Datasheet GDC21D003 Datasheet (HEI)

Download

Page 1

GDC21D003

(VSB Receiver)

Version 1.0

Mar, 99

HDS-GDC21D003-9908 / 10

Page 2

GDC21D003

The information contained herein is subject to change without notice. The information contained herein is presented only as a guide for the applications of our products. No

responsibility is assumed by Hyundai for any infringements of patents or other rights of the third parties which may result from its use. No license is granted by implication or otherwise under any patent or patent rights of Hyundai or others.

These Hyundai products are intended for usage in general electronic equipment (office equipment, communication equipment, measuring equipment, domestic electrification, etc.).

Please make sure that you consult with us before you use these Hyundai products in equipment which require high quality and / or reliability, and in equipment which could have major impact to the welfare of human life (atomic energy control, airplane, spaceship, traffic signal, combustion control, all types of safety devices, etc.). Hyundai cannot accept liability to any damage which may occur in case these Hyundai products were used in the mentioned equipment without prior consultation with Hyundai.

Page 3

GDC21D003

TABLE OF CONTENTS

1. General Description.................................................................................................................8

2. Features....................................................................................................................................8

3. Internal Block Diagram..........................................................................................................11

4. Pin Description.......................................................................................................................11

4.1 Pin Configuration ..............................................................................................................11

4.2 Pin Description..................................................................................................................14

4.3 Pin Assignment.................................................................................................................16

5. I2C Bus I/F & Registers..........................................................................................................17

5.1 I2C Bus I/F Description......................................................................................................17

5.1.1 Write Operation...........................................................................................................17

5.1.2 Read Operation...........................................................................................................17

5.2 I2C Bus Register Configuration.........................................................................................18

5.3 I2C Bus Register Description ............................................................................................20

6. Functional Description..........................................................................................................29

6.1 ADC ..................................................................................................................................29

6.1.1 Electrical Characteristics.............................................................................................30

6.1.2 Timing Diagram...........................................................................................................32

6.1.3 Application Circuits......................................................................................................33

6.2 Clock Divider.....................................................................................................................35

6.3 Synchronizer.....................................................................................................................36

6.3.1 Input Control................................................................................................................36

6.3.2 DC Reduction..............................................................................................................39

6.3.3 Auto Gain Control(AGC) .............................................................................................39

6.3.4 Polarity Correction.......................................................................................................40

6.3.5 Data Segment Sync Recovery....................................................................................41

6.3.6 Polarity Decision..........................................................................................................42

6.3.7 Timing Recovery .........................................................................................................43

6.3.8 Field Sync Recovery...................................................................................................46

6.3.9 VSB Mode Detect........................................................................................................47

6.3.10 NTSC Rejection.........................................................................................................48

6.4 Equalizer...........................................................................................................................50

6.4.1 Block Diagram.............................................................................................................50

6.4.2 Training/Data Mode Equalization................................................................................51

6.4.3 Error Estimation...........................................................................................................52

6.4.4 Adaptive Filter .............................................................................................................53

6.4.5 Equalizer Clock Scheme.............................................................................................53

6.4.6 I2C Bus I/F...................................................................................................................54

6.4.7 Coefficient Reading/Writing.........................................................................................56

Page 4

GDC21D003

6.5 Phase Tracker ..................................................................................................................56

6.5.1 Error Detection............................................................................................................57

6.5.2 Gain & Offset Loop......................................................................................................58

6.5.3 Phase Loop.................................................................................................................58

6.5.4 I2C Bus I/F ..................................................................................................................59

6.6 Channel Decoder..............................................................................................................60

6.6.1 12 Symbol Intrasegment Deinterleaver.......................................................................61

6.6.2 Segment Sync Suspension.........................................................................................61

6.6.3 Viterbi Decoder............................................................................................................62

6.6.4 Symbol-to-Byte Converter...........................................................................................64

6.6.5 Convolutional Deinterleaver........................................................................................65

6.6.6 Reed-Solomon Decoder..............................................................................................65

6.6.7 Data Derandomizer.....................................................................................................66

6.6.8 I/F to Transport Demultiplexer.....................................................................................67

6.7 PLL....................................................................................................................................74

7. Electrical Characteristics......................................................................................................75

8. Package Dimensions.............................................................................................................77

9. Application Notes ..................................................................................................................78

Figures

Figure 3.1 Functional Block Diagram ..............................................................................10

Figure 5.1.1 I2C Write Operation Example ..........................................................................17

Figure 5.1.2 I2C Read Operation Example ..........................................................................18

Figure 6.1.1 The Block Diagram of ADC .............................................................................29

Figure 6.1.2 Timing Diagram of ADC ..................................................................................32

Figure 6.1.3 ADC Application Circuit ...................................................................................33

Figure 6.1.4 Equivalent Circuits ..........................................................................................34

Figure 6.3.1 The Block Diagram of Input Selection ............................................................36

Figure 6.3.2 Digital Input Setting Up & Chip I/F Circuit(1) ..................................................37

Figure 6.3.3 Digital Input Setting Up & Chip I/F Circuit(2) ..................................................38

Figure 6.3.4 The Block Diagram of DC Reduction ..............................................................39

Figure 6.3.5 The Block Diagram of AGC ............................................................................40

Figure 6.3.6 AGC Signal I/F for DTV System .....................................................................40

Figure 6.3.7 The Block Diagram of Polarity Correction .......................................................41

Figure 6.3.8 The Block Diagram of Data Segment Sync Recovery ....................................41

Figure 6.3.9 Polarity signal I/F Circuit .................................................................................42

Figure 6.3.10 Timing Recovery Block ....................................................................................43

Figure 6.3.11 Timing Recovery I/F Circuit(1) .........................................................................44

Figure 6.3.12 Timing Recovery I/F Circuit(2) .........................................................................45

Figure 6.3.13 Field Sync Structure ........................................................................................46

Figure 6.3.14 The Block Diagram of Field Sync Recovery ....................................................46

Figure 6.3.15 Comb Filter Block ............................................................................................48

Page 5

GDC21D003

Figure 6.3.16 The Block Diagram of NTSC Rejection............................................................49

Figure 6.4.1 Channel Equalizer ...........................................................................................50

Figure 6.4.2 VSB Slicer .......................................................................................................51

Figure 6.4.3 Training/Data Equalization ..............................................................................51

Figure 6.4.4 VSB Slice Level ..............................................................................................52

Figure 6.4.5 Coefficient Update Filter .................................................................................53

Figure 6.4.6 I2C Bus I/F .......................................................................................................54

Figure 6.5.1 Phase Tracker .................................................................................................57

Figure 6.5.2 Error Detection ................................................................................................57

Figure 6.5.3 Coefficients of Hilbert Transform Filter ...........................................................58

Figure 6.5.4 Complex Multiplier ..........................................................................................58

Figure 6.6.1 Block Diagram of Channel Decoder ...............................................................60

Figure 6.6.2 12-Symbol Intrasegment Deinterleaver ..........................................................61

Figure 6.6.3 Segment Sync Suspension .............................................................................62

Figure 6.6.4 Viterbi Decoding with and without NTSC Rejection Filter ...............................63

Figure 6.6.5 Internal Block Diagram of Viterbi Decoder ......................................................63

Figure 6.6.6 Convolutional Deinterleaver ............................................................................65

Figure 6.6.7 Derandomizer Polynomial ...............................................................................66

Figure 6.6.8 I/F to Transport Demultiplexer when Register64[7:0] is set to

Default Value ....................................................................................................68

Figure 6.6.9 I/F to Transport Demultiplexer when Register64[3] (Derand_on)

is set to ‘ 0’ ........................................................................................................69

Figure 6.6.10 I/F to Transport Demultiplexer when Register64[2](Errorflg_ins)

is set to ‘ 0’ ........................................................................................................69

Figure 6.6.11 I/F to Transport Demultiplexer when Register64[1](Vsbdvalid_pol)

is set to ‘ 0’ ........................................................................................................70

Figure 6.6.12 I/F to Transport Demultiplexer when Register64[0](Vsbclk_sup)

is set to ‘ 0’ ........................................................................................................70

Figure 6.6.13 I/F to Transport Demultiplexer(MMDS 8VSB Mode) .......................................71

Figure 6.6.14 I/F to Transport Demultiplexer at Serial Output Mode .....................................71

Figure 6.6.15 Connection with VSB Receiver and Transport Demultiplexer

Chip(GDC21D301A) ........................................................................................72

Figure 6.6.16 Connection with VSB Receiver and Transport Demultiplexer

Chip(L64007) ...................................................................................................72

Figure 6.6.17 Connection with VSB Receiver and Transport Demultiplexer

Chip(AVIA-MAX) ..............................................................................................73

Figure 6.7.1 Clock Scheme .................................................................................................74

Figure 7.1 Clock Reset Stabilization Timing .....................................................................76

Figure 7.2 Input and Output Timing ..................................................................................76

Figure 8.1 Physical Dimensions ........................................................................................77

Figure 9.1 VSB Receiver Application Circuit .....................................................................78

Page 6

GDC21D003

Tables

Table 6.2.1 Register Setting Up for Clock Divider .............................................................35

Table 6.3.1 Input Signal Path Setting Up ..........................................................................36

Table 6.3.2 DATAPOLP & DATAPOLN Output ................................................................42

Table 6.3.3 VSB Mode Data for Each VSB Mode .............................................................47

Table 6.3.4 VSB Mode Signal Control ..............................................................................47

Table 6.3.5 Comb Filter Control through I2C Bus ..............................................................49

Table 6.4.1 Contents of I2C Bus Register33 for Equalizer ................................................55

Table 6.4.2 Contents of I2C Bus Register33 for Filter Control ..........................................55

Table 6.5.1 Contents of I2C Bus Register33 for Phase Tracker ........................................59

Table 6.5.2 Contents of I2C Bus Register33 for Gain Control ...........................................59

Table 6.6.1 Bypassed Sub-blocks by the Values of I2C Bus Register64 ..........................60

Table 6.6.2 Symbol-to-Byte Conversion ...........................................................................64

Table 6.6.3 I2C Register64 Flags controlling Transport Demultiplexer I/F ........................67

Page 7

GDC21D003

VSB Receiver

GDC21D003

1. General Description

The VSB Receiver(GDC21D003) is an ATSC compliant single chip communications device that synchronizes, equalizes, and corrects errors of ATSC 8/16 VSB and MMDS (Multichannel Multipoint Distribution System) 2/4/8/16 VSB modulated signal. The on-chip 10-bit 10.76Msps Analog-to-Digital Converter has an input sample-and-hold amplifier. By implementing a multistage pipelined architecture with output correction logic, the ADC offers accurate performance and guarantees no missing codes over the full operating temperature. Clock divider divides output clock of external VCXO and generates symbol clock (CLKFS) and ADCCLK. The CLKFS has 10.76MHz frequency as symbol frequency used in DTV transmitter, ADCCLK is used for external A/D converter. At this time, if you use digital signal as input of chip, CLKFS or ADCCLK are used for external A/D converter clock. Synchronizer removes DC entered from transmitter and DC generated by analog circuit used in receiver. Also it checks gain of input signal and sends it to demodulator, detects polarity, and corrects it. It recovers Data Segment Sync period and Field Sync period entered from transmitter. It detects VSB mode of current input signal and removes NTSC co-channel interference in channel. Equalizer corrects linear distortion created during transmission. It uses Least-Mean-Square algorithm and has decision feedback equalizer structure. It uses adaptive filter having coefficient update structure consisted of multiplier, adder, and memory structure in every tap. Phase Tracker compensates phase distortion due to phase noise and it consists of gain correction loop for gain error, offset correction loop for offset error, and phase correction loop for phase error. Channel Decoder consists of Viterbi Decoder, Convolutional Deinterleaver, Reed-Solomon Decoder, Data Derandomizer, and etc. It decodes ATSC 8/16 VSB signal and MMDS 2/4/8/16 VSB signal. Also it has internal segment error counter that send out the number of segment errors per second and offers tri-state parallel/serial Transport Demultiplexer interface.

2. Features

General features

• ATSC compliant 8/16 VSB receiver

• MMDS 2/4/8/16 VSB receiver

• SNR threshold 14.9 dB on AWGN channel

• Tri-state parallel/serial MPEG-2 transport

interface

• Supports I2C bus interface

• Boundary Scan Test circuit complies with

IEEE Std. 1149.1

ID-Code = 0D0031C1

• Operating voltage : 3.3V

• 0.35µm CMOS technology

• 128 pin HQFP package ADC

• Resolution : 10bits (≤ ±1⁄2 LSB DNL error)

• Sampling rate : 10.76 Msps

• Differential input range : 2Vpp(1.7 ± 0.5V

differential)

Clock Divider

• Generates symbol clock(10.76MHz)

• Uses one of two VCXOs, fs(10.76MHz) and

2fs(21.52MHz) as input

Synchronizer

• Input control

• DC reduction and polarity correction

- Correction of polarity ambiguity caused by FPLL

• Non-coherent and coherent automatic gain control (AGC)

• Data Segment Sync and Field Sync recovery

• Timing recovery

• Polarity decision

- Polarity decision after Data Segment

Sync is locked

• VSB mode detection

• Comb control

- Comb filter for the rejection of NTSC

co-channel interface

Page 8

Equalizer

• Decision feedback equalizer

• Supports training sequence and blind

equalization

• Concurrent coefficients update in symbol time

• Available 3 different step-size

• Capability of reading equalizer coefficients

• Ghost cancellation in the range from -2.86µs to 20.76µs

Phase Tracker

• Intelligent loop control according to noise environment

• Phase tracking from -60° to 60° with resolution of 0.004 degree

• Phase, offset, and gain correction at a time

Channel Decoder

• Concatenated Viterbi/Reed-Solomon Decoder with Deinterleaver and Derandomizer

• Internal segment error counter

• Tri-state parallel/serial MPEG-2 Transport

Demultiplexer interface

GDC21D003

Page 9

3. Internal Block Diagram

GDC21D003

NADTONDATA

INN

INP

DIN[9:0]

GUP

GDN

DATAPOLP DATAPOLN

NCHGUP NCHGDN

DOUT[9:0]

TM[2:0]

VCXO ADCCLK

CLKFS

I2C BUS

ADC

Mux

Clock

Divider

I2C

Interface

Input

SelectionDCReduction

AGC

Polarity

Decision

Timing

Recovery

Phase

Loop

Polarity

Correction

Data

Segment

Sync

Recovery

Field Sync

Recovery

Offset

Loop

Gain

Loop

Comb

Filter

Comparator

NTSC Rejection

192 Tap

Feedback

Filter

Mux

MSE

VSB

Mode

Detect

64 Tap

Forward

Filter

TRST TMS TCK TDI TDO

SYMCLK

JTAG

Error

Detect

Phase Tracker

Deinterleaver/

Viterbi Decoder

Convolutional

Deinterleaver

Reed-Solomon

Decoder

Data

Derandomizer

Figure 3.1 Functional Block Diagram

Error

Control

Equalizer

Transport

Demultiplexer I/F

VSBCLK

VSBDVALID

VSBDATA[7:0]

4x Clock

4X PLL

VSBSOP

NVSBERRFLG

Page 10

4. Pin Description

4.1 Pin Configuration

GDC21D003

∗

128 PIN HQFP, 28X28 mm BODY, 1.60/0.33 mm FORM, 3.37mm THICK, 0.8 mm PITCH

Page 11

4.2 Pin Description

Clock/Reset ; 6 Pins

PIN NAME TYPE DESCRIPTION

A/D Converters ; 7 Pins

PIN NAME TYPE DESCRIPTION

121 123 124 125 126 128

NRESET I

VCXO I

ADCCLK O Clock for Off-chip ADC(21.52MHz or 10.76MHz);

CLKFS O

CLK4FS I/O Test clock/4x symbol clock;

SYMCLK I System clock input(10.76MHz)

REF I

COM I Common voltage(1.5V) REFN I Reference voltage(bottom: 1.2V) REFP I Reference voltage(top: 2.2V) INN I Analog data input(negative) INP I Analog data input(positive) BIAS I

GDC21D003

System reset(active low); This signal should be

activated on channel change or power on. Clock input generated in VCXO; This pin can be connected to one of two VCXOs whose output frequencies are fs(10.76MHz) and 2fs(21.52MHz).

