2022

SMPS Design Laboratory

Geoff Hilton
University of Bristol
23rd September 2022
Version 1.0
Table of Contents
1. SWITCHED MODE POWER SUPPLY DESIGN OVERVIEW .............................................................................. 1
1.1 Introduction ............................................................................................................................................... 1
1.2 Resources .................................................................................................................................................. 2
1.3 Group Organisation and Assessments....................................................................................................... 2
1.4 Keeping notes for this design .................................................................................................................... 3
1.5 Practical Laboratory work.......................................................................................................................... 3
1.6 Design Specifications ................................................................................................................................. 4
2. INTRODUCTION TO CIRCUIT MODELLING USING MATLAB SIMULINK AND SIMSCAPE .............................. 5
2.1 The series LCR resonator circuit ................................................................................................................ 5
2.2 Quick overview of time-domain modelling ............................................................................................... 6
2.3 Simscape modelling ................................................................................................................................... 7
2.3.1 Finding components and making interconnections ........................................................................... 7
2.3.2 Setting component values – typical example but will be different for each component. ................. 7
2.3.3 Labelling components......................................................................................................................... 7
2.4 Simulink components and interface .......................................................................................................... 8
2.4.1. Adding Scope and Solver Configuration blocks ................................................................................. 8
2.4.2. Solver Configuration and Configure Parameters ............................................................................... 8
2.4.3 Changing Scope settings and data logging ......................................................................................... 9
2.5 Practical versus simulation ...................................................................................................................... 10
2.5.1 Scope Probes .................................................................................................................................... 10
2.5.2 Component Polarity.......................................................................................................................... 11
2.5.3 Preferred component values ............................................................................................................ 11
2.5.4 Maximum operating conditions of components .............................................................................. 11
3. CIRCUIT DESCRIPTION AND TASKS ........................................................................................................... 12
4. CIRCUIT MODELS ...................................................................................................................................... 13
4.1 Full-wave rectification of AC supply voltage ........................................................................................... 13
4.2 PWM generator ....................................................................................................................................... 14
4.3 Gate-drive circuit ..................................................................................................................................... 15
4.4 Completing the Main Transformer Primary Winding circuit ................................................................... 17
4.4.1. MOSFET switching ........................................................................................................................... 18
4.4.2. MOSFETs, Freewheeling Diodes and Main Transformer primary winding inductance. .................. 19
4.4.3 Nonlinear Inductor ........................................................................................................................... 19
4.4.4 Heatsink requirements for active components ................................................................................ 21

1
 4.5 Secondary Circuit and ‘100KHz’ filter ...................................................................................................... 21
4.6 Complete open loop SMPS circuit ........................................................................................................... 23
4.7 Closed loop control and other considerations ........................................................................................ 24
5. PWM Practical Laboratory ......................................................................................................................... 25
Appendix A: Laboratory Notebook ............................................................................................................... 26

Document Version
This document may need to be updated throughout TB1 if more clarification of sections is required.

This table details these updates.

Version Date Comment

1.0 23rd September 2022 Initial document on BB

2
1. SWITCHED MODE POWER SUPPLY DESIGN OVERVIEW
1.1 Introduction
Switched Mode Power Supplies (SMPS) provide a constant, stabilised DC voltage for a wide range of
electrical and electronic equipment such as computers, televisions etc. For the second year SMPS design
exercise, students have to design a highly efficient regulated DC power supply that is fed from a 50Hz AC
supply. The design exercise considers the use of a variety of electronic components as well as the practical
implementation of electronic control techniques in order to provide students with an appreciation of the
challenges and compromises involved with any product design.

The SMPS design exercise covers a wide variety of topics found in Electrical and Electronic Engineering,
including:

• Power management
• Classical control theory
• Use of power semiconductor devices
• System modelling
• Mixed analogue and digital ICs
• Magnetics
• Filter design
• Thermal management
• Measurement techniques
• Practical limitations of real components
• EMC and power quality
• Use of data sheets

A block diagram of the system is shown in Figure 1.1. Low voltage digital and analogue circuitry, and high
current (high voltage) analogue circuitry are used in the design. The input is 50 Hz AC (30 Vrms in this case),
while the DC output is designed to supply between 9 V and 18 V into a resistive load between 2 Ω and 21 Ω.
The resistive load has to the switched to show how the supply copes with a change in current load.

Open Loop SMPS

AC Rectification
PWM and Power Output Filter
Electronics ‘100KHz’
AC INPUT
50Hz supply
+
-
Feedback
Set voltage
output level DC Link DC OUTPUT
Resistive Load
Stabilised DC voltage output
Figure 1.1: SMPS overview including closed loop control.

1
The majority of the work will be a simulation exercise using Matlab Simulink/Simscape combined with
practical work on one aspect of the design. The main objectives of the project would not be significantly
different to a practical SMPS design and build, only that you will build and test a software design rather than
a hardware design. There are differences - getting hardware to work correctly is undoubtedly more difficult
than getting software to produce ‘something’, through this is not necessarily the correct ‘something’. It is
possible to produce a completely meaningless output in software (assuming the model is ‘stable’ and able to
run in the first place).

Even though a simulation-based exercise would generally be significantly less time consuming than a
hardware-based exercise only 3 out of the 4 design stages (shown in Figure 1.1) will be formally required and
assessed (the feedback loop will not be). There is opportunity to complete the full design with the feedback
loop, but support will not be available. In addition to the simulation work, you will be required to complete
the practical design of the Pulse Width Modulator (PWM) circuit.

