Serenade Data Sheet
ENT3002—Wire-Speed Router IC for
Access Networks
Revision 1.2
Entridia Corporation Phone: (949) 823-3600
131 Theory, Sui te 150 Fax: (949 )75 2-14 62
Irvine, CA 92612 Email: info@entridia.com
USA http://www.entridia.com/
1998-2000 Entridia Corporation. All rights reserved. No part of this publication may be reproduced,
transmitted, transcribed, stored in a retrieval system, or translated into any language, in any form or by
any means, electronic, mechanical, photocopying, recording, or otherwise, without prior written permis-
sion from Entrid ia Cor poration.
The information furnished herein is believed to be accurate and reliable. However, no responsibility is as-
sumed by Entridia Corporation for its use, nor for any infringements of patents or other rights of third par-
ties resulti ng from its use.
Entridia, OPERA, Serenade, Chorus, Sonata, OptiStream, EdgeStream, and OCPort are trademarks of En-
tridia Corporation. All other trademarks are the property of their respective owners.
Revision 1.2 Preface iii
Preface
This document describes the design features and functional descriptions for the Entridia Sere-
nade™ wire-speed router IC for Access Networks.
Serenade is ideally suited for IP based aggregation systems in Access networks, and is part of
Entridia’s OPERA™ (Optical Edge Routing Architecture) family of products targeted at high
performance aggregation routers. Serenade incorporates two OC-3c Packet-over-Sonet (PoS)
ports via Entridia’s OCPort™ interface and offers connectivity to legacy WAN networks via
Entridia’s EdgeStream™ interface and glueless expansion via the OptiStream™ expansion
bus.
References
POS-PHY Saturn Compatible Packet Over SONET Interface Specification (Level2),
PMC-Sierra Inc. / Saturn Development Group, PMC-971147 Issue 5, December 1998
IETF Network Management RFCs
RFC 1812 Requirements for IP version 4 Routers
RFC 1213 Management Information Base for Network Management of TCP/IP
based internets: MIB-II
RFC 2474 Definition of th e Differentiated Services Field (DS Field) in the IPv4
and IPv6 Headers
RFC 2475 An Architecture for Differen tiated Series
iv ENT300 2 Serena de Data Sheet Revisio n 1.2
Revision 1.2 Table of Contents v
P R E L I M I N A R Y
Table of Contents
Preface- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - iii
Chapter 1 Functional Description - - - - - - - - - - - - - - - - - - - - 1
1.1 Data Link Layer (OSI Layer 2) Frame Processing- - - - - - - - - - - - - - - - - -1
1.2 Network Layer (OSI Layer 3) Packet Processing - - - - - - - - - - - - - - - - - -1
1.3 EdgeStream Interface - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -2
1.4 Route Table- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -2
1.5 Flow Table - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -3
1.6 Queue Management - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -4
1.7 OptiStream Expansion Bus - - - - - - - - - - - - - - - - - - - - - - - - - - - -4
1.8 Software Interface- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -5
1.9 System Integration - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -5
Chapter 2 Hardware Description- - - - - - - - - - - - - - - - - - - - - 7
2.1 Pin-Out- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -9
2.1.1 Package Details - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 10
2.2 Pin Signal Descriptions - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 14
2.2.1 OptiStream Interface Pin Signal Descriptions - - - - - - - - - - - - - - - 14
2.2.2 EdgeStream Interface Pin Signal Descriptions - - - - - - - - - - - - - - - 16
2.2.3 SRAM Interface Pin Signal Descriptions - - - - - - - - - - - - - - - - - 17
2.2.4 Network Interface Pin Signal Descriptions- - - - - - - - - - - - - - - - - 18
2.2.5 Analog, Power, Ground, and Utility Pin Signal Descriptions - - - - - - - - 20
2.3 External Components - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 20
2.3.1 External OptiStream Bus Routing - - - - - - - - - - - - - - - - - - - - - 21
2.3.2 External Reference Clock- - - - - - - - - - - - - - - - - - - - - - - - - 21
2.3.3 Band Gap Reference - - - - - - - - - - - - - - - - - - - - - - - - - - - 22
2.3.4 GTL+ Receiver Reference - - - - - - - - - - - - - - - - - - - - - - - - 22
2.3.5 External SRAM - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 23
Chapter 3 System Interface Operation - - - - - - - - - - - - - - - - - -29
3.1 Electrical Characteristics - - - - - - - - - - - - - - - - - - - - - - - - - - - - 29
3.1.1 Absolute Maximum Ratings - - - - - - - - - - - - - - - - - - - - - - - 29
3.1.2 Power Sequencing - - - - - - - - - - - - - - - - - - - - - - - - - - - - 29
3.1.3 D.C. Operating Characteristics - - - - - - - - - - - - - - - - - - - - - - 30
3.1.4 I/O Pin Types and Electrical Characteristics - - - - - - - - - - - - - - - - 31
3.2 External Timing Specification - - - - - - - - - - - - - - - - - - - - - - - - - - 31
vi ENT3003 Chorus Data Book Revision 1.2
P R E L I M I N A R Y
Chapter 4 Programming Guide - - - - - - - - - - - - - - - - - - - - - 33
4.1 Theory of Operation - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 33
Chapter 5 Register Reference - - - - - - - - - - - - - - - - - - - - - 35
5.1 Overview - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - 35
Revision 1.2 Chapter 1, Functional Description 1
Chapter 1 Functional Description
Entridia Corporations Serenade chip is a fully i nteg rated, wire-sp eed c l assical IP router chip
targeted at high performance routing systems. The architecture of the Serenade chip supports
Data Link (OSI Layer 2) and Physical (OSI Layer 1) interface technologies, such as Packet-
Over-Sonet and ATM at OC-3c rates, DSL, T1/T3, and broadband cable. A high-speed expan-
sion port on the chip supports cascading four Serenade devices to form large port-count sys-
tems. W ith aggr egate throughp ut of 1.2 Gbps and guaranteed in gress-to-e gress late ncy of eight
microseconds, the Serenade chip is ideally suited for converged access aggregation edge rout-
ing systems.
The primary function of the Serenade chip is to route IP packets using information contained
in the heade r of ea ch pac ket a nd knowl edge ( route s) of which groups of I P addres ses ( subnet s)
are reachable through its network interfaces.
This chapter describes the features and functions supported by the Serenade chip.
1.1 Data Link Layer (OSI Layer 2) Frame
Processing
Serenade has two OC-3c line-rate ports that provide the connection to the network. Through
these port interfaces Serenade transmits and receives PPP encapsulated IP packets with
optional 32-bit tags attached. An external Layer 2 device is required to add the appropriate
framing/deframing to PPP packets transmitted from Serenade. The maximum packet size for
the OC-3c ports is 1,536 bytes.
1.2 Network Layer (OSI Layer 3) Packet
Processing
Once a Layer 2 frame has been successfully received without errors by a Network Interface,
the Network Layer processing logic takes over. For IP packets, the Layer 3 processing logic
includes the following functions:
Verify the header checksum of an incoming packet.
Verify the validity of the IP Time to Live (TTL) f iel d, decrement the TTL field appropri-
ately, and recalculate the IP checksum field.
Look up dest inati on inte rface in the r oute t able (r efer to Route Table - Sect ion 1.4) using
the destination address and IP type of service field and queue the packet for transmis-
sion.
EdgeStre am Interface
2 ENT3002 Serenade Data Sh eet Revision 1.2
Prioritize the transmission of packets based on the contents of the IP Type of Service
(TOS) field.
If enabled, search the flow table (refe r to Flow Table Section 1.5) for rule s that apply to
the source and destination IP address, source and destination port (TCP, UDP or other
Layer 4 port) of each packet. Then, perform the specified action (forward, drop, or
invoke processing hooks in software running on the companion microprocessor).
Notify the companion microprocessor of IP Multicast, ICMP, and non-IP protocol pack-
ets. Permi t sof twa re runni ng on the microproc es sor to det er m ine handl in g for such pack-
ets and queue them for transmission from the appropriate output interface.
1.3 EdgeStream Interface
The EdgeS tream Inte rface is the m icroprocess or interface of th e Sere nade device. It co nsists of
a 32-bit data bus and a 16-bit address bus.
This inte rface present s Serenad e as a memory- like device to th e micropr ocesso r. An extensi ve
register map enables software running on the microprocessor to initialize and dynamically re-
configure the Serenade chip.
This interface also inclu des an interrup t line , which is used to notify the micropro cess or of
excepti on condit ions a nd peri odic events . Notif icat ion of speci fic condit ion s can be maske d or
enabled by programming appropriate interrupt control registers.
