Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
Chris Carr
Space & Atmospheric Physics Group
Room 6M65 Huxley
c.m.carr@imperial.ac.uk
January 2018
An atom of copper (Cu) contains 29 protons and 29 electrons. The most abundant isotope of
copper also contains 34 neutrons, however neutrons will play no part in this course; in fact we will
pay little attention to protons either, since this is a course all about electrons. Electronics makes
use of our ability to control the motion of electrons, providing us with the ability to transfer energy
and information over distance scales from the sub-micron (e.g. within an integrated circuit) to the
country-wide (e.g. the National Grid).
1.1 Charge
Charge is a fundamental property of the electron and proton which manifests itself as a force on
the particle when it is exposed to an electric field. The magnitude of the electron and proton charge
is e = 1.6 × 10−19 C; the electron charge is defined to be negative and the proton charge positive
therefore overall our copper atom is electrically neutral. The unit of charge is the coulomb (symbol
C) which is an SI derived unit.
Coulomb’s Law
Consider a charge Q at the origin and a second charge q located at a distance r along the positive
x axis. Assuming Q and q are like charges, q will experience a force
1 Qq
F~ = î
4π0 r 2
and according to Newton’s laws, Q will experience a force −F~ . Clearly, like a compressed spring,
this arrangement is storing energy. If the charges are unlike, the forces act to pull the charges
together (like a stretched spring).
1
Current
The coulomb is a SI derived unit. The relevant SI base unit for our studies is the ampere (or ’amp’,
symbol A) which is a measure of current. The coulomb is derived from the amp as follows:
One coulomb is the amount of charge transported by a constant current of one
amp flowing for one second. Hence 1 C = 1 As.
Current is ’rate of flow of charge’ such that
dq
I=
dt
where I is the current in amps. If we observe a point on a wire, and find that 1 C flows past that
point each second, then the current is 1 A.
1 V = 1 J/C
2
The volt is another SI derived unit and involves resistance for its definition, so we will return to this
later. Before leaving this section it is important to note that the strength of the electric field is
related to the spatial gradient of the electric potential
dV
E=− (3)
dr
We can conclude:
• A positive charge will experience a force tending to move it from a higher
electric potential to a lower electric potential.
• An external agent would have to do work on the charge to make it go
from lower to higher potential
Gravitational Analogy
This situation is exactly analogous to a mass in a gravitational field. To raise a mass m by a height
h requires us to do some work mgh. Unless held in place, the mass will naturally act so as to reduce
its gravitational potential energy. Here, we assume that the gravitational field is uniform over the
distance h so we don’t have to do any integration. Uniform electric fields will be quite common
on electronics, so the simple gravitational analogy can be useful in the understanding of electrical
circuits.
A complication is that electric charges can be both positive and negative. So we can also write
• A negative charge will naturally tend to move from a lower electric poten-
tial to a higher electric potential
• An external agent would have to do work on the charge to make it go
from higher to lower potential
3
because both sheets act to pull a charge in the same direction. This field is again constant. Outside
the sheets, the fields act in opposite directions and so cancel each other out and the net field is
zero
E=0
Another way of thinking about this is to consider that, at some distance from the sheets, the net
charge is zero. These rather surprising results need Gauss’s theorem to fully explain, and since this
is the highlight of the E&M course we won’t spoil that by going into this in any detail, but the
relevance for electronics is this:
∆U = Uab = Ua − Ub = W
where Ua is the system’s energy at a, Ub at b, and Uab means the difference in electrical potential
energy between q at a and q at b.
To move the charge from b back to a, some external agent would have to do an amount of work
W to increase the system’s energy
4
Conservative Field
It’s worth pointing out that the electric field is conservative so it doesn’t matter what path the
charge takes to get from b to a. Work is only done against the component of the force parallel to
the direction of travel. We can write this as
Uba = −Uab
Gravitational Analogy
Again, this is quite analogous to a mass m in a uniform gravitational field g. Potential energy
gained by raising the mass through height d (by some external agent) would be returned by the
gravitational force as the mass falls back to y = 0.
Voltage
Electrical potential (an absolute quantity) and electrical potential difference (a relative quantity)
are well-defined. However, in electronics, we will typically use the term voltage. Unfortunately,
voltage is more loosely-defined and can often refer to either the electric potential at a point in a
circuit or the electrical potential difference between two points. This usage is so common that it is
impossible to avoid and we just have to get used to it. So long as we understand the context, this
will not cause any problems.
Voltage is synonymous with electrical potential difference. One joule is the
energy required to move one coulomb of charge through an electrical potential
difference of one volt.
1J
1V =
C
Thus, the voltage measured across the terminals of a component - say a resistor - is the energy
required to move one coulomb of charge through the resistor.
But what is the ’voltage’ at either of the resistor’s terminals? It depends on how we define the
’zero’ of electrical potential. In section 1.3 we saw that point charges have zero potential when
5
they are infinitely separated. This is simply a matter of definition, and for electronics we will define
a point in our circuit to be the zero of electrical potential and this point is called the ground.
In this way, the use of the term voltage is acceptable and self-consistent, but it requires a firm-
understanding of the circuit’s ground potential. We will look into this in more detail in Lecture
2.
While this may seem confusing, the use of conventional current is ubiquitous in
electronics and we have to remember that currents in electronics are taken to
be positive charges flowing from higher to lower electric potentials. Rarely, you
may need to think about what the electrons are actually doing, and always it is
the opposite direction to conventional current.
6
It is conventional to use the letter I to represent current and q for charge so we can write
dq
I=
dt
The Ampère
When one coulomb of charge flows past a point in one second then the current is one amp
1 A = 1 C/s
however as mentioned previously, charge is not a base SI unit but current is and it’s definition is
based on the force associated with the magnetic field generated by moving charges.
One amp is the current that will produce an attractive force of 2×10−7 newtons,
per metre of length, between two straight, parallel conductors of infinite length.
Drift Speed
Consider a short section of wire with length dL, cross-sectional area A, made from a metal where
the conduction electron density is n, and carrying a steady current I. The conduction electron
charge in this piece of wire is dq = nAdLe and assuming it moves left to right at an average speed
vd then the time taken for all this charge to leave the volume is dt = dL/vd . Using I = dq/dt we
can re-arrange to get
I
vd = (6)
nAe
Since a good conductor has a very large value for n, the drift speed is typically of the order mm/s or
less. Note that this represents an electron current moving left-to-right but we would more properly
describe this as a positive conventional current I moving right-to-left.
1.7 Resistance
This picture of conduction ignores the fact that our electronics is typically operated at room tem-
perature which means that the ionic-lattice and the electrons have a random thermal motion. For
the highly mobile electrons, this can be something like 106 m/s, which is about ten orders of mag-
nitude greater than the drift speed. We have to think of the drift speed as being an extremely
slow movement of the average position of the electrons. In an ideal conductor the electrons would
flow with no hindrance but in practice they keep bumping into the atomic ions (also in random
7
thermal motion about their lattice location). Each time an electron impacts an ion it shares some
of it’s small extra energy; the electron slows down and the ion heats up. This means that, in order
to maintain a constant current, we need to keep putting energy into the system. This we do by
maintaining an electrical potential difference across the conductor such that the electrons always
feel some small acceleration to maintain the average drift speed.
The drift speed of electrons is analogous to the terminal velocity of an object falling in air. The
object experiences a constant force due to gravity, but this is balanced by an equal and opposite
force due to the air-drag. The object loses potential energy without gaining kinetic energy, since
energy is dissipated as heat.
Ohm’s Law
This rather simplified conceptual representation gives rise to the idea of resistance of a conductor
as being the voltage necessary to maintain a current of 1 A.
V
R=
I
The unit of resistance is the ohm (Ω). Put another way, as volts have the same dimensions as
joules per coulomb, a wire with a resistance 1 Ω carrying a current of 1 A requires a continuous
energy input of 1 J per second (or 1 W). This results in heating of the conductor.
