Electronics 10 02119 v2
Electronics 10 02119 v2
Electronics 10 02119 v2
Article
Development of a High-Power-Factor Power Supply for an
Atmospheric-Pressure Plasma Jet
Chi-Feng Su 1 , Chih-Tung Liu 2 , Jong-Shinn Wu 2 and Ming-Tzu Ho 1, *
1 Department of Engineering Science, National Cheng Kung University, Tainan City 70101, Taiwan;
[email protected]
2 Department of Mechanical Engineering, National Yang Ming Chiao Tung University,
Hsinchu 30010, Taiwan; [email protected] (C.-T.L.); [email protected] (J.-S.W.)
* Correspondence: [email protected]
Abstract: This paper presents the design and implementation of a miniaturized high-voltage power
supply with power factor correction (PFC) for atmospheric-pressure plasma jet (APPJ) applications.
The sinusoidal output frequency and voltage of the power supply can be controlled independently
from 16 to 24 kHz and from 1 to 10 kVpeak , respectively. A helium APPJ load is used to assess the
performance of the developed power supply. It is shown that the developed high-voltage power
supply operates effectively, and the designed PFC converter improves the input current distortion of
the power supply. Not only the power factor of the power supply is increased from 0.41 to 0.95, but
it also provides a low-ripple DC voltage, which reduces the high-voltage ripple of the output from
730 to 50 Vp-p . In this paper, the proposed design integrates the PFC converter into the high-voltage
power supply so that the developed power supply has better electrical characteristics and the overall
power supply can be significantly miniaturized.
Keywords: atmospheric-pressure plasma jet (APPJ); plasma power supply; power factor correction (PFC)
Citation: Su, C.-F.; Liu, C.-T.; Wu,
J.-S.; Ho, M.-T. Development of a
High-Power-Factor Power Supply for
an Atmospheric-Pressure Plasma Jet. 1. Introduction
Electronics 2021, 10, 2119. https://
By continually improving the design of plasma generators for decades, the atmospheric-
doi.org/10.3390/electronics10172119
pressure plasma has been compliant with the low-temperature work to ensure safe opera-
tions. Among atmospheric-pressure plasma designs, the atmospheric-pressure plasma jet
Academic Editor: Hani Vahedi
(APPJ) [1,2] is the most commonly used for many applications. The atmospheric-pressure
plasma jet produces plasma with a narrow area and high power density, suitable for single-
Received: 2 July 2021
Accepted: 25 August 2021
point processing. Therefore, many novel applications related to APPJ can be found in the
Published: 31 August 2021
biomedical field, including root canal disinfection [3], tooth bleaching [4], and treatment
of human extremities [5]. Handheld commercial APPJ devices will become popular for
Publisher’s Note: MDPI stays neutral
biomedical sterilization or surface treatment in the near future. For ease of use during
with regard to jurisdictional claims in
treatment, it is crucial to miniaturize the plasma power supply.
published maps and institutional affil- This paper aims to design a miniaturized plasma power supply. The expected output
iations. of the designed power supply has a high-voltage AC sinusoidal waveform, the output
voltage is adjustable at a fixed frequency, and it has a burst mode feature to protect the
plasma generator when the temperature is too high. We propose to improve the conven-
tional design in the commercially available product [6] by changing the autotransformer
Copyright: © 2021 by the authors.
and high-voltage transformer to improve the disadvantages of having a large volume so
Licensee MDPI, Basel, Switzerland.
that the space for circuit composition is significantly reduced.
This article is an open access article
This study first introduces the model of the plasma load. In [7–11], the plasma
distributed under the terms and models were simplified as combinations of several capacitors and threshold voltages.
conditions of the Creative Commons Here, we use the model introduced in [12] with the relationships between the structure
Attribution (CC BY) license (https:// and capacitors in [11]. The value of the capacitor at a non-discharged state is calculated
creativecommons.org/licenses/by/ approximately by measuring the dimensions of the APPJ structure and the arrangement
4.0/). of the ground electrode. The obtained parameters of the model are then used in the
simulation. In addition, the high-voltage transformer model [13] and the inverter circuit are
also built in the simulation. Based on the simulation-based design, the feasible transformer
specifications are determined by simulating the gain curves of the input and output. Finally,
the plasma model is combined with the designed high voltage transformer circuit, and the
design of the high voltage power supply is validated by the simulation.
There are many advantages in our proposed design over conventional designs. In
the design, we redesign the high-voltage transformer to have a higher resonant frequency
so that it does not have a drastically changing gain in the operating range; this way we
can have advantages such as reduced operating current, inductance value, and coil wire
diameter. The heat and Electromagnetic Interference (EMI) issues can also be improved
by reducing the transformer’s size. Although this reduces the boost ability of the inverter,
the front-end converter can compensate for this reduced capability. Through such a
design compromise, we can reduce the volume of the back-end inverter. Furthermore,
the external DC power supply is integrated into the designed high-voltage power supply.
Most plasma power supplies are based on stand-alone DC power combined with high-
voltage power supplies [14,15] or Marx generators [16,17], and are thus bulky and heavy,
with limited portability. To overcome these disadvantages, in this study, a flyback PFC
converter [14,18] is use to replace the front-end DC power supply. Since the output
voltage of a flyback converter is arbitrarily step-up/down, it is different from the boost
system that is only used when the output voltage is higher than the input voltage. It
is very suitable for miniaturization applications as a front-end circuit. It can not only
replace the external DC power supply but also improve the capability of power factor
correction [19]. Based on the above two improvements, the overall power supply can be
further significantly miniaturized.
