RESEARCH

PROJECT
ERGODICITY IN REVERBERANT
CHAMBERS
1 • Renato Alvarez
• Juan Valverde
Tutor: Prof. Nathalie Raveu

January, 2019
 2
REVERBERANT CHAMBERS

 Uniformity of the EM field inside the chamber

 Electromagnetic compatibility tests

 Inmunity tests
 Emission measurements
 Shielding effectiveness
Image Source: https://www.researchgate.net  Material Characterization

ERGODICITY!
3
OUTLINE
 Ergodicic Hypothesis
 Reverberant Chambers Theory
 Statistical Model of Electromagnetic Field
 Ergodicity Measurement Experiment
 Conclusions and Perspectives
 ERGODIC HYPOTHESIS
4

Spatial Average = Time [1]

Average
= ⟨ 𝑿 (𝒕) ⟩
=𝟏
⟨ 𝑿 ( 𝑵 )⟩
• Example: Measurement of Capacitance

1 Day production Part of the ensemble each day

D1 D2 DN
B

= …….

https://www.online-stopwatch.com/spanish/countdown-eggtimer.php

𝑪𝟏 + 𝑪𝟐 +… 𝑪 𝑴 𝑪 𝑫 𝟏 + 𝑪 𝑫 𝟐 +… 𝑪 𝑫𝑵
=
𝑴 (¿ 𝒐𝒇 𝒔𝒂𝒎𝒑𝒍𝒆𝒔 ) 𝑵 (¿ 𝒐𝒇 𝒅𝒂𝒚𝒔 )

[1] KEANE, M.. (2010). Understanding Ergodicity. Weleyan University. pp 3-8

5
OUTLINE
 Ergodic Hypothesis
 Reverberant Chambers Theory
 Statistical Model of Electromagnetic Field
 Ergodicity Measurement Experiment
 Conclusions and Perspectives
 6
ERGODICITY IN REVERBERATION
CHAMBERS
h

w
STIRRE [1]

RS
l
https://www.ditutor.com
• c – speed of light
• m
• n coefficient
modes
• p
• Resonant frequency
• Electrically Big chambers
Image Source: https://www.researchgate.net • Stirrers introduces different boundary
conditions
• Standing Wave pattern[2]
[1] HILL, D. (2009). Electromagnetic Fields in Cavities. USA. Wiley and Sons. pp 44-46
[2] THIAN, Z, (2017) Efficient Measurement Techniques in Reverberation Chamber, Liverpool, Univeristy of Liverpool, pp
 7
REVERBERATION CHAMBERS (RC)
• Weyl’s formula, cumulative number of modes depending on frequency (f) [2]:

[1][2]

• Density of modes (number of modes per frequency interval)[2]

[1][2]

• Many modes Deterministic analysis is impractical and difficult. [1][3]

• EM fields varies randomly with Stirrer positions then statistical approach is the
solution. [1]

• Many modes EM field becomes statistically uniform and isotropic. [2] [4]
[1] HILL, D. (2009). Electromagnetic Fields in Cavities. USA. Wiley and Sons. pp 30-31
[2] THIAN, Z, (2017) Efficient Measurement Techniques in Reverberation Chamber, Liverpool, Univeristy of Liverpool, pp 5-24.
[3] KOSTAS, J – BOVERIE, B. (1991).Statistical Model for a Mode-Stirred Chamber. IEEE. pp. 2
[4] TSIGROS, C. (2014) Efficiency analysis of innovative tuning methods for immunity testing in reverberation chamber and comparison to anechoic
room, Leuven university. pp. 59-60.
 REVERBERANT CHAMBERS
8

One mode – no uniformity of Many Modes– statistic

Electrical Field uniformity of Electrical Field

[1]

IMPORTANT: According to IEC61000-4-24

• Lowest Usable frequency (first frequency which has at least 100 cumulative
modes and a density of[2]1.5 modes/MHz)
[1] TSIGROS, C. (2014) Efficiency analysis of innovative tuning methods for immunity testing in reverberation chamber and comparison to anechoic room, Leuven university. pp. 36-
37
9
OUTLINE
 Ergodic Hypothesis
 Reverberant Chambers Theory
 Statistical Model of Electromagnetic Field
 Ergodicity Measurement Experiment
 Conclusions and Perspectives
 STATISTICAL MODEL
10  Based on plane-wave integration model and its statistic properties

