Analog CMOS/VLSI Circuits

Independent Study

Isi Oamen

CWID# 894816099

EGEE 599

Dr. Young D. Kwon

Fall 2011

Introduction:

In modern electronics, VLSI technology plays an important part in everyday applications. The

potential to decrease chip size while simultaneously improving speed is forever improving, as

can be seen with the computers and cellular phones of today. The movement from analog to

digital circuitry led us to replace vacuum tubes with solid-state silicon devices. Due to the

immense decrease in size and manufacturing cost, Metal-Oxide Semiconductors (MOS) and

Bipolar Junction Transistors (BJT) took the lead to be the popular choice in the industry. The

idea is that they can act as switches or voltage/current amplifiers, which can improve power

efficiency as well as performance.

Although these digital devices have been the primary choice for corporations and the consumer

in general, it is important to note that it is necessary for many real world systems to process an

input that originates from a continuously changing source. Because there is no electronic

substitute for what is generated in nature, analog input devices are required because they allow

signals to be handled at the same rate that they are generated. For this reason, a combination of

analog and digital technology is the best solution for those particular situations. This is

commonly referred to as mixed-signal electronic design.

Mixed-signal circuits are involved in the operation of many devices, which include temperature

sensors, radar systems, and amplifiers. The manner in which many of these devices work is that

they receive an analog or continuous input and then convert it to a digital signal. This digital

signal can then be utilized in a multitude of ways, whether as a digital readout or as an input of

another device. One example of this process can be viewed in the block diagram in Figure C-1.

Figure C-1: Temperature Sensor

This temperature sensor works in the following manner: first the analog input is received by a

sensor diode, which outputs a particular voltage based on the ambient temperature. This voltage

is then passed through an analog-to-digital converter which changes the voltage to a particular

binary value. At this point the binary value is sent to a value register. A value register in

essence acts as a table which contains all of the possible binary values and their corresponding

outputs. This output can then be sent to a digital readout or any device in which it can be used as

needed. This same concept is similarly used in pressure and altitude sensors, light sensors, and

seismometers. Another example of a mixed signal system is a Doppler radar/navigation system,

seen in Figure C-2.

Figure C-2: Doppler Radar/Navigation System

The way a Doppler Navigation System works is that it first accepts an analog waveform through

an antenna. The antenna sends this signal to a receiver which measures the frequency or

performs any other necessary calculations. In the event the signal is going the other direction, it

is sent out as an analog value by a transmitter to the antenna. The incoming measurement is

passed along as a voltage which is then converted by the signal data converter into a digital

value. That value is then displayed on the unit. As you can see, the process works in a similar

fashion to that of the temperature sensor: an analog input is converted to a digital output and

displayed to a device. One type of analog circuit that does not work in the same way is an audio

amplifier, displayed in Figure C-3.

Figure C-3: Audio Amplifier System

In this scenario, the microphone contains a transducer which converts the analog audio input to a

voltage, which is sent to the pre-amp. The pre-amp increases the waveform to a higher voltage

so that there is more signal to work with. At this point the tone and volume controls are accessed

to modify the high and low frequencies, as well as to increase the amplitude. The purpose of the

power amplifier is to strengthen the signal by increasing the amount of current that is transferred

to the speakers or any particular audio component. In turn, the component converts the signal

back to sound.

This system differs from the other two because it maintains an analog signal throughout the

process, and simply deals with the amplification of voltages and currents. For purposes of

staying within the scope of the study, I will focus mainly on these types of analog circuits.

Before that can be done, I must first explain the operational characteristics of the transistor and

MOS process.

nMOS Transistor:

The nMOS transistor is the most basic of all the MOS transistors. MOS stands for Metal Oxide

Semiconductor, which means there is a metal gate, an oxide insulator, and silicon semiconductor.