Current Density
Returning to our idea of a wire with cross-sectional area A we can make an alternative statement of
ohm’s law by defining the current density J as being the current per unit area at any point inside
the conductor.
I
J = = nvd e (7)
A
Formally, ohm’s law states that J ∝ E, the electric field inside the conductor. The constant
of proportionality is the conductivity but it’s more usual to use it’s inverse ρ such that in this
formulation, ohm’s law becomes
E
J= (8)
ρ
Resistivity
ρ is the resistivity of the conductor and has units of Ω m. By the description of resistance given
here, it is clear that resistivity will be a function of temperature as increasing the thermal motion of
the ionic lattice will further impede electron flow. This is modelled as being linear for practical tem-
peratures encountered in real-world electronics and we can write that the resistivity at temperature
T is given by
ρ(T ) = ρ0 [1 + α(T − T0 )] (9)
where ρ0 is the resistivity at T0 (typically 20 ◦ C) and α is the dimensionless temperature coefficient.
Table 1 gives some values.
Note that carbon, a non-metal, has a negative temperature coefficient since heating releases addi-
tional conduction electrons.
Finally, using E = Jρ gives
V Iρ
=
L A
8
Table 1: Resistivities of Conductors
Conductor ρ0 (Ωm) at 20 ◦ C α
Copper 1.7 × 10−8 4.0 × 10−3
Aluminium 2.7 × 10−8 3.9 × 10−3
Tungsten 5.6 × 10−8 4.5 × 10−3
Carbon 7.8 × 10−6 −5 × 10−4
Further Reading
Sears and Zemansky’s University Physics (Young and Freedman) is a good reference text for a
more in-depth treatment of these foundation topics. The relevant chapters are:
Chapter 21 : Electric Charge and Electric Field
Chapter 23 : Electric Potential
Chapter 25 : Current, Resistance and Electromotive Force
9
Lecture 2 Circuit Theory
Topics:
Electromotive force (EMF), ideal voltage sources, ideal circuits, ground poten-
tial, power, equivalent resistance, Kirchhoffs laws, current sources, principle of
superposition, ideal meters.
The idea that maintaining a constant current in a resistive conductor requires a continuous input
of energy raises the question of the energy source.
Conservation of Charge
When we construct a circuit we have a fixed number of conduction electrons which we can cause
to move round in loops when current flows. The circuit neither makes nor destroys charge, it just
moves it round the loop. Remember that the circuit remains overall neutral.
Water Electrical
Mass (kg) Charge (C)
Flow rate (kg/s) Current (A = C/s)
PE per mass (J/kg) Electrical potential (V = J/C)
PE difference per mass (J/kg) Potential Difference (V = J/C)
Resistance (Js/kg2 ) Resistance (Ω = V/I = Js/C2 )
Clearly, the pump requires some effort to drive, and the work done by the pump results in heating of
the resistance. The water flows round the loop, is neither created nor destroyed, and simply serves
as a medium for moving energy from the pump to the resistance.
10
through it. The EMF it produces is fixed and independent of any resistance connected to it, so the
ideal voltage source can provide any current from zero to ∞.
The Battery
An AA battery (strictly speaking an individual cell, as a battery is a series of cells connected
together) has an EMF E = 1.5 V. The battery stores energy is chemical form and releases it into
electrical form as charge passes through. Each coulomb of charge passing through results in 1.5
J being converted from chemical to electrical form. E is determined by the chemistry of the cell,
with different types having different EMF (for example, a NiCd rechargeable cell only generates 1.2
V). The battery is quite a good approximation to an ideal voltage source, and we’ll look at the
deviations later.
Direct Current
DC stands for direct current which refers to a constant, non time-varying current I. Often we talk
about a ’DC voltage source’ by extension, which is a source with a fixed EMF. By convention, DC
quantities are often (not always) given as upper-case variables e.g. I, E or V , while time-varying
quantities are represented by lower-case variables e.g. I = dq/dt for a constant current or i (t) if
time-varying.
For the rest of this section, and then next, we’ll be considering DC Circuits for simplicity, but it’s
important to note that all of the principles introduced can also be applied to time-varying circuits,
also described as AC circuits (for Alternating Current) which will be covered in depth later in the
course.
11
Figure 2: Ideal Circuit
2.5 Power
As introduced in section 1.7, an important result of Ohm’s law is that it requires a source of energy
to drive a current through a resistance, and as a result we observe a difference in potential across
the resistance. In our ideal circuit, the source of energy is the EMF of the voltage-source, which
provides the power to pump current round the circuit.
Energy is Conserved
Consider a charge +q at b. It has zero potential energy, by definition (this point is grounded). As
it moves to a it gains a PE of Eq. It has the same PE at c, which it loses as heat into R as it falls
to d. Importantly, the charge gains no net energy as it passes round the loop, otherwise we would
have created an unlimited source of power, and we would be able to retire; pausing only to collect
our Nobel prize. Overall, the system conserves energy, as well as charge.
12
Power in DC Circuits
In time t a charge q = It has passed through the battery and energy Eq = EIt has been converted
to electrical form.
Per second, this is a power
P = EI
Whenever a charge flows through a difference in potential there is a change in
energy. The power is always the potential difference times the current. The
voltage source delivers energy into the circuit by converting chemical energy
into electrical energy; the resistor removes electrical energy from the circuit
and converts it into heat.
By conservation of energy, the power dissipated in the resistor is the same, which we might write
as
V2
PR = EI = Vcd I = VR I = R
R
Resistors in Series
If we have R1 and R2 in series then and the voltage across them is E then the same current must
pass through them both.
IR1 = IR2
This leads to the general case that the equivalent resistance of series resistors is
Req = R1 + R2 + ...
The overall resistance will always increase, since the current has to pass through a greater quantity
of ’resistive’ material.
13
Resistors in Parallel
If R1 and R2 were to be in parallel then each would have the same voltage across them
VR1 = VR2
The total current must go up compared to a single resistor on its own. This leads to the general
case that the equivalent resistance of parallel resistors is
1 1 1
= + + ...
Req R1 R2
Around any circuit loop, the sum of the potential gains due to EMFs must equal
the sum of the potential drops across the components.
14
We have to be very careful about the sense of the measurements. Imagine working clockwise around
the left-side loop in figure 3. Sitting in the centre, with a meter (black lead in left hand, red in right)
we would measure a positive EMF E across the source and rotating clockwise we’d measure the
voltages across the resistors as negative quantities. The chosen sense of our measurement reverses
the arrows.
E − VR1 − VR2 = 0
Or, in the ’easier form’,
E = VR1 + VR2
ΣI = 0 (12)
At any junction, the sum of currents into the junction must be equal to the sum
of the currents leaving.
In figure 3, at junction c
I = Ia + Ib
also at d
Ia + Ib = I
15
to ∞. This means that if we ’switch off’ the source then it looks like a break in the circuit (often
called an open circuit). For figure 4
VR = IR
While it’s intuitively right that a voltage source like a battery has (ideally) no resistance it’s not
so simple to understand why a current source has (ideally) infinite resistance. One example of a
current source is the van de Graaff generator, which works by moving charge on an insulating belt
from the bottom to the top. The sphere at the top charges to whatever voltage is necessary in
order to force the charge to flow through air in the form of a spark. (If we connected the sphere
to ground through a resistor then we wouldn’t get a spark). The current is carried by the belt, and
it should be clear that if we switch the generator off and measure the resistance of the belt with
an ohm-meter then we would get an almost arbitrarily high reading. Putting this the other way
around, if the belt wasn’t insulating, it wouldn’t be possible to use it to move charge from ground
to a higher potential. So, a high resistance is essential in order for a current source to work.
Superposition
Superposition allows us to consider the effect of each source independently, then get the overall
effect by summation. The principle can be summarised as
For each source in turn, replace all other sources by their circuit equivalent, and
calculate the behaviour due to the remaining source. The overall behaviour is
then the summation of the individual behaviours.
