Thursday, 11 October 2012

Digital logic - Sequential Circuits, by Goh Hooi Kuan


Sequential Circuits

4.1            Introduction

Combination circuits whose outputs are function of only the inputs were discussed. However, in many digital designs there is a need for logic circuits whose outputs depend not only on the present inputs, but also on the past values of the input and output, of the circuit.
In other words, the concept of time has to be introduced, and memory circuits are required to store information about the past history of the circuit. Such circuits are known as sequential logic circuits, because they follow a predetermined sequence. There are numerous examples of sequential circuits, one of the most widely met being the traffic light controller, which may be regarded as a black box, as shown in Figure 4.1.



There are 3 outputs, labeled Red, Yellow and Green, to drive the 3 traffic lights, and a single input, labeled clock. A light will be turned on if there is a 1 on its output line. Fundamental to the design of sequential circuits is the concept of internal states.
The simple traffic light controller requires 4 states, labeled S0, S1, S2 and S3, as shown in Figure 4.2. The lines drawn from 1 state to another show the transition between states. The input signal (or input signal combinations) that gives rise to each transition is indicated above the arrow. Each transition between states is initiated by a clock signal, clk, as there are no other input signals in this simple example. Those outputs that are turned on in a given state are shown either alongside, or inside the state circles. For example, in state S1, Red and Yellow are on (1), while Green is off (0).


                    

4.2 Synchronous and Asynchronous Sequential Circuits

There are 2 fundamentally different types of sequential circuit, namely synchronous and asynchronous. In a asynchronous (or event driven) circuit there is no clock, and transitions between states are initiated by changes in the appropriate input signals. In a synchronous circuit, transitions between states are initiated by a pulse from a single clock. It is normally assumed that any changes in the input signals occur between clock pulses, so that the input signals are stable when a clock pulse is applied. Synchronous circuits are easier to design, and so this text will be restricted to them.
Figure 4.3 shows 2 typical clock signals. The first is a repetitive signal, derived from an oscillator circuit, and has a constant frequency, f. At any instant in time, its value is either 0 or VH where VH is the voltage that corresponds to a 1 for the being used to implement the sequential design (typically 5 volts). The period T of this clock signal is related to its frequency by T = 1/f. The second clock signal although repetitive, does not have a constant frequency. It is often referred to as a pulsed clock, the pulses being generated in response to a change in one or more of the input signals. A synchronous circuit is designed to change state either when the clock goes from low to high, a positive edge transition, or when the clock goes from high to low, a negative edge transition. The arrows in Figure 4.3 indicate the positive transitions of the clock.



4.3 D-type flip-flops

There are 4 states, labeled S0 to S3, 2 state variables are required, X and Y, say.
Each state variable is implemented with a D-type flip-flop, so in this example 2 flip-flops are used. Both the inputs and the outputs of the flip-flops are labeled with the suffixes X and Y, respectively. The clock inputs are connected together, and are driven from a single clock source, There are now 2 remaining problems, both involving combinational logic, namely the interconnections required to the D inputs of each flip-flop, and the logic required to generate the Red, Green and yellow output signals from the state variables X and Y. The interconnections required to the Dinputs of the flip-



flops are found by slightly modifying the state table. This table for the traffic light controller is shown in Table 4.7. The states are now shown by their state variables; the assignment of each variable combination to a given state is up to the designer but it is simplest, at least initially, to follow the normal binary sequence. Note, however, that the choice of state variables can alter the amount of hardware require to implement the necessary combinational logic; other assignments may result in less hardware.





This equation shows that the next output, or state, of the flip-flop is just equal to the present value of the D input. Thus the entries in Table 4.7 for the next state are just those required for the D inputs, Dx and Dy.
Scanning down the column labelled Y’ shows that there are only two 1 entries, corresponding to the present states S1 (Y’.X) and S2 (Y,X’), so that :
DY  =  Y’ . X + Y .X’ = X+ Y
Similarly for the X’ column there are also two 1 entries corresponding to states S0 and S2, and so :
DX  =  Y’ . X + Y .X’ = X’
Figure 4.12 shows the two D-type flip-flops with the logic required to generates DX and DY. Finally the logic required to generate the output signals must be determined. Returning to the state diagram, it will be seen that the following output signals are required.


