Digital PDFs

DEC-8E-ALUMA-B-D

December 1973

278 pages

Original

5.2MB

Document:	LAB-8E SoftwSysUsersMan
Order Number:	DEC-8E-ALUMA-B-D
Revision:
Pages:	278
Original Filename:	https://svn.so-much-stuff.com/svn/trunk/pdp8/src/dec/dec-8e-aluma/dec-8e-aluma-b-d.pdf

OCR Text

• II II*

users manual

digital equipment corporation

DEC-8E-ALUMA-B-D

L A B - 8 / E

SOFTWARE
USER'S

SYSTEM

MANUAL

from Software
For additional copies, order No. DEC-8E-ALUMA-B-D
Distribution Center, Digital Equipment Corporation, Maynard,
Massachusetts 01754

First Printing, August, 1972
First Revision, February, 1973

Your attention is invited to the last two pages
of this document.
The "How to Obtain Software
Information" page explains how to keep up-to-date
with DEC's software. The "Reader's Comments" page,
when filled in and mailed, is beneficial to both
you and DEC; all comments received are acknowledged
and considered when documenting subsequent manuals.

(c)

1972, 1973 by Digital Equipment Corporation

The material in this document is for information purposes and is subject to change without
notice.

The following are trademarks of Digital Equipment
Corporation, Maynard, Massachusetts:
CDP
Computer Lab
Comtex
DEC
DECtape
Dibol
FOCAL

Digital
DNC
Flip Chip
IDAC
Indac
KA10
RASP

LAB-8/e
OMNIBUS
OS/
PDP
PHA
PS/8
Quickpoint

RADRSTS
RSX
RTM
SABR
Typeset
Unibus

PREFACE

This manual supersedes the following manuals:

Advanced Averager User's Manual
(DEC-LB-U61A-D)

Basic Averager User's Manual
(DEC-LB-U60B-D)
PST, TIH and Latency Histogram Programs User's Manual
(DEC-LB-U40B-D)

Auto- and Cross-Correlation User's Manual
(DEC-LB-U41B-D)

DAQUAN User's Manual
(DEC-LB-U80B-D)
BASIC/RT User's Manual
(DEC-LB-U70B-D)

NMR Simulator User's Manual
(DEC-LB-U53A-D)

NMR Averager User's Manual
(DEC-LB-U52A-D)

and includes the LAB8/E Data Conversion program.

Advanced versions of the Advanced Averager; Basic Averager; TIH, PST
and Latency; Auto- and Cross-Correlation; and DAQUAN programs have
been written to run under OS/8.
MS

(Mass Storage)

These programs are referred to as the

versions.

XXI

CONTENTS
Page

BOOK ONE

ADVANCED AVERAGER

1-1

BOOK TWO

BASIC AVERAGER

2-1

BOOK THREE

PST, TIH AND LATENCY HISTOGRAM PROGRAMS

3-1

BOOK FOUR

AUTO- AND CROSS-CORRELATION

4-1

BOOK FIVE

DAQUAN

5-1

BOOK SIX

CONVERT PROGRAM

6-1

BOOK SEVEN

BASIC/RT

7-1

BOOK EIGHT

NMR SIMULATOR

8-1

BOOK NINE

NMR AVERAGER

9-1

APPENDIX A
APPENDIX B

LOADING PROCEDURES
A-l
RUNNING THE LAB8/E PAPERTAPE ADVANCED AVERAGER
FROM DISK OR DECTAPE (4K DISK MONITOR SYSTEM)
B-l
SCHMITT TRIGGER
C-l
EQUATIONS FOR STATISTICAL CALCULATIONS
D-l
DATA POINT STRUCTURE
E-l
HOW TO ACCESS DATA
F-l
MEMORY ORGANIZATION
G-l
ATTACHMENT OF AN ANALOG X-Y PLOTTER
H-l
DATA BLOCK STRUCTURE
1-1
THE GENERATION AND MODIFICATION OF CONTROL
TAPES FOR THE LAB8/E BASIC AVERAGER
J-l
EXAMPLE OF PARAMETER SETUP
K-l
IMPLEMENTATION OF USER COMMANDS FOR PST
L-l
HISTOGRAM BINARY OUTPUT FORMAT
M-l
SUMMARY OF COMMANDS FOR PST, TIH AND LATENCY
HISTOGRAM PROGRAMS
N-l
IMPLEMENTATION OF USER COMMANDS FOR AUTOAND CROSS-CORRELATION
0-1
CORRELOGRAM BINARY OUTPUT FORMAT
P-l
CHART RECORDER INTERFACING INFORMATION TO LAB8/E
VC8-E
Q-l
DAQUAN COMMAND SUMMARY
R-l
CALCULATE ROUTINE
S-l
CORE MAPS
T-l
BASIC/RT COMMAND AND FUNCTION SUMMARY
U-l
SUMMARY OF THE USE OF NMR AVERAGER
V-l
LAB8/E MASS STORAGE PROGRAMS BINARY DECTAPE
W-l

APPENDIX C
APPENDIX D
APPENDIX E
APPENDIX F
APPENDIX G
APPENDIX H
APPENDIX I
APPENDIX J

APPENDIX K
APPENDIX L
APPENDIX M
APPENDIX N
APPENDIX
APPENDIX P
APPENDIX Q
APPENDIX R
APPENDIX S
APPENDIX T
APPENDIX U
APPENDIX V
APPENDIX W

BOOK

ONE

ADVANCED AVERAGER

CONTENTS
BOOK ONE
Page

1.0

ADVANCED AVERAGER FOR PAPER TAPE
GENERAL DESCRIPTION

1-1

1.1

Program Tapes

1-2

1.2

Basic System Requirements

1-2

RUNNING THE ADVANCED AVERAGER
SYSTEM INITIALIZATION AND PROGRAM LOADING

1-4

CHAPTER 1

CHAPTER 2
2.0

2.1
2.2

2.3
2.4

CHAPTER

The LAB8/E Compiler (Section I)
On-Line Signal Averaging* Confidence
Limits and Trend Data Collection
Completion of Confidence Limits and
Trend Computations (Section III)

Output (Analog X-Y Recorder or Scope
Photo; Digital Values in Actual Input
Signal Units) (Section IV)

1-5

1-20
1-2 3

1-24

ADVANCED AVERAGER MS

3.0

1-37

3.1

GENERAL DESCRIPTION
Using Advanced Averager MS

3.2

New Commands

1-39

3.3

Errors

1-40

3.4

Restart Program

1-41

3.5

Data Format of Output File

1-41

3.6

Types of Data Blocks on Mass Storage

1-41

3.7

Data Storage Example for Three Averages

1-42

1-37

TABLES
1-1
1-2
1-3
1-4

On-Line Control Characters
Correction Factors for Confidence Limits
Threshold Values of the Trend Function
Section IV Commands

1-i

1-22
1-25
1-27
1-35

CHAPTER 1
ADVANCED AVERAGER FOR PAPER TAPE

1.0

GENERAL DESCRIPTION

The LAB8/E Advanced Averaging Program digitizes, displays, and averages analog signals at rates from 175 ysec to

seconds/point.

Pre-

stimulus averaging, dual resolution sweeps, on-line computation of 95

percent confidence limits and trend, signal editing and sorting, preset sweep count, as well as a conversational compiler for selecting
the parameters of an average, are provided.

Before the user proceeds to the actual running of the Advanced Averager, he should try to gain an overall understanding of the program.

The Advanced Averager was written to be as flexible a signal averager
as possible.

Practically all of the parameters that define an average
are specified by the user.
For this reason, an entire program has

been devoted to the setup phase of the averaging process. This is
the first of four programs (or sections) that make up the Advanced
Averager.
Section I asks and records the answers to the parameter
questions

When all of the questions pertaining to the average have
been answered, either the answers are left in the memory of the com.

puter for use by the remaining three sections , or they are punched
on paper tape (a control tape) which may be used in place of Section
at a later time.

The latter feature provides a fast means of entering

the averaging parameters to the remaining three sections of the Advanced Averager.

Once the average has been defined by either Section I or a control
tape, the program that performs the data acquisition is read in.
This
is Section II.

Since the primary purpose of Section II is to acquire
and sum the data, the command structure is limited to data acquisition
and display commands, such as:
start summing, stop summing, start
again, display the input or the sum, and expand or contract the sum.

When the data has been collected, Section III is called in to process
it.
At this time the true average is calculated and the final computations for the statistics are performed.
Because this section is

devoted entirely to computation, there is no display and no keyboard
interaction.
l-i

upon the completion of the
calculations, the output program (Section
IV) is read.
This program is designed to make the data
available to
the user in as many useful ways as possible.
Display of each average
and its associated statistics is provided for
photography along with
selective type-out and integration routines.
Plotter calibration and
plotter output routines are also provided.

When output has been completed, another average may
be taken by reading in a control tape or Section I, followed
by Section II.
It should be understood that while one section
is present in core,

other sections are not present and may not be
referenced.
1. 1

Program Tapes

There are seven binary tapes
_which_ make up _the LAB8/E Advanced
Averaging System.
These are:
DEC-8E-AAP1A-A-PB, DEC-8E- AAP2A- A-PB DEC-8E
AAPTA-A-PB1-PB4, and DEC-8E-AAAPA-A-PB. DEC- 8E-AAP1AA-PB contains all
four sections of the Advanced Averager,
it should be used if the system
does not have disk or DECtape. Although the
sections are contained on
one tape, they are separated by "leader/trailer"
code.
This signals the
loader program that the end of a section has been
reached and reading
should stop until the next section is required.
DEC-8E-AAP2A-A-PB contains Sections II-IV and should always be used if
a control tape is
used instead of Section I and the system does not
have disk or DECtape.
DEC- 8E-AAPTA- A-PB contains Sections i-iv and should
not be used when the
program is to be run from either disk or DECtape. Control
tapes
should be used in conjunction with DEC-8E-AAAPA-A-PB
when disk or DECtape is used.
If the system is run from disk or DECtape, refer
to
instructions in Appendix B for loading and storing the
programs.
The
Averager can be run under OS/8 and data can be written
to mass storage.
Refer to Chapter 3 for program tapes and instructions.
,

1.2

Basic System Requirements
PDP 8/E Computer with at least 4K of core.
Additional core will
be utilized automatically by the program.
D

L m8 EA

t
L^ tc
+ am8

10 bit analog to di gital (A/D) converter with
pre-amplifiers and eight channel multiplexer.

DK8-ES

Real time programmable clock.

Vc8-E

10 bit display controller.

1-2

H945AA

Mounting panel and power supplies for A/D,

VR14, m
Tektronix
ttt,i a

.l.

-.clock, and display controller.
}

602,603,604,RM503L
or other accept- ("Display oscilloscope.
able scope
J
ASR33
Teletype
1

High-speed reader Strongly recommended.
The OS/8 version requires the standard OS/8 configuration.

teletype is a registered trademark of the Teletype Corporation.

1-3

CHAPTER 2
RUNNING THE ADVANCED AVERAGER

SYSTEM INITIALIZATION AND PROGRAM LOADING

2.0

For the LAB8/E programs to run properly,
the PDP-8/E must have its
knob and switches set in certain positions.
Before running any programs, the user should check to see that the
knob and switches are
set in their correct positions.
The user starts the system by turning
the key on the left-hand side
of the PDP-8/E console to POWER.
The key must not be turned as far
as PANEL LOCK.
Next, the user must ensure that the HALT and
SING STEP
keys are in the up position.
Then, the status knob in the right-hand
center of the console is turned to AC.
This step enables the contents
of the accumulator (AC) to be seen in the
lower of the two rows of
lights.
Next, the Teletype switch is turned to LINE,
and, finally,
the scope is turned on.
The LAB8/E system is now ready to receive'

programs
All of the LAB8/E programs are loaded with the
Binary Loader program
(BIN)
If the BIN Loader is not in core (field
0) , it must be loaded
with the RIM Loader.
See Appendix A for details of loading BIN
and
RIM
.

To load the Advanced Averager program from
paper tape:
a.

Place the tape (DEC-8E-AAP1A-A-PB, Sections i-iv)
in
the tape reader so that the leader/trailer
(a single
channel 8 punch) is on top of the reading mechanism.

Turn the Teletype control knob to LINE.

Turn the scope to ON.

Set the Switch Register
an the up position.)

to 7777 Q

(SR)

(All switches

Press the ADDR LOAD key.

e d read r is USed
turn on reader and set SR
It
! fl
« xn ?
to ?7?7~^
3777 (bit
down position).
'

W " SPeed readSr

g
"

START!°

" USSd

SSt thS reader switch to

Press the CLEAR and CONT switches on the
computer console.
1-4

The tape should start moving through the reader.
If it does not, BIN is not present and must be
loaded per the instructions in Appendix A.

When the tape stops, check that the AC lights are
all off (to see the accumulator, turn the status
knob to AC).
If they are not all off, a read error
has occurred and steps a. through j. should be repeated.

When loading is complete (AQ*0) , press CONT.

The following message appears on the scope:

LAB8/E IS READY
HIT RETURN TO PROCEED

RETURN is the button on the far right of the Teletype keyboard. Press
this key to go to the next display.
RETURN is used throughout the
LAB8/E program to go on to the next step or to "execute" an answer
which has just been typed.
2.1

The LAB8/E Compiler (Section

Section

defines the average to be taken.
The answers the program
receives are used throughout the remainder of the Advanced Averager.
1.

Typing the RETURN key after the introductory display causes
the following question to appear on the oscilloscope:

DIGITAL I/O?
(Y OR N)
(Y FOR YES, N FOR NO)
This question asks if the digital I/O option (12 bits in,
12 bits out) is part of the system configuration.
If N
for NO is typed, no further questions are asked regarding
digital I/O. If the option is part of the system, but the
contingency feature is not to be used in a particular experimental run, N can be answered to this question. RETURN
must be typed after either Y or N to execute the answer.
If Y is answered, then question 10 is asked later.

NOTE

If YES is answered to the above question,
the digital I/O must be the unmodified
DR8-EA for the program to function properly.

In all conversational mode YES or NO questions, SPACE, RETURN, or any character other
than Y is interpreted as NO.

After the RETURN key is typed, the following will appear regardless of the answer to question 1.
SYNC ON INPUT:

1-5

The blank can be filled in with
0, 1, 2, or 4.
This is
number of the phone jack connector on the front of the the
laboratory panel to which the synchronization cable is
attached.
Attach the lead now. SI, S2 and S4 are inputs for
either
fast pulses or slowly varying analog signals (a coupling
can
be used to lead a signal to both a sync input and one
of the
raaxn analog sxgnal inputs, allowing the LAB8/E
to synchronize
on the signal itself).
The knobs above SI, S2, and S4 control
the levels at which the input pulse (or signal) causes
the
Schmitt triggers to fire. For details of Schmitt trigger
firing, see Appendix C.
If S0 is chosen for the sync pulse, no sync lead is
required.
The program will automatically trigger itself every
Least
Sweep Interval (for definition, see question 17).
Simultaneously with start of the sweep, a +3 to
volt pulse, the
duration of a sample (see sampling rate, question 9) will
be
presented to bit
of the digital I/O.
This enables the
LAB8/E to control the stimulation in this particular
trial.

NOTE
The user should always verify that the scope
displays the intended answer before typing
the RETURN key.
The user can correct errors
by using the RUBOUT key.
3.

The following display appears next:

STANDARD RESOLUTION:
-

DATA POINTS

Standard resolution refers to the main signal epoch which
this
and the following two displays define.
It takes the following
three parameters to define a signal epoch for the LAB8/E.
a.

Number of data points into which the waveform
is to be digitized,

Length (in milliseconds or seconds) , and,

Delay (amount of time either before or after
the sync pulse that the signal epoch begins).

Any number from 65 to 1024 (using the basic system requirements)
can be typed.
However, standard resolution is only one of the
two basic building blocks ot the LAB8/E system.
It may be associated with specific analog input and sort code (with digital
I/O) before signals defined by this signal epoch can be
averaged.
Therefore, it should be remembered that (in the basic configuration) if 1024 data points are used to define the standard
resolution epoch, there will be no room for high resolution,
standard deviation, or trend. A reasonable choice might be
200
data points for the standard resolution epoch.
4,

After entering the number of data points, followed by a carriaqe
RETURN, the following will be displayed:
LENGTH

SEC

1-6

This asks the user to define the length of the signal epoch
he wishes to look at or analyze.
Note that there is a space
between the fifth and sixth blanks. This indicates that the
first five blanks are for one to five numbers (or one to tour
plus decimal point) which must be followed by a carriage
RETURN, and the sixth is for either an M to indicate milliseconds or SPACE or RETURN to indicate seconds.

Numbers may be entered in any position (leading or trailing
blanks and leading J?'s are not required). The sampling rate
(epoch length/number of data points) can be between 175
ysec/point and 4 sec/point. For faster sampling rates the
Basic Averager must be used (refer to Book Two of this
manual)
5.

The only parameter yet to be specified for the first basic
signal epoch is the time at which the epoch begins (with
reference to the sync pulse)
DELAY:

_SEC

Note that this display is similar to the preceding; however,
a minus sign (hyphen on keyboard) is allowed to indicate a
signal epoch beginning before the sync pulse.
If a negative
delay is used, the sync pulse must arrive no later than the
time equivalent of the last data point.
In other words, a
negative delay cannot be longer than the length of the signal
epoch.
Otherwise, the following error message will appear:
?

SWEEP END EARLY

indicating that the sweep would end before sync pulse
arrived.
Type the RETURN key to respecify the delay.
LINE FEED to change both length and delay.

Type

A negative delay is of use if the user suspects his system
of anticipating the stimulus and reacting before the stimulus actually arrives, or if the user is interested in detecting any CNV that may be present. Specifying a negative
delay will enable him to view the state of his system before
the stimulus arrives.
Thus, he sees any pre-stimulus response
which otherwise would have remained obscure and the base line
before the onset of response.
A positive delay would be used if the elicited response did
not occur immediately following the stimulus, but at a later
time (i.e., the time until response was long in comparison
to the length of the sweep) .
Using a positive delay would
yield a sweep whose focus was on the area of interest and
not over the entire time range.
NOTE
In any display requiring a numerical answer,
typing the RETURN key only is interpreted as
an answer of 0.
If only one resolution is given, it is referred to as the
high resolution; however, to change it refer to the standard resolution input.

1-7

Defining the Second (or High Resolution)
Basic Signal Epoch
6.

The next display is:

HIGH RESOLUTION EPOCH
DATA POINTS

Defining two epochs (standard and high resolution)
permits
simultaneous averaging of signals in two different time
intervals at two different sampling rates on the same or
different signals. Averaging is done at the standard
resolution only if this question is answered with
or RETURN.
If the high resolution epoch is used, at least
65 points
must be requested. The maximum number of points
that can
be specified by the combined epochs is 1024; if
points
400
were used in the standard resolution, up to 624
points could
be specified here.

NOTE
There are 3072
a core locations available
±v
for storing
data.
It takes three locations to take an average of one point
(one for temporary storage and two for the
average).
Therefore, if only an average
is taken, the maximum number of points that
may be averaged is 1024 (3072/3).
If more
core is available, the maximum possible number of points is increased.
7.

Typing RETURN will now bring up the following display:
HI LENGTH:

_SEC

If the high resolution epoch length typed results in a
per-point sampling rate which is slower than that resulting
from the definition of the standard resolution epoch,
the
following error message appears:
?

BAD SAMPLING RATIO

This means that either the number of data points in
the high
resolution epoch or the length must be redefined so as to
give a new sampling rate faster than or equal to the standard
resolution sampling rate.
Type RETURN to change the high
resolution parameters.
Type LINE FEED to change both hiqh
and standard (or low)
8.

After typing RETURN, the following is displayed:
HI DELAY:

_SEC

Again a delay (positive or negative) is permitted. This
allows the user to position a high resolution window
about an
area of particular interest.
RETURN must be typed after the
numerical part of the answer and again after the unit portion
(M for milliseconds or SPACE for seconds).

1-8

If the epoch lengths and delays selected for standard and
high resolution are such that there is no overlap between
them, the following error message appears:
?

NO COMMON INTERVAL

Type RETURN to respecify both high and standard (low) resolution epoch lengths and delays.
They must be respecified
so that the epochs overlap by at least one point.
Type
LINE FEED to change only the high resolution parameters.
As noted, the number of high and low resolution data points
If 64 or less points are specified for the
must exceed 64.
standard resolution, the question is reasked.
If 64 or less
points are specified for the high resolution, a straightline display results.
9.

After successfully defining the three parameters for both
the standard and high resolution signal epochs, the following
sweep summary appears (these figures are examples only)
BEGINS

-19.50M
-: CHANGE (H,L

AT RATE

ENDS

500U
1.000M

179. 5M

99.50M

)

In the above example, the standard resolution epoch samples
the first point at 19.50 ms before the sync pulse, the sampling
rate is 1.000 ms per point, and the last data point is 179.5
ms after the sync pulse.

The high resolution epoch begins at the sync pulse, the
sampling rate is 500 ps/point (twice the rate of the standard resolution epoch) and the epoch ends 99.50 ms after
the sync pulse.

When the high resolution is displayed, it is shown to have
the same length as the standard resolution, so it is effectively blown up in the horizontal direction.
NOTE
a.

All measurements (such as DELAY, SORT AT,
etc.) in the use of the LAB8/E are referenced to the leading edge of the sync
pulse.

Both standard resolution and high resolution epoch lengths and delays should
be integer multiples of the fastest sweep
rate selected.
If the user selects fractional multiples, the LAB8/E calculates and
displays the closest approximation. This
is the sweep summary display.

There is an uncertainty of 1 standard
resolution sampling interval in the sweep
start time.
This is expressed in the
sweep summary by not rounding the low
resolution begin and end times.

1-9

If only a standard (low)
resolution is

specified along with a positive delay
time, then the high resolution becomes
the standard (low) resolution and the
BEGIN and END values of the low resolution
sweep summary equal the positive delay
time.
(A zero or negative delay appears
as zero in the low resolution sweep summary.)
These low resolution values are
ignored by the remaining sections of the
program.
Only the high resolution figures
future calculations.

There are only two data rates in the system,
the high and low
80
011 rateS
Then for a11 channels used, there can be
ff!
the
high resolution or the low resolution, or both
rates on
each channel.
The system requires that the low resolution
a m ltx P le of the high resolution rate.
Namely, if
?
It „ t
1S R micros econds between points, then the
5;
w
•?i
will be M*R microseconds between points where M=l,2,3. L.R.R
Hence,
if the H.R.R. is 200 microseconds between
points then the L r'r
must be 200 or 400 or 600, etc. microseconds
between points
If the user puts in a high and low resolution
length and num&
f P °. ntS SUCh that the L-R-R- is not
a multiple of the
S
u
l
h.r.r. , then the low resolution rate will be
adjusted to the
nearest larger rate so that the rates are multiples.
For
example:
-

LOW RES.

300 points
200 M seconds

RATE=666y

HIGH RES.

200 points
100 M seconds

RATE=500u

Then the rate for the low resolution will be
set to
lisecond.
10.

mil-

If YES was typed in answer to DIGITAL I/O?
(Question 1) the
following will appear, otherwise Question 11 will
appear:

SORT AT

SEC

This refers to the contingency feature and specifies
the sort
time with reference to the sync pulse.
The sort time is the
instant at which the computer will look at the states
of
rightmost eight bits of the digital contingency inputs; the
the
states of any of these inputs can have changed at
any time
during the interval between the sync pulse and
the sort time.
At sort time the computer stores the pattern of
0's and l's
lng thS states of ^e eight digital contingency
;"£
inputs.
When !u
the averages are updated with the current sweep,
this
pattern is compared with the sort code specified
for each
discrete average waveform.
(The sort code will be specified
later at the same time that individual analog
inputs are
matched up with the standard and high resolution
epochs.)
If the pattern contains l's in the same
positions as in the
corresponding sort code, the current sweep will be
added to
the average corresponding to this sort code;
if not
the
sweep will not be used in the average.

1-10

NOTE

11.

For the program to read a "1" from any
of the digital inputs at sort time, the
digital input must have seen a ground at
any time from the sync time to the sort
time.
If no ground is seen during this
time period, 0's will be read.

Since the computer will not look for the
next sync pulse until after sort time,
this feature can be used to artificially
lengthen the sweep for purposes of disregarding sync pulses occurring too soon
after the end of the previous signal
epoch.

After typing RETURN, the following scope display will appear:

AVG #1
ANALOG INPUT
In this display, ANALOG INPUT
asks which analog input will
be used for Average Number 1." If the first input is used,
can be typed, followed by a carriage RETURN.

Analog inputs can be associated with discrete averages in
any order.

Example:

AVG #1
ANALOG INPUT

AVG #2
ANALOG INPUT

AVG #3
ANALOG INPUT

Since octal notation is used to specify the inputs, there are
no inputs 8 and 9.
In the basic configuration only inputs 0-7
can be specified.
If more inputs are implemented, inputs 0-17
may be specified.
NOTE
For further explanation of octal and binary
numbers, refer to Chapter 1 of Introduction
to Programming , Volume 1.
12.

The next display (if both high and standard resolutions have
been defined) is:

RESOLUTION:

(H,L)

which asks which of the two basic signal epochs (windows) is
to be used to define average #1.
H indicates high resolution;
L indicates standard or low resolution. All averages using
high resolution should be specified before low resolution averages so that the positions of the averages on the screen are
in the same order as for their corresponding raw input (Monitor)
displays.
1-11

13.

If YES was answered to the original DIGITAL I/O?
question,
the following is displayed after typing RETURN, otherwise
question 14 is displayed.

SORT CODE:

NOTE

Analog input numbers and sort codes are
the only parameters which must be specified in octal notation. All others are
in the familiar decimal notation.
SORT AT
and SORT CODE:
are the two displays which
specify the parameters of the contingency condition. With
the digital I/O option the contents of the temporary storage
buffer are not automatically emptied into the series of memory
locations allocated for the accumulation of a certain average.
This is done only if certain conditions, or contingencies,
are fulfilled.
The sort code is what specifies the condition
for the addition to memory of the most recently digitized
data.
For example, if a sort code of 001 were specified for
average #1, data acquired during each sweep would only be
added to those memory locations allocated to average #1, if,
at sort time, the status of the inputs in the contingency
register were as follows:
C9

CIO

Cll

(binary)
(octal)

C4-C11 refer to the rightmost
digital input register.

A logical AND is performed between the sort
code and the contents of the contingency
register.
If the result of the AND procedure
is equal to the sort code, then the contingency
condition is satisfied.

In the above example, therefore, the contingency
condition is also met when the contingency register contains a 00 000 111, or a 10 110 111, or,
in fact, whenever Cll contains a 1 regardless
of the contents of the remaining bits in the
register.

bits of the

The contingency feature can be used to generate a wide series
of different average waveforms from the same analog input
based on changes in experimental conditions. The changes in
experimental conditions, in turn, cause changes in the contingency register.
This feature could be used to sort responses to different stimuli so that several different
stimuli could be presented in random order.
This feature
could also be used to edit abnormal responses.
If an artifact were present in the response, the contingency register
could be set so that the particular response would be discarded or saved in a separate average (the signal would be discarded if the contingency register were set so that it did
not correspond to any sort code)
If the contingency feature
.

1-12

is not to be used for a certain average,
should be typed for the sort code.
14.

or RETURN

After the RETURN key is typed, the following is displayed:
CONFIDENCE LIMITS?:
(Y:

YES)

asking if confidence limits are desired for each point on
this average.
NOTE

15.

A single confidence limit data point requires three memory locations, whereas
an average data point requires only two.
In order to conserve memory locations
confidence limits should be requested for
averages consisting of as few locations as
Relatively few data points could
possible.
be allocated to the standard resolution
epoch, for example, and confidence limits
Twice or
could be requested for this.
three times the data points could then be
allocated to the high resolution epoch,
If the lengths
without confidence limits.
and delays were identical, this arrangement
would produce one confidence limits point
for every two or three average points, a
ratio which is usually acceptable.

Confidence limits and trend points are
displayed in Section IV (Output Section)

See Appendix D for confidence limit equation.

If Y (YES) is answered in response to the above display and
the RETURN key is typed, the following appears:

COMPUTE TREND
(Y:

YES)

The significance of trend will be explained in the description
of Section IV of the program.

NOTE
a.

Each trend data point requires four memory
locations in addition to the three for confidence limits and two for its associated
average.
Trend cannot be calculated without confidence limits.

See Appendix D for trend equation.

1-13

To optimize the power of LAB8, a few minutes
might be spent calculating the maximum number of points available to the user in various situations.
The following equation allows simple determination of this.
Keep in
mind that the number of memory locations
used must not exceed 3072 + n X 4096, where
n = the number of additional memory stacks.

Number of memory locations used =
x
2x
+
3x
+
+
4x
+
y
+2y + 3y + 4y+..
TEMP BUFFER AVG. TERM CONF. TERM TREND TERM TEMP BUFFER AVG.
CONF. TREND

where x = the number of standard resolution points
and y = the number of high resolution points.
If some of the statistics are not used, their respective
terms
should be dropped from the equation.
16.

When the above question is answered and carriage RETURN is
typed, the following display appears:
AVG #2
ANALOG INPUT
This can be answered in the same manner as in steps 11-15.
If two resolutions are being used and average #1 is reserved
for the high resolution, average #2 could be associated with
the same input signal and question 12 should be answered with
an L for low resolution.
Additional averages can be set up
as required until the following display appears (only
after
the last question for an individual average)
?

INSUFFICIENT MEMORY

If an average with confidence limits resulted in this error
message, or if trend was specified as well, it might be possible to set up just the average, and possibly confidence
limits without exceeding the memory available.

Type the RETURN key to respecify this average.
Type the
LINE FEED key to respecify all averages.
In certain cases,
where a sampling rate close to the maximum possible was
specified, the following error message may appear:
?
TOO MANY INPUTS
FOR SWEEP RATE

It takes a finite amount of time for the computer to switch
back and forth between a series of inputs, and this time may
make it no longer possible to maintain the sampling rate.
To
remedy this, either a slower sampling rate should be specified
or fewer inputs should be requested.
Type the RETURN key to
delete the last average.
Type the LINE FEED key to delete
all averages.
Type CTRL/Z to start from the beginning.

1-14

NOTE
a.

Under most circumstances when in the Conversational Mode Compiler section (Section I)
it is possible to go back to the previous
display by typing the LINE FEED key, in which
case the previously- typed answer will be discarded and must be respecified.

In case it is desired to respecify the parameters of the averages , type the CTRL/Z keys
(CTRL key depressed at the same time as the
letter Z) to return to the display:

LAB8/E IS READY
HIT RETURN TO PROCEED
in which case all parameters can be redefined.
17.

When sufficient averages have been set up, typing the LINE FEED
key discontinues this section of the compiler and brings up
the following display:
LEAST SWEEP INTERVAL
_S

(>238.0 MS)

(238.

MS is an example only)

This indicates that to do the calculations which the LAB8/E
makes between sweeps , the sync pulses can be no closer together than 238.0 ms (in this example). External sync pulses
can thus be at regular 1/4, 1/2, 1 second intervals, etc.,
or they can be at irregular intervals the shortest of which
is 239.0 ms (in this example).
If the S0 sync ouptut is
used, the LAB8/E supplies pulses at exactly the interval
specified in answer to the above question. Sync pulses occurring at a shorter interval from the last acknowledged
sync will be ignored (and not acknowledged)

NOTE
a.

Any time between the least sweep interval
and the user-specified minimum interval
typed above is used by the LAB8/E to stabilize the display of inputs.
The shortest
interval desired or expected under experimental conditions and not the shortest interval allowed by the LAB8/E should normally
be selected.

Because of differences between individual
machines, trouble may develop if the user
specifies a least sweep interval that is
very close to the shortest interval allowed.
This is especially so in a case of multiple
inputs.
The symptoms of the trouble will be
a halt of the averaging process before the
preset number of sweeps is reached. This is
remedied by increasing the least sweep interval.

1-15

18.

Typing the RETURN key brings up the following Preset Sweep
Counter display:
AVERAGE

SWEEPS

The largest number that can be typed as an answer to this
display is 4095. However, answering
or RETURN is equivalent to requesting 4096 sweeps.
In the Signal Averaging
section (Section II), typing the CTRL/R keys allows repetition of the present number of sweep counts.
19.

The next display shows the positions the actual inputs
occupy on the oscilloscope screen during the View Input
mode of the on-line operation. As an example:

INPUTS
1:0, H

5:0, L

2:5, H

6:3, L

3:2, H

7:4, L

4:1, H
.

The numbers 1-7 indicate that, in this example, there will
be seven different inputs and seven different oscilloscope
traces in the View Input mode.
The middle numbers indicate
the input channel numbers associated with each trace, and
the letters H or L repres ent high or low resolution signal
epochs.
Note that all hi gh resolution traces precede low
resolution traces in this INPUTS display.
20.

Typing the RETURN key brings up the following display:

AVERAGES
1
2
3
4

If no more than one contingency condition for the same
signal epoch (either high or standard resolution) on the
same analog input has been selected, the number of inputs
will be the same as the number of averages. The numbers
in the above display correspond to AVG 1, AVG 2, AVG3, etc.

Typing the LINE FEED key causes a return to the previous
question between the Sync question and the rate display
(shown in step 9).
To change answers after a RETURN is
typed to question 9, CTRL/Z should be typed to
return to question 1. CTRL/Z does not work while displays 19 and 20 are shown, and once the sweep summary
has started to print.
If any key is hit once the sweep
summary starts to type, the "TRIGGER" section is skipped,
and Section II of the tape is read.
1-16

21.

Typing the RETURN key brings up the following display:
PUNCH CONTROL TAPE?
YES)

(Y:

Punching a control tape allows the user to:
a.
Establish a precisely repeatable setup for
averaging.
Avoid reading in the compiler (Section I)
b.
each time a particular set of parameters
chosen is desired for an average.
c.
Vary certain parameters during the averaging
section (Section II) by reading in different
control tapes and restarting Section II at 6530
(set 6530 in the switches, press ADDR LOAD and
then CLEAR, CONT)
If a control tape is not desired, type N or the
RETURN key and a summary of the averaging parameters will be typed. At the end of the summary
1
a line similar to the following is printed
:

(VAP,0, 230-6277

VAP indicates the version being used - in this
case, version A.

indicates the number of data fields (blocks
means that the
of 4K of core) minus 1.
machine is a 4K machine. 2 30-6277 spots the
area of core that is available for the storage
of data.

Trigger Adjustment

After this message, the following is typed 2

TRIGGER
The operator should now consult the sweep summary typeout
Included
which corresponds to the parameters now in memory.
followed by 0, 1,
will be the statement SYNC ON INPUT S
If S0 is being used, this section of the program
2, or 4.
actually has no function, since the sweep will begin as the
result of an internal, computer-generated pulse which needs
Depending on which of the numbers 1, 2, or
no adjustment.
4 was selected, the operator should make sure that a cable
carrying the sync pulse is connected to the proper phone
connector on the front panel of the laboratory peripheral:
The operator should then make
SI, S2, or S4, respectively.
sure that the input pulse, or analog signal, is in the
range of ±5V at the time it is desired that the Schmitt
trigger fire, and that this voltage is actually being delivered by the cable. An oscilloscope could be used for
,

this purpose.
'This message is meaningless when output
program.

by the Advanced Averager MS

If the number of channels and size of the channel buffers are such that
the end of buffer link is stored in locations 4272 8 -4365 g , the trigger
If this situasection is destroyed and the results are unpredictable.
tion occurs, create a control tape of the Q & A section and run the Averager with that tape.
2

1-17

An analog signal in the range of ±lv should now
be connected
to Analog Input 0.
If the Schmitt trigger is firing, this
signal is displayed on the scope every time a sync pulse is
received.
For those acquainted with the wired-program averaging devices, the characteristic sweep of the signal from
left to right beginning at the sync pulse will be familiar.

NOTE
This is true, however, only in this section of the program.
In the on-line averaging section, the movement of the illuminated
part of the display from left to right
across the waveform has nothing to do with
the arrival of a sync pulse or the actual
beginning of the sweep; it is merely the
beginning of another display cycle.

The clock knob with the same number as the Schmitt trigger
being used should now be adjusted to ensure reliable, regular
firing of the Schmitt trigger every time a sync pulse is received.
If displays appear on the scope periodically regardless of the knob (threshold) setting, the sync pulse apparently goes beyond ±5V each time.
If it is important to synchronize the sweep with a particular point on the waveform
arriving at the sync connector, the amplitude of this point
should first be measured and then the Schmitt trigger threshold adjusted to this value.
Once the synchronization has
been satisfactorily adjusted, type any key to go on to the
on-line averaging section (Section II). (Refer to Section 2.2.)

Ignore the dot on the X axis of the scope while the TRIGGER
and channel
are being viewed in the TRIGGER section of
Section I.
It is not part of the input signal.

NOTE
To reset the timing before proceeding to
Section II, stop the computer and reload
Section I via the Binary Loader.

An
and a
(on the next line) are printed to reguest the loading of Section II. Turn the reader
ON and Section II will begin loading.
Proceed to
the description of Section II. 1
If Y for YES is typed followed by the RETURN key, the following display appears:
d.

22.

[

TITLE

Note that there are twenty-one blanks which can be filled in.
In addition to letters, numbers and normal punctuation, the
following characters can be used:
*Not done in MS version.

1-18

_
<
>

Exclamation Mark
Percent Sign
Space
Angle Brackets
(both directions)
Multiplication Sign

Number Sign
Division Sign
Parentheses
Up Arrow
Back Arrow
Equal Sign

/
(

)

*
=

(*)

After 21 (or fewer) characters are entered, type the
RETURN key to indicate the end of the title identification.
At this point the oscilloscope screen will go blank and
t will be printed.
To punch out a control or parameter tape

1
;

Push the button marked ON on the box at the upper left of
the ASR33 Teletype, making sure that the leftmost switch
(bit 0) of the Switch Register is up (a binary 1), or,

With the high-speed paper tape punch, make sure that the
leftmost switch (bit 0) of the Switch Register is down
(a binary 0)
Now type CTRL/P (hold down the CTRL key while typing P)
This causes the punch to begin punching out the control
(parameter) tape.
The first part of the tape to be punched
is the 21 (or fewer) character title in hand-readable form
(holes are punched in the tape in the same configuration
as dots for the corresponding characters displayed on the
scope)
The strip of tape immediately following the handreadable title will contain the ASCII (Teletype) characters
making up the sweep summary. At the same time that the
control tape is being punched, the teleprinter types the
sweep summary corresponding to the parameters contained
on this tape.
This sweep summary can be torn off and
attached to the control tape for future reference. As
many control tapes as required can be generated with one
loading of the compiler (Section I)
The compiler returns
to display 1 to permit the user to specify a new set of
parameters if Y was answered to PUNCH CONTROL TAPE?
.
(To determine what parameters are contained
in a particular control tape, switch the Teletype to LOCAL
and read in this first section of the tape. The punched
tape characters immediately following the title should
produce the sweep summary.)
The last section (separated
by a length of tape with continuous holes on the right side)
contains the actual parameters required by the computer
in concentrated form.
Only this last section should be
read followed immediately by the shorter of the two LAB8/E
program tapes marked Use with Control Tapes Only
.

NOTE
The user may return to display 1 at any time
during Section I (except between questions 19
and 20) by executing the CTRL/Z command.

Reading a Control Tape for Paper Tape
The condensed information of a control tape (described above) is used
to drive the averaging process according to the responses the user made
in Section I of the averaging program.
'Not applicable to MS version.

1-19

b.
c.

d.
e.
f.

Place the control tape in the reader with
the leader/
trailer codes over the reading mechanis
Lsm.
Turn the Teletype control knob to LINE.
Turn the scope to ON.
Set the SR to 7777 Q
O
Press the STOP and ADDR LOAD keys.
.

If using a low-speed reader, set the
reader switch to START.
If using a high-speed reader, set the
SR to 3777 and
turn on the high-speed reader.
8

e S
LEAR and C0NT switch es.
lu
thatt the

If error has occurred (Ac/0) , return
to step a.
(If
this fails, verify that the Binary Loader
is in core.
(Refer to Appendix A.)

Place tape DEC-8E-AAP2A-A-PB in the reader.
Repeat steps b. to i.

S accumulator displays 0. When tape stops, verify

When loading is complete, check to see that the
twelve lights of the
accumulator are off (this corresponds to the condition
AC=0)
If
this is not the case, a read error has occurred
and the tape should
be reloaded until such time that the AC does
equal zero. When AC=0,
press the CONTinue key on the computer console.
Section II starts.
.

2.2

On-Line Signal Averaging, Confidence Limits and Trend Data
Collection

After the tape has been read, turn the reader to STOP 1
Verify that the
accumulator is 0; if not, reread Section II 1
Connect the leads to the
analog inputs if this has not already been done.
Press CONTinue on the
console to proceed with Section II 1
The LAB8/E system types < and
starts in the View Input Mode.
This means that whatever is displayed on
the scope initially is the actual analog input to
the system and not
an accumulating average.
This is done so that the operator may adjust
the gains on his pre-amplifiers to produce outputs
as close to ±1V as
possible without clipping (flattening) the highest
amplitude peaks.
This can be done before the actual averaging is begun.
.

The traces which are displayed when the on-line
section of the program
starts have the same heights and widths and are in
the same positions
as in the INPUTS display (the INPUTS display
was one of the last displays in the compiler section of the program)
if many inputs are
being used, or if their positions on the scope are
somewhat unusual,
it may be advisable to take a Polaroid scope photograph
of the INPUTS
display during the compiler operation to use as a reference
during online operation.
.

Not applicable in MS version.
1-20

NOTE
If a relatively fast sampling rate has been specified, a large number of inputs selected, confidence

limits and trend requested, or relatively little
time allowed between sync pulses, the LAB8/E may
have so many operations to perform that it does
not have time to display more than every fourth or
fifth data point.
Display is considered a low
priority operation in the LAB8/E because A/D
conversion, multiplexing, storing the results
of A/D conversions, confidence limit and trend
computation (if requested) are considered more
important.
For this reason, the quality of the
display may be somewhat degraded under certain
circumstances during on-line operation.
If the V command is now typed, the LAB8/E will switch the view mode
to show the contents of the memory locations in which the average
way.eiorms will be accumulated.
(Confidence limits and trend are not

displayed during the on-line operation.)
The average is automatically
zeroed so that all of the traces in this display should now appear
as
straight, horizontal lines.
Typing V again will switch the view mode
back to the digitized inputs.
Type RETURN to begin the averaging process. Waveforms should
now appear to grow where the straight lines had been in the previous
display.
If the straight lines do not change, the slops and
threshold
adjustments on the Schmitt trigger should be altered until the
straight lines do change, see Appendix C for adjusting the
Schmitt
trigger.
Typing V will change the view mode to View Input.

Typing V again will change it back to view averages.
If a very fast
sampling rate was specified, only a flash of light at
the beginning
of the display cycle may be visible.
When the display appears to go
off scale, type the command C for contract, which will
divide the
vertical display scale by a factor of two.
This may be necessary since the data is saved as a 24-bit sum.
X can also be
typed to multiply the vertical display scale by a factor
of two
.(X stands for expand)
.

The character

attained.

is printed when the preset number of sweeps has been
This display is the non-normalized sum of the data taken
>

from each separate sweep.
To erase this data and begin a new average,
type CTRL/Z followed by the RETURN key.
The computer will stop averaging and display the new waveform after the same preset number
of
sweeps

1-21

If it is desired to take another group of sweeps, type CTRL/R
and the RETURN key and the preset number of sweeps will be taken

and added to the first group.

Table 1-1 is a summary of the control characters used in the on-line
section of the LAB8/E Advanced Averager Program.
Table 1-1

On-Line Control Characters

Character

Explanation

CTRL/Z

Zeros (reinitialized) data points used for
accumulation of average waveforms, sets
view mode to View Input, prints <.

RETURN

Starts or resumes averaging process, sets
view mode to View Average, types carriage
RETURN and LINE FEED; returns display to
zero scale.

LINE FEED

Interrupts averaging process (without initializing average to zero) , sets view mode
to View Input, types two LINE FEEDS.

Changes view mode (either from View Input
to View Average, or vice versa)

Expands display of average by a factor of

Contracts display of average by a factor
of 2.

CTRL/R

Resets sweep counter to begin preset number of sweeps again. Previous average is
not erased so that new sweeps are added to
old after RETURN is typed. CTRL/R or CTRL/Z
is required to resume averaging after > has
been printed.

CTRL/P

Pulls in next section of the LAB8/E Advanced
Averager Program tape.

The following characters are printed only:

Indicates average is done (preset number of
sweeps has been attained)
Indicates a requirement to read more tape.
Turn Teletype reader to START if the lowspeed reader is used.
Does not apply to MS version.
Requests next section to be read.

1-22

NOTE

Although confidence limits and trend can
be computed during Section II, the subroutines for displaying confidence limits and
trend data points (made up of three and
four memory locations, respectively) are
not included in the on-line averaging section.
These are contained in the Output
Section (Section IV)
2.3

Completion of Confidence Limits and Trend Computations
(Section III)

In Section II the LAB8/E stores the accumulating sum in the core

locations reserved for the average, the sums of the squares in the
core locations reserved for confidence intervals, and the sums of
the squares of successive differences in the core locations reserved
for trend.
Section III completes the calculations necessary to con-

vert the sums of the squares into confidence intervals (95%)
and
the sum of the squares of successive differences into the trend
,

function (see Appendix

The user may proceed from Section II by
typing CTRL/P (holding down the CTRL key and typing the P key)
D) .

When Section III has been read in, check the accumulator lights on
the console to make sure that they are all out all 0's) 2
If this
is not the case, i.e., if a reading error is indicated, reposition
.

the tape to the beginning of Section III and press the CONTinue key
on the console twice. 2 If Section III has been read properly,

press CONTinue once to start Section III 2 . To compute standard
deviation instead of confidence limits, modify the contents
of locations PATCH

and PATCH

(which will start the section)
gram at location 7603.

Location
PATCH 1 (7304)
PATCH 2 (7360)

as follows before pressing CONTinue

After modification, start the pro-

Change to

4506
1265

7410
1264

Section III will take between 20 and 50 seconds to complete the confidence interval and trend calculations (with 4K memory) depending on
the number of data points involved.

The only indication that the
computer is calculating will be that the computer console lights will

When the calculation is complete, an
and a
(on the
2
next line) will be typed on the Teletype
At this point, the reader

be flashing.

[

*For the Advanced Averager MS refer to section 3.2 for patch instructions.
2

Not applicable to MS version.
1-23

switch on the Teletype should be pushed to START if
the low-speed
reader is used and Section IV will be read 2
In the case of the
high-speed paper tape reader, Section IV will be read automatically 2
.

2.4

Output (Analog X-Y Recorder or Scope Photo; Digital Values in
Actual Input Signal Units)
(Section IV) 1

When Section IV has been read in, verify that the accumulator lights
are all out (all 0's), i.e., that there has been no error during read. 2
(If an error is indicated, reposition the tape to the
beginning of
Section IV and press CONTinue twice) 2 . When the accumulator indicates
all zeros, press CONTinue to start up Section IV 2
.

As soon as Section IV is started, disconnect the leads to
the analog
inputs.
One to four traces should appear on the oscilloscope.
Only
one trace will be displayed if N was answered to the following
question (during Section I):

CONFIDENCE LIMITS?:
If Y is answered to this question, but N is answered to
the following
question:

COMPUTE TREND?
three waveforms are displayed. At first, it may appear that only
one trace is visible even if confidence limits have been requested.
If this is the case, turn up intensity on the oscilloscope
to increase
the persistence of the image and type X (for expand), if necessary,
to separate the three curves to a greater degree.
The Teletype prints
a record of the times X has been typed so that C can be
typed the

same number of times later to restore calibration of the digital
typeout.
Typing RETURN will also restore the display to zero scale.
If confidence limits were requested during the preparation
of the
particular control tape being used, the scope intensity was
turned
up, and X typed as needed, three curves should be clearly
visible
in the upper part of the screen.
These curves may appear to be one
curve which jumps up and down between three different
positions.

