TI TLK2226

TLK2226
www.ti.com
SLLS689 – JANUARY 2006
6 PORT GIGABIT ETHERNET TRANSCEIVER
FEATURES
•
•
•
•
•
•
•
•
•
•
•
•
•
Six 1.25 Gigabits Per Second (Gbps)
Synchronizable Transceivers (Support for 100
Mbps 100Base-FX Mode)
Configurable 1, 2, 3, 4, 5, or 6 Port Operation
via MDIO
Low Power Consumption <1.5W at 1.25Gb/sec
IEEE 802.3z Gigabit Ethernet Compliant
Differential VML Transmit Outputs With no
External Components Necessary
Programmable High Speed Output
Pre-Emphasis Levels
Nibble Wide RTBI/RGMII Compliant Interface
Selectable Clock Tolerance Compensation
Selectable on Chip Physical Coding Sublayer
(PCS) Functions Including 8b/10b IEEE 802.3z
Compliant Encoder and Decoder
1/10th Rate Capability for 125 Mbps Operation
With Automatic Rate Sense Capability
JEDEC 1.5V HSTL (Extendable to 1.8 V) on
Parallel Data Busses
JEDEC 1.8/2.5/3.3 V LVCMOS on REFCLK and
Control Pins
Internal Series Termination on HSTL Outputs
to Drive 50 Ω Lines
•
•
•
•
•
•
•
•
•
•
•
•
•
•
Standard Parallel Interface Timing:
– Source Centered/Aligned Timing on Inputs
– Source Centered/Aligned Timing on Outputs
Comprehensive Suite Of Built-In Testability
IEEE 802.3 Clause 22 Management Data
Interface (MDIO) Support
IEEE 1149.1 JTAG Support
Hot Plug Protection on Serial I/O
No External Filter Components Required for
PLLs
Small Footprint 15 × 15 mm, 196-Pin, 1,0 mm
Ball Pitch BGA
Advanced Low Power 0.18 µm CMOS
Technology
Commercial Temperature Rating (0°C to 70°C)
Repeater Mode:
– In the Serial Domain Using Recovered Clock
(Non-Retiming)
– In the Parallel Domain Using REFCLK
(Retiming)
Jumbo Packet (9300 Byte) Support
Clause 36A Test Pattern Gen/Ver
SGMII Mode Support (10/100/1000 Mbps)
Parallel Port Swap Mode
FUNCTIONAL OVERVIEW
The TLK2226 is the third generation of Gigabit Ethernet transceivers from Texas Instruments combining high port
density and ultra-low power in a small form factor footprint. The TLK2226 provides for high-speed full-duplex
point-to-point data transmissions based on the IEEE802.3z 1000Mbps Ethernet specification. The TLK2226
supports data rates from 1.0 Gbps through 1.3 Gbps. Each channel is capable of operating at 125 Mbps for IEEE
Ethernet 100FX mode, and the rate for each channel may be set through MDIO or automatically sensed and set
by the TLK2226.The primary application of this device is to provide building blocks for developing point-to-point
baseband data transmission over controlled impedance media of 50 Ω. The transmission media can be printed
circuit board traces, copper cables or fiber-optical media. The ultimate rate and distance of data transfer is
dependent upon the attenuation characteristics of the media and the noise coupling to the environment.
This device performs the data encoding, decoding, serialization, deserialization, clock extraction and clock
tolerance compensation functions for a physical layer interface device. Each channel operates at up to 1.3 Gbps
providing up to 7.5 Gbps of aggregate data bandwidth over a copper or optical media interface.
The TLK2226 has six channels of SERDES each with 5 bit busses clocked double data rate (DDR). The parallel
interface accepts un-encoded or 5-bit wide encoded data aligned to both the rising and falling edge of the
transmit clock. The data is clocked low-order bits first, (i.e. bits 0–4 of the 8b/10b encoded data) on the rising
edge of the transmit clock and the high-order bits (i.e. bits 5–9 of the 8b/10b encoded data) are clocked on the
falling edge of the clock. The receive path interface is clocked in the same manner.
Please be aware that an important notice concerning availability, standard warranty, and use in critical applications of Texas
Instruments semiconductor products and disclaimers thereto appears at the end of this data sheet.
PRODUCTION DATA information is current as of publication date.
Products conform to specifications per the terms of the Texas
Instruments standard warranty. Production processing does not
necessarily include testing of all parameters.
Copyright © 2006, Texas Instruments Incorporated
TLK2226
www.ti.com
SLLS689 – JANUARY 2006
The TLK2226 supports two modes that can repeat data from a channel’s serial receiver to the adjacent channel’s
serial transmitter: serial repeater mode and retiming repeater mode. In serial repeater mode, the clock/data
recovery (CDR) function will lock onto the incoming data stream and will pass this data stream over to the
adjacent channel to be retransmitted. In serial repeater mode, the data stream stays in the same clock domain
as the incoming data; there is no clock tolerance compensation function performed to align the data stream to
the reference clock domain. In retiming repeater mode, the CDR function will lock onto the incoming data stream
but the data will be deserialized and passed through the CTC FIFO and then serialized for transmission on the
adjacent channel. In both repeater modes, the received data for Channel A is retransmitted on the output for
channel B while the received data for Channel B is retransmitted on the output for Channel A. Likewise, Channel
C is paired with Channel D, and Channel E is paired with Channel F. The repeater modes can be enabled on a
channel by channel basis through the use of register control bits programmed through the MDIO interface.
The recovered data clock frequency can be aligned to the reference clock on each channel by means of a clock
tolerance compensation circuit and internal FIFO that will insert or drop Idle 20-bit IDLE codes as needed in the
absence of data. The received data for all channels are aligned to a single receive data clock that is a buffered
version of the reference clock.
The TLK2226 supports a selectable IEEE802.3z compliant 8b/10b encoder/decoder in all its 1.25Gbps modes of
operation, and a IEEE802 compliant 4b/5b encoder/decoder in its 125Mbps mode of operation.
The TLK2226 supports selectable RTBI and RGMII interface to supervision ASIC.
The TLK2226 automatically locks onto incoming data without the need to pre-lock.
Detection of whether the incoming data stream is at 1.25 Gbps or 125 Mbps data rate is automatic.
A comprehensive series of built-in tests for self-test purposes including loopback and PRBS generation and
verification is provided. An IEEE 1149.1 JTAG port is also supported to aid in board manufacturing test.
This device is housed in a small form factor 15 × 15 mm, 196-pin, BGA with 1,0 mm ball pitch and is
characterized to support the commercial temperature range of 0°C to 70°C.
Expect the TLK2226 to consume less than 1.5 W, when operating at 1.25 Gbps.
The TLK2226 is designed to be hot plug capable. A power-on reset puts the parallel side output signal pins in a
high-impedance state during power-up as well as pulls both TX+ and TX– to VDDA through 500 Ω.
Line Card
6
RCLK
RD..F[4:0]
System Backplane
6
TDA..F[4:0]
RX+
RX−
TX−
TX+
TLK2226
6
TCLK
MAC/
PACKET
PROCESSOR
TLK2226
TCLK
TDA..F[4:0]
RDA..F4:0]
RX−
RX+
RX+
RX−
Electrical to
Optical Array
RCLK
TX+
TX−
RCLK
RDA..F4:0]
6
6
6
SWITCH
FABRIC
TLK2226
6
TCLK
TDA..F[4:0]
TX+
TX−
Figure 1. TLK2226 System Implementation Diagram
2
Submit Documentation Feedback
6
TLK2226
www.ti.com
SLLS689 – JANUARY 2006
CTC
RX RGMII A
TX RGMII A
5bit DDR
to
10bit
SDR
Interface
+
TXA
Decode
PCS
Logic
_
Encode
channel A
TX Analog
Core
TX RGMII B
RX RGMII B
PCS
Logic
+
RXA
_
channel B
+
RXB
_
RX Analog
Core
channel B
Encode
5bit DDR
to
10bit
SDR
Interface
channel A
+
Decode
TXB
_
CTC
2 of the 6 channels of TLK2226 shown
REFCLK
data path for RGMII mode
Clock
Synthesizer
Common Control
logic
MDIO/MDC
Figure 2. TLK2226 Block Diagram
ABSOLUTE MAXIMUM RATINGS
over operating free-air temperature range (unless otherwise noted)
(1)
VALUE
UNIT
VDDQ
I/O Supply voltage (2)
–0.3 to 2.5
V
VDDS
I/O Supply voltage (2)
–0.3 to 3.6
V
–0.3 to 2.5
V
VDD , VDDA
VI
Core Supply
voltage (2)
LVCMOS
–0.5 to 3.6
HSTL
–0.5 to 2.5
DC Input voltage (I/O )
–0.3 to 2.5
V
Storage temperature
–65 to 85
°C
Input voltage
Electrostatic discharge
Characterized free-air operating temperature range
(1)
(2)
V
HBM: 2KV, CDM:750
V
0 to 70
°C
Stresses beyond those listed under absolute maximum ratings may cause permanent damage to the device. These are stress ratings
only, and functional operation of the device at these or any other conditions beyond those indicated under recommended operating
conditions is not implied. Exposure to absolute-maximum-rated conditions for extended periods may affect device reliability.
All voltage values, except differential I/O bus voltages, are with respect to network ground terminal
RECOMMENDED OPERATING CONDITIONS
VDD
VDDQ
VDDS
Core supply voltage
HSTL I/O supply voltage
LVCMOS I/O supply voltage
MIN
NOM
MAX
Peak-peak AC noise in the 1–10 MHz range may not
exceed 100 mV
1.7
1.8
1.9
V
1.5 V HSTL Class 1 Peak-peak AC noise may not
exceed 150 mV
1.4
1.5
1.6
V
1.8 V HSTL Class 1 Peak-peak AC noise may not
exceed 150 mV
1.7
1.8
1.9
V
1.8 V HSTL Class 1 Peak-peak AC noise may not
exceed 150 mV
1.7
1.8
1.9
V
2.5 V HSTL Class 1 Peak-peak AC noise may not
exceed 150 mV
2.375
2.5
2.625
V
3.3 V HSTL Class 1 Peak-peak AC noise may not
exceed 150 mV
3.14
3.3
3.47
V
Submit Documentation Feedback
UNIT
3
TLK2226
www.ti.com
SLLS689 – JANUARY 2006
RECOMMENDED OPERATING CONDITIONS (continued)
VDDA
Analog supply voltage
VREF
Input reference voltage (1)
IDD
Core supply current
I_HSTL
HSTL I/O supply current
IDDS
LVCMOS I/O supply current
MIN
NOM
MAX
Peak-peak AC noise in the 1–10 MHz range may not
exceed 100 mV
1.7
1.8
1.9
1.5 V HSTL Class 1
0.7
0.75
0.8
1.8 V HSTL Class 1
0.85
0.9
0.95
Rω = 125 MHz, VDD = 1.8 V
390
Rω = 125 MHz, VDDQ = 1.5 V
146
Rω = 125 MHz, VDDQ = 1.8 V
204
Rω = 125 MHz, VDDS = 1.8 V
1
Rω = 125 MHz, VDDS = 2.5 V
3
Rω = 125 MHz, VDDS = 3.3 V
7
IDDA
Analog supply current
Rω = 125 MHz, VDDA = 1.8 V
183
PD
Total power consumption
Rω = 125 MHz, VDDQ = 1.8 V
1.50
IDDQ
Shutdown current
ENABLE = Low
(1)
30
UNIT
V
V
mA
mA
mA
mA
W
mA
Typically, the value of VREF is expected to be 0.5 VDDQ of the transmitting device and VREF is expected to track variations in VDDQ.
LVCMOS ELECTRICAL CHARACTERISTICS (1)
over recommended operating conditions (unless otherwise noted)
PARAMETER
TEST CONDITIONS
VOH
High-level output voltage
IOH = –400 µA, VDDS = MIN
VOL
Low-level output voltage
IOL = 1 mA, VDDS = MIN
VIH
High-level input voltage
VIL
Low-level input voltage
MIN
TYP
VDDS– 0.2
0
0.25
MAX
UNIT
VDDS
V
0.5
V
VDDS = 1.8 V
1.4
VDDS +0.2
VDDS = 2.5 V
1.55
VDDS +0.2
VDDS = 3.3 V
2.0
VDDS +0.2
VDDS = 1.8 V
–0.2
0.63
VDDS = 2.5 V
–0.2
0.7
VDDS = 3.3 V
–0.2
1.4
V
V
IH
High input current
VDDS = MAX, VIN = 2.0 V
400
µA
IL
Low input current
VDDS = MAX, VIN = 2.0 V
–600
µA
CIN
Input capacitance
4
pF
(1)
Unused inputs that do not hold an integrated pull up or pull down circuit need to be terminated to either GND or VDDQ respectively to
avoid extensive currents and life time degradation.
HSTL ELECTRICAL CHARACTERISTICS
over recommended operating conditions (unless otherwise noted)
PARAMETER
TEST CONDITIONS
MAX
UNIT
VDDQ–0.4
MIN
VDDQ
V
0.4
V
VREF+0.10
VDDQ+0.3
V
–0.5
VREF–0.10
V
VOH(dc)
High-level output voltage
DC output, logic high
VOL(dc)
Low-level output voltage
DC output, logic low
VIH(dc)
High-level input voltage
DC input, logic high
VIL(dc)
Low-level input voltage
DC input, logic low
VIH(ac)
High-level input voltage
VIL(ac)
Low-level input voltage
IOH(dc)
High input current
VDDQ = 1.5 V
–8
IOL(dc)
Low input current
VDDQ = 1.5 V
8
CIN
Input capacitance
4
AC input, logic high, VDDQ = 1.5 V
VREF+0.20
AC input, logic high, VDDQ = 1.8 V
VREF+0.35
TYP
V
AC input, logic low, VDDQ = 1.5 V
VREF–0.20
AC input, logic low, VDDQ = 1.8 V
VREF–0.35
mA
mA
4
Submit Documentation Feedback
V
pF
TLK2226
www.ti.com
SLLS689 – JANUARY 2006
REFERENCE CLOCK TIMING REQUIREMENTS (REFCLK)
(1)
over recommended operating conditions (unless otherwise noted)
PARAMETER
Rω
Frequency
TEST CONDITIONS
NOM
MAX
UNIT
TYP–0.01%
125
TYP+0.01%
MHz
Maximum data rate
TYP–0.01%
125
TYP+0.01%
MHz
100
ppm
50%
60%
Accuracy
–100
Duty cycle
40%
Jitter
(1)
MIN
Minimum data rate
Random and deterministic
40
pspp
This clock should be crystal referenced to meet the requirements of this table ( Contact TI for specific clocking recommendations). Rate
activity of REFCLK is internally monitored. As soon as REFCLK is disconnected the device will initiate a global power-down mode; PLLs
will be turned off, output divers disabled and other current sources such as pull-up’s and pull down’s will be disabled to avoid power
consumption. To re-activate the device, restart REFCLK.
SERIAL TRANSMITTER/RECEIVER CHARACTERISTICS
over operating free-air temperature range (unless otherwise noted)
PARAMETER
VOD(p)
TX output voltage magnitude
TX Amp. Control 16.14 = 1
VOD(d)
VOD(p)
TX output voltage magnitude
TX Amp. Control 16.14 = 0
VOD(d)
VOD(pp)
VOD(pd)
TX output differential
peak-to-peak voltage swing
TX Amp. Control 16.14 = 1
MIN
TYP
MAX
UNIT
Pre-Emphasis 16.1:0 = 11. See Figure 3.
Differential and common-mode output
voltage definitions.
TEST CONDITIONS
650
950
1250
mV
Pre-Emphasis 16.1:0 = 10. See Figure 3.
Differential and common-mode output
voltage definitions.
625
900
1125
mV
Pre-Emphasis 16.1:0 = 01. See Figure 3.
Differential and common-mode output
voltage definitions.
575
850
1075
mV
Pre-Emphasis 16.1:0 = 00. See Figure 3.
Differential and common-mode output
voltage definitions.
550
800
1000
mV
Pre-Emphasis 16.1:0 = 11. See Figure 3.
Differential and common-mode output
voltage definitions.
Not
Supported
Not
Supported
Not
Supported
mV
Pre-Emphasis 16.1:0 = 10. See Figure 3.
Differential and common-mode output
voltage definitions.
Not
Supported
Not
Supported
Not
Supported
mV
Pre-Emphasis 16.1:0 = 01. See Figure 3.
Differential and common-mode output
voltage definitions.
375
500
675
mV
Pre-Emphasis 16.1:0 = 00. See Figure 3.
Differential and common-mode output
voltage definitions.
275
400
500
mV
Pre-Emphasis 16.1:0 = 11. See Figure 3.
Differential and common-mode output
voltage definitions.
1300
1900
2500
mVp-p
Pre-Emphasis 16.1:0 = 10. See Figure 3.
