Preliminary specification
File under Integrated Circuits, IC01
2000 Mar 21
Philips SemiconductorsPreliminary specification
Channel encoder/decoder CDR60SAA7392
CONTENTS
1FEATURES
2GENERAL DESCRIPTION
3QUICK REFERENCE DATA
4ORDERING INFORMATION
5BLOCK DIAGRAM
6PINNING INFORMATION
6.1Pinning
6.2Pin description
7FUNCTIONAL DESCRIPTION
7.1Microprocessor interfaces
7.2Register map
7.3System clocks
7.4HF analog front-end
7.5Bit recovery
7.6Decoder function
7.7Subcode interface
7.8Digital output
7.9Serial output interface
7.10Motor control
7.11The serial in function
7.12The subcode insert function
7.13The data encoder block
7.14Encode control block
7.15The EFM modulator
7.16The EFM clock generator
7.17The Wobble processor
8LIMITING VALUES
9OPERATING CHARACTERISTICS
9.1ADC and AGC parameters
10APPLICATION INFORMATION
10.1Write startcontrol of encoder in CD-ROM mode
10.2Write start control of encoder in Audio mode
10.3Start-up of encode in flow-control operation
10.4Start-up of encoder in synchronous stream
mode
10.5Operating CDR60 in CAV mode, flow control
on input stream
10.6Operating in CLV Mode, Flow Control on Input
Stream
10.7Operating in CLV Mode, Synchronous Stream
Operation
11PACKAGE OUTLINE
12SOLDERING
12.1Introduction to soldering surface mount
packages
12.2Reflow soldering
12.3Wave soldering
12.4Manual soldering
12.5Suitability of surface mount IC packages for
wave and reflow soldering methods
13DEFINITIONS
14LIFE SUPPORT APPLICATIONS
15PURCHASE OF PHILIPS I2C COMPONENTS
2000 Mar 212
Philips SemiconductorsPreliminary specification
Channel encoder/decoder CDR60SAA7392
1FEATURES
• Very high speed Compact Disc (CD) compatible
decoding and encoding device
• On-chip Analog-to-Digital Converter (ADC) and
Automatic Gain Control (AGC) for HF data capture
• Eight-to-Fourteen Modulation (EFM)
• Advanced motor control loop to allow CAV, CLV and
pseudo-CLV playback
• Integrated FIFO for de-coupling of mechanism speed
and application speed
• Versatile output interface allowing different I2S-bus and
Electronic Industries Association of Japan (EIAJ)
formats
• Device is fully compatible with ELM, PLUM and Sanyo
CD-ROM block decoders
• Quad-pass CIRC correction for CD mode
(C1-C2-C1-C2)
• Subcode/header processing for CD format
• Frequency multiplier allows use of a 8 MHz crystal.
The bit detector recovers the individual bits from the
incoming signal, correcting asymmetry, performing noise
filtering and equalisation, and recovering the channel bit
clock using a digital PLL. The demodulator converts the
EFM bits to byte-wide data symbols, before passing them
onto the decoder for subcode extraction, de-interleaving
and error correction. The decoded data is then made
available via the multi-function serial output interface.
The encode datapath takes data symbols from the block
encoder/decoder via the serial data and subcode input
functions, encoding them via the encoder block. The
encoded data stream is passed to the EFM modulator,
which generates the required EFM signal, output as a
digital bitstream. Theencode process iscontrolled viathe
Wobble processor, encode control and EFM clock
generator functions.
2GENERAL DESCRIPTION
As well as these two data processing sections, three
further blocks support overall device operation. The
CDR60 is a channel encoder/decoder for
CD/CD-R/CD-RW/CD Audio Recorder systems. It
incorporates all logic and RAM required for the complete
encoding and decoding processes.
There aretwo main datapathsthrough the CDR60 device.
system clockgenerator provides all digitalclocks required
by the CDR60. The motor servo allows the CDR60 to
control the spindle motor and is controlled by the
microprocessor interface. This interface can be accessed
either via a parallel (80C51) or a serial (I2C-bus) interface.
The decode datapath captures the incoming EFM data
stream via the HF ADC and AGC functions.
