HYNIX HDM8515P

HDM8515 Users Manual
DVB/DSS Compliant Receiver
Dec. 2000
Priliminary
1
Direct Broadcast Satellite (DBS) has been one of the most successful new product
introductions in the history of consumer electronics. This product represents the first
application of digital video compression for broadcast television. Originally intended to
provide cable quality television services to remote areas, this product is now offering a
competitive replacement to cable services in many urban areas.
The first operational systems employ closed proprietary signaling structures. The
European Broadcasting Union (EBU) has developed the first open standard (DVB-S) for
DBS services. The broadcasting community has embraced this standard which is now
being adopted for new systems throughout the world. This widely accepted open
standard is essential for DBS to achieve full market potential.
The HDM8515TM is a fully DVB-S&DSS compliant ADC/QPSK demodulator/FEC device
which provides an MPEG-2 stream to be processed by the conditional access and video
decompression circuits. The demodulator clocked with a fixed frequency is true variable
rate over the range of 1 to 55M symbols-per-second. This product achieves the highest
performance and flexibility. It minimizes the cost of external circuits, thus reducing
overall system cost.
2
Hynix Semiconductor Co., Ltd reserves the right to make changes to its products or
specifications to improve performance, reliability, or manufacturability. Information
furnished by Hynix Semiconductor Co., Ltd is believed to be accurate and reliable.
However, no responsibility is assumed by Hynix Semiconductor Co., Ltd for its use; nor
for any infringement of patents or other rights of third parties which may result from its
use. No license is granted by its implication or otherwise under any patent rights of
Hynix Semiconductor Co., Ltd.
For more information contact:
Address: Youngdong Bldg. 891, Daechi-dong, Kangnam-gu, Seoul, 135-738, Korea
Tel: 82-2-3459-3188
Fax: 82-2-3459-5843
E-mail: [email protected]
3
TABLE OF CONTENTS
1. INTRODUCTION TO THE HDM8515...................................................................................................................7
1.1 FEATURES AND BENEFITS..................................................................................................................................8
2. HARDWARE SPECIFICATION..............................................................................................................................9
3. TECHNICAL OVERVIEW..................................................................................................................................... 19
3.1 DUAL CHANNEL A NALOG TO DIGITAL CONVERTER .................................................................................. 19
3.2 VARIABLE RATE DEMODULATOR.................................................................................................................. 21
3.3 NOISE MEASUREMENT CIRCUIT .....................................................................................................................23
3.4 VITERBI DECODER.............................................................................................................................................25
3.5 A UTONOMOUS A CQUISITION ..........................................................................................................................26
3.6 REED SOLOMON DECODER.............................................................................................................................. 28
3.7 CLOCK GENERATION PLL.................................................................................................................................30
3.8 DBS RECEIVER................................................................................................................................................... 35
3.9 DISEQC INTERFACE ...........................................................................................................................................36
4. MECHANICAL SPECIFICATIONS..................................................................................................................... 37
4.1 100 PIN QUAD FLAT PACK ................................................................................................................................37
4.2 64 PIN THIN QUAD FLAT PACK........................................................................................................................39
4.3 RECOMMENDED A NALOG PIN CONNECTION............................................................................................... 41
4.4 RECOMMENDED CLOCK GENERATION CIRCUIT ...........................................................................................41
5. SIGNAL DESCRIPTION....................................................................................................................................... 42
5.1 INPUTS..................................................................................................................................................................42
5.2 OUTPUTS............................................................................................................................................................. 42
5.3 M ONIT OR AND CONTROL INTERFACE ...........................................................................................................45
5.4 I2C MODE............................................................................................................................................................. 46
6. REGISTER DEFINITIONS..................................................................................................................................... 48
6.1 W RITE REGISTERS..............................................................................................................................................48
6.2 READ REGISTERS................................................................................................................................................61
APPENDIX.................................................................................................................................................................... 66
A1.
A2.
A3.
A4.
4
LOOP FILTER PROGRAMMING APPLICATION NOTE ................................................................................67
FALSE LOCK ESCAPE A PPLICATION NOTE .................................................................................................70
PERFORMANCE WITH INTERFERENCE .......................................................................................................... 71
NYQUIST CRITERIA CONSIDERATIONS......................................................................................................... 75
LIST OF FIGURES
FIGURE 1: TOP LEVEL BLOCK DIAGRAM....................................................................................................................7
FIGURE 2: INPUT DATA TIMING DIAGRAM .............................................................................................................10
FIGURE 3: INTEL 80C88A READ TIMING DIAGRAM ............................................................................................... 11
FIGURE 4: INTEL 80C88A WRITE TIMING DIAGRAM............................................................................................. 12
FIGURE 5: INTEL 8051 READ TIMING DIAGRAM .....................................................................................................13
FIGURE 6: INTEL 8051 WRITE TIMING DIAGRAM...................................................................................................14
FIGURE 7: MOTOROLA READ TIMING DIAGRAM....................................................................................................15
FIGURE 8: MOTOROLA W RITE TIMING DIAGRAM .................................................................................................16
FIGURE 9: OUTPUT TIMING DIAGRAM FOR NORMAL PARALLEL ....................................................................... 17
FIGURE 10: OUTPUT TIMING DIAGRAM FOR NORMAL SERIAL ...........................................................................17
FIGURE 11: OUTPUT TIMING DIAGRAM FOR REGULATED PARALLEL............................................................... 18
FIGURE 12: OUTPUT TIMING DIAGRAM FOR REGULATED SERIAL .....................................................................18
FIGURE 13: ADC BLOCK DIAGRAM............................................................................................................................ 20
FIGURE 14 DEMODULATOR BLOCK DIAGRAM........................................................................................................ 21
FIGURE 15: NOISE M EASUREMENT CIRCUIT ...........................................................................................................23
FIGURE 16: NOISE A CCUMULATOR AS A FUNCTION OF SNR AND TIME............................................................ 24
FIGURE 17: VITERBI DECODER...................................................................................................................................25
FIGURE 18: REED SOLOMON DECODER.................................................................................................................... 29
FIGURE 19: TYPICAL SET TOP BOX DEMODULATOR............................................................................................ 35
FIGURE 20: M ECHANICAL CONFIGURATION ...........................................................................................................38
FIGURE 21: M ECHANICAL CONFIGURATION ...........................................................................................................40
FIGURE 22: A NALOG PIN CONNECTION .................................................................................................................... 41
FIGURE 23: CLOCK GENERATION CIRCUIT...........................................................................................................41
FIGURE 24: I2C WRITE TO THE HDM8515..............................................................................................................46
FIGURE 25: I2C READ FROM THE HDM8515............................................................................................................47
FIGURE A1: SYMBOL TIMING RECOVERY TRANSIENT RESPONSE....................................................................... 67
FIGURE A2: CARRIER PHASE RECOVERY TRANSIENT RESPONSE ........................................................................ 68
FIGURE A3: CARRIER PHASE RECOVERY TRANSIENT RESPONSE WITH LOW SNR ..........................................69
FIGURE A4: A DJACENT CHANNEL INTERFERENCE OF 10 DB, 1.35 SPACING.................................................... 72
FIGURE A5: PERFORMANCE WITH INTERFERER AT DIFFERENT CARRIER SPACINGS .....................................73
FIGURE A6: PERFORMANCE WITH +10 DB INTERFERER......................................................................................74
5
LIST OF TABLES
TABLE 1: ABSOLUTE M AXIMUM RATINGS ...............................................................................................................9
TABLE 2: DC CHARACTERISTICS.................................................................................................................................9
TABLE 3: DEMODULATOR SPECIFICATIONS...........................................................................................................10
TABLE 4: AC CHARACTERISTICS............................................................................................................................... 10
TABLE 5: INTEL 80C88A READ CYCLE TIMING PARAMETERS (BUSMODE = 1)................................................11
TABLE 6: INTEL 80C88A WRITE CYCLE TIMING PARAMETERS (BUSMODE = 1) .............................................12
TABLE 7: INTEL 8051 READ CYCLE TIMING PARAMETERS (BUSMODE = 1)......................................................13
TABLE 8: INTEL 8051 WRITE CYCLE TIMING PARAMETERS (BUSMODE = 1)................................................... 14
TABLE 9: M OTOROLA READ CYCLE TIMING PARAMETERS (BUSMODE =0).................................................... 15
TABLE 10: M OTOROLA W RITE CYCLE TIMING PARAMETERS (BUSMODE =0).................................................16
TABLE 11: OUTPUT TIMING....................................................................................................................................... 17
TABLE 12: EXAMPLE OF ACQUISITION TIMING .....................................................................................................27
TABLE 13: I2C SLAVE A DDRESS..................................................................................................................................47
6
1. Introduction to the HDM8515
The HDM8515 digital demodulator for direct broadcast satellite receivers is a single chip solution fully
compliant with the European Telecommunications Standards Institute (ETSI) specification ETS 300
421. This chip integrates an A/D converter, a variable rate matched filter, a variable rate QPSK
demodulator with a Viterbi decoder, a deinterleaver and a Reed Solomon decoder.
The HDM8515, which is implemented in a 0.25 micron CMOS, Four Layer Metal Process, provides
variable rate capability while operating with a fixed frequency sampling clock. Digital samples of
baseband I and Q data are generated by an internal A/D converter, then provided to the demodulator at
a fixed sample rate. The root raised cosine filter is implemented internally with fully digital techniques.
Similarly, the symbol timing recovery and carrier phase tracking functions are performed entirely in the
digital domain. This approach provides minimum constraints on external circ uits, thus reducing overall
system costs.
The HDM8515 may be configured by an external processor for a specific symbol rate, and carrier
frequency along with loop gain parameters. The HDM8515 provides an external AGC signal which is
used to control the gain of the analog signal which is applied to the down-converters. And it also
provides a digital AGC internally which controls the gain of the signal out of the matched filters. In
addition, the HDM8515 provides fully programmable sweep circuitry to aid in initial acquisition when
large frequency offsets may be present.
The digital frequency translation capability of the HDM8515 permits this part to be used in frequency
multiplexing applications. In this application, an entire transponder bandwidth containing many signals
is sampled at a fixed rate. The digital oscillator within the HDM8515 is programmed to the specific
desired carrier frequency within that band to permit the selected signal to be passed through the
baseband filter and processed by the demodulator circuits.
