Skip to content

Digital Techniques: Logic, Registers, Counters and Comparators

CSIR NET electronics: Boolean algebra, K-maps, magnitude comparators, flip-flops, shift registers, ripple and synchronous counters, with practice.

  • 18 min read
  • Updated
On this page · 15 sections
  1. Prerequisites
  2. The Hook: Counting Photons One Clock at a Time
  3. Number Systems and Codes
  4. Boolean Algebra
  5. Combinational Building Blocks
  6. Sequential Logic: Latches and Flip-Flops
  7. Registers
  8. Counters
  9. Memory Tricks
  10. Don’t Confuse
  11. Common Mistakes
  12. Practice
  13. Quick Revision Table
  14. Key Takeaways
  15. What’s Next

Prerequisites

The Hook: Counting Photons One Clock at a Time

Number Systems and Codes

CodeRuleExample
Binarybase 2\( 1011_2 = 11 \)
Hexadecimalbase 16, groups of 4 bits\( 1011\,1110_2 = \text{BE}_{16} = 190 \)
BCDeach decimal digit in 4 bits\( 59 = 0101\,1001 \)
2’s complementinvert all bits and add 1\( -5 \) in 4 bits \( = 1011 \)
Gray codeadjacent values differ in one bit; \( g_i = b_i \oplus b_{i+1} \), MSB unchangedbinary 0110 → Gray 0101

An \( n \)-bit 2’s-complement number ranges from \( -2^{n-1} \) to \( 2^{n-1} - 1 \). Overflow in addition occurs when two numbers of the same sign give a result of the opposite sign. Gray code is used in shaft encoders because only one bit changes at each step, so a misread at a boundary is off by at most one count.

Boolean Algebra

LawAND formOR form
Identity\( A\cdot1 = A \)\( A + 0 = A \)
Null\( A\cdot0 = 0 \)\( A + 1 = 1 \)
Idempotent\( AA = A \)\( A + A = A \)
Complement\( A\bar A = 0 \)\( A + \bar A = 1 \)
Absorption\( A(A + B) = A \)\( A + AB = A \)
Redundancy\( A(\bar A + B) = AB \)\( A + \bar AB = A + B \)
De Morgan\( \overline{AB} = \bar A + \bar B \)\( \overline{A + B} = \bar A\bar B \)

Universal gates: NAND alone (or NOR alone) can build any function. NOT is a NAND with tied inputs; AND is NAND followed by NOT; OR is NAND of inverted inputs (De Morgan).

XOR \( A \oplus B = A\bar B + \bar AB \) is 1 when inputs differ; XNOR \( A \odot B \) is 1 when they are equal. XOR is the parity function and the sum bit of an adder; XNOR is the equality detector of a comparator.

Karnaugh maps

Arrange minterms in Gray-code order so that adjacent cells differ in one variable. Group 1s in rectangles of 1, 2, 4, 8 … cells (wrapping round the edges); each group of \( 2^k \) cells eliminates \( k \) variables. “Don’t care” cells may be included to enlarge groups.

Combinational Building Blocks

BlockFunctionKey fact
Half adder\( S = A \oplus B \), \( C = AB \)no carry in
Full adder\( S = A \oplus B \oplus C_{in} \), \( C_{out} = AB + C_{in}(A \oplus B) \)two half adders and an OR
Multiplexer (\( 2^n \):1)selects one of \( 2^n \) inputs with \( n \) select linescan realise any function of \( n \) variables (or \( n + 1 \) with the extra variable on data inputs)
Decoder (\( n \) to \( 2^n \))one output high per input codegenerates all minterms; memory chip selection
Encoder / priority encoderreverse of decoderpriority version reports highest active input
Parity generatorXOR chainodd or even parity bit for error detection

Digital comparators

A 1-bit comparator produces three outputs:

$$G = A\bar B \ (A > B), \qquad E = \overline{A \oplus B} = A \odot B \ (A = B), \qquad L = \bar AB \ (A < B)$$

For two 4-bit numbers \( A_3A_2A_1A_0 \) and \( B_3B_2B_1B_0 \), define \( x_i = A_i \odot B_i \) (bit \( i \) equal). Compare from the most significant bit down:

$$E = x_3x_2x_1x_0$$$$G = A_3\bar B_3 + x_3A_2\bar B_2 + x_3x_2A_1\bar B_1 + x_3x_2x_1A_0\bar B_0$$$$L = \bar A_3B_3 + x_3\bar A_2B_2 + x_3x_2\bar A_1B_1 + x_3x_2x_1\bar A_0B_0$$

The first unequal bit, scanning from the MSB, decides. Exactly one of \( G \), \( E \), \( L \) is 1. Comparators are cascaded for wider words by feeding the lower stage’s \( G \), \( E \), \( L \) into the next stage’s cascade inputs, which matter only when all of the higher bits are equal.

