PHILIPS TZA3015HW

INTEGRATED CIRCUITS
DATA SHEET
TZA3015HW
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
Preliminary specification
Supersedes data of 2003 Oct 06
2003 Dec 16
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
FEATURES
General
• A-rabitte(1): supports any bit rate from 30 Mbit/s to
3.2 Gbit/s with one single reference frequency
• 4-bit parallel interface
• Selectable Double Data Rate (DDR, half clock rate) or
Single Data Rate (SDR) clocking scheme on parallel
interface, enabling easy interfacing with FPGA devices
Multiplexer
• 4 : 1 multiplexing ratio
• I2C-bus and pin programmable
• Supports co-directional and contra-directional clocking
• Six selectable reference frequency ranges
• 4-stage FIFO for wide tolerance to clock skew
• Transmitter, receiver and transceiver modes
• Rail-to-rail parallel inputs compliant with LVPECL,
Current-Mode Logic (CML) and LVDS
• Clean-up loop back mode
• Programmable parity checking
• Line loop back mode
• CML data and clock outputs.
• Diagnostic loop back mode
• Serial loop timing mode
Demultiplexer
• Single 3.3 V power supply.
• 1 : 4 demultiplexing ratio
Limiter
• Adjustable LVDS output swing
• Frame detection for SDH/SONET and Gigabit Ethernet
• Limiting amplifier with typical 5 mV input sensitivity
(GE) frames.
• Received Signal Strength Indicator (RSSI)
• Loss Of Signal (LOS) indicator with adjustable threshold
I2C-bus configurable options
• Differential overvoltage protection.
• Programmable frequency resolution of 10 Hz
• Independent receive and transmit bit rate
Data and clock recovery and synthesizer
• Supports any bit rate from 30 Mbit/s to 3.2 Gbit/s when
using I2C-bus interface
• Slice level adjustment to improve Bit Error Rate (BER)
• Supports eight pre-programmed (pin selectable) bit
rates:
• Adjustable swing for CML serial data and clock outputs
• Six reference frequency ranges
• Programmable polarity of RF I/Os
• Clock versus data swap for optimum connectivity
– SDH/SONET rates at 155.52 Mbit/s, 622.08 Mbit/s,
2488.32 Mbit/s and 2666.06 Mbit/s
(STM16/OC48 + FEC)
• Swap of parallel bus for optimum connectivity
• Mute function for a forced logic 0 output state
– Gigabit Ethernet at 1250 Mbit/s and 3125 Mbit/s
• Programmable parity
– Fibre Channel at 1062.5 Mbit/s and 2125 Mbit/s.
• Programmable 32-bit frame detection.
• Provides stable clock signal at LOS
• Frequency lock indicator for DCR
• Loss Of Lock (LOL) indicator for synthesizer
• ITU-T compliant jitter tolerance for Data and Clock
Recovery (DCR)
• ITU-T compliant jitter transfer for DCR in clean-up loop
back mode
• ITU-T compliant jitter generation for synthesizer.
(1) A-rate is a trademark of Koninklijke Philips Electronics N.V.
2003 Dec 16
2
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
APPLICATIONS
GENERAL DESCRIPTION
• Any optical transmission system with line rates between
30 Mbit/s and 3.2 Gbit/s
The TZA3015HW is a fully integrated optical network
transceiver containing a limiter, data and clock recovery
circuit, clock synthesizer, 1 : 4 demultiplexer and 4 : 1
multiplexer.
• Physical interface IC in receive and transmit channels
• Transponder applications
The A-rate feature allows the IC to operate at any bit rate
between 30 Mbit/s and 3.2 Gbit/s with one single
reference frequency.
• Dense wavelength division multiplexing systems
• Due to DDR clocking option, the ultimate physical
interface for FPGA based designs.
All clock signals are generated using a fractional N
synthesizer with 10 Hz resolution offering a true
continuous rate operating. For full configuration flexibility
the transceiver can be programmed by pin and via the
I2C-bus.
ORDERING INFORMATION
TYPE
NUMBER
TZA3015HW
2003 Dec 16
PACKAGE
NAME
DESCRIPTION
VERSION
HTQFP100
plastic thermal enhanced thin quad flat package; 100 leads; body
14 × 14 × 1 mm; exposed die pad
SOT638-1
3
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
BLOCK DIAGRAM
CREF
FIFORESET
PAREVEN
OVERFLOW
81
FREF0
FREF1
ENRX
ENTX
93 94
42
89, 90 91
79, 80
31
LM1
LM2
TXSD
TXSDQ
TXSC
TXSCQ
ENTXSC
RXSD
RXSDQ
45
86, 87
64
65
CLK
MUX
÷4
32
LM
MUX
27
LM0
TXPARERR/
TXPARERRQ
PHASE
SHIFT
CLOCK
SYNTHESIZER
÷R
95
TXPCO/
TXPCOQ ENDDR
LOL
CREFQ
83
82
TXPRSCL/
TXPRSCLQ
LOOP
MODE
SELECT
28
29
c
LOOP
d
5
BUF
6
PARITY
CHECK
AND
BUS SWAP
MUX
4:1
d
BUF
MUX
c
3
DLB
MUX
16
PARITY
GENERATOR
AND
BUS SWAP
PHASE
DETECTOR
LIM
17
DMX
1:4
RSSI
RSSI
LOSTH
RREF
LOS
SCL(DR2)
SDA(DR1)
CS(DR0)
UI
VCO
LPF
20
ENDDR
19
TZA3015HW
LOS
14
FREQUENCY
WINDOW DETECTOR
21
76
77
TXPAR/
TXPARQ
73
74
TXPD3/
TXPD3Q
71
72
TXPD2/
TXPD2Q
69
70
TXPD1/
TXPD1Q
67
68
48
49
TXPD0/
TXPD0Q
61
62
RXPD3/
RXPD3Q
59
60
RXPD2/
RXPD2Q
57
58
RXPD1/
RXPD1Q
55
56
RXPD0/
RXPD0Q
52
53
RXPC/
RXPCQ
46
47
43
RXFP/
RXFPQ
RXPAR/
RXPARQ
ENBA
24
23
CLEAN-UP
PLL
I2C-BUS
22
INTERRUPT
CONTROLLER
13
25
VDD
(1)
VEE
(2)
VCCD
96
98, 99
38
37
IPUMP
INWINDOW
WINSIZE
RXPRSCL/
RXPRSCLQ
(1) Connected to pins 2, 12, 26, 33, 35, 40, 50, 84 and 100.
(2) Connected to pins 4, 7, 8, 11, 36, 39, 44, 51, 54, 63, 66, 75, 78, 85 and 97.
Fig.1 Block diagram.
2003 Dec 16
CLKDIR
FIFO
9
10
88
TXPC/
TXPCQ
4
1, 34
VCCO
15, 18,
92
VCCA
30
INT
41
LOWSWING
mgu679
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
PINNING
SYMBOL
VEE
SYMBOL
PIN
die
pad
DESCRIPTION
PIN
DESCRIPTION
VCCD
36
supply voltage (digital part)
common ground plane
WINSIZE
37
wide and narrow frequency
detect window select input
VCCO
1
supply voltage (clock generator)
INWINDOW
38
frequency window detector output
VEE
2
ground
VCCD
39
supply voltage (digital part)
ENTXSC
3
enable serial clock
VEE
40
ground
VCCD
4
digital supply voltage
LOWSWING
41
enable low LVDS swing
TXSD
5
serial data output
FREF0
42
reference frequency select
input 0
TXSDQ
6
serial data output inverted
VCCD
7
supply voltage (digital part)
ENBA
43
enable byte alignment
44
supply voltage (digital part)
VCCD
8
supply voltage (digital part)
VCCD
TXSC
9
serial clock output
ENDDR
45
enable DDR
TXSCQ
10
serial clock output inverted
RXFP
46
frame pulse output
VCCD
11
supply voltage (digital part)
RXFPQ
47
frame pulse output inverted
48
parity output
VEE
12
ground
RXPAR
UI
13
user interface select input
RXPARQ
49
parity output inverted
RREF
14
reference resistor input
VEE
50
ground
51
supply voltage (digital part)
52
parallel clock output
VCCA
15
supply voltage (analog part)
VCCD
RXSD
16
serial data input
RXPC
RXSDQ
17
serial data input inverted
RXPCQ
53
parallel clock output inverted
54
digital supply voltage
VCCA
18
supply voltage (analog part)
VCCD
LOSTH
19
loss of signal threshold input
RXPD0
55
parallel data output 0
RSSI
20
received signal strength indicator
output
RXPD0Q
56
parallel data output 0 inverted
RXPD1
57
parallel data output 1
58
parallel data output 1 inverted
LOS
21
loss of signal output
RXPD1Q
CS(DR0)
22
chip select output (data rate
select input 0)
RXPD2
59
parallel data output 2
RXPD2Q
60
parallel data output 2 inverted
SDA(DR1)
23
I2C-bus serial data input and
output (data rate select input 1)
RXPD3
61
parallel data output 3
SCL(DR2)
24
I2C-bus serial clock input (data
rate select input 2)
RXPD3Q
62
parallel data output 3 inverted
VCCD
63
supply voltage (digital part)
TXPC
64
parallel clock input
TXPCQ
65
parallel clock input inverted
VCCD
66
supply voltage (digital part)
TXPD0
67
parallel data input 0
TXPD0Q
68
parallel data input 0 inverted
VDD
25
supply voltage (digital)
VEE
26
ground
LM0
27
loop mode select input 0
LM1
28
loop mode select input 1
LM2
29
loop mode select input 2
INT
30
interrupt output
ENRX
31
enable receiver
ENTX
32
enable transmitter
VEE
33
ground
VCCO
34
supply voltage (clock generator)
VEE
35
ground
2003 Dec 16
5
TXPD1
69
parallel data input 1
TXPD1Q
70
parallel data input 1 inverted
TXPD2
71
parallel data input 2
TXPD2Q
72
parallel data input 2 inverted
TXPD3
73
parallel data input 3
TXPD3Q
74
parallel data input 3 inverted
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
SYMBOL
PIN
TZA3015HW
DESCRIPTION
SYMBOL
PIN
DESCRIPTION
VCCD
75
supply voltage (digital part)
TXPRSCL
89
prescaler synthesizer output
TXPAR
76
parity input
TXPRSCLQ
90
TXPARQ
77
parity input inverted
prescaler synthesizer output
inverted
VCCD
78
supply voltage (digital part)
LOL
91
loss of lock output
TXPCO
79
transmitter parallel clock output
VCCA
92
supply voltage (analog part)
TXPCOQ
80
transmitter parallel clock output
inverted
CREF
93
reference clock input
CREFQ
94
reference clock input inverted
FREF1
95
reference frequency select
input 1
IPUMP
96
clean-up PLL charge pump
output
VCCD
97
supply voltage (digital part)
RXPRSCL
98
prescaler DCR output
RXPRSCLQ
99
prescaler DCR output inverted
VEE
100
ground
PAREVEN
81
parity select input (odd or even)
OVERFLOW
82
FIFO overflow alarm output
FIFORESET
83
FIFO reset input
VEE
84
ground
VCCD
85
supply voltage (digital part)
TXPARERR
86
parity error output
TXPARERRQ
87
parity error output inverted
CLKDIR
2003 Dec 16
88
selection input between co- and
contra-directional clocking
6
Philips Semiconductors
Preliminary specification
76 TXPAR
78 VCCD
77 TXPARQ
79 TXPCO
81 PAREVEN
80 TXPCOQ
82 OVERFLOW
83 FIFORESET
85 VCCD
84 VEE
86 TXPARERR
87 TXPARERRQ
88 CLKDIR
89 TXPRSCL
TZA3015HW
90 TXPRSCLQ
92 VCCA
91 LOL
93 CREF
94 CREFQ
95 FREF1
96 IPUMP
97 VCCD
98 RXPRSCL
100 VEE
handbook, full pagewidth
99 RXPRSCLQ
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
VCCO
1
75 VCCD
VEE
2
74 TXPD3Q
ENTXSC
3
73 TXPD3
VCCD
4
72 TXPD2Q
TXSD
5
71 TXPD2
TXSDQ
6
70 TXPD1Q
VCCD
7
69 TXPD1
VCCD
8
68 TXPD0Q
TXSC
9
67 TXPD0
TXSCQ 10
66 VCCD
65 TXPCQ
VCCD 11
VEE 12
64 TXPC
TZA3015HW
UI 13
63 VCCD
RREF 14
62 RXPD3Q
VCCA 15
61 RXPD3
RXSD 16
60 RXPD2Q
RXSDQ 17
59 RXPD2
VCCA 18
58 RXPD1Q
LOSTH 19
57 RXPD1
RSSI 20
56 RXPD0Q
Fig.2 Pin configuration.
2003 Dec 16
7
VEE 50
RXPAR 48
RXPARQ 49
RXFPQ 47
RXFP 46
ENDDR 45
ENBA 43
VCCD 44
FREF0 42
LOWSWING 41
VCCD 39
VEE 40
INWINDOW 38
VCCD 36
WINSIZE 37
VCCO 34
VEE 35
VEE 33
ENTX 32
51 VCCD
ENRX 31
52 RXPC
VDD 25
INT 30
SCL(DR2) 24
LM2 29
54 VCCD
53 RXPCQ
LM1 28
SDA(DR1) 23
LM0 27
55 RXPD0
CS(DR0) 22
VEE 26
LOS 21
MGU680
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
Some functions of the TZA3015HW can be controlled both
using pre-program mode and via the I2C-bus. In these
cases, an extra I2C-bus bit called I2C<pinname> is
available to set the programming precedence to
pre-programmed or I2C-bus bit (default is selection by
pre-programmed).
FUNCTIONAL DESCRIPTION
The TZA3015HW contains the following main blocks:
• General part: configuration via I2C-bus mode or
pre-programmed mode
• Receiver part: limiting amplifier, data and clock recovery
and demultiplexer
PRE-PROGRAMMED MODE
• Transmitter part: clock synthesizer and multiplexer.
The TZA3015HW is primarily intended to be programmed
via the I2C-bus. If no I2C-bus control is present in the
application, the TZA3015HW can be used in the
pre-programmed mode (pin UI = LOW), with reduced
functionality. The TZA3015HW functions that are
accessible in the pre-programmed mode and their
associated pins are:
• All pre-programmed modes are supported by one single
reference frequency
General
CONFIGURATION
The IC features two types of user interface: I2C-bus or
direct pin programming of eight predefined modes. The
mode selection is set by pin UI.
The I2C-bus mode is operational and A-rate functionality is
enabled if pin UI is left open-circuit or connected to VCC
(see Table 1). If pin UI is connected to VEE, the eight
pre-programmed modes can be selected with pins
CS(DR0), SDA(DR1) and SCL(DR2).
Table 1
• The redefined pins DR0 to DR2 act as standard CMOS
inputs that select any of the desired data rates; see
Table 3
• Transceiver mode (transceiver, transmitter, receiver,
off) (ENRX and ENTX)
Truth table for pin UI
UI
MODE
LOW
pre-programmed
HIGH
I2C-bus
• Enable serial clock output (ENTXSC)
PIN 22
PIN 23
PIN 24
DR0
DR1
DR2
• Loss of signal threshold setting (LOSTH)
CS
SDA
SCL
• Select loop mode (LM0 to LM2)
• Automatic byte alignment for SDH/SONET or Gigabit
Ethernet (ENBA)
I2C-BUS MODE
• Frame detection for SDH/SONET or Gigabit Ethernet
In I2C-bus mode the IC can be configured by using pins
SDA and SCL. Pin CS has to be HIGH during the I2C-bus
read or write actions. When pin CS is made LOW, the
programmed configuration remains active, but signals
SDA and SCL are ignored. In this way, all ICs in the
application with the same I2C-bus address (e.g. other
TZA3015HWs) are individually accessible.
The
I2C-bus
• Even parity generation (PAREVEN)
• In window detection (INWINDOW)
• Sizeable frequency window: 1000 or 0 ppm (WINSIZE)
• Temperature alarm (INT, open drain)
• Co-directional or contra-directional clocking scheme
(CLKDIR)
address is given in Table 2.
• Enable DDR for both receiver and transmitter (ENDDR)
Table 2
Device address of the TZA3015HW
• CML serial RF outputs with typical 300 mV (p-p)
single-ended signal (DC-coupled load)
DEVICE ADDRESS BITS
R/W
A6
A5
A4
A3
A2
A1
A0
1
0
1
0
1
0
0
• Loss of lock detection (LOL)
• FIFO overflow indication (OVERFLOW)
X
• FIFO reset (FIFORESET)
• Supported reference frequencies: 19.44, 38.88, 155.52
and 622.08 MHz.
After power-up, the TZA3015HW initiates a Power-On
Reset (POR) sequence to restore the default settings of
the I2C-bus registers, regardless of the user interface. See
Table 21 for the defaults and a detailed list of all I2C-bus
registers and the meaning of their contents.
