.~l\\a- fiafi‘ntd an».

- 'r

.1‘ \\1

 

N

. M. A ...u pip »
_ ,, .y g... . ,
7r 5 n :21".

   

AS.

.LH':
'\ ..'

can

I .
F

mu; .‘Hv

        
     

PA"

H1“ Iwfla

\"5 ~j.
‘ Aka

,.. u.

;..‘.u
.w

23-?

3: ...- .4
. .a‘ -I... . fi‘ ..r

 

3 .K’
M.) ‘hh‘n‘u
5’34 .

   

1w”
»...n
n- In

5M: r‘n’rwxaf. Apr, 7,;

.‘i

«LLHHHUH

éw’w

 

..
A£;..:.v.‘

ln-v'l v! “'4 - livn
11p.

 

 

T‘“‘% if";

.A‘ §n ». ~
; ,u a“:-
n .u, ,u: . n:
” ‘4'.le : .3.
3‘ . .r.‘ . a

1». '

7:3":

at”: b u

 

Jru‘rv.
.n‘
1 ‘
r.“ . .4

37'» '3

a: ‘

 

‘1 w
2..
«'2‘!

,.
F
r
~
,
‘.

 

u‘

...

 

 

   

 

 

   

 

{HF-31‘s

MICHIGAN STATE

III I III

 

 

 

 

 

 

 

 

 

 

 

 

 

 

IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII

008918421

This is to certify that the

thesis entitled

Precision Automated Measurements

presented by

Robert Jay Randel

has been accepted towards fulfillment
of the requirements for

Master's Electrical

degree in

 

 

Engineering

m ‘ //

A0 / a flajoru pr essor
Date 0C [7/ /77‘)
//

0.7 639 MS U is an Affirmative Action/Equal Opportunity Institution

 

 

’—

LIBRARY
“tomcat: State
University

w v ._ ~ ...l'

 

 

‘—

PLACE IN RETURN BOX to remove this checkout from your record.
T AVOID FINES return on or before date due.

0
I DATE DUE DATE DUE DATE DUE II

 

 

 

 

 

 

 

 

 

 

 

 

 

 

___— #fi!—

I___________I

A VT"?

MSU Is An Atflrmdlve ActiorVEqual Opportunity Institution
cw

 

 

 

 

 

 

 

 

 

 

 

 

 

. __ -,__,_..

PRECISION AUTOMATED MEASUREMENTS
By
Robert Jay Randel

A THESIS

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

MASTER OF SCIENCE

Department of Electrical Engineering

1990

ABSTRACT
PRECISION AUTOMATED MEASUREMENTS

By
Robert Jay Randel

Due to their IEEE-488 GPIB capabilities, the latest series of Fluke/Philips
instruments can be programmed to perform a wide variety of automated
measurements. The Test'l‘eam software package supplied by Philips contains
drivers which make this programming directly possible from Microsoft
QuickBASIC. Techniques and software are developed using these drivers to
automate tedious steady-state measurements. The topics covered are bode
magnitude and phase plots, locating cutofl' (-3 dB) frequencies, measuring gain-
bandwidth-products of op-amps, and finding the second order filter parameters

fl, Ho, and Q0. This software is tested on numerous circuits.

Copyright by
ROBERT JAY RANDEL
1990

ACKNOWLEDGEMENTS

My appreciation to Ken Noren for his support and advice.
Much thanks to Jim MacKay for his assistance and support.
My infinite gratitude to Professors G.M. Wierzba and C.R. MacCluer for their

guidance and assistance in completing this document.

iv

TABLE OF CONTENTS

List of Tables , vii
List of Figures viii
Chapter 1 Introduction to TestTeam Software 1

1. 1 Introduction and Preview of Results, 1

1.2 Choosing a Programming Language, 2

1.3 Hardware Requirements, 3

1.4 Installation and Configuration of TestTeam Software, 4

1.5 Analysis of the Configuration Program, 6

1.6 Programming the Instruments, 7

1.7 Using the TestTeam Drivers, 7

1.8 Summary, 8
Chapter 2 Steady State Analysis of

Passive Networks 9

2.1
2.2
2.3
2.4
2.5
2.6

2.7
2.8
2.9

Introduction, 9

Creating a Bode Magnitude Plot of a Passive Circuit, 9
Speeding Up Operation of This and Other Routines, 12
Creating a Bode Phase Plot, 14

Defects and Potential Improvements, 18

Implementation of Improvements, 19

2.6.1 Access to Raw Data, 19

2.6.2 Correcting for Input Voltage Errors, 19

2.6.3 Disabling the Counter to Prevent Erroneous Data, 20
2.6.4 Correction of Phase Angle, 20 '
2.6.5 Finding a -3 dB Point, 21

The Program BODE5C.BAS, 22

Sample Data from BODE5C.BAS, 26

Summary, 28

Chapter 3 Steady State Analysis of

Active Networks 29
3.1 Introduction, 29
3.2 Enhancements For the Measurement of Active Circuits, 29
3.2. 1 Addition of a Second Voltmeter, 29
3.2.2 Modification of -3 dB Routine for Active Circuits, 34
3.2.3 Measuring the Gain-Bandwidth-Product of an Op-Amp, 34
3.2.4 Aborting the Program, 37
3.3 The Listing of BODE6C.BAS, 38
3.4 Sample Data From BODE60.BAS, 43
3.5 Testing for Linearity, 44
3.6 Automated Measurement of Filter Parameters, 45
3.6.1 Location of a Second -3 dB Point, 45
3.6.2 Calculation of the Center Frequency of a Second-order
Notch Filter, 46
3.6.3 Finding the Parameters of Band-pass Filters, 47
3.7 Accuracy of These Routines, 48
3.8 The Program BODE7C.BAS, 49
3.9 Summary, 56
Chapter 4 Experimental and Theoretical
Verification 57
4. 1 Introduction, 57
4.2 A Diflicult Test for BODE7C.BAS, 57
4.3 Measuring the Gain-Bandwidth-Products, 58
4.4 Testing the Filter, 62
4.5 Theoretical Verification of Measured Results, 65
4.6 Summary, 70
Chapter 5 Conclusions and Future Research 71
5.1 Conclusions, 71
5.2 Future Research, 71
List of References 79

Table 1.1
Table 1.2
Table 1.3
Table 2.1
Table 4.1
Table 4.2
Table 4.3
Table 4.4
Table 4.5
Table 4.6
Table 4.7
Table 4.8

LIST OF TABLES

Directories of TestTeam Files ........................ 4
Variable Names for Instruments ..................... 7
Summary of Important TestTeam Functions ............ 8
Raw Data For High Pass Filter ..................... 28
Gain-Bandwidth—Products of Measured Op-Amps ........ 61
Gain-Bandwidth-Product vs. Time ................... 62
Component Values for the Tow-Thomas Filter .......... 63
Measured Results of Relocated Tow-Thomas Filter ....... 63
Measured Results of Original Tow-Thomas Filter ........ 65
Measured Output Resistance of Op-Amps .............. 68
Parameter Shifting Predicted by Sspice ............... 69
Predicted Parameters and Errors .................... 69

'vii

Figure2.1
Figure2.2
Figure2.3
Figure2.4
Figure2.5
Figure2.6
Figure2.7
Figure3.1
Figure3.2
Figure3.3
Figure3.4
Figure3.5
Figure3.6
Figure3.7
Figure3.8
Figure3.9

Figure 3.10

Figure4.1
Figure4.2
Figure4.3
Figure4.4
Figure5.1
Figure5.2
Figure5.3

LIST OF FIGURES

Low-Pass Filter ................................ 11
Bode Magnitude Plot of Low-Pass Filter .............. 12
High-Pass Filter ................................ 17
Bode Magnitude Plot of High Pass Filter ............. 17
Bode Phase Plot of High Pass Filter ................. 18
Bode Magnitude Plot of High Pass Filter ............. 27
Bode Phase Plot of High Pass Filter ................. 27
Thevenin Equivalent of Function Generator ........... 31
Passive Notch Filter ............................. 31
Incorrect Bode Magnitude Plot for Notch Filter ........ 32
Bode Phase Plot for Notch Filter ................... 33
Correct Bode Magnitude Plot for Notch Filter ......... 33
Inverting Amplifier ............................. 35
Controlled Source Model of Inverting Amplifier ........ 35
Non-inverting Amplifier .......................... 35
Bode Magnitude Plot of Non-inverting Amplifier ....... 43
Bode Phase Plot of Non-inverting Amplifier ........... 44
Relocated Tow-Thomas Filter ...................... 59
Original Tow-Thomas Filter ....................... 60
Bode Magnitude Plot of Relocated Tow-Thomas Filter . . . 64
Bode Phase Plot of Relocated Tow Thomas Filter ....... 65
Oscilloscope Output for Positive Pulse ............... 76
Oscilloscope Output for Triangle Wave ............... 77
Oscilloscope Output for Square Wave ................. 78

CHAPTER 1
INTRODUCTION TO TES'ITEAM SOFTWARE

1.1 Introduction and Preview of Results

One tedious task a modern electrical engineer faces is the testing of a
circuit. Despite advances in computer modelling and simulation, there is no
substitute for building a circuit. Often problems are discovered that did not
show up in the simulation. Many of the tests that are run on these circuits are
very tedious measurements that can take even the most skilled of engineers
many hours. Until recently automation of these measurements was not
possible. However many newer test instruments now include some kind of
computer interface. When properly connected to a computer, the computer can
control the test equipment and perform the measurements for the engineer.
Of course the computer and the test equipment both need to be programmed.
One such set of instruments and software is the newest line of Fluke/Philips
instruments and the Philips TestTeam sofiware package.

TestTeam can be an extremely powerful tool to aid in the automation of
circuit measurements. To unleash this power, however, it is necessary to
program the package. The package is composed of a programming
environment (called LabWindows) and a complete set of drivers which support
many Fluke/Philips instruments. When these drivers are combined with some

programming, they allow the experimenter to automate measurements. The

2

documentation that accompanies this package can be described as sparse at
best. It consists of descriptions of the drivers and an introduction to
LabWindows, but contains no usefirl examples or techniques. The goal of this
document is to indicate an approach to the deve10pment of software to
automate precision measurements using these drivers. In Chapter 2, sample
software is designed for the automated measurements of Bode magnitude and
phase plots and cutoff frequencies of passive networks. This sofizware is tested
on low and high-pass filters. In Chapter 3, the software is expanded to include
active circuits and perform such measurements as gain-bandwidth-products of
op-amps and the second order filter parameters f,, Ho, and Q. In Chapter 4,
the software is utilized to perform measurements to analyze a high-Q band-
pass filter. Complete listings of the developed programs are provided to assist
the user in further development. The examples herein will enable a user to

fully exploit the powerful device drivers included in TestTeam.

1.2 Choosing a Programming Language

The TestTeam drivers and the programming environment are compatible
with two programming languages, Microsoft QuickBASIC and Microsoft C.
This study will use Microsoft QuickBASIC. Although not a widely used
language in the academic community, BASIC is widely used in industry.

BASIC achieved great popularity in the late 197 0’8 because it was included in

3
the ROM of virtually every personal computer of that time. These older

versions of BASIC are best suited to simple tasks. Critics complained that it
is difficult or impossible to write structured, easy-to-follow code since there
were no "while-until" or "repeat-until" commands nor were there subroutines
with parameters. Also, since it is an interpreted language (as opposed to
compiled) it could be too slow for many problems. Microsoft QuickBASIC has
obviated all of these objections. It is a free-form language (requiring no line
numbers) with a compiler, supporting subroutines and functions similar to
Pascal, and has numeric coprocessor support. It would appear to be an ideal
choice for a study whose goal is to make TestTeam more accessible to industry

users.

1.3 Hardware Requirements

The TestTeam software requires an BM PC/XT/AT, IBM PS/2, or
compatible computer with one floppy disk drive, one hard disk drive, 640 KB
of RAM, and the Philips PM 2201 IEEE-488 GPIB Bus Interface. For
graphical output, a graphics card and monitor are also required. This study
employed a 6 MHz Zenith 80286-based computer with monochrome CGA and

an 80287 coprocessor.

4

1.4 Installation and Configuration of TestTeam Software

One should begin by making backup copies of the TestTeam distribution

disks. To install TestTeam onto a hard drive, insert the first diskette, make

floppy drive A the current drive and type

A>setup

and press the Enter key. The program now prompts the user to insert all of

the TestTeam disks for installation. The directories into which the setup

program will place the files are shown in Table 1.1 [1].

Table 1.1 - Directories of TestTeam Files

 

I

 

 

 

 

 

\LW System files
\LW\FONTS Font files required for graphics I.
operations I
\LW\LIBRARY Library files for linking with
standalone programs
I \LW\INCLUDE Include files associated with
I libraries
: \LW\PROGRAMS Source code to sample programs
I \LW\INSTR Instrument modules

The setup program also performs some necessary changes to the

CONFIG.SY S file in the root directory of the hard disk. These include adding

the lines FILES=20 and DEVICE=\LW\GP1B.COM. The former allows a

5
maximum of 20 files to be open simultaneously, while the latter allows the

system to be aware of the IEEE-488 port.

