AN390 DNR Applications Of The LM1894 AN 0390

User Manual: AN-0390

Open the PDF directly: View PDF .
Page Count: 10

Download
Open PDF In Browser	View PDF

National Semiconductor
Application Note 390
Martin Giles
Kerry Lacanette
March 1985

INTRODUCTION
The operating principles of a single-ended or non-complementary audio noise reduction system, DNR, have been
covered extensively in a previous application note AN384,
Audio Noise Reduction and Masking. Although the system
was originally implemented with transconductance amplifiers (LM13600) and audio op-amps (LM387), dedicated I/Cs
have since been developed to perform the DNR function.
The LM1894 is designed to accommodate and noise reduce
the line level signals encountered in video recorders, audio
tape recorders, radio and television broadcast receivers,
and automobile radio/cassette receivers. A companion device, the LM832, is designed to handle the lower signal levels available in low voltage portable audio equipment. This
note deals chiefly with the practical aspects of using the
LM1894, but the information given can also be applied to
the LM832.

In the majority of these applications the circuit used is identical to that shown in Figure 1, and this is the basic stereo
Dynamic Noise Reduction System. Although a split power
supply can be used, a single positive supply voltage is
shown, with ac coupled inputs and outputs common in many
consumer applications. This supply voltage can be between
4.5 VDC and 18 VDC but operation at the higher end of the
range (above 8 VDC) is preferred, since this will ensure adequate signal handling capability. The LM1894 is optimized
for a nominal input signal level of 300 mVrms but with an
8 VDC supply it can handle over 2.5 Vrms at full audio bandwidth. Smaller nominal signal levels can be processed but
below 100 mVrms there may not be sufficient gain in the
control path to activate the detector with the source noise.
In this instance, and where battery powered operation is
desired, the LM832 is a better choice. The LM832 has identical operating principles and a similar (but not identical) pinout. It is optimized for input levels around 30 mVrms and a
supply voltage range from 1.5 VDC to 9.0 VDC.
The capacitors connected at Pins 12 and 3 determine the
range of b3 dB cut off frequencies for the audio path filters.
Increasing the capacitor value scales the range downward –
the minimum frequency becomes lower and the maximum
or full bandwidth frequency will decrease proportionally.
Similarly, smaller capacitors will raise the range.
fb3 dB e IT/9.1C (IT e 33 mA MIN)
(1)

THE BASIC DNR APPLICATION CIRCUIT
At the time of writing, the LM1894 has already found use in
a large variety of applications. These include:
AUTOMOTIVE RADIOS
TELEVISION RECEIVERS
HOME MUSIC CENTERS
PORTABLE STEREOS (BOOM BOXES)
SATELLITE RECEIVERS
AUDIO CASSETTE PLAYERS
AVIONIC ENTERTAINMENT SYSTEMS
HI-FI AUDIO ACCESSORIES
BACKGROUND MUSIC SYSTEMS
ETC.

( e 1.05 mA MAX)
For normal audio applications the recommended value of
0.0033 mF should be adhered to, producing a frequency
range from 1 kHz to 35 kHz.

TL/H/8420 – 1

FIGURE 1. Complete DNR Application Circuit
is a trademark of National Semiconductor Corp.

C1995 National Semiconductor Corporation

TL/H/8420

RRD-B30M115/Printed in U. S. A.

AN-390

*R1 a R2 e 1 kX total

DNRTM

DNR Applications of the LM1894

DNR TM Applications
of the LM1894

TL/H/8420 – 18

TL/H/8420–2

(b)

(a)
FIGURE 2. Two Methods of DNR IN/OUT Switching

Apart from the basic circuit shown in Figure 1, all applications of the DNR system have another feature in common –
the location of the LM1894 in the signal chain. As Figure 3
shows, the LM1894 is always placed right after the signal
source pre-amplifier and before any circuit that includes
user adjustable controls for volume or frequency response.
The reasons for this are obvious. If the gain of the signal
amplifier preceding DNR is changed arbitrarily, the noise
input level to the LM1894 will not be at the correct point to
begin activation of the audio path filters. Either reduced
noise reduction will be obtained, or the high frequency content of the program material will be affected. A change in
system gain prior to the LM1894 requires a corresponding
change in the control path threshold sensitivity. Similarly
modifying the frequency response, by heavy boost or cut of
the mid to high frequencies, will have the same effect of
changing the required threshold settingÐapart from modifying the masking qualities of the program material.