1.2 Resources
The design exercises uses a series of (mainly) audio PowerPoint videos. These are used to explain the
operation of the circuit and how to do a ‘paper’ design (i.e. specify the value of the components needed for
correct operation). In a hardware design, these component values would then be used in the practical
circuit build, while here the simulation environment will be used to verify the correct calculation of the
components and the correct operation of the circuit. The PowerPoints include information relating to a
practical supply build as any simulation should (where possible) match that of the practical design needs.
This is not always possible and any simulation will only be an approximation of the practical design. The
emphasis is therefore on your understanding of the circuit operation and ability to produce a workable
design.

Some components are specified. These are usually the active components (diodes or MOSFETs), and you will
need to look up the device parameters in data sheets. Other components you are free to choose depending
on the design. These are the passive components (resistors, capacitors and inductors). As with any design,
the initial calculations may not work as expected (or only give a first-order estimate) and hence these can be
fine-tuned (and optimised) using the model.

The following PowerPoints give an overview of the SMPS:

1. Introduction.mp4
2. How does SMPS work.mp4

There is a lot of information to take in here, so at this stage all that is required is a general ‘feel’ for what the
project is about. The design itself is done in ‘bite-size’ stages so that you will build up a full understanding of
the circuit operation by the time you complete the final tasks.

1.3 Group Organisation and Assessments

This is a group design exercise and normally each group will comprise 5 students. Each student has to
undertake their own design based upon the information provided. Then as a group you will discuss the
design options and agree on a final choice of components – this is the same general process that would be
required for a practical design project. There is support for this unit from Teaching Assistants (TAs) and
academic staff. Each group will be able to meet both online and in-person at timetabled sessions.
Furthermore, each group will have regular meetings with the Unit Organiser to discuss progress and any

2
further clarification with the design and expected outcomes. The group progress will be assessed and forms
part of the overall SMPS mark.

There are two assessments: Individual (60% weighting) and Group (40% weighting).

The individual assessment comprises an online Blackboard multiple choice test. The design will need to be
completed and fully written up in your notebooks by the time that the individual Blackboard test has to be
undertaken. The test will last 1 hour and is envisaged to take place during a timetabled slot in TB1, though
there may be a need to organise some special arrangements. This is an Open Book test with the ‘book’ being
the Laboratory book as detailed in Section 1.4. Students must use only their OWN notebook and the use of
another student’s notebook will be regarded as CHEATING.

For the Group mark, each group will have to fill in designs and results in a worksheet (one per group) based
on each of the models developed in the laboratory and discuss the results with the Unit Organiser.

The information required for the worksheet is typically:

• Circuit diagram (with components clearly labelled on the circuit)

• Brief description of operation
• Waveforms to demonstrate working part of the system and comment on performance
• Choice of components and questions

An Introductory Talk (with Q&A) will be scheduled for all students in Week 1 with the unit formally starting
in Week 2. The timetable for support and meetings with TAs from Weeks 2 to 12 will be posted on BB.

1.4 Keeping notes for this design

All students must keep a note of their design in their own ‘Laboratory Notebook’. The notebook is very
important. It is the equivalent of the Engineers Notebook that you keep if you are working in industry. It is
important that you record the steps you take and the design calculations that you make so that you have a
continuous record that you can refer to as a design develops. The individual notebooks will be required for
discussions with Unit Organiser and the TAs and are used for the Individual Assessment at the end of the unit
(Section 1.3). Therefore, you must write-up as you go along.

You will need to keep your own notes on the design. An A4 booklet or pad (with soft or hard covers) should
ideally be used. In some cases an electronic equivalent may be more appropriate. If so, please discuss with
the unit organiser. This notebook must be dedicated to this project (not used for any other work) and the
Unit Organiser will check and clarify this with each Group. Notes should be handwritten in English and can
be supplemented with data sheet information, diagrams etc. that are firmly affixed into the booklet.

An example of information recorded in a laboratory notebook is shown in Appendix A.

1.5 Practical Laboratory work

Although most of the assessment will cover the Simulink design, one stage of the practical laboratory activity
will have to be undertaken. This involves the breadboard design for the PWM controller and use of the
oscilloscope to record data (see Section 5). This requires about 3 hours of laboratory time. Further details
will be available in TB1.

3
1.6 Design Specifications
Not all groups will carry out the same design and there are 4 output specifications that will be randomly
assigned. The specifications for circuit design are shown in Table 1.1 (general specifications) and Table 1.2
(Individual Group specifications).

The PowerPoint videos give general guidance and these are for a 15V, 45W design. Since there are a number
of individual group specifications, these designs will be to be adjusted accordingly. Given the number of
degrees of freedom, no two designs should be identical (there is no ‘correct answer’).

Specification Comment
Input voltage 30 V rms AC Although the SMPS would use an AC mains supply (single phase
240 V rms in the UK) the practical laboratory was limited to a 30
V rms AC supply and so here the same will be assumed for this
design exercise.
Output voltage Variable See Table 1.2
Power output (max.) Variable See Table 1.2
Power output (min) Variable See Table 1.2
Output regulation 0.1% This was with the feedback loop.
(min to max. load) Without feedback performance will be compromised.
Output ripple 1% max Split between 100 Hz ripple and PWM switching ripple.
Without feedback performance will be compromised.
Efficiency at full load 75% The ratio of DC power out to AC power in
Ambient temperature 0 – 40 °C
Table 1.1: SMPS design specifications (all groups).