Differential and common-mode output
voltage definitions.
1250
1800
2250
mVp-p
Pre-Emphasis 16.1:0 = 01. See Figure 3.
Differential and common-mode output
voltage definitions.
1150
1700
2150
mVp-p
Pre-Emphasis 16.1:0 = 00. See Figure 3.
Differential and common-mode output
voltage definitions.
1100
1600
2000
mVp-p
Submit Documentation Feedback
5
TLK2226
www.ti.com
SLLS689 – JANUARY 2006
SERIAL TRANSMITTER/RECEIVER CHARACTERISTICS (continued)
over operating free-air temperature range (unless otherwise noted)
PARAMETER
VOD(pp)
TX output differential
peak-to-peak voltage swing
TX Amp. Control 16.14 = 0
VOD(pd)
TEST CONDITIONS
MIN
TYP
MAX
UNIT
Pre-Emphasis 16.1:0 = 11. See Figure 3.
Differential and common-mode output
voltage definitions.
Not
Supported
Not
Supported
Not
Supported
mVp-p
Pre-Emphasis 16.1:0 = 10. See Figure 3.
Differential and common-mode output
voltage definitions.
Not
Supported
Not
Supported
Not
Supported
mVp-p
Pre-Emphasis 16.1:0 = 01. See Figure 3.
Differential and common-mode output
voltage definitions.
750
1000
1350
mVp-p
Pre-Emphasis 16.1:0 = 00. See Figure 3.
Differential and common-mode output
voltage definitions.
550
800
1000
mVp-p
750
1200
mV
200
900
mV
400
2500
mVp-p
–10
10
µA
2
pF
150
220
ps
165
260
ps
0.75
UI (1)
0.462
UI
0.24
UI
0.10
UI
192
Bits
136
Bits
32
Bits
14
Bits
VCMT
TX output common mode
voltage range
See Figure 3. Differential and
common-mode output voltage definitions.
VID
RX input voltage magnitude
away from common mode
See Figure 5
VID(p)
RX input differential
peak-to-peak voltage swing
See Figure 5
ILKG
RX input leakage current
CI
RX input capacitance
RL = 50 Ω, CL = 5 pF, See Figure 3
tr, tf
Differential output signal rise,
RL = 50 Ω, CL = 5 pF, See Figure 3 For Half
fall time (20% to 80%)
amplitude condition without Pre-Emphasis
JTOL
Jitter Tolerance, Total Jitter
at Serial Input
Zero crossing, See Figure 6
JDR
Serial Input Deterministic
Jitter
Zero crossing, See Figure 6
JT
Serial Output Total Jitter
Alternating disparity
K28.5 1.25 Gbps
JD
Serial Output Deterministic
Jitter
Alternating disparity K28.5 1,25 Gbps
From MDI /T/ to RGMII
R(LATENCY) RX_DV
de-assert
T(LATENCY)
From the assertion of TX_EN
to MDI output of /S/
R(LATENCY)
MDI Input to RX_DV
de-assert
T(LATENCY)
(1)
6
TX_EN sampled to MDI
output
0.15
1000Base-X Operating Mode
100Base-FX Operating Mode
Unit Interval = one serial bit time (min 800 ps)
Submit Documentation Feedback
TLK2226
www.ti.com
SLLS689 – JANUARY 2006
VOD(p)
VOD(d)
VCMT
VOD(pd) VOD(pp)
VOD(d)
VOD(p)
bit
time
tr , tf
Figure 3. Differential and Common-Mode Output Voltage Definitions
Differential Amplitude
(mV)
+1000
+550
Transmit
0
−550
−1000
.66 .86 1.0
.14 .34
Normalized Time
Figure 4. Transmit Template
Differential Amplitude
(mV)
+1000
+200
0
Receive
−200
−1000
0
0.7
0.3
Normalized Time
1.0
Figure 5. Receive Template
J DR
JR
JR
J TOL
NOTE: JTOL = JR + JDR, where JTOL is the receive jitter tolerance, JDR is the received deterministic jitter, and JR is the
Gaussian random edge jitter distribution at a maximum BER = 10-12.
Figure 6. Input Jitter
Submit Documentation Feedback
7
TLK2226
www.ti.com
SLLS689 – JANUARY 2006
HSTL OUTPUT SWITCHING CHARACTERISTICS
over recommended operating conditions (unless otherwise noted)
PARAMETER
TEST CONDITION
MIN
NOM
MAX
UNIT
Tr
Clock and data rise time
80% to 20% output voltage, see Figure 7,
C = 10 pF
1.5
ns
Tf
Clock and data fall time
80% to 20% output voltage, see Figure 7,
C = 10 pF
1.8
ns
TOSU
RD[4:0] valid (setup) prior to RCLK
transition high or low
See Figure 7, timing relative to 0.5×VDDQ,
25 MHz and 125 MHz
1.2
ns
TOHD
RD[4:0] valid (hold) after RCLK transition
high or low
See Figure 7, timing relative to 0.5×VDDQ,
25 MHz and 125 MHz
1.2
ns
Tsuav
RCLK transition high or low to data valid
window beginning (RD[4:0]).
See Figure 8, timing relative to 0.5×VDDQ, 25
MHz and 125 MHz
Thdav
RCLK transition high or low to data valid
window ending (RD[4:0]).
See Figure 8 , timing relative to 0.5×VDDQ,
25 MHz and 125 MHz
800
3.2
ps
ns
VDDQ/2
RCLK[A−F]
tr
tf
t OSU
t OHD
LSB Nibble
RDx[4:0]
t OSU
t OHD
MSB Nibble
Figure 7. HSTL Receive Output Timing Requirements (Centered)
Thdav
VDDQ/2
RCLK[A−F]
Thdav
Tsuav
LSB Nibble
RDx[4:0]
Tsuav
MSB Nibble
Figure 8. HSTL Receive Output Timing Requirements (Aligned)
HSTL INPUT TIMING REQUIREMENTS
over recommended operating conditions (unless otherwise noted)
PARAMETER
TEST CONDITION
MIN
TYP (1)
MAX
UNIT
TISU
TDx[4:0] setup prior to TCLK transition
high or low
Figure 19 timing relative to 0.5×VDDQ,
25 MHz and 125 MHz
0.6
ns
TIHD
TDx[4:0] hold after TCLK transition high or
low
Figure 19, timing relative to 0.5×VDDQ,
25 MHz and 125 MHz
0.6
ns
Tsuav
RCLK transition high or low to data valid
window beginning (RD[4:0]).
Figure 18, timing relative to 0.5×VDDQ,
25 MHz and 125 MHz
Thdav
RCLK transition high or low to data valid
window ending (RD[4:0]).
Figure 18, timing relative to 0.5×VDDQ,
25 MHz and 125 MHz
(1)
8
All typical values are at 25°C and with a nominal supply.
Submit Documentation Feedback
1.4
2.6
ns
ns
TLK2226
www.ti.com
SLLS689 – JANUARY 2006
VDDQ/2
TCLKx
tr
TDx[4:0]
tf
t ISU
t IHD
LSB Nibble
t ISU
t IHD
MSB Nibble
Figure 9. HSTL Data Input Timing Requirements (Centered)
Thdav
VDDQ/2
TCLK[A−F]
Thdav
Tsuav
TDx[4:0]
LSB Nibble
Tsuav
MSB Nibble
Figure 10. HSTL Data Input Timing Requirements (Aligned)
Submit Documentation Feedback
9
TLK2226
www.ti.com
SLLS689 – JANUARY 2006
MDIO TIMING REQUIREMENTS
over recommended operating conditions (unless otherwise noted)
PARAMETER
TEST CONDITIONS
MIN
tperiod
MDC period
See Figure 11
tsetup
MDIO setup to ↑ MDC
See Figure 11
10
thold
MDIO hold to ↑ MDC
See Figure 11
10
Tdata
MDC ↑ to MDIO valid
NOM
0
MDC
t PERIOD
t SETUP
t HOLD
MDIO
Figure 11. MDIO Read/Write Timing
TXP
RXP
50 ohm transmission line
VDDA
50
4.2K
6.3K
50
GND
50 ohm transmission line
RXN
TXN
TRANSMITTER
MEDIA
RECEIVER
Figure 12. Example High Speed I/O AC Coupled Mode
10
Submit Documentation Feedback
MAX
UNIT
500
ns
ns
ns
300
ns
TLK2226
www.ti.com
SLLS689 – JANUARY 2006
A
B
C
D
E
F
G
H
J
K
L
M
N
P
14
GNDA
GNDA
GNDA
GNDA
GNDA
GNDA
GNDA
GNDA
GNDA
GNDA
GNDA
GNDA
GNDA
GNDA
14
13
TXA+
RXA+
RXB+
TXB+
GNDA
TXC+
RXC+
RXD+
TXD+
GNDA
TXE+
RXE+
RXF+
TXF+
13
12
TXA−
RXA−
RXB−
TXB−
VDDA
TXC−
RXC−
RXD−
TXD−
VDDA
TXE−
RXE−
RXF−
TXF−
12
11
VDDA
VDDA
VDDA
VDDA
VDDA
VDDA
VDD
VDD
VDDA
VDDA
VDDA
VDDA
VDDA
VDDA
11
10
TDA0
TDA1
TDA2
TDA3
RESET
VDD
VDD
VDD
VDD
SYNC
EN
TDF3
TDF2
TDF1
TDF0
10
9
TCLKA
VDDQ
TDA4
RDA4
ENABLE
GND
GND
GND
GND
DVAD4
RDF4
TDF4
VDDQ
TCLKF
9
8
RDA0
GND
RDA3
VDDQ
TCLK
SEL
GND
GND
GND
GND
DVAD3
VDDQ
RDF3
GND
RDF0
8
7
RCLKA
RDA1
RDA2
GND
SGMII
GND
GND
GND
GND
SWAP
GND
RDF2
RDF1
RCLKF
7
6
TDB0
VDDQ
TDB1
TDB2
CODE
GND
GND
GND
GND
REPEA
TB
TDE2
TDE1
VDDQ
TDE0
6
5
TCLKB
GND
TDB3
VDDQ
LPBK
PRBSE
N
TRSTN
CTC_E
N
RSVD
REPEA
TA
VDDQ
TDE3
GND
TCLKE
5
4
RDB0
RDB1
TDB4
GND
RDC2
RDC4
TDI
TDO
RDD4
RDD2
GND
TDE4
RDE1
RDE0
4
3
RCLKB
VDDQ
RDB2
RDB3
RDC1
RDC3
TCLK
TMS
RDD3
RDD1
RDE3
RDE2
VDDQ
RCLKE
3
2
RDB4
GND
TDC2
TDC4
RDC0
MDIO
VREF
GND
VDDS
RDD0
TDD4
TDD2
GND
RDE4
2
1
TDC0
TDC1
TCLKC
TDC3
RCLKC
MDC
VDDS
REFCL
K
GND
RCLKD
TDD3
TCLKD
TDD1
TDD0
1
A
B
C
D
E
F
G
H
J
K
L
M
N
P
NOTE: Unused inputs that do not hold an integrated pull up or pull down circuit need to be terminated to either GND or
VDDQ respectively to avoid extensive currents and life time degradation.
Figure 13. Terminal Diagram
Submit Documentation Feedback
11
TLK2226
www.ti.com
SLLS689 – JANUARY 2006
SIGNAL TERMINAL DESCRIPTION
SIGNAL
LOCATION
TYPE
DESCRIPTION
Serial I/O Signals
TXA+, TXA–
TXB+, TXB–
TXC+, TXC–
TXD+, TXD–
TXE+, TXE–
TXF+, TXF–
A13, A12,
D13, D12,
F13, F12,
J13, J12,
L13, L12,
P13, P12
VML Output
Differential Output Transmit. TX[A-F]+ and TX[A-F]- are differential serial outputs
that interface to a copper or an optical I/F module. TX[A-F]+ and TX[A-F]- are pulled
to VDDA through 500 ohms when LPBK = high or when the LOOPBACK (deep only)
bit for a particular channel in the MDIO registers is set.
RXA+, RXA–
RXB+, RXB–
RXC+, RXC–
RXD+, RXD–
RXE+, RXE–
RXF+, RXF–
B13, B12,
C13, C12,
G13, G12,
H13, H12,
M13, M12,
N13, N12
PECLcompatible
Input
Differential input receive. RX[A-F]+ and RX[A-F]- are the differential serial input
interface from a copper or an optical I/F module. Differential resistive termination is
built-in for these terminals.
Transmit Data Bus and Clock Signals
REFCLK
H1
LVCMOS Input
Reference Clock. REFCLK is an external input clock that provides the clock
reference.
TCLKA,
TCLKB,
TCLKC,
TCLKD,
TCLKE,
TCLKF
A9,
A5,
C1,
M1,
P5,
P9
HSTL Input with Transmit Data Clock Channels A, B, C, D, E, and F.
P/U
RTBI/RGMII mode:
When in Independent Transmit channel mode (TCLKSEL = 1) these pins are
used to latch data for their respective channels on both the rising and falling
edges. TCLKA latches data for channel A. TCLKB latches data for channel
B. TCLKC latches data for channel C. TCLKD latches data for channel D.
TCLKE latches data for channel E. TCLKF latches data for channel F.
When in synchronous transmit channel mode (TCLKSEL = 0) only TCLKC is
used as the transmit clock for all 6 parallel transmit busses.
TDA[4:0]
C9, D10, C10,
B10, A10
HSTL Input with Transmit Data Channel A.
P/U
RTBI/RGMII mode:
The parallel data is clocked into the transceiver on the rising and falling
edge of TCLKx and transmitted as a serial stream with TDx0 rising edge
data sent as the first bit. Data is input low-order bits first aligned to the rising
edge of TCLKx following by high order bits aligned to the falling edge.
TDB[4:0]
C4, C5,
D6, C6, A6
HSTL Input with Transmit Data Channel B.
P/U
TDC[4:0]
D2, D1, C2, B1,
A1
HSTL Input with Transmit Data Channel C.
P/U
TDD[4:0]
L2, L1, M2, N1,
P1
HSTL Input with Transmit Data Channel D.
P/U
TDE[4:0]
M4, M5, L6, M6, HSTL Input with Transmit Data Channel E.
P6
P/U
TDF[4:0]
M9, L10, M10,
N10, P10
HSTL Input with Transmit Data Channel F
P/U
Receive Data Bus and Clock Signals
RCLKA
RCLKB
RCLKC
RCLKD
RCLKE
RCLKF
A7,
A3,
E1,
K1,
P3,
P7
HSTL Output
Individual Receive Byte Clock Channels A through F.
RTBI/RGMII mode:
Recovered byte clock for channels A through F. These clocks are used by
the protocol device to latch the received data output for channels A through
F. Data is aligned to both the rising and falling edges.
If the internal CTC FIFO is disabled for a channel, the clock for that channel
is 1/10th the clock recovered from the incoming data stream. If CTC is
enabled for a channel, the clock for that channel is a buffered version of
REFCLK.
These pins are internally series terminated to provide direct connection to a
50-Ω transmission line.
12
Submit Documentation Feedback
TLK2226
www.ti.com
SLLS689 – JANUARY 2006
SIGNAL TERMINAL DESCRIPTION (continued)
SIGNAL
LOCATION
TYPE
DESCRIPTION
RDA[4:0]
D9, C8, C7, B7,
A8
HSTL Output
Receive Data Channel A.
RTBI/RGMII mode:
Output data is aligned to the rising and falling edge of the RCLKA. Channel
data is output low-order bits first aligned to the rising edge of RCLKA
followed by high order bits aligned to the falling edge. These pins are
internally series terminated to provide direct connection to a 50-Ω
transmission line.
RDB[4:0]
A2, D3, C3, B4,
A4
HSTL Output
Receive Data Channel B
RDC[4:0]
F4, F3, E4, E3,
E2
HSTL Output
Receive Data Channel C
RDD[4:0]
J4, J3, K4, K3,
K2
HSTL Output
Receive Data Channel D
RDE[4:0]
P2, L3, M3, N4,
P4
HSTL Output
Receive Data Channel E
RDF[4:0]
L9, M8, M7, N7, HSTL Output
P8
Receive Data Channel F
Management Data Interface Signals
MDIO
F2
LVCMOS I/O
with P/U
Management Data I/O. MDIO is the bi-directional serial data path for the transfer of
management data to and from the protocol device.
MDC
F1
LVCMOS with
P/U
Input Management Data Clock. MDC is the clock reference for the transfer of
management data to and from the protocol device.
DVAD[4-3]
K9, K8
LVCMOS Input
with P/D
Management Address. Device Address: DVAD[4:3] is the externally set physical
address given to this device used to distinguish one device from another.
JTAG Interface Signals
TCK
G3
LVCMOS input
Test Clock. IEEE1149.1 (JTAG) TCK is used to clock state information and test data
into and out of the device during the operation of the test port.