3QUICK REFERENCE DATA
SYMBOLPARAMETERMIN.TYP.MAX.UNIT
V
V
V
I
f
T
T
DDD
DDA
DDE
DD
xtal
amb
stg
supply voltage (core and pad ring)3.03.33.6V
supply voltage (analog)3.03.33.6V
supply voltage (output drivers)3.03.33.6V
supply current−200−mA
crystal frequency88.467233MHz
operating ambient temperature0−70°C
storage temperature−55−+125°C
WREFLO1Owobble ADC analog reference voltage
WREFHI2Owobble ADC analog reference voltage
WREFMID3Owobble ADC analog reference voltage
V
SSA1
V
DDA1
WIN6Iwobble analog input
VREF7Oanalog voltage reference
IREF8Oanalog current reference
n.c.9−not connected
HREFHI10OHFADC analog reference voltage
HREFLO11OHF ADC analog reference voltage
AGCREF12IAGC analog reference voltage
HIN13IHF analog data input
HREFMID14OHFADC analog reference voltage
V
SCL32Isub-CPU clock
SDA33I/Obidirectional sub-CPU data
INT34Osub-CPU interrupt
DA735I/Obidirectional sub-CPU parallel data bus
DA636I/Obidirectional sub-CPU parallel data bus
DA537I/Obidirectional sub-CPU parallel data bus
DA438I/Obidirectional sub-CPU parallel data bus
4supplyanalog ground
5supply3 V analog supply voltage 1; note 2
15supply3 V analog supply voltage 2; note 2
16supplyanalog ground
19, 43, 62, 71supplyoutput driver ground
20supplyoutput driver 3 V supply voltage
30, 49, 53, 76supplycore and pad ring 3 V supply voltage; note 2
31, 48, 52, 75supplycore and pad ring ground
2000 Mar 216
Philips SemiconductorsPreliminary specification
Channel encoder/decoder CDR60SAA7392
SYMBOLPINTYPEDESCRIPTION
DA339I/Obidirectional sub-CPU parallel data bus
DA240I/Obidirectional sub-CPU parallel data bus
DA141I/Obidirectional sub-CPU parallel data bus
DA042I/Obidirectional sub-CPU parallel data bus
V
DDE
WRi45Isub-CPU write enable; active LOW
RDi46Isub-CPU read enable; active LOW
ALE47Isub-CPU address latch enable
CSi50Isub-CPU chip select
PCAin51IPCA input
STOPCK54Ostop clock output
V455Oserial subcode output
EBUOUT56Odigital output
SYNC57OI
FLAG58OI
DATAO59OI
WCLK60I/Obidirectional I
BCLK61I/Obidirectional I
V
DDE
DATAI64II
SUB65IEIAJ subcode data
RCK66OEIAJ subcode clock
SFSY67IEIAJ subcode sync
CFLG68Ocorrection statistics; open-drain
T269Itacho control input 2
T170Itacho control input 1
V
DDE
MOTO2/T373I/Omotor output 2/tacho input 3
MOTO174Omotor control output 1
LASERON77Olaser write control
XEFM78OEFM clock output
EFMDATA79OEFM data output
MEAS180Ofront end telemetry; open-drain
44supplyoutput driver 3 V supply voltage
2
S sector sync output
2
S correction flag
2
S data output
2
S word clock
2
S bit clock
63supplyoutput driver 3 V supply voltage
2
S data input
72supplyoutput driver 3 V supply voltage
Notes
1. No signal may be applied to this device when it is not powered.
2. The analog and digital supply pins (V
DDA
and V
) must be connected to the same external supply.
DDD
2000 Mar 217
Philips SemiconductorsPreliminary specification
Channel encoder/decoder CDR60SAA7392
7FUNCTIONAL DESCRIPTION
7.1Microprocessor interfaces
The SAA7392 is programmed via two independent
microprocessor interfaces:
• Parallel 80C51 compatible
– DA(7:0) = address/data bus
– ALE = address latch enable; latches the address
information on the bus
– WRi = active LOW write signal; write to SAA7392
– RDi = active LOW read signal; read from SAA7392
– CSi = chip select signal; gates the RDi and WRi
signals.
