Microelectronics and Solid State Electronics 2016, 5(1): 7-13

DOI: 10.5923/j.msse.20160501.02

Design and Analysis of Low Power Universal Line

Encoder & Decoder
Anjali Taya, Balwinder Singh*, Hitesh Pahuja

ACS Division, Center for Development of Advanced Computing, Mohali, India

Abstract Communication plays an important role in day to day life. The information or data is transmitted through
various techniques and line coding is one of the finest techniques for sending data. The selection of these techniques
depends on the bandwidth requirement, DC level, bit error rate performance and the inbuilt error detection property. In this
line coding techniques whose encoder and decoder have been designed and analyzed are Unipolar RZ, NRZ-I, NRZ-L,
Manchester, Differential Manchester, AMI, Pseudoternary, B8ZS, HDB3 coding. Any one of these techniques can be
access with the mode of selection. Switching activity is the one of the main factors that is responsible for the dynamic
power dissipation. Power consumption by the encoders and decoders is directly proportional to switching activity. To
optimize the power of the universal line encoder – decoder, Bus Shift (BS) coding scheme is applied that circularly shifts
the data to minimize the transition. Simulation results show the average saving margin of power in universal encoder is
22% while in universal decoder saving margin is 35%.
Keywords NRZ-I (Non Return To Zero Invert), NRZ-L (non return to zero level), AMI (alternate mark inversion),
Manchester formats, pseudo ternary encoded format, B8ZS (bipolar eight zero substitution) and HDB3 (high density
bipolar zero)

density, low probability of error and transparency etc.

1. Introduction Present applications of line codes are astronomic in data
communication networks and storage of information
Digital communication is the transmitting of data over a systems [5].
physical transmission media such as twisted pair, fibre optic
link and coaxial cable etc, in the form of digital signals. The
process of sending the data one bit at time, sequentially 2. Related Works
over a communication channel is known as serial Mohammed Alamgir et. al [1] purposed bus shift (BS)
communication. The data maintenance is required as it is coding scheme that circularly shifts the data to minimize
transmitted through echo chancellors, repeaters and another transition. Power saving of the bus invert is poor in average
electronically equipment. Data reconstruction is required cases while BS scheme better in both maximum and average
time to time for maintained the data integrity. Transmission cases of power save 14% for a 32 bit bus.
coding (also known as line coding) operation used in high Shankaranarayana Bhat et. al [3] purposed a technique for
speed systems for digital transmission of telephony and bus encoding which reduce number of transition on data bus,
television. The propose modifications into the signal which acts better than the current methods such as bus invert coding
ease synchronization, error detection etc [10]. and shift invert coding for random data in terms of switching
Converting a sequence of 1’s and 0’s to a time domain activity. Here reduction in switching activity irrespective of
signal suitable for communication over a channel is known bus width and more power efficient.
as line coding. The main goal of the line code is to Amrinder Kaur et. al [5] describes the implementation of
minimize noise effects on the information existing on the various line encoders and decoders using VERILOG on a
channel by finding the suitable form of the transmitted single chip.
information signal. The primary factors should be Anjali.v et. al [6] in this the implementation of Manchester
considered when choosing the line code: self encoder and decoder circuit along with irrational detection
synchronization, transmission power, power spectral and clock recovery unit (CRU) using HDL. This encoding
technique is broadly used in fields like biomedical
* Corresponding author:
[email protected] (Balwinder Singh)
applications, satellite communication etc.
Published online at http://journal.sapub.org/msse Wei Dang [4] purposed the rules based on the HDB3
Copyright © 2016 Scientific & Academic Publishing. All Rights Reserved encoder and decoder. For the detection of four zero’s
8 Anjali Taya et al.: Design and Analysis of Low Power Universal Line Encoder & Decoder

