Lecture 15 Finite State Machine Implementation

Prith Banerjee ECE C03 Advanced Digital Design Spring 1998

ECE C03 Lecture 15 1

Outline
Mapping FSM to random logic Mapping FSM to ROMS Mapping FSM to PLAs Mapping FSM to Programmable Logic Devices (Xilinx) READING: Katz 10.1.1, 10.1.2, 10.3, Dewey 9.5

ECE C03 Lecture 15

FSM Implementation Strategies

Discrete Gate Logic Emphasis so far MSI Logic (e.g., Counters) Structured Logic (e.g., PLA/PAL, ROM) Field Programmable Gate Arrays (FPGAs) Function can be configured "on the fly" or in the field Flipflops/Registers plus discrete gates on the same chip

ECE C03 Lecture 15

FSM Design with Structured Logic

Combinational Logic Inputs Output Function Next State Function Registers Outputs

Block Diagram for Synchronous Mealy Machine

State

ROM-based Realization
ROM A0 Inputs An-1 An Dk-1 Dk D0 Outputs Registers

Inputs & Current State form the address ROM data bits form the Outputs & Next State

An+m-1 Dk+m-1

State
ECE C03 Lecture 15 4

ROM-Based Implementation

Example: BCD to Excess 3 Serial Converter BCD Excess 3 Code 0000 0011 0001 0100 0010 0101 0011 0110 0100 0111 0101 1000 0110 1001 0111 1010 1000 1011 1001 1100

Conversion Process Bits are presented in bit serial fashion starting with the least significant bit

Single input X, single output Z

ECE C03 Lecture 15

BCD to Excess-3 Converter

Present State S0 S1 S2 S3 S4 S5 S6 Next X=0 S1 S3 S4 S5 S5 S0 S0 State X=1 S2 S4 S4 S5 S6 S0 -Output X=0 X=1 1 0 1 0 0 1 0 1 1 0 0 1 1 --

State Transition Table

Reset

0/1
S1 0/1 1/0

S0

1/0
S2 0/0, 1/1 S4 1/0 S6 0/1 ECE C03 Lecture 15 6

Derived State Diagram

S3 0/0, 1/1 S5 0/0, 1/1 0/1

ROM-Based Implementation
BCD to Excess 3 Converter
ROM Address X Q2 Q1 Q0 0 0 0 0 0 0 0 1 0 0 1 0 0 0 1 1 0 1 0 0 0 1 0 1 0 1 1 0 0 1 1 1 1 0 0 0 1 0 0 1 1 0 1 0 1 0 1 1 1 1 0 0 1 1 0 1 1 1 1 0 1 1 1 1 ROM Outputs Z D2 D1 D0 1 0 0 1 1 0 1 1 0 1 0 0 0 1 0 1 1 1 0 1 0 0 0 0 1 0 0 0 X X X X 0 0 1 0 0 1 0 0 1 1 0 0 1 1 0 1 0 1 1 0 1 0 0 0 X X X X X X X X

1 CLK 1 0 X converter ROM X Q2 Q1 Q0 Z D2 D1 D0 1 0 9 13 12 5 4

CLK
D C B A

QD 175 QD QC QC QB QB

15 14 10 11 7 6

1 CLR \Reset

2 QA 3 QA

Circuit Level Realization 74175 = 4 x positive edge triggered D FFs

Truth Table/ROM I/Os

In ROM-based designs, no need to consider state assignment ECE C03 Lecture 15

ROM Based Implementation

BCD to Excess-3 Converter LSB MSB Timing Behavior for input strings 0 0 0 0 (0) and 1 1 1 0 (7)

0 1

0 1

0 0

0 0

1 0

1 1

1 0

0 1

0000
LSB

1100

1110

0101

LSB

ECE C03 Lecture 15

PLA Based Design

BCD to Excess 3 Converter

State Assignment with NOVA 0 1 0 1 0 1 0 1 0 1 0 1 0 S0 S0 S1 S1 S2 S2 S3 S3 S4 S4 S5 S5 S6 S1 S2 S3 S4 S4 S4 S5 S5 S5 S6 S0 S0 S0 1 0 1 0 0 1 0 1 1 0 0 1 1 S0 = 000 S1 = 001 S2 = 011 S3 = 110 S4 = 100 S5 = 111 S6 = 101 NOVA derived state assignment 9 product term implementation
ECE C03 Lecture 15 9

