ETHERNET-10 [ETC]
Ethernet-10 Core technical manual 2/97 ; 以太网- 10的核心技术手册2/97\n型号: | ETHERNET-10 |
厂家: | ETC |
描述: | Ethernet-10 Core technical manual 2/97
|
文件: | 总100页 (文件大小:452K) |
中文: | 中文翻译 | 下载: | 下载PDF数据表文档文件 |
Ethernet-10 Core
Technical Manual
February 1997
®
Order Number R14006
ii
February 1997 - Rev. A
Copyright © 1994-1997 by LSI Logic Corporation. All rights reserved.
This document is preliminary. As such, it contains data derived from functional
simulations and performance estimates. LSI Logic has not verified either the
functional descriptions, or the electrical and mechanical specifications using
production parts.
Document DB14-000063-00, First Edition (February 1997)
This document describes revision A of LSI Logic Corporation’s Ethernet-10
(E-10) Core and will remain the official reference source for all revisions/releases
of this product until rescinded by an update.
To receive product literature, call us at 1.800.574.4286 (U.S. and Canada);
+32.11.300.531 (Europe); 408.433.7700 (outside U.S., Canada, and Europe)
and ask for Department JDS; or visit us at http://www.lsilogic.com.
LSI Logic Corporation reserves the right to make changes to any products herein
at any time without notice. LSI Logic does not assume any responsibility or
liability arising out of the application or use of any product described herein,
except as expressly agreed to in writing by LSI Logic; nor does the purchase or
use of a product from LSI Logic convey a license under any patent rights,
copyrights, trademark rights, or any other of the intellectual property rights of LSI
Logic or third parties.
Copyright © 1994-1997 by LSI Logic Corporation. All rights reserved.
TRADEMARK ACKNOWLEDGMENT
LSI Logic logo design, CoreWare, are registered trademarks and G10, and Right-
First-Time are trademarks of LSI Logic Corporation. All other brand and product
names may be trademarks of their respective companies.
Ethernet-10 (E-10) Core Technical Manual
iii
February 1997 - Rev. A
Copyright © 1994-1997 by LSI Logic Corporation. All rights reserved.
iv
February 1997 - Rev. A
Copyright © 1994-1997 by LSI Logic Corporation. All rights reserved.
Contents
Preface
Chapter 1
Introduction
1.1
1.2
CoreWare Program
Local Area Networks Overview
1.2.1 Ethernet Standards
1-1
1-3
1-3
1.3
1.4
General Description
Features
1-4
1-8
1.4.1
1.4.2
1.4.3
1.4.4
1.4.5
General Features and Benefits
1-8
Media Access Control (MAC) Features
Encoder/Decoder (ENDEC) Features
Twisted-Pair Transceiver Interface Features
Attachment Unit Interface Features
1-8
1-9
1-9
1-10
1-10
1.5
Other Key E-10 Features
Chapter 2
Signal Descriptions
2.1
2.2
2.3
2.4
2.5
2.6
2.7
2.8
2.9
Logic Symbol
2-1
2-3
Transceiver Interface Signals
Receiver Signals
2-6
Receiver Status Signals
Multicast Filter Signals
Transmitter Signals
2-8
2-9
2-10
2-13
2-15
2-17
Transmitter Status Signals
General MAC Signals
Timing and Test Signals
Chapter 3
E-10 Functional Description
3.1 Media Access Control (MAC)
3-1
v
Ethernet-10 (E-10) Core Technical Manual
February 1997 - Rev. A
Copyright © 1994-1997 by LSI Logic Corporation. All rights reserved.
3.1.1
3.1.2
3.1.3
Frame Data Encapsulation and Decapsulation
Frame Transmission
3-1
3-4
Frame Reception
3-9
3.2
3.3
Encoder/Decoder (ENDEC)
3.2.1 PLL Characteristics
Twisted-Pair Interface
3-13
3-13
3-14
3-14
3-16
3-19
3-20
3-20
3-20
3-21
3-23
3-25
3.3.1
3.3.2
3.3.3
3.3.4
3.3.5
4-Wire Twisted-Pair Interface
8-Wire Twisted-Pair Interface
Squelch
Link Test Function
Polarity Detection And Correction
3.4
Attachment Unit Interface (AUI)
3.4.1
3.4.2
3.4.3
6-Wire AUI Interface
9-Wire AUI Interface
Signal Quality Error (SQE) Function (Heartbeat)
Chapter 4
Functional Timing
4.1
MAC Transmission and Reception Timing
4-1
4-1
4.1.1
4.1.2
4.1.3
MAC Transmission Timing
MAC Reception Timing
ENDEC Timing
4-5
4-9
4.2
4.3
E-10 Clock Timing
4-10
4-11
4-11
4-11
4-12
4-12
Multiple Core Clocking
4.3.1
4.3.2
4.3.3
4.3.4
Core Clocks
80-MHz Clock
10- and 20-MHz Clocks
E-10 Testing
Chapter 5
Specifications
5.1
5.2
5.3
Derivation of AC Timing and Loading
AC Timing
5-1
5-1
5-4
Pin Summary
Appendix A
Glossary
Customer Feedback
vi
Contents
February 1997 - Rev. A
Copyright © 1994-1997 by LSI Logic Corporation. All rights reserved.
Figures
1.1
1.2
2.1
3.1
3.2
3.3
3.4
3.5
3.6
3.7
3.8
3.9
Ethernet-10 MAC Relationship to the OSI Model
E-10 Block Diagram
1-3
1-7
E-10 Logic Symbol
2-2
MAC Frame Format
3-2
Alternate One-Zero Manchester Encoding
Host Multicast Filter Implementation
4-Wire Twisted-Pair Interface Using Mixed Signal Cores
8-Wire Twisted-Pair Interface
Manchester Encoding
3-3
3-11
3-15
3-17
3-18
3-19
3-21
3-22
3-24
4-2
Squelch Signal
10BASE-5 AUI Connection
AUI 6-Wire Interface
3.10 AUI 9-Wire Interface
4.1
4.2
4.3
4.4
4.5
4.6
4.7
4.8
4.9
Successful Transmission
Transmission Aborted by a Collision
Transmission Aborted by TABORTP
Successful Reception
4-3
4-4
4-5
Reception With No Individual Address Filter Match
Reception With External Multicast Filter Match
Reception With No External Multicast Filter Match
Twisted Pair Serial Output Data
4-6
4-7
4-8
4-9
Twisted Pair Link Integrity Pulses
4-9
4.10 E-10 Clock Buffers
4-10
4-13
5-4
4.11 Multiple-Core Clock Multiplexing
5.1
5.2
A.1
E-10 Core Setup and Hold Timing Diagram
E-10 Core Reset Timing Diagram
Router OSI Interconnect Level
5-4
A-7
Tables
5.1
5.2
E-10 Core AC Timing Parameters
E-10 Pin Summary
5-2
5-5
Ethernet-10 (E-10) Core Technical Manual
vii
February 1997 - Rev. A
Copyright © 1994-1997 by LSI Logic Corporation. All rights reserved.
viii
Contents
February 1997 - Rev. A
Copyright © 1994-1997 by LSI Logic Corporation. All rights reserved.
Preface
This book is the primary reference and Technical Manual for the
Ethernet-10 Core. It contains a complete functional description for the
E-10 core and includes complete physical and electrical specifications.
Audience
This document assumes that you have some familiarity with
microprocessors and related support devices. The people who benefit
from this manual are:
♦
Engineers and managers who are evaluating the E-10 core for
possible use in value-added local area network (LAN) applications,
such as network hubs, routers, and adapter cards
♦
Engineers with motherboard designs requiring system integration of
an Ethernet core function
Organization
This document has the following chapters and appendixes:
♦
♦
♦
♦
♦
Chapter 1, Introduction, describes the general characteristics and
capabilities of the E-10 core.
Chapter 2, Signal Descriptions, describes each of the input and
output signals of the E-10 core.
Chapter 3, E-10 Functional Description, provides information on
the architecture and functional blocks of the E-10 core.
Chapter 4, Functional Timing, provides a detailed description of the
E-10 Media Access Control (MAC) timing and the E-10 clock timing.
Chapter 5, Specifications, contains the complete electrical and
mechanical specifications of the E-10 core.
Ethernet-10 (E-10) Core Technical Manual
ix
February 1997 - Rev. A
Copyright © 1994-1997 by LSI Logic Corporation. All rights reserved.
♦
♦
Appendix A, Glossary, gives a list of terms applicable to the E-10
core.
Appendix B, Customer Feedback, includes a form that you may use
to fax us your comments about this document.
Related Publications
IEEE Standard 802.3 10 Mbits/s Ethernet Specification
CoreWare® Mixed Signal 10 Mbits/s Ethernet Transceiver for 10BASE-T
and AUI Datasheet, Order Number R15003
Conventions Used in This Manual
The first time a word or phrase is used in this manual, it is italicized. See
the glossary at the end of this manual for definitions of italicized words.
The word assert means to drive a signal true or active. The word
deassert means to drive a signal false or inactive.
The following signal naming conventions are used throughout this
manual:
♦
♦
♦
♦
A signal whose name ends with a P (positive) is true or valid when
the signal is HIGH.
A signal whose name ends in an N (negative) is true or valid when
the signal is LOW.
If a signal can be either active-HIGH or active-LOW and is an input, the
signal name ends in I (input).
If a signal can be either active-HIGH or active-LOW and is an output,
the signal name ends in O (output).
The following register bit conventions are used throughout this manual:
♦
♦
A register bit that is set has a binary value of one.
A register bit that is cleared has a binary value of zero.
Hexadecimal numbers are indicated by the prefix “0x” before the
number—for example, 0x32CF. Binary numbers are indicated by the
prefix “0b” before the number—for example, 0b0011.0010.1100.1111.
x
Preface
February 1997 - Rev. A
Copyright © 1994-1997 by LSI Logic Corporation. All rights reserved.
Chapter 1
Introduction
LSI Logic’s CoreWare® library consists of leading-edge microprocessors,
floating-point processors, digital signal processing (DSP) functions, and
peripheral functions. Designers can combine CoreWare building blocks
with other LSI Logic library elements to implement highly integrated
systems in silicon for a wide variety of applications. This document
describes the function, operation, and features of the Ethernet-10 (E-10)
building block that is part of the CoreWare library.
This chapter provides an overview of LSI Logic’s CoreWare program and
the CoreWare building blocks, introduces some networking concepts,
describes in general terms the E-10 core, and then provides a summary
of the core’s features.
The chapter contains the following sections:
♦
♦
♦
♦
Section 1.1, “CoreWare Program”
Section 1.2, “Local Area Networks Overview”
Section 1.3, “General Description”
Section 1.4, “Features”
1.1 CoreWare Program
An LSI Logic core is a fully defined, optimized, and reusable block of
logic. It supports industry-standard functions and has predefined timing
and layout. The core is also an encrypted RTL simulation model for a
wide range of VHDL and verilog simulators.
The CoreWare library contains an extensive set of complex cores for the
communications, consumer, and computer markets. The library consists
of high-speed interconnect functions such as the GigaBlaze G10™-p
core, MIPS embedded microprocessors, MPEG-2 decoders, a PCI core,
and many more.
Ethernet-10 (E-10) Core Technical Manual
1-1
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
The library also includes megafunctions or building blocks, which provide
useful functions for developing a system on a chip. Through the
CoreWare program, you can create a system on a chip uniquely suited
to your applications.
Each core has an associated set of deliverables, including:
♦
♦
♦
♦
♦
♦
RTL simulation models for the Verilog and HDL environments
A System Verification Environment (SVE) for RTL-based simulation
Synthesis and timing shells
Netlists for full timing simulation
Complete documentation
LSI Logic ToolKit support
LSI Logic's ToolKit provides seamless connectivity between products
from leading electronic design automation (EDA) vendors and LSI Logic's
manufacturing environment. Standard interfaces for formats and
languages such as VHDL, Verilog, Waveform Generation Language
(WGL), Physical Design Exchange Format (PDEF), and Standard Delay
Format (SDF) allow a wide range of tools to interoperate within the
LSI Logic ToolKit environment. In addition to design capabilities, full scan
Automatic Test Pattern Generation (ATPG) tools and LSI Logic's
specialized test solutions can be combined to provide high-fault coverage
test programs that assure a fully functional design.
Because your design requirements are unique, LSI Logic is flexible in
working with you to develop your system-on-a-chip CoreWare design.
Three different work relationships are available:
♦
♦
♦
You provide LSI Logic with a detailed specification and LSI Logic
does all of the design
You design some functions while LSI Logic provides you with the
cores and megafunctions, and LSI Logic completes the integration
You perform the entire design and integration, and LSI Logic
provides the core and associated deliverables
Whatever the working relationship, LSI Logic’s advanced CoreWare
methodology and ASIC process technologies consistantly produce
Right-First-Time™ silicon.
1-2
Introduction
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
1.2 Local Area Networks Overview
If you are already familiar with networking and local area networks, you
may skip this section and proceed directly to Section 1.3 on page 1-4.
1.2.1 Ethernet Standards
Standards organizations govern the development of specifications for
Ethernet LAN technology. The standard having the greatest influence
over LAN design and specifications is the ANSI/IEEE 802.3 Ethernet
standard, which is available from the IEEE in published form.
Each major LAN system technology has gained acceptance by the
IEEE 802 Standards Committee. The IEEE 802 standards are compati-
ble with the International Standards Organization’s (ISO) Open System
Interconnect (OSI) reference model, shown in Figure 1.1.
Figure 1.1 Ethernet-10 MAC Relationship to the OSI Model
OSI
REFERENCE
MODEL
LAN
CSMA/CD
LAYERS
LAYERS
HIGHER LAYERS
APPLICATION
LLC - LOGICAL LINK
CONTROL
PRESENTATION
MAC - MEDIA ACCESS
CONTROL
SESSION
RECONCILIATION
MII
TRANSPORT
NETWORK
DATA LINK
PCS
PMA
PMD
PHY
E-10
Core
PHYSICAL
MDI
MEDIUM
MDI = Media-Dependent Interface
MII = Media-Independent Interface
PCS = Physical Coding Sublayer
PMA = Physical Medium Attachment
PHY = Physical Layer Device
PMD = Physical Medium Dependent
Ethernet-10 (E-10) Core Technical Manual
1-3
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
The E-10 core complies with the Media Access Control (MAC) sublevel
within the Data Link layer of the OSI reference model and also contains
a Manchester Encoder/Decoder (ENDEC) and transceivers for
10BASE-T or AUI interfaces. The ENDEC and transceivers are part of
the Physical layer of the OSI model.
1.3 General Description
The IEEE 802.3 E-10 core provides the designer flexibility in integrating
an Ethernet port into an LSI Logic application-specific design. An
Ethernet port includes the following:
♦
♦
♦
Media Access Control (MAC)
Manchester Encoder/Decoder (ENDEC)
Network Transceivers
The E-10 core is part of the LSI Logic CoreWare Library. The CoreWare
program provides the critical building blocks, CAD tools, and applications
support necessary to achieve high levels of system integration and
product differentiation using LSI Logic’s Right-First-Time ASIC design
methodology.
As a CoreWare product, the E-10 core provides a highly-integrated
Ethernet core implementation using submicron CMOS process
technology. This optimization of size and functionality enables designers
to easily integrate multiple Ethernet cores onto a single chip.
The E-10 core is fully synchronous, therefore providing predictable and
reliable operation.
The E-10 core processes data at 10 Mbit/s at half-duplex and 10 Mbit/s
simultaneously in each direction in full-duplex mode, provides a
comprehensive solution for 10BASE-T based systems, and is designed
for easy interface to 10BASE-2, 10BASE-5, or 10BASE-F (fiber) media
through an Attachment Unit Interface (AUI). The E-10 core engine gives
the designer a very effective systems integration building block for
developing next-generation networking products for switched hub or
router applications. The E-10 core also enables the highest form of
system integration for workstation or high-end PCs by converting all the
I/O and system logic into a single-chip motherboard solution.
1-4
Introduction
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
The functional blocks within the E-10 core are:
♦
Media Access Control (MAC). The Media Access Control function
provides simple and independent frame transmission and reception
control by means of parallel eight-bit data interfaces.
♦
Manchester Encoder/Decoder (ENDEC). The E-10 core contains an
integrated ENDEC function that performs Manchester encoding and
decoding and utilizes a digital phase-locked loop for decoding data
at 10 Mbit/s.
♦
Network Transceivers. LSI Logic provides two methods for integrating
AUI/TP transceivers:
–
–
LSI library cells (best for single-port applications)
Mixed signal cells (best for multiport applications)
Ethernet-10 (E-10) Core Technical Manual
1-5
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
LSI Logic provides the following mixed signal cores for interfacing to
a network:
–
–
–
10BASE-T/AUI driver
10BASE-T/AUI receiver
Bias generator.
These cores perform all the analog interface functions for a 10BASE-T
four-wire twisted pair or AUI six-wire connection, reduce requirements for
external components, and reduce emissions, which is especially useful
for multiport devices. The E-10 core provides inputs and outputs that are
easily connected to the mixed signal cores, which are compatible with
IEEE 802.3 networks. The core has pins that can be connected to an
8-wire twisted-pair interface or a 9-wire AUI interface using transmit
drivers and differential receivers.
Figure 1.2 is a block diagram of the E-10 core.