These positions represent the average waveform (in the middle)
the
upper (95%) confidence limit, and the lower confidence limit.
The
1
Refer to Appendix F, How to Access Data, for more detailed
information.
2
Not applicable to MS version.
1-24
,

confidence limits indicate the range within which 95 percent of
These limits are accurate to +1 percent.
(Strictly

the data falls.

speaking, the probability is 95% that all the data fall within this
range
.

A figure (J) will appear on the left side of the screen. The horizontal bar represents ten sampling intervals; the vertical bar represents 100 millivolts as seen by the A/D converter.
Two bright spots will appear on each curve.
The positions of these
spots (cursors) can be controlled by moving analog input knobs
and 1.

The case of seemingly nonexistent cursors can be solved by

a combination of two methods

Turn the intensity down so that the waveform almost
disappears.
The cursors should be visible if they
are on the trace.

Type the command I.
Data will be printed on the
Teletype in the following format:
#

Sweeps

Scale Factor

Lo Lim

Hi Lim

XXXX

The significance of this will be explained later.
Of interest now is Lo Lim and Hi Lim.
These are
the point positions of the cursors on the waveform.
If 0000 is typed for a limit, the cursor is off the
left-hand end of the trace and the knob must be
turned counterclockwise.
If both cursors are lost
off the right-hand end of the trace, only the number of sweeps and the scale factor are typed.

Table 1-2 shows the different correction factors which the user can
apply to the confidence limits calculated by the LAB8/E Advanced
Averager Program (which are accurate to ±1%, assuming 40 to 128
sweeps) to obtain the absolute 95 percent confidence limits.
Table 1-2

Correction Factors for Confidence Limits
Number of Sweeps

1.15
1.06
1.02
1.01
1.00
1.00
0.99
0.99
0.98
0.98

16
32
40

60
64

120
128
256
512

1-25

True 95 percent Confidence Limits = m±pc

where

mean

correction factor

confidence limits calculated by LAB8/E
(see Appendix D)

As indicated above, the LAB8/E goes several steps beyond the variance calculations.
In all but cases involving a very small number

of sweeps, the confidence limits computed by the LAB8/E can be
used in lieu of the true confidence without applying the correction

factor P.
If the trend function is requested, one additional waveform appears
on the bottom of the screen.
This is the trend function which indi-

cates

(a) if the successive sweeps which are added together to make
up the average waveform are steadily drifting in amplitude,
(b) if
the successive sweeps are alternately increasing and then decreasing

(oscillating), or

(c)

if there is no measurable consistent change

between the sweeps
Table 1-3 indicates the threshold values of the trend function for
several values of N (number of sweeps).
The lower trend threshold,
t , is the value above which one can be 95 percent confident that
d
there was no drift, or below which one can be 95 percent confident

that there was consistent change (in one direction) in the amplitude
from one sweep to the next, or from one group of sweeps to the next.
The upper trend threshold, t , is the value above which one can be
95 percent confident that there was a measurable, repeatable increase

followed by a decrease in amplitude from one sweep or group of sweeps
to the next.
The grid calibration routine in Section IV draws lines
indicating trend = 2.000, trend = 2.500, and trend = 1.500, values

which approximate the significant ranges shown in Table 1-3.

The trend function provides an important means for evaluating the
validity of the data represented by the average waveform. An average waveform accompanied by fairly wide confidence limits can
indicate
any of three radically different conditions which can only be distinguished by means of the trend function. These conditions are:
1.

A signal of interest oscillating up and down in
amplitude between fairly wide limits.

1-26

Table 1-3

Threshold Values of the Trend Function

N
8

16
32
40

60
64

120
128
256
512

98
23
1 44
1 49
1. 58
1. 58
1. 62
1. 62
1. 62
1. 62
1

A signal of interest which has increased or decreased significantly in amplitude during the
averaging run.
A signal of interest which has not changed at
all during the averaging run but which happens
to be very small in amplitude compared to the
noise in which it is buried.
Trend between t and 4.00 indicates Condition 1, trend between 0.01
and t, indicates Condition 2, and trend between t, and t indicates
u
Condition 3. Condition 1 suggests that the averaging run be broken

up into a series of shorter runs to try to establish the period of
oscillation. Condition 2 also suggests that the averaging run be

broken up into a series of shorter runs to establish the period of
time in which the greatest change occurred.
Condition 3 suggests
that the averaging run be made even longer to narrow the change between the confidence limits and further improve the signal to noise

An average waveform accompanied by a fairly narrow confidence
limit and a trend function between t and t indicates that sufficient
sweeps have already been taken.
ratio.

All of the curves displayed on the oscilloscope at one time in Section IV belong to the same average (AVG #1, AVG #2, or AVG #3, etc.).
This is in contrast to the display in Section II in which all the
averages (but not confidence limits or trend) were displayed on the
oscilloscope at one time. The displays in Section IV are like pages
in a book.
Pages might be thought of as labeled AVG #1, AVG #2, etc.

Each contains its own discrete average and all the waveforms (either
2 or 3) associated with it.
To display, plot, or print any average,

1-27

turn to the correct page and proceed
with the desired routine. Pages
may be turned forward by typing the RETURN key
(a RETURN on the last
page will bring up the first page; CTRL/Z will
return the display
to the first page)

Plotter Calibration Routine
Section IV includes calibration and adjustment
routines for the X-Y
recorder.
Appendix H describes how to connect the plotter to
the
LAB8/E.
To calibrate the X-Y recorder, type K (calibrate).
The following message is typed:
BIAS

and the maximum voltage appears at both the X and
Y outputs of
the oscilloscope connections.
The bias controls on the X-Y recorder
should now be adjusted until the pen is at the
appropriate corner of
the plotter field for maximum voltage (usually
the upper right-hand
corner)
Type the RETURN key.
The following message is
typed:
.

GAIN

Minimum voltage appears at both the X and Y outputs
of the oscilloscope.
The gain controls on the X-Y recorder should now
be adjusted
so that the pen is either at the lower left-hand
corner of the printed
section of X-Y recorder paper (or at some multiple of
ten chart divisions from the right-hand edge in the X direction and
some multiple
of eight chart divisions from the upper border in
the Y direction)
Because adjusting the gain controls of most X-Y
recorders affects the
bias setting
(and vice versa)
it will usually be necessary to type
RETURN several more times (repeating the "BIAS" and
"GAIN" adjustment
cycle) until the cycle can be repeated with the
pen alternately at
the upper right-hand corner and then at the lower
left position without readjustment of the controls.
To EXIT from the calibrate routine
type CTRL/Q.
,

Drawing a Grid on the X-Y Chart Paper
In order to draw in the borders on the X-Y
chart paper indicating
full scale positive, 0, and full scale for
negative for the average

1-28

and confidence limits as well as lines indicating trend = 2.500,

trend = 2.000 and trend = 1.500 (see diagram below), type the command
G on the Teletype.
AVERAGE = 1.024V

AVERAGE = 0.000V
TREND = 2.500
TREND = 2.000

TREND = 1.500

AVERAGE = 1.024V
The following message is typed:

LIFT
Turn the X-Y recorder on and type the RETURN key on the Teletype.
The following message is printed:

LOWER

Follow this instruction by lowering the pen against the X-Y chart
paper, turning analog knob 3 fully counterclockwise, and typing the
RETURN key. This causes the grid to be plotted. The speed of
the plot can be made slower by turning knob

3 clockwise.
grid is completed, the following message is printed:

When the

LIFT

Lift the pen away from the chart paper before any other character
is typed on the Teletype, turn off the X-Y recorder, and type

CTRL/Q to resume the display of the average.
NOTE
If the X-Y recorder is left on while waveforms
are being displayed on the scope, the X-Y recorder pen will shake furiously as it attempts
to move at the very rapid rate of the oscilloscope display.
If the pen is against the chart
paper at this time, the chart will be ruined.

1-29

Plotting the Average, Confidence Limits and Trend
The plotting of curves can be started by typing P (for
plot) on
the Teletype.
This stops the oscilloscope display and causes
the following message to be printed:

LIFT

Turn on the plotter and lift the pen away from the chart paper.
Type the RETURN key, and the following message is printed:

LOWER

Disregard this message the first time. Keep the pen raised away
from the paper and type the RETURN key.
The pen begins
tracing the average waveform except that the pen is not touching the paper.

Quickly adjust analog knob 3 for the fastest
slew rate possible without running the risk of rounding off the
steeply rising or falling parts of the waveform. When this waveform is completely traced, the following message is printed:

LIFT
(If only the average was computed,

OFF is printed.)
If
the slew rate is adjusted for the X-Y recorder in use, type
CTRL/Q followed by a P command. Answer LIFT by typing the RETURN
key.
The following is printed:

LOWER
If analog knob

3 is set for the proper slew rate, lower the
pen against the paper and type the RETURN key.
The average is
traced on the paper.

After this waveform is completed, the following is printed:
LIFT
If any further adjustment of the slew rate is desired,
knob 3 can
be adjusted at this time (clockwise to make the pen move
slower;

counterclockwise to make it move faster)
following is printed:
1-30

Type RETURN and the

LOWER

Lower the pen against the paper and type the RETURN key
to begin plotting the upper confidence limits on the average.
Repeat the process for the lower confidence limits and trend waveform (if specified)
When the last waveform specified in this average is written,
the following is printed:
OFF

Lift the pen and turn off the X-Y recorder.
to start the display.

Type the RETURN key

Interpreting the Result of the X-Y Recorder Trend Readout
The following is an example of a typical trend readout at the bottom
of the X-Y recorder chart paper (it is assumed that there were more

than 40 sweeps in the average)
TREND > 2.5
TREND = 2.500

\Ai/w1/KAa A/

TREND = 2.000

1.5<TREND<2.5

TREND = 1.500
TREND < 1.5

In this example (with 95% confidence) the signal remained constant
during the first one-third of each sweep. The trend = 3 in the
second one-third of the sweep indicates (with 95% confidence) an
oscillating condition in this part of the waveform. The signal was

alternately higher and then lower in amplitude from one sweep or
group of sweeps to the next sweep or group of sweeps in this section of the waveform.
this example, indicates

The last section of the trend function, in
(with 95% certainty)

a steady drift in ampli-

tude from one sweep or group of sweeps to the next sweep or group of
sweeps.
Trends of 4 and
instead of approximately 3 and 1 as in
this example, would have indicated a greater certainty of oscillation, and a greater certainty of drift, respectively.

1-31

Digital Typeout and Punched Paper Tape Output on ASR33 Teletype
To begin the typeout of the data corresponding to the average,
confidence limits, and trend for the average being displayed, type the
command T on the Teletype. For punched paper tape output, merely
push the button marked ON on the control box on the left-hand side
of the Teletype before typing T.
The average confidence limits and

trend for the data points between and including the two cursors
are
The same data printed is also punched on paper tape.

printed.

The format of the data printed on the teleprinter
is similar to
the following:

JW61,
0010,
0011,
0012,

0000
-0344,
-0338,
-0327,
" t Ce •

•

0022,
0029,
0023,

2075,
109 3,
2311,

• •

The first number in the upper left-hand corner is n, the number of
sweeps included in this average.
The second number on the same line
is k,
the scale factor (expressed as a power of 2).
In order for
the following numbers in the left-hand column to be in actual calib-

rated input signal units, the scale factor must be 0.
The scale factor may be decreased or increased by typing the commands C or X
respectively.
The numbers in the four columns beginning at the second line are
the data point number, the mean, the confidence interval, and
1000
times the trend, from left to right respectively.
The confidence limits are the mean ± the scaled confidence interval.
In the example:

-0344 + 0022 = -0322 = Upper 95 Percent Confidence Limit
-0344 - 0022 = -0366 = Lower 95 Percent Confidemce Limit

Confidence limits for each of the other average points can be calculated in similar fashion.
For its interpretation, trend merely requires a decimal point between
the first and the next three digits in each four-digit group in
the

1-32

third column.

In the example, trend for the first data point is

2.075, for the second data point is 1.093, and for the third is
2.311, etc.
If trend was not computed for this particular average, there will

be only three columns in the printout.

If confidence limits were

not computed, there will be only two columns

(the data points making

up the average)

NOTE
1.

Only the number of sweeps and the scale
factor will be typed out if both cursors
are off the right-hand end of the trace.

If the input is constant and there is no
variance, the computation of trend will
involve division by zero.
This returns a
zero answer.
If this happens, the trend
value is set to 2 which indicates no
trend has taken place.
,

Commas are provided to make any output
tape compatible with BASIC/RT (refer to
Book Seven)

When the last requested point in this particular average has been
typed, the display of this average is resumed.
To discontinue digital
readout before all the data points are read, type the CTRL/Q command,

which also resumes the display of this average.

Finding the Area Under the Curve
Type

to print the sum of the data points between the two cursors

The format of the printout is:

sweeps, scale factor, lo lim, hi lim, sum avg's, sum conf. lims.,
The high and low limits identify the bounds on the data points

included in the integration.

The sum of the averages and sum of

confidence limits are suffixed with t if overflow occurs.
To
get the true summation (supposing overflow occurred) scale the average
down by typing the command C and take the new scale factor into
account when noting the result is within scale.
To get the true
integral, multiply the summation by the clock rate.
The sum of
the confidence limits is not reported when none has been computed.

1-33

To fxnd the latency of a particular
part of the waveform displayed
from the beginning of the waveform,
position the left-hand cursor on
the point in question and type the I
command.
The third number
printed is the number of the point covered by
the cursor.
To find
the latency of that point, multiply the
point number by the sampling
rate.

It should be noted that there is a
certain amount of error inherent
in the sum of the averages printed by
the I command.
The sum of the
averages is done as single precision (12 bits)
add on scaled numbers.
If the sum exceeds plus or minus 204
8 during the summing the
up arrow
is printed, indicating an overflow.
The sum is obtained as follows
The double precision data point is
first scaled by the scale factor
then added to the single precision sum.
This type of adding creates
the sum's error. As a result the
difference

between two identical
sums of N numbers taken at a scale
factor differing by 1 can be as
much as N.
For example if the first sum of N numbers
is at a scale
of -3 and the resultant scaled values
are all even positive values,
then the resultant N numbers at a scale of -4
will be exactly 1/2
the number at the scale of -3.
The sum at a scale of -4 will be
exactly 1/2 the sum at a scale of -3, and
no error will be seen between
the two sums.
If the N numbers at a scale of -3 are
all positive odd
numbers then double the sum of the N numbers
at a scale of -4 will
differ from the sum at a scale of -3 by N.
Hence, if the result of
the sum is small in comparison to the
number of values summed, the
comparison of two sums taken with different
scale factors can be
relatively large. For example if the values
are both positive and
negative, the sum can be small in comparison
to the number of values
summed.
If 200 points were summed, the sum
on

identical data could
be 100 for the scale of -3 and
for a scale of -4.
Hence it is
important to keep the sums as large as
possible (small scale factor)
but this is not always possible. When
positive and negative values
are added, the resultant sum can be small
no matter what is done.
For the above reasons, comparisons between
averages should be made at the same scale

values of different
value.
Two averages can
I

be almost identical and one will overflow
at a scale of -3 for example
and the other one will not overflow.
Hence it is important to assure
that all values being compared be calculated
at a scale that will
not overflow following averages.

1-34

In Section IV of the averager the cursors may shift between two

points.

This is due to a one bit jitter in reading the pots.

will not affect the T or

commands since the limits are output.

If a large jitter is seen, check if all inputs are disconnected,

and when only the sync is in, the sync input may be miswired, causing a ground loop.

Proceeding to the Next Display of an Average
Type the RETURN key to proceed from one display of an average and

associated statistics to the next. Type CTRL/Z to return to the first
average.
If all averages computed have been displayed, the display
returns to the first average.
To begin computing new averages, remount either a control tape followed by DEC-8E-AAP2A-A-PB or
DEC-8E-AAP1A-A-PB and type CTRL/P.
Table 1-4 summarizes the Teletype keyboard commands during Section
IV operation.

1-35

Table 1-4

Section IV Commands

Command

Explanation
Contract display; divide data to be printed, punched
or written by a factor of 2.

Expand display; multiply data to be printed, punched
or written by a factor of 2.

—

Calibrate ("Gain.

Plot grid on analog plotter (average and confidence
limits =
+ 1.024V, -1.024V; trend = 2.000, 2.500,
and 1.500)

"Bias

etc.) analog plotter.

Plot data on analog plotter.
NOTE
For commands G and P above, analog knob
should be turned fully counterclockwise
then clockwise a bit before executing
these commands.

Type (and punch, if Teletype punch is on) on teleprinter the digital data between cursors.
I

Sum of the data between the cursors.

RETURN

Proceed to the next step or next average; return display to zero scale.

CTRL/Q

Quit current operation, return to display of current
average

CTRL/Z

Return to the display of the first average.

CTRL/P

Read in next section of tape.

1-36

CHAPTER

ADVANCED AVERAGER MS

3.0.

GENERAL DESCRIPTION

The Advanced Averager MS

(Mass Storage)

for OS/8 will write data on

mass storage and may operate from control tapes stored on mass storage.
Otherwise, the operation is the same as paper tape system.
The binaries of the Advanced Averager MS come in five sections 1

Section
Section
Section
Section
Section

Paoertane
Papertape
DEC-8E-AAA1A-A-PB
DEC-8E-AAA2A-A-PB
DEC-8E-AAA3A-A-PB
DEC-8E-AAA4A-A-PB
DEC-8E-AAA5A-A-PB

1
2
3
4
5

DECtape
_
(DEC-8E-ALMSA-A-UB)
~*
AAVG1 BN
AAVG2.BN
AAVG3.BN
AAVG4.BN
AAVG5.BN
.

All five sections are "CHAINED" together under OS/8 so that each
section calls the next from the system device on command or
automatically.
3.1

Using Advanced Averager MS

Loading the Programs
The programs are distributed on paper tape or DECtape in binary format.
If desired, the files can be PIPed (with /B) onto the system.
(Note
that the files are binary and a .BN extension must be assigned to the
files.)
The binaries must be saved on the OS/8 system device as follows:

Load each of the overlays in turn via the ABS Loader and save it with
the SAVE command.
The SAVE commands for the overlays are:
2

SAVE SYS

AAVG2 SV

6 400-75 77; 06 52 7=0

Section

SAVE SYS

AAVG3.SV

Section

SAVE SYS

AAVG4.SV

6400-7577;06777=0
0-177, 6400-7577;00017=0

SAVE SYS

AAVG5.SV

6 400- 71 7 7,-06 3

Section

Section
and 403.

7=0

of the Advanced Averager has two starting locations 400
Start at 400 to do parameter initialization of Section 1;

start a t 403 when a control tape is being used to run Section 2.
Sections 1, 2, 3 and 4 are the same as I, II, III and IV of the Advanced Averager.
1-37

Section
SAVEd.

can be run after being loaded with ABS
Loader or after being
If Section 1 is to be SAVEd, use the SAVE command:
1

To start at 400,
SAVE SYS

AAVG1.SV

0-7577, 12000-13177 -.00400=0

AAVG1.SV

0-7577 ,12000-13177 00 403=0

or to start at 403,
SAVE SYS

Then to start the program respond to the dot with

AAVG1

STARTING LOCATIONS:
Section

Sections
2/

to run Section 1
to use control tapes

have the same starting addresses as the paper
tape system.

Section

400
403

6400.

NOTE

When a SAVEd module is being chained under OS/8,
the starting address in the SAVE command is its
true starting address -1.

CONTROL FILES:
Control files replace control tapes of the paper tape system.
When a control file is desired, YES is answered to the question "PUNCH CONTROL TAPE".
The program prints an asterisk to
which the user gives the output file name in OS/8 format:
*dev:FILE.5<
The file is written to the device specified as file FILE.
and control is given to OS/8. When producing a control file

or running the Averager from Section 1, the starting address
is 400.

To run the Averager from a control file, Section 1 of the
Averager is started at 403. The program then asks for the
input control file by typing an asterisk.
The file is given
in OS/8 format:

dev:FILE.5
1-38

This would load FILE. 5 from the device specified.

The

TRIGGER section would then start.
3.2

New Commands

The following paragraphs explain the features of Advanced Averager MS

which differ from the Advanced Averager.
Section
1.
2.
3.

Typing CTRL/C returns control to OS/8.
Typing the RETURN key in the TRIGGER section loads Section
"PUNCH A CONTROL TAPE" refers to a control file, answer Y
for YES, N for NO.

If a control file is to be output to mass storage, an asterisk
is printed.
In response, the OS/8 file definition is given
as shown above
4.

If the program is started at 403, an asterisk is printed and
the control file is given as shown above.

Section

CTRL/C returns control to OS/8 and replaces the CTRL/U, user
command.

CTRL/P loads Section 3.
If a user command is desired, the CTRL/C can be replaced or
adjusted as desired.

Section

runs by itself after being loaded.
pleted, Section 4 is loaded.

When it is com-

The following procedure must be used to patch the program
to compute
standard deviation instead of confidence limits. When section
3
(AAVG3.SV) is to be saved, load AAVG3.BN via the Absolute
Loader.
Call ODT via the OD command to OS/8 and change the two
locations as
shown.

Location

Change to

PATCH 1 (7304)
PATCH 2 (7360)

4506
1265

7410
1264

SAVE this module as explained in section 3.1.

1-39

Section

CTRL/C returns the user to OS/8.

CTRL/W replaces CTRL/P. CTRL/W loads Section
the data collected to a mass storage device.
Section

which outputs

An asterisk is printed requesting the output file and device.
For example,

*dev:FILE.l<

outputs the averaged data on the device specified in file
FILE.l.
3.3

Errors

Most errors are I/O errors, assuming the five sections were saved
correctly.
For more information on errors, refer to the Software
Support Manual for OS/8.

Section

If an error is detected in Section 1 while linking the

chains of the program, the Program prints USER ERROR 1, and
returns to OS/8.
This will happen if Sections 2-5 were not
SAVEd correctly.
An error when reading or writing a control file returns
the program to OS/8.
This can be caused by giving an output
file when an input file is requested or an input file when
an output file is requested.
In addition, if the device
does not exist or the file is not found, an error occurs.

Note that if a file exists with the same file name given
the new file, the old file is deleted.
Sections

If the programs were not saved correctly, the program crashes,
This is true for Sections 1 and 5 also.

1-40

Section

The following errors can occur:

USER ERROR 1

the output device is not found.

USER ERROR 2

an output file cannot be created.

USER ERROR

a write error occurs while writing the
output data.

the file cannot be closed or the device
is full.

Other errors result in Monitor Error messages.

Chapter

3.4

(OS/8) of Introduction to Programming

(Refer to
.

Restart Program

To run the Averager again,

type CTRL/C then R

AAVG1 when the dot

is printed.

3.5

Data Format of Output File

The output file created by the CTRL/W command is in OS/8 format as

follows

BLOCK

WORD

1=1

WORD

=
= 1

WORD

256 = Job lists from Advanced Averager.
A
first word of the Job list is
the end of the Job lists.

If WORD 2 of previous block = 1, then the format is
the same as for Block 1.
If WORD
data.

1st DATA BLOCK

3.6

if no more Job list blocks
if the next block is a Job list block.

of previous block = 0, this block starts

WORD 1 = number of sweeps
WORD 2 and on = data as stored in core of
Advanced Averager.

Types of Data Blocks on Mass Storage

START OF ADVANCED AVERAGER BLOCK DATA

WORD 1 = number of sweeps
WORD 2 and on = data as stored in core

1-41

CONTINUATION OF DATA INTO NEXT MASS STORAGE BLOCK
WORD 1 and on = data continues from previous block
as stored in core.

LINKED ADVANCED AVERAGER DATA BLOCK
WORD

and on = data starts with the average value
of a data point as stored in core.
The first data of a linked data
block in core always starts a new
output block.

Data blocks or linked data blocks as written by the Advanced Averager
are written on mass storage starting with a new block, in a
contiguous
fashion.

The data is ended with the length of the next linked data

block or a 0.

ADVANCED AVG. CORE
Loc. 230
2 37

246

DATA START

(as it

Job List
Job List
Job List

(7
(7
(7

exists in Section IV)
words)
words)
words)

for AVERAGE 1
for AVERAGE 2
for AVERAGE 3

# of Sweeps of Average 1
DATA of average 1 in one field, 100 locations

end of data

DATA START

# of sweeps of average 2
Data of average 2 in one field, 300 locations

end of data

DATA START

# of sweeps of Average 3
Data of average 3 in multiple fields, 100 locations

-# of points in linked block

CDF
Start of data -1

DATA CONTINUATION

Data of average

continued, 300 locations

Data as written to output device by CTRL/W command:

Refer to Appendix

for more information on data block structure.

1-42

Block

Word

1
2

Job list of average
Job list of average
Job list of average

10
17

Block

1
2
3

Word

1
2

Block

# of sweeps of average
Data

of sweeps of average

Word

Data

256

Block

Word 1

continuation of data

Block

Word

Block

of sweeps of average 3

Data

102

-# of points in linked block = -300

continuation of data

Word

256

Block

Word

continuation of data

1-43

BOOK

BASIC

TWO

AVERAGER

CONTENTS
BOOK TWO
Page

CHAPTER 1

2-1

1.1

INTRODUCTION
Program Tapes

1.2

Basic System Requirements

2-1

1.0

CHAPTER 2

CHAPTER

BASIC AVERAGER
2-1

RUNNING THE BASIC AVERAGER

2.1

Connecting Leads

2-3

2.2

System Initialization and Program Loading

2-3

2.3

Loading the Basic Averager

2-4

COMPUTATIONS OF STATISTICS AND OUTPUT
COMPLETION OF CONFIDENCE LIMITS AND TREND
COMPUTATION

2-10

3.0

CHAPTER 4

BASIC AVERAGER MS

4.0

GENERAL DESCRIPTION

2-11

4.1

Using the Basic Averager MS

2-11

FIGURES
2-1

Basic Averager:

User Flow Chart

2-i

2-8

CHAPTER 1
BASIC AVERAGER
1.0

INTRODUCTION

The LAB8/E Basic Averager Program digitizes, displays, and averages
analog signals at rates from 33 ' usee per point per channel. The

program is entirely core resident and its output is compatible with
the statistics calculations and output section (3 and 4) of the LAB8/E
The user can select and
Advanced Averaging program (DEC-8E-AAP1A-A-PB)
adjust, on-line, the sampling rate, the number of sweeps, and the
delay via the Teletype. By using control tapes (LB-U01B-PB through
the user shapes LAB8/E to his particular needs and is
LB-U08B-PB)
Refer
able to perform sophisticated statistical manipulation on data.
.

to Chapter

1.1

for the OS/8 version of the Basic Averager.

Program Tapes

The Basic Averager program consists of one large paper tape and a set
The large tape, DEC-8E-ACA0A-A-PB, consists of
of eight small tapes.

three programs separated from each other by leader/trailer code.
This code signals the loader program that a program has been read

and that reading should stop until the next program is required.
The other two
The first program of the three is the Basic Averager.

programs are taken from Sections III and IV of the Advanced Averager
program.
These two programs are optional and are provided as a con-

venience to the user.
The eight short tapes are called "control tapes".
be used when the Basic Averager is run.

One of these must

See Appendix J for details

of control tapes.

1.2

Basic System Requirements

PDP8/E computer with 4K of core
AD8-ES single input A/D converter
or
AD8-E + AM8-EA + AM8-EC A/D converter with
multiplexer and preamps

channel

DK8-ES real time programmable clock
*33 microseconds is the maximum rate for multiple channels.
However,
a single channel can be run at a maximum rate of 25 microseconds.

2-1

VC8-E 10 bit display controller

H945AA mounting panel and power supply
VR14
or Tektronix

")
I

f display oscilloscope
602,603
604 or
J
RM503
ASR33 Teletype

High speed reader is used when available.
The OS/8 version requires the standard OS/8
configuration.

2-2

CHAPTER 2
RUNNING THE BASIC AVERAGER

Connecting Leads

The leads from the experiment to the LAB8/E should be connected to
the proper inputs to assure proper functioning.
The sync pulse should

always be connected to SI.

analog input

If only one signal is being averaged,

is the only one used.

If the eight channel multiplexer

is implemented and more than one channel is being averaged,

tional channels should be connected to analog inputs
secutive inputs starting with input
must be used.
2.2

the addi-

2, etc.

Con-

System Initialization and Program Loading

For the LAB8/E programs to run properly, the PDP-8/E must have its
knob and switches set in certain positions.
Before running any programs , check to see that the knob and switches are in their correct
positions.

Turn the system on by turning the key on the lefthand side of the
PDP8/E console to POWER. Do not turn as far as PANEL LOCK. Next,
ensure that the HALT and SING STEP keys are in the up position.

Turn the status knob in the right hand center of the console to AC.
This step enables the contents of the accumulator to be seen in the
lower of the two rows of lights. Turn the Teletype knob to LINE, and
finally turn the scope on. The LAB8/E system is now ready to receive

programs
All of the LAB8/E programs are loaded with the Binary Loader (BIN)
program.
If BIN is not in core it must be loaded with the RIM Loader.
See Appendix A for details of loading BIN and RIM.

Select a control tape from the following chart:
TAPE

CHANNELS

STATISTICS

DATA POINTS

LB-U01A-PB

Average

1000

LB-U02A-PB

Average

500

LB-U03A-PB

Average

250

2-3

TAPE

CHANNELS

STATISTICS

DATA POINTS

LB-U04A-PB

Average

Confidence Limits

500

LB-U05A-PB

Average

Confidence Limits

250

LB-U06A-PB

Average

Confidence Limits

125

LB-U07A-PB

Average,
& Trend

Confidence Limits

300

LB-U08A-PB

Average, Confidence Limits
& Trend

150

NOTE

Only the first control tape can be used
if the system is equipped with a single
A/D input.
To modify any of the control tapes, the user must
obtain the ASCII
version of the tape and change it, using the Symbolic
Tape Editor
(DEC-08-ESAB-PB)
When the desired changes are made, the tape must
be assembled using the PAL III Symbolic Assembler
(DEC-08-ASAC-PB)
.

before it can be used as a control tape with the Basic
Averager.
Refer to Appendix J for detailed instructions.
2.3

Loading the Basic Averager

To run the Basic Averager both the Averager program
itself and a
control tape must be read.
Using the BIN Loader, load the control
tape first:
a.

Place a control tape in the tape reader so that
the
leader/trailer code is over the read mechanism.

Turn the Teletype to LINE.
Turn the scope to ON.
Set the Switch Register (SR) to 7777. (all
switches
°
xn the up position)
Press the ADDR LOAD key.

c.
d.

e.
f*

cm*™
d
.TART

1"9

a low-s P eed reader, set the reader switch
to

•

If using a high-speed reader, set the
SR to 3777 and
turn on the high-speed reader.
8
h.
i'

Press the CLEAR and CONT switches on the
computer console.
Pe hould move trough the reader,
S» is notf present and must be loaded as if it does not,
BIN
described in
"
Appendix A.

When the tape stops, verify that the
accumulator lights
a
SSe thS accui™lator, turn the status
11°
knob
ar
rl
to AC).
If they are not all off, a read error
has occurred
and steps a. through j. should be repeated.

2-4

k.
1.

When the control tape has been correctly loaded, follow
the above steps to load the Basic Averager, DEC-8E-ACA0A-A-PB.
When loading is complete, press CONT.
NOTE

When the LAB8/E Basic Averager is loaded, it modifies
the Binary Loader to allow a program start by simply
pressing CONTinue. This feature requires the user to
start the Binary Loader at 7777 g each time a tape is
loaded.

See Figure 2-1 for a flow diagram of the operating procedure.

The

LAB8/E responds with
R:
1
The number typed is considered by
Type the sampling rate to be used
the program to be in units of microseconds. A number between 33 and
This sets the time
4096 should be typed, followed by a RETURN key.
.

If a slower sampling rate is required, the Advanced
between samples.
The program types
Averager should be used.
N:

In response to N: , type the number of sweeps desired in the average folIf an error has been made in typing, type the
lowed by the RETURN key.

CTRL/Q command and the question will be asked again.

N=0 is treated

as 4096.

The program will now print
D:

Answer with the delay in terms of the clock rate. That is, if the sampling rate is 50 ysec and it is desired to delay 500 usee after the
sync pulse before sampling begins, type 10, followed by the RETURN key.
Again, if a typing error is made, type CTRL/Q and the number of sweeps
and delay can be respecified.

NOTE

When a sync pulse is sensed via the Schmitt trigger,
the program waits one sample interval before the
This timing wait is in
first A/D sample is taken.
addition to the delay specified in answer to D:.
It is possible to alter these parameters at any time during the experi-

mental run by typing CTRL/A.
for
The rate specified is the total rate on all channels, not the rate
Control
for
specified
is
rate
same
if
the
example,
For
each channel.
is four times
Tape 1 and Control Tape 3, the sampling rate on channel
is
used.
faster when Control Tape 1
However, a
2
33 microseconds is the maximum rate for multiple channels.
single channel can be run at a maximum rate of 25 microseconds.
x

2-5

The program now begins showing
the input signal on the CRT display,
a
colon and two LINE FEEDs are typed and the input
signal appears on the
oscilloscope. Adjust the threshold for the
Schmitt trigger until a
partial display is shown. The partial display
assures that the trigger
is firing.
To begin averaging, type the RETURN
key.
The summation will
then be shown as it accumulates.

The accumulated sum may be cleared at any
time by typing CTRL/Z.

Type V at any time to switch from viewing
the input to viewing the
summation of the sweeps or vice versa. Type X
to expand the viewing
scale.
Type C to contract the viewing scale of the
summation.
The
input data cannot be scaled.
Type LINE FEED to pause averaging at any point,
and then type the
RETURN key to resume. Typing LINE FEED always
forces the display to
the view inputs mode.
The sampling continues until the number of
sweeps requested in the parameter setup have
been taken.
The average
may be prematurely terminated by typing CTRL/Q.
If the average has been completed, it can
be
command P to enter plot mode, turning on the
typing the RETURN key to move the pen to the
Lower the pen and type the RETURN key. Lift

plotted by typing the
X-Y plotter, and then
beginning of the plot.

the pen and type the
RETURN key again.
The pen moves to the beginning of the plot.
Lower
the pen and type the RETURN key again to replot
the data or plot the
next of a series of curves.
Terminate plot mode by typing P and turning off the plotter.
Type the RETURN key.
The display will again
b e come ac t i ve

A completed average can be output on the Teletype
by typing the command T.
The message

is typed.

If it is desired to dump all of the
accumulated averages,
answer this with a
or the RETURN key.
If it is desired to type
selectively, one entire channel or selective
portions of a channel,
type 1, 2, 3, or 4 plus the RETURN key to type
out the first, second,
third, or fourth channel in the display.
The message
L:

2-6

Answer this with the low limit in terms of the data point
number in which you are interested and the RETURN key. The message
is typed.

Answer this by typing the high limit data point to be output and the RETURN key.
is typed.

The first two numbers typed out represent the number of sweeps in-

cluded in the average and the scale factor, respectively.

The scale

factor is the power of two by which the data has been scaled.

If the

number of sweeps is a power of two, then, by typing C and X while the
display is active, the gain can be adjusted until the scale factor is

appropriate for the number of sweeps.

That is,

factor power is equal to the number of sweeps.

raised to the scale

The typeout of data

points will then be in millivolts as seen at the input to the A/D
Converter.

Following the typeout of the number of sweeps and the scale factor,
the value of each point in the average is typed out in millivolts.
NOTE
If a single input A/D is used (input voltage ±5V)
the point values should be multiplied by a factor
of 5 to yield the true value.

To terminate typeout at any time, simply type CTRL/Q.
If another set of sweeps is to be added to the average already accumu-

lated after a group of sweeps has been completed, type CTRL/R to retain the data and reinitialize for more averaging.

To begin averaging after reinitialization has taken place, type the
RETURN key.
If it is desired to start averaging with a zero buffer, type CTRL/Z
and the RETURN key.

2-7

READ CONTROL
TAPE

Parameters
No. of channels
No. of points

Desired statistics

SPECIFY DELAY
(POSITIVE ONLY
IN MULTIPLES OF
SAMPLING RATES)

READ BASIC
AVERAGER

PRESS CONTINUE
START PROGRAM

\~<D
TYPE RETURN

TYPES R:
(ASKING FOR
SAMPLING RATE
IN y SEC/POINT)

AVERAGE IS
ZEROED

SPECIFY
SAMPLING RATE

SIGNAL INPUT
IS VIEWED

TYPE RETURN

TO START
AVERAGING
TYPE RETURN

TYPES N:
(ASKING FOR
NUMBER OF
SWEEPS)

USER SEES
ACCUMULATION
OF THE SWEEPS

SPECIFY NUMBER
OF SWEEPS

TYPE RETURN

TYPES D:
(ASKING FOR
DELAY, IF ANY)

Figure 2-1

Basic Averager:

2-8

User Flow Chart

KD
<D

READS IN
DISPLAY
SECTION

TYPES > WHEN
DONE AND
DISPLAYS THE
SUM OF THE
SWEEPS

PRESS CONTINUE
TO START
SECTION 4

DISPLAYS THE
AVERAGE AND
STATISTICAL
FUNCTIONS

TYPE CONTROL P
READER READS
SECTION 3 OF
THE ADVANCED
AVERAGER

TYPE I TO SUM
BETWEEN
CURSORS

PRESS CONTINUE
TO START
SECTION 3

20 TO 50 SECOND

WAIT FOR
COMPUTATIONS

Figure 2-1

Basic Averager:

2-9

User Flow Chart (Cont)

CHAPTER

COMPUTATIONS OF STATISTICS AND OUTPUT

3.0

COMPLETION OF CONFIDENCE LIMITS AND TREND COMPUTATION

Although not strictly a part of the Basic Averager
system, statistical
analysis of data can be accomplished, if the
eight-channel multiplexer
is part of the system.
If the eight channel multiplexer is not
implemented, the user is urged not to proceed, but he
may do so at his
own risk.
Typing the CTRL/P command advances the user to
two statistical sections following the Basic Averager on
the paper tape provided.
Do not proceed

Control
Tapes

Basic
Averager

Figure 2-2

if A/D multiplexer is not
part of the
system

Advanced
Averager
Section
III Status

Advanced
Averager
Section
IV Output

Block Diagram of the Basic Averager

After the first statistical section (Section III)
has completed loading and the accumulator equals zero, press
CONTinue.
Computations
will then be done on the raw statistics calculated
while the averager
was in operation (provided control tapes requesting
confidence limits
and trend were used)
If the accumulator does not equal zero after
reading, an error has
occurred.
Reposition tape to the beginning of Section III and
press
the CONTinue key on the console twice.
The

rest of the Basic Averager

is the same as the Advanced Averager

(Book One, Section 2.3).

On output (Section IV) the number of traces (1-4)
which appear on the
oscilloscope depends on the control tape used.
To begin computing new averages, remount a
control tape, type the
CTRL/P command, and load the Basic Averager,

DEC-8E-ACA0A-A-PB.

2-10

CHAPTER 4
BASIC AVERAGER MS

4.0

GENERAL DESCRIPTION

Basic Averager MS is the same as the papertape version except that
it will chain in Sections 3, 4 and 5 of the Advanced Averager.

When the Basic Averager is done and in output mode, Section 3 of the
Advanced Averager can be loaded by typing a CTRL/P. The binary tape
of the Basic Averager is DEC-8E-ABA0A-A-PB and the binaries of the
control tape are
Control tape

DEC-8E-ABA1A-A-PB
DEC-8E-ABA2A-A-PB
DEC-8E-ABA3A-A-PB
DEC-8E-ABA4A-A-PB
DEC-8E-ABA5A-A-PB
DEC-8E-ABA6A-A-PB
DEC-8E-ABA7A-A-PB
DEC-8E-ABA8A-A-PB

1
2

3
4
5
6
7
8

4.1

Using the Basic Averager MS

Loading the Programs
NOTE
Mass Storage output from BASIC Averager is done
by Section 5 of the Advanced Averager.
If no
MS output is desired, the BASIC Averager can
be run without Sections 3, 4, and 5 of the Advanced Averager.

If Sections 3,

and

of the Advanced Averager are to be used to

output data as in the Advanced Averager, the following programs must
be saved as SAVE files under OS/8.

Sections 3, 4 and 5 of the Advanced Averager must be saved after being loaded with the Absolute
Loader. (Refer to Book One.)

Section

4 -

SAVE SYS AAVG3.SV 6400-7577 ;06777=0
SAVE SYS AAVG4.SV 0-177 6400-7577 ,-00017=0

Section

5 -

SAVE SYS AAVG5.SV 6400-7177,-06377=0

2-31

The Basic Averager can be loaded
with the Absolute Loader in the order

*CONT.BN, BAD2.BN

where CONT.BN is the control tape and
BAD2.BN is the Basic Averager.
As with the paper tape system the
control tape binary must be loaded
then the binary of the Basic Averager.
The binaries of the control
tapes and the Basic Averager are stored
on the user's OS/8 device.
Any one of the eight control tapes can
be used.

STARTING LOCATIONS:
If Sections 3, 4 and
at 400, field 0.

of the Advanced Averager are to be used,
start

If they are not going to be used, the
program can be started at 6505.
It still can be started at 400 if the
saved Sections 3, 4 and 5 of
the Advanced Averager are present on the
system device. A control
tape and the Basic Averager can be saved after
being loaded, if desired.

The save locations are:

BASIC Averager

0-577,

Control Tape
Control Tape
Control Tape
Control Tape
Control Tape
Control Tape
Control Tape
Control Tape

6134
4175,
3233,
6134
3603,
2444,
6134
3437,

1
2
3
4

5
6

7
8

6200-7577

6147
4221,

5207,

6175

6147
3627,

5012,

6175

6147

Running the MS Basic Averager

Operation is the same as the paper tape system
except the programs
come off mass storage.
CTRL/P in output mode will chain in Section
3 of the Advanced Averager.
From there on the Advanced Averager commands
are active.
For example, to save Control Tape

SAVE SYS BASAV2.SV

the command would be:

0-577, 6200-7577, 4175 , 6147; 400=0

2-12

Errors

;

The BASIC Averager has only one error message:

USER ERROR 1
which indicates a chaining problem.

2-13

BOOK

TI H

PST

THREE

AND

LATENCY

HISTOGRAM PROGRAMS

PREFACE

the Post
a)
Book Three describes the operation of three programs:
Stimulus Time Histogram, b) the Latency Histogram, and c) the Time
Interval Histogram program. Programs a) and b) are contained on one

binary tape (DEC-8E-APLHA-A), program c) is contained on DEC-8E-ATINA-A-PB.
DEC-8EIn addition, the mass storage version of each is discussed:

ATIHA-A-PB for TIH and DEC-8E-APSTA-A-PB for PST.
Histogram program records all signals
crossing a set threshold after a given event. This program, which
produces a histogram representing the sum of all activity following a
particular stimulus, is most frequently used in experiments dealing
The LAB8/E PST and Latency

with the response of a single nerve cell to a stimulus. The histogram
is displayed on an oscilloscope as a frequency of event versus time
plot.

The LAB8/E Time Interval Histogram program records the time intervals
between events. These events are defined as a crossing of a set threshold.

The program records the intervals by means of a histogram or

frequency distribution graph; where the time interval is plotted
against its frequency of occurrence. The program is generally used
to record the spontaneous activity of single nerve cells or to establish the mean time interval between peaks of an analog signal like an
e lect rocar diogram

3-i

CONTENTS

Page

CHAPTER 1

GENERAL DESCRIPTION

3-1

1.1

3-1

1.2

Hardware Requirements
The Program

1.3

Connections

CHAPTER 2

CHAPTER

TIH, PST AND LATENCY HISTOGRAM
PROGRAMS

1.0

3-2

3-3

PROGRAM EXECUTION

2.1

Program Loading

3-5

2.2

Running the Program

3-5

MASS STORAGE

3.0

GENERAL DESCRIPTION
Using MS TIH, MS PST and Latency
Programs
Data Output Format on Mass Storage

3.1
3.2

TIH, PST AND LATENCY HISTOGRAM
3 _ lg

3-l 8
3-19

TABLES
3-1
3-2

Data Collection Commands
Output Mode (OPM) Commands

3-12
3-14

FIGURES
3-1

Timing Relationships

3-ii

3-10

CHAPTER 1
TIH, PST AND LATENCY HISTOGRAM PROGRAMS

1.0

GENERAL DESCRIPTION

The LAB8/E Post Stimulus Time (PST) and Latency Histogram program
records all signals crossing a set threshold after a given event.
The program is used most frequently in experiments dealing with the
response of a single nerve cell to a stimulus. After the primary

event (SI or a stimulus) the program measures the time of occurrence
of all subsequent events (responses) above a certain analog threshold.
The PST histogram represents the sum of all activity following a
It is not an analog summation (signal averaging) but an
stimulus.

event summation which can be interpreted as a distribution of the
probability that an event will occur at time T following a stimulus.
The histogram is displayed on the oscilloscope as a plot of frequency
The program allows the user to specify
of event (y) versus time (x)
.

the number of firings following a stimulus to be recorded, the resolution of the time axis (bin width), the duration of the time axis,

(number of bins) , and the range of the time axis (the minimum time
which must pass after the stimulus before analysis begins)

The Time Interval Histogram (TIH) program records the time between
events occurring on one input (i.e., Schmitt trigger 2). The display
is a frequency distribution showing the duration of the time interval

versus its frequency of occurrence.
1.1

Hardware Requirements
4K

PDP-8/E

AD8-E

10 bit A/D converter

AM8-EA

16 channel multiplexer,

DK8-ES

Real-time clock

VC8-E

10 bit point plot display

Tektronix
602,

603, 604

or VR14
or RM503

}

Display oscilloscope

3-1

channels implemented

An OS/8 configuration is necessary to run
OS/8 versions of histograms.
1.2

The Program

The PST and Latency Histogram program provides
information about the
activity which follows a stimulus or synchronizing
event.
The activity
is defined by a signal which will fire
a Schmitt trigger (the characteristics of a signal which cause a firing are
described in Section
1.3).
It is this event (the firing) and its time of
occurrence which
are detected by the program.
In addition to recording the time of
occurrence of events following a stimulus, the PST
and Latency Histogram program can show a zero th histogram. This is
a graph showing how
much activity followed each stimulus. The Y axis
is the frequency of
occurrence where the X axis is divided into units of
epochs (1 epoch =
This feature is useful in checking to see if
1 unit).
the overall
activity following a stimulus remains constant through
time or decreases
due to habituation or fatigue.
The TIH program also has a zero th histogram.

This display shows how
much activity occurred during each time segment
of the experiment.
(The duration of the run is specified by defining
a basic time unit
and then specifying the number of units to be
considered.

Before data collection can start, the scheme
for recording events
and their time of arrival must be defined.

Latency

A post stimulus time histogram is a record of all
activity
following a stimulus for a set period of time.
In certain studies, it
is desirable to have a record of only the
first few nerve potentials
following a stimulus.
This is called a latency histogram.
The number
of potentials recorded following the stimulus
is called the order of
the latency histogram.
The program can form either a PST or latency
:

histogram.
Time Scale
Since the LAB8/E is a digital device, time must
be counted
in increments rather than as a continuous
function.
The finer the increments, the more closely the real situation
is approached.
The
majority of biological phenomena occur slowly
enough that this departure
from the real time world does not impose severe
restrictions.
Therefore, the first parameter we must assign is
the time scale increment,
or resolution.
This is called the bin width.
Since approximately 100
:

3-2

ysec are required to acknowledge the tick of a clock, bin widths are

defined in multiples of 100 psec.
Once the units of the time scale are defined, the length and range of
scale must be assigned.
This is done by supplying the number of bins

which should be recognized, i.e.,

(bin width)

x (number of bins) yields

time scale length and time at which the time scale should begin.