Reference clk
XTAL1_IN
C/N
Estimator
Symbol Clock
Viterbi Bit Clock
PLL
T
u
n
e
r
AIN_I
AIN_Q
WB_AGC
Data Clock
DATA_CLK
I
Q
A/D
Converter
AGC
DiSEqC
Interface
6
6
Variable
Rate
QPSK
Demodulator
4
4
Viterbi
Decoder
Synchronization
and
Deinterleaving
8
AGC_Detector
I2C
Interface
MCU
Interface
SCL_I2C
SDA_I2C
8
DATA[7:0]
QPSK Lock
QPSK
Lock
Node
Sync
QPSK_LOCK
Frame
Sync
Viterbi
Data
Byte Sync
DISEQC
Reed
Solomon
Decoder
Data
BER
Monitoring
FRAME_SYNC
HI_ADDR[5:0]
HI_DATA[7:0]
FIGURE 1: TOP LEVEL BLOCK DIAGRAM
7
1.1 Features and Benefits
* Fully DVB&DSS compliant
* Dual 6bit A/D converters
* Continuously variable symbol rate from 1Msps to 66Msps (90MHz clock)
* Internal digital root raised cosine filter
* Less than 0.5 dB implementation loss
* Frequency multiplexing capability
* Automated frequency search
* Internal bias cancellation
* Both wideband and narrowband AGC
* Noise calibration for antenna steering
* Output data rate as high as 82Mbps
* Fixed frequency sampling clock
* Simple interface with tuner and analog processing
* Microcontroller interface
* Eight bit parallel or I2C monitor and control interface
* I2C by-pass mode
* DiSeqC 1.2 interface support
* Dual Carrier Loop Filter
Part code
HDM8515P
8
Package
100PQFP
2. Hardware Specification
Table 1: Absolute Maximum Ratings
Rating
Ambient Temperature under Bias
Storage Temperature
Ambient Humidity under Bias
Thermal Resistance(Ja)
Junction Temperature
Voltage on Any Pin
VDD, IOVDD
Package Material
Value
-10 to 70
-65 to 150
85(85 c ,500hrs)
45
120
Vss - 0.3V to V DD + 0.5V
4.5
- Compound : CEL-4630SX
- Lead Frame : Copper
Unit
c
c
%
c/W
c
V
V
Table 2: DC Characteristics
Symbol
IDD
IOVDD
VDD
V
VIL
VIH
VOL
VOH
IIH
IIL
CIN
COUT
Parameter
Dynamic Current
Interface Power Supply
Voltage
Core Power Supply
Voltage
ADC Power Supply
Voltage
Input Low Voltage
Input High Voltage
Output Low Voltage
Output High Voltage
Input High Current
Input Low Current
Input Capacitance
Output Capacitance
Min.
3
Max.
390
3.6
Units
mA
V
2.3
2.7
V
Normal Operation
2.3
2.7
V
Normal Operation
0
0.7V DD
0.3V DD
VDD+
0.5
0.4
10
10
10
10
V
V
2.4
-10
-10
-
V
V
uA
uA
pF
pF
Test Conditions
VDD=2.7, Freq=90Mhz
Normal Operation
IOL = 4 mA
IOH = 4 mA
VIN =3.6, VDD =3.6
VDD = 3.6, VIN =0
Typical 5.75pF
Typical 5.97pF
9
Table 3: Demodulator Specifications
Parameter
Sampling Clock Frequency
Analog Input Full Scale Range
Symbol Rate
Viterbi Data Rate
Reed Solomon Data Rate
Implementation Loss
Symbol Rate Resolution
Carrier Frequency Resolution
Acquisition Sweep Range
Min.
1MHz
0.9 Vpp
1Msps
20
Clock/(2 )
20
Clock/(2 )
-
Max.
90MHz
1.1 Vpp
66Msps
90Mbps
82Mbps
0.5 dB
+ or - Clock/2
Table 4: AC Characteristics
Symbol
tsu1
Parameter
Input Data Setup before Clock
Min.
6
Max.
-
Unit
ns
th1
Input Data Hold after Clock
2
-
ns
tpw1
Low Pulse Width of Clock
8.7
-
ns
tpw2
High Pulse Width of Clock
8.1
-
ns
tpw1
CLOCK
tpw2
tsu1 th1
I_IN [5:0]
or Q_IN [5:0]
FIGURE 2: INPUT DATA TIMING DIAGRAM
10
Table 5: Intel 80C88A Read Cycle Timing Parameters (Busmode = 1)
Symbol
tsu1
Parameter
Input Address and /CE Setup before /RE Inactive
Min.
35
Max.
-
Unit
ns
5
-
ns
th1
Input Address and /CE Hold after /RE Inactive
tpw1
td1
/RE Low Duration
200
-
ns
Delay from /CE to DTACK Active
-
35
ns
tdoz1
Delay from /RE Inactive to DTACK in Tristate Mode
-
10
ns
tdoz2
Delay from /RE Inactive to HI_DATA [7:0] Tristate Mode
10
-
ns
Valid
HI_ADDR [4:0]
/CE
t pw1
/RE
td1
th1
Z
Z
DTACK
tdoz1
HI_DATA[7:0]
tsu1
tdoz2
FIGURE 3: INTEL 80C88A READ TIMING DIAGRAM
Note: HI_ADDR[4:0] is derived from the processor(80C88A) A15-A8 bus and HI _DATA[7:0] is
connected to the AD7 - AD0 bus.
#This page is only for HDM8515P.
11
Table 6: Intel 80C88A Write Cycle Timing Parameters (Busmode = 1)
Symbol
tsu1
th1
Parameter
Input Data Setup before /WE Inactive
Input Address, Data and /CE Hold after /WE Inactive
Min.
20
Max.
-
Unit
ns
8
-
ns
200
-
ns
tpw1
/WE Low Duration
td1
tdoz1
Delay from /CE to DTACK Active
-
35
ns
Delay from /WE Inactive to DTACK in Tristate Mode
-
15
ns
Valid
HI_ADDR [4:0]
/CE
tpw1
/WE
td1
th1
DTACK
tdoz1
HI_DATA[7:0]
tsu1
FIGURE 4: INTEL 80C88A W RITE TIMING DIAGRAM
Note: HI_ADDR[4:0] is derived from the processor(80C88A) A15-A8 bus and HI_DATA[7:0] is
connected to the AD7 - AD0 bus.
#This page is only for HDM8515P.
12
Table 7: Intel 8051 Read Cycle Timing Parameters (Busmode = 1)
Symbol
tsu1
th1
Parameter
Input Address Setup before /CE Active
Input Address and /CE Hold after /RE Inactive
tpw1
/RE Active Duration
tpd1
Delay from /RE Active to HI_DATA [7:0] Valid
tdoz1
Delay from /RE Inactive to HI_DATA[7:0] Tristate Mode
Min.
5
Max.
-
Unit
ns
5
-
ns
400
-
ns
-
40
ns
10
-
ns
Valid
HI_ADDR [4:0]
th1
tsu1
/CE
tpw1
/RE
HI_DATA[7:0]
tpd1
tdoz1
FIGURE 5: INTEL 8051 READ TIMING DIAGRAM
#This page is only for HDM8515P.
13
Table 8: Intel 8051 Write Cycle Timing Parameters (Busmode = 1)
Symbol
tsu1
Parameter
Input Address and Data Setup before /WE Active
Min.
5
Max.
-
Unit
ns
5
-
ns
400
-
ns
th1
Input Address and Data Hold after /WE Inactive
tpw1
tsu2
/WE Active Duration
/CE Setup to /WE Active
5
-
ns
th2
/CE Hold after /WE Inactive
5
-
ns
Valid
HI_ADDR [4:0]
tsu1
t h1
tsu2
th2
/CE
t pw1
/WE
Valid
HI_DATA[7:0]
FIGURE 6: INTEL 8051 WRITE TIMING DIAGRAM
#This page is only for HDM8515P.
14
Table 9: Motorola Read Cycle Timing Parameters (Busmode =0)
Symbol
tsu1
Parameter
Setup Time of R/W with respect to /CE Active
Min.
5
Max.
-
Unit
ns
tsu2
td1
Address Setup with respect to /DS Active
5
-
ns
Delay from DTACK Active to Data Valid
-
30
ns
th1
R/W Hold with respect to /DS Inactive
5
-
ns
th2
Address Hold with respect to /DS Inactive
5
-
ns
th3
Data Hold with respect to /DS Inactive
10
-
ns
HI_ADDR[4:0]
Valid
tsu2
th2
/CE
/DS
th1
tsu1
R/W
td1
DTACK
th3
HI_DATA[7:0]
FIGURE 7: MOTOROLA READ TIMING DIAGRAM
Note:
External pull-up resistor is required on DTACK.
#This page is only for HDM8515P.
15
Table 10: Motorola Write Cycle Timing Parameters (Busmode =0)
Symbol
tsu1
Parameter
Data Setup to /DS Active
Min.
5
Max.
-
Unit
ns
tsu2
td1
R/W Setup to /CS and Address
3
-
ns
/DS Delay from R/W
5
-
ns
td2
DTACK Delay from /DS Active
-
40
ns
td3
DTACK Delay from /DS Inactive
-
10
ns
tpw1
th1
/DS Active Duration
5
-
ns
Address, /CS and R/W Hold from /DS Inactive
5
-
ns
th2
Data Hold from /DS Inactive
5
-
ns
Valid
HI_ADDR[4:0]
/CS
tsu2
tpw1
/DS
td1
th1
R/W
td2
td3
DTACK
HI_DATA[7:0]
tsu1
th2
Valid
FIGURE 8: MOTOROLA W RITE TIMING DIAGRAM
Note:
External pull up resistor is required on DTACK.
#This page is only for HDM8515P.
16
Table 11: Output Timing
Symbol
tsu
thd
Parameter
Output Data Setup before DATA_CLK and DATA_STB
Min.
5
Max.
-
Unit
ns
10
-
ns
Output Data Hold after DATA_CLK and DATA_STB
t hd
t su
DATA_CLK
DATA_STB
FRAME_SYNC
DATA_VALID
n-3
DATA
n-2
n-1
n
xx
xx
xx
xx
xx
xx
xx
xx
xx
1
2
3
4
FIGURE 9: OUTPUT TIMING DIAGRAM FOR NORMAL PARALLEL
t hd
t su
DATA_CLK
DATA_STB
FRAME_SYNC
DATA_VALID
DATA[0]
8n-8
xx
xx
8n-7 8n-6 8n-5 8n-4
xx
8n-3 8n-2
8n-1
8n
xx
1
2
3
4
FIGURE 10: OUTPUT TIMING DIAGRAM FOR NORMAL SERIAL
NOTE : In case of DVB, n is 188
In case of DSS, n is 144
17
t hd
t su
DATA_CLK
DATA_STB
FRAME_SYNC
DATA_VALID
DATA
n-3 n-2 n-1 n
xx
xx xx
xx xx xx
xx xx xx
xx xx
xx xx xx
xx xx
1
2
3
4
FIGURE 11: OUTPUT TIMING DIAGRAM FOR REGULATED PARALLEL
t hd
t su
DATA_CLK
DATA_STB
FRAME_SYNC
DATA_VALID
DATA [0]
8n-8 xx xx 8n-7 8n-68n-5 8n-4 xx 8n-3 8n-28n-1 8n
xx
xx xx
xx xx xx
xx xx
FIGURE 12: OUTPUT TIMING DIAGRAM FOR REGULATED SERIAL
NOTE : In case of DVB, n is 188
In case of DSS, n is 144
18
1
2
3
4
3. Technical Overview
3.1 Dual Channel Analog to Digital Converter
The block diagram shown below illustrates internal configuration of the Dual Channel ADC.