7.1.1S
ERIAL I
C-BUS INTERFACE
2
Data is transferred over the interface in single bytes, via
write data or read data commands.
The sequence for a write data command is as follows:
1. Send START condition
2. Send address 3EH (write)
3. Write register address byte
4. Write data byte
5. Send STOP condition.
The sequence for a read data command is as follows:
1. Send START condition
2. Send address 3EH (write)
3. Write status register address byte
4. Send STOP condition
5. Send address 3FH (read)
6. Read data byte
7. Send STOP condition.
7.1.2P
ARALLEL INTERFACE
The parallelinterface hasa multiplexed address/databus.
Information can be written to or read from the SAA7392
using the protocols shown inFigs 3 and 4; specific timings
are shownin Table 2. Notethat only thelower six address
bits are decoded; so writing to address 40H would have
the same effect as writing to address 00H.
t
handbook, full pagewidth
ALE
WRi
CSi
DA0 to DA7
address (0:7)
IN
t
su1
d1
t
d2
data (0:7)
t
t
h2
su2
Fig.3 Microprocessor write protocol.
2000 Mar 218
t
WRiL
t
IN
h1
MGR793
Philips SemiconductorsPreliminary specification
Channel encoder/decoder CDR60SAA7392
t
handbook, full pagewidth
ALE
RDi
CSi
d1
RDi
L
t
d2
t
h1
DA0 to DA7
address (0:7)
IN
t
su1
t
h2
data (0:7)
OUT
Fig.4 Microprocessor read protocol.
Table 2 Parallel interface timing
SYMBOLDESCRIPTIONMIN.
t
d1
t
d2
t
h1
t
su1
t
h2
t
su2
t
h3
t
WRiL
t
h4
t
d3
t
d4
t
RDiL
Delay ALE falling to RDi/WRi falling.17−ns
Delay CSi rising to RDi/WRi falling.17−ns
CSi hold time after RDi/WRi falling.2T
Address setup time before ALE falling.17−ns
Address hold time after ALE falling.17−ns
Data setup time before WRi falling.0−ns
Data hold time after WRi falling.2T
WRi LOW time.1T
ALE LOW hold time after WRi LOW.3T
Delay data valid after RDi LOW.−3T
Delay RDi HIGH to data out high-impedance.−17ns
RDi LOW time.3T
MGR794
(1)
+17−ns
clk
+17−ns
clk
+17−ns
clk
+17−ns
clk
+ 128−ns
clk
(1)
MAX.
+17ns
clk
UNIT
Note
1. T
is the system clock period.
clk
2000 Mar 219
This text is here in white to force landscape pages to be rotated correctly when browsing through the pdf in the Acrobat reader.This text is here in
_white to force landscape pages to be rotated correctly when browsing through the pdf in the Acrobat reader.This text is here inThis text is here in
white to force landscape pages to be rotated correctly when browsingthrough the pdf inthe Acrobat reader. white to force landscape pages to be ...
0EMotor Control Register 3 (Motor3)WriteMotor integrator presetmotor/tacho
0FMotor Control Register 4 (Motor4)WriteMotor controlmotor/tacho
10Motor Control Register 5 (Motor5)Read/Write Motor integrator valuemotor/tacho
11Motor Control Register 6 (Motor6)Read/Write Motor integrator valuemotor/tacho
2
S Output Register 2 (Output2)WriteI2S output 2serial out
2
S Output Register 3 (Output3)WriteI2S output 3serial out
REGISTER NAMETYPEFUNCTION
Read8-bit PLL frequencybit detector
Read8-bit asymmetry signalbit detector
Read8-bit jitter signalbit detector
ReadObserve internal lock flagsbit detector
communication
communication
communication
Read8-bit slicer compensation
value
ReadOpening of eye patternbit detector
ReadRead back of motor frequency motor/tacho
sub-CPU
sub-CPU
sub-CPU
bit detector
BLOCK
RESPONSIBLE
This text is here in white to force landscape pages to be rotated correctly when browsing through the pdf in the Acrobat reader.This text is here in
_white to force landscape pages to be rotated correctly when browsing through the pdf in the Acrobat reader.This text is here inThis text is here in
white to force landscape pages to be rotated correctly when browsingthrough the pdf inthe Acrobat reader. white to force landscape pages to be ...