1.4 Route Table
The Serenade chip has a built-in 512 entry route table with full support for longest prefix
match look up, essential for Classless Interdomain Routing (CIDR). As route lookup can be
performed in three cycles using this approach Serenade can guarantee wire-speed processing
with deterministic latency. The Serenade chip route table also supports using the IP TOS field
in making Quality of Service (QOS) policy based destination route selection.
The on-chip 512 entry table can be extended at the system level by using a 64K external table
mem ory that communicates with the Sere nade c hip via the OptiStream Interface or a memory
manager like the Entridia ENT2003 Sonata controller.
Flow Table
Revision 1.2 Chapter 1, Functional Description 3
Configuration software running on a companion microprocessor can install static routes
(inclu ding a def aul t gat ewa y entry) and per i odically age and rebuild other entrie s d ynami cal l y
as a result of routing protocol messages (such as OSPF and BGP4). Also, the companion
microprocessor can maintain very large route tables in system memory and use the on-chip
512 entry table as a cache for frequently used routes.
Table 1.1 Ro ute Tab le Entr y
1.5 Flow Table
The Serenade chip has a 256 entry flow table that can be extended to 64K flows via off chip
memory connected to the OptiStream interface, or by a memory manager like the Entridia
ENT2003 Sonata controller. The flow table maintains filtering rules to determine which types
of IP packets a network interface can forward or discard, and which types require further
examination by software running on a companion microprocessor.
Each flow table entry can specify a source and destination IP subnet, source and destination
port (TCP, UDP or other Layer 4 port) along with an action to take if any or all of these fields
match those of the packet being filtered. The action can specify one of the following:
Forward: Accept the packet, and forward it normally.
Priority Forward: Accept the packet, and forward it with priority over normal packets.
Drop: Discard the packet and account for it in the management counters.
Examine: Accept the packet and notify a companion microprocessor of the match and
the starting address of the packet data in the Packet Buffer memory. This enabl es add i-
tional filtering software running on the microprocessor to further process the packet.
Table 1.2 Flow Table Entry
Lookup Terms Lookup Result
Valid
1-bit DestIP/Tag
32-bit TOS
8-bit NextHopIP
32-bit NextHop Interface Number
8-bit
VLSM
32-bit
Lookup Terms Lookup Results
Valid
1-bit Layer 3
Parameters Layer 4
Parameters Chip Interface Number
4-bit Action
2-bit Priority
6-bit
See Table 1.3 See Table 1.4
Queue Management
4 ENT3002 Serenade Data Sh eet Revision 1.2
Table 1.3 Layer 3 Para meters
Table 1.4 Layer 4 Para meters
1.6 Queue Management
The Serenade chip architecture implements a hybrid input and output queuing model, which
enables extensible end-to-end quality of service in a router system.
One OC-3c port has 8 levels of priority and the other port 2 levels of priority. Additional
scheduli ng can b e a dde d via th e En tr idia Opt iStrea m Interface. The pr io rity out put q ueue s can
be selected based on the type of service, source or destination IP addresses, and TCP, UDP or
other Layer 4 p orts, or a va lid combi nation of the af orementioned paramete rs, maki ng it highly
flexible for system designers to implement policy-based wire-speed routing rules for the OC-
3c ports. Serenade has a built in weighted round-robin scheduler with user programmable
weights.
1.7 OptiStream Expansion Bus
The expansion bus of the Serenade chip is an enhanced glueless Gunning Transceiver Logic
(GTL+) interface which can operate at frequencies up to 100 Mhz. This 64-bit wide data bus
supports a sustained throughput of 6.4 Gbps, allowing the system designers to cascade up to
four Serenade devices, thereby forming an 8-port OC3c PoS, wire-speed IP router and the
attachment of an external memory m anager like the Entr idia ENT2003 Sonata cont roller.
DestIP 32-bit SrcIP 32-bit Protocol 8-bit TOS 8-bit
VLSM
32-bit VLSM
32-bit
DestPort Number
(TCP/UDP) SrcPort Number
(TCP/UDP)
Lower Limit
16-bit Upper Limit
16-bit Lower Limit
16-bit Upper Limit
16-bit
Software Interface
Revision 1.2 Chapter 1, Functional Description 5
1.8 Software Interface
Entridi a provid es complete sof tware drive rs and a Service Appl ication Programming Inte rface
(API) to enabl e syst em designe rs to q uick ly migrate their exis ting so ftwa re to t he Ser enad e
platform.
The software d rivers are targeted to the Wind Rive r Syst ems VxWorks real-time operating
system. However, the drivers use a hardware abstraction model that makes it very easy to port
to other real-time operating s ystems.
To ena ble q uick i ntegr ation with existing softw are in routers, the Entridi a driv er presents the
Serenade network interface as two independent network devices. The router software that
deals wit h Network Layer communi cati ons can tr eat ea ch of the inter faces as a dedi cated Dat a
Link Layer device and interoperate with them without modification.
In addit ion t o t he dri vers, t he Ser vice API is a li brary of C func tions that simp lify access to the
more advance d feature s of the Serenad e chip such as filterin g rule s and traffic shaping param-
eters.
1.9 System Integration
The Serenade chip is designed to integrate into Access/Service Provider edge routing systems
without significant modification to the software currently running on those systems, while
accelerating their performance and increasing functionality.
The Serenade device can handle the following functions, which are currently handled by a
microprocessor or discrete programmable logic, without any software intervention:
Packet for warding: The Serenade chip transf ers all pac ket data int o and out of the Pack et
Buffer memory.
Route look up: The Serena de chip searc hes for t he longes t prefi x match for a destin ation
IP addres s in th e route table.
Packet filtering rules: The Serenade chip forwards or discards packets based on rules in
the flow table.
Packet Buffer management: The Serenade chip allocates and reclaims space for the
packets in the Pack et Bu ffe r memory.
IP Multica st: The Serenade chip manages trans mission of mult icast packets to all the
recipient ports without unnecessary data copying.
IP based QoS: Priority Scheduling; weighted round-robin.
System Integration
6 ENT3002 Serenade Data Sh eet Revision 1.2
Revision 1.2 Chapter 2, Hardware Description 7
Chapter 2 Hardware Description
This chapter provides a hardware description of the Serenade device. This chapter also
describes the Serenade chip external signals. It provides detailed pin-out diagrams and
describes each of the signals.
The Serenade chip integrates two PoS OC-3c ports. Multiple Serenade devices can be cas-
caded together, via OptiSt ream, to build a policy based, wire-speed router supporting both
LAN and WAN protocols. The device is packaged in a 520-pin HPBGA.
The device has four major external interfaces:
OptiStream interface
Edge Stream interface
SRAM interface
Netw ork OC 3-c interfa ce
OptiStream Interface. This proprietary bus operates at frequencies of up to 100 MHz. It is a
low voltage swing interface that uses terminated Gunning Transceiver Logic (GTL+) and
implements a Time Division Multiplexed (TDM) bus access mechanism. It communicates
control information and data between multiple Serenade devices.
In a system wher e s everal S ere nad e devices ar e c as caded, a single mast er dev ice is designate d
to control access to the OptiStream Interface. The total throughput on the bus is divided into
single cycle transactions, defined by the systems clock frequency. The master device grants
access to the bus sequentially to initiators for a single cycle. If an initiator does not have a
transaction pending, bus access is granted to the next pending initiator in sequence. This cir-
cuit-switched concept provides guaranteed bandwidth and latency to all packets in the system.
The in teg ra ted GTL+ transcei ver s used on Ser enad e en abl e di rect inter connection of cas cad ed
Serenade devices , via the Opti St ream Int erface, with out any ext ernal sup port device s. The bus
however, must be terminated externally.
The OptiStream interface also allows for expansion of the Route and Flow tables to support
64K flows at full line-ra tes.
EdgeStream Interface. This generic memory interface i s desi gned so tha t syst em OEMs can
manage the Serenade chip as a memory mapped device. This interface supports the use of an
external CPU. In this way, system integrators can preserve their existing investment in Net-
work Operat ing Syst em so ftware . In addi tion, s yst em inte grator s can f urt her di f fere ntiat e thei r
products by providing additional policy-based routing, bridging, VLAN, QOS, and/or VPN
functionality layered with Serenade devices.
SRAM Interface. The synchr onous SRAM memory inter fa ce opera tes at 100MHz, and is used
8 ENT3002 Serenade Data Sheet Revision 1.2
to interface with industry standard SRAMs available from various suppliers. The external
SRAM is used as a Packet Buffer for storing network packet data in queued data structures.
Packets ar ri vi ng via the Network Interface , the OptiStream Interface, or the Edge Stream Int er -
face are always buffered in the Packet Buffer to permit additional processing before transmis-
sion.
