A Single-Chip Universal Digital Satellite Receiver with 480

A Single-Chip Universal Digital Satellite Receiver with 480
1634
IEEE JOURNAL OF SOLID-STATE CIRCUITS, VOL. 34, NO. 11, NOVEMBER 1999
A Single-Chip Universal Digital Satellite
Receiver with 480-MHz IF Input
Alan Y. Kwentus, Member, IEEE, Patrick Pai, Steve Jaffe, Ray Gomez, Sean Tsai, Tom Kwan,
Hing-Tsun Hung, Young J. Shin, Vin Hue, Darwin Cheung, Member, IEEE, Raheel A. Khan,
Christopher M. Ward, Mong-kai Ku, Kenneth Choi, Jim Searle, Klaas Bult, Kelly Cameron, Member, IEEE,
Jason Demas, Charles Reames, and Henry Samueli, Member, IEEE
Abstract—This paper presents a complete single-chip universal
digital satellite receiver supporting all current DBS system
standards. The mixed-signal device accepts a modulated data
stream at up to 90 Mbps and delivers a demodulated, errorcorrected output data stream. The IC features an analog front
end with 480-MHz intermediate-frequency downconversion
and dual 8-bit analog-to-digital converters, an all-digital
BPSK/QPSK/OQPSK variable-rate receiver supporting 1–45MBaud operation with phase/frequency recovery, variable-rate
digital filters, square-root Nyquist matched filters, acquisition
and tracking loops, and a DVB/DSS/DigiCipher I/II-compliant
concatenated Viterbi/Reed–Solomon forward error correction
decoder with on-chip deinterleaver RAM. All required clocks
are generated on chip from a single reference crystal. The chip
contains 1.2 million transistors in a die area of 22 mm2 and
was implemented in a single-poly 0.35-m CMOS process with
four layers of metal.
I. INTRODUCTION
Fig. 1. Satellite TV system block diagram.
A
BLOCK diagram of a typical satellite TV system is
shown in Fig. 1. The system contains an antenna that
receives the K-band signal (11–12 GHz) transmitted by the
satellite and a low noise block (LNB) that downconverts the
signal for transmission over a coaxial cable to the set-top
box. In Europe, the signal is downconverted by the LNB to
950–2150 MHz; in North America, the LNB downconverts the
signal to 950–1450 MHz. An L-band tuner then selects the
desired transponder channel. Fig. 2 shows a block diagram
of a typical L-band satellite TV tuner. Most tuners in use
today downconvert the signal to a 480-MHz intermediatefrequency (IF) and then use either a 36- or 55-MHz surface
acoustic wave (SAW) filter to remove adjacent channels and
out-of-band noise. An analog I/Q demodulator then produces
a quadrature baseband (I/Q) output. Next, the digital data
bitstream is recovered by a receiver, and error correction is
performed to improve the bit error rate of the data. The data
stream is then passed to a conditional access module that
controls access to pay-per-view and premium channels. Last,
an MPEG decoder recovers the digitized video data, and a
video encoder produces an NTSC or phase alternation line
(PAL) video output for a television.
Manuscript received May 12, 1999; revised June 15, 1999.
The authors are with Broadcom Corp., Irvine, CA 92618 USA (e-mail:
[email protected]).
Publisher Item Identifier S 0018-9200(99)08336-5.
II. CHIP ARCHITECTURE
Fig. 3 shows a block diagram of the universal digital
satellite receiver IC. The chip contains an analog front end, an
all-digital receiver, and a concatenated Viterbi/Reed–Solomon
(RS) forward error correction (FEC) decoder. The analog
front end consists of a 480-MHz IF downconversion block,
dual 8-bit analog-to-digital (A/D) converters that operate up
to 128 MHz, and a low-jitter phase-locked loop (PLL) that
generates all required clocks for the chip from a single external
reference crystal.
The digital receiver supports variable symbol rates from
1 to 45 MBaud with BPSK, QPSK, or OQPSK modulation.
The 8-bit A/D converter(s) operate with a free-running sample
clock, and symbol timing is recovered digitally using multirate
digital filters and a digital symbol timing recovery loop. The
receiver also contains dual square-root Nyquist matched filters
and digital carrier frequency recovery and phase-tracking loops
with additional circuitry for automatic carrier acquisition that
avoids all false lock points.
The error-correction block consists of a 64-state Viterbi
decoder concatenated with a
RS decoder. The Viterbi
decoder supports puncture modes for rate 5/11, 1/2, 3/5,
2/3, 3/4, 4/5, 5/6, 6/7, and 7/8 operation with circuitry to
automatically scan the data for the correct puncture mode.
