LABORATORY 1-2

PSIM EVALUATION OF PERFORMANCE INDICES

FOR A DC-DC BUCK CONVERTER

1. Theory

The simplified circuit for converting a DC voltage to another DC voltage with a lower value is shown
within Fig. 1. This is called buck converter and it can represent the main component within a power
supply.
V Voltmeter:

Gating Block

Fig. 1 Simplified circuit for a buck converter

The input DC voltage is transformed in a sequence of pulses of frequencies usually within the ultrasonic
range. These pulses are next applied to a low-pass LC filter which stores energy for the interval when
the main switch is in OFF-state. The output capacitor in Fig. 1 is usually made up a group of ceramic and
electrolytic capacitors. The level of the output DC voltage depends on the pulse frequency and the
duration of the ON-state. The ratio between the ON-state duration and the pulse period is called duty
ratio.
The switch is usually built with a MOSFET transistor. During the MOSFET’s ON state, the diode D is
seeing negative voltage and stays in OFF-state. The input current also charges the output capacitor and
supplies the load. The inductor is seeing a roughly constant voltage during the ON-state of the MOSFET
and this means a linear increase of the inductor current. Simultaneously, energy is stored within the
magnetic field of this inductor.
After the switch turns-off by control, the inductor tries to maintain the current circulation and forces
the turning-on of the diode D. The current flows from inductor into the output capacitor bank and load
and closes through diode. The inductor is seeing a negative voltage equal to the load voltage. The
inductor current is linearly decreasing ceasing energy from magnetic field into the capacitor bank and
load.
As long as the current is lower than the current required at output, the capacitor would deliver an
additional current, that means it would discharge. Conversely, when the inductor current is larger than
the load demand, the output capacitor bank would charge.
Since the MOSFET’s current becomes zero after its turn-off, the energy stored within the magnetic
field of the inductance would tend to produce an overvoltage. The inductor would release as much
energy as necessary to maintain or create a path for the current. Such current closes through diode. If
1
the diode would be missed, the MOSFET transistor would get damaged by the created overvoltage. For
this reason, this diode is also called recovery diode.
The inductance current fluctuates in between two limits. Depending on the levels of the minimum and
maximum currents, two operation modes are possible: continuous and discontinuous conduction
modes. Performance indices for both operation modes are discussed herein. Examples for waveforms
are shown in Figs. 2-3.

Vout

(a) 3.34
3.335
3.33
3.325
3.32

Gate

(b) 0.8

0.4

I(L1)
(c) 17.2
16.8
16.4
16

I(MOS1)
(d)
15
10
5
0
0.1 0.10002 0.10004 0.10006 0.10008 0.1001
Time (s)

Fig. 2 Continuous Conduction Mode (a) load voltage and ripple; (b) gate control signals; (c) inductor current; (d) transistor
current

Vout

6.888
(a) 6.884
6.88

Gate

0.8

0.4
(b) 0

I(L1)
0.8
0.6
0.4
0.2
0
(c) I(MOS1)
0.8
0.6
0.4
0.2
0
0.1 0.10002 0.10004 0.10006 0.10008 0.1001
Time (s)
(d)

Fig. 3 Discontinuous Conduction Mode (a) load voltage and ripple; (b) gate control signals; (c) inductor current; (d) transistor current

The LC filter can be designed for a continuous conduction mode. For instance, the maximum level of
the inductor current ripple can be established (usually around 5%) as well as the maximum level of the
capacitor voltage ripple (possibly under 10 %). The following design relationship yields:

2
 1
Cf = ΔvCf  ·(1-D) (1)
2 
8·L·fS ·  
 VCf 
 
2
VL
Lf = ·(1-D) (2)
 ΔiLf 
P·fs ·  
 ILf 
 

Where VL represents the average value for the load voltage (V), P represents the active power in load
(W), fs represents the switching frequency (Hz), D represents the nominal duty ratio, ΔiLf/ILf represents
the normalized ripple of the inductor current (5%), ΔvCf/VCf represents the normalized ripple of the
capacitor voltage (10%), Lf represents the filter inductor (H), Cf represents the filter capacitor (F).

2. Definition of Performance Indices

Performance indices for a dc/dc converter can be defined with a simulation tool.
For comparison, their analytical definition is provided herein.