Next one must configure the TestTeam software to the instruments.
The instruments used in this study are the Philips PM 3365 Oscilloscope, PM
5183 Programmable Synthesizer/Function Generator, PM 6666 Programmable
Timer/Counter, and the Fluke 8840A Digital Multimeter. Before we can
connect the instruments to the computer, their IEEE address must be set.
This is done by using the front panel keys (PM 5193, PM 3365) or by setting
DIP switches (PM 6666, Fluke 8840A). Any address can be chosen as long as
there are no conflicts. Using standard IEEE-488 cables, connect each of the
four instruments to the computer’s interface. A typical way to do this is to
"daisy-chain" the instruments to the computer; this makes the detection of a
bad cable easier should that problem arise. Once this is complete turn on the
instruments (making sure there are no front panel connections present) and

type the following commands at the DOS prompt:

c:
cd\lw

contig/a

The last command selects the configuration program, and the la stands for
automatic. The automatic configuration seems to work well. It recognizes the

instruments attached to the GPIB bus and configure itself accordingly.

6
1.5 Analysis of the Configuration Program

One of the files created by the configuration program is called
TESTTEAMBAS. A sample listing is shown below:

REM File : TESTTEAM.BAS

REM Setup file for Philips/Fluke Instrument drivers
REM This will assign logical names to the instruments
REM and initialize the instruments

REM Following include statements can be removed in Interactive

window

REM SINCLUDE: '\lw\instr\generato.inc'
REM SINCLUDE: ’\lw\instr\counter.inc’
REM SINCLUDE: '\lw\instr\multimet.inc'
REM SINCLUDE: '\lw\instr\scope.inc’
REM $INCLUDE: '\lw\instr\general.inc’
DEFINT A-Z

COMMON SHARED /DMM1/ DMMI AS INTEGER
COMMON SHARED /GNR1/ GNRl AS INTEGER
COMMON SHARED /OSC1/ OSC1 AS INTEGER
COMMON SHARED /CNT1/ CNTI AS INTEGER

CALL res.glb

CALL reset.config

CALL config(DMM1,"8840A,A 706,N DMMl")

CALL confithNRl,"PM5193/V2.5,A 707,N GNRl")

CALL config(OSC1,"PM3365/V07V04,A 708,N OSC1")

CALL config(CNT1,"PM6666/22,A 710,N CNTl")

CALL allinittDEFAULT.SET)

IF glb.stat > 1 THEN
PRINT "Error: ",glb.str :REM Print global error string
call res.glb :REM Reset global error status

ELSE

IF glb.stat - 1 THEN
PRINT "Warning: ",glb.str :REM.Print global warning string
call res.glb :REM Reset global error status

END IF

END IF

This file serves as the basis for all applications developed. Its purpose is to
initialize the GPIB bus and the instruments so they will be ready for the

commands.

7
1.6 Programming the Instruments

Within the QuickBASIC environment, the instruments are given integer
variable names consistent with their function. For example the 8840A is given
the variable name dmml % (which stands for Digital Multi-Meter). Ifmore than
one of a given type of instrument is connected, the last digit will increase, e.g.

dmm2%. A summary of these variable names is given in Table 1.2.

Table 1.2 - Variable Names for Instruments

   
   

s...1.v.....1. Name I

 

 

 

 

 

 

Digital Multimeter dmm1%
Function Generator gnr1%
Timer/Counter cnt1%
Oscilloscope osc1% I

1.7 Using the TestTeam Drivers

The real power of the TestTeam package lies in its driver functions.
These drivers, when properly utilized, can be used to configure the
instruments, perform measurements, and display graphs. The QuickBASIC
syntax for these functions is

CALL function.name (...)

where the ellipses stand for whatever parameters are required for the function.

A summary of these functions is provided in Table 1.3. [2]

Table 1.3 - Summary of Important TestTeam Functions

 

 

 

 

 

 

 

 

 

 

 

 

 

.amplitude Sets output amplitude of function generator
. set.function Sets function of instrument (e.g. VOLTS—DC or
‘ VOLTS-AC for multimeter)
I set.speed Sets speed of instrument (note that speed is
. inversely proportional to significant figures)
set.coupling Allows AC or DC coupling to be selected
. set.sensitivity Sets sensitivity of counter
1 measure Triggers instrument and performs measurement
grfreset Clears the graphics display
setxdatatype Sets data type for x-axis (integer, real, etc.)
setydatatype Sets data type for y-axis
grfcuerd Creates x-y plot
grflreset I Resets graphics library
iolocal I Returns instrument to local control

 

setaxname Labels x or y axis

 

1.8 Summary

In this chapter the goals of the document have been outlined. An
introduction to using TestTeam software was given, and Microsoft QuickBASIC
was chosen as the programming language to use. Finally the TestTeam

software was installed, preparing us for what lies ahead.

CHAPTER 2
STEADY STATE ANALYSIS OF PASSIVE NETWORKS

2.1 Introduction

In this chapter routines are designed to employ TestTeam for steady
state passive circuit analysis. In particular these routines automate common
tedious measurements such as Bode magnitude and phase plots, and to

automatically locate a -3 dB point.

2.2 Creating a Bode Magnitude Plot of a Passive Circuit

To use TestTeam to perform a measurement, one starts by envisioning
how to perform that measurement manually, and then translate those steps

into a program. To make a bode magnitude plot by hand, one would:

Connect the circuit to the function generator and multimeter.
Set the function generator to the beginning frequency.

Set the output of the function generator to 1 V rms.

Set the multimeter to read rms ac voltage.

Wait at least 5 t for the circuit to settle.

Measure the output on the voltmeter.

Convert the output to dB.

Increase the frequency.

Ifnot done, go to 5.

PPSP’P‘PS‘F’!‘

10
It would also be wise to choose the frequencies logarithmically if we intend to

make a semilog plot. Translating these steps into QuickBASIC:

CALL set.amplitude (GNR1%, VRMS%, 1.0) ’set generator to 1V RMS
CALL set.function (DMM1%, VOLT.AC%) 'set DMM to read.RMs AC v
pdeS ’points per decade
strt !-1 .5 ' log of starting frequency
stpl-4 . 0 ' log of stopping frequency
for il-strt! to stp! step l/ppd
p-(il-strt!)*ppd 'offset in array
xttp)-10“i! ’actual frequency
CALL set.frequency (GNR1%, x#(p)) ’set freq of generator
CALL measure (DMM1%, y#(p)) 'take voltage measurement
y#(p)-20*log(y#(p))/log(10) 'convert to dB
print x#(p),y#(p) 'print frequency and d3
x#(p)-i! 'store log of frequency
next i!
CALL GrfReset (4) 'set up graphics screen
np-(stpt-strt!)*ppd 'number of points read
CALL GrfCuerD (x#(). y#()' np) 'do x-y'plot/wait for key
CALL GrfLReset (0, 0, 1, 2) 'return to text mode

This program may be run directly from the LabWindows environment,
provided it is added to the configuration program, TESTTEAMBAS. To enter-
the LabWindows program, type the following commands at the DOS prompt:
cd\ 1w
1w

This program is relatively straightforward in its operation. The first two lines
set the function generator to 1 V rms and the multimeter to AC voltage,
respectively. The next three set the number of points to take per decade and
the starting and stopping frequencies. After this comes the loop which takes
the frequencies. Since the FOR-NEXT loop chooses points linearly, if we take .
the antilog of each point, we will have the actual frequency. This is done in
the line x# (p) =1 0 " i . Next the frequency of the function generator is set and

the measurement is taken. Note that the internal settling time of the

11
multimeter (551 ms in VAC mode) [3] is sufiicient for many circuits to settle,

thus no delay is included. Then the measurement is converted to dB (log in
QuickBASIC is to the base e, thus we must divide by log( 10) to convert to base .
10) and the result is displayed. The log of the fi'equency is stored so we can
plot dB vs. log(frequency). Finally the data is plotted using the graphics
library drivers.

A test run of this program was done using the first-order Butterworth
low-pass filter shown in Figure 2.1. The actual values of the components (as
measured by the HP 4284A Precision LCR Meter) are R = 1.0169 k!) for the
resistor, and R3 = 7.6 Q and C8 = 100.11 nF (measured at 4 kHz) for the
capacitor. The generated plot appears in Figure 2.2. This plot appears to be

what would be expected from such a circuit.

 

 

 

R
0——_W\r O
+ +
VIN ;:C VOUT
C ‘ C

Figure 2.1 - Low-Pass Filter

12

 

.. r.

2!!! 2&1 Elsi '51-] an ill 21!)
Figure 2.2 - Bode Magnitude Plot of Low-Pass Filter

2.3 Speeding Up Operation of This and Other Routines

The LabWindows interactive environment, an interpreted subset of
QuickBASIC, is too slow by modern standards. Each measurement and
calculation takes several seconds. For professional lab measurements, one
must employ the full Microsoft QuickBASIC 4.5 compiler. An additional driver
package is then required: the PM2233 Instrument Drivers. This package
allows the interface of compiled QuickBASIC with the necessary drivers to
control the instruments. The installation of these drivers is identical to the
installation of the TestTeam software. It will place its files in the directory
\DRIVERS. The only necessary changes to the program are in the INCLUDE

statements at the beginning of the program. These changes can be seen in the

13
listing of the program BODE3C.BAS in the next section.

It is possible to program directly in the QuickBASIC environment using
these libraries, however it is limited to small programs. Thus the QuickBASIC
editor was used to type the programs, and then they were compiled and linked
from DOS. To automate this procedure, the batch file DOIT.BAT was created.
Its listing is shown below.

bc/o %1;
\qb45\link %1,,@linkme.lnk

The command bc invokes the QuickBASIC compiler, and the / o parameter
tells it to compile stand-alone, i.e. without a run-time module. The % 1
represents the parameter passed to the batch file. Finally the @linkmelnk for
the linker tells it to get its input fi'om the file linkmelnk. The contents of this

file are shown below.

NUL.MAP/NOE/NOD/SEGMENTS:IOOO/STACK:10000,
\LW\LIBRARY\formatio+

\LW\LIBRARY\graphics+

\LW\LIBRARY\gpib qn+\LW\LIBRARY\1wqb1+\qb45\bcom45+
\LW\LIBRARY\ttdrqub+\DRIVERS\drivers+\LW\LIBRARY\lwqbZ:

The first line tells the linker not to create a cross-reference file and to set aside
10 KB for the stack. The other lines tell the linker where to find the necessary
libraries. For example, to compile and link the program BODE3C.BAS, type
at the DOS prompt: ’ .

doit b04034:

and press ENTER. Afier the process is complete, to run the program simply
type bodeac and press ENTER.

14
2.4 Creating a Bode Phase Plot

A method of measuring phase angle is via an oscilloscope by finding a
common point on the two waveforms and measuring the time delay between
them. The ratio of this time delay to the period of the waveform is then
proportional to the phase angle. Symbolically this is

ALL.
T 360°
The problem is obtaining an accurate measure of AT. Fortunately there is a
simple solution. The PM6666 counter has a measurement mode called "TIME
A-B". This mode gives the time delay between a positive slope on channel A
and a positive slope on channel B [4]. Since the frequency (and thus the
period), is known, the calculation of phase follows. The only other
consideration is the coupling of the counter. For low frequencies, below about
500 Hz., the counter should be on DC coupling to avoid errors. Above 500 Hz.
the counter should be set for AC coupling. The QuickBASIC implementation

of this is shown below.

REM File : BODE3C.BAS

REM SINCLUDE: '\lw\instr\generato.inc'
REM SINCLUDE: '\lw\instr\counter.inc'
REM SINCLUDE: '\lw\instr\multimet.inc'
REM $INCLUDE: '\1w\instr\scope.inc'

REM SINCLUDE: '\lw\instr\general.inc'
REM SINCLUDE: '\lw\include\phildecl.inc'
REM SINCLUDE: '\lw\inc1ude\graphics.inc’
DEFINT A-Z

COMMON SHARED ldmml/ dmml AS INTEGER
COMMON SHARED Ignrl/ gnrl as INTEGER
COMMON SHARED /osc1/ oscl AS.INTEGER

l5
COMMON SHARED /cntl/ cntl AS INTEGER
CLEAR ’ ' 2048

DIM SHARED x#(100), y#(100): ph#(100)

CALL getmemIZOOOO)

overlay.memory.size& - 98304

memory.size& - SETMEM(640000)

memory.size& - SETMEM(-overlay.memory.size&)

CLS

CALL res.glb

CALL reset.config

PRINT "Initializing Multimeter..."

CALL config(dmm1, "8840A,A 706,N DMMl")

PRINT "Initializing Function Generator..."

CALL config(gnr1, "PM5193/V2.$,A 707,N GNRl")

PRINT "Initializing Oscilloscope..."

CALL config(oscl, ”PM3365/V07V04,A 708,N OSC1")

PRINT "Initializing Counter..."

CALL config(cnt1, “PM6666/22,A 710,N CNTI”)

PRINT "Setting up defaults...”

CALL allinit(DEFAULT.SET) 'or ACTUAL.SET to leave unchanged

IF Glb.Stat > 1 THEN
PRINT "Error: ", Glb.Str: REM Print global error string
CALL res.glb: REM Reset global error status
ELSE

IF Glb.Stat - 1 THEN
PRINT "Warning: ", Glb.Str: REM Print global warning string
CALL res.glb: REM Reset global error status

END IF

END IF

PRINT
StartOfProg:

CALL set.amplitude(gnrl%, VRMSt, 1!)