Baseband signals, in-phase(I) and quadrature phase(Q), which are generated by down converters,
are applied to the dual channel ADC and quantized to 6-bit digital codes respectively. The ADC is
optimized to allow AC coupled inputs with full scale input range of 1V + or - 10%. An LSB weight is
approximately 15.6 mV.
The full scale input analog conversion range (Vpp) is determined by the voltages of VTOP and
VBOT and simply equal to (VTOP - VBOT). The full scale range is defined as the voltage range that
accommodates 63 codes of equally spaced LSBs. Also the ADC supplies its own reference
voltages for A/D conversions. The voltages can be monitored by external reference pins. The
VTOP, VBOT represent top and bottom reference voltages respectively. REF_I, REF_Q represent
middle reference voltages for each channel. All these 4 reference voltage pins should be by-passed
to GND via 0.1uF capacitors. The values of internally generated voltage of VTOP and VBOT are
2.0V and 1.0V respectively. Vpp can be adjusted by externally applying voltages to both VTOP
and VBOT pins respectively when different conversion ranges are necessary. VTOP can be
adjusted as high as 2.3V and VBOT can be as low as 0.5V. A larger input range can be
established by taking VTOP higher and VBOT lower than on-chip generated voltages.
To supply necessary bias voltages for AC coupled applications, REF_I and REF_Q, which are
middle reference voltages for I and Q channel, are connected to the analog input pins (AIN_I and
AIN_Q ) respectively through 40 kohm resistors, as shown in the block diagram. For DC coupled
applications, these voltages can be used to feed back offset compensation signals.
To insure optimum performance, a low impedance analog ground plane is recommended and
should be separated from other digital ground planes. The analog power supplies should be by passed at device to analog ground through 0.1uF ceramic capacitors.
19
VTOP
AIN_I
6-bit ADC
REF_I
DI
Ref.
Voltage
Gen.
CLOCK
AIN_Q
6
6-bit ADC
6
DQ
REF_Q
VBOT
FIGURE 13: ADC BLOCK DIAGRAM
20
3.2 Variable Rate Demodulator
The block diagram illustrates the overall configuration of the variable rate QPSK demodulator.
Baseband in-phase (I) and quadrature (Q) inputs are applied to the demodulator at a fixed sampling
rate. These digital samples are produced by A/D converters which employ AC coupling to minimize
DC offset.
Lock Detect
QPSK Lock
Signal
Strength
Symbol
Tracking
I_in
Q_in
First
Frequency
Trans.
Dual FIR
NB AGC
Second
Frequency
Trans.
I_out
Q_out
Carrier
Tracking
Frequency
Sweeper
FIGURE 14 DEMODULATOR BLOCK DIAGRAM
The only significant change to this configuration over the HDM8513A is the addition of the Second
Frequency Translator. The carrier tracking block produces two outputs, one is the frequency
correction which is provided to the First Frequency Translator. This insures that the input to the
Dual FIR is always centered at zero frequency error, although there may be a phase error at this
point. The second output of the Carrier Tracking function provides the phase correction to the
Second Frequency Translator.
The carrier frequency error associated with these samples is removed digitally during tracking
operations by a complex multiplier and a digitally controlled oscillator, sometimes called a
numerically controlled oscillator (NCO). During initial acquisition, coarse frequency error is
removed by a combination of the digital AGC within the HDM8515 and external analog tuning
circuits.
A Dual filter performs the root raised cosine filtering of the frequency corrected baseband samples.
This filter, which implements the function of equation (1), is always configured to have an impulse
response duration of 8 symbols regardless of the programmed symbol rate. For low symbol rates,
a large number of samples are used, while for high symbol rates a relatively low number of samples
are processed for each filter output. The outputs of the daul filters are applied to a digital
21
narrowband AGC which insures that the signal is optimally scaled to the Viterbi decoder to an
accuracy of + or - 0.5 dB to insure optimum FEC performance.
y[k] = Σ h[n] x[k-n]
(1)
In addition to optimizing performance of the Viterbi decoder, the digital narrowband AGC also
insures that the performance of the symbol timing and carrier tracking loops is independent of
signal level variations. An analog wideband AGC is also employed to insure that the analog signal
applied to the A/D converters is properly scaled.
Both the symbol timing and carrier tracking loops are implemented digitally, which eliminates the
need for external connections to analog tuning components during steady state operation. This
causes the requirements on the analog presampling filter to be relaxed, permitting a lower cost
analog front end. For systems which require a narrow band presampling filter, and have the
potential for significant frequency error in the LNB (several MHz) the HDM8515 provides a high
resolution measure of carrier frequency to permit periodic readjustment of the front end tuner
frequency to compensate for drift. The host processor periodically reads the frequency register,
then computes appropriate correction to the tuner frequency.
The nominal symbol rate and the nominal carrier frequency are programmed into the demodulator
to an accuracy provided by 20 bits of resolution, and the system accuracy is equivalent to that of
the fixed frequency sampling clock.
During initial acquisition, the HDM8515 provides an automated sweep program to facilitate carrier
acquisition. The host processor loads a 20 bit register which determines the initial carrier
frequency. A 16 bit register is programmed with the number of symbol times the receiver will dwell
at each frequency. If the receiver remains at the initial frequency for the programmed number of
symbol times without achieving lock, the carrier frequency is incremented by the step frequency
value programmed into another 16 bit register. If no lock is achieved, the receiver will continue to
increment the frequency until the maximum number of search frequencies, as determined by the
value in an 8 register, is achieved. When the max imum number of search frequencies is reached,
the carrier frequency returns to the initial value and the entire process is repeated. Once the host
processor determines that lock is achieved by observing the lock flag, it then inhibits the sweep
function and programs loop bandwidth parameters which are optimized for steady state
performance.
22
3.3 Noise Measurement Circuit
When the DBS system is being installed in any place, the most difficult part of the installation is
accurate pointing of the antenna toward the satellite. Inaccurate pointing results in loss of margin
and greater potential for outages in adverse weather conditions. Existing systems use information
from the demodulator forward error correction circuits to provide a measure of antenna pointing.
Unfortunately, this method is useful over a range of only several dB above system threshold.
The HDM8515 employs a unique circuit for accurate measure of signal strength over a 20 dB range
of signal to noise ratio. This method, illustrated in the block diagram, makes use of the fact that
the demodulator provides 8 bits of resolution for each of the quadrature output components. This
high resolution provides a means of measuring the noise component with great accuracy.
The eight bit in-phase demodulator filter output is detected by an absolute value circuit, then
passed through an IIR to provide a measure of average signal amplitude. Each sample is then
subtracted from this average amplitude to provide an instantaneous noise sample. The absolute
value of these noise samples are then averaged by a second IIR to provide a measure of the noise
which is roughly proportional to the noise power and inversely proportional to signal to noise ratio.
Finally, the Figure 16 illustrates the results of simulations under different noise conditions. This
figure illustrates that for signal-to-noise ratio as high as 19 dB, the noise measurement circuit
provides a meaningful measure of signal power with worst case resolution of 1 dB.
In-Phase
Component
8
Absolute
Value
R
8
Average
Magnitude
255
256
16
8
Average
Deviation
Absolute
Value
R
8
Instantaneous
Deviation
255
256
16
FIGURE 15: N OISE MEASUREMENT CIRCUIT
23
FIGURE 16: NOISE A CCUMULATOR AS A FUNCTION OF SNR AND TIME
24
3.4 Viterbi Decoder
The Viterbi decoder accepts 4 bit soft decision samples of the in-phase (I) and quadrature (Q)
components of the received signal. Once QPSK lock has been achieved, the decoder searches for
the correct code rate, starting with rate 3/4, then proceeding to rate 2/3, 5/6, 7/8 and finally rate
1/2. Each of the possible synchronization phases at each rate is tested as well as the two
possible carrier phase ambiguity conditions. Polarity reversal is corrected in the word
synchronization logic. Viterbi lock is achieved when the trellis traceback algorithm converges, on
the average, within a prescribed number of symbols.
Although the algorithm automatically tests for carrier phase ambiguity, there is no provision to
automatically correct for phase reversal. Phase reversal can occur if the receiver chain, consisting
of an LNB and the tuner, provides an odd number of high side frequency translation operations. A
system may be required to operate with different LNBs, some of which provide phase reversal.
This condition may be corrected by the host processor, which can set a bit in the down converter
to correct for phase reversal.
The Viterbi decoder employs the radix two algorithm. The output buffer reserializes the data which
is made available, along with the Viterbi data clock as external signals. These signals permit
verification of the DVB specification which is referenced to the Viterbi decoder output.
ACS Array
64
Trace-back
RAM
Traceback Memory
Controller
I
Q
4
4
G1
Depuncturing
Logic
G2
Branch Metric
Calculator
Decoder
Quality
Estimate
Change Puncture Phase
Last-In
First-Out
Buffer
Clock
Out
Data
Out
Viterbi
Lock
Change Carrier Phase
FIGURE 17: VITERBI DECODER
25
3.5 Autonomous Acquisition
The HDM8515 provides several features to permit signal acquistion with minimal interaction with
the host microcontroller. The host microcontroller must configure the HDM8515 for a specific
symbol rate, carrier frequency, carrier sweep conditions, and tracking loop bandwidth. The
microcontroller also must monitor lock status to determine when acquisition is achieved. There are
many provisions in the HDM8515 to enable the system designer to implement custom algorithms
for specific requirements.
The microcontroller first must set the lower edge of the carrier search range in the Carrier
Frequency registers (04, 05 and 06). Then the processor configures the Carrier Sweep Step Size
register (09, 0A) to a value which is less than two times the carrier pull-in range. The number of
symbols per dwell is defined in registers (0B,0C), and is typically set to a value of 500 to 1000.
The total search range is set by the Number of Search Frequencies as defined in register 0D. The
total sweep frequency range is this number times the Carrier Sweep Step Size. The sweep
process stops once QPSK carrier lock is detected. If no lock is detected, the sweep process
continuously repeats.
The QPSK demodulator may lock to any one of four different phase reference states, only one of
which produces true I and Q data as it was modulated at the transmitter. If the local phase
reference is plus 90 degrees or minus 90 degrees with respect to the true phase, the information
provided to the Viterbi decoder will be unintelligible. If the Viterbi decoder is unable to achieve valid
lock, it will reattempt lock with a 90 degree phase shift, without external intervention.
In the event that the local phase is 180 degrees from the true phase, the data provided to the
Viterbi decoder will be inverted, but otherwise valid. The code employed by the Viterbi decoder is
transparent, thus the data from the Viterbi decoder will be inverted if the input is inverted. This
situation is corrected in the word synchronization circuit. This circuit searches for the
unscrambled sync word which occurs once per frame (every 204 bytes at the Viterbi output). Once
correlation with the sync word is found, the data is reformatted as a series of bytes with the
beginning of each 204 byte frame identified to provide the synchronization information required for
the deinterleaver and the Reed Solomon decoder. If the polarity of the sync word is incorrect, the
data is inverted before further processing without external interaction.