2003 Dec 16
8
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
Table 3
TZA3015HW
Truth table for pins DR2 to DR0 (pin UI = VEE)
PROTOCOL
BIT RATE
(Mbit/s)
DR2
DR1
DR0
LOW
LOW
LOW
STM1/OC3
155.52
LOW
LOW
HIGH
STM4/OC12
622.08
LOW
HIGH
LOW
STM16/OC48
2488.32
handbook, halfpage
VCCA
IN
50 Ω
LOW
HIGH
HIGH
STM16 + FEC 2666.06
HIGH
LOW
LOW
GE
50 Ω
1250.00
HIGH
LOW
HIGH
10GE
HIGH
HIGH
LOW
Fibre Channel 1062.50
HIGH
HIGH
HIGH
Fibre Channel 2125.00
INQ
3125.00
VEE
MDB385
Fig.3 Limiter input termination configuration.
Receiver
LIMITING AMPLIFIER
The TZA3015HW contains a limiting amplifier (see Fig.3).
Received Signal Strength Indicator (RSSI)
To achieve optimum receiver sensitivity for any bit rate, the
bandwidth of the amplifier is automatically scaled with the
bit rate. Wideband noise of the optical front-end (photo
detector and transimpedance amplifier) is thus reduced for
lower bit rates. When using the I2C-bus, the bandwidth of
the amplifier can be set independently of the bit rate with
bits AMP[2:0] in register LIMCON (D3h).
The signal strength at the input is measured with a
logarithmic detector. The logarithmic detector converts the
input signal amplitude into a voltage which can be
measured at pin RSSI. The RSSI reading has a dynamic
range of 40 dB with a sensitivity (SRSSI) of 17 mV/dB
(typical) for a Vi(p-p) range of 5 to 500 mV (see Fig.4). VRSSI
can be calculated using the following formula:
The highest bandwidth is selected as default at power-up.
V i(p-p)
V RSSI = V RSSI(32mV) + S RSSI × 20log ----------------32 mV
where: VRSSI(32mV) = 680 mV (typical).
MCE412
handbook, full pagewidth
1.2
VRSSI
(V)
SRSSI
0.9
0.68
0.6
0.3
0
5
10
32
102
300
500
Vi(p-p) (mV)
Fig.4 VRSSI as a function of Vi(p-p).
2003 Dec 16
9
103
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
Loss Of Signal (LOS) indicator
Besides the analog RSSI output, a digital LOS indication is
present on the TZA3015HW. The RSSI level is internally
compared with a LOS threshold, which can be set by
connecting an external resistor to pin LOSTH or by means
of an internal DAC which is accessible via the I2C-bus.
VCCA
RSSI
RSSI
Bit I2CLOSTH of register LIMLOSCON (D1h) enables the
8-bit DAC, of which the value needs to be programmed
into register LIMLOSTH (D0h). The threshold level is
adjustable in 256 steps from 0 to 1.2 V.
LOS
LOS
compare
1.2 V
Vref
If the received signal strength is below the threshold value,
pin LOS will be HIGH. A default hysteresis of 3 dB is
applied in the comparator. The hysteresis can be set with
bits HTLC[2:0] in register LIMLOSCON (D1h). The
programmable range is 0 to 7 dB.
VEE
RREF
I
R2
ground
The polarity of the LOS output can be inverted by bit
LOSPOL of register LIMLOSCON (D1h) to provide more
flexibility in the application.
MGU681
Fig.5
LOSTH reference setting by external resistor
If the built-in DAC is not used, the reference voltage level
to pin LOSTH can be set by connecting an external resistor
(R2) between pin LOSTH and ground. VLOSTH is
determined by the resistor ratio between R2 and R1 (see
Fig.5). For resistor R1 a value of 10 to 20 kΩ is
recommended, yielding a current of 120 to 60 µA
through R1.
Setting the LOSTH reference level by
external resistors.
Slice level adjustment
Due to asymmetrical noise in some optical transmission
systems, a pre-detection signal-to-noise ratio
improvement can be achieved by adding a DC offset to the
input signal. This is done by the slice level circuit in the
TZA3015HW. The required offset depends on the photo
detector characteristics in the optical front-end and the
amplitude of the received signal. The slice level is
adjustable between −50 and +50 mV in 512 steps of
0.2 mV.
R2
VLOSTH = ------- × V ref
R1
Vref represents a temperature stabilized and accurate
reference voltage of 1.2 V. The minimum threshold level
corresponds to 0 V and the maximum to 1.2 V. Hence, the
value of R2 may not be higher than R1. The accuracy of
VLOSTH depends mainly on the matching of the two
external resistors.
Bit SLEN of register LIMLOSCON (D1h) enables the slice
function. The slice level is set by sign and magnitude
convention. The polarity sign is set by bit SLSGN in
register LIMLOSCON (D1h). The magnitude is set by an
8-bit DAC, accessible via register LIMSL (D2h), from
0 to 50 mV in 256 steps.
Apart from using resistors (R1 and R2) to set the LOS
threshold, an accurate external voltage source may also
be used.
The introduced offset is not present on input pins
RXSD(Q), in order not to affect the logarithmic RSSI
detector, which would detect the offset as a valid input
signal.
If no resistor is connected or an external voltage higher
than 2/3 × VCC is applied to pin LOSTH, the LOS detection
circuit (including the RSSI reading) is automatically
switched off to reduce power dissipation. This ‘auto power
off’ function only works in the pre-programmed mode.
I2C-bus mode allows flexible configuration.
2003 Dec 16
R1
10 kΩ
LOSTH
10
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
and takes over control if the VCO frequency drifts outside
the predefined frequency window. This might occur during
a ‘loss of signal’ situation. Due to the FWD, the VCO
frequency is always close to the required bit rate, enabling
rapid phase acquisition when the lost input signal returns.
DATA AND CLOCK RECOVERY (DCR)
The TZA3015HW recovers the clock and data contents
from the incoming bit stream; see Fig.6. The DCR uses a
combined frequency and phase locking scheme, providing
reliable and quick data acquisition on any bit rate between
30 Mbit/s and 3.2 Gbit/s.
Due to the loose coupling of 1000 ppm, the reference
frequency does not need to be highly accurate or stable.
Any crystal-based oscillator that generates a reasonably
accurate frequency (e.g. 100 ppm) will do. This only holds
if the TZA3015HW is used as a receiver since the
synthesizer of the transmitter uses the same reference
clock. The transmitter does need a very accurate
reference frequency.
At power-up, coarse adjustment of the free running
Voltage Controlled Oscillator (VCO) frequency is required.
This is achieved by the Frequency Window Detector
(FWD) circuit. The FWD is a conventional frequency
locked PLL. The FWD checks the VCO frequency, which
has to be within a 1000 ppm window around the required
frequency. The FWD then compares the divided VCO
frequency, also available on pins RXPRSCL(Q), with the
reference frequency on pins CREF(Q), usually 19.44 MHz.
If the VCO frequency is outside this window, the FWD
disables the Data Phase Detector (DPD) and forces the
VCO to a frequency within the window. As soon as the ‘in
window’ condition occurs, which is visible on pin
INWINDOW, the DPD is enabled and will lock on the
incoming bit stream. Since the VCO frequency is very
close to the expected bit rate, the phase acquisition will be
almost instantaneous, resulting in quick phase lock to the
incoming data stream.
Fractional N synthesizer in the DCR
The DCR section contains a fractional N synthesizer as
frequency acquisition aid for the A-rate functionality. This
allows the DCR to synchronize on incoming data,
regardless of the received bit rate. Any combination of bit
rate and reference frequency is possible, due to the 22 bits
fractional N synthesizer, allowing approximately 10 Hz
frequency resolution. The LSB (bit K0) should be set to
logic 1 to avoid limit cycles (cycles of less than maximum
length). This leaves 21 bits (bits K[21:1]), available for free
programming.
Although the VCO is now locked to the incoming bit
stream, the FWD is still supervising the VCO frequency
handbook, full pagewidth
recovered data
from limiting
amplifier and
DLB MUX
to
demultiplexer
recovered clock
DATA PHASE
DETECTOR
up
down
OCTAVE
DIVIDER
CHARGE PUMP
÷M
VOLTAGE
CONTROLLED
OSCILLATOR
LOOP FILTER
+
MAIN Frac
DIVIDER N, K ÷N
REFERENCE
DIVIDER
divided
CREF(Q)
up
FREQUENCY
WINDOW
DETECTOR
down
CHARGE PUMP
RXPRSCL(Q)
PRESCALER BUFFER
MGU683
INWINDOW
WINSIZE
Fig.6 Functional diagram of data and clock recovery.
2003 Dec 16
11
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
DCR programming
Table 5
Programming the DCR involves four dividers:
Most-common optical transmission protocols
PROTOCOL
• Reference divider R
BIT RATE
(Mbit/s)
OCTAVE
• Main divider N
10GE
3125.00
0
• Fractional divider K
2xHDTV
2970.00
0
• Octave divider M.
STM16/OC48 + FEC
2666.06
0
STM16/OC48
2488.32
0
DV-6000
2380.00
0
The first step is to determine in which octave the desired
bit rate fits, see Fig.7 and Tables 4 and 5. Figure 7 shows
the position of the most commonly used line rates in
relation to the defined octaves of the TZA3015HW.
Table 5 lists the most commonly used standards together
with the associated line rates. Table 4 clarifies the octave
definitions. This yields the value for the octave divider M.
The value for R is determined by the reference frequency
and the received bit rate (see Section “Reference clock
programming”).
handbook, 6halfpage 5
28.125 56.25
4
112.5
3
225
2
450
1
900
0
1800
3200
Mbits/s
MGU316
Fig.7
Table 4
Commonly used line rates and allocation of
octaves along a logarithmic bit rate scale.
Fibre Channel
2125.00
0
HDTV
1485.00
1
D-1 video
1380.00
1
DV-6010
1300.00
1
Gigabit Ethernet
1250.00
1
Fibre Channel
1062.50
1
OptiConnect
1062.50
1
ISC
1062.50
1
STM4/OC12
622.08
2
DV-6400
595.00
2
Fibre Channel
425.00
3
OptiConnect
265.63
3
Fibre Channel
212.50
4
ESCON/SBCON
200.00
4
STM1/OC3
155.52
4
FDDI
125.00
4
Fast Ethernet
125.00
4
Fibre Channel
106.25
5
OC1
51.84
6
The values for N and K are derived from the division
ratio (n.k). The division ratio (n.k) can be calculated with
the following formula:
Octave definition
bit rate × M × R
n.k = ---------------------------------------f ref
OCTAVE
M
LOWEST BIT
RATE (Mbit/s)
HIGHEST BIT
RATE (Mbit/s)
0
1
1800
3200
1
2
900
1800
2
4
450
900
k = fractional part of the division ratio
where:
n = integer part of the division ratio
3
8
225
450
bit rate = bit rate at serial input in Mbit/s
4
16
112.5
225
M = octave divider M
5
32
56.25
112.5
R = reference divider R
6
64
28.125
56.25
fref = reference frequency in MHz.
2003 Dec 16
12
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
handbook, full pagewidth
TZA3015HW
CALCULATE
N and K
n.k =
bit rate × M × R
fref
n is integer part
k is fractional part
yes
k=0?
no
RXNILFRAC = 1
RXNILFRAC = 0
0.25 < k < 0.75
no
yes
no
k ≤ 0.25 ?
yes
k ≥ 0.75 ?
no
yes
k = k + 0.5
N=2×n
N=2×n
N=2×n−1
k = k − 0.5
N=2×n+1
j = 21
k=k×2
k≥1?
no
yes
Kj = 1
Kj = 0
decimal to binary
conversion of
fractional part
k=k−1
j=j−1
j=0?
no
yes
K0 = 1
Write Kj into registers C3h, C4h, C5h
or E3h, E4h, E5h
Convert N to binary
and write into registers C1h, C2h
or E1h, E2h
END
MCE413
Fig.8 Flowchart for calculating N and K.
2003 Dec 16
13
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
Having calculated the division factor (n.k), the values for N and K can be calculated according to the flow depicted in the
flowchart of Fig.8.
The value of the octave divider M is programmed by bits RXDIV_M[2:0] in register RXOCTDIV (C0h). The value for the
main divider N is programmed by bits RXN[8:0] in registers RXMAINDIV1 (C1h) and RXMAINDIV0 (C2h). The value for
the fractional divider K is programmed by bits RXK[21:0] in registers RXFRACN2 to RXFRACN0 (C3h to C5h). Bit
RXNILFRAC in register RXFRACN2 (C3h) must be set depending on whether there is a fractional part or not.
Example 1: An SDH or SONET link has a bit rate of 2488.32 Mbit/s (STM16/OC48) and consequently fits in octave
number 0, so M = 1. Suppose the reference frequency provided at pins CREF(Q) is 77.76 MHz. This means that the
reference division R needs to be 4. The values of n and k can be calculated from the flowchart:
bit rate × M × R
2488.32 Mbits × 1 × 4
n.k = ---------------------------------------- = --------------------------------------------------------- = 128
f ref
77.76 MHz
Since k = 0 in this example, no fractional functionality is required, bit RXNILFRAC (register C3h), should be logic 1.
N = 2 × n and no correction is required. Consequently the appropriate values are: R = 4 (register A1h), M = 1
(register C0h) and N = 256 (registers C1h and C2h).
Example 2: An SDH STM16 or SONET OC48 link with FEC has a bit rate of 2666.057143 Mbit/s
(15/14 × 2488.32 Mbit/s) and consequently fits in octave number 0, so M = 1. Suppose the reference frequency provided
at pins CREF(Q) is 38.88 MHz. This means that the reference division R needs to be 2. The values of n and k can be
bit rate × M × R
2666.05714283 Mbits × 1 × 2
calculated from the flowchart: n.k = ---------------------------------------- = ----------------------------------------------------------------------------- = 137.1428571
f ref
38.88 MHz
This means that n = 137, k = 0.1428571 and bit RXNILFRAC (register C3h) should be logic 0. Since k < 0.25, k is
corrected to 0.6428571, while the corrected N becomes N = 273. Consequently the appropriate values are: R = 2
(register A1h), M = 1 (register C0h), N = 273 (registers C1h and C2h) and K = 10 1001 0010 0100 1001 0011 (registers
C3h to C5h). The FEC bit rate is usually quoted to be 2666.06 Mbit/s. Due to round off errors, this leads to a slightly
different value for k than in the example.
Example 3: A Fibre Channel link has a bit rate of 1062.50 Mbit/s and consequently fits in octave number 1, so M = 2.
Suppose the reference frequency provided at pins CREF(Q) is 19.44 MHz. This means that the reference division R
needs to be 1. The values of n and k can be calculated from the flowchart:
bit rate × M × R
1062.50 Mbits × 2 × 1
n.k = ---------------------------------------- = --------------------------------------------------------- = 109.3106996
f ref
19.44 MHz
This means that n = 109, k = 0.3107 and bit RXNILFRAC should be logic 0 (register C3h). Since k is between 0.25 and
0.75, k does not need to be corrected and N = 2 × n = 218. Consequently the appropriate values are: R = 1 (register
A1h), M = 2 (register C0h) and N = 218 (registers C1h and C2h). K = 01 0011 1110 0010 1000 0001 (registers C3h to
C5h).
Example 4: A non standard transmission link has a bit rate of 3012 Mbit/s and consequently fits in octave number 0, so
M = 1. Suppose the reference frequency provided at pins CREF(Q) is 20.50 MHz. This means that the reference
division R needs to be 1. The values of n and k can be calculated from the flowchart:
3012 Mbits × 1 × 1
bit rate × M × R
n.k = ---------------------------------------- = ------------------------------------------------ = 146.9268293
20.50 MHz
f ref
This means that n = 146, k = 0.9268293 and bit RXNILFRAC should be logic 0 (register C3h). Since k is larger than 0.75,
k needs to be corrected to 0.4268293 and N = 2 × n + 1 = 293. Consequently the appropriate values are: R = 1 (register
A1h), M = 1 (register C0h) and N = 293 (registers C1h and C2h). K = 01 1011 0101 0001 0010 1011 (registers C3h to
C5h).
If the I2C-bus is not used, the DCR can be set up for the eight pre-programmed bit rates by pins DR0 to DR2 with an
applied reference frequency of 19.44 MHz (see Table 3).
2003 Dec 16
14
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
Reference clock programming
Reference input
The reference clock, connected to pins CREF(Q), is used
for both the DCR frequency window detector and the
transmitter synthesizer. The reference clock is divided by
divider R. Pre-programmed operating in an SDH/SONET
application assumes the use of a reference clock with a
frequency that is a multiple (R) of 19.44 MHz. For other
applications, any reference frequency between
18 and 21 MHz may be used. If a reference frequency is
selected, any bit rate between 30 Mbit/s and 3.2 Gbit/s is
supported.
For optimum jitter performance and Power Supply
Rejection Ratio (PSRR), the sensitive reference input
should be driven differentially (see Fig.9). If the reference
frequency source (fref) is single-ended, the unused CREF
or CREFQ input should be terminated with an impedance
which matches the source impedance Rsource. The PSRR
can be improved by AC coupling the reference frequency
source to inputs CREF and CREFQ. Any low frequency
noise injected from the fref power supply will be attenuated
by the resulting high-pass filter. The low cut-off frequency
of the AC coupling must be lower than the reference
frequency, otherwise the reference signal will be
attenuated and the signal to noise ratio will be reduced.
