ISSN 2277 – 3916 CVR Journal of Science and Technology, Volume 15, December 2018

Implementation of Low Power Sigma Delta ADC

for Broad band Communications
B Satish1and Prabhu G Benakop2
1
Assoc. Professor, CVR College of Engineering/EIE, Hyderabad, India
Email: satishbojjawar78@gmail.com
2
Principal, INDUR Institute of Engineering & Technology/ECE, Siddipet, India
Email: pgbenakop@rediffmail.com

Abstract: For the broadband communication systems, a) Front-end Amplifier

the high data rate converters are required to achieve It is difficult to achieve specifications simultaneously,
high accuracy and higher bandwidth. The mixed signal high gain and high speed due to degradation of output
approach using CMOS technology is mostly preferred for conductance in short channel transistors. By including a
these applications. Now a days, the ΣΔ converters are gaining cascade stage, a better compromise between speed and gain
popularity with their robustness. This design is implemented is obtained, so the DC gain of the transistors with minimum
mainly in digital domain with small analog content. In this length is increased to 80 dB and the gain bandwidth product
paper, a high speed 14-bit ΣΔ modulator with 2 MHz signal of 250 MHz, with 37 mW power consumption.
bandwidth with low oversampling ratio is designed. To
increase the resolution, the dual quantization technique TABLE. 1
is used. It simplified the design and reduces the SPECIFICATIONS OF BUILDING BLOCKS FOR ΣΔ ADC
hardware circuitry and power consumption.
Index Terms: ΣΔ modulator, ADC, DAC, quantization, Over Modulator Clock Frequency 60 MHz
Sampling, Low power and CMOS.
Differential Reference 1V
Voltage
I. INTRODUCTION
Oversampling ratio 16
The high performance data converters requires accuracy Clock jitter 15ps (0.1%)
and high resolution of 12 to 16 bits, higher bandwidth with Dual quantization 1bit /4 bit
less production cost for broadband applications. The ΣΔ Front-End Sampling capacitor 0.5pF
converters are mostly preferred for pipelining because of Integrator Switch on resistance 250Ω
oversampling and robustness of the converter. The CMOS Capacitor non-linearity 25ppm/V
implementation of ΣΔ converter increases the inherent Amplifier Open Loop DC gain 2500
complexity but, improve the speed of the conversion Slew rate 380V/µs
process. Gain bandwidth product 235MHz
A high speed ΣΔ modulator for signal bandwidth of Comparat Offset ±10mV
2MHz is implemented in deep sub-micron 45 nm CMOS ors Hysteresis 10mV
technology. The fourth order ΣΔ architecture is considered
Resolution time 3.5ns
for the design [1]. The fourth order ΣΔ modulator consists of
ADC and Resolution 4-bit
four integrators, comparator and amplifiers, ADC and DAC
DAC DAC INL 0.4% FS
circuits. The amplifier and comparator are the analog blocks
and remaining blocks are implemented in the digital form.
Hence the mixed signal approach is used for the Figure 1 shows a two stage operational amplifier
implementation of ΣΔ modulator in this paper. (OpAmp1) with Miller compensation is used to achieve high
gain and high bandwidth is achieved. .Miller compensation
The specifications of the ΣΔ modulator are considered for converts an internal pole pertaining to specific node, to
the applications of broad band communications. The table.1 become dominant pole by adding a small physical
represents all the specifications of important building blocks capacitance, which will be simulated as a large capacitance
used in the modulator. due to Miller effect and brings the effect of pole with low
frequency. This technique is used for the transistors M12
and M13,so that high bandwidth as well as high speed can
be achieved.
II. BUILDING BLOCKS OF ΣΔ MODULATOR
In this section the important building blocks of ΣΔ One more compensation technique called nulling resistor
modulator are designed and implemented in the 45nm along with miller capacitor can be used to improve stability
CMOS technology. All these blocks are operated at the [2] [3]. The nulling resistor eliminates the zero present in the
power supply of 1 Volt. right half of s-plane or even it brings the zero present on RH
side into LH side, so that excellent stability can be obtained.
But nulling resistor implemented with MOSFET when