1-6
Introduction
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
Figure 1.2 E-10 Block Diagram
1
10BASE-2, 10BASE-5, or 10BASE-F Network
Medium Attachment Unit (MAU)
Attachment Unit Interface (AUI)
Transceivers
E-10 Core
Predistort Shaping
and Link Integrity
Generation
Squelch and Link
Integrity Detection
Manchester Encoder/Decoder (ENDEC)
Function
Phase-Locked Loop
Data Recovery Unit
Manchester Encoder
Media Access Control (MAC)
Function
802.3 Receive Engine
- CRC and Other Error Checking
- Collision Detection Unit
802.3 Transmit Engine
- CRC Generation
- Deferral and Collision Backoff Control
Control
Host Interface Unit
1. With the proper transceiver implementation, the E-10 can be compatible with 10BASE-F, 10BASE-FL and 10BASE-FX.
Ethernet-10 (E-10) Core Technical Manual
1-7
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
1.4 Features
This section summarizes the key features of the E-10 core.
1.4.1 General Features and Benefits
The E-10 core contains the following general features and associated
benefits:
♦
♦
♦
Certification to IEEE 802.3 Ethernet requirements by a recognized
Ethernet Interoperability Lab. Compliance with other Ethernet
products at the MAC and PHY layers of the OSI Model is assured.
Deep submicron CMOS process technology. Submicron technology
allows optimized core size and functionality, which enables
integration of multiple cores.
System Verification Environment (SVE) for E-10 simulation. Verilog
SVE allows designers to rapidly simulate with the E-10 core, and
includes an Ethernet emulator, packet analyzer, and host interface
bus emulator for performing system analysis.
♦
♦
VHDL and Verilog HDL simulation model support.
LSI Logic-created core test vectors. The core test vectors are
guaranteed to provide a minimum 95% fault coverage.
♦
CoreWare engineering support for applications, development, and
test needs. Expert engineering assistance is available for integrating
the E-10 core into an ASIC design.
1.4.2 Media Access Control (MAC) Features
The MAC contains the following features:
♦
♦
♦
♦
IEEE 802.3 Ethernet standards compliance
Functionality at the MAC sublevel of OSI layer 2
Support for an external Multicast Hash filter
Automatic internal Frame Check Sequence (FCS) generator and
checker
1-8
Introduction
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
♦
♦
♦
Programmable source address insertion
Programmable pad insertion
Separate, parallel eight-bit host interfaces for transmit and receive data
1.4.3 Encoder/Decoder (ENDEC) Features
The ENDEC contains the following features:
♦
Manchester encoding/decoding
10 Mbit/s receive or transmit (half-duplex)
10 Mbit/s receive and transmit (full-duplex—twisted-pair mode only)
–
–
♦
♦
♦
Frame decoding up to 4,500 bytes
Digital phase-locked loop with synchronous data recovery
Manchester data decoding of incoming signals containing up to ± 18
ns of jitter
♦
♦
PLL with fast lock time (seven bit times)—one bit at 10 Mbit/s equals
100 ns
Randomized collision backoff algorithm
1.4.4 Twisted-Pair Transceiver Interface Features
The E-10 twisted-pair transceiver interface has the following features:
♦
Integrated transceiver, which provide:
Transmitter and receiver (four-wire or eight-wire interface) for
–
10BASE-T
–
–
Collision detection
Link integrity test
♦
♦
♦
Internal loopback test capability
Link polarity detection and correction
Smart receive squelch for receive data
Ethernet-10 (E-10) Core Technical Manual
1-9
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
1.4.5 Attachment Unit Interface Features
The E-10 AUI interface contains the following features:
♦
Signal Quality Error test (Heartbeat)
Smart receive squelch for receive data
♦
♦
♦
Smart SQE squelch for collision detection
10BASE-2, 10BASE-5, or 10BASE-F interface
1.5 Other Key E-10 Features
Following are some additional key features of the E-10 core:
Three Address Filtering Modes:
♦
–
–
–
Individual address filtering (matches 48-bit destination address
to 48-bit host-supplied address)
Multicast address filtering (provides nine-bit polynomial output for
matching by an external multicast hashing table)
Promiscuous address filtering (matches to any individual
address)
♦
Status Pins
–
–
–
–
Frame size checking/runt frame detection
Dribble bit error
Frame check sequence error
Phase-locked loop error
1-10
Introduction
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
Chapter 2
Signal Descriptions
This chapter provides detailed descriptions of the E-10 signals. These
descriptions are useful for designers who are interfacing the E-10 core
with other core logic or logic external to the core. This chapter contains
the following sections:
♦
♦
♦
♦
♦
♦
♦
♦
♦
Section 2.1, “Logic Symbol”
Section 2.2, “Transceiver Interface Signals”
Section 2.3, “Receiver Signals”
Section 2.4, “Receiver Status Signals”
Section 2.5, “Multicast Filter Signals”
Section 2.6, “Transmitter Signals”
Section 2.7, “Transmitter Status Signals”
Section 2.8, “General MAC Signals”
Section 2.9, “Timing and Test Signals”
Please see the subsection entitled “Conventions Used in This Manual,”
in the preface of this manual for a description of how signals are named.
2.1 Logic Symbol
The logic diagram for the E-10 core is shown in Figure 2.1.
Ethernet-10 (E-10) Core Technical Manual
2-1
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
Figure 2.1 E-10 Logic Symbol
E-10 Core
Transceiver Interface
DATAP
DATAN
PREEP
PREEN
DATAI
DATATP
DATATN
COLLINP
COLLINN
LCORPP
LFORCEP
DRIVEENN
LPBADP
LPASSP
Receiver
RLOCKEDP
RCVNGP
RENABP
RMCENP
RDONEP
RPMENP
RSNOOPENP
RCLEANP
RBYTEP
RDATAO[7:0]
RSPLLEP
RSFCSEP
Status RSDRBLEP
RSMATIP
RSMATMP
RSTARTP
Multicast Filter
MDONEP
MBITP
MSTARTP
MINDEXO[8:0]
Transmitter
TRNSMTP
TSTARTP
TDONEP
TAVAILP
TFINISHP
TDATAI[7:0]
TLASTP
TREADDP
TSECOLP
TSLCOLP
TSMCOLP
TSOCOLP
TSSQEEP
TSNCLBP
TSIDEFP
TABORTP
TCTRI[3:0]
Status
General MAC Functions
SELIO
COLLOUTP
NOADDRI[47:0]
TP/AUI
DUPLEXP
LOOPBKP
Timing/Test Functions
CLK80I
CLK20I
CLK10I
RESETN
TEST1N
TEST2N
CLK20O
CLK10O
TESTSO
TMODE
TESTSEP
TESTSIP
FASTESTN
GNDN
VCC
2-2
Signal Descriptions
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
2.2 Transceiver Interface Signals
The E-10 transceiver interface signals connect to Ethernet transceivers
or transceiver control lines that are external to the E-10 core. The
descriptions of each individual signal are given below.
COLLINN
COLLINP
DATAI
Negative Collision Threshold
Input
COLLINN is one of two pins (the other is COLLINP) used
to detect collisions. It is used for the AUI mode only and
must be connected to the negative threshold output of the
off-core collision differential receiver.
Positive Collision Threshold
Input
COLLINP is one of two pins (the other pin is COLLINN)
used to detect collisions. It is used for the AUI mode only
and must be connected to the positive threshold output
of the off-core collision differential receiver.
Data In
Input
This input pin receives the serial data from the network
and is used by the E-10 ENDEC PLL logic for data
recovery. It should be connected to the zero offset output
of the off-core differential input data receiver.
DATAP
Data Positive
Output
The E-10 core ENDEC transmits positive Manchester-
encoded data on the DATAP pin. DATAP is used in
conjunction with the PREEP signal and operates in both
AUI and TP modes. DATAP must be connected to the
data input of the off-core transmit driver.
DATAN
Data Negative
Output
The E-10 core ENDEC transmits negative Manchester-
encoded data on the DATAN pin. DATAN is used in test
mode and is not needed for interfacing to the mixed
signal transmit driver.
DATATN
Data Threshold Negative
Input
DATATN is used in conjunction with DATATP to control
the E-10 core ENDEC squelch function and link integrity
pulse detection. It should be connected to the negative
threshold output of the off-core differential input data
receiver.
Ethernet-10 (E-10) Core Technical Manual
2-3
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
DATATP
Data Threshold Positive
Input
DATATP is one of two pins (the other pin is DATATN) that
control the E-10 core ENDEC squelch function and link
integrity pulse detection. It should be connected to the
positive threshold output of the off-core differential input
data receiver.
DRIVEENN
LCORPP
Drive Enable
Output
DRIVEENN is the enable output signal needed for
enabling the mixed signal drivers in a four-wire solution
when frame data or link pulses are to be transmitted from
the E-10 core. DRIVEEN is an active-LOW signal and
must be connected to the enable input of the driver.
Link Correct Polarity
Input
LCORPP is used in the Twisted-Pair mode only. When
this input is HIGH in Twisted-Pair mode and the polarity
of the link integrity pulses is inverted, the received data
is inverted to correct for inverted wiring. The E-10 core
ENDEC monitors Link Integrity pulses, which may be
sent on a regular basis from any other device on the
network when the channel is idle. The Link Integrity
pulses received by the core normally must be positive-
going pulses spaced from 6 to 96 milliseconds apart. If
the E-10 core ENDEC detects eight consecutive pulses
of the wrong polarity and the LCORPP pin is HIGH, the
E-10 core ENDEC inverts the received data. When
LCORPP is LOW, the E-10 core ENDEC does not correct
for inverted wiring.
LFORCEP
LPASSP
Link Control Force
Input
LFORCEP is used in the Twisted-Pair mode only. When
LFORCEP is asserted HIGH, the E-10 core ENDEC
allows a transmission even in a link integrity fail situation
(occurs when link pulses are not present due to a broken
or missing wire).
Link State Pass
Output
LPASSP is used in the Twisted-Pair mode only to indicate
the link status. LPASSP is HIGH when the link is active
(data or link integrity pulses received). LPASSP is LOW
when there is a link failure (no data or link integrity pulses
received for more than 96 milliseconds). When in the link
fail state (LPASSP is LOW), the transmit function within
the E-10 core ENDEC is disabled. The disabling of the
2-4
Signal Descriptions
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
receive operation has to be done in the MAC logic, using
this output. The collision output is invalid during the link
fail state.
LPBADP
Link Polarity Bad
Output
LPBADP is used in the Twisted-Pair mode only to
indicate the link polarity status. LPBADP is HIGH when
eight consecutive link integrity pulses of inverted polarity
are received. If LCORPP is HIGH, the E-10 core ENDEC
automatically inverts the data coming from the link to
compensate for inverted wiring.
PREEP
Positive Data Pre-Emphasis
Output
This pin is used in Twisted-Pair mode only and is the
positive pre-emphasis output from the core. The purpose
of PREEP is to increase the amplitude of the higher
frequencies in the output signal. Data pre-emphasis
guarantees that Twisted-Pair mode can drive signals
properly at 10 Mbit/s on cable up to 100 meters long.
PREEP is used in conjunction with DATAP and must be
connected to the pre-emphasis input of the off-core mixed
signal transmit driver. The mixed signal cell uses DATAP
with PREEP to generate the proper transmit signals.
PREEN
Negative Data Pre-Emphasis
Output
This pin is used in Twisted-Pair mode only and is the
negative pre-emphasis output from the core. The
purpose of PREEN is to increase the amplitude of the
higher frequencies in the output signal. It is used in test
mode and is not needed for interfacing to the mixed sig-
nal transmit driver.
Ethernet-10 (E-10) Core Technical Manual
2-5
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
2.3 Receiver Signals
The receiver signals connect to the parallel interface and the control
circuitry of the system. The description of each signal follows.
RLOCKEDP Receiver Locked
Output
The E-10 core ENDEC asserts this output HIGH when
the receiver phase-locked loop is locked onto an input
signal and is receiving frame bits. It takes a minimum of
six bits and a maximum of eight bits for lock to occur. At
least three bits are required for the smart squelch
operation and the remaining bits are required for the PLL
to operate.
RBYTEP
Received Byte
Output
A 100-ns positive pulse on this pin indicates that a new
receive data byte is valid and available on the RDATAO[7:0]
pins. RBYTEP pulses for 100 ns every 800 ns (a byte time)
as long as the E-10 core is receiving data.
RCLEANP
Receiver Cleanup
Output
The E-10 core drives RCLEANP with a 100-ns positive
pulse when the end of an invalid frame is encountered. An
invalid frame is a frame that failed the address filter or was
shorter than a minimum length frame. RCLEANP indicates
to the host that the bytes received and stored by the host
since the last RSTARTP pulse can be discarded.
RCVNGP
Receiving
Output
The E-10 core asserts RCVNGP HIGH when the receiver
is receiving valid Manchester-encoded input data. Every
valid Manchester-encoded bit contains a transition (either
HIGH-to-LOW or LOW-to-HIGH) in the middle of the 100-ns
bit time. The E-10 core asserts RCVNGP HIGH directly
following the start of frame delimiter and deasserts it LOW
when a bit cell is received without a transition in the center
of the bit time.
RDATAO[7:0] Received Data
Output
The E-10 core drives the RDATAO[7:0] receive data bus
with a new data byte when the RBYTEP pin is pulsed
HIGH. The data on these pins is valid and stable for
800 ns until the next RBYTEP pulse is received.
2-6
Signal Descriptions
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
RDONEP
Receiver Done
Output
The E-10 core drives RDONEP with a 100-ns positive
pulse after the reception of a complete frame. The
RDONEP pulse indicates that all data bytes have been
written and all receive status bits are valid.
RENABP
RMCENP
Receive Enable
Input
When RENABP is HIGH, the receive engine is enabled,
frames are able to be received, and address checking is
performed.
Receive Multicast Enable
Input
When RMCENP is HIGH and the multicast filter logic
external to the E-10 receiver is enabled, the E-10 core
continues to receive the frame. Multicast addressing is
used when it is necessary to transmit to a selected group
of destinations with a single individual address. The first
bit of a multicast address is always a one. However, even
if RMCENP is HIGH, an individual address (first address
bit is a zero) can still be recognized.
The external multicast filter logic supplied by the host
uses the nine-bit polynomial supplied on the
MINDEXO[8:0] E-10 pins to create an index into a hash
table. The nine bits are the nine MSBs of the E-10 linear
feedback shift register (LFSR), which generates the
32-bit FCS. The host hash table result (MBITP,
MDONEP) determines whether or not the E-10 core
should continue to receive the frame.
RPMENP
Receive Promiscuous Mode Enable
Input
When this input is HIGH, the E-10 core disables individual
address checking. All frames with an individual address
are accepted. The least-significant bit of the most-
significant byte of an individual address is always a zero.
RSNOOPENP
Receiver Lookback Enable
Input
When RSNOOPENP is HIGH, all transmitted frames are
routed back through the receiver. Operation is similar to
that of loopback except that the phase-locked loop is
bypassed. A lookback frame is indicated when both the
RSMATIP and RSMATMP receiver output pins are
asserted HIGH.
Ethernet-10 (E-10) Core Technical Manual
2-7
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
2.4 Receiver Status Signals
The E-10 receiver status signals may be used by the host to monitor
various conditions regarding the reception of a data frame.
RSDRBLEP Receive Status Dribble Bits
Output
When RSDRBLEP is HIGH, the number of bits in the
received frame was not a multiple of eight, or contained
dribble bits.
RSFCSEP
Receive Status Frame Check Sequence Error Output
When this signal is HIGH, the frame check sequence
(FCS), which consists of the last four bytes of the
received frame, did not correspond with the expected
value. The FCS is also known as a Cyclic Redundancy
Check (CRC). The E-10 core preloads a 32-bit LFSR with
all ones. It then uses the incoming source address,
destination address, length, data, pad, and FCS fields to
update the 32-bit LFSR. If the contents of the 32-bit
LFSR are incorrect after all bits are received, the E-10
core asserts RSFCSEP HIGH.
RSMATIP
Receive Status Match Individual Address
Output
When this signal is HIGH, all 48 destination address bits
of the received frame matched the 48 bits of the node or
individual address. In promiscuous mode, RSMATIP goes
HIGH only if the address received matches the individual
address. RSMATIP is not asserted for multicast
addresses.
RSMATMP
RSPLLEP
Receive Status Match Multicast Address
Output
When this signal is HIGH, the frame was received
because the external multicast filter logic accepted the
frame. RSMATMP is not asserted for individual or
promiscuous address modes.
Receive Status Phase-Locked Loop (PLL)
Error
Output
When this pin is HIGH, the phase-locked loop
encountered a clock rate error during the reception of the
frame. The state of RSPLLEP is a “don’t care” prior to the
time that the PLL acquires phase-lock.
2-8
Signal Descriptions
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
A phase-locked loop error signals that the data recovery
circuit ran out of elasticity during the reception of the
frame. Elasticity is a measure of the tolerance of the
phase-locked loop to the cumulative bit shift due to small
differences in the free-running clock oscillator frequencies
of the transmitting source and receiving destination. If the
oscillator frequencies are not exactly the same, it is
possible that the centers of the bits during lengthy frames
will drift apart so much that the phase-locked loop cannot
properly recover the data. The PLL runs out of elasticity
after a bit shift between the transmitter and receiver of
exactly six bit times or a clock period difference of
± .05%. For more information on elasticity, see the
subsection entitled “PLL Characteristics” on page 3-13.
