SGS Thomson Microelectronics TEA7540FP Datasheet



NOISE/SPEECH DISCRIMINATION IN EMIS­SION ANDRECEPTION

INTEGRATED SIGNAL GAIN COMPRESSOR
IN BOTHMODES

PROGRAMMABLE ATTENUATORS IN BOTH
MODES

ADAPTED TO ACOUSTIC PARAMETERS OF
ALL CABINETS

LOW OPERATINGVOLTAGE2.5V
LOW OPERATINGCURRENT 2.1mA
CHIP SELECT BETWEEN HANDSFREE AND

MONITORINGMODES

TEA7540

HANDSFREECONTROLLER

SO28

DESCRIPTION

This 28 pins IC is an innovativeapproach to quality
handsfreetelephonesets.Itresultsfromanextensive
researchonspeechsignal.

PIN CONNECTION

(topview)

HYST1 ATR
HYST2

MICOUT

GND

CS

OUTE

ATTE

TIM

CCE

CE3
CE2
CE1

INE

VREF

1
2
3
4
5
6
7
8
9
10
11
12
13
14

ORDERING NUMBERS:

28
27
26
25
24

22
21
20
19
18
17
16
15

V+
IREF
OUTR
ATTR

VOL23
AMP SUP
AMP IN
CCR
CR3
CR2
CR1
RECOUT
INR

TEA7540FP

D93TL005

July 1998

T is advanced informationon a new product now indevelopment or undergoing evaluation. Details are subject to change without

1/12

TEA7540

BLOCK DIAGRAM

2/12

PIN FUNCTION

N° Name Function

1 GND Ground
2 CS Chip Select
3 OUTE Transmit Attenuator Output
4 ATTE Transmit Attenuator Input
5 HYST1 Transmit Channel Hysteresis
6 HYST2 Receive Channel Hysteresis
7 TIM RC Timer
8 CCE Time Constant of the Transmit Signal Compressor

9 CE3 Transmit Background Noise Memorization Output
10 CE2 Transmit Peak Detector Output
11 CE1 Transmit Rectifier Input
12 MICOUT Transmit Signal Compressor Output
13 INE Transmit Signal Compressor Input
14 V

15 INR Receive Signal Compressor Input

16 RECOUT Receive Signal Compressor Output
17 CR1 Receive Rectifier Input
18 CR2 Receive Peak Detector Output
19 CR3 Receive Background Noise Memorization Output
20 CCR Time Constant of the Receive Signal Compressor
21 AMP IN Handset Preamplifier Input
22 AMP SUP Handset Preamplifier Power Supply
23 VOL Volume Control
24 ATR Attenuation Value
25 ATTR Receive Attenuator Input
26 OUTR Receive Attenuator Output
27 I
28 V+

ref

ref

V+/2 - Reference Voltage

Reference Current Source

TEA7540

FUNCTIONAL DESCRIPTION

SWITCHEDATTENUATORS
Fig.A represents a block diagram of a handsfree
subset with attenuators in signal mode. To pre­vent the system from howling, the total loop gain,
including acoustic feedback through the housing
and sidetone coupling, must be less than 0dB.
For this purpose, two switchedattenuatorsare in­serted in each mode (emission and reception).
The attenuation is shifted from one mode to the
other, resulting from the speech level comparison
between each way.
To prevent the circuit to switch continuously in
one way, the operation of the IC must be fully
symetrical in both ways. This involves signal com­parison, attenuationvalue.

GAIN COMPRESSOR
In TEA7540, two signal compressors are inserted
in each mode before the signal comparison, so

the signal coming from each end has the same
level (100mV peak), the losses in each way (for
instance losses resulting from the line length in
receiving mode) are compensatedand the signal
comparison is fully symetrical. The time constant
of each signal compressor decreases 80 times
more quickly than it increases to prevent from
noise increasing between words. The compress­ing depth is 38dB.

BACKGROUND NOISEDISCRIMINATION
An additional feature provided in TEA7540 is
background noise level discrimination in each
way. The IC stores the background sound level
with a long time constant (3 to 5 seconds depend­ing on an external RC) and compares it with the
incoming signal in order to distinguish a useful
signal (speech) from the background noise. This
background noise memorization is also used to
compensatethe noise in each mode before signal
comparison: the noise level in each mode is sub-

3/12

TEA7540

stracted from the incoming signal before the com­parison. So very high noise level in one mode
cannot troublethe comparisonbetween the useful
signals.
The result of the comparisonmanages the attenu­ators in the following way:

- The maximum attenuation is switched on
the mode where the speech signal is the
lowest. The maximum attenuation is fixed
by two external resistor (maximum 52dB).

Figure A

The time constant of the switch is fixed by
the timer via an externalcapacitor.

- When neither party is talking both attenu­ators are set to a medium attenuation.
Thus each mode is in idle mode. The time
constant of the switch from active mode to
idle mode must be long enough to prevent
from switching to idle mode between two
words (see fig.B). This time constant is
fixed by an externalRC.

Figure B

SPEECH/NOISE

4/12

Em Rec

A max dB

1

0

1

0

0dB

IDLE (A max/2)

D93TL009A