Type1390 BRandom Noise Generator Operating Instructions 1965
User Manual:
Open the PDF directly: View PDF 
.
Page Count: 24
OPERATING INSTRUCTIONS
TYPE
1390-B
RANDOM-NOISE
GENERATOR
GENERAL
RADIO
COMPANY
K
-
(...)
~
0
I
G"
OPERATING INSTRUCTIONS
TYPE
1390-8
RANDOM-NOISE
GENERA
TOR
Form
1390-0100-K
November,
1965
GENERAL
RADIO
COMPANY
WEST
CONCORD,
MASSACHUSETTS,
USA
SPECIFICATIONS
Frequency
Range:
5
c/s
to
5
Mc/s.
Output
Voltage:
Maximum
open-circuit
output
is
at
least 3 V for
20-kc range, 2 V for 500-kc range,
and
1 V for 5-Mc range.
Output
Impedance:
Source impedance for maximum
output
is
approximately 900
fl.
Output
is
taken
from a 2500-fl
potentiom-
eter. Source impedance for
attenuated
output
is 200
fl.
One
output
terminal is grounded.
Typical Spectrum Level
Range with
1-
V,
rrns,
output)
20
kc/s
5
mV
for 1-cycle
band
500
kc/s
1.2
mV
for 1-cycle
band
5
Me/
s 0.6
mV
for 1-cycle
band
Spectrum Level Uniformity*
within ± 1 dB,
20
c/s
to
20
kc/s
within ± 3 dB,
20
c/s
to
500
kc/s
within ± 3 dB,
20
c/
s to 500
kc/s;
within ± 8 dB, 500
kc/s
to
5
Mc/s
• Noise
energy
also
present
beyond
these
limits. Level is
down
3
dB
at
5
cfs.
See
plot.
Waveform:
Noise source has good normal,
or
Gaussian,
distribu-
tion of amplitudes for ranges of
the
frequency
spectrum
that
are
5
~
0
UJ
aJ
u
-5
UJ
0
I r
I
narrow compared
to
the
band
selected. Over wide ranges
the
distribution is less symmetrical because of dissymmetry
intro-
duced
by
the
gas tube. Some clipping occurs on
the
500-kc
and
5-Mc ranges.
Voltmeter: Rectifier-type averaging
meter
measures
output.
It
is
calibrated to
read
rms value of noise.
Attenuator:
Multiplying factors of 1.0, 0.1, 0.01, 0.001,
and
0.0001. Accurate
to
±3%
to
100
kc/s,
within ±
10%
to
5
Mc/s.
Power
Required:
105
to
125
or
210
to
250
V,
50
to
400
c/s,
50 W.
Accessories
Supplied:
TYPE
CAP-22 Power Cord, spare fuses.
Accessories
Available:
Rack-adaptor
panel
(panel height 7 in).
Mechanical
Data:
Convertible-Bench
Cabinet
Net
Width Height Depth Weight
Shipping
Weight
kg
7.5
For
additional information, ask for General
Radio
Reprint
E-110.
5Mc l
RANGE'\
\
./
~
;
\
~fOkc
'~500kc
RANGE
~RANGE
,\
-10
lc 5c IOc
IOOc
Ike IOkc IOOkc I
Me
lOMe
FREQUENCY
TABLE
Section 1 INTRODUCTION
1.
1
Purpose
. . . .
1.2
Description
. . .
OF
Section 2 THEORY OF OPERATION
2. 1 General . . . . . . . .
2.2
Output
Voltage
. . . . . .
2.3
Characteristics
of
Noise Output
CONTENTS
2.4
Departures
of Output
from
True
Randomness
2.5
Frequency
Spectrum
of
Noise
. . . . .
2.6
Analysis
of Noise by
Constant-Percentage
Analyzers
2.7
Type
1390-P2
Pink-Noise
Filter
Section 3 INSTALLATION .
3.
1 Bench Mounting
3.2
Relay-Rack Mounting .
3.3
Power Supply Connection
Section 4 OPERATING PROCEDURE
4.
1 Start-Up . . . .
4.2
Frequency
Control
4.3
Output Control
4.4
Voltmeter
4.5
Load . . .
4.6
Hum
. . .
4.7
Applications
Section 5 SERVICE
AND
MAINTENANCE
5. 1 General . . . .
5.2
Access
to Components
5.3 Preliminary
Checks
.
5.4
Meter Does Not Read .
5.5
Tube
Replacement
5.6
Hearer Voltage of Type 6D4
Gas
Tube
5.7
Voltage Measurements
PARTS
LIST.
. . . . . .
1
1
2
2
2
2
4
4
4
4
6
6
6
6
6
6
6
6
6
7
7
7
11
11
11
11
11
11
11
11
13
Figure
1.
Type
1390-B
Random-Noise
Generator
. (
See
Section
3 for
different
mounting
arrangements).
View
at
lower
right
shows
Type
1390-P2
Pink-Noise
Filter
mounted
on
the
generator
.
TYPE
1390-B
RANDOM-NOISE
GENERATOR
Section 1
INTRODUCTION
1.1
PURPOSE.
The
Type
1390-B
Random-Noise
Gen-
erator
(Figure
1)
provides
a
high
level
of
electrical
noise
at
its
output
terminals.
This
type
'of
signal
is
useful
in
room
acoustic
measurements,
loudspeaker
and
micro-
phone
tests,
psychoacoustic
tests,
filter
tests,
cross-
talk
measurements
for
multichannel
carrier
systems,
calibration
checks
on
recording
systems,
modulation
of
signal
generators
and
test
oscillators,
tests
of
rms
re-
sponse
of
meters,
observation
of
resonance
in
systems,
electrical
averaging
of
resonant
responses,
and
compar-
isons
of
effective
band
width.
A
pair
of
th~se
genera-
tors
can
be
used
as
signal
sourc.es
for
demonstration
of
various
degrees
of
correlation,
possible
errors
of
random
sampling,
and
other
concepts
of
statistical
theory.
1.2
DESCRIPTION.
1. 2.1 CONNECTIONS.
Two
jack-top
binding
posts,
used
as
output
terminals,
are
provided
on
the
panel
of
the
Type
1390-B
Random-Noise
Generator.
1.
2. 2
CONTROLS.
The
following
controls
are
on
the
panel
of
the
Type
1390-B
Random-Noise
Generator:
Name
Descriotion
Positions
_lls_e_
RANGE
3-position
rotary
selector
20
kc,
500
kc,
5
Me
Selects
network
used
for
shaping
switch
noise
spectrum
POWER
2-position
toggle
OFF,
POWER
Energizes
instrument.
switch
None
2-position
rotary
selector
LOW, HIGH In
LOW
position,
introduces
a 10:1
switch
resistance
pad
after
gas-tube
noise
source
to
reduce
effect
of
amplitude
limitations
of
amplifier
and
to
re-
duce
noise
field
radiated
by
the
in-
strument.
OUTPUT
Continuous
rotary
control
Varies
output
voltage.
OUTPUT
Rectifier-type,
averaging
Indicates
rms
value
of
noise
at
VOLTS
voltmeter
output
terminals.
MULTIPLY
5-position
rotary
selector
0.0001,
0.001,
0.01,
0.1,
1.0
Attenuates
output
voltage.
BY
switch
GENERAL RADIO
COMPANY
Section 2
THEORY
OF
2.1
GENERAL
(See
Figure
2.)
The
Type
1390-B
Random-Noise
Generator
uses
a
gas-discharge
tube
as
its
noise
source.
A
transverse
magnetic
field
is
ap-
plied
to
the
tube
to
eliminate
the
oscillations
usually
associated
with a
gas
discharge
and
to
increase
the
noise
level
at
high
frequencies
1•
The
noise
output
from the
gas
tube
is
amplified
in a
two-stage
amplifier.
Between
the
first
and
second
stages
the
noise
spectrum
is
shaped
in
one
of
three
different
ways,
depending
on
the
setting
of
the
RANGE
switch.
At
the
20-kc
posi-
tion,
a
low-pass
filter
is
inserted,
which
has
a
gradual
roll-off
above
30
kc,
with
the
audio
range
to
20
kc
uni-
form
in
spectrum
level.
At
the
500-kc
setting,
a low-
pass
filter
is
inserted,
which
rolls
off
above
500 kc.
At
the
5-Mc
setting,
a
peaking
network
is
used.
This
network
approximately
compensates
for
the
drop
in
noise
output
from
the
gas
tube
at
high
frequencies,
so
that
a
reasonably
good
spectrum
is
obtainable
to
5 Me.
2.2
OUTPUT
VOLTAGE.
The
maximum
open-circuit
output
voltage
on
the
20-kc
band
is
at
least
3
volts,
on
the
500-kc
band
at
least
2
volts,
and
on
the
5-Mc
band
at
least
i
volt.
This
corresponds
to
a
relatively
high
noise
level,
since
the
output
impedance
at
maxi-
mum
output
is
only
about
900
ohms.
This
level
can
be
expressed
in
terms
of
the
resistance
noise
correspond-
ing
to
900
ohms
at
room
temperature.