19Tacho Control Register (Tacho3)WriteTacho control settingsmotor/tacho
1BSoft Reset Register (SoftReset)WriteSub-block resetsub-CPU
1DMotor Control Register7 (Motor7)WriteControl coefficients selectmotor/tacho
1EInput Configuration Register (InputConfig)WriteEBU clock frequency and
38ATIP Error Register (ATER)ReadCounter for ATIP CRC errorssub-CPU
39C1 Block Error Register (C1BLER)ReadCounter for C1 errorssub-CPU
3AC2 Block Error Register (C2BLER)ReadCounter for C2 errorssub-CPU
3CEFM Preset Count Register (EFMPresetCount)WriteEFM frame position for outputEFM modulator
3DEFM Modulator Configuration Register (EFMModConfig)WriteXEFM control and output data
3EEFM Modulator Configuration Register 2 (EFMModConfig2)WriteXEFM control and output data
REGISTER NAMETYPEFUNCTION
data
ReadIntegrator valueEFM clock generator
format
format
BLOCK
RESPONSIBLE
Wobble processor
EFM modulator
EFM modulator
Philips SemiconductorsPreliminary specification
Channel encoder/decoder CDR60SAA7392
Philips SemiconductorsPreliminary specification
Channel encoder/decoder CDR60SAA7392
7.2.1INTERRUPT PIN
The interrupt pin (INT) is the AND-OR-INVERT of the Status and Interrupt Enable Registers, i.e. INT will become active
when corresponding bits are set at the same time in the Status and Interrupt Enable Registers.
7.2.2THE SEMAPHORE REGISTERS (SEMA1, SEMA2 AND SEMA3)
The Semaphore Registers are intended for inter-microprocessor communications. For example, microcontroller 1 can
writedata tomicrocontroller 2 viaSema1 andmicrocontroller 2 can writedata tomicrocontroller 1 viaSema2. TheStatus
Register ofthe SAA7392offers a mechanismso that both microcontrollers cansee when newdata hasbeen written and
whenit hasbeen readby lookingat thecontents ofthe Semaphore Registers. Version M3of theCDR60 canbe identified
by writingand reading register Sema3. In version M3,bit 1 of Sema3is always read as logic 0, whereas inother CDR60
versions this bit reads the same value as what was written to it before.
7Sema1If Sema1 = 1, change in register Sema1 has been detected. Reset if register Sema1 read.
6Sema2If Sema2 = 1, change in register Sema2 has been detected. Reset if register Sema2 read.
5Sema3If Sema3 = 1, change in register Sema3 has been detected. Reset if register Sema3 read.
4LockInIf LockIn = 1, then channel data PLL in lock (not latched).
3HeaderValHeaderVal is set when new header/subcode is available; reset on reading SubReadEnd.
2MotorOverflow If MotorOverflow = 1, then a motor overflow is occurring (not latched).
1FIFOOvIf FIFOOv = 1, then the FIFO has overflowed.
0−This bit is reserved.
7Sema1EnIf Sema1En = 1, then Semaphore Register 1 interrupt is enabled.
6Sema2EnIf Sema2En = 1, then Semaphore Register 2 interrupt is enabled.
5Sema3EnIf Sema3En = 1, then Semaphore Register 3 interrupt is enabled.
4LockInEnIf LockinEn = 1, then channel data PLL in lock interrupt is enabled.
3HeaderValEnIf HeaderValEn = 1, then new header/subcode available interrupt is enabled.
2MotorOverflowEn If MotorOverflowEn = 1, then motor overflow interrupt is enabled.
1FIFOOvEnIf FIFOOvEn = 1, then FIFO overflow interrupt is enabled.
0−This bit is reserved.
7.2.5S
Table 11 Status Register 2 (address 20H) - READ/WRITE
7BankSwitchEnIf BankSwitchEn = 1, then BankSwitch interrupt is enabled.
6SyncErrorEn If SyncErrorEn = 1, then SyncError interrupt is enabled.
