09 Spru302b DMA TMS320C54XX

TMS320C54x DSP
Reference Set
Volume 5: Enhanced Peripherals
Literature Number: SPRU302B

March 2007
Printed on Recycled Paper

IMPORTANT NOTICE
Texas Instruments Incorporated and its subsidiaries (TI) reserve the right to make corrections,
modifications, enhancements, improvements, and other changes to its products and services at any
time and to discontinue any product or service without notice. Customers should obtain the latest
relevant information before placing orders and should verify that such information is current and
complete. All products are sold subject to TI’s terms and conditions of sale supplied at the time of order
acknowledgment.
TI warrants performance of its hardware products to the specifications applicable at the time of sale
in accordance with TI’s standard warranty. Testing and other quality control techniques are used to the
extent TI deems necessary to support this warranty. Except where mandated by government
requirements, testing of all parameters of each product is not necessarily performed.
TI assumes no liability for applications assistance or customer product design. Customers are
responsible for their products and applications using TI components. To minimize the risks associated
with customer products and applications, customers should provide adequate design and operating
safeguards.
TI does not warrant or represent that any license, either express or implied, is granted under any TI
patent right, copyright, mask work right, or other TI intellectual property right relating to any
combination, machine, or process in which TI products or services are used. Information published by
TI regarding third-party products or services does not constitute a license from TI to use such products
or services or a warranty or endorsement thereof. Use of such information may require a license from
a third party under the patents or other intellectual property of the third party, or a license from TI under
the patents or other intellectual property of TI.
Reproduction of information in TI data books or data sheets is permissible only if reproduction is without
alteration and is accompanied by all associated warranties, conditions, limitations, and notices.
Reproduction of this information with alteration is an unfair and deceptive business practice. TI is not
responsible or liable for such altered documentation.
Resale of TI products or services with statements different from or beyond the parameters stated by
TI for that product or service voids all express and any implied warranties for the associated TI product
or service and is an unfair and deceptive business practice. TI is not responsible or liable for any such
statements.
Following are URLs where you can obtain information on other Texas Instruments products and
application solutions:
Products Applications
Amplifiers amplifier.ti.com Audio www.ti.com/audio
Data Converters dataconverter.ti.com Automotive www.ti.com/automotive
DSP dsp.ti.com Broadband www.ti.com/broadband
Interface interface.ti.com Digital Control www.ti.com/digitalcontrol
Logic logic.ti.com Military www.ti.com/military
Power Mgmt power.ti.com Optical Networking www.ti.com/opticalnetwork
Microcontrollers microcontroller.ti.com Security www.ti.com/security
Low Power Wire- www.ti.com/lpw Telephony www.ti.com/telephony
less
Video & Imaging www.ti.com/video
Wireless www.ti.com/wireless
Mailing Address: Texas Instruments

Post Office Box 655303 Dallas, Texas 75265
Copyright  2007, Texas Instruments Incorporated

Preface
Read This First
About This Manual
The TMS320C54x is a fixed-point digital signal processor (DSP) in the Texas

Instruments (TI) TMS320 family. These devices are characterized by a low-
power, enhanced architecture core. This book, the fifth volume of a 5-volume
set, provides information about the enhanced peripherals available on some
of these devices.
Many device references are shown with an apostrophe ( ’ ) replacing the usual
alphanumeric prefix (ex. ’5420 instead of TMS320VC5420). Unless otherwise
specified, all references to the ’54x in this book apply to the TMS320VC54x.
This user’s guide describes the enhanced peripherals available on the ’5402,
’5410, and ’5420 devices, and explains their operations.
The main topics discussed are:

- Host port interface (HPI)
- Multi-channel buffered serial ports (McBSPs)
- Programmable clock generator with a multiple phase-locked loop (PLL)
- DMA controller (DMA)
- Enhanced external input/output interface (EnhXIO)
- Interprocessor FIFO
For all non-enhanced peripheral information related to ’54x devices, see

TMS320C54x DSP, CPU and Peripherals, Volume 1, referenced in the section
titled Related Documents from Texas Instruments.
iii
Notational
Notational Conventions
Conventions // Information
InformationAbout
AboutCautions
Cautions
Notational Conventions
This book uses the following conventions.
- The TMS320C54x DSP can use either of two forms of the instruction set:
a mnemonic form or an algebraic form. This book uses the mnemonic form
of the instruction set. For information about the mnemonic form of the in-
struction set, see TMS320C54x DSP Reference Set, Volume 2: Mnemon-
ic Instruction Set. For information about the algebraic form of the instruc-
tion set, see TMS320C54x DSP Reference Set, Volume 3: Algebraic
Instruction Set.
- Program listings and program examples are shown in a special type-

face.
Here is a segment of a program listing:
STL A,*AR1+ ;Int_RAM(I)=0
RSBX INTM ;Globally enable interrupts
B MAIN_PG ;Return to foreground program
- Square brackets, [ and ], identify an optional parameter. If you use an op-

tional parameter, specify the information within the brackets; do not type
the brackets themselves.
Information About Cautions

This book contains cautions.
This is an example of a caution statement.

A caution statement describes a situation that could potentially
damage your software or equipment.
The information in a caution is provided for your protection. Please read each
caution carefully.
iv
Related Documentation From Texas Instruments
Related Documentation from Texas Instruments
The following books provide related documentation for the TM320C54x. To

obtain a copy of any of these TI documents, call the Texas Instruments Litera-
ture Response Center at (800) 477-8924. When ordering, please identify the
book by its title and literature number.
Many of these documents are located on the internet at http://www.ti.com.
TMS320C54x DSP Reference Set, Volume 1: CPU and Peripherals

(literature number SPRU131) describes the TMS320C54x 16-bit
fixed-point general-purpose digital signal processors. Covered are
its architecture, internal register structure, data and program
addressing, the instruction pipeline, and on-chip peripherals. Also
includes development support information, parts lists, and design
considerations for using the XDS510 emulator.
TMS320C54x DSP Reference Set, Volume 2: Mnemonic Instruction

Set (literature number SPRU172) describes the TMS320C54x
digital signal processor mnemonic instructions individually. Also
includes a summary of instruction set classes and cycles.
TMS320C54x DSP Reference Set, Volume 3: Algebraic Instruction

Set (literature number SPRU179) describes the TMS320C54x
digital signal processor algebraic instructions individually. Also
includes a summary of instruction set classes and cycles.
TMS320C54x DSP Reference Set, Volume 4: Applications Guide

(literature number SPRU173) describes software and hardware
applications for the TMS320C54x digital signal processor. Also
includes development support information, parts lists, and design
considerations for using the XDS510 emulator.
TMS320C54x DSP Reference Set, Volume 5: Enhanced Peripherals

(literature number SPRU302) describes the enhanced peripherals
available on the TMS320C54x digital signal processors. Includes
the multichannel buffered serial ports (McBSPs), direct memory
access (DMA) controller, the HPI-8 and HPI-16 host port inter-
faces, and the interprocessor.
TMS320C54x, TMS320LC54x, TMS320VC54x Fixed-Point Digital Signal

Processors (literature number SPRS039) data sheet contains the
electrical and timing specifications for these devices, as well as signal
descriptions and pinouts for all of the available packages.
Read This First v

Related Documents From Texas Instruments
TMS320C54x DSKplus User’s Guide (literature number SPRU191)

describes the TMS320C54x digital signal processor starter kit (DSK),
which allows you to execute custom ’C54x code in real time and debug it
line by line. Covered are installation procedures, a description of the
debugger and the assembler, customized applications, and initialization
routines.
TMS320C54x Assembly Language Tools User’s Guide (literature number
SPRU102) describes the assembly language tools (assembler, linker,
and other tools used to develop assembly language code), assembler di-
rectives, macros, common object file format, and symbolic debugging di-
rectives for the ’C54x generation of devices.
TMS320C5xx C Source Debugger User’s Guide (literature number
SPRU099) tells you how to invoke the ’C54x emulator, evaluation
module, and simulator versions of the C source debugger interface. This
book discusses various aspects of the debugger interface, including
window management, command entry, code execution, data
management, and breakpoints. It also includes a tutorial that introduces
basic debugger functionality.
TMS320C54x Code Generation Tools Getting Started Guide (literature
number SPRU147) describes how to install the TMS320C54x assembly
language tools and the C compiler for the ’C54x devices. The installation
for MS-DOS, OS/2, SunOS, Solaris, and HP-UX 9.0x systems
is covered.
TMS320C54x Evaluation Module Technical Reference (literature number
SPRU135) describes the ’C54x evaluation module, its features, design
details and external interfaces.
TMS320C54x Optimizing C Compiler User’s Guide (literature number
SPRU103) describes the ’C54x C compiler. This C compiler accepts
ANSI standard C source code and produces TMS320 assembly lan-
guage source code for the ’C54x generation of devices.
TMS320C54x Simulator Getting Started (literature number SPRU137) de-
scribes how to install the TMS320C54x simulator and the C source
debugger for the ’C54x. The installation for MS-DOS, PC-DOS,
SunOS, Solaris, and HP-UX systems is covered.
TMS320 Third-Party Support Reference Guide (literature number
SPRU052) alphabetically lists over 100 third parties that provide various
products that serve the family of TMS320 digital signal processors. A
myriad of products and applications are offered—software and hardware
development tools, speech recognition, image processing, noise can-
cellation, modems, etc.
vi
TechnicalArticles
Related Documentation from Texas Instruments / Technical Articles
TMS320C548/C549 Bootloader Technical Reference (literature number

SPRU288) describes the process the bootloader uses to transfer user
code from an external source to the program memory at power up. (Pres-
ently available only on the internet.)
TMS320 DSP Development Support Reference Guide (literature number
SPRU011) describes the TMS320 family of digital signal processors and
the tools that support these devices. Included are code-generation tools
(compilers, assemblers, linkers, etc.) and system integration and debug
tools (simulators, emulators, evaluation modules, etc.). Also covered are
available documentation, seminars, the university program, and factory
repair and exchange.
Technical Articles
A wide variety of related documentation is available on digital signal processing.
These references fall into one of the following application categories:
- General-Purpose DSP
- Graphics/Imagery
- Speech/Voice
- Control
- Multimedia
- Military
- Telecommunications
- Automotive
- Consumer
- Medical
- Development Support
In the following list, references appear in alphabetical order according to

author. The documents contain beneficial information regarding designs, op-
erations, and applications for signal-processing systems; all of the documents
provide additional references. Texas Instruments strongly suggests that you
refer to these publications.
General-Purpose DSP:
1) Chassaing, R., Horning, D.W., “Digital Signal Processing with Fixed and
Floating-Point Processors” , CoED, USA, Volume 1, Number 1, pages 1-4,
March 1991.
2) Defatta, David J., Joseph G. Lucas, and William S. Hodgkiss, Digital Sig-
nal Processing: A System Design Approach, New York: John Wiley, 1988.
Read This First vii

Technical Articles
3) Erskine, C., and S. Magar, “Architecture and Applications of a Second-

Generation Digital Signal Processor,” Proceedings of IEEE International
Conference on Acoustics, Speech, and Signal Processing, USA, 1985.
4) Essig, D., C. Erskine, E. Caudel, and S. Magar, “A Second-Generation

Digital Signal Processor,” IEEE Journal of Solid-State Circuits, USA, Vol-
ume SC-21, Number 1, pages 86-91, February 1986.
5) Frantz, G., K. Lin, J. Reimer, and J. Bradley, “The Texas Instruments

TMS320C25 Digital Signal Microcomputer,” IEEE Microelectronics, USA,
Volume 6, Number 6, pages 10-28, December 1986.
6) Gass, W., R. Tarrant, T. Richard, B. Pawate, M. Gammel, P. Rajasekaran,

R. Wiggins, and C. Covington, “Multiple Digital Signal Processor Environ-
ment for Intelligent Signal Processing,” Proceedings of the IEEE, USA,
Volume 75, Number 9, pages 1246-1259, September 1987.
7) Jackson, Leland B., Digital Filters and Signal Processing, Hingham, MA:
Kluwer Academic Publishers, 1986.
8) Jones, D.L., and T.W. Parks, A Digital Signal Processing Laboratory Using
the TMS32010, Englewood Cliffs, NJ: Prentice-Hall, Inc., 1987.
9) Lim, Jae, and Alan V. Oppenheim, Advanced Topics in Signal Processing,

Englewood Cliffs, NJ: Prentice- Hall, Inc., 1988.
10) Lin, K., G. Frantz, and R. Simar, Jr., “The TMS320 Family of Digital Signal
Processors,” Proceedings of the IEEE, USA, Volume 75, Number 9, pages
1143-1159, September 1987.
11) Lovrich, A., Reimer, J., “An Advanced Audio Signal Processor” , Digest of
Technical Papers for 1991 International Conference on Consumer Elec-
tronics, June 1991.
12) Magar, S., D. Essig, E. Caudel, S. Marshall and R. Peters, “An NMOS Digi-
tal Signal Processor with Multiprocessing Capability,” Digest of IEEE Inter-
national Solid-State Circuits Conference, USA, February 1985.
13) Oppenheim, Alan V., and R.W. Schafer, Digital Signal Processing, Engle-
wood Cliffs, NJ: Prentice-Hall, Inc., 1975 and 1988.
14) Papamichalis, P.E., and C.S. Burrus, “Conversion of Digit-Reversed to Bit-

Reversed Order in FFT Algorithms,” Proceedings of ICASSP 89, USA,
pages 984-987, May 1989.
15) Papamichalis, P., and R. Simar, Jr., “The TMS320C30 Floating-Point Digi-
tal Signal Processor,” IEEE Micro Magazine, USA, pages 13-29, Decem-
ber 1988.
viii
Technical Articles
16) Papamichalis, P.E., “FFT Implementation on the TMS320C30,” Proceed-

ings of ICASSP 88, USA, Volume D, page 1399, April 1988.
17) Parks, T.W., and C.S. Burrus, Digital Filter Design, New York, NY: John Wiley
and Sons, Inc., 1987.
18) Peterson, C., Zervakis, M., Shehadeh, N., “Adaptive Filter Design and
Implementation Using the TMS320C25 Microprocessor” , Computers in
Education Journal, USA, Volume 3, Number 3, pages 12-16, July-Sep-
tember 1993.
19) Prado, J., and R. Alcantara, “A Fast Square-Rooting Algorithm Using a

Digital Signal Processor,” Proceedings of IEEE, USA, Volume 75, Number
2, pages 262-264, February 1987.
20) Rabiner, L.R. and B. Gold, Theory and Applications of Digital Signal Pro-
cessing, Englewood Cliffs, NJ: Prentice-Hall, Inc., 1975.
21) Simar, Jr., R., and A. Davis, “The Application of High-Level Languages to
Single-Chip Digital Signal Processors,” Proceedings of ICASSP 88, USA,
Volume D, page 1678, April 1988.
22) Simar, Jr., R., T. Leigh, P. Koeppen, J. Leach, J. Potts, and D. Blalock, “A
40 MFLOPS Digital Signal Processor: the First Supercomputer on a Chip,”
Proceedings of ICASSP 87, USA, Catalog Number 87CH2396-0, Volume 1,
pages 535-538, April 1987.
23) Simar, Jr., R., and J. Reimer, “The TMS320C25: a 100 ns CMOS VLSI Digi-
tal Signal Processor,” 1986 Workshop on Applications of Signal Processing
to Audio and Acoustics, September 1986.
24) Texas Instruments, Digital Signal Processing Applications with the TMS320
Family, 1986; Englewood Cliffs, NJ: Prentice-Hall, Inc., 1987.
25) Treichler, J.R., C.R. Johnson, Jr., and M.G. Larimore, A Practical Guide
to Adaptive Filter Design, New York, NY: John Wiley and Sons, Inc., 1987.
Graphics/Imagery:
1) Reimer, J., and A. Lovrich, “Graphics with the TMS32020,” WESCON/85

Conference Record, USA, 1985.
Read This First ix

Technical Articles
Speech/Voice:
1) DellaMorte, J., and P. Papamichalis, “Full-Duplex Real-Time Implementa-

tion of the FED-STD-1015 LPC-10e Standard V.52 on the TMS320C25,”
Proceedings of SPEECH TECH 89, pages 218-221, May 1989.
2) Gray, A.H., and J.D. Markel, Linear Prediction of Speech, New York, NY:
Springer-Verlag, 1976.
3) Frantz, G.A., and K.S. Lin, “A Low-Cost Speech System Using the
TMS320C17,” Proceedings of SPEECH TECH ’87, pages 25-29, April
1987.
4) Papamichalis, P., and D. Lively, “Implementation of the DOD Standard

LPC-10/52E on the TMS320C25,” Proceedings of SPEECH TECH ’87,
pages 201-204, April 1987.
5) Papamichalis, Panos, Practical Approaches to Speech Coding, Engle-

wood Cliffs, NJ: Prentice-Hall, Inc., 1987.
6) Pawate, B.I., and G.R. Doddington, “Implementation of a Hidden Markov

Model-Based Layered Grammar Recognizer,” Proceedings of ICASSP
89, USA, pages 801- 804, May 1989.
7) Rabiner, L.R., and R.W. Schafer, Digital Processing of Speech Signals,

Englewood Cliffs, NJ: Prentice-Hall, Inc., 1978.
8) Reimer, J.B. and K.S. Lin, “TMS320 Digital Signal Processors in Speech
Applications,” Proceedings of SPEECH TECH ’88, April 1988.
9) Reimer, J.B., M.L. McMahan, and W.W. Anderson, “Speech Recognition

for a Low-Cost System Using a DSP,” Digest of Technical Papers for 1987
International Conference on Consumer Electronics, June 1987.
Control:
1) Ahmed, I., “16-Bit DSP Microcontroller Fits Motion Control System Applica-
tion,” PCIM, October 1988.
2) Ahmed, I., “Implementation of Self Tuning Regulators with TMS320 Family

of Digital Signal Processors,” MOTORCON ’88, pages 248-262, Septem-
ber 1988.
3) Ahmed, I., and S. Lindquist, “Digital Signal Processors: Simplifying High-

Performance Control,” Machine Design, September 1987.
4) Ahmed, I., and S. Meshkat, “Using DSPs in Control,” Control Engineering,

February 1988.
x
Technical Articles
5) Allen, C. and P. Pillay, “TMS320 Design for Vector and Current Control of
AC Motor Drives” , Electronics Letters, UK, Volume 28, Number 23, pages
2188-2190, November 1992.
6) Bose, B.K., and P.M. Szczesny, “A Microcomputer-Based Control and

Simulation of an Advanced IPM Synchronous Machine Drive System for
Electric Vehicle Propulsion,” Proceedings of IECON ’87, Volume 1, pages
454-463, November 1987.
7) Hanselman, H., “LQG-Control of a Highly Resonant Disc Drive Head Posi-

tioning Actuator,” IEEE Transactions on Industrial Electronics, USA, Vol-
ume 35, Number 1, pages 100-104, February 1988.
8) Lovrich, A., G. Troullinos, and R. Chirayil, “An All-Digital Automatic Gain

Control,” Proceedings of ICASSP 88, USA, Volume D, page 1734, April
1988.
9) Matsui, N. and M. Shigyo, “Brushless DC Motor Control Without Position

and Speed Sensors” , IEEE Transactions on Industry Applications, USA,
Volume 28, Number 1, Part 1, pages 120-127, January-February 1992.
10) Meshkat, S., and I. Ahmed, “Using DSPs in AC Induction Motor Drives,”
Control Engineering, February 1988.
11) Panahi, I. and R. Restle, “DSPs Redefine Motion Control” , Motion Control
Magazine, December 1993.
Multimedia:
1) Reimer, J., “DSP-Based Multimedia Solutions Lead Way Enhancing

Audio Compression Performance” , Dr. Dobbs Journal, December 1993.
2) Reimer, J., G. Benbassat, and W. Bonneau Jr., “Application Processors:

Making PC Multimedia Happen” , Silicon Valley PC Design Conference,
July 1991.
Military:
1) Papamichalis, P., and J. Reimer, “Implementation of the Data Encryption

Standard Using the TMS32010,” Digital Signal Processing Applications,
1986.
Read This First xi

Technical Articles
Telecommunications:
1) Ahmed, I., and A. Lovrich, “Adaptive Line Enhancer Using the

TMS320C25,” Conference Records of Northcon/86, USA, 14/3/1-10,
September/October 1986.
2) Casale, S., R. Russo, and G. Bellina, “Optimal Architectural Solution Us-

ing DSP Processors for the Implementation of an ADPCM Transcoder,”
Proceedings of GLOBECOM ’89, pages 1267-1273, November 1989.
3) Cole, C., A. Haoui, and P. Winship, “A High-Performance Digital Voice

Echo Canceller on a SINGLE TMS32020,” Proceedings of ICASSP 86,
USA, Catalog Number 86CH2243-4, Volume 1, pages 429-432, April
1986.
4) Cole, C., A. Haoui, and P. Winship, “A High-Performance Digital Voice

Echo Canceller on a Single TMS32020,” Proceedings of IEEE Internation-
al Conference on Acoustics, Speech and Signal Processing, USA, 1986.
5) Lovrich, A., and J. Reimer, “A Multi-Rate Transcoder,” Transactions on

Consumer Electronics, USA, November 1989.
6) Lovrich, A. and J. Reimer, “A Multi-Rate Transcoder” , Digest of Technical

Papers for 1989 International Conference on Consumer Electronics, June 7-9,
1989.
7) Lu, H., D. Hedberg, and B. Fraenkel, “Implementation of High-Speed Voice-

band Data Modems Using the TMS320C25,” Proceedings of ICASSP 87,
USA, Catalog Number 87CH2396-0, Volume 4, pages 1915-1918, April
1987.
8) Mock, P., “Add DTMF Generation and Decoding to DSP- µP Designs,”

Electronic Design, USA, Volume 30, Number 6, pages 205-213, March
1985.
9) Reimer, J., M. McMahan, and M. Arjmand, “ADPCM on a TMS320 DSP

Chip,” Proceedings of SPEECH TECH 85, pages 246-249, April 1985.
10) Troullinos, G., and J. Bradley, “Split-Band Modem Implementation Using the
TMS32010 Digital Signal Processor,” Conference Records of Electro/86 and
Mini/Micro Northeast, USA, 14/1/1-21, May 1986.
Automotive:
1) Lin, K., “Trends of Digital Signal Processing in Automotive,” International

Congress on Transportation Electronic (CONVERGENCE ’88), October
1988.
xii
Technical Articles
Consumer:
1) Frantz, G.A., J.B. Reimer, and R.A. Wotiz, “Julie, The Application of DSP
to a Product,” Speech Tech Magazine, USA, September 1988.
2) Reimer, J.B., and G.A. Frantz, “Customization of a DSP Integrated Circuit

for a Customer Product,” Transactions on Consumer Electronics, USA,
August 1988.
3) Reimer, J.B., P.E. Nixon, E.B. Boles, and G.A. Frantz, “Audio Customiza-
tion of a DSP IC,” Digest of Technical Papers for 1988 International Con-
ference on Consumer Electronics, June 8-10 1988.
Medical:
1) Knapp and Townshend, “A Real-Time Digital Signal Processing System

for an Auditory Prosthesis,” Proceedings of ICASSP 88, USA, Volume A,
page 2493, April 1988.
2) Morris, L.R., and P.B. Barszczewski, “Design and Evolution of a Pocket-

Sized DSP Speech Processing System for a Cochlear Implant and Other
Hearing Prosthesis Applications,” Proceedings of ICASSP 88, USA, Vol-
ume A, page 2516, April 1988.
Development Support:
1) Mersereau, R., R. Schafer, T. Barnwell, and D. Smith, “A Digital Filter Design

Package for PCs and TMS320,” MIDCON/84 Electronic Show and Conven-
tion, USA, 1984.
2) Simar, Jr., R., and A. Davis, “The Application of High-Level Languages to

Single-Chip Digital Signal Processors,” Proceedings of ICASSP 88, USA,
Volume 3, pages 1678-1681, April 1988.
Read This First xiii

Trademarks
Trademarks
HP-UX is a trademark of Hewlett-Packard Company.
MS-DOS is a registered trademark of Microsoft Corporation.
OS/2 and PC-DOS are trademarks of International Business Machines Corpo-

ration.
PAL is a registered trademark of Advanced Micro Devices, Inc.
Solaris and SunOS are trademarks of Sun Microsystems, Inc.
SPARC is a trademark of SPARC International, Inc., but licensed exclusively

to Sun Microsystems, Inc.
Windows is a registered trademark of Microsoft Corporation.
320 Hotline On-line, TI, XDS510, and XDS510WS are trademarks of Texas
Instruments Incorporated.
Micro Star is a trademark of Texas Instruments Incorporated.
SPI is a trademark of the Motorola Corporation.
xiv
Contents
Contents
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-1
Provides an overview of the enhanced peripherals available on the ’5402, ’5410, and ’5420
devices.
1.1 Overview of the ’54x Enhanced Peripherals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-2
1.1.1 Multi-channel Buffered Serial Ports (McBSPs) . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-2
1.1.2 Direct Memory Access (DMA) Controller . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-2
1.1.3 Host Port Interfaces (HPI-8 and HPI-16) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1-2
2 Multichannel Buffered Serial Port (McBSP) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-1

Describes the features and operation of the multichannel buffered serial ports (McBSPs).
2.1 McBSP Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-2
2.2 McBSP General Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-3
2.2.1 Serial Port Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-6
2.2.2 Receive and Transmit Control Registers: RCR[1,2] and XCR[1,2] . . . . . . . . . 2-15
2.3 Data Transmission and Reception Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-22
2.3.1 Resetting the Serial Port: (R/X)RST, and RESET . . . . . . . . . . . . . . . . . . . . . . . 2-22
2.3.2 Determining Ready Status . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-26
2.3.3 CPU Interrupts: (R/X)INT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-27
2.3.4 Frame and Clock Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-27
2.3.5 McBSP Standard Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-38
2.3.6 Frame-Synchronization Ignore . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-40
2.3.7 Serial Port Exception Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-43
2.3.8 Receive Data Justification and Sign-Extension: RJUST . . . . . . . . . . . . . . . . . . 2-52
2.4 µ-LAW/A-LAW Companding Hardware Operation: (R/X)COMPAND . . . . . . . . . . . . . . 2-53
2.4.1 Companding Internal Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-55
2.5 Programmable Clock and Framing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-57
2.5.1 Sample Rate Generator Clocking and Framing . . . . . . . . . . . . . . . . . . . . . . . . . 2-58
2.5.2 Data Clock Generation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-62
2.5.3 Frame-Sync Signal Generation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-66
2.5.4 Clocking Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-69
2.6 Multichannel Selection Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-72
2.6.1 Multichannel Operation Control Registers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-73
2.6.2 Enabling Multichannel Selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-76
2.6.3 Enabling and Masking of Channels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-76
2.6.4 A-bis interface functionality (available on ’5410 only) . . . . . . . . . . . . . . . . . . . . 2-83
xv
Contents
2.7 SPI Protocol: McBSP Clock Stop Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-86

2.7.1 Clock Stop Mode Configuration and Signal Descriptions . . . . . . . . . . . . . . . . . 2-88
2.7.2 McBSP Operation as an SPI Master . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-90
2.7.3 McBSP Operation as an SPI Slave . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-92
2.7.4 McBSP Initialization for SPI Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-93
2.8 Emulation FREE and SOFT Bits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-95
2.9 McBSP Pins as General Purpose I/O . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-96
2.10 McBSP Operation in Power-Down Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-98
2.11 McBSP Programming Example Code . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-99
3 Direct Memory Access (DMA) Controller . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-1

Describes the direct memory access (DMA) controller, channels, and registers.
3.1 DMA Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-2
3.2 DMA Operation and Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-4
3.2.1 Register subaddressing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-4
3.2.2 DMA Channel Priority and Enable Control Register . . . . . . . . . . . . . . . . . . . . . . 3-7
3.2.3 Channel-Context Registers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-11
3.3 Extended Addressing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-33
3.4 DMA Memory Maps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-34
3.4.1 ’5402 DMA Memory Map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-34
3.4.2 ’5410 DMA Memory Map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-35
3.4.3 ’5420 DMA Memory Map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-36
3.5 DMA Transfer Latency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-39
3.6 Enhanced Host Port Interface Access through the DMA Controller . . . . . . . . . . . . . . . 3-42
3.7 Interprocessor FIFO Communication on the ’5420 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-43
3.8 DMA Operation in Power-Down Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-43
3.9 Programming Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-44
4 Enhanced 8-Bit Host Port Interface (HPI-8) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-1

Describes the operation and function of the enhanced 8-bit host port interface (HPI-8).
4.1 Introduction to the Enhanced 8-Bit Host Port Interface (HPI-8) . . . . . . . . . . . . . . . . . . . . 4-2
4.2 HPI-8 Basic Functional Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-4
4.3 Details of HPI-8 Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-6
4.3.1 HPI-8 Address Register and Memory Map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-9
4.3.2 Extended HPI-8 Addressing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-10
4.3.3 Address Autoincrement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-11
4.3.4 HPI-8 Control Register Bits and Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-12
4.4 Host Read/Write Access to HPI-8 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-14
4.4.1 Latency of HPI-8 Accesses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-16
4.4.2 Access Sequence Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-19
4.5 DSPINT and HINT Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-23
4.5.1 Host Device Using DSPINT to Interrupt the ’54x . . . . . . . . . . . . . . . . . . . . . . . . 4-23
4.5.2 ’54x Using HINT to Interrupt the Host Device . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-23
4.6 Considerations for HPI-8 Transfers While Changing Clock Modes . . . . . . . . . . . . . . . . 4-24
xvi
Contents
4.7 Considerations in IDLE Use . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-26

4.7.1 HPI-8 Accesses During IDLE1 and IDLE2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-26
4.7.2 HPI-8 Accesses During IDLE3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-26
4.8 Effects of Reset on HPI-8 Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-28
4.8.1 Accesses to HPI-8 After Reset . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-28
4.8.2 Access to HPI-8 During Reset (’5410 Only) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-28
4.9 HPI-8 Data Pins as General Purpose I/O Pins (Not Available on ’5410) . . . . . . . . . . . 4-30
4.9.1 Using the GPIO feature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-36
5 Enhanced 16-Bit Host Port Interface (HPI-16) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-1

Describes the operation and function of the enhanced 16-bit host port interface (HPI-16).
5.1 HPI-16 Operational Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-2
5.2 Multiplexed Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-7
5.2.1 Host Accesses With HAS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-10
5.2.2 Host Accesses Without HAS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-11
5.2.3 Autoincrement Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-12
5.3 Non-Multiplexed Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-15
5.4 HPI-16 Memory Map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-18
5.5 HPI-16 and DMA Interaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-19
5.6 HPI-16 Operation During Reset . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-21
5.7 HPI-16 Operation During IDLEn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-21
5.8 Changes in DSP Clock Modes That Affect the HPI-16 . . . . . . . . . . . . . . . . . . . . . . . . . . 5-22
6 Interprocessor Communications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-1

Describes multi-core communications, including core-to-core FIFO communications and
external memory interface- to-host port interface (EMIF-to-HPI) communications.
6.1 Communication Within Multi-Core DSPs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-2
6.2 The Bi-Directional FIFO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-3
6.3 Accessing the HPI-16 From External Memory Space . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-7
6.4 Subsystem Communication Using McBSP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-10
6.5 Interprocessor Interrupts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-11
7 Enhanced External Bus Interface (EnhXIO) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-1

Describes the improved functionality of the enhanced external parallel interface.
7.1 Introduction to the Enhanced External Parallel Interface (XIO2) . . . . . . . . . . . . . . . . . . . 7-2
7.1.1 Additional Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-2
7.2 Bus Sequences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-3
A Glossary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . A-1
B Revision History . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . B-1
Contents xvii
Figures
Figures
2−1 McBSP Internal Block Diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-3
2−2 Serial Port Control Register 1 (SPCR1) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-7
2−3 Serial Port Control Register 2 (SPCR2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-10
2−4 Pin Control Register (PCR) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-12
2−5 Receive Control Register 1 (RCR1) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-16
2−6 Receive Control Register 2 (RCR2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-17
2−7 Transmit Control Register 1 (XCR1) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-19
2−8 Transmit Control Register 2 (XCR2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-20
2−9 Frame and Clock Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-28
2−10 Receive Data Clocking . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-30
2−11 Dual-Phase Frame Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-30
2−12 Single-Phase Frame of Four 8-Bit Words . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-33
2−13 Single-Phase Frame of One 32-Bit Word . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-34
2−14 Data Delay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-34
2−15 Two-bit Data Delay Used to Discard Framing Bit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-35
2−16 AC97 Dual-Phase Frame Format . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-36
2−17 AC97 Bit Timing Near Frame-Synchronization Example . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-36
2−18 DX Enabler in Normal Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-37
2−19 DX Enabler in A-bis mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-37
2−20 McBSP Standard Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-38
2−21 Receive Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-39
2−22 Transmit Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-39
2−23 Maximum Frame Frequency Receive/Transmit (R/X)DATDLY = 0) . . . . . . . . . . . . . . . . . . 2-40
2−24 Maximum Packet Frequency Operation with 8-bit Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-41
2−25 Data Packing at Maximum Packet Frequency With (R/X)FIG=1 . . . . . . . . . . . . . . . . . . . . 2-42
2−26 Unexpected Frame Synchronization With (R/X)FIG=0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-43
2−27 Unexpected Frame Synchronization With (R/X)FIG = 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-43
2−28 Serial Port Receive Overrun . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-45
2−29 Serial Port Receive Overrun Avoided . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-46
2−30 Response to Receive Frame-Synchronization Pulse . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-47
2−31 Unexpected Receive Synchronization Pulse . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-48
2−32 Transmit With Data Overwrite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-48
2−33 Transmit Empty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-49
2−34 Transmit Empty Avoided . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-50
2−35 Response to Transmit Frame Synchronization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-51
2−36 Unexpected Transmit Frame-Synchronization Pulse . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-52
xviii
Figures
2−37 Companding Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-54

2−38 µ-Law Transmit Data Companding Format . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-54
2−39 A-Law Transmit Data Companding Format . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-54
2−40 Companding of Internal Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-55
2−41 Clock and Frame Generation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-57
2−42 Sample Rate Generator . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-58
2−43 Sample Rate Generator Register 1 (SRGR1) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-59
2−44 Sample Rate Generator Register 2 (SRGR2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-60
2−45 CLKG Synchronization and FSG generation when GSYNC = 1 and
CLKGDV = 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-63
2−46 CLKG Synchronization and FSG generation when GSYNC = 1 and
CLKGDV = 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-64
2−47 Programmable Frame Period and Width . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-67
2−48 ST-BUS and MVIP Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-69
2−49 Single-Rate Clock Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-70
2−50 Double-Rate Clock Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-71
2−51 Multichannel Control Register 1 (MCR1) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-73
2−52 Multichannel Control Register 2 (MCR2) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-75
2−53 Channel Enabling by Blocks in Partition A and B . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-77
2−54 XMCM Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-79
2−55 Receive Channel Enable Register Partition A (RCERA) . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-80
2−56 Receive Channel Enable Register Partition B (RCERB) . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-81
2−57 Transmit Channel Enable Register Partition A (XCERA) . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-81
2−58 Transmit Channel Enable Register Partition B (XCERB) . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-82
2−59 A-bis Mode Receive Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-84
2−60 A-bis Mode Transmit Operation (z Indicates High-Impedance) . . . . . . . . . . . . . . . . . . . . . 2-84
2−61 Typical SPI Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-86
2−62 SPI Configuration: McBSP as the Master . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-87
2−63 SPI Configuration: McBSP as the Slave . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-87
2−64 SPI Transfer with CLKSTP = 10b, and CLKXP = 0 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-88
3−1 Register Subaddressing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-4
3−2 DMA Channel Priority and Enable Control (DMPREC) Register . . . . . . . . . . . . . . . . . . . . . 3-7
3−3 DMA Sync Event and Frame Count (DMSFCn) Register . . . . . . . . . . . . . . . . . . . . . . . . . . 3-13
3−4 DMA Transfer Mode Control (DMMCRn) Register . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-17
3−5 Data Sorting Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-23
3−6 DMA Source Program Page Address Register (DMSRCP) . . . . . . . . . . . . . . . . . . . . . . . . . 3-33
3−7 DMA Destination Program Page Address Register (DMDSTP) . . . . . . . . . . . . . . . . . . . . . 3-33
4−1 Host Port Interface Block Diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-3
4−2 Generic System Block Diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-4
4−3 HPI Strobe and Select Logic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-8
4−4 HPI-8 Memory Maps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-9
Contents xix
Figures
4−5 HPIC Diagram — Host and ’54x Accesses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-13

4−6 HPI-8 Timing Diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-15
4−7 HPI-8 Transfers While Switching to DIV Clock Modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-24
4−8 HPI-8 Transfers While the ’54x is in IDLE3 Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-27
4−9 General Purpose I/O Control Register (GPIOCR) MMR Address 003Ch . . . . . . . . . . . . . 4-30
4−10 General Purpose I/O Status Register (GPIOSR) MMR address 003Dh . . . . . . . . . . . . . 4-32
4−11 GPIO Code Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-36
5−1 HPI Strobe and Select Logic . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-3
5−2 Interfacing to the HPI-16 in Multiplexed Mode (’VC5420) . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-7
5−3 HPIC Register . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-8
5−4 HPIA Write Using HAS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-10
5−5 HPIC Read Without Using HAS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-11
5−6 HPID Read using Autoincrement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-12
5−7 HPID Write Using Autoincrement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-13
5−8 Interfacing to the HPI-16 in Non-Multiplexed Mode (’VC5420) . . . . . . . . . . . . . . . . . . . . . . 5-15
5−9 HPID Read in Non-Multiplexed Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-16
5−10 HPID Write in Non-Multiplexed Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-17
5−11 ’VC5420 Memory Map Relative to the HPI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-18
5−12 HPI and DMA Interaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-19
5−13 HPI-16 Operation During the PLL to DIV Clock Mode Change . . . . . . . . . . . . . . . . . . . . . 5-22
6−1 ’VC5420 — The 2-Subsystem DSP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-2
6−2 ’VC5420 FIFO Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-3
6−3 DMA Configuration for FIFO Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-5
6−4 HPI-16 and External Memory Interface Connection (’VC5420) . . . . . . . . . . . . . . . . . . . . . . 6-7
6−5 McBSP-to-McBSP Connection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-10
7−1 Non-Consecutive Memory Read and I/O Read Bus Sequence . . . . . . . . . . . . . . . . . . . . . . 7-3
7−2 Consecutive Memory Read Bus Sequence (n = 3 reads) . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-4
7−3 Memory Write and I/O Write Bus Sequence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-5
xx
Tables
Tables
2−1 McBSP Interface Signals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-4
2−2 McBSP Registers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-5
2−3 McBSP CPU Interrupts and DMA Event Synchronization . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-6
2−4 Serial Port Control Register 1 (SPCR1) Bit-Field Descriptions . . . . . . . . . . . . . . . . . . . . . . 2-7
2−5 Serial Port Control Register 2 (SPCR2) Bit-Field Descriptions . . . . . . . . . . . . . . . . . . . . . . 2-10
2−6 Pin Control Register (PCR) Bit-Field Descriptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-12
2−7 Receive Control Register 1 (RCR1) Bit-Field Descriptions . . . . . . . . . . . . . . . . . . . . . . . . . 2-16
2−8 Receive Control Register 2 (RCR2) Bit-Field Descriptions . . . . . . . . . . . . . . . . . . . . . . . . . 2-17
2−9 Transmit Control Register 1 (XCR1) Bit-Field Descriptions . . . . . . . . . . . . . . . . . . . . . . . . . 2-19
2−10 Transmit Control Register 2 (XCR2) Bit-Field Descriptions . . . . . . . . . . . . . . . . . . . . . . . . . 2-20
2−11 Reset State of McBSP Pins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-23
2−12 RCR[1,2]/XCR[1,2] Bit-Fields Controlling Words per Frame and Bits per Word . . . . . . . 2-31
2−13 McBSP Receive/Transmit Frame Length (1,2) Configuration . . . . . . . . . . . . . . . . . . . . . . . 2-31
2−14 McBSP Receive/Transmit Word Length Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-32
2−15 Use of RJUST Field With 12-Bit Example Data 0xABC . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-52
2−16 Use of RJUST Field With 20-Bit Example Data 0xABCDE . . . . . . . . . . . . . . . . . . . . . . . . . 2-52
2−17 Sample Rate Generator Register 1 (SRGR1) Bit-Field Descriptions . . . . . . . . . . . . . . . . . 2-59
2−18 Sample Rate Generator Register 2 (SRGR2) Bit-Field Descriptions . . . . . . . . . . . . . . . . 2-60
2−19 Receive Clock Selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-65
2−20 Transmit Clock Selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-65
2−21 Receive Frame-Synchronization Selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-67
2−22 Transmit Frame-Synchronization Selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-68
2−23 Multichannel Control Register 1 (MCR1) Bit-Field Descriptions . . . . . . . . . . . . . . . . . . . . 2-73
2−24 Multichannel Control Register 2 (MCR2) Bit-Field Descriptions . . . . . . . . . . . . . . . . . . . . 2-75
2−25 Receive Channel Enable Register Partition A (RCERA) Bit-Field Descriptions . . . . . . . . 2-80
2−26 Receive Channel Enable Register Partition B (RCERB) Bit-Field Descriptions . . . . . . . . 2-81
2−27 Transmit Channel Enable Register Partition A (XCERA) Bit-Field Descriptions . . . . . . . 2-81
2−28 Transmit Channel Enable Register Partition B (XCERB) Bit-Field Descriptions . . . . . . . 2-82
2−29 Clock Stop Mode Configurations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-88
2−30 Register Bit Values for SPI Mode Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-90
2−31 Register Bit Values for SPI Master Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-91
2−32 Register Bit Values for SPI Slave Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-92
2−33 McBSP Clock Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-95
2−34 Configuration of Pins as General Purpose I/O . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-97
3−1 DMA Registers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-6
3−2 DMA Channel Priority and Enable Control (DMPREC) Register Bit/Field
Descriptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-8
Contents xxi
Tables
3−3 Multiplexed Interrupt Assignments for the ’5402 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-10

3−4 Multiplexed Interrupt Assignments for the ’5410 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-10
3−5 Multiplexed Interrupt Assignments for the ’5420 (each subsystem) . . . . . . . . . . . . . . . . . . 3-10
3−6 DMA Sync Event and Frame Count (DMSFCn) Register Bit/Field Descriptions . . . . . . . 3-13
3−7 DMA Sync Event Options on the ’5402 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-14
3−8 DMA Sync Event Options on the ’5410 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-15
3−9 DMA Sync Event Options on the ’5420 (each subsystem) . . . . . . . . . . . . . . . . . . . . . . . . . 3-16
3−10 DMA Transfer Mode Control (DMMCRn) Register Bit/Field Descriptions . . . . . . . . . . . . . 3-17
3−11 ABU Buffer Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-24
3−12 ABU Address Indexing Modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-25
3−13 DMA Block Transfer Interrupt Generation Modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-28
3−14 ’5402 DMA Memory Map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-34
3−15 ’5410 DMA Memory Map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-35
3−16 ’5420 DMA Memory Map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-37
3−17 DMA Transfer Cycle Times . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-39
4−1 Main Differences Between Enhanced 8-Bit HPI and Standard 8-Bit HPI . . . . . . . . . . . . . . 4-2
4−2 HPI-8 Register Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-5
4−3 HPI-8 Signal Names and Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-6
4−4 HPI-8 Input Control Signals and Function Selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-9
4−5 HPI Control Register (HPIC) Bit Descriptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-12
4−6 HPIC Host and ’54x Read/Write Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-13
4−7 HPI-8 Internal Delays by Access Type . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-17
4−8 Wait-State Generation Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-18
4−9 Initialization of BOB and HPIA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-20
4−10 Read Access to HPI-8 With Autoincrement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-21
4−11 Write Access to HPI With Autoincrement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-22
4−12 HPI-8 Operation During RESET (’5410 only) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-29
4−13 General Purpose I/O Control Register (GPIOCR) Bit Functions . . . . . . . . . . . . . . . . . . . 4-30
4−14 General Purpose I/O Status Register (GPIOSR) Bit Functions . . . . . . . . . . . . . . . . . . . . 4-32
5−1 HPI-16 Pin Descriptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-4
5−2 HCNTL0/1 Modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-8
5−3 HPIC Bit Descriptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-8
6−1 DMA Configuration to Support FIFO Transfers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-4
6−2 EMIF/HPI-16 Modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-8
6−3 MP/MC Bit Levels at Reset . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-8
xxii
Examples
Examples
2−1 Resetting and Configuring the Transmitter While Receiver is Running . . . . . . . . . . . . . . . 2-26
3−1 Using Register Subaddressing Without Autoincrement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-5
3−2 Using Register Subaddressing With Autoincrement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-5
3−3 Data Sorting by Address Modification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-22
3−4 ABU Buffer Size Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-25
3−5 Wrap Address Calculation for a Single-Word Transfer With Indexed Addressing . . . . . . 3-26
3−6 Wrap Address Calculation for a Double-Word Transfer With Indexed Addressing . . . . . 3-26
3−7 ’5402/’5420 ABU Interrupt Example − Even Size Buffer With +1 Index . . . . . . . . . . . . . . 3-29
3−8 ’5402/’5420 ABU Interrupt Example − Odd Size Buffer With +1 Index . . . . . . . . . . . . . . . 3-29
3−9 ’5402/’5420 ABU Interrupt Example − Even Size Buffer With −1 Index . . . . . . . . . . . . . . . 3-30
3−10 ’5402/’5420 ABU Interrupt Example − Odd Size Buffer With −1 Index . . . . . . . . . . . . . . . 3-30
3−11 ’5410 ABU Interrupt Example − Even Size Buffer With +1 Index . . . . . . . . . . . . . . . . . . . . 3-31
3−12 ’5410 ABU Interrupt Example − Odd Size Buffer With +1 Index . . . . . . . . . . . . . . . . . . . . . 3-31
3−13 ’5410 ABU Interrupt Example − Even Size Buffer With −1 Index . . . . . . . . . . . . . . . . . . . . 3-32
3−14 ’5410 ABU Interrupt Example − Odd Size Buffer With −1 Index . . . . . . . . . . . . . . . . . . . . . 3-32
3−15 DMA Channel Transfer Rate Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-40
3−16 Program Memory to Data Memory Transfer Without Autoincremented
Subaddressing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-46
3−17 Program Memory to Data Memory Transfer Using Autoincremented
Subaddressing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-48
3−18 Program Memory to Data Memory Transfer With Autoinitialization . . . . . . . . . . . . . . . . . . 3-50
3−19 McBSP Data Transfer in ABU Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-52
3−20 McBSP Data Transfer in Double-Word Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-54
3−21 McBSP to Data Memory Transfer With Data Sorting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-56
6−1 DMA Channels 0 and 1 Configured for FIFO Transmit and Receive . . . . . . . . . . . . . . . . . . 6-4
Contents xxiii
Running Title—Attribute Reference
xxiv
Chapter 1
Introduction
The ’54x device is a fixed-point digital signal processor (DSP) in the TMS320
family. The central processing unit (CPU), with its modified Harvard architec-
ture, minimizes power consumption and adds a high degree of parallelism.
System performance is further enhanced by versatile addressing modes and
instruction sets. These and other characteristics allow the ’54x to meet the
specific needs of real-time embedded applications such as telecommunica-
tions.
All ’54x devices have general-purpose I/O pins (XF and BIO), a timer (two on
the ’5402), a clock generator, a software-programmable wait-state generator,
and a programmable bank-switching module. Different types and quantities of
serial ports, host-port interfaces, and clock generators are specific to the vari-
ous ’54x devices.
This chapter discusses the enhanced peripherals available on the ’5402,

’5410, and ’5420 devices.
Topic Page
1.1 Overview of the ’54x Enhanced Peripherals . . . . . . . . . . . . . . . . . . . . . 1-2
1-1
1.1 Overview of the ’54x Enhanced Peripherals
The sections that follow provide an overview of the enhanced peripherals
available on the ’54x.
1.1.1 Multi-channel Buffered Serial Ports (McBSPs)

The’54x family provides high-speed, full-duplex multi-channel buffered serial
ports (McBSPs) that allow direct interface to other ’54x devices, codecs, and
other devices in a system. The MCBSPs are an enhanced version of the stan-
dard serial port interface found on other ’54x devices. Some features of the
McBSP include:
- Double-buffered transmit and triple-buffered receive operation to allow a

continuous data stream.
- Independent framing and clocking for receive and transmit.
- Multi-channel transmit and receive up to 128 channels.
- Data sizes including 8, 12, 16, 20, 24, and 32 bits.
- µ-law and A-law companding.
- Programmable polarity for both frame synchronization and clocks.
- Programmable internal clocks and frame synchronization.
For more information on the McBSPs, see Chapter 2.
1.1.2 Direct Memory Access (DMA) Controller

The 6-channel ’54x direct memory access (DMA) controller transfers data be-
tween points in the memory map without intervention by the CPU. The DMA
allows the following movements of data to occur in the background of CPU op-
eration: data to and from internal program/data memory; internal peripherals
such as the McBSP’s; and external memory devices. The DMA has six inde-
pendent programmable channels allowing six different contexts for DMA op-
eration. For more information on the DMA, see Chapter 3.
1.1.3 Host Port Interfaces (HPI-8 and HPI-16)

There are two enhanced host port interfaces on the ’54x, the HPI-8 and the
HPI-16. These are 8-bit and 16-bit parallel ports that provide an interface to
a host processor. Information is exchanged between the ’54x and the host pro-
cessor through ’54x on-chip memory that is accessible to both the host proces-
sor and the ’54x. For more details about the operation of the HPI-8 and HPI-16,
see Chapters 4 and 5, respectively.
1-2
Chapter 2
Multichannel Buffered Serial PortĂ(McBSP)
Depending on the specific device, the 54x digital signal processor provides
multiple high-speed, full-duplex, multichannel buffered serial ports (McBSPs)
that allow direct interface to other ’54x devices, codecs, and other devices in
a system. The ’5402 provides two, the ’5410 three, and the ’5420 six McBSPs.
They are based on the standard serial port interface found on other ’54x de-
vices.
This chapter describes the operation of the McBSPs, and includes register
definitions and timing diagrams.
Topic Page
2.1 McBSP Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-2

2.2 McBSP General Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-3
2.3 Data Transmission and Reception Flow . . . . . . . . . . . . . . . . . . . . . . . . 2-22
2.4 µ-LAW/A-LAW Companding Hardware Operation: (R/X)
COMPAND . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-53
2.5 Programmable Clock and Framing . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-57
2.6 Multichannel Selection Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-72
2.7 SPI Protocol: McBSP Clock Stop Mode . . . . . . . . . . . . . . . . . . . . . . . . 2-86
2.8 Emulation FREE and SOFT Bits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-96
2.9 McBSP Pins as General-Purpose I/O . . . . . . . . . . . . . . . . . . . . . . . . . . . 2-97
2.10 McBSP Operation in Power-Down Mode . . . . . . . . . . . . . . . . . . . . . . . 2-99
2.11 McBSP Programming Example Code . . . . . . . . . . . . . . . . . . . . . . . . . 2-100
2-1
2.1 McBSP Features
The McBSP is based on the standard serial port interface found on the
TMS320C2x, ’C20x, ’C5x, and ’C54x devices. The McBSP provides:
- Full-duplex communication
- Double-buffered transmit and triple-buffered receive data registers, which

allow a continuous data stream
- Independent framing and clocking for receive and transmit
- Direct interface to industry-standard codecs, analog interface chips (AICs),

and other serially connected A/D and D/A devices
- External shift clock generation, or an internal, programmable-frequency
shift clock
In addition, the McBSP has the following capabilities:
- Direct interface to:
J T1/E1 framers
J MVIP switching compatible and ST-BUS compliant devices including:
H MVIP framers
H H.100 framers
H SCSA framers
J IOM-2 compliant devices
J AC97 compliant devices (The necessary multi-phase frame-synchro-
nization capability is provided.)
J IIS compliant devices
J SPIt devices
- Multichannel transmit and receive up to 128 channels
- A wide selection of data sizes including 8, 12, 16, 20, 24, and 32 bits
Note: Data sizes are referred to as word or serial word throughout this document and
can be 8, 12, 16, 20, 24, or 32 bits, in contrast to the true definition of word which is
32 bits.
- µ-Law and A-Law companding
- 8-bit data transfers with option of LSB or MSB first
- Programmable polarity for both frame synchronization and data clocks
- Highly programmable internal clock and frame generation
2-2
2.2 McBSP General Description
The McBSP consists of a data path and a control path connected to external
devices by seven pins as shown in Figure 2−1.
Figure 2−1. McBSP Internal Block Diagram
McBSP
ÁÁÁÁÁÁÁÁÁÁÁÁÁ
Compand
ÁÁÁ ÁÁÁÁÁÁÁ
DR RSR RBR Expand DRR
ÁÁÁ ÁÁÁÁÁÁÁ
DX XSR Compress DXR
ÁÁÁÁÁÁÁÁÁÁÁ
ÁÁÁÁ
ÁÁÁÁ SPCR
ÁÁÁÁ
CLKX
ÁÁÁÁ
CLKR RCR
Clock and 16-bit
ÁÁÁÁ
FSX frame-sync peripheral
XCR
ÁÁÁÁ
FSR generation bus
and control
ÁÁÁÁ
CLKS SRGR
ÁÁÁÁ PCR
ÁÁÁÁ
ÁÁÁÁ
MCR
ÁÁÁÁ Multichannel RCER
ÁÁÁÁ
selection
ÁÁÁÁ
XCER
RINT
Interrupts to CPU
XINT
REVT
XEVT
Synchronization
events to DMA
REVTA
XEVTA
2-3
Data is communicated to devices interfacing the McBSP via the data transmit
(DX) pin for transmit and the data receive (DR) pin for receive. Control informa-
tion in the form of clocking and frame synchronization is communicated via
CLKX, CLKR, FSX, and FSR. The ’54x communicates with the McBSP through
16-bit-wide control registers accessible via the internal peripheral bus.
The CPU or the DMA controller reads the received data from the data receive
register (DRR[1,2]) and writes the data to be transmitted to the data transmit
register (DXR[1,2]). Data written to DXR[1,2] is shifted out to DX via the trans-
mit shift register (XSR[1,2]). Similarly, receive data on the DR pin is shifted into
the receive shift register (RSR[1,2]) and copied into the receive buffer register
(RBR[1,2]). RBR[1,2] is then copied to DRR[1,2], which can be read by the CPU
or the DMA controller. This allows simultaneous movement of internal and exter-
nal data communications.
DRR2, RBR2, RSR2, DXR2, and XSR2 registers are not utilized (written, read,
or shifted) if the receive/transmit word length, R/XWDLEN[1,2], is specified for
8-, 12-, or 16-bit mode.
The remaining registers accessible to the CPU configure the control mechanism
of the McBSP. These registers are listed in Table 2−2, McBSP Registers, on
page 2-5. The control block consists of internal clock generation, frame-syn-
chronization signal generation, and their control and multichannel selection.
This control block sends notification of important events to the CPU and DMA
controller via the two interrupt and four event signals shown in Table 2−3,
McBSP CPU Interrupts and DMA Event Synchronization, on page 2-6.
Table 2−1. McBSP Interface Signals
Pin I/O/Z† Description