The two resistors connected at Pin 5 set the overall control
path gain, and hence the system sensitivity. A lower tap
point will decrease the sensitivity for high signal level sources, and a higher tap point will accommodate lower level
sources. For purposes of initial calibration it is best to replace the resistors with a 1 kX potentiometer (the wiper arm
connecting through C6 to Pin 6), and follow the procedures
outlined below. Once the correct adjustment point has been
found, the position of the wiper arm is measured and an
equivalent pair of resistors are used to replace the potentiometer. This, of course, can be done only if the source has
a relatively fixed noise floorÐthe output from an audio cassette tape for example. For an add-on audio accessory the
potentiometer should be retained as a front panel control to
allow adjustment for individual sources. Use of DNR with
multiple sources is described later.
SYSTEM CALIBRATION
System calibration can be performed in a number of ways.
With the source connected play a blank but biased section
of the cassette tape. Set the potentiometer so that the wiper
arm is at ground and then steadily rotate it until a slight
increase in the output noise level is heard. Alternatively,
with source program material present, set the potentiometer
with the wiper arm connected to the Pin 5 end of the slider
and again rotate until the high frequency content of the program material appears to begin to be attenuated. Then return the potentiometer wiper slightly towards Pin 5 so that
the music is unaffected.
A third method of adjustment can be done with an oscilloscope monitoring the voltage on the control path detector
filter capacitor, Pin 10. This will show a steady dc voltage
around 1V while the wiper arm of the potentiometer is at
ground. As the wiper arm is rotated, this voltage will start to
increase. About 200 mV above the quiescent value will usually be the right point. Note that this will not be a steady dc
voltage but a random peak, low amplitude sawtooth waveform caused by peak detection of the source noise in the
control path.
Whatever method is used to determine the potentiometer
setting, this setting should be confirmed by listening to a
variety of programs and comparing the audio quality while
switching DNR in and out of the circuit. This is easily accomplished by grounding Pin 9 which will disable the control
path and force the audio filters to maximum bandwidth, Figure 2(a) . Also shown is a second method of ON/OFF
switching that gives an increased maximum bandwidth over
that obtained in normal operation. Although the switch is not
a required front panel control it can be an important feature.
Unlike compander systems, DNR can be switched out leaving the source completely unprocessed in any way. With a
switch, the user can always be assured that the noise reduction is not affecting the program material.

HOW MUCH NOISE REDUCTION?
The actual sensitivity setting that is finally used, and the
amount of noise reduction that is obtained, will depend on a
number of factors. As the data sheet for the LM1894 and
other application notes have explained in some detail, the
noise reduction effect is obtained by audio bandwidth restriction with a pair of matched low-pass filters. A
CCIR/ARM* weighted noise measurement is used so that
the measured improvement obtained with DNR correlates
well to the subjective impression of reduced noise. This is
another way of stating that the source noise spectrum level
versus frequency characteristic can have a large impact on
how ‘‘noisy’’ we judge a source to beÐand concomitantly
how much of the ‘‘noisiness’’ can be reduced by decreasing
the audio bandwidth. Fortunately most of the audio noise
sources we deal with are smooth although not necessarily
flat, resembling white noise. The weighting characteristic referred to above generally gives excellent correlation. For
example, if the source b3 dB upper frequency limit is only
2 kHz (an AM radio), reducing the audio path bandwidth
down to 800 Hz will improve the S/N ratio by only 5 to 7 dB.
On the other hand, if the source bandwidth exceeds at least
8 kHz then from 10 dB to 14 dB noise reduction can be
obtained. Of course, it is always worth remembering that
this is the reduction in the source noiseÐany noise added in
circuits after the LM1894 may contribute to the audible output and prevent the full noise reduction effect. To see how
easily this can happen, we will consider the noise levels at
various points in a typical automotive radio using an I/C
tone and volume control, and an I/C power amplifier, both
with and without noise reduction of the cassette player.
*See pp. 2–9 to 2–10, Audio Handbook, National Semiconductor 1980.