Spec. DC output Load resistance range Output Power Range

A 9V 2Ω–6Ω 14 W – 41 W
B 12 V 3Ω–9Ω 16 W – 48 W
C 15 V 4 Ω – 15 Ω 15 W – 56 W
D 18 V 7 Ω – 21 Ω 15 W – 46 W
Table 1.2: SMPS output voltage specifications (individual groups).

4
2. INTRODUCTION TO CIRCUIT MODELLING USING MATLAB SIMULINK AND SIMSCAPE
This exercise should be able to be completed within around 3-hours (as verified by previous undergraduate
students) and must be undertaken individually by each member of the group before starting the SMPS as a
group design exercise.

The aim is for each student to understand the basic concepts of time-domain circuit modelling and to
understand how to put together a simple passive electrical circuit in Matlab Simulink/Simscape. It will
describe the main features of the software and the best method to present the results.

2.1 The series LCR resonator circuit

The initial stage of the work will be to model a circuit that is familiar to you from the first year, namely the
series LCR resonator, Figure 2.1. The Blue blocks are from Simscape and show the actual circuit diagram
including voltage source (this can set both DC and AC voltages), voltage sensor (ideal voltmeter) and current
sensor (ideal ammeter). The Simscape design is then embedded within Simulink and are shown as the Black
Blocks.

Figure 2.1: Complete circuit in Simscape and Simulink interface.

The analysis of this circuit in the time domain for a step input of 5 V at time t = 0s is shown in figures 2(a) for
the voltage sensors and 2(b) for the current sensor. Each scope input needs to have a ‘PS-Simulink’
converter and each sensor has an output to that. The circuit also has to contain a ‘Solver Configuration’
block - this will all be explained in detail in Section 2.4.

(a) Scope showing voltage levels (b) Scope 1 showing current in the circuit.
Figure 2.2: Simulink analysis with voltage step applied at 0μs.

5
2.2 Quick overview of time-domain modelling
While a detailed understanding of the methodology behind the various ‘solvers’ is not required an
appreciation of the circuit analysis in time steps is important. The circuit in figure 2.1 is described in Eq. 1,
and for time stepping this can be rewritten as Eq. 2
𝑑𝑖 1
𝑉=𝐿 + 𝑅𝑖 + ∫ 𝑖 𝑑𝑡
𝑑𝑡 𝐶
(1)

𝑑𝑉 𝑑2 𝑖 𝑑𝑖 𝑖
=𝐿 2+𝑅 +
𝑑𝑡 𝑑𝑡 𝑑𝑡 𝐶
(2)

Expanding Eq. 2 in discrete time steps gives Eq. 3 as one example, though there are several different versions
for this approximation. In this case, the central difference is used. The subscript n is the discrete time step
number, and the applied stimulus is the voltage at a given time step, 𝑣𝑛 .
𝑣𝑛+1 − 𝑣𝑛−1 𝑖𝑛+1 − 2𝑖𝑛 + 𝑖𝑛−1 (𝑖𝑛+1 − 𝑖𝑛−1 ) 𝑖𝑛
=𝐿 2
+𝑅 +
2∆𝑡 ∆𝑡 2∆𝑡 𝐶
(3)

With initial conditions of 𝑖𝑛 as zero, this can then be written as Equation 4 and implemented directly as a
Matlab script, with the corresponding results shown figure 2.3.

2 ∆𝑡 2 2𝐿 1 𝑅 𝐿 𝑣𝑛+1 − 𝑣𝑛−1
𝑖𝑛+1 = [𝑖𝑛 ( 2 − ) + 𝑖𝑛−1 ( − 2) + ]
𝑅∆𝑡 + 2𝐿 ∆𝑡 𝐶 2∆𝑡 ∆𝑡 2∆𝑡
(4)

There are clearly issues surrounding the time step interval, ∆𝑡, as too short a time step will require a ‘long’
simulating time, while too long a time step may lead to instability. This does not need to be discussed
further here but may be an issue that should be considered with any numerical simulation.

Figure 2.3: Implementation of Equation 4 with 0.1μs time step and 5μs delayed voltage step.

Figure 2.1 can be described as a series of nodes and what Simscape does is to deal with the analysis of the
circuit as well (as the time-step) for any complex circuit.

6
2.3 Simscape modelling
2.3.1 Finding components and making interconnections

Menu for Simulink Library (including

all Simscape components).

You will mainly need to look in these

sections for the components you want.

Connect wires between components

by dragging ‘line’ between terminals Highlight the
component and
Components drag into the
can be rotated Simulink window

2.3.2 Setting component values – typical example but will be different for each component.

Double-left click Actual value in here

on component
This is only used if
the line below is set
Only need this
bit (for now) These are set for
‘ideal’ component

2.3.3 Labelling components.

Right click on
component

Move property
tokens across.

Unit and its value

7
This was a quick start guide to finding your way around Simscape so that you can quickly put together a
model to test. There are a lot of additional features that you may wish to try at a later stage.

2.4 Simulink components and interface

There is now a need to add the Simulink components to control the model. Again, bring up the library and
drag the highlighted components into the Simulink window.

2.4.1. Adding Scope and Solver Configuration blocks

In the PS-Simulink Converter for

the Voltmeter set this to V and for
the ammeter set this to A.

2.4.2. Solver Configuration and Configure Parameters

The ‘Solver Configuration’ needs to be linked to the Simscape model together with the ‘Electrical Reference’
(Ground) and the voltage and current sensors to the PS-Simulink converter again by dragging a line between
terminals.

May be required Generally will use local

solver

Set time for a fixed

step simulation

Set solver parameters

to these values for this
Make sure these are connected. Set model.