CVR College of Engineering

 ISSN 2277 – 3916 CVR Journal of Science and Technology, Volume 15, December 2018

coupled through miller capacitor degrades the transient b) Amplifiers other than stage 1
response of integrator, as the output voltage will be reduced The second stage and next stage integrators require
due to compensation capacitor. This technique is used in this medium DC gain amplifiers. Hence, a cascade single stage
implementation as it needs less power to implement and is amplifiers are used. Telescopic operational amplifiers has
more stable with temperature and process variations. advantage of providing signal path by only NMOS
transistors, non existence of mirror poles and a small bias
current is required to operate with expected output voltage
swing of 1 volt. The disadvantage of telescopic Op-Amp is
that its common mode voltage levels at the both inputs must
be different and also it requires a precisely controllable
voltage reference to get stable and large voltage swing. A
folded cascade operational amplifier is used in this design
[5]. It has more advantages than telescopic Op-Amp like,
superior frequency response, better PSRR even both
consume equal power.

Figure 1: Two stage operational amplifier (OpAmp1) internal diagram with

Miller compensation.

The weights used in implementing integrators strongly

influences the output swing of the amplifier, when the
reference voltage is fixed.
Figure 3. Schematic of Folded-Cascode Amplifier (OpAmp2, OpAmp3,
The transient response of the two stage Operational OpAmp4).
Amplifier used as a front-end integrator in the modulator is
presented in the figure2 [4].It behave like first order system The simulated output results of the OpAmp1,OpAmp2,
and it settles properly in the given time slot. OpAmp3 and OpAmp4 are presented in the table 2. The
order of the ΣΔ modulator refers to the number of loops
with integrator. These four Operational Amplifiers used as
four integrators in the fourth order ΣΔ modulator. The power
consumption, gain and some important parameters are
compared among all the operational amplifiers.
The power consumption of OpAmp4 is higher when
compared with OpAmp2 and OpAmp3, the reason is that the
OpAmp4 has larger load capacitance.
TABLE 2.
SIMALTAED OUTPUT PARAMETERS OF THE FOUR AMPLIFIERS

Parameters OpAm OpAmp OpAmp OpAmp

p1 2 3 4
DC gain 80.0 62.8 55.7 62
(dB)
Figure 2: Transient response of two stage operational amplifier (OpAmp1)
GB (MHz) 250 311 261 167
PM 67 65 71 79
The biasing requirement of an amplifier is fulfilled by (Degrees)
transistors M14 to M19. The amplifier biasing stage is SR (V/µs) 390 410 320 185
implemented by transistors M14to M19. A dynamic CMFB Input Noise 5 3.2 4.1 3.8
net is used to minimize the power consumption as it zero (nV/√Hz)
static power dissipation. Note that an inverting stage (M20, OS (Volts) ±2.5 ±3.1 ±2.9 ±3.1
M221) is added to the basic SC net in order to implement the Power (mV) 38.5 4.5 4.0 6.6
negative common-mode feedback.

CVR College of Engineering

 ISSN 2277 – 3916 CVR Journal of Science and Technology, Volume 15, December 2018

The figure 4 shows the simulated output waveform of the The important features obtained after simulating the
second integrator, which is also used for third and fourth comparator are shown in the below table. 3.
integrators in the modulator implementation. The output is
linear but has limited slew rate as the applied input is large TABLE 3
step signal. This is tolerable for this design as it reduced the COMPARATOR RESULTS AFTER SIMULATION
power dissipation.
Hysteresis < 10 mV
Resolution time, LH 2.5 ns
Resolution time, HL 2.3 ns
Power consumption 0.65 mW