RSTARTP
Receiver Start
Output
The E-10 core drives this pin with a 100-ns positive pulse
immediately following the start of frame delimiter. The
pulse may be used to prepare the external logic for the
reception of data.
2.5 Multicast Filter Signals
The multicast filter signals connect to the parallel interface and the
control circuitry of the system. The description of each signal follows.
MBITP
Multicast Hash Table Bit
Input
While the MDONEP pulse is asserted HIGH, the MAC
samples MBITP. When MBITP is HIGH, it indicates that
the external logic has determined that the destination
multicast address is valid.
MDONEP
Multicast Filter Done
Input
External host logic must examine the E-10 MINDEXO[8:0]
lines to determine if a destination multicast address is
valid. The host may use MINDEXO[8:0] as an index into
a multicast hash lookup table. When the external logic
decides to accept or reject the multicast message based
on the hash table results, the logic indicates it to the E-10
receive engine by asserting the MBITP pin HIGH (if the
multicast address is accepted) or by deasserting the
MBITP pin LOW (if the multicast address is rejected) and
asserting a 100-ns positive pulse on the MDONEP pin.
Ethernet-10 (E-10) Core Technical Manual
2-9
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
MINDEXO[8:0] Multicast Hash Function
Output
These nine output pins contain the nine-bit hash function
computed from the destination address. The E-10 core
uses the nine MSBs of the 32-bit LFSR that is used to
generate the FCS to create the nine-bit hash function.
The nine-bit hash function is generated after the final
destination address bit is received.
MSTARTP
Multicast Filter Start
Output
When multicast is enabled (RMCENP is HIGH) and a
frame is received with a destination multicast address, the
E-10 core places a 100-ns pulse on MSTARTP to indicate
to the external logic that a new nine-bit hash function has
been computed from the destination address.
2.6 Transmitter Signals
The MAC transmitter signals connect to the parallel interface and the
control circuitry of the system. The description of each signal follows.
TABORTP
Transmit Abort
Input
The host asserts TABORTP to indicate that an underflow
condition exists, which signifies that the host is not
supplying transmit data fast enough to the core. The host
is responsible for detecting the underflow condition.
Assertion of TABORTP forces an invalid end to the frame.
The host may assert TABORTP for any reason in order
to stop transmission of data. As soon as TABORTP is
received, the E-10 core transmits 32 bits of alternating
ones and zeros, starting with a one, before stopping
transmission.
TAVAILP
Transmit Packet Available
Input
TAVAILP should be asserted HIGH when the host has
one or more data packets available to be transmitted.
This signal is sampled only once per frame, so it can be
deasserted LOW by the host after the E-10 core asserts
the TRNSMTP and TSTARTP output pins. The E-10 core
continues to attempt to transmit the frame until it is
transmitted successfully, an error occurs, or the
transmission is aborted.
2-10
Signal Descriptions
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
After the TAVAILP signal is first sampled, it is not
sampled again until a minimum of 800 ns after the
TDONEP pulse, to give external logic time to do buffer
housekeeping, and to determine whether or not another
buffer is ready to be transmitted.
TCTRI[3:0]
Transmit Control
Input
The bits in this four-bit control bus are described in the
following table.
Bit Description
0
1
2
3
Insert Source Address Enable
Insert FCS Field Enable
Automatic Frame Padding Enable
SQE Test Enable
The TCTRI[3:0] signals come from the host. They are
normally static signals that the E-10 core uses to
determine transmission operation.
TCTRI[0] is the Insert Source Address Enable signal.
When TCTRI[0] is HIGH, the E-10 core inserts the source
address into a transmitted frame from the 48-bit
NOADDRI[47:0] bus supplied at the E-10 input pins by
the host. If TCTRI[0] is LOW, the host itself must place
the source address in the transmitted frame.
TCTRI[1] is the Insert FCS Field Enable signal. When
TCTRI[1] is HIGH, the E-10 core computes a four-byte
FCS and appends it to the end of a transmitted frame.
When TCTRI[1] is LOW, the host must generate and
append the FCS field to the frame.
TCTRI[2] is the Automatic Frame Padding Enable signal.
When TCTRI[2] is HIGH, the E-10 core examines the
data field supplied by the host. If the data field contains
less than 46 bytes (46 bytes is the minimum amount of
data required for a minimum frame length), the E-10 core
supplies enough bytes filled with zeroes to make a
46-byte data field. If TCTRI[2] is LOW, the E-10 core
does not automatically append any bytes to the data field.
TCTRI[3] is the SQE Test Enable signal. When TCTRI[3]
is HIGH, the E-10 core expects to receive a “heartbeat”
on the COLLINP and COLLINN signal inputs after each
frame is transmitted. The heartbeat is a 10-MHz burst
Ethernet-10 (E-10) Core Technical Manual
2-11
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
and starts six to 16 bit times (0.6 µs to 1.6 µs) after the
last transition of the transmitted signal and lasts for a
duration of five to 15 bit times. If TCTRI[3] is LOW, the
E-10 core expects no such signalling. The SQE Test
Enable signal is not applicable to Twisted-Pair mode.
TDATAI[7:0] Transmit Data
Input
The data on this eight-bit transmit data bus is sampled on
the LOW-to-HIGH transition of CLK10I when TREADDP
is asserted. After TREADDP is asserted HIGH, the
external logic has up to 800 ns to place the next data
byte onto the TDATAI[7:0] bus.
TDONEP
Transmit Done
Output
The E-10 core puts a 100-ns positive pulse on the
TDONEP pin to indicate either the end of a transmission,
the frame was transmitted successfully, excessive
collisions are detected, or the signals TFINISHP or
TABORTP are asserted. When this signal is asserted, the
transmit status pins are valid. Section 2.7, “Transmitter
Status Signals,” defines the transmit status pins.
TFINISHP
Transmit Finish
Input
When TFINISHP is asserted HIGH during a transmission,
the current transmission attempt is completed, no retry is
attempted, and a TDONEP pulse generated even if the
transmission attempt is not successful.
TLASTP
Transmit Last Byte
Input
The host asserts TLASTP HIGH to indicate that
TDATAI[7:0] contains the last byte of data in the frame.
TREADDP
Transmitter Read Data Output
A HIGH pulse on this output pin indicates that the data
byte previously placed on the TDATAI[7:0] lines by the
host is going to be read into the E-10 core on the next
LOW-to-HIGH transition of CLK10I. After the data is read,
the host can remove the data from the TDATAI[7:0] pins.
The host should put the next data byte on the
TDATAI[7:0] pins within 800 ns.
2-12
Signal Descriptions
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
TRNSMTP
TSTARTP
Transmitting
Output
A HIGH on TRNSMTP indicates to external logic for both
the AUI and TP modes that the MAC is in transmit mode
and is putting Manchester-encoded data on the E-10 pins
that drive the off-core transceivers.
Transmit Start
Output
The E-10 core asserts a 100-ns positive pulse on the
TSTARTP pin to indicate the start of a transmission (or
retransmission). The host should supply data on the
TDATAI[7:0] pins within 6.4 µs of the TSTARTP pulse.
External logic may use this signal to reset the data
pointer to the beginning of the transmit data buffer.
2.7 Transmitter Status Signals
The E-10 transmitter status signals may be used by the host to monitor
various conditions regarding the transmission of a data frame.
TSECOLP
TSIDEFP
Transmit Status Excessive Collisions
Output
This signal is asserted HIGH when the frame could not
be transmitted after 16 attempts. The transmission will
then be aborted with no further retry attempts.
Transmit Status Initially Deferred
Output
This signal is asserted HIGH when the first transmission
attempt of a frame is deferred because of activity on the
network and the frame was transmitted without a
collision. The core keeps the TSIDEFP signal asserted
during any subsequent transmission attempt and does
not deassert the TSIDEFP signal until after the end of the
transmission.
TSLCOLP
Transmit Status Late Collision
Output
This signal is asserted HIGH when a collision is detected
more than 512 bits into the frame. The transmit engine
continues to retry. The time period corresponding to
512 bits is a “slot” time (51.2 µs), and is considered to be
the maximum time necessary to transmit data on an
Ethernet channel so that all nodes detect activity. If a late
collision (after 512 bit times) occurs, it indicates that the
Ethernet connection is out of specification and corrective
action should be taken.
Ethernet-10 (E-10) Core Technical Manual
2-13
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
When the E-10 core detects a collision less than 512 bits
into the frame, the TSECOLP, TSMCOLP, or TSOCOLP
signals are asserted HIGH, depending on the
circumstances surrounding the collision (see the
individual signal descriptions for more details).
TSMCOLP
TSNCLBP
Transmit Status Multiple Collisions
Output
This signal is asserted HIGH when more than one
transmission attempt resulted in a collision during the
transmission of the frame. This signal is also asserted
when the transmission is terminated due to the assertion
of TABORTP.
Transmit Status No Carrier Loopback
Output
In AUI mode, TSNCLBP is asserted HIGH when a
transmission results in a transmit carrier loopback or
transmit carrier dropout error. If the E-10 core does not
detect a signal within 512 bit times from the time it was
transmitted, the error condition exists. A transmit dropout
condition could result from a broken or missing coaxial
cable.
In Twisted-Pair mode TSNCLBP is asserted HIGH if a
frame is transmitted while the transceiver is in the link-fail
state. Link pulses are 100-ns in duration and are
transmitted every 16.384 ms in the absence of transmit
data. A link fail state could result from a broken or
missing receive wire pair.
TSOCOLP
TSSQEEP
Transmit Status One Collision
Output
This signal is asserted HIGH when the frame was
transmitted after exactly one collision. This signal is also
asserted when the transmission was terminated due to
the assertion of TABORTP.
Transmit Status SQE Error
Output
A HIGH on this output pin indicates a SQE test error.
2-14
Signal Descriptions
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
2.8 General MAC Signals
The general MAC signals connect to the parallel interface and the control
system circuitry. The description of each signal follows.
COLLOUTP Transmit Collision
Output
COLLOUTP is asserted (HIGH) when a collision is
encountered during a transmission. COLLOUTP is valid
for AUI mode and Twisted-Pair half duplex mode.
COLLOUTP is not valid for Twisted-Pair full duplex mode.
In AUI mode, the collision signal coming from the external
Media Attachment Unit (MAU) at the collision input pins
causes the COLLOUTP signal to be asserted. In Twisted-
Pair mode, when simultaneous transmit and receive
activity takes place, the E-10 core asserts COLLOUTP.
The host must take the operational mode into account
when interpreting this signal.
DUPLEXP
Full Duplex Select
Input
When this input is asserted, the MAC is in full duplex
mode. Full duplex mode is applicable only in the Twisted-
Pair mode of operation.
In Twisted-Pair full duplex mode, collision detection is not
valid. The host must take the operating mode into
account when interpreting the collision output signal
(COLLOUTP).
LOOPBKP
Diagnostic Loopback Enable
Input
When LOOPBKP is HIGH, the E-10 core ENDEC is in
the diagnostic loopback mode and the transmitter output
is fed back to the receiver input inside the core. Loopback
mode allows testing the E-10 transmitter and receiver up
to but not including the off-core transceivers under
program control without the need of an external loopback
connector. This diagnostic loopback capability is
applicable to both the AUI and Twisted-Pair modes. It
does not represent the standard loopback function in
Twisted-Pair mode (as the core presents an internal
on-chip MAU for Twisted-Pair mode)
Ethernet-10 (E-10) Core Technical Manual
2-15
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
NOADDRI[47:0]
Node Address
Input
This 48-bit node address bus is driven with the individual
address of the station. It is up to the host to provide the
appropriate levels on NOADDRI[47:0] to configure the
address.
SELIO
4-Wire and 8-Wire Mixed Signal Cell Select
Input
SELIO must be asserted for four-wire Twisted-Pair mode
and deasserted for eight-wire Twisted-Pair mode. When
SELIO is asserted, the core does not use the PREEP,
PREEN, or DATAN signals. For more information on the
four-wire and eight-wire modes, see Section 3.3,
“Twisted-Pair Interface” on page 3-14.
TP/AUI
Twisted-Pair/AUI Select
Input
When TP/AUI is deasserted (LOW), the MAC core is in
AUI mode. When the signal is asserted (HIGH), Twisted-
Pair mode is selected. Below are listed the principal
differences in the way the E-10 core operates in these
two modes:
Feature
TP Mode AUI Mode
SQE Test
No
Yes
No
No
Link Pulse Detection
Yes
Yes
Link Pulse Polarity Detection and
Correction
Link Integrity Pulses Generated
Yes
No
No
Collision Pair Pins
Yes
(COLLINP, COLLINN) are used
Cause of Loopback Transmit Carrier Link Fail No Carrier
Error Loopback
It makes no difference to the core if the AUI interface is
6-wire or 9-wire. The transceivers and external circuitry
used in either 6-wire or 9-wire AUI operation present the
same signals to the MAC core.
For more information on the AUI mode of operation, see
Section 3.4, “Attachment Unit Interface (AUI)” on
page 3-20. For more information on the twisted-pair mode
of operation, see Section 3.3, “Twisted-Pair Interface” on
page 3-14.
2-16
Signal Descriptions
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
2.9 Timing and Test Signals
The timing and test signals connect to the parallel interface and the
control circuitry of the system. In multiple-core environments, only one
core may be the clock master. All other cores must be slaves. For clock
duty cycles and tolerances, see Section 4.2, “E-10 Clock Timing,” on
page 4-10. The description of each signal follows.
CLK10I
10-MHz Clock In
Input
The E-10 core ENDEC uses CLK10I for MAC and bus
interface timing. CLK10I should be driven by the CLK10O
output pin through a buffer external to the E-10 core.
CLK10O
CLK20I
10-MHz Clock Out
This 10-MHz output is derived from the 80-MHz input
signal, CLK80I.
Output
20-MHz Clock In
Input
The E-10 core uses CLK20I for timing and for the
Manchester encoder clocking. CLK20I should be driven
by the CLK20O output pin through a buffer external to the
E-10 core.
CLK20O
CLK80I
20-MHz Clock Out
This 20-MHz output is derived from the 80-MHz input
signal, CLK80I.
Output
80-MHz Clock In
Input
CLK80I is the master clock source provided to the E-10
core ENDEC. It is the only external timing signal used by
the E-10 core ENDEC. CLK80I may be provided as an
external oscillator or it may be provided with a PLL circuit
as part of some chip logic surrounding the E-10 core.
FASTESTN
GNDN
Fast Test
Input
FASTESTN is used for chip testing purposes only (fast
test) and must be left HIGH during normal operation.
Ground
Input
The GND signal must be asserted (LOW) during normal
core operation. It is used for testing the core.
Ethernet-10 (E-10) Core Technical Manual
2-17
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
RESETN
Core Reset
Input
After power is applied to the chip, RESETN should be
asserted LOW for a minimum of 200 ns (at least 100 ns
after TEST1N goes HIGH) to force initialization.
TESTSEP
Scan Enable
Input
TESTSEP is the scan enable pin for testing the module
in the ENDEC. The ENDEC has a full-scan test chain
built into it.
TESTSIP
TESTSOP
TEST1N
Scan Input
Input
TESTSIP is the scan input pin for full scan testing.
Scan Output
TESTSOP is the scan output pin for full scan testing.
Output
Test 1
Input
TEST1N is for chip testing purposes only (fast test) and
must be left HIGH in normal operation.
TEST2N
Test 2
Input
TEST2N must be left HIGH in normal operation. It must
be asserted LOW for 100 ns to initialize the ENDEC core
logic. The CLK10O AND CLK20O outputs become
operational only after the TES2N signal is deasserted
(HIGH) during initialization RESETN is then asserted
(LOW) for an additional 100 ns to initialize the rest of the
ENDEC logic. In test mode, TEST2N is used for testing
the ENDEC core.
TMODE
VCC
Test Mode
Input
TMODE must be left deasserted (LOW) during normal
operation and asserted (HIGH) in test mode. TMODE is
used during scan testing.
Positive Voltage
Input
The VCC pin should be asserted (HIGH) during normal
operation. It is used for core testing.
2-18
Signal Descriptions
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
Chapter 3
E-10 Functional
Description
This chapter provides a functional description of the E-10 core and
contains the following sections:
♦
♦
♦
♦
Section 3.1, “Media Access Control (MAC)”
Section 3.2, “Encoder/Decoder (ENDEC)”
Section 3.3, “Twisted-Pair Interface”
Section 3.4, “Attachment Unit Interface (AUI)”
3.1 Media Access Control (MAC)
The MAC is the portion of the E-10 core that handles the Carrier Sense
Multiple Access with Collision Detection (CSMA/CD) protocol for trans-
mission and reception of frames. The MAC performs the following func-
tions:
♦
♦
♦
Frame Data Encapsulation and Decapsulation
Frame Transmission
Frame Reception
3.1.1 Frame Data Encapsulation and Decapsulation
The format of the frame generated by the ANSI/IEEE Ethernet 802.3
protocol consists of several fields, as explained in the following
subsections. Figure 3.1 shows the overall structure and order of
transmission of bits and octets of a frame.