The HDM8515 supports five different code rates, including 1/2, 2/3, 3/4, 5/6 and 7/8. When rate
1/2 is employed, there is a one-to-one correspondence between incoming I and Q samples and G1
and G2 terms required by the Viterbi decoder. The higher rates employ punctured coding
techniques which periodically cause either a G1 or G2 term to be deleted. The puncturing pattern
can have 6 possible ambiguity states for rate 2/3, 4 states for rate 3/4, 6 states for rate 5/6 and 8
states for rate 7/8. As part of the Viterbi decoding acquisition process, each puncturing state of
each code must be tested. Total acquisition requires search of 26 different conditions. The
process starts with rate 3/4 coding and proceeds sequentially to rate 2/3, 5/6, 7/8, and finally rate
1/2.
In some systems, it may be possible to experience spectral inversion. This might occur when
different combinations of LNBs and tuners are employed which implement different frequency
translation schemes. Correction of spectral inversion must be corrected with host processor
interaction. If the host processor detects that QPSK lock is achieved, but Viterbi lock has not
occurred within a specified time, then a bit must be set in the demodulator which reverses the
spectrum.
26
The table below illustrates a typical acquisition timing. For this example, the symbol rate is one
half of the clock rate. The code rate is set to 5/6, which requires 13 trial and errors before node
sync is achieved. The carrier search logic requires 10 dwells at different frequencies (500 symbols
per dwell) before demodulator lock is achieved.
Table 12: Example of Acquisition Timing
Carrier Search
Viterbi Node Sync
Byte Sync
Deinterleaver Flush
Reed Solomon
Total Timing
Bit Times
8,333
2,652
16,000
19,584
1,632
48,201
Symbols
5,000
1,591
9,600
11,750
979
26,950
Clock Cycles
10,000
3,182
19,200
23,500
1,958
57,840
The total time required for acquisition could vary widely, depending upon the carrier search range
and the time required for Viterbi node sync. For this example, however, the Byte Sync time and
the time required to flush the deinterleaver dominates the total time. If a 90MHz clock were
employed, the total acquisition time would be 0.642 milliseconds for this example
27
3.6 Reed Solomon Decoder
The serial output from the Viterbi is provided to the Word Sync circuits which searches for the eight
bit frame sync word which occurs every 204 bytes. By detecting the polarity of the sync word, this
module can correct polarity reversals in the data provided by the Viterbi decoder.
Byte serial data is provided to the convolutional deinterleaver, which reorders the received symbols.
This process causes errors, which typically occur in bursts from the Viterbi decoder, to be
distributed randomly over many blocks. This deinterleaved data is then provided to the Reed
Solomon decoder which can reduce an error rate of 2 x10-4 from the Viterbi decoder to less than 1
in 10-10. The Reed Solomon decoder accepts input data in blocks of 204 bytes and produces error
corrected blocks of 188 bytes. Maximum 8 bytes per a RS block can be corrected in RS decoder.
Reedsolomon block includes on-chip BER calculator at the output of Viterbi to monitor signal
quality or estimate the SNR of incoming signal. The calculated value can be read by accessing two
read registers via utility bus such as I2C. It represents the number of errors among 220 data bits.
The next process is descrambling, not to be confused with the descrambling which is part of
conditional access. The purpose of scrambling the transmitted data and performing the inverse in
the receiver is to insure that the spectrum of the transmitted waveform is always evenly distributed
without significant discrete spectral lines. Without the scrambling/descrambling process, a
transmitted sequence of all ones or all zeroes would result in strong spectral components and
could interfere with other signals in the same satellite transponder.
The final process is data regulation. Viterbi Data and Viterbi Clock occur irregularly according to
the code rate. Data clock regulation makes it possible to interface with external common interface
devices. To make external bus interface more flexible, interface mode such as parallel or serial can
be selected by mode selection register.
Parameter
Regulate_data_clk
Mode_serial
Clk_pol
Register
Bit 5 of 14H register
Bit 0 of 18H register
Bit 7 of 14H register
l
NORMAL INTERFACE MODE (parallel/serial)
If regulate_data_clk is reset, both parallel interface and serial interface work in normal
operation which is same as HDM8513A. Parallel interface or serial interface can be alternated
by modifying mode_serial bit (Refer to Figure 9 and Figure 10)
l
REGULATED INTERFACE MODE (parallel/serial)
If regulate_data_clk is set,all interfaces are from internal FIF O designed to regulate irregular
interface signals. Data clock cycle is a little bit faster than the average of cycle of irregular data
clock, so meaningless data can be output in invalid data period. Parallel interface or serial
interface can be alternated by modifying mode_serial bit (Refer to Figure 11 and Figure 12)
l
CLOCK POLARITY
This bit is used to select the DATA_CLK polarity either for serial or parallel transport interface.
If this bit is set to zero(default value), the transport data and control signals are latched at the
positive edge of DATA_CLK. Otherwise, the signals are latched at the negative edge of
DATA_CLK.
28
Error Flag
Deinterleaver
Memory
Viterbi
Data
Viterbi
Clock
8
Word
Sync.
8
Deinterleaver
Memory
Control
8
Reed
Solomon
Decoder
Word Clock
Frame Clock
8
Descrambler
Data
Out
Sync.
Data Clock
FIGURE 18: REED SOLOMON DECODER
29
3.7 Clock Generation PLL
An integrated VCO is locked to MxN times a reference frequency provided by a external clock.
1.Determining Output Frequency
Fully programmable feedback and reference divider capability allows virtually any frequency
to be generated, not just simple multiples of reference frequency.
There are two status exist
(1) PLL Disable mode : The PLL is bypassed and the external clock is directly connected to the
Internal clock.
(2) PLL Enable mode : The internal clock is connected to the generated clock of the PLL.
1.1 PLL disable mode
PLL control setting is as follows
TDM (Bit 7 of 0x23 register) is set to one and BYPASS (Bit 4 of 0x23 register) is set to one.
1.2 Normal Frequency mode
Output frequency range is limited to 160MHz.
PLL control setting is as follows:
TDM (Bit 7 of 0x23 register) is set to zero, and BYPASS (Bit 4 of 0x23 register) is set to
zero.
At this condition, the output frequency, F(ck), is actually determined by the following
equation.
F(ref) x (Feedback divisor)
F(ck) = ---------------------------------------------(Reference divisor)
F(ck) : Frequency of output
F(ref): Frequency of reference input
Feedback divisor : M[13:0]+2 , (0x25 and 0x26 registers)
Reference divisor : N[7:0]+1 , (0x27 register)
1.3 Extended Frequency mode
Output frequency range is limited to 320MHz.
PLL control setting is as follows
TDM (Bit 7 of 0x23 register) is set to zero, and BYPASS (Bit 4 fo 0x23 register) is set to
one.
At this condition, the output frequency, F(ck), is actually determined by the following
Equation
F(ref) x (Feedback divisor) x (Pre divisor)
F(ck) = -------------------------------------------------------(Reference divisor)
Pre divisor : 2p[1:0]+1 , P is Bit 2 and 3 of 0x23 register
2. PLL Control Parameter setting
Besides of M (Feedback divisor), N (Reference divisor), P (Pre divisor) , You must determine vc
(VCO range control vector), lfm (Loop filter mode selector), icp (Charge pump bias current
control
30
vector) values appropriately.
2.1vc value setting
According to Output clock frequency, determine the vc values.
31
vc[1:0]
p[1:0]
00
00
01
10
11
00
01
10
11
00
01
10
11
00
01
10
11
01
10
11
Output Clock Frequency
min
max
40MHz
100MHz
20MHz
50MHz
10MHz
25MHz
5MHz
12.5MHz
60MHz
100MHz
30MHz
50MHz
15MHz
25MHz
7.5MHz
12.5MHz
80MHz
100MHz
40MHz
50MHz
20MHz
25MHz
10MHz
12.5MHz
100MHz
100MHz
50MHz
50MHz
25MHz
25MHz
12.5MHz
12.5MHz
2.2 lfm value setting
According to the table, determine the lfm value
lfm
7
0
1
2
3
4
5
6
(Reference Frequency)/(Reference Divisor)/15
Less than 0.01555
Less than 0.0258
Less than 0.0421
Less than 0.070
Less than 0.114
Less than 0.187
Less than 0.309
Greater than 0.309
2.3 icp value setting
According to the lfm value, you determine zero value, pole value, rlf value.
lfm
7
0
1
2
3
4
5
6
Zero value
External filter used
0.008
0.013
0.021
0.032
0.060
0.100
0.160
Pole value
External filter used
0.03
0.050
0.082
0.135
0.221
0.360
0.600
Step 1: According to the following formula, Kvcop is determined
32
rlf value
External filter used
40
24.1
40
24.1
24.1
14.7
8.9
100
Kvcop = ---------------2p
P: Pre Divisor
33
Step 2: According to the following formula, Kpll is determined
Kpll = (Zero value * Pole value)1/2
Step 3: According to the following formula, Kpd is determined
Kpd = 1000.0 * Kpll * Feedback Divisor / Kvcop / rlf
Step 4: According to the following formula, temp value is determined
Temp value = 2.0 * 3.14 * Kpd
Step 5: Finally, According to the following formula, icp is determined
icp = 16.5 – 16.0 / 40.0 * temp value
If icp value has fraction, truncate it.
34
3.8 DBS Receiver
The HDM8515 DVB Demodulator including a dual A/D converter and the MPEG-2 decoder provide
the core digital processing technology for a DBS receiver conforming with the DVB standard.
DRAM
Inte rfa c e
4 80 MHz
Loo p
Filter
Fixed F requency P LL Control
Da ta
8
I
L-B and
Tuner
4 80 MHz
IF
Do wnconverter
Q
HDM8515
8
Conditional
Access
Clock
Video
MPEG-2
Demultiplexer
Audio
BSFC77GV6
SL171 0
AGC 2
L ow
Pa ss
Filter
AGC 1
3
WB
AGC
Co arse Tuning Step Frequ ency Control
M C68306
(M C68340)
Host Proce ssor
Serial
Interfa ce
FIGURE 19: TYPICAL SET TOP BOX DEMODULATOR
A tuner accepts an L-band RF input from the antenna/LNB assembly located outside the building.
A host processor controls the tuner to the nominal center frequency of the target signal. Baseband
I and Q outputs from the downconverter are applied to an A/D converter pair which is sampled at a
fixed rate, 90MHz as illustrated in this example. The tuner is required to filter the received
baseband signal to a bandwidth less than half the sampling rate, but is not required to perform
matched filtering.