5DataNotVali
dEn
4QSyncEnIf QSyncEn = 1, then QSync interrupt is enabled.
3ATIPSyncEn If ATIPSyncEn = 1, then ATIPSync interrupt is enabled.
2LaserOnEnIf LaserOnEn = 1, then LaserOn interrupt is enabled.
1LaserOffEnIf LaserOffEn = 1, then LaserOff interrupt is enabled.
0XErrorLargeEnIf XerrorLarge = 1, then XErrorLarge interrupt is enabled.
If DataNotValidEn= 1, then DataNotValid interrupt is enabled.
0SReset1When set, synchronisation with PLUM on subcode transfer has failed; reset when
a logic 1 is written to this bit (Status2).
This bit is an active HIGH reset to the following blocks: Encoder/decoder, EFM
modulator, Encode control block, Serial input/output block and Encode subcode insert
block. The clock control, EFM PLL, tacho, motor interface and wobble interface remain
running.
Soft reset will reset the following registers: EFMPresetCount, EFMModulateConfig,
EFMModulateConfig2, EncodeXOffset, EncodeWriteControl, EncodeStartOffset,
EncodeStopOffset, SubPresetCount, SubConfig1, Subconfig2,SubStartData, SubData,
InputConfig, DecoMode, Output1, Output2 and Output3.
A soft reset is mandatory in the following cases:
1. After programming the BCLK clock
2. When switching from encode to decode
3. When switching from decode to encode.
2000 Mar 2115
Philips SemiconductorsPreliminary specification
Channel encoder/decoder CDR60SAA7392
7.3System clocks
The principleclocks used inthe SAA7392 are derived from the crystaloscillator input pinXTLI (alternatively, anexternal
clock can be connected to this pin). These clocks are the system clock (also used as the ADC clock) and the I2S output
bit clock (BCLK).
The system clock (f
) defines the maximumoperational channel rate for the device. The maximum EFM channel clock
clk
is twice the system clock, for CD it is equivalent tosystem clock/(4.3 × 106) which is approximately 11.5 × CDROM for a
25 MHz system clock.
The other clock in the system is the channel data clock, this is recovered by the front-end bit recovery PLL.
handbook, full pagewidth
crystal
oscillator
MUXSWI
XTLI
XTLO
CL1
×
(1)
CLOCK
MULTIPLIER
CL1
DIVIDER
M × XTLI
XTLI
SYSTEM
CLOCK
DIVIDER
BIT
CLOCK
DIVIDER
system clock
system clock
BCLK
MGR795
(1) M = 1 if MUXSWI is LOW; M = 8 if MUXSWI is HIGH.
7CL1DivIf CL1Div = 0, then CL1 output frequency is
frequency is1⁄2f
6GateBClkIf GateBClk = 0, then I
clk
.
2
S output bit clock gating is disabled. If GateBClk = 1, then I2S
output bit clock gating enabled, BCLK is output, clock is automatically stopped if FIFO
underflows (this is known as Flow control mode).
5Div.1These 2 bits select the system clock frequency (f
4Div.0
should be programmed for the expected disc channel rate (e.g. 4.33 MHz for 1 × CD)
within the following constraints:
. If CL1Div = 1, then CL1 output
clk
); see Table 19. This frequency
clk
Channel rate
---------------------------------2
4 Channel rate×<<
f
clk
In this clock range, reliable bit detection is possible. All data found will be written to the
FIFO. It is the responsibility of the user to select system clock values so that the FIFO
performance is controlled.
3Mux2If Mux2 = 0, then N (bit clock divider pre-scaler) = 1. If Mux2 = 1, then N = M.
2 to 0Div2<2:0>These 3 bits select the BCLK frequency (f
); see Table 20. It is the responsibility of
BCLK
the user to select BCLK values so that the FIFO performance is controlled.