The

latter parameter is defined in answer to the question "Minimum time?"
(where minimum time is in terms of bins).

As an example:

is defined as having bin widths of 5.0 milliseconds.

bins and has a minimum time of

bins.

A histogram

It consists of 100

The time scale would start at

time zero (time zero is defined by the occurrence of the stimulus in
the case of the PST and Latency Histogram program; in the TIH program,
time is relative to each event) , and range in increments of

milli-

If a
seconds to time = 500 milliseconds [5.0 milliseconds x 100).
minimum time had been specified, the starting and ending times of the

range would be incremented by that amount (e.g.

therefore,

minimum time =

bins;

x bin width = 10 milliseconds total delay of the start of

the range; thus the range would start at 10 milliseconds and end at 510
milliseconds.

Minimum time is used usually for one of two reasons.

Either it is

used to avoid artifacts which the stimulus might cause or it is used
to move the time scale to the time of interest (i.e., if precise

measurements had to be obtained at a specified time after the stimulus
and no other time was of importance in the study)
Once the time scale is defined, the only other parameter that must be

specified before data taking can start is the duration of the experiment.
In the case of PST and Latency this is defined by specifying the number
of stimuli to be recorded.
The TIH has its duration defined by specifying a basic time unit; and the number of times this time unit should
occur from the start of data collection. This timing feature is useful
in studies where the nerve cell is prone to fatigue.

After the preset

time has elapsed, the program will notify the user and he can either

output his data or return and take more data.
1.3

Connections

A lead which contains a synchronization pulse

pulse occurring at

the same time as the stimulus) should be connected to Schmitt trigger

number

A second lead which contains the response activity or the

3-3

spontaneous activity should be
connected to the Schmitt trigger number
2
The Schmitt triggers are threshold
detectors and can accept pulse or
analog inputs.
In either case, the circuit will fire
when the trigger
threshold has been exceeded after having been
0.3 volts above or below
the trigger threshold.
The trigger threshold is set with the
knob corresponding to the phone jack input connector.
This threshold varies
The input to the Schmitt trigger must be
± 5 volts.
kept to within
A technical description of the Schmitt trigger
± 50 volts.
is contained in Appendix C.
During the trigger initialization section,
a lead containing an analog signal may
be connected to analog input 1
or 2 but this is not necessary during the
remainder of the program.

3-4

CHAPTER
PROGRAM

EXECUTION

Program Loading

The Binary Loader (BIN) is used to load either of the histogram

programs.

If BIN is not present, refer to Appendix A for loading

instructions.

(For loading the program from the Disk Monitor system,

follow the instructions given in the Monitor handbook. The SAVE
command should be SAVE PST! 0-4177, 7420-7577; 200 and SAVE TIH!
0-4177, 7430-7577; 200).
To load the histogram program, use the following procedure:

Place the tape in the tape reader.
Set the Switch Register to 7777.

Press the ADDR LOAD key.

d.
e.

If using a high-speed reader, set the Switch Register to 3777.
Turn the Teletype control knob to LINE.

Turn the scope to ON.

g.
h.

If using a low-speed reader, set the reader switch to START.
If using a high-speed reader, press reader ON switch.

Press the CLEAR and CONT keys on the computer.

See Chapter

for OS/8 version of the programs.

Running the Program

After the tape has loaded in, verify that the accumulator lights are
all off.
If they are not all off, there has been a reader error and
the program should be reloaded.
After loading is complete and correct,
set 200 in the Switch Register, press ADDR LOAD, then press CLEAR and
CONT.
At this point, the message

TRIGGER
is printed.

3-5

Trigger Initialization
At this point, adjust the Schmitt trigger
thresholds (described in Section 1.3) by moving the clock knobs corresponding
to the clock input
connections so that the triggers are fired on data
and not noise spikes.
The trigger thresholds should be adjusted
individually to avoid confusion, so during the initialization only one
lead should be connected to
the Schmitt trigger inputs at a time.

When the trigger fires, the scope will
show a line across its face.
The presence or absence of this line
indicates whether or not the
Schmitt trigger is firing.
If desired, a lead may be attached
in
parallel (with the line going to the Schmitt
trigger input) to the
analog input number 1 or 2
Now, whenever the trigger fires, two
traces will appear on the oscilloscope.
One is a sweep of the analog input and the other is a baseline to
represent the voltage at
which the trigger fired. This feature is useful
if a slowly varying
analog signal is used to fire the trigger.
By viewing the input, you
can see where on the waveform the trigger is
firing.
The sampling rate
of the analog channel is set by analog knob
3 on the front of the
laboratory peripheral panel: counterclockwise
to sample faster; clockwise to sample slower.
.

Once the trigger thresholds are set
satisfactorily, all leads except SI
and S2 should be removed to avoid
confusion.
Type the RETURN key to
advance the program.
NOTE

After the program has left the trigger section,
it
is impossible to return to this
section under program control.
The program must be restarted by the
switches at location 7424 for PST and
7432 for
TIH.
If data has been collected, this
section
may have been overwritten by data.
If this is
the case, the program must be reloaded to
recalibrate the Schmitt triggers.
Parameter Setup
The user must now define the PST or
Latency Histogram by answering a
series of questions.
The questions will be asked and the
user must
respond in the format described below.
The user must type the RETURN
key after all answers.
Leading zeros are not required.
If the question is improperly answered a ? is printed
and the question is repeated.
To back up one question to redefine a
parameter, type the LINE FEED key.

3-6

If a mistake in answering is made, type the RUBOUT key to repeat the

current question.

To redefine all of the parameters, type CTRL/A

(this is read as "control A").

To type it, hold down the CTRL key

CTRL/C returns control to the Monitor, if one exists,
If no Monitor exists, refrain from
at any time during the program.
and type A.

issuing this command.

Question

(not asked in TIH)

First of all, the program wants to know if it
is to form a Latency Histogram or a PST Histogram (for description, see Section 1.2).

PST?

Answer Y (for yes) N (for no) followed by the
RETURN key (RETURN is signified by J
,

)

If Y is the answer, question

asked.

Question

will be skipped and question

will be

If N is typed, the following question will be asked.

(not asked in TIH)

For a Latency Histogram, the program must know
the number of theshold crossings following the
If a frequency distribustimulus to record.
tion of the first pulse following a stimulus
For a distribuis desired, 1 should be typed.
tion of the first two pulses, 2 should be typed,

ORDER=

etc.

Answer with a number between
and 4095.
This
signifies the number of pulses to record following the stimulus.
If all of the pulses following are to be recorded, Question 1 should have
been answered Y.

Question

(Question

BINW=

for TIH)
The
The question is asking for the bin width.
bin width is the basic time unit for the time
scale used in building the histogram.

Answer with a number between 0.1 and 409.5. The
number is interpreted as milliseconds and must
be in the form x.x \.
Question

BINS=

(Question

for TIH)
The number of bins and bin width determine the
range of the time scale.

3-7

Answer with a number between
and 1890.
if more
bins are requested, there will not be
enough core
for data storage; the message CORE? will
be typed
then the question will be asked again.
Question

(not asked for TIH)

EPOCHS=

This question is to define the number
of times
the stimulus will be presented (the
number of
epochs)

Answer with a number between
and (1890 - #
Again, it is possible for this
answer to
request more core than is av ailable.
ze r 8
histogram is a graph showing how many
firings
were recorded consequent to each
stimulus.
Therefore a core location is needed
to record the
count for each epoch.
if core is exceeded, the
message CORE? is typed and the
question is askld
again.
Simply reduce the number of
number of bins until the answer is epochs or th*
accepted?
bins)

Question

(Question

MINTIM=

i V

for TIH)

This question asks the starting time of the
time
range under consideration.

Answer with a number between
units are bin widths.

Question

(for TIH only.

TIME\UNIT=

and 4094.

The

Not asked for PST)

This question asks what the basic time unit
is
for timing the experimental trail.
It is unrelated to the histogram time units.

Answer with a number between jZf.l and 409.5
The
number is interpreted as milliseconds and must
be
in the form x.x
J.
Question

(

#UNITS=

for TIH only.

Not asked for PST)

This question requests the number of basic
time
units defined in the previous question which is
to be used for the experimental trial.

Answer with a number between
and (1890 - # bins)
Again, it is possible for this answer to
request
more core than is available.
The zero th histogram
is a graph showing how many firings were
recorded
consequent to each stimulus.
Therefore, a core
location is needed to record the count for each
epoch.
If core is exceeded, the message CORE'
is typed and the question is asked
again.
Simply reduce the number of epochs or the
number of bins until the answer is accepted.

After the last question is answered, the
program waits for CTRL/S to
start data collection.
(Type CTRL/A to repeat all questions; CTRL/C
to return to the Monitor.)
For an example of parameter setup, see
Appendix K.

3-8

Data Taking

After the questions have been answered, type CTRL/S to start data
taking.
The LAB8 will now start displaying the first histogram.
This is the PST or Latency Histogram which records the responses and
their times of arrival following the stimulus; or the TIH displaying
the occurrence of various intervals
The X axis of the display is the time range.

At the start of data
taking, the time range consists of all the bins which were requested.
These bins are shown with zero content at first so that their heights
are coincident with the base line. As responses are registered, the
bins will show themselves.

If a minimum time is specified, the first

bin on the left-hand edge of the X axis is the underflow bin.
That
is, the bin records all responses that occur at some time before the

specified minimum time has elapsed. If no minimum time is requested,
there will be no underflow bin and the first division of the X axis
will be the first bin of the time range. The rightmost bin in the TIH
is for all inter-pulse intervals greater than the largest bin.

It is

called the overflow bin.
NOTE
The counts that fall into the underflow or overflow bins are not recorded in the zero tn histogram.
The zerotn histogram records only those
counts that occur during the specified time range.
The Y axis is also shown as a straight line.
It differs from the X
axis in that it contains 25 hatch marks spaced 40 raster points apart.

These marks are provided so that bin content can be estimated visually
during data taking. These hatch marks remain the same distance apart
regardless of the Y scale factor.
There is one other display which can be viewed. This is called the
zero
histogram.
To see this display, type <.
This histogram is a
graphical representation of the number of pulses which are received
after each stimulus or during each time unit. The X axis here has an

increment for each epoch (rather than for each bin). The first epoch
is shown at the left-hand end of the X axis.
This histogram is useful
for determining the gross activity of the neuron through time.

The display will remain active during data collection.

algorithm for PST and Latency is the following.
3-9

The collection

The program waits for

a stimulus,

(Schmitt trigger 1).
Until a stimulus is recognized,
no responses are acknowledged.
When SI is received, the program accepts only responses (S2's) until the end of the
time range is reached
(i.e., total time = minimum time + (# of bins
x bin width)).
During
this time, no Si's are acknowledged, so the
user should make sure that
the stimulus repetition rate is slower than
the length of the analysis
time period.
Figure 3-1 describes the timing relationships and
the
signals that are acknowledged.

Stimulus

Illegal Pulse
Error Condition
Noted
I

Only Si's
accepted

(SI)

Response (S2)
,

Ignored

Only S2's
accepted

Ignored
End of
Epoch

fr=0

Figure 3-1

T=0

Timing Relationships

From the diagram it is clear that only those
response pulses occurring
during the analysis time are counted.
An error condition is noted if
an SI (stimulus) is received during the
analysis period.
The time
counter does not reset to zero so the SI is
effectively ignored. The
occurrence of the untimely SI is noted and is
reported at a later

time.

The collection algorithm for the TIH is the
following:
the time between
events seen on Schmitt trigger 2 is measured and
then recorded in the
appropriate bin.
If the time interval measured were 2.5
milliseconds
and the histogram had no minimum time constraints
with a bin width of
1 millisecond, the third bin of the histogram would
receive a count.
The first bin is bin 0; so bin 2 would be
incremented by one.
If a
minimum time is specified, the underflow bin, or
the minimum time bin,
is bin 0.
The overflow bin is always the last bin.
If at any time during data collection the
contents of a bin in the histogram exceeds 4096, the overflow is noted and
the contents are set to
Every time a bin is filled over 409 6, the
4096.
bin is set to its maximum and the overflow is noted. The errors are reported when data

collection is finished.

3-10

Commands During Data Collection
During data collection there are several commands which can be
executed. Most of these commands operate on the display.
Table 3-1
is a list of possible "Group I Commands."

Note that, although the
display is changed, the data remains unaltered.

After data collection is complete, the program transfers to the outgut mode (OPM)
This transfer takes place after the user-specified
.

number of epochs is complete or after CTRL/Q orCTRL/P is typed during
data collection.
The computer outputs a carriage RETURN/LINE FEED,
the message OPM, another carriage RETURN/LINE FEED, two four-digit

numbers

separated by commas, and then resumes the display.

ERROR CONDITION NOTIFICATION:
The two four-digit numbers that are typed represent the number of
times two types of errors occurred during data collection.
Error

condition one (the first number) arises if an SI or stimulus pulse
is received during an epoch (refer to Figure 3-1)
These SI pulses
.

have no effect on the data collection, but are noted as errors and
reported here. Error condition two (the second number) arises if
the contents of any bin exceed 4096,-.
Each time the overflow occurs,
it is noted, and the bin content is reset to 4096.

Only error condi-

tion two is reported in the TIH.
At this time, all of the Group I commands are still active as well
as the set of Group II commands described in Table 3-2.

Before using the program, the user should thoroughly familiarize
himself with the Group I and Group II commands summarized in Appendix N.

3-11

Table 3-1

Data

COMMAND

Collection Commands
ACTION

Display the next highest histogram.
In
the present histogram programs there are
only two possible histograms to display.
Therefore this command.will only move the
display from the zero™ to the first
histogram.

Basically the same as > but moves the display
back one histogram. Note that the display will not
be forced to display nonexistent histograms
by repeating these commands
CTRL/Z

Stop data collection.
Zero out all buffers
and error counters.
Wait for CTRL/S to
restart collection.

CTRL/A

Go back to parameter setup and Alter parameters.

CTRL/Q

Quit data taking at the next end of epoch
and transfer to output mode (data taking then
may be resumed, saving all data by CTRL/R)

CTRL/P

Panic Stop.
Immediately transfer to output
Do not wait for end of epoch or anything else.

mode.

CTRL/C

Immediately return to Monitor (this command
should not be used if system Monitor does
not exist)

Each time this command is issued, the display
alternates between displaying Cursors or no
Cursors.
The cursors are two vertical lines
which are controlled by analog knobs
and 1.
Their relative positions are unimportant.
U

Scale the dimension Y Up by a factor of 2.
This does not affect the actual bin values,
only the display.

Scale Y Down by a factor of 2.
be given any number of times.)

and D may

Expand the area of the histogram enclosed by
the cursors so that the area takes up the
entire X axis.
(E may be done any number of
times
.

3-12

Table 3-1

Data Collection Commands (Cont.)

ACTION

COMMAND
J

Return to the Beginning expansion factor
(i.e., no expansion).
In other words,
cancel all E commands.

Before display, Sum adjacent bins. This command adds adjacent bins and displays the resulting histogram. This effectively doubles
the bin width.
It also can be considered a
type of smoothing routine.
At the start of
the program, the sum factor is set to one.
Each time the S command is executed, the sum
factor is increased by one. When sum factor
= 1, each bin is displayed separately; when
Sf=2, two bins are added together and displayed as one. When Sf=3, three bins are
added together and displayed as one, etc.
Set sum factor to
sum factor.

or return to Original

Zero all display changes or set all displays
to unmodified state.

Each time this command is issued, the display alternates between displaying an Axis
or no Axis.
This command is included for
aesthetics only.
Its sole purpose is to
yield a clearer display.
V

Each time this command is given, the display switches between Viewing the histogram
in a bar graph display or a point plot display.

CTRL/U

This is an undefined User command which can
be implemented by following the instructions
in Appendix L.
Unless the program is modified, it will have no effect.

3-13

Table 3-2

OPM COMMANDS

Command

Action
Type out the data enclosed by the cursors.
Before
the T command is issued, the cursors should be placed
so that they surround an area of interest.
When T is
typed, the scope goes blank and the Teletype types
the contents of the bins between the cursors.
If the
cursor line goes through any part of a bin, that bin
is included in the typeout.
To terminate the typeout
before the right cursor is reached, type CTRL/Q. Until the user is familiar with the typeout portion of
the program, the following procedure is suggested.
Bring up a bar graph display.
Position the cursors
on the bins that delimit the area of interest.
Type
T.
A sample of the output is the following:
1,

0001,

0000010, 0000229,

0053, 0006,
0055, 0002,
0056, 0002,

.6000,
.2000,
.2000,

,0436, 0053, 0056,

.0262,
.0087,
.0087,

Starting with the leftmost number on the first line,
the interpretation is:
the histogram being typed out is the firstorder
;r histogram (not
(r
zero t
the
sum factor
0001
0000010 the total number of counts between the cursors
0000229 the total number of counts in the entire histogram
the ratio of the number of counts between the
.04 36
cursors to the number of counts in the entire
histogram is .0436 to 1 or 10/229.
bin number 5 3 is the leftmost bin being typed
005 3
bin number 56 is the rightmost bin
0056
on the second line, bin number 53 contains six
005 3
(0006) counts
the ratio of the number of counts in this bin
.6000
to the number of counts between the cursors
is .6 to 1 or 6/10.
.0262
the rate of the number of counts in this bin to
number of counts in the entire histogram is .0262
to one or 6/229.
Each bin is reported in this manner.
If a bin has zero
contents, its typeout is skipped.
In this example, bin
54 was omitted because it contained nothing.
Commas are
provided to make any output tape compatible with BASIC/RT
(refer to Book Seven)
)

Note that, if a minimum time is specified, bin
represents a count of all S2's occurring before the minimum
time is elapsed.
It is the Minimum Time Bin (or underflow bin)
If no minimum time is specified, bin
is
the first bin of the histogram.
The total number of
bins displayed is always equal to the user-specified
number.
.

3-14

Table 3-2

OPM Commands (Cont.)

Command

Action

Another fact to be considered is that when the histogram
has been subjected to the Sum command, the bin numbers
refer to the bin numbers of the original, unaltered
histogram.
The bin number typed out refers to the leftmost bin that is summed with its neighbors on the right.
Example:
Bin
Bin
Bin
Bin

10
11
12
13

has
has
has
has

4
3
8
5

counts
counts
counts
counts

If S is typed once and the T command is then issued, the
typeout will be the following:

BIN 10;
BIN 12;

COUNTS

13 COUNTS

NOTE
When S has been typed, a skipped bin number does not necessarily mean a bin of zero
contents.
Similarly, when the E command
is executed, followed by T, the bin numbers
refer to the original unaltered histogram.

Calibrate an X-Y analog plotter. When the K command is
given, the outputs of the oscilloscope go to the maximum
positive voltage for the X and Y coordinates. When any
key on the keyboard, except CTRL/Q, is typed, the outputs
go to the maximum negative voltage.
After the outputs
have gone to the maximum negative voltage, typing any
other key causes the outputs to go back to positive
voltage.
This process may be repeated as many times as needed
while the bias and gain controls for the X-Y plotter
are adjusted (see Appendix H for details of attaching
an analog recorder)
When the plotter is calibrated,
type CTRL/Q to resume the display.
.

Plot the histogram being displayed on the analog X-Y
recorder.
The histogram is plotted in the exact scale
as it is displayed, in bar graph form, with an axis and
no cursors.
After the initial command, the program waits
for the user to turn on his plotter and lower the pen.
Type any key on the keyboard and the axis and histogram
will be plotted. The speed of the pen can be controlled
with analog knob 3. Adjustments of the speed can be made
while the plotting is taking place. Turn the knob
counterclockwise to accelerate the plot and clockwise to
slow the plot.

3-15

Table 3-2

OPM Commands (Cont.

Command

Action
When the axis and histogram have been plotted, the
program pauses again. Raise the pen and turn off the
plotter.
If the plotter is on while the display is
active, the pen will shake furiously, trying to keep
pace with the changes in analog voltage.
When the
plotter is turned off, type any key and the display
will be resumed.

CTRL/B

Causes a binary dump of the histogram being displayed.
NOTE
The histogram is returned to normal
size and scale before the binary
dump takes place.

After the initial command of B, the program prints
HIGH?. Answer Y J
(for yes) if the dump is to proceed on a high-speed punch or N
J if the Teletype
punch is to be used. After this initial question,
the program prints ID.
Enter a maximum of five characters followed by the RETURN key for an identification
code.

CAUTION
The following characters may not be used
in the ID code:
RUBOUT, CTRL/C, or CTRL/A.
A LINE FEED at this point repeats the question HIGH? Typing the RUBOUT key causes
the ID question to be repeated.

After the ID code has been entered, 6 4 leader/trailer
punches are punched.
If the low-speed punch is to be
used, it should be turned on during the leader/trailer.
Following the leader/trailer code, the ID code is
punched in ASCII and followed by another 64 leader/
trailer punches
After the identifying information has been output, a
core image of the display list (the pertinent information relating to the histograms for display purposes)
a core image of histogram (which was being displayed
when the CTRL/B command was executed)
and a checksum
are punched.
A core image dump allows the papertape to
be read back into the computer in the same form as it
was originally dumped.
The binary loader can be used
to reload the data.
This feature allows the user to
take data, dump it and then return at any later time to
process it or review it.
The user may write his own
processing program or read the data back into the histogram program. For details of binary format, see Appendix M.
To review the data with the program, read in
the program as described in Section 2.1 (if it is in
core, reloading is not necessary).
After the program
,

3-16

Table 3-2

OPM Commands (Cont.)
Action

Commands

load the data with the binary loader using
the same procedure as if the main program were being
After the data and program are in core, start
loaded.
the program at location 1375 (place 1375 in the front
Now
Press ADDR LOAD and CLEAR and CONT.)
switches.
operate the program is if the output mode had just been
entered.
is in core,

Note
Since the CTRL/B command dumps only
the histogram being displayed, only
one histogram will be displayed
properly after data readin.
If this program is run from disk or

DECtape, the histogram can be dumped
by the following procedure.
1.
2.

Type CTRL/C - Monitor returns.
Save locations 4000-7577 as a binary

3.
4.

Restart the program at location 1375.
Continue as if a binary dump had
taken place.

file.

To review this data, load the program
Start at location 1375.
then the data.

After the binary dump has taken place, the program returns to the active display and all of the commands are
operative
CTRL/I

This is a user Group II command which may be Implemented
Until a
See Appendix L.
in the same manner as CTRL/U.
patch is made to implement it, issuing this command does
nothing.

CTRL/Z

Zero all buffers and error counters.

CTRL/R

Retain the data in the buffers so that data collection
The effect of
may be resumed without loss of data.
is to allow it
histogram
order
first
this command on the
The effect
to accumulate another preset number of epochs.
of this command on the zero th histogram is to set the
That is, the first bin will
epoch pointer back to zero.
be incremented by the number of pulses received during
The error counters are not reset.
the next epoch.

CTRL/S

Follows either CTRL/R or CTRL/Z to start data collection
again.

CTRL/A

Return to redefine parameters.
3-17

CHAPTER

MASS STORAGE TIH, PST AND LATENCY HISTOGRAMS

3.0

GENERAL DESCRIPTION

The mass storage version of the histogram programs is
the same as the
paper tape program except that it runs under OS/8 and can read
and
write the histogram to and from the system device.
The tape numbers
for the binary program are:

3.1

DEC-8E-ATIHA-A-PB

for TIH

DEC-8E-APSTA-A-PB

for PST

Latency

Using MS TIH, MS PST and Latency Programs

Loading the Programs
The programs can be loaded via the Absolute Loader and
started at 200
of field 0.
If the programs are to be saved with the SAVE command
the core locations to be saved for both programs are 0-7577 of field

and 7400-7577
For example, to SAVE PST, load the binary
via ABS Loader,
terminated by ALTMODE and save the core image via
the SAVE command,
SAVE SYS PSTD1.SV 0-7577 , 17400-17577 ;
of field 1.

200.

Starting Locations
The normal start for both programs is 200 of field
if data is not
0.
to be collected but a file is to be read in
and reviewed, the
program can be started at 1375.

Commands

CTRL/W is used to read or write the histogram
data to mass storage.
When CTRL/W is typed, the OS/3 command decoder
is called to request
an input or output file.
An asterisk is printed to request the
file
(use OS/8 command decoder format).
CTRL/W is active only during the
output mode of the program.

3-18

system device.
Output or input can be done only to or from the
Answer the asterisk as follows.

*FILE.X

to read FILE.X or

*FILE.X<
to output to file FILE.X on the system device.

An input file must

have been written by the same program that is calling it.

Errors
An error during data transfer causes the message

USER ERROR

and control returns to OS/8.
3.2

Data Output Format on Mass Storage

When a file is written out by a histogram program, page 0, the zero
histogram and the histogram are written to the system device. The

format of the file is:
Block

Word

256

Block

Word

Block

is a

Location 1 of page
rest of page
up to location 2 56

histogram
4000, location -1 of start of
th histogram
4001, end location of
4002, location -1 of start of histogram
4003, end location of histogram
4004, start of th histogram
End of block = histogram data as in core.

Loc.
Loc.
Loc.
Loc.
Loc.

continuation of data as in core, blocks
continue to end of data

3-19

BOOK

AUTO-

AND

FOUR

CROSS-CORRELATIO

PREFACE

The Auto- and Cross-Correlation program for papertape, DEC-LB-U41B-PB,
is designed to correlate data (five bit samples (4 bits + sign)) at

sampling rates ranging from

to 2)14.7 milliseconds, on line, with

the user controlling all parameters from the Teletype.

It displays

and scales data while computing and provides output that can be used

with FOCAL, BASIC, or user's programs. An OS/8, version,
DEC-8E-AACRA-A-PB, runs under OS/8 and transfers its data to and
from the system device

A listing is available through the Software Distribution Center,
Digital Equipment Corporation, Maynard, Massachusetts 01754.

4-i

CONTENTS

BOOK FOUR

Pa ? e

CHAPTER 1
1.0

GENERAL DESCRIPTION

4-1

1 .

Theory

4-1

1.2

4-5

1.4

Properties and Examples of Correlation
Principle of Operation
Hardware Requirements

4-7

Program Loading

4-7

1.6

4-7

1.7

Using the Auto- and Cross-Correlation Package
Miscellaneous

CORRELATION MS

2.0

GENERAL DESCRIPTION
Use of Correlation MS for OS/8
Data Format

1.3

CHAPTER

AUTO- AND CROSS-CORRELATION PROGRAM

2.1
2.2

4-3

4-11

4-18
4-18
4-19

TABLES
4-1
4-2

Calculations for Cross-Correlation
Commands

4-6

4-12

FIGURES
4-1

Simplified Block Diagram of the LAB8/E
Correlation Program

4-ii

4-5

CHAPTER 1
AUTO- AND CROSS -CORRELATION PROGRAM

1.0

GENERAL DESCRIPTION

A variety of techniques are available for extracting signals from their
noisy environments. These include filtering (frequency domain), signal
averaging (amplitude domain)
and correlation (time domain)
Each
,

technique has its own advantages and disadvantages and hence its own
area of application.
Correlation, as it applies to waveforms, can be used to detect periodic
signals buried in noise or provide a measure of similarity between two
waveforms. Auto-correlation measures the similarity of a signal to a
time delayed version of itself, while cross-correlation measures the
degree of similarity of one source or input to a second source. No
synchronizing events, such as the trigger required in signal averaging,

need be available for the application of correlation techniques.
Typical applications for correlation include:
1.

Determination of the transmission paths and propogation
velocities of mechanical vibrations, acoustic and
seismic waves

Indication of epilepsy through comparison of electroencephalograms from the two halves of the brain.

Measurement of impulse-response characteristics of servo
control systems in the presence of noise.
Analysis of the tremor frequencies of patients afflicted
with Parkinson's disease.
Detection of very weak and distant radio sources in

space.

Speech research in telecommunications, education for the
deaf, and non-computer interfaces.

Theory

Mathematically, the auto-correlation function is defined as:

Rxx (T)=

S k /x(t)x(t-T)dt

*The reference for the statistical equations is:
Bendat, Julius S. and
Piersol, Allan G., Measurements and Analys is of Random Data, John Wiley
& Sons, Inc., N. Y. 1967.
4-1

where x(t), the time function of interest, is multiplied by a time delayed version of itself, x(t-x)

and the product of the two values

averaged over the observation time T.

The resulting average approaches

an exact correlation as T approaches infinity.

Cross-correlation is very similar to auto-correlation, the difference
being that the input is multiplied by a time delayed version of a
The
second signal, rather than a time delayed version of itself.

cross-correlation function of a waveform is defined as:

*(t)y(t-T)dt

T~ -^" /

(T)=

(2)

To apply digital computing techniques to the above equations, it is

convenient to approximate the average by sampling the signal at regular intervals, for instance t, and then summing a finite number, k,

Rewritten in discrete form, equation

of the sample products.

comes

(1)

be-

(T)

*kTT

x<t.)X(t.-T)

.*.

l=(f)

divided into N intervals of Ax
x
max is
samples
is
made
equal to the time delay
At
and the time between
If the maximum time delay,

increment Ax equation

(3)

(nAx)

becomes:

-j-fy
K+x

x(t. )x(t.-nAx)

which is the equation used in the LAB8
(2)

program.

where: n=0,l,2,3 ,

Similarly, equation

can be approximated by:

1
(nAx) ~ *£*

k+1

k
I

x(t.)y(t.-nAx)
i

4-2

(5)

1.2

Properties and Examples of Correlation

The auto-correlation function has the following properties:
1.

Auto-correlation is an even function; i.e., R (x) = R (-t)
xx
xx
therefore, the correlogram need only be computed and dis-

played for positive values of

is displayed at

the left of the screen.

The auto-correlation function has a maximum value for x= 0.
Sometimes it is convenient to set this value corresponding
to t= (J to 1 and express all other data as a fraction.
CTRL/T command of the program does this automatically.

The value of the auto-correlated function for

is the

mean square voltage of the time function, substituting
in equation (1)

T
xx (*)

= i
T-*-°°

The

2
x (t)dt = x (t)

_1_

The T command in the program automatically normalizes the
data and types it out as mr 2
.

If the time function contains periodic components, the auto-

correlation function will contain components having the same
periods.
Example
x(t)

R
i

2tt

XX (x)

-A
2

Example
x(t)

_2jT

4-3

,
'

If the function contains only random components

(noise)

and

no periodic components, the auto-correlation function will

The rate at which it approaches

approach zero as t increases.

zero is a measure of the randomness of the time function and,

therefore, the bandwidth of the noise.
R

(t)

L
6.

If the time function is composed of two or more components,

the resultant auto-correlation function will be the sum of
the auto-correlation functions of each individual component.

Example 4
x(t)

The cross-correlation function has the following properties:

The cross-correlation function is not necessarily an even
(-t).
(t)^R
Consequentxy
xy
ly, a time shift of one input function in one direction with
respect to the other does not produce the same result as an

function; therefore, in general, R

equal shift of the same function in the opposite direction.

function x(t) in one
(t)=R
(-t)
A shift in one input
c
xy
xy
direction must yield the same result as an equal shift of
R

the other input function y(t)

does not necessarily have a maximum at

(t)

maximum value of R
shift

in the opposite direction.

(t)

x=J?

The

will occur for that value of time

for which the two input functions are most alike.

4-4

The following example illustrates the application of the

cross-correlation function:
Example
x(t)

1.3

y(t)

VJ>

Principle of Operation

A simplified block diagram of the LAB8/E correlation program is shown
in Figure 4-1.

PROCESS
x(t)

h3PLE SINGLE PRECISION
STORE

n-1
n-2

B b
y(t)

DOUBLE
PRECISION
STORE

CROSS

Figure 4-1

Simplified Block Diagram of the LAB8/E
Correlation Program

When the program is started, the user is requested to reply to three
questions: rate, number of points, and auto? Once these questions
have been answered, the program does the following automatically:

4-5

Sets the clock to take analog samples at the
specified
rate.
Sampling rate = Ax.

Sets up a single precision buffer to hold the delayed samples.

Sets up a double precision buffer to hold the correlogram data.

Sets the software auto/cross switch to the proper position.

The data collecting part of the program can now begin.
The memory updating process can be visualized as continually
storing
data in a closed, loop. After each revolution of the loop, the oldest
data point in the shift register is replaced by a new sample.
As each sample is converted into an equivalent digital value
by the
analog-to-digital converter, it is stored in the single precision
store.

This buffer acts in a manner similar to a shift register.
That is, for each new sample entered, the oldest sample is lost.
As the data is being shifted in the single precision store, calculations are being performed to update the correlogram.
Each data point
is multiplied by the most recent sample (x=0) and the product
is added
to the corresponding cell in the double precision store.
In crosscorrelation, the multiplier is a sample from another channel.

Table 4-1 shows the calculations formed for cross-correlation.

Table 4-1

Calculations for Cross-Correlation

Samples

Shift Register Number

x(t )y(t
1

x(t )y(t
2

•

V V
)

)

xCt^yttp-A-r)

x(t )y(t -2AT) ...x(t )y(t^-nAT)
J?

x(t )y(t
i
±

y (t^Ax)

x(t )y(t -Ax)

x(t

x(t )y(t -Ax)

x (t )y (t ~2Ax)
2
2

•

)

•

x(t )y(t -Ax)

Ix(t.)y(t.)
x
i=0 x

Zx(t.)y(t.-Ax)
x
i=0 x

•x(t )y (t^nAx)
1

.x(t

)

•

)

x(t.

)

y (t 2 ~nAx)
a

y (t. -2Ax) .

.x(t.

)

y (t. -nAx)

£x(t.

i=0

y (t. -2Ax) .
!

Ex(t. y (t. -nAx)
i
*
i=
.

)

J
4-6

1.4

Hardware Requirements

PDP-8/E
AD8-ES

10 bit A/D converter

DK8-EP

Real-time clock

VC8-E

10 bit point plot display
~\

VR14 Tekni
02 ' pisplay oscilloscope
603, ?„^
604 or
RM503
J
AM8-EA
8-channel multiplexer must be implemented to
perform cross-correlation and to obtain a
printed or plotted output.

1 .

Program Loading

The binary tape of the Auto- and Cross-Correlation program,
DEC-LB-U41B-PB, is loaded using the binary loader. First, place
the tape in the high- or low-speed reader and advance to the leader/
trailer portion. Load 7777 in the Switch Register (SR) and depress

ADDR LOAD.

If loading with the high-speed reader, set bit

to 0.

Depress CLEAR then CONT keys.

After the tape has been loaded, load 0200 in the SR, depress ADDR LOAD,
push CLEAR, then CONT keys.
The binary tape, DEC-LB-U41B-PB, can be loaded with the disk monitor
system by following the disk monitor's instructions for loading tapes,
i.e., SAVE AUTO! 0-3177 ,6200-7577 ;200
1.6

Using the Auto- and Cross-Correlation Package

System Initialization
Load the binary tape as described in Section 1-5.
Once the program has been started, it prints:
RATE ?
Type the sampling rate

(t)

in milliseconds.

be in one of three formats

4-7

The number typed must

XXX.

xx.

x.x

"x" signifies a decimal digit.

be in the range of
the message

to 204.7.

The value of the number typed must
If the number exceeds this range,

000.0<RAT<204.8
is printed.

In practice, the rate specified should be at
least
four times the maximum frequency being sampled.
If something other than a digit is typed, it
is ignored and not

echoed.

The sampling rate is the time between the taking
of samples from
the analog to digital converter.
Once the rate has been entered,
the message

NO. OF POINTS?

is printed.

This parameter defines the number of points that
will
make up the correlogram.
Input is a decimal number of no more than
three digits.
Its range is 4<number<513.
If the number entered is
outside this range, the message

4<# OF POINTS<513
is printed.

In all cases, when typing a number to define
the rate and number of

points, input can be restarted by typing a RUBOUT, which
echoes a
backslash
input can be prematurely terminated when typing
\
the
number of points by typing the RETURN key.
(

)

Once the number of points and the sampling rate
have been defined,
the computer prints:

AUTO?
If the signal is to be auto-correlated, respond
with a Y; if it is to
be cross-correlated, type an N. Any other characters
are ignored and
not echoed.
4-8

After the reply is entered, the program prints either
BATCH MODE

NORMAL MODE

A small digression is necessary to explain the difference
between
BATCH MODE and NORMAL MODE. It was found that
to process one point
took at least 200 usees. To process N points
took at least (200N)
Msecs.
Thus the user, when defining the rate and
number of points
had to make sure he was allocating at least
usees
200
per point. For
,

example, suppose the rate is defined as 20.0
millisecs., and the
number of points as 100.
20.0 millisecs. = 20,000 microsecs.,
20,000
- 2 00 usees per point.
But what if the rate were 20.0
100
msecs., but with 200 points; this would mean
100 ysecs. per point;
which is not enough time to process one point.
When there are 200 or
more microseconds per point, the correlator
operates in NORMAL MODE,
that is, it does one entire multiplication
sweep, reads in a new
value from the ADC, does another sweep, etc.
This provides a steady
sampling rate. However, if the time allocated
to the processing of
one point is less than 200 ysecs., then
BATCH MODE is entered. In
this mode, the entire buffer is filled with
samples taken at the defined rate. Once N+l samples have been
taken, the buffer is multiplied
by the last sample and displayed, after
which new data is taken. The
following graphs illustrate the two modes:

NORMAL MODE

The arrows represent the places on the graph
where samples are taken.
During this time, the

At is the time between samples, or
the rate.
buffer is processed.

4-9

BATCH MODE

Because of its nature, BATCH MODE takes longer to yield a clean
correlogram.
Therefore, the user should operate in NORMAL MODE whenever possible.

After the computer notifies the user of the mode, it displays the empty
buffer and waits for CTRL/S to start data collection.
NOTE

During the time of parameter setup, typing
CTRL/A restarts the parameter setup routine,
or CTRL/C jumps to 7600 and returns to the
Monitor, if it exists.

Input Connections
If doing auto-correlation,

If cross-correlation,

and 1.

the lead must be connected to channel 0.

the two leads must be connected to channels

shift, register is filled with samples from channel 0.

The

either case, the input should be externally adjusted to the range of
+1.0 volt.

For best results, the input should be as close as possible

to +1.0 volt.

Commands
There are two major groups of commands:
Group

Thesp commands are active during data collection and
output.

Group

These commands are active during data output and are
inactive during data collection. They are ignored if
typed while data is being correlated.

4-10

The commands are described in Table 4-2.
1.7

Miscellaneous

Program response to keyboard commands and the integrity of the
display will differ between collection modes during data collection
(refer to System Initialization for description of modes).

Because of the nature of data collection, BATCH MODE allots less
time to keyboard monitoring and display.
This has no effect on
the quality of the data, and a fully active display and responsive
keyboard will return when data collection has ended.

4-11

Table 4-2

Commands

Command

Action

CTRL/S

Starts data taking.
It clears the buffer and all
counters before starting.

CTRL/C

Stops data collection, turns off the clock and
jumps to 7600 in field 0.
If running with the disk
monitor system, this command returns control to the
monitor.

CTRL/A

Immediately transfers control of the parameter setup
routines and clears the buffers.

CTRL/Z

Clears the buffer and the sweep counter. The program then waits for CTRL/S to start data collection.
CTRL/Z must be preceded by a CTRL/Q if data taking
is in progress.
In batch mode, CTRL/Z cannot be
typed until the display appears on the scope

CTRL/R

Restarts data taking, but does not clear the buffer
or the sweep counter.
It is generally used for restarting data taking after a CTRL/Q command.

CTRL/Q

Stops data collection after the current sweep and
enables all Group 2 commands.

CTRL/U

This is an unimplemented user command.
Details of
its implementation may be found in Appendix O.

Expands the display of the correlogram by a factor
of 2.
The scale factor is initially set to 2 4
.

Contracts the display of the correlogram by a factor of 2.
Display of input is not affected.

Determines whether the input or the correlogram is
displayed.
The input is initially displayed.
Typing V will switch the display from input to correlogram, or vice versa.
When switching modes, the
display may deteriorate due to interrupts in the
middle of the display loop.
Effects only the display of the correlogram.
The
base line is shifted by the value of knob 2.
The
base line will be biased by the same amount each
time X is typed.
The correlogram can thus be moved
up and down on the scope.
Restores the correlogram to its unbiased X axis.
It is used to counteract all previous X commands.

Group

Alternately turns on and off the display of cursors.
The two cursors are initially off.
Their locations
on the X-axis are determined by knobs
and 1.

4-12

Table 4-2

Commands

(

Command

Cont

Action
NOTE
The inputs to analog inputs
and
1 must be disconnected before the
cursors are used.

Prints the number of sweeps and the current scale factor in the following format:

xxxxxxx
2 fxxxxxxx

(number of sweeps)
(scale factor)

where x is a decimal digit. The number of sweeps is
defined as the number of times the first position of
the shift register is multiplied by the results of an
A/D sample.
Calibrates an X-Y analog plotter. When the K command
is given, the outputs of the oscilloscope go to the
maximum output voltage for the X and Y scope coordinates.
When any non-command key on the keyboard is
typed, the outputs each go to the minimum voltage.
When any other key is typed, the outputs go back to
the maximum.
This process may be repeated as many
times as needed while the bias and gain controls for
the X-Y plotter are adjusted.
(Refer to Appendix H
for details of attaching an analog recorder.)
When
the plotter is calibrated, type CTRL/Q to resume the
display.
Plots the displayed waveform on an analog plotter.
After P is typed, the first point is displayed, and
the plot routine waits for a character to be typed
before proceeding with the plot. The plotter should
now be turned on. Once a character has been typed,
the data is plotted.
The speed of the plot is determined by knob 3 on the laboratory peripheral. Turn
it counterclockwise to go faster, clockwise to go
slower.
This should be adjusted until the plotting
speed is appropriate for the plotter being used. At
the end of the plot, the routine stops, waiting for a
character to be typed before proceeding. This allows
the user time to turn off the plotter.
After the
character is typed, plot mode is terminated, and the
display returns to normal.

CTRL/I

is an unimplemented user command.

For details of

implementation, see Appendix 0.

CTRL/B

causes a binary dump of the correlogram. (only used
in paper tape system, DEC-LB-U41B-PB

NOTE
The correlogram is returned to
normal size and scale before the
binary dump takes place.

4-13

Table 4-2

Commands

(Cont.

Command

Action
After the initial command of CTRL/B, the computer
prints
OUT?

Answer H if the dump is to proceed on a high-speed
punch or L if the Teletype punch is to be used.
(Any
non-command character is ignored and not echoed.)
After this initial question, the computer prints:
NAME:

Enter a maximum of six characters followed by the
RETURN key for an identification code.
CAUTION
The following characters can not be
used in the identification code:

RUBOUT

CTRL/C

CTRL/

Typing RUBOUT causes input to be
zeroed and then restarted.

After the identification code has been entered, leader/
trailer code is punched.
If the low-speed punch is to
be used, it should be turned on during the leader/
trailer output. Following the leader/trailer code, the
identification code is punched in ASCII and followed by
more leader/trailer code.

After the identifying information has been output, a
core image of the correlogram and a checksum are punched.
A core image dump allows the paper tape to be read
back into the computer in the same form as it was
originally dumped. The binary loader can be used to
reload the data.
This feature allows the user to take
data, dump it, and then return at any later time to
process or review it. The user may write his own processing program or he can read the data back into the
auto- and cross-correlation program.
For details of
binary format, see Appendix P.
To review the data with
the auto- and cross-correlation program, load the program as described in Section 1.5 (if the program is in
core, give the command CTRL/Z; reloading is not necessary)

After the program is in core, load the data with the
binary loader. After the data and programs are in
core, start the program at location 7560 (place 7560
in the front switches , press ADDR LOAD CLEAR and CONT
keys)
Now operate the program as if the output mode
had just been entered.
,

4-14

Table 4-2

Commands (Cont.)
Commands

Action
NOTE
Since only the correlogram is
dumped and not the temporary
buffer, reloading of the data
may cause a bad display of the
temporary buffer. This will not
effect the proper operation of
the program.
If this program is run from disk
or DECtape, the correlogram can
be dumped by the following pro-

cedure

Type CTRL/C.

Save page zero and locations
3200-6177 as binary files.
Reload the program and start
at 200.

Monitor returns.

To review this data, load the autoand cross-correlation program, then
the data.
Start at location 7560.

After the Binary dump has taken place, the program returns to the active display and all of the commands are
operative.
Prints the data between the cursors.
If the cursors
overlap one another, the typeout routine is immediately
exited.
An example of the output follows:

0000001,

0000004,

0000804,

0000004,

0000200

0175872
0057856
0035072
-0015872

0000001
0000004
0000804

0000004

0000200
0175872

signifies the leftmost cursor is on the first
point of the correlogram
the position of the rightmost cursor (i.e.,
the fourth point of the correlogram)
is the sweep count which indicates the number
of times the first position of the shift register was multiplied by an ADC sample
is the number of points whose values are being
printed
200 x 100 ysec is the amount of time represented
by the distance between the cursors
is the normalized value for the first position
of the correlogram

4-15

Table 4-2

Commands

(Cont.)

Command

Action
0057856

is the normalized value for the second posi-

tion of the correlogram.

Before the data is typed, it is normalized by one of
the following standard equations:
For "normal mode" normalization:

Normalized value of the
(

point =

raw value of the i
point
(number of sweeps) - (i)

)

For "batch mode" normalization:

Normalized value of the
(

point =

raw value of the i th point
(number of sweeps)

The units for the data points are

)

(millivolts).

NOTE
If the LAB8/E is equipped with a single
analog input, a typeout can only be obtained by placing a +5 volt signal on
the input and typing T.
The contents
of the entire buffer are dumped.

The typeout continues until all the data between the
cursors has been output.
Typeout can be stopped prematurely by using the CTRL/Q command.

CTRL/T

is similar to the T command in that it also types data.
The first line of output is identical to the first line
printed by the T command. However, the data points are
printed as correlation coefficients, ranging from +1
to -1, rather than as normalized values.

NOTE
This command should only be used while
doing auto-correlation. The results
are meaningless for cross-correlation.
An example of the output follows:

0000001

000000 4,

0000804

+0.99999E+00
+0. 32827E+00
+0.19863E-01
-J3.92564E-01

4-16

000000 4,

0000200

Table 4-2

Commands (Cont.)

Commands

Action
The correlation coefficients are output in floating
point form. The five digit decimal number to the right
of the decimal point should be multiplied by 10 raised
to the power of the number following the E (E is for
exponent)
The equation by which the correlation coefficient is reached is the following:
.

Correlation coefficient of the
(

point =

raw value of the i,th point
(raw value of the first point)
)

In the above example printout, the correloation coefficients can be interpreted as follows:

1st point / 1st point = +0.99999E+00
2nd point / 1st point = +0. 32827E+00
3rd point / 1st point = +0.19863E+00

NOTE
The correlation coefficients obtained
from this routine when cross-correlation
is performed bear no resemblance to
the correct correlation coefficients.
The routine is valid only for obtaining
correlation coefficients from an autocorrelogram.

4-17

CHAPTER 2
CORRELATION MS

GENERAL DESCRIPTION

2.0

The correlation program that runs under OS/ 8 writes the correlogram
to the system device and has the ability to read it back at a later

time

The binary of the program is

DEC-8E-AACRA-PB
2.1

USE OF CORRELATION MS FOR OS/8

Auto and cross-correlation MS is the same as the paper tape system
with the addition of writing data to and reading from the system
device.
The data written by the correlation program can be read by
it at a later time.

Loading
The program's binary can be loaded via the Absolute Loader and started
at location 200.
To save the core image after the binary is loaded, use
the SAVE command

SAVE SYS CORD3.SV 0-3177 ,6200-7577 , 17400-17577 ;200
MS Command

CTRL/W is the command to read or write data. The OS/8 command
decoder is loaded and requests a file.
In response to the * an
output file can be given to write the correlogram,
*FILE.XX<
or an input file can be given to read a previously
written file.
* FILE. XX

NOTE
If a file other than one written by the correlation program is
given on input, the program will
crash.

