( MPRH01TSU 02_603_604_Reference_Design_2.1_Aug95 02 603 604 Reference Design 2.1 Aug95

MPRH01TSU-02_603_604_Reference_Design_2.1_Aug95 MPRH01TSU-02_603_604_Reference_Design_2.1_Aug95

User Manual: MPRH01TSU-02_603_604_Reference_Design_2.1_Aug95

Open the PDF directly: View PDF PDF.
Page Count: 200

Download( MPRH01TSU-02_603_604_Reference_Design_2.1_Aug95 MPRH01TSU-02 603 604 Reference Design 2.1 Aug95
Open PDF In BrowserView PDF
PowerPC 603/604 Reference
Design Technical Specification
Release 2.1

This document provides a detailed
technical description of the PowerPC
603/604 Reference Design. It is intended as a first source of information
for both hardware and software designers. Where appropriate, other
documents are referenced.
Document Number:
MPRH01TSU-02
August, 9 1995

R

 International Business Machines Corporation, 1995. Printed in the United States of America 8/95. All Rights
reserved.
IBM Microelectronics, PowerPC, PowerPC 601, PowerPC 603, PowerPC 603e, PowerPC 604, RISCWatch,
and AIX are trademarks of the IBM corporation. IBM and the IBM logo are registered trademarks of the IBM
corporation. Other company names and product identifiers are trademarks of the respective companies.
This document contains information which is subject to change by IBM without notice. IBM assumes no responsibility or liability for any use of the information contained herein. Nothing in this document shall operate
as an express or implied license or indemnity under the intellectual property rights of IBM or third parties. The
products described in this document are not intended for use in implantation or other direct life-support applications where malfunction may result in physical harm or injury to persons. NO WARRANTIES OF ANY
KIND, INCLUDING BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY OR
FITNESS FOR A PARTICULAR PURPOSE, ARE OFFERED IN THIS DOCUMENT.

Contacts
USA and Canada:
IBM Microelectronics Division
1580 Route 52, Bldg. 504
Hopewell Junction, NY 12533-6531
Tel: (800) PowerPC
Fax: (800) PowerFax

Japan:
IBM
800, Ichimiyake
Yasu–cho, Yasu–gun
Shiga–ken, Japan 520–23
Tel: (81) 775–87–4745
Fax: (81) 775–87–4735

Europe:
IBM
La Pompignane BP 1021
34006 Montpellier, France
Tel: (33) 6713–5757 (Français)
(33) 6713–5756 (Italiano)
Fax: (33) 6713–5750
( from Paris add 16)

Europe:
IBM
Informations Systeme GmbH
Laatzener Str. 1
30539 Hannover, Germany
Tel: (49) 511–516–3444 (English)
(49) 511–516–3555 (Deutsche)
Fax: (49) 511–516–3888

ESD Warning
The motherboard and memory cards contain CMOS devices which are very susceptible
to ElectroStatic Discharge (ESD). DO NOT remove them from the antistatic bags until you
have connected yourself to an acceptable ESD grounding strap. Work in a static free environment and be sure any person or equipment coming into contact with the cards do not
have a static charge. The cards are particularly susceptible until they are placed in a properly designed enclosure. Bench work should be done by persons connected to ESD
grounding straps.

2

MPRH08TSU-02

IBM POWERPCTM 603/604 REFERENCE DESIGN AGREEMENT
BEFORE READING THE REST OF THE DOCUMENT, YOU SHOULD CAREFULLY READ THE FOLLOWING TERMS
AND CONDITIONS. OPENING THE PACKAGE INDICATES YOUR ACCEPTANCE OF THESE TERMS AND CONDITIONS. IF YOU DO NOT AGREE WITH THEM, YOU SHOULD PROMPTLY RETURN THE PACKAGE UNOPENED TO
YOUR IBM SALES OFFICE.
International Business Machines Corporation (”IBM”) agrees to provide you a PowerPC 603/604 Reference Design (Reference Design)
in return for your promise to use reasonable efforts to develop a system based on the technology in the Reference Design. The Reference
Design contains documentation and software listed below:
Documentation
PowerPC 603e RISC Microprocessor Hardware Specification
PowerPC 603e RISC Microprocessor Technical Summary
PowerPC 603/604 RISC Microprocessor Hardware Specification
PowerPC 603/604 Reference Design Technical Specification
IBM PowerPC 603/604 Reference Board Design Files (on 8mm tape)
IBM PowerPC 604 Reference Board Mfg. Data Files (in Gerber format)
IBM14N1372 Data Sheet
IBM11D4360B Data Sheet
Motorola MPC970 Data Sheet
Altera 5130 Data Sheet
Integrated Device Technology IDT71216 Data Sheet

LICENSE TO SOFTWARE
The software is licensed not sold. IBM, or the applicable IBM country organization, grants you a license for the software only in the country
where you received the software. Title to the physical software and documentation (not the information contained in such documentation)
transfers to you upon your acceptance of these terms and conditions. The term ”software” means the original and all whole or partial copies
of it, including modified copies or portions merged into other programs. IBM retains title to the software. IBM owns, or has licensed from
the owner, copyrights to the software provided under this agreement. The terms of this Agreement apply to all of the hardware, software
and documentation provided to you as part of the Reference Design.
With regard to the software provided hereunder, it is understood and agreed that you intend to use the software solely for the purpose of
designing PowerPC compatible products, testing your designs, and making your own independent determination of whether you wish to
eventually manufacture PowerPC compatible products commercially. In accordance with this understanding, IBM hereby grants you the
rights to: a) use, run, and copy the software, but only make such number of copies and run on such number of machines as are reasonably
necessary for the purpose of designing PowerPC compatible products and testing such designs; and b) copy the software for the purpose
of making one archival or backup copy.
With regard to any copy made in accordance with the foregoing license, you must reproduce any copyright notice appearing thereon. With
regard to the software provided hereunder, you may not: a) use, copy, modify or merge the software, except as provided in this license;
b) reverse assemble or reverse compile it; or c) sell, sublicense, rent. lease, assign or otherwise transfer it. In the event that you no longer
wish to use the software, you will return it to IBM.

LICENSE TO DESIGN DOCUMENTATION
With regard to the design documentation provided hereunder, it is understood that you intend to use such documentation solely for the purpose of designing your own PowerPC compatible products, testing your designs, and making your own independent determination of whether you wish to eventually manufacture PowerPC compatible products commercially. In accordance with this understanding, IBM hereby
grants you the right to: a) use the design documentation for the purpose of designing PowerPC compatible products and testing such designs; b) make derivative works of the design documentation for the purpose of designing PowerPC compatible products, and testing such
designs; and c) make copies of the design documentation and any such derivative works, but only such numbers as are reasonably necessary for designing PowerPC compatible products and testing such designs.
With regard to any copy made in accordance with the forgoing license, you must reproduce any copyright notice appearing thereon. With
regard to the design documentation provided hereunder, you may not: a) use, copy, modify, or merge the design documentation as provided
in this license; or b) sell, sublicense, rent, lease, assign, or otherwise transfer it.
In the event you no longer wish to use the design documentation or any derivative versions thereof, you must return them to IBM.

DISCLAIMER OF WARRANTY
IBM does not represent or warrant that the Reference Design (which may contain prototype items): a) meets any particular requirements;
b) operates uninterrupted; c) is error free; or d) is non–infringing of any patent, copyright, or other intellectual property right of any third party.
IBM makes no representation or warranty regarding the performance or compatibility that may be obtained from the use of the Reference
Design or that the Reference Design is adequate for any use. The Reference Design may contain errors and may not provide the level
of completeness, functionality, support, performance, reliability, or ease of use available with other products, whether or not similar to the
Reference Design. IBM does not represent or warrant that errors or other defects will be identified or corrected.

THE REFERENCE DESIGN IS PROVIDED ”AS IS” WITH ALL FAULTS, WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND
FITNESS FOR A PARTICULAR PURPOSE. THE ENTIRE RISK AS TO THE QUALITY AND PERFORMANCE OF THE
REFERENCE DESIGN IS WITH YOU.
Some jurisdictions do not allow exclusion of implied warranties, so the above exclusions may not apply to you.

MPRH08TSU-02

3

LIMITATION OF REMEDIES
IBM’s entire cumulative liability and your exclusive remedy for damages for all causes, claims or actions wherever and whenever asserted
relating in any way to the subject matter of this agreement including the contents of the Reference Design and any components thereof,
is limited to twenty five thousand dollars ($25,000.00) or its equivalent in your local currency and is without regard to the number of items
in the Reference Design that caused the damage. This limitation will apply, except as otherwise stated in this Section, regardless of the
form of the action, including negligence. This limitation will not apply to claims by you for bodily injury or damages to real property or tangible
personal property. In no event will IBM be liable for any lost profits, lost savings, or any incidental damages or economic consequential damages, even if IBM has been advised of the possibility of such damages, or for any damages caused by your failure to perform your responsibilities. In addition, IBM will not be liable for any damages claimed by you based on any third party claim. Some jurisdictions do not allow these
limitations or exclusions, so they may not apply to you.

RISK OF LOSS
You are responsible for all risk of loss or damage to the Reference Design upon its delivery to you.

IBM TRADEMARKS AND TRADE NAMES
This Agreement does not give you any rights to use any of IBM’s trade names or trademarks. You agree that should IBM determine that
any of your advertising, promotional, or other materials are inaccurate or misleading with respect to IBM trademarks or trade names, that
you will, upon written notice from IBM, change or correct such materials at your expense.

NO IMPLIED LICENSE TO IBM INTELLECTUAL PROPERTY
Notwithstanding the fact that IBM is hereby providing design information for your convenience, you expressly understand and agree that,
except for the rights granted under sections 1 and 2 above, no right or license of any type is granted, expressly or impliedly, under any patents, copyrights, trade secrets, trademarks, or other intellectual property rights of IBM. Moreover, you understand and agree that in the
event you wish to be granted any license beyond the scope of the expressly stated herein, you will contact IBM’s Intellectual Property Licensing and Services Office (currently located at 500 Columbus Avenue, Thornwood, N.Y.), or such other IBM offices responsible for the licensing of IBM intellectual property, when you seek the license.

YOUR ASSUMPTION OF RISK
You shall be solely responsible for your success in designing, developing, manufacturing, distributing, and marketing any product(s), or
portion(s), where use of all or any part of the Reference Design is involved. You are solely responsible for any claims, warranties, representations, indemnities and liabilities you undertake with your customers, distributors, resellers or others, concerning any product(s) or portion(s) of product(s) where use of all or any part of the Reference Design is involved. You assume the risk that IBM may introduce other
Reference Design that are somehow better than the Reference Design which is the subject of this Agreement. Furthermore, you accept
sole responsibility for your decision to select and use the Reference Design; for attainment or non-attainment of any schedule, performance,
cost, reliability, maintainability, quality, manufacturability or the like, requirements, or goals, self–imposed by you or accepted by you from
others, concerning any product(s) or portion(s) of product(s), or for any delays, costs, penalties, charges, damages, expenses, claims or
the like, resulting from such non-attainment, where use of all or any part of the Reference Design is involved.

GENERAL
In the event there is a conflict between the terms of this Agreement and the terms printed or stamped on any item or any ambiguities with
respect thereto, including documentation, contained in the Reference Design, the terms of this Agreement control to the extent IBM is afforded greater protection thereby. IBM may terminate this Agreement if you fail to comply with the terms and conditions of this Agreement.
Upon termination of this Agreement, you must destroy all copies of the software and documentation. You are responsible for payment of
any taxes, including personal property taxes, resulting from this Agreement. Neither party may bring an action hereunder, regardless of form,
more than one (1) year after the cause of the action arose. If you acquired the Reference Design in the United States, this Agreement is
governed by the laws of the State of New York. In the event of litigations, trial shall be in New York without a jury. If you acquired the Reference Design in Canada, this Agreement is governed by the laws of the Province of Ontario; otherwise, this Agreement is governed by the
laws of the country in which you acquired the Reference Design. All obligations and duties which, by their nature, survive termination or
expiration of this Agreement, shall remain in effect beyond termination or expiration of this Agreement, and shall bind IBM, you and your
successors and assigns. If any section or paragraph of this Agreement is found by competent authority to be invalid, illegal or unenforceable
in any respect for any reason, the validity, legality, and enforceability of any such section or paragraph in every other respect, and the remainder of this Agreement, shall continue in effect so long as it still expresses the intent of the parties. If the intent of the parties cannot be preserved, the parties will attempt to renegotiate this Agreement and failing renegotiation, this Agreement will then be terminated. The headings
in this Agreement shall not affect the meaning or interpretation of this Agreement in any way. No failure by IBM in exercising any right, power
or remedy under this Agreement shall serve as a waiver of any such right, power or remedy. Neither this Agreement nor any activities hereunder will impair any right of IBM to develop, manufacture, use or market, directly or indirectly, alone or with others, any products or services
competitive with those offered or to be offered by you; nor will this Agreement or any activities hereunder require IBM to disclose any business planning information to you. You agree to comply with all applicable government laws and regulations. Any changes to this Agreement
must be in writing and signed by the parties.

4

MPRH08TSU-02

Table of Contents
Section 1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.1 IBM Reference Products . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.1.1
Reference Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.1.2
Reference Board . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.1.3
Reference Firmware . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.1.4
Reference System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.2 Purpose . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.3 Reference Design Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.3.1
The CPU . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.3.2
IBM27-82660 Bridge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.3.3
L2 Cache . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.3.4
System Memory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.3.5
PCI Bus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.3.6
Flash ROM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.3.7
ISA Bus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.3.8
Time of Day Clock . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.3.9
PS/2t Compatible Keyboard/Mouse Controller . . . . . . . . . . . . . . . . .
1.3.10 System Clocks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.3.11 System I/O EPLD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.4 Quickstart Peripheral List . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.5 Reference Design Level . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

17
17
17
17
17
18
18
19
19
19
19
20
21
21
21
21
21
21
21
22
22

Section 2 CPU and CPU Bus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.1 CPU Bus Masters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.1.1
603e CPU . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.1.2
604 CPU . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.2 System Response by CPU Bus Transfer Type . . . . . . . . . . . . . . . . . . . . .
2.3 System Response by CPU Bus Address Range . . . . . . . . . . . . . . . . . . . .
2.3.1
Address Mapping for Non-Contiguous I/O . . . . . . . . . . . . . . . . . . . . . .
2.3.2
Address Mapping for Contiguous I/O . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.3.3
PCI Final Address Formation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.4 CPU to Memory Transfers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.4.1
LE Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.5 CPU to PCI Transactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.5.1
CPU to PCI Read . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.5.2
CPU to PCI Write . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.5.2.1 Eight-Byte Writes to the PCI (Memory and I/O) . . . . . . . . . . . . . . . . .
2.5.3
CPU to PCI Memory Transactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.5.4
CPU to PCI I/O Transactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.5.5
CPU to PCI Configuration Transactions . . . . . . . . . . . . . . . . . . . . . . . .
2.5.5.1 Preferred Method of Generating PCI Configuration Transactions .

25
25
25
26
26
28
28
30
30
30
31
31
31
31
31
32
32
32
32

MPRH08TSU-02

5

2.5.5.2 650 Bridge compatible method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.5.6
CPU to PCI Interrupt Acknowledge Transaction . . . . . . . . . . . . . . . . .
2.5.7
PCI Lock . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.6 CPU to ROM Transfers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.6.1
CPU to ROM Read . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.6.2
CPU to ROM Write . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.6.2.1 ROM Write Protection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2.6.3
CPU to BCR Transfers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

32
33
33
33
34
34
34
34

Section 3 PCI Bus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.1 PCI Transaction Decoding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.1.1
PCI Transaction Decoding By Bus Command . . . . . . . . . . . . . . . . . . .
3.1.2
PCI Memory Transaction Decoding By Address Range . . . . . . . . . . .
3.1.3
PCI I/O Transaction Decoding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.1.4
ISA Master Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2 PCI Transaction Details . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.1
Bus Snooping on PCI to Memory Cycles . . . . . . . . . . . . . . . . . . . . . . .
3.2.2
PCI to PCI Peer Transactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.2.3
PCI to System Memory Transactions . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.3 Bus Arbitration Logic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.4 Other PCI Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

35
35
36
36
38
38
38
39
39
39
41
41

Section 4 ISA Bus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.1 The ISA Bridge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.2 Address Ranges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.3 ISA Bus Concurrency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.4 ISA Bus Masters and IGN_PCI_AD31 . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.5 DMA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.5.1
Supported DMA Paths . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.5.2
DMA Timing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.5.3
Scatter-Gather . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.6 X-Bus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.6.1
Control Signal Decodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.6.2
Keyboard/Mouse Controller . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.6.3
Real Time Clock (RTC) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.6.4
PCI Adapter Card Presence Detect Register . . . . . . . . . . . . . . . . . . . .
4.6.5
L2 SRAM Identification Register . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.6.6
Planar ID Detection Register . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.6.7
DRAM Presence Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.6.7.1 DRAM SIMM 1-2 Memory ID Register . . . . . . . . . . . . . . . . . . . . . . . . .
4.6.7.2 DRAM SIMM 3-4 Memory ID Register . . . . . . . . . . . . . . . . . . . . . . . . .
4.7 Miscellaneous . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.7.1
Speaker Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

43
43
44
44
44
46
46
46
46
47
47
47
47
47
46
48
49
49
49
50
60

Section 5 System I/O EPLD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.1 System Register Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

51
51

6

MPRH08TSU-02

5.1.1
External Register Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.1.2
Internal Registers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.1.2.1 Storage Light Register . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.1.2.2 Power Management Control Register 1 . . . . . . . . . . . . . . . . . . . . . . .
5.1.2.3 Power Management Control Register 2 . . . . . . . . . . . . . . . . . . . . . . .
5.1.2.4 Freeze Clock Register (FCR) Low . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.1.2.5 Freeze Clock Register (FCR) High . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.2 Signal Descriptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.3 EPLD Design Equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.3.1
Fit File . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
5.3.2
TDF File . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

51
52
52
52
53
53
53
54
57
57
59

Section 6 Memory Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.1 DRAM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.1.1
Refresh . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.1.2
DRAM Presence Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.1.3
Organization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.2 L2 Cache . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.2.1
SRAM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.2.2
TagRAM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.2.3
L2 Cache Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.3 ROM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.4 CPU to ROM Transfers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.4.1
CPU to ROM Read . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.4.2
CPU to ROM Write . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.4.2.1 ROM Write Protection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

67
67
67
67
68
69
69
72
73
74
74
74
74
75

Section
7.1
7.2
7.3
7.4
7.5
7.6
7.7
7.8
7.9
7.10
7.11
7.12

7 Endian Mode Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
What the 603/604 CPU Does . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
What the 660 Bridge Does . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Bit Ordering Within Bytes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Byte Swap Instructions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
603/604 CPU Alignment Exceptions In LE Mode . . . . . . . . . . . . . . . . . . .
Single–Byte Transfers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Two–Byte Transfers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Four–Byte Transfers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Three byte Transfers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Instruction Fetches and Endian Modes . . . . . . . . . . . . . . . . . . . . . . . . . . .
Changing BE/LE Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Summary of Bi–Endian Operation and Notes . . . . . . . . . . . . . . . . . . . . . .

77
77
78
79
79
80
80
84
86
88
89
90
92

Section 8 Exceptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
8.1 Interrupts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
8.1.1
System Interrupt Handler . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
8.1.2
Interrupt Handling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

93
93
93
94

MPRH08TSU-02

7

8.1.3
Interrupt Assignments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
8.1.4
Scatter-Gather Interrupts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
8.2 Error Handling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
8.2.1
Data Error Checking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
8.2.1.1 CPU to Memory Writes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
8.2.1.2 CPU to Memory Reads . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
8.2.1.3 PCI to Memory Parity Errors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
8.2.1.4 CPU to PCI Transaction Data Parity Errors . . . . . . . . . . . . . . . . . . . .
8.2.2
Illegal CPU cycles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
8.2.3
SERR, I/O Channel Check, and NMI Logic . . . . . . . . . . . . . . . . . . . . . .
8.2.4
Out of Bounds PCI Memory Accesses . . . . . . . . . . . . . . . . . . . . . . . . . .
8.2.5
No Response on CPU to PCI Cycles – Master Abort . . . . . . . . . . . . .
8.2.6
CPU to PCI Cycles That Are Target Aborted . . . . . . . . . . . . . . . . . . . .
8.2.7
Error Status Registers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
8.2.8
Reporting Error Addresses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
8.2.9
Errant Masters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
8.2.10 Special Events Not Reported as Errors . . . . . . . . . . . . . . . . . . . . . . . . .

95
95
96
96
96
96
97
97
97
97
97
98
98
98
98
98
99

Section 9 System Setup and Initialization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
9.1 CPU Initialization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
9.2 660 Bridge Initialization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
9.3 ISA Bridge (SIO) Initialization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
9.3.1
Summary of SIO Configuration Registers . . . . . . . . . . . . . . . . . . . . . . .
9.4 PCI Configuration Scan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
9.4.1
Multi-function Adaptors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
9.4.2
PCI to PCI Bridges . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
9.5 Reference Design Combined Register Listing . . . . . . . . . . . . . . . . . . . . . .
9.5.1
Direct Access Registers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
9.5.2
Indexed BCR Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
9.6 ISA Bus Register Suggestions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

101
101
102
103
106
107
107
107
108
108
114
116

Section 10 System Firmware . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
10.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
10.2 Power On System Test . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
10.2.1 Hardware Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
10.3 Boot Record Format . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
10.3.1 Boot Record . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
10.3.1.1 PC Partition Table Entry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
10.3.1.2 Extended DOS Partition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
10.3.1.3 PowerPC Reference Platform Partition Table Entry . . . . . . . . . . . . .
10.3.2 Loading the Load Image . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
10.4 System Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
10.4.1 System Console . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
10.4.2 System Initialization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
10.4.3 Main Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

119
119
119
119
119
120
120
121
122
123
125
125
125
126

8

MPRH08TSU-02

10.4.3.1 System Configuration Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
10.4.3.2 Run a Program . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
10.4.3.3 Reprogram Flash Memory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
10.4.3.4 Exit Options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
10.4.4 Default Configuration Values . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

127
133
134
135
135

Section 11 Electromechanical . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
11.1 Electrical . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
11.1.1 Power Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
11.1.2 Onboard 3.3V Regulator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
11.1.3 Onboard 2.5V Regulator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
11.2 Thermal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
11.2.1 Thermal Requirements for the 603/604 Processor . . . . . . . . . . . . . . .
11.2.1.1 604 Fan-Sink Installation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
11.2.1.2 604 Fan-Sink Experimentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
11.2.1.3 Thermal Requirements for the 3.3V Regulator . . . . . . . . . . . . . . . . .
11.3 Mechanical . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
11.3.1 Reference Design Board Mechanical . . . . . . . . . . . . . . . . . . . . . . . . . . .
11.3.2 Connector Locations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
11.3.3 Connector Locator Diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
11.3.4 Keyboard Connector J14 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
11.3.5 Mouse Connector J15 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
11.3.6 Speaker Connector J13 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
11.3.7 Power Good LED/KEYLOCK# Connector J12 . . . . . . . . . . . . . . . . . . .
11.3.8 HDD LED Connector J11 (1 x 2 Berg) . . . . . . . . . . . . . . . . . . . . . . . . . .
11.3.9 Reset Switch Connector J10 (1 x 2 Berg) . . . . . . . . . . . . . . . . . . . . . . .
11.3.10 Fan Connector J9 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
11.3.11 3.3V Power Connector J5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
11.3.12 Power Connector J4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
11.3.13 AUX5/ON-OFF Connector J6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
11.3.14 PCI Connectors J25, J26, and J27 . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
11.3.15 ISA Connectors J29, J30, J31, J32, and J33 . . . . . . . . . . . . . . . . . . . .
11.3.16 SIMM Connectors J21, J22, J23, and J24 . . . . . . . . . . . . . . . . . . . . . .
11.3.17 Power Switch Connector J8 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
11.3.18 Power Up Configuration Connector J7 . . . . . . . . . . . . . . . . . . . . . . . . .
11.3.19 L2 Cache Data SIMM Connector J3 . . . . . . . . . . . . . . . . . . . . . . . . . . . .
11.3.20 RISCWatch Connector J19 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
11.3.21 Battery Connector BT2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
11.3.22 Reference Design Board Connector Footprint #1 . . . . . . . . . . . . . . . .
11.3.23 Reference Design Board Connector Footprint #2 . . . . . . . . . . . . . . . .
11.4 Enclosure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

137
137
137
138
138
139
139
139
140
140
141
141
142
143
144
144
145
145
145
146
146
147
147
148
148
150
152
154
154
155
157
157
158
159
160

Section 12 Physical Design Guidelines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
12.1 General Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
12.1.1 Construction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

161
161
161

MPRH08TSU-02

9

12.1.2 General Wiring Guidelines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
12.2 Clock Nets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
12.3 CPU Bus Nets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
12.4 Timing Critical Nets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
12.5 PCI Bus Nets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
12.6 Group 2A: Noise Sensitive Wires . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
12.7 8mm Tape Contents and Extract Instructions . . . . . . . . . . . . . . . . . . . . . .
12.7.1 Download Instructions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
12.7.2 Cadence Version . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
12.7.3 Tape Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

164
164
165
166
167
167
168
168
168
168

Section 13 Errata for Reference Design Release 2.1 . . . . . . . . . . . . . . . . . . . . .
13.1 PowerPC 603/604 Reference Design Roadmap . . . . . . . . . . . . . . . . . . . .
13.2 Release 2.1 Board Level Errata . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
13.2.1 Driving 32MB 72-pin DRAM SIMMs . . . . . . . . . . . . . . . . . . . . . . . . . . . .
13.2.2 CPU Data Bus 33 Ohm Series Resistors . . . . . . . . . . . . . . . . . . . . . . .
13.3 Reference Design Errata for Revision 1.1 of the 660 Bridge . . . . . . . . .
13.3.1 Workaround Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
13.3.2 HPP1 Reference Design Workaround PAL . . . . . . . . . . . . . . . . . . . . . .
13.3.3 HCn Reference Design Workaround PAL . . . . . . . . . . . . . . . . . . . . . . .
13.3.4 HBROOM Reference Design Workaround PAL . . . . . . . . . . . . . . . . . .
13.4 660 Bridge Revision 1.1 Errata, 8/9/95 Release . . . . . . . . . . . . . . . . . . . .
13.4.1 Individual 66 Bridge Errata . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
13.4.2 IPAL Design Files . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
13.4.2.1 HPP1 Design File . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
13.4.2.2 C3 PAL Design File . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
13.4.2.3 Broom PAL Design File . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
13.4.3
Workaround PAL Installation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

169
169
171
171
171
172
172
172
173
175
177
177
182
182
184
186
188

Section 14 Bill of Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
14.1 603e/604 Reference Design Bill of Materials . . . . . . . . . . . . . . . . . . . . . . .
14.1.1 603e Bill of Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
14.1.2 604 Bill of Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

189
189
189
194

Section 15 Schematics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
15.1 Reference Board Component Placement . . . . . . . . . . . . . . . . . . . . . . . . . .

199
200

Section 16 Selected Component Data Sheets . . . . . . . . . . . . . . . . . . . . . . . . . . .

10

MPRH08TSU-02

Figures
Figure 1. 603/604 Reference Design Block Diagram . . . . . . . . . . . . . . . . . . .
Figure 2. Non-Contiguous PCI I/O Address Transformation . . . . . . . . . . . . .
Figure 3. Non-Contiguous PCI I/O Address Translation . . . . . . . . . . . . . . . . .
Figure 4. Contiguous PCI I/O Address Translation . . . . . . . . . . . . . . . . . . . . .
Figure 5. Typical External Register . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure 6. DRAM Bank Organization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure 7. Synchronous SRAM, 256K L2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure 8. Synchronous SRAM, 512K L2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure 9. Synchronous SRAM, 1M L2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure 10. Asynchronous SRAM, 256K L2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure 11. Asynchronous SRAM, 512K L2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure 12. Asynchronous SRAM, 1M L2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure 13. Synchronous TagRAM, 512K L2 . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure 14. Synchronous TagRAM, 1M L2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure 15. Endian Mode Block Diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure 16. Example at Address XXXX XXX0 . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure 17. Example at Address XXXX XXX2 . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure 18. Double Byte Write Data ab at Address XXXX XXX0 . . . . . . . . . . .
Figure 19. Word (4-Byte) Write of 0a0b0c0dh at Address XXXX XXX4 . . . . .
Figure 20. Instruction Alignment Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure 21. Wrong Instruction Read When Unmunger is used . . . . . . . . . . . . .
Figure 22. Instruction Stream to Switch Endian Modes . . . . . . . . . . . . . . . . . . .
Figure 23. Interrupt Handling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure 24. PCI Interrupt Connections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure 25. Boot Record . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure 26. Partition Table Entry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure 27. Partition Table Entry Format for an Extended Partition . . . . . . . . .
Figure 28. Partition Table Entry for PowerPC Reference Platform . . . . . . . . .
Figure 29. PowerPC Reference Platform Partition . . . . . . . . . . . . . . . . . . . . . . .
Figure 30. System Initialization Screen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure 31.Configuration Utility Main Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure 32. System Configuration Menu . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure 33. System Information Screen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure 34. Device Configuration Screen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure 35. SCSI Devices Screen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure 36. Boot Devices Screen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure 37. Set Date and Time Screen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure 38. Run a Program Screen . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure 39. Reprogram the Flash Memory Screen . . . . . . . . . . . . . . . . . . . . . . .
Figure 40. 604 Heat Sink Assembly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure 41. Connector Location Diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
MPRH08TSU-02

11

20
29
29
30
51
68
69
70
70
71
71
72
72
72
79
80
81
84
84
89
90
91
94
95
120
120
122
122
123
125
126
127
128
129
130
131
132
133
134
139
143

Figure 42.
Figure 43.
Figure 44.
Figure 45.
Figure 46.
Figure 47.
Figure 48.
Figure 49.
Figure 50.
Figure 51.
Figure 52.
Figure 53.
Figure 54.
Figure 55.
Figure 56.
Figure 57.
Figure 58.
Figure 59.
Figure 60.
Figure 61.
Figure 62.
Figure 63.
Figure 64.

The Keyboard Connector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
The Mouse Connector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1x4 Speaker Connector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1x5 Power Good LED Connector . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1x2 HDD LED Connector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1x2 Reset Switch Connector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1x2 Fan Connector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1x6 3.3V Power Connector J5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1x12 Power Connector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
AUX5/ON-OFF Connector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
PCI Connector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
ISA Connector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
SIMM Connector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1x2 Power Switch Connector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1x2 Power Up Configuration Connector . . . . . . . . . . . . . . . . . . . . . .
L2 SRAM Module Connector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2x8 RISCWatch Connector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Signal and Power Layers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Typical Wiring Channel Top View . . . . . . . . . . . . . . . . . . . . . . . . . . . .
PowerPC 603/604 Board Fabrication . . . . . . . . . . . . . . . . . . . . . . . .
Power Plane Split . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
PCI Interrupt Slot 0 Connections . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Generic Implementation PAL Installation . . . . . . . . . . . . . . . . . . . . .

12

144
144
145
145
145
146
146
147
147
148
148
150
152
154
154
155
157
161
162
163
164
171
188

MPRH08TSU-02

Tables
Table 1. Quickstart Peripheral List . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Table 2. Example Clock Generator Frequencies . . . . . . . . . . . . . . . . . . . . . . . .
Table 3. TT[0:3] (Transfer Type) Decoding by 660 Bridge . . . . . . . . . . . . . . . .
Table 4. 660 Bridge Address Mapping of CPU Bus Transactions . . . . . . . . . .
Table 5. 660 Bridge Address Mapping of CPU Bus Transactions . . . . . . . . . .
Table 6. Reference Design Responses to PCI_C[3:0] Bus Commands . . . . .
Table 7. Mapping of PCI Memory Space, Part 1 . . . . . . . . . . . . . . . . . . . . . . . . .
Table 8. Mapping of PCI Memory Space, Part 2 . . . . . . . . . . . . . . . . . . . . . . . . .
Table 9. Mapping of PCI Master I/O Transactions . . . . . . . . . . . . . . . . . . . . . . .
Table 10. Active CAS# Lines – PCI to Memory Writes, BE or LE Mode . . . . .
Table 11. DMA Assignments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Table 12. DRAM Module Presence Detect Bit Encoding . . . . . . . . . . . . . . . . . .
Table 13. Planar ID Encoding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Table 14. DRAM Module Presence Detect Bit Encoding . . . . . . . . . . . . . . . . . .
Table 15. External Register Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Table 16. Signal Descriptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Table 17. Supported DRAM Modules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Table 18. L2 Configuration Implementation . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Table 19. Endian Mode Byte Lane Steering . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Table 20. Endian Mode 6–3/604 Address Translation . . . . . . . . . . . . . . . . . . . .
Table 21. Memory in BE Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Table 22. Memory in LE Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Table 23. PCI in BE Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Table 24. PCI in LE Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Table 25. Two Byte Transfer Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Table 26. Rearranged 2-Byte Transfer Information . . . . . . . . . . . . . . . . . . . . . .
Table 27. 4-Byte Transfer Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Table 28. Rearranged 4–Byte Transfer Information . . . . . . . . . . . . . . . . . . . . . .
Table 29. Mapping of PCI Memory Space, Part 1 . . . . . . . . . . . . . . . . . . . . . . .
Table 30. Summary of SIO Register Setup
(Configuration Address = 8080 08xx) . . . . . . . . . . . . . . . . . . . . . . . .
Table 31. Summary of SIO Configuration Registers . . . . . . . . . . . . . . . . . . . . .
Table 32. Configuration Address Assignments . . . . . . . . . . . . . . . . . . . . . . . . . .
Table 33. Combined Register Listing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Table 34. 660 Bridge Indexed BCR Listing . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Table 35. Compatible ISA Ports (Not on Reference Board) . . . . . . . . . . . . . .
Table 36. Power Supply Specification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Table 37. Approximate Power Consumption . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Table 38. Specifications for 3.3V Regulator on the Motherboard . . . . . . . . . . .
Table 39. Specifications for 2.5V Regulator on the Motherboard . . . . . . . . . . .
Table 40. Keyboard Connector Pin Assignments . . . . . . . . . . . . . . . . . . . . . . . .
Table 41. Mouse Connector Pin Assignments . . . . . . . . . . . . . . . . . . . . . . . . . .
MPRH08TSU-02

13

23
26
27
28
33
36
36
37
38
40
46
48
48
49
51
54
67
73
78
78
81
82
82
83
85
85
87
88
95
104
106
107
108
114
116
137
137
138
138
144
144

Table 42.
Table 43.
Table 44.
Table 45.
Table 46.
Table 47.
Table 48.
Table 49.
Table 50.
Table 51.
Table 52.
Table 53.
Table 54.
Table 55.
Table 56.
Table 57.
Table 58.
Table 59.
Table 60.
Table 61.
Table 62.
Table 63.
Table 64.
Table 65.

Speaker Connector Pin Assignments . . . . . . . . . . . . . . . . . . . . . . . . .
Power Good LED Connector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
HDD LED Connector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Reset Switch Connector . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Fan Connector Pin Assignments . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.3V Power Connector J5 Pin Assignments . . . . . . . . . . . . . . . . . . .
Power Connector J4 Pin Assignments . . . . . . . . . . . . . . . . . . . . . . . .
AUX5/ON-OFF Connector Pin Assignments . . . . . . . . . . . . . . . . . . .
PCI Connector Pin Assignments . . . . . . . . . . . . . . . . . . . . . . . . . . . .
ISA Connector Pin Assignments . . . . . . . . . . . . . . . . . . . . . . . . . . . .
SIMM Connector Pin Assignments . . . . . . . . . . . . . . . . . . . . . . . . . .
Power Switch Connector Pin Assignments . . . . . . . . . . . . . . . . . . . .
Power Up Configuration Connector Pin Assignments . . . . . . . . . . .
L2 SRAM Module Connector Pin Assignments . . . . . . . . . . . . . . . .
RISCWatch Connector Pin Assignments . . . . . . . . . . . . . . . . . . . . . .
Height Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Clock Net Lengths ( 7  a  10 ) . . . . . . . . . . . . . . . . . . . . . . . . . . . .
CPU Bus Nets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Timing Critical Nets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
PCI Bus Nets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Noise Sensitive Nets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
PowerPC 603/604 Reference Design Roadmap . . . . . . . . . . . . . . . .
603e Bill of Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
604 Bill of Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

14

145
145
145
146
146
147
147
148
148
150
152
154
154
155
157
160
165
165
166
167
167
170
189
194

MPRH08TSU-02

About This Book
Notice:
The 603 version of the PowerPC 603/604 Reference Design is no longer supported. It is replaced by the 603e
version of the 603/604 PowerPC Reference Design. In this document, the term 603 also refers to the PowerPC 603e RISC microprocessor unless otherwise specified.
Power management is beyond the scope of this document.

Audience:
This reference design is designed for engineers and system designers who are interested in implementing
PowerPC systems that are compliant with the PowerPC Reference Platform Specification. The material requires a detailed understanding of computer systems at the hardware and software level.

Reference Material:
Understanding of the relevant areas of the following documents is required for a good understanding of the
reference design:
PowerPC 604 User’s Manual, IBM document MPR604UMU-01
PowerPC 604 Hardware Specification, IBM document MPR604HSU-01
PowerPC 603 User’s Manual, IBM document MPR603UMU-01
PowerPC 603e Hardware Specification, IBM document MPR603EHS-01
PowerPC 603e Technical Summary, IBM document MPR603TSU-04
IBM27-82660 PowerPC to PCI Bridge User’s Manual, IBM document number MPR660UMU–01
PCI Local Bus Specification, Revision 2.1, available from the PCI SIG
PowerPC Reference Platform Specification, Version 1.1, IBM document MPRPRPPKG
The Power PC Architecture, second edition, Morgan Kaufmann Publishers
(800) 745–7323, IBM document MPRPPCARC–02
S Intel 82378ZB System I/O (SIO) Data Book, Intel order number 290473-004.

S
S
S
S
S
S
S
S
S

The following documents are useful as sources of tutorial and supplementary information about the reference
design.
S PowerPC System Architecture, Tom Shanley, Mindshare Press (800) 420-2677.
S IBM27-82650 PowerPC to PCI Bridge User’s Manual, IBM document number MPR650UMU–01

Document Conventions:
Kilobytes, megabytes, and gigabytes are indicated by a single capital letter after the numeric value. For example, 4K means 4 kilobytes, 8M means 8 megabytes, and 4G means 4 gigabytes.
The terms DIMM and SIMM are often used to mean DRAM module.
Hexadecimal values are identified (where not clear from context) with a lower-case letter h at the end of the
value. Binary values are identified (where not clear from context) with a lower-case letter b at the end of the
value.
In identifying ranges of values from and to are used whenever possible. The range statement from 0 to 2M
means from and including zero up to (but not including) two megabytes. The hexadecimal value for the range
from 0 to 64K is: 0000h to FFFFh.
The terms asserted and negated are used extensively. The term asserted indicates that a signal is active
(logically true), regardless of whether that level is represented by a high or low voltage. The term negated
means that a signal is not asserted. The # symbol at the end of a signal name indicates that the active state
of the signal occurs with a low voltage level.

MPRH08TSU-02

15

16

MPRH08TSU-02

Introduction

Section 1
Introduction
This document provides a detailed technical description of the PowerPCt 603/604 Reference Design, and is intended to be used by hardware, software, test, simulation, and other
engineers as a first source of information. Software developers should read through the
entire document because pertinent facts may be located in hardware sections.
The focus of this document is mainly the motherboard electronics and firmware. Where appropriate this document references detailed information in other documents. Consult other
documents for information on specific I/O devices such as hard drives, CD–ROMs, L2
cache cards, video cards, etc. that comprise a total system.
Recommendations for memory mappings, software implementations, and the like are only
recommendations and may or may not represent the algorithms implemented in boot code
or operating systems.
Note: This document contains several references to the 603/604 Reference Design Power
Management Specification; however, the specification was not yet available at the date of
printing.

1.1

IBM Reference Products

IBM offers several different PowerPC reference products for a given PowerPC system.
1.1.1 Reference Design
The PowerPC 603/604 Reference Design (reference design) is composed of both the intangible design and the documentation describing that design. The reference design documentation addresses the motherboard electronics, firmware, and various system related
issues. The reference design contains this reference design Technical Specification, Gerber format physical design files on an 8mm tape, electrical device model files in Cadencet
format, system firmware guidelines, schematics, contact information for commented boot
ROM source code, and such other device and system information as is deemed helpful.
1.1.2 Reference Board
The PowerPC 603/604 Reference Board (reference board) is the physical implementation
of the motherboard part of the reference design. It includes the reference design, the populated motherboard, the boot ROM, and other components as appropriate.
1.1.3 Reference Firmware
The PowerPC 603/604 Reference Firmware (reference firmware) is described in the reference design, and consists of the commented source code of the software contained in the
boot ROM. This is available from IBM as discussed in section 10.1.
MPRH01TSU-02

17

Introduction
1.1.4 Reference System
The PowerPC 603/604 Reference System (reference system) consists of a complete
603/604 PowerPC computer system, including the motherboard, enclosure, power supply,
cooling devices, and such other adaptors and peripherals as are described in the detailed
product offering.

1.2

Purpose

The reference design is aimed at the market for low cost desk top PowerPC personal computers. The motherboard is sized to fit within a BabyAT form factor enclosure, although the
enclosure may need to be modified to provide additional cooling and/or additional space
for I/O cards.
The reference design is intended to help companies develop their own products using the
PowerPC architecture. The reference design may be used:
S As a baseline system in order to gauge the effects of changes on the design
S To test new boot code
S To test operating systems and/or applications
S For performance measuring.
The reference design is:
S A compliant implementation of the PowerPC Hardware Reference Platform Specification, version 1.1
S Tested for functionality to the level of software available at the time of shipping
S A prototype of a system under development which may have prototype ASICs, errata, and/or wiring changes.
The reference design is not:
S A complete market ready design
S Tested for compliance to FCC and other regulatory requirements.

18

MPRH01TSU-02

Introduction

1.3

Reference Design Overview

This section contains an overview of the reference design. The block diagram of the reference design is shown in Figure 1.
The reference design is compliant with the PowerPC Reference Platform Specification Version 1.1.
The core of the system is the PowerPC 603et or PowerPC 604t RISC microprocessor.
The IBM27-82660 Bridge chipset (660 Bridge) interfaces the CPU to the DRAM memory,
and provides L2 cache control for the tagRAM and SRAM components that are located on
the CPU bus.
The 660 Bridge also interfaces the CPU to the PCI bus. On the motherboard, the ISA bus
bridge is located on the PCI bus. The boot ROM is also physically located on the PCI bus,
but is accessed using a special protocol. PCI devices are unable to activate the ROM, and
ROM operations do not interfere with the PCI bus protocol.
The reference design also provides three PCI slots, by which major I/O subsystems, such
as SCSI and video adaptors, can be connected to the system. The reference design also
provides five ISA slots.
The motherboard is designed to an industry standard BabyAT (8.6 in. by 13 in.) form factor.
It requires +5V to power most of the components. The motherboard also requires 12 V
to support some of the peripheral features. Components that require +3.3V (such as the
PCI bus agents) are supported using a regulator mounted on the motherboard to convert
+5V to +3.3V.
1.3.1 The CPU
The reference board can be configured with either the 603e, or the 604 implementation of
the PowerPC architecture. Only one CPU may be installed at a time.
These CPUs use a CPU bus clock, and are in general capable of running their internal clock
at several different multiples of the bus clock frequency. The reference design runs the PCI
bus clock at a fixed frequency multiple of one half of the CPU bus clock frequency. The CPU
bus clock frequency is adjustable, and is nominally 66MHz.
The reference design supports bi-endian operation, and is equipped with an ESP connector to support RISCWatch debugging and monitor systems.
Consult your IBM representative for currently available choices of CPU type and operating
frequency.
1.3.2 IBM27-82660 Bridge
The IBM27-82660 Bridge (660 Bridge) chipset supplies many of the functions of the reference design. The 660 Bridge interfaces the CPU to the L2, system memory, the PCI bus,
the ROM, and other reference design components.
1.3.3 L2 Cache
The reference design supplies an L2 cache controller, located inside the 660 Bridge chipset. The motherboard provides a socket for an SRAM module, and the L2 tag RAM (16K
x 15, synchronous) is supplied installed on the board. The L2 is a unified, write-thru, directMPRH01TSU-02

19

Introduction
mapped, look-aside cache that supports 1M of SRAM to cache the low 1G of CPU memory
space. The L2 supplies data to the CPU bus on write hits and snarfs the data (updates the
SRAM data while the memory controller is accessing DRAM memory) on read/write misses. It snoops PCI to memory transactions. Typical read performance with 9ms SRAM is
3-1-1-1, followed by -2-1-1-1 on pipelined reads for sync.

