188 JOURNAL OF ELECTRONIC SCIENCE AND TECHNOLOGY OF CHINA, VOL. 7, NO.

2, JUNE 2009

Simulation and Experimental Method for Microwave Oven

Han-Ji Ju and Qing Zhao

Abstract⎯The simulation software, HFSS (high fre- In recent years, HFSS (high frequency structure
quency structure simulator), is introduced in microwave simulator), a computer simulation software, gains wide
oven design. In the cold test, a network analyzer is used applications in many domains. By effectively using HFSS,
to measure the reflection coefficient (S11) of the cavity general insights into the true nature of electric field in the
under empty and loaded states over the frequency range microwave oven can be achieved and used to guide the
from 2.448 GHz to 2.468 GHz. In the hot test, a piece of design of ovens. Until now this work has been carried
wet thermal paper and an infrared thermal imaging through only in a few corporations, such as LG Microwave
camera are used to measure the electric field distribu- and Electrical Appliances Manufacturing Company Limited
tions on the mica and turntable. In the cold test, the of Korea and Panasonic Corporation of Japan. We find one
simulation agrees well with the experiment no matter in picture of the Panasonic Company who studied the influ-
empty state or loaded state. In the hot test, the ence of steam to the electric field distribution. The research
simulation agrees well with the experiment in general in of LG mainly focus on the arcing in the magnetron[6] and
empty state and approximately in loaded state. The little the impedance matching[7] with CST Microwave Studio
difference in both cold and hot test may be due to that software. Another research by University of Bologna
the model in simulation is not absolutely identical with simulated the electric field distribution on load of different
that in experiment or the inadequate precision of temperatures in a rectangular cavity. But they did not give
infrared thermal imaging camera. the experimental result. In China, Guangdong Province has
launched this research as a Scientific and Technological
Index Terms⎯Far-infrared thermal imaging camera, Development Program in 2004.
microwave oven simulation, network analyzer, wet This paper aims at verifying the feasibility of HFSS
thermal paper. simulation method. The simulation is compared with the
experiment both in the cold and hot state. A challenging
1. Introduction goal for the future is to introduce this method to contrive a
best structure before designing the microwave oven and
Microwave oven has been used for heating/cooking solve some problems such as wave leakage and striking fire.
applications for more than 50 years. Researchers and
technicians have struggled long against the principles
involved in the interaction of microwaves with food. In
2. Simulation and Experimental Setup
1976, Osepchuk attributed the lack of microwave oven A 25 L microwave oven is used in the simulation and
theory to “the over-whelming complexity of the oven-food experiment. The configuration of the oven and coupling
configuration as an object for study involving Maxwell’s window is shown in Fig. 1. And the model of the magne-
equations”[1]. Since then, the microwave field distribution tron and waveguide used in simulation are shown in Fig. 2.
and heat transfer in microwave oven have been investigated The cavity dimensions (X, Y, Z)=(320 mm, 327 mm, 200
with numerical analysis[2]-[4] or experimental methods[1],[5]. mm), with some protrudes on the side wall and waveguide
However, the former can only analyze a regular rectangular of 84 mm wide, 93 mm high and 33 mm thick on the right
cavity which is different from the actual microwave oven
sold at the market, so it is not helpful in guiding the design. Cavity
The design primarily remains an empirical art.
Wave port
Manuscript received January 17, 2008, revised April 10, 2008. This
work was supported by the National Natural Science Foundation of China Coupling
under Grant No. 10775029, Vacuum Electronics National Laboratory window
Foundation under Grant No. NKLC001-063 and Postdoctoral Foundation
under Grant No. 20070411149.
H.-J. Ju and Q. Zhao are with School of Physical Electronics,
University of Electronic Science and Technology of China, Chengdu, Waveguide
610054, China. (e-mail: [email protected] and [email protected] ) Mica Glass tumtable
Color versions of one or more of the figures in this paper are available
online at http://www.xb.uestc.edu.cn/Default_je.aspx. Fig. 1. Schematic view of the microwave oven.
JU et al.: Simulation and Experimental Method for Microwave Oven 189