Polysilicon is more often used today than metal for the gate. The ‘n’ represents n-type silicon,

which is the material that composes the source and drain. It is doped with a material to give it

more negatively charged ions. When the device conducts the channel also becomes n-type. A

diagram of an nMOS transistor is shown below in Figure C-4.

Figure C-4: nMOS Transistor Process

The reason nMOS is simple is because it requires less steps in manufacturing and utilizes much

less space than other types of transistors. The silicon substrate, which is generally p-type, is

doped with two n-wells to become source and drain. Field oxide is grown across the substrate; a

polysilicon gate and much thinner gate oxide are also applied. Metal contacts are added to the

source, gate and drain.

When zero voltage is applied to the gate, there is no current flowing through the channel. As the

voltage is increased positively, negatively charged ions are attracted to the surface of the p-type

material while the positively charged ions are moved away. Once the gate voltage reaches a

level that collects enough negative ions, the surface of channel becomes n-type and conducts

between the source and drain. This point is called the threshold voltage. Eventually as the drain

voltage increases the current will approach a maximum point and no longer be linear. When that

situation occurs the mode is called saturation, and all of this is characteristic of an enhancement

type nMOS transistor. This can be seen in Figure C-5(a).

The other type of nMOS transistor is one which contains a small amount of doping at the surface

of the substrate, creating an n-type channel. This channel is already conducting when there is

zero voltage applied to the gate. The only way to eliminate the channel is to apply a negative

gate voltage, which is depleting the channel. This is called a depletion type nMOS transistor,

which is shown in Figure C-5(b).

Figure C-5: Enhancement and Depletion Mode nMOS Characteristic

There are a few equations that will come in useful when deciding which mode the transistor is in,

as well as the amount of current flowing through the drain, or Id. The three modes are cutoff,

triode (or linear), and saturation. The equations along with their constraints are below.

Id = 0 Vgs < Vto (cut-off)

Id = μCox(W/L)Vds*(Vgs – Vto – Vds/2) Vds < Vgs – Vto (triode)

Id = μCox/2(W/L)*(Vgs – Vto)² Vds > Vgs – Vto (saturation)

Where

μ – carrier mobility Cox – gate oxide capacitance W, L – channel length/width

Vgs – gate-source voltage Vds – drain-source voltage Vto – threshold voltage

To understand how these transistors would operate together in a circuit, Figure C-6 shows a

depletion-load nMOS inverter and its drain current and output voltage characteristic.

Figure C-6: Depletion-load nMOS Inverter

Due to the negative threshold voltage and the gate being shorted to the source (Vgs1 = 0),

Transistor 1 is always on. Beginning with Vi at 0V, Transistor 2 is off. This means the output

voltage across the load is at maximum value or Vdd. As Vi is increased, Transistor 2 eventually

reaches its threshold voltage and begins to conduct. This draws current away from the load and

Vo approaches 0V. The second graph shows the effects of increasing Vdd as well as Vgs2 (Vi).

Because the depletion transistor acts as a resistor in the circuit, an increase in Vdd also increases

Id. A higher value of Vgs allows more current to flow through Transistor 2 which also raises Id.

When this condition is true, these equations can be used:

Id = uCox(W2/L2)(Vo)(Vi – Vt2 – Vo/2)

Vo = (Vi – Vt2) – [(Vi – Vt2)² - (W1L2/L1W2)*(-Vt1)²]½

Now that it is better understood how the components of analog circuits work, I can better explain

the amplifier and its operation.

Operational Amplifier:

One of the most popular devices in analog electronics is the operational amplifier (op-amp).

They are very versatile units that have many applications. The low cost of manufacturing makes

them a good candidate for large scale production, and this also allows them to be used for

circuits that contain a system where many are needed. A typical op-amp has three main stages in

its function, which include: differential amplifier, voltage amplifier and output amplifier. These

stages can be seen in Figure C-7.