There are a number of ways to set up the solver and once working you can try other solver types and sample
times to see how these may affect the outcomes. In addition to this, Configuration Parameters needs to be
set whereby a fixed or variable time step can be selected and the Stop time for the model (leave Start time
as 0). Once this is done the simulation can be run.

8
 Run simulation

Model Configuration
Parameters button

Simulation time limits.

Keep start as zero but
adjust stop as required.

This sets the model to run

with a fixed time step

Both the Solver Configuration and the Configuration Parameters may need to be adjusted during the
modelling process to allow fast, yet stable, operation of the models. In all designs and summary table of the
initial set up will be included (Table 2.1 shows this one) but you are free to change parameters to see how
these may affect accuracy, stability and running efficiency.

Solver Configuration Configuration Parameters

Local solver: Backward Euler Type: Fixed-step
Sample time: 1e-6 [Seconds] Stop time: 5e-5 [seconds]. Adjust as necessary.
Table 2.1: Summary of suggested model parameters

Since initial conditions for the components are set to zero, the simulation will yield both the transient and
the steady-state responses. In the physical circuit build for the practical laboratory, the oscilloscope will
trigger on events (i.e. rising/falling edges of repetitive signals) and hence what is captured on the
oscilloscope is usually the steady-state outcome (though transients are possible). The model will therefore
need to be ‘run’ for the steady state to be reached.

2.4.3 Changing Scope settings and data logging

On running the simulations for the first time, to bring up the Scope outputs, just click on the relevant Scope
icon. The default background colour is black, and the first trace colour is yellow. While this is what
oscilloscopes do usually look like, it is not particularly useful for presentation in documents. While the
default is not to preserve colours when copying, the colours will change. You will then need to set/change
both the horizontal and vertical axes on graphs.

You will need to export the figures you have generated. This must not be a PC screenshot but you must
copy the figure (‘copy figure’ from ‘Tools’) and paste into the word document that will be used in the
laboratory assessment and in your Laboratory Notebooks.

9
 Where to find
‘copy figure’

Set time-axis
labels

Change these
to white.

Tick this box so that

Add legend Default black plot colours do not
and relabel background change when copying
to documents
Set suitable colour and line
for each plot on the graph DON’T FORGET

Manually set limits here

Must add label and units

You can also output the

data from Simulink for
processing elsewhere

You can also configure data logging that will allow you to export your data for use in other applications.

2.5 Practical versus simulation

2.5.1 Scope Probes
In the simulation you can apply both (ideal) voltage monitors and (ideal) current monitors, but in reality,
these are not ideal, and their use will influence the circuit operation in some way. In the practical laboratory
you would be using a 4-channel oscilloscope with x10 probes for monitoring voltages at various parts of the
circuit. You would tend not to use current probes (although available) for monitoring current but may
instead use a low-value (known) resistor, ‘current sensing resistor’, and use the voltage across it to
determine the current level.

A realistic model for the x10 scope probes and the oscilloscope’s input impedance also includes the
compensation capacitance for tuning the scope probe so that a 10KHz square wave applied at the probe
does look like a 10KHz square wave on the oscilloscope display.

Also note the Earth (Ground) connections on the circuit. A scope probe has two connections: the probe tip
and a Crocodile clip that is connected to Ground. So, wherever this clip is attached will ‘ground’ that point
and if multiple probes are used and clips placed at different parts of the circuit, then these points will be
grounded (and therefore connected together). This may cause problems with the circuit operation, so this
is something to be aware of.

10
 Probe
X10 Probe Oscilloscope
tip

Crocodile
clip

10kHz squarewave for calibration In practice, the oscilloscope is Earthed

generated by the oscilloscope Tuning capacitor and therefore so are all the probes!

In all the simulation models you will use the ideal voltage and current monitors that will not influence the
operation of the circuit, but you will need to appreciate that in a practical circuit this will not be the case.

2.5.2 Component Polarity

It can be seen that many of the model components have a polarity attached to them. In some cases, this is
necessary and in others it is not. This is summarised in Table 2.2.

Component Comment
Resistor Not relevant for this design
Capacitor The practical design does use Electrolytic Capacitors where polarity is
shown on the component and is VERY important.
For all other capacitors, polarity is not relevant.
Inductor Not relevant for this design
Transformer This is important but is effectively replicating the ‘dot convention’
used in circuit diagrams.
Diode May be shown. This is important but the anode and cathode are
already identified in the component’s symbol.
Voltage and current sources This is important
Voltage and current monitors This is important
Table 2.2: Importance of polarity for the electrical components.

2.5.3 Preferred component values

While the simulation allows you to specify any component value (and down to any number of decimal
places) in reality, components can only be sourced in ‘preferred values’ and these have known tolerances
(e.g. ±5%). So, having calculated the ‘exact’ value for any component, you must then choose the nearest
preferred value to use for your design. This may change the operating conditions of the circuit slightly.

2.5.4 Maximum operating conditions of components

Finally, it is impossible in modelling to actually blow up components! This is not the case with the practical
work where resistors that are, say, rated at 0.25W get very stressed when sinking 0.5W and at that point will
‘fail’. Fault conditions can be applied to modelling but will not be implemented here. In any modelling work
you should therefore take account of this for before attempting to build a practical circuit.

11
3. CIRCUIT DESCRIPTION AND TASKS
Figure 3.1 shows the circuit diagram for the open loop SMPS. The aim here is not to just build the circuit
using Simulink/Simscape but to understand the operation of each of the six highlighted blocks that make up
the complete circuit.