Network Interface. This interface uses two OC-3c ports using a Packet-Over-Sonet POS-PHY
Level 2 interface.
Figure 2.1 Serenade System Block Diagram
Serenade Chip
OC-3 Framer
External
Packet
Status
Buffer
SRAM
External
Packe t Da ta
Buffer SRAM
Packet
SRAM
Bus
Status
SRAM
Bus
External OptiStream Interface
(more Serenade devices,
EdgeStream
cascaded)
CPU Bus
Bridge
Logic
External
CPU
CPU
Memory
OC-3 Framer
POS
PHY
External 64K
Route Table External 64K
Flow Tabl e
OptiStream
OC3 Port OC3 Port
POS
PHY
Pin-Out
Revision 1.2 Chapter 2, Hardware Description 9
2.1 Pin-Out
Figure 2.2 is a ball placement diagram of the Serenade chip. The labels on the x and y axes
provide reference between this diagram and the signal descriptions. The maximum height
(above board) of the package including solder balls is 1.4mm.
Figure 2.2 Ball Placement Diagram (looking from top, through package)
A
B
C
D
E
F
G
H
J
K
L
M
N
P
R
T
U
V
W
Y
1 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20221 22 23 24 25 26
AA
AB
AC
AD
AE
AF
27 28 29 30 31
AL
AG
AJ
AK
AH
40.0mm
1.27mm
40.0mm
1.27mm
Pin-Out
10 ENT3002 Serenade Data Sheet Revision 1.2
2.1.1 Package Details
Figure 2.3 Package Dimensions
D,D Body
Size (mm)
A Max
Mounted
Pkg Height
(mm) C,C Ball
Matrix (mm) E Ball
Pitch (mm) Pin Count / Packaging
40x40 1.4 31x31 1.27 520 HPBGA
Table 2.1 Package Dimension Details
Pin-Out
Revision 1.2 Chapter 2, Hardw are Description 11
Pin-Out
12 ENT3002 Serenade Data Sheet Revision 1.2
12345678910111213141516
Apkt_ce3# pktce2# pktce1# pktce0# VSS stat_ce3# statclkout statclkin statdata12 statdata8 statdata4 statdata0 stataddr17 stataddr13 stataddr9 stataddr5
Bpktclkout pkt_we# pkt_rwoe# pktclkin VSS VSS stat_ce0# statrwoe# statdata13 statdata9 statdata5 statdata1 stataddr18 stataddr14 stataddr10 stataddr6
Cpktaddr3 pktaddr2 pktaddr1 pktaddr0 VSS VSS stat_ce1# test_mode statdata14 statdata10 statdata6 statdata2 stataddr19 stataddr15 stataddr11 stataddr7
Dpktaddr7 pktaddr6 pktaddr5 pktaddr4 VSS VSS stat_ce2# stat_bw# statdata15 statdata11 statdata7 statdata3 stataddr20 stataddr16 stataddr12 stataddr8
Epktaddr11 pktaddr10 pktaddr9 pktaddr8 VSS VSS VSS VSS VDDram VDDram VDDram VDDram VSS VSS VSS VSS
Fpktaddr15 pktaddr14 pktaddr13 pktaddr12 VDDram
Gpktaddr19 pktaddr18 pktaddr17 pktaddr16 VDDram
Hpktdata2 pktdata1 pktdata0 pktaddr20 VDDram
Jpktdata6 pktdata5 pktdata4 pktdata3 VDDram
Kpktdata10 pktdata9 pktdata8 pktdata7 VSS
Lpktdata14 pktdata13 pktdata12 pktdata11 VSS
Mpktdata18 pktdata17 pktdata16 pktdata15 VSS
Npktdata22 pktdata21 pktdata20 pktdata19 VSS
Ppktdata26 pktdata25 pktdata24 pktdata23 VSS
Rpktdata27 pktdata28 pktdata29 pktdata30 VSS
Tpktdata31 pktdata32 pktdata33 pktdata34 VSS
Upktdata35 pktdata36 pktdata37 pktdata38 VSS
Vpktdata39 pktdata40 pktdata41 pktdata42 VSS
Wpktdata43 pktdata44 pktdata45 pktdata46 VDDram
Ypktdata47 pktdata48 pktdata49 pktdata50 VDDram
AA pktdata51 pktdata52 pktdata53 pktdata54 VDDram
AB pktdata55 pktdata56 pktdata57 pktdata58 VDD
AC pktdata59 pktdata60 pktdata61 pktdata62 VDD
AD osc25 pktdata63 VSSAF1 VDDAF1 VDDAF2
AE gtl_vref bgr_res VSSAF2 VSSAB VDDAB
AF VSS VSS VSS VSS VSS
AG VSS VSS VSS VSS VSS VSS VSS VSS VDD VDD OS_3V OS_3V VSS VSS VSS VSS
AH OS_dat0 OS_dat4 OS_dat8 OS_dat12 OS_dat16 OS_dat20 OS_dat24 OS_dat28 OS_dat32 OS_dat36 OS_dat40 OS_dat44 OS_dat48 OS_dat51 OS_dat55 OS_dat62
AJ OS_dat1 OS_dat5 OS_dat9 OS_dat13 OS_dat17 OS_dat21 OS_dat25 OS_dat29 OS_dat33 OS_dat37 OS_dat41 OS_dat45 OS_dat49 OS_dat52 OS_dat56 OS_dat61
AK OS_dat2 OS_dat6 OS_dat10 OS_dat14 OS_dat18 OS_dat22 OS_dat26 OS_dat30 OS_dat34 OS_dat38 OS_dat42 OS_dat46 OS_dat50 OS_dat53 OS_dat57 OS_dat60
AL OS_dat3 OS_dat7 OS_dat11 OS_dat15 OS_dat19 OS_dat23 OS_dat27 OS_dat31 OS_dat35 OS_dat39 OS_dat43 OS_dat47 VSSAB OS_dat54 OS_dat58 OS_dat59
Figure 2.4 Ball Assignment Chart (Left Half looking from top, through package)
Pin-Out
Revision 1.2 Chapter 2, Hardw are Description 13
17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
stataddr4 stataddr0 ES__data31 ES__data27 ES__data23 ES__data19 ES__data15 ES__data11 ES__data7 ES__data3 ES__par3 ES__addr6 ES__addr2 ES__clk ES__ce# A
stataddr3 ES__perr# ES__data30 ES__data26 ES__data22 ES__data18 ES__data14 ES__data10 ES__data6 ES__data2 ES__par2 ES__addr5 ES__addr1 ES__wait tx0_pa B
stataddr2 ES__int# ES__data29 ES__data25 ES__data21 ES__data17 ES__data13 ES__data9 ES__data5 ES__data1 ES__par1 ES__addr4 ES__addr0 ES__we# tx0_par C
stataddr1 sysreset# ES__data28 ES__data24 ES__data20 ES__data16 ES__data12 ES__data8 ES__data4 ES__data0 ES__par0 ES__addr3 ES__oe# N/C rx0_err D
VSS VDD VDD VDD VDDIO VDDIO VSS VSS VSS VSS VSS VSS VSS VSS VSS E
VDD VDD VSS VSS VSS F
VDD scan_en test1 test0 rx0_par G
VDD VDD rx0_pa rx0_val rx0_eop H
VDD VDD rx0_enb# txd_data0 txd_data1 J
VDDIO VDDIO rx0_sop rx0_data0 rx0_data1 K
VDDIO VDDIO tx0_data2 tx0_data3 tx0_data4 L
VSS VSS rx0_data2 rx0_data3 rx0_data4 M
VSS VSS tx0_data5 tx0_data6 tx0_data7 N
VSS VSS rx0_data5 rx0_data6 rx0_data7 P
VSS VSS tx0_enb# tx0_sop tx0_eop R
VSS VSS VSS VSS nirefclk T
VDD VDD rx1_eop rx1_val rx1_pa U
VDD VDD tx1_data1 tx1_data0 rx1_enb V
VDD VDD rx1_data1 rx1_data0 rx1_sop W
VDD VDD tx1_data4 tx1_data3 tx1_data2 Y
VDD VDD rx1_data4 rx1_data3 rx1_data2 AA
VDD VDD tx1_data7 tx1_data6 tx1_data5 AB
VSS VSS rx1_data7 rx1_data6 rx1_data5 AC
VSS VSS tx1_eop tx1_sop tx1_enb AD
OS_3V OS_3V rx1_err tx1_prty rx1_prty AE
VSS VSS VSS VSS VSS AF
VSS VSS VSS VDD VDD VDD VSS VSS VSS VSS VSS VSS VSS VSS VSS AG
OS_pktid2 OS_pktid6 OS_pktid10 OS_pktid14 OS_pktid18 OS_pktid22 OS_pktid26 OS_trans3 OS_bsy0# OS_bsy4# OS_rq0# OS_rq4# OS_gnt0# OS_gnt4# OS_parity AH
OS_pktid1 OS_pktid5 OS_pktid9 OS_pktid13 OS_pktid17 OS_pktid21 OS_pktid25 OS_trans2 VSSAB OS_bsy3# OS_bsy7# OS_rq3# OS_rq7# OS_gnt3# OS_gnt7# AJ
OS_pktid0 OS_pktid4 OS_pktid8 OS_pktid12 OS_pktid16 OS_pktid20 OS_pktid24 OS_trans1 OS_trans5 OS_bsy2# OS_bsy6# OS_rq2# OS_rq6# OS_gnt2# OS_gnt6# AK
OS_dat63 OS_pktid3 OS_pktid7 OS_pktid11 OS_pktid15 OS_pktid19 OS_pktid23 OS_trans0 OS_trans4 OS_bsy1# OS_bsy5# OS_rq1# OS_rq5# OS_gnt1# OS_gnt5# AL
Figure 2.5 Ball Assignment Chart (Right Half looking from top, through package)
Pin Signal Descriptions
14 ENT3002 Serenade Data Sheet Revision 1.2
2.2 Pin Signal Descriptions
The Serenade chip pins are grouped as follows:
OptiStream Interface pins
EdgeStream Interface pins
SRAM Interface pins
Network Interface pins
Analog, power, ground, and utility pins
The following sections provide signal descriptions for each of these groups of pins.