For practical applications, we used the helium APPJ as an output load, compared the
difference between simulation and experimental results, and validated the effectiveness
of the electrical function through measurement data. The performance of APPJ load was
assessed by optical emission spectroscopy (OES) [20–22]. Finally, in comparison with the
case study, the test results show that it is a successful miniature integrated design.
The rest of this paper is organized as follows. Section 2 discusses the structure
and equivalent electrical model of a plasma jet. Section 3 describes the design steps of
power supply circuits. In Section 4, a description of the experimental setup is given and
then the results of simulations and experiments are compared. Section 5 presents the
concluding remarks.
the plasma discharge current rises quickly as long as the applied voltage is greater than the
TVS voltage level. The TVS stays at the cut-off condition when the applied voltage is lower
Electronics 2020, 9, x FOR PEER REVIEW
than the TVS voltage level.
Figure 2. Equivalent model of a plasma jet: (a) dependent voltage source, (b) LC circuit and
Figure 2. Equivalent model of a plasma jet: (a) dependent voltage source, (b) LC circuit and
(c) TVS component.
component.
In this paper, we used the third type model to do the simulation-based design to
quickly determine the approximate working range of the system. Since the plasma jet
The parameter α adjusts the slope of the discharge current 𝐼 . This par
configuration does not use the same material along the entire length in this study, the value
depends onbethe
of Cg cannot plasmadirectly.
calculated load, whose
Instead,range
we usedis the
typically between
approximate method1 and 12. In addit
proposed
total
in [11]current ILoad(t)
to calculate theof the equivalent
capacitance model
values of can be obtained
this structure, by Kirchhoff’s
through assuming that the rule, a
given
surfacesbydefined by the specific radius have the same potentials (boundary conditions).
Capacitances of Cd and Cg in the non-discharged state can be obtained by
dVpower t dVd t
ILoad t =ICs +ICd =Cs · +Cd ·
2π ·ε 0 ·ε r_quartz ·l dt dt
Cd = (3)
ln bc
where Vpower(t) represents the input voltage and Vd(t) represents the voltage of Cd
such a model is accurate, the analysis of this model becomes more complicated.
The second type of model [11] consists of a capacitor Cg, an inductor Lg in p
and a series resistor Rg , as shown in Figure 2b. Here Rg represents the conductivit
current return paths. The model reflects the transient characteristics in the sim
Electronics 2021, 10, 2119 5 of 21
2π ·ε 0 ·ε r_air ·l
Cg = (4)
ln ba
where ”0 is the dielectric constant of vacuum, ”r_quartz is the relative dielectric constant of
quartz, and ”r_air is the dielectric constant of air. The parameter Cs can be obtained from
the connector specifications or by measuring the connection line.
Figure 3. Schematic diagrams of input current distortion: (a) before improvement and (b) ideal improvement.
Figure 3. Schematic diagrams of input current distortion: (a) before improvement and (b) ideal
improvement. In this paper, the solutions mentioned above are adopted in the design of the power
supply. A schematic overview of the proposed power supply circuit is shown in Figure 4,
in which
In this paper, the adjustable
the solutions voltage and
mentioned frequency
above can be controlled
are adopted by resistances
in the design Rv and
of the power
Rf , respectively. In this design, the high-voltage secondary side only has the output
supply. A schematic overview of the proposed power supply circuit is shown in Figure 4,
voltage sensing, and there are no control and no other feedback circuits. The design of the
in which the adjustable
transformer voltage and frequency
of the high-voltage powercan be controlled
supply and the PFC by resistances
converter is givenRbelow.
v and Rf,
respectively. In this design, the high-voltage secondary side only has the output voltage
sensing, and there are no control and no other feedback circuits. The design of the
transformer of the high-voltage power supply and the PFC converter is given below.
In this paper, the solutions mentioned above are adopted in the design of the power
supply. A schematic overview of the proposed power supply circuit is shown in Figure 4,
in which the adjustable voltage and frequency can be controlled by resistances Rv and Rf,
respectively. In this design, the high-voltage secondary side only has the output voltage
Electronics 2021, 10, 2119 sensing, and there are no control and no other feedback circuits. The design of 6 ofthe
21
transformer of the high-voltage power supply and the PFC converter is given below.
This design can reduce the resonance current such that the transformer size can be
reduced. Assuming that the orange line at the bottom of Figure 5 is the gain curve of the
transformer, and the blue line is the DC voltage (Vdc_Bus) controlled by the PFC converter
from the previous stage, the gain variation of the transformer is compensated to achieve
regulation of the output voltage amplitude (as illustrated by the green dashed line). The
design of the transformer with the desired resonance characteristic curve is the primary
concern here. Considering the transformer equivalent circuit in [15], the typical equivalent
model of a transformer is shown in Figure 6a. The model includes an ideal transformer
(with turns ratio N), primary side inductance LP, magnetizing inductance Lm, leakage
inductance Llea, secondary-side inductance LS, winding resistances (Rp and Rs), and
parasitic capacitances on the primary side (C1), secondary side (C2), and both sides (C1,2).
The simplified circuit [13] shown in Figure 6b is used for the transformer design.
