Professional Documents
Culture Documents
fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2019.2919384, IEEE Access
Date of publication xxxx 00, 0000, date of current version xxxx 00, 0000.
Digital Object Identifier 10.1109/ACCESS.2017.Doi Number
ABSTRACT This paper deals with the fast S-parameter modelling of multi-port lumped structures. The
developed model is based on the unfamiliar formalism using the tensorial analysis of networks (TAN). The
modelling methodology is described with general abstract topology and different application cases. The
methodology consists first in elaborating the equivalent graph topology of the considered problem. Then, it
is followed by the TAN mathematical abstraction including successively the branch and mesh space
analyses. The treated problem metric can be written with the tensorial Ohm’s law expressed in function of
the covariant voltage, contravariant current and the twice covariant impedance in the mesh spaces. The
equivalent Z-matrix of the considered multi-port structure is established from an innovative reduction
method of the mesh impedance. Then, the S-parameter model is extracted from the Z-to-S matrix transform.
The effectiveness of the established fast S-parameter TAN modelling is validated with three cases of proof-
of-concept constituted by TT-cell, TTT-cell two-port circuit and four port structure inspired from the 3D
coaxial shielded cable. As expected, an excellent agreement between the S-parameters calculated from
TAN model and simulated from the commercial tools from DC to some hundred’s megahertz is obtained. In
the future, the developed model is outstandingly beneficial for fast and accurate applications notably for the
conduced shielded cable electromagnetic compatibility analysis.
INDEX TERMS Circuit theory, cable shielding, S-parameter modelling, tensorial analysis of networks
(TAN)
2169-3536 (c) 2019 IEEE. Translations and content mining are permitted for academic research only. Personal use is also permitted, but republication/redistribution requires IEEE permission. See
http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2019.2919384, IEEE Access
As reported in [26-32], the TAN formalism is implemented • The contravariant parameter Ib represented by the
based on the graph topology followed by the problem metric branch current vector:
elaboration with tensorial approach. Fig. 2 indicates the I1
fundamental phases of TAN approach for solving electrical 2
I
problems. I b = , (7)
This analytical process must start with the topological
B
analysis by identifying the quintuplet integers: I
• The network number, Nw, • And the twice covariant metric represented by the
• The branch number, B, branch impedance matrix which is systematically a
• The node number, N, diagonal matrix for the lumped circuit with inter-
• The mesh number, M < B, branch coupling:
• And the port number, P < M. Z11 0 0
0 Z
Z bb = 22 . (8)
0
0 0 Z BB
The mesh space analysis is derived from these branch
space variables via the connectivity matrix Cbm which will
be described in the next paragraph.
2) MESH SPACE VARIABLES
The mesh space analysis consists in expressing:
• The mesh voltage is the covariant parameter Um
represented by the vector:
U m = U1 U 2
UM
, (9)
U m = Cm Vb U m = CmVb
b b
It is worth to point out that the voltage vector
component Um=0 in all meshes without external
sources.
t
• Knowing that Cbm = Cmb , the mesh current is
FIGURE 2. Main phases of the TAN method.
the contravariant parameter Jn represented by the
Before the analytical exploration, the main vector vector:
variables in branch and mesh spaces are recalled in the next J1
paragraphs. It is worth to recall that in the following section, 2
J n = J
the Ohm’s law V=ZI rewritten and implemented in tensorial
notation, must respect the tensor algebra rule: M
, (10)
J
• the voltage V (noticed with a subscript as Va) is a
covariant parameter which is represented as a row J n = Cbn I b J n = Cbn I b
vector,
• And the twice covariant metric represented by the
• the current I (noticed with a superscript as Ib) is a
mesh impedance matrix:
contravariant parameter which is represented as a
column vector, Z11 Z12 Z1M
Z
• and the associated impedance Z must be a twice Z = 21 Z 22 Z2M
covariant parameter which is denoted with two mn
. (11)
subscripts Zab which is represented as a square
M1
Z Z Z MM
M2
matrix.
1) BRANCH SPACE VARIABLES Z mn = Cm Z bb Cn Z mn = Cmb Z bbCnb
b b
The branch space analysis consists in expressing: The fundamental metric of the problem can be formulated by
• The covariant parameter Vb represented by the the compact Einstein relation:
branch source voltages vector: J n = Y nm U m J n = Y nmU m , (12)
Vb = V1 V2 VB , (6)
where:
2169-3536 (c) 2019 IEEE. Translations and content mining are permitted for academic research only. Personal use is also permitted, but republication/redistribution requires IEEE permission. See
http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2019.2919384, IEEE Access
Y nm = Z mn .
