This is only a preview of the June 1991 issue of Silicon Chip. You can view 41 of the 96 pages in the full issue, including the advertisments. For full access, purchase the issue for $10.00 or subscribe for access to the latest issues. Articles in this series:
Articles in this series:
Items relevant to "Four-Channel Lighting Desk, Pt.1":
Items relevant to "13.5V 25A Power Supply For Transceivers, Pt.2":
Articles in this series:
Articles in this series:
Articles in this series:
|
Howl suppression for
public address systems
Acoustic feedback is a problem with sound
reinforcement systems. At a particular sound
level, enough of the signal will be coupled back
to the microphone so that a positive feedback
loop is set up. This experimental circuit reduces
this problem by shifting the whole audio
spectrum up by 10Hz.
By DARREN YATES
If you've ever been in a lecture
theatre or even at a concert, then
you've probably heard the loud squeal
that can occur due to acoustic feedback. This effect limits the number of
locations where you can place loudspeakers, as well as how loud you
can drive them.
Acoustic feedback occurs when the
resonance of a room at a particular
frequency induces a positive feedback
between the microphone and loudspeaker. Since theory says we only
need a gain of slightly more than unity
to start and sustain oscillation, the
inevitable effect is that the PA system
turns into a high power oscillator at
the particular frequency of resonance.
35
0
30
25
a,
~ 2ll
--- - -
~
n
I
n(\
MEANJlESPONg_
How it works
This circuit concept was originally
a design idea published a few years
ago in "Electronics & Wireless World".
That circuit used a couple ofMC1495
analog multipliers. We recently decided to take another look at the concept and use some lower cost and
more readily available devices such
as the LM13600/LM13700.
A simplified block diagram of the
frequency shifter is shown in Fig.2
while the full circuit is shown in
Fig.3.
It consists of, firstly, an audio stage
using IC1a and IC1b, which incorporates a passive quadrature filt er net-
t---
r
"'
z
0
c..
~
The common way used by PA operators to overcome this problem is to
introduce a dip in gain at the frequency of resonance, using a parametric equaliser or a one-third octave
equaliser. This lets the overall gain of
the system be raised until acoustic
feedback again becomes the limiting
factor. In practice , several dips may
have to be used to make the system
work at a sufficiently high level.
This article discusses a different
technique. If we take the entire audio
frequency spectrum and shift it up by
about 10Hz, we change the frequency
response relationship of the amplifier and the room so that what used to
be a frequency of resonance now becomes a frequency of something like
"anti-resonance".
To make this a little clearer, the
graph in Fig.1 shows a typical room
response to a section of the lower
audio spectrum. As you can see, some
of the p eaks are well above the average room response. It is at these fre quencies that we're likely to get feedback.
You'll also notice that every few Hz
or so above each frequency of resonance there is one of anti -resonance;
ie, a dip. The idea behind this circuit
is to shift the frequency spectrum up
so that the frequency which is accentuated by the room response is now
reproduced as a frequency which corresponds to a dip in the response, or
so the theory goes.
The end result is that we can increase the gain further before instability occurs.
/
.-----ix
15
"'
10
\I
V
OUTPUT
500
AUDIO
MIXER
AUDIO
STAGE
0
520
540
560
580
FREQUENCY=
INPUT•+ 1OHz
600
FR EOUENCY (Hz)
'-----.-tX
Fig.1: this graph shows a typical room
response for the lower part of the
audio spectrum. Acoustic feedback
problems occur at the peaks.
XY/101-----'
Fig.2: block diagram of the frequency shifter. It shifts the
incoming frequency by lOHz.
JUN E 1991
5!
al
Q
,--------------<l>------+12V
CJ)
11
r
n
0
z
n
::r:
"O
.001
0.1
10k
.001
VR3
50k
- 12V
0.1
+12V
~
0.1
202
560k
01
WI,,
r270k
.,.
VR6
100k
270k
270k
~ .I
-:120k
270k
...
011'.
-1 2V
- 12V
01!
VR5
50k
+12V
D1
1N4002
OUT
12VAC
PLUG -PACK
,I:-
I
r
+12V
0.1
100
-
D2
1N4002
.1
I
---
o,Il
OUT
GNO
I
I I
....
I
+V
- 12V
Fig.3: ICla & IClb generate two quadrature signals which drive the X inputs of 4-quadrant
multiplier stages IC2a & IC2b. The Y inputs are driven by a lOHz quadrature sinewave
oscillator consisting of IClc & ICld & the outputs then mixed & fed to amplifier stage IC3.
- 12V
0.47
OUTPUT
V
OUTPUT IC2a
OUTPUT IC2b
Fig.4: trimpots VR3 & VR5 should be
adjusted to obtain amplitude
modulated waveforms at pins 8 & 9 of
IC2, as shown here.
work. The two quadrature signals (ie,
out of phase by 90 degrees) are then
connected to the X inputs of two 4quadrant multipliers, IC2a and IC2b.
A quadrature sinewav·e source of
about 10Hz is produc ed by IClc and
ICld , and is fed into the Y-inputs of
the two multipliers (IC2a & IC2b). The
outputs are then mixed together and
amplified by IC3 , where the final output is then connected directly to the
PA system.