sequence in source information, there are four modules in HDB3 (High density bipolar three): In this encoded format
encoder. For the generator of “V” and generator of “B”, each string of four zeros are represented according to the odd
obeying the alternate mark inversion (AMI) rule. Simulated and even parities define in table 1 [12].
output of HDB3 encoder & decoder proved that results are
identical with theory analysis. 3.1. Encoding Algorithms
Balwinder singh et. al [13] In this paper read/write cycles The signed binary characters are used to implement the
are required for memory testing. The result shows the power design. By using that, zero voltage level represented by 00,
reduction of 30% and 74.8% in data bus and address buses while high (positive) voltage level represented by 01 and
encoding schemes. low voltage (negative) level is represented by 11. Therefore
+1 =00
-1 = 11
3. Line Coding Techniques 0 = 00
Line coding techniques can be widely divided into Table 2. Algorithms
following different classes:
OUTPUT(y0) OUTPUT(y1)
Unipolar RZ: In this encoded format binary 1 is S.NO. TECHNIQUE INPUT
(0 - t/2) (t/2 – t)
represented by 01 for first half interval and it return to 00 for 0 00 00
second half interval. Binary 0 is represented by 00 for both 1 Unipolar RZ
1 01 00
intervals. 0 11 01
Unipolar NRZ-I: In this encoded format binary 1 is 2 Manchester
1 01 11
represented by high level and for next bit position it
represented by low level. Binary 1 is alternating bit by bit. The algorithm for Manchester and Unipolar RZ encoding
Binary 0 represented by same level as binary 1. technique is given in table 2. The differential Manchester
Unipolar NRZ-L: In this encoded format binary 1 is algorithm is given in table 3. The truth table for AMI
represented by low level and for binary 0 is represented by encoding technique is shown in table 4, and for
high level. Pseudoternary encoding is shown in table 5.
Manchester Encoded: In this binary 1 is represented by 01
Table 3. Differential Manchester
(high positive level) for first half bit interval and 11 (low
negative level) for second half bit interval. Binary 0 is Present OUTPUT(y0) OUTPUT(y1) Next
INPUT
represented by 11 (low negative level) for first half bit state (0 - t/2) (t/2 - t) state
interval and 01 (high positive level) for second half bit 0 1 01 11 1
interval [3]. 0 0 11 01 0
Differential Manchester Encoded: In this binary 1 is 1 1 11 01 0
represented by 01 (high positive level) for first half and 11 1 0 01 11 1
(low negative level) for second half of bit interval. For next
bit 1 is represented by 11 (low negative level) first half and Table 4. AMI Truth table
01 (high positive level) second half. Binary 1 is alternating
PRESENT STATE INPUT OUTPUT NEXT STATE
bit by bit. Binary 0 is represented by same level as binary 1.
0 0 00 0
AMI (alternate mark inversion): In this encoded format bit
1 is represented by 01 and 11 alternatively for bit duration 0 1 01 1
while bit 0 is represented by 00 for whole bit interval. 1 0 00 1
Pseudo ternary Encoded: In this encoded format bit 1 is 1 1 11 0
represented by 00 for whole bit interval while bit 0 is
represented by 01 and 11 alternatively for bit duration. Table 5. Pseudoternary truth table
B8ZS (binary eight zero sequence): In this encoded format PRESENT STATE INPUT OUTPUT NEXT STATE
each string of eight zeros are represented by 000+-0+- if it 0 0 00 0
comes after positive pulse while 000-+0+- if it comes after 0 1 01 1
negative pulse.
1 0 11 1
Table 1. HDB3 pattern 1 1 00 0
Parity of -/+ bits since
Pattern Previous pulse Coded 3.2. Decoding Algorithm
previous V
+ 000+ The decoding rules for the line coding techniques are
ODD 000V
- 000- reverse from the encoding algorithms. The rules for
+ -00- decoding the line coded scheme differ for each technique
EVEN B00V defined below.
- +00+
 Microelectronics and Solid State Electronics 2016, 5(1): 7-13 9

Unipolar RZ: IN this decoding format if input is 01 then CL = load capacitance

it decoded the signal as 1 and if observe 00 then it decoded
signal as 0.
NRZ-I: The conversion from the previous voltage level 5. Universal Line Encoder & Decoder
would be decoded as bit 1 while no shift or no conversion with Low Power Technique
would be decoded as bit 0.
NRZ-L: In this high voltage level decoded as bit 0 and In the universal line encoder-decoder by using structural
low voltage level decoded as bit 1. style modelling we have integrated all the 9 line coding
Manchester decoded: In this at the positive edge if the technique together. In this we use the four bit select line for
input is 01 then it would be decode as bit 1 and if it detect the selection of 9 line coding technique from 0000 to 1001.
11 then it would be decode as bit 0. Given user data gets converted in the bit format to the desired
Differential Manchester decoded: In this if there is line coding and additionally at the receiver end, converted
transition occur in the input level from the previous bit then it back line coding signal to bit format [10].
decoded as bit 0 and if there is no transition occur then it Table 6. Block selection table
decoded as bit 1.
AMI decoded format: In this if it detect either 01 or 11 VALUE OF SEL[3:0] CORRESPONDING ENCODER
then it would be decode as bit 1 while if it detect 00 then it SIGNAL DECODER BLOCK SELECTED
would be decode as bit 0. 0000 UNIPOLAR RZ
Pseudoternary decoded format: In this if 00 is detected 0001 NRZ-L
then it would be the decode as bit 1 while if it detect either 01
0010 NRZ-I
or 11 then it would be decode as bit 0.
B8ZS decoded format: In this if the sequence of 000+-0+- 0011 MANCHASTER
or 000-+0+- is detect it would be decode as sequence of eight 0100 DIFFERENTIAL MANCHASTER
zeros.
0101 AMI
HDB3 decoded format: In this format according to the odd
and even parities, pattern generated 000v or b00v it would be 0110 PSEUDOTERNARY
decode as sequence of four zeros [4]. 0111 B8ZS
1001 HDB3

