Digital Circuits

Prof. Santanu Chattopadhyay

Lecture - 01
Introduction

Welcome to this course on Digital Circuits. So, if you look into this VLSI design process
or VLSI industrial development where this I integrated circuit chips are being
manufactured. So, you will see that most of the designs that we have so, they are digital in
nature.

So, the basis of all these developments are on these type of circuits where we consider the
digital circuits as the building block.

(Refer Slide Time: 00:45)

So, the start of this modern electronics era if you look back in to the history so, we will
see that it started around the revolution started in 1947, where a group of a scientist at bell
labs they invented the bipolar transistors in 1947. So, 1947 is a very important year for
India as well we got freedom, and similarly the on the electronics industry. So, this is
another land mark that we had achieved in 1947.

So, from that point onwards the electronics design it has advanced a lot. And today we are
at a stage where this the complexity of systems and their performance have gone up by
leaps and bounds. And now it is it is almost whatever functionality we think about so, we
will look for an integrated circuit chip which can do that. And majority of them are digital
in nature.

So, we will why this has happened. So, if you look into the, these slides on the next in the
right side, it shows the first germanium bipolar transistor. So, this left side. So, this was
silicon, based on silicon. So, this is the bipolar transistor based on germanium. So, 50 years
later so, this electronics account for about 10 percent of the world’s total GDP.

(Refer Slide Time: 02:14)

So, this is the milestones that we have in the electronics industry. In 1874, Braun invented
the solid state rectifier. Then 1906 Deforest, Deforest invented triode vacuum tube. So,
that way this the whole development has taken place. So, it started with solid state rectifier
then the vacuum tube diode the triodes, then the radio circuits, then field effect devices the
patent was filled in 1925.

1947 as we already said so, there is a bipolar transistor was invented. Then 1952, Texas
instruments they started making these bipolar transistors commercially. 1956 that there
was a Nobel Prize given to Bardeen, Brattain and Shockley. So, they receive the Nobel
Prize for this transistor.

Then IC was developed in 1958 by Kilby and Noyce, and then the first commercial IC
came from the Fairchild semiconductor in 1961. So, as this industry was advancing so, so
the scientists also started doing lots of research works, and there was an institution of
engineers ie which was formed - IEEE so, electrical and electronics engineers.

So, that was found by merger of IRE and AIEE. So, IEEE is one of the very popular
societies now that we have and they apart from just bringing peers close to each other.
They make a lot of standards for the new devices and systems that are coming up, and that
way they are helping in the electronics industry.

So, whenever we are looking for some development of some device or system, we
normally look for some standards and that standardization is generally done by IEEE
whatever we have today. The operational amplifier; so, this is the analog part. So, though
we will not discuss much about the analog revolution, but this analog part also has started
be getting developed from 1968.

So, first operational amplifier IC came up in 1968. The other important part that we have
in it is system, digital system; so, apart from during computation. So, you need to store lot
of information and in that direction the memory chips. So, they are the main part that we
have. So, they so, transistors are again are used for making the memory chips. So, this
dynamic RAM that we see in almost all the computer systems today. So, they are they
were invented in 1970 at IBM by Dennard.

So, in the involution process as this computations and computations and memory storage.
So, they went on improving over the generations. So, this computational module. So, they
were clubbed together into a structure called microprocessor, where it can do some set of
operations in taking help of some registers are may be taking executing some stored
program, which it can access through some address and data lines. So, that microprocessor
came in 1971. So, it was Intel 4004 microprocessor.

So, and now you see that after that it has gone a long way, and today we have got this
Pentium processors and other RISC machines and all so, they are all very advanced
versions of this microprocessors. So, this is another digital component that has involved a
lot. Then this commercially memory chips they started coming in 1978. So, first one kilo
bit memory chip was available in 1978. 1974 another very popular microprocessor 8080
came up.
So, this is one of the, I should say with the one of the first well known microprocessor
8080 of the, again from Intel. And that so, based on this microprocessor there has been
they lot of developments and all this later microprocessors that are available; so, they are
actually developed from this microprocessor architecture.

We will see one such microprocessor in this particular course which is 8085. So, it has it
is bit improved version over 8080. In 1984, we have got mega bit memory chip that was
introduced so, memory capacity went up. So, this was a because of because of this we
could we can now stored lots of memory and today storage is not a problem. Because this
memory chips they have advanced so much like even this tiniest devices that we have in
our day to day life, electronic devices. So, they have got large amount of memory that we
can have in the system. So, they can store lots of information.

So, that is the, that is the point so, we this memory chip development that is also digital in
nature. So, this digital circuits so, they helped us in developing this memory chips to get
the functionality implemented by the systems. Then the, this another Nobel prize came up
in 2000 to Alferov, Kilby and Kromer. So, they have got another Nobel Prize for this
memory and all.