The
rms
voltage
in a
one-cycle
band
due
to
thermal
agitation
in a 900-
ohm
resistor
at
room
temperature
is
about
3.8
x 10-9
volt.
The
level
from
the
Type
1390-B
Random-Noise
Generator
is
about
five
millivolts
for a
one-cycle
band
when
there
is
a
total
output
voltage
of
one
volt
on
the
20-kc
band.
This
level,
then,
is
about
1,300,000
umes
the
corresponding
voltage
for
resistance
noise,
LOW-PASS
"'!!'
B+
FILTERS
OPERATION
or
about
122
decibels
above
resistance
noise
at
the
same
impedance
level.
2.3
CHARACTERISTICS
OF NOISE
OUTPUT.
As
shown
in
Figure
3, no
regular
pattern
appears
in
the
output
waveform;
it
is
characterized
by
randomness
rather
than
by
regularity.
Noise
is
therefore
described
by
statistical
means,
2
and
is
characterized
by
its
distri-
bution
of
instantaneous
amplitudes
and
by
its
frequen-
cy
spectrum.
A random
noise
is
often
defined
as
a
noise
that
has
a
"normal"
or
"Gaussian"
distribution
of
ampli-
tudes.
This
concept
is
illustrated
by
the
following
simple
experiment
performed with
the
noise
generator.
Set
the
noise
generator
to
the
20-kc
band
and
to
maximum
output.
Connect
a
small
capacitor
(about
1000 pf)
across
the
output.
Suddenly
disconnect
the
capacitor.
Measure
its
voltage
with
an
electrometer
or
I
1
J.
D.
Cobine
and
J.
R. Curry,
"Electrical
Noise
Gen-
erators",
Proc.
IRE,
Vol. 35, No.
9,
September
1947,
pp. 875-879.
2 S.
0.
Rice,
"Mathematical
Analysis
of
Random
Noise';
Bell
System
Technical
Journal,
Vol. 23, No. 3,
July
1944, pp. 282-332; Vol. 24, No. 1,
January
1945,
pp.
46-156.
A.
van
der
Ziel,
Noise,
New York,
Prentice-Hall,
Inc.,
1954.
W.
B.
Davenport,
Jr.
and
W.
L.
Root,
An
Introduc-
tion
to
the
Theory
of Random
Signals
and
Noise,
New York, McGraw-Hill, 1958.
W.
R.
Bennett,
Electrical
Noise,
New
York, McGraw-
Hill,
1960.
Fi_gur~
2.
Elementary
Circuit
Diagram.
2
TYPE
1390-B
RANDOM-NOISE
GENERATOR
Figure
3. Oscillograms
of
Three
Different Samples
of
the
Output Voltage
Wave.
(A
single
sweep
was
used
for
each.
Middle
trace
is
at
four
times,
and lower
trace
20
times
sweep
speed
of
upper
trace.)
its
charge
with a
ballistic
galvanometer.
Record
this
value,
which
is
the
instantaneous
amplituae
of
the
noise
voltage
at
the
time
the
capacitor
is
disconnected.
A
series
of
these
values
can
be
obtained,
and a graph
prepared,
with
instantaneous
amplitude
versus
the
percentage
of
time during
which
any
amplitude
is
ex-
ceeded.
A
large
number
of
amplitudes
must
be
deter-
mined
in
this
way
before
a
reliable
distribution
results.
Two
or
three
hundred
observations
are
usually
suffi-
cient
to
show
the
trend
for
demonstration
purposes,
while
several
thousand
will
give
a
reliable
curve
for
the
important
part
of
the
range.
Because
of
the
large
number
of
observations
required,
automatic
apparatus
is
helpful
in
making
these
determinations.
3
Figure
4
illustrates
a
chart
made up
after
400
observations.
This
chart
shows
the
fraction
of
obser-
vations
in
each
interval
of
0.2
volt.
It
is
seen
that
most
of
the
observations
were
relatively
low
values,
although
some
relatively
high
values
were
observed.
These
results
are
also
shown
in
a
qualitative
way
in
the
oscillograms
in
Figure
3.
3 McKnudtzon,
"Experimental
Study
of
Statistical
Char-
acteristics
of
Filtered
Random
Noise",
Technical
Report
No. 115, M.I.
T.
Research
Laboratory
of
Elec-
tronics,
July
15, 1949.
L.
W.
Orr, "Wide-Band
Amplitude
Distribution
Anal-
ysis
of
Voltage
Sources",
Review
of
Scientific
In-
struments,
Vol. 25, No. 9,
Sept.,
1954, pp. 894-898.
3
.2sr----,------,r---
.....
-----.-----,
~
0
~
~
~
.15W.71":~17+?1\---+----I---+---l
Ill
0
u.
0
~
.IOV7'7fo.~f7t,"*~'A-'71-----f---+---1
§
g
2.0
2.5
Figure
4.
Results
of
Voltage-Sampling Experiment.
(Continuous
curve
is
a normal probability
distribution
curve
adjusted
according
to
r-m-s
value
of
noise
vol-
tage
and
size
of
intervals
used
in
plot.)
0.4
zw
Q~0.3
1-:.J
=>o
~>
tr
1-UJ
en
en
0
~0.2
)-
1-u.
:Jo
ii'i)-
<(1-
ID-
~~0.1
0
j
I
I \
1/
\
1\
\
\
v
0 -3.0
v
~
-2.0
-1.0
0
1.0
2.0 3.0
INSTANTANEOUS
NOISE
VOLTAGE
IN
TERMS
OF
R-M-S
VOLTAGE
Figure
5.
Normal Distribution Curve of True Random Noise.
The
normal
(Gaussian
or
Laplacian)
distribution•
curve
is
also
shown
in
Figure
4.
It
has
been
adjusted
according
to
the
computed
r-m·s
value
of
the
data
(the
standard
deviation)
and
the
size
of
the
interval
used
in
plotting
the
data.
The
experimental
data
fit
the
normal
curve
very
closely.
Departures
from
the
normal
curve
are
almost
entirely
the
result
of
so
few
obser-
vations.
Had
more
observations
been
made,
the
result
would
have
been
even
closer
to
the
expected
values.
3
In
Figure
5,
the
probability
that
a
voltage
be-
tween
two
limits
will
be
observed
is
given
by
the
area
under
the
normal
curve
between
those
two
limits.
Ex-
GENERAL RADIO
COMPA~Y
pressed
in
other
terms,
if
the
output
voltage
is
ob-
served
over
long
periods
of
time,
the
fraction
of
the
total
time
that
the
voltage
is
between
the
two
voltage
limits
is
given
by ·the
corresponding
area
under
the
probability
cutve.
4
For
example,
the
instantaneous
voltage
magnitude
will
be
no more
than
one-tenth
the
r-m-s
value
for
about
eight
percent
of
the
time,
and
will
be
greater
than
three
times
the
rms
value
only
about
0.
26
percent
of
the
time.
2.4
DEPARTURES
OF
OUTPUT
FROM
TRUE
RAN-
DOMNESS.
The
cutve
in
Figure
5
is
a
theoretical
curve
and
is
symmetrical
about
the
origin.
The
noise
of
the
generator
has
a
similar
distribution,
but
is
somewhat
asymmetrical
because
of
the
gas
tube.
In
addition,
the
inherent
amplitude
limitations
of
the
vacuum-tube
am-
plifiers
limit
the
distribution
cutve
at
high
levels.
Clipping
is
most
serious
on
the
500-kc
and
5-Mc
ranges.
When
a
narrow-band
filter
is
used
at
the
output,
the
distribution
becomes
more
nearly
random.
2.5
FREQUENCY
SPECTRUM
OF
NOISE.
The
mean-
ing
of
the
term
"frequency
spectrum
of
noise"
is
illus-
trated
in
the
following experiment; .
If
a
wave
analyzer,
such
as
the
Type
1900-A,
set
to
a
50-cps
bandwidth,
is
used
to
analyze
the
output
of
the
noise
generator,
a
fluctuating
meter
reading
will
be
observed
at
any
set-
ting
of
the
analyzer.
If
an
average
value
of
this
reading
is
taken
over
a
period
of
time,
this
average
value
is
an
estimate
of
the
level
in
that
50
-cycle
-wide
band.
This
level,
determined
on
any
ranges
of
the
noise
generator,
is
essentially
independent
of
the
frequency
setting
of
the
Type
1900-A Wave
Analyzer.
Thus
the
spectrum
in
this
region
is
uniform.
The
relative
spectrum
on
the
noise
can
be
determined
by
the
use
of
suitable
an-
alyzers
to
cover
the
full
range
of
the
principal
energy
regions
of
the
noise.
A
typical
result
of
such
an
anal-
ysis
is
shown
in
Figure
6 for
the
three
bands
of
the
Type
1390-B Random
Noise
Generator.
When
the
spec-
trum
is
uniform
over
a
broad
band,
as
shown
in
Figute
6,
it
is
commonly
called
"white
noise".