Ethernet-10 (E-10) Core Technical Manual
3-1
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
Figure 3.1 MAC Frame Format
7 Octets
1 Octet
Preamble
SFD
6 Octets
6 Octets
2 Octets
Destination Address
Source Address
Length/Type
1 Frame
Octets within a Frame are
Transmitted Top-to-Bottom
Data
Pad
Frame Check
Sequence
4 Octets
LSB
MSB
Bits within each Octet are
Transmitted Left-to-Right
3.1.1.1 Preamble Field
The preamble field is a 7-octet (56-bit) alternating one-zero pattern
(starting with a one), which is Manchester-encoded before being
transmitted onto the media. The alternating one-zero pattern ensures
that there are no transitions between bits in the Manchester-encoded
signal, which results in a 5-MHz signal (see Figure 3.2). Because the
twisted-pair Ethernet cable attenuates higher frequencies, the 5-MHz
signal creates a perfect signal for acquisition by the digital Phase-Locked
Loop (PLL). Any pattern other than an alternating one-zero pattern would
create a higher-frequency signal and make acquisition more difficult. In
addition, an alternating one-zero data pattern, when Manchester-
encoded, has data transitions at the same clock edge, which makes it
easier for the PLL to achieve lock.
3-2
E-10 Functional Description
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
Figure 3.2 Alternate One-Zero Manchester Encoding
1 Bit
10-MHz Clock
Alternate Ones
and Zeroes Data
0
1
0
1
0
1
0
1
0
1
0
1
0
1
0
1
Manchester-Encoded
Alternate Ones and
Zeroes Data (5 MHz)
3.1.1.2 Start Frame Delimiter (SFD) Field
The eight-bit start of frame delimiter field contains the 10101011 bit
2
pattern. This bit pattern allows alignment of the data received in the rest
of the frame.
3.1.1.3 Address Fields
The frame has two 48-bit address fields. The first field contains the
destination address and specifies the address(es) for which the frame is
intended. If the first bit of the destination address is a zero, the address
is an individual address. A one in the first bit indicates a multicast or
group address. The second address is the source address and should
contain the individual address of the station from which the message
originated.
3.1.1.4 Length/Type Field
The two-byte length field indicates the number of data bytes in the data
field. The MAC does not interpret the data in the length field and does
not treat it any differently than the data field.
3.1.1.5 Data and Pad Field
The data field contains a sequence of fully transparent data bytes; that
is, the data bytes are not analyzed or interpreted in any way by the MAC.
The size of the data field may be between zero and 1,500 bytes, but can
be extended up to 4,500 bytes or more, depending on the clock accuracy
of the transmitting and receiving stations.
Ethernet-10 (E-10) Core Technical Manual
3-3
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
The function of the pad field is to ensure minimum frame size. If the
supplied data field contains less than 46 bytes of data, the MAC can add
a pad field to make the sum of bytes in the data and pad field equal to
46. Appended pad data consists of bytes filled with zeroes.
3.1.1.6 Frame Check Sequence (FCS) Field
The transmit and receive algorithm uses the standard 802.3 four-byte
frame check sequence (FCS) field to ensure data integrity. The E-10 core
uses a linear feedback shift register to compute the value in the FCS
field. The FCS is a function of the content of the source address,
destination address, length, data, and pad fields. The generating
polynomial shown in Equation 3.1 defines the encoding. The nine MSBs
of the FCS may be used by an external hashing table function to perform
multicast address filtering.
Equation 3.1
FCS Encoding
G(x) = x32 + x26 + x23 + x22 + x16 + x12 + x11 + x10 + x8 + x7 + x5
+ x4 + x3 + x2 + x + 1
3.1.2 Frame Transmission
Frame transmission is enabled with the TAVAILP signal. TAVAILP
indicates that at least one frame is ready to be transmitted. All
transmissions start by complying with the rules of deference; that is, they
monitor the physical medium for activity. If no activity exists, the transmit
cycle starts.
During the start of a transmission, the transmitter asserts a 100-ns
positive start of transmission pulse (TSTARTP). The start pulse resets
the transmit pointer to the beginning of the transmit buffer. When the
TREADDP signal is pulsed (HIGH), it indicates to the external control
logic that data is read on the LOW-to-HIGH transition of the CLK10I
input. If a collision is detected during a transmission, the collision
enforcement rules are followed, and the transmission is aborted. If the
signals TABORTP and TFINISHP are not asserted, and the frame was
not retried 15 times, backoff and deference rules are followed and the
frame transmission is retried.
The external control logic asserts the TLASTP signal to indicate that the
last byte of data to be transmitted is present at the TDATAI[7:0] pins.
When a frame is transmitted successfully, or when a serious error is
3-4
E-10 Functional Description
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
detected during the transmission, the transmitter asserts the transmitter
done (TDONEP) pulse. The TDONEP pulse indicates to the external logic
that either the transmission is complete, or the TABORTP or TFINISHP
signals were asserted before the end of the transmission, or that the
transmitter exhausted its ability to transmit the frame successfully.
External logic has a minimum of 800 ns to deassert or assert TAVAILP
indicating whether or not a new frame is ready for transmission.
3.1.2.1 Carrier Deference Rules
Even when it has nothing to transmit, the transmit portion of the MAC
monitors the traffic on the physical medium. When there is no traffic (a
carrier is not detected by the MAC and the MAC is not transmitting) the
MAC starts a 9.6 µs interframe timer. The interframe timer is reset if the
MAC senses a carrier during the first 5.6 µs of the interframe gap after
a reception. During the time from 5.6 µs to 9.6 µs, the interframe timer
is not reset, to ensure fair access to the medium.
If, at the end of the interframe gap, a frame is waiting to be transmitted,
transmission is started regardless of the presence of a carrier. If a carrier
is present, a collision occurs so that a node constantly transmitting is
forced to back off. If, at the end of the interframe gap, there is no frame
waiting to be transmitted, the MAC starts monitoring for a carrier again,
and resets the interframe timer as soon as it detects a carrier.
3.1.2.2 Interframe Spacing
The carrier deference method described above ensures that the E-10
provides an interframe gap of 9.6 µs. The interframe gap ensures fair
arbitration among all sources trying to transmit and prevents one
transmitter from dominating the transmission channel.
3.1.2.3 Collision Enforcement Rules
When a collision is detected during a frame transmission (for the definition
of a frame, see the subsection entitled “Frame Data Encapsulation and
Decapsulation” on page 3-1), the transmission is not terminated
immediately. The preamble and start of frame delimiter (SFD) are
transmitted unmodified regardless of the collision state. If a collision is
detected before the preamble and SFD have finished transmitting, a 32-bit
jam pattern of alternating ones and zeros is transmitted immediately after
the SFD and then the transmission ends. If the collision is detected after
Ethernet-10 (E-10) Core Technical Manual
3-5
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
the preamble and SFD, while data is being transmitted, the data
transmitted changes to a jam pattern of alternating ones and zeros for
32 bits before the transmission is terminated.
3.1.2.4 Transmission Backoff Algorithm
When a transmission attempt has terminated due to a collision, the
transmission is retried until either it is successful or 15 retries (16
attempts total) have been terminated due to collisions. A controlled
randomization process called “truncated binary exponential backoff”
schedules the retransmission. At the end of enforcing a collision
(jamming), the transmit engine delays before attempting to retransmit the
frame. The delay is in 51.2 µs increments, or slot times. The number of
slot times to delay before the nth retransmission attempt is chosen as a
uniformly distributed random integer r in the range:
k
-1 < r < 2
, where
k = min (n, 10)
Note that min (n, 10) signifies that the value is the minimum of n or 10,
whichever is less.
A 25-bit Linear Feedback Shift Register (LFSR) generates a pseudo-
random number. Because of potential synchronization problems between
multiple stations on the network using the same pseudo-random number
generator, the number is further randomized every transmission attempt
with the station’s individual address.
3.1.2.5 Source Address Insertion
The hardware can insert the source address during transmission from
the individual address input pins. In this mode, the host is responsible for
supplying the first six bytes (destination address). The six address bytes
are transmitted from the host on the E-10 transmit data input pins
(TDATAI[7:0]). The next six bytes transmitted are read from the individual
address registers (source address). After these two addresses are sent,
the transmitter returns to the transmit data pins and continues
transmitting subsequent data bytes from there. The byte counter in the
transmit data buffer must not include the six source address bytes in this
mode, but must include the six destination address bytes. The transmitter
is in source address insertion mode when the TCTRI[0] pin is HIGH.
3-6
E-10 Functional Description
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
3.1.2.6 Automatic Pad Insertion
During transmission, a pad field can be inserted by the hardware. In this
mode enough bytes are added just before the FCS field to increase the
total data field length to the IEEE 802.3 minimum of 46 bytes. The data
inserted in the pad field is all zeroes. The transmitter is in automatic pad
insertion mode when the TCTRI[2] pin is HIGH.
3.1.2.7 Frame Check Sequence (FCS) Insertion
During transmission the E-10 hardware can compute and transmit the
FCS field automatically. In this mode, a 32-bit polynomial is computed
from the source address, destination address, length, data, and pad
fields and transmitted immediately following the data or pad field. The
transmitter is in FCS insertion mode when the TCTRI[1] pin is HIGH.
If the host instead is going to supply the FCS field and the E-10 core is
going to insert the source address, the host must ensure that the FCS
is computed based on the 48-bit source address being supplied to the
E-10 core on the NOADDRI[47:0] pins.
3.1.2.8 Transmit Status Indications
The six E-10 pins listed below indicate the status of Ethernet
transmission with regard to collisions and errors:
♦
Late Collision (TSLCOLP). The bit durations for transmit and receive
data are each 100 ns. When a collision onset is detected after the
allowed window of 512 bit times (51.2 µs), the E-10 core asserts the
late collision error bit (TSLCOLP) HIGH. A late collision is treated
just like a normal collision, following the collision enforcement rules,
and the transmission is retried.
♦
Excessive Collisions (TSECOLP). When the transmission engine has
attempted to transmit a frame 16 consecutive times, and all
attempted transmissions have terminated due to a collision, the E-10
core asserts the excessive collision error pin (TSECOLP) HIGH, and
no more transmit retries are performed.
Ethernet-10 (E-10) Core Technical Manual
3-7
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
♦
Transmit Carrier Loopback Error (TSNCLBP). In AUI mode, when an
error is detected in the receive carrier signal that is required to be
fed back during the transmission of a frame, the E-10 core asserts
the transmit carrier loopback error pin (TSNCLBP) HIGH. Two
different conditions can cause the transmit carrier loopback error pin
to be asserted in AUI mode:
–
During transmission, no carrier is received during the first 512 bit
times (51.2 µs) of the transmission.
–
During transmission, carrier was received but was interrupted
before transmission stopped.
The transmit carrier loopback error has no effect on the transmission/
retransmission process, but no SQE test is performed.
In Twisted-Pair mode, the E-10 core asserts the transmit carrier loopback
pin HIGH to indicate a link failure during the transmission of the frame.
A link failure can be caused by a twisted-pair transceiver failure or by a
broken twisted-pair cable. In the link fail state, the E-10 core does not
route the Manchester-encoded data to the transmitter pins, but it still
transmits link integrity pulses.
♦
SQE Error (TSSQEEP). The E-10 core asserts the signal quality
error pin (TSSQEEP) HIGH when the MAC detects all the following
conditions:
–
–
–
–
–
–
The E-10 core is in AUI mode
The E-10 core is in half-duplex mode
SQE testing is enabled (TCTRI[3] pin is HIGH)
The late collision status pin (TSLCOLP) is not HIGH
The excessive collisions status pin (TSECOLP) is not HIGH
The transmit carrier loopback error status pin (TSNCLBP) is not
HIGH
–
No SQE signal was received in the first 5.6 µs of the interframe gap
♦
One Collision (TSOCOLP). The E-10 core asserts the one collision
status pin (TSOCOLP) HIGH when the frame is transmitted after
exactly one collision. If more than one collision occurred before the
frame was transmitted, the TSMCOLP pin is asserted (HIGH) and
the TSOCOLP pin is deasserted (LOW). Assertion of the TSOCOLP
pin is not an error indication, but is intended only for information, and
3-8
E-10 Functional Description
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
can be used to update network statistics counters. When the
transmission of a frame is terminated with the signal TABORTP, the
E-10 core asserts the TSOCOLP signal (together with TSMCOLP) to
indicate this fact to external logic. Section 2.6, “Transmitter Signals,”
contains a description of the TABORTP signal. TSOCOLP is valid
when the TDONE pin pulses.
♦
Multiple Collisions (TSMCOLP). The E-10 core asserts the multiple
collision status pin (TSMCOLP) HIGH when the frame is transmitted
after more than one collision. This pin is not an error indication, but
is intended only for information, and can be used to update network
statistical counters. When the transmission of a frame was
terminated with the signal TABORTP, the E-10 core asserts the
TSMCOLP signal (together with TSOCOLP) to indicate this fact to
external logic. TSMCOLP is valid when the TDONEP pin pulses.
3.1.3 Frame Reception
Frame reception is allowed after the receiver enable signal (RENABP) is
asserted. After the transceiver squelch circuitry detects a valid signal,
and the receiver PLL locks onto the incoming signal as indicated by
assertion of the RLOCKEDP signal, the receiver starts searching for a
valid start of frame delimiter. The PLL takes a minimum of six and a
maximum of eight bit times (600 to 800 ns) to achieve lock.
A valid start of frame delimiter is defined as 0b10101011. When the start
of frame delimiter is detected, the receiver pulses the RSTARTP pin. A
pulse on the RSTARTP pin informs the external logic to make available
a new receive buffer, because received data will start arriving in 800 ns.
Every time a new byte of received data is available on the RDATAO[7:0]
bus, the E-10 receiver pulses the RBYTEP pin.
The first 48 bits of the frame contain the destination address. The MAC
receiver portion of the E-10 core compares the destination address bits
with the data on the 48-bit node address bus (NOADDRI[47:0]), and
when the bits are equal, allows the continued reception of the frame.
When the receiver promiscuous mode pin (RPMENP) is asserted, any
individual destination address validates the frame, and allows the
reception process to continue. When the destination address is a
multicast address, including the broadcast address (the address contains
all ones), the receiver computes a nine-bit hash function from the
destination address, and leaves the decision to accept or reject the frame
to external logic.
Ethernet-10 (E-10) Core Technical Manual
3-9
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
When an end-of-frame delimiter is detected, the receiver accepts the
frame and generates a 100-ns pulse on the RDONEP pin when all the
following conditions are true:
♦
♦
♦
At least one address filter passed
The frame was at least 64 bytes in length
The receiver is still enabled
When the end-of-frame delimiter does not generate an RDONEP pulse,
the receiver instead pulses the RCLEANP pin, which indicates to the
external logic that it must ignore the data received. It is then up to the
host logic to manage the receive buffer space properly (either give up the
buffer space so it can be used for other transmit or receive data or
allocate another buffer).
3.1.3.1 Address Filtering
Three address filter mechanisms are present in the receiver, two for
individual addresses, and one for multicast addresses. At least one of
these address filters must pass to accept the received frame. The
address filters are listed below:
♦
Individual Address Filter. The first bit of an individual address is a
zero. The incoming destination address is compared with the data
from the individual address pins. When all 48 bits match and the
receiver is enabled (RENABP pin asserted HIGH), the address filter
passes.
♦
♦
Individual Address Promiscuous Mode. When the external logic
asserts the individual address promiscuous mode enable pin
(RPMEN) HIGH and the destination address is an individual address,
the address filter always passes, and the incoming frame is received.
Multicast Address Filter. The first bit of an multicast address is a one.
When the destination address bits that are received in the frame
contain a multicast address, the E-10 core uses its built-in FCS
generator to compute a nine-bit polynomial (the nine MSBs of the
32-bit FCS generator) from the incoming address. The value of this
polynomial can be used as an index into an external multicast filter
hash table. External logic can decide to either accept or reject the
incoming frame. The external logic asserts the RMCENP pin (HIGH)
to enable the multicast address filter. Figure 3.3 shows the general
3-10
E-10 Functional Description
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
host implementation of a multicast filter. Section 2.5, “Multicast Filter
Signals,” in Chapter 2, “Signal Descriptions,” explains the functions of
the individual signals shown in the figure.
When the destination address is all ones, it indicates a broadcast
address. When the E-10 core detects a broadcast address, it receives
the incoming frame.
Figure 3.3 Host Multicast Filter Implementation
RMCENP (HIGH to Enable E-10 Multicast Filter Logic)
E-10
Core
32-Bit LFSR
MINDEX8
MINDEX0
MINDEX0
MINDEX8
Bit 32
Register
Bit 24
Multicast Filter
Hash Table
MSTARTP
MDONEP
MBITP
3.1.3.2 FCS Checking
During reception, the FCS value is checked against the value computed
by the receive CRC checker. If the two values do not match, the frame
is received and the E-10 core asserts the FCS error pin (RSFCSEP)
HIGH to indicate data corruption in the frame. The decision to accept or
reject the frame is left to external logic.
3.1.3.3 Frame Size Checking
Frame size checking is performed on every received frame. If a frame is
less then 64 bytes in length, the receiver cleanup pulse (RCLEANP)
discards the received data (collision fragment). IEEE 802.3 specifies a
maximum frame size of 1,518 bytes, but the MAC does not do any
maximum length checking, and much larger frames (for instance 4,500
bytes) can be received, as long as care is taken not to exceed PLL
elasticity. The application is responsible for checking the maximum frame
size and reporting any errors.