Table 19 Selection of system clock frequency
Div.1Div.0SYSTEM CLOCK FREQUENCY (f
00M×f
010.5 × M × f
100.25 × M × f
110.125 × M × f
XTLI
XTLI
XTLI
XTLI
clk
Table 20 Selection of BCLK frequency
Div2.1Div2.1Div2.0BCLK FREQUENCY (f
000N×f
001N×f
010
011
100
101
110
111
XTLI
XTLI
1
/
(N × f
2
1
/
(N × f
3
1
/
(N × f
4
1
/
(N × f
6
1
/
(N × f
8
1
/
(N × f
12
XTLI
XTLI
XTLI
XTLI
XTLI
XTLI
)
)
)
)
)
)
BCLK
)
)
2000 Mar 2117
Philips SemiconductorsPreliminary specification
Channel encoder/decoder CDR60SAA7392
7.4HF analog front-end
The HF ADC in the SAA7392 encodes the EFM high
frequency signal from the disc light pen assembly. These
signals are pre-processed, externally to the SAA7392, by
either AEGER-2 or a DALAS equivalent. The dynamic
range of the ADC is optimized by the inclusion of an
AC coupled AGC function under digital control.
In order to make use of the whole digital front-end
resolution, the output of the gain control amplifier should
constantly deliver 1.4 V
ofthe ADC isapproximately14 dB, with32 steps. Thegain
control for the variable gain amplifier is controlled by an
on-chip digital gaincontrol block (AGC). This blockallows
for both automatic and microprocessor gain control. The
gain control block will detect ADC extreme conditions
(00H or FFH outputs); on these values the gain control
block will decrement the gain. If no extreme codes occur
the gain is incremented.
7.4.1FIXED GAIN
Control of the gain is as follows:
1. Writing XX1X XXXX to the Anaset1 register
(address 15H) increases the AGC gain by 1.1 dB
2. Writing XX0X XXXX to the AnaSet1 register
(address 15H) decreases the AGC gain by 1.1 dB
output signal. The gain range
(p-p)
3. Instructions to increment/decrement gain are ignored
when the AGC gain limits of −4/+12 dB are reached.
7.4.2AUTOMATIC GAIN CONTROL (AGC)
The gain of theAGC cellis adjusteduntil the analog signal
at the ADC input extends over the complete range of
the ADC.Detection ofthis conditionis in thedigital domain
where the maximum and minimum ADC codes are
measured. The dynamics of the AGC system are as
follows.
1. If the ADC output codes are not full scale (i.e.
000 0000 and 111 11111) the AGC gain is
incremented in 1.1 dB steps with a time constant of
1000/n µs, where n is the over-speed factor i.e. n = 1
for basic audio CD.
2. When full scale is detected at the output of the ADC
the AGC gain is fixed provided that full scale is
maintained and clipping does not occur for greater
than 20% of the time.
3. If clipping occurs for more than 20% of the time, then
the AGC gain is reduced in 1.1 dB steps with a time
constant of 60/n µs.
The ADC and AGC electrical characteristics are specified
in Chapter 9.
7.4.3ANALOG SETTINGS REGISTER 1(ANASET1)
Table 21 Analog Settings Register 1 (address 15H) - WRITE
76543210
GainControlMaxGainStepUpStepDownPowerDown−−−
Table 22 Description of AnaSet1 bits
BITSYMBOLDESCRIPTION
7GainControl If GainControl = 0, then gain control is in Hold mode. If GainControl = 1, then automatic
gain control is on.
6MaxGainIf MaxGain = 0, then there is no gain limit. If MaxGain = 1, then the maximum gain is
7.66 dB.
5StepUpIf StepUp = 1, then step up gain by one LSB.
4StepDownIf StepDown = 1, then step down gain by one LSB.
3PowerDown If PowerDown = 0, then analog blocks are powered up. If PowerDown = 1, then analog
blocks are powered down.
2to0−These 3 bits are reserved and must be set to a logic 0s.
2000 Mar 2118
Philips SemiconductorsPreliminary specification
Channel encoder/decoder CDR60SAA7392
7.5Bit recovery
The bit recovery block (shown in Fig.6) contains the slice
levelcircuitry, anoise filtertolimit theHF-EFMsignal noise
contribution, an adaptive slicer circuit and a digital PLL.
These blocks can be controlled via the microprocessor.
The channel rate should always obey the following
constraints:
• It should be less than 2 × the system clock
• It should be greater than 0.25 × the system clock.