CLKR I/O/Z Receive clock
CLKX I/O/Z Transmit clock
CLKS I External clock
DR I Received serial data
DX O/Z Transmitted serial data
FSR I/O/Z Receive frame synchronization
FSX I/O/Z Transmit frame synchronization

† I = Input, O = Output, Z = High-impedance
2-4
Table 2−2. McBSP Registers
Hex Address
Sub-
McBSP 0 McBSP 1 McBSP 2 Address Acronym Register Name{ Section
— — — RBR[1,2] McBSP receive buffer register 1 and 2 2.2
— — — RSR[1,2] McBSP receive shift register 1 and 2 2.2
— — — XSR[1,2] McBSP transmit shift register 1 and 2 2.2
0020 0040 0030 — DRR2x McBSP data receive register 2 2.2
0021 0041 0031 — DRR1x McBSP data receive register 1 2.2
0022 0042 0032 — DXR2x McBSP data transmit register 2 2.2
0023 0043 0033 — DXR1x McBSP data transmit register 1 2.2
0038 0048 0034 — SPSAx McBSP sub-address register
0039 0049 0035 0x0000 SPCR1x McBSP serial port control register 1 2.2.1
0039 0049 0035 0x0001 SPCR2x McBSP serial port control register 2 2.2.1
0039 0049 0035 0x0002 RCR1x McBSP receive control register 1 2.2.2
0039 0049 0035 0x0003 RCR2x McBSP receive control register 2 2.2.2
0039 0049 0035 0x0004 XCR1x McBSP transmit control register 1 2.2.2
0039 0049 0035 0x0005 XCR2x McBSP transmit control register 2 2.2.2
0039 0049 0035 0x0006 SRGR1x McBSP sample rate generator register 2.5.1.1
1
0039 0049 0035 0x0007 SRGR2x McBSP sample rate generator register 2.5.1.1
2
0039 0049 0035 0x0008 MCR1x McBSP multichannel register 1 2.6.1
0039 0049 0035 0x0009 MCR2x McBSP multichannel register 2 2.6.1
0039 0049 0035 0x000A RCERAx McBSP receive channel enable register 2.6.3.1
partition A
0039 0049 0035 0x000B RCERBx McBSP receive channel enable register 2.6.3.1
partition B
† RBR[1,2], RSR[1,2], and XSR[1,2] are not directly accessible via the CPU or DMA.
2-5
Table 2−2. McBSP Registers (Continued)
Hex Address
Sub-
McBSP 0 McBSP 1 McBSP 2 Address Acronym Register Name{ Section
0039 0049 0035 0x000C XCERAx McBSP transmit channel enable 2.6.3.1
register partition A
0039 0049 0035 0x000D XCERBx McBSP transmit channel enable 2.6.3.1
register partition B
0039 0049 0035 0x000E PCRx McBSP pin control register 2.2.1 &
2.9
† RBR[1,2], RSR[1,2], and XSR[1,2] are not directly accessible via the CPU or DMA.
Table 2−3. McBSP CPU Interrupts and DMA Event Synchronization

Interrupt
Name Description Section
RINT Receive interrupt to CPU 2.3.3
XINT Transmit interrupt to CPU 2.3.3
REVT Receive synchronization event to DMA 2.3.2.1
XEVT Transmit synchronization event to DMA 2.3.2.2
REVTA Receive synchronization eventA to DMA 2.6.4
XEVTA Transmit synchronization eventA to DMA 2.6.4
2.2.1 Serial Port Configuration

The serial port is configured via the two 16-bit serial port control registers 1 and
2 (SPCR[1,2]) and the Pin Control Register (PCR). These registers are shown
in Figure 2−2, Figure 2−3 and Figure 2−4, respectively. SPCR[1, 2] and the
PCR contain McBSP status information and also bits that can be configured
for desired operation. The operation of each bit-field is discussed in the sec-
tions listed in Table 2−4, Serial Port Control Register 1 (SPCR1) Bit-Field Des-
cription, on page 2-7; Table 2−5, Serial Port Control Register 2 (SPCR2) Bit-
Field Description, on page 2-10; and Table 2−6, Pin Control Register (PCR)
Bit-field Description, on page 2-12.
2-6
In addition to the PCR being used to configure the McBSP pins as inputs or
outputs during normal serial port operation, it is used to configure the serial
port pins as general purpose inputs or outputs during receiver and/or transmit-
ter reset. This is described in section 2.9, McBSP Pins as General Purpose I/O,
on page 2-97.
Figure 2−2. Serial Port Control Register 1 (SPCR1)

15 14 13 12 11 10 8
DLB RJUST CLKSTP reserved
RW,+0 RW,+0 RW,+0 RW,+0 R,+0
7 6 5 4 3 2 1 0
DXENA ABIS RINTM RSYNCERR RFULL RRDY RRST
RW,+0 RW,+0 RW,+0 RW,+0 R,+0 R,+0 RW,+0{
Note: R = Read, W = Write, +0 = Value at reset
† R, +0 means read-only, reset value is 0. RW, +0 means read and write allowed, reset value is 0.
Table 2−4. Serial Port Control Register 1 (SPCR1) Bit-Field Descriptions

Bit Name Function Section
15 DLB Digital Loop Back Mode 2.5.2.5
DLB = 0 Digital loop back mode disabled
DLB = 1 Digital loop back mode enabled
14−13 RJUST Receive Sign-Extension and Justification Mode 2.3.8
RJUST = 00 Right-justify and zero-fill MSBs in DRR[1,2]
RJUST = 01 Right-justify and sign-extend MSBs in DRR[1,2]
RJUST = 10 Left-justify and zero-fill LSBs in DRR[1,2]
RJUST = 11 Reserved
2-7
Table 2−4. Serial Port Control Register 1 (SPCR1) Bit-Field Descriptions (Continued)
12−11 CLKSTP Clock Stop Mode 2.7
CLKSTP = 0X Clock stop mode disabled. Normal clocking for

non-SPI mode.
Various SPI modes when:
CLKSTP = 10 and Clock starts with rising edge without delay

CLKXP = 0
CLKSTP = 10 and Clock starts with falling edge without delay

CLKXP = 1
CLKSTP = 11 and Clock starts with rising edge with delay

CLKXP = 0
CLKSTP = 11 and Clock starts with falling edge with delay

CLKXP = 1
10−8 reserved Reserved
7 DXENA DX Enabler. 2.3.4.8
DXENA = 0 DX enabler is off
DXENA = 1 DX enabler is on
6 ABIS ABIS Mode 2.6.4
ABIS = 0 A-bis mode is disabled
ABIS = 1 A-bis mode is enabled
5−4 RINTM Receive Interrupt Mode 2.3.3
RINTM = 00 RINT driven by RRDY (i.e. end of word) and end of

frame in A-bis mode.
RINTM = 01 RINT generated by end-of-block or end-of-frame in

multichannel operation
RINTM = 10 RINT generated by a new frame synchronization
RINTM=11 RINT generated by RSYNCERR
3 RSYNCERR Receive Synchronization Error 2.3.7.2
RSYNCERR = 0 No synchronization error
2-8
RSYNCERR = 1 Synchronization error detected by McBSP.
2 RFULL Receive Shift Register (RSR[1,2]) Full 2.3.7.1
RFULL = 0 RBR[1,2] is not in overrun condition
RFULL = 1 DRR[1,2] is not read, RBR[1,2] is full and RSR[1,2]

is also full with new word
1 RRDY Receiver Ready 2.3.2
RRDY = 0 Receiver is not ready.
RRDY = 1 Receiver is ready with data to be read from

DRR[1,2].
0 RRST Receiver reset. This resets and enables the receiver. 2.3.1
RRST = 0 The serial port receiver is disabled and in reset state.
RRST = 1 The serial port receiver is enabled.
2-9
Figure 2−3. Serial Port Control Register 2 (SPCR2)
15 14 13 12 11 10 9 8
reserved{ FREE SOFT
R,+0 RW,+0 RW,+0
7 6 5 4 3 2 1 0
FRST GRST XINTM XSYNCERR} XEMPTY XRDY XRST
RW,+0 RW,+0 RW,+0 RW,+0 R,+0 R,+0 RW,+0
† Note: This and all reserved bit-fields have NO storage associated with them; however, they are always read as 0.
‡ CAUTION: Writing a 1 to this bit sets the error condition; thus, it is mainly used for testing purposes or if this operation is desired.
Table 2−5. Serial Port Control Register 2 (SPCR2) Bit-Field Descriptions

15−10 rsvd Reserved
9 FREE Free Running Mode 2.8
FREE = 0 Free running mode is disabled
FREE = 1 Free running mode is enabled
8 SOFT Soft Bit 2.8
SOFT = 0 SOFT mode is disabled
SOFT = 1 SOFT mode is enabled
7 FRST Frame-Sync Generator Reset 2.3.1
FRST = 0 Frame-synchronization logic is reset. Frame-sync

signal FSG is not generated by the sample-rate
generator.
FRST = 1 Frame-sync signal FSG is generated after

(FPER+1) number of CLKG clocks; i.e., all frame
counters are loaded with their programmed values.
6 GRST Sample-Rate Generator Reset 2.3.1
GRST = 0 Sample rate generator is reset
GRST = 1 Sample rate generator is pulled out of reset. CLKG

is driven as per programmed value in sample rate
generator registers (SRGR[1,2]).
2-10
5−4 XINTM Transmit Interrupt Mode 2.3.3
XINTM = 00 XINT driven by XRDY (i.e., end of word) and end of

frame in A-bis mode.
XINTM = 01 XINT generated by end-of-block or end-of-frame in

multichannel operation
XINTM = 10 XINT generated by a new frame synchronization
XINTM=11 XINT generated by XSYNCERR
3 XSYNCERR Transmit Synchronization Error 2.3.7.2
XSYNCERR = 0 No synchronization error
XSYNCERR = 1 Synchronization error detected by McBSP.
2 XEMPTY Transmit Shift Register (XSR[1,2]) Empty 2.3.7.4
XEMPTY = 0 XSR[1,2] is empty
XEMPTY = 1 XSR[1,2] is not empty
1 XRDY Transmitter Ready 2.3.2
XRDY = 0 Transmitter is not ready.
XRDY = 1 Transmitter is ready for new data in DXR[1,2].
0 XRST Transmitter reset. This resets and enables the transmitter. 2.3.1
XRST = 0 The serial port transmitter is disabled and in reset

state.
XRST = 1 The serial port transmitter is enabled.
2-11
Figure 2−4. Pin Control Register (PCR)
15 14 13 12 11 10 9 8
reserved XIOEN RIOEN FSXM FSRM CLKXM CLKRM
R,+0 RW,+0 RW,+0 RW,+0 RW,+0 RW,+0 RW,+0
7 6 5 4 3 2 1 0
reserved CLKS_STAT DX_STAT DR_STAT FSXP FSRP CLKXP CLKRP
R,+0 R,+0 R,+0 R,+0 RW,+0 RW,+0 RW,+0 RW,+0
Table 2−6. Pin Control Register (PCR) Bit-Field Descriptions

15−14 reserved Reserved
13 XIOEN Transmit general purpose I/O mode only when XRST = 0 in SPCR[1,2] 2.9
XIOEN = 0 DX, FSX and CLKX are configured as serial port pins and
do not function as general-purpose I/Os.
XIOEN = 1 DX pin is a general purpose output. FSX and CLKX are

general purpose I/Os. These serial port pins do not
perform serial port operation.
12 RIOEN Receive general purpose I/O mode only when RRST = 0 in SPCR[1,2] 2.9
RIOEN = 0 DR, FSR, CLKR and CLKS are configured as serial port
pins and do not function as general-purpose I/Os.
RIOEN = 1 DR and CLKS pins are general purpose inputs; FSR and
CLKR are general purpose I/Os. These serial port pins do
not perform serial port operation. The CLKS pin is affected
by a combination of RRST and RIOEN signals of the
receiver.
2-12
Table 2−6. Pin Control Register (PCR) Bit-Field Descriptions (Continued)
11 FSXM Transmit Frame-Synchronization Mode 2.5.3.3
and 2.9
FSXM = 0 Frame-synchronization signal derived from an external

source
FSXM = 1 Frame synchronization is determined by the sample rate

generator frame-synchronization mode bit FSGM in
SRGR2.
10 FSRM Receive Frame-Synchronization Mode 2.5.3.2

and 2.9
FSRM = 0 Frame-synchronization pulses generated by an external

device. FSR is an input pin
FSRM = 1 Frame synchronization generated internally by sample

rate generator. FSR is an output pin except when
GSYNC=1 in SRGR (see section 2.5.1.1).
9 CLKXM Transmitter Clock Mode 2.5.2.7

and 2.9
CLKXM = 0 Transmitter clock is driven by an external clock with CLKX

as an input pin.
CLKXM = 1 CLKX is an output pin and is driven by the internal sample

rate generator.
During SPI mode (when CLKSTP is a non-zero value):
CLKXM = 0 McBSP is a slave and clock (CLKX) is driven by the SPI

master in the system. CLKR is internally driven by CLKX.
CLKXM = 1 McBSP is a master and generates the clock (CLKX) to

drive its receive clock (CLKR) and the shift clock of the
SPI-compliant slaves in the system.
2-13
8 CLKRM Receiver Clock Mode 2.5.2.6
and 2.9
Case 1: Digital loop back mode not set (DLB = 0) in SPCR1
CLKRM = 0 Receive clock (CLKR) is an input driven by an external

clock.
CLKRM = 1 CLKR is an output pin and is driven by the internal sample

rate generator.
Case 2: Digital loop back mode set (DLB=1) in SPCR1
CLKRM = 0 Receive clock (not the CLKR pin) is driven by transmit

clock (CLKX) which is based on the CLKXM bit in the PCR.
CLKR pin is in high-impedance.
CLKRM = 1 CLKR is an output pin and is driven by the transmit clock.

The transmit clock is derived based on the CLKXM bit in
the the PCR.
7 rsvd Reserved
6 CLKS_ STAT CLKS pin status. Reflects value on CLKS pin when selected as a general 2.9
purpose input.
5 DX_STAT DX pin status. Reflects value driven on to DX pin when selected as a 2.9
general purpose output.
4 DR_STAT DR pin status. Reflects value on DR pin when selected as a general

purpose input.
3 FSXP Transmit Frame-Synchronization Polarity 2.3.4.1

and 2.9
FSXP = 0 Frame-synchronization pulse FSX is active high
FSXP = 1 Frame-synchronization pulse FSX is active low
2 FSRP Receive Frame-Synchronization Polarity 2.3.4.1

and 2.9
FSRP = 0 Frame-synchronization pulse FSR is active high
FSRP = 1 Frame-synchronization pulse FSR is active low
2-14
1 CLKXP Transmit Clock Polarity 2.3.4.1
and 2.9
CLKXP = 0 Transmit data sampled on rising edge of CLKX
CLKXP = 1 Transmit data sampled on falling edge of CLKX
0 CLKRP Receive Clock Polarity 2.3.4.1

and 2.9
CLKRP = 0 Receive data sampled on falling edge of CLKR
CLKRP = 1 Receive data sampled on rising edge of CLKR
2.2.2 Receive and Transmit Control Registers: RCR[1,2] and XCR[1,2]

The receive and transmit control registers (RCR[1,2] and XCR[1,2]) configure
various parameters of the receive and transmit operations, respectively. They
are shown in Figure 2−5, Receive Control Register 1 (RCR1), on page 2-16;
Figure 2−6, Receive Control Register 2 (RCR2), on page 2-17; Figure 2−7,
Transmit Control Register 1 (XCR1), on page 2-19; and Figure 2−8, Transmit
Control Register 2 (XCR2), on page 2-20.
The operation of each bit-field is discussed in the sections listed in Table 2−7,
Receive Control Register 1 (RCR1) Bit-Field Description, on page 2-16;
Table 2−8, Receive Control Register 2 (RCR2) Bit-Field Description, on page
2-17; Table 2−9, Transmit Control Register 1 (XCR1) Bit-Field Description, on
page 2-19; and Table 2−10, Transmit Control Register 2 (XCR2) Bit-Field De-
scription, on page 2-20.
2-15
Figure 2−5. Receive Control Register 1 (RCR1)
15 14 8 7 5 4 0
rsvd RFRLEN1 RWDLEN1 reserved
R,+0 RW,+0 RW,+0 R,+0
Table 2−7. Receive Control Register 1 (RCR1) Bit-Field Descriptions

15 rsvd Reserved
14−8 RFRLEN1 Receive Frame Length 1 2.3.4.3
RFRLEN1 = 000 0000 1 word per frame
RFRLEN1 = 000 0001 2 words per frame
7−5 RWDLEN1 Receive Word Length 1 2.3.4.4
RWDLEN1 = 000 8 bits
RWDLEN1 = 11X Reserved
4−0 rsvd Reserved
2-16
Figure 2−6. Receive Control Register 2 (RCR2)
15 14 8 7 5 4 3 2 1 0
RPHASE RFRLEN2 RWDLEN2 RCOMPAND RFIG RDATDLY
RW,+0 RW,+0 RW,+0 RW,+0 RW,+0 RW,+0
Table 2−8. Receive Control Register 2 (RCR2) Bit-Field Descriptions

15 RPHASE Receive Phases 2.3.4.2
RPHASE = 0 Single-phase frame
RPHASE = 1 Dual-phase frame
14−8 RFRLEN2 Receive Frame Length 2 2.3.4.3
RFRLEN2 = 000 0000 1 word per frame
7−5 RWDLEN2 Receive Word Length 2 2.3.4.4
RWDLEN2 = 11X Reserved
2-17
Table 2−8. Receive Control Register 2 (RCR2) Bit-Field Descriptions (Continued)
4−3 RCOMPAND Receive companding mode. Modes other than 00b are only enabled 2.4
when the appropriate RWDLEN is 000b, indicating 8-bit data.
RCOMPAND = 00 No companding, data transfer starts with

MSB first.
RCOMPAND = 01 No companding, 8-bit data, transfer starts

with LSB first.
RCOMPAND = 10 Compand using µ-law for receive data.
RCOMPAND = 11 Compand using A-law for receive data.
2 RFIG Receive Frame Ignore 2.3.6.2
RFIG = 0 Receive frame-synchronization pulses

after the first restarts the transfer.
RFIG = 1 Receive frame-synchronization pulses

after the first are ignored.
1−0 RDATDLY Receive data delay 2.3.4.6
RDATDLY = 00 0-bit data delay
RDATDLY = 11 Reserved
2-18
Figure 2−7. Transmit Control Register 1 (XCR1)
15 14 8 7 5 4 0
rsvd XFRLEN1 XWDLEN1 rsvd
R,+0 RW,+0 RW,+0 R,+0
Table 2−9. Transmit Control Register 1 (XCR1) Bit-Field Descriptions

15 rsvd Reserved
14−8 XFRLEN1 Transmit Frame Length 1 2.3.4.3
XFRLEN1 = 000 0000 1 word per frame
XFRLEN1 = 000 0001 2 words per frame
7−5 XWDLEN1 Transmit Word Length 1 2.3.4.4
XWDLEN1 = 000 8 bits
XWDLEN1 = 11X Reserved
4−0 rsvd Reserved
2-19
Figure 2−8. Transmit Control Register 2 (XCR2)
15 14 8 7 5 4 3 2 1 0
XPHASE XFRLEN2 XWDLEN2 XCOMPAND XFIG XDATDLY
RW,+0 RW,+0 RW,+0 RW,+0 RW,+0 RW,+0
Table 2−10. Transmit Control Register 2 (XCR2) Bit-Field Descriptions

15 XPHASE Transmit Phases 2.3.4.2
XPHASE = 0 Single-phase frame
XPHASE = 1 Dual-phase frame
14−8 XFRLEN2 Transmit Frame Length 2 2.3.4.3
XFRLEN2 = 000 0000 1 word per frame
7−5 XWDLEN2 Transmit Word Length 2 2.3.4.4
XWDLEN2 = 11X Reserved
2-20
Table 2−10. Transmit Control Register 2 (XCR2) Bit-Field Descriptions (Continued)
4−3 XCOMPAND Transmit companding mode. Modes other than 00b are only enabled 2.4
when the appropriate XWDLEN is 000b, indicating 8-bit data.
XCOMPAND = 00 No companding, data transfer starts with

MSB first.
XCOMPAND = 01 No companding, 8-bit data, transfer starts

with LSB first.
XCOMPAND = 10 Compand using µ-law for transmit data.
XCOMPAND = 11 Compand using A-law for transmit data.
2 XFIG Transmit Frame Ignore 2.3.6.2
XFIG = 0 Transmit frame-synchronization pulses

after the first restarts the transfer.
XFIG = 1 Transmit frame-synchronization pulses

after the first are ignored.
1−0 XDATDLY Transmit Data Delay 2.3.4.6
XDATDLY = 00 0-bit data delay
XDATDLY = 11 Reserved
2-21
2.3 Data Transmission and Reception Flow
As shown in Figure 2−1, McBSP Internal Block Diagram, on page 2-3, the re-
ceive operation is triple buffered and the transmit operation is double buffered.
Receive data arrives on DR and is shifted into RSR[1,2]. Once a full word (8-,
12-, 16-, 20-, 24-, or 32-bit) is received, RSR[1,2] is copied to the receive buffer
register, RBR[1,2], only if RBR[1,2] is not full. RBR[1,2] is then copied to
DRR[1,2], unless DRR[1,2] is not read by the CPU or DMA.
Transmit data is written by the CPU or DMA to DXR[1,2]. If there is no data in

XSR[1,2], the value in DXR[1,2] is copied to XSR[1,2]; otherwise, DXR[1,2] is
copied to XSR[1,2] when the last bit of data is shifted out from DX. After trans-
mit frame synchronization, XSR[1,2] begins shifting out the transmit data from
DX.
2.3.1 Resetting the Serial Port: (R/X)RST, and RESET

The serial port can be reset in the following two ways:
1) Device reset (RS = 0) places the receiver, transmitter and the sample rate
generator in reset. When the device reset is removed (RS = 1),
GRST = FRST = RRST = XRST = 0, keeping the entire serial port in the
reset state.
2) The serial port transmitter and receiver can be independently reset by the
RRST and XRST bits in the serial port control registers. The sample rate
generator is reset by the GRST bit in SPCR2.
Table 2−11, Reset State of McBSP Pins, on page 2-23 shows the state of
McBSP pins when the serial port is reset due to device reset and receiver/
transmitter reset (XRST = RRST = FRST = 0).
2-22
Table 2−11. Reset State of McBSP Pins
McBSP Device Reset
Pins Direction (RESET = 0) McBSP Reset
Receiver Reset (RRST = 0 and GRST = 1)
DR I Input Input
CLKR I/O/Z Input Known state if Input; CLKR running if output
FSR I/O/Z Input Known state if Input; FSRP inactive state if output
CLKS I/O/Z Input Input
Transmitter Reset (XRST = 0 and GRST = 1)
DX O Hi-Impedance Hi-Impedance
CLKX I/O/Z Input Known state if Input; CLKX running if output
FSX I/O/Z Input Known state if Input; FSXP inactive state if output
CLKS I Input Input
- Device reset or McBSP reset: When the McBSP is reset in any of the
above two ways, the state machine is reset to its initial state. This initial
state includes resetting all counters and status bits. The receive status bits
include RFULL, RRDY, and RSYNCERR. The transmit status bits include
XEMPTY, XRDY, and XSYNCERR.
- Device reset: When McBSP is reset due to device reset (device pin
RS = 0), the entire serial port including the transmitter, receiver, and the
sample rate generator is reset. All input-only pins and three-state pins
should be in a known state. The output-only pin, DX, is in the high-imped-
ance state. Since the sample rate generator is also reset (GRST = 0), the
sample rate generator clock, CLKG, is driven by the divide-by-2 CPU
clock, whereas the frame-sync signal, FSG, is not generated. For more in-
formation on sample rate generator reset, see section 2.5.1.2, Sample
Rate Generator Reset Procedure, on page 2-61. When the device is
pulled out of reset, the serial port remains in reset condition ([R/X]RST = 0
and FRST = 0) and in this condition the DR and DX pins may be used as
general purpose I/O as described in section 2.9, McBSP Pins as General
Purpose I/O, on page 2-97.
- McBSP reset: When the receive and transmitter reset bits, RRST and
XRST, are written with a zero, the respective portions of the McBSP are
reset and activity in the corresponding section of the serial port stops. All
input-only pins, such as DR and CLKS, and all other pins that are
2-23
configured as inputs, are in a known state. FS(R/X) is driven to its inactive
state (same as its polarity bit FS[R/X]) if it is an output. If CLK(R/X) is pro-
grammed as an output, it will be driven by CLKG, provided that GRST = 1.
Lastly, the DX pin will be in the high-impedance state when the transmitter
and/or the device is reset. During normal operation, the sample rate gen-
erator can be reset by writing a zero to GRST. GRST should be low only
when neither the transmitter nor the receiver is using the sample rate gen-
erator. In this case, the internal sample rate generator clock (CLKG) and
its frame-sync signal (FSG) are driven inactive low. When the sample rate
generator is not in the reset state (GRST = 1), pins FSR and FSX are in
an inactive state when RRST = 0 and XRST = 0, respectively, even if they
are outputs driven by FSG. This ensures that when only one portion of the
McBSP is in reset, the other portion can continue operation when
FRST = 1 and its frame sync is driven by FSG. For more information on
sample rate generator reset, see section 2.5.1.2, Sample Rate Generator
Reset Procedure, on page 2-61.
- Sample rate generator reset: As noted earlier, the sample rate generator
is reset when the device or its reset bit, GRST, is written with a zero. In the
case of device reset, the sample rate generator clock, CLKG, is driven by
a divide-by-2 CPU clock, whereas the frame-sync pulse, FSG, is driven
inactive low. If you want to reset the sample rate generator when neither
the transmitter nor the receiver is fed by CLKG and FSG, you can program
GRST in SRGR2 to zero. Here, CLKG and FSG are driven inactive-low.
When GRST = 1, CLKG comes up running as programmed in SRGR1.
Later, if FRST = 1, FSG is driven active-high after the programmed frame
period (FPER + 1) number of CLKG cycles has elapsed.
After device reset is complete (RS = 1), the serial port initialization procedure
is as follows:
1) Set XRST = RRST = FRST = 0 in SPCR[1,2]. If coming out of device re-

set, this step is not required.
2) Program only the McBSP configuration registers (and not the data regis-
ters) listed in Table 2−2, McBSP Registers, on page 2-5, as required
when the serial port is in reset state (XRST = RRST = FRST = 0).
3) Wait for two bit clocks. This is to ensure proper synchronization internally.
4) Set up data acquisition as required such as writing to DXR.
5) Set XRST = RRST= 1 to enable the serial port. Note that the value written
to SPCR[1,2] at this time should have only the reset bits changed to 1, and
the remaining bit-fields should have the same value as in step 2 above.
2-24
6) Set FRST = 1, if internally generated frame sync is required.
7) Wait two bit clocks for the receiver and transmitter to become active.
Alternatively, on either write (steps 1 and 5), the transmitter and receiver may
be placed in or taken out of reset individually by modifying the desired bit. Note
that the necessary duration of the active-low period of XRST or RRST is at
least two bit-clocks (CLKR/CLKX) wide.
The above procedure for reset initialization can be applied in general when the
receiver or transmitter has to be reset during its normal operation, and also
when the sample rate generator is not used for either operation.
Notes:
(a) The appropriate bit-fields in the serial port configuration registers,
SPCR[1,2], PCR, RCR[1,2], XCR[1,2], and SRGR[1,2], should only be modi-
fied by the user when the affected portion of the serial port is in reset.
(b) Data Transmit Register, DXR[1,2], should be loaded by the CPU or DMA
only when the transmitter is not in reset (XRST = 1). An exception to this rule
is during digital loop back mode described in section 2.4.1, Companding In-
ternal Data, on page 2-55.
(c) The multichannel selection registers, MCR, XCER[A/B], and RCER[A/B],
can be modified at any time as long as they are not being used by the current
block in the multichannel selection. See section 2.6.3.2 on page 2-82 for fur-
ther details in this case.
Example 2−1 shows values in the control registers that reset and configure the
transmitter while the receiver is running.
2-25
Example 2−1. Resetting and Configuring the Transmitter While Receiver is Running
SPCR1 = 0x0001 Transmitter reset, transmit interrupt (XINT to CPU) generated

SPCR2 = 0x0030 by XSYNCERR; receiver is running with RINT driven by
RRDY.
PCR = 0x0A00 FSX determined by FSGM in SRGR, transmit clock driven by

external source, receive clock continues to be driven by
sample rate generator.
SRGR1 = 0x0001 CPU clock drives the sample rate generator clock (CLKG)
SRGR2 = 0x2000 after a divide-by-2. A DXR[1,2]-to-XSR[1,2] copy generates
the transmit frame-sync signal.
XCR1 = 0x0840 Dual-phase frame; phase 1 has eight 16-bit words; phase 2
XCR2= 0x8421 has four 12-bit words, and 1-bit data delay
SPCR2 = 0x0031 Transmitter taken out of reset
2.3.2 Determining Ready Status

RRDY and XRDY indicate the ready state of the McBSP receiver and transmit-
ter, respectively. Serial port writes and reads may be synchronized by polling
RRDY and XRDY, or by using the events to DMA (REVT and XEVT in normal
mode, and REVTA and XEVTA in A-bis mode), or by interrupts to CPU (RINT
and XINT), which the events generate. Note that reading DRR[1,2] and writing
to DXR[1,2] affects RRDY and XRDY.
2.3.2.1 Receive Ready Status: REVT, RINT, and RRDY
RRDY = 1 indicates that the RBR[1,2] contents have been copied to DRR[1,2]
and that the data can be read by the CPU or DMA. Once that data has been
read by either the CPU or DMA, RRDY is cleared to 0. Also, at device reset
or serial port receiver reset (RRST = 0), RRDY is cleared to 0 to indicate no
data has yet been received and loaded into DRR[1,2]. RRDY directly drives
the McBSP receive event to the DMA (REVT). Also, the McBSP receive inter-
rupt (RINT) to the CPU may be driven by RRDY, if RINTM = 00b in SPCR1.
2.3.2.2 Transmit Ready Status: XEVT, XINT, and XRDY
XRDY = 1 indicates that the DXR[1,2] contents have been copied to XSR[1,2]
and that DXR[1,2] is ready to be loaded with a new data word. When the trans-
mitter transitions from reset to non-reset (XRST transitions from 0 to 1), XRDY
also transitions from 0 to 1 indicating that DXR[1,2] is ready for new data. Once
new data is loaded by the CPU or DMA, XRDY is cleared to 0. However, once
this data is copied from DXR[1,2] to XSR[1,2], XRDY transitions again from 0
to 1. Now once again, the CPU or DMA can write to DXR[1,2] although
2-26
XSR[1,2] has not been shifted out on DX yet. XRDY directly drives the transmit
synchronization event to the DMA (XEVT or XEVTA). In addition, the transmit
interrupt (XINT) to the CPU may also be driven by XRDY, if XINTM = 00b in
SPCR2.
2.3.3 CPU Interrupts: (R/X)INT

The receive interrupt (RINT) and transmit interrupt (XINT) signals the CPU of
changes to the serial port status. Four options exist for configuring these inter-
rupts. The options are set by the receive/transmit interrupt mode bit-field,
(R/X)INTM, in SPCR[1,2].
1) (R/X)INTM=00b. Interrupt on every serial word by tracking the (R/X)RDY

bits in SPCR[1,2]. Sections 2.3.2.1 and 2.3.2.2 describe the RRDY and
XRDY bits.
2) (R/X)INTM=01b. Interrupt after every 16-channel block boundary (in mul-

tichannel selection mode) has been crossed within a frame. In any other
serial transfer case, this setting is not applicable; and therefore, no inter-
rupts are generated. For details, see section 2.6.3.3, Update Interrupt, on
page 2-83.
3) (R/X)INTM=10b. Interrupt on detection of frame-synchronization pulses.

This generates an interrupt even when the transmitter/receiver is in reset.
This is done by synchronizing the incoming frame-sync pulse to the CPU
clock and sending it to the CPU via (R/X)INT. This is described in section
2.5.3.4, Frame Detection for Initialization, on page 2-68.
4) (R/X)INTM = 11b. Interrupt on frame-synchronization error. Note that if

any of the other interrupt modes are selected, (R/X)SYNCERR may be
read when servicing the interrupts to detect this condition. See sections
2.3.7.2 and 2.3.7.5 for more detail on synchronization error.
Note that the last three options listed above are applicable as interrupts to
the CPU, and not as events to the DMA.
2.3.4 Frame and Clock Configuration

Figure 2−9 shows the typical operation of McBSP clock and frame-sync sig-
nals. Serial clocks CLKR, and CLKX define the boundaries between bits for
receive and transmit, respectively. Similarly, frame-sync signals FSR and FSX
define the beginning of a serial word.
2-27
The McBSP allows configuration of various parameters for data frame syn-
chronization. This can be done independently for receive and transmit, which
includes the following items:
- Polarities of FSR, FSX, CLKX, and CLKR
- A choice of single- or dual-phase frames
- For each phase, the number of words per frame
- For each phase, the number of bits per word
- Subsequent frame synchronization may restart the serial data stream or

be ignored.
- The data bit delay from frame synchronization to first data bit can be 0-,
1-, or 2-bit delays.
- Right- or left-justification as well as sign-extension or zero-filling can be

chosen for receive data.
Figure 2−9. Frame and Clock Operation

Internal
CLK(R/X)
Á Á Á
Internal
Á Á Á
FS(R/X)
D(R/X) A1 A0 B7 B6 B5 B4 B3 B2 B1 B0
2.3.4.1 Frame and Clock Operation
Receive and transmit frame-sync pulses can be generated either internally by

the sample rate generator (see section 2.5.1, Sample Rate Generator Clock-
ing and Framing, on page 2-58) or driven by an external source. The source
of frame sync is selected by programming the mode bit, FS(R/X)M, in the PCR.
FSR is also affected by the GSYNC bit in SRGR2 (for details, see section
2.5.3.2, Receive Frame-Sync Selection: DLB, FSRM, GSYNC, on page 2-67).
Similarly, receive and transmit clocks can be selected to be inputs or outputs
by programming the mode bit, CLK(R/X)M, in the PCR.
When FSR and FSX are inputs (FSXM=FSRM=0, external frame-sync

pulses), the McBSP detects them on the internal falling edge of clock, internal
CLKR, and internal CLKX, respectively (see Figure 2−41, Clock and Frame
Generation, on page 2-57). The receive data arriving at the DR pin is also
sampled on the falling edge of internal CLKR. Note that these internal clock
signals are either derived from external source via CLK(R/X) pins or driven by
the sample rate generator clock (CLKG) internal to the McBSP.
2-28
When FSR and FSX are outputs, implying that they are driven by the sample
rate generator, they are generated (transition to their active state) on the rising
edge of internal clock, CLK(R/X). Similarly, data on the DX pin is output on the
rising edge of internal CLKX. See section 2.3.4.6, on page 2-34, for further de-
tails.
FSRP, FSXP, CLKRP, and CLKXP configure the polarities of FSR, FSX, CLKR,
and CLKX signals as shown in Table 2−6, Pin Control Register (PCR) Bit-Field
Description, on page 2-12. All frame-sync signals (internal FSR, internal FSX)
that are internal to the serial port are active high. If the serial port is configured
for external frame synchronization (FSR/FSX are inputs to McBSP), and
FSRP = FSXP = 1, the external active-low frame-sync signals are inverted
before being sent to the receiver (internal FSR) and transmitter (internal FSX).
Similarly, if internal synchronization (FSR/FSX are output pins and
GSYNC = 0) is selected, the internal active-high sync signals are inverted, if
the polarity bit FS(R/X)P = 1, before being sent to the FS(R/X) pin.
Figure 2−41, on page 2-57 shows this inversion using XOR gates.
On the transmit side, the transmit clock polarity bit, CLKXP, sets the edge used
to shift and clock out transmit data. Note that data is always transmitted on the
rising edge of internal CLKX. If CLKXP=1, and external clocking is selected
(CLKXM = 0 and CLKX is an input), the external falling-edge triggered input
clock on CLKX is inverted to a rising-edge triggered clock before being sent
to the transmitter. If CLKXP = 1, and internal clocking selected (CLKXM = 1
and CLKX is an output pin), the internal (rising-edge triggered) clock, internal
CLKX, is inverted before being sent out on the CLKX pin.
Similarly, the receiver can reliably sample data that is clocked with a rising
edge clock (by the transmitter). The receive clock polarity bit, CLKRP, sets the
edge used to sample received data. Note that the receive data is always
sampled on the falling edge of internal CLKR. Therefore, if CLKRP = 1 and ex-
ternal clocking is selected (CLKRM = 0 and CLKR is an input pin), the external
rising edge triggered input clock on CLKR is inverted to a falling-edge before
being sent to the receiver. If CLKRP = 1, and internal clocking is selected
(CLKRM = 1), the internal falling edge triggered clock is inverted to a rising
edge before being sent out on the CLKR pin.
Note that CLKRP = CLKXP in a system where the same clock (internal or ex-
ternal) is used to clock the receiver and transmitter. The receiver uses the op-
posite edge as the transmitter to ensure valid setup and hold of data around
this edge. Figure 2−10 shows how data, clocked by an external serial device
using a rising edge, may be sampled by the McBSP receiver with the falling
edge of the same clock.
2-29
Figure 2−10. Receive Data Clocking
Internal
CLKR
Data setup
ÁÁ
Data hold
DR
ÁÁ B7 B6
2.3.4.2 Frame-Synchronization Phases
Frame synchronization indicates the beginning of a transfer on the McBSP.

The data stream following frame synchronization may have two phases,
phase 1 and phase 2. The number of phases can be selected by the phase bit,
(R/X)PHASE, in RCR2 and XCR2. The number of words per frame and bits
per word can be independently selected for each phase via (R/X)FRLEN[1,2]
and (R/X)WDLEN[1,2] respectively. Figure 2−11 shows an example of a frame
where the first phase consists of 2 words of 12 bits each followed by a second
phase of three words of 8 bits each. Note that the entire bit stream in the frame
is contiguous. There are no gaps either between words or between phases.
Table 2−12 shows the bit-fields in the receive/transmit control register
(RCR[1,2]/XCR[1,2]) that control the number of words per frame and bits per
word for each phase, for both the receiver and transmitter. The maximum num-
ber of words per frame is 128 for a single-phase frame and 256 for a dual-
phase frame. The number of bits per word can be 8, 12, 16, 20, 24, or 32 bits.
Figure 2−11.Dual-Phase Frame Example

Phase 2
ÁÁÁ
ÁÁ ÁÁÁ
ÁÁ
ÁÁÁÁ
ÁÁÁ
ÁÁÁÁÁ
ÁÁÁ
ÁÁÁÁ
ÁÁÁ
ÁÁÁÁÁÁ
ÁÁÁ
ÁÁÁÁÁÁÁÁÁ
ÁÁÁÁÁÁÁÁÁ
ÁÁÁÁÁÁÁÁÁ
ÁÁ
Phase 2
Phase 1 word 1 Phase 2 word 3
Phase 1 word 2 word 1
word 2
CLK(R/X)
ÁÁÁÁÁ
ÁÁÁ
ÁÁ
ÁÁÁÁ
ÁÁÁ
ÁÁÁÁÁ
ÁÁÁ
ÁÁÁÁ
ÁÁÁ
ÁÁÁÁÁÁ
ÁÁÁ
ÁÁÁÁÁÁÁÁÁ
ÁÁÁÁÁÁÁÁÁ
ÁÁÁÁÁÁÁÁÁ
ÁÁ
Á Á
FS(R/X)
D(R/X)
2-30
Table 2−12. RCR[1,2]/XCR[1,2] Bit-Fields Controlling Words per Frame and Bits per
Word
RCR[1,2]/XCR[1,2] Bit-Field Control
Serial Port Frame
McBSP0/1 Phase Words per Frame Bits per Word
Receive 1 RFRLEN1 RWDLEN1
Receive 2 RFRLEN2 RWDLEN2
Transmit 1 XFRLEN1 XWDLEN1
Transmit 2 XFRLEN2 XWDLEN2
2.3.4.3 Frame Length: (R/X)FRLEN[1,2]

Frame length can be defined as the number of serial words (8-, 12-, 16-, 20-,
24-, or 32-bit) transferred per frame. The length corresponds to the number of
words or logical time slots or channels per frame-synchronization signal. The
7-bit (R/X)FRLEN[1,2] field in (R/X)CR[1,2] supports up to 128 words per
frame as shown in Table 2−13. (R/X)PHASE = 0 represents a single-phase
data frame and (R/X)PHASE = 1 selects a dual phase for the data stream.
Note that for a single-phase frame, FRLEN2 is a don’t care. The user is
cautioned to program the frame length fields with [w minus 1], where w repre-
sents the number of words per frame. For the example in Figure 2−11,
(R/X)FRLEN1 = 1 or 0000001b, and (R/X)FRLEN2 = 2 or 0000010b.
Table 2−13. McBSP Receive/Transmit Frame Length (1,2) Configuration

(R/X)
PHASE (R/X)FRLEN1 (R/X)FRLEN2 Frame Length
0 0 ≤ n ≤ 127 X Single-phase frame; (n+1) words per
frame
1 0 ≤ n ≤ 127 0 ≤ m ≤ 127 Dual-phase frame; (n+1) plus (m+1)

words per frame
2.3.4.4 Word Length: (R/X)WDLEN[1,2]

The 3-bit (R/X)WDLEN[1,2] fields in the receive/transmit control register deter-
mine the word length in bits-per-word for the receiver and transmitter for each
phase of the frame, as shown in Table 2−12. Table 2−14 shows how the value
of these fields selects particular word lengths in bits.
For the example in Figure 2−11 on page 2-30, (R/X)WDLEN1 = 001b, and
(R/X)WDLEN2 = 000b.
2-31
Notes:
(a) If (R/X)PHASE = 0 indicating a single-phase frame, (R/X)WDLEN2 is not
used by the McBSP, and its value is a don’t care.
(b) If the specified word length is larger than 16 bits, D(X/R)R2 must be writ-
ten or read before D(X/R)R1.
Table 2−14. McBSP Receive/Transmit Word Length Configuration
(R/X)WDLEN[1,2] McBSP Word Length (bits)

000 8
001 12
010 16
011 20
100 24
101 32
110 reserved
111 reserved
2.3.4.5 Data Packing Using Frame Length and Word Length
The frame length and word length can be manipulated to effectively pack data.
For example, consider a situation where four 8-bit words are transferred in a
single-phase frame as shown in Figure 2−12. In this case:
- (R/X)FRLEN1 = 0000011b, 4-word frame
- (R/X)PHASE = 0, single-phase frame
- (R/X)FRLEN2 = X
- (R/X)WDLEN1 = 000b, 8-bit word
In this case, four 8-bit data elements are transferred to and from the McBSP
by the CPU or DMA. Thus, four reads from DRR1 and four writes to DXR1 are
necessary for each frame.
2-32
Figure 2−12. Single-Phase Frame of Four 8-Bit Words
ÁÁÁ
ÁÁÁÁÁ
ÁÁÁÁÁÁ
ÁÁÁÁ
ÁÁÁÁÁÁ
ÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁ
ÁÁÁÁÁ
ÁÁÁÁÁÁÁÁÁÁ
ÁÁÁ
Word 1 Word 3
Word 2 Word 4
CLKR
ÁÁÁ
ÁÁÁÁÁ
ÁÁÁÁÁÁ
ÁÁÁÁ
ÁÁÁÁÁÁ
ÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁ
ÁÁÁÁÁ
ÁÁÁ
Á Á
FSR
DR
Á Á
ÁÁÁÁ
ÁÁÁÁÁ
RSR1 to RBR1 copy
ÁÁÁÁÁÁÁÁÁÁÁÁ
ÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁ
RSR1 to RBR1 copy
ÁÁÁ
RSR1 to RBR1 copy RSR1 to RBR1 copy
CLKX
FSX
ÁÁÁÁ
ÁÁÁÁÁ
ÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁ
ÁÁÁ
DX Á
DXR1 to XSR1 copy DXR1 to XSR1 copy
Á
The example in Figure 2−12 can also be treated as a data stream of a single-
phase frame consisting of one 32-bit data word, as shown in Figure 2−13. In
this case:
- (R/X)FRLEN1 = 0b, 1-word frame
- (R/X)FRLEN2 = X
In this case, two 16-bit data words are transferred to and from the McBSP by
the CPU or DMA. Thus, two reads from DRR2 and DRR1 and two writes to
DXR2 and DXR1 are necessary for each frame. This results in only one-half
the number of transfers compared to the previous case. This manipulation re-
duces the percentage of bus time required for serial port data movement.
Note:
In this case, D(X/R)R2 must be written or read before D(X/R)R1.
2-33
Figure 2−13. Single-Phase Frame of One 32-Bit Word
ÁÁÁÁÁ
ÁÁÁÁÁÁÁÁ
ÁÁÁÁ
ÁÁÁÁÁ
ÁÁÁÁÁÁ
ÁÁÁÁÁÁÁÁÁ
Word 1
ÁÁÁÁÁ
CLKR
ÁÁÁÁÁÁÁÁ
ÁÁÁÁ
ÁÁÁÁÁ
ÁÁÁÁÁÁ
ÁÁÁÁÁÁÁÁÁ
Á ÁÁ
FSR
DR
ÁÁÁ
ÁÁÁÁÁÁÁÁÁ
ÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁ ÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁ
RBR1 to DRR1 copy RBR2 to DRR2 copy
CLKX
FSX
Á ÁÁ
DX
Á DXR1 to XSR1 copy
ÁÁ DXR2 to XSR2 copy
2.3.4.6 Data Delay: (R/X)DATDLY
The start of a frame is defined by the first clock cycle in which frame synchro-
nization is found to be active. The beginning of actual data reception or trans-
mission with respect to the start of the frame can be delayed if required. This
delay is called data delay. RDATDLY and XDATDLY specify the data delay for
reception and transmission, respectively. The range of programmable data
delay is zero to two bit-clocks ([R/X]DATDLY = 00b −10b), as described in
Table 2−7, Receive Control Register 1 (RCR1) Bit-Field Description, on page
2-16, and Table 2−8, Receive Control Register 2 (RCR2) Bit-Field Description,
on page 2-17, and shown in Figure 2−14, Data Delay. Typically a 1-bit delay
is selected, since data often follows a one-cycle active frame-sync pulse.
Figure 2−14. Data Delay
CLK(R/X)
ÁÁ
FS(R/X)
ÁÁ
0−Bit Period
D(R/X) B7 B6 B5 B4 B3
Á
Data Delay 0
1−Bit Period
D(R/X)
Á B7 B6 B5 B4
Á
Data Delay 1
2−Bit Periods
D(R/X)
Data Delay 2
Á B7 B6 B5
2-34
Normally, frame-sync pulse is detected or sampled with respect to an edge of
serial clock internal CLK(R/X) (see section 2.3.4.1, Frame and Clock Opera-
tion, on page 2-28). Thus, on the following cycle or later (depending on data
delay value), data may be received or transmitted. However, in the case of
zero-bit data delay, the data must be ready for reception and/or transmission
on the same serial clock cycle. For reception, this problem is solved, since re-
ceive data is sampled on the first falling edge of CLKR where an active-high
internal FSR is detected. However, data transmission must begin on the rising
edge of the internal CLKX clock that generated the frame synchronization.
Therefore, the first data bit is assumed to be present in XSR1, and thus DX.
The transmitter then asynchronously detects the frame synchronization, FSX,
going active high, and immediately starts driving the first bit to be transmitted
on the DX pin.
Another common mode is a data delay of two. This configuration allows the
serial port to interface to different types of T1 framing devices where the data
stream is preceded by a framing bit. During reception of such a stream with
data delay of two bits (framing bit appears after one-bit delay and data appears
after 2-bit delay), the serial port essentially discards the framing bit from the
data stream as shown in Figure 2−15. In transmission, by delaying the first
transfer bit, the serial port essentially inserts a blank period (high-impedance
period) in place of the framing bit. Here, it is expected that the framing device
inserts its own framing bit or that the framing bit is generated by another de-
vice. Alternatively, you can pull up or pull down DX to achieve the desired
value.
Figure 2−15. Two-Bit Data Delay Used to Discard Framing Bit
CLKR
FSR
Á
2-bit periods
DR
2.3.4.7
Á
Multi-Phase Frame Example: AC97
Framing bit B7 B6 B5
Figure 2−16 shows an example of the Audio Codec ‘97 (AC97) standard which
uses the dual-phase frame feature. The first phase consists of a single 16-bit
word. The second phase consists of twelve 20-bit words. The phases are con-
figured as follows:
- (R/X)PHASE = 1b, dual-phase frame
- (R/X)FRLEN1 = 0b, 1 word per frame in phase 1
2-35
- (R/X)WDLEN1 = 010b, 16 bits per word in phase 1
- (R/X)FRLEN2 = 0001011b, 12 words per frame in phase 2
- (R/X)WDLEN2 = 011b, 20 bits per word in phase 2
- CLK(R/X)P = 0, receive data sampled on falling edge of internal CLKR;
transmit data clocked on rising edge of internal CLKX.
- FS(R/X)P = 0, active-high frame-sync signals
- (R/X)DATDLY = 01b, data delay of one bit-clock
Figure 2−16. AC97 Dual-Phase Frame Format

P1W1 P2W1 P2W2 P2W3 P2W4 P2W5 P2W6 P2W7 P2W8 P2W9 P2W10 P2W11 P2W12
FS(R/X)
1-bit data delay

16 bits
Á
20 bits
D(R/X)
Á PxWy = Phase x Word y
Figure 2−16 shows the timing of AC97 near frame synchronization. First, no-
tice that the frame-sync pulse itself overlaps the first word. In McBSP opera-
tion, the inactive to active transition of the frame-synchronization signal actual-
ly indicates frame synchronization. For this reason, frame synchronization
may be high for an arbitrary number of bit-clocks. Only after the frame synchro-
nization is recognized to have gone inactive, and then active again, is the next
frame synchronization recognized.
Also, notice that there is a one-bit data delay in Figure 2−17. Regardless of the
data delay, transmission can occur without gaps. The last bit of the previous
(last) word in phase 2 is immediately followed by the first data bit of the first
word in phase 1 of the next data frame.
Figure 2−17. AC97 Bit Timing Near Frame-Synchronization Example
CLKR
FSR
ÁÁ
1-bit data delay
ÁÁ
DR P2W12B1 P2W12B0 P1W1B15 P1W1B14 P1W1B13 P1W1B12
PxWyBz = Phase x Word y Bit z
2-36
2.3.4.8 Delay Enable/Disable on the DX Pin
Figure 2−18 and Figure 2−19 show the timing of the DX pin when DXENA bit
is set to 1 to enable extra delay for turn-on time. This bit controls the high-im-
pedance (hi-Z) enable on the DX pin, not the data itself; so only the first bit will
be delayed in the normal mode. In the A-bis mode, any bit can be delayed since
any bit can go from hi-Z to valid. This bit should be set to avoid conflict when
tying the DX pins together.
Figure 2−18. DX Enabler in Normal Mode

CLKX
FSX
DX
te
Note: te = extra delay for turn on time with DXENA = 1.
Figure 2−19. DX Enabler in A-bis mode