Figure C-7: Op-Amp Stages of Operation

The differential amplifier stage receives two input voltages and then outputs the difference in

voltages multiplied by a gain factor, which is usually unity. A single output can be generated by

combining it with a source follower or current mirror. The voltage gain amplifier uses the output

of the first stage and increases the voltage. After this happens, the output amplifier increases the

current output with a low impedance and unity voltage gain.

Generally an ideal op-amp will have a few theoretical characteristics such as infinite input

resistance, infinite open loop gain, and 100% dynamic range. What this equates to in voltage

and current calculations is that there will be minimal current through the input terminals and

their voltages at the point of entry will be the same. The range of output voltages will follow that

of the voltage rails and have a high value of Vdd and a low value of Vss.

An op-amp can be utilized as a voltage comparator, which outputs Vdd when the positive

terminal is a higher value and Vss when the negative terminal is greater. Another use is a non-

inverting amplifier, which will increase the input voltage with a gain based on the two resistors

used on the negative terminal. It can also act as an inverting amplifier, which operates in a

similar fashion but the output is negative. They can be popular for analog-to-digital and digital-

to-analog converters as well.

When creating the differential amplifier portion of the op-amp, it is necessary to know what

combination of devices will give the desired effect. A current mirror circuit is the basic building

block of many of the amplifier stages. The way it operates is that there are two nMOS transistors

with the same gate-source voltage. If both transistors have identical aspect ratios (the width and

length of the channel are the same), then the drain current of one would equal the other. One is

considered the input current (Ii) and the other one is the output current (Io). If the aspect ratios

are different then so will be the Io/Ii ratio. The drain-source voltage also has an effect on the two

currents, which is why a second pair of transistors can be added to help normalize the inequality.

The core of the differential amplifier has opposing depletion nMOS transistors in series with

enhancement mode nMOS transistors. The enhancement transistors should have large aspect

ratios compared to the depletion transistors for the gain to be high. One problem that arises is

that the output voltage will lean towards Vss because of the large resistance of the depletion

transistors. The additional transistors which make up a current mirror provide a current source to

bias the differential portion away from Vss. This differential amplifier combined with current

mirror can be seen in Figure C-8.

Figure C-8: Differential Amplifier

As you can see, transistors M1 and M2 control which side of the circuit has a higher potential. If

the gate of M1 receives a higher voltage than M2 (Vi is high) then M1 draws a current, which

pulls Vo towards Vss (in this case, ground). Similarly with the other side, if Vj is high then Vp

pulls towards Vss. The purpose of M5 in this case is to provide a current source, and to help

attain a reasonable common-mode rejection ratio. In order to show the effects of various

voltages on the differential amplifier combined with current mirror (M5-M7, the PSpice

waveform for this stage of the circuit is in Figure C-9.

Figure C-9: Differential Amplifier Waveform

In this circuit, the voltage applied to M1 remained constant at 2 volts while the voltage to M2

slowly increased from 0 to 5 volts. All values are DC. As you can see, when Vi = 2V and Vj =

0V, Vp holds steady at Vdd = 10V while Vo = 1.5V. Vss = 0V but the Vds value of M5 is what

keeps Vo from reaching Vss. As Vj approaches 2V, the difference in voltage works its way

towards 0, which causes the two outputs to begin reversing their polarities. At exactly Vj = 2V is

when the waveforms change to greater and lesser values. From this point on the difference in Vi

and Vj keep the outputs Vo and Vp high and low, respectively.

At this point we currently have two outputs, but the differential output needs to be converted to a

single output. A source follower or differential to single ended amplifier helps achieve this goal.

Figure C-10 shows a version of this circuit.

Figure C-10: Differential to Single-ended Amplifier

The previous outputs are sent to two depletion transistors, which depending on the voltage can

turn them on or off. Each of them leads in series to an enhancement nMOS, of which both share

a gate voltage. This means that they are either both on or both off. If M12 is on and M13 is off,

then the output is low. If M12 is off and M13 is on, then the output is a voltage differential of

M13 and M11. If both of them are off then the output is high.