4.4  Register

            A register is a device in which a number of binary digits (bits) can be stored and retrieved, or read back. The use of a D-type flip-flop to store a single bit of information. The data bit to be stored in, or written to, the memory (0,1) is placed on the D input and a clock pulse is then applied to the clock input. This results in the data bit being stored in the flip-flop. Its value can be ascertained at any time from the memory by reading the value of the Q output. A set of n of these D-type flip-flops, fabricated as a single unit, can be used to store n bits of information, and is known as a register. Frequently, n is a power of 2, usually 4, 8, 16, 32 or 64. Figure 4.22 is a schematic diagram of a single 4-bit register.
The inputs are labeled D0 to D3, and the outputs Q0 to Q3. The pattern of bits stored in the register will be referred to as a (4-bit) word. The 4 clock inputs to the individual flip-flops within the register are connected to a common clock line. When a pulse is applied to this line the data on the output lines are written into the memory, over-writing the previous word stored.



4.5  Counter

            Counters are a special class of finite state machine. A counter follows a preset sequence and repeatedly outputs the same pattern. The simplest counter is the binary-up counter in which the counter is incremented by one each time a clock pulse occurs, until the maximum count is reached. Figure 4.19 shows the state diagram for a modulo-8 counter.



By convention the count begins at 0 and is then incremented by 1 on the next clock pulse until the maximum of 7 occurs. When the next clock pulse occurs the counter is reset to zero, and the sequence starts again. There are no inputs, apart from the clock. Eight states require 3 state variables, X, Y and Z. Where possible the state assignments are made so that the circuit outputs are just the state variables, X, Y and Z. Thus, S0 = 000, S1 = 001 etc.
The design of a synchronous counter proceeds in exactly the same way as discussed above for any synchronous sequential circuit. For example, consider the design of a modulo-3 up/down synchronous counter. This counter has a single input labelled up/down, in addition to the normal clock. If up/down = 1 the counter is to count up in the normal way, following the sequence 0, 1, 2, 0, …, while if up/down = 0 it is required to count down following the sequence 0, 1, 2, 0, … There are 3 states, so two state variables are required, X and Y. In normal operation the 4th state, state 4, is never entered and so could be regarded as a ‘don’t care’ state. However, if state is ever entered (Eg : as the random state entered when power is first switched on) then the next state should be one of the 3 states, state 0 say, so that the count sequence is properly resumed.



In this example it will be assumed that each of the state variables, X and Y, is stored in a JK flip-flop, as shown in Figure 4.20. As before, the problem comes down to designing the combinational logic required for the connections to the J and k inputs of the 2 flip-flops, and the logic to generate the output signals. The logic is determined by expanding the state table into a new modified table, called an excitation table, Table 4.10 (b). Again the normal binary pattern has been assumed for the state variables. The column labelled JK flip-flop controls is filled in with the aid of the JK flip-flop state change table, table 4.6.




The decision on whether to use D-type or JK flip-flops is really a matter of designer choice. A JK flip-flop implementation is slightly more difficult to design, but will often result in less hardware being required for the combinational logic, than an equivalent design using D-type flip-flops.
Another type of counter that is frequently used is the asynchronous or ripple counter. These counters use JK flip-flops connected in the toggle mode, with J = K = 1.Figure 4.21 (a) shows a modulo-16 ripple counter. The pulses to be counted are applied to the clock input of the first JK flip-flop. The output of this flip-flop, QA, is applied to the clock input of the second flip-flop. Its output is applied in turn to the clock input of the third flip-flop, and so on.



The first flip-flop toggles, that is, changes state, on the rising edge of each of the pulses to be counted. It effectively divides the pulse frequency by 2. The second flip-flop toggles on every rising edge of QA, so effectively dividing the input pulse frequency by 4. The outputs of the flip-flops are thus a binary count of the input pulses, with QA being the least significant digit, and QD the most significant digit. Clearly the design can be extended to any number of stages to produce a modulo-N counter, but with the limitation that N is a power of two. It is also possible to design ripple counters where n is not a power of two, but their design is not considered here. Note also that a modulo-N counter may also be used as a divide-by-N frequency divider, dividing the input frequency by N, N2, N4, etc. The name ‘ripple counter’ comes from the fact that changes in the state of the counter ripple through the chain of flip-flops, with the result that there is a propagation delay between the outputs of the flip-flops changing. If this is important, then a synchronous counter must be used. In such a counter, all the changes occur instantaneously coincident with the clock pulse.


Written by,
Goh Hooi Kuan
B031210043

No comments:

Post a Comment