TDI
G4
LVCMOS Input
with P/U
Test Data Input. IEEE1149.1 (JTAG) TDI is used to serially shift test data and test
instructions into the device during the operation of the test port.
TDO
H4
LVCMOS
Output
Test Data Output. IEEE1149.1 (JTAG) TDO is used to serially shift test data and
test instructions out of the device during operation of the test port. When the JTAG
port is not in use, TDO is in a high impedance state.
TMS
H3
LVCMOS Input
with P/U
Test Mode Select. IEEE1149.1 (JTAG) TMS is used to control the state of the
internal test-port controller.
TRSTN
G5
LVCMOS Input
with P/U
JTAG Reset. IEEE1149.1 (JTAG) TRSTN is used to reset the internal JTAG
controller.
Miscellaneous Signals
CODE
E6
LVCMOS Input
with P/U
PCS Functions Enable. When high, the Physical Coding Sublayer (PCS) functions
are enabled. This pin is logically AND'd with MDIO register PCS_EN.
ENABLE
E9
LVCMOS Input
Device Enable. Pulling this pin high will enable all outputs of the device. A low on
this pin will tri-state all outputs of the device, and causes the serial outputs to be
pulled to VDDA through 500 Ω. This signal must be externally held high during JTAG
manufacturing test on the customer application board. See Appendix A for more
details.
CTC_EN
H5
LVCMOS Input
with P/U
CTC Enable. This pin is AND'd with the CTC register bit. Setting this pin to logic 1
activates the CTC FIFO. A low on this pin disables the clock tolerance compensation
function. This function should only be activated in transceiver mode (CODE=1).
TCLKSEL
E8
LVCMOS Input
with P/D
Independent Channel Transmit Clock Select. A logic 0 selects latching of all
channels synchronous to TCLKC. A logic 1 selects individual clocking. This pin bit is
logically OR'd with the TCLKSEL input MDIO register bit.
RESET
E10
LVCMOS Input
with P/D
Chip Reset (FIFO Clear) Pulling this terminal high re-centers the transmit skew
buffers, receive channel synchronization FIFO's, and will reset MDIO flags to default
settings. Reset should be asserted for 100 µs after all power supplies and input clock
frequencies are within specification. Alternatively, a soft reset could be issued after
clock and power stabilization.
LPBK
E5
LVCMOS Input
with P/D
Loopback Enable. When set high, TX+/– outputs are pulled to VDDA through 500
ohms. The serial TX+/– output data is internally looped to RX+/– inputs. External
data applied to RX+/– inputs is ignored.
Submit Documentation Feedback
13
TLK2226
www.ti.com
SLLS689 – JANUARY 2006
SIGNAL TERMINAL DESCRIPTION (continued)
SIGNAL
LOCATION
TYPE
DESCRIPTION
PRBSEN
F5
LVCMOS Input
with P/D
PRBS Enable. When this terminal is asserted high, the pseudo-random bit stream
generator and comparator circuits are inserted into the transmit and receive data
paths on all channels, respectively.
If this pin is not used it can be tied to the GND reference. TX+– are transmitting 27-1
PRBS bit streams. RX+– are comparing incoming data to 27-1 PRBS bit stream.
Results of the RX comparison can be read from the MDIO.
SYNCEN
K10
LVCMOS Input
with P/U
SYNC Enable. When enabled, the TLK2226 byte-aligns incoming data on RX+/– on
positive RD (running disparity) commas (001 1111b). This pin is AND'd with the
COMM_DET register bit.
SWAP
K7
LVCMOS Input
with P/D
Parallel Port Swap. When enabled, the parallel port for channel B is routed to the
serial port for channel A, and the parallel port for channel A is routed to the serial
port for channel B. Likewise, the parallel ports for channels C and D, E and F are
swapped. This pin is logically OR'd with the PORT_SWAP register bits.
REPEATA
K5
LVCMOS Input
with P/D
Repeater Mode Ports A, C, E. When enabled, the serial data on the channel B
receiver pins is repeated out on the channel A serial transmit pins. Likewise, Channel
D serial receive data is repeated out on channel C serial transmit pins, and Channel
F serial receive data is repeated out on channel E serial transmit pins.
This pin is logically OR'd with the REPEATER register bits for channels A, C, and E.
REPEATB
K6
LVCMOS Input
with P/D
Repeater Mode Ports B, D, F. When enabled, the serial data on the channel A
receiver pins is repeated out on the channel B serial transmit pins. Likewise, Channel
C serial receive data is repeated out on channel D serial transmit pins, and Channel
E serial receive data is repeated out on channel F serial transmit pins.
This pin is logically OR’d with the REPEATER register bits for channels B, D, and F.
RSVD
J5
RESERVED
Reserved: This signal is reserved for TI internal testing. These should be connected
to GND in customer applications.
SGMII
E7
LVCMOS Input
with P/D
SGMII Mode When asserted, enables SGMII mode. When low, SGMII mode is
deasserted. This pin sets the default value of the SGMII register bit (16.4) on the
falling (deasserting) edge of the RESET pin.
Power and Reference Pins
VDDQ
B3, B6, B9, D5,
D8, L5, L8, N3,
N6, N9
Supply
HSTL I/O Supply Voltage. 1.5 V or 1.8 V
VDD
F10, G10, G11,
H10, H11, J10
Supply
Digital Logic Power: Provides power for all digital circuitry. Nominally 1.8 V.
VDDS
G1, J2
Supply
Secondary voltage LVCMOS I/O Supply Voltage 1.8 V OR 2.5 V OR 3.3 V
VDDA
A11, B11, C11,
D11, E11, E12,
F11, J11, K11,
K12, L11,M11,
N11, P11
Supply
Analog Power. VDDA provides a supply reference for the high-speed analog
circuits, receiver and transmitter. Nominally 1.8 V.
VREF
G2
Reference
Voltage Reference– for HSTL reference
GNDA
A14, B14, C14,
D14, E13, E14,
F14, G14, H14,
J14, K13, K14,
L14, M14, N14,
P14
Ground
Analog Ground. GNDA provides a ground reference for the high-speed analog
circuits, RX and TX.
GND
B2, B5, B8, D4, Ground
D7, F6, F7, F8,
F9, G6, G7, G8,
G9, H2, H6, H7,
H8, H9, J1, J6,
J7, J8, J9, L4,
L7, N2, N5, N8
14
Digital Logic Ground. Provides a ground for the logic circuits and digital I/O buffers.
Submit Documentation Feedback
TLK2226
www.ti.com
SLLS689 – JANUARY 2006
DETAILED DESCRIPTION
REFERENCE CLOCK SYNTHESIZER (PLL)
The TLK2226 employs a mature Phase Lock Loop (PLL) design in use for Gigabit Ethernet transceivers and
high-speed serial links by Texas Instruments since 1997 on both standard products and custom ASIC designs.
This PLL design is used to synthesize the serial line rate bit clock from the REFCLK input as well as generate
clocks for the receiver sampling circuitry. The PLL and associated high speed circuitry is powered by the analog
power supply pins (VDDA) with isolated grounds (GNDA). Care should be taken in providing a low noise
environment in a system. It is recommended to supply the VDDA reference by separate isolated plane within the
system printed circuit board (PCB). It is recommended that systems employing switching power supplies provide
proper filtering of the fundamental and harmonic components in the 1MHz-10Mhz band to avoid bit errors from
injected noise. It is strongly recommended that no PLL based clock synthesizer circuit be used as the source for
the REFCLK. This could cause accumulation of jitter between the two PLL's
OPERATING MODES
The TLK2226 has two operational modes with respect to the Physical Coding Sublayer (PCS) functions called
RTBI mode and RGMII mode. The RTBI or RGMII mode is selectable via the CODE bit in MDIO register space
or the CODE pin as described in Table 1. The CODE pin is gated with the internal CODE register such that
setting the pin to a logic zero disables the PCS functions, setting the CODE pin to a logic one enables PCS
functions and permits these functions to be disabled on a channel by channel basis via MDIO control. In addition,
all PCS functions except the CTC FIFO may be enabled if the CTC FIFO is to be implemented in the protocol
device. The CTC FIFO may be disabled by way of the CTC_EN pin or the CTC_EN bit in the MDIO register
space. Besides the RTBI or RGMII modes, additional modes of operation include the serial repeater mode with
CTC FIFO mode, serial repeater without CTC FIFO mode, RGMII port swap mode, and 1/10 rate for 100FX
mode.
Table 1. PCS Operational Modes
CODE
OPERATING MODES
Low
RTBI mode. Disables PCS functions for each channel. Refer to Byte Alignment Logic section, for additional description over
this mode.
High
RGMII mode. Enables PCS functions for each channel. Data on the transmit and receive data bus is treated as un-coded data.
RTBI MODE
In RTBI mode, the transmit data bus for each channel accepts 5-bit wide encoded data on the transmit data bus
pins. Data is latched on the rising and falling edge of the transmit data clock. The encoded data is then
phase-aligned, serialized and transmitted sequentially beginning with bit 0 over the differential high speed serial
transmit pins. The receive data bus for each channel outputs 5-bit wide data. Data is output relative to both the
rising and falling edge of the receive clock.
CTC
RX RTBI A
to
TX RTBI A
+
_
5bit DDR
PCS
Logic
10bit
SDR
Interface
TXA
Decode
Encode
channel A
TX Analog
Core
TX RTBI B
5bit DDR
RX RTBI B
10bit
SDR
Interface
Encode
to
PCS
Logic
channel A
+
RXA
_
channel B
+
RXB
_
RX Analog
Core
channel B
Decode
+
_
TXB
CTC
2 of the 6 channels of TLK2226 shown
data path for RTBI mode
Figure 14. RTBI Mode Block Diagram
Submit Documentation Feedback
15
TLK2226
www.ti.com
SLLS689 – JANUARY 2006
RGMII MODE
In RGMII mode, the transmit data bus for each channel accepts 4-bit wide un-coded parallel data plus the two
GMII control bits TX_EN and TX_ER (on bit4 of TDx[4:0]). Data is sampled on the rising and falling edge of the
transmit clock. The data is first aligned to the reference clock (REFCLK), then 8B/10B encoded and passed to
the serializer. The receive data bus for each channel outputs un-coded 4-bit wide parallel data plus the two GMII
status bits RX_DV and RX_ER (on bit4 of RDx[4:0]).
CTC
RX RGMII A
to
TX RGMII A
+
_
5bit DDR
10bit
SDR
Interface
TXA
Decode
PCS
Logic
Encode
channel A
TX Analog
Core
TX RGMII B
5bit DDR
RX RGMII B
10bit
SDR
Interface
Encode
to
PCS
Logic
channel A
+
RXA
_
channel B
+
RXB
_
RX Analog
Core
channel B
Decode
+
_
TXB
CTC
2 of the 6 channels of TLK2226 shown
data path for RGMII mode
Figure 15. RGMII Mode Block Diagram
In RGMII mode, all of the PCS functions are implemented as required by the IEEE802.3 specification, including
the 8b/10b encoding, decoding, Clock Tolerance Compensation, insertion of the Start of Frame and End of
Frame codes, and the auto-negotiation state machine. In RGMII mode, it is also possible to bypass the CTC
FIFOs in which case the received data will be output with a recovered byte clock that remains in the clock
domain of the received serial data stream.
SERIAL REPEATER MODE (WITHOUT CTC)
The serial repeater mode operation can be selected on a per channel basis. The serial inputs on one channel are
transmitted back out serially on the adjacent channel. Channel A receive data is repeated out on channel B
transmit while channel B receive data is repeated out on channel A transmit. In this manner channels A and B
can form a full-duplex repeater. Likewise, channels C and D, and E and F can be used to form full-duplex
repeaters.
Repeater mode for channels A and B can be chosen separately, if desired. That is, serial data received on
channel A can be repeated out on channel B transmit pins without also choosing to repeat channel B serial
receive data out on channel A. In fact, the repeater mode selection bit for a given channel may be thought of as
simply selecting whether the data to be transmitted for that port comes from the parallel port for that channel or
the serial data from the adjacent channel. In this way, a channel may be used as a normal RGMII or RTBI
channel while the incoming serial data is also repeated out on the adjacent channel. When repeater mode for a
given channel is selected, any data presented to the parallel transmit bus for that channel will be ignored.
While in repeater mode, the incoming serial data will still go through the deserializer function, and the PCS
functions if enabled, and the deserialized data will still be presented out on the parallel transmit bus. Since the
Clock Tolerance compensation functions are disabled in this mode, the data presented on the parallel port will
reflect the CTC function being disabled as well.
Repeater mode without CTC is chosen by setting the repeater mode register bit active while also setting the
CTC_EN register bit inactive at the same time.
16
Submit Documentation Feedback
TLK2226
www.ti.com
SLLS689 – JANUARY 2006
CTC
RX RGMII A
to
TX RGMII A
+
_
5bit DDR
PCS
Logic
10bit
SDR
Interface
TXA
Decode
Encode
channel A
TX Analog
Core
TX RGMII B
5bit DDR
RX RGMII B
10bit
SDR
Interface
to
PCS
Logic
+
RXA
_
channel B
+
RXB
_
RX Analog
Core
channel B
Encode
channel A
+
_
Decode
TXB
CTC
2 of the 6 channels of TLK2226 shown
data path for Serial Repeater mode with out retiming
Figure 16. Serial Repeater Mode Block Diagram
In serial repeater mode, a clock/data recovery function (CDR) locks to the incoming data stream and then sends
that data out on the transmit port of the adjacent channel. The CDR function will lock to the data stream so that
Inter-Symbol Interference (ISI) jitter is filtered out from the retransmitted data. Retransmitted data will remain in
the clock domain of the incoming data stream, since there is no clock tolerance compensation FIFO to adjust the
data stream to the REFCLK clock domain. Since the CDR function will track any frequency wander on the data
stream that is lower than the bandwidth of the CDR, any such wander components will remain in the
retransmitted data stream. Also, since the jitter tolerance specification for the CDR function is such that the
recovered bit clock must work with a minimum eye opening of 0.2513UI (equal to 201ps at 1.25GB/s), and the
jitter generation specification for a transmitted data stream is 0.24UI (equal to 192 ps at 1.25Gb/s), it is not
guaranteed that the re-transmitted serial data will meet the IEEE802.3 jitter generation specifications.
SERIAL RETIMING REPEATER MODE (WITH CTC)
The TLK2226 also provides another serial repeater mode serial repeater mode except the data stream is
deserialized and passed through a Clock Tolerance Compensation FIFO. The CTC FIFO will add or drop IDLE
codes as needed (or CONFIG codes) to match the data stream to the REFCLK clock domain. The data stream is
then retransmitted out the TX of the adjacent channel. In serial retiming repeater mode the CTC FIFO ensures
that any frequency wander passed through the CDR function will not appear in the transmit data stream. The
retransmitted data can be guaranteed to meet the transmit jitter generation specifications for IEEE802.3.
Repeater mode with CTC is chosen by setting the repeater mode register bit active while also setting the
CTC_EN register bit active at the same time. As in serial repeater mode, the repeater path from channel A to
channel B may be selected separately from the repeater path for channel B to channel A.
CTC
RX RGMII A
to
TX RGMII A
+
_
5bit DDR
PCS
Logic
10bit
SDR
Interface
TXA
Decode
Encode
channel A
TX Analog
Core
TX RGMII B
Encode
5bit DDR
to
RX RGMII B
PCS
Logic
10bit
SDR
Interface
channel A
+
RXA
_
channel B
+
RXB
_
RX Analog
Core
channel B
Decode
+
_
TXB
CTC
2 of the 6 channels of TLK2226 shown
data path for Serial Repeater mode with CTC
Figure 17. Serial Retiming Repeater Mode Block Diagram
Submit Documentation Feedback
17
TLK2226
www.ti.com
SLLS689 – JANUARY 2006
RGMII PORT-SWAP MODE
The TLK2226 provides a set of multiplexors on the parallel busses between pairs of channels such that the
RGMII port for Channel A can be steered to the serial port for Channel B, while at the same time the RGMII port
for Channel B can be steered to the serial port for Channel A. See Figure 18 for the block diagram illustrating this
mode. Likewise, the RGMII ports for channels C and D may be swapped, as may the RGMII ports for Channels
E and F,
The selection of swapping a pair of RGMII ports is accomplished by setting the appropriate register bit in MDIO
register space.
