Download Trade-off between EMI Separator and D. Sakulhirirak , V. Tarateeraseth

Trade-off between EMI Separator and
RF Current Probe for Conducted EMI Testing
D. Sakulhirirak1, V. Tarateeraseth2, W. Khan-ngern1, and N. Yoothanom3
1
Research Center for Communications and Information Technology (ReCCIT),
Faculty of Engineering, King Mongkut’s Institute of Technology Ladkrabang (KMITL), Bangkok, Thailand.
2
Srinakharinwirot University, Faculty of Engineering, Ongkharak, Thailand.
3
Sripatum University, Faculty of Engineering, Bangkok, Thailand.
E-mail: [email protected], [email protected], [email protected], [email protected]
Abstract-This paper presents the trade-off between EMI
separator and the RF current probe for conducted EMI
measurement. EMI toolkit is used as a noise source for
investigating effects from differential-mode (DM) and commonmode (CM). Paul-Hardin noise separation network or EMI
separator is used as a benchmark to compare with the RF current
probe. The conducted electromagnetic interference between RF
current probe and EMI separator is compared by the experiment.
Finally, the trade-off between two measurement methods is
verified by both theoretical and experimental results.
I.
ports of LISN when measuring by RF current probe but when
using Paul-Hardin network these terminators are input
impedance of the network. VP and VN stand for phase and
neutral voltage signal, respectively. CS is a parasitic
capacitance caused by MOSFET (Q ) between drain and frame
ground via its heat-sink.
INTRODUCTION
Electromagnetic Compatibility (EMC) is an electronic
system which is able to function compatibly with other
electronic systems and not produce or be susceptible to
interference with its environment [1]. The interference can be
divided widely into two groups: conducted and radiated
interferences. In this paper, the conducted emission is
emphasized and mentioned to separate noise components
(Differential-Mode: DM and Common-Mode: CM) for
electromagnetic environment. It is very useful to know noise
components in every interested circuit [2]. Generally, there are
two methods, often used for separating noise components. The
first method is a classical measurement using radio frequency
current probe (RF current probe). The second method is
measurement CM and DM by using separating devices such as
CM/DM discrimination network [3], Paul-Hardin noise
separation network [4] and noise separator [5] etc. However,
all separation networks in this paper are called as the EMI
separator.
The Equipment Under Test (EUT) in this paper is EMI
toolkit, used for conducted EMI studying in terms of theory
and practice. Fig. 1 shows outside of EMI toolkit. Boost
converter is a main part which generates EMI. In addition, a lot
of study functions are combined in the toolkit and it can be
selected functional operation such as self-resonant frequency
(SRF) of the passive components, reverse recovery time of
diodes, switching frequencies, and gate drive control [6].
To understand noise behaviors in a boost converter, Fig. 1
can be rewritten to Fig. 2 which includes Line Impedance
Stabilization Network (LISN) schematic and two 50
terminators. These terminators are connected to the RF output
Figure 1. Top view of EMI toolkit configuration.
Q
Figure 2. Schematic of LISN and boost converter under consideration.
II. CONDUCTED EMI MEASUREMENTS: RF CURRENT PROBE
AND EMI SEPARATOR CONCEPTS
Conducted EMI emission is measured using a LISN as a
50 impedance. The DM noise current ( I DM ) flows out from
line and returns via neutral while the CM noise current ( I CM )
flows out from line and neutral and returns via ground wire as
shown in Fig. 3 Eqns. (1)-(3) are line, neutral and ground
current and Eqns. (4)-(5) are line and neutral voltages,
respectively [7].
ECTI-CON 2007
The 2007 ECTI International Conference
___________________________________________________________
101
transformers (1:1 ratio) and single-pole-double-throw (SPDT)
switch, which operate simultaneously [7].
The line-ground voltage %VLG & and neutral-ground voltage
%VNG & are connected to EMI separator, using Eqns. (8) and (9)
for voltage across switch at A and B position respectively.
VCM ! VDM
VCM " VDM
Figure 3. DM and CM currents from LISN.