Assuming ideal conditions (i.e., neglecting winding resistance RSC and conversion loss
resistance RC), representing by the Norton equivalent circuit shown in Figure 6c, the
resonant frequency in the circuit is determined by the equivalent inductance L1 and the
total stray capacitance Cstr.
Figure 6. Simplified analysis of a transformer with: (a) typical model, (b) transferring to the primary
Figure 6. Simplified
side circuit and (c) the analysis of a transformer
Norton equivalent circuit. with: (a) typical model, (b) transferring to the p
side circuit and (c) the Norton equivalent circuit.
The simplified circuit [13] shown in Figure 6b is used for the transformer design.
Assuming ideal conditions (i.e., neglecting winding resistance RSC and conversion loss
resistance RC ), representing by the Norton equivalent circuit shown in Figure 6c, the
resonant frequency in the circuit is determined by the equivalent inductance L1 and the
total stray capacitance Cstr .
Therefore, as long as the primary-side inductance LP and the natural resonant fre-
quency fr are determined, the simplified transformer model can be derived using the
following equations.
L P = Llea + Lm (5)
Llea · Lm
L1 = (6)
Llea + Lm
1
fr = √ (7)
2π L1 ·Cstr
where the resonant parameters L1 and Cstr of the transformer are obtained from the sim-
ulation results. These two parameters can define the manufacturing specifications of the
transformer. In addition, the stray parameters of the transformer depend on the structure,
the manufacturing method, and the operating frequency. Especially, most magnetic compo-
nents still need manual production, and thus the manufacturing error must be considered.
Generally, the manufacturer only gives the specification of the parameter LP , the range of
the leakage inductance Llea and the turns ratio N.
For example, LP is approximately 5 mH, Llea is about 15% of LP , and N is 20. Here,
there is no limit to the stray capacity range, and this may have a significant impact on the
performance of the high voltage transformer. To ensure that the actual characteristics agree
with the simulated results, the resonant frequency and stray capacitance must be tested
after the transformer is manufactured.
Next, various resonant frequencies are calculated to obtain the corresponding stray
capacitance Cstr . Table 1 shows the stray parameters under various resonant frequencies.
Electronics 2020, 9, x FOR PEER REVIEW 9 of 21
In the simulation circuit, the effects of high-voltage probes and external capacitors are
considered. The simulation circuit model of the overall inverter is shown in Figure 7.
Table 1. Stray parameters with different resonant frequencies.
Table 1. StrayFrequency
Resonance parameters with different resonant frequencies.
Lm Llea L1 Cstr
(fr) Frequency
Resonance
Lm Llea L1 Cstr
30 kHz (fr ) 44.2 nF
35 kHz
30 kHz 32.5
44.2 nF nF
35 kHz
4.25 mH 750 μH 638 μH 32.5 nF
40 kHz 4.25 mH 750 µH 638 µH 24.9 nF
40 kHz 24.9 nF
45 kHz
45 kHz
19.6 nF
19.6 nF
Figure
Figure 7. 7. Simulationcircuits
Simulation circuits of
of the
the high-voltage
high-voltagepower
powersupply.
supply.
Figure 8 shows the simulation results of the gain curves with operating frequencies
of 16 to 24 kHz under various resonant frequencies (simulation results for 30, 35, 40, and
45 kHz are shown with solid lines and actual sample test results are shown with the red
dashed line). The simulation results imply that if a gentle output gain curve is desired for
the operating frequency range, the fabricated transformer must have a resonant frequency
Figure 7. Simulation circuits of the high-voltage power supply.
Electronics 2021, 10, 2119 9 of 21
Figure 8 shows the simulation results of the gain curves with operating frequencie
of 16 to 24 kHz under various resonant frequencies (simulation results for 30, 35, 40, and
45 kHz Figure 8 showswith
are shown the simulation
solid linesresults of the gain
and actual curves
sample testwith operating
results frequencies
are shown with the red
dashed line). The simulation results imply that if a gentle output gain curve isand
of 16 to 24 kHz under various resonant frequencies (simulation results for 30, 35, 40, desired fo
45 kHz are shown with solid lines and actual sample test results are shown with the red
the operating frequency range, the fabricated transformer must have a resonant frequency
dashed line). The simulation results imply that if a gentle output gain curve is desired for
ofthe
40operating
to 45 kHz. This means that the stray capacitance Cstr should be controlled to be
frequency range, the fabricated transformer must have a resonant frequency
approximately
of 40 to 45 kHz. 19.6
Thisto means
24.9 nF. Thus,
that the manufacturing
the stray capacitance Cstrparameters of the transformer
should be controlled to be are
defined in Table 1. Finally, the actual parameters of the fabricated
approximately 19.6 to 24.9 nF. Thus, the manufacturing parameters of the transformer transformer are a
follows: the in
are defined resonant
Table 1. frequency
Finally, the =actual
40.3 parameters
kHz, Llea = 0.77
of themH, Lm = 4 transformer
fabricated mH, and Cstr are= as
24.2 nF.
follows: the resonant frequency = 40.3 kHz, Llea = 0.77 mH, Lm = 4 mH, and Cstr = 24.2 nF.