Clock Gen

Power
Mgt Cntrlr
87C750

MPC970

System I/O
EPLD
EPM5130C

X-Bus
Strobes
RTC
DC1385

ESP Connector
PD

KYBD
Mouse

Data
Addr/Cntrl

603
604
Data

Keyboard
Mouse
8042

X-Bus
Address

X-Buffers
L2 SRAM
Module

ISA Expansion
Slots (5)

PD
Speaker

TagRAM
C

60X Bus

ISA Bus
Bridge

C

IBM27-82664
Controller

Data
Control
Address

SIO

ISA Bus
PCI Bus

IBM27-82663
Buffer

PCI Expansion
Slots (3)

PD
Control
DRAM Module
Address/Data

DRAM Module
5v to 3.3v
Regulator

PCI_AD

DRAM Module

Flash ROM
29F040

DRAM Module

Figure 1. 603/604 Reference Design Block Diagram
1.3.4 System Memory
The reference design memory subsystem can support up to 128M of 70ns DRAM memory
on four 72 pin modules via sockets. Each SIMM socket can support an 8M or 32M 72 pin
SIMM. The DRAM subsystem is 72 bits wide: 64 data bits and eight parity bits. One parity
bit is generated for each byte of data written. The 660 Bridge can also be configured to per20

MPRH01TSU-02

Introduction
form ECC memory data checking and correction using standard parity DRAM modules. Or
it can be configured to disable DRAM parity checking for systems using non-parity DRAM.
The 660 Bridge also provides DRAM refresh and supports EDO hyper-page mode DRAM.
Memory access performance from the CPU bus at 66MHz with 70ns DRAM is typically:
S Pipelined burst read: 4-4-4-4 CPU bus clocks—16 CPU clocks for 32 bytes of data
S Pipelined burst write: 5-4-4-4 CPU bus clocks—17 CPU clocks for 32 bytes of data
Memory access performance from the PCI bus at 33MHz with 70ns DRAM is typically:
S Read bursts 5-1-1-1 -1-1-1-1 6-1-1-1 -1-1-1-1 6-1-1-1 -1-1-1-1 ... 6-1-1-1 -1-1-1-1,
S Write bursts 5-1-1-1 -1-1-1-1 3-1-1-1 -1-1-1-1 3-1-1-1 -1-1-1-1 ... 3-1-1-1 -1-1-1-1.
1.3.5 PCI Bus
The 660 Bridge includes the interface between the PCI bus and the rest of the system. The
reference design allows CPU to PCI access and PCI bus master to memory access (with
snooping), and handles all PCI related system memory cache coherency issues. Three PCI
expansion slots are provided.
The reference design also supports memory block locking, types 0 and 1 configuration
cycles, and ISA master access to system memory thru the ISA bridge.
1.3.6 Flash ROM
The reference design uses an AMD AM29F040-120 Flasht ROM to contain the POST and
boot code. It is recommended that Vital Product Data (VPD) such as the motherboard
speed and native I/O complement be programmed into in this device. It is possible to program the Flash before or during the manufacturing process.
After power on, the initial code fetched is supplied from this device. The 660 Bridge manages ROM access and control. The reference design supports a 512K Flash.
1.3.7 ISA Bus
The ISA bridge function is provided by an Intel 82378ZB chip (SIO). It provides a PCI to
ISA bus bridge where the native I/O and the ISA slots reside, and it provides system services such as ISA bus DMA, PCI bus arbitration, and interrupt control.
1.3.8 Time of Day Clock
The reference design uses a Dallas Semiconductort DS1385S to provide the real time
clock (TOD or RTC) function. This device is PC compatible and resides on the X-bus. It features an additional 4K of NVRAM and a replaceable battery.
1.3.9 PS/2t Compatible Keyboard/Mouse Controller
The reference design uses an Intel 8042AH as a keyboard and mouse controller.
The code used is the same version as used in IBM Personal System/2 machines. This microcode may differ from other 8042 type keyboard controllers.
1.3.10 System Clocks
The primary clock generation is accomplished with a Motorolat MPC970 PLL clock generator, which uses a seed oscillator to generate the CPU and PCI clocks needed by the system.
MPRH01TSU-02

21

Introduction
1.3.11 System I/O EPLD
The system I/O EPLD is a programmable logic device that uses the X-bus signals and the
partial decode signals from the SIO to decode chip selects for various components.
1.3.12 Power Management
Power management hardware is included on the board; however, power managenent capability will be implemented at a later date and will be described in a separate document.

1.4

Quickstart Peripheral List

The reference design is intended for typical PC peripherals. Products from a large number
of manufacturers should work satisfactorily (see Table 1 for a list of peripherals and materials). Reference boards do not come with the all of the required peripherals, cables, speaker, indicator LEDs, switches, and such that are needed to configure a properly working system.
Table 1 outlines the generic requirements for peripherals and gives examples of some devices that have been used for testing. It is not a recommendation of any particular vendor.
The purpose of this table is to outline at least one set of peripherals that may be used to
begin testing.
Table 1 does not include cables for a parallel port, indicators, a switch, or a speaker.
An IBM 3101 asynchronous terminal or equivalent is required for testing with the bring up
driver (BUD) code. Settings are 8-bit, no parity, one stop bit, and 9600 baud. VT100 or VT52
emulator terminals may be acceptable. It is desirable to also have a video monitor for BUD
tests. The boot code will boot with either an async console, a video on motherboard, or
both.

1.5

Reference Design Level

This documentation supports the release 2.1 version of the 603/604 reference design and
reference board. Except as noted, the information herein is believed to be correct for the
release 3.0 version of the reference design, once all of the errata are cured (see section
13, Errata for the reference design roadmap).
The reference board schematics are of the release 2.0 reference board, and correctly show
the workarounds that are installed on that board in order to work around certain errata.

22

MPRH01TSU-02

Introduction
Table 1. Quickstart Peripheral List
Generic Description

Example Device

L2 SRAM card, 256KB

Alliance Semiconductor AS7M64P3256–15C

Loctite 384 adhesive, 300ml cartridge

Loctite Corp. 17041

Loctite 384 activator, 10 liter can

Loctite Corp. 17101

Video adapter card, PCI S3

Diamond Stealth (S3) 864

SCSI adaptor card, PCI

NCR 8100S with 609–039–1635 controller

Super I/O adaptor ( IDE, floppy, serial
ports, parallel ports, etc.)

Acculogic sIDE–4/HP (110–00139–00E00)

Audio adaptor card

Creative Labs Soundblaster 16

Floppy disk drive, 3.5”x1.44MB

Alps DFR723F, IBM 73G4514, Mitsubishi MF355F–258UG

Hard disk drive, SCSI–2, 8 bit

Quantum LPS270/5405, Maxtor MXT–540SL, IBM WDS–3200 (79F4042)

Hard disk drive, SCSI 1GB

IBM 94G3187

Hard disk drive, IDE 1GB

Maxtor HDMC71260AC

CD_ROM drive, internal SCSI

Toshiba XM–4101BMY

CD_ROM drive, internal SCSI, 4x

Toshiba 5301–4x

CD_ROM drive, internal IDE, 4x

Chinnon CDS5451

Chassis, Baby AT

Olsen Metal Products CC300249–17

Power supply, 200W Energy Star

API-3186S, IBM 06H2968

Box fan

Panaflow FBA08T12M

Box fan shock mounts

IBM 81F7977

Internal cables, floppy, SCSI, and CD– Standard cables
ROM
Speaker, internal 8W .5W 2pin
LED, 2.5 ma drive
Asynchronous terminal

IBM 3101

Super VGA monitor

IBM 6324, 6325, 6327, 9524, 9525, 9527, 9521

Keyboard, PS/2 compatible
Mouse, PS/2 compatible

MPRH01TSU-02

23

Introduction

24

MPRH01TSU-02

CPU

Section 2
CPU and CPU Bus
This section discusses topics that are directly related to the CPU, including how the 660
bridge decodes CPU initiated transfers as a function of the transfer type and address range.
For more information, refer to the 660 Bridge User’s Manual.
The reference design supports CPU bus speeds up to 66MHz, and PCI bus speeds up to
33MHz. The reference design is initially configured with a CPU:PCI bus speed ratio of 2:1.
The CPU:PCI clock ratio can be changed to 1:1 or 3:1 by reconfiguring the clock generator
and the 660 bridge, as long as other system considerations are handled correctly.

2.1

CPU Bus Masters

The reference design uses a single CPU, of either the PowerPC 603e or PowerPC 604 family, and an external L2 cache is not allowed. Thus there are only two bus masters on the
CPU bus, the CPU and the 660 bridge. CPU bus arbitration is greatly simplified, and the
multi-processor capabilities of the 660 bridge are not used. The remaining arbitration on
the CPU bus is between the CPU and the snoop broadcasting logic in the 660 bridge. Since
the 660 bridge parks the CPU bus on CPU1 whenever the bus is idle, CPU latency is minimized.
One level of address bus pipelining is supported, and most data writes are posted. Precise
exceptions are reported via TEA#, and imprecise exceptions are reported via MCP#. PIO,
or programmed I/O transactions (XATS# type) are not supported.
2.1.1 603e CPU
The 603e version of the reference design operates the 603e in 64 bit data bus mode.
The reference design is initially configured for DRTRY# mode, and can be reconfigured to
no-DRTRY# mode by populating R440 = 0 ohm. In DRTRY# mode, data is assumed to
have been speculatively presented to the CPU, and so is held for one clock in an internal
CPU latch before being presented to the CPU data consumers. In no-DRTRY# mode, the
data is assumed to be good when TA# is sampled active, so it is immediately forwarded
without the delay cycle. Use of this mode mainly speeds up reads from the L2.
The 603e version of the reference design runs at 3:2 603e internal clock to bus clock ratio
at 99MHz:66MHz. CPU PLL_CFG[0:3] is set to 1100.
On the 603e version of the reference board, the following are populated: R425 and R426
with 1k ohm resistors, and R423 and R424 with with 10k ohm resistors. The following are
not populated: R421, R422, R427, and R428, R377.
MPRH01TSU-02

25

CPU
2.1.2 604 CPU
The reference design is initially configured for DRTRY# mode, and can be reconfigured to
no-DRTRY# mode by populating R440 = 0 ohm.
The 604 version of the reference design is initially configured to run at 2:1 604 internal clock
to bus clock ratio at 132MHz:66MHz. CPU PLL_CFG[0:3] is set to 0100. The 604 can be
configured to run with different internal to bus clock ratios as described in the 604 User’s
Manual.
The clock generator can be configured to produce different frequencies. Some examples
are shown in Table 2.
Table 2. Example Clock Generator Frequencies
CPU Internal Clock (MHz)

CPU Bus Clock (MHz)

PCI Bus Clock (MHz)

Crystal Y2A (MHz)

100

50

25

12.5

66

33

16.5

120

60

30

15

132

66

33

16.5

On the 604 version of the reference board, the following are populated: R423 (10k) and
R425, R426, and R428 with with 1k ohm resistors, and R377 with 0 ohm. The following are
not populated: R421, R422, R424, and R427.

2.2

System Response by CPU Bus Transfer Type

All access to the rest of the system is provided to the CPU by the 660 Bridge. Table 3 shows
the 660 Bridge decoding of CPU bus transfer types. Based on TT[0:3], the 660 Bridge responds to CPU bus master cycles by generating a read transaction, a write transaction, or
an address-only response. The 660 Bridge ignores TT[4] when it evaluates the transfer
type.
The bridge decodes the target of the transaction based on the address range of the transfer
as shown in Table 4. The transfer type decoding shown in Table 3 combines with the target
decoding to produce the following:
S
S
S
S
S
S
S

System memory reads and writes
PCI I/O reads and writes
PCI configuration reads and writes
PCI interrupt acknowledge reads
PCI memory reads and writes
System ROM reads and writes
Various bridge control register (BCR) reads and writes.

26

MPRH01TSU-02

CPU
Table 3. TT[0:3] (Transfer Type) Decoding by 660 Bridge
60X Operation

60X Bus
Transaction

0000

Clean block or lwarx

Address only Asserts AACK#. No other response. No PCI transaction.

0001

Write with flush

SBW(1)
or burst

0010

Flush block or stwcx

Address only Asserts AACK#. No other response. No PCI transaction.

0011

Write with kill

SBW or burst Memory write operation. L2 invalidates PCI write transaction.
addressed block.

0100

sync or tlbsync

Address only Asserts AACK#. No other response. No PCI transaction.

0101

Read or read with no
intent to cache

SBR(1)
or burst

0110

Kill block or icbi

Address only Asserts AACK#.
addressed block.

0111

Read with intent to
modify

Burst

1000

eieio

Address only Asserts AACK#. No other response. No PCI transaction.

1001

Write with flush atomic,
stwcx

SBW

Memory write operation.

1010

ecowx

SBW

Asserts AACK# and TA# if the transaction is not claimed by another 60X bus
device. No PCI transaction. No other response.

1011

Reserved

1100

TLB invalidate

Address only Asserts AACK#. No other response. No PCI transaction.

1101

Read atomic, lwarx

SBR or burst

1110

External control in,
eciwx

Address only 660 asserts all ones on the CPU data bus. Asserts AACK# and TA# if the
transaction is not claimed by another 60X bus device. No PCI transaction. No other
response.

1111

Read with intent to
modify atomic, stwcx

Burst

TT[0:3]

660 Bridge Operation For CPU to
Memory Transfers

Memory write operation.

Memory read operation.
L2

Memory read operation.

660 Bridge Operation For CPU to
PCI Transactions

PCI write transaction.

PCI read transaction.
invalidates Asserts AACK#. No other response.
PCI read transaction.

PCI write transaction.

Asserts AACK#. No other response. No PCI transaction.

Memory read operation.

Memory read operation.

PCI read transaction.

PCI read transaction.

Note:
(1)As used in this table, SBR means Single-Beat Read, and SBW means Single-Beat
Write.

Transfer types in Table 3 that have the same response are handled identically by the
bridge. For example, if the address is the same, the bridge generates the same memory
read transaction for transfer types 0101, 0111, 1101, and 1111.
The 660 Bridge does not generate PCI or system memory transactions in response to address only transfers. The bridge does drive all-ones onto the CPU bus and signals TA# during an eciwx if no other CPU bus agent claims the transfer.
References in the remainder of this document to a CPU read, assume one of the transfer
types in Table 3 that produce the read response from the 660 bridge. Likewise, references
to a CPU write refer to those transfer type that produce the write response.

MPRH01TSU-02

27

CPU

2.3

System Response by CPU Bus Address Range

The 660 Bridge determines the target of a CPU bus master transaction based on the CPU
bus address range as shown in Table 4. The acronym BCR means bridge control register.
Table 4. 660 Bridge Address Mapping of CPU Bus Transactions
CPU Bus Address

Other
Conditions

Target Transaction

Target Bus Address

Notes

System Memory

0 to 2G
0000 0000h to 7FFF FFFFh

(1)(2)

PCI I/O Transaction, BCR Transaction,
or
PCI Configuration
(Type 1) Transaction

0 to 8M
0000 0000h to 007F FFFFh

(3)

0 to 64K
0000 0000h to 0000 FFFFh

(4)

2G + 8M to 2G + 16M
8080 0000h to 80FF FFFFh

PCI Configuration
(Type 0) Transaction

PCI Configuration Space
0080 0000h to 00FF FFFFh

2G + 16M to 3G – 8M
8100 0000h to BF7F FFFFh

PCI I/O Transaction

16M to 1G – 8M
0100 0000h to 3F7F FFFFh

3G – 8M to 3G
BF80 0000h to BFFF FFFFh

BCR Transactions
and PCI Interrupt
Ack. Transactions

1G – 8M to 1G
3F80 0000h – 3FFF FFFFh

3G to 4G – 2M
C000 0000h to FFDF FFFFh

PCI Memory
Transaction

0 to 1G – 2M
0000 0000h to 3FDF FFFFh

Direct Attach
ROM Read,
Write, or Write
Lockout

BCR Transaction

0 to 2M
0000 0000h to 001F FFFFh
(ROM Address Space)

(2)

Remote ROM

PCI Memory Transaction to I/O Bus
Bridge

1G – 2M to 1G
3FE0 0000h to 3FFF FFFFh

(2)

0 to 2G
0000 0000h to 7FFF FFFFh
2G to 2G + 8M
8000 0000h to 807F FFFFh

Contiguous
Mode
Non-Contiguous
Mode

4G – 2M to 4G
FFE0 0000h to FFFF FFFFh

(3)(6)

Notes for Table 4:
1.
2.
3.
4.

System memory can be cached. Addresses from 2G to 4G are not cacheable.
Memory does not occupy the entire address space.
Registers do not occupy the entire address space.
Each 4K page in the 8M CPU bus address range maps to 32 bytes in PCI I/O
space.
5. Registers and memory do not occupy the entire address space. Accesses to unoccupied addresses result in all one-bits on reads and no-ops on writes.
6. A memory read of BFFF FFF0h generates an interrupt acknowledge transaction
on the PCI bus.
2.3.1 Address Mapping for Non-Contiguous I/O
Figure 2 shows the address mapping that the 660 Bridge performs in non-contiguous
mode. The I/O map type register (address 8000 0850h) and the bridge chip set options 1
register (index BAh) control the selection of contiguous and non-contiguous I/O. In noncontiguous mode, the 8M address space of the 60X bus is compressed into 64K of PCI address space, and the 60X CPU cannot create PCI I/O addresses from 64K to 8M.
In non-contiguous I/O mode, the 660 Bridge partitions the address space so that each 4K
page is remapped into a 32-byte section of the 0 to 64K ISA port address space, so that
60X CPU protection attributes can be assigned to any of the 4K pages. This provides a flex28

MPRH01TSU-02

CPU
ible mechanism to lock the I/O address space from change by user-state code. This partitioning spreads the ISA I/O address locations over 8M of CPU address space.
In non-contiguous mode, the first 32 bytes of a 4K page are mapped to a 32-byte space
in the PCI address space. The remainder of the addresses in the 4K page are mirrors into
the the same 32-byte PCI space. Each of the 32 contiguous port addresses in each 4K page
has the same protection attributes in the CPU.
MSB

00 =
01 =
02 =
03 =
04 =
05 =
06 =
07 =
08 =
09
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31

6
0
X
C
P
U
A
d
d
r
e
s
s
B
u
s

LSB

1
0
0
0
0
0
0
0
0

Forced to zero

Discarded

31
30
29
28
27
26
25
24
23
22
21
20
19
18
17
16
15
14
13
12
11
10
09
08
07
06
05
04
03
02
01
00

MSB

P
C
I
I
/
O
A
d
d
r
e
s
s

LSB

The three least-significant address bits are unmunged during the transformation if little-endian mode is
selected.

Figure 2. Non-Contiguous PCI I/O Address Transformation

For example in Figure 3, 60X CPU addresses 8000 0000h to 8000 001Fh are converted
to PCI I/O port 0000h through 001Fh. PCI I/O port 0020h starts in the next 4K page at 60X
CPU address 8000 1000h.
ISA I/O

4K Page

4K Page

0000
0001
0002
.
.

60X Address

001E
001F

8000 0000
8000 0001
8000 0002
.
.
.
8000 001E
8000 001F

0020
0021
.
.

8000 1000
8000 1001
.
.

60X Addresses
8000 0020 to 8000 0FFF
Are Wrapped and Should
Not Be Used.

Figure 3. Non-Contiguous PCI I/O Address Translation

MPRH01TSU-02

29

CPU
2.3.2 Address Mapping for Contiguous I/O
In contiguous I/O mode, CPU addresses from 2G to 2G + 8M generate a PCI I/O cycle on
the PCI bus with PCI_AD[29:00] unchanged. The low 64K of PCI I/O addresses are forwarded to the ISA bus unless claimed by a PCI agent.
Memory page protection attributes may only be assigned by 4K groups of ports, rather than
by 32-port groups as in the non-contiguous mode. This is the power-on default mode.
Figure 4 gives an example of contiguous I/O partitioning.
ISA I/O

60X Address

0000
0001
0002
.
.
.
001E
001F

8000 0000
8000 0001
8000 0002
.
.
.
8000 001E
8000 001F
Contiguous 603/604
addresses (No gaps)

0020
0021
.
.

8000 0020
8000 0021
.
.

.
.
FFFE
FFFF

.
.
8000 FFFE
8000 FFFF

Figure 4. Contiguous PCI I/O Address Translation

2.3.3 PCI Final Address Formation
The 660 Bridge maps 60X CPU bus addresses from 2G to 4G as PCI transactions, error
address register reads, or ROM reads and writes. The 660 Bridge manipulates 60X bus
addresses from 2G to 4G to generate PCI addresses as follows:
S PCI_AD[31:30] are set to zero.
S PCI_AD[2:0] are unmunged if little-endian mode is selected.
S After unmunging, PCI_AD[1:0] are set to 00b except for PCI I/O cycles.

2.4

CPU to Memory Transfers

The system memory address space is from 0 to 2G. Physical memory does not occupy the
entire address space. When the CPU reads an unpopulated location, the 660 Bridge returns all-ones and completes the transfer normally. When the CPU writes to an unpopulated location, the Bridge signals normal transfer completion to the CPU but does not write
the data to memory. The memory select error bit in the error status 1 register (bit 5 in index
C1h) is set in both cases.
All CPU to memory writes are posted and can be pipelined.
The 660 Bridge supports all CPU to memory bursts, and all single-beat transfer sizes and
alignments that do not cross an 8-byte boundary, which includes all memory transfers initiated by the 603/604 CPU.
30

MPRH01TSU-02

CPU
2.4.1 LE Mode
The bridge supports all transfer sizes and alignments that the CPU can create in LE mode;
however, all loads or stores must be at natural alignments in LE mode (or the CPU will take
an alignment exception). Also, load/store multiple word and load/store string word instructions are not supported in the CPU in LE mode.

2.5

CPU to PCI Transactions

Since all CPU to PCI transactions are CPU memory mapped, software must in general utilize the EIEIO instruction which enforces in-order execution, particularly on PCI I/O and
configuration transactions. Some PCI memory operations can be sensitive to order of access also. See the 660 Bridge User’s Manual.
All addresses from 2G to 4G (including ROM space) must be marked non-cacheable. See
the PowerPC Reference Platform Specification. The reference design supports all PCI bus
protocols during CPU to PCI transactions.
The reference design supports all CPU to PCI transfer sizes that do not cross a 4-byte
boundary, and, to support the 604 store multiple instruction, the reference design supports
8-byte CPU to PCI writes that are aligned on an 8-byte boundary. The bridge does not support CPU bursts to the PCI bus.
When the 660 bridge decodes a CPU access as targeted for the PCI, the 660 bridge requests the PCI bus. Once the SIO grants the PCI bus to the 660 bridge, the bridge initiates
the PCI cycle and releases the bus.
CPU to PCI transactions that the PCI target retries, cause the 660 Bridge to deassert its
PCI_REQ# (the Bridge follows the PCI retry protocol). The Bridge stays off of the PCI bus
for two PCI clocks before reasserting PCI_REQ# (or FRAME#, if the PCI bus is idle and
the PCI_GNT# to the Bridge is active).
2.5.1 CPU to PCI Read
If the CPU to PCI cycle is a read, a PCI read cycle is run. If the PCI read cycle completes,
the data is passed to the CPU and the CPU cycle is ended. If the PCI cycle is retried, the
CPU cycle is retried. If a PCI master access to system memory is detected before the PCI
read cycle is run then the CPU cycle is retried (and no PCI cycle is generated).
2.5.2 CPU to PCI Write
If the CPU to PCI cycle is a write, a PCI write cycle is run. CPU to PCI I/O writes are not
posted, as per the PCI Local Bus Specification version 2.1. If the PCI transaction is retried,
the Bridge retries the CPU.
CPU to PCI memory writes are posted, so the CPU write cycle is ended as soon as the data
is latched. If the PCI cycle is retried, the Bridge retries the cycle until it completes.
2.5.2.1
Eight-Byte Writes to the PCI (Memory and I/O)
The 660 Bridge supports 1-byte, 2-byte, 3-byte, and 4-byte transfers to and from the PCI.
The 660 Bridge also supports 8-byte memory and I/O writes (writes only, not reads) to the
PCI bus. This enables the use of the 604 store multiple instruction to PCI devices.When
an 8-byte write to the PCI is detected, it is not posted initially. Instead the CPU waits until
the first 4-byte write occurs, then the second 4-byte write is posted. If the PCI retries on
MPRH01TSU-02

31

CPU
the first four byte transfer or a PCI master access to system memory is detected before the
first 4-byte transfer then the CPU is retried. If the PCI retries on the second 4-byte transfer
then the 660 Bridge retries the PCI write.
2.5.3 CPU to PCI Memory Transactions
CPU transfers from 3G to 4G–2M are mapped to the PCI bus as memory transactions.
2.5.4 CPU to PCI I/O Transactions
CPU transfers from 2G+16M to 3G–8M are mapped to the PCI bus as I/O transactions. In
compliance with the PCI specification, the 660 Bridge master aborts all I/O transactions that
are not claimed by a PCI agent.
2.5.5 CPU to PCI Configuration Transactions
The reference design allows the CPU to generate type 0 and type 1 PCI configuration
cycles. The CPU initiates a transfer to the appropriate address, the 660 bridge decodes the
cycle and generates a request to the PCI arbiter in the SIO. When the PCI bus is acquired,
the 660 bridge enables its PCI_AD drivers and drives the address onto the PCI_AD lines
for one PCI clock before it asserts PCI_FRAME#. Predriving the PCI_AD lines for one clock
before asserting PCI_FRAME# allows the IDSELs to be resistively connected to the
PCI_AD[31:0] bus at the system level.
The transfer size must match the capabilities of the target PCI device for configuration
cycles. The reference design supports 1-, 2-, 3-, and 4-byte transfers that do not cross a
4-byte boundary.
Address unmunging and data byte swapping follows the same rules as for system memory
with respect to BE and LE modes of operation. Address unmunging has no effect on the
CPU address lines which correspond to the IDSEL inputs of the PCI devices.
2.5.5.1
Preferred Method of Generating PCI Configuration Transactions
The preferred method for generating PCI configuration cycles is via the 660 Bridge indexed
Bridge Control Registers (BCR). This configuration method is described in section 4 of the
660 User’s Manual. The IDSEL assignment and their respective PCI_AD lines are shown
in Table 5. The addresses used for configuration are assigned as shown in Table 5.
2.5.5.2
650 Bridge compatible method
If it is not possible to use indexed BCRs to generate PCI configuration cycles, they can be
generated by an alternate method known as the 650 bridge compatible method. CPU accesses to the address range 2G+8M to 2G+16M cause the bridge to arbitrate for the PCI
bus and then to execute a type 0 PCI configuration transaction as described in the PowerPC Reference Platform Specification and implemented by the IBM27–82650 PowerPC
to PCI Bridge. This is referred to as the 650 compatible configuration method. This method
of accessing PCI configuration space does not allow access to the PCI configuration registers in the bridge chip, and it should not be used unless required to maintain 650
compatibility.
When using the 650 bridge compatible configuration method, use only the specified address. Using certain other addresses could cause bus contention because multiple PCI
slots could be selected. For example, using any CPU address in the range 8080 0000 to
80FF FFFF with both AD11 = 1 and AD12 = 1 causes selection of both the SIO and any
device in slot 1, possibly resulting in damage.
32

MPRH01TSU-02

CPU

Device

Table 5. 660 Bridge Address Mapping of CPU Bus Transactions
IDSEL Line
60X Address*

PCI Address

Intel SIO 82378ZB

A/D 11

8080 08XXh

080 08XX

PCI slot 1

A/D 12

8080 10XXh

080 10XX

PCI slot 2

A/D 13

8080 20XXh

080 20XX

PCI slot 3

A/D 14

8080 40XXh

080 40XX

Reserved config scan

A/D 18

8084 00XXh

084 00XX

Note: *This address is independent of contiguous I/O mode.

2.5.6 CPU to PCI Interrupt Acknowledge Transaction
Reading the interrupt acknowledge address (BFFF FFF0h) causes the bridge to arbitrate
for the PCI bus and then to execute a standard PCI interrupt acknowledge transaction. The
system interrupt controller in the ISA bridge claims the transaction and supplies the 1-byte
interrupt vector. There is no physical interrupt vector BCR in the bridge. Other PCI bus masters can initiate interrupt acknowledge transactions, but this may have unpredictable effects.
2.5.7 PCI Lock
The 660 Bridge does not set PCI locks when acting as the PCI master. The PCI_LOCK#
signal in the 660 Bridge supports resource locking of one 32-byte cache sector (block) of
system memory. Once a PCI lock is established, the block address is saved. Subsequent
accesses to that block from other PCI bus masters or from the CPU bus are retried until
the lock is released.
The bridge generates a flush-sector snoop cycle on the CPU bus when a PCI bus master
sets the PCI lock. The flush-sector snoop cycle causes the L1 and L2 caches to invalidate
the locked block, which prevents cache hits on accesses to locked blocks. If the L1 contains
modified data, the PCI cycle is retried and the modified data is pushed out to memory.
Note: The 60X processors do not have bus-locking functions. Instead, they use the load
reserve and store conditional instructions (lwarx and stwcx) to implement exclusive access.
To work with the lwarx and stwcx instructions, the 660 Bridge generates a flush-sector operation to the CPU in response to the PCI read that begins a PCI lock.

2.6

CPU to ROM Transfers

The PowerPC Reference Platform Specification allocates the upper 8M of the 4G CPU address space as ROM space. The reference design implements a 2M ROM space from
4G–2M to 4G. The actual ROM is a 512K device located at 4G–2M. The ROM is attached
to the 660 bridge via the PCI_AD lines. This mode is required when using the Intel SIO.
ROM device writes and write-protect commands are supported. See the 660 Bridge User’s
Manual for more information.
The ROM device attaches to the 660 bridge by means of control lines and the PCI_AD[31:0]
lines. When a CPU bus master reads from the ROM, the bridge masters a BCR transaction,
during which it reads the ROM and returns the data to the CPU. CPU writes to the ROM
and ROM write-protection operations are also forwarded to the ROM device.
Although connected to the PCI_AD lines, the ROM is not a PCI agent. The ROM and the
PCI agents do not interfere with each other because the ROM is under bridge control, and
MPRH01TSU-02

33

CPU
the bridge does not enable the ROM except during ROM cycles. The bridge accesses the
ROM by means of BCR transactions. Other PCI devices cannot read or write the ROM because they cannot generate BCR transactions.
2.6.1 CPU to ROM Read
At power-on, the 603/604 CPU comes up in BE Mode with the L1 cache disabled, and begins fetching instructions (using 8-byte single beat reads) at address FFF0 0100 (4G – 1M
+ 100h). The reference design logic also resets to BE mode.
The system ROM address space is from 4G – 2M to 4G. Since the size of the installed ROM
is less than 2M (512K), it is mirrored every 512K throughout the ROM space. Location 0
of the 512K ROM is mapped to CPU bus addresses 4G–2M, 4G–1.5M, 4G–1M, and
4G–.5M.
The Flash is located on the PCI bus physically but not logically, and is 8 bits wide. This requires the 660 Bridge to decode Flash address, run 8 cycles to PCI bus without activating
FRAME, accumulate the 8 single bytes of read data into an 8-byte group and generate a
TA# and an AACK# to complete the cycle. The CPU can also read the ROM using bursts,
but it receives the same 2 instructions from the ROM on each beat of the burst. For more
information, see the 660 Bridge User’s Manual.
Software can lock out the ROM using a 660 bridge BCR. When the CPU writes to any ROM
location while the ROM is locked out, the bridge signals normal transfer completion to the
CPU but does not write the data to the ROM. The CPU bus write to the locked flash bit in
the 660 bridge error status 2 register (bit 0 in index C5h) is set.
2.6.2 CPU to ROM Write
Writing to Flash is another very specialized cycle. Only one address (FFFF FFF0) is used
for writing data to Flash. The Flash address and data are both encoded into four bytes and
written using a 4-byte write transfer. Eight byte and burst transfers to the ROM are not supported. See the 660 Bridge User’s Manual.
Writes to Flash may be performed in either BE or LE mode. The data byte swapper in the
660 Bridge is gated according to endian mode. Writes in BE mode occur in natural sequence. However, address unmunging in LE mode has no effect on the cycle because the
addresses are ignored. Therefore, software must reverse the byte significance of the data
and address encoded into the store instructions for LE mode writes to the ROM.
2.6.2.1
ROM Write Protection
Flash write protection must be implemented within software. Port FFFF FFF1 can be used
to lock out all Flash writes. Writing any data to this port address locks out all Flash writes
until the 660 Bridge is hardware reset. In addition, the Flash itself has means to permanently lock out changing certain sectors by writing control sequences. Consult the Flash Specification for details.
2.6.3 CPU to BCR Transfers
The 660 Bridge can be extensively programmed by means of the Bridge Control Registers
(BCR). See the 660 Bridge User’s Manual for a description of the operation and programming of the 660 bridge BCRs.
34

MPRH01TSU-02

PCI Bus

Section 3
PCI Bus
The reference design includes a 32-bit PCI bus at frequencies up to 33MHz. The PCI bus
is compatible with the PCI Specification, revisions 2.0 and 2.1. The power supply voltage
is 3.3v, and the reference design components have 5v tolerant I/O. Up to three PCI cards
may be added to the system, which provides full hardware and software support for PCI
2.0 and 2.1 compliant PCI agents.
The PCI bus can be run an one-third, one-half, or the same frequency as the 60X CPU bus,
and is initially configured to run at one half of the CPU bus speed.
PCI bus activity initiated by the CPU is discussed in section 2. This section describes PCI
bus transactions initiated by a (non-660-bridge) PCI bus master.

3.1

PCI Transaction Decoding

When a PCI bus master initiates a transaction on the PCI bus, the transaction either misses
and is master aborted, or it is claimed by a PCI target. The target can be either the 660
bridge or another PCI target.

MPRH01TSU-02

35

PCI Bus
3.1.1 PCI Transaction Decoding By Bus Command
Table 6 shows the responses of the 660 bridge and other agents to various PCI bus transactions initiated by a PCI bus master other than the 660 bridge. As shown in Table 6, the
660 bridge ignores (No response) all PCI bus transactions except PCI memory read and
write transactions, which it decodes as possible system memory accesses.
Table 6. Reference Design Responses to PCI_C[3:0] Bus Commands
C[3:0]

PCI Bus Command

Can a PCI Bus Master
Initiate this Transaction?

660 Bridge Response to
the Transaction

Can Another PCI Target
Claim the Transaction?

0000

Interrupt Acknowledge

No. Only the 660 Bridge is
allowed to initiate.

None

Yes. The ISA bridge is
intended to be the target.

0001

Special Cycle

Yes

None

Yes

0010

I/O Read

Yes

None

Yes

0011

I/O Write

Yes

None

Yes

0100

Reserved

No. Reserved

None

n/a

0101

Reserved

No. Reserved

None

n/a

0110

Memory Read

Yes

System memory read

Yes, if no address conflict.

0111

Memory Write

Yes

System memory write.

Yes, if no address conflict.

1000

Reserved

No. Reserved

None

n/a

1001

Reserved

No. Reserved

None

n/a

1010

Configuration Read

No. Only 660 Bridge.

None

Yes

1011

Configuration Write

No. Only 660 Bridge.

None

Yes

1100

Memory Read Multiple

Yes

System memory read

Yes, if no address conflict.

1101

Dual Address Cycle

Yes

None

Yes

1110

Memory Read Line

Yes

System memory read

Yes, if no address conflict.

1111

Memory Write and Invalidate

Yes

System memory write

Yes, if no address conflict.

3.1.2 PCI Memory Transaction Decoding By Address Range
When a PCI bus master transaction is decoded by bus command as a system memory read
or write, the 660 bridge checks the address range. Table 7 shows the mapping of PCI bus
master memory accesses to system memory. This is the mapping that the 660 bridge uses
when it decodes the bus command to indicate a system memory access.
Table 7. Mapping of PCI Memory Space, Part 1
PCI Bus Address

Other Conditions

Target Cycle Decoded

Target Address

Notes

0 to 2G

IGN_PCI_AD31
Deasserted

Not Decoded

N/A

No Response.

IGN_PCI_AD31
Asserted

System Memory *

0 to 2G

Snooped by caches.

System Memory *

0 to 2G

Snooped by caches.

2G to 4G

Note:
*Memory does not occupy this entire address space. Accesses to unoccupied space
are not decoded.
Unless the IGN_PCI_AD31 signal is asserted, PCI memory accesses in the 0 to 2G address range are ignored by the 660 Bridge. There is no system memory access, no snoop
36

MPRH01TSU-02

PCI Bus
cycle, and the 660 bridge does not claim the transaction. When the IGN_PCI_AD31 signal
is asserted, the 660 Bridge maps PCI memory accesses from 0 to 2G directly to system
memory at 0 to 2G. PCI memory accesses from 2G to 4G are mapped to system memory
from 0 to 2G.
PCI memory access that are mapped to system memory cause the 660 bridge to claim the
transaction, access system memory, and arbitrate for the CPU bus and broadcast a snoop
operation on the CPU bus. A detailed description of the snoop process is presented in the
660 Bridge User’s Manual.
Table 8 gives a more detailed breakdown of the reference design response to PCI memory
transactions in the 0 to 2G range. Note that the preferred mapping of PCI memory, so that
it can be accessed both by the CPU and by PCI bus masters, is from 16M to 1G–2M.
Table 8. Mapping of PCI Memory Space, Part 2
PCI Bus Address

Target Resource

System Memory Address

Snoop Address
0 to 2G

2G to 4G

System memory (1)

0 to 2G

1G–2M to 2G

Reserved (2)

16M to 1G–2M

PCI Memory

0 to 16M

PCI/ISA Memory (3)

No system memory access. No snoop.
The 660 bridge ignores PCI
memory transactions in this
range.

Notes:
1) The 660 bridge maps PCI bus master memory transactions in the 2G to 4G range to system memory, and the CPU is unable to initiate PCI memory transactions to this address
range, so do not map devices to this PCI memory address range.
2) The CPU (thru the 660 bridge) can not access the 1G–2M to 2G address range, so do
not map PCI devices herein unless the CPU will not access them.
3) Transactions initiated on the PCI bus by the ISA bridge on behalf of an ISA bus master
only (IGN_PCI_AD31 asserted for an SIO), are forwarded to system memory and broadcast snooped to the CPU bus from 0 to 16M. If this is not an ISA bus master transaction,
then the 660 bridge ignores it. Note that the 660 bridge will also forward PCI transactions
from 16M to 2G if IGN_PCI_AD31 is asserted during an ISA-bridge-mastered transaction, and that this capability is not normally used.

MPRH01TSU-02

37

PCI Bus
3.1.3 PCI I/O Transaction Decoding
The 660 Bridge initiates PCI I/O transactions on behalf of the CPU. Other PCI bus masters
are also allowed to initiate PCI I/O transactions. Table 9 shows the reference design mapping of PCI I/O transactions. The 660 bridge ignores PCI I/O transactions.
PCI/ISA I/O is mapped to PCI I/O space from 0 to 64K. The ISA bridge subtractively decodes these transactions (and also PCI memory transactions from 0 to 16M). Other devices
may actively decode and claim these transactions without contention.
PCI I/O is assigned from 16M to 1G–8M.
Table 9. Mapping of PCI Master I/O Transactions
PCI Bus Address

Target Resource

1G–8M to 4G

Reserved (1)

16M to 1G–8M

PCI I/O devices

8M to 16M

Reserved (1)

64K to 8M

Reserved (2)

0 to 64K

PCI/ISA I/O

Other System Activity
The 660 Bridge ignores I/O transactions initiated by PCI
bus masters.

Notes:
1) The CPU (thru the 660 bridge) can not access this address range, so do not map PCI
devices herein unless the CPU will not access them.
2) In contiguous mode, the CPU (thru the 660 bridge) can create PCI I/O addresses in the
64K to 8M range. In non–contiguous mode, the CPU can only access PCI addresses
from 0 to 64K.
3.1.4 ISA Master Considerations
Since the reference design implements IGN_PCI_AD31 and uses an Intel SIO, memory
transactions produced on the PCI bus by the ISA bridge on behalf of an ISA master, are
forwarded to system memory at the corresponding address (0 —16M).
If ISA masters are utilized and the SIO is programmed to forward their cycles to the PCI
bus, then no other PCI device (e.g. video) is allowed to be mapped at the same addresses
because contention would result.
The SIO chip contains registers to control which ranges of ISA addresses are forwarded
to the PCI bus.
ISA masters cannot access any PCI memory.
For more information on the handling of ISA bus master operations, see the 660 Bridge
User’s Manual and the SIO data book.

3.2

PCI Transaction Details

Details of the reference design implementation of various PCI transactions, including sequencing, timing, and interactions with the CPU bus, are found in the 660 Bridge User’s
Manual.
38

MPRH01TSU-02

PCI Bus
PCI bus masters are not able to access the boot ROM, the BCRs in the 660 bridge, or the
CPU bus.
3.2.1

Bus Snooping on PCI to Memory Cycles

Each time a PCI (or ISA) bus master accesses memory, (and once again for each time a
PCI burst crosses a cache block boundary) the 660 bridge broadcasts a snoop operation
on the CPU bus. If the CPU signals an L1 snoop hit by asserting ARTRY#, the 660 bridge
retries the PCI transaction. The ISA bridge then removes the grant from the PCI agent, who
(according to PCI protocol) releases the bus for at least one cycle and then arbitrates again.
Meanwhile, the 660 bridge grants the CPU bus to the CPU, allowing it to do a snoop push.
Then the PCI agent again initiates the original transaction.
During the transaction, the 660 bridge L2 cache is monitoring the memory addresses. The
L2 takes no action on L2 misses and read hits. If there is an L2 write hit, the L2 marks that
block as invalid, does not update the block in SRAM, and does not affect the PCI transaction. L2 operations have no effect on PCI to memory bursts.
3.2.2

PCI to PCI Peer Transactions

Peer to peer PCI transactions are supported consistent with the memory maps of Table 6,
Table 7, Table 8, and Table 9, which together show the ranges of different bus command
transactions that are supported.
3.2.3

PCI to System Memory Transactions

PCI to system memory transactions are described in detail in the 660 Bridge User’s Manual.
Single and burst transfers are supported. Bursts are supported without special software
restrictions. That is, bursts can start at any byte address and end on any byte address and
can be of arbitrary length. Also, the arbitration logic insures that the PCI does not monopolize the PCI bus.
As per the PCI specification, the byte enables are allowed to change on each data phase.
This has no practical effect on reads, but is supported on writes. The memory addresses
linearly increment by 4 on each beat of the PCI burst. All PCI devices must use only linear
burst incrementing.
Table 10 shows which CAS# lines are activated when a PCI master writes memory. Note
that CAS[0]# refers to byte addresses 0 mod 8, CAS[1]# refers to byte addresses 1 mod
8, etc.. For read cycles, eight bytes of memory data are read on each access, but the master
receives only the desired 4 bytes. The bytes are read or written to memory independently
of BE or LE mode (the endian mode byte swappers are situated between the CPU and the
rest of the system, not between the PCI and the rest of the system).
In ECC mode, PCI to memory transactions that result in less than 8-byte writes, cause the
memory controller in the 660 bridge to execute a read-modify-write operation, during which
8 bytes of memory data are read, the appropriate bytes are modified, the ECC byte is modified, and then the resulting 8 bytes are written to memory.
MPRH01TSU-02

39

PCI Bus

PCI_
AD[2]

Table 10. Active CAS# Lines – PCI to Memory Writes, BE or LE Mode
Byte Enables BE[ ]#
Column Address Selects CAS[ ]#
3

2

1

0

0

1

1

1

1

0

1

1

1

0

0

1

1

0

1

0

1

1

0

0

0

1

0

1

1

0

1

0

1

0

0

1

0

0

1

0

1

0

0

0

0

0

1

1

1

0

0

1

1

0

0

0

1

0

1

0

0

1

0

0

0

0

0

1

1

0

0

0

1

0

0

0

0

0

1

0

0

0

0

0

1

1

1

1

1

1

1

1

1

0

1

1

1

0

1

1

1

1

0

0

1

1

0

1

1

1

1

0

1

0

1

1

0

0

1

1

1

0

0

0

1

0

1

1

1

1

0

1

1

0

1

0

1

0

1

1

0

1

0

0

1

0

0

1

1

1

0

0

1

0

1

0

0

0

1

1

0

0

0

0

0

1

2

3

4

5

6

7

X
X
X

X
X

X
X

X
X

X

X

X
X

X
X

X
X

X

X

X
X

X

X

X

X

X

X

X

X

X

X
X

X
X
X

X
X

X
X

X
X

X

X

X
X

X
X

X
X

X

X

X
X

X

X

X

X

X

X

X

X

X

X
X

Note:
X = active. Blank = inactive. Byte enables would normally represent contiguous addresses. This table shows what would happen for all cases.

40

MPRH01TSU-02

PCI Bus

3.3

Bus Arbitration Logic

The reference design uses the Intel SIO as the PCI bus arbiter. The PCI arbiter sees the
660 bridge as one of several PCI agents. The order of priority for PCI arbitration is programmable, and is initially set to be:
1. 660 bridge (the SIO normally parks the bus on the 660 bridge)
2. PCI slot 1
3. PCI slot 2
4. PCI slot 3
For more information on arbitration, see section 9 on reference design initialization, and
see the PCI Arbitration Controller section of the SIO data book. See the reference design
schematics for the connection of the various PCI requests and grants.
There may be concurrency of cycles on the ISA bus (caused by DMA or ISA masters) with
PCI or CPU transactions as long as the ISA bus operations are not forwarded to the PCI
bus. Forwarding of ISA bus operations must wait for the ISA bridge to grant the PCI bus
to its ISA interface.

3.4

Other PCI Considerations

The reference design motherboard presents from 7 to 10 PCI loads to the bus, and at least
three additional PCI compliant agents/loads can be added to the PCI bus without exceeding the loading parameters.

MPRH01TSU-02

41

PCI Bus

42

MPRH01TSU-02

ISA Bus

Section 4
ISA Bus
The reference design includes an ISA bus that is interfaced to the PCI bus by the ISA bus
bridge. Five ISA expansion slots are provided by the reference design. The reference design uses a buffered subset of the ISA bus, called the X-bus, to host onboard native I/O,
such as the real time clock and the keyboard and mouse controller.