Ethernet-10 (E-10) Core Technical Manual
3-11
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
3.1.3.4 Start of Frame Delimiter Checking
The start of frame delimiter, a bit sequence of 0b10101011 that occurs
following the preamble, starts the reception of bytes.
3.1.3.5 Receive Status Indications
The five E-10 pins listed below indicate the status of Ethernet reception
with regard to collisions and errors.
♦
Frame Check Error (RSFCSEP). The E-10 core asserts the frame
check error pin (RSFCSEP) HIGH when bits of the incoming frame
do not generate a CRC identical to the one in the received FCS field,
indicating data corruption has occurred in the received frame.
♦
Dribble Bit Error (RSDRBLEP). The E-10 core asserts the dribble bit
error pin (RSDRBLEP) HIGH at the end of a frame when less than
a complete byte has been received, indicating the reception of from
one to seven “dribble” bits. The E-10 core asserts the RSDRBLEP
pin HIGH independently of the frame check error bit. RSDRBLEP
can be ignored if the RSFCSEP pin is not set.
The reason RSDRBLEP can be ignored is that the FCS is updated
every eight bits. If there are less than eight dribble bits (from one to
seven bits) at the end of the frame and the current FCS is correct,
any extra dribble bits (less than eight) do not matter. However, if
there are eight or more dribble bits, the FCS will be updated and will
most likely be wrong due to the dribble bits. In this case, RSDRBEL
should not be ignored.
♦
Phase-Locked Loop Error (RSPLLEP). The E-10 core asserts the
phase-locked loop error pin (RSPLLEP) when, during reception of a
frame, the digital phase-locked data recovery circuit elasticity is
exhausted. The elasticity becomes exhausted due to drift between
the 10-MHz clock frequency of the device sending the data and the
ENDEC 10-MHz clock frequency. See the subsection entitled “PLL
Characteristics” on page 3-13 for an explanation of elasticity.
RSPLLEP is asserted when a frame is received with a transmit clock
error sufficiently larger than the IEEE 802.3 requirement of 0.01%
that the PLL elasticity is exhausted within a normal frame, or when
a frame length sufficiently larger than that specified in IEEE 802.3
has caused the elasticity to be exhausted even when the clock is
within tolerance.
3-12
E-10 Functional Description
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
♦
♦
Match Individual Address (RSMATIP). The E-10 core asserts the
match individual address pin (RSMATIP) HIGH when the destination
address matches the node address.
Match Multicast Address (RSMATMP). The E-10 core asserts the
match multicast address pin (RSMATMP) HIGH or deasserts it LOW
as a result of the decision of the external multicast filter logic to
accept or reject the multicast address.
3.2 Encoder/Decoder (ENDEC)
The ENDEC consists of the Manchester encoder, Manchester decoder,
and the digital phase-locked loop.
3.2.1 PLL Characteristics
The digital phase-locked loop used for data recovery has the following
characteristics:
♦
Sampling
–
–
Sampling Frequency: 160 MHz
Sampling Interval: 6.25 ns
With the above sampling parameters, each bit of 100 ns is sampled
16 times.
♦
♦
Jitter Tolerance: ± 18.00 ns
This parameter indicates that the incoming data bits may have a jitter
up to ±18.00 ns and the digital PLL will still work acceptably.
Tracking speed
–
–
Initial (first 4.8 µs): 0.39%
Steady-state (after 4.8 µs): 0.098%
When first attempting to lock on to an oncoming data stream (during
the first 4.8 µs), the PLL looks every 16 bits to see if a bit transition
is occurring in the center of the sampling window. Dividing the
sampling interval of 6.25 ns by the 16-bit interval (1600 ns) yields
0.0039 or 0.39%.
Ethernet-10 (E-10) Core Technical Manual
3-13
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
After the first 4.8 µs, the PLL maintains a window of 64 bits for
finding a transition. Dividing the sampling interval of 6.25 ns by the
64-bit interval (6400 ns) yields 0.00098 or 0.098%.
♦
Elasticity
–
–
–
Maximum clock drift time 600 ns (6 bit times)
Maximum clock error @ 1518 byte frame: 0.05%
Maximum clock error @ 4500 byte frame: 0.016%
Elasticity indicates the tolerance of the PLL to slightly different clock
rates between a transmitter and a receiver. The E-10 core allocates
a buffer of 13 bits to accommodate clock drift. The 13-bit buffer
allows a clock drift of six bits (600 ns) from center in either direction
during a frame transfer.
For a 1518-byte frame, consisting of six bytes of destination address,
six bytes of source address, two bytes of length, 1500 bytes of data,
and four bytes of FCS (6 + 6 + 2 + 1500 + 4 = 1518), the maximum
clock drift allowed is 0.05% (0.05% x 1518 bytes x 8 bits/byte = 6 bits).
For a 4500 byte frame, the maximum clock drift allowable during
frame transfer is 0.016% (0.016% x 4500 bytes x 8 bits/byte = 6 bits).
3.3 Twisted-Pair Interface
The E-10 core supports two twisted-pair configurations:
♦
Four-wire
Eight-wire
♦
3.3.1 4-Wire Twisted-Pair Interface
The four-wire E-10 twisted-pair interface is shown in Figure 3.4. The
four-wire interface implements CoreWare mixed signal transmitter and
receiver cores to prepare the E-10 core’s I/O signals for connection to the
network. An ASIC implementation of an Ethernet port using the E-10 core
and mixed signal I/O cores provides a simple interface to the network.
3-14
E-10 Functional Description
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
A four-wire interface is usually best for multiport applications. The mixed
signal I/O cores reduce the number of ASIC device pins and external
passive components.
Figure 3.4 4-Wire Twisted-Pair Interface Using Mixed Signal Cores
ASIC
E-10 Core
DATAP
TX+
TX-
Drivers
10BASE-T
Magnetics and
Passive
Components
RJ-45
Connector
DATATP
DATAI
RX+
Receivers
DATATN
RX-
3.3.1.1 4-Wire Twisted-Pair Transmit Interface
With the four-wire twisted-pair interface, the E-10 core outputs a single
Manchester-encoded transmit signal, DATAP, to a mixed signal
differential driver core. The driver produces low common-mode noise
differential outputs, which provide the desired output swing to the
magnetics (transformer). The mixed signal driver reduces the generation
of high-frequency signal components, thereby minimizing external
filtering requirements and reducing external component costs.
3.3.1.2 4-Wire Twisted-Pair Receive Interface
The four-wire receiver interface complies with the receiver specifications
of the IEEE 802.3 10BASE-T standard, including noise immunity,
propagation delays, jitter requirements, and received signal rejection
criteria (Smart Squelch).
Ethernet-10 (E-10) Core Technical Manual
3-15
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
The four-wire twisted pair receive interface has three signals at the E-10
core interface: DATATP, DATATN, and DATAI. The three signals are inputs
to the E-10 core and come from a two-input differential mixed signal
receiver.
The receiver converts the two differential analog data signals from the
network to DATATP, DATATN, and DATAI, which are CMOS-level
single-ended digital outputs. The three signals switch in such a way that
the E-10 core can properly perform data detection, reverse polarity
detection, and smart squelch.
3.3.2 8-Wire Twisted-Pair Interface
The eight-wire E-10 twisted-pair interface is shown in Figure 3.5. This
interface uses data pre-emphasis on data transmitted from the core and
threshold detection circuitry on the data being received by the core.
3-16
E-10 Functional Description
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
Figure 3.5 8-Wire Twisted-Pair Interface
ASIC
E-10 Core
DATAP
TX+
PRE+
PREEP
Drivers
PRE-
PREEN
10BASE-T
DATAN
TX-
Magnetics and
Passive
Components
RJ-45
Connector
RX+
REF
DATATP
Receivers
DATAI
REF
RX-
DATATN
3.3.2.1 8-Wire Twisted-Pair Transmit Interface
With the eight-wire twisted-pair interface, the core outputs four transmit
signals, the true and complemented (differential) Manchester-encoded
data (DATAP and DATAN) and those signals delayed by 50 ns (PREEP
and PREEN). These four signals are sent through four individual transmit
buffers and resistively combined (DATAP with PREEP and DATAN with
PREEN). The technique of resistively combining the signals is known as
digital pre-emphasis and provides the necessary electrical driving
capability and predistortion control for transmitting signals over a
maximum length twisted-pair cable, as specified by the IEEE 802.3
10BASE-T standard. Digital pre-emphasis compensates for the twisted-
pair cable, which acts like a low-pass filter and causes greater
attenuation to the 10-MHz (50-ns) pulses of the Manchester encoded
waveform than the 5-MHz (100-ns) pulses. When a data stream of all
ones or all zeroes is present, Manchester encoding causes a transition
Ethernet-10 (E-10) Core Technical Manual
3-17
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
every bit time, which creates a 10-MHz bitstream. With a data stream of
alternate ones and zeroes, Manchester encoding causes a transition
every other bit time, which creates a 5-MHz bitstream. Figure 3.6 shows
how different data patterns are Manchester encoded.
The differential Manchester-encoded signal is passed to an external
transmit filter. The transmit function meets the propagation delays and
jitter specified by the standard.
Figure 3.6 Manchester Encoding
1 Bit
10-MHz Clock
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
0
All Ones
Data (10 MHz)
0
0
0
1
0
0
0
1
0
0
0
1
0
0
0
1
0
0
0
0
0
1
0
0
0
1
0
0
0
1
All Zeroes
Data (10 MHz)
Alternate Ones and
Zeroes Data (5 MHz)
0
0
1
0
1
1
1
0
1
0
1
0
1
1
1
1
Random Data
3.3.2.2 8-Wire Twisted-Pair Receive Interface
The eight-wire receiver interface complies with the receiver specifications
of the IEEE 802.3 10BASE-T standard, including noise immunity,
propagation delays, jitter requirements, and received signal rejection
criteria (Smart Squelch).
With the eight-wire twisted-pair receive interface, there are three input
signals at the E-10 core interface: DATATP, DATATN, and DATAI. These
input signals are the outputs from three distinct differential receivers.
In a similar fashion to the four-wire interface, the three input signals
switch in such a way that the E-10 core can properly perform data
detection, reverse polarity detection, and smart squelch.
3-18
E-10 Functional Description
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
Following is a functional description of the twisted-pair interface. It
includes these sections:
♦
♦
♦
Squelch
Link Test Function
Polarity Detection and Correction
3.3.3 Squelch
An intelligent (smart) receive squelch is implemented on the receiver
differential inputs to ensure that impulse noise on the receiver inputs is
not mistaken for a valid signal. The smart squelch logic uses a
combination of amplitude and timing measurements to determine the
validity of data on the twisted-pair inputs.
The squelch circuitry operation is illustrated in Figure 3.7. Squelching
determination starts when the input signal crosses the positive or
negative threshold (a). To continue a valid signal sequence, the signal
then has to cross the other threshold (b) within 150 ns. To complete the
valid signal process, the input signal has to cross the original threshold
(c) again within 150 ns. If the signal proves to be valid, the smart squelch
logic does not suppress the signal. However, if not all of the conditions
just outlined are met, the signal is suppressed.
Figure 3.7 Squelch Signal
< 150 ns
Positive
Threshold
(+ 585 mV)
a
c
b
Negative
Threshold
(- 585 mV)
< 150 ns
Ethernet-10 (E-10) Core Technical Manual
3-19
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
3.3.4 Link Test Function
The link test function is implemented as specified by the IEEE 802.3
10BASE-T standard. During periods of transmit pair inactivity, Link Test
pulses are periodically sent from any transmitting device over the twisted-
pair medium to allow constant monitoring of medium integrity. Link
pulses, which are 100 ns wide, are transmitted from the E-10 core every
16.384 ms in the absence of transmit data. The core expects to receive
a link pulse every 96 ms.
The absence of serial Link Integrity pulses on the receiver inputs causes
a Link Fail State to occur. In the Link Fail State, data transmission, data
reception, and the collision detection functions are disabled and remain
disabled until valid data or eight consecutive Link Test pulses appear on
the receiver inputs.
When the E-10 core identifies the link as functional, the E-10 core
asserts the LPASSP pin HIGH to indicate that the link status is good.
3.3.5 Polarity Detection And Correction
The receive function includes the ability to invert the polarity of the
signals appearing at the receiver pair if the polarity of the received signal
is reversed, as in the case of a wiring error. This feature allows frames
received from a reversed twisted-pair interface to be corrected prior to
transfer to the E-10 core. The function uses link pulses to determine
polarity of the received signal. A reversed polarity condition is detected
when seven consecutive opposite-polarity receive link pulses are
detected without receipt of a link pulse of the expected polarity.
3.4 Attachment Unit Interface (AUI)
An AUI provides the link from a 10BASE-T network to a 10BASE-5,
10BASE-2, or 10BASE-F network. As shown in Figure 3.8, an AUI
provides the appropriate interface to a Medium Attachment Unit (MAU)
so that the MAU can interface to a variety of physical media.
3-20
E-10 Functional Description
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
Figure 3.8 10BASE-5 AUI Connection
MAU
AUI
ENDEC
Transceiver
DTE (PC)
AUIs provide transmit, receive, and collision control functions. The E-10
core supports two different Attachment Unit Interface (AUI) configurations:
♦
♦
Six-wire
Nine-wire
3.4.1 6-Wire AUI Interface
The six-wire E-10 AUI interface shown in Figure 3.9 implements CoreWare
mixed signal transmitter and receiver cores to prepare the
E-10 core’s I/O signals for connection to the network. An ASIC
implementation of an Ethernet port using the E-10 core and mixed signal
cores, along with a few discrete components provide impedance matching,
isolation, filtering and electromagnetic interference (EMI) suppression.
Ethernet-10 (E-10) Core Technical Manual
3-21
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
Figure 3.9 AUI 6-Wire Interface
ASIC
E-10 Core
DATAP
AUI
DB-15
Connector
TX+
TX-
Data
Driver
1
9
2
10
3
11
RX+
DATATP
4
12
Magnetics and
Passive
Components
Data
Receiver
5
DATAI
13
6
14
DATATN
RX-
7
15
8
COLLINP
COLLINN
COLL+
COLL-
Collision
Receiver
Without the integrated mixed signal I/O cores, multiple-port interfaces
produce high common-mode coherent noise, which requires the use of
expensive filtering to meet FCC requirements.
3.4.1.1 6-Wire AUI Transmit Interface
With the six-wire AUI interface, the E-10 core outputs a single
Manchester-encoded signal, DATAP, to a mixed signal differential driver
core. The driver produces low common-mode noise differential outputs.
Using controlled impedance voltage, the driver produces the desired
output swing to the magnetics (transformer).
3-22
E-10 Functional Description
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
Using mixed signal drivers reduces the generation of high-frequency signal
components, which propagate easily and cause interference. High-
frequency components must otherwise be eliminated with external filtering.
Using mixed signal drivers minimizes external filtering requirements and
thereby reduces the associated external component costs.
3.4.1.2 6-Wire AUI Receive Interface
The six-wire twisted pair receive interface provides a receiver for data
and a receiver for collision control.
The data receiver has three signals at the E-10 core interface: DATATP,
DATATN, and DATAI. The three signals are inputs to the E-10 core and
come from a two-input differential mixed signal receiver.
The data receiver converts the two differential analog data signals from
the network to DATATP, DATATN, and DATAI. The three signals switch in
such a way that the E-10 core can properly perform data detection,
reverse polarity detection, and smart squelch.
The collision receiver has two signals at the E-10 core interface:
COLLINP and COLLINN. These two signals are inputs to the core and
come from a two-input differential mixed signal receiver
3.4.2 9-Wire AUI Interface
The nine-wire E-10 AUI interface is shown in Figure 3.10.
Ethernet-10 (E-10) Core Technical Manual
3-23
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
Figure 3.10 AUI 9-Wire Interface
ASIC
AUI
DB-15
Connector
E-10 Core
DATAP
Data
Driver
1
9
DATAN
2
10
3
11
DATATP
Data
4
12
Magnetics and
Passive
Components
DATAI
5
Receiver
13
6
DATATN
14
7
15
8
COLLINP
Collision
Receiver
COLLINN
3.4.2.1 9-Wire AUI Transmit Interface
With the nine-wire AUI interface, the E-10 core outputs two transmit
signals, the true and complemented (differential) Manchester-encoded
data signals DATAP and DATAN. The transmit circuitry sends these
differential signals to the AUI cable through a transformer.
A transformer isolates the transmit pins from the transceiver cable,
achieving system isolation.
3-24
E-10 Functional Description
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
3.4.2.2 9-Wire AUI Receive Interface
The nine-wire twisted pair receive interface provides a receiver for data
and a receiver for collision control.
The data receiver has three signals at the E-10 core interface: DATATP,
DATATN, and DATAI.
The data receiver converts the two differential analog data signals from
the network to DATATP, DATATN, and DATAI. The three signals switch in
such a way that the E-10 core can properly perform data detection,
reverse polarity detection, and smart squelch.
The collision receiver has two signals at the E-10 core interface:
COLLINP and COLLINN.
3.4.3 Signal Quality Error (SQE) Function (Heartbeat)
The Signal Quality Error (SQE) diagnostic feature has been specified for
the MAU in the IEEE 802.3 standard. The feature is supported by the
E-10 core when it operates in the AUI mode with an external MAU
(see Figure 3.8).