NOVA input file

PLA Implementation
BCD to Excess 3 Converter .i 4 .o 4 .ilb x q2 .ob d2 d1 .p 16 0 000 001 1 000 011 0 001 110 1 001 100 0 011 100 1 011 100 0 110 111 1 110 111 0 100 111 1 100 101 0 111 000 1 111 000 0 101 000 1 101 --0 010 --1 010 --.e Espresso Inputs .i 4 .o 4 .ilb x q2 q1 q0 .ob d2 d1 d0 z .p 9 0001 0100 10-0 0100 01-0 0100 1-1- 0001 -0-1 1000 0-0- 0001 -1-0 1000 --10 0100 ---0 0010 .e

q1 q0 d0 z
1 0 1 0 0 1 0 1 1 0 0 1 1 -

Espresso Outputs

ECE C03 Lecture 15

10

PLA Implementation
BCD to Excess 3 Converter D2 = Q2 Q0 + Q2 Q0 D1 = X Q2 Q1 Q0 + X Q2 Q0 + X Q2 Q0 + Q1 Q0 D0 = Q0 Z = X Q1 + X Q1
1 CLK 9 1 0 X conv erter PLA X Q2 Q1 Q0 Z D2 D1 D0 1 0 13 12 5 4 CLK D C B A 15 QD 14 QD 10 QC 11 QC 7 QB 6 QB 2 QA 3 QA

175

1 CLR \Reset

ECE C03 Lecture 15

11

PAL Implementation
BCD to Excess 3 Serial Converter 10H8 PAL: 10 inputs, 8 outputs, 2 product terms per OR gate
D1 = D11 + D12 D11 = X Q2 Q1 Q0 + X Q2 Q0

D12 = X Q2 Q0 + Q1 Q0
0 1 2 3 X 45 89 12 13 16 17 20 21 24 25 28 29 30 31

0. Q2 Q0 1. Q2 Q0 8. X Q2 Q1 Q0 9. X Q2 Q0 16. X Q2 Q0 17. Q1 Q0 24. D11 25. D12 32. Q0 33. not used 40. X Q1 41. X Q1

0 1 Q2 8 9 Q1 16 17 Q0 24 25 D11 32 33 D12 40 41

D2

D11

D12

D1

D0

ECE C03 Lecture 15

12

PAL Implementation
BCD to Excess 3 Serial Converter
0 1 2 3 X 0 1 Q2 8 9 Q1 16 17 Q0 24 25 D11 32 33 D12 40 41 Z D0 D1 D12 D11 D2 45 89 12 13 16 17 20 21 24 25 28 29 30 31

ECE C03 Lecture 15

13

More Advanced PAL Architectures

Registered PAL Architecture
Buffered Input or product term
CLK Q2 Q0 + Q2 Q0 OE

Q2 Q0 Q2 Q0

D2 DQ Q

Q2+

Q2+

Q2+ X Q2 Q0 + Q2 Q0

Q2 Q2

Q0 Q0

D2 = Q2 Q0 + Q2 Q0

Negative Logic Feedback

D1 = X Q2 Q1 Q0 + X Q2 + X Q0 + Q2 Q0 + Q1 Q0 D0 = Q0 Z = X Q1 + X Q1
ECE C03 Lecture 15 14

Advanced PAL Architectures

Programmable Output Polarity/XOR PALs
CLK OE

DQ Q

Buried Registers: decouple FF from the output pin

Advantage of XOR PALs: Parity and Arithmetic Operations

AB AB AB AB AB AB AB AB C C C C C C C C D D D D D D D D A B C D AB AB CD CD A B C D

ECE C03 Lecture 15

15

Examples of XOR and Registered PALs

Example of XOR PAL
INCREMEN T 1 0 FIRST FUSE NUMBER 0 40 80 120 Q 2
2

Example of Registered PAL

INCREMEN T 1 0 4 8 12 16 20 24 28

12

16

20

24

28

32

36 D Q 23
FIRST FUSE NUMBER S

0 32 64 96 128 160 192 224

19

160 200 240 280

Q Q

22

3 320 360 400 440

256 288 320 352 384 416 448 480 3 512 544 576 608 640 672 704 736 4

Q Q

18

Q Q

21

4 480 520 560 600

Q Q

17

Q Q

20

5 640 680 720 760

Q Q

19

768 800 832 864 896 928 960 992 5

Q Q

16

6 800 840 880 920

1024 1056 1088 1120 1152 1184 1216 1248 6

Q Q

18

Q Q

15

7 960 1000 1040 1080 Q 8 1120 1160 1200 1240 Q 9 1280 1320 1360 1400 Q 10 1440 1480 1520 1560 Q 11 INCREMEN T 0 4 8 12 16 20 24 28 32 36 13
9