2.2.1 OptiStream Interf ace Pin Signal Descriptions
This secti on provi des a detai led descr ipti on of the sign als for ea ch of the OptiS trea m Inte rface
pins. For the location of these pins, see cells AH1 through AL31 on 12 and 13.
Signal Name Type Description
OS_Data[63:0] GTL+
[In/Out]
An initiator uses these signals to communicate control and data to targets.
These signals are driven by initiators and are inputs for targets. They are sampled
on every rising edge of the phase and frequency locked, 100MHz internal clock.
When the bus is idle, these signals are pulled up to the terminating voltage (VT).
OS_Trans[5:0] GTL+
[In/Out] An initiator uses these signals to define the transaction occurring on the OptiSt -
ream Interface to the targets. These transactions are described in the OptiS tream
Interface Transactions section.
These signals are driven by the initiators and are inputs to targets. They are sam-
pled on every rising edge of the phase and frequency locked, 100MHz internal
clock. When the bus is idle, these signals are pulled up to the terminating voltage
(VT).
OS_Pktid[26:0] GTL+
[In/Out] Each device in the system uses these signals to define a unique identity for each
packet. To create this unique identity, the device encodes its identity on
OS_Pktid[26:23] and encodes a packet sequence number on OS_Pktid[22:0]. The
packet sequence number is the start address of the Packet Buffer assigned to
store the packet in the local SRAM.
These signals are driven by the initiator and are inputs to targets. They are sam-
pled on every rising edge of the phase and frequency locked, 100MHz internal
clock. When the bus is idle, these signals are pulled up to the terminating voltage
(VT).
Table 2.2 OptiStream Interface Pins
Pin Signal Descriptions
Revision 1.2 Chapter 2, Hardw are Description 15
OS_Rq[7:0]# GTL+
[In/Out] An initiator uses these active low signals to request use of the OptiStream Interface
from the master.
Each device uses the value in its device ID register to determine which OS_Rq#
signal to drive as an output. For instance, if a device ID register is programmed to
01H, the device drives OS_Rq[1]# as an output and floats the other OS_Rq# sig-
nals, which are pulled up to the terminating voltage (VT).
These signals are driven active on the rising edge of the clock as soon as an initia-
tor has a transaction pending. The OS_Rq# signal stays active as long as there are
transactions pending that require use of the OptiS t ream Interface. When there are
no additional transactions pending that require use of the OptiStream Interface, the
OS_Rq# signal is driven inactive when the final transfer is initiated after receiving
the OS_Gnt# signal.
These signals are driven by the initiators and are inputs to the master. They are
sampled on every rising edge of the phase and frequency locked, 100MHz internal
clock.
OS_Gnt[ 7 :0]# GTL +
[In/Out] The master uses these active low signals to respond to OS_Rq# signals from initia-
tors.
Each device uses the value in its device ID register to determine which OS_Gnt#
signal to monitor as an input. For instance, if the device ID register is programmed
to 02H, the device monitors OS_Gnt[2]# as an input and ignores the other
OS_Gnt# signals.
These signals are driven active of the rising edge of the clock by the master. This
signal communicates to an initiator that the next cycle has been reserved for its
use. An initiator checks that its OS_Rq# and OS_Gnt# signals are active before
starting a transaction on the OptiStream Interface. The OS_Gnt# signal for a partic-
ular initiator can continue to be active for consecutive cycles if no other OS_Rq#
signals are active and the corresponding OS_Rq# signal stays active. If other
OS_Rq# signals are active, a OS_Gnt# signal is driven inactive on the rising edge
of the internal clock one cycle after being driven active.
These signals are outputs from the master and are inputs to the initiators. They are
sampled on every rising edge of the phase and frequency locked, 100MHz internal
clock.
OS_Busy[7:0]# GTL+
[In/Out] Each target uses these signals to communicate to initiators that the target is unable
to accept data transfer transactions. If the target cannot successfully latch the
transaction, it drives OS_Busy# low on the rising edge of the clock. If the target can
accept the transaction, the OS_Busy# signal stays at the terminating voltage (VT).
The initiator samples the OS_Busy# in the cycle before the transaction will occur to
determine if it will successfully complete.
OS_Parity GTL+
[In/Out] The Initiator generates this signal, and the targets check it. This signal is used for
OptiS tream Interface error detection. It can be programmed to support odd or even
parity.
If an error occurs, a target registers the error and generates an interrupt to the sys-
tem via the OptiStream Interface Interrupt register.
Signal Name Type Description
Table 2.2 OptiStream Interface Pins (continued)
Pin Signal Descriptions
16 ENT3002 Serenade Data Sheet Revision 1.2
2.2.2 EdgeStream Interface Pin Signal Descriptions
This sect ion provide s a detail ed descript ion of the si gnals for each of the Edge Stre am interfa ce
pins. For the location of these pins, see cells A18 through D31 on 12 and 13.
Signal Name Type Description
ES_Addr[6:0] Std TTL
[Input] These signals provide 7-bit, non-multiplexed address input signals to address the
Serenade chip registers. Normally, ES_A ddr[6 :0] should be tied to address pins
[8:2] of a 32-bit processor bus.
There are two classes of register accesses, direct and indirect. The EdgeStream
Interface registers are addressed directly. The registers for the rest of the chip are
accessed using an indirect address register (Control Bus Address Register) and an
indirect data register (Control Bus Data Register).
ES_Data[31:0] 24 ma
[In/Out] These signals provide 32-bit, non-multiplexed data signals, used by the external
CPU to read/write configuration, status, control, and packet data from or to the Ser-
enade chip. These signals are placed in a high impedance state when ES_CE# or
ES_OE# signals are de-asserted.
ES_Clock Std TTL
[Input] The Serenade chip uses this clock signal to allow burst read/write operations from
the packet read and write FIFOs. This clock allows use of a single register access
and the subsequent burst read /write of the FIFOs on every leading edge of the
clock.
ES_CE#
ES_WE#
ES_OE#
St d TTL
[Input] The external CPU uses the Chip Enable, Write Enable, and Output Enable signals
to control the read/write of data from and to the Serenade chip.
A write to the Serenade chip is accomplished by asserting the chip enable
(ES_CE#) and the write enable (ES_WE#) signals while negating the output
enable signal (ES_OE#). The pattern on the ES_Data[31:0] pins is then written
into the location specified on the address pins (ES_Addr[6:0]) and latched at the
location by driving ES_WE# high.
A read from the Serenade chip is accomplished by asserting the chip enable
(ES_CE#) and output enable (ES_OE#) while negating the write enable
(ES_WE#) signals. The contents of the register specified by the address pins
appears on the data pins.
The ES_Data[31:0] pins are placed in a high impedance state when the ES_CE#
signal is negated.
If wait states are generated by the Serenade chip during the read/write access, the
CPU (or any accessing agent) must hold the ES_CE#, ES_WE#, and ES_OE# sig-
nal levels constant until the wait is negated.
These are active-low signals.
ES_WAIT 24 ma
[In/Out] The Serenade chip uses this signal to communicate to an external processor that
the register access being processed requires additional cycles to complete. The
ES_WAIT signal is active-high.