The SQE test is a self-test feature supported in the MAU that is invoked
after the end of each transmission by the DTE, where the E-10 core is
located. When enabled, the SQE test consists of a 10-MHz burst sent
from the MAU over the collision pair and starts six to 16 bit times (0.6 µs
to 1.6 µs) after the last transition of the transmitted signal and lasts for
a duration of five to 15 bit times. This test is an indication to the E-10
core that the MAU has recognized the end of the transmission and the
MAU collision circuitry is intact and operational.
Ethernet-10 (E-10) Core Technical Manual
3-25
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
3-26
E-10 Functional Description
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
Chapter 4
Functional Timing
This chapter describes the Media Access Controller (MAC) and E-10
clock timing and contains the following sections:
♦
♦
♦
Section 4.1, “MAC Transmission and Reception Timing”
Section 4.2, “E-10 Clock Timing”
Section 4.3, “Multiple Core Clocking”
4.1 MAC Transmission and Reception Timing
This section explains MAC transmission and reception timing, and
Manchester encoder-decoder (ENDEC) timing.
4.1.1 MAC Transmission Timing
Figure 4.1 shows the MAC transmission timing. The transmission
process starts with external logic asserting TAVAILP to indicate that a
complete frame is ready to be transmitted (a). The transmitter starts
transmitting (the E-10 core asserts TRNSMTP HIGH) as soon as it is
finished deferring, and then waits an additional 9.6 µs for the interframe
gap (b). At this time, the E-10 core pulses the TSTARTP pin to indicate
that transmit data is needed within 6.4 µs on the transmit data bus
(TDATAI[7:0]). At this point, the E-10 core pulses the TREADDP pin
every 800 ns for every byte of transmit data read from the transmit data
bus. There are gaps in the stream of TREADDP pulses when the
transmitter does an automatic insertion of the source address, pad field,
or frame check sequence. External logic asserts the TLASTP signal,
accompanied by the last byte of transmit data.
Ethernet-10 (E-10) Core Technical Manual
4-1
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
The E-10 core stops transmission after the optional autoinsertion of the
pad and FCS fields and after all data has been transmitted to the
network (c). As soon as all transmit status bits are valid, the E-10 core
pulses the TDONEP pin to indicate that transmission of the frame is
complete (d). The TAVAILP signal is not sampled again until a minimum
of 800 ns after the TDONEP pulse, to give external logic time to do buffer
housekeeping, and to determine whether or not another buffer is ready
to be transmitted.
Figure 4.1 Successful Transmission
a
b
c
d
TAVAILP
TRNSMTP
TSTARTP
TLASTP
TREADDP
TDATAI[7:0]
TDONEP
D3 D4
DL
D1 D2
TSXXXX
(Status)
Note:TSXXXX consists of the following Transmit Status signals:
TSECOLP, TSIDEFP, TSLCOLP, TSMCOLP, TSNCLBP, TSOCOLP, and TSSQEEP
4-2
Functional Timing
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
4.1.1.1 Transmission Aborted by a Collision
A collision can abort a transmission. The timing for this condition is
shown in Figure 4.2. The transmission starts normally as shown in (a)
and (b). The transmitter transmits the complete preamble and start of
frame delimiter while it latches any collision. As soon as the transmitter
detects the presence of a collision after the SFD is transmitted (c), it
stops requesting additional transmit data bytes, transmits four bytes of
alternating ones and zeros and then terminates the transmission without
asserting a TDONEP pulse (d). After the transmission is terminated, the
transmitter starts the backoff algorithm, and then, after deferring, starts
a frame retransmission.
Figure 4.2 Transmission Aborted by a Collision
a
b
c
d
TAVAILP
TRNSMTP
TSTARTP
TLASTP
TREADDP
TDATAI[7:0]
TDONEP
D1 D2
D3
TSXXXX
(Status)
Note:TSXXXX consists of the following Transmit Status signals:
TSECOLP, TSIDEFP, TSLCOLP, TSMCOLP, TSNCLBP, TSOCOLP, and TSSQEEP
4.1.1.2 Retransmission
The host asserts AVAILP to the E-10 core when the host has one or
more frames available to be transmitted. A transmission that was aborted
by a collision initiates a retransmission. After the backoff algorithm and
deferral, the E-10 core starts the retransmission. The E-10 core does not
sample the TAVAILP signal during a retransmission, but the rest of the
timing is identical to any normal transmission. The host asserts the
TLASTP signal to indicate to the core that the TDATA[7:0] signals from
the host contain the last byte of data in the frame.
Ethernet-10 (E-10) Core Technical Manual
4-3
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
4.1.1.3 Transmission Aborted by TABORTP
When external logic decides it wants to terminate a transmission (for
example, after a transmission data underrun), it can assert the signal
TABORTP instead of sending a new transmit data byte. The timing for
this condition is shown in Figure 4.3. The transmission starts normally as
shown in (a) and (b). The transmitter terminates the transmission when
it detects the assertion of TABORTP (c) after transmitting four bytes of
alternating ones and zeros (d). This operation is identical to a
transmission aborted by a collision.
Figure 4.3 Transmission Aborted by TABORTP
a
b
c
d
TAVAILP
TRNSMTP
TSTARTP
TABORTP
TREADDP
TDATAI[7:0]
TDONEP
D1 D2
XX
TSXXXX
(Status)
Note:TSXXXX consists of the following Transmit Status signals:
TSECOLP, TSIDEFP, TSLCOLP, TSMCOLP, TSNCLBP, TSOCOLP, and TSSQEEP
4.1.1.4 Transmission Terminated With TFINISHP
When the signal TFINISHP is asserted at the end of a transmission, the
MAC terminates the transmission with the assertion of TDONEP,
independent of the detection of a collision. The frame is not retried, even
if it normally would have been. Assertion of TFINISHP can prevent a
frame from being retried independent of the normal default conditions.
For example, the transmit status pin TSLCOLP can be connected to
TFINISHP to prevent frames with a late collision from being retried.
4-4
Functional Timing
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
4.1.2 MAC Reception Timing
MAC reception timing is shown in Figure 4.4. When the E-10 core
receives a frame from the network, the data recovery digital PLL locks
onto the transitions of the preamble. Phase-lock occurs within six to eight
bit times (three bits for intelligent squelch operation and the remainder
for the PLL to operate). When phase-lock occurs, the E-10 core asserts
the RLOCKEDP pin (a). The receiver then goes into receiving mode and
asserts RCVNGP (b) when the start of frame delimiter is detected (b).
The beginning of a reception is marked with an RSTARTP pulse, which
prepares the off-core logic for the arrival of new data. The off-core logic
has 800 ns to make a new receive buffer available before new data bytes
start arriving (indicated by RBYTEP pulses). Reception is stopped when
the end of frame delimiter is detected and the receive status pins
(RSxxxx) are valid. Section 2.4, “Receiver Status Signals,” lists the
individual receive status pins. The RDONEP pulse indicates to external
logic that the received frame is complete (c).
Figure 4.4 Successful Reception
a
b
c
RLOCKEDP
RCVNGP
RSTARTP
RBYTEP
D1 D2 D3
DL-1
DL
RDATAO[7:0]
RDONEP
RCLEANP
RSXXXX
(Status)
Note:RSXXXX consists of the following Transmit Status signals:
RSDRBLEP, RSFCSEP, RSMATIP, RSMATMP and RSPLLEP
Ethernet-10 (E-10) Core Technical Manual
4-5
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
4.1.2.1 Reception With No Individual Address Filter Match
Figure 4.5 shows the timing of the address filters during reception.
Packet reception begins at (a). When all address filters do not match
during a reception of a frame from the network (b), the receiver stops
transferring received data to the host. When reception stops, the receiver
terminates with the assertion of the RCLEANP pulse (c).
Figure 4.5 Reception With No Individual Address Filter Match
a
b
c
RCVNGP
RSTARTP
RBYTEP
D1 D2 D3 D4 D5
D6
RDATAO[7:0]
RDONEP
RCLEANP
RSXXXX
(Status)
Note:RSXXXX consists of the following Transmit Status signals:
RSDRBLEP, RSFCSEP, RSMATIP, RSMATMP and RSPLLEP
4-6
Functional Timing
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
4.1.2.2 Reception With External Multicast Filter Match
Figure 4.6 shows the external multicast filter match timing. Packet
reception begins at (a). When multicast is enabled, and a multicast frame
is received from the network, the receive engine calculates a nine-bit hash
function (MINDEXO[8:0]) from the 48-bit multicast destination address.
The receiver indicates to external logic to either accept or reject the
message with the MSTARTP pulse (b). External logic has until the end of
a minimum length frame (51.2 µs from the SFD) to make a decision.
When the external logic decides to accept the frame, it asserts the MBITP
signal and pulses the MDONEP input (c). The receiver now permanently
asserts the RSMATMP signal and accepts the rest of the frame, and when
the receiver is finished, it asserts the RDONEP pulse (d).
Figure 4.6 Reception With External Multicast Filter Match
a
b
c
d
RCVNGP
RSTARTP
RBYTEP
D1 D2 D3 D4 D5 D6 D7 D8
DL-1 DL
RDATAO[7:0]
MSTARTP
MINDEXO[8:0]
MDONEP
MBITP
RSMATMP
RDONEP
Ethernet-10 (E-10) Core Technical Manual
4-7
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
4.1.2.3 Reception With No External Multicast Filter Match
Figure 4.7 shows the no external multicast filter match timing. Packet
reception begins at (a). A pulse on MSTARTP (b) indicates to the external
logic that a new nine-bit hash function has been computed from the
destination address. When the external logic decides to reject the frame,
it deasserts the MBITP signal and pulses the MDONEP input to indicate
to the receiver that it has made the decision (c). The receiver now
deasserts the RSMATMP signal and rejects the rest of the frame. When
the receiver is finished, the E-10 core pulses the RCLEANP pin (d).
Figure 4.7 Reception With No External Multicast Filter Match
a
b
c
d
RCVNGP
RSTARTP
RBYTEP
D1 D2 D3 D4 D5 D6
D7
RDATAO[7:0]
MSTARTP
MINDEXO[8:0]
MDONEP
MBITP
RSMATMP
RCLEANP
4-8
Functional Timing
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
4.1.3 ENDEC Timing
Examples of ENDEC timing are shown in Figure 4.8 and Figure 4.9. The
bit times shown in the diagrams are 100 ns in duration.
Figure 4.8 Twisted Pair Serial Output Data
1
Idle
0
1
1
0
0
EOF (300 ns)
Idle
DATAP
DATAN
PREEP
PREEN
Externally
Combined
Signal
Figure 4.9 Twisted Pair Link Integrity Pulses
100-ns
Idle
Idle
Pulse
DATAP
DATAN
PREEP
PREEN
Externally
Combined
Signal
Ethernet-10 (E-10) Core Technical Manual
4-9
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
4.2 E-10 Clock Timing
This section explains the E-10 clock, receiver, and transmitter timing.
The input clock to the E-10 core is 80 MHz. The duty cycle should have
a nominal duty cycle of 50/50, with a worst-case duty cycle of 60/40
because both the LOW-to HIGH and the HIGH-to-LOW transitions of this
signal are used in the digital phase-locked loop that samples the input
data. The MAC timing logic divides the 80 MHz down to 20 MHz and
10 MHz. These derived frequencies are available on the E-10 output pins.
Noninverting buffers have to be put outside of the E-10 core. The outputs
of these buffers have to be routed back to the E-10 input pins as shown
in Figure 4.10. The E-10 core expects all input signals from the host to
be synchronous with its own 10-MHz clock rate. This requirement for
synchronism dictates that the external logic use the same synchronous
10-MHz clock driver that E-10 uses. If more than one E-10 core resides
on the same chip, only the clock drivers from a single E-10 core can drive
all clock inputs and the external logic.
Figure 4.10 E-10 Clock Buffers
E-10 Core
CLK80
from clock_driver
or PLL
CLK80
CLK20I
CLK10I
CLK20
CLK10
CLK20O
CLK10O
Because the transmit clock is derived from the 80 MHz input clock, the
maximum clock error has to comply with the IEEE 802.3 maximum error
of ± 0.01% or 100 ppm.
4-10
Functional Timing
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
4.3 Multiple Core Clocking
Multiple E-10 cores can be used in an ASIC. This section describes how
multiple E-10 core clocking can best be accomplished.
4.3.1 Core Clocks
The E-10 core uses three different clocks:
♦
80 MHz. This clock is used as the primary timing input, and is used
in several places, as follows:
–
–
–
Clocking the data recovery digital PLL
Clocking the Ethernet signal squelch circuits
Generating the lower frequency clocks (20 and 10 MHz)
♦
♦
20 MHz. This clock is used for the Manchester encoding.
10 MHz. This main clock drives the MAC, ENDEC, and Transceiver
state machines.
The E-10 core does not contain any clock drivers. The designer should
add two clock buffers (see Figure 4.10) of sufficient strength to drive all
E-10 cores in parallel. The 10- and 20-MHz clock buffers should be
driven by the 10- and 20-MHz output pins of only one of the E-10 cores.
4.3.2 80-MHz Clock
The 80-MHz clock can be generated by a Phase-Locked Loop (PLL),
which consists of a phase comparator, external loop filter, and voltage-
controlled oscillator (see Figure 4.11). The reference input can be a
10- or 20-MHz clock or can be driven by an on chip 10- or 20-MHz
crystal oscillator. Generating the 80-MHz clock on the chip reduces the
higher frequencies on the printed circuit board and should allow for
easier passing of the FCC emissions test.
Ethernet-10 (E-10) Core Technical Manual
4-11
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
4.3.3 10- and 20-MHz Clocks
The 10- and 20-MHz inputs should be synchronized to the 80-MHz input
clock. The E-10 core generates 20- and 10-MHz clocks and makes these
three signals available on core output pins. The 20- and 10-MHz output
pins of only one of the E-10 cores should be used to drive all E-10 cores
on the chip. Most of the input signals of the E-10 core are synchronous
and should be clocked with the same 10-MHz clock as is used by the
cores. The 10- and 20-MHz clock buffers should be strong enough to
allow for a delay of no more than 9 ns from CLK80 input to the CLK10
and CLK20 inputs.
4.3.4 E-10 Testing
To allow testing of the E-10 cores, the three clocks used by the cores
should be multiplexed with three signals coming directly from the input
pins. The E-10 test vectors need direct control over the three test inputs
plus the multiplexer control input. The clock multiplexing scheme is
shown in Figure 4.11.
4-12
Functional Timing
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
Figure 4.11 Multiple-Core Clock Multiplexing
Test_CLK80
Clock
E-10
Multiplexers Buffers
Core #1
80-MHz
VCO
CLK80I
CLK20O
CLK20I
CLK10O
Not Connected
Not Connected
Test_CLK20
Test_CLK10
CLK10I
E-10
Core #2
Loop
Phase
FiLter
Compare
CLK80I
CLK20I
CLK10I
Clock Reference
10- or 20-MHz
Test_Enable
CLK20O
CLK10O
Not Connected
Not Connected
E-10
Core #n
10- or 20-MHz
External Clock
CLK80I
CLK20I
CLK10I
CLK20O
CLK10O
Clocks for
Synchronous
Interface
Ethernet-10 (E-10) Core Technical Manual
4-13
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
4-14
Functional Timing
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
Chapter 5
Specifications
This chapter provides specifications for the E-10 core, including the
AC timing as well as input and output loading, driver type, and power
consumption.
This chapter has three sections:
♦
♦
♦
Section 5.1, “Derivation of AC Timing and Loading”
Section 5.2, “AC Timing”
Section 5.3, “Pin Summary”
5.1 Derivation of AC Timing and Loading
Every customer that buys a core from LSI Logic for incorporation into an
ASIC also receives delay predictor software, which gives you an input
loading report so you can plan for buffer strengths that drive core inputs.
When you integrate the core into the rest of your logic and run other
simulation software, you get a ramptime violation report that indicates if
a core output is too heavily loaded. You can then adjust buffering, wire
length, and other parameters to eliminate the violation.
As a result, there are no specific numbers in this chapter for AC timing
and loading, because the numbers depend on the technology you use
and the design layout.
5.2 AC Timing
This section gives a list of the signals that must operate properly with
respect to the CLK10I and CLK20O signals. The relationship between
the signals is depicted in Figure 5.1 and Figure 5.2.
Ethernet-10 (E-10) Core Technical Manual
5-1
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
The numbers in Figure 5.1 and Figure 5.2 refer to the AC timing
parameters listed in the first column of Table 5.1. The core has been
verified under worst-case process voltage and ambient temperature.