CLKX
FSX
DX
te te te
Note: te = extra delay for turn on time with DXENA = 1.
2-37
2.3.5 McBSP Standard Operation
During a serial transfer, there are typically periods of serial port inactivity be-
tween packets or transfers. The receive and transmit frame-synchronization
pulse occurs for every serial transfer. When the McBSP is not in reset state and
has been configured for the desired operation, a serial transfer can be initiated
by programming (R/X)PHASE = 0, for a single-phase frame with required
number of words programmed in (R/X)FRLEN1. The number of words can
range from 1 to 128 ([R/X]FRLEN1 = 0x0 to 0x7F). The required serial word
length is set in the (R/X)WDLEN1 field in (R/X)CR1. If dual-phase is required
for the transfer, RPHASE = 1, (R/X)FRLEN[1,2] can be set to any value be-
tween 0x0 to 0x7F, which represents 1 to 128 words.
Figure 2−20 shows an example of a single-phase data frame comprising one

8-bit word. Since the transfer is configured for one data bit delay, the data on
the DX and DR pins are available one bit-clock after FS(R/X) goes active. This
figure, as well as all others in this section, make the following assumptions:
- (R/X)FRLEN1 = 0b, 1 word per frame
- (R/X)FRLEN2 = X, (R/X)WDLEN2 = X, don’t care
- CLK(X/R)P = 0, receive data clocked on falling edge; transmit data

clocked on rising edge
- FS(R/X)P = 0, active-high frame-sync signals
- (R/X)DATDLY = 01b, one-bit data delay
Figure 2−20. McBSP Standard Operation
CLK(R/X)
Á Á Á ÁÁ
FS(R/X)
D(R/X) A1 A0 B7 B6 B5 B4 B3 B2 B1 B0 C7 C6 C5
2.3.5.1 Receive Operation
Figure 2−21 shows an example of serial reception. Once receive frame-syn-

chronization (FSR) transitions to its active state, it is detected on the first falling
edge of CLKR of the receiver. The data on the DR pin is then shifted into the
receive shift register (RSR[1,2]) after the appropriate data delay as set by
RDATDLY. The contents of RSR[1,2] is copied to RBR[1,2] at the end of every
2-38
word on the rising edge of clock, provided RBR[1,2] is not full with the previous
data. Then, an RBR[1,2]-to-DRR[1,2] copy activates the RRDY status bit to 1
on the following falling edge of CLKR. This indicates that the receive data reg-
ister (DRR[1,2]) is ready with the data to be read by the CPU or DMA. RRDY
is deactivated when DRR[1,2] is read by the CPU or DMA.
Figure 2−21. Receive Operation
CLKR
Á ÁÁ Á Á
FSR
DR A1 A0 B7 B6 B5 B4 B3 B2 B1 B0 C7 C6 C5
RRDY
RBR1 to DRR1 copy(A) Read from DRR1(A) RBR1 to DRR1 copy(B) Read from DRR1(b)
2.3.5.2 Transmit Operation
Once transmit frame synchronization occurs, the value in the transmit shift
register, XSR[1,2], is shifted out and driven on the DX pin after the appropriate
data delay as set by XDATDLY. XRDY is activated on every DXR[1,2]-to-
XSR[1,2] copy on the following falling edge of CLKX, indicating that the data
transmit register (DXR[1,2]) is written with the next data to be transmitted.
XRDY is deactivated when DXR[1,2] is written by the CPU or DMA.
Figure 2−22 shows an example of a serial transmission. See section 2.3.7.4,
Transmit Empty: XEMPTY, on page 2-48, for transmit operation when trans-
mitter is pulled out of reset (XRST = 1).
Figure 2−22. Transmit Operation
CLKX
FSX
Á ÁÁ Á Á
DX A1
XRDY
A0
Á ÁÁ B7 B6 B5 B4 B3 B2 B1
Á
B0
Á C7 C6 C5
DXR1 to XSR1 copy(B) Write to DXR1(C) DXR1 to XSR1 copy(C) Write to DXR1
2.3.5.3 Maximum Frame Frequency
The frame frequency is determined by the period between frame-synchroniza-

tion signals:
Bit−Clock Frequency
Frame Frequency +
Number of Bit−Clocks Between Frame Sync Signals
2-39
The frame frequency may be increased by decreasing the time between
frame-synchronization signals in bit clocks (limited only by the number of bits
per frame). As the frame transmit frequency is increased, the inactivity period
between the data packets for adjacent transfers decreases to zero. The mini-
mum time between frame synchronization is the number of bits transferred per
frame. The maximum frame frequency is defined as follows:
Bit−Clock Frequency
Maximum Frame Frequency +
Number of Bits Per Frame
Figure 2−23 shows the McBSP operating at maximum packet frequency. At

maximum packet frequency, the data bits in consecutive packets are trans-
mitted contiguously with no inactivity between bits. If there is a one-bit data
delay as shown, the frame-synchronization pulse overlaps the last bit trans-
mitted in the previous frame.
Figure 2−23. Maximum Frame Frequency Receive/Transmit (R/X)DATDLY = 0)
CLK(R/X)
FS(R/X)
D(R/X) A2 A1 A0 B7 B6 B5 B4 B3 B2 B1 B0 C7 C6
Effectively, this permits a continuous stream of data; and thus, the frame-syn-
chronization pulses are essentially redundant. Theoretically, only an initial
frame-synchronization pulse is required to initiate a multipacket transfer. The
McBSP supports operation of the serial port in this fashion by ignoring the
successive frame-sync pulses. Data is clocked in to the receiver, or clocked
out of the transmitter, on every clock{. The frame ignore bit, (R/X)FIG, in
(R/X)CR can be programmed to ignore the successive frame-sync pulses until
the desired frame length or number of words is reached. This is explained in
section 2.3.6.1, Data Packing using Frame-Sync Ignore Bits, on page 2-41 .
2.3.6 Frame-Synchronization Ignore

The McBSP can be configured to ignore transmit and receive frame-synchro-
nization pulses. The (R/X)FIG bit in (R/X)CR2 can be programmed to zero to
recognize frame-sync pulses, or set to one to ignore frame-sync pulses. The
user can use (R/X)FIG bit to either pack data or ignore unexpected frame-sync
pulses. Section 2.3.6.1 explains data packing and McBSP operation on unex-
pected frame-sync pulses.
† For (R/X)DATDLY=0, the first bit of data transmitted is asynchronous to internal CLKX.
2-40
2.3.6.1 Data Packing Using Frame-Sync Ignore Bits
Section 2.3.4.5, Data Packing using Frame Length and Word Length, on page
2-32, describes one method of changing the word length and frame length to
simulate 32-bit serial word transfers, thus requiring much less bus bandwidth.
This example works when there are multiple words per frame. Now consider
the case of the McBSP operating at maximum packet frequency as shown in
Figure 2−24. Here, each frame only has a single 8-bit word. This stream takes
one read and one write transfer for each 8-bit word. Figure 2−25 shows the
McBSP configured to treat this stream as a continuous 32-bit word. In this ex-
ample, (R/X)FIG is set to 1 to ignore subsequent frames after the first. Only
two read- or two write-transfers are needed every 32 bits. This configuration
effectively reduces the required bus bandwidth to one-half of the bandwidth
needed to transfer four 8-bit words.
Figure 2−24. Maximum Packet Frequency Operation With 8-bit Data
Word 1 Word 3
ÁÁÁÁÁ
ÁÁ ÁÁÁÁÁÁ
ÁÁÁÁ
ÁÁÁÁÁÁ
ÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁ
ÁÁÁ
Word 2 Word 4
ÁÁÁÁÁÁ
Á ÁÁÁÁÁÁ
ÁÁÁÁ
ÁÁÁÁÁÁ
ÁÁÁ
CLKR
FSR
Á Á
DR
Á Á
ÁÁÁÁÁÁ
Á ÁÁÁÁÁÁ
ÁÁÁÁÁ
ÁÁÁÁÁ
RBR1 to DRR1 copy
RBR1 to DRR1 copy
ÁÁÁ
CLKX
ÁÁÁÁÁÁ
Á ÁÁÁÁÁÁ
ÁÁÁÁÁ
ÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁ
ÁÁÁ
Á Á
FSX
DX
ÁDXR1 to XSR1 copy DXR1 to XSR1 copy Á
2-41
Figure 2−25. Data Packing at Maximum Packet Frequency With (R/X)FIG=1
Word 1
ÁÁÁÁÁÁ
ÁÁÁÁÁ
ÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁ
ÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁ
ÁÁÁÁÁÁ
ÁÁÁÁÁ
ÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁ
ÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁ
CLKR
Á Á
FSR Frame ignored Frame ignored Frame ignored
DR
Á Á
ÁÁÁÁÁ
ÁÁÁÁÁÁ
ÁÁÁÁÁÁÁÁÁ
ÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁ
ÁÁÁÁÁ
ÁÁÁÁÁÁ
ÁÁÁÁÁÁÁÁÁ
ÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁ
CLKX
FSX Frame ignored Frame ignored Frame ignored
Á Á
Á Á
DX
2.3.6.2 Frame-Sync Ignore and Unexpected Frame-Sync Pulses
The previous section explained how frame ignore bits can be used to pack data
and efficiently use the bus bandwidth. (R/X)FIG bit can also be used to ignore
unexpected frame-sync pulses. Thus, any frame-sync pulse that occurs one
bit-clock earlier than the programmed data delay ([R/X]DATDLY) is considered
unexpected. Setting the frame ignore bits to one causes the serial port to ig-
nore these unexpected frame-sync signals.
In reception, if not ignored (RFIG = 0), an unexpected FSR pulse will discard
the contents of RSR[1,2] in favor of the new incoming data. Therefore, if
RFIG = 0, an unexpected frame-synchronization pulse aborts the current data
transfer, sets RSYNCERR in SPCR1 to 1, and begins the transfer of a new
data word. For further details, see section 2.3.7.2, Unexpected Receive Frame
Synchronization: RSYNCERR, on page 2-46. When RFIG = 1, reception con-
tinues, ignoring the unexpected frame-sync pulses.
In transmission (if not ignored [XFIG = 0]), an unexpected FSX pulse aborts
the present transmission, sets XSYNCERR to 1 in SPCR2 , and re-initiates
transmission of the current word that was aborted. For further details, see sec-
tion 2.3.7.5, Unexpected Transmit Frame Synchronization: XSYNCERR, on
page 2-50. When XFIG = 1, normal transmission continues with unexpected
frame-sync signals ignored.
2-42
Figure 2−26. Unexpected Frame Synchronization With (R/X)FIG=0
CLK(R/X)
Frame sync aborts current transfer
ÁÁ
FS(R/X)
New data received
ÁÁ
DR A0 B7 B6 C7 C6 C5 C4 C3 C2 C1 C0 D7 D6
Current data re-transmitted
(R/X)SYNCERR
DX A0 B7 B6 B7 B6 B5 B4 B3 B2 B1 B0 C7
ÁÁ
C6
Figure 2−26 shows an example wherein word B is interrupted by an unex-

pected frame-sync pulse when (R/X)FIG = 0. In the case of reception, the re-
ception of B is aborted (B is lost), and a new data word (C in this example) is
received after the appropriate data delay. This condition is a receive synchro-
nization error, and thus sets the RSYNCERR bit. However, for transmission,
the transmission of B is aborted, and the same data (B) is retransmitted after
the appropriate data delay. This condition is a transmit synchronization error
and thus sets the XSYNCERR bit. Synchronization errors are discussed in
sections 2.3.7.2 and 2.3.7.5. In contrast, Figure 2−27 shows McBSP operation
when unexpected frame-synchronization signals are ignored by setting
(R/X)FIG = 1. Here, the transfer of word B is not affected by an unexpected
frame synchronization.
Figure 2−27. Unexpected Frame Synchronization With (R/X)FIG = 1
CLK(R/X)
Frame synchronization ignored
ÁÁ
FS(R/X)
D(R/X)
(R/X)SYNCERR
A0 B7 B6 B5 B4 B3 B2 B1 B0 C7 C6 C5
ÁÁ
C4
2.3.7 Serial Port Exception Conditions

There are five serial port events that may constitute a system error:
1) Receive Overrun (RFULL = 1). This occurs when DRR[1,2] has not been
read since the last RBR[1,2]-to-DRR[1,2] copy. Consequently, a new word
in RBR[1,2] will not be transferred to DRR[1,2], and RSR[1,2] is now full
with another new word shifted in from DR. Therefore, RFULL indicates an
error condition wherein any new data that may arrive at this time on DR
will replace the contents in RSR[1,2], and thus, the previous word is lost.
RSR[1,2] continues to be overwritten as long as new data arrives on DR
and DRR[1,2] is not read.
2-43
2) Unexpected Receive Frame Synchronization (RSYNCERR=1). This
can occur during reception when RFIG = 0 and an unexpected frame-
sync pulse occurs. An unexpected frame-sync pulse is defined as that
which occurs RDATDLY minus 1 bit-clock earlier than the first bit of the
next associated word. This causes the current data reception to abort and
restart. If new data has been copied into RBR[1,2] from RSR[1,2] since the
last RBR[1,2]-to-DRR[1,2] copy, this new data in RBR[1,2] will be lost. This
is because no RBR[1,2]-to-DRR[1,2] copy occurs as the reception has
been restarted.
3) Transmit Data Overwrite. Here the user overwrites data in DXR[1,2] be-
fore it is copied to XSR[1,2]. The data previously in DXR[1,2] is never
transferred on DX, since it never got copied to XSR[1,2].
4) Transmit Empty (XEMPTY = 0). If a new frame-synchronization signal

arrives before new data is loaded into DXR[1,2], the old data in DXR[1,2]
will be sent again. This will continue for every new frame-sync signal that
arrives on the FSX pin until DXR[1,2] is loaded with new data.
5) Unexpected Transmit Frame Synchronization (XSYNCERR = 1). This

can occur during transmission when XFIG = 0 and an unexpected frame-
sync pulse occurs. Again, an unexpected frame-sync pulse is defined as
that which occurs XDATDLY minus 1 bit-clock earlier than the first bit of
the next associated word. This causes the current data transmission to
abort and restart the current transfer. If new data had been written to
DXR[1,2] since the last DXR[1,2]-to-XSR[1,2] copy, the current value in
XSR[1,2] will be lost.
These events are described in more detail in the sections that follow.
2.3.7.1 Reception With Overrun: RFULL
RFULL = 1 in SPCR1 indicates that the receiver has experienced overrun and
is in an error condition. RFULL is set when all of the following conditions are
met:
1) DRR[1,2] has not been read since the last RBR[1,2]-to-DRR[1,2] transfer
(RRDY = 1).
2) RBR[1,2] is full and an RBR[1,2]-to-DRR[1,2] copy has not occurred.
3) RSR[1,2] is full and an RSR[1,2]-to-RBR[1,2] transfer has not occurred.
2-44
Data arriving on DR is continuously shifted into RSR[1,2]. Once a complete
word is shifted into RSR[1,2], an RSR[1,2]-to-RBR[1,2] transfer can occur only
if an RBR[1,2]-to-DRR[1,2] copy is complete. Therefore, if DRR[1,2] has not
been read by the CPU or DMA since the last RBR[1,2]-to-DRR[1,2] transfer
(RRDY = 1), an RBR[1,2]-to-DRR[1,2] copy will not take place until RRDY = 0.
At this time, new data arriving on the DR pin is shifted into RSR[1,2] and the
previous contents of RSR[1,2] is lost. This data loss occurs because comple-
tion of a serial-word reception triggers an RBR[1,2]-to-DRR[1,2] transfer only
when RRDY = 0. Note that after the receive portion starts running from reset,
a minimum of three words must be received before RFULL is set.
The data loss of the contents in RSR[1,2] can be avoided if DRR[1,2] is read
not later than two and one-half cycles before the end of the third word in
RSR[1,2].
Either of the following events clears the RFULL bit to 0 and allows subsequent
transfers to be read properly:
- Reading DRR[1,2]
- Resetting the receiver (RRST = 0) or the device
Another frame synchronization is required to restart the receiver.
Figure 2−28 shows the receive overrun condition. Because serial word A is not
read before the reception of serial word B is complete, B is not transferred to
DRR1 yet. Another new word C arrives and RSR1 is full with this data. DRR1
is finally read, but not earlier than two and one-half cycles before the end of
word C. Therefore, new data D overwrites the previous word C in RSR1. If
RFULL is still set after arrival of D, the next word can overwrite it, if DRR is not
read in time.
Figure 2−28. Serial Port Receive Overrun
CLKR
ÁÁ ÁÁÁ ÁÁÁ
FSR
DR A1
RRDY
ÁÁ
A0 B7 B6 B5 B4 B3 B2 B1 B0
RBR1 to DRR1 copy(A)

ÁÁÁ C7 C6 C5 C4 C3 C2 C1 C0
ÁÁÁD7
No RSR1 to RBR1 copy(C)
No read from DRR1(A)

No read from DRR1(A) No RBR1 to DRR1 copy(B)
RFULL
2-45
Figure 2−29 shows the case where RFULL is set, but the overrun condition is
averted by reading the contents of DRR1 at least two and one-half cycles be-
fore the next serial word C is completely shifted into RSR1. This ensures that
an RBR1-to-DRR1 copy of data B occurs before the next serial word (C) is
transferred from RSR1 to RBR1.
Figure 2−29. Serial Port Receive Overrun Avoided
CLKR
Á Á ÁÁ Á
FSR
DR A1 A0 B7 B6 B5 B4 B3 B2 B1 B0 C7 C6 C5 C4 C3 C2 C1 C0
RBR1 to DRR1(B)
RRDY RBR1 to DRR1 copy(A)
No read from DRR1(A) No RBR1 to DRR1 copy(B) Read from DRR1(A)
RFULL
2.3.7.2 Unexpected Receive Frame Synchronization: RSYNCERR
Figure 2−30 shows the decision tree that the receiver uses to handle all incom-
ing frame-synchronization pulses. The diagram assumes that the receiver has
been started, RRST = 1. Unexpected frame-sync pulses can originate from an
external source or from the internal sample rate generator. An unexpected
frame-sync pulse is defined as a sync pulse which occurs RDATDY bit-clocks
earlier than the last transmitted bit of the previous frame. Any one of four cases
can occur:
- Case 1: Unexpected internal FSR pulses with RFIG = 1. This case is dis-
cussed in section 2.3.6.2 on page 2-42 and shown in Figure 2−27 on page
2-43. In Case 1, receive frame-sync pulses are ignored and the reception
continues.
- Case 2: Normal serial port reception. There are three possible reasons
why a receive might NOT be in progress:
J The FSR is the first after RRST = 1.
J The FSR is the first after DRR[1,2] is read clearing an RFULL condi-
tion.
J The serial port is in the interpacket intervals. The programmed data
delay (RDATDLY) for reception may start during these interpacket in-
tervals for the first bit of the next word to be received. Thus, at maxi-
mum frame frequency, frame synchronization can still be received
RDATDLY bit-clocks before the first bit of the associated word.
For Case 2, reception continues normally since these are not unexpected
frame-sync pulses.
2-46
- Case 3: Unexpected receive frame synchronization with RFIG = 0 (unex-
pected frame not ignored). This case was shown in Figure 2−26 on page
2-43 for maximum frame frequency. Figure 2−31 on page 2-48 shows this
case during normal operation of the serial port, with time intervals between
packets. An unexpected frame-sync pulse is detected when it occurs at
or before RDATDLY minus 1 bit-clock before the last bit of the previous
word is received on the DR pin. In both cases, the RSYNCERR bit in
SPCR1 is set. RSYNCERR{ can be cleared only by a receiver reset or by
the user writing a 0 to this bit in SPCR1. You should note that if
RINTM = 11b in SPCR1, RSYNCERR drives the receive interrupt (RINT)
to the CPU.
ÁÁÁÁÁÁÁÁ
Figure 2−30. Response to Receive Frame-Synchronization Pulse
ÁÁÁÁÁÁÁÁ
Receive frame-sync
ÁÁÁÁÁÁÁÁ pulse occurs
ÁÁÁÁÁÁÁÁ
ÁÁÁÁÁÁÁÁ ÁÁÁÁÁÁÁÁÁ
ÁÁÁÁÁÁÁÁ ÁÁÁÁÁÁÁÁÁ Case 2:
Unexpected No
Normal Reception
frame-sync
Start receiving data
pulse
ÁÁÁÁÁÁÁÁ
?
Yes
ÁÁÁÁÁÁÁÁÁ
ÁÁÁÁÁÁÁÁÁ Case 3:
ÁÁÁÁÁÁÁÁÁ
Without Frame Ignore
ÁÁÁÁÁÁÁÁÁ
No Abort reception
RFIG=1 SET RSYNCERR
ÁÁÁÁÁÁÁÁÁ
? Start next reception
immediately
Yes ÁÁÁÁÁÁÁÁÁPrevious word is lost
ÁÁÁÁÁÁÁÁ
ÁÁÁÁÁÁÁÁ Case 1:
ÁÁÁÁÁÁÁÁ
With Frame Ignore
Ignore frame pulse
ÁÁÁÁÁÁÁÁ
Receiver continues
running
† The RSYNCERR bit in SPCR1 is a read/write bit. Therefore, writing a 1 to it sets the error condi-
tion. Typically, writing a 0 is expected.
2-47
Figure 2−31. Unexpected Receive Synchronization Pulse
CLKR
Unexpected frame synchronization
Á Á ÁÁ
FSR
DR A1 A0 B7 B6 B5 B4 C7 C6 C5 C4 C3 C2 C1 C0
RBR1 to DRR1(B)
RRDY
RBR1 to DRR1 copy(A) Read from DRR1(A) RBR1 to DRR1 copy(C) Read from DRR1(C)
RSYNCERR
2.3.7.3 Transmit with Data Overwrite

Figure 2−32 depicts what happens if the data in DXR1 is overwritten before
being transmitted. Initially, the programmer loaded DXR1 with data C. A sub-
sequent write to DXR1 overwrites C with D before it is copied to XSR1. Thus, C
is never transmitted on DX. The CPU can avoid data overwrite by polling
XRDY before writing to DXR1 or by waiting for an XINT programmed to be trig-
gered by XRDY (XINTM = 00b). The DMA can avoid overwriting by synchro-
nizing data transfers with XEVT.
Figure 2−32. Transmit With Data Overwrite
CLKX
Á Á ÁÁ ÁÁ
FSX
DX A1
XRDY
A0
Á ÁB7 B6 B5 B4 B3 B2 B1
ÁÁ
B0
ÁÁD7 D6 D5
Write to DXR1(C) Write to DXR1(D) DXR1 to XSR1 copy(D) Write to DXR1(E)
2.3.7.4 Transmit Empty: XEMPTY
XEMPTY indicates when the transmitter has experienced underflow. Any of

the following conditions causes XEMPTY to become active (XEMPTY = 0):
1) During transmission. DXR[1,2] has not been loaded since the last
DXR[1,2]-to-XSR[1,2] copy, and all bits of the data word in XSR[1,2] have
been shifted out on DX.
2) The transmitter is reset (XRST = 0, or device is reset) and then restarted.
During an underflow condition, the transmitter continues to transmit the old

data in DXR[1,2] for every new frame-sync signal that arrives on FSX until a
new word is loaded into DXR[1,2] by the CPU or DMA. XEMPTY is deactivated
2-48
(XEMPTY = 1) when the new word in DXR[1,2] is transferred to XSR[1,2]. In
the case of an internal frame generation, the transmitter regenerates a single
internal FSX initiated by a DXR[1,2]-to-XSR[1,2] copy (FSXM = 1 in the PCR
and FSGM=0 in SRGR2). Otherwise, the transmitter waits for the next frame
synchronization.
When the transmitter is taken out of reset (XRST = 1), it is in a transmit ready
(XRDY=1) and transmit empty (XEMPTY = 0) condition. If DXR[1,2] is loaded
by the CPU or DMA before internal FSX goes active high, a valid DXR[1,2]-to-
XSR[1,2] transfer occurs. This allows for the first word of the first frame to be
valid even before the transmit frame-sync pulse is generated or detected. Al-
ternatively, if a transmit frame sync is detected before DXR is loaded, zeros
will be output on DX.
Figure 2−33 depicts a transmit underflow condition. After B is transmitted, the

programmer fails to reload DXR[1,2] before the subsequent frame synchro-
nization. Thus, B is again transmitted on DX. Figure 2−34 shows the case of
writing to DXR1 just before a transmit underflow condition that would other-
wise occur. After B is transmitted, C is written to DXR1 before the next transmit
frame-sync pulse occurs so that C is successfully transmitted on DX, thus
averting a transmit empty condition.
Figure 2−33. Transmit Empty
CLKX
Á Á Á Á
FSX
DX A1
XRDY
Á
A0
ÁB7 B6
DXR1 to XSR1 copy(B)

B5 B4 B3 B2 B1
Á
B0
Á B7 B6 B5
Write to DXR1(C)
XEMPTY
2-49
Figure 2−34. Transmit Empty Avoided
CLKX
ÁÁ Á ÁÁ ÁÁ
FSX
DX A1
XRDY
ÁÁ
A0
Á B7
DXR1 to XSR1 copy

B6 B5 B4 B3 B2 B1
ÁÁ
B0
Write to DXR1(C)
ÁÁ C7 C6 C5
DXR1 to XSR1 copy(C)
XEMPTY
2.3.7.5 Unexpected Transmit Frame Synchronization: XSYNCERR
Figure 2−30 on page 2-47 shows the decision tree that the transmitter uses
to handle all incoming frame-synchronization signals. The diagram assumes
that the transmitter has been started (XRST = 1). An unexpected transmit
frame-sync pulse is defined as a sync pulse which occurs XDATDLY bit-clocks
earlier than the last transmit bit of the previous frame. Any one of three cases
can occur:
- Case 1: Unexpected FSX pulses with XFIG = 1. This case is discussed in

section 2.3.6.2 on page 2-42 and shown in Figure 2−27 on page 2-43.
- Case 2: Normal serial port transmission. This case is discussed in section

2.3.5.2 on page 2-39. You should note that there are two possible reasons
why a transmit might NOT be in progress:
J This FSX pulse is the first after XRST = 1.
J The serial port is in the interpacket intervals. The programmed data
delay (XDATDLY) may start during these interpacket intervals before
the first bit of the previous word is transmitted. Thus, if operating at
maximum packet frequency, frame synchronization can still be re-
ceived XDATDLY bit-clocks before the first bit of the associated word.
2-50
Figure 2−35. Response to Transmit Frame Synchronization
Transmit frame-sync
pulse occurs
No Case 2:
Unexpected Normal Transmission
frame-sync Start new transmit
pulse
?
Yes
Case 3:
Without Frame Ignore
No Abort transfer.
XFIG=1
? Set XSYNCERR.
Restart current
transfer
Yes
Case 1:
With Frame Ignore
Ignore frame pulse
Transmit stays
running
- Case 3: Unexpected transmit frame synchronization with XFIG = 0. The

case for subsequent frame synchronization with XFIG = 0 at maximum
frame frequency is shown in Figure 2−26. Figure 2−36 shows the case for
normal operation of the serial port with interpacket intervals. In both cases,
the XSYNCERR bit in SPCR2 is set. XSYNCERR can only be cleared by
transmitter reset or by the user writing a 0 to this bit in SPCR2. Note that
if XINTM=11b in SPCR2, XSYNCERR drives the receive interrupt (XINT)
to the CPU.
Note:
The XSYNCERR bit in SPCR2 is a read/write bit. Therefore, writing a 1 to
it sets the error condition. Typically, writing a 0 is expected.
2-51
Figure 2−36. Unexpected Transmit Frame-Synchronization Pulse
CLKX
Unexpected frame synchronization
Á ÁÁ Á
FSX
DX A1
XRDY
Á
A0
ÁÁ B7 B6 B5 B4 B7 B6 B5 B4 B3 B2
Á
B1 B0
DXR1 to XSR1 copy(B) Write to DXR1(C) DXR1 to XSR1 (C) Write to DXR1(D)
XSYNCERR
2.3.8 Receive Data Justification and Sign-Extension: RJUST

RJUST in SPCR1 selects whether data in RBR[1,2] is right or left justified (with
respect to the MSB) in DRR[1,2]. If right-justification is selected, RJUST further
selects whether the data is sign-extended or zero-filled. Table 2−15 shows the
effect various modes of RJUST have on an example 12-bit receive-data value
0xABC. Table 2−16 shows the effect various modes of RJUST have on an ex-
ample 20-bit receive-data value 0xABCDE.
Table 2−15. Use of RJUST Field With 12-Bit Example Data 0xABC
RJUST Justification Extension Value in DRR2 Value in DRR1

00 Right Zero-fill MSBs 0x0000 0x0ABC
01 Right Sign-extend MSBs 0xFFFF 0xFABC
10 Left Zero-fill LSBs 0x0000 0xABC0
11 Reserved Reserved Reserved Reserved
Table 2−16. Use of RJUST Field With 20-Bit Example Data 0xABCDE
RJUST Justification Extension Value in DRR2 Value in DRR1

00 Right Zero-fill MSBs 0x000A 0xBCDE
01 Right Sign-extend MSBs 0xFFFA 0xBCDE
10 Left Zero-fill LSBs 0xABCD 0xE000
11 Reserved Reserved Reserved Reserved
2-52
2.4 µ-LAW/A-LAW Companding Hardware Operation: (R/X)COMPAND
Companding (COMpress and exPAND) hardware allows compression and ex-
pansion of data in either µ-law or A-law format. The companding standard
employed in the United States and Japan is µ-law. The European companding
standard is referred to as A-law. The specification for µ-law and A-law log PCM
is part of the CCITT G.711 recommendation. A-law and µ-law allows 13 bits
and 14 bits of dynamic range, respectively. Any values outside this range will
be set to the most positive or most negative value. Thus, for companding to
work best, the data transferred to and from the McBSP via the CPU or DMA
must be at least 16-bit wide data.
The µ-law and A-law formats encode data into 8-bit code words. Companded
data is always 8-bits wide; therefore, the appropriate (R/X)WDLEN[1,2] must
be set to 0, indicating 8-bit wide serial data stream. If companding is enabled
and either phase of the frame does not have 8-bit word length, then compand-
ing continues as if the word length is eight bits.
When companding is used, transmit data is encoded according to specified

companding law, and receive data is decoded to 2’s complement format. Com-
panding is enabled and the desired format selected by appropriately setting
(R/X)COMPAND in (R/X)CR2 as shown in:
- Table 2−7, Receive Control Register 1 (RCR1) Bit-Field Description, on

page 2-16
- Table 2−8, Receive Control Register 2 (RCR2) Bit-Field Description, on

page 2-17
- Table 2−9, Transmit Control Register 1 (XCR1) Bit-Field Description, on

page 2-19
- Table 2−10, Transmit Control Register 2 (XCR2) Bit-Field Description, on

page 2-20
Compression occurs during the process of copying data from DXR1-to-XSR1

and from RBR1-to-DRR1, as shown in Figure 2−37.
2-53
Figure 2−37. Companding Flow
8 16 16
DR RSR RBR Expand RJUST DRR1 To CPU/DMA
8 16
DX XSR Compress DXR1 From CPU/DMA
For transmit data to be compressed properly, the data should be left-justified

when it is written to DXR1. When using µ-law, the 14 data bits are left-justified
in the register, with the remaining two low-order bits filled with zeros as shown
in Figure 2−38.
Figure 2−38. µ-Law Transmit Data Companding Format
15 2 1 0
µ−law in DXR1 Value 0
When using A-law, the 13 data bits are left-justified in the register, with the re-
maining three low-order bits filled with zeros as shown in Figure 2−39.
Figure 2−39. A-Law Transmit Data Companding Format
15 3 2 0
A−law in DXR1 Value 0
For reception, the 8-bit compressed data in RBR1 is expanded to left-justified

16-bit data in DRR1. Note that RJUST is ignored when companding is used.
2-54
2.4.1 Companding Internal Data
If the McBSP is otherwise unused (serial port X/R sections are reset), the com-
panding hardware can compand internal data. This can be used to:
- Convert linear to the appropriate µ-law or A-law format.
- Convert µ-law or A-law to the linear format.
- Observe the quantization effects in companding by transmitting linear

data, and compressing and re-expanding this data. This is only useful if
both XCOMPAND and RCOMPAND enable the same companding format.
Figure 2−40 shows two methods by which the McBSP can compand internal
data. Data paths for these two methods are used to indicate:
1) When both the transmit and receive sections of the serial port are reset,
DRR1 and DXR1 are internally connected through the companding logic.
Values from DXR1 are compressed, as selected by XCOMPAND, and
then expanded, as selected by RCOMPAND. Note that RRDY and XRDY
bits are not set. However, data is available in DRR1 within four CPU clocks
after being written to DXR1. The advantage of this method is its speed.
The disadvantage is that there is no synchronization available to the CPU
and DMA to control the flow. Note that DRR1 and DXR1 are internally con-
nected if the (X/R)COMPAND bits are set to 1xb, i.e., compand using A-
law or µ-law.
2) The McBSP is enabled in digital loop back mode with companding appro-
priately enabled by RCOMPAND and XCOMPAND. Receive and transmit
interrupts (RINT when RINTM = 0 and XINT when XINTM = 0) or synchro-
nization events (REVT and XEVT) allow synchronization of the CPU or
DMA to these conversions, respectively. Here, the time for this compand-
ing depends on the serial bit rate selected.
Figure 2−40. Companding of Internal Data
DR RSR1 RBR1 Expand RJUST DRR1 To CPU/DMA

(2) (DLB) (1)
DX XSR1 Compress DXR1 From CPU/DMA
2-55
2.4.1.1 Bit Ordering
Normally, all transfers on the McBSP are sent and received with the MSB first.
However, certain 8-bit data protocols (that do not use companded data) re-
quire the LSB to be transferred first. By setting (R/X)COMPAND = 01b in
(R/X)CR2, the bit ordering of 8-bit words is reversed (LSB first) before being
sent to the serial port. Similar to companding, this feature is only enabled if the
appropriate (R/X)WDLEN[1,2] is set to 0, indicating 8-bit words are to be trans-
ferred serially. If either phase of the frame does not have an 8-bit word length,
the McBSP assumes the word length is eight bits, and LSB-first ordering is
done.
2-56
2.5 Programmable Clock and Framing
The McBSP has several means of selecting clocking and framing for both the
receiver and transmitter. Clocking and framing can be sent to both portions by
the sample rate generator. Each portion can select external clocking and/or
framing independently. Figure 2−41 shows a block diagram of the clock and
frame selection circuitry. The features that are enabled by this logic are ex-
plained in the sections that follow.
Figure 2−41. Clock and Frame Generation

Clock selection Frame selection
CLKXM CLKXP FSXP FSXM
CLKX pin See Inset See Inset FSX pin
Internal Internal
CLKXP 0 CLKX FSX 0 DXR to XSR FSXP
Transmit transfer
1 1 0
1
CLKRM CLKRP FSRP FSRM and IGSYNC
CLKXM Internal FSXM
CLKR FSGM
CLKR pin Internal FSR pin
See inset See Inset
FSR
1 1
Receive
CLKRP 0 0 0 0 FSRP
1 1
DLB DLB Inset
CLKRM CLKG FSG FSRM (R/X)RST
Sample Internal Out
CPU clock rate Yyy
CLKS pin
generator
FSRG Internal Yyy is either internal CLKX,
internal CLKR, internal FSR, or
Internal FSX
2-57
2.5.1 Sample Rate Generator Clocking and Framing
The sample rate generator is composed of a three-stage clock divider that al-
lows programmable data clocks (CLKG) and framing signals (FSG) as shown
in Figure 2−42. CLKG and FSG are McBSP internal signals that can be pro-
grammed to drive receive and/or transmit clocking (CLKR/X) and framing
(FSR/X). The sample rate generator can be programmed to be driven by an
internal clock source or an internal clock derived from an external clock source.
The three stages of the sample rate generator circuit compute the following:
- Clock divide down (CLKGDV): The number of input clocks per data bit-
clock.
- Frame period divide down (FPER): The frame period in data bit-clocks.
- Frame width count down (FWID): The width of an active frame pulse in
data bit-clocks.
In addition, a frame pulse detection and clock synchronization module allows

synchronization of the clock divide down with an incoming frame pulse. The
operation of the sample rate generator during device reset is described in sec-
tion 2.3.1, Resetting the Serial Port: (R/X)RST, and RESET, on page 2-22.
Figure 2−42. Sample Rate Generator

CLKGDV FPER FWID
÷ ÷
CPU clock 1 CLKSRG Frame
CLKS Pulse FSG
0
CLKSP
CLKG
CLKSM
Frame pulse
detection
GSYNC and clock
FSR synchronization
2-58
2.5.1.1 Sample Rate Generator Register (SRGR [1,2])
Sample rate generator registers 1 and 2 (SRGR[1,2]) control the operation of

various features of the sample rate generator. These registers and their bit-
field descriptions are shown in Figure 2−43 and Table 2−17, and Figure 2−44
and Table 2−18, respectively. The sections that follow these figures and tables
describe how you can configure the operation of SRGR using the SRGR[1,2]
bit-fields.
Figure 2−43. Sample Rate Generator Register 1 (SRGR1)

15 8 7 0
FWID CLKGDV
RW,+0 RW
Table 2−17. Sample Rate Generator Register 1 (SRGR1) Bit-Field Descriptions

15−8 FWID Frame Width. This field plus 1 determines the width of the frame-sync pulse, 2.5.3.1
FSG, during its active period.
Range: up to 2; 1 to 256 CLKG periods.
7−0 CLKGDV Sample Rate Generator Clock Divider 2.5.2.2

This value is used as the divide-down number to generate the required
sample rate generator clock frequency. Default value is 1.
2-59
Figure 2−44. Sample Rate Generator Register 2 (SRGR2)
15 14 13 12 11 0
GSYNC CLKSP CLKSM FSGM FPER
RW,+0 RW,+0 RW RW,+0 RW,+0

15 GSYNC Sample Rate Generator Clock Synchronization 2.5.2.4
Only used when the external clock (CLKS) drives the sample rate generator
clock (CLKSM=0).
GSYNC = 0 The sample rate generator clock (CLKG) is free running.
GSYNC = 1 The sample rate generator clock (CLKG) is running. But

CLKG is resynchronized and frame-sync signal (FSG) is
generated only after detecting the receive
frame-synchronization signal (FSR). Also, frame period,
FPER, is a don’t care because the period is dictated by the
external frame-sync pulse.
14 CLKSP CLKS Polarity Clock Edge Select 2.5.2.3
Only used when the external clock CLKS drives the sample rate generator
clock (CLKSM = 0).
CLKSP = 0 Rising edge of CLKS generates CLKG and FSG.
CLKSP = 1 Falling edge of CLKS generates CLKG and FSG.
13 CLKSM McBSP Sample Rate Generator Clock Mode 2.5.2.2

CLKSM = 0 Sample rate generator clock derived from the CLKS pin.
CLKSM = 1 Sample rate generator clock derived from CPU clock.
2-60
(Continued)
12 FSGM Sample Rate Generator Transmit Frame-Synchronization Mode 2.5.3.3
Used when FSXM=1 in the PCR.
FSGM = 0 Transmit frame-sync signal (FSX) due to
DXR[1,2]-to-XSR[1,2] copy. When FSGM = 0, FPR and
FWID are ignored.
FSGM = 1 Transmit frame-sync signal driven by the sample rate
generator frame-sync signal, FSG.
11−0 FPER Frame Period. This field plus 1 determines when the next frame-sync signal 2.5.3.1
becomes active.
Range: 1 to 4096 CLKG periods.
2.5.1.2 Sample Rate Generator Reset Procedure
The sample rate generator reset and initialization procedure is as follows:
1) During device reset, GRST = 0. Otherwise, during normal operation, the

sample rate generator can be reset with GRST = 0 in SPCR2, provided
CLKG and/or FSG is not used by any portion of the McBSP. If GRST = 0
due to device reset, CLKG is driven by the divide-by-2 CPU clock, and
FSG is driven inactive-low. If GRST = 0 as programmed by the user,
CLKG and FSG are driven inactive-low. If necessary, set (R/X)RST = 0.
2) Program SRGR[1,2] as required. If necessary, other control registers can

be written with desired values, provided the respective portion (R/X) is in
reset.
3) Wait two CLKSRG clocks. This ensures proper synchronization internally.
4) Set GRST = 1 to enable the sample rate generator.
5) Wait two CLKG bit-clocks.
6) Pull the receiver and/or transmitter out of reset ([R/X]RST = 1), if required.
7) On the next rising edge of CLKSRG, CLKG transitions to 1 and starts

clocking with a frequency equal to (CPU clock/(1+CLKGDV), if
CLKSM = 1, or CLKS clock/(1+CLKGDV) if CLKSM = 0).
8) After the required data acquisition set up is done (DXR[1/2]) is loaded with
data), FRST can be written with 1 if internally generated frame-sync pulse
2-61
is required. FSG is generated with an active-high edge after the pro-
grammed number of eight CLKG clocks have elapsed.
2.5.2 Data Clock Generation

When the receive/transmit clock mode is set to 1 (CLK[R/X]M = 1), the data
clocks (CLK[R/X]) are driven by the internal sample rate generator output
clock, CLKG. You can select from a variety of data bit-clocks independently for
the receiver and transmitter. These options include:
- The input clock to the sample rate generator can be either the CPU clock
or a dedicated external clock input (CLKS).
- The input clock (CPU clock or external clock CLKS) source to the sample
rate generator can be divided down by a programmable value (CLKGDV)
to drive CLKG. Regardless of the source to the sample rate generator, the
rising edge of CLKSRG (see Figure 2−42) generates CLKG and FSG
(also, see section 2.5.2.3).
2.5.2.1 Input Clock Source Mode: CLKSM
The CLKSM bit in SRGR2 selects either the CPU clock (CLKSM = 1) or the
external clock input (CLKSM = 0), CLKS, as the source for the sample rate
generator input clock. Any divide periods are divide-downs calculated by the
sample rate generator and are timed by this input clock selection. When
CLKSM = 1, the minimum value of CLKGDV should be 1.
2.5.2.2 Sample Rate Generator Data Bit Clock Rate: CLKGDV
The first divider stage generates the serial data bit clock from the input clock.
This divider stage utilizes a counter that is preloaded by CLKGDV which con-
tains the divide ratio value. The output of this stage is the data bit-clock which
is output on sample rate generator output, CLKG, and serves as the input for
the second and third divider stages.
CLKG has a frequency equal to 1/(CLKGDV+1) of sample rate generator input

clock. Thus, sample generator input clock frequency is divided by a value be-
tween 1 and 256. When CLKGDV is odd or equal to 0, the CLKG duty cycle
is 50%. When CLKGDV is an even value, 2p, representing an odd divide-
down, the high-state duration is p+1 cycles and the low-state duration is p
cycles.
2.5.2.3 Bit Clock Polarity: CLKSP
External clock (CLKS) is selected to drive the sample rate generator clock di-
vider by selecting CLKSM=0. In this case, the CLKSP bit in SRGR2 selects the
2-62
edge of CLKS on which sample rate generator data bit-clock (CLKG) and
frame-sync signal (FSG) are generated. Since the rising edge of CLKSRG
(see Figure 2−42) generates CLKG and FSG, the rising edge of CLKS when
CLKSP = 0, or the falling edge of CLKS when CLKSP = 1, causes the transi-
tion on the data bit-rate clock (CLKG) and frame sync (FSG).
2.5.2.4 Bit Clock and Frame Synchronization
When CLKS is selected to drive the sample rate generator (CLKSM = 0),
GSYNC can be used to configure the timing of CLKG relative to CLKS.
GSYNC = 1 ensures that the McBSP, and the external device that it is commu-
nicating to, are dividing down CLKS with the same phase relationship. If
GSYNC = 0, this feature is disabled and therefore CLKG runs freely and is not
re-synchronized. If GSYNC = 1, an inactive-to-active transition on FSR trig-
gers a resynchronization of CLKG and generation of FSG. CLKG always be-
gins with a high state after synchronization. Also, FSR is always detected at
the same edge of CLKS that generates CLKG, no matter how long the FSR
pulse is. Although an external FSR is provided, FSG can still drive internal re-
ceive frame synchronization when GSYNC = 1. Note that when GSYNC = 1,
FPER is a don’t care because the frame period is determined by the arrival of
the external frame-sync pulse.
Figure 2−45 and Figure 2−46 show the bit clock and frame-synchronization
operation with various polarities of CLKS and FSR. These figures assume
FWID = 0, for an FSG one CLKG wide.
Figure 2−45. CLKG Synchronization and FSG Generation When GSYNC = 1 and
CLKGDV = 1
CLKS (CLKSP=1)
CLKS (CLKSP=0)
FSR external
(FSRP=0)
FSR external
(FSRP=1)
CLKG
(No need to
resync)
CLKG
(needs resync)
FSG
2-63
Figure 2−46. CLKG Synchronization and FSG Generation When GSYNC = 1 and
CLKGDV = 3
CLKS (CLKSP=1)
CLKS (CLKSP=0)
FSR external
(FSRP=0)
FSR external
(FSRP=1)
CLKG
(No need to
resync)
CLKG
(needs resync)
FSG
Note that FPER is not programmed since it is determined by the arrival of the
next external frame-sync pulse. Figure 2−45 and Figure 2−46 show what hap-
pens to CLKG both when it is initially synchronized and GSYNC = 1, and when
it is not initially synchronized and GSYNC = 1.
When GSYNC = 1, the transmitter can operate synchronously with the receiv-
er, providing:
1) FSX is programmed to be driven by the sample rate generator frame sync,

FSG (FSGM = 1 in SRGM and FSXM = 1 in the PCR). If the input FSR has
appropriate timing so that it can be sampled by the falling edge of CLKG,
it can be used, instead, by setting FSXM = 0 in the PCR and connecting
FSR to FSX externally.
2) The sample rate generator clock should drive the transmit and receive bit
clock (CLK[R/X]M = 1 in SPCR[1,2]). Therefore, the CLK(R/X) pin should
not be driven by any other driving source.
2.5.2.5 Digital Loop Back Mode: DLB
Setting DLB = 1 in SPCR1 enables digital loop back mode. During DLB mode,
DR, FSR, and CLKR are internally connected through multiplexers to DX,
FSX, CLKX, respectively, as shown in Figure 2−41. DLB mode allows testing
of serial port code with a single DSP device. Figure 2−42 shows the multiplex-
ing of receiver control inputs during digital loop back mode.
2-64
2.5.2.6 Receive Clock Selection: DLB, CLKRM
Table 2−19 shows how the digital loop back bit (DLB) and the CLKRM bit in the
PCR can select the receiver clock. In digital loop back mode (DLB =1), the
transmitter clock drives the receiver. CLKRM determines whether the CLKR
pin is an input or an output.
Table 2−19. Receive Clock Selection
DLB in CLKRM
SPCR1 in PCR Source of Receive Clock CLKR Pin
0 0 CLKR pin acts as an input driven by Input
external clock and inverted as determined
by CLKRP before being used.
0 1 Sample Rate Generator Clock (CLKG) Output. CLKG inverted as determined by

drives CLKR. CLKRP before being driven out on CLKR.
1 0 Internal CLKX drives the receive clock High Impedance

internal CLKR as selected and inverted as
shown in Table 2−20.
1 1 Internal CLKX drives internal CLKR as Output. CLKR (same as transmit) inverted
selected and inverted as shown in as determined by CLKRP before being
Table 2−20. driven out.
2.5.2.7 Transmit Clock Selection: CLKXM
Table 2−20 shows how the CLKXM bit in the PCR selects the transmit clock
and indicates whether the CLKX pin is an input or output.
Table 2−20. Transmit Clock Selection
CLKXM
in PCR Source of Transmit Clock CLKX Pin
0 External clock drives the CLKX input pin. CLKX is Input
inverted as determined by CLKXP before being used.
1 Sample rate generator clock, CLKG, drives transmit Output. CLKG inverted as determined by
clock CLKXP before being driven out on
CLKX.
2-65
2.5.3 Frame-Sync Signal Generation
Similar to data bit clocking, data frame synchronization is also independently

programmable for the receiver and transmitter for all data delays. When set
to one, the FRST bit in SPCR2 activates the frame-sync generation logic to
generate a frame-sync signal, provided FSGM = 1 in SRGR2. The frame-sync
programming options are:
- A frame pulse with a programmable period between sync pulses and pro-
grammable active width, using the sample rate generator register
(SRGR1).
- The transmit portion may trigger its own frame-sync signal generated by
a DXR[1,2]-to-XSR[1,2] copy. This causes a frame sync to occur on every
DXR1 to XSR1 copy. The data delays can be programmed as required;
however, maximum frame frequency cannot be achieved in this method
for data delays one and two. This limitation can be overcome by program-
ming the frame ignore bit (R/X)FIG = 1.
- Both the receive and transmit sections may independently select an exter-
nal frame synchronization on the FSR and FSX pins, respectively.
2.5.3.1 Frame Period and Frame Width: FPER and FWID
FPER and FWID are implemented as down-counters. The FPER stage is a

12-bit down-counter that counts down the generated data clocks from 4095 to
0. FPER controls the period of active frame-sync pulses. The FWID stage in
the sample rate generator is an 8-bit down counter. The FWID field controls
the active width of the frame-sync pulse. Both these counters gets loaded with
their respective programmed value in FPER and FWID.
When the sample rate generator comes out of reset, FSG is in its inactive state.
Then, when FRST = 1 and FSGM = 1, a frame sync is generated. The frame
width value (FWID+1) is counted down on every CLKG cycle until it reaches
zero, when FSG goes low. Thus, the value of FWID + 1 determines an active
frame pulse width ranging from 1 to 256 data bit-clocks. At the same time, the
frame period value (FPER+1) is also counting down. When this value reaches
zero, FSG goes high, again indicating a new frame.
It is recommended that FWID be programmed to a value less than

WDLEN[1,2].Thus, the value of FPER+1 determines a frame length from 1 to
4096 data bits. When GSYNC = 1, FPER is a don’t care value. Figure 2−47
shows a frame of period 16 CLKG periods (FPER = 15 or 00001111b), and a
frame with an active width of 2 CLKG periods (FWID = 1).
2-66
Figure 2−47. Programmable Frame Period and Width
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19
CLKG
Frame period: (FPER+1) x CLKG
Frame width: (FWID + 1) x CLKG
FSG
2.5.3.2 Receive Frame-Sync Selection: DLB, FSRM, GSYNC
Table 2−21 shows how you may select various sources to provide the receive
frame-synchronization signal. Note that in digital loop back mode (DLB = 1),
the transmit frame-sync signal is used as the receive frame-sync signal, and
that DR is connected to DX internally.
Table 2−21. Receive Frame-Synchronization Selection

DLB FSRM GSYNC
in in in Source of Receive Frame
SPCR1 PCR SRGR2 Synchronization FSR Pin
0 0 x External frame-sync signal drives the Input
FSR input pin. This is then inverted as
determined by FSRP before being
used as internal FSR.
0 1 0 Internal FSR driven by sample rate Output. FSG inverted as determined

generator Frame-Sync signal (FSG), by FSRP before being driven out on
FRST = 1 FSR pin.
0 1 1 Internal FSR driven by sample rate Input. The external frame-sync input
generator Frame-Sync signal (FSG), on FSR is used to synchronize CLKG
FRST = 1 and generate FSG.
1 0 0 Internal FSX drives internal FSR. FSX High Impedance.

is selected as shown in Table 2−22.
1 X 1 Internal FSX drives internal FSR and is Input. External FSR not used for frame
selected as shown in Table 2−22. synchronization but still used to
synchronize CLKG and generate FSG,
since GSYNC = 1.
1 1 0 Internal FSX drives internal FSR and is Output. Receive (same as transmit)
selected as shown in Table 2−22. frame synchronization inverted as
determined by FSRP before being
driven out.
2-67
2.5.3.3 Transmit Frame-Sync Signal Selection: FSXM, FSGM
Table 2−22 shows how you can select the source of transmit frame-synchro-
nization pulses. The three choices are:
1) External frame-sync input.
2) The sample rate generator frame-sync signal, FSG.
3) A signal that indicates a DXR[1,2]-to-XSR[1,2] copy has been made.
Table 2−22. Transmit Frame-Synchronization Selection
FSXM FSGM
in in Source of Transmit Frame
PCR SRGR Synchronization FSX Pin
0 x External frame-sync input on FSX pin. This Input.
is inverted by FSXP before being used as
internal FSX.
1 1 Sample rate generator frame-sync signal Output. FSG inverted by FSXP before being
(FSG) drives internal FSX. FRST = 1 driven out on FSX pin.
1 0 A DXR[1,2]-to-XSR[1,2] copy activates Output. One bit-clock wide signal inverted

transmit frame-sync signal. as determined by FSXP before being driven
out on FSX pin.
2.5.3.4 Frame Detection for Initialization
To facilitate detection of frame synchronization, the receive and transmit CPU

interrupts (RINT and XINT) may be programmed to detect frame synchroniza-
tion by setting RINTM = XINTM=10b in SPCR[1,2]. Unlike other types of serial
port interrupts, this mode can operate while the associated portion of the serial
port is in reset (such as activating RINT when the receiver is in reset). In this
case, FS(R/X)M and FS(R/X)P still select the appropriate source and polarity
of frame synchronization. Thus, even when the serial port is in reset state,
these signals are synchronized to the CPU clock and then sent to the CPU in
the form of RINT and XINT at the point at which they feed the receive and trans-
mit portions of the serial port. Consequently, a new frame-synchronization
pulse can be detected, and after this occurs the CPU can take the serial port
out of reset safely.
2-68
2.5.4 Clocking Examples
The following sections provide clocking examples:
2.5.4.1 Double-Rate ST-BUS Clock
Figure 2−48 shows McBSP configuration to be compatible with the Mitel ST-
Bus. Note that this operation is running at maximum frame frequency.
- CLK(R/X)M = 1, internal CLK(R/X) generated internally by sample rate
generator
- GSYNC = 1, synchronize CLKG with external frame-sync signal input on
FSR. Note that CLKG is not synchronized until the frame-sync signal is ac-
tive. Also, note that FSR is regenerated internally to form a minimum pulse
width.
- CLKSM = 1, External clock (CLKS) drives the sample rate generator
- CLKSP = 1, falling edge of CLKS generates CLKG and thus internal
CLK(R/X)
- CLKGDV = 1, receive clock (shown as CLKR) is half of CLKS frequency
- FS(R/X)P = 1, active-low frame-sync pulse
- (R/X)FRLEN1 = 11111b, 32 words per frame

- (R/X)WDLEN1 = 0, 8-bit word
- (RX)PHASE = 0, single-phase frame and thus

(R/X)FRLEN2 = (R/X)WDLEN2 = X
- (R/X)DATDLY = 0, no data delay
Figure 2−48. ST-BUS and MVIP Example
4.096 MHz CLKS
FSR external
Internal FSG, FSR,

internal FSX
2.048 MHz CLKG,
ÁÁ
internal CLKR,
internal CLKS
ÁÁ
(first FSR)
DR, DX (first FSR) W1B7 W1B6 W1B5 W1B4 W1B3 W1B2 W1B1 W1B0 W2B7
Internal CLKG, CLKR,

(subsequent FSR)
DR, DX W32B0 W1B7 W1B6 W1B5 W1B4 W1B3 W1B2 W1B1 W1B0 W2B7
(subsequent FSR)
WxBy = Word x Bit y
2-69
2.5.4.2 Single-Rate ST-BUS Clock
This example is the same as the ST-BUS example except for the following:
- CLKGDV = 0, CLKS drives internal CLK(R/X) without any divide down

(single-rate clock).
- CLKSP = 0, rising edge of CLKS generates internal clocks CLKG and in-
ternal CLK(R/X).
Figure 2−49. Single-Rate Clock Example
CLKS
FSR external
Internal FSG, FSR,

internal FSX
Á
internal CLKX
Á
(first FSR)
DR, DX (first FSR) W1B7 W1B6 W1B5 W1B4 W1B3 W1B2 W1B1 W1B0 W2B7

(subsequent FSR)
DR, DX
W32B0 W1B7 W1B6 W1B5 W1B4 W1B3 W1B2 W1B1 W1B0 W2B7
(subsequent FSR)
WxBy = Word x Bit y
The rising edge of CLKS is used to detect the external FSR. This external
frame-sync pulse is used to resynchronize internal McBSP clocks and gener-
ate frame-sync for internal use. Note that the internal frame sync is generated
so that it is wide enough to be detected on the falling edge of internal clocks.
2-70
2.5.4.3 Double-Rate Clock Example
This example is the same as the ST-BUS example except for the following:
- CLKSP = 0, rising edge of CLKS generates CLKG and thus CLK(R/X).
- CLKGDV = 1, CLKG and thus internal CLKR and internal CLKX frequency
is half CLKS.
- GSYNC = 0, CLKS drives CLKG. CLKG runs freely and is not resynchro-
nized by FSR.
- FS(R/X)M = 0, frame synchronization is externally generated. The fram-

ing pulse is wide enough to be detected.
- FS(R/X)P = 0, active-high input frame-sync signal.
- (R/X)DATDLY = 1, data delay of one bit.
Figure 2−50. Double-Rate Clock Example
CLKS
Internal FS(R/X)
Internal CLK(R/X)
D(R/X) W32B0 W1B7 W1B6 W1B5 W1B4 W1B3 W1B2 W1B1 W1B0 W2B7
WxBy = Word x Bit y
2-71
2.6 Multichannel Selection Operation
Multiple channels can be independently selected for the transmitter and re-
ceiver by configuring the McBSP with a single-phase frame. Each frame repre-
sents a time-division multiplexed (TDM) data stream. The number of words per
frame represented by (R/X)FRLEN1, denotes the number of channels avail-
able for selection.
When using TDM data streams, the CPU may need to process only a few of
them. Thus, to save memory and bus bandwidth, multichannel selection al-
lows independent enabling of particular channels for transmission and recep-
tion. Up to 32 channels can be enabled in an up-to-128-channel bit-stream.
If a receive channel is not enabled:
- RRDY is not set to 1 upon reception of the last bit of the word.
- RBR[1,2] is not copied to DRR[1,2] upon reception of the last bit of the
word. Thus, RRDY is not set active. This feature also implies that no inter-
rupts or synchronization events are generated for this word.
If a transmit channel is not enabled:
- DX is in the high-impedance state.
- A DXR[1,2]-to-XSR[1,2] transfer is not automatically triggered at the end

of serial transmission of the related word.
- XEMPTY and XRDY similarly are not affected by the end of transmission
of the related serial word .
A transmit channel which is enabled can have its data masked or transmitted.
When masked, the DX pin will be forced to the high-impedance state even
though the transmit channel is enabled.
2-72
2.6.1 Multichannel Operation Control Registers
The following control registers are used in multichannel operation:
1) Multichannel control 1 and 2 (MCR[1,2]) registers
2) Transmit channel enable partition A/B (XCER[A/B]) registers
3) Receive channel enable partition A/B (RCER[A/B]) registers
The use of these registers in controlling multichannel operation is described

in the sections that follow.
Figure 2−51. Multichannel Control Register 1 (MCR1)

15 9 8 7 6 5 4 2 1 0
rsvd RPBBLK RPABLK RCBLK rsvd RMCM
R,+0 RW,+0 RW,+0 R,+0 R,+0 RW,+0
Table 2−23. Multichannel Control Register 1 (MCR1) Bit-Field Descriptions

8−7 RPBBLK Receive Partition B Block 2.6.3
RPBBLK = 00 Block 1. Channel 16 to channel 31
6−5 RPABLK Receive Partition A Block 2.6.3
RPABLK = 00 Block 0. Channel 0 to channel 15
2-73
(Continued)
4−2 RCBLK Receive Current Block 2.6.3.2
RCBLK = 000 Block 0. Channel 0 to channel 15
1 rsvd Reserved
0 RMCM Receive Multichannel Selection Enable 2.6.2
RMCM = 0 All 128 channels enabled.
RMCM = 1 All channels disabled by default. Required channels are

selected by enabling RP(A/B)BLK and RCER(A/B)
appropriately.
2-74
Figure 2−52. Multichannel Control Register 2 (MCR2)
15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0
RSVD XPBBLK XPABLK XCBLK XMCM
R,+0 RW,+0 RW,+0 R,+0 RW,+0

8−7 XPABLK Transmit Partition A Block 2.6.3
XPABLK = 00 Block 0. Channel 0 to channel 15
6−5 XPBBLK Transmit Partition B Block 2.6.3
XPBBLK = 00 Block 1. Channel 16 to channel 31
4−2 XCBLK Transmit Current Block 2.6.3.2
XCBLK = 000 Block 0. Channel 0 to channel 15
2-75
(Continued)
1−0 XMCM Transmit Multichannel Selection Enable 2.6.2
XMCM = 00 All channels enabled without masking (DX is always

driven during transmission of data{).
XMCM = 01 All channels disabled and therefore masked by default.