4. Bus Shift Encoding Technique In this universal line encoder & decoder by using the bus
In this shift the data word on a bus to get minimum shift coding technique power has been reduced. This
transition and also it reduce the power. For simplicity, we use technique based on the hamming distance between the
only right shift but in circular manner. Given presentdata or presentdata and nextdata. In the universal encoding method
nextdata, to determine how many bits of nextdata to shift, do least number of transitions is used to encode the data and
the following: transmit it over the bus. The universal decoder has ability to
coded back signal to bit format at receiver end by using least
1. For j = 1 to m, shift nextdata to right by j-bit and transitions. The optimized universal encoder & decoder has
determine the hamming distance between the low power, high speed, low complexity, and useful for all
presentdata and shifted word. coding techniques.
2. Let us assume HD(min) is the number of bits that result
in minimum hamming distance.
3. Shift the nextdata by HD(min) bits and send the shifted 6. Result and Discussion
word [1].
Example of bus shift ISIM simulator is used for the analysis of universal line
On a 10 bit bus, let us assume presentdata = 1011001001 Encoder & Decoder as shown in Figs 1-3. In this sel [3:0] pin
and the nextdata = 0011011001; here HD(min) between the is used to change line coding technique.
words is 3. The Bus shift will find that on the right shifted Figure 1 shows the results for universal line encoder when
nextdata by 7 bits to get 1011001001 results in zero the sel[3:0] = 0000, due to this value Unipolar RZ is enabled
hamming distance as compare to presentdata [1]. and get the results at the signal z[1:0].
After that power reduced on the bus, is calculated by In the Fig.2 values of signal sel[3:0] = 0110 change and
applying all the coding techniques [3]. To determine the get the output of Pseudoternary encoder block.
power dissipation is given by: Figure 3 shows the results of universal line decoder when
the sel[3:0] = 0001, due to this value NRZ-I is enabled and
Pd = α CL V2DD f (1) get the results at the output signal z[1:0]. Similarly by
α = Switching activity changing the values of the sel[3:0] signal, then the selected
VDD = Supply Voltage blocks also change the corresponding values on sel[3:0]
f = Frequency signal and get outputs at selected blocks on z[1:0] signal.
10 Anjali Taya et al.: Design and Analysis of Low Power Universal Line Encoder & Decoder

The power analysis of optimized Universal line encoder & saving margin of 22% in average cases of universal line
decoder by compared with the conventional Universal line encoder while the saving margin of 35% in average cases of
encoder & decoder. Performance analysis shows the power universal line decoder by using bus shift coding technique.

Figure 1. Universal encoder when sel[3:0] = 0000

Figure 2. Universal encoder when sel[3:0] = 0110

 Microelectronics and Solid State Electronics 2016, 5(1): 7-13 11

Figure 3. Universal decoder when sel[3:0] = 0001

In the Figure 4 given below, this graph shows the power Table 7. Encoder Power analysis
comparison between the different line coding techniques of
Power
encoders and optimized encoders as given in Table 7. Encoder Optimal Encoder
Technique

700 RZ 413 380

600 NRZ-L 405 352
500 NRZ-I 396 360
400
300 Encoder Manchester 421 376

200 Power(mW) Differential Manchester 556 392

100 AMI 457 414
0 Pseudoternary 417 626
Optimal
Differential…

Pseudoternary
AMI

B8ZS 425 417

B8ZS
NRZ-I
RZ

HDB3
Mamchester
NRZ-L

Encoder
Power(mW) HDB3 440 430
Universal Encoder 6604 5114

In Figure 5 this graph shows the power comparison

between the various line coding techniques of decoders and
optimized decoders by using the Table 8.
Figure 4. Encoder Power (mw) graph
12 Anjali Taya et al.: Design and Analysis of Low Power Universal Line Encoder & Decoder

7. Conclusions
3000
2500 In this research work, universal line encoder & decoder
2000 has been designed and analyzed by using the various line
1500 Decoder
1000 Power(mW) coding techniques. After that binary shift coding technique is
500 used to reduce dynamic power dissipation in data bus by
0 circularly shifts the bits to minimize the transition.
Optimal Optimized universal encoder & decoder were compared with
Decoder the conventional encoder & decoder, which shows the power
Power(mW) saving margin of 22% in case of encoder while 35% in case
of decoder. The design has low complexity, high speed, low
power and efficient for all the encoding and decoding
schemes.
Figure 5. Decoder Power (mw) graph

