Interface for T1 and E1 Applications
Provides Line Driver, Jitter Attenuator
••
and Clock Recovery Functions
Transmit Side Jitter Attenuation
••
Starting at 3 Hz, with > 300 UI of Jitter
Tolerance
B8ZS/HDB3/AMI Encoders/Decoders
••
Compatible with SONET, M13 , CCITT
••
G.742, and Other Asynchronous Muxes
50 mA Transmitter Short-Circuit
••
Current Limiting
General Description
The CS61305A combines the complete analog transmit
and receive line interface for T1 or E1 applications in a
low power, 28-pin device operating from a +5V supply.
The CS61305A is a pin-compatible replacement for the
LXT305A in most applications.
The CS61305A provides a transmitter jitter attenuator
making it ideal for use in asynchronous multiplexor
systems with gapped transmit clocks. The transmitter
features internal pulse shaping and a low impedance
output stage allowing the use of external resistors for
transmitter impedance matching. The receiver uses a
digital Delay-Locked-Loop clock and data recovery circuit which is continuously calibrated from a crystal
reference to provide excellent stability and jitter tolerance.
Applications
•
Interfacing network transmission equipment such as
SONET multiplexor and M13 to a DSX-1 cross connect.
TV+
Input Voltage, Any Pin(Note 1)V
Input Current, Any Pin(Note 2)I
Ambient Operating TemperatureT
Storage TemperatureT
in
in
A
stg
-
-
6.0
(RV+) + 0.3
RGND-0.3(RV+) + 0.3V
-1010mA
-4085
-65150
WARNING: Operations at or beyond these limits may result in permanent damage to the device.
Normal operation is not guaranteed at these extremes.
Notes: 1. Excluding RTIP, RRING, which must stay within -6V to (RV+) + 0.3V.
2. Transient currents of up to 100 mA will not cause SCR latch-up. Also TTIP, TRING, TV+ and TGND
can withstand a continuous current of 100 mA.
RECOMMENDED OPERATING CONDITIONS
ParameterSymbolMinTypMaxUnits
DC Supply(Note 3) RV+, TV+4.755.05.25V
Ambient Operating TemperatureT
Power Consumption(Notes 4,5)P
Notes: 3. TV+ must not exceed RV+ by mor e than 0.3V.
4. Power consumption while driving line load over oper ating temperature range. Includes IC and load.
Digital input levels are within 10% of the supply rails and digital outputs are driving a 50 pF
capacitive load.
5. Assumes 100% ones density and maximum line length at 5.25V.
A
C
-402585
--350mW
V
V
°C
°C
°C
DIGITAL CHARACTERISTICS (TA = -40°C to 85°C; TV+, RV+ = 5.0V ±5%; GND = 0V)
1:1.26 transformer and 120Ω load
Load Presented To Transmitter Output(Note 10)-75Jitter Added by the Transmitter(Note 15)
10Hz - 8kHz
8kHz - 40kHz
10Hz - 40kHz
Broad Band
Power in 2kHz band about 772kHz(Notes 10, 16)12.61517.9dBm
Power in 2kHz band about 1.544MHz(Notes 10, 16)
(referenced to power in 2kHz band at 772kHz)
Positive to Negative Pulse Imbalance(Notes 10, 16)
T1, DSX-1
E1 amplitude at center of pulse
E1 pulse width at 50% of nominal amplitude
E1 Transmitter Return Loss(Notes 10, 16, 17)
51 kHz to 102 kHz
102 kHz to 2.048 MHz
2.048 MHz to 3.072 MHz
E1 Transmitter Short Circuit Current(Notes 10, 18)--50mA RMS
Notes: 10. Using a 0.47 µF capacitor in series with the primary of a tr ansformer recommended
in the Applications Section.
11. Pulse amplitude measured at the output of a 1:1 tr ansformer across a 75 Ω load for
line length setting LEN2/1/0 = 0/0/0.