The
"white-
ness"
always
applies
to
a
definite
band
only.
For
ex-
ample,
if
the
noise
spectrum
is
uniform from 100
to
500
kc,
the
noise
is
referred
to
as
white
in
that
band.
It
is
customary
to
adjust
the
measuted
value
of
analyzed
noise
to
that
corresponding
to
an
ideal
filter
of
one-cycle
band
width.
Since
noise
voltage
in-
creases
as
the
square
root
of
the
band
width,
the
value
determined
on
the
Type
1900-A Wave
Analyzer
is
then
Y
50
cycles
divided
by
to
obtain
what
is
called
"spec-
1
cycle
tral
voltage
density".
This
can
be
defined
as
the
4 E. R.
Neinburg
and
T.
F.
Rogers,
"Amplitude
Dis-
tribution
Analyzer",
Radio-
Electronic
Engineering,
Vol. 46, No.
6.
December
1951, pp. 8-10.
4
I I
~~~G~
,....
'(I~
\
~
~
0
"'
"'
u
-5
I
~Okc
l'1~0kc
~
"'
0
-10
lc
5c
JOe
IOOc
Ike
JOke
FREQUENCY
RANGE
RANGE
/OOkc
I
Me
I
OM<
Figure
6.
Typical
Spectrum Level
Characteristics
tor
Type
1390-B Random-Noise Generator.
rms
voltage
corresponding
to
~he
energy
contained
within
a
band
one
cycle
per
second
wide.
NOTE
"Spectral
voltage
densiry",
although
a
conven-
ient
term,
is
sometimes
not
used
because
most
work
on
noise
deals
with
energy
level.
The
transfer
from
one
to
the
other
requires
a know-
ledge
of
the
impedance
level
in
the
circuit.
It
should
be
remembered
that
separate
noise
sig-
nals
add
on
an
energy
basis
and
that
the
noise
energy
increases
directly
with
the
noise
band-
width,
while
the
noise
voltage
increases
as
the
square
root
of
the
bandwidth.
Furthermore,
the
concept
used
here
assumes
a uniform
densiryof
the
noise
signal
over
the
band
of
the
analyzer.
It
should
not
be
used
for
discrete
co'mponents.
2.6
AJ':lALYSIS
OF
NOISE BY
CONSTANT-PERCEN-
TAGE
ANALYZERS.
If
the
ou~ut
of
the
Type
1390-B
Random-Noise
Generator
is
analyzed
by a
Type
1564-A
Sound
and
Vibration
Analyzer,
the
results
will
be
sim-
ilar
to
those
shown
in
Figure
7.
Here
the
indicated
level
increases
10
decibels
for
each
decade
increase
in fre-
quency.
This
result
can
be
understood
by
realizing
that
this
analyzer
has
a
bandwidth
that
is
essentially
a
con-
stant
percentage
of
the
center
frequency.
For
example,
at
5 kc
the
effective
band
width for
noise
is
about
160
cps,
and
at
500
cps
is
about
16
cps.
2.7
TYPE
1390-P2
PINK-NOISE
FILTER.
2.7.1
DESCRIPTION.
The
Type
1390-P2
Pink-Noise
Filter
(Figute
8)
converts
the
electrical
noise
output
0
-I
0
,?'
0 v
~
0 v
~
-5
0 •
.
-6
0
25
JO
25
100
250
1000
2500
10,000
25f>OO
FREQUENCY
IN
CYCLES
PER
SECOND
Figure 7.
Results
of
Analysis
of
Noise-Generator
Output
Voltage
by a
Type
1564-A Analyzer. {Straight
line
drawn
at
slope
of
10
db per frequency
decade.)
TYPE
1390-B
RANDOM-NOISE
GENERATOR
of
the
Type
1390-B
Random-Noise
Generator
to
"pink
noise"
(constant
energy
per
octave)
which
facilitates
measurements
with
constant-percentage-bandwidth
an-
alyzers.
It
is
designed
to
plug
into
the
output
binding
posts
of
the
Type
1390-B
Random-Noise
Generator,
but
can
also
be
used
at
any
point
in
a
system
where
this
filter
characteristic
is
needed
.
The
filter
is
an
RC
low-pass
filter
with
a
slope
of
-3
db
per
octave
from
20
cycles
to
20
kc
and
a
slope
of
-6
db
at
higher
fre-
quencies
(See
Figure
9).
For
shielding,
the
case
of
the
filter
is
grounded
to
the
LO
input
and
output
ter-
minals
.
Figure
10
is
a
schematic
diagram
of
the
filtP.r.
The
input
terminals
of
the
Type
1390
-P2
Filter
are
re-
cessed
plugs
at
the
rear
and
the
output
terminals
are
binding
posts
on
the
front
.
Figure
8.
Type
1390-P2
Pink-Noise
Filter.
2.7.2
USE WITH
THE
TYPE
1390-B RANDOM-NOISE
GENERATOR.
Plug
the
Type
1390-P2
Pink-Noise
Filter
into
the
output
terminals
of
the
Type
1390-B
Random-Noise
Generator
.
The
impedance
of
the
load
connected
to
the
output
terminals
of
the
filter
should
not
be
less
than
20
kilohms.
On
the
Random-Noise
Generator,
set
the
RANGE
switch
to
20
kc
,
the
LOW-HIGH
switch
to
HIGH,
and
the
MULTIPLY
BY
switch
to
1.0.
The
output
of
the
Pink-Noise
Filter
will
be
approximately
30
db
below
its
input
and
the
level
in
each
one-third-octave
band
will
be
approximately
17
db
below
that.
Thus,
when
the
output
meter
of
the
Random-Noise
Generator
indi-
cates
3
volts,
the
output
of
the
filter
will
be
approx-
imately
0.1
volt
and
the
level
in
each
one-
third-
octave
band
will
be
approximately
15
millivolts.
2.7.3
USE IN
OTHER
APPLICATIONS.
When
the
Type
1390-P2
Pink-Noise
Filter
is
used
in
a
system
at
some
point
other
than
the
output
terminals
of
the
Random-
Noise
Generator,
the
input
source
to
the
filter
should
have
an
impedance
of
less
than
1
kilohm.
Input
con-
nections
can
be
made
with
clip
leads
or
Type
274-MB
Double
Plugs
to
the
recessed
input
terminals.
The
im-
pedance
of
the
l
oad
connected
to
the
output
terminals
should
not
be
less
than
20
kilohms
.
2.
7.4
FREQUENCY-RESPONSE
MEASUREMENTS.
In
many
acoustical
systems,
frequency
response
mea-
surements
made
with
a
sine-wave
tone
source
are
dif-
ficult
to
interpret
because
of
the
large
amplitude
flue-
5
· 10
r""-
I'
r-....
I'
r-....
~
'
~
'\
\
·00
\
1\
IOe IOOe
llr.c
IOII.c
100
•
•
f"AEEU!NCY
Figure
9.
Typical
frequency
response
of
the
Type
1390
-
P2
Pink-Noise
Filter.
tuations
that
may
occur
. When
the
measurements
are
made
by
effectively
avera
g
ing
the
data
over
a
narrow
range
of
frequencies,
response
curve
is
considerably
smoother
and
much
easier
to
use.
In
the
past,
"warble
tones"
have
been
used
for
this
purpose
. A more
con-
venient
method,
however,
is
to
use
pink-noise
as
the
tone
source
and
a
constant-percentage-bandwidth
.
an-
alyzer
(such
as
the
Type
1564-A
Sound
and
Vibration
Analyzer)
with
one-thi
rd
octave
bandwidth
as
the
fre-
quency-determining
element
in
the
receiving
system.
5
Theresults
of
these
frequency-response
measurements
can
be
conveniently
recorded
on
the
Type
1521
Graphic
Level
Recorder.
6
2.7.5
USE AS A NOISE SIMULATOR. Some
noises
that
occur
in
nature
are
closer
in
spectral
character
-
istics
to
pink
noise
th
an
to
white
noise.
This
is
true,
for
instance,
of
the
low-frequency
noise
in
semicon-
ductors
and
of
some
acoustical
background
noises.
To
simulate
electrical
signals
generated
in
such
cases,
it
is
convenient
to
use
pink
noise.
r--
-------
---
- -
---
-l
I
/",:C
/
0/
I
c/
/0/
I @
I
I
I
I
I
HI
I so
I
URC£
I
I
I
I
LO
I
IP.
'1:
/
02
AVO/
6.49K
.eto%
.3.09K
:: C /
01
J
I
.R
/
0.3
~~0~
I
!JS3
309
I
I
::
t.~J
I
I
::
CI
OZ
::
~
C /
03
I
(7.287
0.08Gfl
I
I
I
..,~
I
L
_______
--------
----
--
_
_j
'"
L OAD
20K
)
LO
Figure
10.
Schematic
diagram
of
the
Pink-Noise
Filter.
5
..
A
New
Analyzer
for
Sound
and
Vibration"
,
General
Radio
Experimenter
,
Volume
33
,
Number
12,
December,
J959.