17

1280 1312 1344 1376 1408 1440 1472 1504 7 1536 1568 1600 1632 1664 1696 1728 1760 8 1792 1824 1856 1888 1920 1952 1984 2016

Q Q

14

16

Q Q

13

15

12

14

11

NOTE: FUSE NUMBER = FIRST FUSE NUMBER + INCREMENT

ECE C03 Lecture 15

16

FSM Design With Counters

Synchronous Counters: CLR, LD, CNT

Four kinds of transitions for each state: (1) to State 0 (CLR)

0 no signals asserted LD m

CLR
(2) to next state in sequence (CNT) (3) to arbitrary next state (LD) (4) loop in current state

CNT n+1

Careful state assignment is needed to reflect basic sequencing of the counter

ECE C03 Lecture 15

17

Implementation Strategies
FSM Design with Counters Excess 3 Converter Revisited

Reset 0/1 1 0/1 1/0 0 1/0 4 0/0, 1/1 5 0/1 1/0 6 0/1
ECE C03 Lecture 15 18

Note the sequential nature of the state assignments

2 0/0, 1/1 3 0/0, 1/1

Implementation Strategies
FSM Design with Counters Excess 3 Converter
Inputs/Current Next State State X Q2 Q1 Q0 Q2+ Q1+ 0 0 0 0 0 0 0 0 0 1 0 1 0 0 1 0 0 1 0 0 1 1 0 0 0 1 0 0 1 0 0 1 0 1 0 1 0 1 1 0 0 0 0 1 1 1 X X 1 0 0 0 1 0 1 0 0 1 1 0 1 0 1 0 0 1 1 0 1 1 0 0 1 1 0 0 1 0 1 1 0 1 1 1 1 1 1 0 X X 1 1 1 1 X X Outputs Q0+ 1 0 1 0 1 1 0 X 0 1 1 0 1 0 X X Z CLR LD 1 1 1 1 1 1 0 1 1 0 0 X 1 1 1 0 1 0 1 0 X X X X 0 1 0 0 1 0 1 1 1 1 0 X 0 1 1 1 1 1 X X X X X X EN 1 1 1 X 1 X X X X X 1 X 1 1 X X C X X X X X 0 X X 1 1 X X X X X X B X X X X X 1 X X 0 0 X X X X X X A X X X X X 0 X X 0 1 X X X X X X

CLR signal dominates LD which dominates Count

ECE C03 Lecture 15 19

Implementation FSM With Counters

Excess 3 Converter .i 5 .i 5 Espresso Input File .o 7 .o 7 .ilb res x q2 q1 q0 .ilb res x q2 q1 q0 .ob z clr ld en c b a .ob z clr ld en c b a .p 17 .p 10 1---- -0----0-001 0101101 00000 1111---0-01 1000000 00001 1111---11-0 1000000 00010 0111--0-0-0 0101100 00011 00-----000- 1010000 00100 0111---0--0 0010000 00101 110-011 0-10- 0101011 00110 10------11- 1000000 00111 -------11-- 0010000 01000 010-100 -1-1- 1010000 01001 010-101 .e 01010 1111--Espresso Output File 01011 10----01100 1111--01101 0111--01110 ------01111 ------ECE C03 Lecture 15 20 .e

Implementing FSM with Counters

Excess 3 Converter Schematic
CL K 1 0 1 0 exce ss 3 PL A X Re set X Q2 Q1 Q0 Z \CLR \LD EN C B A 7 P 10 163 T RCO 15 2 CL K 6 D QD 11 5 C QC 12 4 B QB 13 3 A 14 QA 9 LOA D 1 CL R D Q C Q Z

Synchronous Output Register

ECE C03 Lecture 15 21

FSM Design with FPGAs

Programmable Logic Devices = PLD PALs, PLAs = 10 - 100 Gate Equivalents

Field Programmable Gate Arrays = FPGAs Altera MAX Family Actel Programmable Gate Array Xilinx Logical Cell Array 100 - 1000(s) of Gate Equivalents!