The value of coupling capacitor C is calculated using the
1
formula: C > ----------------------------------2πR source f ref
The division ratio and reference frequency can be
programmed by the bits FREFI2C[2:0] of register REFDIV
(A1h) or by pins FREF0 and FREF1. Internally, the
reference frequency is always divided to the lowest
frequency range between 18 and 21 MHz and for
SDH/SONET applications to 19.44 MHz. This is done by
divider R which is set by the described pins and bits.
In the pre-programmed mode (Table 6) four ranges of
clock frequencies can be used by programming R through
pins FREF0 and FREF1. In I2C-bus mode (Table 7) two
additional ranges of clock frequencies can be used by
programming R through bits FREFI2C[2:0].
Table 6
VCCD
handbook, halfpage
Truth table for reference divider R in
pre-programmed mode
50 Ω
50 Ω
43 CREF
REFERENCE
FREQUENCY
PIN
DIVISION
FACTOR
SDH/SONET
R
FREF1 FREF0
(MHz)
HIGH
1
HIGH
LOW
2
38.88
36 to 42
LOW
HIGH
8
155.52
144 to 168
LOW
LOW
32
622.08
576 to 672
Table 7
FREF
I2C1
FREF
I2C0
REFERENCE
FREQUENCY
RANGE (MHz)
0
0
0
1
18 to 21
0
0
1
2
36 to 42
0
1
0
4
72 to 84
0
1
1
8
144 to 168
1
0
0
16
288 to 336
1
0
1
32
576 to 672
2003 Dec 16
C
Rsource
fref
18 to 21
DIVISION
FACTOR
R
C
Rsource
on-chip
Truth table for reference divider R in I2C-bus
mode
BIT
FREF
I2C2
19.44
42 CREFQ
RANGE
(MHz)
HIGH
VCC
Fig.9
off-chip
MDB060
Reference input with single-ended clock
source.
Prescaler outputs
The prescaler output RXPRSCL(Q) is the VCO frequency
of the DCR divided by the main division factor N. It can be
used as an accurate reference for another PLL, since it
corresponds to the recovered data rate. If needed, the
polarity of the prescaler outputs can be inverted by bit
RXPRSCLINV of register DDR&RXPRSCL (D5h).
If no prescaler information is desired, the output can be
disabled by bit RXPRSCLEN of the same register. Apart
15
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
INWINDOW output
from these settings, the signal amplitude can be set. This
parameter follows the settings of the LVDS outputs. For
programming details, see Section “LVDS outputs”.
The status of the FWD circuit is reflected in the state of pin
INWINDOW; HIGH for an ‘in window’ situation and LOW
whenever the VCO is outside the defined frequency
window. Due to the fact that the device enters the
frequency acquisition mode when out of window is
detected, the INWINDOW pin will have an intermittent
value when the input signal is not within the defined
window boundary.
FWD programming
The default width of the window for frequency acquisition
is 1000 ppm around the required bit rate. This window size
can be changed between 4000 and 250 ppm by bits
WINSIZE[2:0] of register DCRCON (C6h). This allows for
loose or tight coupling of the VCO to the applied reference
clock. Another feature is to define a window width of
0 ppm, by means of pin WINSIZE, see Table 8. This
effectively removes the dead zone from the FWD,
rendering the FWD into a classical PLL.
DEMULTIPLEXER
The demultiplexer converts the serial input bit stream to a
parallel format. The output data is available on a 4-bit
LVDS-bus, thus reducing the data frequency by a factor
four. Apart from the de-serializing function, the
demultiplexer comprises a parity calculator and a frame
header detection circuit.
The VCO will be directly locked to the reference signal
instead of the incoming bit stream. Apart from pin
WINSIZE, this mode can be invoked by bits I2CWINSIZE
and WINSIZE of register DCRCON(C6h).
Table 8
The calculated parity (even) is available at output pins
RXPAR(Q), whereas occurrence of the frame header
pattern in the data stream results in a one clock cycle
(parallel clock output) wide pulse on output pins RXFP(Q).
Truth table for pin WINSIZE
WINSIZE
FREQUENCY WINDOW (ppm)
LOW
0
HIGH
1000
Accurate clock generation during loss of signal
If pin ENBA is HIGH, automatic byte (word) alignment
takes place, formatting the parallel output to logical
nibbles. Apart from pin ENBA, this mode can be invoked
by bits I2CENBA and ENBA of register DMXCON (B8h).
A zero window size is especially interesting in the absence
of input data, since the frequency of the ‘recovered clock’
will be equal to the programmed line clock rate.
To support most commonly used transmission protocols,
the frame header pattern can be programmed to any 32-bit
pattern (see Section “Frame detection”).
Bit AUTOWIN of register DCRCON (C6h) (see Table 9)
makes the window size dependent on the LOS status of
the limiter. If the optical input signal is lost, the FWD
automatically selects the 0 ppm window size; i.e. a direct
lock to the reference frequency. This results in a stable and
defined output clock during LOS situations, while
automatically reverting back to normal DCR operating
when the input signal returns.
If required, the demultiplexer output can be forced into a
fixed logic state by bit DMXMUTE of register DMXCON
(B8h).
The highest supported parallel bus speed is 800 Mbit/s.
Frame detection
Byte alignment is enabled if the enable byte alignment
input (pin ENBA) is forced HIGH. Whenever a 32-bit
sequence matches the programmed header pattern, the
incoming data is formatted into logical bytes (being output
as nibbles) and a frame pulse is generated on differential
output pins RXFP(Q). Any header pattern can be
programmed through registers HEADER3 to HEADER0
(B0h to B3h). It is possible to enter a ‘don’t care’ for any bit
position, e.g. to program a header pattern that is much
shorter than 32 bits or to program a pattern with a gap in it.
The accuracy of the reference frequency needs to be
better than 20 ppm if the application has to comply with
ITU-T recommendations.
Table 9
Truth table for bit AUTOWIN
AUTOWIN
FREQUENCY WINDOW
0
FWD user defined
1
FWD dependent on LOS
2003 Dec 16
16
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
MSB HEADER
handbook, full pagewidth
LSB HEADER
bit 0
bit 31
HEADER3
0
0
0
1
0
1
1
1
0
1
1
0
0
0
1
0
HEADER0
HEADERX3
1
0
0
0
0
1
0
0
0
0
0
0
0
0
1
1
HEADERX0
X
0
0
1
0
X
1
1
0
1
1
0
0
0
X
X
received
data
data stream
MGU548
Fig.10 Example of programming the frame pattern (the symbol ‘X’ represents a don’t care).
For this, it is necessary to program registers HEADERX3
to HEADERX0 (B4h to B7h). Programming a logic 1 into
the HEADERX register will turn the corresponding bit in the
HEADER register into a don’t care bit, in this way the
HEADER register is masked. An example of programming
the framing pattern is shown in Fig.10.
boundary detection is enabled on the rising edge of ENBA
and remains enabled while ENBA is HIGH. Boundaries are
recognized on receipt of the second A2 byte and RXFP
goes HIGH for one RXPC clock cycle.
The four most significant bits of the first A2 byte in the
frame header are the first bits that appear on the outgoing
data bus (RXPD0 to RXPD3) with the correct alignment.
The default frame header pattern is F6F62828h,
corresponding to the middle section of the standard
SDH/SONET frame header (the last two A1 bytes plus the
first two A2 bytes).
When interfacing with a section terminating device, ENBA
must remain HIGH for a full frame after the initial frame
pulse. This is to allow the section terminating device to
verify internally that frame and byte alignment are correct
(see Fig.12). Byte boundary detection is disabled on the
first RXFP pulse after ENBA has gone LOW.
If signal ENBA is LOW, no active alignment takes place.
However, if the framing pattern happens to occur in the
formatted data, a frame pulse will continue to be output on
pins RXFP(Q).
Figure 13 shows frame and byte boundary detection
activated on the rising edge of ENBA and deactivated by
the first RXFP pulse after ENBA has gone LOW.
Receiver framing in SDH/SONET applications
Figure 11 shows a typical SDH/SONET re-frame
sequence involving byte alignment. Frame and byte
2003 Dec 16
17
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
handbook, full pagewidth
serial clock
ENBA
serial data
A1
A1
A1
A2
RXPD0 to
RXPD3
A2
A2, bits 0-3
valid
data
invalid data
RXPC
RXFP
MGU342
Fig.11 Frame and byte detection in SDH/SONET application.
handbook, halfpage
boundary detection enabled
boundary detection
enabled
handbook, halfpage
ENBA
ENBA
RXFP
RXFP
MCE414
MCE415
Fig.12 ENBA operating time with section
terminating device.
Fig.13 Alternate ENBA timing.
Receiver framing in other applications
Parity generation
In other applications frame headers may be used that are
shorter than 32 bits, e.g. 10 bits for Gigabit Ethernet. The
position of the frame header in the header register can be
chosen freely, but determines the boundary of the parallel
data on pins RXPD0(Q) to RXPD3(Q). After alignment, the
header bits that are programmed by bits H12 to H15 of
register HEADER1 (B2h), appear at the RXPD(Q) outputs.
A frame pulse appears at output RXFP(Q) at the same
time.
Outputs RXPAR(Q) provide the even parity of the nibble
that is currently available on the parallel bus. With bit
RXPARINV of register RXMFOUTC0 (D4h), the parity can
be made odd. If no parity check is required, bit RXPAREN
of register RXMFOUTC0 (D4h) can be programmed to
disable this output, to reduce power dissipation.
2003 Dec 16
18
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
Transmitter
Because of the 22 bits fractional N capability, any
combination of bit rate (30 Mbit/s to 3.2 Gbit/s) and
reference frequency between 18 and 672 MHz is possible.
The LSB (bit k0) of the fractional divider, should be set to
logic 1 to avoid limit cycles. These are cycles of less than
maximum length, which generate spurs in the frequency
spectrum. This leaves bits k[21:1] available for
programming the fraction, allowing approximately 10 Hz of
frequency resolution without altering the reference
frequency.
CLOCK SYNTHESIZER
The transmitter frequency can be set independently of the
receiver frequency. For this a clock synthesizer is provided
that drives the multiplexer. Just like the DCR the clock
synthesizer is built around a fractional N synthesizer
offering A-rate functionality for the transmit path.
The clock synthesizer consists of a VCO, several dividers,
a phase frequency detector, an integrated loop filter, a lock
detection circuit and a prescaler output buffer (see Fig.14).
To meet most transmission standards, the reference
frequency should be very accurate. In order to be able to
synthesize a clean RF clock that is compliant with the most
stringent jitter generation requirements, it should also be
very clean in terms of phase noise.
The internal VCO is phase-locked to the reference clock
signal provided at pins CREF(Q). This frequency is
internally scaled down (if necessary) to a frequency in the
range of 18 to 21 MHz by divider R.
handbook, full pagewidth
LOL
OCTAVE
DIVIDER
up
divided
CREF(Q)
PHASE
FREQUENCY
DETECTOR
down
CHARGE PUMP
AND
LOOP FILTER
VCO
÷M
to LM MUX
and
multiplexer
MAIN
DIVIDER N, K
÷N
TXPRSCL(Q)
MGU682
Fig.14 Schematic diagram of the clock synthesizer.
2003 Dec 16
19
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
Parallel bus clocking schemes
All parts of the PLL are internal; no external components
are required. This allows for easy application.
The TZA3015HW supports both co-directional and
contra-directional clocking schemes for the parallel data
bus. The clocking application can be selected by pin
CLKDIR or by the bit CLKDIR of register MUXCON0 (F1h).
Co-directional clocking is default.
Programming the clock synthesizer involves four dividers:
• Reference divider R
• Main divider N
• Fractional divider K
• Octave divider M.
Table 10 Truth table for clocking scheme
PIN CLKDIR BIT CLKDIR
This is essentially the same as for the DCR.
The first step is to determine in which octave the desired
bit rate fits, see Tables 4 and 5 and Fig.7. Figure 7 shows
the position of the most commonly used line rates in
relation to the defined octaves of the TZA3015HW.
Table 5 clarifies the octave definitions; this yields the value
for the octave divider M. The value for R is determined by
the reference frequency and the received bit rate (see
Section “Reference clock programming”).
LOW
0
contra-directional clocking
HIGH
1
co-directional clocking
In the co-directional clocking mode, the parallel clock
signal is applied to pins TXPC(Q). The parallel clock signal
is generated in the data processing device (e.g. a framer).
The co-directional application is depicted in Fig.15. The
data processing device may be clocked by an external
crystal or by the parallel clock output TXPCO(Q) of the
TZA3015HW. This clock output is internally derived from
the synthesizer. If the parallel clock output TXPCO(Q) is
not required, it can be disabled in order to save dissipation.
This is done by programming bit TXPCOEN of register
TXMFOUTC (F2h).
Prescaler output
The prescaler output TXPRSCL(Q) is the VCO frequency
of the synthesizer divided by the main division factor N.
If the synthesizer is in-lock, the frequency is equal to the
reference frequency at CREF(Q) divided by R. It can be
used as an accurate reference for another PLL. If needed,
the polarity of the prescaler outputs can be inverted by bit
TXPRSCLINV of register TXMFOUTC (F2h).
In a contra-directional clock application, no clock is
provided on pin TXPC (see Fig.16). The clock that
samples the input data on the parallel bus, is an internal
clock derived from signal TXPCO. In this application, the
part providing the parallel data has to be clocked with the
clock signal TXPCO(Q). In order to alleviate timing
problems, the phase of clock TXPCO(Q), with respect to
the internal clock, can be shifted in 90° steps. Bit
TXPCOINV (180°) of register TXMFOUTC (F2h) together
with bit TXPOPHASE (90°) of register MUXCON0 (F1h)
sets the phase shift (see Table 11).
If no prescaler information is desired, the output can be
disabled by bit TXPRSCLEN of the same register. Apart
from these settings, the signal amplitude can be set. This
parameter follows the settings of the LVDS outputs. For
programming details, see Section “LVDS outputs”.
Loss of lock
During operating, the loss of lock output pin LOL should be
LOW which means that the clock synthesizer is in-lock and
the output frequency corresponds to the programmed
value. If pin LOL goes HIGH, phase and/or frequency lock
is lost and the output frequency may deviate from the
programmed value. The LOL condition is also available in
the registers INTERRUPT (00h) and STATUS (01h).
Table 11 Truth table for bits TXPCINV and TXPOPHASE
On demand (interrupt is default masked), it generates an
interrupt signal at pin INT.
MULTIPLEXER
The multiplexer comprises a high-speed input register, a
4-stage First In First Out (FIFO) elastic buffer, a parity
check circuit and the actual multiplexing tree.
2003 Dec 16
APPLICATION
20
TXPCOINV
TXPOPHASE
PHASE SHIFT
0
0
0°
0
1
90°
1
0
180°
1
1
270°
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
FRAMER
handbook, full pagewidth
TZA3015HW
TXPAR
TX_PARITY
TXPARQ
4
TX_DATA
4
TXPD0 to TXPD3
TXPD0Q to TXPD3Q
TXPC
TX_CLK
TXPCQ
TXPCO
TX_CLK_SRC
TXPCOQ
FIFORESET CREF
MGU684
system
clock
Fig.15 Co-directional clocking diagram.
FRAMER
handbook, full pagewidth
TZA3015HW
TXPAR
TX_PARITY
TXPARQ
4
TX_DATA
4
TXPD0 to TXPD3
TXPD0Q to TXPD3Q
TXPCO
TX_CLK_SRC
TXPCOQ
FIFORESET CREF
MGU685
system
clock
Fig.16 Contra-directional clocking diagram.
2003 Dec 16
21
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
Double data rate mode
Table 15 Truth table for bit TXPCODDREN
Usually the parallel clock frequency (TXPC, RXPC and
TXPCO) equals the parallel data rate (for example when
the serial bit rate is 2.488 Gbit/s, the parallel bit rate is
622 Mbit/s and the data is clocked with a 622 MHz clock).
This is the default operating mode.
TXPCODDREN
In I2C-bus mode the three parallel clocks can be set
separately in the DDR mode by bits RXPCDDREN,
TXPCDDREN and TXPCODDREN of registers
DDR&RXPRSCL (D5h), MUXCON0 (F1h) and
TXMFOUTC (F2h) respectively (see Tables 13, 14 and
15).
Table 12 Truth table for pin ENDDR
TXPC, RXPC and TXPCO in normal
mode
HIGH
TXPC, RXPC and TXPCO in DDR
mode
TXPCO in normal mode
The asynchronous signal FIFORESET is re-timed by the
internal clock from the clock generator. Two clock cycles
after signal FIFORESET has been made HIGH, the FIFO
initializes. Two clock cycles after signal FIFORESET has
been made LOW, the FIFO will be operational again.
To initialize automatically, when an overflow has occurred,
it is possible to connect pin OVERFLOW to pin
FIFORESET directly or via a resistor.
Table 13 Truth table for bit RXPCDDREN
RXPCDDREN
0
The overflow alarm persists until the FIFO is reset by a
HIGH-level on pin FIFORESET or by setting bit
FIFORESET of register MUXCON0 (F1h) to logic 1.
A FIFORESET also initializes the FIFO. I2C-bus control of
the FIFORESET function is obtained by programming bit
I2CFIFORES of register MUXCON0 (F1h). To fully benefit
from the FIFO, it should be reset whenever there has been
a LOL condition, or when bit rates have changed.
The DDR mode is functional for the whole bit-rate range,
so it is true A-rate.