16
Circuit Equivalent
Circuit equivalent means removing the effect of the source without otherwise changing the circuit
topology. This is equivalent to ’switching-off’ the source and replacing it with it’s circuit equivalent.
Voltage Source E = 0 and resistance R = 0 (the source ’looks like’ a short-circuit).
Current Source I = 0 and resistance R → ∞ (the source ’looks like’ a open-circuit).
Ideal Voltmeter
The ideal voltmeter would be connected in parallel across a resistor in order to measure the ’voltage’
(electrical potential difference) across it. So as not to disturb the measurement, the ideal meter has
infinite resistance (recall parallel resistances), which means it draws no current into it’s terminals.
Ideal Ammeter
The ideal ammeter would be connected in series in a circuit in order to measure the current through
that point. So as not to disturb the measurement, the ideal meter has zero resistance (recall series
resistances), which means it drops no potential difference across it’s terminals.
Like real sources, real meters have finite resistances, and we will come back to this in the next part
of the course.
Further Reading
The relevant chapters in Sears and Zemansky’s University Physics (Young and Freedman) are :
Chapter 25 : Current, Resistance and Electromotive Force
Chapter 26 : Direct Current Circuits
17
Lecture 3 Equivalent Circuits
Topics:
Internal resistance, real voltage sources, loading, maximum power theorem,
Thévenin equivalent circuit, Norton equivalent circuit.
with an ideal meter, such that no current is drawn and hence the voltage drop across r is zero.
3.3 Loading
For all practical purposes, connecting any real device or resistor (such as a bulb) to the battery puts
a resistance in series with r and results in a voltage drop over r . If we connect a load resistor RL
across the output then the voltage across the load, using the voltage divider formula, is
ERL
VL =
r + RL
Equivalently,
VL = E − Ir
The reduction in output voltage depends on the relative values of the resistances and clearly it’s
best if the internal resistance is small.
18
This effect is known as loading; whenever a real-world voltage-source is con-
nected to some device which causes it to supply current then the actual volt-
age observed across the terminals will be reduced and the source is said to be
’loaded’.
Any source of EMF can be described as a voltage source, so this includes not only batteries but
also signal generators, lab power-supplies, sensors with a voltage output, dynamos etc.
Efficiency
Current through the internal resistance means that there is some power dissipated in the battery.
If we define efficiency as the power developed in the load divided by the total power in the circuit
(which is the chemical energy converted to electrical energy per second) then the efficiency is
PL V 2 /RL RL
η= = 2 L = (15)
P E /(r + RL ) r + RL
Efficiency is never 100% since any real world source has some internal resistance.
RL = r
This surprising result is called the maximum power theorem. Efficiency is only 50%, but if our
load is a bulb then the brightest it can be is if the bulb resistance equals the internal resistance. We
have to accept that half the power will be wasted inside the battery as heat. An equivalent theorem
crops up across many fields of study from acoustics and optics to the physics of heat transfer.
19
Figure 6: Circuit (top) and its Thévenin Equivalent (bottom)
In figure 6, the two circuits are exactly equivalent; if they were inside ’black-boxes’, with only the
terminals showing, there’s nothing we could do to test which is which. To find the equivalent circuit,
we need to perform two measurements (or calculations) on the original circuit:
1. Find the open-circuit voltage VOC across terminals a and b, equivalent to ’measuring’ with
an ideal volt-meter which draws no current.
2. Find the short-circuit current ISC , which we get by ’measuring’ the current drawn with an
ideal ammeter connected across a to b.
The Thévenin voltage is then
VT h = VOC
and the Thévenin resistance is
VOC
RT h =
ISC
Thévenin circuits are useful ways of representing the outputs of real-world sources such as batteries,
signal-generators and sensors. For context, an AA battery looks like a 1.5 V source in series with a
0.25 Ω resistor, while a lab signal generator has an output resistance RT h = 600 Ω.
20
circuit can be converted into a Norton equivalent. The description of the ’real current source’, in
section 3.3, is a Norton equivalent representation.
The equivalence between Thévenin and Norton representations is given below.
RNo = RT h
VT h
INo =
RT h
Consequently, it’s easy to switch between Thévenin and Norton representations for any source. The
equivalence between the two is illustrated in figure 7; in this figure, the dashed box illustrates how,
from the point of view of a two-terminal ’black-box’, both circuits are indistinguishable.
Generally, it makes sense to represent a source with a low internal resistance (e.g. a battery) by its
Thévenin equivalent. A source with a high internal resistance (e.g. a lab current-source) is most
intuitively represented by a Norton equivalent.
LTSpice Exercises
From Blackboard, download the Python script MaxPower.py and the LTSpice
schematics MaxPowerThevenin.asc and MaxPowerNorton.asc
Figure 8 shows a ’real-world’ voltage source with EMF Vs and internal resistance Rs . Remember
that this is a Thévenin equivalent representation of the real source, which could be a battery such
as a PP3 (Vs = 9 V and Rs = 2 Ω). The voltage source is connected to a ’load’ resistor RL .
VS2 RL
PL =
(Rs + RL )2
21
Figure 8: Voltage source with internal resistance Rs connected to a load resistance RL
VS2
Ps =
Rs + RL
22
We can verify this result using LTSpice. The schematic MaxPowerThevenin.asc is setup with the
load resistance as a variable (shown as {R} on the schematic) and the text of the spice directive
.step param R .1 10 .1 tells the simulation to vary this parameter from 0.1 Ω to 10 Ω in steps
of 0.1 Ω. The .op directive evaluates the DC voltages and currents at each step. Investigate the
voltages and currents in the circuit, as a function of RL . Note that you can plot the power in any
component by holding down the ALT key while clicking the component.
1. Modify the component values to match your Python script and check that the results are
consistent.
2. Convert the Thévenin representation of the voltage source to a Norton equivalent and verify
that the behaviour is identical. The schematic MaxPowerNorton.asc is setup as a starting
point for this.
Further Reading
The relevant chapter in Sears and Zemansky’s University Physics is :
Chapter 26 : Direct Current Circuits
For a more general introduction from an electronics point of view, see also The Art of Electronics
(Horowitz and Hill):
Chapter 1 : Foundations
23
Lecture 4 Capacitors and Inductors
Topics:
Charge stored by the parallel plate capacitor, capacitance, equivalent capaci-
tance of series and parallel capacitors, energy stored, current-voltage relation,
RC circuit, charging/discharging characteristic curves, time-constant, magnetic
flux in a current-carrying coil, induction and self-inductance, inductors and the
current-voltage relation, energy stored.
The capacitor is a device for holding sheets of charge at a separation. This idea was introduced
in section 1.3. Equation 5 shows that the electric field is uniform between the plates, and so
the electric potential varies linearly between the plates, and overall there is an electric potential
difference across the device. We also saw that by separating charges, and raising them through a
potential difference, energy is stored.
4.1 Capacitance
Because the electric field between the plates is proportional to the charge-stored,
there is also a linear relationship between charge and voltage, and the constant
of proportionality is the capacitance, with units of farad (coulombs per volt).
Q
C= (16)
V
Physically, the capacity for charge storage is a function of only the dimensions and material properties
of the capacitor. For the case of the vacuum between the plates, equation 5 applies, and we find that
we can decrease the electric field between the plates by using a material with a relative permittivity
r such that the amount of charge stored per volt increases. Equation 5 becomes
σ
E=
0 r
As E = V /d (d is the plate separation) and σ = Q/A (A is the plate area), this re-arranges to
0 r A
C= (17)
d
Given the magnitude of 0 it should be clear that it’s difficult to make a 1 farad capacitor and
even values above a few mF are hard to achieve. A typical capacitor consists of two sheets of foil
sandwiching a thin plastic dielectric. This can be rolled-up into a cylindrical shape. For smaller
values, other constructions are possible.
