374 ACES JOURNAL, VOL. 28, NO.

5, MAY 2013

High Accuracy Time Domain Modeling of Microstrip

Discontinuities by Using Modified TDR Based on Barker Codes
with Flat Spectrum and Integrated Side-Lobes
Mohammad Ojaroudi and Esfandiar Mehrshahi

Faculty of Electrical and Computer Engineering,

Shahid Beheshti University, Tehran, Iran
[email protected] and [email protected]

Abstract ─ In this paper, a novel time domain consequence of the Fourier transform, the energy
method by using a modified time domain in the spectrum of a step falls as the frequency
reflectometry is presented for high accuracy increases. On the other hand, an ideal impulse
modeling of microstrip discontinuities. The (Dirac delta) test signal has a theoretically flat
ordinary stimulus signals used in the TDR bandwidth. Recent studies have demonstrated that
technique are voltage step or voltage impulse. In step TDR can successfully be used to characterize
this paper, we propose an alternative technique, the reflection scattering parameter S11 of antennas
whereby a modified excitation pulse based on [4-6].
Barker-Code orthogonal pulses is employed as Several factors affect a TDR system's ability
the stimulus signal in TDR. The advantage to resolve closely-spaced discontinuities. If a TDR
conferred by “Barker codes TDR” is that more system has insufficient resolution, small or
energy is available at higher frequencies in closely-spaced discontinuities may be smoothed
contrast with conventional step or impulse TDR, together into a single aberration in the waveform.
and so a higher bandwidth and higher accuracy of This effect may not only obscure some
the line impedance is achieved. It can also be discontinuities, but it also may lead to inaccurate
advantageous when the user is looking for impedance readings. Rise time, settling time, and
precision in spatial localization, say in a pulse aberrations of the stimulus signal can also
connector or similar in-line structure, as the significantly affect a TDR system's resolution.
increased energy at higher frequencies can help. Two neighboring discontinuities may be
Simulated results are presented to validate the indistinguishable to the measurement instrument if
usefulness of the proposed method for calculating the distance between them amounts to less than
the precise amount of time-dependent equivalent half the system rise time [2, 3]. Also, many factors
circuit of microstrip discontinuity. contribute to the accuracy of a TDR results. These
include the TDR system’s step response, probe
Index Terms ─ Finite difference time domain and interconnect reflections and DUT losses, step
(FDTD), microstrip discontinuity, time domain amplitude accuracy, baseline correction, and the
reflectometry (TDR), and time domain modeling. accuracy of the reference impedance (ZO) used in
the measurements. All TDR measurements are
I. INTRODUCTION relative; they are made by comparing reflected
Time domain reflectometry (TDR) is a well amplitudes to incident amplitude.
known technique that is typically used to measure In this paper, we explore the advantages of
the impedance of discontinuities as a function of making modified impulse TDR results, similar to a
time (or distance) in electronic systems [1-3]. A traditional step TDR but employing a modified
TDR instrument consists primarily of an signal by using orthogonal codes instead of a step
oscilloscope and a test signal generator, where the like or impulse like (Gaussian) signal. This allows
test signal is traditionally a voltage step. As a us to compare theory with simulation. The

Submitted On: Sept. 11, 2012

Accepted On: Mar. 17, 2013
1054-4887 © 2013 ACES
OJAROUDI, MEHRSHAHI: HIGH ACCURACY TIME DOMAIN MODELING OF MICROSTRIP DISCONTINUITIES BY USING MODIFIED TDR 375