RX RGMII A
5bit DDR
TX RGMII A
10bit
SDR
Interface
to
PCS
Logic
+
_
CTC
function
and
8b10
coding
TXA
channel A
TX Analog
Core
5bit DDR
RX RGMII B
10bit
SDR
Interface
to
PCS
Logic
+
RXA
_
channel B
+
RXB
_
RX Analog
Core
channel B
TX RGMII B
channel A
CTC
function
and
8b10
coding
+
_
TXB
2 of the 6 channels of TLK2226 shown
data path for RGMII mode (with port swapping)
Figure 18. RGMII Mode with Port Swapping
100FX 125Mbps MODE
The TLK2226 can be set to operate at 125 Mbps on a per channel basis. While in 100FX mode, the PCS
functions are enabled for RGMII mode, but the transmit and recovered byte clocks are at 25 MHz. When a
channel is in 100FX mode the PCS functions for that channel will comply with the IEEE802.3 100 base-x
specifications, including the 4b/5b coding, NRZI Encoding, and nibble alignment.
While a channel is in 100FX mode, the reference clock for that channel will still be the 125MHz reference clock
input for the TLK2226. For this reason, the transmit clock for that channel’s RGMII port must be a 25MHz
derivative of the reference clock without any ppm frequency offset. The recovered byte clock for that channel’s
RGMII port will be divided down to 25 MHz.
Auto rate select for a channel will operate primarily off of transition density of the incoming serial data stream.
The Clock and Data Recovery (CDR) function of the channel receiver will always operate at the 1.25Gbps rate. If
the incoming data stream is properly coded 8b/10b data for 1.25Gbs data then the transition density of the
incoming data will be detected and that channel will operate at 1.25Gbps. If properly 4B/5B and NRZI Encoded
data for 100FX operation received, then the incoming transition density will be detected as a maximum 1
transition per 10bit input word, and that channel will operate at 125 Mbps. When a channel is operating at
125Mbps, the actual SERDES front end will still be operating at 10 times that rate by transmitting every 100FX
bit 10 times in a row and by over-sampling every incoming 100FX bit by a factor of 10.
SGMII (Serial GMII) (1000 Mbps/100 Mbps/ 10 Mbps) MODE
The TLK2226 can be set to operate in SGMII mode (register bit 16.4). In this mode, the serial side of TLK2226
always operates at 1.25 Gbps. The parallel interface speed is directly selected through mdio register bits 0.6 and
0.13. In 10 Mbps mode, the parallel interface runs at 2.5 MHz. In 100 Mbps mode, the parallel interface runs at
25 MHz. In 1000 Mbps mode, the parallel interface runs at 125 MHz. In all three modes, the slightly modified (per
the SGMII document) auto-negotiation state machine (if enabled) is active.
In 1000 Mbps SGMII mode the parallel interface operates identical to the 1000Base-X case with SGMII disabled.
In both 10 and 100 Mbps SGMII modes, the parallel data is transferred on the parallel interface at the lower
provisioned rate, and replicated either 10 or 100 times for transmission (and deleted accordingly in the receive
direction) by the PCS layer out the serial interface. Note that in both the transmit and receive directions for both
18
Submit Documentation Feedback
TLK2226
www.ti.com
SLLS689 – JANUARY 2006
10 and 100 Mbps SGMII modes, the nibble transferred on the rising edge is repeated on the negative edge of
the parallel interface clock. For a given packet byte, the lower nibble is placed on the parallel interface during the
first clock period (the data is duplicated on falling edge), followed by the upper nibble of the packet byte on the
next clock period (data is duplicated on the falling edge). Also note that the control bits for the byte are
duplicated according to the RGMII standard across the two clock periods.
All test related and data path functional modes for TLK2226 are capable of operating in the SGMII mode.
Specifically, all three loopback modes, transceiver, port swap, and all repeater modes operate properly.
PARALLEL INTERFACE MODES
The TLK2226 provides an RTBI compliant interface or an RGMII compliant interface. Each parallel channel
provides 5 bit busses clocked in Double Data Rate (DDR) fashion. The five least significant bits of data are
latched on the rising edge of the transmit clock are put together with the 5 most significant bits of data latched on
the falling edge of the clock.
The parallel transmit busses may be configured such that a single transmit clock (TCLKC) is used to clock in the
parallel data for all 6 channels, or each parallel transmit bus may have its own transmit clock. The transmit clocks
must not have any frequency deviation relative to REFCLK. Each transmit channel employs a small FIFO to
allow for arbitrary phase relationships between the transmit clock and REFCLK, but the transmit FIFO will not
account for frequency variance between the transmit clocks and REFCLK. On the receive side, if the CTC FIFO
is used for a given channel, then the receive data bus for that channel will be clocked out by a buffered version
of REFCLK. If the CTC FIFO is not used, then the recovered byte clock for that channel will be 1/10th the rate of
the incoming serial data stream.
The interface modes can be controlled via CODE and CTC. A summary of the TLK2226 parallel interface modes
is shown in Table 2.
Table 2. Parallel Interface Modes
CODE
Low=RTBI
High=RGMII
Low
CTC
Low=disable
High=enable
RTBI Mode
Bit[4]
Bit[3]
Bit[2]
Bit[1]
Bit[0]
CLK
High
Data[4]
Data[3]
Data[2]
Data[1]
Data[0]
↑
CTRL1
Data[3]
Data[2]
Data[1]
Data[0]
↑
RTBI channel mode:
• No clock tolerance compensation is performed on Receive data.
• Data towards MAC are aligned to individual RCLKs on a
source-centered/aligned mode per channel basis
Data[9]
Data[8]
Data[7]
Data[6]
Data[5]
↑
RGMII Mode
Bit[4]
Bit[3]
Bit[2]
Bit[1]
Bit[0]
CLK
x
DESCRIPTION
CTRL2
Data[7]
Data[6]
Data[5]
Data[4]
↑
Low
RGMII (without CTC) channel mode:
• No clock tolerance compensation is performed
• Data towards MAC are aligned to individual RCLKs on a
source-centered/aligned per channel basis.
High
RGMII channel mode
•
•
All data towards MAC are aligned to RCLKx thru CTC FIFO. (TLK2226
accepts +/-200ppm frequency offset on the receive data relative to the
transmit data.)
Each RCLKx is a buffered version of REFCLK, source centered/aligned
relative to the receive data.
TRANSMIT DATA PATH
The transmit logic converts parallel data into an NRZ serial bit stream with a differential VML output at a rate of
from 1.0 to 1.3Gbps
Transmit Clock Interface
Transmit clocks for all channels is expected to be the same frequency as the reference clock, REFCLK, but of
arbitrary phase relationship to REFCLK.
Submit Documentation Feedback
19
TLK2226
www.ti.com
SLLS689 – JANUARY 2006
Transmit Parallel Interface
In synchronous channel mode, parallel data to be transmitted is latched by both the rising and falling edges of
TCLKC. TCLKC must be frequency synchronous with REFCLK (0 ppm), but may have any phase relationship
with respect to REFCLK. While in synchronous channel mode, it is recommended that the unused individual
channel clocks TCLKx be grounded to minimize noise.
In independent channel mode, parallel data to be transmitted on channel A is latched on the rising and falling
edges of TCLKA. The transmit data bus for channel B is latched on the rising and falling edges of TCLKB, the
transmit data bus for channel C is latched on the rising and falling edges of TCLKC and so on.
t PULSE
VDDQ/2
TCLKx
tr
TDx[4:0]
tf
t SETUP
t HOLD
t SETUP
LSB Nibble
t HOLD
MSB Nibble
Figure 19. Transmit Timing (Clock Centered)
Serializer
The parallel-to-serial shift register on each channel takes in an internally generated 10-bit wide word from either
the 8b/10b encoders, if enabled, or directly derived from the 5-bit to 10-bit converter and converts it to a serial
stream. The shift register is clocked by the internally generated bit clock, which is 10 times the reference clock
(REFCLK) frequency. The least significant bit (LSB) for each channel is transmitted first.
Serial Output Drivers
The TLK2226 serial transmitter utilizes a Voltage Mode Logic (VML) driver that drives both the high level and low
level of the output signal swing. Thus, no external pull-up resistors are needed as in the case of Current Mode
Logic (CML) drivers, nor are pull-down resistors needed as in the case of Emitter-Coupled Logic (ECL). The
output voltage swing is set at a nominal 800 mV on each leg of the differential signal, making the output signal
swing compatible with PECL systems.
The TLK2226 output driver has a pre-emphasis option (selectable through MDIO) for improved performance over
lossy media such as backplane traces.
RECEIVE DATA PATH
The receiver input data must be ac-coupled and have a rate of 1.0–1.3 Gbps (or 1/10th that rate for 100FX
mode). Resistive termination is implemented on-chip to match 50Ω.traces. The clock recovery circuitry retimes
the input data by extracting a clock from the input data, and passes on the serial data and this recovered clock to
the deserializer. Byte alignment is performed on K-characters per IEEE 802.3z while in 1.25Gbps operation (see
byte alignment logic section for details). Byte alignment is performed on the NRZI Encoded IDLE pattern while in
125Mbps mode.
Serial Input Samplers
The TLK2226 strobes the received serial data with an internal high-speed clock that is locked to the data stream
by the clock recovery function. The differential serial receive data buffers internally terminate the transmission
line with 50Ohm termination resistors. It is expected that the serial receive data is AC coupled, so the internal
termination also biases the signal to the common mode level that is optimum for the input samplers. No external
components are needed apart from the AC coupling capacitors themselves
20
Submit Documentation Feedback
TLK2226
www.ti.com
SLLS689 – JANUARY 2006
Clock Recovery
A baud rate clock is extracted from the 10-bit encoded serial data stream independently on each channel. After
powering up the device, it takes approximately 8 µs before data is received and the device is ready to transmit.
The receive clock locks to the input within 2 µs after a valid input data stream is applied. The received data is
de-serialized and byte aligned to IEEE802.3z /K28.5-characters (while in 1.25Gbps mode). In the absence of
input data, the clock recovery circuit will lock onto the reference clock frequency REFCLK; once valid data is
again presented at the serial inputs, a time will elapse before the channel reacquires lock and performs data and
clock recovery functions.
Deserializer
For each channel, serial data is received on the RXx+/RXx– pins. The interpolator and clock recovery circuit will
lock to the data stream if the clock to be recovered is within ±200 PPM of the internally generated bit rate clock,
assuming a ±100 PPM tolerance on the reference clock. The recovered clock is used to retime the input data
stream. The serial data is then clocked into the serial-to-parallel shift registers. If enabled, the 10-bit wide parallel
data is then fed into 8b/10b decoders. The parallel data for each channel is fed into a FIFO buffer where the
output is synchronized to REFCLK when CTC is on. In case CTC is off data is synchronized to 1/10th of the
received bit clock.
Byte Alignment Logic (1.25 Gbps mode)
Under default conditions, the TLK2226 uses the IEEE802.3z defined 10-bit K28.5 character (comma character,
positive disparity) word alignment scheme . The following sections explain how this scheme works and how it
realigns itself. When parallel data is clocked into a parallel to serial converter, the byte boundary that was
associated with the parallel data is lost in the serialization of the data. When the serial data is received and
converted to parallel format again, a method is needed to be able to recognize the byte boundary again.
Generally, this is accomplished through use of a synchronization pattern. This is a unique a pattern of 1’s and 0’s
that either cannot occur as part of valid data or is a pattern that repeats at defined intervals. 8B/10B encoding
makes allowance for a special pattern called the comma (B’0011111’) which is used by the byte alignment circuit
to align the received serial data back to its original byte boundary. The decoder detects the comma pattern in the
K28.5 character, generating a synchronization signal aligning the data to their 10-bit boundaries for decoding. It
then converts the data back into 8-bit data. It is important to note that the comma can be either a (B’0011111’) or
the inverse (B’1100000’) depending on the running disparity. The TLK2226 decoder will detect only the
(B’0011111’) pattern. Therefore, since synchronization is achieved on the positive comma, two consecutive
K-codes containing commas are required to guarantee byte boundary synchronization (see Table 3 Valid K
characters for K-codes containing positive commas).
During all operations, the TLK2226 receive clocks (RCLKx) are a constant duty cycle and frequency. There are
no stretched nor shortened clock pulses.
Submit Documentation Feedback
21
TLK2226
www.ti.com
SLLS689 – JANUARY 2006
Table 3. Valid K Characters
K CHARACTER
RECEIVE DATA BUS
ENCODED K-CODE
NEGATIVE RUNNING
DISPARITY
K28.0
000 11100
001111 0100
1001 (1)
POSITIVE RUNNING
DISPARITY
110000 1011
K28.1
001 11100
001111
K28.2
010 11100
001111 0101
110000 1010
K28.3
011 11100
001111 0011
110000 1100
K28.4
100 11100
001111 0010
110000 1101
K28.5
101 11100
001111 1010 (2)
110000 0101
K28.6
110 11100
001111 0110
110000 1001
1000 (2)
110000 0110
K28.7
111 11100
001111
K23.7
111 10111
111010 1000
000101 0111
K27.7
111 11011
110110 1000
001001 0111
K29.7
111 11101
101110 1000
010001 0111
K30.7
111 11110
011110 1000
100001 0111
(1)
(2)
110000 0111
Setting COMMA_DET=0, or SYNCEN pin K10 by changing its value via MDIO 17.8 would disable
comma detection, and byte alignment would take place on any bit boundary; this permits external byte
alignment on different bit sequences, and allows for the use of different bit balancing algorithms.
A comma is contained within this K-code.
Receive Parallel Interface
The receive data bus for all channels may be output in either a source-centered fashion (with the bus clock
transitions in the center of the data eye) or a source-aligned fashion (with the bus clock transitions aligned to the
data transitions) allowing direct connection to the protocol device.
Parallel data to be transferred to the protocol device is output referenced to both the rising and falling edges of
RCLKx. RCLKx is frequency synchronous with REFCLK when the Clock Tolerance Compensation logic is
enabled, but RCLKx has no set phase relationship with respect to REFCLK. Each channel is output on RDx[4:0]
respectively, as shown in Figure 20.
Parallel data to be transferred to the protocol device on channel A (RDA[4:0]) is output referenced to both the
rising and falling edges of the RCLKA, Channel B (RDB [4:0]) is output referenced to both the rising and falling
edges of RCLKB. Channels C through F are output in the same fashion with their respective clocks. Nibble
ordering is such that least significant bits are output first and on the rising edge of the clock.
It is important to note, that when the Clock Tolerance Compensation logic is enabled, all recovered byte clocks
(RCLKx) are buffered versions of REFCLK. When the Clock Tolerance Compensation circuit is disabled, the
RCLKx for a channel will be the true 1/10th rate clock recovered from the incoming data stream. Like transmit
timing options, receive clocks RCLKx are either Source Centered or Aligned meeting specified output setup and
hold times. Clock Tolerance Compensation is only available in RGMII mode.
VDDQ/2
RCLK[A−F]
tr
RDx[4:0]
tf
t SETUP
t HOLD
t SETUP
LSB Nibble
Figure 20. RTBI Receive Timing (Clock Centered)
22
Submit Documentation Feedback
t HOLD
MSB Nibble
TLK2226
www.ti.com
SLLS689 – JANUARY 2006
PHYSICAL CODING SUBLAYER (PCS) FUNCTIONS (1.25 Gbps MODE)
8b/10b Encoder
All true serial interfaces require a method of encoding to insure sufficient transition density for the receiving PLL
to acquire and maintain lock. The encoding scheme also maintains the signal DC balance by keeping the number
of ones and zero’s the same which allows for AC coupled data transmission. The TLK2226 uses the 8B/10B
encoding algorithm that is used by Fibre Channel and Gigabit Ethernet. This provides good transition density for
clock recovery and improves error checking. The 8B/10B encoder function is enabled for all channels by the
assertion of the CODE bit. When enabled, the TLK2226 will internally encode the data such that the user actually
writes 8-bit data out on each channel. The encoded 10 bit data will then be transmitted.
When enabled (CODE = high), the 8B/10B encoder converts 8-bit wide data to a 10-bit wide encoded data
character to improve its transition density. This transmission code includes D-characters, used for transmitting
data, and K-characters, used for transmitting protocol information. Each K or D character code word can also
have both a positive and a negative disparity version. The disparity of a code word is selected by the encoder to
balance the running disparity of the serialized data stream.
The TLK2226 will transmit and receive all of the valid K-characters used in the IEEE802.3 Gigabit Ethernet
ordered sets.
8b/10b Decoder
The 8B/10B decoder function is enabled for all channels by the assertion of the CODE bit. When enabled, the
TLK2226 will internally decode the deserialized data such that the user actually reads the decoded bytes out on
each channel.