III. PROGRAMMABLE A/D/A CONVERTER

The switched capacitor implementation of 4-bit A-D-A
converter is implemented. It can also be used for 2 or 3 bit
resolutions. A fully differential parallel ADC is designed
and used to compare the differential voltage as illustrated in
the figure.6

a) A/D converter
Figure 4.Simulated small- and large-signal transient response of the fourth
integrator.
The fully differential flash architecture is used for A to D
converter, which compares the differential input voltage
c) Comparators with the reference voltages [7]. The reference voltages are
The comparators used in this design needs low resolution produced by DAC. During the Φ1phase,Vref positive and
time with large hysteresis voltage of up to 10 mV. Vref negative values are stored in the input capacitors,
Generally, this can be implemented with traditional which are then used to compute the difference
comparator with pre-amplifier stage, positive feedback duringphaseΦ2. At the end ofΦ2, comparators are activated
circuit used as decision making circuit and output buffer. to find the sign for that difference. The thermometer output
But it takes large area and more power required for code of the 15 comparators is then translated to a code using
implementation. Instead a dynamic comparator with AND gates.
regenerative latch is used as it does not require the pre- Comparators in the ADC are identical to those used in the
amplification stage and has zero static power dissipation. 1st- and 2ndstage.Multi-metal sandwich capacitors of value
The regenerative latch used acts as decision making circuit are used and the analog switches are identical to those in the
while comparing the two inputs. SC. The timing scheme of the switches has been adapted to
reduce the capacitive load to the fourth integrator.
Figure.5 shows the architecture employed for the Nevertheless, it suffers from input dependent charge
implementation of comparators [6], which has been widely injection from switches controlled byΦ2. This problem has
used in ΣΔ modulator design. In practice, this topology is been overcome making these switches considerably smaller
capable of achieving resolutions about that required with no (for both NMOS and PMOS), with no degradation of the
pre-amplifying stage. converter performance.

Figure 5. Dynamic comparator with Regenerative latch Figure 6. Programmable A/D/A converter: Partial view of the Switched
Capacitor implementation.

CVR College of Engineering

 ISSN 2277 – 3916 CVR Journal of Science and Technology, Volume 15, December 2018

b) D/A Converter d) Clock phase generator

For the proposed ADC, the ladder type DAC is used. This The different clock phases are generated using the circuit
DAC provides the reference voltage for the ADC. The same diagram shown in the figure 9 from external clock. The
DAC is also used for the conversion of digital voltage to delayed version of these clock phases are used to reduce
analog output voltage in overall A/D/A converter. The signal dependent charge injection[8]. The complimentary
ladder resistor network will generate the equivalent output clock signals for the phases also generated by this circuit.
analog voltage based on the selection of resistors
corresponding to digital code.

The Resistor ladder network is comprises of 30 unit

resistors, which are connected between positive and negative
reference voltages, +Vref and –Vref respectively. The each
unit resistor value is 50Ω and differential full scale current is
1.33mA. The resistors are implemented by unsilicided poly
layers to provide high resistivity, good matching and
independent on variations of temperature and it also
provides small settling error.

c) Control Circuitry
The 4 bit A/D/A converter is provide 2 or 3 bit resolution
with the help of digital control circuitry, which is shown in
the figure 7. The selection of the desired resolution is done
with two signals, S3b and S2b. The partial view of the digital Figure 9. Clock phase generator and drivers.
control circuitry, which is used in the programmable A/D/A
converter is shown in the figure 8. The complete timing of the different clock phases are used
in the ΣΔ modulator is shown in the figure.10

Figure 7. Block diagram of the programmable A/D/A converter.

Figure 10.Timing of the clock phases in the ΣΔ modulator.