Once the HDM8515 has locked to the target signal, the host processor may read the internal
registers to determine the steady state frequency error. This error would be used to make period
corrections to the programmed frequency of the tuner PLL.
The HDM8515 provides an output which can be used to control the analog AGC in the tuner. This
digital signal must be filtered and amplified before applying it to the AGC control element. When
the loop is closed, the signal applied to the A/D converters is optimally scaled.
35
3.9 DiSEqC Interface
The DiSEqC system is a communication bus between satellite receivers and satellite peripheral
equipment, using only the existing coaxial cable.
1.1 DiSEqC mode
According to the value of DiSEqC_mode of 0x31 register, DiSEqC mode can be changed
0: 22KHz tone off
1: 22KHz tone on
2: Burst mode - on for 12.5ms =’0’
3: Burst mode - modulated 1:2 for 12.5ms =’1’
4: Modulated with bytes from DiSEqC instruction
1.2 DiSEqC instruction
Up to eight instruction data bytes are loaded into a bank of registers(0x29 -0x30). I2C
automatic
register address incrementing is turn on. The number of bytes in the DiSEqC instruction must
be defined in the DiSEqC_length of 0x31 register.
When the DiSEqC instruction data bytes have been loaded, set DiSEqC_mode of 0x31 register.
At the same time, program DiSEqC_length of 0x31 register. The instruction data is modulated
onto
22KHz signal and output from the DISEQC pin.
36
4. Mechanical Specifications
4.1 100 Pin Quad Flat Pack
4.1.1 Pin Assignment
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
DATA_CLK
FRAME_ERROR
FRAME_SYNC
VDD
VSS
DISEQC
SIGMADELTA
SYMBOL_CLOCK
WB_AGC
QPSK_LOCK
IOVDD
IOVSS
TEST15
TEST14
TEST13
TEST12
VDD
VSS
TEST11
TEST10
TEST9
TEST8
IOVDD
IOVSS
TEST7
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
TEST6
VDDA
VSSA
VTOP
AIN_I
IOVDDA
IOVSSA
REF_I
REF_Q
AIN_Q
VBOT
TEST5
TEST4
TEST3
VDDP
VSSP
TEST2
TEST1
TEST0
CLOCK
XTAL1_IN
XTAL1_OUT
IOVDD
IOVSS
RESET
51
HI_ADDR5
76
52
HI_ADDR4
77
53
HI_ADDR3
78
54
HI_ADDR2
79
55
HI_ADDR1
80
56
HI_A DDR0
81
57
VDD
82
58
HI_DATA7
83
59
HI_DATA6
84
60
HI_DATA5
85
61
HI_DATA4
86
62
IOVDD
87
63
IOVSS
88
64
HI_DATA3
89
65
HI_DATA2
90
66
HI_DATA1
91
67
HI_DATA0
92
68
VDD
93
69
VSS
94
70 VB_NODESYNC 95
71
VB_CLOCK
96
72
VB_DATA
97
73
VDD
98
74
VSS
99
75
BUSMODE
100
DTACK
SDA_I2CO
SDA_I2C
SCL_I2CO
SCL_I2C
VDD5
VSS
R/W(/RE)
/CE
/DS(/WE)
VDD
VSS
DATA7
DATA6
DATA5
LOCK
DATA4
DATA3
IOVDD
IOVSS
DATA2
DATA1
DATA0
DATA_VALID
DATA_STB
37
4.1.2 Package Dimensions
23.340
23.090
20.100
19.900
80
51
50
81
17.880
17.908
HDM8515
14.100
13.900
DVB Demodulator
100
31
1
30
0.380
0.220
0.650 Typ.
All Dimensions in mm
3.350
3.000
0.230
0.130
0.7
0.950
0.650
0.500
0.250
1.950 Typ.
FIGURE 20: M ECHANICAL CONFIGURATION
38
4.2 64 Pin Thin Quad Flat Pack
4.2.1 Pin Assignment (will be changed)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
FRAME_ERROR
FRAME_SYNC
LNB_SYNC
WB_AGC
IOVDD
IOVSS
TEST13
TEST12
VDD
VSS
TEST11
TEST10
TEST9
TEST8
VTOP
AIN_I
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
VSSA
VDDA
REF_I
REF_Q
AIN_Q
VBOT
VSSA
VDD
VSS
N/C
N/C
N/C
XTAL1
IOVDD
IOVSS
RESET
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
VDD
VSS
I2C_ADD2
IOVDD
IOVSS
I2C_ADD1
I2C_ADD0
VDD
VSS
VB_CLOCK
VB_DATA
BUSMODE
SDA_I2C0
SDA_I2C
SCL_I2C0
SCL_I2C
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
IOVSS
VDD
VSS
DATA7
DATA6
DATA5
LOCK
DATA4
DATA3
IOVDD
DATA2
DATA1
DATA0
DATA_VALID
DATA_STB
DATA_CLK
39
4.2.2 Package Dimensions
12.00
10.00
48
33
32
49
HDM8515T
10.00 12.00
0.17 Min.
17
64
0.27 Max.
1
16
0.50
All Dimensions in mm
0 Min.
0.95 Min.
1.00 Typ.
1.05 Max.
0-7
0.15 Max.
0.45 Min.
0.08R Min.
1.00 Ref.
FIGURE 21: M ECHANICAL CONFIGURATION
40
4.3 Recommended Analog Pin Connection
1.2uH
VDD
VDD
Electrolyte
Capacitor
47uF
0.1uF
0.1uF
I
0.1uF
Down
AIN_I
Converter
VTO
VBO
AIN_Q
Q
REF_Q
REF_I
0.1uF
0.1uF
VSS
0.1uF
VSS
0.1uF
HDM8515
0.1uF
AGND
FIGURE 22: A NALOG PIN CONNECTION
4.4 Recommended Clock Generation Circuit
4MHz Crystal
to XTAL1_IN
22pF
to XTAL1_OUT
22pF
FIGURE 23: CLOCK GENERATION CIRCUIT
41
5. Signal Description
5.1 Inputs
XTAL1
XTAL1 can be configured either for sampling clock input or PLL
reference clock input . The sampling clock rate must be a minimum of
1.33 times the symbol rate of the signal to be processed and at least
equal to the total bandwidth of the signal to be processed.
RESET
A low on this signal causes the chip to be initialized. I/O registers are
not cleared by this signal. This signal is asynchronous with respect to
the clock.
AIN_I
Analog Input Signal for I channel. This should be AC coupled with
Analog Input Source via 0.1uF capacitor.
AIN_Q
Analog Input Signal for I channel. This should be AC coupled with
Analog Input Source via a 0.1uF capacitor.
5.2 Outputs
VTOP
Top Reference Voltage Output of about 2.0V. It should be bypassed to
GND by 0.1uF capacitor. External bias voltage can be applied if
necessary.
VBOT
Bottom Reference Voltage Output of 1.0V. It should be bypassed to
GND by a 0.1uF capacitor. External bias voltage can be applied if
necessary.
REF_I
Middle Reference Voltage for I Channel. It should be bypassed to GND
by a 0.1uF capacitor.
REF_Q
Middle Reference Voltage for Q Channel. It should be bypassed to
GND by a 0.1uF capacitor.
DATA [7:0]
The eight bit output data is provided in parallel format to be handed to
an MPEG decoder for video and audio decompression.
42
DATA_CLK
The DATA_CLK is used to latch data and control signal of transport
stream. The data and control signals can be programmed to be
latched either at positive or negative edge of DATA_CLK. This signal is
used in conjunction with DATA_VALID to transfer data from the
HDM8515. The DATA_CLK will continue to toggle during the 16
bytes that the DATA_VALID signal indicates that no data is available
(see figure 9 and 10).
DATA_VALID
When this signal is true, data is valid. This signal is not true during
the time the 16 bytes of redundancy information is transmitted for the
Reed Solomon decoder.
FRAME_SYNC
This signal is true at the first byte of a block of 188/144 bytes.
DATA_STB
This signal is used to transfer data from the HDM8515 to an MPEG
decoder. This signal goes from low to high when a new byte of a 188
/144byte MPEG2 data stream block is available. This signal is
inactive during the time the 16 redundancy bytes are transferred.
FRAME_ERROR
This signal goes true when the Reed Solomon decoder detects that an
uncorrectable number of errors have occurred. The error flag in the
MPEG2 output stream is also set when this flag goes high.
WB_AGC
This one bit output provides a measure of the external analog gain
required for optimizing the signal applied to the analog to digital
converters. This signal must be filtered, then applied to the analog
gain control.
CLOCK
This is a buffered clock output signal which may be used to drive other
devices with the same clock which drives the HDM8515.
QPSK_LOCK
This signal goes true when the QPSK demodulator has achieved
phase lock.
VB_NODESYNC
This signal goes true when the Viterbi decoder has achieved node
synchronization.
LOCK
This signal goes true when the output data is valid and all
synchronization functions have been performed.
SYMBOL_CLOCK
This signal, used for test purposes, goes true for a duration of one
clock cycle for each received symbol. For symbol rates equal or
greater than half the clock frequency, this signal at times may remain
high for two successive clock cycles to indicate that two symbols have
occurred.
VB_DATA
The serial output of the Viterbi Decoder is provided on this pin. The
information rate at this point is less than the rate of the input clock
( less than 60Mbps if a 60MHz clock is employed). As long as valid
convolutional encoding is employed, there is no constraint that the
input signal adheres to MPEG2 format. This data is tapped priod to
the polarity correction circuitry, so the data at this point may be
inverted.
43
VB_CLOCK
The positive edge of this signal indicates that VB_DATA is valid.
SIGMADELTA
This is an one bit Sigma Delta D/A converter which has 8 bits of
resolution. This output must be filtered with an analog low pass filter
off the chip. This output may be used for any external analog control.
DISEQC
This is a DiSEqC output to control the LNB.
TEST[15:0]
The data provided on the test output signals is defined by data value of
register 14H. Refer to register 14H.
44
5.3 Monitor and Control Interface
Three different modes are supported for the monitor and control interface. Two of the modes are 8
bit parallel interfaces, one which supports Intel microcontrollers and the other intended for Motorola
microcontrollers. The third mode is a serial interface conforming to the I2C standard.
The I2C mode is activated by placing BUSMODE high at the same time both /RE and /WE are low
simultaneously. When this mode is active, the seven bit I2C slave address of the HDM8515 is
configured by the seven least significant bits of the HI_DATA[7:0] bus.
HI_DATA [7:0]
This bi-directional data bus is used for transferring control parameters to
the demodulator and for reading the status registers within the
demodulator.
/CE
Chip enable is an active low input to the demodulator which signifies that
the other control signals are active.
/RE
Read Enable is an active low input to the device which, when active at the
same time chip enable is true, permits the device to drive the HI_DATA
[7:0] lines. When BUSMODE i s 0 (Motorola), this pin is read / not write
(see timing diagrams).
