ANALOG DEVICES AN-746 Service Manual

AN-746

APPLICATION NOTE

One Technology Way • P.O. Box 9106 • Norwood, MA 02062-9106 • Tel: [phone] • Fax: [phone] • www.analog.com

Supporting FDDI with the ADN2812

by Kevin Buckley

INTRODUCTION

The ADN2812 continuous-rate clock and data recovery
device (CDR) can suppor t all NRZ and NRZI random
data patterns between 12.3 Mbps and 2.7 Gbps, and is
protocol agnostic. The default operating mode of the
ADN2812, lock to data mode, is for the CDR to lock on
to any incoming data rate automatically, without the
need of programming or ex ternal reference clocks.
This requires that the incoming data is random NRZ or
NRZI data. In lock to data mode, it is required that the
bit pattern have transitions at the desired frequency
(e.g., 0101), so that the frequency-locked loop can detect
the correct frequency of the bit pattern.

If the data pattern does not meet this requirement, then it
is possible to program the ADN2812 to lock to an external
reference clock so that nonrandom bit patterns can be
accommodated at the desired frequency. An example of
a situation where the ADN2812 needs to be placed in lock
to REFCLK mode is when an FDDI port is attempting to
connect to another FDDI port.

FDDI CONNECTION PROTOCOL

Within every FDDI station there are SMT (station man­agement) entities called PCM (physic al conne ction
management) . The number of PCM entities within a
station is exactly equal to the number of ports that the
station has. This is because each PCM is responsible for
one port.

The PCM entities are the part of SMT that controls the
ports. In order to make a connection, two ports must be
physically connected to each other by means of a fiber­optic or copper cable. When this happens, the PCMs
that are responsible for those ports can recognize each
other’s existence and begin communicating. They do
this by sending line states out of the port and onto the
fiber. The PCM at the other end of the connection will
recognize the line state and respond accordingly.

When the PCM sees another PCM on the other end of
the connection, they will synchronize and communicate
with each other. During this communication, a couple of
important things happen:

1. The PCMs figure out the types of each port, and deter­mine if they are compatible.

2. The PCMs perform an LCT (link confidence test).

REV. 0

The LCT determines if the quality of the link is good
enough to establish a connection on. If it is not, the PCMs
will not make a connection. If everything goes well dur­ing this signaling, the PCMs will establish a connection
and place the ports on the token path that goes through
that station. At this point, data (in the form of frames)
can be sent through these ports and the por ts become
part of the network.

PCM entities have a number of internal states that they
can be in. While in any state, the PCM has the port send
out a certain line state. This line state will be received by
the PCM on the other side of the connection. The other
PCM entity (at the other end of the connection) will be
able to tell which state the original PCM is in. PCMs use
this information to signal data across the connection.

The connection attempt begins with the PCM state:
break. The break state is the entry point in the start
of a PCM connection. In the break state, a continuous
stream of Q symbols (QLS; 00000) is transmitted to
force the other end of the connection to break any exist­ing connection and restart the connection initialization
sequence. The break state is entered upon the star t of a
connection attempt, or if there was a problem with the
previous connection attempt.

Once the PCM has begun to receive the correct symbols
from the other side, the link will transition to the connect
state. The connect state is used to synchronize the ends
of the connection for the signaling sequence.

In the connect state, the PCM transmits a continuous
stream of H symbols (HLS; 00100). The PCM stays in
this state until it receives HLS from the other end of
the connection. The reception of HLS indicates that the
other PCM is also in the connect state. If the PCM is not
attached to anything at all, then it will be receiving QLS
and will remain here until it is connected to another PCM
that sources HLS.

Once the connect state is successful, the link transitions
to the next state. While in the next state, a continuous
stream of I symbols is transmitted (ILS; 11111). The next
state is used to separate the signaling performed in the
signal state and it is where the LCT (link confidence test)
is performed.

AN-746

0 0 00 0 0 00 0 0 00 0 0 00

NRZ

Q

0

00 0

QQQ

BIT PATTERN

NRZI

NRZ

BIT PATTERN

NRZI

0 0 01 0 0 01 0 0 01 0 0 01

H

0

00 0

HHH

NRZ

BIT PATTERN

NRZI

1 1 11 1 1 11 1 1 11 1 1 11

I

1

11 1

III

NRZI ENCODING

FDDI uses NRZI encoding. The ADN2812 inherently sup­ports NRZI encoded data without the need of reference
clocks or any external programming as long as the NRZI
data is random. Due to the nonrandom nature of the
FDDI connection protocol in addition to the NRZI encod­ing of the data, it is required that the ADN2812 be put into
lock to REFCLK mode to support FDDI.

