NEW CONCEPTS FOR DIGITAL POWER ELECTRONICS

CLASSICAL CHARACTERISTICS PARAMETERS

 Traditional parameters used in power electronics are the
power factor (PF), power transfer efficiency (η), total
harmonic distortion (THD) and ripple factor (RF). Using
these parameters has successfully described the
characteristics of power (generation, transmission,
distribution, protection and harmonic analysis) systems
and most drive (AC and DC motor drives) systems.

 Power Factor (PF)

Power factor is defined by the ratio of real power P over
the apparent power S:

PF = P / S = cos θ

 Power-Transfer Efficiency (η)

Power-transfer efficiency (η) is defined by the ratio of
output power PO over the input power Pin:

η = PO / Pin

The output power PO is received by the load. The input

power Pin is usually generated by the power supply
source. Both the input power Pin and output power PO are
real power.

 Total Harmonic Distortion (THD)

Aperiodical AC wave form usually possesses various order
harmonics. Since the instantaneous value is periodically
repeating in fundamental frequency f (or ω=2πf ), the
corresponding spectrum in the frequency domain consists
of discrete peaks at the frequencies nf (or nω=2nπf ),
where n=1, 2, 3,…∞. The first-order component (n=1)
corresponds to the fundamental component V1. The total
harmonic distortion (THD) is defined by the ratio of the
 sum of all higher-order harmonics over the fundamental
harmonic V1:

where all Vn (n=1, 2, 3,…∞) are the corresponding rms

values.

 Ripple Factor (RF)

A DC waveform usually possesses DC component Vdc
and various high-order harmonics. These harmonics make
the variation (ripple) of the DC waveform. Since the
instantaneous value is periodically repeating in
fundamental frequency f (or ω=2πf ), the corresponding
spectrum in the frequency domain consists of discrete
peaks at the frequencies nf (or nω=2nπf ), where n=0, 1, 2,
3,…∞. The zeroth-order component (n=0) corresponds to
the DC component Vdc. The ripple factor (RF) is defined
by the ratio of the sum of all higher-order harmonics over
the DC component Vdc:

where all Vn (n=1, 2, 3,…∞) are the corresponding rms

values.
NEW CHARACTERISTICS PARAMETERS
INTRODUCTION

 From the introduction in previous concepts, we have got

the impression of the four important factors: power factor
(PF), power transfer efficiency (η), total harmonic
distortion (THD) and ripple factor (RF) that well describe
the characteristics of power systems.
 Unfortunately, all these factors are not available to be used
to describe the characteristics of power DC/DC converters
and other high-frequency switching circuits.

 Power DC/DC converters have usually equipped by a DC

power supply source, pump circuit, filter and load. The
load can be of any type, but most of the investigations are
concerned with resistive load, R, and back electromagnetic
force (EMF) or battery.

 It means that the input and output voltages are nearly pure
DC voltages with very small ripple (e.g. output voltage
variation ratio is usually less than 1%). In this case, the
corresponding RF is less than 0.001, which is always
ignored.

 Since all power is real power without reactive power jQ,

we cannot use power factor PF to describe the energy-
transferring process.

 Since DC components exist without harmonics in input

and output voltage, THD is not available to be used to
describe the energy-transferring process and waveform
distortion.

 To simplify the research and analysis, the condition

without power losses during power-transferring process is
assumed to investigate power DC/DC converters.

 Consequently, the efficiency η=1 is 100% for most of the

description of power DC/DC investigation. Otherwise,
efficiency η must be considered for special investigations
regarding the power losses.

 In general conditions, all four factors are not available to

apply in the analysis of power DC/DC converters. This
 situation makes the designers of power DC/DC converters
confusing for very long time.

 So it is necessary to find other new parameters to describe

the characteristics of power DC/DC converters.

 Energy storage in power DC/DC converters has been paid

attention long time ago.

 Unfortunately, there is no clear concept to describe the

phenomena and reveal the relationship between the stored
energy and the characteristics of power DC/DC converters.

 Since 2005 a new concept, “energy factor (EF)” is

presented, and researched the relationship between EF and
the mathematical modeling of power DC/DC converters.