24
Material r
Vacuum (by definition) 1
Air 1.0005
Polythene 2.25
Glass 4
Water 80
Capacitors in Parallel
Consider two of these capacitors connected in parallel. Both capacitors will see the same voltage
V across their plates. The parallel arrangement doubles the charge-storing capability (the area of
the plates has been doubled) and the charge stored is 2CV = 2Q. The equivalent capacitance is
2C. From this follows the general rule for capacitors in parallel
Ceq = C1 + C2 + ... (18)
Capacitors in Series
Placing two identical capacitors in series means that the voltage applied is across both capacitors
in series so we can assume that the voltage across each is halved (V /2). Each capacitor appears
to store a charge CV /2 = Q/2 so we might first think that the total charge stored is the same as
for a single capacitor. However, the section of circuit in-between the two capacitors is electrically
isolated from the rest of the circuit and carries a net charge of zero, so the overall charge stored is
only Q/2 and the effective capacitance is C/2. From this follows the general rule for capacitors in
series
1 1 1
= + + ... (19)
Ceq C1 C2
For a more intuitive explanation, consider also the physical arrangement of two parallel plate capac-
itors placed ’back-to-back’. The two inner plates are touching each other, isolated from the rest
of the circuit, and play no active part. What we have is an equivalent capacitor where the area of
the plates is the same as before but the distance between them has doubled, so by C ∝ 1/d the
capacitance halves.
25
Neutrality
In charging the capacitor we have moved charge from one plate to another. If the top plate carries
a surplus of positive charge +Q then the bottom plate has a deficit of positive charge and carries
−Q. A method for achieving this would be to use a source of EMF such as a battery to ’pump’
charge from the bottom plate to the top. Overall, the capacitor remains electrically neutral. There
is, of course, a force between the plates.
4.5 RC Circuit
Since all real sources have some internal resistance it is useful to study the behaviour of the series
RC circuit shown in figure 11, where it is assumed that the capacitor is initially uncharged. Applying
Kirchhoff’s Voltage Law gives
q
E = vR + vC = i R +
C
26
leading to the first-order differential equation for charge
dq q
E =R + (22)
dt C
Note that vC ∝ q and vR ∝ i = dq/dt. Since E is a constant, as the voltage on the capacitor
increases, the voltage across the resistor will reduce. Before solving the equation, we can make
predictions based on the initial and final states of the circuit.
1. At time t = 0, at the instant the switch is closed, the capacitor is uncharged and so vC = 0
and all the EMF appears across the resistor. The instantaneous initial current I0 = E/R.
2. For time t → ∞, vR → 0 and vC → E.
Charging
dq −1
= (q − EC)
dt RC
Z q Z t
dq 0 −dt 0
0
=
0 q − CE 0 RC
q − CE −t
ln =
−CE RC
q = CE(1 − e −t/RC )
The capacitor voltage is found from q = Cvc
27
which satisfies the second prediction. The current in the circuit is found by differentiating the
charge q
dq
i= = I0 e −t/RC (24)
dt
which satisfies the first prediction. The time-constant
τ = RC
is the time taken for the voltage to rise to a factor (1 − 1/e) = 63.2% of the final value. After
3τ , vc = 0.95E. If we now open the switch then the capacitor remains charged to E. We can
now ’switch-off’ the voltage source (set E = 0) and the capacitor will remain charged as the open
switch prevents current from flowing.
Discharging
Closing the switch again (at a ’new’ time t = 0) results in the discharging behaviour
i = −I0 e −t/RC
vC = Ee −t/RC
We can see that the current is now negative, which is to say that conventional current is now
flowing back from the capacitor and into the (switched-off) source.
Energy
At any instant in time, during charging or discharging, Kirchhoff gives:
E = vR + vC
Since power in a component is always voltage × current [J/C × C/s], the instantaneous power in
the source is
Ei = vR i + vC i
When charging the capacitor, the source is delivering positive power to the circuit, part of which is
stored in the capacitor, and part of which is dissipated in the resistor. We know that for t → ∞
the capacitor energy stored will be CE 2 /2 and we can find the energy dissipated in the resistor by
Z ∞ Z ∞ Z ∞
CE 2
ER = vR i dt = 2 2
i Rdt = I0 R e −2t/RC dt =
0 0 0 2
This is equal to the amount of the energy stored in the capacitor.
28
Magnetic Flux Density
Like the electric field, the magnetic field is a vector quantity and its magnitude and direction can
be described by the magnetic flux density B ~ with units of tesla (T ).
Magnetic Flux
~ through an area A
The flux of a field B ~ is
~A
Φ = B. ~ = BA cos θ
B = µ0 nI (25)
where n is the number of turns per unit length. The field must be perpendicular to the current
so is directed along the axis of the solenoid. Real solenoids are not infinite, but equation 25 is a
good approximation at the centre of a real solenoid, as long as the length is at least ten times the
diameter.
29
Energy
Since passing a current through an EMF either stores or releases energy it is clear that changing
the flux through a coil requires work to be done. An ideal circuit carrying a constant (DC) current
will be content to maintain the current. It takes some work to either increase or reduce the current.
This is similar to Newtonian mechanics: an object will continue in constant motion until an external
force does work against the motion.
4.9 Inductance
There is a linear relationship between flux and current and so it is convenient to define the self-
inductance (L) of a coil by
di
E = −L (27)
dt
The unit of self-inductance is the henry (H). If a coil with N turns carrying a current I generates a
flux Φ then its self-inductance is given by
NΦ
L=
I
Applying this to a solenoid, assuming its length x is long compared to its diameter, gives
µ0 N 2 A
L= (28)
x
30
Figure 12: Sense of the EMF with respect to Current
Inductors in Parallel
The inductors must share the current, but each has to maintain the same potential difference across
its terminals, as for resistors.
1 1 1
= + + ...
Leq L1 L2
31
4.12 Energy Stored
In figure 13 our source does work Ei per unit time and all of this energy will be stored in the inductor
as potential energy U.
dU
Ei =
dt
dU di
= Li
dt dt
An inductor energised to a current I0 stores potential energy U.
dU = Li di
Z I0
LI 2
U=L i di = 0 (30)
0 2
LTSpice Exercises
Download the LTSpice schematic ChargingRC.asc and the Python script
ChargingRC.py
Run the LTSpice simulation which applies a 10 V pulse for a duration of 10 s. Plot the current
through the capacitor and observe how, according to equation 24, the current decays exponentially
as the capacitor charges. When the applied voltage returns to zero, the capacitor discharges,
resulting in an equal but opposite current.
32
Figure 14: Capacitor Power: Note the charging/discharging asymmetry
Lecture 5 Oscillation
The RC circuit is governed by a first-order differential equation and so it can never exhibit self-
sustained oscillation. The RL circuit is governed by a similar equation. However, circuits which
contain both inductors and capacitors will be described by a second-order equation, and we will see
that this leads to oscillating behaviour. First, we need to complete our study of the inductor by
describing the RL circuit.
Topics:
RL circuit, energising characteristic curve, time-constant, LC circuit, oscillation,
series LCR circuit, under-damped case, equivalence to the mechanical harmonic
oscillator.
5.1 RL Circuit
Figure 13 does not represent a practical circuit since all real-world inductors have some resistance1 ,
and are usually used in circuits with resistors. As with the RC circuit, it is useful to study the
behaviour of the series RL circuit. In figure 15 Kirchhoff’s voltage law gives
di
E = iR + L
dt
Comparing with the RC circuit (equation 22) we can see that this is a differential-equation of the
same form and so we predict similar exponential behaviour.
1 Note that LTSpice will not allow an ideal inductor and always includes a small amount of series resistance in the
simulation.
33
Figure 15: Series RL Circuit
As i increases so vR increases and di /dt must decrease and so we can predict that for t → ∞,
i → E/R. As with the capacitor we can solve the differential equation to obtain the following
relations for the ’energising’ of the inductor.