In order to analyze the maximum load capacitance, we refer to the reference [23]. The
In order to analyze the maximum load capacitance, we refer to the reference [23]. The
conversion ratio of the capacitance value between the primary and secondary sides of an
conversion ratio of
ideal transformer the
is 1: N2capacitance value
(N is turn ratio). between
Affected theturns
by the primary and
ratio, it cansecondary
be seen thatsides
as of an
Electronics 2020, 9, x FOR PEER REVIEWideal transformer
output is 1: Nincreases,
load capacitance
2 (N is turn
the ratio). Affected
capacitance valueby the primary
of the turns ratio, 10
it can be seen tha
side conversion
asalso
output loadso
increases capacitance increases,
that the resonance the capacitance
frequency valueside
of the primary of the
willprimary
be lower.side conversion
We use
frequency,
also as shown
the increases
parameters thein
soofthat Figure
the
actual 9, andfrequency
resonance
transformerit to
can be seen
draw that
of the
the gain the lower
primary
curve of side the
willresonance
the lower be lower. frequen
resonance We use
the frequency,
thehigher as shown
the
parameters gain in Figure
value.
of the actual9,transformer
and it can be to
seen that the
draw thelower
gain the resonance
curve of the frequency,
lower resonance
the higher the gain value.
Figure
Figure 9. 9. Selecting the
Selecting the frequency
frequency curve of the
curve of maximum gain. gain.
the maximum
In order to avoid the large current of the high gain damaging the transformer, we
the actual transformer parameters in the simulation and change the output capacitanc
obtain the gain curves of different resonant frequencies. The frequency curve of
maximum gain is selected, then the maximum load capacitance value can be obtai
Electronics 2021, 10, 2119 10 of 21
In order to avoid the large current of the high gain damaging the transformer, we use
the actual transformer parameters in the simulation and change the output capacitance
to obtain the gain curves of different resonant frequencies. The frequency curve of the
maximum gain is selected, then the maximum load capacitance value can be obtained
through reverse calculation. The maximum gain is assumed not to exceed twice the original
gain. In Figure 9, the maximum gain is about 70 for 29.8 kHz, almost double the original
maximum gain for 40.3 kHz. Therefore, the frequency of 29.8 kHz is the lowest resonance
frequency allowed. Table 2 shows the corresponding calculation results. The maximum
capacitance of the secondary side is 110 pF, the no-load output capacitance is 60.2 pF, and
the high voltage probe and other calibration capacitance is 2 pF. Thus, it is shown that
the secondary side capacitance margin is 47.8 pF, and it is recommended to use a plasma
generator with a load capacitance less than this value.
Primary Side Resonance Primary Side Resonant Capacitor Secondary Side Total
Frequency (Cstr ) Capacitance
29.8 kHz 44 nF 110 pF
40.3 kHz 24.1 nF 35 pF
In addition, one can compensate the resonant frequency through a passive filter to
solve the matching problem between the power supply and the plasma jet. Due to a high
number of turns in the output winding of the high voltage transformer, the inductance
value of the output side is considerable. If an inductance device is used in a passive filter
to compensate for the resonance frequency on the secondary side, its inductance value
must be immense, and the volume of the output stage hence increases. This problem can
be solved by increasing the inductance on the primary side, or a pF-scale high-voltage
capacitor can be used on the secondary side to do the compensation. This compensation
method can make the circuit volume relatively small.
If the duty ratio adjustment is based on the input sinusoidal voltage Vmult, the input
current will be transformed to have the same amplitude and waveform as those of the
voltage signal during the repeated power on/off switching process, as shown by the red
current waveform Ip in Figure 11. An electromagnetic interference (EMI) filter, which
filters high-frequency signals, is placed ahead of the PFC converter. The input current Iin
is then averaged to a sinusoidal-like wave. Ideally, the input current waveform is
corrected to approximate the input voltage waveform, as shown by the fine dotted line in
Figure 11.
The AC main voltage input is passed through a bridge rectifier to obtain half-wave
voltage VB, and a voltage divider circuit is employed to detect the low-voltage half-wave
signal Vmult. In addition, the signal Fb passes through an optocoupler circuit from the
Figure Figure
11. 11. The
The PFC
secondary-side PFC
circuit circuit working
working
feedback waveform.side, and is then connected to the inverting port
to waveform.
the primary
of the error amplifier.
Therefore, when the Theinputcontrol
andreference Vc candiffer,
output voltages be determined
the voltage bylevelmultiplying
of V c will the
Therefore,
output
change.of thewhen
V c error
the input
amplifier
is compared with
and output
bythe
Vmult.
signal
voltages differ, the voltage level of Vc will
Vcs and then assigned to the input R of the RS
change. flip-flop,
Vc is compared
where Vwith the signal Vcs and then assigned to the input R of the RS flip-
cs is the obtained voltage signal by the resistor Rs , proportional to the
flop, where V
primary inductor currentvoltage
cs is the obtained signalatby
Ip . The signal thethe resistor
input Rs, proportional
S is defined to thedetection
as the zero current primary
(ZCD), and the output Q is the PWM signal of the control. In a switching cycle, the PFC
circuit working process is described as below.
where Vin_pk and Iin_pk are the peak values of the input voltage and current, respectively.
The term f is the frequency of the input, which is either 50 or 60 Hz. Assuming that
the transformer of the PFC circuit is ideal, the current of the secondary side is defined
as follows:
Is (t) = n· I p (t) (n : primary-to- sec ondary turns ratio)
because this PFC converter operates in discontinuous current boundary mode [18], a single
switching period Tsw equals the sum of the turn-on time ∆ton and the turn off time ∆toff , is
shown by the current control condition in Figure 11.