4-18

After the file is given, the program reads or writes the file and
returns to display mode with the correlogram on the display.
MS I/O Errors

An error to the command decoder causes an OS/8 error message. An
error reading or writing the data, closing the file or filling the system device causes an error message to be printed. After an error
OS/8 is running, restart the correlation program at 2556, field 0.

Running with no Data Collection
If the program is to be run by reading a file and not collecting data,

the following procedure can be followed.
Start at location 200 and give
dummy answers to the questions asked. Type CTRL/S followed by CTRL/Q.
The program is now in output mode.
Type CTRL/W to read in the file.
2 . 2

Data Format

The data written out is page 0, the temporary buffer at 3200 followed
by the correlogram.

Block

Page 0,
1

Word

1=3, words 2 to 256 are locations

on.

Block

temporary buffer starting at 3200 in single precision followed by the correlogram in double precision.

Block

is written as needed to get all current data.

etc.

The temporary buffer contains the A/D readings and the correlogram
contains the summed product as described in Table 4-1.
,

The temporary buffer is N+l words long and the correlogram is 2N
words long where N is the answer to the setup question "No. of Points".

The data in the correlogram is a double precision number, referred
to as the raw data under the T command.

When a T command is given, the

printed

(i)

value printed is

Raw Data(i) x 4096
# of Sweeps - i

The data in core which is written to the system device is "RAW DATA (i)".

4-19

BOOK

FIVE

D A Q U A N

CONTENTS
BOOK

Chapter

FIVE

Pa ? e

USING DAQUAN

1.0

GENERAL DESCRIPTION

5-1

1.1

Hardware Requirements

5-2

1.2

Initial Starting Procedure

5-2

1.3

Inputting Data

5-3

1.4

Stopping and Restarting DAQUAN

5-4

DAQUAN COMMANDS

2.0

INTRODUCTION

5-5

2.1

AVERAGE

5-6

2.2

TIME

5-8

2.3

PAPER TAPE

5-8

2.4

CALCULATE

5-9

2.5
2.6

ALTER
COPY

5-12

2.7

XINVERT

5-12

2.8

YINVERT

5-12

2.9

SCALE

5-12

2.10

MULTIPLY

5-13

2.11

SMOOTH

5-13

2.12

CURSORS

5-13

2. 13

LIST

5-14

2. 14

INTEGRATE

5-14

2. 15

STRIP

5-15

2. 16

DIFFERENTIATE

5-16

2. 17

PLOT

5-17

2. 18

SUBTRACT

5-18

2.19

ADD

5-18

2.20

SWAP

5-18

2.21

SQUEEZE

5-19

2.22

MODIFY

5-19

2.23

OUTPUT

5-19

2.24

RESTART

5-20

2.25

PEAK Report

5-20

DAQUAN EXAMPLE

5-23

5-11

5-i

CONTENTS
Page

Chapter

DAQUAN MS FOR OS/8

4.0

GENERAL DESCRIPTION

5-26

4.1

Loading DAQUAN MS

5-26

4.2

Using DAQUAN MS

5-27

4.3

Data Format

5-28

4.4

Errors

5-28

DAFFT OVERLAY FOR DAQUAN MS

5 .

5-30

5.2

Operation

5-30

5.3

Command Description

5-31

5.4

DAFFT Example

5-34

PAFFT OVERLAY FOR DAQUAN MS

6.1

5-35

6.2

Operation

5-35

6.3

Command Description

5-36

6.4

PAFFT Example

5-37

5-ii

CHAPTER 1
USING DAQUAN

1.0

GENERAL DESCRIPTION

DAQUAN is used to acquire data by boxcar, multi-sweep time averaging
from one instrument at a time and to display the results. A wide

variety of subsequent processing techniques such as smoothing, differentiation, curve simulation, etc., are then used to reduce the data.
Input

Data rate

one channel

seconds/point to 8.2KHz
(using Schmitt triggers)
Maximum Number of Points = 2047
30

DAQUAN is intended for the following applications:
spectra comparison, stripping, and simulation
Gaussian and/or Lorentzian fitting
deconvolution of fused peaks (by Gaussian,

Lorentzian or mixed technique)
integration

differentiation

multiplication
scaling
plotting (standard X-Y recorder)
peak detection and summary reporting of
positions and percent areas.

Using DAQUAN' s data handling commands, the researcher can interpret
the data as required by his experiment in seconds.
A sloping baseline
can be aligned, a spectrum can be scaled, integrations can
be performed, and two spectra can be compared simultaneously, each
operation
requiring only a single command.
An X-Y recorder can be interfaced parallel to a scope channel of
the
computer for accurate plotting of the results of any operation.
The applications of DAQUAN can include limited gas chromatography,
slow mass spec, autoanalyzers atomic absorption, infra-red, and
U.V.
,

spectrometers
5-1

The output from DAQUAN is fully compatible
with BASIC/RT so that the
latter can be used for more sophisticated processing of the
data acquired and/or processed by DAQUAN. It is possible to
successively read
2047 point blocks from paper tape, process the data with output
to the
high-speed punch and, when complete, read the output tape
into the
BASIC/RT program and perform a wide variety of subsequent
analyses.
1.1

Hardware Requirements

The minimum configuration for using DAQUAN is:

Advanced LAB8/E - 15 system
An X-Y Analog Recorder is recommended for hard
copy,
using a relay driver controlled by the channel
select
bit of the display register (see Appendix
Q)
1.2

Initial Starting Procedure

Both the RIM and Binary Loaders must be loaded before
DAQUAN can be
loaded.
Refer to Appendix A for Loading Procedures for RIM
and

BIN.

To load DAQUAN use the following procedure:
1.

Place the DAQUAN tape (DEC-8E-ADQNA-A-PB) in the proper
reader (with leader/trailer under read head)
either
Teletype or photoelectric. NOTE:
The binaries of
DAQNFP and DAQUAN. 8 E are combined on the paper tape
DEC-8E-ADQNA-A-PB.
,

Set the Switch Register (SR) to
and press EXTD ADDR LOAD.
Set the SR to 7777 (the starting address of the BIN Loader)
Press the ADDR LOAD switch and turn on the reader.

If using the high-speed reader, set SR bit
CLEAR then CONT keys.

When the reader stops, DAQUAN is in core memory.

If the accumulator

to 0.

Press

does not contain zero when the reader halts or
if the tape does not stop on trailer code, a checksum
discrepancy exists
(AC)

indicating that the DAQUAN tape has been read incorrectly
tape again by repeating the procedure described above.

Load the

The OS/8 save locations for DAQUAN (paper tape)
are 0-7577.
To start the DAQUAN program initially, set the SR to
0200 and press
ADDR LOAD. Clear the Switch Register, press EXTD ADDR LOAD
switch,
next press CLEAR, then CONT keys.

DAQUAN indicates it has been successfully loaded into the computer
by
printing the message

5-2

DAQUAN LIVES!
TITLE:

Commentary of any length can be typed after TITLE. The information
is not stored with the data and is used only to supply a titled paper
copy record of data operations performed.
Type CTRL/A to exit from
the title phase.
A command may now be issued (refer to Chapter 2)

Refer to Appendix T for the core map of DAQUAN.
1.3

Inputting Data

Most calculations are performed with a Floating Point Package (refer to
Chapter 8 of Introduction to Programming for a description of Floating
but final array results are single precision (12 bit) integers,
normally scaled to the range 0-100£L
When a numerical value is requested,
Point),

most conventional formats are acceptable. Thus, for example, the decimal
value 10 may be entered as 10. J3, 10, 1E1, or 1E+02, etc., as described in
.

the above manual. After a numeric value has been typed in response to a
command, any character except
to 9, E, or
terminates input for
.

that entry.

A space is recommended as the terminator. Typing the
RETURN key does not automatically generate a line feed.)
In response
to questions, only Y or N is an acceptable answer.
Any other response
generates a question mark on the Teletype and is ignored. No terminator is required after a Y or N response.
In all cases, typing RUBOUT before a terminator deletes all input up
to the preceding terminator and allows the correct value to be entered.
A RUBOUT during numeric input echoes as an exclamation mark on the
Teletype.

If, during scope display, a command unacceptable to DAQUAN is
typed,
a ? is printed on the Teletype and the program returns to
the same

scope display.

Because most DAQUAN commands replace the previous contents on the
scope, it is recommended that any raw or interpreted data that may

possibly be used at a later time be saved on tape (see PAPER TAPE command, Section 2.3) to prevent accidental elimination from the display
channel.

5-3

1.4

Stopping and Restarting DAQUAN

If an operation must be halted immediately, press the console HALT
This should not be used haphazardly; if arrays were being

switch.

modified, the data may be lost.
Routines requiring input parameters
or initial dialogue can be halted during that stage and before the
input is complete.
To restart DAQUAN after an emergency stop, set
HALT switch to run position (switch in up position)
set Switch
Register to 0201. Push ADDR LOAD switch, clear Switch Register and
push EXTD ADDR LOAD switch. Next, push the CLEAR switch then the
,

CONT switch.

A new DAQUAN command can be issued when the display

is restarted.

To Reassemble:

The sources for the paper tape DAQUAN system are:
DAQUAN. 8

DAQNFP
on the LAB8/E paper tape source DECtape,

DEC-8E-ASPTA-A-UA4.

DAQNFP is the floating point package used by DAQUAN.
}

To reassemble,

each file should be assembled separately (with PAL8) to create the
binary paper tapes. The floating point binary must be loaded (via
the binary loader) before the DAQUAN binary.

5-4

CHAPTER 2
DAQUAN COMMANDS

INTRODUCTION

After starting DAQUAN (refer to Section 1.2), a scope buffer should
be designated to accommodate the incoming data which is loaded into
the computer's memory area directly from an analog input channel, the
Teletype keyboard, and the high-speed or Teletype reader. Any of
DAQUAN 's commands can then be used to interpret the data. For example,
the baseline can be linearized and the resulting spectrum integrated.

Each DAQUAN command is described in this chapter.
The values requested,
acceptable ranges, formulas used and available options for each are
included in the discussion. After each command is completed, the correct spectrum is displayed (except the MODIFY command)
Another command may be issued at that time.
.

Scope Buffer

DAQUAN provides three buffers for viewing data:

buffer 1 - assianed bv user
buffer 2 - assigned by user
buffer 3 - buffers 1 and 2 simultaneously
Data supplied from the associated instrument or from paper tape is assigned to buffer 1 by typing the number 1 on the Teletype; similarly,
data is assigned to buffer 2 by typing the number 2 on the Teletype.
The choice of buffer 1 or 2 determines which data will be used.
If
a number between 4 and

is typed,

buffer

is assumed.

Typing

produces a response of question mark on the Teletype.
Each buffer can accommodate up to 2047 data points. By typing 1 or 2
a buffer can be assigned to accommodate data when a new spectrum is
being collected or when another DAQUAN command has been completed.
The present contents of that buffer are displayed on the scope. When

new data is assigned to a buffer which already contains data, it
replaces the old data.

5-5

Buffer

permits the user to view the data in buffers 1 and 2 together. When buffer 3 is being used, analog knobs
and 1 control the
X and Y offset for buffer 2 to allow comparison of data sets.
It can
3

be viewed at any time between commands, but should be followed
by a 1
or 2 before calling a command.
If left at buffer 3, the next command
assumes buffer 2.

Most of the remaining DAQUAN commands are called by typing at least
the first two letters of the command name and a colon after the
termination of the preceding command. Refer to Appendix R for a summary
of the commands.
Commands such as AV, PA, CA and CO require specification of a channel, i.e., 1AV or 2CA.
2. 1

AVERAGE

The AVERAGE command (AV:
implements time averaged analog input from
the associated instrument.
The parameters of the scan are specified
)

on the Teletype in response to AVERAGE'S questions and then
that
operation is performed. The data collected is displayed from
the
buffers specified before the AVERAGE command was initiated.
Because of
the need to store the intermediate sum in double precision
format, if
more than 1024 points are requested, storage will overlap the
alternate

display buffer and destroy any data kept there.
To collect data, proceed as follows:
1.

Prepare the analog instrument for the experiment
and connect the analog signal line to any available channel and the sync or trigger to Schmitt
input 4.

Specify a scope buffer number
modate the data.

Type AV: indicating an averaging operation is to be
performed.

The computer prints:

to accom-

CHANNEL =

requesting the analog channel number (decimal) to
which the instrument is connected.
Type in the
analog channel number to be used by this experiment and press the RETURN key.
5.

The computer responds with the message
NO.

POINTS =

5-6

Type in the number of points to be collected in the
scan where 0<POINTS<2048 and then press the SPACE bar.
6.

The next question is

SEC/SCAN =
Type in a value for the length in seconds of each scan
where SEC/SCAN= seconds/point x points/scan. Remember
that a minimum of 0.123 millisecond/point and max of
30
sec/pt. is required.
Press the SPACE bar.
If the resulting rate is too high [(SEC/SCAN/NO. POINTS)
<0. 123
msec], the routine restarts, asking for
NO. POINTS =
7.

When the rate is acceptable, the next question is:
DELAY (SEC) =
asking for the delay from receipt of the trigger
pulse
to the start of data acquisition for each scan.
Values
from
to 20 seconds are legal; resolution is 10 milliseconds.
Press the SPACE bar after typing the value.

The message
NO.

SCANS =

is printed last and the reply must be in the
range 1
to 2047.
Type in the value and press the SPACE bar.
Input will begin with the arrival of the first triqqer
pulse.
9.

Activate the instrument to begin the experiment.
The
AVERAGE command will now collect the data according
to
the specified parameters.

The AVERAGER implements two display routines of
its own, totally separate from the main scope routine of DAQUAN.
This was necessary because
of internal timing and intermediate data structure
peculiar to the
AVERAGER.

During data acquisition, the value for each point
in the current scan
is displayed as soon as it is available.
Note that the X axis is not
scaled to the number of points; thus, if more than
1024 points were requested, the trace will wrap-around such that the
1025th point will be
displayed at the first location, the 1026th at
the second location, etc.

Between the end of each sweep and the trigger for
the next, the current
summed data is displayed. By typing C for check,
subsequent triggers
can be ignored.
The Y amplitude can be magnified at this time
by typing M; or decreased by typing D.
Type P to proceed or S to stop the
run.
When S is typed the run is terminated, the number
of scans completed to that time is indicated by the message
NO. SCANS =

5-7

the summed data is divided by the number of scans taken, and the average is displayed on the scope.

TIME

The constant corresponding to the computer's memory cycle time used to
determine timing accuracy for averaging during data acquisition can be
reset, by using the TI command.
The AVERAGE routine can be calibrated
as follows:
1.

Disconnect all inputs to Schmitt triggers. A dummy
run is made to measure the actual time required for
a two minute sweep vs. the time based on an assumed
memory cycle time of 1.2 microseconds. This is done
by manually triggering the scan by flipping the
Schmitt trigger polarity switch and measuring via a
stopwatch the elapsed time until the static scope
display appears after the sweep.

Call the AVERAGER by typing AV: then use the values
CHANNEL 0, 1000 POINTS, 120 SEC/SCAN,
for DELAY,
and 2 for NO. SCANS.

Time the program from when the Schmitt trigger slope
switch is struck to when the static display begins,
approximately 120 seconds later. Call this value T.
The memory cycle time, MCT, is then determined by

MCT = T x 1.2 x 10~ 6 = T x 10~ 8 sec.

12j
4.

Enter the desired value by typing TI: and supplying
the value of MCT after the message
MCT (SEC)

2.3

PAPER TAPE

Data can be input to the computer from papertape via the high- or lowThe data may have been punched directly by the analog
speed reader.
1

instrument or may have been prepared by the DAQUAN command OUTPUT.
The procedure for inputting data from the tape is as follows:
1.

Type PA: to call the routine.

Answer the message
NO. POINTS=

with a value less than 2048, not including skipped
points.
3.

The next message is
PTS TO SKIP=

DAQUAN accepts data in USA-ASCII format with and without spaces and/
or non-numeric characters before a numeric value (each number must be
properly terminated, as described in Section 1.3).
5-8

Type in the number of points to be skipped from
the start of the tape before data is retained
(less than 2048) and press the SPACE bar.
4

Two parameters are requested by

YRANGE

MIN:

The range on the Y axis should be at least as
large as the difference between the largest and
smallest values on the tape. Type the Y- range,
a comma, and the minimum value, which should be
no larger than the smallest Y value in the data.
Press the SPACE bar.
5.

The last question is

TTY I/O?

Prepare the reader before replying because the
tape will be read in immediately after the reply.
Answer Y if the data is coming from the low-speed
reader.
A reply of N will start the high-speed
reader.
2. 4

CALCULATE

Spectra can also be calculated from the Teletype keyboard. DAQUAN requests information on the peaks, parameters, number of points, and
baseline.

(Refer to Appendix S for definition of the equations used.)

The spectrum resulting from this calculation is displayed.
Interaction from a peak that is more than eight times the half-width

away from its position (or crest) is ignored.
This gives a considerable saving in time for narrow width, multi-peak spectra.

Note that all peak widths must be less than 1/4 x (X RNG) (step 4).
For ease of input, the peak width for only the first peak is entered (step 5) .
See Appendix S if it is desired to change these
conditions.

The CALCULATE procedure is as follows:

Call the routine by typing CA:

Respond to the message
NO. POINTS=

with a number less than 2048 and press the SPACE
bar.
Note that both buffers 1 and 2 are affected
by this value; thus, if the CALCULATE operation is
using buffer 1, only that number of points will
be available for display of the data in buffer 2.
3.

After the message
NO. PEAKS=

type the desired number of peaks, in the range
0<PEAKS<30.

5-9

More data is requested by
X RNG, XI,

INT MPLR, BASE, LOR FR.

(0-1):

The first three parameters allow the
user to define
any coordinate grid in which to contain
the calculated
SSe P arameter s scale the data to a
1000 x
r2S2* point
1000
grid for compact integer storage.
This
does, however, impose a minimum on
resolution,
S
f ° r Peak P° sition and w^th, defined
as
v DMr/?iiX
X
RNG/1000.
Type in the parameters:
X range, initial
X location, multiplier which is the Y
axis scaler
the location of the base line on the scaled
Y axis,
and the Lorentzian fraction where
is a 100% Gaussian
tit and 1 is a complete Lorentzian fit.
The
must be in the range 0.0 to 1.0. A value of fraction
1 for
INT MPLR implies a range of
to 1000 on the Y axis,
a value of 10 implies a range of
to 100, etc.
Base
and peak heights should then fall within
this scaled
range.
Peak positions should be in the range XI to
Xl+X RNG. A negative X RNG is legal; e.g.,
for
Xl=500 and X RNG= -500, the X axis runs from
500 on
the left to
on the right.
5.

The dimensions of each peak, requested
next by the
message,

HGT,H-WIDTH,POS:
are height, half width at half height, and
position
on the scaled X axis.
Type in each value followed
by a comma or space and then press RETURN after
the
parameters for each peak.
Thus, a two peak spectrum
^
can include the data:

678,5,35
789,50

Enter a width for the first peak only. That
value
is assumed for all other peaks.
A carriage return line feed is issued when the parameters for
each are
complete.
6.

After all the necessary data has been typed
in, the
computer indicates it is busy performing the
calculation by the message
COOL IT.
on the Teletype.
scribed by steps

When calculated, the spectrum dethrough 5 is displayed on the scope.

After a display buffer has been selected and data
has been brought
into the computer's memory, any of the following
commands can be
issued.
Any number of them can be requested and in
any order.
After
the data has been interpreted by any of
the available commands, it
can be stored on tape for later use.

5-10

NOTE

Because of parameter storage requirements use
of the CALCULATE routine will remove MODIFY
(Section 2.22)
from DAQUAN.
,

2. 5

ALTER

The parameters specified by the most recent CALCULATE command can be
modified by an ALTER command (AL:). The command is implemented as

follows
1.

Type AL: to call the ALTER routine.

The message

PK,PA,VL:
is printed on the Teletype, requesting the peak index number.
PK = peak number where the first peak
typed in during CALCULATE is 1, the second is 2,
etc.
PA = parameter index where height is 1, halfwidth is 2 and position is 3. VL = new value for
that parameter.
Type the peak number, a comma or
space, the parameter index, a comma or space, and
the new value.
Press the SPACE bar. All parameters
except X RNG and XI may be altered. The equal widths
option has no effect at this time; any width may be
altered.
Consider the following sequence:

PK,PA,VL: 12,3,674
This series will position the twelfth peak at 674
on the scaled X axis.

Other spectral parameters can be modified by typing one of the sequences
listed below and then the new value in response to the above message.
Sequence

Parameter to be Modified

0rl''

intensity multiplier

0,2:

baseline

0/3:
n,0:

Lorentzian fraction
print parameters for peak n

0i0-

print all parameters

The last two sequences exit to the display without recalculating the
spectrum.

Type 0,-1 and a SPACE to signal the end of
alterations.
The computer indicates it is recalculating the
spectrum by printing the message
COOL IT.
5-11

on the Teletype. When it is completed, the
new spectrum is displayed.
2. 6

COPY

The contents of the currently displayed buffer can be copied
into the
other buffer, leaving the two with identical data, by typing CO:.
This is particularly useful if the manipulated
data is to be compared
with the original data.
2 . 7

XINVERT

The left to right X axis relationship of the
displayed array can be
inverted by typing XI
:

2. 8

YINVERT

Each Y value of the displayed buffer is subtracted
from 1000 and the
differences stored in that buffer, thus effectively
inverting the Y
axis by the command YI
:

2. 9

SCALE

If, when the data is displayed on
the scope, the maximum and minimum

poxnts "wrap around" because they are out
of the scope's range or if
the range is smaller than desired, SCALE
(scs) can be used to bring the
Y data into the range 0<Y<1000.
The scaling parameters used, original
minimum value, original maximum value, and
multiplier, are printed on
the Teletype after the computation, as,
for example:

MIN=
MAX=
MPY=

-54
75

7.752

where
MPY = 1000
(MAX-MIN)

and

Y. =
•

(Y.-MIN)
1

x MPY

NOTE
If the range between minimum and maximum
exceeds 2047, the Multiply (not the Scale)
Command must be used to adjust the data.

5-12

2.10

MULTIPLY

Data can be scaled to an arbitrary range other than the
to 1000 range
assumed by the SCALE routine. The MULTIPLY routine is used
mainly in
spectrum stripping when a standard and/or background spectrum
is to be
subtracted from a raw data set after appropriate scaling.
The command is used as follows:
1.

Type MU: to call the MULTIPLY routine.

The parameter

MIN=
is then requested.

This value is to be subtracted
from each data point before MPY is applied.

When
MPY =
is printed, type in the desired multiplier.
the following computation is performed:

(Y -MIN)
i

•

Then

MPY

The final data should be in the range -2047 to
2 ^ 47
Va lues outside of this range are truncated
10"
to
these limits. Final values less than -5
or greater than 1019 will "wrap around" when displayed.

2.11 SMOOTH

Collected data can be smoothed by an 11 point digital filter routine
via the SMOOTH command (SM:). The new spectrum is displayed on the
scope.
S

The first and last five points are not altered.
for functional equation.

2.12

See Appendix

CURSORS

Typing CU: displays two movable bright dots on the scope referred
to
as cursors.
They are used to implement the INTEGRATE and STRIP commands by delimiting the data points to be modified.
The two cursors
are controlled by four analog channel knobs as follows:
5-13

CURSOR

Analog Channel
knob

Direction of
movement

left

horizontal
vertical

right

horizontal
vertical

2
3

When the two dots appear on the display initially, knobs
always position one of them to the left of the other.

and

must

A grid of 1000
x 1000
on the scope is assumed.
The actual range
10
10
extends slightly beyond, but the data points should be within
that
range.

Typing CU: a second time removes the cursors from the scope.
Note that the cursors can be used to inspect peak amplitudes, valleys,
widths, etc., by positioning them appropriately and noting the
values,
using the LIST command.
2.13

LIST

To get a Teletype listing of the decimal values of the cursors,
type
LI:

The values are printed in the following order:
2 . 14

Y left; X, Y right.

INTEGRATE

The INTEGRATE command (IN:) provides two options:
1.

Integrate the data between the cursors by using them
to set a pivot or baseline, print out the area, and
leave the data array unchanged.

Integrate all data using the cursors (extrapolated)
to define a pivot line and store a scaled running
integral in the data array, then print out a scale

down multiplier.

The data is defined as lying on a 1000 x 1000 grid, independent
of the
total number of points in the array.
The area at a given point, i, is

expressed as

5-14

j-1

i -

i + l- i

}

[(Y

i + l"

i + l>

]

where X^ = 0, X = 1000 and s is the starting point which is equal to
n
the X value of the left cursor for option 1 or is equal to
for option 2.
Similarly, j E x value of the right cursor and P. (=P
is
.

)

the value of the pivot line at that point.

An integration is performed as follows.
1.

Type CU: and position the cursors as required for the
desired option (refer to Section 2.12).

Type IN: to call the routine.

The message

SCAN INT'L?
is printed.
Type Y if a running integral (option 2)
is desired; type N is a partial integral and area

printout (option

are desired.

If, when the routine is first called, the message

BAD X POINTS!
is printed, the left-right relationship of the cursors has been
inverted.
The display is restarted to allow correct positioning.
When corrected,

type IN: to call the routine again.
2.15

STRIP

Three options are available for altering portions of the displayed
spectrum.

Replace the data between the cursors with the best
straight line.

Subtract a sloping baseline from all data points.

Subtract a straight line interpolated between the
cursors from the data between the cursors.

The STRIP command is implemented as follows:
1.

Type CU: and position the cursors as explained
below.

5-15

Type ST: to call the STRIP program. 1

The message

STRIP PEAK?
is printed on the Teletype and the two dots are
displayed on the scope.
If some of the data points are
to be replaced by the best straight line, position
the left dot at the first data point to be replaced,
and the right dot at the last data point to be replaced.
Then type Y and the delimited portion of the spectrum
is replaced by a straight line.
The STRIP program is
exited.
Continue with another DAQUAN operation.
If
one of the other STRIP operations is desired, type a
response of N.
4.

A reply of N generates the message
FULL BASE?
If a (sloping) baseline is to be subtracted from all
the data points, the two dots should have been positioned on the scope with the desired slope and amplitude.
The dots need not be on the data curve nor on
the X axis extrema because they are extrapolated.
Type Y and the new display will appear.
If a straight
line is to be subtracted from some data points, i.e
partial baseline restoration is desired, type N.
•

If, when the STRIP routine is called, the message

BAD X POINTS!
is printed, the left and right dots have been inverted.

The original
display with the cursors appears.
Correct the dots by adjusting the
knobs; and type ST: to call the routine again.
2.16

DIFFERENTIATE

The derivative curve of a spectrum is computed by the
DIFFERENTIATE
command (DE:). Derivatives to a depth of at least six can be
calculated for some spectra with minimal distortion of the data.
These
are produced, not by adjacent point differences, but by
the following
procedure.

If STRIP is called without calling CURSORS first, stop
and restart
the program (refer to Section 1.4), call the CURSORS and
then STRIP
routines.

5-16

Y.=Y., -Y. .
1+2
i-2
1

with
Y.
i

Y, = Y, and Y
2

n-2

Yn ,
n-i

If multi-depth derivatives are desired, scaling and smoothing are sug-

gested before each level to reduce quantization error or "stairsteps"
It may also be necessary to strip
that result from integer arithmetic.
out the first and last five points if they interfere with scaling (they
are unaffected by smoothing)

2.17

PLOT

Any displayed spectrum can be plotted on an X-Y recorder interfaced to
a scope channel for a hard copy of that spectrum using the PLOT command
(PL:).
The size and rate of the plot are controlled by the user; the
pen is controlled by a relay driver triggered by the channel select
To
See Appendix Q for details.
bit of the display enable register.

generate a plot of the presently displayed spectrum, proceed as follows:
1.

Type PL: to call the plotting routine.

The first message printed on the Teletype is

LINE PLOT?

Type Y if a line plot is desired or type N if a
point plot is preferred.
3.

The plotting routine is now waiting for the user
A small
to calibrate an area on the X-Y recorder.
dot appears on the scope and will make the same
movements as the plotter pen. The speed for drawing the axes, as well as for the actual plotting,
is controlled by knob 3 of the analog channel controls and can be adjusted at any time while using
Turning knob 3 clockwise inthe PLOT command.
creases the rate of plotting; turning it counterclockwise decreases the rate. The axes are calibrated by typing the following letters to perform
the indicated operations.

Letter

Operation

locates maximum X coordinate

locates maximum Y coordinate

returns pen to X-Y origin
The pen on the X-Y recorder should be set initially
Type X and then manually reposition
to the origin.

5-17

the pen to the desired maximum
X coordinate.
After
P° Sltlon has been determined, type Y.
x
Set
thP
nS3 +#
the pen
tS the maximum Y coordinate (X
Y
larly.
The axes have now been determine1 * raax simifor the
Plot.
Type o to move the pen automatically
back
to
uu
the origin (X-Y,*)
)

e
n
dra
ma erS ° f the gra ? h can now be
q
marked of?
off on the plot, if desired.
Type M to mark off
the frame and quadrants.
Knob 3 controL the pen spefd.
When the axes have been marked, the
spectrum is readv
to be plotted.
Type G to initiate the plot
As before
analog channel knob 3 is used to
adjustthe plotting

^^^ ^

Any characters other than X, Y,
0, M or G during this
11
*
°" the Telet
arfignore'r

sage ifprfnt'ed™

^ »"

*** ***" Pl ° ttSd

the

*>"«*»* -s-

PLOTTER OFF?
Sh
P
d be ^ned °ff or set to stand-by
and
tnen Y°tvoed
then
typed. °ri
(N is also a legal reply.)
Typing any

^ "^
-

If the
Srne^orf
turned off, the pen will * start to moveploite/i"^?*
wildly
r

plf? starts ?

2.18

ll0W thS SC ° Pe anal ° g

in its

-gnalswhL^hfdis-

SUBTRACT

Using the SUBTRACT command (SU:), data
in the displayed buffer will
be subtracted from data in the other
buffer.
The result will appear
in the displayed buffer replacing
the previous data.
2.19

ADD

The two buffers can be averaged
and the result seen in the
dispiayed
buffer by issuing an ADD command
(AD:).
As with SUBTRACT, the data
in the displayed buffer will
be lost when the ADD command is

executed

2.20

SWAP

The data currently in buffer
1 can be placed in buffer
2 and the data
in buffer 2 placed in buffer
1 with the SWAP command
(SW:).
This

facility is especially useful with
ADD and SUBTRACT.
5-18

2.21

SQUEEZE

compresses the data array by a factor of 2
It is recommended that this comby averaging adjacent data points.
mand be used immediately after data input because the number of points
If the current buffer
is halved for each issuance of this command.
The SQUEEZE command (SQ:

)

is 1 or 2, only that buffer is affected, but if both buffers are

The
was typed during display) both are halved.
purpose of this command is to reduce the number of points in order to
diminish flicker and allow faster smoothing, scaling, etc.

being displayed

2.22

MODIFY

The DAQUAN program itself can be modified by a routine that is similar
(See
to ODT-8 (DEC-8-COCO-D) which makes changes in core locations.
Note on CALCULATE routine (Section 2.4). MODIFY is not available if

After calling (with MO:) the MODIFY routine, any
Each must
of the following can be performed in a logical sequence.
begin at the left margin.

CALCULATE was used.)

open location

Type the location (octal) and press the
The contents of that locaSPACE bar.
tion are printed followed by a colon.

change contents

After the present contents of a location
have been printed, type the new contents
and the letter C to enter the correction
and press the SPACE bar.

inspect next location

After pressing the SPACE bar, type N.
The next location and its contents are
printed followed by a colon.

reinspect same

After pressing the SPACE bar, the contents of the last location opened can
be printed again by typing S.

transfer control to

Type R and the location and press the
Execution of the program
SPACE bar.
will continue from that location
.

Pressing RUBOUT at any time will terminate the operation and restart
the routine.
2.2 3

OUTPUT

Collected data can be listed and/or output on paper tape by using the
OUTPUT command (OU:). After calling the routine, the question

5-19

TTY I/O?
is asked.

A response of Y implies the Teletype
punch, and a response
of N implies the high-speed punch.
Be sure the device is prepared before responding to the question.
The format is 10 columns of four
digit integers.
2.24

RESTART

The DAQUAN program can be restarted
from the beginning at any time.
The message

DAQUAN LIVES!
is printed.

or (under OS/8)

This command (RE:)

as described in Section 1.2.

2.25

DAQUAN OS-8!

is equivalent to restarting at
0200

PEAK Report

The PEAK processor will request X axis
scaling information (to allow
setting up X as time, wavelength, mass units,
furlongs, etc.) and a

detection sensitivity parameter.
Then a pass is made through the stored
data to detect the peaks and integrate
between the minima found.
Finally, a summary report is printed containing
the minima, maxima,
their scaled X value, the scaled raw area
between minima, the intensities normalized to the largest peak of
100, and the % area.
The data
array is left unaltered and can be subsequently
processed as desired.
Before calling the PEAK processor routine,
the user should call the
CURSOR (Section 2.12) and STRIP (Section 2.15) routines
to remove any
sloping or non-linear baseline deviation and
to remove spurious points
SMOOTHing is also advantageous.
It is not necessary to set the baseline down to zero, only that there be
a constant offset.
This is corrected for when areas and intensities are
created.
Type

PE:

DAQUAN prints
X1,XMPY,GATE:

5-20

requesting the X value of the first data point for XI, the X axis

multiplier or delta X for XMPY, and the detection GATE. The detection
algorithm requires a GATE'S worth of consistently increasing data
points followed by a GATE'S worth of consistently decreasing values.
This does not imply consecutive changes, just that there be no major
deviations from the trend.

Thus, the GATE is usually set to about 1/3

to 1/5 the number of points over the narrowest peak in the data with

values of 3-6 being typical.

Small values give added sensitivity

whereas large values allow small or narrow peaks to remain undetected.
Type RETURN after specifying these three values.

The computer prints:

MIN,X,MAX,X,A,I,%A:
TTY I/O?
The top line labels the subsequent columns where MIN, the first column,

minimum value for the peak, the second column is its
scaled X value.
Similarly, columns 3 and 4 are the maximum and X
value, then column 5 is the scaled raw area, column 6 the intensity
is the leading

normalized to the largest peak with a value of 100, and finally, column
7, the percent area.
Areas and intensities are computed based on the
lowest or minimum point in the data array which is thus defined as
baseline.
The TTY I/O question requests the output device to be used.

A response

of Y specifies Teletype punch; N specifies high-speed punch.

Typing

any other character causes a

to be printed.

Be sure the device

specified is prepared before responding to the question.
For example, this display

1000

5-21

when processed by PEAK would report the four peaks with
leading
minima, maxima, and area as K^ c^ A
m
c , A , etc. with areas
1
2
2
2
computed above the horizontal baseline through M
the lowest point,
,

using perpendicular line drop at the valleys.
I.e. ,

Pt.

where M

M.
J

Pt. M.

(Y.-M
1

)

•

XMPY

is the lowest point between C

and the last data point.
4
normalized intensities would be computed as
x

C.-M
4

5-22

The

CHAPTER

DAQUAN EXAMPLE

The printout on the left side of the page is an actual DAQUAN sequence.
The commentary on the right was added to indicate the opera-

tion performed.

Underlined information on the left is that typed by

the user.

DAQUAN LIVES!

PROGRAM IS READY

TITLE: DEMO

NAME THIS SEQUENCE "DEMOl"

OK, HIT ME!

READY FOR A COMMAND.
1 IS ASSUMED.

1CA:

CALCULATE SPECTRUM.

NO. POINTS=100_0
NO. PEAKS=5
XRNG, XI, INT MPLR, BASE, LOR FR (0-1)

1000 POINTS
PEAKS
PARAMETERS FOR SPECTRUM

BUFFER

1000,0,1,0,0
HGT, H-WIDTH,POS:

100,5,100
200,200
300,300
400,400
500,500

SPECIFY WIDTH FIRST PEAK ONLY
SUBSEQUENT PEAKS ASSUME SAME WIDTH

COOL IT
CO:
1AL:
PK, PA, VL: 1, 1,500

SPECTRUM BEING COMPUTED
COPY DATA ONTO BUFFER 2
ALTER SPECTRUM ON BUFFER 1
ALTER HEIGHT OF FIRST PEAK TO 500
CHANGE HALF WIDTH OF PEAK 2 TO 10
CENTER THIRD PEAK AT LOCATION 800
ALL THE CHANGES REQUIRED

PK,PA,VL: 2,2,10
PK,PA,VL: 3,3,800
PK,PA,VL:0,-1

COOL IT

SPECTRUM BEING COMPUTED
HALVE DATA POINTS ON BUFFER 1
DISPLAY CURSORS (POSITION CURSORS)
INTEGRATE BUFFER 1
PARTIAL INTEGRAL REQUESTED
AREA UNDER CURVE BASED ON 1000 X
1000 GRID
SMOOTH DATA

1SQ:
1CU:

UN:
SCAN INT'L? N
AREA= 0.1103J9E+06
SM:

5-23

CO:
CU:

COPY DATA INTO BUFFER 2
DISPLAY CURSORS (POSITION CURSORS)
STRIP OUT A

ST:

STRIP PEAK? N
FULL BASE? Y

(SLOPING) BASELINE

MU:

SCALE DATA

MIN= -52
MPY+.5

WITH THIS MINIMUM AND THIS
MULTIPLIER (AXIS IS HALVED)
ADD DATA IN DISPLAY BUFFER 2
OUTPUT THIS DATA
ON HIGH-SPEED PUNCH
ACQUIRE DATA. IN DISPLAY BUFFER 2
INPUT FROM BUFFER 1

21AD;
OU;

TTY I/O? N
2AV:

CHANNEL=1
N. POINTS=1000

1000 POINTS,
10 SECONDS/SCAN
10 SECOND DELAY,

SEC/SCAN=lj^~
DELAY (SEC) =10
SCANS=2

NO.

AV:

CALL AVERAGER TO TIME MEMORY CYCLE
USE THESE PARAMETERS

CHANNEL=0
POINTS= 1000
SEC/SCAN=12j
DELAY (SEC) =0
NO. SCANS=10

SCANS MAXIMUM

NO.

TIME THE PROGRAM FROM WHEN THE
SCHMITT POLARITY SWITCH IS STRUCK
TO WHEN STATIC DISPLAY BEGINS.
CHANGE TIME OF THIS PROGRAM
SET CONSTANT TO 1.22 MICROSECONDS

1TI:

MCT (SEC) 1.22E-6
12SW:

PLOTTER OFF? Y

VIEW AND THEN SWAP BUFFERS
SUBTRACT BUFFER 2 FROM BUFFER 1
PLOT SPECTRUM ON X-Y RECORDER
LINE PLOT
CALIBRATE AXES THEN MARK BOX
AND QUADRANTS
PLOT
DONE PLOTTING

RE:

RESTART DAQUAN

DAQUAN LIVES!

PROGRAM IS READY

TITLE DEMQ2

NAME THIS SEQUENCE "DEM02"

OK, HIT ME!

READY FOR A COMMAND

PA:
NO. POINTS= 600
—

READ IN FROM PAPERTAPE
600 POINTS
START WITH FIRST POINT
400 IS RANGE OF DATA
WITH MIN VALUE OF
VIA HIGH SPEED READER
INVERT LEFT-RIGHT RELATIONSHIP
ON X-AXIS
INVERT Y-AXIS

SU:

1PL:

LINE PLOT? Y
OXYOM

PTS TO SKIP=|F
YRANGE & MIN: 400 ,0

TTY I/O? N
XI:
YI:

5-24

XF"XMPY,GATE:1 1 3
MIN,X MAX,X,ATT7%A:

WANT PEAKS PROCESSED WITH FIRST
POINT HAVING AN X VALUE OF 1,
DELTA X OF 1 , AND A SENSITIVITY OF

TTY I/0?Y

TO BE PRINTED ON TELETYPE

448,
13,
44,
935,
160,

1.00,
22.00,
32.00,
56.00,
80.00,

506,
81,

957,
943,
799,

3.00,
28.00,
53,00,
67.00,
97.00,

0. 622200E+04 ,
0. 358000E+03,
0. 134590E+05 ,
0. 125840E+05 ,
0. 866200E+04 ,

52.22,
7.20,
100.00,
98.52,
83.26,

15.0708,
0.8671,
32.6002,
30.4808,
20.9810,

CALCULATE FIRST DERIVATIVE OF
DISPLAYED DATA

DE:"

5-25

CHAPTER 4
DAQUAN MS FOR OS/

4.0

GENERAL DESCRIPTION

DAQUAN MS is DAQUAN DEC-8E-ADQNA-A-PB
changed to run under OS/8
DAQUAN MS can be loaded by OS/8 and
returns to OS/8 via a CTRL/C
In
addition, data can be read and written
via mass storage (system device)
To accomplish this the CALCULATE
command has been taken out.
Hence
the commands CALCULATE and ALTER
are not present.
A new command MA:'
(mass storage) has been added.
The binary for MS DAQUAN is
DEC-8E-ADAQA-A-PB, or on DEC t ape
ALMSA-A-UB, the binary file is DAQD5.BN.
4.1

DEC-8E-

Loading DAQUAN MS

DAQUAN MS can run with the EAE or
non-EAE floating-point package.
two sources that can be used are
1

DEC 8E NFPPA
dIcIsTnSpa

The

* AE float ing-point package
non-EAE floating-point package

The floating-point binaries provided
with DAQUAN are:
PaperDEC-8E-ADFEA-A-PB ,.
EAE Fln^i™ t>
t,
tape:
D E c- 8E - MPPa - A - PB .
•

S-STSSTtMS'SK^

DEC-8E-ALMSA-A-UB:

Reassembling the Floating-Point Sources:
The EAE package is ready to be
assembled.
However, the non-EAE sources
must be edited before assembly. Edit
the origin statement

*FLPT-2600
to read

*FLPT-2000
Delete the origin statement *FLPT-2500.

Reassemble the new source with PALIII
or PAL8.
The source of the
floatxng-point package is distributed
on three paper tapes (DEC-08NFPPA-A-PA1, -PA2, and -PAS).
Only tapes 2 and 3 are used and
should
5-26

be assembled with PAL III or equivalent in the order they are numbered (tape 2 first, then 3) . Tape 2 ends with a PAUSE statement while
Tape 1 is the extended functions which
tape 3 ends with a dollar sign.
are not used.

several
If the package is assembled with PAL 8, the user must define

floating-point and several 8/E instructions which are used by the
interpreter, but not contained in the PAL 8 symbol table. The user
should create a paper tape which looks like:
FIXMRI FADD=1000
FIXMRI FSUB=2000
FIXMRI FMPY=3000
FIXMRI FDIV=4000
FIXMRI FGET=5000
FIXMRI FPUT=6000
FEXT=0000
FNOR=7000
SWP=7521
CAM=7621
MQA=7501
MQL=7421
SGT=6006
PAUSE

and use it as the first tape of the assembly (before tape

of the

package.
To load DAQUAN, use the Absolute Loader as follows.

In response to

the asterisk, type

*FPP.BN,DAQD5.BN

where FPP.BN is the binary of the floating-point package being used
and DAQD5.BN is the binary of DAQUAN.
The OS/8 save locations for the floating-point package and DAQUAN MS
are 0-7577.

Starting Address
Start DAQUAN at location 200 of field 0, restart at 203 and start at
Do not restart at 200.
202 to reenter the display routine.

5-27

4.2

Using DAQUAN MS

Under OS/8 DAQUAN MS prints DAQUAN OS-8!
when started. All commands are
the same as paper tape DAQUAN except
that CALCULATE and ALTER are not
present.
To write or read from mass storage
(system device), the MAcommand is given.
The OS/8 command decoder is called
to get the input
or output file.
In response to the asterisk, type
the standard OS/8
reply

*FILE.xx
for an input file, and

*FILE.xx<
for an output file.

If an output file is given, the
current channel
being shown is written to the system device.
The format of the data on
disk is a core image, the first word
of the data starting in the first

word of the mass storage file.

If the file is an input file,
two more questions are asked:
PTS. TO SKIP=

and
NO. POINTS=

Respond with a number less than 2048
followed by the RETURN key for
each question.
If more than 2047 points need to
be skipped, the file
can be reconfigured by the CONVERT
program.
4.3

Data Format

Data stored on mass storage from DAQUAN
is in single precision, the
first word on mass storage is the first
word of the channel buffer that
is written.
Data read by DAQUAN is single precision
data.
The values
should be between
and 1000.
if not, the MULTIPLY command
should be
used on the data to scale it down.
4.4

Errors

Any errors in the command string to
the command decoder cause an OS/8
error message. An error with I/O writing
or reading, or closing of
the file causes an error message
to be printed.
Check to see if there
is room on the system device and
try again.
Restart DAQUAN at 202.

5-28

CHAPTER 5
DAFFT OVERLAY FOR DAQUAN MS

DAFFT runs with DAQUAN MS and the EAE floating point package DEC-8ENEAEA.

DAFFT (Data Acquisition and FFT) contains all functional

aspects of DAQUAN MS except that:
1.

the MODIFY function is removed.

the SQUEEZE function is removed.

the number of points is limited to 1024.

It includes these added functions:

Performs complex FFT from single-precision data to give
complex single precision results.

Performs complex inverse FFT similarly.

Computes scaled power spectrum.

Allows temporary storage of 1024 values in a third buffer
area.

Has the option of 3-point or 11-point digital filtering
in time domain and Hanning filtering in frequency domain.

DAFFT allows signal averaging in the time domain (up to 1024 doubleprecision points) , FFT into frequency domain giving up to 1024 real
and 1024 complex coefficients (half of which are in positive and half
in negative frequency) , computation of the power versus frequency
spectrum as the scaled sum of the squares of those coefficients, input

and/or output of any of these arrays using the Teletype or high-speed
reader/punch, and allows retention of an extra 1024 point array in
core for recall of data from a previous step.
The software package consists of the DAQUAN MS, EAE floating point

package and DAFFT OVERLAY binaries.

The DAFFT OVERLAY contains a

copy of DECUS 8-144, J. W. Rothman's FFTS-C, and the modifications
and additions to DAQUAN MS. Also incorporated is the patch for FFTS-C
to correct a roundoff error.

5-29

5.1

To load DAFFT, the Absolute Loader is used.
asterisk, type

In response to the

*FPPEAE.BN, DAQD5.BN, DAFFT3.BN

where FPPEAE.BN is the binary of the EAE version of the
floating-point
package DEC-8E-NEAEA. DAQD5 BN is the binary of DAQUAN
MS followed by
the binary of DAFFT.
The OS/ 8 save locations for DAFFT are:
0-7577,10000-11575.
The start location is 200 of field 0, the same as
DAQUAN. MS.
If no mass storage is available, the programs
can be
loaded by the binary loader and started at 200.
The program then runs
as long as CTRL/C or the MA: command is not given.
The binary of
DAFFT is DEC-8E-ADAFA-PB.
.

Refer to Appendix T for core maps of DAFFT.
5.2

Operation

Since MODIFY and SQUEEZE are not available, requesting
them causes typeout of a question mark.
Similarly, do not at any time request more
than 1024 points. Also, any array upon which FFT or
IFFT is to be done
must be some power of two in length, i.e., 4,
8, 16... 256, 512, or 1024
points.
If the array is not a power of two, the
program halts and
must be reloaded with OS/8. Also, any array
to be transformed must
reside in buffer 1.

Relationship to Acquisition
The FFT of

real or complex input points yields N
2
complex coefficients, half of which are positive
frequency and half of which are
negative frequency. Most applications
use only positive frequency,
therefore, 2
points result from the transform.
For a desired final
result of 0-f max bandwidth, the input
should be band limited at twice
fmax and sample rate should equal
twice fmax to prevent aliasing and to
allow accurate frequency recovery.
2

In order to remove DC offset from
the data, before computing the FFT
the mean value of the array is
subtracted from each datum.
This zero
mean" correction forces the integral
of the array to zero.

