INTEGRATEDHIGH-PASSFILTER
NOISE RECTIFIER OUTPUT FOR QUALITY
DETECTION
PROGRAMMABLETRIGGERTHRESHOLD
DEVIATION AND FIELD STRENGTH DE-
PENDENT TRIGGERADJUSTMENT
PAUSE DETECTOR:
PROGRAMMABLETHRESHOLD
ALLFUNCTIONS PROGRAMMABLEVIA I
DESCRIPTION
The TDA7340G I
processor contains all signal processing blocks of
2
C bus controlled audio signal
2
CBUS
TDA7340G
AUDIO SIGNAL PROCESSOR
PQFP44
ORDERING NUMBER: TDA7340G
a high performance car radio, including audioprocessor, stereodecoder, noise blanker, pause
detectorand differentmute functions.
The use of BICMOS technology allows the implementation of several filter functions with switched
capacitor techniques like fully integrated, adjustment free PLL Loop filter, pilot detector with integratorand pilotcancellation.
This minimizes the number of external components.
Due to a highly linear signal processing, using
CMOS-switching techniques instead of standard
bipolar multipliers, very low distortion and very
low noise are obtained also in the stereodecoder
part. The audioprocessor contains several new
features like softmute, zero-crossing mute and
pausedetector.
Very low DC stepping is obtained by use of a
BICMOStechnology.
September 1999
1/27
TDA7340G
AUDIO PROCESSORPART
FEATURES:
Input Multiplexer:
DIFFERENTIAL CD STEREOINPUT
CASSETTESTEREOINPUT
FM STEREO INPUT FROM STEREODE-
CODER
AM INPUT:
MONO OR STEREO MODE (PROGRAMMABLE)
BEEP INPUT(ONLY IN AM MONO MODE)
TELEPHONEDIFFERENTIAL MONOINPUT
GAIN PROGRAMMABLE IN 3 x 3.75dB
STEPS
Loudness:
FULLY PROGRAMMABLE
15 x 1.25dB STEPS
Volume Control:
1.25dBCOARSE ATTENUATOR
0.31dBFINE ATTENUATORS
MAX GAIN 20dB
MAX ATTENUATION 59.7dB (PLUS LOUD-
measuredwith : 81% mono signal; 9% pilot signal;fm=1KHz;
10% SCA- subcarrier(f
= 67KHz, unmodulated)
s
4) ACI(ADJACENT CHANNEL INTERFERENCE)
α114 =
α
190
(
V
signal)(at1KHz
O
(spurious)(at4KHZ)
V
O
(signal)(at1KHz)
V
O
=
(
V
spurious)(at4KHZ
O
)
;f
= 110KHz- (3 x 38KHz)
s
= 186KHz- (5 x 38KHz)
;f
s
)
measuredwith : 90% mono signal; 9% pilot signal;fm=1KHz;1% spurioussignal
(f
= 110KHzor 186KHz, unmodulated)
s
TDA7340G
5) Controlrangetyp 11%of V
6) Controlrangetyp 30%of V
7) Allthresholdsare measuredby usinga pulsewith T
R (see figure 2)
R (see figure 1)
=2µs, T
R
=2µs and TF=10µs.
HIGH
The repetitionrate must not increasethe PEAKvoltage.
8) NBTrepresentthe STDEC bit pairD
NAT representthe SPKR_LFbit pair D
9) OVDrepresent the SPKR_LR bit pair D
10) FSCrepresent the SPKR_RFbit pair D
for the noiseblanker trigger threshold
6,D5
for the noise controlledtrigger threshold
7,D5
forthe over deviationdetector
7,D6
for thefield strength control
7,D6
11) The TDA7340G has a dedicatedinternalcircuitryproviding a softpower-on. The I2C bus data
programmationmust start after the reference DC level has reached the targetVs/2 value,
otherwisea pop can be generated.The Crefpin and Out pins rise time at poweron are riported
in Figg.4,5, 6 for Cref valuesof 4.7uF,10uF, 22uF.
12) The CDL- and CDR- can be shortcircuitedin applications providing3 wiresCD signal.