The decoder operates as a rate 1/2 based decoder for European
0018–9200/99$10.00  1999 IEEE
KWENTUS et al.: SINGLE-CHIP UNIVERSAL DIGITAL SATELLITE RECEIVER
1635
Fig. 2. Block diagram of a typical L-band satellite TV tuner.
Fig. 3. Universal digital satellite receiver IC block diagram.
Digital Video Broadcast (DVB), or DirecTV Digital Satellite
System (DSS), and as a rate 1/3 based decoder for Digicipher
I/II operation. The RS decoder implements the DVB, DSS, and
Digicipher I/II standards. It contains an on-chip deinterleaving RAM and can output either DSS or MPEG-2 formatted
packets.
The chip architecture supports three tuner interface modes:
baseband I/Q, low IF, and 480-MHz IF. For the baseband
I/Q mode, the dual 8-bit A/D converters are used with a
conventional tuner supplying quadrature baseband outputs, as
shown in Fig. 4(a). Alternatively, the low-IF tuner interface
mode allows the I/Q demodulator in the tuner to be replaced
with a simple mixer that downconverts the signal to a low
IF frequency within the bandwidth of the A/D converter,
as shown in Fig. 4(b). A single A/D converter sampling at
128 MHz is used in this mode. For the lowest overall system
cost, the I/Q demodulator in the tuner can be eliminated
entirely and the chip can accept the 480-MHz IF signal
directly, as shown in Fig. 4(c). In this mode of operation,
the signal is downconverted on-chip to a 32-MHz IF using
a 512-MHz mixer clock generated by the PLL and then
directly sampled by a single 8-bit A/D converter operating at a
128-MHz sample rate. The I/Q demodulation is then performed
digitally in the receiver.
III. ANALOG FRONT END
The integrated analog front end (AFE) accepts an analog input waveform in one of three formats: 480-MHz IF,
low-frequency IF, or baseband I/Q. Fig. 5(a) shows a block
diagram of the AFE, which switches the input signal according
1636
IEEE JOURNAL OF SOLID-STATE CIRCUITS, VOL. 34, NO. 11, NOVEMBER 1999
(a)
(b)
Fig. 4. Tuner interface modes: (a) baseband I/Q and (b) low IF.
to the desired mode of operation. The signal path for each of
the three modes is indicated in the figure.
In the 480-MHz IF mode, the chip accepts a
BPSK/QPSK/OQPSK input signal centered at 480 MHz.
This signal is internally mixed to a lower IF frequency of
32 MHz using a 512-MHz on-chip PLL. It is then low-pass
filtered and converted to a digital signal by an 8-bit A/D
converter operating at a sampling rate of 128 MHz. A
simplified schematic of the CMOS mixer and a block diagram
of the PLL are shown in Fig. 5(b) and (c). The on-chip
PLL is a four-stage differential ring oscillator operating at
an internal frequency of 512 MHz. The 512-MHz clock
is divided by 32 and phase-locked to a 16-MHz off-chip
crystal reference.
In the low-frequency IF mode of operation, the chip accepts
an IF input signal with a frequency spectrum within the A/D
converter’s Nyquist bandwidth. In this low-IF mode, the onchip 480-MHz mixer is bypassed, and the signal is converted
by the 8-bit A/D converter operating at a sampling rate of
128 MHz. Last, in the baseband I/Q mode, the chip accepts
KWENTUS et al.: SINGLE-CHIP UNIVERSAL DIGITAL SATELLITE RECEIVER
1637
(c)
Fig. 4. (Continued.) Tuner interface modes: (c) 480-MHz IF.
baseband I and Q input signals, which are sampled by dual
8-bit A/D converters operating at a maximum sampling rate
of 128 MHz.
The ADC utilizes folding, interpolation, and averaging
to achieve good integral- and differential-nonlinearity with
relatively low complexity and power [1]. It differs from the
ADC described in [1] in that it operates at a much higher
sample rate (128 versus 50 Msample/s) with reduced resolution
(8 versus 10 bits). Also the noise environment on the chip is
quite different due to the integrated 45-MBaud digital receiver
and FEC circuits. The ADC uses two cascaded stages of three
times folding to reduce the number of comparators required
from 256 to 32. Fig. 6(a) and (b) shows the measured integral
nonlinearity (INL) and differential nonlinearity (DNL) for the
8-bit ADC. Fig. 7 shows the measured ADC output frequency
spectrum for an input tone at 11 MHz. All measured data were
taken with the chip in 480-MHz IF mode and with the digital
receiver and FEC locked (i.e., all digital circuitry on the chip
operating). The power dissipation of the ADC is approximately
160 mW from a 3.3-V supply. Since the 512-MHz mixer was
designed to achieve 10-bit linearity, the performance of the
analog front end is dominated by that of the ADC.