This clock is generated by dividing VCXO input signal. System clock; This clock is generated from dividing VCXO input signal. Its frequency is the same of symbol rate(10.76MHz).

When PLLEN(pin35) input is set to ‘1’, this pin is used

as 4x symbol clock output.

When PLLEN(pin35) input is set to ‘0’, this pin is used

as test clock(43.04MHz) input.

Bias register for internal ADC; This pin should be connected to AVDD(3.3V) via 12k ohm register.

1.65 ± 0.5V

differential

Bias input(2V typical) for On-chip ADC; This pin should be connected to AVSS via 0.1µF capacitor.

Page 12

GDC21D003

Sync Recovery ; 20 Pins

PIN NAME TYPE DESCRIPTION

20, 22-24, 26-28, 30-32

(NOTE) * These five I/O pins are used as input pin only for chip test.

DIN[9:0]

Bit9 : MSB

DATAPOLP I/O

DATAPOLN I/O*Inverted polarity signal of input data(active high);

SEGSYNCLOCK I/O*Stability Indication of Data Segment Sync

GUP I/O

GDN I/O

NFSYNC O

NCHGUP O Charging signal for charge pump in the timing

NCHGDN O Discharging signal for charge pump in the timing

FOE O

NSEGSYNC O

Equalizer ; 6 Pins

PIN NAME TYPE DESCRIPTION

35 108

110

114

116

117

(NOTE) When I2C is enabled, operation is performed either using these pins or via I2C register, but when disabled, only these pins are used. If you want to control Equalizer fast, use these external input pins. Otherwise use I2C bus registers.

PLLEN I NCOUPDTWIN O

EQSTAT O

NADTONDATA I Data mode coefficient update(active low);

NINITEQ I

NFREEZEEQ I

Digital data input; This data input comes from

external ADC.

Polarity signal of input data(active high); If this

output value is ‘1’, it means the plus polarity. This signal should be applied to demodulator IC.

This signal should be applied to demodulator IC.

recovery(active high)

Input data gain increasing signal(active high); This

signal should be applied to demodulator IC.

Input data gain decreasing signal(active high); This

signal should be applied to demodulator IC. Field Sync(active low); If this output value is ‘0’, it means Field Sync interval.

recovery block(active low)

recovery block(active low) Field status indicator(active high); If this output value

is ‘1’, it means inverted field. Data Segment Sync(active low); If this output value is ‘0’, it means Data Segment Sync interval.

PLL enable(active high); This pin should be set to ‘1’. Coefficient update window(active low); If this output

value is ‘0’, the Equalizer adapts its coefficients. Otherwise, it doesn't adapt its coefficients. Equalizer status; If this output value is ‘1’, the Equalizer is in normal status. Otherwise, the Equalizer has diverged.

If this input is set to ‘0’, the Equalizer adapts its coefficients during training sequence and data interval. Otherwise, the Equalizer adapts its coefficients during only training sequence interval. Equalizer initialization(active low); If this input is set to ‘0’, the Equalizer is initialized. Equalizer freeze(active low); If this input is set to ‘0’, the Equalizer coefficient does not be adapted.

Page 13

GDC21D003

Phase Tracker ; 2 Pins

PIN NAME TYPE DESCRIPTION

111

112

NINITPH I

NFREEZEPH I

Channel Decoder ; 12 Pins

PIN NAME TYPE DESCRIPTION

68,69,71,72,74,75,77, 80

85 89

VSBDATA[7:0]

Bit7 : MSB

NVSBERRFLG O

VSBDVALID O Valid data indication flag;

VSBSOP O Start byte indicator of a packet VSBCLK O Data clock of packet data

Phase tracker initialization(active low); If this input

is set to ‘0’, the Phase Tracker is initialized. Phase tracker freeze(active low) ; If this input is set to ‘0’, the Phase Tracker stops phase tracking..

O Data output to transport multiplexer;

VSBDATA[7] : Used as start bit indicator of a byte in

serial output mode.

VSBDATA[0] : Used as serial data output in serial

output mode.

Packet error indication flag(active low); This output

indicates whether the packet has error or not. 0 : with error 1 : without error

1: valid when register64[1](Vsbdvalid_pol = ‘1’) 0 :valid when register64[1](Vsbdvalid_pol = ‘0’)

I2C Bus Interface ; 4 Pins

PIN NAME TYPE DESCRIPTION

9 10 39

119

SCL I I2C bus serial clock input SDA I/O I2C bus serial data input/output NI2CEN I

I2CSEL I I2C bus device address selection;

Boundary Scan Signal ; 5 Pins

PIN NAME TYPE DESCRIPTION

3 4 5 6 7

TRST I Boundary scan test reset TMS I Boundary scan test mode selection TCK I Boundary scan test clock TDI I Boundary scan test data input TDO O Boundary scan test data output

C bus enable(active low); If this input is set to ‘0’,

the chip is controlled by I2C bus.

Otherwise, stand alone mode.

0 : I

C device address is set to b"1011001".

1 : I

C device address is set to b"0001110".

(default value)

Page 14

Miscellaneous ; 18 Pins

PIN NAME TYPE DESCRIPTION

58,59,63

91, 93, 95-97, 100, 102-104, 106 16,48,50,61,87

TM[2:0]

Bit2 : MSB

DOUT[9:0]

Bit9 : MSB

NC No connection

Supply Voltages ; 48 Pins

PIN NAME TYPE DESCRIPTION

8, 13, 17, 25, 42, 45, 49,60,64,65,73,78, 79,86,94,101,105, 109,118

11,15,19,21,29,40,

VDD Digital positive supply voltage(3.3V)

VSS Digital negative supply voltage(ground)

47,51,54,62,67,70, 76,81,82,88,92,98 99,107,113,115

1,36,122 33,34,120,127

AVDD Analog positive supply voltage(3.3V) AVSS Analog negative supply voltage(ground)

GDC21D003

I Tmode;

“111” : normal mode with ADC output “011” : normal mode with Phase Tracker output others : reserved for chip test Data output; Either ADC or Phase Tracker block

output is selected by tmode as dout[9:0] output.

Page 15

GDC21D003

4.3 Pin Assignment

PIN NAME TYPE PIN NAME TYPE PIN NAME TYPE

1 AVDD 44 DATAPOLP I/O 87 NC 2 REF I 45 VDD 88 VSS 3 TRST I 46 SEGSYNCLOCK I/O 89 VSBCLK O 4 TMS I 47 VSS 90 NSEGSYNC O 5 TCK I 48 NC 91 DOUT[0] O 6 TDI I 49 VDD 92 VSS 7 TDO O 50 NC 93 DOUT[1] O 8 VDD 51 VSS 94 VDD

9 SCL I 52 GUP I/O 95 DOUT[2] O 10 SDA I/O 53 GDN I/O 96 DOUT[3] O 11 VSS 54 VSS 97 DOUT[4] O 12 VCXO I 55 NFSYNC O 98 VSS 13 VDD 56 NCHGUP O 99 VSS 14 ADCCLK O 57 NCHGDN O 100 DOUT[5] O 15 VSS 58 TM[2] I 101 VDD 16 NC 59 TM[1] I 102 DOUT[6] O 17 VDD 60 VDD 103 DOUT[7] O 18 CLKFS O 61 NC 104 DOUT[8] O 19 VSS 62 VSS 105 VDD 20 DIN[9] I 63 TM[0] I 106 DOUT[9] O 21 VSS 64 VDD 107 VSS 22 DIN[8] I 65 VDD 108 NCOUPDTWIN O 23 DIN[7] I 66 FOE O 109 VDD 24 DIN[6] I 67 VSS 110 EQSTAT O 25 VDD 68 VSBDATA[0] O 111 NINITPH I 26 DIN[5] I 69 VSBDATA[1] O 112 NFREEZEPH I 27 DIN[4] I 70 VSS 113 VSS 28 DIN[3] I 71 VSBDATA[2] O 114 NADTONDATA I 29 VSS 72 VSBDATA[3] O 115 VSS 30 DIN[2] I 73 VDD 116 NINITEQ I 31 DIN[1] I 74 VSBDATA[4] O 117 NFREEZEEQ I 32 DIN[0] I 75 VSBDATA[5] O 118 VDD 33 AVSS 76 VSS 119 I2CSEL I 34 AVSS 77 VSBDATA[6] O 120 AVSS 35 PLLEN I 78 VDD 121 COM I 36 AVDD 79 VDD 122 AVDD 37 CLK4FS I/O 80 VSBDATA[7] O 123 REFN I 38 NRESET I 81 VSS 124 REFP I 39 NI2CEN I 82 VSS 125 INN I 40 VSS 83 NVSBERRFLG O 126 INP I 41 SYMCLK I 84 VSBDVALID O 127 AVSS 42 VDD 85 VSBSOP O 128 BIAS I 43 DATAPOLN I/O 86 VDD

Page 16

GDC21D003

5. I2C Bus I/F & Registers

5.1 I2C Bus I/F Description

When NI

may be controlled over I2C bus interface which consists of two signals, serial data(SDA) and serial clock(SCL) that can control a large number of devices on a common bus. The Device Address of this chip is “1011001”b or “0001110”b which can be selected by I2CSEL pin. The data on the I2C bus can be transferred at a rate up to 100 kbits/s in the standard mode, or up to 400 kbits/s in the fast mode. In the GDC21D003, SDA is bi-directional but SCL is only used as input, since the IC can only act as a slave device. In normal operations, data transfers are clocked by the SCL signal with one SCL pulse per data bit, and SDA is required to be stable during the high period of the SCL signal. Transitions of SDA while SCL is high are performed by the interface signals of start(S), stop(P), and repeated start(Sr) conditions. The start condition is defined as a high-to-low transition of SDA while SCL is high, and the stop condition is the low-to-high transition of SDA while SCL is high. Data transmissions are always proceeded by a start condition and ended with a stop condition, and may contain repeated starts within the transmission to alter the direction of the data flow or to change register base addresses. All data transmission

CEN pin is set to Low, the GDC21D003

operations occur in 8-bit blocks with each block acknowledged through the designated receiver by the generation of an acknowledge signal(A). This signal is generated on the ninth pulse of SCL for each transferred block.

5.1.1 Write Operation

In order to perform a write operation, the interface

is accessed in following manner. The master first generates a start condition by pulling SDA down to low while SCL is high. The master next sends a 7bit Device Address and a one bit R/W signal, and each slave compares this address with its own address and acknowledges the master if the device address sent by the master coincides with that of its own. If not so, the slave ignores the rest of current data being transmitted. If the master is writing to the GDC21D003, the chip interprets the next data byte as a register base address. This is used as the location to store the next received data byte. This base address increases as each data byte is received allowing a contiguous register block to be programmed in a single transmission. Noncontiguous blocks may be programmed in multiple transmissions or by using a repeated start condition, which allows a new Device Address and register base address to be specified without the master giving up control of the bus. The transmission is terminated with the receipt of a stop condition.

A BASE ADDRESS A DATA #1

ISSUED BY MASTER

SDA

DEV. ADDRESS W

Figure 5.1.1 I

5.1.2 Read Operation

Read operation is performed in a manner similar

to write operation. The master first generates a start condition and then sends the Device Address and R/W signal. The master will acknowledge each byte as receiving if it desires another byte to be sent. At the end of the transmission, the master will not acknowledge the slave and will then be

ISSUED BY GDC21D003

C Write Operation Example

free to generate a stop condition to terminate the transmission. The base address register contents are used to determine the location to be read, and once again this address will be increased with each successive read. Because the base address register can only be programmed through a write operation, a general read will require two accesses or a single access with a embedded repeated start to change the direction of transmission.

