
AD6652
AUTOMATIC GAIN CONTROL
The AD6652 is equipped with two independent automatic
control (AGC) loops for direct interface with a Rake receive
Each AGC circuit has 96 dB of range. It is important that the
decimating filters of the AD6652 preceding the A
undesired signals, so that each AGC loop is operating on o
the carrier of interest and carriers at other frequencies do not
affect the ranging of the loop.
Rev. 0 | Page 41 of 76
gain
r.
GC reject
nly
f
he
he AGC strives to maintain a constant mean
signal of interest. T
output power despite input signal fluctuations. This permits
operation in environments where the dynamic range of the
signal exceeds the dynamic range of the output resolution.
multiplier.
:
uncation of bits below the output range. Overflow is caused by
pping errors when the output signal exceeds the output range.
Modulation error occurs when the output gain varies during the
reception of data.
Set the desired signal level based on the probability-density
function of the signal, so that the errors due to underflow and
overflow are balanced. Set the gain and damping values of the
loop filter so that the AGC is fast enough to track long-term
amplitude variations of the signal that might cause excessive
underflow or overflow, but slow enough to avoid excessive loss
of amplitude information due to the modulation of the signal.
AGC LOOP
The AGC loop is implemented using a log-linear architecture. It
performs four basic operations: power calculation, error calcu-
lation, loop filtering, and gain multiplication. The AGC can be
configured to operate in one of the following modes:
Desired signal level mode
Desired clipping level mode as set by Bit 4 of AGC control
word (0x0A, 0x12)
The AGC adjusts the gain of the incoming data according to
how far its level is from the desired signal level or desired
clipping level, depending on the mode of operation selected.
n
h
Two datapaths to the AGC loop are provided: one before the
clipping circuitry and one after the clipping circuitry, as show
in Figure 51. For desired signal level mode, only the I/Q pat
before the clipping is used. For desired clipping level mode, the
difference of the I/Q signals before and after the clipping
circuitry is used.
The AGC compresses the 23-bit complex output from the
interpolating half-band filter into a programmable word size o
4 to 8, 10, 12, or 16 bits. Because the small signals from the
lower bits are pushed into higher bits by adding gain, the
clipping of the lower bits does not compromise the SNR of t
CLIP
I
23 BITS
Q
CLIP
MEAN SQUARE (I + jQ)
GAIN
MULTIPLIER
I
PROGRAMM
BIT WIDTH
Q
USED ONLY FOR
DESIRED
CLIPPING LEVEL
MODE
–
–
2
x
The AGCs and the interpolation filters need not be linked
together. Either can be selected without the other. The AGC
section can be bypassed, if desired, by setting Bit 0 of the AGC
control word. When bypassed, the I/Q data is still clipped to a
desired number of bits, and a constant gain can be provided
through the AGC gain
AVERAGE 1 – 16384 SAMPLES
DECIMATE 1 – 4096 SAMPLES
SQUARE ROOT
K
×
z
–1
1 – (1 + P)
×
z
–1
+ P
×
z
–2
ERROR
K GAIN
P POLE
+
–
R DESIRED
ABLE
LOG
2
(X)
0
DESIRED SIGNAL LEVEL MODE
In this mode of operation, the AGC strives to maintain the
output signal at a programmable set level. This mode of opera-
tion is selected by writing AGC control word (0x0A:4, and
0x12:4) to Logic 0. First, the loop finds the square (or power) of
the incoming complex data signal by squaring I and Q and
adding them. This operation is implemented in exponential
domain using 2
x
.
The AGC loop has average and decimate blocks that operate on
power samples before the square root operation, as shown in
Figure 51. The average block can be programmed to average
1 to 16,384 power samples, and the decimate block can be pro-
grammed to update the AGC once every 1 to 4096 samples. The
limitations on the averaging operation are that the number of
averaged power samples must be an integer multiple of the
decimation value, and the only allowable multiple values are
1, 2, 3, or 4.
The averaging and decimation effectively mean that the AGC
can operate over averaged power of 1 to 16,384 output samples.
The choice of updating the AGC once every 1 to 4096 samples
and operating on average power facilitates the implementation
of a loop filter with slow time constants, where the AGC error
converges slowly and makes infrequent gain adjustments. It
would also be useful where the user wants to keep the gain
scaling constant over a frame of data (or a stream of symbols).
Figure 51. Block Diagram of the AGC
Three sources of error can be introduced by the AGC function
underflow, overflow, and modulation. Underflow is caused by
tr
cli