1
TSW = ∆ton + ∆toff = (8)
f SW
∆IL
The voltage of the inductor is: VL = L ∆t on
. When the power switch is turned on, L
becomes Lp , the primary inductance of the transformer. The voltage VL of the primary-side
coil equals the input voltage Vin_pk . ∆IL is the current change from zero to Ip_pk , as shown
by the a-b segment in Figure 11. Then, ∆ton is given as:
L p · Ip_pk
∆ton = (9)
Vin_pk
∆IL
and when the power switch is turned off, VL = L ∆t . The voltage VL of the primary-side
off
coil equals the output voltage Vout plus the forward voltage Vout_diode of the output diode,
multiplied by the coil ratio n. Let ∆IL be the current change from zero to Ip_pk , as shown by
the c-d segment in Figure 11. Then, ∆toff is given as:
L p · Ip_pk
∆toff = (10)
n·(V out +V out_diode )
Electronics 2021, 10, 2119 13 of 21
where Vin_pk is the maximum voltage and Ip_pk is the current in a sinusoidal waveform.
Finally, the inductance value of PFC at the working frequency can be derived from
Equation (11).
4. Test Results
The internal configuration of the implemented high-voltage power supply of this
paper is shown in Figure 12. There are two independent power stages in the box. In the
first power stage, the DC bus voltage is roughly regulated in the range of 140 to 250 Vdc
from the main input voltage of 90 to 130 Vac by the PFC converter. In the second power
stage, the inverter is combined with a full-bridge circuit and a high-voltage transformer to
perform a boost function, resulting in the generation of a high-voltage sinusoidal wave.
The test results without load are shown in Figure 13. It shows that the peak voltage
value can reach 10 kVpeak at different frequencies. Moreover, the output voltage uses the
waveform controlled by the burst mode, as shown in Figure 14, and it can have 10~90%
Electronics 2020, 9, x FOR PEER REVIEW 14 of 21
control options.
Figure
Figure 12.12. Power
Power supply
supply photo
photo and
and schematic
schematic diagram.
diagram.
Figure 15 shows the working waveforms of the full-bridge circuit, and we measured
the drain to source voltages of low-side switches (S3_VDS , S4_VDS ), primary side current
(Ip_ inv ) and output voltage (Vo_peak ). It can be seen from the waveforms that the loop current
is not zero when the switches are switching, so this circuit has no soft-switching ability.
This working current is small, and the power consumption is also low despite switching
Electronics 2021, 10, 2119 14 of 21
loss. 12.
Figure The primary
Power side
supply only
photo andneeds a 0.4diagram.
schematic A peak current for the output voltage on the
Figure 12. Power supply photo and schematic diagram.
secondary side to reach 4 KVpeak without load.
Figure
Figure13.
13.Output
Outputvoltages
voltagesat different
differentfrequencies:
frequencies:(a)
(a)16 kHz
kHz(b)
(b)20 kHz
kHzand
and(c)
(c)24 kHz.
Figure 13. Output voltages atatdifferent frequencies: (a) 1616kHz (b) 2020kHz and (c) 2424kHz.
kHz.
Figure
Figure 15
15 shows
shows thethe working
working waveforms
waveforms of of the
the full-bridge
full-bridge circuit,
circuit, and
and we
we measured
measured
the drain to source voltages of low-side switches (S3_V DS, S4_VDS), primary side current
the drain to source voltages of low-side switches (S3_VDS, S4_VDS), primary side current
(I
(Ip_inv ) and output voltage (Vo_peak). It can be seen from the waveforms that the loop current
p_inv) and output voltage (Vo_peak). It can be seen from the waveforms that the loop current
is
is not zero
not zero when
when the
the switches
switches are
are switching,
switching, so
so this
this circuit
circuit has
has no
no soft-switching
soft-switching ability.
ability.
This
This working current is small, and the power consumption is also low despite
working current is small, and the power consumption is also low despite switching
switching
loss.
loss. TheThe primary
primary side
side only
only needs
needs aa 0.4
0.4 A
A peak
peak current
current for
for the
the output
output voltage
voltage on
on the
the
secondary side to reach 4 KV peak without load.
secondary side to reach 4 KVpeak without load.
design is smaller than that of PVM500, the proposed design outperforms PVM500 in all
other items. It demonstrates the practicality of the proposed design.
Figure
Figure Testsetup
16.Test
16. setup photos:
photos:(a)
(a)setup
setupbefore packaging
before and and
packaging (b) packaged prototype.
(b) packaged prototype.
4.1. Measurement and Simulation Verification
4.1. Measurement andhelium
In the tests, the Simulation Verification
flow rate was 5 slm in the APPJ. The operating peak voltage of
theIn
power supplythe
the tests, washelium
in the range
flow of 4 towas
rate 6 kV5peak
slm , and the frequency
in the APPJ. The was about 20.9peak
operating kHz.voltage
of the power supply was in the range of 4 to 6 kVpeak, and the frequency was about 20.9
kHz. The output waveforms were measured using an oscilloscope (LeCroy WS44XS-A)
via a high-voltage probe (Tektronix P5100) and a current probe (Pearson 3972). Figure 17
kHz. The output waveforms were measured using an oscilloscope (LeCroy WS44XS-A
via a high-voltage probe (Tektronix P5100) and a current probe (Pearson 3972). Figure
shows the measured voltage and current waveform. By observing the waveform, it is se
that the current phase of the APPJ is about 90° ahead of the voltage phase, when driv
Electronics 2021, 10, 2119 16 of 21
by a sinusoidal AC signal. Such a waveform is similar to the result given in [7], whi
means the APPJ is a capacitive load.