I Line
I CM ! I DM
(1)
I Neutral
I CM " I DM
(2)
I Ground
2 # I CM
(3)
VLine
VNeutral
50 # ( ICM ! I DM )
50 # ( I CM " I DM )
(4)
(5)
VCM ! VDM
'VCM
VDM
A. RF Current Probe
RF current probe is a clamp-on RF current transformer
designed for use with EMI Test Receivers/Spectrum Analyzers,
2VCM or 2VDM
or with any similar instrument having a 50 input impedance,
Figure 5. Paul-Hardin noise separation network schematic.
to determine the intensity of RF current present in an electrical
conductor or group of conductors [7]. Fig. 4 shows the RF
(8)
VLG " VNG VCM ! VDM " VCM " VDM
2VDM
current probe within bandwidth 10 kHz - 250 MHz. The
maximum primary current from DC - 400 Hz is up to 100 A
(9)
VLG ! VNG VCM ! VDM ! VCM " VDM
2VCM
and the transfer impedance ( ZT ) is 5 [8]. The performance
of the RF current probe may be expressed in terms of sensor
or 2VDM
(10)
VLG " VNG 50 # 2 I DM
transfer impedance: Eqn. (6). Where Vout is the voltage
developed across a 50 termination on the output and I cond is
(11)
2VCM VLG ! VNG 50 # 2 I CM
the current flowing through the conductor being measured. The
probe transfer impedance is often expressed in terms of dB
The output is VLG " VNG for 2VDM defined by Eqn. (8) and
which can be calculated the measured current from Eqn. (7) [9]. VLG ! VNG for 2VCM defined by Eqn. (9) and lastly, leakage
between the DM and CM at the output should be small because
noise measurement between the DM and CM must be
guaranteed small interference [5]. The impedance of two inputs
are 50 within 20 percent tolerance according to CISPR 16-1
standard because both of them have to connected with LISNs.
The CM output impedance of network is very closed to 50
(a) Top view
(b) Side view
(in range 45
– 50
at 150 kHz – 30MHz) but for DM
Figure 4. Toroidal RF current probe.
output impedance is nearly 20% tolerance from
50 (35 – 50 ) as shown in Fig. 7. The performance of
Vout
(6) EMI separator can be described in term of rejection attenuation
ZT
I cond
value of both modes; some papers call “CM/DM rejection
(7) ratio” [2], [10] as shown in Fig. 8.
I cond ( dB $ A) Vout (dB $V ) " Z ( dB)
B. EMI separator
Many papers have been discussed and proposed EMI
separators [3-5], [10-11]. They can be separated roughly in two
groups based on magnitude of output signal, single and double
output noise. Moreover, in group of double output [4] has no
guarantee that it can be used representative of RF current probe.
Fig. 5 shows the Paul-Hardin noise separation network
schematic. There are two important elements; two wideband
(a) Outside
(b) Inside
Figure 6. Paul-Hardin noise separation network.
ECTI-CON 2007
The 2007 ECTI International Conference
___________________________________________________________
102
LISNs
EUT
Supply
Figure 7. CM and DM output impedance of EMI separator.
Figure 10. EMI separator measurement method.
IV. EXPERIMENTAL RESULTS
Figure 8. CM and DM rejection.
III. MEASUREMENT METHODS
A. RF Current Probe Measurement
Figs. 9 (a)-(b) show basically method to measure DM and
CM noise, respectively. However, the measured results of DM
and CM are double ( 2 I DM and 2 I CM ) caused by sum of current
vectors in same direction. The current probe is usually clamped
between the EUT and the LISN as near as possible. In some
EMC standard current probes, such as the DEF STAN 59-41
DCE01, have to away 5 cm from the LISN connection on the
power lead [7].
In this section, noise floor between the RF current probe and
EMI separator are measured first for using as reference levels.
There are three noise floors: RF current probe, CM EMI
separation and DM EMI separation which are the same level of
noise floors about 25 dB.
Measured results of 2VCM and 2VDM are divided into three
sections by comparison results between RF current probe and
EMI separator. Firstly, noise source is measured without filter
components this is a full noise condition as shown in Figs. 1112. Secondly, adding two X-capacitors between line and
neutral, 0.47 !F, as shown in Figs. 13-14, which show DM and
CM comparing between RF current probe and EMI separator
with and without filter components. A dash line prefers to EMI
measured result when filter components are added in the
circuit. Finally, two Y-capacitors are added across line-ground
and neutral-ground, 0.2 !F, as shown in Figs. 15-16.
Figure 11. The comparison of 2VDM between RF current probe and
EMI separator.
(a) Measured 2DM noise
(b) Measured 2CM noise
Figure 9. RF Current probe measurement method.
B. EMI separator Measurement
The second method to measure DM and CM noise
components are displayed in Fig. 10. LISN-1 couples voltage
from line-ground while LISN-2 couples voltage from neutralground. Output of separator is connected with 50 terminal of
spectrum analyzer. In addition, coaxial cables (50 ) with
BNC connector are connected with two RF output terminals of
LISNs for coupling line-ground and neutral-ground signal
simultaneously.
Figure 12. The comparison of 2V
between RF current probe and
CM
EMI separator.