Required channels are selected by enabling XP(A/B)BLK
and XCER(A/B) appropriately. Also, these selected
channels are not masked and therefore DX is always
driven.
XMCM = 10 All channels enabled, but masked. Selected channels

enabled via XP(A/B)BLK and XCER(A/B) are unmasked.
XMCM = 11 All channels disabled and therefore masked by default.

Required channels are selected by enabling
RP(A/B)BLK and RCER(A/B) appropriately. Selected
channels can be unmasked by RP(A/B)BLK and
XCER(A/B). This mode is used for symmetric transmit
and receive operation.
† DX is masked or driven to hi-Z during (a) interpacket intervals, (b) when a channel is masked regardless of whether it is enabled,
or (c) when a channel is disabled.
2.6.2 Enabling Multichannel Selection

The multichannel mode can be enabled independently for receive and trans-
mit by setting RMCM = 1 and XMCM to a non-zero value in MCR[1,2], respec-
tively.
2.6.3 Enabling and Masking of Channels

A total of 32 out of the available 128 channels may be enabled at any given
point in time. The 128 channels comprise eight blocks (0 through 7) and each
block has 16 contiguous channels. Further, even-numbered blocks 0, 2, 4, and
6 belong to Partition A, and odd-numbered blocks 1, 3, 5, and 7 belong to Parti-
tion B.
The number of channels enabled can be updated during the course of a frame
to allow any arbitrary group of channels to be enabled. This update is accom-
plished using an alternating scheme controlling two blocks (one odd-num-
bered and one even-numbered) of 16 contiguous channels each, at any given
time within the frame. One block belongs to Partition A and the other to Parti-
tion B.
2-76
Any two out of the eight 16-channel blocks may be selected, yielding a total
of 32 channels that can be enabled. The blocks are allocated on 16-channel
boundaries within the frame as shown in Figure 2−53. (R/X)PABLK and
(R/X)PBBLK fields in MCR[1,2] determine the blocks that get selected in parti-
tion A and B, respectively. This enabling is performed independently for trans-
mit and receive.
Figure 2−53. Channel Enabling by Blocks in Partition A and B

Block 0 1 2 3 4 5 6 7 0
Partition A 0−15 32−47 64−79 96−111 0−15
ÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁ
Partition B 16−31 48−63 80−95 112−127
FS(R/X)
ÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁ
Transmit data masking allows a channel enabled for transmit to have its DX
pin set to the high-impedance state during its transmit period. In systems
where symmetric transmit and receive provides software benefits, this feature
allows transmit channels to be disabled on a shared serial bus. A similar fea-
ture is not needed for receive, since multiple receptions cannot cause serial
bus contention.
Note:
DX is masked or driven to high-Impedance during (a) interpacket intervals,
(b) when a channel is masked regardless of whether it is enabled, or (c) when
a channel is disabled.
The following gives a description of the multichannel operation during transmit

for various XMCM values:
- XMCM = 00b: The serial port transmits data over the DX pin for as many
number of words programmed in XFRLEN1. Thus, DX is driven during
transmit.
- XMCM = 01b: Required channels, or only those words that need to be
transmitted, are selected via XP(A/B)BLK and XCER(A/B). Therefore,
only these selected words will be written to DXR[1,2] and ultimately trans-
mitted. In other words, if XINTM = 00b, which implies that an XINT will be
generated for every DXR1-to-XSR1 copy, the number of XINT generated
will be equal to the number of channels selected via XCER(A/B) (and NOT
equal to XFRLEN1).
- XMCM = 10b: For this case, all channels are enabled which means all the
words in a data frame (XFRLEN1) will be written to DXR[1,2] and
2-77
DXR[1,2]-to-XSR[1,2] copy occurs at their respective times. However, DX
will be driven only for those channels that are selected via XP(A/B)BLK
and XCER(A/B), and is placed in the high-impedance state otherwise. In
this case, if XINTM = 00b, the number of interrupts generated due to every
DXR1-to-XSR1 copy would equal the number of words in that frame
(XFRLEN1).
- XMCM = 11b: This mode is basically a combination of XMCM = 01b and

10b cases so that symmetric transmit and receive operation is achieved.
For receive, an RBR[1,2]-to-DRR[1,2] copy occurs only for those channels
that are selected via RP(A/B)BLK and RCER(A/B). If RINT were to be gen-
erated for every RBR[1,2]-to-DRR[1,2] copy, it would occur as many times
as the number of channels selected in RCER(A/B) (and NOT the number
of words programmed in RFRLEN1). For transmit, the same block that is
used for reception is used in order to maintain symmetry and thus
XP(A/B)BLK is a don’t care. DXR[1,2] is loaded and DXR[1,2]-to-XSR[1,2]
copy occurs for all the channels that are enabled via RP(A/B)BLK. Howev-
er, DX will be driven only for those channels that are selected via
XCER(A/B). Note that the channels enabled in XCER(A/B) can only be a
subset, or same as, those selected in RCER(A/B). Therefore, if
XINTM = 00b, transmit interrupts to the CPU would be generated as many
times as the number of channels selected in RCER(A/B) (not XCER[A/B]).
Figure 2−54 shows the activity on the McBSP pins for all the previously de-
scribed modes with the following conditions:
- (R/X)PHASE = 0, single-phase frame for multichannel selection enabled
- FRLEN1 = 011b, 4-word frame
- WDLEN1 = 000b, 8-bit word
Please note that in the following illustrations, the arrows showing where the
various events occur are only sample indications. Wherever possible, there is
a time window in which these events can occur.
2-78
Figure 2−54. XMCM Operation
(a) XMCM = 00b
ÁÁ Á
Internal FSX
DX
XRDY
ÁÁ Á W0 W1 W2 W3
Write to DXR1(W3)
Write to DXR1(W1) DXR1 to XSR1 copy(W2)
DXR1 to XSR1 copy(W0) DXR1 to XSR1 copy(W3)
Write to DXR1(W2)
DXR1 to XSR1 copy(W1)
(b) XMCM = 01b; XPABLK = 00b, XCERA = 1010b

Internal FSX
DX
XRDY
ÁÁ Á W1
Á Á W3
Write to DXR1(W3) DXR1 to XSR1 copy(W3)

(c) XMCM = 10b; XPABLK = 00b, XCERA = 1010b
ÁÁ Á Á Á Á
Internal FSX
DX
XRDY
ÁÁ Á W1
Á Á W3
Á
Write to DXR1(W1) Write to DXR1(W3)
DXR1 to XSR1 copy(W0) DXR1 to XSR1 copy(W3)
Write to DXR1(W2)
(d) XMCM = 11b; RPABLK = 00b, XPABLK = X, RCERA = 1010b, XCERA = 1000b
Internal FS(R/X)
Á Á Á Á Á
Á Á Á Á Á
DR W1 W3
RRDY
Read from DRR1(W3) Read from DRR1(W1)
Á Á Á
RBR1 to DRR1 copy (W3) RBR1 to DRR1 copy (W1)
Á Á Á
RBR1 to DRR1 (W3)
DX W3
XRDY
Write to DXR1(W3) DXR1 to XSR1 copy (W3)

DXR1 to XSR1 copy (W1)
2-79
2.6.3.1 Channel Enable Registers: (R/X)CER(A/B)
The Receive Channel Enable Partition A and B (RCER[A/B]) and Transmit

Channel Enable Partition A and B (XCER[A/B]) registers are used to enable
any of the 32 channels for receive and transmit, respectively. Out of the 32
channels, 16 channels belong to a block in partition A and the other 16 belong
to a block in partition B. They are shown in Figure 2−55, Figure 2−56,
Figure 2−57, and Figure 2−58.
(R/X)CERA and (R/X)CERB register fields shown in Table 2−25, Table 2−26,
Table 2−27, and Table 2−28 enable channels within the 16-channel-wide
blocks in partitions A and B, respectively. The (R/X)PABLK and (R/X)PBBLK
fields in the MCR select which 16-channel blocks get selected.
Figure 2−55. Receive Channel Enable Register Partition A (RCERA)

15 14 13 12 11 10 9 8
RCEA15 RCEA14 RCEA13 RCEA12 RCEA11 RCEA10 RCEA9 RCEA8
RW,+0 RW,+0 RW,+0 RW,+0 RW,+0 RW,+0 RW,+0 RW,+0
7 6 5 4 3 2 1 0
RCEA7 RCEA6 RCEA5 RCEA4 RCEA3 RCEA2 RCEA1 RCEA0
Table 2−25. Receive Channel Enable Register Partition A (RCERA) Bit-Field

Descriptions
Bit Name Function
15−0 RCEA(0:15) Receive Channel Enable-
RCEA n = 0 Disables reception of nth channel in an even-numbered block in

partition A
RCEA n = 1 Enables reception of nth channel in an even-numbered block in

partition A
2-80
Figure 2−56. Receive Channel Enable Register Partition B (RCERB)
15 14 13 12 11 10 9 8
RDEB15 RCEB14 RCEB13 RCEB12 RCEB11 RCEB10 RCEB9 RCEB8
7 6 5 4 3 2 1 0
RCEB7 RCEB6 RCEB5 RCEB4 RCEB3 RCEB2 RCEB1 RCEB0
Table 2−26. Receive Channel Enable Register Partition B (RCERB) Bit-Field

Descriptions
Bit Name Function
15−0 RCEB(0:15) Receive Channel Enable-
RCEB n = 0 Disables reception of nth channel in an even-numbered block in

partition B
RCEB n = 1 Enables reception of nth channel in an even-numbered block in

partition B
Figure 2−57. Transmit Channel Enable Register Partition A (XCERA)
15 14 13 12 11 10 9 8
XCEA15 XCEA14 XCEA13 XCEA12 XCEA11 XCEA10 XCEA9 XCEA8
7 6 5 4 3 2 1 0
XCEA7 XCEA6 XCEA5 XCEA4 XCEA3 XCEA2 XCEA1 XCEA0
Table 2−27. Transmit Channel Enable Register Partition A (XCERA) Bit-Field

Descriptions
Bit Name Function
15−0 XCEA(0:15) Transmit Channel Enable-
XCEA n = 0 Disables transmission of nth channel in an even-numbered block in

partition A
XCEA n = 1 Enables transmission of nth channel in an even-numbered block in

partition A
2-81
Figure 2−58. Transmit Channel Enable Register Partition B (XCERB)
15 14 13 12 11 10 9 8
XCEB15 XCEB14 XCEB13 XCEB12 XCEB11 XCEB10 XCEB9 XCEB8
7 6 5 4 3 2 1 0
XCEB7 XCEB6 XCEB5 XCEB4 XCEB3 XCEB2 XCEB1 XCEB0
Table 2−28. Transmit Channel Enable Register Partition B (XCERB) Bit-Field

Descriptions
Bit Name Function
15−0 XCEB(0:15) Transmit Channel Enable-
XCEB n = 0 Disables transmission of nth channel in an even-numbered block in

partition B
XCEB n = 1 Enables transmission of nth channel in an even-numbered block in

partition B
2.6.3.2 Changing Channel Selection
Using the multichannel selection feature, a static group of 32 channels may

be enabled, and will remain enabled with no CPU intervention until this alloca-
tion requires modification. An arbitrary number, group, or all of the words/chan-
nels within a frame can be accessed by updating the block allocation registers
during the course of the frame, in response to the end-of-block interrupts (see
section 2.6.3.3, Update Interrupts).
Note:
When changing the selection, the user must be careful not to affect the cur-
rently selected block.
The currently selected block is readable through the RCBLK and XCBLK fields
in MCR1 for receive and MCR2 for transmit, respectively. The associated
channel enable register cannot be modified if it is selected by the appropriate
(R/X)P(A/B)BLK register to point toward the current block. Similarly, the
(R/X)PABLK and (R/X)PBBLK fields in MCR[1,2] cannot be modified while
pointing to, or being changed to point to, the currently selected block. Note that
if the total number of channels is 16 or less, the current partition is always
pointed to. In this case, only a reset of the serial port can change enabling.
2-82
2.6.3.3 Update Interrupts
At the end of every 16-channel block boundary during multichannel operation,

the receive interrupt (RINT) or transmit interrupt (XINT) to the CPU is gener-
ated if RINTM = 01b or XINTM = 01b in SPCR[1,2], respectively. This interrupt
indicates that a new partition has been crossed. You can then check the cur-
rent partition and change the selection of blocks in the A and/or B partitions,
if they do not point to the current block. These interrupts are two CPU-clock
long, active-high pulses. If RINTM = XINTM = 01b when (R/X)MCM = 0 (non-
multichannel operation), interrupts are not generated.
2.6.4 A-bis interface functionality (available on ’5410 only)

In the A-bis mode (ABIS = 1), the McBSP can receive and transmit up to 1024
bits on a PCM link. The receive section can extract all 1024 bits from a 1024-bit
PCM frame according to a given receiving pattern, and generate an interrupt
to the CPU when DRR is compacted with a 16-useful-bit word, or when a
receive frame is ended. In addition, the transmit section can expand up to 1024
bits into a 1024-bit PCM frame at a specific position, according to a given
transmitting pattern, and generate an interrupt when a 16-bit word is
transmitted or a transmit frame is ended. In this mode, a word length must be
specified as 16-bit mode (WDLEN1 = 010b); otherwise, the A-bis mode is
undetermined.
In A-bis mode, the channel enable registers, (R/X)CER(A/B), have a different

function than in normal multichannel operation. Instead of indicating which
channels will be enabled, in A-bis mode these registers indicate which bits in
the data stream will be enabled. A 1 in a given position in the (R/X)CER(A/B)
register enables a corresponding bit in the receive or transmit data stream.
On the receiver, bits that are not enabled in the RCER(A/B) registers are ig-
nored and are not compacted in the receiver. Bits that are enabled are received
and compacted into a useful data word. When 16 useful-data bits have been
received, the received word is copied from RSR1 to DRR1 and the McBSP
generates an interrupt to the CPU. RCERA and RCERB alternate specifying
the receive masking pattern for each of the 16 receive clocks. Figure 2−59
shows an example bit sequence for the receiver.
2-83
Figure 2−59. A-bis Mode Receive Operation
RCERA 0 1 0 1 0 1 0 1 0 1 0 1 0 1 0 1
RCERB 0 0 1 0 0 1 1 1 0 0 1 0 0 1 1 1
DR pin† 1 0 1 1 0 1 0 1 0 0 1 1 0 1 0 1 0 1 1 1 0 1 1 0 1 1 0 1 1 0 1 1
DRR1‡ − 0 − 1 − 1 − 1 − 0 − 1 0 − − 1 − − 1 − − 1 1 0 − − 0 − − 0 1 1 75E3h
† Data arriving on the DR pin
‡ Received data in DRR1
On the transmitter, only bits that are enabled in XCER(A/B) are transmitted out
from the DX pin. Bits that are not enabled are not transmitted and the DX pin
is in the high-impedance state during that clock cycle. XCERA and XCERB al-
ternate specifying the receive masking pattern for each 16 receive clocks.
When 16 useful bits have been shifted out, the McBSP generates an interrupt
to the CPU. Figure 2−60 shows an example bit sequence for the transmitter.
Figure 2−60. A-bis Mode Transmit Operation (z Indicates High-Impedance)

XCERA 0 1 1 1 1 1 0 0 0 1 1 0 0 1 1 1
XCERB 0 0 1 1 0 0 0 1 1 0 0 0 1 1 0 0
DXR1§ 1 0 1 1 0 1 0 1 0 0 0 1 1 1 1 1 1 0 1 1 0 1 1 0 1 1 0 0 0 0 0 0
DX pin¶ z 0 1 1 0 1 z z z 0 0 z z 1 1 1 z z 1 1 z z z 0 1 z z z 0 0 z z
§ Data written to DXR1
¶ Data transmitted on the DX pin
Two special events, XEVTA (transmit A-bis event) and REVTA (receive A-bis
event), can be used by the DMA to load patterns into the (R/X)CER(A/B) regis-
ters. This capability is used for bit sequences that are longer than the 32 bits
covered by (R/X)CER(A/B). These two events are generated every 16 CLKR/
CLKX cycles.
As an example, the following gives a description of the A-bis operation on a

256-bit PCM link:
- ABIS bit is enabled in SPCR1.
- The initial pattern of bits that must be enabled is loaded into the
(R/X)CER(A/B) registers.
- (R/X)PHASE = 1111b, single-phase frame.
- FRLEN1 = 1111b, 16-word frame.
- WDLEN1 = 010b, 16-bit words (required for A-bis mode).
2-84
Two DMA channels (one for transmit, one for receive) are used to update the
pattern selection data to (R/X)CER(A/B) as the operation proceeds. One
16-word block in memory contains the bit pattern selections for the receiver.
Sixteen words of 16-bits each contain the entire receive selection pattern for
the 256-bit PCM link. On each REVTA event, the DMA copies new receive
selection pattern data from memory to RCER(A/B) and automatically toggles
its destination pointer from RCERA to RCERB, or vice versa, as necessary.
In other words, the DMA channel is initially set to RCERA as a destination. Af-
ter the first access to RCERA, the destination automatically toggles to
RCERB. After the next RCERB access, the destination automatically toggles
back to RCERA. Since the toggling between RCERA and RCERB is handled
automatically, you do not need to configure the DMA controller to modify the
destination address by other means. As the A-bis transfer proceeds, the re-
ceiver alternates using RCERA and RCERB to specify the enable pattern for
each group of 16 serial port clocks.
Another 16-word block in memory contains the bit pattern selections for the
transmitter. Sixteen words of 16-bits each contain the entire transmit selection
pattern for the 256-bit PCM link. On each XEVTA event, the DMA copies the
new transmit selection pattern data from memory to XCER(A/B) and automati-
cally toggles its destination pointer from XCERA to XCERB, or vice versa, as
necessary. In other words, the DMA channel is initially set to XCERA as a des-
tination. After the first access to XCERA, the destination automatically toggles
to XCERB. After the next XCERB access, the destination automatically
toggles back to XCERA. Since the toggling between XCERA and XCERB is
handled automatically, you do not need to configure the DMA controller to
modify the destination address by other means. As the A-bis transfer pro-
ceeds, the receiver alternates. The transmitter alternates using XCERA and
XCERB to specify the enable pattern for each group of 16 serial port clocks.
The result is operation of a continuous 256-bit sequence with a unique

selection pattern for both the transmitter and the receiver.
2-85
2.7 SPI Protocol: McBSP Clock Stop Mode
The SPI protocol is a master-slave configuration, with one master device and
one or more slave devices. The interface consists of the following four signals:
- Serial data input (also referred to as Master In − Slave Out, or MISO)
- Serial data output (also referred to as Master Out − Slave In, or MOSI)
- Shift-clock (also referred to as SCK)
- Slave-enable signal (also referred to as SS)
A typical SPI interface with a single slave device is shown in Figure 2−61.
Figure 2−61. Typical SPI Interface
SPI-compliant SPI-compliant
master slave
SCK SCK
MOSI MOSI
MISO MISO
SS SS
The master device controls the flow of communication by providing shift-clock

and slave-enable signals. The slave-enable signal is an optional active-low
signal that enables the serial data input and output of the slave device (device
not outputting clock). In the absence of a dedicated slave-enable signal,
communication between the master and slave is determined by the presence
or absence of an active shift-clock. In such a configuration, the slave device
must remain enabled at all times, and multiple slaves cannot be used.
The clock stop mode of the McBSP provides compatibility with the SPI
protocol. When the McBSP is configured in clock stop mode, the transmitter
and receiver are internally synchronized, so that the McBSP functions as an
SPI master or slave device. The transmit clock signal (BCLKX) corresponds
to the serial clock signal (SCK) of the SPI protocol, while the transmit
frame-synchronization signal (BFSX) is used as the slave-enable signal (SS).
The receive clock signal (BCLKR) and receive frame-synchronization signal
(BFSR), are not used in the clock stop mode, since these signals are internally
connected to their transmit counterparts, BCLKX and BFSX.
2-86
When the McBSP is configured as a master, the transmit output signal (BDX)
is used as the MOSI signal of the SPI protocol, and the receive input signal
(BDR) is used as the MISO signal. An SPI interface with the McBSP used as
the master is shown in Figure 2−62.
Figure 2−62. SPI Configuration: McBSP as the Master
McBSP-master SPI-compliant
slave
BCLKX SCK
BDX MOSI
BDR MISO
BFSX SS
Similarly, when the McBSP is configured as a slave, BDX is used as the MISO
signal and BDR is used as the MOSI signal. An SPI interface with the McBSP
used as a slave is shown in Figure 2−63.
Figure 2−63. SPI Configuration: McBSP as the Slave
McBSP-slave SPI-compliant
master
BCLKX SCK
BDX MISO
BDR MOSI
BFSX SS
2-87
2.7.1 Clock Stop Mode Configuration and Signal Descriptions
The CLKSTP bit-field of serial port control register 1 (SPCR1) and the CLKXP
bit of the pin configuration register (PCR) are used to configure the clock stop
mode. The CLKSTP bit-field enables clock stop mode and selects one of two
possible timing variations, while the CLKXP bit configures the polarity of the
BCLKX signal. Together, these bits provide four possible clock-stop mode con-
figurations, as shown in Table 2−29.
Table 2−29. Clock-Stop Mode Configurations
CLKSTP CLKXP Clock Scheme

0X X Clock stop mode disabled. Clock enabled for non-SPI mode
10 0 Low inactive state without delay: The McBSP transmits data on the rising edge of CLKX
and receive data on the falling edge of CLKR.
11 0 Low inactive state with delay: The McBSP transmits data one-half cycle ahead of the
rising edge of CLKX and receives data on the rising edge of CLKR.
10 1 High inactive state without delay: The McBSP transmits data on the falling edge of CLKX
and receives data on the rising edge of CLKR.
11 1 High inactive state with delay: The McBSP transmits data one-half cycle ahead of the
falling edge of CLKX and receives data on the falling edge of CLKR.
The CLKXM bit-field of the PCR designates the McBSP as a master or a slave
by configuring the BCLKX signal in output mode (master), or input mode
(slave). The timing diagrams for the four possible clock stop mode configura-
tions are shown in Figure 2−64 through Figure 2−67.
Figure 2−64. SPI Transfer With CLKSTP = 10b, and CLKXP = 0
ÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁ
Á Á
BCLKX/SCK
BDX or BDR/MOSI
Á B7
Á B6 B5 B4 B3 B2 B1 B0
Á Á
(from master)†
Á Á
BDX or BDR/MISO
B7 B6 B5 B4 B3 B2 B1 B0
(from slave)‡
ÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁ
BFSX/SS
ÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁ
† If the McBSP is the SPI master (CLKXM = 1), MOSI = BDX. If the McBSP is the SPI slave (CLKXM = 0), MOSI = BDR.
‡ If the McBSP is the SPI master (CLKXM = 1), MISO = BDR. If the McBSP is the SPI slave (CLKXM = 0), MISO = BDX.
2-88
ÁÁ Á
BCLKX/SCK
BDX or BDR/MOSI
ÁÁ B7
ÁB6 B5 B4 B3 B2 B1 B0
ÁÁ Á
(from master)†
ÁÁ Á
BDX or BDR/MISO
B7 B6 B5 B4 B3 B2 B1 B0
(from slave)‡
BFSX/SS
BCLKX/SCK
ÁÁ Á
ÁÁ Á
BDX or BDR/MOSI
(from master)†
B7 B6 B5 B4 B3 B2 B1 B0
BDX or BDR/MISO
ÁÁ Á
B7 B6 B5 B4 B3 B2 B1 B0
(from slave)‡
BFSX/SS
ÁÁ Á
BCLKX/SCK
BDX or BDR/MOSI
ÁÁ ÁB6 B4 B3
ÁÁ Á
(from master)† B7 B5 B2 B1 B0
ÁÁ Á
BDX or BDR/MISO B7 B6 B5 B4 B3 B2 B1 B0
(from slave)‡
BFSX/SS ÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁ
† If the McBSP is the SPI master (CLKXM = 1), MOSI=BDX. If the McBSP is the SPI slave (CLKXM = 0), MOSI = BDR.
‡ If the McBSP is the SPI master (CLKXM = 1), MISO=BDR. If the McBSP is the SPI slave (CLKXM = 0), MISO = BDX.
2-89
Notice that the frame-synchronization signal used in clock stop mode is active
throughout the entire transmission. The frame-synchronization waveform is
described in more detail in the next two sections. Although the timing diagrams
show 8-bit transfers, the packet length can be set to 8-, 12-, 16-, 20-, 24-, or
32-bits per packet. The packet length is selected by the receive word length
(RWDLEN1) in receive control register 1 (RCR1), and the transmit word length
(XWDLEN1) in transmit control register 1 (XCR1). Note that for clock stop
mode, the values of RWDLEN1 and XWDLEN1 must be set to the same value
since the McBSP transmit and receive circuits are synchronized to a single
clock. The bit-fields required to configure the McBSP as an SPI device are
shown in Table 2−30.
Table 2−30. Register Bit Values for SPI Mode Configuration

Bit-Field Value Description Register Reference
CLKSTP 1xb Enables clock stop mode, and selects one of SPCR1 Table 2−4,
two timing variations. page 2-7
CLKXP 0 or 1 Configures the BCLKX signal polarity. PCR Table 2−6,

page 2-12
CLKXM 0 or 1 Configures the BCLKX signal as an input (slave) PCR Table 2−6,
or an output (master). page 2-12
RWDLEN1 000−101b Configures the receive packet length. Must be RCR1 Table 2−7,
equal to XWDLEN1. page 2-16
XWDLEN1 000−101b Configures the transmit packet length. Must be XCR1 Table 2−9,
equal to RWDLEN1. page 2-19
2.7.2 McBSP Operation as an SPI Master

When the McBSP functions as the SPI master, it controls the transmission of
data by producing the serial clock signal. The clock signal on the BCLKX pin
is only enabled during packet transfers. When packets are not being trans-
ferred, the BCLKX pin remains high or low depending on the polarity used. An
example of an SPI interface with the McBSP used as the master is shown in
Figure 2−62. For SPI master operation, the BCLKX pin must be configured as
an output. The sample rate generator is then used to derive the BCLKX signal
from the CPU clock. The clock stop mode internally connects the BCLKX pin
to the BCLKR signal so that no external signal connection is required on the
BCLKR pin, and both the transmit and receive circuits are clocked by the mas-
ter clock (BCLKX).
The McBSP can also provide a slave-enable signal (SS) on the BFSX pin. If
a slave-enable signal is required, the BFSX pin must be configured as an out-
2-90
put, and the frame generator must be configured so that a frame-synchroniza-
tion pulse is generated each time a packet is transmitted. The data delay pa-
rameters of the McBSP (XDATDLY and RDATDLY) must be set to 1 for proper
SPI master operation, since data delay values of 0 or 2 are undefined during
this mode. The polarity of the BFSX pin is programmable high or low; however,
in most cases the pin should be configured active-low.
In this configuration, the frame generator bit fields (FPER and FWID) in the
sample rate generator registers are overridden, and custom frame-synchro-
nization waveforms are not allowed. The resulting waveform produced on the
BFSX pin is shown in Figure 2−64 though Figure 2−67. The signal becomes
active before the first bit of a packet transfer, and remains active until the last
bit of the packet is transferred. After the packet transfer is complete, the BFSX
signal returns to the inactive state.
It is important to note that even if multiple words are consecutively transferred,

the BCLKX signal is always stopped and the BFSX signal returns to the inac-
tive state after a packet transfer. When consecutive packet transfers are per-
formed, this leads to a minimum idle time of two bit-periods between each
packet transfer. The register bit values required to configure the McBSP as a
master are listed in Table 2−31.
Table 2−31. Register Bit Values for SPI Master Operation

Bit-Field Value Description Register Reference
CLKXM 1 Configures the BCLKX pin as an output. PCR Table 2−6,
page 2-12
CLKSM 1 Sample rate clock derived from CPU clock. SRGR2 Table 2−18,
page 2-60
CLKGDV 1−255 Defines the divide factor for the sample rate clock. SRGR1 Table 2−17,
page 2-59
FSXM 1 Configures the BFSX pin as an output. PCR Table 2−6,

page 2-12
FSGM 0 The BFSX signal is activated for each packet trans- SRGR2 Table 2−18,
fer. page 2-60
FSXP 1 Configures the BFSX pin as active-low. PCR Table 2−6,

page 2-12
XDATDLY 01b Provides correct setup time on the BFSX signal. XCR2 Table 2−10,
page 2-20
RDATDLY 01b Provides correct setup time on the BFSX signal. RCR2 Table 2−8,
page 2-17
2-91
2.7.3 McBSP Operation as an SPI Slave
When the McBSP is used as an SPI slave, the master clock and slave-enable
signals are generated externally by a master device. Accordingly, the BCLKX
and BFSX pins must be configured as inputs. The BCLKX pin is internally con-
nected to the BCLKR signal, so that both the transmit and receive circuits of
the McBSP are clocked by the external master clock. The BFSX pin is also in-
ternally connected to the BFSR signal, and no external signal connections are
required on the BCLKR and BFSR pins. An example of an SPI interface with
the McBSP used as the slave is shown in Figure 2−63.
Although the BCLKX signal is generated externally by the master and is asyn-
chronous to the McBSP, the sample rate generator of the McBSP must be en-
abled for proper SPI slave operation. The sample rate generator should be
programmed to its maximum rate of half the CPU clock rate. The internal sam-
ple rate clock is then used to synchronize the McBSP logic to the external mas-
ter clock and slave-enable signals.
The McBSP requires an active edge of the slave-enable signal on the BFSX
input for each transfer. This means that the master device must assert the
slave-enable signal at the beginning of each transfer, and deassert the signal
after the completion of each packet transfer; the slave-enable signal cannot
remain active between transfers. The data delay parameters of the McBSP
must be set to 0 for proper SPI slave operation, and values of 1 or 2 are unde-
fined during this mode of operation. The register bit values required to config-
ure the McBSP as a slave are listed in Table 2−32.
Table 2−32. Register Bit Values for SPI Slave Operation
Bit Field Value Description Register Reference

CLKXM 0 Configures the BCLKX pin as an input. PCR Table 2−6,
page 2-12
CLKSM 1 Sample rate clock derived from CPU clock. SRGR2 Table 2−18,
page 2-60
CLKGDV 1 Select a divide factor of 2 for the sample rate clock. SRGR1 Table 2−17,
page 2-59
FSXM 0 Configures the BFSX pin as an input. PCR Table 2−6,
page 2-12
FSGM 0 The BFSX signal is activated for each packet trans- SRGR2 Table 2−18,
fer. page 2-60
FSXP 1 Configures the BFSX pin as active-low. PCR Table 2−6,
page 2-12
2-92
Bit Field Value Description Register Reference
XDATDLY 0 Must be 0 for SPI slave operation. XCR2 Table 2−10,
page 2-20
RDATDLY 0 Must be 0 for SPI slave operation. RCR2 Table 2−8,
page 2-17
2.7.4 McBSP Initialization for SPI Mode

The operation of the McBSP during device reset, transmitter reset, or receiver
reset is described in section 2.3.1, Resetting the Serial Port, on page 2-22. To
configure the McBSP for SPI master or slave operation, the following steps
must be performed:
1) Set the transmitter reset bit (XRST) in serial port control register 2
(SPCR2) to zero, to reset the transmitter. Set the receiver reset bit (RRST)
in serial port control register 1 (SPCR1) to zero, to reset the receiver.
(XRST=RRST=0)
2) Program the McBSP register fields shown in Table 2−30. Also program the
register fields shown in Table 2−31 or Table 2−32, depending on whether
the McBSP is the master or the slave. All other McBSP register fields can
be programmed to their default values.
3) Set the sample rate generator reset bit (GRST) in serial port control regis-
ter 2 (SPCR2) to one, to release the sample rate generator from reset.
Note that the value written to SPCR2 should have only the GRST bit
changed to one, and the remaining bit-fields should have the same values
written in Step 2.
(GRST=1)
4) Wait for the duration of two sample rate generator clock-periods for the
McBSP logic to stabilize.
5) Depending on whether the CPU or DMA controller services the McBSP

transmit and receive buffers, either Step(a) or Step (b) should be
performed. If the CPU services the McBSP buffers, Step (a) should be
performed, or if the DMA controller services the McBSP buffers, Step (b)
should be performed.
a) CPU services the McBSP buffers. Set the XRST and RRST bit-fields
to one, to enable the transmitter and receiver. Note that the values
written to SPCR1 and SPCR2 should have only the reset bits changed
to one, and the remaining bit-fields should have the same values writ-
ten in Step 2.
2-93
(XRST=RRST=1)
2-94
b) DMA controller services the McBSP buffers. First, configure the DMA
controller and enable the channels that service the McBSP buffers.
Then set the XRST and RRST bit-fields to one, to enable the transmit-
ter and receiver. Note that the values written to SPCR1 and SPCR2
should have only the reset bits changed to one, and the remaining bit-
fields should have the same values written in Step 2.
(XRST=RRST=1)
6) Wait for the duration of two sample rate generator clock-periods for the
McBSP logic to stabilize.
2-95
2.8 Emulation FREE and SOFT Bits
FREE and SOFT are special emulation bits that determine the state of the seri-
al port clock when a breakpoint is encountered in the high-level language de-
bugger. If the FREE bit is set to one, then upon a software breakpoint, the clock
continues to run (free runs) and data still shifts out. When FREE = 1, the SOFT
bit is a don’t care. If the FREE bit is cleared to zero, then the SOFT bit takes
effect. If the SOFT bit is cleared to zero, then the clock stops immediately, thus
aborting a transmission. If the SOFT bit is set to one and a transmission is in
progress, the transmission continues until completion of the transfer, and then
the clock halts. These options are listed in Table 2−33.
The receiver-side functions in a similar fashion. Note that if an option other

than immediate stop (SOFT = FREE = 0) is chosen, the receiver continues
running and an overflow error is possible.
Table 2−33. McBSP Clock Configuration
FREE SOFT McBSP Clock Configuration

0 0 Immediate stop, clocks are stopped. (Reset value)
0 1 Transmitter stops after completion of the current word. The receiver is not affected.
1 don’t Free run.

care
2-96
2.9 McBSP Pins as General Purpose I/O
Two conditions allow the serial port pins (CLKX, FSX, DX, CLKR, FSR, and
DR) to be used as general purpose I/O rather than serial port pins:
1) The related portion (transmitter or receiver) of the serial port is in reset;

(R/X)RST = 0 in SPCR[1,2].
2) General purpose I/O is enabled for the related portion of the serial port;
(R/X)IOEN = 1 in the PCR.
Figure 2−4, Pin Control Register (PCR), on page 2-12, has bits that configure
each of the McBSP pins as general purpose inputs or outputs. Table 2−34
shows how this is achieved. In the case of FS(R/X), FS(R/X)M=0 configures
the pin as an input and FS(R/X)M = 1 configures that pin as an output. When
configured as an output, the value driven on FS(R/X) is the value stored in
FS(R/X)P. If configured as an input, FS(R/X)P becomes a read-only bit that re-
flects the status of that signal. CLK(R/X)M and CLK(R/X)P work similarly for
CLK(R/X). When the transmitter is selected as general purpose I/O, the value
of the DX_STAT bit in the PCR is driven onto DX. DR is always an input and
its value is held in the DR_STAT bit in the PCR. To configure CLKS as a general
purpose input, both the transmitter and receiver must be in reset state and
(R/X)IOEN = 1, because CLKS is always an input to the McBSP and affects
both transmit and receive operations.
2-97
Table 2−34. Configuration of Pins as General Purpose I/O
General Purpose I/O
Enabled by Selected as Output Value Selected as Input Value
Pin Setting Both Output Driven From Input Readable On
CLKX XRST = 0 CLKXM = 1 CLKXP CLKXM = 0 CLKXP

XIOEN = 1
FSX XRST = 0 FSXM = 1 FSXP FSXM = 0 FSXP

XIOEN = 1
DX XRST = 0 Always DX_STAT Never Does not apply

XIOEN = 1
CLKR RRST = 0 CLKRM = 1 CLKRP CLKRM = 0 CLKRP

RIOEN = 1
FSR RRST = 0 FSRM = 1 FSRP FSRM = 0 FSRP

RIOEN = 1
DR RRST = 0 Never Does not apply Always DR_STAT

RIOEN = 1
CLKS RRST = XRST = 0 Never Does not apply Always CLKS_STAT

RIOEN = XIOEN = 1
2-98
2.10 McBSP Operation in Power-Down Mode
’54x devices offer several power-down modes which allow all or part of the de-
vice to enter a dormant state, and consequently, dissipates considerably less
power than when running normally. Power-down modes may be invoked in
several ways, including executing the IDLE instruction or driving the HOLD in-
put low with the HM status bit set to one. The McBSP, like other peripherals,
can take the CPU out of IDLE using a transmit or receive interrupt.
When in IDLE1 or HOLD modes, the McBSP continues to operate normally

with no restrictions.
In IDLE2 or IDLE3 modes, the internal device clocks provided to the peripher-
als are stopped. Consequently, some limitations are placed on operating the
McBSP in these modes. If external clock and frame-sync are provided, the
McBSP can continue to operate, and receive and transmit interrupts can be
used to exit the IDLE state. If either clocks or frame-syncs are internal (pro-
vided by the clock and frame-sync generator), the McBSP will stop in IDLE2/3
because there are no internal clocks available to operate the clock and frame-
sync generator.
In IDLE2/3, the internal clocks to the McBSP and the DMA controller are
started automatically when a transfer begins, and stopped after the transfer
is completed. This feature provides additional power savings since neither
module is active when no activity is occurring.
In an autobuffering configuration, the McBSP and the DMA can be operated

such that the CPU remains in the IDLE state until a half- or full-buffer of data
has been transferred. The McBSP receive and transmit interrupts should be
disabled to avoid interrupt generation on the reception/transmission of each
word. The DMA should be configured in autobuffering (ABU) mode and DMA
transfers should be triggered on McBSP receive and transmit events. The
DMA should then be configured to interrupt the CPU based on completion of
half-buffers or full-buffers, whichever is desired. Since the CPU remains in the
IDLE state until an interrupt occurs, an entire block of data can be transferred
without causing the CPU to exit the IDLE state. For more information on the
DMA operation, see Chapter 3, Direct Memory Access (DMA) Controller.
2-99
2.11 McBSP Programming Example Code
McBSP programming example code is presented in Chapter 3 as follows:
Example 3−19, McBSP Data Transfer in ABU Mode, is shown on page 3-53;
Example 3−20, McBSP Data Transfer in Double-Word Mode, is shown on
page 3-55; and Example 3−21, McBSP to Data Memory Transfer With Data
Sorting, is shown on page 3-57.
2-100
2-101
Chapter 3
Direct Memory Access (DMA) Controller
The direct memory access (DMA) controller is available on selected members

of the ’54x family of devices. This chapter describes the configuration and op-
eration of the DMA available on the ’5402, ’5410, and ’5420 devices.
Topic Page
3.1 DMA Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-2

3.2 DMA Operation and Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-4
3.3 Extended Addressing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-34
3.4 DMA Memory Maps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-35
3.5 DMA Transfer Latency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-40
3.6 Enhanced Host Port Interface Access Through the DMA . . . . . . . . 3-43
3.7 Interprocessor FIFO Communication on the ’5420 . . . . . . . . . . . . . . 3-44
3.8 DMA Operation in Power-Down Mode . . . . . . . . . . . . . . . . . . . . . . . . . . 3-44
3.9 Programming Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3-45
3-1
3.1 DMA Overview
The direct memory access (DMA) controller transfers data between regions
in the memory map without intervention by the CPU. The DMA allows move-
ment to and from internal memory, internal peripherals, or external devices to
occur in the background of CPU operation. The DMA has six independent pro-
grammable channels allowing six different contexts for DMA operation. The
DMA controller also services requests from the host port interface peripherals
(HPI-8 or HPI-16) to utilize the DMA bus. Throughout this chapter, the abbrevi-
ation HPIx is used to denote either the HPI-8 or HPI-16.
The terms used in discussing DMA operations are defined as follows:
- Read transfer. The DMA reads a data element from a source location in
memory. The source may be memory or a peripheral device, and may be
located in program, data, or I/O space.
- Write transfer. The DMA writes the data elements that were read during
a read transfer to its destination in memory location. The destination may
be memory or a peripheral device, and may be located in program, data,
or I/O space.
- Element transfer. The combined read and write transfer for a single data
element.
- Frame transfer. Each DMA channel has an independently programmable
number of elements per frame. In completing a frame transfer, the DMA
moves all elements in a single frame.
- Block Transfer. Each DMA channel also has an independently program-
mable number of frames per block. In completing a block transfer, the
DMA moves the total number of frames defined for the block.
The DMA has the following features:
- Background operation. The DMA operates independently of the CPU.