IV. DIGITAL RECEIVER
The digital receiver is a BPSK/QPSK/OQPSK receiver
capable of variable-rate operation from 1 to 45 MBaud. It
consists of a frequency recovery loop, variable-rate digital filters controlled by a timing recovery loop, square-root Nyquist
matched filters, a phase tracking loop, and circuitry to control
loading the inputs to the Viterbi decoder. A block diagram of
the receiver is shown in Fig. 8. The inputs to the frequency
recovery loop and phase tracking loop can be taken from
several different points in the system, allowing the loops to
be operated independently or nested together under software
control. This flexibility provides optimal performance of the
receiver under varying operating conditions.
The automatic gain control (AGC) loop is a noncoherent
loop designed to maintain a constant input power level into
the A/D converters. The loop produces a 1-bit
-modulated
output that is filtered off chip by a simple RC low-pass filter,
providing a voltage control for the tuner’s AGC amplifier.
Input for the loop can be taken from either the I-channel or
the Q-channel data, or an average of the two. A programmable
loop threshold allows the input power level to be specified by
the user.
The carrier frequency recovery and phase tracking loops
are high-speed all-digital phase/frequency circuits capable of
tracking out relatively large amounts of frequency offsets
and phase noise such as those contributed by conventional
tuners and LNB’s. The loops are each filtered by an integralplus-proportional filter. Programmable integrator and linear
coefficients are provided to set the bandwidth of the loops,
and a carrier sweep capability is included to reduce acquisition time and avoid all false lock points [2]. The output
of the loop filter is used to control a complex derotator
[3], providing extremely fine phase and frequency resolution.
The phase/frequency recovery loops are capable of removing
residual phase and frequency offsets in the signal of up to
16 MHz. An important parameter for satellite receivers is
acquisition time. The automated frequency recovery loop in
this receiver can acquire an input signal with 16-MHz offset
in less than 50 ms.
1638
IEEE JOURNAL OF SOLID-STATE CIRCUITS, VOL. 34, NO. 11, NOVEMBER 1999
(a)
(b)
(c)
Fig. 5. (a) Analog front end block diagram, (b) PLL block diagram, and (c)
CMOS mixer schematic.
The timing recovery loop consists of a timing error discriminant, a loop filter, and a digital timing recovery block, which
controls the digital resampler. The timing error discriminant
outputs a new value each symbol, which is filtered by a digital
integral-plus-proportional low-pass filter. The output of the
loop filter is applied to a numerically controlled oscillator,
which allows the symbol rate to be varied from 1 to 45 MBaud.
The variable-rate interpolation filters [4], which are under the
control of the timing recovery loop, resample the input data at
the correct frequency and phase to ensure that ideally sampled
symbols are input to the matched filters. The matched filters
consist of dual square-root Nyquist filters with a programmable
20% or 35% excess bandwidth factor. The output of the
Nyquist filters is then scaled to optimally load the Viterbi
decoder.
The receiver can also be configured to perform spectral
analysis of the input signal with a resolution less than 50 kHz
to aid in “blind” acquisition when the carrier frequency and/or
symbol rate are not known.
Although the satellite-TV QPSK receiver architecture is
similar to that of a quadrature amplitude modulation (QAM)
receiver for cable-TV applications [5], there are some significant differences. The first major architectural difference is that
the carrier recovery loop must be closed at the front end of the
receiver to compensate for the large allowable frequency offset
in the input signal (up to 5 MHz). This frequency offset
is introduced in the LNB when the signal is downconverted
from 11–12 GHz to 950–2150 MHz. This offset is very small
as a percentage of the K-band carrier frequency; however, it
is a rather large absolute offset that must be removed at the
front of the receiver to ensure that the signal is within the
passband of the matched filters. In the case of a cable system,
the transmitted signal is typically on a carrier less than 1 GHz
so the frequency offsets at the input of the receiver are much
smaller. The second major architectural difference is that little
or no equalization is required in the satellite-TV receiver. In
a cable system, the channel contains many impairments that
cause notches in the spectrum and intersymbol interference
(ISI). An equalizer is typically required to remove these
impairments. In the satellite-TV system, however, cable-type
impairments are not typical because the signal is transmitted
on a very high-frequency carrier (11–12 GHz) through a
“line-of-sight” channel to the receiving antenna. The main
impairment in the satellite-TV system is noise in the channel.