ECTI-CON 2007
The 2007 ECTI International Conference
___________________________________________________________
103
Where Vout ( CM ) and Vout ( DM ) are common
Figure 13. The comparison of 2VDM when adding Cx filter between
RF current probe and EMI separator.
mode and
differential mode output voltage measured by RF current probe
respectively.
Then, multiplication Eqn. (12) by two to compare with Eqns.
(10)-(11) and inverting them to dB values using logarithm
function.
The compared results show the different value between
using RF current probe and EMI separator about 13.98 dB.
The experimental results as shown in Figs. 11-12, can be
realized by the theoretical expression as mention. Fig. 13
shows the performance of X capacitor filters. The DM EMI is
decreased about 20 dB while the CM EMI has a little changed
as shown in Fig. 14 and the difference between RF current
probe and EMI separator still about 14 dB. Figs. 15-16 show
EMI results which are mitigated by Y capacitors. The
performance of Cy can suppress noise about 5 dB for common
mode and nearly 0 dB for differential mode
VI. CONCLUSIONS
Figure 14. The comparison of 2VCM when adding Cx filter between
RF current probe and EMI separator.
The separation of conducted EMI measurement, measured
by RF current probe and EMI separator, is compared. The
attenuation value between RF current probe and EMI separator
is about 14 dB, realized by the theoretical and experimental
results. The performance of EMI separator has been realized
using the convenient EUT as EMI toolkit. However, it should
be noted that the EMI separator, which is low cost, can be used
at low-medium power level because the saturation of wideband
transformer while the RF current probe can be covered the high
level.
REFERENCES
[1]
Figure 15 The comparison of 2V when adding Cy filter results between
DM
RF current probe and EMI separator.
Figure 16 The comparison of 2V when adding Cy filter results between
CM
RF current probe and EMI separator.
V. ANALYSIS
From Eqn. (6), the transfer impedance of the current probe
and the measured current I cond is represented
by 2 I DM or 2 I CM as follows;
( ZT ) is 5
Vout (CM )
(5) # (2 I CM ) and Vout ( DM )
(5) # (2 I DM )
(12)
Clayton R. Paul, Introduction to Electromagnetic Compatibility. WileyInterscience, A John Wiley & Sons, Inc., publication, 1992.
[2] Ting Guo, Dan Y. Chen and Fred C. Lee “Separation of the CommonMode-and Differential-Mode-Conducted EMI Noise”, IEEE Trans.
Power Electron, vol. 11, no. 3, pp. 480-488, 1996.
[3] See Kye Yak “Network for EMI”, Electronic Letter, vol. 35, no. 17, pp.
1446-1447, 19th Aug 1999.
[4] Clayton R. Paul and Keith B. Hardin “Diagnosis and Reduction of
Conducted Noise Emissions”, IEEE Trans. on Electromagnetic
Compatibility, vol. 30, no. 4, pp. 553-560, Nov 1988.
[5] Shou Wang, Fred. C Lee and Willem Gerhardus Odendaal.
“Characterization Evaluation, and Design of Noise Separator for
Conducted EMI Noise Diagnosis”, IEEE Trans. on Power Electronics,
vol. 20, no. 4, pp. 974–982, July 2005.
[6] Werachet Khan-ngern, Vuttipon Tarateeraseth. “Self-learning EMC
Toolkit for Electronic and Electrical Engineers”, ICEMC Conf. on EMC
on Education, July 2005, session 3C-1.
[7] V. Prasad Kodali, Engineering Electromagnetic Compatibility. The
Institute of Electrical and Electronics Engineers, Inc., New York, 2001.
[8] Fischer Custom Communications, Inc. (2004). Instrumentation:
Available:http://www.fischercc.com/instframe.html [Online].
[9] David Morgan, A handbook for EMC testing and measurement. WileyInterscience, Short Run Press Ltd., Exeter, 1994.
[10] Marco Chiad" Caponet and Francesco Profumo “Devices for the
Separation of the Common and Differential Mode Noise: Design and
Realization”, Applied Power Electronics Conference and Exposition
(APEC) Seventeenth Annual IEEE, vol. 1, pp. 100-105, March 2002.
[11] Mark J. Nave “A Novel Differential Mode Rejection Network for
Conducted Emissions Diagnostics”, IEEE National Symposium on
Electromagnetic Compatibility, pp. 223–227, May 1989.