/WE
Write enable is an active low input to the device which, when true at the
same time chip enable is true, causes input data on the HI_DATA [7:0] bus
to be transferred to the register defined by the HI_ADDR [4:0] bus. When
BUSMODE is 0 (Motorola), this pin is not data strobe (see timing
diagrams).
HI_ADDR [4:0]
The address bus defines which location within the device is to be accessed
during a read or write operation.
BUSMODE
BUSMODE selects the type of microcontroller/processor used to setup the
chip. When high, an Intel processor/microcontroller interface is used.
When low, a Motorola processor interface is used.
DTACK
Data Acknowledge/Data Ready is a tristate output signal which informs
the controlling processor that a data transfer has been acknowledged by
the HDM8515.
SCL_I2C
This pin provides the clock for the I2C interface when that mode is active.
SDA_I2C
This pin is the data for the I2C interface and requires an external pull-up
resistor as per the I2C standard.
SDA_I2CO
This pin, which can be by-passed, is the data for the I2C interface.
SCL_I2CO
This pin, which can be by-passed, provides the clock for the I2C interface.
45
5.4 I2C Mode
The HDM8515 utilizes the subaddress technique when the I2C mode is employed. In all cases,
the HDM8515 behaves as the slave device (transmitter or receiver), whilst the host behaves as the
master device. The seven bit slave address of the HDM8515 is user selectable, being defined by
the inputs to HI_DATA[6:0] when the HDM8515 is in I2C mode.
Further information on the I2C bus formats and protocols is contained in the Philips
Semiconductors I2C specification.
In a 100pin configuration, SDA_I2CO and SCL_I2CO are added to provide a by-passing function.
When I2C bypass bit is set to zero, SDA_I2CO and SCL_I2CO are disabled.
5.4.1 I2C Write to HDM8515
The master initiates communication with the HDM8515 by generating a start condition and then
sending the HDM8515 the slave address defined by the seven bit hardwired address on HI_DATA
[6:0]. Per I2C convention, the eighth bit in the address byte is a read/not write bit, and should be
set to zero. The HDM8515 will acknowledge the correctly sent slave address, following which the
master sends an eight bit word address; this is the address of the first HDM8515 register to be
written to. Once the word address has been acknowledged by the HDM8515, the master can then
transmit the byte to be written to the word address. Once this byte is acknowledged by the
HDM8515, the word address is automatically incremented and further data bytes may be
transmitted by the master as necessary; one transmission may therefore contain a number of
bytes of data to be written to a sequential set of addresses (dummy bytes should be written to
addresses not defined in the HDM8515 register set to continue this process). The process is
terminated by the master generating a stop condition. Figure 24 depicts this protocol.
acknowledgement
from slave
S
7 Bits
SLAVE ADDRESS 0 A
acknowledgement
from slave
acknowledgement
from slave
8 Bits
WORD ADDRESS A
DATA BYTE
A
P
repeat if
necessary
R/W
S - Start Condition
A - Acknowledge
P - Stop Condition
auto increment
memory word address
FIGURE 24: I2C W RITE TO THE HDM8515
5.4.2 I2C Read from the HDM8515
To read information from the HDM8515, the master must first write the desired word address.
Hence the master must first generate a start condition and transmit the seven bit HDM8515 slave
address defined on HI_DATA[6:0], with the eighth bit (read/not write) set to zero. Once this has
been acknowledged by the HDM8515, the master transmits the first word address from which it
wishes to read information. The master must then generate a second start condition and
46
retransmit the HDM8515 slave address, this time with the read/not write bit set to one (read). This
will be acknowledged by the HDM8515, which then assumes the role of slave transmitter and
transmits the requested byte. This byte should be acknowledged by the master receiver. If no
stop condition is generated by the master, the HDM8515 will increment its word address pointer
and transmit the next byte of information. This process is detailed in Figure 25.
acknowledgement
from slave
S
7 Bits
SLAVE ADDRESS 0 A
acknowledgement
acknowledgement
from slave
from slave
8 Bits
WORD ADDRESS
acknowledgement
from master
7 Bits
A S
SLAVE ADDRESS 1 A
R/W
DATA BYTE
A
R/W
auto increment
memory word address
HDM8515 becomes
slave transmitter
acknowledgement
from master
S: Start Condition
A: Acknowledge
P: Stop Condition
DATA BYTE
A
P
last byte
auto increment
memory word address
FIGURE 25: I2C READ FROM THE HDM8515
Table 13: I2C Slave Address
I2C_ADD0
0
1
0
1
0
1
0
1
I2C_ADD1
0
0
1
1
0
0
1
1
I2C_ADD2
0
0
0
0
1
1
1
1
I2C Address
0000000
0000011
0001100
0001111
0110000
0110011
0111100
0111111
47
6. Register Definitions
6.1 Write Registers
ADDRESS (Hex)
00, 01, 02
Symbol Timing Frequency
The 20 bit straight binary number in this field establishes the symbol
timing frequency utilized within the demodulator. Bit 7 of address 00 is
the MSB and bit 4 of address 02 is the LSB. If Rs is the symbol rate
and fc is the clock frequency, the value to be stored in this 20 bit field is
the integer portion of Rs (220)/fc.
03
Symbol Timing Loop Gain Control
This field establishes the K1 and K2 gain values for the second order
loop filter of the symbol tracking loop. Bits 0,1,2 and 3 determine the
straight-through gain, and bits 4,5,6 and 7 determine the integration
path gain. The nominal value of this parameter in Hex, is expressed
below for different ranges of symbol rate to clock rate ratios:
Clock/Symbol Rate
2
4
8
16
32
64
04, 05, 06
Value
FB
DA
B9
98
77
56
Carrier Frequency
The 20 bit, two's complement number in this field establishes the
nominal carrier frequency of the demodulator. Bit 7 of address 04 is the
MSB and bit 4 of address 06 is the LSB. The number in this 20 bit field
multiplied by the clock frequency divided by 220 is the carrier frequency
in Hertz. When the carrier sweep function is active, this value defines
the starting frequency.
48
07, 08
Carrier Loop Filter Control
This field establishes the K1 and K2 gain values for the second order loop
filter of the carrier tracking loop. Bits 0,1,2 and 3 determine the straightthrough gain, and bits 4,5,6 and 7 determine the integration path gain.
The nominal value of this parameter in Hex, is expressed below for
different ranges of symbol rate to clock rate ratios. Two loop filter
configurations are provided at each symbol rate, one for steady state
operation(08) and one which is used only for acquisition(07) to permit
greater frequency pull-in. Initially the gains are set to acquisition values.
When QPSK_LOCK is achieved, they are automatically switched to
steady state values.
Clock/Symbol Rate Steady State Acqu.
2
4
8
16
32
64
09, 0A
C7
A7
87
67
47
27
C7
A7
87
67
47
27
Carrier Sweep Step Size
This 16 bit value defines the size of the step of each carrier frequency
dwell. Bit 7 of address 09 is the MSB and bit 0 of address 0A is the
LSB. The number in this register is divided by 216, and multiplied by the
clock frequency to determine the frequency step increment.
0B, 0C
Symbols Per Dwell
This 16 bit value defines the time, in symbol periods, for which the
demodulator will dwell before making the next frequency step in a sweep.
Bit 7 of address 0B is the MSB and bit 0 of address 0C is the LSB.
0D
Number of Search Frequencies
This 8 bit field determines the number of frequency steps which occur
during the frequency sweeping process. Combined with the frequency
step size, this determines the frequency span of the carrier sweep.
0E
Narrow Band AGC initial value
This 8 bit field establishes the initial gain of the narrow band AGC. High
numbers correspond to low gain associated with low symbol rates. If the
narrowband AGC function is enabled, this number is used as a starting
point and the closed loop will seek the optimum setting without
processor interaction.
49
0F
Control Parameters
Bit 0. Binary/Two’s Complement
When this bit is a zero, the system expects the six bit modulation input
samples in two’s complement format, otherwise the input should be in
offset binary format.
Bit 1. Spectrum Invert
When this bit is set to zero, the spectrum of the received signal is
inverted. This has the effect of complementing the in-phase channel
only.
Bit 2. Bias Cancel Enable
When this bit is a one, the internal circuit which cancels DC bias on the I
and Q inputs is enabled. When this function is enabled, it is assumed
that the input signal is scrambled with no significant DC component on
either the I or Q.
Bit 3. Symbol Track Enable
When this bit is set to one, the symbol tracking function is enabled.
When this bit is zero the symbol tracking frequency is forced to the
nominal 20 bit programmed value.
Bit 4. Carrier Track Enable
When this bit is set to one, the carrier phase tracking function is
enabled. When this bit is zero, the carrier frequency is forced to the 20
bit programmed value.
Bit 5. Sweep Hold
When this bit is set to one, the sweeping process is inhibited, and the
nominal carrier frequency remains at the last value.
Bit 6. Narrowband AGC Mode 1 Enable
When this bit is set to one and the narrowband AGC is in Mode 1, the
narrowband AGC self-adjusts to the optimum gain setting. When the bit
is set to zero, the most recent value is held without updating.
Bit 7. Automatic Detection of Spectrum Inversion
When this bit is set to one, the spectrum inversion is detected
automatically.
50
10
Reset Functions
Bit 0. Symbol Timing Frequency Accumulator
When this bit is set to zero, the frequency accumulator in the symbol
tracking loop is cleared to zero. This bit must be set to one in normal
tracking operation to implement a second order tracking loop, otherwise
the loop is first order.
Bit 1. Carrier Phase Tracking Frequency Accumulator
When this bit is set to zero, the frequency accumulator in the carrier
phase tracking loop is cleared to zero. This bit must be set to one in
tracking operation to implement a second order loop filter otherwise the
loop is first order.
Bit 2. Wideband AGC Accumulator
When this bit is set to zero, the accumulator in the wideband AGC is
cleared to zero. In normal operation, this bit is set to one. When the
wideband AGC is set to Mode 1, this bit has no effect as the integrator
must be implemented in the external analog circuits.
Bit 3. Narrowband AGC Accumulator
When this bit is set to zero, the accumulator in the narrowband AGC is
cleared to the initial value defined in location 0E. In normal operation,
this bit is set to one.
Bit 4. Unused
Bit 5. Carrier Sweep Function
When this bit is set to zero, the sweep function is disabled and the
carrier frequency is forced to the preset value defined in register
locations 04, 05 and 06.
Bit 6. Viterbi Reset
When this bit is set to zero, the accumulator for the signal quality is
cleared to zero. In normal operation, this bit is set to one.
Bit 7. Reed Solomon Error Counter
When this bit is set to zero, the counters for the number of corrected
errors and the number of uncorrected code words are cleared to zero.
51
11
Wideband AGC Control
Bit 0. Wideband AGC Mode
When this bit is set to one (Mode 0), the WB AGC output must be
filtered with an external integrating analog filter to implement a first order
feedback loop. When this bit is zero (Mode 1), a digital integrator within
the HDM8515 performs this function and the only external analog
function required is a low pass filter to remove the high frequency
components of the sigma delta converter output.