Since the signal is transmitted from a satellite in orbit many
miles from the receiver, it typically has a very low signalto-noise ratio (SNR) at the input to the receiver (typically
5–10 dB). The cable channel, on the other hand, typically has
a very high SNR ( 30 dB). The low SNR of the satelliteTV system mandates the use of a lower order modulation
format (QPSK) versus the higher order modulation formats
used in the cable-TV systems (64/256/1024-QAM). Because
of the lower order modulation format, the symbol rate must be
increased to achieve comparable bit rates, thus requiring the
use of more dedicated hardware (i.e., less hardware sharing in
the architecture).
Fig. 9 shows the receiver constellation for a 1-MBaud
input signal with no noise, clearly illustrating the wide dynamic range of the 8-bit A/D converters and the datapath
in the receiver. The receiver performance including a conventional tuner is typically within 0.2 dB from theory at
quasi-error-free operating conditions. Fig. 10 shows measured
bit error rate (BER) performance of the receiver for 1-, 20-,
and 45-MBaud input signals. All data were taken using the
480-MHz IF input mode with the A/D converter sampling
at 128 MHz and include the degradation of a conventional
480-MHz L-band tuner. For all symbol rates, the overall
KWENTUS et al.: SINGLE-CHIP UNIVERSAL DIGITAL SATELLITE RECEIVER
1639
(a)
(b)
Fig. 6. Measured (a) ADC DNL and (b) ADC INL.
degradation (receiver tuner) is well within the required DVB
limit of 0.8 dB. The best performance occurs at 20 MBaud,
approximately in the center of the 1–45-MBaud range. The
performance degrades slightly at the two extremes of 1 and
45 MBaud. The degradation at 1 MBaud is primarily due
to phase noise in the tuner and finite wordlength effects in
the digital filters that recover the symbol timing from the
highly oversampled input data (128-MHz sample rate versus
1-MBaud input signal) when adjacent channel interferers are
present. The primary cause for the performance degradation
at 45 MBaud is that the tuner’s 55-MHz SAW filter cuts off
some of the excess bandwidth of the input signal. (Note that a
45-MBaud signal shaped by a square-root Nyquist filter with
a 0.35 alpha has a bandwidth of 60.75 MHz.) Additionally,
the performance of the chip’s analog front end degrades
as the signal approaches the Nyquist frequency of the A/D
converters.
Typical satellite systems operate in extremely noisy environments. Thus, receiver performance under low SNR conditions
(i.e., high noise levels) is extremely important. Also, since the
tuner’s SAW filter bandwidth must be wide enough to pass
a 45-MBaud signal, at lower symbol rates significant outof-band energy is allowed to pass through to the receiver,
and sufficient filtering must be performed in the variable-rate
1640
IEEE JOURNAL OF SOLID-STATE CIRCUITS, VOL. 34, NO. 11, NOVEMBER 1999
Fig. 7. Measured ADC frequency spectrum for 11-MHz input tone.
Fig. 8. BPSK/QPSK/OQPSK receiver.
receiver to remove the out-of-band energy. Fig. 11 shows
a 4.42-MBaud signal in a 3-dB SNR (0-dB Eb/No) noise
environment with two interfering signals within the bandwidth
of the A/D converters, each 8 dB higher in power. Measured
data show that the receiver’s BER degradation due to the
interferers is less than 0.1 dB. This is due to the 8-bit
resolution of the A/D converters and the advanced multirate
digital filtering in the datapath. Under high-noise operating
conditions, QPSK receivers also suffer from a phenomenon
known as cycle slips, where the receiver instantaneously
“slips” phase lock by 90 , 180 , or 270 . The cycle slip
cannot be detected by the receiver because it remains in
lock, just with a different phase. Cycle slips occur when
transmitted symbols repeatedly appear to the receiver as an
adjacent symbol because of the additive noise in the channel.
As the noise level in the channel increases, the cycle slip rate
of the receiver also increases. The IC described in this paper
uses two approaches to improve the cycle-slip performance of
the receiver. The first is to use a low delay loop for carrier
phase tracking (see Fig. 8), which allows for a much wider
loop bandwidth and improved phase noise performance. The
second improvement is the addition of a nondecision directed
phase detector in the carrier tracking loop that has been
optimized for low SNR operation [6]. Using these techniques,
the receiver is capable of reliably acquiring an input signal at
an SNR of 2.0 dB (Eb/No
dB) with measured cycleslip performance better than one slip every three hours using a
conventional satellite tuner. Fig. 12 shows measured cycle-slip
KWENTUS et al.: SINGLE-CHIP UNIVERSAL DIGITAL SATELLITE RECEIVER
Fig. 9. Receiver constellation for a 1-MBaud input signal with no noise.