ES_PA R[3:0] 24 ma
[In/Out] This parity signal generates a parity bit computed over each byte of the
ES_Data[31:0] signals during valid data phases, if the en_prty bit in the ES_ Control
register is set (the default after a reset is parity off). The polarity depends on the
odd_prty bit in the ES_ Control register.
As an input, this line is sampled when the Serenade chip is placed in a write cycle
and driven when the Serenade chip is executing a read cycle.
These signals are placed in a high impedance state when the ES_CE# signal is de-
asserted or when the ES_OE# signal is high.
Table 2.3 EdgeStream Interface Pins
Pin Signal Descriptions
Revision 1.2 Chapter 2, Hardw are Description 17
2.2.3 SRAM Interface Pin Signal Descriptions
This sect ion provide s a detailed descri ption of the sign als for each of the SRAM Int erface pins.
For pin location, see cells A1 through AD4 and A6 through D18 on 12 and 13.
ES_INT# 24 ma
[Emulated
Open Drain]
The Serenade chip asserts this interrupt request signal to indicate that one or more
interrupt events have occurred and that the corresponding bits in the Interrupt Reg-
ister have been set. The Serenade chip continues to assert this signal until all inter-
rupt flags are cleared in the respective registers.
This signal drives low when an interrupting event occurs. The output emulates an
open-drain circuit. When transitioning from low to high (INT cleared), it drives high
for a short period before going to a high impedance. This line is not internally pulled
up so that it can be OR-tied with INT# pins from multiple Serenade devices. An
external pull-up resistor is required.
This is an active low signal.
ES_PER R# 24 ma
[Tri-stated
Out]
The Serenade chip asserts this Parity Error signal during a write cycle to indicate
that a parity error was detected on the ES_Data[31:0] signals. If an error is
detected, the write operation is either aborted or goes through anyway, depending
on the pewrt_inhibit bit in ES_ Control register (this bit defaults to write through
after reset).
This is an active low signal.
SysReset# Std TTL
[Input] This System Reset is asserted to perform a hardware reset of the Serenade chip.
The SysReset# input uses a Schmidt trigger.
This is an active low signal.
Signal Name Type Description
PktClockOut,
StatusClockOut 24 ma
[Output] These signals provide the SRAM clock. They are 100MHz output from the Sere-
nade chip to the SRAM (data and status respectively). They are used to register
the address, data, and chip enable inputs on the rising edge. All synchronous out-
puts meet setup and hold times around the clocks rising edge.
PktClockIn,
StatusClockIn 24 ma
[Input] These clock input signals are connected to the clock output signals. They are
available to allow the system designer to match the data signal traces to the clock
trace to minimize skew on memory read operations.
PacketAddr[20:0] 24 ma
[Output] These signals provide 21-bit, non-multiplexed address signals, which the Sere-
nade chip outputs to address the SRAM. Since the Serenade chip does not
access bytes on the data side of the SRAM interface, the byte access signals on
the SRAM must always be enabled for full access.
In conjunction with the eight independent chip-enable signals, these address sig-
nals allow the Serenade chip to logically address a maximum of 8 Mbytes of mem-
ory for packet data while using 256K x 32 SRAMs.
For information on the electrical restrictions that limit the amount of RAM that can
be connected, see Section 2.3.5, External SRAM, on page 23.
Note: The Synchronous Load (LD#) signal on the SRAM, used to clock in a new
address, is permanently enabled when used with the Serenade chip.
PacketData[63:0] 24 ma
[In/Out] T hes e signals provide 64-bit, non-multiplexed Packet Data signals used by the
Serenade chip to read and write packet data from and to the SRAM.
Table 2.4 SRAM Bus Signal Pins
Signal Name Type Description
Table 2.3 EdgeStream Interface Pins (continued)
Pin Signal Descriptions
18 ENT3002 Serenade Data Sheet Revision 1.2
2.2.4 Network Interface Pin Signal Descriptions
This section provides a detailed description of the signals for each of the Network Interface
pins. For pin location, see cells B29 through AE31 on 12 and 13.
PacketCE[ 3:0 ]# 24 ma
[Output] The Serenade chip uses these data buffer chip enable signals to enable the read-
ing and writing of data. To maximiz e memory cycle time and minimize decode
time, four individual PacketCE# signals are available.
The decode space enabled by each PacketCE# signal is controlled by setting
appropriate bits in the RAM Parameter Register.
PacketR/W#
/PacketOE# 24 ma
[Output] This data RAM read/write/output enable signal has a dual function depending on
the RAM type specified in the Packet Buffer RAM Control Register.
In RAM mode, this line serves as a Read/Write# signal. It signifies a read cycle
when it is high and a write cycle when it drives low.
PacketWEe# 24 ma
[Output] This packet RAM write enable signal sends a write command (when low) to the
packet data SRAM. This signal is asserted in the same cycle as PacketR/W#/
PacketOE# and PacketA ddr[20:0] signals, described above.
One pin is used for all 64-bit wide RAM devices.
StatusAddr[20:0] 24 ma
[Output] These signals provide 21 non-multiplexed address signals, which the Serenade
chip outputs to addresses the Status Buffer. These signals are placed in a high
impedance state when the corresponding StatusCE[n]# signals are de-asserted.
Note: The Synchronous Load (LD#) signal on the SRAM, used to clock in a new
address, is permanently asserted when used with the Serenade chip.
StatusData [15:0] 24 ma
[In/Out] T hes e signals provide non-multiplexed bidirectional packet status signals, which
the Serenade chip uses to read and write packet status to the Packet Buffer.
StatusCE[3:0]# 24 ma
[Output] The Serenade chip uses these status buffer chip enable signals to enable reading
or writing of status. To minimize decode time, four StatusCE# signals are avail-
able.
The decode space enabled by each StatusCE# signal is controlled by setting
appropriate bits in the Packet Buffer RAM Control Register.
StatusBW# 24 ma
[Output] The S tatusBW# signal is asserted on the same cycle as the StatusAddr[20:0].
StatusR/W#
/StatusOE# 24 ma
[Output] This status buffer read/write/output enable signal has a dual function depending
on the RAM type specified in the Packet Buffer RAM Control Register.
In RAM mode, this line serves as a Read/Write# signal. It signifies a read cycle
when it is high and a write cycle when it drives low.
Signal Name Type Description
POS-Phy Interface Pins
NIRefClock 8 ma
[Output] This Network Interface reference clock is a continuous clock that provides
the timing reference for the RXn_Data[7:0] and TX[1:0] Data[7:0] signals.
This clocks frequency is 50 MHz.
TX0_Enb# for np0
TX1_Enb# for np1
8 ma
[Output] OC-3c port Transmit Data Write Enable:
This active-low signal indicates the data transmit activity on TXn_Data[7:0]
bus.
Table 2.5 Network Interface Pins
Signal Name Type Description
Table 2.4 SRAM Bus Signal Pins (continued)
Pin Signal Descriptions
Revision 1.2 Chapter 2, Hardw are Description 19
TX0_PA for np0
TX1_PA for np1
St d TTL
[Input] OC-3c port Direct Transmit Packet Available:
This active-high signal indicates that a programmed number of data bytes
are available to transfer in the Transmit FIFO when it transitions to high.
Once high, it indicates that the Transmit FIFO is not (almost) full.
TX0_SOP for np0
TX1_SOP for np1
8 ma
[Output] OC-3c port Transmit Start of Packet:
This active-high signal indicates the first byte of a packet on TXn_Data[7:0]
bus when asserted with TXn_Enb# at the rising edge of NIRefClock.
TX0_EOP for np0
TX1_EOP for np1
8 ma
[Output] OC-3c port Transmit End of Packet:
This active-high signal marks the end of a packet on TXn_Data[7:0] bus.
The last byte of the packet is on the bus when this pin and TXn_Enb# are
both asserted at the rising edge of NIRefClock.
TX0_Data [7:0] for np0
TX1_Data[7:0] for np1
8 ma
[Output] OC-3c port Tr ansmit Packet Data Bus (8-bit):
This bus carries the packet bytes that are output to the external optical PHY.
The output is valid when TXn_Enb# is also asserted at the rising edge of
NIRefClock.
TX0_Par for np0
TX1_Par for np1
8 ma
[Output] OC-3c port Transmit Packet Data Parity:
This bit carries the parity (odd or even) for the 8-bit TXn_Data bus during the
data packet transmission, if parity generation is enabled.
RX0_Enb# for np0
RX1_Enb# for np1
8 ma
[Output] OC-3c port Read Data Enable:
This active-low signal initiates the input data transfer on RXn_Data [7:0]
bus.