Clock Tolerance Compensation (CTC) FIFOs
The TLK2226 compensates for the possibility that incoming serial data rate on any channel can be as much as
100 ppm faster or slower than the REFCLK frequency (±100ppm). Each channel independently and dynamically
compensates for any frequency difference by use of an elasticity buffer. If the incoming data rate is faster than
the REFCLK frequency, the elasticity buffer will fill. As it approaches the FILL limit, it will delete or drop a 20-bit
IDLE code (1) found in the IPG (inter packet gap) between Ethernet packets. If the incoming data rate is slower
than the REFCLK, the elasticity buffer will empty. As it approaches the EMPTY limit, it will add or insert the 20-bit
IDLE code found in the gap between Ethernet packets. Idle code selection defaults to IDLE2, a K28.5 followed
by a D16.2. No running disparity is affected due either the addition or the deletion of the IDLE2 code as the
IDLE2 code has a balanced number of 1's and 0's. Note that a deletion of a 20-bit IDLE2 code could reduce the
inter-packet gap below the minimum inter-packet gap of 12 bytes (120 bits).
The CTC function will only add or delete IDLE codes in the inter-packet gap. Thus the CTC FIFO depth is sized
to insure that maximum size Ethernet packets (9300 bytes) can be continuously received at the frequency offset
extremes without loss of data or synchronization and still meet latency requirements per the standard. The CTC
function can be disabled for all channels via the MDIO registers or through its dedicated pin. The CTC function is
available only when the TLK2226 is set in the transceiver mode by MDIO or CODE pin or activation.
During Auto-negotiation the TLK2226 will support clock tolerance during config word (which are converted to
idles by the PCS receive state machine) reception, so that FIFO overrun or underrun does not occur during
auto-negotiation.
AutoNegotiation (AN)
AN is initialized for all channels at power-up once all PLL’s have locked. AN can be restarted on a per channel
basis by writing a logic HIGH to the MDIO reg0x00[9] for the specific channel. The AN process is implemented
according to IEEE 802.3z clause 37. The Status and results of the AN process can be read from the MDIO.
TLK2226 also supports the transmit of NextPage (NP) information provided by it’s management device via
MDIO.
CODE needs to be asserted and ‘AN Enable’ register in the MDIO 0x00[12] is set to a logic HIGH.
(1)
IEEE802.3z specifies an IDLE as a 20-bit code consisting of an IDLE1 code (/K28.5/D5.6) or an IDLE2 code (/K28.5/D16.2).
Submit Documentation Feedback
23
TLK2226
www.ti.com
SLLS689 – JANUARY 2006
Framing (Encapsulation)
Upon the assertion of TX_EN, TLK2226 encapsulates data asserted by the MAC on the parallel interface (TDx)
into a code_group stream according to IEEE 802.3z clause 36.
Upon the receipt of a valid code_group stream at the serial RX interface, TLK2226 decodes the incoming data,
extracts packets and passes the decoded data to the MAC via the parallel RDx interface according to IEEE
802.3z clause 36.
The 5-bit parallel interfaces of TLK2226 (RX and TX) are compliant to the RGMII specification version 2.0 from
04/01/2002.
CODE needs to be asserted a logic HIGH to perform framing.
PHYSICAL CODING SUBLAYER (PCS) FUNCTIONS (125 Mbps MODE)
4b/5b Encoder/Decoder
The TLK2226 shall implement the 4b/5b encoder and decoder functions according to IEEE802.3 Clause 24 when
in 100FX mode instead of the 8b/10b coding functions.
NRZI Encoder/Decocer
The TLK2226 shall implement the cipher scrambling and descrambling functions according to ISO/IEC
9314-3:1990.
Nibble Alignment
The TLK2226 will align the first properly NRZI decoded /I/J/K/ pattern.
PER CHANNEL RATE DETECTION DETERMINATION
The actual rate per channel is determined by Table 4.
Table 4. Per Channel Rate Determination
0.12
Auto Negotiation
Enable
16.5
Rate Sense
Enable
0.6
Speed Selection
(MSB)
0.13
Speed Selection
(LSB)
16.4
SGMII Mode
Enable
TLK2226 Behavior
1
1
X
X
0
Auto Rate Sense (Default Behavior)
1
0
X
X
0
1000Base-X Rate
0
1
X
X
0
Auto Rate Sense
0
0
1
0
0
1000Base-X Rate
0
0
0
1
0
100Base-FX Rate
X
X
1
0
1
SGMII (1000 Mbps Data path Operation)
X
X
0
1
1
SGMII (100 Mbps Data path Operation)
X
X
0
0
1
SGMII (10 Mbps Data path Operation)
The resulting value of the per channel rate may be read in read only register bit 28.14. This bit indicates the
actual serial rate of a channel as a function of the above bits shown in Table 4.
CONTROL LOGIC
MDIO Management Interface
The TLK2226 supports the Management Data Input/Output (MDIO) Interface as defined in Clause 22 of the IEEE
802.3 Ethernet specification. The MDIO allows register-based management and control of the serial links. Normal
operation of the TLK2226 is possible without use of this interface since all of the essential signals necessary for
operations are accessible via the device pins. However, some additional features are accessible only through the
MDIO. The PHY address contained in the MDIO protocol allows to individually access/address 4 TLK2226
24
Submit Documentation Feedback
TLK2226
www.ti.com
SLLS689 – JANUARY 2006
devices connected to the same MDIO bus. The address of the individual device is set by the pins DVAD[4:3].
PHYAD[4:3] correspond to DVAD[4:3] and is used to select the device. In order to read/write information from/to
the IEEE defined registers for a specific channel within TLK2226, PHYAD[2:0] are to be used in the PHY
Address section. Channel A would be addressed by setting PHYAD[2:0] to 3’b000, Channel B would be
addressed by setting PHYAD[2:0] to 3’b001 and so on.
IEEE defined registers (from 0x00 to 0x08 and 0x0F) are duplicated for each individual channel of TLK2226.
Whether or not additional registers are specific to a single channel or to all channels is lined out in the register
map below.
The MDIO Management Interface consists of a bi-directional data path (MDIO) and a clock reference (MDC). The
timing required to read from the internal registers is shown in Figure 21. The timing required to write to the
internal registers is shown in Figure 22.
MDC
MDIO
High
Z
32 ”1”s
0
1
1
Read
Code
SFD
Preamble
0
A4
A0
PHY
Address
R4
R0
Register
Address
D15
0
Turn
Around
D0
Idle
Data
Figure 21. Management Interface Read Timing
MDC
MDIO
32 ”1”s
Preamble
0
1
SFD
0
1
Write
Code
A4
A0
PHY
Address
R4
R0
1
Register
Address
0
Turn
Around
D15
D0
Data
Idle
Figure 22. Management Interface Write Timing
The MDIO Interface allows up to 32 (16-bit) internal registers. Sixteen registers are defined by the IEEE 802.3
Clause 22 specification. Additional registers are allowed for expanded functionality. The TLK2226 employs all
IEEE defined registers for PCS layer devices. The TLK2226 also implements additional registers for expanded
functionality. The IEEE defined registers and the expanded functionality registers are outlined in Table 5.
All latching bits need to be read twice after device initialization to read current/correct status.
The legend for the register map summary is:
1. Self Clearing
2. ReadWrite
3. Read Only
4. Latched High
5. Latched Low
Submit Documentation Feedback
25
TLK2226
www.ti.com
SLLS689 – JANUARY 2006
Table 5. MDIO Register Map Summary
DEC
HEX
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
0
Name (Hex Reset
Val)
0
0
Reset (1)
Loopbac
k(2)
Speed
Sl lsb(2)
AN
Enable(2)
Power
Down(2)
Isolate
Restart
AN (1)
Duplex
Mode(3)
Collision
Tst(3)
Speed
Sl msb(2)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
Control 1 (0x1140)
1
1
1000Bas 100Base 100Base
e T-4(3)
XFD(3)
-XHD(3)
10Mb/s
FD(3)
10Mb/s
HD(3)
Extende
d Stat(3)
RSVD(3)
MF Prea
Supp(3)
AN
Complet
e(3)
Remote
Fault(3)
AN
Ability(3)
Link
Status(5)
Jabber
Detect(3)
Extende
d Cap(3)
Status 1 (0x4149)
2
2
3
3
100Base 100Base
-T2
-T2 FD(3)
HD(3)
Organizationally Unique Idendifier (19-24)(3)
0
1
0
1
Manufacturer Model Number(3)
0
0
0
0
1
1
0
0
0
0
0
PHY ID1 (0x5060)
4
Next
Page(2)
RSVD(3)
Remote Fault[1:0](2)
RSVD(3)
RSVD(3)
RSVD(3)
Technology Ability[3:0] (2)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
ANEG ADV
(0x0020)
5
5
LP Next
Page(3)
LP
Ack(3)
LP Remote Fault(3)
RSVD(3)
RSVD(3)
RSVD(3)
LP Technology Ability[3:0](3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
AN LP Ability
(0x0020)
6
6
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
Next
Page
Able(3)
Page
Receive
d(4)
RSVD(3)
AN Expansion
(0x0004)
7
7
Next
Page(2)
RSVD(3)
Msg
Page(2)
Acknowl.
NP
2(2)
Toggle(3)
8
8
LP Next
Page(3)
RSVD(3)
LP Msg
Page(3)
LP Ack
2(3)
LP
Toggle(3)
9
—
14
9
—
0E
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
Reserved Registers
(0x0000)
15
0F
1000Bas 1000Bas 1000Bas 1000Bas
e-X
e-X
e-T
e-T FD(3)
FD(3)
HD(3)
HD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
Extended Status
(0x8000)
16
10
Global
Write (1)
TX
Amp(2)
CTC
PCS
Swap
Repeat
Shallow
Enable(2) Enable(2) Enable(2) Enable(2) Loopbk(2
Farend
Loopbac
k(2)
PRBS
Ver
Enab(2)
PRBS
Enable(2)
Rate
Sense
EN(2)
SGMII
EN(2)
TCLK
Sel(2)
Comma
Det(2)
Premphasis[1:0](2)
CH Contrl 0
(0x7024)
17
11
Global
Write (1)
SyncStat
Override
RX_LOC
KREFN(2
RSVD(2)
Chan
Soft
Reset
RSVD(2)
TX DP
Override
PSC
1GX TX
EN(2)
PCS
1GX RX
EN(2)
TX Src
Centere
d(2)
RX Src
Centere
d(2)
CH Control 1
(0x3F0F)
TX PMA
Order(2)
RX PMA
Order(2)
RSVD(2)
RSVD(2)
RSVD(2)
(2)
26
0
Manufacturer Revision Number(3)
4
)
(1)
PHY ID0 (0x4000)
Organizationally Unique Identifier[3:18](3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
Msg/Unformatted Code Field[10:0](2)
AN Next Page
(0x2001)
LP Msg/Unformatted Code Field[10:0](3)
LP AN Next Page
(0x0000)
)
(2)
18
12
RX CTC
Reset(4)
RX CTC
Insert(4)
RX CTC
Delete(4)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
Rx CTC Status
(0x0000)
19
13
TX FIFO
Rst
1GX(4)
TX FIFO
Rst
100FX(4)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
TX FIFO Stat
(0x0000)
20
14
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
TP Gen
En(2)
TP Ver
En(2)
21
15
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
(1) - Self Clearing, (2) - Read/Write, (3) - Read Only, (4) - Latched High, (5) - Latched Low
Submit Documentation Feedback
Test Pattern Sel[2:0](2)
Test Pattern
Control (0x0000)
Test Pat
Sync(3)
Test Pattern Sync
Status (0x0000)
RSVD(3)
CRPAT
Sync(3)
TLK2226
www.ti.com
SLLS689 – JANUARY 2006
Table 5. MDIO Register Map Summary (continued)
DEC
HEX
22
16
Fixed Test Pattern Counter [15:0](3)
Test Pattern
Counter (0xFFFD)
23
17
CRPAT Error Counter1 [31:16] MSB(3)
CRPAT Error
Counter1 (0xFFFF)
24
18
CRPAT Error Counter2 [15:0] LSB(3)
CRPAT Error
Counter2 (0xFFFD)
25
19
RSVD(3)
SYNC_1
GX(5)
RSVD(3)
RSVD(3)
RSVD(3)
26
1A
PG1.0
vs
PG2.0(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
27
1B
28
1C
29
1D
30
1E
31
1F (2)
(2)
15
14
13
12
Test MUX Select [3:0](2)
TI Test(3)
Actual
Rate(3)
Encoder
Inval(4)
Decoder
Inval(4)
11
10
RSVD(3)
9
RSVD(3)
8
RSVD(3)
TX DLL Offset[2:0](2)
7
RSVD(3)
RSVD(3)
6
RSVD(3)
5
RSVD(3)
4
RSVD(3)
RX DLL Offset[2:0](2)
3
2
1
0
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
Descrambler Status
(0x0000)
RSVD(2)
RSVD(2)
RSVD(3)
RSVD(3)
Scrambler/DLL
Control (0x0004)
Test
MUX
En(2)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
Test Mode Control
(0x0000)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
RSVD(3)
Channel Status
(0x4000)
PRBS HS Test
Counter (0xFFFD)
PRBS High Speed Test Counter [15:0](3)
RSVD(3)
RSVD(3)
RX
RAMP[1:
0](2)
RX
Ramp
Dir(2)
Name (Hex Reset
Val)
RX
Ramp
Dir(2)
Rx
Ramp
EN(2)
RX
Tstclk
EN(2)
TX
Testclk
EN(2)
Clock
Watch(2)
RSVD(3)
PRBS Low Speed Test Counter [15:0](3)
TCLK
Gating(2)
Global
Write(2)
DLL
Filter(2)
RSVD(3)
Link Time[1:0](2)
Global Test Control
(0x008B)
PRBS LS Test
Counter (0xFFFD)
(1) - Self Clearing, (2) - Read/Write, (3) - Read Only, (4) - Latched High, (5) - Latched Low
Submit Documentation Feedback
27
TLK2226
www.ti.com
SLLS689 – JANUARY 2006
Table 6. Control 1 Register (0x00) Bit Definitions
Bit(s)
0.15
0.14
Default Value
1B’0
1B’0
Name
Description
Reset
Loopback
Read/Write
Logically OR'd with the RESET pin.
Read/Write
1= Global Resets including FIFO clear
0= Normal operation (default)
Self Clearing (1)
Logically OR'd with LPBK pin. (Also referred to as deep
loopback)
Read/Write
1=enable loopback mode per channel
0=disable loopback mode per channel
0.13
1B’0
Speed Selection(LSB)
This is the least significant bit of the speed selection bits
(MSB is 0.6).
Read/Write
{0.6,0.13} = 2’b10 1000Base-X Rate
{0.6,0.13} = 2’b01 100Base-FX Rate
{0.6,0.13} = 2’b00 10 Mbps (SGMII Mode Only)
TLK2226 will not allow writes of 2’b11 to {0.6,0.13}, since it
is an unsupported rate.
0.12
1B’1
Auto-Negotiation Enable
It allows the device to perform an Auto-Negotiation as
described in IEEE802.3 clause 37.
Read/Write
1= Auto-Negotiation enabled (default)
0= Auto-Negotiation disabled
0.11
1B’0
Power Down
Setting this bit high powers down the device, with
exception that MDIO interface stays active.
Read/Write
All zero’s are sent on the receive parallel interface when
this bit is set.
1 = Power Down mode is enabled.
0 = Normal operation (default).
0.10
1B’0
Isolate
Setting this bit high isolates the channel from the RGMII
interface. Inputs are ignored; Outputs are set to high
impedance.
Read/Write
1 = Isolate is enabled
0 = Normal operation (default)
0.9
1B’0
Restart Auto-Negotiation
Restarts an Auto-Negotiation sequence as described in
IEEE802.3 clause 37.
Read/Write
Self Clearing
1= Restart Auto-Negotiation
0= Normal operation (default)
0.8
1B’1
Duplex Mode
Only Full Duplex is supported. Write is ignored,
Read will return a ‘1’.
Read Only
0.7
1B’0
Collision Test
Not Applicable. Read will return a 0.
Read Only
0.6
1B’1
Speed Selection (MSB)
This is the most significant bit of the speed selection bits
(LSB is 0.13).
Read/Write
{0.6,0.13} = 2’b10 1000Base-X Rate
{0.6,0.13} = 2’b01 100Base-FX Rate
{0.6,0.13} = 2’b00 10 Mbps (SGMII Mode Only)
TLK2226 will not allow writes of 2’b11 to {0.6,0.13}, since it
is an unsupported rate.
0.5:0
(1)
28
6B’00 0000
Reserved
Write is ignored and returns 0 when read.
After reset bit is set to one, it automatically sets itself back to zero on the next MDC clock cycle.
Submit Documentation Feedback
Read only
TLK2226
www.ti.com
SLLS689 – JANUARY 2006
Table 7. Status 1 Register (0x01) Bit Definitions
Bit(s)
Default Value
Name
Description
Read/Write
1.15
1B’0
1000Base-T4
Read will return 0, writes will be ignored.