Fig. 10.
Receiver BER performance for 1, 20, and 45 MBaud.
1641
1642
IEEE JOURNAL OF SOLID-STATE CIRCUITS, VOL. 34, NO. 11, NOVEMBER 1999
(a)
(b)
Fig. 11. Input spectrum and QPSK constellation plot for 4.42-MBaud input
with 8-dB interferers in 3-dB SNR noise environment.
+
performance of the receiver versus SNR using the 480-MHz
IF input mode with a conventional L-band tuner.
V. FORWARD ERROR CORRECTION
Since satellite systems operate in extremely noisy environments, powerful error correction is required to achieve bit
error rates appropriate for transmission of digitally compressed
video signals. This IC contains an FEC decoder compatible
with the European Digital Video Broadcast (DVB), DirecTV
Digital Satellite System (DSS), and Primestar DigiCipher I/II
forward error correction and framing scheme for satellite
transmission. To achieve the level of error detection and
correction required, a concatenated FEC code is used. This
code is based on convolutional encoding for the inner code
and RS encoding for the outer code. Protection against burst
errors is achieved using byte interleaving. To support these
functions, the decoder consists of five layers: Viterbi decoding,
frame synchronization, convolutional deinterleaving, RS error
correction, and derandomization.
The Viterbi decoder can decode two families of constraint
length
, 64-state punctured convolutional codes: the
DVB/DSS rate 1/2 (171, 133) code and the DigiCipher I/II
rate 1/3 (117, 135, 161) code. The Viterbi decoder consists
of a depuncturing block, a dual-rate capable 64-state Viterbi
core, and a node synchronization block, as shown in Fig. 13.
The depuncturing block supports puncture rates 1/2, 2/3,
3/4, 5/6, 6/7, or 7/8 for the DVB/DSS rate 1/2 code, and
puncture rates 5/11, 1/2, 3/5, 2/3, 3/4, 4/5, 5/6, or 7/8 for the
DigiCipher I/II rate 1/3 code. The depuncture block accepts
QPSK, BPSK, or offset QPSK (OQPSK) soft decisions from
the receiver and inserts dummy cycles to restore the original
rate 1/2 or 1/3 timing relationship before sending the data
stream to the Viterbi core. The 64-state Viterbi core design
employs a dual-rate capable add-compare-select block that can
handle both codes with minimal chip area penalty. The node
synchronization block corrects the receiver phase ambiguity
and depuncturing timing ambiguity in the decoding process.
In the case of a receiver cycle slip, the node synchronization
logic will attempt to correct the cycle slip before triggering the resynchronization process. The node synchronization
block also provides the capability to automatically scan for
the correct puncture rate. The frame synchronization block
will recover either MPEG-2 or DSS transport framed data
streams as defined in the DVB, DSS, and DigiCipher I/II
specifications.
In order to deal with burst errors in the channel, the
coded data stream is interleaved so that the bursty channel
is transformed into a channel having independent channels.
The convolutional deinterleaver supports all of the standards
by providing programmable interleaving depths of 1, 2, 3, 4, 6,
12, and 13 and programmable block sizes. The deinterleaving
RAM is included on-chip.
The RS decoder supports
,
or
; or
, (125 115) systematic Reed–Solomon
codes. For DVB, DSS, or DigiCipher II applications, the RS
code is a systematic code with 8-bit symbols, shortened from
a block length of 255 symbols and correcting up to
symbols per block. The finite field GF (256) is constructed
.
from the primitive polynomial
The generator polynomial is
. The data randomization block scrambles data with a
pseudorandom number sequence for energy dispersal and to
ensure a high data transition density for timing recovering
purposes.
VI. CHIP PERFORMANCE
The IC has been extensively tested and is being deployed
in DBS systems around the world. BER performance of the
device has been measured to be well within all DVB, DSS,
and DigiCipher I/II specifications. Fig. 14 shows BER performance for 1-, 20-, and 45-MBaud operation at three different
Viterbi puncture rates using the 480-MHz IF input mode
and a conventional L-band tuner. These performance data
span the chip’s operating range from 1 to 45 MBaud symbol
KWENTUS et al.: SINGLE-CHIP UNIVERSAL DIGITAL SATELLITE RECEIVER
Fig. 12.
Receiver cycle-slip performance.
Fig. 13.
Viterbi decoder block diagram.
