Digital Decode Correction Logic Preliminary design review PAYAL

  • Slides: 24
Download presentation
Digital Decode & Correction Logic Preliminary design review PAYAL DAVE

Digital Decode & Correction Logic Preliminary design review PAYAL DAVE

Outline • Introduction • Why Digital Decode & Error correction logic requires • Main

Outline • Introduction • Why Digital Decode & Error correction logic requires • Main Block diagram • Individual block explanation • Sims result • Conclusion

A 7 D 7 B 7 A 6 D 6 B 6 A 5

A 7 D 7 B 7 A 6 D 6 B 6 A 5 D 5 B 5

Need for Digital Decode & Error Correction • To eliminate the redundancy of ½

Need for Digital Decode & Error Correction • To eliminate the redundancy of ½ bit in each stage • Digital correction takes raw output data of the ADC as input and outputs the digital representation. • Error correction logic circuitry takes 14 bit input from ADC stages and eliminated the redundancy of ½ bit in each stage and gives 8 bit digital output. • Digital circuits are fast and take low power.

How it is done? • Requirements: § § Align ADC (comparator) decisions Correct ADC

How it is done? • Requirements: § § Align ADC (comparator) decisions Correct ADC output bit pattern Add digital data in a 1. 5 -bit fashion Output a 8 -bit digital word • Implementation: § Shift register using Delayed D flip flop § Ripple carry adder

Main Block Diagram A 7 B 7 A 6 B 6 A 5 B

Main Block Diagram A 7 B 7 A 6 B 6 A 5 B 5 A 4 B 4 A 3 B 3 A 2 B 2 A 1 B 1 CLK 2 SHIFT REGISTER Ripple Carry Adder CLK 2 Output register D 7 D 6 D 5 D 4 D 3 D 2 D 1 D 0

Shift register to align ADC decisions STG -1 CLK 1 DFF CLK 2 DFF

Shift register to align ADC decisions STG -1 CLK 1 DFF CLK 2 DFF STG -2 STG -4 STG -6 Shift register using D flipflop DFF DFF DFF CLK 1 CLK 2 STG -7 DFF CLK 1 CLK 2 STG -5 DFF CLK 1 CLK 2 STG -3 DFF DFF DFF RIPPLE CARRY ADDER DFF

SHIFT REGISTER TEST BENCH SHIFT REG

SHIFT REGISTER TEST BENCH SHIFT REG

ALIGNING OF 14 BITS WITHIN 31/2 CLK

ALIGNING OF 14 BITS WITHIN 31/2 CLK

Digital correction logic A 7 A 6 B 6 A 5 STAGE - 3

Digital correction logic A 7 A 6 B 6 A 5 STAGE - 3 B 5 B 4 A 4 Concept of Digital Correction + STAGE - 1 STAGE - 2 B 7 A 3 STAGE -4 B 3 A 2 D 7 D 6 D 5 D 4 D 3 * Z = D Y = B XOR C X = A + BC (CARRY BIT) * D 2 STAGE - 5 STAGE -6 B 2 A 1 B 1 D 1 STAGE 7 D 0 * A B C D Y Z + Mathematical analysis Of Digital Correction X

Logic Equations used for RCA • Total gate delay for RCA C 0 =

Logic Equations used for RCA • Total gate delay for RCA C 0 = A 2 + A 1 B 2 C 1 = B 3 C 0 + A 3 C 2 = B 4 C 1 + A 4 C 3 = B 5 C 2 + A 5 C 4 = B 6 C 3 + A 6 D 0 = B 1 D 1 = A 1 XOR B 2 D 2 = C 0 XOR B 3 D 3 = C 1 XOR B 4 D 4 = C 2 XOR B 5 D 5 = C 3 XOR B 6 D 6 = C 4 XOR B 7 D 7 = B 7 C 4 + A 7

RIPPLE CARRY ADDER CARRY BIT A + BC

RIPPLE CARRY ADDER CARRY BIT A + BC

SS With worst case for 4. 5 v & 85 C

SS With worst case for 4. 5 v & 85 C

The delay in SS with worst case is 6. 34 ns

The delay in SS with worst case is 6. 34 ns

SIMS RESULTS FOR DIFFERENT CORNERS FOR RCA 5. 5 v 0 27 5. 0

SIMS RESULTS FOR DIFFERENT CORNERS FOR RCA 5. 5 v 0 27 5. 0 v 85 0 27 4. 5 v 85 0 TT SS FS FF SF 4. 26 n 4. 63 n 5. 44 n 4. 94 n 3. 70 n 2. 69 n 27 85 3. 97 n 4. 65 n 5. 38 n 6. 34 n 4. 02 n 4. 73 n 2. 89 n 3. 36 n 3. 92 n 4. 60 n

ALIGNING OF 14 BITS WITHIN 31/2 CLK

ALIGNING OF 14 BITS WITHIN 31/2 CLK

Error Correction Logic A 7 B 7 A 6 B 6 A 5 B

Error Correction Logic A 7 B 7 A 6 B 6 A 5 B 5 A 6 B 6 A 4 B 4 A 3 B 3 A 2 B 2 A 1 B 1 01 01 D 7 D 6 D 5 D 4 D 3 D 2 D 1 D 0 01 1 1 1 1 01 01 10 000000 00 11 11 00000001 10 10 01111110 10 01 01111101

SHIFT REGISTER WITH RCA AND OUTPUT REGISTER SHIFT REGISTER RCA WITH OUTPUT REG

SHIFT REGISTER WITH RCA AND OUTPUT REGISTER SHIFT REGISTER RCA WITH OUTPUT REG

30 n 35 n Pulse width linear input to shift reg

30 n 35 n Pulse width linear input to shift reg

Conclusion • 14 bits coming from the comparator are aligning with the help of

Conclusion • 14 bits coming from the comparator are aligning with the help of the shift register within 41/2 clock cycle. • Ripple carry adder takes 6. 34 ns for worst case condition which is fast enough to get desired output • Error correction logic is working for different combination of input bits.