…

DATA #N A

Page 17

GDC21D003

P S DEV. ADDRESSDEV. ADDRESS

DEV. ADDRESS

DATA #2

ISSUED BY GDC21D003

ASDA SrW

…

R DATA #1

DATA #N

SDA

DEV. ADDRESS

A BASE ADDRESS

DATA #1

A BASE ADDRESS

ISSUED BY MASTER

C Read Operation Example

Figure 5.1.2 I

5.2 I2C Bus Register Configuration

Add

ress

D7 D6 D5 D4 D3 D2 D1 D0

Dinmode Dinsel ADCCLKSEL ADCCLKPH DCbypass DChold AGChold AGCoffsetW 11110000

1 2

VSBmod[2:0] “ 10010 ” 10110010

3 4 5

“ 00 “ PolarityW Polarity “ 10 “ nSyncLockrst VSBmodW 00001000

VCXOSEL[1:0] nCombW nComb Combouthalf NoCombgain[1:0] “ 0 “ 00011010

6 7 8

X nSyncLock nSegLock Combstat nVSBmodstart DATAPOLN nPolLock nFldLock

9 10 11

X nFrmLock nVSBLock “ 0101 “ XXXX0101

X VSBmodA[2:0]

12 13 14 15

nSyncLockPH U U nCombPH nSyncLockEQ nDSsycnEQ nFsyncEQ nCombEQ

nFreezePHI2 InitPHI2 PHASmodeIN[1:0] nFreezeEQI2 InitEQI2 STEPsizeIN[1:0] 10001011

EQmodeIN TRAINmodeIN[1:0]

nIIR16ONIN nIIRONIN nAdtOnDataI2 BLIDmodeIN[1:0] nRingENIN nCoefRead nMakeRingIN 10100101

35 36

PredicIN[1:0] LoopgainIN[2:0] nOPERmodeIN nDNgainIN nDNgainThIN 11000011

DATA BYTE

AGCoffset[7:0] 01100000

“ 11001000 “ 11001000

DCvalue[7:0]

X X

X nCombLock

CombOutHalfI N

TapAddress[7:0] 00000000

nDSadptIN nEQoutIN FLTtestIN[1:0] 00010000

A P

Initial Value

Page 18

GDC21D003

Add

ress

38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 64 65 66 67

128

DATA BYTE

D7 D6 D5 D4 D3 D2 D1 D0

WrCoef[11:8] RdCoef[11:8] 0000XXXX

WrCoef[7::0] 00000000

RdCoef[7:0]

UpdtRngIN[9:8] MeanErrINE[18:16] MeanErrOUTE[18:16] 00XXXXXX

UpdtRngIN[7:0] 10010000

MeanErrINE[15:8]

MeanErrINE[7:0]

MeanErrOUTE[15:8]

MeanErrOUTE[7:0]

MenErrOUTP[18:16] UPlimitIN[9:8] UDlimitIN[9:8] DCinformRD[8] XXX0111X

MenErrOUTP [15:8]

MenErrOUTP [7:0]

UPlimitIN[7:0] 00000000 UDlimitIN[7:0] 00000000

DCinformRD[7:0]

U GAINcnt[2:0] “ 1101 “ XXXX1101

Pase Viterbi_on Deint_on RSdec_on Derand_on Errorflag_ins Vsbdvalid_pol Vsbclk_sup 11111111

Err_count[7:0]

U U U Data_out_en X XXX1XXXX

Err_count[15:8]

“ 0 “ X 0XXXXXXX

Initial Value

Where U: unused register bit, X: don’t care

Page 19

5.3 I2C Bus Register Description

Address 0:

Most significant bit (MSB) inversion control signal of data input (DIN[9:0]).

7 Dinmode W

6 Dinsel W

5 ADCCLKSEL W

4 ADCCLKPH W

3 DCbypass W DC remove block bypass (active high). Initial value is ‘ 0’. 2 DChold W DC remove block hold (active high). Initial value is ‘0’ . 1 AGChold W AGC block hold (active high). Initial value is ‘0’. 0 AGCoffsetW W AGC offset write enable (active high). Initial value is ‘0’.

Address 1:

AGCoffset

[7:0]

If data input form is unsigned, the MSB of digital data input should be inverted because all of functions in this chip use their complement data. If this bit is set to ‘1’, it indicates the inversion of MSB. Initial value is ‘ 1’ . (refer to table 6.3.1) Digital data input path selection signal. If this bit is set to ‘1’, it indicates output of the internal ADC. Initial value is ‘1’. (refer to table 6.3.1) ADC Clock Select. When the frequency of VCXO is 2fs(21.52MHz), the output frequency of ADCCLK can be one of the two following frequencies, fs(10.76MHz) and 2fs. If this bit is set to ‘1’, the frequency of ADCCLK is always fs. Initial value is ‘1’. (refer to table 6.2.1) ADC Clock Phase Select. This signal can choose one of the ADCCLK output phases. If this bit is set to ‘0’, the ADCCLK output phase is rotated 180° off with respect to CLKFS phase, and otherwise 0°. Default value is ‘1’. (refer to table

6.2.1)

AGC offset value. If AGCoffsetW is set to ‘1’, this signal is used for the

reference of AGC block. Default value is “01100000”.

GDC21D003

Address 2:

[7:5]

VSBmod[2:0] W

[4:0] W Initial value is “10010”. It would be better set to “10110”.

VSB mode signal. If VSBmodW is set to ‘1’, this signal is used for VSB mode signal. Otherwise the VSBmod[2:0] signal is generated internally. Initial value is “101”.

Address 3:

[7:0] W Always set to “11001000”.

Page 20

Address 4:

[7:6] W Initial value is “00”. It would be better set to “01”.

5 PolarityW W

4 Polarity W

[3:2] W Always set to “10”.

1 nSyncLockrst W

0 VSBmodW W

Polarity signal write enable. If this bit is set to ‘1’, it means write enable. Initial value is ‘0’. Polarity signal for polarity control and DATAPOLP/DATAPOLN. If PolarityW is ‘1’, this signal is used for polarity control and the generation of DATAPOLP/DATAPOLN signal output. Otherwise the polarity control block uses internally calculated signal. Initial value is ‘0’. (refer to table 6.3.2)

nSyncLock reset control signal. If this bit is set to ‘0’, nSyncLock signal isn't initialized by the change of VSB mode. Otherwise, nSyncLock signal is initialized and changed to ‘0’ for the next Field sync duration. Initial value is ‘0’. VSB mode write enable. If this bit is set to ‘1’, it means write enable. Initial value is ‘0’.

Address 5:

VCXO Selection. These pins should be set as follows according to the output

[7:6]

5 nCombW W

4 nComb W

3 Combouthalf W

[2:1]

0 W Always set to ‘0’.

VCXOSEL [1:0]

NoCombgain [1:0]

frequency of VCXO. VCXOSEL[1:0] The output frequency of VCXO

00 fs(10.76Mhz) 01 2fs(21.52Mhz)

Initial value is “00”. nComb signal write enable. If this bit is set to ‘1’, it means write enable. Initial value is ‘0’. Comb filter ON/OFF signal. If nCombW are ‘1’, this signal is used for Comb filter ON/OFF signal. Otherwise the Comb filter ON/OFF signal is generated internally. Initial value is ‘1’. Comb filter output gain selection signal. If this bit is set to ‘0’, the gain of the Comb filter output is 1. Otherwise its gain is 1/2. When Comb filter is activated this signal is valid. Initial value is ‘1’. NoComb path gain selection signal. The gain is as follows; NoCombgain[1:0] the gain of normal path “00” 1(0dB)

“01” 1.125(1.023dB) “10” 1.1875(1.493dB) “11” 1.25(1.938dB) Initial value is “01”.

GDC21D003

Address 6:

[7:0] DCvalue[7:0] R Calculated DC value of input data.

Address 7:

[7:0] R don’ t care

Page 21

GDC21D003

Address 8:

7 R don’t care

Stability Indication of Data Sync Recovery block (active low). If the value of

6 nSyncLock R

5 nSegLock R Stability Indication of Data Segment Sync Recovery (active low). 4 Combstat R

3 nVSBmodstart R 2 DATAPOLN R Inverted polarity signal of input data.

1 nPolLock R Stability Indication of Polarity Decision (active low). 0 nFldLock R Stability Indication of Field Sync Recovery (active low).

Address 9:

[7:0] R don’ t care

Address 10:

[7:6] R don’ t care 5 nFrmLock R Stability Indication of inverted/non-inverted Field decision (active low). 4 nVSBLock R Stability Indication of current VSB mode detection (active low).

this bit is ‘0’, Data Segment Sync Recovery and Field Sync Recovery blocks are stable.

Comb filter ON/OFF status. If the value of this bit is ‘0’, it indicates the Comb filter is ON. Start indication of VSB mode detector which in the Equalizer. If the value of this bit is ‘1’, it means detector is reset.

Address 11:

[7:3] R don’ t care [2:0]

VSBmodA[2:0]

R Internally decided VSB mode.

Address 12, 13:

[7:0] R don’ t care [7:0] R don’ t care

Address 14:

[7:1] R don’ t care 0 nCombLock R Stability Indication of Comb filter ON/OFF decision (active low).

Address 15:

[7:0] R don’ t care

Address 32 :

7 nSyncLockPH R the state of the nSyncLock at the output of Phase Tracker 4 nCombPH R indicates whether comb filter is on or not at the output of Phase Tracker 3 nSyncLockEQ R the state of the nSyncLock at the output of Equalizer 2 nDSsyncEQ R the state of Data Segment Sync at the output of Equalizer 1 nFsyncEQ R the state of Field Sync at the output of Equalizer 0 nCombEQ R indicates whether Comb filter is on or not at the output of Equalizer

Page 22

Address 33 :

7 nFreezePHI2

6 InitPHI2

PHASmodeIN

[5:4]

[1:0]

3 nFreezeEQI2

2 InitEQI2

STEPsizeIN

[1:0]

GDC21D003

‘0’ : Freezes the Phase Tracker in the device, which means phase tracking does not occur. ‘1’ : normal operation

W/R

If you want to control Phase Tracker fast, use external input pins. Initial value is ‘1’. ‘1’ : Initialize the Phase Tracker in the device. ‘0’ : normal operation

W/R

If you want to control Phase Tracker fast, use external input pins. Initial value is ‘0’. There are three loops in the Phase, which are gain, offset, and phase loop Tracker. 00 : all loops on 01 : offset loop off

W/R

10 : offset and gain loops off 11 : all loops off Initial value is “00”. ‘0’ : Freezes the Equalizer in the device, which means coefficient update does not occur.

W/R

‘1’ : normal operation If you want to control Equalizer fast, use external input pins. Initial value is ‘1’. ‘1’ : Initializes the Equalizer in the device. ‘0’ : normal operation

W/R

If you want to control Equalizer fast, use external input pins. Initial value is ‘0’. There are three available step-sizes in the Equalizer. 10,11 : smallest step-size

W/R

01 : middle step-size 00 : largest step-size Initial value is “11”.

Page 23

Address 34 :

7 EQmodeTIN

[6:5]

4 CombOutHalfIN

3 nDSadptIN

2 nEQoutIN

[1:0] FLTtestIN[1:0]

TRAINmodeIN [1: 0]

GDC21D003

Updating range of the training sequence. The range value can be changed with

W/R

combination of these three bits. Initial value is “000”. EQmodeTIN is ‘0’ 00 : 574 symbols of field sync are used for equalization. 01 : 637 symbols of field sync are used for equalization. 10 : 700 symbols of field sync are used for equalization.

W/R

11 : 820 symbols of field sync are used for equalization. EQmodeTIN is ‘1’ 00 : 574 symbols of field sync are used for equalization. 01 : 637 symbols of field sync are used for equalization. 10,11 : 700 symbols of field sync are used for equalization. Comb filter output gain selection signal. ‘0’ : the gain of the Comb filter output is 0 ‘1’ : the gain of the Comb filter output is 1/2

W/R

When Comb filter is activated this signal is valid. Initial value is ‘1’. ‘0’ : uses data segment during equalization. ‘1’ : not uses data segment during equalization.

W/R

Default value is ‘0’. ‘0’ : noise removed output from Equalizer

W/R

‘1’ : bypassed output Initial value is ‘0’. The location of center tap can be changed using these two bits. Initial value is “00”. 00 : Center tap is 32nd tap

W/R

01 : Center tap is 44th tap 10 : Center tap is 52nd tap 11 : Center tap is 60th tap

Page 24

Address 35 :

7 nIIR16ONIN

6 nIIRONIN

5 nAdtOnDataI2

[4:3]

2 nRingENIN

1 nCoefRead

0 nMakeRingIN

BLNDmodeIN [1:0]

Address 36:

[7:6] PredicIN[1: 0]

[5:3]

2 nOPERmodeIN

1 nDNgainIN 0 nDNgainThIN

LOOPgainIN [2 : 0]

GDC21D003

‘0’ : feedback filter is on in 16 VSB mode

W/R

‘1’ : feedback filter is off Initial value is ‘1’. ‘0’ : feedback filter is on in 2, 4 and 8 VSB mode ‘1’ : feedback filter is on only in 8 VSB mode

W/R

Initial value is ‘0’. ‘0’ : coefficient adaptation during training sequence and data interval ‘1’ : coefficient adaptation during training sequence interval only

W/R

Initial value is ‘1’. If you want to control Equalizer fast, use external input pins. “00” : does not use blind equalization “01” : uses blind equalization with 4-level data

W/R

“10”, “11” : uses blind equalization with 2-level data Initial value is “00”. ‘1’ : can not change nMakeRingIN ‘0’ : can change nMakeRingIN

W/R

Initial value is ‘1’. ‘0’ : read coefficient

W/R

‘1’ : write coefficient Initial value is ‘1’. ‘0’ : can read and write the coefficients ‘1’ : normal operation

W/R

Initial value is ‘1’.

Determines whether to use slice predictor in Phase Tracker. Initial value is “11”. “00” : Slice Prediction is OFF.

W/R

“01” : not use. “10” : not use. “11” : Slice Prediction is ON. Determines use of automatic gain routine and type of loop gain to be used in Phase Tracker. Initial value is “000”. “000” : Automatic gain change. “001” : phase tracker is OFF.

W/R

“010” : smaller gain. “011” : normal gain. “1xx” : not use. Sets the operation mode ‘0’ : -60° ~ 60°

W/R

‘1’ : -45° ~ 45° Initial value is ‘0’. Choose the value of loop gain. Initial value is ‘1’.

W/R

Choose the threshold value of loop gain when gain loop is used in automatic

W/R

mode. Initial value is ‘1’.

Address 37:

TapAddress

[7:0]

[7: 0]

Filter tap address in Equalizer. Initial value is “00000000”. address 0 to address 63 : feed forward filter

W/R

address 64 to address 255 : feed back filter

Page 25

Address 38:

[7:4] WrCoef[11: 8] [3:0] RdCoef[11: 8] R Coefficients to be read from the Equalizer filter.

Coefficients to write to the Equalizer. Default value is “0000”.

W/R

Address 39:

[7:0] WrCoef[7: 0]

Coefficients to write to the Equalizer filter. Initial value is “00000000”.

W/R

Address 40:

[7:0] RdCoef[7: 0] R Coefficients to be read from the Equalizer filter.

Address 41:

[7:6]

[5:3]

[2:1]

UpdtRngIN[9: 8]

MeanErrINE [18: 16] MeanErrOUTE[ 18: 16]

Data range to be updated when nAdtOnDataI2 is ‘0’. Initial value is “00”. This

W/R

is 10-bit number.

R Mean squared error at the input of equalizer. This is 19-bit number.

R Mean squared error at the output of equalizer. This is 19-bit number.

