HSP50210JC-52Z Intersil, HSP50210JC-52Z Datasheet - Page 8
![IC DEMODULATOR COSTAS 84-PLCC](/photos/14/15/141534/84-plcc_sml.jpg)
HSP50210JC-52Z
Manufacturer Part Number
HSP50210JC-52Z
Description
IC DEMODULATOR COSTAS 84-PLCC
Manufacturer
Intersil
Datasheet
1.HSP50210JC-52Z.pdf
(51 pages)
Specifications of HSP50210JC-52Z
Function
Demodulator
Frequency
52MHz
Rf Type
AM, FM
Package / Case
84-PLCC
Lead Free Status / RoHS Status
Lead free / RoHS Compliant
NCO/Mixer
The NCO/Mixer performs a complex multiply between the
baseband input and the output of a quadrature NCO
(Numerically Controlled Oscillator). When the HSP50210
(DQT) is used with the HSP50110 (DCL), the NCO/Mixer
shortens the Carrier Tracking Loop (i.e., minimizes pipeline
delay around the loop) while providing wide loop
bandwidths. This becomes important when operating at
symbol rates near the maximum range of the part.
There are three configurations possible for closing the
Carrier Tracking Loop when the DQT and the DCL are used
together. The first configuration utilizes the NCO on the DQT
and bypasses the NCO in the DCL. The Data Path
Configuration Control Register (see Table 15 on page 32),
Bit 10, and Carrier Loop Filter Control Register #1
(see Table 21 on page 34), Bit 6, are used to bypass the
DCL NCO/Mixer and route the Loop filter outputs,
respectively. The DQT provides maximum flexibility in NCO
control with respect to frequency and phase offsets.
The second configuration feeds the lead Carrier Loop filter
term to the DCL NCO/Mixer, and the lag Loop filter Term to
the DQT NCO. This reduces the loop transport delay while
maintaining wide loop bandwidths and reasonable loop
damping factors. This configuration is especially useful in
SATCOM applications with medium to high symbol rates.
The Carrier Loop Filter Control Register #1, Bit 5 is where
the lead/lag destination is set.
The final configuration feeds both the lead and lag Carrier
Loop Filter terms back to the DCL NCO/Mixer. This provides
the shortest transport delay. The DCL NCO/Mixer provides
only for frequency/phase control from the Carrier Loop filter.
The center frequency of this NCO/Mixer is set to the average
of the Upper and Lower Carrier Loop Limits programmable
parameters. These parameters are set in the two control
registers bearing their names (see Tables 23 and 24 on
page 35).
The NCO/Mixer uses a complex multiplier to multiply the
baseband input by the output of a quadrature NCO. This
operation is represented by Equations 2 and 3:
Equation 3 illustrates how the complex multiplier implicitly
performs the summing function when the DCL is configured
as a modulator. The quadrature outputs of the NCO are
generated by driving a sine/cosine look-up table with the
output of a phase accumulator, as shown in Figure 3 on
page 7. Each time the phase accumulator is clocked, its sum
is incremented by the contents of the Carrier Frequency (CF)
Register. As the accumulator sum increments from 0 to 2
the SIN/COS ROM produces quadrature outputs whose
phase advances from 0 to 360°. The CF Register contains a
I OUT
Q OUT
=
=
I IN
I IN
cos
sin
(
ω C
(
ω C
)
–
)
Q IN
+
Q IN
sin
cos
(
ω C
8
(
ω C
)
)
(EQ. 2)
(EQ. 3)
32
HSP50210
,
32-bit phase increment, which is updated with the output of
Carrier Tracking Loop. Large phase increments take fewer
clocks to step through the sine wave cycle, which results in a
higher frequency NCO output.
The CF Register sets the NCO frequency using Equation 4:
where f
complement hexadecimal value loaded into the Carrier
Frequency Register. As an example, if the CF Register is
loaded with a value of 4000 0000 (Hex), and the CLK
frequency is 40MHz, the NCO would produce quadrature
terms with a frequency of 10MHz. When CF is a negative
value, a clockwise cos/sin vector rotation is produced. When
CF is positive, a counterclockwise vector rotation is
produced.
Note: The NCO is set to a fixed frequency by programming
the upper and lower limits of the Carrier Tracking Loop Filter
to the same value and zeroing the lead gain.
Matched Filtering
The HSP50210 provides two selectable matched filters: a
Root Raised Cosine Filter (RRC) and an Integrate and Dump
(I and D) filter. These are shown in Figure 3. The RRC filter
is provided for shaped data pulses and the I and D filter is
provided for square wave data. The filters may be cascaded
for better adjacent channel rejection for square wave data. If
these two filters do not meet baseband filtering
requirements, then they can be bypassed and an external
digital filter (such as the HSP43168 Dual FIR Filter or the
HSP43124 Serial I/O Filter) used to implement the desired
matched filter. The desired filter configuration is set in the
Data Path Configuration Control Register, bits 1 through 7
(see Table 15 on page 32).
The sample rate of the baseband input depends on the
symbol rate and filtering configuration chosen. In
configurations which bypass both filters or use only the RRC
Filter, the input sample rate must be twice the symbol rate. In
configurations which use the I and D Filter, the input sample
rate is decimated by the I and D Filter, down to two samples
per symbol. I and D configurations support input sample
rates up to 32x the input symbol rate.
The RRC filter is a fixed coefficient 15 Tap FIR filter. It has
~40% excess bandwidth beyond Nyquist, which equates to
α = ~0.4 shape factor. The filter frequency response is
shown in Figures 4 and 5. In addition, the 9-bit filter
coefficients are listed as integer values in Table 1. The noise
equivalent bandwidth of the RRC filter and other filter
configurations possible with the HSP50110/210 chipset are
given in Appendix A.
F
CF
C
=
=
f
INT F
CLK
CLK
[
(
×
is the CLK frequency, and CF is the 32-bit two’s
(
C
CF
⁄
f
CLK
) 2 ⁄
32
)2
32
]H
July 2, 2008
FN3652.5
(EQ. 4)