12. Pulse amplitude measured at the output of a 1:1.26 tr ansformer across a 120 Ω load for line length
setting LEN2/1/0 = 0/0/0 or at the output of a 1:1 transformer across a 120 Ω load for LEN2/1/0 = 001.
13. Pulse amplitude measured at the output of a 1:1.15 tr ansformer across a 100 Ω load for
line length setting LEN2/1/0 = 0/1/0.
14. Pulse amplitude measured at the DSX-1 Cross -Connect across a 100 Ω load for all line length
settings from LEN2/1/0 = 0/1/1 to LEN2/1/0 = 1/1/1 using a 1:1.5 transformer.
15. Input signal to RTIP/RRING is jitter free. Values will reduce slightly if jitter free clock is input to TCLK.
16. Not production tested. Parameters guaranteed by design and characterization.
17. Return loss = 20 log
= impedance of line load. Measured with a repeating 1010 data pattern with LEN2/1/0 = 0/0/0
z
0
and a 1:2 transformer with two 9.4 Ω series res istors terminated by a 75Ω load,
or for LEN2/1/0 = 0/0/1 with a 1:2 transformer and two 15 Ω ser ies resistors terminated by a
120Ω load.
18. Measured broadband through a 0.5 Ω resistor across the secondary of the transmitter transformer
during the transmission of an all ones data pattern for LEN2/1/0 = 0/0/0 or 0/0/1with a 1:2 transformer
and the series resistors specified in Table A1.
ABS((z1 +z0)/(z1-z0)) where z1 = impedance of the transmitter, and
10
2.14
2.7
2.7
2.4
-0.237
-0.3
-
-
-
-
-29-38-dB
-
-5
-5
20
20
20
2.37
3.0
3.0
3.0
-
-
-
-
-
-
0.2
-
-
28
28
24
2.6
3.3
3.3
3.6
0.237
0.3
0.01
0.025
0.025
0.05
0.5
5
5
-
-
-
V
V
V
V
V
V
Ω
UI
UI
UI
UI
dB
%
%
dB
dB
dB
DS157PP33
CS61305A
ANALOG SPECIFICATIONS (TA = -40°C to 85°C; TV+, RV+ = 5.0V ±5%; GND = 0V)
T1, (FCC Part 68) and E1(Note 21)
Allowable Consecutive Zeros before LOS160175190bits
Receiver Input Jitter Tolerance(Note 22)
10kHz - 100kHz
2kHz
10Hz and below
Loss of Signal Threshold-0.30-V
Notes: 19. Attenuation measured with input jitter equal to 3/4 of measured jitter tolerance. Circuit attenuates
jitter at 20 dB/decade above the corner frequency. See Figure 10. Output jitter can increase
significantly when more than 12 UI’s are input to the attenuator. See discussion in the text section.
20. For input amplitude of 1.2 V
21. For input amplitude of 1.05 V
to 4.14 Vpk.
pk
to 3.3 Vpk.
pk
22. Jitter tolerance increases at lower frequencies. See Figure 12.
Rise Time, All Digital Outputs(Note 26)t
Fall Time, All Digital Outputs(Note 26)t
TCLK Frequencyf
TCLK Pulse Width(Notes 16, 27, 28)
c
pwh3/tpw3
aclki
pw1
t
pwh1
t
pwl1
r
f
tclk
t
pwh2
(Notes 29, 30)
TPOS/TNEG (TDATA) to TCLK Falling Setup Timet
TCLK Falling to TPOS/TNEG (TDATA) Hold Timet
RPOS/RNEG Valid Before RCLK Falling(Note 27)t
RDATA Valid Before RCLK Falling(Note 29)t
RPOS/RNEG Valid Before RCLK Rising(Note 28)t
RPOS/RNEG Valid After RCLK Falling(Note 27)t
RDATA Valid After RCLK Falling(Note 29)t
RPOS/RNEG Valid After RCLK Rising(Note 28)t
su2
h2
su1
su1
su1
h1
h1
h1
Notes: 23. Crystal must meet specifications described in CXT6176/CXT8192 data sheet.