L+
L+
L-∼R-
CDTDA7340G
=
R+
L-
R-
R+
D95AU352
13)The AGND and DGND layout wires must bekept separated. A 50Ω resistor is recommend to be put
as far as possible from the device.
9/27
TDA7340G
Figure 1: High Cut Control
Figure 3
V
MPX
V
TH
Figure2: StereoBlend
SEP
(dB)
50
VR=3.6V
40
30
20
10
0
-0.4-0.3-0.2-0.1VSB-VR(V)
D94AU056
DC-LEVEL
D94AU185
2
C BUS INTERFACE PROTOCOL
I
TRT
HIGH
T
F
Time
The interfaceprotocol comprises:
A start condition(s)
A chip address byte, (theLSB bit determines
read/writetransmission).
CHIP ADDRESS
MSBLSBMSBLSBMSBLSB
S1000100R/W ACKIACKDATAACKP
D95AU216
SUBADDRESSDATA 1 ... DATA n
XXTA3A2A1A
A subaddressbyte
A sequenceof data (N-bytes+ acknowledge)
A stop condition (P)
0
ACK = Acknowledge
S = Start
P = Stop
I = Autoincrement
MAX CLOCK SPEED 500kbits/s
Autoincrement
If bit I in the subaddressbyte is setto ”1”, the autoincrementof subaddressis enabled.
For example to select -17.5dBloudnessthe Data Byte is:XXX01110
Note (1):
If the loudness is switched OFF, the loudness stage is acting like a volume attenuator with flat frequency response. D0 to D3 determine the
attenuation level
LOUDNESS
MUTE
MSBLSB
D7D6D5D4D3D2D1D0
1Soft Mute On
01Soft Mute with fast slope (I = I
11Soft Mute with slow slope (I = I
1Direct Mute
01Zero CrossingMute ON
00ZC Mute OFF(delayed until next zero crossing)
1Zero Crossing Mute and Pause Detector Reset(*)
DESCRIPTIONOF THE NOISE BLANKER
In the normal automotive environment the MPX
signal is disturbed by ignition spikes, motors and
high frequency switchesetc.
The aim of the noise blanker part is to cancelthe
influence of the spikes produced by these components. Therefore the output of the stereodecoder
is switched off for a time of 40µs (average spike
duration).
In a first stagethe spikes must be detectedbut to
avoid a wrong triggering on high frequency noise
a complex trigger controlis implemented.
Behind the trigger stage a pulseformer generates
the 40µs ”blanking”pulse.
In the following section all of thesecircuits are described in their function and their programming,
too (see fig.4).
TDA7340G
1.1 Trigger Path
The incoming MPX signal is highpass-filtered,
amplifiedand rectified(block RECT-PEAK).
The second orderhighpass-filterhas a corner-frequency of 140KHz.
The rectifier signal, RECT, is used to generate by
peak-rectification a signal called PEAK, which is
available at the PEAK pin.
Also noise with a frequency >100KHz increases
the PEAK voltage.The value of the PEAK voltage
influences the trigger threshold voltage Vth (block
ATC).
Both signals, RECT and PEAK+Vth are fed to a
comparator (block PEAK-COMP) which outputs a
sawtooth-shapedwaveform at theTBLANK pin. A
second comparator (block BLANK-COMP) forms
the internal blanking durationof 40µs.
The noise blanker is supplied by his own biasing
circuit (block BIAS-MONO).
1.2 Automatic NoiseControlled Threshold
Control (ATC)
The are two independent possibilities for programmingthe triggerthreshold:
a)the minimum threshold in 8 steps(bits D6, D5
of the STD-byte and bit D5 of the SPKR_LF
byte)
b)and the noise adjusted threshold in 4 steps
(bits D6, D5of the SPKR_LFbyte, (see fig.5)
The minimum threshold is used in combination
with a goodMPX signalwithout anynoise.
The sensitivityin this operation is high, depending
only onthe programmed ”minimum trigger threshold”, bits NTB of thenoise blanker byte 1.
It is independentof the PEAK voltage.
If the MPX signal is noisy (low fieldstrength) the
PEAK signal increases due to the higher noise,
which is also rectified(see part1.1).