4.1

The ISA Bridge

The ISA bridge function is provided by an Intel 82378ZB chip (SIO). It provides a PCI to
ISA bus bridge, with the following major functions:
S Bridge between PCI and ISA
S 8/16 bit ISA devices
S 24 bit addressing on ISA
S Partially decodes native I/O addresses
S Unclaimed PCI memory address below 16MB forwarded to the ISA bus
S Unclaimed PCI I/O address below 64K forwarded to the ISA bus
S Powers up to an open condition (i.e., cycles may be passed to the ISA bus)
S Generates ISA clock, with a programmable divide ratio of three or four
S Allows ISA mastering and has programmable decodes that map ISA memory
cycles to the PCI bus
S 32–bit posted memory write data buffer (no I/O buffering)
S Seven channel ISA DMA controller
S Function of two 83C37s with 32–bit extensions
S Supports 8– bit or 16–bit devices on the ISA bus
S Supports 32–bit addressing for ISA to PCI memory transfers
S 8–byte bidirectional buffer for DMA data
S Timer block (function of 82C54)
S Interrupt Controller (function of two 8259s)
S PCI bus arbiter.
S Functions as PCI target during programming and ISA target cycles, and as bus master during DMA or ISA master cycles
S Generates ISA_REFRESH# signal to refresh ISA bus DRAM.
MPRH01TSU-02

43

ISA Bus

4.2

Address Ranges

The ISA bus address ranges which may be separately enabled in the ISA bus bridge for
forwarding to the PCI bus are:
0 - 512K
512K - 640K
640K - 768K
768K - 896K ( in 8 ranges of 16 K each)
896K - 960K
1M - xM (where x < or = 16) with the hole (see the SIO data book). The hole may
be 64K or 8M.
If an ISA DMA produces an address in the 0-16M range and this address is enabled in the
ISA bridge for forwarding to the PCI, the ISA bridge will initiate a PCI transaction which the
660 bridge will forward to system memory.
S
S
S
S
S
S

The 660 bridge uses medium timing when claiming ISA master originated cycles on the
PCI. It does not use subtractive decoding. Hence, ISA masters can only communicate with
other ISA devices or system memory. They may not communicate with PCI devices.
Warning: The software should not map any PCI memory at PCI addresses which ISA masters can create (those addresses between 0-16M which are programmed for forwarding
from ISA to PCI). This is because contention would result between the device mapped at
that address and the 660 bridge. Alternatively stated, ISA masters should not be allowed
to create accesses to system memory using any address between 0 and 16M that is
mapped to a PCI device, such as video.

4.3

ISA Bus Concurrency

ISA bus cycles which are not enabled for forwarding, including the hole, remain on the ISA
bus. That is, DMA or ISA bus master cycles on the ISA bus can run concurrently with PCI
or CPU cycles.

4.4

ISA Bus Masters and IGN_PCI_AD31

The ISA bridge supports ISA bus masters. System memory accesses from an ISA bus master are designed to be mapped to the 0 to 16M range, and the ISA bridge forwards them
to the PCI bus at the same range, which is not compliant to the PowerPC Reference Platform specification. Other PCI to system memory accesses however, are correctly mapped
to the 2G to 4G range (for system memory address range from 0 to 2G). In some architectures this problem is handled by using the ISA_MASTER# signal, which is active during the
ISA bus master operation.
However, the ISA bridge allows ISA masters to run posted writes to system memory, without latching in the accompanying ISA_MASTER# signal. In this situation, the ISA_MASTER# signal is no longer synchronized to the ISA bus master operation.
To overcome this challenge, the reference design detects PCI memory transactions that
are initiated by the SIO. The reference design ANDs together GNT0#, GNT1#, and GNT2#,
44

MPRH01TSU-02

ISA Bus
to generate IGN_PCI_AD31, which is active high during the address phase of any PCI
transaction that is not initiated by one of the three possible PCI agents (besides the 660
bridge and the SIO). If the PCI transaction was not initiated by the 660 bridge, and if it is
a memory transaction, then the 660 bridge assumes that it is a system memory transaction
initiated on the PCI bus by the SIO on behalf of an ISA bus master, and so forwards it to
the correct system memory address in the 0 to 16M range.
As a consequence of this design, the ISA bridge must be programmed to map ISA DMA
(that is bound for system memory) to a PCI memory transaction using the 0 to 2G address
range, rather than the apparently correct 2G to 4G range. Since the DMA sourced PCI
transaction also causes IGN_PCI_AD31 to be asserted during the address phase of a PCI
transaction initiated by the ISA bridge, the 660 bridge will not do the usual inversion of the
highest order address bit, but will forward the transaction to system memory in the 0 to 2G
range.
Another consequence of the design is that the ISA bridge can not initiate peer to peer PCI
memory transactions, because no matter what the PCI address is, it will be claimed by the
660 bridge (if the address is that of a populated memory location) and mapped to system
memory, possibly causing various inappropriate results.
These are the only limitations on the normal operation of the ISA bridge that are caused
by the IGN_PCI_AD31 design, and there are no implications for other PCI or ISA agents,
which are totally unaffected by the situation.

MPRH01TSU-02

45

ISA Bus

4.5

DMA

The DMA controller in the ISA bridge consists of the functionality of two 82C37A DMA controllers with 32-bit addressability extensions and enhanced functionality. The DMA request/
grant lines are connected on the reference design as shown in Table 11.
Table 11. DMA Assignments
DMA Channel

Assignment or Connection

0

ISA connectors

1

ISA connectors

2

ISA connectors

3

ISA connectors

4

Cascade in

5

ISA connectors

6

ISA connectors

7

ISA connectors

4.5.1 Supported DMA Paths
DMA operations can be performed only:
S From ISA I/O mapped devices to ISA memory mapped devices, and
S From ISA I/O mapped devices to system memory (via the PCI bus). In these transfers, the system memory address must be mapped to the PCI address range 0 to
2G (see section 4.4).
The DMA source device can be located on the X-bus. If the DMA target is ISA memory
mapped, it can also reside on the X-bus.
S ISA DMA to PCI I/O devices is not allowed.
S ISA DMA to PCI memory devices is not allowed.
S ISA DMA from ISA memory mapped devices is not allowed.
4.5.2 DMA Timing
The DMA controller runs compatible cycles for all ISA to ISA DMA transfers. Type A, type
B and type F timing is available only for ISA I/O to system memory (via the PCI) DMA transfers.
4.5.3 Scatter-Gather
The reference design permits the use of independent scatter-gather (SG) operations on
DMA channels 0-3 and 5-7. This operation chains together a number of DMA transfers to
different memory locations so that they appear as one DMA transfer. The SG command,
descriptor table, and status registers are relocatable via a configuration register in the ISA
bridge. The termination of an operation may be signaled to the software by configuring any
(or all) of the SG channels to use IRQ13 or by configuring the channel to signal end of process (aka Terminal Count) to the DMA device. See the SIO data book for details.
46

MPRH01TSU-02

ISA Bus

4.6

X-Bus

The X-bus is a utility bus, an 8 bit buffered version of the ISA bus, implemented on the reference board to support the native I/O devices that are located on the reference board. X-bus
data transceiver U12 is controlled by the ISA bridge via XDIR (UBUSTR) and XDEN#
(UBUSOE#). Various devices are located on the X–bus.
4.6.1 Control Signal Decodes
The System I/O EPLD (Chandra) is a programmable logic device that uses the X-bus signals and the partial decode signals from the ISA bridge to decode chip selects for various
components. For more information on the System I/O EPLD, see that data sheet.

4.6.2 Keyboard/Mouse Controller
The reference design uses an Intel 8042AH as a keyboard and mouse controller, which
resides on the X-bus. The code used is the same version as used in IBM Personal System/2
machines. This microcode may differ from other 8042 type keyboard controllers. These differences are usually only significant when porting AIX to the system (for more information
contact your IBM representative. This device contains several registers. See the data sheet
for more information.

4.6.3 Real Time Clock (RTC)
The reference design uses a Dallas Semiconductort DS1385S to provide the real time
clock (TOD or RTC) function. This device is PC compatible and resides on the X-bus. It features an additional 4K of NVRAM and a replaceable battery. This device contains several
registers. See the data sheet for more information.

4.6.4 PCI Adapter Card Presence Detect Register
The reference design uses U20 to buffer the PCI adaptor card presence detection bits onto
the X-bus under control of the system I/O EPLD. These bits report in pairs, and do not contain any information about the identity of the card. They merely report on its presence.
ISA Port

080C

Read Only

Reset n/a

MSB D7 D6 D5 D4 D3 D2 D1 D0 LSB
PCI Slot 0 Presence Detect
PCI Slot 1 Presence Detect
PCI Slot 2 Presence Detect
Reserved

Bits 7:6
Bits 5:4
Bits 3:2
Bits 1:0

Reserved
PCI Slot 2 Presence Detect Bits. 00 = Present, 11 = No PCI card installed.
PCI Slot 2 Presence Detect Bits. 00 = Present, 11 = No PCI card installed.
PCI Slot 2 Presence Detect Bits. 00 = Present, 11 = No PCI card installed.

MPRH01TSU-02

47

ISA Bus
4.6.5

L2 SRAM Identification Register

The reference design uses U19 to buffer the SRAM identification/presence detect bits from
the SRAM socket onto the X-bus under control of the system I/O EPLD.
ISA Port

080D

Read Only

Reset n/a

MSB D7 D6 D5 D4 D3 D2 D1 D0 LSB
SRAM ID[2:0]
Reserved
Table 12. DRAM Module Presence Detect Bit Encoding
SRAM Module Identification

SRAM ID [2:0]
ID3

ID2

ID1

0

0

0

512K Synchronous

0

0

1

256K Synchronous

0

1

0

Reserved

0

1

1

Reserved

1

0

0

512K Asynchronous

1

0

1

256K Asynchronous

1

1

0

1M Asynchronous

1

1

1

L2 SRAM module not present

4.6.6

Planar ID Detection Register

Revision information on the planar (motherboard) is buffered onto the X-bus by U18, under
control of the system I/O EPLD.
ISA Port

0852

Read Only

Reset n/a

MSB D7 D6 D5 D4 D3 D2 D1 D0 LSB
ID Bits[7:0]
Table 13. Planar ID Encoding
Planar
ID[7:0]

CPU

CPU Internal Clock/
CPU Bus Clock/
PCI Bus Clock

Schematic

0C

603

66/66/33

MPRH02SCU–01

0D

603e

99/66/33

MPRH02SCU–01

0E

604

132/66/33

MPRH02SCU–01

other

—

—

Reserved

48

MPRH01TSU-02

ISA Bus
4.6.7 DRAM Presence Detection
DRAM module presence and identification data is hard coded into the pinout of the SIMM
(or DIMM) by shorting particular pins to ground or no-connecting them on the SIMM itself.
The reference design uses U17 and U40 to buffer the presence detect bits from the DRAM
sockets onto the X-bus under control of signals from the system I/O EPLD. This information
appears as the DRAM SIMM 1-2 Memory ID Register and the DRAM SIMM 3-4 Memory
ID Register.

4.6.7.1

DRAM SIMM 1-2 Memory ID Register

ISA Port

0880

Read Only

Reset n/a

This register indicates the ID bits associated with SIMMs 1 and 2.
MSB D7 D6 D5 D4 D3 D2 D1 D0 LSB
ID Bits[7:0]

Bits 7:4
Bits 3:0

SIMM 2 ID bits. See Table 14.
SIMM 1 ID bits. See Table 14.

4.6.7.2

DRAM SIMM 3-4 Memory ID Register

ISA Port

0881

Read Only

This register indicates the ID bits associated with SIMMs 3 and 4.
MSB D7 D6 D5 D4 D3 D2 D1 D0 LSB
ID Bits[7:0]

Bits 7:4
Bits 3:0

SIMM 4 ID bits. See Table 14.
SIMM 3 ID bits. See Table 14.

Table 14. DRAM Module Presence Detect Bit Encoding
Presence Detect Bits

DRAM Module Identification

PD4

PD3

PD2

PD1

1

0

0

0

4 MB (1M x 36b) 70ns

1

0

1

1

8 MB (2M x 36b) 70ns

1

0

1

0

16 MB (4M x 36b) 70ns

1

1

0

1

32 MB (8M x 36b) 70ns

1

1

1

1

No DRAM module installed.

Other encodings

MPRH01TSU-02

Not currently defined.

49

Reset n/a

ISA Bus

4.7

Miscellaneous

4.7.1 Speaker Support
The reference design has a connector for a small speaker. The speaker output is driven
by the timer 2 signal from the ISA bridge. The intended speaker is a typical PC type, 8 ohms
and .5W.

50

MPRH01TSU-02

EPLD

Section 5
System I/O EPLD
Note:
The System I/O EPLD generates system control register access signals from X-bus I/O
port transactions, supports power management, and provides various other system functions.
In this section, the system I/O EPLD is referred to as the EPLD.
The EPLD is a programmed Alterat EPM5130QC100 electrically programmable logic device. For timing and electrical specifications, see that data sheet.

5.1

System Register Support

The EPLD supports both internal and external registers.
5.1.1 External Register Support
The EPLD supports a group of external registers, which are latches or other devices that
are physically located in a device other than the EPLD. As shown in Figure 5, the EPLD
supplies control signals to the external registers, based on a decode of the address and
control lines of the X-bus (or ISA bus). In response to the signals from the EPLD, the external register either reads or writes data to the X-bus.
X-bus or ISA
A C D

Address

External Register
or Device

EPLD
Strobes

Control

Figure 5. Typical External Register

MPRH01TSU-02

51

Possible
Other
Circuitry

EPLD
For the external registers that the EPLD supports, Table 15 shows the external register,
the ISA I/O port address, and the supplied control signal(s).
Table 15. External Register Support
Register

ISA Port
Address

Read/
Write

Register
Location

Signal

Strobe or
Function

0060 or
0062 or
0064 or
0066

Keyboard Controller
Registers

R/W

Keyboard
Controller

KBD_CS# (asserted for an access to any of
these addresses)

Address Decode

0070

RTC Address Latch Enable

W/O

RTC

RTC_ALE

Write strobe

0071

RTC Data

Write

RTC

RTCWR#

Write strobe

RTCDS#

Read strobe

Read
0074

NVRAM Address Low Byte

W/O

NVRAM

AS0#

Address Decode

0075

NVRAM Address High Byte

W/O

NVRAM

AS1#

Address Decode

0077

NVRAM Data

Write

NVRAM

NVRAMWE#

Write Strobe

NVRAMOE#

Read Strobe
Read Strobe

Read
080C

Equipment Present

R/O

Board

PRSNT_RD#

080D

L2 Cache Status Register

R/O

Board

L2_STATUS_RD# Read Strobe

0852

Planar ID

R/O

Board

PLANAR_ID_RD# Read Strobe

0880

DRAM Presence Detect 1/2

R/O

Board

DRAM_PD_RD1#

Read Strobe

0881

DRAM Presence Detect 3/4

R/O

Board

DRAM_PD_RD2#

Read Strobe

5.1.2 Internal Registers
The EPLD contains a group of internal registers, which are accessed via the ISA bus I/O
port address shown for each register.
5.1.2.1
ISA Port

Storage Light Register
0808

Read/Write

Reset to xxxx xxx0

This register controls the HDD_LED# output of the EPLD. This signal normally controls the
hard disc drive activity LED.
MSB D7 D6 D5 D4 D3 D2 D1 D0 LSB
Hard Disk Activity Light
Reserved

Bit 0 . . . Hard Disk Activity Light:
0 = Turn light off (negate HDD_LED#).
1 = Turn light on (assert HDD_LED#).

52

MPRH01TSU-02

EPLD
5.1.2.2
ISA Port

Power Management Control Register 1
082A

Read/Write

Reset to xxxx xxxx

MSB D7 D6 D5 D4 D3 D2 D1 D0 LSB
Serial Data Bit
Command to 87C750
I/O Strobe Key
Reserved

This register is part of the power management system. For a detailed functional description
of the operation of this register, see the 603/604 Reference Design Power Management
Specification.
5.1.2.3
ISA Port

Power Management Control Register 2
082B

Read/Write

Reset to 0xxx 0xx0

MSB D7 D6 D5 D4 D3 D2 D1 D0 LSB
Reset PM Controller/Status
Reserved
IRQ12 Mask
Reserved
Reserved
Reserved
CPU1 ROM Completed

This register is part of the power management system. For a detailed functional description
of the operation of this register, see The 603/604 Reference Design Power Management
Specification.
5.1.2.4
ISA Port

Freeze Clock Register (FCR) Low
0860

Read/Write

Reset to 0000 0000

MSB D7 D6 D5 D4 D3 D2 D1 D0 LSB
Freeze Clock Register[7:0]

See section 5.1.2.5.
5.1.2.5
ISA Port

Freeze Clock Register (FCR) High
0862

Read/Write

Reset to xxx0 0000

MSB D7 D6 D5 D4 D3 D2 D1 D0 LSB
Freeze Clock Register [12:8]
Reserved

MPRH01TSU-02

53

EPLD
The freeze clock register (FCR) is a 13 bit register that is accessed by the system via the
X-bus as two 8-bit registers, as shown in this and the previous sections. Once triggered,
the FCR data is shifted out as serial data on FRZ_DATA_OUT.
This function is intended for use with the MPC970 clock chip, as is done in the reference
design. The MPC970 contains an input shift register, the input of which (Frz_Data) is connected to FRZ_DATA_OUT. ISA_CLK is used to clock the data from the EPLD to the
MPC970.
EPLD will shift the freeze clock data out of the FCR in response to either one of two triggers:
1. A write to ISA I/O port 862, or
2. A low to high transition on the UNFREEZE input iff bit 0 and bit 1 of the FCR are
high.
After triggering the data transfer, wait at least 2.3us for the transfer to complete before
accessing the FCR or retriggering the data transfer.
This is a one way serial data transfer between the two devices. Reading the FCR returns
the contents of the FCR, and does not cause a read of the data in the MPC970 register.
A 1 in a bit position freezes the corresponding clock output of the MPC970. For details of
the data transfer operation and the meaning of the data, see the MPC970 data sheet.

5.2

Signal Descriptions

Table 16 shows the active signals of the EPLD. Pins not shown should not be connected.
Table 16. Signal Descriptions
Signal Name

Pin

I/O

Description

ECS[2:0]

21, 22, 59

I

ECSEN#

60

I

Encoded Chip Select [2:0]. Encoded chip selects for peripheral devices supported by the ISA I/O Bridge. Used by
EPLD X-bus I/O port address decoders. From SIO.
Encoded Chip Select Enable. Asserted to enable the base
decoder. Negated to select the option decoder. Used by
EPLD X-bus I/O port address decoders. From SIO.
X–address bus [7:0]. Used by EPLD X-bus I/O port address decoders.

X-bus Interface Signals

XA[7:0]

61 ,64, 65,
I
66, 67, 70,
71, 72.
XD[7:0]
31, 32, 33, I/O
X-data bus [7:0]. Used by EPLD to transfer data.
92, 73, 74, 24mA
34, 95.
XIOW#
16
I
X-bus I/O Write. This signal indicates that the system is
writing to an X-bus I/O device. Used by EPLD X-bus I/O
port address decoders.
XIOR#
9
I
X-bus I/O Read. This signal indicates that the system is
reading from an X-bus I/O device. Used by EPLD X-bus
I/O port address decoders.
External Register Support Signals
AS0#

50

AS1#

49

O
6mA
O
6mA

NVRAM address register low byte write strobe. EPLD asserts this signal to write to X–bus port 0074.
NVRAM address register high byte write strobe. EPLD
asserts this signal to write to X–bus port 0075.

54

MPRH01TSU-02

EPLD
Table 16. Signal Descriptions (Continued)
Signal Name

Pin

I/O

Description

O
6mA
O
6mA
O
6mA
O
6mA
O
6mA

DRAM SIMM presence detect read enable 1. EPLD asserts this signal to read X-bus port 0880.
DRAM SIMM presence detect read enable 2. EPLD asserts this signal to read X-bus port 0881.
Keyboard chip select. EPLD asserts this signal to read
X-bus ports 0060, 0062, 0064, and 0066.
L2 cache status register read strobe. EPLD asserts this
signal to read X-bus port 080D.
NVRAM output enable (read strobe). EPLD asserts this
signal to read X–bus port 0076. This normally causes a
read of the NVRAM data stored at the location contained
in the NVRAM address register.
NVRAM data write strobe. EPLD asserts this signal to
write to X–bus port 0076. This normally causes the data
associated with this write to be written into the NVRAM
location contained in the NVRAM address register.
Planar ID read. EPLD asserts this signal to read X-bus
port 0852.
Equipment present register read. EPLD asserts this signal
to read X-bus port 080C.
Real time clock address latch enable. EPLD asserts this
signal to write X-bus port 0070.
Real time clock read strobe. EPLD asserts this signal to
read X–bus port 0071.
Real time clock write strobe. EPLD asserts this signal to
write to X–bus port 0071.

External Register Support Signals
DRAM_PD_RD1#

57

DRAM_PD_RD2#

56

KYBD_CS#

98

L2_STATUS_RD#

54

NVRAMOE#

1

NVRAMWE#

53

O
6mA

PLANAR_ID_RD#

52

PRSNT_RD#

41

RTC_ALE

48

RTCDS#

51

RTCWR#

35

O
6mA
O
6mA
O
6mA
O
6mA
O
6mA

Interrupt Signals
IRQ1

86

O

IRQ1_IN

89

I

IRQ12

36

I

Interrupt request 1. Latched active when IRQ1_IN is asserted. Negated when KYBD_CS# is asserted.
Keyboard interrupt. EPLD is designed to intercept the
keyboard interrupt between the keyboard and the ISA bus
bridge. Either connect IRQ1 and IRQ1_IN as shown in the
reference design, or disconnect both signals from the system (routing the keyboard interrupt to the ISA bus bridge),
or see the 603/604 Reference Design Power Management Guide.
Interrupt request 12 input. Connect to system IRQ12, the
mouse interrupt. Also see the 603/604 Reference Design
Power Management Guide.

System Clock Interface Signals
FRZ_DATA_OUT

Freeze data out. Serial data stream to MPC970 clock
chip. See the MPC970 data sheet.
ISA_CLK
20
ISA clock. Used to clock the freeze serial data stream to
the MPC970 clock chip. See the MPC970 data sheet.
Also see the 603/604 Reference Design Power Management Guide.
Power Management Signals (not used in release 2.0 or 2.1)

MPRH01TSU-02

90

O
6mA
I

55

EPLD
Table 16. Signal Descriptions (Continued)
Signal Name

Pin

I/O

Description

Power Management Signals (not used in release 2.0 or 2.1)
83CX_RESET

ACTIVITY#
CMD_STATE#

EXT_ACTVTY#

IO_STROBE#
PROC_RDY

RWD0

UNFREEZE

91

O
6mA

Power management controller reset. No-connect or connect as shown in the 603/604 Reference Design Power
Management Guide.
58
O
Activity. No-connect or connect as shown in the 603/604
6mA Reference Design Power Management Guide.
8
I
Power manangement controller command state. No-connect or connect as shown in the 603/604 Reference Design Power Management Guide.
3
I
External activity. No-connect or pull low or connect as
shown in the 603/604 Reference Design Power Management Guide.
99
O
I/O strobe. No-connect or connect as shown in the
6mA 603/604 Reference Design Power Management Guide.
83
I
Power management controller ready. No-connect or connect as shown in the 603/604 Reference Design Power
Management Guide.
96 I/O
Power management controller serial read/write data bit.
24mA No-connect or connect as shown in the 603/604 Reference Design Power Management Guide.
82
I
Unfreeze. No-connect or connect as shown in the
603/604 Reference Design Power Management Guide.
100

O
6mA

10

I

18, 19, 43,
44, 68, 69,
93, 94
12, 13, 37,
38, 62, 63,
87, 88

I

Hard disk drive activity light. EPLD asserts this signal
while bit 0 of the storage light register (port 0808) is 1.
This signal normally indicates hard disk drive activity.
System reset. Used by EPLD to reset internal state machines and internal registers.
+5V:

I

GROUND:

Other Signals
HDD_LED#

RESET#
VCC

GND

56

MPRH01TSU-02

EPLD

5.3
5.3.1

EPLD Design Equations
Fit File

–– MAX+plus II Compiler Fit File
–– Version 5.0 8/5/94
–– Compiled: 05/05/95 14:44:22
BEGIN
DEVICE = ”EPM5130WC–1”;
”A0”
”A1”
”A2”
”A3”
”A4”
”A5”
”A6”
”A7”
”/CMD_STATE”
”/ECSEN”
”ECS0”
”ECS1”
”ECS2”
”/EXT_ACTVTY”
”IRQ1”
”IRQ12”
”ISA_CLK”
”PROC_RDY”
”/RESET”
”UNFREEZE”
”/XIOR”
”/XIOW”
”/ACTIVITY”
”/AS0”
”/AS1”
”/DRAM_PD_RD1”
”/DRAM_PD_RD2”
”FRZ_DATA_OUT”
”/HDD_LED”
”/IO_STROBE”
”IRQ1_OUT”
”/KYBD_CS”
”/L2_STATUS_RD”
”/NVRAMOE”
”/NVRAMWE”
”/PLANAR_ID_RD”
”/PRSNT_RD”

:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:

MPRH01TSU-02

57

INPUT_PIN
INPUT_PIN
INPUT_PIN
INPUT_PIN
INPUT_PIN
INPUT_PIN
INPUT_PIN
INPUT_PIN
INPUT_PIN
INPUT_PIN
INPUT_PIN
INPUT_PIN
INPUT_PIN
INPUT_PIN
INPUT_PIN
INPUT_PIN
INPUT_PIN
INPUT_PIN
INPUT_PIN
INPUT_PIN
INPUT_PIN
INPUT_PIN
OUTPUT_PIN
OUTPUT_PIN
OUTPUT_PIN
OUTPUT_PIN
OUTPUT_PIN
OUTPUT_PIN
OUTPUT_PIN
OUTPUT_PIN
OUTPUT_PIN
OUTPUT_PIN
OUTPUT_PIN
OUTPUT_PIN
OUTPUT_PIN
OUTPUT_PIN
OUTPUT_PIN

=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=

72
71
70
67
66
65
64
61
8
60
59
22
21
3
89
36
20
83
10
82
9
16
58
50
49
57
56
90
100
99
86
98
54
1
53
52
41

;
;
;
;
;
;
;
;
;
;
;
;
;
;
;
;
;
;
;
;
;
;
;
;
;
;
;
;
;
;
;
;
;
;
;
;
;

–– LC8

–– LC3
–– LC103
–– LC38
–– LC99
–– LC98

––
––
––
––
––
––
––
––
––
––
––
––
––
––
––

LC72
LC56
LC55
LC71
LC70
LC104
LC120
LC119
LC102
LC118
LC68
LC1
LC67
LC66
LC49

EPLD
”RTC_ALE”
”/RTCDS”
”/RTCWR”
”83CX_RESET”
”RWD0”
”XD0”
”XD1”
”XD2”
”XD3”
”XD4”
”XD5”
”XD6”
”XD7”
”CLKFF0”
”CLKFF1”
”CLKFF2”
”CLKFF3”
”CLKFF4”
”CLKFF5”
”CLKFF6”
”CLKFF7”
”CLKFF8”
”CLKFF9”
”CLKFF10”
”CLKFF11”
”CLKFF12”
”CNTR0”
”CNTR1”
”CNTR2”
”CNTR3”
”DOUBLE_FRZ”
”DOUBLE_SNC”
”FREEZEFF”
”GEN_STOP_BITFF”
”GPCS0”
”PWR_REG21”
”PWR_REG22”
”SHIFT_ENFF”
”SNC_FREEZE”
”SNC_SHIFT_ENFF”
”SNC_UNFREEZE”
”START_SHIFTFF”
”UNFREEZEFF”
END;

:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:
:

58

OUTPUT_PIN
OUTPUT_PIN
OUTPUT_PIN
OUTPUT_PIN
BIDIR_PIN
BIDIR_PIN
BIDIR_PIN
BIDIR_PIN
BIDIR_PIN
BIDIR_PIN
BIDIR_PIN
BIDIR_PIN
BIDIR_PIN
LOCATION
LOCATION
LOCATION
LOCATION
LOCATION
LOCATION
LOCATION
LOCATION
LOCATION
LOCATION
LOCATION
LOCATION
LOCATION
LOCATION
LOCATION
LOCATION
LOCATION
LOCATION
LOCATION
LOCATION
LOCATION
LOCATION
LOCATION
LOCATION
LOCATION
LOCATION
LOCATION
LOCATION
LOCATION
LOCATION

=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=
=

48
51
35
91
96
95
34
74
73
92
33
32
31
LC48
LC96
LC95
LC94
LC47
LC46
LC45
LC44
LC43
LC93
LC92
LC91
LC112
LC111
LC110
LC109
LC108
LC90
LC107
LC106
LC105
LC89
LC80;
LC79
LC103
LC99
LC98
LC101
LC100
LC97

; –– LC54
; –– LC65
; –– LC37
; –– LC113
; –– LC116
; –– LC115
; –– LC36
; –– LC82
; –– LC81
; –– LC114
; –– LC35
; –– LC34
; –– LC33
;
;
;
;
;
;
;
;
;
;
;
;
;
;
;
;
;
;
;
;
;
;
––IRQ12_MASK
;
; –– PIN 89
; –– PIN 83
; –– PIN 82
; –– PIN 85
; –– PIN 84
; –– PIN 81

MPRH01TSU-02

EPLD
5.3.2

TDF File

%***************************************************************************%
%****** AHDL SOURCE CODE FOR: SIO/XBUS INTERFACE CONTROL *****************%
%***************************************************************************%
SUBDESIGN chandra(
%%
%
DEFINE PRIMARY INPUTS AND OUTPUTS
%%
A[7..0]
:INPUT; % A7 is MSB and A0 is LSB %
ECS[2..0]
:INPUT; % ECS2 is MSB and ECS0 is LSB
/ECSEN
:INPUT;
/XIOR
:INPUT;
/XIOW
:INPUT;
/RESET
:INPUT;
/CMD_STATE
:INPUT;
PROC_RDY
:INPUT;
/EXT_ACTVTY
:INPUT;
IRQ1
:INPUT;
IRQ12
:INPUT;% Freeze Clock %
ISA_CLK
:INPUT;
UNFREEZE
:INPUT;
/PLANAR_ID_RD
/PRSNT_RD
/L2_STATUS_RD
/DRAM_PD_RD1
/DRAM_PD_RD2
/KYBD_CS
IRQ1_OUT
RTC_ALE
/RTCDS
/RTCWR
/AS0
/AS1
/NVRAMWE
/NVRAMOE
/HDD_LED
83CX_RESET
/ACTIVITY
/IO_STROBE
FRZ_DATA_OUT
XD[7..0]
RWD0

MPRH01TSU-02

%

:OUTPUT;
:OUTPUT;
:OUTPUT;
:OUTPUT;
:OUTPUT;
:OUTPUT;
:OUTPUT;
:OUTPUT;
:OUTPUT;
:OUTPUT;
:OUTPUT;
:OUTPUT;
:OUTPUT;
:OUTPUT;
:OUTPUT;
:OUTPUT;
:OUTPUT;
:OUTPUT;
:OUTPUT;
:BIDIR;
:BIDIR;

% XD7 is MSB and XD0 is LSB
)

VARIABLE
GPCS0
HDD_LEDFF
D[7..0]
XD_TRI_OE
LIGHT_STRB
INT1FF
CLRINT1

%

% Misc. Variables%
:NODE;
:SRFF;
:NODE;
:NODE;
:NODE;
:DFF;
:NODE;

% Keyboard interput latch %
% Power Management Variables%

59

%

EPLD
PWR_REG1_STRB
PWR_REG2_STRB
PWR_REG2[2..0]
83CX_CS
RWD0_STRB
IRQ12_MASK
CLKFF[12..0]
SHIFT_ENFF
CNTR[3..0]
SNC_SHIFT_ENFF
START_SHIFTFF
CLKFF_WR
CLKFF_STRB
CLKFF_SELL
CLKFF_SELH
GEN_START_BIT
GEN_STOP_BITFF
STOP_SHIFT
UNFREEZEFF
SNC_UNFREEZE
DOUBLE_SNC
FREEZEFF
SNC_FREEZE
DOUBLE_FRZ

:NODE;
:NODE;
:SRFF;
:NODE;
:NODE;
:NODE;
:DFFE;
:SRFF;
:DFF;
:DFFE;
:SRFF;
:NODE;
:NODE;
:NODE;
:NODE;
:NODE;
:DFF;
:NODE;
:DFFE;
:DFF;
:DFF;
:DFFE;
:DFF;
:DFF;

% Freeze Clock Variables %

%
%
%
%
%
%

Unsynchronized UNFREEZE
Synchronized UNFREEZE
Double Synchronized UNFREEZE
Unsynchronized
FREEZE
Synchronized
FREEZE
Double Synchronized
FREEZE

%
%
%
%
%
%

BEGIN%*****%
% CHANDRA uses SIO’s General Purpose Register Decode 0 –– Software sets
% to 800–8FF to enable the following decoded signals.

%
%

GPCS0 = LCELL (!ECS[2] & ECS[1] & !ECS[0] & /ECSEN);
% Hard Disk Light I/O address range : 0808

%

LIGHT_STRB = !A[7] & !A[6] & !A[5] & !A[4] & A[3] & !A[2]
& !A[1] & !A[0] & GPCS0; HDD_LEDFF.s = XD[0] & LIGHT_STRB;
HDD_LEDFF.r = !XD[0] & LIGHT_STRB;
HDD_LEDFF.clk= GLOBAL(/XIOW);
HDD_LEDFF.clrn= (/RESET);
/HDD_LED
= !HDD_LEDFF.q;
%%
% Equipment present Read Command 1 (I/O address range: 080C)
%%
/PRSNT_RD
= ! (!A[7] & !A[6] & !A[5] & !A[4] & A[3] & A[2]
& !A[1] & !A[0] & GPCS0 & !/XIOR);
%%
% Equipment present Read Command 2 (I/O address range: 080D)
%%
/L2_STATUS_RD = !(!A[7] & !A[6] & !A[5] & !A[4] & A[3] & A[2]
& !A[1] & A[0] & GPCS0 & !/XIOR);
%%
% External/Internal Planar ID I/O address range: 0852
%%
/PLANAR_ID_RD = !(!A[7] & A[6] & !A[5] & A[4] & !A[3] & !A[2]
& A[1] & !A[0] & GPCS0 & !/XIOR);
%%

60

%

%

%

MPRH01TSU-02

EPLD
% Keyboard Chip Select I/O address range = 0060, 0062, 0064, 0066–
%%
/KYBD_CS = !(!ECS[2] & ECS[1] & !ECS[0] & !/ECSEN);

%

% Keyboard interupt (INT1) Clear = KBD_CS qualified with XIOR %
CLRINT1 =!ECS[2] & ECS[1] & !ECS[0] & !/ECSEN & !/XIOR; INT1FF.d=VCC;
INT1FF.clk = IRQ1;
% Set Keyboard INT1 latch on rising edge of IN1 %
INT1FF.clrn = /RESET & !CLRINT1;
IRQ1_OUT = INT1FF.q;
%%
% RTC_ALE I/O address 0070
%
%%
RTC_ALE = (!A[2] & !A[1] & !A[0] & !ECS[2] & !ECS[1] & !/ECSEN & !/XIOW) ;
%%
% RTCWR I/O address 0071
%%
/RTCWR = !(!A[2] & !A[1] &
%%
% RTCDS I/O address 0071
%%
/RTCDS = !(!A[2] & !A[1] &
%%

%
A[0] & !ECS[2] & !ECS[1] & !/ECSEN & !/XIOW);

%
A[0] & !ECS[2] & !ECS[1] & !/ECSEN & !/XIOR);

% Nvram AS0 I/O address 0074
%
%%
/AS0
= !(A[2] & !A[1] & !A[0] & !ECS[2] & !ECS[1] & !/ECSEN & !/XIOW);
%%
% Nvram AS1 I/O address range: 0075
%
%%
/AS1
= !(A[2] & !A[1] & A[0] & !ECS[2] & !ECS[1] & !/ECSEN & !/XIOW);
%%
% Nvram RTC WER I/O address 0077
%
%%
/NVRAMWE = !(A[2] & A[1] & A[0] & !ECS[2] & !ECS[1] & !/ECSEN & !/XIOW);
%%
% Nvram RTC OER I/O address range: 0077
%
%%
/NVRAMOE = !(A[2] & A[1] & A[0] & !ECS[2] & !ECS[1] & !/ECSEN & !/XIOR)
; %%
% DRAM_PD_RD1 enables Present Detect I/O address range: 0880
%
%%
/DRAM_PD_RD1 = !(A[7] & !A[6] & !A[5] & !A[4] & !A[3] & !A[2] & !A[1]
& !A[0] & GPCS0 & !/XIOR);

MPRH01TSU-02

61

EPLD
%%
% DRAM_PD_RD2 enables Present Detect I/O address range: 0881
%
%%
/DRAM_PD_RD2 = !(A[7] & !A[6] & !A[5] & !A[4] & !A[3] & !A[2] & !A[1]
& A[0] & GPCS0 & !/XIOR);
%%
%*************************************************************************%
%
POWER MANAGEMENT
%
%
%
%*************************************************************************%
% Write Power Control Register 1 I/O address: 082A
%
%
%
% (MSB)
Bits 7–5
Reserved
%
%
Bit 4
I/O Strobe Key
(W/O)
%
%
Bits 3–1
Data/Command B[3..1] to 83C750
(W/O)
%
% (LSB)
Bit 0
83C750 D0
(R/W)
%
%*************************************************************************%
83CX_CS =
(!A[7] & !A[6] & A[5] & !A[4] & A[3] & !A[2] & A[1] & GPCS0);
PWR_REG1_STRB = 83CX_CS & !A0;
%%
% –Write ZERO to 83C750
%%
RWD0_STRB
= (PWR_REG1_STRB & !/XIOW & !XD[0]);
RWD0
= TRI (GND, RWD0_STRB);

%

%%
% 83C750 I/O Strobe Key
%%
/IO_STROBE
= !(XD[4] & PWR_REG1_STRB & !/XIOW & PROC_RDY
& !83CX_RESET);
%IO_CHRDY

%

= TRI (GND, IO_STRB);%

%*************************************************************************%
% Write Power Control Register 2 I/O address: 082B
%
%
%
% (MSB)
Bit 7
CPU1 ROM Completed
(Reset by /RESET,
W/R)
%
%
Bit 6
Reserved
%
%
Bit 5
Reserved
%
%
Bit 4
Reserved
%
%
Bit 3
IRQ12 Mask
(Reset by /RESET)
(W/R)
%
%
Bits 2–1
Reserved
%
% (LSB)
Bit 0
Reset 83C750
(W/O)
%
%*************************************************************************%
PWR_REG2_STRB = 83CX_CS & A0;
PWR_REG2[1].s
= XD[3] & PWR_REG2_STRB;
PWR_REG2[1].r
= !XD[3] & PWR_REG2_STRB;
PWR_REG2[1].clrn = (/RESET);
PWR_REG2[2].s
= XD[7] & PWR_REG2_STRB;
PWR_REG2[2].r
= !XD[7] & PWR_REG2_STRB;
PWR_REG2[2].clrn = (/RESET);
PWR_REG2[0].s
= XD[0] & PWR_REG2_STRB;
PWR_REG2[0].r
= !XD[0] & PWR_REG2_STRB;

62

MPRH01TSU-02

EPLD
PWR_REG2[0].clrn
PWR_REG2[].clk
83CX_RESET
IRQ12_MASK

=
=
=
=

VCC;
GLOBAL(/XIOW);
PWR_REG2[0].q;
PWR_REG2[1].q;

%*************************************************************************%
% This output should be NORed with IRQ1 and EXTACTIV signals externally to%
% provide a real Activity Alert to the 87C350 (because of restriction of %
% CHANDRA’s IO pins.)
%********************%
/ACTIVITY = !(IRQ1 # !IRQ12_MASK & IRQ12 # !/EXT_ACTVTY);

%

%*************************************************************************%
% Freeze Clock Logic – Refer to the Motorola MC88LV970 PLL Clock Driver
%
% Specification for information on Freeze Data protocol.
%
%
%
% Freeze Clock Logic contains a 12 Bit serial shift register
%
% Decode the low order CLKFF[7..0] on addresses: 0860 –0861
%
% Decode the high order CLKFF[12..8] on addresses: 0862 –0863
%
%*************************************************************************%CLK
FF_STRB
= (!A[7] & A[6] & A[5] & !A[4] & !A[3] & !A[2] & GPCS0);
CLKFF_WR
= CLKFF_STRB & !/XIOW;
CLKFF_SELL = CLKFF_STRB & !A[1];
CLKFF_SELH = CLKFF_STRB & A[1];
CLKFF[12..0].prn = VCC;
CLKFF[12..0].clrn = /RESET;
CLKFF[12..0].ena = START_SHIFTFF.q # CLKFF_WR;
CLKFF[12..0].clk = ISA_CLK;
if START_SHIFTFF.q then
CLKFF[11..0].d = CLKFF[12..1].q;
CLKFF[12].d
= CLKFF[0].q;
else if CLKFF_SELL then
CLKFF[7..0].d = XD[7..0];
CLKFF[12..8].d = CLKFF[12..8].q;
else if CLKFF_SELH then
CLKFF[7..0].d = CLKFF[7..0].q;
CLKFF[12..8].d = XD[4..0];
else
CLKFF[12..0].d = CLKFF[12..0].q;
end if;
end if;
end if;

% Shift with wraparound %

% Write to CLKFF[7..0]

%

% Write to CLKFF[12..8] %

% Do nothing

%

%*************************************************************************%
% UNFREEZE Flip Flop is cleared asynchronously when the UNFREEZE signal
%
% makes a low to high transition. It is set once shifting has been enabled%
% The CHANDRA ignores the UNFREEZE signal if the 601 PCLK clocks are not %
% frozen.
%
%*************************************************************************%UNFREEZEFF.prn
= !SNC_SHIFT_ENFF.q & /RESET;
UNFREEZEFF.clrn = VCC;
UNFREEZEFF.ena = !SHIFT_ENFF.q ;
UNFREEZEFF.d
= !(CLKFF[0].q # CLKFF[1].q);
UNFREEZEFF.clk = UNFREEZE ;

MPRH01TSU-02

63

EPLD

%*************************************************************************%
% This signal is the UNFREEZE Flip Flop synchronized to the ISA clock.
%
%*************************************************************************%
SNC_UNFREEZE.prn = /RESET;
SNC_UNFREEZE.clrn = VCC;
SNC_UNFREEZE.d
= UNFREEZEFF.q;
SNC_UNFREEZE.clk = ISA_CLK;
%****************************************************************************%
% This signal is the UNFREEZE Flip Flop double synchronized to the ISA clock.%
%****************************************************************************%
DOUBLE_SNC.prn = /RESET;
DOUBLE_SNC.clrn = VCC;
DOUBLE_SNC.d
= SNC_UNFREEZE.q;
DOUBLE_SNC.clk = ISA_CLK;
%*************************************************************************%
% FREEZE Flip Flop is cleared asynchronously when the UNFREEZE
signal
%
% makes a high to low transition. It is set once shifting has been enabled%
% The CHANDRA ignores the UNFREEZE signal if the 601 PCLK clocks are
%
% already frozen.
%
%*************************************************************************%
FREEZEFF.prn = !SNC_SHIFT_ENFF.q & /RESET;
FREEZEFF.clrn = VCC;
FREEZEFF.ena = !SHIFT_ENFF.q ;
FREEZEFF.d
= CLKFF[0].q & CLKFF[1].q;
FREEZEFF.clk = !UNFREEZE ;
%*************************************************************************%
% This signal is the FREEZE Flip Flop synchronized to the ISA clock.
%
%*************************************************************************%
SNC_FREEZE.prn = /RESET;
SNC_FREEZE.clrn = VCC;
SNC_FREEZE.d
= FREEZEFF.q;
SNC_FREEZE.clk = ISA_CLK;
%****************************************************************************%
% This signal is the
FREEZE Flip Flop double synchronized to the ISA clock.%
%****************************************************************************%
DOUBLE_FRZ.prn = /RESET;
DOUBLE_FRZ.clrn = VCC;
DOUBLE_FRZ.d
= SNC_FREEZE.q;
DOUBLE_FRZ.clk = ISA_CLK;
SHIFT_ENFF.s
= CLKFF_SELH;
% Start shifting upon write to 862
%
SHIFT_ENFF.r
= GND;
SHIFT_ENFF.prn = DOUBLE_SNC.q & DOUBLE_FRZ;
SHIFT_ENFF.clrn = !GEN_STOP_BITFF.q & /RESET; %Clr when it reaches 15 & rst%
SHIFT_ENFF.clk = /XIOW;SNC_SHIFT_ENFF.d = SHIFT_ENFF.q;
% One clock after SHIFT_ENFF%
SNC_SHIFT_ENFF.ena = /RESET;
SNC_SHIFT_ENFF.clrn = /RESET;
SNC_SHIFT_ENFF.clk = ISA_CLK;
CNTR[].clk = ISA_CLK;
CNTR[].clrn = /RESET;
% Counter resets to zero
%

64

MPRH01TSU-02

EPLD
If SNC_SHIFT_ENFF.q Then
CNTR[].d = CNTR[].q + 1;
Else
CNTR[].d = CNTR[].q;
End If;

% Count while SNC_SHIFT_ENFF is 1

GEN_START_BIT = !CNTR[3].q & !CNTR[2].q & !CNTR[1].q & CNTR[0].q; % 1 %
STOP_SHIFT
= CNTR[3].q & CNTR[2].q & CNTR[1].q & !CNTR[0].q; % 14 %
GEN_STOP_BITFF.d = STOP_SHIFT; %Generate Stop Bit when counter is 15 %
GEN_STOP_BITFF.prn = VCC;
GEN_STOP_BITFF.clrn = /RESET;
GEN_STOP_BITFF.clk = ISA_CLK;
START_SHIFTFF.s
= GEN_START_BIT;
START_SHIFTFF.r
= STOP_SHIFT;
START_SHIFTFF.clrn = /RESET;
START_SHIFTFF.clk = ISA_CLK;
%*************************************************************************%
% FRZ_DATA_OUT is a 1 when the counter = 0. It is also a 1 whenever
%
% START_SHIFTFF is 1 and CLKFF[0] is a 1 and then when the counter is
%
% Fifteen to leave it in the high state.
%
%*************************************************************************%
FRZ_DATA_OUT = !CNTR[3].q & !CNTR[2].q & !CNTR[1].q & !CNTR[0].q
# GEN_STOP_BITFF.q
# START_SHIFTFF.q & !CLKFF[0].q;

%*************************************************************************%
% Read Storage Light Status Register I/O address: 0808
%
% (MSB)
Bits 7–1 Reserved
%
% (LSB)
Bit 0
Hard Disk Active Light
(W/R) %
%
%
% Read Power Control Register 1 I/O address: 082A
%
% (MSB)
Bits 7–1
Reserved
%
% (LSB)
Bit 0
83C750 D0
(W/R) %
%
%
% Read Power Control Register 2 I/O address: 082B
%
% (MSB)
Bit 7
CPU1 ROM Completed
(Reset by /RESET,
R/W)
%
%
Bit 6
Reserved
%
%
Bit 5
Reserved
%
%
Bit 4
Reserved
%
%
Bit 3
IRQ12 Mask
(Reset by /RESET,
W/R)
%
%
Bits 2–1
Reserved
%
% (LSB)
Bit 0
83C750 Status
(R/O)
%
%
%
%*************************************************************************%
XD_TRI_OE = ((LIGHT_STRB
# PWR_REG1_STRB # PWR_REG2_STRB
# CLKFF_STRB) & !/XIOR);
XD[0]
= TRI (D[0], XD_TRI_OE);
D[0]
= HDD_LEDFF & LIGHT_STRB
# RWD0 & PWR_REG1_STRB
# PROC_RDY & /CMD_STATE & PWR_REG2_STRB
# CLKFF[0].q & CLKFF_SELL
# CLKFF[8].q & CLKFF_SELH;

MPRH01TSU-02

65

%

EPLD
XD[1]
D[1]
XD[2]
D[2]
XD[3]
D[3]

XD[4]
D[4]
XD[5]
D[5]
XD[6]
D[6]
XD[7]
D[7]

= TRI (D[1], XD_TRI_OE);
= CLKFF[1].q & CLKFF_SELL
# CLKFF[9].q & CLKFF_SELH;
= TRI (D[2], XD_TRI_OE);
= CLKFF[2].q & CLKFF_SELL
# CLKFF[10].q & CLKFF_SELH;
= TRI (D[3], XD_TRI_OE);
= PWR_REG2[1].q & PWR_REG2_STRB
# CLKFF[3].q & CLKFF_SELL
# CLKFF[11].q & CLKFF_SELH;
= TRI (D[4], XD_TRI_OE);
= CLKFF[4].q & CLKFF_SELL
# CLKFF[12].q & CLKFF_SELH;
= TRI (D[5], XD_TRI_OE);
= CLKFF[5].q & CLKFF_SELL;
= TRI (D[6], XD_TRI_OE);
= CLKFF[6].q & CLKFF_SELL;
= TRI (D[7], XD_TRI_OE);
= PWR_REG2[2].q & PWR_REG2_STRB
# CLKFF[7].q & CLKFF_SELL;

END;

66

MPRH01TSU-02

Memory

Section 6
Memory Systems
The reference design contains four distinct memory systems: system memory (DRAM), L2
cache memory (SRAM and tagRAM), ROM for boot and POST code, and NVRAM.