Read Only
1.14
1B’1
100Base-X FD
Read will return 1, writes will be ignored
Read Only
1.13
1B’0
100Base-X HD
Read will return 0, writes will be ignored
Read Only
1.12
1B’0
10Mb/s FD
Read will return 0, writes will be ignored
Read Only
1.11
1B’0
10Mb/s HD
Read will return 0, writes will be ignored
Read Only
1.10
1B’0
100Base-T2 FD
Read will return 0, writes will be ignored
Read Only
1.9
1B’0
100Base-T2 HD
Read will return 0, writes will be ignored
Read Only
1.8
1B’1
Extended Status
Read will return 1 indicating extended status information is held in
register 0x0F.
Read Only
1.7
1B’0
Reserved
Read will return 0, writes will be ignored
Read Only
1.6
1B’1
MF Prea Supp
Read will return 1 indicating the capability to accept a MDIO
command without preceding preamble. Writes will be ignored .
Read Only
1.5
1B’0
AN Complete
Read will return the status if Auto Negotiation is completed.
Read Only
0 = AN not completed
1 = AN has been completed successful
1.4
1B’0
Remote Fault
Not Applicable. Read will return a 0.
Read Only
1.3
1B’1
AN Ability
Read will return 1, indicating, that device is able to perform a Auto
Negotiation, writes will be ignored
Read Only
1.2
1B’0
Link Status
Read will return the Link Status of the selected channel past auto
negotiation has completed according to IEEE 802.3z clause 37 in
1000Base-X mode.
Read Only/
Latched Low
Will return the Link Status of the link monitor state machine when in
100Base-FX mode.
Note, the link status will always indicate high when in deep loopback
or shallow loopback. In remote loopback mode, the bit represents the
normal bit function.
1 = Link UP
0 = Link DOWN
1.1
1B’0
Jabber Detect
Read will return 0, writes will be ignored
Read Only
1.0
1B’1
Extended Capability
Read will return 1 indicating extended register capability
Read Only
The identifier code is composed of bits 3-24 of the 25-bit organizationally unique identifier (OUI) assigned to
Texas Instruments by the IEEE. The 6-bit Manufacturer Model Number is unique to the TLK2226. The
Manufacturer Revision Number denotes the current revision of the TLK2226.
Table 8. PHY ID0 Identifier (0x02) Bit Definitions
OUI Address Bits 3–18
2.15
2.14
2.13
2.12
2.11
2.10
2.9
2.8
2.7
2.6
2.5
2.4
2.3
2.2
2.1
2.0
0
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Table 9. PHY ID1 Identifier (0x03) Bit Definitions
OUI Address Bits 19-24
Manufacturer Model Number
Manufacturer Revision Number
3.15
3.14
3.13
3.12
3.11
3.10
3.9
3.8
3.7
3.6
3.5
3.4
3.3
3.2
3.1
3.0
0
1
0
1
0
0
0
0
0
1
1
0
0
0
0
0
Submit Documentation Feedback
29
TLK2226
www.ti.com
SLLS689 – JANUARY 2006
Table 10. Autonegotiation Advertisement Register (0x04) Bit Definitions (1)
Bit(s)
4.15
Default Value
1B’0
Name
Description
Next Page
Writing a “1” will signal the link partner that a next page of
information is available to be sent after the base page during
autonegotiation.
Read/Write
Read/Write
0 = normal operation.
See IEEE 802.3 clause 37. Section 37.2.1.7
4.14
1B’0
Reserved
Read will return 0, writes will be ignored
Read Only
4.13:12
2B’00
Remote Fault
00 =
01 =
10 =
11 =
Read/Write
No Error (Link OK)
Offline
Link Failure
Auto-Negotiation Error
See IEEE 802.3 clause 37. Section 37.2.1.5
4.11:9
3B’000
Reserved
Read will return 0, writes will be ignored
Read Only
4.8:5
4B’0001
Technology Ability
Field
0001 = Full Duplex
Read/Write
Reserved
Read will return 0, writes will be ignored
4.4:0
(1)
5B’0_0000
See IEEE 802.3 clause 37. Sections 37.2.1.2 to 37.2.1.4
Read Only
In SGMII mode, values written to the Autonegotiation Advertisement Register (0x04) will not be used to configure the functions of the
local device. Instead, 16'h40001 value is advertised as the ability of the local device per the SGMII specification.
Table 11. AN Link Partner Ability Base Page Register (0x05) Bit Definitions
Bit(s)
5.15
Default Value
1B’0
Name
Description
LP Next Page
A “1” will indicate that the link partner has information for a next page at
autonegotiation.
Read/Write
Read Only
0 = LP has no information for a next page
See IEEE 802.3 clause 37. Section 37.2.1.7
5.14
1B’0
LP Acknowledge
1 = Indicates that the link partner has successfully received a base or next
page
0 = Link partner has not received a base or next page
Read Only
5.13:12
2B’00
LP Remote Fault
5.11:9
3B’000
Reserved
Reading these three bits will reflect the Base page ability bits sent by the
link partner through rx_config.
Read Only
5.8:5
4B’0001
Link Partner
Technology
Ability Field
0001 = Full Duplex
Read Only
Reserved
Reading these five bits will reflect the Base page ability bits sent by the link Read Only
partner through rx_config.
See IEEE 802.3 clause 37. Section 37.2.1.6
00
01
10
11
=
=
=
=
No Error (Link OK)
Offline
Link Failure
Auto-Negotiation Error
Read Only
See IEEE 802.3 clause 37. Section 37.2.1.5
5.4:0
5B’0_0000
See IEEE 802.3 clause 37. Sections 37.2.1.2 to 37.2.1.4
Table 12. AN Expansion Register (0x06) Bit Definitions
Bit(s)
Default Value
Name
Description
Read/Write
6.15:3
13B’0000_0000_0000_0
Reserved
Read will return 0, writes will be ignored
Read Only
6.2
1B’1
Next Page Able
Read will return 1, writes will be ignored
Read Only
6.1
1B’0
Page Received
0 = no new page was received
1 = a new page has been received
Read Only/
Latched High
6:0
1B’0
Reserved
Read will return 0, writes will be ignored
Read Only
30
Submit Documentation Feedback
TLK2226
www.ti.com
SLLS689 – JANUARY 2006
Table 13. AN Next Page Transmit Register (0x07) Bit Definitions
Bit(s)
7.15
Default Value
Name
1B’0
Description
Next Page
Read/Write
Writing a “1” will signal the link partner that a next page of
Read/Write
information is available to be sent after the base page during
autonegotiation.
See IEEE 802.3 clause 37. Section 37.2.1.7
7.14
1B’0
Reserved
Read will return 0, writes will be ignored
Read Only
7.13
1B’1
Msg Page
0 = unformatted page
1 = message page
Read/Write
7.12
1B’0
Acknowledge 2
0 = cannot comply with message
1 = will comply with message
Read/Write
7.11
1B’0
NP Toggle
0 = prev. value of transmitted link code words equals 1’b1
1 = prev. value of transmitted link code words equals 1’b0
Read Only
7.10:0
11B’000_0000_0001
Msg/Unformated
Code Field
See IEEE 802.3 annex 28C for details
Read/Write
Table 14. AN Link Partner Ability Next Page Register (0x08) Bit Definitions
Bit(s)
8.15
Default Value
Name
1B’0
Description
LP Next Page
A “1” will indicate that the link partner has information for
a next page at autonegotiation.
Read/Write
Read Only
0 = LP has no information for a next page
See IEEE 802.3 clause 37. Section 37.2.1.7
8.14
1B’0
Reserved
Reserved. Ignore on read.
Read Only
8.13
1B’0
LP Msg Page
0 = unformatted page
1 = message page
Read Only
8.12
1B’0
LP Acknowledge 2
0 = cannot comply with message
1 = will comply with message
Read Only
8.11
1B’0
LP Toggle
0 = prev. value of transmitted link code words equals 1’b1
Read Only
1 = prev. value of transmitted link code words equals 1’b0
8.10:0
11B’000_0000_0000
LP Msg/Unformated
Code Field
See IEEE 802.3 annex 28C for details
Read Only
Table 15. Reserved Registers (0x09…0x0E) Bit Definitions
Bit(s)
9.15:0…
14.15:0
Default Value
Name
16B’0000_0000_0000_0000
Reserved
Description
Read will return 0, writes will be ignored
Read/Write
Read Only
Table 16. Extended Status Register (0x0F) Bit Definitions
Bit(s)
Default Value
Name
Description
Read/Write
15.15
1B’1
1000Base-X FD
Always reads 1, indicating device supports Full Duplex mode.
Read Only
15.14
1B’0
1000Base-X HD
Read will return 0, writes will be ignored.
Read Only
15.13
1B’0
1000Base-T FD
Read will return 0, writes will be ignored.
Read Only
15.12
1B’0
1000Base-T HD
Read will return 0, writes will be ignored.
Read Only
15.11:0
12B’0000_0000_0000
Reserved
Read will return 0, writes will be ignored.
Read Only
Submit Documentation Feedback
31
TLK2226
www.ti.com
SLLS689 – JANUARY 2006
Table 17. (Vendor Specific) CH Control Register 0 (0x10) Bit Definitions
Bit(s)
Default Value
Name
Description
Read/Write
16.15
1B’0
Global write
When written as 1 the settings in 16.14:0 will affect all channels of one
Read/Write
device simultaneously.
Self Clearing
When written as 0 the settings in 16.14:0 are only valid for the addressed
channel. This value always reads zero.
16.14
1B’1
TX Amplitude
Control
1 = Transmitter Amplitude is at full value.
0 = Transmitter Amplitude is half value.
This should be set to 0 during SGMII mode.
Read/Write
16.13
1B’1
CTC Enable
Logically AND'd with the CTC_EN pin.
0 = clock tolerance compensation is disabled
1 = CTC is enabled (default)
Read/Write
16.12
1B’1
PCS Enable
Logically AND'd with the CODE pin.
0 = PCS functions are disabled
1 = PCS functions are enabled (default)
Read/Write
16.11
1B’0
Swap Enable
When logic high the parallel interfaces of two adjacent channels are
swapped. For example data coming in at TDA will be sent out at TXB.
Data coming in at RXA will be sent out at RDB in this example.
Read/Write
This bit is always valid for a channel pair like AB, CD, or EF.
Logically OR'd with the SWAP pin.
0 = Swap function disabled.
1 = Swap function enabled.
16.10
1B’0
Repeat Enable
When logic high the channel will ignore the data presented to its transmit Read/Write
parallel interface but will send the data coming in at the serial receive
interface of its partner channel.
The partner channel is the other one of the pair AB, CD, or EF.
Logically OR'd with REPEATA or REPEATB pins.
16.9
1B’0
Serial Shallow
Loopback
Applicable only when standard loopback bit (Register 0.14) is low and
the external loopback pin (LPBK) is low.
Read/Write
When asserted high the data presented at the parallel transmit interface
is looped back to the parallel receive interface of the same channel.
0 = Serial shallow loopback is disabled.
1 = Serial shallow loopback is enabled.
16.8
1B’0
Farend Loopback
When asserted high the data presented at the serial receive interface is
looped back to the serial transmit interface of the same channel via the
deserializer, the serializer and if enabled the PCS function. If 1GX PCS
is not enabled (i.e. RTBI or 100Base-FX modes), the incoming data rate
must be frequency locked (ppm 0) with REFCLK.
Read/Write
Also referred to as remote loopback.
0 = Farend Loopback is disabled.
1= Farend loopback is enabled.
16.7
1B’0
PRBS Verifier
Enable
16.6
1B’0
PRBS Enable
A logic 1 enables the PRBS verifier in the receive data path.
Read/Write
Logically OR'd with the PRBSEN pin.
A logic 1 enables the PRBS generator in the transmit data path.
Read/Write
Logically OR'd with the PRBSEN pin.
16.5
1B’1
Rate Sense Enable A logic 1 will enable auto rate sensing between 1.25 Gbps data and 125
Mbps data.
Read/Write
16.4
SGMII Pin Value
at Rising Edge of
RESET.
SGMII Mode
Enable
Read/Write
When set to 1, enables SGMII mode.
When set to 0 (default), disables SGMII mode.
When set to 1, the standard based bits 0.6 and 0.13 directly determine
the parallel data rate of operation, according to the “Per Channel Rate
Determination” table.
16.3
1B’0
TCLK Select
Logically OR'd with the TCLKSEL pin.
Read/Write
0 = parallel transmit data is latched synchronous to TCLKC (default)
1 = parallel transmit data is latched synchronous to TCLK of the
according channel.
16.2
32
1B’1
Comma Detect
A logic 1 enables the comma detection in the serial receive circuit.
Submit Documentation Feedback
Read/Write
TLK2226
www.ti.com
SLLS689 – JANUARY 2006
Table 17. (Vendor Specific) CH Control Register 0 (0x10) Bit Definitions (continued)
Bit(s)
Default Value
Name
Description
Read/Write
When set to logic 0 the byte boundary of the received data stream is
ignored.
Logically AND'd with the SYNCEN pin.
16.1:0
2B’00
Preemp
Pre-emphasis control
2B’00 =
2B’01 =
2B’10 =
2B’11 =
Read/Write
no Pre-emphasis (default)
low Pre-emphasis
medium Pre-emphasis
high Pre-emphasis
Submit Documentation Feedback
33
TLK2226
www.ti.com
SLLS689 – JANUARY 2006
Table 18. (Vendor Specific) CH Control 1 Register (0x11) Bit Definitions
Bit(s)
Default Value
Name
Description
Read/Write
17.15
1B’0
Global write
When written as 1 the settings in 17.14:0 will affect all channels of one
device simultaneously.
When written as 0 the settings in 17.14:0 are only valid for the addressed
channel. This bit always reads zero.
17.14
1B’0
Sync Status
Override
1 = Causes an override of the sync state 1000Base-X synchronization state Read/Write
machine to reflect a “1” in the sync_1GX (25.14) bit.
0 = Original (normal operation) sync_1GX value is represented in bit 25.14.
(Default)
17.13
1’B1
TX PMA Bit
Order
When asserted, allows the ten bits of data given to the parallel side of the
Read/Write
SERDES TX macro to be flipped. This is normally set since the SERDES
transmits MSB first, and the 1000Base-X standard requires LSB to be
transmitted first. For standard based operation, the customer may leave this
bit alone.
17.12
1’B1
RX PMA Bit
Order
When asserted, allows the ten bits of data received from the parallel side of Read/Write
the SERDES RX macro to be flipped. This is normally set since the
SERDES receives MSB first, and the 1000Base-X standard requires LSB to
be received first. For standard based operation, the customer may leave
this bit alone.
17.11: 3B’111
9
Reserved
Reserved for TI test.
17.8
RX_LOCKREFN
1B’1
1 = Enables clock recovery loop to lock onto incoming data rate.
Read/Write
Self Clearing
Read/Write
Read/Write
0 = Clock recovery loop ignores incoming data rate and locks to REFCLK.
17.7
1B’0
Reserved
17.6
1’B0
Channel Soft
Reset
17.5
1B’0
Reserved
17.4
1B’0
TX Datapath
Override
Reserved for TI test.
1 = Resets channel logic excluding MDIO registers.
Reserved for TI test.
1 = Allows Transmit Datapath to Send Normal Data Traffic Regardless of
Errors in the Receive Datapath (In 1000Base-X Mode,
Auto-Negotiation should also be disabled).
Read/Write
Read/Write
Self-Clearing
Read Only
Read/Write
0 = Normal Chip Operation (Standard Based)
17.3
1’B1
PCS 1GX TX
Enable
17.2
1’B1
PCS 1GX Rx
Enable
17.1
1B’1
TX Source
Centered
1 = Enables 1000Base-X PCS TX Function
Read/Write
0 = Disables 1000Base-X PCS TX Function
1 = Enables 1000Base-X PCS Rx Function
Read/Write
0 = Disables 1000Base-X PCS Rx Function
0 = Source synchronous timing on transmit parallel interface. Data is
latched 2ns after clock edge.
Read/Write
1 = Source centered timing on transmit parallel interface. Data is latched
at clock edge.
17.0
1B’1
RX Source
Centered
0 = Source synchronous timing on receive parallel interface. Data
changes at clock edge.
1 = Source centered timing on receive parallel interface. Data changes
2ns after clock edge.
34
Submit Documentation Feedback
Read/Write
TLK2226
www.ti.com
SLLS689 – JANUARY 2006
Table 19. (Vendor Specific) RX CTC FIFO Status (0x12) Bit Definitions
Bit(s)
Default Value
Name
Description
Read/Write
18.15
1B’0
RX_CTC_Reset
When high indicates overflow or underflow has occurred in CTC Read Only/
FIFO and FIFO has been reset.
Latched High
18.14
1B’0
RX_CTC_Insert
When high indicates RX CTC has inserted at least one ordered
set.
Read Only/
Latched High
18.13
1B’0
RX_CTC_Delete
When high indicates RX CTC has deleted at least one ordered
set.