ECE C03 Lecture 15

22

Xilinx Logic Cell Arrays

CMOS Static RAM Technology: programmable on the fly! All personality elements connected into serial shift register Shift in string of 1's and 0's on power up
IOB IOB IOB IOB

IOB

General Chip Architecture: Logic Blocks (CLBs) IO Blocks (IOBs) Wiring Channels

IOB

CLB

CLB

Wiring Channels

IOB

CLB
ECE C03 Lecture 15

CLB
23

IOB

Xilinx LCA Architecture

Inputs: Tri-state enable bit to output input, output clocks Outputs: input bit Internal FFs for input & output paths Fast/Slow outputs 5 ns vs. 30 ns rise Pull-up used with unused IOBs
Enable Output Program Controlled Options OUT INV TS INV OUTPUT SOURCE SLEW RATE PASSIVE PULLUP Vcc

MUX Out D Q Output Buffer

PAD

Direct In Q Registered In

D TTL or CMOS Input Buffer R

Clocks

Global Reset

ECE C03 Lecture 15

24

Xilinx LCA Architecture

Configurable Logic Block: CLB

2 FFs Any function of 5 Variables Global Reset Clock, Clock Enb

Reset DIN Mux D RD Q CE

Q1 A B C D E

Mux

Combinational Function Generator Q2 G Mux D RD Q CE Mux Y

Clock

Mux

Independent DIN

Clock Enable

ECE C03 Lecture 15

25

Xilinx CLB Function Generator

CLB Function Generator

Q1 A B Mux Mux Mux Q2 Q1 A B C D E Mux Mux Mux

Function of 4 Variables Function of 4 Variables

Q1 A B Mux
Function of 5 Variables

C Mux D E Q2

C D E

Any function of 5 variables

ECE C03 Lecture 15

Q2 Two Independent Functions of 4 variables each 26

Xilinx CLB Function Generator

Q1 A B C D
Certain Limited Functions of 6 Variables

Mux Mux
Function of 4 Variables

E F Mux G

Q2 Q1 A B C D Q2 Mux Mux
Function of 4 Variables

ECE C03 Lecture 15

27

Xilinx Interconnect Architecture

Interconnect Direct Connections Global Long Line Horizontal/Vertical Long Lines Switching Matrix Connections
Horizontal Long Line DI CE A B X C CLB2 K Y E D R DI CE A B X C CLB3 K Y E D R Horizontal Long Line Switching Matrix Direct Connections DI CE A B X C CLB0 K Y E D R DI CE A B X C CLB1 K Y E D R

Vertical Long Lines

Global Long Line

ECE C03 Lecture 15

28

Implementing FSM with Xilinx LCA

Implementing the BCD to Excess 3 FSM Q2+ = Q2 Q0 + Q2 Q0 Q1+ = X Q2 Q1 Q0 + X Q2 Q0 + X Q2 Q0 + Q1 Q0

Q0+ = Q0
Z = Z Q1 + X Q1

No function more complex than 4 variables 4 FFs implies 2 CLBs

Synchronous Mealy Machine Global Reset to be used Place Q2+, Q0+ in once CLB Q1, Z in second CLB maximize use of direct & general purpose interconnections
ECE C03 Lecture 15 29

Implementing FSM with Xilinx LCA

Implementing the BCD to Excess 3 FSM
Clk Clk

CE CE DI B C Y K E D RES Q0 FG Q0 Q2 Q0 FG A X DI Q2 B C K E D RES CE X Q2 Q1 Q0 X Q1 A X FG Q1

Y FG

CLB1

CLB2

ECE C03 Lecture 15

30

Summary
Mapping FSM to random logic Mapping FSM to ROMS Mapping FSM to PLAs Mapping FSM to Programmable Logic Devices (Xilinx) NEXT LECTURE: VHDL Language READING: Dewey 11.2, 11.3, 11.4, 11.5, 11.6, 12.2, 12.2

ECE C03 Lecture 15

31