6.1

DRAM

The reference design has slots for up to 128M of system memory (DRAM), arranged as
four industry standard 72-pin SIMMs (see Table 17). When using more than one pair of
32MB SIMMs however, it is necessary to insert a buffer on the WE[1:0] and MA[11:0] lines
between the SIMMs and the 664, in order to drive the increased capacitive load presented
by the 32MB SIMMs. The reference design does not buffer these signals, and does not provide circuit board patterns for such buffers.
S 70ns
S 4 data bytes plus 1 parity bit per byte, 72 pin SIMM
S presence detect bits.
Table 17. Supported DRAM Modules
Size

Organization IBM Part Number

DRAM Data Sheet

4MB

1M x 36b

IBM11E1360BA

MMDS14DSU–00

8MB

2M x 36b

IBM11E2360BA

MMDS22DSU–00

16MB

4M x 36b

IBM11E4360B

MMDS26DSU–00

32MB

8M x 36b

IBM11E8360B

MMDS27DSU–00

6.1.1 Refresh
The memory controller in the 660 bridge provides a flexible refresh capability for the reference design.
The ISA bus bridge provides the ISA_REFRESH# signal to refresh ISA bus memory. Refer
to the SIO data book for more information.

6.1.2 DRAM Presence Detection
The reference design includes a method for software to detect and identify installed DRAM
modules using the presence detect bits, PD[15:0]. See section 4.6.7.
MPRH01TSU-02

67

Memory

MRE3#

MRE2#

Memory
Bank 1

MRE1#

MRE0#

Memory
Bank 0

PD1

SIMM 1 PD2
PD3
PD4

CAS0
CAS1
CAS2
CAS3
WE

PD3
PD4

MPD8
MPD9
MPD10
MPD11

Data & Par

D[0:31]

PD1

SIMM 3 PD2
CAS0
CAS1
CAS2
CAS3
WE

SIMM 2

PD1
PD2
PD3
PD4

RAS0
RAS1
RAS2
RAS3

CASP

RAS0
RAS1
RAS2
RAS3

Data & Par

MCE4#
MCE5#
MCE6#
MCE7#
MWE1#

CASP

Presence
Detect Bits
MPD0
MPD1
MPD2
MPD3

MPD4
MPD5
MPD6
MPD7

CAS0
CAS1
CAS2
CAS3
WE

SIMM 4

CASP

MCE0#
MCE1#
MCE2#
MCE3#
MWE0#

GND
RAS0
RAS1
RAS2
RAS3

CASP

RAS0
RAS1
RAS2
RAS3

GND

PD1
PD2
PD3
PD4

MPD12
MPD13
MPD14
MPD15

CAS0
CAS1
CAS2
CAS3
WE
Data & Par

Data & Par

D[32:64]

Figure 6. DRAM Bank Organization

6.1.3 Organization
The 4-byte DRAM modules (SIMMs) are arranged in pairs to form 8-byte memory banks,
as shown in Figure 6. Each socket supports a 32-bit (non-parity) or a 36-bit (4 data bytes
plus 4 parity bits) DRAM SIMM. Bank 0 is composed of SIMM 1 and SIMM 2. Bank 1 is composed of SIMM 3 and SIMM 4. The 2 SIMMs in each bank must be the same size.
Combining the internal organization of the SIMMs (found in the respective data sheets) with
the bank organization of Figure 6 shows that each parity bit is accessed with the associated
data byte. The 660 bridge uses this standard organization for either parity or ECC mode
operation.

68

MPRH01TSU-02

Memory

6.2

L2 Cache

The reference design supplies an L2 cache controller, located inside the 660 Bridge chipset. The motherboard provides a socket for an SRAM module, and the L2 tag RAM (two
16K x 15 synchronous devices) is supplied installed on the board. The L2 is a unified, writethru, direct-mapped, look-aside level 2 cache that caches the low 1G of CPU memory
space.
The reference design is initially configured to use a 512K synchronous SRAM module, but
can be configured to use 256K, 512K, or 1M module populated with either synchronous or
asynchronous SRAM.
The L2 supplies data to the CPU bus on write hits and snarfs the data (updates the SRAM
data while the memory controller is accessing DRAM memory) on read/write misses. It
snoops PCI to memory transactions. Typical synchronous SRAM read performance with
9ns SRAM is 3-1-1-1, followed by -2-1-1-1 on pipelined reads. Typical asynchronous
SRAM read performance with 15ns SRAM is 3-2-2-2, followed by -3-2-2-2 on pipelined
reads. For more information on the operation and capabilities of the L2, see the 660 Bridge
User’s Manual.

6.2.1

SRAM

As initially configured (Figure 8), the reference design features a 512K synchronous SRAM
SIMM. A synchronous 256K SRAM module is shown in Figure 7. A synchronous 1M SRAM
module is shown in Figure 9.
It is also possible to use asynchronous SRAM modules with the reference design.
Figure 10 shows a 256K asynchronous SRAM module. Figure 11 shows a 512K asynchronous SRAM module. Figure 12 shows a 1M asynchronous SRAM module.

CPU_ADDR[14:28]

32k x 18
SRAM
Address

CPU_DATA[0:63]

Data

CPU_DPAR[0:7]

Data
CS
CS
CS
CS

Figure 7. Synchronous SRAM, 256K L2

MPRH01TSU-02

69

Memory

64k x 18
SRAM
Address

CPU_ADDR[13:28]
CPU_DATA[0:63]

Data

CPU_DPAR[0:7]

Data
CS
CS
CS
CS

Figure 8. Synchronous SRAM, 512K L2

CPU_ADDR[13:28]
CPU_DATA[0:63]
CPU_DPAR[0:7]

64k x 18
SRAM
Address

64k x 18
SRAM
Address

Data

Data

Data

Data

CS

CS

CS

CS

CS
CPU_ADDR[12]

CS

CS

CS

Figure 9. Synchronous SRAM, 1M L2

70

MPRH01TSU-02

Memory

CPU_ADDR[14:28]

32k x 9
SRAM
Address

CPU_DATA[0:63]

Data

CPU_DPAR[0:7]

Data
CS
CS
CS
CS
CS
CS
CS
CS

Figure 10. Asynchronous SRAM, 256K L2

CPU_ADDR[13:28]

64k x 18
SRAM
Address

CPU_DATA[0:63]

Data

CPU_DPAR[0:7]

Data
CS
CS
CS
CS

Figure 11. Asynchronous SRAM, 512K L2

MPRH01TSU-02

71

Memory

CPU_ADDR[12:28]

128k x 9
SRAM
Address

CPU_DATA[0:63]

Data

CPU_DPAR[0:7]

Data
CS
CS
CS
CS
CS
CS
CS
CS

Figure 12. Asynchronous SRAM, 1M L2

6.2.2 TagRAM
The tagRAM, a pair of IDT71216S10 devices, is installed on the reference board. The tagRAM configuration for a 512K L2 (initially supplied on the reference design) is shown in
Figure 13. The tagRAM configuration for a 1M L2 is shown in Figure 14.

CPU_ADDR[13:26]

16k x 15
TagRAM
Address

CPU_ADDR[2:12]

Data

TAG_VALID

Valid

Match

VDD

TAG_MATCH

CS
Figure 13. Synchronous TagRAM, 512K L2

CPU_ADDR[13:26]
CPU_ADDR[2:11]
TAG_VALID
CPU_ADDR[12]

16k x 15
TagRAM
Address

VDD

Data
Valid
CS

Match

CS

TAG_MATCH

Match

Figure 14. Synchronous TagRAM, 1M L2

72

MPRH01TSU-02

Memory

6.2.3 L2 Cache Configuration
To implement the desired L2 configuration on the reference design, refer to Table 18.
Table 18. L2 Configuration Implementation
Configuration Component Values

MPRH01TSU-02

Cache Size

SRAM Type

256K

512K

1M

Burst

Asynch

R 463

0W

0W

no-pop

R 465

0W

no-pop

no-pop

R 466

no-pop

10k

10k

R 467

no-pop

0W

0W

R 475

10k

no-pop

no-pop

R 477

no-pop

no-pop

10k

R 478

no-pop

no-pop

0W

R 479

1k

1k

no-pop

U 37

populate

populate

populate

U 38

nopop

nopop

populate

R 168

0W

no-pop

R 169

no-pop

0W

R 170

no-pop

0W

R 171

0W

no-pop

R 172

0W

no-pop

R 474

no-pop

0W

R 476

no-pop

0W

73

Memory

6.3

ROM

The reference design uses an AMD AM29F040-120 Flasht ROM to contain the POST and
boot code. It is recommended that Vital Product Data (VPD) such as the motherboard
speed and native I/O complement be programmed into in this device. It is possible to program the Flash before or during the manufacturing process.

6.4

CPU to ROM Transfers

The PowerPC Reference Platform Specification allocates the upper 8M of the 4G CPU address space as ROM space. The reference design implements a 2M ROM space from
4G–2M to 4G. The actual ROM is a 512K device located at 4G–2M. The ROM is attached
to the 660 bridge via the PCI_AD lines. This mode is required when using the Intel SIO.
ROM device writes and write-protect commands are supported. See the 660 Bridge User’s
Manual for more information.
The ROM device attaches to the 660 bridge by means of control lines and the PCI_AD[31:0]
lines. When a CPU bus master reads from the ROM, the bridge masters a BCR transaction,
during which it reads the ROM and returns the data to the CPU. CPU writes to the ROM
and ROM write-protection operations are also forwarded to the ROM device.
Although connected to the PCI_AD lines, the ROM is not a PCI agent. The ROM and the
PCI agents do not interfere with each other because the ROM is under bridge control, and
the bridge does not enable the ROM except during ROM cycles. The bridge accesses the
ROM by means of BCR transactions. Other PCI devices cannot read or write the ROM because they cannot generate BCR transactions.
6.4.1 CPU to ROM Read
At power-on, the 603/604 CPU comes up in BE Mode with the L1 cache disabled, and begins fetching instructions (using 8-byte single beat reads) at address FFF0 0100 (4G – 1M
+ 100h). The reference design logic also resets to BE mode.
The system ROM address space is from 4G – 2M to 4G. Since the size of the installed ROM
is less than 2M (512K), it is mirrored every 512K throughout the ROM space. Location 0
of the 512K ROM is mapped to CPU bus addresses 4G–2M, 4G–1.5M, 4G–1M, and
4G–.5M.
The Flash is located on the PCI bus physically but not logically, and is 8 bits wide. This requires the 660 Bridge to decode Flash address, run 8 cycles to PCI bus without activating
FRAME, accumulate the 8 single bytes of read data into an 8-byte group and generate a
TA# and an AACK# to complete the cycle. The CPU can also read the ROM using bursts,
but it receives the same 2 instructions from the ROM on each beat of the burst. For more
information, see the 660 Bridge User’s Manual.
Software can lock out the ROM using a 660 bridge BCR. When the CPU writes to any ROM
location while the ROM is locked out, the bridge signals normal transfer completion to the
CPU but does not write the data to the ROM. The CPU bus write to the locked flash bit in
the 660 bridge error status 2 register (bit 0 in index C5h) is set.
6.4.2 CPU to ROM Write
Writing to Flash is another very specialized cycle. Only one address (FFFF FFF0) is used
for writing data to Flash. The Flash address and data are both encoded into four bytes and
74

MPRH01TSU-02

Memory
written using a 4-byte write transfer. Eight byte and burst transfers to the ROM are not supported. See the 660 Bridge User’s Manual.
Writes to Flash may be performed in either BE or LE mode. The data byte swapper in the
660 Bridge is gated according to endian mode. Writes in BE mode occur in natural sequence. However, address unmunging in LE mode has no effect on the cycle because the
addresses are ignored. Therefore, software must reverse the byte significance of the data
and address encoded into the store instructions for LE mode writes to the ROM.
6.4.2.1
ROM Write Protection
Flash write protection must be implemented within software. Port FFFF FFF1 can be used
to lock out all Flash writes. Writing any data to this port address locks out all Flash writes
until the 660 Bridge is hardware reset. In addition, the Flash itself has means to permanently lock out changing certain sectors by writing control sequences. Consult the Flash Specification for details.

MPRH01TSU-02

75

Memory

76

MPRH01TSU-02

Endian Modes

Section 7
Endian Mode Considerations
The 603 and 604 normally operate with big endian (BE) byte significance. They have a
mode of operation designed to more efficiently process code written with little endian (LE)
byte significance. The reference design supports this ability, and can operate with BE operating systems such as AIXt or LE operating systems such as WindowsNTt.
When the system is in BE mode, data is stored in memory with BE ordering. When the system is in little endian mode, data is stored in memory with LE ordering. The 660 Bridge has
hardware to select the proper bytes in the memory and on the PCI, and to steer the data
to the correct CPU data lane.
A number represented in storage is said to be in big endian (BE) order when the most significant part of the number is in the lowest numbered storage location and less significant
parts are in successively higher numbered locations. AIX is an example of an operating
system that stores data in memory and on media in BE order. A number is said to be in little
endian (LE) order when it is stored with the order of bytes reversed from that of BE order.
WindowsNT is an example of an operating system that uses LE order in memory and on
media. The endian order of data never extends past an 8-byte group of storage.

7.1

What the 603/604 CPU Does

The 603/604 CPU assumes that the significance of memory is BE. When it operates in internal LE mode, it internally generates an effective address the same as the LE code would
generate. Since it assumes that the memory is stored with BE significance, it transforms
(munges) the three low order addresses when it activates the address pins. For example,
in the 1-byte transfer case, address 7 is munged to 0, 6 to 1, 5 to 2, and so on. The data
transfer occurs on the byte lanes identified by the address pins and transfer size (TSIZ) pins
in either BE or LE mode. The CPU ”shifts” the data to the correct byte lane(s). Note that
if the TSIZ is 1 and the address pins are 000b, then byte lane 0 (cpu data lines 0:7) must
be used for the data transfer in either mode. For a 4-byte transfer, to add 4 in LE mode, the
CPU munges the add to 0 and drives the data onto byte lanes 0 through 7.
The 603/604 performs the following munge operation in LE mode:
S
S
S
S

TSIZ = 8(0)=> None
TSIZ = 4 => 603 Internal Address 29:31 XOR with 100b => address at pins of CPU
TSIZ = 2 => 603 Internal Address 29:31 XOR with 110b => address at pins of CPU
TSIZ = 1 => 603 Internal Address 29:31 XOR with 111b => address at pins of CPU.

MPRH01TSU-02

77

Endian Modes

7.2

What the 660 Bridge Does

Data is stored in system memory in the same endian mode as the mode in which the CPU
operates. That is, the byte significance in memory is BE in BE mode and it is LE in LE mode.
Because of this, hardware is included in the 660 bridge to steer the data bytes to the correct
byte lanes according to mode, and to unmunge the addresses coming from the 603/604
CPU in LE mode. This unmunge merely applies the same transformation to the three low–
order address lines as the 603/604 CPU reversing the effect of the munge that occurs within
the 603/604 CPU.
The hardware cannot tell the endian mode of the CPU directly. There is a control bit located
in ISA I/O space (port 0092) that the CPU can write to in order to set the endian mode of
the motherboard. This signal is applied to the 660 Bridge, which performs the operations
shown in Table 19.
Table 19. Endian Mode Byte Lane Steering
CPU Byte Lane

BE Mode Connection

LE Mode Connection

CPU byte lane 0 (MSB)

Memory byte lane 0, PCI lane 0

Memory byte lane 7, PCI lane 7*

CPU byte lane 1

Memory byte lane 1, PCI lane 1

Memory Byte lane 6, PCI lane 6*

CPU byte lane 2

Memory byte lane 2, PCI lane 2

Memory byte lane 5, PCI lane 5*

CPU byte lane 3

Memory byte lane 3, PCI lane 3

Memory byte lane 4, PCI lane 4*

CPU byte lane 4

Memory byte lane 4, PCI lane 4*

Memory byte lane 3, PCI lane 3

CPU byte lane 5

Memory byte lane 5, PCI lane 5*

Memory byte lane 2, PCI lane 2

CPU byte lane 6

Memory byte lane 6, PCI lane 6*

Memory byte lane 1, PCI lane 1

CPU byte lane 7 (LSB)

Memory byte lane 7, PCI lane 7*

Memory byte lane 0, PCI lane 0

Note:
* In this table, PCI byte lane 3:0 refers to the data associated with PCI C/BE# 3:0 when
the third least significant bit of the target PCI address is 0b, as coded in the instruction.
PCI byte lane 7:4 refers to the data associated with PCI c/BE# 3:0 when this bit is b1b.
Since the reference design logic maintains the memory in BE mode during BE operation
and in LE mode during LE mode operation, no address translations are necessary in BE
mode. However, the CPU addresses must be unmunged in LE mode. This is accomplished
in the 660 Bridge by applying the same XOR function mentioned above whenever the CPU
accesses either PCI or memory. The munge effect nullifies the address translation that occurs within the CPU. For example, if the CPU executes a one-byte load coded to access
byte 0 of memory in LE mode, it will emit address 7. The 660 Bridge will change the external
address to 0. This is summarized in Table 20.
Table 20. Endian Mode 6–3/604 Address Translation
Mode

Effective Address at Memory or PCI

BE

No change (Same address as coded)

LE

Unmunge Address emitted (Transform to same address as coded)

When a PCI master accesses memory, the bridge chip set does not make any address
translations (in the address fields affected by endian mode) in either endian mode. In other
words, data is stored or fetched at the address presented on the PCI bus in either endian
mode.
78

MPRH01TSU-02

Endian Modes
The way the 660 Bridge implements the endian mode logic is conceptualized in Figure 15.

Memory

Address

CPU

Byte Swap
and
Unmunge

Data
1=DO,0=PASS

PCI Bus

Port 92

LE Mode Bit

Figure 15. Endian Mode Block Diagram

7.3

Bit Ordering Within Bytes

The BE/LE discussion in this section applies only to the ordering of bytes. The LE convention of numbering bits (the least significant bit having the lowest number within that byte)
is followed throughout the 603/604 Board schematics, including the PCI bit numbering. The
only exception is the CPU data and address buses; these are numbered with BE nomenclature. The 603/604 CPU buses are connected to the 663 buffer buses with the significance
of the bits maintained so that MSb connects to MSb and so on. Note that the pin numbering
convention on the 660 Bridge chip is BE at the CPU side and LE at the memory and PCI
sides.

7.4

Byte Swap Instructions

The Power PC architecture defines both word and half–word load/store instructions that
have byte swapping capability. Programmers will find these instructions valuable for dealing with the BE nature of this architecture. For example, if a 32–bit configuration register
of a typical LE PCI device is read in BE mode, the bytes will appear out of order unless the
”load word with byte swap” instruction is used. The byte swap instructions are:
S lhbrx (load half word byte–reverse indexed)
S lwbrx (load word byte–reverse indexed)
S sthbrx (store half word byte–reverse indexed)
S stwbrx (store word byte–reverse indexed)
The byte–reverse instructions should be used in BE mode to access LE devices and in LE
mode to access BE devices.

MPRH01TSU-02

79

Endian Modes

7.5

603/604 CPU Alignment Exceptions In LE Mode

The CPU does not support a number of instructions and data alignments in the LE mode
that it supports in BE mode. When it encounters an unsupportable situation, it takes an internal alignment exception (machine check) and does not produce an external bus cycle.
See the latest 603/604 CPU documentation for details. Examples include:
S LMW instruction
S STMW instruction
S Move assist instructions (LSWI, LSWX, STSWI, STWX)
S Unaligned loads and stores.

7.6

Single–Byte Transfers

Figure 16 is an example of byte write data a at address XXXX XXX0.

603/604

data
lane

3 low
addr
bits
T size

7
6
5
4
3
2
1
0

Big Endian
Swap Off
LSB
7
7
6
6
5
5
4
4
3
3
2
2
1
1
a
0++++++0
MSB
0

Memory

603/604

7
LSB
6
5
4
3
2
1
0 a MSB

data
lane

3 low
addr
bits
T size

addr
0

Unmunge
Off

OR

1

7
6
5
4
3
2
1
0

Little Endian
Swap On
LSB
a
7
7
6
6
5
5
4
4
3
3
2
2
1
1
0
0
MSB
7

Memory
7
MSB
6
5
4
3
2
1
0 a LSB
addr
0

Unmunge
On

OR

1

PCI

A/D 2

PCI

A/D 2

3210
0

a

3210
0

a

Figure 16. Example at Address XXXX XXX0

80

MPRH01TSU-02

Endian Modes

603/604
7
6
5
4
3
2
1
0

data
lane

Big Endian
Swap Off
LSB
7
7
6
6
5
5
4
4
3
3
a 2++++++2
1
1
0
0
MSB

3 low
addr
bits
T size

2

Memory
7
LSB
6
5
4
3
2a
1
0 MSB

Off

7
6
5
4
3
2
1
0

data
lane

Little Endian
Swap On
LSB
7
7
6
6
a
5
5
4
4
3
3
2
2
1
1
0
0
MSB

3 low
addr
bits
T size

addr
2

Unmunge

1

603/604

OR

5

Memory
MSB
7
6
5
4
3
2a
1
LSB
0
addr
2

Unmunge
On

OR

1

PCI

A/D 2

PCI

A/D 2

3210
a

0

3210
0

Figure 17. Example at Address XXXX XXX2

Figure 17 is an example of byte write data a at address XXXX XXX2.
For single byte accesses to memory in BE mode, Table 21 applies.
Table 21. Memory in BE Mode
603/

603/

603

663

604

604

BYTE

BYTE

MEM BYTE

CAS

add

LANE

LANE*

LANE

ACTIVE

A31 30 29
0

0

0

0

0

1

0

0

1

0

1

0

1

1

0

MSB

0

0

0

1

1

1

1

2

2

2

2

2

0

3

3

3

3

3

0

1

4

4

4

4

4

1

0

1

5

5

5

5

5

0

1

1

6

6

6

6

6

1

1

1

7

7

7

7

7

LSB

NOT MUNGED

SWAP
OFF

Note:
* – At the CPU side.

MPRH01TSU-02

81

NOT UNMUNGED

a

Endian Modes
For single byte accesses to memory in LE mode, Table 22 applies.
Table 22. Memory in LE Mode
603/

603/

603

604

604

BYTE

BYTE

MEM BYTE

add

LANE

LANE*

LANE

A31 30 29

663
CAS
ACTIVE

0

0

0

0

0 MSB 0

7

7

1

0

0

1

1

1

6

6

0

1

0

2

2

2

5

5

1

1

0

3

3

3

4

4

0

0

1

4

4

4

3

3

1

0

1

5

5

5

2

2

0

1

1

6

6

6

1

1

1

1

1

7

7 LSB 7

0

0

MUNGED

SWAP
ON

UNMUNGED

Note:
* – At the CPU side.

For single byte accesses to PCI in BE mode, Table 23 applies.
Table 23. PCI in BE Mode
603/

603/

603

604

604

BYTE BYTE

add

LANE LANE

A31 30 29

663*
PCI BYTE
LANE

A/D**

BE#

2 1 0

3 2 1 0
(0=active byte enable)

0

0

0

0

0 MSB 0

0

0 0 0

1 1 1 0

1

0

0

1

1

1

1

0 0 1

1 1 0 1

0

1

0

2

2

2

2

0 1 0

1 0 1 1

1

1

0

3

3

3

3

0 1 1

0 1 1 1

0

0

1

4

4

4

0

1 0 0

1 1 1 0

1

0

1

5

5

5

1

1 0 1

1 1 0 1

0

1

1

6

6

6

2

1 1 0

1 0 1 1

1

1

1

7

7 LSB 7

3

1 1 1

0 1 1 1

NOT MUNGED

SWAP
OFF

NOT UNMUNGED

Note:
* – At the CPU side.
**– AD[0:1] set to 00 for all PCI transactions except I/O cycles.

82

MPRH01TSU-02

Endian Modes
For single byte accesses to PCI in LE mode, Table 24 applies.
Table 24. PCI in LE Mode
603/

603/

603

663*

604

604

BYTE

BYTE

add

LANE

LANE

A31 30 29

PCI BYTE
LANE

A/D **

BE#

2 1 0

3 2 1 0
(0=active byte enable)

0

0

0

0

0 MSB 0

3

1 1 1

0 1 1 1

1

0

0

1

1

1

2

1 1 0

1 0 1 1

0

1

0

2

2

2

1

1 0 1

1 1 0 1

1

1

0

3

3

3

0

1 0 0

1 1 1 0

0

0

1

4

4

4

3

0 1 1

0 1 1 1

1

0

1

5

5

5

2

0 1 0

1 0 1 1

0

1

1

6

6

6

1

0 0 1

1 1 0 1

1

1

1

7

7 LSB 7

0

0 0 0

1 1 1 0

MUNGED

SWAP
ON

UNMUNGED

Notes:
* – At the CPU side.
**– AD[0:1] set to 00 for all PCI transactions except I/O cycles.

MPRH01TSU-02

83

Endian Modes

7.7

Two–Byte Transfers

Figure 18 gives an example of double byte write data ab at address XXXX XXX0.
603/604

data
lane

3 low
addr
bits
T size

7
6
5
4
3
2
1
0

Big Endian
Swap Off
LSB
7
7
6
6
5
5
4
4
3
3
2
2
b
1++++++1
a
0++++++0
MSB
0
2

Memory

603/604

7
LSB
6
5
4
3
2
1b
0 a MSB

data
lane

3 low
addr
bits
T size

addr
0,1

Unmunge
Off

OR

7
6
5
4
3
2
1
0

Little Endian
Swap On
LSB
b
7
7
a
6
6
5
5
4
4
3
3
2
2
1
1
0
0
MSB
6

Memory
7
MSB
6
5
4
3
2
1a
0 b LSB
addr
0,1

Unmunge
On

OR

2

PCI

A/D 2

PCI

A/D 2

3210
0

ba

3210
0

ab

Figure 18. Double Byte Write Data ab at Address XXXX XXX0

Table 25 and Table 26 illustrate all cases that can occur. The columns of Table 25 have
these meanings:
S The first column indicates target address (e.g. the address of the byte coded into
a store half–word instruction).
S The next two columns show the state of the address pins for BE mode.
S The next two columns show the state of the address pins for the same target data
when the machine is in LE mode.
S The remaining columns show the CASs and the PCI byte enables associated with
the target data.
S The notes indicate which combinations either do not occur at the 603 pins because
of internal exceptions, or are not supported externally.

84

MPRH01TSU-02

Endian Modes
For 2-byte transfers, Table 25 holds:
Table 25. Two Byte Transfer Information
PROG

BE MODE

LE MODE

BE OR LE

BE OR LE

BE OR LE

BE

(X or w 110)

Target

CAS# 0:7

PCI CBE#

Add a29:31

bytes

0

AD2 3210

TARG

603/604

ADDR

add

a29:31

0

0

000

6

110

0–1

0011 1111

1

1

001

7 E 111

1–2

E 1001 1111

2

2

010

4

100

2–3

1100 1111

3

3

011

5 E 101

3–4

E 1110 0111

4

4

100

2

010

4–5

1111 0011

5

5

101

3 E 011

5–6

E 1111 1001

6

6

110

0

000

6–7

1111 1100

7

N

NNN

1 E 001

NNN

E NNNN NNNN

7

0

1100

0 E 1001
0

0011

1 E PPPP
1

1100

1 E 1001
1

0011

N E NNNN

Notes:
N= not emitted by 60X because it crosses 8 bytes (transforms to 2 singles in BE, machine CH in LE)
P= not allowed on PCI (crosses 4 bytes)
E= causes exception (does not come out on 603/604 bus) in LE mode

Table 26 contains the same information as found in Table 25, but it is arranged to show the
CAS and PCI byte enables that activate as a function of the address presented at the pins
of the 603/604 and as a function of BE/LE mode.
Table 26. Rearranged 2-Byte Transfer Information
2 BYTE XFERS
60X ADDRESS PINS

BE

BE

LE

CAS#0:7

PCI CBE#

CAS#0:7

0

7

A2

3210

0

LE
7

PCI

CBE#

AD2

3210

0

000

0011

1111

0

1100

1111 1100

1

0011

1

001

1001

1111

0

1001

E NNNN NNNN

N

E NNNN

2

010

1100

1111

0

0011

1111 0011

1

1100

3

011

1110

0111

0

PPPP

E 1111 1001

1

E 1001

4

100

1111

0011

1

1100

1100 1111

0

0011

5

101

1111

1001

1

1001

E 1110 0111E

0

E PPPP

6

110

1111

1100

1

0011

0011 1111

0

1100

7

111

NNNN

NNNN

N

NNNN

E 1001 1111E

0

E 1001

Notes:
N= not emitted by 60X because it crosses 8 bytes (transforms to 2 singles in BE, machine CH in LE)
P= not allowed on PCI (crosses 4 bytes)
E= causes exception (does not come out on 603/604 bus) in LE mode

MPRH01TSU-02

85

Endian Modes

7.8

Four–Byte Transfers

Figure 19 gives an example of Word (4–BYTE) Write of 0a0b0c0dh AT ADDRESS XXXX
XXX4.

Big Endian
Swap Off

data
lane

3 low
addr
bits
T size

LSB
7 0d
6 0c
5 0b
4 0a
3
2
1
0

Memory

7
6
5
4
3
2
1
0

7
6
5
4
3
2
1
0

7
6
5
4
3
2
1
0

MSB
4
4

Little Endian
Swap On

603/604

Memory

LSB
++++
++++
++++
++++

603/604

d LSB
c
b
a

data
lane

MSB
3 low
addr
bits
T size

addr
4-7

Unmunge
Off

7
6
5
4
3
2
1
0

OR

0d
0c
0b
0a

7
6
5
4
3
2
1
0

7
6
5
4
3
2
1
0

7
6
5
4
3
2
1
0

MSB
0

data
A/D31

On

OR

4

PCI

3210

A/D 2

1

data
A/D31

A/D24

A/D24

0d
0 0 0 0 1 0 1 0

23

23

16

3210
1

0 0 0 0 1 1 0 1

0a

16
0b

0c
0 0 0 0 1 0 1 1

0 0 0 0 1 1 0 0
15

LSB

addr
4-7

Unmunge

PCI

A/D 2

a MSB
b
c
d

15

8

8
0c

0b
0 0 0 0 1 0 1 1

0 0 0 0 1 1 0 0

7

7

0

0
0d

0a
0 0 0 0 1 0 1 0

0 0 0 0 1 1 0 1

Figure 19. Word (4-Byte) Write of 0a0b0c0dh at Address XXXX XXX4

86

MPRH01TSU-02

Endian Modes
Table 27 and Table 28 illustrate the cases that can occur. The columns of Table 27 have
these meanings:
S The first column indicates the target address (e.g. the address of the byte coded into
a store word instruction).
S The next two columns show the state of the address pins for BE mode.
S The next two columns show the state of the address pins for the same target data
when the machine is in LE mode.
S The remaining columns show the CASs and the PCI byte enables associated with
the target data.
S The notes indicate which combinations either do not occur at the 603/604 pins because of internal exceptions, or are not supported externally.
For 4-byte transfers, Table 27 holds:
Table 27. 4-Byte Transfer Information
PROG

BE MODE

LE MODE

BE OR LE

BE OR LE

BE OR LE

TARG

603/604 BE

(X or w 100)

Target

CAS# 0:7

PCI CBE#

ADDR

add

add a29:31

bytes

0

a29:31

7

AD2 3210

0

0

000

4

100

0–3

0000 1111

1

1

001

5

E 101

1–4

E 1000 0111

0 E PPPP

2

2

010

6

E 110

2–5

E 1100 0011

0 E PPPP

3

3

011

7

E 111

3–6

E 1110 0001

1 E PPPP

4

4

100

0

000

4–7

1111 0000

1

5

5

NNN

1

E NNN

N–N

NNNN NNNN

1 E NNNN

6

6

NNN

2

E NNN

N–N

NNNN NNNN

1 E NNNN

7

7

NNN

3

E NNN

N–N

NNNN NNNN

1 E NNNN

Notes:
N= not emitted by 60X because it crosses 8 bytes (transformed into 2 bus cycles)
P= not allowed on PCI (crosses 4 bytes)
E= causes exception (does not come out on 603/604 bus) in LE mode

MPRH01TSU-02

87

0

0000

0000

Endian Modes
Table 28 contains the same information as found in Table 27, but it is arranged to show the
CAS and PCI byte enables that activate as a function of the address presented at the pins
of the 603/604 and as a function of BE/LE mode.
Rearranging Table 27 for 4-byte transfers:
Table 28. Rearranged 4–Byte Transfer Information
4 BYTE XFERS
60X ADDRESS PINS

BE

BE

CAS#0:7
0

7

LE

PCI CBE#

CAS#0:7

A2

3210

0

LE
PCI CBE#
7

AD2 3210

0

000

0000 1111

0

0000

1111 0000

0

1

001

1000 0111

0

PPPP

E NNNN NNNN

0 E NNNN

2

010

1100 0011

0

PPPP

E NNNN NNNN

0 E NNNN

3

011

1110 0001

0

PPPP

E NNNN NNNN

E NNNN

4

100

1111 0000

1

0000

0000 1111

5

101

NNNN NNNN

1

NNNN

E 1000 0111

1 E PPPP

6

110

NNNN NNNN

1

NNNN

E 1100 0011

1 E PPPP

7

111

NNNN NNNN

1

NNNN

E 1110 0001

1 E PPPP

1

0000

0000

Notes:
N= not emitted by 60X because it crosses 8 bytes (transformed into 2 bus cycles)
P= not allowed on PCI (crosses 4 bytes)
E= causes exception (does not come out on 603/604 bus) in LE mode
X= not supported in memory controller (crosses 4-byte boundary

7.9

Three byte Transfers

There are no explicit Load/Store three–byte instructions; however, three–byte transfers occur as a result of unaligned four–byte loads and stores as well as a result of move multiple
and string instructions.
The TSIZ=3 transfers with address pins = 0, 1, 2, 3, 4, or 5 may occur in BE. All of the other
TSIZ and address combinations produced by move multiple and string operations are the
same as those produced by aligned or unaligned word and half–word loads and stores.
Since move multiples, strings, and unaligned transfers cause machine checks in LE mode,
they are not of concern in the BE design.

88

MPRH01TSU-02

Endian Modes

7.10 Instruction Fetches and Endian Modes
Most instruction fetching is with cache on. Therefore memory is fetched eight bytes wide.
Figure 20 shows the instruction alignment.
Example: 8 byte instruction fetch I1=abcd, I2=efgh at address XXXX XXX0

603/604

data
lane

7
6
5
4
3
2
1
0

Big Endian
Swap Off
LSB
h
7
7
g
6
6
f
5
5
e
4
4
d
3
3
c
2
2
b
1
1
a
0
0

Memory

603/604

7 h
6g
5 f I2
4e
3d
2c
1 b I1
0a

data
lane

MSB
3 low
addr
bits
T size

0
8

Unmunge

3 low
addr
bits
T size

addr
0-7

Off

7
6
5
4
3
2
1
0

Little Endian
Swap On
LSB
d
7
7
c
6
6
b
5
5
a
4
4
h
3
3
g
2
2
f
1
1
e
0
0
MSB
0
8

Figure 20. Instruction Alignment Example

MPRH01TSU-02

89

Unmunge
no
effect

Memory
7a
6b
5c
4d
3e
2 f
1g
0h
addr
0-7

I2

I1

Endian Modes
It is possible, in rare cases, to fetch instructions with 4–byte aligned transfers when the
cache is turned off. In that case the 603/604 does not munge the address in LE mode. The
memory controller does not differentiate between instruction and data fetches, but the unmunger is ineffective because the memory is always read 8–byte wide, and data is presented on all 8–byte lanes. If the unmunger were used, the wrong instruction would be read.
The net result is illustrated in Figure 21.
Example: 4 byte instruction fetch, I2=efgh at address XXXX XXX4

603/604

data
lane

7
6
5
4
3
2
1
0

Big Endian
Swap Off
LSB
h
7
7
g
6
6
f
5
5
e
4
4
x
3
3
x
2
2
x
1
1
x
0
0

Memory

603/604

7 h
6g
5 f I2
4e
3x
2x
1 x I1
0x

data
lane

MSB
3 low
addr
bits
T size

4
4

Unmunge

3 low
addr
bits
T size

addr
0-7

Off

7
6
5
4
3
2
1
0

Little Endian
Swap Off
LSB
x
7
7
x
6
6
x
5
5
x
4
4
h
3
3
g
2
2
f
1
1
e
0
0
MSB
0

Unmunge

Memory
7e
6 f
5g
4h
3x
2x
1 x
0 x

I2

I1

addr
0-7

On
4

Figure 21. Wrong Instruction Read When Unmunger is used

7.11 Changing BE/LE Mode
There are two BE/LE mode controls. One is inside the 603/604 CPU and the other is a register bit on the motherboard. The 603/604 CPU interior mode is not visible to the motherboard
hardware. The BE mode bit referred to in this document is the register bit on the motherboard. It is a bit in I/O space which is memory mapped just like other I/O registers. It defaults
to BE mode.
The 603/604 CPU always powers up in the BE mode and begins fetching to fill its cache.
Consequently, at least the first of the ROM code must be BE code. It is beyond the scope
of this document to define how the system will know to switch to LE mode. However, great
care must be made during the switch in order to synchronize the internal and external mode
bits, to flush all caches, and to avoid executing extraneous code.
The following process switches the system from BE to LE mode when used in this system:
1.
2.
3.
4.

Disable L1 caching
Disable L2 caching
Flush all system caches
Turn off interrupts immediately after a timer tick so no timer interrupts will occur
during the next set of cycles
5. Mask all interrupts
90

MPRH01TSU-02

Endian Modes
6. Set the CPU state and the motherboard to LE (see Figure 22). Note that CPU is
now in LE mode. All instructions must be in LE order.
7. Put interrupt handlers and CPU data structures in LE format
8. Enable caches
9. Enable Interrupts
10. Start the LE operating system initialization
Figure 22 shows the instruction stream to switch endian modes.
x

mfspr

R2,1008

;Load the HDO register

;Instructions to set the Little–Endian bit in R2
0

sync

4

sync

8

sync

C

mtspr

10

sync

14

sync

18

sync

1c

sync

20

Store to external Endian control port (X8000 0092)

1008,R2

;Moves to HID0 register

;The above instruction must be on a double word boundary
;So the following instruction is executed first (due to pipeline)
24

eieio

; To this point all instructions are in Big Endian format
; The following instructions look the same in either Endian mode
28

X38010138

2C

X38010138

...

;Enough of these instructions must be executed

...

;to guarantee the above store has occurred.

;
;before any memory or I/O cycles are listed.
xx

X38010138
Figure 22. Instruction Stream to Switch Endian Modes

MPRH01TSU-02

91

Endian Modes

7.12 Summary of Bi–Endian Operation and Notes
S When the 603/604 CPU is in BE mode, the memory is in BE mode, and data flowing
on the PCI is in BE order so that it is recorded on the media in BE order. Byte 0 is
the most significant byte.
S When the 603/604 processor is in LE mode, the memory is in LE mode, and data
flowing on the PCI is in LE order, so that it is recorded on the media in LE order. Byte
0 is the least significant byte.
S The PCI bus is addressed in the same manner as memory is when the 603/604 CPU
runs a cycle. The unmunging in LE mode changes the effective low–order address
bits (the byte enables and A/D 2). On all but I/O cycles, the two low–order A/D lines
are set to zero. On PCI I/O cycles, A/D 1,0 are also transformed by the unmunge
operation
S No translations are made when PCI accesses memory so that the byte with address
0 on the PCI flows to byte 0 in memory –– 1 to 1, 2 to 2, and so on. For example,
if BE0# and BE1# are active and A/D 2 is a 0, then memory byte lanes 0 and 1 are
addressed (cas 0 and cas 1 active on writes).
S Note that the LE devices which interpret data structures in the memory require that
their control data be arranged in LE order even in BE mode. For example, SCSI
scripts in memory must always be arranged in LE order because that is what the
device expects.
S Devices such as video may require the bytes to be swapped unless these devices
have byte swap capability.

92

MPRH01TSU-02

Exceptions

Section 8
Exceptions
The reference design handles two classes of exceptions, interrupts and errors. Interrupts
are handled primarily by the interrupt controller in the ISA bridge, the 660 bridge, the CPU,
and the firmware. Errors are handled primarily by the 660 bridge, the CPU, and the firmware.

8.1

Interrupts

8.1.1 System Interrupt Handler
There are two 8259 type interrupt controllers located in the ISA bridge. These controllers
receive and prioritize reference design interrupts, which can be asserted by motherboard
logic, PCI devices, or ISA devices. The interrupt controller then asserts an interrupt to the
660 bridge.
The interrupt controller handles both ISA and PCI interrupts using the correct protocols,
under software control. Much of the operation of the interrupt controller is programmable.
See the SIO data book for more information.