LOW
TXPCO in DDR mode
In the co-directional clocking scheme, the input register
samples the parallel bus data on the rising edge of the
clock signal TXPC(Q). The same clock writes this data into
the FIFO register. Data is retrieved from the FIFO by an
internal clock, derived from the clock generator of the
actual multiplexing tree. This provides for large jitter
tolerance on the parallel interface; the FIFO absorbs
momentary phase disturbances. Excessively large phase
disturbances may stretch the elastic buffer to its limits,
causing a FIFO overflow or underflow. Pin OVERFLOW
and the registers STATUS (01h) and INTERRUPT (00h)
indicate this situation. On demand (i.e to programmed in
the register INTMASK [A0h]) it generates an interrupt
signal at pin INT.
The DDR functionality can be enabled by pin ENDDR (see
Table 12) or via the I2C-bus. I2C-bus control is enabled by
setting bit I2CDDR of register DDR&RXPRSCL (D5h).
MODE
1
FIFO register
However, in some applications it is required to use a
parallel clock operating at a frequency that is half of the
parallel data rate. This is the DDR mode (for example
when the serial bit rate is 2.488 Gbit/s, the parallel bit rate
is 622 Mbit/s and the data is clocked at both the rising as
well as the falling edge of the 311 MHz clock). The timing
for the parallel input interface is in accordance with the
SFI4 specification.
ENDDR
MODE
MODE
1
RXPC in DDR mode
Multiplexing bus swap
0
RXPC in normal mode
Bit TXBUSSWAP of register MUXCON1 (F0h) swaps the
bus order of the parallel data input bus TXPD0(Q) to
TXPD3(Q). Bit TXBUSSWAP reverses the order of bits
from MSB to LSB, or vice versa, to allow for optimum
connectivity on the PCB.
Table 14 Truth table for bit TXPCDDREN
TXPCDDREN
MODE
1
TXPC in DDR mode
0
TXPC in normal mode
2003 Dec 16
22
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
Parity checking
independently. This saves power when only one half of the
functionality is needed. The TZA3015HW can also be
configured as a clean-up PLL. This is described in the
Section “Loop modes”. The operating modes can be
selected with pins ENRX and ENTX, these pins enable the
receiver and the transmitter. This also offers the possibility
to power-down the complete IC. Operating (or enable)
modes are listed in Table 17.
In order to check the integrity of the data provided on the
parallel input bus, a parity checking function has been
implemented in the TZA3015HW. The calculated parity,
based on the data currently on the bus, is compared to the
expected parity provided at pins TXPAR(Q). If these do not
match, i.e. a parity error has occurred, the output pins
TXPARERR(Q) are HIGH during the next parallel bus
clock (TXPC) period.
Table 17 Truth table for the operating modes
Odd or even parity checking can be selected by pin
PAREVEN or by bit TXPAREVEN of register MUXCON1
(F0h). I2C-bus control of the parity type is enabled by
setting bit I2CTXPAREVEN of register MUXCON1 (F0h).
A HIGH-level on pin PAREVEN corresponds with even
parity (default for bit TXPAREVEN), see Table 16.
ENRX
ENTX
LOW
LOW
OPERATING MODE
power-down
LOW
HIGH
transmitter
HIGH
LOW
receiver
HIGH
HIGH
transceiver (or transponder)
Table 16 Truth table for parity setting
LOOP MODES
PIN PAREVEN
BIT TXPAREVEN
PARITY TYPE
LOW
0
odd
HIGH
1
even
The TZA3015HW supports four loop modes:
• Line loop back
• Diagnostic loop back
• Serial loop timing
Jitter performance
• Clean-up loop back.
The clock synthesizer has been optimized for lowest jitter
generation and the data and clock recovery has been
optimized for the best jitter tolerance. For all SDH/SONET
line rates, the jitter tolerance and the jitter generation is
compliant with ITU-T standard G.958, provided the
reference clock is clean enough. For optimum jitter
generation, the single-sideband phase noise of the
reference frequency should be less than −140 dBc/Hz, for
frequencies greater than 12 kHz from the carrier. If the
reference divider R is used, this requirement elevates with
approximately 20 × log R.
Selecting the loop modes
The required loop mode can be selected either by
pins LM0, LM1 and LM2 or by I2C-bus control.
The pin settings for the loop mode selection can be seen
in Table 18.
Table 18 Loop mode selection; note 1
Configuring the main functionality
OPERATING MODES
The TZA3015HW can be configured in several operating
modes. It can be configured as:
• Transceiver
• Transmitter
LM2
LM1
LM0
MODE
LOW
LOW
LOW
normal
LOW
LOW
HIGH
line loop back
LOW
HIGH
LOW
diagnostic loop back
HIGH
LOW
HIGH
serial loop timing
HIGH
HIGH
LOW
clean-up loop back
HIGH
HIGH
HIGH
normal
• Receiver
Note
• Transponder with clean-up PLL.
1. The loop mode can be also programmed by setting
bits LM[2:0] in register LOOPMODE (A3h).
The transceiver configuration is the default operating
mode. The transmitter and receiver part can be enabled
2003 Dec 16
23
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
Line loop back mode
transmit clock is generated. The parallel output clock
signal is recovered from the serial output data. This loop
mode is used to test the connection between the
transceiver and the data processing unit and the system
itself. No external fibre optic connection is needed to test
the system (see Fig.18).
This mode feeds back the received serial data to the serial
data output together with the recovered serial clock. This
allows testing of the serial data path including the optic
fibres. The received serial data that is fed back is also
available in parallel format at the parallel output bus (see
Fig.17).
Serial loop timing mode
Diagnostic loop back mode
This mode feeds back the recovered clock to the clock
synthesizer in order to run the receiver and transmitter at
the same clock frequency (see Fig.19).
This mode feeds back the parallel input data to the parallel
outputs together with a parallel clock. The parallel data is
serialized and available at the serial output. Also a serial
handbook, full pagewidth
data
serial
data
clock
serial
clock
4
4
4
MULTIPLEXER
parallel
data
parallel
clock
SYNTHESIZER
DCR
serial
data
data
4
parallel
data
DEMULTIPLEXER
clock
LIMITER
4
parallel
clock
MCE416
Fig.17 Line loop back mode.
handbook, full pagewidth
data
serial
data
clock
serial
clock
4
4
4
MULTIPLEXER
parallel
data
parallel
clock
SYNTHESIZER
DCR
serial
data
LIMITER
data
4
4
parallel
data
DEMULTIPLEXER
clock
parallel
clock
MCE417
Fig.18 Diagnostic loop back mode.
2003 Dec 16
24
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
data
handbook, full pagewidth
serial
data
clock
serial
clock
4
4
4
MULTIPLEXER
parallel
data
parallel
clock
SYNTHESIZER
DCR
serial
data
LIMITER
data
4
4
parallel
data
DEMULTIPLEXER
clock
parallel
clock
MCE418
Fig.19 Serial loop timing mode.
Clean-up loop back mode
components. The PLL consists of a phase frequency
detector, a charge pump, an external loop filter (R, C1 and
C2), a VCXO and a reference divider. The combination of
R and C1 is mandatory and will transform the current at the
output of the charge pump into a control voltage for the
VXCO. Capacitor C2 is optional.
The TZA3015HW can be used in transponder
applications. In this application, the transmitter is locked
onto the recovered clock from the DCR (RXPRSCL).
Without preparations, the jitter transfer of this application is
determined by cascading the transfer functions of the DCR
and the clock synthesizer. This transfer function is not well
controlled and may not meet the required specification in
terms of bandwidth and/or jitter peaking. A second
drawback is that the jitter generation of the synthesizer is
degraded because the frequency reference (i.e. the DCR)
is not very clean in terms of phase-noise.
The internal clock and data path in the TZA3015HW is
clarified in Fig.21. As can be seen in the clean-up
application, the received (and transmitted) data is also
available in parallel format at the parallel output bus.
Two bits are available to ease the design of the clean-up
PLL. The loop is designed to work with a VCXO that has a
positive gain. That is an increasing voltage on the VCXO
control input will increase the output frequency. By means
of bit CLUPPLLINV of register REFDIV (A1h) the loop is
inverted and will work with VCXOs which have a negative
gain. Bit CLUPPLLHG of register REFDIV (A1h) will
change the gain of the charge pump. If bit CLUPPLLHG is
logic 0, the charge pump current ICP is 100 µA. If bit
CLUPPLLHG is logic 1, the charge pump current ICP is
1 mA. This eases choosing suitable component values for
R and C1.
To improve both the jitter transfer and jitter generation in
transponder applications, an external low-noise reference
oscillator is locked onto the DCR recovered clock by
means of a small band PLL, i.e. the clean-up PLL. The
low-noise oscillator, e.g. a Voltage Controlled Crystal
Oscillator (VCXO), acts as the reference for the clock
synthesizer. If appropriately designed, the jitter will be
dominated by the clean-up PLL. This PLL can be
optimized for bandwidth and jitter peaking, while the jitter
generation is optimized by choosing the appropriate
VCXO.
Figure 20 shows a typical clean-up PLL application. For
ease of use, all components are integrated in the
TZA3015HW, except for the VCXO and the loop filter
2003 Dec 16
25
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
external components
handbook, full pagewidth
from
DCR
PHASE
FREQUENCY
DETECTOR
I2C bit:
CLUPPLLINV
CHARGE
PUMP
CREF REFERENCE
DIVIDER
IPUMP
VCXO
e.g. Vectron
VDSGLA type
C1
C2
R
I2C bit:
CLUPPLLHG
to
synthesizer
MCE419
Fig.20 Clean-up PLL application with the TZA3015HW.
handbook, full pagewidth
data
serial
data
clock
serial
clock
4
4
4
MULTIPLEXER
parallel
data
parallel
clock
SYNTHESIZER
DCR
serial
data
LIMITER
data
4
4
parallel
data
DEMULTIPLEXER
clock
parallel
clock
MCE420
Fig.21 Clean-up loop back mode.
I/O configuration
• Transmitter parallel clock output; pins TXPCO(Q)
LVDS OUTPUTS
• Prescaler DCR output; pins RXPRSCL(Q)
• Prescaler synthesizer output; pins TXPRSCL(Q).
Several options exist that allow flexible configuration of the
LVDS outputs: output amplitude, signal polarity, bus order,
mute and selective enable/disable of various outputs. All
these options can be set in the registers MFOBCON (A4h),
DMXCON (B8h), RXMFOUTC0 (D4h), DDR&RXPRSCL
(D5h) and TXMFOUTC (F2h). Affected by these registers
are:
The output swing of all LVDS outputs can be set by pin
LOWSWING or by programming bit LOWSWING in
register MFOBCON (A4h). I2C-bus control is enabled by
programming bit I2CLOWSWING in register MFOBCON
(A4h). The typical voltage levels are given in Table 19. See
also Figs 34 and 35.
• Parallel clock output; pins RXPC(Q)
Table 19 Truth table for pin LOWSWING
• Parallel data output; pins RXPD0(Q) to RXPD3(Q)
• Frame pulse output; pins RXFP(Q)
LOWSWING
LVDS OUTPUT VOLTAGE SWING
• Parity output; pins RXPAR(Q)
LOW
500 mV
• Parity error output; pins TXPARERR(Q)
HIGH
300 mV
2003 Dec 16
26
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
Parallel clock output
Prescaler DCR output
Bit RXPCINV of register RXMFOUTC0 (D4h) sets the
polarity of the parallel clock output RXPC(Q), effectively
shifting the clock edge by half a clock cycle and changing
the rising edge to a falling edge. This might resolve a
parallel bus timing problem. The parallel clock output can
be disabled by programming bit RXPCEN of register
RXMFOUTC0 (D4h).
The polarity of the receiver prescaler output RXPRSCL(Q)
is set by bit RXPRSCLINV of register DDR&RXPRSCL
(D5h). The receiver prescaler output can be disabled by
programming bit RXPRSCLEN of register
DDR&RXPRSCL (D5h).
Prescaler synthesizer output
The polarity of the transmitter prescaler output
TXPRSCL(Q) is set by bit TXPRSCLINV of register
TXMFOUTC (F2h). The transmitter prescaler output can
be disabled by programming bit TXPRSCLEN of register
TXMFOUTC (F2h).
Parallel data output
The parallel output bus data RXPD0(Q) to RXPD3(Q) can
be swapped by bit RXBUSSWAP of register DMXCON
(B8h). The mute option forces the parallel output bits to a
logic 0 state. This is done by programming bit DMXMUTE
of register DMXCON (B8h). The polarity of the data
RXPD0(Q) to RXPD3(Q) can be set by bit RXPDINV of
register RXMFOUTC0 (D4h). The data outputs can be
disabled by programming bit RXPDEN of register
RXMFOUTC0 (D4h).
LVDS INPUTS
The available LVDS inputs are:
• Parallel clock input; pins TXPC(Q)
• Parallel data input; pins TXPD0(Q) to TXPD3(Q)
• Parity input; pins TXPAR(Q).
Frame pulse output
The differential LVDS inputs can handle any input swing
with a minimum of 100 mV (p-p) single-ended. The inputs
accept any value between VEE and VCC, i.e. the input
buffers are true rail-to-rail. The limiting value of the LVDS
input current is 25 mA. A differential hysteresis of 25 mV is
implemented; see Fig.33.
The polarity of the frame pulse output RXFP(Q) is set by
bit RXFPINV of register RXMFOUTC0 (D4h). The frame
pulse output can be disabled by programming bit RXFPEN
of register RXMFOUTC0 (D4h).
Parity output
Parallel clock input
The polarity of the parity output RXPAR(Q) is set by bit
RXPARINV of register RXMFOUTC0 (D4h). The parity
output can be disabled by programming bit RXPAREN of
register RXMFOUTC0 (D4h).
Bit TXPCINV of register MUXCON1 (F0h) sets the polarity
of the parallel clock input TXPC(Q), effectively shifting the
clock edge by half a clock cycle and changing the rising
edge to a falling edge. This could be used to resolve a
parallel bus timing problem.
Parity error output
The polarity of the parity error output TXPARERR(Q) is set
by bit TXPARERRINV of register TXMFOUTC (F2h). The
parity error output can be disabled by programming bit
TXPARERREN of register TXMFOUTC (F2h).
Parallel data input
The order of the parallel output bus data TXPD0(Q) to
TXPD3(Q) can be programmed by bit TXBUSSWAP of
register MUXCON1 (F0h).
Transmitter parallel clock output
Bit TXPDINV of register MUXCON1 (F0h) sets the polarity
of the parallel data inputs TXPD0(Q) to TXPD3(Q).
Bit TXPCOINV of register TXMFOUTC (F2h) sets the
polarity of the parallel clock output TXPCO(Q), effectively
shifting the clock edge by half a clock cycle and changing
the rising edge to a falling edge. The phase of the clock
can be shifted by 90° by programming bit TXPCOPHASE
of register MUXCON0 (F1h). The combination of these two
bits offers a phase shift range of 0 to 360°, adjustable in
four steps (step size 90°). This might resolve a parallel bus
timing problem. The parallel clock output can be disabled
by programming bit TXPCOEN of register TXMFOUTC
(F2h).
2003 Dec 16
RF OUTPUTS
The serial RF outputs are CML type outputs (see Figs 31
and 32). Several options exist that allow flexible
configuration of the RF outputs: output amplitude
adjustment, signal polarity, data-clock swap, output
termination and selective enable/disable of the clock
output. Thus, the TZA3015HW can be configured so that
27
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
connectivity problems with other ICs are avoided. Unused
outputs can be disabled.
REFERENCE CLOCK INPUT
The reference clock CREF(Q) input is shown in Fig.36
These options can be programmed in registers
TXRFOUTC1 (F3h) and TXRFOUTC0 (F4h). The
following RF outputs are available:
RF INPUT
The serial data inputs are pins RXSD(Q). These pins are
differential CML type serial RF data inputs. There are no
special settings for these inputs.
• Serial data output; pins TXSD(Q)
• Serial clock output; pins TXSC(Q).
The RF CML data and clock outputs have an adjustable
signal amplitude between 70 and 1100 mV (p-p)
single-ended in 16 steps. The amplitude can be
programmed by setting bits RFS[3:0] of register
TXRFOUTC0 (F4h). The default amplitude is
300 mV (p-p) single-ended.
CMOS OUTPUTS
The CMOS outputs are all used as logic outputs to indicate
the status of the TZA3015HW.
• Loss of signal output; pin LOS
• Frequency window detector output; pin INWINDOW
The clock and data outputs can be swapped by
programming bit TXSDSCSWAP of register TXRFOUTC1
(F3h). Allowing full flexibility in the PCB design.
• Interrupt output; pin INT
• Loss of lock output; pin LOL
• FIFO overflow alarm output; pin OVERFLOW.
The data and clock outputs can be DC- or AC-coupled to
the laser driver. The TZA3015HW serial RF outputs can be
adapted to this for optimal connectivity by appropriately
setting bit RFOUTTERMAC of register TXRFOUTC0
(F4h). DC termination is default.
A LOW state equals the ground potential and a HIGH state
equals the supply voltage. The INT output can be
configured as CMOS output or as open-drain output (see
Sections “Open-drain output” and “Interrupt generation”).
The output is configured as open-drain output by default.