With increasing of the PEAK voltage the trigger
threshold voltage increases, too. This particular
gain is programmablein 4 steps(see fig.2).
1.3 Automatic ThresholdControl by the
Stereoblendvoltage (ATC-SB)
Besides the noisecontrolled threshold adjustment
there is an additional possibilityfor influencingthe
trigger. It is controlled by the difference between
Vsb and Vr, similar to the Stereoblend. The reason for implementing such a second control will
be explainedin the following:
The pointwhere the MPXsignal starts to become
noisy is fixed by the RF part. Therefore also the
starting point of the normal noise controlled trigger adjustmentis fixed (fig.6).
But in some cases the behaviour of the noiseblanker can be improved by increasing the
threshold even in a region of higher fieldstrength,
for the MPX signal often shows distortion in this
range.
Because of the overlap of this range and the
range of the stereo/mono transition it can be controlledby Vsband Vr.
This threshold increase is programmable in 3
steps or switchedoff (see fig.6).
1.4 OverDeviation Detector (MPX-RECT)
Sometimes when listening to stations with a
higher deviation than 75KHz the noiseblanker
triggerson thehigh frequencymodulation.
To avoid this blanking, which causes noise in the
output signal, the noiseblankeroffers a deviationdependentthreshold adjustment.
By rectifying the MPX signal a further signal representing the actual deviation is obtained. It is
used to increasethe PEAKvoltage.
Offset and gain of this circuit are programmablein
3 steps (the first step turns off the detector, see
fig.7).
1.5 BlendMode
Another possibility to avoid a disturbingtriggering
on modulation is to use the spikes on the fieldstrengthsignal (LEVELpin).
But in the range of higher fieldstrength the signal
saturates and no more spike detection is possible. For this reason the TDA7340G offers the
”BLEND MODE”. When ”BLEND MODE” is activated a smooth transition between the LEVELand the MPX-signal is used to detect the spikes
eitheron LEVELor on MPX.
In the lower fieldstrength range mainly the
LEVEL-signal is used whereas in the higher
range mainly the MPX is used. This switching is
controlled also by the normal Stereoblend signal
to avoid additionalpins.
With ”BLEND MODE OFF” both signals are used
to detect spikesin thewhole fieldstrength range.
1.6 Input Mode
The NB of TDA7340G offers two input modes.
The first one uses the internal trigger path and
optional the LEVEL input. But the TDA7340G offers also an external trigger mode.
During this mode the internal MPX trigger path is
disabled whereas the high pass at the LEVELpin
is bypassed.
By using an external highpass at the LEVEL-pin
one can adjust the NB’s behaviour to the desired
one.
17/27
TDA7340G
Figure 4: Block Diagram of the Noise Blanker
MPX IN
LEVEL
V
R
V
SB
80KHz LP
I2C-BUS
140KHz HP
120KHz HP
D95AU330
SIGNAL PATH
AMP
2
ADDITIONAL THRESHOLD
(ATC-SB, MPX_RECT)
RECT-PEAK
4
CONTROL
R
PEAK
82KΩ
BUF
RECT
PEAK
AUTOMATIC
THRESHOLD
CONTROL
C
PEAK
47nF
ATC
5
PEAK COMP
-
+
PEAK+VTH
C
BLANK
330pF
BLANK COMP
REF.
RIGHT
+
LEFT
to OUTPUTS
40µs
Figure 5: Trigger Thresholdvs. V
8 STEPS
peak
VTH
65mV
30mV
MIN. TRIG. THRESHOLD
0.9V
D95AU331
NOISE
ADJUSTED
TRIG. THRESHOLD
1.5V
V
PEAK(V)
260mV(01)
220mV(10)
180mV(11)
140mV(00)
18/27
Figure 6: Behaviourof the Field StrengthControlledThreshold Adjustment
V
PEAK
MONOSTEREO
≈3V
TDA7340G
TRIG. THRESHOLD
NOISE
noisy signalgood signal
D95AU333A
ATC_SB OFF (11)
2.2V(00)
1.8V(01)
1.2V(10)
0.9V
E’
Figure 7: Behaviourof the DeviationDependent ThresholdAdiust (Over DeviationDetector)
V
PEAK
(V)
=00
OVM
3.2
2.3
1.3
0.9
DETECTOR OFF (11)
OVM
OVM
=01
=10
D95AU332A
20
32.54575
MUTE & PAUSEFEATURES
The TDA7340G provides three types of mute,
controlled via I2C bus (see pag.12, MUTE BYTE
register).