Table 5.1 E-10 Core AC Timing Parameters
Parameter
Description
TFINISHP setup to CLK10I
TFINISHP hold from CLK10I
TAVAILP setup to CLK10I
TAVAILP hold from CLK10I
TDATAI[7:0] setup to CLK10I
TDATAI[7:0] hold from CLK10I
TABORTP setup to CLK10I
TABORTP hold from CLK10I
TLASTP setup to CLK10I
TLASTP hold from CLK10I
RENABP setup to CLK10I
RENABP hold from CLK10I
MDONEP setup to CLK10I
MDONEP hold from CLK10I
MBITP setup to CLK10I
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
MBITP hold from CLK10I
TSTARTP delay from CLK10I
TDONEP delay from CLK10I
TREADDP delay from CLK10I
TRNSMTP delay from CLK10I
RSTARTP delay from CLK10I
5-2
Specifications
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
Table 5.1 (Cont.) E-10 Core AC Timing Parameters
Parameter
Description
RDONEP delay from CLK10I
RCLEANP delay from CLK10I
RBYTEP delay from CLK10I
RDATAO[7:0] delay from CLK10I
RCVNGP delay from CLK10I
RLOCKEDP delay from CLK10I
DATAP delay from CLK20O
DATAN delay from CLK20O
PREEP delay from CLK20O
PREEN delay from CLK20O
LPASSP delay from CLK10I
LPBADP delay from CLK10I
COLLOUTP delay from CLK10I
MSTARTP delay from CLK10I
MINDEXO[8:0] delay from CLK10I
RESETN pulse width
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
RESETN LOW before CLK10I rise
RESETN HIGH before CLK10I rise
Ethernet-10 (E-10) Core Technical Manual
5-3
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
Figure 5.1 E-10 Core Setup and Hold Timing Diagram
CLK10I
1, 3, 5, 7, 9, 11, 13, 15
2, 4, 6, 8, 10, 12, 14, 16
Valid
Inputs
17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 32, 33, 34, 35, 36
Valid
Outputs
CLK20O
Outputs
28, 29, 30, 31
Valid
Figure 5.2 E-10 Core Reset Timing Diagram
37
RESETN
38
39
CLK10O
5.3 Pin Summary
Table 5.2 summarizes the E-10 core input and output signals. The table
provides the signal names and types for both outputs and inputs.
5-4
Specifications
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
Table 5.2 E-10 Pin Summary
Mnemonic
Description
Type
Active
CLK80I
80-MHz Clock
Input
Input
Input
Output
Output
Input
Input
–
–
–
–
–
–
–
CLK10I
10-MHz Clock
CLK20I
20-MHz Clock
CLK10O
CLK20O
COLLINN
COLLINP
COLLOUTP
DATAI
10-MHz Clock
20-MHz Clock
Negative Collision Threshold
Positive Collision Threshold
Transmit Collision
Data Input from Network
Positive Data
Output HIGH
Input
–
–
–
–
–
DATAP
Output
Output
Input
DATAN
Negative Data
DATATP
Positive Data Threshold
Negative Data Threshold
Drive Enable
DATATN
Input
DRIVEENN
DUPLEXP
FASTESTN
LCORPP
LFORCEP
LOOPBKP
LPASSP
LPBADP
MBITP
Output LOW
Full Duplex Select
Fast Test
Input
Input
Input
Input
Input
HIGH
LOW
HIGH
HIGH
HIGH
Link Correct Polarity
Force Link
Loopback Enable
Link Pass
Output HIGH
Output HIGH
Link Polarity Bad
Multicast Hash Table Bit
Multicast Filter Done
Input
Input
HIGH
HIGH
MDONEP
(Sheet 1 of 3)
Ethernet-10 (E-10) Core Technical Manual
5-5
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
Table 5.2 (Cont.) E-10 Pin Summary
Mnemonic
Description
Type
Active
MINDEXO[8:0] Multicast Hash Function
Output
–
MSTARTP
Multicast Filter Start
Node Address
Output HIGH
NOAD-
Input
–
DRI[47:0]
PREEP
Positive Data Pre-Emphasis
Negative Data Pre-Emphasis
Received Byte
Output
Output
–
–
PREEN
RBYTEP
Output HIGH
Output HIGH
Output HIGH
RCLEANP
RCVNGP
RDATAO[7:0]
RDONEP
RENABP
RESETN
RLOCKEDP
RMCENP
RPMENP
RSDRBLEP
RSFCSEP
Receiver Cleanup
Receiving
Received Data
Output
–
Receiver Done
Output HIGH
Receive Enable
Input
Input
HIGH
LOW
Reset
Receiver Locked Output
Receive Multicast Enable
Receive Promiscuous Mode Enable
Receive Status Dribble Bits
Output HIGH
Input
Input
HIGH
HIGH
Output HIGH
Output HIGH
Receive Status Frame Check Sequence
Error
RSMATIP
RSMATMP
RSPLLEP
RSTARTP
Receive Status Match Individual Address Output HIGH
Receive Status Match Multicast Address Output HIGH
Receive Status Phase-Locked Loop Error Output HIGH
Receiver Start Output HIGH
RSNOOPENP Receiver Lookback Enable
Input
Input
HIGH
–
SELIO
4-Wire/8-Wire Mixed Signal Cell Select
(Sheet 2 of 3)
5-6
Specifications
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
Table 5.2 (Cont.) E-10 Pin Summary
Mnemonic
Description
Type
Active
TABORTP
TAVAILP
Transmit Abort
Input
Input
Input
Input
HIGH
HIGH
–
Transmitter Packet Available
Transmit Control
TCTRI[3:0]
TDATAI[7:0]
TDONEP
TESTSE
Transmit Data
–
Transmit Done
Output HIGH
Scan Enable
Input
Input
Output
Input
Input
Input
Input
Input
Input
HIGH
–
TESTSI
Scan Input
TESTSO
TEST1N
Scan Output
–
Fast Test
LOW
LOW
HIGH
HIGH
HIGH
–
TEST2N
Reset B
TFINISHP
TLASTP
Transmit Finish
Transmit Last Byte
Test Mode
TMODE
TP/AUI
Twisted-Pair/AUI Select
Transmitter Read Control
Transmitting
TREADDP
TRNSMTP
TSECOLP
TSIDEFP
TSLCOLP
TSMCOLP
TSNCLBP
TSOCOLP
TSSQEEP
TSTARTP
Output HIGH
Output HIGH
Output HIGH
Output HIGH
Output HIGH
Output HIGH
Output HIGH
Output HIGH
Output HIGH
Output HIGH
Transmit Status Excessive Collisions
Transmit Status Initially Deferred
Transmit Status Late Collision
Transmit Status Multiple Collision
Transmit Status No Carrier Loopback
Transmit Status One Collision
Transmit Status SQE Error
Transmit Start
(Sheet 3 of 3)
Ethernet-10 (E-10) Core Technical Manual
5-7
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
5-8
Specifications
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
Appendix A
Glossary
Address Filtering (Individual, Multicast, Promiscuous) – Address
filtering is the process the E-10 core performs to match a destination
address within a received frame with an address derived at the receiving
station. Three common types of address filtering are:
♦
Individual—the first bit of an individual address is a zero. The
incoming destination address is compared with the data from the
individual address pins. When all 48 bits match and the receiver is
enabled, the address filter passes.
♦
Multicast—the first bit of an multicast address is a one. When the
destination address bits that are received in the frame contain a
multicast address, the E-10 core uses its built-in FCS generator to
compute a nine-bit polynomial (the nine MSBs of the 32-bit FCS
generator) from the incoming address. The value of this polynomial
can be used as an index into an external multicast filter hash table.
External logic can decide to either accept or reject the incoming
frame. The external logic enables the multicast address filter.
♦
Promiscuous—when the external logic asserts the individual address
promiscuous mode enable pin (RPMEN) HIGH and the destination
address is an individual address, the address filter always passes,
and the incoming frame is received.
Attachment Unit Interface (AUI) – The AUI is the interface between the
Medium Attachment Unit (MAU) and the data terminal equipment (DTE)
device or repeater. A typical AUI interface consists of a 15-pin “D” connector.
Backoff – In IEEE 802.3 networks, backoff occurs when two or more
nodes attempt a transmission and collide. The function of stopping
transmission, and waiting a specified random time before retrying the
transmission is considered backoff. In IEEE 802.3 networks a “truncated
binary exponential backoff” algorithm is employed.
Ethernet-10 (E-10) Core Technical Manual
A-1
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
Coaxial Cable – Coaxial cable is a transmission medium with a central
copper-wire conductor surrounded by concentric layers of plastic/polyvinyl
chloride, aluminum or aluminized mylar and a copper tube that acts as
an insulator (ground) and source of shielding from electromagnetic and
radio frequency interference (CMI/RFI). Two types of coaxial cable—
known as “thick” and “thin” for their respective diameters—are used in
Ethernet data transmissions.
Collision – When an Ethernet station is operating in the AUI mode, it
can sense a change in the energy level of the communication channel
and interpret the phenomenon as a collision. A collision is caused by two
stations attempting to transmit at the same time.
Collision Detection – Collision detection is the ability of a transmitting
node on an Ethernet LAN to sense a change in the energy level of the
channel and to interpret the phenomenon as a collision
CRC (Cyclic Redundancy Check) – CRC is a basic error-checking
mechanism for link-level data transmission; a characteristic link-level
feature of (typically) bit-oriented data communications protocols. The
data integrity of a received frame or packet is checked via a polynomial
algorithm based on the content of the frame and then matched with the
result that is performed by the sender and included in a (most often,
16-bit) field appended to the frame.
CSMA/CD (Carrier Sense Multiple Access with Collision
Detection) – CSMA/CD is a LAN protocol access method where the
nodes are attached to a cable. When a node transmits data onto the
network and raises the carrier, the remaining nodes detect the carrier
(Carrier Sense) and “listen” for the information to detect if it is intended
for them. The nodes have network access (Multiple Access) and can
send if no other transmission is taking place. If two nodes attempt to
send simultaneously, a collision takes place (Collision Detection) and
both nodes must retry at random intervals.
Data Recovery – The data sent over an Ethernet LAN is Manchester-
encoded and must be recovered by the receiving station. In the E-10
core, a digital phase-locked loop recovers the data through a sampling
process and has the necessary elasticity to allow for the bit drift that
occurs over lengthy frame sizes.
A-2
Glossary
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
Dribble Bit – A dribble bit is a bit that is “extra” and occurs when, at the
end of a frame, less than a complete byte is received. It is possible to
have from one to seven dribble bits at the end of a frame if a complete
byte is not received.
Dropout Error – A dropout error occurs when the carrier on a
transmission channel stops due to a broken or missing cable.
Elasticity – Elasticity is a measure of the tolerance of the phase-locked
loop to the cumulative bit shift due to small differences in the free
running clock oscillator frequencies of the transmitting source and
receiving destination.
End-of-Frame Delimiter – An end-of-frame delimiter is a series of three
bits of ones at the end of a frame that are not Manchester-encoded; that is,
they are non-return-to-zero bits that do not contain a transition in the
middle of the bit time. The purpose of the end-of-frame delimiter is to
cause the PLL at the receiving station to lose lock so the host can be
notified that the frame has ended.
When a receiving node detects a bit that is not Manchester-encoded, it
considers the previous four bytes to be the FCS.
ENDEC – ENDEC is short for Encoder/Decoder and is a functional block
within network adapters that performs two basic functions. First, this
function encodes the data from the controller to be transmitted over the
network. Second, it decodes the data on the network to a form suitable
for the network controller chip. In the case of Ethernet, the ENDEC
converts NRZ controller data to Manchester data (and back again).
Ethernet Local Area Network (LAN) – An Ethernet LAN is a branching
broadcast communications system for carrying digital data packets
among locally distributed computing stations. Ethernet is a 10 Mbit/s
baseband, Local Area Network that has evolved into the IEEE 802.3
specification and is defined by a data-link protocol that specifies how
data is placed on and retrieved from a common transmission medium.
Ethernet is used as the underlying transport vehicle by several upper
level protocols, including TCP/IP and Xerox Network System (XNS).
See IEEE 802.3.
Ethernet-10 (E-10) Core Technical Manual
A-3
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
Frame – An Ethernet frame contains the following eight fields:
♦
♦
♦
♦
♦
♦
♦
♦
Preamble
Start-of-Frame Delimiter
Destination Address
Source Address
Length
Data
Pad (if necessary)
Frame Check Sequence
Frame Check Sequence (FCS) – The Ethernet transmit and receive
algorithm uses the standard IEEE 802.3 four-byte frame check sequence
(FCS) field to ensure data integrity. During data transmission, the E-10
core uses a 32-bit linear feedback shift register (LFSR) to compute the
value that will be sent in the FCS field. The FCS value is a function of
the content of the source address, destination address, length, data, and
pad fields. On data reception, the E-10 core preloads the LFSR with all
ones and updates this value with each byte received, including the
received FCS. If the final value in the LFSR does not match a
predetermined value after all bytes including the FCS are received, the
E-10 core flags an FCS error.
Heartbeat – In IEEE 802.3 networks, a heartbeat is a short burst of
collision signal that is transmitted from the MAU to the DTE after every
packet. See Signal Quality Error (SQE) Test.
Hub – The E-10 core operating in 10BASE-T mode uses a hub to
concentrate connections to multiple terminals. Hubs come in various
sizes, with 4-port, 8-port, and 12-port being the most common. The
number of ports indicates the number of terminals that can be connected
to the hub. Most hubs have built-in transceivers and network
management features. Hubs are mainly used in star topology LANs. The
E-10 core is specifically designed for multiple-core integration, especially
for “hub-on-a-chip” implementations.
A-4
Glossary
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
IEEE 802.3 – IEEE 802.3 is a standard set by the IEEE for CSMA/CD
network protocol, which is a Physical Layer definition including
specifications for cabling in addition to transmitting data and controlling
cable access. See Ethernet LAN.
Jam – In IEEE 802.3 networks, when a collision occurs the colliding
nodes ensure that the collision is seen by the entire network by
continuing to transmit for a minimum time during a collision. This
occurrence is known as jamming.
Jitter – Jitter is the slight movement of a transmission signal in time or
phase that can introduce errors and loss of synchronization in high-
speed synchronous communications.
LAN (Local Area Network) – A LAN is a communications system
linking computers together to form a network whose dimensions typically
are less than five kilometers. Transmissions within a Local Area Network
generally are digital, carrying data among stations at rates usually above
one megabit per second. A LAN is an assembly of computing resources
such as microcomputers (for example, PCs), printers, minicomputers and
mainframes linked by a common transmission medium, including coaxial
cable or twisted-pair wiring.
Link Failure – In the Twisted-Pair mode, a link failure can be caused by
a twisted-pair transceiver failure or by a broken twisted-pair receive
cable. A link failure occurs when there are eight consecutive missing link
integrity pulses.
Link Integrity Pulses – The E-10 core monitors Link Integrity pulses,
which may be sent on a regular basis from any other device on the
network when the channel is idle. The purpose of the pulses is to ensure
that the channel is functional even in the absence of active data. The
Link Integrity pulses should be positive-going pulses spaced
approximately 50 milliseconds apart.
Link Polarity Detection/Correction – Link Integrity pulses should be
positive-going pulses. If they are not, it is an indication that the wiring is
inverted. The E-10 core can optionally, under control of the host, invert
the data on the receive pair when it detects that eight consecutive Link
Integrity pulses have been received with inverted polarity.
Ethernet-10 (E-10) Core Technical Manual
A-5
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
Lock Time – Lock time is the amount of time it takes the receiver phase-
locked loop to achieve phase lock with an input signal, which may be
either data bits or preamble bits. It takes approximately six to eight bits
for the E-10 digital phase-locked loop circuit to achieve lock. Three bits
are required for the smart squelch operation and the remaining bits are
required for the PLL to operate.
Loopback – When the E-10 core is in loopback mode, the transmitter
output is fed back to the receiver input inside the chip. Loopback mode
allows testing of the E-10 transmitter and receiver up to but not including
the off-core transceivers under program control without the need of an
external loopback connector. Loopback is applicable to both the AUI and
Twisted-Pair modes.
Another loopback mode is known as “lookback.” When the E-10 core is
in lookback mode, all transmitted frames are routed back through the
receiver. Operation is similar to that of normal loopback described in the
previous paragraph except that the phase-locked loop is bypassed.
MAC (Media Access Control) – The data link sublayer that is
responsible for transferring data to and from the physical layer.
Manchester Encoding/Decoding – A Manchester-encoded bit contains
a transition (either HIGH-to-LOW or LOW-to-HIGH) in the center of the
bit time. Having a transition every bit time allows a phase-locked loop at
a receiving station to maintain lock and extract both data and clock from
the waveform.
With Manchester encoding, each bit is divided into two complementary
halves. A LOW-to-HIGH voltage transition in the middle of the bit period
designates a binary one, while a HIGH-to-LOW transition represents a
binary zero.
MAU (Media Attachment Unit) – An MAU is the physical and electrical
component that provides the means of attaching a terminal to a local
area network medium.
Multicast Hash Function – See address filtering (multicast).
Packet – A packet is data sent in the data field of an Ethernet frame.
A-6
Glossary
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
Phase-Locked Loop Circuit – A phase-locked loop circuit operates
from a local free-running oscillator and is able to achieve and maintain
phase and frequency with a signal received from a source that is
operating on a different free-running oscillator. The purpose of a phase-
locked loop is to provide a clock source that is synchronous with data
that is derived from a different clock source. Such a circuit is used
commonly in telecommunications systems where data is transmitted
without an accompanying clock.
Pre-Emphasis – The purpose of pre-emphasis is to increase the
amplitude of the higher frequencies in an output signal. Data pre-emphasis
guarantees that an Ethernet system operating Twisted-Pair mode can drive
signals properly at 10 Mbit/s on cable up to 100 meters long.
Router – A router is similar to a bridge but has the ability to connect
networks at the OSI network level, as shown in Figure A.1.