CALL set.function(dmm1%, VOLT.AC%)

CALL set.speed(dmm1, low)

CALL set.function(cntl%, TIMEINTERVAL.A.B)
CALL set.coupling(cntl%, chall%, dc%)

CALL set.sensitivity(cnt1%, cha%, .02)
CALL set.sensitivity(cnt1%, chb%, .02)

INPUT "Enter starting frequency ->", strt!

INPUT "Enter ending frequency ->', stp!

INPUT "Enter number of points per decade ->", ppd
PRINT

PRINT "Frequency dB Angle"

PRINT

strt! - LOG(strt!) / LOG(10)
stp! - LOG(stp!) / LOG(10)

FOR i! - strt! TO stp! STEP 1 / ppd
p - (i! - strt!) * ppd
X#(p) - 10 “ 1!
CALL set.frequency(gnr1%, x#(p))
CALL measure(dmm1%, yi(p))

16

IF x#(p) > 500 THEN 'set AC coupling above 500 82
CALL set.coupling(cnt1%, chall%, ac%)

ELSE
CALL set.coup1ing(cntl%, chall%, dc%)

END IF

CALL measure(cnt1%, diff!)

ph#(p) - -di£f# * x#(p) * 360*
IE ph#(p) < -180 THEN

ph#(p) - ph#(p) + 360!
END IF

y#(p) - 20 * LOG(y#(p)) / LOG(10) 'convert to dB

LOCATE CSRLIN - 1

PRINT USING "##.###“““ +###.### +###.###"; x#(p); y#(p);
Ph*(P)

x#(p) - i!
NEXT i!

CALL GrfReset(4)

np - (stp! - strt!) * ppd 'number of points
CALL SetXDataType(4)

CALL SetYDataType(4)

CALL GrfCurv2D(x#(), y#(). np) ’x-y plot of dB vs. log(f)
CALL GrfReset(4) 'clear the graphics display
CALL GrfCuerD(x#()p ph#(), np) 'x-y plot of phase vs.log(f).
CALL GrfLReset(0, 0, 1, 2) ’return to text mode

CLS

PRINT "Again (y/n)?";

a$ - INPUT$(1)

PRINT : PRINT

IF UCASE$(a$) - "Y" THEN GOTO StartOfProg
REM Now return to local control

CALL ioloca1(dmm1%)

CALL ioloca1(osc1%)

CALL iolocal(cnt1%)
CALL ioloca1(gnr1%)

END

REM $DYNAMIC
SUB ReportError
END SUB

The output of this program for a the high-pass R—C circuit in Figure 2.3 is
shown in Figures 2.4 and 2.5. The values of the components used were Cs =
105.48 nF, R5 = 11.9 9 (measured at 400 Hz), and R = 4.028 kn. The frequency

range was from 20 Hz to 50 kHz, taking 9 points per decade.

17

 

7:0

 

 

O o
+ +
VIN R VOUT
O 0

Figure 2.3 - High-Pass Filter

l’s.’

 

E a

“Iii 111 m E] w ill an ill all
Figure 2.4 - Bode Magnitude Plot of High Pass Filter

18

EE‘LEEE

E“

Figure 2.5 - Bode Phase Plot of High Pass Filter

2.5 Defects and Potential Improvements

The program BODE3C.BAS discussed above suffers from several defects.
When measuring a low-pass filter, as the output voltage falls below a certain
level, the counter can no longer lock onto the signal and erroneous data is
taken. Also no access is given to the raw data in case an error is suspected.
In addition the counter may occasionally give a time delay that is ofi‘ by a
period or two, e.g. yielding a phase angle of -362° as opposed to the correct
value of -2°. Still another problem is that the output of the function generator
may not be exactly 1 V RMS. It would also be useful to label the axes of the
graph. Another desirable feature would be to implement a routine that would

locate a -3 dB point.

19
2.6 Implementation of Improvements

2.6.1 Access to Raw Data

Providing the user with access to the raw data accumulated would be a
convenient feature. This way the user could verify the data by hand if an error
is suspected, or even use this data in another program. It would be best to
supply all the calculated quantities as well as the measured values. The
values to be stored in the file are: frequency, dB, phase, measured voltage,

and measured time delay.

2.6.2 Correcting for Input Voltage Errors

When the PM5193 Function Generator is set for 1 V RMS output, the
open-circuit output is not exactly 1 V RMS. This effect does depend upon
frequency somewhat, although if we were to assume a constant error over a
range of fiequencies (which may or may not be accurate depending upon the
range), we could divide all our voltage measurements by a correlation to
approximate the actual transfer function. The optimum solution, of course,
would be to use a second voltmeter, then the transfer fimction would simply
be output divided by input. The former shall be implemented for the time

being, assuming that access to a second voltmeter is not possible. The

20
implementation of the second voltmeter is addressed in Section 3.2. 1.

2.6.3 Disabling the Counter to Prevent Erroneous Data

When the output voltage falls below 10 mV RMS, the counter becomes
unable to lock onto the signal even when it is set for maximum sensitivity. To
solve this problem there are several options: (1) discard phase measurements
when the output falls below the 10 mV threshold (although some value must
be assigned to the array), or (2) we could stop taking phase measurements

below the threshold. The latter shall be implemented.

2.6.4 Correction of Phase Angle

Occasionally the counter gives a time delay that is off by a period or two.
This would result in a phase angle of say -722° or -362° instead of -2° (it
always appears to be off in the negative direction). The simplest method to
correct this is to add 360° to the phase angle as long as it is less that -180°.

Translated to QuickBASIC:

while phttp) < —180
ph#(P) - phttp) + 360!
wend

21
2.6.5 Finding a -3 dB Point

Finding a -3 dB point is a common required measurement. It may
represent the bandwidth or a cutofi‘ frequency of a circuit. By hand it is a
tedious procedure, which makes it a particularly desirable feature to
implement by machine. The method chosen here is the linear interpolation
method. This method assumes a Bode plot has already been performed and
that the data from this plot is stored in an array. The steps followed are
detailed below.

1. Find two points whose magnitudes are "close" to- ‘12/2 ~ 0.70711.
If there are none, inform the user to re-run the bode plot to
include such points.

2. Use the linear interpolation formula (from basic calculus)
flx +Ax)~flx) +f’(x)Ax

where fix), flx+Ax), f(x), and x are known and Ax is the only
unknown. (f(x+Ax) = ‘12/2 and f(x) can be approximated by using
the two points found in step 1). That is, if the two known
frequencies are f1 and f2, and the two magnitudes are t1 and t2,
respectively, then

‘2 Itl
f2 ‘f 1
and the target frequency is

f’(x)-

 

[3:2

f’(x) +f"

 

3. Set the target frequency to the value found in step 2. Read the
output voltage from the voltmeter. If the voltage does not equal
0.70711 V (the 8840A displays 5 decimal points) then use one of

22

the original points and point found in step 2 and return to step
2.

Translating steps 2 and 3 into QuickBASIC:

DO
slope! - (t2! - t1!) / (f2! - f1!)
f0? - (0.70711 - t1?) / slope? + £1!
CALL set.£requency(gnr1%, f0#)
CALL measure (dmm1%, t0#)
IE (f2? - £0!) < (f0# - f1?) THEN
f1¥ - f0¥
tlt - t0#
ELSE
f2# - f0#
t2# - t0#
END IF
LOOP UNTIL t0# - 0.70711

In practice this routine seems to require an average of 6 or 7 tries before it

finds the cutofi' frequency. Also we will want to include this value in the raw

data file.

2.7 The Program BODE5C.BAS

A listing of the finished code with the modifications discussed in

Sections 2.5 - 2.6 and some miscellaneous cosmetic changes is shown below:

REM File : BODE5C.BAS

REM- SINCLUDE: ’\lw\instr\generato.inc’
REM SINCLUDE: '\lw\instr\counter.inc'
REM $INCLUDE: '\lw\instr\multimet.inc'
REM SINCLUDE: '\lw\instr\scope.inc’

REM SINCLUDE: '\1w\instr\genera1,inc’
REM SINCLUDE: '\lw\include\phildec1.inc’
REM SINCLUDE: '\1w\include\graphics.inc’
DEFINT A-Z

COMMON SHARED ldmml/ dmml AS INTEGER
COMMON SHARED lgnrl/ gnrl AS INTEGER
COMMON SHARED /osc1/ oscl AS INTEGER
COMMON SHARED /cnt1/ cntl AS INTEGER

23

CLEAR , , 2048

DIN SHARED X#(500)I Y‘(500)I Ph#(500)

CALL getmethOOOO)

overlay.memory.size& - 98304

memory.size& - SETMEM(640000)

memory.size& - SETMEM(-over1ay.memory.sizeE)

CLS

CALL res.glb

CALL reset.config

PRINT "Initializing Multimeter..."

CALL configtdmml, "8840A,A 706,N DMMl”)

PRINT "Initializing.Function Generator...”

CALL configtgnrl, "PM5193/V2.5,A 707,N GNRl")

PRINT "Initializing Oscilloscope..."

CALL config(oscl, "PM3365/V07V04,A 708,N OSC1”)

PRINT "Initializing Counter..."

CALL config(cnt1, "PM6666/22,A 710,N CNTl")

PRINT "Setting up defaults..."

CALL allinit(DEFAULT.SET) 'or ACTUAL.SET to leave unchanged

IF Glb.$tat > 1 THEN
PRINT "Error: ”, Glb.$tr: REM Print global error string
CALL res.glb: REM Reset global error status
ELSE

IF Glb.$tat - 1 THEN
PRINT "Warning: ", Glb.$tr: REM Print global warning string
CALL res.glb: REM Reset global error status

END IF

END IF

PRINT
StartOfProg:
OPEN "rawdata." FOR OUTPUT AS #1 'save data in file

CALL set.amp1itude(gnr1%, VRMS%, 1!)

CALL set.function(dmml%, VOLT.AC%)

CALL set.speed(dmm1, low)

CALL set.function(cnt1%, TIMEINTERVAL.A.B)
CALL set.coup1ing(cnt1%, chall%, dc%)

CALL set.sensitivity(cnt1%, cha%, .02)
CALL set.sensitivity(cnt1%, chb%, .02)

INPUT "Enter starting frequency ->", strt!

INPUT ”Enter ending frequency ->", stp!

INPUT ”Enter number of points per decade ->", ppd

INPUT ”Enter actual RMS output of function generator ->”, ff?
IF ff! - 0 THEN ff? - 1 'assume 1 if they hit return

PRINT

PRINT "Frequency dB Angle"

PRINT

strt! - LOG(strt!) / LOG(lO)

stp! - LOG(stp!) / LOG(lO)

np - (stp! - strt!) * ppd 'number of magnitude points
nph - np 'number of phase points

FOR 1! - strt! TO stp! STEP 1 / ppd

24

p - (i! - strt!) * ppd

x#(p) - 10 “ i!

CALL set.frequency(gnr1%, xttp))
CALL measure(dmm1%, dvmt)

dvmi - dvmi / ff#

IF dvmt > .01 AND nph - np THEN ’take phase reading only if
voltage
'is greater than 10 mv RMS
IF xttp) > 500 THEN
CALL set.coup1ing(cnt1%, chall%, ac%)
ELSE
CALL set.coup1ing(cntl%, chall%, dc%)
END IF

CALL measure(cntl%, difft)
phitp) - -diff# * x#(p) * 360i

WHILE ph#(p) < -180
phttp) - ph#(p) + 360!

WEND
ELSE
IF nph - np THEN nph - p
END IF
y#(p) - 20 * LOG(dvmt) / LOG(lO) ’convert to dB

LOCATE CSRLIN - 1

PRINT USING "##.###“‘““ +###.### +###.###": xttp); yttp);
ph#(p)

PRINT #1, USING "#§.§t#“‘“ +§##.### ‘+###.##i -+###.#t#
##.####“‘“‘": x#(p): yitp): phttp): dvmt; diff!

x#(p) - i!
NEXT i!

PRINT
'Now look for -3 dB point

targeti - .70711# / ff#
1 - 1
WHILE ABS(EXP(LOG(10) * y#(i) / 20) - target#) / target¥ > .1 AND 1
< np
i - i + 1
WEND

IF i - np THEN
PRINT ”Couldn't find -3 dB frequency... run again with different
points"
PRINT #1, ""
PRINT #1, ”Couldn’t find -3 dB frequency... run again with
different points"
ELSE
flt - 10 ‘ x#(i - 1) 'starting frequencies
f2! - 10 ‘ x#(i)
t1! - 10 ‘ (y#(i - 1) / 20) 'starting target values
t2# - 10 “ (ytti) / 20)
DO
slope} - (t2# - t1!) / (f2! - fl!)
f0! - (target! - t1!) / slope! + £1! ,
PRINT USING ”Locating -3 dB frequency: #####.##": f0!
LOCATE CSRLIN - 1
'now find to
CALL set.frequency(gnr1%, f0!)

25

CALL measuretdmml%, t0!)

IF (f2# - f0#) < (f0! - flt) THEN
f1! - f0#
t1} - t0#

ELSE
f2! - f0#
t2! - t0!

END IF

LOOP UNTIL t0# - target?

PRINT USING "The cutoff frequency is ##i##.## Hz.": f0!