−1 equation (11). According to the circuit and system theory,
(13)
the structure under study can be assumed as a P-port
3) METHODOLOGY diagram. This type of system is usually represented by the
Fig. 3 describes the successive steps of the analytical impedance Z matrix or admittance:
processes of the proposed fast S-parameter TAN model. This
workflow indicates the overview of the overall routine Ymatrix = Zmatrix −1 , (15)
algorithm to be applied in the next two sections. which are assumed as p-size square matrices, linked by the
classical definition:
t
V1 I1 C11
C 1p J 1
= Z matrix = Z matrix .
V p I p C1p
C p J p
p
(16)
The equivalent impedance matrix relation is given by:
t
J1 V1
(
)
−1
2169-3536 (c) 2019 IEEE. Translations and content mining are permitted for academic research only. Personal use is also permitted, but republication/redistribution requires IEEE permission. See
http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2019.2919384, IEEE Access
( M ) : 0 = Z M 1 J 1 + Z M 2 J 2 + + Z Mp J p + Z M ( p +1) J p +1 + + Z MM J M
Z11 Z12 Z1 p Z1( p +1) Z1M
Z 21 Z 22 Z2 p Z 2( p +1) Z2M
Z mn = Z p1 Z p 2 Z pp Z p ( p +1) Z pM . (19)
Z ( p +1)1 Z ( p +1)2 Z ( p +1) p Z ( p +1)( p +1) Z ( p +1) M
Z M 1 ZM 2 Z Mp Z M ( p +1) Z MM
( ( p +1) ) Z ( p +1)1 Z ( p +1) p J 1 Z ( p +1)( p +1) Z ( p +1) M J p +1
− =
(M ) ZM1 Z Mp J p Z M ( p +1) Z MM J M
−1
. (20)
J p +1 Z ( p +1)( p +1) Z ( p +1) M Z ( p +1)1 Z ( p +1) p J 1
= −
JM Z M ( p +1) Z MM ZM1 Z Mp J p
( 1 ) U 1 Z11 Z1 p J 1 Z1( p +1) Z1M J p +1
= + . (21)
( p ) p
U Z p1 Z pp J p Z p ( p +1)
Z pM J M
2169-3536 (c) 2019 IEEE. Translations and content mining are permitted for academic research only. Personal use is also permitted, but republication/redistribution requires IEEE permission. See
http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2019.2919384, IEEE Access
Z 01 + Z 3 + Z 4 0 −Z4 0
0 Z 02 + Z8 + Z 9 0 − Z8
Z mn = .
(27)
−Z4 0 Z 4 + Z5 + Z6 −Z6
0 − Z8 −Z6 Z 6 + Z 7 + Z8
The excitation voltage sources V1 and V2 are inserted in the [ZC] and [ZD] as formulated in equations (19) and (21). After
access branches referred by Port P1 and Port P2 with internal the calculation of the impedance matrix from relation (23),
impedances Z01 and Z02. The internal branches I3…9 are the S-parameter analytical expression can be obtained from
supposed presenting impedances Z3…9. equation (4).
The TAN equivalent graph of the circuit shown in Fig.
4(a) is drawn in Fig. 4(b). It can be understood from the B. Application Case 1: TT-Circuit
topological analysis that this circuit is defined with the A particular case of the TT-cell is treated in the present
following parameters: Nw=1, B=9, N=5, P=2 and M=4. subsection. The validation is based on the comparison
2) TAN MODELLING between the circuit simulation from the electronic and
The branch space analysis enables to write the branch source RF/microwave circuit simulator ADS® from Keysight
voltage vector, the branch current vector and the branch Technologies®. After the brief description of the proof-of-
impedance matrix given by, respectively: concept, the comparative results will be discussed.
Vb = V1 V2 0 0 , (24) 1) POC DESCRIPTION
Fig. 5 depicts the proof-of-concept representing the TT-cell
I1 of the lumped two-port circuit. It is mainly constituted by R
2
I and C components. In this case, the TAN modelling of
I = ,
b
(25)
subsection III-A was applied by defining the branch
9 impedance parameters addressed in Table I.
I
Z 01 0 0 0
0 Z 0 0
02
Z bb = 0 0 Z 3 . (26)
0
0 0 Z 9
0
The connectivity matrix is expressed as:
1 0 0 0
0 1 0 0 FIGURE 5. Proof-of-concept of SPICE circuit schematic.