Figure A.1 Router OSI Interconnect Level
Application
Presentation
Session
Application
Presentation
Session
Transport
Network
Transport
Network
LLC Sublevel
MAC Sublevel
Data Link
Physical
Data Link
Physical
Router
Source: In-Stat
A router is an intelligent device that plans the most efficient route for data
to move between networks. Routers generally have the ability to recover
lost data packets. The end user must address the router to send data
through it. Routers are slower in the transfer of data than bridges and
are more expensive to purchase. For this reason, routers are not used
to connect LANs (although it is possible) but are more commonly used
to connect LANs to wide area networks.
Ethernet-10 (E-10) Core Technical Manual
A-7
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
Runt Packet – In IEEE 802.3 networks, a runt packet is a special case of
a fragment packet where the length of the packet is less than 512 bit times.
Signal Quality Error (SQE) Test – The Signal Quality Error (SQE) test
is a self-test feature supported in the MAU that is invoked after the end
of each transmission by the station. When enabled, the SQE test
consists of a 10-MHz burst and starts six to sixteen bit times (0.6 µs to
1.6 µs) after the last transition of the transmitted signal and lasts for a
duration of five to 15 bit times. This test is an indication to the station
that the MAU has recognized the end of the transmission and the MAU
collision circuitry is intact and operational.
Slot Time – A slot time is 512 bit times (51.2 µs), and is the time it takes
to “fill up” an Ethernet cable with data. In AUI mode, if the receive
circuitry of a station does not detect a carrier within one slot time from
the initiation of transmission (preamble start), it is an indication of trouble
on the cable.
Smart Receive Squelch – An intelligent (smart) receive squelch is
implemented on the receiver differential inputs to ensure that impulse
noise on the receive inputs is not mistaken for a valid signal. The smart
squelch logic uses a combination of amplitude and timing measurements
to determine the validity of data.
SQE Burst (Heartbeat) – See Signal Quality Error (SQE) Test.
Start of Frame Delimiter (SFD) – The eight-bit start of frame delimiter
field follows the seven-octet preamble at the beginning of a frame and
contains the 10101011 bit pattern. This pattern allows synchronization
2
of the data received in the rest of the frame.
Tracking Speed – Tracking speed refers to the speed with which a PLL
responds to changes in the incoming signal. In the E-10 core, when first
attempting to lock on to an oncoming data stream (during the first 4.8
µs), the PLL looks every 16 bits (1.6 µs) to see if a bit transition is
occurring in the center of the sampling window. After the first 4.8 µs, the
PLL maintains a window of 64 bits (6.4 µs) for finding a transition.
A-8
Glossary
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
Transceiver – A transceiver is a combined transmitter and receiver and
is an essential element of all LAN networks. When an Ethernet LAN
operates in the 10BASE-2 or 10BASE-5 mode, the transceivers are
generally located in the MAU and connect directly to the coaxial cable.
In the Twisted-Pair mode, the transceivers are generally located in the
data terminal equipment and connect directly to the twisted-pair cable.
Transmit Carrier Loopback – Transmit Carrier Loopback occurs in the
AUI mode. When a station transmits onto the wire, it should detect its
own activity on the line within 512 bit times (51.2 µs), which is the
maximum amount of time to “fill the wire” under normal circumstances.
Transmit Carrier Loopback Error – Transmit Carrier Loopback Error
occurs when a station does not detect its own activity on the line within
512 bit times (51.2 µs).
Transmit Carrier Dropout Error – A transmit carrier dropout error
occurs when the carrier on a transmission channel stops due to a broken
or missing cable, or bad transceivers.
Twisted-Pair Transmission System – In 10BASE-T terminology, this
term refers to the twisted-pair wire link and its two attached MAUs.
Twisted-Pair Wire – Twisted-pair wire is a cable comprised of two 18 to
24 AWG (American Wire Gauge) solid copper strands twisted around
each other. The twisting provides a measure of protection from
electromagnetic and radio-frequency interference (EMI/RFI). Two types
are available: shielded and unshielded. The former is wrapped inside a
metallic sheath that provides protection from EMI/RFI. The latter, also
known as telephone wire, is covered with plastic or PVC, which provides
no protection from EMI/RFI.
10BASE-2 – 10BASE-2 is the IEEE 802.3 Physical Layer Standard for
thin wire Ethernet (sometimes called cheapernet). This standard uses
RG58 standard coaxial cable. 10BASE2 stands for: 10 = 10 Mbit/s data
rate, BASE = Baseband, 2 = 185 meter segment length.
10BASE-5 – 10BASE-5 is the IEEE 802.3 Physical Layer Standard for
thick cable Ethernet, utilizing thick double shielded coaxial cable.
10BASE-5 stands for: 10 = 10 Mbit/s data rate, BASE = Baseband,
5 = 500 meters segment length.
Ethernet-10 (E-10) Core Technical Manual
A-9
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
10BASE-F – 10BASE-F is the IEEE 802.3 Physical Layer Standard for
fiber-based Ethernet. 10BASE-F stands for; 10 = 10 Mbit/s data rate,
BASE = baseband, F= fiber.
10BASE-T – 10BASE-T is the IEEE 802.3 Physical Layer Standard for
the new twisted-pair Ethernet in a star topology. 10BASE-T stands for;
10 = 10 Mbit/s data rate, BASE = baseband, T = Twisted-pair over 100
meters nominal segment length.
A-10
Glossary
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
Customer Feedback
We would appreciate your feedback on this document. Please copy the
following page, add your comments, and fax it to us at the address on
the following page.
If appropriate, please also fax copies of any marked-up pages from this
document.
Important: Please include your name, phone number, fax number, and
company address so that we may contact you directly for
clarification or additional information.
Thank you for your help in improving the quality of our documents.
Ethernet-10 (E-10) Core Technical Manual
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
Reader’s Comments
Fax your comments to:
LSI Logic Corporation
Technical Publications
M/S F-112
Fax: 408.433.4333
Please tell us how you rate this document: Ethernet-10 (E-10) Core
Technical Manual. Place a check mark in the appropriate blank for each
category.
Excellent Good Average Fair
Poor
Completeness of information
Clarity of information
Ease of finding information
Technical content
Usefulness of examples and
illustrations
____
____
____
____
____
____
____
____
____
____
____
____
____
____
____
____
____
____
____
____
____
____
____
____
____
Overall manual
____
____
____
____
____
What could we do to improve this document?
If you found errors in this document, please specify the error and page
number. If appropriate, please fax a marked-up copy of the page(s).
Please complete the information below so that we may contact you
directly for clarification or additional information.
Name
Date
Telephone
Title
Fax
Department
Company Name
Street
Mail Stop
City, State, Zip
Customer Feedback
February 1997 - Rev. A
Copyright © 1996, 1997 by LSI Logic Corporation. All rights reserved.
U.S. Distributors
by State
Alabama
Colorado
Kansas
New York
Texas
Huntsville
Hamilton Hallmark
Tel: 800.633.2918
Colorado Springs
Hamilton Hallmark
Tel: 719.637.0055
Overland Park
Hamilton Hallmark
Tel: 800.332.4375
Hauppauge
Hamilton Hallmark
Tel: 516.434.7471
Austin
Hamilton Hallmark
Tel: 512.258.8848
Wyle Electronics
Tel: 800.964.9953
Wyle Electronics
Tel: 800.365.9953
Denver
Kentucky
Lexington
Hamilton Hallmark
Tel: 800.235.6039
Long Island
Wyle Electronics
Tel: 516.293.8446
♦ Wyle Electronics
Tel: 303.457.9953
Arizona
Phoenix
Hamilton Hallmark
Tel: 800.528.8471
Dallas
Hamilton Hallmark
Tel: 214.553.4302
Englewood
Hamilton Hallmark
Tel: 303.790.1662
Rochester
Hamilton Hallmark
Tel: 800.462.6440
Maryland
Baltimore
Wyle Electronics
Tel: 410.312.4844
Wyle Electronics
Tel: 800.955.9953
Wyle Electronics
Tel: 602.804.7000
Connecticut
Cheshire
Hamilton Hallmark
Tel: 203.271.5700
North Carolina
Raleigh
Hamilton Hallmark
Tel: 919.872.0712
Houston
Hamilton Hallmark
Tel: 713.787.8300
Tempe
Hamilton Hallmark
Tel: 602.414.7705
Columbia
Hamilton Hallmark
Tel: 800.638.5988
Wyle Electronics
Tel: 713.784.9953
Wyle Electronics
Tel: 919.469.1502
Florida
California
Culver City
Hamilton Hallmark
Tel: 310.558.2000
Fort Lauderdale
Hamilton Hallmark
Tel: 305.484.5482
Massachusetts
Boston
San Antonio
Wyle Electronics
Tel: 210.697.2816
Ohio
Cleveland
Wyle Electronics
Tel: 216.248.9996
♦ Wyle Electronics
Tel: 800.444.9953
Wyle Electronics
Tel: 305.420.0500
Irvine
Hamilton Hallmark
Tel: 714.789.4100
Peabody
Utah
♦ Hamilton Hallmark
Salt Lake City
Hamilton Hallmark
Tel: 801.266.2022
Largo
Hamilton Hallmark
Tel: 800.282.9350
Dayton
Hamilton Hallmark
Tel: 800.423.4688
Tel: 508.532.3701
♦ Wyle Electronics
Michigan
Plymouth
Hamilton Hallmark
Tel: 313.416.5800
Tel: 714.789.9953
Wyle Electronics
Tel: 801.974.9953
Wyle Electronics
Tel: 513.436.9953
Orlando
Wyle Electronics
Tel: 407.740.7450
Los Angeles
Wyle Electronics
Tel: 818.880.9000
Washington
Redmond
Hamilton Hallmark
Tel: 206.881.6697
Solon
Hamilton Hallmark
Tel: 216.498.1100
Minnesota
Bloomington
Hamilton Hallmark
Tel: 612.881.2600
Tampa/N. Florida
Wyle Electronics
Tel: 800.395.9953
Rocklin
Hamilton Hallmark
Tel: 916.624.9781
Toledo
Wyle Electronics
Tel: 419.861.2622
Seattle
Wyle Electronics
Tel: 800.248.9953
Winter Park
Hamilton Hallmark
Tel: 407.657.3317
Minneapolis
Wyle Electronics
Tel: 800.860.9953
Sacramento
Wyle Electronics
Tel: 916.638.5282
Worthington
Hamilton Hallmark
Tel: 614.888.3313
Wisconsin
Milwaukee
Wyle Electronics
Tel: 800.867.9953
Georgia
Atlanta
Wyle Electronics
Tel: 800.876.9953
Missouri
Earth City
Hamilton Hallmark
Tel: 314.291.5350
San Diego
Hamilton Hallmark
Tel: 619.571.7540
Oklahoma
Tulsa
Hamilton Hallmark
Tel: 918.254.6110
Wyle Electronics
Tel: 619.565.9171
New Berlin
Hamilton Hallmark
Tel: 414.780.7200
Duluth
Hamilton Hallmark
Tel: 800.241.8182
New Jersey
Mt. Laurel
Hamilton Hallmark
Tel: 609.222.6463
San Jose
Oregon
Beaverton
Hamilton Hallmark
Tel: 503.526.6200
♦ Hamilton Hallmark
Illinois
Tel: 408.435.3500
Arlington Heights
♦ Hamilton Hallmark
Tel: 847.797.7300
No. New Jersey
Wyle Electronics
Tel: 201.882.8358
Santa Clara
Wyle Electronics
Tel: 408.727.2500
Portland
Wyle Electronics
Tel: 503.643.7900
Chicago
Wyle Electronics
Tel: 847.620.0969
Parsippany
Hamilton Hallmark
Tel: 201.515.1641
Woodland Hills
Hamilton Hallmark
Tel: 818.594.0404
Pennsylvania
Philadelphia
Wyle Electronics
Tel: 800.871.9953
Iowa
Carmel
Hamilton Hallmark
Tel: 800.829.0146
New Mexico
Albuquerque
Hamilton Hallmark
Tel: 505.293.5119
♦ Distributors with
Design Resource
Centers
Sales Offices and Design
Resource Centers
LSI Logic Corporation
Corporate Headquarters
Tel: 408.433.8000
Fax: 408.433.8989
New Jersey
Edison
Denmark
LSI Logic Development
Centre
Korea
LSI Logic Corporation of
Korea Ltd
Seoul
♦ Tel: 908.549.4500
Fax: 908.549.4802
Ballerup
Tel: 45.44.86.55.55
Fax: 45.44.86.55.56
♦ Tel: 82.2.561.2921
NORTH AMERICA
New York
Fax: 82.2.554.9327
New York
California
Irvine
Tel: 716.223.8820
Fax: 716.223.8822
France
LSI Logic S.A.
Paris
Singapore
LSI Logic Pte Ltd
Singapore
Tel: 65.334.9061
Fax: 65.334.4749
♦ Tel: 714.553.5600
Fax: 714.474.8101
North Carolina
Raleigh
♦ Tel: 33.1.34.63.13.13
Fax: 33.1.34.63.13.19
San Diego
Tel: 619.635.1300
Fax: 619.635.1350
Tel: 919.783.8833
Fax: 919.783.8909
Germany
LSI Logic GmbH
Munich
Electronic Resources Ltd
Tel: 65.298.0888
Fax: 65.298.1111
Oregon
Silicon Valley
Sales Office
Tel: 408.433.8000
Fax: 408.954.3353
Design Center
Beaverton
Tel: 503.645.0589
Fax: 503.645.6612
♦ Tel: 49.89.4.58.33.0
Fax: 49.89.4.58.33.108
Spain
LSI Logic S.A.
Madrid
Stuttgart
Tel: 49.711.13.96.90
Fax: 49.711.86.61.428
Texas
Austin
Tel: 512.388.7294
Fax: 512.388.4171
♦ Tel: 34.1.3672200
♦ Tel: 408.433. 8000
Fax: 34.1.3673151
Fax: 408.433.2820
Hong Kong
Sweden
Colorado
Boulder
AVT Industrial Ltd
Hong Kong
LSI Logic AB
Stockholm
Dallas
Tel: 303.447.3800
Fax: 303.541.0641
♦ Tel: 214.788.2966
Tel: 852.2428.0008
Fax: 852.2401.2105
♦ Tel: 46.8.444.15.00
Fax: 214.233.9234
Fax: 46.8.750.66.47
Florida
Houston
India
Switzerland
Boca Raton
Tel: 407.989.3236
Fax: 407.989.3237
Tel: 713.379.7800
Fax: 713.379.7818
LogiCAD India Private Ltd
Bangalore
Tel: 91.80.526.2500
Fax: 91.80.338.6591
LSI Logic Sulzer AG
Brugg/Biel
Tel: 41.32.536363
Fax: 41.32.536367
Washington
Georgia
Bellevue
Atlanta
Tel: 770.395.3808
Fax: 770.395.3811
Tel: 206.822.4384
Fax: 206.827.2884
Israel
LSI Logic
Ramat Hasharon
♦ Tel: 972.3.5.403741
Fax: 972.3.5.403747
Taiwan
LSI Logic Asia-Pacific
Taipei
Canada
Ontario
♦ Tel: 886.2.718.7828
Illinois
Fax: 886.2.718.8869
Schaumburg
Ottawa
♦ Tel: 847.995.1600
♦ Tel: 613.592.1263
Netanya
Cheng Fong Technology
Corporation
Tel: 886.2.910.1180
Fax: 886.2.910.1175
Fax: 847.995.1622
Fax: 613.592.3253
♦ Tel: 972.9.657190
Fax: 972.9.657194
Kentucky
Toronto
Bowling Green
Tel: 502.793.0010
Fax: 502.793.0040
♦ Tel: 416.620.7400
Italy
LSI Logic S.P.A.
Milano
Fax: 416.620.5005
Lumax International
Corporation, Ltd
Quebec
Montreal
♦ Tel: 39.39.687371
Tel: 886.2.788.3656
Fax: 886.2.788.3568
Maryland
Fax: 39.39.6057867
Bethesda
♦ Tel: 514.694.2417
♦ Tel: 301.897.5800
Fax: 514.694.2699
Japan
Macro-Vision
Fax: 301.897.8389
LSI Logic K.K.
Tokyo
Technology Inc.
Tel: 886.2.698.3350
Fax: 886.2.698.3348
INTERNATIONAL
Massachusetts
Waltham
♦ Tel: 81.3.5463.7821
Australia
Fax: 81.3.5463.7820
♦ Tel: 617.890.0180
Reptechnic Pty Ltd
New South Wales
Tel: 612.9953.9844
Fax: 612.9953.9683
United Kingdom
LSI Logic Europe plc
Bracknell
Fax: 617.890.6158
Osaka
♦ Tel: 81.6.947.5281
Minnesota
Minneapolis
Fax: 81.6.947.5287
♦ Tel: 44.1344.426544
Fax: 44.1344.481039
♦ Tel: 612.921.8300
Fax: 612.921.8399
♦ Sales Offices with Design
Resource Centers
Printed in USA
Printed on
Recycled Paper
ISO 9000 Certified
相关型号:
ETHERNET_AU1000_REV003
Writing an Ethernet Device Driver for the AMD Alchemy? Solutions Au1000? Processor?
ETC
©2020 ICPDF网 联系我们和版权申明