Do not confuse this digital magnitude comparator with the analogue voltage comparator (an op-amp without negative feedback), which decides whether one voltage exceeds another and is the front end of every flash ADC in A/D and D/A Converters.

Sequential Logic: Latches and Flip-Flops

A latch is level-sensitive (transparent while enabled); a flip-flop changes state only on a clock edge.

Flip-flopInputs → next state \( Q^+ \)Characteristic equation
SR00 hold, 01 reset, 10 set, 11 forbidden\( Q^+ = S + \bar RQ \), with \( SR = 0 \)
D\( Q^+ = D \)\( Q^+ = D \)
JK00 hold, 01 reset, 10 set, 11 toggle\( Q^+ = J\bar Q + \bar KQ \)
T0 hold, 1 toggle\( Q^+ = T \oplus Q \)

For design you need the JK excitation table (inputs required for a given transition; X = don’t care):

\( Q \to Q^+ \)\( J \)\( K \)
0 → 00X
0 → 11X
1 → 0X1
1 → 1X0

Race-around: a level-triggered JK with \( J = K = 1 \) toggles repeatedly while the clock is high if the clock pulse is longer than the propagation delay. Cures: a master–slave arrangement (master follows the inputs while the clock is high, slave copies the master when it falls) or edge triggering. Real flip-flops also need data stable for a set-up time before and a hold time after the active edge.

Registers

A register is a set of flip-flops sharing a clock, storing a word. Shift registers move the word one place per clock.

TypeLoadReadClocks to load \( n \) bits
SISO (serial in, serial out)serialserial\( n \) (and \( n \) more to read out)
SIPOserialparallel\( n \)
PISOparallelserial1 to load, \( n \) to shift out
PIPOparallelparallel1
Universal (bidirectional)eithereithermode select lines choose hold, shift left, shift right, load

Uses: serial–parallel conversion (UARTs, the serial port of a microcontroller), delay lines (an \( n \)-stage SISO delays data by \( n \) clock periods), and arithmetic: shifting a binary number left multiplies by 2, shifting right divides by 2 (with an arithmetic shift preserving the sign bit for 2’s-complement numbers).

Counters

Asynchronous (ripple) counters

Connect T flip-flops (or JKs with \( J = K = 1 \)) in a chain: the external clock drives the first, and each flip-flop’s output clocks the next. Each stage divides frequency by 2, so an \( n \)-stage ripple counter is mod-\( 2^n \) and its last output runs at \( f_{clk}/2^n \).

The weakness is accumulated delay: the last stage changes \( n\,t_{pd} \) after the clock edge. For all outputs to settle before the next edge,

$$\boxed{f_{clk,\max} = \frac{1}{n\,t_{pd}}}$$

During the ripple the outputs pass through transient wrong states, which cause glitches if decoded.

Mod-\( N \) by reset: use \( \lceil\log_2N\rceil \) flip-flops and a gate that detects the count \( N \) and clears all flip-flops. A decade (mod-10) counter detects 1010 (\( Q_3 \) and \( Q_1 \) both 1) and resets; the state 1010 exists momentarily.

Synchronous counters

All flip-flops share one clock; logic decides which ones toggle. For a binary up-counter with T flip-flops, stage \( i \) toggles when all lower bits are 1:

$$T_0 = 1, \quad T_1 = Q_0, \quad T_2 = Q_0Q_1, \quad T_3 = Q_0Q_1Q_2$$

All outputs change together (one \( t_{pd} \) after the edge), so there are no ripple glitches and the speed does not fall with \( n \). For a down counter, use \( \bar Q \) instead of \( Q \).

Designing a synchronous counter: write the state table, use the excitation table to find each flip-flop’s required inputs, simplify each with a K-map, and check unused states (a counter is self-starting if every unused state leads eventually into the main sequence).

Shift-register counters

CounterFeedbackStates with \( n \) flip-flopsRemarks
Ring counter\( Q_{last} \to D_{first} \)\( n \)one hot bit circulates; needs presetting to one 1
Johnson (twisted ring)\( \bar Q_{last} \to D_{first} \)\( 2n \)decoded with 2-input gates; glitch-free
Maximal-length LFSRXOR of chosen taps \( \to D_{first} \)\( 2^n - 1 \)all-zero state locks up; pseudo-random sequences
NeedFlip-flopsExample
Mod-\( N \) binary or decoded counter\( \lceil\log_2N\rceil \)mod-12 needs 4
Johnson mod-\( N \) (\( N \) even)\( N/2 \)mod-10 needs 5
Ring mod-\( N \)\( N \)mod-10 needs 10