In order to understand why the ADN2812 needs to be
used in lock to REFCLK mode to support this FDDI pro­tocol, an explanation of NRZI encoded data is required.
NRZI encoded data is a type of NRZ data that is encoded
such that long strings of consecutive ones are broken up
with transitions. This provides more transitions to CDR
(clock data recovery) devices which make it easier for the
CDR to lock and stay locked to incoming NRZ type pat­terns. A NRZI encoded data pattern will have a transition
every time there is a 1 in the pattern and there will be no
transition every time there is a 0. This is illustrated in the
following diagrams which show patterns of consecutive
Q, H, and I symbols in an FDDI link.

Figure 1. Consecutive “Q” Symbols During PCM
State: Break

Figure 2. Consecutive “H” Symbols During PCM
State: Connect

Figure 3. Consecutive “I” Symbols During PCM
State: Next

ADN2812 DEFAULT MODE: LOCK TO DATA

The ADN2812 is a con tinuous rate clock and data
recovery device that is designed to automatically lock
on to any random NRZ or NRZI data pattern between

12.3 Mbps and 2.7 Gbps. The frequency-locked loop of
the ADN2812 frequency locks the VCO to the incoming
bit stream and then the delay and phase-locked loop
aligns the phase of the VCO with the incoming bit stream.
This is the default mode of operation of the ADN2812.
There is no need for programming or any ex ternal
reference clocks to operate in this mode.

The reason that this default mode cannot support the
FDDI protocol is because, during the initial PCM states,
the data that is presented to the ADN2812 is not ran­dom and it does not contain any transitions at the FDDI
data rate of 125 Mbps. During the PCM states, the data
that is presented to the ADN2812 is conse cutive Q
symbols followed by consecutive H symbols, followed
by consecutive I symbols in that order. The period of
time that each of these states lasts could be hundreds
to thousands of milliseconds.

The first data presented to the ADN2812 is the Q symbols,
which are all zeroes (Figure 1). Obviously, the ADN2812
will not lock to this when it’s in lock to data mode since
there are no transitions on the data. The subsequent
data that the ADN2812 sees will be the consecutive H
symbols. The H symbols, with NRZI encoding, look like
a 25 Mbps NRZ data pattern of 10101010 … (Figure 2).
Therefore, in lock to data mode, the ADN2812 will lock
to 25 Mbps during the transmission of the consecutive
H symbols. When the link transitions to the PCM state:
next, consecutive I symbols are present at the input to
the ADN2812. NRZI encoded, the I symbols will look like
a true 125 Mbps bit pattern (Figure 3). The ADN2812’s
lock detector will sense that the frequency of the input
data has changed from 25 Mbps to 125 Mbps and a new
frequency acquisition will be initiated. At this point, the
ADN2812 will lock to 125 Mbps. This temporary loss of
synchronization by the ADN2812 is unacceptable. The
result is that the FDDI link will never be able to achieve
a valid connection.

SOLUTION: LOCK TO REFCLK MODE

A reference clock is not required to per form clock and
data recover y on random NRZ and NRZI data patterns
with the ADN2812. However, suppor t for an optional
reference clock is provided for situations where there is
extended periods of time where the input data is non­random, repetitive, and does not have transitions at the
desired frequency.

In lock to REFCLK mode, the ADN2812 uses an external
reference clock as an acquisition aid. The VCO will be
frequency locked with respect to an external reference
clock and the lock detector will compare the VCO fre­quency against the reference clock frequency, not the
input data. Using the ADN2812 in this mode will allow the
part to lock to 125 Mbps when the consecutive H symbols
are being received at the ADN2812 inputs even though
there are no transitions at 125 Mbps present. When the
PCM state transitions to the next state, the ADN2812 will
already be locked at the correct frequency, 125 Mbps.
When the input data switches from consecutive H sym­bols (PCM state: connect) to consecutive I symbols (PCM
state: next) the ADN2812 will simply stay locked at 125
Mbps and there will not be a loss of synchronization.

–2–

REV. 0