Bit 1. WB AGC Invert
When this bit is set to zero a high duty factor on the WB AGC output
corresponds to too much gain. When the control bit is set to one, high
duty factor corresponds to not enough gain.
Bit 2. WB AGC Hold
During normal tracking operation, this bit is set to one. When this bit is
set to zero and the wideband AGC is in Mode 1, the digital integrator is
held to the most recent value and loop updates are inhibited.
Bit 3. LNB Hold
When this bit is set to one, the output of LNB-Tone is held on zero.
Bit 4. I2C By-pass
When this bit is set to zero, SCL_I2CO and SDA_I2CO are disabled.
The default is one and Data/clock can be by -passed.
Bits [7:5]. WB AGC Gain
This three bit field defines the time constant of the WB AGC in Mode 1.
A value of zero corresponds to the shortest time constant and 7
corresponds to the slowest time constant.
12
LNB Tone
This eight bit value establishes the control for LNB tone generator. If fL
is the desired frequency and fC is the clock frequency, the value to be
stored in this 8 bit field is the integer portion of fL(217)/fC. The default
value(20H) generates 22KHz tone at 90MHz sampling clock.
13
Sigma Delta
This eight bit input value establishes the control for Sigma Delta
converter. This function is independent of other demodulator functions
and is provided as control for external analog components.
52
14
Test Set-up
The eight bit data written to this location defines the data presented on
the 16 bit test bus. For configurations where the data is updated once
per symbol period, the data changes at the rising edge of
SYMBOL_CLOCK
(in the case that SYMBOL_CLOCK remains high for consecutive
CLOCK
cycles, the test port data will also change accordingly during the high
period of SYMBOL_CLOCK due to the arrival of another symbol).
Bits [2:0]. Test port configuration
00H Output is tristate.
01H Test bits [13:8] provide the I baseband filter output. Test bits [5:0]
provide the Q baseband filter output. This information is updated once
per symbol period.
02H Test bits [15:0] provide the sixteen most significant bits of the
demodulator carrier phase test bits. This information is updated once
per
symbol period.
03H Test bits [15:0] provide the sixteen most significant bits of the
demodulator symbol phase test bits. This information is updated once
per
symbol period.
04H Test bits [15:8] provide the Reed Solomon output data. Test bits
[7:0] provide the deinterleaver output data. This information is updated at
the Reed Solomon clock rate; when the transport stream output is
configured to parallel output mode, DATA_CLK may be used as an
output clock for this data.
05H All Zero.
06H Test bits [13:8] provide the six bit I-channel data from the ADC.
Test bits [5:0] provides the six bit Q-channel data from the ADC. This
information is updated at the fixed rate sample clock.
07H In this mode the test pins are used as input pins. The internal
ADC is disabled, and the inputs at the test pins are fed directly to the
demodulator. Test bits [13:8] are used as I-channel input and test bits
[5:0] are used as Q-channel input. This information is updated at the
fixed rate sample clock.
Bit 3. Transport error Indicator Enable/Disable
Enables/Disables the transport error indicator,1 bit indicator in transport
header. When this bit is set to 1 and if transport error is internally
detected the transport error indicator bit is set to 1. When zero this
functionality is disabled.
53
Bit 4. This bit should be fixed to zero
Bit 5. Regulated Data Clock
Enables/Disables the data and data clock regulator. When this bit is set
to 1, data output and data clock are regulated by FIFO operation. When
this bit is set to 0, internal data output and internal data clock are bypassed
Bit 6. This bit should be fixed to zero.
Bit 7. Clock Polarity
This bit is used to select the DATA_CLK polarity either for serial or
parallel transport interface. If this bit is set to zero(default value), the
transport data and control signals are latched at the positive edge of
DATA_CLK. Otherwise, the signals are latched at the negative edge of
DATA_CLK.
15
Viterbi Lock Threshold
Register 15 to 17 contain control parameters for synchronization in
Viterbi decoder. Ordinary users are recommended to use the default
value.
Bit[7:4] defines the lock threshold for VB_NODESYNC. Viterbi decoder
decides that the correct code rate has been found. A large number
means it takes longer to find the correct code rate in automatic
detection mode. It should be greater than 7. The default value is 12.
Bit[3:0] defines the lock fail threshold. Viterbi decoder rejects a code
rate and moves on to the next code rate. A small number means Viterbi
decoder tries more data before it moves to the next code rate. It should
be less than 7. The default value is 2.
16
Viterbi Unlock Threshold
This number defines the threshold to maintain the Viterbi lock state. A
large number means it needs more bad data to get out of the viterbi lock
state and re-start searching the correct code rate. The default value is
1.
54
17
Viterbi Byte-Sync control
Once the viterbi lock(VB_NODESYNC) is achieved, the Viterbi decoder
tries to find the byte-sync. This 8-bit register is used to set “unlockthreshold” for the byte-sync. Large number means it needs more baddata to get out of the byte-sync state, i.e. less sensitive to noise. The
default value is 1.
55
18
Control Parameters for Viterbi and RS Decoders
Bit 0. Parallel or Serial Output
Controls the transport stream output of the 8515 to serial or parallel
mode.
“0” (default) means the 8515 MPEG output is parallel.
“1” means the 8515 MPEG output is serial. The LSB of the data
bus(data[0] - pin 98) is used as the serial output pin.
Bit 1. MPEG2 Data
“0” (default) means the incoming data is MPEG2 decoded. In this mode
a sync byte is expected every 188 bytes.
“1” means non-MPEG2 data. The Viterbi decoder doesn’t check the
existence of the sync byte.
Bit [4:2]. Depuncturing Rate
It defines the depuncturing rate of the Viterbi decoder. When
vb_autocode is disabled, the depuncturing rate is set to this value.
0 --- 1/2
1 --- 2/3
2 --- 3/4
3 --- 5/6
4 --- 7/8
5 --- 6/7
Bit 5. Viterbi Auto Decoding Mode
When this bit is set to 1, the Viterbi decoder automatically finds the
correct code rate of the incoming signal. When this bit is set to 0, the
code rate is set to the user-defined value at bit[4:2]. The default is 0.
Bit 6. DSS Mode
When this bit is set to 0, this device operates as DVB mode. When
this bit is set to 1, this device operates as DSS mode. In that case, the
roll-off factor of the Nyquist filter is set to 0.2. The default is 0 (DVB).
Bit 7. BPSK Mode
When this bit is set to 0, the demodulator assumes the incoming data
is QPSK-modulated. When this bit is set to 1, the demodulator
assumes the incoming data is BPSK-modulated.
19
Rate 1/2 Threshold Select
This seven bit parameter defines the threshold used in the Viterbi
decoder node synchronization process. For rate 1/2, the nominal value
is 30 (1EH).
56
1A
Rate 2/3 Threshold Select
This seven bit parameter defines the threshold used in the node
synchronization process. For rate 2/3, the nominal value is 30 (1EH).
1B
Rate 3/4 Threshold Select
This seven bit parameter defines the threshold used in the node
synchronization process. For rate 3/4, the nominal value is 40 (28H).
1C
Rate 5/6 Threshold Select
This seven bit parameter defines the threshold used in the node
synchronization process. For rate 5/6, the nominal value is 60 (3CH).
1D
Rate 6/7 Threshold Select
This seven bit parameter defines the threshold used in the node
synchronization process. For rate 6/7, the nominal value is 60 (3CH).
1E
Rate 7/8 Threshold Select
This seven bit parameter defines the threshold used in the node
synchronization process. For rate 7/8, the nominal value is 60 (3CH).
Bit 7. This bit should be fixed to zero.
57
1F
Bit 7. This bit should be fixed to zero.
20
Unused
21
Wideband AGC threshold
Bits [5:0]. Wideband AGC threshold
It determines the threshold of wide band AGC accumulator. This value
controls the magnitude of ADC input.
Bit 6. Unused
Bit 7. Wideband AGC Frequency down
It regulates wide band AGC frequency. When this bit is set to zero,
system clock for wide band AGC frequency is sampling clock.
Otherwise, system clock for wide band AGC frequency become sixteen
times of sampling clock frequency.
22
Scaling factor
Bits [2:0]. Scaling factor
This value manages to scale the soft decision demodulator outputs to
the proper levels for the 4 bit soft decision Viterbi inputs. If an overflow is
detected, the output is limited to maximum or minimum 6 bit values.
The upper four bits of this result are passed to the Viterbi decoder.
00H QPSK output
01H QPSK output * 1.25
02H QPSK output * 1.5
03H QPSK output * 1.75
04H QPSK output * 2.00
05H QPSK output * 2.25
06H QPSK output * 2.5
07H QPSK output * 2.75
58
23
Clock Generation PLL Control Parameter-1
Bits [1:0]. VCO range control vector
Default value is 1.
Bits [3:2]. Pre divisor
In case of Extended Frequency mode, this value is used the calculation
of
output Frequency. Default value is 0.
Bit 4. Digital part test mode
When this bit is set to 1and Bit 7 of this register is set to 1, the PLL is
bypassed and the external clock signal is directly connected to the
internal clock. When this bit is set to 0, the generated clock of the PLL
is connected to the internal clock. The default is 1.
Bit 5. VCO power down mode
When this bit is set to 1, VCO power down and does not oscillate
Bit 6. PLL power down mode except VCO
When this bit is set to 1, PLL power down and digital circuits do not
operate and charge pump is disabled.
Bit 7. PLL By-pass
24
If this bit is set to 0 and Digital part test mode (Bit 4 of this register) is
set to 0, then PLL Normal frequency mode is selected. Else if this bit is
set to 1 and digital part test mode is set to 0, then PLL Extended
frequency mode is selected.
Clock Generation PLL Control Parameter-1
Bits [1:0]. Unused
Bit 2. Counter toggle test
Internal used Only. The default is 0.
Bits [5:3]. Loop filter mode selector
The default is 5H.
Bits [7:6]. Charge pump test mode
25, 26
Internal used Only. The default is 0.
M Divider Ratio
This 14 bit value defines a feedback divider with a divider ratio M. Bit 5 of
address 25 is the MSB and bit 0 of address 26 is the LSB. The default
value is 002Bh
59
27
N Divider Ratio
It defines a reference divider with a divider ratio N. The default value is
01h
28
Charge pump bias current control vector
Bits [3:0]. The default value is 1.
29
DiSEqC message frame byte
It defines the format of frame in DiSEqC message
2A
DiSEqC message address byte
It defines the format of slave address in DiSEqC message
2B
DiSEqC message command byte
It defines the format of command in DiSEqC message.
2C, 2D, 2E, 2F, 30
DiSEqC message data byte(s)
For some DiSEqC message, additional data is carried in one or more
subsequent data byte(s).
31
DiSEqC Mode Control
Bit 0. DiSEqC By-pass
When this bit is set to 1, the function of DiSEqC interface is disabled.