So, this way this electronic industry came. So, after 2000 so, we do not have the
information here, but as you are all familiar. So, lots of development are taking place every
day where this electronics industry is advancing towards better and better processors with
higher and higher capabilities, and better and higher memory storage capacities.
(Refer Slide Time: 08:22)

So, if pictorially so, this shows that. So, this is the first generation where we have got this
is the vacuum tube diode triodes and all. Then they came that discrete transistors. So, you
see the volumes or size of the devices. So, that has decreased price so much you see. So,
this is a power, power transistor so, this is another transistor. So, these are some diode
some transistors and all so, they are all discrete transistors.

Then after that came this one. So, where we have got this small scale integrated circuits
and medium scale integrated circuits, MSI and SSI. So, they are actually this IC chips
started coming. So, they do not have large number of transistors in them, SSI chips they
will have about 10 transistors, MSI will have about thousand, 100 transistors. LSI will
have thousand transistors, and then VLSI is going beyond that.

So, that way we have got large number of transistors. So, today this we have got even
further terms like ULSI and all so, that is actually having larger and larger number of
transistors in them. And now this is not a limitation; like, if we want to pack large number
of transistors on a single chip that is possible because of the technology advancement. So,
we can have transistor size is very small so, we can pack large number of transistors on a
silicon floor.

So, that way size is not a problem, but the difficulty comes in terms of power consumption
in terms of heat generation and all. So, those are of course, not part of this course
discussion. So, they will go to some course on VLSI design.
(Refer Slide Time: 10:09)

So, integrated circuit it was invented in 1958, then this world transistor production has
more than doubled every year for the past 20 years. So, every year so, total transistor
productions so, it has doubled. So, you can imagine the amount of functionality that is
becoming digital every year, digital or electronic going to electronic components every
year.

So, more transistors are produced every year more transistors are produced which is more
than all the previous year previous years combined. So, if you sum up are all the transistors
that are produced till last year. So, this year it is going to be double of that sum. So, that is
an exponential growth in terms of transistor production you can say.

So, approximately 109 transistors were produced can be, in some recent year. So, though
it is not mentioned which year. So, it is around that. So, we can and roughly 50 transistors
for every ant in the world. So, this is another statistical information. So, if you think about
all the ants that are available in the world. So, each of them that multiplied by 50 so, that
give the total number of transistors in the world, this is just a speculation of course.

So, this is obtained from 2003 source so, so, you all this figures. So, they have gone up
significantly by the time today.
(Refer Slide Time: 11:40)

So, there are a few laws which actually predicted the growth of this electronic industry.
So, first one is the Moore’s law, it says that the number of transistors on a chip doubles
annually. So, this is a prediction and more or less it has been followed, then there is a
Rock’s law it says that the cost of semiconductor tools doubles every 4 years.

So, this is another problem because as the complexity of the system is going up so, you
cannot think about designing the system by hand. So, you cannot just take draw say 10 12
transistors, connect them in some fashion and get your circuit design. So, that is not
possible so, we have to take help of these semiconductor tools; that may be based on, that
maybe for this design specification, that maybe for simulation, that maybe for synthesis,
that may be for testing.

But the costs are going up for every 4 years. Because the complexity of this systems are
going up, complexity of this chips are going up. So, testing them the corresponding testing
and implementing them the corresponding tools so, they are also becoming very complex,
and the cost is going up significantly.

Then there is another law due to Machrone, which says that the pc you want to buy will
always be dollar 5000. So, what does it mean it means two things, first of all it says that
the cost will not go up significantly, because of the advancement in the semiconductor
industry. So, the cost of production will go down so, that way it is going like this. So, you
can say the cost are producing a transistor is coming down. So, that way it is it will in some
sense say.

So, it will balance the cost increase in the in the in the tools that we have, in the cost of the
semiconductor tools that we have so, that that will get balanced. So, if we say on the other
side that your cost of this manufacturing has gone down, but what has gone up is the cost
of the software. So, software’s are becoming more and more complex, as a result this the
PC will always be a more or less the same price so, if you look at any point of time.

Then there is a Metcalfe’s law which says that the network’s value grows proportionately
to the number of users of it is users squared so, that you know very well. Today in the in
the era of this mobile networks so, you know so, if they are as the number of subscribers
to the network increase we value that network further because that gives us connectivity
both in our day today life and social life.

So, that way this is another law which says that if you if you are making the network
complex, and you can accommodate more number of users. So, you are going to be paid
back at a square rate. So, that is also very important thing.

(Refer Slide Time: 14:43)

Then there is a Wirth’s law it is say that software is slowing faster than hardware is
accelerating. So, this is another important issue which says that the software is slowing.
So, why because the computational power is increasing. So, as a software designer people
think that we can pushing lot of functionalities into it, and when we do that, the overall
software speed that is coming down. So, so, essentially again there are research which will
try to make that software faster.