SOFT MUTE
Bit D0=1
Bit D0=0
→ Soft Mute ON
→Soft Mute OFF
It allows an automatic soft muting and unmuting
of thesignal.
The time constant is fixed by an external capacitor Csm inserted between pin Csm andground.
Once fixed the external capacitor, two different
slopes (time constant) are selectable by programmationof bit D1.
DEVIATION(KHz)
Bit D1=1
Bit D1=0
→ fast slope (I=Imax)
→
slow slope (I=Imin)
The soft mute generates a gradual signal decreasing avoiding big click noise of an immediate
high attenuation, without necessity to program a
sequence of decreasing volume levels. A response example is reported in Fig.12 (mute) and
Fig.13 (unmute). The final attenuation obtained
with soft muteON is60dB typical.
The used reference parameter is the delay time
taken to reach 20dB attenuation (no matter what
the signal levelis).
In application, the soft mute ON programmation
should be followed by programmation of DIRECT
MUTE ON (see later) in order to achieve a final
100dB attenuation.
Beside the I2C bus programmation,the Soft Mute
ON can be generated in a fast way by forcing a
LOW level at pin phone GND, controlled by the
µP through a transistor. This approach is recommended for fast RDS AF switching.
The Soft Mute status can be detected via I2C
bus, reading the Transmitted Byte, bit SM (see
data sheet pag.11).
read bit SM = 1 soft mute status ON
read bit SM = 0 soft mute status OFF
DIRECT MUTE
bit D3 = 1 Direct mute ON
bit D3 = 0 Directnute OFF
The direct mute bit forces an internal immediate
signal connectionto ground.
It is located just before the Volume/Loudness
stage, and gives a typical100dB attenuation.
SPEAKERSMUTE
An additional direct mute function is included in
the speakersattenuatorsstage.
The four output LF, RF, LR, RR can be separately
muted by setting the speaker attenuator byte to
the value 11111111 binary.
Typical attenuationlevel 100dB. This mute is useful for fader and balance functions. It should not
be applied for system mute/unmute, because it
can generate noise due to the offset of previous
stages (bass / treble).
ZEROCROSSINGMUTE
bit D2=1 D4=0 zero crossingmute ON
bit D2=0 D4=0 zero crossingmute OFF
The mute activation/deactivation is delayed until
the signal waveform crosses the DC zero level
(Vref level).
The detection works separately for the left and
the right channels (see Figg. 14, 15). Four different windows threshold are software selectable by
two dedicated bits.
bit D6 bit D5WINDOW
00Vref DC +/-160mV
01Vref DC +/-80mV
10Vref DC +/-40mV
11Vref DC +/-20mV
The zero crossing mute activation/deactivation
starts when the AC signal level falls inside the selected window (internal comparator).
The ZEROCROSS Mute (and Pause) detector is
always active. It can be disabled, if the feature is
not used, by forcing the bit D4=1 Zero crossing
and Pause detector reset.
In this way the internal comparator logic is
stopped,eliminating its switching noise.
The zero cross mute status is detected reading
the TransmittedByte bit ZM.
bit ZM = 1 zerocross mute status ON
bit ZM = 0 zerocross mute status OFF
PAUSEFUNCTION
On chip is implementeda pausedetectorblock.
It uses the same 4 windows threshold selectable
for the zero crossing mute, bit D6,D5 byte MUTE
(see above). The detector can be put in OFF by
forcingbit D4=1,otherwise it is active.
The Pause detector info is available at PAUSE
pin. A capacitor must be connected between
PAUSEpin and Ground.
When the incoming signal is detected to be outside the selected window, the external capacitor
is discharged. When the signal is inside the window, the capacitor is integrating up (see Figg.16
and 17).
