminal,

terminal, the following set of instructions:

<< FLOW
PUSH PSW
MVI A,4
STA 16000
POP PSW
>>

The instruction '<<' tells the CONVERS—M monitor that a new command

     
   

"(113.1“: In I-I I-.i ‘.'.'i.l1!I'IH:]n'."! If .- -1.. II a.r'.":::-.nr1m 1+3.-

' .I- I-I’L'X":

137

named FLOW is about to be defined. The next four lines are assembly
language instructions which, when executed, interact with the hard-
ware and cause the valve to open. The final instructin, '>>', sig-
nals the monitor that the definition has been completed.

The completed command can be used in two different ways. When
only the command is typed at the terminal it is being used in the run
mode. In this mode the command is executed, causing the valve to
open. This allows testing of the new command and, later on, provides
a simple method of testing the valve and associated electronics in
case of a hardware failure. The other mode in which FLOW can be used
is in the creation of new commands. This mode, defined as the compile
mode, will be illustrated through the use of another example. A more
complex task that must sometimes be performed in the stopped-flow
system is to empty the sample and reagent syringes. The syringes can
be partially emptied by opening the flow valve, then closing it and
allowing the reacted chemicals to be expelled to the waste container.
If this is repeated a number of times, the syringes will be emptied.
A command called EMPTY was created to accomplish this task by typing

the following:

<< EMPTY
PUSH PSW
MVI A,4
:10 WAIT
FLOW
WAIT

WASTE

-. . ‘1' ,u.-._ . -‘
J?" W-:- _4‘, I.

I... mm” f"

.-: : I: '1 1' ... ._._ _II.‘ -_'=' .I. .53 Hull-“mgr, Stag-f

.I III I‘ III .9150“

ELI-'- ”“qu

 

   

138

DCR A
JNZ .10
POP PSW
>>

The definition of the EMPTY command differs in only two ways from
the simple definition of the command FLOW. First there are three
commands, including FLOW, which are not assembly language commands.
The other two non—assembly level commands are WAIT, which causes the
system to wait until the valves have switched and the solutions have
stopped moving, and WASTE, which causes the solution to be emptied to
the waste container. Both of these commands have been previously
defined, just as FLOW was. In this mode the FLOW command is not ex-
ecuted immediately. Rather, a subroutine call to the routine FLOW is
linked into the EMPTY command.

The other new technique which is illustrated in the EMPTY command
is the use of branching within a command definition. In this case it
is desired to execute the command sequence, WAIT, FLOW, WAIT and
WASTE, four times. This is accomplished by three assembly language
instructions. First the instruction 'MVI A,4' loads the accumulator
register in the computer with 4. The command sequence is then exe-
cuted once, the 'DCR A' instruction causes the accumulator to be
decremented by one and if the result is non-zero the 'JNZ .10'
instruction causes the system to branch back to the WAIT command
referenced by the ‘:10' label. In general, any location within a

command definition can be labeled with an octal number between 1 and

 

139

177, preceeded by a colon (e.g. :10, :4, :103, etc.). This location
can then be referenced by repeating the octal number, this time
preceeded by a decimal point (e.g. .10, .4, .103, etc.).

In all the preceeding examples, when a command was used in the
compile mode, a subroutine call to the specified command was linked
into the definition. This is the simplest and most common type of
command. A more powerful and versatile type of command, the macro
command, is any command which causes more than a simple subroutine
call to be linked into the definition. One way in which a macro
command might be used would be to speed up the execution of a command
by linking the actual assembly code for the command into the definition
instead of linking in a subroutine call to the command. For example,
forty percent of the execution time of the FLOW command is involved
in the subroutine call and the return from the subroutine. If these
steps were eliminated, the command would execute much more rapidly.
The FLOW command could be made into a macro command by defining it in

the following way:

<< FLOW

MODE FLOW 300

MAC 7
PUSH PSW
MVI A,4
STA 16000
POP PSW

>>

    
  

    
 
  

 
   

_ _I. .3.“
i-""-‘3 .L I --

    
 
      

..”; his“, my. -. ("Em -.aow E's-Ififl'fi II1 I;

"III

M .IT.I.!. mm Ian-salam- 1.2111594? -.-.'.' ..-T LIE-5 5'173'4‘5WI‘ 5‘ .obdu - - .

:- :---'I I-I":'!I.-:".‘I 'T.r-I'.-'.: .1..-:: - - .:_".. 1'. - 1'1 .--_..1:;;.1‘I'sb

 

140

Only two new lines of code are required to make FLOW into a macro
command. The first new line of code, starting with MODE, informs the
monitor that the command FLOW should execute even in the compile mode.
The second new line of code starting with MAC, causes the four assem-
bly language instructions (7 bytes of code) to be linked into a def-
inition when the FLOW command is used.

CONVERS—M was designed to operate in the multi-microcomputer
system described in Chapter 3. CONVERS—M enables most of the details
of the interaction between the microcomputers to be kept transparent
to the programmer. Each microcomputer is assigned a number with the
coordinator designated as microcomputer 0. To create commands in any
of the microcomputers, it is first necessary to specify in which
microcomputer the commands are to be created. For example, the FLOW
command described earlier must reside in the stopped—flow microcompu-
ter which is microcomputer 4. By typing 'MICRO 4' at the terminal,
any commands, including FLOW, which are now created will reside in
microcomputer 4. Once the commands are created in the slave micro-
computers, it is necessary to be able to create routines in the
coordinator which are able to cause these routines in the slaves to
be executed. This is accomplished through the use of four macro
commands. An example of a program segment, which would cause the

command FLOW to be executed by the stopped-flow slave, is the following:

SETUP 4,FLOW
RUN 4
:10 FLAG 4

JZ .10

L—

9 I
I

4

».x-‘

Lanna-Ian I

t ,"-

 

141

STOP 4

The first line specifies that the command FLOW on slave 4 is to be
executed. The next line creates a code which will cause slave 4 to
begin execution of the FLOW command. The next two lines form a loop
which checks the 'task completed flag' of slave 4 to determine when
the FLOW command has completed execution. This final line causes slave
4 to halt any further program execution.

For further details on programming with CONVERS—M, refer to
Appendix A (The CONVERS—M User's Guide) and Appendix B (CONVERS-M

Basic Command List).
C. The Structure of CONVERS—M

The command library of CONVERS—M is structured to allow for simple
expansion. As new commands are created, they are entered into the
library. Figure 8-2 shows the way in which a typical library command
is stored in memory. The body of the command definition is preceeded
by the command header. This header contains the command name, a two
byte link address used during library searching, and a mode byte
which specifies the mode in which the command is used.

Whenever a command is used, it is necessary to search the library
in order to find the command. The search of the library is accom-
plished through the use of the link addresses stored with each
command. Figure 8—3 shows how the library search is performed. The
library is searched beginning with the last command that was entered.

If that command is not the one sought, then the search is transferred

to the address stored in the link address following the command name.

     

       
    

      

Hg-I:

._- I - '"I-L' I " : 'Il-IIIJ' ' -' ‘21:"
'5‘me ’5 5 ‘-

.' "'. . . .. " ..." .. ."' ' ‘1,
tom's; Inn-fi- 291ml“ w: 1qu. :II‘!‘

  

runs InrmwnTrb I PTnI? In REIT Barmlnflft IraI' s41 ai‘afl

"I'll. :' l-.' . .: I-.I- -_ -.".'ll. --- ' ‘-" 5.";‘1 -."I'-:.".'I"

 

 

Storage

Location

KOOO\IO\UI-I><NNHO

10
ll
12
l3
14
15

Figure 8-2.

142

Contents

# Characters
F
L
0
W
Low Byte Link Address
High Byte Link Address
Mode
Assembly Code

Return

Command
Definition

<< FLOW
PUSH PSW
MVI A,4
STA 16000
POP PSW
>>

The Structure of a Command Definition.

 

 

SUBROUTINE

NKDDE
LINK

 

lINK

 

‘ Figure 8-3. CONVERS-M Library Searching Procedure.

 

144

The link address points to the command name which is the next lowest
in the library. This process is continued until either the command
being sought is found or the entire library is searched with no
match being found. The end of the library is found when the link
address points to a command name whose length is zero. If the char-
acter string being searched for is not found in the library, then the
attempt to interpret the character string continues on through a
number of levels. For example, in the compile mode, the character
string is compared to assembly language instructions to determine if
there is a match. The entire process of character string interpreta—
tion is shown, in detail, in Appendix C (The CONVERS—M Flow Chart).
When CONVERS-M operates in the multi-microcomputer system each
microcomputer contains its own separate library of commands. The
coordinator microcomputer contains almost all of the CONVERS—M
monitor overhead for the entire system. The basic memory allocations
for CONVERS-M in the coordinator and a typical slave microcomputer
are illustrated in Figure 8-4. The first 1 K (the first 1024 memory
locations) in the coordinator contain the CONVERS—M system routines.
These are routines that are not stored in the normal library format.
This is the heart of the CONVERS-M monitor. The first section of
code found here is the system initialization routine, which is exe-
cuted only when the entire system is reset. Next comes the character
string interpreter, followed by the assembler routines, the label
decoding routines, and finally, the assembly language conversion
tables. The next 1 K of memory contains the minimal CONVERS-M library.

Here are found routines, stored in the normal library format, which

 

145

COORDINATOR SLAVE

MEMORY MEMORY
TOP OF
MEMORY

Figure 8-4.

SUBROUTINE
RETURN

STACK

T

SOFTWARE
STACK

 

SPECIFIC
APPLICATION

LIBRARY

1

MINIMAL
CONVERS

LIBRARY

CONVERS
SYSTEM
ROUTINES

 

2K

IK

0

CONVERS-M Memory Allocation.

SUBROUTINE
RETURN
STACK

T

SOFTWARE
STACK

 

SPECIFIC

APPLICATION
LIBRARY

CONVERS
OVERHEAD

 

”at“
13A“

       
 

QI-F.‘ _ a .—..an.

 

 

146

are essential to the operation of the monitor. Above this is free
memory, available for the storage of specific application libraries.
The top of the memory in each of the microcomputers is reserved for
the subroutine return stack. Below this is found the software stack
which is used in one type of parameter passing. (See the section on
parameter passing in Appendix A). The memory allocation in the slave
microcomputers, as indicated in Figure 8—4, is much simpler. Here
the CONVERS—M monitor overhead requires only about 20 locations. The
upper half of this area contains monitor parameter storage locations.
In the lower half of this area reside the four to six assembly lan—
guage instructions which are required to enable any routine in the
slave to be executed. These instructions first load the hardware
stack pointer and then perform a subroutine call to the desired
routine. Upon return from this routine, the slave signals the coor—
dinator that it has completed the routine and then halts. In this
way the coordinator can cause any routine in the slave to be executed
by simply placing the address of that routine in the subroutine call
and then running that microcomputer. Since the monitor overhead

is so small in the slave microcomputers, most of the memory is left

free for the specific application libraries for the slaves.

w-E 9H3 : ' .--.' 2.3.: .-_..uu~:~.-

 

 

 

 

APPENDICES

 

 

 

 

APPENDIX A

CONVERS-M USER'S GUIDE

 

 

 

APPENDIX A

CONVERS-M USER'S GUIDE

A. Introduction

This Appendix is only designed to give details on some of the
programming conventions used in CONVERS-M and is, therefore, not a
complete manual for the use of CONVERS-M. The structure of CONVERS-M
and some examples of its use can be found in Chapter 8. Appendix B
contains a listing of all the commands which are found in the minimum
CONVERS-M library and a flow chart of the operation of the CONVERS-M

monitor appears in Appendix C.

B. Special Characters

Most operating systems input an entire line of characters before
actually beginning to interpret the character string. CONVERS-M oper-
ates On character strings which are smaller than this. Here, a char-
acter string is defined as a group of characters followed by a separator
which is either a comma, space, tab, or carriage return. The monitor
does not differentiate among the four separators, therefore, in any
specific application, it does not matter which of the four separators
is used. The following two program segments differ only in the separ-
ators which are used and, therefore, are compiled identically by the

monitor.

MVI A,10

:10 DCR A

147

    
 

  
 

3:11 ":0 “WWW 33:3. _EIIEIJD amig- cu .."rr';: '

,.', "' (".Il‘pu'fl ..1.‘ '--.- ' J'I'fl': H'- . |.1 i:.\'..'l.'I-L'J'fial"'7.'l Enimm

 

 

148

JNZ .10
MVI,A 10,:10 DCR

A,JNZ .10

The semicolon is a special character which is used to indicate
a comment. Comments are character strings which appear on the term-

inal, but are ignored by the CONVERS-M monitor. A comment begins with

a semicolon and is terminated by a carriage return. There are no re-

strictions on the characters which may appear within a comment.

**

; Examples of Comments

**

MVI A,10 ; At the end of a line

; Between Lines

10 DCR A
JNZ .10
C. Labels

Labels are character strings which are used to define or reference
locations within a command definition. They are always octal numbers
between 1 and 177, preceeded by either a period or a colon (e.g. :10,
.26, :170, etc.). If the label starts with a colon, then it iS defin-
ing a location. A single octal number can define only a single location.
If a second label definition is attempted with the same octal number,
then the monitor will signal an error by outputting a 'P', indicating

that the number was used in a previous definition. These defined

_ .. 'I ;,.... .....:. . n.1,. L5
“_f ...: "I... ......” , .,_ ._ .. . .

 

149

locations can now be referenced by using the same octal number, this

time preceeded by a period.

When a command is being created, a temporary library of the label
definitions and references is stored in memory. This temporary library
is located just below the software stack. Each time a label is defined,
the memory address at which the compiler will next store code is entered
into the temporary library. Whenever that label is referenced, this

address is stored in the next two bytes of the command definition. An
example of the use of labels is shown in the simple timing loop routine

shown below.

LDA .20
:10 DCR A
JNZ .10

RET

:20 10

In this example the 'A' register is loaded with the octal value 10.
This register is decremented until it is zero, and then command is

returned to the calling routine.
D. Parameter Passing

It is often necessary to transfer data or parameters from one
routine to another. There are a number of ways that this can be
accomplished. The most common method is to use the software stack.
The software stack is a last-in—first—out data buffer located near the

top of memory below the subroutine return stack. The address, at which

  
       
 

"{'ni'r'lIgfiz-a‘jmiffimui .. 'I I: ..

"- a"? --‘. '21. .-n.1uf'3-F 9111i: dyad JIM-Jr -.=;.-«.--a’-toa :21” "Mm-”I 3'3"! I

I. . . .--.-.-.-.-_- 7...... ..-,:...-: 1v .2.-'::"‘.5"I “vat-"Mi“

 

.3:

 

150

the data is currently being stored, is kept in the stack pointer which
occupies two bytes of storage in the monitor. As this data buffer
fills, the data is stored in successively lower memory addresses. This
leaves the bottom of memory free for the user's software. There are
four commands in the basic monitor which simplify the use of the soft-
ware stack. To move a single byte of data onto the software stack, the
'ON' command is used. This command moves a single byte of data from

the 'A‘ register onto the stack. To retrieve this byte, the 'OFF'
command is used. This command removes a byte from this stack and places
it in the 'A' register. In the following example, the first program
segment places the octal number 10 onto the stack and the second segment

retrieves that number, and moves it to the 'B' register.