Duty Cycle Control Secondary filter

and DC output
Gate Drive Circuit

30V AC supply and

PWM Generator full wave rectifier
MOSFET switches
Gate drive transformer Main transformer

Figure 3.1: Open loop SMPS circuit diagram

The relevant PowerPoint presentations are listed under the following stages:

1. PWM Controller
2. Full wave rectification of 30Vrms AC supply; Gate Drive Circuit, MOSFET switches and Main
transformer (primary winding)
3. Main transformer (secondary winding) and secondary filter
4. Feedback (closed loop). This is not shown here and will not be assessed this year.

In a practical design, these would have to be tackled in this order to be able to get each section of the design
to operate due to interdependencies. With the simulation design, ‘sources’ and ‘sinks’ (voltage and current
monitors) can be inserted into the part-constructed circuits, and hence there is more flexibility in the
verification of component values. Therefore, the calculation of component values in each of the highlighted
blocks will be separate tasks, with the integration of the various circuit parts being the final stage.

Section 4 will detail the various Simulink/Simscape circuits that will be constructed and identify the relevant
PowerPoints for the calculations. Although the input voltage will be 30V rms AC in all cases, Table 1.2, there
will be 4 possible DC output voltages ranging from 9V to 18V as listed in Table 1.3, with one of these
assigned to each group. In addition, the PWM frequency can be chosen by groups to be anywhere between
80 KHz and 120 KHz. Therefore, it is very unlikely that any two circuits will be identical in choice of
components and circuit performance.

The PowerPoint presentations that you will need to access and take notes from assume a 35Vrms AC input
and a 15V DC output, so all your calculations will need to be scaled accordingly.

12
4. CIRCUIT MODELS
For each of the designs there will be a worksheet to record the design and simulation outputs as a group.
This is in addition to individually recording the information (in more detail) in your Laboratory Notebooks.
Before attempting to calculate the values of the components and construct the model, all PowerPoints
identified in each of the sections must be read and design information written into Laboratory Notebooks.

4.1 Full-wave rectification of AC supply voltage

The PowerPoint that needs to be used for calculations prior to starting the modelling:

• Reservoir Capacitor.mp4

This is an equivalent
resistance representing
+ the rest of the SMPS

One of 4 diodes in a
-
single package.

Reservoir Capacitor. This is an Electrolytic

Capacitor, so polarity does matter here

Figure 4.1: Input AC rectification.

Solver Configuration Configuration Parameters

Local solver: Backward Euler Type: Fixed-step
Sample time: 1e-6 Stop time: 50e-3
Table 4.1: Summary of suggested model parameters

Requirements for the full-wave rectification of the AC input, figure 4.1:

1. Here you are calculating the value of the reservoir capacitor based upon the information given in the
PowerPoints for the power requirements of the SMPS. Make sure to include ‘efficiency’.
2. In addition to this (not in PowerPoints), to get this model to operate you will have to determine an
Equivalent Resistance for the rest of the SMPS circuit. You need to just consider the power needed
to be ‘dissipated’ into this load being equivalent to the power flowing the rest of the SMPS (if it was
connected).
3. A voltage sensor has been connected across the output. Add any other appropriate sensors.
4. Record all relevant data and fill in the Group Worksheet on this.

13
4.2 PWM generator

The PowerPoint that needs to be used for calculations prior to starting the modelling:

• 3524 PWM.mp4

The PowerPoint concentrates on the use of the 3524 IC to generate the PWM waveform. This circuit design,
Figure 4.2, has been slightly simplified in that the PWM frequency is set within the ‘Controlled PWM Voltage
block’, Figure 4.3, and not by external CR timing components. However, these CR components must be
calculated as they would be required in the optional practical work described in Section 5. A transistor has
also to be added to the circuit to model the open-collector output stage of the 3524 IC.

Sets potentiometer position (not duty cycle) This is a simplified representation

of the 3524 PWM Chip showing
equivalent pin numbers and no
external CR timing components

Pin 9 Pin 12

Pin 11

Assume 15V DC supply here for now

Figure 4.2: Basic PWM circuit.

You decide this

May need to adjust to work

with PWM frequency.

Limits maximum Duty Cycle to 50%

Figure 4.3: Parameters for the PWM controller and open-collector output transistor

14
 Solver Configuration Configuration Parameters
Local solver: Backward Euler Type: Fixed-step
Sample time: 1e-9 Stop time: 1e-4
Table 4.2: Summary of suggested model parameters

Requirements:

1. Setting the PWM frequency to your own choice (80-120 KHz) in the model and determining the CR
values that would be required for the practical 3524 PWM IC.
2. Insert an open collector NPN transistor that is actually the output stage of the 3524 IC and
identifying a suitable pull-up resistor to be used. This resistor may be changed later.
3. Checking range of duty cycle with the potentiometer
4. Record all relevant data and fill in the Group Worksheet on this.

4.3 Gate-drive circuit

The PowerPoints that need to be used for calculations prior to starting the modelling:

1. Gate Drive Transformer.mp4

2. PNP-NPN Transistors.mp4

The next stage, Figure 4.4, requires a number of new components, and the circuit should be constructed and
tested in stages rather than all at once.

Although a transistor has been used in the PWM controller, the values required were given to you. Here the
PNP-NPN follower pair of transistors will need to be specified from the data sheets with the parameters
required to be entered shown as the blanked-out sections of Figure 4.5 for the NPN transistor. The PNP has
similar, but not identical, parameters as the PNP transistor and will also have to be entered into the model.