- Six channels. The DMA can keep track of the context of six independent
block transfers.
- Host port interface access. The DMA provides a path to the memory for
host port interface data without intervention from the CPU. This allows the
host port interface access to a greater memory space (the entire on-chip
memory) than ever before.
- Multiframe transfer. Each block transfer can consist of multiple frames
of a programmable size. See section 3.2.3.3, DMA Sync Event and Frame
Count Registers, on page 3-14.
- Programmable priority. Each channel can be programmed indepen-
dently for high or low priority. See section 3.2.2, DMA Channel Priority and
Enable Control Register, on page 3-7.
3-2
- Programmable address generation. Each channel’s source and des-
tination address registers can have configurable indexes for each read
and write transfer. The address may remain constant, increment, decre-
ment, or be adjusted by a programmable value. These programmable val-
ues allow a separate index for the last transfer in a frame from the preced-
ing transfers, and can be used to achieve data sorting (discussed later in
this chapter).
- Full address range. The DMA can access the full extended address
range implemented on the device. The DMA access extends to the follow-
ing:
J On-chip memory
J On-chip peripherals
J External memory (selected devices). Access to certain areas of the
memory map may be restricted depending on the device. For more in-
formation, see Chapter 3, Memory, in the the user guide titled
TMS320C54x DSP, CPU and Peripherals, Reference Set Volume 1
(literature number SPRU131).
- Programmable width transfers. Each channel can be independently
configured to transfer in single-word mode (16-bit), or double-word mode
(32-bit). See section 3.2.3.3, DMA Sync Event and Frame Count Regis-
ters, on page 3-14.
- Autoinitialization. Once a block transfer is complete, a DMA channel
may automatically re-initialize itself for the next block transfer. See section
3.2.3.4, Transfer Mode Control Register, on page 3-18.
- Event synchronization. Each element transfer may be initiated by se-
lected events. See section 3.2.3.3, DMA Sync Event and Frame Count
Registers, on page 3-14.
- Interrupt generation. On completion of each frame transfer or block
transfer, each DMA channel can send an interrupt to the CPU. See section
3.2.2.2, Multiplexed Interrupt Control, on page 3-9.
3-3
3.2 DMA Operation and Configuration
The DMA configuration and operation is achieved by a set of memory-mapped
control registers. This scheme is similar to that used by other ’54x device pe-
ripherals. The DMA registers are memory-mapped using a register subad-
dressing scheme.
3.2.1 Register Subaddressing

Register subaddressing involves multiplexing a set of registers to a single
location in the memory map. A subbank addressing register is used to control
the multiplexer. A subbank data register is used to actually read or write data
to the desired subaddressed register. A visual representation of this arrange-
ment is shown in Figure 3−1. In this way, a large number of registers can be
mapped into a small memory space.
Figure 3−1. Register Subaddressing
DMSDI
DMSDN
Subaddressed
registers
DMSA
To access a specific subaddressed register, first the desired subaddress is

written to the subbank address register (DMSA). This directs the multiplexer
to connect the subbank access registers (DMSDI and DMSDN) to the desired
physical location. When an access (read or write) is made to the subbank data
registers, the data is moved to or from the desired embedded data register as
previously specified in the subbank address register.
3-4
The DMA controller provides two types of subbank access registers. The first,
DMSDI, is a subbank access register with automatic increment of the subad-
dress. This means that the subaddress will be incremented automatically after
each access. If the entire set of DMA configuration registers needs to be con-
figured, DMSDI provides a convenient means to set the subbank address
once and then postincrement it after each access until the entire set of regis-
ters has been configured. If a single register access is desired, or if there is a
need to prevent modification of the subaddress, DMSDN is used. DMSDN pro-
vides access to the subaddressed register without modifying the subaddress.
In this way, the entire register set used to program the DMA is contained in four
memory-mapped register locations.
Examples of the use of subaddressing are shown in Example 3−1 and

Example 3−2.
Example 3−1. Using Register Subaddressing Without Autoincrement
The following code sample illustrates how the DMA channel 5 source address
register is initialized with the value 1000h.
DMSA .set 55h ;subbank address register location
DMSDN .set 57h
DMSRC5 .set 19h
STM DMSRC5,DMSA ;initialize the subbank address
;register to point to DMSRC5
STM #1000h,DMSDN ;write 1000h to DMSRC5
Example 3−2. Using Register Subaddressing With Autoincrement

The following code sample illustrates how DMA channel 5 context registers
can be initialized. The values written to the registers in this example are arbi-
trary and do not represent a particular configuration.
DMSA .set 55h ;set register locations
DMSDI .set 56h
DMSRC5 .set 19h
DMDST5 .set 1Ah
DMCTR5 .set 1Bh
DMSFC5 .set 1Ch
DMMCR5 .set 1Dh
STM DMSRC5,DMSA ;initialize the subbank address
;register to point to DMSRC5
STM #1000h,DMSDI ;write 1000h to DMSRC5
3-5
Example 3−2 Using Register Subaddressing With Autoincrement (Continued)
STM #2000h,DMSDI ;write 2000h to DMDST5
STM #0010h,DMSDI ;write 10h to DMCTR5
STM #0002h,DMSDI ;write 2h to DMSFC5
STM #0000h,DMSDI ;write 0h to DMMCR5
The DMA registers are shown in Table 3−1. Only the channel priority and en-
able control register (DMPREC) is directly addressed. All other DMA registers
are subaddressed.
Table 3−1. DMA Registers
Address SubAddress Name Function
54h — DMPREC Channel Priority and Enable Control Register
55h — DMSA Subbank Address Register
56h — DMSDI Subbank Access Register With Autoincrement
57h — DMSDN Subbank Access Register Without Autoincrement
— 00h DMSRC0 Channel 0 Source Address Register
— 01h DMDST0 Channel 0 Destination Address Register
— 02h DMCTR0 Channel 0 Element Count Register
— 03h DMSFC0 Channel 0 Sync Select and Frame Count Register
— 04h DMMCR0 Channel 0 Transfer Mode Control Register
— 0Ah DMSRC2 Channel 2 Source Address Register
— 0Bh DMDST2 Channel 2 Destination Address Register
— 0Ch DMCTR2 Channel 2 Element Count Register
— 0Dh DMSFC2 Channel 2 Sync Select and Frame Count Register
— 0Eh DMMCR2 Channel 2 Transfer Mode Control Register
— 0Fh DMSRC3 Channel 3 Source Address Register
3-6
Table 3−1. DMA Registers (Continued)
Address SubAddress Name Function
— 1Ah DMDST5 Channel 5 Destination Address Register
— 1Bh DMCTR5 Channel 5 Element Count Register
— 1Ch DMSFC5 Channel 5 Sync Select and Frame Count Register
— 1Dh DMMCR5 Channel 5 Transfer Mode Control Register
— 1Eh DMSRCP Source Program Page Address (all channels)
— 1Fh DMDSTP Destination Program Page Address (all channels)
— 20h DMIDX0 Element Address Index Register 0
— 21h DMIDX1 Element Address Index Register 1
— 22h DMFRI0 Frame Address Index Register 0
— 23h DMFRI1 Frame Address Index Register 1
— 24h DMGSA Global Source Address Reload Register
— 25h DMGDA Global Destination Address Reload Register
— 26h DMGCR Global Element Count Reload Register
— 27h DMGFR Global Frame Count Reload Register
3.2.2 DMA Channel Priority and Enable Control Register
The channel priority and enable control (DMPREC) register controls several
functions of the overall operation of the DMA system including:
- Selective enabling of each of the DMA channels

- Control of the multiplexed interrupts
- Control of channel priorities
The structure of DMPREC is shown in Figure 3−2. DMPREC is located at ad-

dress 0054h in data space and is not subaddressed. Following reset,
DMPREC is initialized to 0000h.
Figure 3−2. DMA Channel Priority and Enable Control (DMPREC) Register
15 14 13 8 7 6 5 0
FREE RSVD DPRC INTOSEL DE[5:0]
3-7
Table 3−2. DMA Channel Priority and Enable Control (DMPREC) Register Bit/Field
Descriptions
Reset
Bit Name Value Function
15 FREE 0 This bit controls the behavior of the DMA controller during emulation. When
FREE = 0, DMA transfers are suspended when the emulator stops. When
FREE = 1, DMA transfers continue even during emulation stop.
14 RSVD 0 Reserved. Values written to this field have no effect.
13 DPRC[5] 0 DMA channel 5 priority control bit.
DPRC[5] = 1 High priority
DPRC[5] = 0 Low priority
7−6 INTOSEL 0 Interrupt multiplex control bits. The INTOSEL bits control how the DMA
interrupts will be assigned in the interrupt vector table and IMR/IMF registers.
The effects of this field on the operation are device-specific (refer to Table 3−3,
Table 3−4, and Table 3−5.)
5 DE[5] 0 DMA channel 5 enable bit.
DE[5] = 1 Enables DMA channel 5
DE[5] = 0 Disables DMA channel 5
3-8
Table 3−2. DMA Channel Priority and Enable Control (DMPREC) Register Bit/Field
Descriptions (Continued)
Reset
3 DE[3] 0 DMA Channel 3 enable bit.
2 DE[2] 0 DMA Channel 2 enable bit.
3.2.2.1 DMA Channel Enable Control
Each of the six DMA channels can be independently enabled through the DE
field in the DMPREC. Bits 0−5 in this register correspond to each of the six
DMA channels, with bit 0 representing channel 0, bit 1 representing channel
1, etc. A logic 1 enables the associated channel, and a logic 0 disables the
channel.
Each enable bit can be reset by the DMA controller upon completion of a block
transfer. DMPREC can be polled to determine when a block transfer on a given
channel is complete. If the DMA controller and the CPU both attempt to change
the state of the DE field bits, the DMA controller has priority.
3.2.2.2 Multiplexed Interrupt Control
DMA events can trigger interrupts to the CPU upon completion of a transfer.
Due to a limited number of interrupts in the ’54x memory map, some DMA inter-
rupts are multiplexed with other peripheral interrupts on the device. The inter-
rupt functions available are controlled by the INTOSEL field (bits 7 and 6) in
DMPREC. The INTOSEL field can configure up to 6 interrupts to be assigned
to DMA interrupts, depending on the device being used.
The DMA interrupt assignments available on each device are shown in

Table 3−3, Table 3−4, and Table 3−5. Following reset, the interrupts are con-
figured as in the 00b column in each table.
3-9
Table 3−3. Multiplexed Interrupt Assignments for the ’5402
Interrupt
Number INTOSEL [1:0] Value
(IMR/IFR #) 00b 01b 10b 11b
6 Reserved Reserved DMA Channel 0 Interrupt Reserved
7 Timer 1 interrupt Timer 1 interrupt DMA Channel 1 Interrupt Reserved
10 McBSP 1 RINT DMA Channel 2 Interrupt DMA Channel 2 Interrupt Reserved
11 McBSP 1 XINT DMA Channel 3 Interrupt DMA Channel 3 Interrupt Reserved
Table 3−4. Multiplexed Interrupt Assignments for the ’5410

Interrupt
(IMR/IFR #) 00b 01b 10b 11b
4 McBSP 0 RINT McBSP 0 RINT McBSP 0 RINT Reserved
5 McBSP 0 XINT McBSP 0 XINT McBSP 0 XINT Reserved
6 McBSP 2 RINT McBSP 2 RINT DMA Channel 0 Interrupt Reserved
7 McBSP 2 XINT McBSP 2 XINT DMA Channel 1 Interrupt Reserved
12 DMA Channel 4 Interrupt DMA Channel 4 Interrupt DMA Channel 4 Interrupt Reserved
Table 3−5. Multiplexed Interrupt Assignments for the ’5420 (each subsystem)
Interrupt
(IMR/IFR #) 00b 01b 10b 11b
4 McBSP 0 RINT McBSP 0 RINT McBSP 0 RINT Reserved
5 McBSP 0 XINT McBSP 0 XINT McBSP 0 XINT Reserved
6 McBSP 2 RINT McBSP 2 RINT DMA Channel 0 Interrupt Reserved
7 McBSP 2 XINT McBSP 2 XINT DMA Channel 1 Interrupt Reserved
3-10
Table 3−5. Multiplexed Interrupt Assignments for the ’5420 (each subsystem)
Interrupt
(IMR/IFR #) 00b 01b 10b 11b
3-11
3.2.2.3 DMA Channel Priority Control
Each DMA channel can be independently assigned low or high priority. All high
priority DMA channels are serviced before any low priority channels are servi-
ced. When multiple channels are enabled and assigned the same priority lev-
el, each of the enabled channels are serviced in a circular pattern.
Priority level for each channel is set in the 6-bit DPRC field of the DMPREC.
Each bit position is associated with a DMA channel. Bits 8 through 13 are asso-
ciated with channels 0 though 5, respectively. A logic 0 in the bit position as-
signs low priority to the associated channel. A logic 1 assigns high priority to
the associated channel.
3.2.2.4 Emulation Control for DMA Functions

The FREE field of DMPREC controls the behavior of the DMA controller when
the emulator is stopped. If FREE is cleared, DMA transfers are suspended
when the emulator stops. If FREE is set, DMA transfers continue even if the
emulator stops.
3.2.3 Channel-Context Registers

Each DMA channel has a set of 5 channel-context registers which configure
the operation for only that channel. These five registers are:
- Source address register
- Destination address register
- Element count register
- Sync event / frame count register
- Transfer mode control register
The operation and effects of each of these registers is discussed in detail in

the sections that follow.
3.2.3.1 Source and Destination Address Registers

Each DMA element transfer involves a read followed by a write. The address
of the data to be read is called the source address, and the address where the
data will be written is called the destination address. Two channel-context reg-
isters per channel store the source and destination addresses: the DMA chan-
nel n source address register (DMSRCn) and the DMA channel n destination
address register (DMDSTn). Each is a 16-bit register and contains the least
significant 16 bits of the extended address of the source or destination loca-
tion.
3-12
The source and destination address registers are initialized prior to starting the
DMA transfer in software, and updated automatically during transfers by the
DMA controller. If necessary, the CPU can read either of these addresses dur-
ing transfers to monitor the status of the transfer. These registers can also be
written by the CPU during transfers. Changes to the address registers made
during block transfers are effective immediately and will affect the progress of
the transfer. Care should be taken to prevent unintentional writes to these reg-
isters during transfers.
The locations of the source and destination address registers for each channel
are shown in Table 3−1.
3.2.3.2 Element Count Register
Each DMA Channel has a 16-bit element counter that keeps track of the num-
ber of DMA transfers to be performed. The element counter is initialized with
the 16-bit unsigned number contained in the DMA channel element count reg-
ister (DMCTRn) that represents the number of elements to be transferred. The
element count register should be initialized with one less than the desired num-
ber of element transfers. For example, if DMA channel 2 is to move nine data-
elements, DMCTR2 should be initialized to eight.
In normal multi-frame transfer mode, the DMCTRn is automatically decrem-

ented after each transfer by the DMA controller. When the last element in each
frame is reached, the element counter is reloaded with the original contents
of the DMCTRn, which has been stored in a shadow register.
In autobuffering (ABU) mode, the contents of the DMCTRn contains the buffer
size. The DMCTRn is not decremented during transfers in this mode. ABU
mode is included in the DMA controller to perform the same function as the
autobuffering units of the ’54x buffered serial port (BSP). The McBSP, in con-
junction with the DMA in ABU mode, can mimic the BSP ABU operation. ABU
mode is discussed in more detail in section 3.2.3.5, Addressing Modes, on
page 3-22.
If the CPU and the DMA controller attempt to modify the element count register
at the same time, the CPU has priority.
The locations of the element count registers for each channel are shown in
Table 3−1.
3-13
3.2.3.3 DMA Sync Event and Frame Count Registers
The DMA sync event and frame count register (DMSFCn) serves three pur-
poses:
- Determines which synchronization events will be

used to trigger DMA transfers
- Determines the transfer word size (16-bit or 32-bit)
- Determines the number of frames to be transferred
The structure of the DMSFCn is shown in Figure 3−3. The register locations
of the DMA sync event and frame count registers for each channel are shown
in Table 3−1.
Figure 3−3. DMA Sync Event and Frame Count (DMSFCn) Register
15 12 11 10 8 7 0
DSYN[3:0] DBLW rsvd Frame Count
Table 3−6. DMA Sync Event and Frame Count (DMSFCn) Register Bit/Field Descriptions
Reset
15−12 DSYN[3:0] 0 DMA sync event. Specifies which sync event is used to initiate DMA transfers
for the corresponding DMA channel. See the DMA Synchronization Events
section following this table for available sync events.
11 DBLW 0 Double-word mode.

DBLW = 0 Single-word mode. Each element is 16 bits.
DBLW = 1 Double-word mode. Each element is 32 bits.
10−8 rsvd 0 Reserved. Values written to this field have no effect.
7−0 Frame 0 Frame count. Specifies the total number of frames to be transferred. See the
Count Frame Count section following Table 3−9 for more details.
DMA Synchronization Events
DMA element transfers can be initiated by a variety of events including McBSP

receive or transmit events, timer interrupt events, and external interrupt
events. A DMA channel can also run unsynchronized (not associated with any
event). When a DMA channel is synchronized to a particular event, each ele-
ment transfer waits for the occurrence of that event. If the DMA channel is not
synchronized to any event, the transfers proceed as rapidly as possible.
3-14
If sync events are used, one sync event is required for each transfer. In single-
word mode, one sync event initiates one 16-bit single-word transfer. In double-
word mode, one sync event initiates one 32-bit double-word transfer (as two
16-bit transfers).
The types of sync events available depend on the ’54x device. Not all of the
sync events mentioned above are available on all devices. Table 3−6,
Table 3−7, Table 3−8, and Table 3−9 show the available sync events on each
device. The sync event is chosen by loading the DSYN field of the DMSFCn
with the appropriate value as shown in the Tables. Only one sync event can
be chosen for each DMA channel. The DSYN field is set to 0000h (no sync
event) on reset.
Table 3−7. DMA Sync Event Options on the ’5402
DSYN [3:0] value DMA Synchronization Mode

0000 No sync event (nonsynchronized operation)
0001 McBSP 0 receive event (REVT0)
0010 McBSP 0 transmit event (XEVT0)
0011 Reserved
0100 Reserved
0111 Reserved
1000 Reserved
1001 Reserved
1010 Reserved
1011 Reserved
1100 Reserved
1101 Timer 0 interrupt event
1110 External interrupt 3 (INT3) event
1111 Timer 1 interrupt event
3-15
Table 3−8. DMA Sync Event Options on the ’5410
0111 McBSP 0 receive event − ABIS mode (REVTA0)
1000 McBSP 0 transmit event − ABIS mode (XEVTA0)
1101 Timer interrupt event
1110 External interrupt 3 (INT3) event
1111 Reserved
3-16
Table 3−9. DMA Sync Event Options on the ’5420 (each subsystem)
0111 FIFO receive buffer not empty event
1000 FIFO transmit buffer not full event
1001 Reserved
1010 Reserved
1011 Reserved
1100 Reserved
1101 Reserved
1110 Reserved
1111 Reserved
Double-Word Mode
DMA transfers can be configured as single-word 16-bit transfers, or double-
word 32-bit transfers. In double-word mode, each 32-bit transfer is considered
one element. Double-word transfers are implemented as two back-to-back
16-bit transfers, and administered by the DMA controller (including appropri-
ate address modification) without any intervention from the CPU. Double-word
mode is available with any of the address indexing modes. For a detailed dis-
cussion of address indexing modes, see section 3.2.3.5, Addressing Modes,
on page 3-22.
In double-word mode, the address of the second transfer is always the ad-
dress of the first transfer with the LSB inverted. If the first transfer address is
an even address, the second transfer will be to the next higher address. If the
first transfer address is an odd address, the second transfer will be directed
to the next lower address.
3-17
Double-word mode is configured by setting the DBLW bit in the DMSFCn. On
reset, DBLW is cleared (single-word mode).
Frame Count
The frame count is an 8-bit field in the DMSFCn that specifies the number of
frames to be included in a block transfer. The frame count should be initialized
to one less than the desired number of frames. For example, if the desired
number of frames to be transferred in a block is eight, the frame count field
should be initialized to seven. If only one frame is desired, the frame count field
should be set to zero. On reset, the frame count field is set to zero (single frame
operation). The maximum number of frames is 256.
The frame count field is decremented by the DMA controller upon the comple-
tion of each frame. Once the last frame is transferred, the frame count can be
reloaded with the contents of the DMA global frame count register (DMGFR),
if autoinitialization mode is enabled. For more information on autoinitialization,
see section 3.2.3.6, Autoinitialization, on page 3-28.
The element count and frame count can be used together to allow up to 65536
transfers. The total number of transfers is the product of the element count and
the frame count.
If the CPU and the DMA controller attempt to modify the frame count field at
the same time, the CPU has priority.
3.2.3.4 Transfer Mode Control Register

The DMA transfer mode control register (DMMCRn) is a 16-bit register that
controls the transfer mode for the channel. The structure of the DMMCRn is
shown in Figure 3−4, DMA Transfer Mode Control (DMMCRn) Register, on
page 3-18. Function descriptions of the register fields are shown in Table 3−10
on page 3-18, and discussed in the sections that follow.
Figure 3−4. DMA Transfer Mode Control (DMMCRn) Register

15 14 13 12 11 10 8 7 6 5 4 2 1 0
AUTOINIT DINM IMOD CTMOD rsvd SIND DMS rsvd DIND DMD
Table 3−10. DMA Transfer Mode Control (DMMCRn) Register Bit/Field Descriptions
Reset
15 AUTOINIT 0 DMA autoinitialization mode bit.
AUTOINIT = 0 Autoinitialization is disabled
AUTOINIT = 1 Autoinitialization is enabled
3-18
(Continued)
Reset
14 DINM 0 DMA interrupt generation mask bit.
DINM = 0 No interrupt generated
DINM = 1 Interrupts generated based on IMOD bit
13 IMOD 0 DMA interrupt generation mode bit.
In ABU mode (CTMOD = 1):
IMOD = 0 Interrupt at buffer full only
IMOD = 1 Interrupt at half buffer full and buffer full
In multiframe mode (CTMOD = 0):
IMOD = 0 Interrupt at completion of block transfer
IMOD = 1 Interrupt at end of frame and end of block
12 CTMOD 0 DMA Transfer Counter Mode Control Bit.

CTMOD = 0 Multiframe mode
CTMOD = 1 ABU mode
11 Reserved 0 Reserved. Values written to this field have no effect.
10 - 8 SIND 0 DMA source address transfer index mode bit.
SIND = 000 No modification
SIND = 001 Postincrement
SIND = 010 Post-decrement
SIND = 011 Postincrement with index offset (DMIDX0)
SIND = 100 Postincrement with index offset (DMIDX1)
SIND = 101 Postincrement with index offset (DMIDX0 and DMFRI0)
SIND = 110 Postincrement with index offset (DMIDX1 and DMFRI1)
SIND = 111 Reserved
7-6 DMS 0 DMA source address space select bit.
DMS = 00 Program space
DMS = 01 Data space
DMS = 10 I/O space
DMS = 11 Reserved
5 Reserved 0 Reserved. Values written to this field have no effect.
3-19
(Continued)
Reset
4-2 DIND 0 DMA destination address transfer index mode bit.
DIND = 000 No modification
DIND = 001 Postincrement
DIND = 010 Post-decrement
DIND = 011 Postincrement with index offset (DMIDX0)
DIND = 100 Postincrement with index offset (DMIDX1)
DIND = 101 Postincrement with index offset (DMIDX0 and DMFRI0)
DIND = 110 Postincrement with index offset (DMIDX1 and DMFRI1)
DIND = 111 Reserved
1-0 DMD 0 DMA Destination Address Space Select Bit.
DMD = 00 Program space
DMD = 01 Data space
DMD = 10 I/O space
DMD = 11 Reserved
3-20
Source and Destination Address Selection and Modification
Two fields in the DMMCRn, DMS and DMD, control the selection of the source
and destination address spaces for DMA transfers. DMS specifies whether the
source data comes from program space, data space, or I/O space. DMD speci-
fies whether the destination data goes to program space, data space, or I/O
space. Bit configurations for each of these modes is shown in Table 3−10.
In addition, the DMMCRn controls how addresses are modified for the source
and destination. The SIND field controls how the source addresses are in-
dexed as transfers proceed. The DIND field controls how destination address-
es are indexed as transfers proceed. The following options are available for
address modification:
- No modification (address remains constant for each transfer)
- Postincrement by 1
- Post-decrement by 1
- Postincrement by the offset value contained in the element index register
0 (DMIDX0)
- Postincrement by the offset value contained in the element index register
1 (DMIDX1)
- Postincrement by the offset values contained in the element index register
0 (DMIDX0) and the frame index register 0 (DMFRI0)
- Postincrement by the offset values contained in the element index register
0 (DMIDX1) and the frame index register 0 (DMFRI1)
The latter four modes utilize the element and frame index registers. The ele-
ment index registers (DMIDX0 and DMIDX1) are used to index the source and
destination addresses during transfers. The index contained in these registers
is used to modify the source or destination address following the transfer of
each element. DMIDX0 and DMIDX1 are not associated with either the source
or destination address. Either register can be used to index the source, des-
tination, or both.
The frame index registers (DMFRI0 and DMFRI1) are used to index the source
and destination addresses following completion of blocks (or frames) of ele-
ment transfers. When both element and frame indexes are used, the address
is modified by the element index after each transfer, and then modified by the
frame index at the end of each frame. This capability can be used to implement
circular buffers and sorting functions (see the related sections that follow).
DMFRI0 and DMFRI1 are not associated with either the source or destination
address. Either register can be used to index the source, destination, or both.
DMMCRn is a channel-context register, so each channel can be configured
separately. DMIDX0, DMIDX1, DMFRI0 and DMFRI1 are not channel-context
registers. These registers configure index options for the entire DMA system.
3-21
3.2.3.5 Addressing Modes
The CTMOD field in DMMCRn controls the operation of the DMA transfer
counter for each channel. In multiframe mode, the element and frame indexes
are used to modify the source and destination addresses following each trans-
fer. This mode is convenient for transferring data formatted as frames or
blocks. ABU mode is used to implement autobuffering functions. In this mode,
the element counter for each channel (DMCTRn) represents the buffer size
and is not modified during the transfers. Although the McBSP is mentioned
here as an example of a source/destination, any source/destination can be
used.
Multiframe Mode
In multiframe mode, the data transfers can be structured as multiple elements
in a frame, and multiple frames in a block. The element index registers
(DMIDX0, DMIDX1) are used to modify the addresses of elements within a
frame, and the frame-index registers (DMFRI0, DMFRI1) are used to modify
addresses following the completion of frames. The number of elements trans-
ferred per frame is determined by the channel element count (DMCTRn) regis-
ter and the number of frames transferred per block is determined by the chan-
nel-frame count (frame count field of the DMSFCn register). The two counters
are decremented following each element or frame transfer, respectively.
The element count is an unsigned 16-bit integer. The element count register
should be initialized with one less than the desired number of elements to be
transferred. The number of elements to be transferred per frame can be be-
tween 1 (0000h) and 65536 (FFFFh).
The frame count is an unsigned 8-bit integer. The frame count register should
be initialized with one less than the desired number of frames to be transferred.
The number of frames to be transferred per block can be between 1 (00h) and
256 (FFh). The total number of elements to be transferred is called the block
size, and is the product of the frame count and the element count.
The element counter is decremented following each transfer. If all elements

in a frame have been transferred, the element counter is reloaded with the
original value (via a shadow register) and the frame count is decremented. If
the last element in the last frame has been transferred, both the element count-
er and the frame counter will contain 0000h and will remain unmodified, unless
the autoinitialization mode is enabled (AUTOINIT= 1). For more information on
the autoinitialization mode, see section 3.2.3.6, on page 3-28.
All of the address indexing modes mentioned in the Source and Destination
Address Selection and Modification section are valid in multiframe mode, and
3-22
the indexing methods can be applied to the source and destination addresses
separately. The following describes the operation of each of the addressing
modes during multiframe operation:
- No modify. The address is not modified following transfers, and main-

tains its initialized value.
- Postincrement. The address is incremented by 1 following each transfer.

The address is incremented by 2 if double-word operation is selected.
- Post-decrement. The address is decremented by 1 following each trans-

fer. The address is decremented by 2 if double-word operation is selected.
- Postincrement With Index Offsets (DMIDX0 or DMIDX1). The

address is modified by the amount in the referenced element index regis-
ter. The element index register contains a 16-bit signed (2s complement
value). Following a transfer, the index value is added (or subtracted) from
the current address to generate the next address. In double-word mode,
the address is modified by twice the value of the index.
- Postincrement With Index Offset (DMIDX0 and DMFRI0, or

DMIDX1 and DMFRI1), or Sorting Mode. If the current transfer is
not the last element in the frame, the address is modified by the contents
of the referenced element index register. If the current element is the last
element in the frame, the address in modified by the contents of the frame
index register. The frame-index registers contain a 16-bit signed value.
This mode is useful for sorting data such as the information in a T1 frame.
Example 3−3. Data Sorting by Address Modification
For this example, assume a data stream where the data is structured as four
elements per frame, and four frames per block. The source data arrives in the
order shown on the left in Figure 3−5. It may be convenient to store or process
the data sorted by element number instead of frame number. This is achieved
by selecting sorting mode for the address modification, setting the destination
element index register to increment by four after each transfer, and setting the
elements and frame index to decrement by 11 after each frame. When the en-
tire block of 16 transfers has been completed, the data will be sorted by ele-
ment number instead of frame number as shown in Figure 3−5.
To configure the element index to increment by four, the signed value 0004h
is stored in the element index register. To configure the frame index to decre-
ment by 11, the signed value 0FFF5h (-11 decimal) is stored in the frame index
register.
3-23
The source and destination addresses are not updated after the last transfer
in the block. They contain the last source and destination address. The ele-
ment and frame counters both contain 0000h after the last transfer. The config-
uration of the DMA registers required to implement this example is shown in
Example 3−21, McBSP to Data Memory Transfer with Data Sorting, on page
3-57.
Figure 3−5. Data Sorting Example
Data Receive Order 0h Frame 0 element 0

Frame 0 Element 0 1h Frame 1 element 0 Element index =
Frame 0 Element 1 0004h (4)
2h Frame 2 element 0
Frame 0 Element 2
Frame 0 Element 3
Frame 1 Element 0 5h Frame 1 element 1
Frame 1 Element 1
Frame index = 6h Frame 2 element 1
Frame 1 Element 2
FFF5h (−11)

Frame 2 Element 2 Ah Frame 2 element 2
Frame 2 Element 3
Bh Frame 3 element 2
Frame 3 Element 0 Ch Frame 0 element 3
Frame 3 Element 1
Dh Frame 1 element 3
Frame 3 Element 2
Frame 3 Element 3 Eh Frame 2 element 3
Fh Frame 3 element 3
ABU Mode
ABU, or autobuffering mode, provides automatically controlled circular buffer

capability for DMA transfers. ABU can be used to duplicate the autobuffering
function implemented on the buffered serial port (BSP) using the McBSP in
conjunction with the DMA controller.
In this mode, either the source or destination is configured in autobuffering

mode, and the other location is specified as an unmodified address. For exam-
ple, a circular buffer for McBSP receive data requires the DMA to be configured
with the data receive register of the McBSP as the source location, and the cir-
cular buffer controlled by the DMA as the destination location.
3-24
During ABU mode operation, the element count register contains the buffer
size. The frame count register does not have a function in ABU mode. When
the address reaches the end of the buffer, it wraps back to the beginning auto-
matically. The number of transfers is not specified in this mode, so the address
wraps indefinitely until the DMA channel is disabled. The element count regis-
ter is interpreted as a 16-bit unsigned integer and valid buffer sizes range from
0002h to 0FFFFh. The buffer can be any size in this range and is not limited
to powers of two.
The buffer has a minimum address, called the base address, and a maximum
address. The difference between the base address and the maximum address
is the buffer size. Although the buffer size is not limited to powers of two, the
base address must be based on powers of two. The required location depends
on the buffer size. The base address must be located on an address boundary
that corresponds to one power-of-two higher than the most significant bit posi-
tion of the buffer size. The address boundaries for all available buffer sizes are
shown in Table 3−11. Circular buffers cannot cross 64k address boundaries.
Don’t care values in the table are indicated by X.
Table 3−11. ABU Buffer Examples
ABU Buffer Size Buffer Base Address

[hexadecimal (decimal)] [binary]
0002h−0003h (2−3) XXXX XXXX XXXX XX00 b
0004h−0007h (4−7) XXXX XXXX XXXX X000 b
0008h−000Fh (8−15) XXXX XXXX XXXX 0000 b
0010h−001Fh (16−31) XXXX XXXX XXX0 0000 b
0020h−003Fh (32−63) XXXX XXXX XX00 0000 b
0040h−007Fh (64−127) XXXX XXXX X000 0000 b
0080h−00FFh (128−255) XXXX XXXX 0000 0000 b
0100h−01FFh (256−511) XXXX XXX0 0000 0000 b
0200h−03FFh (512−1023) XXXX XX00 0000 0000 b
0400h−07FFh (1024−2047) XXXX X000 0000 0000 b
0800h−0FFFh (2048−4095) XXXX 0000 0000 0000 b
1000h−1FFFh (4096−8191) XXX0 0000 0000 0000 b
2000h−3FFFh (8192−16383) XX00 0000 0000 0000 b
4000h−7FFFh (16384−32767) X000 0000 0000 0000 b
8000h−7FFFh (32768−65535) 0000 0000 0000 0000 b
3-25
Example 3−4. ABU Buffer Size Examples
- For a buffer size of eight (decimal), the next higher power of two is 16, so
the buffer base address must be aligned with 16 word boundaries (ex.
0000h, 0010h, 0020h).
- For a buffer size of 5 (decimal), the next higher power of two is 8, so the
buffer base address must be aligned with 8 word boundaries (ex. 0000h,
0008h, 0010h, 0018h, 0020h).
- For a buffer size of 200 (decimal), the next higher power of two is 256, so
the buffer base address must be aligned with 256 word boundaries (ex.
0000h, 0100h, 0200h, 0300h).
The DMA transfer always starts at the addresses specified in the channel
source or destination address registers. The address register can point to any
location inside the defined range of the buffer. After each access, the appropri-
ate address register is modified according to the specified address indexing
mode. A limited set of address indexing modes are available for ABU mode
operation. Table 3−12 shows the modes available for ABU operation.
To configure the ABU mode properly, one side of the transfer (the source or
destination, but not both) must be configured with no address modification
(SIND/DIND=000b). The other side of the transfer must be configured with a
modified addressing mode. All legal combinations of SIND and DIND for ABU
mode are shown in Table 3−12. Use of combinations other than those shown
may cause unpredictable behavior.
Table 3−12. ABU Address Indexing Modes
SIND Address Index Mode DIND Address Index Mode
001 Postincrement
010 Postdecrement
000 No modification 011 Postincrement with
index offset (DMIDX0)
100 Postincrement with
001 Postincrement
010 Postdecrement
011 Postincrement with 000 No modification
100 Postincrement with
3-26
As transfers proceed, the address in the circular buffer can be incremented or
decremented by one, or modified by the contents of one of the element index
registers (DMIDX0 or DMIDX1). If the DMA channel is configured for double-
word mode, the address is modified either by two, or by twice the index-offset
value. If the calculated next address for the buffer is greater than the maximum
buffer size, the address wraps to the base address. If a greater-than-one index
is used, the appropriate offset from the base address is calculated.
You should note that in double-word mode, the address of the second transfer
is always the address of the first transfer with the LSB inverted, regardless of
whether the maximum address has been crossed. The buffer wrap address
is always calculated with regard to the first transfer address of the double-word
transfer.
Example 3−5 shows the buffer wrap address calculated for a single-word
transfer with indexed addressing. Example 3−6 shows the buffer wrap ad-
dress calculated for a double-word transfer with indexed addressing.
Example 3−5. Wrap Address Calculation for a Single-Word Transfer With Indexed
Addressing
Buffer location: 0580h−059Eh
Current Address: 059Eh
Index: 2
Next Address: 0581h
Example 3−6. Wrap Address Calculation for a Double-Word Transfer With Indexed
Addressing
Buffer location: 0580h−059Eh
Current Address: 059Eh
Index: 2 (a total of + 4 since double-word mode)
Next Address: 0583h
Negative indexes can also be specified. In this case, the base address and the
maximum address calculation remains the same as described above. The ad-
dress is modified according to the negative address index. When the base ad-
dress is crossed, the address wraps to the maximum address. If an index of
less than -1 is being used, the appropriate offset from the maximum address
is calculated.
In ABU mode, interrupts can be configured to occur when the buffer is half full
or full. A detailed explanation of the interrupt operation is provided in section
3.2.3.7, Interrupt Generation, on page 3-28.
3-27
3.2.3.6 Autoinitialization
The AUTOINIT field of the DMMCRn controls the autoinitialization capability

for each DMA channel. When autoinitialization is enabled (AUTOINIT = 1), the
channel-context registers are re-initialized upon the completion of a block
transfer. Autoinitialization mode eliminates the need for the CPU to intervene
when block transfers are completed. When autoinitialization occurs, the fol-
lowing channel-context registers are modified:
- DMSRCn is loaded with the contents of DMGSA, the global source ad-
dress register. DMGSA contains a 16-bit source address.
- DMDSTn is loaded with the contents of DMGDA, the global destination ad-
dress register. DMGDA contains a 16-bit destination address.
- DMCTRn is loaded with the contents of DMGCR, the global element count
register. DMGCR contains a 16-bit unsigned element count value.
- The frame count field of the DMSFCn is loaded with the contents of
DMGFR, the global frame count register. DMGFR contains an 8-bit un-
signed frame count value. The upper 8 bits of the DMGFR are reserved
and have no effect. The reserved bits are always read as zeros.
Autoinitialization is only available in multiframe mode (CTMOD = 0).
3.2.3.7 Interrupt Generation
Two fields in the DMMCRn (DINM and IMOD) control how interrupts are han-
dled during DMA transfers. DINM is used to enable or disable a generation of
interrupts based on the completion of part or all of a transfer. If DINM = 0, inter-
rupts are disabled and no interrupt generation occurs. If DINM = 1, interrupt
generation is enabled and occurs based on the configuration of the IMOD bit.
The interrupt mask register (IMR) and the INTM bit in the CPU still control
whether an interrupt from the DMA is serviced. DINM enables the ability of the
DMA to generate an interrupt. For more information on operation of the IMR,
INTM, and interrupt processing, see the user’s guide titled TMS320C54x DSP,
CPU and Peripherals Reference Set, Volume 1 (literature number SPRU131).
If interrupt generation is enabled (DINM=1), then the IMOD field in DMMCRn

controls when the interrupt is generated within a block transfer. The available
operation modes are shown below in Table 3−13.
3-28
Table 3−13. DMA Block Transfer Interrupt Generation Modes
MODE CTMOD DINM IMOD Interrupt Generation
ABU 1 1 0 At buffer full only
ABU 1 1 1 At half buffer full and buffer full
Multiframe 0 1 0 At block transfer complete
Multiframe 0 1 1 At end of frame and end of block
Either X 0 X No interrupt generated
In multiframe mode, interrupts can be generated either at the end of each

frame (and the end of each block of frames), or only at the end of the entire
block. The end of a frame occurs when the frame count equals zero. The end
of a block occurs when the element count and the frame count both equal zero.
In ABU mode (CTMOD = 1), interrupts can be generated either when the entire
buffer has been transferred or when each half of the buffer has been transfer-
red. The half-buffer option can be used to emulate the operation of the buffered
serial port (BSP) on the ’548 and ’549 devices.
The logic that generates the interrupt is based on the calculated next address.
The determination of the half-buffer point depends on the ’54x DSP being
used. The half-buffer interrupt generation for each device is described in detail
in the sections that follow.
3.2.3.8 ’5402/’5420 Half-Buffer Interrupt Generation in ABU Mode
On the ’5402 and ’5420, the half-buffer interrupt is generated when the next
address is equal to or greater than the halfway point in the buffer (if positive
indexing is used), or when the next address is less than the halfway point when
negative indexing is being used. The interrupt point for odd and even size buff-
ers differ accordingly. The full-buffer interrupt is generated when the next ad-
dress wraps back to the base address (positive indexing), or back to the maxi-
mum address (negative indexing). The examples that follow illustrate when
the interrupts are generated in each case.
3-29
Example 3−7. ’5402/’5420 ABU Interrupt Example − Even Size Buffer With +1 Index
Buffer size = 8 Half-buffer = 4
Current Address Next Address Interrupt

0000h 0001h
0001h 0002h
0002h 0003h
0003h 0004h Interrupt generated
0004h 0005h
0005h 0006h
0006h 0007h
0007h 0000h Interrupt generated (wrap)
Example 3−8. ’5402/’5420 ABU Interrupt Example − Odd Size Buffer With +1 Index

0000h 0001h
0001h 0002h
0003h 0004h
0004h 0005h
0005h 0006h
3-30
Example 3−9. ’5402/’5420 ABU Interrupt Example − Even Size Buffer With −1 Index

0007h 0006h
0006h 0005h
0005h 0004h
0003h 0002h
0002h 0001h
0001h 0000h
Example 3−10. ’5402/’5420 ABU Interrupt Example − Odd Size Buffer With −1 Index

0006h 0005h
0005h 0004h
0004h 0003h
0002h 0001h
0001h 0000h
3-31
3.2.3.9 ’5410 Half-Buffer Interrupt Generation in ABU Mode
On the ’5410, the half-buffer interrupt is generated when the next address is
greater than the halfway point in the buffer if positive indexing is used, or when
the next address is less than the halfway point when negative indexing is used.
For odd and even size buffers, the interrupt point differs accordingly. The full-
buffer interrupt is generated when the next address wraps back to the base
address (positive indexing), or back to the maximum address (negative index-
ing). The examples that follow illustrate when interrupts are generated in each
case.
Example 3−11. ’5410 ABU Interrupt Example − Even Size Buffer With +1 Index

0000h 0001h
0001h 0002h
0002h 0003h
0003h 0004h
0005h 0006h
0006h 0007h
Example 3−12. ’5410 ABU Interrupt Example − Odd Size Buffer With +1 Index
0000h 0001h
0001h 0002h
0002h 0003h
0004h 0005h
0005h 0006h
3-32
Example 3−13. ’5410 ABU Interrupt Example − Even Size Buffer With −1 Index
0007h 0006h
0006h 0005h
0005h 0004h
0003h 0002h
0002h 0001h
0001h 0000h
Example 3−14. ’5410 ABU Interrupt Example − Odd Size Buffer With −1 Index
Current Address6 Next Address Interrupt
0006h 0005h
0005h 0004h
0004h 0003h
0002h 0001h
0001h 0000h
3-33
3.3 Extended Addressing
The DMA controller has the ability to perform transfers to and from the ex-
tended program memory space. Two subaddressed registers are employed
to provide this capability: the DMA source program page address register
(DMSRCP) and the DMA destination program page address register
(DMDSTP). These registers contain the extended program page number for
the source and destination locations, respectively. In each of these registers,
the least significant seven bits are used to store the extended address page
of the source and destination addresses, as shown in Figure 3−6 and
Figure 3−7. Following reset, DMSRCP and DMDSTP are initialized to 0000h,
or program page zero. The reserved bit locations are always read as zeros.
DMSRCP and DMDSTP are not channel specific. These registers control the
source and destination address pages in program memory for all DMA chan-
nels. In addition, the program page addresses stored in these registers are not
modified during source and destination address modifications. Consequently,
program space transfers cannot cross 64K page boundaries. If a program
page boundary is crossed during a transfer, the next transfer wraps to the
same page.
Not all ’54x devices support 128 pages of extended program memory. You
should consult the device-specific data sheet for detailed information on ex-
tended memory support.
Figure 3−6. DMA Source Program Page Address Register (DMSRCP)

15 7 6 0
Reserved Source Program Page Address
Figure 3−7. DMA Destination Program Page Address Register (DMDSTP)

15 7 6 0
Reserved Destination Program Page Address
3-34
3.4 DMA Memory Maps
This section provides DMA memory maps for the ’5402, ’5410, and ’5420 de-
vices.
3.4.1 ’5402 DMA Memory Map

The ’5402 DMA supports internal accesses only. The DMA memory map for
the ’5402 is shown in Table 3−14. Memory regions marked as reserved
(shaded) are not accessible through the ’5402 DMA. The DMA memory map
is not affected by the MP/MC, DROM, or OVLY bits.
Table 3−14. ’5402 DMA Memory Map
Address Range
(Hex) Description
Program space 00 0000 0F FFFF Reserved
Data space 0000 001F Reserved
0020 McBSP0 data receive register (DRR20)
0022 McBSP0 data transmit register (DXR20)
0024 003F Reserved
0044 005F Reserved
0060 007F Scratchpad RAM
0080 3FFF DARAM
4000 FFFF Reserved
I/O space 0000 FFFF Reserved
3-35
The ’5410 supports DMA accesses to internal and external memory. The DMA
memory map for the ’5410 is shown in Table 3−15. Memory regions marked
as reserved (shaded) are not accessible through the ’5410 DMA. The DMA
memory map is not affected by the MP/MC, DROM, or OVLY bits.

Address Range
(Hex) Description
Program space 00 0000 00 007F Reserved
00 0080 00 BFFF External memory
00 C000 00 FFFF On-chip ROM
10 0000 01 7FFF External memory
01 8000 01 FFFF On-chip SARAM2
02 0000 7F FFFF External memory
0024 002F Reserved
0034 0035 Reserved
0036 RCERA2
0037 XCERA2
0038 0039 Reserved
003A RCERA0
003B XCERA0
3-36
Table 3−15. ’5410 DMA Memory Map (Continued)
Address Range
(Hex) Description
003C 003F Reserved
0044 0049 Reserved
004A RCERA1
004B XCERA1
004C 005F Reserved
0060 7FFF On-chip RAM
8000 FFFF External memory
I/O space 0000 FFFF External memory

The ’5420 DMA supports internal accesses only. The DMA memory map for
the ’5420 is shown in Table 3−16. Memory regions marked as reserved
(shaded) are not accessible through the ’5420 DMA. The entire I/O space on
both subsystems is devoted to the interprocessor FIFO; consequently, the
’5420 DMA does not have access to the external I/O ports. The DMA memory
map is not affected by the MP/MC, DROM, or OVLY bits.
3-37
Address Range
(Hex) Description
Program space 00 0000 00 001F Reserved
00 0020 McBSP0 data receive register (DRR20)
00 0022 McBSP0 data transmit register (DXR20)
00 0024 00 002F Reserved
00 0034 00 003F Reserved
00 0044 00 005F Reserved
00 0060 00 3FFF On-chip DARAM 0
00 4000 00 7FFF On-chip SARAM 1
00 8000 00 FFFF On-chip SARAM 2
01 0000 01 005F Reserved
01 8000 01 FFFF On-chip SARAM 3
02 0000 02 005F Reserved
02 8000 02 EFFF Reserved
02 F000 02 FFFF On-chip SARAM 4
3-38
Table 3−16. ’5420 DMA Memory Map (Continued)
Address Range
(Hex) Description
03 0000 03 005F Reserved
03 8000 03 FFFF Reserved
McBSP0 data receive register (DRR20)
0024 002F Reserved
0034 003F Reserved
0044 005F Reserved
0060 007F On-chip scratchpad RAM
0080 3FFF On-chip DARAM 0
4000 7FFF On-chip SARAM 1
8000 FFFF On-chip SARAM 2
I/O space 0000 FFFF DMA FIFO for interprocessor communications
3-39
3.5 DMA Transfer Latency
All 16-bit DMA transfers are composed of a read followed by a write. The time
to complete this activity depends on the source and destination locations, ex-
ternal interface conditions (such as wait states and bank-switching cycles), the
number of active DMA channels, and the activity level of the host port interface
(HPIx).
A single DMA transfer from an internal (on-chip) source location to an internal

destination location requires four CPU clock cycles. The read portion of the
transfer takes two cycles and the write portion of the transfer takes two cycles.
If either the read, the write, or both attempt to access an external (off-chip)
location, the total transfer time is extended by the number of cycles necessary
to complete the access on that interface. This time determines the maximum
transfer rate for transfers from/to a given source/destination. All ’54x devices
support internal-to-internal transfers of four cycles. Therefore, the maximum
transfer rate for a channel is the CPU clock rate divided by four. For example,
a processor operating with a CPU clock rate of 100 MHz can support a maxi-
mum of 25M 16-bit transfers per second (25 Mwords per second), or 12.5M
double-word (32-bit) transfers per second. A summary of maximum DMA
transfer cycle times is shown in Table 3−17. Accesses to on-chip memory or
memory-mapped registers are both considered internal. For devices that do
not support external accesses through the DMA, the table entries are marked
not applicable (N/A).
Table 3−17. DMA Transfer Cycle Times

DMA Source DMA Destination ’5402 Transfer ’5410 Transfer ’5420 Transfer
Location Location Cycles Cycles Cycles
Internal Internal 4 4 4
Internal External N/A 5 N/A
External Internal N/A 5 N/A
External External N/A 6 N/A
On transfers that involve an external access, the number of cycles required to

complete the transfer is affected by conditions on the external interface includ-
ing software and hardware wait states, additional cycles due to bank-switch-
ing, and the speed of the interface if CLKOUT division is available. For exam-
ple, on the ’5410, CLKOUT (and the external parallel interface) can be config-
ured to operate as slow as one-fourth the speed of the CPU clock. In this case,
external DMA transfers also slow down accordingly. The conditions present in
the actual application should be considered when estimating the maximum
DMA transfer rate involving external accesses. The ’5410 column in
3-40
Table 3−17 assumes the external interface is configured with no hardware or
software wait states and CLKOUT is in divide-by-1 mode (CLKOUT is equal
to CPU clock).
The transfer rate on a particular channel is affected by the activity on the other
channels. Since all high-priority channels are serviced first in a circular pattern,
the data rate on a particular channel is governed by the number of other chan-
nels that are active at the same priority level. So while the high-priority chan-
nels are still active, the maximum data rate on each high-priority channel is di-
vided by the total number of high-priority channels. The transfer rate on the
low-priority channels is zero until there is no high priority activity.
Example 3−15 illustrates such a situation.
Example 3−15. DMA Channel Transfer Rate Example

Assume the ’54x processor is running at 100 MHz CPU clock. Assume the fol-
lowing DMA channel configuration:
- DMA channel 1 is configured for 100 high priority internal-to-internal single-

word transfers.
- DMA channel 2 is configured for 200 high priority internal-to-internal single-
word transfers.
- DMA channel 3 is configured for 50 low priority internal-to-internal single-word
transfers.
- DMA channel 4 is configured for 75 low priority internal-to-internal single-word
transfers.
- DMA channels 5 and 6 are disabled and HPI-8/-16 is inactive.
Assuming channels 1, 2, 3, and 4 are enabled at the same time and none of
the channels wait for sync events, channels 1 and 2 are serviced first and
channels 3 and 4 wait until channels 1 and 2 are completed. The transfer rates
on each of the channels proceed as follows.
While channels 1 and 2 are still active:
- DMA channel 1 data transfer rate = 25MWps/2 active channels = 12.5 MWps
- DMA channel 3 data transfer rate = 0 due to priority
3-41
When channel 1 is completed but channel 2 is still active:
- DMA channel 1 data transfer rate = 0 (completed)

- DMA channel 2 data transfer rate = 25MWps /1 active channel = 25 MWps
When channels 1 and 2 are completed, channels 3 and 4 become active:

When channels 1, 2, and 3 are completed but channel 4 is still active:

- DMA channel 4 data transfer rate = 25MWps/1 active channels = 25 MWps
3-42
3.6 Enhanced Host Port Interface Access Through the DMA Controller
The enhanced host port interface (HPI-8 or HPI-16, depending on the DSP)
uses the DMA bus to gain access to on-chip memory. The HPI has a dedicated
port on the DMA controller and makes requests to use the DMA bus. The DMA
controller arbitrates these requests based on the current activity. When the
HPIx makes a request, the current DMA transfer in progress is completed and
the DMA controller grants the HPIx access to the DMA bus. All pending high-
and low-priority transfers are suspended until the HPIx releases the bus.
When the HPI releases the bus, the pending DMA transfers continue as be-
fore. For this reason, HPIx activity affects the transfer rates on the DMA chan-
nels.
Double-word transfers are not interrupted by an HPIx request. Both 16-bit

transfers are completed before the DMA controller grants access to the HPIx.
Although the HPIx has a dedicated port on the DMA controller, it does not have
a dedicated channel. No channel context programming is required for the HPIx
to use the DMA bus. The HPIx is operated normally by the host, while access
to the internal buses is arbitrated by the DMA controller. Due to this arrange-
ment, use of the HPIx does not consume any of the available DMA channels.
For more information on the operation of the enhanced host port interfaces
(HPI-8 and HPI-16), refer to Chapter 4, Enhanced 8-Bit Host Port Interface
(HPI-8), and Chapter 5, Enhanced 16-Bit Host Port Interface (HPI-16).
3-43
Interprocessor FIFO Communication on the ’5420 / DMA Operation in Power-Down Mode
3.7 Interprocessor FIFO Communication on the ’5420

The ’5420 has two CPU subsystems that are independent of each other, and
can operate separately with complete functionality.
One method of exchanging data between the two subsystems uses the DMA
controller in conjunction with an on-chip first-in, first-out (FIFO) unit. The FIFO
is mapped to the DMA I/O space. Any access to DMA I/O space on the ’5420
writes to, or reads from, this FIFO. All addresses are mapped to the same FIFO
location. For this reason, the ’5420 DMA does not access I/O space in the
same manner as the other ’54x devices.
For more information on ’5420 subsystem communication methods, see

Chapter 6, Interprocessor Communications.
3.8 DMA Operation in Power-Down Mode

The ’54x devices offer several power-down modes that allow part or all of the
device to enter a dormant state and dissipate less power than when running
normally. Power-down modes can be invoked in several ways, including exe-
cuting the IDLE instruction, and driving the HOLD input low with the HM status
bit set to one. The DMA, like the other peripherals, can take the CPU out of
IDLE using interrupts.
The DMA continues to operate while in all IDLE modes (IDLE1/2/3). When the
DMA is not running (not performing a transfer), the DMA controller automati-
cally stops its internal clocks to conserve power. When a transfer needs to be
performed, the clocks are switched on, the transfer is completed, and the
clocks are switched off again. This clock management operates in each of the
three IDLE modes.
In IDLE1 and IDLE2, the clock source to the DMA comes from the same source
used to generate the CPU and system clocks (either the PLL or a divided ver-
sion of the X2/CLKIN). In IDLE3, the PLL is stopped, and therefore, cannot be
used as a clock source for the DMA. In this case, X2/CLKIN is used directly
to clock the DMA controller. For this reason, a clock must always be present
at the X2/CLKIN pin for the DMA to operate in IDLE 3 mode.
When X2/CLKIN is used to clock the DMA in IDLE 3 mode, it is not multiplied
or divided. As a result, the speed of DMA transfers in IDLE 3 mode is affected
accordingly.
3-44
3.9 Programming Examples
Below are several DMA controller programming examples. Each example
contains a description of the desired operation and the code used to set up the
DMA controller. Only the configuration code is included − additional code such
as interrupt service routines are not included. For clarity, the code sections re-
fer to DMA register names. The addresses for these registers can be identified
in assembly code as shown on the facing page.
3-45
**************************************************
*
* 54x Register Definitions for the DMA Controller
*
**************************************************
DMPREC .set 0054h ;Channel Priority and Enable Control Register
DMSA .set 0055h ;Sub-bank Address Register
DMSDI .set 0056h ;Sub-bank Data Register with autoincrement
DMSDN .set 0057h ;Sub-bank Data Register without modification
DMSRC0 .set 00h ;Channel 0 Source Address Register
DMDST0 .set 01h ;Channel 0 Destination Address Register
DMCTR0 .set 02h ;Channel 0 Element Count Register
DMSFC0 .set 03h ;Channel 0 Sync Select and Frame Count Register
DMMCR0 .set 04h ;Channel 0 Transfer Mode Control Register
DMSRC2 .set 0Ah ;Channel 2 Source Address Register
DMDST2 .set 0Bh ;Channel 2 Destination Address Register
DMCTR2 .set 0Ch ;Channel 2 Element Count Register
DMSFC2 .set 0Dh ;Channel 2 Sync Select and Frame Count Register
DMMCR2 .set 0Eh ;Channel 2 Transfer Mode Control Register
DMSRC3 .set 0Fh ;Channel 3 Source Address Register
DMDST5 .set 1Ah ;Channel 5 Destination Address Register
DMCTR5 .set 1Bh ;Channel 5 Element Count Register
DMSFC5 .set 1Ch ;Channel 5 Sync Select and Frame Count Register
DMMCR5 .set 1Dh ;Channel 5 Transfer Mode Control Register
DMSRCP .set 1Eh ;Source Program Page Address
DMDSTP .set 1Fh ;Destination Program Page Address
DMIDX0 .set 20h ;Element Address Index Register 0
DMIDX1 .set 21h ;Element Address Index Register 1
DMFRI0 .set 22h ;Frame Address Index Register 0
DMFRI1 .set 23h ;Frame Address Index Register 1
DMGSA .set 24h ;Global Source Address Reload Register
DMGDA .set 25h ;Global Destination Address Reload Register
DMGCR .set 26h ;Global Element Count Reload Register
DMGFR .set 27h ;Global Frame Count Reload Register
3-46
Example 3−16. Program Memory to Data Memory Transfer Without Autoincremented
Subaddressing
This example transfers a block of data from program space to data space via
direct memory access. The DMA controller is configured to perform 1000h
single-word transfers with both the source and destination addresses increm-
ented by one after each element transfer. The transfers are not associated with
any sync event (free-running). After the block transfer is complete (autoinitial-
ization is off), the DMA channel is disabled.
Transfer mode: Multi-frame mode

Source address: 18000h in program space
Destination address: 03000h in data space
Transfer size: 1000h single (16-bit) words
Sync event: None (free-running)
Channel use: DMA channel #0
3-47
***********************************
stm DMSRCP,DMSA ;set source program page to 1
stm #1h,DMSDN
stm DMSRC0,DMSA ;set source program address to 8000
stm #8000h,DMSDN ; (lower 16-bit of 18000h)
stm DMDST0,DMSA ;set destination address to 3000
stm #3000h,DMSDN
stm DMCTR0 ,DMSA ;set for 1000h transfers
stm #(1000h−1) ,DMSDN
stm DMSFC0 ,DMSA
stm #0000000000000000b ,DMSDN
;0000~~~~~~~~~~~~ (DSYN) No sync event
;~~~~0~~~~~~~~~~~ (DBLW) Single-word mode
;~~~~~000~~~~~~~~ Reserved
;~~~~~~~~00000000 (Frame Count) Frame Count=0h (one frame)
stm DMMCR0 ,DMSA
stm #0000000100000101b ,DMSDN
;0~~~~~~~~~~~~~~~ (AUTOINIT) Autoinitialization disabled
;~0~~~~~~~~~~~~~~ (DINM) Interrupts masked
;~~0~~~~~~~~~~~~~ (IMOD) N/A
;~~~0~~~~~~~~~~~~ (CTMOD) Multi-frame mode
;~~~~0~~~~~~~~~~~ Reserved
;~~~~~001~~~~~~~~ (SIND) Post increment source address
;~~~~~~~~00~~~~~~ (DMS) Source in program space
;~~~~~~~~~~0~~~~~ Reserved
;~~~~~~~~~~~001~~ (DIND) Post increment destination address
;~~~~~~~~~~~~~~01 (DMD) Destination in data space
stm #0000000100000001b ,DMPREC

;0~~~~~~~~~~~~~~~ (FREE) DMA stops on emulation stop
;~0~~~~~~~~~~~~~~ Reserved
;~~0~~~~~~~~~~~~~ (DPRC[5]) Channel 5 low priority
;~~~0~~~~~~~~~~~~ (DPRC[4]) Channel 4 low priority
;~~~~0~~~~~~~~~~~ (DPRC[3]) Channel 3 low priority
;~~~~~0~~~~~~~~~~ (DPRC[2]) Channel 2 low priority
;~~~~~~0~~~~~~~~~ (DPRC[1]) Channel 1 low priority
;~~~~~~~1~~~~~~~~ (DPRC[0]) Channel 0 high priority
;~~~~~~~~00~~~~~~ (INTOSEL) N/A
;~~~~~~~~~~0~~~~~ (DE[5]) Channel 5 disabled
;~~~~~~~~~~~0~~~~ (DE[4]) Channel 4 disabled
;~~~~~~~~~~~~0~~~ (DE[3]) Channel 3 disabled
;~~~~~~~~~~~~~0~~ (DE[2]) Channel 2 disabled
;~~~~~~~~~~~~~~0~ (DE[1]) Channel 1 disabled
;~~~~~~~~~~~~~~~1 (DE[0]) Channel 0 enabled
***********************************
3-48
Example 3−17. Program Memory to Data Memory Transfer Using Autoincremented
Subaddressing
This example performs the same transfer as the previous example but uses
autoincremented subaddressing (the DSMBAI register) during the setup. This
subaddressing scheme allows the channel-context registers to be configured
using a single instruction instead of two, saving time and memory.