MVI A,lO
0N
OFF
MOV B,A

It is also possible to use the software stack for two byte data. In

this case the command 'PUSHZ” puts the two bytes contained in the 'H‘

and 'L' registers on the stack. The command 'POPZ' removes two bytes

of data from the stack and places it in the 'H‘ and 'L' registers.
Another method of passing data to another routine, is to store the

data directly after the call to that routine. An example of an assembly

language routine which could accept this type of data transfer, is the

following:

, ‘L ‘ ‘h‘t’

J1 ,
.7 "f

‘ ', W -

5,9375.) {M id?“ $5M»! 4 .WII ’dlfi’lf

 

 

151

<< WAIT
XTHL
MOV A,M
INX H
XTHL

:10 DCR A

JNZ .10

This routine causes a time delay whose length is determined by the byte
of data following a call to this routine. For example, if the line
'WAIT 6' appeared in a command definition, then the internal loop in the
'WAIT' routine would be executed six times. The actual transfer of the
data to the 'WAIT' routine is accomplished by the assembly language
instructions which form the second through fifth lines of the definition.
These four lines of code move the byte into the 'A' register and also
increment the return address to insure that the return will occur to the
location after the data byte. This type of parameter passing is used
primarily when the data is predefined and will never change. The above
'WAIT' routine would be a good example in an application where a number
of software delays, each of a different predetermined length, were
needed. Variable data is best handled through the software stack.

When it is necessary to pass data very rapidly between routines,

the internal registers of the computer can be used. Another technique

which is almost as fast is to use a predetermined address in memory to
store the common data. These last two methods should only be used when

no other method is fast enough, since it requires the programmer to

 

152

remember in which registers or memory addresses the data must be stored.
E. Mode Word

Each command definition has a one byte mode word stored with it.
The default mode specifies that when the monitor is in the run state
the routine will execute only if the coordinator microcomputer's
library is currently loaded and the routine is found in that library.
If the monitor is in the compile state, a call to the routine will be
compiled only if the routine is found in the currently loaded library.
Bit 6 and bit 7 in the mode word are the 'run only' bits. They are used
to override the default actions and cause routines to be executed when
they normally would not be. Since the monitor can only execute routines
found in the coordinator's library, these mode bits only affect routines
found in that library. Bit 7 is the 'run only' bit for the coordinator's
library. If the coordinator's library is currently loaded and bit 7
is set, the routine will execute regardless of whether the monitor is in
the run or compile state. Bit 6 is the 'run only' bit for a slave's
library. If a slave's library is currently loaded and the routine is
found in the coordinator's library, the routine will execute regardless
of the state of the monitor. If both bits are set, then the routine

will always run.

F. Using the Floppy Disk Drive

Mass data storage for the multi—microcomputer system is accomplished
through the use of floppy disks. They are used for storing binary

programs, command definitions and data. Most of the programmer's

 

153

interaction with the floppy disk drive is handled by the 'FLOPY'
routine. The floppy disk drive is interfaced through the Persci Model
1070 Intelligent Diskette Controller. Since the controller contains a
powerful set of commands for manipulating files on the disks, the
'FLOPY' routine simply serves as a means of passing these commands

from the user to the controller. A complete description of the con—
troller's commands can be found in the product specifications for the
Model 1070 controller (73). The 'FLOPY' routine interacts in a special
way with two of the controller's commands. The controller's 'LOAD' (L)
command causes the contents of the specified file to be listed on the
terminal. The 'SAVE' (S) command can be used to create files on the
disk. The process of creating a file is started when the 'SAVE' command
is entered. The file name specified with the command is entered on the
disk as the name of the new file which is being created. From this
point on, everything that is typed at the terminal will be entered into
a data buffer. Each time a carriage return is typed, the contents of
the buffer are written into the disk file. When a control/C is typed,
the input is terminated and the disk file is closed. An example of the

creation of the disk file, WAIT.CON, is shown below.

FLOPY
* S WAIT.CON
<< WAIT
MVI A,10
:10 DCR A

JNZ .10

. . I.
'p r" r|.

. a

I. sum-w

_LL‘: : r 7

 

154

>>

(CTRL/C)

*

In the above example the asterisks (*) are output by the 'FLOPY'

routine whenever it is inputting command strings to be sent to the

controller. To terminate the 'FLOPY' routine and return to the monitor,

control/C is typed at the terminal.

G. Assembly (Compilation of stored routines)

It would be extremely tedious if the programmer had to enter, at
the keyboard, the definitions for all the commands not already found in
the monitor, which he planned to use during that session. To alleviate
this, it is possible to enter command definitions that were previously
stored on floppy disks. This is done through the use of the 'ASSM'
routine. The 'ASSM' routine interacts with the floppy disk controller
in much the same manner as the 'FLOPY' routine described in the preceed-
ing section. In this case, when the controller's 'LOAD' (L) command is
used, the contents of the specified file are compiled into the library
which is currently loaded. For example, to compile the 'WAIT' routine,

which was used as an example in the last section, the following would be

typed:

ASSM

# L WAIT.CON

The number sign (#) is output by the 'ASSM' routine to indicate that it

‘ ' .... .
'l'lll‘ " -. u;
I

 

 

155

is ready for a controller command. As the 'WAIT' routine is read from
the disk, it is output to the terminal along with any output (normally
error messages) from the monitor itself. As soon as the end of the disk
file is encountered, the system returns to the monitor.

Since the entire disk file is output to the terminal, it is often
difficult to notice any error messages which the monitor might output.
This problem can be overcome by the 'DCRT' command which allows only
output from the monitor to be sent to the terminal when the 'ASSM'
routine is used. The 'DCRT' command inputs an octal number from the
terminal. If the number is 6, then all input from the disk will be
sent to the terminal also. If the number is not 6, the terminal out-
put is disabled.

The 'ASSM' routine functions by forcing all input to the monitor
to come from the floppy disk instead of the terminal as it normally
does. Upon encountering the end of file from the disk, monitor input
once again comes only from the terminal. For this reason, almost any
commands which are allowed at the terminal can be input from floppy
disk as well. The only exceptions are commands which also use the

floppy controller. This means that batch streams can be stored on the

disks and executed at any time.

H. Communication with the PDP 11/40

It is often necessary to send files to and from the PDP 11/40
minicomputer. This communication is accomplished by the 'PIP' routine
in the multi—microcomputer system. (This routine is not found in the

basic CONVERS—M library, it must be loaded from floppy disk). When

    

"..I- -. _"-. -r"
'ij‘xfiflifnt-m sin: m raw-H: maxi. :r {W ._ ' *

      
  
 

-,.‘ 'ZIIIZ'.’ n' .Hn,':;j-". "' all? 'JEIII ”HENRI-r “.7955“?

I -H"- i . ' “‘ .
I .. ,_ :I.u _- -.- - ... : :. «M -.r.;1_-_- -. '- ri L": I"? 11
I I I" .I I?!”
' ':!'.z

 

 

156

this routine is run, it first outputs a dollar sign ($) indicating that
it is requesting the mode of the current transfer. This mode that must
be input is made up of two characters. The first character indicates the
destination of the transfer. If the character which is typed is a 'U',
then the destination of the transfer is the floppy disk in the multi—
microcomputer system. If this character is an 'M', then the destination
is the PDP 11/40. The second character defines the type of transfer
which will take place. If the character is an 'N', it indicates that
the transfer will be a 'normal' transfer. This means that an exact
copy of the file will be transferred. The character 'A' indicates that
the transfer will be an ASCII transfer. This is necessary since the
CONVERS-M system uses only a carriage return to terminate a line, where-
as, the PDP 11/40 uses both a carriage return and a line feed to termin-
ate a line. This problem is corrected during an ASCII transfer.

When the 'PIP' routine outputs a number sign (#), it indicates
that it is ready to input a standard file specification for the PDP
11/40. After that has been input, the routine will output an asterisk
(*), this time requesting a file name to be used on the floppy disk in
the multi—microcomputer system. Once this file name has been input,
the transfer will take place and the routine will be ready to input the
next transfer mode from the user. As example of an ASCII transfer from
a floppy disk on the PDP ll/40 to the multi—microcomputer system is

shown below.

$UA

# FL:WAIT.CON

 

 

157

* WAIT.CON

$

In the above example the file 'WAIT.CON' is transferred from the PDP
11/40 to the microcomputer system and placed in a file also named
'WAIT.CON'. To terminate the 'PIP' routine and return to the monitor

control/C is typed at the terminal directly after the dollar sign ($).

1. Error Codes

Due to the necessity of keeping the CONVERS-M monitor 'short and
simple', the error codes that are output by the system are limited.
The monitor itself has only three different error codes. By far the
most common error code is the question mark (?). This error can mean
nearly anything. Some typical examples are: input of an unknown
character string, illegal micro number, illegal label, illegal register
during assembly, attempt to input something besides an octal number
when an octal number is required, etc. At times the system can appear
to 'lock up' and output question marks no matter what is typed. This
can be caused by one of two problems. First, there are some routines
that require the input of an octal number and they will continue to
output question marks until an octal number is typed at the terminal.
The second problem which can cause the system to 'lock up' is when an
assembly language instruction is used which requires the use of a
register code (e.g. MOV A,B etc.). The system will output question
marks until a correct register code is input.

The other two error codes which the monitor uses involve improper

use of labels or command names. The three error codes used by the

158

monitor are listed in Table A-1.

More detailed error codes are often encountered when using any of
the routines which interact with the floppy disk controller. These er-

ror codes are part of the software which comes with the controller and

all are self-explanatory (2).

Table A—1. Monitor Error Codes.

Error Code Error

Nearly anything

P Indicates a previously
defined command name
or label definition

nnnU Indicates an undefined

label with 'nnn' being
the label in question

Sometimes detailed error codes are also found when using the 'PIP'
routine to transfer routines to and from the PDP ll/40. The detailed
error codes are sent by the PDP ll/4O when it encounters an error in
the transfer. The 'PIP' routine also outputs a number of error codes,
all of which begin with the letter 'F'. These 'PIP' error messages

are listed in Table A—2.

 

_. —— .
-\ll ...

 

159

Table A-2. 'PIP' Error Codes.

Error Code

F0

F1

F2

F3

F4

F5

Error

Impossible? error code
No ACK/EOT from PDP 11/40 on start up

NAK from the PDP 11/40 (ll/40 found an
error)

Checksum error during transfer to
PDP 11/40

Checksum error during transfer from
PDP 11/40

Unknown code sent from PDP ll/4O

 

 

APPENDIX B

CONVERS-M BASIC COMMAND LIST

ASCII

ASSM

BININ

CONVERS-M BASIC COMMAND LIST

A. Coordinator's Basic Command List

Performs octal to ASCII or ASCII to octal conversions.
(i.e. adds or subtracts 6O octal to or from the data.)
The top of the stack must contain a mode byte.

Mode = O: ASCII to Octal

 

Mode = l: Octal to ASCII
INPUT OUTPUT
Top of Stack: Mode # Char.
# Char. Etc.
Etc.

Is used to compile commands which are stored on floppy disk.

Details of its use are given in section G of Appendix A.

Reads an octal number (between 0 and 177777) from the termin—
al. The octal number is converted to binary and the 2—byte

result is stored on the stack.

Top of Stack: Low Byte
High Byte
If the input from the terminal is not octal, a question mark

is output and a new number is accepted.

160

 

BINOCT

BINOUT

CLEAN

CRLF

DEP

161

Converts a 2—byte binary value on the stack to octal and

puts the resultant octal digits on the stack.

 

INPUT OUTPUT
Top of Stack: Low Byte # of Digits
High Byte Least Sig. Digit
Etc.

Most Sig. Digit

Leading zeros are suppressed on output.

Converts a 2-byte binary number found on the stack to octal

and outputs it to the terminal with leading zeros suppressed.

Top of Stack: Low Byte

High Byte

Is used to remove unwanted data strings from the stack. The

top of the stack must contain the number of bytes to be re-

moved.

Causes a carriage return / line feed to be output to the

terminal.

(Run only mode) — is used to alter a location in memory.

DEP AAA DDD

 

EXM

FLAG

FLOPY

LP

MAC

162

AAA - Octal address in memory

DDD — Octal data to store in AAA

(Run only mode) — is used to examine a location in memory.

EXM AAA
AAA — Octal address to examine

The octal contents of AAA will be output to the terminal.

Is a macro routine which creates a code segment to check
the task completed flag of the specified slave microcomputer.

If the task is completed the internal zero flag will be reset.

Handles file manipulations on the floppy disks. It is

described in section F of Appendix A.

Inputs an octal number from the terminal. If it is a 5, then
the output of the 'FLOPY' routine is sent to the TTY instead
of the terminal. If any other number is input, then the

'FLOPY‘ routine functions normally.

Simplifies the creation of macro routines. When run, it

enables strings of code to be linked into a definition.

MAC 3, CALL 1106

The above example causes the three byte instruction, CALL

 

 

 

163

1106, to be linked into the definition.

MCHG Loads the CONVERS-M library of the micro corresponding to

the l-byte number on the top of the stack, (i.e. 0,1,2,3, etc.)

 

MESSAG Causes a message to be written on the terminal.
INPUT EXAMPLE

Top of Stack: # Char. 4

Last char. sent '2'

Etc. 'T'

First char. sent 'U'

'0'

Output of example: OUT2

MICRO Loads the CONVERS-M library of the micro corresponding to

the octal number input from the terminal.

MICRO DD
DD — Octal micro number between 0 and NMICRO inclusive

NMICRO — Current number of micros in the system

This value is stored at octal location 52.

MODE Allows the mode word of a routine to be changed.

MODE NAME DATA

NAME = Name of the routine

      

. {EH-Ir. _ ..
- .1 245‘ m' I
1a: - " —
-:-.l [‘5' 1‘. I
“HAW! "‘1' 'U‘;
..., -{--"" T'-

 

 
 
 

1.-

 

 

 

 

OCTBIN

OFF

ON

164

DATA = The new mode

Converts the octal digits (0—7) found on the stack to a

single 2-byte binary word.

 

INPUT OUTPUT*

Top of Stack: # Char. Low Byte
Most Sig. Digit High Byte
Etc.

Least Sig. Digit
* If there are too many digits or if the data is not octal,

the carry flag is set (otherwise clear) and the output is:

OUTPUT

 

Top of Stack: # Char. left

Etc.

Pops one byte off of the software stack and puts it into the

'A' register.

A < = ((Stack))

(Stack) < = (Stack)+l

Pushes one byte from the 'A' register onto the software stack.

((Stack)-l) < = A

(Stack) < = (Stack)-l

 

 

POPZ

PUSHZ

RUN

SEARCH

165

Pops two bytes off the software stack and puts them in the

'H and L' registers.

L<

((Stack))

H<

((Stack)+l)

(Stack) < = (Stack)+2

Pushes two bytes from the ‘H and L' registers onto the soft-

ware stack.

((Stack)—l <

II

H

((Stack)-2 <

L

(Stack) < = (Stack)-2

Is a macro routine for creating program segments which reset

and run the specified slave micro.

RUN 2

This example will create a program segment which will reset

and run slave micro number 2.