The second new component is the gate-drive transformer. This is a toroid and probably strictly should be
represented as a non-linear three-way transformer, but a suitable Simscape model is not available in the
library. Given that this is to give isolated (equal) feeds to the MOSFET gates and add some voltage
amplification, this representation is sufficient. The PowerPoint show the number of turns that should be
used on the primary and secondary windings and the corresponding value of primary inductance, L1, is given
in Figure 4.6. From the relationship between number of turn an inductance you will need to calculate the
secondary inductance. You may need to increase the number of secondary turns later. Note also the ‘dot
convention’ used on the circuit.

The circuit also includes an equivalent gate charge capacitance for the MOSFETs. This will be discussed in
detail in Section 4.4.1, so for now just include the values indicated in Figure 4.6.

15
 Add 1MΩ resistors for 10Ω damping
model ‘continuity’ only resistors

0.22μF (DC Block)

Equivalent MOSFET Gate Drive Transformer

gate-charge capacitance (as ideal transformer)

Figure 4.4: PWM and gate-drive transformer circuit.

Figure 4.5: NPN transistor parameters. Double left-click the icon to access these.

Note the ‘dot convention’.

To test this circuit initially

make these ‘lossy’
capacitors by setting:

Need to work these out C = 200pF, R = 200Ω

given the turns ratio.

As the secondary circuit is isolated there is a need to add the

high value resistors (say 1MΩ) to allow the model to run.

‘Comment Out’ later when more of the circuit is added.

An alternative is to add an Electrical Reference to both output

windings, though in a practical circuit this could be an issue.

Figure 4.6: Gate-drive transformer and load conditions.

16
 Solver Configuration Configuration Parameters
Local solver: Backward Euler Fixed time step
Sample time: 0.01e-6 Stop time: 3e-4
Table 4.3: Summary of suggested model parameters

Requirements:

1. Build up the circuit in stages by first adding PNP-NPN transistor pair and DC blocking capacitor, and
then checking waveforms.
2. Add the gate-drive transformer and again check voltages at both outputs
3. Add final components and checking range of duty cycle with the potentiometer
4. Record all relevant data and fill in the Group Worksheet on this.

4.4 Completing the Main Transformer Primary Winding circuit

The PowerPoints that need to be used for calculations prior to starting the modelling:

1. MOSFETs.mp4
2. Primary Winding.mp4
3. Circuit Operation.mp4

To complete the circuit to operate the main transformer primary winding, MOSFETS, freewheeling diodes
and the main transformer primary winding need to be added, Figure 4.7. Before embarking on this it is
necessary to identify a number of features of both components. The MOSFET are just on-off switches, but to
get them to operate (power) efficiently, current needs to be injected quickly into their gates at turn on and
quickly removed at turn off – hence the need for the gate-drive circuit that is a ‘current amplifier’.

Damping resistors ‘Freewheeling’ Diode Primary winding inductance

1MΩ resistors now not needed

Current sensing resistor Ideal DC link voltage

Figure 4.7: Complete circuit up to the Main Transformer Primary Winding.

17
The diodes are also just acting as on-off switches, but it is the voltages applied across their terminals that
determines this operation. The primary transformer at this stage is simply an inductance. This is, however, a
non-ideal inductance as it exhibits hysteresis (and other losses). Initially this can be modelled as ideal and
then changed to non-ideal once the model is working.

4.4.1. MOSFET switching

The MOSFET capacitance between its terminals plays an important part in the operation and efficiency of
the device. While the device will be ‘on’ once the gate-drain threshold voltage is reached just applying a
voltage of that level to the gate will not instantaneously turn the device on due to this capacitance. Delays
in turning the device on (and off) considerably increase the losses (‘switching’ losses) and these add to the
loss due to the ‘on resistance’ of the device, 𝑅𝑑𝑠(𝑜𝑛) . The MOSFET loss calculations in the PowerPoints deal
only with the latter.

Once the N-channel MOSFET is loaded into the circuit you will need to double-left click on the icon, and you
can then bring up the two input windows to enter the necessary data, Figure 4.8. Again, you are expected to
input the data from the relevant data sheet – the blanked-out parts are the figures required. For the
‘Measurement temperature’ use the 25°C value for 𝑅𝑑𝑠(𝑜𝑛) shown on the data sheet for now, but in all your
calculations for losses and heatsink requirements give in the PowerPoints use the scaled value for you
chosen junction temperature.

Figure 4.8: MOSFET parameters.

For the MOSFET capacitance three figures are required to be entered in the Junction Capacitance Table.
These do not necessarily give the capacitance between terminals but are measured as a combination of
them according to equations 5 – 7.

𝐶𝑖𝑠𝑠 = 𝐶𝑔𝑑 + 𝐶𝑔𝑠 with 𝐶𝑑𝑠 shorted (5)

𝐶𝑟𝑠𝑠 = 𝐶𝑔𝑑 (6)

𝐶𝑜𝑠𝑠 = 𝐶𝑔𝑑 + 𝐶𝑑𝑠 (7)

Figure 4.9 shows a model to test both the threshold and the capacitance effects of the MOSFET. The input is
a ramp function that can be used to see what happens when threshold is reached. The pulse generator can
also be directly connected to the circuit to see the usual switching operation of the MOSFET and the delayed
switching due to the capacitance. This circuit build is optional.

18
 Pulse generator can be directly
connected to the circuit.

Figure 4.9: Testing the MOSFET.