5-30

5.3

Command Description

Table 5-1 summarizes the new commands.
Table 5-1

DAFFT Commands

Action

Command
FT:

ZT:

IT;

Calculate FFT of complex time domain data. Real coefficients are in buffer 1 and imaginary coefficients
are in buffer 2 before calling.
Real
Calculate FFT of real valued time domain data.
coefficients are in buffer 1. Buffer 2 is set to
zero before transform.

Calculate inversed FFT of complex frequency domain
Real coefficients in buffer 1 and imaginaries
data.
in buffer 2.

PO:

Compute power versus frequency spectrum as sum of
Buffer 1
squares of real and imaginary coefficients.
contains real coefficients and buffer 2 contains imaginary coefficients before calling.

SH:

Shift to or from buffer

SM:

Perform Hanning, 3-point, or 11-point filtering.

OF:

Move the display's x axis from the bottom to the middle
This
or from the middle to the bottom of the scope.
enables plus and minus y values to be displayed with
zero at mid-scopt.

and storage buffer.

FT and ZT create arrays including both positive and negative frequencies.

Zero frequency is set at left margin with increasing values to

the most positive frequency in the center of the scope.

The most

negative frequency is adjacent to that and remainder of display is
negative frequency going to zero again on the right margin. The power
display is analogous.
If an FT or ZT command is used to obtain an FFT of the data and then an
IT command is used to return to the original data, the result of the IT
operation must be multiplied (not scaled) so that the data displayed

represents the original data.

After data is entered (by AVERAGE or PAPER TAPE commands) ensure tnat
the real coefficients are in buffer 1 and imaginary coefficients (if
any)

are in buffer 2.

The real and imaginary coefficients are output

or input as separate arrays.

5-31

Shift Command

Typing
SH:

causes
SAVE?

to be printed.

Answer Y if the data in buffer 1 is
to be placed in
Answer N if the data in the storage buffer
is to be
loaded into buffer 1.
Type the RETURN key after either response,
to
continue execution.
a storage buffer.

SMOOTH Command

A choice of two time domain filters or a
frequency domain filter is
available. When
SM:

is typed,

the message
H, 3, 11 (0,1, 2):

is printed, requesting a value of

to execute Hanning smoothing of
the real and imaginary Fourier coefficients
stored in buffers 1 and
2, respectively.
The expression used is:
Y

= -1/4

Yi-1 + 1/2

•

1/4

•

i+1

It is designed to remove fractional
frequencies and to enhance resolution in the frequency domain.

A response of 1 causes the data in the
currently-displayed buffer to
be smoothed with a three-point filter defined
as follows:
Y

= 1/4

i-i

•

+ 1/4

•

i+1

Responding with
2

2 executes the 11-point filter as described
in Chapter
(Section 2.11) and Appendix S.

Complete the desired response with the RETURN key.

5-32

FFT Commands

Typing FT:, IT:, or ZT: causes the action described in Table 5-1.
When the transform is completed, DAFFT prints:
FACTOR= xx

where xx is a power of two by which the stored values are to be
multiplied to get the actual Fourier coefficients. All stored
values were divided by that power of two to prevent overflow of the
single-precision format.

The coefficients are inferred to be frac-

tions between +1 and -1 but for all intents and purposes can be considered as integers between +2048 and -2048. Then the actual Fourier

coefficient is found as
FC.

(Y.j/2048)

•

FACT0R

to
Since DAFFT, for purposes of display, assumes a number range of
display of the transforms will show scope wraparound. This
1000,
is allowed in order to retain maximum resolution of the data.

desired the arrays can be scaled down using the MULTIPLY command

with a MIN of 2000 (or 1000) and MPY of 0.25 (or 0.5).
This should be done for both the real array in buffer 1 and the
Use of the SCALE command is discouraged
imaginaries in buffer 2.
('Section 2.10)

since that routine assumes a data range of +1024 to -1024 and the
FFT or IFFT gives a range of +2048 to -2048.

However, if the power

spectrum is to be computed, it is recommended that no scaling be done.
This then allows the full resolution to be used.
POWER Command
To compute the power spectrum from the stored real and imaginary co-

efficients, type
PO:

DAFFT then requests

FACTOR=

Enter a value from 6-12 followed by the RETURN key. This is to allow
scaling to the desired range. The larger the FACTOR the smaller the

5-33

result.

The power coefficient is
computed as a 24-bit value then
divxded by 2 ACT0R
Use of a large FACTOR allows keeping
all power
coefficients on scale whereas a smaller
FACTOR allows presentation
of the weak power components.
Only the low order 12 bits are
retained.
.

The expression for computing
power is
P

i =

+ Y

-FACTOR

where Y^ is the real coefficient
(in buffer 1) and Y,,. is
the iraag mary coefficient (in buffer 2)
The power array replaces the
data
in buffer 1.

5.4

DAFFT Example

DAQUAN LIVES!
TITLE: DAFFT DEMO (+A)
OK, HIT ME]
AV:

CHANNEL=

ACQUIRE 10 SCANS OF 1024 POINTS
CHANNEL 2 WITH NO PRESCAN DELAY. AT 1024H7 from

NO. POINTS= 1024

SEC/SCAN=1
DELAY (SEC) =0
NO. SCANS= 10
SH;

SAVE ?

SAVE TIME AVERAGED DATA IN
STORAGE BUFF-fr

21ZT:

FACTOR=

2 IPO:

FACTOR=

1SH:

SHIFT STORAGE BUFFER BACK
TO BUFFER
RESTORE TIME AVERAGED DATA.

SAVE?N
1ZT:

FACTOR=

SM:
H,3,ll(j2(,l,2)=0f
PO:

FACTOR=

COMPUTE POWER SPECTRUM WITH
FACTOR

COMPUTE FFT AS ABOVE

APPLY HANNING FILTER
COMPUTE POWER SPECTRUM AS
ABOVE.

5-34

OF 2^'.
1

CHAPTER 6
PAFFT OVERLAY FOR DAQUAN MS

PAFFT (Power Average by FFT has all the features of DAFFT (Chapter 5)
but instead of time domain averaging, frequency domain averaging of the
power spectrum is done. This is accomplished by acquiring a synced
sweep of data in time, resettling to a zero mean, doing the FFT, applying
Hanning filtering, computing power, adding to the double precision power
)

average, and returning to allow the next sweep. A 1024 average requires
PAFFT is also
about four seconds dead time per sweep (set by FFT time).
limited to 1024 points and does not have the MODIFY or SQUEEZE functions.

6 .

The Absolute Loader is used to load PAFFT.

In response to the asterisk,

type

*FPPEAE.BN, DAQD5.BN,PAFFT2.BN

where FPPEAE.BN is the binary of the EAE version of the floating-point
package DEC-8E-NEAEA, DAQD5.BN is the binary of DAQUAN MS followed by
The OS/8 save locations for PAFFT are: 0-7577,
the binary of PAFFT.
The start location is 200 of field 0, the same as DAQUAN
10000-11575.
If no mass storage is available, the programs can be loaded by the
MS.
binary loader and started at 200. The program then runs as long as
CTRL/C or the MA: command is not given. The binary of PAFFT is
DEC- 8E- APAF A-PB

Refer to Appendix T for core maps of PAFFT.
6.2

Operation

The AVERAGE routine as described in Chapter

(Section 2.1) was modi-

fied by inserting the FFT and power average computations after each
sweep.

Also, when the run is complete the number of points for dis-

play is halved.

This will present only the positive frequency power

spectrum.

5-35

Also, when average is requested
and after NO. POINTS= (power of
2) is
specified, PAFFT requests

FACTOR=
The response is a value of 6-12
followed by RETURN.
This FACTOR is
analogous to that described in Chapter
4
(Power Command)

PAFFT requires using buffer

1 for the sweep acquisition,
both buffers 1 and 2 for the Fourier
coefficients, and the storage buffer
to
retain the double-precision power
average.
Therefore, no other arrays
preceding data, etc., can be contained
in core when a power average
is'
being computed.

When the average is terminated the
double precision power average
is divided by the number of
sweeps taken, truncated to the
limits
of +2047, and stored in buffer
1.

6 . 3

Command Description

The PAFFT commands are the same as the DAFFT commands
described in
detail in paragraph 5.3. Table 5-2 summarizes the new
commands.

Table 5-2

PAFFT Commands

Command
FT:

ZT:

IT:

PO:

SH
SM
OF

Action
Calculate FFT of complex time domain data. Real coefficients are in buffer 1 and imaginary coefficients
are in buffer 2 before calling.
Calculate FFT of real valued time domain data. Real
coefficients are in buffer 1. Buffer 2 is set to zero
before transform.
Calculate inversed FFT of complex frequency domain data
Real coefficients in buffer 1. and imaginaries in
buffer 2.
Compute power versus frequency spectrum as sum of
squares
of real and imaginary coefficients. Buffer 1 contains
real
real coefficients and buffer 2 contains imaginary
coefficients before calling.
Shift to or from buffer 1 and storage buffer.

Perform Banning, 3-point, or 11-point filtering.
Move the displays x axis from the bottom to the middle,
or
from the middle to the bottom of the scope.
This enables
plus and minus y values to be displayed with zero
at midscope
.

5-36

6. 4

PAFFT Example

DAQUAN LIVES!
TITLE
PAFFT DEMO (tA)
:

OK, HIT ME!

AV:

CHANNEL=0
NO. POINTS= 1024

ACQUIRE 1024 POINT POWER AVERAGE WITH SCALE
FACTOR OF 8 AND 10 SCANS MAXIMUM. FINAL ARRAY
IS 512 POINTS WITH
- 170 HZ BANDWIDTH (1024/3

FACTOR= 8
SEC/SCAN= 3
DELAY SEC )=
NO. SCANS= if

NO. SCANS=

AVERAGE STOPPED AT FOUR SCANS

(

SCALE DATA TO SIZE FOR PLOTTING

SC;

MIN=
MAX= 2047
MPY= 0.489

APPLY

SM;

POINT FILTER

H,3,ll(0,l,2):l
PL;

LINE PLOT7Y

MAKE A LINE PLOT OF DATA

PLOTTER OFF7Y
AV:

CHANNEL=

ACQUIRE POWER AVERAGE WITH SCALE FACTOR OF 9 TO
GIVE 512 POINTS WITH BANDWIDTH OF
- 512 HZ

NO. POINTS= 1024

FACTOR= 9
SEC/SCAN= 1
DELAY (SEC) =
NO. SCANS= 10"

5-37

BOOK

CONVERT

SIX

PROGRAM

CONTENTS

BOOK SIX
Page

Chapter

LAB8/E DATA CONVERSION PROGRAM

1.1

Loading and Starting

6-1

1.2

Commands

6-1

1.3

File Types

6-4

1.4

Input Data Formats

6-5

1.5

6-6

1.7

Output Data Formats
Histogram CONVERT Output
Basic and Advanced Average CONVERT Output

1.8

Correlation CONVERT Output

6-10

1.6

6-8
6-9

TABLES
6-1

6-3

Switches

6-i

CHAPTER

LAB8/E DATA CONVERSION PROGRAM

The conversion program is used to change data from one format to

another.

It reads a file from mass storage and writes the new file

on mass storage.

The program accepts single, double, or floating-

point data, FORTRAN IV and FORTRAN II data, and data from the LAB8/E
programs. The output of the data file can be single, double, or
The binary tape of

floating point, FORTRAN IV or FORTRAN II format.
this program is DEC-8E-ACVTA-A-PB.

The program name on DECtape is

COND14.BN(DEC-8E-ALMSA-A-UB) and C0ND14 (DEC-8E-AMASSA-A-UA1)
1.1

Loading and Starting

The CONVERT program can be loaded by the Absolute Loader and starts
at location 2000 of field 1.

To run the program, type .R ABSLDR to

OS/8; in response to the asterisk, answer

*CON/G=12000$

where CON is the binary of the CONVERT program and
ALT MODE.
1.2

The OS/8 save locations are:

represents an

12000-17177.

Commands

Once the CONVERT program is loaded and started, it asks two questions:
the FILE question and the CONVERT question.

FILE Question
The OS/8 command decoder is used to request the files.

In response

to the asterisk, type

*OUTFIL<INFIL^
where OUTFIL is the new file being created and INFIL is the old file.
If a file already exists with the name of the output file, it will
be deleted when the new file is closed.

6-1

For example, the command

*FILE1<FILE1^
takes FILE1 and converts it to
the new format with the name
FILE1
When
1
1
ClOS6d thS ° ld
1S delet6d
The -t-ndard OS/8
rubout format xs active and CTRL/C
takes the program back to OS/8
If
any errors are detected, the
command decoder outputs an OS/8
error
messa ge and r easks the
q uestion.
The fi les are specified ±a
QS/8 format
andean be read from and written to any OS/8 device
that handles binary

" ^
LT/

CONVERT Question
After the files are entered, the
program prints an asterisk.
with the CONVERT type in the format

Respond

*XX<YYj
or

*XX<YY/Z /Z

where XX is the output type and YY is
the input type.
z
z
etc
are conversion switches and are
discussed later. After the command'
is gxven the program changes
the file and closes the new file.
When
it returns to process another file
an asterisk is printed on the
Teletype for file input.
,

A RUBOUT deletes the last character
typed and types a backslash.
cessive RUBOUTs delete the next preceding
characters.

Suc-

Switches
The five switches, H, T, D, S, and R,
which can be specified to the
CONVERT question are described in
Table 6-1.

6-2

Table 6-1
Switches

Explanation

Switch

If
TIH output).
used only with Histogram input (PST or divided by 2.
is
present, each value of the histogram
histogram program
is used because the output of the
would be processed as
is a positive 12-bit number, which
were set.
negative if bit

file from
Allows an ID TITLE to be put on the output
ASCII in
packed
OS/8
in
written
is
The TITLE
CONVERT.
the data.
start
then
block 1 of the file. Block 2 will
PIP
that
so
CTRL/Z,
a
with
The ID TITLE is terminated
and reTeletype
or
printer
on
the
TITLE
the
can list
question ID- is
turn to OS/8. When a /T is given, the
deletes the
RUBOUT
A
title.
the
with
Respond
printed.
preceding
delete
RUBOUTs
Successive
last character.
typed to start
characters. A RETURN, LINE FEED must be
CTRL/Z
input.
be
can
Up to 255 characters
a new line.
S
the
with
used
be
not
can
/T
ends the title input.
switch.

Block 1
Used to delete block 1 from the input file.
to be one
have
not
does
but
should be an ID title block
ot
file.
a
on
Z ? ca
changed
be
to
title
/D allows the
AV,
with
used
be
not
be used with the S switch and can
CO or HI input.
This
file.
used to skip the first block of the input
changing
without
allows a titled file to be converted
switches
/S cannot be used with the /D or /T
the title.
input.
HI
or
AV
nor with CO,

™

front of
Used to delete any number of numbers from the
in the
numbers
of
number
the file and to write a desired
asked:
are
questions
two
given,
is
/R
output file. When

SKIP #=

and
READ #=

skipped and the
In response the number of numbers to be
and terminated
entered
number of numbers to be read are
consists of tne
then
file
new
The
key.
with the RETURN
on SI or DO
only
used
number of numbers read. /R can be
,

input

Error Conditions
are:
The four types of errors which may occur

Meaning

Message

USER ERROR 3
USER ERROR 2

An incorrect command was given
A mass storage read or write error occurred.
6-3

Mess age

USER ERROR
USER ERROR

1.3

Meaning
1

The output device is
full
11
dUrlng a ° ;eate

haSle?" "^

close or in the

File Types

Input

FORTRAN IV Floating-Point
format
FORTRAN II Floating-Point
format
Double precision
Single precision

DO
SI

DAQUAN data

FORTRAN II single precision
data
Histogram (TIH, PST) mass storage

output
Advanced and Basic Averager mass
storage output
Correlation mass storage output

AV
CO

Output
F4

FORTRAN IV Floating-Point
format
FORTRAN II Floating-Point
format
Double precision

DO
SI

Single precision
DAQUAN data

DA
FS

FORTRAN II single precision
data

Note that FORTRAN IV is RTPS
FORTRAN and FORTRAN II is OS/8
8K FORTRAN.

Convert Chart
Output
Input

where Y = yes and blank means
no
6-4

Input Data Formats

1.4

AV, HI, and CO are the output of the averagers, histograms and correla-

tion programs , as written to mass storage by the mass storage WRITE
command.
The format of the data is described in the write-up of those

programs
F4 input is the output of FORTRAN IV or OS/8 BASIC.

The data is
3-word floating-point numbers, 85 data points to a 256-word mass
storage block.
The last word of the block is not used.
The floatingpoint data has the format, word 1 is a 2's complement exponent. Words
2

and

are a 24-bit 2's complement mantissa with the decimal point

between bits

and 1 of word

when the number is normalized.

F2 input is the output of OS/8 FORTRAN II floating-point numbers.

The

numbers are 3-word floating-point numbers written under an A6 format
and then packed in standard OS/8 ASCII packing, three ASCII characters
per two words.

Hence, a 3-word floating-point number occupies four
words of the mass storage block. A maximum of 6 3 data points can be

stored per block.
The data must be written in A6 format, and an example of a WRITE statement is

WRITE (4,10) X
FORMAT (120A6)

where X is an array of 120 points.
ASCII and packed,
WORD

WORD
WORD

where

word

1
1

Each X value would be written in

is the 6-bits,

of output.

0-5, of word

The six bits of

1
3

the top four bits go in word

and is stored as 8-bit ASCII in
are changed to 8-bit ASCII, then

1 (3H) and the low four bits in word 2
At the end of each write there is an 8-bit packed ASCII 215,212.
The next write starts with the next 8-bit ASCII character.

(3L).

6-5

An F2 floating-point number is 3
words, 27 bits of mantissa,
bits of exponent, and a sign bit as
follows

A is the sign bit, B the exponent, C
the high-order 3 bits and D
the low 24 bits of mantissa.
The value is expressed as sign
magnitude
times an exponent.
The exponent is stored as exponent
-200 octal,
and
is expressed as all zeros.
DO, Double precision, input is
assumed to be 24 bit 2's complement

numbers, with the high order first
followed by the low order.
are 12 8 double precision numbers to
a block.

There

SI, single precision, input is 12
bit 2's complement numbers, 256
numbers to a block.

DA is single precision numbers from
DAQUAN.
FS is single precision numbers from
OS/8 FORTRAN II.
The numbers must
have been written in an A2 format by
FORTRAN. Like F2 output, the 6bit ASCII is packed as 8-bit ASCII by
OS/8, the record ended by an 8bit packed ASCII 215,212.
The next record follows immediately.
An
exmaple of a write by FORTRAN II is

WRITE (4,10) IX
FORMAT (120A2)

where IX is an array of 120 points.
NOTE
The output of CONVERT in F2 and FS
formats can
b e later converted by the CONVERT
program.
2
u
Each block output by CONVERT has two
dummy characters at the end and as a result is
not the
same as the output format of F2 and
FS data.
1.5

Output Data Formats

F4 output is the same as input.
Data is written 85 floating-point numbers per block until the input file is
exhausted.

6-6

F2 output is written in OS/8 packed 8-bit ASCII so that it is readable

under A6 format by OS/8 FORTRAN II. Each block of output contains up
The last four words of the block are
to 6 3 floating-point numbers.
Hence, the longest
used to fill out the block and end the record.
The format of the output block is

record is 63 values.
Words

Words 253

floating-point values in A6
format

252

256

0, 0, 0, 0, 215, 212 in packed

8-bit ASCII
If the last block is not full the above four words follow the last

floating-point number.

To read 200 data points of converted data,

the following FORTRAN II statements can be used,

DO 1 1=1,3
READ (4,10) (X(J+(I-1)*63) , J=l ,63) IX, IY
FORMAT(6 3A6,2A2)

CONTINUE
READ (4,11)

(X(I) ,1=190,200)

FORMAT (11A6)

Note that IX and IY are dummy variables to read the two single-

precision numbers at the end of the data before the end of record
mark.

DO ,

double precision, output is the same as for DO input.

SI , single precision, output is the same as for SI input.

DA , DAQUAN, output is single-precision numbers.
FS , single precision, output for FORTRAN II is in OS/8 packed 8-bit

ASCII in A2 format.

There are 189 numbers per block.

The last four

words of the block are the same as for F2 output.
If the last block
of output is not full, the last number is followed by the four words to
mark the end of records as in F2 output.
To read 200 words of single-

precision output, the following FORTRAN II statements can be used.

10
II

READ (4,10) (IX(I) ,1=1,189) ,IY,IZ
FORMAT (191A2)
READ (4,11) (IX(I) ,1=190,200)
FORMAT (11A2)

6-7

NOTE
IY and IZ are dummy variables to read
the
two single precision dummy values at the
end of the block.

1.6

Histogram CONVERT Output

The output of the histogram program is
page
and two histograms.
The
histogram values are single precision 12-bit positive
numbers.
However, if the convert output is to be used by the
DAQUAN program or is
single precision data, the histogram values are
treated as a 2's complement number. The H switch on the convert type causes
the histogram
data to be divided by two before conversion to ensure
all positive results.

NOTE
Only the numbers in the histogram are
divided, not any other output of the
histogram program.

Output of Histogram Input

1st Number - number of points in zero th histogram
2nd Number - zero

Parameters

Values

histogram bin width

3rd Number - number of points in histogram
4th Number - bin width of histogram
5th Number - min. time
,6th Number - is the ORDER for PST (unimportant value for TIH)
7th Number ~ is the firs t value of the zero th histogram
f
\^8th Number - is the second value of the zero th histogram
followed directly by the histogram

The values accepted by the histogram program during
setup go up to
4095, hence if a value was given that is greater than
2047 and the
histogram data is converted to single precision, the input
values
will be treated as 2's complements and converted.
If the output is
double precision or floating-point, the setup value will
always be
positive.
The output of the setup value is:

INPUT TO HIST.

ORDER
BIN WIDTH
BINS
EPOCHS
MIN TIME
TIME/UNIT
UNITS

to
.1 to
to
to
to
.1 to
to

OUTPUT OF CONVERSION

4095
409.5
1890
1890
4094
409.5
1890

6-8

to
to
to
to
to
to
to

4095
4095
1890
1890
4094
4095
1890

The histogram numbers range from

to 409 5 as output of the histogram

If the numbers are not scaled by /H, they will be treated

program.

as 2's complement numbers if converted to single precision but will

always be positive

to 4095 if converted to double precision or

floating point.
1.7

Basic and Advanced Average CONVERT Output

The Averager program writes out the job lists followed by the Average

The job lists are seven words long and describe the average.

data.

The average data point can have three different formats:

average

alone; average and confidence limits; or average, confidence limits

and trend.

The average and confidence limit are in double precision

and the trend is single precision.

The high-order part of the average

and confidence limit is the integer part and the low order the frac-

When the data is typed out by the Averager program in Section 4,
with a scale factor of 0, the high order plus one bit of low order
is typed for the average and the confidence limit.
tion.

If the output of the CONVERT program for the Averager is single pre-

cision or DAQUAN, the high order part of the average and confidence
limit is output.

The trend is single precision to begin with.

For

double precision or floating-point output, both the high and low parts
are converted by the CONVERT program without outputting the decimal

point.

The CONVERT program treats this number as a 24-bit integer

while converting it; the user must divide the result by 4096 to find
the average integer.
The output file format is as follows:
1st Number

1 if

low resolution average

if high resolution average

2nd Number

Channel number

3rd Number

4th Number

Data type =1 if Average only
=2 if Confidence limits
=3 if Average, Confidence Limit and Trend
Sort code

5th Number

and up

6th Number

Number of points in the average
Number of sweeps

7th Number
etc.

Data points

Following the data of an average will be the first number of the next
average or -1 if no more averages.

6-9

A data point

is as follows for Type 3 data:

Average value
Confidence Limit
Trend
and for Type

data:

Average Value
Confidence Limit
For Type

data it is:

Average
If the output is going to DAQUAN the data is rearranged
so that all
the average values are together followed by the confidence

limits and

then the trends, if they exist, as follows:
1 -

6th numbers

7th

number

average data header
number of sweeps

Xth numbers

average data

confidence limits, if they exist
trends if they exist

X+l-Y

Y+l-Z
Z+l

1.8

numbers
numbers

number

first number of next average, or
-1 if no more averages.

Correlation CONVERT Output

The output of the Correlation programs is page 0, the temporary
buffer and the correlogram.
The temporary buffer starts at the first
word of block 2 and is N+l words long, where N is the number of
points,
The output of the conversion program is as follows:
1st Number

2nd Number

3rd Number

zero if Auto, -1 if Cross-correlation
RATE input to program, 1 to 2047 represents 1 to 204.7
number of points

4th Number

sweep count

5th, 6th

dummy numbers

7th Number
etc.

correlogram

6-10

The sweep count is in double precision and the correlogram data is
in double precision; hence, if the output is in single precision, the

low order part of the correlogram and sweep count is obtained.

However, if the output is in single precision format, the sweep count
low-order is in word 4 and the sweep count high order in word 5.
If
N is the number of points collected for the correlation, the N numbers

starting at number 7 are the correlogram. The correlogram numbers
which are converted are the double precision core values of the correlation program.

The value typed by the correlation program under the

/T command is

VALUE (i) = RAW DATA of i th point
(number of sweeps) -i

.-„,

The data which is output by the CTRL/W command in the correlation
program and then converted by the convert program is the "RAW DATA
th
of i
point".

6-11

BOOK

SEVE

BASIC

CONTENTS

Pa ? e

INTRODUCTION

USING BASIC/RT

7-1

1.0

SCOPE COMMANDS

7-2

1.1

PLOT

7-2

1.2

DELAY

7-3

1.3

CLEAR

7-3

1.4

Display Buffer

7-4

1.5

USE

7-4

CLOCK COMMANDS

7-6

2.0

GENERAL DESCRIPTION

7-6

2.1

SET RATE

7-6

2.2

SET CLOCK

7-7

2.3

TIM

7-7

2.4

WAITC

7-8

2.5

CLS and CLC

7-8

A/D COMMANDS

7-10

3.1

ADC

7-10

3.2

REAL TIME

7-10

3.3

ACCEPT and REJECT

7-11

3.4

Buffer Capacity

7-11

3.5

ADB

7-12

TEST AND PAUSE COMMANDS

7-14

TST

7-14

WAIT

7-14

USER COMMANDS AND FUNCTIONS

7-15
7-15

User Command (UCOM)
User Coded Function

ERROR MESSAGES

7-18

CHAPTER 1

CHAPTER

CHAPTER 3

CHAPTER 4

CHAPTER

7-17

TABLES
Table

7-1
7-2

Clock Enable Register Functions
BASIC/RT Error Messages

7-i

7-9
7-18

CHAPTER 1
USING BASIC/RT

INTRODUCTION
BASIC/RT provides the researcher with a powerful and complete software package for total programmable control of all LAB8/E peripheral
analog channels, Schmitt triggers, crystal clock and display
devices:
In addition, commands to specify parameters such as pulse rate
scope.
Buffer aland response time permit optimum experimental flexibility.
location commands for the display and analog channels facilitate computer efficiency. Using the timing commands, a delay before sampling,
pause until a user response is typed on the Teletype, or a pause
until an interrupt occurs, can easily be included in a program. Another
feature, the user's command, permits customized system software.
a

All of the features provided by 8K BASIC are also implemented in
BASIC/RT, permitting total programmable I/O control and user coded
The entire BASIC/RT command and function set is included
functions
For additional information on BASIC
in Appendix U of this manual.
.

fundamentals, refer to the EduSystem-10 System User's Guide for
details on 8K BASIC refer to the 8K BASIC User's Manual DEC-08-SKXA-D.
BASIC/RT does not run under the OS/8 system.
;

Loading Procedure
Use the following procedure to load BASIC/RT:

Either the high- or low-speed
Load the RIM Loader into field 1.
Set the data and instruction fields to 1
reader can be used.
(use the EXTD ADDR LOAD key) before toggling in the first inNote that after pressing DEP or EXAM, the LAB8/E
always indicates the Program Counter location, not the Memory
To see what conAddress location, in the upper row of lights.

struction.

tent has just been loaded, set the multiple register knob to MD
(Memory Buffer)

Load the BIN Loader into instruction field 1.
ADDR LOAD key to load in this field setting.)

7-1

(Use the EXTD

Steps

and

can be replaced by loading
the HELP Bootstrap.

Read in the BASIC/RT paper
tape.
struction field 1. ^en loading

Use data field 1 and invia the BIN Loader, press

START CLEAR then START CONT
keys to initiate reading of
the
tape,
when program is loaded, it
prints the message READY.

1.0

SCOPE COMMANDS

The display scope on the LAB8/E
can be programmed to plot
points on
its screen.
The scope commands provide
complete control for graph
location and size, display time
and number of points displayed.
1.1

PLOT

When plotting on the scope, which
is rectangular, BASIC/RT
uses these
dimensions for its perimeter:
0<_X< 1

0<_Y<1.0

Thus, any plot must be within
the above limits, which can
easily be
accomplished by inserting a scaling
factor.
The PLOT command causes the
appropriate point to be displayed
on the
scope.
It is issued in the format:

PLOT x,y

where x and y are any expressions
which equal the actual X ard Y
coordinates of the point to be plotted.
An acceptable sequence for
plotting a straight line across
the middle of the scope is:

7-2

B=.5
FOR A=0 TO 1.1 STEP .01
PLOT A,B
NEXT A

10
20
30

Remember that every X,Y set must be within the specified ranges.

it is not, that data set is simply not displayed.

BASIC/RT displays points on the scope when it is not doing any internal calculations.

Data is displayed, for example, while waiting for

input or output.

If a calculation is required during a plotting se-

quence

pleted.

the data is not displayed until all the calculations are com-

Thus, in the following example, plotting a decaying sine wave,

the function is not displayed until all the points have been calculated.

200
210
220

FOR S=0 TO 1.10 STEP .006
PLOT S, SIN(S*35)*EXP(-S*2.5)/3+.5
NEXT S

DELAY

BASIC/RT provides a command that refreshes the scope after each cal-

culation so that the progress of the graph can be seen.

This command,

DELAY, causes BASIC to display all x,y points calculated up to this

statement.

Thus, the decay of the sine wave above can be viewed after

each point is calculated by adding to the above example the command
215

DELAY

The DELAY command provides the additional time for BASIC to display
the point before continuing to the next statement.

CLEAR

BASIC/RT also permits erasing the scope under program control.
inserting the statement:

CLEAR

7-3

all points currently displayed are removed from the screen.
Thus, if
more than one plot is required by a user program and it is not neces-

sary for them to overlap, a CLEAR command between calculations erases
the scope for the second plot.
In the next example, two compounded interest sums, $400 at 7% and
$450
at 6.25% per annum compounded yearly for 30 years, are plotted.
If
the intercept point is to be noted, then line 120 can be omitted and

at completion the two curves will be displayed together.

100
102
104
110
120
130
134
140
150
500
510
520
530
540
550
560

I=.07
P=400
T=30
GOSUB 500
CLEAR
I=.0625
P=450
GOSUB 500
STOP
FOR N=l TO T
X=N/35
Y=(P*((I+l)+N)/4000)
PLOT X,Y
DELAY
NEXT N
RETURN

Lines 510 and 520 include scaling factors for the scope.
1.4

Display Buffer

The technique for plotting points employed by BASIC/RT includes creating a buffer in the user's area of core to store all the calculated

points before they are displayed.
(This buffer area is considered as
a dimensioned variable; thus, executing the commands SCRATCH, RUN,
and END removes the buffer from core.
The CLEAR command does not delete the buffer, it merely erases its contents.)
When a PLOT command
is encountered, BASIC/RT checks to see if a display buffer has already
been assigned and, if not, then space sufficient for about 500 points
is allocated.
If this amount of room is not available in core, the

error message TOO BIG is printed.
1.5

USE

The area created by the PLOT command is approximately equal to a 333dimensional array.
In order to conserve space, if less than 500 points
are to be plotted, or to plot more than 500 points, a buffer dimensioning

7-4

command is provided so that core can be allocated optimally. This
command, USE x, is implemented as an array, as follows (x is always
a variable)

DIM P(200)
USE P

20
30

Line 30 says

use P as a storage

buffer for a future PLOT command;

Line 20 creates
If a user-assigned or BASIC/RT
enough room for about 300 data points
during execution
overflows
generated buffer is not large enough and
do not create an additional array at that time.
1

the error message TOO BIG is printed.

Only one USE statement is effective for any plot sequence and, if it
The variable
is to be used, it must be issued before the PLOT command.

associated with the USE command is active until one of the buffer removing statements (RUN, SCRATCH, END) is encountered, but it can be
used for another purpose when not currently required for displaying.
In the example in Section 1.3, the maximum number of points to be

plotted is 30, so a considerable amount of core can be conserved by
specifying a smaller buffer. The following code can be added to
optimize the program.
10
20

DIM M(40)
USE M

*To determine the minimum array size for point displays, consider that
each point requires two values (X and Y) and that every three values
require one word of memory. To display P points an array dimensioned
for W words is necessary:
W=.67(P)

7-5

CHAPTER 2
CLOCK COMMANDS

2.0

GENERAL DESCRIPTION

The hardware clock provided in the LAB8/E, like all the other
devices,
can be handled under program control to maximize its
performance, by
specifying parameters such as pulse rate and initial pulse
time.
2.1

SET RATE

To set the clock to interrupt at a specific rate, use the
command

SET RATE mode, time

where mode 1 is the desired clock speed (0-7) and time is the
number
of clock "ticks" between interrupts, up to 4096 counts.
The appropriate mode value is derived from the next chart 2
.

mode

1
2
3

5
6
7

rate
Stop
external input
10~2 seconds
seconds
1J?~
10~ 4 seconds
10~5 seconds
10~6 seconds
Stop

Thus, for the clock to interrupt at

second intervals, an acceptable

command is
SET RATE 3,1000

which causes the clock to wait 1000 one msec, ticks. If the
specified
clock rate is too fast, so that the interrupt cannot be serviced
in
time, the error message RATE ERROR AT (line number) is
printed and the
clock stopped.
The line number printed is that of the statement
cur-

rently being executed.
*If an illegal mode is requested, BASIC tries to use
the value- -3
will be taken as 3, .03 as 0, etc.
*Refer to DK8-EP Clock Description in the PDP-8/e Small C omputer
Hand*

book.

7-6

Note that in the 100-200 ysec elapsed total time range BASIC/RT is
servicing the interrupt correctly, but is not executing any BASIC/RT
commands because of the high rate; in this case, the processing has

been suspended. Type CTRL/C, which will cause the RATE ERROR message
to be printed, to restart.
2.2

SET CLOCK

BASIC/RT provides another command for setting the clock rate for the
Its format is the same as that for the SET RATE
Schmitt triggers.
command, namely
SET CLOCK mode, time

except that mode is a 12-bit decimal number which will be used to load
1
thus permitting the user to enable any
the clock enable register
function he chooses. Refer to Table 1 to determine the appropriate
The time parameter is specified in the same manner as with
value.
,

2
the SET RATE command .

Any time either of the SET statements is encountered, the time counter
Then, any time a clock interrupt occurs, this counter is
is zeroed.
Up to approximately 16,000,000 counts can occur before
incremented.
this counter resets itself to zero.
2.3

TIM

At any time in the program, the current count (number of elapsed interrupts) can be determined via the function TIM(n).
This function can be used in conjunction with any of the BASIC commands
so that the value can be printed or the next action to be performed by
The format of the function
the program can be dependent on the count.
call is TIM(n) where n is any argument (the argument is not checked by
In the following program the elapsed time for the plot
BASIC/RT)
determines the next action; print the count and halt or, for 50 data
.

points, print the sines and the terminating count and then stop.
In the first run, the time elapsed.

By changing line 10, the sine

table was generated.

BASIC/RT always forces the interrupt request bit on.
The 2's complement of the value of time is performed so that it is
in a form suitable for the clock.
1

7-7

20 SET RATE 2,20
30 FOR M=0 TO 1 STEP .01
40 PLOT M,M+2
50 DELAY
60 NEXT M
IF A>TIM(0) THEN 200
70
80 PRINT TIM (23)
90 ST0P
200 FOR Z=0 TO 1 STEP .02
210 PRINT SIN(Z)
220 NEXT Z
2 30 PRINT TIM(C)
240 END
2. 4

READY
10 A=12
RUN
nt

01999867
03998933
-03998933

WAITC

Another application of the clock is
to halt program execution until
a
clock interrupt occurs.
The WAITC command performs this
function
thereby permitting BASIC/RT to display
on the scope while waiting for
the interrupt to signal continuation
of program execution.
2.5

CLS and CLC

There are two other functions that are
used with the Schmitt triggers
The first, CLS(n), obtains the
status bits of the clock after the
previous interrupt (those obtained from
the PDP-8 CLSA (Clock Status
to AC) instruction).
This value can be tested, for
example, to determine how the interrupt occurred.
The other, CLC(n), performs a
CLCA
(Counter to buffer to AC) and returns
the value as a floating point
number.
Refer to PDP-8/e Small Computer HanHhnoV

7-8

Table 7-1

Clock Enable Register Functions

Function

AC Bit

Enables clock, overflow to cause an interrupt.
1

Mode

Counter runs at selected rate. Overflow occurs
every 4096 counts. Flag remains set.

Counter runs at selected rate. Overflow causes
Clock Buffer to be transferred to the Clock
Counter, which continues to run. Overflow remains set until cleared with IOT 6135.

Counter runs at selected rate. When an enabled
event occurs, the Clock Counter is transferred
to the Clock Buffer, and the Counter continues.

Counter runs at selected rate. When an enabled
input occurs on channel 3, the Clock Counter is
transferred to the Clock Buffer, and the Clock
Counter continues to run from zero.

Count Rate
Stop
1 MHz
100 KHz
10 KHz
1 KHz
100 KHz
External input
Stop

111
110
101
100
011
010
001
000

(When the Clock Counter overflows,
Overflow starts ADC.
type AD8-EA, is started.)
converter,
the analog-to-digital

When set to a 1, the Crystal Clock is inhibited from
generating clock pulses that increment the counter.
Events in Channels 1, 2, or
quest.
9,

10 & 11

Enable Events
9

10 11 -

Event
Event
Event

and

2
1

7-9

cause an interrupt re-

CHAPTER

A/D COMMANDS

3. 1

ADC

Any A/D channel can be sampled at any
time by using the single function ADC(n), where n is the channel
number,
to 15 , to be sampled
in a direct or indirect statement.
The ADC function performs an immediate conversion; the clock, however,
can be incorporated so that
samplxng occurs at an established clock
rate.
In the next program,
BASIC/RT waits for a clock tick and then
prints the value of the clock
using the TIM function, and of A/D
channels 3 and 4 for fifty samples.
300
310
320
321
322
323
330
340
350

3. 2

SET RATE 2,60
FOR P=l TO 50
WAITC
A1=ADC(3)
A2=ADC(4)
T1=TIM(0)
PRINT T1,A1,A2
NEXT P
STOP

REAL TIME

The ADC function is restricted to
non-time critical work because the
finite amount of time elapsed between
clock ticks may not be sufficient
to perform the tasks requested between
ticks (for example, printing
three values above). Also, more than
one channel cannot be sampled
the same time quanta (for example,
sampling channels 3 and 4 above).

For time critical operations, the REAL
TIME command should be used
because it provides a buffer to hold
the sampled value prior to processing.
Its format is:

REAL TIME v,cl,nl,n2

where v is a subscripted variable to
be used as the data buffer.
The
variable is assigned in a manner analogous
to the USE command for the
scope, namely, as a dimensioned array.
Because only one value is to
be taken per sample, three samples are
stored per buffer word.
Thus
a dimension of 100 can store 300 data
items.
The array cannot be
dimensioned larger than approximately
750, or a maximum of approximately
2,200 points.
The parameter CI is the first channel
to be sampled,
7-10

is the
nl is the number of consecutive channels to be sampled, and n2
sample
number of clock ticks for which to sample. To prepare to
suitable
channels 1 and 2 once every millisecond for 150 milliseconds,

code is
SET RATE 4,10
DIM G(100)
REAL TIME G, 1,2,150

Operation of the REAL TIME command is independent of the BASIC/RT
statement processing speed; as long as there is sufficient buffer
space, the REAL TIME statement will work.
3.3

ACCEPT and REJECT

To
The REAL TIME statement only creates the specified data buffer.
actually initiate sampling, the statement ACCEPT is required. Then
sampling will start at the next clock tick. There must be an active

REAL TIME statement or the ACCEPT is ignored. A REAL TIME command
becomes inactive when the clock count equals zero. To suspend sampling
use the command REJECT. This command is also useful for executing

subsequent REAL TIME statements.
In the next example, the ACCEPT and REJECT statements are used to

be sure sampling occurs at the specified rate for only the designated
Statement 500 stops the clock, 530 prepares it so
number of counts.
that the first sample is taken after processing statement 540. The
REJECT at 560 assures that after 100 counts no extra samples are

taken at the rate of 10 mseconds.

500
510
520
5 30

540
550
560
570
580
590

3.4

SET RATE 7,0
DIM G(100)
REAL TIME G, 1,2, 150
ACCEPT
SET RATE 2,10
IF TIM(4)<100 THEN 550
REJECT
SET RATE 3,10
ACCEPT
END

Buffer Capacity

In the last REAL TIME example, the buffer, G, is subscripted to accom-

modate 3*100=300 items.

The program, however, could have been required

7-11

amp ia a greater total
number of points.
If the 520 was REAL TIME
G, 1,6,150 then a total of
6*150=900 items must be accommodated.
To
handle this variation without having
to allocate large amounts of
core
as buffer space, BASIC/RT
uses a ring buffer so that
data can be
removed from the buffer before it
is overwritten by a new,
incoming
item.
To meet this condition, the
sampling can be:
a

'•

** *

Jamp i?ng7 of*

tine
'

—

-tent time

At a fast rate for a short time.

If a fast rate is required
for a longer period, a large
buffer can be
created, provided this space is
not required by BASIC/RT for
variables
etc.
If the buffer becomes
completely full such that data is
being
overwritten, the message A-D FULL is
printed
The max±mum
rate ia about 4 KHz or 250 s.;
the steady state long term rate
M
is
about 50 points/second.
.

3.5

ADB

To retrieve data collected by
REAL TIME and ACCEPT sequences
that are
Placed in a buffer, the ADB(n) function
is required.
Data is withdrawn from the buffer in the same
order in which it was entered.
Thus
if four A/D channels, 1-4, are
being sampled, the order of the
data in'
the buffer is

1,2.3.4.

iiii

l+l i+l J i+l 4 i+l

i
1

i+2'

The argument in the ADB function
is ignored.
The items will be removed
consecutively,
if there is no REAL TIME or
ACCEPT statement or if
there is no data remaining in the
buffer (because the number of
clock
ticks, „ 2 in the REAL TIME
statement had expired), the error
message
NO A-D and a line number are
printed.
,

This example is an expansion of
the one in Section 3.3 to
incorporate
the ADB function.
Lines 590 through 610 have been
added so that the
sampled values can be printed.

7-12

SET RATE 7,0
DIM G(100)
REAL TIME G, 1,2, 150
ACCEPT
SET RATE 2,10
IF TIM(4)<100 THEN 550
REJECT
SET RATE 3,10
ACCEPT
FOR A=l TO 150
PRINT ADB(l) ,ADB{2)
NEXT A
END

500
510
520
5 30

540
550
560
570
580
590
600
610
6 30

requested
The user must keep track of the items in the buffer (number
the
In
and number removed) and their ordering by channel number.
above example there will be 75 lines printed, each with two data
values in the format:

\
h+l

2
2

i+l

h+2
If line 140 was PRINT ADB (0) ,ADB (6) ,ADB (2)

would be in the order:

2.
l

h+2

i+l

i+ 2

7-13

the information printed

CHAPTER 4
TEST AND PAUSE COMMANDS

4.1

TST

For automated environments
requiring a response typed on the
Teletype
to determine the subsequent
action, a testing function, TST(n)
is
available which checks the Teletype
to see if a character has been
typed.
If lt has, a non-zero is
returned; if it has not, a zero
is
returned.
The TST function is particularly
handy in conjunction with
the TIM function for specifying a
maximum response time. The TST
function. must be reset by doing
a GETC function for the TST
statement
to be used more than once in
a program.
4 . 2

WAIT

If a pause time is required
by a program, the WAIT command
can be

xncluded in the program.

interrupt occurs.
reactivate BASIC/RT.
any.

BASIC/RT processing will be halted
until
Note that a clock interrupt is
sufficient to

7-14

CHAPTER 5
User Commands and Functions

5.1

User Command (UCOM)

The USER COMMAND is a feature in BASIC/RT which allows the user to
insert his own code and to effectively communicate with BASIC/RT.

This command differs from the UUF function in that it is more sophisticated and hence more complicated in its operation. To briefly
explain the differences: UUF is a function , such as SIN and COS, while
A function takes an expression in parentheses as
UCOM is a command
.

For example,
its argument and returns a value to be used later.
Y=INT(4.5) takes 4.5 and returns 4.0. A command is a statement which
may take arguments (such as READ or INPUT) or may not have any (such
In a
as CLEAR or END) , or they may be optional (as with PRINT)
,

function such as UUF, BASIC/RT has already evaluated the expression
and left the results in the FAC before calling the function so the
function has no control over what is in it. With a command, it is up
To successfully
to the user code to scan and determine what is wanted.
use this command, a knowledge of PDP-8 code and the floating point
It is suggested that the coding for commands
package is essential.
such as CLEAR, PLOT, USE, (and any others) be studied in order to
understand the nature of the commands. Some brief examples are illus-

trated next to show how to insert some simple commands.
In this example, a UCOM is created such that whenever it is issued,
The first
it executes a 6123 (some random IOT) with the AC all 7's.
thing to do is allocate space. There is room in upper and lower core.

Put a
In this case, let us overlay the EXP function with the UCOM.
pointer in the execute table to the location of the command. The table
entry will be about location 306 (refer to a listing for the exact
Now, whenever a UCOM is given, control will transfer
where this location points. The code will then look like:

location).

*EXP

CLA CMA
612 3

CLA
JMP

(DEVCOM

/Set the AC to all 7's
/Execute the IOT
/Clear the AC
/and continue with BASIC

7-15

The routine devcom checks
to see that the next item on
the line is a
CR (carriage RETURN). If it is not/
then an error message w±n
given.
This is to prevent illegal user
code from getting throughUCOMING HOME FOR DINNER? The previous
sentence would call the UCOM
processor, but since the next item is
not a carriage RETURN, an error
message would be given.

There are two more routines that can
be used:
DEVCON and SKIPIT.
DEVCON does the same thing as DEVCOM
except that it checks the present
xtem for CR rather than the next item.
SKIPIT skips past the rest
of the line until it finds a CR and
never gives an error message.

To modify the previous example to
take two expressions and multiply them together and then send out
the low 12 order bits to
the device via a 6123, the command
would look like:

UCOM A,B

where A and B are any expressions.
*EXP
JMS
JMS I

GETWD

The coding is:

(MEVAL

/skip past the UCOM element,
/evaluate next element of line.

TEMP

/save away

(MEVAL

/evaluate next element of line.
/enter the FPP package
/multiply A*B.
/exit now.
/fix the AC now.
/clear the AC.
/pick up low word.
/send to the device.

FENTER
FSTA
FEXIT
JMS I
FENTER
FMU
FEXIT
JMS I
CLA
TAD

AC 3

6123
CLA
JMP I

(DEVCON

TEMP
(FIX

/and exit.