MPRH01TSU-02

93

Exceptions
8.1.2

Interrupt Handling
INT_CPU#
ISA
Interrupts
PCI
Interrupts

INT_REQ
ISA
Bridge

NMI_REQ

660
Bridge

CPU
CPU Bus

PCI Bus
Figure 23. Interrupt Handling

As shown in Figure 23, the reference design interrupts are routed to the interrupt controller
located inside the ISA bridge. When a device signals an interrupt (which is not masked in
the interrupt controller):
1. The ISA bridge asserts either INT_REQ or NMI_REQ to the 660 bridge.
2. The 660 bridge asserts INT_CPU# to the CPU.
3. The CPU recognizes the interrupt signal (INT#) immediately (or as soon as the
MSR(EE) interrupt enable bit in the CPU is set to 1), saves its state, and then takes a
precise external interrupt exception, branching to either 500h or FFF0 0500h, depending upon the Exception Prefix (EP) bit in the MSR. The MSR(EE) bit is automatically
set to 0 at this time.
4. The code at the vector location requests a single-byte read of memory address BFFF
FFF0h.
5. In response to the read, the 660 bridge arbitrates for the PCI bus and then generates
an interrupt acknowledge transaction on the PCI bus.
6. The ISA bridge decodes and claims the PCI interrupt acknowledge transaction, and
returns the 8-bit vector which has been preprogrammed for the active interrupt, and
then negates the interrupt output (whichever of INT_REQ or NMI_REQ that it asserted).
7. The 660 bridge accepts the interrupt vector on the PCI bus and returns it to the CPU.
If the ISA bridge signalled the interrupt via INT_REQ, the 660 bridge asserts TA# to
terminate the CPU transfer normally. However, if the interrupt was signalled via
NMI_REQ, the 660 bridge terminates the CPU transfer with TEA#.
Since the CPU does not require that the interrupt signal (INT_CPU#) be deactivated between interrupts, another interrupt is allowed to occur as soon as software sets the
MSR(EE) bit back to 1. For this reason, software should enable interrupts as soon as possible after receiving the vector. Note that the load instruction that fetches the interrupt vector
is subject to out-of-order execution in the same way as any other load instruction. After
servicing the interrupt, execute a return from interrupt (RFI) instruction to return to the program that was interrupted. For more information on interrupts, see the Exceptions chapters
of the PowerPC 603 User’s Manual and of the PowerPC 604 User’s Manual.
Note that other PCI bus masters can initiate interrupt acknowledge transactions, but this
may have unpredictable effects.
94

MPRH01TSU-02

Exceptions
8.1.3 Interrupt Assignments
In general, program ISA interrupts are edge sensitive. Program PCI interrupts as level sensitive. Interrupts are assigned to priority levels per ISA conventions. Table 29 shows interrupt assignments. IRQ[0:7] connect to the master controller, and IRQ[8:15] connect to the
cascaded controller. Figure 24 shows the connection of the PCI interrupts.
Table 29. Mapping of PCI Memory Space, Part 1
SIO IRQ # Connects to Priority Assignment or (Comment)
0
no pin
1
Timer 1 Counter 0 (Internal to SIO).
1
Sys I/O EPLD
2
Keyboard
2
no pin
(3-10) Cascade from controller 2
3
ISA IRQ3
11
(COM 2 or COM 4)
4
ISA IRQ4
12
(COM 1 or COM 3)
5
ISA IRQ5
13
(Parallel LPT 1 or 2)
6
ISA IRQ6
14
(Floppy)
7
ISA IRQ7
15
(Parallel LPT 2 or 3)
8#
RTC
3
TOD (aka Real Time Clock)
9
ISA IRQ9
4
10
ISA IRQ10
5
(Audio)
11
ISA IRQ11
6
12/M
ISA IRQ12
7
(Mouse)
13/FERR
—
8
Pulled up. Also see section 13, Errata.
14
ISA IRQ14
9
(IDE)
15
ISA IRQ15
10
PIRQ2#
PCI_INTC#
See Figure 24.
PIRQ1#
PCI_INTB#
See Figure 24.
PIRQ0#
PCI_INTA#
See Figure 24. Also see section 13, Errata.
Vdd

Vdd

ISA
Bridge

PCI Slot 1, J26
INTA#
PIRQ1#
INTB#
PCI_INTB#
INTC#
INTD#
Vdd

ISA
Bridge

PCI Slot 0, J25
INTA#
PCI_INTA# PIRQ0#
INTB#
INTC#
Also see section 13, Errata.
INTD#

PCI Slot 2, J27
INTA#
PIRQ2#
INTB#
PCI_INTC#
INTC#
INTD#

Figure 24. PCI Interrupt Connections

8.1.4 Scatter-Gather Interrupts
Where possible, set up the scatter-gather function to use the ISA bridge end of process
(EOP output) indicator for the termination of ISA bus DMA in which scatter-gather is
employed. The reference design is initially configured to use this scheme. The EOP signal
from the ISA bridge is used as the terminal count (ISA_TC) signal on the ISA bus.
MPRH01TSU-02

95

Exceptions

8.2

Error Handling

There are two methods which the reference design (660 bridge) uses to report errors to the
CPU, the TEA# method, and the MCP# method.
Errors that are detected while the CPU is running a cycle that can be terminated immediately are reported using TEA#. Errors reported in this way are a direct result of the CPU transfer
that is currently in progress. For example, when the 660 detects a transfer size error, it terminates the CPU transfer with TEA# instead of with TA#.
Errors that are detected while a CPU transfer is not in progress, and errors that occur because of a CPU transfer but which are detected too late to be reported using TEA#, and
errors that are not a direct result of the current CPU transfer, are reported using MCP#. For
example, parity errors occuring while a PCI bus master is accessing memory are reported
using MCP#.
For more information on error handing, see the 660 Bridge User’s Manual.
8.2.1 Data Error Checking
The reference design is initially configured to use parity, and this documentation reflects
that configuration. However, the 660 bridge can be programmed to execute an error checking and correction (ECC) algorithm on the memory data, generating ECC check bits during
memory writes, and checking-correcting the data during memory reads. ECC can be implemented using normal parity DRAM. See the 660 Bridge User’s Manual for more information.
Note that for each memory read operation, eight bytes of memory are read, and parity on
eight bytes is checked regardless of the transfer size. Therefore, all of memory must be
initialized (at least up to the end of any cache line that can be accessed).
The reference design does not generate or check CPU bus address parity.
8.2.1.1
CPU to Memory Writes
During CPU to memory writes, the CPU drives data parity information onto the CPU data
bus. Correct parity is then generated in the 660 and written to DRAM memory along with
the data. The L2 SRAM is updated (when required) with the data and the parity information
that the CPU drove onto the CPU data bus.
During CPU to memory writes, the 660 bridge checks the data parity sourced by the CPU,
and normally reports any detected parity errors via TEA#.
8.2.1.2
CPU to Memory Reads
When the CPU reads from memory, the data and accompanying parity information can
come from either the L2 SRAM or from DRAM memory. If the data is sourced from the L2,
the parity information also comes from the L2.
If the data is sourced by memory, the parity information also comes from memory. The L2
SRAM is updated (when required) using the data and parity from memory.
During CPU to memory reads, the 660 bridge samples the DPE# output of the CPU to determine parity errors, and reports them back to the CPU via MCP#. The particular memory
read data beat will be terminated normally with TA#.
96

MPRH01TSU-02

Exceptions
8.2.1.3
PCI to Memory Parity Errors
During PCI to memory writes, the 660 bridge generates the data parity that is written into
DRAM memory. The bridge also checks the parity of the data, and asserts PCI_PERR# if
it detects a data parity error.
During PCI to memory reads, the 660 bridge checks the parity of the memory data, and then
generates the data parity that is driven onto the PCI bus. If there is a parity error in the data/
parity returned to the 660 bridge from the DRAM, the bridge drives PCI_PAR incorrectly
to propagate the parity error (and also reports the error to the CPU via MCP#). The data
beat with the bad parity is not target aborted because doing so would slow all data beats
for one PCI clock (TRDY# is generated before the data is known good). However, if the
agent is bursting and there is another transfer in the burst, the next cycle is stopped with
target abort protocol.
During PCI to memory reads, the 660 bridge also samples the PCI_PERR# signal, which
other agents can be programmed to activate when they detect a PCI parity error.
8.2.1.4
CPU to PCI Transaction Data Parity Errors
During CPU to PCI writes, the 660 bridge sources the PCI parity information, and monitors
PCI_PERR#, which other agents can be programmed to activate when they detect a PCI
parity error.
During CPU to PCI reads, the 660 bridge checks the data parity and asserts PCI_PERR#
if it detects a data parity error.
8.2.2 Illegal CPU cycles
Whenever a CPU transfer which is not supported for memory or for the PCI is detected,
the cycle is terminated with a TEA# and the illegal transfer register is set. No memory or
PCI cycle is initiated. Read data returned is all 1’s. The CPU address is captured in the Error
Address Register.
8.2.3 SERR, I/O Channel Check, and NMI Logic
The PCI bus defines a signal called SERR# which any agent can pulse. This signal is to
report error events within the devices, not bus parity errors. The signal is wired to the ISA
bus bridge. The ISA bus signal IOCHCK is also wired to the ISA bridge. If either of these
lines activate, the ISA bridge asserts NMI to the 660 bridge unless the condition is masked
by a register within the ISA bridge. The NMI signal causes the 660 bridge to generate an
interrupt to the CPU, and to assert MCP# to the CPU. The ISA bridge contains status registers to identify the NMI source. Software may interrogate the ISA bridge and other devices
to determine the source of the error.
No address is associated with this type of error; therefore, the contents of the error address
register are not defined.
8.2.4 Out of Bounds PCI Memory Accesses
If a PCI bus master runs a cycle to a system memory address above the top of physical
memory, no one will respond, and the initiator master aborts the cycle. The initiating bus
master must be programmed to notify the system of master aborts as needed. The system
logic does not notify the CPU.

MPRH01TSU-02

97

Exceptions
8.2.5 No Response on CPU to PCI Cycles – Master Abort
The 660 bridge master aborts if no agent responds with DEVSEL# within eight clocks after
the start of a CPU to PCI cycle. The cycle is ended with a TEA# response to the CPU, all
1’s data is returned on reads, the Illegal Transfer Error register is set, and the Error Address
register is held.
The 660 bridge also checks for bus hung conditions. If a CPU to PCI cycle does not terminate within approximately 60 usec after the PCI is owned by the CPU, the cycle is terminated with TEA#. This is true for all CPU to PCI transaction types except configuration
transactions. This feature may be disabled via a 660 bridge control register.
In the case of configuration cycles that do not receive a DEVSEL# (no device present at
that address), the PCI cycle is master aborted, and TA# (normal response) is returned.
Write data is thrown away and all 1’s are returned on read cycles. No error register is set
and no address is captured in the error address register.
8.2.6 CPU to PCI Cycles That Are Target Aborted
When any CPU to PCI cycle of any sort receives a Target Abort from its target PCI agent,
the CPU cycle is terminated with a TEA#, the Error Address register is held, and the Illegal
Transfer Error register is set.
8.2.7 Error Status Registers
Error status registers in the 660 Bridge may be read to determine the types of outstanding
errors. Errors are not accumulated while an error is outstanding; however, there will be one
TEA# or MCP# for each error that occurs. For example, if an illegal transfer error causes
a TEA#, a memory parity error can occur while the CPU is processing the code that handles
the TEA#. The second error can occur before the error status registers are read. If so, then
the second error status is not registered, but the MCP# from the memory parity error is asserted.
8.2.8 Reporting Error Addresses
One register holds the address for any memory parity error, multi-bit ECC error, or illegal
transfer error when either the CPU or a PCI agent reads memory. It is also loaded on CPU
cycles that cause the Illegal Transfer Error register to be set.
See the System Error Address BCR information in the Bridge Control Registers section of
The IBM27-82660 PowerPC to PCI Bridge User’s Manual for more information.
8.2.9 Errant Masters
Either PCI or ISA masters can access certain motherboard and ISA bridge registers. For
example, various control registers such as the I/O Map Type register, the BE/LE mode bit,
the Memory Control registers, etc. are accessible. Faulty code in the PCI or ISA masters
can defeat password security, read the NVRAM, and cause the system to crash without
recovery. Take care when writing device drivers to prevent these events.

98

MPRH01TSU-02

Exceptions
8.2.10 Special Events Not Reported as Errors
S A PCI to memory cycle at any memory address above that programmed into the top
of memory register.
The 660 bridge ignores this cycle and the initiator master aborts. No data is written into
system memory on writes, and the data returned on reads is indeterminate. The bus
master must be programmed to respond to a target abort by alerting the host.
S CPU to PCI configuration cycles to which no device responds with a DEVSEL# signal within 8 clocks (no device at this address)
The data returned on a read cycle is all 1’s and write data is discarded. This allows software to scan the PCI at all possible configuration addresses, and it is also consistent
with the PCI specification.
S A CPU read of the IACK address having a transfer size other than 1, or having other
than 4–byte alignment.
These conditions return indeterminate data. The ISA bridge requires the byte enables,
CBE#3:0, to be 1110 in order to place the data on the correct byte lane (0). Accesses
other than one byte at the address BFFF FFF0h are undefined.
S A read of the IACK address when no interrupt is pending
A DEFAULT 7 vector is returned in this case. This is the same vector that is returned
on spurious interrupts.
S Parity error in Flash/ROM.
Parity is not stored in the Flash ROM. Therefore the memory parity error signal and the
DPE signal are ignored during ROM reads. The Flash or ROM should include CRC with
software checking to insure integrity.
S Write to Flash with TSIZ other than 4.
This will cause indeterminate data to be written into the Flash at an indeterminate address.
S Caching ROM space.
An L1 or CB–L2 cast out will cause indeterminate results.
S Running any cycle to the PCI configuration space with an undefined address.
Some of these could potentially cause damage. See the warning under the PCI configuration cycle section.
S Accessing any ISA device with the wrong data size for that device.
Indeterminate results will occur.

MPRH01TSU-02

99

Exceptions

100

MPRH01TSU-02

Setup

Section 9
System Setup and Initialization
9.1

CPU Initialization

The 603/604 CPU exits the reset state with the L1 cache disabled and bus error checking
disabled.
All memory pages 2G to 4G must be marked as non-cacheable.
The Segment Register T bit, bit 0, defaults to 0 which is the normal storage access mode.
It must be left in this state for the hardware to function. Direct store (PIO) segments are not
supported.
Set the bit that controls ARTRY# negation, HID0[7], to 0 to enable the precharge of
ARTRY#. It may be necessary set HID0[7] to 1 to disable the precharge of ARTRY# for reference design configurations having a CPU bus agent (such as an added L2) that drives
the ARTRY# line. Software must set this bit before allowing any CPU bus traffic to which
the CPU agent might respond. Note that PCI to memory transactions cause the 660 bridge
to broadcast snoop operations on the CPU bus.
HID0 bit 0, Master Checkstop Enable, defaults to 1 which is the enabled state. Leave it in
this state so that checkstops can occur.
Reference design errors are reported through the 660 bridge by way of the TEA# and
MCP# pins. Because of this, the bus error checking in the CPU must be disabled by setting
HID0 bits 2 and 3 to zero (in the 604, enable L1 cache parity checking by setting HID0(1)
to 1).

MPRH01TSU-02

101

Setup

9.2

660 Bridge Initialization

Before DRAM memory operations can begin, the software must:
1. Read the SIMM presence detect and SIMM type registers.
2. Set up and check the memory-related registers in the 660 bridge (see the 660
Bridge User’s Manual).
3. Program the timer in the ISA bridge register which controls ISA refresh timing. In
SIO compatible bridges it should be programmed to operate in Mode 2 with an
interval of approximately 15 usec.
4. Make sure 200 usec has elapsed since starting the refresh timer so that sufficient
refresh cycles have occurred to properly start the memory. This will be hidden if
approximately 120 Flash accesses occur after the timer is started and before the
memory initialization starts.
5. Initialize all of memory so that all parity bits are properly set. (The CPU may
cache unnecessary data; hence, all of memory must be initialized.) The 660
bridge does not require reconfiguration when port 4Dh in the ISA bridge is utilized to reset the native I/O and the ISA slots.

102

MPRH01TSU-02

Setup

9.3

ISA Bridge (SIO) Initialization

The reference design uses an Intel 82378ZB SIO as the ISA bridge. The following information applies to SIO compatible ISA bridges.
The SIO chip should be configured prior to any other PCI bus agent. The SIO PCI arbiter
is automatically enabled upon power-on reset. During power-on reset, the SIO drives the
A/D(31:0), C/BE#(3:0), and par signals on the PCI bus.
The system I/O EPLD uses the decode circuits in the SIO that produce the signals ECSADDR[2:0] and UBUSCOE# to decode the motherboard register addresses. For this reason, utility bus A and B decode registers must be initialized as shown in Table 30.
The ISA clock divisor must be set as indicated prior to running any CPU to PCI transactions.
If the configuration information is stored in Flash, this should pose no problem.
The SIO must be programmed so that interval timer 1 operates in mode 2 with a period of
approximately 15 microseconds. This timer controls the ISA refresh interval. It must be programmed at least 200 microseconds before any access to ISA DRAM is attempted.
PCI memory write cycles destined for ISA can use a 32-bit posted write buffer in the SIO.
Bit 2 of the PCI control register controls the enabeling of the posted write buffer. The default
(power-on reset) state for the posted write buffer is disabled. It is required that the posted
write buffer be enabled.
Note that PCI burst transactions are not supported by the SIO. For burst transactions, the
SIO will always target abort after the first data phase. The system will not allow the CPU
to burst to the SIO (or any other PCI agent). No PCI master should be programmed to attempt burst transactions to the SIO.
The SIO defaults (after power-on reset) to the slow sampling point (bits 4:3 of the PCI Control Register) for its subtractive decode. Of the three choices for the sampling point: slow
(5 PCI cycles), typical (4 PCI cycles) and fast (3 PCI cycles), one should be chosen that
is one clock after the slowest I/O device on the PCI bus. If the PCI agents are all memory
mapped above 16M Byte and all I/O mapped above 64K, then the fast sampling point for
the subtractive decode can be chosen. This insures that no other PCI agent except the SIO
will claim these addresses. Configure PCI agents in this manner to improve performance.
The SIO automatically inserts a 4 ISA clock cycle delay between PCI originated back-toback 8 and 16 bit I/O cycles to the ISA bus. In addition, the ISA Controller Recovery Timer
Register (configuration register, address offset=4Ch) enables a number of additional ISA
clock cycles of delay to be inserted between these types of back-to-back I/O cycles. The
ISA Controller Recovery Timer Register defaults (after power-on reset) to 2 additional ISA
clock cycles of delay, making the total delay equal to 6 ISA clock cycles, for both the 8 and
16 bit I/O recovery times. Since none of the native I/O devices on the reference design require such long recovery times, the additional cycles specified by the ISA Controller Timer
Register can be disabled. If an ISA card requiring a long recovery time is supported, the
driver should insure that the recovery time is met.
Disable scatter/gather mode and GAT mode.
Do not attempt to access DMA channel 4 address and byte count registers.
MPRH01TSU-02

103

Setup
Always enable the ISA master and DMA buffers. In order to isolate slow ISA Bus I/O devices
from the PCI bus, the DMA controller uses the DMA/ISA master Line Buffer. This buffer can
operate in single transaction or in 8-byte mode. Bits 0-LE/7-BE and 1-LE/6-BE of the PCI
Control register configure the line buffer for DMA and ISA masters separately. It is required
that the 8-byte mode be enabled for both (Bits = 1,1).
The registers in Table 30 must be set in order for the reference design I/O hardware to operate properly. Vendors use LE bit nomenclature, and nomenclature within CPU registers is
BE.
Table 30. Summary of SIO Register Setup (Configuration Address = 8080 08xx)
Register

Addr Bit

Set Reset DESCRIPTION
To Value

PCI Control Register

40h

1

0

Enable PCI Memory Posted Write Buffer.

PCI Control Register

40h

1

0

Enable ISA Master Line buffer.

PCI Control Register

40h

1

0

Enable DMA Line Buffer

PCI Arbiter Control (Config/
PCI)

41h

0

0

Disable GAT (Guaranteed Access Time
Mode). Note: GAT does not work in SIO.

ISA Clock Divisor (Config/
PCI)

4Dh

0

0

Disable Coprocessor Error Support.

ISA Clock Divisor (Config/
PCI)

4Dh

1

0

Enable IRQ12/M Mouse Support.

ISA Clock Divisor (Config/
PCI)

4Dh

2-LE
5-BE
1-LE
6-BE
0-LE
7-BE
0-LE
7-BE
5-LE
2-BE
4-LE
3-BE
3-LE
4-BE

*

0

ISA Clock Divisor (Config/
PCI)

4Dh

2:0-LE
5:7-BE

*

0

Utility Bus Chip Select A
(Config/PCI)

4Eh

1

0

Utility Bus Chip Select A
(Config/PCI)

4Eh

1

1

Enable keyboard addresses (60h, 62h,
64h, 66h).

Utility Bus Chip Select A
(Config/PCI)

4Eh

1

1

Enable TOD Addresses (70h, 71h).

Utility Bus Chip Select B
(Config/PCI)

4Fh

1

0

Enable access to the motherboard registers in the 0800-08FF address range.

Utility Bus Chip Select B
(Config/PCI)

4Fh

1

1

Enable Port 92h access.

Utility Bus Chip Select B
(Config/PCI)

4Fh

4-LE
4-BE
1-LE
6-BE
0-LE
7-BE
7-LE
0-BE
6-LE
1-BE
5:4-LE
2:3-BE

* This bit should be set to 1 before changing or loading the PCI ISA Clock Divisor
value. Setting this bit to 1 will assert the
RSTDVR signal (which resets the System
I/O EPLD and any devices on the ISA bus
slots). All these devices will require reconfiguration after this bit has been asserted.
Software must guarantee that RSTDVR be
asserted for a minimum of 1 ms after the
clock divisor value is set.
* Set this field to 000b (divisor = 4). (If PCI
clock is slower than 33 MHz, then this field
would be 001b (divisor=3).
Disable generation of ECSADDR(2:0) and
UBUSOE# for the IDE and Floppy decode.

11

00

Disable generation of default address for
Parallel Port.

104

MPRH01TSU-02

Setup
Table 30. Summary of SIO Register Setup (Configuration Address = 8080 08xx) (Continued)
Register

Addr Bit

Set Reset DESCRIPTION
To Value

Utility Bus Chip Select B
(Config/PCI)

4Fh

11

00

Disable generation of default address for
Serial Port B.

Utility Bus Chip Select B
(Config/PCI)

4Fh

11

00

Disable generation of default address for
Serial Port A.

Interrupt Controller 1 ICW1 (I/O /PCI)

20h

0

x

Set Interrupt Controller 1 to edge triggered
mode.

Interrupt Controller 1 ICW1 (I/O /PCI)

20h

0

x

Set Interrupt Controller 1 to cascade
mode.

Interrupt Controller 2 ICW1 (I/O /PCI)

A0h

0

x

Set Interrupt Controller 2 to edge triggered
mode.

Interrupt Controller 2 ICW1 (I/O /PCI)

A0h

0

x

Set Interrupt Controller 2 to cascade
mode.

0

0

IOCHK# NMI enabled.

0

0

PCI SERR# NMI enabled.

0

1

NMI interrupt enabled.

0

0

DACK# Assert Level set to low.

0

0

DREQ Sense Level set to high.

NMI Status and Control (I/O 61h
/PCI)
NMI Status and Control (I/O 61h
/PCI)
NMI Enable and TOD Address (I/O /PCI)

70h

DMA Command (I/O /PCI)

08h,
D0h

DMA Command (I/O /PCI)

08h,
D0h

MPRH01TSU-02

3:2-LE
4:5-BE
1:0-LE
6:7-BE
3-LE
4-BE
1-LE
6-BE
3-LE
4-BE
1-LE
6-BE
3-LE
4-BE
2-LE
5-BE
7-LE
0-BE
7-LE
0-BE
6-LE
1-BE

105

Setup
9.3.1

Summary of SIO Configuration Registers
Table 31. Summary of SIO Configuration Registers

Address

Description

Type

Reset
Value

8080 0800 Vendor Identification
R/O
86h
8080 0801 Vendor Identification
R/O
80
8080 0802 Device Identification
R/O
84
8080 0803 Device Identification
R/O
04
8080 0804 Command
R/W
07
8080 0805 Command
R/W
00
8080 0806 Device Status
R/W
00
8080 0807 Device Status
R/W
02
8080 0808 Revision Identification
R/W
00
8080 0840 PCI Control
R/W
20
8080 0841 PCI Arbiter Control
R/W
00
8080 0842 PCI Arbiter Priority Control
R/W
04
8080 0843 PCI Arbiter Priority Control Extension
R/W
00
8080 0844 MEMCS# Control
R/W
00
8080 0845 MEMCS# Bottom of Hole
R/W
10
8080 0846 MEMCS# Top of Hole
R/W
0F
8080 0847 MEMCS# Top of Memory
R/W
00
8080 0848 ISA Address Decoder Control
R/W
01
8080 0849 ISA Address Decoder ROM Block
R/W
00
8080 084A ISA Address Bottom of Hole
R/W
10
8080 084B ISA Address Top of Hole
R/W
0F
8080 084C ISA Controller Recovery Timer
R/W
56
8080 084D ISA Clock Divisor
R/W
40
8080 084E Utility Bus Chip Select A
R/W
07
8080 084F Utility Bus Chip Select B
R/W
4F
8080 0854 MEMCS# Attribute Register #1
R/W
00
8080 0855 MEMCS# Attribute Register #2
R/W
00
8080 0856 MEMCS# Attribute Register #3
R/W
00
8080 0857 Scatter/Gather Relocation Base
R/W
04
8080 0860 PIRQ Route Control 0
R/W
80
8080 0861 PIRQ Route Control 1
R/W
80
8080 0862 PIRQ Route Control 2
R/W
80
8080 0863 PIRQ Route Control 3 (unused)
R/W
80
8080 0880 BIOS Timer Base Address
R/W
78
8080 0881 BIOS Timer Base Address
R/W
00
* If the entry in this column is blank, then the boot firmware does not write to this register.

106

Set
To *

0F
00

21
00
04
00
00
10
0F
00
F1
00
10
0F
56
10
07
FF

0F
0F
80
80

MPRH01TSU-02

Setup

9.4

PCI Configuration Scan

The reference design enables the software to implement a scan to determine the complement of PCI devices present. This is because the system returns all ones rather than an
error when no PCI device responds to initialization cycles. The software may read each
possible PCI device ID to determine devices present.
Table 32. Configuration Address Assignments
Device

IDSEL Line

60X Address*

PCI Address

ISA bus bridge (SIO)
PCI Slot 1
PCI Slot 2
PCI Slot 3

A/D 11
A/D 12
A/D 13
A/D 14

8080 08XXh
8080 10XXh
8080 20XXh
8080 40XXh

080 08XX
080 10XX
080 20XX
080 40XX

Note: This address is independent of contiguous I/O mode.
Software must use only the addresses specified. Using any addresses that causes more
than one IDSEL to be asserted (high) can cause bus contention, because multiple PCI
agents will be selected.
9.4.1 Multi-function Adaptors
The 660 Bridge supports multi-function adapters. It passes the address of the load or store
instruction that causes PCI configuration cycle unmodified (except the three low-order bits
are unmunged in little endian mode and the two low-order address bits are set to zero in
either endian mode). Therefore, addresses may be selected with non-zero CPU address
bits (21:23)—corresponding to PCI bits (10:8)—to configure multi-function adaptors. For
example, to configure device 3 in slot 1 use address 80C0 03XXh. To configure device 7
in slot 2 use address 8084 07XXh.
9.4.2 PCI to PCI Bridges
The 660 bridge supports both Type 0 and Type 1 configuration cycles.

MPRH01TSU-02

107

Setup

9.5
9.5.1

Reference Design Combined Register Listing
Direct Access Registers

Table 33 contains a summary listing of the registers that are physically located in the reference design motherboard. These registers are in general accessed using single CPU
transfers. There is an additional set of registers (see Table 34) located in the 660 bridge,
which are accessed using pairs of CPU transfers.
Table 33. Combined Register Listing
ISA
Port
0000
0001
0002
0003
0004
0005
0006
0007
0008
0009
000A
000B
000C
000D
000E
000F
0020
0021
0040
0041
0042
0043
0060
0061
0062
0064
0066
0070
0071
0074
0075
0077
0078
0079

Contiguous Non-Contig Description
Mode Addr Mode Addr
8000 0000
8000 0001
8000 0002
8000 0003
8000 0004
8000 0005
8000 0006
8000 0007
8000 0008
8000 0009
8000 000A
8000 000B
8000 000C
8000 000D
8000 000E
8000 000F
8000 0020
8000 0021
8000 0040
8000 0041
8000 0042
8000 0043
8000 0060
8000 0061
8000 0062
8000 0064
8000 0066
8000 0070
8000 0071
8000 0074
8000 0075
8000 0077
8000 0078
8000 0079

8000 0000
8000 0001
8000 0002
8000 0003
8000 0004
8000 0005
8000 0006
8000 0007
8000 0008
8000 0009
8000 000A
8000 000B
8000 000C
8000 000D
8000 000E
8000 000F
8000 1000
8000 1001
8000 2000
8000 2001
8000 2002
8000 2003
8000 3000
8000 3001
8000 3002
8000 3004
8000 3006
8000 3010
8000 3011
8000 3014
8000 3015
8000 3017
8000 3018
8000 3019

DMA1 CH0 Base and Current Addr
DMA1 CH0 Base and Current Cnt
DMA1 CH1 Base and Current Addr
DMA1 CH0 Base and Current Cnt
DMA1 CH2 Base and Current Addr
DMA1 CH2 Base and Current Cnt
DMA1 CH3 Base and Current Addr
DMA1 CH3 Base and Current Cnt
DMA1 Status(R) Command(W)
DMA1 Soft Request
DMA1 Write Single Mask Bit
DMA1 Write Mode
DMA1 Clear Byte Pointer
DMA1 Master Clear
DMA1 Clear Mask
DMA1 R/W All Mask Register Bits
INT1 Control
INT1 Mask
Timer Counter 1 - Counter 0 Cnt
Timer Counter 1 - Counter 1 Cnt
Timer Counter 1 - Counter 2 Cnt
Timer Counter 1 Command Mode
Reset X-Bus (mse) IRQ12 and Kbd
NMI Status and Control
Reserved for Keyboard/Mouse
Keyboard/Mouse
Reserved for Keyboard/Mouse
TOD Addr and NMI Enable
TOD Read/Write
NV RAM Addr Strobe 0
NV RAM Addr Strobe 1
NV RAM Data Port
BIOS Timer
BIOS Timer

108

R/W

Set
To (6)

Loc
(4)

R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
W
W
W
W
W
W
R/W
R/W
R/W
R/W
R/W
R/W
W
R
R/W
R/W
R
R/W
W
R/W
W
W
R/W
R/W
R/W

—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—

SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
KBD
KBD
KBD
SIO
RTC
NVR
NVR
NVR
SIO
SIO

No
te

(3)
(3)
(3)
(3)
(3)
(3)
(3)

MPRH01TSU-02

Setup
Table 33. Combined Register Listing (Continued)
ISA
Port

Contiguous Non-Contig Description
Mode Addr Mode Addr

R/W

Set
To (6)

Loc
(4)

007A
007B
0080
0081
0082
0083
0084
0085
0086
0087
0088
0089
008A
008B
008C
008D
008E
008F
0090
0092
0094
0095
0096
0098
009C
009D
009E
009F
00A0
00A1
00C0
00C2
00C4
00C6
00C8
00CA
00CC
00CE
00D0
00D2
00D4
00D6

8000 007A
8000 007B
8000 0080
8000 0081
8000 0082
8000 0083
8000 0084
8000 0085
8000 0086
8000 0087
8000 0088
8000 0089
8000 008A
8000 008B
8000 008C
8000 008D
8000 008E
8000 008F
8000 0090
8000 0092
8000 0094
8000 0095
8000 0096
8000 0098
8000 009C
8000 009D
8000 009E
8000 009F
8000 00A0
8000 00A1
8000 00C0
8000 00C2
8000 00C4
8000 00C6
8000 00C8
8000 00CA
8000 00CC
8000 00CE
8000 000D
8000 00D2
8000 00D4
8000 00D6

R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
W
W
W

—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—

SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
660
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO

MPRH01TSU-02

8000 301A
8000 301B
8000 4000
8000 4001
8000 4002
8000 4003
8000 4004
8000 4005
8000 4006
8000 4007
8000 4008
8000 4009
8000 400A
8000 400B
8000 400C
8000 400D
8000 400E
8000 400F
8000 4010
8000 4012
8000 4014
8000 4015
8000 4016
8000 4018
8000 401C
8000 401D
8000 401E
8000 401F
8000 5000
8000 5001
8000 6000
8000 6002
8000 6004
8000 6006
8000 6008
8000 600A
8000 600C
8000 600E
8000 6010
8000 6012
8000 6014
8000 6016

BIOS Timer
BIOS Timer
DMA Page Register Reserved
DMA Channel 2 Page Register
DMA Channel 3 Page Register
DMA Channel 1 Page Register
DMA Page Register Reserved
DMA Page Register Reserved
DMA Page Register Reserved
DMA Channel 0 Page Register
DMA Page Register Reserved
DMA Channel 6 Page Register
DMA Channel 7 Page Register
DMA Channel 5 Page Register
DMA Page Register Reserved
DMA Page Register Reserved
DMA Page Register Reserved
DMA Low Page Register Refresh
DMA Page Register Reserved
Special Port 92 Register
DMA Page Register Reserved
DMA Page Register Reserved
DMA Page Register Reserved
DMA Page Register Reserved
DMA Page Register Reserved
DMA Page Register Reserved
DMA Page Register Reserved
DMA Low Page Register Refresh
INT2 Control Register
INT2 Mask Register
DMA2 CH0 Base and Current Addr
DMA2 CH0 Base and Current Cnt
DMA2 CH1 Base and Current Addr
DMA2 CH1 Base and Current Cnt
DMA2 CH2 Base and Current Addr
DMA2 CH2 Base and Current Cnt
DMA2 CH3 Base and Current Addr
DMA2 CH3 Base and Current Cnt
DMA2 Status(R) Command(W)
DMA2 Soft Request
DMA2 Write Single Mask Bit
DMA2 Write Mode

109

No
te

(2)

Setup
Table 33. Combined Register Listing (Continued)
ISA
Port

Contiguous Non-Contig Description
Mode Addr Mode Addr

R/W

Set
To (6)

Loc
(4)

00D8
00DA
00DC
00DE
00F0
040B
0410
0411
0412
0413
0415
0416
0417
0418
0419
041A
041B
041D
041E
041F
0420
0421
0422
0423
0424
0425
0426
0427
0428
0429
042A
042B
042C
042D
042E
042F
0434
0435
0436
0437
0438
0439

8000 00D8
8000 00DA
8000 00DC
8000 00DE
8000 00F0
8000 040B
8000 0410
8000 0411
8000 0412
8000 0413
8000 0415
8000 0416
8000 0417
8000 0418
8000 0419
8000 041A
8000 041B
8000 041D
8000 041E
8000 041F
8000 0420
8000 0421
8000 0422
8000 0423
8000 0424
8000 0425
8000 0426
8000 0427
8000 0428
8000 0429
8000 042A
8000 042B
8000 042C
8000 042D
8000 042E
8000 042F
8000 0434
8000 0435
8000 0436
8000 0437
8000 0438
8000 0439

W
W
W
R/W
R/W
W
W
W
W
W
W
W
W
R
R
R
R
R
R
R
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W

—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—

SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO

8000 6018
8000 601A
8000 601C
8000 601E
8000 7010
8002 0006
8002 0010
8002 0011
8002 0012
8002 0013
8002 0015
8002 0016
8002 0017
8002 0018
8002 0019
8002 001A
8002 001B
8002 001D
8002 001E
8002 001F
8002 1000
8002 1001
8002 1002
8002 1003
8002 1004
8002 1005
8002 1006
8002 1007
8002 1008
8002 1009
8002 100A
8002 100B
8002 100C
8002 100D
8002 100E
8002 100F
8002 1014
8002 1015
8002 1016
8002 1017
8002 1018
8002 1019

DMA2 Clear Byte Pointer
DMA2 Master Clear
DMA2 Clear Mask
DMA2 R/W All Mask Register Bits
Coprocessor Error Reg - Reserved
DMA1 Extended Mode
CH0 Scatter/Gather Command
CH1 Scatter/Gather Command
CH2 Scatter/Gather Command
CH3 Scatter/Gather Command
CH5 Scatter/Gather Command
CH6 Scatter/Gather Command
CH7 Scatter/Gather Command
CH0 Scatter/Gather Status
CH1 Scatter/Gather Status
CH2 Scatter/Gather Status
CH3 Scatter/Gather Status
CH5 Scatter/Gather Status
CH6 Scatter/Gather Status
CH7 Scatter/Gather Status
CH0 Scatter/Gather Pointer
CH0 Scatter/Gather Pointer
CH0 Scatter/Gather Pointer
CH0 Scatter/Gather Pointer
CH1 Scatter/Gather Pointer
CH1 Scatter/Gather Pointer
CH1 Scatter/Gather Pointer
CH1 Scatter/Gather Pointer
CH2 Scatter/Gather Pointer
CH2 Scatter/Gather Pointer
CH2 Scatter/Gather Pointer
CH2 Scatter/Gather Pointer
CH3 Scatter/Gather Pointer
CH3 Scatter/Gather Pointer
CH3 Scatter/Gather Pointer
CH3 Scatter/Gather Pointer
CH5 Scatter/Gather Pointer
CH5 Scatter/Gather Pointer
CH5 Scatter/Gather Pointer
CH5 Scatter/Gather Pointer
CH6 Scatter/Gather Pointer
CH6 Scatter/Gather Pointer

110

No
te

MPRH01TSU-02

Setup
Table 33. Combined Register Listing (Continued)
ISA
Port

Contiguous Non-Contig Description
Mode Addr Mode Addr

R/W

Set
To (6)

Loc
(4)

043A
043B
043C
043D
043E
043F
0481
0482
0483
0487
0489
048A
048B
04D0
04D1
04D6
0808
080C
080D
0814
081C
0821
082A
082B
0840
0842
0843
0844
0850
0852
0860
0862
0880
0881

8000 043A
8000 043B
8000 043C
8000 043D
8000 043E
8000 043F
8000 0481
8000 0482
8000 0483
8000 0487
8000 0489
8000 048A
8000 048B
8000 04D0
8000 04D1
8000 04D6
8000 0808
8000 080C
8000 080D
8000 0814
8000 081C
8000 0821
8000 082A
8000 082B
8000 0840
8000 0842
8000 0843
8000 0844
8000 0850
8000 0852
8000 0860
8000 0862
8000 0880
8000 0881
8000 0CF8
8000 0CFC
8080 08xx
8080 0800
8080 0801
8080 0802
8080 0803
8080 0804

R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
W
R/W
R
R
W
R/W
R/W
R/W
R/W
R
R
R
R
W
R
R/W
R/W
R
R
R/W
R/W
R/W
R
R
R
R
R/W

—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
(7)
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
0F

SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
EPLD
logic
logic
660
660
660
EPLD
EPLD
660
660
660
660
660
logic
EPLD
EPLD
logic
logic
660
660
660
SIO
SIO
SIO
SIO
SIO

MPRH01TSU-02

8002 101A
8002 101B
8002 101C
8002 101D
8002 101E
8002 101F
8002 4001
8002 4002
8002 4003
8002 4007
8002 4009
8002 400A
8002 400B
8002 6010
8002 6011
8002 6016
8004 0008
8004 000C
8004 000D
8004 0014
8004 001C
8004 1001
8004 100A
8004 100B
8004 2000
8004 2002
8004 2003
8004 2004
8004 2010
8004 2012
8004 3000
8004 3002
8004 4000
8004 4001

CH6 Scatter/Gather Pointer
CH6 Scatter/Gather Pointer
CH7 Scatter/Gather Pointer
CH7 Scatter/Gather Pointer
CH7 Scatter/Gather Pointer
CH7 Scatter/Gather Pointer
DMA CH2 High Page
DMA CH3 High Page
DMA CH1 High Page
DMA CH0 High Page
DMA CH6 High Page
DMA CH7 High Page
DMA CH5 High Page
Interrupt Control 1
Interrupt Control 2
DMA2 Extended Mode
HDD Light
Equipment Present
L2 Cache Status Reg
L2 Flush
System Control 81C
Memory Controller Misc
Power Mgmt Control Reg1
Power Mgmt Control Reg2
Memory Parity Error Status
L2 Error Status
L2 Parity Read & Clear
Unsupported Transfer Type Error
I/O Map Type
Board ID
Freeze Clock Reg Low
Freeze Clock Reg High
SIMM Presence Detect Slot 1/2
SIMM Presence Detect Slot 3/4
PCI/BCR Configuration Address
PCI/BCR Configuration Data
PCI Type 0 Configuration Addr
Vendor Identification
Vendor Identification
Device Identification
Device Identification
Command

111

No
te

(3)
(3)

(5)
(5)

(3)
(5)
(5)
(3)
(3)

Setup
Table 33. Combined Register Listing (Continued)
ISA
Port

Contiguous Non-Contig Description
Mode Addr Mode Addr

R/W

Set
To (6)

Loc
(4)

8080 0805
8080 0806
8080 0807
8080 0808
8080 0840
8080 0841
8080 0842
8080 0843
8080 0844
8080 0845
8080 0846
8080 0847
8080 0848
8080 0849
8080 084A
8080 084B
8080 084C
8080 084D
8080 084E
8080 084F
8080 0854
8080 0855
8080 0856
8080 0857
8080 0860
8080 0861
8080 0862
8080 0863
8080 0880
8080 0881
8080 10xx
8080 20xx
8080 40xx
8080 80xx
8081 00xx
8082 00xx
8084 00xx
8088 00xx
8090 00xx
80A0 00xx
80C0 00xx
BFFF EFF0

R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R

00
—
—
—
21
00
04
00
00
10
0F
00
F1
00
10
0F
56
10
07
FF
—
—
—
—
0F
0F
80
80
—
—
—
—
—
—
—
—
—
—
—
—
—
—

SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
SIO
660
660
660
660
660
660
660
660
660
660
660
660

BFFF EFF0

Command
Device Status
Device Status
Revision Identification
PCI Control
PCI Arbiter Control
PCI Arbiter Priority Control
PCI Arbiter Priority Control Extension
MEMCS# Control
MEMCS# Bottom of Hole
MEMCS# Top of Hole
MEMCS# Top of Memory
ISA Address Decoder Control
ISA Address Decoder ROM Block
ISA Address Bottom of Hole
ISA Address Top of Hole
ISA Controller Recovery Timer
ISA Clock Divisor
Utility Bus Chip Select A
Utility Bus Chip Select B
MEMCS# Attribute Register #1
MEMCS# Attribute Register #2
MEMCS# Attribute Register #3
Scatter/Gather Relocation Base
PIRQ Route Control 0
PIRQ Route Control 1
PIRQ Route Control 2
PIRQ Route Control 3 (unused)
BIOS Timer Base Address
BIOS Timer Base Address
PCI Type 0 Configuration Addr
PCI Type 0 Configuration Addr
PCI Type 0 Configuration Addr
PCI Type 0 Configuration Addr
PCI Type 0 Configuration Addr
PCI Type 0 Configuration Addr
PCI Type 0 Configuration Addr
PCI Type 0 Configuration Addr
PCI Type 0 Configuration Addr
PCI Type 0 Configuration Addr
PCI Type 0 Configuration Addr
System Error Addr

112

No
te

(5)
(5)
(5)
(5)
(5)
(5)
(5)
(5)

MPRH01TSU-02

Setup
Table 33. Combined Register Listing (Continued)
ISA
Port

Contiguous Non-Contig Description
Mode Addr Mode Addr
BFFF FFF0
FFFF FFF0
FFFF FFF1

BFFF FFF0
FFFF FFF0
FFFF FFF1

Interrupt Vector
Flash Write Addr/Data
Flash Lock Out

R/W

Set
To (6)

Loc
(4)

R
W
W

—
—
—

660
660
660

No
te

Notes:
1) The first 5 hex digits in the contiguous and non–contiguous mode columns represent the
memory page number for which the protection attributes may be set in contiguous I/O
mode. That is, devices having the same first five digits in this column will have the same
attributes in the memory page table.
2) Port 94 may be used by certain video controllers (e.g. Weitekt 9100). The SIO chip positively decodes this port. Therefore bus contention may arise when both devices claim
the PCI cycle to this port address. Bus contention results in invalid data and possibly
harm to the hardware.
3) The control signals for these ports are partially decoded by the SIO. The System I/O
EPLD completes the decodes, and issues control signals to the registers, which are usually X-bus buffers.
4) KBD = Keyboard / Mouse Controller
RTC = Real Time Clock, also known as the TOD (Time Of Day clock)
NVR = Non-Volatile RAM, in the same package as the RTC
660 = The 660 Bridge.
5) Not used.
6) In the Set To column, a long dash — means that the initialization firmware does not write
to this register. The register is either not used, not written to, or the value of it depends
on changing circumstances.
If the word Memory appears, please refer to the System Memory section of the 660
User’s Manual.
7) Set register 81C to C0h if an L2 is installed, else leave at reset value.

MPRH01TSU-02

113

Setup
9.5.2

Indexed BCR Summary

Table 34 contains a summary listing of the indexed BCRs. Access to these registers is described in the 660 Bridge User’s Manual.
Table 34. 660 Bridge Indexed BCR Listing
Index

R/W

Bytes

Set To (1)

PCI Vendor ID

Index 00 – 01

R

2

—

PCI Device ID

Index 02 – 03

R

2

—

PCI Command

Index 04 – 05

R/W

2

—

PCI Device Status

Index 06 – 07

R/W

2

—

Revision ID

Index 08

R

1

—

PCI Standard Programming Interface

Index 09

R

1

—

PCI Subclass Code

Index 0A

R

1

—

PCI Class Code

Index 0B

R

1

—

PCI Cache Line Size

Index 0C

R

1

—

PCI Latency Timer

Index 0D

R

1

—

PCI Header Type

Index 0E

R

1

—

PCI Built-in Self-Test (BIST) Control

Index 0F

R

1

—

PCI Interrupt Line

Index 3C

R

1

—

PCI Interrupt Pin

Index 3D

R

1

—

PCI MIN_GNT

Index 3E

R

1

—

PCI MAX_LAT

Index 3F

R

1

—

PCI Bus Number

Index 40

R

1

—

PCI Subordinate Bus Number

Index 41

R

1

—

PCI Disconnect Counter

Index 42

R/W

1

—

PCI Special Cycle Address BCR

Index 44 –45

R

2

—

Memory Bank 0 Starting Address

Index 80

R/W

1

Memory

(2)

Memory Bank 1 Starting Address

Index 81

R/W

1

Memory

(2)

Memory Bank 2 Starting Address

Index 82

R/W

1

Memory

(2)

Memory Bank 3 Starting Address

Index 83

R/W

1

Memory

(2)

Memory Bank 4 Starting Address

Index 84

R/W

1

Memory

(2)

Memory Bank 5 Starting Address

Index 85

R/W

1

Memory

(2)

Memory Bank 6 Starting Address

Index 86

R/W

1

Memory

(2)

Memory Bank 7 Starting Address

Index 87

R/W

1

Memory

(2)

Memory Bank 0 Ext Starting Address

Index 88

R/W

1

Memory

(2)

Memory Bank 1 Ext Starting Address

Index 89

R/W

1

Memory

(2)

Memory Bank 2 Ext Starting Address

Index 8A

R/W

1

Memory

(2)

Memory Bank 3 Ext Starting Address

Index 8B

R/W

1

Memory

(2)

Bridge Control Register

114

Note

MPRH01TSU-02

Setup
Table 34. 660 Bridge Indexed BCR Listing (Continued)
Index

R/W

Bytes

Set To (1)

Note

Memory Bank 4 Ext Starting Address

Index 8C

R/W

1

Memory

(2)

Memory Bank 5 Ext Starting Address

Index 8D

R/W

1

Memory

(2)

Memory Bank 6 Ext Starting Address

Index 8E

R/W

1

Memory

(2)

Memory Bank 7 Ext Starting Address

Index 8F

R/W

1

Memory

(2)

Memory Bank 0 Ending Address

Index 90

R/W

1

Memory

(2)

Memory Bank 1 Ending Address

Index 91

R/W

1

Memory

(2)

Memory Bank 2 Ending Address

Index 92

R/W

1

Memory

(2)

Memory Bank 3 Ending Address

Index 93

R/W

1

Memory

(2)

Memory Bank 4 Ending Address

Index 94

R/W

1

Memory

(2)

Memory Bank 5 Ending Address

Index 95

R/W

1

Memory

(2)

Memory Bank 6 Ending Address

Index 96

R/W

1

Memory

(2)

Memory Bank 7 Ending Address

Index 97

R/W

1

Memory

(2)

Memory Bank 0 Ext Ending Address

Index 98

R/W

1

Memory

(2)

Memory Bank 1 Ext Ending Address

Index 99

R/W

1

Memory

(2)

Memory Bank 2 Ext Ending Address

Index 9A

R/W

1

Memory

(2)

Memory Bank 3 Ext Ending Address

Index 9B

R/W

1

Memory

(2)

Memory Bank 4 Ext Ending Address

Index 9C

R/W

1

Memory

(2)

Memory Bank 5 Ext Ending Address

Index 9D

R/W

1

Memory

(2)

Memory Bank 6 Ext Ending Address

Index 9E

R/W

1

Memory

(2)

Memory Bank 7 Ext Ending Address

Index 9F

R/W

1

Memory

(2)

Memory Bank Enable

Index A0

R/W

1

Memory

(2)

Memory Timing 1

Index A1

R/W

1

0001 0010

(2)

Memory Timing 2

Index A2

R/W

1

1000 1010

(2)

Memory Bank 0 & 1 Addressing Mode

Index A4

R/W

1

Mode 2

(2)

Memory Bank 2 & 3 Addressing Mode

Index A5

R/W

1

—

Memory Bank 4 & 5 Addressing Mode

Index A6

R/W

1

—

Memory Bank 6 & 7 Addressing Mode

Index A7

R/W

1

—

Cache Status

Index B1

R/W

1

—

Refresh Cycle Definition

Index B4

R

1

—

Refresh Timer B5 (Not used – see Indexed BCR D0)

Index B5

R

1

—

RAS Watchdog Timer

Index B6

R/W

1

—

PCI Bus Timer (Not used)

Index B7

R

1

—

Single-Bit Error Counter

Index B8

R/W

1

—

Single-Bit Error Trigger Level

Index B9

R/W

1

—

Bridge Options 1

Index BA

R/W

1

—

Bridge Control Register

MPRH01TSU-02

115

Setup
Table 34. 660 Bridge Indexed BCR Listing (Continued)
Index

R/W

Bytes

Set To (1)

Bridge Options 2

Index BB

R/W

1

—

Error Enable 1

Index C0

R/W

1

—

Error Status 1

Index C1

R/W

1

—

Error Simulation 1

Index C2

R/W

1

—

CPU Bus Error Status

Index C3

R

1

—

Error Enable 2

Index C4

R/W

1

—

Error Status 2

Index C5

R/W

1

—

Error Simulation 2

Index C6

R/W

1

—

PCI Bus Error Status

Index C7

R/W

1

—

CPU/PCI Error Address

Index C8–CB

R/W

4

—

Single-Bit ECC Error Address

Indx CC – CF

R/W

4

—

Refresh Timer Divisor

Index D0 – D1 R/W

2

01F8h

Suspend Refresh Timer

Index D2 – D3 R/W

2

—

1

0000 1000

Bridge Control Register

Bridge Chip Set Options 3

Index D4

R/W

Note

1) In this column, a long dash — means that the initialization firmware does not write to this
register. The register is either not used, not written to, or the value of it depends on
changing circumstances.
If the word Memory appears, please refer to the System Memory section of the 660
User’s Manual.
2) The initialization firmware sets these registers depending on the information reported
by the DRAM presence detect registers.