TABLE I
IC SUMMARY
rates and puncture rates from 1/2 to 7/8. Since the Viterbi
and Reed–Solomon decoders are all-digital operations whose
performance degradation is independent of the symbol rate,
1643
the chip’s performance loss is dominated by the performance
loss of the receiver. Thus the best performance for the overall
chip occurs in the 20-MBaud example with some performance
degradation at the two extremes of 1 and 45 MBaud, as was the
case for the receiver performance by itself. For all cases, the
chip’s performance is within the required DVB limit of 0.8-dB
loss, including the effects of the L-band 480-MHz tuner.
Fig. 15 shows the chip micrograph. The IC is packaged in an
80-pin PQFP package. Power dissipation is 2.1 W at worst case
supply voltage of 3.465 V (3.3 V 5%) with 128-MHz sample
rate and 45-MBaud symbol rate under worst case temperature
and process. The power dissipation breakdown is 0.7 W for
the analog circuitry and 1.4 W for the digital circuitry.
Table I summarizes the details of the IC.
VII. CONCLUSION
Currently, most satellite tuners output analog I- and Qwaveforms at baseband. The 480-MHz IF mode for the IC
1644
Fig. 14.
IEEE JOURNAL OF SOLID-STATE CIRCUITS, VOL. 34, NO. 11, NOVEMBER 1999
Overall chip BER performance for 1-MBaud rate 1/2, 20-MBaud rate 3/4, and 45-MBaud rate 7/8 QPSK.
performance data include the effects of digital switching noise
on the 480-MHz analog front end.
ACKNOWLEDGMENT
The authors wish to thank J. Echtenkamp, K. Kindsfater,
S. Tollefsrud, R. Hawley, A. Buchwald, M. Hatamian, V.
Nguyen, D. Bogosh, I. Tsiperfal, F. Syed, and P. LaRocca
for contributions to the development of this design.
REFERENCES
Fig. 15.
Chip micrograph.
described above allows the I/Q demodulator in the tuner to
be removed, significantly reducing the overall system cost.
Similar solutions accepting 480-MHz IF input reported previously have been implemented as two or more chips, with the
high-IF analog signal processing being realized in a bipolar
IC and the digital signal processing in a CMOS IC [7]. All
components of the AFE described above are implemented in
a standard digital CMOS process, and all measured system
[1] K. Bult, A. Buchwald, and J. Laskowski, “A 170mW 10b 50MSample/s
CMOS ADC in 1mm2 ,” in ISSCC Dig. Tech. Papers, San Francisco,
CA, Feb. 1997, pp. 136–137.
[2] R. van der Wal and L. Montreuil, “QPSK and BPSK demodulator chipset for satellite applications,” IEEE Trans. Consumer Electron., vol. 41,
pp. 30–41, Feb. 1995.
[3] L. Tan and H. Samueli, “A 200-MHz quadrature digital synthesizer/mixer in 0.8-m CMOS,” IEEE J. Solid-State Circuits, vol. 30,
pp. 193–200, Mar. 1995.
[4] T. A. Ramstad, “Digital methods for conversion between arbitrary
sampling frequencies,” IEEE Trans. Acoust., Speech, Signal Processing,
vol. ASSP-32, pp. 577–591, June 1984.
[5] L. Tan, J. Putnam, F. Lu, L. D’Luna, D. Mueller, K. Kindsfater, K.
Cameron, R. Joshi, R. Hawley, and H. Samueli, “A 70-Mb/s variablerate 1024-QAM cable receiver IC with integrated 10-b ADC and FEC
decoder,” IEEE J. Solid-State Circuits, vol. 33, pp. 2205–2218, Dec.
1998.
[6] M. Simon, “On the optimality of the MAP estimation loop for carrier
phase tracking BPSK and QPSK signals,” IEEE Trans. Commun., vol.
COM-27, pp. 158–165, Jan. 1979.
[7] T. Tewksbury, W. Palmer, K. Fernald, and J. Liebetreu, “A 480 MHz
variable rate QPSK demodulator for direct broadcast satellite,” in ISSCC
Dig. Tech. Papers, San Francisco, CA, Feb. 1997, pp. 86–87.
KWENTUS et al.: SINGLE-CHIP UNIVERSAL DIGITAL SATELLITE RECEIVER
Alan Y. Kwentus (S’93–M’95), for for photograph and biography, see p.
1043 of the August 1999 issue of this JOURNAL.
Patrick Pai, photograph and biography not available at the time of publication.
1645
Young J. Shin received the B.S. degree in electrical
engineering and computer science from the University of California, Berkeley, in 1994 and the M.S.
degree in electrical engineering from the University
of California, Los Angeles, in 1996.
In 1996, he joined Broadcom Corp., Irvine, CA,
as a Design Engineer, where he is currently working
on high-speed CMOS analog front-end circuits for
communication systems.
Steve Jaffe, photograph and biography not available at the time of publication.