Read Only/
Latched High
18.12:0
13B’0_0000_0000_0000 Reserved
Read will return 0, writes will be ignored
Read Only
Table 20. (Vendor Specific) TX FIFO Status (0x13) Bit Definitions
Bit(s)
Default Value
Name
Description
Read/Write
19.15
1B’0
TX_FIFO_Reset_1Gx
When high indicates collision has occurred in TX FIFO
and the FIFO is reset in 1gx mode.
Read Only/
Latched High
19.14
1B’0
TX_FIFO_Reset_100F
x
When high indicates collision has occurred in TX FIFO
and the FIFO is reset in 100fx mode.
Read Only/
Latched High
19.13:0
14B’00_0000_0000_0000
Reserved
Read will return 0, writes will be ignored
Read Only
Table 21. Vendor Specific) Test Pattern Control Registers (0x14) Bit Definitions
Bit(s)
Default Value
Name
Description
Read/Write
20.15:5
11B’000_ 0000_0000
Reserved
Read will return 0, writes will be ignored
Read Only
20.4
1B’0
Test Pattern
Generator Enable
When high activates the generator selected by bits 20.2:0. Read/Write
20.3
1B’0
Test Pattern
Verifier Enable
When high activates the verifier selected by bits 20.2:0.
Read/Write
20.2:0
2B’00
Pattern Select
Test Pattern Selection
000 = High Frequency Test Pattern
001 = Low Frequency Test Pattern
010 = Mixed Frequency Test Pattern
011 = CRPAT Long
100 = CRPAT Short
Others = Reserved
Read/Write
Table 22. (Vendor Specific) Test Pattern Sync Status (0x15) Bit Definitions
Bit(s)
Default Value
Name
Description
Read will return 0, writes will be ignored
Read/Write
21.15:2
14B’00_0000_0000_0000
Reserved
Read Only
21.1
B’0
Test Pattern Sync When high indicates alignment has been determined and a Read Only
correct pattern has been received for fixed test patterns.
21.0
1B’0
CRPAT Sync
When high indicates alignment has been determined and a Read Only
correct pattern has been received for continuous test
patterns.
Table 23. (Vendor Specific) Test Pattern Counter (0x16) Bit Definitions
Bit(s)
Default Value
Name
22.15:0
16B’1111_1111_1111_1101
Fixed Test Pattern
Error Counter
Description
Read/Write
This counter reflects error count for high, Mixed, and Low Read Only
Frequency test patterns. Counter increments for each
received character that has an error. Counter clears upon
read.
Table 24. (Vendor Specific) CRPAT Error Counter1 (0x17) Bit Definitions
Bit(s)
Default Value
23.15:0
16B’1111_1111_1111_1111
Name
Reserved
Description
MSB of CRPAT Error Counter. This counter reflects errors
for CRPAT Test. Counter increments by one for each
received character that has an error. Counter clears upon
read.
Submit Documentation Feedback
Read/Write
Read Only
35
TLK2226
www.ti.com
SLLS689 – JANUARY 2006
Table 25. (Vendor Specific) CRPAT Error Counter2 (0x18) Bit Definitions
Bit(s)
Default Value
Name
24.15:0
16B’1111_1111_1111_1101
Description
Reserved
Read/Write
LSB of CRPAT Error Counter. This counter reflects errors
for CRPAT Test. Counter increments by one for each
received character that has an error. Counter clears upon
read.
Read Only Note:
Register 23 has
to be read first.
Table 26. (Vendor Specific) Descrambler Status (0x19) Bit Definitions
Bit(s)
Default Value
Name
Description
Read/Write
25.15
1B’0
Reserved
Read will return 0, writes will be ignored.
Read Only
25.14
1’B0
Sync_1GX
When low, indicates PCS synchronization has lost sync.
When high, indicates PCS synchronization is currently in
sync.
Read Only
Latched Low
25.13:0
14B’000_0000_0000_0000
Reserved
Read will return 0, writes will be ignored
Read Only
Table 27. (Vendor Specific) Scrambler/DLL control (0x1A) Bit Definitions
Bit(s)
Default Value
Name
Description
Read/Write
26.15
PG1.0 : 1’b1
PG2.0 : 1’b0
PG1.0 –vs- PG2.0
PG1.0 : Read/Write Default to One
PG2.0 : Read Only, Read will return 0
PG1.0 : Read/Write
PG2.0: Read Only
26.14:11
4B’0000
Reserved
Read will return 0, writes will be ignored
Read Only
26.10:8
3B’000
TX DLL Offset[2:0]
Unit delay offset setting for TX DLL.
Read/Write
If [2] (MSB) is 1’b1, then delay can be increased by
[1:0] (LSB’s) unit delays.
If [2] (MSB) is 1’b0, then delay can be decreased by
[1:0] (LSB’s) unit delays.
26.7
1B’0
Reserved
Read will return 0, writes will be ignored
Read Only
26.6:4
3B’000
RX DLL Offset[2:0]
Unit delay offset setting for Rx DLL.
Read/Write
If [2] (MSB) is 1’b1, then delay can be increased by
[1:0] (LSB’s) unit delays.
If [2] (MSB) is 1’b0, then delay can be decreased by
[1:0] (LSB’s) unit delays.
26.3:0
4B’0100
Reserved
Read will return 0, writes will be ignored
Read Only
Table 28. (Vendor Specific) Reserved Test Mode Control Registers (0x1B) Bit Definitions
Bit(s)
Default Value
Name
Description
Read/Write
27.15.12
4’B1111
Test MUX Select
Must be set to 4’B1010 to allow PRBS_PASS and
PRBS_SYNC to be output on the partner RD[1] and RD[2]
parallel interface pins.
Read/Write
27.11
1’b0
Test MUX Enable
When asserted, allows the test MUX data to be driven onto Read/Write
the parallel interface (data from the partner channel).
27.10:0
11B’000_0000_0000
Reserved
Read will return 0, writes will be ignored
Read Only
Table 29. (Vendor Specific) Reserved Channel Status Registers (0x1C) Bit Definitions
Bit(s)
Default Value
Name
Description
Read/Write
28.15
1B’0
TI Test
Reserved for TI test
Read Only
28.14
1’B1
Actual Channel Rate
When high, indicates the serial side of the channel is
currently operating in 1.25 Gbps mode.
When low, indicates the channel is currently operating in
125 Mbps mode.
Read Only
28.13
1’B0
Encoder Invalid Code
Word
When high, indicates that the 1000Base-X encoder
received an invalid control word.
Read Only
Latched High
28.12
1’B0
Decoder Invalid Code
Word
When high, indicates that the 1000Base-X decoder
received an invalid code word.
Read Only
Latched High
28.11:0
12B’0000_0000_0000
Reserved
Read will return 0, writes will be ignored
Read Only
36
Submit Documentation Feedback
TLK2226
www.ti.com
SLLS689 – JANUARY 2006
Table 30. (Vendor Specific) PRBS High Speed Test Counter (0x1D) Bit Definitions
Bit(s)
Default
Value
Name
29.15:0
16B’1111_1111_1111_1101
Description
PRBS High
Speed Test
Counter
Read/Write
This counter reflects errors for high speed PRBS test
Read Only
pattern (1000Base-X). Counter increments by one for each
received character that has error. This counter saturates at
16’hffff. When read, it resets to zero and continues to
count.
Table 31. (Vendor Specific) Global Test Control Register (0x1E) Bit Definitions
Bit(s)
Default Value
Name
Description
Read/Write
30.15:14
2B’00
Reserved
Read will return 0, writes will be ignored
Read Only
30.13:12
2’B00
RX Ramp Rate
Ramp Rate control during Ramp test mode
Read/Write
This bit value can be changed per serdes macro by addressing
to the lead channel of a particular macro (Channel A or C or E)
30.11
2’b0
RX Ramp Direction
Ramp Rate direction during Ramp test mode
Read/Write
This bit value can be changed per serdes macro by addressing
to the lead channel of a particular macro (Channel A or C or E)
30.10
1B’0
Rx Ramp Enable
When set enables Ramp test mode in serdes macro
Read/Write
This bit value can be changed per serdes macro by addressing
to the lead channel of a particular macro (Channel A or C or E)
30.9
1B’0
RX Testclock Enable
30.8
1B’0
TX Testclock Enable
When set enables test clock in all receive SERDES macros
Read/Write
This bit value can be changed by addressing to any channel.
When set enables test clock in all transmit SERDES macros
Read/Write
This bit value can be changed by addressing to any channel.
30.7
1’B1
Clock Watch
When set enables Reference Clock Watchdog function
Read/Write
30.6
1B’0
Reserved
Read will return 0, writes will be ignored
Read Only
30.5
1’B0
TCLK Gating
For TI Test Purposes Only
Read/Write
This bit value can be changed by addressing to any channel.
This bit value can be changed per serdes macro by addressing
to the lead channel of a particular macro (Channel A or C or E)
30.4
1’B0
Global Write
If set all write accesses to MDIO Control (R/W) registers between Read/Write
0x00 to 0x0F and registers 0x14, 0x1A, and 0x1E affect all
channels in the device.
30.3
1B’1
DLL Filter
30.2
1B’0
Reserved
Read will return 0, writes will be ignored
Read Only
30.1:0
2B’11
Link Time
Sets the link time used for 1000Base-X autonegotiation. Be
aware: only default setting complies with IEEE802.3.
Read/Write
This bit value can be changed by addressing to any channel.
For TI test purposes
Read/Write
This bit value can be changed by addressing to any channel.
2B’00
2B’01
2B’10
2B’11
=
=
=
=
500 ns
10 µs
1.6 ms
10 ms (default)
In 100Base-FX mode, these bits are also used. They are used to
set the Far End Fault Link Monitor period. Only the default setting
will comply with the IEEE802.3 standard.
2B’00
2B’01
2B’10
2B’11
=
=
=
=
120 ns
131 µs
262 µs
524 µs (default)
This bit value can be changed per channel.
Submit Documentation Feedback
37
TLK2226
www.ti.com
SLLS689 – JANUARY 2006
Table 32. (Vendor Specific) PRBS Low Speed Test Counter (0x1F) Bit Definitions
Bit(s)
31.15:0
Default Value
16B’1111_1111_
1111_1101
Name
Description
Read/Write
PRBS Low Speed This counter reflects errors for low speed PRBS test pattern Read Only
Test Counter
(100Base-FX). Counter increments by one for each received
character that has error. This counter saturates at 16’hffff.
When read, it resets to zero and continues to count.
JTAG Interface
The TLK2226 provides the full five pin JTAG interface as defined in IEEE 1149.1 to support manufacturing test.
SOFTWARE PROCEDURES
The following are per channel sequences.
• 1000Base-X Transceiver Mode (PCS On/CTC on)
1. write 0x8000 to location 0x00
2. verify actual speed detected bit (28.14). If wrong,
3. poll bit 2 set at location 0x01 (link status)
4. channel is carrying traffic
• 1000Base-X Transceiver Mode (PCS On/CTC Off)
1. write 0x8000 to location 0x00
2. write 0x1027 to location 0x10
3. verify actual speed detected bit (28.14). If wrong,
4. poll bit 2 set at location 0x01 (link status)
5. channel is carrying traffic
• 1000Base-X Transceiver Mode (PCS Off)
1. write 0x8000 to location 0x00
2. write 0x0027 to location 0x10
3. verify actual speed detected bit (28.14). If wrong,
4. poll bit 2 set at location 0x01 (link status)
5. channel is carrying traffic
• 100Base-FX Transceiver Mode
1. write 0x8000 to location 0x00
2. verify actual speed detected bit (28.14). If wrong,
3. poll bit 2 set at location 0x01 (link status)
4. channel is carrying traffic
• 1000Base-X Port Swap (PCS On/CTC On)
1. write 0x8000 to location 0x00
2. write 0x3827 to location 0x10
3. verify actual speed detected bit (28.14). If wrong,
4. poll bit 2 set at location 0x01 (link status)
5. channel is carrying traffic
• 1000Base-X Port Swap (PCS On/CTC Off)
1. write 0x8000 to location 0x00
2. write 0x1827 to location 0x10
3. verify actual speed detected bit (28.14). If wrong,
4. poll bit 2 set at location 0x01 (link status)
5. channel is carrying traffic
38
manually provision.
manually provision.
manually provision.
manually provision.
manually provision
manually provision.
Submit Documentation Feedback
TLK2226
www.ti.com
SLLS689 – JANUARY 2006
•
•
•
•
•
•
•
1000Base-X Port Swap (PCS Off)
1. write 0x8000 to location 0x00
2. write 0x0827 to location 0x10
3. verify actual speed detected bit (28.14). If wrong, manually provision.
4. poll bit 2 set at location 0x01 (link status)
5. channel is carrying traffic
1000Base-X Retiming Repeater (PCS On/CTC On)
1. write 0x8000 to location 0x00
2. write 0x3427 to location 0x10
3. verify actual speed detected bit (28.14). If wrong, manually provision.
4. poll bit 2 set at location 0x01 (link status)
5. channel is carrying traffic
1000Base-X Non-Retiming Repeater (PCS On/CTC Off)
1. write 0x8000 to location 0x00
2. write 0x1427 to location 0x10
3. verify actual speed detected bit (28.14). If wrong, manually provision.
4. poll bit 2 set at location 0x01 (link status)
5. channel is carrying traffic
1000Base-X Non-Retiming Repeater (PCS Off)
1. write 0x8000 to location 0x00
2. write 0x0427 to location 0x10
3. verify actual speed detected bit (28.14). If wrong, manually provision.
4. poll bit 2 set at location 0x01 (link status)
5. channel is carrying traffic
1000Base-X Remote Loopback (PCS On/CTC On)
1. write 0x8000 to location 0x00
2. write 0x3127 to location 0x10
3. verify actual speed detected bit (28.14). If wrong, manually provision.
4. poll bit 2 set at location 0x01 (link status)
5. channel is carrying traffic
1000Base-X Remote Loopback (PCS On/CTC Off)
1. write 0x8000 to location 0x00
2. write 0x1127 to location 0x10
3. verify actual speed detected bit (28.14). If wrong, manually provision.
4. poll bit 2 set at location 0x01 (link status)
5. channel is carrying traffic
1000Base-X Remote Loopback (PCS Off)
1. write 0x8000 to location 0x00
2. write 0x0127 to location 0x10
3. verify actual speed detected bit (28.14). If wrong, manually provision.
4. poll bit 2 set at location 0x01 (link status)
5. channel is carrying traffic
Submit Documentation Feedback
39
TLK2226
www.ti.com
SLLS689 – JANUARY 2006
•
•
•
•
•
•
•
40
100Base-FX Remote Loopback
1. write 0x8000 to location 0x00
2. write 0x3127 to location 0x10
3. verify actual speed detected bit (28.14). If wrong, manually provision.
4. poll bit 2 set at location 0x01 (link status)
5. channel is carrying traffic
1000Base-X Local Loopback (Shallow) (PCS On/CTC On)
1. write 0x8000 to location 0x00
2. write 0x3227 to location 0x10
3. verify actual speed detected bit (28.14). If wrong, manually provision.
4. poll bit 2 set at location 0x01 (link status)
5. channel is carrying traffic
1000Base-X Local Loopback (Shallow) (PCS On/CTC Off)
1. write 0x8000 to location 0x00
2. write 0x1227 to location 0x10
3. verify actual speed detected bit (28.14). If wrong, manually provision.
4. poll bit 2 set at location 0x01 (link status)
5. channel is carrying traffic
1000Base-X Local Loopback (Shallow) (PCS Off)
1. write 0x8000 to location 0x00
2. write 0x0227 to location 0x10
3. verify actual speed detected bit (28.14). If wrong, manually provision
4. poll bit 2 set at location 0x01 (link status)
5. channel is carrying traffic
100Base-FX Local Loopback (Shallow)
1. write 0x8000 to location 0x00
2. write 0x3227 to location 0x10
3. verify actual speed detected bit (28.14). If wrong, manually provision.
4. poll bit 2 set at location 0x01 (link status)
5. channel is carrying traffic
1000Base-X Local Loopback (Deep) (PCS On/CTC On)
1. write 0x8000 to location 0x00
2. write 0x4000 to location 0x00
3. write 0x3027 to location 0x10
4. verify actual speed detected bit (28.14). If wrong, manually provision.
5. poll bit 2 set at location 0x01 (link status)
6. channel is carrying traffic
1000Base-X Local Loopback (Deep) (PCS On/CTC Off)
1. write 0x8000 to location 0x00
2. write 0x4000 to location 0x00
3. write 0x1027 to location 0x10
4. verify actual speed detected bit (28.14). If wrong, manually provision.