Address 42, 43, 44, 45, 46:

[7:0]

UpdtRngIN[7:0]

MeanErrINE [15 : 8] MeanErrINE [7 : 0] MeanErrOUTE[ 15: 8] MeanErrOUTE[ 7: 0 ]

Data range to be updated when nAdtOnDataI2 is set Initial value is

W/R

“10010000”.

Mean squared error at the input of Equalizer

Mean squared error at the output of Equalizer

GDC21D003

Address 47:

[7:5]

[4:3]

[2:1] UDlimitIN[9 : 8]

0 DcinformRD[8] R Information of DC value. This shows the current DC value in DC reduction. 0 DcinformRD[8] R Information of DC value. This shows the current DC value in DC reduction.

MeanErrOUTP [18 : 16] UPlimitIN [9 : 8]

R Mean squared error at the output of Phase Tracker. This is 19-bit number.

10-bit 2’s complementary number and this should be positive number. If error of

W/R

equalizer is larger than this limit, forces error to zero. Default value is “01”. 10-bit 2’s complementary number and this should be negative number. If error of equalizer is smaller than this limit, it is forced to be zero. Default value is

W/R

“11”.

Page 26

Address 48, 49, 50, 51, 52:

[7:0] [7:0]

[7:0] UPlimitIN[7 : 0]

[7:0]

MeanErrOUTP[15 : 8] MeanErrOUTP

[7:0]

UDlimitIN[7 : 0] DcinformRD

[7:0]

Mean squared error at the output of Phase Tracker

2’s complementary number. If error of Equalizer is larger than this limit, it is

W/R

forced to be zero. Initial value is “00000000”. 2’s complementary number. If error of Equalizer is smaller than this limit, it is

W/R

forced to be zero. Initial value is “00000000”.

R Information of DC value. This shows the currently DC value in DC reduction.

Address 53:

Shows the type of loop gain used in Phase Tracker. These bits are the results of gain loop setting in LOOPgainIN

[6:4] GAINcnt [2:0] R

[3:0] W Always set to “1101”.

“001” : phase tracker is OFF. “010” : smaller gain is used. “011” : normal gain is used.

Address 64:

Parallel/serial output selection

7 Pase W

6 Viterbi_on W

5 Deint_on W

4 RSdec_on W

3 Derand_on W

2 Errorflag_ins W

1 Vsbdvalid_pol W

0 Vsbclk_sup W

‘1’ : parallel ‘ 0’ : serial If this bit is set to ‘0’, VSBDATA[0] pin is used as serial data output and VSBDATA[7] is used as start bit indicator of a byte. Initial value is ‘1’. Viterbi Decoder on/off selection ‘1’ : on ‘ 0’ : off If this bit is set to ‘0’, hard decision decoding is performed instead of viterbi decoding. Initial value is ‘1’. Deinterleaver on/off selection ‘1’ : on ‘ 0’ : off If this bit is set to ‘0’, deinterleaver is bypassed. Initial value is ‘ 1’ . RS Decoder on/off selection ‘1’ : on ‘ 0’ : off If this bit is set to ‘0’, RS decoder is bypassed. Initial value is ‘ 1’ . Derandomizer on/off selection ‘1’ : on ‘ 0’ : off If this bit is set to ‘0’, derandomizer is bypassed. Initial value is ‘ 1’ . Error flag bit insertion on/off selection. Valid only when Derand_on is set to ‘1’. ‘1’ : MSB of first data byte is set to ‘ 1’ when the packet has an uncorrected errors(when NVSBERRFLG is ‘0’) ‘0’ : nothing is done at the MSB of first data byte although the packet has an uncorrected errors(when NVSBERRFLG is ‘0’) Initial value is ‘1’. VSBDVALID polarity indicator ‘1’ : VSBDATA[7:0] is valid at VSBDVALID = ‘ 1’ interval and invalid at VSBDVALID = ‘0’ interval. ‘0’ : polarity is inverted Initial value is ‘1’. VSBCLK suppression indicator ‘1’ : VSBCLK is not suppressed at VSBDVALID = “invalid” interval ‘0’ : VSBCLK is suppressed(set to 0) at VSBDVALID = ”invalid” interval Initial value is ‘1’.

GDC21D003

Page 27

Address 65:

[7:0]

Err_count[7 : 0]

R Lower 8 bits of Segment error counter output

Address 66:

Tri-state data out

4 Data_out_en W

[3:0] R don’ t care

‘1’ : normal operation ‘0’ : high impedance Initial is ‘1’.

Address 67:

[7:0]

Err_count[15 : 8]

R Upper 8 bits of segment error counter output

Address 128

7 W Always set to ‘0’. [6:0] R don’ t care

GDC21D003

Page 28

GDC21D003

6. Functional Description

6.1 ADC

The ADC is a 10-bit 10.76Msps analog-to-digital converter. The ADC takes samples of the differential analog input signals at rising clock edge and converts them into digital values. The ADC consists of four main function blocks; SHA block, A/D & D/A block, Correction Logic block, and Output Buffer block. The detailed description is as follows.

INP

INN

REFP REFN COM

BIAS

REF

SYMCLK

SHA SHA SHA SHA SHAGain Gain Gain Gain

A/D D/A

The SHA(sample & hold amplifier) block samples the differential analog inputs at rising clock edge. And the following A/D & D/A(sub-A/D converter & multiplying D/A converter) block compares the input signal with the reference voltage and multiplies the residue signal as the determined gain. These processes are repeated in the following stage. The digital outputs are generated at each stage, corrected in the Correction Logic block, and come out through the Output Buffer block. Figure 6.1.1 shows the block diagram of ADC.

A/D D/A

Correction Logic

Output Buffer

A/D

D/A

A/D

data output

(to synchronizer block)

Figure 6.1.1 Block Diagram of ADC

Page 29

6.1.1 Electrical Characteristics DC & AC Characteristics

(Temp = 25°C, DVDD = 3.3V, AVDD = 3.3V, COM = 1.5V)

Symbol Parameters Conditions Min. Typ. Max. Measure

INN,INP

Analog Input Voltage Differential

1.7± 0.5

Normal Operation Fs = 10.76 Msps

Input = 1.7± 0.5V

INL

Integral Non-Linearity

Fs = 10.76 Msps Input = 1.7± 0.5V (Fin*=12.2KHz sine)

DNL

Differential Non-Linearity

Fs = 10.76 Msps Input = 1.7± 0.5V (Fin =12.2KHz sine)

SNR

Signal-to-Noise Ratio Fs = 10.76 Msps

Fin = 100KHz

Sampling Clock Frequency

Ci**

Input Pin Capacitance INP = 1.7V

INN = 1.7V Reference Resistance 12 Reference

***

REFP

ref

Top Voltage

REFN

Reference Bottom

***

Voltage

(NOTE) * Fin is input frequency. ** Ci = 1pF(pin capacitance) + 5pF(I/O pad capacitance) + 4.5pF(capacitance at the input sampling) *** V

REFP

, V

values should be used as a pair.

REFN

1 2 Vpp

25 mA

± 4 ±6

± 1 ± 2

44 48 dB

10.76 MHz

10 pF

2.15 2.2 2.4 V

1.15 1.2 1.4 V

GDC21D003

Unit

Value

LSB

kΩ

Analog Input Voltage Level vs. Digital Output Code

Input Voltage Level Step

REFP

. . . . . .

REFN

1023

. 512 511

MSB Digital Output Code LSB

1 1 1 1 1 1 1 1 1 1

1 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1

0 0 0 0 0 0 0 0 0 0

. .

Page 30

Recommended Operating Range

GDC21D003

SYMBOL PARAMETERS MIN MAX

AVDD DGND-AGND* REFP REFN R

(NOTE) * The difference between analog ground and digital ground should be less than 100 mV

Power Supply

Reference Top Voltage 2.15 2.4 V Reference Bottom Voltage 1.15 1.4 V Clock Duty Ratio 40 60 %

3.15 3.45 V 0 100 mV

MEASURED

VALUE

UNIT

Page 31

6.1.2 Timing Diagram

SYMCLK

GDC21D003

Analog Input

Data Output

N + 2

N + 1

N-3

TDL : Output Delay

: Analog Input Sampling Point

¡Ü

N-2

Figure 6.1.2 Timing Diagram of ADC

N + 3 N + 4

N-1

TDL =3ns

N+1

Page 32

6.1.3 Application Circuits Application Circuits

GDC21D003

SANYO

LA7785M

IOUT+ IOUT -

GUP

GDN POLN POLP

20 22

3.3V

1.5K

3.3K

2.0K

3.3V

0.1uF

3.3V

3.3K

2.7K

2.2V

1.2V

3.3V

12K

0.1uF

1.5V

Differential

(1.7 0.5V)

3.3V

0.1uF

125

INP

126

INN

GUP

GDN

DATAPOLN

DATAPOLP

GDC21D003

124

REFP

123

REFN

REF

121

COM_0

128

BIAS

AVDD AVSS

AVDD = 3.3V, AVSS = 0V DVDD = 3.3V, DVSS = 0V

Figure 6.1.3 ADC Application Circuit

Page 33

Equivalent Circuits for Bias and Input Pins

GDC21D003

REFP,REFN COM

Differential

(1.7¡¾0.5V)

1.7V

2.2V

1.2V

2.2V

1.2V

VDD

330~700 ohm

22pF

INP

INN

VDD

REF

BIAS

450~700 ohm

22pF

VDD

AVDD

3.6K

6.3K

10bit A/D Converter

100uA

Figure 6.1.4 Equivalent Circuits

Page 34

GDC21D003

6.2 Clock Divider

VSB receiver(GDC21D003) can use external VCXO which outputs symbol frequency and its multiple frequency because it has internal clock divider. Multiples can be 2 times. Also, this block has 2 output signals, CLKFS and ADCCLK. CLKFS is used as symbol clock for DTV transmitter, ADCCLK is used as A/D converter’s clock when external A/D converter is used for

Table 6.2.1 Register Setting for Clock Divider

Frequency of

external

VCXO

10.76MHz “00” ‘0’ ‘0’ 10.76MHz

10.76MHz “00” ‘0’ ‘1’ 10.76MHz

10.76MHz “00” ‘1’ ‘0’ 10.76MHz

10.76MHz “00” ‘1’ ‘1’ 10.76MHz

21.52MHz “01” ‘0’ ‘0’ 21.52MHz

21.52MHz “01” ‘0’ ‘1’ 10.76MHz

21.52MHz “01” ‘1’ ‘0’ 21.52MHz

21.52MHz “01” ‘1’ ‘1’ 10.76MHz

VCXOSEL[1:0]

(in I2C bus

register5)

ADCCLKPH

(in I2C bus

register0)

digital input. The phase of ADCCLK against CLKFS can vary through I2C bus. When external VCXO output frequency is 2 times(21.52MHz) of symbol frequency, the output frequency of ADCCLK signal becomes the same as symbol frequency(10.76MHz) or 2 times of symbol frequency(21.52MH). ADCCLK should be set through I2C bus according to output frequency of external VCXO. The setting of each case is described in following tale 6.2.1.

ADCCLKSEL

(I2C bus

register0)

ADCCLK

regarding

(phase: 180o)

(phase: 0o)

(phase: 180o)

(phase: 0o)

(phase

CLKFS)

CLKFS

10.76MHz

Page 35

6.3 Synchronizer

GDC21D003

6.3.1 Input Control

VSB receiver(GDC21D003) use 10bits signal generated in internal A/D converter or

INN INP

A/D Converter

DIN[9:0]

Input Selection

Figure 6.3.1 The Block Diagram of Input Selection

Figure 6.3.1 is internal block diagram of Input Selection. You have to choose either the output of internal A/D converter or that of external A/D converter as digital input signal (Dinsel). If you decide to use the output of external one,

A/D Converter

MSB Control

Internal

Dinmode Dinsel

10bits(DIN[9:0]) external data input to perform digital processing. Therefore, the data input path must be set according to input before digital processing.

Digital Input Signal

Multiplexer

MSB(DIN[9]) of 10bits input (DIN[9:0]) should be set according to its output characteristics(Dinmode). Setting of Dinsel and Dinmode is performed through I2C bus and it is as following Table 6.3.1.

Table 6.3.1 Input Signal Path Setting

Input signal path Dinmode

(in I2C bus register0)

From internal A/D converter Don’ t care ‘1’

From DIN[9:0] (Signed signal) ‘0’ ‘0’

From DIN[9:0] (Unsigned signal) ‘1’ ‘0’

Figure 6.3.2 shows the relationship between chip and external device in case that internal ADC output is used for digital processing. Figure 6.3.3

device in case that ADC digital output is used for on-chip digital processing with using other external ADC instead of using internal ADC.

shows the relationship between chip and external

Dinsel

Page 36

GDC21D003

INN

ADCCLK

126

125

INP

LA7785M

(SANYO)

IOUT-

IOUT+

GDC21D003

SYMCLK

CLKFS

OPEN

VCXO

(NOTE) - Regarding external VCXO output frequency, VCXOSEL[1:0] should be set as Table 6.2.1.

- Since internal ADC is used, peripheral circuit should be set according to ADC.

- Digital input signal path should be set as Table 6.3.1 to use internal ADC.

- If VCXO output frequency is equal to symbol frequency, VCXO(pin12) can be connected to GND, and SYMCLK(pin41) and VCXO output can be connected directly.

Figure 6.3.2 I/F Circuit Diagram between VSB Receiver & Demodulator (Internal A/D Input)

Page 37

126

125

LA 7785M

GDC21D003

INN

INP

Demodulator

(SANYO)

A/D

DIN[9:0]

SYMCLK

CLKFS

GDC21D003

VCXO

(NOTE)- Regarding external VCXO output frequency, VCXOSEL[1:0] should be set as Table 6.2.1.

- Since internal ADC is used, frequency and phase of ADCCLK that is its input clock should be set as Table 6.2.1.

- Digital input signal path should be set as Table 6.3.1 to use external ADC output.

- If VCXO output frequency is equal to symbol frequency, VCXO(pin12) can be connected to GND, and SYMCLK(pin41), A/D clock, and VCXO output can be connected directly.