Serial clock output
CMOS INPUTS
The polarity of the serial clock output TXSC(Q) can be
programmed by bit TXSCINV of register TXRFOUTC1
(F3h). The serial clock output can be disabled by setting
pin ENTXSC or by programming bit TXSCEN of register
TXRFOUTC1 (F3h) (see Table 20). This saves power
dissipation in applications where the serial clock is not
needed
The CMOS inputs are all used as logic inputs to configure
the TZA3015HW:
• User interface selection input; pin UI
• Data rate selection inputs; pins DR0 to DR2
• Loop mode selection inputs; pins LM0 to LM2
• Enable receiver input; pin ENRX
Table 20 Truth table for serial clock enable
• Enable transmitter input; pin ENTX
PIN ENTXSC
BIT ENTXSC
LOW
0
disabled
• Wide and narrow frequency detect window selection
input; pin WINSIZE
HIGH
1
enabled
• Enable low LVDS swing output input; pin LOWSWING
SERIAL CLOCK
• Reference frequency selection inputs; pins FREF0 and
FREF1
In order to control the enabling of the serial clock output by
the I2C-bus, bit I2CTXSCEN of register TXRFOUTC1
(F3h) must be programmed.
• Enable byte alignment input; pin ENBA
• FIFO reset input; pin FIFORESET
• Odd or even parity check input; pin PAREVEN
Serial data output
• Co-directional or contra-directional clocking selection
input; pin CLKDIR
The polarity of the serial data output TXSD(Q) can be
programmed by bit TXSDINV of register TXRFOUTC1
(F3h). The data output can be disabled by programming bit
TXSDEN of register TXRFOUTC1 (F3h).
2003 Dec 16
• Enable serial clock input; pin ENTXSC.
28
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
interrupt bits where the ‘alarm’ is no longer present. If the
‘alarm’ is still set, the interrupt bit is not cleared after the
read action. If an interrupt bit remains set (and if it is not
masked) the INT pin will keep its interrupt condition active;
it will not generate a pulse nor a spike. The I2C-bus status
register is not reset since it always shows the present
status of the receiver. It is important to note that the three
reserved bits of the STATUS and INTERRUPT registers
can take any value and that they can change during
operating. These bits can not be used to obtain information
on the status of the IC.
The CMOS inputs have an internal pull-up resistance; if
the input is left open, a logic HIGH state will be forced
internally. In the pre-programmed mode (UI = LOW), pins
DR0 to 2 act as regular CMOS inputs. In the I2C-bus mode
(UI = HIGH), pins SCL and SDA comply with the I2C-bus
interface standard.
OPEN-DRAIN OUTPUT
The TZA3015HW contains one open-drain interrupt output
pin INT. The output type of the interrupt controller can be
configured by programming bit INTOUT of register
INTCONF (A5h). The output can be configured as a
push-pull CMOS output or as an open-drain output. For the
open-drain configuration an external pull-up resistor of
3.3 kΩ is recommended. The polarity can be set by
programming bit INTPOL of register INTCONF (A5h).
Power supply connections
Four separate supply domains (VDD, VCCD, VCCO and
VCCA) provide isolation between the various functional
blocks. Each supply domain should be connected to a
common VCC via separate filters. All supply domains
should be powered synchronously.
INTERRUPT GENERATION
The TZA3015HW features a fully configurable interrupt
generator. An interrupt signal can be generated in the
following events:
All supply pins, including the exposed die pad, must be
connected. The die pad should be connected with the
lowest inductance possible. Since the die pad is also used
as the main ground return of the chip, the connection
should have a low DC impedance as well. The voltage
supply levels should be in accordance with the values
specified in Chapter “Characteristics”.
• Loss Of Signal (LOS)
• INWINDOW
• Temperature alarm
• Loss Of Lock (LOL)
All external components should be surface mounted
devices, preferably of size 0603 or smaller. The
components must be mounted as closely to the IC as
possible.
• FIFO overflow or underflow.
The aforementioned events generate flags which can be
read in register STATUS (01h). Each of these flags will
generate an interrupt in the INTERRUPT register (00h).
If programmed so in the register INTMASK (A0h) the
INTERRUPT register bit(s) will generate an interrupt on
pin INT. In this mask register each interrupt bit can be
masked by writing a logic 0 in the corresponding bit
position.
I2C-BUS
I2C-bus characteristics
The I2C-bus is a 2-line communication between different
ICs or modules. The two lines are a serial data line (SDA)
and a serial clock line (SCL). Data transfer may be initiated
only when the line is not busy.
The STATUS register shows the present status of the
receiver. The INTERRUPT register shows the history of
the interrupts and is not affected by the INTMASK register.
START AND STOP CONDITIONS
Bit INTOUT of register INTCONF (A5h) determines the
output type of pin INT: standard CMOS output or
open-drain output. The latter is the default which provides
for multiple receivers sharing a common interrupt signal
wire with a 3.3 kΩ pull-up resistor (INT is active LOW in
this case). The polarity can be set by programming bit
INTPOL of register INTCONF (A5h).
Figure 22 shows the definition of the start and stop
conditions. Both data and clock lines remain HIGH when
the bus is not busy. A HIGH-to-LOW transition of the data
line, while the clock is HIGH is defined as the start
condition (S). A LOW-to-HIGH transition of the data line
while the clock is HIGH is defined as the stop condition (P).
The interrupt and status register can be polled by an
I2C-bus read action. After the read action on the interrupt
register the interrupt register is reset by clearing the
2003 Dec 16
29
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
ACKNOWLEDGE
The device that acknowledges must pull down the SDA
line during the acknowledge clock pulse, so that the SDA
line is stable LOW during the HIGH period of the
acknowledge related clock pulse (set-up and hold times
must be taken into consideration). A master receiver must
signal an end of data to the transmitter by not generating
an acknowledge on the last byte that has been clocked
out of the slave. In this event the transmitter must leave the
data line HIGH to enable the master to generate a stop
condition.
Figure 23 shows the definition of an acknowledgement on
the I2C-bus. Only one data byte is transferred between the
start and stop conditions during a write from the transmitter
to the receiver. Each byte of eight bits is followed by an
acknowledge bit. The acknowledge bit is a HIGH level
signal put on the bus by the transmitter during which time
the master generates an extra acknowledge related clock
pulse. A slave receiver which is addressed must generate
an acknowledge after the reception of each byte. Also a
master receiver must generate an acknowledge after the
reception of each byte that has been clocked out of the
slave transmitter.
handbook, full pagewidth
SDA
SDA
SCL
SCL
S
P
START condition
STOP condition
MBC622
Fig.22 Start and stop conditions.
handbook, full pagewidth
DATA OUTPUT
BY TRANSMITTER
not acknowledge
DATA OUTPUT
BY RECEIVER
acknowledge
SCL FROM
MASTER
1
2
8
9
S
clock pulse for
acknowledgement
START
condition
MBC602
Fig.23 (Not) acknowledge condition on the I2C-bus.
2003 Dec 16
30
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
I2C-bus protocol
READ PROTOCOL
Figure 24 shows the definition of the bytes. If bit R/W = 1
the master reads from the read register, if bit R/W = 0 the
master writes to the write register. It is not possible to write
and read the same register.
Figure 26 shows the protocol for reading from one or more
registers.
After the start command (S) the receiver sends the
address of the slave device, waits for an acknowledge
from the transmitter slave, receives data from the slave
(slave, TZA3015HW, starts sending data after generating
the acknowledge), after receiving the data, the receiver
(master) sends an acknowledge, or if finished a
not-acknowledge followed by a stop condition (P).
WRITE PROTOCOL
Figure 25 shows the protocol for writing to one single
register. After the start command (S) the transmitter sends
the address of the slave device, waits for an acknowledge
from the slave, sends the register address, waits for an
acknowledge, sends data, waits for an acknowledge from
the master followed by a stop condition (P).
handbook, full pagewidth
MSB
LSB
MSB
R/W
1
LSB
Slave address
Register address
MCE425
Fig.24 Definition of slave- and register address (= instruction byte); slave and register addresses are 7 bits.
acknowledge
from slave
handbook, full pagewidth
acknowledge
from slave
R/W MSB
S
SLAVE
ADDRESS
0 A 1
acknowledge
from master
MSB
REGISTER
ADDRESS
A
LSB
DATA
A
P
one byte transferred
MDB071
Fig.25 Write protocol.
handbook, halfpage
R/W
S
acknowledge
from master (1)
acknowledge
from slave
SLAVE
ADDRESS
MSB
1 A
LSB
DATA
A/A
P
n bytes
MDB072
(1) The master receiver must signal an end of data to the slave transmitter by not generating an acknowledge on the last byte that has been clocked
out of the slave.
Fig.26 Read protocol.
2003 Dec 16
31
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
I2C-BUS REGISTERS
The TZA3015HW can be programmed via the I2C-bus if pin UI = HIGH or leaving the pin open-circuit. The I2C-bus
registers can be accessed via the 2-wire I2C-bus interface using pins SCL and SDA if pin CS = HIGH during read or write
actions. The I2C-bus address of the TZA3015HW can be found in Table 2.
Table 21 I2C-bus registers
ADDRESS
(HEX)
NAME
FUNCTION
DEFAULT
RANGE
R/W
General part
00
INTERRUPT
interrupt register (see Table 22)
XXXX XXXX
n.a.
R
01
STATUS
status register (see Table 23)
XXXX XXXX
n.a.
R
A0
INTMASK
interrupt mask register (see Table 24)
0000 0100
n.a.
W
A1
REFDIV
reference divider and clean-up PLL
(see Table 25)
0000 0000
n.a.
W
A3
LOOPMODE
loop mode and enable register (see Table 26)
0110 0111
n.a.
W
A4
MFOBCON
LVDS output buffer configuration (see Table 27)
0101 0000
n.a.
W
A5
INTCONF
interrupt output configuration (see Table 28)
0000 0001
n.a.
W
B0
HEADER3
programmable header; MSB (see Table 29)
1111 0110
n.a.
W
B1
HEADER2
programmable header (see Table 30)
1111 0110
n.a.
W
B2
HEADER1
programmable header (see Table 31)
0010 1000
n.a.
W
B3
HEADER0
programmable header; LSB (see Table 32)
0010 1000
n.a.
W
B4
HEADERX3
programmable header don’t care; MSB
(see Table 33)
0000 0000
n.a.
W
B5
HEADERX2
programmable header don’t care (see Table 34)
0000 0000
n.a.
W
B6
HEADERX1
programmable header don’t care (see Table 35)
0000 0000
n.a.
W
B7
HEADERX0
programmable header don’t care; LSB
(see Table 36)
0000 0000
n.a.
W
B8
DMXCON
demultiplexer configuration register
(see Table 37)
0000 0000
n.a.
W
C0
RXOCTDIV
DCR octave M divider (see Table 38)
0000 0000
n.a.
W
C1
RXMAINDIV1
VCO frequency N divider (see Table 39)
0000 0001
128 to 511
W
C2
RXMAINDIV0
VCO frequency N divider (see Table 40)
0000 0000
128 to 511
W
C3
RXFRACN2
fractional division (see Table 41)
1000 0000
n.a.
W
C4
RXFRACN1
fractional division (see Table 42)
0000 0000
n.a.
W
C5
RXFRACN0
fractional division (see Table 43)
0000 0000
n.a.
W
C6
DCRCON
DCR configuration register (see Table 44)
0000 1100
n.a.
W
D0
LIMLOSTH
limiter loss threshold
0000 0000
0 to 255
W
D1
LIMLOSCON
limiter loss of signal configuration register
(see Table 45)
0000 1101
n.a.
W
D2
LIMSL
limiter slice level
0000 0000
0 to 255
W
D3
LIMCON
limiter amplifier configuration (see Table 46)
0000 0000
n.a.
W
D4
RXMFOUTC0
disable/invert parallel outputs (see Table 47)
1010 1010
n.a.
W
Transceiver
2003 Dec 16
32
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
ADDRESS
(HEX)
D5
NAME
TZA3015HW
FUNCTION
DEFAULT
DDR&RXPRSCL disable/invert parallel outputs (see Table 48)
RANGE
R/W
0010 0000
n.a.
W
Transmitter part
E0
TXOCTDIV
synthesizer octave divider (see Table 49)
0000 0000
n.a.
W
E1
TXMAINDIV1
VCO frequency (N divider) (see Table 50)
0000 0001
128 to 255
W
E2
TXMAINDIV 0
VCO frequency (N divider) (see Table 51)
0000 0000
128 to 255
W
E3
TXFRACN2
fractional division (see Table 52)
1000 0000
n.a.
W
E4
TXFRACN1
fractional division (see Table 53)
0000 0000
n.a.
W
E5
TXFRACN0
fractional division (see Table 54)
0000 0000
n.a.
W
F0
MUXCON1
multiplexer configuration byte 1 (see Table 55)
0110 0010
n.a.
W
F1
MUXCON0
multiplexer configuration byte 0 (see Table 56)
0000 0010
n.a.
W
F2
TXMFOUTC
disable/invert LVDS outputs (see Table 57)
1010 1000
n.a.
W
F3
TXRFOUTC1
disable/invert RF outputs (see Table 58)
0100 1011
n.a.
W
F4
TXRFOUTC0
RF output configuration register (see Table 59)
0000 0011
n.a.
W
Table 22 Register INTERRUPT (address: 00h)
BIT
7
6
5
4
PARAMETER
3
2
1
0
DESCRIPTION
loss of signal
1
no signal present
0
signal present
INWINDOW
1
frequency out of window
0
frequency in window
temperature alarm
1
junction temperature ≥130 °C
0
junction temperature <130 °C
loss of lock
x
x
LOS
INWINDOW
TALARM
LOL
1
synthesizer out of lock
0
synthesizer out of lock
x
reserved
FIFO overflow or underflow
1
FIFO overflow or underflow occurred
0
FIFO normal operating
2003 Dec 16
NAME
33
OVERFLOW
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
Table 23 Register STATUS (address: 01h)
BIT
7
6
5
4
PARAMETER
3
2
1
0
DESCRIPTION
loss of signal
LOS
1
no signal present
0
signal present
INWINDOW
INWINDOW
1
frequency out of window
0
frequency in window
temperature alarm
TALARM
1
junction temperature ≥130 °C
0
junction temperature <130 °C
loss of lock
x
x
NAME
LOL
1
synthesizer out of lock
0
synthesizer out of lock
x
reserved
FIFO over- or underflow
OVERFLOW
1
FIFO under- or underflow occurred
0
FIFO normal operating
Table 24 Register INTMASK (address: A0h, default value: 04h)
