
REV. M
Page 5
September 1998
AS3502
Functional Description
Power-On Reset
When power is applied first a power on reset signal is
generated on chip which initializes AS3502:
The on chip programmable AFE registers are set to
their default values (those values are defined in the
register allocation section), the tone control register is
set to the default status and the serial interface is ini-
tialized. AS3502 remains in power down state until a
software start-up command. An active Low signal with
a duration of min. 25 μs on the power on reset pin can
be used to externally reset the device AS3502. For
normal operation this pin must be pulled High.
Power Up Mode
AS3502 is powered up through a one byte start-up
command. The byte written into the Digital Control
Register DC allows to individually enable the transmit
and the receive section. If the transmit channel is en-
abled first, the receive channel may be enabled any
time without any restrictions. On enabling the receive
channel and subsequent enabling of the transmit
channel the PCM strobe signals TXS and RXS have
to be tied together. The configuration information
written into the AC and AG define which analogue
transducer interfaces will be enabled on power up.
The PCM output TXD remains in Tristate until the
second frame synchronization signal after start-up.
Any of the programmable registers may be modified
while AS3502 is in active mode.
Power Down Mode
In power down mode all chip functions except the se-
rial interface are kept inactive. All analogue functions
are powered down and all digital outputs are put into
Tristate mode. In this operating state the internal
registers are normally configured to the desired
values prior to the start-up command. The chip can
be brought into power down mode any time through a
power down command written into the DC Register. In
this case all programmable registers retain their pro-
grammed values.
Analogue Input interface
The AS3502 input interface provides two identical dif-
ferential inputs e.g. for a handset microphone and for
a handsfree microphone. The input sources are se-
lected through the AG register. Clipping of signals
with arbitrary DC offset must be avoided by capacitive
coupling. The input impedance of 2 x 30 k
is
compatible
with
both
microphones. Each input is connected through an
analogue input multiplexer to a low noise high gain
preamplifier. The gain is software programmable
through register AG from +16 to +46 dB in 6 dB steps
with a tolerance of ±0.2 dB. This wide range
electret
and
dynamic
guarantees optimum usage of the A/D converter dy-
namic range with various transducers.
Analogue Output Interface
The AS3502 output interface provides differential
outputs for an earpiece, for a loudspeaker and for a
toneringer. The output stages are selected through
the AC register. The earpiece output driver is a fully
differential amplifier that is capable of driving 3.2Vpp
into a 150
transducer directly and is gain
programmable in three steps from -12 dB to +6 dB
through the AG register. The +6 dB step allows to
drive ceramic earpiece transducers or to boost the re-
ceive amplitude. The loudspeaker driver is a fully
differential power amplifier with a peak output power
of 50 mW into a 50
loudspeaker. This output allows
loudhearing and handsfree operation under software
control.
The tone ringer outputs are digital push/pull outputs
with rail to rail voltage swing that capable of driving
various toneringers. For volume control the output
signal may be either pulse density modulated or pulse
width modulated under software control.
Transmit Section
The scaled analogue input signal enters a 1st order
RC antialiasing filter with a corner frequency of
approx. 40 kHz. This filter eliminates the need for any
off chip filtering as it provides sufficient attenuation at
1.024 MHz to avoid aliasing. From there the bandlim-
ited signal is fed to a 2nd order Sigma Delta modula-
tor with a sampling frequency of 1.024 MHz. A factory
trimmed voltage reference guarantees accurate
absolute transmit gain (0 dBm0 reference level). The
modulator is followed by a digital decimation filter that
transforms the resolution in time to resolution in
amplitude. The decimation filter is followed by a mini-
mum phase 5th order IIR filter implementing the
CCITT lowpass portion of the encoder bandpass fre-
quency characteristics. Finally a 3rd order IIR high
pass filter implements the highpass portion of the en-
coder bandpass frequency characteristics according
to CCITT specifications.
The digitally filtered signal is further fed to a digital
gain setting stage which allows to program the gain
from -38 to +10 dB with a tolerance of better than
±0.05 dB from 0 to +6 dB to compensate for
transducer sensitivity variations. The same stage may
additionally be used for digital volume control for
transmit volume attenuation. This feature may be
used for software based handsfree voice switching
algorithms.
In case of 16 bit linear mode the voice band signals
are converted to a PCM two's complement 12 data bit
plus sign bit format with a sample rate of 8 kHz and