Table 8. Decoder Power analysis

Power (mW) Optimal

Technique
Decoder
Decoder REFERENCES
RZ 681 2588 [1] M. Alamgir, I. I. Basith, T. Supon and R. Rashidzadeh,
NRZ-L 325 352 "Improved bus-shift coding for low-power I/O," 2015 IEEE
International Symposium on Circuits and Systems (ISCAS),
NRZ-I 325 322
Lisbon, 2015, pp. 2940-2943.
Manchester 2358 955
[2] Abdelrahman H. Elsayed, Ramy N. Tadros, Maged
Differential Manchester 1028 915
Ghonemia, yeheaIsmail “Low-Power All-Digital
AMI 396 370 Manchester-Encoding- Based High-Speed Transceiver for
Pseudoternary 410 1190 On-Chip Networks”, IEEE Conference Publications, Pages:
2752 - 2755, 2014.
B8ZS 356 347
HDB3 378 375 [3] Shankaranarayana Bhat M and D. Yogitha Jahnavi
“Universal rotate invert bus encoding for low power VLSI”,
Universal Decoder 2508 1624
International Journal of VLSI design & Communication
Systems (VLSICS) Vol.3, No.4, pp.97-109, August 2012.
The final result of power is shown by the universal
encoder & decoder as shown in figure 6, where we analyze [4] Wei Dang, "Implement of HDB3 coder in base-band system,"
that power is saved 22-33% by using the bus shift technique. in Consumer Electronics, Communications and Networks
Delay: For the both universal encoder and decoder speed (CECNet), 2012 2nd International Conference, pp.3273-3274,
21-23 April 2012.
grade is -3 and minimum period required or delay for the
encoder is 2.556ns and for universal decoder 1.129 ns. [5] Amrinder Kaur, Mandeep Singh, Balwinder Singh “VHDL
Optimized universal encoder and decoder has better Implementation of Universal Encoder for communication”,
ISP Journal of Electronic Engineering, Vol.1, Issue.2, ISSN
minimum period requirement as compare to conventional
2250-0537(online), pp. (37-41). Dec 2011.
encoder decoder. Speed grade is -3 and the delay for the
optimized encoder is 2.153ns and for the optimized decoder [6] Anjali. V, P. Satishkumar, "Manchester Encoder and Decoder
is 0.766ns. with Clock Recovery Unit and Invalid Detector”, ISSN
2222-1719 (Paper) ISSN 2222-2863 (Online) Vol.6, No.2,
2015.
[7] Sridhar, K.P.; Agalya, R.; Narmatha, D.; Vignesh, B.;
Saravanan, S., "Test data compression using Hamming
Encoder and Decoder for system on chip (SOC) testing," in
Circuit, Power and Computing Technologies (ICCPCT),
2014 International Conference, pp.1094-1098, 20-21 March
2014.
[8] Mircea R. Stan, Wayne P. Burleson, “Bus-invert coding for
low power I/O”, IEEE transactions on VLSI systems, vol. 3,
no. I, March 1995.
[9] C.nithya, d.kowsalya devi, “Efficient and high speed VLSI
modelling of fm0/manchester encoding using sols technique
and clock gating technique for dedicated short range
communication applications”, International Journal for
science and Advance Research In Technology(IJSART) vol.
Figure 6. Universal encoder & decoder graph 1 Issue 6, pp.2395-1052, June 2015.
 Microelectronics and Solid State Electronics 2016, 5(1): 7-13 13

[10] Prabhat Kumar Mishra, Sahaj Saxena, “A Novel Approach digital data in communication”, International Journal of
for VHDL Implementation of Universal Line encoder for Research in Engineering and Technology, e-ISSN:
Communication”, International Journal of Electronics 2319-1163 pISSN: 2321-7308, vol. 03, April 2014.
Engineering, 2(2), pp. 339-340, 2010.
[13] Balwinder Singh, Arun Khoshla, S. Bindra. Narang “Low
[11] Suchitra Suresh, “VHDL implementation of Manchester power bus Encoding Techniques for memory testing”,
encoder and Decoder”, International Journal of Electrical, Microelectronics and solid state Electronics 2(3), pp. 45-51,
Electronics and Data Communication, vol-1, ISSN 2013S. M. Metev and V. P. Veiko, Laser Assisted
2320-2084, April 2013. Microtechnology, 2nd ed., R. M. Osgood, Jr., Ed. Berlin,
Germany: Springer-Verlag, 1998.
[12] Vivek Singh, Brijesh Mishra, “FPGA implementation of
various lines coding technique for efficient transmission of