In this clock range reliable bit clock detection is possible.
All data found will be written to the FIFO. It is the
responsibility ofthe user toselect BCLK and system clock
values so that the FIFO operation is controlled.
The digital noise filter runs on the PLL bit clock and limits
the bandwidthof the incomingsignal to 0.25 ofthe PLL bit
clock frequency. The characteristics of the filter are:
• Passband: 0 to 0.22 f
• Stopband: 0.28 fb to (f
b
− 0.28 fb)
clk
• Rejection: −28 dB.
The slice level determination circuit compensates the
incoming signalasymmetry component. Thebandwidth of
this circuit is programmable via register PLLSet.
A programmable (one tap presetable, asymmetrical)
equaliser is used in the bit detection circuit. The first and
last tap settings are different. Possible tap values are
settable via register PLLEqu.
The advanced detector has two extra detection circuits
(adaptive slicer and run length 2 push-back) which are
controlled via the VitSet register, that allow improved
margin in the bit detector.
The adaptive slicer does a second stage slice operation;
thebandwidth ishigherthan thefirst slicer. Itcan beturned
on/off via the VitSet register.
If the advanced detector is switched on all run length 2
symbols are pushed back to run length 3. The circuit will
determine thetransition that was most likely to bein error,
and shift the transition on that edge.
GAIN CONTROLLED
handbook, full pagewidth
AMPLIFIER
HIN
GAIN CONTROL
BLOCK
ADC
+
+
−
NOISE
FILTER
SLICE LEVEL
DETERMINE
DIGITAL
EQUALIZER
clocked on PLL clock
DIGITAL
PLL
RMS JITTER
MEASUREMENT
Fig.6 Block diagram of bit recovery block.
VITERBI
DETECTOR
ZERO TRANS
DETECTOR
PLL frequency
jitter value
slice level
MULTIPLEXER
MEAS1
MGR796
2000 Mar 2119
Philips SemiconductorsPreliminary specification
Channel encoder/decoder CDR60SAA7392
7.5.1DIGITAL PLL
The digital PLL will recover the channel bit clock. As the
capturerange ofthe PLLitselfis limited,lockdetectors and
2 capture aids are present. In total three different PLL
operation modes exist: In-lock, Inner-lock aid and
Outer-lock aid.
The PLL behaviour during in-lock (the normal on-track
situation) can be best explained in the frequency domain.
The PLL operation is completely linear during in-lock
situations. The open-loop response of the PLL is given in
Fig.7. The three frequencies, f0(integrator cross-over
frequency), f1(PLL bandwidth) and f2(low-pass
bandwidth) are programmable via register PLLSet.
To extend the PLL capture range two lock aids are used:
• Inner lock aid: has a capture range of ±10% and will
bring the PLL frequency to the lock point
• Outer lock range: has no limitation on capture range,
and will bring the PLL within the range of the inner lock
range.
Two outer lock aids can be used:
• Run length 3 deviation detector: this circuit is known to
be sensitive to systematic over/under equalization; this
over/under equalizationcan becounter-acted by writing
a non-zero phase offset value to register PLLLock.
• Frequency measurement detector: this circuit regulates
the PLL frequency so that the average number of EFM
transitions is a fixed fraction of the PLL bit clock; the
transition frequency is settable via register PLLFMeas.
• PLL frequency signal: the most significant 8 bits are
available via register PLLLock
• Asymmetry signal: the 8-bit signal in 2’s complement
form is available via register PLLSet
• Jitter signal: the most significant 8 bits are available via
register PLLFreq. This gives an impression of the
detection jitter after all processing is done.
jitter<9:0> = average ((jitter individual
transition)2× 8192)
To obtain the jitter in the bit clocks the jitter<9:0> value
must be divided by 8192 and square routed. Note that
the jitter<9:0> overestimates thejitter (byapproximately
rms jitter increase of 0.03 bit clock), because the
quantization of the zero transitions is in 4 intervals.
Note the jitter is measured before the bit detection and
contains contributions due to various imperfections in
the complete signal path; i.e. disc, preamplifier, ADC,
limited bitwidths, PLL performance, internal filter noise,
asymmetry compensation, equalizer.