PRINT #1, ""

PRINT #1, USING "The cutoff frequency is #####.## 32.”; f0!
END IF

WHILE INKEY$ - ”"z WEND

CLOSE #1 'output raw data file

CALL GrfResett4)

CALL SetAxGridVist-l, 1) 'enable grid lines for both
axes

CALL SetAxAutot-l, 21) 'auto-scale both axes every
time

CALL SetXDataTYPe(4)
CALL SetYDataTYP6(4)

CALL SetAxNametO, ”log(f)") 'set name of x-axis

CALL SetAxName(1, ”dB') 'set name of y-axis

CALL GrfCurv2D(x#(), y#(), np) 'x-y plot of dB vs. log(f)
CALL GrfReset(4) 'clear the graphics display
CALL SetAxName(1, "Phase Angle”) 'set name of y-axis

CALL GrfCurv2D(x#(), phtt), nph) 'x-y plot of phase vs. log(f)
CALL GrfLReset(0, 0, 1, 2) 'return to text mode

CLS

PRINT "Again (y/n)?":

as - INPUTS (1)

CLS

IF UCASE$(a$) - "Y" THEN GOTO StartOfProg

REM Now return to local control

CALL iolocal(dmm1%)
CALL iolocaltosc1%)
CALL iolocaltcnt1%)
CALL iolocaltgnr1%)

END
REM $DYNAMIC

SUB ReportError
END SUB

26
2.8 Sample Data from BODE5C.BAS

The program was run for the high pass circuit shown in Figure 2.3. The
resulting Bode plots are shown in Figures 2.6 - 2.7, and the raw data is shown
in Table 2.1. The actual values of the components are R = 3.801 kn, R8 =
12.120, and Cs = 105.31 nF (measured at 400 Hz.). The data was collected
using 10 points per decade over a frequency range of 20 Hz to 50 kHz. The
cutoff frequency was found to be 416.59 Hz by the program. Note that this is
not the same as the theoretical cutofl‘ frequency (396.34 Hz) but rather is the
frequency at which the voltmeter read 0.70711. The error is due to the
presence of R8 of the capacitor and the output impedance of the function

generator. Correction of these errors is dealt with in the next chapter.

can.

CID-“3:9 OVI’!‘

27

9:;

 

D;
C:

I
31h] 1!! an ill 1!.) Lil! Ell ill SI!)
. 09
Figure 2.6 - Bode Magnitude Plot of High Pass Filter

EEEE

EE

& ill '5!!! 2m 1% ill 21!] 211 an
09
Figure 2.7 - Bode Phase Plot of High Pass Filter

28

Table 2.1 - Raw Data For High Pass Filter

2.000D+01 -26.225 +82.968 +0.049 3.8477D-02
2.518D+01 -24.164 +83.496 +0.062 3.0505D-02
3.1?0D+01 -22.159 +82.968 +0.0?8 2.4277D-02
3.991D+01 -20.16? +82.852 +0.098 1.9292D-02
S.024D+01 -18.18? +81.544 +0.123 1.5397D-02
6.325D+01 -16.225 +80.142 +0.154 1.2292D-02
?.962D+01 -14.288 +78.08? +0.193 9.8352D-03
1.002D+02 -12.386 +75.480 +0.240 7.8846D-03
1.262D+02 -10.S43 +72.680 +0.29? 6.3246D-03
1.589D+02 -8.??S +68.233 +0.364 5.1016D-03
2.000D+02 -?.120 +63.631 +0.441 4.1162D-03
2.518D+02 -5.613 +58.315 +0.524 3.3283D-03
3.1?0D+02 -4.294 +52.302 +0.610 2.6965D-03
3.991D+02 -3.191 +45.??9 +0.693 2.18730-03
5.024D+02 -2.315 +38.062 +0.?66 1.?801D-03
6.325D+02 -1.653 +32.418 +0.82? 1.4388D-03
7.962D+02 -1.1?4 +26.988 +0.8?4 1.1618D-03
1.002D+03 -0.842 +22.121 +0.908 9.3633D-04
1.262D+03 -0.616 +18.199 +0.932 7.5239D-04
1.589D+03 -0.465 +14.543 +0.948 6.0404D-04
2.000D+03 -0.366 +11.691 +0.959 4.8376D-04
2.518D+03 -0.299 +9.530 +0.966 3.86650-04
3.170D+03 -0.255 +7.596 +0.9?1 3.08820-04
3.99ID+03 -0.224 +6.383 +0.9?S 2.4615D-04
5.024D+03 -0.203 +5.30? +0.97? 1.96120-04
6.325D+03 -0.189 +4.190 +0.979 1.5627D-04
7.962D+03 -0.177 +3.380 +0.980 1.2442D-04
1.002D+04 -0.168 +2.?31 +0.981 9.900?D-05
1.262D+04 -0.159 +2.1?3 '+0.982 ?.8?6?D-05
1.589D+04 -0.150 +1.?01 +0.983 6.2649D-05
2.000D+04 -0.140 +1.282 +0.984 4.9822D-OS
2.518D+04 -0.138 +0.954 +0.984 3.9611D-05
3.1?0D+04 -0.14? +0.?95 +0.983 3.1478D-05
3.991D+04 -0.154 +0.?31 +0.982 2.5009D-05
The cutoff frequency is 416.59 Hz.

2.9 Summary

In this chapter TestTeam was utilized to analyze passive networks.
Routines for creating Bode magnitude and phase plots were developed, and an
algorithm for locating a -3 dB point was presented. These programs were

tested as they were developed and sample data was provided.

CHAPTER 3
STEADY STATE ANALYSIS OF ACTIVE NETWORKS

3.1 Introduction

In the previous chapter we developed techniques to analyze passive
networks. Now techniques are developed to measure active circuits. The
topics covered include gain-bandwidth-products of op-amps and the second

order filter parameters ft, Ho, and Q0.

3.2 Enhancements for the Measurement of Active Circuits
3.2.1 Addition of a Second Voltmeter

There are many circuits to which the 50!) output impedance of the PM
5193 function generator is significant. That is, for the one-port network shown
in Figure 3.1, V5' is significantly less than Vs. Until now, we have been
assuming that Vs a V8, which can result in significant errors. Consider the
passive notch filter shown in Figure 3.2 with component values R=116.9 0, C3
= 206.2 nF, Rec = 0.87 9, L3 = 907.34 uH, Ra = 12.45 9 (measured at 10 kHz).
The output of the program as it is now is shown in Figures 3.3 - 3.4. The data

was collected using 80 points per decade over a frequency range of 1 kHz to

'29

30
100 kHz. The magnitude plot, however, is incorrect. At the center (notch)

frequency, the input impedance of this circuit is R+R3L+Rsc = 130.22 Q. Thus
assuming the 50 52 output impedance of the fimction generator is purely

resistive,

Vg= 13022 (1 V)=0.723 V
130.22+5o

 

which is already nearly 3 dB down. The best solution to this problem is to add
a second voltmeter to measure Vs’. As stated earlier, as long as the IEEE
address is different, there will be no conflict on the bus. Afier re-running the

configuration program, a new TESTTEAMBAS file is generated.

REM File : TESTTEAM.BAS

REM Setup file for Philips/Fluke Instrument drivers
REM This will assign logical names to the instruments
REM and initialize the instruments

REM Following include statements can be removed in Interactive

window

REM SINCLUDE: '\lw\instr\generato.inc’
REM SINCLUDE: '\lw\instr\counter.inc'
REM SINCLUDE: '\lw\instr\multimet.inc'
REM $INCLUDE: '\1w\instr\scope.inc'
REM $INCLUDE: '\lw\instr\general.inc’
DEFINT A-Z

COMMON SHARED /DMM1/ DMMl AS INTEGER
COMMON SHARED /DMM2/ DMMZ AS INTEGER
COMMON SHARED /GNR1/ GNRl AS INTEGER
COMMON SHARED /OSC1/ OSC1 AS INTEGER
COMMON SHARED /CNT1/ CNTl AS INTEGER

CALL res.glb

CALL reset.config

CALL config(DMM1,"8840A,A ?05,N DMMl”)

CALL config(DMM2,"884OA,A 706,N DMMZ")

CALL confithNRl,"PM5193/V2.5,A 707,N GNRl")

CALL config(OSC1,"PM3365/V07V04,A 708,N OSC1")

CALL configtCNTl,"PM6666/22,A 710,N CNTl")

CALL allinit(DEFAULT.SET)

IF glb.stat > 1 THEN
PRINT "Error: ",glb.str :REM Print global error string
call res.glb . :REM Reset global error status

31

ELSE
IF glb.stat - 1 THEN
PRINT "Warning: ",glb.str :REM Print global warning string
call res.glb :REM Reset global error status
END IF
END IF

There are only three difi‘erent lines in this code. First is the declaration of the
new variable ('11ng . TestTeam recognizes the two meters as dmml and dmmz .
Then in the two CALL config lines the proper IEEE address is assigned to
each meter (address 5 is assigned to the new meter). These three changes may

be made directly to the BODE5C.BAS program.

 

+, 1-Port
3 VS Network

 

 

 

 

Figure 3.1 - Thevenin Equivalent of

 

Function Generator
R
O——’VVV O
+ +
L
V I N VOUT

_ C _
o ’f‘ 0

Figure 3.2 - Passive Notch Filter

32
The only necessary changes to the code to utilize the two meters are to

take the ratio of the two measurements instead of taking only one
measurement. The correct magnitude plot for the notch filter is shown in
Figure 3.5.

Another advantage of using a second voltmeter is increased bandwidth.
The upper frequency limit for the 8840A is 100 kHz. When using a second
meter though, if its measurements roll ofi' at the same rate as the first meter,
their ratio will still be correct. For the set of instruments used, an efi‘ective

increase in bandwidth to 400 kHz within 1% (0.1 dB) was observed.

5::

ah!!!” mmmlm&)mm ll-l mm
09
Figure 3.3 - Incorrect Bode Magnitude Plot for Notch Filter

33

[L

O—Iflap temps-01:

1% S!!! mmmlm&)mm an 113 all
, 09
Figure 3.4 - Bode Phase Plot for Notch Filter

EEEEEEBEEE

3% ill I!!! mmlzrg)mm II! was
09
Figure 3.5 - Correct Bode Magnitude Plot for Notch Filter

34
3.2.2 Modification of .3 dB Routine for Active Circuits

In previous sections we have found the -3 dB point with respect to 1 V.
For many circuits this may not be true. Let us examine a one-pole active low-
pass filter. In this case we will measure the DC gain of the circuit, then find
the -3 dB point with respect to it. Of course this will not work for high-pass
or band-pass filters since their DC gain will be small. A solution to this

problem is presented in section 3.6.3.

3.2.3 Measuring the Gain-Bandwidth-Product of an Op-Amp

Measuring the gain-bandwidth—product of an op-amp is another common
required measurement. However certain circuit configurations make the
measurement easier than others. Consider the inverting amplifier in Figure
3.6. By inserting a controlled source model of the op-amp (Figure 3.7) it can
be shown that
V _ _ Ii!2 1

TI: TI: _1_R,+R,' . (3.1)
A4,. R1

 

1+

 

35

R2

V10-—W\r— }
——-OV2

Figure 3.6 - Inverting Amplifier

 

 

 

 

V10—MN—

 

5I|

0 V2

Figure 3.7 - Controlled Source
Model of Inverting Amplifier

""1.

V e
IN —0

—~w—~

R2
R, l

Figure 3.8 - Non-inverting
Amplifier

 

 

36
Likewise for the non-inverting amplifier (Figure 3.8) it can be shown that

V: R2+R1 1
V, R, 1 R,+R,' (3.2)

1+...
A4... 31

 

 

Note that both of these configurations (and indeed all one op-amp circuits [5])

have a transfer function of the form

 

1
-K
F(s) __ 1

1+ 1 ' (3.3)

Ade

 

For low to mid frequencies, the open-loop difi'erential mode gain A,“ is
not constant, rather it acts as a one-pole transfer function. A one-pole model

for Ad... is given by

 

Adm- A°‘°°. (3.4)
8+0)o

The frequency (on is the frequency at which the gain is down by 3 dB. From
the Fairchild data sheet for the 741 op-amp, A0 is 200,000 and f, is 5 Hz. The
product Aof0 is the gain-bandwidth-product. For frequencies much greater than
1;, Ah can be approximated as

A°‘°°. (3.5)
8

Ad,"-

 

Inserting this expression into Equation 3.3 yields

37

F(s)-K 1 .
1+ 8 .1;

A000 B
and making the substitution 3 = jm gives

 

(3.6)

 

pom-K 1 ,
1&2.

 

(3.7)

where

" 1 (3.8)

Note that if K = 1/B, then the product K-fe = GBP. The non-inverting
amplifier has this property; the inverting amplifier does not. Thus to find the
gain-bandwidth-product of an opoamp, configure it as a non-inverting amplifier.
Then use BODE6C.BAS to find the -3 dB frequency of the circuit. The product
of the DC gain and the cutofi' frequency is then the gain-bandwidth-product.
This is true as long as the one-pole approximation of Adm holds. Thus K must

be chosen large enough that the second pole of Adm does not efi'ect the results.

3.2.4 Aborting the Program

At certain times it may be desirable to abort operation of the program.

Perhaps the circuit is not responding as expected, the wrong number of points

38
per decade was selected, or a different sweep is desired. It would be

convenient to simply press the ESCape key to stop operation of the program.
This feature will also be implemented below.

3.3 The Listing of BODE6C.BAS

A listing of the program as it stands now is shown below.