2

TL/H/8420 – 3

FIGURE 3. Location of DNR in Audio Systems

TL/H/8420 – 19

(a) Without Noise Reduction

TL/H/8420 – 4

(b) With Noise Reduction
FIGURE 4. Signal and Noise Levels in the Audio Path
3

If we assume that the tape head pre-amplifier gain is such
that the nominal output level (corresponding to O‘‘VU’’) is
300 mVrms, then for a typical cassette tape the noise will be
50 dB lower, or 949 mV. The gain of the tone and volume
control (an LM1036) is unity or 0 dB at maximum volume
setting, with an output noise level of 33 mV with no signal
applied. With the tape pre-amplifier connected, the output
noise from the LM1036 will be Vn where
Vn e 10b6 0(33)2 a (949)2 e 949.6 mV

MODIFICATIONS TO THE STANDARD APPLICATIONS
CIRCUIT
1. TAPE DECKS WITH EQUALIZATION SWITCHES:
Many modern cassette tape decks and automotive radio
cassette players offer at least two types of equalization in
the head-preamplifier in order to optimize the frequency response of various tape formulations. These are often identified on the equalization switch as ‘‘Normal’’ and ‘‘CrO2’’
corresponding to 120 ms and 70 ms time constants in the
equalization network. This difference in time constants can
mean that the noise floor from a cassette tape in the
‘‘CrO2’’ mode can be up to 4 dB lower than for a tape requiring the ‘‘Normal’’ mode, Figure 5 .

(2)

Clearly, the LM1036 has caused an insignificant increase in
the background noise level (0.006 dB). Even when the volume control is set at b20 dB overall gain, the LM1036 intrinsic noise level is 22 mV. The tape noise level is now
94.9 mV ( b20 dB) and the output noise Vn is
Vn e 10b6 0(22)2 a (94.9)2 e 97.4 mV

(3)

Once more an insignificant contribution on the part of the
LM1036 (0.23 dB).
Now we add noise reduction between the tape head amplifier and the LM1036. Usually this will mean over 10 dB reduction in the tape noise so that the input of the LM1036 sees
300 mV noise. At 0 dB gain we have
Vn e 10b6 0(33)2 a (300)2 e 301.8 mV

(4)

TL/H/8420 – 5

FIGURE 5. Tape Playback
Equalization Including Integration
Although a compromise setting can be found for the DNR
threshold setting to accommodate both types of tape, a single pole, double throw switch ganged to the equalization
switch will optimize performance for each mode. In the example given in Figure 6, the resistor values shown are from
an application that yielded a 400 mVrms input to the
LM1894 when the tape flux density was 200 nW/m. For
different tape-head amplifiers the resistors R1 and R2 are
selected using a ‘‘Normal’’ tape as a source, and then R3 is
selected according to the relationship given in Equation (5).

But at b20 dB
Vn e 10b6 0(22)2 a (30)2 e 37.2 mV

(5)