9.6

ISA Bus Register Suggestions

The following port assignments are designed to be compatible with the reference design
firmware and Super I/O type chips. These registers and functions are not implemented on
the reference design motherboard.
Table 35. Compatible ISA Ports (Not on Reference Board)
ISA
Port
01F0
01F1
01F2
01F3
01F4
01F5
01F6
01F7
0278

Contiguous Non-Contig Description
Mode Addr Mode Addr
8000 01F0
8000 01F1
8000 01F2
8000 01F3
8000 01F4
8000 01F5
8000 01F6
8000 01F7
8000 0278

8000 F010
8000 F011
8000 F012
8000 F013
8000 F014
8000 F015
8000 F016
8000 F017
8001 3018

IDE Data
IDE Error/Features
IDE Sector Count
IDE Sector Number
IDE Cylinder Low
IDE Cylinder High
IDE Drive Head
IDE Status/Command
Parallel Port 2

116

Note

R/W

(1)

R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W

MPRH01TSU-02

Setup
Table 35. Compatible ISA Ports (Not on Reference Board) (Continued)
ISA
Port

Contiguous Non-Contig Description
Mode Addr Mode Addr

0279
027A
027B
027C
027D
02F8
02F9
02FA
02FB
02FC
02FD
02FE
02FF
0370
0371
0372
0373
0374
0375
0376
0377
0376
0377
0378
0379
037A
037B
037C
037D
0398
0399
03BC
03BD
03BE
03F0
03F1
03F2
03F3
03F4
03F5
03F6
03F7

8000 0279
8000 027A
8000 027B
8000 027C
8000 027D
8000 02F8
8000 02F9
8000 02FA
8000 02FB
8000 02FC
8000 02FD
8000 02FE
8000 02FF
8000 0370
8000 0371
8000 0372
8000 0373
8000 0374
8000 0375
8000 0376
8000 0377
8000 0376
8000 0377
8000 0378
8000 0379
8000 037A
8000 037B
8000 037C
8000 037D
8000 0398
8000 0399
8000 03BC
8000 03BD
8000 03BE
8000 03F0
8000 03F1
8000 03F2
8000 03F3
8000 03F4
8000 03F5
8000 03F6
8000 03F7

MPRH01TSU-02

8001 3019
8001 301A
8001 301B
8001 301C
8001 301D
8001 7018
8001 7019
8001 701A
8001 701B
8001 701C
8001 701D
8001 701E
8001 701F
8001 B010
8001 B011
8001 B012
8001 B013
8001 B014
8001 B015
8001 B016
8001 B017
8001 B016
8001 B017
8001 B018
8001 B019
8001 B01A
8001 B01B
8001 B01C
8001 B01D
8001 C018
8001 C019
8001 D01C
8001 D01D
8001 D01E
8001 F010
8001 F011
8001 F012
8001 F013
8001 F014
8001 F015
8001 F016
8001 F017

Parallel Port 2
Parallel Port 2
Parallel Port 2
Parallel Port 2
Parallel Port 2
Serial Port 2
Serial Port 2
Serial Port 2
Serial Port 2
Serial Port 2
Serial Port 2
Serial Port 2
Serial Port 2
Secondary Floppy Digital Output
Secondary Floppy Digital Output
Secondary Floppy Digital Output
Secondary Floppy Digital Output
Secondary Floppy Digital Output
Secondary Floppy Digital Output
Secondary Floppy Digital Output
Secondary Floppy Digital Output
Secondary IDE Alt Status/Device Ctl
Secondary IDE drive address
Parallel Port 1
Parallel Port 1
Parallel Port 1
Parallel Port 1
Parallel Port 1
Parallel Port 1
Super I/O Index Address
Super I/O Data Address
Parallel Port 3
Parallel Port 3
Parallel Port 3
Primary Floppy Digital Output (Media Sense)
Primary Floppy Digital Output
Primary Floppy Digital Output
Primary Floppy Digital Output (Also Media Sense)
Primary Floppy Digital Output
Primary Floppy Digital Output
Primary Floppy Digital Output
Primary Floppy Digital Output

117

Note

(1)

R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
R/W
W
W
W
W
W
W
W
W
R/W
W
W
W
W
W
W
W
R/W
R/W
R/W
R/W
R/W
W/O
W/O
W
W
W
W
W
W

Setup
Table 35. Compatible ISA Ports (Not on Reference Board) (Continued)
ISA
Port

Contiguous Non-Contig Description
Mode Addr Mode Addr

03F6
03F7
03F8
03FA
03FB
03FC
03FD
03FE
03FF

8000 03F6
8000 03F7
8000 03F8
8000 03FA
8000 03FB
8000 03FC
8000 03FD
8000 03FE
8000 03FF

8001 F016
8001 F017
8001 F018
8001 F01A
8001 F01B
8001 F01C
8001 F01D
8001 F01E
8001 F01F

Primary IDE Alt Status/Device Ctl
Primary IDE Drive Address
Serial Port 1
Serial Port 1
Serial Port 1
Serial Port 1
Serial Port 1
Serial Port 1
Serial Port 1

Note

R/W
R/W
R
R/W
R/W
R/W
R/W
R/W
R/W
R/W

1) This is a preferred location for this function.

118

MPRH01TSU-02

Firmware

Section 10
System Firmware
10.1 Introduction
The firmware on the PowerPC 603/604 reference board handles three major functions:
S Test the system in preparation for execution,
S Load and execute an executable image from a bootable device, and
S Allow user configuration of the system.
Section 10.2 briefly discusses the power on system test function.
Section 10.3 details a structure for boot records which can be loaded by the system firmware.
Section 10.4 describes the system configuration utility.
To obtain a copy of the commented source code of the firmware on diskette, contact
your IBM representative. This material is available free of charge with a signed license agreement.

10.2 Power On System Test
The Power On System Test (POST) code tests those subsystems of the reference board
which are required for configuration and boot to ensure minimum operability. Tests also assure validity of the firmware image and of the stored system configuration.
10.2.1 Hardware Requirements
In addition to the reference board, the firmware requires the following peripherals to be
installed as adapter cards:
S Serial Port 1
S Serial Port 2
S Floppy Controller

Address: 0x3F8 (COM1:)
Address: 0x2F8 (COM2:)
Address: 0x3F0 (Primary Floppy)

S IDE Controller

Address: 0x1F0 (Primary IDE)

Interrupt: IRQ 4
Interrupt: IRQ 3
Mode: PC/AT or
PS/2

10.3 Boot Record Format
The firmware will attempt to boot an executable image from devices specified by the user.
See section 10.4 for details on specifying boot devices and order.
MPRH01TSU-02

119

Firmware
The PowerPC Reference Platform Specification details a structure for boot records which
can be loaded by the system firmware. This specification is described in the following sections.
10.3.1 Boot Record
The format of the boot record is an extension of the PC environment. The boot record is
composed of a PC compatibility block and a partition table. To support media interchange,
the PC compatibility block may contain an x86-type program. The entries in the partition
table identify the PowerPC Reference Platform boot partition and its location in the media.
The layout of the boot record must be designed as shown in Figure 5. The first 446 bytes
of the boot record contain a PC compatibility block, the next four 16-byte entries make up
a partition table totalling 64 bytes, and the last 2 bytes contain a signature.
0

0
PC Compatibility
Block

0x1BE

446
Partition Entry 1

0x1CE

462
Partition Entry 2

0x1DE

478
Partition Entry 3

0x1EE

494
Partition Entry 4

0x1FE

510
0x55

0xAA

Figure 25. Boot Record
10.3.1.1
PC Partition Table Entry
To support media interchange with the PC, the PowerPC Reference Platform defines the
format of the partition table entry based on that for the PC. This section describes the format
of the PC partition table entry, which is shown in Figure 26.
0
Partition Begin

4

Boot Ind

Head

Sector

Cyl

Partition End

8

Sys Ind

Head

Sector

Cyl

Beginning Sector
Number of Sectors

12

Low Word (LE)

High Word (LE)

Low Word (LE)

High Word (LE)

LE = Little-Endian

Figure 26. Partition Table Entry
120

MPRH01TSU-02

Firmware

S Partition Begin
S Partition End
S Beginning Sector

S Number of Sectors

The beginning address of the partition in head, sector, cylinder notation.
The end address of the partition in cylinder, head, sector
notation.
The number of sectors preceding the partition on the disk.
That is, the zero-based relative block address of the first
sector of the partition.
The number of sectors allocated to the partition.

The subfields of a partition table entry are defined as follows:
S Boot Ind

S Head
S Sector

S Cyl

S Sys Ind

Boot Indicator. This byte indicates if the partition is active. If
the byte contains 0x00, then the partition is not active and
will not be considered as bootable. If the byte contains 0x80,
then the partition is considered active.
An eight-bit value, zero-based.
A six-bit value, one-based. The low-order six bits are the
sector value. The high-order two bits are the high-order bits
of the 10-bit cylinder value.
Cylinder. The low-order eight-bit component of the 10-bit cylinder value (zero-based). The high-order two bits of the cylinder value are found in the sector field.
System Indicator. This byte defines the type of the partition.
There are numerous partition types defined. For example,
the following list shows several:
0x00
0x01
0x04
0x05
0x41

Available partition
DOS, 12-bit FAT
DOS, 16-bit FAT
DOS extended partition
PowerPC Reference Platform partition

10.3.1.2
Extended DOS Partition
The extended DOS partition is used to allow more than four partitions in a device. The boot
record in the extended DOS partition has a partition table with two entries, but does not contain the code section. The first entry describes the location, size and type of the partition.
The second entry points to the next partition in the chained list of partitions. The last partition in the list is indicated with a system indicator value of zero in the second entry of its
partition table.

MPRH01TSU-02

121

Firmware
Because of the DOS format limitations for a device partition, a partition which starts at a
location beyond the first 1 gigabyte is located by using an enhanced format shown in
Figure 27.

0
Partition Begin

Boot Ind

–1

–1

–1

Sys Ind

–1

–1

–1

4
Partition End
8
32–bit start RBA (zero–based) (LE)

Beginning Sector
12

32–bit RBA count (one–based) (LE)

Number of Sectors

–1 = All ones in the field.
RBA = Relative Block Address in units of 512 bytes.

LE = Little-Endian

Figure 27. Partition Table Entry Format for an Extended Partition
10.3.1.3
PowerPC Reference Platform Partition Table Entry
The Power PC Reference Platform partition table entry (see Figure 28) is identified by the
0x41 value in the system indicator field. All other fields are ignored by the firmware except
for the Beginning Sector and Number of Sectors fields. The CV (Compatible Value – not
shown) fields must contain PC-compatible values (i.e. acceptable to DOS) to avoid confusing PC software. The CV fields, however, are ignored by the firmware.

0
Partition Begin

Boot Ind

Head

Sector

Cyl

Sys Ind

Head

Sector

Cyl

4
Partition End
8
32–bit start RBA (zero–based) (LE)

Beginning Sector
12
Number of Sectors

32–bit RBA count (one–based) (LE)

RBA = Relative Block Address in units of 512 bytes.

LE = Little-Endian

Figure 28. Partition Table Entry for PowerPC Reference Platform
The 32-bit start RBA is zero-based. The 32-bit count RBA is one-based and indicates the
number of 512-byte blocks. The count is always specified in 512-byte blocks even if the
physical sectoring of the target devices is not in 512-byte sectors.

122

MPRH01TSU-02

Firmware
10.3.2 Loading the Load Image
This section describes the layout of the PowerPC 0x41 type partition and the process of
loading the load image.
0
PC Compatibility
Block
512
Entry Point Offset (LE)
516
Load Image Length (LE)
Load Image
Flag Field
OS_ID

520
521
522

Partition Name
554
Reserved1
1024
OS–Specific Field
(Optional)

Code Section of the
Load Image

Reserved2
RBA Count * 512

Figure 29. PowerPC Reference Platform Partition
The layout for the 0x41 type partition is shown in Figure 29. The PC compatibility block in
the boot partition may contain an x86-type program. When executed on an x86 machine,
this program displays a message indicating that this partition is not applicable to the current
system environment.
The second relative block in the boot partition contains the entry point offset, load image
length, flag field, operating system ID field, ASCII partition name field, and the reserved1
area. The 32-bit entry point offset (little-endian) is the offset (into the image) of the entry
point of the PowerPC Reference Platform boot program. The entry point offset is used to
MPRH01TSU-02

123

Firmware
allocate the Reserved1 space. The reserved1 area from offset 554 to Entry Point – 1 is reserved for implementation specific data and future expansion.
The 32-bit load image length (little-endian) is the length in bytes of the load image. The load
image length specifies the size of the data physically copied into the system RAM by the
firmware.
The flag field is 8 bits wide. The MSb in the field is allocated for the Open Firmware flag.
If this bit is set to 1, the loader requires Open Firmware services to continue loading the
operating system.
The second MSb is the endian mode bit. If the mode bit is 0, the code in the section is in
big-endian mode. Otherwise, the codes is in little-endian mode. The implication of the endian mode bit is different depending on the Open Firmware flag. If the Open Firmware flag
is set to 1, the mode bit indicates the endian mode of the code section pointed to by the
load image offset, and the firmware has to establish the hardware endian mode according
to this bit. Otherwise, this bit is just an informative field for firmware.
The OS_ID field and partition name field are used to identify the operating system located
in the partition. The OS_ID field has the numeric identification value of the operating system
located in the partition. The 32 byes of partition name field must have the ASCII notation
of the partition name. The name and OS_ID can be used to provide to a user the identification of the boot partition during the manual boot process.
Once the boot partition is identified by the PowerPC Reference Platform boot partition table
entry, the firmware:
S Reads into memory the second 512-byte block of the boot partition
S Determines the load image length for reading in the boot image up to but not including the reserved2 space
S Allocates a buffer in system RAM for the load image transfer (no fixed location)
S Transfers the load image into system RAM from the boot device (the reserved2
space is not loaded).
The load image must be fully relocatable, as it may be placed anywhere in memory by the
system firmware. Once loaded, the load image may relocate itself anywhere within system
RAM.

124

MPRH01TSU-02

Firmware

10.4 System Configuration
This section describes the utilities in the system firmware which allow the system to be customized. These utilities allow viewing of the system configuration, as well as the ability to
change I/O device configurations, console selection, boot devices, and the date and time.
These functions are described in the following sections.
10.4.1 System Console
The system console can be either a screen-oriented video display or a line-oriented serial
terminal. The example screens shown in this section show the S3 video/keyboard interface. When using a serial terminal, the configuration utilities will prompt for numeric input
for each prompt instead of using the arrow keys. All choices and options are the same as
for the screen-oriented menus.
The configuration of the reference board as shipped is set for S3 video / Keyboard console.
In the case that either the video adapter or the keyboard fails the power-on test, the system
console will default to serial port 1. The baud rate for the serial console is specified in the
configuration menus. The value as shipped is 9600 baud.
10.4.2 System Initialization
The logo screen, shown in Figure 30, is displayed at power-on. The logo screen is active
while the system initializes and tests memory and performs a scan of the SCSI bus to determine what SCSI devices are installed.

PowerPC 603/604 Reference Board System Firmware
(C) Copyright 1994 IBM Corp. All Rights Reserved.

######
######
#####
#
#
####
#
# ###### ##### #
# #
#
#
# #
# #
# #
#
# #
# #
######
#
# #
# #####
#
# ###### #
#
#
# # ## # #
##### #
#
#
#
# ## ## #
#
# #
#
#
#
####
#
# ###### #
# #
#####
PowerPC 603/604 Reference Board
Press C during memory test for configuration utilities
Testing 8192K of memory:

8192 KB OK

Figure 30. System Initialization Screen

MPRH01TSU-02

125

Firmware
While the logo screen is displayed, pressing the ’C’ key on the console will enter the system
configuration utility. The configuration menu will also be entered if there is no bootable device present, or if the configuration stored in the system non-volatile RAM is not initialized
or is corrupt.
10.4.3 Main Menu
Figure 31 shows the main menu for the system configuration utility. Selections on the menu
are highlighted by using the up and down arrow keys on the keyboard, and are chosen with
the Enter key. Each choice is detailed in the following sections.
PowerPC 603/604 Reference Board System Firmware
(C) Copyright 1994 IBM Corp. All Rights Reserved.

Main Menu

System Configuration Menu
Run a Program
Reprogram Flash Memory
Save and exit
Exit without saving

Press ↑ ↓ to select item
Press Enter to perform action

Figure 31. Configuration Utility Main Menu

126

MPRH01TSU-02

Firmware
10.4.3.1
System Configuration Menu
Figure 32 shows the System Configuration menu, which has choices to display and change
the default state of the reference board on boot. Each menu item is discussed in the following sections.
PowerPC 604 Reference Board System Firmware
(C) Copyright 1994 IBM Corp. All Rights Reserved.

System Configuration Menu

System Information
Configure I/O Devices
View SCSI Devices
Set Boot Devices
Set Date and Time
Previous Menu

Press ↑ ↓ to select item
Press Enter to perform action

Figure 32. System Configuration Menu

MPRH01TSU-02

127

Firmware
System Information
The system configuration option shows the hardware configuration of the system at powerup—including processor, installed options, and firmware revision level. A sample screen
is shown in Figure 33.
PowerPC 603/604 Reference Board System Firmware
(C) Copyright 1994 IBM Corp. All Rights Reserved.

System Configuration

System Processor
PowerPC 604
Installed Memory
8 MB
Second-Level Cache
Not Installed
Upgrade Processor
Not Installed
Boot Firmware Revision 1.0
Go to Previous Menu

Press ↑ ↓ to select item
Press Enter to perform action

Figure 33. System Information Screen

128

MPRH01TSU-02

Firmware
Configure I/O Devices
The configure I/O devices option allows the customization of system I/O ports and the system console. The menu is shown in Figure 34. Options are highlighted by using the up and
down arrow keys on the keyboard and are changed with the left and right arrow keys. Options on the menu are discussed below.

PowerPC 603/604 Reference Board System Firmware
(C) Copyright 1994 IBM Corp. All Rights Reserved.

Device Configuration

Select Console Device
Set Serial Port 1 Speed
Set Serial Port 2 Speed

[S3 Video / Keyboard]
[ 9600 Baud]
[ 9600 Baud]

Go to Previous Menu

Press ↑ ↓ to select item
Press ← → to change item

Figure 34. Device Configuration Screen
Any changes made in I/O device configuration are saved when the Save and Exit option
on the main menu is selected. Exiting the system configuration utility in any other manner
will cause device configuration changes to be lost.
Select Console Device
The console selection box allows the selection of an option for the system console
S Serial Port 1 or 2

Console input and output will be transmitted and received
through a serial port on an adapter card. Console input and
output will be transmitted and received at the baud rate selected with Serial Port Speed.

S S3 Video/Keyboard

Console output will be displayed on a video monitor connected to an S3 PCI video adapter; console input will be received from a keyboard connected to the keyboard connector on the reference board

MPRH01TSU-02

129

Firmware
Set Serial Port 1 or 2 Speed
The serial port speed selection box sets the speed of each serial port. Baud rates for the
two serial ports are independent. If a serial port is used as the system console, set this value
to match the baud rate of the terminal.
View SCSI Devices
The SCSI devices screen shows the devices found on the SCSI bus during power-on initialization. The string shown is the SCSI device’s response to the SCSI inquiry command. According to the SCSI specification, this data comprises the manufacturer’s ID, device model
number, and device revision level. A sample screen is shown in Figure 35.
PowerPC 603/604 Reference Board System Firmware
(C) Copyright 1994 IBM Corp. All Rights Reserved.

SCSI Devices

SCSI
SCSI
SCSI
SCSI
SCSI
SCSI
SCSI

Device
Device
Device
Device
Device
Device
Device

0
1
2
3
4
5
6

None
None
None
None
None
None
IBM MXT-540SL H

Previous Menu

Press ↑ ↓ to select item
Press Enter to perform action

Figure 35. SCSI Devices Screen

130

MPRH01TSU-02

Firmware
Set Boot Devices
The boot devices menu allows the user to select which devices are queried for boot images
and in what order they are selected for boot. Allowable selections are one of the two floppy
disk drives, any of six SCSI drive ID numbers, either of two IDE disk drives, or no device
selected. The default configuration is shown in Figure 36. In this configuration, the system
will attempt to find a boot image on the first floppy disk drive. If this fails, the system will
attempt to boot from the SCSI device programmed to SCSI ID 6. If this fails, the system will
attempt to boot from IDE drive zero (master).
PowerPC 603/604 Reference Board System Firmware
(C) Copyright 1994 IBM Corp. All Rights Reserved.

Boot Device Selection

Set
Set
Set
Set

Boot
Boot
Boot
Boot

Device
Device
Device
Device

1
2
3
4

[Floppy 1
]
[SCSI ID 6 ]
[IDE Drive 0 ]
[None
]

Go to Previous Menu

Press ↑ ↓ to select item
Press ← → to change item

Figure 36. Boot Devices Screen
If the system fails to find a valid boot image (as discussed in section 10.3) on any of the
selected boot devices, or if no boot device is selected, the user will be prompted to enter
the configuration menu to select a valid boot device.
Any changes made in boot device selection is saved when the Save and Exit option on the
main menu is selected. Exiting the system configuration utility in any other manner will
cause boot device changes to be lost.

MPRH01TSU-02

131

Firmware
Set Date and Time
The set date and time screen allows the date and time stored in the battery-backed real
time clock to be updated. The screen is shown in Figure 37. To change the time, the left
and right arrow keys are used to select the digit to modify, and the digit is then typed over
with the number keys. The date or time will be updated when Enter or either the up or down
arrow is pressed. Changing the date or time is immediate, and is not affected by either the
Save and Exit or Exit Without Saving options on the main menu.
PowerPC 603/604 Reference Board System Firmware
(C) Copyright 1994 IBM Corp. All Rights Reserved.

Set Date and Time

Set Date
Set Time

[03/01/94]
[11:30:00]

Go to Previous Menu

Press ↑ ↓ to select item
Enter data at cursor

Figure 37. Set Date and Time Screen

132

MPRH01TSU-02

Firmware
10.4.3.2
Run a Program
The Run a Program option on the main menu loads and executes a program from a FAT
(DOS) disk or from a CD-ROM in ISO-9660 format. The program is loaded at location
0x00400000 (4 MB) and control is passed with a branch to the first address.
All boot devices specified in the Boot Devices Menu will be searched in order for FAT and
CD-ROM file systems, and the first matching file on a boot device will be loaded.
The Run a Program screen is shown in Figure 38. To run a program, enter the file name
in the Specify Program Filename field and select the Run the Program option.
PowerPC 603/604 Reference Board System Firmware
(C) Copyright 1994 IBM Corp. All Rights Reserved.

Run a Program

Specify Program Filename [
Run the Program

]

Go to Previous Menu

Press ↑ ↓ to select item
Enter data at cursor

Figure 38. Run a Program Screen

MPRH01TSU-02

133

Firmware
10.4.3.3
Reprogram Flash Memory
The PowerPC 603/604 reference board stores its system firmware in a reprogrammable
flash memory on the system board. The reprogram flash memory option on the main menu
allows the reprogramming of the flash device with a DOS-formatted diskette. This allows
future revisions of the system firmware to be provided on diskette without the need for removal of the device from the board.
If done improperly, reprogramming the flash memory can cause the system to become unusable until external means are available to reprogram the device. Use this option with
care.
All boot devices specified in the Boot Devices Menu will be searched in order for FAT and
CD-ROM file systems, and the first matching file on a boot device will be loaded.
The Reprogram the Flash Memory screen is shown in Figure 39. To reprogram the flash,
enter the file name in the Specify Image Filename field and select the Reprogram the
Memory option.
PowerPC 603/604 Reference Board System Firmware
(C) Copyright 1994 IBM Corp. All Rights Reserved.

Reprogram Flash Memory

Image Filename
[
Reprogram the Memory

]

Go to Previous Menu

Press ↑ ↓ to select item
Enter data at cursor

Figure 39. Reprogram the Flash Memory Screen

134

MPRH01TSU-02

Firmware
10.4.3.4
Exit Options
The two exit options at the bottom of the main menu leave the system configuration utility.
The two options are:
S Save and Exit
S Exit without Saving

Saves any changes made in the Configure I/O Devices and
Set Boot Devices screens, and restarts the system.
Proceeds with the boot process as if the configuration utility
had not been entered. Any changes made in Configure I/O
Devices or Set Boot Devices are lost.

10.4.4 Default Configuration Values
When the PowerPC 603/604 reference board is shipped from the factory, it has the following default configuration:
S3 Video / Keyboard
9600 Baud
9600 Baud
Device 1 - Floppy 1
Device 2 - SCSI ID 6
Device 3 - IDE Drive 0
These default values also take effect whenever the system configuration in system nonvolatile RAM becomes corrupted.
S
S
S
S

Console Device
Serial Port 1
Serial Port 2
Boot Devices

MPRH01TSU-02

135

Firmware

136

MPRH01TSU-02

Electromechanical

Section 11
Electromechanical
11.1 Electrical
11.1.1 Power Requirements
This section sets out the power supply requirements for the motherboard. They are achievable with low-cost PC power supplies (see Table 36 for specifications and Table 37 for
approximate power consumption).
Table 36. Power Supply Specification
Tolerance
Maximum Ripple P–P

Output
+5V

+5% –4%

50 mV

+12V

+5% –5%

120 mV

–12V

+10% –9%

120 mV

–5V* (1)

+10% –10%

120 mV

3.3V (2)

+5% –4%

50 mV

Element

Table 37. Approximate Power Consumption
+5V
+12V
–12V
(Amp)
(Amp)
(Amp)

–5V
(Amp) (1)

3.3V
(Amp) (2)

0.02

0

0

Base motherboard with 1 8M SIMM
running typical code

3.0

0.1

8M DRAM standby/refresh (each)

0.18

0

0

0

0

32M DRAM standby/refresh (each)

0.16

0

0

0

0

256K SRAM L2 Cache

1.6

0

0

0

0

Each PCI Slot – allocation (3)

3.0 max

0.3 max

0.06 max

0

3.3 max

Each ISA Slot – allocation (3)(4)

4.5 max

1.5 max

0.3 max

0.2 max

0

Main Power Connector Capacity (at
20.0
5.0
5.0
0.5
0
5.0 amp./pin––20 amp. DC return)
Notes for Table 36 and Table 37:
1. The –5v is not used on the planer, but it is routed from the power supply connector to the ISA slots.
2. The 3.3v to the PCI bus slots comes from the power supply connector. The 3.3v for the CPU and
the 660 bridge is a separate supply, generated from the +5v supply by a linear regulator.
3. These power requirements are allocated by the system designer. The currents specified per slot
in Table 37 are also the maximum currents which may be consumed by the bus. Because of this,
the total currents for all the cards used for either the PCI slot or the ISA slot must not exceed the
amounts listed in Table 37 for those slots.
4. ISA slot information is taken from IEEE P996.

MPRH01TSU-02

137

Electromechanical
Other requirements are:
1.
2.

3.

4.
5.

Overshoot on any voltage must be less than 10% of nominal and must decay to
within the regulation band within 50 msec.
In any failure situation, the power supply must shut down before the +5v output
reaches 6.5V to give the motherboard a reasonable chance of surviving; however, damage may occur at any voltage above 5.5V.
Power_Good Signal Requirements
S The signal must be at a TTL down level when power is applied until >100
msec to 500 msec after the 5V supply has reached its minimum regulation
level, and at TTL high level thereafter as long as outputs are within regulation.
S At turn–off, the Power Good signal must drop to a TTL low level before any
output drops below its regulation limits.
S The driver must be capable of driving 400 microamps or sinking 5 milliamps.
The rise time/fall time must be less than 1 usec, 10––90%
The +5V rise time (10%-90%) shall be 3 msec to 100 msec with a maximum
slope of 0.75 volts/msec for voltages above 1.5 volts for all loadings.
All supply voltages shall track within 50 msec of each other measured at the 50%
point.

11.1.2 Onboard 3.3V Regulator
There is a 3.3 volt regulator (U30) on the reference design to support the CPU and the 660
Bridge. See Table 38. Note that the 3.3v on the PCI bus slots is sourced via the power supply connector, not by U30.
Table 38. Specifications for 3.3V Regulator on the Motherboard
Specification

Value

Output Voltage

3.3 V  3%

Output Current

0.01 A to 5 A

Input voltage

4.75 V to 5.25 V

Pass element maximum case temperature

110 °C

Tracking

In regulation <1ms after +5 reaches 4.75V.

Overcurrent

No current limit feature

11.1.3 Onboard 2.5V Regulator
There is an uninstalled 2.5 volt regulator (U34) on the reference design to support possible
future 604ev installation. See Table 39.
Table 39. Specifications for 2.5V Regulator on the Motherboard
Specification

Value

Output Voltage

2.5 V  3%

Output Current

0.01 A to 5 A

Input voltage

3 V to 5.25 V

Pass element maximum case temperature

110 °C

Tracking

In regulation <1ms after +5 reaches 4.75V.

Overcurrent

No current limit feature

138

MPRH01TSU-02

Electromechanical

11.2 Thermal
The most thermally active components are outlined in this section. Designers should verify
that the temperature limits of all component are not exceeded in their application.
11.2.1 Thermal Requirements for the 603/604 Processor
When the motherboard is operated in an open environment for testing, a fan should be
placed so that there is a good air flow over the 603/604 CPU at all times. Enclosure design
is up to the manufacturer. It is likely that a fan will be required near the 603/604 CPU in
order to meet the requirements of operating junction temperature equal to or less than
105°C.
Under some conditions, release 2.1 of the reference design requires the 604 to be actively
cooled. A fan-sink was chosen to satisfy this requirement. A high quality model with a larger
than necessary thermal mass was chosen. This will tend to reduce the rate (and range) of
temperature changes of the 604 package once thermal management is implemented. A
Sanyot 109P5412H2026 was chosen, and is bonded to the 604 package with thermally
conductive adhesive. Similar devices are expected to perform in a similar manner.
11.2.1.1

604 Fan-Sink Installation

ÉÉÉÉÉ
ÉÉÉÉÉ
ÉÉÉÉÉ
Fan Sink

604 Chip

Adhesive dot

604 Substrate
Thermostrate

Figure 40. 604 Heat Sink Assembly
The procedure for bonding the fan-sink to the 604 package is as follows (see Section 1.4
for a list of materials and resources):
1.Dispense Loctitet 384 dots on the 604 substrate using four dots located at the four corners of the flat pack (see Figure 40). Maximize these dots, but make sure that the Loctite
does not contact the chip or the substrate leads.
2.Place the Thermostratet material on the chip.
3.Brush the activator onto the Loctite dots, and, within 15-20 seconds, place the fan-sink
onto the flat pack. Note that once the activator is applied to the 384, it will set up within
15 to 30 seconds. Normally, the activator is applied to the mating surface and not directly
to the 384. However, applying the activator directly onto the 384 ensures that the activator and the 384 are well mixed, and that there is a good cure over the large gap between
the fan-sink and the substrate (normally the Loctite does not cure well with a gap of over
0.020 inch).
4.Place a 5-6 lb weight on top of the fan-sink/substrate assembly for at least five minutes,
taking care that the fan-sink remains parallel to the substrate.
5.Remove the weight and allow the assembly to sit undisturbed for at least another five minutes.
The Loctite takes about 10 minutes to fully set up. It then takes 24 hours at room temperature to cure fully, but the assembly can be handled after the initial 10 minute setup time.
MPRH01TSU-02

139

Electromechanical
11.2.1.2
604 Fan Sink Experimentation
Some experimentation may be required to determine the best placement and size of the
adhesive dots required to attach the fan-sink to the 604. The following experimental procedure is recommended:
1.Dispense Loctite 384 dots on the substrate using four dots located at the four corners of
the substrate.
2.Place a piece of mylar over the flat pack (a piece of overhead projection material works
well).
3.Place the fan-sink in position.
4.Remove the fan-sink and look at how the dots spread out.
5.Remove the mylar and wipe the excess Loctite from the flat pack.
11.2.1.3
Thermal Requirements for the 3.3v Regulator
The 3.3v voltage regulator is assembled with a heat sink on the pass element which must
be maintained at an operating junction temperature equal to or less than 150°C. At the specified maximum current of 5A, the heat sink assembly combination provided results in an
assembled q js of 2.44°C/watt. Assembled q js is such that in operation, the case of the pass
element must be maintained at or below 110°C under any use condition (combination of
ambient temperature, altitude, air flow, or system configuration). Any change in heat sink
(type, style, manufacturer) or bonding technique (epoxy, glue, etc.) may alter the q js and
should be investigated to insure that the specified operating case temperature of 110°C is
not exceeded.

140

MPRH01TSU-02

Electromechanical

11.3 Mechanical
11.3.1 Reference Design Board Mechanical

MPRH01TSU-02

141

Electromechanical
11.3.2 Connector Locations

142

MPRH01TSU-02

Electromechanical
11.3.3 Connector Locator Diagram

pwr switch

J21 J22 J23 J24

PCI slots

Figure 41. Connector Location Diagram

MPRH01TSU-02

143

Electromechanical
11.3.4 Keyboard Connector J14
The keyboard connector uses a 6-pin miniature DIN connector (see Figure 42). Pins are
assigned as shown in Table 40 (as viewed from the back of the machine).

6

5
3
4
2 1
Figure 42. The Keyboard Connector
Table 40. Keyboard Connector Pin Assignments
Pin

I/O

SIGNAL NAME

1

I/O

DATA

2

NA

RESERVED

3

NA

GROUND

4

NA

+ 5V DC

5

I/O

CLOCK

6

NA

RESERVED

11.3.5 Mouse Connector J15
The mouse connector uses a 6-pin miniature DIN connector (see Figure 43). Pins are assigned as shown in Table 41 (as viewed from the back of the machine).

6

5
3
4
2 1
Figure 43. The Mouse Connector
Table 41. Mouse Connector Pin Assignments
Pin

I/O

Signal Name

1

I/O

DATA

2

NA

RESERVED

3

NA

GROUND Ã

4

NA

+ 5V DC

5

I/O

CLOCK

6

NA

RESERVED

144

MPRH01TSU-02

Electromechanical
11.3.6 Speaker Connector J13
Top view
4

x

x

x

x

1

Figure 44. 1x4 Speaker Connector
Table 42. Speaker Connector Pin Assignments
Pin No.

Signal Name

1

MINUS (–) INPUT TO SPEAKER

2

NO CONNECT

3

NO CONNECT

4

GND

11.3.7 Power Good LED/KEYLOCK# Connector J12
Top view
5

x

x

x

x

x

1

Figure 45. 1x5 Power Good LED Connector
Table 43. Power Good LED Connector
Pin No.

Signal Name

1

LED_POWER_GOOD/RESET#

2

NO CONNECT

3

GND

4

KEYLOCK* (Not Used)

5

GND

11.3.8 HDD LED Connector J11 (1 x 2 Berg)
Top view
2

1

Figure 46. 1x2 HDD LED Connector
Table 44. HDD LED Connector
Pin No.

Signal Name

1

+ LED VOLTAGE

2

HDD LED drive signal (low to drive LED)

MPRH01TSU-02

145

Electromechanical
11.3.9 Reset Switch Connector J10 (1 x 2 Berg)
Top view
2

1

Figure 47. 1x2 Reset Switch Connector
Table 45. Reset Switch Connector
Pin No.

Signal Name

1

RESET N/O CONTACT (close to reset)

2

GROUND

11.3.10 Fan Connector J9
Top view
1

2

Figure 48. 1x2 Fan Connector
Table 46. Fan Connector Pin Assignments
Pin No.

Signal Name

1

+12 VOLTS

2

GROUND

146

MPRH01TSU-02

Electromechanical
11.3.11 3.3V Power Connector J5
Top view
P10
1

x

x

x

x

x

x

6

plastic tabs

Figure 49. 1x6 3.3V Power Connector J5

Table 47. 3.3V Power Connector J5 Pin Assignments
Pin No.

Signal Name

1

+3.3 V

2

+3.3 V

3

+3.3 V

4

GROUND

5

GROUND

6

GROUND

11.3.12 Power Connector J4
Top view
open side
P1
1

x

x

x

P2
x

x

x

x

x

x

x

x

x

12

plastic tabs

Figure 50. 1x12 Power Connector
Table 48. Power Connector J4 Pin Assignments
Pin No.

Signal Name

1

POWER GOOD

2

+5 VOLTS

3

+12 VOLTS

4

–12 VOLTS

5

GROUND

6

GROUND

7

GROUND

8

GROUND

9

–5 VOLTS

10

+5 VOLTS

11

+5 VOLTS

12

+5 VOLTS

MPRH01TSU-02

147

Electromechanical
11.3.13 AUX5/ON-OFF Connector J6
Top view
1

2

3

Figure 51. AUX5/ON-OFF Connector
Table 49. AUX5/ON-OFF Connector Pin Assignments
Pin No.

Signal Name

1

AUX +5V

2

ON/OFF

3

GROUND

11.3.14 PCI Connectors J25, J26, and J27

A1

A49 A52
..
x x x x x..
..x x
x x x x
x x x x x..
..x x
x x x x

B1

B49

B52

A62

B62

Figure 52. PCI Connector
Pins are assigned as shown in Table 50.
Table 50. PCI Connector Pin Assignments
Pin

Function

Pin

Function

A1

TRST#

B1

–12 VOLTS#

A2

+12 VOLTS

B2

TCK

A3

TMS

B3

GROUND

A4

TDI

B4

TDO

A5

+5 VOLTS

B5

+5 VOLTS

A6

INTA#

B6

+5 VOLTS

A7

INTC#

B7

INTB#

A8

+5 VOLTS

B8

INTD#

A9

RESERVED

B9

PRESENT 1# *

A10

+5 VOLTS

B10

RESERVED

A11

RESERVED

B11

PRESENT 2# *

A12

GROUND

B12

GROUND

A13

GROUND

B13

GROUND

A14

RESERVED

B14

RESERVED

A15

RESET#

B15

GROUND

A16

+5 VOLTS

B16

CLK

148

MPRH01TSU-02

Electromechanical
Table 50. PCI Connector Pin Assignments (Continued)
Pin

Function

Pin

Function

A17

GNT#

B17

GROUND

A18

GROUND

B18

REQ#

A19

RESERVED

B19

+5 VOLTS

A20

A/D(30)

B20

A/D(31)

A21

+3.3V

B21

A/D(29)

A22

A/D(28)

B22

GROUND

A23

A/D(26)

B23

A/D(27)

A24

GROUND

B24

A/D(25)

A25

A/D(24)

B25

+3.3V

A26

IDSEL

B26

C/BE#(3)

A27

+3.3V

B27

A/D(23)

A28

A/D(22)

B28

GROUND

A29

A/D(20)

B29

A/D(21)

A30

GROUND

B30

A/D(19)

A31

A/D(18)

B31

+3.3V

A32

A/D(16)

B32

A/D(17)

A33

+3.3V

B33

C/BE#(2)

A34

FRAME#

B34

GROUND

A35

GROUND

B35

IRDY#

A36

TRDY#

B36

+3.3V

A37

GROUND

B37

DEVSEL#

A38

STOP#

B38

GROUND

A39

+3.3V

B39

LOCK#

A40

SDONE

B40

PERR#

A41

SBO#

B41

+3.3V

A42

GROUND

B42

SERR#

A43

PAR

B43

+3.3V

A44

A/D(15)

B44

C/BE#(1)

A45

+3.3V

B45

A/D(14)

A46

A/D(13)

B46

GROUND

A47

A/D(11)

B47

A/D(12)

A48

GROUND

B48

A/D(10)

A49

A/D(9)

B49

GROUND

A50



B50



A51



B51



A52

C/BE#(0)

B52

A/D(8)

A53

+3.3V

B53

A/D(7)

A54

A/D(6)

B54

+3.3V

A55

A/D(4)

B55

A/D(5)

A56

GROUND

B56

A/D(3)

A57

A/D(2)

B57

GROUND

A58

A/D(0)

B58

A/D(1)

MPRH01TSU-02

149

Electromechanical
Table 50. PCI Connector Pin Assignments (Continued)
Pin

Function

Pin

Function

A59

+5 VOLTS

B59

+5 VOLTS

A60

REQ64#

B60

ACK64#

A61

+5 VOLTS

B61

+5 VOLTS

A62

+5 VOLTS

B62

+5 VOLTS

Notes:
* The two card type bits, B9 and B11, are connected on the riser.
** Int A-D, A6, A7, B7, and B8 are connected together on the riser.

11.3.15 ISA Connectors J29, J30, J31, J32, and J33
A1

A31 C1
..
x x x x x..
..x x
x x x x
x x x x x..
..x x
x x x x

B1

B31

D1

C18

D18

Figure 53. ISA Connector
Pins are assigned as shown in Table 51.
Table 51. ISA Connector Pin Assignments
Pin

Function

Pin

Function

A1

IO CHCK#

B1

GROUND

A2

SD(7)

B2

RESET_DRV

A3

SD(6)

B3

+5 VOLTS

A4

SD(5)

B4

IRQ9

A5

SD(4)

B5

–5 VOLTS

A6

SD(3)

B6

DRQ2

A7

SD(2)

B7

–12 VOLTS

A8

SD(1)

B8

ZERO WS#

A9

SD(0)

B9

+12 VOLTS

A10

IO CHRDY

B10

GROUND

A11

AEN

B11

SMEMW#

A12

SA(19)

B12

SMEMR#

A13

SA(18)

B13

IOW#

A14

SA(17)

B14

IOR#

A15

SA(16)

B15

DACK3#

A16

SA(15)

B16

DRQ3

A17

SA(14)

B17

DACK1#

A18

SA(13)

B18

DRQ1

A19

SA(12)

B19

REFRESH

A20

SA(11)

B20

CLK

A21

SA(10)

B21

IRQ7

150

MPRH01TSU-02

Electromechanical
Table 51. ISA Connector Pin Assignments (Continued)
Pin

Function

Pin

Function

A22

SA(9)

B22

IRQ6

A23

SA(8)

B23

IRQ5

A24

SA(7)

B24

IRQ4

A25

SA(6)

B25

IRQ3

A26

SA(5)

B26

DACK2#

A27

SA(4)

B27

T/C

A28

SA(3)

B28

BALE

A29

SA(2)

B29

+5 VOLTS

A30

SA(1)

B30

Oscillator

A31

SA(0)

B31

GROUND

C1

SBHE#

D1

MEMCS16#

C2

LA(23)

D2

IO CS16#

C3

LA(22)

D3

IRQ10

C4

LA(21)

D4

IRQ11

C5

LA(20)

D5

IRQ12

C6

LA(19)

D6

IRQ15

C7

LA(18)

D7

IRQ14

C8

LA(17)

D8

DACK0#

C9

MEMR#

D9

DRQ0

C10

MEMW#

D10

DACK5#

C11

SD(8)

D11

DRQ5

C12

SD(9)

D12

DACK6#

C13

SD(10)

D13

DRQ6

C14

SD(11)

D14

DACK7#

C15

SD(12)

D15

DRQ7

C16

SD(13)

D16

+5 VOLTS

C17

SD(14)

D17

MASTER#

C18

SD(15)

D18

GROUND

MPRH01TSU-02

151

Electromechanical
11.3.16 SIMM Connectors J21, J22, J23, and J24
Top view

1

36 37
x ..........x
x...........x
1

72

x..............x
x..............x

36 37

72

Figure 54. SIMM Connector

Pins are assigned as shown in Table 52.
Table 52. SIMM Connector Pin Assignments
SIMM
Pin

motherboard Pin

Function

1

1

GROUND

2

2

DQ0

3

3

DQ18

4

4

DQ1

5

5

DQ19

6

6

DQ2

7

7

DQ20

8

8

DQ3

9

9

DQ21

10

10

+5 V

11

11

CASP

12

12

A0

13

13

A1

14

14

A2

15

15

A3

16

16

A4

17

17

A5

18

18

A6

19

19

A10

20

20

DQ4

21

21

DQ22

22

22

DQ5

23

23

DQ23

24

24

DQ6

25

25

DQ24

26

26

DQ7

27

27

DQ25

152

MPRH01TSU-02

Electromechanical
Table 52. SIMM Connector Pin Assignments (Continued)
SIMM
Pin

motherboard Pin

Function

28

28

A7

29

29

BS0/A11

30

30

+5V

31

31

A8

32

32

A9

33

33

RAS3#

34

34

RAS2#

35

35

DQ26

36

36

DQ8

37

37

DQ17

38

38

DQ35

39

39

GROUND

40

40

CAS0#

41

41

CAS2#

42

42

CAS3#

43

43

CAS1#

44

44

RAS0#

45

45

RAS1#

46

46

BS1

47

47

WE#

48

48

RESD1

49

49

DQ9

50

50

DQ27

51

51

DQ10

52

52

DQ28

53

53

DQ11

54

54

DQ29

55

55

DQ12

56

56

DQ30

57

57

DQ13

58

58

DQ31

59

59

+5 V

60

60

DQ32

61

61

DQ14

62

62

DQ33

63

63

DQ15

64

64

DQ34

MPRH01TSU-02

153

Electromechanical
Table 52. SIMM Connector Pin Assignments (Continued)
SIMM
Pin

motherboard Pin

Function

65

65

DQ16

66

66

BS2

67

67

PD1

68

68

PD2

69

69

PD3

70

70

PD4

71

71

BS3

72

72

GROUND

11.3.17 Power Switch Connector J8

Top view
2

1

Figure 55. 1x2 Power Switch Connector
Table 53. Power Switch Connector Pin Assignments
Pin No.