REM File : BODE6C.BAS

REM Support for second voltmeter added
REM and support for active circuits
REM Following include statements can be removed in Interactive
window'

REM $INCLUDE: '\lw\instr\generato.inc’
REM $INCLUDE: '\lw\instr\counter.inc'

REM $INCLUDE: '\lw\instr\multimet.inc'
REM $INCLUDE: '\lw\instr\scope.inc'

REM $INCLUDE: ’\lw\instr\general.inc’

REM $INCLUDE: '\1w\include\phildecl.inc'
REM $INCLUDE: '\lw\include\graphics.inc'
DEFINT A-Z

COMMON SHARED ldmmll dmml AS INTEGER
COMMON SHARED /dmm2/ dmm2 AS INTEGER
COMMON SHARED /gnr1/ gnrl As INTEGER
COMMON SHARED loscll oscl AS INTEGER
COMMON SHARED /Cnt1/ Cntl AS INTEGER

CLEAR , , 2048

DIM SHARED x#(500), y#(500). PhttSOO)

CALL getmem(20000)

overlay.memery.size& - 98304

memory.size& - SETMEM(640000)

memory.size& - SETMEM(-overlay.memory.size&)

CLS

CALL res.glb

CALL reset.config

PRINT "Initializing Multimeters..."

CALL config(dmml, ”8840A,A ?05,N DMMl“)

CALL config(dmm2, "8840A,A 706,N DMMZ")

PRINT "Initializing Function Generator..."
CALL config(gnrl, ”PMSl93/V2.5,A ?0?,N GNRl")
PRINT "Initializing Oscilloscope..."

CALL config(oscl, "PM3365/V0?V04,A 708,N OSC1")
PRINT "Initializing Counterr..”

39

CALL config(cntl, "PM6666/22,A ?10,N CNTl")
PRINT "Setting up defaults..."
CALL a11init(DEFAULT.SET) ’or ACTUAL.SET to leave unchanged
IF Glb.$tat > 1 THEN
PRINT "Error: ”, Glb.$tr: REM Print global error string
CALL res.glb: REM Reset global error status
ELSE
IF Glb.$tat - 1 THEN
PRINT "warning: ", Glb.$tr: REM Print global warning string
CALL res.glb: REM Reset global error status
END IF
END IF

PRINT

StartOfProg:
OPEN "rawdata." FOR OUTPUT AS #1 'save data in file

CALL set.function(dmm1%, volt.ac%)

CALL set.function(dmm2%, volt.ac%)

CALL set.speed(dmml, low)

CALL set.speed(dmm2, low)

CALL set.function(cnt1%, TIMEINTERVAL.A.B)
CALL set.coupling(cnt1%, chall%, dc%)

CALL set.sensitivity(cnt1%, cha%, .02)
CALL set.sensitivity(cnt1%, chb%, .02)

INPUT "Enter starting frequency ->", strt!

INPUT ”Enter ending frequency ->", stp!

INPUT "Enter number of points per decade ->", ppd

INPUT "Enter desired RMS function generator voltage (default 1 V)
->", gnv!

IF gnv! - 0 THEN gnv! - 1

CALL set.offset(gnr1, 0%)
CALL set.amplitude(gnr1, vrms, gnv!)

PRINT
PRINT "Frequency dB Angle"
PRINT

strt! - LOG(strt!) / LOG(10)

stp! - LOG(stp!) / LOG(10)

np - (stp! - strt!) * ppd 'number Of magnitude points
nph - np 'number Of phase points

FOR i! - strt! TO stp! STEP 1 / ppd
p - (i! - strt!) * ppd
x#(p) - 10 ‘ i!

IF INKEY$ - CHR$(27) THEN
PRINT
PRINT ”Run aborted..."
PRINT
PRINT #1, "”
PRINT #1, "Run aborted..."
CLOSE
GOTO Again

END IF

40

CALL set.frequency(gnr1%, x#(p))
CALL measure(dmm1%, dvmli)

CALL measure(dmm2%, dvm2i)

dvm! - ABStdva! / dvmli)

'input
'output
'find transfer fn output/input

IF dvm2# > 0 AND nph - np THEN
output

'take phase reading only if

'is greater than 10 my RMS
IF x#(p) > 500 THEN
CALL set.coupling(cnt1%, Chall%, ac%)
ELSE
CALL set.coup1ing(cnt1%, Chall%, dc%)
END IF

CALL measure(cntl%, diff!)

ph#(p) - -diff# * x#(p) * 360*
WHILE ph#(p) < -180
ph#(p) - phttp) + 360*
WEND
ELSE
IF nph - np THEN nph - p
END IF

y#(p) - 20 * LOG(dvmi) / LOG(lO) 'convert to dB

LOCATE CSRLIN - 1
+##§.### +§§#.#§#”; x§(p):

PRINT USING "##.##§““
ph#(p)

PRINT #1, USING ”##.###““““ +###.i## +t##.###
##.##i#““‘": x#(p): yttp): ph#(p): dvmi: difft

x#(p) - i!
NEXT i!

Yttp):
+###.###

PRINT

'Now

CALL
CALL

CALL
CALL

CALL
CALL

hdci

CALL
CALL
CALL
CALL

find DC gain

'turn Off AC voltage from gen
'turn on DC volts from gen

set.amp1itude(gnr1, vrms, 0)
set.offset(gnr1, gnv!)

set.function(dmm1, volt.dc)
set.function(dmm2, volt.dc)

measure(dmm1, dvml!)
measure(dmm2, dvm2t)
- dvm2§ / dvmli 'DC gain (NOT dB)

'return DMM to its original
’state

set.offset(gnr1, Oi)
set.function(dmm1, volt.ac)
set.function(dmm2, volt.ac)
set.amplitude(gnr1, vrms, gnv!)

PRINT

PRINT #1,

PRINT USING "DC Gain - i#§.###"; hdci:

IF hdc’ <> 0 THEN PRINT USING ' - ##.##i dB": 20 * LOG(ABS(hdc§)) /
LOG(IO)

PRINT #1, USING ”DC Gain - ###.###": hdci:

IF hdci <> 0 THEN PRINT

#1, USING " - ##.### dB": 20 *

LOG(ABS(hdc#)) / LOG(lO)

41

'Now look for -3 dB point from.left side

PRINT
PRINT #1, ”"

target# - hdc? / SQR(2#)
IF target! < .001 THEN

PRINT "DC Gain is too low - Cannot sweep for -3 dB point from
left"

PRINT #1, "DC Gain is too low - Cannot sweep for -3 dB point from
left"

GOTO graphit
END IF
i - 1
WHILE ABStEXP(LOG(10#) * ytti) / 20%) - targett) / target! > .1# AND
i < np

i - i + 1
WEND

IF 1 - np THEN
PRINT "Couldn't find -3 dB frequency... run again with different

points"
PRINT #1, ""
PRINT #1, "Couldn't find -3 dB frequency... run again with
different points"
ELSE
flt - 10 “ x#(i - 1) 'starting frequencies
f2! - 10 “ x#(i)
t1! - 10 ‘ (y#(i - 1) / 20) 'starting target values
t2? - 10 “ (y#(i) / 20)
DO

slope# - (t2t - t1!) / (f2! - f1!)

f0! - (target# - t1!) / slope! + f1#

PRINT USING "Locating -3 dB frequency: #####.##”; f0#
LOCATE CSRLIN - 1

IF INKEYS - CHR$(27) THEN .
PRINT "Location of -3 dB frequency aborted... . "
PRINT #1, "”

PRINT #1, "Location of -3 dB frequency aborted..."
GOTO graphit
END IF

'now find to
CALL set.frequency(gnr1%, f0!)
CALL measure(dmm1, dvmlt)
CALL measure(dmm2, dvm2#)
t0! - dvm2§ / dvml!
IF (f2! - £0!) < (f0# - f1!) THEN
fli - f0!
t1? - t0!
ELSE
f2! - f0!
t2! - t0#
END IF
hgs - 100000! * t0! / hdci
LOOP UNTIL hg& - 70711

PRINT USING ”The cutoff frequency is #####.#t Hz.'; f0!
PRINT USING "GBP - ##§!#####.## Hz"; f0! * hdct

PRINT #1, USING "The cutoff frequency is #####.## Hz.'; f0#
PRINT #1, USING "GBP - ###}#####.## Hz"; f0! * hdct

42

PRINT #1, "Actual cutoff ratio - ": t0# / hdc#
END IF
graphit:
CLOSE #1 'output raw data file

CALL set.amp1itude(gnr1, vrms, 0) 'turn off AC voltage from gen
'to facilitate changing components

PRINT
PRINT "Press any key to see graphs...";

WHILE INKEYS ' "": WEND

CALL GrfReset(4)
CALL SetAxGridVis(-1, 1)
axes

'enable grid lines for both

CALL
time
CALL
CALL
CALL
CALL

SetAxAutO(’1, 21)

SetXDataType(4)
SetYDataType(4)
SetAxName(0, "log(f)")
SetAxName (1, "dB")

'auto-scale both axes every

'set name of x-axis
'set name Of y-axis

CALL GrfCurv2D(x#(), y#(). np) ’x-y plot of dB vs. log(f)
CALL GrfReset(4) ‘ 'clear the graphics display
CALL SetAxName(1, "Phase Angle”) 'set name of y-axis

CALL GrfCurv2D(x#()p ph#(), nph) 'x-y plot of phase vs. log(f)
CALL GrfLReset(0, 0, 1, 2) ’return to text mode

CLS

Again:

PRINT "Again (y/n)?";

a$ - INPUT$(1)

CLS

IF UCASE$(a$)

REM

CALL
'CALL
CALL
CALL
CALL

END

Now return to local control

ioloca1(dmml%)
iolocal(dmm2%)
iolocal(osc1%)
ioloca1(cnt1%)
iolocal(gnr1%)

REM SDYNAMIC
SUB ReportError
END SUB

- "Y" THEN GOTO StartOfProg

43
3.4 Sample Data from BODE6C.BAS

As stated earlier, one useful application of this program is to measure
the Gain-Bandwidth-Product of an op-amp. A non-inverting amplifier (Figure
3.8) is configured using an LM741 op-amp with R1 = 1 k!) and R2 = 47 k9
(nominal values). The resulting Bode plots are shown in Figures 3.9 - 3.10.
Unfortunately, the bandwidth of the meters is only 400 kHz, since by
examining the phase plot there appears to be a second pole around 4 MHz.
The DC gain was found to be 48.01, with a GBP of 828.3 kHz. The Fairchild
data sheet for the 7 41 lists a GBP 'of 1 MHz, however this characteristic varies

widely from sample to sample. Thus the results are considered acceptable.

 

I

h '5!!! ll] m 31-1619 ii! an: all m
09

Figure 3.9 - Bode Magnitude Plot of Non-inverting Amplifier

 

til
31
P
h
a!!!)
S
e
:19
(t
0
gm]
1
e

“E! as as '5!!! fling!!! 1&1 a!!! 2!] m
09
Figure 3.10 - Bode Phase Plot of Non-inverting Amplifier

3.5 Testing for Linearity

The measurements we have been taking are only valid when performed
on linear systems. If a circuit is not behaving linearly, the measurements may
be worthless. One possibility is that an op-amp in a circuit has saturated. A
method of checking for linearity is to reduce the input by some ratio (perhaps
50%) and check that the output is also reduced by the same ratio. If not, a
warning to the user is displayed. The QuickBASIC code to add to the program

to do this is given below.

CALL set.amplitude(gnr1, vrms, gnv! / 2!)

CALL measure(dmml, dvmlfl

CALL measure(dmm2, dvm2#)

IF Assmvair / dvmli - dvmi 7 2+) / (dvmt / 2:) > .01 THEN

45

PRINT USING "Warning - Linearity Check Failed at f - ##.####"“"":
pr)

LOCATE CSRLIN - 1

PRINT #1, USING "Warning - Linearity Check Failed at f -
#§,*§##eecc"; pr)
END IF
CALL set.amplitude(gnr1, vrms, gnv!)

It may also be desirable to do this only once every five or ten points, since it
slows down the program considerably. Due to the delays imposed by this

routine, it is not included in any other listings.

3.6 Automated Measurement of Filter Parameters

Another tedious measurement that would be useful to implement is the
measurement of the second order filter parameters f,, Ho, and Q0. Finding Q0
of a high-Q filter by hand is particularly dificult. When properly programmed,

however, it becomes as simple as the touch of a button.

3.6.1 Location of a Second -3 dB point

If the program BODEGCBAS was run on a notch filter, it would locate
the -3 dB fiequency to the left of the notch. This is because we begin our
sweep from the left. If we were to sweep for a second -3 dB point from the
right, we would be able to calculate the bandwidth of this filter. One question

that rises immediately is what to find this second -3 dB point with respect to:

46
the DC value or the highest fi'equency point taken. There are some notch

filters that do not return to the DC gain to the right of the notch. We will
assume that all notch filters swept by this program do return to the DC gain.
To ensure compatibility of this program with other filters, we will allow the

user to look for 0, 1, or 2 cutofi‘ frequencies.

3.6.2 Calculation of the Center Frequency of a Second-order Notch
Filter

If the notch filter is a second-order filter, then the center frequency can

be calculated by the formula

fcgfi (3.9)

where f, and f, are the two cutofi‘ fi'equencies, and the Q of the filter can be

found by

f
,=._¢_. (3.10)
Q f2 ‘f1

Calculating f; is certainly faster than searching for the center of the notch, and
is sufficiently accurate for many applications. In the next section we will
develop a technique for locating the peak of a band-pass filter; if desired that

routine could be altered to locate the center of the notch.