Figure 6.3.3 I/F Circuit Diagram between VSB Receiver & Demodulator (External A/D Input)

Page 38

GDC21D003

6.3.2 DC Reduction

Current DTV transmission system uses pilot for restoration of carrier signal. Inserted DC value at transmitter is transformed into pilot in frequency domain. In receiver, A/D is used to convert

Input Selection

DC Reduction

Figure 6.3.4 The Block Diagram of DC Reduction

Figure 6.3.4 is block diagram of DC Reduction. Output of Input Selection is sent to DC Reduction block and removes pilot as well as all DCs in analog system. And DC value is calculated from all data of input signal. Through I2C bus this block can be bypassed and DC value doesn’t be updated from current state. Among I2C bus control signals, if DC bypass signal(in I2C bus register0) is ‘0’, DC removed signal is output, and if it is ‘1’, input signal is bypassed and output. Also, when DChold signal(in I2C bus register0) is ‘0’, DC value is calculated using continuously input signals, when it is ‘1’, DC value at the moment is saved without updating DC value. Calculated DC value can be read through I2C bus. (DCvalue[7:0] in I2C bus register address6)

6.3.3 Auto Gain Control(AGC)

There are 2 AGC modes currently used, one is Non-coherent AGC mode and the other is Coherent AGC mode. Figure 6.3.5 is the block diagram of

baseband analog signal that completed the demodulation of RF signal into digital signal, A/D output has the same DC value inserted from transmitter. Also DC components arise through various analog processing.

10 10

Accumulator

DChold

Multiplexer

DCbypass

AGC. Non-coherent AGC mode is performed during initialization state and before finding the Data Segment Sync interval in receiver. During initialization state, it increases Gain continuously to take the maximum value.(GUP = ‘1’ during initialization state). Also, since at the moment of reset completion, received signal has the maximum gain, it is highly possible that A/D converter output has either maximum or minimum value. A/D converter output can also have either maximum or minimum value from too much noise within channel. When input signal value has maximum/minimum value for 8 symbols in a row, control signal (GDN = ‘1’ during 2 symbols) is generated to reduce system gain. Non-coherent AGC mode doesn’t stop before it finds timing of Data Segment Sync signal. Coherent AGC mode calculates the average of input signals and controls the GUP/GDN signal so that this average value can have the desired value.

Page 39

Input Selection

DC Reduction Accumulator

Coherent AGC

AGC offset

Figure 6.3.5 The Block Diagram of AGC

PWM

Non-Choerent

AGC

RESET

GDC21D003

GUP

GDN

GUP/GDN signal is transmitted to off-chip demodulator and controls the gain of Tuner and demodulator. Through I2C bus, AGC block can be held(When AGChold signal in register0 is ‘1’). And for input data to have desired optimum value, Coherent AGC uses reference value which can be input through I2C bus (AGCoffset[7:0] in I2C bus

TUNER

(NOTE) RFAGC signal, an input signal to Tuner should be connected referring to LA7785M.

LA7785M

(SANYO)

RFAGC10GDN20GUP

register1) or can use already defined value. If you want to change Coherent AGC reference value through I2C bus, AGCoffset[7:0](in I2C bus register1) value should be set and AGCoffsetW(in register0) signal should be set to ‘1’. If AGCoffsetW signal is ‘0’, already defined value is used as a reference value. Figure 6.6 shows the connection of chip output, GUP/GDN signal.

Figure 6.3.6 AGC Signal I/F for DTV System

6.3.4 Polarity Correction

phase) or 180ophase(out of phase). When carrier is locked in 0o phase, there is no problem, but when

For currently used demodulator algorithm, FPLL(Frequency & Phase Locked Loop) is used. Due to FPLL algorithm’s property, demodulator

carrier is locked in 180o phase, polarity of baseband data are all inverted. In this case, input data polarity should be reverted.

(carrier recovery) lock can occur in 0o phase(in

GDC21D003

(LG)

GUP52GDN

Page 40

GDC21D003

DC Reduction

Polarity Correction

Fig 6.3.7 The Block Diagram of Polarity Correction

Figure 6.3.7 is Polarity Correction block diagram. Output signal polarity is corrected by control of polarity from Polarity Decision block.

6.3.5 Data Segment Sync Recovery

At DTV transmitter part Data Segment Sync is inserted for 4-symbol period in every Data Segment. This Data Segment Sync has (1,0,0,1) pattern. DTV receiver compares inputted signal pattern with Data Segment Sync pattern (1, 0, 0, 1) inputted from transmitter and acknowledges the most similar pattern as Data Segment Sync interval. But because the inputted signal at the beginning of activation has polarity ambiguity generated by FPLL, inverted polarity pattern of Data Segment Sync (1, 0, 0, 1) from transmitter is acknowledged as Data Segment Sync interval. After completion of

Polarity Inverse

10 10

Multiplexer

polarity

initial Data Segment Sync detection, input signal polarity is detected from Polarity Decision block at the end. Also after polarity correction in Polarity Correction block, only positive polarity is acknowledged as Data Segment Sync interval. And Data Segment Sync detection becomes easier under close Ghost environment by improving its performance. Figure 6.3.8 shows the improved Data Segment Sync Recovery block diagram. At first input signal is integrated by the segment passing through Segment Correlator. Slicer extracts the information of Data Segment Sync interval from output of integration period. This information undergoes its reliability check in Confidence Counter and then creates Data Segment Sync. When reliability builds up to some level, nSegLock signal is generated to inform the completion of Data Segment Sync Recovery. This signal can read through I2C bus(nSegLock in I2C bus register8).

output of Polarity correct

Segment

Correlator

Segment

Integrator

Figure 6.3.8 The Block Diagram of Data Segment Sync Recovery

If Data Segment Sync is found in this block, input polarity is detected in the Polarity Decision block and Timing Recovery block sets SEGSYNCLOCK

nSegLock

Segment

Slicer

control signal

Confidence

Counter

nSegSync

signal to ‘1’ which informs the start of activation (Timing recovery start status). This signal is ‘0’ until it finds Data Segment Sync.

Page 41

GDC21D003

6.3.6 Polarity Decision

It decides input signal polarity after it finds Data Segment Sync timing input from transmitter and checks Data Segment Sync pattern. Data Segment Sync pattern input from transmitter is (1,0,0,1). Therefore when Data Segment Sync pattern of transmitted signal is (1,0,0,1), baseband signal has positive polarity, and when (0,1,1,0), it has negative polarity. In this block, input Data Segment Sync pattern is checked and output signal is

Table 6.3.2 DATAPOLP & DATAPOLN Output

(I2C bus register4)

After Data Segment

Sync Lock

PolarityW : ‘1’

(I2C bus register4)

generated by this result. Then it is sent to internal Polarity Correction and demodulator. Also, regardless of inputted Data Segment Sync pattern, through I2C bus, DATAPOLP and DATAPOLN are decided. The output of Polarity Decision, DAPOLP and DATAPOLN are shown in Table 6.3.2. On the other hand, DATAPOLN signal can be read through I2C bus.(DATAPOLN in I2C bus register8) Figure 6.3.9 describes the connection of DATAPOLP & DATAPOLN which is output of chip and Demodulator chip.

DATAPOLP DATAPOLN

Positive polarity ‘ 0’ ‘0’Before Data Segment Sync Lock

Negative polarity ‘0’ ‘0’

Positive polarity ‘ 1’ ‘ 0’PolarityW : ‘0’

Negative polarity ‘0’ ‘1’

Polarity : ‘0’

‘1’ ‘0’

(I2C bus register4)

Polarity : ‘1’

‘0’ ‘1’

(I2C bus register4)

LA7785M

(SANYO)

POLP

POLN

DATAPOLP

DATAPOLN

Figure 6.3.9 Polarity Signal I/F Circuit

GDC21D003

(LG)

Page 42

GDC21D003

6.3.7 Timing Recovery

In current DTV system, timing information is extracted from inputted Data Segment Sync signal and the clock used for digital processing and A/D. Timing Recovery structure used in this chip is similar to typical PLL structure. PLL generally

output of polarity correct

Correlator PWM

Digital Phase Detector

Figure 6.3.10 Timing Recovery Block

On-chip Phase Detector extracts phase information only in Data Segment Sync interval, and it is activated after finding Data Segment Sync interval inserted in transmitter. SEGSYNCLOCK(PIN46) shows whether Data Segment Sync is found or not. If this signal is ‘0’, Data Segment Sync isn’ t found, if this is ‘1’, it is found. Phase detector uses filter whose coefficient is (-1,1,1,1) because it uses Data Segment Sync pattern(1,0,0,1) to get phase information. Phase

consists of phase detector, loop filter, and VCO(VCXO). In this chip, Timing Recovery is consisted of internal Digital Phase Detector, external analog device, Charge Pump, Loop Filter and VCXO. Figure 6.3.10 is Timing Recovery block diagram.

CHGUP

CHGDN

Charge

Pump

Loop Filter

VCXO

information is converted again to PWM signal and output from chip(NCHGUP/ NCHGDN). While SEGSYNCLOCK signal is ‘0’, NCHGUP/NCHGDN signal outputs are all ‘0’. In this time, external Charge Pump is charged to freerun voltage of VCXO. When SEGSYNCLOCK signal is changed to ‘1’, NCHGUP/NCHGDN signals are all changed to ‘1’, and once per every Data Segment, charge of external Charge Pump is charged up and down.

CLOCK

Page 43

6.8K

+3.3V

0.01uF

NCHGUP

141841

VCXO

GDC21D003

Analog

S/W

10K

VCXO

(LG)

NCHGDN

Charge Pump Loop Filter

OPEN

(NOTE)- Regarding external VCXO output frequency, VCXOSEL[1:0] should be set as Table 6.2.1.

Analog

S/W

10.76MHz, 21.52MHz

- In case of using digital input by external ADC, ADCCLK frequency and CLKFS phase should be set as Table 6.2.1.

- Each discrete device value in figure is just recommended value.

- PNP type transistor can be used instead of Analog S/W.

Figure 6.3.11 Timing Recovery I/F Circuit(1)

Page 44

GDC21D003

6.8K

0.01uF

ADCCLK

CLKFS

VCXO

Phase information from chip passes through Charge Pump and is input to Loop Filter. Loop Filter output is connected to VCXO to control clock phase. Figure 6.3.11 shows the connection diagram when external VCXO output frequency is

NCHGUP

GDC21D003

(LG)

NCHGDN

equal to 1 and 2 times output frequency of symbol frequency. And Figure 6.3.12 shows the connection diagram when external VCXO output frequency is equal to that of symbol frequency.

+3.3V

Analog

S/W

10K

VCXO

Analog

S/W

Charge Pump Loop Filter

10.76MHz

OPEN

To ADC

(NOTE) - ADCCLK frequency and CLKFS phase should be set as Table 6.2.1 when digital input using external ADC is used.

OPEN

- Each discrete device value in this figure is just recommended value.

- Analog S/W should be ON when input signal (NCHGUP/NCHGDN) is ‘0’ .

- PNP type transistor can be used instead of Analog S/W.

Figure 6.3.12 Timing Recovery I/F Circuit(2)

Page 45

GDC21D003

6.3.8 Field Sync Recovery

Transmitter inserts 1 Data Segment long Field Sync every 313th segment. Inserted Field Sync structure is described in Figure 6.3.13. Field Sync is found

Segment

sync

(4 symbols)

Among three PN63 signals exceptionally the second one changes its polarity in every Field. Equalizer uses this data interval, so polarity information(FOE : pin number 66) should be detected. This polarity information detection starts

PN511

(511 symbols)

Error

Calculation

PN63

(63 symbols)

Figure 6.3.13 Field Sync Structure

PN63*

(63 symbols)

Minimum

Error

Detection

using PN511 as shown in Figure 6.3.13. Cause pattern of PN511 input from transmitter is already acknowledged, by comparing with transmitted signal, the interval having the most similar pattern is acknowledged as Field Sync interval.

PN63

(63 symbols)

VSB mode

(24 symbols)

reserved

(92 symbols)

12 symbol

data

(12 symbols)

after finding Field Sync. Used algorithm is similar to that of Field Sync case. If the second PN63 phase is the same as others, FOE signal outputs ‘0’, otherwise it outputs ‘1’.

nSyncLock

Confidence

Counter

nFsync

nFldLock

Reference

Generation

Figure 6.3.14 The Block Diagram of Field Sync Recovery

Figure 6.3.14 is the block diagram of Field Sync Recovery. Input signal calculates error value by the Data Segment, comparing with Reference Generation output. Among these error values, the smallest Data Segment is acknowledged as Field Sync interval, Confidence Counter checks its confidence and generates Field Sync. And after finding Field Sync it detects Polarity Error of the

Polarity

Error

Detection

Foe nFrmLock

second PN63 interval in Error Calculation output and generates Foe signal. Field Sync Recovery generates nVSBmodstart(in I2C bus register8) and nSyncLock(in I2C bus register8) signal. nVSBmodstart signal is used as initialization signal of VSB Mode Detect block. When this signal is ‘1’, it means the initialization state, when it is ‘0’, it means the operation state.

Page 46

GDC21D003

This signal is changed to ‘0’ after detection of the Field Sync signal. Then VSB Mode Detect is activated. nSyncLock signal is used for initialization of NTSC Rejection, Equalizer, Phase Tracker, and Channel Decoder(FEC). If this signal is ‘1’, it means the initialization state, if it is ‘ 0’ , it means the operation state. This signal is changed to ‘0’ after detection of the Field Sync signal and FOE signal. If detected VSB mode from VSB mode detect is different from existing one, NTSC Rejection, Equalizer, Phase Tracker, and Channel Decoder(FEC) are operated in new VSB mode. But since Equalizer and Phase Tracker are consisted of lots of feed back loops, these 2 blocks operated in existing VSB mode can be invalid when they suddenly should be operated on new VSB mode. In this case, nSyncLock signal can initialize again these blocks for normal operation, which may bring the increase of system Lock up time. These 2 cases have trade-off respectively that nSyncLock signal can be initialized through nSyncLockrst(in I2C bus register4) signal or not on the change of VSB mode. If nSyncLockrst signal is ‘0’, it isn’ t initialized on every change of VSB mode, if it’s ‘1’ , it is initialized.

Table 6.3.3 VSB Mode Data for each VSB Mode

VSB Mode VSB Mode Data

MMDS 2 VSB “000011110000111100001111” MMDS 4 VSB “000011110000111110010110” MMDS 8 VSB “000011110000111110100101”

MMDS 16 VSB/ATSC 16VSB “000011110000111111000011”

ATSC 8 VSB “000010100101111101011010”

6.3.9 VSB Mode Detect

After Field Sync Recovery completed, VSB mode of current transmitted signal should be searched. There are 6 types of VSB mode suggested till now, they are ATSC(terrestrial) 8 VSB, ATSC 16VSB, MMDS(cable) 2VSB, MMDS 4 VSB, MMDS 8VSB, and MMDS 16 VSB. But as ATSC 16VSB and MMDS 16VSB are identical, in fact there are 5 VSB modes. Field Sync in figure

6.3.13 has the information of current transmitted VSB mode. 24-symbol VSB Mode signal is detected in field Sync interval of input signal and VSB Mode is generated. VSB mode data on each mode is shown in Table 6.3.3. Among VSB mode data, bold letter number is original VSB mode signal, and italic is derivative signal. Mode signal of MMDS 16 VSB and ATSC 16 VSB mode signal can use “100” and “001”, but in this chip output is always “100” regardless of input.