Frequency division

Counters in cascade multiply their moduli: a 1 MHz clock through a mod-10 counter and then a mod-8 counter gives 1 MHz / 80 = 12.5 kHz at the last output. A mod-\( N \) binary counter’s MSB is not a square wave unless \( N \) is a power of 2: in a decade counter \( Q_3 \) is high only for counts 8 and 9, a 20% duty cycle.

graph LR
    A[Clock 1 MHz] --> B[Mod-10 counter]
    B -->|100 kHz| C[Mod-8 counter]
    C -->|12.5 kHz| D[Register latches count]
    D --> E[Magnitude comparator]
    E -->|count above threshold| F[Flag]

Memory Tricks

Don’t Confuse

ThisWith thisThe difference
LatchFlip-floplevel-sensitive, transparent while enabled; edge-triggered
Ripple counterSynchronous counterclock passes stage to stage, delays add, glitches; common clock, outputs change together
Ring counterJohnson counter\( Q \) fed back, \( n \) states; \( \bar Q \) fed back, \( 2n \) states
Magnitude comparatorVoltage comparatordigital words, outputs \( G, E, L \); analogue voltages, op-amp without feedback
SR forbidden stateJK toggle\( S = R = 1 \) undefined; \( J = K = 1 \) toggles
Mod-\( N \)number of flip-flopsnumber of states in the cycle; \( \lceil\log_2N\rceil \) for a binary counter

Common Mistakes

Practice

All questions below are original practice questions written for this page. They are not from any CSIR NET paper.

Level 1: Foundation

  1. How many flip-flops does a mod-12 binary counter need?
  2. How many states does a Johnson counter built from 4 flip-flops have?
  3. Write \( -6 \) in 4-bit 2’s complement.
Answers
  1. 4. \( 2^3 = 8 < 12 \leq 16 = 2^4 \).
  2. 8 (\( 2n \)).
  3. 1010. \( 6 = 0110 \); invert to 1001; add 1: 1010.

Level 2: Part B Style

Practice Q1. The function \( f(A, B, C, D) = \sum m(0, 2, 5, 7, 8, 10, 13, 15) \) simplifies to:

(a) \( A \oplus C \) (b) \( B \odot D \) (c) \( \bar B\bar D \) (d) \( B \oplus D \)

Answer: (b) B ⊙ D

Corners (0, 2, 8, 10) give \( \bar B\bar D \); the central square (5, 7, 13, 15) gives \( BD \). Sum: \( \bar B\bar D + BD = B \odot D \). (c) is only one of the two groups; (d) is the complement.

Practice Q2. A 4-bit ripple counter uses flip-flops with propagation delay 10 ns each. If all outputs must settle within one clock period, the maximum clock frequency is:

(a) 100 MHz (b) 50 MHz (c) 25 MHz (d) 6.25 MHz

Answer: (c) 25 MHz

Worst-case settling \( = 4 \times 10 = 40 \) ns, so \( f_{\max} = 1/40\ \text{ns} = 25 \) MHz. (a) is the limit for a synchronous counter (one delay); (d) divides by \( 2^4 \) as well.

Practice Q3. A 1 MHz clock drives a mod-10 counter whose most significant output drives a mod-8 counter. The frequency at the most significant output of the second counter is:

(a) 125 kHz (b) 12.5 kHz (c) 1.25 kHz (d) 55.6 kHz

Answer: (b) 12.5 kHz

Total division \( 10 \times 8 = 80 \): \( 10^6/80 = 12.5 \) kHz. (d) adds the moduli (\( 10^6/18 \)).

Practice Q4. A 4-bit serial-in parallel-out shift register \( Q_AQ_BQ_CQ_D \) is cleared to 0000. Data enter at \( Q_A \) and shift towards \( Q_D \). The serial bits 1, 1, 0, 1 are applied in that order, one per clock. After the fourth clock, \( Q_AQ_BQ_CQ_D \) is:

(a) 1101 (b) 1011 (c) 0111 (d) 1110

Answer: (b) 1011
ClockInput\( Q_AQ_BQ_CQ_D \)
111000
211100
300110
411011

The first bit entered has reached \( Q_D \). (a) is the input order, the classic reversal trap.

Level 3: Part C Style

Practice Q5. A synchronous counter uses three JK flip-flops \( A \) (MSB), \( B \), \( C \) (LSB) with

$$J_A = BC,\ K_A = 1; \qquad J_B = C,\ K_B = C + A; \qquad J_C = \bar A,\ K_C = 1$$

Consider:

  1. Starting from 000, the counter goes 0, 1, 2, 3, 4, 0, …
  2. The counter is self-starting.
  3. From the unused state 5 (101) the next state is 6 (110).
  4. Output \( A \) has frequency \( f_{clk}/5 \) and is high for one clock period in five.