BIT
7
6
5
4
PARAMETER
3
2
1
0
DESCRIPTION
mask LOS signal
1
not masked
0
masked; note 1
mask INWINDOW signal
1
not masked
0
masked; note 1
mask temperature alarm
1
not masked
0
masked; note 1
mask LOL signal
x
x
1
not masked
0
masked; note 1
x
1
not masked
0
masked; note 1
0
2003 Dec 16
0
0
MLOS
MINWINDOW
MTALARM
MLOL
reserved
mask FIFO overflow or underflow
0
NAME
0
1
0
0
MOVERFLOW
default value
34
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
Note to Table 24
1. Signal is not processed by the interrupt controller.
Table 25 Register REFDIV (address: A1h, default value: 00h)
BIT
7
6
5
4
PARAMETER
3
2
1
0
DESCRIPTION
reference frequency division ratio divider R; octave
selection
0
0
0
R=1
0
0
1
R=2
0
1
0
R=4
0
1
1
R=8
1
0
0
R = 16
1
0
1
NAME
FREFI2C[2:0]
R = 32
reference frequency division programming by I2C-bus I2CFREF
1
enable I2C-bus programming
0
enable programming by pins
x
reserved
high gain clean-up PLL
1
enable high gain
0
normal gain
CLUPPLLHG
invert charge pump currents of the clean-up PLL
1
clean-up PLL inverted
0
clean-up PLL normal operating
enable clean-up PLL
CLUPPLLEN
1
clean-up PLL enabled
0
clean-up PLL disabled (except in clean-up loop
back mode)
0
0
0
0
0
0
0
CLUPPLLINV
0
default value
Table 26 Register LOOPMODE (address: A3h, default value: 67h)
BIT
7
6
5
4
PARAMETER
3
2
1
0
DESCRIPTION
0
0
0
normal mode
0
0
1
line loop back mode
0
1
0
diagnostic loop back mode
0
1
1
reserved
1
0
0
reserved
1
0
1
serial loop timing mode
1
1
0
clean-up loop back mode
1
1
1
normal mode
loop mode selection
2003 Dec 16
35
NAME
LM[2:0]
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
BIT
7
6
5
4
PARAMETER
3
2
1
0
DESCRIPTION
loop mode programming by I2C-bus
1
enable I2C-bus programming
0
enable programming by pins
x
I2CLM
reserved
enable receiver
1
ENRX
receiver enabled
0
receiver disabled
enable transmitter
1
ENTX
transmitter enabled
0
transmitter disabled
transmitter/receiver enable by I2C-bus
1
enable
0
0
NAME
I2C-bus
I2CENTRX
programming
enable programming by pins
1
1
0
0
1
1
1
default value
Table 27 Register MFOBCON (address: A4h, default value: 50h)
BIT
7
6
PARAMETER
5
4
3
2
1
0
x
x
x
x
x
x
DESCRIPTION
NAME
reserved
parallel output voltage swing
1
low swing (300 mV)
0
high swing (500 mV)
LOWSWING
parallel output voltage swing programming by I2C-bus I2CLOWSWING
1
enable I2C-bus programming
0
enable programming by pins
0
1
0
1
0
0
0
0
default value
Table 28 Register INTCONF (address: A5h, default value: 01h)
BIT
7
6
5
4
PARAMETER
3
2
1
0
DESCRIPTION
interrupt output polarity
1
inverted
0
normal operating
interrupt output configuration
x
x
x
x
x
x
0
0
0
0
0
0
2003 Dec 16
1
push-pull output
0
open drain output
NAME
INTPOL
INTOUT
reserved
0
1
default value
36
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
Table 29 Register HEADER3 (address: B0h, default value: F6h)
BIT
7
6
5
4
PARAMETER
3
2
1
0
x
x
x
x
x
x
x
x
1
1
1
1
0
1
1
0
DESCRIPTION
programmable header; H31 = MSB
NAME
H[31:24]
default value
Table 30 Register HEADER2 (address: B1h, default value: F6h)
BIT
PARAMETER
7
6
5
4
3
2
1
0
x
x
x
x
x
x
x
x
1
1
1
1
0
1
1
0
DESCRIPTION
programmable header
NAME
H[23:16]
default value
Table 31 Register HEADER1 (address: B2h, default value: 28h)
BIT
PARAMETER
7
6
5
4
3
2
1
0
x
x
x
x
x
x
x
x
0
0
1
0
1
0
0
0
DESCRIPTION
programmable header
NAME
H[15:08]
default value
Table 32 Register HEADER0 (address: B3h, default value: 28h)
BIT
PARAMETER
7
6
5
4
3
2
1
0
x
x
x
x
x
x
x
x
0
0
1
0
1
0
0
0
DESCRIPTION
programmable header; H00 = LSB
NAME
H[07:00]
default value
Table 33 Register HEADERX3 (address: B4h, default value: 00h)
BIT
PARAMETER
7
6
5
4
3
2
1
0
x
x
x
x
x
x
x
x
0
0
0
0
0
0
0
0
DESCRIPTION
don’t care; HX31 = MSB
NAME
HX[31:24]
default value
Table 34 Register HEADERX2 (address: B5h, default value: 00h)
BIT
PARAMETER
7
6
5
4
3
2
1
0
x
x
x
x
x
x
x
x
0
0
0
0
0
0
0
0
2003 Dec 16
DESCRIPTION
don’t care
NAME
HX[23:16]
default value
37
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
Table 35 Register HEADERX1 (address: B6h, default value: 00h)
BIT
7
6
5
4
PARAMETER
3
2
1
0
x
x
x
x
x
x
x
x
0
0
0
0
0
0
0
0
DESCRIPTION
don’t care
NAME
HX[15:08]
default value
Table 36 Register HEADERX0 (address: B7h, default value: 00h)
BIT
PARAMETER
7
6
5
4
3
2
1
0
x
x
x
x
x
x
x
x
0
0
0
0
0
0
0
0
DESCRIPTION
don’t care; HX00 = LSB
NAME
HX[07:00]
default value
Table 37 Register DMXCON (address: B8h, default value: 00h)
BIT
7
6
5
4
PARAMETER
3
2
1
0
x
x
x
x
DESCRIPTION
reserved
parallel bus swapping
1
RXPD0 = MSB; RXPD3 = LSB (swapped)
0
RXPD3 = MSB; RXPD0 = LSB (normal)
mute parallel outputs
1
enable mute; parallel outputs forced to logic 0
0
disable mute
enable byte alignment
1
byte alignment enabled
0
byte alignment disabled
ENBA programming by I2C-bus
I2C-bus
1
enable
0
enable programming by pins
0
0
2003 Dec 16
0
0
0
0
0
NAME
0
RXBUSSWAP
DMXMUTE
ENBA
I2CENBA
programming
default value
38
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
Table 38 Register RXOCTDIV (address: C0h, default value: 00h)
BIT
7
6
5
4
PARAMETER
3
2
1
0
DESCRIPTION
0
0
0
M = 1; octave number 0
0
0
1
M = 2; octave number 1
0
1
0
M = 4; octave number 2
0
1
1
M = 8; octave number 3
1
0
0
M = 16; octave number 4
1
0
1
M = 32; octave number 5
1
1
0
M = 64; octave number 6
division ratio octave divider M; octave selection
x
x
x
x
x
0
0
0
0
0
NAME
RXDIV_M[2:0]
reserved
0
0
0
default value
Table 39 Register RXMAINDIV1 (address: C1h, default value: 01h)
BIT
7
6
5
4
PARAMETER
3
2
1
0
x
x
x
x
x
x
x
x
0
0
0
0
0
0
0
DESCRIPTION
division ratio divider N; RXN8 = MSB
NAME
RXN8
reserved
1
default value
Table 40 Register RXMAINDIV0 (address: C2h, default value: 00h)
BIT
PARAMETER
7
6
5
4
3
2
1
0
x
x
x
x
x
x
x
x
0
0
0
0
0
0
0
0
DESCRIPTION
division ratio divider N; RXN0 = LSB
NAME
RXN[7:0]
default value
Table 41 Register RXFRACN2 (address: C3h, default value: 80h)
BIT
7
6
PARAMETER
5
4
3
2
1
0
x
x
x
x
x
x
DESCRIPTION
fractional divider; RXK21 = MSB
x
RXNILFRAC control bit (NF)
no fractional N functionality
0
fractional N functionality
0
2003 Dec 16
RXK[21:16]
reserved
1
1
NAME
0
0
0
0
0
0
RXNILFRAC
default value
39
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
Table 42 Register RXFRACN1 (address: C4h, default value: 00h)
BIT
7
6
5
4
PARAMETER
3
2
1
0
x
x
x
x
x
x
x
x
0
0
0
0
0
0
0
0
DESCRIPTION
fractional divider
NAME
RXK[15:8]
default value
Table 43 Register RXFRACN0 (address: C5h, default value: 00h)
BIT
PARAMETER
7
6
5
4
3
2
1
0
x
x
x
x
x
x
x
x
0
0
0
0
0
0
0
0
DESCRIPTION
fractional divider; RXK0 = LSB
NAME
RXK[7:0]
default value
Table 44 Register DCRCON (address: C6h, default value: 0Ch)
BIT
7
6
5
4
PARAMETER
3
2
1
0
DESCRIPTION
frequency window size; relative to bit rate
0
1
1
2000 ppm
1
0
0
1000 ppm
1
0
1
500 ppm
1
1
0
1
window size according to bits WINSIZE[2:0]
(default value 1000 ppm); PLL frequency loosely
coupled to reference crystal
0
window size is 0 ppm; PLL frequency directly
synthesized from reference crystal
WINSIZE control bit
through
0
I2C-bus
x
x
0
WINSIZE
I2CWINSIZE
interface
through external pin WINSIZE
automatic frequency window size selection
0
WINSIZE[2:0]
250 ppm
manual frequency window size selection
1
NAME
1
enabled
0
disabled
AUTOWIN
reserved
0
0
1
1
0
0
default value
Table 45 Register LIMLOSCON (address: D1h, default value: 0Dh)
BIT
7
6
5
4
PARAMETER
3
2
1
0
DESCRIPTION
enable loss of signal detection
2003 Dec 16
1
LOS detection enabled
0
LOS detection disabled
40
NAME
LOSEN
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
BIT
7
6
5
4
PARAMETER
3
2
1
0
DESCRIPTION
LOS threshold level programming by I2C-bus
1
enable I2C-bus programming; set level by
register D0h
0
set level by applying analog reference voltage on
pin LOSTH
loss of signal detection hysteresis
0
0
0
0 dB
0
0
1
1 dB
0
1
0
2 dB
0
1
1
3 dB
1
0
0
4 dB
1
0
1
5 dB
1
1
0
6 dB
1
1
1
7 dB
enable slice level
slice level enabled
0
slice level disabled
positive slice level
0
negative slice level
LOS level polarity
LOSPOL
1
inverted polarity
0
normal polarity
0
0
1
1
0
HTLCB[2:0]
SLSGN
1
0
I2CLOSTH
SLEN
1
slice level sign
0
NAME
1
default value
Table 46 Register LIMCON (address: D3h, default value: 00h)
BIT
7
6
5
4
PARAMETER
3
2
1
0
DESCRIPTION
0
0
0
octave number 0; 1800 to 3200 Mbit/s
0
0
1
octave number 1; 900 to 1800 Mbit/s
0
1
0
octave number 2; 450 to 900 Mbit/s
0
1
1
octave number 3; 225 to 450 Mbit/s
1
X
X
octave number 4; 30 to 225 Mbit/s
amplifier octave selection
x
x
x
x
x
0
0
0
0
0
2003 Dec 16
NAME
AMP[2:0]
reserved
0
0
0
default value
41
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
Table 47 Register RXMFOUTC0 (address: D4h, default value: AAh)
BIT
7
6
5
4
PARAMETER
3
2
1
0
DESCRIPTION
parallel data output polarity
1
inverted
0
normal
RXPDINV
parallel data output enable
1
enabled
0
disabled
RXPDEN
parallel clock output polarity
1
inverted
0
normal
RXPCINV
parallel clock output enable
1
enabled
0
disabled
RXPCEN
parity output polarity
1
inverted
0
normal
RXPARINV
parity output enable
1
enabled
0
disabled
RXPAREN
frame pulse output polarity
1
RXFPINV
inverted
0
normal
frame pulse output enable
1
RXFPEN
enabled
0
1
NAME
disabled
0
1
0
1
0
1
0
default value
Table 48 Register DDR&RXPRSCL (address: D5h, default value: 20h)
BIT
7
6
5
4
PARAMETER
3
2
1
0
x
x
x
x
DESCRIPTION
reserved
invert RX prescaler output
1
0
RXPRSCLINV
inverted
normal
enable RX prescaler output
2003 Dec 16
NAME
1
enabled
0
disabled
42
RXPRSCLEN
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
BIT
7
6
5
4
PARAMETER
3
2
1
0
DESCRIPTION
DDR clock frequency mode for RXPC
1
DDR mode enabled
0
normal operating mode
DDR programming by I2C-bus
1
enable I2C-bus programming
0
enable programming by pin ENDDR
0
0
1
0
0
0
0
0
NAME
RXPCDDREN
I2CDDR
default value
Table 49 Register TXOCTDIV (address: E0h, default value: 00h)
BIT
7
6
5
4
PARAMETER
3
2
1
0
DESCRIPTION
0
0
0
M = 1; octave number 0
0
0
1
M = 2; octave number 1
0
1
0
M = 4; octave number 2
0
1
1
M = 8; octave number 3
1
0
0
M = 16; octave number 4
1
0
1
M = 32; octave number 5
1
1
0
M = 64; octave number 6
division ratio octave divider M; octave selection
x
x
x
x
x
0
0
0
0
0
NAME
TXDIV_M[2:0]
reserved
0
0
0
default value
Table 50 Register TXMAINDIV1 (address: E1h, default value: 01h)
BIT
7
6
5
4
PARAMETER
3
2
1
0
x
x
x
x
x
x
x
x
0
0
0
0
0
0
0
DESCRIPTION
division ratio divider N; TXN8 = MSB
NAME
TXN8
reserved
1
default value
Table 51 Register TXMAINDIV0 (address: E2h, default value: 00h)
BIT
PARAMETER
7
6
5
4
3
2
1
0
x
x
x
x
x
x
x
x
0
0
0
0
0
0
0
0
2003 Dec 16
DESCRIPTION
division ratio divider N; TXN0 = LSB
NAME
TXN[7:0]
default value
43
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
Table 52 Register TXFRACN2 (address: E3h, default value: 80h)
BIT
7
6
PARAMETER
5
4
3
2
1
0
x
x
x
x
x
x
DESCRIPTION
fractional divider: TXK21 = MSB
x
TXK[21:16]
reserved
TXNILFRAC control bit (NF)
1
TXNILFRAC
no fractional N functionality
0
1
NAME
fractional N functionality
0
0
0
0
0
0
0
default value
Table 53 Register TXFRACN1 (address: E4h, default value: 00h)
BIT
7
6
5
4
PARAMETER
3
2
1
0
x
x
x
x
x
x
x
x
0
0
0
0
0
0
0
0
DESCRIPTION
fractional divider
NAME
TXK[15:8]
default value
Table 54 Register TXFRACN0 (address: E5h, default value: 00h)
BIT
7
6
5
4
PARAMETER
3
2
1
0
x
x
x
x
x
x
x
x
0
0
0
0
0
0
0
0
DESCRIPTION
fractional divider; TXK0 = LSB
NAME
TXK[7:0]
default value
Table 55 Register MUXCON1 (address: F0h, default value: 62h)
BIT
7
6
5
4
PARAMETER
3
2
1
0
DESCRIPTION
parallel INPUT bus swapping
1
TXPD0 = MSB; TXPD3 = LSB (swapped)
0
TXPD3 = MSB; TXPD0 = LSB (normal)
parity polarity
1
even parity
0
odd parity
parity programming by I2C-bus
I2C-bus
1
by
0
by external pin PAREVEN
2003 Dec 16
inverted
0
normal
TXBUSSWAP
TXPAREVEN
I2CTXPAREVEN
interface
parallel clock input polarity
1
NAME
44
TXPCINV
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
BIT
7
6
5
4
PARAMETER
3
2
1
0
DESCRIPTION
parallel data input polarity
x
x
x
0
1
1
1
inverted
0
normal
NAME
TXPDINV
reserved
0
0
0
1
0
default value
Table 56 Register MUXCON0 (address: F1h, default value: 02h)
BIT
7
6
5
4
PARAMETER
3
2
1
0
DESCRIPTION
parallel clock output phase
1
0
0° phase shift
contra-directional clocking
parallel clock direction programming I2C-bus
1
by
0
I2C-bus
by external pin CLKDIR
1
FIFORESET
reset FIFO
0
normal mode
FIFO reset programming by I2C-bus
1
by
0
I2C-bus
x
2003 Dec 16
I2CFIFORES
interface
by external pin FIFORESET
DDR clock frequency mode for TXPC
0
I2CLKDIR
interface
FIFO reset
x
CLKDIR
co-directional clocking
0
0
TXPCOPHASE
90° phase shift
parallel clock direction
1
NAME
1
DDR mode enabled
0
normal mode
TXPCDDREN
reserved
0
0
0
0
1
0
default value
45
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
Table 57 Register TXMFOUTC (address: F2h, default value: A8h)
BIT
7
6
5
4
PARAMETER
3
2
1
0
DESCRIPTION
x
reserved
DDR clock frequency mode for TXPCO
1
DDR mode enabled
0
normal mode
parallel clock output polarity
1
inverted
0
normal
1
enabled
0
disabled
TXPCOEN
prescaler output polarity
1
inverted
0
normal
TXPRSCLINV
prescaler output enable
1
enabled
0
disabled
TXPRSCLEN
parity error output polarity
1
inverted
0
normal
TXPARERRINV
parity error output enable
1
enabled
0
disabled
0
1
0
1
0
0
TXPCODDREN
TXPCOINV
parallel clock output enable
1
NAME
TXPARERREN
0
default value
Table 58 Register TXRFOUTC1 (address: F3h, default value: 4Bh)
BIT
7
6
5
4
PARAMETER
3
2
1
0
x
x
DESCRIPTION
reserved
serial output data polarity
1
inverted
0
normal
enable serial data output
1
enabled
0
disabled
clock and data output swap
2003 Dec 16
NAME
1
swapped clock and data output
0
normal clock and data output
46
TXSDINV
TXSDEN
TXSDSCSWAP
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
BIT
7
6
5
4
PARAMETER
3
2
1
0
DESCRIPTION
serial clock output polarity
1
inverted
0
normal
TXSCINV
enable serial clock output
1
enabled
0
disabled
NAME
TXSCEN
serial clock output enable programming by I2C-bus I2CTXSCEN
1
by I2C-bus interface
0
by external pin TXSC
0
1
0
0
1
0
1
1
default value
Table 59 Register TXRFOUTC0 (address: F4h, default value: 03h)
BIT
7
6
5
4
PARAMETER
3
2
1
0
DESCRIPTION
0
0
0
0
minimum; 70 mV (p-p)
0
0
1
1
default; 300 mV (p-p)
1
1
1
1
maximum; 1100 mV (p-p)
serial output signal amplitude
x
x
RFS[3:0]
reserved
serial output termination
1
AC-coupled
0
DC-coupled
x
0
NAME
RFOUTTERMAC
reserved
0
2003 Dec 16
0
0
0
0
1
1
default value
47
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
LIMITING VALUES
In accordance with the Absolute Maximum Rating System (IEC 60134).
SYMBOL
PARAMETER
MIN.
MAX.