When we compare the results of Equation (3) and (5) we
see that at b20 dB gain setting we are getting only 8.4 dB
noise reduction compared to 10 dB at maximum gain! Since
the volume control is not normally set to maximum, this is a
significant loss.
Active tone and volume controls are not the only circuits
that can contribute to a loss in noise reduction. Most modern automotive radios use I/C power amplifiers delivering in
excess of 6 watts into 4X loadsÐand even more if bridge
amplifiers are employed. With a 12 VDC supply, the output
signal swing is limited to less than 4 Vrms if clipping is
avoided. Typical amplifiers have an input referred noise level of 2 mVrms, and with a gain of 40 dB (a typical value) the
intrinsic output noise level is 200 mVrms, or 86 dB below
clipping. For a normal listening level, the signal amplitude
will be 20 dB below clipping which yields a S/N ratio of only
66 dBÐwhich is just better than the noise reduced input to
the amplifier.
Many manufacturers recommend using I/C power amplifiers
with gains of 60 dB. This will always result in unacceptable
noise performance at moderate listening levels since the
amplifier generated noise is now over 2 mV. For a signal
20 dB below clipping the output S/N ratio is only 46 dB!
It is interesting to note that the inclusion of just 10 dB noise
reduction is sufficient to put pressure on the performance
standards of the remaining circuits in the audio path of an
automotive radio. If more noise reduction is available, such
as a combination of Dolby B and DNR, or Dolby C, then the
subsequent gain distribution must be considered even more
carefully. The power amplifier gain may have to be reduced
to 20 dB to avoid degrading the noise performance. In fact it
may be impractical to realize the full noise performance capability of systems providing high levels of noise reduction
in many automotive stereo radios.

TL/H/8420 – 6

FIGURE 6. Optimizing the Control Path Threshold
for Different Tape Formulations
Notice that only one additional resistor is required over the
standard application, and it is easy to substitute transistor
switching in place of the spdt switch.
R1/(R1 a R2) e 0.63 R3/(R1 a R3)
(5)
2. TAPE DECKS WITH COMPLEMENTARY
NOISE REDUCTION:
Most cassette decks available today employ some form of
complementary (companding) noise reduction system, usually Dolby B Type. DNR can be used in conjunction with
these noise reduction systems as a means to provide yet
more noise reduction on decoded tapes and still provide
4

through modest or better Hi-Fi systems. Although the mono
track width (twice as wide as an audio cassette stereo track)
should help the S/N ratio, the slower tape speed does not,
as shown in the curves of Figure 9. For the SP mode the
S/N ratio is approximately 5 to 10 dB lower than the audio
cassette and worsens by 3 to 5 dB in the extended play
modes. Some ‘‘spurs’’ or ‘‘spikes’’ may be observed at harmonics of the video field frequency (60 Hz) and at the video
line scan frequency of 15.734 kHz. The low frequency
spikes will not affect DNR operation since the control path
sensitivity decreases sharply below 1 kHz, but the presence
of the 15.734 kHz component could cause improper sensitivity settings to be obtained. If this is the case, the pilot
frequency notch filter for FM, described later, can be retuned by changing the capacitor from 0.015 mF to 0.022 mF.

noise reduction for unencoded tapes. The LM1894 is located after the companding system and provision must be
made for the drop in noise level when the compandor is
being used. The DNR threshold sensitivity is increased by
the appropriate amount so that the lower noise levels are
still able to activate the audio filters. For example, the circuit
in Figure 7 shows a switching arrangement to compensate
for the 9 dB lower noise floor from a Dolby B decoded tape.
Notice the change in resistor values R1 through R3 to raise
the sensitivity (yet keeping the sum of R1 and R2 to 1k) and
the 9 dB pad formed by the 3 kX resistor and the 1.5 kX
resistor in parallel with the control path input Pin 6, for use
when the compandor is switched off. Since the output level
from the compandor is usually around 580 mV for a flux
density of 200 nW/m, the ratio of R1 to R2 and R3 is
changed by only 5.6 dB compared to that shown in the previous Figure where the input level was 400 mVrms.

TL/H/8420 – 9

FIGURE 9. Video Tape Noise Spectrum Levels
Figure 9 also shows the noise spectrum with the new Beta
Hi-Fi format. This is clearly superior to both the standard
format and audio cassette tapes and is realized by using the
two video record/play heads simultaneously for audio, thus
taking advantage of the substantially higher relative tape
speed. The audio is added in the form of four FM carriers,
Figure 10. Four carriers are necessary for two audio channels since the azimuth loss between the normal video
heads (reducing crosstalk between the heads at video frequencies) is not enough at the lower audio carrier frequencies. Each head therefore uses different carriers for the left
and right channel signals.