RX0_Val for np0
RX1_Val for np1
St d TTL
[Input] OC-3c port Receive Data Valid:
This active-high signal indicates the validity of the receive data signals.
When RXn_Val is high and RXn_Enb# is low, the receive signals
(RXn_Data, RXn_SOP, RXn_EOP and RXn_Err) are valid.
RX0_PA for np0
RX1_PA for np1
St d TTL
[Input] OC-3c port Receive Packet Data Available:
This active-high signal indicates that either some programmable amount of
data or an end of packet is available in the Receive FIFO in the PHY.
RX0_SOP for np0
RX1_SOP for np1
St d TTL
[Input] OC-3c port Receive Start of Packet:
This active-high signal marks the first byte of a packet transfer on
RXn_Data[7:0] bus, when asserted simultaneously with RXn_Enb# at the
rising edge of NIRefClock.
RX0_EOP for np0
RX1_EOP for np1
St d TTL
[Input] OC-3c port Receive End of Packet:
This active-high signal marks the end of packet on the RDAT[7:0] bus, when
asserted simultaneously with RXn_Enb# at the rising edge of NIRefClock.
RX0_Data [7:0] for np0
RX1_Data[7:0] for np1
St d TTL
[Input] OC-3c port Receive Packet Data Bus (8-bit):
This bus carries the packet bytes that are read from the FIFO in the external
PHY device. The input is valid when RXn_Enb# is also asserted at the rising
edge of NIRefClock.
RX0_Par for np0
RX1_Par for np1
St d TTL
[Input] OC-3c port Receive Packet Data Parity:
This bit reads the parity (odd or even) for the 8-bit RXn_Data during the data
packet reception, if parity check is enabled.
RX0_Err for np0
RX1_Err for np1
St d TTL
[Input] OC-3c port Receive Error Indicator:
The PHY device uses this active-high signal to indicate that the current
packet is aborted and should be discarded. It can be asserted only during
the last byte transfer of the packet.
Signal Name Type Description (continued)
Table 2.5 Network Interface Pins (continued)
External Components
20 ENT3002 Serenade Data Sheet Revision 1.2
2.2.5 Analog, Powe r, Ground, and Utilit y Pin Signal Descriptions
This section provides a detailed description of the signals for each of these pins.
2.3 External Components
This section describes the characteristics and restrictions that apply to the Serenade chips
external components. It covers the following topics:
External OptiStream In terface routing
Signal Name Type Description
Osc25 Analog
[Input] This external reference clock provides the reference timing for the internal clock synthe-
sizer circuit of Serenade chip. This 25MHz clock is used to derive the internal clocks.
The internally-generated100MHz clock is kept in phase lock with the 25MHz clock input
to ensure system synchronization in a application with multiple Serenade devices. This
is achieved by the following means:
sharing of the same Osc25 clock source between all Serenade devices involved.
careful board layout to achieve matching traces.
BGR_Res Analog The band gap reference pin requires a 12.4K 1% resistor to be attached between it and
the (analog) ground to insure proper operation of internal current sources.
GTL_Vref Analog This GTL+ Receiver Reference Voltage pin must have nominal (2/3 * Vterm) voltage
applied to it for the OptiStream Interface receivers t o operate properly.
Testmode Std TTL
[Input] This scan test mode pin must be grounded for the normal mode of operation.
Scan_En Std TTL
[Input] This scan shift enable pin must be grounded for the normal mode of operation.
Te st[1 :0 ] Std TTL
[Input] These test mode selection pins must both be grounded for the normal mode of opera-
tion.
VDDAF1,
VDDAF2 Pwr This frequency synthesizer power pad must have 3.3V power applied to it to drive inter-
nal analog circuitry. A clean, filtered 3.3V should be used.
VDDAB Pwr This band gap reference power pad must have 3.3V power applied to it to drive internal
analog circuitry. A clean, filtered 3.3V should be used.
VSSAF1,
VSSAF2 Gnd This frequency synthesizer ground pad is dedicated to the internal analog circuitry.
VSSAB Gnd This band gap reference ground pad is dedicated to the internal analog circuitry.
VDDOS_3V Pwr These OptiStream Interface 3.3V power pads are exclusively for OptiStream Interf ace
transceivers. They should be combined together and shorted to the board 3.3V VDD at
one place only.
VDDram Pwr These RAM 3.3V power pads are exclusively for SRAM Interface transceivers. They
should be combined together and shorted to the board 3.3V VDD at one place only.
VDDIO Pwr These I/O VDD power pads exclusively for the EdgeStream Interface transceivers. They
should be combined together and shorted to the board 3.3V VDD at one place only.
VDD Pwr These VDD 2.5V power pins drive the core logic.
VSS Gnd These ground pins are for the digital logic circuits.
Table 2.6 Analog, Power, Ground, and Utility Signal Pins
External Components
Revision 1.2 Chapter 2, Hardw are Description 21
External refer ence c lock
Band gap reference
GTL+ rec eiver reference
External SRAM
2.3.1 External OptiStream Bus Routing
Up to four Serenade devices can be tied to the OptiStream bus to form a system of up to eight
OC-3c ports. Due to the high speed nature of this bus, some physical restrictions apply.
The wire length for the OptiStream Interface cannot exceed 8 inches.
Each OptiStream Interface signal line must be terminated to 1.5V, with a maximum trace
length of 8 in ches.
The circuit board wires for OptiStream Interface should run directly underneath the
chips without adding any significant stub length.
The maximum pitch between the devices is two inches. The distance from the edge
device t o the te rminat or must be one i nch or le ss. A st raight lin e thro ughou t the trace run
is desirable.
Figure 2.6 illustrates these requirements:
Figure 2.6 OptiStream Interface Wiring
2.3.2 External Reference Clock
The Serenade chip uses one external 25MHz reference clock (Osc25) to generate all other
clocks necessary to run the chip, except the ES_Clock (which is supplied from the CPU sys-
tem, see Table 2.3 on page 16).
Serenade
chip Serenade
chip Serenade
chip Serenade
chip
1 Max 2 Max 2 Max
8 Max
V
T
= 1.5V V
T
= 1.5V
(up to total of four Serenade chips)
(wires go through directly
underneath the chips;
no stubs)
External Components
22 ENT3002 Serenade Data Sheet Revision 1.2
The external reference clock may be derived from a discrete oscillator, or from a precision
PLL frequency generator. However, it should be free of ripple noise and jitter. The Serenade
chips Osc25 pin requires 2.5V CMOS swing instead of 3.3V. If a 5 or 3.3V source is used,
perform the level translation before connecting the source to the Serenade chips Osc25 pin.
The Osc25 pin is i n location AD1 in Fig ure 2.4 on page 12. For a descript ion of the Osc25 pin
see Table 2.6 on page 20.
2.3.3 Band Gap Reference
A 1% precisi on r esistor pro vides a ref er enc e po int for t he i nte rna l band-gap reference. Pl ace a
12.4K 1% resistor between the BGR_Res pin and the analog ground as illustrated in Figure
2.7.
Figure 2.7 Band Gap Reference Resistor
The BGR_Res pin is in location AE2 in Figure 2.4 on page 12. For a description of the
BGR_Res pin, see Ta ble 2.6 on page 20.
2.3.4 GTL+ Receiver Reference
To maximi ze the timin g margin on the OptiStream Interface, the receiver is biased at a lev el
where transitions are most quickly detected. To set the appropriat e bias level, th e GTL _Vref
pin should be driven to 2/3 of VT, the OptiStream Interface termination voltage. This is nomi-
BGR_Res
Serenade chip
12.4K 1%
VSSAB
External Components
Revision 1.2 Chapter 2, Hardw are Description 23
nally 1.0V. Figure 2.8 illustrates this requirement.
Figure 2.8 GTL+ Receiver Reference Voltage
The GTL_Vref pin is in location AE1 in Figure 2.4 on page 12. For a description of the
GTL_Vref pin, see Table 2.6 on page 20.
2.3.5 External SRAM
The Serenade chip is designed to work with pipelined (recommended) ZBT SRAM for the
packet da ta buf fe r an d the st atus buffer. It can l ogi ca ll y addr es s up to four ban ks of dat a RAM
devices, up to a 21-bit address range per device, with a maximum addressable RAM capacity
of 64MB. In reality, the capacitive loading and RAM availability limit the configuration. Spe-
cifi cally, the Serenade chip can di rectly dri ve up t o eigh t packet data RAM chips i n various
configurations, or more if the RAM clock can be buffered-phase locked externally.