Input side
• Average value of the input voltage
1
= ∙

• Average value of the input current

1
= ∙ = ∙

• Input active power

1
= ∙ ∙ = ∙

Power transistor switch:

• Average value of the voltage drop across transistor:
1 1−
= ∙ = ∙

• Average current through power switch:

1
= ∙ = ∙

• Peak voltage on switch = supply voltage

• Peak current on switch = supply current

3
 Recovery diode (this is NOT the anti-parallel diode to the MOSFET):
• Average voltage drop
1
= ∙ =

• Average current through diode

1
= ∙ = 1− ∙ = 1− ∙

• Peak voltage on diode = supply voltage

• Peak current on diode = load current

LC Power Filter
• Average voltage across capacitor
1
= ∙ =

• Average current through inductor

1
= ∙ =

Load:
• Average voltage on the load
1
= ∙ = ∙

• Average current in load

1
= ∙

• Active power in load

1
= ∙ ∙ = ∙

3. PSIM MODEL

In order to simplify the analysis, generic models for transistor and diode are used from PSIM library.
There are several possibilities for modeling of the gating control signal. The provided model herein
uses the simplest solution, that is a gating block from library. This gating block includes the gate driver
and the logic for control. If a logical circuit is used instead to generate pulses, an actual gate driver model
needs to be used in simulation.
The sequence of commands used to build the model is included in Table I.

4
 Elements > Sources > Voltage > DC
Elements > Power > Switches > MOSFET
Elements > Power > switches > Diode
Elements > RLC Branches > Inductor
Elements > RLC Branches > Capacitor
Elements > RLC Branches > Resistor
Elements > Other > Probes > Voltage Probe

Simulate > Simulation Control

TABLE I

4. Analysis by simulation

To start off, the analysis is performed for the two operation modes and performance indices for the
two cases are collected from simulation. The main circuitry should look as shown in Fig. 4 and the only
difference in between the two cases is the load resistance value of 0.2 Ohm and 40 Ohm, respectively.
For data collection, the current through MOSFET, diode and inductor currents need to be validated by
clicking on each of these components and setting the current flag to “1”.

Fig. 4 PSIM Model and validation of currents.

The gate control generator should be set for a switching frequency of 100 kHz, and the “simulation
controller” should see a simulation step of at least 1/100 of the switching period. This setting goes to 0.1
µsec. The analysis window between the “Print time” and the “Total time” should correspond to several
cycles, for instance 100 µsec.

5
5. Collecting data

After the simulation is performed, an analysis window opens, offering a possibility to select curves
from currents in diode, transistor, inductor, the gate signal and the output voltage. A good practice is to
select one curve per screen so that these are properly scaled.
Since the performance indices need averaging calculation, a good practice is to allow several cycles as
the width of the window for a calculation. After a waveform (curve) is uploaded, the right analysis tool
is selected from the top-down menu.

Fig. 5 Selection of analysis tools.

5.1. Continuous Conduction Mode

Consider a load resistance of 0.2 Ohm for CCM. Next, Table II has to be completed with measurement
data, not calculated.

Performance Index Value

• Average value of the input voltage
• Average value of the input current
• Input active power
• Average value of voltage drop on transistor
• Average current through power switch:
• Peak voltage on switch = supply voltage
• Peak current on switch = supply current
• Average voltage drop on diode
• Average current through diode
• Peak voltage on diode = supply voltage
• Peak current on diode = load current
• Average voltage across capacitor
• Average current through inductor
• Average voltage on the load
• Average current in load
• Active power in load
TABLE II
6
 5.2. Discontinuous Conduction Mode
Consider a load resistance of 40 Ohm for DCM. Next, Table III has to be completed with measurement
data, not calculated.

Performance Index Value

• Average value of the input voltage
• Average value of the input current
• Input active power
• Average value of voltage drop on transistor
• Average current through power switch:
• Peak voltage on switch = supply voltage
• Peak current on switch = supply current
• Average voltage drop on diode
• Average current through diode
• Peak voltage on diode = supply voltage
• Peak current on diode = load current
• Average voltage across capacitor
• Average current through inductor
• Average voltage on the load
• Average current in load
• Active power in load
TABLE III

5.3. Analysis of results

(1) As required at 5.1 and 5.2, complete the two tables.
(2) Explain differences between collected data in the two cases.
(3) Apply FFT analysis to the input current waveform. Copy the result and explain the harmonic
components.
(4) Verify equations (1) and (2) for the calculation of the L,C components using data from simulated
waveforms, for the two cases. Wherefrom the difference arises?
(5) Specify data able to select the recovery diode.
(6) How much is the system efficiency in the two cases? Why?
(7) Do we have power loss in this case?
(8) For the first case, change the duty ratio D and observe current ripple. When do we have the largest
value of the ripple?