TL/H/8420 – 7

FIGURE 7. Switching with Other NR Systems

TL/H/8420 – 8

FIGURE 8. Video Magnetic Tape Format
3. VIDEO TAPE RECORDERS:
The audio track of a video cassette tape is similar to an
audio cassette and appears along one edge of the tape.
Although provision is made for two tracks, each 0.35 mm
wide, a large number of recordings are monaural with a
track width of 1 mm (0.04 inches).
Unlike the video heads, which are mounted on a rotating
drum and angled to the direction of tape travel in order to
give a much higher recording speed, the audio is recorded
longitudinally with a separate head at 33.35 mm/sec for
standard play, 16.88 mm/sec for long play, and
11.12 mm/sec for the very long play mode (VHS format
tape machines). The noise spectrum is similar to an audio
cassette but with a couple of differences. The typical frequency response from the head pre-amplifier does not extend beyond 10 kHz in the SP mode and is less in the LP
and VLP modes. Even so, this bandwidth is enough to ensure the presence of the familiar tape ‘‘hiss’’ when played

TL/H/8420 – 10

FIGURE 10. Beta Hi-Fi Carrier Frequencies
A quite different technique is used for VHS Hi-Fi, which is
similar to that for 8 mm video. Separate audio heads are
mounted on the same rotating drum that is carrying the video heads, but with a much larger azimuth angle compared
to the video heads. The sound signal is written deep into the
tape coating and then written over by the video signal which
causes partial erasure of the audioÐabout a 10 dB to 15 dB
loss. The difference in azimuth angle prevents crosstalk and
the much greater writing speed still yields an S/N of over
80 dB.
Both Hi-Fi formats provide excellent sound quality with hardly any need for noise reduction but DNR can still play a role.
Conventionally recorded tapes are and will be popular for
quite a while, and even with Hi-Fi recording capability much
5

TL/H/8420 – 11

FIGURE 11. VHS Hi-Fi Recording Format
normal component value tolerances ( g 7% inductance,
g 10% capacitance) the pilot tone will be attenuated by at
least 15 dB.
Handling the signal strength dependence of the FM signal
noise floor is not quite as easy – at least if pre-set DNR
sensitivity settings are used. A look at the quieting curves
for an FM radio will show why. At strong signal levels, greater than 1 mV/meter field strength at the antenna, the IF
amplifier of the radio is in full limiting and the noise floor is
between 60 dB and 80 dB below the audio signal. However,
as the field strength starts to decrease below 1 mV/meter,
the noise level begins to increase, even though the IF amplifier is still in limiting. Worse yet, since the demodulated output includes the noise from the stereo difference signal
channel (L-R), the noise level is increasing more rapidly in
the stereo mode than in the monaural mode. By the time the
field strength has fallen to 100 mV/m the stereo noise is
over 20 dB higher than the equivalent mono noise. If the
DNR sensitivity is pre-set such that noise at the b45 dB to
b 55 dB level is activating the control path detector, when
weaker stations are tuned in the noise level will increase
and less noise reduction will be obtained. On the other
hand, for stronger stations the noise level will drop below
the detector threshold and a possibility exists that high frequency signals will be attenuated. Fortunately this latter occurrence is unlikely with commercial FM broadcasts since
substantial signal compression is common, and the relatively high mid-band signals will be adequate enough to open
the audio bandwidth sufficiently. In any event, with very
strong r.f. signals, the need for noise reduction is minimal
and DNR can be switched out.