"A
Graphic
Level
Recorder
with
High
Sensitivity
and
Wide
Ranges"
,
General
Radio
Experimenter,
Vol-
ume
33
, Number 6,
June
,
1959.
GENERAL
RADIO
COMPANY
Section 3
INSTALLATION
3.1
BENCH
MOUNTING.
To
set
the
instrument
in
a
tilted
position
(shown
in
inset
of
Figure
1),
simply
pull
each
front
leg
down
as
far
as
possible
and
then
turn
the
leg
so
that
its
notch
faces
the
back
of
the
instru-
ment.
To
restore
the
leg
to
its
retracted
position,
turn
it
to
release
the
catch
and
push
the
leg
up.
3.2
RELAY-RACK
MOUNTING.
Type
480-P412
Panel
Extensions
are
available
to
adapt
the
Type
1390-B
Random-Noise
Generator
for
relay-rack
mounting.
To
mount
the
Type
1390-B
Random-Noise
Generator
in
a
relay
rack,
first
attach
the
two
panel
extensions
to
the
instrument.
Remove
the
two
screws
in
the
upper
and
lower
corners
on
one
side
of
the
panel.
These
screws
fasten
the
panel
to
the
aluminum
end
frames.
Place
one
of
the
extensions
in
the
front
of
the
panel
so
that
the
corner
holes
on
the
plate
line
up
with
those
on
the
instrument
and
replace
the
two
screws.
Attach
the
second
extension
on
the
other
side
of
the
instrument
panel
in
the
same
manner.
The
instru-
ment
can
then
be
mounted
in
a
standard
19-inch
relay
rack.
3.3
CONNECTION-
TO
POWER
SUPPLY.
Connect
the
Type
1390-B
to
a
source
of
power
as
indicated
by
the
legend
at
the
input
socket
at
the
rear
of
the
instrument,
using
the
power
cord
provided.
While
in-
struments
are
normally
supplied
for
115-volt
operation,
the
power
transformer
can
be
reconnected
for
230-volt
service
(see
schematic
diagram,
Figure
12). When
changing
connections,
be
sure
to
replace
line
fuses
with
those
of
current
rating
for
the
new
input
voltage
(refer
to
Parts
List).
Appropriate
measures
should
be
taken
so
that
the
legend
indicates
the
new
input
volt-
age.
On
instruments
changed
from 230
to
115
volts,
this
simply
means
removal
of
the
230-v
nameplate;
a
115-v
legend
is
marked
beneath.
For
instruments
changed
to
230
volts,
a
nameplate
{Type
5590-1664)
may
be
ordered
from
General
Radio.
Section 4
OPERATING
4.1
START-UP.
Turn
the
POWER
switch
on.
After
30
seconds,
when
the
heater
of
the
Type
6D4
thyratron
tube
has
warmed up,
plate
voltage
is
applied
by
an
in-
ternal
time-delay
relay.
(Simultaneous
application
of
heater
and
plate
voltage
would
shorten
the
useful
life
of
the
thyratron
tube
and
increase
the
drift
in
noise-
output
level
on
warm-up.)
4.2
FREQUENCY
CONTROL
The
RANGE
switch
selects
the
network
used
for
shaping
the
noise
spec·
trum. Markings
indicate
the
upper
frequency
limits
for
which
the
noise
spectrum
is
reasonably
uniform.
4.3
OUTPUT
CONTROL.
Output
controls
are
a
switch
for
selecting
LOW
or HIGH
output,
an
OUTPUT
level
control,
and
an
output
attenuator.
In
the
LOW
position,
the
switch
introduces
a 10:1
resistance
pad
after
the
gas-tube
noise
source.
This
reduces
the
effect
of
the
6
PROCEDURE
unavoidable
amplitude
limitations
of
the
vacuum-tube
amplifier
and
also
reduces
the
noise
field
radiated
ex-
ternally
by
the
instrument.
To
keep
hum
and
microphon-
ics
to
a minimum,
however,
it
is
generally
advisable
to
operate
the
instrument
in
the
HIGH
position.
The
OUTPUT
level
control
is
a
continuous-type
control
that
is
used
to
vary
the
output
voltage
from a
very
low
value
to
maximum for
either
setting
of
the
output
switch.
The
MULTIPLY
BY
switch
is
used
to
provide
low
output
levels.
It
has
multiplying
factors
of
1.0,
0.1,
0.01,
0.001,
and
0.0001.
4.4
VOLTMETER.
A
rectifier-type,
averaging
meter
measures
the
output
voltage.
It
is
calibrated
to
indi-
cate
the
rms
value
of
the
noise.
When
the
MULTIPLY
BY
switch
is
at
1.0,
the
meter
indicates
directly
the
open-circuit
voltage
at
the
output
terminals.
In
the
other
positions
of
the
MULTIPLY
BY
switch,
the
open-
TYPE 1390·8
RANDOM-NOISE
GENERATOR
circuit
output
voltage
is
the
product
of
the
meter
read-
ing
and
the
multiplier
reading.
The
spectral
voltage
density
of
the
noise
at
a
given
frequency
is
the
r·m·s
voltage
corresponding
to
the
energy
contained
within
a
band
1-cps
wide
centered
on
that
frequency.
The
typical
spectral
voltage
den-
sity
at
1
kc
with
one
volt
output
is
approximately
as
follows:
(a)
20-kc band: 5
millivolts
for
one-cycle
band.
(b) 500-kc band:
1.2
millivolts
for
one-cycle
band.
(c)
5··MC:
band:
0.6
millivolt
for
one-cycle
band.
When
an
accurate
value
is
quency,
it
should
be
measured.
are
intended
only
as
a
guide.
desired
at
any fre-
The
values
given
4.5
LOAD.
The
output
is
taken
from a 2500-ohm po-
tentiometer,
and
one
output
terminal
is
grounded.
For
a
truly
resistive
load
with
the
MULTIPLY
BY
switch
at
1.0,
the
apparent
source
impedance
is
zero when a
reading
of
the
voltmeter
is
taken
with
the
load
con-
nected,
since
the
voltmeter
reads
the
voltage
across
the
load.
As
the
output
control
is
varied
from
the
maxi-
mum
to
the
minimum
setting,
the
actual
source
imped-
ance
varies
from
about
900 ohms
to
nearly
zero. When
the
MULTIPLY
BY
switch
is
in
any
position
other
than
1.0,
the
source
impedance
is
200 ohms.
A
load
that
is
not
independent
of
frequency
will
affect
the
frequency
spectrum
of
the
output
noise.
For
example,
a
capacitor
shunted
across
the
output
termin-
als
will
decrease
the
level
of
the
high-frequency
noise
more
than
it
decreases
the
level
of
the
noise
at
low
frequencies.
The
voltmeter
is
then
less
indicative
of
the
spectrum
level
than
it
is
for a
resistive
load.
4.6
HUM.
The
hum
level
is
u&ually more
than
40 db
below
the
over-all
noise
level
in
the
HIGH
output
po·
sltwn.
This
hum
level
is
sufficiently
low
so
that
for
most
applications
there
is
no
effect
from hum,
even
when an
analyzer
with a narrow
band
is
used
for
analy-
sis.
The
relative
hum
level
in
the
HIGH
output
posi-
tion
is
lower
than
that
in
the
LOW
position.
4. 7 APPLICATIONS.
4.
7.1
GENERAL. Some
applications
of
a
noise
gen-
erator
depend
on
its
amplitude
distribution
character-
istics
(Figures
4
and
5.)
For
example,
the
amplitude
distribution
is
similar
to
that
of
speech,
music,
and
many
other
sounds
or
electrical
disturbances
that
occur
naturally
7,
while
the
amplitude
distribution
of
a
sine
7
H.
K.
Dunn
and
S. D. White,
"Statistical
Measure-
ments on
Conversational
Speech",
Journal
of
the
Acoustical
Society
of America, Vol. 11, No. 3,
Jan-
uary 1940, pp. 278·288.
7
2.0
0
SPEECH
IN
ANECHOIC
CHAMBER
-2.0
-1.0
0
1.0
INSTANTANEOUS
AMPLITUDE
2.0
Figure
lla.
Amplitude Distribution
Curves
for
Various
Sounds. (Curves
labeled
"Speech"
are
for
particular
cases
of
sounds
produced from
readings
of printed
matterS; curve
labeled
"Music"
is
an
analysis
of an
orchestral
selection
made
in
a large theaterS.
1.5
SINUSOIDAL
WAVE
INSTANTANEOUS
AMPLITUDE
Figure
llb.
Distribution Curves of a Single Sinusoidal
Wave and a Random Noise.
wave
is
entirely
different.
These
similarities
and
dif-
ferences
can
be
seen
by
comparison
of
the
distributions
of
Figure
11.
Because
of
this
characteristic,
random
noise
is
an important
signal
for
psychoacoustic
tests.
Psychoacoustic
tests
include
masking
or
interference
tests,
loudness
measurements,
determination
of
critical
bandwidths,
and
audiometric
tests.