Which are correct?

(a) 1 and 2 only (b) 1, 2 and 4 only (c) 1, 3 and 4 only (d) 1, 2, 3 and 4

Answer: (b) 1, 2 and 4 only

Use \( Q^+ = J\bar Q + \bar KQ \).

\( ABC \)\( J_AK_A \)\( J_BK_B \)\( J_CK_C \)Next
0000 10 01 1001
0010 11 11 1010
0100 10 01 1011
0111 11 11 1100
1000 10 10 1000
1010 11 10 1010
1100 10 10 1000
1111 11 10 1000
  1. True: a mod-5 sequence. 2. True: 5 → 2, 6 → 0, 7 → 0 all enter the cycle. 3. False: 5 goes to 2. 4. True: \( A = 1 \) only in state 4.

Practice Q6. A 3-bit shift register \( Q_2Q_1Q_0 \) shifts \( Q_0 \to Q_1 \to Q_2 \) on each clock, with new \( Q_0 = Q_2 \oplus Q_1 \). It starts at 001. Consider:

  1. The sequence is 001, 010, 101, 011, 111, 110, 100, 001, …
  2. The sequence length is 7.
  3. If the register is ever cleared to 000 it stays there.
  4. Replacing the XOR by XNOR would make 111 the lock-up state.

Which are correct?

(a) 1 and 2 only (b) 1, 2 and 3 only (c) 2, 3 and 4 only (d) 1, 2, 3 and 4

Answer: (d) 1, 2, 3 and 4

1–2. From 001: new \( Q_0 = 0 \oplus 0 = 0 \), shift gives 010; then \( 0 \oplus 1 = 1 \): 101; \( 1 \oplus 0 = 1 \): 011; \( 0 \oplus 1 = 1 \): 111; \( 1 \oplus 1 = 0 \): 110; \( 1 \oplus 1 = 0 \): 100; \( 1 \oplus 0 = 1 \): 001. Seven distinct states, the maximal \( 2^3 - 1 \). True. 3. \( 0 \oplus 0 = 0 \): 000 maps to itself. True. 4. With XNOR, 111 gives \( 1 \odot 1 = 1 \) and maps to itself. True.

Practice Q7. A decade ripple counter is made from four negative-edge-triggered flip-flops \( Q_3Q_2Q_1Q_0 \), with a NAND gate on \( Q_3 \) and \( Q_1 \) driving the asynchronous clear of all four. A 50 kHz clock is applied. Consider:

  1. The counter cycles through 0000 to 1001.
  2. The state 1010 appears only momentarily.
  3. \( Q_3 \) is a 5 kHz square wave with 50% duty cycle.
  4. Decoding the count 0000 with a gate may produce narrow glitches.

Which are correct?

(a) 1 and 2 only (b) 1, 2 and 4 only (c) 1, 3 and 4 only (d) 1, 2, 3 and 4

Answer: (b) 1, 2 and 4 only
  1. On reaching 1010 (\( Q_3 = Q_1 = 1 \)) the NAND output goes low and clears the counter: the stable states are 0 to 9. True.
  2. 1010 exists only until the clear propagates. True.
  3. \( Q_3 \) has frequency \( 50/10 = 5 \) kHz, but it is high only for counts 8 and 9: duty cycle 20%. False.
  4. Ripple transitions and the brief 1010 state pass through intermediate codes; decoded outputs can glitch. True.

Quick Revision Table

IdeaFormula or fact
2’s complement range\( -2^{n-1} \) to \( 2^{n-1} - 1 \)
Gray code\( g_i = b_i \oplus b_{i+1} \), one bit changes per step
De Morgan\( \overline{AB} = \bar A + \bar B \), \( \overline{A + B} = \bar A\bar B \)
Universal gatesNAND, NOR
1-bit comparator\( G = A\bar B \), \( E = A \odot B \), \( L = \bar AB \)
4-bit equality\( E = x_3x_2x_1x_0 \), \( x_i = A_i \odot B_i \)
JK\( Q^+ = J\bar Q + \bar KQ \); 11 toggles
D, T\( Q^+ = D \); \( Q^+ = T \oplus Q \)
Race-aroundcured by master–slave or edge triggering
Shift register\( n \) clocks to load serially; left shift ×2
Ripple countermod-\( 2^n \); \( f_{\max} = 1/nt_{pd} \)
Mod-\( N \)\( \lceil\log_2N\rceil \) flip-flops
Ring / Johnson / LFSR\( n \) / \( 2n \) / \( 2^n - 1 \) states
Cascademoduli multiply

Key Takeaways

What’s Next