recording will be done with television sound as a sourceÐ
and the source noise will dominate now instead of the tape
noise. As discussed later, DNR can be very effective in
dealing with television S/N ratios, allowing much of the benefit of improved recording techniques to be enjoyed.
4. FM RADIOS:
FM sources can present special problems to DNR users.
The presence of the 19 kHz stereo pilot tone can be detected in the DNR control path and cause improper threshold
settings (the problem is not so much that the 19 kHz tone
gives the wrong setting, but that if the threshold is adjusted
with the tone present, then the threshold is wrong when the
tone is absentÐas in a monaural broadcast). Secondly, for
FM broadcasts the noise level at the receiver detector output is dependent on the r.f. field strength when this field
strength is under 100 mV/meter at the antenna terminal.
With a fixed DNR threshold, as the noise level increases
with decreasing field strength, the minimum audio bandwidth becomes wider and a loss in noise reduction is perceived. This latter problem occurs primarily with automobile
radios where the signal strength can vary dramatically as
the radio moves about. For the home receiver, re-adjustment of the DNR threshold setting for an individual station
will compensate for the weaker signals.
To understand how much the pilot tone can affect the DNR
control path, we can take a look at some typical signal levels. For an FM broadcast in the U.S., the maximum carrier
deviation is limited to g 75 kHz with a pilot deviation that is
10% of this value. A high quality FM I/C such as the
LM1865 will produce a 390 mVrms output at the detector
with this peak deviation, so the pilot level at 19 kHz will be
39 mVrms. If the receiver does not include a multiplex filter,
after de-emphasis 4 mV will appear at the inputs to the
LM1894. Typically for FM signal noise floors, the resistive
divider at Pin 5 will attenuate the pilot by 20 dB leaving
0.4 mVrms at Pin 6. This input level to the LM1894 control
path is sufficient to cause the audio bandwidth to increase
by over 1 kHz compared to the monaural minimum bandwidth. Of course, if the receiver does have a multiplex filter,
which is common in high quality equipment or receivers that
include Dolby B Type noise reduction, this problem will not
happen, but otherwise we require an extra 15 dB to 20 dB
attenuation at 19 kHz. This is obtained with a notch filter
tuned to the pilot frequency connected between Pins 8 and
9 of the LM1894. Although a tuned inductor is shown, a
fixed coil of similar inductance and Q can be used since with

L

FM

TV

4.7 mH

4.7 mH

C 0.015 mF 0.022 mF

TL/H/8420 – 12

FIGURE 12. Control Path Notch Filter
6

tude is 7 dB to 10 dB below the picture carrier amplitude
and for cable services the typical sound/picture carrier ratio
is b15 dB. However, due to the FM improvement factor
(45.4 dB for equal amplitude carriers compared to the AM
picture carrier) audio S/N ratios do not degrade as rapidly
as the picture S/NÐeven with the lower audio carrier amplitudes. Figure 14 shows the increase in audio noise level as
both carrier amplitudes are reduced from the picture carrier
level that produces a noise-free picture. When the picture
noise is already objectionable the audio noise level has remained virtually unchanged, even for an audio carrier 30 dB
below the picture carrier. By the time an unacceptable picture noise level has been reached, the audio noise has increased by less than 3 dB for sound carriers at b10 dB and
b 20 dB relative to the picture carrier. Therefore it is unlikely
that a perceptible increase in noise compared to a strong
channel will occur before the viewer switches to another
channel.

Recognizing that a fixed threshold setting is necessarily a
compromise for FM, the designer can still elect to use a preset adjustment for convenience. The set-up procedure is a
little more complicated than for an audio tape source and
involves the use of an FM signal generator. The carrier frequency from the generator (between 88 MHz and 108 MHz)
is unmodulated except for the stereo pilot tone, and the
receiver is tuned to this carrier frequency. Then the carrier
level is increased until the stereo demodulator output S/N
ratio is that desired for the DNR threshold setting. For example, if the recovered audio output is 390 mVrms for
75 kHz deviation of the carrier frequency, the stereo noise
level is 2.2 mVrms for a 45 dB S/N ratio. The generator
level is increased until this noise voltage is measured at the
demodulator output and the resistive divider at Pin 5 of the
LM1894 adjusted correspondingly. A multiplex filter should
be inserted between the decoder output and the S/N meter
to prevent the pilot tone from giving an erroneous reading.
At no time should the pilot tone be switched off since this
will allow the decoder to switch into the nomaural mode,
decreasing the noise level b65 dB instead. A S/N ratio of
45 dB is chosen since many modern receivers incorporate
blending stereo demodulators. As the dashed curve of Figure 13 shows, when the stereo S/N ratio falls to 45 dB, the
decoder starts to blend into monaural operation, thus keeping a constant S/N ratio. The loss in stereo separation that
inevitably accompanies this blending is far less objectionable than abrupt switching from stereo to mono operation at
weak signal levels.