The
techniques
used
in
making
such
tests
are
discussed
in
the
var-
ious
numbers
of
the
Journal
of
the
Acoustical
Society
W.
B.
Davenport,
Jr. "A Study
of
Speech
Probability
Distributions",
M.I. T.
Laboratory
of
Electronics,
Technical
Report
No. 148, August 25, 1950.
GENERAL RADIO
COMPANY
of
America
(for which
there
are
two
comprehensive
indexes
available)
and
in
various
psychological
jour-
nals.
A
useful
bibliography
for
these
applications
is
S. S.
Stevens,
J.
G.
S.
Loring,
and
Dorothy
Cohen,
Bibliography
on
Hearing,
Harvard
University
Press,
Cambridge,
1955,
particularly
those
references
listed
in
Sections
139 (p. 571), 157 (p. 573)
and
222-228
(pp. 579 f).
Other
applications
depend
on
the
various
pos-
sible
frequency
spectra
of
noise.
The
frequency
spec-
trum
is
independent
of
the
amplitude
distribution,
in
the
sense
that
a normal
distribution
of
amplitudes
is
possible
with
any
frequency
spectrum
-
flat,
broad,
narrow,
sloping,
or
peaked.
Systems
that
affect
one
characteristic,
however,
may
also
affect
the
other.
For
example,
nonlinear
clipping
affects
both
the
amplitude
distribution
and
the
frequency
spectrum.
Linear
filter
networks
used
on
purely
random
noise
do
not
affect
the
randomness
but
alter
the
frequency
characteristic
and
correspondingly
the
time
scale.
Linear
filter
net-
works
used
after
clipped
noise
alter
the
frequency
spectrum
and
also
tend
to
make
the
noise
more
nearly
random.
4.7.2
ELECTROACOUSTIC
TESTS.
The
Type
1390-B
Random-Noise
Generator
is
a
useful
signal
source
for
many
types
of
electroacoustic
tests,
including
loud-
speaker-response
tests.
Some
useful
discussions
of
these
tests
using
a
noise
source
are
given
in
the
fol-
lowing:
Leo.
L
Beranek,
Acoustic
Measurementsr
New York,
John
Wiley
and
Sons,
1949,
pp.
639-640,
665
f,
697-702;
and
RMA
Standard
SE-103,
Speakers
for Sound
Equipment,
April
1949,
p.
6,
Standard
Test
Signal
BA.
Other
General
Radio
instruments
useful
in
elec-
troacoustic
tests
are
the
Type
1551
Sound-Level
Me-
ter,
the
Type
1551-P1
Condenser
Microphone
System,
the
Types
1550-A
and
1558
Octave-Band
Noise
Analyzers,
the
Type
1564-A
Sound
and
Vibration
Analyzer,
and
the
Type
1521
Graphic
Level
Re-
corder.
4.
7. 3
ROOM
ACOUSTICS
TESTS.
The
noise
generator
is
a
useful
signal
source
for many
types
of
tests
in
room
acoustics.
These
include
reverberation
tests,
panel
(wall
and
floor)
transmission
measurements,
measurement
of
space
irregularities,
and
measurement
of
steady-state
signal
transmission.
For
details,
con-
sult
the
following:
Leo.
L
Beranek,
Acoustic
Mea·
surements,
New York,
John
Wiley
and
Sons,
1949,
pp.
804ff,
826f,
831
and
883.
The
Type
1551-B
Sound-Level
Meter,the
Types
1550-A
and
1558-A
Octave-Band
Analyzers,
the
Type
1564-A Sound
and
Vibration
Analyzer,
and
the
Type
8
1521
Graphic
Level
Recorder
are
useful
elements
in
the
over-all
set-up
for
these
tests.
4.7.4
STATISTICAL
DEMONSTRATIONS.
The
prop-
erties
o(
noise
that
concern
the
amplitude-time
rela-
tionship
are
usually
described
by
statistical
means
2.
(Refer
to
paragraph
2.
3.)
Random-noise
generators
can
be
used
to
demon-
strate
some
concepts
of
statistical
theory.
The
equip-
ment
and
methods
for
demonstrating
various
degrees
of
correlation
and
possible
errors
of
random
sampling
are
described
by
J.
C.
R.
Licklider
and
E.
Dzendolet,
"Oscillographic
Scatterplots
Illustrating
Various
De-
gree~
of
Correlation",
Science,
January
30, 1948, Vol.
107, No. 2770,
pp.
121-124.
4.
7.5
NOISE AT HIGH
FREQUENCIES.
The
noise
generator
can
be
used
to
modulate
an
r-f
car:ier
when
a
noise
signal
is
desired
at
a
frequency
above
5Mc.
The
Type
1000-P6
Crystal
Diode
Modulator8
is
a
suit-
able
instrument
for
wide-band
modulation,
and
the
Type
1208-B
VHF
Unit0scillators
9
and
the
Type
1209-B
UHF
Unit
Oscillators
are
suitable
rf
oscillators
cov-
ering
the
range
from 65
to
920
Me.
Because
of
the
two
sidebands
that
result
from
the
standard
modulation
techniques,
the
noise
band
can
be
made
to
extend
over
a 10•Mc
range,
5
Me
on
each
side
of
the
carrier.
For
some
applications
it
may
be
desuable
to
use
a
suppressed-carrier
or
balanced-type
modulator
(see
Terman,
Radio
Engineering
Handbook,
New
York, McGraw-Hill Book
Co.,
1943,
pp.
551-553).
It
is
also
possible
to
use
a
series
of
carriers
and
mod-
ulators
to
combine
to
give
a much
broader
band
of
noise
than
10 Me.
Some
signal
generators
and
oscillators
include
modulating
circuits,
so
that
an
external
source
such
as
the
Type
l390-B
Random-Noise
Generator
can
be
used
to
modulate
the
signal.
Instruments
of
this
type
are
the
Types
1001-A, 1021-AU,
and
1021-AV
Standard
Signal
Generators.
For
these
generators
the
modulation
produced
is
limited
to
the
audio
range
and
to
about
5
to 10
percent
rms
noise
modulation,
with
peaks
much
higher.
When a
wider
frequency
band
is
desired,
the
Type
1000-P6
or
some
other
external
modulator
should
be
used
as
described
above.
8
Byers,
W.
F.,
~An
Amplitude
Modulator for
Video
Frequencies",
General
Radio
Experimenter,
March
1950, Vol. 24, No. 10, pp. 6-8.
9
E.
Karplus,
"V-H-F
and
U-H-F
Unit
Oscillators",
General
Radio
Experimenter,
May 1950,
Vol.
24,
No. 12, pp. 7-11.
TYPE
1390-B RANDOM-NOISE GENERATOR
4.7.6
VERY HIGH
NOISE
LEVELS.
When
noise
levels
even
higher
than
those
provided
by
the
Type
1390-B
Random-Noise
Generator
are
desired,
an
amplifier
should
be
used,
such
as
the
Type
1206-B
Unit
Ampli-
fier.
The
Type
1233-A
Power
Amplifier
is
useful
when
a
wide
frequency
range
is
desired.
4.
7. 7
INTERFERENCE
TESTS.
Since
noise
is
a
common form
of
interfering
or
disturbing
signal
or
sig-
nal
that
limits
the
threshold
of
detectability,
the
noise
generator
can
be
used
to
check
receivers,
communi-
cation
systems,
and
detection
systems
for
suscepti-
bility
to
interference.
It
can
also
be
used
as
a
train-
ing
aid
for
operators
who
must
communicate
through
interference.
4.7.8
OVER-ALL
CALIBRATION
TESTS.
The
noise
generator
can
be
used
as
an
over-all
calibration
device
because
of
the
wide
frequency
range
available
at
the
output.
This
calibration
signal
can
be
particularly
use-
ful
in
audio
systems
th_at
involve
a
recording
technique,
and
its
use
can
frequently
simplify
the
calibration
pro-
cedure
when an
analyzer
forms
part
of
the
system.1°
For
example,
when a
magnetic
tape
recorder
is
used
to
record
a
signal
to
be
measured
on
playback,
reference
signals
must
be
recorded
before
and
after
the
unknown
signal
is
recorded.
These
reference
signals
permit
one
to
fix
levels
and
to
determine
response
char-
acteristics,
which
can
vary
from
time
to
time
depend-
ing
on
the
condition
of
the
tape
and
the
machine.
These
reference
signals
are
usually
a
series
of
tones
at
var·
ious
points
in
the
frequency
range
of
interest.
The
noise
generator,
due
to
the
broad
frequency
band,
per-
mits
the
use
of
a more
versatile
reference
signal.
Thus
a
useful
set
of
reference
signals
would
be
a
burst
of
noise
of
about
one-half
minute
duration
and
a
burst
of
a
400-cycle
tone
of
about
the
same
length.
These
two
signals
would
permit
the
determination
of
frequency
response,
signal-to-noise
ratio,
harmonic
distortion
(at
one
level
and
one
frequency),
and
flutter.