The external capacitor value fixes the time constant.
The pull up current is 25uV typical
Withinput signal
Vin = 1Vrm --; Vdc pin pause= 15mV
Vin = 0Vrms--; Vdc pin pause= 5.62V
For example choosing Cpause = 100nF the
charge up constant is about 22ms. Instead with
Cpause = 15nF the charge up constant is about
360us.
The Pause detection is useful in applications like
RDS, to performnoiseless tuning frequeny jumps
avoidingto mute the signal.
NO SYMMETRICAL BASS CUT RESPONSE
bit D7=0 No symmetrical
bit D7=1 Symmetrical
The Bass stage has the option to generate an
unsymmetrical response, for cut mode settings
(basslevel from-2db to - 14dB)
For example using a T-type band pass external
20/27
TDA7340G
filter, the bass cut response becomes a low pass
filter, while the response in bass boost condition
is unchanged.
The feature is useful for human ear equalization
in noisy enviromentslike cars etc.
See examples in Fig. 18 (symmetrical response)
and Fig. 19 (unsymmetricalresponse).
TRANSMITTED DATA (SEND MODE)
bit P=0Pause active
bit P=1No pausedetected
bit ZM = 1Zero cross mute ON
bit ZM = 0Zero cross mute OFF
bit SM = 1 Soft mute ON
bit SM = 0 Soft mute OFF
bit ST = 1Stereosignaldetected(inputMPX)
bit ST = 0Mono signal detected (inputMPX)
The TDA7340G allows the reading of four info
bits.
The type (Stereo/Mono)of received broadcasting
signal is easily checked and displayed by using
theST bit.
The P bit check is useful in tuning jumps without
signal muting.
The SM soft mute status becomesactive immedi-
ately, when bit D0 is set to 1 (soft mute ON,
MUTE byte) and not when the signal level has
reached the 60 dB final attenuation.
TDA7340G I
The protocol is standard I
2
C BUS PROTOCOL
2
C, using subaddress
byte plus data bytes (see pagg.11 to 16).
The optional Autoincrement mode allows to re-
fresh all the bytesregisters with transmissionof a
single subaddress, reducing drastically the total
transmissiontime.
Without autoincrement, subaddress bit I=0,to
refresh allthe bytes registers (10), it is necessary
to transmit10 times the chip address, the subaddress and the data byte.
Working with a 100Kb/sclock speed the total time
would be :
[(9*3+2)*10]bits*10us=2.9ms
Instead using autoincrement mode, subaddress
bit I=1, the total time will be:
(9*12+2)*10us=1.1ms.
The autoincrementmode is useful also to refresh
partially the data. For example to refresh the 4
speakers attenuatorsit is possibleto programthe
subaddress Spkr LF (code XX010100), followed
by the data byte of SPKR LF, LR, RF, RR in sequence.
Note:
that the autoincrement mode has a module 16
counter, whereas the total used register bytes are
10.
It is not correct to refresh all the 10 bytes starting
from a subaddress differentthan XX010000.
For example using subaddress XX010010 (vol-
ume) the registersfrom Volume to Stereodecoder (see pag.11) are correctly updated but the
next two transmitted bytes instead to refer to the
wanted Input selector and Loudness are discharged. (the solution in this case is to send two
separated pattern in autoinc mode, the first composed by address, subaddressXX010010, 8 data
bytes, and the second composed by address,
subaddressXX010000,2 data bytes).
With autoincrement disabled, the protocol allows
the transmission in sequenceof N data bytes of a
specific register, without necessity to resend each
time the addressand subaddressbytes.
This feature can be implemented, for example, if
a gradual Volume change has to be performed (
the MCU has not to send the STOP condition,
keeping activethe TDA7340G communication).
WARNING
The TDA7340G always needs to receive a STOP
condition, before beginning a new START condition. The device doesn’t recognize a START condition if a previously active communication was
not ended by a STOP condition.
2
C BUS READ MODE
I
The TDA7340G gives to the master a 1 byte
”TRANSMITTED INFO” via I2C bus in read
mode. The read mode is Master activated by
sending the chip address with LSB set to 1, followedby acknowledgebit.