UNIT
VCC
analog supply voltage
−0.5
+3.6
V
VDD
digital supply voltage
−0.5
+3.6
V
Vn
DC voltage
0.7
VCC + 0.5 V
on pins RXPC(Q), RXPD0(Q) to RXPD3(Q), RXFP(Q),
RXPAR(Q), TXPARERR(Q), TXPCO(Q), RXPRSCL(Q) and
TXPRSCL(Q)
on pins RXSD(Q), CREF(Q), TXPC(Q), TXPD0(Q) to TXPD3(Q), −0.5
TXPAR(Q), UI, RREF, LOSTH, RSSI, LOS, CS, SDA, SCL, LM0
to LM2, INT, ENRX, ENTX, WINSIZE, INWINDOW, ENDDR,
LOWSWING, ENBA, PAREVEN, OVERFLOW, FIFORESET,
ENTXSC, TXSD(Q), TXSC(Q), LOL, FREF0, FREF1, CLKDIR
and IPUMP
VCC + 0.5 V
input current
In
on pins RXPC(Q), RXPD0(Q) to RXPD3(Q), RXFP(Q),
RXPAR(Q), TXPARERR(Q), TXPCO(Q), RXPRSCL(Q) and
TXPRSCL(Q)
−20
+20
mA
on pins RXSD(Q) and CREF(Q)
−30
+30
mA
on pin INT
−2
+2
mA
on pins TXPC(Q), TXPD0(Q) to TXPD3(Q) and TXPAR(Q)
−25
+25
mA
Tamb
ambient temperature
−40
+85
°C
Tj
junction temperature
−
125
°C
Tstg
storage temperature
−65
+150
°C
THERMAL CHARACTERISTICS
SYMBOL
Rth(j-a)
PARAMETER
CONDITIONS
thermal resistance from junction to ambient
notes 1 and 2
VALUE
UNIT
16
K/W
Notes
1. In compliance with JEDEC standards JESD51-5 and JESD51-7.
2. Four-layer Printed-Circuit Board (PCB) in still air with 36 plated vias connected with the heatsink and the second and
fourth layer in the PCB.
2003 Dec 16
48
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
CHARACTERISTICS
VCCA = VCCD = VCCO = 3.14 to 3.46 V; Tamb = −40 to +85 °C; Rth(j-a) < 16 K/W; all characteristics are specified for the
default test settings (see Table 60); all voltages are referenced to VEE; positive currents flow into the device; unless
otherwise specified.
SYMBOL
PARAMETER
CONDITIONS
MIN.
TYP.
MAX.
UNIT
Supplies
ICCA
analog supply
current
13
21
29
mA
ICCD
digital supply
current
350
395
456
mA
IDD
digital supply
current
0
0.3
1
mA
ICCO
supply current for
clock generator
41
55
64
mA
ICC(tot)
total supply
current
notes 1 and 2
404
471
550
mA
Ptot
total power
dissipation
notes 1 and 2
1.3
1.6
1.8
W
notes 1 and 2
CMOS inputs: pins UI, CS, DR0 to DR2, LM0 to LM2, ENRX, ENTX, PAREVEN, WINSIZE, LOWSWING, FREF0,
FREF1, ENBA, FIFORESET, CLKDIR, ENTXSC and ENDDR
VIL
LOW-level input
voltage
−
−
0.2VCC
V
VIH
HIGH-level input
voltage
0.8VCC
−
−
V
IIL
LOW-level input
current
VIL = 0 V
−200
−
−
µA
IIH
HIGH-level input
current
VIH = VCC
−
−
10
µA
0
−
0.2
V
VCC − 0.2
−
VCC
V
CMOS outputs: pins LOS, INT, INWINDOW, LOL and OVERFLOW
VOL
LOW-level output
voltage
IOL = 1 mA
VOH
HIGH-level output IOH = −0.5 mA
voltage
Open-drain output: pin INT
VOL
LOW-level output
voltage
IOL = 1 mA
0
−
0.2
V
IOH
HIGH-level output VOH = VCC
current
−
−
10
µA
220
300
380
mV
Serial outputs: pins TXSD(Q) and TXSC(Q)
single-ended with 50 Ω
external load; DC swing;
note 3
Vo(p-p)
default output
voltage swing
(peak-to-peak
value)
Zo
output impedance single-ended to VCC
40
50
60
Ω
tr
rise time
−
60
90
ps
2003 Dec 16
20% to 80%
49
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
SYMBOL
PARAMETER
TZA3015HW
CONDITIONS
MIN.
TYP.
MAX.
UNIT
tf
fall time
80% to 20%
−
60
90
ps
tD-C
data-to-clock
delay
between differential
crossovers; see Fig.27
−50
−
+50
ps
δ
duty cycle signal
TXSC(Q)
between differential
crossovers
40
50
60
%
fbit
output bit rate
30
−
3200
Mbit/s
Serial input: pins RXSD(Q)
Vi(p-p)
input voltage
swing
(peak-to-peak
value)
single-ended; note 4;
PRBS (27 − 1)
12
−
500
mV
Vi(sens)(p-p)
input voltage
sensitivity
(peak-to-peak
value)
single-ended;
PRBS (27 − 1)
−
5
12
mV
Vsl
typical slice level
range
note 5
−50
−
+50
mV
Zi
input impedance
differential
80
100
120
Ω
fbit
input data rate
30
−
3200
Mbit/s
LVDS outputs: pins RXPD0(Q) to RXPD3(Q), RXPC(Q), RXPAR(Q), TXPARERR(Q), RXPRSCL(Q),TXPRSCL(Q),
RXFP(Q) and TXPCO(Q)
Vo(dif)
differential output
voltage
RL = 100 Ω; DC-coupled
low swing mode, DC
250
300
360
mV
high swing mode, DC
400
500
600
mV
Vo(cm)
common mode
output voltage
RL = 100 Ω, DC-coupled
1.10
1.22
1.33
V
tr, tf
rise and fall time
CL = 1 pF
100
200
250
ps
tD-C
data to clock
delay
normal mode; see Fig.28 −200
+200
ps
DDR mode; see Fig.28
1/
duty cycle
RXPC(Q)
normal mode
45
50
55
%
DDR mode
47
50
53
%
δTX
duty cycle
TXPCO(Q)
normal mode
45
50
55
%
DDR mode
47
50
53
%
skew
channel to
channel skew
RXPD0 to RXPD3,
RXPAR and RXFP;
note 6
−
−
100
ps
0
−
VCC
mV
−100
−
+100
mV
δRX
4Tclk
−
− 250
1/
4Tclk
− 50
1/
4Tclk
+ 150
ps
LVDS inputs: pins TXPD0(Q) to TXPD3(Q), TXPAR(Q) and TXPC(Q)
Vi
input voltage
range
Vi(th)(dif)
differential input
voltage threshold
2003 Dec 16
DC
50
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
SYMBOL
PARAMETER
TZA3015HW
CONDITIONS
MIN.
TYP.
−
MAX.
UNIT
Vi(p-p)
input voltage
swing
(peak-to-peak
value)
single-ended; note 6
−
Vi(hys)
differential
hysteresis input
voltage
Tamb = 0 °C to 85 °C
25
−
−
mV
Tamb = −40 °C to 0 °C
15
−
−
mV
80
100
120
Ω
1000
mV
zi(dif)
differential input
impedance
th(co)
hold time
co-directional
clocking
see Fig.29
−
150
300
ps
tsu(co)
set-up time
co-directional
clocking
see Fig.29
−
20
300
ps
th(contra)
hold time contradirectional
clocking
see Fig.29
−
−1100
−850
ps
tsu(contra)
set-up time
contra-directional
clocking
see Fig.29
−
1300
1450
ps
th(co)DDR
hold time
co-directional
clocking in DDR
mode
fbit = 124 to 800 Mbit/s;
see Fig.29
−
0.3Tclk + 40
0.3Tclk + 240
ps
fbit = 30 to 124 Mbit/s;
see Fig.29; note 6
−
4780
5000
ps
fbit = 124 to 800 Mbit/s;
see Fig.29
−
−1/4Tclk − 130
−1/4Tclk + 200
ps
fbit = 30 to 124 Mbit/s;
see Fig.29; note 6
−
−4560
−3700
ps
hold time
contra-directional
clocking in DDR
mode
see Fig.29
−
−1/4Tclk − 1200 −1/4Tclk − 1000 ps
tsu(contra)DDR set-up time
contra-directional
clocking in DDR
mode
see Fig.29
−
1/
δ
note 6
40
50
60
%
tsu(co)DDR
th(contra)DDR
set-up time
co-directional
clocking in DDR
mode
duty cycle clock
TXPC(Q)
4Tclk
+ 1400
1/
4Tclk
+ 1600
ps
Reference frequency input; pins CREF(Q)
Vi(p-p)
input swing
(peak-to-peak
value)
single-ended
50
−
1000
mV
Vi
input voltage
range
note 6
VCC − 1
−
VCC + 0.25
V
Zi
input impedance
single-ended to VCC
40
50
60
Ω
2003 Dec 16
51
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
SYMBOL
PARAMETER
TZA3015HW
CONDITIONS
MIN.
TYP.
MAX.
UNIT
∆fCREF
reference clock
frequency
accuracy
SDH/SONET
requirement
−20
−
+20
ppm
fCREF
reference clock
frequency
see Section “Reference
clock programming”;
R = 1, 2, 4, 8, 16 or 32
18 × R
19.44 × R
21 × R
MHz
10 kΩ resistor to VEE
1.17
1.21
1.26
V
Reference voltage; pin RREF
Vref
reference voltage
Received signal strength indicator; pin RSSI
Vi(p-p)
detectable input
voltage swing on
serial data input
(peak-to-peak
value)
single-ended
5
−
500
mV
SRSSI
RSSI sensitivity
see Fig.4
15
17
20
mV/dB
VRSSI(32mV)
output voltage
serial data input voltage
Vi = 32 mV;
PRBS(231 − 1)
580
680
780
mV
∆Vo
output voltage
variation
input 30 to 3200 Mbit/s;
PRBS(231 − 1);
VCC = 3.14 to 3.47 V;
∆T = 120 °C
−50
−
+50
mV
Zo
output impedance
−
1
10
Ω
Io(source)
output source
current
−
−
1
mA
Io(sink)
output sink
current
−
−
0.4
mA
LOS detector
hys
hysteresis
note 7
−
3
−
dB
ta
assert time
∆Vi(p-p) = 3 dB
−
−
5
µs
td
de-assert time
∆Vi(p-p) = 3 dB
−
−
5
µs
CLUPPLLHG = 0
−
−0.1
−
mA
CLUPPLLHG = 1
−
−1
−
mA
charge pump sink CLUPPLLHG = 0
current
CLUPPLLHG = 1
−
0.1
−
mA
−
1
−
mA
Clean-up PLL: pin IPUMP
Icp(source)
Icp(sink)
2003 Dec 16
charge pump
source current
52
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
SYMBOL
PARAMETER
TZA3015HW
CONDITIONS
MIN.
TYP.
MAX.
UNIT
Jitter characteristics
Jtol(p-p)
jitter tolerance to
serial data input
signal
(peak-to-peak
value)
STM1/OC3 mode;
PRBS(223 − 1)
f = 6.5 kHz
3
>10
−
UI
f = 65 kHz
0.3
>1
−
UI
f = 1 MHz
0.3
>0.5
−
UI
f = 25 kHz
3
>10
−
UI
f = 250 kHz
0.3
>1
−
UI
f = 5 MHz
0.3
>0.5
−
UI
f = 100 kHz
3
10
−
UI
f = 1 MHz
0.3
1
−
UI
f = 20 MHz
0.3
0.5
−
UI
−
−
16
mUI
−
−
4
mUI
−
−
4
mUI
f = 1 kHz to 5 MHz
−
−
63
mUI
f = 12 kHz to 5 MHz
−
−
13
mUI
f = 250 kHz to 5 MHz
−
−
13
mUI
f = 5 kHz to 20 MHz
−
32
250
mUI
f = 12 kHz to 20 MHz
−
30
50
mUI
f = 1 MHz to 20 MHz
−
6
50
mUI
STM4/OC12 mode;
PRBS(223 − 1)
STM16/OC48 mode;
PRBS(223 − 1)
Jgen(p-p)
jitter generation at STM1/OC3 mode;
serial data and
notes 8 and 9
clock output
f = 500 Hz to 1.3 MHz
(peak-to-peak
f = 12 kHz to 1.3 MHz
value)
f = 65 kHz to 1.3 MHz
STM4/OC12 mode;
notes 8 and 9
STM16/OC48 mode;
note 8
PLL characteristics receiver
tacq
acquisition time
30 Mbit/s; note 6
−
−
200
µs
tacq(pc)
acquisition time at 30 Mbit/s; note 6
power cycle
−
−
10
ms
tacq(o)
acquisition time
octave change
−
−
10
µs
30 Mbit/s; note 6
I2C-bus input and output: pins SCL and SDA
VIL
LOW-level input
voltage
−
−
0.2VCC
V
VIH
HIGH-level input
voltage
0.8VCC
−
−
V
2003 Dec 16
53
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
SYMBOL
PARAMETER
TZA3015HW
CONDITIONS
MIN.
TYP.
MAX.
UNIT
Vhys
hysteresis of
Schmitt-trigger
inputs
note 6
0.05VCC
−
−
V
VOL
LOW-level output
voltage on pin
SDA (open-drain)
IOL = 3 mA
0
−
0.4
V
ILI
input leakage
current
−10
−
+10
µA
Ci
input capacitance note 6
−
−
10
pF
I2C-bus timing; note 6
fSCL
SCL clock
frequency
−
−
100
kHz
tLOW
SCL LOW time
1.3
−
−
µs
tHD;STA
hold time START
condition
0.6
−
−
µs
tHIGH
SCL HIGH time
0.6
−
−
µs
tSU;STA
set-up time
START condition
0.6
−
−
µs
tHD;DAT
data hold time
0
−
0.9
µs
tSU;DAT
data set-up time
100
−
−
ns
tSU;STO
set-up time STOP
condition
0.6
−
−
µs
tr
SCL and SDA
rise time
20
−
300
ns
tf
SCL and SDA fall
time
20
−
300
ns
tBUF
bus free time
between STOP
and START
1.3
−
−
µs
Cb
capacitive load for
each bus line
−
−
400
pF
tSP
pulse width of
allowable spikes
0
−
50
ns
VnL
noise margin at
LOW-level
0.1VCC
−
−
V
VnH
noise margin at
HIGH-level
0.2VCC
−
−
V
Notes
1. For the typical specification LVDS outputs: RXPAR(Q), RXPRSCL(Q), TXPARERR(Q), TXPCO(Q) and
TXPRSCL(Q) are disabled. Also serial output TXSC(Q) is disabled.
2. The following conditions are valid for the maximum specification and are additional to the default settings: bit
CLUPPLLEN = 1 (clean-up PLL is enabled); bit CLUPPLLHG = 1 (high gain); line loop back is enabled; pin
LOWSWING = LOW (high swing for LVDS outputs); bits RFS[3:0] = 1111 (maximum output swing for TXSD(Q) and
2003 Dec 16
54
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
TXSC(Q). These maximum settings yield the following maximum specification values: ICCD = 680 mA,
ICC(tot) = 774 mA and Ptot = 2.7 W.
3. The output swing is adjustable between 70 mV (typical) and 1100 mV (typical) in 16 steps controlled by bits RFS[3:0]
of the register TXRFOUTC0 (F4h).
4. The RF input is protected against a differential overvoltage; the maximum input current is 30 mA.
5. The slice level is adjustable in 256 steps controlled by register LIMSL (D2h).
6. Guaranteed by design.
7. The hysteresis is adjustable in 8 steps controlled by bits HTLCB[2:0] of register LIMLOSCON (D1h).
8. Reference frequency of 19.44 MHz, with a phase-noise of less than −140 dBc for frequencies of more than 12 kHz
from the carrier (measured during 60 seconds, within the appropriate bandwidth).
9. For bit rates lower than 1.8 Gbit/s, the jitter decreases by the octave division ratio M.
Table 60 Default test settings
PIN
SETTING
UI = LOW
pre-programmed mode
DR0 = LOW, DR1 = HIGH, DR2 = LOW
STM16/OC48
LM0 = HIGH, LM1 = HIGH, LM2 = HIGH
normal mode
ENRX = HIGH
receiver enabled
ENTX = HIGH
transmitter enabled
ENDDR = LOW
DDR mode disabled
LOWSWING = HIGH
low LVDS swing
FREF0 = HIGH, FREF1 = HIGH
19.44 MHz reference
RREF
RRREF = 10 kΩ to VEE
IPUMP
open circuit
RSSI
open circuit
LOSTH
VLOSTH = 0.6 V
ENTXSC = HIGH
serial output clock enabled
WINSIZE = HIGH
1000 ppm
ENBA = HIGH
automatic byte alignment
PAREVEN = HIGH
even parity
CREF(Q)
AC-coupled, fi = 19.44 MHz, Vi = 0.2 V (p-p) single-ended
RXSD(Q)
input STM16; PRBS (223 − 1)
RXPD0(Q) to RXPD3(Q), RXFP(Q), RXPAR(Q),
RXPC(Q), TXPCO(Q), TXPARERR(Q), TXPRSCL(Q)
and RXPRSCL(Q)
100 Ω differential outputs
TXPC(Q), TXPD0(Q) to TXPD3(Q) and TXPAR
open circuit
FIFORESET = LOW
normal mode
CLKDIR = HIGH
co-directional clocking
TXSD(Q) and TXSC(Q)
external load of 50 Ω to VCC
CMOS outputs
not loaded
2003 Dec 16
55
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
handbook, full pagewidth
TXSC
t D-C
TXSD
MGX390
The timing is measured from the crossover point of the clock output signal to the crossover point of the data output (all signals are differential).