motivation behind this work is driven by the fact Generally, we can write the i th order of the
that a TDR is less expensive than a vector network Barker-code stimulus signal as [8],
analyzer (VNA), but more importantly the time N
localization of the energy in the transient test Bi t     i .h(t   i ) (7)
signal means that the user can dispense with the i 1
anechoic chamber that is required for antenna where,  and  are the amplitude and the order of
measurements with a sine wave exciting signal. shifting in the Barker-codes sequence,
The advantage conferred by “Barker codes TDR” respectively. Generally, Barker-codes are subsets
is that more energy is available at higher of pseudo-noise (PN) sequences [8]. The origin of
frequencies than with conventional step or impulse the name pseudo-noise is that the digital signal has
TDR, and so a higher bandwidth and higher an autocorrelation function, which is very similar
accuracy of the line impedance is achieved. to that of a white noise signal: Impulse like. PN
Simulated results are presented to validate the sequence may also be periodic. Such sequences
usefulness of the proposed method for calculating are known as Barker sequences. Barker-codes are
precise amount of time-dependent equivalent commonly used for frame synchronization in
circuit of microstrip discontinuity. digital communication systems. Barker-codes have
a length of at most 13 and have low correlation
II. FREQUENCY DOMAIN side lobes. Barker sequences are too short to be of
COMPARISON OF THE STIMULUS practical use for spectrum spreading. A correlation
SIGNALS SPECTRUM side lobe is the correlation of a code word with a
The unit impulse (Dirac delta) function is time-shifted version of itself. The correlation side
defined as having zero amplitude for all time lobe, Ck for a k-symbol shift of an N-bit code
except at t = 0, where it has infinite amplitude, sequence, {Xj} is given by,
N k
0 t  0 Ck  X X jk
 t    . (1) j (8)
 t  0
j 1
where, Xj is an individual code symbol taking
The Fourier transform for the unit impulse is,
values +1 and -1, for 0 < j < N, and the adjacent
F  t   1 . (2) symbols are assumed to be zero. The Barker-code
In general the unit impulse is a theoretical generator block provides the codes with various
construct and cannot physically exist [7], it is used order listed in Table 1.
as a limiting case for when the width of a pulse
approaches zero. Derived from the convolution of Table 1: Barker-codes for various orders.
Barker-Code  i 
two rectangular (“rect”) functions, the trapezoid
Order (i)
function provides an approximation of a realistic
impulse with finite rise and fall times [7], 1 [-1]
1 t t [-1 1]
ht   rect   A.rect 
2
(3)
   T  3 [-1 -1 1]
in which A is the trapezoid amplitude, T is the full 4 [-1 -1 1 -1]
width at half maximum (FWHM), and  is the 5 [-1 -1 -1 1 -1]
rise/fall time from 0 % to 100 % of the amplitude. 7 [-1 -1 -1 1 1 -1 1]
The Fourier transform of ht  is given by, 11 [-1 -1 -1 1 1 1 -1 1 1 -1 1]
F ht   AT sin c f sin c fT  .
(4) 13 [-1 -1 -1 -1 -1 1 1 -1 -1 1 -1 1 -1]
The second order of the Barker-code functions is
defined as, Figure 1 presents a number of stimulus signals
B2 t   h(t  1)  h(t  1) . (5) with their spectra that will be compared with the
spectrum of Barker-codes. It can be seen that even
The Fourier transform of B2 t  is given by, the realistic, limited Barker signal contains more
F B2 t   (e  j 2f  e j 2f ). AT sin c f sin c fT  . (6) energy than an ideal rectangular pulse. When the
step waveform is not ideal, but similar to what is
376 ACES JOURNAL, VOL. 28, NO. 5, MAY 2013

practically available today, the comparison induce a voltage difference on the microstrip line.
becomes even more favorable, as can be seen in These two effects can be modelled as a T-shaped
the same figure in [9]. Finally, as shown in Fig. 1, LC circuit, as shown in Fig. 2 (b) [12]. The
the Barker-codes with higher orders have a flat equivalent circuit parameters are defined as,
frequency response with integrated side-lobes  reff  W2   reff  0.3  W1 h  0.264  (9)
[10]. C  0.00137 h 1    
Z C1  W1   reff  0.258  W1 h  0.8 

L1 
LW 1
L
, and L2 
LW 2
L
LW 1  LW 2 LW 1  LW 2
where, 2
 Z  ,
LWi  Z Ci  reff c , and L  0.000987h1  C1 
 ZC 2 

where LWi for i =1, 2 are the inductances per unit

length of the appropriate microstrips, having
widths W1 and, W2, respectively. Zc and εreff denote
the characteristic impedance and effective
dielectric constant corresponding to width, and h is
the substrate thickness in micrometres [12]. The
optimal dimensions of the equivalent circuit model
parameters for the proposed microstrip
transmission line with step-like discontinuity are
specified in Table 2.

Fig. 1. An ideal unit impulse, rectangular pulse,

and some practical Barker-codes of unit amplitude
and its Fourier transform.

III. RESULTS AND SISCUSSIONS

This section describes the simulated line
impedance of the microstrip discontinuity using (a)
the proposed Barker-codes TDR and compares it
with the simulated line impedance obtained using
the conventional step TDR and the equivalent
circuit methods. In order to illustrate the proposed
method performance, a step-like microstrip
discontinuity with mentioned design parameters (b)
were simulated, and the TDR results of the input
impedance for them in equivalent circuit and full- Fig. 2. (a) Geometry of the proposed microstrip
wave analysis cases are presented and discussed. transmission line with step-like discontinuity and
The simulated full-wave TDR results are obtained (b) step-like microstrip discontinuity and its
using the Ansoft simulation software high- equivalent circuit model [12].
frequency structure simulator (HFSS) [11].
The proposed microstrip transmission line Table 2: The dimensions of proposed microstrip
with step like discontinuity is shown in Fig. 2 (a), transmission line with step-like discontinuity.
which is printed on an FR4 substrate of thickness Param. mm Param. mm Param. mm
0.8 mm, permittivity 4.4, and loss tangent 0.018. WSub 12 LSub 18 Wf 1.5
This defected structure on the 50-Ω microstrip line
h 0.8 WS 4 LS 2
will perturb the incident and return currents and
OJAROUDI, MEHRSHAHI: HIGH ACCURACY TIME DOMAIN MODELING OF MICROSTRIP DISCONTINUITIES BY USING MODIFIED TDR 377