E
i= (1 − e −Rt/L ) (31)
R
vL = Ee −Rt/L (32)
The time-constant is
L
τ=
R
5.2 LC Circuit
Figure 16 shows an ideal capacitor connected to an ideal inductor. We will define that the charge
on the capacitor for all time t ≥ 0 is q(t). At some arbitrary time t < 0, with the switch open, the
capacitor is connected to a power supply E (not shown!) and so charged to Q0 = CE. The power
supply E is then disconnected such that the capacitor remains charged. Consequently, we have the
initial condition q(0) = Q0 . At time t = 0, the switch is then closed so that current i can flow.
Kirchhoff requires
v C + vL = 0
34
Note how the sense of the arrow for vC is drawn, in order to ensure consistency with the current
through the capacitor.
q di d 2q
= −L = −L 2
C dt dt
d 2q −q
2
= (33)
dt LC
This is the equation for harmonic oscillation of the charge at frequency
r
1
ω0 =
LC
Since q oscillates, so does the voltage and the current. A solution to equation 33, valid for time
t > 0, is
q = Q0 cos(ω0 t)
From which the current
dq
i= = −ω0 Q0 sin(ωt)
dt
and the voltage
di q
vL = L = −Lω02 Q0 cos(ω0 t) = − = −vC
dt C
We can see that the voltage and current are always π/2 out of phase, and that the oscillation
continues forever, as illustrated in figure 17.
Energy
At time t = 0, the circuit’s energy is stored on the capacitor.
Q20
U0 =
2C
35
Figure 17: Oscillation of the LC Circuit (L = 0.1 H, C = 0.1 F, E = 10 V). From top to bottom:
capacitor voltage; inductor voltage; current; capacitor energy; inductor energy.
36
For time t > 0 the energy is shared between the capacitor and the inductor
q̈ + γ q̇ + ω02 q = 0
where γ = R/L and ω0 , as before, is the natural frequency of oscillation, which we get for R = 0.
The solution, which we won’t go through in detail, was covered in the Vibrations & Waves course,
and starts by anticipating q of the form
q̃ = Ãe α̃t
Under-damped Case
ω02 > γ 2 /4 makes α̃ complex and results in an oscillating form for q̃. This will occur for low values
of the damping term γ, in which case
37
where r
γ2
ωd = ω02 −
4
To find the physical solution, we take the real part of q̃ to get
q = A0 e −γt/2 cos(ωd t + φ)
We can find A0 and φ from the initial conditions. For the case that we start, as described in section
5.2, with q = Q0 at t = 0, then Q0 = A0 cos φ.
Q0 −γt/2
q= e cos(ωd t + φ) (34)
cos φ
which is exponentially-decaying oscillation. Note that the frequency of the damped oscillation ωd
is slightly less than the natural frequency ω0 , which we can interpret by understanding that the
damping slows the oscillation. For the lightly-damped case, with R = C = L = 0.1, the response
is shown in figure 19. Increasing the value of R increases the damping. The dashed line gives the
envelope of the exponential term.
38
Figure 19: Decaying oscillation for the under-damped LCR circuit. The dashed-line shows the
envelope of the exponential term.
39
LTSpice Exercises
Download the Python script LCR.py and LTSpice schematic LCR.asc
The under-damped LCR circuit will exhibit decaying oscillations. Consider a circuit with E = 1 V,
L = 10 mH, C = 100 µF and R = 1 Ω.
1. Calculate γ, ω0 and ωd
2. From equation 34, derive an expression for the current.
3. Using the initial condition that the current must be zero at time t = 0 (the instant the switch
is closed), show that tan φ = −γ/2ωd .
4. LCR.py will plot the circuit current and the voltage across the capacitor. Extend this code
to plot also the resistor and inductor voltages.
5. LCR.asc will simulate this circuit. Check the voltages and currents versus your Python results.
6. For γ 2 ≥ 4ω02 , α̃ will be real and hence the solution is non-oscillatory. Observe the circuit
response for various values of R in
(a) the under-damped case (4ω02 > γ 2 )
(b) the over-damped case (4ω02 < γ 2 )
(c) the critically-damped case (4ω02 = γ 2 )
40
Lecture 6 AC Circuits
Topics:
Alternating current, amplitude, phase, AC power, root-mean square, linearity,
phasors, complex representation
AC Signals
We could imagine an almost arbitrary number of repetitive waveforms. A typical signal generator
such as the ones used in the lab offer three outputs: sinusoidal, triangle and square-wave. The
sinusoidal signal is the simplest case since it represents oscillation at a single frequency, though
as will be shown in the year 2 Fourier course any repetitive waveform can be constructed from a
summation of sinusoidal signals at different frequencies. The characteristics of a sinusoidal signal
such as
v = A cos(ωt + φ) (35)
are summarised below.
Amplitude A, the difference between the maximum and zero.
Peak-to-peak Amplitude 2A, often written Vpp = 2A, the difference between maximum and min-
imum.
√
Root Mean Square (RMS) Amplitude A/ 2 (NB only for a sinusoid), the equivalent DC voltage
which would result in the same power dissipation as a sinusoid of amplitude A.
Period T , the minimum time between repeat of the waveform.
Frequency f = 1/T with units of hertz (Hz).
Angular Frequency ω = 2πf , with units of radians per second.
Phase φ, with units of radians, where −π < φ ≤ π; positive φ shifts the waveform earlier in time.
41
v(t)
a(t)
t →
Figure 1.5: Plots of x(t) = A cos(!t+ ), v (t) = !A sin(!t+ ), and a(t) = !2 A cos(!t+ ).
Note that acceleration is always in the opposite direction to displacement, but its magnitude
is proportional to the displacement. Maximum speed (|v (t)|) is when displacement and
acceleration are zero. Speed is zero when the magnitudes of displacement and acceleration
are at their maxima.
Figure 1.6: A plot of x(t) = A cos(!t + ) showing the amplitude (A), period (T = 1/f ) and
Figure
the shift 20: Amplitude,
backwards in time ( Period and Phase
/!) relative for a .Sinusoidal Signal [V&W notes]
to A cos(!t)
The maximum and minimum values of x are A and A, respectively. A is called the am-
plitude of the oscillation. The motion repeats itself after a time T , known as its period. If,
at any time t0 , its displacement and velocity are x(t0 ) and v (t0 ), respectively, this means
that at all times tm = t0 + mT , where m = 0, ±1, ±2, · · · , it is in precisely the same state:
x(tm ) = x(t0 ), v (tm ) = x(t0 ).
The value of T depends on the argument of cos and sin in the expressions for x(t) and v (t).
Both functions repeat when their argument (!t + ) changes by an integer multiple of 2⇡
) !tm + = !t0 + + 2m⇡ ) t0 tm = 2m⇡/!. The period is the length of the shortest
interval (m = 1) after which the system returns to the same state.
2⇡
T = (1.10)
!
The number of complete oscillations in a time interval t is n = t/T , and so the rate at
42
6.2 AC Power
In the circuit of figure 22 the voltage across the resistor is in-phase (zero phase difference) with
the current. A sinusoidal voltage
v (t) = V0 cos ωt
across a resistor R results in a current
i (t) = I0 cos ωt
By our earlier definitions of voltage and current, equation 36 is always true and represents the
instantaneous power associated with a component. For the particular case of figure 22 we can
further write
p(t) = v i = V0 I0 cos2 ωt
43
from which we observe (figure 23) that the power is a positive quantity varying at twice the frequency
of the applied voltage, with a peak value
V02
Ppk = V0 I0 = = I02 R
R
For AC circuits the instantaneous power is not normally useful so a better measure of the power
dissipated is had by taking the average over 1 cycle
V0 I0 V2 I2
Pav = hp(t)i = V0 I0 cos2 ωt = = 0 = 0
(37)
2 2R 2R
We can get this result knowing that the average value of the cos2 function is 21 , but in the general
case
1 t+T
Z
hf (t)i = f (t 0 )dt 0
T t
where T is the period of f (t). Here, the average power is half the peak power.