Source address: 18000h in program space
3-49
***********************************
stm #1h,DMSDN
stm DMSRC0,DMSA ;set source program address to 8000 and

stm #8000h,DMSDI ; point DMSA to next sub-address (DMDST0)
stm #3000h,DMSDI ;set destination address to 3000 and

; point DMSA to next sub-address (DMCTR0)
stm #(1000h−1),DMSDI ;set for 1000h transfers and point

; DMSA to next sub-address (DMSFC0)
stm #00000h,DMSDI ;configure DMSFC0 and point DMSA

; to next sub-address (DMMCR0)
stm #00105h,DMSDI ;configure DMMCR0 and point DMSA

; to next sub-address (DMSRC0)
stm #00101h,DMPREC ;configure DMPREC

***********************************
3-50
Example 3−18. Program Memory to Data Memory Transfer With Autoinitialization
This example transfers a block of data from program space to data space via
direct memory access. The DMA controller is configured to perform 1000h
single-word transfers per block, with both the source and destination address-
es incremented by one after each element transfer. The transfers are not asso-
ciated with any sync event (free-running). When the first block transfer is com-
pleted, the DMA channel is autoinitialized to begin again using the contents
of the global reload registers (DMGSA, DMGDA, DMGCR, DMGFR). The
global reload registers direct the destination to 02000h in data space, instead
of 03000h, on the first block transfer. The global reload registers can also be
used to autoinitialize the identical conditions as the first block transfer.

Initial source address: 18000h in program space
Initial destination address: 03000h in data space
Initial transfer size: 1000h single (16-bit) words
(1 frame, 1000h elements)
Autoinitialize source address: 18000h in program space
Autoinitialize destination address: 02000h in data space
Autoinitialize element count: 1000h single (16-bit) words
Autoinitialize frame count: 000h (1 frame)
***********************************
stm #1h,DMSDN
stm DMSRC4,DMSA ;set source program address to 18000h
stm #8000h,DMSDN
stm DMDST4,DMSA ;set destination address to 3000h
stm #3000h,DMSDN
stm DMCTR4 ,DMSA ;set for 1000h transfers
stm #(1000h−1) ,DMSDN
stm DMSFC4 ,DMSA
stm #0000000000000000b ,DMSDN
;0000~~~~~~~~~~~~ (DSYN) No sync event
;~~~~~000~~~~~~~~ Reserved
;~~~~~~~~00000000 (Frame Count) Frame Count = 0h (one frame)
3-51
stm DMMCR4 ,DMSA
stm #1000000100000101b ,DMSDN
;1~~~~~~~~~~~~~~~ (AUTOINIT) Autoinitialization enabled
;~0~~~~~~~~~~~~~~ (DINM) Interrupts masked
;~~0~~~~~~~~~~~~~ (IMOD) N/A
;~~~~0~~~~~~~~~~~ Reserved
;~~~~~001~~~~~~~~ (SIND) Post increment source address
;~~~~~~~~00~~~~~~ (DMS) Source in program space
;~~~~~~~~~~0~~~~~ Reserved
stm DMGSA,DMSA ;set global source address to 8000h

stm #8000h,DMSDN ; (lower 16-bits of 18000h in program space)
stm DMGDA,DMSA ;set global destination address to 2000h
stm #2000h,DMSDN
stm DMGCR,DMSA ;set global element count to move 1000h words
stm #(1000h−1),DMSDN
stm DMGFR,DMSA ;set global frame count to 0h
stm #0000000000000000b,DMSDN
;00000000~~~~~~~~ Reserved
;~~~~~~~~00000000 Global Frame count
stm #0001000000010000b ,DMPREC

;~0~~~~~~~~~~~~~~ Reserved
;~~~1~~~~~~~~~~~~ (DPRC[4]) Channel 4 high priority
;~~~~~~0~~~~~~~~~ (DPRC[1]) Channel 1 low priority
;~~~~~~~0~~~~~~~~ (DPRC[0]) Channel 0 low priority
;~~~~~~~~00~~~~~~ (INTOSEL) N/A
;~~~~~~~~~~~1~~~~ (DE[4]) Channel 4 enabled
;~~~~~~~~~~~~~~0~ (DE[1]) Channel 1 disabled
;~~~~~~~~~~~~~~~0 (DE[0]) Channel 0 disabled
***********************************
3-52
Example 3−19. McBSP Data Transfer in ABU Mode
This example transfers 16-bit words received on McBSP0 to data space via
the DMA controller. The destination is configured as a circular buffer. The ad-
dress in the destination buffer is incremented by one after each transfer. ABU
mode is used to implement the circular buffer. In ABU mode, the element count
register represents the buffer size, and the frame count is not used. Only the
DMA configuration is shown, which in this example, is to generate an interrupt
to the CPU when the 100h buffer is full. The interrupt service routine for this
interrupt is not shown in this example.
Transfer mode: ABU (non-decrement) mode

Source address: McBSP0 data receive register (DRR10)
Destination buffer: 03000h − 030FFh in data space
Buffer size: 100h single (16-bit) words
Sync event: McBSP0 receive event
3-53
***********************************
stm DMSRC1,DMSA ;set source address to DRR10
stm DRR1_0,DMSDN
stm #3000h,DMSDN
stm DMCTR1 ,DMSA ;set buffer size to 100h words
stm #100h ,DMSDN
stm DMSFC1 ,DMSA
stm #0001000000000000b ,DMSDN
;0001~~~~~~~~~~~~ (DSYN) McBSP0 receive sync event
;~~~~~000~~~~~~~~ Reserved
;~~~~~~~~00000000 (Frame Count) Frame count is not
; relevant in ABU mode
stm DMMCR1 ,DMSA

stm #0101000001001101b ,DMSDN
;~1~~~~~~~~~~~~~~ (DINM) DMA Interrupts enabled
;~~0~~~~~~~~~~~~~ (IMOD) Interrupt at full buffer
;~~~1~~~~~~~~~~~~ (CTMOD) ABU (non-decrement) mode
;~~~~0~~~~~~~~~~~ Reserved
;~~~~~000~~~~~~~~ (SIND) No modify on source address (DRR10)
;~~~~~~~~01~~~~~~ (DMS) Source in data space
;~~~~~~~~~~0~~~~~ Reserved
; with DMIDX0
stm DMIDX0,DMSA ;set element address index to +1

stm #0001h,DMSDN
stm #0000001000000010b ,DMPREC

;~0~~~~~~~~~~~~~~ Reserved
;~~~~~~1~~~~~~~~~ (DPRC[1]) Channel 1 high priority
;~~~~~~~~00~~~~~~ (INTOSEL) N/A
;~~~~~~~~~~~~~~1~ (DE[1]) Channel 1 enabled
***********************************
3-54
Example 3−20. McBSP Data Transfer in Double-Word Mode
This example transfers 32-bit words received on McBSP0 to data space via
direct memory access in multi-frame mode. The DMA controller is configured
to automatically read DRR20 and DRR10 and transfer the data to sequential
locations starting at 03000h in data space.

Source address: McBSP0 data receive registers
(DRR20/DRR10)
Transfer size: 50h double (32-bit) words
3-55
***********************************
stm DRR2_0,DMSDN
stm #3000h,DMSDN
stm DMCTR1 ,DMSA ;set buffer size to 50h words
stm #050h−1,DMSDN
stm DMSFC1 ,DMSA

stm #0001100000000000b ,DMSDN
;~~~~1~~~~~~~~~~~ (DBLW) Double-word mode
;~~~~~000~~~~~~~~ Reserved
;~~~~~~~~00000000 (Frame Count) Single Frame
stm DMMCR1 ,DMSA

stm #0100000001001101b ,DMSDN
;~~0~~~~~~~~~~~~~ (IMOD) Interrupt at complete block transfer
;~~~~0~~~~~~~~~~~ Reserved
;~~~~~~~~~~0~~~~~ Reserved
; with DMIDX0
stm #0001h,DMSDN
stm #0000001000000010b ,DMPREC

;~0~~~~~~~~~~~~~~ Reserved
;~~~~~~~~00~~~~~~ (INTOSEL) N/A
***********************************
3-56
Example 3−21. McBSP to Data Memory Transfer With Data Sorting
This example implements the sorting example shown in Figure 3−5 on page
3-24. Single words received on the McBSP0 are transferred to data memory
starting at address 03000h. The incoming data is structured as four frames of
four elements each. The DMA sorts the incoming data by modifying the des-
tination addresses with the element index register (DMIDX0) and the frame in-
dex register (DMFRI0). The resulting stored data is sorted by element number
instead of frame number.

Source address: McBSP0 data receive register (DRR10)
3-57
***********************************
stm DRR1_0,DMSDN
stm #3000h,DMSDN
stm DMCTR1 ,DMSA ;set elements per frame to 4h
stm #0004h−1 ,DMSDN
stm DMSFC1 ,DMSA
stm #0001000000000011b ,DMSDN
;~~~~~000~~~~~~~~ Reserved
;~~~~~~~~00000011 (Frame Count) Frames per block = 4h−1
stm DMMCR1 ,DMSA
stm #0100000001010101b ,DMSDN
;~~0~~~~~~~~~~~~~ (IMOD) Interrupt at complete block transfer
;~~~~0~~~~~~~~~~~ Reserved
;~~~~~~~~~~0~~~~~ Reserved
; with DMIDX0 and DMFRI0
stm #0004h,DMSDN
stm DMFRI0,DMSA ;set frame address index to −11

stm #0FFF5h,DMSDN
stm #0000001000000010b ,DMPREC

;~0~~~~~~~~~~~~~~ Reserved
;~~~~~~~~00~~~~~~ (INTOSEL) N/A
***********************************
3-58
3-59
Chapter 4
Enhanced 8ĆBit Host Port Interface

(HPIĆ8)
The enhanced 8-bit host port interface, also referred to as HPI-8, is an im-
proved version of the standard 8-bit HPI designed to interface a host device
or host processor to the ’54x.
The HPI-8 is available on selected devices in the ’54x family of TI DSPs.
Topic Page
4.1 Introduction to the Enhanced 8-Bit Host Port Interface (HPI-8) . . . . 4-2
4.2 HPI-8 Basic Functional Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-4
4.3 Details of HPI-8 Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-6
4.4 Host Read/Write Access to HPI-8 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-15
4.5 DSPINT and HINT Operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-24
4.6 Considerations for HPI-8 Transfers While Changing
Clock Modes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-25
4.7 Considerations in Idle Use . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-27
4.8 Effects of Reset on HPI-8 Operations . . . . . . . . . . . . . . . . . . . . . . . . . . 4-29
4.9 HPI-8 Data Pins as General Purpose I/O Pins
(Not Available on’5410) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4-31
4-1
4.1 Introduction to the Enhanced 8-Bit Host Port Interface (HPI-8)
The HPI-8 is an 8-bit parallel port that interfaces a host device or host proces-
sor to the ’54x. Information is exchanged between the ’54x and the host device
through on-chip ’54x RAM.
The main differences between the enhanced 8-bit HPI and the standard 8-bit
HPI are listed in Table 4−1.
Table 4−1. Main Differences Between Enhanced 8-Bit HPI and Standard 8-Bit HPI
Enhanced 8-bit HPI (HPI-8) Standard 8-bit HPI
- Access to all on-chip RAM locations Access to only a fixed 2K portion of on-chip RAM
- Host accesses always synchronized to Host-only mode allows asynchronous host accesses
the’54x clock (no host-only mode)
- Both host and ’54x always have access to Host-only mode gives host exclusive access to RAM
on-chip RAM (no host-only mode)
The HPI-8 functions as a slave and enables the host processor to access the
on-chip memory of the ’54x. The interface consists of an 8-bit, bidirectional
data bus and various control signals. Sixteen bit transfers are accomplished
in two parts with the HBIL input designating high or low byte. The host commu-
nicates with the HPI-8 through dedicated address and data registers, which
the ’54x cannot directly access. The HPI control register, which is accessible
by both the host and the ’54x, includes bits for configuring the protocol and for
controlling communication (handshaking). A simple block diagram of the
HPI-8 is shown in Figure 4−1.
4-2
Figure 4−1. Host Port Interface Block Diagram
’54x data bus
’54x address bus
HPI-8
HPIC
register ’54x
16 on-chip
RAM
HD(7−0) HPID
register
16
DMA bus
HPIA
register
16
Interface HPI
control control
signals logic
The HPI-8 provides 16-bit data to the ’54x while maintaining the economical
8-bit external interface. Successive bytes transferred are automatically com-
bined into 16-bit words. When the host device performs a data transfer with
HPI-8 registers, the HPI-8 control logic automatically performs an access to
the internal ’54x RAM to complete the transaction. The ’54x can then access
the data within its memory space.
Both the ’54x and the host can access the entire on-chip RAM of the ’54x. The
’54x clock must be active for host accesses to occur, and the HPI-8 is not func-
tional while the ’54x device is in reset mode (except the ’5410 − see section
4.8.2, Access to HPI-8 During Reset (’5410 Only), on page 4-29). The host
accesses are synchronized to the ’54x clock internally to ensure proper ar-
bitration of the on-chip RAM accesses. In the case of a conflict between a ’54x
and a host cycle where both accesses involve the same memory location, the
host has access priority and the ’54x CPU waits one ’54x clock cycle.
4-3
4.2 HPI-8 Basic Functional Description
The external HPI-8 interface can connect to a variety of host devices with little
or no additional logic necessary. The 8-bit data bus (HD0 – HD7) exchanges
information with the host. The two control inputs (HCNTL0 and HCNTL1) indi-
cate which internal HPI-8 register is accessed. These inputs, along with HBIL,
are commonly driven by host address-bus bits or a function of these bits.
Figure 4−2 shows a simplified diagram of a connection between the HPI-8 and
a host device.
Figure 4−2. Generic System Block Diagram
ÁÁÁÁÁÁÁÁÁÁÁ ÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁ
Host Device ’54x
ÁÁÁÁÁÁÁÁÁÁÁ ÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁ
8
Data HD0−HD7
ÁÁÁÁÁÁÁÁÁÁ Á ÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁ 2
ÁÁÁÁÁÁÁÁÁÁ Á ÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁ
HCNTL0/1 (address)
Address Sampled by internal
ÁÁÁÁÁÁÁÁÁÁ ÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁ
Á
HBIL 1st/2nd
strobe or HAS
Read/write HR/W
ÁÁÁÁÁÁÁÁÁÁ Á ÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁ
ÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁ HDS1
Á
Data strobe HDS2 Internal strobe (controls transfer)
ÁÁÁÁÁÁÁÁÁÁ ÁÁ ÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁ
HCS
ÁÁÁÁÁÁÁÁÁÁ ÁÁ ÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁ
Address latch HAS (Samples address and signals,
(if used) if used)
READY HRDY
Á ÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁ
Interrupt HINT
Using the HCNTL0/1 inputs, the host can specify an access to the HPI control
register (HPIC), the HPI address register (HPIA) (which serves as the pointer
into ’54x RAM), or the HPI data register (HPID). Because of the 16-bit word
structure of the ’54x, all HPI-8 transfers must consist of two consecutive bytes.
The dedicated HBIL pin indicates whether the first or second byte is trans-
ferred. An internal control register bit determines whether the first or second
byte is placed the most significant byte of a 16-bit word.
The HPID register can also be accessed with an optional automatic address
increment feature. The autoincrement feature provides a convenient way of
reading from, or writing to, consecutive word locations. In autoincrement
mode, the HPIA register is automatically incremented during consecutive
transfers. This feature is described further in section 4.3.3, Address Autoincre-
ment, on page 4-11.
4-4
The HPI-8 includes interrupt logic to facilitate software handshaking. The host
can interrupt the ’54x CPU by writing to a dedicated bit in the HPIC. Similarly,
the ’54x can use the HINT output pin to interrupt the host. The HINT output is
asserted when the ’54x writes to a dedicated bit in the HPIC. The host can also
acknowledge and clear HINT pin by writing to the HINT bit in the HPIC register.
Table 4−2 lists and describes the three registers that the HPI-8 utilizes for com-
munication between the host device and the ’54x CPU.
Table 4−2. HPI-8 Register Description
’54x
Name Address Description
HPIA — HPI address register. Directly accessible only by the host. Contains the address
in the ’54x on-chip RAM where the current access occurs.
HPIC 002Ch HPI control register. Directly accessible by either the host or by the ’54x. Contains
control and status bits for HPI-8 operations.
HPID — HPI data register. Directly accessible only by the host. Contains data that is read
from the ’54x on-chip memory if the current access is a read, or data that is written
to on-chip memory if the current access is a write.
The various strobe signals – data strobes (HDS1 and HDS2), read/write strobe
(HR/W), and the address strobe (HAS) – enable the HPI-8 to interface to a vari-
ety of host devices.
The HPI ready pin (HRDY) provides a convenient way to automatically adjust
the host access rate to the HPI-8 access rate. The HRDY pin allows insertion
of host wait states for hosts that support a ready input. This enables deferred
completion of host accesses when the host bus cycle times are faster than the
HPI-8 access rate.
All of these features combined provide a flexible and efficient interface to a

wide variety of industry-standard host devices.
4-5
4.3 Details of HPI-8 Operation
This section includes a detailed description of each HPI-8 external interface
pin function, as well as descriptions of the control register bit functions and the
HPI-8 memory map. The logical interface timings, HPI-8 initialization, and
read/write sequences are discussed in section 4.4, Host Read/Write Access
to HPI-8, on page 4-15.
The external HPI-8 interface signals implement a flexible interface to a variety

of host device types. Devices with single- or multiple-data strobes, with or with-
out address latch enable (ALE) signals, can easily be connected to the HPI-8.
Table 4−3 gives a detailed description of each HPI-8 external interface pin.
Table 4−3. HPI-8 Signal Names and Functions

HPI Pin Host Pin State† Signal Function
HAS Address latch I Address strobe input. Hosts with a multiplexed address
enable (ALE) or and data bus connect HAS to their ALE pin or equivalent.
address strobe or HBIL, HCNTL0/1, and HR/W are then latched on the
unused (tied high) HAS falling edge. When used, HAS must precede the
latter of HCS, HDS1, or HDS2 (see ’54x data sheet for
detailed HPI-8 timing specifications). Hosts with
separate address and data bus can connect HAS to a
logic-1 level. In this case, HBIL, HCNTL0/1, and HR/W
are latched by the latter of HDS1, HDS2, or HCS falling
edge while HAS stays inactive-high.
HBIL Address or control I Byte identification input. Identifies first or second byte of
lines transfer (but not most significant or least significant —
this is specified by the BOB bit in the HPIC register,
described later in this section). HBIL is low for the first
byte and high for the second byte.
HCNTL0, HCNTL1 Address or control I Host control inputs. These inputs select a host access to
lines the HPIA register, the HPI data latches (with optional
address increment), or the HPIC register.
HCS Address or control I Chip select. Serves as the enable input for the HPI-8 and
lines must be low during an access but may stay low between
accesses. HCS normally precedes HDS1 and HDS2, but
this signal also samples HCNTL0/1, HR/W, and HBIL if
HAS is not used and HDS1 or HDS2 is already low (this
is explained in further detail later in this section).
Figure 4−3, HPI Strobe and Select Logic, on page 4-8
shows the equivalent circuit of the HCS, HDS1, and
HDS2 inputs.
4-6
Table 4−3. HPI-8 Signal Names and Functions (Continued)
HPI Pin Host Pin State† Signal Function
HD0−HD7 Data bus I/O/Z Parallel bidirectional 3-state data bus. HD7 (MSB)
through HD0 (LSB) are placed in the high-impedance
state when the HPI is not active (HDSx OR HCS = 1) or
when EMU1/OFF is active-low. These pins can also be
used for general purpose input/output when the HPI-8 is
disabled. For more details on this feature, refer to section
4.9, HPI-8 Data Pins as General Purpose I/O Pins, on
page 4-31.
HDS1, HDS2 Read strobe and I Data strobe inputs. Control transfer of data during host
write strobe or access cycles. Also, when HAS is not used, these inputs
data strobe sample HBIL, HCNTL0/1, and HR/W when HCS is
already low (which is the case in normal operation).
Hosts with separate read and write strobes connect
those strobes to either HDS1 or HDS2. Hosts with a
single data strobe connect it to either HDS1 or HDS2,
connecting the unused pin high. Regardless of HDS
connections, HR/W is still required to determine direction
of transfer. Because HDS1 and HDS2 are internally
exclusive-NORed, hosts with a high true data strobe can
connect this to one of the HDS inputs with the other HDS
input connected low. Figure 4−3 on page 4-8 shows the
equivalent circuit of the HDS1, HDS2, and HCS inputs.
HINT Host interrupt O/Z Host interrupt output. Controlled by the HINT bit in the
input HPIC. Driven high when the ’54x is being reset. Placed
in high-impedance when EMU1/OFF is active-low.
HRDY Asynchronous O/Z HPI ready output. When high, indicates that the HPI-8 is
ready ready for a transfer to be performed. When low, indicates
that the HPI-8 is busy completing the internal portion of
the previous transaction. Placed in high-impedance
when EMU1/OFF is active-low. HCS enables HRDY; that
is, HRDY is always high when HCS is high.
HR/W Read/write strobe, I Read/write input. Hosts must drive HR/W high to read
address line, or HPI-8 and low to write HPI-8. Hosts without a read/write
multiplexed strobe can use an address line for this function.
address/data
The HCS input serves primarily as the enable input for the HPI-8, and the
HDS1 and HDS2 signals control the HPI-8 data transfer; however, the logic
with which these inputs are implemented allows their functions to be inter-
4-7
changed if desired. The equivalent circuit for these inputs is shown in
Figure 4−3.
Figure 4−3. HPI Strobe and Select Logic

HDS1
HDS2
Internal
HCS HPI strobe
HRDY Internal
HRDY signal
Figure 4−3 shows that the internal HPI strobe signal, which is used to control
transfers, is derived from all three input signals — HCS, HDS1, and HDS2.
Note that if HCS is used in place of HDS1 and HDS2 to control HPI-8 access
cycles, HRDY operation is affected (because HCS enables HRDY, and when
HCS is high, HRDY is forced high). It is also important to note that because
HDS1 and HDS2 are exclusive-NORed, driving both of these inputs low does
not constitute an enabled condition.
The falling edge of the internal strobe is used to sample the HCNTL0/1, HBIL,
and HR/W inputs whenever the HAS input is not used. Therefore, when HAS
is not used, the latest of HCS, HDS1, or HDS2 is the signal that actually con-
trols sampling of the HCNTL0/1, HBIL, and HR/W inputs. In addition to sam-
pling the control inputs, the internal strobe defines the boundaries of an HPI-8
cycle. This function of the internal strobe is described in section 4.4, Host
Read/Write Access to HPI-8, on page 4-15.
When using the HAS input to sample HCNTL0/1, HBIL, and HR/W, these sig-
nals can be removed earlier in an access cycle, thus allowing more time to
switch bus states from address to data information. This additional time facili-
tates interface to hosts with multiplexed address and data busses. In this type
of system, an address latch enable (ALE) signal is often provided and can be
used to drive the HAS input.
The two control pins (HCNTL0 and HCNTL1) indicate which internal HPI-8
register is accessed. Table 4−4 describes the HCNTL0/1 pin functions.
4-8
Table 4−4. HPI-8 Input Control Signals and Function Selection
HCNTL1 HCNTL0 Description
0 0 Host can read from or write to the HPI control register, HPIC.
0 1 Host can read from or write to the HPI data latches. HPIA is automatically
postincremented each time a read is performed, and preincremented each time a write
is performed.
1 0 Host can read from or write to the address register, HPIA. This register points to the ’54x
on-chip RAM.
1 1 Host can read from or write to the HPI data latches. HPIA is not affected.
4.3.1 HPI-8 Address Register and Memory Map

The host uses the HPIA register as a pointer to ’54x on-chip memory, and all
on-chip RAM locations are accessible through the HPI-8. Because the internal
memory map of each ’54x device is unique, the address range that the HPI-8
can access varies from device to device. For example, the ’VC5410 includes
much more on-chip RAM than the ’VC5402. The HPI-8 memory maps for the
’5402 and ’5410 devices are shown in Figure 4−4.
Figure 4−4. HPI-8 Memory Maps

TMS320VC5402 TMS320VC5410
0000h 0000h
Reserved Reserved
005Fh 005Fh
0060h 0060h
Scratch-Pad RAM Scratch-Pad RAM
007Fh 007Fh
0080h 0080h
On-chip DARAM
On-chip DARAM
(16K × 16 bits)
3FFFh 1FFFh
4000h 2000h On-chip SARAM1
Undefined
FFFFh 7FFFh (24K × 16 bits)
8000h
Undefined
17FFFh
18000h On-chip SARAM2
1FFFFh (32K × 16 bits)
4-9
All on-chip RAM blocks—program RAM and data RAM—are mapped to one
contiguous address range within the HPI memory map. The addresses within
this memory map cannot be dynamically re-mapped by the user (that is, the
HPI-8 memory map is not affected by any programmable register bits).
4.3.2 Extended HPI-8 Addressing

For devices that include on-chip RAM mapped beyond the first 64K address-
es, the HPI-8 includes an extended addressing feature. Seven extended
address bits are internally available to allow HPI-8 access to extended pages
of on-chip RAM. The host accesses this extended addressing feature using
the extended HPI address (XHPIA) bit of the HPIC register. When the host sets
the XHPIA bit, a 7-bit register representing the extended address bits
(HPIA16:22) is accessible in place of the HPIA register. To initialize these ex-
tended address bits, the host must perform a write access to the HPIA register,
with the seven LSBs of each byte designating the value of HPIA 16:22. Note
that during this write access, the first and second byte values are both written
to the same register. Therefore, if the same value is not written for both bytes,
the second byte value is used to initialize the extended addresses, and the first
byte value is discarded.
After initializing the extended address bits, the host must clear the XHPIA bit
to regain access to the lower sixteen HPI address bits in the HPIA register. The
XHPIA bit must also be cleared to zero for proper operation of the autoincre-
ment feature. The autoincrement feature does not function properly when the
XHPIA bit is set to one.
It is important to note that neither the XHPIA nor the seven extended address
bits are initialized after reset, so the host should always initialize these bits af-
ter the ’54x is reset. The XHPIA bit is described in more detail in section 4.3.4,
HPI-8 Control Register Bits and Functions, on page 4-12.
4-10
4.3.3 Address Autoincrement
The HPI-8 address autoincrement feature provides a convenient way of ac-
cessing consecutive word locations in on-chip RAM. When the autoincrement
feature is enabled, the HPIA register is automatically incremented for each ac-
cess. Although the access times are not changed, performance is increased
because the host does not have to update the HPIA register between each
memory access. The autoincrement feature is enabled when the HCNTL0 pin
is set to one and the HCNTL1 pin is set to zero. Note that for devices with ex-
tended on-chip RAM, the XHPIA bit of the HPIC register must be set to one for
proper autoincrement operation.
When autoincrement is enabled, a data read causes a postincrement of the

HPIA, and a data write causes a preincrement of the HPIA. Therefore, to write
a particular address using autoincrement, the HPIA register should be initial-
ized to address-1. The increment function affects all 16 bits of the HPIA, and
on devices with extended on-chip RAM (except the ’5410), the increment fea-
ture also affects the extended addresses. If the HPIA is set to FFFFh and auto-
increment is enabled, the next access will change the HPI address to 010000h.
Because the autoincrement circuitry of the ’5410 does not affect the extended
HPI addresses, the above example will change the ’5410 HPI address to
000000h.
4-11
4.3.4 HPI-8 Control Register Bits and Functions
The bits of the HPIC register control and monitor HPI-8 operation. These bits
are: BOB (selects first or second byte as most significant), DSPINT and HINT
(can be used to generate ’54x and host interrupts, respectively), XHPIA (used
by the host to access extended addresses), and HPIENA (indicates HPI-8 en-
abled or disabled status). Table 4−5 presents a detailed description of the
HPIC bit functions.
Table 4−5. HPI Control Register (HPIC) Bit Descriptions

Reset
Bit Value Function
BOB 0 Byte-order bit. This bit determines the placement for the two bytes of a transfer. If
BOB = 1, the first byte of a transfer is least significant. If BOB = 0, the first byte is most
significant. This bit can only be accessed (written or read) by the host, and it must be
initialized before the first data or address register access.
DSPINT 0 Host-to-’54x interrupt. When the host writes a 1 to this bit, a ’54x interrupt is generated.
The bit can only be written to by the host, and is always read as 0 by both the host and
the ’54x. When the host writes to HPIC, both bytes must write the same value. For a
detailed description of the DSPINT function, see section 4.5 on page 4-24.
HINT 0 ’54x-to-host interrupt. This bit determines the state of the ’54x HINT output, which can
be used to interrupt the host. When the HINT bit is set to 1, the HINT output is driven low,
and when the bit is cleared to 0, the output is driven high. The HINT bit can only be set
by the ’54x, and it can only be cleared by the host. The host clears the bit by writing a
1 to it. For a detailed description of the HINT function, see section 4.5 on page 4-24.
XHPIA X Extended address enable. When XHPIA=1, host writes to the HPIA register are loaded
into the most significant bits HPIA[n:16]. If XHPIA=0, host writes to HPIA are loaded into
HPIA[15:0]. All n+1 address bits are incremented in the autoincrement mode. Reading
the HPIA register is performed in the same manner. Only the host has access to this bit.
HPIENA X HPI enable status bit. This bit latches the reset value of the HPIENA pin, and can be used
by the ’54x to determine if the HPI-8 is enabled or disabled. This bit is not affected by
writes, and is not available to the host.
Note: This bit is not available on all devices. For more details, see the specific HPIC
register diagrams that follow.
4-12
The HPIC is organized on the host side as a 16-bit register with the same high
and low byte contents (although access to certain bits is limited, as described
previously). The upper 8 bits of the HPIC are unused on the ’54x side. The host
accesses the HPIC register using the appropriate selection of HCNTL0/1, and
performs two consecutive byte accesses to the 8-bit HPI-8 data bus. When the
host writes to HPIC, both the first and second byte written must be the same
value. The ’54x accesses the HPIC as a memory-mapped register at address
002Ch in data memory space.
The layout of the HPIC bits is shown in Figure 4−5. In this figure, a zero is spe-
cified for a read operation to indicate that the bit is always read as zero; similar-
ly, an X specifies that the read value for this bit is indeterminate. For write op-
erations, a one is specified to indicate that the bit must only be written with a
one; an X is specified for a write operation to indicate that the bit can be written
with a zero or a one. Note that unused bits are reserved for future expansion
and it is recommended that they be written as zero, unless specified otherwise.
Figure 4−5. HPIC Diagram — Host and ’54x Accesses

Host Reads From HPIC
15−13 12 11 10−9 8 7−5 4 3 2 1 0
0 XHPIA† HINT 0 BOB 0 XHPIA† HINT 0 X BOB
Host Writes to HPIC
15−13 12 11 10 9 8 7−5 4 3 2 1 0
X XHPIA† HINT DSPINT X BOB X XHPIA† HINT DSPINT X BOB
’54x Reads From HPIC
15−8 7 6−4 3 2 1 0
0 HPIENA‡ 0 HINT 0 X 0
’54x Writes to HPIC
15−4 3 2 1 0
X HINT X 1 X
X denotes bits that are unaffected by writes, or bits that can be read as either 1 or 0.
† This bit is only available on ’54x devices with on-chip RAM mapped in extended addresses.
‡ This bit is not available on the ’5410.
Because the ’54x can write to the HINT bit, which is read twice on the host inter-
face side, the first and second byte reads by the host may yield different data
if the ’54x changes the state of this bit in between the two host read operations.
The characteristics of host and ’54x HPIC read/write cycles are summarized
in Table 4−6.
4-13
Table 4−6. HPIC Host and ’54x Read/Write Characteristics
Device Read Write

Host 2 bytes 2 bytes (Both bytes must be equal)
’54x 16 bits 16 bits
4-14
4.4 Host Read/Write Access to HPI-8
An HPI-8 transfer is composed of an external portion in which the host ex-
changes data with the HPI-8 registers, and an internal portion in which the
HPI-8 logic exchanges data between the registers and on-chip RAM. The ex-
ternal portion of the transfer always requires two bytes, regardless of the type
of access — HPIA, HPIC, or data access. During the external portion of the
transfer, the host drives the HBIL input low for the first byte, and then high for
the second byte. If the host breaks this sequence of first byte/second byte
(HBIL low/high) during an ongoing HPI-8 transfer, data may be lost and an un-
predictable operation may result. To recover from such a condition, the host
must repeat the access with the correct HBIL polarities for each byte.
Also for each byte, the host controls the duration of the access using the strobe
and select inputs — HDS1, HDS2, and HCS. The derivation of the internal HPI
strobe from these inputs is described in section 4.3, Details of HPI-8 Opera-
tion, on page 4-6. The falling edge of the HPI strobe marks the beginning of
the first or second byte of a transfer. This event usually occurs at the beginning
of the host bus cycle. The rising edge of the HPI strobe marks the end of the
first or second byte of an HPI transfer. This event usually occurs at the end of
the host bus cycle. During the exchange of the second byte, the rising edge
of the HPI strobe marks the end of the external portion of the cycle and initiates
the internal portion of the cycle. Consequently, the HPID value read during the
external portion of a memory read is the contents of the address specified in
the previous access. A typical HPI-8 access is shown in Figure 4−6.
4-15
Figure 4−6. HPI-8 Timing Diagram
Byte 1 Byte 2
HCNTL0/1 Valid Valid
HR/W
HBIL
HCS
HAS
(if used)
Á Á ÁÁ ÁÁ
HDS1,
HDS2
Á Á ÁÁ ÁÁ
Á Á Á ÁÁ Á ÁÁ
HD Valid Valid
read
HD
Á Á Á ÁÁ
Á Á Á ÁÁ
Valid Valid
write
Internal portion
HRDY External portion of transfer of transfer
Internal portion of previous transfer
The cycle begins with the host driving HCNTL0/1, HR/W, HBIL, and HCS, indi-
cating the type of transfer and whether the cycle is to be a read or write. Then
the host asserts the HAS signal (if used) followed by one of the data strobe
signals. If HRDY is not already high, it goes high when the previous internal
cycle is complete, allowing data to be transferred. Following the external HPI-8
cycle, HRDY goes low (HPI-8 not ready) and stays low while the internal por-
tion of the transfer is accomplished. HRDY then switches high again when the
internal portion of the cycle is complete. You should note, however, that HRDY
is always high when HCS is high.
4-16
4.4.1 Latency of HPI-8 Accesses
HPI-8 access time is composed of the time required for the external portion of
a transfer and the time required for the internal portion. The time required for
the external portion of an HPI-8 transfer is a fixed delay based on the setup
times, hold times, and buffer delay times of the various signal pins. The internal
delay of an HPI-8 transfer usually accounts for the majority of the access time
and is related to the ’54x clock frequency. This internal delay varies depending
on the type of access. An accurate analysis of the HPI-8 access time requires
consideration of both the internal and external delays. Because the HPI strobe
edges define the boundaries of an access, analysis of the timing requirements
for these signals is sufficient. The timing requirements for HDS1,2 are given
in the device-specific data sheets, and these timings incorporate both the ex-
ternal and internal delays of an HPI-8 access. The relationship between the
HPI-8 internal access delay and the ’54x clock rate is explained in the para-
graphs that follow.
The fastest accesses are:

- Read accesses to the HPIC and HPIA registers
- Write accesses to the HPIC register with the HINT and DSPINT bits written
as zeros
These accesses only exchange data between the host and the internal regis-
ters, so no additional internal access time is required. The HRDY signal is nev-
er driven low for these accesses, since another HPI-8 access can begin as
soon as the external portion of the access is complete.
Write accesses to the HPIC register with a one written to either the HINT or
DSPINT bits have a longer latency. These accesses require some time after
the external portion of the transfer completes for the internal interrupt logic of
the HPI-8 to perform its function. The HRDY signal is held low for approximate-
ly three ’54x clock cycles during the internal portion of these accesses.
The slowest and most common HPI-8 accesses are memory accesses. A write
access to the HPIA register, or a read/write access to the HPID register, initi-
ates an internal memory transfer that exchanges the desired data between the
HPID and the on-chip ’54x memory. This process requires several ’54x clock
cycles each time an HPI-8 memory access is made. The minimum time re-
quired for the internal portion of an HPI-8 memory transfer is five ’54x clock
cycles.
4-17
The maximum duration for the internal portion of an HPI-8 memory transfer de-
pends on several factors. The HPI-8 and direct memory access (DMA) control-
ler share the DMA bus for internal memory accesses, and only one of the mod-
ules can access the bus at a time. (For more details on the DMA controller, see
Chapter 3.) When HPI-8 accesses and DMA controller accesses contend for
bus usage, internal arbitration logic resolves the conflict. If the HPI-8 and DMA
controller initiate an access at the same time, the HPI-8 has priority and the
DMA controller transfer is postponed for one ’54x clock cycle. If the HPI-8
initiates an access while a DMA transfer is in progress, the HPI-8 access is
delayed (HRDY stays low) until the current word of the DMA transfer is com-
plete. This delay time, during which the host waits for the DMA controller, de-
pends on the type of DMA transfer in progress. The internal delay for each of
the HPI-8 access types are summarized in Table 4−7.
Table 4−7. HPI-8 Internal Delays by Access Type
Access Type Internal Delay
Read access to HPIA or HPIC or No internal delay

write access to HPIC with HINT and
DSPINT cleared to zero.
Write access to HPIC with either 3 CPU clock cycles

HINT or DSPINT set to one.
Memory access when DMAC is inac- 5 CPU clock cycles

tive
16 bit DMAC 5 CPU clock cycles + 4 CPU clock cycles

transfer to/from
Memory access
on-chip RAM†
when DMAC is
active 32 bit DMAC 5 CPU clock cycles + 8 CPU clock cycles
transfer to/from
on-chip RAM†
† Note: additional latency can occur due to wait-states when the DMAC accesses off-chip
memory.
Because the HPI-8 access times vary by access type, the host should always
sample the HRDY output to adjust its bus cycle time for the varying HPI-8 rate.
This sampling can be accomplished by connecting HRDY to the host READY
input (if available) to generate host wait-states. Alternatively, the host bus
cycle time can be set to the slowest possible HPI-8 access time. The function
of HRDY for the various HPI-8 access types is summarized in Table 4−8.
4-18
Table 4−8. Wait-State Generation Conditions
Wait States Generated (HRDY driven low)

Register Reads Writes
HPIC No 1 to DSPINT/HINT − Yes
All other cycles − No
HPIA No Yes
HPID Yes Yes
Due to the prefetch nature of internal HPI-8 accesses, special care must be
taken under certain conditions. During random (non-sequential) transfers, or
sequential accesses with a significant amount of time between them, it is pos-
sible for the ’54x to change the contents of the location being accessed during
the time between a host read and the previous host access. If this occurs, the
data read may be different from the actual memory contents being accessed.
Where this is of concern in a system, two reads from the same address, or an
address write prior to the read access, can be made to ensure that the most
recent data is read.
4-19
4.4.2 Access Sequence Examples
Before accessing data, the host must first initialize HPIC, in particular BOB (bit
0 and bit 8), and then the HPIA register. The initialization must occur in this or-
der because the state of BOB affects the HPIA register access.
On devices with extended on-chip RAM, the host should also initialize the
XHPIA bit of the HPIC before accessing the HPIA register. The XHPIA bit can
be initialized in the same write access to the HPIC that initializes BOB. By writ-
ing a one to the XHPIA bit, the host gains access to the seven extended HPI
addresses. The host then writes the HPIA register with the seven LSBs desig-
nating the value of the extended addresses (HPIA 16:22). When initializing the
extended HPI addresses, the same value should be written for the first and
second bytes of the access. After initializing the extended addresses, the host
must perform another access to the HPIC, writing a zero in the XHPIA bit to
regain access to the lower sixteen address bits in the HPIA register. The ex-
tended address feature is discussed in more detail in section 4.3.2. on page
4-10.
After initializing BOB, the host can then write to the HPIA register with the cor-
rect byte alignment. When the host writes to the HPIA, the on-chip ’54x
memory is automatically read and the contents at the given address are trans-
ferred to the two 8-bit data latches — the first and second bytes of the HPID
register.
Table 4−9 illustrates the sequence involved in initializing the BOB and XHPIA
bits of the HPIC and HPIA registers for an HPI-8 memory read. Note that the
first six rows of the table show the initialization of the extended address fea-
ture, which is only required for devices with extended on-chip RAM. In this ex-
ample, BOB is set to zero and a read is requested of the first on-chip memory
location (in this case 0060h) which contains FFFEh.
4-20
Table 4−9. Initialization of BOB and HPIA
HPID
Event HRDY HD HR/W HCNTL1/0 HBIL HPIC HPIA Latch1 Latch2

Host writes HPIC, 1 10 0 00 0 10xx xxxx xxxx xxxx
1st byte†
Host writes HPIC, 1 10 0 00 1 1010 xxxx xxxx xxxx

2nd byte†
Host writes HPIA, 1 00 0 10 0 1010 xxxx xxxx xxxx

1st byte†
Host writes HPIA, 1 00 0 10 1 1010 xx00 xxxx xxxx

2nd byte†
Internal delay† 0 xxxx xxxx
Internal HPI RAM 1 xxxx xxxx

read complete†
Host writes HPIC, 1 00 0 00 0 00xx xxxx xxxx xxxx

1st byte
Host writes HPIC, 1 00 0 00 1 0000 xxxx xxxx xxxx

2nd byte
Host writes HPIA, 1 00 0 10 0 0000 00xx xxxx xxxx

1st byte
Host writes HPIA, 1 60 0 10 1 0000 0060 xxxx xxxx

2nd byte
Internal delay 0 0060 xxxx xxxx
Internal RAM read 1 0060 FF FE

complete
† These accesses initialize the extended addressing feature and are only required for devices with extended on-chip RAM.
In the example shown in Table 4−9, the BOB and XHPIA bits of HPIC are initial-
ized first; then the write access to the seven LSBs of the HPIA register initial-
izes the extended HPI address bits. The first and second bytes of this write ac-
cess must write the same values for proper initialization of the extended HPI
address bits. Notice that the write access to the HPIA register automatically
initiates an internal read access; however, the data read is indeterminate be-
cause the lower sixteen address bits are not yet initialized. These steps are
only required for devices with extended on-chip RAM.
4-21
Next, the host initializes the lower 16 address bits by writing a zero to the
XHPIA bit and writing the 16-bit address to the HPIA register. As before, writing
to the HPIA register automatically initiates an internal HPI-8 memory access.
This time, all address bits are initialized so the internal read retrieves the con-
tents of the specified memory location. The last line of Table 4−9 shows the
condition of the HPI-8 after the internal RAM read is complete. That is, after
a delay following the end of the HPIA register write, the read is completed and
the data is placed in the HPID register. The host must perform an additional
read of HPID to actually retrieve this data. The sequence involved in this ac-
cess is shown in Table 4−10.
Table 4−10. Read Access to HPI-8 With Autoincrement

HPID
Host reads data, 1 FF 1 01 0 0000 0060 FF FE

1st byte
Host reads data, 1 FE 1 01 1 0000 0061 FF FE
2nd byte
Internal delay 0 0061 FF FE

Internal HPI RAM 1 0061 6A BC
read complete
In the access shown in Table 4−10, the data obtained from reading HPID is the
data from the read initiated in the previous cycle (the one shown in Table 4−9).
During this HPID read access, the contents of the first byte data latch are driv-
en on the HD pins when HBIL is low, and the contents of the second byte data
latch are driven on the HD pins when HBIL is high. The access performed in
Table 4−10 also initiates another read of location 0061h (because autoincre-
ment was specified in this access by setting HCNTL1/0 to 01). When autoin-
crement is selected, the increment occurs with each 16-bit word transferred
(not with each byte); therefore, as shown in Table 4−10, the HPIA is increm-
ented by one. The last line of Table 4−10 indicates that after some delay follow-
ing the read of the second byte, the contents of location 0061h (6ABCh) are
read and placed in the HPID register.
During a write access to the HPI, the first byte data latch is overwritten by the
data coming from the host while the HBIL pin is low, and the second byte data
latch is overwritten by the data coming from the host while the HBIL pin is high.
At the end of this write access, the bytes in both data latches are transferred
as a 16-bit word to the on-chip RAM at the address specified by the HPIA regis-
ter. The address is incremented prior to the memory write if autoincrement is
selected.
4-22
An HPI write access is illustrated in Table 4−11. In this example, autoincrement
is enabled and the HPIA register is incremented before the write occurs (prein-
cremented). Because the previous read access caused a postincrement of the
HPIA register, address 0061h is skipped. After the internal portion of the write
is completed, location 0062h of on-chip RAM contains 1234h. If a read of ad-
dress 0062h follows this write, the same data (1234h) is read back.
Table 4−11. Write Access to HPI With Autoincrement

HPID

Host writes data, 1 12 0 01 0 0000 0061 12 FE
1st byte
Host writes data, 1 34 0 01 1 0000 0062 12 34

2nd byte
Internal delay 0 0062 12 34
Internal RAM 1 0062 12 34

write complete
4-23
4.5 DSPINT and HINT Operation
The host and the ’54x can interrupt each other using bits in the HPIC register.
The sections that follow explain this process.
4.5.1 Host Device Using DSPINT to Interrupt the ’54x

A ’54x interrupt is generated when the host writes a one to the DSPINT bit in
HPIC. This interrupt can be used to wake up the ’54x from IDLE. The host and
the ’54x always read this bit as zero. A ’54x write has no effect. After a one is
written to DSPINT by the host, it is not necessary for a zero to be written before
another interrupt can be generated. Writing a zero to this bit has no effect. The
host should not write a one to the DSPINT bit while writing to BOB or HINT;
otherwise, an unwanted ’54x interrupt is generated.
On the ’54x, the host-to-’54x interrupt vector address is xx64h. This interrupt
is located in bit nine of the IMR/IFR registers. Because the ’54x interrupt vec-
tors can be re-mapped into the on-chip RAM, the host can instruct the ’54x to
execute pre-programmed functions by initializing the DSPINT interrupt vector.
When the ’54x interrupts are re-mapped to on-chip RAM, the host can: 1) write
the opcode for a branch instruction at address xx64h, and 2) write the start ad-
dress of a function at address xx65h in the interrupt vector table prior to inter-
rupting the ’54x. When using this technique, care must be taken to prevent the
host from corrupting the other interrupt vectors.
4.5.2 ’54x Using HINT to Interrupt the Host Device

When the ’54x writes a one to the HINT bit in HPIC, the HINT output is driven
low, and the HINT bit is read as a one. Consequently, the HINT signal can be
used as an active-low interrupt source for a host device. The host can clear
this interrupt and cause the HINT signal to be deasserted by writing a one to
the HINT bit. The HINT bit is cleared by this write, and the HINT signal is driven
high. If the ’54x or host writes a zero, the HINT bit remains unchanged.
4-24
4.6 Considerations for HPI-8 Transfers While Changing Clock Modes
Unlike the standard 8-bit HPI, the HPI-8 does not include an asynchronous
host-only mode (HOM). All HPI-8 transfers are synchronized to the ’54x clock.
This requires special considerations for HPI-8 transfers while changing clock
modes because a change in the ’54x clock frequency causes a change in the
HPI-8 access rate. For more information on changing the ’54x clock generator
modes, refer to TMS320C54x DSP Reference Set, Volume 1: CPU and Pe-
ripherals (literature number SPRU131). The relationship between the ’54x
clock rate and the HPI-8 access rate is described in section 4.4.1 on page 4-17.
When the ’54x clock mode is changed from PLL mode to DIV mode, the result-
ing clock rate is slower, and the HPI-8 access rate is also slower. In this case,
the host can use the HRDY output to generate wait-states and adjust to the
change in access rate. An example of HPI-8 accesses while changing from
multiply-by-1 (MULT1) clock mode to divide-by-2 (DIV2) clock mode is shown
in Figure 4−7.
Figure 4−7. HPI-8 Transfers While Switching to DIV Clock Modes

MULT1 clock mode DIV2 clock mode
CLKIN
CLKOUT
HDS1,2
HRDY
ÁÁ ÁÁÁÁ Byte 1
Á Byte 1
ÁÁÁ
HD[0:7]
Host
wait-
ÁÁ ÁÁÁÁ
Host
wait-
Host
wait-
Byte 2 Host
wait-
Host
wait-
HostÁ
wait-
Host
wait-
ÁÁÁByte 2
state state state state state state state
In the example shown in Figure 4−7, the difference in the ’54x clock frequency
before and after switching clock modes is only a factor of two, and the resulting
effect on HPI-8 access time is small. In this case, the host can compensate for
the change in access rate with one additional wait-state. When higher clock
multiply ratios are used, switching to a divide-by mode causes a significant de-
crease in the HPI-8 access rate, and several additional host wait-states may
be required.
4-25
When the ’54x clock generator is switched from a divide-by mode to a multiply
mode, the resulting ’54x clock rate and HPI-8 access rate are increased. In this
case, if host wait-states are generated using HRDY, then the host automatical-
ly reduces the number of wait-states in a similar manner to that shown in
Figure 4−7. If the resulting HPI-8 access rate is faster then the host bus cycle,
then no wait-states are required and the host bus-cycle time defines the new
access rate.
4-26
4.7 Considerations in IDLE Use
The HPI-8 allows the host to access on-chip RAM while the ’54x is in an IDLE
mode. The only requirement for HPI-8 access during these modes is that the
’54x input clock (CLKIN) must remain active. For more information on the IDLE
modes of the ’54x, refer to the TMS320C54x DSP Reference Set, Volume 1:
CPU and Peripherals (literature number SPRU131).
4.7.1 HPI-8 Accesses During IDLE1 and IDLE2
Because the IDLE1 and IDLE2 modes do not affect the ’54x clock, no special
considerations are required for HPI-8 transfers during these modes. However,
in some cases, you may wish to disable the PLL circuit prior to initiating these
power-down modes to further reduce power consumption. In these cases, the
HPI-8 access rate is affected and special considerations must be made. For
more information on considerations for HPI-8 usage during clock mode
changes, refer to section 4.6 on page 4-25.
4.7.2 HPI-8 Accesses During IDLE3
The HPI-8 access rate is usually affected by the IDLE3 power-down mode
because this mode affects the internal clock generation circuitry of the ’54x.
When the ’54x enters the IDLE3 mode, the on-chip clock generator is disabled
and HPI-8 accesses are automatically synchronized to the input clock source
(CLKIN). The host can continue to access the HPI-8 while the ’54x is in the
IDLE3 mode as long as an active clock is applied to the ’54x CLKIN pin. HPI-8
accesses remain synchronized to CLKIN until the ’54x is taken out of IDLE3
mode. After the ’54x is taken out of IDLE3 mode by an interrupt source, HPI-8
accesses are automatically synchronized to the output of the ’54x clock gener-
ator. If the clock generator is configured in PLL mode, HPI-8 accesses remain
synchronized to the input source until the PLL lock timer counts down. For
more information on the ’54x clock generator and programmable lock timer,
refer to TMS320C54x DSP Reference Set, Volume 1: CPU and Peripherals
(literature number SPRU131).
When the ’54x enters the IDLE3 mode while a clock multiply ratio of one is se-
lected, the HPI-8 access rate remains the same (×1) during IDLE3. Also, if the
clock generator is configured in DIV (divider) mode when the ’54x enters
IDLE3, then the HPI-8 access rate is increased (×1) while the ’54x is in IDLE3.
In both of these cases, no special considerations are required because the
HPI-8 access rate is neither unaffected nor improved during IDLE3.
4-27
When the ’54x enters IDLE3 mode while a clock multiply ratio greater than one
is selected, the HPI-8 access rate is reduced. In this case, the host can use
the HRDY output to generate additional wait-states and adjust to the change
in access rate. An example of this case is shown in Figure 4−8.
Figure 4−8. HPI-8 Transfers While the ’54x is in IDLE3 Mode
IDLE3 mode IDLE3 mode

entered exited PLL locked
PLL DISABLED
MULT2 clock mode MULT2 clock mode
CLKIN
CLKOUT
HDS1,2
Á ÁÁÁÁ ÁÁÁÁ ÁÁÁÁ