Ray Gomez received the B.S. degree in biomedical
engineering from Case Western Reserve University,
Cleveland, OH, in 1981, the M.S.E.E. degree from
Stanford University, Stanford, CA, in 1982, and the
Ph.D. degree from the University of California, Los
Angeles (UCLA), in 1993.
In 1982, he joined TRW, Inc., Redondo Beach,
CA, where he designed high-performance frequency
synthesizers. His research at UCLA involved analog
CMOS integrated circuits for disk-drive read channels. From 1993 to 1995, he was a member of the
Cirrus Logic/Crystal Semiconductor Read Channel Design Team. He joined
Broadcom Corp., Irvine, CA, in 1995, where he has focused on CMOS analog
and RF circuits for DBS satellite and cable television receivers.
Sean Tsai received the B.S. and M.S. degrees
in electrical engineering from the University of
Southern California, Los Angeles, in 1992 and 1994,
respectively.
He was with Lincom Corp., Los Angeles, CA,
from 1994 to 1996 as a Systems Engineer. He
joined Broadcom Corp., Irvine, CA, in 1996. His
work is mainly focused on the integrated circuit
development of DVB/DSS satellite receivers, xDSL
receivers, and VSB terrestrial receivers.
Tom Kwan received the B.A.Sc. degree in engineering science from the University of Toronto,
Toronto, Ont., Canada, in 1984 and the M.S. and
Ph.D. degrees in electrical engineering from the
University of California, Los Angeles, in 1986 and
1990, respectively.
He joined Analog Devices, Inc., in 1990 as a
Design Engineer working in the field of mixedsignal audio products. In 1996, he joined Broadcom
Corp., Irvine, CA, as a Staff Engineer. His technical
interests include analog and digital filter design,
signal processing, and data converters.
Hing-Tsun Hung received the B.A. degree in mathematics and the B.S. degree in electrical engineering
from Ohio State University, Columbus, in 1990 and
1991, respectively, and the M.S. degree in electrical
engineering from the University of California, Los
Angeles, in 1992.
From 1992 to 1996, she was an IC Design Engineer at Rockwell Semiconductor Systems, where
she worked on custom and semicustom general
purpose DSP. In 1996, she joined Broadcom Corp.,
Irvine, CA. She has worked on implementation of
satellite receivers, xDSL modems, and design of adaptive equalizers in
QAM/VSB receivers for HDTV applications.
Vin Hue received the B.S. degree in electrical
engineering from Case Western Reserve University,
Cleveland, OH, in 1990 and the M.S.E.E. degree
from the University of Southern California, Los
Angeles, in 1993.
Since 1990, he has performed custom circuit
design and ASIC design, some of which are processors for satellite applications, communications,
and disk servo. In 1996, he joined Broadcom Corp.,
Irvine, CA, where he has been designing ASIC’s for
satellite and cable set-top box applications.
Darwin Cheung (S’92–M’96) was born in Hong
Kong in 1972. He received the B.Eng.EE. (with
honors) and M.Phil.EEE. degrees in electronic engineering from Hong Kong University of Science and
Technology, Clear Water Bay, in 1994 and 1996,
respectively.
In 1996, he joined Broadcom Corp., Irvine, CA,
where he is an Analog Circuit Designer. He is
currently focused on analog and mixed-mode signal
processing, including high-speed data converters
and continuous-time filters.
Raheel A. Khan was born in Islamabad, Pakistan,
in December 1970. He received the B.E.E.
and Ms.E.E. degrees from the Georgia Institute
of Technology, Atlanta, in 1991 and 1993,
respectively.
From 1993 to 1995, he was with the Subscriber
Systems Group in Scientific Atlanta, where he
worked on system design and VLSI implementation
for cable TV systems. He is presently with
Broadcom Corp., Irvine, CA, where he is working
on advanced receiver algorithms. His main interests
lie in DSP techniques for modulation, synchronization, and detection of
signals and modeling and simulation of communications systems.
Christopher M. Ward was born in Los Angeles,
CA, in 1972. He received the B.S. degree from
the California Institute of Technology, Pasadena, in
1994 and the M.S. degree from the University of
California, Los Angeles, in 1996, both in electrical
engineering.
He joined Broadcom Corp., Irvine, CA, in 1996,
where he is currently an Analog Design Engineer.
His research interests are in analog and mixedsignal circuit design for communication and signal
processing, as well as RF circuit design.
1646
IEEE JOURNAL OF SOLID-STATE CIRCUITS, VOL. 34, NO. 11, NOVEMBER 1999
Mong-kai Ku was born in Taipei, Taiwan, R.O.C., on September 12, 1967.