1 0 0 0 TABLE I
PARAMETERS OF THE PROOF-OF-CONCEPT CIRCUIT SHOWN IN FIG. 5
1 0 −1 0
Cbm = 0 0 1 0 . (28)
Parameter Nature Value
R01=0 Ω
0 0 1 −1
Z01 Resistor
Z02 Resistor R02=0 Ω
0 0 0 1 Z3 Resistor R3=55 Ω
Z4 Capacitor C4=10 pF
0 1 0 1 Z5 Resistor R5=30 Ω
0 −1 0 0 Z6 Capacitor C6=5 nF
Z7 Resistor R7=75 Ω
The mesh voltage derived from equation (9) is transformed Z8 Resistor R8=∞
as: Z9 Resistor R9=0
Um = V1 −V2 0 0 . (29)
Following equation (10), the mesh impedance is given in 2) VALIDATION RESULTS
equation (27).
3) S-PARAMETER RESULTS As numerical application, the routine algorithm of the
The S-parameter model is derived from the impedance proposed fast S-parameter TAN model inspired from the
matrix which is calculated from mesh sub-matrices [ZA], [ZB], workflow of Fig. 3 was implemented in Matlab program.
2169-3536 (c) 2019 IEEE. Translations and content mining are permitted for academic research only. Personal use is also permitted, but republication/redistribution requires IEEE permission. See
http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2019.2919384, IEEE Access
The present S-parameter analyses have been performed parallel network. The parameters of the circuit are addressed
from DC to 300 MHz. Fig. 6 plots the obtained results of the in Table II.
reflection and transmission coefficients from the proposed
TAN computations and the ADS® simulations. To highlight
the comparison over the whole frequency band, linear (in
top) and semi-logarithmic (in bottom) plots are presented. It
can be emphasized that the TAN computation and simulation
results are in excellent correlation in the whole considered
frequency band. In this figure, the TAN computed results are
plotted dotted lines and the ADS® simulation results are
plotted in solid lines. Acting as symmetrical passive circuit, FIGURE 7. Proof-of-concept of SPICE circuit schematic with parasitic
the input and output reflection coefficients, S11=S22 and the effects.
2) S-PARAMETER RESULTS
2169-3536 (c) 2019 IEEE. Translations and content mining are permitted for academic research only. Personal use is also permitted, but republication/redistribution requires IEEE permission. See
http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2019.2919384, IEEE Access
A. Theoretical Investigation
After the problem description, the TAN modelling will be
introduced in the present subsection.
1) PROBLEM FORMULATION
As explained in Fig. 9(a), in this case, we have a four-port
structure represented by d-length two cylindrical conductors
separated by an insulating dielectric structure presenting
inner and external diameters Φi and Φe, respectively. The
resistive loads R1 (resp. R2) and R4 (resp. R3) are connected
between ports P1-P4 (resp. P2-P3) and ports P4-GND (resp. P3-
GND). From DC up to hundred megahertz, as reported in
[33-34], the coaxial cable is equivalent to RLC lumped
circuits. Accordingly, we considered the lumped circuit
model shown in Fig. 9(b).
It is noteworthy that the source internal impedances Z1…4
connected at port P1…4 respectively, were considered to start
the TAN modelling.
2) TAN MODELLING: BRANCH SPACE ANALYSIS
To start the TAN modelling, the graph topology of Fig. 10
FIGURE 8. S-parameter comparison results from the circuit shown in has been elaborated. This topological analysis reveals that
Fig. 7. this circuit is defined with the parameters: Nw=1, B=14, N=6,
P=4 and M=8.
FIGURE 9. Four-port structure: (a) 3D design and (b) equivalent circuit schematic.
2169-3536 (c) 2019 IEEE. Translations and content mining are permitted for academic research only. Personal use is also permitted, but republication/redistribution requires IEEE permission. See
http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2019.2919384, IEEE Access
1 0 0 0 0 0 0 0
The branch space analysis enables to write the branch 0 −1 0 0 0 0 0 0
source voltage vector, the branch current vector and the
0 0 1 0 0 0 0 0
branch impedance matrix given by, respectively:
Vb = V1 V2 0 0 , (30) 0 0 0 −1 0 0 0 0
0 0 0 0 1 0 0 0
I1
2
I 1 0 0 0 −1 0 0 0
I b = , (31) 1 0 0 −1 0 0 −1 0
Cbm = . (33)
14 0 0 0 0 −1 0 1 0
I 0 0 0 0 0 1 0 0
0 0 0 0 1 −1 0 0
Z1 0 0 0 0 0 0 0 0 0 1 −1
0 Z 0 0
2 0 0 0 0 0 −1 0 −1
Z bb = 0 0 Z 3 , (32)
0 −1 0 0 0 1 0 0
0 0 −1 1 0 0 0 0 −1
0 0 0 Z14
with Z5=Z9=Zi, Z6=R3, Z7=R1, Z8=Z12=Ze, Z10=1/(C1s),
B. Numerical Application with Full Wave Simulated
Z11=1/(C2s), Z13=RL and Z14=R2 by denoting s the Laplace Validation Results
variable.