TL/H/8420 – 14

FIGURE 14. Increase in Audio Noise with
Decreasing Carrier Levels

TL/H/8420 – 13

FIGURE 13. FM Radio Quieting Curves
5. TELEVISION RECEIVERS:
At first it might be thought that television broadcast signals,
with an FM sound carrier located 4.5 MHz above the picture
carrier frequency, will present the same difficulties as FM
radio broadcasts to a DNR system with a pre-set threshold.
This conclusion is modified by two considerations. First the
TV receiver is unlikely to be mobile and the received signal
strength will be relatively constant from an individual broadcast station. Secondly another subjective factor, the picture
quality, will largely determine whether the signal strength is
adequate enough for the viewer to stay tuned to that station.
A representative television receiver will have a VHF Noise
Figure between 6 dB and 7 dB such that, with a 75X antenna impedance, the picture will be judged noise-free at an
input signal level of just above 0.5 mVrms – i.e. a picture
signal to noise ratio of 43 dB. Noise will become perceptible
to most viewers at a S/N ratio of 38 dB and become objectionable at 28 dB to 30 dB. Therefore 13 dB below 1 mVrms
the picture noise is objectionable, and at b25 dB to
b 30 dB it will probably be totally unacceptable to the majority of viewers. For off-air broadcasts, the audio carrier ampli-

TL/H/8420 – 15

FIGURE 15. TV Noise Spectrum Level

Figure 15 shows the noise spectrum level of a strong audio
carrier (1 mVrms) referred to 7.5 kHz carrier deviation. The
standard peak deviation in the U.S. is 25 kHz so that the
spectrum level will be 10 dB lower when referred to the
peak audio level, meaning that the noise is not much better
than the cassette tape noise levels shown previously. Only
the relatively small power capability and limited bandwidth
of audio amplifiers and speakers in conventional receivers
has made this noise level acceptable. Unfortunately for the
listener who hooks up the audio to his Hi-Fi system, or buys
a new receiver with wider audio bandwidth and high output
power (in anticipation of the proposed BTSC stereo audio
broadcasts for television), TV sound will exhibit this noise.
7

Examples of both arrangements are shown in Figure 16(a)
and (b). To set up the multiple source system of 16(b) , the
DNR control path sensitivity is adjusted for the source with
the lowest noise floor. Measure the peak detector voltage
(Pin 10) produced by this noise source and then switch to
the next source. Adjust (attenuate) the input level of the new
source to match the previous Pin 10 detector voltage and
repeat this procedure for each subsequent source.

Because the noise floor will be relatively constant, a pre-set
threshold can be used for the LM1894 control path (although broadcast of older movies with unprocessed and
noisy optical soundtracks might increase the received
noise), and the only modification to the standard application
circuit is to shift the control path notch filter down to
15.734 kHz. This is done with sufficient accuracy simply by
changing the 0.015 mF tuning capacitor to 0.022 mF.
Note: The introduction of a stereo audio broadcast (the BTSC-MCS proposal) does not substantially modify the above conclusions, even though
dbx noise processing is used. The dbx-TV noise reduction is applied
only to the new stereo difference signal channel (L-R) to decrease the
additional noise intrinsic in the use of an AM subcarrier along with the
normal (L a R) monaural channel. This means that the new stereo
signal should have roughly the same characteristics as the present
monaural signal.