4.4.2. MOSFETs, Freewheeling Diodes and Main Transformer primary winding inductance.
The final components can now be added to the model, Figure 4.7. In addition to the MOSFETs, Figure 4.10
shows the Block Parameter details that need to be entered from the data sheet for the diodes and the
Primary winding inductance that you have calculated. The circuit shows a non-linear inductance that can be
operated in linear inductance mode, and you should use this first in the model.

Calculate this using the ‘Test Initially do not include May be able to calculate a
Conditions’ information, IF the non-linearity suitable value for this later

Figre 4.10: Freewheeling diode and linear inductance parameters.

Solver Configuration Configuration Parameters

Local solver: Backward Euler Type: Fixed-step
Sample time: 1e-9 Stop time: 0.3e-3
Table 4.4: Summary of suggested model parameters

4.4.3 Nonlinear Inductor

While both the turns and (unsaturated) inductance have been calculated, the model also requires the
Saturated Magnetic Flux and Saturated Inductance. These has not been described on the PowerPoint slides,
so details are given here.

Figure 4.11 shows the mapping of the B-H curve axes to a Φ − I axes using equations 8 and 9, where A is the
ferrite core area and l is the ferrite core length (both given on data sheets).

Φ = 𝐵𝐴 (8)

𝐻𝑙 (9)
𝐼=
𝑁

19
 From equation 10 we can then have a relationship between the magnetic flux in the inductor and the
applied current

𝑑𝜙 𝑑𝑖 (10)
𝑉=𝑁 = 𝐿
𝑑𝑡 𝑑𝑡

𝑁Φ = 𝐿𝐼 (11)

Or more strictly this can be applied as a piecewise linear model looking at the asymptotic inductance
slopes, equation 12, and hence the unsaturated inductance and saturated inductance can be defined
separately with the intercept point at the saturate magnetic flux level, Φ𝑠𝑎𝑡

𝑑Φ 𝐿 (12)
=
𝑑𝐼 𝑁

The equation for the asymptote for the unsaturated inductance can therefore be written as equation 13
and therefore these parameters can be calculated for figure 4.10 using the magnetisation, Φ𝑟𝑒𝑠 .

𝐿 (13)
Φ= 𝐼 + Φ𝑟𝑒𝑠
𝑁

𝑑Φ 𝐿𝑠𝑎𝑡
B Φ =
𝑑𝐼 𝑁

Φ𝑠𝑎𝑡
′
Φ𝑠𝑎𝑡
Φ𝑟𝑒𝑠 𝑑Φ 𝐿
=
𝑑𝐼 𝑁
H I
Figure 4.11: Piecewise linear approximation including saturation (positive Φ − 𝐼 quadrant)

While it would be more applicable to use the non-linear mode, this should not be essential since you should
be operating the inductor within its saturation limits and only in one quarter of the B-H curve (positive B,
positive H). Figure 4.12 show the Block Parameters required for the non-linear inductor using the model in
Figure 4.11(a). The saturation inductance and saturation limit will need to be estimated from the B-H curve.
The latter should be above your operating limits. If not, the primary winding current will not be a uniform
ramp waveform.

20
 Now include basic non-linearity
(but not hysteresis)

You have calculated both of these

These will need to be

calculated from B-H curves.

Φ𝑠𝑎𝑡 : use ferromagnetic

core data sheet to find this

Figure 4.12. Non-linear inductor mode with saturation needing to be set.

4.4.4 Heatsink requirements for active components

Many active components (such as MOSFET) have provision to attach a heatsink in order to control the
junction temperature of the device and hence its efficiency. For this laboratory, you will need to determine
the losses in the MOSFET and the requirement for a heatsink. This information can be found in:

• MOSFET Heatsinks.mp4

Requirements:

1. The MOSFETs, freewheeling diodes and primary inductance will all need to be added before
checking the circuit.
2. The DC link voltage will need to be set to the average value based upon your reservoir capacitor
choice, Section 4.1.
3. Include damping resistors to both MOSFET gates. Initially use a series resistance of 10Ω and a shunt
resistance of 5.6KΩ. These may be modified later.
4. Use ideal (linear) inductance setting first.
5. In addition to monitoring voltages, monitor the primary winding inductance current using the
current sensing resistor. Amplify the signal so that this signal can be seen on the same scope vertical
axis scales as those for the other monitored parts of the circuit.
6. Include the non-linear transformer model parameters
7. Identify heatsink requirements for MOSFETS
8. Record all relevant data and fill in the Group Worksheet on this

4.5 Secondary Circuit and ‘100KHz’ filter

The PowerPoints that need to be used for calculations prior to starting the modelling:

1. Secondary Winding.mp4
2. Filter Inductor Design.mp4
3. Filter Capacitor.mp4

Although not assessed in the laboratory, information on the practical construction of the inductor can be
found at:
• Inductor Construction.mp4

21
This section considers the filtering of the transformer secondary waveform to remove the AC components to
yield the DC voltage output for the SMPS, Figure 4.13. This will be initially modelled independently from the
previous models for two main reasons:

• The waveforms generated in the earlier circuit produce a lot of oscillations that do happen in
practice though makes it very difficult to understand the basic operation of the filter, so the
switching waveform is ‘cleaned up’.
• The filter characteristics have a long transient time given the high frequency of the square wave
input. The model has to be run for a long time period before the steady state is reached and this run
time needs to be adjusted manually.

The new component being introduced here is the non-linear transformer representation of the main
transformer. As with previous magnetic components this can be used in both linear and non-linear modes,
Figure 4.14. The ‘cleaned-up’ switching waveforms use two pulse generators whose outputs are combined.
Figure 4.15 shows a potential set-up to generate the stepped waveform driving the main transformer
primary winding though these will need to be adjusted for your design specifications.