The present subsection is focused on the validation of the
3) TAN MODELLING: MESH SPACE ANALYSIS developed fast TAN S-parameter model with circuit and full
The connectivity matrix related to the branch and mesh wave simulations. After the description of the equivalent
currents of the graph of Fig. 10 is expressed in (33). circuit proof-of-concept, comparisons between the simulated
Following equation (9), the mesh voltage is given by: and computed S-parameters will be performed.
Um = V1 −V2 V3 −V4 0 0 0 0 , (34) 1) POC DESCRIPTION
and the mesh impedance Z mn is calculated from equation Fig. 11(a) depicts the HFSS® 3D design of the coaxial cable.
(11). Then, the mesh sub-matrices [ZA], [ZB], [ZC] and [ZD] Fig. 11(b) shows the hybrid circuit design integrating the full
can be deduced from the previous expression. After the wave simulation results as a black box. Similar to the model
calculation of the impedance matrix from relation (23), the S- suggested in [33-34], the the lumped equivalent model of the
parameter model can be obtained from equation (4). coaxial cable constituted by the R, L and C lumped elements
of the simulated ADS® circuit schematic shown in Fig. 12
was calculated.
2169-3536 (c) 2019 IEEE. Translations and content mining are permitted for academic research only. Personal use is also permitted, but republication/redistribution requires IEEE permission. See
http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2019.2919384, IEEE Access
(b)
FIGURE 11. HFSS® (a) 3D and (b) mixed structure design of the proof- 0
of-concept structure introduced in Fig. 9.
S11 (dB)
-20
TAN
-40 ADS
HFSS
-60
0 100 200 300
0
S21 (dB)
-10
2169-3536 (c) 2019 IEEE. Translations and content mining are permitted for academic research only. Personal use is also permitted, but republication/redistribution requires IEEE permission. See
http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2019.2919384, IEEE Access
-20
TAN - Possibility
ADS to operate up
-40 to several
HFSS gigahertz
-60
0 1 2
10 10 10
0 V. GENERAL CONCLUSION
A fast TAN S-parameter modelling method is developed and
S21 (dB)
2169-3536 (c) 2019 IEEE. Translations and content mining are permitted for academic research only. Personal use is also permitted, but republication/redistribution requires IEEE permission. See
http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
This article has been accepted for publication in a future issue of this journal, but has not been fully edited. Content may change prior to final publication. Citation information: DOI
10.1109/ACCESS.2019.2919384, IEEE Access
Compatibility (EMC EUROPE), 2013, Brugge, Belgium, 2-6 Sept. Issue: Network Methods in Field Modelling, Jan./Feb. 2002, pp. 59-
2013, pp. 241-246. 72.
[9] T. Kley, “Optimized single-braided cable shields,” IEEE Trans. [22] I. Timmins and K.-L. Wu, “An efficient systematic approach to
Electromagn. Comp., vol. 35, no. 1, pp. 1-9, Feb. 1993. model extraction for passive microwave circuits,” IEEE Trans.
[10] S. A. Schelkunoff, “The electromagnetic theory of coaxial Microw. Theory Techn., vol. 48, no. 9, Sept. 2000, pp. 1565-1573.
transmission lines and cylindrical shields,” Bell System Technical [23] T. L. Moss and Y. Chen, “Mesh analysis for extracting the S-
Journal, vol. 13, no. 4, pp. 532-579, Oct. 1934. parameters of lumped element RF and microwave circuits,” Int.
[11] R. B. Schulz, V. C. Plantz, and D. R. Brush, “Shielding theory and Journal of Electrical Engineering Education, vol. 51, no. 4, pp. 330–
practice,” IEEE Trans. Electromagn. Comp., vol. 30, pp. 187-201, 339, Oct. 2014.
1988. [24] J. Yan and X. Liao, “Equivalent lumped circuit model and S-
[12] S. Sali, “An improved model for the transfer impedance calculations parameter of indirect-heating thermoelectric power sensor,” Sensors
of braided coaxial cables,” IEEE Trans. Electromagn. Comp., vol. and Actuators A: Physical, Vol. 240, Apr. 2016, pp. 110-117.