Notice that the exit is via DEVCON rather
than DEVCOM.
This is because MEVAL has evaluated the expression
until it found a delimiter.
This must be a comma in the first
place and a CR in the second
Place.
If it is not, then DEVCON gives
an error message.
As stated
previously, the user should look at PLOT
and USE and perhaps a few
other commands to fully appreciate
the UCOM command.
It is possible to put the UCOM
code in field 1 with only some minor
changes.
Since some commands are dispatched
again in field
1,

7-16

place

in 306 a pointer to 7743.

(7743 is the upper core dispatcher.)

Then,

On
in field 1 at 744, place the address of the UCOM command code.
Cross-field
UCOM, BASIC/RT will now go to this location in field 1.

communication is done through a variety of functions, all of which are
mentioned in the internal documentation. The novice user is advised
against putting instructions in field 1 until he has successfully run
functions and commands from field 0. Note that functions can also
reside in field

The user should examine ACCEPT, TIM and possibly

REAL TIME to get a better understanding of cross-field communications.
It is possible to allocate more space in field 1 for functions/commands
LIMIT is the first location of upper
by changing the value of LIMIT.

core used by BASIC/RT.

By changing PLIMIT at 2562 to, say, 2000, about

locations are gained for user functions and commands. Note that
this space is lost from the user program and variable area, but very
often this is worth the sacrifice.
1000

5.2

User Coded Function

As in 8K BASIC, the user can include his own function in BASIC/RT,
The UUF function
to perform a special task.
calling it as UUF (n)
,

is coded in the same manner as described in the 8K BASIC User's Manual ,

except that the suggested locations for the code are 0064 to 0177 in
and many more can
Other smaller areas are available in field
field 1.
be created by deleting some of the mathematical functions.

7-17

CHAPTER 6
ERROR MESSAGES

The computer checks all commands before
executing them.
If for some
reason it cannot execute the command, it reports the
fact by printing
one of the messages in Table 7-2 and the number of
the line in which
the error was found.

Table 7-2

BASIC/RT Error Messages

Message

Explanation

SYNTAX ERROR

The command does not correspond to the
language
syntax.
Common examples of syntax errors are
misspelled commands, unmatched parentheses,
and
other typographical errors.

FUNCTION ERROR

A function was used which was deleted when
the
system was loaded and hence is not available.
A DEF statement will produce this error
if the
DEF capacity has been deleted.

TOO-BIG ERROR

The combination of program and variables
exceeds
the capacity of the computer.
Reducing one or
the other may help.
If the program has undergone
extensive revision, try punching it out,
typing
tvpina
SCRATCH, and reloading!

SUBSCRIPT ERROR

A subscript has been used which is outside
the
bounds defined in the DIM statement.

LINENO ERROR

A GOTO, GOSUB, or IF references a nonexistent
line.

FOR ERROR

FOR loops are nested too deeply.

NEXT ERROR

FOR and NEXT statements are not properly
paired.

GOSUB ERROR

Subroutines are nested too deeply.

RETURN ERROR

GOSUB and RETURN statements are not properly
x
paired.

DATA ERROR

There are no more items in the data line.

ARGUMENT ERROR

A function has been given an illegal argument,
for example, SQR(-l).

RATE ERROR

Clock rate was too fast to service interrupts.

A-D FULL

No more room in the data buffer for
remaininq
items

NO A-D

No input coming from specified analog
c hannel
~~
7-18
'

BOOK

EIGHT

SIMULATOR

CONTENTS

Pa ? e

CHAPTER 1

USING NMR SIMULATOR

INTRODUCTION

8-1

1.1

Storage Requirements

8-1

1.2

Hardware Requirements

8-1

1.3

Use Of NMR SIMULATOR

8-2

CHAPTER 2

DISPLAY MODE COMMANDS

8-5

CHAPTER .3

METHOD OF CALCULATION

8-7

CHAPTER 4

PROGRAM DESCRIPTION

8-9

CHAPTER

TELETYPE AND PLOTTER OUTPUT FOR THE
CALCULATION OF THE NMR SPECTRUM

8-11

LOGICAL FLOW OF NMR SIMULATOR
FOR THE LAB8/E

8-12

CHAPTER 6

TABLES
8-1

DISPLAY MODE COMMANDS

8-i

8-5

CHAPTER

USING NMR SIMULATOR

INTRODUCTION

•

NMR SIMULATOR 1 will

calculi

coupling constats,
are sweep offset,
sweep width and «„
a-ired. punch the
* 1U '
°"
-lY-i.. Whiie the Prineip^le
'" Xn
th
°
Pr
9ram
U1
be in proton NMR, it
°
"
undoubtedly
i, not „„,,,
to proton range
=°n Pli „ g constan ; s
valueSi
shift ='
,
!,,
°" Set Can be anywhere from
10 616 Hs.
lo" 6
to

'"^"^ "
J^t^T^
" ir '*" *"
\

^"rLd

The prinoipel value

speetre which ere so

of hmr simulator
is ln «,„
closely ooupled that

thT

""' "'^

approbations no longer hold
.1
**"
longer lerge compared
* hift ° *"
to
Tcoueu'
°
nst
«"s. This program pro'
vides e method of
=.,«
oirectly i n the
"*
°" Pr09
S
theoretical or analytical
*
t001
oo ' it his 'rest
utility
aid.
as an educational

£
,1 T

^"^

^IrZ^'tTtZT^"
U ^aT

The

„

1-1

Storage Requirements

SIHUMT0R program floes

^^ "^
™'

NMR SIMULATOR occupies

locations 0-7577 in field
locations 20-7245 of
field 1.
1-2

Hardware Requirements

The minimum confi
guration is a
and VC8E display
control.
In order to
-ting options, a

^ased on "NMR SIM"

H. E.

" St° res data in

AS * 33 Telet

^-^^^T^^^—

^^'

j„,=,

Zimmerman, DECUS 8-194
Add^^ 1
and J. w. Cooper,
^Tahedron Let-wft^
8-1

^l?™* ^
"

Hul l *nd
G ' F ~en k el

1.3

USE OF NMR SIMULATOR
as follows:
Load the binary tape of the NMR simulator

Place the tape in the reader.
Set the Switch Register (SR) to 7777.
Press ADDR LOAD.
bit
If the high-speed reader is used, set
Press CLEAR then CONT.

o
o
o
o
o

Start the program as follows:
Set the SR to 0200 (Binary
o
Press LOAD ADDR
o
Press CLEAR and CONT.
o

i».

000 010 000 000)

The Teletype will type:

NMR-SIMULATOR
NUMBER OF SPINS=

Enter any value from 2 through 6. Do NOT follow with
If you answer with 0, the program
a carriage RETURN.
to a display of the contents of
directly
will shift
program
the display array. This is most useful if the
been
has
spectrum
a
after
restarted
was inadvertently
calculated.
4.

The next question is

PUNCHED OUTPUT?
transiIf you wish a binary tape containing all of the
Answering with
tions and their intensities, answer Y.
Since 5- and 6
any other character produces no tape.
time to calcuof
amount
significant
a
take
cases
spin
_if you
option
this
late some time is saved by using
For smaller
width.
sweep
and
offset
the
will be changing
computer
systems, it is usually just as fast to allow the
scratch.
from
spectrum
to recalculate the
,

The program then types

SPECTROMETER FREQ.

(MHZ)

i-e.,
Answer with the frequency of your spectrometer,
This frekey.
RETURN
the
by
followed
60., 90., 100., etc.,
coupling
ana
quency is used only if you enter shifts
If you plan to
constants on the delta or tau scales.
this paramscale,
frequency
the
on
values
these
enter
key
RETURN
eter can be answered by pressing the

8-2

The next two questions are
SWEEP OFFSET =
and

SWEEP WIDTH =

Set these to the closest approximation possible of the
experimental spectrum you wish to simulate. A positive
value for the offset produces an upfield shift of the
spectrum. These two values must be in Hz (cps)
Follow these two values with a RETURN.
.

The program next asks for the shifts

CHEMICAL SHIFTS
H(l) =

These values can be entered according to one of three
possible scales.

The Frequency scale -

This is defined so that TMS is
Hz and the
frequency increases positively in the downfield direction. Just type the decimal number
key.
followed by the RETURN

The Tau scale -

This is defined so that TMS is 10 tau, and
so that tau values decrease in the downf ield
direction. To use this scale, terminate each
shift value with a T instead of a RETURN.
The Delta scale -

This is defined so that TMS is
delta, and
so that delta values increase in the downfield direction.
To use this scale, type D
after each shift value instead of RETURN.

These three scales are illustrated below for a 60 MHz
spectrometer
1

180

120

1
I

i
I

The Frequency scale

10t

8-3

The Tau scale

The Delta scale

The final parameters to be entered
are the couplina
constants

COUPLING CONSTANTS
J(l,2) =
J(l,3) =
.

.etc.

Answer these values with decimal numbers in the
frequency scale (or, if you wish, in the tau or delta
scales)
The subscripts following each J refer to
the numbers of the nuclei as determined by their
order of entry. Thus, J (1,2) is the coupling
constant between H(l) and H(2).
.

If you make a mistake while entering any
decimal number, typing RUBOUT will delete the entire
number.
You can then re-enter the number starting from
its
txrst digit.
The program does not print any character for RUBOUT.

All decimal numbers may also be entered according
to
the exponential or E format.
This is described
in Introduction to Programming 1 and in Programmingfully
Languages •
"

10.

and 6_s P in cases, the computer will type
to assure you that the calculation is
proceeding normally. Be sure that the high
speed
punch is turned on if the PUNCHED OUTPUT?
option was
specified.

*°*t~'
COOKING.

5~
.

Approximate execution times for various size cases
are given below:

Number of

Execution Time
in Seco nds

— P ins
s

<1
2--3
8-•10

3
4

11.

30-90

300-•600

When the calculation is complete,
the Teletype will
type
DONE!

and ring its chime.
The spectrum will then be displayed on the scope of the LAB8/E.

'Digital Equipment Corporation, Introduction
to Programmina
programming
(Maynard, Mass., Sept. 1970), p. 1-35.

2rf
2d

pv.
Ed.

Digital Equipment Corporation, Programming
La
nguagess, 1st Ed.
^"juo^e
Mass., May, 1970), p. 16-11.

8-4

•

;

(«»«
a
(Maynard,

CHAPTER

DISPLAY MODE COMMANDS

While the spectrum is being displayed, the commands in Table 8-1 can
be issued by pressing the indicated key on the Teletype.

Table 8-1

Display Mode Commands

Command

Meaning

Restart the program from the beginning.

Computer asks COUPLING CONSTANTS, etc. Enter
new coupling constants as described in Section 2.0.
Computer asks CHEMICAL SHIFTS, etc. Enter new
chemical shifts as described in Section 2.0.
Computer asks SWEEP OFFSET =, SWEEP WIDTH =.
Enter new offset and width as described in Section
2.0.
After asking for the offset and width values,
the computer asks TAPE?.
If you have punched no
tape, answer N and wait for the program to recalculate the spectrum.
If a tape was punched, answer
Y.
The computer responds with LOAD AND CONT. and
halts.
Load the punched tape into the high-speed
reader and press CONT on the console.

Computer types the transitions and intensities.
It is necessary that a tape be punched in advance
to use thj.s option.
The computer will type

MINIMUM INTENSITY =
For small cases, or for accurate fitting, answer
and RETURN to get all of the transitions.
To list
only the major peaks, the minimum intensity may be
increased.
The computer responds with LOAD AND CONT.
At this point, load the punched tape into the highspeed reader and press CONT on the console. The
computer will respond with ENERGY INTENSITY, and the
listing will be typed out on the Teletype. The display
will return when the listing is complete.

8-5

Table 8-1

Display Mode Commands (Cont.)

Command

Meaning

Plot out the displayed spectrum.
voltages driving the scope display By connectina the
to an x-v plotter
01
har copy °f the theo«ti«l P
ipictrum.
The computer ?types L.L.
and displays a
dot in the lower left hand
corner.
AdjSstthJ plotter bias and sensitivity to
place the pen in Ih?a
thlS
corner.
Then type one of the fo?lowLg?

LINE FEED

The computer types U.R. and
displays
a dot in the upper right
hand corner

RETURN

The display returns to the lower
left.

The plot begins at a rate specified by Knob
0.
if
the plot rate is too fast, interrupt it by
typing
CTRL/0
When the plot is complete, or when it is
interrupted by CTRL/0, the computer types TURN
OFF
PLOTTER. The program waits until RETURN is
typed
before returning to the display.
This is done so
that the rapidly varying voltages of the display
will
not cause the plotter pen to gyrate wildly and
scribble
all over the freshly plotted spectrum.
.

Lorentzian line shape calculation.

The computer types:

LINE WIDTH =

Answer with any decimal number Sallowed by RETURN
program begins to calculate the curve. The lenqth The
of
time required for this calculation varies
from
about 2 to 30 seconds, depending on the
number
of lines and the line width. When the
calculation
is complete, the computer types:
DONE!

and displays the curves on the scope.
the following commands are available

At this time

R - restart the program from the beginning
J - change the coupling constants
- change the chemical shifts
U - shift the display up
D - shift the display down
CTRL/O return to the stick figure display
P
plot out the Lorentzian curves
CTRL/C return to the Monitor (DECtape or Disk
systems only)
S

Thus, one can vary the line width by returning
to the
stick display via CTRL/O and re-entering the
Lorentzian
routine by typing L.

8-6

CHAPTER

METHOD OF CALCULATION
The simplest method of hand calculation of an NMR spectrum requires
that all possible spin functions be written and sorted according to
the number of +1/2 and -1/2 spins in each.

spin -1/2 is represented by

Using the usual notation

and spin +1/2 by a; the order in which

the spins are written represents the nuclei to which they belong.

Thus,

for a three spin system, the spin functions can be written as follows:

aaa

aaB
aBa
Baa

aBB
BaB
BBa

BBB

Since quantum mechanics requires the nuclei to be indistinguishable,
the difference between aaB and aBa and Baa cannot be determined. Thus,
these spin functions must be mixed 2 3 and an equal number of new,
mixed ones created. This is done by the computer in a manner somewhat
'

different from hand calculation.
For each grouping of spin functions, an H-matrix is constructed as follows
:

The diagonal elements, H

are calculated from

L (SiV U

-1/2 for a B spin
+1/2 for an a spin

the frequency of the absorption in Hz

T—
i]

+1/4 if nuclei i and j have the same spin
-1/4 if they have opposite spins

the cou P lin g constant between

where

ii
ID

TijJij)

(1)

and

in Hz

The off-diagonal elements are calculated from the followi ng:
H

= 1/2
'

U. .J.

(2)

J. A. Pople, W. G. Schneider and H. J. Bernstein, High Resolution
Nuclear Magnetic Resonance
(McGraw-Hill, New York, 1959), pp. 103ff.
,

J. D. Roberts, An Introduction to the Analysis of Spin-Spin on High
Resolution Nuclear Magnetic Resonance
(W. A. Benjamin, New York, 1961),
,

pp.
3

37ff.

J. D. Roberts, Notes on Molecular Orbital Calc ulations.
New York, 1962).

o— 7

(W.

A. Beniamin,

where
J
u

is as above

•

1 if the basic functions differ only in the interchange of i and j
if the condition is not satisfied
-0

This matrix is then diagonalized, and the diagonal
elements (or trace)
of the matrix become the energies (eigenvalues) of
the "mixed" spin
functions.
The mathematical method of matrix diagonalization 1
produces
a second matrix, the U matrix, which contains
the final spin eigenvectors.
These are used to calculate the probability of each
transition
between the calculated energy levels.

Transitions occur only between spin functions, the sum of whose
spins,
F
differs
by
1.
Thus,
z
,

n
z

and

(3)

=1 x

must =1 for a transition to occur.

The probability (intensity) of this transition is given
by
I =

C„

A)2

(4)

where
C

the u

element of one spin function's eigenvectors,

the v

element of the other set of eigenvectors.

1 if c

V
A

and c
u
v differ b Y one spin
in all other cases.
'

These matrices are formed and diagonalized, a group at
a time, and the
transitions calculated. A reduced copy of these transitions
is stored
in a 500 word display array, with all transitions
within the specified
range placed in one of these 500 bins.
The Lorentzian line shape calculation is performed by setting
T = 1/Cline width), and applying the
equation of Pople: 2
2
line height
3
~=
g(v)
,

1 +

(2tt

T (v-v
2

P. R. Bevington, Data Reduction and Error An alysis
for the Phvsical
Sciences
(McGraw-Hill, New York, 1969).
2
Pople, et al, pp. 103ff.
,

8-8

CHAPTER 4

Program
1.

Description

The program types its name andsks for the number of
spins.

Enter the number of spins. The program tests for a
value between 2 and 6 and stores it in N.

Set switch if punched output is desired.

Enter spectrometer frequency. The program stores it
for use in calculating Hz from delta or tau values.

The program sets up basic product functions (BPF's).
a.

It calculates NARRY array which contains
relative addresses of the BPF's of different F„ value.
z

It calculates NUSE array. This array actually contains the BPF's. They are stored
in groups each group having F z value one
greater than the preceding group.
,

Enter sweep offset and sweep width.

Enter chemical shifts.

Enter coupling constants.

SPEC array is initialized

10.

This array eventually contains the displayed
spectrum

except for 25 calibraIt is initialized to
tion points which are set to 4j2f0J0g.

First H and U matrices are calculated.

which does not

First H matrix is always a
require diagonalization.

First U matrix is always a 1 X 1 with its only
element equal to 1.

11.

The program sets FLAG which is the negative of the number
of times the loop from 12 to 18 must be executed.

12.

The program sets N2, calculates EN
a.
b.
c.

and UOLD arrays.

It copies the trace of the H matrix into the EN array.
It copies the U matrix into the UOLD array.
It sets N2 equal to the size of the H matrix just copied.

13.

The size, of the next H matrix is calculated and put into Nl.

14.

The program determines which transitions will be allowed
between the BPF's represented by UOLD and those that will
be represented by U. This is stored in the array called
TABLE.
8-9

BOOK

NMR

NINE

SIGNAL AVERAGER

CONTENTS
Page

CHAPTER 1
1.0
1.1
1.2
1.3
1.4
1.5

NMR AVERAGER
GENERAL DESCRIPTION
Hardware
Storage
Loading
Execution Time
Abbreviations and Conventions Used in this
Book

9-1
9-1
9-1
9-1
9-2
9-2

CHAPTER 2

INTRODUCTION TO SIGNAL AVERAGING

9-3

CHAPTER

CONNECTING THE COMPUTER AND THE SPECTROMETER
Homogeneity
Filtering

9-6
9-7
9-7

DETAILED DESCRIPTION OF THE USE OF THE
AVERAGING PROGRAM, NMR AVERAGER

9- 8

CHAPTER 5

ON-LINE AVERAGING

9-11

CHAPTER 6

THE DISPLAY MODE

9-12

CHAPTER

ERROR RECOVERIES

9-18

MAINTENANCE OF NMR AVERAGER FOR THE LAB8/E
Introduction
Storage
Startup
Calculation of the Pause Time
Calculation of the Sweep Time
Sweeping and Sampling
Interrupt
Pre-Normalization Display
Contraction and Expansion
Binary Punch
Decimal Typeout
Normalization
Plotting
N Command
Integration
Analog Integration
Digital Integration
Filter
Calibration

9-20
9-20
9-20
9-21
9-21
9-21
9-22
9-22
9-22
9-22
9-23
9-23
9-23
9-23
9-24
9-24
9-25
9-25
9-26

3.1
3.2

CHAPTER 4

CHAPTER 8
8.1
8.2
8.3
8.4
8.5
8.6
8.7
8.8
8.9
8.10
8.11
8.12
8.13
8.14
8.15
8.16
8.17
8.18
8.19

TABLES
9-1
9-2

Pre-normalization Mode Commands
Post-Normalization Mode Commands

9-i

9-13
9-16

CHAPTER 1
NMR AVERAGER
1.0

GENERAL DESCRIPTION

NMR AVERAGER is a signal averaging program designed for use with any
NMR spectrometer configuration in which the application of a sawtooth
voltage from -5 to +5 volts will cause a linear sweep of the spectrometer frequency (or field)
Since the computer actually does the sweep.

ing, no trigger signal is required.

This leads to a highly reproducible relationship between memory channel and spectrometer frequency,
and an extremely accurate method of calibration. Once the average
is accumulated, NMR AVERAGER can be used for display manipulation,

plotting, calibration, digital and analog integration, differentiation,
curve smoothing and decimal and binary output.
1.

Hardware

The minimum computer configuration includes a LAB8/E, Teletype, and an
X-Y analog plotter.
An NMR spectrometer with an internal lock and auto-

shim is required.
It must have some modification such that the application of a -5 to +5 sweep voltage will cause a frequency sweep in an
adjustable range. One method of doing this is by substituting a digital
frequency synthesizer or voltage controlled oscillator for
band modulation oscillator.

Optional Equipment:

a digital frequency synthesizer,

the side-

frequency counter

and high-speed punch.
1.2

Storage

NMR AVERAGER occupies locations 0-3377 and uses locations 3600-7577 to
store 1024 double precision data points. After data normalization (+E) 1
locations 3600-5577 are used as a display buffer.
in single precision from 5600-7577.

1.3

The data are stored

NMR AVERAGER is loaded using the standard Binary Loader.
address is 0200.

Its starting

(Refer to Chapter 4.)

NMR Averager can not be run under OS/8.
t 8 C0

™ and USed t0

enter Normalized Display Mode and is typed by
ho?,^
yy
holding 5
down the CTRL key and typing E.
9-1

1.4

Execution Time

The execution time is entirely I/O bound.

The sweep time is variable
in one second increments from 1 to 4095 seconds.
The actual timing
is calculated by an integer approximation of the correct number of
crystal clock pulses per sample point. Consequently, while the time
may be slightly slower or faster than that specified, it will be re-

Plotting time is variable by adjusting Knob 3 on the
Digital output is via the Teletype (10 characters per second)
or the high-speed punch (50 characters per second). All display comproducible.

LAB8/E.

mands are executed essentially instantaneously.
1.5

Abbreviations and Conventions Used in this Book

CR/LF is used to mean a carriage return, line feed combination output
by computer.

CTRL/H or +H is used to mean the non-printing character produced by
holding down the CTRL key and simultaneously typing the alphabetic

character H.
ADC and DAC are used to mean analog to digital and digital to analog
converters

9-2

CHAPTER 2
INTRODUCTION TO SIGNAL AVERAGING

The signal averaging method assumes that a very weak, but coherent,
signal is buried under random noise.

Analog to digital converters

are used to convert the varying analog voltage output of the

(ADC's)

spectrometer into digital information. After the passage of a specified time interval, the signal is again examined and digitized. These
signals separated in time by the total sweep time divided by the
,

number of memory channels used, are stored in successive locations in
the computer's memory during the first sweep.
During succeeding sweeps,
the signal is added to the information already contained in the mem-

ory locations.

The noise, if random in nature, averages to zero while
signal
the
to noise ratio grows in magnitude 1
According to theory,
.

signal enhancement is of the order /N, where N is the number of sweeps
taken.
Obviously, then, the greater the number of sweeps, the greater the
signal enhancement.
If there were no other constraints it would seem

apparent that signal averaging should be carried out at the highest
possible sweep rates.
Unfortunately, in NMR spectroscopy, this is
not possible because of the necessity of satisfying conditions of slow
passage 2
Thus, since there is a minimum desirable sweep time, a
.

multiple sampling technique was developed in the NMR AVERAGER to increase the statistical significance of each sweep.
In this method,
each data point is examined several times (up to 20) before the computer goes on to the next channel. These twenty A-D conversions are
summed and placed in memory. This permits the performance of
twenty sweeps in the time of one, without sacrificing resolution to
rapid passage conditions.
The next problem is:

how can one legitimately sample several

times unless he is sure that the exact same frequency is being

examined?

In most signal averaging systems, the computer

and spectrometer sweep simultaneously but independently once they
start.
This means that the spectrometer continues to sweep while the
R. R. Ernst, "Sensitivity Enhancement in Magnetic Resonance", Advance
in Magnetic Resonance 2, J. S. Waugh, Ed., (Academic Press, New York,
1966)
,

J. A. Pople, W. G. Schneider and H. J. Bernstein, High Resolution
Nuclear Magnetic Resonance
(McGraw-Hill, New York, 1959) pps. 37, 38.
,

9-3

computer "stands still".

The answer is, of course, that multiple
sampling techniques (here known as boxcar integration)
cannot be performed accurately under these conditions. This problem
has been overcome, however,
in NMR AVERAGER, since the sweeping is controlled
by
the computer, and it can hold its output voltage at
the exact same
level as long as necessary to obtain all 20 samples,
before incrementing it to the next frequency level and channel.
Thus the multiple
sampling technique is a form of "digital filtering" rather
than of
boxcar integration.
Once the signal data has been accumulated, there are still
additional
problems which must be solved.
It is obviously easy to plot out the
spectrum using a pair of DAC's to drive an X,Y plotter, but
once the
data has been plotted, how can it be calibrated? In the ideal
case,
there are two or more reference lines in the spectrum which make
it
easy to obtain approximate chemical shifts for each spectral
peak.
But in the less-than-ideal, that is to say the usual case,
there may
be no known lines at all, although the frequency values
of the extreme
limits of the plot may be known. The obvious solution, which
some
chemists have always used, is linear interpolation between two
known
points.
However, when these known points are the extreme ends of the
spectrum, this interpolation can be disastrous. Neither the
spectrometer sweep nor the DAC computer sweep is exactly
linear, and in the
case of simultaneous, but independent sweeping,
the two sweep voltages
may even be bowed in different directions.
To see just how nonlinear some of these sweep voltages are, plot two
of them against each
other on an X,Y plotter, and lay a ruler alongside.
However, since NMR AVERAGER actually does the sweeping itself,
there
is always a direct correspondence between memory channel
and sweep
voltage.
This fact is then used in the calibration routine, in

which

the sweep voltage can be advanced manually until the
plotter pen hovers
over the peak of interest, and the corresponding frequency can
be read
from a frequency counter.
To make such a calibration routine completely accurate,
it must be
recognized that the computer spends some of its time putting
out the

DAC voltage and some examining and converting the input
signal.
During the time spent in calculation and conversion,
there is no voltage
coming from the DAC.
This means that the average voltage "seen" at
the output of the DAC depends on the sweep time.
Since the examining
and conversion time is constant and ttie voltage output
time varies with

9-4

the sweep time, it is necessary to take this "wasted" time into account

in the calibration routine.

Thus, in NMR AVERAGER, the amount of time

spent in putting out the calibration voltage is proportional to the
sweep time used to acquire the data.
calibration.

9-5

This ensures a highly accurate

CHAPTER 3
CONNECTING THE COMPUTER AND THE SPECTROMETER

There are undoubtedly numerous methods of interfacing
various spectrometers with the LAB8/E.
The most successful methods have used a
digital frequency synthesizer or voltage controlled
oscillator as
either the modulation frequency source or as the radio
frequency source.
In the case of Varian HA-model instruments the technique
is extremely
simple.
The x-axis scope output of the LAB8/e is connected to both
the scope input and the sweep input of the synthesizer, using
a BNC

or UHF T-connector, and appropriate adaptors. The frequency
output of
the synthesizer is connected to MAN OSC OUT at the
rear of the lock
box, and the MANUAL OSCILLATOR tuning network card
is removed.
The
synthesizer has now replaced the manual oscillator as a
modulation
source.
The spectrometer is then set up as for FIELD SWEEP, so
that
the irradiating channel is the just modified manual
oscillator channel.
The output voltage which sweeps the synthesizer will then
cause a frequency sweep of the sample.
An alternative method, for synthesizers whose frequency
is high enough
for the nuclei being studied, is to use the synthesizer
as an actual
radio frequency source. The actual connections made will
depend on the
characteristics of the probes and detectors used.
For the JEOL C60-HL and PS-100, the input voltage
necessary to sweep
the spectrometer is
to 190 volts on older models, and
to +25 volts
on the newer models.
Interfacing thus requires an inexpensive differential amplifier. The output to JEOL recorders is 10
millivolts full
scale and will thus also need amplification to the ±1
volt range required by the LAB8/E.
For the Perkin Elmer-Hitachi R20B, the input voltage
is -15 to +15
volts.
The output voltage should require no amplification.
Digital
Equipment Corporation will be glad to furnish to its users
all available information on spectrometer interfacing.
Simply contact the
Chemistry Marketing Manager.

9-6

Regardless of the frequency input method, the output remains the same.
Virtually all spectrometers have an analog output jack marked "external
recorder" .
panel.

This is connected to analog input
on the LAB8/E front
If, for some reason, an input other than
is used, it is neces-

sary to deposit the number of this input channel in location 126 after
NMR AVERAGER has been loaded. The input jacks are numbered 0-3 from
top to bottom.
Still another approach is to modify the spectrometer such that the sawtooth output is used to cause the direct sweeping of the spectrometer's
sweep coils. The actual modifications depend to a large degree on the
spectrometer used.
3.

Homogeneity

NMR AVERAGER does not control the homogeneity of the spectrometer.
Consequently, it is necessary that both field and homogeneity drift be
minimized by the spectrometer itself. This requires that the spectrometer have an internal lock and some sort of autoshim circuitry. For
spectrometers without automatic shim control, it is recommended that
signal averaging not be carried out over a period greater than two
hours, unless experiments show that drift is negligible over a longer
period.
3.2

Filtering

The

response time, or amount of "filtering" should be adjusted to a
moderate to large value for the compound being examined. Ernst has
1

reported that it is preferable to perform filtering before signal averaging rather than afterwards, to avoid down-conversion of high
frequency
noise.
If, after the spectrum has been accumulated, it is found
that
the filtering was insufficient, the curve can be smoothed
by use of
the F command.

^rnst,

1-135.

9-7

CHAPTER 4
DETAILED DESCRIPTION OF THE USE OF
THE AVERAGING PROGRAM, NMR AVERAGER

Load the binary tape of the averaging program as follows:
a.

Turn on the power to the computer and Teletype and
switch the Teletype to LINE.

Place the binary tape in the tape reader while the
reader is set to FREE.

Turn the reader to START.
Set switches 6-8 to field in which BIN is loaded.
Set
switches 9-11 to field in which program is to be loaded.
Press EXTD ADDR LOAD.

e.
f

Set the Switch Register to 7777.
This is the position
with all 12 switches having their tips pointing up.
Depress ADDR LOAD. All the lights in the Program
Counter should come on.
Depress CLEAR then CONT. (To use the high-speed reader,
change bit
from 1 to
before pressing CLEAR. Bit f5
is the leftmost bit of the Switch Register.)
The tape
should begin reading in.
If it does not, the Binary
Loader program is not resident and must be reloaded.
Refer to DEC instructions (PDP8/E Small Computer Handbook 1971) on reloading the Binary Loader.
When the tape finishes reading in, it should stop
at the last set of holes and the computer should halt
with all 12 lights in the accumulator out.
(The link
may be either on or off.)
If the accumulator is not
equal to 0, there is a tape-reading error.
To rectify
this, reload the tape, starting at instruction d.
If
this error persists, reload the Binary Loader program.

Start up the averaging program as follows.
a.

Set the Switch Register to 0200 (Binary 000 010 000 000)
Depress ADDR LOAD and then CLEAR and CONT. The
Teletype should type:

NMR-AVERAGER
SWEEPS=
c.

Enter any value between 1 and 4095 for the number
of sweeps.
If a mistake is made, type RUBOUT.
The
computer types ? followed by a carriage return and
line feed (CR/LF)
Enter a new value.
Typing any
noninteger character, or entering a number larger
than 4095 will also cause this error response. You
can tell the computer to accumulate more than 4095
scans by answering Y to REPEAT? (see 2.g).
.

9-8

If the response to this question is
or the RETURN key,
the program immediately shifts to the display
section.
If normalization has been performed
since memory was last erased, the display is
the Post-Normalization Display. If no normalization has been performed, the display is the
Pre-rNormalization one.

This is most useful as an error recovery, when
the program has been inadvertently restarted.

When a correct value has been entered, type the
RETURN key.
d.

After responding with a CR/LF, the computer types:
SWEEP TIME =

Answer with any value less than 4096. Correct
any error as above by typing RUBOUT, and reentering the value. Then type RETURN. If the
value is more than 4095 the computer responds
with a ? and reasks the question.
e.

If the value of the sweep number is acceptable,
the Teletype will follow the CR/LF with

SAMPLES =
This parameter refers to the number of times
each point is sampled during a single sweep.
Respond with any number between 1 and 20. The
purpose of this parameter is to increase the
statistical significance of each sweep, by examining each point several times. Follow the
typed value with RETURN.
f

The program types out

PAUSE TIME =
This is the number of seconds between sweeps.
During this time a voltage is generated corresponding to the left end of the X-axis. This
allows time for the oscillator and/or the recorder pen carriage to return to the left end
before the next sweep. Answer with any integer
between
and 40 and type RETURN.
g.

The Teletype does a CR/LF and types
REPEAT?
If Y is typed, the computer sweeps indefinitely,
typing "END" after e/ery N sweeps. If any other
character is typed, the sweeping is not repeated.

The Teletype does a CR/LF and types
TYPE RETURN
If RETURN is typed, the memory is erased and aver-

aging is

begun.

9-9

If LINE FEED is typed, the memory is not erased,
and additional signals are added to whatever data
has already been accumulated.
This feature allows
change of the sweep parameters without disturbing
the accumulated data, by using the R command.

Typing either RRTURN or LINE FEED causes a CR/LF
on the Teletype, and displays a dot at the left
hand edge of the scope, representing the rate and
position of the sweep. This linear scope sweep is
the actual voltage which is also sweeping the spectrometer.
If the averaging is to take place over
a long period, turn down the scope intensity to
avoid burning a dot on the scope, and turn off the
Teletype.

Refer to Appendix V for a summary of use of the NMR
Averager.

9-10

CHAPTER 5
ON-LINE AVERAGING

During the actual averaging, no other operations can be performed by
the computer, but, at the end of any sweep, the computer can be stopped
and the accumulated data displayed (refer to Chapter

Stop
Commands

Meaning

Causes the sweeping to be interrupted at the end
of the current sweep. After viewing the display
type G to continue or R to restart the program.

(CTRL/H)
Causes an immediate halt. This command is only for use as an error recovery , if
sweeping has been inadvertently started. It
should not be used for an interrupt if the data
acquired so far is to be saved. The reason for
this warning is obvious; the computer adds each
new data point into memory immediately after receiving it, and, if it is interrupted in the middle
of a sweep, part of memory will contain one more
sweep than the rest.

9-11

CHAPTER 6
THE DISPLAY MODE

When all the sweeps are completed, the computer prints
END
and shifts to a display of the accumulated data.
This section of the
program is called the Display Mode.
If sweeping is interrupted either
by the S or CTRL/H commands, the computer shifts to Display Mode with-

out printing anything.
If a large number of sweeps have been taken, the computer displays a
group of disconnected dots.
These can be reconnected, however, by
contracting the display several times.
Display Mode commands are

described in Tables 9-1 and 9-2.

9-12

Table 9-1

PRE-NORMALIZATION MODE COMMANDS

Command

Meaning

tjped?^

thS display by a factor of

expand the display by a factor of

each time C is

each time X is

Go - continue averaging without
erasing the memory.
s P ecifled number of sweeps have
not been
*
ILt
pleted,
sweeping continues until this number is comreached.
If all sweeps have been completed,
an additional N
sweeps are performed.
If the answer to REPEAT' was
Y, sweeping continues indefinitely;
if the Teletype
is turned on, END is typed after
every N sweeps.

CTRL/Z

Zero (erase) the memory and the display
buffer. When
the memory is erased, the number of
sweeps is also
Zer °'
nSW s P ectra can be accumulated using
£hf«J;f
f°
y
the
same parameters.

Restart the program at the beginning.
This is equivalent to pressing HALT, ADDR=0200, and
CLEAR then CONT.
Print the Number of sweeps completed.
Thus, to find out
how many sweeps have been completed,
type S to Stop the
n
Continue *" sweeping Ey
*"
?yp?ng I.'

tiSe, it

?ints

When thS COmputer enters the p lot rou-

LOWER LEFT

^^
^;^^^
f

1315
lQWer left ~hand corner of the
scope
outputs to the scope from the
J"*'
t^
llll/k are -5,-5)
*
volts.
Refer to the LAB8/E U ser's
Sn„i? ffor instructions
Manual
on connecting th e LAB8/E to th e
er
ThS Pl tter should be ^^sted so
that Sis
vi?^a
i
? PSn
laC6S
thS
the lower left-hand corner of
Si plotter
?n^ paper. Now ±ntype
the
the LINE FEED key
'

The computer responds with

UPPER RIGHT
in th Upper right-hand corner
of the
?
scope^^r^?^
scope.
The LAB8/E outputs
.
are now (+5, +5) volts
After

making tentative gain and zeroing
adjustments to 'the
Plotter, type RETURN, which wilAgain
the pen to
By
P±ng the LINE EEED ™* Carriage rItuRN
RETURN keys alternately, the plotter
can be adiusted
ad:USted
to plot a spectrum of any
desired size and shape?

Sng

Lvf^f

9-13

Table 9-1

PRE-NORMALIZATION MODE COMMANDS (Cont.)
Command

Meaning
When the plotter adjustments are complete, retype P to
begin the plot.
The rate of plotting is controlled by
knob 2 on the LAB8/E. Adjust this knob until an appropriate rate is found.
If the plotting is too fast or
too slow during the first trial, adjust knob 2 to vary
the plot rate and type P to restart the plot.

When an appropriate rate is found, lower the pen to make
a hard copy of the spectrum.
When the plot is complete or when it is interrupted by
typing CTRL/0, the message
,

TURN OFF PLOTTER

will be printed on the Teletype. The program waits for
the user to type the RETURN key before returning to the
display.
This is so that the rapidly changing voltages
of the display will not cause the plotter pen to gyrate
wildly.
Calibrate. The calibration routine can be entered
either from the plot subroutine or from the Display
Mode.
When K is typed, a single dot whose coordinates
are controlled by knobs 2 and 3 is displayed on the
scope.
This subroutine provides an extremely accurate
method of calibration of the plotted spectrum.

calibrate a plotted spectrum, simply use the knobs
adjust the plotter pen so it coincides with the peak
interest on the spectrum.
Since the x-axis voltage
connected to the frequency synthesizer, it is only
necessary to connect a frequency counter to the synthesizer to find the frequency of the plotted line.
To
to
of
is

After using this calibration routine several times, it
will become apparent that the speed of response of the
dot on the scope to the control knobs varies with the
sweep time.
This is entirely normal, since the average
output voltage varies somewhat with the sweep time.
Thus , in order to make the calibration as accurate as
possible, the computer calculates how long to spend displaying each point before checking the knobs to see if
they have been changed.
This rate is based directly
on the sweep time currently set in the computer.
It is
therefore necessary to calibrate a spectrum with the
sweep time set to the same value that it had during the
actual averaging process.
Exit from the calibration routine can be accomplished by
typing CTLR/O on the Teletype.
The computer then returns to either the pre-normalization or post normalization display.

9-14

Table 9-1

PRE-NORMALIZATION MODE COMMANDS (Cont.)

Command

Meaning
In order to calibrate a spectrum approximately from the
display alone, use the digital I integration routine
to adjust cursor 2 to be on top of the line of interest.
Then type I again, wait for the numerical integral to be
typed, and type K.
The x-coordinate of the dot has now
been adjusted to exactly the same point on the scope as
the cursor occupied.
Then read the frequency from the
frequency counter.

Binary punch. This command tells the computer to punch
out the contents of its memory in binary form, which
can be read by the Binary Loader.
It includes a load
address of 3600 and a checksum at the end. When B is
typed, the computer asks
H.S. PUNCH?

Turn on the punch to be used, and answer Y for output
on the high-speed punch.
Input other than Y causes
output on the low-speed punch. Type CTRL/O to abort
the punching.
The computer responds with CR/LF and
returns to the Display Mode.
Type out the contents of memory as 1024 double-precision
decimal words. As above, the computer first asks
H.S. PUNCH?
If Y is answered, the output will be in ASCII code
on
the high-speed punch.
This saves time, since the tape
can be listed later off-line, while the computer is
doing something else. Be sure to turn on the punch
before answering Y. As in the binary punch routine,
the printing/punching can be interrupted at any time
by typing CTRL/O and the RETURN key.

CTKL/C

Shifts from the display mode to location 7600 octal.
This command is used with the Disk System Monitor or
the OS/8 Monitor, as a method of direct return to the
Monitor.
If the system does not have one of these
Monitors (available with DECtape or Disk systems) typing CTRL/C will probably wipe out the Binary Loader
and NMR AVERAGER as well.

CTRL/E

Enters the normalized Display Mode.
The double-precision
numbers stored in memory are converted to single precision numbers based on the number of contractions which
have been performed on the display before normalization.
Thus, the display must be on scale and contracted so
that there is no "wraparound" before normalization is
performed.
This command is irreversible: the double
precision pre-normalization mode cannot be reentered
once this command has been issued.
The additional 1024
points of memory gained by this normalization are used
as a display buffer in which data can be modified
without destroying the original spectrum.
Thus, curve smoothing, integration and differentiation are
performed non9-15

Table 9-2

POST-NORMALIZATION MODE COMMANDS
Command

Meaning

CTRL/D

Differentiates the spectrum as displayed.
The digital
approximation of the first derivative is calculated
and displayed.

Shifts the display up by

Shifts the display down by

Filters (or smooths) the data by the use of a threepoint smoothing routine.
This routine operates only
on the display buffer, leaving the original data intact.
This smoothing can be repeated as many times
as desired, and its result plotted by typing P or
integrated by typing I or A.
To return to the unsmoothed data, type CTRL/0 and the RETURN key.

percent.
5

percent.

Care should be taken that the entire display is on
scale and not "wrapped around" before smoothing, since
the computer will attempt to connect lines off scale
at the top of the scope with lines off scale at the
bottom of the scope.
The commands C, X and S also return to the unsmoothed
display, once contracted, expanded or in unmodified
form.
Thus, if the display is off scale when F is
typed, just retype C and F.
The filtered spectrum can
be shifted up or down without destroying it.
The filter routine uses the simple equation
y'

(x)

1
:

= l/4(y(x-l)+2y(x)+y(x+l)

to replace points 2-1023 of the 1024 points.
Thus,
while points 1 and 1024 are included in each calculation, they are never replaced.
Instead, points 1
and 1024 are set equal to points 2 and 1023 after each
smooth

The computer calculates the analog integral so it will
fit on scale, but the display may need vertical shifting.
This can be accomplished, as usual, by the U and
D commands

Once the integral has been calculated, and is on scale,
it can be filtered, plotted or again integrated by typing A.
To return to the unintegrated display, type
CTRL/O and the RETURN key.

"P. R. Bevihgton, Data Reduction and Error Analysis for the Physical
Sciences
(McGraw-Hill, New York, 1969), pp. 255-259.
,

9-16

Table 9-2

POST-NORMALIZATION MODE COMMANDS (Cont.
Command

Meaning
When I is typed, two new lines will appear on the scope
These are two vertical cursors and are controlled by
knobs 2 and 3. Adjust the two cursors to enclose the
area of interest and type I again.
This will produce
the printout of a number on the Teletype.
This
number is proportional to the area surrounded by the
cursors.
In order for this digital integral to be
meaningful, it is, of course, necessary to integrate
two or more areas and compare them.

A rough calibration can be performed on a displayed
spectrum, without plotting it, by typing I, adjusting
cursor 2 to the point of interest, typing I again,
ignoring the number printed and typing K.
Since the
x-position of the calibration dot is also controlled
by knob 2, the position of the dot will be the same as
that of the cursor.
The frequency can then be recorded
from the frequency counter.
The Display Mode is reentered by typing CTRL/0 and the RETURN key.
CTRL/0

Means "whatever you are doing - stop!" and has the
effect of halting several operations performed by commands from the Display Mode
If the computer is punching or typing the contents of
memory in decimal or binary form, the computer does a
CR/LF and returns to Display Mode.
If a smoothed spectrum, an analog integral, a first
derivative, or an analog integral of a smoothed spectrum is being displayed, the computer does a CR/LF and
returns to the unsmoothed, unintegrated display.
If the computer is plotting, it ceases plotting and

prints TURN OFF PLOTTER. Typing the RETURN key shifts
to the display of whatever data was being plotted.
Smoothed or integrated data is not destroyed unless
CTRL/0 is typed again.
If CTRL/0 is typed when normal, unmodified data are
being displayed, the command will have no effect.

Other
Commands

The following commands from the pre-normalization mode
are also active
C
X

CTRL/Z
K
P

CTRL/C
N
R
G

Contract
expand
Zero memory
calibrate
plot the spectrum currently displayed
return to the OS/8 Monitor
Number of sweeps completed
Restart
Go - start averaging - used only after

CTELZi

9-17

CHAPTER

ERROR RECOVERIES

This Chapter lists methods of recovering from
errors made while running the NMR AVERAGER.

7.1

AVERAGING SECTION ERRORS
Error

Recovery

/"number of sweeps

Typed the wrong I samples

l^sweep time

REPEAT?

answered with Y

Type RUBOUT.
Teletype types ?, and
does a CR/LF.
Enter a new value.

Restart at 200, or if already sweeping, interrupt with S or CTRL/H.
Change by restarting with R command.

In answer to TYPE RETURN:

Typed LINE FEED so
memory was not zeroed

Interrupt with CTRL/H then zero
with CTRL/Z and type G.

Typed carriage RETURN
instead of LINE FEED
so memory was accidentally zeroed

None.

Sweeping started inadvertently:
If you want the result

Type

If you do not want the result

to get to Display

Type CTRL/H, CTRL/Z and, when ready,
G.

9-18

7.2

Display Section Errors
Error

Recovery

Typed too many C commands

Type an X for each extra C.

Typed too many X commands

Type a C for each extra X.

G was typed inadvertently

Type S if you want data in memory,
CTRL/H if you do not. Type CTRL/Z
to erase, and G when ready.

R was typed inadvertently

Answer J? to SWEEPS= to return to
the display

P was accidentally typed

Type CTRL/0 and the RETURN key to
return to display.

K was accidentally typed

Type CTRL/0 and the RETURN key to
return to display.

Plot rate is too fast (or too slow) Type CTRL/0, the RETURN key, and
then P to start over or type P to
restart the plot immediately.

Spectrum has been filtered too much Type CTRL/0, the RETURN key, then
E to start filtering again.
I

was accidentally typed

Type

again to return to display.

B or T was accidentally typed

Type CTRL/0 and the RETURN key to
return to display.

CTRL/Z or CTRL/E was accidentally
typed

If binary tape of data has been
punched, reload it with the Binary
Loader.
If not, recovery is not
possible.

9-19

CHAPTER

MAINTENANCE OF NMR AVERAGER FOR THE LAB8/E

8.1

Introduction

NMR AVERAGER is a signal averaging program specifically designed for
cases where the x-axis DAC voltage can be used to sweep the laboratory
instrument, in this case an NMR spectrometer.
It acquires 1024 data
points in double precision, and will plot and display them without
further action.

Other routines are accessible by normalizing the data
to single precision, and using the space gained as a display buffer.
It differs from the usual signal averager in that it can sample each
of the 1024 data points up to 20 times before proceeding.
Given that
the computer is sweeping the spectrometer, the result is that each
of
the 2
samples are taken at the same point in the spectrum rather than
the sequential points obtained when the spectrometer continues to sweep

independently.

