Main WRK PDF

INTERNATIONAL JOURNAL OF CIRCUIT THEORY AND APPLICATIONS
Int. J. Circ. Theor. Appl. 2017; 45:595–612

Published online 21 September 2016 in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/cta.2250
Fractional-order electronically controlled generalized filters
Georgia Tsirimokou1, Costas Psychalinos1,*,† and Ahmed S. Elwakil2

1
Electronics Laboratory, Physics Department, University of Patras, Rio Patras GR 26504, Greece
2
Department of Electrical and Computer Engineering, College of Engineering, University of Sharjah, Sharjah, P.O. Box
27272, United Arab Emirates
SUMMARY
Novel topologies of fractional-order generalized filters are introduced in this paper. These offer the follow-
ing benefits: (1) realization of lowpass, highpass, bandpass, allpass, or bandstop filter functions by the same
topology; (2) resistorless realizations; (3) electronic adjustment of their frequency characteristics as well as
their order; and (4) employment of only grounded capacitors. All the above have been achieved using
Operational Transconductance Amplifiers as active elements and appropriate multi-feedback topologies.
The behavior of the proposed designs is verified through simulation results using the Cadence IC design
suite and the Design Kit provided by the Austrian Micro Systems 0.35-μm complementary metal–
oxide–semiconductor process. Copyright © 2016 John Wiley & Sons, Ltd.
Received 25 January 2016; Revised 18 July 2016; Accepted 26 July 2016
KEY WORDS: fractional-order filters; fractional-order circuits; generalized filters; CMOS analog integrated
circuits
1. INTRODUCTION
The utilization of the fractional-order calculus, instead of the conventional integer-order calculus, in
electronic engineering offers attractive features [1]–[12]. For example, the magnitude responses of
these filters exhibit a stopband attenuation equal to 6(n + a) dB/oct, where n and a (0 < α < 1) are
the integer and fractional parts of the order of the filter. Compared with the 6n dB/oct attenuation
offered in the case of integer-order filters, fractional filters offer a more precise control of the
attenuation gradient [4].
As a result, significant research effort has been conducted for developing fractional-order filter
structures. Implementation of fractional-order filters in fully digital form [13]–[15] offers the
advantages of easy design, reliability, programmability, and better noise rejection in comparison
with the corresponding analog realizations. On the other hand, the digital implementation suffers
from the high power consumption associated with the required analog-to-digital (A/D) converter.
Analog realizations of fractional-order filters have been already introduced in discrete component
form in [4], [16]–[26]. The used active elements were Operational Amplifiers (op-amps), second-
generation Current Conveyors (CCIIs), and Current Feedback Operational Amplifiers (CFOAs).
Because of the employment of passive resistors, the realized time-constants have the form τ = RC,
and, consequently, an additional automatic tuning circuitry is required for compensating the
deviations from the desired frequency response. Another important drawback is the absence of
programmability, making these structures not capable of fulfilling the demand for realizing
*Correspondence to: Costas Psychalinos, Electronics Laboratory, Physics Department, University of Patras, Rio Patras
GR 26504, Greece.
†
E-mail: cpsychal@physics.upatras.gr
Copyright © 2016 John Wiley & Sons, Ltd.

596 G. TSIRIMOKOU, C. PSYCHALINOS AND A.S. ELWAKIL
programmable analog filters. Alternative fractional-order filters which do not suffer from the
aforementioned drawbacks have been introduced in [27]–[29].
All the above designs offer one type of filter function, and, therefore, the existence of filter
topologies which are capable for implementing various types of filter functions without modifying
their core is very important from the design flexibility point of view.
The contribution made in this paper is that, analog implementations of fractional-order generalized
filters are introduced for first time in the literature, where the filter functions are implemented without
changing the topology but through an electronic tuning of the corresponding bias current. It should be
mentioned at this point that, for demonstration purposes, only the standard filter functions will be
presented even though the proposed topologies are capable of realizing arbitrary fractional-order
filter functions. In addition, they are resistorless because of the employment of Operational
Transconductance Amplifiers (OTAs) as active elements. As a result, the designer has only to
choose the appropriate values of the dc bias currents in order to realize the desired transfer function,
and, therefore, the proposed schemes offer design flexibility and programmability. Another feature is
the employment of only grounded capacitors.
The paper is organized as follows: the characteristics of various fractional-order filter functions are
presented in Sections 2 and 3, while the corresponding OTA-C realizations are introduced in Section 4.
The performance of the proposed schemes is evaluated in Section 5, through the utilization of the
Cadence IC design suite and metal–oxide–semiconductor (MOS) transistors models provided by the
Austrian Micro Systems (AMS) 0.35-μm complementary MOS (CMOS) process.
2. GENERALIZED FILTERS OF ORDER a
2.1. Fractional-order lowpass filter (FLPF)