 EF is a new concept in power electronics and conversion

technology, which thoroughly differs from the traditional
concepts such as power factor (PF), power transfer
efficiency (η), total harmonic distortion (THD) and ripple
factor (RF). EF and the sub-sequential other parameters
can illustrate the system stability, reference response and
interference recovery. This investigation is very helpful for
system design and DC/DC converters characteristics
foreseeing.
Before the definition of EF we will introduce some factors:
PUMPING ENERGY (PE)

 Assuming the instantaneous input voltage and current of a

DC/DC converter are, v1(t) and i1(t), and their average
values are V1 and I1, respectively. The instantaneous
output voltage and current of a DC/DC converter are,
respectively, v2(t) and i2(t), and their average values are
V2 and I2, respectively. The switching frequency is f , the
 switching period is T =1/f , the conduction duty cycle is k
and the voltage transfer gain is M =V2/V1.

 All power DC/DC converters have pumping circuit to

transfer the energy from the source to some energy-storage
passive elements, e.g. inductors and capacitors.

 The pumping energy (PE) is used to calculate the input

energy in a switching period T. Its calculation formula is:

where is the average value of the input current if

the input voltage V1 is constant. Usually, the input average
current I1 depends on the conduction duty cycle.

STORED ENERGY (SE)

 Energy storage in power DC/DC converters has been paid

attention long time ago. Unfortunately, there is no clear
concept to describe the phenomena and reveal the
relationship between the stored energy and the
characteristics of power DC/DC converters.

 Stored Energy in Continuous Conduction Mode

If a power DC/DC converter works in the continuous
conduction mode (CCM), then all inductor’s currents
and capacitor’s voltages are continuous (not to be equal
to zero).
 Stored Energy (SE) in inductor

The stored energy in an inductor is:

The stored energy across a capacitor is,

Therefore, if there are nL inductors and nC capacitors, the total
stored energy in a DC/DC converter is:

Usually, the stored energy (SE) is independent from the

switching frequency f (as well as the switching period T). Since
the inductor currents and the capacitor voltages rely on the
conduction duty cycle k, the stored energy does also rely on the
conduction duty cycle k. The stored energy (SE) is used as a
new parameter in further description.

 Capacitor–Inductor Stored Energy Ratio (CIR)

Most power DC/DC converters consist of inductors and
capacitors. Therefore, the capacitor–inductor stored energy
ratio (CIR) can be defined as follows:

Energy Losses (EL)

Usually, most analyses applied in DC/DC converters are
assuming no power losses, i.e. the input power is equal to the
output power, Pin =Po or V1I1 =V2I2, so that pumping energy is
equal to output energy in a period PE =V1I1T =V2I2T. It
corresponds to the efficiency η=V2I2T/PE =100%.
Particularly, power losses always exist during the conversion
process. They are caused by the resistance of the connection
cables, resistance of the inductor and capacitor wire, and power
losses across the semiconductor devices. We can sort them as
the resistance power losses Pr , passive element power losses Pe
and device power losses Pd. The total power losses are:

Ploss = Pr + Pe + Pd
and
 Pin = PO + Ploss = PO + Pr + Pe + Pd = V2I2 + Pr + Pe + Pd
Therefore,

EL = Ploss × T = (Pr + Pe + Pd)T

The energy losses (EL) is in a period T:

Since the output energy in a period T is (PE – EL)T, we can

define the efficiency η to be:

If there are some energy losses (EL>0), then the efficiency η is

smaller than unity. If there are no energy losses during
conversion process (EL=0), then the efficiency η is equal to
unity.

ENERGY FACTOR (EF)

 As described previously that the input energy in a period T

is the pumping energy PE =Pin ×T =Vin Iin ×T.

 The energy factor (EF) is the ratio of the stored energy

(SE) over the pumping energy (PE):
  Energy factor (EF) is a very important factor of a power
DC/DC converter. It is usually independent from the
conduction duty cycle k, and proportional to the switching
frequency f (inversely proportional to the) since the
pumping energy (PE) is proportional to the switching
period T.

VARIATION ENERGY FACTOR (EFV)

 We also define that the energy factor for the variation of
stored energy (EFV) is the ratio of the variation of stored
energy over the pumping energy:

Energy factor (EF) and variation energy factor (EFV) are

available to be used to describe the characteristics of power
DC/DC converters.

TIME CONSTANT, τ ,AND DAMPINGTIME CONSTANT, τ d

We define the time constant, τ, and damping time constant, τd,

of a power DC/DC converters.