ECTI-CON 2007
The 2007 ECTI International Conference
___________________________________________________________
104
II. EXPERIMENTAL
Ceramic sample of Gd 1:2:3 were prepared by conventional
powder processing from high-purity oxides and carbonates,
calcining at 930oC, the powder were pressed into pellets with
a pressure of 1 ton/cm2 .The pellets were sintered at 900oC,
920oC, 925oC, 930oC, 935oC, 940oC, 945oC, 950oC, 955oC,
960oC, 965oC and 970oC for 10 hours.
We had experimentally investigated the following effect of
external magnetic field on current-voltage characteristics and
magnitude of negative resistances. The current-voltage
characteristics were measured by the four probe technique
with indium electrodes at 77 K. The magnetic field is applied
to samples perpendicular with the direction of current flow.
III. EXPERIMENTAL RESULTS
2.5
0.8
!V
IC( A)
B=0.35 mT
7
B=0.2 mT
6
B=0 T
5
4
3
0.6
0.4
1
B=0.6 mT
8
0.7
0.5
1.5
9
2
V(mV)
Ic
2
B. The effect of magnetic field on magnitude of negative
resistance
The influences of BEXT on the magnitude of differential
voltage (!V) are studied. Sample used here show the various
values of critical currents which are 1.55A, 1.3A, 1.05A,
0.78A, 0.52A, 0.39A and 0.12A respectively. Current-voltage
relations and the dependences of !V on BEXT are shown in
Fig.4. It’s found that the magnitude of !V depends on BEXT
[4]. For example, for the sample with IC = 1.05 A, when we
applied BEXT from 0 to 0.2 mT, !V was increased. However if
applied BEXT exceeded 0.2 mT, then !V was decreased
continuously. The maximum !V (!VMAX) was obtained at 1.4
mV.
V(mV)
A. The Optimum condition of critical current for the
occurrence of negative resistance
From Fig. 2 shown the relationship between sintering
temperature and critical current (IC). It’s found that, the
highest IC is 2.2 A at sintering temperature 930oC. At the
highest IC sample must be applied external magnetic field
(BEXT) higher than the low IC sample, to destroy the
superconducting state. However the negative resistance could
not be observed at this highest IC. But negative resistance
obviously occurred as IC is reduced from maximum point.
The maximum magnitude of difference voltage(!V) is
obtained at IC = 1.05 A. !V will be to zero while IC is
decreased to zero as shown in Fig. 2 and Fig. 3. Thus, it can
be seen that the negative resistance phenomena is apparently
found in the range of low IC.
1
0
0.3
0
1
2
3
4
5
6
0.2
0.5
I(A)
0.1
0
0
900 920 925 930 935 940 945 950 955 960 965 970
1.6
Sintering temp. ( OC )
1.4
VMAX
1.2
V vs sintering temperature
V(mV)
Figure 2. The plot of IC,
0.8
0.8
0.6
0.7
0.4
0.6
V(mV)
1
0.2
0.5
0
0.4
0
0.3
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
BEXT (mT)
0.2
0.1
Figure 4. The effect of BEXT on samples with various IC
1.05 A
0
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
2.2
Fig. 5 shows !V versus BEXT relations obtained from
sample with various critical current. It is found from Fig. 6
that the sample with IC = 1.05 A shows the highest value of
!VMAX comparing with other IC samples.
B
IC(A)
Figure 3. Dependence of
V on IC
ECTI-CON 2007
The 2007 ECTI International Conference
___________________________________________________________
106
1.6
V
Ic = 0.12 A
Ic = 0.39 A
Ic = 0.52 A
Ic = 0.78 A
Ic = 1.05 A
Ic = 1.30 A
Ic = 1.55 A
1.4
1.2
V(mV)
1
I
0.8
c) Sample with IC = 1.05 A
d) Sample with IC = 0.52 A
0.6
Figure 7. Illustration of a macrostructure model at various IC
0.4
0.2
From Fig. 8, when the applied current (I) is equal to (or less
than) its critical current(IC) value, both sides of the sample are
connected by the superconduction parts. Then, the resistance
of sample does not appear. But when I>IC, superconduction
part will be cut-off, because weak point region is destroyed.
The resistance appears in this condition. Because the volume
of upper destroyed parts are less than lower destroyed parts,
all current will flow over the upper part only. Then, the small
voltage drop will appear across the upper part. When the
current reaches IN, the upper superconduction part, where has
magnetic substance, is destroyed. Thus, overall cut-off region
of the upper part is more than lower part. The resistance of
upper part increases quickly. Then all current flows to lower
part, which has low resistance. Therefore, the voltage drop
across sample decreases immediately. This phenomenon is
called Negative-resistance. The magnitude of differential
voltage ( V ) is about 0.71 mV.