Under these conditions this corresponds not to boxcar
averaging, but to a form of digital filtering which is referred to here
as multiple sampling.
To better understand the following paragraphs,
obtain a listing by running source tapes. DEC-LB-U31A-UA, DEC-UB-U32A-UA,
and DEC-LB-U33A-UA available from DEC's Software Distribution Center.
8.2

Storage

NMR AVERAGER occupies locations 0-3377. Locations 3600-7577 are used
to store the double precision data before normalization.
The high and
low order halves of each word are stored separately as two lists, with
the low order words starting at 3600 and the high order words starting
at 5600.
After normalization, locations 3600-5577 are the display buffer, and 5600-7577 contain the normalized data.
8.3

Startup

The program is loaded as usual with the Binary Loader and starts at
0200.
Each message is unpacked using the character unpacking routineThe program asks for the number of sweeps and, if zero, jumps
directly to the pre-normalization display.
It asks for the sweep
time in seconds and rejects zero.
It asks for the number of

samples per point, accepting only values between 1 and 20.
The pause
time is acquired and any value between
and 40 seconds is accepted.
9-20

8.4

Calculation of the Pause Time

If PAUSET is non-zero, it is multiplied by 100
and its negative saved
10
in PAUSET. This value is loaded into the clock buffer register and
the clock is set to tick at 10~ 2 seconds.

number of ticks = PT sec x 100 ticks/sec
8.5

Calculation of the Sweep Time

The minimum time between clock ticks equals the shortest sweep
time divided by the number of points times the number of samples per
point:
= 1

tmin

(

e
^ ?

5
x 10~ seconds/sample

points) (2J3 samples/point)

1)32 4

Thus, the base clock rate is 5 x 10~ 5 seconds per sample.
The actual
number of ticks per sample for a given clock rate is calculated as
follows:

number of ticks/sample =
(T sec/sweep)
,-5
(5 x 10
sec/tick) (1/124 points/sweep)

— ticks/sample

T (2 x 10

}

(ljT)

20T

-%±s

samples/point)

ticks/sample
F

This double precision number is then divided by 1000
The result is
ia
the number of 5 x 10
second ticks per sample, and its negative is
stored as DELAY2. The remainder is the number of 5 x 1j0~ 5 ticks per
sample and its negative is stored in DELAY1. During actual sampling,
.

both counters are incremented to
8.6

between each sample.

Sweeping and Sampling;

The sweep and sample counters are set, the clock started at
100 KHz,
and the clock buffer set to -5.
The x-axis register is loaded with
1000 and displayed until the clock overflows.
The clock overflow is

DLL When DL1 reaches zero, DL2 is tested. If it is nonzero, the counter TCNT, initially set to
-1000^, is incremented. When
both DL1 and DL2 are zero, an A/D conversion is initiated.
counted by

The A/D conversion is taken from the channel specified by CHANEL.
CHANEL is set to zero in the released version of the program,
but it
9-21

can be changed by hand and a new tape dumped using a binary punch
routine

Following the A/D conversion, the value is tested for sign and added
into the double precision storage in the usual manner. The sample
per point counter AGN is incremented and, if zero, the pointers and
x-axis display are advanced using the subroutine PADV. If AGN is not
zero, another sample is taken at the same output voltage.
8.7

Interrupt

The sweep routine does not use the interrupt facility, as it would take
more instructions to detect a clock interrupt, and the sweep rate could
not quite make it at the one second rate. After each set of samples
are taken on a given point, the keyboard flag is tested, and if it is
set, a character is read.
If it is S, the register STOP is set to one.
If the struck character is CTRL/H, the sweeping is interrupted immedi-

ately and the pre-normalization display begun.

Any other character is

ignored and sweeping is resumed.
8. 8

Pre-Normalization Display

During this display the low order words are loaded one-by-one into the
y-axis register and displayed. When the display is contracted, this
causes an arithmetic right shift of each double precision word CCNT
times, where typing C increments CCNT once. The low order result of
this rotation is loaded into the y-axis register.
The x-axis is incremented by one after each display by the routine PADV.
8.9

Contraction and Expansion

The register CCNT is incremented by one each time C is typed and decremented by one each time X is typed. If CCNT becomes negative, the

rotation is to the left instead of to the right.
formed by the routine pointed to by CONTRC.
8.10

Rotations are per-

Binary Punch

The binary punch routine is called by B, and it produces a tape of locations 3600-7577 with a load address of 3600 and a checksum at the
end.
The double precision storage is not reshuffled to the standard
high- low order. This tape can, of course, be reloaded using the

standard Binary Loader.
9-22

Entering the routine asks the question, H.S. PUNCH?, which if answered
by Y sets the register HPUN to 1. The actual punching is done by
BPUNCH which tests the register and uses the high speed punch if HPUN
= 1.
If CTRL/0 is typed,
It also allows interrupts from the keyboard.
the punching halts and the pre-normalization display returns.
8.11 Decimal Typeout

The contents pointed to by HIWD and LOWD are loaded into HISAV and
LOSAV and typed out using the signed decimal point routine Digital- 825-U.Sym.

The numbers are typed eight to a line, separated by

commas.

As in the binary punch routine, either punch can be used.

8.12

Normalization

When CTRL/E is typed, the program counts the number of contractions to
This corresponds to the difference between the
date, and subtracts 2.
10-bit display and the 12-bit memory word. Each double precision word
is then rotated this many times right.

If CCNT-2 is negative, no left

The array starting at 3600 becomes the display buffer and the normalized data is stored at 5600-7577. Commands

rotations are performed.

operating on the data, such as smoothing, integration, differentiation,
and vertical shifting, place their results in the display buffer,
leaving the original data untouched.

The flag EDONE is set to 1, so

that the program can test whether normalization has been performed.
8.13

Plotting

The clock is set to tick at the rate of 100 Hz.

Knob

is read and

biased to a positive number by the addition of 1001 ... This value is
converted to its negative and loaded into the clock buffer register.
The next point in the display buffer (if normalized) or the next

double precision word, appropriately contracted (if not normalized)
is obtained,

loaded into the y register and the display intensified
until the clock overflows.
Then the knob is read again, the next

data point is loaded into the buffer, and the process continues
until all 1024 data points have been displayed.
8.14

N Command

The location LOSAV is loaded with the sum of NSWP and SWEEP.

HISAV is cleared.

Location

NSWP is minus the number of sweeps remaining and

9-23

SWEEP is the total number of sweeps asked for. Their
sum is the number completed. This value is printed by SDPRNT as the
number of sweeps.
8.15

Integration

The same subroutines are used by both A and I to calculate
the actual
integral.
Different driver routines are used, however. The integration itself will be discussed first.

Integration under a curve can be simulated by the rectangular rule,
which states that the area under the curve is approximately equal
to
the sum of the areas of very small rectangles drawn under the
curve,
or

1024

^1024

B(x)dx
Ja

x B(x)

x=l

Integral = Area under the curve

Integral = sum of the areas of the
rectangles

The integration routines assume that the display on the scope
is properly contracted, and that it has been shifted so that it does not
"wrap around" from top to bottom.
8.16

Analog Integration

Analog integration in this case refers to the simulation of the RC
integral of a varying voltage, using digital means.
It proceeds as
follows
a)

The number of times the display has been shifted
vertically from zero is found in UCNT and the buffer rebiased by this amount.
9-24

The sum of all buffer words is calculated and stored
in INTHI-INTLO

INTHI is loaded into the AC and it is rotated left
until the first non-zero bit is found. If no nonzero bit is found, the same testis performed on
This number of rotations is subtracted from
INTLO.
15,-.
The difference is the number of times the sum
of ^ the buffer words must be rotated so that the
final sum will be on scale for a 10-bit DAC display.
The negative of this value is stored in RSET.

The sum of the buffer is again calculated a word at
After each word is added, the sum is copied
a time.
into HITEMP-LOTEMP and this pair of words rotated
right -RSET times. This right-shifted sum is contained entirely in LOTEMP and each time is loaded
back into the buffer location just examined.
Thus, when this operation is complete, the display
buffer contains the running sum of all data points
in the original display - so divided that the total
sum is never greater than the maximum value of the
display.

This sum of data points is equivalent to the rectangular rule integration with the x-distance assumed equal to one.
8.17

Digital Integration

The buffer is displayed once and then Knobs 2 and 3 are read. A cursor is displayed at the x-value represented by each knob and 61 points
above to 61 points below the y-value of the buffer at that point.

When I is typed again, these cursor values

are converted to buffer

addresses

Buffer Address = Knob value + 777 + LOSTRT
and the points between these values summed and typed out using SDPRNT.
8.18

Filter

The equation y' (x) = 1/4 (y(x-l) + 2y(x) + y(x+l)) is applied to points
This is an extremely simple routine since 1/4 and 2
2 through 1023.
The result of this calculacan be represented by bit rotation alone.
tion is placed into the buffer, and the double precision memory is unSince neither x, nor x
touched.
4 are operated on (since there is
102
no x a or x 1(- 25 to use in the calculation), there is some tailing of
the baseline after an extensive number of filterings.

9-25

8.19

Calibration

The calibration routine basically controls the x and y coordinates of
a point by reading Knobs 2 and 3.
The added features are those designed
to make the voltage as close as possible to that produced during the

actual sweeping.

This is accomplished by re-reading the two knobs only

as often as DL1 reaches 0, and all multiple samples have been "taken".

9-26

APPENDIX A
LOADING PROCEDURES 1
A.l

READ IN MODE (RIM) LOADER

The RIM Loader is a program used to load the Binary Loader. The RIM
Loader must be toggled into memory using the switches on the computer
console.

Load the RIM Loader as follows:
Set 00xx in the Switch Register (xx=field in which RIM
is to be loaded) and press EXTD ADDR LOAD

Follow the procedure below for each location listed.
Set the Switch Register (SR) to the location value.
Press ADDR LOAD switch on the computer console.

Set the SR to the appropriate reader instruction value.
Press DEPosit switch. Repeat steps c. and d. until
loading is completed.

c.
d.

Location

ASR33 Reader

7756
7757
7760
7761
7762
7763
7764
7765
7766
7767
7770
7771
7772
7773
7774
7775
7776

6032
6031
5357
6036
7106
7006
7510
5357
70Q6
6031
5367
6034
7420
3776
3376
5356
0000

Refer to Introduction to Programming
mation
.

A-l

High-Speed Reader
6014
6011
5357
6016
7106
7006
7510
5374
7006
6011
5367
6016
7420
3776
3376
5357
0000

Volume

for detailed infor-

A. 2

BINARY (BIN) LOADER

The BIN Loader is a program used
to load the Advanced Averager
program into memory.
The BIN Loader tape is loaded
after the RIM Loader
To load the BIN Loader, follow the
procedure below:

Put the BIN Loader tape in the reader
(always put
cod over reader head, never blank tape). leader/
Set 00xv
I*,
X
Whi 1 RIM WaS l03ded y " fi«l* iJ
'
bin is to"bf
B?N
?
be ?iVSi
loaded) and press EXTD ADDR LOAD key.

•

25h

Set SR to 7756
8

(the starting address of the RIM Loader),

Press ADDR LOAD on the computer console.

If the low-speed reader (ASR33) is used, turn
the Teletype control knob to LINE and set the reader switch
to
START.
If using the high-speed reader, set the
highspeed reader switch to ON.

Depress the CLEAR and CONT switches on the computer
conPe should be 9 in reading in.
If not, check the
d?mt*
RIM Loader and repeat steps a. to d.

After the program is read, depress the HALT switch
on
the computer console.
BIN is now loaded.

A-

APPENDIX B

RUNNING THE LAB8/E PAPERTAPE ADVANCED AVERAGER FROM DISK OR DECTAPE
(4K DISK MONITOR SYSTEM)
To run the LAB8/E Advanced Averaging program from disk or DECtape,

use only tapes DEC-8E-AAPTA-A-PB and DEC-8E-AAAPA-A-PB.

These tapes
contain the necessary changes to enable the programs to run from
disk or DECtape.

DEC-8E-AAPTA-A-PB is analogous to DEC-8E-AAP1A-A-PB in that it contains all four sections of the Advanced Averager (separated by leader/
This tape should be used for operation of the program
trailer)
without control tapes. DEC-8E-AAAPA-A-PB is an expanded version of
.

Section II.

This should be saved separately and used only in con-

junction with control tapes.
B.l

SAVING THE PROGRAM ON THE MASS STORAGE DEVICE

DEC-8E-AAPTH-A-PB
Save each section of the program as system programs.
Consult the
monitor document for your system for the specific procedure.

The locations to be saved in each section and the starting address
for each section are:

Locations to be Saved

Section I
Section II
Section III
Section IV

Starting Address

0-277; 600-7577
6400-7577
6400-7577
0-177; 6400-7577

600
6530
7000
6400

0-177; 6260-7577

6530

DEC - 8 E-AAAPA-A-PB

Section II
NOTE

Save DEC-8E-AAAPA-A-PB as a binary file whose
starting address is specified at load time.
B.2

RUNNING THE ADVANCED AVERAGER FROM DISK OR DECTAPE

To run the Advanced Averager from Disk or DECtape, call Section I
into core.
If the program was saved as a system file, it will start

itself.

Run Section

as described in this manual.

B-l

At the end of

Section I, control will return to monitor. Call
in Section II and
proceed as described. Do this for all sections.
If it is desired to run the program using a control
tape, call
the system loader program and load Section II first;

then load the
control tape from the reader.
Specify a starting address of 6530.
Proceed as described in the manual.

B-2

APPENDIX

SCHMITT TRIGGER

The Schmitt triggers of the LAB8/E a re part of the DK8-ES real time
clock.
They can interact with the clock in many different ways (see
description of the DK8-ES in the LAB8/E User's Handbook DEC-LB-HRZA-D
,

for details)

The Schmitt trigger circuit produces standard, predictable levels as
a result of external activity at their inputs.

(or fire)

if it sees the input go through a .3 volt excursion and then

cross a preset threshold.

switch.
knob.

The trigger will set

The direction of crossing is set with a

The threshold is variable between ±5 volts and is set with a
The .3 volt excursion, threshold crossing, and return crossing

may take a minimum of

microseconds.

pulse through the threshold and back

In other words, a 100 nanosecond

will not fire the trigger.

Re-

cross is not necessary to fire the trigger.

Detailed specifications

for the Schmitt trigger are:

Nominal input voltage range
Input type
Input resistance
Minimum duration input pulse
Maximum permissible input voltage

+ or - 5V.

Hysteresis
Common Mode rejection
Propagation

0.3 V
35 db
600 ns

Input threshold
Source

Variable between + and - 5V
+, -, or line frequency
zero crossing.

C-l

Differential
50K
2ys
+ or - 50V

APPENDIX D
EQUATIONS FOR STATISTICAL CALCULATIONS

The 95 percent confidence limits are the following:
C. =2S./N
l

1/2

where N is the number of sweeps and S.
2

S.={l/(N-l)[l

where A.

is the value of the

)

(standard deviation) is
'

-(Z A..)

(l/N)]}

point of the

1/2

sweep.

The trend function is:
N
T
i

(N-l)S. 2

The reference for these equations is E.L. Crow, F.A. Davis,
M.W. Maxfield, Statistics Manual Dover Publications Inc. N.Y., N.Y.

D-l

APPENDIX E
DATA POINT STRUCTURE

SECTION

II
Representation

Sum
(of jth point

for

N sweeps)

Low

=Z A..M

Sum of Squares

High

Low
Mid

(of jth point

for

Word Order

N sweeps)

High

Value of jth.
point in (i~l)tb.

sweep

A,.

, ..

where A_. =

Low

(i-l)l

Sum of Squares
of Successive

Differences
(of jth, point for

v|,(vW

Low
Mid
High

N sweeps)

Values are stored in either single, double, or triple precision
as indicated by the Word Order.

E-l

Sections III and IV

Representation

Word Order

AVERAGE
(of jth point

for

N sweeps)

High

a./N
I

N= Number of sweeps
C. =2S./N

LIMITS
(of jth point

for

N sweeps)

where,
High

S.={l/(N-I)
(S. is

TREND

1/2

CONFIDENCE

[p\-(a. /N)]

1/2

N sweeps)

Low

}

standard deviation)

=K y ./(N-l)S.

(of jth point

for

Low

Low
where K = 1000 (output scale factor)

The values of AVERAGE, CONFIDENCE LIMITS, and TREND are stored
in
contiguous locations for each data point. Locations are allotted
only for those functions requested in Section I (compiler)
If
CONFIDENCE LIMITS and TREND were not requested, a. would be
.

immediately followed by a
;that is, all values pertaining to one
j+1
point are kept together in the format shown. These would be followed
by the values for the next data point.
locations between data points.

There are no unused core

Values are stored in either single or double precision as
indicated by the Word Order.

E-2

APPENDIX F
HOW TO ACCESS DATA
Core Map
For a generalized core map of the LAB8/E Averager, see Appendix G
(Memory Organization)
The structure pictured is for the Advanced
.

Averager.

A and B sweeps refer to high and low resolution capability
of the Advanced Averager.
The locations of the areas can be found in certain locations on page
zero.
Below is a table stating the name of the area and where its

location can be found.

Example:

lists is found in location 23.

The starting location -1 of the job
+ l should be read, "the

C(23)

contents of location 23 plus 1."
Area
Basic Averager
Job List
C(23) + 1

Advanced Averager
C(23) + 1

MPLXR List
Chan Display List

C(24) +

C(24) + 1

ADC Buffers A Sweep

C(47) + 1

C(56) +

B sweep

C(54)

Calculation (storage)
Buffers

+ 1

the location of these is obtained from
job list (explained below)

The LAB8/E Averaging Programs utilize what is called a "job list".
The job list contains seven words which define the average. Each
average has a job list particular to itself. If three averages
are taken, three separate job lists are created by the compiler.

They would be stored sequentially, starting at address
At the end of the last job list, there is a word containing
230g.
zero to denote the last job list.
This list is used whenever the
data is referenced by any part of the program.
230,

Word

Bit

for high resolution sweep; 1 for
]ov:

Bits 1-5:

Bit 6:
Bits 7-11:

Word

Bits 0-3:
Bits 4-11:

Word

arralog channel # pertaining to this
average
always 1 to prevent possiblity of zero word
position of this average in sampling
list

job type (Avg. only = 0400); Conf.
lim.= 1000; (TRN = 1400)
sort code

Minus the number of data points in
the data block containing this
average (all the numbers pertaining
to an average are stored in a data

F-l

Word

CDF (change data field) instruction for
the data block.

Word

address-1 of this block

Word

defines the scaling of the display
of this average

Word

position on a display scope where
x =
and y =
for the averager

block). Each average has its own data
block (does not include points that are
in block extension)

cont.

Of interest here is Word

(location

This gives the location

234,.) .

-1 of the data block which pertains to this average and its associated

numbers.

Contents of Word

+ 1 is the address of this data block.

The first location contains the number of sweeps taken.

The follow-

ing locations contain the sum (in double precision) , the confidence

limits (in triple precision)

and the trend (in triple precision

for each point.

+ 1 location)

A data block during Section III would

look something like the following:
c

(word 5) = 1263

1263\

start of block

6 \ hi
"7

of sweeps)
° rder of sum of a11 point l's

(

1270 \midl order of the sura of the squares for
a11 P oint l's
1 \hi J
value
of last point:
2\
necessary in trend calculation
^
4 \ mid f
5 \ hi J
6 \lo
I
7 \hi

order of sums necessary for trend calculation
order of sum of all point 2's

etc.

In Sections III and IV the arrangement of the data changes.

From the
raw statistics gathered during the averaging section the final
statistics are calculated. Sections III and IV contain the data in
the following format:

Starting address of the data block is the same
1263\
4 N
# of sweeps
6 \ lo
"7

V»

}

order of sum of point

1270 \ lo)
1 \

2N hil
3 \ lo >

order of confidence limits for point
1000 x trend for point 1 x

order of sum of point

Note the reversal of the order.

F-2

APPENDIX G
MEMORY ORGANIZATION

Memory Organization

MEMORY ORGANIZATION
SECT

SECT

BASIC

SECT

SECI

LOADERS

SECTION

SECTION 3

ADVANCED
AVERAGING

PREPARES CTUL
TAPE OR IMAGE

BASIC

AVG.

STATISTICS

ADVANCED

FINALS

OUTPUT
PACKAGE

RAW STATS

FLT PNT SUBS

FLTPNT

COMMON SUBROUTINES
1

HI-USAGE
SUBROUTINES

DATA BLOCK STRUCTURE

PARAMETER
SETUP

DISPLAY

FORMATS
1000

BUFFER -B
INITIALIZE

BUFFER - A

600

CD - LISTS

BUFFER

MX - LISTS

CD LIST

JOB LIST

230

200

PARAMETERS, LINKS, CONSTANTS, TEMP STORE

G-l

Jl: A/Ml),

CHAN' (5), l(l).ORD(5)

J2: TYPEI4).

MEMORY ORGANIZATION

J3: LI: -'

SORT CODE (8)

OF POINTS IN BIK

Jl: L2: CDF N TO BLK

LAB-S AVERAGER

J5: L3:

LOC-1 FOR lit PNT

J6: Dl:

AX (8), Y-SCALEI4)

J7: D2: X0(6),Y0(6)

INTERRUPT BRANCH

2nd CHAN OF A

PARAMETERS

3rd CHAN OF A

BASIC LINKAGES
nth

TEMP STORAGE

CHAN OF A

4000 - 1st CHAN OF A
1st

CHAN OF B

2nd CHAN OF B

nth

CHAN OF B

4000 * lit CHAN OF A

CHAN - A SWEEP
CHAN - B SWEEP
Hi CHAN OF A *4000

ADC BUFR
SWEEP - A

ADC BUFR

CHAN DISPLAY LISTS

SWEEP B

CALCULATION
BUFFERS

SECTION

PROGRAMMING

OF POINTS IN NEXT BLK

DF TO NEXT BLK: CDF N

BOOTSTRAP-SYSIO

LOC -1 FOR 1st LOC OF NEXT BLK

CALCULATION
BUFFERS

G-2

APPENDIX H

ATTACHMENT OF AN ANALOG X-Y PLOTTER
Plotting of a displayed waveform may be easily accomplished by the use
Since the display is itself an analog device,
of an analog plotter.
the information carrying signal is available at the X and Y input con-

nections of the scope.

The plotter should have an effective analog

range of ±5 Volts.

To attach the plotter to the scope, simply run two wires from the

plotter to the scope

inputs for each of the X and Y connections and

their respective grounds.

The wires need not be shielded.

Now,

whatever voltage is available to the scope is also available to the
plotter. To plot a waveform, one slow display cycle should be gone
through
CAUTION

While the display is running at its
normal speed, the plotter should be
disconnected or turned off because
the pen will shake furiously attempting to follow the display.

H-l

APPENDIX

DATA BLOCK STRUCTURE
To provide for multiple field machines, the LAB8/E Averager saves data
in linked blocks.
In general, LAB8/E will end one block and link to

another whenever it comes to an end of field.
Section I

Linking is done in

An example is shown for N sweeps and M data points broken into two data
blocks.
In the example, the block bridges Field
to Field 1.
Block
-# of points in 1st block:

CDF to 1st block:

-L

BLK1

CDF J?

in Data Field

of sweeps

1st data point

Location -1 of 1st word
Block -1

2nd data point
3rd data point

Words J3-J5 of Job
List Entry for this
Job.

Lth data point
-# of points in block
extension: L-M

CDF to block extension:
CDF 10

Location -1 of 1st word
of block extension:
BLK EXT-1
BLK1 EXTENSION

L+l data point

.Mth data point

-# in next block:

BLOCK

EXTENSION

in Data Field 1
i-i

Each data point consists of information pertaining
to the mean and, if
requested, standard deviation and trend for the signal at that
point
in time.

The number of locations used for each of these functions
for each data point and the contents of each location
depends on
which section of the LAB8/E Averager is in core.

NOTE
If the data block does not have an extension,
there is a zero word for termination. This
terminator must appear at the logical end of
every data block or data block extension.

1-2

INDEX

Abbreviations, 9-2
Axis inversion, 5-12
Abnormal responses, 1-13
Axis on or off, 3-12
Abstract, 9-1
ACCEPT function, 7-10, 7-12, 7-17 Backslash (\)
4-9
Accessing data, 1-43
Bin
#0, 3-15
Accumulated sum clearing, 2-5
Bandwidth of noise, 4-4
Accumulator (AC) , 5-2
Baseline, 5-21
Accumulator lights, 3-1
Baseline tailing, 9-23
Acquisition, 5-30
Basic and advanced average CONVERT
A/D channel sampling, 7-10
output, 6-9
A/D commands, 7-10
Basic average CONVERT output, 6-9
ADB function, 7-13, 7-14
Basic time unit, 3-8
ADC (analog to digital converter) ,BASIC/RT compatibility, 5-2
9-2, 9-3
Batch mode, 4-10, 4-11, 4-17
ADC function, 7-8, 7-10
normalization, 4-16
ADD command, 5-18
Best straight line, 5-16
Address of histogram, M-l
Bias controls on X-Y recorder, 1-28
Advanced average CONVERT output, Bin width, 3-7
6-9
doubled, 3-13
Algorithm PST and Latency collec- Binary
,

tion, 3-9, 3-10

dump, 4-13

ALTER command, 5-11, 5-27
loader, 4-7
Alter parameters, 3-11
Binary (BIN) Loader, 1-4, 1-21,
5-2
Altering displayed spectrum, 5-15
Amplifier, differential, 9-6
program, 2-3
Analog
Binary dump, 3-16, 3-17
channel, 3-6
Binary loader reload, 9-8
plotter calibration, 3-15
Binary numbers, 1-12
signal, 3-6
Binary output format, 3-M-l
Analog integral calculation, 9-17 Binary punch, 9-15, 9-22
Analog integration, 9-22
Block diagram of program 4-5
Analog plotter, 4-1
Boxcar averaging, 9-20
calibration, 4-12
boxcar integration, 9-4
Analog X-Y recorder output, 1-24 Buffer capacity, 7-12
Analysis time period, 3-10
Applications, DAQUAN, 5-1
C (change contents of location Applications for correlation, 4-1
MODIFY routine)
5-19
Area under curve, 1-33
C (check - AVERAGER routine)
5-7
ASCII code, 9-15
C (contract display) 2-6, 2-7
ASCII format, 5-6
CALCULATE command, 5-9, 5-26
ASR33 Teletype output, 1-32
Calculation of
Assembly, 5-26
analog integral, 9-17
Auto-correlation, 4-1
maximum number of data points,
examples, 4-3, 4-4
1-14
mathematical definition, 4-1
pause time, 9-21
Automatic shim control, 9-7
sweep time, 9-21
Average
timing, 9-2
definition, 1-5
Calculation methods, 8-7
display, 1-35
Calculation, Tau-Delta, 8-12
plotting, 1-29
Calculation from Teletype keyrespecification, 1-14
board, 5-9
AVERAGE command, 5-6, 5-35
Calculations for crossAverage termination, 2-5
correlation, 4-6
Average, typing of completed, 2-6 Calibrate X-Y analog plotter, 3-15
AVERAGER display routines, 5-7
Calibration, 9-26
Averager MS, 2-11
routine, 9-3, 9-14
Averaging
routine exit, 9-15
on line, 9-11
X-Y analog plotter, 4-12
program
CCNT incrementation, 9-22
load, 9-8
Change core location, 5-19
,

startup, 9-8

INDEX-

Channel combination, 5-6
Channel sampling, 7-10
5-7
Check (C)
Checksum, 3-17, 3-M-l, 5-2
Chemical shifts, 8-3, 8-4
CLC function, 7-10
CLEAR command, 7-5
Clearing accumulated sum, 2-5
Clock, 9-21
buffer request, 9-21
interrupt facility, 9-22
overflow, 9-21
tick rate, 9-23
Clock commands, 7-6
Clock enable register functions,
,

Configuration, 5-2
Configuration, hardware, 9-1
Connecting computer and spectometer,
9-6

Connecting leads, 2-3
Connections, input, 4-10
Contingency condition, 1-12, 1-13
register, 1-13
Contraction of display, 9-13, 9-22
Contraction of viewing scale, 2-5
Control character summary, on-line
section, 1-22
Control tape, 2-1, 2-11
punching, 1-19
reading, 1-19
Conventions, writing, 9-2
Conversion program, 6-1
Conversion switches, 6-2
CONVERT program, 6-1
CONVERT, save locations, 6-1
Coordinates of point, 7-3
COPY command, 5-12
Core exceeded, 3-9
Core and image dump, 4-14
Core image dump, 3-17
Core location changes, 5-19
Core map, 1-44
Correct (C) contents of location,

7-9
CLS function, 7-8, 7-10
CNV detection, 1-8
Code, ASCII, 9-15
Coefficient output, 4-16
Collection alogrithm, PST and
latency, 3-10, 3-11
TIH, 3-10
Colon usage, 5-6
Combining channels, 5-6
Commands, 3-10 - 17, 5-5 - 22, 6- 1
9-13 - 17
CLEAR, 7-5
5-19
clock, 7-6
data collection and output,
Correction factor, 2-11, 2-12
1-26
4-10, 4-11, 4-12 - 17
definition, 7-14
Correlation
DELAY, 7-3
applications, 4-1
END, 7-5, 7-6
CONVERT output, 6-10
output only, 4-13 - 17
examples, 4-3
PLOT, 7-2, 7-6
mass storage for OS/8, 4-18
REAL TIME, 7-10
properties, 4-3
REJECT, 7-11
theory, 4-1
response, 7-14
Counter incrementation, 9-22
scope, 7-1
Coupling constants, 8-4, 8-5
SCRATCH, 7-5, 7-6
Cross field communication, 7-16
SET CLOCK, 7-9
Cross-correlation
SET RATE, 7-9
calculations, 4-6
summary, 5-31
examples, 4-5
USE, 7-6, 7-11
mathematical definition, 4-2
user, 7-12
CR/LF, 9-2
WAIT, 7-14
CRT display input signal, 2-5
WAITC, 7-10
CTRL commands, 4-11 - 17
Comments, 5-3
CTRL key, 8-6
Common interval, 1-9
CTRL keys
Compatibility, BASIC/RT, 1
Q, 2-5
Completed average typing, 2-6
2-7
R,
Compression of. data array, 5-19
2-7
Z,
Computation of
CTRL/A, 3-7, 3-8, 5-3
derivative curve, 5-16
CTRL/C, 3-7, 3-8, 5-26
standard deviation, 2-9
CTRL/H (+H), 9-2
statistics, 2-9
CTRL/S, 3-8
trend, 1-24
Cursors, 1-25, 2-11, 4-12, 4-14
Confidence limits, 1-15, 1-21,
5-14
1-23, 1-25
CURSORS command, 5-13
calculation, 1-32
curves, 2-10
plotting, 1-29
trend computation, 2-9

INDEX-

Data
array compression, 5-19
display, 5-7
format, 5-28
input, 5-8
rate, 5-1
reload, 3-16, 3-17
retained in buffers, 3-17
scaling, 5-13
typeout, 3-14
Data accessing, 1-44
Data block structure, 1-4

Disk dumping, 3-17
Disk monitor system loading, 4-7
Disk operation, 1-38
Disk system monitor, 9-15
Display
buffer, 9-1
contraction, 9-13

9-10, 9-14
expansion, 9-13
histogram, 3-1
list, 3-M-l
moving, 3-12
Data collection
shift, 9-16
commands, 3-11, 3-12, 3-13
Display
array, 8-2
restart, 3-17
7-4
Display
buffer,
start, 3-9
Display of data, 5-7
3-12
stop,
Display high/standard resolution,
trend, 1-20, 1-21
1-10
Data collection and output comDisplay
integrity, 4-17
mands, 4-11, 4-12, 4-19
Display
mode
commands, 8-4, 9-13 - 17
Data normalization, 9-1
Display program, 9-9
Data point structure
Display quality, scope, 1-21
Section II, 1-42
Display routines, AVERAGER, 5-7
Section III, 1-43
Display on scope, 8-4, 8-5
Section IV, 1-43
Display time, 7-1
1-7
Data points,
Dot display, 9-10, 9-14
calculating maximum number, 1-14
disconnected 9-13
Data validity, 1-27
Dots
on scope, (see cursors)
Decimal numbers, 8-4
Double
precision, 9-1, 9-16, 9-20,
Decimal typeout, 9-23
9-23
Decrease (D) amplitude, 5-7
Double precision format, 5-6
DECtape dumping, 3-17
Drawing grid X-Y chart paper, 1-29
DEC tape operation, 1-38
Dump histogram, 3-17
Define average, 1-5
Duration of
Delay
experiment, 3-3
negative, 1-7, 1-9, 1-10
run specification 3-2
positive, 1-8, 1-9
zero, 1-10
EAE, 5-26,5-30
DELAY command, 7-3
END command, 7-3, 7-4
2-5
Delay, sampling,
Epoch, 1-9, 1-10, 3-2, 3-8
DELAYl, DELAY2, 9-21
definition, 1-6
Deleting input, 5-3
length, 1-7
Delta scale, 8-3, 8-4
second, 1-8
Derivative curve computation ,5 -16
Equations,
Gaussian curve, 5-1
Derivatives, multi-depth, 5-17
Equations
for
Statistical CalculaDescription, general, 1-1
tions, 1-40
Description (stepwise) of program
Equipment, 9-1
8-8
Erase memory, 9-13
DEVCOM routine, 7-16
Erasing scope, 7-3
DEVCON routine, 7-16
Error, 6-2
Diagonal elements, 8-8
recoveries, 9-18
Differential amplifier, 9-6
in tape reading, 9-8
DIFFERENTIATE routine, 5-16
Error condition, 3-11
Differentiate spectrum, 9-16
on readin, M-l
Differentiation, 9-23
Error correction, 1-6
Digital
Error message, 4-8, 4-19
filtering, 9-4
Error messages, 7-18
frequency synthesizer, 9-6
RATE ERROR, 7-18
integration, 9-25
TOO BIG, 7-18
output, 9-2
Error, readin (see Read- in Error)
typeout ASR33 Teletype, 1-32
Dimensions of scope, 7-1

dot,

INDEX-

Error, read-in (see Read-in Error)Grid (cont'd)
loading, 1-5
drawing, X-Y chart paper, 1-29
Section IV, 1-24
plotting, 1-30
Error, typing, 8-3
range, 5-14
Errors
Gross activity of neuron, 3-9
command string, 5-28
Group I commands, 3-11 - 13
quantization, 5-17
Group II commands, 3-14 - 17
read-in, 5-2
Example, 5-23
H (halt)
2-6
to 5-25
HALT switch, 5-4
auto-correlation, 4-3, 4-4, 4-18Halting operations, 9-17
cross-correlation, 4-5, 5-18
Hard copy, 5-17
Exclamation mark usage, 5-3
of spectrum, 9-14
Execution time, 9-2
Hardware, 5-2, 9-1
Exit from calibration routine,
Hardware requirements, 1-2, 2-1,
9-15
3-1, 4-7, 8-1
EXP function, 7-15
High limit data point, 2-6
Expand histogram, 3-12
High resolution epoch, 1-9, 1-10,
Expand waveforms (X) , 1-25
1-11
Expansion of
High-speed punch, 5-21, 9-15
curves, 2-11
output, 9-2
display, 9-13, 9-22
High-speed reader, 1-4, 2-3, 5-2,
viewing scale, 2-5
9-8
Expansion factor, 4-11
Histogram
Exponential format (E)
8-3
CONVERT output, 6-8
External recorder connection, 9-7
display, 3-1
expansion, 3-12
F command, 9-7
Homogeneity drift, 9-7
Field drift, 9-7
Hyphen (-) symbol, 1-7
File types, 6-4
Filtering, 9-3, 9-7, 9-25
Increment counter, 9-22
routine, 9-16
Initialization,
Finding area under curve, 1-33
of trigger, 3-6
Flattening of peak, 1-21
system, 1-4, 4-8
Floating point package, 5-3
Input, 8-1
modification, 8-13
Input connections, 4-10
Frequency, 8-12
Input data, 5-1, 5-3, 5-8
counter, 9-4
deletion, 5-3
output, 9-6
formats, 6-5
scale, 8-3, 8-4
Input signal, CRT display, 2-5
source, 9-6
Integers, single precision, 5-3
Functions
Integral calculation, analog, 9-17
ACCEPT, 7-11
INTEGRATE command, 5-14
ADB, 7-12, 7-13
Integration, 5-15, 9-23, 9-24
CLC, 7-10
analog, 9-24
CLS, 7-10
digital, 9-25
EXP, 7-15
Intensity, 8-4, 8-8
REJECT, 7-11
Interpolation, 9-4
TIM, 7-8, 7-10
Interpreting X-Y recorder readout,
TST, 7-14
1-31
UUF, 7-16
Interrupt facility, 9-22
Intervals, sync pulse, 1-16
Gain controls, X-Y recorder, 1-29 Inversion of axis, 5-12, 5-15, 5-16
GATE, 5-21
Gaussian fit, 5-10
Job List, 1-44
Generating control tape, 1-20
JEOL recorder, 9-7
GETC statement, 7-14
Keyboard commands, Section IV, 1-34
Graphs, 7-1
Knob positions, 2-3
Grid
calibration routine, 1-27
,

INDEX-

Latency, 1-34, 3-2
Leads, connecting, 2-3,
3-3

Leader/trailer code, 3-16, 3-M'-1
Leader/trailer tape, 1-4
Least sweep interval, 1-6
Length of signal epoch, 1-7

Nerve cell fatigue, 3-3
Neuron gross activity, 3-9
Noise bandwidth, 4-4
Normal mode, 4-9, 4-10
Normalization, 4-16, 9-23
Normalized display mode, 9-16

NSWP, 9-23
LIMIT, 7-17
Number of points, 4-8
LINE FEED, 2-5, 3-7
Number of sweeps, 2-5, 4-12, 9-13
Numbers, octal/binary, 1-12
Linear interpolation, 9-4
Numeric value format, 5-3
Listing off-line, 9-15
Loader, binary, 4-7
Octal numbers, 1-12
Loading, 8-2, 8-12, 9-1, 9-8
Off line listing, 9-15
basic averager MS, 2-11
On line averaging, 9-11
binary (BIN) loader, 1-36
LAB8/E advanced averager, 1- 36 On-line signal averaging, confidence limits, trend data
program, 1-4
collection, 1-20
read- in mode (RIM) loader, 1 -35
Loading in disk monitor system ,4-70pen location, 5-19
Loading procedures, 1-35, 5-2
Operating procedure flow diagram,
2-7, 2-8
DAQUAN MS, 5-26
DAFFT, 5-30
Operation principle, 4-5
PAFFT, 5-30
Options, 8-1
Loading program, 2-3, 4-7, 3-5
Oscillator, 9-6, 9-9
Oscilloscope, 3-1
Loading and starting, 6-1
Locations, storage, 9-1
Oscilloscope display, 2-10
Oscilloscope output, 4-12
Lorentzian fit, 5-10
Lorentzian line shape, 8-6, 8-8
Output, 5-19, 8-1, 8-8, 8-11
9-7, 9-15
curve calculation, 8-13
Low limit data point, 2-6
analog X-Y recorder, 1-32
Low-speed reader, 1-4, 2-3
ASR33 Teletype, 1-32
coefficient, 4-16
Mass storage device, 1-36 - 41
commands, 4-11 - 18
Mass storage for OS/8, 5-26, 5 -29
mode, 4-12
Mathematical definitions
oscilloscope, 4-12
auto-correlation, 4-1
scope photo, 1-24
cross-correlation, 4-2
section IV, 1-24
Maximum number of points, 5-1
Output data formats, 6-6
Memory organization, 1-47
OutPut Mode (OPM) , 3-13
Methods of calculation, 8-7
Overflow, 1-33, 3-9, 3-10
Minimum time, 3-3, 3-15
Minus sign (-) symbol, 1-7
P (proceed) , 5-7
Mode
PAFFT, 5-29
batch, 4-17
example, 5-37
output, 4-12
overlay, 5-35
plot, 4-13
PAL III symbolic assembler, 2-4
view switching, 4-12
PAPER TAPE command, 5-8
MODIFY command, 5-19, 5-30
Parameter
Modifying parameters, 5-11
altering, 3-12
Modulation frequency source, 9 -6
modification, 5-11
Monitors, disk system or PS/8, 9-15 redefinition, 1-15, 3-7, 3-18
Move display, 3-12
setup, PST/latency £ 3-6
Moving pen, 2-5, 2-6
Pause averaging, 2-5
Multi-depth derivatives, 5-17
Pause time, 9-20, 9-21
Multiple sampling, 9-4, 9-20
PAUSET, 9-21
Multiplexer, 2-9
Peak flattening, 1-21
Multiply amplitude (M) , 5-7
Peak index number, 5-11
MULTIPLY command, 5-12
PEAK report, 5-20
Pen, 9-9
N command, 9-23
movement, 2-5, 2-6
N (inspect next location)
5-l£
shaking, 1-30, 3-2-11, 5-17, 5-18,
Negative delay, 1-8, 1-9, 1-10
9-14
,

INDEX-

Perimeter dimensions of scope, 7-1
Perkin Elmer-Hitachi R2-B, 9-6
Phone jack connector number, 1-6
PLIM1T, 7-17
PLOT, 5-17
PLOT command, 7-2, 7-6
Plot, histogram on recorder, 3-5,
3-15
Plot mode, 2-5
Plot routine, 8-13, 9-13
Plot termination, 2-6

Plotter
calibration, 4-12
Plotter adjustment, 9-14
Plotter, attachment of analog X-Y,
1-30
Plotter calibration routine, 1-28
Plotter output, 8-11
Plotter pen, 9-4
Plotting, 7-2, 8-6, 9-23
Plotting average, confidence limit
limits, and trend, 1-20
Points, maximum number, 5-1
Points, number of, 4-8
Positive delay, 1-9, 1-10
Post normalization
commands, 9-16
display, 9-9
Power command, 5-33
POWER key, 2-3
Pre-normalization
commands, 9-3
display, 9-9, 9-20, 9-22
Pre-stimulus response, 1-8
Program
block diagram, 4-5
loading, 1-4, 1-35, 2-3, 4-7
response, 4-17
start questions, 4-5
starting address, 1-36
startup, 4-6
tapes, 1-2, 2-1

Program description (stepwise)
8-9

Program loading, 3-5
Program restart, 9-13
Properties of correlation, 4-3,
4-4

Pulse, frequency distribution, 3-8
Punch, binary, 9-15
Punched output, 8-2, 8-4
Punched tape options, 8-1
Punching control tape, 1-20

Quantization error, 5-17
Question mark (?) usage, 3-6, 5-18
Questions to define average, 1-5
Questions to establish parameters,

R (transfer control) , 5-19
Radio frequency source, 9-6
Random noise, 9-2
RATE ERROR error message, 7-6
Rate, sampling, 2-5
Rate of sampling, 4-8, 4-9
Read-in
error,
2-4, 3-5, 5-2
loading,
1-5, 1-37, 3-5, 5-2
Mode (RIM) loader, 1-4, 1-3 5, 2-3
section IV,
1-24
Reader, high or low speed, 1-4, 2-3
Reader punch, 8-1
Reading control tape, 1-20
Readout, interpreting X-Y
recorder, 1-31
REAL TIME command, 7-10, 7-12
REAL TIME statement, 7-17
Recalibration, Schmitt trigger, 3-6
Recorder pen, 9-9
Redefine parameters, 3-18
Refreshing scope, 7-2
Reinitialization, 2-6
REJECT command, 7-10, 7-12
Relational operators, 7Requirements, hardware, 4-7
Requirements, system, 1-2, 2-1
Resolution, 3-2
high, 1-9, 1-10, 1-11
standard, 1-7, 1-10, 1-11
Respecification of average, 1-15
Response commands, 7-15
Response, pre-stimulus, 1-8
Response time, 9-7
Responses, 3-1
abnormal, 1-13
RESTART, 5-20
Restart, 8-2, 8-4
Restart program, 9-13
Restarting DAQUAN, 5-4
RIM Loader, See Read
Loader
RETURN, 1-5, 2-5, 2-6, 3-6
Routine
calibration, 9-4, 9-14
filter, 9-16
plotting, 9-13
Routine, grid calibration, 1-27
Routine, plotter calibration, 1-2 8
Routines, 5-5 - 22
averager display, 5-7
DEVCOM, 7-16
DEVCON, 7-16
SKIPIT, 7-16
RUBOUT, 1-6, 3-7, 4-9, 4-14, 5-3
5-19, 8-3
RUN command, 7-4, 7-5
Run duration specification, 3-2
Running from disk or DECtape, 1-38
Running MS basic averager, 2-12

3-7
Questions, program start, 4-5

INDEX-

(reprint contents last location) ,SKIPIT routine, 7-16
5-19
Slew rate, 1-31
Samples per point, 9-20
SMOOTH command, 5-13, 5-32
Sampling A/D channel, 7-10, 7-12
Smoothing, 5-17
rate, 7-13
Smoothing data, 9-16, 9-23
Sampling parameters, 2-5
Software, 8-9
Sampling rate, 1-7, 1-10, 1-21, 2- 5Software requirements, 1-2
3-6, 4-8, 4-9
Sort code, 1-6, 1-11, 1-12
Save locations, 2-12
Sort time (definition), 1-11, 1-12
SAVE question, 5-32
Space (terminator)
5-3
Saving programs, 1-38
Spectrometer connection, 9-6
Sawtooth output, 9-7
Spectrometer frequency, 8-12
Sawtooth voltage, 9-1
Spectrum dif ferenciation, 9-16
SCALE command, 5-12
Spectrum display alteration, 5-15
Scale dimension, 3-12, 3-13
Spin functions, 8-7
Scale factor, 2-7, 4-11, 4-12
Spin number, 8-12
contraction, 4-12
SQUEEZE, 5-19, 5-30
expansion, 4-11
Stairsteps, 5-17
Scale factor, increase/decrease,
Standard deviation, 2-9
1-32
Standard resolution, 1-7, 1-10, 1-11
Scales, 8-2
Start data taking, 3-9
Scaling data, 5-13, 5-17
Start program questions, 4-5
Scaling factor, 7-1
Starting, 8-12
Schmitt trigger, 1-6, 2-5, 3-2, 3- 3 Starting procedure, 5-27
7-10
Starting time, 3-8
adjustment, 1-19, 1-20
Startup, 9-8, 9-20
recalibration, 3-6
Statement, GETC, 7-12
Scope
Stimulus, repetition rate, 3-10
channel, 5-5
Stop commands, 9-11
commands, 7-2
Stop data taking, 3-11
erasing, 7-3
Stop (S) run, 5-7
grid range, 5-14
Stopping DAQUAN, 5-4
refreshing, 7-2
Stopping operations, 9-17
Scope display, 8-4, 8-5
Storage, 1-36 - 41, 5-6, 8-1, 8-12
Scope display quality, 1-21
9-1, 9-20
Scope photo output, 1-24
STRIP command, 5-15
SCRATCH command, 7-3, 7-4
SUBTRACT command, 5-18
SDPRNT, 9-24
Subtract sloping baseline, 5-16
Section I, compiler, 1-5
Subtract straight line, 5-16
Section II, on-line signal averag- Sum adjacent bins, 3-12
ing, confidence limits, trend
Summary, control characters ondata collection, 1-20, 1-21
line section, 1-23
Section III, completion of confi- Summary Teletype keyboard commands,
dence limits , trend computation
Section IV,
1-24
SWAP command, 5-18
Section IV, output, analog X-Y
SWEEP, 9-24
recorder or scope photo, digital Sweep offset, 8-2, 8-4
values in actual input signal
Sweep time, 9-20, 9-21
units, 1-24
Sweep voltage, 9-4
SET CLOCK command, 7-7
Sweep width, 8-2, 8-4
SET RATE command, 7-6
Sweeping, 9-3, 9-4
Shaking pen, 1-30, 5-17, 5-18
direct, 9-7
SHIFT command, 5-31
and sampling, 9-21
Shift display, 9-16
Sweeps, 2-5
Shifting, 9-23
Sweeps, number of, 4-12
Shifts, chemical, 8-3, 8-4
Switch positions, 1-4, 2-3
Shim control, 9-7
Switching view mode, 4-12
Signal averaging, on-line, 1-20
Symbolic tape editor, 2-4
Signal enhancement theory, 9-13
Sync pulse, 1-6, 1-7, 1-17
Signal epoch length, 1-7
Sync pulse connection, 2-3, 2-5
Single precision, 9-1, 9-16
Synchronization pulse, 3-3
Single precision integers, 5-3
Synthesizer, 9-6
S

INDEX-

System
initialization, 1-4, 2-3, 4-8
requirements, 1-2, 2-1
(type digital data on teleprinter) , 2-6
Tailing of baseline, 9-25
Tape-reading error, 9-8
Tapes
control, 2-1, 2-4
modifying, 2-4
program, 2-1
Tau-Delta calculation, 8-12
Tau scale, 8-3, 8-4
Teletype
keyboard calculation, 5-9
listing, 5-14
punch, 5-21
Teletype keyboard commands,
section IV, 1-34
Teletype output, 1-33, 9-2
Teletype and blotter output, 8-11
Termination of
average, 2-5
plot, 2-6
typeout, 2-6
Terminator, 5-3
Theory of correlation, 4-1
Theory of signal enhancement, 9-3
Threshold, 1-19, 3-1, 3-3
adjustment, 3-6
values, 1-27
TIM function, 7-17
Time
execution, 9-2
plotting, 9-2
range, 3-9
zero, 3-3
TIME command, 5-8
Time scale, 3-3
increment, 3-2
range, 3-7
Title identification, 1-19
TOO BIG error message, 7-3, 7-4
Trace, 8-8
Transitions, 8-4, 8-8
listing option, 8-1
Trend, 1-14
completion, 1-20
computation, 1-20, 1-33
function, 1-26
interpretation, 1-33
plotting, 1-29
Trigger initialization, 3-6
Trigger threshold adjustment, 3-6
TST function, 7-15
Turn on system, 2-3
Type digital data on teleprinter,
2-6
Typeout termination, 2-6
Typing completed average, 2-6
Typing error, 8-3
T

UCOM command code, 7-16, 7-17, 7-18
Underflow bin, 3-9, 3-16
Underlining, 9-2
USE command, 7-5, 7-9
User coded function, 7-17
User command, 7-15
User commands implementation, 4UUF function, 7-15, 7-17
V (switch view between input
and output) , 2-6
Validity of data, 1-27
Varian HA-model instruments, 9-6
Vertical shifting, 9-23
View input mode, 1-17
View mode switching, 1-21, 4-12
Viewing scale, 2-5
Viewing switched between graph
and plot displays, 3-13
Voltage controlled oscillator, 9-6
WAIT command, 7-15
WAITC command, 7-8
Waveforms, 1-25
Wraparound, 5-7, 5-12, 5-13, 9-16,
9-24

Writing abbreviations and conventions, 9-3

X (expand display)
2-6, 2-7
X-invert command, 5-12
X-Y analog plotter, 4-12
X-Y analog plotter attachment, 1X-Y chart, drawing grid, 1-28
X,Y coordinates, 7-2
X-Y plotter, 2-6
X-Y recorder,
bias controls, 1-28
gain controls, 1-29
readout, 1-31
,

Y-invert command, 4-12
Zero all buffers, 3-18
Zero buffer, 2-7
Zero delay, 1-10
Zero memory, 9-13
Zero th histogram, 3-1

INDEX-

APPENDIX J
THE GENERATION AND MODIFICATION OF CONTROL
TAPES FOR THE LAB8/E BASIC AVERAGER
When the LAB8/E Basic Averager is used, it must be preceded or followed by control tapes. The purpose of these tapes is to define the
specific parameters involved in the average to be taken. The Basic
Averager Program is general in that it can be adapted to a variety of
situations. The control tapes accomplish specific adaptation.
There are eight different control tapes available from the DEC Program
Library. These allow the maximum number of points to be taken, in
all cases, as the number of channels and types of statistics are
In some cases, the user may want to decrease the number of
varied.
points taken to realize a shorter period of stimulus repetition or
to allow himself core for his own programs or he may want to sample
some number of channels not covered by the existing control tapes.
If the machine contains more than 4K of memory, the user may wish to
fully utilize the core available to him.