In addition, a high-frequency surge occurred in the current waveform. The cause
The output waveforms were measured using an oscilloscope (LeCroy WS44XS-A) via a
that the switching frequency is different from the natural frequency of the transforme
high-voltage probe (Tektronix P5100) and a current probe (Pearson 3972). Figure 17 shows
When the inverter
the measured voltageis and
switching, the output
current waveform. current istheforced
By observing to change
waveform, it is seendirection.
that T
parasitic inductor
the current and
phase of thecapacitor in the
APPJ is about 90 circuit
◦ loop
ahead of the produce the high-frequency
voltage phase, when driven by aresonan
in sinusoidal
the transient changes;
AC signal. thus,
Such there are
a waveform high surges
is similar in thegiven
to the result current.
in [7], which means
the APPJ is a capacitive load.
Figure18.
Figure 18.Simulation
Simulationcircuits
circuitsofofpower
powersupply,
supply,transform
transformmodel,
model,and
andplasma
plasmamodel.
model.
The
Thehelium
heliumAPPJ
APPJused
usedin
in this
this work
work has
has a breakdown voltage of about
about 44 to
to 55 kV
kVpeak
peakat
atthe
thenormal
normal pressure,
pressure, according
according toto previous
previous experimental
experimental results
results [3].[3]. Therefore,
Therefore, wewesetset
the
breakdown voltage to 4.5 kVpeak. Second, the manufacturer’s transformer parameters are
as follows: (1) the transformer turns ratio is 1:20; (2) the equivalent parameters of the
transformer are a magnetizing inductance of LP = 4.77 mH and leakage inductance of Lr =
770 μH; and (3) the resonant frequency is about 40.3 kHz.
The helium APPJ used in this work has a breakdown voltage of about 4 to 5 kV
the normal pressure, according to previous experimental results [3]. Therefore, we s
breakdown voltage to 4.5 kVpeak. Second, the manufacturer’s transformer paramete
Electronics 2021, 10, 2119 17 of 21
as follows: (1) the transformer turns ratio is 1:20; (2) the equivalent parameters o
transformer are a magnetizing inductance of LP = 4.77 mH and leakage inductance o
770 μH; and (3) the resonant frequency is about 40.3 kHz.
the breakdown voltage to 4.5 kVpeak . Second, the manufacturer’s transformer parameters
Figure 19 shows the results of the experiment and simulation. The solid lines a
are as follows: (1) the transformer turns ratio is 1:20; (2) the equivalent parameters of the
measured
transformervoltage and current
are a magnetizing (Vout_meaof
inductance , ILout_mea), and the dashed lines are the simu
P = 4.77 mH and leakage inductance of
voltage
Lr = 770and
µH;current (Vresonant
and (3) the out_sim, Iout_sim ). It can
frequency be 40.3
is about seenkHz.that the voltage in the simulatio
Figure 19 shows the results of the experiment
good agreement with the experimental measurement, and and simulation. Thethere
solid lines
is aare the
difference i
measured voltage and current (Vout_mea , Iout_mea ), and the dashed lines are the simulated
amplitude of output current between the simulation and experimental result. We
voltage and current (Vout_sim , Iout_sim ). It can be seen that the voltage in the simulation
this
hasdeviation may with
good agreement be because the discharge
the experimental consumption
measurement, and there isisa not defined
difference in the pl
in the
model, andofthe
amplitude actual
output impedance
current between the ofsimulation
the gas inand the dischargeresult.
experimental stateWe is unknown.
think this Alth
deviation may be because the discharge consumption is not defined
this paper uses a general and less precise plasma model in the simulation-based desin the plasma model,
canandallow
the actual impedance of the gas in the discharge state is unknown. Although this paper
the designer to quickly determine the gas discharge load’s approx
uses a general and less precise plasma model in the simulation-based design, it can allow
working range and waveform.
the designer to quickly determine the In gas
further analysis,
discharge we suggestworking
load’s approximate addingrangethe load
and imped
and adjusting
waveform. the connecting
In further analysis, we wire’s
suggeststray
addingcapacitance Cs to correct
the load impedance the simulation
and adjusting the re
connecting wire’s stray capacitance Cs to correct the simulation results.
Figure
Figure Actual measurement
19.Actual
19. measurement andand
simulation.
simulation.
4.2. PFC and Output Ripple
Following the PFC circuit design given in Section III, the specifications of the PFC
converter are: (1) an input voltage of 90 to 264 Vac and an output voltage of 140 to 250 Vdc;
(2) a maximum power of 60 W; (3) a minimum operating frequency of the PFC circuit of
25 kHz; and (4) a transformer core type of PQ3230 with the specifications of le = 67.8 mm,
Ae = 153.8 mm2 , Ve = 10,440 mm3 , core material: SAMWHA PL-9, maximum magnetic
flux setting: 2900 G, leakage inductance: about 80 µH, a primary turns to the secondary
turns ratio of 66:60 with an inductance value of 730 µH, and air gap = unclear. We tested
the input-end performance of the designed plasma power supply under a 110 Vac input
voltage. Figure 20a shows that the input current waveform deforms dramatically without
PFC, resulting in a power factor of just 0.41. We then replaced the autotransformer with the
PFC converter. Figure 20b shows that the input current becomes a nearly sinusoidal wave
that is in phase with the input voltage and the resulting power factor is 0.95. Figure 21
also shows that the difference in the harmonic distortion of the current before and after
improvement. With PFC, the 1st harmonic component is significantly increased, and the
others are reduced. These results verify that satisfactory power quality can be achieved
using the proposed plasma power supply.