Now consider the circuit of figure 24; the voltage across the capacitor is out-of-phase (π/2 radian
phase difference) with the current.
An applied voltage
v (t) = V0 cos ωt
across a capacitor C results in a current
dv (t)
i (t) = C = I0 cos(ωt + π/2)
dt
where I0 = ωCV0 . At any time t the instantaneous power dissipation is, as before p(t) = v (t)i (t),
with an average value < p(t) >= 0. As the ideal capacitor has no resistance it dissipates no average
power; positive instantaneous power is associated with ’charging’ while negative power is associated
with ’discharging’.
The same is true for an AC voltage applied to an inductor.
In any circuit, power is only dissipated in the resistive components. It’s usually
easiest to calculate either the voltage across the resistor, or the current through
it, and the use equations 36 or 37 to find the power dissipation.
44
Figure 25: Capacitive circuit: Voltage out-of-phase with current (arbitrary units)
and
VRMS
IRMS = (40)
R
However, in the more general case (a circuit with some combination pf resistors, capacitors, induc-
tors) there may be a phase difference φ between applied voltage v (t) and resultant current i (t) in
which case
45
Waveform Amplitude RMS
DC A (constant) A√
Sinusoid A (peak ±A) A/√2
Triangle A (peak ±A) A/√3
Sawtooth A (peak ±A) A/ 3
Square A (peak ±A) A
6.5 Phasors
We have already seen in section 5.3 that is was mathematically convenient to assume a complex
solution for the charge q̃ associated with the capacitor in the LCR circuit. There is nothing special
about the use of complex notation; we could use trigonometric functions (sines, cosines) to solve
all possible circuit problems, however the use of a complex representation makes the maths simpler.
46
Vector Representation
A voltage such as v = A cos(ωt + φ) can be represented as a vector rotating anti-clockwise in the
complex plane. In cartesian coordinates, the vector is
ṽ = x + jy = A(cos θ + j sin θ)
where the angle θ = ωt + φ (figure 26). The projection of this vector onto the real (or imaginary)
axis describes a sinusoid. The amplitude of the vector is A and the ’real’ voltage, i.e. that which
we would physically observe and measure, is the projection onto the real axis (x = A cos(ωt + φ)).
According to the linearity of our circuits, all signals will be oscillating at the same frequency,
consequently we can choose to remove the time-dependence and represent the signal simply by its
amplitude and phase. The term phasor (for phase-vector ) is used to describe this representation.
The phasor is a vector in the complex plane with a magnitude representing the amplitude of the
oscillation and an angle representing the phase. This is equivalent to saying that the phasor is a
’snapshot’ of the vector at time t = 0. Thus, in figure 26, the shaded vector at angle φ is the
phasor. We can establish a set of rules for the phasor representation:
1. The phasor is a vector in the complex plane.
2. The magnitude of the phasor is the amplitude of the oscillation.
3. The angle of the phasor is the phase angle φ of the oscillation.
4. The phasor can be thought of as a ’snapshot’ of the oscillation at time t = 0.
5. The physically observable quantity (at time t = 0) is the projection onto the real axis.
6. The same angular frequency ω is assumed for all phasors represented on the diagram.
We will usually plot several phasors on a single phasor diagram since phasors can be drawn to
represent all the oscillating voltages and currents in our circuit. The phasor diagram gives a simple
view of the relative phaes and amplitudes of these quantities. The phasor diagram makes it much
easier to see which quantity leads/lags which (compared to, for example, figure 21).
√
Note the use of j for the imaginary unit −1. This is common in electronics
and electrical engineering to avoid confusion with i for current.
v = V cos(ωt + φ)
47
Figure 26: Complex representation of an oscillating quantity; for constant ω, the projection onto
the x-axis is A cos(ωt + φ)
where oscillation at a frequency ω is assumed. Since e jπ/2 = j we can create a new vector v˜0 = j ṽ
with the appropriate rotation. To get the real physical observable associated with this new vector
we simply multiply by e jωt to include the time-dependence, and take the real part.
48
To summarise, we create a phasor representation of of a signal as follows:
1. Express the signal as a complex exponential.
2. Remove the time-dependent part.
3. The result will be a vector in the complex plane with the length of the
signal’s amplitude and the angle of the signal’s phase. For example
v = V cos(ωt + φ) → ṽ = V e jφ
Operation Multiply by
Scale by A (without rotation) A
Rotation by π/2 e jπ/2 (or j)
Rotation by π e jπ (or −1)
−jπ/2
Rotation by −π/2 e (or −j)
Rotation by arbitrary angle θ e jθ
LTSpice Exercises
Download the Python script sawtooth.py and LTSpice schematic
drivenLCR.asc
The Python script aims to show that the average power generated in a resistor by an arbitrary
voltage waveform is the same as the power generated by a DC voltage of the waveform’s RMS. To
illustrate this, the script generates a sawtooth waveform, calculates the power in a 10 Ω resistor
and then finds the average by integration. Importantly, it then calculates the equivalent DC voltage
which would generate this power, and then compares this to the waveform’s theoretical RMS value.
1. Run sawtooth.py
49
2. Change the amplitude and/or period and observe that the results remain consistent with
theory.
3. You could try modifying the waveform shape (e.g. square-wave, sinusoid) but note that line
50 will need to be updated for consistency with table 6.3.
The LTSpice schematic drivenLCR.asc drives the series LCR circuit used previously with a periodic
sinusoidal waveform.
50
Lecture 7 Impedance
Topics:
Reactance of capacitor and inductor, complex impedance, equivalent
impedances, AC analysis
Our aim here is to produce a generalised form of Ohm’s law which includes also capacitors and
inductors. Returning to the discussion of linearity in section 6.4, we have
vR = i R R
diL
vL = L
dt
dvC
iC = C
dt
7.1 Reactance
For the resistor, the amplitudes of the voltage and current are linearly related, and of course the
current is always in-phase with the voltage. The constant of proportionality in this linear relation is
the resistance, with units of ohms. The equivalent quantity for capacitors and inductors is called
the reactance, also with units of ohms.
|vc | 1
XC = = (43)
|iC | ωC
51
7.2 Complex Impedance
Equations 43 and 45 would be sufficient if we just want to know about the amplitudes of the signals,
but we can use the complex representation to additionally encode the phase changes involved.
Z̃R = R (48)
Note that the resistor impedance is entirely real, representing the in-phase nature of the voltage
and current signals, while the capacitor and inductor impedances are entirely imaginary, encoding
the π/2 phase shifts between voltage and current.
52
Impedances in Series
The equivalent impedance of series impedances is
Z̃eq = Z̃1 + Z̃2 + ... (49)
For example, in figure 27 (left), the combined total impedance Z̃ is given by
j
Z̃ = Z̃R + Z̃C = R −
ωC
Impedances in Parallel
The equivalent impedance of parallel impedances (e.g. figure 27, right) is
1 1 1
= + + ... (50)
Z̃eq Z̃1 Z̃2
Consequently, combining resistances with reactive components results in an equivalent impedance
with both real and imaginary parts.
7.4 AC Analysis
As long as we follow all the rules for manipulation of complex quantities, we can use complex
impedances just like resistances with Ohm’s law. Using the phasor representation for voltage ṽ and
current ĩ
ṽ
Z̃ =
ĩ
So we can find an unknown voltage from a current by
ṽ = ĩ Z̃
53
Circuit Rules
The circuit rules of section 2 can be extended to AC signals. Kirchhoff’s voltage law always applies
in that the instantaneous sum of voltages around any loop must always be zero. On a phasor
diagram, this means that the vector sum of voltages is zero, i.e. both the real and imaginary
voltage components sum to zero. We can also plot currents as phasors and Kirchhoff’s current
law also applies; the vector sum of currents into a junction is zero. We can also use the complex
equivalents of the voltage and current divider rules. The use of these principles is best understood
by some examples as we will see in section 8. Finally, the principle of superposition still applies, and
we can make Thévenin equivalent circuits with complex impedances.