47
3.6.3 Finding the Parameters of Band-pass Filters

Locating the cutofi' frequencies of a band-pass filter is much more
difiicult than it was for the notch filter. The hardest part is finding the center
fi-equency. Unlike the notch filter, where we could use the DC gain as the
reference for the -3 dB frequencies, there is no reference to begin with. Thus
the only way to find the cutoff frequencies is to first find the peak of the filter.
An algorithm to find the peak is detailed below.

Sweep the bode plot as before.

Find the maximum value in the array y#. This will be the

starting point. Call this point Q.

3. Set the variable Alogf = 0.1. This will be the logarithmic window
in which we will look.

4. Measure the voltage at the two endpoints of the window,

P!"

f13=100°l fM

5. If the voltage at either f1 or f2 is larger than at f0, then set f, to
the frequency at which the voltage is highest, and go to step 4.
6. Otherwise divide Alogf by 2.
7. If Alogf > 0.00001 (or some acceptable tolerance), go to step 4.
Otherwise, f, = fo is the center frequency.
Once we have found )2. we can find H(, by measuring the output voltage at that
frequency, and we can find the cutoff frequencies relative to H0 in the same
manner as we have done all along. The formula given for Q0 in Equation 3.10
is still valid, and now we have all the filter parameters. Note that once these
modifications are made to the program, it can be used to locate the cutoff

frequency of a high-pass filter by sweeping the bode plot to the highest

48
allowable frequency (400 kHz for the set of instruments used), then finding one

cutofi' frequency with respect to the peak output.

3.7 Accuracy of These Routines

The accuracy of these routines are limited mainly by the accuracy of the
instruments used. For a Fluke 8840A DMM calibrated within one year in True
RMS AC volts mode, the worst case accuracy (fi'om 50 - 100 kHz) is given by
the manual to be :(0.5% + 400 counts). In the 2 V range (full scale reading
1.99999 V), if the voltmeter reads 1 V, this says that the actual voltage lies in
the range

0.991 S V 5 1.009.
So the worst possible readings of a transfer function of 1 when using the ratio

of two meters are

flame

0.991

and

L991=oaaa
1.009

Based upon the agreement of theoretical and experimental data presented in
chapter 4, however, it appears that these numbers may be quite pessimistic.

The program does appear to be quite consistent in its results, within one count

49
at the fourth significant figure. Henceforth all measured values are reported

to four significant figures.

3.8 The Program BODE7C.BAS

A Hating of the program BODE7C is given below. The program will be

tested in the next chapter.

REM File : BODE7C.BAS

REM Added ability to search for two -3 dB points

REM and calculate Ho, 00 for band-pass & notch filters
REM Following include statements can be removed in Interactive
window

REM $INCLUDE: '\lw\instr\generato.inc'

REM . $INCLUDE: '\lw\instr\counter.inc’

REM $INCLUDE: '\1w\instr\multimet.inc'

REM $INCLUDE: ’\lw\instr\scope.inc'

REM $INCLUDE: '\lw\instr\genera1.inc'

REM $INCLUDE: '\lw\include\phildecl.inc'

REM $INCLUDE: '\lw\include\graphics.inc'

DEFINT A-Z

COMMON SHARED /dmml/ dmml AS INTEGER
COMMON SHARED /dmm2/ dmmZ AS INTEGER
COMMON SHARED /gnr1/ gnrl AS INTEGER
COMMON SHARED /Oscl/ oscl AS INTEGER
COMMON SHARED Icntll cntl AS INTEGER

CLEAR , , 2048

DIM SHARED x#(500), y#(500), Ph#(500)

CALL getmem(20000)

overlay.memery.size& - 98304

memery.size& - SETMEM(640000)

memory.size& - SETMEM(-overlay.memery.sizes)

CLS

CALL res.glb

CALL reset.config

PRINT "Initializing Multimeters..."

CALL config(dmml, “8840A,A ?05,N DMMl")

CALL config(dmm2, "8840A,A 706,N DMMZ")
PRINT "Initializing Function Generator..."
CALL config(gnrl, "PM5193/V2.5,A ?0?,N GNRl")

50

PRINT "Initializing Oscilloscope..."
CALL config(oscl, "PM3365/V0?V04,A 708,N OSC1")
PRINT "Initializing Counter..." '
CALL config(cntl, "PM6666/22,A ?10,N CNTl")
PRINT "Setting up defaults..."
CALL a11init(DEFAULT.SET) 'or ACTUAL.SET to leave unchanged
IF Glb.$tat > 1 THEN
PRINT "Error: ”, Glb.$tr: REM Print global error string
CALL res.glb: REM Reset global error status
ELSE
IF Glb.$tat - 1 THEN
PRINT "Warning: ", Glb.$tr: REM Print global warning string
CALL res.glb: REM Reset global error status
END IF
END IF

PRINT

StartOfProg:
OPEN "rawdata." FOR OUTPUT AS #1 'save data in file

CALL set.function(dmm1%, volt.ac%)

CALL set.function(dmm2%, volt.ac%)

CALL set.speed(dmm1, low)

CALL set.speed(dmm2, low)

CALL set.function(cnt1%, TIMEINTERVAL.A.B)
CALL set.coupling(cnt1%, chall%, dc%)

CALL set.sensitivity(cnt1%, cha%, .02)
CALL set.sensitivity(cnt1%, chb%, .02)

INPUT "Enter starting frequency ->", strt!

INPUT "Enter ending frequency ->", stp!

INPUT "Enter number of points per decade ->", ppd

INPUT "Enter desired RMS function generator voltage (default 1 V)
IF gnv! - 0 THEN gnv! - 1

CALL set.offset(gnr1, 0%)
CALL set.amplitude(gnrl, vrms, gnv!)

PRINT
PRINT "Frequency dB Angle"
PRINT

strt! - LOG(strt!) / LOG(lO).

stp! - LOG(stp!) / LOG(lO)

np - (stp! - strt!) * ppd 'number of magnitude points
nph - np 'number of phase points

' FOR i! - strt! TO stp! STEP 1 / ppd
p ' (i! - strt!) * ppd
x#(p) - 10 “ i!

IF INKEY$ - CHR$(27) THEN
PRINT
PRINT "Run aborted..."
PRINT
PRINT #1, ""
PRINT #1, "Run aborted...“

51

CLOSE
GOTO Again
END IF
CALL set.frequency(gnr1%, x#(p))
CALL measure(dmml%, dvml#) 'input
CALL measure(dmm2%, dvm2#) 'output
dvm# - ABS(dvm2# / dvm1#) 'find transfer fn output/input

IF dva# > 0 AND nph - np THEN 'take phase reading only if output
'is greater than 10 mV RMS
IF x#(p) > 500 THEN
CALL seg,coupling(cnt1%, chall%, ac%)
ELSE
CALL set.coupling(cnt1%, cha11%, dc%)
END IF

CALL measure(cnt1%, diff#)
ph#(p) - -diff# * x#(p) * 360#

WHILE ph#(p) < -180
ph#(p) - ph#(p) + 360#

WEND
ELSE
IF nph - np THEN nph - p
END IF
y#(p) - 20 * LOG(dvm#) / LOG(lO) ’convert to dB

LOCATE CSRLIN - l

PRINT USING "##.###“‘“ +###.### +###.###”; x#(p); y#(p);
ph‘lI (p)

PRINT #1, USING "##.###““ +###.### +###.### . +###.###
##.####““‘"; x#(p): y#(p); ph#(p); dvm#; diff#

x#(p) - i!
NEXT i!

PRINT
’Now find out what to do with this data

DO
INPUT "Find (P)eak or (D)C gain? ", g5
95 - UCASE$(g$)

LOOP UNTIL 95 3 ”P" OR g5 - "D"

DO
INPUT "Search for how many -3 dB frequencies (0,1,2) ? ", cf
LOOP UNTIL Cf - 0 OR of - 1 OR cf - 2

IF g$ - "D” THEN GOTO dcgain
'Find peak value of the points we took first

k - 1 'pointer to maximum value
FOR i - 2 TO np - 1
IF y#(i) > y#(k) THEN k - i 'reset pointer if we found a new
max.
NEXT i

IF k - np - 1 THEN 'if rightmost point is max, use it
as h0

h0# - 10# ‘ (y#(k) / 20#)

GOTO sweepleft
END IF

52

PRINT

deltalogf# - .1

f0# - x#(k) ’log of freq. of peak magnitude
t0! - y#(k) ’dB of peak magnitude

WHILE deltalogf! > .00001
PRINT USING "Locating center frequency: ######.##"; 10! “ f0!
LOCATE CSRLIN - 1

CALL set.frequency(gnr1, 10% “ (f0# - deltalogf#))
CALL measure(dmm1%, dvmlt) 'input
CALL measure(dmm2%, dvm2#) ’output
tlt - 20# * LOG(dvat / dvmlt) / LOG(lOt) ’gain at lower freq.

CALL set.frequency(gnr1, 10% ‘ (f0# + deltalogf#))

CALL measure(dmm1%, dvmlt) 'input

CALL measure(dmm2%, dvm2#) ’output

t2# - 20% * LOG(dvm2# / dvmlt) / LOG(10#) 'gain at upper freq.

IF t1§ < t0? AND t2? < t0? THEN
deltalogft - deltalogft / 2i

ELSE
IF t1# > t0# THEN
t0! - t1}
f0! - f0# - deltalogft 'set new freq. lower
END IF
IF t2# > t0! THEN
t0# - t2#
f0! - £0} + deltalogf! 'set new freq. higher
END IF
END IF
WEND
h0i - 10* ‘ (t0# / 20*) ’convert back from dB

fcnf - 10* “ f0§

PRINT USING "fc - ######.### Hz. ”; fcni
PRINT USING "HO - ##.*§§§§"; h0#

PRINT #1, USING "fc - ##§#i§.### 32."; fen?
PRINT #1, USING "HO - ##.#i#§§"; h0§

GOTO sweepleft
'Now find DC gain
dcgain:

CALL set.amp1itude(gnr1, vrms, 0) 'turn off AC voltage from gen
CALL set.offset(gnr1, gnv!) ’turn on DC volts from gen

CALL set.function(dmm1, volt.dc)
CALL set.function(dmm2, volt.dc)

CALL measure(dmml, dvmli)
CALL measure(dmm2, dvm2#)

h0# - dvm2# / dvmlt ’DC gain (NOT dB)
CALL set.offset(gnr1, 0!) ’return DMM to its original
CALL set.function(dmm1, volt.ac) 'state

CALL set.function(dmm2, volt.ac)

53

CALL set.amplitude(gnr1, vrms, gnv!)

PRINT
PRINT #1, ""

PRINT USING "DC Gain - ##§.#####"; h0#:

IF hat <> 0 THEN PRINT USING " - ##.### dB"; 20 * LOG(ABS(h0#)) /
LOG(lO)

PRINT #1, USING "DC Gain - ###.#####"; hat;

IF hat <> 0 THEN PRINT #1, USING " - ##.### dB"; 20 * LOG(ABS(h0#))
/ LOG(lO)

'Now look for -3 dB point from left side
sweepleft:
IF cf - 0 THEN GOTO graphit

PRINT
PRINT *1, ""

target! - h0# / SQR(2#)
IF target¥ < .001 THEN

PRINT "DC Gain is too low - Cannot sweep for -3 dB point from
left"

PRINT #1, "DC Gain is too low - Cannot sweep for -3 dB point from
left”

f0! - 0
GOTO sweepright
END IF
i - 1
WHILE ABS(EXP(LOG(10#) * y#(i) / 20%) - target!) / target! > .2# AND
1 < np
i - i + 1
WEND

IF i - np THEN

PRINT "Couldn't find left -3 dB frequency... run again with
different points"

PRINT #1, "”

PRINT #1, "Couldn’t find left -3 dB frequency... run again with
different points”

ELSE
flt - 10 ‘ x#(i - 1) 'starting frequencies
f2! - 10 ‘ x#(i)
t1! - 10 “ (y#(i - 1) / 20) 'starting target values
t2! - 10 “ (y#(i) / 20)
DO

slope! - (t2# - t1!) / (f2! - f1!)

f0! - (target# - t1?) / slope§ + £1}

PRINT USING "Locating left -3 dB frequency: #####.##"; f0!
LOCATE CSRLIN - 1

IF INKEYS - CHR$(27) THEN
PRINT ”Location of left -3 dB frequency aborted...