Signal Name

1

SWITCH_P1

2

SWITCH_GD

11.3.18 Power Up Configuration Connector J7

Top view
2

1

Figure 56. 1x2 Power Up Configuration Connector
Pins are assigned as shown in Table 54.
Table 54. Power Up Configuration Connector Pin Assignments
Pin No.

Signal Name

1

PWR_CFG1

2

PWR_CFG2

154

MPRH01TSU-02

Electromechanical
11.3.19 L2 Cache Data SIMM Connector J3

Top view

160

124 122

82

x.....................x x.............x
x.....................x x.............x
79

43

41

1

Figure 57. L2 SRAM Module Connector

Pins are assigned as shown in Table 55.
Table 55. L2 SRAM Module Connector Pin Assignments
Pin

Function

Pin

Function

81

GND

1

GND

82

D63

2

D62

83

VCC5

3

VCC3

84

D61

4

D60

85

VCC5

5

VCC3

86

D59

6

D58

87

D57

7

D56

88

GND

8

GND

89

DP7

9

DP6

90

D55

10

D54

91

D53

11

D52

92

D51

12

D50

93

GND

13

GND

94

D49

14

D48

95

D47

15

D46

96

D45

16

D44

97

D43

17

D42

98

GND

18

GND

99

D41

19

D40

100

DP5

20

DP4

101

D39

21

D38

102

D37

22

D36

103

D35

23

D34

104

GND

24

GND

105

D33

25

D32

106

D31

26

D30

107

D29

27

D28

108

D27

28

D26

109

D25

29

D24

110

GND

30

GND

111

DP3

31

DP2

MPRH01TSU-02

155

Electromechanical
Table 55. L2 SRAM Module Connector Pin Assignments (Continued)
Pin

Function

Pin

Function

112

D23

32

D22

113

D21

33

D20

114

VCC5

34

VCC3

115

D19

35

D18

116

GND

36

GND

117

D17

37

D16

118

VCC5

38

VCC3

119

D15

39

D14

120

D13

40

D12

121

GND

41

GND

122

D11

42

D10

123

VCC5

43

VCC3

124

D9

44

D8

125

DP1

45

DP0

126

VCC5

46

VCC3

127

D7

47

D6

128

D5

48

D4

129

D3

49

D2

130

D1

50

D0

131

GND

51

GND

132

A0B

52

A0A

133

A1B

53

A1A

134

A2B

54

A2A

135

A3B

55

A3A

136

A4

56

A5

137

GND

57

GND

138

A6

58

A7

139

A8

59

A9

140

A10

60

A11

141

A12

61

A13

142

A14

62

A15

143

GND

63

GND

144

A16

64

PD0

145

PD1

65

PD2

146

CLK0

66

CLK1

147

CLK2

67

CLK3

148

GND

68

GND

149

WE7*

69

WE6*

150

WE5*

70

WE4*

151

WE3*

71

WE2*

152

WE1*

72

WE0*

153

GND

73

GND

156

MPRH01TSU-02

Electromechanical
Table 55. L2 SRAM Module Connector Pin Assignments (Continued)
Pin

Function

Pin

Function

154

ADSC1*

74

ADSC0*

155

CE1*

75

CE0*

156

ADV1*

76

ADV0*

157

OE1*

77

OE0*

158

VCC5

78

VCC3

159

ADSP1*

79

ADSP0*

160

GND

80

GND

11.3.20 RISCWatch Connector J19

Top view
1
9

2
–

3 4 5 6 7 8
11 12 13 14 15 16

Figure 58. 2x8 RISCWatch Connector
Table 56. RISCWatch Connector Pin Assignments
Pin No.

Signal Name

1

CHECK_STOP#

2

HDWR_RESET#

3

RESET_INTERRUPT#

4

CNTL/SCAN_DATA

5

SHIFT_CLK

6

+RUN/–BREAKPOINT

7

SCAN_IN

8

SCAN_OUT

9

GROUND

10



11

GROUND

12

RESERVED

13

RESERVED

14

+5 VOLTS

15

OCS_OVERRIDE

16

RESERVED

11.3.21 Battery Connector BT2
The battery type is CR2032 3 Volt.
Insert the battery with + side up.
MPRH01TSU-02

157

Electromechanical
11.3.22 Reference Design Board Connector Footprint #1

158

MPRH01TSU-02

Electromechanical
11.3.23 Reference Design Board Connector Footprint #2

MPRH01TSU-02

159

Electromechanical

11.4 Enclosure
The Reference Design Board has a standard BabyAT form factor. It will fit within enclosures
manufactured by a variety of companies. The keyboard/mouse connector type should be
considered when selecting a enclosure. Additional cooling options may be required.
The following venders supply enclosures:
Olson Metal Products Company Inc.
Attn: Michelle Seay
1903 N. Austin Street
Seguin, Texas
78155
1-800-951-9517 or (210) 379-7000
AT Desktop model number:
Medium Tower model number:

CC300249
CC400000

Altex Electronics
11342 IH-35 North
San Antonio, Texas
1-800-531-5367 or FAX (210) 637-3264
Mini Tower model number
Medium Tower model number
Full Tower model number

STC–05
STC–08
STC–16

Mega-Tech Marketing Inc.
3900-D Drossett Dr.
Austin, Texas
Mini Tower model number
Full Tower model number

A6601
A5561

Note: IBM makes no recommendations regarding vendors of any components.
In planning the layout of an enclosure, the height of the items in Table 57 should be considered.
Table 57. Height Considerations
Item

Approximate Height

3.3V regulator heat sink

26.0 mm

SIMMs (seated)

28.4 mm

L2 SRAM SIMM (seated)

28.4 mm

160

MPRH01TSU-02

Physical Design

Section 12
Physical Design Guidelines
12.1 General Considerations
These guidelines are given to aid designers with the physical design phase of their PowerPC reference design board. The guidelines are not intended to replace good physical design practices for the signal types and frequencies discussed, but to supplement standard
practice by pointing out sensitive and critical areas. Some discussion of the IBM implementation of the reference board is also included to establish the context for the wiring
guidelines.
Construction
The general construction of the PowerPC 603/604 reference board (reference board) is
shown in Figure 59. It is constructed with two high frequency signal layers in the center, two
power planes, and two external general purpose signal layers.
12.1.1

Signal layer, general purpose
Power plane
Signal layer, clocks and high frequency
Signal layer, clocks and high frequency
Power plane (ground)
Signal layer, general purpose
Figure 59. Signal and Power Layers

MPRH01TSU-02

161

Physical Design
A top view of a typical wiring channel, as implemented on the reference board, is shown
in Figure 60 (all dimensions are shown in inches). Minimum trace width is .006” (at 1:1) and
minimum space width is .004” (at 1:1).
Space .004
CL

Trace .006
Space
Wiring Channel
Adjacent Wiring Channel

Figure 60. Typical Wiring Channel Top View
For fabrication information, see Figure 61.

162

MPRH01TSU-02

Physical Design

Signal

Power

Signal

Signal

Ground

Signal

ÉÉÉÉÉÉÉÉÉÉ
ÉÉÉÉÉÉÉÉÉÉ
ÉÉÉÉÉÉÉÉÉÉ
ÉÉÉÉÉÉÉÉÉÉ
ÉÉÉÉÉÉÉÉÉÉ
ÉÉÉÉÉÉÉÉÉÉ
ÉÉÉÉÉÉÉÉÉÉ
ÉÉÉÉÉÉÉÉÉÉ
ÉÉÉÉÉÉÉÉÉÉ
ÉÉÉÉÉÉÉÉÉÉ
ÉÉÉÉÉÉÉÉÉÉ
ÉÉÉÉÉÉÉÉÉÉ
ÉÉÉÉÉÉÉÉÉÉ
ÉÉÉÉÉÉÉÉÉÉ
ÉÉÉÉÉÉÉÉÉÉ
ÉÉÉÉÉÉÉÉÉÉ
ÉÉÉÉÉÉÉÉÉÉ
ÉÉÉÉÉÉÉÉÉÉ
ÉÉÉÉÉÉÉÉÉÉ
ÉÉÉÉÉÉÉÉÉÉ

ÉÉÉ
ÉÉÉ
ÉÉÉ
ÉÉÉ

0.5 oz. Foil
0.0045
2 oz. Foil
0.0062

0.00288 A
1 oz. Foil
0.0062

0.0045

A

0.5 oz. Foil

Fiberglass
Epoxy
Resin

C Trace to ground impedance 70 Ω ± 10%
D Finished signal trace width: 0.00425
E Overall thickness: 0.060
after plating: 0.062
F Plating 7070

Figure 61. PowerPC 603/604 Board Fabrication

163

A

2 oz. Foil

B All dimensions in inches

MPRH01TSU-02

A

1 oz. Foil

Notes:

A

A

Physical Design
General Wiring Guidelines
1. A power (ground or voltage) plane split occurs where there is a discontinuity in the plane.
As shown in Figure 62, route wires across splits in a power plane in the most perpendicular manner possible. Do not run wires parallel to the split in close proximity to the split.

12.1.2

Plane 1

Plane 2

Not ok
Ok
Figure 62. Power Plane Split
2. Minimize the number of wires on the top surface of the board that cross under the clock
generator. Do not run any wires close to the crystal connections.
3. If a wire is routed near a via which is part of a clock net, there must be at least one vacant
wiring channel between the wire and the clock via.
4. Several groups of nets require special attention to ensure correct system operation.
They are listed below in order of the importance of meeting the design rules that are suggested for that group. For example, ensuring that the clock nets are routed according
to the design rules suggested in section 12.2 is more important than ensuring that the
PCI bus nets are routed according to the design rules suggested in section 12.5. The
ideal system design will meet all of the suggested design rules, but this information is
included to guide the designer in case some tradeoffs have to be evaluated.
S Clock nets.
S CPU bus nets and timing critical nets.
S PCI bus nets.
S Noise sensitive nets.
S Other nets.

12.2 Clock Nets
1. Clock nets are the most critical wiring on the board. Their wiring requirements should
be given priority over the requirements of other groups of signals.
2. Clock nets are to have a minimum number of vias.
3. No clock wires may be routed closer than one inch to the edge of the board.
4. Clock nets with more than two nodes (devices connected to them) are to be daisychained. Stubs and star fanouts are not allowed.
164

MPRH01TSU-02

Physical Design
5. Clock nets are to be routed as much as possible on internal signal planes.
6. Where series termination resistors are required, place them as close as possible to the
clock generator.
7. Route a ground trace as a shield in the adjacent wiring channel on both sides of the clock
trace. It is a good practice to periodically (every inch or so) connect these shield traces
to the ground plane. Completely surround the clock trace with shield traces.
8. Table 15 shows the length and tolerances of the clock nets. All dimensions are shown
in inches. The Design Rule column shows the guideline, and the Reference Board column shows the actual length of the net on the reference board.
Table 58. Clock Net Lengths ( 7  a  10 )
Net

Design Rule

Reference Board

Tolerance (Inch)

SYSCLK/PCLK_EN

a

9.5

0.3

L2_BCLK<4>

a

9.5

0.3

a–5

4.5

0.3

663_BCLK

a

9.5

0.3

664_BCLK

a

9.5

0.3

664_PCI_CLK

a

9.5

0.3

PCI_CLK<4:1>

a+0.5

10.0

0.3

L2_BCLK<3:0>

ISA_CLK

As Short as possible

As Short as possible

12.3 CPU Bus Nets
1. The CPU bus nets shown in Table 59 are to be daisy-chained. Stubs and star fanouts
are not allowed.
2. The wiring order is U4, U33, U32, U5, U37, U38, and J3.
3. Route nets so as to minimize noise reception. Make these nets as short as possible.
Table 59. CPU Bus Nets
A<31:0>
D<63:0>
DP<7:0>
AACK_60X#
ARTRY_60X#
BR_60X#
BG_60X#
SHD_6014#
TS_60X#
XATS_60X#
TT<4:0>
TSIZ<2:0>
TBST_60X#
GBL_60X#

MPRH01TSU-02

165

Physical Design
Table 59. CPU Bus Nets (Continued)
DRTRY_60X#
TA_60X#
TEA_60X#

12.4 Timing Critical Nets
The timing of the nets shown in Table 60 is critical. Make these nets as short as possible
to reduce board delays as much as possible.
Table 60. Timing Critical Nets
CKSTP_OUT_60X#
INT_60X#
DPE_60X#
MCP_6034#
SMI_6034#
ESP_TMS_60X
ESP_TCK_60X
ESP_TDI_60X
ESP_TDO_60X
HALTED/RUN_NSTOP
SRAM_ADS/ADDR0
SRAM_CNT_EN/ADDR1
SRAM_ALE
SRAM_WE#
SRAM_OE#
TAG_MATCH
TAG_WE#
TAG_VALID
TAG_CLEAR#

166

MPRH01TSU-02

Physical Design

12.5 PCI Bus Nets
1. The PCI bus nets shown in Table 61 are to be daisy-chained. Stubs and star fanouts are
not allowed.
2. The wiring order is M2, U4, U5, U7, J25, J26, and J27.
3. Route these nets so as to minimize noise reception and timing delays. Make these nets
as short as possible.
Table 61. PCI Bus Nets
AD<31:0>
PCI_FRAME#
PCI_TRDY#
PCI_IRDY#
PCI_STOP#
PCI_DEVSEL#
PCI_PERR#
PCI_SERR#
PCI_C/BE<3:0>#
PCI_LOCK#
PCI_PAR

12.6 Group 2A: Noise Sensitive Wires
These nets are noise sensitive. Route them so as to minimize noise reception. Make these
nets as short as possible.
Table 62. Noise Sensitive Nets
BUFF_PS_POWER_GOOD
MAN_RESET_R
BUFF_MAN_RESET
POWER_GOOD/RESET
SYS_RESET
PCI_RESET
HRESET_60X
SRESET_60X
ESP_TRST
HDWR_RESET
ESP_TRST_60X
RESETDRV
ISA_RESET

MPRH01TSU-02

167

Physical Design

12.7 8mm Tape Contents and Extract Instructions
Part Number = MPRH07INU-01
Download Instructions
The enclosed 8mm tape was created using the tar command. First find the address of the
8mm tape drive by executing the following command:
12.7.1

isdev –C –c tape.
Then change the block size to 1024 by executing the following command:
chdev –l rmt0 –a block_size=1024
To extract the data, create a directory and ensure that at least 80M of free space is available. Use the cd command to get to the created directory and type the following:
tar –xvf/dev/rmt0
Cadence Version
S Concept=version 1.7–S1
S Allegro=version 8.0 or later
S packagerXL

12.7.2

Tape Contents
The tape contains the following directories:
12.7.3

PXL
S Pst.* files, packager files
S *View* files, Allegro feedback files
S bom.* files, bill of material files

SCR
S Script files

Flatlab
S Logic models used for this design

Harley (603/603 Reference Board)
S Schematic data
S Postscript plots Each page=plot.* total=*.ps

tscr
S Cross reference sheets.

168

MPRH01TSU-02

Errata

Section 13
Errata for Reference Design Release 2.1
This is the errata section of release 2.1 of the 603/604 Reference Design. This is not the
final release of this reference design. This release contains errata that will be cured in subsequent releases according to the reference design roadmap. Some of these errata are
due to board level errata, and some are due to individual device errata.
Note that all of the workarounds for the errata described herein were implemented during
the physical design of the board. Circuitry, devices, and connections shown in this section
are shown in the schematics in section 15, and do not need to be added to the board.
Section 13.1 discusses the roadmap for the reference design.
Section 13.2 describes board level errata and workarounds.
Section 13.3 describes the logic workarounds associated with 660 bridge errata and gives
the logic design files for the PALs that were chosen to implement the workarounds. The
connectivity of these PALs is shown in the schematics in section 15. Descriptions of the
individual errata can be found in section 13.4. Note that due to physical design constraints,
the PALs used to implement the workarounds on the reference design are slightly different
than those used to implement the workarounds on the generic system discussed in section
13.4, although the resulting logical function of both sets of PALs is identical.
Section 13.4 is exerpted from the 8/9/95 release of the 660 Bridge Revision 1.1 Errata Summary. This section describes each of the 660 bridge errata, the suggested workarounds,
and sample logic design files for PALs to implement the workarounds in a generic system.

13.1 PowerPC 603/604 Reference Design Roadmap
Table 63 shows the release roadmap for the reference design. There are three releases
expected for the 603/604 Reference Design, 2.0, 2.1, and 3.0. Each release consists of
both a reference design, consisting of the intangible design and the documentation thereof
(see section 1.1), and a reference board, which consists of the actual populated circuit
board.
Each release of the reference design contains functional specifications that describe the
intended behavior of the final (release 3.0) design, the schematic and BOM of the current
release, and an errata section that details the functional and implementation differences
between the current level and the final level.
MPRH01TSU-02

169

Errata
Table 63. PowerPC 603/604 Reference Design Roadmap
Reference Board Level

Release 2.0

Release 2.1

Release 3.0

Circuit Board Level

2.0

2.1

3.0

IBM27-82664 Controller Level

1.0

1.1

1.2

IBM27-82663 Buffer Level

1.0

1.0

2.0

(1.0)

(1.1)

(2.0)

Limited

Full

Full

Significant

Limited

None

IBM Internal Use

Limited

General

Release 2.0

Release 2.1

Release 3.0

Functional Spec Level

3.0

3.0

3.0

BOM and Schematic Level

2.0

2.1

3.0

(IBM27-82660 Chipset Level)
Reference Board Function
Errata Quantity
Intended Customer Availability
Reference Design (Documentation) Level

Errata Section Content

Reference Board Reference Board
Release 2.0
Release 2.1
Errata and
Errata and
Workarounds
Workarounds

None

Release 2.0 of the reference design contains version 1.0 of the 660 Bridge chipset, and
requires a significant quantity of workarounds for the corresponding errata. Release 2.0
has limited functionality and is intended for IBM internal use. The schematic and the BOM
match the actual release 2.0 reference board, while the functional specification describes
the expected operation of the final design. The errata section describes the errata and workarounds required for the release 2.0 reference board, and details the differences between
the actual operation of the release 2.0 board and the expected operation of the final design.
Likewise, release 2.1 of the reference design contains version 1.1 of the 660 Bridge chipset, and requires a limited quantity of workarounds for the corresponding errata. Release
2.1 has full functionality and is intended for limited release. The schematic and the BOM
match the actual release 2.1 reference board, while the functional specification describes
the expected operation of the final design. The errata section describes the errata and workarounds required for the release 2.1 reference board, and details the differences between
the actual operation of the release 2.1 board and the expected operation of the final design.
Release 3.0 of the reference design contains version 2.0 of the 660 Bridge chipset, requires
no workarounds and contains no errata. Release 3.0 has full functionality, has been guardband tested, and is intended for general release. The schematic and the BOM and the functional specification all match the final, release 3.0 reference board.
This document contains several references to the 603/604 Reference Design Power Management Specification, which was not available at the date of printing of the reference design documentation.

170

MPRH01TSU-02

Errata

13.2 Release 2.1 Board Level Errata
13.2.1 Driving 32MB 72-pin DRAM SIMMs
Release 2.0 of the reference design documentation incorrectly stated that up to 128MB of
DRAM could be installed on the reference board. To do so requires the use of four 32MB
(72 pin) devices, which presents a greater load to the memory controller than is supported.
To drive four 32MB SIMMs requires the installation of buffers on the WE[1:0] and MA[11:0]
lines. These buffers are not supplied on the reference board.

13.2.2 CPU Data Bus 33 Ohm Series Resistors
Due to a change in the recommended implementation of the 660 bridge, 33 ohm series resistors have been added to the reference design. These resistors are placed between the
663 buffer and all other connections to the CPU data (and parity) bus. These resistors will
remain in place for the final version of the reference design.

MPRH01TSU-02

171

Errata

13.3 Reference Design Errata for Revision 1.1 of the 660 Bridge
Release 2.1 of the reference design uses revision 1.1 of the 660 bridge (663 rev. 1.0 and
664 rev. 1.1). This section describes the workarounds associated with 660 bridge errata
and gives the logic design files for the PALs that were chosen to implement the workarounds. The connectivity of these PALs is shown in the schematics in section 15. Descriptions of the individual errata can be found in section 13.4. Note that due to physical design
constraints, the PALs used to implement the workarounds on the reference design are
slightly different than those used to implement the workarounds on the generic system discussed in section 13.4, although the resulting logical function of both sets of PALs is identical.
13.3.1 Workaround Implementation
The following items were implemented on the reference design to work around the errata
associated with revision 1.1 of the 660 bridge.
1.
2.
3.
4.

Advisory 1 is implemented by using the corrected revision of the SCSI controller.
Errata 3 is not applicable since no CPU bus targets are used.
Errata 4 workaround is implemented by not executing ECIWX or ECOWX instructions.
Errata 9 workaround is implemented as suggested depending on the CPU family. For
604 boards, TT[2] connects between the CPU bus and the 664, and has no pull-down
resistors. For 603e boards, TT[2] is cut (by not populating a jumper) between the CPU
and the 660. TT[2]_CPU is pulled down using a 500 ohm resistor. TT[2]_660 is pulled
down using a 500 ohm resistor.
5. Errata 12 is not applicable, since the 2.1 reference design uses revision 1.0 of the 663,
which does not require the 100 ohm series resistor.
6. Errata 14 is implemented in 604 designs by not installing pullup resistors on TT[0:4].
On 603e reference designs, TT[2] is connect as described in errata 9, and TT[0,1,3,4]
have no pull-up resistors.
7. Errata 15 workaround is not comprehensive, and involves changing ROM accesses
from 1-byte reads to 4-byte reads. Other workarounds are TBD.
8. Errata 16 workaround is to not use PCI_LOCK#.
9. Three PALs implement the remaining workarounds for the 660 bridge.
S PAL HPP1 is located at M6.
S PAL HCn is located at M7.
S PAL HBROOM is located at M8.
Note that this set of PALs is used for both the 603e and 604 versions of the reference design.

13.3.2 HPP1 Reference Design Workaround PAL
This PAL is identical to the HPP1 PAL described in section 13.4.

172

MPRH01TSU-02

Errata
13.3.3 HCn Reference Design Workaround PAL
TITLE
HC.pds
PATTERN none
REVISION 1.0
COMPANY IBM
DATE
08/07/95
CHIP _HC PALCE16V8
;–––––––––––––––––––––––––––– PIN Declarations ––––––––––––––––––––
;
; Predefined
;
;PIN 1
CPU_CLK
; CLOCK
;PIN 10
GND
; GROUND
;PIN 11
REG_OE#
; OUTPUT ENABLE FOR REGISTERED OUTPUTS
;PIN 20
VCC
; VCC
;
;

Inputs

PIN
PIN
PIN
PIN
PIN
PIN
PIN
PIN
;PIN

2
3
4
5
6
7
8
9
9

BG_IN_
ABB_
BR_IN_
MASK_BG2
MASK_BG1
MASK_BR
DBB_
MASK_BG3
PCI_GNT_IN_

;
; outputs
;
PIN 12
CPU_A0
PIN 13
BR_OUT_
;PIN 17
PCI_GNT_OUT_
PIN 19
BG_OUT_
;
; Internal Registers
;
; PIN 14
PIN 15
DEL_ABB_
PIN 16
DEL_BG_
PIN 18
MASK_PCI_GNT

MPRH01TSU-02

; INPUT
; INPUT
; INPUT
; INPUT
; INPUT
; INPUT
; INPUT
; INPUT
; INPUT

COMB
COMB
COMB
COMB

; COMBINATIONAL
; COMBINATIONAL
; COMBINATIONAL
; COMBINATIONAL

REG
REG
REG

; REGISTERED
; REGISTERED
; REG

173

Errata
;––––––––––––––––––––––– Boolean Equation Segment –––––––––––––––––––
;
EQUATIONS
;
;
Output enables for Comb logic (Reg OE’s controllered by pin 11)
;
CPU_A0.TRST = ABB_ * /DEL_ABB_
;
;
;

;for Errata #6

Equations for internal registers

DEL_BG_

= BG_IN_

DEL_ABB_ = ABB_
MASK_PCI_GNT = /ABB_ * DEL_ABB_ * CPU_A0 * /DBB_ ;for Errata #11
+ MASK_PCI_GNT * /DBB_
; Delay granting PCI to
; the CPU when a CPU to
;PCI_GNT_OUT_ = PCI_GNT_IN_
; PCI cycle is pipelined
;
+ MASK_PCI_GNT
; over the data tenure of
; the previous cycle
BG_OUT_

=
+
+
+

BG_IN_
MASK_BG1
MASK_BG2
MASK_BG3

;for Errata #5
;for Errata #5
;for Errata #7

CPU_A0 = GND

;for Errata #6

BR_OUT_ = BR_IN_
+ MASK_BR

;for Errata #8

;––––––––––––––––––––––––– END OF FILE ––––––––––––––––––––––––––––

174

MPRH01TSU-02

Errata
13.3.4 HBROOM Reference Design Workaround PAL
TITLE
HBROOM.pds
PATTERN none
REVISION 1.0
COMPANY IBM
DATE
08/07/95
CHIP _HBROOM PALCE16V8
;–––––––––––––––––––––––––––– PIN Declarations ––––––––––––––––––––
;
; Predefined
;
;PIN 1
CPU_CLK
; CLOCK
;PIN 10
GND
; GROUND
;PIN 11
REG_OE#
; OUTPUT ENABLE FOR REGISTERED OUTPUTS
;PIN 20
VCC
; VCC
;
; Inputs
;
;PIN 2
; INPUT
PIN 3
PCI_GNT_IN_
; INPUT
PIN 4
ABB_
; INPUT
;PIN 5
; INPUT
;PIN 6
; INPUT
PIN 7
DBB_
; INPUT
PIN 8
TA_
; INPUT
PIN 9
SRAM_OE_
; INPUT
PIN 16
MASK_PCI_GNT
; INPUT
;
; Outputs
;
PIN 17
PCI_GNT_OUT_
COMB
; COMB
;
; Internal Registers
;
PIN 12
DEL_ABB_
REG
; REG
PIN 13
MASK_7A
REG
; REGISTERED
PIN 14
MASK_7B
REG
; REGISTERED
PIN 15
MASK_BG3
REG
; REG
PIN 18
TA_CNT0
REG
;
PIN 19
TA_CNT1
REG
;

MPRH01TSU-02

175

Errata
;––––––––––––––––––––––– Boolean Equation Segment –––––––––––––––––––
;
Output enables for Comb logic (Reg OE’s controllered by pin 11)
;
Equations for internal registers
;
;********************************************************************
DEL_ABB_

= ABB_

;********************************************************************
;
; Below is fix for Errata #7B, 11, 13,
;
TA_CNT0 = /TA_CNT0 * /TA_
+ TA_CNT0 * /DBB_ * TA_
TA_CNT1 = TA_CNT0 * /TA_
+ TA_CNT1 * /DBB_
MASK_BG3 =
+
+
+
+

(/TA_CNT0 + /TA_CNT1) * SRAM_OE_ * /DBB_
/ABB_ * /SRAM_OE_ * /DBB_
PCI_GNT_IN_ * /DBB_
MASK_7A * /ABB_ * DEL_ABB_* /DBB_
MASK_7B * /(/ABB_ * DEL_ABB_) * /DBB_

;ERR
;ERR
;ERR
;ERR
;ERR

13
7B
11
7A
7A

;*******************************************************************
;ERRATA 7 (fix for Async caches)
;MASK_7B feeds wires to CPUPAL MASK_BG3 input to be
; included in BG equation shown below:
;
;BG_OUT_ = BG_IN_
;
+ MASK_BG2
;for Errata #5
;
+ MASK_BG1
;for Errata #5
;
+ ABB_ * /DEL_ABB_
;for Errata #7
;
+ MASK_BG3 * /DBB_
;for Errata #7 (Asych cache fix)
;******************************************************************
MASK_7A = PCI_GNT_IN_
+ MASK_7A * /(/ABB_ * DEL_ABB_)
MASK_7B = MASK_7A * /ABB_ * DEL_ABB_
+ MASK_7B * /(/ABB_ * DEL_ABB_)
;******************************************************************
PCI_GNT_OUT_ = PCI_GNT_IN_
+ MASK_PCI_GNT
;––––––––––––––––––––––––– END OF FILE –––––––––––––––––––––––––––

176

MPRH01TSU-02

Errata

13.4 660 Bridge Revision 1.1 Errata, 8/9/95 Release
This section contains a subset of the 8/9/95 release of the 660 bridge revision 1.1 errata
summary. The errata summary contains generic example PALs which are not identical to
those used on the reference design. The function of the two sets of PALs is identical; only
the parsing of the logic among the devices is different.
Section 13.4.1 describes the invididual 660 bridge errata and their associated workarounds.
Section 13.4.2 contains the logic design files of the PALs chosen to implement the workaround logic.
Section 13.4.3 describes the installation of the PLDs in a generic system.
13.4.1 Individual 660 Bridge Errata
The following sections contain a detailed breakdown of individual 660 Bridge Errata. All equations are shown
in PALASMt format: INVERT is /, OR is +, and AND is *. The term performance refers to the number of clock
cycles required to complete an operation.

Advisory 1 . . . NCR 53C810 SCSI Controller
Date . . . . . . . . . . . . 6/21/95
Functional . . . . . . On certain PCI configuration writes to the SCSI controller, the 664 will wait several
( >3 ) PCI clocks before asserting IRDY#. Under these conditions the SCSI controller
appears to latch in the write data when it asserts TRDY# even if IRDY# is not yet asserted. The manufacturer states that this problem exists on NCR part number
609–0391399, but that it has been corrected on later parts, starting with part number
609–1391635.
Workaround . . . . Use part number 609–1391635 or later.
Impact . . . . . . . . . . None.

Errata 1 . . . . . . MIO Test
Date . . . . . . . . . . . . 6/09/95
Functional . . . . . . The 664 MIO test function is incorrect.
Workaround . . . . None.
Impact . . . . . . . . . . None. This errata only affects chip-level testing.

Errata 2 . . . . . . PCI Busmaster Timing
Date . . . . . . . . . . . . 6/19/95
Functional . . . . . . When a PCI busmaster accesses system memory, the 664 CPU bus arbiter samples
FRAME#, AD[31:30], and C/BE[3:0] on CPU clock edges rather than on PCI clock
edges. If these signals are switching at the time of sampling, then the PCI bus may hang
(the 664 may not assert TRDY#).
Workaround . . . .
Option 1: Set the CPU bus frequency below 50Mhz to ensure that the signals are not sampled while
switching.

MPRH01TSU-02

177

Errata
Option 2: Separate the FRAME# signal to the 664 from the rest of the PCI bus by a PAL. When
a PCI busmaster runs a cycle, hold FRAME# (as an input to the 664) deasserted until
the second (low) half of the PCI clock, so that the 664 samples it active on the rising edge
of PCI_CLK.
Impact . . . . . . . . . . Option 1 affects maximum operating frequency. Option 2 has no effect on performance,
but may affect the maximum operating frequency of the PCI bus.

Errata 3 . . . . . . TA# Not Tristated for CPU Bus Targets
Date . . . . . . . . . . . . 6/09/95
Functional . . . . . . The 664 fails to tristate TA# when a cycle is claimed (using CPU_BUS_CLAIM#) by a
CPU bus target and the CPU bus is not pipelined and the L2 is enabled.
Workaround . . . . Disable the L2 when using a CPU bus target. Other workarounds TBD.

Errata 4 . . . . . . ECIWX, ECOWX and TBST#
Date . . . . . . . . . . . . 6/09/95
Functional . . . . . . If an ECIWX or ECOWX bus operation is run with TBST# active, then the 664 incorrectly
treats the cycle as a burst by providing four TA# assertions. However, when an ECIWX
or ECOWX bus operation is run, the TBST# and TSIZE do not actually define the tranfer
size, the size is always 4 bytes.
Workaround . . . . Do not execute ECIWX or ECOWX instructions.
Impact . . . . . . . . . . This workaround has no effect on performance or maximum frequency.

Errata 5 . . . . . . PCI Busmaster Disconnect on Refresh
Date . . . . . . . . . . . . 6/19/95
Functional . . . . . . The 664 asserts spurious TA#s when a PCI busmaster read from memory is target–disconnected due to a memory refresh request while a CPU to PCI address tenure is outstanding.
Workaround . . . . Separate the BG1# (and BG2# if MP) signal between the 664 and the CPU by a PAL.
When a PCI busmaster access to memory is target disconnected, deassert BG1# (and
BG2#) for six PCI CLKs.
Impact . . . . . . . . . . This workaround has negligible impact on performance but may affect the maximum operating frequency of the CPU bus.

Errata 6 . . . . . . PCI Busmaster Hang
Date . . . . . . . . . . . . 6/19/95
Functional . . . . . . PCI to memory reads hang when the PCI busmaster initiates the transaction while a CPU
to memory data tenure is outstanding and a CPU to PCI address tenure is outstanding.
Workaround . . . . Drive CPU_ADDR[0] low on the CPU bus clock following the assertion of AACK#.
Impact . . . . . . . . . . This workaround has no effect on performance or maximum frequency.

178

MPRH01TSU-02

Errata
Errata 7 . . . . . . Bus Hang During L2 Read Hit + CPU to Memory + PCI to Memory
Date . . . . . . . . . . . . 7/11/95
Functional . . . . . . If a PCI busmaster begins a system memory access while a CPU to memory address
tenure is pending and an L2 read hit data tenure is pending, then the CPU and PCI
busses will hang (TA# and TRDY# are never asserted).
Workaround . . . . Separate the BG1# (and BG2# if MP) signal between the 664 and the CPU by a PAL.
Deassert BG1# (and BG2#) on the clock following the assertion of AACK#. If asynchronous SRAM is used in the L2, then also deassert BG1# (and BG2# if MP) while DBB#
is active, from the time PCI_GNT# goes inactive until ABB# has asserted twice.
Impact . . . . . . . . . . This workaround has negligible impact on performance but may affect the maximum operating frequency of the CPU bus.

Errata 8 . . . . . . Back-To-Back PCI Busmaster Write
Date . . . . . . . . . . . . 6/22/95
Functional . . . . . . The CPU and PCI busses hang if a PCI busmaster write to memory begins while the CPU
is asserting BR# and the CPU address bus is becoming free from the broadcast snoop
caused by a previous PCI busmaster write to memory.
Workaround . . . . Separate the BR1# (and BR2# if MP) signal from the 664 to the CPU by a PAL. Deassert
BR1# (and BR2#) while PCI_CLK is high and a PCI busmaster is accessing memory.
Impact . . . . . . . . . . This workaround has negligible impact on performance but may affect the maximum operating frequency of the CPU bus.

Errata 9 . . . . . . 603 Snoop on PCI Busmaster Write
Date . . . . . . . . . . . . 6/27/95
Functional . . . . . . The snoop operation that the 664 broadcasts on the CPU address bus when a PCI busmaster begins a write to memory is Write–with–Kill (00110b). For a 603, the snoop is supposed to be Write–with–Flush (00010b).
Workaround . . . . In 603/603e designs, cut the TT[2] signal between the 664 and the CPU bus. Pull both
both signals (664_TT[2] and CPU_TT[2]) down with 500 ohm resistors. This workaround
is not required in 604 designs.
Impact . . . . . . . . . . This workaround has no effect on performance or maximum operating frequency.

Errata 10 . . . . CPU to PCI Bus Clocks at 3:1
Date . . . . . . . . . . . . 7/21/95
Functional . . . . . . If the CPU:PCI clocks are running at a frequency ratio of 3:1, and a CPU to memory data
tenure is pending, then PCI busmaster to memory accesses hang.
Workaround . . . . Do not operate the CPU:PCI bus clocks at 3:1.
Impact . . . . . . . . . . This workaround has no effect on performance or maximum frequency.

MPRH01TSU-02

179

Errata
Errata 11 . . . . . PCI Bus Hang On Pipelined CPU to PCI Master Abort
Date . . . . . . . . . . . . 8/08/95
Functional . . . . . . Normally, when a CPU initiates a transfer to the PCI bus, and the 660 bridge initiates the
PCI transaction before the CPU data tenure begins (the CPU bus is pipelined), the
FRAME# is held asserted (more than the usual 2 PCI_CLKs) until the CPU data tenure
begins. However, if the 660 bridge detects a PCI master abort condition (due to no DEVSEL#) before the CPU data tenure begins, then the 660 bridge will hang the PCI bus.
The CPU bus transfer will complete normally.
Note that if the PCI target watchdog timer is enabled then the busses will not hang, since
the hung PCI data phase will cause the PCI cycle to terminate after 2000 PCI clocks.
However, if the CPU attempts to again access the PCI bus before this point, then the CPU
cycle will be lost (terminated on the CPU bus without executing on the PCI bus).
Workaround . . . . Separate the PCI_GNT# signal from the PCI arbiter to the 664 by a PAL. Deassert
PCI_GNT# (to the 664) from the time the CPU starts a cycle to the PCI until the CPU is
no longer pipelined (DBB# goes inactive).
Impact . . . . . . . . . . This workaround has negligible impact on performance but may affect the maximum operating frequency of the CPU bus.

Errata 12 . . . . 663 Memory Read
Date . . . . . . . . . . . . 7/24/95
Functional . . . . . . Under certain conditions (CPU_RDL_OPEN goes low before MEM_RD_SMPL goes
low) during a CPU to memory read transfer, the 663 may supply corrupted data to the
CPU bus on CPU_DATA[1, 3, 5, ...].
Workaround . . . . Add a 100 ohm series resister to the CPU_RDL_OPEN signal to slow its falling edge.
This workaround is still under analysis.
Impact . . . . . . . . . . This workaround has no effect on performance but may affect the maximum operating
frequency of the CPU bus.

Errata 13 . . . . Extraneous CPU to PCI Cycle
Date . . . . . . . . . . . . 8/8/95
Functional . . . . . . Normally, if a CPU to PCI memory write is pending, and a PCI busmaster initiates a system memory access, then the CPU cycle is retried. Under certain conditions however,
the CPU to PCI transaction is also posted inside the 660, causing the CPU to PCI transaction to be executed once by the 660 and once by the CPU. Additionally, the 660
executes the posted transaction with random levels on the C/BE# lines.
Workaround . . . . Separate the BG1# (and BG2# if MP) signal between the 664 and the CPU by a PAL.
Deassert BG1# (and BG2#) from the beginning of each CPU bus address tenure until
4 CPU_CLKs before the associated data tenure ends. This can be done by waiting until
the 3rd TA# is received for memory cycles, or until one clock after AACK# on L2 cache
hits.
Impact . . . . . . . . . . This workaround has a less than 2% impact on performance, and may affect the maximum operating frequency of the CPU bus.

180

MPRH01TSU-02

Errata
Errata 14 . . . . TT[0:4] Float to ECIWX/ECOWX
Date . . . . . . . . . . . . 8/8/95
Functional . . . . . . If the CPU TT[0:4] lines float to the value 1x10xb while a CPU data tenure is pending,
then the 664 will malfunction, usually by exhibiting incorrect behavior on the TA# line.
Details . . . . . . . . . During a CPU bus transfer, TT[0:4] are driven to a correct value, and then allowed to float
for up to 16 CPU clocks in certain situations. The 664 incorrectly samples their value during the above conditions. As long as there are no current sources operating on the TT
lines during the float, these lines will maintain their value due to bus capacitance.
Workaround . . . . Do not put put pullup or pulldown resistors, TTL–type inputs, or any other current sources
on TT[0, 2, 3]. For 603 or 603e designs only, workaround 9 overrides this workaround;
pull down TT[2]_664 and TT[2]_CPU as shown in errata 9.
Impact . . . . . . . . . . This workaround has no effect on performance or maximum frequency.

Errata 15 . . . . Spurious Set of the Flash Write Lockout Bit
Date . . . . . . . . . . . . 8/1/95
Functional . . . . . . While the 660 is configured for direct-connect ROM (PCI-based), and a ROM read cycle
is pipelined over a single-beat CPU to memory write to an odd memory location, then
the flash write lockout bit (located at register address FFFF FFF1h) may be inadvertantly
set (without a write to the aforementioned address). Once written, this bit can only be
reset via power on reset. If the bit sets the PCI-based flash ROM cannot be updated.
Workaround . . . . A hardware workaround is TBD.
A possible software workaround is to prevent the above scenario, possibly by adding
sync cycles before the ROM reads.
Impact . . . . . . . . . . This workaround has no effect on performance or maximum frequency.

Errata 16 . . . . PCI Lock in 603 1:1 Mode
Date . . . . . . . . . . . . 8/1/95
Functional . . . . . . While the 660 bridge is configured for 603 1:1 mode (which ensures a wait state between
TS# and AACK#) the assertion of PCI_LOCK# may cause a system hang.
Details . . . . . . . . . This bug has only been predicted by system simulation, and has not been observed in
the lab.
Workaround . . . . Do not use PCI_LOCK# while in 603 1:1 mode.
Impact . . . . . . . . . . This workaround has no effect on performance or maximum frequency.

Errata 17 . . . . 663 Parity Error Detection
Date . . . . . . . . . . . . 8/4/95
Functional . . . . . . Parity errors on the CPU and memory data busses are only reported by the 663 if there
are an odd number of bytes containing parity errors. All single single bit errors (in the
number of bytes transferred that beat) are reported, but double bit arrors may not be reported, even if the two errors are in different bytes.
Workaround . . . . TBD.
Impact . . . . . . . . . . This errata has no effect on performance or maximum frequency.

MPRH01TSU-02

181

Errata
13.4.2 PAL Design Files
The following PAL equations describe the 660 team’s implementation of the example logic equations contained in the workarounds for the errata. These are PALASM design files. Note that the PAL pinouts may not
match the reference design PAL pinouts.