To
determine
the
frequency
response
by
use
of
a
noise
signal,
perform
the
following
operations:
1.
Set
the
noise
generator
to
the
20-kc
range.
Con-
ne~t
it
to
the
'input
of
the
system
under
test,
at
such
a
level
that
the
r-rn-s
input
is
at
least
14
decibels
below
the
sine-wav'e
overload
point.
2.
Make a
frequency
spectrum
analysis
of
the
input
noise
signal
and
of
the
output
noise
signal
from
the
device
under
test.
The
relative
level
of
input
and
out-
put
as
a
fuaction
of
frequency
is
then
the
frequency
10
s.
S.
Stevens,
J.
P.
Egan,
and
G.
A.
Miller, "Meth-
ods
of
Measuring
Speech
Spectra",
Journal
of
the
Acoustical
Society
of
America,
Vol. 19, No. 5,
Sep·
tember 1947, pp.
771·780.
9
response
of
the
device
under
test,
unless
spurious
sig·
nals
are
present
in
the
output
of
the
device.
3.
Test
for
spurious
signals
by
making
an
analysis
of
the
output
with
no
input
signal
applied.
When
these
measurements
are
made,
the
input
and
output
must
be
analyzed
by
analyzers
of
the
same
effective
bandwidth.
The
bandwidth
of
the
analyzer
should
also
be
appreciably
smaller
than
the
bandwidth
of
the
device
under
test.
Furthermore,
the
ultimate
attenuation
of
the
analyzer
should
be
much
greater
than
variation
in
response
that
one
expects
to
measure,
so
that
it
will
not
limit
the
observed
response.
Distortion
and
background
noise
in
the
device
under
test
will
also
limit
the
range
of
variation
in
response
that
can
be
measured
by
this
method,
and
it
is
therefore
impor-
tant
to
select
the
proper
level
for
input
signal.
4.
7.9
ANALYSIS
OF
NOISE. In
the
course
of
meas-
urements
with
a
noise
generator,
it
is
often
necessary
to
make
a
frequency
spectrum
analysis
of
noise.
The
Type
1900-A Wave
Analyzer,
the
Types
1558-A
and
1558-AP
Octave-Band
Noise
Analyzers,
and
the
Type
1564-A
Sound
and
Vibration
Analyzer
are
usefulacces·
sories
for
this
analysis
in
the
audio-frequency
range.
The
results
of
noise
analyses
by
these
different
analyzers
cannot
be
compared
directly;
the
results
must
be
modified
because
of
the
different
bandwidths.
Refer
to
paragraph
2.5 for a
discussion
of
the
fre-
quency
spectrum
of
noise.
40
m
-'
w
35
30
5
20
"''
5
~
0
10
5
5
/
/ L
/
v/
~~/
/ v
~
\
~"
»o"'<>
/ ~
/
//
/ / /
//
50
2..5
10
25
100
250
1000
2500
10,000 2 •
00
FREQUENCY
IN
CYCLES
PEA
SECOND
Figure
12.
Decibels
to Be
Subtracted
from
Type
1564-A
Reading
to
Obtain
Spectral-Density
Level.
The
bandwidths
of
the
Octave-Band
Noise
Analyzers
increase
in
cycles
directly
with
the
mean
frequency
of
the
band.
For
that
reason,
a
noise
signal
that
is
uniform
in
spectral-energy
density
over
the
frequency
range
will
give
higher-level
readings
for
the
higher-frequency
bands
than
for
the
lower-frequency
bands.
The
following
table
shows
the
values
to
be
subtracted
from
the
readings
of
the
analyzer
to
obtain
the
spectral-density
levels.
GENERAL
RADIO
COMPANY
Type
DB
Type
1550-A
to
be
1558-A
Band sub- Band
tracted
20-75 18 18.75-37.5
75-150 19 37.5-75
150-300 22 75-150
300-600 25 150-350
600-1200 28 350-600
1200-2400 31 600-1200
2400-4800 34 1200-2400
4800-10kc 38 2400-4800
4800-9600
9600-19,200
LP
7.5
The
effective
bandwidth
of
the
Type
1564-A
Sound
and
Vibration
Analyzer
increases
with
increase
in
the
frequency
to
which
the
analyzer
is
tuned.
The
graph
in
Figure
12
shows
the
value
in
decibels
that
must
be
subtracted
from
the
reading
of
the
analyzer
to
obtain
the
spectral-density
level.
This
value
is
determined
on
the
basis
of
initial
calibration
of
the
instrument
by
a
sine-wave
signal.
The
corrections
for
spectral
density
level
for
the
Type
1900-A Wave
Analyzer
are
independent
of
the
center
frequency
to
which
it
is
tuned
but
do
de-
pend
on
the
bandwidth
used.
For
the
3-cycle
band-
width
subtract
3. 7 db;
10-cycle,
subtract
9 db;
50-cy-
cle,
subtract
15.9
db
to
obtain
the
spectral
density
level.
(These
correction
numbers
take
into
account
the
metering
characteristic
as
well
as
the
bandwidth.)
4.
7.10
FREQUENCY
RESPONSE
MEASUREMENT.
The
noise
generator
can
be
used
in
place
of
the
usual
sine-wave
oscillator
for
measuring
the
response
of
circuits
and
systems.
In
this
application
the
selective
characteristics
of
generator
and
detector
are
reversed
from
those
ordinarily
used
in
point-by-point
measure-
ments;
the
wide-band
noise
source
and
a
selective
detector
replace
the
single-frequency
source
and
wide-
band
detector.
For
speech
and
music
circuits,
this
technique
provides
a much
closer
approximation
to
operating
conditions
than
does
the
older
system.
This
approach
is
particularly
useful
in
testing
recording
systems.
10
The
usual
sweeping
sinusoidal
tests
are
sometimes
inconvenient
because
of
the
problem
of
de-
termining
the
recorded
frequency
during
playback.
This
11
L. L.
Beranek,
Acoustic
Measurements,
New York.
John
Wiley, 1949, pp. 665-668
and
697-702.
12 Emory
Cook,
"White-Noise
Testing
Methods",
Audio
Engineering,
Vol. 34, No. 3, March
~950,
pp. 13-15.
13
J.
P.
Vasseur,
"Les
foisceaux
hertziers
a
courants
10
DB
Type
DB
to
be
1558-AP
to
be
sub- Band sub-
tracted
Center
tracted
freq.
13
31.5 13.5
16 63 16.5
19 125 19.5
22
250 22.5
25
500 25.5
28
1,000 28.5
31
2,000 31.5
34 4,000 34.5
37
B,OOO
37.5
40 16,000 40.5
18 I
problem
is
eliminated
by
a
recorded
n01se
signal
that
is
analyzed
on
playback.
Because
of
its
broad
frequency
spectrum,
noise
is
often
used
to
avoid
the
marked
resonance
effects
that
can
occur
when
vibrations
in
mechanical
struc-
tures
and
acoustical
systems
are
measured.
The
use
of
noise
as
a
source
in
measuring
the
reverberation
characteristics
of
rooms
and
the
transmission
charac-
teristics
of
building
structures
results
in
a
type
of
elettrical
averaging
of
the
characteristics,
provided
a
reasonably
broad
band
is
us·ed.
This
averaging
often
simplifies
the
comparison
of
the
characteristics
of
dif-
ferent
structures.
The
noise
generator
is
useful
in
response
meas-
urements
on
loudspeaker
systems
in
rooms.11
The
elec-
trically
averaged
response
can
be
used
to detet:mine
the
optimum
characteristic
for
equalizing
networks,
to
set
the
relative
levels
of
woofer
and
tweeter
units,
and
to
adjust
levels
of
multiple-speaker
units
installed
in
different
loca,tions
in
a
large
hall.
4.
7.11
RESONANCE
TESTS.
Because
of
its
broad
frequency
spectrum,
noise
can
sometimes
simplify
the
h f d. . .
12
h
searc
or
resonant
con
ttwns
1n
a
system.
T e
resonance
produces
a
peak
in
the
frequency
spectrum,
v.-hich
can
be
observed
in
oscillographic
displays.
4.7.12
OTHER
USES.
The
noise
generator
can
also
be
used
in
crosstalk
measurements
13,
for
masking
crosstalk
in
multichannel
communication
systems
14,
to
drive
vibrators
in
component
testing
15, for
noise
factor
comparison
tests
16, or
distortion
measurements
17•
porteurs
devant
les
recommandations
du
C.C.I.
F.",
Annales
de
Radioelectricite,
Vol. 9, No. 35,
Janu-
ary
1954, pp.
47-82and
EIA
Standard
RS-252,"Base
band
Characteristics
of
the
Microwave
Radio
and
Multiplex
Equipment,"
October,
1961.
TYPE
1390-B
RANDOM-NOISE
GENERATOR
14
A.
J.
Aikens
and
C. S.
Thaeler,
"Noise
and
Cross-
talk
on
N1
Carrier
Systems",
Electrical
Engineering,
Vol.
72, No. 12,
December
1953, pp.
1075-1080.
15
J.
Robbins.