So, hardware engineer they come up with some assisted tools by which you can make this
hardware faster to run those operating system and software modules, but it is actually the
software which is pooling back the developments in hardware, the speed advancement in
hardware to a limit so that you can always you will always have the more or less the similar
type of speed if you are if you are using if you are not using very advanced software.

So, this is a reference that talks about these commandments further.

(Refer Slide Time: 15:46)

So, if you look into the Moore’s law; so, Moore predicted that the number of transistors
that can be integrated on a silicon, on a on a die so that would grow exponentially with
time. So, it is says that it will double every year, every 2 years or so.

So, what was happened is a it is amazingly visionary law that Moore has predicted. So,
you see that million transistor per chip barrier was crossed in the 1980s. So, and then I
mean Ultra Sparc III, had 16 million transistors, then HP’s PA 8500 140 million
transistors. So, that way the number of transistors that we have so that is more or less
following the Moore’s law.
(Refer Slide Time: 16:29)

So, this is the evolutions if you see that this is a. So, from 1980 to above 2010. So, you we
if we just look into this DRAM capacity. So, dram how this DRAM capacity is growing
so, you see that this is it was about say 64 kilo bit per chip in the around 19 in the year
around 1980 so, which was sufficient for holding say one page of information.

Then around 1989 1990 around that time we have got the capacity which us sufficient for
a book. Then in the year 2001, it reached to support suppose as a it can hold the entire
encyclopedia. So, 2 hours of CD audio or 30 second of HDTV, high definition television
high definition television signals.

Then this then in 2010. So, the capacity has increased so much that even the total human
memory or human DNA, we can store in that chip. So, memory IC is it is a becoming 4
times every 3 years. The capacity of the memory chips are it is becoming 4 times every 3
years roughly. So, that way it is memory capacity is going to increase.
(Refer Slide Time: 17:51)

If you plot the number of transistors per integrated circuit so the trend is like this. So, it is
started in 1960 so, if you say that it just started then it has gone up.

So, this Moore’s law, which says that the complexity for minimum component cost has
increased at a rate of a roughly a factor of 2 per year. So, it was later amended to 18 to 24
months ok. So, this is a 2 year so, this was shortened further. So, if this was a brought
down to 18 months also.

So, this way where is the transistor per integrated circuit so, number transistor so, it has
increased following Moore’s law and that is going on, and as the technology is advancing
further and further. So, whenever it comes to a saturation point you see that the technology
advances is advances it to the next state, next stage and this Moore’s law still continue to
hold in the next generation.
(Refer Slide Time: 18:54)

Feature size or device size so, feature size feature, when you say feature so, it is a size of
a transistor you can say. So, that is the feature size in the VLSI terminology so, this feature
size is reducing that will enable by process innovations. So, you we in the VLSI
manufacturing process so, it is advancing so, we can produce a smaller and smaller sized
devices, transistors.

So, you can have say some something like say 16 nano meter designs now. So, that way it
is a big advantage that we have so, the feature size is reducing so, overall size of the system
is coming down so on the same silicon floor so, you can put in lot of transistors.

So, they put more transistors per unit area and therefore, we can have higher density. So,
this is a plot till 2000 predicted up to 2020. So, after 2005 so, these are the goals that is
said up to this much this was fine and then it is so, it is dynamic memory inter feature size
in micron so that has come down significantly.
(Refer Slide Time: 20:06)

So, if you look in to this memory chip density versus time so, that is also density is also
increasing significantly, as you say see over the year. So, that is following a more or less
an exponential development and this microprocessor complexity. So, over this if you
measure in terms of number of transistor so, that is also going up exponentially.

So, you see that from 1970’s where we have we if you just start around 1975 or so, this
4004, Intel 4004 was introduced, then 1980 this 8008, was introduced so, like that it went.
So, there are lots of developments that are taking place and they more or less following
this Moore’s law in the rate of increase of this device density and this computational
power.
(Refer Slide Time: 21:01)

So, if you look into these developments in the electronics industry so, you find
development both in the analog domain and in digital domain. So, to on the analog domain
it is mainly the operational amplifier, introduction of operational amplifier and circuit
development around that, so, which is which has got the momentum on the digital domain
of course, we have got different type of logic families that are being developed and we can
realize some switching circuitry based on those logic on those logic elements, and
accordingly we can have some digital design implemented in the IC.