13.4.2.1

HPP1 Design File

TITLE
HPP.pds
PATTERN none
REVISION 1.0
COMPANY IBM
DATE
06/29/95
CHIP _FPP PALCE16V8
;–––––––––––––––––––––––––––– PIN Declarations ––––––––––––––––––––
;
; Predefined
;
;PIN 1
PCI_CLK
; CLOCK
;PIN 10
GND
; GROUND
;PIN 11
REG_OE#
; OUTPUT ENABLE FOR REGISTERED OUTPUTS
;PIN 20
VCC
; VCC
; Inputs
;
PIN 2
PIN 3
PIN 4
PIN 5
PIN 6
;PIN 7
;PIN 8
;PIN 9

IRDY_
TRDY_
PCI_CLK
STOP_
PCI_GNT_

; outputs
;
PIN 12
PCI_FRAME_
bus
PIN 13
664_FRAME_
PIN 19
MASK_BR
;
; Internal Registers
;
PIN 14
664_OWN_PCI
PIN 15
DEL_PCI_FRAME_
PIN 16
MASK_BG0
PIN 17
MASK_BG2
PIN 18
MASK_BG1

; INPUT from
; INPUT from
; INPUT
; INPUT from
; INPUT from
; INPUT
; INPUT
; INPUT

COMB

PCI bus
PCI bus
PCI bus
PCI bus arbiter

; CONBINATIONAL

COMB
COMB

; COMBINATIONAL
; COMBINATIONAL

REG
REG
REG
REG
REG

;
;
;
;
;

REGISTERED
REGISTERED
REGISTERED
REGISTERED
REGISTERED

182

goes to/from PCI

goes to/from Kauai
goes to H603C PAL

goes to H603C PAL
goes to H603C PAL

MPRH01TSU-02

Errata
;––––––––––––––––––––– Boolean Equation Segment –––––––––––––––––––
;
EQUATIONS
;
;
Output enables for Comb logic (Reg OE’s controllered by pin 11)
;
PCI_FRAME_.TRST
664_FRAME_.TRST
;
;
;

= 664_OWN_PCI
= /664_OWN_PCI

Combinational and registered equations

DEL_PCI_FRAME_ = PCI_FRAME_
664_OWN_PCI = /PCI_GNT_ * PCI_FRAME_ * IRDY_
+ 664_OWN_PCI * /(PCI_FRAME_ * IRDY_)
PCI_FRAME_ = 664_FRAME_

; Errata #2

664_FRAME_ = PCI_FRAME_
+ PCI_CLK * DEL_PCI_FRAME_

; Errata #2

;*************************************************************
MASK_BG2 = /PCI_FRAME_ * /IRDY_ * /TRDY_ * /STOP_
+ MASK_BG2 * MASK_BG1
+ MASK_BG2 * MASK_BG0

; Errata #5

MASK_BG1 = /PCI_FRAME_ * /IRDY_ * /TRDY_ * /STOP_
+ MASK_BG2 * /MASK_BG1 * /MASK_BG0
+ MASK_BG1 * MASK_BG0

; Errata #5

MASK_BG0 = /PCI_FRAME_ * /IRDY_ * /TRDY_ * /STOP_
+ MASK_BG2 * /MASK_BG0
+ MASK_BG1 * /MASK_BG0

; Errata #5

;*************************************************************
MASK_BR = /PCI_CLK * /PCI_FRAME_ * /664_OWN_PCI

;for Errata #8

;––––––––––––––––––––––––– END OF FILE ––––––––––––––––––––––––

MPRH01TSU-02

183

Errata
13.4.2.2
TITLE
PATTERN
REVISION
COMPANY
DATE
CHIP

C3 PAL Design File
C.pds
none
1.0
IBM
07/31/95

_C2

PALCE16V8

;–––––––––––––––––––––––––––– PIN Declarations ––––––––––––––––––––
; Predefined
;
;PIN 1
CPU_CLK
; CLOCK
;PIN 10
GND
; GROUND
;PIN 11
REG_OE#
; OUTPUT ENABLE FOR REGISTERED OUTPUTS
;PIN 20
VCC
; VCC
;
;
;

Inputs

PIN 2
PIN 3
PIN 4
PIN 5
PIN 6
PIN 7
PIN 8
PIN 9
PIN 17

BG_IN_
ABB_
BR_IN_
MASK_BG2
MASK_BG1
MASK_BR
DBB_
PCI_GNT_IN_
MASK_BG3

; outputs
;
PIN 12
CPU_A0
PIN 13
BR_OUT_
PIN 18
PCI_GNT_OUT_
PIN 19
BG_OUT_
;
; Internal Registers
;
PIN 14
MASK_PCI_GNT
PIN 15
DEL_ABB_
PIN 16
DEL_BG_

;
;
;
;
;
;
;
;
;

INPUT
INPUT
INPUT
INPUT
INPUT
INPUT
INPUT
INPUT
INPUT

COMB
COMB
COMB
COMB

;
;
;
;

COMBINATIONAL
COMBINATIONAL
COMBINATIONAL
COMBINATIONAL

REG
REG
REG

; REG
; REGISTERED
; REGISTERED

184

MPRH01TSU-02

Errata
;–––––––––––––––––––––– Boolean Equation Segment –––––––––––––––––––
;
EQUATIONS
;
; Output enables for Comb logic (Reg OE’s controllered by pin 11)
;
CPU_A0.TRST = ABB_ * /DEL_ABB_
;
;
;

;for Errata #6

Equations for internal registers

DEL_BG_

= BG_IN_

DEL_ABB_ = ABB_
MASK_PCI_GNT = /ABB_ * DEL_ABB_ * CPU_A0 * /DBB_ ;for Errata #11
+ MASK_PCI_GNT * /DBB_
; Delay granting PCI to
; the CPU when a CPU to
PCI_GNT_OUT_ = PCI_GNT_IN_
; PCI cycle is pipelined
+ MASK_PCI_GNT
; over the data tenure of
; the previous cycle
BG_OUT_

=
+
+
+

BG_IN_
MASK_BG1
MASK_BG2
MASK_BG3

CPU_A0 = GND

;for Errata #5
;for Errata #5
;for Errata #7
;for Errata #6

BR_OUT_ = BR_IN_
+ MASK_BR
;for Errata #8
;––––––––––––––––––––––––– END OF FILE –––––––––––––––––––––––––––

MPRH01TSU-02

185

Errata
13.4.2.3

Broom PAL Design File

TITLE
BROOM1.pds
PATTERN none
REVISION 1.0
COMPANY IBM
DATE
08/02/95
CHIP _BROOM1 PALCE16V8
;–––––––––––––––––––––––––––– PIN Declarations ––––––––––––––––––––
;
; Predefined
;
;PIN 1
CPU_CLK
; CLOCK
;PIN 10
GND
; GROUND
;PIN 11
REG_OE#
; OUTPUT ENABLE FOR REGISTERED OUTPUTS
;PIN 20
VCC
; VCC
;
; Inputs
;
PIN 2
DBB_
; INPUT
PIN 3
PCI_GNT_IN_
; INPUT
PIN 4
ABB_
; INPUT
PIN 5
TA_
; INPUT
PIN 6
SRAM_OE_
; INPUT
;PIN 7
; INPUT
;PIN 8
; INPUT
;PIN 9
; INPUT
;
; Outputs
;
;PIN 13
COMB
; COMBINATIONAL
;PIN 14
COMB
; COMBINATIONAL
;
; Internal Registers
;
PIN 12
DEL_ABB_
REG
; REG
PIN 15
MASK_BG3
REG
; REG
PIN 16
MASK_7A
REG
; REGISTERED
PIN 17
MASK_7B
REG
; REGISTERED
PIN 18
TA_CNT0
REG
;
PIN 19
TA_CNT1
REG
;

186

MPRH01TSU-02

Errata
;––––––––––––––––––––––– Boolean Equation Segment –––––––––––––––––––
;
; Output enables for Comb logic (Reg OE’s controllered by pin 11)
;
;
Equations for internal registers
DEL_ABB_

= ABB_

;**********************************************************************
;
; Below is fix for Errata #7B, 11, 13,
;
TA_CNT0 = /TA_CNT0 * /TA_
+ TA_CNT0 * /DBB_ * TA_
TA_CNT1 = TA_CNT0 * /TA_
+ TA_CNT1 * /DBB_
MASK_BG3 =
+
+
+
+

(/TA_CNT0 + /TA_CNT1) * SRAM_OE_ * /DBB_
/ABB_ * /SRAM_OE_ * /DBB_
PCI_GNT_IN_ * /DBB_
MASK_7A * /ABB_ * DEL_ABB_* /DBB_
MASK_7B * /(/ABB_ * DEL_ABB_) * /DBB_

;ERR
;ERR
;ERR
;ERR
;ERR

13
7B
11
7A
7A

;**********************************************************************
;ERRATA 7 (fix for Async caches)
;MASK_7B feeds wires to CPUPAL MASK_BG3 input to be included in BG
equation
;
shown below:
;
;BG_OUT_ = BG_IN_
;
+ MASK_BG2
;for Errata #5
;
+ MASK_BG1
;for Errata #5
;
+ ABB_ * /DEL_ABB_
;for Errata #7
;
+ MASK_BG3 * /DBB_
;for Errata #7 (Asych cache fix)
;*********************************************************************
MASK_7A = PCI_GNT_IN_
+ MASK_7A * /(/ABB_ * DEL_ABB_)
MASK_7B = MASK_7A * /ABB_ * DEL_ABB_
+ MASK_7B * /(/ABB_ * DEL_ABB_)
;––––––––––––––––––––––––– END OF FILE –––––––––––––––––––––––––––

MPRH01TSU-02

187

Errata
13.4.3 Workaround PAL Installation
Figure 63 shows a typical installation of the workaround PALs in a generic 603e/604 uniprocessor system.
PAL connections shown without an asterisk do not require any nets to be cut. Merely connect the PAL pin to
the an existing net (eg: TA#) or in the case of the clocks, to another clock driver output.
* Cut these nets between the 664 and everything else. Then connect the PALs to each side of the cut as
shown. For example, cut the PCI_FRAME# net at the 664, connect only the 664 PCI_FRAME# pin to the
664_FRAME# pin of the HPP1 PAL, and connect all other devices on the FRAME# net to the PCI_FRAME#
pin of the HPP1 PAL.
CPU_CLK

BROOM PAL

CPU
TA#

TA#
PCI_GNT_IN#
DBB#
ABB#

ABB#
DBB#
A0

ABB#
DBB#
A0
BR#
BG#

*
*

664
SRAM_OE#

SRAM_OE#

MASK_BG3

MASK_BG3

C3 PAL
BR_OUT#

BR_IN#
BG_OUT#

BG_IN#

*
*

CPU_GNT1#

*

PCI_GNT#

*

PCI_FRAME#

CPU_REQ1#

PCI Arbiter
CPU_GRANT#

*

PCI_GNT_IN#

PCI_GNT_OUT#

MASK_BR
MASK_BG1
MASK_BG2

PCI Bus
IRDY#
TRDY#
STOP#
FRAME#
PCI_CLK

*

PCI_GNT_IN#
IRDY#
TRDY#
STOP#

MASK_BR
MASK_BG1
MASK_BG2

PCI_FRAME#

664_FRAME#

HPP1 PAL

Tie OE# (p11) low on each PAL.

Figure 63. Generic Implementation PAL Installation

188

MPRH01TSU-02

BOM

Section 14
Bill of Materials
14.1 603e/604 Reference Design Bill of Materials
14.1.1 603e Bill of Materials
Table 64. 603e Bill of Materials
No

Part Name

Part
Number

Ref Des

Qty

Manufacturer
Name

Manufac- JEturer
DEC_TYPE
Part
Number

Part Description

1

10MQ040–SOIC

87F4917

CR1, CR7

2

29F040ROM_
MECH–MECH

82G6496

X3

2

SANKEN

SFBP–54YL

D64

40V 1A SCHTK REC

1

AMP

AM29F040–
120JC

XPLCC32

512KX8 FLASH ROM

3

2N7002–SMD

31F2311

Q1

1

MOTOROLA

2N7002

SOT23

3 PIN FET TRANSISTOR

4

603–SMD

26H4196

U32

1

IBM

IBM25-PPC
603EFC-100-X

QFP240

603e 100MHZ

5
6

7406–SMD

17F7776

U3, U28

2

TI

7406

14SO

HEX INVERTER OC

74F08–SMD

61X9236

U1, U29

2

NATL

FO8

14SO

QUAD 2 INPUT AND GATE

7

74F11–SMD

17F7745

U13

1

NATL

F11

14SO

TRI 3 INPUT AND GATE

8

74F125–SMD

68X2888

U16

1

NATL

74F125

14SO

74F125 BUFFERS 3 STATE

9

74F244–SMD

6480438

U10, U11,U17–U20,
U40

7

PHILIPS

74F244

SO20W

OCTAL BUFF/DRIVER

10

74F245–SMD

55X8091

U12

1

PHILIPS

74F245

SO20W

OCTAL TRANSCEIVER

11

74F74–SMD

61X9238

U21

1

NATL

F74

14SO

DUAL”D”FF W/CLEAR &
PRESET

12

74HCT14–SMD

37F9034

U2

1

PHILIPS

74HCT14

14SO

74HCT14 SCHMT
TRIGGER INV

13

8042H–SMD

1054195

U27

1

INTEL

8042H

PLCC44

KEYBOARD
CONTROLLER

14

8G4756–SMD

08G4756

CR6

1

MOTOROLA

MMBD914L
T1

SOT23

70V 200MA SWC DIODE

15

BATTERY_CONN–
CONN

19G2441

M5

1

SONY

CR2–032Q

CONN_
19G2441

BATTERY HOLDER

16

BATTERY_MECH–
MECH

15F8409

B1

1

SONY

CR2–032Q

PART

BATTERY 3.0V

17

BERG1X2–DIP

6181127

J1,J8–J11,J35

6

BERG

69190–502

BERG1X2

1X2 100MIL HEADER
VERTICAL

18

BERG1X4–DIP

6359315

J13

1

BERG

69190–504

BERG1X4

1X4 100MIL HEADER
VERTICAL

19

BERG1X5–DIP

88G4908

J12

1

AMP

104345–3

BERG1X5

BERGSTICK 1X5

MPRH01TSU-02

189

BOM
Table 64. 603e Bill of Materials (Continued)
No

Part Name

Part
Number

Ref Des

Manufacturer
Name

Manufac- JEturer
DEC_TYPE
Part
Number

Part Description

20

CAPACITOR—
0.001UF, 20%

98F1292

C23– C26,
C96,C97,
C129–C148,
C202–205

30

KYOCERA

08055C102
KAT2A

SMC0805

CAPACITOR

21

CAPACITOR—
0.01UF, 20%

41F0313

C3–C6,C8,C9,C12–
C15,C18–C21,C28–
C31,C33–C37,C39,
C40,C42,C46–C55,
C57–C65,C67–C75,
C77–C82,C84,C86,
C93,C95,
C117–C120,C157–
C164,C167–C169,
C172–C174,C208–
C215,C217–C220,
C227–C231,C233,
C235–C243,C245–
C248,C251–C279

142

KYOCERA

0805X103M
2B05

SMC0805

CAPACITOR

22

CAPACITOR—
0.1UF,20%

41F0316

C43,C91,C92,
C94,C222

5

KYOCERA

0805Y104Z
1B05

SMC0805

CAPACITOR

23

CAPACITOR—
10UF,20%

71F7911

C221

1

NEC

TESVEC1C106
M12R

71F7911

CAPACITOR

24

CAPACITOR—
2200PF, 20%

42G3220

C32,C83,C85,C123,
C125

5

MURATA

GRM40XTR
222J050AD

SMC0805

CAPACITOR

25

CAPACITOR—
68PF, 20%

62G4724

C87–C90

4

KYOCERA

08055A680
KAT2

SMC0805

CAPACITOR

26

CAPACITORDO_NOT_ POP,20%

–NONE–

C10,C16,C66,C98,
C99,C111,C149,150,
152–154, C156,
C216, C226

14

NONE

NONE

SMC0805

CAPACITOR

27

CHANDRA–SMD

–NONE–

M1

1

ALTERA

100 QFP
SOCKET

100QFP

GTP W/128 MACROCELLS

28

CONNDIN5DO_NOT_POP

–NONE–

J14A

1

NONE

NONE

CONN_5DIN

AT KEYBOARD
CONNECTOR

29

CONNMDIN6
-CONNMDIN6

15F6890

J14,J15

2

AMP

749180–1

CONN_
15F6890

6 POS CIRCULAR MINI
DIN

30

CONNPOWER-DIP

55X8085

J4

1

MOLEX

15–48–0212

CONN_
55X8085

1X12 0 156 CL FRET LCK
HDR

31

CONNPOWER2–DIP

–NONE–

J5

1

BURNDY

GTC6R–1

CONN_1X6

1X6 0 156 CL FRET LCK
HDR

32

CONN_DIMMSRAM160–DIP

—

J3

1

BURNDY

CELP2X80S
C3Z48

CONN_
DIMMSRA

160 PIN CELP CONN

33

CONN_ISA–DIP

6137473

J29–J33

5

AMP

645169–4

ISA2X49

2X49 ISA CONNECTOR

34

CONN_PCI–DIP

72G0316

J25–J27

3

AMP

646255–1

PCI2X605V

2X60 32 BIT PCI
CONNECTOR

35

CONN_SIMM72DIP

–NONE–

J21–J24

4

AMP

822032–4

SIMM72

72 PIN SIMM CONNECTOR

36

CRYSTL32.768KHZ

03G9527

Y3

1

EPSON

MC405–
32.768KHZ–
6PF

XTALSMT1

CRYSTAL

37

CRYSTL–
DO_NOT_POP

DO_NOT_
POP

Y2

1

NONE

NONE

XTALSMT2

CRYSTAL

38

DO–NOT–POP_14–14

–NONE–

S6

1

NONE

NONE

SO14

DO–NOT–POP

39

DO–NOT–POP_16–16

–NONE–

S4

1

NONE

NONE

SO16

DO–NOT–POP

40

DO–NOT–
POP_24W–24W

–NONE–

S3

1

NONE

NONE

SO24W

DO–NOT–POP

41

DO–NOT–POP_
PLCC20–PLCC20

–NONE–

S1,S5

2

NONE

NONE

PLCC20

DO–NOT–POP

Qty

190

MPRH01TSU-02

BOM
Table 64. 603e Bill of Materials (Continued)
No

Manufacturer
Name

Manufac- JEturer
DEC_TYPE
Part
Number

Part Description

1

NONE

NONE

PLCC28

DO–NOT–POP

U26

1

DALLAS

DS1385S

SO28W

RTC+ 4KX8SRAM

C109,110

2

SPRAGUE

293D107X9
6 R3D2T

SMC2816

CAPACITOR

57G9281

C1,C2,C7,C11,C17,
C22,C27,C38,C56,
C112–C116,C121,
C122,C165,C166,
C170,C171,C175C183,C187-C201,
C207,C232,C234,
C244,C249

49

KEMET

T491D336M
O16AS

CAP33UF3SMT

CAPACITOR

EPM5130FP–SMD

–NONE–

X2

1

ALTERA

EPM5130–
QC–100–1

XQFP100

GTP W/128 MACROCELLS

47

ESP_CONN–CONN

42F6867

J2

1

BERG

79282–516

CONN_
42F6867

2X8 PIN HEADER CONN

48

FCT162373–
FCT162373

–NONE –

U14

1

IDT

74FC16237
3CTPV

SSOP48

16 BIT TRANS LATCH

49

F_BEAD–BEAD

26F4865

FB1,FB2,FB4–FB6

5

TDK

HF50ACB–3
21611T

SMC1206

FERRITE BEAD

50

IBM82663–SMD

94G0235

U5

1

IBM

IBM27–
82663

QFP240_5MM

DATA MAPPING & BUFFER

51

IBM82664–SMD

94G0176

U4

1

IBM

IBM27–
82664

QFP208_5MM

SYSTEMS & MEMORY
CONTROLLER

52

IDT71216–
DO_NOT_POP

–NONE–

U38

1

IDT

IDT71216 S
10 PF

QFP80_
65MM–2M

TAG RAM(IDT71216

53

IDT71216–
IDT71216

05H1054

U37

1

IDT

IDT71216 S
10 PF

QFP80_
65MM–2M

TAG RAM(IDT71216

54

IRFZ44–SMD

03G9500

Q6,Q7(DO NOT
POP)

1

IR

IRFZ44

TO220X

TRANSISTOR 50V 35A

55

LT1431REG

31F2428

U22

1

LINEARTECH

LT1431CS8

SO8

VOLT REGULATOR

56

LT1431REGDO_NOT_POP

31F2428

U23

1

LINEARTECH

LT1431CS8

SO8

VOLT REGULATOR

57

MC88LV970–DIP

05H1509

U6

1

MOTOROLA

MC88LV970
FA

QFP52_65MM

CLOCK CHIP

58

MTGHOLE–
DO_NOT_POP

?

MH1,MH3–MH5,
MH7,MH8

?

?

157TOOLS

?

59

OSCLR–
14.3181MHZ

87F5263

Y1

1

EPSON

SG–615P–
14.31818M
C

OSCSMT4

OSCLR

60

OSCLR–24.0MHZ

87F5265

Y5

1

EPSON

SG–615P–
24.0000MC

OSCSMT4

OSCLR

61

PALCE16V8_SKT–
SKT

19G5840

M6,M7

2

AMP

822014–3

PLCC20

20 PIN PLCC SOCKET

62

PLCC32SKT–SMD

10G7624

M2

1

AMP

821977–1

PLCC32SKT

32 PIN PLCC SOCKET

63

POLYSWITCH–SMD

34G3113

F1

1

RAYCHEM

SMD100

POLYSWITCH

TAPE RESISTOR

64

POWER1X3–DIP

65G3724

J6

1

MOLEX

705–43–003
7

POWER1X3

AUX POWER
CONNECTOR

65

REGHSINK–DIP

NONE

M4

1

AAVID

533402B
21552

REGHSINK

HEATSINK VOLTREG

Part Name

Part
Number

Ref Des

42

DO–NOT–POP_
PLCC28–PLCC28

–NONE–

S2

43

DS1385–SMD

70G6764

44

ELCAP–100UF,
10%/16V

75G8253

45

ELCAP–33UF,
20%/16V

46

MPRH01TSU-02

Qty

191

BOM
Table 64. 603e Bill of Materials (Continued)
No

Qty

Manufacturer
Name

Manufac- JEturer
DEC_TYPE
Part
Number

Part Description

R18,R21–R32,R45,
R46,R52–60,R171,
R207,R438,R439,
R444–R446,R448,
R449,R463,R465,
R474, R476,
R482–R484

40

PANASONIC

ERJ8GVJ0
R

SMC0805

RESISTOR

98F1741

R436

1

ROHM

MCR10EZH
LJW152

SMC0805

RESISTOR

RESISTOR–10,5%

58F1831

R82,R85,R87,R89,
R96,R97,R429

7

PANASONIC

ERJ6GVYJ1
00S

SMC0805

RESISTOR

69

RESISTOR–
100,5%

41F0328

R176,R238,R240,
R472

10

PANASONIC

ERJ6VJ101
S

SMC0805

RESISTOR

70

RESISTOR–
10K,5%

41F0337

R1,R2,R4,R7,R19,
R47–R50,R61,R80,
R81,R90-R93,R95,
R102,R107,R108,
R119,R138-R145,
R147,R148,R155–
R166,R174,R237,
R253-R255,R257R260,R287-R351,
R364,R368-R376,
R378-R381,R384,
R387-R392,R394R397,R399-R401,
R403-R416,R423,
R424,R430-R434,
R452,R461,R468,
R471,R473,R475,
R488-R495,R497,
R510

182

PANASONIC

ERJ6GVJ10
3S

SMC0805

RESISTOR

71

RESISTOR–
12.7K,1%

40G7066

R65

1

PANASONIC

ERJ–6VNF
1272S

SMC0805

RESISTOR

72

RESISTOR–
150,5%

98F1671

R123

1

ROHM

MCR10EZH
MJW151

SMC0805

RESISTOR

73

RESISTOR–1K,5%

41F0333

R20,R51,R62,R100,
R173,R230,R239,
R280,R386,R393,
R418,R425,R426,
R435,R451,R464,
R470,R479,R505

19

PANASONIC

ERJ6GVY
272S

SMC0805

RESISTOR

74

RESISTOR–
2.7K,5%

09G9748

R3,R261-R272

13

PANASONIC

ERJ–6GVYJ
272S

SMC0805

RESISTOR

75

RESISTOR–
200,1%

40G6920

R459

1

PANASONIC

ERJ–6VNF
2000S

SMC0805

RESISTOR

76

RESISTOR–22,5%

98F1736

R5,R6,R8,R10-R17,
R33-R44,R88,R101,
R109,R111–R113,R
175,R177,R178,R19
5–R197,R208,R511

37

PANASONIC

ERJ6GVYJ2
20S

SMC0805

RESISTOR

77

RESISTOR–
220K,5%

61F2952

R122

1

PANASONIC

ERJ6GVJ22
4S

SMC0805

RESISTOR

78

RESISTOR–
249,1%

40G7233

R437

1

PANASONIC

ERJ6VNF24
90S

SMC0805

RESISTOR

79

RESISTOR–
3.9K,5%

42G3067

R120

1

PANASONIC

ERJ6GVYJ3
92S

SMC0805

RESISTOR

80

RESISTOR–
300,5%

98F1674

R103,R106,R180,
R181,R183,R225,
R236,R241,R251,
R281-R286

15

ROHM

MCR10EZH
MJW301

SMC0805

RESISTOR

Part Name

Part
Number

Ref Des

66

RESISTOR–0,5%

98F1665

67

RESISTOR–
1.5K,5%

68

192

MPRH01TSU-02

BOM
Table 64. 603e Bill of Materials (Continued)
No

Qty

Manufacturer
Name

Manufac- JEturer
DEC_TYPE
Part
Number

Part Description

R83,R84,R86,R104,
R105,R114,R226R229,R243,R250

12

PANASONIC

ERJ6GVYJ3
30S

SMC0805

RESISTOR

09G9165

R187,R188

2

PANASONIC

ERJ-3VKF
54R9S

SMT0603

RESISTOR

RESISTOR–
4.7K,5%

41F0336

R231-R235,R273R279,R460,R462,
R469

15

PANASONIC

ERJ–6GVJ
472S

SMC0805

RESISTOR

84

RESISTOR–
4.99K,1%

40G7034

R64

1

PANASONIC

ERJ6VNF
4991S

SMC0805

RESISTOR

85

RESISTOR–
470K,5%

03G9709

R124

1

ROHM

MCR10
EZHLJW47
4

SMC0805

RESISTOR

86

RESISTOR–
5.6K,5%

98F1737

R352-R358

7

PANASONIC

ERJ6GVJ
562S

SMC0805

RESISTOR

87

RESISTOR–
510,5%

41F0331

R118,R245

2

PANASONIC

ERJ6GVYJ
511S

SMC0805

RESISTOR

88

RESISTOR–
54.9,1%

09G9165

R66-R78,R244

14

PANASONIC

ERJ–3VKF
54R9S

SMC0603

RESISTOR

89

RESISTOR–
620,5%

61F2960

R63

1

PANASONIC

ERJ6GVJ
621S

SMC0805

RESISTOR

90

RESISTOR–75,5%

61F2961

R79,R359-R361

4

PANASONIC

ERJ6GVJ
750S

SMC0805

RESISTOR

91

RESISTOR–
80.6,1%

40G6887

R9

1

PANASONIC

ERJ–6VNF
80R6S

SMC0805

RESISTOR

92

RESISTOR–
DO_NOT_ POP,5%

–NONE–

R121,R146,R149–
R152,R168–R170,
R172,R179,R185,
R186,R189,R191,
R193,R198,R382,
R383,R385,R417,
R419-R422,R427,
R428,R440,R441,
R443,R447,R466,
R467,R477,R478

35

NONE

NONE

SMC0805

RESISTOR

93

RLS4148–SOIC

87F4920

CR2

1

ROHM

RLS4148

MELF

40V 1A SCHTK REC

94

RPAC24_1–33,5%

NONE

RP1-RP3,RP7,RP9,
RP10

6

CMD

PRN 110
2433 ROJ

QSOP24

RESISTOR R–PACK

95

S82378ZB–SMD

82G6542

U7

1

INTEL

S82378ZB

QFP208_5MM

PCI TO ISA BRIDGE CONT

96

XTAL–16.50MHZ

89G3833

Y2A

1

ECLIPTEK

ECX–2900–
16.500MHZ

CLP_XTAL

16.5 MHZ CRYSTAL

97

XTAL–DO_NOT_POP

–NONE–

Y3A

1

NONE

NONE

CLP_XTAL

NONE

98

X_PART–CPUHSINK

6567199

X6

1

THERMALLOY

2330B

XLOOSE

CPU HEAT SINK

99

X_PART–H604C

8185187

X1

1

AMD

PALCE16V8
H-5

XPLCC20

PROGRAMMED PAL
PLCC20

100

X_PART–HPP

8185187

X4

1

AMD

PALCE16V8
H-5

XPLCC20

PROGRAMMED PAL
PLCC20

101

THERMOSTRATE

X6

1

POWER
DEVICES

AL-079-079

Part Name

Part
Number

Ref Des

81

RESISTOR–33,5%

41F0327

82

RESISTOR-33,603

83

MPRH01TSU-02

193

HEAT SINK INTERFACE
MATERIAL

BOM
14.1.2 604 Bill of Materials
Table 65. 604 Bill of Materials
Part Name

Part
Number

Ref Des

Qty

Manufacturer
Name

Manufac- JEturer
DEC_TYPE
Part
Number

Part Description

1

10MQ040–SOIC

87F4917

CR1, CR7

2

29F040ROM_
MECH–MECH

82G6496

X3

2

SANKEN

SFBP–54YL

D64

40V 1A SCHTK REC

1

AMD

AM29F040–
120JC

XPLCC32

512KX8 FLASH ROM

3

2N7002–SMD

31F2311

Q1

1

MOTOROLA

2N7002

SOT23

3 PIN FET TRANSISTOR

4

604FP–SMD

50H5239

U33

1

IBM

IBM25-PPC
604-F-X-133
-X

QFP304_5MM

604 132MHZ IN QFP

5
6

7406–SMD

17F7776

U3, U28

2

TI

7406

14SO

HEX INVERTER OC

74F08–SMD

61X9236

U1, U29

2

NATL

FO8

14SO

QUAD 2 INPUT AND GATE

7

74F11–SMD

17F7745

U13

1

NATL

F11

14SO

TRI 3 INPUT AND GATE

8

74F125–SMD

68X2888

U16

1

NATL

74F125

14SO

74F125 BUFFERS 3 STATE

9

74F244–SMD

6480438

U10, U11,U17–U20,
U40

7

PHILIPS

74F244

SO20W

OCTAL BUFF/DRIVER

10

74F245–SMD

55X8091

U12

1

PHILIPS

74F245

SO20W

OCTAL TRANSCEIVER

11

74F74–SMD

61X9238

U21

1

NATL

F74

14SO

DUAL”D”FF W/CLEAR &
PRESET

12

74HCT14–SMD

37F9034

U2

1

PHILIPS

74HCT14

14SO

74HCT14 SCHMT
TRIGGER INV

13

8042H–SMD

1054195

U27

1

INTEL

8042H

PLCC44

KEYBOARD
CONTROLLER

14

8G4756–SMD

08G4756

CR6

1

MOTOROLA

MMBD914
LT1

SOT23

70V 200MA SWC DIODE

15

BATTERY_CONN–
CONN

19G2441

M5

1

SONY

CR2–032Q
834

CONN_
19G2441

BATTERY HOLDER

16

BATTERY_MECH–
MECH

15F8409

B1

1

SONY

CR2032

PART

BATTERY 3.3V

17

BERG1X2–DIP

6181127

J1,J8–J11,J35

6

BERG

69190–502

BERG1X2

1X2 100MIL HEADER
VERTICAL

18

BERG1X4–DIP

6359315

J13

1

BERG

69190–504

BERG1X4

1X4 100MIL HEADER
VERTICAL

19

BERG1X5–DIP

88G4908

J12

1

AMP

104345–3

BERG1X5

BERGSTICK 1X5

20

CAPACITOR—
0.001UF, 20%

98F1292

C23–C26,C96,C97,
C129–C148,
C202–C205

30

KYOCERA

08055C102
KAT2A

SMC0805

CAPACITOR

21

CAPACITOR—
0.01UF, 20%

41F0313

C3–C6,C8,C9,C12–
C15,C18–C21,C28–
C31,C33–C37,C39,
C40,C42,C46–C55,
C57–C65,C67–C75,
C77–C82,C84,C86,
C93,
C95,C117–
C120,C157–C164,
C167–C169,C172–
C174,C208–C215,
C217–C220,C227–
C231,C233,C235–
C243,C245–C248,
C251–C279

142

KYOCERA

0805X103M
2B05

SMC0805

CAPACITOR

22

CAPACITOR—
0.1UF,20%

41F0316

C43,C91,C92,
C94,C100,C222

6

KYOCERA

0805Y104Z
1B05

SMC0805

CAPACITOR

23

CAPACITOR—
10UF,20%

71F7911

C221

1

NEC

TESVEC1C
106M12R

71F7911

CAPACITOR

24

CAPACITOR—
2200PF, 20%

42G3220

C32,C83,C85,C123,
C125

5

MURATA

GRM40XTR
222J050AD

SMC0805

CAPACITOR

No

194

MPRH01TSU-02

BOM
Table 65. 604 Bill of Materials (Continued)
No

Part Name

Part
Number

Ref Des

Manufacturer
Name

Manufac- JEturer
DEC_TYPE
Part
Number

Part Description

25

CAPACITOR—
68PF,20%

62G4724

C87–C90

4

KYOCERA

08055A680
KAT2

SMC0805

CAPACITOR

26

CAPACITORDO_NOT_ POP,20%

–NONE–

C10,C16,C66,C98,
C99,C111,C149,
C150,C152–C154,
C156,C216,C226

16

NONE

NONE

SMC0805

CAPACITOR

27

CHANDRA–SMD

–NONE–

M1

1

ALTERA

100 QFP
SOCKET

100QFP

GTP W/128 MACROCELLS

28

CONNDIN5DO_NOT_POP

–NONE–

J14A

1

NONE

NONE

CONN_5DIN

AT KEYBOARD
CONNECTOR

29

CONNMDIN6
-CONNMDIN6

15F6890

J14,J15

2

AMP

749180–1

CONN_
15F6890

6 POS CIRCULAR MINI
DIN

30

CONNPOWER-DIP

55X8085

J4

1

MOLEX

15–48–0212

CONN_
55X8085

1X12 0 156 CL FRET LCK
HDR

31

CONNPOWER2–DIP

–NONE–

J5

1

BURNDY

GTC6R–1

CONN_1X6

1X6 0 156 CL FRET LCK
HDR

32

CONN_DIMMSRAM160–DIP

—

J3

1

BURNDY

CELP2X80S
C3Z48

CONN_
DIMMSRA

160 PIN CELP CONN

33

CONN_ISA–DIP

6137473

J29–J33

5

AMP

645169–4

ISA2X49

2X49 ISA CONNECTOR

34

CONN_PCI–DIP

72G0316

J25–J27

3

AMP

646255–1

PCI2X605V

2X60 32 BIT PCI
CONNECTOR

35

CONN_SIMM72DIP

–NONE–

J21–J24

4

AMP

822032–4

SIMM72

72 PIN SIMM CONNECTOR

36

CRYSTL32.768KHZ

03G9527

Y3

1

EPSON

MC405–
32.768KHZ–
6PF

XTALSMT1

CRYSTAL

37

CRYSTL–
DO_NOT_POP

DO_NOT_
POP

Y2

1

NONE

NONE

XTALSMT2

CRYSTAL

38

DO–NOT–POP_14–14

–NONE–

S6

1

NONE

NONE

SO14

DO–NOT–POP

39

DO–NOT–POP_16–16

–NONE–

S4

1

NONE

NONE

SO16

DO–NOT–POP

40

DO–NOT–
POP_24W–24W

–NONE–

S3

1

NONE

NONE

SO24W

DO–NOT–POP

41

DO–NOT–POP_
PLCC20–PLCC20

–NONE–

S1,S5

2

NONE

NONE

PLCC20

DO–NOT–POP

42

DO–NOT–POP_
PLCC28–PLCC28

–NONE–

S2

1

NONE

NONE

PLCC28

DO–NOT–POP

43

DS1385–SMD

70G6764

U26

1

DALLAS

DS1385S

SO28W

RTC+ 4KX8SRAM

44

ELCAP–
100UF,10%/16V

75G8253

C109,110

2

SPRAGUE

293D107X
96R3D2T

SMC2816

CAPACITOR

45

ELCAP–33UF,
20%/16V

57G9281

C1,C2,C7,C11,C17,
C22,C27,C38,C56,
C112–C116,C121,
C122,C165,C166,
C170,C171,C175C183,C187-C201,
C207,C232,C234,
C244,C249

49

KEMET

T491D336M
O16AS

CAP33UF3SMT

CAPACITOR

46

EPM5130FP–SMD

–NONE–

X2

1

ALTERA

EPM5130–
QC–100–1

XQFP100

GTP W/128 MACROCELLS

47

ESP_CONN–CONN

42F6867

J2

1

BERG

79282–516

CONN_
42F6867

2X8 PIN HEADER CONN

48

FCT162373–
FCT162373

–NONE –

U14

1

IDT

74FC16237
3CTPV

SSOP48

16 BIT TRANS LATCH

49

F_BEAD–BEAD

26F4865

FB1,FB2,FB4–FB6

5

TDK

HF50ACB–
321611T

SMC1206

FERRITE BEAD

MPRH01TSU-02

Qty

195

BOM
Table 65. 604 Bill of Materials (Continued)
No

Manufacturer
Name

Manufac- JEturer
DEC_TYPE
Part
Number

Part Description

1

IBM

IBM27–
82663

QFP240_5MM

DATA MAPPING & BUFFER

U4

1

IBM

IBM27–
82664

QFP208_5MM

SYSTEMS & MEMORY
CONTROLLER

–NONE–

U38

1

IDT

IDT71216
S 10 PF

QFP80_
65MM–2M

TAG RAM(IDT71216

IDT71216–
IDT71216

05H1054

U37

1

IDT

IDT71216
S 10 PF

QFP80_
65MM–2M

TAG RAM(IDT71216

54

IRFZ44–SMD

03G9500

Q6,Q7

2

IR

IRFZ44

TO220X

TRANSISTOR 50V 35A

55

LT1431REG

31F2428

U22

1

LINEARTECH

LT1431CS8

SO8

VOLT REGULATOR

56

LT1431REGDO_NOT_POP

31F2428

U23

1

LINEARTECH

LT1431CS8

SO8

VOLT REGULATOR

57

MC88LV970–DIP

05H1509

U6

1

MOTOROLA

MC88LV
970FA

QFP52_65MM

CLOCK CHIP

58

MTGHOLE–
DO_NOT_POP

?

MH1,MH3–MH5,MH
7,MH8

6

?

?

157TOOLS

?

59

OSCLR–
14.3181MHZ

87F5263

Y1

1

EPSON

SG–615P–
14.31818
MC

OSCSMT4

OSCLR

60

OSCLR–24.0MHZ

87F5265

Y5

1

EPSON

SG–615P–
24.0000MC

OSCSMT4

OSCLR

61

PALCE16V8_SKT–
SKT

19G5840

M6,M7

2

AMP

822014–3

PLCC20

20 PIN PLCC SOCKET

62

PLCC32SKT–SMD

10G7624

M2

1

AMP

821977–1

PLCC32SKT

32 PIN PLCC SOCKET

63

POLYSWITCH–SMD

34G3113

F1

1

RAYCHEM

SMD100

POLYSWITCH

TAPE RESISTOR

64

POWER1X3–DIP

65G3724

J6

1

MOLEX

705–43–
0037

POWER1X3

AUX POWER
CONNECTOR

65

REGHSINK–DIP

NONE

M4

1

AAVID

533402
B22552

REGHSINK

HEATSINK VOLTREG

66

RESISTOR–0,5%

98F1665

R18,R21–R32,R45,
R46,R52–R60,R110,
R125,R171,R207,
R377,R438,R439,
R444–R446,R448,
R449,R463,R465,
R474,R476,R482–
R484

43

PANASONIC

ERJ8GV
J0R

SMC0805

RESISTOR

67

RESISTOR–
1.5K,5%

98F1741

R436

1

ROHM

MCR10EZH
LJW152

SMC0805

RESISTOR

68

RESISTOR–10,5%

58F1831

R82,R85,R87,R89,
R96,R97,R429

7

PANASONIC

ERJ6GVYJ1
00S

SMC0805

RESISTOR

69

RESISTOR–
100,5%

41F0328

R153,R154,R176,
R190,R192,R194,
R238,R240,R472,
R485

10

PANASONIC

ERJ6VJ
101S

SMC0805

RESISTOR

Part Name

Part
Number

Ref Des

50

IBM82663–SMD

94G0235

U5

51

IBM82664–SMD

94G0232

52

IDT71216–
DO_NOT_POP

53

Qty

196

MPRH01TSU-02

BOM
Table 65. 604 Bill of Materials (Continued)
No

Qty

Manufacturer
Name

Manufac- JEturer
DEC_TYPE
Part
Number

Part Description

181

PANASONIC

ERJ6GVJ10
3S

SMC0805

RESISTOR

R65

1

PANASONIC

ERJ–6VNF
1272S

SMC0805

RESISTOR

98F1671

R123

1

ROHM

MCR10EZH
MJW151

SMC0805

RESISTOR

RESISTOR–1K,5%

41F0333

R20,R51,R62,R100,
R173,R230,R239,
R280,R386,R393,
R418,R425,R426,
R428,R435,R451,
R464,R470,R479,
R505

20

PANASONIC

ERJ6GVJ
102S

SMC0805

RESISTOR

74

RESISTOR–
2.7K,5%

09G9748

R3,R261-R272

13

PANASONIC

ERJ–6GVYJ
272S

SMC0805

RESISTOR

75

RESISTOR–
200,1%

40G6920

R459

1

PANASONIC

ERJ–6VNF
2000S

SMC0805

RESISTOR

76

RESISTOR–22,5%

98F1736

R5,R6,R8,R10-R17,
R33-R44,R88,R101,
R109,R111–R113,
R175,R177,R178,
R195–R197,R208,
R511

38

PANASONIC

ERJ6GVYJ
220S

SMC0805

RESISTOR

77

RESISTOR–
220K,5%

61F2952

R122

1

PANASONIC

ERJ6GVJ
224S

SMC0805

RESISTOR

78

RESISTOR–
249,1%

40G7233

R437

1

PANASONIC

ERJ6VNF
2490S

SMC0805

RESISTOR

79

RESISTOR–
3.9K,5%

42G3067

R120

1

PANASONIC

ERJ6GVYJ
392S

SMC0805

RESISTOR

80

RESISTOR–
300,5%

98F1674

R103,R106,R180,
R181,R183,R225,
R236,R241,R251,
R281-R286

15

ROHM

SMC0805

RESISTOR

81

RESISTOR–33,5%

41F0327

R83,R84,R86,R104,
R105,R114,R226R229,R243,R250

12

PANASONIC

ERJ6GVYJ
330S

SMC0805

RESISTOR

82

RESISTOR-33,603

09G9165

R187,R188

2

PANASONIC

ERJ-3VKF
54R9S

SMT0603

RESISTOR

83

RESISTOR–
4.7K,5%

41F0336

R231-R235,R273R279,R460,R462,
R469

15

PANASONIC

ERJ–6GVJ
472S

SMC0805

RESISTOR

84

RESISTOR–
4.99K,1%

40G7034

R64

1

PANASONIC

ERJ6VNF
4991S

SMC0805

RESISTOR

85

RESISTOR–
470K,5%

03G9709

R124

1

ROHM

MCR10
EZH
LJW474

SMC0805

RESISTOR

Part Name

Part
Number

Ref Des

70

RESISTOR–
10K,5%

41F0337

R1,R2,R4,R7,R19,
R47–R50,R61,R80,
R81,R90-R93,R95,
R102,R107,R108,
R119,R138-R145,
R147,R148,R155R166,R174,R237,
R253-R255,R257R260,R287-R351,
R364,R368-R376,
R378-R381,R384,
R387-R392,R394R397,R399-R401,
R403-R416,R423,
R430-R434,R452,
R461,R468,R471,
R473,R475,R488R495,R497,R510

71

RESISTOR–
12.7K,1%

40G7066

72

RESISTOR–
150,5%

73

MPRH01TSU-02

197

BOM
Table 65. 604 Bill of Materials (Continued)
No

Part Name

Part
Number

Ref Des

Qty

Manufacturer
Name

Manufac- JEturer
DEC_TYPE
Part
Number

Part Description

86

RESISTOR–
5.6K,5%

98F1737

R352-R358

7

PANASONIC

ERJ6GVJ
562S

SMC0805

RESISTOR

87

RESISTOR–
510,5%

41F0331

R118,R245

2

PANASONIC

ERJ6GVYJ
511S

SMC0805

RESISTOR

88

RESISTOR–
54.9,1%

09G9165

R66-R78,R244

14

PANASONIC

ERJ–3VKF
54R9S

SMC0603

RESISTOR

89

RESISTOR–
620,5%

61F2960

R63

1

PANASONIC

ERJ6GVJ
621S

SMC0805

RESISTOR

90

RESISTOR–75,5%

61F2961

R79,R359-R361

4

PANASONIC

ERJ6GVJ
750S

SMC0805

RESISTOR

91

RESISTOR–
80.6,1%

40G6887

R9

1

PANASONIC

ERJ–6VNF
80R6S

SMC0805

RESISTOR

92

RESISTOR–
DO_NOT_ POP,5%

–NONE–

R116,R117,R121,
R146,R149-R152,
R168–R170,R172,
R179,R185,R186,
R189,R191,R193,
R198,R382,R383,
R385,R417,R419R422,R424,R427,
R440,R441,R443,
R447,R466,R467,
R477,R478

34

NONE

NONE

SMC0805

RESISTOR

93

RLS4148–SOIC

87F4920

CR2

1

ROHM

RLS4148

MELF

40V 1A SCHTK REC

94

RPAC24_1–33,5%

NONE

RP1-RP3,RP7,RP9,
RP10

6

CMD

PRN 110
2433 ROJ

QSOP24

RESISTOR R–PACK

95

S82378ZB–SMD

82G6542

U7

1

INTEL

S82378ZB

QFP208_5MM

PCI TO ISA BRIDGE CONT

96

XTAL–16.50MHZ

89G3833

Y2A

1

ECLIPTEK

ECX–2900–
16.500MHZ

CLP_XTAL

16.5 MHZ CRYSTAL

97

XTAL–DO_NOT_POP

–NONE–

Y3A

1

NONE

NONE

CLP_XTAL

NONE

98

X_PART–CPUHSINK

11H6666

X6

1

SANYO

REMP5412
H2026

XLOOSE

FAN

99

X_PART–H604C

8185187

X1

1

AMD

PALCE16V8
H-5

XPLCC20

PROGRAMMED PAL
PLCC20

100

X_PART–HPP

8185187

X4

1

AMD

PALCE16V8
H-5

XPLCC20

PROGRAMMED PAL
PLCC20

101

THERMOSTRATE

X6

1

POWER
DEVICES

AL-079-079

198

HEAT SINK INTERFACE
MATERIAL

MPRH01TSU-02

Schematics

Section 15
Schematics
This section contains the schematics of the reference design and a component placement
drawing of the reference board.
The schematics are numbered separately from the rest of the reference design technical
specification.

MPRH01TSU-02

199

Schematics

POWER GOOD LED

SIO
INTEL 82378ZB

SRAM

15.1 Reference Board Component Placement

200

MPRH01TSU-02



Source Exif Data:
File Type                       : PDF
File Type Extension             : pdf
MIME Type                       : application/pdf
PDF Version                     : 1.4
Linearized                      : No
Page Count                      : 200
XMP Toolkit                     : XMP toolkit 2.9.1-13, framework 1.6
About                           : e234e986-f00e-11ed-0000-adecf58f94ce
Producer                        : GPL Ghostscript 8.57
Modify Date                     : 2013:05:08 15:17:02Z
Create Date                     : 2013:05:08 15:17:02Z
Creator Tool                    : \(Interleaf, Inc.\)
Document ID                     : e234e986-f00e-11ed-0000-adecf58f94ce
Format                          : application/pdf
Title                           : \(Document:\)
Creator                         : \(elson\)
Author                          : (elson)
EXIF Metadata provided by EXIF.tools

Navigation menu