(“000011110000111100111100”)

VSB Mode Detect block can be variously controlled through I2C bus. And internally detected VSB mode can be read through I2C bus.

Table 6.3.4 VSB Mode Signal Control

NI2CEN

(pin number 39)

‘1’ don’t care don’ t care ATSC 8VSB(“101”) ‘0’ ‘ 1’ VSBmod[2:0] VSB mod[2:0] ‘0’ ‘ 0’ don’ t care Internally detected

VSBmodW

(in I2C bus register4)

(VSBmodA[2:0] : in I2C bus register11) Table

6.3.4 shows the creation of VSB mode signal through I2C bus.

VSBmod[2:0]

(in I2C bus register2)

VSB Mode

VSB mode

Page 47

GDC21D003

6.3.10 NTSC Rejection

In the beginning of HDTV broadcasting, it will be serviced with conventional NTSC broadcasting. But in case that NTSC broadcasting exists in the same channel where HDTV broadcasting is on the air, at the HDTV signal’s point of view NTSC

10 10

Figure 6.3.15 Comb Filter Block

Figure 6.3.16 is the block diagram of NTSC Rejection. PN511 signal section contained in Field Sync helps to get MSE(Mean Square Error). Difference(error) of input signal and reference signal generated in Reference Generator is summed during PN511 signal section and difference(error) of input signal passed through Comb Filter and reference signal passed through Comb Filter is summed during PN511 section. These sums of signal difference(error) are compared and the smaller one is selected. The SNR of signal passed through Comb Filter is decreased to 3dB low.

signal is interference. This interference is called as co-channel interference. 3 carrier signals contained in this co-channel interference can be removed by using Comb Filter shown in Figure 6.3.15, but should be checked first if there is co-channel interference in current channel.

12 symbol delay

Therefore, if input signal has no co-channel interference, the sum of error without passed through Comb Filter is always smaller than the other one that Comb Filter isn’t used. But if cochannel interference exists on channel, even though the SNR of signal passed through Comb Filter is 3dB low, 3 carriers by co-channel interference are removed. Therefore the signal passed through Comb Filter has smaller sum than signal without passed through Comb Filter and it is output as a result.

Page 48

GDC21D003

Comb

Filter

Reference

Generation

Figure 6.3.16 The Block Diagram of NTSC Rejection

When NTSC co-channel interference is checked, Comb Filter ON/OFF timing can be changed through I2C bus. Its control signal is NoCombgain[1:0](in I2C bus register5) and controls this by changing error gain which doesn’t use Comb Filter. When NoCombgain[1:0] is “00”, gain is ‘1’. As a result, sum of error value during PN511 section, which doesn’t use Comb Filter is compared with the sum of the other during PN511 section, passed through Comb Filter. When it is

Table 6.3.5 Comb Filter Control through I2C Bus

MSE

Comparator

“01”, each error that doesn’ t use Comb Filter is multiplied by the gain of ‘1.125’(1.023dB) and the results are compared. When it is “10”, gain is ‘1.1875’ (1.493dB), and when “11”, gain is ‘ 1.25’ (1.983dB). By this manner, it is determined whether Comb Filter will be used or not. The signal to inform this to the next part is Combstat. And it can be read through I2C bus. (in I2C bus register8) Various controls can be done through I2C bus as shown in Table 6.3.5.

Multiplexer

Combstat

NCombW

(in I2C bus register5)

‘1’ ‘ 1’ ‘ 1’ (Comb filter OFF) ‘1’ ‘ 0’ ‘0’(Comb filter ON) ‘0’ don’t care Internally detected signal

NComb

(in I2C bus register5)

Combstat

(in I2C bus register8)

Page 49

GDC21D003

6.4 Equalizer

The equalizer compensates for linear channel distortions, such as tilt and ghosts. These distortions can come from the transmission channel or the imperfect components within the receiver. The Equalizer uses a Least-Mean-Square algorithm and decision feedback equalizer structure. The adaptive equalization is performed with training sequence, and blind equalization is also supported.

6.4.1 Block Diagram

The Equalizer has a decision feedback structure with 64-tap feed-forward filter and 192-tap feedback filter. Its internal diagram is Figure 6.4.1 and it consists of 256-tap adaptive filter, training sequence generator, slicer, delays for data storage,

I Channel Input Symbols

10.76

MS/sec.

Forward

Tap Index

Filter Taps

Delay 1

64 Tap

Forward

Filter

Tap

Storage

Tap

Storage

Estimated

Error

192 Tap

Feedback

Filter

Filter Taps

subtractor, and control block. Each tap of 256-filter tap has its own update part which is composed of tap coefficient storage, adder, and multiplier. This makes the equalizer converge into the channel condition very fast. This filter is able to initialize its coefficients through register33[2] (InitEQI2) bit and update the coefficients internally during one system clock(10.76 MHz) cycle as well as download the coefficients through I2C register by user. The outputs of feed-forward filter and feedback filter are summed to produce the output. This filter is applicable to 8 VSB terrestrial broadcasting mode and 2, 4, 8, and 16 VSB cable mode. Figure 6.4.2 shows the slicer of equalizer. This decision feedback structure with a coefficient update filter and LMS algorithm can optimize the VSB equalizer in converging time, remaining error and stability.

Equalized I Channel Output

10.76

Decision

Data

M U X

Feedback

Tap Index

Delay 2

Training

Sequence

Slicer

MS/sec.

Control

nFreezeEQ

Figure 6.4.1 Channel Equalizer

Page 50

VSB mode

GDC21D003

Equalizer Output

Figure 6.4.2 VSB Slicer

6.4.2 Training/Data Mode Equalization

The Equalizer can achieve equalization through three means: it can do adaptation on the binary training sequence; it do adaptation on data symbols throughout the frame when the eyes are open; or it

field interval

training sequence

data mode equalization

Segment

sync

(4 symbols)

PN511

(511 symbols)

PN63

(63 symbols)

training sequence mode equalization

training sequence mode equalization data mode equalization

2VSB slicer 4VSB slicer 8VSB slicer

16VSB slicer

PN63*

(63 symbols)

MUX

Slicer Output

can do adaptation on data when the eyes are closed(blind equalization). The principal difference among these three methods is how the error estimate is generated, and Figure 6.4.3 shows the difference of training and data mode equalization.

PN63

(63 symbols)

...

Segment

sync

(4 symbols)

data ...data ...

Figure 6.4.3 Training/Data Equalization

Page 51

GDC21D003

6.4.3 Error Estimation

The Equalizer uses a Least Mean Square(LMS) algorithm and can do adaptation on the transmitted binary training sequences as well as on the transmitted data. The LMS algorithm computes

-4

(a) 2VSB slice (b) 4VSB slice (c) 8VSB slice (d) 16VSB slice

Figure 6.4.4 VSB Slice Level

how to adjust the filter taps in order to reduce the current error at the output of the Equalizer, by generating an estimate of the current error in the equalizer output signal using the slice level of Figure 6.4.4.

7 5

3 1

-1

-2

-3

-5

-6

-7

15 13 11 9 7 5 3 1

-1

-3

-5

-7

-9

-11

-13

-15

: slice level

The estimated error is used in the adaptive filter to update its coefficients as following equations. One is for feed-forward adaptive filter and the other is for feedback adaptive filter. The error is

¥Ä

New j th filter tap coefficient of equalizer

Current j th filter tap coefficient of equalizer

-sizestep

Current error value of equalizer

Current data stored in j th filter tap of Feedforward

New jth filter tap coefficient of equalizer

Current jth filter tap coefficient of equalizer

-sizestep

Current error value of equalizer

Current decision data stored in jth filter tap of Feedback

¥Ä

kjkk

VeC

^1+

kjkk

IeC

multiplied by data and adjusted by step-size, and then added to the stored coefficient value to get new coefficient value.

Page 52

GDC21D003

6.4.4 Adaptive Filter

The 256-tap adaptive filter consists of two subfilters. One is feed-forward filter, and the other is feedback filter. The feed-forward filter is 64-taps long and the feedback filter is 192-taps long. The data of feed-forward filter is I channel input symbol. The data of feedback filter is training sequence or sliced filter output. The coefficients of this filter are updated internally by means of the product of the estimated error by data delayed in all taps. And the coefficients can be updated at

data

estimated error

decision data FLTout

Data Bank

Coefficient Bank /

Coefficient Updater

Data Bank 192-tap feedback filter

every clock. The multiplexing scheme can be introduced for 4 taps to share one multiplier. The number of multipliers is reduced to a quarter. The products of data and coefficients are added and accumulated to make output. The output of feedforward filter and the output of feedback filter are summed to make the filter output. These outputs are transferred to the phase tracker. And they are used to calculate the estimated error with training sequence and sliced filter output. Figure 6.4.5 shows the block diagram of the filter.

64-tap feedforward filter

Multiplier / Accumulator

memory

estimated error

memory

Coefficient Bank /

Coefficient Updater

Figure 6.4.5 Coefficient Update Filter

6.4.5 Equalizer Clock Scheme

The 256-tap filters are used in the equalizer. Generally every tap has its own multiplier and the portion of the multipliers is vast. The multiplexing scheme can be introduced for several taps to share one multiplier in this case. The 4-times

Multiplier / Accumulator

multiplexing scheme is used in our design. Therefore one multiplier is shared to 4 taps. A clock, 4-times multiplied by SYMCLK(symbol clock) is needed for multiplexing. The PLL is used for generating the 4-times multiplying clock(CLK4EQ).

Page 53

6.4.6 I2C Bus I/F

GDC21D003

I Channel Input Symbols

10.76

MS/sec.

256 tap adaptive filter

Coefficient

I2C Control

RdCoef

nCoefRead

nMakeRingEn

WrCoef

TAPaddress

Estimated Error

Figure 6.4.6 I2C Bus I/F

The adaptive filter is able to initialize and freeze its coefficients through register33[2](InitEQI2) bit and register33[3](nFreezeEQI2) bit. The control block located at the front of filter controls the estimated error value according to each operating mode and sends it to the adaptive filter. These bits are in the table 6.4.1. For example, if register33[3](nFreezeEQI2) bit is set to “low”, Equalizer Control block forces error value to be ‘0’, which makes the coefficients of the filter unchanged. The equalizer supports three different

Control

nInitEQ

nFreezeEQ

nAdtOnDataI2

Equalized I Channel Output

10.76

MS/sec.

M U X

Training

Sequence

Slicer

I2C BUS

step-sizes; the smallest, the middle, and the largest step size. The difference is converging time and remaining error. If the largest step-size is selected, equalizer converges fast but the remaining error is large. And if the smallest step-size is selected equalizer converges slowly but the remaining error is small. The decision directed mode can be used for moving ghost by just setting nAdtOnDataI2. When this signal is low, the decision directed equalization is ON. When this pin is high, the decision directed equalization is OFF.

Page 54

7 nFreezePHI2

6 InitPHI2

3 nFreezeEQI2

2 InitEQI2

[1:0]

STEPsizeIN [1:0]

Table 6.4.1 Contents of I2C Bus Register33 for Equalizer

‘0’ : Freezes the Phase Tracker in the device, which means phase tracking does not occur. ‘1’ : normal operation

W/R

If you want to control Phase Tracker fast, use external input pins. Default value is ‘1’. ‘1’ : Initializes the Phase Tracker in the device. ‘0’ : normal operation

W/R

If you want to control Phase Tracker fast, use external input pins. Default value is ‘0’. ‘0’ : Freezes the Equalizer in the device, which means coefficient update does not occur. ‘1’ : normal operation

W/R

If you want to control Equalizer fast, use external input pins. Default value is ‘1’. ‘1’ : Initializes the Equalizer in the device. ‘0’ : normal operation

W/R

If you want to control Equalizer fast, use external input pins. Default value is ‘0’. There are three available step-sizes in the Equalizer. 00,01 : smallest step size 10 : middle step size

W/R

11 : largest step size Default value is “11”.

GDC21D003

The mean squared errors are calculated at equalizer input and output, and these can be read by using I2C bus. The location of center tap could be changed using I2C bus, so the equalization range

Table 6.4.2 Contents of I2C Bus Register33 for Filter Control

‘0’ : feedback filter is on in 16 VSB mode

7 nIIR16ONIN

6 nIIRONIN

‘1’ : feedback filter is off

W/R

Default value is ‘1’. ‘0’ : feedback filter is on in 2, 4 and 8 VSB mode

W/R

‘1’ : feedback filter is on in only 8 VSB mode Default value is ‘0’.

could be adjusted according to channel condition. If you set nIIR16ONIN or nIIRONIN, the decision feedback filter is turned off for 2, 4, and 16 VSB mode.

Page 55

GDC21D003

6.4.7 Coefficient Reading/Writing

The adaptive filter is able to initialize and freeze its coefficients as well as download the coefficients through I2C register by user. The following are coefficient reading/writing routines through I2C register, and the I2C control block in figure 6.4.6 supports this coefficient reading/writing function

Equalizer / Phase Tracker Coefficient Write Routine

1) Freeze an equalizer and a phase tracker. (Sets nFreezeEQI2 and nFreezePHI2 to ‘0’)

2) Sets the nCoefRead to a write mode. (set this bit to ‘1’)

3) Writes the filter tap address to write the coefficients.(TapAddress : Register37)

4) Write the coefficients. (WrCoef : Register38, 39)

5) Sets the nMakeRingIN to ‘0’.

6) Sets the nMakeRingIN to ‘1’.

7) To write another coefficients, goes to the 3rd step.

8) Sets nFreezeEQI2 and nFreezePHI2 to HIGH and quits the write routine.

Equalizer Coefficient Read Routine

1) Freeze an equalizer and a phase tracker.

(Sets nFreezeEQI2 and nFreezePHI2 to ‘0’)

2) Set the nCoefRead to a read mode. (Set this bit to ‘0’)

3) Writes the filter tap address to read the coefficients. (TapAddress : Register37)

4) Set the nMakeRingIN to ‘0’.

5) Read the coefficients.(RdCoef : Register38, 40)

6) Set the nMakeRingIN to ‘1’.

7) To read another coefficients, goes to the 3rd step.

8) Set nFreezeEQI2 and nFreezePHI2 to HIGH and quit the read routine.