7. CASCADING THE LM1894 AUDIO FILTERS
The LM1894 has two matched audio lowpass filters which
can be cascaded, providing a single channel filter per I/C
with a 12 dB/octave roll-off. This produces slightly more
noise reduction (up to 18 dB) but because the steeper filter
slope may in some cases produce audible effects on high
frequency material, cascaded filters are best used for sources with a relatively restricted h.f. content. When the filters
are cascaded the combined corner frequency decreases by
64% according to Equation (6), for n e 2
(6)
fc e fo 0100.3/n b 1
Therefore, to retain the original frequency range, the capacitor values must be reduced by the same factor to
0.0022 mF. One of the audio outputs is connected over to
the other audio filter input and the summing amplifier in the
control path is by-passed by moving the 0.1 mF coupling
capacitor from Pin 5 over to the single audio input. If the
audio source is unable to drive the 1 kX impedance of the
control path input network, this can be scaled up by using a
0.01 mF capacitor and a 10 kX potentiometer.

6. MULTIPLE SOURCES:
Multiple sources are best accommodated by keeping the
potentiometer in the LM1894 control path and allowing the
user to optimize each source. Nevertheless, for convenience, pre-sets are often desired and these can be done in
two ways.
1) If the sources have widely different S/N ratios, the resistive divider at Pin 5 should be tapped at the appropriate
point for each source noise level. This assumes that the
source signal levels have been matched at the input to
the LM1894 for equal volume levels.
2) If the source S/N ratios are not too far different, then the
input levels can be trimmed individually to produce the
same noise level in the LM1894 control path. A single
sensitivity setting is used, and an additional switch pole
ganged to the source selector switch is avoided.

(a) Input and Control
Path Switching for
Two Sources

(b) Eliminating the
Control Path Switch

TL/H/8420 – 16

TL/H/8420 – 20

FIGURE 16. Multiple Programme Source Switching

8

TL/H/8420 – 17

FIGURE 17. Cascading the Audio Filters of the LM1894

9

DNR Applications of the LM1894
LIFE SUPPORT POLICY
NATIONAL’S PRODUCTS ARE NOT AUTHORIZED FOR USE AS CRITICAL COMPONENTS IN LIFE SUPPORT
DEVICES OR SYSTEMS WITHOUT THE EXPRESS WRITTEN APPROVAL OF THE PRESIDENT OF NATIONAL
SEMICONDUCTOR CORPORATION. As used herein:

AN-390

1. Life support devices or systems are devices or
systems which, (a) are intended for surgical implant
into the body, or (b) support or sustain life, and whose
failure to perform, when properly used in accordance
with instructions for use provided in the labeling, can
be reasonably expected to result in a significant injury
to the user.
National Semiconductor
Corporation
1111 West Bardin Road
Arlington, TX 76017
Tel: 1(800) 272-9959
Fax: 1(800) 737-7018

2. A critical component is any component of a life
support device or system whose failure to perform can
be reasonably expected to cause the failure of the life
support device or system, or to affect its safety or
effectiveness.

National Semiconductor
Europe
Fax: (a49) 0-180-530 85 86
Email: cnjwge @ tevm2.nsc.com
Deutsch Tel: (a49) 0-180-530 85 85
English Tel: (a49) 0-180-532 78 32
Fran3ais Tel: (a49) 0-180-532 93 58
Italiano Tel: (a49) 0-180-534 16 80

National Semiconductor
Hong Kong Ltd.
13th Floor, Straight Block,
Ocean Centre, 5 Canton Rd.
Tsimshatsui, Kowloon
Hong Kong
Tel: (852) 2737-1600
Fax: (852) 2736-9960

National Semiconductor
Japan Ltd.
Tel: 81-043-299-2309
Fax: 81-043-299-2408

National does not assume any responsibility for use of any circuitry described, no circuit patent licenses are implied and National reserves the right at any time without notice to change said circuitry and specifications.



---

Source Exif Data:

File Type                       : PDF
File Type Extension             : pdf
MIME Type                       : application/pdf
PDF Version                     : 1.3
Linearized                      : Yes
Create Date                     : 1995:10:26 12:22:50
Producer                        : Acrobat Distiller 2.0 for Windows
Title                           : AN390 DNR - Applications Of The LM1894
Subject                         : AN-390
Author                          : 
Keywords                        : Application Notes, Audio Circuits
Modify Date                     : 2001:11:23 11:42:45+05:30
Page Count                      : 10

EXIF Metadata provided by EXIF.tools