 Time Constant, τ
The time constant, τ, of a power DC/DC converter is a
new concept to describe the transient process of a DC/DC
converter. If there are no power losses in the converter, it
is defined as:

This time constant (τ) is independent from switching

frequency f (or period T =1/f ). It is available to estimate
 the converter responses for a unit-step function and
impulse interference.

If there are power losses and η<1, it is defined as:

The time constant (τ) is still independent from switching

frequency f (or period T =1/f ) and conduction duty cycle
k.

If there is no power loss and η=1, then Equation (11)

becomes Equation (10). Usually, the higher the power
losses (the lower efficiency η), the larger the time constant
τ since usually CIR>1.

 Damping Time Constant, τ d

The damping time constant, τd, of a power DC/DC
converter is a new concept to describe the transient process
of a DC/DC converter. If there are no power losses, it is
defined as:

This damping time constant (τd) is independent from

switching frequency f (or period T =1/f ). It is available to
estimate the oscillation responses for a unit-step function
and impulse interference.

If there are power losses and η<1, then it is defined as:

 The damping time constant (τd) is also independent from
switching frequency f (or period T =1/f ) and conduction
duty cycle k. If there is no power loss and η=1, then
Equation (13) becomes Equation (12). Usually, the higher
the power losses (the lower efficiency η), the smaller the
damping time constant (τd) since usually CIR>1.

 Time Constant Ratio, ξ

The time constant ratio, ξ, of a power DC/DC converter
is a new concept to describe the transient process of a
DC/DC converter. If there are no power losses, it is
defined as:

ξ = τd / τ = CIR ………….…… (14)

 This time constant ratio is independent from switching

frequency f (or period T =1/f ). It is available to estimate
the oscillation responses for a unit-step function and
impulse interference.

If there are power losses and η<1, it is defined as:

 The time constant ratio is still independent from switching

frequency f (or period T =1/f ) and conduction duty cycle
k. If there is no power loss and η=1, then Equation (15)
becomes Equation (14). Usually, the higher the power
losses (the lower efficiency η), the smaller the time
constant ratio (ξ) since usually CIR>1. From this analysis,
most power DC/DC converters with lower power losses
possess the output voltage oscillation when the converter
operation state changed. Vice versa, power DC/DC
converters with high power losses will possess the output
 voltage smoothening when the converter operation state
changed.

 we can estimate the unit-step function response using the

ratio ξ. If the ratio ξ is equal to or smaller that 0.25 the
corresponding unit-step function response has no
oscillation and overshot. Vice versa, if the ratio ξ is greater
that 0.25 the corresponding unit-step function response has
oscillation and overshot. The higher the value of ratio ξ,
the heavier the oscillation with higher overshot.

Mathematical Modeling for Power DC/DC Converters

The mathematical modeling for all power DC/DC converters is:

where M is the voltage transfer gain (M =V2/V1); τ, the time

constant in Equation (11); τd, the damping time constant in
Equation (13), and s, the Laplace operator in the s-domain.

Using this mathematical model of power DC/DC converters, it

is significantly easy to describe the characteristics of power
DC/DC converters.

In order to verify this theory, few converters are investigated to

demonstrate the characteristics of power DC/DC converters and
applications of the theory.

EXAMPLES OF APPLICATIONS

In order to demonstrate the parameters’ calculation some

examples are presented in this section. A buck converter, super-
lift Luo-converter, boost converter, buck–boost converter and
positive-output Luo-converter are used for this purpose.
1. A Buck Converter in CCM

We will carefully discuss the mathematical model for buck

converter in various conditions in this sub-section.

Buck Converter without Energy Losses (rL =0_)

A buck converter shown in Figure 1 has the components values:

V1 =40V, L=250μH with resistance rL =0Ω, C =60μF, R=10Ω,
the switching frequency f =20 kHz (T =1/f =50μs) and
conduction duty cycle k =0.4. This converter is stable and works
in CCM.

Figure 1 Buck converter.

Therefore, we have got the voltage transfer gain M =0.4, i.e. V2

=VC =MV1 =0.4× 40=16V, IL =I2 =1.6A, Ploss =0W and I1 =0.64
A. The parameter EF and others are listed below:
By cybernetic theory, since the damping time constant (τd) is
much larger than the time constant (τ), the corresponding ratio
(ξ) is 25>>0.25. The output voltage has heavy oscillation with
high overshot.

Substituting the above calculated parameters in equation 16 and

applying a unit step input to see the response of the system
(Buck converter), the result is shown in figure (2). Figure (3)
shows the practical result of the same case.