contour line of power and frequency in a Polar or Smith

Antenna
Antenna cap chart, then this drawing is called Rieke diagram. A typical
Rieke diagram of magnetron is shown in Fig. 3. From the
Wave port
Rieke diagram, we can see the change of output power and
frequency deviation for various load impedances.
Ceramic According to experience, the magnitude of S11 should
satisfy the following relationship roughly: 0.6 ≤| S 11 |≤ 0.9
Coupling in the empty state, and | S 11 |≤ 0.25 in the loaded state. Also,
window Waveguide
the phase of S11 is important in the loaded state, it is best
located at the first and forth quadrant. In the empty state,
there is no special requirement on the phase.
Fig. 2. The simulation model of magnetron and waveguide.
The simulation results are compared with the
Table 1: The material’s characteristics experiment both in the empty and loaded state, as shown in
Fig. 4 and in Fig. 5, respectively.
relative permittivity
Material dielectric loss tangent
(F/m)
Mica 3.54 0.00191 Δf=0 MHz To load
pps 3.2 0.0009 0.45 0.05
Polystyrene 2.6 0 600W −5MHz
Ceramic 10.2 0.00067 +5 MHz
Boron silicon glass 6.9 0.001 700W
0.10
Water 76.7 0.157 0.40
800W
Table 2: The metal’s characteristics
+10 MHz −10 MHz
Metal Conductivity (S/m) relative permeability 840W
Stainless steel 1500000 1 0.15
0.35
Copper 58000000 0.999991

wall. The cavity with a turntable is considered as an empty 0.30 0.20

cavity in engineering. 0.25
In order to observe the field distribution in the empty
Fig. 3. Rieke diagram of magnetron.
cavity, a piece of mica with radius of 157.5 mm is placed
25 mm over the turntable. In the loaded state, the load is a
bottle of water of 300 mm.
In cold test, an Agilent Technologies E50708 network
analyzer is used to measure the S11 over the frequency range
from 2.448 GHz to 2.468 GHz. The bandwidth is 20 MHz.
The frequency range is selected according to the resonant
frequency of magnetron. The central frequency is the
resonant frequency. It must be noted that the resonant
frequency of different manufacturers’ products differs from
each other. Also, the resonant frequency will change (a) (b)
slightly with different load in the oven. Fig. 4. The Rieke diagram of empty cavity state: (a) simulation
In hot test, a SAT HY-6700 series infrared thermal result and (b) experimental result.
imaging camera is used to get the temperature photographs
of the mica, turntable and water.
We all know that material’s characteristic has great
influence on results, no matter in simulation or experiment,
so the boundary conditions and relative permittivity of
materials should be selected accurately. Their parameters
are presented in Table.1 and Table. 2.

3. Cold Test
In cold test, the Rieke diagram is observed. The micro- (a) (b)
wave tubes’ load characters are generally stated by some Fig. 5. The Rieke diagram of loaded cavity state: (a) simulation
feature curves. When these feature curves are presented as result and (b) experimental result.
190 JOURNAL OF ELECTRONIC SCIENCE AND TECHNOLOGY OF CHINA, VOL. 7, NO. 2, JUNE 2009

Fig. 4 shows that in empty state, the plots behave in and Fig. 6 shows that the load markedly changes the field
almost the same manners except a little difference in phase. distribution. The field concentrates on the center of the
In simulation, the phase starts from about −67.5° to 22.5°. In turntable. This conforms to the fact that the electric field is
experiment, the phase starts from about −57° to 10°. The inclined to focus on high permittivity material. So the field
phase of cross-section in simulation is almost 15° ahead of on the surface of the material will be large.
experiment.
Fig. 5 shows that in loaded state, the Rieke diagram of
the oven changes significantly compared with the empty
state, either in simulation or experiment. The phase range
becomes narrow. This means the reflection coefficient S11
becomes insensitive to frequency when the oven is loaded.
The magnitude of S11 in simulation agrees well with
experiment. The phase expands from about 30° to 45° both
in simulation and experiment.
In the cold test, the simulation agrees well with the
(a) (b)
experiment no matter in empty state or loaded state.
Fig. 6. The field distribution on the wet paper: (a) simulation
result and (b) experimental result.
4. Hot Test
In the hot test, the field distributions are recorded with
two methods: wet thermal paper[6] and infrared thermal
imaging camera. The field distribution on wet thermal
paper and turntable are shown in Fig. 6 to Fig. 9. In
experiment, the paper and turntable do not rotate.
4.1 Empty Cavity
To record the field distributions in the oven, a piece
of mica is placed 25 mm over the turntable. The (a) (b)
microwave heats up the paper and the color of the paper
Fig. 7. The field distribution on turntable: (a) simulation result and
will change according to the different microwave density (b) experimental result.
distribution on the paper. That is to say, the color can
reflect the electric field intensities. In simulation, the
white means the strongest field and the blue represents
the weakest. Correspondingly, in the experiment, the
white still means the strongest field but the black
represents the weakest.
Fig. 6 shows the views of the field distribution on the
wet thermal paper both in simulation and experiment. The
heat distribution in simulation is highly similar to that in
experiment. Some hot spots occur on the paper, especially (a) (b)
at the edge of the paper, corresponding color is bright in Fig. 8. The field distribution on turntable with a glass of water: (a)
simulation and black in experiment. The center of the paper simulation result and (b) experimental result.
is also hot but less than the edge. Fig. 7 shows views of the
field distribution on the turntable both in simulation and
experiment. The hottest spots are all around the edge of the
turntable. The center of the turntable is cold both in
simulation and experiment. The simulation agrees with the
experiment approximately.
In general, the simulation agrees well with the
experiment in the empty state.
4.2 Loaded Cavity
(a) (b)
Fig. 8 shows the field distributions on the turntable with
the load of a glass of water; Fig. 9 shows the field distribu- Fig. 9. The field distribution on water face: (a) simulation result
tions on the top surface of water. The comparison of Fig. 8 and (b) experimental result.
JU et al.: Simulation and Experimental Method for Microwave Oven 191