HRDY Byte 1 Byte 1 Byte 1
HD[0:7]
Byte 2 Byte 2 Byte 2
Host Host Host Host Host Host Host Host Host Host
wait- wait- wait- wait- wait- wait- wait- wait- wait- wait-
state state state state state state state state state state
In Figure 4−8 the clock generator is configured in PLL mode with a multiply ra-
tio of two when the ’54x enters IDLE3 mode. HPI-8 accesses are automatically
synchronized to the input clock source, and the host must insert an additional
wait-state for memory accesses based on the state of the HRDY pin. The fig-
ure also shows that HPI-8 accesses are automatically re-synchronized to the
faster clock after the ’54x is taken out of IDLE3 mode and the clock generator
stabilizes.
4-28
4.8 Effects of Reset on HPI-8 Operation
HPI-8 operation is affected when the ’54x device is reset. This section de-
scribes the effect of reset on the HPI-8 operation.
4.8.1 Accesses to HPI-8 After Reset

The HPI-8 does not allow a host access while the ’54x device is in reset (except
’5410, see section 4.8.2). After the ’54x device is released from reset, the host
can access the HPI-8. After the ’54x device is released from the reset, the host
can access the HPI-8. Because the HPI-8 access rate depends on the ’54x
clock rate, and the ’54x clock rate is re-initialized by reset, care must be taken
to match the speed of host accesses to the HPI-8 access rate selected after
reset. The ’54x clock generator is initialized at reset to the mode selected by
the clock mode pins (CLKMD1−3). If the host has control of the ’54x reset pin
and clock mode pins, then it can configure the HPI-8 access rate after reset.
For more information on the clock generator and clock mode pins, refer to the
device-specific data sheets and to TMS320C54x DSP Reference Set, Volume
1: CPU and Peripherals (literature number SPRU131).
4.8.2 Access to HPI-8 During Reset (’5410 Only)

The HPI-8 of the ’5410 supports the host-only-mode (HOM) during reset for
compatibility with previous versions of the 8-bit HPI. The ’5410 is not opera-
tional during reset, but the host can access the HPI-8, allowing program or data
downloads to the on-chip memory. Accesses during reset are not internally
synchronized to the ’54x clock, and the HRDY output is never de-asserted (ex-
cept when writing a one to the DSPINT or HINT bits of the HPIC). The host
should not write a one to either the DSPINT or HINT bits of the HPIC register
while the ’5410 is in reset because the HPI-8 will not function properly, and the
HRDY output will be de-asserted indefinitely. For the HPI-8 access time during
reset, refer to the ’5410 data sheet (see Related Documentation from Texas
Instruments in the Preface).
When HOM is used during reset, it is often convenient for the host to control
the ’5410 reset input. The sequence of events for resetting the ’5410 and
downloading a program to on-chip RAM while the ’5410 is in reset is summa-
rized in Table 4−12.
4-29
Table 4−12. HPI-8 Operation During RESET (’5410 only)
Host ’5410 Mode ’5410 CLK
Waits 6 ’5410 clock periods Running X Running
Brings RESET low and waits 4 clocks Goes into reset HOM Running
Can stop ’5410 clock In reset HOM Stopped or running
Writes program and/or data in on-chip RAM In reset HOM Stopped or running
Turns on DSP clock if it was stopped† In reset HOM Running
Brings RESET high In reset HOM Running
Waits 20 ’5410 clock periods Comes out of reset SAM Running
Can access HPI-8 Running SAM Running

† Sufficient wake-up time must be ensured when the ’5410 on-chip PLL is used.
Initially, the host stops accessing the HPI-8 at least six ’5410 clock periods be-
fore driving the ’5410 reset line low. The host then drives the ’5410 reset line
low and can start accessing the HPI-8 after a minimum of four ’5410 clock
periods. The HPI-8 mode is automatically set to HOM during reset, allowing
a high-speed program download. The ’5410 clock can even be stopped at this
time; however, the clock must be running when the reset line falls and rises for
proper reset operation of the ’5410.
Once the host has finished downloading into on-chip RAM, the host stops ac-
cessing the HPI-8 and drives the ’5410 reset line high. At least 20 ’5410 peri-
ods after the reset line rising edge, the host can again begin accessing the
HPI-8. This number of periods corresponds to the internal reset delay of the
’5410. The HPI-8 mode is automatically set to SAM upon exiting reset. If the
host writes a one to DSPINT while the ’5410 is in reset, the interrupt is lost
when the ’5410 comes out of reset. The ’5410 HPI-8 boot mode can then be
used to start execution from address 02000h.
4-30
4.9 HPI-8 Data Pins as General Purpose I/O Pins (Not Available on ’5410)
The 8-bit bidirectional data bus of the HPI-8 can be used as general-purpose
input/output (GPIO) pins. This feature is only available when the HPI-8 is dis-
abled; that is, when the HPIENA pin is driven low during reset. Two memory-
mapped registers are used to control the GPIO function of the HPI-8 data pins:
the general-purpose I/O control register (GPIOCR), and the general-purpose
I/O status register (GPIOSR). The GPIOCR is shown in Figure 4−9, and its bits
are described in Table 4−13.
The direction bits (DIRx) of the GPIOCR are used to configure HD0-HD7 as
inputs or outputs. The timer1 output bit (TOUT1) is available on devices that
include two timers to enable the timer1 output on the HINT pin. When the HPI-8
is enabled, the TOUT1 bit and DIRx bits are forced to zero, and the general
purpose I/O feature functions in input mode only.
Figure 4−9. General Purpose I/O Control Register (GPIOCR) MMR Address 003Ch
15 14 8
TOUT1† rsvd
R/W−0
7 6 5 4 3 2 1 0
DIR7 DIR6 DIR5 DIR4 DIR3 DIR2 DIR1 DIR0
R/W−0 R/W−0 R/W−0 R/W−0 R/W−0 R/W−0 R/W−0 R/W−0
† Only available on devices with a second on-chip timer.
Note: R = Read, W= Write
Table 4−13. General Purpose I/O Control Register (GPIOCR) Bit Functions
Reset
No. Name Value Function
15 TOUT1 0 Timer1 output enable bit. The TOUT1 bit enables or disables the timer1
output on the HINT pin. The timer1 output is only available when the HPI-8
is disabled. Note, this bit is reserved on devices that have only one timer.
TOUT1 = 0 The timer1 output is not available externally.
TOUT1 = 1 The timer1 output is driven on the HINT pin.
14−8 rsvd 0 These pins are reserved and unaffected by writes.
7 DIR7 0 I/O pin 7 direction bit. DIR7 configures the HD7 pin as input or output.
DIR7 = 0 The HD7 pin is configured as an input.
4-31
(Continued)
Reset
DIR7 = 1 The HD7 pin is configured as an output. When the HPI-8 is
enabled, this bit is forced to 0, and is not affected by writes.

enabled, this bit is forced to 0 and is not affected by writes.





0 DIR0 0 I/O pin 0 direction bit. DIR0 configures the HD0 pin as input or
output.
4-32
(Continued)
Reset

The status of the GPIO pins (HDx, x=0:7) can be monitored using the bits of
the general purpose I/O status register (GPIOSR). When an HDx pin is config-
ured as an input (by writing a one to the DIRx bit of the GPIOCR), the corre-
sponding bit in the GPIOSR can be read to determine the logic value sensed
at the pin. Similarly, when an HDx pin is configured as an output, the logic value
to be driven by the pin is written to the corresponding bit in the GPIOSR. The
GPIOSR is shown in Figure 4−10, and its bits are described in Table 4−14.
Figure 4−10. General Purpose I/O Status Register (GPIOSR) MMR address 003Dh
15 8
rsvd
7 6 5 4 3 2 1 0
IO7 IO6 IO5 IO4 IO3 IO2 IO1 IO0
R/W−x R/W−x R/W−x R/W−x R/W−x R/W−x R/W−x R/W−x
Note: R = Read, W= Write
† Note, when the HD pins are configured as inputs, writes to the IOx bits have no effect.
Table 4−14. General Purpose I/O Status Register (GPIOSR) Bit Functions
Reset
No. NAME Value Function
15−8 rsvd 0 These pins are reserved, and unaffected by writes.
7 IO7 X IO7 - I/O pin 7 status bit. This bit reflects the logic level on the HD7 pin. When
the HD7 pin is configured as an input (DIR7 = 0 in the GPIOCR), the IO7 bit
latches the logic value of the pin (1 or 0). Writes to the IO7 bit have no effect
when the HD7 pin is configured as an input. When the HD7 pin is configured
as an output (DIR7 = 1 in GPIOCR), the HD7 pin is driven to the logic level
written in the IO7 bit (1 or 0).
IO7 = 0 The HD7 input is externally driven low, or the HD7 output is
internally driven low.
4-33
(Continued)
Reset
IO7 = 1 The HD7 input is externally driven high, or the HD7 output is
Internally driven high.
4-34
(Continued)
Reset
4-35
(Continued)
Reset
4-36
(Continued)
Reset
0 IO0 X IO0 - I/O pin 0 status bit. This bit reflects the logic level on the
HD0 pin. When the HD0 pin is configured as an input (DIR0 = 0
in the GPIOCR), the IO0 bit latches the logic value of the pin (1
or 0). Writes to the IO0 bit have no effect when the HD0 pin is
configured as an input. When the HD0 pin is configured as an
output (DIR0 = 1 in GPIOCR), the HD0 pin is driven to the logic
level written in the IO0 bit (1 or 0).
4.9.1 Using the GPIO feature

To use the HPI data pins as general purpose inputs or outputs, the pins must
first be configured appropriately. Afterwards, the pins can be monitored or ma-
nipulated by reading and writing the GPIOSR. Figure 4−11 shows a code seg-
ment example for using the GPIO feature of the HPI-8.
Figure 4−11.GPIO Code Example

GPIOCR .set 3Ch ;MMR address for GPIOCR is 3Ch
GPIOSR .set 3Dh ;MMR address for GPIOSR is 3Dh
.text
STM #0F0h, GPIOCR ;Configure HD0-3 as in, and
;HD4-7 as out.
. . .
LDM GPIOSR, A ;Get GPIOSR value.
AND #0Fh, A ;Mask off MSBs.
STLM A, AR3 ;Store value of HD0-3 in AR3.
STM #050h, GPIOSR ;Set HD4-7 to 0101b.
. . .
In this example, the four LSBs of the HPI data bus (HD0−3) are configured as
general purpose inputs, while the four MSBs (HD4−7) are configured as out-
puts. The status of HD0−3 is read and stored in AR3; then the HD4−7 MSBs
are set to 0101b.
4-37
Chapter 5
Enhanced 16ĆBit Host Port Interface

(HPIĆ16)
The enhanced 16-bit host port interface, also referred to as HPI-16, is an im-
proved version of the HPI designed to interface a variety of host processors
to the ’54x.
Topic Page
5.1 HPI-16 Operational Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-2

5.2 Multiplexed Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-7
5.3 Non-Multiplexed Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-15
5.4 HPI-16 Memory Map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-18
5.5 HPI-16 and DMA Interaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-19
5.6 HPI-16 Operation During Reset . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-21
5.7 HPI-16 Operation During IDLEn . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5-21
5.8 Changes in DSP Clock Modes That Affect the HPI-16 . . . . . . . . . . . 5-22
5-1
5.1 HPI-16 Operational Overview
The HPI uses a parallel bus interface that provides a host processor with ac-
cess to internal DSP memory. The HPI is used to transfer data between the
host and DSP allowing data buffering, real-time data logging, and message
processing.
There are several variations of HPIs designed by Texas Instruments. Current-

ly, the 8-bit HPIs support 16-bit data, but interface to an 8-bit bi-directional bus.
Two consecutive byte transfers with byte identification for the most and least
significant bytes are required to make a 16-bit data word.
The HPI-16 is an enhanced 16-bit version of the ‘54x 8-bit host port interface
that, like its predecessors, functions as a slave and allows the host processor
to access internal memory without DSP CPU intervention. It provides a full
16-bit bi-directional data bus that does not require byte identification. In addi-
tion, only one host transfer is required to complete the access.The HPI-16 was
designed to interface a wide variety of host processors to a TMS320 DSP.
Key features of the HPI-16 include:
- 16-bit address bus for direct connection to the host address bus (18-bit ad-
dress bus on the ’VC5420).
- 16-bit data bus that requires no byte sequencing.
- Flexible interface that includes several strobe and control signals suitable
for a variety of 16-bit hosts.
- Both multiplexed and non-multiplexed operation for additional interfacing

flexibility.
- Memory accesses that are synchronized with the direct memory access
(DMA) controller providing access to the complete internal memory ad-
dress range.
- Software polling of the HRDY pin.
- Software control of data fetching.
- Strobe and control signals.
There are two strobe inputs and an HPI select pin available for interfacing.
These three input pins control the strobing of the HPI peripheral. The internal
strobing logic, which is referred to as internal HSTRB throughout this chapter,
functions as the actual strobe signal to the HPI peripheral.
5-2
As illustrated in Figure 5−1, the strobing logic is a function of three key strobe
inputs — HCS, HDS1 and HDS2. The HCS, also known as the HPI chip select
pin, must be low during strobe activity on the HDS pins. If HCS remains high,
activity on the HDS pins is ignored and the HRDY pin goes high. The HRDY
signal provides feedback to the host regardless of whether the current transfer
is completed (HRDY goes high). Therefore, if the host is much slower than the
HPI access and HRDY is not needed, HCS and one HDS strobe input can be
tied together. This effectively chip-selects the HPI and provides the strobe si-
multaneously.
Figure 5−1. HPI Strobe and Select Logic

HDS1
HDS2
Internal
HCS HSTRB
HRDY Internal
HRDY signal
There are several ways to interface to a host, but all interfacing schemes fall
into two general modes of operation: multiplexed and non-multiplexed. Multi-
plexed operation allows the host processor to control the HPI operation via two
control signals called HCNTL0 and HCNTL1, while non-multiplexed mode per-
forms data read/writes using only the HPI address and data buses. The HPI
mode is hardware configurable using the HMODE pin. Multiplexed mode is se-
lected when HMODE = 0.
Some devices may or may not support HMODE. You should refer to the de-
vice-specific data sheet to determine if both modes are supported. Table 5−1
contains a list of all possible HPI-16 pins and a description of their functions.
5-3
Table 5−1. HPI-16 Pin Descriptions
HPI-16 I/O/Z
Signal State† Host Connection Signal Description
HMODE I Usually static and tied high or HPI mode select pin. The logic level of HMODE
low. determines the operational mode of the HPI. A logic
level 0 selects multiplexed mode (HCNTL0/1 are
required). Logic level 1 selects non-multiplexed mode.
This is a device-specific pin. To determine if HMODE is
supported, see the device-specific data sheet.
HAS I ALE − Address latch enable or Address strobe. Used only in multiplexed mode. Host
address strobe with multiplexed address/data buses connect this pin to
their ALE. The falling edge of this input signal is used
to latch the logic levels of the HR/W, HCNTL0, and
HCNTL1 pins. When used, the HAS signal must
precede an active strobe (HCS or HDS). Hosts with
separate address/data buses must tie this signal high.
As a result, the HR/W, HCNTL0, and HCNTL1 inputs
are latched on the falling edge of HDS when HCS = 0.
HCS I Address or select line HPI chip select. Serves as an HPI select input. HCS
must be low for the HPI module to be selected. HCS
may stay low between accesses. HCS normally
precedes an active HDS strobe, but can be connected
to an HDS for simultaneous select and strobe. For an
illustration of the internal HPI strobing logic, see
Figure 5−1.
HDS1 I Read strobe and write strobe HPI data strobe pins. Used for strobing data in/out of
HDS2 or any data strobe. the HPI module. The direction depends on the logic
level of HR/W signal. The HDS signals are also used to
latch control information (if HAS is tied high) on the
falling edge. During an HPID write access, data is
latched into the HPID register on the rising edge of
HDS. During read operations, these pins act as
output-enables of the data bus.
HR/W I R/W strobe HPI read/write signal. Indicates to the HPI on the falling
edge of HAS or HDS whether the current access is to
be a read or write operation. A logic 1 indicates the
transfer is a read-from-HPI operation, while a logic 0 is
a write-to-HPI.
5-4
Table 5−1. HPI-16 Pin Descriptions (Continued)
HPI-16 I/O/Z
HCNTL0 I Address or control lines. HPI access control inputs. The logic level of these pins
HCNTL1 is latched-in on the falling edge of HAS or HDS. The
four binary states of these pins determine the access
mode of the current transfer (i.e. HPIC, HPIDinc, HPIA,
and HPID).
HA[n:0] I Host address bus HPI address bus. The n+1pins must be connected to
the host address bus if the HPI is to operate in
non-multiplexed mode. However, these pins can be
tied to a logic level if the host has a multiplexed
address/data bus and uses the HCNTL0/1 pins in
multiplexed mode. The n value may vary depending on
the DSP address range. For example, the ’VC5420
DSP has an 18-bit address range (A17:A0); therefore,
n = 17.
HD[15:0] I/O/Z Host 16-bit data bus HPI data bus. The HPI data bus is 16-bits wide and
carries the data to/from the HPI module. These pins are
Hi-Z when EMU1/OFF is active-low or when there are
no read accesses occurring.
HRDY O/Z Asynchronous ready input. HPI ready signal. Logic level 1 indicates the current
transfer is complete. Logic level 0 indicates the HPI is
not ready for the next access. The host must wait until
HRDY goes high. The logic level can be
software-polled by reading the HRDY bit in the HPIC
register. The HRDY signal is forced high when HCS
goes high. The pin is Hi-Z when EMU1/OFF is
active-low.
HINT O/Z Host interrupt pin. Host Interrupt. The DSP can interrupt the host
processor by writing a 1 to the HINT bit of the HPIC
register. Before subsequent HINT interrupts can occur,
the host must clear previous interrupts by writing a 1 to
the HINT bit of the HPIC register. This pin is driven high
when the DSP is in reset. This pin is active-low and
inverted from the HINT bit value in the HPIC register.
The pin is Hi-Z when EMU1/OFF is active-low.
HPIRS I Control pin or tied high/low. HPI reset pin. Places the HPI module in reset. No HPI
accesses can occur, HD[15:0] are placed into the
high-impedance state.
This is a device-specific pin. To determine if HPIRS is
5-5
Table 5−1. HPI-16 Pin Descriptions (Continued)
HPI-16 I/O/Z
HPIENA I Normally tied to VCC HPI enable. The HPIENA pin is used to completely
deselect the HPI by shutting the module off. Connecting
this pin to ground deactivates the HPI module and
consumes no power. Otherwise, this pin should be tied
to +VCC.
This is a device-specific pin. To determine if HPIENA is
HPI16 I Usually static and pulled high HPI16 select pin. This pin allows the HPI to support
or low 8-bit host (0) or 16-bit hosts (1). For information on the
8-bit operation, see Chapter 4.
This is a device specific pin. To determine if HPI-16 is
5-6
5.2 Multiplexed Mode
In multiplexed mode, the HPI operates with the HCNTL0/1 pins and is used
for hosts that require the use of host/DSP interrupts, or when the address and
data lines of the host are multiplexed. A host may control the HCNTL0/1 and
HR/W pins in two ways. The first case is when the host drives the HCNTL0/1
and HR/W pins with dedicated output or address pins. In this case, the
HCNTL0/1 and HR/W logic levels are latched on the falling edge of the host
driven strobe (HDS) signal. The second case is when a host has a multiplexed
address and data bus. Since there is no dedicated address bus, the host uses
the same bus to drive the control signals as it does for the data access. As a
result, the HCNTL0/1 and HR/W signals are latched using the HPI HAS signal
allowing time for the host to subsequently perform a write or read operation.
Figure 5−2 shows a typical block diagram of a host controlling the HPI in multi-
plexed mode using HAS.
Figure 5−2. Interfacing to the HPI-16 in Multiplexed Mode (’VC5420)
HOST HPI-16
HD[15:0]
Data[15:0]
Internal memory
HPID[15:0]
HCNTL0 DMA
Access
Address HCNTL1 HPIA[17:0]
Mode
Inc Mode
HPIC[15:0]
ALE HAS
R/W HR/W
Data HDS1, HDS2, HCS
strobes DSPINT
HINT ’54x
IRQ CPU
HRDY
READY
Notice that Figure 5−2 shows only one data bus and no address bus. The
HCNTL and HR/W signals are driven by the host data bus and latched using
the address latch enable pin connection to HAS. Figure 5−2 and Table 5−2
illustrate how the HCNTL0/1 signals control the HPI. Higher data throughput
can be achieved by initializing the HPIA register once, and then performing
contiguous data memory accesses using the HPIA increment mode.
5-7
Table 5−2. HCNTL0/1 Modes
HCNTL1 HCNTL0 Host Access Mode
0 0 Host has R/W capability to the HPIC register.
0 1 Host has R/W capability to the HPID register. HPIA is

automatically postincremented each time an HPI access
is performed.
1 0 Host has R/W capability to the HPIA register.
1 1 Host has R/W capability to the HPID register. The HPIA

register is not modified.
The host port interface configuration register (HPIC) is an internal configura-

tion register accessible by the HPI and DSP CPU. This register provides the
HPI and DSP with the capability to perform DSP-to-host and host-to-DSP in-
terrupts and software polling of the HRDY pin, The DSP can access this regis-
ter at location 0x2C, while the host can only access this register when in the
specific HCNTL mode (00).
Figure 5−3. HPIC Register

15−6 5 4 3 2 1 0
RSVD XHPIA FETCH HRDY HINT DSPINT RSVD
The HPIC register is accessible by the host processor and the DSP CPU.
Table 5−3 lists the bit-fields and a description of each.
Table 5−3. HPIC Bit Descriptions
Access By
Bit Host DSP Description
DSPINT R−0/W R−0 Host-to-DSP interrupt. The host can interrupt the DSP by writing a 1 to DSPINT.
This bit is always read as 0 by the host and DSP. DSP writes to this bit have no
effect.
HINT R/W−1 R/W−1 DSP-to-host interrupt. The DSP writes a 1 to the HINT bit to generate a host
interrupt.
The host interrupt HINT bit has an inverted logic level to the HINT pin. The host
must write a 1 to HINT to clear the HINT pin. Writing a 0 to the HINT bit by the host
or DSP has no effect.
HRDY R R The logic level of the HRDY pin appears in this field. The host and DSP can read
this bit for software polling of the HRDY pin. If HRDY = 0, the HPI-16 has not
completed the current data access.
5-8
Table 5−3. HPIC Bit Descriptions (Continued)
Access By
Bit Host DSP Description
FETCH R−0/W R−0 Host data fetch request. When a host writes a 1 to this bit, data located at the
current HPIA address is fetched and loaded into the HPID register. This bit is
always read as 0 by the host and DSP.
XHPIA R/W R Extended address enable. When XHPIA=1, host writes to the HPIA register are
loaded into the most significant bits, HPIA[n:16]. If XHPIA=0, host writes to HPIA
are loaded into HPIA[15:0]. All n+1 address bits are incremented in the
autoincrement mode. Reading the HPIA register is performed in the same
manner. Only the host has access to this bit.
The HPIA register contains the address where the next data access occurs.
The HPIA register is automatically incremented in the HCNTL0/1 = 10b mode.
The FETCH = 1 bit issues a request to read data pointed to by the HPIA regis-
ter. The FETCH bit can be used with the HRDY bit to perform software polling
to determine when the internal access is complete. The FETCH and HRDY
pins are useful when an update to the HPID register is desired. The FETCH
bit is always read as 0. Performing a FETCH does not increment the HPIA reg-
ister.
Interrupts between the host and DSP are possible by setting and polling
DSPINT and HINT bits in the HPIC register. The DSPINT bit, when set to 1 by
the host, posts an interrupt to the DSP CPU. The DSP can process this inter-
rupt if desired by enabling the interrupt in the CPU’s IMR register. The DSP can
post a host interrupt by writing a one to the HINT bit. In this case, the host must
also perform a write to the HINT bit to acknowledge the previous HINT request.
The DSPINT bit is always read as zero. Writing zeros to the HINT and DSPINT
bits has no effect.
The XHPIA bit logic level determines which address bits in the HPIA are initial-
ized. Because the data bus is 16 bits and the address reach is greater than 16
bits (18 bits on the ’VC5420), the XHPIA bit loads the lower 16 bits of the HPIA
when set to zero, and the extended address bits (>16) are loaded when
XHPIA = 1. Initialization of all address bits is recommended before performing
data accesses. XPIA must be changed in the same manner to read the con-
tents on the HPIA register.
5-9
5.2.1 Host Accesses With HAS
The host address strobe (HAS) is only valid in multiplexed mode. The HAS sig-
nal enables glueless interfacing for host processors with multiplexed address
and data buses. The falling edge of HAS is used to latch the HCNTL0/1 and
HR/W states into the HPI. First, the host latches the transfer mode in the HPI
by driving the HCNTL0/1 and HR/W pins to the desired access mode. HAS is
not gated by HCS; and therefore, allows time for the host to perform the subse-
quent access. The HAS signal may be brought high after the HDS is driven low,
indicating the data access is about to occur. HAS is not required to be high,
but must eventually transition high when there is a change in access type.
Figure 5−4 illustrates a write-access using HAS.
Figure 5−4. HPIA Write Using HAS
HCS
Internal
HSTRB
HAS
HR/W
HCNTL[1:0] 10b
HD[15:0] HPIA
HRDY
Latch access Latch HPIA
type using HAS address
5-10
5.2.2 Host Accesses Without HAS
In cases where the host processor has dedicated signals (or bit I/O) capable
of driving the HCNTL0/1 pins, the HAS signal is not required. These dedicated
signals can be directly connected to the HCNTL0/1 and HR/W pins. This sim-
plifies the access considerably because there is no setup and strobing of the
HCNTL0/1 data relative to HAS. For example, an HPIC read without using
HAS is performed by driving the HCNTL0/1 value (00) and then strobing the
HPI. After the falling edge of the internal HSTRB signal, the HPI drives the
HPIC register data onto the data bus. Figure 5−5 shows the read-access se-
quence without using the HAS signal.
Figure 5−5. HPIC Read Without Using HAS
HCS
Internal
HSTRB
HAS
HR/W
HCNTL[1:0] 00
HD[15:0] HPIC
HRDY
Latch access type using
internal HSTRB
5-11
5.2.3 Autoincrement Operation
All of the HPI peripherals include a feature designed to increase the data
throughput of the HPI. This feature, called autoincrement, is used to pre-fetch
data and point to the next higher data location (post-modified) after the current
access is complete. This feature automatically modifies the HPIA register so
that the host can perform consecutive HPI read/write transfers to a contiguous
memory space without having to modify the HPIA register contents. This fea-
ture is enabled when the HCNTL0/1 mode = 01b.
Figure 5−6. HPID Read Using Autoincrement
HCS
Internal
HSTRB
HR/W
HCNTL[1:0] 01 01
HD[15:0] Data 1 Data 2
HRDY
DMA
active
HPIA
register HPIA HPIA + 1 HPIA + 2
HPID
register Valid 1 Valid 2
Initial Pre-fetch Read Subsequent

data read data pre-fetch pre-fetch
data
5-12
Figure 5−7. HPID Write Using Autoincrement
HCS
Internal
HSTRB
HR/W
ÁÁ
HCNTL[1:0]
ÁÁ 01 01
HD[15:0] Data 1 Data 2
HRDY
ÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁ
DMA
ÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁÁ
active
HPIA
register HPIA HPIA + 1
HPID
register Valid 1 Valid 2
Write data at HPIA Write data at

HPIA + 1
5-13
Due to the nature of pre-fetching, post-modified reads can cause invalid (old)
data to be read by the host CPU. This occurs when the DSP has updated the
next higher data location after the host has performed a read access. As a re-
sult of the post-modify and pre-fetch mechanism, the next data read by the
host will not have the updated data value. If both DSP and HOST are writing
to the same data locations, it is suggested that FETCH be performed before
reading the data. FETCH is performed by setting the FETCH bit in the HPIC
register. Subsequently, the HPIC register can be read to poll the HRDY bit be-
fore reading the data from the HPI. Figure 5−6 and Figure 5−7 show read and
write operations with the autoincrement feature. For the initial read (with auto-
increment) and for all reads without the autoincrement feature, the HRDY sig-
nal goes not-ready when the access begins (falling edge of internal HSTRB).
Subsequent reads in increment mode perform data pre-fetching, which in-
creases the HPI throughput.
Write operations are always performed on the rising edge of the internal
HSTRB signal. The increased throughput on write operations is realized when
making consecutive writes and using the not-ready state of HRDY to initiate
the next write operation. Since the host must transition and hold HSTRB high
for at least 10ns before making the next write, the hold time can be satisfied
while HRDY is not-ready. This is illustrated in Figure 5−7.
5-14
5.3 Non-Multiplexed Mode
In non-multiplexed mode, hosts with separate address and data buses can ac-
cess the HPI data (HPID) register by using the 16-bit bi-directional data bus
and the HPIA address register via the n-bit address bus. There is no HPIC reg-
ister in the non-multiplexed mode; therefore, no interrupts (DSPINT and HINT)
can be passed between the host and DSP. Due to the fact that a host address
bus is available to drive the HPIA, there is no need for the autoincrement fea-
ture, which precludes the necessity of the FETCH and XHPIA bits. Software
polling of the HRDY is not available, but the HRDY pin is fully functional and
capable of holding off a subsequent host access before the access is com-
pleted internally to the DSP. The HRDY pin functions as previously stated in
the multiplexed mode. Figure 5−8 illustrates how to interface to the HPI in non-
multiplexed mode.
Figure 5−8. Interfacing to the HPI-16 in Non-Multiplexed Mode (’VC5420)
ÎÎ
ÎÎ
ÎÎ
ÎÎ
HOST HPI-16
ÎÎ
HD[15:0]
memory
Data[15:0]
Internal
HPID[15:0]
ÎÎ
ÎÎ
DMA
HA[ n :0]
Address[n:0]
VCC
x
ÎÎ
ÎÎ
HCNTL0
x HCNTL1
HAS
R/W HR/W ’54X
Data strobes HDS1, HDS2, HCS CPU
HRDY
Ready
The HPI select pin (HCS) and strobe signals (HDS1 and HDS2) function as
stated in the multiplexed mode. The HAS, HCNTL0, and HCNTL1 pins are not
included because they have no function in the non-multiplexed mode.
A typical access in non-multiplexed mode is shown in Figure 5−9 and

Figure 5−10. Data throughput is the same as in the multiplexed mode when
not using the HAS signal or autoincrement feature. Bus arbitration with the
DMA controller is handled in the same way.
5-15
Figure 5−9. HPID Read in Non-Multiplexed Mode
HCS
Internal
HSTRB
HR/W
HA[n:0] n+1-bit address
Á Á
HD[15:0]
Á HPID contents
Á
HRDY
DMA
active
HPI HPI
not ready ready
5-16
Figure 5−10. HPID Write in Non-Multiplexed Mode
HCS
Internal
HSTRB
HR/W
HA[n:0] n+1-bit address
HD[15:0] HPID contents
HRDY
DMA
active
HPI
ready
5-17
5.4 HPI-16 Memory Map
Depending on the logic level of the HMODE pin, the HPI-16 peripheral can per-
form accesses to DSP memory (internal memory only on the ’VC5420 ) in mul-
tiplexed or non-multiplexed modes. As previously noted, multiplexed mode
uses the HPIA register for access to the memory map. Non-multiplexed mode
uses the n-bit HPI address bus to directly access the memory map. Both of
these modes are capable of accessing all DSP memory locations that are ac-
cessible via the DMA controller. Figure 5−11 shows the HPI memory map for
the ’VC5420, a device having an 18-bit address bus.
Figure 5−11.’VC5420 Memory Map Relative to the HPI

Program Page 0 Program Page 1 Program Page 2 Program Page 3
Hex Hex Hex Hex
0000 10000 20000 30000
Reserved
001F
Reserved Reserved Reserved
0020 McBSPDXR/DRR
005F MMRegs only 005F 005F 005F
0060 On-chip 0060 On-chip 0060 On-chip 0060 On-chip
DARAM 0 DARAM 0 DARAM 0 DARAM 0
(overlayed) (overlayed) (overlayed) (overlayed)
3FFF prog/data 13FFF prog/data 23FFF prog/data 33FFF prog/data
4000 14000 24000 34000
On-chip On-chip On-chip On-chip
SARAM 1 DARAM 1 SARAM 1 SARAM 1
(overlayed) (overlayed) (overlayed) (overlayed)
prog/data prog/data prog/data prog/data
7FFF 17FFF 27FFF 37FFF
8000 18000 28000 38000
Reserved
2EFFF
On-chip On-chip 2F000
SARAM 2 SARAM 3 On-chip Reserved
(32K words) (32K words) SARAM 4
prog/data program (32K words)
program
FFFF 1FFFF 2FFFF 3FFFF
Note: I/O Space is not accessible.

All internal memory is divided into 8K blocks with the exception of the 4K word block on P2 (0x2F000−0x2FFFF)
DROM and OVLY bits do not affect the memory map.
The HPI peripheral has no knowledge of any memory configuration bits con-
trolled by the CPU. For this reason, the values of the OVLY and DROM bits do
not affect the data location relative to the HPI. In the case of the ’VC5420, ac-
cesses to the internal DSP memory are always considered overlayed program
and data. There is no way to distinguish between program and data spaces.
I/O space is not accessible by the HPI.
You should refer to the device-specific data sheet for information on other HPI
memory maps and configurations.
5-18
5.5 HPI-16 and DMA Interaction
The HPI-16 operates with the DMA controller granting the host access to any
memory location (internal memory only on the ’VC5420). Each time the host
CPU request an access, the DMA finishes any current transfers and releases
the DMA bus to the HPI module. As a result, there are latencies associated
with HPI accesses due to other DMA activity. The host CPU can access the
DSP memory (writes) at the maximum rate of once every six DSP cycles, if no
other DMA channels are active. However, HPI throughput decreases to one
transfer every 10 cycles when there are two or more active DMA channels. The
HPI has the highest priority, including the CPU. As a result, the HPI is granted
access even if the CPU or other DMA channels are competing for the same
memory location. Figure 5−12 illustrates the HPI access sequence relative to
the DMA controller.
Figure 5−12. HPI and DMA Interaction
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26
DSP
CLK
DMACH0 W1 B R2 W2 B R3 W3
DMACH3 R1 W1 B R2 W2 B R2 W2
HPI-16 S B XFER D S B XFER D S B

1 2 3 4 5 6 7 8 9 10
Note: S = Sync to DSP clock

B = Bid for DMA bus
D = Delay for propagation
Rn = Read cycle on the nth access
Wn = Write cycle of the nth access
The HPI transfer latency increases if one or more DMA channels are active;
however, the HPI latency never increases beyond the latency of one active
DMA channel. The HPI bus grants are interlaced between the revolving DMA
channels. For this reason, data throughput on other DMA channels is sacri-
ficed if high HPI data rates are required. It should be noted that even without
HPI intervention, it is possible to lock out low priority DMA channels if all high
priority DMA channels are continuously serviced. You should refer to the de-
vice-specific data sheet for HPI read and write transfer latencies for the various
’54x devices.
5-19
The HPIA register is not susceptible to latencies due to DMA controller arbitra-
tion. Accesses to the HPIA register do not use the DMA bus because this regis-
ter is part of the HPI peripheral. HPIC register accesses have no latencies
when writing values that have the DSPINT and HINT bits set to zero. HRDY
is never driven low for these accesses because subsequent accesses can be-
gin immediately. However, writing a one to the DSPINT or HINT bits introduces
a latency due to the required access to the DSP CPU module. Consequently,
the HRDY signal is held low for three DSP clock cycles.
5-20
HPI-16 Operation During Reset / HPI-16 Operation During IDLEn
5.6 HPI-16 Operation During Reset

The HPI-16 peripheral can operate while the DSP CPU is in reset. On the
’VC5420 device, the RESET signals are separated between core CPUs and
the HPI. The HPI is placed in reset when the dedicated HPI reset pin, HPIRST,
is equal to zero. When the HPIRST signal is one, the HPI has full access to the
internal memory space of the DSP. Accesses still require six DSP clock cycles,
allowing the maximum data throughput of 33MB/s @ 100MIPS.
The host processor can place the DSP in reset by controlling the RS pin. In this
case, the host must complete any current memory access and wait six DSP
clock cycles before driving the RS pin low. During reset, the internal memory
is available to the host. To release the DSP from reset, the host can drive the
RS pin high, but is required to wait at least 20 DSP clock cycles before making
subsequent memory accesses.
An additional method for releasing the DSP from reset is to leave each core’s
RS pin high and cycle HPIRST low, then high. When this occurs, the host can
write to location 0x2F (any value), which will internally reset that particular
core.
5.7 HPI-16 Operation During IDLEn

The HPI-16 can continue to operate during the IDLE 1 and IDLE 2 states by
using special clock management circuitry. This circuitry turns on relevant
clocks to perform synchronous memory accesses and then turns these clocks
off to conserve power. The host processor can wake the DSP from the IDLE
states when in multiplexed mode via the DSPINT bit in the HPIC register. In
non-multiplexed mode, external interrupts (ex. INT0, INT1) can be used to
wake the DSP.
The host processor can wake the DSP from IDLE 3 by writing a one to the
DSPINT bit of the HPIC register. Memory accesses during IDLE 3 are not sup-
ported.
5-21
5.8 Changes in DSP Clock Modes That Affect the HPI-16
As stated earlier, the HPI is synchronized to the DSP clock. The HPI can ac-
cess internal DSP memory every six DSP clock cycles. Therefore, changes to
the DSP clock mode affect the time required to complete memory accesses.
If the DSP clock is slowed, the HPI access rate is slowed proportionally.
For example, if the clock mode is changed from a PLL mode to a DIV mode,
the host can use the HRDY pin to effectively insert host wait-states. In
Figure 5−13, the relative timing for a change from multiply-by-1 to DIV mode
is shown. Notice that by using the HRDY pin, wait-states are effectively in-
serted to hold off the host until the current access is complete.
Figure 5−13. HPI-16 Operation During the PLL to DIV Clock Mode Change
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
DSP
CLK
HPI-16 S B R1 W1 S B R2 W2 S B R3 W3
HRDY
PLLx 1 mode DIV2 mode
Note: S = Sync to DSP clock

B = Bid for DMA bus
Rn = Read cycle on the nth access
Wn = Write cycle of the nth access
The effective HPI transfer rate is increased by changing the DSP clock to a
PLLxn multiply mode. In this case, fewer host wait-states are required. In fact,
as the multiplier mode is increased, it is possible that the HPI transfer rate will
exceed the latency of the host access rate. When this occurs, no host wait-
state is required.
5-22
Chapter 6
Interprocessor Communications
This chapter describes interprocessor communications associated with

multi-core DSPs, and includes core-to-core FIFO communications and exter-
nal memory interface-to-host port interface (EMIF-to-HPI) communications.
Topic Page
6.1 Communication Within Multi-Core DSPs . . . . . . . . . . . . . . . . . . . . . . . . 6-2

6.2 The Bi-Directional FIFO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-3
6.3 Accessing the HPI-16 From External Memory Space . . . . . . . . . . . . . 6-7
6.4 Subsystem Communication Using McBSP . . . . . . . . . . . . . . . . . . . . . 6-10
6.5 Interprocessor Interrupts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6-11
6-1
6.1 Communication Within Multi-Core DSPs
Multi-core (multi-CPU), multi-subsystem DSPs were introduced by Texas
Instruments in 1999. These devices are capable of doubling performance
either by executing duplicate application code on multiple cores or by running
one application that divides tasks between the cores. The subsystems, each
containing its own core, are independent of each other.
Dividing tasks between multiple cores requires data transfers and message
processing between the subsystems. The methods for implementing interpro-
cessor communications include:
- Use of FIFOs
- Host port interfaces (HPIs)
- Multichannel buffered serial ports (McBSPs)
- Mutual memory spaces
This chapter explains the implementation of these task management methods

using the ’VC5420 as an example. For device specific information, you should
refer to the appropriate data sheet.
Figure 6−1. ’VC5420 — The 2-Subsystem DSP
ÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎÎ
Core Core
Sub- Sub-
system system
A B
ÏÏÏÏÏÏÏÏÏÏÏÏÏÏÏÏÏÏ
6-2
6.2 The Bi-Directional FIFO
Communications between subsystems can be accomplished using the first-in,
first-out (FIFO) method described in this section. A second method is dis-
cussed in section 6.4 on page 6-10.
The FIFO peripheral consists of two uni-directional, 8-element-deep channels.
Each channel supports 16-bit data (element) transfers and is dedicated to
either transmit or receive operation. In the case of the ’VC5420, the FIFO is
shared by both subsystems. One FIFO channel is dedicated for transmission
of data from A to B, and the other channel is dedicated for data transmission
from B to A. The subsystem DMA controller that is transmitting data can write
eight words to the FIFO before the FIFO becomes full; however, if the receiving
subsystem reads from its receive channel simultaneously, the transmitting
subsystem can continue indefinitely.
Figure 6−2. ’VC5420 FIFO Configuration

Transmit FIFO not full Receive FIFO not empty
8-Element FIFO Subsystem B

Subsystem A
DMA Controller DMA Controller
Word 7
Word 6
Word 5
Word 4
Word 3
Word 2
Word 1
Word 0
DMA channel (write to FIFO) DMA channel (read from FIFO)

DMA channel DMA channel
DMA channel Shared DMA I/O space DMA channel
Word 0
Word 1
Word 2
Word 3
Word 4
Word 5
Word 6
Word 7
DMA channel (read from FIFO) DMA channel (write to FIFO)

Receive FIFO not empty Transmit FIFO not full
Subsystem A Subsystem B
internal memory internal memory
space space
Special internal handshaking is provided for handling overrun and underrun

conditions of the 8-element FIFO buffer; and for this reason, it is impossible
to lose data by either condition. This implies that data throughput is dependent
upon both transmit and receive subsystem activity.
On the ’VC5420, handshaking is performed between the FIFO and DMA pe-
ripherals, and no intervention by the CPU is required once the DMA receive
6-3
and transmit channels are configured and enabled. Example 6−1 shows DMA
channel information that is used to configure DMA channels 0 and 1 for FIFO
transmit and receive, respectively.
Example 6−1. DMA Channels 0 and 1 Configured for FIFO Transmit and Receive
Transmit − DMA channel 0
DMSRC0 = #200h −source address
DMDST0 = #xxxh −this is a don’t care
DMCTR0 = #0Fh −element count (buffer size)
DMSFC0 = #8000h −DMA sync mode
DMMCR0 = #0142h −DMA increment mode,
memory space information.
Receive − DMA channel 1

DMSRC1 = #xxxh −this is a don’t care
DMDST1 = #400h −destination address
DMCTR1 = #0Fh −element count (buffer size)
DMSFC1 = #7000h −DMA sync mode
DMMCR1 = #085h −DMA increment mode,
memory space information.
Table 6−1. DMA Configuration to Support FIFO Transfers
Transmit Channel (DMA Channel 0) Receive Channel (DMA Channel 1)

Source Destination Source Destination
Location of source and internal FIFO Location of source and FIFO internal
destination memory destination memory
Memory space data/prog I/O Memory space I/O data/prog
Memory address 200h don’t care Memory address don’t care 400h
DMA index mode (buffer post- no DMA index mode (buffer no post-
pointer increment mode) increment increment pointer increment mode) increment increment
Internal DMA sync event transmit FIFO Internal DMA sync event receive FIFO
not full not empty
Element count 0Fh Element count 0Fh

(buffer size) (buffer size)
Interrupt selection DMA CH0 = Interrupt selection DMA CH1 =

IMR/IFR bit 6 IMR/IFR bit 7
6-4
Figure 6−3 illustrates the physical setup for the DMA configuration shown in
Table 6−1. Transfer buffers (located in either internal data or program space)
must be initialized so that the DMA can process the data stream without CPU
intervention. The data buffers have a total of 32 words (a 16-word receive buff-
er and a 16-word transmit buffer) and are located in internal memory at 200h
and 400h.
Figure 6−3. DMA Configuration for FIFO Operation
Subsystem A Subsystem B
DMA controller DMA controller
0x200 DMAC0 FIFO XMT DMAC0 FIFO XMT 0x200

FIFO
0x400 DMAC1 FIFO RCV DMAC1 FIFO RCV 0x400
DMA channel 2 DMA channel 2
The FIFO is located in a shared DMA I/O space; that is, both subsystem’s DMA
share a common I/O space. As a result, the DMA memory space selection
must be configured as I/O space to access the FIFO. The memory space
selection is made in the DMMCR register.
The source/destination address that points to the FIFO is a don’t care, since
all writes and reads are from one FIFO memory location. Consequently, the
indexing mode for the FIFO side of the DMA should be configured for no-incre-
ment. The buffer size, also known as the element count in the DMCTR register,
is always programmed as buffer size-1. The element count is used for both
source and destination buffers.
Interrupts for DMA channels 0 and 1 are mapped to IMR bits 6 and 7. However,
in Example 6−1 and Table 6−1, no interrupts are programmed or enabled. To
enable interrupts, the DMA channels must be mapped to the IMR bits as
shown. In addition, the DMA must be programmed to generate interrupts at a
specific time while filling the FIFO.
6-5
Interrupts can be generated when the FIFO buffers become 50% or 100% full.
These interrupts are turned on in the DMA’s DMMCR register and enabled in
the CPU’s interrupt mask register (IMR). If an interrupt is generated on a half-
full buffer, the interrupt is generated when the receipt of the fourth word is com-
pleted. A full-buffer interrupt occurs after receipt of the eighth word is com-
pleted.
The latency to write a word to the FIFO is four CPU cycles. If the receive sub-
system immediately reads the word, it requires four additional CPU cycles to
read the data from the FIFO. Therefore, the maximum throughput is one 16-bit
word every eight CPU cycles. Additional latency can be incurred when the
DMA controller is processing other DMA channels or HPI transfers.
6-6
6.3 Accessing the HPI-16 From External Memory Space
The standard HPI, HPI-8, and HPI-16 peripherals are all designed to connect
to another DSP’s external memory space. Since the HPI is always a slave de-
vice, the master DSP must perform writes and reads to the external memory
space to strobe the HPI peripheral. On multi-core DSPs, this connection is
made internally. Figure 6−4 illustrates the internal connection between the ex-
ternal memory interface and the HPI-16 peripheral on the ’VC5420.
Figure 6−4. HPI-16 and External Memory Interface Connection (’VC5420)
Subsystem
Master DMA controller
Slave
Subsystem DMA channel 0 Subsystem
DMA channel 1
DMA channel 2
Master DMA channel 3

16−bit data/program bus
subsystem Subsystem
CPU DMA channel 4
16−bit DMA bus

internal
DMA channel 5 memory
space
External memory HPI-16 peripheral

interface
PPA[17:0] HA[17:0]
Slave
PPD[15:0] HD[15:0]
subsystem
R/W HR/W CPU
MSTRB HCS
PS HDS1
DS HDS2
Ready HRDY
The HPI-16 is a 16-bit device capable of performing data accesses from inter-
nal memory space. Due to the 18-bit address bus, the HPI-16 can address the
full address range. The HPI-16 operates in non-multiplexed mode and is acti-
vated by strobing the HDS lines. The master device must perform an access
to the external memory. The external program/data memory interface signals
are mapped to the HPI-16 control pins.
6-7
The subsystems of the multi-core DSP share the HPI-16 peripheral and exter-
nal memory interface. Because of this, external arbitration logic is required if
direction of the communication (determined by the SELA/B pin) must be
changed. The HPI-16 must operate in the non-multiplexed mode requiring that
the HMODE pin be pulled high. The XIO pin must also be pulled high to enter
a special state where the EMIF pins are connected internally to the HPI-16
pins. Table 6−2 summarizes the operation for each pin configuration.
Table 6−2. EMIF/HPI-16 Modes
HMODE SELA/B EMIF Modes (XIO = 1)

1 0 EMIF-to-HPI master is subsystem A
EMIF-to-HPI slave is subsystem B
1 1 EMIF-to-HPI master is subsystem B

EMIF-to-HPI slave is subsystem A
An additional requirement when changing the direction of communication is

the value of the MP/MC bit in the PMST register. The handshaking between
subsystems must not only request/acknowledge a state change of the SELA/B
pin, but also change the value of the MP/MC bit in both subsystems. The only
time when initialization is not required is after reset. Table 6−3 lists the MP/MC
bit levels at reset as a function of the XIO, HMODE, and SELA/B pins.
Table 6−3. MP/MC Bit Levels at Reset
‘5420 Pins MP/MC Bit

XIO HMODE SELA/B Subsystem A Subsystem B
0 X X 0 0
1 0 X 1 1
1 1 0 1 0
1 1 1 0 1
After reset, it is up to the handshaking logic and subsystems to correctly config-

ure the SELA/B pin and MP/MC bit levels. The MP/MC value of subsystem A
must always be set to the inverted logic level of the SELA/B pin. Furthermore,
subsystem B must always have the MP/MC bit set to same logic level as the
SELA/B pin.
6-8
Due to the state of the XIO and HMODE pins, the master device must execute
from external memory. The slave subsystem is configured to operate in HPI
mode; therefore, all internal memory is enabled. As a result, the external
memory location that the master subsystem accesses is the location that is ac-
cessed on the slave subsystem.
The maximum throughput of the EMIF/HPI-16 channel is one transfer every

six DSP clock cycles, assuming there is no additional DMA activity. Transfer
rates can decrease to one word every 14 DSP clock cycles if there is additional
DMA activity. However, the master device does not need to program wait-
states because the HPI ready pin (HRDY) is connected to the DSP CPU
READY pin.
6-9
6.4 Subsystem Communication Using McBSP
Even though the ’VC5420 multi-core DSP does not support McBSP-to-McBSP
communication by design, simply connecting a McBSP from each subsystem
externally provides a very robust interprocessor communication mechanism.
Figure 6−5 illustrates this external connection. There is only a 6-line connec-
tion required to implement the bi-directional/full-duplex communication chan-
nel. No master/slave configuration is required.
Figure 6−5. McBSP-to-McBSP Connection
McBSP
peripheral
CLKX
FSX
DX
DR
FSR
CLKR
McBSP
peripheral
CLKX
FSX
DX
DR
FSR
CLKR
The transmit section of the McBSPs must provide a shift-clock and frame-sync
to the receive section of the other subsystem’s McBSP. Therefore, the McBSP
CLKX and FSX pins are configured as output signals. The CLKR and FSR pins
are configured as input signals.
The McBSP can be programmed to support many different data transfer
modes. The subsystem’s DMA controller can be used for autobuffering trans-
mit and receive data buffers. In addition, the McBSP peripheral can be pro-
grammed to generate CPU interrupts to indicate when certain conditions exist.
When used for interprocessor communications, the McBSP serial port is one
of the simplest forms of subsystem communication. The McBSP-to-McBSP
connection is not supported internally — it is the responsibility of the system
designer to implement the connections at the system board level.
6-10
6.5 Interprocessor Interrupts
If DMA, HPI, and McBSP interrupts do not provide enough flexibility, an addi-
tional interrupt called the interprocessor Interrupt is provided. This interrupt is
generated internally to notify the other subsystem of a certain event. To gener-
ate the interrupt, the CPU must write a one to the interprocessor interrupt re-
quest bit (IPIRQ) in the BSCR register. Each subsystem’s BSCR register con-
tains the IPIRQ bit. Furthermore, each subsystem can respond to the respec-
tive interrupt named IPINT.
The interrupting subsystem must write a one to the IPIRQ bit followed by a
zero. Consequently, the interrupted subsystem’s IFR register (IPINT bit) is up-
dated and the CPU processes the IPINT interrupt, if enabled in the IMR. The
interrupting subsystem may leave the IPIRQ bit set high indefinitely. This does
not generate multiple interrupts on the interrupted subsystem. The IFR flag
representing the IPINT interrupt is cleared automatically when the CPU pro-
cesses the associated interrupt service routine. The IPIRQ bit must be set to
zero before subsequent interrupts can be generated.
The interprocessor interrupt mechanism is not supported on all multi-core