Figure 2 Buck converter unit-step function response.

 Figure 3 Unit-step function responses of buck converter (experiment).

Buck Converter with Small Energy Losses (rL =1.5_)

A buck converter shown in Figure 1 has the components values:
V1 =40V, L=250μH with resistance rL =1.5Ω, C =60μF, R=10Ω,
the switching frequency f =20 kHz (T =1/f =50μs) and
conduction duty cycle k =0.4. This converter is stable and works
in CCM.
Therefore, we have got the voltage transfer gain M
=0.35(k*RL/(RL+rL)), i.e. V2 =VC=MV1=0.35×40=14V,
2
IL=I2=1.4A, Ploss=IL ×rL = 1.42 x1.5 = 2.94W and I1 =0.564A.
The parameter EF and others are listed below:
By cybernetic theory, since the damping time constant (τd) is
much larger than the time constant (τ), the corresponding ratio
(ξ) is 1.31>>0.25. The output voltage has heavy oscillation with
high overshot
Substituting the above calculated parameters in equation 16 and
applying a unit step input to see the response of the system
(Buck converter), the result is shown in figure (4). Figure (5)
shows the practical result of the same case.

Figure 4 Buck converter unit-step function response (rL =1.5Ω)

 Figure 5 Unit-step function responses of buck converter (rL
=1.5Ω experiment).

Buck Converter with Energy Losses (rL =4.5 Ω)

A buck converter shown in Figure 1 has the components values:
V1 =40V, L=250μH with resistance rL =4.5Ω, C =60μF, R=10Ω,
the switching frequency f =20 kHz (T =1/f =50μs) and
conduction duty cycle k =0.4. This converter is stable and works
in CCM.

Therefore, we have got the voltage transfer gain M =0.27586

(0.4*10/14.5), i.e. V2 =VC =MV1 = 0.27586×40=11V, IL =I2
=1.1A, Ploss =IL2L ×rL =1.12 ×4.5=5.445W and I1 =0.4386A.
The parameter EF and others are listed below:
By cybernetic theory, since the damping time constant (τd) is the
critical value, the corresponding ratio (ξ) is equal to 0.25. The
output voltage has no oscillation.

Substituting the above calculated parameters in equation 16 and

applying a unit step input to see the response of the system
(Buck converter), the result is shown in figure (6).

Figure 6 Buck converter unit-step function response (rL =4.5Ω).

2. A Boost Converter in CCM (no Power Losses)

A boost converter shown in Figure 7 has the components values:

V1 =40V, L=250μH, C =60μF, R=10Ω, the switching frequency
f =20 kHz (T =1/f =50μs) and conduction duty cycle k =0.6. This
converter is stable and works in CCM.

Figure 7 Boost converter.

Therefore, we have got the voltage transfer gain M

=1/(1−k)=2.5, i.e. V2 =VC =V1/(1−k)=100V, I2 =10A, and I1 =IL
=25A. The parameter EF and others are listed below:
By cybernetic theory, since the damping time constant (τd) is
much larger than the time constant (τ), the corresponding ratio
(ξ) is 3.84>0.25. The output voltage has heavy oscillation with
high overshot.
Substituting the above calculated parameters in equation 16 and
applying a unit step input to see the response of the system
(Boost converter), the result is shown in figure (8).

Figure 8 Boost converter unit-step responses.

3. A Buck–Boost Converter in CCM (No Power Losses)

A boost converter shown in Figure 9 has the components values:
V1 =40V,
L=250μH, C =60μF, R=10Ω, the switching frequency f =20 kHz
(T =1/f =50μs) and conduction duty cycle k =0.6. This converter
is stable and works in CCM.

Figure 9 Buck–boost converter.

Therefore, we have got the voltage transfer gain M
=k/(1−k)=1.5, i.e. V2 =VC =-kV1/(1−k)=60V, I2 =6A, I1 =9A and
IL =15A. The parameter EF and others are listed below:

By cybernetic theory, since the damping time constant (τd) is

much larger than the time constant (τ), the corresponding ratio
(ξ) is 3.84>0.25. The output voltage has heavy oscillation with
high overshot.

Substituting the above calculated parameters in equation 16 and

applying a unit step input to see the response of the system
(Buck-Boost converter), the result is shown in figure (10).
Figure 10 Buck–boost converter unit-step responses.