Table 2.7 illus t rates some of th e pos sible combinations f or the memory configu rations. In this
table:
A bank is a set of RAM chips controlled by the same Pkt_CEx# pin. Since the width of
the Packet Data bus is fixed at 64-bit (8-byte), one bank of 256K x 32 bit x 2 chips
SRAM represents 256K x 8 = 2MBytes of storage space. Four of these rows comprise
the capacity of eight MBytes.
The Minimum Chip Count column indicates how many SRAM chips are required to
implement a si ngle bank RAM system (including the status SRAM).
The Maximum Chip Count column indicates the RAM chip count with maximum
allowed number of banks.
GTL_Vref
Serenade Chip 3.32K 1%
VSS
VT = 1.5V
6.65K 1%
External Components
24 ENT3002 Serenade Data Sheet Revision 1.2
As indicated in Table 2.7, an independent set of data RAM (64-bit wide) and status RAM (16-
bit) is required. For the data port, depending on whether the RAM is 16-bit wide or 32-bit
wide, four or two chips per row are needed. Each Serenade device can logically support up to
eight rows, but the driver electrical properties limit the number of RAMs directly attached to
the bus to eight as shown in the table s Data SRAM Chip Count column.
Note: In genera l, for a gi ven siz e RAM, use a 32 -bit de vice to limit t he RAM chip count ; use
a 16-bit device to maximize the buffer capacity.
The SRAM inte rface transceiver pads ar e designed with following loading assu mptions:
The Serenade chip can connect more SRAM devices than described above. For example, six-
teen 128K x 32 RAMs instead of eight 256K x 32 can be used to get an eight MByte capacity
by using PacketAddr[20] as one of the chip enable signals (see Table 2.4 on page 17 for a
description of PacketAddr[20]). Since such applications require heavier electrical loading,
careful electrical and timing analysis must be followed to insure its proper operation.
Data SRAM Device Buffer Size Dat a SRAM
Chip Count Status SRAM
Chip Count Max.a number
of banks
a. Max imum number of banks based on the four-chip loading limit specified below. The device allows up to
eight rows of RAMs to be attached, but careful electrical considerations must be made to go beyond this
limit.
Min. Chip
Count Max. Chip
Count
128K x 16
e.g., Micron MT55L128L18 1MB x row
(2MB Max.) 4 x row 1 x row 2 4 + 1 8 + 2
64K x 32
e.g., Micron MT55L64L32 512KB x row
(2MB Max.) 2 x row 1 x row
(use 64K x 16) 4 2 + 1 8 + 4
256K x 16
e.g., Micron MT55L256L18 2MB x row
(4MB Max.) 4 x row 1 x row 2 4 + 1 8 + 2
128K x 32
e.g., Micron MT55L128L32 1MB x row
(4MB Max.) 2 x row 1 x row
(use 128K x 16) 4 2 + 1 8 + 4
256K x 32 2M x bank
(8MB Max.) 2 x bank 1 x bank
(use 256K x 16) 4 2 + 1 8 + 2
Table 2.7 Recommended RAM Configurations
RAM Control Pin Input Capacitance: 4pF Max
SRAM Data Pin Input Capacitance: 4pF Max
SRAM Address Pin Input Capacitance: 3.5pF Max
Maximum Number of Data SRAMs: 8
PC Board Trace Capacitance Allowance: 3pF / bank Max
Maximum Loading on Address Pins: 34pF Max
Maximum Loading on Data Pins: 23pF Max
External Components
Revision 1.2 Chapter 2, Hardw are Description 25
The type and size of the RAM chips being using in the Packet Buffer RAM Control register
must be programmed. Figure 2.9 illustrates how to program this type of configuration:
Figure 2.9 R AM Conf igu rat io n Exampl e
Serenade is designed to work with SRAMs of various types rated at 100MHz or faster. To
accommoda te cost sen sit i ve applications, the fo ll owing various prog ra mmab le paramete rs are
provided to interface to slower RAMs. Contact Entridia for the most current SRAM support
information:
Device Types
Pipelined ZBT SRAM is currently recomm ended for o ptimal per formance for Serenade:
00 Reserved
01 Reserved
10 Reserved
11 Pipeline mode ZBT SRAM
Device Address Space
The number of the co lumn addr ess the specifie d RAM chip has. This d efines the size of each
RAM bank. Each column is 8-bytes wide (packet) and 2-bytes wide (status).
000 15-bit -- 32K columns
001 16-bit -- 64K columns
010 17-bit -- 128K columns
011 18-bit -- 256K columns
100 19-bit -- 512K columns
101 20-bit -- 1M columns
110 21-bit -- 2M columns
256K x 32
256K x 32
256K x 32
256K x 32
256K x 32
256K x 32
256K x 32
256K x 32
PktCE3#
PktCE2#
PktCE1#
PktCE0#
Addr[16:0]
Data[63:0]
Type: Pipelined ZBT SRAM
Device Size: 256Kx32
# of Banks: 4
Wait States: 0
= 0x004F in RAM
Data Capacity = 8MB
PktClkOut
PktClkIn
Control Reg.
External Components
26 ENT3002 Serenade Data Sheet Revision 1.2
Number of Memory Banks
The number of banks of packet/status RAMs. Each bank is selected by one of the four chip
enable (p acke tCE[3:0]#, statusCE[3:0]#) p ins.
0000 Reserved
0001 1 bank
0010 2 banks
0011 3banks
0100 4 banks
0101 - 1111 Reserved
Chip Enable Setup Time (CSS)
When PacketCE[3:0]#/StatusCE[3:0]# and other control pins are negated for a while, some
RAM chips enter a sleep mode. They may take some extra time before they wake up, though
such timings are often not specified. Serenade accommodates such RAM chips by inserting a
clock cycle delay between PacketCE[3:0]#/StatusCE[3:0]# assertion to setting of all other
address/data/control signals.
0 No extra chip enable setup time is set. PacketCE[3:0]#/StatusCE[3:0]#, address, control
and/or data are coincident.
1 One c lock delay is inse rted b etween PacketCE[3:0]#/ StatusCE [3:0]# assertion fro m the
rest of the control signals.
The RAM inte rf ac e as sumes that a g ive n app li ca ti on us es the same size RAM for data and sta-
tus stor age . For exa mple, if 128K (x16 or x32bit) devices ar e use d for the packet data SRAM,
use a 128K x 16 device for the status SRAM.
Actual wiring of RAM interface pins to the RAM devices depends on the type and configura-
tion of the RAM de vice. Table 2.8 shows wir ing ex amples f or ca ses us ing RAM usi ng the pop-
ular pin-out from those devices by Micron Technology, IDT, and Motorola.
Serenade Chip Pin-Out RAM
Both Status and Data RAMs
LD = VSS
LBO = VDD
CKE = VSS
CE2 = VDD
CE2 = VSS
OE = VSS
ZZ = VSS
Status RAMs
StatusClockOut CLK
Table 2.8 Device to RAM Wiring Chart
External Components
Revision 1.2 Chapter 2, Hardw are Description 27
Due to the large capacitive load it represents, the board layout of the RAM array must use a
thick trac e to al low for a relatively la rge driv ing current for each addres s and data p in, and
must have an ample ground and power plane to insure proper de-coupling and minimize the
ground bounce potential.
An external clock distributor chip may be used to drive more RAM chips, as shown in Figure
2.10. The Data RAM cl ock outpu t (Pk tCl ockOu t) pin has the heavi es t lo ad at the hi ghe st duty
cycle. If that pin can be buffered effectively, the number of RAMs can be doubled. However,
take care to critically analyze the timing margin w hen designing a system in this way. In this
configuration, it is estimated that up to 16 chips can be driven, bringing the total capacity to
16MB, using 256K x 32 RAM.
StatusClockIn (Clock return from the end of the RAM array)
StatusCE# CE
StatusR/W# / StatusOE# R/W
StatusWE#
WE = VSS
OE = VSS
StatusAddr[20:0] SA, SA1, SA0
StatusData[7:0]
StatusData[15:8] DQa [7:0]
DQb[7:0]
Packet Data RAMs
PacketClockOut CLK
PacketClockIn (Clock return from the end of the RAM array)
PacketCE# CE
Pac ke t R /W# / Packet OE # R/W
PacketWE#
WE = VSS
OE = VSS
PacketAddr[20:0] SA, SA1, SA0
PacketData[63:32]
PacketData[31:0] DQa, DQb, DQc, DQd
DQa, DQb, DQc, DQd
Serenade Chip Pin-Out RAM
Table 2.8 Device to RAM Wiring Chart (continued)
External Components
28 ENT3002 Serenade Data Sheet Revision 1.2
Figure 2.10 External RAM Clock Buffering
Clock
Distributor
RAM RAM RAM RAM
RAM RAM RAM RAM
RAM RAM RAM RAM
RAM RAM RAM RAM
PktClockOut
PktClockIn
Out1
Out2
Out3
Skew between PktClockOut, Out1, Out2 and PktClockIn, including the wire delay, should be
less than 2nsec
Revision 1.2 Chapter 3, System Interface Operation 29
Chapter 3 System Interface Operation
This chapter describes the electrical characteristics and timing specifications for the Serenade
chip.