DSPs. You should refer to the device-specific data sheet to determine if this
feature is supported on a particular device.
6-11
6-12
Chapter 7
Enhanced External Bus Interface

(EnhXIO)
This chapter describes the improved functionality of the enhanced external

parallel interface (XIO2) available on the ’5410 device.
Topic Page
7.1 Introduction to the Enhanced External Bus Interface (XIO2) . . . . . . 7-2

7.2 Bus Sequences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7-3
7-1
7.1 Introduction to the Enhanced External Parallel Interface (XIO2)
The enhanced external parallel interface is available only on select ’54x de-
vices. Improvements to the ’VC5410’s external interface (XIO2) include:
- Simplification of bus sequences
- More immunity to bus contention when transitioning between read and

write operations
- External memory access to the DMA controller
- Optimization of the power-down modes
The bus sequence on the ’VC5410 still maintains all of the same interface sig-
nals as on previous ’54x devices, but the signal sequence has been simplified.
Most external accesses now require three cycles composed of a leading cycle,
an active (read or write) cycle, and a trailing cycle.
The leading and trailing cycles provide additional immunity against bus con-
tention when switching between read and write operations. To maintain high-
speed read access, a consecutive read mode that performs single-cycle
reads, as on previous ’54x devices, is available.
7.1.1 Additional Features

The XIO2 also provides the ability for DMA transfers to extend to external
memory. For more information on DMA capability, see Chapter 3.
The XIO2 improves low-power performance, already present on ’54x devices,
by switching off the internal clocks to the interface when it is not being used.
This power-saving feature is automatic, requires no software setup, and
causes no latency in the operation of the interface.
Additional features integrated in the XIO2 include:
- The ability to automatically insert bank-switching cycles when crossing
32K memory boundaries
- The ability to program up to 14 wait states through software
- The ability to divide down CLKOUT by a factor of 1, 2, 3, or 4. Dividing

down CLKOUT provides an alternative to wait states when interfacing to
slower external memory or peripheral devices.
Although inserting wait states extends the bus sequence during read or write
accesses, it does not slow down the bus signal sequences at the beginning
and end of the access. Dividing down CLKOUT provides a method of slowing
the entire bus sequence when necessary. The CLKOUT divide-down factor is
controlled through the DIVFCT field in the bank-switching control register
(BSCR).
7-2
7.2 Bus Sequences
Figure 7−1 shows the bus sequence for three cases: all I/O reads, memory
reads in non-consecutive mode, and single memory reads in consecutive
mode. The accesses shown always require three CLKOUT cycles to com-
plete.
Figure 7−1. Non-Consecutive Memory Read and I/O Read Bus Sequence
CLKOUT
A[22:0]
D[15:0] Read
R/W
MSTRB or IOSTRB
PS/DS/IS
Leading
cycle
Read
cycle
Trailing
cycle
7-3
Figure 7−2 shows the bus sequence for repeated memory reads in con-
secutive mode. The accesses shown require (2+n) CLKOUT cycles to com-
plete, where n is the number of consecutive reads performed.
Figure 7−2. Consecutive Memory Read Bus Sequence (n = 3 reads)
CLKOUT
A[22:0]
D[15:0] Read Read Read
R/W
MSTRB
PS/DS
Leading
cycle
Read
cycle
Read
cycle
Read
cycle
Trailing
cycle
7-4
Figure 7−3 shows the bus sequence for all memory and I/O writes. The ac-
cesses shown always require three CLKOUT cycles to complete.
Figure 7−3. Memory Write and I/O Write Bus Sequence
CLKOUT
A[22:0]
D[15:0] Write
R/W
MSTRB or IOSTRB
PS/DS/IS
Leading
Cycle
Write
Cycle
Trailing
Cycle
7-5
7-6
Appendix
AppendixAA
Glossary
A
A-bis mode: In the A-bis mode (ABIS = 1), the McBSP can receive and
transmit up to 1024 bits on a PCM link.
AC97: Audio Codec ’97. A standard which uses a dual-phase frame feature
with the first phase consisting of a single 16-bit word, and the second
phase consisting of twelve 20-bit words.
Access mode: HPI access mode is determined by the logic levels of the
HCNTL0/1 pins. Access mode can be an HPIC access, HPIA access, or
HPID access with autoincrement.
ALE: address latch enable. Hosts that use a multiplexed address/data bus
may use this to latch address/control data. Usually connected to HAS.
AUTOINIT: DMA auto-initialization bit. This bit configures the channel to au-
tomatically initialize after a frame by loading from the global DMA regis-
ters. Located in the DMMCR register.
B
Bit Ordering: A feature that allows the LSB to be transferred to the serial
port first when companded data is not used.
BOB: HPI byte-order bit. Selects which byte is transferred first. Only on
HPI-8 HPI versions. Located in the HPIC register.
A-1
C
CLKG: programmable data clocks. Internal McBSP signals that can be pro-
grammed to drive receive and/or transmit clocking and framing.
CLKGDV: A programmable value that can be used to divide down the input
clock source to the sample rate generator.
CLKR: receive clock. A McBSP interface signal.
CLKRM: receive clock mode. Located in the PCR register.
CLKRP: receive frame clock polarity bit. Located in the PCR.
CLK(R/X): data clocks(receive/transmit).
CLKS: external clock input. A McBSP interface signal.
CLKSM: clock source mode. The CLKSM bit in the SRGR2 selects either the
CPU clock (CLKSM = 1) or the external clock input (CLKSM = 0) CLKS,
as the source for the sample rate generator input clock.
CLKSP: clock polarity. The rising edge of CLKS when CLKSP = 0, or the fal-
ling edge of CLKS when CLKSP = 1, causes the transition on the data
bit-rate clock (CLKG) and frame sync (FSG).
CLKSRG: clock sample rate generator. The rising edge of CLKSRG gener-
ates clocking (CLKG) and framing (FSG).
CLKS_STAT: CLKS pin status bit. Indicates the logic level of the CLKS pin
when CLKS is configured as a general-purpose input.
CLKSTP: clock stop mode bit. This bit is located in the SPCR1 register and
is used to stop the serial port clock.
CLKX: transmit clock. A McBSP interface signal.
CLKXM: transmit clock mode. Located in the PCR register.
CLKXP: transmit clock polarity bit. Located in the PCR.
Companding: compressing and expanding. A quantization scheme for au-

dio signals in which the input signal is compressed, and after processing,
is reconstructed at the output by expansion. There are two distinct com-
panding schemes: A-law, used in Europe, and µ-law, used in the United
States.
CTMOD: DMA transfer counter mode control bit. Selects multiframe or ABU
mode. Located in the DMMCR register.
A-2
D
DATDLY: data delay (R/X). A delay at the beginning of actual data reception
or transmission with respect to the start of the frame. The range of pro-
grammable data delay is 0 to 2 bit-clocks.
DBLW: double-word mode. Double-word mode is used when DMA data is

32-bits wide. Located in the DMSFC register.
DE: DMA channel enable bit. Enables/disables the DMA channel operation.
Located in the DMPREC register.
DIND: DMA destination address transfer index mode bit. Offers several dif-
ferent indexing modes. Located in the DMMCR register.
DINM: DMA interrupt generation mask bit. Enables/disables DMA transfer

interrupts. Located in the DMMCR register.
DIR: direction bit. Selects the direction of a general-purpose I/O pin. Located
in the GPIOCR register.
Direct Memory Access (DMA) Controller: The direct memory access

(DMA) controller transfers data between regions in the memory map
without intervention by the CPU. The DMA allows movement to and from
internal memory, internal peripherals, or external devices to occur in the
background of CPU operation.
DLB: digital loop-back mode. DLB mode allows testing of serial port code
with a single DSP device by internally connecting DR, FSR, and CLKR
through multiplexers to DX, FSX, and CLKX.
DMA: direct memory access (controller).
DMCTR: DMA channel element count register.
DMD: DMA destination address space select bit. Located in the DMMCR
register.
DMDST: DMA channel destination address register.
DMDSTP: DMA destination program page address (all channels).
DMFRI0: DMA frame address index register 0.
DMFRI1: DMA frame address index register 1.
DMGCR: DMA global element count reload register.
A-3
DMGDA: DMA global destination address reload register.
DMGFR: DMA global frame count reload register.
DMGSA: DMA global source address reload register.
DMIDX0: DMA element address index register 0.
DMIDX1: DMA element address index register 1.
DMMCR: DMA channel transfer mode control register.
DMPREC: DMA priority and enable control register.
DMS: DMA source address space select bit. Located in the DMMCR
register.
DMSA: The address register for DMA sub-addressed registers.
DMSDI: DMA sub-address data register with autoincrement.
DMSDN: DMA sub-address data register without autoincrement.
DMSFC: DMA channel sync select and frame count register.
DMSRC: DMA channel source address register.
DMSRCP: DMA source program page address (all channels).
DPRC: Configures the DMA channel priority level. Located in the DMPREC
register.
DR: data receive. Receives data from devices interfacing the McBSP.
DRR[1,2]: data receive register. Two 16-bit registers used to receive data
through the synchronous serial ports (McBSPs).
DSPINT: host port interface’s DSP CPU interrupt. The host can set this to in-
terrupt the DSP. Located in the HPIC register.
DSYN: DMA sync event control bits. Configures the DMA to synchronize to
a particular event. Located in the DMSFC register.
DX: data transmit. Communicates data to devices interfacing the McBSP.
DXENA: data transmit delay bit. When this bit is set, the transmitted data is
delayed.
DXR[1,2]: data transmit registers 1 and 2. Two 16-bit registers used to
transmit data through the synchronous serial ports (McBSPs).
DX_STAT: DX pin status bit. Indicates the logic level of the DX pin when DX
is configured as a general-purpose input.
A-4
E
Extended Software-Programmable Wait-State Generator: Extends
external bus cycles up to 14 machine cycles to interface with slower off-
chip memory and I/O devices.
F
FETCH: A bit located in the HPIC register that is used to fetch the data at the
current HPIA address.
FIFO: first-in-first-out. A hardware mechanism that allows several elements

to be stored in order. The first to be written to the FIFO is the first to be
read out.
FPER: frame period register. The FPER determines when the next frame-
sync signal becomes active.
Frame Count: DMA frame count. Specifies the number of frames to be

transferred. Located in the DMSFC register.
FREE: free running mode. Disabled when FREE = 0, and enabled when
FREE = 1.
FRLEN (R/X)[1,2]: frame length. The number of serial words (8-, 12-, 16-,
20-, 24-, or 32-bit) transferred per frame. The length corresponds to the
number of words, or logical time slots, or channels per frame-synchro-
nization signal.
FRST: frame-sync generator reset. Frame-sync logic is reset when FRST =

0. Frame-sync signal FSG is generated after (FPER + 1) number of
CLKG clocks when FRST = 1.
FSG: frame-sync signal. Generated by the frame-sync generation logic acti-

vated by the FRST bit.
FSGM: sample rate generator transmit frame-synchronization mode bit.

Located in the SRGR2 register.
FSR: receive frame synchronization. A McBSP interface signal.
FSRM: receive frame-synchronization mode bit. Located in the PCR

register.
FSRP: receive frame-synchronization polarity bit. Located in the PCR.
A-5
FSX: transmit frame synchronization. A McBSP interface signal.
FSXM: transmit frame-synchronization mode bit. Located in the PCR

register.
FSXP: transmit frame-synchronization polarity bit. Located in the PCR.
FWID: frame width. An 8-bit down counter that controls the active width of
the frame-sync pulse.
G
general-purpose input/output pin (GPIO): Pins that can be used to supply
input signals from an external device or output signals to an external
device. These pins are not linked to specific uses; rather, they provide
input or output signals for a variety of purposes, and include the general-
purpose BIO input pin and XF output pin.
GRST: sample rate generator reset. The sample rate generator is reset by
the GRST bit in SPCR2 when GRST = 0, and pulled out of reset when
GRST = 1.
GSYNC: sample rate generator clock synchronization.
H
HINT: DSP-to-host interrupt. The DSP can interrupt the host by writing to this
bit. Located in the HPIC register.
HMODE: HPI mode. Valid for the 16-bit HPI only. Allows the HPI to operate
in multiplexed/non-multiplexed modes.
Host: External device that drives the HPI peripherals.
Host Port Interface (HPI-8): An enhanced 8-bit parallel port that interfaces
a host device or host processor to the ’54x. Information is exchanged
between the ’54x and the host device through all on-chip ’54x RAM.
Host Port Interface (HPI-16): An enhanced 16-bit version of the ’54x 8-bit
host port interface that provides a full 16-bit bi-directional data bus, which
does not require byte identification. Memory accesses are synchronized
with the direct memory access (DMA) controller providing access to the
complete internal memory address range.
HPIA: host port address register. A register that is loaded with the address
that points to the data location to access. Accessible only by the host.
A-6
HPIC: host port control register. A register used by the host and DSP.
Location of the HINT, XHPIA, DSPINT, BOB, and HPIENA bits.
HPID: host port data register. Only accessible by the host. Data passes
through this register when using the HPI.
HPIENA: HPI enable status bit. Allows the host to determine the logic level
of the HPIENA pin. Located in the HPIC register.
HPI increment mode: The HPI address register is automatically modified

after the transfer completes.
HRDY: A bit located in the HPIC register that is used for software polling of
the HRDY pin.
HSTRB: Internal strobe that controls the HPI peripheral. The internal
HSTRB signal is a function of external pins HDS1, HDS2, and HCS.
I
IMOD: DMA interrupt generation mode bit. Configures the interrupt to occur
on full/half-buffer boundaries. Located in the DMMCR register.
INTOSEL: interrupt multiplexed control bits. Determines which IFR bits will
be associated with the DMA interrupts. Located in the DMPREC register.
IPINT: interprocessor interrupt. Located in the IFR and IMR registers.
IPIRQ: interprocessor interrupt request. Used to interrupt between subsys-

tems on multi-core DSPs.
L
latency: The delay between when a condition occurs and when the device
reacts to the condition. Also, in a pipeline, the necessary delay between
the execution of two instructions to ensure that the values used by the
second instruction are correct.
LSB: least significant bit. The lowest order bit in a word.
A-7
M
Master: A device that takes control of a peripheral, bus, etc.
MSB: most significant bit. The highest order bit in a word.
Multichannel Buffered Serial Port (McBSP): McBSPs are high-speed,

full-duplex, multichannel buffered serial ports that allow direct interface
to other ’54x devices, codecs, and other devices in a system. In addition,
the McBSP offer double-buffered registers, which allow a continuous
data stream, and independent framing and clocking for receive and
transmit.
Multiplexed mode: The HPI operates in a multiplexed address/data bus
fashion.
N
Non-multiplexed mode: The HPI operates with separate address and data
buses.
P
Pin Control Register (PCR): Used to configure the McBSP pins as inputs
or outputs during normal serial port operation. Also used to configure the
serial port pins as general purpose inputs or outputs during receiver and/
or transmitter reset.
Programmable Bank-Switching Module: Allows the ’54x to switch

between external memory banks without requiring external wait states
for memories that need several cycles to turn off. Bank-switching logic
automatically inserts one cycle when accesses cross a 32K word
memory-bank boundary inside program or data space.
R
RBR[1,2]: receive buffer registers 1 and 2. Receives copied data from the
receive shift register (RSR).
RCBLK: receive current block.

RCER[A,B]: McBSP receive channel enable register for partitions A and B.
RCOMPAND: receive companding mode. Configure the McBSP to

decompress receive data.
A-8
RDATDLY: receive data delay bit. Delays reception of data for 0, 1, or 2 bits
of delay.
Receive and Transmit Control Registers (RCR[1,2] and XCR[1,2]): Con-

figure various parameters of the receive and transmit operations.
Receive Buffer Register (RBR[1,2]): See RBR[1,2].
RESET: A means of bringing the CPU to a known state by setting the

registers and control bits to predetermined values and signaling execu-
tion to start at a specified address.
REVT: receive synchronization event to DMA.
REVTA: receive synchronization event to DMA in A-bis mode.
RFIG: receive frame ignore bit. Ignore subsequent FSRs after first.
RFRLEN1: receive frame length 1. Sets the receive frame length for first
phase.
RFRLEN2: receive frame length 2. Sets the receive frame length for second
phase.
RFULL: reception with overrun. Indicates that the receiver has experienced
overrun and is in an error condition.
RINT: receive interrupt to CPU.
RIOEN: receive section general-purpose I/O mode. The DX, FSX, and
CLKX pins are general-purpose I/Os.
RJUST: receive data justification and sign-extension. RJUST in the SPCR1

selects whether data in the RBR(1,2) is right or left justified (with respect
to the MSB) in the DRR(1,2). If right-justification is selected, RJUST
further selects whether the data is sign-extended or zero-filled.
RMCM: receive multichannel selection enable.
RPABLK: receive partition A block. Selects which block of channels to

receive. Located in the MCR1 register.
RPBBLK: receive partition B block. Selects which block of channels to

receive. Located in the MCR1 register.
RPHASE: receive phase bit. McBSP can be configured for one or two
phases.
RRDY: receive ready bit. Indicates the ready status of the McBSP receiver.
A-9
RRST: receive reset bit.
RSR[1,2]: receive shift registers 1 and 2. Receives shifted data from the DR
pin after the appropriate data delay.
rsvd: reserved.
RSYNCERR: unexpected receive frame synchronization. A sync pulse
which occurs RDATDLY minus 1 bit-clock earlier than the first bit of the
next associated word. This causes the current data reception to abort
and restart.
RWDLEN1: receive word length 1. Sets the receive word length for first
phase.
RWDLEN2: receive word length 2. Sets the receive word length for second
phase.
S
Sample Rate Generator: The sample rate generator is composed of a
three-stage clock divider that allows programmable data clocks (CLKG)
and framing signals (FSG), which are McBSP internal signals that can
be programmed to drive receive and/or transmit clocking (CLKR/X) and
framing (FSR/X). The sample rate generator can be programmed to be
driven by an internal clock source or an internal clock derived from an
external clock source.
Serial Port Control Register(SPCR[1,2]): Memory-mapped registers that
contains status and control bits for the serial-port interface.
SIND: DMA source address transfer index mode bit. Offers several different
indexing modes. Located in the DMMCR register.
Slave: A device that is controlled by another device.
SOFT: soft bit. SOFT mode is disabled when SOFT = 0, and enabled when
SOFT = 1.
Software-Programmable Wait-State Generator: See Extended Software-
Programmable Wait-State Generator.
SPCR[1,2]: serial port control registers 1 and 2.
SPSA: The address register for the McBSP sub-address registers.
SRGR[1,2]: sample rate generator registers 1 and 2.
Subsystem: Multi-core DSPs. Referred to as an independent system
consisting of a CPU, memory, and peripherals.
A-10
T
TDM Data Stream: Multiple channels can be independently selected for the
transmitter and receiver by configuring the McBSP with a single phase
frame. Each frame represents a time-division multiplexed (TDM) data
stream.
W
WDLEN(R/X)[1,2]: word length. 3-bit fields in the receive/transmit control
register that determine the word length in bits-per-word for the receiver
and transmitter for each phase of a frame.
X
XCBLK: transmit current block.
XCER[A,B]: McBSP transmit channel enable register for partitions A and B.
XCOMPAND: transmit companding mode. Configure the McBSP to

compress transmit data.
XDATDLY: transmit data delay bit. Delays transmission of data for 0, 1, or

2 bits of delay.
XEMPTY: transmit shift register (XSR[1,2]) empty. Indicates when the trans-
mitter has experienced underflow. When XSR[1,2] is empty, XEMPTY =
0. When XSR[1,2] is not empty, XEMPTY = 1.
XEVT: transmit synchronization event to DMA.
XEVTA: transmit synchronization event to DMA in A-bis mode.
XFIG: transmit frame ignore bit. Ignore subsequent FSXs after first.
XFRLEN1: transmit frame length 1. Sets the transmit frame length for first
phase.
XFRLEN2: transmit frame length 2. Sets the transmit frame length for
second phase.
XHPIA: extended HPIA. Select either the lower 16 bits or the most significant
bits to be accessed from the HPIA register. Located in the HPIC register.
XINT: transmit interrupt to CPU.
A-11
XINTM: receive/transmit interrupt mode.
XIOEN: transmit section general-purpose I/O mode. The DX, FSX, and
CLKX pins are general-purpose I/Os.
XMCM: transmit multichannel selection enable.
XPABLK: transmit partition A block. Selects which block of channels to

transmit. Located in the MCR1 register.
XPBBLK: transmit partition B block. Selects which block of channels to

transmit. Located in the MCR1 register.
XPHASE: transmit phase bit. McBSP can be configured for one or two
phases.
XRDY: transmit ready. Indicates the ready state of the McBSP transmitter.
When XRDY = 0, the transmitter is not ready. When XRDY = 1, the trans-
mitter is ready with data in DXR[1,2].
XRST: transmitter reset. This resets and enables the serial port transmitter.
Disables the transmitter and in reset state when XRST = 0, and enables
the transmitter when XRST = 1.
XSR[1,2]: transmit shift registers 1 and 2. Data written to the DXR[1,2] is

shifted out to DX via the transmit shift registers (XSR[1,2]).
XSYNCERR: unexpected transmit frame synchronization − transmit syn-

chronization error. A sync pulse which occurs earlier than XDATDLY bit-
clocks before the last transmit bit of the previous frame. When
XSYNCERR = 0, no synchronization error is detected. When
XSYNCERR = 1, an error is detected by the McBSP.
XWDLEN1: transmit word length 1. Sets the transmit word length for first
phase.
XWDLEN2: transmit word length 2. Sets the transmit word length for second
phase.
A-12
Appendix
AppendixBA
Revision History
Table B−1 lists the changes made since the previous version of this document.
Table B−1. Document Revision History

Page Additions/Modifications/Deletions
3-10 Updated Table 3−3.
B-1
B-2
Index
Index
HPID write using 5-13

A autoincrement operation 5-12
A-bis interface functionality 2-83 AUTOINIT, definition A-1
A-bis mode autoinitialization 3-27
definition A-1 automotive applications vii, xii
receive operation 2-84
transmit operation 2-84 B
ABIS Mode 2-8
BCLKX signal 2-92
ABU (autobuffering) mode 3-23
bi-directional FIFO 6-3
ABU address indexing modes 3-25
bit clock and frame synchronization 2-63
ABU buffer examples 3-24
bit clock polarity (CLKSP) 2-62
ABU buffer size examples 3-25
bit ordering 2-56
AC97 definition A-1
bit timing near frame-sync example 2-36 bits, FREE and SOFT 2-95
definition A-1
block diagram, McBSP 2-3
dual-phase frame format 2-36
multi-phase frame example 2-35 BOB, definition A-1
BOB and HPIA, initialization 4-20
access mode, definition A-1
BSP serial port control extension register (SPCE)
access sequence examples, HPI-8 4-19
CLKP bit 2-15
accesses to HPI-8 after reset 4-28 FSP bit 2-14
accesses to HPI-8 during reset 4-28 HALTR bit 2-74, 2-76
address autoincrement 4-11 buffered serial port 1-2
addressing modes 3-21
multiframe mode 3-21 C
ALE, definition A-1
changing channel selections 2-82
application(s)
automotive vii, xii channel enable control, DMA 3-9
consumer vii, xiii channel enable registers:, (R/X)CER(A/B) 2-80
development support vii, xiii channel enabling by blocks in partition A
general-purpose vii and B 2-77
medical vii, xiii channel-context registers, DMA 3-11
speech/voice vii, x CLK(R/X), definition A-2
autobuffering (ABU) mode 3-23 CLKG, definition A-2
autobuffering configuration 2-98 CLKGDV, definition A-2
autoincrement CLKGDV (sample rate generator data bit clock
HPID read using 5-12 rate) 2-62
Index-1
Index
CLKR, definition A-2

CLKRM
D
definition A-2 data clock generation 2-62
receive clock selection 2-65 data delay 2-34
CLKRP, definition A-2 2-bit used to discard framing bit 2-35
CLKS, definition A-2 data packing 2-32
at maximum packet frequency 2-42
CLKS_STAT, definition A-2
using frame-sync ignore bits 2-41
CLKSM, definition A-2 data receive (DR) pin 2-4
CLKSM (input clock source mode) 2-62 data receive register (DRR) 2-4
CLKSP, definition A-2 data sorting
CLKSP (bit clock polarity) 2-62 ABU mode 3-23
by address modification 3-22
CLKSRG, definition A-2 example 3-23
CLKSTP 2-90 data transmission and reception flow 2-22
definition A-2 data transmit (DX) pin 2-4
CLKSTP bit field 2-88 DATDLY, definition A-3
CLKX, definition A-2 DBLW, definition A-3
CLKXM 2-90 DE, definition A-3
definition A-2 development support applications vii, xiii
transmit clock selection 2-65 device clocks 2-98
CLKXP 2-90 device reset 2-23
definition A-2 McBSP operation 2-93
clock and frame generation 2-57 digital loop-back mode (DLB) 2-64
clock modes, DSP, changes that affect the DIND, definition A-3
HPI-16 5-22 DINM, definition A-3
clock stop mode 2-8 DIR, definition A-3
clock stop mode configurations 2-88 direct memory access (DMA) controller 1-2
definition A-3
clock stop mode configurations 2-88
DIV clock modes, HPI-8 transfers while switching
clocking and framing, sample rate generator 2-58 to 4-24
clocking examples 2-69 DLB, definition A-3
companding DLB (digital loop back mode)
definition A-2 receive clock selection 2-65
flow 2-54 receive frame-sync selection 2-67
internal data 2-55 DLB (digital loop-back mode) 2-64
companding hardware operation, DMA (direct memory access) controller 1-2
(R/X)COMPAND 2-53 block transfer interrupt generation modes 3-28
channel 0 and 1, configured for FIFO transmit
configuration, serial port 2-6
and receive 6-4
configuration of pins as general purpose I/O, channel enable control 3-9
table 2-97 channel priority control 3-11
consumer applications vii, xiii channel transfer rate example 3-40
control applications vii, x channel-context registers 3-11
element count register 3-12
CPU, interrupts 2-27 source and destination address
CTMOD, definition A-2 registers 3-11
Index-2
Index
sync event and frame count register DMA memory map

(DMSFCn) 3-13 ’5402 3-34
channels 2-85 ’5410 3-35
configuration for FIFO operation 6-5 ’5420 3-37
configuration to support FIFO transfers 6-4 DMA registers 3-6
definition A-3 DMA channel priority and enable control
destination program page address register (DMPREC), bit-field descriptions 3-8
(DMDSTP) 3-33 DMA sync event and frame count register
emulation control 3-11 (DMSFCn) 3-13
enhanced HPI access 3-42 bit-field descriptions 3-13
frame count 3-17
DMA synchronization events 3-13
memory map
DMA transfer mode control register
’5402 3-34
(DMMCRn) 3-17
’5410 3-35
bit-field descriptions 3-17
’5420 3-37
multiplexed interrupt control 3-9 DMCTR, definition A-3
operation and configuration 3-4 DMD, definition A-3
operation in power-down mode 3-43 DMDST, definition A-3
overview 3-2 DMDSTP, definition A-3
programming examples 3-44 DMDSTP (DMA destination program page address
register subaddressing 3-4 register) 3-33
with autoincrement 3-5
DMFRI0, definition A-3
without autoincrement 3-5
registers 3-6 DMFRI1, definition A-3
channel priority and enable control register DMGCR, definition A-3
(DMPREC) 3-7 DMGDA, definition A-4
destination program page address register DMGFR, definition A-4
(DMDSTP) 3-33 DMGSA, definition A-4
source program page address register
DMIDX0, definition A-4
(DMSRCP) 3-33
sync event and frame count register DMIDX1, definition A-4
(DMSFCn) 3-13 DMMCR, definition A-4
transfer mode control register DMMCRn (DMA transfer mode control
(DMMCRn) 3-17 register) 3-17
source program page address register bit-field descriptions 3-17
(DMSRCP) 3-33 DMPREC, definition A-4
sync event options DMPREC (DMA channel priority and enable control
’5402 3-14 register) 3-7
’5410 3-15 bit-field descriptions 3-8
’5420 3-16 DMS, definition A-4
synchronization events 3-13
DMSA, definition A-4
transfer cycle times 3-39
transfer latency 3-39 DMSDI, definition A-4
transfers, double-word mode 3-16 DMSDN, definition A-4
DMSFC, definition A-4
DMA block transfer interrupt generation
modes 3-28 DMSFCn (DMA sync event and frame count
register) 3-13
DMA channel priority and enable control register bit-field descriptions 3-13
(DMPREC) 3-7 DMSRC, definition A-4
DMA channels 2-85 DMSRCP, definition A-4
Index-3
Index
DMSRCP (DMA source program page address extended software-programmable wait-state

register) 3-33 generator, definition A-5
dormant state 2-98 external data communications 2-4
double-rate clock example 2-71 external master clock 2-92
double-rate ST-BUS clock 2-69 external memory interface, HPI-16 connection,
double-word mode, DMA transfer 3-16 ’VC5420 6-7
DPRC, definition A-4 external parallel interface (XIO2)
bus sequences 7-3
DR, definition A-4 consecutive memory read 7-4
DRR, definition A-4 memory write and I/O write 7-5
DSP, articles vii, ix, x, xi, xii, xiii non-consecutive memory read and I/O
DSPINT, definition A-4 read 7-3
introduction 7-2
DSPINT and HINT operations 4-23
DSYN, definition A-4
dual-phase frame example 2-30 F
DX, definition A-4 FETCH, definition A-5
DX Enabler 2-8 FIFO
in ABIS mode 2-37 bi-directional 6-3
in normal mode 2-37 configuration, ’VC5420 6-3
DX pin, delay enable/disable 2-37 configured for transmit and receive, DMA
DX_STAT, definition A-4 channels 0 and 1 6-4
definition A-5
DXENA, definition A-4
DMA configuration to support transfers 6-4
DXR, definition A-4 operation, DMA configuration for 6-5
DXR2 2-5 FPER, definition A-5
FPER (frame period) 2-66
E frame and clock
configuration 2-27
element count register 3-12 operation 2-28
EMIF/HPI-16 modes 6-8 frame count
definition A-5
emulation bits 2-95
DMA 3-17
emulation control for DMA functions 3-11 frame detection for initialization 2-68
emulation FREE and SOFT bits 2-95 frame example, multi-phase: AC97 2-35
enabling and masking of channels 2-76 frame format, dual-phase: AC97 2-36
enabling multichannel selection 2-76 frame length 2-31
enhanced 16-bit host port interface (HPI-16) 5-1 frame period (FPER) 2-66
enhanced 8-bit host port interface (HPI-8) 4-1 frame width (FWID) 2-66
enhanced 8-bit host post interface (HPI-8) frame-sync generator reset 2-10
basic functional description 4-4 frame-sync signal generation 2-4, 2-66
introduction 4-2 frame-synchronization ignore 2-40
enhanced external bus interface (EnhXIO) 7-1 and unexpected frame-sync pulses 2-42
EnhXIO (enhanced external bus interface) 7-1 frame-synchronization phases 2-30
events, XEVTA and REVTA 2-84 FREE, definition A-5
extended addressing, DMA 3-33 FREE bits 2-95
extended HPI-8 addressing 4-10 free running mode 2-10
Index-4
Index
free runs 2-95

FRLEN(R/X), definition A-5
H
FRST, definition A-5 half-buffer interrupt generation in ABU mode
FSG, definition A-5 ’5402/’5420 3-28
FSGM ’5410 3-31
definition A-5 HAS
transmit frame-sync signal selection 2-68 host accesses with 5-10
FSR, definition A-5 host accesses without 5-11
HPIA write using 5-10
FSRM
HPIC read without using 5-11
definition A-5
HCNTL0/1 modes 5-8
receive frame-sync selection 2-67
HINT, definition A-6
FSRP, definition A-5
HMODE, definition A-6
FSX, definition A-6
host, definition A-6
FSXM
host accesses with HAS 5-10
definition A-6
transmit frame-sync signal selection 2-68 host accesses without HAS 5-11
FSXP, definition A-6 host and ’54x accesses, HPIC diagram 4-13
FWID, definition A-6 host device using DSPINT to interrupt the
’54x 4-23
FWID (frame width) 2-66
host port interface 1-2, 4-2
block diagram 4-3
G generic system block diagram 4-4
host port interface (HPI-16) A-6
general purpose I/O, McBSP pins 2-96 host port interface (HPI-8), definition A-6
general purpose I/O control register host read/write access to HPI-8 4-14
(GPIOCR) 4-30 HPI
bit functions 4-30 See also host port interface
general purpose I/O status register (GPIOSR) 4-32 strobe and select logic 5-3
bit functions 4-32 HPI control register (HPIC), bit descriptions 4-12
general-purpose input/output pin (GPIO), HPI increment mode, definition A-7
definition A-6
HPI-16 1-2
general-purpose applications vii HPI-16 (enhanced 16-bit host port interface)
generator reset bit (GRST) 2-93 accessing from external memory space 6-7
GPIO changes in DSP clock modes 5-22
code example 4-35 DMA interaction 5-19
definition A-6 during the PLL to DIV clock mode change 5-22
GPIO feature, using 4-35 EMIF modes 6-8
GPIOCR (general purpose I/O control external memory interface connection,
register) 4-30 ’VC5420 6-7
bit functions 4-30 interfacing in non-multiplexed mode,
’VC5420 5-15
GPIOSR (general purpose I/O status register) 4-32
memory map 5-18
bit functions 4-32
relative to the HPI, ’VC5420 5-18
graphics/imagery applications vii, ix operation
GRST, definition A-6 during IDLEn 5-21
GSYNC during reset 5-21
definition A-6 operational overview 5-2
receive frame-sync selection 2-67 pin descriptions 5-4
Index-5
Index
HPI-8 1-2 HPID read using autoincrement 5-12

HPI-8 (enhanced 8-bit host port interface) 4-1 HPID write in non-multiplexed mode 5-17
access sequence examples 4-19 HPID write using autoincrement 5-13
accesses HPIENA, definition A-7
during IDLE1 and IDLE2 4-26
HRDY, definition A-7
during IDLE3 4-26
accesses after reset 4-28 HSTRBint, definition A-7
accesses during reset, ’5410 4-28
address register and memory map 4-9
basic functional description 4-4
I
control register bits and functions 4-12 IDLE 2-98
data pins, as general purpose I/O pins 4-30 IDLE use, considerations 4-26
details of operation 4-6
IDLE1 and IDLE2, HPI-8 accesses during 4-26
extended addressing 4-10
host read/write access 4-14 IDLE3, HPI-8 accesses during 4-26
input control signals and function selection 4-9 IDLEn, HPI-16 operation during 5-21
internal delays by access type 4-17 IMOD, definition A-7
introduction 4-2 initialization of BOB and HPIA 4-20
latency of accesses 4-16 input clock source mode (CLKSM) 2-62
memory maps, ’VC5402 and ’VC5410 4-9 internal data movement 2-4
operation
interprocessor communications 6-1
during reset (’5410) 4-29
within multi-core DSPs 6-2
effects of reset 4-28
read access with autoincrement 4-21 interprocessor FIFO communication, ’5420 3-43
register descriptions 4-5 interprocessor interrupts 6-11
signal names and functions 4-6 interrupt generation, DMA 3-27
strobe and select logic 4-8 interrupts, interprocessor 6-11
timing diagram 4-15 INTOSEL, definition A-7
transfers
introduction 1-1 to 1-8
considerations while changing clock
modes 4-24 IPINT, definition A-7
while switching to DIV clock modes 4-24 IPIRQ, definition A-7
while the ’54x is in IDLE3 mode 4-27
transfers while changing clock modes,
considerations 4-24
L
wait-state generation conditions 4-18 latency, definition A-7
write access with autoincrement 4-22
latency of HPI-8 accesses 4-16
HPIA, definition A-6
LSB, definition A-7
HPIA write using HAS 5-10
HPIC
bit descriptions 5-8 M
definition A-7
register 5-8 master, definition A-8
master−slave configuration 2-86
HPIC (HPI control register), bit descriptions 4-12
maximum frame frequency 2-39
HPIC diagram, host and ’54x accesses 4-13
receive/transmit 2-40
HPIC host and ’54x, read/write characteristics 4-13
McBSP 2-1
HPIC read without using HAS 5-11 block diagram 2-3
HPID, definition A-7 clock configuration 2-95
HPID read in non-multiplexed mode 5-16 CPU interrupts 2-6
Index-6
Index
definition A-8 MCR2 (multichannel control register 2) 2-75

DMA event synchronization 2-6 medical applications vii, xiii
features 2-2 memory map
general description 2-3 HPI-16 5-18
initialization for SPI mode 2-93 relative to the HPI, ’VC5420 5-18
interface signals 2-4
military applications vii, xi
maximum frame frequency 2-39
McBSP to McBSP connection 6-10 MISO (Master in - Slave out) 2-86
MCR1 (multichannel control register 1), bit-field Mitel ST-BUS 2-69
descriptions 2-73 MOSI (Master out - Slave in) 2-86
multichannel control register 2 (MCR2) 2-75 MP/MC bit levels at reset 6-8
bit-field descriptions 2-75 MSB, definition A-8
operation
multi-channel buffered serial ports 1-2
as an SPI master 2-90
as an SPI slave 2-92 multi-core DSPs, communications within 6-2
in power-down mode 2-98 multi-subsystem DSPs 6-2
pin configurations, general-purpose I/O 2-97 multichannel buffered serial port (McBSP)
programming example code 2-99 definition A-8
receive operation 2-38 features 2-2
receive/transmit frame length configuration 2-31 multichannel buffered serial ports (McBSPs 2-1
receive/transmit word length configuration 2-32 multichannel control register 1 (MCR1) 2-73
registers 2-5 bit-field descriptions 2-73
MCR1 (multichannel control register 1) 2-73
multichannel control register 2 (MCR2) 2-75
reset 2-23
reset state of pins 2-23
standard operation 2-38 multichannel operation control registers 2-73
subsystem communication using 6-10 multichannel selection operation 2-72
transmit operation 2-39 multimedia applications vii, xi
McBSP clock configuration 2-95 multiplexed interrupt assignments
McBSP CPU interrupts and DMA event ’5402 3-10
synchronization, table 2-6 ’5410 3-10
’5420 3-10
McBSP initialization for SPI mode 2-93
multiplexed interrupt control, DMA 3-9
McBSP operation
as an SPI master 2-90 multiplexed mode 5-7
as an SPI slave 2-92 definition A-8
interfacing to the HPI-16, ’VC5420 5-7
McBSP operation in power-down mode 2-98
McBSP pin configurations, general purpose I/O,
table 2-97 N
McBSP pins as general purpose I/O 2-96
non-multiplexed mode 5-15
McBSP programming example code 2-99 definition A-8
McBSP registers 2-5 HPID read 5-16
McBSP reset 2-23 HPID write 5-17
McBSP to McBSP connection 6-10
McBSPs 1-2 P
MCR1 (multichannel control register 1) 2-73
bit-field descriptions 2-73 packet length 2-90
MCR2 (multichannel control register 2), bit-field PCR, definition A-8
descriptions 2-75 PCR (pin control register) 2-12
Index-7
Index
peripherals receive channel enable register partition A

host port interface 1-2 (RCERA) 2-80
host port interface (HPI) 4-2 bit-field descriptions 2-80
pin configuration register (PCR) 2-88 receive channel enable register partition B
pin control register (PCR) A-8 (RCERB) 2-81
Pin Control Register (PCR) 2-6, 2-12
bit-field descriptions 2-12 receive clock selection: DLB, CLKRM 2-65
receive clock signal (BCLKR) 2-86
PLL to DIV, clock mode change, HPI-16 operation
during 5-22 receive control register 1 (RCR1) 2-16
power down modes 2-98
receive control register 2 (RCR2) 2-17
programmable bank-switching logic, definition A-8
programmable clock and framing 2-57 receive data clocking 2-30
programmable frame period and width 2-67 receive data justification and sign extension,
programming example code 2-99 RJUST 2-52
receive frame synchronization signal (BFSR) 2-86
receive frame-sync selection:, DLB, FSRM,
R GSYNC 2-67
receive frame-synchronization pulse, response
RBR 2-5
to 2-47
definition A-8
receive operation 2-22
RCBLK, definition A-8
receive ready status 2-26
RCER, definition A-8
receive shift register (RSR) 2-4
RCERA (receive channel enable register
receive/transmit word length configuration 2-32
partition A) 2-80
bit-field descriptions 2-80 receiver reset, McBSP operation 2-93
reception with overrun, RFULL 2-44
RCERB (receive channel enable register
partition B) 2-81 register bit values for SPI master operation 2-91
bit-field descriptions 2-81 register bit values for SPI mode configuration 2-90
RCOMPAND, definition A-8 register bit values for SPI slave operation 2-92
RCR, definition A-9 register subaddressing 3-4
with autoincrement 3-5
RCR1 (receive control register 1) 2-16
without autoincrement 3-5
registers
RCR2 (Receive Control Register 2) 2-17
data transmit register (DXR) 2-4
RCR2 (receive control register 2), bit-field pin control register (PCR) 2-12
descriptions 2-17 receive buffer register (RBR) 2-4
RDATDLY 2-34 receive control register 1 2-16
definition A-9 receive control register 2 2-17
read access with autoincrement, HPI-8 4-21 receive shift register (RSR) 2-4
serial port control register 1 2-7
read/write characteristics, HPIC host and ’54x 4-13
serial port control register 2 2-10
ready state of the McBSP receiver 2-26 transmit shift register (XSR) 2-4
ready status, determining 2-26 RESET 2-22
receive and transmit control registers (RCR and definition A-9
XCR), definition A-9 reset
receive buffer register (RBR) 2-4 effects on HPI-8 operation 4-28
definition A-9 HPI-16 operation during 5-21
Index-8
Index
HPI-8 operation during, ’5410 4-29 sample rate generator register 2 (SRGR2) 2-60
MP/MC bit levels at 6-8 bit-field descriptions 2-60
reset state of McBSP pins 2-23 SCK (shift clock) 2-86
REVT 2-26 Serial data input (Master In−Slave Out, or
definition A-9 MISO) 2-86
REVTA, definition A-9 serial data output (Master Out−Slave In, or
REVTA (receive A-bis event) 2-84 MOSI) 2-86
RFIG, definition A-9 serial port
configuration 2-6
RFRLEN(1,2) 2-31
exception conditions 2-43
RFRLEN1, definition A-9 receive overrun 2-45
RFRLEN2, definition A-9 receive overrun avoided 2-46
RFULL 2-44 reset 2-22
definition A-9 serial port control register (SPC)
RINT 2-26 DLB bit 2-7
definition A-9 RSRFULL bit 2-9
RIOEN 2-12 serial port control register (SPCR), definition A-10
definition A-9 Serial Port Control Register 1 (SPCR1)
RJUST, definition A-9 bit-field descriptions 2-7
RMCM, definition A-9 Figure 2-7
RPABLK, definition A-9 serial port control register 1 (SPCR1) 2-88
RPBBLK, definition A-9 serial port control register 2 (SPCR2) 2-93
RPHASE, definition A-9
serial port control registers 2-6
RRDY 2-26
serial port interface 2-2
definition A-9
buffered serial ports 1-2
RRST, definition A-10
serial ports 1-2
RRST bit 2-93
shift clock (SCK) 2-86
RSR, definition A-10
SIND, definition A-10
rsvd, definition A-10
Single- Phase Frame, one 32-bit word 2-34
RSYNCERR 2-46
single-phase frame, four 8-bit words 2-33
definition A-10
single-rate clock example 2-70
RWDLEN(1,2) 2-31
single-rate ST-BUS clock 2-70
RWDLEN1 2-90
definition A-10 slave, definition A-10
RWDLEN2, definition A-10 slave enable signal (SS) 2-86
SOFT, definition A-10
SOFT bits 2-10, 2-95
S software-programmable wait-state generator,
definition A-10
sample rate generator 2-92
source and destination address, selection and
clocking and framing 2-58
modification 3-20
data bit clock rate (CLKGDV) 2-62
definition A-10 source and destination address registers,
diagram 2-58 DMA 3-11
reset procedures 2-61 SPCR, definition A-10
sample rate generator register 1 (SRGR1) 2-59 SPCR1, bit-field descriptions 2-7
bit-field descriptions 2-59 SPCR2, bit-field descriptions 2-10
Index-9
Index
SPI Configuration transmit channel enable register partition A

McBSP as the master 2-87 (XCERA) 2-81
McBSP as the slave 2-87 bit-field descriptions 2-81
SPI master, McBSP operation as 2-90 transmit channel enable register partition B
SPI mode, McBSP initialization 2-93 (XCERB) 2-82
SPI modes 2-8
transmit clock selection: CLKXM 2-65
SPI protocol, McBSP clock stop mode 2-86
transmit clock signal (BCLKX) 2-86
SPI slave, McBSP operation as 2-92 transmit control register 1 (XCR1) 2-19
SPI Transfer with CLKSTP = 10b, and CLKXP = 0 bit-field descriptions 2-19
2-88 Transmit Control Register 2 (XCR2), bit-field
SPI Transfer with CLKSTP = 10b, and CLKXP = 1 descriptions 2-20
2-89 transmit control register 2 (XCR2) 2-20
SPI Transfer with CLKSTP = 11b, and CLKXP = 0 transmit empty
2-89 avoided 2-50
SPI Transfer with CLKSTP = 11b, and CLKXP = 1 diagram 2-49
2-89 XEMPTY 2-48
SPSA, definition A-10 transmit frame synchronization, response to 2-51
transmit frame-sync signal selection:, FSXM,
SRGR, definition A-10
FSGM 2-68
SRGR1 (sample rate generator register 1) 2-59 transmit operation 2-22
transmit ready status 2-26
SRGR2 (sample rate generator register 2 2-60
transmit shift register (XSR) 2-4
transmit with data overwrite 2-48
SS (slave enable) 2-86
transmitter reset, McBSP operation 2-93
ST-BUS and MVIP example 2-69 transmitter reset bit (XRST) 2-93
ST-BUS clock typical SPI interface 2-86
double-rate 2-69
single-rate 2-70
subsystem, definition A-10
U
subsystem communication using McBSP 6-10 unexpected receive frame-synchronization,
RSYNCERR 2-46
unexpected receive synchronization pulse 2-48
T unexpected transmit frame synchronization,
XSYNCERR 2-50
TDM data stream, definition A-11 unexpected transmit frame-synchronization
TDM serial port control register (TSPC) pulse 2-52
MCM bit 2-60 update interrupts 2-83
RRDY bit 2-9
RRST bit 2-9
TXM bit 2-13, 2-16, 2-17, 2-18, 2-19, 2-20, W
2-21, 2-61 wait-state generation conditions 4-18
XRDY bit 2-11
WDLEN(R/X), definition A-11
XRST bit 2-11
word length 2-31, 2-32
telecommunications applications vii, xii
wrap address calculation
transfer cycle times, DMA 3-39 double-word transfer with indexed
transfer latency, DMA 3-39 addressing 3-26
Index-10
Index
single-word transfer with indexed XHPIA, definition A-11

addressing 3-26 XINT 2-26
write access to HPI with autoincrement 4-22 definition A-11
XINTM, definition A-12
X XIO2 (external parallel interface)
bus sequences 7-3
XCBLK, definition A-11 consecutive memory read 7-4
XCER, definition A-11 memory write and I/O write 7-5
non-consecutive memory read and I/O
XCERA (transmit channel enable register
read 7-3
partition A) 2-81
introduction 7-2
XIOEN 2-12
XCERB (transmit channel enable register
definition A-12
partition B) 2-82
bit-field descriptions 2-82 XMCM, definition A-12
XCOMPAND, definition A-11 XMCM operation, diagram 2-79
XCR, definition A-9 XPABLK, definition A-12
XCR1 (transmit control register 1) 2-19 XPBBLK, definition A-12
bit-field descriptions 2-19 XPHASE, definition A-12
XCR2 (transmit control register 2) 2-20 XRDY 2-26
bit-field descriptions 2-20 definition A-12
XDATDLY 2-34 XRDY bit 2-11
definition A-11 XRST, definition A-12
XEMPTY A-11 XRST bit 2-11, 2-93
XEMPTY bit 2-11 XSR 2-5
XEVT 2-26 definition A-12
definition A-11 XSREMPTY bit 2-11
XEVTA, definition A-11 XSYNCERR 2-50
XEVTA (transmit A−bis event) 2-84 definition A-12
XEVTA event 2-85 XSYNCERR bit 2-11
XFIG, definition A-11 XWDLEN(1,2) 2-31
XFRLEN(1,2) 2-31 XWDLEN1 2-90
XFRLEN1, definition A-11 definition A-12
XFRLEN2, definition A-11 XWDLEN2, definition A-12
Index-11
Index-12

09 Spru302b DMA TMS320C54XX

Uploaded by

Copyright:

Available Formats

09 Spru302b DMA TMS320C54XX

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

09 Spru302b DMA TMS320C54XX

Uploaded by

Copyright:

Available Formats

TMS320C54x DSP

Volume 5: Enhanced Peripherals

Literature Number: SPRU302B

Printed on Recycled Paper

Mailing Address: Texas Instruments

Copyright  2007, Texas Instruments Incorporated

Read This First

About This Manual

The TMS320C54x is a fixed-point digital signal processor (DSP) in the Texas

The main topics discussed are:

For all non-enhanced peripheral information related to ’54x devices, see

- Program listings and program examples are shown in a special type-

- Square brackets, [ and ], identify an optional parameter. If you use an op-

Information About Cautions

This is an example of a caution statement.

Related Documentation from Texas Instruments

The following books provide related documentation for the TM320C54x. To

Many of these documents are located on the internet at http://www.ti.com.

TMS320C54x DSP Reference Set, Volume 1: CPU and Peripherals

TMS320C54x DSP Reference Set, Volume 2: Mnemonic Instruction

TMS320C54x DSP Reference Set, Volume 3: Algebraic Instruction

TMS320C54x DSP Reference Set, Volume 4: Applications Guide

TMS320C54x DSP Reference Set, Volume 5: Enhanced Peripherals

TMS320C54x, TMS320LC54x, TMS320VC54x Fixed-Point Digital Signal

Read This First v

TMS320C54x DSKplus User’s Guide (literature number SPRU191)

TMS320C548/C549 Bootloader Technical Reference (literature number

In the following list, references appear in alphabetical order according to

Read This First vii

3) Erskine, C., and S. Magar, “Architecture and Applications of a Second-

4) Essig, D., C. Erskine, E. Caudel, and S. Magar, “A Second-Generation

5) Frantz, G., K. Lin, J. Reimer, and J. Bradley, “The Texas Instruments

6) Gass, W., R. Tarrant, T. Richard, B. Pawate, M. Gammel, P. Rajasekaran,

9) Lim, Jae, and Alan V. Oppenheim, Advanced Topics in Signal Processing,

14) Papamichalis, P.E., and C.S. Burrus, “Conversion of Digit-Reversed to Bit-

16) Papamichalis, P.E., “FFT Implementation on the TMS320C30,” Proceed-

19) Prado, J., and R. Alcantara, “A Fast Square-Rooting Algorithm Using a

1) Reimer, J., and A. Lovrich, “Graphics with the TMS32020,” WESCON/85

Read This First ix

1) DellaMorte, J., and P. Papamichalis, “Full-Duplex Real-Time Implementa-

4) Papamichalis, P., and D. Lively, “Implementation of the DOD Standard

5) Papamichalis, Panos, Practical Approaches to Speech Coding, Engle-

6) Pawate, B.I., and G.R. Doddington, “Implementation of a Hidden Markov

7) Rabiner, L.R., and R.W. Schafer, Digital Processing of Speech Signals,

9) Reimer, J.B., M.L. McMahan, and W.W. Anderson, “Speech Recognition

2) Ahmed, I., “Implementation of Self Tuning Regulators with TMS320 Family

3) Ahmed, I., and S. Lindquist, “Digital Signal Processors: Simplifying High-

4) Ahmed, I., and S. Meshkat, “Using DSPs in Control,” Control Engineering,

6) Bose, B.K., and P.M. Szczesny, “A Microcomputer-Based Control and

7) Hanselman, H., “LQG-Control of a Highly Resonant Disc Drive Head Posi-

8) Lovrich, A., G. Troullinos, and R. Chirayil, “An All-Digital Automatic Gain

9) Matsui, N. and M. Shigyo, “Brushless DC Motor Control Without Position

1) Reimer, J., “DSP-Based Multimedia Solutions Lead Way Enhancing

2) Reimer, J., G. Benbassat, and W. Bonneau Jr., “Application Processors:

1) Papamichalis, P., and J. Reimer, “Implementation of the Data Encryption

Read This First xi

1) Ahmed, I., and A. Lovrich, “Adaptive Line Enhancer Using the

2) Casale, S., R. Russo, and G. Bellina, “Optimal Architectural Solution Us-

3) Cole, C., A. Haoui, and P. Winship, “A High-Performance Digital Voice