PRINT #1, "”
PRINT #1, "Location of left -3 dB frequency aborted...”
f0! - 0
GOTO sweepright
END IF

54

'now find to
CALL set.frequency(gnr1%, f0#)
CALL measure(dmml, dvm1#)
CALL measure(dmm2, dvm2#)
t0# - dvm2# / dvm1#
IF (f2# - f0#) < (f0# - f1#) THEN
f1# - f0#
t1# - t0#
ELSE
f2# - f0#
t2# - t0#
END IF
hg& - 100000# * t0# / h0#
LOOP UNTIL hg& - 70711

fcl# - f0# ’left cutoff frequency
PRINT USING "The left cutoff frequency is #####.## Hz.'; f0#
PRINT #1, USING "The left cutoff frequency is #####.## Hz.': f0#

IF Cf - 1 THEN
PRINT USING "GBP - #########.## Hz": f0# * h0#
PRINT #1, USING "GBP - #########.## Hz"; f0# * h0#
GOTO graphit

END IF

END IF
sweepright:
'Now look for -3 dB point from right side

PRINT
PRINT #1, ""

IF target# < .001 THEN
PRINT "Rightmost Gain is too low - Cannot sweep for -3 dB point
from right"
PRINT #1, "Rightmost Gain is too low - Cannot sweep for -3 dB
point from right"
GOTO graphit
END IF
1 - np - 1
YHILE ABS(EXP(LOG(10#) * y#(i) / 20#) - target#) / target# > .2# AND
> 0
i - i - 1
WEND

IF 1 - 0 THEN

PRINT "Couldn’t find right -3 dB frequency... run again with
different points"

PRINT #1, ”"

PRINT #1, "Couldn’t find right -3 dB frequency... run again with
different points"

ELSE
f1# - 10 “ x#(i - 1) 'starting frequencies
f2# - 10 “ x#(i)
t1# - 10 ‘ (y#(i - 1) / 20) 'starting target values
t2# - 10 “ (y#(i) / 20)
DO

slope# - (t2# - t1#) / (f2# - f1!)
f0# - (target# - t1#) / slope# + f1#
PRINT USING "Locating right -3 dB frequency: #####.##"; f0#

55

LOCATE CSRLIN - 1

IF INKEY$ " CHR$ (27) THEN

PRINT "Location of right -3 dB frequency aborted...

PRINT #1,

PRINT #1,
' GOTO graphit
END IF

’now find to
CALL set.frequency(gnr1%,
CALL measure(dmml, dvm1#)
CALL measure(dmm2, dvm2#)
t0# - dvm2# / dvm1#
IF (f2# - f0#) < (f0# - f1#)
f1# - f0#
t1# - t0#
ELSE
f2# - f0#
t2# - t0#
END IF
hg& - 100000# * t0# / h0#

f0#)

LOOP UNTIL hg& - 70711

fc

PRINT
PRINT
IF 93

r#‘- f0#

PRINT

PRINT

- fc

PRINT
PRINT
(fcr# -

END

USING "fc - #####.##
l#)

*1, an
#1, USING
fc1#)

”fc = #####.##

IF

graphit:

CLOS

CALL set.amplitude(gnr1, vrms,

E #1

0)

PRINT

PRINT "Press any key to see graphs...

WHILE INKEY$ - "":

CALL
CALL
axes
CALL
time
CALL
CALL
CALL
CALL
CALL
CALL

WEND

GrfReset(4)
SetAxGridVis(-1, 1)

SetAxAuto(-1, 21)
SetXDataType(4)
SetYDataType(4)
SetAxName(0, "log(f)")
SetAxName (1, "dB")
GrfCurv2D(x#(), y#(), np)
GrfReset(4)

Oo - ###.####”:

"Location of right -3 dB frequency aborted...”

THEN

'right cutoff frequency

USING "The right cutoff frequency is #####.## Hz.":
#1, USING "The right cutoff frequency is #####.## Hz."; f0#
- "D” THEN fcn# - SQR(fc1# * fcr#)

f0#

fcn#; fcn# / (fcr#

do = ###.####"; fcn#: fcn# /

'output raw data file

’turn off AC voltage from gen
'to facilitate changing
'components

'enable grid lines for both

'auto-scale both axes every

'set name of x-axis
'set name of y-axis
'x-y plot of dB vs. log(f)
'clear the graphics display

56

CALL SetAxName(1, "Phase Angle") 'set name of y-axis

CALL GrfCurv2D(x#(), ph#(), nph) 'x-y plot of phase vs. log(f)
CALL GrfLReset(0, 0, 1, 2) 'return to text mode

CLS

Again:

PRINT ”Again (y/n)?";

as - INPUT$(1)

CLS

IF UCASE$(a$) - "Y" THEN GOTO StartOfProg
REM Now return to local control
CALL iolocal(dmml%)

CALL iolocal(dmm2%)

CALL iolocal(osc1%)

CALL iolocal<cntl%)

CALL iolocal(gnr1%)

END

REM SDYNAMIC

SUB ReportError
END SUB

3.9 Summary

In this chapter techniques were developed to measure active circuits.
A second voltmeter was added to improve accuracy, along with the necessary
changes to the code. A method of measuring the gain-bandwidth-product of an
Op-amp was also given. Finally a program to measure the second order filter

parameters fc, Ho, and Q0 was designed.

CHAPTER 4
EXPERIMENTAL AND THEORETICAL VERIFICATION

4.1 Introduction

In this chapter the routines developed in the previous chapters are used
to investigate and verify the properties of a second-order active band-pass filter

presented in a recent paper.
4.2 The Filters to be Tested

The most dificult band-pass filter to measure by hand is a high-Q filter.
In a recent paper [6] a new relocated Tow-Thomas filter was presented (Figure
4.1) derived from the original Tow-Thomas filter (Figure 4.2). It claims to be
insensitive to the gain-bandwidth-product of the op-amps used provided that
they are matched. We will use BODE6C.BAS and BODE7C.BAS to attempt
to verify these results.

The new filter was generated from the original by a method called op-
amp relocation. This method generates circuits that have the same idea]
characteristics but may react difi'erently to a real (non-ideal) op—amp. The
basis upon which this method is founded is that the voltage across the
terminals of an ideal op-amp is zero (in the presence of feedback). So if one

terminal of an op-am'ii is grounded, the other is forced to zero potential. Thus

'57

58

any other ground connection in the circuit may be made to the Op-amp node
instead of ground. Consider the original Tow-Thomas filter. The non-inverting
node of Op-amp 2 is connected to ground. Thus there will be no change in the
ideal characteristics of the circuit if it is connected to the inverting node of op-
amp 1 or op-amp 3 instead. Now the non-inverting node of op-amp 3 may be
connected to the inverting node of either op-amp 1 or op-amp 2. Certain
connections can make the circuit unstable, however. There are software
programs that perform op-amp relocation and check for stability; one was used
to design the relocated filter and led to the addition of R, and R,. The non-
ideal characteristics (due to finite GBP, et. al.) of a relocated circuit may be
dramatically difi'erent. In fact, the original Tow-Thomas filter is unstable

above about 6 kHz, whereas the relocated filter is stable to a much higher

frequency.

4.3 Measuring the Gain-Bandwidth-Products

The paper [6] calls for the use of the Texas Instruments TL084 quad Op-
amp. This integrated circuit contains four op-amps whose gain-bandwidth-
products are supposed to be "reasonably" close. The data sheet for this IC
claims a typical GBP of about 3 MHz. Measurements were taken for six
TL084’s with varying gain-bandwidth-products. The summary of the measured

results are shown in Table 4.1.

59

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

R3
—~w
R6
VS ‘1 I

 

 

 

Figure 4.1 - Relocated Tow-Thomas Filter

 

60

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

R I“; was
3 __
1w» 3+
R1 _
"Ac/3W— -
1} C2
M :e
_ R2
1 .—
+ +2
T VBP _

Figure 4.2 - Original Tow-Thomas Filter

 

61

Table 4.1 - Gain-Bandwidth-Products of Measured Op-Amps

 

 

 

 

 

 

GBP 1 Room
(MHz) (MHz) Temp. (°C)

1

2 2.971 2.376 2.909 24

3 2.761 2.903 2.690 24

4 2.979 2.930 2.357 24

16 3.356 3.613 3.479 23

21 2.373 2.716 2.305 24

 

 

 

 

 

 

 

While measuring these gain-bandwidth-products a crucial error was almost
made. The chips were being measured by placing them into the circuit,
applying power, taking the measurement, and then removing them. While
doing this, it was noticed that subsequent runs of the program were producing
significantly different results - as much as 10%. This was caused by the
failure to allow the IC to reach its normal operating temperature. From the
data sheet for the TL084, 168 mW is being dissipated by the biasing current
when using :15 V power supplies. Since the thermal resistance is 131.6° C/W,
we can expect an increase of 22° C above ambient. Out of curiosity, several
runs were performed at varying intervals after applying power. These results
are shown in Table 4.2. They are not conclusive since they only apply to one
op-amp on one 10, but they are interesting nonetheless. Based upon these
results, all op—amps were allowed at least 15 minutes to reach operating

temperature before any measurements were taken.

62
Table 4.2 - Gain-Bandwidth-Product vs. Time

Time After Power
Application (min) GBP (MHz) g
I 3
| .

 

 

 

 

 

 

 

1 2.915

5 2.332 ‘
II 10 2.373 f

15 2.377 r

20 2.376 ’
| 25 2.876 1';
‘ , ,, _ _ 2.37 , E

4.4 Testing the Filter

Now we are ready to build and test the filter. A list of the component
values recommended in the paper and the actual values used are shown in
Table 4.3. The results of the measurements are summarized in Table 4.4. The
filter performed as the authors claimed. Despite the varying gain-bandwidth-
products of the op-amps used, the filter parameters remained relatively
constant. The consistency of the center frequency was most impressive,
varying only in the least significant digit. The maximum deviation in Q0
among the filters, between op-amps 1 and 16, was only 2.94%, and the
maximum deviation of Ho was 2.49%. A typical set of generated Bode plots

are shown in Figures 4.3 - 4.4.

63

Table 4.3 - Component Values for the Tow-Thomas Filter

j Recommended
._ Component __ _ Value

 

 

 

 

 

 

 

 

 

 

 

1 Actual Value i
; R1 100 kg 101.05 m
f R, 4 m 3.355 m
l R, 1 m 1.030 m
IR, 1 kn 1.020 m
I R, 1 kn 1.014 kg
R, 100 m 104.25 m
R, 2.47 m 2.240 m
R, 1.175 m 1.203 kg
(31 7.953 nF C, = 7.370 nF
R, = 4.67 M12
6 10 kHz.
7.953 nF cP -.- 7.951 nF
R9 = 4.53 MO
@ 10 kHz. j

 

  

 

 

 

 

 

 

 

 

 

 

 

 

Room Temp.

Op-Amp f I H - (°C)

1 l 10.12 49.61 0.9513 24

2 10.13 49.43 0.9501 24 I

3 10.13 49.60 0.9532 24 I

4 10.12 49.54 0.9539 24 I

16 10.13 43.15 0.9343 24 n

21 10.12 49.33 0.9573 23 l

64
As a measure Of how good this circuit is, the original Tow-Thomas circuit

was Constructed and tested with two TLO84s. Unfortunately it is unstable at
10 kHz, simply acting as an oscillator. The solution to this is to lower 1; by
making the two capacitors larger. In this case the capacitors were increased
toCl= 19.73nF(Rp=2.04MQ@4kHz)andC,= 19.88nF(Rp=2.90MO@
4 kHz). The results are given in Table 4.5. In the next section, it is shown
that when using an ideal Op-amp model, ,Q0 and Ho are not a function of the
capacitor values. Thus it is clear that in the original design Q0 and H0 are

dependent upon the gain-bandwidth-products of the op-amps used.

(a;

09
Figure 4.3 - Bode Magnitude Plot of Relocated Tow-Thomas Filter

65

 

gm
£311 _
:11
a.
9
II
Ina-11
i
e:13:
1111 \

 

%MMMMI%)MMMMH
0!

Figure 4.4 - Bode Phase Plot of Relocated Tow Thomas Filter

Table 4.5 - Measured Results of Original Tow-Thomas Filter

Room Temp. l'
Op-Amp f (kHz) (°C)

4. 039 71.03 1. 357
21 4.035 82.35 1.589 23

       

 

4.5 Theoretical Verification of Measured Results

There is a symbolic SPICE (Sspice) computer program under
development at Michigan State University [7]. We will use this program to

verify the measured results of the relocated Tow-Thomas filter. The input to

Sspice is similiar to PSpice, and is given below.

Rel
vs
:1
fl
r5
r4
r3
r6

ocated
10 0

O

mbNWHOUQHHUCflHt—‘UQNH
\DmQUIUI-‘bbNNOOO-‘QOUICON

66

Tow-Thomas Filter
ac

101.05k

3.855k

Note the addition of the r0 resistors, representing the output resistance of the

TL084. They are included because later in this section we will show that they

affect the filter parameters in the non-ideal case. These values do not afl’ect

the filter parameters in the ideal case so they may be ignored for now. Also

note that since an ideal op-amp model is being used, the formulas found for

this filter will also apply to the original Tow-Thomas (R, and R3 do not appear

in the formulas for the parameters). The Sspice output file is given below.

Relocated Tow-Thomas Active Filter

[0
[0
[0
[0
[O
[0
[1

Y(
Ii
Y(
Y(

hUNH
““
NNHH
vvvv
I I I I

,1 0 -G3 0
,1 0 0 -G0
Y 3,2 0 0
Y 4,2 0 0
-GS -G4 0
0 Y 6,3 0
0 o 0
-sC1-GP1-G1
+sC1+GP1+GZ+G1+GO
-sC2-GP2
+sC2+GP2+GS+Go

O I 00000

GG

HOOOOO

1 [v2
][V4
][V6
][V7
][V8
1 [v9
1 [v10

i—ll—JHHHL—JO—l

67

Y( 6,3 ) - +G4+G3+GO

*Ignore nodes 11 and higher if present. They are used for internal
numbering.