IV. RESULTS
The functionality of the entire A/D/A converter operation
is tested and measured using specific samples. The ADC,
DAC and clock driver also measured using specific test
samples.
The code histogram method of sine waves is used to
measure the ADC performance [9].An amplitude of 1 volt
and 103 KHz frequency was applied to the ADC. The
corresponding digital codes for 100 input periods were
measured by digital tester. The analog and digital voltages
are estimated from the 16 digital output code.
The 16-bit digital code is applied using digital tester to
DAC and measured the performance, and the corresponding
analog output voltages of DAC using high accuracy multi
meter [10]. The performance measurement results are shown
in the table 4. From that table we can conclude that the
Figure 8. Partial view of the control circuitry.
proposed ΣΔ modulator meets the specifications.

CVR College of Engineering

 ISSN 2277 – 3916 CVR Journal of Science and Technology, Volume 15, December 2018

TABLE 4. [8] Y. B. Hsieh, Y. H. Kao, "A fully integrated spread spectrum

PERFORMANCE MEASUREMENTS OF THE A/D/A CONVERTER clock generator using two-point delta-sigma modulation",
Proc. IEEE International Symposium. Circuits Systems., pp.
parameter ADC DAC 2156-2159, May 2007.
%FS LSB %FS LSB
Offset error -2.014 -0.203 0.302 0.045 [9] A.R. Feldman, B.E. Boser, and P.R. Gray, “A 13-Bit, 1.4-MS/s
Sigma-Delta Modulator for RF Baseband Channel
Applications”. IEEE Journal of Solid-State Circuits, vol. 33,
Gain error 4.595 0.659 -0.633 -0.098 pp. 1462-1469, October 1998.

DNL 2.894 0.453 0.153 0.016 [10] K. Bult, G.J.G.M. Geelen, “A Fast-Settling CMOS Op Amp
INL 1.652 0.256 0.156 0.018 for SC Circuits with 90-dB DC Gain”. IEEE Journal of Solid-
State Circuits, vol. 25, pp. 1379-1384, December 1990.

V. CONCLUSIONS

In this paper, the low power ΣΔ modulator is implemented

with resolution of 14 bits for the ADC with signal
bandwidth of 2MHz. A 2-1-1 ΣΔ Modulator with 4th order 3
stage cascade topology employing dual-quantization was
selected. The dual quantization is only used last stages of
modulator. The quantization power was reduced without any
change in the linearity of the modulator. All the building
blocks of the modulator are implemented in SC form using
CAD tools. The topology of all the blocks and designs in
transistor and gate level are described in this paper.
The entire modulator is implemented in the CMOS
technology with power supply voltage of 1V. The low Vt
and multi metal capacitors, gates and switches with thick
oxide are used to reduce the power in the modulator

REFERENCES

[1] Karema T., Ritoniemi T., and Tenhunen H., "Fourth Order
Sigma-Delta Modulator Circuit for Digital Audio and ISDN
Applications", Proc. of European Circuit Theory and Design
Conference, 1989, pp 223-227.

[2] D.A. Johns, K. Martin, Analog Integrated Circuit Design.

John Wiley&Sons, 1997.

[3] B. Razavi, Design of Analog CMOS Integrated Circuits.

McGraw-Hill, 2000

[4] HenilLangalia, Sarthak Lad, MangeshLolge, SurendraRathod,

"Analysis of two-stage CMOS Op-Amp for Single-Event
Transients", International Conference on Communication
Information & Computing Technology (ICCICT), pp. 1-4,
Oct. 19–20, 2012

[5] S. C. Liu, K. T. Tang, "A low-voltage low-power sigma-delta

modulator for bio-potential signals", 2011 IEEE/NIH Life
Science Systems and Applications Workshop (LiSSA), pp.
24-27, 7–8 April 2011.

[6] Yukawa, “Analysis and Design of 2nd order sigma-delta

modulator for audio applications”. IEEE conference on
Advances in computing, communications and informatics, at
New Delhi, India, Dec2014.

[7] Xian Tang , Kong Pang Pun, “New fully-differential amplifier-

less pipelined ADC with wide power scalability and ERBW,
Analog Integrated Circuits and Signal Processing”, v.80 n.3,
p.427-435, September 2014

CVR College of Engineering