voltage.
the Figure 20a shows
PFC converter. Figurethat
20bthe input that
shows current
thewaveform deforms
input current dramatically
becomes without
a nearly sinusoidal
PFC, resulting in a power factor of just 0.41. We then replaced the autotransformer
wave that is in phase with the input voltage and the resulting power factor is 0.95. with Figure
21the PFC
also converter.
shows Figure
that the 20b shows
difference thatharmonic
in the the input distortion
current becomes
of the acurrent
nearly sinusoidal
before and after
wave that is in phase with the input voltage and the resulting power factor is 0.95. Figure
improvement. With PFC, the 1st harmonic component is significantly increased, and the
21 also shows that the difference in the harmonic distortion of the current before and after
Electronics 2021, 10, 2119 others are reduced. These results verify that satisfactory power quality can be 18 ofachieved
21
improvement. With PFC, the 1st harmonic component is significantly increased, and the
using the proposed plasma power supply.
others are reduced. These results verify that satisfactory power quality can be achieved
using the proposed plasma power supply.
Figure 20. AC input waveforms obtained: (a) without PFC and (b) with PFC.
Figure
Figure20.
20.AC
ACinput
input waveforms obtained:(a)(a)
waveforms obtained: without
without PFC
PFC andand (b) with
(b) with PFC.PFC.
Figure21.
Figure Total harmonic
21.Total harmonic distortion
distortion in
in the
the AC
AC input
input current.
current.
Furthermore, the advantage of using a PFC converter is that it further reduces the
Figure 21.voltage
output Total harmonic distortion
ripple. The rangeinofthe AC input
voltage current.
ripple is reduced from 750 Vp-p to below
50 Vp-p , as shown in Figure 22a,b. Figure 22a shows that the high-voltage power supply
uses a rectifier. The output voltage ripple constantly fluctuates with the AC frequency
(50~60 Hz). When the output power increases or the internal capacitors age, the issue of
the output ripple becomes more serious. However, Figure 22b shows that the high-voltage
power supply uses an active PFC converter. Since the converter can generate a smooth DC
voltage on the input of the inverter, it thus significantly reduces the voltage variations on
the output ripple.
5. Conclusions
This study successfully miniaturized the plasma power supply by the proposed re-
designed high-voltage transformer and combination of a PFC converter and an inverter.
The designed power supply is smaller than the conventional design, improves the input
power factor, and reduces the output voltage ripples. In the experiments, they have been
verified and quantitatively compared. Such features of the proposed design bring advan-
tages to the high-voltage power drivers but can also facilitate the commercial development
of the applications of plasma systems. Future design of this power supply, may consider
using the soft-switching technique in the inverter. The LLC resonant circuit design concept
can be used to improve switching loss on power switches in the inverter, and unlike the
DC/DC LLC resonant circuit, the DC/AC high voltage resonant power supply must pay
special attention to the load capacitance on the secondary side. It may be necessary to limit
the range of load conditions (load and capacitance value) to maintain the advantage of soft
switching on the primary side.
Author Contributions: Conceptualization, J.-S.W. and M.-T.H.; Methodology, C.-F.S. and M.-T.H.;
Validation, C.-F.S. and C.-T.L.; Formal analysis, C.-F.S. and C.-T.L.; Data curation, C.-F.S.; Writing—
original draft, C.-F.S.; Writing—review & editing, J.-S.W. and M.-T.H.; Supervision, J.-S.W. and M.-T.H.
All authors have read and agreed to the published version of the manuscript.
Funding: This research received no external funding.
Data Availability Statement: Data collected through research presented in the paper are available
on request from the corresponding author.
Conflicts of Interest: The authors declare no conflict of interest.
References
1. Kang, H.R.; Chung, T.H.; Joh, H.M.; Kim, S.J. Effects of Dielectric Tube Shape and Pin-Electrode Diameter on the Plasma Plume in
Atmospheric Pressure Helium Plasma Jet. IEEE Trans. Plasma Sci. 2017, 45, 691–697. [CrossRef]
2. Das, S.C.; Majumdar, A.; Mukherjee, S.; Katiyal, S.; Shripathi, T. Development of power supply for atmospheric pressure plasma
jet at room temperature for bio-medical applications. In Proceedings of the 3rd International Conference on Computing for
Sustainable Global Development (INDIACom), New Delhi, India, 16–18 March 2016; pp. 1207–1209.
3. Liu, C.-T.; Wu, C.-J.; Lin, Z.-H.; Wu, J.-Y.; Wu, J.-S. Production Enhancement of Reactive Oxygen and Nitrogen Species at Interface
of Helium Plasma Jet and Agar. IEEE Trans. Plasma Sci. 2016, 44, 3112–3116. [CrossRef]
4. Kim, G.C.; Han, M.S.; Nam, S.H. Evaluation of tooth staining with red wine after tooth bleaching with nonthermal atmospheric
pressure plasma. Biomed. Res. 2018, 29, 396–400.