6.5 Phase Tracker

The Phase Tracker corrects the untracked phase noise in the receiver. It uses slice prediction information that comes from the Viterbi Decoder so that this phase tracker increases its performance in severe noise environment. There are three loops; gain correction loop, offset correction loop, and phase correction loop. Phase Error Correction block corrects the distortion caused by a phase noise using an estimated error in the output of the Phase Tracker. User can control the gain of Phase Tracker by using I2C register. There is a trade-off relation between gain and noise enhancement in phase tracker. So the phase tracker in this device change the gain according to the noise condition. To increase the performance the phase tracker is using the slice predict information that comes from viterbi decoder. Phase tracking range is -60° ~ 60° with resolution of 0.004°. This means that phase error estimation part can estimate phase error degree with 0.004° resolution and phase tracker can track the phase noise from -60° to 60° successfully. A block diagram of the Phase Tracker is shown in Figure 6.5.1.

Page 56

GDC21D003

(input)

Gain

Accumulator

Limiter

Delay

Hilbert

Transform

Filter

Accumulator

Gain Error

Figure 6.5.1 Phase Tracker

6.5.1 Error Detection

The result I signal is output to channel decoder and used at error detection with result Q signal to get estimated phase error degree. Figure 6.5.2 shows the block diagram of error detection. Result

Offset

Limiter

I’

Q’

Offset Error

(output)

Slice Prediction from

Viterbi Decoder

Derotator

cos sine

Sine

Cosine

Table

Accumulator

Limiter

I ”

Q”

Error Decision

Phase Error

I signal is input to VSB mapper which is a Look-up table and the output of this look-up table is a kind of decision error. The 2nd Look-up Table converts decision error to gain error, offset error, and phase error. And these error information are accumulated and used to remove phase noise.

I”

Q”

VSB mode

2VSB mapper 4VSB mapper

MUX

8VSB mapper

16VSB mapper

Figure 6.5.2 Error Detection

Error

Lookup

Table

Gain Error Offset Error Phase Error

Page 57

GDC21D003

6.5.2 Gain & Offset Loop

There are three loops in phase tracker and these loops are designed to be ON/OFF through I2C bus. The first loop is Gain Loop. The Gain Loop consists of multiplier and accumulator, and corrects gain error caused by phase error with error value from an error decision. The output of equalizer is the first gain controlled by a multiplier. The second

Figure 6.5.3 Coefficients of Hilbert Transform Filter

loop is Offset Loop which consists of adder and accumulator. The Offset Loop corrects offset error caused by phase error with error value from an error decision. The gain-corrected output is used to remove offset error at Offset Loop and used at the Hilbert transform filter to generate an approximation of the Q signal. Figure 6.5.3 shows the shape of coefficients of Hilbert Transform filter.

amplitude

time

6.5.3 Phase Loop

The last loop is Phase Loop which consists of adder, accumulator, Sine Cosine Table, and Complex

I’

Q’

Figure 6.5.4 Complex Multiplier

Multiplier. The gain & offset of corrected I and Q signal is derotated at complex multiplier to correct the phase error and Figure 6.5.4 shows the diagram of complex multiplier for a derotation.

cos

sine

cos

I”

Q”

Page 58

GDC21D003

6.5.4 I2C Bus I/F

User can initialize the resulting phase tracking value through I2C register33[6] (InitPHI2) bit. There are three loops in phase tracker; gain correction loop, offset correction loop, and phase

Table 6.5.1 Contents of I2C Bus Register33 for Phase Tracker

‘0’ : Freezes the Phase Tracker in the device, which means phase tracking does not occur.

7 nFreezePHI2

6 InitPHI2

PHASmodeIN

[5:4]

[1:0]

There are 4 different modes for phase tracking gain control as seen in the table 6.5.2. There is a tradeoff relation between gain and noise enhancement in phase tracker. So the phase tracker in this device changes the gain according to the noise condition. As shown in table, if LOOPgainIN is set to “000” gain is changed automatically according to the

‘1’ : normal operation

W/R

If you want to control Phase Tracker fast, use external input pins. Default value is ‘1’. ‘1’ : Initializes the Phase Tracker in the device. ‘0’ : normal operation

W/R

If you want to control Phase Tracker fast, use external input pins. Default value is ‘0’. There are three loops in the Phase Tracker, gain, offset, and phase loop. 00 : all loops on 01 : offset loop off

W/R

10 : offset and gain loops off 11 : all loops off Default value is “00”.

correction loop. The following table shows contents of register address 33 and these three loops can be turned on/off through this register[5:4] (PHASmodeIN). The mean squared errors are calculated at phase tracker output and can be read using I2C bus through MeanErrOUTPH.

noise condition. This means that this routine has noise calculator and selects a gain by using the calculated noise information. User can also control the gain of phase tracker by setting to “001”, “010”, or “011”. The default is “000” with this automatic gain mode.

[5:3]

LOOPgainIN [2 : 0]

Table 6.5.2 Contents of I2C Bus Register36 for Gain Control

Determines the use of automatic gain routine and the type of loop gain used in Phase Tracker. Default value is “000”. “000” : Automatic gain change. “001” : phase tracker is OFF.

W/R

“010” : smaller gain. “011” : normal gain. “1xx” : not use.

Page 59

GDC21D003

6.6 Channel Decoder

Channel Decoder consists of Cable Slicer, Viterbi Decoder, Symbol-to-Byte Converter, Convolutional Deinterleaver, Reed-Solomon Decoder, Data Derandomizer, Transport Demultiplexer Interface, and etc. as shown in Figure 6.6.1. Channel Decoder input comes from Phase Tracker. MMDS 2/4/8/16 VSB signal and ATSC 16 VSB signal are sliced in Cable Slicer, and ATSC 8 VSB signal is decoded in Viterbi Decoder. Viterbi Decoder is bypassed when I2C

Table 6.6.1 I2C bus register64 values for sub-block bypass

Viterbi Decoder on/off selection

Viterbi_on

Deint_on

RSdec_on

Derand_on

‘1’ : on ‘ 0’ : off If this bit is set to ‘0’, hard decision decoding is performed instead of viterbi decoding. Default value is ‘1’. Deinterleaver on/off selection ‘1’ : on ‘ 0’ : off If this bit is set to ‘0’, deinterleaver is bypassed. Default value is ‘ 1’ . RS Decoder on/off selection ‘1’ : on ‘ 0’ : off If this bit is set to ‘0’, RS decoder is bypassed. Default value is ‘ 1’ . Derandomizer on/off selection ‘1’ : on ‘ 0’ : off If this bit is set to ‘0’, derandomizer is bypassed. Default value is ‘ 1’ .

register64[7](Viterbi_on) value is ‘0’. In this case, signal is sliced by 4-level slicer. Symbol stream, the output of Cable Slicer and Viterbi Decoder is converted into Byte Stream by Symbol-to-Byte Converter and sent to Convolutional Deinterleaver. Convolutional Deinterleaver is bypassed when I2C register64[6] (Deint_on) value is ‘0’. ReedSolomon Decoder and Data Derandomizer are bypassed when I2C register64[5] (RSdec_on) and I2C register64[4] (Derand_on) are set to ‘0’ respectively. I2C register values bypassing each block are shown in Table 6.6.1.

Phase

Tracker

Cable Slicer

Viterbi

Decoder

Slicer

Vsbmode

1 0

Viterbi_on

Symbol-

to-Byte

Converter

Convolutional

Deinterleaver

1 0

Deint_on

Reed-Solomon

Decoder

1 0

RSdec_on

Figure 6.6.1 Block Diagram of Channel Decoder

Data

Derandomizer

Transport

Demultiplexer

I/F

1 0

Derand_on

Page 60

GDC21D003

6.6.1 12 Symbol Intrasegment Deinterleaver

To help protect the Viterbi decoder against short burst interference, such as impulse noise or NTSC co-channel interference, 12-symbol code intrasegment interleaving is employed in the transmitter. As shown in figure 6.6.2, the Viterbi

Equalized &

Phase Corrected

Symbols

decoding block uses 12 Viterbi decoders in parallel, where each Viterbi decoder sees every 12th symbol. This code interleaving has all the same burst noise benefits of a 12-symbol interleaver, but also minimizes the resulting code expansion when the NTSC rejection comb filter is active (when the NTSC rejection comb filter is active, the I2C register8[4](Ncomb) is set to ‘0’).

Viterbi Decoder

Viterbi

Decoded

Data

Figure 6.6.2 12 Symbol Intrasegment Deinterleaver

6.6.2 Segment Sync Suspension

Before the 8 VSB signals can be processed by the Viterbi decoder, it is necessary to suspend the Segment Sync. The presence of the Segment Sync character in the data stream passed through the comb filter presents a complication that must be dealt with, because Segment Sync is not trellis encoded or precoded. Figure 6.6.3 shows the technique that has been used. It shows the receiver processing that consists of the comb filter in the VSB Synchronizer/Equalizer and Segment Sync Removal block in this chip. The multiplexer in the Segment Sync Removal block is normally in the

Viterbi Decoder

#10

Viterbi Decoder

#11

upper position. This presents data that has been filtered by the comb filter to the Viterbi Decoder. However, because of the sync character in the data stream, the multiplexer selects its lower input during the 4 symbols that occurs 12 symbols after the Segment Sync. The effect of this sync removal is to present to the Viterbi Decoder a signal that consists of the subtraction of two adjacent data symbols that come from the same Viterbi Decoder, one transmitted before and one after the Segment Sync. The interference introduced by the Segment Sync symbol is removed in this process, and the overall channel response seen by the Viterbi Decoder is the single-delay partial response filter.

Page 61

Comb Filter Segment Sync Removal

GDC21D003

Equalizer

¢²

£-

Phase Tracker

VSB Synchonizer/Equalizer

Figure 6.6.3 Segment Sync Suspension

6.6.3 Viterbi Decoder

The Viterbi decoder performs the task of slicing and convolutional decoding. It has two modes; one is for the case when the NTSC rejection filter is used to minimize NTSC co-channel, and the other is when it is not used. This is illustrated in Figure

6.6.4. The insertion of the NTSC rejection filter is determined automatically in the comb filter control block(before the equalizer, see Figure 6.6.3), with this information passed to the Viterbi decoder . When there is little or no NTSC co-channel interference, the NTSC rejection filter is not used and an optimal Viterbi decoder is used to decode the 4-state trellis-encoded data. Serial bits are recreated in the same order as they were created in the encoder. With significant NTSC co-channel interference, if the NTSC rejection filter(12 symbol, feed-forward subtractive comb) is employed, an optimized 8-state Viterbi decoder for this partial response channel is used. The Viterbi decoder can be bypassed for performance tests by setting I2C register64[5](Viterbi_on) to ‘0’. Then, just 4-level slicing is done. Also, on ATSC 16 VSB and MMDS 2/4/8/16 VSB mode, the Viterbi decoder is bypassed and only slicing is done. Figure 6.6.5 is internal block diagram of Viterbi decoder. Each branch metric of trellis calculated in branch metric

(D = 12 Symbols Delay)

block is 6bits, memory depth of path memory block is 16. Branch metric block and ACS(Add, Compare, and Select) block are adapted to be 4state Viterbi decoder if I2C register84[4](Ncomb) value is ‘1’, or 8-state Viterbi decoder if it is ‘ 0’ . 12 Vitervi decoders are implemented through 12 shift registers sharing one Branch metric block and one ACS block.

¢²

U X

Viterbi Decoder

Page 62

NTSC Rejection Filter

¢²

GDC21D003

15 Level Symbols + Noise + Interference

Partial

Response

Viterbi

Decoder

Received Symbols

From

Segment Sync

Suspension

(D = 12 Symbols Delay)

Optimal

Viterbi

Decoder

8 Level Symbols + Noise + Interference

Figure 6.6.4 Viterbi Decoding with and without NTSC Rejection Filter

Shift

Register1

Shift

Register2

Branch

Metric

ACS

MUX

Sync

Ncomb

Path Memory

Symbol-to-byte

Sync

Converter

Data Out

Figure 6.6.5 Internal Block Diagram of Viterbi Decoder

Page 63

GDC21D003

6.6.4 Symbol-to-Byte Converter

The symbol-to-byte converter converts incoming symbols to bytes. On MMDS 2 VSB mode, each received symbol has one bit of data. The data from MSB to LSB {7,6,...,1,0}is entered and 8 symbols compose one byte. On MMDS 4 VSB and ATSC 8 VSB mode, each received symbol has two bits of data. The data from MSB to

Table 6.6.2 Symbol-to-Byte Conversion

Bits in Byte 7 6 5 4 3 2 1 0 7 6 5 4 3 2 1 0 7 6 5 4 3 2 1 0

Symbol

MMDS 2VSB

MMDS 4VSB/ ATSC 8VSB

MMDS 8VSB

MMDS 16VSB/ ATSC 16VSB

0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7 0 1 2 3 4 5 6 7

0 1 2 3 0 1 2 3 0 1 2 3

0 1 2 3 4 5 6 7

0 1 0 1 0 1

LSB {76,54,32,10} is entered and 4 symbols compose one byte. On MMDS and ATSC 16 VSB mode, each received symbol has four bits of data. The data from MSB to LSB {7654,3210} is entered and 2 symbols compose one byte. On MMDS 8 VSB mode, each received symbol has three bits of data. The data from MSB to LSB {765,432,107,654,321,076,543,210} is entered and 8 symbols compose three bytes. See Table 6.6.2

Page 64

GDC21D003

(

)

6.6.5 Convolutional Deinterleaver

Convolutional deinterleaver performs the opposite function of the transmitter convolutional interleaver. Even strong NTSC co-channel signals passing through the NTSC rejection filter and creation of short bursts due to NTSC vertical edges

From

Symbol-to-Byte Converter

(B=)52

M=4, B=52, N=208, R-S Block = 207,B¡¿M=N

are reliably handled due to the interleaving and RS coding process. The deinterleaver uses Data Field Sync for synchronizing with the first data byte of the data field. The structure is shown in Figure

6.6.6. The deinterleaver can be bypassed for performance tests by setting I2C register64[5] (Deint_on) to ‘0’.

(B-1)M

(B-2)M

Reed-Solomon Decoder

M (=4 Bytes)

Figure 6.6.6 Convolutional Deinterleaver

6.6.6 Reed-Solomon Decoder

and RS error correction. The RS decoder can be turned off for performance test by setting the I2C

Reed-Solomon decoder can correct up to 10byte errors per Data Segment that is encoded by RS(207, 187) code. Any burst errors created by impulse noise, NTSC co-channel interference, or Viterbi decoding errors, are greatly reduced by the combination of the convolutional deinterleaving

G X X X X X X X X X X X

20 19 18 17 16 15 14 13 12 11

= + + + + + + + + +

+ + + + + + + + + + +

152 185 240 5 111 99 6 220 112

10 9 8 7 6 5 4 3 2 1

X X X X X X X X X X

150 69 36 187 22 228 198 121 121 165 174

register64[4](rsdec_on) to ‘0’. The number of segment errors per one second is written in I2C register65 and register67 so that they can be read through I2C register control. The encoder polynomial is as follows. The coefficients of the polynomial are defined in Galois field GF(256).