Using complex impedances requires care and a bit of practice, but so long as we follow all the rules
for complex numbers and the relationships described above, the results fall-out remarkably easily.
The expressions for complex impedance (equations 46 to 48) are usually easiest to remember in
the cartesian form, but it’s often easiest to convert combined impedances into exponential form in
order to simplify the maths.
LTSpice Exercises
Download the LTSpice schematics drivenC.asc and drivenL.asc from Black-
board
The first schematic shows a capacitor C = 1000 µF driven by a source of the form v = 10 cos 100t.
1. What’s the capacitor’s impedance at this frequency?
2. Given the magnitude of the impedance and the amplitude of the applied voltage, predict the
amplitude of the resulting current.
3. Run the simulation to verify this. Plot the current through the capacitor.
4. Write the applied voltage in phasor form and, using ĩ = ṽ /Z̃C , show that this method gives
the amplitude and phase of the current consistent with the simulation.
5. Does the current lead or lag the voltage, and by how many radians?
The second schematic shows an inductor L = 1 mH driven by a current source of the form i =
10 cos 100t.
1. What’s the inductor’s impedance at this frequency?
2. Predict the amplitude of the resulting voltage.
3. Run the simulation to verify this. Plot the voltage across the inductor.
4. Write the applied current in phasor form and, using ṽ = ĩ Z̃L , show that this method gives
the amplitude and phase of the voltage consistent with the simulation.
5. Does the current lead or lag the voltage, and by how many radians?
54
Lecture 8 Filters
Topics:
Impedance divider, gain, Bode plot, decibels, roll-off, low-pass and high-pass
RC filters, low-pass and high-pass RL filters, filter applications.
We will now re-visit the series RC circuit in order to understand its behaviour in response to AC
signals. The frequency-dependence of the capacitor’s impedance implies that we can use the RC
circuit as a filter, that is to say, a circuit which will pass through some frequencies and block others.
The RC filter of figure 28 consists of a resistor and capacitor in series; the input signal Vin is applied
to the series combination while the output Vout is the voltage across the capacitor alone. Since the
capacitor’s impedance is inversely proportional to frequency, we can make some predictions about
the behaviour:
1. For low frequencies, ω → 0, the capacitor’s impedance |ZC | → ∞ so no current will be drawn
from the supply, there will be no voltage drop across R, and hence vout = vin .
2. For high frequencies, ω → ∞, the capacitor’s impedance |ZC | → 0 so there will be a voltage
drop across R equal to vin , and hence vout = 0.
Consequently, figure 28 is described as a low-pass filter since it passes low-frequencies and blocks
high-frequencies.
8.2 Gain
The filter is usually characterised by the gain, which is the ratio of output to input voltages.
ṽout 1
g̃ = = (51)
ṽin 1 + jωRC
It is useful to know the magnitude and argument of the complex gain.
1
g = |g̃| = p (52)
1 + (ωCR)2
tan φ = −ωRC
φ = − tan−1 ωRC (53)
Equation 52 validates the two earlier predictions about the behaviour for low and high frequencies.
55
Figure 28: Low-Pass RC Filter
Decibels
The logarithm of the gain magnitude is, by convention, expressed in dB (decibels).
Cut-off Frequency
All low-pass RC filters will have the same general shape of figure 29. For low frequencies, g is always
1 (0 dB) but the frequency at which the filter starts to attenuate can be adjusted by changing the
56
Figure 29: Magnitude and Gain of the Low Pass Filter (normalised to RC = 1).
57
values R and C. Since there is no sharp transition from the pass-band to the roll-off, by convention
we choose the frequency where the output power is half the input which can be found by solving
1 1
=
2 1 + (ωc CR)2
Roll-Off
An ideal low-pass filter would pass all frequencies below the cut-off and block all frequencies above.
This is impossible to achieve in real-world filters. For the RC low-pass filter, it can be seen that for
ω ωc ,
1
g≈
ωRC
which results in a gradient of -20 dB/decade, where a decade is a factor of 10 in frequency. For
example, the gain falls by -20 dB between the frequencies 10 and 100 rad/s.
Phase
Conventionally, the phase of the Bode plot is expressed linearly in degrees. The phase of the low-
pass filter expresses the phase change between the output versus the input and the negative sign
indicates that the phase of the output lags the input. We can see this by re-arranging equation 51:
ṽout = ṽin g̃
Taking an arbitrary input signal ṽin = V0 (i.e. an input signal with amplitude V0 , phase angle zero,
and an assumed frequency ω) then the output will be
ṽout = V0 ge φ
where g and φ are functions of R, C and ω and defined by equations 52 and 53 respectively. For
the low-pass RC filter of figure√ 28, φ will be negative. The amplitude of the output sinusoid is
vout = V0 g and vout RMS = V0 g/ 2.
ωRC
g=p
1 + (ωRC)2
58
φ = tan−1 ωRC
The cut-off frequency is again
1
ωc =
RC
For low frequencies, g ∝ ω (the gradient is +20 dB/decade) while for ω ωc , g → 1.
8.5 RL Filters
Since the impedance of the inductor is also frequency-dependent, we can build RL filters with
low and high-pass characteristics. In the circuit of figure 28 the capacitor can be replaced with an
inductor. |ZL | ∝ ω so this is a high-pass RL filter. The same principles, starting with the impedance
divider, can be used to derive the gain of the RL filter, and again we can make a low-pass RL filter
by swapping the resistor and inductor.
Audio Example
The source could be a microphone, which is a sensor with characteristics of a voltage source. The
microphone might be sensitive to frequencies up to 50 kHz, but the human ear can’t hear hear much
above 16 kHz or so. Consequently, it makes sense to ’filter-out’ all frequencies above 16 kHz before
amplifying the signal. The amplifier, typically known as a ’pre-amp’ in audio applications, would
appear as a load resistance. If we wish to study the setup, the circuit would look like figure 31.
The microphone is represented as a voltage source VS (note that in practice this may have some
small output resistance, typically RS = 150 Ω or so). The pre-amp is represented by a load resistor
RL . A typical pre-amp would have an input resistance RL = 10 kΩ. The time-constant τ = RC
would be chosen to give a filter-cut-off frequency about 16 kHz. Typically, we would use a value of
R which satisfies RS R RL so as to avoid the filter loading the source and the ’load’ loading
the filter.
59
Figure 31: A low-pass filter used in an audio example. The source may be a microphone and the
load may be a pre-amplifier.
LTSpice Exercises
Download the LTSpice schematics LowPassFilter.asc and LowPassFilter-
Bode.asc
The schematic of LowPassFilter.asc uses components from the set of standard values for resistors
and components to build a low-pass filter with a cut-off frequency at approximately 16 kHz.
1. Run the simulation and plot the applied voltage and the output voltage. At 1 kHz (well below
the cut-off) the traces should be very similar.
2. Edit the source properties to increase the frequency in factors of two (2 kHz, 4 kHz, etc. . . 32
kHz) and note the approximate amplitude of the output at each frequency.
√
3. Check the result that, at the cut-off frequency, the output amplitude should be a factor 1/ 2
times the input (equivalent to -3 db).
4. Again at the cut-off frequency, make an estimation of the phase difference between input and
output. Which leads? Does this agree with theory?
5. Construct a phasor diagram showing the filter at the cut-off frequency.
To make the simulation more ’realistic’ you could repeat the above after placing a 10 kΩ ’load’
resistor in parallel with the capacitor, between Vout and GND, and check the result that the power
in the load resistor reduces with frequency (to about one-half at the cut-off).
LTSpice can automate these frequency scans to create a Bode Plot. Open the schematic Low-
PassFilterBode.asc.
1. Note the use of the .ac dec. . . spice directive, which tells the simulation to perform an AC
analysis.