The following stage is that of the voltage gain amplifier. This can be a very simple stage which

involves two nMOS transistors in a traditional inverter setup, but the input voltage is tailored so

that the inverter is continuously in the linear mode. When an inverter is in this mode it has a

very large current draw and its negative voltage gain is extremely high when combined with a

resistor or an nMOS transistor acting as a resistor.

The final stage of the op-amp is the output amplifier. This involves two more depletion nMOS

transistors and two enhancement nMOS transistors. Shown in Figure C-11 is a power output

stage which is set up in this fashion.

Figure C-11: Power Output Stage of Op-Amp

The way this process works is when M14 receives a low input it remains off. This sends a high

input to M16 which then turns it on and pulls Vout to ground. The low Vout value means that

M15 and M17 are both conducting, but Id = 0 for M15, while M17 has a very low Id due to the

lower Vgs. The load will not receive much power. On the other hand, if the input to M14 is

high then M14 conducts, which sends a low output to M16 and turns it off. This creates a high

voltage at Vout which then supplies the load. The finished product is shown in Figure C-12.

Figure C-12: Complete Op-Amp Schematic

When looking at this circuit from a whole, you can see that V1 and V4 serve as the high and low

voltage rails (Vdd and Vss) for the amplifier, while V2 and V3 are the inverting and non-

inverting inputs respectively. Vout is the output from the final stage. Unfortunately the trial

version of PSpice Student Edition doesn’t allow circuits this large to be simulated, so I could

only find waveforms for the differential amplifier which is the largest and most versatile portion

of the circuit. The best I could do was to replace my circuit with the op-amp symbol to run a few

simulations. Figures C-13, C-14, C-15, and C-16 show the operation of both an inverting and

non-inverting op-amp with their waveforms.

Figure C-13: Non-inverting Op Amp Circuit

Figure C-14: Waveform for Non-inverting Op Amp

Figure C-15: Inverting Op Amp Circuit

Figure C-16: Waveform for Inverting Op Amp

As you can see, the gain on the non-inverting op-amp is roughly Vi(1+R1/R2) = Vi(1+1) = 2Vi,

which starts to deviate slightly as Vo reaches Vdd. The inverting op-amp has a gain of

(-R1/R2)Vi = (-1k/1k)Vi = -Vi.

In conclusion, over the course of the semester I was able to learn about analog VLSI circuits

through the analyzing of an operational amplifier and its components. I believe they still have

many useful purposes in today’s mixed-signal technology. Although I ran into some road blocks

with the software because of the large nature of the circuit, I was still able to break the circuit

down into smaller pieces and analyze those individually. Due to the absence of an analog course

at CSUF, I was grateful to be given the opportunity to study something not offered at the

university.

Works Cited:

Ayers, John E. Digital Integrated Circuits, Analysis and Design. Boca Raton: CRC Press, 2004

Haskard, Malcolm R., and Ian C. May. Analog VLSI Design, nMOS and CMOS. New York: Prentice Hall, 1988.

Supplemental links:

http://www.ladyada.net/learn/sensors/tmp36.htmlhttp://en.wikipedia.org/wiki/Analog_circuithttp://en.wikipedia.org/wiki/Analog_signalhttp://howto.circuitdiagram.net/basic-theory/audio-amplifier-block-diagram/http://www.kpsec.freeuk.com/bdiags.htmhttp://people.seas.harvard.edu/~jones/es154/lectures/lecture_6/lecture_6.htmlhttp://www.allaboutcircuits.com/vol_3/chpt_8/2.htmlhttp://www.national.com/AU/design/courses/268/the02/03the02.htm

Exercise 1.1 – Name some analog type circuits to study.

Heat/temperature sensors, Doppler radar, pressure/altitude sensors, audio amplifiers, light/IR sensors, seismographs, etc.