0
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
BEXT (mT)
Figure 5. The plot of
V vs BEXT at various IC
1.6
1.4
VMAX (mV)
1.2
1
0.8
0.6
0.4
0.2
0
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
3.5
IC (A)
3
2.5
VMAX and IC
V(mV)
Figure 6. The relationship between
IV. DISCUSSIONS
The experiment al results can be explained by the
macrostructure model of Gd-Ba-Cu-O. Since the high-IC
sample exhibits superconducting state more completely than
the low-IC sample, then the connection of weak point region
must be stronger than the low-IC sample as shown in Fig. 7.
a) Sample with IC = 2.2 A
2
V
VN
1.5
1
0.5
IC
IN
0
0
0.5
1
1.5
2
2.5
3
3.5
4
4.5
I(A)
b) Sample with IC = 1.55 A
Figure 8. Illustration of current-voltage characteristics of sample with
IC = 1.05 A
ECTI-CON 2007
The 2007 ECTI International Conference
___________________________________________________________
107
5
We will consider the effect of BEXT on V as follows. In the
case of the sample showing IC=1.55A, V was increased by
BEXT in the range of 0 ~ 0.23 mT, as shown in Fig. 5. This
phenomenon can be explained by considering the
macrostructure model as shown in Fig. 9 a). Since the cut-off
region which has magnetic substance, was broadened by BEXT,
V increases, that is, current path changes from the upper high
resistance path to the lower low resistance path. When the
external magnetic flux BEXT exceeds 0.23 mT, V was
decreased. Since BEXT also destroys partially the lower
superconduction part as shown in Fig. 9 a), the difference of
electrical resistance between the upper path and the lower path
became small. Then, V decreases. For sample IC = 1.05 A,
when BEXT = 0.2 mT, the highest V is obtained.
Consequently, the cut-off region, containing magnetic
substance, is broadened until magnetic substance is uncovered
completely.
In the case of sample showing IC = 0.52 A, when BEXT =
0.16 mT, !VMAX is obtained. Due to the cut-off region, which
has magnetic substance, is broadened until magnetic substance
is uncovered completely. Nevertheless, BEXT will also destroy
partially the lower part similarly as shown in Fig. 9c). Then,
!VMAX in this case less than the sample with IC = 1.05 A as
shown in Fig. 6.
V = VN
B
I = IN
BEXT = 0.16 mT
B
B
c) Sample with IC = 0.52 A
Figure 9. Illustration of a macrostructure model at various applied
magnetic field
V. CONCLUSIONS
From the investigation of Current and Voltage
characteristics in GdBa2Cu3O7-x superconducting ceramic
material, it's found that the negative resistance phenomenon is
apparently found in range of low critical current. Moreover,
the magnitude of differential voltage ( V) depends on external
magnetic field and critical current. The largest differential
voltage !VMAX is obtained by the employment of suitable
BEXT. Moreover, the occurrence of !VMAX in the high-IC
sample must be applied BEXT higher than the low-IC sample.
Some experimental results can be quite explained by the
macrostructure model of GdBa2Cu3O7-x superconducting
ceramic materials.
B
ACKNOWLEDGMENT
We are indebted to A. Keawcharoen
Suriyaamaranont for technical assistance
and,
C.
REFERENCES
a)
Sample with IC = 1.55 A
[1]
[2]
[3]
[4]
V = VN
W. Wongsuttitum, W. Titiroongruang, “Macrostructure Model of GdBa-Cu-O Superconducting Ceramic Materials,” Proc. Of Thailand’s 24th
Electrical Engineering Conference, (2001), pp. 992- 997
W. Titiroongruang, Y. Akiba, S. Supadech, T. Kurosu
and M. Iida, “Macrostructure Model of Y-Ba-Cu-O System,”
Proceeding of Faculty of Engineering, Tokai University. (1991) 17, pp.
7-13.
Y. Akiba, W. Titiroongruang, T. Kurosu, M. Iida and T. Nakamura,
“Electromagnetic Memory Effect in Superconductivity Y-Ba-Cu-O
System,” Oyo Butsuri 58 No. 1 (1989) 151. [in Japanese]
W. Wongsuttitum, R. Piyananjaratsri, W. Titiroongruang and M. Iida,
“The Effect of The External Magnetic Field on Negative Resistance
Phenomena of GdBa2Cu3O7-x Superconducting Ceramic Materials,”
KMITL SCIENCE JOURNAL, Vol. 6, No.1 , pp. 169-175, Jan-Apr 2006.
I = IN
BEXT = 0.6 mT
b) Sample with IC = 1.05 A
ECTI-CON 2007
The 2007 ECTI International Conference
___________________________________________________________
108