When this bit is set to 0, DiSEqC interface is enabled, The default is 1.
Bits [3:1] DiSEqC mode
It determines one of following DiSEqC modes.
0: 22KHz off
1: 22KHz on continuous
2: Burst mode - on for 12.5ms =’0’
3: Burst mode - modulated 1:2 for 12.5ms =’1’
4: Modulated with bytes from DiSEqC instruction.
5-7: Reserved
Bits [7:4] DiSEqC message length
Number of byte in DiSEqC instruction, to output on DISEQC pin.
60
6.2 Read Registers
ADDRESS (Hex)
00
Narrowband AGC Accumulator
The current value of the 8 bit AGC accumulator may be read from this
location.
01, 02, 03
Symbol Timing Frequency Accumulator
The current value of the 24 bit frequency accumulator in the symbol timing
loop filter may be read from these 3 locations.
04, 05, 06
Phase Tracking Frequency Accumulator
The current value of the 24 bit frequency accumulator in the carrier phase
loop filter may be read from these 3 locations.
07
QPSK Lock Status
Bit 0. QPSK Lock Flag
When this bit is set to one, the QPSK demodulator is phase locked.
08
Wide Band AGC Accumulator
This eight bit value represents the most significant bits of the accumulator
in the first order wideband AGC loop. This data only has meaning when
the wideband AGC is in Mode 0.
09, 0A
Sweep Frequency
The 16 bit sweep accumulator is available at this location. Bit 7 of address
09 is the MSB and bit 0 of address 0A is the LSB. The receiver frequency
is determined by adding the Sweep Frequency with the carrier frequency
accumulator (read addresses 04, 05 and 06) and the nominal carrier start
frequency (write addresses 04, 05 and 06).
0B
In-Phase
The six LSB bit output of the In-phase baseband filter is available at this
location. This data is updated once per symbol.
61
0C
Quadrature
The six LSB bit output of the quadrature baseband filter is available at this
location. This data is updated once per symbol.
0D
Noise Power
This eight bit output provides a measure of the noise component of the
signal when QPSK lock is achieved. Higher numbers correspond to lower
signal-to-noise ratio conditions. The quality of this metric is improved if
the narrowband AGC is disabled for a minimum of 1000 symbol periods
before this parameter is read.
0E, 0F
BER Calculator
The current value of the 16bit BER is used to monitor the signal quality or
estimate the SNR of incoming signal at the output of Viterbi. Bit 7 of
address 0E is the MSB and bit 0 of address 0F is the LSB. It represents
the number of errors among 220 data bits.
10, 11, 12
Carrier Frequency_1
This 24 bit value represents the carrier frequency of the first frequency
translator.
13, 14, 15
Carrier Frequency_2
This 20 bit value represents the carrier frequency of second frequency
translator. Bit 7 of address 13 is the MSB and bit 4 of address 15 is the
LSB.
16,17,18
Signal Quality
This 24 bit signal provides a measure of quality of the signal processed by
the Viterbi decoder. This parameter can be used to infer bit error rate and
input signal-to-noise ratio for signals which are within a few dB of threshold.
Bit 7 of address 16 is the MSB and bit 0 of address 18 is the LSB.
The specific definition of this signal for each coding rate is TBD.
19
Viterbi Rate
This three bit number represents the code rate of the Viterbi decoder.
Rate 1/2
Rate 2/3
Rate 3/4
Rate 5/6
Rate 7/8
62
0
1
2
3
4
1A
Reed Solomon Errors
The four bit number at this location indicates the number of errors
corrected in the most current block of 188 bytes. This number may range
from 0 to 8.
63
1B
FEC Lock
Bit 0. Viterbi Node Sync
When this bit is set to one, the Viterbi decoder has successfully
established node synchronization.
Bit 1. Frame Sync
When this bit is set to one, the FEC chip has successfully established
word sync and frame sync.
Bit 2. Viterbi Byte Sync
When this bit is set to one, the Viterbi decoder has successfully
established byte-synchronization.
Bit 3. Pi Ambiguity
When this bit is set to one, the Viterbi decoder has successfully
resolved pi ambiguity in the input data. (i.e inverted data)
Bit 4. Pi/2 Ambiguity
When this bit is set to one, the Viterbi decoder has successfully
resolved pi/2 ambiguity in the input data
1C,1D
Accumulated Reed Solomon Errors
These two registers present a count of corrected errors since it was last
reset. Bit 7 of address 1C is the MSB and bit 0 of address 1D is the
LSB. These registers are reset by writing value to address 10H.
1E
Accumulated Reed Solomon Data
This register presents a count of the uncorrected code words since it
was last reset. When it reaches its maximum count(255), it rolls back
to zero
and starts counting again. This register is reset by writing value to
address 10H.
1F
Device Identifier
23
This register present device identifier. The current value of this register
is
F0H
Reference Divider Test Output
Internal used Only.
24, 25
Feedback Divider Test Output
Internal used Only.
64
26
PLL Lock Indicator
Bit 0. This 1 bit value represents the staus of PLL lock
If PLL is locked, this value is 1, else 0.
65
Appendix
66
A1. Loop Filter Programming Application Note
To illustrate that the symbol timing recovery loop and the carrier phase recovery loop are both
programmable, several simulations were performed with different loop parameter conditions. These
simulations were performed with a symbol rate of two samples per symbol, corresponding to 30M
symbols-per-second if a 60MHz clock were utilized.
Figure A1 illustrates the transient response of the symbol phase with three different loop conditions
(K1=5, K2=10; K1=4, K2=9; and K1=8, K2=7). The vertical scale represents phase over a 360
degree range (524,287 to -524,288). All test cases were run at high signal -to-noise ratio. The
highest gain condition could be used for fast acquisition as well as for steady state with high code
rate conditions, while the intermediate gain is a suitable steady state setting for rate 1/2 codes
(minimum Eb/N0 of 4 dB). The lowest gain setting corresponds to ultra low loop bandwidth and
may be considered for maintaining lock without phase jumps during deep signal fades.
FIGURE A1: SYMBOL TIMING RECOVERY TRANSIENT RESPONSE
67
Figure A2 illustrates the transient response of the carrier tracking loop with the same loop
bandwidth settings at high signal-to-noise ratio. The phase step for this test corresponds to 45
degrees. The actual bandwidth of the carrier loop is greater than that of the symbol loop for the
same settings because the carrier loop must cope with greater dynamics (such as frequency offset
and drift). Figure A3 illustrates the transient response of the carrier phase tracking loop under the
same conditions at minimum signal-to-noise ratio (E b/N0 of 4 dB with rate 1/2 coding). The
highest bandwidth case will pull in with a carrier frequency error of + or - .005 of the symbol rate at
this minimum signal level. Higher loop bandwidth may be programmed to provide greater pull-in
with higher signal-to-noise ratio conditions.
FIGURE A2: CARRIER PHASE RECOVERY TRANSIENT RESPONSE
68
FIGURE A3: CARRIER PHASE RECOVERY TRANSIENT RESPONSE WITH LOW SNR
69
A2. False Lock Escape Application Note
A QPSK signal will have inherent false lock states at frequency offsets of + or - n/4 of the symbol
rate. Mos t DBS signals which have symbol rates of 20M symbols-per-second or higher will not
experience false lock because the carrier frequency uncertainty is less than 1/4 of the symbol rate.
The HDM8515 is designed to process low data rate signals which may experience false lock,
particularly at high signal-to-noise ratio conditions. The HDM8515 will permit recovery from false
lock with some added host processor interaction. Specifically, the processor must initialize the
internal carrier frequency search hardware to search over a carrier frequency range of 1/4 of the
symbol rate. If QPSK lock is achieved, but no Viterbi lock is achieved, the processor would
assume this is a false carrier lock, then program the HDM8515 to search another carrier frequency
range covering 1/4 of the symbol rate. When both QPSK lock and Viterbi lock have been achieved,
the search is completed. This technique is reliable because the HDM8515 utilizes a fixed
frequency clock which is not subject to inaccuracy associated with analog VCOs.
This
accuracy insures that the multiple search ranges are perfectly continuous with respect to each
other with no overlap.
70
A3. Performance with Interference.
In order to evaluate the filter employed within the HDM8515 with respect to attenuating out-of-band
interference, a test was performed utilizing the COSSAP simulator. The desired signal, at zero
frequency, was configured to utilize 16 samples-per-symbol (corresponding to 3.75MHz symbol
rate if a 60MHz clock is employed). An interfering signal was added with the same characteristics,
except that the amplitude was made to be 10dB higher than that of the desired signal, the data
pattern was different and the carrier frequency was offset from that of the desired signal. Several
offset frequencies were evaluated for this case. Figure A4 illustrates the spectrum of the test
condition when the offset frequency is 1.35 times the symbol rate.
Figure A5 illustrates the measured bit error rate for various conditions. The error rate on the I
channel was measured separately from that of the Q channel, and the horizontal axis is scaled in
dB for one component (I or Q of the signal). For example, the point labeled 1dB corresponds to
SNR (noise bandwidth = symbol rate) of 4dB or Eb/N0 of 4dB if rate 1/2 coding is employed.
The theoretical performance for coherent PSK is shown with the solid line. The curve closest to
theoretical is the demodulator performance with no other interferers and corresponds to an
implementation loss of about 0.2dB. When the interferer was placed at a frequency of either 2.0 or
1.35 times the symbol rate away from the desired carrier, there is an additional degradation ranging
from 0dB to 0.1dB. The worst case occurs when the interferer is placed at only 1.28 times the
symbol rate from the carrier of the desired signal. In this case, the performance has degraded with
respect to the no interference case by 0.3 to 0.5dB.
Figure A6 illustrates performance with an interferer which is 20dB higher than the desired signal
and separated in frequency by 2 times the symbol rate. In this case, the performance has
degraded by 0.7 to 0.8dB from the case with no interferer.
71
FIGURE A4: A DJACENT CHANNEL INTERFERENCE OF 10 DB, 1.35 SPACING
72
FIGURE A5: PERFORMANCE WITH INTERFERER AT DIFFERENT CARRIER SPACINGS
73
FIGURE A6: PERFORMANCE WITH +10 DB INTERFERER
74
A4. Nyquist Criteria Considerations
The HDM8515 is clocked at 60MHz, yet processes signals with symbol rates as high as 45M
symbols-per-second. At first thought, this might appear to be violating the Nyquist criteria which
states that the sampling rate must be at least twice the highest frequency component. The total
bandwidth of the 45Msps signal, with 35% excess bandwidth, is about 60MHz.
The samples provided to the HDM8515 are complex samples, which is equivalent to 120M
samples-per-second, which does satisfy the Nyquist criteria. Another way of looking at this is to
examine the baseband signal. The signal bandwidth covers 60MHz, but the baseband spectrum
covers from -30MHz to +30MHz. There are no baseband frequency components greater than
30MHz, and the 60MHz clock is adequate as long as complex samples are taken.
75