Searches the loaded CONVERS—M library for a match to the
character string found on the stack. If a match is found

the carry flag is set, otherwise it is cleared.

T_§_.3££ 3J1“!- 1- i

 

 

166

 

INPUT OUTPUT
Top of Stack: # Char. No Change

First char. of name

Etc.

SETUP Is a macro routine for creating program segments which
enable specified routines on slave micros to be executed

when the 'RUN' command is used.

SETUP 1,INIT
In the above example, the library of micro number 1 is
searched and the address of the routine, 'INIT‘, is compiled
into a program segment which, when run, will load that address

into the initial call of micro number 1.

STRING Inputs a character string from the terminal.
OUTPUT EXAMPLE

Top of Stack: # Char. 4

First char. in '0'

Etc. 'U'

'T'

'2:

Input from terminal for example: OUT2

WHAT Outputs a question mark to the terminal.

 

 

/DI

/FR

/1<1

<<

167

(Run only mode) - lists all the names of the commands in the
CONVERS—M library which is currently loaded. The first com-

mand listed is the last command that was entered.

(Run only mode) — outputs the difference between the address
of the bottom of the stack and LASTAD. This gives a rough
idea of the free core available. (Since the bottom of the

stack will decrease during program execution, this value is

larger than the actual core available.)

(Run only mode) — removes the last entered command from the

CONVERS-M library.

(Run only mode) - outputs the current value of LASTAD to the
terminal. (i.e. the address at which code will next be

stored.)

(Run only mode) — is used to begin the creation of a new

command. It switches the mode to compile mode and requires
the input of a command name which has not been previously
defined. The command name is then entered and linked into

the CONVERS—M library.

<< NAME

NAME — Command name of any length.

 

 

 

>>

ASM

CRT

FCMD

168

(Run only mode) - terminates the creation of an instruction

by placing a return at the end of the routine and flags any
undefined labels as errors. The system then returns to the

run state .

B. Communicator's Basic Command List

Is used by the 'ASSM' routine in the coordinator's library

and has no use outside of this context.

Outputs a single byte, from the 'A' register, to the term—
inal. It also handles the special characters: carriage
return, tab, delete and page. The contents of the 'A'

register are unchanged upon return.

Causes the command stored on the software stack to be sent
to the floppy disk controller. If an error is encountered
the carry will be set. The 'A' register indicates the state

of the controller after accepting the command.

 

'A' Reg. State
0 Command completed.
3 Controller ready to

accept data to place

in a file.

5 Controller ready to

output data from a file.

 

FLOPl

INIT

INPUT

LINE

169

Is used by the coordinator's 'FLOPY' routine and is of no

other use to the programmer.

Initializes the communicator micro. The keyboard is cleared,

the floppy disk drive motor is disabled, both drives are
write protected, the terminal is paged and the software stack

is reset.

Is the character string input routine for the monitor. It
inputs characters until it finds one of the separator char-

acters (space, comma, tab or carriage return).

OUTPUT

Top of Stack: # Char.
Last char. typed
Etc.

First char. typed

Inputs an entire line of characters from the terminal.

OUTPUT
Top of Stack: # Char.
First char. typed
Etc.
Last char. typed
If a control/C character is encountered, the software stack

is cleared and the carry bit is set.

 

 

 

OFF

ON

PRINT

READ

TOUT

TTY

170

See the Coordinator's Basic Command List.

See the Coordinator's Basic Command List.

Causes a character string to be output to the terminal.

INPUT

 

Top of Stack: # Char.
First char. printed
Etc.

Last char. printed

Reads a byte from the keyboard and places it in the 'A'

register.

Sends a byte from the 'A' register to the teletype. The

contents of the 'A' register are unchanged upon return.

Outputs the 'A' register to the teletype. It also handles

the tab and carriage return special characters.

' ".3235 fir‘n%snaift

'4

 

 

 

 

 

 

APPENDIX C

 

CONVERS-M FLOW CHART

 

 

 

APPENDIX C

CONVERS—M FLOW CHART

       

INPUT CHAR.
STRING
SEARCH CURRENT
LIBRARY

      

SEARCH

COORDINAYOI'S
UBRARY

  
   
   
 

COORD.
LIBRARY
LOADED

EXECUTE
ROUTINE

 
    
   
  
   

LABEL
OR ASSM.

    

COORD.

RUN ONLY
BIT SET ?

ENTER CALL TO
ROUTINE

Figure C-l. Flow Chart of the CONVERS-M Command Decoding Software.

      
   
    
 

171

 

 

172

       

SET CARRY

         
      
  
   
   

LOAD ASSM. CODE
TABLE WITH CORRECT
If OF CHARACTERS.

FIRST TYPE OF ASSM.
CODE.

   
 
 

  

SEARCH TABLE OF
CURRENT TYPE OF ASSM.
CODE.

NEXT TYPE OF ASSM. CODE

EXECUTE APPROPRIATE

HANDLER DEPENDING ON
TYPE OF ASSM. CODE
FOUND.

  
     
   

CLEAR CARRY

   

Figure C-Z. Flow Chart of the CONVERS-M Assembler.

 

 

 

 

Figure C-3.

SEARCH LABEL
LIBRARY

DEFINITION
FOUND?

SEND 'P' FOR
PREVIOUSLY DEF.

173

GET FIRST CHAR.

ENTER IT IN
THE LIBRARY

ENCODE
LABE L

SEARCH LABEL
LIBRARY

DEFINITION
FOUND ?

STORE ADDRESS
AT LASTAD

 

Flow Chart of the Label Decoding Software.

 

 

 

 

 

APPENDIX D

 

PROGRAM LISTINGS

 

 

 

 

i

I#########################

I

I#####################fi##fi

I

IBATCH STREAM TO INITIALIZE MICRO 2

MICRO

0

APPENDIX D

A. Microcomputer Programs

ILOAD THE ROUTINE

:10

kc BOOT2

MVI
STA
MVI
LXI
LXI
LDAX
MOV
INX
INX
DCR
JNZ
RET

A:

m
N

I

x -

'GUIZUUIU

ya
0

2
106

.40000

>
0
O

IMICRO 2 CONVERS-M OVERHEAD

l

:30

LXI
CALL
RST
HLT
40120
40016
11500
0:0
MV
STA
STA
HLT

SP.11777

O

2

A;
12
12

1
0003
0002

IMICRO 2 PARAMETER STORAGE AREA

l

---c~>wo-
“o o
oo

OOH0~OOOHHV00MOO
OOOOOOFA‘O‘OO

U
U

532
£34
i36
;4O
i42
:44
i46
550
352
i54
556
560
:62
ib4
;65

DPTS
AVGS
PTS
INT
FREQ
BEAM
DELY
#D
#C
GPTS
AVGA

SCNS

IIIIIIIIIIIIIII

174

iBOOT MICRO 2
iTHO MICROS

iSEND 100 BYTES
iSTART OF MICRO 2
iSTART OF TABLE

iSTORE IT

sMICRO # 2

iLASTAD
5LINKAD
iSOFTNARE STACK

sENABLE HOLD
iSEND THE FLAG

DATAO

TOTAL POINTS IN ONE D
NUMBER OF SCAN AVERAGES
HORIZONTAL SCAN INC.
NUMBER OF INTEGRATIONS
VERTICAL SCAN FREQUENCY
BEAM CURRENT

ELECT. DELAY AND INT.
NUMBER OF DATASETS
NUMBER OF CUTS

NUMBER OF GRAPH POINTS
NUMBER OF ARRAY AVERAGES
ERROR STATUS

NUMBER OF SCANS

DATA1 (POST-INT FOR LAG)
DATA2 (PRE-SCANS FOR LAG

 

 

 

 

0 366 -
O 567 -
0.0 570 - CPT
0,0 z72 - OVL
0,0 ;74 - NAV
0,0 :76 -
51000 5100-
010 .102— MIN
210 I 104' MAX
iRUN THE ROUTINE
BOOT2
iREMOVE THE ROUTINE
I
IKI
i#####fi#################%#
I
MICRO 2
i#########################
<< VSET
PUSH H
PUSH PSN
MVI All
STA 120010 5
STA 120013 ;
STA 120014 F
LDA 44 3
STA 140000 5
LDA 46 .
STA 140001 5
MVI A.13
STA 140002 ,
LDA 42 .
STA 140004 I
STA 140003 5
LDA 36 .
STA 140005 .
POP PSW
POP H
>>
.< SFLAG I
PUSH PSW I
:10 LDA 140011 I
RAL
JNC .10
STA 140010 .
pr Pew .
CC ADDR
PUSH PSW
LHLD 10 i
MOV AIH
ANI 77 i
MOV HIA
pop Psu
<< BSCAN i
PUSH PSN
SFLAG ;
SFLAG I
MVI A121
STA 120013 I

'44 TO 58 USEC.
'SCAN FLAG

DATAG (POST-SCANS

DATA4

TOTAL POINTS IN ONE CUT
OVERLAP POINTS PER DATA
CURRENT MONOCHROMATOR
NEW MONO. HAVELENGHT
FREE ADDRESS FOR CUTS
MINIMUM

MAXIMUM

S
p

5INITIALIZE VIDICON

BY ES OF PRECISION
ET AIN & TURN F.A.
EAT. F.A. COUNTERS

BEAM CURRENT

2 T

S G OFF
I
B

‘DELY

ELECT. DELAY AND INT.

'ggég ON THE BEAM
‘VERTICAL SCAN FREQUENCY

RESET FLOPS

'PTS
‘HORIZONTAL SCAN INC.

'RETURN ON SCAN FLAG

TOTAL

'CLEAR SCAN FLAG
‘30 TO 44 USEC.

TO RETURN
iLOAD DATASET MAP

LASTAD

CLEAR MICRO BIT

SETUP SCAN
CLEAR SCAN FLAG

'SYNCHRONIZE WITH SCAN
‘SET F.A. SCAN BIT

 

POP
33

PS“

i********************

gROUTINE TO TAKE MULTIPLE SCANS FROM THE
IVIDICON. ALL PARAMETERS PASSED THROUGH
{REGISTERS FOR SPEED. ON INPUT:

SCAN ADDRESS POINTER

l
i
i
i
i
i
i
i
i
i
i
i
i

H&L
B

C
A

ON RETURN:

H&L
B&C
D&E
A

SCAN AVG.
SCAN INT.

176

NUMBER OF SCANS TO TAKE

NEW SCAN ADDRESS POINTER

UNCHANGED
UNCHANGED
CLEARED

i********************

%< MSCAN

10

:30

:40

:50

:60
c< ENDS,

PUSH
STA
XRA
MOV
CMP
JZ
INX
MOV
INX
XCHG
SHLD
MOV
MOV
XRA
CMP
JZ
MVI
STA
MVI
STA
SFLAG
DCR
JNZ
MVI
STA
MVI
STA
HOV
SFLAG
DCR
JNZ
MOV
XCHG
76

O
DCR
JNZ
PUP
\ ‘>

PUSH
MVI
STA
SFLAG
LDA
ANI
STA
STA

‘ 0 ‘ m
3 O 3 0

-m
moo
O
...
H

A‘ M‘ h
OHOHO
ONO
O pd
N w

um- #sw
Ovow
r‘ O
0 m

m'm OH)H)‘0 hflNfiV ODIFH IMI'OOD‘U
FOMOHO

~

N
IO

A
.10
D

PSW
AI1
120013

120016
3

60
120017

ITAKE MULTIPLE SCANS
INUMBER OF DATASETS
iHIGH BYTE OF ADDRESS
iIF 0 - END OF MAP
iLON BYTE OF ADDRESS

iDATASET ADDRESS
ISAVE # INT.
ISAVE # AVG.

iTEST IF INT.
iJUMP IF NO INT.

iCLEAR F.A. SCAN BIT

iCLEAR BEAM SCAN BIT
IINT. STARTS ON FLAG
ITEST IF LAST INT

iSET BEAM SCAN BIT

iSET F.A. SCAN BIT
iREPLACE # OF INT
ISCAN STARTS ON FLAG
iTEST IF LAST AVG.

iREPLACE # AVG.
iMVI AI

5% SCANS
ITEST IF LAST SCAN

IROUTINE TO TERMINATE

iCLEAR F.A. SCAN BIT
iTERMINATES ON FLAG
IGET ERROR STATUS

ISTORE IN ERROR STATUS
iCLEAR ERROR FLAG

 

 

 

177

POP PS“
4’)
<< OPEN2 IOPEN THE SHUTTER
PUSH PS“
:10 AB? $20004 ITEST IF ALREADY OPEN
JNZ .20
STA 120007
UMP .10
:20 POP Psw
I,”
{{ SHUT2 ICLOSE THE SHUTTER
PUSH PSU
:10 LDA 120004 iTEST IF ALREADY SHUT
ANI 2
J2 .20
STA 120007
JMP .10
:20 POP Psw
<< SCAN2 ISETUP AND TAKE MULTIPLE SCAN
LDA 55 IAVGA (ARRAY AVG.)
:10 PUSH PSN
LDA 34 IAVGS (SCAN AVG.)
MOV BIA
LDA 4o IINT
MOV CIA
LDA 62 ISCANS
STA 120017 ICLEAR ERROR FLAG
ADDR ILOAD DATASET MAP
BSCAN iSETUP SCANS
MSCAN iTAKE THE SCANS
ENDS IEND SCANS
POP PSH
DCR A ICHECK IF LAST ARRAY AVG.
9N2 .10
4’ J
{C LAG2 ISTUDY VIDICON LAG
LDA 56 iAVGA (ARRAY AVG.)
:10 STA .20
LDA 34 IAVGS (SCAN AVG.)
MOV BIA
LDA 4o IINT (PRE-INT)
MOV CIA
LDA 64 IPOST—INT
MOV OIA
LDA 66 IPOST—SCANS
MOV EIA
LDA 65 IPRE—SCANS
STA 120017 ICLEAR ERROR FLAG
ADDR ILOAD SCAN TABLE
BSCAN ISETUP SCANS
MSCAN iTAKE INIT. SCANS
MOV CID IPOST—INT
MOV AIE iPOST-SCANS
MSCAN iTAKE FINAL SCANS
ENDS iEND SCANS
7e iMVI A.
:20 o IAVGA
DCR A
JNZ .10

\.
4'}

I###########u##########a##
' MICRO 1
Z#############0«##########

I
I'

 

 

178

<< R1140
:10 LDA 16000
ANI 2
JNZ .10
LDA 16300
>>
<< “1140
PUSH PSW
:10 LDA 16000
RRC
JNC .10
POP PSW
STA 16300
>>
;**********

ITHIS ROUTINE OPENS FLOPPY FILES FOR I/O
IUSING THE FLOPPY LOAD OR SAVE COMMANDS.
iAN ASCII ’5’ IN THE AC OPENS THE FILE
iFOR OUTPUT. AN ASCII ’L’ IN THE AC
IOPENS THE FILE FOR INPUT. (I.E. THE FILE
iNAME IS INPUT FROM THE KEYBOARD I THE
IAC IS PUT IN FRONT AND THE STRING IS

iSENT TO THE FLOPPY.
;**********

{C FILE

PUSH B
MOV CI
MVI AI
CRT
LINE
JC .5

OFF iREMOVE # CHAR.

MOV
MOV IPUT COMMAND ON

ON
MOV
INR
ON
FCMD ISEND & DECODE IT

:50 POP B

; **********

ITHIS ROUTINE DECODES MESSAGES FROM THE

i11/40. THIS RESULTS ARE:

ICARRY ACC MESSAGE FROM 11/40

OI)

iSAVE THE CHAR.
2 I’*’

O

I» 3’?
w 0)

I CLR 0 ACK/EOT - COMMAND RECEIVED & COMPLETED

I CLR 1 ENQ/EOT - 11/40 EXPECTS DATA FROM MICRO

I CLR 2 SOH - 11/40 SENDING DATA TO MICRO

I CLR 3 BEL/EDT - 11/40 EXPECTS FILESPEC FROM MICRO
I SET 25 NAK/EOT - ERROR MESSAGE - ANY CHARS. BETWEEN
I NAK & EOT ARE SENT TO THE CRT.

i ————— --— ——————————————————

iANY CHARS. NOT MATCHING THE ABOVE ARE ASSUMED TO BE A MESS

IAND ARE SENT TO THE CRTI UNTIL ONE OF THE ABOVE CODES
iARE FOUND.

;**§*******
{C MDEC
PUSH B
:10 R1140 IGET A CHAR.
MVI BIO IINIT. RETURN CODE
CPI 6 I’ACK’
JZ .20
CPI 25 I’NAK’
JZ .50 I**ERROR**
INR B
CPI 5 I’ENG’
JZ .20
INR B

 

 

 

CPI 1
J2 .30
INR B
CPI 7
J2 .20
CRT
JMP .10
:20 R1140
CPI 4
JNZ .60
:30 MOV AIB
POP B
STC
CMC
RET
:50 R1140
CPI 4
JZ .60
CRT
JMP .50
:60 POP B
MVI AI25
ETC

;******%**§

179

I’SOH’

i’BEL’

iNEXT CHAR.
i’EOT’
5**ERROR**

iGET NEXT CHAR.
I’EOT’

iTHIS ROUTINE SENDS THE CHAR STRING
iON THE STACK TO THE 11/40 AND
iUSES MDEC TO DECODE THE REPLY

;***§******
{{ MCMD
PUSH B
OFF
MOV BIA
:10 OFF
w114o
DCR B
JNZ .10
MVI AI4
w114o
POP B
UPEC

i*******§**

iTHIS ROUTINE SENDS THE

iGET THE DATA
iSEND IT

i’EOT’

B REG. TO THE 11/40

iTHE 11/40’S REPLY IS DECODED AND THE
ICARRY SET IF IT IS NOT ACK/EOT

i******%*§*

<£ CKSUM
MOV AIB
W1140
MDEC
ORA A
RZ
SIC

;******%***

iSEND IT
iDECODE ANSWER
IACK/EOT? (ZERO)

ITHIS IS A ROUTINE WHICH HANDLES TRANSFERS
iBETWEEN A MICRO FLOPPY AND THE 11/40.
iIT SHOULD BE LOADED INTO THE COMMUNICATOR

;**********

<< PIP

:10 MVI AI44
CRT
LINE
RC
OFF
CPI 2
J2 .

MOV BIA

:14 OFF

DCR B

51$!
IGET A 2 LETTER COMMAND

i2 LETTERS?
iREMOVE IF NOT 2 LETTERS

 

 

 

180

JNZ .14
MVI AI77 i’?’
CRT
JMP .10
:20 ON iREPLACE # CHAR
MVI AI4 iSETUP COMMUNICATION
STA 16000
MVI AI6 i’ACK’
W1140
MDEC IWHAT RESPONSE?
ORA A IZERO?
CNZ .150 i**ERROR** NO ACK/EOT ON STAR
:30 MCMD ISEND COMMAND &DECODE RESP
CC .150 5**ERROR** NAK FROM 11/40
:151 CPI 0 iACK/EOT ?
JZ .10 iDONE
CPI 1 511/4O EXPECTS DATA
JZ .40
ggI 260 Ill/4O SENDING DATA
S§I 3100 511/40 WANTS FILESPEC
CALL .150 5**ERROR** UNKNOWN CODE
2153 JMP .10
i **********
iDATA FROM FLOPPY TO 11/40
; **§*******
:40 MVI AI114 i’L’
FILE
JC .40 iIF ERROR TRY AGAIN
:41 LXI BI1 iCHECKSUM=0IBYTE COUNT=1
LXI HI177750 5-30
DAD SP IFIND ADDRESS FOR DATA BUFFER
SHLD .46
42 CALL 1106 iDINP
JC .44 iEOF
DCX H
MOV MIA iSTORE DATA
INR C
JNZ .42 5377 MAX
244 MOV EIC
DCR C
32 .50 IALL DONE
MOV AIC
21140 iSEND BYTE COUNT
I 46 OI 0
DC H
:47 MOV AIM
W1140 ISEND DATA
DCX H
XRA B iFORM CHECKSUM
MOV BIA
DCR C iDONE?
JNZ .47
CKSUM iSEND & CHECK CHECKSUM
CC .150 i**ERROR** CHECHSUM ERROR
:152 XRA A ,
CMP E iZERO FOR CONTINUE
JZ .41
:50 XRA A
W114O iZERO BYTE COUNT TO TERMINATE
JMP .10
i *******%**
iDATA FROM 11/40 TO FLOPPY
; *§*******§
:60 MVI AI123 i’S’
FILE
JC .60 ITRY AGAIN
:62 R1140 iGET BYTE COUNT
ORA A iZERO?

 

 

181

JZ .70 iALL DONE
MOV CIA
MVI BIO iINITIALIZE CHECKSUM
:64 R1140 iGET THE DATA
CALL 1123 iDOUT
XRA B iFORM CHECKSUM
MOV BIA
DCR C iDONE? .
JNZ .64
R1140 iGET CHECKSUM
CMP B SMATCH?
32 .66
MVI AI25 iDON'T MATCH!
W1140 iSEND ’NAK’
MVI AI4 i’EOT’
W1140
CALL 1171 iEOTS - CLOSE FILE
CALL .150 i**ERROR** CHECKSUM ERROR
:66 MVI AI6 i’ACK’
W1140
MVI AI4 i’EOT’
W1140
JMP .62 iNEXT
:70 CALL 1171 iEOTS
JMP .10
,' **********
iFILESPEC FOR 11/40
5 **********
1100 MVI AI43 i’#’
CRT
LINE
JC .100
JMP .30 iGO SEND IT
i *********%
iERROR CODE ROUTINE
; **********
3150 LXI HI.160 iADDRESS TABLE
POP D iRETURN ADDRESS
MVI BI5 iMAX. # OF ERROR CODES
:156 MOV AIM iGET LOW BYTE ADDRESS
CMP E iMATCH?
JZ .157
INX H
INX H iPOINT TO NEXT ADDRESS
DCR B
JNZ .156
1157 MVI AI106 i’F’
CRT
MOV AIB
ADI 60 iCONVERT TO ASCII
CRT
JMP .10
{C DFILE iOPEN DATA FILE
PUSH PSW
:10 MVI AI123 i’S’
FILE iOPEN THE FILE
JC .10 iERROR
POP PSW
CC DATAF ISEND BYTE TO FLOPY
PUSH PSW
LDA 7
CALL 1123 iDOUT
POP PSW
{C AXIS IROUTINE TO DRAW AXIS
PUSH PSW
PUSH H
PUSH B

 

PUSH
XRA
STA
STA
MVI
STA
LXI
CALL
LXI
CALL
POP
POP
POP
POP
RET
121I3I2
21I31I2
MVI
PUSH
MVI
MOV
INX
MOV
INX
:14 STA
DCR
JNZ
DCR
JNZ
POP
DCR
9N2

\
J} J.

PUSH
PUSH
PUSH
PUSH
LHLD
MOV
INX
MOV
INX
PUSH
MOV
INX
MOV
INX
MOV
CMA
STA
XTHL
MOV
STA
MVI
STA
DAD
XTHL
POP
LHLD
MOV
INX
MOV

.bH

' EZII‘ lTl' OHIOIDNIGH-
PJ J) [U 3 34> “(13‘

H
0

LL ALL;
I: 0 ISO
00

‘ ‘ U"
[0‘ H
hm
HH
0‘0

-

I—I-Ho
O
O

‘ ‘O U]
33 2

~

IUImUIUImHIUGU

(It~
IUHPJI PJ 3 3 3

0 O

H [U

0“ 0“
O
0

cm

s

'UUJUIZIZHU UHDH) H D

U]
2

>13
‘ 0’)
NZ
\1

182

IgLEAR THE GRAPHICS
IY
iY AXIS TABLE

I41I31
I241I3
iLOW LEVEL PLOTTING ROUTINE
iLASTAD
iXINIT
iSAVE XINIT
iXINC

INEXT Y DATA
IY VALUE

5X VALUE

ITURN ON POINT
IINR X

ILASTAD
IUPDATE XINIT

iGRAPHICS ON UNTIL CTRL/C

 

183

STA 16500 ITURN ON GRAPHICS
:10 READ

CPI 3 ITEST FOR CTRL/C
JNZ .10

MVI A.17

STA 15500 ITURN ON VRAM
POP PSW

{C LOCIT iDRAWS LINE AT SPECIFIED X
MVI AI27
STA 16500 ITURN ON GRAPHICS
LHLD 10 ILASTAD

MOV ILOAD X POINT IN B&C
INX

MOV
INX
MOV
INX
MOV
LXI
:10 DAD
DCX
MOV
ORA
JNZ
MOV
STA
XRA
STA
LXI
MVI
:20 STA
DCR
JNZ
DSP IVRAM ON AT CTRL/C

I’d

iXINC

03 I! 3 3

CC

iXINIT

C

5X VALUE

iY VALUE
IC=5
ILINE DOWN

[04:09.) PJIO G!
O O
N H

OH
O

-o»>w»>HD-n)woxoxm1wxo
o-~o~ 0w»

m
0

#########################
MICRO 0

#########################

C. C. C. C.

(*)(*)(*)(*)(*)(*)(*)(*)

IGENERAL PURPOSE CONVERS-M
IROUTINES

(*)(*)(*)(*)(*)(*)(*)(*)

C. C. C. C. C. C. C. C.

.. ;‘C.C.C.C

MCHG
STRING
SEARCH

 

 

184

CLEAN

JC .20

POP PSW

ON

MCHG

JMP .5
:20 LHLD 36

MOV AIH

ANI 37

MOV HIA

POP PS

ON

MCHG

PUSH2

MOV AIB

ON

RMICRO

xx

CC SUB2 ;(D&E) = (D&E) - (H&L
PUSH
LDAX
SUB
STAX
INX
INX
LDAX
SBB
STAX
INX
JNC
DCX
XRA
STAX
INX
STAX

:10 INX

(I)
Z

'UUUUUIDU' IUZUUIUZU‘O
H
0

0')
C

{< DIV 316-BIT BY B-BIT
iSAVE STATUS AND REGISTERS

.0
C
U]
I
1)
[I]
Z

IGET DIVISOR

IGET DIVIDEND
iBIT COUNTER

C

H

:10 MOV

CC

iROTATE LOW BYTE

CC

IROTATE HIGH BYTE
iDECREMENT COUNTER

C

19' OI 2hr" DUDO ta COUI

‘LJ
UODIDF'DI‘J)

:20 SUB

HHHM
0000

:30 MOV iCOMPLEMENT HIGH BYTE

DI D” G“ U‘ 51'
I

F)

CMA ICOMPLEMENT LOW BYTE

 

 

18S

MOV LIA
PUSH2 IGUOTIENT ON STACK
MOV AID
ON IREHAINDER ON STACK
g8; IRESTORE STATUS AND REGISTERS
POP
POP
>>

{< DIV3 I24—BIT X 16—BIT
PUSH ISAVE STATUS AND REGISTERS
PUSH
PUSH
PUSH
POP2
MOV
MOV
POP2
MVI
STA
XRA
MOV

:10 MOV
MOV
RAL
MOV
INX
MOV
RAL
MOV
INX
MOV
RAL
MOV
DCX
DCX
LDA
DCR
JZ
STA
MOV
RAL
MOV
MOV
RAL
MOV
JNC
MOV
SUB
MOV
MOV
SBB
STC
CMC
JMP

:20 MOV
SUB
MOV
MOV
SBB
JNC
MOV
MOV
ADD
MOV
MOV
ADC
JMP

:25 000

:30 MOV
CMA

VIUU
(n
5:

U)
2

iGET DIVISOR

iLOW BYTE TO REG C

iHIGH BYTE TO REG B

IADDRESS OF 3-BYTE DIVIDEND
1 IBIT COUNT

D no (3011)
'0‘ ~\

DZDZDZDDUIL-JII"

CCC

IROTATE DIVIDEND LEFT ONE BIT

C

C

C

iMSB OF DIVIDEND INTO CARRY

C

N
U!

IDECREMENT BIT COUNT
IIF A=0I JUMP TO EXIT
IELSE STORE COUNT AND CONT

iMSB OF DIVIDEND INTO REM

'D' II: DIZ DIX DUND‘

‘ T00
"1010

CC

IJUMP IF BORROW ON SUBTRACT
iSUBTRACT LOWER BYTES

CC

EJDI’TIODiup pm D‘
UD ITIOD OD

iSUBTRACT HIGH BYTES

iLOOP
iSUBTRACT LOW BYTES

CH

CC

ISUBTRACT HIGH BYTES
iJUMP IF NO BORROW
IRESTORE D

iELSE ADD DIVISOR BACK

CCH

C C

' GDMODU‘ Roma)-

H

iAND LOOP
iSTORAGE LOCATION FOR BIT
IGET QUOTIENT AND COMPLEMENT

3 0 0D FIDO CD "10

D

 

 

186

MOV
INX
MOV
CMA
MOV
INx
MOV
CMA
MOV
INX
PUSH2
POP
POP
POP
POP

***§***************

OUTINE TO INPUT ASCII TO LASTAD
"' STRING

iTHE ABOVE EXAMPLE WOULD CAUSE THE
iASCII REPRESENTATION OF STRING TO
iBE STORED AT LASTAD PRECEEDED

iBY THE NUMBER OF CHARACTERS IN THE
ISTRING. THE STRING IS STORED FIRST
iCHARACTER FIRST.

i********************
i

C

DZDZD

C

iADDRESS OF NEXT DATA WORD
iRESTORE STATUS AND REGISTERS

1061 I: DI: DIZ

712! *V

0

IJ

3
'0
U)
2

CC

CC.C

CC " IINPUT ASCII TO LASTAD
MODE ".Soo ICHANGE TO RUN ONLY MODE
SESING IGET THE CHARACTER STRING
MOV CIA Ia CHAR.
LHLD 32 iLASTAD
MOV MIA
INx H
:10 OFF
MOV MIA iSTORE NEXT CHAR.
INX H
DCR C
JNZ .10
SHLD 32 IUPDATE LASTAD
CC BINDEC I16-BIT BINARY TO 5 DIGIT BCD
PUSH H
PUSH B
PUSH PSW
MVI C,5 IINITIALIZE BCD DIGIT COUNTER
MOV BIC
:10 SNI A,12 IDIVISOR FO 10
DIV IDIVIDE BINARY NUMBER BY 10
OFF IGET REMAINDER
PUSH PSW IAND SAVE
DCR C IDECREMENT BCD COUNT
JNZ .10 iJUMP IF MORE LEFT
POP2 IGET GUOTIENT
MOV CIB IREINITIALIZE BCD COUNTER
:15 POP PSW IRECOVER REMAINDER
CPI 0 518 IT A ZERO?
JNZ .25 iNOI REST OF DIGITS ARE SIGN
383 g iYESI DECREMENT COUNT
JNZ .15 iJUMP IF MORE TO DO
INR C
INR B
JMP .25

:20 POP PSW

 

 

 

187

:25 ON IPUT REMAINDER ON STACK
DCR C
UNz .20
MOV AIB IGET NUMBER OF NON—ZERO DIGIT
ON iAND PUT ON STACK
POP PSW
POP B
POP H
>>
CC DECBIN IDECIMAL TO BINARY CONVER
PUSH H
PUSH B
PUSH D
PUSH PSW
OFF IR CHARACTERS
MOV BIA
LXI H.o IINITIALIZE THE BUFFER
:10 MVI 0.0
OFF IGET A DIGIT
MOV EIA
CPI 12 518 IT BCD?
UNC .30 iJUMP IF ERROR
DAD D IADD THE DIGIT TO THE BUFF
UC .30 ITOO LARGE IF CARRY
DCR B ILAST DIGIT?
Uz .20 IUUMP IF DONE
MVI C.11 IIF NOT THEN BUFFER x 10
MOV EIL
MOV DIH
:15 DAD D IADD 10 TIMES FOR MULTIPLY
UC .40 ITOO LARGE IF CARRY
DCR C IAODED 10 TIMES?
UNz .15
UMP 1o IBACK TO GET A DIGITEE
:20 PUSH2 IPUT THE RESULT ON THE STAC
POP PSW
POP D
POP B
POP H
STC
g2? ICLEAR THE CARRY (NO ERROR)
:30 DCR B
:40 MOV AIB
ON I# CHARACTERS LEFT
POP PSW
POP D
POP B
POP H
STC ISET THE CARRY (ERROR FLAG)
<< DECOUT
PUSH PSW
BINDEC ICONVERT BINARY NUMBER TO DEC
SNI AII
ASCII ICONVERT TO ASCII CHARACTERS
MESSAG iOUTPUT
‘ POP PSW
CC DECIN IINPUT A DECIMAL #
PUSH PSW
:10 SgglNG A IINPUT ASCII STRING
ON IASCII =\ OCTAL MODE
ASCII ICONVERT TO BCD
DECBIN ICONVERT TO BINARY
Uc .20 IERROR FLAG SET?
POP PS

RET

 

 

 

188

:20 CLEAN
NHAT
UMP 10
>>
{4 RUN2
PUSH PSW
RUN 2
:10 FLAG 2
U2 .10
POP PSW

>>
{a

(*)(*)(#)(*)(*)(*)(*)(*)

'ROUTINES TO INTIIALIZE AND
'ACCESS DATA MEMORY

'(*)(*)(*)(*)(*)(*)(*)(*)

CCCCC.C.C.C.

C

C

{C MINIT
LXI H,o
SHLD 40050
SHLD 40052
SHLD 40054
SHLD 40032
SHLD 40070
LXI H.51000
SHLD 40100
LHLD 40010
MVI M.o

{{ DMEM
DECIN
DECIN
POP2
SHLD 40072
POP2
SHLD 40032
MOV E.L
MOV D,H
DAD D
SHLD .30
XCHG
LHLD 40010
SHLD .20
LXI H.77777

:10 MOV A.H
CPI 140
RC
XRA A
MOV MIA
DCX H
DCX D
MOV AID
ORA E
UNz .10
INX H
XCHG
41

20 0,0
MO AID
ANI 77
MOV MIA
INX H
MOV MIE
INX H
MVI M.o
SHLD .20
LXI H.40050

IREMOVE GARBAGE
ISEND A ’7’ TO TTY
ITRY AGAIN

IRUN VIDICON MICRO

iINITIALIZE DATA & CUT MEMORY
I#D
IfiC
IGPTS
IDPTS
iCPTS
iTDP OF CUT MEMORY
IFREE ADDRESS - CUTS
iLASTAD
IZERO DATASET MAP
IDMEM DPTS OVLP
IDPTS
IOVLP
iDVLP

iDPTS

I2 BYTES PER POINT

iLASTAD
iEND OF DATA MEMORY
iBOTTOM OF DATA MEMORY

ILAST POINT?

ILXI HI
iLASTAD TABLE

iREMOVE MICRO BIT

iSTORE LOW BYTE
iEND OF TABLE MARK
I#D (# OF DATASETS)

 

189

INR M
LHLD 40072 IOVLP (OVERLAP OF DATASETS)
XCHG
DCX H
DAD D
DAD D 2
21 -L
:30 0.0 'D
UMP .10

{{ MCUT iMCUT #
DECIN i#
POP2
MOV CIL
LHLD 40050 i#D (# OF DATASETS)
SHLD 40070 iCPTS (# OF POINTS PER CUT)
MOV EIL
MOV DIH
DAD
SHLD
XCHG
LHLD
XRA
210 CMP
JZ
INX
INX
JMP
:20 INX
SHLD .
LHLD
230 XRA
MOV
DCX
DCX
MOV
ORA
JNZ
INX
XCHG
41 TLXI H!
:40 OIO iADDRESS OF CUT MAP
MO iHIGH BYTE FIRST

INX
MOV ILOW BYTE
IMOVE END OF DATA MARK

YT
S

S PER POINT
'2

CC.

B E
XI D
PT *

C

i2 BYTES PER POINT
ITOTAL # OF BYTES

ILASTAD
iSKIP OVER DATASET MAP

b ' U
0 w
0 O
H

O

I'II'ZD
N
O

p»
O O

iADDRESS OF CUT MAP
0100 iFREE ADDRESS

D

U

iTEST IF ZERO

IQUUNJIXDD
O

C

INX
MVI
SHLD
LXI
INR iINC. # OF CUTS
XCHG
21 ILXI DI

:50 860 H I# OF BYTES
DCR C i# OF CUTS TO CREATE
JNZ .30
SHLD 40100 iNEW FREE ADDRESS

{C TAB SPARAMETER TABLE

AVGS 40034

AVGA 40056

INT 40040

PTS 40036

FREQ 40042

BEAM 40044

DELY 40046

DPTS 40032

CPTS 40070

GPTS 40054

SCNS 40062

#D 40050

#00 TO U

KITIKIKIZ
C‘hC

CCCCCCCCCCCC
CCCCCCCCCCCC

 

 

 

I \l
(‘I

:10
:40

:50

:60

:10

:30

\\\‘\\

O‘JOCCCCCC

O
I
m
X ‘Y’

‘0‘
3O

#- UN}
3 :30 (DH

C

C

U]

@0130 DIN' II' 00000' 3DU
Z

WUUU

U)
E

D ' DIDUI'U
HC
...:

- UP
30(1) 04:
(.J

‘ ‘0
300

U

D' G! DIET OD0‘
00

C

190

40052
40072
40074
40102
40104
0I0

IPARAM. NAME TO CRT

I# CHAR.

IRETURNS WITH PARAMETER

iSTACK

IADDRESS OF PARAMETER TABLE
HITAB
ISAVE STACK

ILXI

iEND OF TABLE
5% CHAR.

IRESTORE STACK
ILCMP

IMATCH

idUMP OVER ADDRESS
iGET THE ADDRESS
iGET RID OF STACK

ICLEAR CARRY - FOUND

ISET CARRY — NOT FOUND

iPARAMETER OUTPUT ROUTINE

HITAB PARAMETER TABLE

191

ON
MESSAG
MVI AI40 iSPACE
CALL .50
INX H
MOV EIM iADDRESS OF DATA
INX H
MOV DIM
INX H
PUSH H iSAVE POINTER
XCHG
MOV EI IGET THE DATA
INX H
MOV D
XCHG
PUSH2
DECOUT iOUTPUT THE PARAMETER
POP H
MVI AI11 iTAB
CALL .50
UMP .20
250 ON
MVI AI1
ON
MESSAG
RET
:60 POP B
POP D
POP H
139*” P
{{ SET iROUTINE TO SETUP VIDICON
PUSH H
PUSH P
:10 41I1I 52 5 1i *
PUSH2
MESSAG
STRING
CHECK
JC .10 IERROR
MOV AIH iALL DUNE
ORA L
d2 .20 iALL DONE IF ZERO
DECIN iGET THE VALUE
XCHG
PDP2
XCHG
MOV
INX
MOV
.JMP
:20 POP
’39P

II
'4

C

ISTORE IT
iNEXT PARAMETER

200 m

ID‘ZIZ
mH‘

(*)(*)(*)(*)(*)(*)(*)(*)

DATA FILE MANIPULATION
AND STORAGE ROUTINES

(*)(*)(*)(*)(*)(*)(*)(*)

C. C. C. C. C. C. C. C. C.

C

{C CFIND ICFIND NUMBER

PUSH PSW
PUSH H
PUSH D
:10 DECIN
LDA 40052 i#C (NUMBER OF CUTS)

POP2

 

CMP
JNC
WHAT
JMP
:20 XCHG
LHLD
XRA
:30 CMP
JZ
INX
INX
JMP
:40 INX
MOV
DCR
RAL
MOV
DAD
MOV
INX
MOV
XCHG
PUSH2
LHLD
SHLD
POP
POP
’39P

CC DFIND
PUSH
PUSH
PUSH

:10 DECIN
LDA
POP2
CMP
\JNC
WHAT
UMP

:20 MOV
DCR
RAL
MOV
MVI
LHLD
DAD
INX
MOV
DCX
MOV
ORI
MOV
LHLD
XCHG
DAD
DAD
PUSH2
LHLD
MOV
SUB
MOV
MOV
SBB
MOV
SHLD
POP
POP
{39"

{c TFINB

'F
M
O

a
h (D H
O (D O
p—C
O

w
0

C

mIgom DDI‘II‘ZD
33) m

40070
40054

PSW

PSW

40050

‘ F
C H m
PC) 0

or ~ 0“
SD03 3 co»
m

‘0
U
m

OD ZID ["0

0C

DJHUDJHJDFWUDJh 0C1 hCh‘DJHnICLhUFI DD“
0
h

m
2

192

ILASTAD
ISKIP OVER DATASET MAP

iPOINT TO CUT ADDRESS

iCUT ADDRESS TO STACK
iGPTS = CPTS

iDFIND NUMBER

INUMBER OF THE DATASET
ICURRENT NUMBER OF DATASETS

ILASTAD
IPOINT TO DATASET ADDRESS

iGET LOW BYTE OF ADDRESS

iHIGH BYTE OF ADDRESS
iMICRO ADDRESS 3

IOVLP (DATASET OVERLAP)

I2 BYTES PER POINT
IDATASET ADDRESS TO STACK
IDPTS (POINTS PER DATASET)

iGPTS = DPTS - OVLP

iTFIND TYPEINUMBER

 

 

PUSH
PUSH
:10 STRING
POP2
PUSH2
CLEAN
DCR
JZ
:20 WHAT
3MP
:30 MOV
CPI
JNZ
DFIND
JMP
:40 CPI
JNZ
CFIND
:50 POP
'39?

>

aa‘

LHLD
MOV
ORI
MOV
INX
MOV
LHLD
MOV
CMA
MOV
MOV
CMA
MOV
INX
CFIND
DECIN
POP2
DCX
XCHG
DAD
BAD
XCHG
POP2
XCHG
LDA

:10 PUSH
MOV
STAX
INX
INX
MOV
STAX
INX
DCX
DAD
DAD
POP
DCR
UNz
LHLD
BOLD

{C STORE
TFIND
SETUP
CALL
SETUP
LXI
MVI

A.H
10

00 I

Oh
mo
20
U1
0

IM

DflmeUUDUIOD
3

H m
0 E

40050
40070

1IDFILE
.60
1IDATAF
HI40032
BI44

193

iGET THE
IC = CUT

TYPE
(CFIND) I D = DATA

iSHOULD BE 1

ILASTAD
IREPLACE

5D&E = F
iDPTS (P

5138(C = '-
iINPUT D
iPOINT

iH
52

&L POI
BYTES

ID&E POI
iRD

ICUT DST POINT
MICRO ADDRESS BITS

IRST DATASET
OINTS PER DATASET)

DPTS
ESTINATION

NT TO SOURCE
PER POINT

NT TO DESTINATION

(# OF DATASETS)

iPOINT TO NEXT DATASET
i2 BYTES PER POINT

iCHECK I

iCPTS

F LAST DATASET

#D

ISTORE DATASET ON FLOPY
IDATASET ADDRESS
iOPEN THE FILE

IRUN IT

iDATA BYTE TO FLOPY
IPARAMETER STORAGE

;# BYTES

 

 

194

:20 MOV AIM IGET A BYTE
INX H
CALL .50 iSTORE IT
DCR B
JNZ .20
LHLD 40054 IGPTS
XCHG
POP2 IDATASET ADDRESS => H&L
:30 gay SIM ILOW BYTE OF DATA
CALL .50 iSEND LOW BYTE TO THE FLOPPY
TNX SIM IHIGH BYTE OF THE DATA
CALL .50 iSEND HIGH BYTE TO THE FLOPPY
DCX D iTEST IF END OF DATA
MOV AID
ORA E
JNZ .30
LXI HI1171 iSETUP 1IEOTS
SHLD 20004
CALL .60
MVI AI166 iREPLACE THE HLT
STA 20007
RET
:50 STA 20007 IDATA BUFFER
:60 RUN 1
:64 FLAG 1 IMICRO DONE YET
9.2 - 64
CC TSTORE ISTORE ALL DATASETS
SETUP 1IDFILE IOPEN THE FILE
CALL .60
SETUP 1IDATAF
LXI HI40032 iSTART OF PARAMETERS
MVI CI44
:10 TRX SIM iGET A BYTE
CALL .50 ISEND IT TO THE FLOPPY
DCR C
JNZ .10
LHLD 40072 iOVLP
MOV CIL
MOV BIH iB&C = OVERLAP
LHLD 40032 IDPTS
MOV AIL
SUB C
MOV LIA
MOV AIH
SBB B
MOV HIA
SHLD .30 I#POINTS = DPTS - OVLP
LHLD 40010 ILASTAD
SHLD .24 IADDRESS POINTER
:20 41 ILXI HIADDRESS POHINTER
:2 OIO IADDRESS POINTER
MOV AIM iHIGH BYTE OF DATASET ADDRESS
CPI 0 I0 IF END OF DATASETS
JZ 40 iEND OF DATASETS
ORI 140 IREPLACE MICRO ADDRESS BITS
T8: aIA iHIGH BYTE OF DATASET ADDRESS
MSV EIM iLOW BYTE OF DATASET ADDRESS
X H
SHLD .24 IREPLACE ADDRESS POINTER
XCHG
DAD B
DAD B iH&L = DATASET ADDRESS + 2
21 iLXI DI#POINTS
:30 0:0 I#POINTS
:34 MO AIM ILOW BYTE OF DATA

CALL .50 iSEND LOW BYTE TO FLOPPY

 

:40

:50
:60
:70

{c MAX

:10

:20

:30

:100

:110

INX
MOV
CALL
INX
DCX
MOV
ORA
UNZ
UMP
LXI
SHLD
CALL
MVI
STA
RET
STA
RUN
FLAG
dz

4' .r’

TFIND
LHLD
MOV
MOV
POP2
MOV
INX
MOV
XCHG
SHLD
SHLD
XCHG
LXI
INX
INX
DCX
MOV
ORA
UZ
CALL
UC
DCX
DCX
CALL
UC
DCX
DCX
MOV
STAX
INX
INX
MOV
STAX
UMP
LHLD
SHLD
LHLD
SHLD
RET
LDAX
CMP
RNZ
DCX
DCX
LDAX
CMP
INX
INX
RET
OIO

GI
03

‘mDUI'DI

NI
OO‘NU
OOHOA U

H
0
‘0

"0
U
S

" 01m 50%
NH 7
NH 3 3 I70

H 00
m
p

HH
00
O

3

C

‘UDUIUDUI

h'k
OHOHH
fiNHHO 3
0000

h m

UI3UUI 30

195

iHIGH BYTE OF DATA
ISEND HIGH BYTE TO FLOPPY

iTEST IF LAST POINT

iNEXT DATASET
iSETUP 1IEOTS

IREPLACE HLT ON MICRO 1

iDATA BUFFER

iMAX TYPEINUMBER

iGPTS

EB&C = GPTS

IH&L = ADDRESS
iMIN

IMAX

IHIGH BYTE OF MAX

IPOINT TO NEXT DATA POINT
iTEST IF LAST POINT

iCHECK FOR NEW MAX
iUUMP IF NEW MAX
iPOINT TO MIN

iUUMP IF NOT NEW MIN

IGET LOW BYTE

iGET THE HIGH BYTE

iMIN
iMAX

IMIN

- ...-C".-

BRIAB ngHa

- -- 17:2

 

:10

." A...
\ \

<. "C

.’-.’.
C \.

:10

TFIND
POP2

TFIND
PUSH2
RBACK

RCLEAR
PUSH
PUSH
PUSH
PUSH
LHLD
XCHG
POP2
MVI
MOV
XRA
MOV
INX
DCR
UNZ
DCX
MOV
ORA
UNZ
POP
POP
POP

TFIND
RCLEAR

CLR8K
PUSH
PUSH
LXI
XRA
MOV
INX
MOV
CPI
UC
POP
POP

UI'U

40054

I‘DI
I

'UIUI'

PSW
H

«>530
O
O
m
p

DCUIU

'UIUGT MDU‘ OI3DOw
H ‘ R]
O U 0

U]
2

PSW
H
HI60000

'UI' IUDI3D
HO‘
OOI

[D
Z:

iRO

iDS
iGP
iSR

;#
IDO

iCL

IGP
IDA

i2

ICL
iNE

196

IMAX
UTINE TO SUBTRACT TWO DATA SETS
T
TS
C
PTS.
NE
IBACK DSTISRC
I<DST) = (DST) - (SRC)
EARS SCAN MEMORY

TS
TASET ADDRESS
BYTES

EAR IT
XT LOCATION

ICLEAR TYPEINUMBER
IGET DATASET ADDRESS

ICLEARS ALL BK OF DATA MEMORY

ISTART OF DATA MEMORY

iEND OF DATA MEMORY

 

 

‘> ">
U 4

{C MPLOT

:10

:30

TFIND
POP2
SHLD
SETUP
LXI
PUSH2
XRA
MOV
INX
MOV
INX
MVI
LHLD
PUSH2
SHLD
DIV3
POP2
LHLD
MOV
INX
MOV
INX
LXI
INX
LDAX
MOV
INX
INX
LDAX
MOV
INX
SHLD
LHLD
SHLD
LXI
LXI
SUB2
LHLD
MVI
INR
MOV
RAL
MOV
MOV
RAL
MOV
UNC
MOV
INR
STA
LHLD
MOV
CMA
MOV
MOV
CMA
MOV
INX
41
0I0
MO
INX
MOV
INX
SHLD
XCHG
DAD
76

0

m
Uh
r
H

I

I.100

C

83I3I3D IH‘
OLIJ D D

U|\l

AN

HI40102

C

CHC CC

‘00
D ID F‘HO MOD ID F'

C C

C

Uta D0 D0" DD' I D!— pmm'

C

10212131
3 3

.24

CD

197

ISRC
IPLOT H
iBUFFER

IX SCAL

iGPTS

i250/GP
iREMOVE
ILASTAD
IXINIT

IX SCAL

IDESTIN
iMAX

iSCALE

iROTATE

iSRC

iDATA =
iMVI
iSHIFT

IPLOTTING ROUTINE

ANDLER

ING

TS
ADDRESS

= 0

E HIGH BYTES

ATION

= MAX - MIN

SCALE UNTIL CARRY

DATA - MIN

AISHIFT

 

 

:34

:40
:44

:50

ICC

‘C-

:10

C. C. C. C. C. C. C. C. C. C.

J-
\. \.

\

:10

.’.

.l .
\C‘.

:10

LOC

OPEN
PUSH
SETUP
RUN
FLAG
U2
‘3’?"

SHUT
PUSH
SETUP
RUN
FLAG
U2

{”9”

J-I'

.40
EIL
DIH

.34
AID

.60

50
IH

.20
IDSP

HA» FD'I

.64

1IAXIS

HH

.10

HHH3I3PJU

.10

(*)(*)(*)(*)(*)(*)(*)(*)

HIGH LEVEL VIDICON
SPECTROPHOTOMETER ROUTINES

(*)(*)(*)(*)(*)(*)(*)(*)

SW
IOPEN2

11‘ [00.11013
[nu-
£0

‘UJ
(DZ

‘0‘ IUI'JI'J'U
I
C
4
PJ

U)»
20

198

ISTA
iDST
IPLOT IT
iLXI
iPOINTS

iTURN ON

DST

HIPOINTS

GRAPHI

IROUTINE

iDRAW THE AXIS

iPLOT THE DATA

ILOCATES

CTRL/C ENDS

TO DRAW AXIS

SPECIFIED

IRELATIVE LOCATION (1-NPTS)

iMICRO 1

LASTAD

IOPEN THE SHUTTER

iCLOSE THE SHUTTER

VINIT
SETUP
89“?

C’ VERR

PUSH

PUSH

LHLD

XRA

CMP

UZ

PUSH2

BINOUT

POP

POP
33

{C SCAN
SETUP
RUN2
YERR
{{ LAG
DECIN
POP2
MOV
STA
LXI
MVI
DECIN
XCHG
POP2
XCHG
MOV
INX
DCR
UNZ
SETUP
RUN2
VERR

CC
‘5')

r' ‘f
\ \.

:10

:10

C. C. C. C.

MICRO
-<.< BOOT4
MVI
STA
MVI
LXI
LXI
LDAX
MOV
INX
INX
DCR
UNZ
RET
LXI
CALL
EI
HLT
100050
100016
105300

0
0

:10

:30

2ISCAN2

AIL
40040
HI40064
CI3

2I LAG2

#########################

0

#########################

000

- - C M.
‘ HI—A .h
00C
00

‘ $01300]:me
D

y-A
0

SPI3777

HHHHH
\IO‘JDNO

199

IINIT

iVIDICON ERROR STATUS

IPRE-

IINT
IDATA

IMICR

LASTAD
LINKAD
SOFTWAR
END OF

IALIZE VIDICON

IOUTPUTS VIDICON ERR

ITAKE A SET OF SCANS

ILAG PRE-INTIPOST-INTI

INTEGRATIONS

STORAGE

LIZE MICRO #4

IA

CROS
17 BYTES

T OF MICRO 4

O #4

E STACK
LIBRARY

TIMES TO FLUSH

 

 

V0

>
BOOT4
/KI

200

520 - # VEL. PTS

I##4####fi#######fi#######fi#

MICRO

4

I##4######################

{C DRIVE4
PUSH
XRA
STA
POP
P)

{C FILL4
PUSH
MVI
STA
E’PP

{C FLOW4
PUSH
MVI
STA
‘39P

{< WAIT4
PUSH
PUSH
PUSH
MVI

:10 MVI

:20 MVI

:30 DCR
UNZ
DCR
UNZ
DCR
UNZ
POP
POP
POP
>>

{< PUSHIT
FILL4
WAIT4
DRIVE4
WAIT4
FLOW4

{C EMPTY
MVI

:10 WAIT4
FLOW4
WAIT4
DRIVE4
DCR
.sz

CC FLUSH4
PUSH
PUSH
LDA
MOV

:10 WAIT4
PUSHIT

PSW

A
16000
PSW

‘UHD‘U
0
O

[00“ (.0
200.12

"U
U)
Z

AI4
16000

'0

U) U)

(.0th Z 8
\IV
\IV

Q ~\\
0

'UGJU' 0' Cd' OOUUUU'U
H N
O O

U)
E

D

.10

DUI-A511)
‘ \l U)
D Z

iRETURN STOPPED FLOW

iFILL S.F. SYRINGES

IRUN A PUSH ON S.F.

ILONG WAIT (APPROX. 3

iFILL THE SYRINGES AND DO A S.F. PUSH
IFILL THE SYRINGES

IPRESSUREIZE THE SYRINGES
IDO THE PUSH
iEMPTY THE DRIVE SYRINGES

iFLUSH OUT THE S.F.

iTIMES TO FLUSH

iTAKE A PUSH

201

WAIT4
DRIVE4 IEXHAUST TO WASTE
EMPTY IEMPTY THE SYRINGES
DCR B
UN: .10
POP B
POP PSW
>>
{C CLEAR1 ICLEAR CLOCK # 1
PUSH PSW
MVI AI377 IDONN COUNTERS
STA 12000
STA 12000
POP PSW
CC CLEAR2 ICLEAR CLOCK # 2
PUSH PSW
MVI A.377 IDOWN COUNTERS
STA 12001
STA 12001
POP PSW
CC VEL§R A IRECORD A TIME PROFILE
A
STA 20 I» POINTS = 0
MVI AIeo IINIT COUNTER # 1
STA 12003
MVI AI160 IINIT COUNTER # 2
STA 12003
CLEARI ICLEAR COUNTER # 1
CLEAR2 ICLEAR COUNTER # 2
LHLD 10 ILASTAD
MOV AIH
ANI 37 IREMOVE MICRO ADDRESS
MOV HIA
LXI D.2o IADDRESS OF THE 4 OF COUNTS
LDA 16000 INATCH POSITION FLAG
RAR IBIT 0 IS THE FLAG
UNC .20 iWATCH COUNTER # 2
:10 LDA 16000 IPOSITION FLAG
RAR
UC .10
LDA 12000 IGET THE LOW BYTE
MOV MIA ISTORE IT
INX H
LDA 12000 IGET THE HIGH BYTE
MOV MIA ISTORE IT
INX H
XCHG
1939 M IINC. THE 5 OF POINTS
CLEAR1 ICLEAR COUNTER # 1
:20 R23 15000 IPOSITION FLAG
UNC .20
LDA 12001 IGET THE LOW BYTE
MOV MIA ISTORE IT
INX H
LDA 12001 IGET THE HIGH BYTE
MOV MIA ISTORE IT
INX H
XCHG
IDEG M IINC. THE # OF POINTS
CLEAR2 ICLEAR COUNTER # 2
UMP .10 iWATCH COUNTER # 1

C ..
J’)

##3##################8###

C. C. C. C.

 

 

MICRO

202

O

I#########################

l
i-
<<

{C

:10

A
A

A
Al

A
A

:10

1C

STOP ISTOP ALL SLAVE MICROS
PUSH PSW
XRA A
STA 16000 IOUTPUT
STA 46 iOUTPUT SAVE
POP PSW
RUN4 IRUN MICRO #4
RUN 4
FLAG 4
U2 .10
LDA 46
ANI 77 ICLEAR RUN BIT
STA 16000
STA 46
DRIVE IRETURN STOPPED-FLOW
PUSH PSW
SETUP 4IDRIVE4
RUN4
POP PSW
WASTE
DRIVE IRETURN TO DEFAULT
4’.)
FILL IFILL S.F. SYRINGES
PUSH PSW
SETUP 4IFILL4
RUN4
POP PSW
FLOW IRUN A PUSH ON S.F.
PUSH PSW
SETUP 4IFLOW4
RUN4
POP PSW
FLUSH iFLUSH THE S.F.
BININ I4 OF TIMES TO FLUSH
POP2
MOV AIL
STA 100017
SETUP 4IFLUSH4
39N4
NAIT IA CALL HERE CAUSES 99.999
PUSH H
PUSH PSW
41,367.50 iLXI HI367I50
SHLD .100
LXI H..100
MVI AI3S7
STA .100
DCR M
UNz .20
INX H
DCR M
UNz .10
NOP
NOP
POP PSW
POP H
RET
0 9x
VEL IRECORD A TIME PROFILE

203

CFIND IINPUT DATASET TO STORE IT IN
SETUP 4IPUSHIT
RUN4 iSTART THE PUSH
SETUP 4IVEL4
RUN 4 iSTART TAKING DATA
MVI AI12 iWAIT 1 SEC TO COLLECT DATA
:10 WAIT
DCR A
UNZ .10
STOP
SETUP 4IEMPTY
RUN4 IEMPTY THE DRIVE SYRINGES
LDA 100020 5% POINTS
STA 40070 ICPTS
STA 40054 iGPTS
MOV CIA
MVI BIO IDON’T CAUSE TROUBLE
:gfig iGET DATASET ADDRESS (FROM CFIND)
LHLD 100010 iSTART OF DATA
:20 MOV AIM iGET A BYTE
CMA iTHEY’RE DOWN COUNTERS
STAX D ISTORE THE DATA
INX H
INX D
MOV AIM iHIGH BYTE
CMA
STAX D
INX H
INX D
DCR C IDONE?
UNZ .20
MVI AI1
STA 40056 iAVGA = 1
STA 40034 IACGS = 1

J7 a)

I#4###4######4444444#44444
' MICRO 4
2#########################

<< VCNT
LDA 2o
MOV C.A
LDA 16000
RAR
UNC .20
:10 LDA 16000
RAR
UC .10
DCR C
R2
:20 LDA 16000
RAR
UNC .20
DCR C
Rz
UMP .10
J’s"
<< VSF4
PUSHIT ITAKE A PUSH
VCNT iWAIT FOR THE TRIGGER POINT

i

;##################8######

I
I

I
,I- r
C“.

:10

:30

:40

:60

MICRO

VSF2

LDA
MOV
LDA
MOV
LDA
MOV
MVI
LHLD
MOV
ANI
MOV
STA
LDA
RAL
UNC
STA
MVI
STA
MVI
MSCAN
DCR
U2
CMP

UZ
PUSH
MVI
STA
LDA
RAL
UNC
STA
DCR
UNZ
POP
UMP
ENDS

{C SFMAXE

:10

i

I##4######################

I

I#########################

I

LHLD
LDA
MOV
MVI
LDA
MOV
STA
LDA
RAL
UNC
STA
DCR
UNZ
MVI
STA
MVI
MSCAN
ENDS

C C
J) In"

MICRO

2

'##4######################

50
DIA

[.0010~
JP‘ «5
D

HuIijI-som

00
HH-
HO

0H0

HOFJOO OOD I OD
H H-
LJ 0

~ru~ 4»- A» \r o~ ~

HHDU‘ mGU >H1>w
o

I}
p

H
.3)
H0030 OO OOD OD

H
00
HH-
HO

O
C.A
O

‘IU‘H DH 4h
OH
H
(J

DHD' NH'

0

204

i# SCANS TO TAKE
5# SCANS TO SKIP
i# AVERAGES

;# INT. = 0
iDATASET POINTER

IREMOVE MICRO BITS
ICLEAR SCAN FLAG
iREAD SCAN FLAG
iCLEAR SCAN FLAG
iSET F.A. SCAN BIT
iTAKE 1 SCAN

ILAST SCAN?

iSKIP = O?

ISTOP F.A.
iREAD THE SCAN FLAG

ICLEAR SCAN FLAG
iLAST SCAN TO SKIP?

iDATASET POINTER
iNUMBER OF AVERAGES

ifi INT. = 0
I# SCANS TO SKIP

iCLEAR SCAN FLAG
iREAD THE SCAN FLAG

iCLEAR THE SCAN FLAG
ISKIP?

iENABLE F.A.

51 SCAN

205

I
{C SFSgAgP iSFSCAN # SKIP:TRIGGER.#
DECIN 5% TO SKIP
POP2
MOV A:L
STA 40064
DECIN iTRIGGER POINT
POP2
MOV AIL
STA 100020
FLUSH iFLUSH IT
SETUP 2IVSF2
SETUP 4IVSF4
RUN4 iRUN THE STOPPED-FLOW
QUE? iTAKE THE SPECTRA
SETUP 4IEMPTY iEMPTY THE SYRINGES
sg~4

<< SFMSTOP ;SFMAX # SKIP.DATASET.TRIG
DECIN 5% SKIP
POP2
MOV A.L
STA 40064

:10 DECIN 5DATASET #
LDA 40050 FCURRENT # 0F DATASETS
POP2
CMP
JNC
WHAT
JMP

:20 NOV
DCR
RAL
MOV
MVI
LHLD
DAD
MOV
ANI
MOV
SHLD
DECIN iTRIGGER POINT
POP2
MOV
STA
FLUSH
SETUP FMAX2
SETUP
RUN4 iRUN THE STOPPED-FLOW
Sggg ;TAKE THE SCAN
gaggP 4,EMPTY ;EMPTY THE SYRINGES

Ir, J,

'l—
‘Hm
[—00

0-

iDATASET TABLE

iREMOVE MICRO ADDRESS BITS

bIVDUbUm DD'

0‘ \1‘

OD I 00D
H
0

\l
0~

iFLUSH IT

bl‘J PD
‘ h on
(U) 0!"
U] 0
'11 [U
b O

206

B . Minicomputer Programs

Cit-*4?***********§**********

CONV2.FTN

THIS ROUTINE IS USED
TO CONVERT STANDARD
BINARY FILES FROM THE
FIEESDN SYSTEM INTO ASCII

*****fi********§**********

BYTE IFILE(24);OFILE(24)
INTEGER PARAM(20):COUNT,DATA<1024)
10 NRITE(6.900)
FORMAT(’ INPUT FILE:'/)
READ(6I905)IFILE
905 FORMAT(24A1)
IERR=0
CALL FORPIP<IFILE;IERR)
IF(IERR)20,30:40

0000000000

0
O
O

20 HRITE(6.904)

904 FORMAT(’ FILE ERROR’)
GOTO 10

30 WRITE(6;906)

906 FORMAT(’ FILE NOT FOUND’)
GOTO 10

40 CALL UNFMO<IERRIIFILE,0.1)

IF(IERR.EG.O)GOTO 44
WRITE(6:907)

907 FORMAT(’ UNFMO ERROR’)
GOTO 10

44 WRITE(6.910)

910 FORMAT(’ OUTPUT FILE:’/)
READ(6;905)OFILE
IERR=O
CALL FORPIP(OFILEIIERR)
IF(IERR)50.70,60

50 NRITE(6:904)

GOTO 44

60 WRITE(6.915)

915 FORMAT(’ FILE ALREADY EXISTS’)
GOTO 44

70 CALL ASSIGN(4.0FILE.24.IERR)

IF(IERR.EG.O)GOTO 80
NRITE(6.920)

920 FORMAT(’ ASSIGN ERROR’)
GOTO 44

SO COUNT= 36
CALL UNFMRN
AVG=(FLOAT(
WRITE(6 925

925 FORMAT(’ XI
READ(6 927)

927 FORMAT(2F15.
COUNT= 2048
CALL UNFMRN( RRIDATAICOUNT)
COUNT=COUNT/2
DO 100 I=11COUNT
Y=FLOAT<DATA(I))
IF(Y.GE.0)GOTO 9O
Y=65536.0 + Y

90 Y=Y/AVG
WRITE(4:930)X;Y

930 FORMAT(’RD’,2015.5)
X=X+XINC

100 CONTINUE

AR AM COUNT)
)) )*(FLOAT(PARAM(11)))

(
P
)
N C:’/)
X

207

CALL UNFMC(IERR)
ENDFILE 4
NRITE(6.940)
940 FORMAT(’ CONVERSION COMMPLETE’)
GOTO 10

END
*******************%**********

ERIK CARLSON 19-dULY-197S

THIS ROUTINE ALLONS EVALUATION
OF DATA FROM THE STOPPED—FLOW
VIDICON RAPID SCANNING SYSTEM.
THE OUTPUT OF THIS ROUTINE

CAN BE FILES CONTAINING EITHER
ABSORBANCE VS TIME OR ABSORBANCE
VS WAVELENGTH.

if'31.i-'21-'11-*fl-{ufi-fl'**************%*****
BYTE IFILE(24).OFILE(24)
INTEGER RPTSIDPTS;NPTS.TSCNS
INTEGER PARAM(20);COUNT.DATA(1024)
DIMENSION REF(512):DARK(512).NORK(512)
##########

FILE INPUT SECTION FOR REFERENCE AND DARK SIGNALS

##########
0 NRITE(6.900)
00 FORMAT(’ REFERENCE’)
CALL INPUT(REF:RPTS)
NRITE(6I9O2)
902 FORMAT(’ DARK SIGNAL’)
CALL INPUT(DARK;DPTS)
IF(RPTS.EQ.DPTS)GOTO 20
NRITE(6I904)
904 FORMAT(’ NUMBER OF POINTS NOT EOUAL!!’)

0000000000000

C
C
C
C
C
1
9

GOTO 10

20 DO 25 I=11RPTS

25 REF(I) = REF(I) - DARK(I)

22 NRITE(6.906)

906 FORMAT(’ INPUT FILE ’/)
READ(6;908)IFILE

908 FORMAT(24A1)
IERR=0
CALL FORPIP<IFILE:IERR)
IF(IERR)30:40,50

30 NRITE(6:910)

910 FORMAT(’ FILE ERROR’)
GOTO 22

4O WRITE(6:912)

912 FORMAT(’ FILE NOT FOUND’)
GOTO 22

50 CALL UNFMO(IERRIIFILE,0,1)
IF(IERR.EG.0)GOTO 60
NRITE(6.914)

914 FORMAT(’ UNFMO ERROR’)
GOTO 22
60 COUNT=36

CALL UNFMRN(IERR:PARAMICOUNT)
AVG=<FLOAT(PARAM(2)))*(FLOAT(PARAM(11)))
NPTS=PARAM(1)-PARAM(17)

TSCNS=PARAM(S)

IF(NPTS.EG.RPTS)GOTO 70

NRITE(6:904)
GOTO 80
g####%#####

g SELECTION OF TYPE OF DATA ANALYSIS
C##########

208
70 “RITE(6I916)
916 FORMAT(’ 1=ABS VS NAV: 2=ABS VS TIME. 3=END’/)
READ<6I918)IERR
918 FORMAT(IS)
GOTO(100.200)IERR
80 CALL UNFMC(IERR)
GOTO 10
g####fi#####
g ABSORBANCE VS WAVELENGTH
C##########

100 WRITE<6;920)

920 FORMAT(’ SCAN# SKI
READ(6I922)ISCA ANIIS

922 FORMAT(2I6.F10.0)

THRESHOLD:’/)

53. HR

X‘O

NRITE(6I924)

924 FORMAT(’ INIT. NAV.. NAV. INC :’/)
READ(6I926)XINIT:XINC

926 FORMAT(2F10 O)
IF(TSCNS.GE. ISCAN)GOTO 110

NRITE(6.928)TSCNS
928 FORMAT(’ SCAN# GREATER THAN TOTAL OF:’;I7)
GOTO 100
110 DO 120 I=1,ISCAN
COUNT=NPTS§2
CALL UNFMRN(IERR.DATA,COUNT)
120 CONTINUE
DO 130 I=ISKIPINPTS
XY=FLOAT(DATA(I))
IF(XY.GE.O O)GOTO 180
XY=65536.0 + XY

130 NORK(I)=(XY/AVG)-DARK(I)
8##########
g CALCULATE THE ABSORBANCES
C##########
500 DO 560 I=ISKIP.NPTS
R=REF(I)
S=WORK(I)
IF((R- -S)-THR)550:540:54O
540 T=S/R
IF(1. 0- T)550;550.545
545 IF(T) )550.550;547
547 WORK( =-1 O*ALOG10(T)

GOTO 560
550 NORK(I)=0.0
560 CONTINUE

GOTO 600

g##########

8 ABSORBANCE VS TIME

C##########

200 WRITE(6:9 930)

930 FORMAT(’ POINT#: THRESHOLD
READ(6;932)IPT.THR

9S2 FORMAT(I6IF10. 0)
NRITE(6.9S4)

934 FORMAT(’ INIT. TIME: TIME INC.:’/)
READ(6I926)XINIT.XINC
DO 220 I=1.TSCNS
COUNT=NPTS*2
CALL UNFMRN(IERRIDATA;COUNT)
XY FLOAT(DATA(IPT))

IF(XY.GE.0. 0)GOTO 220
XY=65536.0 + XY
220 NORK(I)=(XY/AVG)-DARK(IPT)
8##########

 

 

209

g CALCULATE THE ABSORBANCES

C##########
R=REF(IPT)
DO 460 I=1.TSCNS
S=NORK(I)
IF((R— S)-THR)450,440144O
440 T=S/R

IF(1. O-T)450,450,445
445 IF(T) )450;4501447
447 NORK( I)=—1.0*ALOG10(T)
GOTO 460
450 NORK(I)=0.0
460 CONTINUE
ISKIP=1
NPTS=TSCNS
##########

C
C
8 FILE OUTPUT SECTION

C##########

600 NRITE<6I936)

936 FORMAT(’ OUTPUT FILE:’/)
READ(6;908)OFILE
IERR=O
CALL FORPIP(OFILE.IERR)
IF(IERR)650:670;660

650 WRITE(6.910)
GOTO 600

660 NRITE(6.938)

938 FORMAT(’ FILE ALREADY EXISTS’)
GOTO 600

670 CALL ASSIGN(4IOFILEI24IIERR)
IF(IERR.EG.O)GOTO 700
NRITE(6.940)

940 FORMAT(’ ASSIGN ERROR’)
GOTO 600

700 DO 720 I=ISKIP;NPTS
NRITE(4.942)XINIT;WORK(I)

942 FORMAT(’RD’.2G15.5)

720 XINIT=XINIT+XINC
ENDFILE 4
CALL UNFMC(IERR)
GOTO 50

END
*******%************%%********

ERIK CARLSON 27*dULY-1978
LSLOPE.FTN

THIS ROUTINE CALCULATES THE

SLOPE OF A SET OF DATA IN

A STANDARD DATA FILE AND

OUTPUTS A NEW DATA FILE

OF SLOPE VS X—DATA.

THE SLOPE IS CALCULATED BY

A LEAST-SQUARES FIT TO A

PORTION OF THE DATA.
ROUTINES NEEDED:

FORPIP
LEASTS

******§-***********************

BYTE IFILE(24)
INTEGER COUNT
DIMENSION X(200).Y(200).Z(200)
10 NRITE(6.900)
900 FORMAT(’ INPUT FILE:’/)
READ(6:901)IFILE

00000000000000000000

 

 

901

20
902

30
903

40

so
904

80

90
908

100

909
120

00000000000000

210

FORMAT(24A1)
IERR=O
CALL FORPIP(IFILE:IERR)
IF(IERR)20.30.4O
WRITE(6:902)
FORMAT(’ FILE ERROR’)
GOTO 10
WRITE(6:903)
FORMAT(’ FILE NOT FOUND’)
GOTO 10
CALL ASSIGN(4IIFILE124IIERR)
COUNT=1
READ(4.904.END=60)X(COUNT).Y(COUNT)
FORMAT(2X.2015.5)
COUNT=COUNT+1
GOTO 50
ENDFILE 4
COUNT=COUNT—1
WRITE<6;920)COUNT
FORMAT(’ TOTAL PTS:’:I5:/’ START OF FIT. END OF FIT
1; FIRST POINT. WIDTH OF FIT:’/)
READ(6,922)IFS,IFE:IPT.IWID
FORMAT(4I5)
JPT=IPT
COUNT=0
IE=IFS+IWID
CALL LEASTS(X(I)IY(I),5;SLOPE:YINIT:CORR)
DO 220 I=IFS.IE
( (I).Y(I);1.SLOPE;YINIT.CORR)
I

)IY(I):4;Z(IPT):YINIT;CORR)

CALL LEASTS X
CONTINUE

CALL LEASTS(X(
COUNT=COUNT+1

IPT=IPT+1

IFS=IFS+1
IF(IPT.LE.IFE.AND.IE.LT.IFE)GOTO 200
WRITE(6,907)

FORMAT(’ OUTPUT FILE:’/)
READ(6,901)IFILE

IERR=0

CALL FORPIP<IFILE;IERR)
IF(IERR)80,100.90
WRITE(6.902)

GOTO 75

WRITE(6;908)

FORMAT(’ FILE ALREADY EXISTS’)
GOTO 75

CALL ASSIGN(4; IFILE: 24; IERR)
IE=JPT+COUNT-1

DO 120 I= JPTIIE

WRITE(4 909)X(I).Z (I)
FORMAT(’RD’; 015

CONTINUE

ENDFILE 4

GOTO 10

END

*************************

TFREQ.FTN

THIS ROUTINE CONVERTS FILES
OF RELATIVE INTENSITY VS.
FREQUENCY TO FILES WHICH
HAVE BEEN CORRECTED FOR THE
VARYING EXPOSURE TIMES.

THIS IS DONE BY SIMPLY
MULTIPLYING THE ORIGINAL
INTENSITY BY THE FREQUENCY.

************************4?

BYTE IFILE(24)
INTEGER COUNT

J

“:152-31V ..' f“

 

 

10
900

901

20
902

30
903

40
50
904
60

75
907

80

90
908

100

0fn30ffl50fflfi0fH30

(fit-4
0

900
905

20

211

DIMENSION X(200):Y(200)
WRITE(6.900)

FORMAT(' INPUT FILE ’/)
READ(6»901)IFILE

FORMAT(24A1)

IERR=O

CALL FORPIP(IFILE:IERR)
IF(IERR)20:30.4O

WRITE(6:902)

FORMAT(’ FILE ERROR’)

GOTO 10

WRITE(6:903)

FORMAT(’ FILE NOT FOUND’)

GOTO 10

CALL ASSIGN(4;IFILE:24;IERR)
COUNT=1
READ(4:904;END=60)X(COUNT).Y(COUNT)
FORMAT(2X;2G15.5)

COUNT=COUNT+1

GOTO 50

ENDFILE 4

COUNT=COUNT-1

WRITE(6:907)

FORMAT(’ OUTPUT FILE ’/)
READ(6:901)IFILE

IERR=0

CALL FORPIP(IFILE.IERR)
IF(IERR)80;100.90

WRITE(6,902)

GOTO 75

WRITE(6:908)

FORMAT(’ FILE ALREADY EXISTS’)

GOTO 75

CALL ASSIGN FILEl24IIERR)
DO 120 I=1
Y(I)=Y(I)*
WRITE(4.9O
FORMAT(’RD
CONTINUE
ENDFILE 4
GOTO 10
END

**§**********************

NOISE2.FTN

THIS ROUTINE IS USED TO
STUDY NOISE IN THE VIDICON
SYSTEM. IT TAKES A STANDARD
VIDICON BINARY DATA ARRAY
FILE AND CALCULATES THE
RELATIVE STANDARD DEVIATION
OF THE DATA AT EACH CHANNEL.

******%******************

BYTE IFILE<24)

INTEGER PARAM(20).COUNT;DATA(1024)
DIMENSION XDATA(3.1024)

DO 5 1:153

DO 5 d=1.1024
XDATA(I;J)=0.0
WRITE(6.900)

FORMAT(’ INPUT FILE:’/)
READ<6I905)IFILE;IERR
FORMAT(24A1.Q)

IERR=0

CALL FORPIP(IFILE:IERR)
IF(IERR)10.10;2O

CALL UNFMO(IERR;IFILE,0;1)
IF<IERR.NE.0)GOTO 10

COUNT=36

212
CALL UNFMRW(IERR:PARAM,COUNT)
AVG=(FLOAT(PARAM(2)))*(FLOAT(PARAM(11)))
NPTS=PARAM(1)-PARAM(17)
NSCNS=PARAM(8)

YAVG=FLOAT(NSCNS)

DO 30 J=11NSCNS
COUNT=NPTS*2

CALL UNFMRW<IERR.DATA.COUNT)
DO 30 I=1INPTS
X=FLOAT(DATA(I))

30 XDATA(1:I)=XDATA(1.I)+(X/AVG)
DO 50 I=1INPTS
50 XDATA(1.I)=(XDATA(1.I))/YAVG

CALL UNFMC<IERR)
CALL UNFMO(IERR:IFILE;0.1)
COUNT=36

CALL UNFMRW<IERR:PARAMICOUNT)
DO 60 J=11NSCNS

COUNT=NPTS*2

CALL UNFMRW(IERR;DATAICOUNT)
DO 60 I=11NPTS
X=(FLOAT(DATA(I)))/AVG

6O XDATA(2.I)= XDATA(2.I)+(X-XDATA(1.I))*(X-XDATA(1.I))
DO 80 I=1 NPTS
XDATA(2:I)=S SGRT((XDATA(2.I))/(YAVG-1.0))
80 XDATA(3;I)=X DATA(2.I)/XDATA(1.I)
DO 90 I=1.3
85 WRITE(6.910)I
910 FORMAT(’ OUTPUT FILE’:I3:’:’:/)
READ(6 905)IFILE;IERR
IF(IERR. E0. O)GOTO 90
IERR= 0
CALL FORPIP(IFILE;IERR)
IF(IERR)85:88;85
88 CA ALL ASSIGN(4;IFILE.24IIERR)
IF (IERR. NE. O)GOTO 85
DO 89 J=1.N PTS
X= FLOAT(J)
=XDATA(I.J)
WRITE(4 920)X Y
920 FORMAT(’ RD’.2 G15. 5)
89 CONTINUE
ENDFILE 4
9O CONTINUE
GOTO 10
END
C*************************
C
g VELC FTN
C THIS ROUTINE IS USED
C TO PERFORM VARIOUS TYPE OF
C DATA MANIPULATIONS ON THE
C DATA FROM THE VIDICON STOPPED-
C FLOW POSITION/VELOCITY TRANSDUCER.
C THE DATA CAN BE OUTPUT AS
C VELOCITY VS DISTANCEIVELOCITY VS.
8 TIME OF DISTANCE VS TIME.
C*************************

BYTE IFILE(24)
INTEGER COUNT
DIMENSION X(200):Y(200);XO(200):YO(200)
WRITE(6.906)

906 FORMAT(’ LINES PER INCH:’/)
READ(6:905)XLN

905 FORMAT(F10.0)
DX=2.54/XLN

10 WRITE(6:900)

900 FORMAT(’ INPUT FILE:’/)
READ<6:901)IFILE

901

20
902

30
903

40
so
904
60

65
910

911

200
210
300

310
400

FORMAT(24A1)

IERR=0

213

CALL FORPIP<IFILEIIERR)
IF(IERR)20.30:40
WRITE(6.902)
FORMAT(’ FILE ERROR’)

GOTO 10

WRITE(6.903)
FORMAT(’ FILE NOT FOUND’)

GOTO 10

CALL ASSIGN(4:IFILE:24:IERR)

COUNT=1

READ(4.904:END=60)X(COUNT).Y(COUNT)
FORMAT(2X.2G15.5)
COUNT=COUNT+1

GOTO 50
ENDFILE 4

COUNT=COUNT—1
WRITE(6:910)
FORMAT(’ 1=V/D:

READ(6:911)ITYPE

FORMAT

GOTO(
GOTO

' U|A>< 00A

.4
U)
m
0
ll
'maa<
"UJAHO UUJAHO DAD-c OH
XMHH
*OVH

(MDT! 0 flHmH II

A II II AOO\I II II AOOV II I! OHM
"'10 OOH HOOVOVH

ox4<

I

FORMAT(’
READ(6I9
IERR=O
CALL FOR
IF(IERR)
WRITE(6.
GOTO 75
WRITE(6.9
FORMAT(’
GOTO 75
CALL ASSIG
DO 120 I=1
WRITE(4:90
FORMAT(’RD
CONTINUE
ENDFILE 4
GOTO 65
END

l
/
1

"U VCVflI \‘
0h1'fl1
OTIOHC

£2
8
I
N

I

7

COUNT
1000.0) + TSEC

000000.0)/Y(I)

C
1
D
)
T
I
(
1
)

)
LE ALREADY EXISTS’)

(
)

I

OUNT
000.0) + TSEC

X

4
O
X
2

U
0
G

~H r4
m

L
9

IF
NT
(I
15

2=V/T. 3=D/T.

FILE:’/)

EIIERR)
0

ILEI24IIERR)

);YO(I)
.5)

4=END’/)

 

 

REFERENCES

 

10.

ll.

12.

13.

14.

15.

16.

17.

18.

REFERENCES

R.E. Dessy, Anal. Chem., 39, 1100A (1977).

C.A. Korn, Computer Desigp, 177 (June, 1977).

W.J. Lennon, in Computer Networking and Chemistr , P. Lykos,
Ed., Amer. Chem. Soc., Washington, D.C., 1975, pp. 67-84.

P. Day, in Computer Networking and Chemistry, P. Lykos, Ed.,
Amer. Chem. Soc., Washington, D.C., 1975, pp. 85-107.

R.L. Ashenhurst, in Computer Networking and Chemistry, P. Lykos,
Ed., Amer. Chem. Soc., Washington, D.C., 1975, pp. 108-117.

J.C. Emery, in Computer Networking and Chemistry, P. Lykos, Ed.,
Amer. Chem. Soc., Washington, D.C., 1975, pp. 219-230.

B.C. Searls and D.E. Freberg, Computer, 8, 22 (1975).
D.H. Chung, Electronic Desigp, 133 132 (1976).

D.R. McGlynn, Microprocessors: Technology, Architecture, and
Applications, John Wiley and Sons, Inc., 1976, pp. 182-187.

D.H. Chung, Electronics, 102 (January, 1977).

P.M. Russo, Computer, 193 67 (1977).

A.J. Weissberger, Computer Desigp, 151 (June, 1977).

D. Phillips and A. Goodman, Electronics, 109 (July, 1977).
C.J. Sipple, Digital Desigp, 68 (August, 1977).

D.J. Mueller, Computer Desigp, 81 (October, 1977).

G. Adams and T. Rolander, Computer Desigp, 81 (March, 1978).

S. Hourdakis, Ph.D. Thesis, Michigan State University, 1978.

Intel Corporation, Intel 8080 Microcomputer Systems User's
Manual, Intel Corporation, Santa Clara, CA., 1976.

214

 

19.
20.
21.
22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

215

M.J. Milano and H.L. Pardue, Anal. Chem., 41, 25 (1975).
R.B. Coolen, Ph.D. Thesis, Michigan State University, 1974.
M.J. Milano and H.L. Pardue, Clin. Chem., 21, 211 (1975).
G.M. Rider and D.W. Margerum, Anal. Chem., 42, 2098 (1977).

T.A. Nieman, F.J. Holler, and C.G. Enke, Anal. Chem., 48, 899
(1976).

G.C. Pimentel, Appl. Opt., Z, 2155 (1968).

R.E. Santini, M.J. Milano, and J.L. Pardue, Anal. Chem., 45,
915A (1973).

Y. Talmi, Anal. Chem., 47, 658A (1975).

Y. Talmi, Anal. Chem., 423 697 (1975).

J.L. Dye and L.H. Feldman, Rev. Sci. Instrum., El, 154 (1966).

N. Papadakis, R.B. Coolen, and J.L. Dye, Anal. Chem., 42, 1644
(1975).

R.B. Coolen, N. Papadakis, J. Avery, C.G. Enke, and J.L. Dye,
Anal. Chem., 423 1649 (1975).

N. Papadakis, Ph.D. Thesis, Michigan State University, 1974.

R.M. Wightman, R.L. Scott, C.N. Reilley, and R.W. Murray,
Anal. Chem., 4Q, 1492 (1974).

J.W. Strojek, C.A. Gruver, and T. Kuwana, Anal. Chem., 41, 481
(1969).

M.R. Hollaway and H.A. White, Biochem. J., 149, 221 (1975).

C.J. Faini, R.J. Desa, and J. Lee, in Flavins and Flavoproteins,
T.P. Singer, Ed., Elsevier Scien. Pub. Co., Amsterday, Neth.,

1976, pp. 82-86.

MJ.Mflmm,HJHIQNMm PE.Cmm,R£.Smmhd,DM.Mmgmwm
and J.M.T. Raycheba, Anal. Chem., fig, 374 (1974).

T.A. Nieman and C.G. Enke, Anal. Chem., 483 619 (1976).
T.A. Nieman, Ph.D. Thesis, Michigan State University, 1975.
R.E. Dessy, in Computers in Chemical Research and Education,

D. Hadi, Ed., Elsevier Scien. Pub. Co., Amsterdam, Neth.,
1973, pp. 121—134.

 

 

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

53.

54.

55.

56.

57.

58.

59.

60.

61.

216

D.A. Yates and T. Kuwana, Anal. Chem., 4§3 510 (1976).

G. Horlick and E.G. Codding, Anal. Chem., 4a, 1490 (1973).

G. Horlick, Appl. Spect., g9, 113 (1976).

J.E. Hornshuh, Ph.D. Thesis, Michigan State University, 1978.

D.G. Mitchell, K.W. Jackson, and K.M. Aldous, Anal. Chem., 4;,
1215A (1973).

R.E. Johnson, Applications of RCA Silicon Diode Array Target
Vidicons, Camera Tube Application Note AN-4623, RCA Electronics

Components, Harrison, N.J., 1971.

B. Chance, J. Franklin Inst., 232, 455 (1940).
B. Chance, J. Franklin Inst., 2323 613 (1940).
B. Chance, J. Franklin Inst., 3223 737 (1940).

S.R. Crouch, F.J. Holler, P.K. Notz, and P.M. Beckwith,
Appl. Spect. Rev., 123 165 (1977).

D. Sanderson, J.C. Bittikofer, and H.L. Pardue, Anal. Chem.,
44, 1340 (1970).

G.E. Mieling, R.W. Taylor, L.G. Hargis, J. English, and
H.L. Pardue, Anal. Chem., 4§3 1686 (1976).

K.R. O'Keefe and H.V. Malmstadt, Anal. Chem., 707 (1975).

B.W. Renoe, K.R. O'Keefe, and H.V. Malmstadt, Anal. Chem.,
48, 661 (1976).

P.K. Notz, Ph.D. Thesis, Michigan State University, 1977.

F.J. Holler, S.R. Crouch, and C.G. Enke, Anal. Chem., 48,
1429 (1976).

E.G. Douglas, IEEE Trans. Electron. Devices, ED—22, 224 (1975).

Data Sheet on Delfection Assembly Type CYlOO, Penn—Tran Corp.,
Bellefonte, Penn.

L.A. Selke, IEEE Trans. Electron Devices, ED—l6, 625 (1969).
R.L. Felkel, Jr., and H.L. Pardue, Anal. Chem., 42, 1112 (1977).
T.B. McCord and M.J. Frankston, Appl. gptics, 443 1437 (1975).

J.E. Stewart, Durrum Application Notes No. 4, Durrum Instr.
Corp., Palo Alto, CA.

 

 

 

62.

63.

64.

65.

66.

67.

68.

69.

70.

71.

72.

73.

217

J.F. Below, Jr., R.E. Connick, and C.P. Coppel, J. Amer. Chem.
Soc., §Q, 2961 (1958).

J.L. Dye and V.A. Nicely, J. Chem. Ed., 4§, 443 (1971).

N.W. Tietz, Fundamentals of Clinical Chemistry, W.B. Saunders Co.,
Philadelphia, PA., 1976, pp. 298-304.

G.R. Kingsley, J. Biol. Chem., 131, 197 (1939).

A.G. Gornall, C.J. Bardawill, and M.M. David, J Biol. Chem.,
177, 751 (1949).

J.W. Mehl, E. Pacovska, and R.J. Winzler, J Biol. Chem., 177,
13 (1949).

T.R. Koch, C.P. Johnson, and M.E. Chilcote, Clin. Chem., 39,
392 (1974).

J. Savory, M.G. Heintges, M. Sonowane, and R.E. Cross, Clin. Chem.,
24, 1102 (1976).

Wai-Tak Law, Ph.D. Thesis, Michigan State University, 1978.
M.B. Denton and 5.8. Tilden, paper no. 225 presented at the

Pittsburgh Conference on Analytical Chemistry and Applied
Spectros Copy, 1978.

C.H. Moore, Astron. Astrophys. Suppl., 45, 497 (1974).

Product Specifications on Model 1070 Intelligent Diskette
Controller, Persci, Marina Del Ray, CA.

w

 

 

 

 

 

 

 

 
 

 

 

 

 

 

 

 

   

”'TITIITTITTJMTMMTTMTMTTWF