3.1 Electrical Characteristics
This section describes the following electrical characteristics of the Serenade chip:
The absolute maximum ratings
Power sequencing for power-up and reset
D.C. operating characteristics
I/O pin types and electrical characteristics
3.1.1 Absolute Maximum Ratings
The absolute maximum ratings for the Serenade chip are as follows:
3.1.2 Power Sequencing
The 2.5 V power supply should always be lower than the 3.3 V power supply, even at power-
up. For a pro per power-up re se t, the SysReset# si gnal must be low unti l all power-supp ly vol t -
ages reach at least to the lowest normal range. Figure 3.1 il lustrates the po wer s upply rise
curves and the SysReset# signal.
Note: The clock signal Osc25 must be within specified tolerance before the 2.5 V and 3.3 V
power su pplies re ach normal ope rating range. Thereafter, SysR eset# must remain asse rted
until the power supplies have stabilized for a minimum of 1 millisecond. Serenade requires a
maximum of 10 micro seconds after SysRes et# is negated to compl ete the internal initi alization
phase. After initialization is complete all generated clocks will be stable and within specified
tolerance.
For more information on the SysReset# signal, see Table 2.3 on page 16.
Storage Temperature -55°C to +125C°
Operating Power Supply Voltage (VDD pin) 3.6 V
Input Voltage with Respect to Ground -0.5V to 5.5V
ESD Immunity 2.0 KV human model
Electri cal Characteristics
30 ENT3002 Serenade Data Sheet Revision 1.2
Figure 3.1 Power Supply Rise Curves and System Reset
3.1.3 D.C. Operating Charact eristics
Table 3.1 provides a detailed description of the D.C. operating characteristics.
Symbol Parameter Limits
Min Typ Max Unit Test Condition/Comment
Ta Operating Ambient Temperature 0 70 °C
Vdd Operating Supply Volt ages:
VDD
IOVDD
RAMVDD
MXB3V
VDDAF1, 2
VDDAB
2.375 2.5 2.625
3.15 3.3 3.45
3.15 3.3 3.45
3.15 3.3 3.45
3.15 3.3 3.45
3.15 3.3 3.45
V
V
V
V
V
V
± 5%
Idd VDD currenta
IOVDD+RAMVDD+MXB3V
VDDAF1+VDDAF2+VDDAB
TBD
TBD
TBD
mA
mA
mA
All outputs high-Z, All inputs negated,
osc25 = 25MHz
@All xVDD = 3.3 V / 2.5 V
Table 3.1 D.C. Operating Characteristics
t
Vo
3.3V
2.5V
0
3.3V Supply
2.5V Supply
(always below 3.3V)
3.15V
2.375V
SysReset#
Tsra~ 10µsec
Osc25_Clock_In
Clock Outputs
Internal_Reset#
External Timing Specification
Revision 1.2 Chapter 3, System Interface Operation 31
3.1.4 I/O Pin Types and Electrical Characteristics
Table 3.2 shows the I/O pin types and electrical characteristics of the Serenade chip:
3.2 External Timing Specification
Refer to the applicable sections in the Serenade Data Book for details on timings for the
EdgeStream, Network, OptiStream, and SRAM interfaces.
Pd Maximum Power Dissipationb8.5 Watt All outputs active, All inputs active,
osc25 = 25MHz
@All xVDD at maximum
Iil Input Leakage Current
(All pins w/o pull-ups) -10 +10 µA -0.0V Vi VDD
Vi(max)
Vi(min)
Max. operating input voltage
Min. operating input voltage
5.0
3.6
0
V
V
V
5V tolerant pins
Non-5V tolerant & analog pins
all pins
Vih(Osc25)
Vil(Osc25) I nput high voltage for Osc25
Input low voltage for Osc25 1.6 0.9 V
VOsc25 clock input is at 2.5V logic
swing.
Vih/Vil
Ioh/Iol
VHyst
Various I/O parameters
(except for Osc25 pin above) See Table 3.2
Vbgr Band Gap Reference Voltage 1.23 1.25 1.26 V At BGR_Res pin, with 12.4Kto
ground attached
VTOptiS t ream Bus Termination
Voltage 1.35 1.5 1.65 V
a. Total current through all VDD pins at the specified condition.
b. The p ackage alone i s not expected to dissipate this much heat. Heatsink attachment and f orced air cooling is
required.
Type Dir. VHyst
(min) Vih / Vil
(min) (max) Internal
Pullup Ioh Iol 5V
Tolrnt Tri-
State Cload
low drive I /O no 0.65/0.35 VDD no -8.0mA
@Vo=2.4V 8.0mA
@Vo=0.4V yes yes 30pF
high drive I/O no 0.65/0.35 VDD no -24mA
@Vo=2.4V 24mA
@Vo=0.4V yes yes
input only I no 0.65/0.35 VDD no - - yes n/a n/a
hysterisis
input I 1V 2.15V/ 1.05V no - - yes n/a n/a
GTL open
drain I/O no 1.05/0.95 no (Open Drain) 40mA
@Vo=0.5V no yes
analog - n/a n/a no n/a n/a no n/a
Table 3.2 I/O Pin Types and Electrical Characteristics
Table 3.1 D.C. Operating Characteristics (continued)
External Timing Specification
32 ENT3002 Serenade Data Sheet Revision 1.2
Revision 1.2 Chapter 4, Programming Guide 33
Chapter 4 Programmi ng Guide
This part of the document describes the function, interface, and implementation of the config-
uration and management software for the Serenade chip. The software components and inter-
faces described here are defined using WindRiver Corporations VxWorks Real-Time
Operating System (RTOS).
The control softwar e for the Serenade ch ip consists of the foll owing major components :
The Serenade device driver: This driver consti tutes the lo w-level so ftware interface to the
programmable functions of the Serenade device. This component consists of an RTOS-depen-
dent sublay er and a Serena de devi ce-depend ent subla yer.
The Serenade service programming interface: This component consists of a high- l evel
library of C functions that presents an abstraction layer built on top of the facilities provided
by the RTOS device driver.
4.1 Theory of Operation
Refer to the Serenade Data Book for complete details on the theory of operation for program-
ming the Serena de chip. Included in th e Data Book is specific information for the follo wing
areas of operation.
Serenade Device Configuration
OptiStream Interface Operation
OptiStream Interface Physi cal C harac teristics
Inter-Serenade Packet Transfer
External Route / Filter Table Support
Multicast Transaction Handling
OptiStream Interface Chip ID Assignment
Internal OptiStream Tr an sactions
Transaction Flow Management
OptiStream Bus Arbitration Protocol
GTL Driver Slew Rate
OptiStream Port Transactions
External Scheduler Support
Theory of Operation
34 ENT3002 Serenade Data Sheet Revision 1.2
Packet Forwarding Process using External Scheduler
Inter-chip Data Transfer Prio ritization using Externa l Sche duler
Physical Requirem ents for External Units
OptiStream Transactions to support External Scheduler
EdgeStream Interface Operation
Accessing the Serenade chip
EdgeStream Interface Pack et Transfers
Reading packets from the Serenade chip
Writing Packets to the Serenade chip
Multicast Packet Handling
Register Read and Write Operations
Route Table
Packet Buffer Manager Operation
Packet Buffer Organization
Packet Buffer Queue Types
Using the Queues
Packet Buffer Operations
Operation of the Network Interface
PHY Interface
Layer 2 Filtering
Layer 3 Functions
Packet Receive Process
Network Management Process
Revision 1.2 Chapter 5, Register Reference 35
Chapter 5 Register Reference
5.1 Overview
The Serenade chip is mapped into 512 bytes (32-bit wide) of CPU address space. These regis-
ters are described in the Edge Stream Interface Registers section in the Serenade Data Book.
There are seve ral hu ndred more inter nal regist ers to this d evice. The se reg isters ca nnot be
accessed directly; they are accessed indirectly via the Control Bus Address and Control Bus
Data regi sters (80 and 84 hex r espec tively). T hese internal re gisters are also describ ed in th e
Inte rnal Registers section in the Se renade Data B ook.
Overview
36 ENT3002 Serenade Data Sheet Revision 1.2
Aria
Wire-Speed Router IC for Access Networks
Data Sheet Revision 1.1