
2000 Mar 21
16
Philips Semiconductors
Preliminary specification
Speech and handsfree IC
TEA1098A
Decision logic (pins IDT and SWT)
handbook, full pagewidth
FCA153
13 mV
13 mV
TENV
TNOI
RENV
from logic
from dynamic
limiter
RNOI
Vdt
X
X
1
1
10
μ
A
+
10
μ
A
+
10
μ
A
X
1
0
X
1
X
0
X
X
X
1
0
0
0
0
0
X
0
Vref
RIDT
CSWT
SWT
(25) 28
(24) 27
8 (3)
7 (2)
12 (7)
10 (5)
IDT
DUPLEX CONTROLLER
LOGIC
(1)
ATTENUATOR
Fig.12 Decision logic.
(1) When DLC < 0.2 V,
10
μ
A is forced.
The TEA1098A selects its mode of operation (transmit,
receive or Idle mode) by comparing the signal and the
noise envelopes of both channels. This is executed by the
decision logic. The resulting voltage on pin SWT is the
input for the voice-switch.
To facilitate the distinction between signal and noise, the
signal is considered as speech when its envelope is more
than 4.3 dB above the noise envelope. At room
temperature, this is equal to a voltage difference
V
ENV
V
NOI
= 13 mV. This so-called speech/noise
threshold is implemented in both channels.
The signal on pin TXI contains both speech and the signal
from the loudspeaker (acoustic coupling). When receiving,
the contribution from the loudspeaker overrules the
speech.
As a result, the signal envelope on pin TENV is formed
mainly by the loudspeaker signal. To correct this, an
attenuator is connected between pin TENV and the
TENV/RENV comparator. Its attenuation equals that
applied to the microphone amplifier.
When a dial tone is present on the line, without monitoring,
the tone would be recognized as noise because it is a
signal with a constant amplitude. This would cause the
TEA1098A to go into the Idle mode and the user of the set
would hear the dial tone fade away. To prevent this, a dial
tone detector is incorporated which, in standard
applications, does not consider input signals between
HFRX and GND as noise when they have a level greater
than 25 mV (RMS). This level is proportional to R
RSEN
.