5. Bernhardt, T.; Semmler, M.L.; Schäfer, M.; Bekeschus, S.; Emmert, S.; Boeckmann, L. Plasma Medicine: Applications of Cold
Atmospheric Pressure Plasma in Dermatology. Oxidative Med. Cell. Longev. 2019, 4, 1–10. [CrossRef] [PubMed]
6. 1–40 kV 20–70 kHz 10–300 W Adjustable Power Supply. 1–40 kV 20–70 kHz 10–300 W Adjustable Power Supply (<50 pfd
loads)—Information Unlimited. Available online: amazing1.com (accessed on 4 August 2021).
7. Fuentes, A.F.; Eguiluz, R.P.; Callejas, R.L.; Cabrera, A.M.; Alvarado, R.V.; Delgado, S.B.; de la Piedad-Beneitez, A. Electrical Model
of an Atmospheric Pressure Dielectric Barrier Discharge Cell. IEEE Trans. Plasma Sci. 2009, 37, 128–134. [CrossRef]
8. Judée, F.; Vaquero, J.; Guégan, S.; Fouassier, L.; Dufour, T. Atmospheric pressure plasma jets applied to cancerology: Correlating
electrical configurations with in vivo toxicity and therapeutic efficiency. J. Phys. D Appl. Phys. 2019, 52, 1–16. [CrossRef]
9. Wang, R.; Yang, Y.; Chen, S.; Jiang, H.; Martin, P. Martin. Power Calculation of Pulse Power-Driven DBD Plasma. IEEE Trans.
Plasma Sci. 2021, 49, 2210–2216. [CrossRef]
10. Florez, D.; Diez, R.; Piquet, H. DCM-Operated Series-Resonant Inverter for the Supply of DBD Excimer Lamps. IEEE Trans. Ind.
Appl. 2014, 50, 86–93. [CrossRef]
11. Joh, H.M.; Kang, H.R.; Chung, T.H.; Kim, S.J. Electrical and Optical Characterization of Atmospheric-Pressure Helium Plasma
Jets Generated With a Pin Electrode: Effects of the Electrode Material, Ground Ring Electrode, and Nozzle Shape. IEEE Trans.
Plasma Sci. 2014, 42, 3656–3667. [CrossRef]
12. Bonnin, X.; Brandelero, J.; Videau, N.; Piquet, H.; Meynard, T. A High Voltage High Frequency Resonant Inverter for Supplying
DBD Devices with Short Discharge Current Pulses. IEEE Trans. Plasma Sci. 2014, 29, 4261–4269. [CrossRef]
13. Nong, C.Y.; Ming, K.C. Design of Plasma Generator Driven by High-frequency High-voltage Power Supply. J. Appl. Res. Technol.
2013, 11, 225–234. [CrossRef]
14. Diaz, F.J.; Azcondo, F.J.; Casanueva, R.; Brañas, C. Power Source Electronic Ballast for Electrodeless Fluorescent Lamps. Electronics
2020, 9, 34. [CrossRef]
Electronics 2021, 10, 2119 21 of 21
15. Cheng, C.A.; Cheng, H.L.; Ku, C.W.; Chung, T.Y. Design and Implementation of a Single-Stage Acoustic-Resonance-Free HID
Lamp Ballast with PFC. IEEE Trans. Power Electron. 2014, 29, 1966–1976. [CrossRef]
16. Rabkowski, J.; Lasica, A.; Zdanowski, M.; Wrona, G.; Starzynski, J. Portable DC Supply Based on SiC Power Devices for
High-Voltage Marx Generator. Electronics 2021, 10, 313. [CrossRef]
17. Taherian, M.; Allahbakhshi, M.; Farjah, E.; Givi, H. An Efficient Structure of Marx Generator Using Buck–Boost Converter. IEEE
Trans. Plasma Sci. 2018, 46, 117–126. [CrossRef]
18. Nara, K.; Channegowda, J. Single Stage PFC Flyback AC-DC Converter Design. In Proceedings of the IEEE International
Conference on Electronics, Computing and Communication Technologies (CONECCT), Bangalore, India, 2–4 July 2020;
Volume 46, pp. 1–6.
19. Das, S.; Sudhakar, P. Power Factor Correction & Harmonic Distortion Control for AC-DC Fault Tolerant Power Distribution
System & Power Quality Qualification Test Results as per IEC 61000-3-2. BARC Newsl. 2017, 356, 5–13.
20. Liu, C.T.; Kumakura, T.; Ishikawa, K.; Hashizume, H.; Takeda, K.; Ito, M.; Hori, M.; Wu, J.S. Effects of Assisted Magnetic Field to
An Atmospheric-Pressure Plasma Jet on Radical Generation at The Plasma-Surface Interface and Bactericidal Function. Plasma
Sources Sci. Technol. 2016, 25, 065005. [CrossRef]
21. Kostov, K.G.; Nishime, T.M.C.; Castro, A.H.R.; Toth, A.; Hein, L.R.O. Surface Modification of Polymeric Materials by Cold
Atmospheric Plasma Jet. Appl. Surf. Sci. 2014, 314, 367–375. [CrossRef]
22. Witte, M.B.; Kiyama, T.; Barbul, A. Nitric Oxide Enhances Experimental Wound Healing in Diabetes. Br. J. Surg. 2002, 89,
1594–1601. [CrossRef]
23. Kwon, J.; Kim, R.-Y. High Power Density, High-Voltage Parallel Resonant Converter Using Parasitic Capacitance on the Secondary
Side of a Transformer. Electronics 2021, 10, 1736. [CrossRef]