Main Transformer. Can be

operated in linear mode

Linear inductor
+
Diode pair in a
-
single package

Idealised switching waveform to Total ESR Capacitance

test main transformer and filter Load Resistance
and Resistance

Figure 4.13: Secondary filter and main transformer (simplified configuration).

Primary winding inductance

you have already calculated

Now need to calculate

this in same way as for
the primary winding

Figure 4.14: Parameters required for the transformer operated in ideal (linear) and non-ideal modes.

22
 Subtracting this one

These set the same for both Pulse

Generators.

Adjust Amplitude to your (average)

DC link voltage (including MOSFET
voltage drops).

Adjust Period for to your PWM

frequency

Adjust Width for Duty Cycle.

Adjust delay to get correct

composite-pulse shape

Figure 4.15: Setup for pulse generators.

Solver Configuration Configuration Parameters

Local solver: Backward Euler Type: Fixed-step
Sample time: 0.5e-9 Stop time: 1.5e-3
Table 4.5: Summary of suggested model parameters

Requirements:

1. Determine a suitable number of turns for the main transformer secondary winding and hence
equivalent circuit model parameters for the transformer.
2. Calculate the critical values of the filter inductor and hence a suitable final value for the output filter
inductor.
3. Calculate the minimum capacitance requirement for the DC ripple and the ESR requirement for the
ripple and then decide on a suitable output filter capacitance
4. Check circuit operation with both minimum and maximum load resistance and measure the ‘100KHz’
ripple.
5. Record all relevant data and fill in the Group Worksheet on this.

4.6 Complete open loop SMPS circuit

The relevant PowerPoint for this section of work that needs to be used:

• Circuit Test - Operation.mp4

It is now time to combine the Full wave rectifier (section 4.1) and the secondary circuit (section 4.5) with the
primary circuit (section 4.4). Since all the subsections have been fully tested the whole circuit should work
perfectly first time!

The complete open loop circuit is shown in Figure 4.16 and the integration requires some removal in interim
components used in the individual stage tests. The expectation is that while the majority of students will

23
complete up to section 4.5 and hence have satisfied the calculation of the basic design components and
have recorded the majority of the waveforms in the circuit, not all students will get the final model working
in the allotted time for the project. This is entirely in keeping with the practical SMPS build that around 75%
of the students partially or fully completed the final stage of the project. With the knowledge you have
gained as a group from this laboratory will now need to work together with minimum HPT support to get the
circuit to operate.

Set the output voltage using this (maximum = 1) Replace ideal DC link voltage with full wave rectifier

In a practical circuit this would

now be replaced with a current Monitor
sensing transformer Replace transformer primary winding inductance output
with complete transformer and filter

Figure 4.16: Complete open-loop SMPS.

Requirements:

In this section you are not calculating any component values but checking to make sure that components
you have previously specified still suitably work. There may need to be some changes to component values
to get the circuit to operate correctly. Once this is done, you need to consider the following the open-loop
system performance metrics:

1. Linearity of the circuit (DC in: DC out)

2. DC regulation (changing the resistive load from its minimum to its maximum value and seeing how
this affects the DC voltage)
3. The noise being generated in the circuit (‘100Hz’ and ‘100KHz’ ripple on the DC output).
4. Record all relevant data and fill in the Group Worksheet on this.

4.7 Closed loop control and other considerations

1. For those that are interested in taking this design further the feedback loop (see Figure 1.1) can be
included in the model to operate the SMPS in closed loop mode. This would be the usual
configuration to allow regulation of the output voltage and control noise levels in the system.
2. You will see that there are many DC voltages sources that have to be used throughout the design.
These need to be ‘removed’ as such since this is a power supply and the only input is from the AC
supply and the only output is the DC.

These are both outside the scope of this laboratory and the design is not assessed. If anyone is interested,
then this can be discussed with the Unit Organiser.

24
5. PWM Practical Laboratory
All groups must undertake the practical build of the PWM control circuit using the 3524 PWM IC and the
prototyping board, with the results entered into the Group worksheet.

Before undertaking the laboratory, you will need to familiarise yourself with this circuit, make all appropriate
notes in your laboratory book (including the connections and operation of the oscilloscopes). It would be
useful, but not essential, to consider how the PWM waveform is actually being generated within the 3524 IC
used in the laboratory, figure 4.17, and is slightly more complex than the shown in Section 4.2.

Figure 4.17: A representation of the internal operation of a PWM.

For the practical build you will just need to read the following Blackboard information which show the
internal workings of the 3524 IC (in Black) and the external connections you need to make (in Blue):

• 524QuickStart.pdf

You have already calculated the timing components C and R and so you will ned to add a potentiometer for
controlling the PWM duty cycle, a pull up resistor on one of the open-collector outputs (you can choose
either one) and a decoupling capacitor on the 12 V power rails. You only need one DC source (12 V) and this
comes from the prototype variable voltage output. There is also a need for 5 V DC to connect to the duty-
cycle and where this comes from is one of the questions posed on the information sheet. Component will
just need to be plugged into the board and wires (single strand) cut to link the components.

25
Appendix A: Laboratory Notebook
This shows an example of the information that should be recorded in your notebook. Pages should be
numbered, titled, dated and with an index used at the beginning so you can locate information quickly. Each
student must keep their own laboratory notebook even though you are sharing information in your group.
Laboratory notebooks will be checked.

26