24. ACLKI provided by an external source or TCLK but not RCLK.
25. RCLK duty cycle will vary with extent by which pulses are displaced by jitter. Specified under worst
case jitter conditions: 0.4 UI AMI data displacement for T1 and 0.2 UI AMI data displacement for E1.
Key Enhancements of the CS61305A Relative
to the LXT305A
•12.5% lower power consumption,
•50 mA
transmitter short-circuit current
RMS
limiting for E1 (per OFTEL OTR-001),
•Optional AMI, B8ZS, HDB3 encoder/de-
coder or external line coding support,
•Receiver AIS (unframed all ones) detection,
•Improved receiver Loss of Signal handling
(LOS set at power-up, reset upon receipt of
3 ones in 32 bit periods with no more than
15 consecutive zeros),
•Transmitter TTIP and TRING outputs are
forced low when TCLK is static.
Introduction to Operating Modes
The CS61305A supports three operating modes
which are selected by the level of t he MODE pin
as shown in Tables 1 and 2, Figure 7, and Figures
A1-A3 of the Applications section.
There are thirteen multi-function pins whose
functionality is determined by the operating
mode. (see Table 2). The modes are Hardware
Mode, Extended Hardware Mode, and Host
Mode. In Hardware and Extended Hardware
Modes, discrete pins are used to configure and
monitor the device. The Extended Hardware
Mode provides a parallel chip select input which
latches the control inputs allowing individual ICs
to be configured using a common set of control
lines. In the Host Mode, an external processor
monitors and configures the device through a serial interface.
The transmitter takes digital T1 or E1 input data
and drives appropriately shaped bipolar pulses
onto a transmission line. The transmit data (TPOS
& TNEG or TDATA) is supplied synchronously
and sampled on the falling edge of the input
clock, TCLK.
Either T1 (DSX-1 or Network Interface) or E1
CCITT G.703 pulse shapes may be selected.
Pulse shaping and signal level are controlled by
"line length select" inputs as shown in Table 3.
The output options in Table 3 are specified with a
1:1.15 transmitter transformer turns ratio for T1
and a 1:1 turns ratio for E1 without external series resistors. Other turns ratios may be used if
approriate resistors are placed in series with the
TTIP and TRING pins. Table A1 in the applications section lists other combinations which can
be used to provide transmitter impedance matching.
For T1 DSX-1 applications, line lengths from 0 to
655 feet (as measured from the transmitter to the
DSX-1 cross connect) may be selected. The five
partition arrangement in Table 3 meets ANSI
T1.102-1993 and AT&T CB-119 requirements
when using #22 ABAM cable. A typical output
pulse is shown in Figure 8. These pulse settings
can also be used to meet CCITT pulse shape requirements for 1.544 MHz operation.
LEN2 LEN1 LEN0Option SelectedApplication
0110-133 ft
100133-266 ft
101266-399 ft
110399-533 ft
111533-655 ft
000
001
010FCC PART 68, OPT. ANetwork
011ANSI T1.403
Table 3. Line Length Selection
75Ω coax
120Ω twisted-pair
DSX-1
ABAM
(AT&T 600B
or 600C)
E1
CCITT G.703
Interface
The CS61305A transmitter provides short-circuit
current limiting protection and meets OFTEL
OTR-001 short-circuit current limiting requirements for E1 applications.
The CS61305A will detect a static TCLK, and
will force TTIP and TRING low to prevent transmission when data is not present. When any
transmit control pin (TAOS, LEN0-2 or LLOOP)
is toggled, the transmitter outputs will require approximately 22 bit periods to stabilize. The
transmitter will take longer to stabilize when
RLOOP is selected because the timing circuitry
must adjust to the new frequency.
NORMALIZED
AMPLITUDE
1.0
ANSI T1.102,
AT&T CB 119
SPECIFICATIONS
For T1 Network Interface applications, two additional options are provided. Note that the optimal
0.5
pulse width for Part 68 (324 ns) is narrower than
the optimal pulse width for DSX-1 (350 ns). The
CS61305A automatically adjusts the pulse width
based upon the "line length" selection made.