5. poll bit 2 set at location 0x01 (link status)
6. channel is carrying traffic
Submit Documentation Feedback
TLK2226
www.ti.com
SLLS689 – JANUARY 2006
•
•
1000Base-X Local Loopback (Deep) (PCS Off)
1. write 0x8000 to location 0x00
2. write 0x4000 to location 0x00
3. write 0x0027 to location 0x10
4. verify actual speed detected bit (28.14). If wrong, manually provision
5. poll bit 2 set at location 0x01 (link status)
6. channel is carrying traffic
100Base-FX Local Loopback (Deep)
1. write 0x8000 to location 0x00
2. write 0x4000 to location 0x00
3. verify actual speed detected bit (28.14). If wrong, manually provision.
4. poll bit 2 set at location 0x01 (link status)
5. channel is carrying traffic
Submit Documentation Feedback
41
TLK2226
www.ti.com
SLLS689 – JANUARY 2006
JITTER TEST PATTERN GENERATION AND VERIFICATION
Use one of the following procedures to generate and verify the respective jitter test pattern:
• High/Mixed/Low Frequency Test Pattern:
– Issue a hard or soft reset
– Write a zero to 0.12 (disable auto-negotiation).
– Write a zero to 16.5 (disable auto rate sense enable).
– Write a one to 0.6 (1000Base-X rate).
– Write a zero to 0.13 (1000Base-X rate).
– Write “0” to the PCS 1Gx Rx Enable bit (bit 2) of the Channel Control_1 register (Reg 17) to disable
decoder/ PCS 1Gx RX function.
– Write “0” to Comma Detect bit (bit 2) of the Channel Control_0 register (Reg 16) to disable comma detect
function.
– Write “000” to the Pattern_select(bit 2:0) field of the Test Pattern Control register (Reg 20) to select high
frequency test pattern, or “001” to select the low frequency test pattern, or “010” to select the mixed
frequency test pattern. .
– Write “1” to the Generator Enable (20.4).
– Read the test pattern error counter register(Reg 22) for the channel to clear the counters.
– Write “1” to the Verifier Enable (20.3)
– In order for the Test Pattern Verifier to start checking the test pattern, it has to achieve sync to the test
pattern. To make sure that the test pattern checking has started, read bit 1(Test pattern sync) of Test
pattern Sync Status (Reg 21) register. Make sure that the Fixed Test Pattern Sync bit is HIGH for the
corresponding lane. If the sync status is not high, this indicates that the verifier never achieved sync, which
may indicate a more severe link problem.
– At this point the pattern verification is in progress
– Errors are reported in the error counter at Test Pattern Counter Register (Reg 22).
– Reading the counters has no effect on the test except clearing the counters, i.e. the verification of the
pattern continues until the test pattern enable bit(bit 2) of the Test Pattern Control register(Reg 20) is
cleared.
• Long/Short Continuous Random Test Pattern (Long/Short CRPAT
– Issue a hard or soft reset
– Write a zero to 0.12 (disable auto-negotiation).
– Write a zero to 16.5 (disable auto rate sense enable).
– Write a one to 0.6 (1000Base-X rate).
– Write a zero to 0.13 (1000Base-X rate).
– Write “011” to the Pattern_select(bit 2:0) field of the Test Pattern Control register (Reg 20) to select
CRPAT Long test pattern, or “100” to select the CRPAT Short test pattern.
– Write “1” to the Generator Enable (20.4).
– Read the CRPAT Error Counter_1 register and CRPAT Error Counter_2 register (Reg 23 & Reg 24) to
clear
– Write “1” to the Verifier Enable (20.3).
– In order for the Test Pattern Verifier to start checking the test pattern, it has to receive the Preamble /SFD
that is sent at every packet from the test pattern generator. To make sure that the test pattern checking
has started, read bit 0 of the Test Pattern Sync Status Register(Reg 21). Make sure that the CRPAT Sync
bit is HIGH. If the sync status is not high, this indicates that the verifier never received the Preamble, which
may indicate a more severe link problem.
– At this point the pattern verification is in progress. Perform the test as long as desired.
– Read the MSB of the error counter by reading CRPAT Error Counter_1 (Reg 23), then read the LSB of the
error counter at CRPAT Error Counter_2 (Reg 24). If user reads register 24 without reading register 23
first, the count value read through 24 may not be correct.
– If another test is to be performed go to the first step
42
Submit Documentation Feedback
TLK2226
www.ti.com
SLLS689 – JANUARY 2006
PRBS GENERATION AND VERIFICATION PROCEDURE
•
•
•
•
•
•
•
•
•
•
•
•
•
•
Issue a hard or soft reset
Write a zero to 0.12 (disable auto-negotiation).
Write a zero to 16.5 (disable auto rate sense enable).
Write a one to 0.6 (1000Base-X rate) or zero to 0.6 (100Base-FX rate).
Write a zero to 0.13 (1000Base-X rate) or one to 0.13 (100Base-FX rate).
Write “0” to the PCS Enable bit (bit 12) of the Channel Control_0 register (Reg 16) to disable PCS function.
Write “0” to Comma Detect bit (bit 2) of the Channel Control_0 register (Reg 16) to disable comma detect
function.
Write “1” to the PRBS Generator Enable (bit 6) field of the Channel Control_0 register (Reg 16) to enable
PRBS generator.
For 1000Base-X mode, Read the PRBS high speed test error counter (Reg 29) to clear the counter
For 100fx mode, Read the PRBS low speed test error counter (Reg 31) to clear the counter
Write “1” to the PRBS Verifier Enable (bit 7) fields of the Channel Control_0 register (Reg 16) to enable
PRBS verifier
At this point the pattern verification is in progress
Errors are reported in PRBS high speed test error counter (Reg 29) for 1000Base-X and PRBS low speed
test error counter (Reg 31) for 100fx mode.
Reading the counters has no effect on the test except clearing the counters, i.e. the verification of the pattern
continues until the PRBS Generator/PRBS Verifier Enable (bit 7:6) of the Channel Control_0 register (Reg
16) is cleared
TESTABILITY
Serial Loopback (Shallow and Deep)
The TLK2226 can provide a self-test function by enabling the internal loop-back path for all channels with the
assertion of LPBK. The loopback for individual channels can be enabled via the MDIO registers or for all
channels through the LPBK pin. The parallel data output can be compared to the parallel input data for that
channel to provide functional verification. The external differential output is held in a high-impedance state (pulled
to VDDA through 500 Ω) during deep loop-back testing. Loopback for all channels (called deep loopback) may be
enabled by assertion of the LPBK pin (putting all channels into deep loopback simultaneously) or setting register
bit 0.14 (on an individual channel basis). This causes transmit serial data from a channel to be input on the serial
input for that channel, testing the entire path of the serializer/deserializer (except the output driver/receiver).
Shallow loopback may be enabled by setting the shallow loopback register bit 16.9. This will loop the parallel
transmit data bus for an individual channel to the parallel receive bus on that channel.
Far-End Loopback
The TLK2226 can provide a self-test function by enabling the internal far end loop-back (also referred to as
remote loopback) path for a channel with the assertion of MDIO register bit 16.8. The serial data output can be
compared to the serial input data for selected channels to perform functional verification of high speed RX and
TX. The parallel input data during far end loopback test are disregarded. The external parallel outputs are
optionally in a high-impedance state via MDIO control bit (0.10).
Submit Documentation Feedback
43
TLK2226
www.ti.com
SLLS689 – JANUARY 2006
P
a
r
a
l
l
e
l
I
n
t
e
r
f
a
c
e
Per Channel Loopback Definitions
PCS Transmit
PCS Receive
SERDES Transmit
SERDES Receive
Deep Loopback
Shallow Loopback
Far End (Remote) Loopback
S
e
r
i
a
l
I
n
t
e
r
f
a
c
e
Figure 23. Per Channel Loopback Definitions
Power-On Reset
Upon application of minimum valid power, the TLK2226 generates an internal power-on reset. During the
power-on reset the receive data outputs are tri-stated. The length of the power-on reset cycle is dependent upon
the power supply’s ramp curve. The power-on reset is sourced by the digital core supply voltage VDD.
PRBS generator and comparator
The TLK2226 has a built-in 27-1 PRBS (Pseudo Random Bit Stream) self test function available on each
channel. Compared to all 8B/10B data pattern combinations, the PRBS is a worst-case bit pattern. The self-test
function is enabled using the PRBSEN pin or setting the PRBS Enable bit in the MDIO registers. When the
self-test function is enabled, a PRBS is generated and fed into the 10-bit parallel-to-serial converter input
register. Data on the Transmit Data Bus inputs to the TLK2226 are ignored during the PRBS test mode. The
PRBS pattern is then fed through the transmit circuitry as if it were normal data and sent out to the transmitter.
The output can be sent to a BERT (Bit Error Rate Tester) or the receiver of another TLK2226 channel.
Setting the PRBSEN pin has the following effect:
A PRBS datastream is generated on the TX ±ports.
The RX ± ports expect PRBS pattern at their inputs and status signals could be monitored on RD bus or
MDIO. While the PRBS test function is in operation, a real-time signal PRBS_PASS is available on RDx[1] or
RDx[2] of partnering channel depends on whether the rate is 1.25 Gbps or 125Mbps; other signals on the
RDx[4:0] busses have no meaning during PRBS testing. Alternately, the result can be read from the MDIO
vendor specific register 29 and 31. Register 29 is the error counter for high speed(1.25Gbps) PRBS Verifier.
Register 31 is the error counter for low speed(125Mbps) PRBS. These are counters that increment by one
for each received character that has error. Reading the MDIO register clear the counter information in the
MDIO.
The PRBS test function can also be activated through MDIO by activating the PRBS_EN Register 16.7:6.
The PRBS function can be used in conjunction with the serial loopback function. This can be done by asserting
the PRBS generators via the PRBSEN pin or MDIO register 16.7:6, and assertion of the LPBK pin or loopback
control provided via MDIO register 0.14. The result from PRBS verification at the RX ports of the device can be
read from the MDIO registers 29 and 31.
Simultaneous assertion of the PRBS_EN and PRBS_VE register bits in MDIO register 16.7:6, and the serial
loopback enable, puts the device in a test mode which provides PRBS generation and verification at the local
transmitter and receiver internal to the TLK2226
44
Submit Documentation Feedback
TLK2226
www.ti.com
SLLS689 – JANUARY 2006
POWER SUPPLY SEQUENCING
TI ASIC I/O design allows either the core supply (VDD) or the I/O supply (VDDS) to be powered up for an indefinite
period of time while the other power supply is not powered up if all of these constraints are met:
• All maximum ratings and recommended operating conditions
• All warnings about exposure to maximum rated and recommended conditions, particularly junction
temperature. These apply to power transitions as well as normal operation.
• Guidelines for the number and placement of I/O pins
• No bus contention
Bus contention while VDDS is powered up must be limited to 100 hours over the projected lifetime of the device.
Bus contention while VDDS is powered down may violate the absolute maximum ratings.
A supply bus is powered up when the voltage is within the recommended operating range. It is powered down
when it is below that range, either stable or in transition.
Table 33. Device Thermal Characteristics
Package Thermal Theta Values
Airflow (m/s)
Value
0
32.7 (C/W)
1
28.7 (C/W)
θJA
2.5
27.5 (C/W)
θJB
19.8 (C/W)
θJC
14.8 (C/W)
Device Maximum Characteristics
Maximum Junction Temperature
115°C
Maximum Power Dissipation
1.5 Watts
Submit Documentation Feedback
45
TLK2226
www.ti.com
SLLS689 – JANUARY 2006
APPENDIX A
Known issues with the TLK2226 device:
1. ENABLE Input Signal – ENABLE signal must be held high during JTAG operation. Failure to do so will cause
data compare errors on the parallel side data path input signals. A possible solution is that a pull-up resistor
exclusively drives the ENABLE signal on the customer application board. Alternatively, the customer must
guarantee ENABLE is asserted high during JTAG operation.
46
Submit Documentation Feedback
TLK2226
www.ti.com
SLLS689 – JANUARY 2006
APPENDIX B – SGMII 10/100 ODD NIBBLE PACKET/IPG HANDLING
SGMII 10/100 Nibble Processing Assumptions
Case
1
Input (from Mac)
Output(Rate adaptation)
Odd nibble/cycle EXTEND during burst
packet, 3 cycles or more
Assumptions
Odd nibble CARRIER EXTEND (3 or more)
possible during burst. We will stuff or delete to
octet boundary for PCS encoding. Packets are not
corrupted. (1)
clk
en
clk
err
en
or
clk
err
en
err
2
Odd nibble/cycle regular packet
Always deleted to Octet
clk
clk
Packet ends on non octet boundary. We drop the
nibble to form last octet for PCS encoding. (non
valid packet anyway)
en
en
err
err
3
Odd nibble EOP EXTEND
clk
clk
en
Odd nibble EOP EXTEND. We do not know if this
is valid thus output to the PCS will be either NO
EXTEND or EXTEND. (1)
err
en
or
err
clk
en
err
4
Odd nibble ERROR during packet
clk
clk
en
TX_ERR asserted only during 1 nibble. We do not
know if this is a real error, thus output to PCS
encoding will either have an error indicated on a
byte(2 cycles) or no error.
err
en
or
err
clk
en
err
5
1 nibble/cycle EXTEND during burst packet
clk
clk
en
End of Packet extend only 1 nibble during burst.
Output to PCS will either be a corrupted burst with
nibble extended or joined into 1 big corrupted
packet. We cannot always extend as there is no
CTC on the TX side. Following packets will not be
affected
err
en
or
err
clk
en
err
(1)
Note that in the case of an odd EXTEND and if it is stuffed, the PCS TX state machine will embed an |V| (error). The last nibble alone
does not contain enough bits to build a meaningful encoding byte (8’h0F). The packets themselves are not corrupted.
Submit Documentation Feedback
47
TLK2226
www.ti.com
SLLS689 – JANUARY 2006
APPENDIX B – SGMII 10/100 Odd Nibble Packet/IPG Handling (continued)
Case
6
Input (from Mac)
regular packet with odd nibble/cycle
interframe spacing
clk
Output(Rate adaptation)
Interframe gap will be padded or shortened by 1
nibble to meet octet boundary requirements.
Packets will not be corrupted unless interframe
gap falls below 3 nibbles.
clk
en
err
en
err
or
clk
en
err
48
Assumptions
Submit Documentation Feedback
IMPORTANT NOTICE
Texas Instruments Incorporated and its subsidiaries (TI) reserve the right to make corrections, modifications,
enhancements, improvements, and other changes to its products and services at any time and to discontinue
any product or service without notice. Customers should obtain the latest relevant information before placing
orders and should verify that such information is current and complete. All products are sold subject to TI’s terms
and conditions of sale supplied at the time of order acknowledgment.
TI warrants performance of its hardware products to the specifications applicable at the time of sale in
accordance with TI’s standard warranty. Testing and other quality control techniques are used to the extent TI
deems necessary to support this warranty. Except where mandated by government requirements, testing of all
parameters of each product is not necessarily performed.
TI assumes no liability for applications assistance or customer product design. Customers are responsible for
their products and applications using TI components. To minimize the risks associated with customer products
and applications, customers should provide adequate design and operating safeguards.
TI does not warrant or represent that any license, either express or implied, is granted under any TI patent right,
copyright, mask work right, or other TI intellectual property right relating to any combination, machine, or process
in which TI products or services are used. Information published by TI regarding third-party products or services
does not constitute a license from TI to use such products or services or a warranty or endorsement thereof.
Use of such information may require a license from a third party under the patents or other intellectual property
of the third party, or a license from TI under the patents or other intellectual property of TI.
Reproduction of information in TI data books or data sheets is permissible only if reproduction is without
alteration and is accompanied by all associated warranties, conditions, limitations, and notices. Reproduction
of this information with alteration is an unfair and deceptive business practice. TI is not responsible or liable for
such altered documentation.
Resale of TI products or services with statements different from or beyond the parameters stated by TI for that
product or service voids all express and any implied warranties for the associated TI product or service and
is an unfair and deceptive business practice. TI is not responsible or liable for any such statements.
Following are URLs where you can obtain information on other Texas Instruments products and application
solutions:
Products
Applications
Amplifiers
amplifier.ti.com
Audio
www.ti.com/audio
Data Converters
dataconverter.ti.com
Automotive
www.ti.com/automotive
DSP
dsp.ti.com
Broadband
www.ti.com/broadband
Interface
interface.ti.com
Digital Control
www.ti.com/digitalcontrol
Logic
logic.ti.com
Military
www.ti.com/military
Power Mgmt
power.ti.com
Optical Networking
www.ti.com/opticalnetwork
Microcontrollers
microcontroller.ti.com
Security
www.ti.com/security
Mailing Address:
Telephony
www.ti.com/telephony
Video & Imaging
www.ti.com/video
Wireless
www.ti.com/wireless
Texas Instruments
Post Office Box 655303 Dallas, Texas 75265
Copyright  2006, Texas Instruments Incorporated