Fig.27 Serial bus output timing.
handbook, full pagewidth
RXPC
t D-C
RXPD0 to RXPD3
RXFP, RXPAR
MGX478
The timing is measured from the crossover point of the clock output signal to the crossover point of the data output (all signals are differential).
Fig.28 Parallel bus output timing.
T clk
handbook, full pagewidth
TXPCO, TXPC
th
t su
TXPD0 to TXPD3
valid data
MCE422
The timing is measured from the crossover point of the reference signal to the crossover point of the input.
Fig.29 Parallel bus co-directional (TXPC) and contra-directional (TXPCO) timing.
2003 Dec 16
56
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
APPLICATION INFORMATION
The transfer has a first order roll-off (i.e. 20 dB/decade), up
to the bandwidth of the VCXO control input. If a second
order roll-off is required C2 may be added, as long as
C1 + C2
----------------------------------------- > 2 × f –3dB
2 × π × C1 × C2
Calculations on the clean-up PLL
The important specifications of the clean-up PLL are the
bandwidth (f−3dB) and the jitter peaking. If these are known,
the component parameters can be calculated. First
assume that the bandwidth of the VCXO control input
(f−3dB(vcxo)) is much higher than f−3dB and C2 is left out. This
simplifies the loop into a second order, type II PLL. In a
second order PLL, the damping factor ζ determines the
amount of peaking. To obtain peaking of less than 0.1 dB,
ζ must be higher than 4.3. For peaking of less than
0.05 dB, ζ must be higher than 6. See Fig.30 for an
example.
Example: The clean-up PLL uses a VCXO with a
frequency of 20 MHz and has a gain KVCXO = 2000 Hz/V.
The bandwidth of the control input is f−3dB(VCXO) = 10 kHz.
Since the reference frequency is 20 MHz, the reference
divider ratio RDIV = 1. According to the specification, the
maximum allowed jitter peaking is 0.1 dB. To add some
margin the design is for less than 0.05 dB peaking, so
ζ = 6. Also according to the specification, f−3dB should be
less than 100 kHz. To satisfy the conditions as previously
described, f−3 dB < 0.5 × f−3 dB(VCXO) < 5 kHz. To cope with
component tolerances, f−3dB(VCXO) = 2.5 kHz is chosen.
Now R and C1 may be calculated with the following
formulas:
RDIV × 2 × π × f –3dB
R = ----------------------------------------------------K VCXO × I CP
1 × 2 × π × 2500
R = ----------------------------------------------Ω = 78540Ω
–6
2000 × 100 × 10
2
K VCXO × I CP × ζ
C 1 = -----------------------------------------------------2
2
RDIV × π × ( f –3dB )
2000 × 100µF × 6
C 1 = ------------------------------------------------- = 116.7nF
2
2
1 × π × 2500
Where:
Choosing ICP = 1 mA yields R = 7854 Ω and
C1 = 1.167 µF.
2
RDIV = reference divider ratio (1, 2, 4, 8, 16 or 32)
To calculate f-3dB and ζ, if R and C1 are known, use the
following formulas:
f−3dB = clean-up PLL bandwidth in Hz
KVCXO = VCXO gain in Hz/V
K VCXO × I CP × R
f –3dB = -----------------------------------------2 × π × RDIV
ICP = charge pump current in A (100 µA or 1 mA,
depending on I2C-bus bit CLUPPLLHG)
ζ = damping factor.
K VCXO × I CP × C1
R
ζ = ---- × ----------------------------------------------RDIV
2
These formulas are valid if:
ζ >> 1 and
f−3dB(VCXO) > 2 × f−3dB and
C1 + C2
---------------------------------------------------- > 2 × f –3dB .
2 × π × R × C1 × C2
2003 Dec 16
57
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
handbook, full pagewidth
MCE423
Jtransfer [dB]
jitter peaking
f−3 dB
f−3 dB (VCXO)
log(f) [Hz]
−3
20 dB/decade
40 dB/decade
Fig.30 Clean-up PLL jitter transfer.
I/O CONFIGURATIONS
handbook, full pagewidth
SWING CONTROL
VCC
50 Ω
Vbias
50 Ω
50 Ω
transmission
lines
OUT
50 Ω
to highimpedance
input
50 Ω
OUTQ
Iswing
in
on-chip
off-chip
Fig.31 Serial RF output (AC-coupled).
2003 Dec 16
58
50 Ω
MDB068
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
handbook, full pagewidth
SWING CONTROL
VCC
50 Ω
50 Ω
50 Ω
transmission
lines
OUT
50 Ω
to highimpedance
input
50 Ω
OUTQ
Iswing
in
on-chip
off-chip
MDB069
Fig.32 Serial RF output (DC-coupled).
VCCD
handbook, halfpage
D
50 Ω
30 kΩ
300 Ω
50 Ω
1.8 pF
30 kΩ
DQ
VEE
MGX391
Fig.33 LVDS input.
2003 Dec 16
59
50 Ω
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
handbook, full pagewidth
VCC
COMMON MODE CONTROL
transmission
lines
OUT
50 Ω
Vref
to highimpedance
input
100 Ω
OUTQ
50 Ω
in
MGX392
swing-setting
on-chip
off-chip
Fig.34 LVDS output (DC-coupled).
handbook, full pagewidth
VCC
COMMON MODE CONTROL
Vbias
AC
coupling
transmission
lines
50 Ω
50 Ω
OUT
50 Ω
Vref
OUTQ
to highimpedance
input
50 Ω
in
MGX393
swing-setting
on-chip
off-chip
Fig.35 LVDS output (AC-coupled).
2003 Dec 16
60
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
VCC
50 Ω
CREF
50 Ω
CREFQ
VEE
Fig.36 Reference clock input.
2003 Dec 16
61
001aaa056
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
PACKAGE OUTLINE
HTQFP100: plastic thermal enhanced thin quad flat package; 100 leads;
body 14 x 14 x 1 mm; exposed die pad
SOT638-1
c
y
exposed die pad side
X
Dh
A
75
51
76
50
ZE
e
E HE
Eh
A
A2
(A3)
A1
wM
θ
bp
Lp
pin 1 index
L
detail X
26
100
1
25
bp
e
w M
ZD
v M A
D
B
HD
v M B
0
10 mm
scale
DIMENSIONS (mm are the original dimensions)
A
UNIT max.
mm
1.2
A1
A2
A3
bp
c
D(1)
Dh
E(1)
Eh
e
0.15
0.05
1.05
0.95
0.25
0.27
0.17
0.20
0.09
14.1
13.9
7.1
6.1
14.1
13.9
7.1
6.1
0.5
HD
HE
16.15 16.15
15.85 15.85
L
Lp
v
w
y
1
0.75
0.45
0.2
0.08
0.08
ZD(1) ZE(1)
θ
1.15
0.85
7°
0°
1.15
0.85
Note
1. Plastic or metal protrusions of 0.25 mm maximum per side are not included.
OUTLINE
VERSION
REFERENCES
IEC
JEDEC
JEITA
ISSUE DATE
01-03-30
03-04-07
SOT638-1
2003 Dec 16
EUROPEAN
PROJECTION
62
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
To overcome these problems the double-wave soldering
method was specifically developed.
SOLDERING
Introduction to soldering surface mount packages
If wave soldering is used the following conditions must be
observed for optimal results:
This text gives a very brief insight to a complex technology.
A more in-depth account of soldering ICs can be found in
our “Data Handbook IC26; Integrated Circuit Packages”
(document order number 9398 652 90011).
• Use a double-wave soldering method comprising a
turbulent wave with high upward pressure followed by a
smooth laminar wave.
There is no soldering method that is ideal for all surface
mount IC packages. Wave soldering can still be used for
certain surface mount ICs, but it is not suitable for fine pitch
SMDs. In these situations reflow soldering is
recommended.
• For packages with leads on two sides and a pitch (e):
– larger than or equal to 1.27 mm, the footprint
longitudinal axis is preferred to be parallel to the
transport direction of the printed-circuit board;
– smaller than 1.27 mm, the footprint longitudinal axis
must be parallel to the transport direction of the
printed-circuit board.
Reflow soldering
Reflow soldering requires solder paste (a suspension of
fine solder particles, flux and binding agent) to be applied
to the printed-circuit board by screen printing, stencilling or
pressure-syringe dispensing before package placement.
Driven by legislation and environmental forces the
worldwide use of lead-free solder pastes is increasing.
The footprint must incorporate solder thieves at the
downstream end.
• For packages with leads on four sides, the footprint must
be placed at a 45° angle to the transport direction of the
printed-circuit board. The footprint must incorporate
solder thieves downstream and at the side corners.
Several methods exist for reflowing; for example,
convection or convection/infrared heating in a conveyor
type oven. Throughput times (preheating, soldering and
cooling) vary between 100 and 200 seconds depending
on heating method.
During placement and before soldering, the package must
be fixed with a droplet of adhesive. The adhesive can be
applied by screen printing, pin transfer or syringe
dispensing. The package can be soldered after the
adhesive is cured.
Typical reflow peak temperatures range from
215 to 270 °C depending on solder paste material. The
top-surface temperature of the packages should
preferably be kept:
Typical dwell time of the leads in the wave ranges from
3 to 4 seconds at 250 °C or 265 °C, depending on solder
material applied, SnPb or Pb-free respectively.
• below 225 °C (SnPb process) or below 245 °C (Pb-free
process)
A mildly-activated flux will eliminate the need for removal
of corrosive residues in most applications.
– for all BGA, HTSSON-T and SSOP-T packages
– for packages with a thickness 2.5 mm
Manual soldering
– for packages with a thickness < 2.5 mm and a
volume ≥ 350 mm3 so called thick/large packages.
Fix the component by first soldering two
diagonally-opposite end leads. Use a low voltage (24 V or
less) soldering iron applied to the flat part of the lead.
Contact time must be limited to 10 seconds at up to
300 °C.
• below 240 °C (SnPb process) or below 260 °C (Pb-free
process) for packages with a thickness < 2.5 mm and a
volume < 350 mm3 so called small/thin packages.
When using a dedicated tool, all other leads can be
soldered in one operation within 2 to 5 seconds between
270 and 320 °C.
Moisture sensitivity precautions, as indicated on packing,
must be respected at all times.
Wave soldering
Conventional single wave soldering is not recommended
for surface mount devices (SMDs) or printed-circuit boards
with a high component density, as solder bridging and
non-wetting can present major problems.
2003 Dec 16
63
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
Suitability of surface mount IC packages for wave and reflow soldering methods
SOLDERING METHOD
PACKAGE(1)
WAVE
REFLOW(2)
BGA, HTSSON..T(3), LBGA, LFBGA, SQFP, SSOP..T(3), TFBGA,
USON, VFBGA
not suitable
suitable
DHVQFN, HBCC, HBGA, HLQFP, HSO, HSOP, HSQFP, HSSON,
HTQFP, HTSSOP, HVQFN, HVSON, SMS
not suitable(4)
suitable
PLCC(5), SO, SOJ
suitable
suitable
not
recommended(5)(6)
suitable
SSOP, TSSOP, VSO, VSSOP
not
recommended(7)
suitable
CWQCCN..L(8), PMFP(9), WQCCN..L(8)
not suitable
LQFP, QFP, TQFP
not suitable
Notes
1. For more detailed information on the BGA packages refer to the “(LF)BGA Application Note” (AN01026); order a copy
from your Philips Semiconductors sales office.
2. All surface mount (SMD) packages are moisture sensitive. Depending upon the moisture content, the maximum
temperature (with respect to time) and body size of the package, there is a risk that internal or external package
cracks may occur due to vaporization of the moisture in them (the so called popcorn effect). For details, refer to the
Drypack information in the “Data Handbook IC26; Integrated Circuit Packages; Section: Packing Methods”.
3. These transparent plastic packages are extremely sensitive to reflow soldering conditions and must on no account
be processed through more than one soldering cycle or subjected to infrared reflow soldering with peak temperature
exceeding 217 °C ± 10 °C measured in the atmosphere of the reflow oven. The package body peak temperature
must be kept as low as possible.
4. These packages are not suitable for wave soldering. On versions with the heatsink on the bottom side, the solder
cannot penetrate between the printed-circuit board and the heatsink. On versions with the heatsink on the top side,
the solder might be deposited on the heatsink surface.
5. If wave soldering is considered, then the package must be placed at a 45° angle to the solder wave direction.
The package footprint must incorporate solder thieves downstream and at the side corners.
6. Wave soldering is suitable for LQFP, TQFP and QFP packages with a pitch (e) larger than 0.8 mm; it is definitely not
suitable for packages with a pitch (e) equal to or smaller than 0.65 mm.
7. Wave soldering is suitable for SSOP, TSSOP, VSO and VSSOP packages with a pitch (e) equal to or larger than
0.65 mm; it is definitely not suitable for packages with a pitch (e) equal to or smaller than 0.5 mm.
8. Image sensor packages in principle should not be soldered. They are mounted in sockets or delivered pre-mounted
on flex foil. However, the image sensor package can be mounted by the client on a flex foil by using a hot bar
soldering process. The appropriate soldering profile can be provided on request.
9. Hot bar or manual soldering is suitable for PMFP packages.
ADDITIONAL SOLDERING INFORMATION
The die pad has to be soldered to the PCB for thermal and grounding reasons.
2003 Dec 16
64
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
DATA SHEET STATUS
LEVEL
DATA SHEET
STATUS(1)
PRODUCT
STATUS(2)(3)
Development
DEFINITION
I
Objective data
II
Preliminary data Qualification
This data sheet contains data from the preliminary specification.
Supplementary data will be published at a later date. Philips
Semiconductors reserves the right to change the specification without
notice, in order to improve the design and supply the best possible
product.
III
Product data
This data sheet contains data from the product specification. Philips
Semiconductors reserves the right to make changes at any time in order
to improve the design, manufacturing and supply. Relevant changes will
be communicated via a Customer Product/Process Change Notification
(CPCN).
Production
This data sheet contains data from the objective specification for product
development. Philips Semiconductors reserves the right to change the
specification in any manner without notice.
Notes
1. Please consult the most recently issued data sheet before initiating or completing a design.
2. The product status of the device(s) described in this data sheet may have changed since this data sheet was
published. The latest information is available on the Internet at URL http://www.semiconductors.philips.com.
3. For data sheets describing multiple type numbers, the highest-level product status determines the data sheet status.
DEFINITIONS
DISCLAIMERS
Short-form specification  The data in a short-form
specification is extracted from a full data sheet with the
same type number and title. For detailed information see
the relevant data sheet or data handbook.
Life support applications  These products are not
designed for use in life support appliances, devices, or
systems where malfunction of these products can
reasonably be expected to result in personal injury. Philips
Semiconductors customers using or selling these products
for use in such applications do so at their own risk and
agree to fully indemnify Philips Semiconductors for any
damages resulting from such application.
Limiting values definition  Limiting values given are in
accordance with the Absolute Maximum Rating System
(IEC 60134). Stress above one or more of the limiting
values may cause permanent damage to the device.
These are stress ratings only and operation of the device
at these or at any other conditions above those given in the
Characteristics sections of the specification is not implied.
Exposure to limiting values for extended periods may
affect device reliability.
Right to make changes  Philips Semiconductors
reserves the right to make changes in the products including circuits, standard cells, and/or software described or contained herein in order to improve design
and/or performance. When the product is in full production
(status ‘Production’), relevant changes will be
communicated via a Customer Product/Process Change
Notification (CPCN). Philips Semiconductors assumes no
responsibility or liability for the use of any of these
products, conveys no licence or title under any patent,
copyright, or mask work right to these products, and
makes no representations or warranties that these
products are free from patent, copyright, or mask work
right infringement, unless otherwise specified.
Application information  Applications that are
described herein for any of these products are for
illustrative purposes only. Philips Semiconductors make
no representation or warranty that such applications will be
suitable for the specified use without further testing or
modification.
2003 Dec 16
65
Philips Semiconductors
Preliminary specification
30 Mbit/s to 3.2 Gbit/s A-rate
4-bit fibre optic transceiver
TZA3015HW
PURCHASE OF PHILIPS I2C COMPONENTS
Purchase of Philips I2C components conveys a license under the Philips’ I2C patent to use the
components in the I2C system provided the system conforms to the I2C specification defined by
Philips. This specification can be ordered using the code 9398 393 40011.
2003 Dec 16
66
Philips Semiconductors – a worldwide company
Contact information
For additional information please visit http://www.semiconductors.philips.com.
Fax: +31 40 27 24825
For sales offices addresses send e-mail to: [email protected].
SCA75
© Koninklijke Philips Electronics N.V. 2003
All rights are reserved. Reproduction in whole or in part is prohibited without the prior written consent of the copyright owner.
The information presented in this document does not form part of any quotation or contract, is believed to be accurate and reliable and may be changed
without notice. No liability will be accepted by the publisher for any consequence of its use. Publication thereof does not convey nor imply any license
under patent- or other industrial or intellectual property rights.
Printed in The Netherlands
R56/04/pp67
Date of release: 2003
Dec 16
Document order number:
9397 750 12216