[1] [2][3]

Where Ω is the solid angle, k the wavenumber vector, F is angular spectrum

vector
MEAN VALUE

 Applying the statistical properties of the angular spectrum [1][2]

MEAN SQUARE VALUE OF E

FIELD, INDEPENDENT OF
[1] [2][3] POSITION

[1] HILL, D. (2009). Electromagnetic Fields in Cavities. USA. Wiley and Sons. pp 94-101
[2] TSIGROS, C. (2014) Efficiency analysis of innovative tuning methods for immunity testing in reverberation chamber and comparison to anechoic room, Leuven university. pp. 206-211
[3] THIAN, Z, (2017) Efficient Measurement Techniques in Reverberation Chamber, Liverpool, Univeristy of Liverpool, pp 36-41
11
STATISTICAL MODEL
 − ∞¿ ∞
If the mean and variance are specified for a PDF over the , then we get Gaussian PDF.

Same the real and imaginary part of Ey

and Ez

For𝐄 𝐱 𝐄𝐲 , 𝐄𝐳 and and for Total Electric Field E

 Chi – is the statistics that fits the best for many independent random variables.

• m independent Gaussian random variables a1, a2, am

PDF • zero mean
• standard deviation σ
• m number of degrees of freedom (dof)
• Γ is the Gamma function, Γ(m)=(m-1)!

|𝑬 𝒙|=√ 𝑬 𝟐 𝟐 PDF

𝒙𝒓 +𝑬 𝒙𝒊
Same for Ey and Ez

PDF
Total Electric
Field
 VALIDATION OF STATISTICAL MODEL
12
 Liverpool University, f=2 GHz, with 120 samples
 Two stirrers are stepwise rotated at the same time

[1]

 Electric Field magnitude Ex has a chi square or Rayleigh Distribution confirming

the statistical properties of EM fields in Reverberant Chambers.
[1] THIAN, Z, (2017) Efficient Measurement Techniques in Reverberation Chamber, Liverpool, University of Liverpool, pp 39
13
OUTLINE
 Ergodic Hypothesis
 Reverberant Chambers Theory
 Statistical Model of Electromagnetic Field
 Ergodicity Measurement Experiment
 Conclusions and Perspectives
 MEASUREMENT SET-UP
14
APPLICATION EXAMPLE
 Leuven University.
 Cubic chamber
 LUF=276 MHz.
 2 moving antennas each
 Antenna A (83 rail steps)
 Antenna B (67 rail steps)
 150 total steps
 The two antennas never work
at the same time
 22 test frequencies from 800 a
2500 MHz
 158400 measurements (79200
for space and 79200 for time).
 1 probe or sensor measuring Ex,
l=2.48 m.
Ey, Ez.
[1] TSIGROS, C. (2014) Efficiency analysis of innovative tuning methods for immunity testing in reverberation chamber and comparison to anechoic room, Leuven University. pp. 213-
222
 15 AVERAGE OVER
 SPACE
Fixed tunner position and moving probe over space

Tuner steps Measurements

1 <ExS1> <EyS1> <EzS1> <ErmsS1>
2 <ExS2> <EyS2> <EzS2> <ErmsS2>
…. …. …. …. ….
…. …. …. …. …
150 <ExS150 <EyS150 <EzS150 <ErmsS150
> > > >
𝐸 𝑥 + …+ 𝐸 𝑥
⟨ 𝐸 𝑥 𝑆𝑛 ⟩ = 𝑝1

8
𝑝8
(8 probe locations)

√ 2
⟨ 𝐸 𝑟𝑚𝑠 ⟩ = ⟨ 𝐸 𝑥𝑆𝑛 ⟩ +⟨ 𝐸 𝑦 𝑆𝑛 ⟩ + ⟨ 𝐸 𝑧𝑆𝑛 ⟩
𝑺𝒏
2 2

𝐸 𝑥 … .. 𝐸𝑥
𝑝1 𝑝8
- Electric field in “x” axis, positions p1 to p8
• n – Number of tuner position
[1] TSIGROS, C. (2014) Efficiency analysis of innovative tuning methods for immunity testing in reverberation chamber and comparison to anechoic room, Leuven university. pp. 213-
 16
AVERAGE OVER TIME
 Fixed probe in one position and tuner steps variation
Probe Measurements
locations
P1 <Ext1> <Eyt1> <Ezt1> <Ermst1>
P2 <Ext2> <Eyt2> <Ezt2> <Ermst2>
…. …. …. …. ….
…. …. …. …. …
P8 <Ext8> <Eyt8> <Ezt8> <Ermst8>