The E1 G.703 pulse shape is supported with line
length selections LEN2/1/0 = 0/0/0 and 0/0/1.
The pulse width will meet the G.703 pulse shape
template shown in Figure 9, and specified in Table 4.
10DS157PP3
0
OUTPUT
PULSE SHAPE
-0.5
02507501000
Figure 8. Typical Pulse Shape at DSX-1 Cross Connect
500
TIME (nanoseconds)
120
110
100
90
80
50
Percent of
nominal
peak
voltage
269 ns
244 ns
194 ns
Attenuation in dB
0
10
20
30
40
50
CS61305A
a) Minimum Attenuation Limit
AT&T 62411
Requirements
b) Maximum
Attenuation
Limit
-10
-20
10
0
219 ns
488 ns
Nominal Pulse
Figure 9. Mask of the Pulse at the 2048 kbps Interface
Transmit All Ones Select
The transmitter provides for all ones insertion at
the frequency of ACLKI. Transmit all ones is selected when TAOS goes high, and causes
continuous ones to be transmitted on the line
(TTIP and TRING). In this mode, the TPOS and
TNEG (or TDATA) inputs are ignored. A TAOS
request will be ignored if remote loopback is in
effect. ACLKI jitter will be attenuated. TAOS is
not available on the CS61305A when ACLKI is
grounded.
60
1101001 k10 k
Frequency in Hz
Measured Performance
Figure 10. Typical Jitter Attenuation Curve
Jitter Attenuator
The jitter attenuator is designed to reduce wander
and jitter in the transmit clock signal. It consists
of a 192 bit FIFO, a crystal oscillator, a set of
load capacitors for the crystal, and control logic.
The jitter attenuator exceeds the jitter attenuation
requirements of Publications 43802 and REC.
G.742. A typical jitter attenuation curve is shown
in Figure 10.
The jitter attenuator works in the following manner. Data on TPOS and TNEG (or TDATA) are
For coaxial cable,
75Ω load and
transformer specified
in Application Section.
For shielded twisted
pair, 120Ω load and
transformer specified
in Application Section.
Nominal peak voltage of a mark (pulse)2.37 V3 V
Peak voltage of a space (no pulse)
0 ±0.237 V0 ±0.30 V
Nominal pulse width244 ns
Ratio of the amplitudes of positive and negative
pulses at the center of the pulse interval
Ratio of the widths of positive and negative
pulses at the nominal half amplitude
0.95 to 1.05*
0.95 to 1.05*
* When configured with a 0.47 µF nonpolarized capacitor in series with the TX transformer
primary as shown in Figures A1, A2 and A3.
Table 4. CCITT G.703 Specifications
DS157PP311
CS61305A
written into the ji tt er at tenu at o r’s FIFO by TCLK.
The rate at which data is read out of the FIFO and
transmitted is determined by the oscillator. Logic
circuits adjust the capacitive loading on the crystal to set its oscillation frequency to the average
of the TCLK frequency. Signal jitter is absorbed
in the FIFO.
Jitter Tolerance of Jitter Attenuator
The FIFO in the jitter attenuator is designed to
neither overflow nor underflow. If the jitter amplitude becomes very large, the read and write
pointers may get very close together. Should the
pointers attempt to cross, the oscillator’s divide
by four circuit adjusts by performing a divide by
3 1/2 or divide by 4 1/2 to prevent the overflow
or underflow. When a divide by 3 1/2 or 4 1/2
occurs, the data bit will be driven on to the line
either an eighth bit period early or an eighth bit
period late.
The FIFO of the jitter attenuator in the transmit
path is 192 bits deep. This FIFO will typically be
near the half full point under normal operating
conditions, buffering about 96 bits of data. The
number of bits actually buffered depends on the
relationship of the nominal TCLK frequency to
the center frequency of the crystal oscillator. As
these frequencies deviate, a few bits of FIFO
depth will be lost.