"Standardized
White
Noise
Tests",
Elec·
tronic
Industries
&
Tele-Tech,
Vol16,
No. 2,
February,
1957'
pp.
68-69.
S. H.
Crandall,
ed.,
Random
Vibration,
Cambridge,
Massachusetts,
The
Technology
Press
of
MIT, 1958.
16
IRE
"Standards
on
Electron
Devices:
Methods
of
Measuring
Noise",
Proceedings
of
the
IRE,
Vol. 41,
No.
7,
July
1953, pp.
891-896.
17 A.
P.
G.
Peterson,
"Intermodulation
Distortion,"
1957
IRE
National
Convention
Record,
Vol.
5,
Part
7,
March,
1957,
pp
51·58.
J.
S.
Murray
and
J.
M.
Richards,
"Non-linearity
Distortion
Measurements,"
Wireless
World,
Vol.
69,
No.4,
April,
1963,
pp
161-165.
Section 5
SERVICE
AND MAINTENANCE
5.1
GENERAL.
The
two-year
warranty
given
with
every
General
Radio
instrument
attests
the
quality
of
materials
and
workmanship
in
our
products.
When
dif-
ficulties
do
occur,
our
service
engineers
will
assist
in
any
way
possible.
In
case
of
difficulties
that
cannot
be
eliminated
by
the
use
of
these
service
instructions,
please
write
or
phone
our
Service
Department,
giving
full
informa-
tion
of
the
trouble
and
of
steps
taken
to
remedy
it.
Be
sure
to
mention
the
serial
and
type
numbers
of
the
in-
strument.
Before
returning
an
instrument
to
General
Radio
for
service,
please
write
to
our
Service
Department
or
nearest
district
office
(see
back
cover),
requesdng
a
Returned
Material
Tag.
Use
of
this
tag
will
insure
proper
handling
and
identification.
For
instruments
not
covered
by
the
warranty,
a
purchase
order
should
be
forwarded
to
avoid
unnecessary
delay.
5.2
ACCESS
TO
COMPONENTS.
To
remove
the
shield,
loosen
the
two
fluted
locking
screws
on
the
back
of
the
instrument.
These
will
unlock
on
the
first
turn,
but
should
be
loosened
the
maximum amount
before
the
shield
is
removed.
Pull
the
shield
straight
back
from
the
panel.
5.3
PRELIMINARY CHECKS.
If
the
noise
generator
is
inoperative,
make
the
following
checks
before
re-
moving
the
case.
Make
sure
that
the
ac
supply
is
plugged
into
a
live
pow.er
line,
that
the
power
switch
is
turned
to
the
POWER
position,
that
the
output
control
is
turned
up,
that
the
time-delay
relay
is
operative,
and
that
the
fuses
are
not
open.
5.4
METER
DOES
NOT
READ.
Disconnect
any
load
from
the
output
terminals
to
insure
that
the
load
imped·
ance
is
not
too
low
to
permit
a
meter
reading.
Place
the
output
switch
in
the
HIGH
position.
11
5.5
TUBE
REPLACEMENT.
Since
the
vacuum
tubes
have
a
shorter
life
on
the
average
than
the
other
com·
ponents
used
in
the
instrument,
they
should
be
tested
if
the
instrument
is
still
inoperative
after
·the
above
checks
have
been
made.
The
Type
6D4
Gas
Triode
used
as
the
noise
source
is
aged
and
selected
for uniformity
of
the
noise
spectrum
and
for
good
amplitude
characteristics.
If
the
tube
has
deteriorated
and
must
be
replaced,
some
selection
among
different
tubes
of
this
type
may
be
necessary
to
find a
satisfactory
replacement.
5.6
HEATER
VOLTAGE
OF
TYPE
6D4 GAS
TUBE.
The
potentiometer
at
the
rear
of
the
chassis
is
for
set·
ting
the
heater
voltage
of
the
gas
tube.
Over
certain
ranges
of
heater
voltage,
some
gas
tubes
will
"sputter".
The
meter
indication
on
the
20-kc
range
will
often
re·
veal
this
sputter.
Under normal
conditions,
the
meter
reading
fluctuates
two
or
three
percent.
When
sputter-
ing
occurs,
the
meter
reading
may
fluctuate
10
percent
or
more.
The
sputtering
is
more
easily
detected
by
ob·
servation
of
the
noise
pattern
on an
oscilloscope
or
by
ear
with
a
good
pair
of
earphones.
The
heater
voltage
should
be
set
so
that
this
sputtering
does
not
occur.
Some
selection
among
tubes
of
this
type
may
be
neces·
sary
to find a
tube
that
is
free from
sputtering.
5.7
VOLTAGE
MEASUREMENTS.
Thefollowingtables
give
test
voltages
for
aid
in troubleshooting:
T1 TRANSFORMER MEASUREMENTS
Between
Terminals
AC
Volts
10
and
11 120
5
and
6 17
6
and
7 17
8
and
9
6.3
TUBE
V1
(3-4)
V2
(6D4)
V3
(6AQ5)
GENERAL
RADIO
COMPANY
TABLE
OF
VOLTAGES
PIN
VOLTS
TUBE
PIN
2
13.4
V3 6
7
6.3
(6AQ5)
(cont.)
3
6.3
4 0 V4 1
7
16.0
(6AQ5) 2
3
to
4
1 0 5
2
5.5
6
3 20
4
13.4
5 230 (A)
V5
4
215
(B)
(115N030T)
9
45 (C)
NOTES
Voltages
are
measured
with
a
20,000-ohms-per-volt
voltmeter
and
are
de with
respect
to
ground
unless
otherwise
stated.
(A)
S3 = 5
Me
(B) S3
·=
500
kc
(C)
S3 = 20
kc
*
Voltages
are
measured
with a
1000.0/volt
rectifier
meter,
and
are
with
respect
to
terminal
11 on
transformer
Tl.
12
VOLTS
150 (A)
150
(B)
110 (C)
0
13
6.3
ac
165
245
120 ac*
120 ac*
TYPE
1390-B
RANDOM-NOISE
GENERATOR
PARTS
LIST
PART
NO.
(NOTE
A)
RESISTORS
(NOTE
B)
CAPACITORS
(NOTE
C)
R1
470
±
5%
5w
REP0-22
C1A
25}
200
dcwv
COE-51
R2
330
±
5%
5w
REP0-22
C1B
25
R3
220
±
5%
5w
REP0-22
C2A
50}
R4
5.1
k
±10%
1 w REC-30BF
C2B
25
450
dcwv
COE-10
R5
10
k
±10%
2w
REC-41BF
C2C
25
R6
10k
±10%
2w
REC-41BF
C3A
50}
R7
56
±
5%
1!2w
REC-20BF
C3B
25
450
dcwv
COE-10
R8
250
±10%
POSW-3
C3C
25
R9
150
k ±
5%
1!2w
REC-20BF
C4A
800}
R10
16
k ±
5%
1!2w
REC-20BF
C4B
400
25
dcwv
COE-39
Rll
330
±
5%
1!2w
REC-20BF
C4C
400
R12
100
±
5%
5w
REP0-22
C5
1.0
j.J.pJ
±10%
COC-1
R13
2.2 k ±
5%
1/2w
REC-20BF
C6
0,47
±10%
600
dcwv
COP-19
R14
56
±
5%
1!2w
REC-20B.F
C7
15
300
dcwv
COE-27
R15
56
k ±
5%
1/2w
REC-20BF
CSA
50}
R16
330
±
5%
1/2w
REC-20BF
CSB
25
450
dcwv
COE-10
R17
68
±
5%
1/2w
REC-20BF
esc
25
R18
22
k ±
5%
1/2w
REC-20BF
C9
680
f.Lf.Lf
±10%
COM-20B
R19
2.2 k ±
5%
5w
REP0-22
C10
470
f.Lf.Lf
±10%
COM-20B
R20
470
k ±
5%
1/2w
REC-20BF
Cll
0.0022
±10%
600
dcwv
COL-71
R21
56
±
5%
1!2w
REC-20BF
C12
300
f.Lf.Lf
±1
0%
COM-20B
R22
180
±
5%
1/2w
REC-20BF
C13
0.01
±20%
500
dcwv
COC-62
R23
150
±
5%
1/2w
REC-20BF
C14
5-20
f.Lf.Lf
±10%
COT-18
R24
15
k ±
5%
5w
REP0-43
C15
0.33
±10%
600
dcwv
COP-19
R25
1 k ±
5%
1 w REC-30BF
C16A
1500}
R26
1 k ±
5%
1!2w
REC-20BF
C16B
750
10
dcwv
COE-9
R27
220
±
5%
1/2w
REC-20BF
C16C
750
R28
5.1
k
±10%
1 w REC-30BF
C17
0.01
±20%
500
dcwv
COC-62
R32
1.5
±
5%
5w
REP0-22
C18
50
f.Lf.Lf
±10%
COC-21
R33
2.5
k
±10%
POSC-7
C19
50
f.Lf.Lf
±10%
COC-21
R34
15
±10%
2w
REW-3C
C20
so
f.Lf.Lf
±10%
COC-21
R35
2 k ±
1%
1/4w
REF-65
C21
4
70
f.Lf.Lf
±1
0%
COM-20B
R36
2 k ±
1%
1!4w
REF-65
R37
2k
±
1%
1/4w
REF-65
C100
6.8f.Lf.Lf
±
5%
COC-1
R38
2 k ±
1%
1/4w
REF-65
C101
100
3
dcwv
COE-46
R39
246.9
:±0.25%
700-352
C102
100
3
dcwv
COE-46
R40
246.9
:±0.25%
700-352
DIODES
R41
246.9
:±0.25%
700-352
R42
222.2
:±0.25%
700-351
0100
1N995
0101
1N995
FUSES
R100
2.7 k ±
5%
1!2w
REC-20BF
F1
(for
115
v)
0.6 amp Slo-Bio
R101
4.3 k ±
5%
1/2 w REC-20BF
3AG
FUF-1
R102
5.1
k ±
5%
1/2w
REC-20BF
Fl
(for
230
v)
0.3
amp Slo-Bio
R103
5 k
±20%
POSC-22
3AG
FUF-1
Continued
13
F2
F2
L1
L2
L3
Ml
Pl
GENERAL RADIO COMPANY
PARTS
LIST
(Continued)
FUSES
(Continued) RECTIFIERS
{for
115
v)
0.6
amp
Slo-Bio
RXl
1Nl695
RX3
3AG
FUF-1
RX2
1N1695
RX4
(for 230
v)
0.3
amp
Slo-Blo
SWITCHES
3AG
FUF-1
INDUCTORS
51
SWT-333NP
53
52
SWRW-180
54
22
jLh
±10%
CHM-1
TRANSFORMER
50
mh
379-35
0.5
mh
CHA-597A
Tl
METER
TUBES
0-200
JLa
680 n
MEDS-92
Vl 3-4
V4
PiLOT
LAMP
V2
6D4
V5
V3
6AQ5
Mazda
Type
44
2LAP-939
NOTES:
(A)
Type designations
for
resistors
and
capacitors are
as
follows:
COC-
Capacitor, ceramic POSC- Potentiometer, composition
COE
-Capacitor, electrolytic
POSW-
Potentiometer, wire-wound
COL -Capacitor, oil-impregnated
REC
-Resistor, composition
COM- Capacitor, mica
REF-
Resistor,
film
COP-
Capacitor, plastic
REPO-
Resistor, power
COT
-Capacitor, trimmer
(B)
All
resistances are
in
ohms unless otherwise designated
by
k (kilohms).