33, no. 2, pp. 139–143, May 1991. [25] F. Fesharaki, T. Djerafi, M. Chaker and K. Wu, “S-Parameter
[13] IEC 62153-4-15: 2015: Metallic communication cable test methods: Deembedding Algorithm and Its Application to Substrate Integrated
Part 4-15: EMC – Test method for measuring transfer impedance and Waveguide Lumped Circuit Model Extraction,” IEEE Trans.
screening attenuation or coupling with Triaxial cell, Edn. 1, 4 Dec. Microw. Theory Techn., vol. 65, no. 4, Apr. 2017, pp. 1179-1190.
2015. [26] G. Kron, Tensor Analysis of Networks, Wiley, New York, Chapman
[14] B. Demoulin and L. Kone, “Shielded cable transfer impedance & Hall, London, 1939.
measurements high frequency range 100 MHz–1 GHz,” IEEE-EMC [27] O. Maurice, “Elements of theory for electromagnetic compatibility
Newsletter, pp. 42–50, 2011. and systems,” Bookelis, Aix en Provence, France, 2017.
[15] HFSS - 3D Electromagnetic Simulation Software. [ONLINE] [28] O. Maurice, A. Reineix, O. Maurice and A. Reineix, “Link between
Available at: https://www.ansys.com/fr- the free field and the elements of frontiers in a complex structure,”
fr/products/electronics/ansys-hfss. [Accessed 1 February 2019] HAL Id: hal-00265343, 2008.
[16] CST – Computer Simulation Technology. 3D Electromagnetic [29] O. Maurice, A. Reineix, P. Durand, F. Dubois, “Kron’s method and
Simulation Software. [ONLINE] Available at: https://www.cst.com/. cell complexes for magnetomotive and electromotive forces,” Int. J.
[Accessed 1 February 2019] Applied Mathematics, vol. 44, no. 4, pp. 183-191, 2014.
[17] ADS – RF/microwave electronic circuit Simulation Software. [30] O. Maurice, A. Reineix, P. Hoffmann, B. Pecqueux, P. Pouliguen,
[ONLINE] Available at: “A formalism to compute the electromagnetic compatibility of
https://www.keysight.com/main/software.jspx?cc=FR&lc=fre&ckey complex networks,” Advances in Applied Science Research, vol. 2,
=2212036&nid=- no. 5, pp. 439-448, 2011.
34333.804574.02&id=2212036&cmpid=zzfindeesof-ads-downloads- [31] B. Ravelo and O. Maurice, “Kron-Branin Modeling of Y-Y-Tree
and-trials. [Accessed 1 February 2019] Interconnects for the PCB Signal Integrity Analysis”, IEEE
[18] R. Otin, J. Verpoorte, and H. Schippers, “Finite Element Model for Transactions on Electromagnetic Compatibility, Vol. 59, No. 2, Apr.
the Computation of the Transfer Impedance of Cable Shields,” IEEE 2017, pp. 411-419.
Trans. Electromagn. Comp., vol. 53, no. 4, Nov. 2011, pp. 950-958. [32] Z. Xu, Y. Liu, B. Ravelo, J. Gantet, N. Marier and O. Maurice,
[19] K. Kurokawa, “Power waves and the scattering matrix,” IEEE Trans. “Direct Time-Domain TAN Model of 3D Multilayer Hybrid PCB:
Microw. Theory Techn., vol. 13, no. 2, Mar. 1965, pp. 194–202. Experimental Validation,” IEEE Access, Vol. 6, No. 1, Dec. 2018,
[20] L. Zhu and K. Wu, “Unified equivalent-circuit model of planar pp. 60645-60654.
discontinuities suitable for field theory-based CAD and optimization [33] J. C. Rautio, “Synthesis of Compact Lumped Models From
of M(H)MIC’s,” IEEE Trans. Microw. Theory Techn., vol. 47, no. 9, Electromagnetic Analysis Results,” IEEE Trans. Microw. Theory
pp. 1589-1602, Sept. 1999. Techn., vol. 55, no. 12, Dec. 2007, pp. 2548-2554.
[21] R. Mittra, A. Suntives, M. S. Hossain and J. Ma, “A systematic [34] P. Xiao, P.-A. Du and B. Zhang, “An Analytical Method for
approach for extracting lumped circuit parameters of microstrip Radiated Electromagnetic and Shielding Effectiveness of Braided
discontinuities from their S‐parameter characteristics,” Int. J. Num. Coaxial Cable,” IEEE Trans. Electromagn. Comp., vol. 61, no. 1, pp.
Mod.: Electronic Networks, Devices and Fields, vol. 5, no. 1, Special 121-127, Feb. 2019.
2169-3536 (c) 2019 IEEE. Translations and content mining are permitted for academic research only. Personal use is also permitted, but republication/redistribution requires IEEE permission. See
http://www.ieee.org/publications_standards/publications/rights/index.html for more information.