TCLK can have gaps or bursts. As long as the gap
or burst is less than the remaining FIFO depth,
normal operation will continue. For example, if
the nominal TCLK frequency was less than the
oscillator’s center frequency by 40 Hz. The FIFO
will operate 3-4 bits off center or 92 bits full. A
gap in TCLK of 80 cycles would empty the FIFO
by 80 bits but would still not envoke the divide
by 4 1/2 circuitry, as about 12 bits would remain
in the FIFO.
The crystal frequency must be 4 times the nominal signal frequency: 6.176 MHz for 1.544 MHz
operation; 8.192 MHz for 2.048 MHz applica-
tions. Internal capacitors load the crystal, controlling the oscillation frequency. The crystal must be
designed so that over operating temperature, the
oscillator frequency range exceeds the system frequency tolerance. Crystal Semiconductor offers
the CXT6176 & CXT8192 crystals, which yield
optimum performance with the CS61305A.
Receiver
The receiver extracts data and clock from an AMI
(Alternate Mark Inversion) coded signal and outputs clock and synchronized data. The receiver is
sensitive to signals over the entire range of cable
lengths and requires no equalization or ALBO
(Automatic Line Build Out) circuits. The signal is
received on both ends of a center-tapped, centergrounded transformer. The transformer is
center-tapped on the IC side. The clock and data
recovery circuit exceeds the jitter tolerance specifications of Publications 43802, 43801, 62411
amended, TR-TSY-000170, and CCITT REC.
G.823.
A block diagram of the receiver is shown in Figure 11. The two leads of the transformer (RTIP
and RRING) have opposite polarity allowing the
receiver to treat RTIP and RRING as unipolar signals. Comparators are used to detect pulses on
RTIP and RRING. The comparator thresholds are
dynamically established at a percent of the peak
level (50% of peak for E1, 65% of peak for T1;
with the slicing level selected by LEN2/1/0).
The receiver uses an edge detector and a continuously calibrated delay line to generate the
recovered clock. The delay line divides its reference clock, ACLKI or the jitter attenuator’s
oscillator, into 13 equal divisions or phases. Continuous calibration assures timing accuracy, even
if temperature or power supply voltage fluctuate.
The leading edge of an incoming data pulse triggers the clock phase selector. The phase selector
chooses one of the 13 available phases which the
delay line produces for each bit period. The out-
12DS157PP3
CS61305A
RTIP
1 : 2
RRING
Data
Level
Slicer
Edge
Detector
Figure 11. Receiver Block Dia gram
put from the phase selector feeds the clock and
data recovery circuits which generate the recovered clock and sample the incoming signal at
appropriate intervals to recover the data. The jitter
tolerance of the receiver exceeds that shown in
Figure 12.
The CS61305A outputs a clock immediately upon
power-up and will lock onto the AMI data input
immediately. If loss of signal occurs, the RCLK
frequency will equal the ACLKI frequency.
Data
Sampling
&
Clock
Extraction
Clock
Phase
Selector
Continuously
Calibrated
Delay Line
RPOS
RNEG
RCLK
ACLKI or
Oscillator in Jitter
Attenuator
In the Hardware Mode, data at RPOS and RNEG
is stable and may be sampled on the rising edge
of the recovered clock. In the Extended Hardware
Mode, data at RDATA is stable and may be sampled on the fallings edge of the recovered clock.
In the Host Mode, CLKE determines the clock
polarity for which output data is stable and valid
as shown in Table 5.