The TDA7340G recognizes the request. At the
following master generated clocks bits, the
TDA7340Gissues the TRANSMITTED INFO
byte on the SDA data bus line (MSB transmitted
first).
At the nineth clock bit the MCU master can:
- acknowledge the reception, starting in this
way the transmission of another byte from
the TDA7340G.
- no acknowledge, stopping the read mode
communication.
LOUDNESS STAGE
The previousSTMicroelectronicsaudioprocessors
were implementing a fixed loudness response,
only ON/OFFsw programmable.
21/27
TDA7340G
No possibility to change the loud boost rate at a
certain volumelevel.
The TDA7340Gimplements a fullyprogrammable
loudnesscontrol in 15steps of1.25dB.
It allows a customized loudness response for
each application.
The external network connected to the loudness
pins LOUD_Land LOUD_Rfixes the type ofloudness response
1) Simple Capacitor
The loudness effect is only a boost of low frequencies.(see Fig.20)
2)Second order Loudness (boost of low and
high frequencies).
3)Second order decreased type Loudness
(lowerboost of lowand highfrequencies).
4)Second order modified type Loudness(higher
boost of low and high frequencies).
BASS FILTER
Severalbass filter types can be implemented.
Normallyit is usedthe basicT-typeBandpassFilter.
Starting from the filter component values (R1 internal and R2, C1, C2 external), the centre frequency Fc, the gain Av at max bass boost and
the filter Q factor are computed as follows:
TREBLESTAGE
The Treble stage is a simplehigh pass filter which
time constant is fixed by internalresistor (50Kohm
typ) and an external capacitorconnected between
pins TREB_R/TREB_Land Ground.
IN-OUT PINS
The multiplexer output is available atOUT_R and
OUT_L pins for optional connection of external
graphic equalizer(TDA7316/TDA7317), surround
chip (TDA7346)etc.
The signal is fed in again at pinsIN_L and IN-R.
In case of applicationwithout external devices the
pins OUT_L/OUT_R and IN_L/IN_R cannot be
short circuited, but must be decoupledvia capacitor, necessaryto avoid signal DC jumps, generating ”Clicking”output noise.
The input impedance of the next volume stage is
35Kohm typical (minimum 24Kohm). A capacitor
no lower than 1uF shouldbe used.
INPUT SELECTOR
The multiplexer selector can choose one of the
followinginputs:
- a differentialCD stereo input.
- an FM stereo input coming from the on chipstereo decoder.
- a Cassettestereo input.
- a TelephoneDifferentialmono input.
- an AM stereo input or alternatively (sw programmable)an AM mono + BEEPmono.
The signal fed to the input pins must be decoupled via series capacitors. The minimum allowed
value depends on the correspondentinput impedance.
For the CD diff input (Zi=10Kohm worst case ) a
Cin=4.7uFis recommended.
For the other inputs (70Kohm worst case, except
PHONE 14Kohm worst case but speech audio
band) a Cin=1uFis recommended.
Information furnished is believed to be accurate and reliable. However, STMicroelectronics assumes no responsibility for the consequences
of use of such information nor for any infringement of patents or other rights of third parties which may result from its use. No license is
granted by implication or otherwise under any patent or patent rights of STMicroelectronics. Specification mentioned in this publication are
subject to change without notice. This publicationsupersedes and replaces all information previously supplied. STMicroelectronics products
are not authorized for use as critical components in life support devices or systems without express written approval of STMicroelectronics.
The ST logo is a registered trademark of STMicroelectronics
1999 STMicroelectronics – Printed in Italy – All Rights Reserved
STMicroelectronics GROUP OF COMPANIES
Australia - Brazil - China - Finland- France - Germany - Hong Kong - India - Italy - Japan - Malaysia- Malta - Morocco -
Singapore - Spain - Sweden - Switzerland - United Kingdom - U.S.A.
http://www.st.com
27/27
Loading...
+ hidden pages
You need points to download manuals.
1 point = 1 manual.
You can buy points or you can get point for every manual you upload.