• Internal lock flags: The internally generated inner-lock
signal (f_lock_in), lock signal (lock_in) and flag that
indicates when a run length 14 is detected
(long_symbol) are available via register PLLEqu.
handbook, halfpage
amplitude
(dB)
Programmability/observability is built into the PLL. Its
operation can be influenced in two ways:
• It is possible to select the state the PLL is in (in-lock,
near-lock, outer-lock) via register PLLLock
• It is possible to preset the PLL frequency to a certain
value via registers PLLEqu and PLLFreq.
The operation of the bit detector can be monitored by the
microprocessor and via the MEAS1 pin. Four signals are
available for measurement:
2000 Mar 2120
f
2
f
0f1
frequency (Hz)
Fig.7 PLL bode diagram.
MGR797
Philips SemiconductorsPreliminary specification
Channel encoder/decoder CDR60SAA7392
7.5.2MEAS1 PIN
The MEAS1 pin carries the 3 measurement signals: jitter
(sampled twice), PLL frequency, and asymmetry. Each
frame consists of 64 bits (each 4 system clock periods
long), beginning with a start bit, then data bits then pause
bits (see Fig.8). The start bit is always preceded by
17 pause bits; and the intermediate start bits at
locations 12, 24 and 36 guarantee that no other ‘1’ bit is
preceded by 17 ‘0’ bits, making the start detection easy.
The structure of the frame is described in Table 23 and
shown in Fig.8.
Table 23 Frame structure
BITVALUEFUNCTION
0logic 1start bit
1 to 10jitter<9:0>jitter word
11logic 0
12logic 1intermediate start bit
13 to 22pllfreq<9:0> PLL frequency word
23logic 0
24logic 1intermediate start bit
25 to 32assym<7:0> asymmetry word
33logic 0
34logic 1intermediate start bit
37 to 46jitter<9:0>second sample of jitter
word
47 to 63logic 0pause
handbook, halfpage
pausedata bits
bit 0bit 1
start
bit
bit 2bit 3
MGR798
Fig.8 Format on MEAS1 pin.
2000 Mar 2121
Philips SemiconductorsPreliminary specification
Channel encoder/decoder CDR60SAA7392
7.5.3PLL LOCK SELECT REGISTER (PLLLOCK)
The behaviour of this register is dependent upon whether its being read or written. The behaviour for the write operation
is described in Tables 24 to 27. When read the 8 MSBs of the PLL frequency counter are returned; this is described in
Tables 24 and 28.
Table 25 Description of PLLLock bits for write operation
BITSYMBOLDESCRIPTION
7LockOrideWhen LockOride = 0, then automatic lock behaviourselected, PLLForceL<3:0>must be
set to ‘0000’. When LockOride = 1, then PLL manual override, PLLForceL<3:0> must
also be programmed.
6PhaOset.2These 3 bits are used to select the phase override settings; see Table 26.
5PhaOset.1
4PhaOset.0
3PLLForceL.3 These 4 bits are used to select the PLL lock; see Table 27.
2PLLForceL.2
1PLLForceL.1
0PLLForceL.0
0000automatic lock behaviour
0001force PLL in-lock
0100force PLL into outer-lock
0110force PLL into inner-lock
1000force PLL into Hold mode (PLL frequency can be
forced using preset value in register PLLFreq)
XXXXall other combinations are reserved
2000 Mar 2122
Philips SemiconductorsPreliminary specification
Channel encoder/decoder CDR60SAA7392
Table 28 Description of PLLock bits for read operation
BITSYMBOLDESCRIPTION
7 to 0PLLFreq<7:0> This register holds the 8 MSBs of the PLL frequency counter. The PLL frequency is
calculated as shown below:
f
(Hz)
PLL
7.5.4PLL B
The function of this register is dependent upon whether its being read or written. The function for the write operation is
described in Tables 29 to 34. Note the measurement conditions are: system clock = 2.15 MHz, bit clock = 4.3 MHz,
bandwidth is proportional to the system clock.
When read this register returns the 8-bit PLL asymmetry value, see Table 29.