Jitter and Recovered Clock
The CS61305A is designed for error free clock
and data recovery from an AMI encoded data
Minimum
Performance
101k10k1100100k700
JITTER FREQUENCY(Hz)
300
PEAK-TO-PEAK
JITTER
(unit intervals)
300
138
100
AT&T 62411
28
10
1
.4
.1
Figure 12. Minimum Input Jitter Tolerance of Receiver
MODE
(pin 5)
LOW
(<0.2V)
HIGH
(>(V+) - 0.2V)
HIGH
(>(V+) - 0.2V)
MIDDLE
(2.5V)
X = Don’t care
Table 5. Data Output/Clock Relationship
CLKE
(pin 28)
XRPOS
LOWRPOS
HIGHRPOS
DATACLOCKClock Edge
for Valid Data
RCLK
RNEG
RCLK
RCLK
RNEG
SDO
RCLK
SCLK
RCLK
RNEG
SDO
RCLK
SCLK
XRDATARCLKFalling
Rising
Rising
Rising
Rising
Falling
Falling
Falling
Rising
DS157PP313
CS61305A
stream in the presence of more than 0.4 unit intervals of jitter at high frequency. The clock
recovery circuit is also tolerant of long strings of
zeros. The edge of an incoming data bit causes
the circuitry to choose a phase from the delay line
which most closely corresponds with the arrival
time of the data edge, and that clock phase triggers a pulse which is typically 140 ns in duration.
This phase of the delay line will continue to be
selected until a data bit arrives which is closer to
another of the 13 phases, causing a new phase to
be selected. The largest jump allowed along the
delay line is six phases.
When an input signal is jitter free, the phase selection will occasionally jump between two
adjacent phases resulting in RCLK jitter with an
amplitude of 1/13 UIpp. These single phase
jumps are due to differences in frequency of the
incoming data and the calibration clock input to
ACLKI. For T1 operation the instantaneous period can be 14/13 * 648 ns = 698 ns or 12/13 *
648 ns = 598 ns when adjacent clock phases are
chosen. As long as the same phase is chosen, the
period will be 648 ns. Similar calculations hold
for the E1 rate.
The clock recovery circuit is designed to accept at
least 0.4 UI of jitter at the receiver. Since the data
stream contains information only when ones are
transmitted, a clock/data recovery circuit must assume a zero when no signal is measured during a
bit period. Likewise, when zeros are received, no
information is present to update the clock recovery circuit regarding the trend of a signal which is
jittered. The result is that two ones that are separated by a string of zeros can exhibit maximum
deviation in pulse arrival time. For example, one
half of a period of jitter at 100 kHz occurs in 5
µs, which is 7.7 T1 bit periods. If the jitter amplitude is 0.4 UI, then a one preceded by seven zeros
can have maximum displacement in arrival time,
i.e. either 0.4 UI too early or 0.4 UI too late. The
data recovery circuit correctly assigns a received
bit to its proper clock period if i t is displaced by
less than 6/13 of a bit period from its optimal location. Theoretically, this would give a jitter
tolerance of 0.46 UI. The actual jitter tolerance of
the CS61305A is only slightly less than the ideal.
In the event of a maximum jitter hit, the RCLK
clock period immediately adjusts to align itself
with the incoming data and prepare to accurately
place the next one, whether it arrives one period
later, or after another string of zeros and is displaced by jitter. For a maximum early jitter hit,
RCLK will have a period of 7/13 * 648 ns = 349
ns. For a maximum late jitter hit, RCLK will have
a period of 19/13 * 648 ns = 947 ns.
Loss of Signal
Receiver loss of signal is indicated upon receiving 175 consecutive zeros. A digital counter
counts received zeros based on RCLK cycles. A
zero input is determined either when zeros are received, or when the received signal amplitude
drops below a 0.3 V peak threshold.
The receiver reports loss of signal by setting the
Loss of Signal pin, LOS, high. If the serial interface is used, the LOS bit will be set and an
interrupt issued on INT. LOS will go low (and
flag the INT pin again if serial I/O is used) when
a valid signal is detected. Note that in the Host
Mode, LOS is simultaneously available from both
the register and pin 12.
In a loss of signal state, the RCLK frequency will
be equal to the ACLKI frequency since ACLKI is
being used to calibrate the clock recovery circuit.
Received data is output on RPOS/RNEG regardless of LOS status. LOS returns to logic zero
when 3 ones are received out of 32 bit periods
containing no more than 15 consecutive zeros.
Also, a power-up or manual reset will set LOS
high.
14DS157PP3
Loading...
+ 30 hidden pages
You need points to download manuals.
1 point = 1 manual.
You can buy points or you can get point for every manual you upload.