The equivalent circuit model parameters are listed impedance of the microstrip line. At impedance
in Table 3. discontinuities, part of the input signal is reflected.
These reflections, after traveling back, reach
Table 3: Equivalent circuit model parameters. terminal port 1 and are observed there [9]. From
Element DGS these observations, the characteristic impedances
L1 1.131 (nH) along the transmission line can be computed. As
C 1.181 (pF) shown in Fig. 4, when the Trise increases from zero
L2 0.858 (nH) in ideal step case to 30 psec, the line impedance in
the center location of the step discontinuity is
The simulated TDR curves with different varied from 35 GHz to 38 GHz. From these
values of rise time and discontinuity distance from results, we can conclude that the line impedance in
reference port (d) are plotted in Fig. 3. As shown the discontinuity location is controllable by
in Fig. 3, when the distance between the reference changing the Trise. In order to decrease the
port and the discontinuity location increases from discrepancy between the simulated TDR data and
10 mm to 50 mm, the amount of the line the ideal step result and also to achieve the
impedance is varied from 33 Ω to 37.5 Ω. Also, as accurate impedance characteristics for the
the rise time of the excitation signal increases from designed discontinuity, we need a stimulus signal
20 psec to 30 psec, the port reflection over shoot is with lower Trise.
varied from 4 % to 7 %. From these results, we
can conclude that the accuracy of discontinuity is
varied by changing d and rise time [13]. Therefore,
due to variation of the results in various conditions
of distance and Trise, to have an accurate judgment
for the proposed method performance we need to
plot TDR curves in a stable and constant condition
for all of scenarios, as shown in Figs. 4 and 5.

Fig. 4. Time dependent impedance with incident

pulse having different rise time when a 1 mm step
is inserted.

To improve the accuracy of the TDR results,

without decreasing Trise, Fig. 5 shows the
simulated reflection waveform observed at port 1
of the defected structures, with a 50 Ω termination
Fig. 3. Simulated TDR characteristic with different at port 2. The excitation source is a step wave with
values of rise time (Trise) and discontinuity amplitude of 1 V and rise time of 20 psec. The
distance from reference port (d). TDR results for the equivalent circuit with C =
1.181 pF, L1 = 1.131 nH, and L2 = 0.858 nH based
Figure 4 shows the effects of Trise with on the method described in [4] and specified in
different values on the line impedance (TDR) in Table 1 are also shown for comparison. The
comparison with the ideal step for the proposed corresponding results generated by Barker-codes
microstrip transmission line with step-like TDR are also shown in this figure. Two cases are
discontinuity. As shown in Fig. 4, these reports studied for Barker-codes TDR, with different code
start, for t = 0.5 nsec, with an impedance just length. It is clearly shown that our TDR results
under 50 Ω. This is indeed the characteristic with modified excitation signal by Barker-codes
378 ACES JOURNAL, VOL. 28, NO. 5, MAY 2013

gives a very good accuracy, and the error is less [2] Time Domain Reflectometry Theory, Agilent
than 1 % in this case in our simulation. It is Technologies, USA, Application Note 1304-2,
apparent from this figure that the energy in the 2006.
modified TDR reflection exceeds the energy in the [3] A. Cataldo, G. Monti, E. De Benedetto, G.
Cannazza, L. Tarricone, and L. Catarinucci,
step TDR reflection. Therefore, the modified
“Assessment of a TD-based method for
Barker-codes TDR results follows the characterization of antennas,” IEEE Trans.
transmission line reference more closely than does Instrum. Meas., vol. 58, no. 5, pp. 1412-1419, May
the ordinary step TDR results. 2009.
[4] M. Ojaroudi and E. Mehrshahi, “Novel band-stop
small square monopole antenna by using
interdigital strip protruded inside the recangular
slot as a band-stop filter with its equivalen circuit
based in TDR analysis for UWB applications,”
International Symposium on Antennas and
Propagation , ISAP 2012, NAGOYA – JAPAN, 2012.
[5] C. Chiu, W. Chen, K. Liao, B. Chen, Y. Teng, G.
Huang, and L.Wu, “Pad characterization for
CMOS technology using time domain
reflectometry,” in Proc. IEEE Int. RF and
Microwave Conf., in Kuala Lumpur, pp. 215-217,
Dec. 2008.
[6] M. Harper, N. Ridler, and M. Salter, “Comparison
Fig. 5. The reflected waveforms comparison for between root-impulse-energy and vector network
analyzer methods for measuring loss on printed
simulated results by HFSS, equivalent circuit
circuit boards,” in ARFTG Microwave
model, and two cases of Barker-codes for the Measurement Symp., in Portland, OR, pp. 20-25,
proposed defected microstrip structure shown in 2008.
Fig. 2 (a). [7] M. Kowalski, “A simple and efficient
computational approach to chafed cable
IV. CONCLUSION time-domain reflectometry signature prediction,”
Time domain reflectometry (TDR) with finite 25th Annual Review of Progress in Applied
Computational Electromagnetics (ACES),
difference time domain (FDTD) method analysis
Monterey, California, March, 2009.
is a novel technique that is typically used to [8] S. Golomb and R. Scholtz, “Generalized Barker
calculate the time modeling of a microstrip sequences (transformations with correlation
structure. To improve the accuracy of the TDR function unaltered, changing generalized Barker
sequences),” IEEE Trans. Inf. Theory, vol. 11, pp.
results, without decreasing Trise, we proposed an
533-537, Oct. 1965.
alternative technique, whereby a modified [9] C. Lo and C. Furse, “Modeling and simulation of
excitation pulse based on Barker-code orthogonal branched wiring networks,”
pulses is employed instead of the stimulus signal Applied Computational Electromagnetics Society
(ACES) Journal, vol. 23, no. 2,
in TDR. It can also be advantageous when the user pp. 143-146, June 2008.
is looking for precision in spatial localization, say [10] R. N. Bracewell, The Fourier Transform and its
in a connector or similar in-line structure, as the Applications, 3rd Edition, McGraw-Hill
increased energy at higher frequencies can help. International Editions, pp. 61, 2000.
[11] Ansoft High Frequency Structure Simulation
The simulated results showed good agreement (HFSS), Ver. 13, Ansoft Corporation, 2010.
with the numerical prediction. [12] T. C. Edwards, Foundations for Microstrip Circuit
Design, Wiley, Great Britian, 1981.
REFERENCES [13] C. H. Cheng, C. H. Tsai, and T. L. Wu, “A novel
[1] R. Tamas, G. Caruntu, and D. Popa, “A time- time domain method to extract equivalent circuit
domain measuring technique for ultra-wide band model of patterned ground structures,” IEEE
antennas,” Microwave and Optical Technology Microw. Wireless Compon. Lett., vol. 17, no. 11,
Letters, vol. 53, no. 2, pp. 281-286, Feb. 2011. Nov. 2010.
OJAROUDI, MEHRSHAHI: HIGH ACCURACY TIME DOMAIN MODELING OF MICROSTRIP DISCONTINUITIES BY USING MODIFIED TDR 379

Mohammad Ojaroudi was born

on 1984 in Germi, Iran. He
received his B.Sc. degree in
Electrical Engineering from Azad
University, Ardabil Branch and
M.Sc. degree in
Telecommunication Engineering
from Urmia University. From
2010, he is working toward the Ph.D. degree at Shahid
Beheshti University. From 2007 until now, he is a
Teaching Assistant with the Department of Electrical
Engineering, Islamic Azad University, Ardabil Branch,
Iran. Since March 2009, he has been a Research Fellow
(Chief Executive Officer) in the Microwave
Technology Company (MWT), Tehran, Iran. From
2012, Mr. Ojaroudi is a member of the IEEE
Transaction on Antennas and Propagation (APS)
reviewer group and the Applied Computational
Electromagnetic Society (ACES). From 2013 he is a
student member of the IEEE. His research interests
include analysis and design of microstrip antennas,
design and modeling of microwave structures, radar
systems, and electromagnetic theory. He is author and
coauthor of more than 100 journal and international
conferences papers. His papers have more than 300
citations with 10 h-index.

Esfandiar Mehrshahi was born in

Tehran, Iran, in 1964. He received
the B.Sc. degree from the Iran
University of Science and
Technology, Tehran, Iran, in 1987,
and the M.Sc. and Ph.D. degrees
from the Sharif University of
Technology, Tehran, Iran, in 1991
and 1998, respectively. Since 1990, he has been
involved in several research and engineering projects at
the Iran Telecommunications Research Center (ITRC).
He is currently an Associate Professor at Shahid
Beheshti University, Tehran, Iran. His main areas of
interest are the nonlinear simulation of microwave
circuits and microwave oscillator’s spectrum.