2. Run the simulation and plot the output voltage. You should get a Bode Plot with the gain
magnitude (in dB) and the gain phase (in degrees) in the same plot panel.
3. Click on the V(vout) legend above the plot to bring up cursors and find the frequency at
which the gain is -3 dB. Find also the phase at this frequency, and check agreement with
theory.
4. Right-clicking on the left-hand vertical scale allows changing the scale between dB, logarithmic
and linear. Check that you are able to interpret gain in both linear and decibel representations.
60
You could also experiment with a high-pass filter, and/or filters built using inductors instead of
capacitors.
61
Lecture 9 Driven LCR Circuit
Topics:
Equivalence to the mechanical oscillator, Steady-State AC Analysis, Resonance.
We will now look again at the LCR circuit; here the circuit will be connected to an external EMF
source which makes the driven LCR circuit the exact physical analogue of the forced mass-spring-
damper mechanical system. The series LCR circuit is shown in figure 32. Kirchhoff gives
v = vL + vR + vC
Assuming that the circuit is driven by a voltage source with an EMF of the form v = V0 cos ωt
di
V0 cos ωt = L + Ri + vC
dt
Rewriting in terms of charge q
q
V0 cos ωt = Lq̈ + Rq̇ +
C
Using γ = R/L and ω02 = 1/LC gives
V0
cos ωt = q̈ + γ q̇ + ω02 q (55)
L
62
Table 6 summarises the equivalences. The general solution to equation 55 was covered in the
Vibrations & Waves course and will not be part of this course but it is useful to remember that the
general solution was a superposition of the transient (or ’startup’) behaviour with the steady-state.
We will not look at the transient behaviour but will see how we can apply AC circuit analysis to
understand the circuit’s steady-state behaviour.
Mechanical Electrical
Displacement x Charge q
Velocity ẋ Current q̇
Mass m Inductance L
Spring constant k Capacitance 1/C
Damping b Resistance R
Force F0 EMF V0
ṽR = ĩ R
ṽL = ĩ Z̃L
ṽC = ĩ Z̃C
Since the phase angle of the applied voltage ṽ is defined to be zero, the arguments of the complex
current and voltages are their phase angles relative to the applied voltage. It is helpful to plot these
on a phasor diagram, and it will be found that the magnitudes and phases will all be frequency-
dependent.
Frequency Response
The magnitude and phase of the current are represented in figures 33 and 34. It is instructive to
go through the complex algebra to find expressions for |ĩ | and φ assuming ĩ = I0 e jφ . For simplicity,
the figures 33 and 34 are calculated for R = L = C = 1.
63
Figure 33: Magnitude of the Current versus Frequency
64
as the voltage across the resistor, which will always be in-phase with the current. The LCR circuit
is useful for tuning applications where a particular frequency needs to be extracted from a signal.
Tuning circuits in radios used to use this technique where a variable-capacitor was used to adjust
the tuning frequency ω0 .
9.3 Resonance
For ω → 0 the impedance of the capacitor |Z̃C | → ∞ and hence |ĩ | → 0. For ω → ∞ the impedance
of the inductor |Z̃L | → ∞. The real (resistive) part of Z̃ is fixed but the imaginary (reactive) part
is a function of frequency and it is clear that |Z̃| will be minimised for the frequency at which Im(Z̃)
is zero:
j
jωL =
ωC
which occurs at the natural frequency
r
1
ω = ω0 =
LC
Hence, when the voltage source is oscillating at the natural frequency ω0
• The impedance is at a minimum.
• The current is at a maximum.
• The impedance is entirely real and equal to the resistance R
• |ĩ (ω0 )| = V0 /R
• The voltage applied across the LCR series circuit will be in phase with the current.
• The voltage across the resistor will be at a maximum.
• The voltage across the inductor and capacitor will be equal in amplitude but opposite in phase.
Power Dissipation
Over a complete cycle, the average power of the inductor and capacitor are zero. Only the resistor
dissipates power according to
p(t) = vR (t)i (t) = i (t)2 R
We know that the voltage and current in the resistor are always in phase. The average power
dissipated is then
I 2R
hp(t)i = 0
2
Q-factor
The shape of the curve in figure 33 is characterised by a dimensionless parameter Q known as the
Q-factor or quality-factor.
ω0
Q= (56)
∆ω
∆ω is a measure of the width of the peak so it is clear that a high Q-factor is associated with a
narrow peak. ∆ω is defined as the difference in frequencies between the points where the power
dissipated falls to one-half of the maximum. This is illustrated in figure 35 which shows the average
65
power dissipated in an LCR circuit for two different Q-factors. It is left as an exercise to show that,
for an lightly-damped LCR circuit (i.e. γ 2ω0 )
R
∆ω ≈ γ =
L
Consequently, reducing the resistance in an LCR circuit increases the ’sharpness’ of the resonance
peak.
ṽ = 1
Z̃R = 0.1 Ω
Z̃L = j Ω
Z̃C = −j Ω
Therefore the total series impedance is
Z̃ = 0.1 Ω
66
We can find the current to be
ṽ
ĩ = = 10 A
Z̃
The current phasor is entirely real and in phase with the voltage. We use the current to find the
voltages across the components:
ṽR = ĩ Z̃R = 1 V
ṽL = ĩ Z̃L = 10j V
ṽC = ĩ Z̃C = −10j V
We can see that at resonance, the voltage across the resistor is equal to the applied voltage in both
magnitude and phase, which we expect as the impedance is entirely real (resistive). The amplitudes
of the voltages across the reactive components are much larger, equal, and π radians out of phase
with each other. It is helpful to visualise on the phasor diagram (figure 36).
To get the measurable, physical signal, we can add the time-dependence and take the real part
vR = Re 1e j0 e jωt = cos(10t)
n o
vL = Re 10e jπ/2 e jωt = 10 cos(10t + π/2)
n o
vC = Re 10e −jπ/2 e jωt = 10 cos(10t − π/2)
|ṽR |2
hP i = = 5W
2R
67
Figure 37: LCR Circuit Voltages at Resonance (applied voltage 1 V)
Theory predicts that the power should drop to approximately half this value at a frequency
γ
ω = ω0 ±
2
To test this, let’s choose ω = 9.5 rad/s.
Z̃R = 0.1 Ω
Z̃L = 0.95j Ω
−j
Z̃C = Ω
0.95
Total impedance
Z̃ = Ze jφ
where Z = 0.143 Ω and φ = −0.798 rad.
ṽ
ĩ =
Z̃
0.1e −jφ
ṽR = ĩ Z̃R =
Z
|ṽR | = 0.698 V
|ṽR |2
hP i = = 2.4 W
2R
This is not exactly half because of the approximation ∆ω ≈ γ Completing the maths for the other
two voltages results in the phasor diagram shown in figure 38, and for which the physical signals
look like figure 39.
68
Finally, note that the vector sum of the three voltage phasors equals the applied voltage
This is required since their real components must follow Kirchhoff’s voltage law
LTSpice Exercises
Download the LTSpice schematics WorkedExample.asc and
WorkedExampleBode.asc from Blackboard
This schematic WorkedExample.asc implements the circuit described in section 9.4. We can use
this to verify the theory and also some of the predictions for the forced damped-oscillator given in
section 4 of your Vibrations and Waves course notes.
1. Run the simulation and click a component to plot the circuit current.
2. Right-click the plot window and add a new ’plot pane’. Plot the voltage across the capacitor.
3. To plot the inductor voltage it is necessary to perform a simple calculation which can be
done by right-clicking again and selecting ’add trace’ then build the expression ’V(n001)-
V(n002)’. Check the current and voltages agree with the calculations in section 9.4 and that
the capacitor/inductor voltages are π rad out of phase.
4. Similarly, add and check the resistor voltage - this should be (in the steady-state) identical to
the driving voltage.
69
Figure 39: LCR Circuit Voltages at Half-power
CC
13-Jan-18
70