So, if you just try compare between analog and digital electronics so, most observables are
analog in nature. So, why do I say so, because nature is analog so, there is nothing digital
in nature. So, if you look into the temperature. So, it is not the temperature is the digital
quantity, if you look into the amount of light so; it is never a digital quantity. So, that way
it is a most of the observables that we have. So, they are analog in nature so, weight of
some person so, that is also an. So, that is also analog in the sense that. So, if you measure
the weights of different people. So, they will be you cannot always say that after say 75
kg. So, next weight will definitely be 76 kg or 78 kg. So, in between you can get number
of other values.

So, they are observe there will be a values are analog in nature, most of the; but most
convenient way to represent and transmit information electronically is digital. So, we will
see why so, once we have got it in digital format. So, we know that with the transmission
becomes easy from our say if this our devices the electronic devices that we are handling.
So, we are nor- normally going for digital design so, why this digital? So, that will try to
explain in the part of this course.

So, all the processing job that we do so, they are digital in nature whereas, the nature the
environment is inherently analog. So, what is required is that if we are trying to process
something in the analog in the something related to the environment. So, we should have
this interface, which will convert the analog input signal into digital and again that after
processing the actuation should be done in terms of digital values, which should be
converted to analog value, for controlling the environment.

So, this way we have got analog to digital conversion on the input side of this digital
processors and on the output side will we may have a digital to analog conversion. So,
which will convert these digital values to analog actuation signals.

(Refer Slide Time: 24:00)

So, just to have a understanding so, this analog signal; so, suppose we are using this piano
and thus this key A is place so, that is the 440 hertz. So, if you so, you can represent it in
terms of a sinusoidal signals sinusoidal voltage signal which has got a maximum volt of
50 maximum peak of around 50 microvolt, so, this 440 hertz 50 microvolt sinusoidal signal
may be like this. So, in this side if you plot the time in milliseconds and this side if you
plot the volt voltage in microvolts, so the signal generated from the piano key A so, it may
be like this.
So, on the same time, if you say that if you are using microphone voltage with soft pedal,
then this is 20, it is a 25 microvolt signal, frequency may remain same. So, it is 25
microvolt peak to peak, and 25 microvolt peak and we have got 440 hertz signal. Or you
can have this piano A, but one octave bellow the middle A so, that is about 220 hertz so,
frequency is less. So, this is the peak value is 50 microvolt and this frequency is going to
be 220 hertz. So, this way we can have the analog signals which are coming from the
instrument.

(Refer Slide Time: 25:31)

Now, this analog signals that we are talking about. So, they may have a very physical
relationship to information presented. So, naturally this the direct signal that we have so
they are information versus time. So, the so, this is basically the wave form the, that is
generated is in terms of that is giving us some information in terms of the different time
instance. So, in, but it is a normally what is done is that we never have this analog signals
transmitted or analog signals processed directly. So, we convert it into some other form,
which is known as the process of modulation.

So, we with the information that we want to send. So, that or we that we want to transmit.
So, it is modulated on a carrier, such as in amplitude modulation or this frequency
modulation the AM radio or FM radio that we have.

So, what is happening? So, this is an example of this amplitude modulated signal. So, here
these high frequency signal that we have here so, that is basically the carrier signal. So, we
have got this high frequency, this blue colored one so, this is the high frequency carrier
signal. And so, this is used for transmitting a relatively low frequency sinusoidal signal.
So, if you look into this signal so, this high frequency signal so, this it frequency is much
higher. Whereas, this low frequency signals so, it is time period is pretty high.

(Refer Slide Time: 27:20)

So, accordingly so, this if this high frequency carrier signal is multiplied by this low
frequency sinusoidal signal. So, we will get some wave form like this. So, this is the
amplitude modulated signal.

So, this the diagram it shows the parameters like see this is the; so, this parameters of the
sin waves like period time to repeat then frequency. So, these are the parameters for any
sin wave so, we can have so, in this case. So, this is the 25 micro second. So, this is the
period of the modulating signal that we have. So, thus the information that we want to
transmit so, that is the, that is that period is 25 microsecond whereas, the frequency sin
wave. So, it is time period is one microsecond so, this is high frequency.

So, the period of the carrier is one microsecond that is the frequency is one megahertz, and
the period of the modulation is 25 microseconds. So, the sinusoidal signal that we want to
transmits so, that periodicity is 25 microsecond. So, it is frequency is 40 kilohertz, much
lesser than this one megahertz signal.
Then so, this amplitude so what so, amplitude of this carrier signal is getting modulated
by this. So, this amplitude of the modulation is about 50 percent of the maximum possible
value. So, you can just do some mathematics and try to figure out, say like what is the
corresponding equation for this ultimate output signal. So, this is going to be an expression
like this ok.

So, this way we can have this amplitude modulated signal which will be which will be
used for this analog communication purpose.

(Refer Slide Time: 28:55)

On the other hand, when we go for digital signal processing so, we normally use this binary
representation. So, this will help us in making the system much more robust. So, we will
see it in the next class.