Page 65

GDC21D003

6.6.7 Data Derandomizer

The data(excluding Field Sync Data, Segment Sync Data, and parity bytes) is randomized at the transmitter by Pseudo Random Sequence(PRS). The de-randomizer accepts the error-corrected data bytes from the RS decoder, and applies the same PRS randomizing code to the data. The PRS code is generated identically as in the transmitter, using the same PRS generator feedback and output taps

Figure 6.6.7 Derandomizer Polynomial

as shown in Figure 6.6.7. Since the PRS is locked into the reliably recovered Data Field Sync(and not some code words embedded within the potential noisy data), it is exactly synchronized with the data and performs reliably. The initialization(pre-load) to F180H occurs during the Data Segment Sync interval prior to the first Data Segment. The data derandomizer can be bypassed for performance tests by setting the I2C register64[3](Derand_on) to ‘0’.

X9X

Page 66

6.6.8 I/F to Transport Demultiplexer

GDC21D003

The VSB Channel Decoder offers various functions for the interface with the Transport Demultiplexer which are controlled by I2C register64.

Table 6.6.3 I2C Register64 Flags controlling Transport Demultiplexer I/F

Parallel/serial output selection ‘1’ : parallel ‘ 0’ : serial

Pase

Derand_on

Errorflag_ins

Vsbdvalid_pol

Vsbclk_sup

If this bit is set to ‘0’, VSBDATA[0] pin is used as serial data output and VSBDATA[7] is used as start bit of a byte indicator. Default value is ‘1’. Derandomizer on/off selection ‘1’ : on ‘ 0’ : off If this bit is set to ‘0’, derandomizer is bypassed. Default value is ‘ 1’ . Error flag bit insertion on/off selection. Valid only when Derand_on is set to ‘1’. ‘1’ : MSB of first data byte is set to ‘ 1’ when the packet has an uncorrected errors(when NVSBERRFLG is ‘0’) ‘0’ : nothing is done at the MSB of first data byte although the packet has an uncorrected errors(when NVSBERRFLG is ‘0’) Default value is ‘1’. VSBDVALID polarity indicator ‘1’ : VSBDATA[7:0] is valid during VSBDVALID = ‘ 1’ interval and invalid during VSBDVALID = ‘0’ interval. ‘0’ : polarity is inverted Default value is ‘1’. VSBCLK suppression indicator ‘1’ : VSBCLK is not suppressed during VSBDVALID = “invalid” interval ‘0’ : VSBCLK is suppressed(set to 0) during VSBDVALID = ”invalid” interval Default value is ‘1’.

The related I2C register64 flags are given in Table

6.6.3.

The rising edge of VSBCLK signal occurs at the middle of one byte interval of VSBDATA[7:0] signal. The VSBSOP signal indicates the location of the data segment sync(start byte of a data segment) using VSBSOP=‘1’ for one byte interval. At the default value of I2C register64,

1) the signal VSBDVALID=’1’ and VSBDVALID=’ 0’ indicate valid data and invalid data interval respectively,

2) the signal NVSBERRFLG=’0’ indicates that the data segment has uncorrected errors. Otherwise NVSBERRFLG=’1’ ,

3) the data segment sync is set to 47H,

4) the MSB of the first data byte is set to ‘1’ if the data segment has uncorrected errors(NVSBERRFLG=’ 0’),

5) the VSBCLK signal is not suppressed during VSBDVALID=’0’ interval. See Fig. 6.6.8.

Page 67

VSBCLK

VSBSOP

VSBDVALID

GDC21D003

NVSBERRFLG

VSBDATA[7:0]

47H byte0 byte1 byte186 0 47H0 byte0

with error

188 20

MSB of byte0 in the packet with error is set to ‘1’

Figure 6.6.8 I/F to Transport Demultiplexer when Register64[7:0] is set to Default Value

If only I2C register64[3](Derand_on) is set to ‘0’ and other values are reserved as the default,

1) the derandomizer is bypassed,

2) the data segment sync is set to 00H instead of 47H,

3) nothing is done at the MSB of the first data byte although the data segment has uncorrected errors.

See Fig. 6.6.9.

without error

Page 68

VSBCLK

VSBSOP

VSBDVALID

GDC21D003

NVSBERRFLG

VSBDATA[7:0]

00H byte0 byte1 byte186 parity values 00H byte0

with error

188 20

Figure 6.6.9 I/F to Transport Demultiplexer when Register64[3](Derand_on) is set to ‘ 0’

If only I2C register64[2](Errorflag_ins) is set to ‘0’ and other values are reserved as the default,

1) nothing is done at the MSB of the first data byte although the data segment has uncorrected errors. See Fig. 6.6.10.

VSBCLK

VSBSOP

VSBDVALID

NVSBERRFLG

with error

without error

VSBDATA[7:0]

47H byte0 byte1 byte186 0 47H0 byte0

188 20

Nothing is done to the MSB of byte0 in the packet with error (valid only when register64[3] is set to ‘1’)

Figure 6.6.10 I/F to Transport Demultiplexer when Register64[2](Errorflg_ins) is set to ‘ 0’

If only I2C register64[1](Vsbdvalid_pol) is set to ‘0’ and other values are reserved as the default,

1) signal VSBDVALID=’0’ and VSBDVALID=’ 1’ indicate valid data and invalid data interval respectively. See Fig. 6.6.11.

Page 69

VSBCLK

VSBSOP

GDC21D003

VSBDVALID

NVSBERRFLG

VSBDATA[7:0]

polarity is inverted

with error

47H byte0 byte1 byte186 0 47H0 byte0

188 20

Figure 6.6.11 I/F to Transport Demultiplexer when Register64[1] is set to ‘ 0’

If only I2C register64[0](Vsbclk_sup) is set to ‘0’ and other values are reserved as the default,

1) the VSBCLK signal is suppressed to ‘0’ during VSBDVALID=’ 0’ interval. See Fig. 6.6.12.

suppressed to ‘0’

VSBCLK

VSBSOP

without error

VSBDVALID

NVSBERRFLG

VSBDATA[7:0]

47H byte0 byte1 byte186 0 47H0 byte0

with error

188 20

without error

Figure 6.6.12 I/F to Transport Demultiplexer when Register64[0](Vsbclk_sup) is set to ‘ 0’

At MMDS 8VSB mode, the duty cycle of VSBCLK is not 50% of VSBCLK period. Since 8 symbols compose 3 bytes, the duration of first byte equals to those of two symbols, and the duration of second and third byte is 3 symbols each. See Fig. 6.6.13.

Page 70

GDC21D003

H or L

VSBCLK

VSBDATA[7:0]

Figure 6.6.13 I/F to Transport Demultiplexer(MMDS 8VSB Mode)

If only I2C register64[7](pase) is set to ‘0’ and other values are reserved as the default, serial interface is offered.

1) the VSBDATA[7] signal indicates the start bit of a byte,

2) the VSBDATA[0] signal is the serial data output,

3) the VSBDATA[6:1] signals are set to “000000”. See Fig. 6.6.14.

VSBCLK VSBSOP

VSBDVALID NVSBERRFLG VSBDATA[7]

VSBDATA[0]

b7 b6 b5 b4 b3 b2 b1 b0 b7 b6 b5 b40 1 0 0 0 1 1 1

......

Figure 6.6.14 I/F to Transport Demultiplexer at Serial Output Mode

Connection examples with VSB receivers and transport demultiplexer chips are shown in Fig. 6.6.15 - 6.6.17.

Page 71

GDC21D003

VSBDATA[6] VSBDATA[7] VSBDATA[5] VSBDATA[4] VSBDATA[3] VSBDATA[2] VSBDATA[1] VSBDATA[0]

VSBCLK

VSBDVALID

NVSBERRFLG

GDC21D003

VSB Receiver

80 77

72 71

69 68

84 83

FEC_DATA[7]

FEC_DATA[6]

FEC_DATA[5]

FEC_DATA[4]

FEC_DATA[3]

FEC_DATA[2]

FEC_DATA[1]

FEC_DATA[0]

FEC_CLOCK

\D_VALID

\ERR_BLOCK

GDC21D301

Transport Demultiplexer

Figure 6.6.15 Connection with VSB Receiver and Transport Demultiplexer Chip(GDC21D301A)

VSBDATA[6] VSBDATA[7] VSBDATA[5] VSBDATA[4] VSBDATA[3] VSBDATA[2] VSBDATA[1] VSBDATA[0]

VSBCLK

VSBDVALID

NVSBERRFLG

VSBSOP

GDC21D003

CD[7] CD[6] CD[5] CD[4] CD[3] CD[2] CD[1] CD[0] CDCLK CDEN \CDERR CDSYNC

L64007

(LSILOGIC)

VSB Receiver

Transport Demultiplexer

Figure 6.6.16 Connection with VSB Receiver and Transport Demultiplexer Chip(L64007)

Page 72

GDC21D003

VSBDATA[6] VSBDATA[7] VSBDATA[5] VSBDATA[4] VSBDATA[3] VSBDATA[2] VSBDATA[1] VSBDATA[0]

VSBCLK

VSBDVALID

GDC21D003

VSB Receiver

Figure 6.6.17 Connection with VSB Receiver and Transport Demultiplexer Chip(AVIA-MAX)

CHDATA[7] CHDATA[6] CHDATA[5] CHDATA[4] CHDATA[3] CHDATA[2] CHDATA[1] CHDATA[0] CHCLK CHEN

AVIA-MAX

(C-CUBE)

Transport Demultiplexer

Page 73

GDC21D003

6.7 PLL

The 4-times multiplexing scheme is used in the 256-tap filter. The PLL is used for generating the 4-times multiplying clock(CLK4EQ). Basically a little jitter between SYMCLK and CLK4EQ may occur. The clock skew due to the jitter can result in abnormal operation in the paths related to both two clocks. We used the following method to make

SYMCLK

4X-PLL

CLK4EQ

robust design against the clock skew. The SYMCLK is delayed by 4 CLK4EQ-clock cycles. The delayed clock is CLKEQ. Then the skew between CLKEQ and CLK4EQ is fixed in spite of the jitter of PLL. To fix skew is very useful because it is difficult to estimate the accurate amount of jitter. The scheme is shown in figure

6.7.1.

Equalizer

Control

CLKEQ

Clock Delay

Coefficient-

Update

Filter

Figure 6.7.1 Clock Scheme

Page 74

GDC21D003

7. Electrical Characteristics

Absolute Maximum Rating

SYMBOL PARAMETERS RATING UNIT

stg

Recommended Operating Range

SYMBOL PARAMETERS MIN MAX UNIT

opr

DC Characteristics (VDD = 3.15 V ~ 3.45 V, TA = 0 ~ 70 °C )

SYMBOL PARAMETERS MIN MAX UNIT

DDS

AC Characteristics (VDD = 3.3 V ± 0.15, TA = 0 ~ 70 °C )

SYMBOL PARAMETERS MIN TYPE MAX UNIT

* Output Delay time 29 ns

OD1

** Output Delay time 29 ns

OD2

CRH

IPL

Power Supply Voltage - 0.5 ~ 4.6 V DC Input Voltage - 0.5 ~ V

+ 0.5 V

Storage Temperature -25 ~ 150

Power Supply Voltage 3.15 3.45 V Operating Temperature 0 70

Input High Voltage 0.7*VDD VDD + 0.5 V Input Low Voltage - 0.5 0.3*VDD V Output High Voltage 2.4 V Output Low Voltage 0.4 V Dynamic Supply Current 850 mA Standby Current 30 mA

Clock SYMCLK Period 92.9 ns Input Setup Time 10 ns Input Hold Time 10 ns

Chip Reset Hold Time 20 t Internal PLL Lock-up Time 1 ms

°C

* related pins are 68,69,71,72,74,75,77,80,83,84,85,89,108,110 ** related pins are 43,44,46,52,53,55,56,57,66,90,92,93,95,96,97,100,102,103,104,106

Page 75

Timing Specification

SYMCLK

NRESET

POWER

GDC21D003

CRH

IPL

ALL IN/OUT

SYMCLK

Input

Output

Not Valid

Figure 7.1 Clock Reset Stabilization Timing

tISt

OD1

OD2

Valid

Figure 7.2 Input and Output Timing

Page 76

8. Package Dimensions

GDC21D003

Figure 8.1 Physical Dimensions

Page 77

9. Application Notes

GDC21D003

Demodulation Chip

SANYO LA7785M

IOUT+

IOUT-

GUP GDN

POLN POLP

IFIN1

IFIN2

RFAGC

108

SAW Filter

TUNER

VSB Receiver Chip

GDC21D003

3.3V

1.5K

2.2V

1.2V

3.3V

12K

0.1uF

1.5V

0.1uF

3.3K

0.1uF

3.3V

3.3K

2.0K

0.1uF

3.3V

3.3K

2.7K

0.1uF

3.3V

125

INP

126

INN

GUP

GDN

DATAPOLN

DATAPOLP

124

REFP

123

REFN

REF

121

COM_0

128

BIAS

AVDD AVSS

OPEN

ADCCLK

14 18

VSBDATA[7] VSBDATA[6] VSBDATA[5] VSBDATA[4] VSBDATA[3] VSBDATA[2] VSBDATA[1] VSBDATA[0]

VSBCLK

VSBDVALID

NVSBERRFLG

NCHGUP

NCHGDN

SYMCLK

CLKFS

41 12

80 77

74 72

3.3V

Analog

S/W

Analog

S/W

VCXO

10.76MHz, 21.52MHz

10.76MHz, 21.52MHz, 43.04MHz & 75.32MHz

FEC_DATA[7]

FEC_DATA[6]

FEC_DATA[5]

FEC_DATA[4]

FEC_DATA[3]

FEC_DATA[2]

FEC_DATA[1]

FEC_DATA[0]

FEC_CLOCK

\D_VAILD

\ERR_BLOCK

6.8K

10K

6.8K

Transport Chip

GDC21D301

10K

3.6K

0.01uF

0.1uF

2.2uF

VCXO

Figure 9.1 VSB Receiver Application Circuit