In more detail, the input signal is
amplified by the non-inverting amplifier ICla and an inverted copy is
produced by inverting buffer IClb.
IC1 is an RC4136, chosen for its low
noise and low cost.
The passive wideband quadrature
network is connected between the two
outputs of ICla and IClb. This produces two copies of the input audio
signal which are 90 degrees out of
phase.
IC2 is a dual transconductance
amplifier package, the LM13600. Each
amplifier is connected as a 4-quadrant analog multiplier. The outputs
from the two passive networks are
connected via 4. 7kQ resistors to the
X 0 inputs , which in this case are the
inverting pins of IC2a and IC2b .
10Hz oscillator
ICl c and ICld form the 10Hz quadrature sinewave oscillator. Th e two
6.8V zener diodes provide amplitude
regulation and the amplitude itself is
governed by the 100kQ pot, VR4.
The two quadrature sinewaves are
taken from output pins 10 & 12 and
are connected to the Y-inputs of the
two multipliers via two 50kQ pots,
VR3 and VR5. These are really the
inputs to the amplifier bias network,
which control the gain of the amplifier and make it act as a multiplier.
The output of each amplifier section
is then connected external ly to the
internal buffers (from pin 5 to pin 7
and from pin 12 to pin 10 of IC2) to
provide a low -impedance output.
The signals from each of the multipliers are then AC-coup led and mixed
together via 100kQ pot VR6. IC3 is a
TL071 FET-input op amp connected
up as 'a mixing amplifier with a gain
of about four. The output signal at
pin 6 is then connected straight to the
input of the PA system.
The power supply is ±12V DC and
can be derived from a 12VAC plug
pack using a 7812/7912 3-terminal
regulator combination.
Setting up
The setting up procedure is as follows:
(1) Adjust VR4 so that the amplitude of the sinewave outputs at pins
10 and 12 of IC1 is about 9 volts p-p
(about 3 volts RMS).
(2) Connect a lkHz sin ewave source
to the input of ICla. Adjust VR1 and
VR2 so that the respective outputs at
pins 8 and 9 of IC2 are as symmetrical
as possible about the centre-line.
What you'll find is that these outputs
will tend to skew either up or down
initially' and as you rotate the pot,
they will quickly come into line.
(3) Adjust VR3 and VR5 so that you
get an amplitude modulated waveform at the respective outputs . The
signal envelopes you end up with
should be similar to those in Fig.4.
As you adjust VR3 and VR5, you
should get a normal amplitude modulated waveform and as you turn the
pot, the waveform should appear to
fold on top of itself, to produce a
"bow-tie" effect (as in Fig.4).
If you have a CRO handy, connect
one probe to the output at pin 8 ofIC2
and the other to pin 9. What you
should see are two similar waveforms,
except that where one waveform appears to be at a min imum , the other
should be at its maximum amplitude
as shown in Fig.4.
Finally, adjust VR6 so that the output of IC3 has as little or no 10Hz
modulation as possible. Because the
outputs of the two multipliers have
the 10Hz carrier suppressed, mixing
the output should give a steady amplitude signal with its frequency equal
to the input+ 10Hz; ie, 1010Hz.
Note that wh il e this frequency shift
concept do es work, in practice we
found that it was critical to set up and
produced frequency non-l ineari ties of
its own.
SC
PHONE LINE AND MAINS
FILTER/SURGE PROTECTOR
Protection for
•Answering machines
•Cordless phones
,,Facsimiles
· •Computers
•Modems
•Telex
The Arista model CPEP-1
ensures that
lightning, power surges
and spikes do not
reach your costly
communication
equipment as it
constantly monitors the
PHONE line and
the MAINS power line.
Simply connect
to any existing power point,
plug in your fax, phone, video, Hi Fi or any other
240 VAC operated equipment and a series of
indicators shows the status of the mains
power point while several MOV's and gas
arrestors prevents access to damaging energy
sources. The CPEP-1 is fully Telecom and Dept.
of Minerals and Energy tested and approved.
Imported and distributed by:
.ARIST1\
ELECTRONICS
Available through the following retailers:
Telegrafax PIL.
305-307 Sailors
Bay Road.
Northbridge. 2063.
NSW. (02) 958 5137
Geoff Wood Elec.
229Burns
Bay Road.
Lane Cove. 2066.
NSW. (02) 427 1676.
UV PROCESSING
EQUIPMENT
e
KALEX LIGHT
BOXES
UV2
Exposure Area 40cm x 27cm
$650.00
• Portuvee 4
Exposure Area 25cm x 14cm
$225.00
• Portuvee 6
Exposure Area 43cm x 40cm
$625.00
3M Scotchal Photosensitive
Riston 4600 PCB Materal
All prices are plus sales tax if applicable
KALEX
1~1.iMc¼r,i
,I,, - , . .,
40 Wallis Ave
East Ivanhoe
VIC. 3079
(03) 497 3422
(03) 497 3034
Fax (03) 499 2381
ELECTRONIC COMPONENTS & ACCESSORIES
e SPECIALIST SCHOOL SUPPLIERS
Ju,w: 1991
61
|