Numerator of: v2

TERMS SORTED ACCORDING TO POWERS OF 3

s**1 terms:

+ sC2*G6*G4
s**0 terms:

+ GP2*GG*G4
********************t***************************
NUMERICAL VALUE OF ABOVE SYMBOLIC RESULT
+ 7.47731e-017 * s**1 + 2.053330-015 * s**0
***t********************************************
Denominator of: v2
TERMS SORTED ACCORDING To POWERS or s
s**2 terms:

- sC2*sC1*G4
s**1 terms:

- 3C2*GP1*G4 - sC2*G4*Gl - sCl*GP2*G4
s**0 terms:

- GP2*GP1*G4 - GP2*G4*G1 - GS*G3*G2
**************************************i*********
NUMERICAL VALUE OF ABOVE SYMBOLIC RESULT

- 6.134740-020 * s**2 - 8.049480-017 * s**1 - 2.483730-010 * s**0

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

Relocated Tow-Thomas Active Filter

SECOND ORDER FILTER PARAMETERS:

Go is:

SQRT{( + C2*C1)*( + GP2*GP1*G4 + GP2*GQ*GI + GS*G3*GZ)}

( + C2*GP1 + C2*G1 + C1*GP2)*SQRT{ + G4}
8 48.4933

68
Wo**2 is:

( + GP2*GP1*G4 + GP2*G4*G1 + G5*G3*G2)

 

( + C2*C1*G4)
fo I 10126.88:

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

The important parameters in the Sspice output have been highlighted. Sspice
does not give Ho, so a hand calculation was done using the given transfer
function. This calculation yields |T02n-10126.8)| 50.923913.

Sspice also has the ability to predict the shifiing of f, and Q0 due to
finite gain-bandwidth-products. It was stated earlier that the output
resistance also efl‘ects these shifts. Thus it is necessary to measure those
values, and they are given in Table 4.6. The Sspice results are tabulated in

Table 4.7.

Table 4.6 - Measured Output Resistance of Op-Ampsl

 

 

 

 

 

 

 

: Op-Amp H , 1(0) Row 2 (n) Rm 3 (a)
‘ 1 235.0 233.6 306.3
I 2 273.3 272.3 267.4
I 3 279.3 340.6 296.2

4 295.5 237.5 295.5
I 16 263.5 325.4 256.0
I 21 312.5 293.9 237.5

 

1This data was collected by Bassam Hindeleh

 

 

69

Table 4.7 - Parameter Shifting Predicted by Sspice

 

 
      

 

 

 

 

 

 

___. AW
1 0000916259 0.033277500
" 2 0001009520 0.020737300
3 0001061240 0072313000
n 4 0001027560 0.061305000
16 0000740414 0.000161553
21 0001259700 0022055900

 

Table 4.8 - Predicted Parameters and Errors

 

 

 

 

Prediced fc (kHz) Predicted Q0 % Error
1 10.12 0.000 50.11 1.003
2 10.12 l 0.099 49.50 0.040
3 10.12 0.099 52.02 4.379
4 10.12 0.000 51.47 3.396

 

 

 

 

 

 

 

 

16 10.12 0.099 48.50 0.727
21 10.11 0.099 ' 49.56 0.365 I

70
The experimental results agree well with theory for the most part. The most

glaring inconsistencies are the Q0 errors for op-amps 3 and 4. In section 3.7
it was noted that the results of the program are consistent to within one count
at the fourth significant figure. In this case, a change in that last significant
figure of one count in one of the cutofi‘ frequencies is enough to cause the error

in the calculation of Q0.

4.6 Summary

In this chapter the previously developed routines were tested on a high-
Q band-pass filter presented in a recent paper. Sspice was used to provide
theoretical verification of the measured results. The results were found to

compare favorably with theory.

CHAPTER 5
CONCLUSIONS AND FUTURE RESEARCH

5.1 Conclusions

The goal of this document was to indicate an approach to the
development of software using TestTeam to automate precision measurements.
TestTeam has been found to be a powerful and useful toOl in this study. In
Chapter 2, routines were developed using one voltmeter to analyze passive
networks. In particular, a program to read data and create Bode magnitude
and phase plots and -3 dB fi-equencies was designed. In Chapter 3, a second
voltmeter was added to correct errors due to the output impedance of the
function generator. Routines were then developed to analyze active circuits,
such as finding the gain-bandwidth-products of op-amps and finding the
second-order parameters ft, Ho, and Q0 of band-pass and band-stop filters.
Although not explicitly developed for this purpose, the routines presented could
be altered to measure the aforementioned parameters of low and high-pass
filters also. Finally in Chapter 4 the routines were successfully tested on a

high-Q band pass filter.
5.2 Future Research

The next logical step in the progression of automated measurements is

'71

72

to move from steady state to transient measurements. Measurements such as
step response and calculations such as slew rate, step response, and FFTs can
be made from digitized oscilloscope data. In order to use TestTeam with an
oscilloscope, another driver package is required: The PM2235 Oscilloscope
Drivers. An attempt to write a program to capture oscilloscope data was made.
However when feeding the same signal to both channels and performing a
measurement, there was a voltage shift between the two channels, and neither
one had a correct ground. Attempts with other oscilloscopes yielded difl'erent
shifts, but no correct results. Phone consultations with a Fluke engineer have
not provided a satisfactory solution as of this writing.

The code written to capture the oscilloscope data is shown below. The
header on this program is slightly difi'erent due to a replacement for
TES'I'I‘EAMBAS supplied with the oscilloscope drivers.

'************* PM2233 INSTRUMENT DRIVERS INITIALIZATION PROCEDURES
'* File name : SCOPEBC. BAS

'****************************************************************

SDYNAMIC $INCLUDE: '\driver3\DRIVERS.INC'
' $DYNAMIC $INCLUDE: '\lw\include\graphics.inc’

V*******************i********************************************

’* DRIVERS INITIALIZATION
l**t*******************************************t*****************

CLEAR , , 10000 ' Reserve 10K bytes stack
Drivers.Initialization
GOTO Application

Drivers.Report.Error:
IF (ERR <> 11) THEN
PRINT : PRINT "BASIC error"; ERR; "in line": ERL
ELSE
ReportError
END IF
RESUME NEXT
'END Drivers.Report.Error

Application:
' USER APPLICATION STARTS HERE

73

DIM SHARED ca%(512), cb%(512), ma#(512), mb#(512),
reg.settings#(89), t#(512)

CALL getmem(20000)

overlay.memory.size& - 98304

memory.size& - SETMEM(640000)

memory.size& - SETMEM(-over1ay.memory.size&)

CLS
PRINT

StartOfProg:
PRINT "Into my routine...”

CALL Set.Frequency(GNRl, 1000#)
CALL Set.Amplitude(GNR1, PP, 8!)
CALL Set.Waveform(GNR1, triangular)

PRINT "Calling Auto Set..."

CALL Auto.Set(osc1)

CALL set.polarity(oscl, chall, positive)

PRINT ”Setting DC coupling..."

CALL Set.Coupling(oscl, chall, DC)

PRINT "Setting channel format..."

CALL set.chan.trans.format(oscl, 1, real.data, 512)
PRINT ”Setting horizontal mode..."

CALL Set.Horizonta1.Mode(oscl, SINGLE.SHOT)

PRINT "Measuring scope data...”

CALL measure.chan.data(oscl, accu, cha, ca%()) 'measure forces
trigger
CALL read.chan.data(osc1, accu, chb, cb%()) 'read does not

PRINT ”Scaling data..."

CALL get.reg.settings(osc1, accu, reg.settings#())

CALL scale.chan.data(osc1, cha, reg.settings#(), ca%(), ma#())
CALL scale.chan.data(osc1, chb, reg.settings#(), cb%(), mb#())
FOR i - 0 TO 512

t#(i) - i
NEXT 1

graphit:

CALL Set.Amplitude(GNR1, VRMS, 0) 'turn off AC voltage from gen

PRINT
PRINT "Press any key to see graphs...":

WHILE INKEYS ' "": WEND

CALL GrfReset(4)

CALL SetAxAuto(-1, 21) ’auto scale both axes
CALL SetPlotMode(0) 'no wait for keypress
CALL SetXDataType(4)

CALL SetYDataType(4) 'eight-byte floating point

CALL GrfCurv2D(t#(), ma#(),.512)

74

CALL SetPlotMode(2) 'wait for keypress
CALL GrfCurv2D(t#(), mb#(), 512)

CALL GrfLReset(0, 0, 1, 2) 'return to text mode
CLS

Again:

PRINT "Again (y/n)?";

as - INPUT$(1)

CLS

IF UCASE$(a$) - "r" THEN GOTO StartOfProg

REM Now return to local control

CALL Iolocal(DMM1%)
CALL Iolocal(DMM2%)
CALL Iolocal(osc1%)
CALL Iolocal(CNT1%)
CALL Iolocal(GNRl%)

END
REM $DYNAMIC

DEFSNG A-Z
SUB Drivers.Initialization
'******************i************************************i*********

'* DRIVERS INITIALIZATION PROCEDURE

't********************t*******************************************

ON ERROR GOTO Drivers.Report.Error

CLS

CALL Reset.Config

overlay.memory.size& - 141312

memory.size& - SETMEM(640000) ' Try' to get all memory' now
available

memory.size& - SETMEM(-overlay.memory.size&) ' Free memory for
instrument overlays

PRINT "Initializing Multimeters...”

CALL Config(DMM1, ”8840A,N DMM1,A 705")

CALL Config(DMM2, ”8840A,N DMM2,A 706")

PRINT "Initializing Function Generator..."

CALL Config(GNRl, "PM5193/V2.5,N GNR1,A 707")

PRINT "Initializing Oscilloscope..."

CALL Config(oscl, ”PM3365/V07V04,N OSC1,A 708")

PRINT "Initializing Counter..."

CALL Config(CNT1, "PM6666/22,N CNT1,A 710”)

PRINT "Setting up defaults..."

CALL Allinit(actual.SET)
END SUB 'Drivers.Initialization

SUB ReportError
f******************ti***************************************t**t**

'* DEFAULT ERROR HANDLER
I*****************************************************************

GPIB.ERR - Err.Num

IF (GPIB.ERR <> 0) THEN
GPIB.STAT - Err.Stat
GPIB.ERR$ - Err.Str$
GPIB.GLBSTAT - Glb.$tat.

75

GPIB.GLBERR$ - Glb.$tr$

PCIB.ERR - IOErr.Num

PCIB.ERR$ - IOErr.Str$

status = GPIB.STAT

SELECT CASE status
CASE 0: PRINT ”IOdrivers error in line”; ERL: PRINT GPIB.ERR$
CASE 1: PRINT ”Warning in line"; ERL: PRINT GPIB.ERR$
CASE 2: PRINT "Error in line"; ERL: PRINT GPIB.ERR$
CASE 3: PRINT "Fatal error in line"; ERL: PRINT GPIB.ERR$: END

END SELECT

END IF
END SUB 'ReportError

The output of this program with the same signal fed to both channels is shown
in Figures 5.1 - 5.3. It is clear from these plots the ofl'set of the two channels
is not the same, nor is it consistent among different waveforms.

If the two outputs were correct, one measurement which could be made
is the slew rate of an op-amp. The first step is to read the oscilloscope data
when the time base of the oscilloscope was set see the slewing. Then two
points could then be used to calculate the slope, perhaps at 30% and 80% of
the peak voltage values. The sampling rate can be read from the oscilloscope

using a driver, and the calculation of the slew rate follows.

 

76

 

 

 

 

MW wax—u.” .1
,1

W W “-
fl
9
ll
i! “MW

'1] W 51511] $151!] 'ililil iIEIEJ

Figure 6.1 - Oscilloscope Output for Positive Pulse

77

i H till Sill ills]
Figure 5.2 - Oscilloscope Output for Triangle Wave

mm

 

W
W .

78

 

 

 

 

Figure 5.3 - Oscilloscope Output for Square Wave

LIST OF REFERENCES

[1]

[2]
[3]

[4]

[5]

[6]

[7]

LIST OF REFERENCES

"Getting Started with TestTeam," National Instruments Corporation,
1989.

"TestTeam User Manual," National Instruments Corporation, 1988.

"8840A Digital Multimeter Instruction Manual," John Fluke Mfg. 00.,
Everett Washington, 1985.

"Programmable Timer/Counter PM 6666 Operator’s Manual," Philips
Export B.V., The Netherlands, 1988.

G.M. Wierzba, Desigg’ng Electronic Circuits Using Analog ICs, Texas
Instruments Inc, Dallas, TX and Prentice-Hall, Englewood Cliffs, NJ,
in press. . .

G.M. Wierzba and J .A. Svoboda, "An op-amp relocated bandpass filter
with zero center frequency sensitivity to gain-bandwidth-product,"

Invited Paper, Proc. of the 29th Midwest Smposium on Circuits and
Systems, Lincoln, NE, pp. 28-32, August, 1986.

G.M. Wierzba, V. Joshi, A. Srivastava, and J .A. Svoboda, "Sspice -
Symbolic SPICE for Linear Active Circuits," Invited Paper, Proc. of the

32nd Midwest Smmsium on Circuits and Systems, Urbana, IL, pp.
1197-1201, August, 1989.

‘79

NTS RTE UNIV

II II II IIIIII II IIIII III IIIIIIIIII

IIIIIIIIIIIIIIIII II IIIIII