6. Thermal analysis

6.1. Loss Calculation

While the performance indices from the previous section do not change too much with the selection
of the power switch, the thermal aspects are strongly depending on the actual transistor, diode and
inductor used in circuit. This includes power loss, efficiency, junction and case temperature. This section
is analyzing the circuit for these aspects.
In this respect, reconsider the PSIM circuit for an input voltage of 270 Vdc and a 20 A dc constant load,
7
and a switching frequency of 20 kHz. This set of data is generic for a high voltage dc/dc converter used
within a hybrid-electric vehicle.
The generic MOSFET model needs to be replaced with an actual transistor device model, herein
chosen as IRF460. This precise model includes all thermal data for the transistor.
Main datasheet characteristics of the IRF460 device are given in Fig. 6.

Fig. 6 Datasheet information for IRF460.

The thermal model is defined after a loss calculation.

When calculating the loss, using the devices from the device database, PSIM uses the same ideal
switch models as the regular PSIM switch models. For example, in the case of an MOSFET, only the
transistor on-resistance RDS(on) and the anti-parallel diode conduction voltage drop Vd are considered,
and the dynamics of the switch turn-on and turn-off transients are not directly considered.
The on-resistance RDS(on) is a function of the transistor junction temperature, expressed as:

= _! ∙ "1 + $ ∙ % & − &_! '(

where RDS(on)_b and Tj_b are the base values at the test conditions, normally at 25oC.
The temperature coefficient KT is expressed as:

1
$ =) − 1* ∙
_! & − &_!

or in normalized value:

_ +,-. /0 −1
$ =
& − 25

Based on the ideal models, the voltages and currents of the transistor and diode at each instant are
obtained from the simulation. These voltages and currents are then used for the loss calculation. Note
that the new values of RDS(on) and Vd are used in the simulation at the next time step.
The transistor conduction loss is calculated as follows:

3_ _ = 4
∙

8
 where Id is the drain current.
The transistor turn-on loss yields as:

3_ + =5 ∙6

where Eon is the energy spent within a turn-on process and f is the frequency of repeat turn-on
processes.
The transistor turn-off loss is calculated as:

3_ + 77 =5 77 ∙6

where Eoff is the energy spent within a turn-off process and f is the frequency of repeat turn-on
processes.
The energy loss components Eon and Eoff are calculated based on the information of the gate current,
input/output/reverse transfer capacitances, and gate charges of the MOSFET devices. Sometimes these
loss components are plotted as graphs in the datasheet. The actual loss is herein calculated based on
datasheet information rather than slopes of current and voltage transient.
Please note that the results of the calculated loss components, especially the switching loss, are only
approximation and many factors have not been taken into account. The accuracy of the results depends
on the accuracy of the data in the device database, and on the actual circuit operating conditions.

6.2. Junction temperature

The junction temperature during operation with the largest current is required in order to prevent
device overheating and damage. A steady-state model is considered herein without including the
thermal dynamics. The operation is considered in steady-state and each component of power loss is
estimated as a constant DC current source. The loss components add up as currents within a node and
the outcome is applied to the resistive model of the heat transfer. The heat encounters a thermal
resistance from junction to case, next from case to heatsink, and finally from heatsink to the ambient.
The “voltage” in this electric equivalent model would correspond to the actual steady-state
temperature. This is illustrated in Fig. 7. More advanced dynamic models include Cauer and Foster R-C
networks.
Tj
V

A Pcond_Q
Rth_jc Rth_cs R_heatsink
A Psw_Q
0.45 0.24 0.1 T_ambient
A Pcond_D
40
A Psw_D

Fig. 7 Steady-state thermal model from power loss to junction temperature.

9
 An example is shown in Fig. 8. It can be seen that capacitors of infinity value would produce the same
steady-state model. However, dynamic models are not useful for estimating the thermal induce damage
to the device. Rather, they can serve for reliability and lifetime calculation.

Fig. 8 Dynamic thermal model for the MOSFET.

6.3. Analysis of results

The following data needs to be collected (counting is continued from previous list):
(9) Collect data and fill Table IV for 1 A, 5 A, 10 A, 15 A, 20 A, 25 A.
(10) Draw a graphical representation of the variation of each performance index with the load current.
(11) When do we get the largest junction temperature? How would you compare this result with a good
design practice?
(12) How could you change the provided model to also provide efficiency as a direct result?
(13) Back to the case with 20 A load, Change the switching frequency from 20 kHz to 40 kHz. What do you
notice?

Performance Index Value

• Transistor switching loss
• Transistor conduction loss
• Diode switching loss
• Diode conduction loss
• Junction temperature
TABLE IV

Optional:
(14) Change the used MOSFET with an IGBT device and explore changes in power loss and junction
temperature. This model is also provided.

10