Fig. 8 indicates that the field is the largest in the airflow effects on heat and mass transfer in a microwave
turntable center and then decreases radially. A hot spot oven,” Journal of Food Engineering, vol. 59, pp. 181-190,
occurs at the left upward position in simulation which is not Sep. 2003.
obvious in experiment. In Fig. 9, the simulation shows the [5] S. N. Kharkovsky and U. C. Hasar, “Measurement of mode
most intense filed is not in the center of water but a little patterns in a high-power microwave cavity,” IEEE Trans.
Instrumentation and Measurement, vol. 52, no. 6, pp. 1815-
offsets from the center. This observation is not obvious in
1819, 2003.
experiment. The field is uniform in axial direction basically
[6] X.-W. Yang, K.-J. Xie, and J.-L. Shen. “Solving the arcing
and decreases in radial direction, either in simulation or
problem of continuous-wave magnetron,” Materials Review,
experiment. vol. 21, no. 11A, pp. 162-164, 2007.
The little difference may be that the model in [7] C.-Y. Zheng, B.-F. Jia, and G.-Z. Zhang. “Research on
simulation is not absolutely identical with that in impedance matching and energy coupling in microwave
experiment. Another cause may be the inadequate precision oven,” Materials Review, vol. 21, no. 11A, pp. 214-216, 2007.
of infrared thermal imaging camera. So the little difference
is not obvious in experiment.
Han-Ji Ju was born in Qingdao, China, in
5. Conclusions 1982. He received the B.S. degree in electronic
information engineering from Qingdao
In order to find a reasonable structure and solve some University, Qingdao, Shandong, in 2005, the
problems such as wave leakage and striking fire before M.S. degree from University of Electronic
designing microwave ovens. The simulation method with Science and Technology of China (UESTC),
HFSS is introduced to the design of microwave oven Chengdu, Sichuan, in 2008. Now he is
structure. Two cases, cold test and hot test, are simulated in pursuing his Ph.D. degree with Institute of Electronics, Chinese
empty state and loaded state, respectively. The differences Academy of Science, Beijing. His research interests include
between simulation and experiment are discussed. The microwave appliance, high power vacuum microwave electron
presented results show that HFSS is a feasible tool for oven devices, magnetic sensor and ground penetrating radar.
structure design.
Qing Zhao was born in Yiwu, China, in 1964.
He received the B.S. degree in nuclear
References
science from Chengdu University of
[1] T. G.. Mihran, “Microwave oven mode tuning by slab die- Technology, Chengdu, Sichuan, China, in
lectric loads,” IEEE Trans. Microwave Theory and Techniques, 1988, the M.S. degree in nuclear phyics from
vol. 26, no. 6, pp. 381-387, 1978. the Sichuan University, Chengdu, in 1998,
[2] A. Sekkak, L. Pichon, and A. Razek, “3-D FEM magneto- and the Ph.D. degree in applied plasma
thermal analysis in microwave ovens,” IEEE Trans. physics from the Southwestern Institute of Physics, Chengdu, in
Magnetics, vol. 30, no. 5, pp. 3347-3350, 1994. 2001. He joined the High-Power Microwave (HPM) Source
[3] F. Torres and B. Jecko, “Complete FDTD analysis of micro- Group with the School of Physical Electronics, UESTC, Chengdu,
wave heating processes in frequency-dependent and temper- in 2001, where his current research activities include the work on
ature-dependent media,” IEEE Trans. Microwave Theory and high-power gyrotron oscillators, gyroklytron, electron gun, plasma
Techniques, vol. 45, no. 1, pp. 108-117, 1997. effects on HPM sources, low-temperature plasma application, the
[4] P. Verboven, A. K. Datta, and N. TrungAnh, “Computation of surface modification of material, and the plasma diagnose.