All of the situations above are easily reconciled once the user
understands what a control tape is and how it can be generated or
(Although it is possible to generate a control tape using
modified.
no other tools than the Symbolic Editor and PAL III Assembler, it is
easier to modify existing control tapes so that they will conform
to the new situations.)

Before attempting modification, a knowledge of the Symbolic Editor
and PAL III Symbolic Assembler should be obtained.
Briefly, what is done is the following: Using the Editor program,
the necessary changes are made in the ASCII control tape that is
most similar to the desired product. An ASCII tape containing the

changes is punched, and then it and DEC-LB-U00A-PA (a symbol reference
tape) are assembled (i.e., translated into binary code) using the
PAL III Assembler. PAL III then punches out the modified control
tape which can be used with the Basic Averager.

EDITING
Paper tapes marked DEC-LB-U01A-PA through DEC-LB-U08A-PA are the
ASCII versions of the control tapes that come with the LAB-8 system.

J-l

Snard? Ma^f ^

thr ° Ugh

SOftWare Dist ^»tion Center (DEC
NOTE

Programs that are written in the ASCII
language are readily understandable to
us, the users, since they are written
using alphanumeric characters and can
be heavily commented.
Because of the
straightforward nature of ASCII, it is
used exclusively in writing, debugging,
or modifying programs. When the program
is in its corrected form, it must
be
translated into the language understood
by the computer (i.e., binary). Once a
binary version is created it is used to
instruct the computer.
To edit the ASCII control tapes, load
the Editor Program and have it
read the desired ASCII control tape.
For clarity of discussion, a
listing should be created at this time
(type L and the RETURN key)
The first item printed will be:
/ (n) CHAN BASIC (AVGR) CONTROL TAPE

(UJ7-A)

(SD)

(TREND)

The slash denotes a comment following.
The next item printed will be DECIMAL; this
will tell the Assembler
that all numbers following should be considered
decimal numbers
(i.e., numbers of the base 10)

NCHAN

(1,

POINTS =

until the command, OCTAL, is given.
The number should be changed if
a different number of channels are
to be sampled.
Change if a different number of
points are desired.

As with all systems, there is a limit to
the number of points that
can be taken, because there is a finite
amount of core available for
storage.
In the Basic Averager there are 3027 words
available for
data storage. The LAB8/E averaging programs
are set up in such a
manner that the incoming data is first stored
in a temporary buffer
(one core location per point)
The values are then added to the sum,
which is saved in double precision (two core
locations per point)
If confidence limits are requested, the
value is squared and then
added to the sum- of the squares, which is
saved in triple precision
(three core locations per point)
The trend function requires four
additional locations per point.
.

J-

In summary, for one point the following core is required:

Core locations
Temp Buffer
Sum Buffer
Conf Lim Buffer
Trend Buffer

1
2
3
4

To take an average with confidence limits and trend would require
If there were 3000 locations available,
If two
maximum
of
300
points
could
be
taken using a 4K machine.
a
10 core locations per point.

channels were involved and both required data for confidence limits
and trend, 150 points/channel would be maximum,

JLIST=
The LAB8/E Averager programs use what is known as "job list processing".
Each average taken has a job list associated with it that defines the

parameters of that particular average (detailed description follows).
Job lists are seven words long.

If there is more than one, they

follow each other sequentially in memory.
there must be a word containing zeros

Following the last list

The symbol

the total length of the string of job lists.
7 *nchan

JLIST" designates

This is equal to

+ 1 and must be specified as the result of a summation or one

decimal number.

SUMLEN=
This statement defines the core required to store the sum associated
with 1 channel. This is always equal to 2 times the number of points
or POINTS + POINTS.

SDLEN=
This statement defines the core required to store the data for

calculating the Standard deviation or confidence limits for one
channel.
This should remain unchanged.
TRNLEN=
This defines the core required for the trend data.

Do not change this,

BLKLEN=
This defines the amount of core necessary to store all the data for
one channel.

Stored data is kept in blocks, each average having its own block.
The length of the block is determined by the number of data points
and the job type (associated statistics, if any)
Job Type

AVG
AVG

BLKLEN
2*#Points + 1
5*#Points + 1
9*#Points + 1

Confidence Limits
AVG, CLIM and TREND
&

J-

Again, the length must be specified as the result of a summation or
one number (it cannot be specified as the result of a multiplication)
One location is needed at the end for an end-of-the-block word.

OCTAL
The following numbers are to be considered in octal.
The next line of the listing will be the following:

*MEMTOT
The asterisk signifies that MEMTOT is a location that will have its
value specified later by the symbol reference tape. This enables
us to tell the assembler that we want to place a number in a particu-

This
The zero following is that number.
indicates the number of additional 4K stacks of memory on this
machine (i.e., an 8K machine would have *MEMTOT)
lar location in memory.

HIC0RE-L0C0RE-4
Defines the amount of core available for data
*ADJLIS
L0C0RE-1
It is used
Specifies the address -1 of the start of the job list.
by the program to locate the job list.
*ADCHNL
CHLIST=L0C0RE + 3LIST -1
CHLIST
Specifies the address -1 of the start of the channel display list.

*SMASK
SI

This location contains the sync mask.
the proper Schmitt trigger has fired.

This is used to determine if
The Basic Averager is set up

to respond to Schmitt trigger one.

*FAST
ISZ MCHAN or NOP
If only one
This instruction is used in the A/D sampling routine.
channel is sampled NOP should be placed there. More than one

channel requires ISZ MCHAN.

*KMCHAN
-NCHAN
Sets up a location to contain minus the number of channels that was

specified above.
*ASI
-1

This is needed for the compatability of sections

J-

and

*KADC
300

This is the mode setting for the A/D converter.
multiple input, 200 for single input.

300 is used for

*KMODE
5640
This is the mode setting for the clock. This will enable the clock
to tick at one microsecond a tick, to reset itself, and to enable
the A/D converter to start on its overflow.

*SAMA
-POINTS
This is minus the number of samples to be taken
*NCHA
NCHAN
This is the number of channels being sampled.
*ADBUFA
This tells the program where to store the A/D samples.
This is the
temporary buffer and it proceeds the channel display list.

*KBLA
This is equal to minus the total number of samples taken during a
sweep of all of the channels
The next area that may be edited is the job list itself.
Until this
point, all variables that have been defined have been common to all
the averages in this experiment.
Now, each average must have its

particular parameters defined. The first list will always start
at location LOCORE (0230g)
After the end of the last job list, a
zero word is inserted to indicate the end of the lists.
Word 1: Bit
always = J0
Bits 1-5:
specify the analog input number
pertaining to this average.
Bit 6:
always =1
Bits 7-11:
indicate the position of this aver.

age in the sampling list (usually
the same as bits 1-5)

Word

Average Type Code
AVG =
SD =
TRN =

for average only
for average and confidence limits
for average, confidence limits, and
trend (using these abbreviations
will lead to the correct translation)

J-

Word

-POINTS

This will assemble to be the negative of the number of data points
in the block of this average.

Word

CDF n

change data field instruction,
CDF (=62nl), where n is the number

;

of the data field (i.e., the block
of 4K; n =
for 1st 4K; n = 10
for 2nd 4K) where the first data
point for this average is to be
stored.

Word

ADCALn,

Word

Display Word

(see below)

Word

Display Word

(see below)

where n is the number of the average,
First average is always zero. This
is the address -1 of the beginning
of the data block belonging to this
average

Channel Display List
The channel display list contains two words for each average.
These
are the same as words 6 and 7 of the job list.
In assembling, the
channel display list comes directly after the end-of-list word in
the job list.

Display Word

The leftmost eight bits of display word

1 are the x-increment used
between points in the display cycle. The first five of these eight
bits are the integer part of the delta-x; the last three are the
fractional part.

Since the display scope has a maximum resolution of 1024 points in
each direction, it is useful to be able to use a fractional increment
for full screen utilization.
For instance, if the average contained
682 points, an increment of 1.5,- (00001. 100using 8 bits) would
just fill the scope with the display. An increment of 1.5
will
,

display a data point every 1.5 x locations.
A simple formula can be used to determine the DELTAX value to use,
given: a.) POINTS, and b.) PC, the percent or fraction of the screen
that you wish to cover with the display:
1024 * PC
deltax =
POINTS
This number should be translated into a binary integer and fraction
and placed in the left most eight bits of the first display word.

„'**

J-6

The last four bits of display word

are the Y scale factor (YS) for

the particular average. As each data point is displayed, its value
ys
is divided by 2
Thus, for full height display, YS would be zero;
.

for half height it would be one, etc.

Display Word

The first six bits of this word are X origin for the average display;
the last six bits are the Y origin.

Both of these numbers are multi-

plied by sixteen before being used as starting points for the display.
This makes it possible to start the display anywhere on the screen.
The first bit of the origin is interpreted as a sign bit so that it
is possible to start below zero.
For instance, a Y origin of -200 o
o

would imply a 50 _ as the low-order end of display word

Generating the Display Words
Here is an example of how to set up the display words for a 500 point,
two channel, average-only control tape.
The averages are assumed
to be displayed one above the other, the full width of the screen.
DELTAX:

Since there are 500 points in each sweep, our formula
gives DELTAX =1024/500 = 2.00
= 00010. 000
to
,
1Q

8- bit accuracy.

YS:

Each average should cover half the screen, so each
data point value must have been divided by 2
Therefore, YS should be 1.
We now have display word 1 as
.

000100000001

= 0401.

DELTAX YS
XZ:

Since both traces start at the left-hand side of the
scope, XZ will be -777„ for both averages.

YZ:

The first average must cover the top half of the scope
(0 - 777p) and its zero point should then be 400„.
This means that the second half of display word 2 will
be 20g for this average. Similarly,
the second

average must have the Y origin at - 377g and the second
half of the word will be -20 = 60„.
The channel disg

play list will now appear as:

/CHANNEL DISPLAY LIST
*CHLIST + 1
0401
4020
0401
4060
0000

/THESE WORDS ARE FOR
/FIRST AVERAGE.
/THESE ARE
/FOR SECOND.
/THIS IS END OF LIST.

J-

The last symbol to be defined is the end-of-block
symbol.
This is a
zero word at the end of the data block.
Since there must be a zero
following each data block, this is accomplished
by the assembler.
The address is ADCAL + BLKLEN + 1.

Possible Changes
The easiest change to make is the number of
points taken in an average.
The only changes necessary are POINTS and
the DELTAX value.
The
assembler (PAL III) effects the consequent changes.
To enable the Basic Averager to use more
than 4K of core is more
difficult.
The difficulty arises when data must be
stored and retrieved in upper core.
This is easily accomplished if the data
blocks
are in the lower 4K or upper 4K of core (i.e.,
a block is in one or
the other but not both)
Word 4 of the job list defines the data
field of the first point in the data block.
ADCAL
(the address -1 of
n
that data block) directs LAB8/e within that
data field. So, if a
data block is in upper 4K, only the CDF
instruction need be changed.
The user must keep in mind that when addressing
upper core (i.e.,
locations >7777) all addresses- are relative to
location 0000 of
that block of 4K.
.

NOTE
If OS/8 is being used, page 7600 of
all fields can not be used.

To define an address in upper core, the user must
type in the command
FIELD n, where n is the data field. All addresses
specified thereafter will be located in that data field (the first 4K
is field 0)

When a data block lies in both upper and lower core,
the LAB8/E
switches between data fields by a process called
"linking." Linking
is accomplished by a set of three linkage
words at the end of the
block in the lower field. These words inform LAB8/E
of the number
of data points (not core locations or values) that
are needed to
complete this average, the data field in which these
points lie,
and the location -1 of the first word in the next
data block.
These
linkage words must be supplied by the user.
The upper limit of the data storage area in the
lower 4K is 6277
The Basic Averager program is located above this.
So, at some point
before this limit, the linkage words must be inserted.
(After
assembly, the user should check to see that these
words were placed
in locations lower than 6277.)
8

J-

Note that word

of the job list pertaining to this average should

show only the number of points found in the first part of the data
block (i.e., it must not include the point count contained in the
part of the block that is in upper core)
The CDF instruction and
.

ADCAL word of the job list should refer to the address and field
of the first data point contained in the part of the block which is
in lower core.
The linkage words should contain the information on
the points that are in upper core.
Example:

Assume an extra 4K of memory
1 channel
Avg, Conf Lim, and Trend 600 points start with
tape DEC-LB-U07A-PA
.

Changes:
1.)
2.)

Points = 600
*MEMTOT

3.)

In the job list;

Word 3 should be minus the number of
points contained in lower core.
In
this case, 200 points will be stored
in lower core.
DECIMAL
- 200

OCTAL
4.)

Word

DELTAX must be changed to correct
the display for increased number of
points
.

# of points = 600 for full-screen
display.

DELTAX = 1.705
= 1.551g (for
1Q
decimal-to-octal conversion see
Introduction to Programming Fl-4)
so word 6 becomes:
Delta X
000 011 010 000 = 0320
i.e.

y-scale remains zero
The first channel display word must
also be changed to 0320.

5.)
6 . )

1.5 + y scale

Linkage Words

These three words must be located
directly after the last value pertaining to the last point in this
block.
This location is determined
by starting with ADCAL0.
This is
the address -1 of the first word of
the data block.
The first word of
the data block contains the number
of sweeps, and the following words
contain the data. Therefore, the
last location is equal to:

ADCAL0 + 1 + (No. of locations
needed for data storage
+ 1
)

J-9

In this case, 200 points are stored in this block.

Each point needs
two words for the sum, three words for the confidence limits, and
four words for the Trend; a total of nine words per point.
Therefore,
the address of the word following the data is:
ADCALO + 1 + 1800
+ 1
1Q

Linkage Word
-400,6211 8

1
2

/-No. of points in next block
/ CDF for n ext field

/location -1 of first word

"l0
Note

Linkage word

was chosen arbitarily.

Define these in the following manner:
DECIMAL
*ADCAL0 + 1 + 1800 + 1
-400
OCTAL
6211
DECIMAL
99

Now the end-of-block word must be inserted. It lies in Field 1, 400.
n
points or 3600
words from the start of the buffer. There is no
1Q
sweep count included in this block, so the location of the word following the last data word is:
(starting address -1; specified in linkage word) +
words used for storage)

1 + Cnumber

In this case, the location is 3700....

Define this by the following:

FIELD 1
DECIMAL

/Change Fields

*3700
$

The dollar sign signifies the end of the characters to be assembled.

Once the user has familiarized himself with the procedure and logic
of editing and assembling control tapes, he may go on to furthermore

elaborate modifications.

When all modifications are completed, a tape should be punched by
typing P, RETURN, turning on the punch (high speed or low speed), and
pressing CONTinue. After the tape is punched, type F and RETURN, then
T and RETURN.
The output tape may now be assembled by any of the DEC assembly
programs. The output of the assembler may be used as a control tape.

J-10

APPENDIX K
EXAMPLE OF PARAMETER SETUP
PST AND LATENCY
In the following example, the underlined portion of the dialog is gener-

ated by the program.
1.

The reply is not underlined.

Program started at location 0200.
TRIGGER J

PST ?

Check the Schmitt trigger thresholds.
Carriage RETURN tells the program
this has been done.

PST or Latency?

PST chosen.

BINW=

5.0 J

Bin width is set to 5.0 milliseconds.

BINS=

100 J

There are 100 bins in the histogram.

EPOCHS=

40 J

There will be 40 stimulus-response
sequences accepted.

MINTIM=

No minimum time constraints.

This dialog will produce a PST histogram.

It will measure arrival

times of responses from time zero (mintime = 0) up to 500 milliseconds
(binwidth x # of bins) with a resolution of 5 milliseconds following
the stimulus.
2.

Forty stimulus-response sequences will be accepted.

Program started at location 0200.
TRIGGER p
Check the Schmitt trigger thresholds.
Carriage RETURN tells the program
this has been done.
PST ? Nj
PST or Latency? Latency chosen.
ORDER=
BINW=

?BINW=

5 J

0.5

The first five pulses following the
stimulus should be recorded.

First answer was not in correct
format.
Question is repeated.

Bin width is .5 milliseconds or 500
ysec.

BINS=

19P0-J

1900 bins exceed the maximum allowable.
User is notified of this and
the question is repeated.

1800

1800 bins are requested and accepted
bv
by the nrooram.
program.

CORE? J

BINS=

K-l

EPOCHS=

\$$)

Again the core limits are exceeded
and the question is repeated.

70J

70 is accepted.

CORE?*)

EPOCHS =

Minimum time is

10j

MINTIME=

(10 x

milliseconds

.5)

It will be a composite fifth
Here a latency histogram is requested.
order latency histogram (the histogram will be the frequency distribuIt
tion of the first five pulses arriving following the stimulus) .
will measure the time of arrival starting at 5 milliseconds after the

stimulus

(mintime = 10 x bin width) until

9 55

milliseconds after the

bins + mintime) with a resolution of 500 ysec.
70 stimulus-response sequences will be accepted before automatic transfer to output mode.

stimulus (bin width x

TIME INTERVAL HISTOGRAM
1.

Program is started at location 0200

TRIGGER J

Check the Schmitt trigger thresholds.
Carriage RETURN tells the
program this has been done.

BINW= 5.0 J

Bin Width is set to 5.0 milliseconds

BINS= 100 J

There are 100 bins in the histogram.

MINTIM=

No minimum time constraints.

TIME NUNIT= 200.0 J

The basic time unit for timing the
program is 200.0 milliseconds.

#UNITS= 200 J

There are 200 of the basic units
so the experimental run will last
40 seconds (200 ms x 200)

K-2

APPENDIX L
IMPLEMENTATION OF USER COMMANDS
FOR PST

A listing should be consulted before either of the user commands,
CTRL/U or CTRL/I, is utilized.
The user commands are designed to be patched in and will do
anything
that the user can program.
CTRL/U is a Group I command and will be
executed during data acquisition. To utilize CTRL/U, change location
1356 from 1240 to the location of the user routine.
At the end of the
user patch there should be a CLL CLA JMP I 121 (5521). This
will

return control to the main program.
CTRL/I is a Group II command to be executed only after data collection
is finished.
To implement CTRL/I, change location 1357 from 1240 to
the location of the user routine. Again, at the end of the
routine
there should be a CLA CLL JMP I 121.

NOTE
The user patch for CTRL/U must not turn the interrupt ON or OFF.
It also must not disturb
the clock in any way.
To find free areas of core, consult a listing.

The MS version of TIH and PST used CTRL/I for CTRL/W.
If a user command is needed in Group II the histogram size has to be made smaller
and the command put in the command list.

L-i

APPENDIX M
HISTOGRAM BINARY OUTPUT FORMAT

When the CTRL/B command is given, a binary dump of the histogram being
displayed will proceed on either the high or low speed punch. The dump
will start with 64 leader-trailer punches (these are channel 8 punches)
For a description of paper tape punching format, see Introduction to
Programming , Vol. 1, 4-14.

After the first leader-trailer punches, the ID code will be punched in
ASCII format. This will be followed by more leader-trailer.
The core
image is preceded by an address.

punch with the most significant

This is signified by a 7-channel
6

followed by the least significant

bits of the address on the same line
6

bits on the next line.

A core

image follows the address.

The first core image is the display list. This gives the starting location -1 of zero th histogram followed by the endinq address, followed
by the starting address -1 of the first order histogram, followed by

the ending address of the first order histogram.

After the display list, the address of the histogram being dumped is
punched followed by a core image of the histogram. After all information has been punched, a checksum is punched.

A checksum is a means of checking the validity of read-in. It is a 12bit word which represents the sum of all of the punches in the tape
that are not leader-trailer or ASCII information. It is formed by adding every line of punched data (7 channels) to a 12-bit word.
Overflows
are ignored, and the resulting 12-bit sum is punched at the end of the
data.
This is followed by more leader-trailer code.

When the binary loader reads a binary formatted tape, it keeps a running 12-bit sum of all punches between the end of the leader code and
the checksum.
NOTE

The checksum punches are not included in the checksum.
If the result is non-zero, an error on read-in has occurred and the tape should be read-in again.

M-i

APPENDIX N
SUMMARY OF COMMANDS FOR PST, TIH AND LATENCY HISTOGRAM PROGRAMS

Group

Commands for Data Collection and Output

Action

Command

Increase order of display.
Decrease order of display.
Stop, zero buffers wait restart.
Stop, return to alter parameters.
Quit taking data after next epoch.
Return to Monitor.
Panic stop.
Scale display up by power of 2.
Scale display down by power of 2.
Expand area between cursors.
Cancel all E commands, and show beginning display.
Sum n bins before display.
Negate all S commands - return original display.
Return to unaltered display.
Complement axis - no axis switch.
Complement cursor - no cursor switch.
Complement view bar graph or point plot display.
User Group I command.

CTRL/Z
CTRL/A
CTRL/Q
CTRL/C
CTRL/P
U
D

E
B
S
z

A
C

V
CTRL/U
Group II:

Commands Cannot be Used during Data Collection

Action

Command
P

K
T

CTRL/B
CTRL/R
CTRL/I

Plot the histogram being displayed.
Calibrate the plotter.
Type the data between the cursors.
Do a binary core image dump.
Retain the data, ready for restart.
User Group II command.

N-l

APPENDIX
IMPLEMENTATION OF USER COMMANDS
FOR AUTO- AND CROSS-CORRELATION
J

Two user commands are available without reassembly.

and CTRL/O.

CTRL/U is a group

data taking, and CTRL/0

These are CTRL/U

command, that is, it can interrupt

is a group 2 command which indicates that it

operates only after data taking has ceased.
CTRL/U, because it operates while data taking, should be a short, fast
utility routine, because calling it could affect the timing of the

correlogram.

In normal mode, the clock is interrupt enabled, that is,

whenever it ticks an interrupt results. Thus, in normal mode, CTRL/U
could be interrupted in the midst of carrying out its task. This is
why it is recommended that CTRL/U be fast. However, if CTRL/U is long,
precautions must be taken not to affect the data and pointers used by
the routines which actually perform the algorithm.
Batch and normal modes

This applies for

These limitations need not apply to CTRL/0, because it can be called
only after data taking has ceased.
CTRL/0 must not destroy vital flags
or pointers, but has available temporary registers normally used by the

algorithm routines.
How to PATCH in a User Program
The location which specifies which location to jump to in case of a
CTRL/0 is C0MG0+1. The user should put the starting address of his

CTRL/0 routine in this location. NOTE: The routine is not called
with a JMS, but a jump is made to it. Exit is made by one of the two
following methods
1)

JMP I CHECK1 - This will return to the location following the call
to the command checker.
This should be used only
by a routine that does not modify locations used by
the correlator.

JMP

C0NT3

This restarts the display loop, causing all display
variables to be reinitialized. A link to C0NT3 is
on page 0.

The location which specifies the Starting Address of the CTRL/U command
is COMGO+llg.
All other specifications are as above.

*It is assumed that the user has a listing

Software Distribution Center, Maynard)

0-1

(available from the DEC

The user, by looking at the core maps, can see which core is free
for
his use.
To get the octal values of the symbols used, consult a listing. 1

Useful Subroutines
There are several useful subroutines which are available for user
implementation. Many of these are in the form of pseudo-ops. A

description follows

Subroutine

Calling Sequence:

Function:

Subroutine

This routine puts the contents of the
Floating Accumulator, locations EXP,
HORD and LORD in the address pointed
to bv the word after the call.

Calling Sequence:
Function:

NORM
Return

NORMalizes FAC (Floating AC).

Calling Sequence:

Function:

SAVE
Address
Return

DIVF
Address
Return

Divides the FAC by the Floating Point
Word whose exponent is pointed to by
CALL+1.

Subroutine

Calling Sequence:

SORT J
LIST1-1
LIST2-LIST1
Possible Return

Function:

This routine compares the contents of
location CHAR with that of the entries
in LIST1.
If a match is found then a
jump is made to the contents of the
corresponding location in LIST2. LIST1
is terminated by a negative number.
If
a match is not found, then SORTJ returns
to CALL+3.

Example:

SORTJ/SEE IF CHAR = "+";"-"; OR "!"
LIST1-1
LIST2-LIST1
JMP CONT/DOESN'T CONTINUE
LIST1,"+/PUT ASCII VALUES
ii

" !

-1/END LIST
LIST2,ADD/JMP TO ADD IF A"+"
SUB/ JMP TO SUB IF A"-"
IOR/JMP TO IOR IF A"
!

Available from Digital Equipment Corporation, Software Distribution
Center, Maynard, Massachusetts.
0-2

Subroutine

Calling Sequence:

SETDAT
Return

Calling Sequence:

SETUP
Return

Functions

Modifies

Subroutine

NUMER, DENOM

Calling Sequence:
Function:

Subroutine

These routines set up data for display routine.
This data must be set
up before displaying.

Displays the buffer once.

Calling Sequence:

Functions

Example:

DISPLAY
Return

MESAGE
MES-1
Return

Types a CR-LF and then the string of
8-bit ASCII characters pointed to by
CALL+1.
The list is ended by a 0.
Return to CALL+2.

MESAGE/TYPE "GOO"
MES1-1
JMP CONT/CONTINUE
MEST'G
"0
"0
J0/END

Subroutine

Calling Sequence:
Function:

Subroutine 10:

Subroutine 11:

RE ADC
Return

CRLF
Return

Types a CRLF combination

Calling Sequence:

Function:

Reads a CHAR from TTY.
Checks to
see if it is a command, and then
puts the character in location CHAR.

Calling Sequence:
Function:

Subroutine 12:

Prints (AC) if AC not =
(CHAR) if AC = 0.

Calling Sequence
Function:

PRINTC
Return

DNEG
High ADD
Return

This routine performs a double precision negation of the double precision word whose high order word is
pointed to by CALL+1.
It is assumed
that the low order word is located in
core right after the high order.

0-3

Subroutine 13:

Calling Sequence:
Function:

DECIN
Return

This routine inputs a decimal number
from the Teletype.
Two input modes
are possible.
1)

Input with a decimal point.
In this case, input is in one
of the following formats
A)
B)
C)

XXX.
XX.

X.X

In all cases a period and a
digit after it must be typed.
This input mode is used when
location X = 0.
2)

Input without a decimal point.
This input is used when location X<0.
Up to three decimal
digits may be typed.
Input may
be prematurely terminated by a
Carriage Return.
In both cases a RUBOUT restarts
input.
After input is done, the
converted decimal number is left
in location number.

In Case 1, that is input with a decimal point, the value returned is that
of the number typed in if it did not
have a decimal point.
For example,
if 20.0 is typed, the value returned
is 310
or 200
10

Subroutine 14:

Calling Sequence
Function:

Subroutine 15

Subroutine

16:

CURSOR
Return

This reads in the values of knobs
and 1, and displays cursors at the
X point that corresponds to their
values.

Calling Sequence:
Function:

CLEAR
Return

Clears the buffers.

Calling Sequence:
Function:

ADC
Return

This routine samples the A/D channel specified by the AC upon entry.
When it returns, the value read is
in the AC.

Subroutine

17:

Calling Sequence:

0-4

DECOUT
Return

Function:

Subroutine 18:

Calling Sequence:
Function:

Subroutine 19:

Types the decimal value of the
double precision number held in
locations HIGH1, L0W1. This is
signed output, no CR/LF is typed
before or after output.

EFFECTIVE JMS SHIFTL
High Address
Return

Shifts the double precision word
pointed to by CALL+1 the (AC)
times left.
It then returns to
CALL+2.
The (AC) must be positive.

Calling Sequence:
Function:

Subroutine 21:

DOUT
Return

Calling Sequence:

Function:

Subroutine 20:

This routine types out the decimal
value of the number in the AC upon
entry.
The (AC) may be negative.
No CR/LF is typed before or after
output.

EFFECTIVE JMS PNCH
Return

Punches the (AC) on high speed
punch.
Should be called with
interrupt off.

Calling Sequence:

Function:

EFFECTIVE JMS SHIFTR
High Address
Return

Shifts the double precision word
pointed to by CALL+1 (AC) places
right, then returns to CALL+2. The

AO0.
Subroutine 22:

Calling Sequence:
Function:

Subroutine 23:

EFFECTIVE JMS DPDIV
Return

This routine divides the double
precision word, which may be signed,
held in locations HIGHl, LOW1, by
the double precision word, which
must be positive, held in locations
HIGH, LOW. When it returns, the
double precision quotient is in
locations HIGH2, DIGBOX, and the
remainder in locations HIGHl, LOW1.
If there is division by zero, then
an error typeout of "DIVISION BY 0"
results.

Calling Sequence:

EFFECTIVE JMS ADD
High Address 1
High Address 2
Return

-5

Function:

Subroutine 24:

Calling Sequence:
Function:

Subroutine 25;

This routine adds the double precision number pointed to by CALL+1 to
the double precision number pointed
to by CALL+2, putting the result in
the word pointed to by CALL+2.
Return is to CALL+3.

This routine floats the double precision number held in locations HURD,
LORD.
The exponent is left in location EXP. Thus, the result is left
in the FAC.

Calling Sequence:
Function:

EFFECTIVE JMS DFLOAT
Return

EFFECTIVE JMS ACMINS
Return

Negates the FAC.

0-6

APPENDIX P
CORRELOGRAM BINARY OUTPUT FORMAT

When the CTRL/B command is given, a binary dump of the correlogram will
proceed on either the high -or low-speed punch. The dump will start
with leader/trailer code (these are channel 8 punches)
For a description of paper tape punching format, see Introduction to
.

Programming, Vol.

4-14.

After the first leader/ trailer code, the identification code will be
punched in ASCII format. This will be followed by more leader/
trailer.
The core image is preceded by an address. This is signified
by a 7 channel punch with the most significant 6 bits of the address
on the same line followed by the least significant 6 bits on the next
line.

A core image follows the address.

The first locations to be dumped are the parameters of the correlogram
that are necessary for proper display and accurate typeout.
In order
of their dumping, these parameters are:

CLKCNT (minus the clock rate);
WRDCNT (minus the number of points in the correlogram)
ENDTMP (end
the address of temporary buffer -1)
ENDPRM (end address of
;

;

permanent buf fer-1)

;

of the sweep count)

;

SWEEPH and SWEEPL (the high and low order words
NUMBER (number of points) BTCFLG (this =
for
;

normal mode, -1 for batch mode); and FLAG (this =
correlation, -1 for cross-correlation)

for auto-

After the display and typeout parameters are punched, the address of
the correlogram is punched, followed by a core image of the correlogram.

After all information has been punched, a checksum is punched.

A checksum is a means of testing the validity of read-in. It is a
12-bit word formed by adding every line of punched data, which
represents the sum of all of the punches in the tape that are not
leader/ trailer or ASCII information. Overflows are ignored and the
resulting 12-bit sum is punched at the end of the data. This is

followed by more leader/trailer code.

When the binary loader reads a binary formatted tape, it keeps a
12-bit sum of all punches between the end of the leader code and the
checksum (note: the checksum punches are not included in the checkWhen the checksum is read, it is subtracted from the 12 bit sum
accumulated during read-in. If the result is non-zero, an error on
sum) .

read-in has occurred and the tape should be read in again.
P-l

APPENDIX Q
CHART RECORDER INTERFACING INFORMATION TO LAB8/E VC8-E

The VC8-E (10 bit point plot display controller) consists of two modules (M869 and M885) which plug into the OMNIBUS.
The M885 module

includes the "male" portion of a "Berg Connector". The user can plug
the "female" portion (DEC Part No. 1210090 and pin Part No. 1210089 one pin per signal required) with the chart recorder cable assembly
into the male portion. A sketch of the female Berg Connector with the

pin locations is shown below:

BERG
B

W
uu

The pin assignments and their functions follow:

Function

Analog Y output (±5V into lkohm minimum requirement)
Analog Ground

F
L

N
X

V
z

Analog X output (±5V into lkohm minimum requirement)
Analog Ground
output (Not used for chart recorder)
Logic Ground
Z

Channel Select (TTL level into lkohm minimum)
Logic Ground

NOTES

If the chart recorder is differential, run the "X output" and
the "analog ground" into two inputs and use the "logic ground"
as a shield.

If the chart recorder is single-ended, tie the "analog ground"

into the "recorder frame ground".

Q-l

APPENDIX R
DAQUAN COMMAND SUMMARY

At least the first two characters of a command must be typed before
colon is typed.

INPUT-OUTPUT COMMANDS
AVERAGE:

PAPERTAPE
OUTPUT
PLOT:

Accept time averaged analog data
Input data from paper tape or keyboard.
Print/punch paper tape
Plot data on X-Y analog recorder

PROCESSING COMMANDS
CALCULATE
ALTER:

Calculate Lorentzian and/or Gaussian spectrum
Alter parameters input by previous CALCULATE command

COPY:

Copy CDC 1 into NDC 2

XINVERT

Invert X axis

YINVERT:

Invert Y axis

SCALE

Scale to range of 0-1000
Scale to arbitrary range

MULTIPLY:
SMOOTH:

CURSORS
LIST:

Apply eleven point digital filter
Set up two cursors on scope
Print decimal values of cursors

INTEGRATE

Integrate between cursors or running integration (preceded by CURSOR command)

STRIP:

Strip out data or baseline (preceded by CURSOR command)
Form differences (derivatives)

DERIVATIVE:
SUBTRACT:
ADD:

Subtract CDC from NDC: results in CDC
Add NDC to CDC; results in CDC

SWAP:

Swap CDC and NDC

SQUEEZE:

Average adjacent points of displayed channels
Find peaks and report on them

PEAK

REPORT:

MASS STORAGE:

Read or write from mass storage (system

device)

SPECIAL COMMANDS
MODIFY:

ODT-like core modifier
TIME:
Set machine cycle time constant to calibrate AVERAGER
RESTART;
.Restart program
*CDC = currently displayed channel
NDC = non-displayed channel
_— .
R 1

LAB8/E Software
System User's MANUAL
DEC - 8 E -ALUMA -B -D
READER'S COMMENTS
NOTE:

This form is for document comments only. Problems
with software should be reported on a Software
Problem Report (SPR) form (see the HOW TO OBTAIN
SOFTWARE INFORMATION page)

Did you find errors in this manual?

If so,

specify by page.

Did you find this manual understandable, usable, and well-organized?
Please make suggestions for improvement.

c
o

Is there sufficient documentation on associated system programs
required for use of the software described in this manual? If not,
what material is missing and where should it be placed?

Please indicate the type of user /reader that you most nearly represent.
I

l~)

(~)
I

Assembly language programmer
Higher-level language programmer
Occasional programmer (experienced)
User with little programming experience
Student programmer

Non-programmer interested in computer concepts and capabilities

Name.

Date.

Organization.

Street
City.

.State.

.Zip Code.
or

Country
If you do not require a written reply, please check here.

-Fold Here-

Do Not Tear - Fold Here and Staple

FIRST CLASS

PERMIT NO. 33

MAYNARD, MASS.
BUSINESS REPLY MAIL

NO POSTAGE STAMP NECESSARY IF MAILED IN THE UNITED STATES

Postage will be paid by:

Software Communications
P. 0. Box F

Maynard, Massachusetts

01754

APPENDIX

CALCULATE ROUTINE
S.l

FUNCTIONAL EQUATIONS

Functional equations used by CALCULATE:
The equation for a sum of Lorentzian curves at point i is
Y

= B+Z ^

/[(X -P
i

)

]

= B+L

where
P

is number of peaks

is baseline

is intensity of

is half width at half height of

peak

i-h

th
j

peak

is position on X axis of

is X coordinate with resulting net intensity of Y.

peak

The equation for a sum of Gaussian curves is
Y.

P
= B+Z

j=l

„ -1
H e
2

2
i

("RJ^

B+G

with similar parameter definitions except that here W. is the sigma
value.
Combination of the two by means of the LORentzian FRaction is
Y

= B + LORFR

•

(1-LORFR)

•

DIGITAL FILTER EQUATION:
The expression is effectively a low pass filter designed to retain
the integrity of peak areas.
Intensities may be diminished if there
are fewer than 15 points between the peak's inflection points (approxi-

mately the width at half-height)
The equation takes the form of a nonlinear, weighted average of a data
point and its ten nearest neighbors:
5

Y.
1

j=0

C.
3

•

(Y.

.+Y.

i+D

i-:'

S-l

where

0.1038

0.1956

0.1608

0.0488

0.021

l
2

C
3

C
C

4
5

= -0.030

and

S.2

is the resultant, filtered value.

Modifying the CALCULATE ROUTINE

The CALCULATE routine makes these two stipulations:
1.

interaction of a peak beyond the range of eight times
its width is negligible

for initial entry, peak widths are all equal.

These can be removed by changing contents of location 2417 to 5242
and contents of 2 212 to 7001, respectively.
The first change will
include all interaction and is desired for pure Lorentzian spectra.
The second requires the user to enter the width for all peaks.
If desired, these changes should be made using MODIFY

of Book

and before requesting CALCULATE.

S-2

(Section 2.22

APPENDIX T
CORE MAPS

Core Map of DAQUAN

T.l

Field
Page

Contents

200)
400)
(
600)
(
(1000)

(1200)

(1400)
(1600)
(2000)
(2200)
(2400)
(2600)
(3000)
(3200)
(3400)
(3600)

1
2
3

7
8

10
11
12
13
14
15
16
17
18
19
20

Constants, subroutine links, temp, storage
Main display loop, fix and float subroutines
MODIFY*, TTY subroutines
F.P. input, output subroutines, plot axis marker
plotter calibration, answer subroutine, delay, A/D, DAC,
relay subroutines
Leader/trailer subr. X display increment setup, PAPERTAPE input, X-Y analog PLOT setup, Y- INVERT
OUTPUT, SCALE, increment x coordinate subroutine
Analog plotter driver, message output subroutine
Keyboard command decoder, command dispatch table
SUBTRACT, COPY and mass storage
Mass storage
PAFFT routine, 7600 field 1 storage
SWAP, SMOOTH, ADD, CURSOR routines, clockdriver
cursor decoder, STRIP
INTEGRATE
More calculate routines, array initializer, DERIVATIVES,
SQUEEZE
AVERAGE setup, TIME calibration, LIST cursors
Average acquisition routine, output setup for PEAK REPORT
Intersweep average display, X-INVERT, MULTIPLY
Output routine for PEAK REPORT
PEAK REPORT peak detector

(

(4000)
(4200)
(4400)
(4600)
(5000)

5131-5377
5400-7577

6-bit packed ASCII messages
Floating Point Package

MODIFY overwritten by parameters when CALCULATE is used.
Field

0-2111
4000-7777
Field

channel
channel
field

1
2

data
data, note 7600-7777 is saved in 2600-2777 of

Utilization

AVERAGE requires 2N locations starting from base of the channel
where N is the number of points. Hence, a 1024 point average
into channel 1 just fits. More than 1024 points will overrun into
the other channel.

PEAK REPORT requires N+9p locations where N is the number of
points and p is the number of peaks detected. The peak parameters are stored immediately after the data array.

T-l

T. 2

Core Map for DAFFT

Only the major departures from DAQUAN Core
Mao
c
here.

1)
'

are presented

Field
Page

Contents

Field 1 dispatcher, POWER, SHIFT, SMOOTH,
dispatcher
overlay MODIFY
15 (3731-3756) SQUEEZE overlayed by message text
and FACTOR output
i-f"<from FFT's.
Field

(

200)

400-1577
1600-3577
3600-5577
5600-7577
T 3
.

Constants, storage for FFTS-C
riVerS ' Hannin 9 fi lter, zero mean routine
FF? s ^
Channel 1 data
Channel 2 data
Storage buffer

Core Map for PAFFT

below^

PartUreS fr °m C ° re maPS ° f DAQUAN

(T ' 1)

and DAFFT

Field
Page

Contents

None

Field

1( 200)

Hanning filter, zero mean subroutine
FFT drivers, power average subroutine

T-2

are shown

APPENDIX U
BASIC/RT COMMAND AND FUNCTION SUMMARY
U.l

Statements

Start A/C sampling

CLEAR

Erase Scope

DATA

Provide initial data for a program

DEF

Define a function

DELAY

Display scope buffer

DIM

Define subscripted variables

END
FOR TO
STEP

End a program

GOSUB

Go to a subroutine

GOTO

Change order of program execution
Conditionally change order of program execution

IF THEN

IF THEN nn

Set up a program loop

INPUT

Conditionally execute the statement nn
Get variable values from the terminal

LET

Assign a value to a variable (LET is optional)

LPT

Assign line printer as output device
End a program loop

PLOT X,Y

Plot on the scope the point x,y

Print out the indicated information

PTP

Assign high-speed paper tape punch as output device

PTR

Assign high-speed paper tape reader as input device

READ

Initialize variables to value from the data list

REAL TIME

Create buffer for A/D samples

REJECT

Stop A/D sampling

REM

Insert a program comment

RESTORE

Restore the data list

RETURN
SET CLOCK

Returns from a subroutine
Set clock enable register

SET RATE

Set clock interrupt rate

STOP

Stop program execution

TTY IN

Assign Teletype as input device
Assign Teletype as output device
Specify a display buffer
Pause execution until next interrupt

TTY OUT

USE

WAIT
WAITC

Delay program execution until next clock interrupt

U-l

APPENDIX V
SUMMARY OF THE USE OF NMR AVERAGER

Load with DEC Binary Loader, ADDR=0200, CLEAR, CONT.

Averaging -

NMR-AVERAGER
SWEEPS =
SWEEP TIME =
SAMPLES =
PAUSE TIME =
REPEAT?

TYPE RETURN

CTRL/H

(0 shifts directly to display)
0<t<4096
0<s<20
0<p<40
Y for continuous sweeping
N or RETURN to do only N sweeps
RETURN to erase and begin averaging
LINE FEED to begin averaging without
erasing memory
CTRLA) to restart the program from the beginning
Halts the program at the end of the current
sweep
Halts the sweeping immediately (error recovery only)

0<n< 4096

END
3.

Pre-Normalization commands
C
X
P

CTRL/Z
N

K
B
T

CTRL/C
G

CTRL/E
4.

Contract
eXpand
Plot
Zero (erase memory)
type Number of sweeps completed
calibrate
Binary punch of accumulated data
Type out decimal values of accumulated average
return to OS/8 or disk monitor
Go - continue averaging
Restart program from beginning
Enter the normalized section

Post-Normalization commands
C

X
U
D
P

CTRL/Z
N
R

K
F
I

A
CTRL/0
G

CTRL/D
CTRL/C

Contract
eXpand
Up
Down

Plot
Zero memory
Number of sweeps completed
Restart
calibrate
Filter - curve smoothing
cursor Integration
Analog (RC) integration
return unmodified display
Go - start averaging - effective only after CTRL/Z
Differentiate display
return to monitor

V-l

APPENDIX W
LAB/8E MASS STORAGE PROGRAMS
BINARY DECTAPE

All binaries used by the LAB/8E MS programs are contained on a

DECtape DEC-8E-ALMSA-A-UB.

The files are as follows:
the Advanced Averager
the Advanced Averager
the Advanced Averager
the Advanced Averager
the Advanced Averager

AAVG1.BN
AAVG2.BN
AAVG3 BN
AAVG4.BN
AAVG5.BN

Section
Section
Section
Section
Section

BAC1.BN
BAC2.BN
BAC3.BN
BAC4.BN
BAC5.BN
BAC6.BN
BAC7.BN
BAC8.BN

Basic Averager control
Basic Averager control
Basic Averager control
Basic Averager control
Basic Averager control
Basic Averager control
Basic Averager control
Basic Averager control

BAD2.BN

Basic Averager

PSTD1.BN

PST

TIHD2.BN

TIH Histogram

C0RD3.BN

Auto

CONDI 4. BN

Convert Program

DAQD5.BN

DAQUAN Without a Floating Point Package

FPPEAE.BN

EAE Floating Point Package for DAQUAN

FPPNE . BN

non EAE

DAFFT3 BN

DAFFT Overlay to DAQUAN

PAFFT2 BN

PAFFT Overlay to DAQUAN

1
2
3
4
5

of
of
of
of
of

tape
tape
tape
tape
tape
tape
tape
tape

Latency Histogram

Cross Correlation

F.P.P. for DAQUAN

W-l