He received the B.S. degree from National Taiwan University, Taipei, Taiwan,
in 1989 and the M.S. and Ph.D. degrees from the University of California,
Los Angeles, in 1994 and 1997, respectively, all in electrical engineering.
Since 1997, he has been with Broadcom Corp., Irvine, CA. He is currently
a Staff Scientist in the Cable-TV Group. His research interests include
communication systems, coding theory, and VLSI design.
Kenneth Choi was born in Hong Kong in 1971.
He received the B.S. and master of philosophy
degrees in electrical engineering from Hong Kong
University of Science and Technology, Clear Water
Bay, in 1994 and 1997, respectively.
He joined Broadcom Corp., Irvine, CA, in 1996,
where he is currently a Design Engineer. His research interests are in analog and mixed-signal
circuit design for communication and signal processing.
Jim Searle spent 11 years at the Motorola/Codex VLSI Technology Center
in Tempe, AZ, in the custom layout and CAD groups, where he worked on
methods to automate the process of creating dense custom quality layout in less
time. In 1996, he joined Broadcom Corp., Irvine, CA, where he has continued
to develop methodologies to improve physical design flow, concentrating
on the area of timing-driven standard cell place and route. He has been
responsible for the full-chip place and route and timing verification of ten
chips during this time. He is now the Manager of the Cable Television Physical
Design Group, overseeing nine engineers while continuing to automate and
improve the physical design methodologies.
Klaas Bult, photograph and biography not available at the time of publication.
Kelly Cameron (M’98) received the B.S.E.E.
and Ph.D. degrees from the University of Idaho,
Moscow, in 1980 and 1989, respectively.
He worked extensively with the microelectronics
research groups of both the University of Idaho and,
subsequently, of the University of New Mexico
as a Professor. Much of his work has been in
collaboration with Goddard Space Flight Center and
Hewlett Packard Corp. His research areas include
error correction coding, VLSI design methodologies
and languages, image processing, and compression.
He is a Cofounder of Advanced Hardware Architectures. He joined Broadcom
Corp., Irvine, CA, in 1996, where he is currently responsible for the design
and development of error correction codes and systems.
Jason Demas received the B.S. degree in computer
and electrical engineering from Purdue University,
West Lafayette, IN, in 1993 and the M.S. degree
in electrical engineering from the University of
Southern California, Los Angeles, in 1996.
From 1993 to 1996, he was with TRW, Inc.,
Redondo Beach, CA, developing digital ground
station hardware for satellite telemetry and control
processing. Since joining Broadcom Corp., Irvine,
CA, in 1996, he has been engaged in system design
and development for the digital satellite receiver and
MPEG digital video product lines.
Charles Reames, photograph and biography not available at the time of
publication.
Henry Samueli (S’75–M’79–SM’99) was born in
Buffalo, NY, on September 20, 1954. He received
the B.S., M.S., and Ph.D. degrees in electrical
engineering from the University of California, Los
Angeles (UCLA), in 1975, 1976, and 1980, respectively.
From 1980 to 1985, he was with TRW, Inc., Redondo Beach, CA, where he was a Section Manager
in the Digital Processing Laboratory of the Electronics and Technology Division. His group was
involved in the hardware design and development of
military satellite and digital radio communication systems. From 1980 to 1985,
he was also a part-time Instructor in the Electrical Engineering Department of
UCLA. In 1985, he joined UCLA full-time and is currently a Professor in the
Electrical Engineering Department. His research interests are in the areas of
digital signal processing, communications systems engineering, and CMOS
integrated circuit design for applications in high-speed data transmission
systems. In 1998, he cofounded PairGain Technologies, Inc., Tustin, CA,
a telecommunications equipment manufacturer, and in 1991 he cofounded
Broadcom Corp., Irvine, CA, an integrated circuit supplier to the broad-band
communications industry. Since 1995, he has been on leave of absence from
UCLA while serving full-time as Chief Technical Officer of Broadcom, where
he is responsible for all research and development activities for the company.
Dr. Samueli received the 1988/1989 TRW Excellence in Teaching Award
from the UCLA School of Engineering and Applied Science, the Meritorious
Paper Award of the 1991 Government Microcircuit Applications Conference
for the paper entitled “CMOS Integrated Circuits for High Bit-Rate Digital
Modems, Adaptive Equalizers and Frequency Synthesizers,” and the 1995
Best Paper Award from the IEEE JOURNAL OF SOLID-STATE CIRCUITS for the
paper entitled “A 200-MHz Quadrature Digital Synthesizer/Mixer in 0.8-m
CMOS.”
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement