An Impedance Matching Method for Optical Disc-Based

UHF-RFID Tags
Simone Zuffanelli, Pau Aguil, Gerard Zamora, Ferran Paredes, Ferran Martn and Jordi Bonache
GEMMA/CIMITEC, Departament dEnginyeria Electrnica,
Universitat Autnoma de Barcelona,
Bellaterra, Spain
Email: simone.zuffanelli@uab.cat, jordi.bonache@uab.cat


Abstract An impedance matching method applicable to
UHF-RFID tags mounted on optical discs, where the main
radiating element is the disc metallic layer (from now on optical
disc-based tags) is presented in this paper. Such a method
provides a very simple analytical approach and allows us to
obtain any arbitrary matching level at the desired frequency, by
using only one reactive element as a matching network. As a
demonstration, the method is applied to the design of a DVD+R
based tag, using the Alien Higgs 3 RFID ASIC, and forcing
conjugate matching at 915 MHz. The tag design and synthesis
process is explained and a bandwidth optimization analysis is also
provided. The validity of the method is confirmed by EM
simulations and experimental measurements. The measured read
range of the fabricated tag reaches 8.3 m at 915 MHz and
presents 50 MHz half-power bandwidth.

Index Terms DVD tagging, radio frequency identification
(RFID), antennas, impedance matching.

I. INTRODUCTION
Optical discs (CDs, DVDs and Blu-Ray discs) are one
of the most common digital data storage format available
nowadays. They are commonly used to store musical
recordings, videogames and movies, educational material, and
personal data among others. Since they are often used as an
individual product storage systems and their data capacity is
limited to few GBs, a great number of discs are usually
handled at music and video stores, libraries, bookstores and
even at educational facilities (e.g. schools, universities).
Passive UHF-RFID tagging of optical discs facilitates their
inclusion in smart inventory systems, where position and
movements of each item are monitored individually.
Moreover, item level tracking can be provided over the supply
chain, including logistics and payment. Nevertheless, UHF-
RFID tagging of optical discs is relatively uncommon, when
compared to general purpose RFID tagging. The reason is that,
due to the presence of a metal film under the disc surface [1],
optical discs cannot be tagged efficiently by using standard
UHF-RFID labels [2]-[5]. On the other hand, the use of on-
metal RFID tags in this context is not suitable for several
reasons. First of all, the relatively high cost of this kind of tags
would increase significantly the retail price of the discs. Also,
the thickness and weight of such tags would affect the correct

Fig. 1 An optical disc-based tag mounted on a DVD+R disc.

disc operation and storage, since optical discs are very thin
(1.2 mm). Some efforts have been made in the past in order to
design specific UHF-RFID tags for the optical discs. The
approach reported in literature [6]-[7] was mainly to take
advantage of the central region of the disc, which does not
contain the metal layer, by embedding a dipolar-type antenna
based tag. However, this strategy presents some issues which
prevent this kind of tags to exhibit good performance, when
compared to general purpose UHF-RFID tags. Due to the
small dimensions and the annular shape of the available area,
bending [6] or meandering [7] the dipole is necessary.
Moreover, the proximity of the disc metal layer surrounding
the dipole leads to the formation of image currents on the
layer. Therefore radiation efficiency of the tag is seriously
degraded, resulting in a read range in the order of 0.35-0.4 m,
which is inadequate for the aforementioned RFID applications.
Moreover, impedance matching issues arise from the severe
reduction of the input resistance of the dipole antenna. Usually
this problem is solved by means of inductive loop based
matching networks [6]-[7], which however introduce further
ohmic losses due to the reduced width of the inductive paths.
An alternative approach was recently proposed in [8] with
optical disc-based tags (Fig. 1), in order to overcome these
limitations. In this kind of tags the metal layer of the disc is
used as the main radiator, taking advantage of its dimensions
and avoiding the formation of image currents. As a result, read
ranges above 3 m were obtained over the whole UHF-RFID
2014 IEEE International Conference on RFID (IEEE RFID)
978-1-4799-3587-1/14/$31.00 2014 IEEE 15
band (840-930 MHz), with a peak value of 5.2 m at the center
frequency (885 MHz). The system radiates as a two elements
short dipole array, linearly polarized. These promising results,
along with the design simplicity and the possibility to
manufacture the tag by means of serigraphy directly over the
disc surface, encourage the authors to carry on the study of this
solution. An impedance matching analysis is useful in
particular, to provide the design flexibility required to match
the antenna to the various RFID ASICs available on the
market. Indeed, as stated in [8], while the input reactance of
the antenna can be easily matched by varying the tag
geometry, its input resistance (in the range of 8-12 O) cannot
be easily controlled, since it presents a very small variation as
a function of tag geometry. For the case presented in [8],
where the ASIC (NXP UCODE G2XM) resistance (16 O)
matched relatively well to the antenna input resistance, the
power reflection coefficient was acceptable (13 dB at
885 MHz). However, depending on the ASIC input resistance,
an impedance matching network may be needed. For instance,
the more recent and sensitive Alien Higgs 3 ASIC presents a
higher input resistance, in the order of 25 O at 915 MHz [9],
thus providing poor impedance matching when directly
connected to the antenna.
An impedance matching analysis for the optical disc-based
UHF-RFID tags is provided in this work, in section II. Based
on such analysis, a tag design and synthesis process is
explained in section III. As a proof of concept, a prototype
based on the Alien Higgs 3 ASIC is designed and fabricated,
and the measured read range is presented in section IV.

II. MATCHING NETWORK ANALYSIS
Let us now introduce an impedance matching strategy, which
provides a practical method to obtain the required matching
levels analytically, for the tag design described in [8]. As
stated in Section I, such design allows an effective control over
the antenna reactance X
A
at the center frequency. In fact, by
properly adjusting the tag geometry, it is possible to obtain
inductive or capacitive behaviour at the frequency of interest.
On the other hand, the tag geometry permits a very little
control over the antenna resistance R
A
, which can be in general
different from the RFID ASIC input resistance, thus suggesting
the need of a matching network. In general, a minimum of two
elements are needed (e.g. L-section network) in order to match
two given impedances. However, due to the degree of freedom
in the antenna reactance X
A
, the network can be simplified by
using only the shunt reactive element, which matches the
resistive parts of the impedances. An analytical treatment of
the problem is provided as follows.

A. Resistance scaling
Let us assume the case of a generic impedance Z = R+jX in
shunt with a reactance X
p
. Considering Z' = R'+jX' to be the
impedance resulting from the parallel combination of Z and X
p
,
the value of Z' can be easily inferred, yielding:


2 2
2 2 2 2
( )
( ) ( )
'
p p p
p p
X R X X X R X
Z j
R X X R X X
( + +

+
+ + + +
=

(1)



Now, it is possible to imagine the real part of Z' to be equal to
R times a scaling factor o, with o > 0. That is:

' R R o = (2)

This equality leads to a second order equation which solution
provides the value of X
p
as a function of R, X, and o:


2 2 2
1
1 2 ( ) 0
p p
X X X X R
o
| |
+ + + =
|
\ .

(3)

In order to have real solutions (pure reactive values), the
discriminant A of (3) must be positive. This sets a limit on the
maximum value of o, which is:


2
1
X
R
o
| |
s +
|
\ .
(4)

If (4) is satisfied, we have 2 real solutions for (3), given by:



2 2
1,2
(1 )
1
p
X X R
X
o o o
o
( +

=

(5)
It can be demonstrated with algebra that for o < 1 (resistance
decrease), solution 1 is capacitive and solution 2 is inductive.
Moreover, for o > 1 (resistance increase) both solutions are
inductive when X < 0, and capacitive when X > 0. The limiting
case (o = 1) is provided by an infinite reactance (whether
positive or negative) or by the solution X
p
= (X
2
+R
2
)/2X.
It is also possible to demonstrate that the imaginary part of
Z' can be written as:


1,2
'
2
X o
A
=
(6)

showing that for a given value of o, the two solutions for X
p

always give opposite values for X', which are capacitive for
solution 1 and inductive for solution 2.
We conclude that the resistance scaling approach can be
used in order to solve impedance matching problems. In fact,
by properly choosing the scaling factor o, it is possible to
control the matching level between the impedance Z and a load
impedance, assuming that the load reactance resonates with the
transformed reactance X'. This can be very useful for the
design of UHF-RFID tags since that, due to bandwidth
2014 IEEE International Conference on RFID (IEEE RFID)
16



Fig.2 Shunt reactance solutions for the reactive impedance
approximation (a), and transformed reactive values (b) for the case
X > 0. For X < 0 solution 1 and 2 are interchanged.

considerations, a certain matching level could be preferred
over conjugate matching.

B. Reactive impedance approximation
Given that the typical input impedance of an RFID ASIC is
strongly capacitive, it would be useful to find an approximate
solution for (3) in the case of Z being strongly reactive. Under
the assumptions:


2 2
1 X R o
(7a)

2 2
X R
(7b)

the solutions (5) of equation (3) can be simplified to:


1,2
1
p
X X
o
o
~

for X>0 (8a)

1,2
1
p
X X
o
o
~

for X<0 (8b)

suggesting that the value of X
p1,2
can be easily calculated
starting from X and o.
The imaginary part of Z' can be evaluated from (6) by
using the conditions (7), leading to a very simple equation:


1,2
' X X o ~ for X>0 (9a)

1,2
' X X o ~ for X<0 (9b)

A graphical representation of the approximated solutions as a
function of o is depicted in Fig.2.

C. Application example
Let us apply the previous method to obtain conjugate matching
at 915 MHz between the Alien Higgs 3 (SOT323 packaged)
ASIC and an antenna based on a DVD disc. As stated above,
we can take advantage of the reactive impedance
approximation (7) by applying the transformation to the chip
impedance Z
C
, obtaining the transformed

impedance Z'
C
. To
verify this, let us calculate the ASIC input impedance at the
center frequency (f
0
=915 MHz), on the basis of the electrical
parameters of the shunt circuit model declared by the
manufacturer (R
c
= 1.5 kO, C
c
= 0.9 pF) [9]. The resulting
impedance is Z
C
= 25j190 O. In this case the load impedance
is the antenna impedance Z
A
. Hence, the resistance scaling
factor providing conjugate matching can be easily calculated
by using (2), assuming the antenna resistance to be R
A
= 10 O
[8], obtaining o = 0.4. As expected, conditions (7) are
satisfied, as it can be proved by evaluating
X
C
2
= 3.610
4
O
2
,1o|R
C
2
= 3.7510
2
O
2
. Thus, since
X
C
< 0, the solutions for X
p
can be calculated by using (8b),
obtaining X
p1
= 327 O and X
p2
= 74 O. The transformed
reactance can be inferred from (9b), yielding X'
C1
= 120 O
and X'
C2
= 120 O. As a consequence, the required antenna
reactance for each case is X
A1
= 120 O and X
A2
= 120 O.

III. TAG DESIGN, SYNTHESIS AND SIMULATION

A. Design process
The solution adopted in [8], in order to take into advantage the
capability of a DVD disc to work as an efficient antenna at the
UHF-RFID frequencies, is to produce an electric coupling
between the tag and the disc by means of a series stepped-
impedance resonator (SIR) synthetized through semi-lumped
elements. The capacitance between the tag metal (top layer)

b)
a)
2014 IEEE International Conference on RFID (IEEE RFID)
17
a)
b)


Fig. 3 Proposed tag layout (on top). The matching elements are
illustrated at bottom left (solution 1) and bottom right (solution 2).

and the disc metal (inner layer), C
tag
, provides electric
coupling. The series inductance L
tag
, realized through a thin
line (e.g., 0.4 mm), forces a controllable series resonance,
which provides a virtual short circuit between the ASIC and
the metal layer of the disc. In this work, in order to simplify
the design stage, the authors maintain the tag topology
proposed in [8] (Fig. 3). Conjugate matching will be obtained
by introducing the matching element in shunt to the ASIC and
varying the values of the tag geometrical parameters in order
to obtain the required tag reactance at the center frequency.
That is, the matching element and the rest of the tag are
simulated separately, and the tag response is obtained by
joining them in a final simulation.
Let us now introduce the tag equivalent circuit model (Fig.
4b), which is based on the model proposed in [8] (Fig. 4a).
With respect to the latter, the antenna impedance model has
been simplified in this work. In fact, since its first resonance
frequency occurs at 4.8 GHz, the RLC tank which models the
disc impedance Z'
disc
in [8] can be approximated to the series
combination of the disc inductance L'
p
and a resistance, which
is directly the antenna resistance R
A
. The tag (excluding the
ASIC and matching element) is still modeled as an LC series
resonator in series with the disc impedance. The value of C
tag

is given by the capacitance between the disc metal layer and
the tag external branch (with length l
C
and width W
C
(Fig. 3)),
while L
tag
is given by the inductance of the thin electrical line
connecting the external branch to the ASIC. On the left side
(Fig. 4), the transformed chip impedance Z'
C
has been added as
a shunt combination of the ASIC parallel model and the
R
c
C
c
L
ta g
C
ta g
R'
p
L
ta g
C
ta g
X
p1,2
Z'
C
=R'
C
+jX'
C Z
A
=R
A
+jX
A
s
C'
p
L'
p
R
c
C
c
L
tag
+L'
p
/2 C
tag
R
A
L
tag
+L'
p
/2 C
tag
X
p1,2
Z'
C
=R'
C
+jX'
C Z
A
=R
A
+jX
A
s

Fig. 4 Equivalent circuit models of the tag coupled to the disc.

matching reactance X
p
. The proposed model does not take into
account the ohmic losses introduced by the matching network
and the sections L
tag
and C
tag
.
The design process begins with the matching elements,
which are implemented through semi-lumped components, a
solution widely used in the design of UHF-RFID tags. The
capacitive reactance, required for solution X
p1
, is obtained by
means of an interdigitated capacitor, whose parameters are
number of fingers N, length a
1
, width b
1
and e
1
, spacing c
1
and
d
1
(Fig. 3), whereas the inductive solution X
p2
is provided
through a loop inductor of radius a
2
and width b
2
(Fig. 3). The
final layout is obtained in each case by varying the geometrical
parameters until the simulated reactance reaches the required
value (X
p1
or X
p2
). In order to get better accuracy, the simulated
elements are placed at their final positions on the tag substrate
(Rogers RO3010, 0.254 mm thickness), which is placed on top
of the disc. The tag LC series resonator is not included at this
stage.
The next step is to determine, for each solution, the
geometric parameters associated to the LC resonator. This is
accomplished by simulating the resonator (coupled to the disc)
without the matching network, and varying the capacitor area
in order to adjust the antenna reactance to the required value
(X
A1
or X
A2
). As it will be justified in the next paragraph, the
value of the inductor width W
L
(and consequently the value of
L
tag
) will be fixed during the design process. This will be set to
W
L
= 0.6 mm.
Finally, a simulation of the complete layout cascaded to
the impedance Z
C
is performed, in order to verify impedance
matching. At this stage it is useful to point out that, although
the antenna resistance R
A
(which can be inferred from the
previous simulation) cannot be easily controlled by the tag
geometry, a slight variation of its value occurs varying the LC
2014 IEEE International Conference on RFID (IEEE RFID)
18
resonator geometry. As a result, the minimum value of the
power reflection coefficient is expected to present, for each
solution, a deviation from the expected value. In case that poor
impedance matching is obtained, the design of the matching
element should be repeated taking into account a corrected
value of o, which can be easily determined on the basis of the
simulated value of R
A
. Moreover, a small error in the matching
element reactance X
p
(geometrical tolerances might make
difficult obtaining the exact required value) can introduce a
frequency shift Af (with respect to f
0
) of the minimum power
reflection coefficient, due to the deviation of X'
C
from the
expected value. An additional contribution to the frequency
shift could be introduced by the non-zero frequency slope of
the antenna conductance G
A
, which exists because the antenna
cannot be modeled as a pure parallel load. However, the total
frequency shift can be corrected by simply tailoring the value
of l
C
in the final simulation, in order to adjust the antenna
reactance to the required value.

B. Bandwidth considerations
In this section, the final tag bandwidth associated to each
solution will be discussed. For fabrication and measurement,
the solution that provides broader bandwidth will be chosen.
First of all, we discuss the value of L
tag
. Since the tag
introduces a series resonance, bandwidth optimization suggests
minimizing the inductance L
tag
, thus adjusting the reactance by
changing the value of C
tag
. Moreover, it is difficult to
efficiently control the value of L
tag
, since it depends on the
logarithm of the strip width W
L
. On the contrary, the value of
C
tag
can be easily controlled since it linearly depends, in a first
order approximation, on the capacitor area. Therefore, it is
reasonable to adjust the tag reactance by simply varying the
capacitor length l
C
and the capacitor width W
C
.
Let us now compare solutions 1 and 2 from the bandwidth
point of view. To do this, it is convenient to analyze the
frequency slope of the transformed chip susceptance B'
C
,
which is related to the potential tag bandwidth, as stated in
[10]. With a little mathematical effort it is possible to
demonstrate that, under the reactive impedance approximation
(7), the susceptance slopes at e
0
are, in each case:


0
1
1
C
c
B
C
e
e o
' c
~
c

(10)

0
2
1 2
C
c
B
C
e
o
e o
' c +
~
c

(11)

which are valid for o < 1. As a result, the susceptance slope
ratio amounts to 1 2 o + , suggesting that a higher bandwidth
could be obtained by the capacitive matching (solution 1).
However, it is necessary to sum the antenna susceptance slope
A
B
e
c
c
to (10) and (11) in order to estimate the final tag

Fig. 5 Simulated power reflection coefficient.

bandwidth [10]. To do this, it is firstly convenient to analyze
the antenna reactance and its frequency slope, for each case, in
the vicinity oI e
0
. Based on the equivalent circuit model (Fig.
4b), these can be written as:


0
0
0
(2 )
s
A tag p s
s
X L L
e
e e
e
e e
| |
' = +
|
\ .

(12)

0
2
0
(2 ) 1
s A
tag p
X
L L
e
e
e e
(
| | c
' ( = + +
|
c
(
\ .


(13)

where
1/ 2
(L 2)
s tag p tag
L C e

' ( = +

is the tag series resonance
angular frequency. Since an inductive antenna behavior is
required for solution 1 and a capacitive behavior is required
for solution 2, conditions e
s1
< e
0
and

e
s2
> e
0
must be
satisfied, and consequently e
s2
> e
s1
. Therefore, taking into
account that L
tag
is fixed in the design process and L'
p
is
weakly dependent on the tag geometry, a smaller value of (13)
is obtained for solution 1. As a final step, let us write the
general expression of the susceptance slope as a function of
the impedance:


0 0 0
2 2
4 4
2
A A A A A A A
A A
B X R R X R X
Z Z
e e e
e e e
c c c
= +
c c c

(14)

which can be approximated to:


0 0
2
1
A A
A
B X
X
e e
e e
c c
~
c c

(15)

because R
A
<<X
A
. Since X
A1
= X
A2
and the antenna reactance
slope is smaller for solution 1, susceptance slope (15) is also
smaller for solution 1. Therefore it can be concluded that
2014 IEEE International Conference on RFID (IEEE RFID)
19
a)
b)
b)

Fig. 6 Radiation pattern (gain, expressed in dB) at 915 MHz for the
simulated tags: (a) E-plane, (b) H-plane.

capacitive matching provides both smaller ASIC and antenna
susceptance slopes, resulting in a greater bandwidth. An
interesting assumption, which allows us to quantify the
maximum bandwidth achievable for each solution, is to
consider, in terms of susceptance slope, the antenna as a pure
inductor (for solution 1) and a pure capacitor (for solution 2).
As stated in [10], these ideal networks provide the optimal
bandwidth in single resonant UHFRFID tags with conjugate
matching. The resulting 3 dB bandwidths, which are valid for
u < 1, are:


3
1
MAX
dB
c c
f
R C
o
t

A =

(16)


3
1
1
MAX
dB
c c
f
R C
o
t o

A =
+

(17)

for solutions 1 and 2, respectively. Notice that, according to
[10], the factor (aR
c
C
c
)
-1
represents the optimal half-power
bandwidth for a given ASIC. This means that, given o = 0.4,
solution 1 can provide 63% of the ASIC optimal half-power
bandwidth ((aR
c
C
c
)
-1
= 236 MHz for Alien Higgs 3) at most,
while solution 2 can reach only 38%.




Fig. 7 Electric field for solution 1 (a) and solution 2 (b). The cutting
plane is orthogonal to the disc surface and intersects the center of the
disc.

C. Simulation results
The EM simulations are performed by means of the CST
Microwave Studio commercial software (the frequency domain
solver is used). The polycarbonate (c
r
= 2.88, tano = 0.012
[11]) disc substrate is modeled as ring-shaped item, with
external and internal diameters of 120 mm and 15 mm,
respectively, and with 1.2 mm thickness. The metal layer of
the disc, which is buried 0.6 mm under the surface of the
substrate, is modeled as a ring-shaped ohmic sheet
(R
.
= 0.5 O, corresponding to a silver film with 50 nm
thickness [8]), with external and internal diameters of 120 mm
and 40 mm, respectively. The substrate used for the tag is the
Rogers RO3010 (with thickness t = 0.254 mm, c
r
= 10.2,
tano = 0.0023), which has a structural function only. In fact,
the high permittivity of the material does not have a role,
neither in size reduction, nor in increasing the tag
performance. Commercial versions of the tag can be
engineered without any substrate, i.e. screen- printing on the
disc surface, in order to minimize the cost [8].
The first simulations provided a possible set of solutions
for the geometrical dimensions of the matching elements. The
values are found to be N = 6, a
1
= 5 mm, b
1
= 0.5 mm,
c
1
= 0.2 mm, d
1
= 0.2 mm, e
1
= 0.2 mm for the capacitive
solution, and a
2
= 3 mm and b
2
= 0.2 mm for the inductive
solution. The tag LC resonator geometry was then obtained for
each solution, as described at Section III-A. The results are
l
C1
= 13 mm, W
C1
= 2 mm and l
C2
= 3.1 mm, W
C2
= 2 mm. The
values of the simulated antenna resistance are R
A1
= 9 O and
R
A2
= 6 O at 915 MHz. Therefore, since R'
C
= 10 O, the
expected values of the power reflection coefficient at
915 MHz are 25 dB and 12 dB, respectively. Since both
values provide very good performance in terms of read range,
the matching elements do not need to be corrected. The
a)
2014 IEEE International Conference on RFID (IEEE RFID)
20
simulated power reflection coefficients of the complete tag are
shown in Fig. 5. The matching notches (25 dB for solution 1
and 19 dB for solution 2) are clearly visible as expected, and
the frequency shifts are Af
1
= 25 MHz and Af
2
= 34 MHz. In
order to correct the shifts, a tailoring of the capacitor length l
C

is performed as a final step. The resulting geometrical values
are l
C1
= 14.4 mm and l
C2
= 4.1 mm. The final curves for |s|
2
,
centered on f
0
, are depicted in Fig. 5. The expected matching is
achieved in both cases (31 dB and 11 dB for cases 1 and 2,
respectively) and the 3 dB bandwidths are BW
1
= 65 MHz
and BW
2
= 40 MHz. Even if bandwidths are smaller than the
optimal values predicted in (16) and (17) (where an ideal
network is cascaded to the transformed chip impedance), it is
interesting to note that the ratio BW
1
/BW
2
approaches the value
obtained in the optimal case (1+o
1/2
= 1.63 for o = 0.4).
Let us now compare the antenna gain of the tag for the two
different solutions, since its value is strictly related to the tag
read range. Simulated far field (Fig. 6) indicate that, while
directivities are similar in both cases (4 dB approximately) an
important difference in the radiation efficiency exists for the
two cases, providing c
rad1
= 26% and c
rad2
= 14%. There are
mainly two reasons for this difference. Firstly, the inductor is
implemented through a thin line, which introduces more ohmic
losses than an interdigitated capacitance does. Moreover, the
second solution transforms the ASIC impedance, as seen from
the antenna, from capacitive to inductive. This corresponds to
change the sign of the internal electric dipole existing around
the ASIC, with respect to the external dipole created by the
optical disc antenna (Fig. 7). As can be seen in the figure, in
the case of inductive matching the internal dipole is opposite
in sign to the external dipole, causing a reduction of the
radiation efficiency. The simulated antenna gains are
G
1
= 2 dB and G
2
= 4.3 dB, corresponding to the expected
read ranges r
1
= 9.3 m and r
2
= 7.1 m, (considering
EIRP = 4 W, P
th
= 17 dBm), according to the well-known
equation [12]:

4
r
th
EIRP G
r
P
t
t

=

(18)

where EIRP is the equivalent isotropic radiated power, is the
free-space wavelength at the working frequency, G
r
is the gain
of the tag antenna, : = (1-[s|
2
) is the power transmission
coefficient, and P
th
is the ASIC read sensitivity. This
theoretical value corresponds to the read range in free-space,
with both antennas (reader and tag) correctly oriented at the
maximum gain direction.

IV. FABRICATION AND MEASUREMENTS

A. Experimental setup and read range equations
The measurement setup consists of an Agilent N5182A vector
signal generator, capable of generating RFID interrogation
frames, connected to a TEM cell by means of a 50 O coaxial
cable. A circulator is used to send the backscattered signal
from the TEM cell to an Agilent N9020A signal analyzer, in
order to decode the digital RFID frames generated by the tag,
which is placed inside the TEM cell and oriented along the
maximum directivity axis. To measure the tag read range, an
RFID interrogation frame is sent to the TEM cell at different
power levels, in order to determine the minimum power level
P
min
required to activate the tag, that is, to receive a
backscattered response frame. An electric field probe is then
placed at the tag position in order to measure the root mean
square of the electric field E
rms
generated by the interrogation
frame at the power P
min
. Therefore, E
rms
is the minimum
electric field required for the tag operation, and it is possible to
calculate the read range directly from its value. In fact, the
average power density S associated to a plane wave is
determined from the value of the electric field E
rms
according
to:

2
rms
E
S
q
= (19)
where p is the intrinsic impedance of free space. For a
radiating antenna it is possible to calculate the far field
Poynting vector module in a given direction as:

2 2
4 4
t t
P G EIRP
S
r r t t

= =

(20)
where P
t
is the total transmission power, G
t
is the antenna gain
and r is the distance from the antenna. Thus, the read range r
can be obtained by equating (19) and (20), resulting in:

30
rms
EIRP
r
E
=

(21)
The method described above is then repeated for each
frequency of interest (e.g., at 5 MHz steps) in order to obtain
the read range in the whole UHF band.

B. Experimental results
To validate the simulated results, the layout associated to
solution 1 (described in section III-C) was fabricated by means
of a PCB milling machine (LPKF-H100). As explained above,
solution 1 has been chosen because it provides wider
bandwidth and higher antenna gain. The measurements (not
shown) present a frequency shift of the tag response, with a
read range peak at 1 GHz approximately. This effect has been
attributed to the intrinsic indetermination over the electrical
properties of the DVD materials at the UHF band, and to the
simplifications made in the modeling of the disc. In fact,
unlike most substrates used for tag implementation, DVD disc
materials are not intended for microwave design. Therefore,
manufacturers neither report nor control their electrical
properties [8]. Moreover, some of the geometrical parameters
2014 IEEE International Conference on RFID (IEEE RFID)
21


Fig. 8 The fabricated tag, mounted on a DVD-R disc. The matching
network is depicted at bottom left.

(0.2 mm spacing) of the capacitive matching element are very
close to the fabrication tolerances of the PCB milling machine,
introducing further error in the tag response.
Based on the measured results, a second prototype of the tag
was synthetized and fabricated in order to correct the
frequency shift. The matching position has been easily
adjusted by simply tailoring the length l
C
of the tag
capacitance. The corrected value of the geometrical parameter
is l
C
= 22 mm. The tag, mounted on a DVD+R disc, is
depicted in Fig. 8. The read range measurement, depicted in
Fig. 9, shows a 8.3 m peak at the center frequency, in a good
agreement with the simulated value. Based on the read range,
the 3 dB bandwidth has been evaluated to be BW = 50 MHz.

V. CONCLUSION
An analytical impedance matching method for UHF-RFID tags
mounted on optical discs has been presented in this paper.
Such technique allows us controlling the impedance matching
level with a very easy design and synthesis process. The
method has been then applied to design an Alien Higgs 3 based
tag, optimizing the read range at the 902-928 MHz band. The
tag design process has been presented, and the bandwidth
limitations and radiation efficiencies associated to the different
solutions have been discussed. Based on that, the optimal
solution has been chosen for fabrication and measurement. The
measurements agree with simulations, and measured read
range reaches 8.3 meters at the center frequency (915 MHz),
with a 50 MHz half-power bandwidth.


Fig. 9 Measured read range.

REFERENCES
[1] F. R. Byers, Care and Handling of CDs and DVDs - A Guide for
Librarians and Archivists, NIST Special Publication 500-252,
www.nist.gov.
[2] D. M. Dobkins and S. Weigand, Environmental effects on RFID tag
antennas, IEEE MTT-S International Microwave Symposium, Long
Beach, CA, June 2005, pp. 47.
[3] J. D. Griffin, G. D. Durgin, A. Haldi, and B. Kippelen, RF tag antenna
performance on various materials using radio link budgets, IEEE
Antennas and Wireless Propagation Letters, vol. 5, no. 1, pp. 247250,
Dec. 2006.
[4] K. M. Ramakrishnan and D. D. Deavours, Performance benchmarks
for passive UHF-RFID tags, 13th GI/ITG Conference on Measurement,
Modeling, and Evaluation of Computer and Communication Systems,
Nurenberg, Germany, Mar. 2006, pp. 137154.
[5] S. R. Aroor and D. D. Deavours, Evaluation of the state of passive
UHF-RFID: An experimental approach, IEEE Systems Journal, vol. 1,
no. 2, pp. 168176, Dec. 2007.
[6] W. T. Luk and K. N. Yung, Bending dipole design of Passive UHF-
RFID tag antenna for CD/DVD discs, IEEE Asia Pacific Microwave
Conference, Macau, pp. 1-4, 16-20 Dec. 2008.
[7] A. S. Andrenko, M. Kai, T. Maniwa and T. Yamagajo, Compact
printed-on-CD UHF-RFID tag antennas, IEEE Antennas and
Propagation Society International Symposium, Honolulu (USA), pp.
5455-5458, 9-15 June 2007.
[8] S. Zuffanelli, G. Zamora, F. Paredes, F. Martn and J. Bonache, An
approach for the design of passive UHF-RFID tags mounted on optical
discs, IEEE Transactions on Antennas and Propagation, vol. 61, no.
12, pp. 5860-5867, Dec. 2013.
[9] Alien HiggsTM-3 datasheet, Available online at
www.alientechnology.com.
[10] G. Zamora, F. Paredes, F. J. Herraiz-Martnez, F. Martn, J. Bonache,
Bandwidth limitations of ultra high frequencyradio frequency
identification tags, IET Microw. Antennas Propag., vol. 7, no. 10, pp.
788794, July 2013.
[11] C. A. Grosvenor, R. T. Johnk, J. Baker-Jarvis, M. D. Janezic, and B.
Riddle, Time-domain free-field measurements of the relative
permittivity of building materials," IEEE Trans. on Instrum. Meas., Vol.
58, no. 7, pp. 2275-2282, July 2009.
[12] K. V. S. Rao, P. V. Nikitin, S. F. Lam, Antenna Design for UHF RFID
Tags: A Review and a Practical Application, IEEE Transactions on
Antennas and Propagation, vol. 53, no. 12, Dec. 2005.
2014 IEEE International Conference on RFID (IEEE RFID)
22