(C)
All
ca:>acitances are
in
microfarads unless otherwise designated
by
JLJLf
(micromicrofarads).
14
1N1692
1N1692
SWRW-181
SWRW-182
485-497
6AQ5
115N030T
TYPE
1390-B RANDOM-NOISE GENERATOR
Rotary switch
sections
are
shown
as
viewed
from
the
panel end of the
shaft.
The
first
digit
of
the
contact
number
refers
to
the
section.
The
section
nearest
the
panel
is
1,
the
next
section
back is 2,
etc.
The
next two
digits
refer to
the
contact.
Contact
01
is
the
first
position
clockwise
from
a
strut
screw (usu-
ally
the
screw above
the
locating key), and
the
other
contacts
are
numbered
sequentially
(02, 03, 04, etc),,
proceeding
clockwise
around
the
section.
A
suffi~
F or R
indicates
that
the
contact
is
on
the
front or
rear of
the
section,
respectively.
WH-Y£
FOR
Tl
CONNECTIONS: rt
VT
1/SV
OPt:RATION CONNECT
#I
TO#.J
8#2T0#4
230V
OPERATION CONNECTN
TO#iJ
115/230V
50-60c.
IlK
Rl
RXI
470
5W
WH-RO-BR
WH-BL-BK
OAr
.....
15
R2
$$0
5W
MULTIPLY
BY
O.ol
··~o·
54
ENGRAVING
R3
220
5W
R28
5./K
IW
WH-HO-BL
WH-GN-BL
A 0 =
C:JC
C$8
C:JA
25)Jf 5ul 50)Jf
WH-RO-BK
WH-GN-BK
RANGE
SOOKC
"'tf
5:1 ENGRAVING
Rl3
2.2k
RIO
l6K
Rll
330
LOW
HIGH
C6
.47
16A
1500uf
a
52
ENGRAVING
WH-BR-BK
Rl9
R/5
2.2K
56K
c
5W
68
Figure
12.
Schematic W
LOW
HIGH
S2
ENGRAVING
---
WH·8fi-8K
-
-~::::
C9
680
POWER
DELAYED ON
OFF
Sf
ENGRAVING
::::::::
cil
R17
.0022ut._
4..-I\
68
Nr--.
$3
·~
o-
--r----
..
~
1 T
R/8
22K
Rl6
330
R24
R26
IK
R25
IK
IW
~mh
GN
R27
220
RESISTORS
1/2
WATT UNLESS
OTHERWISE
SPECIFIED
RESISTANCE
IN
OHMS UNLESS
OTHERWISE
SPECIFIED
k=IOOO
OHMS
M•l
MEGOHM
CAPACITANCE
VALUES
ON£ AND
OVER
IN
MICRO-MICRO-
FARADS LESS
THAN
ON£
IN
MICROFARDS UNLESS
OTHERWISE
1ffj1FIED
R41
.t:f
246.9
IL
54
"'~'~..::."+......;r--,
..
1
'•.!.
1
'?/~~
...
~
,.,(
~
1
r1Ji.e
'----r-...11
L--~--------------.-------------~-~~--17
V3
1
~6AQ5
''iA{
'f/
=:GN·I*I
~C~6A
/500uf
-::
DA
'~8s,c
50pf
H20
470K
C21
470;.
DA
R2.1
CI6B,C
::::!
50
/500pl
Figure
12.
Schematic Wiring Diagram.
R21
~6
C8A
50pf
.;,..
cf'
15pl
•
R.J3~
2.5K
'"ccw
'-ouTPUT
,Dtoo
,;'"
~~
hRtaJ
~5K
""
0ta
I
~~
I
R.-
I
5./K'
I
I
t
I
~,O.tl
cto2
r-.
I
/()()#
+ I
~.t2
L
________
_j
I•
L-------~+~
r------~
Ml
V2
ca---L
__
_
Cl6 4
V4-~~(Ck-:--
,'
-~
n
CIOI
~
~
~8
RI03
~
~
G
Cl02
Tl
Figure
10
.
Top
Interior
View.
VI
R37
R33
V5
R34
GENERAL
RADIO
COMPANY
R5 R6
RXI
~~~~::::::~::;-L3
~----R20
Cl3 l
Cl?
Figure
11. Bottom
Interior
View
.
16
GENERAL
RADIO
COMPANY
WEST
CONCORD,
MASSACHUSETTS
01781
617
369-4400
617
646-7400
SALES
ENGINEERING
OFFICES
NEW
ENGLAND*
22
Baker
Avenue
West
Concord, Massachusetts
017B1
Telephone
617
646-0550
METROPOLITAN
NEW
YORK*
Broad
Avenue
at
Linden
Ridgefield,
New
Jersey
07657
Telephone N.Y.
212
964-2722
N.
J.201943-3140
SYRACUSE
Pickard Building
East Molloy
Road
Syracuse,
New
York
13217
Telephone
315
454-9323
PHILADELPHIA
Fort
Washington
Industrial Park
Fort
Washington,
Pennsylvania
19034
Telephone
215
646-B030
WASHINGTON*
and
BALTIMORE
11420
Rockville Pike
Rockville, Maryland
20B52
Telephone
301
946-1600
ORLANDO
113
East Colonial Drive
Orlando, Florida
32B01
Telephone
305
425-4671
•
Repair
services
are
available
at
these
offices.
CHICAGO*
6605
West
North
Avenue
Oak
Park, Illinois
60302
Telephone
312
B4B-9400
CLEVELAND
5579
Pearl
Road
Cleveland,
Ohio
44129
Telephone
216
BB6-0150
LOS
ANGELES*
1
000
North
Seward
Street
Los
Angeles,
California
9003B
Telephone
213
469-6201
SAN
FRANCISCO
11B6
Los
Altos
Avenue
Los
Altos, California
94022
Telephone
415
94B-B233
DALLAS*
2600
Stemmons Freeway, Suite
210
Dallas, Texas
75207
Telephone
214
637-2240
TORONTO*
99
Floral Parkway
Toronto
15,
Ontario,
Canada
Telephone
416
247-2171
MONTREAL
1255
Laird Boulevard
Town
of
Mount Royal,
Quebec,
.Canada
Telephone
514
737-3673
General
Radio
Company
(Overseas),
BOOB
Zurich,
Switzerland
General
Radio
Company
(U.K.) Limited, Bourne End, Buckinghamshire, England
Representatives in Principal
Overseas
Countries
P
rin
ted in
USA