𝐸 𝑥 + 𝐸 𝑥 +… + 𝐸 𝑥
⟨ 𝐸 𝑥 𝑡𝑚 ⟩ = 𝑡𝑝1 𝑡𝑝 2 (150 tuner steps)
𝑡𝑝1 50

150

⟨ 𝐸 𝑟𝑚𝑠 ⟩ =√ ⟨ 𝐸 𝑥𝑡𝑚 ⟩ + ⟨ 𝐸 𝑦𝑡𝑚 ⟩ + ⟨ 𝐸 𝑧𝑡𝑚 ⟩

2 2 2
𝒕𝒎

𝐸 𝑥 ….. 𝐸 𝑥
𝑡𝑝 1
- Electric field in “x” axis, for each tunner positi
𝑡𝑝 8

• m – Number of point in space

[1] TSIGROS, C. (2014) Efficiency analysis of innovative tuning methods for immunity testing in reverberation chamber and comparison to anechoic room, Leuven university. pp. 213-
 ERGODICITY MEASUREMENT
17
RESULTS FOR Erms
⟨ 𝑿 (𝒕) ⟩
=𝟏
4 ⟨ 𝑿 ( 𝑵 )⟩
f=800
f=1634MHz
f=2514 MHz
MHz
2

Ergodicity
1

Time step

• Ergodicity approaches to 1 while frequency increase

[1] TSIGROS, C. (2014) Efficiency analysis of innovative tuning methods for immunity testing in reverberation chamber and comparison to anechoic room, Leuven university. pp. 213-
 Ergodicity measurement results
18 f=800 MHz

Ergodicity

f=2514 MHz

Ergodicity
Time step

f=1634 MHz
2
Ergodicity

Time step

Time step
 ANALYSIS
19

• Higher the frequency

the process becomes
Ergodic.
X=2400
• 2 points of interest in Y=1.03
1824 MHz and 2400
MHz in which this rule is
not completely true.

MAX

MIN

Authors: Renato Alvarez and Juan Valverde

 N7 REVERBERANT CHAMBER
20
LUF
Chamber Dimensions: l=3.2 m, w=2.2 m, h=2.4 m

IEC
Standart

IEC
Standart

Authors: Renato Alvarez and Juan Valverde Authors: Renato Alvarez and Juan Valverde

• The lowest Usable frequency is 312 MHz.

• Any measurement should be for frequencies higher than 312 MHz to ensure EM field
uniformity and hence Ergodicity.
 N7 REVERBERANT CHAMBER VS. EXPERIMENT
21 CHAMBER
 N7 Chamber Dimensions:
 l=3.2 m, w=2.2 m, h=2.4 m

 Cubic Chamber: l=2.48 m.

Authors: Renato Alvarez and Juan Valverde

 N7 Reverberant Chamber has slightly more cumulative modes than the experiment one.
 Number of modes are proportional to the dimensions of the chamber.
22
OUTLINE
 Ergodic Hypothesis
 Reverberant Chambers Theory
 Statistical Model of Electromagnetic Field
 Ergodicity Measurement Experiment
 Conclusions and Perspectives
23
CONCLUSIONS
 Statistical analysis is the solution when measuring Ergodicity.

 Analysis methods must be confident with Maxwell’s equations and statistical properties.

 Frequency modes implies uniformity of the Electric Field, which means the system
becomes Ergodic.

 All measurements must be done for frequencies equal or higher than the LUF.
24
PERSPECTIVES
 Few studies about the measurement of Ergodicity in Reverberant Chambers, in fact is a
field that can be more developed.
 A lot of ways to change boundary conditions can be done in order to get many
frequency modes:
 Mechanical Stirrers
 Frequency Stirrers or Electronic Stirrers
 Source Stirring
 Changing the dimensions of the chamber.
 STATISTICAL MODEL
25  E field is defined by 3 rectangular components.

 Based on the statistical property of:

AVERAGE VALUE OF
EACH COMPONENT
REAL AND IMAGINAY
ARE “0” - MEAN

 From the isotropy property, the square mean of these components is derived by:

+ VARIANCE