The transfer function of a FLPF with low-frequency gain K1/K2 and pole frequency ωp = K21/a is given
by
K1
H ðsÞ ¼ : (1)
sa þ K 2
The magnitude and phase responses are given respectively by
K1
jH ðjωÞj ¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ffi (2)
ω2a þ 2K 2 ωa cos aπ 2 þ K 2
2
!
1 ωa sin aπ
argfH ðjωÞg ¼ argfK 1 g tan 2 :
ωa cos aπ
2 þ K2
Note that when ω → 0 the phase is equal to arg{K1} while for ω → ∞ the phase will be equal to
(argfK 1 g aπ
2 ).
The half-power (3 dB) frequency (ωh), defined as the frequency where there is a 0.707 drop of the
low-frequency gain (i.e. K1/K2) is given by
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
aπ ffi aπ
1=a
ωh ¼ K 2 1 þ cos 2 cos (3)
2 2
with a corresponding phase angle of (3)
Copyright © 2016 John Wiley & Sons, Ltd. Int. J. Circ. Theor. Appl. 2017; 45:595–612
DOI: 10.1002/cta
FRACTIONAL-ORDER GENERALIZED FILTERS 597
0 1
aπ
B sin 2 C
argfH ðjωÞgjω¼ωh ¼ argfK 1 g tan1 @ aπ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
aπffiA:
2cos 2 þ 1 þ cos2 2
2.2. Fractional-order highpass filter (FHPF)

The transfer function of a FHPF is given by (4), while its magnitude and phase responses are those
in (5)
K 1 sa
H ðsÞ ¼ (4)
sa þ K2
K 1 ωa
jH ðjωÞj ¼ rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
aπ ffi
ω2a þ 2K 2 ωa cos þ K 22
2
0 aπ 1 (5)
ω a
sin
aπ B C
argfH ðjωÞg ¼ argfK 1 g þ tan1 @ aπ 2 A:
2 ω cos
a þ K2
2
Note that when ω → 0 the phase is equal to (argfK 1 g þ aπ
2 ), while for ω → ∞ the phase will be equal
to arg{K1}. The high-frequency gain is K1/K2 and the corresponding expressions for ωh and the phase
at this frequency are given by
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
aπ ffi aπ
1=a
ωh ¼ K 2 1 þ cos 2 þ cos
2 2
0 aπ 1
(6)
aπ B sin C
argfH ðjωÞgjω¼ωh ¼ argfK 1 g þ tan1 B @ r 2
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
aπ ffiA:
C
2
1 þ cos 2
2
2.3. Fractional-order allpass filter (FAPF)

An FAPF has the transfer function in (7)
sa K 2
H ð sÞ ¼ K 1 (7)
sa þ K 2
and its frequency behavior is described by (8)
vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
aπ ffi
u 2a
uω 2K 2 ωa cos þ K 2
u 2 2
jH ðjωÞj ¼ K 1 t aπ
ω þ 2K 2 ω cos
2a a þ K 22
2
0 aπ 1 0 aπ 1 (8)
ωa sin ωa sin
B aπ 2 C B aπ 2 C
argfH ðjωÞg ¼ argfK 1 g tan1 @ A tan1 @ A:
ωa cos K2 ωa cos þ K2
2 2
Note that when ω → 0 the phase is equal to (arg{K1} + π), while for ω → ∞ the phase is equal to arg
{K1}. For a < 1, the magnitude response has a minimum equal to K1tan(aπ/4) occurred at ωo = K21/a.
The phase at this frequency is equal to (arg{K1} + π/2). In addition, when ω → 0 or ∞ the gain is equal
to K1.
DOI: 10.1002/cta
2.4. Fractional-order bandpass filter (FBPF)

The transfer function of a FBP is:
K 1 sβ
H ðsÞ ¼ : (9)
sa þ K 2
Using (9), it is obtained that
K 1 ωβ
jH ðjωÞj ¼ rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
aπ ffi
ω þ 2K 2 ω cos
2a a þ K2 2
2
aπ 1 0 (10)
βπ ω a
sin
B C
argfH ðjωÞg ¼ argfK 1 g þ tan1 @ aπ 2 A:
2 ω cos
a þ K2
2
In order to obtain a FBPF response, α > β should be fulfilled. Note that when ω → 0 the phase equals
(argfK 1 g þ π2 ), while for ω → ∞ the phase equals (argfK 1 g þ ðβa Þπ
2 ). The peak frequency (ωpeak),
d
calculated using the condition: dω jH ðjωÞjω¼ωp ¼ 0, is given by
8 h qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
i91=a
< cos aπ 2 ð 2β a Þ þ a2 þ 4βðα βÞtan2 aπ 2
=
ωpeak ¼ K 2 : (11)
: 2 ðα β Þ ;
In the case that α = 2β, then ωpeak = ωo.

The magnitude response at ω = ωp = K1/a, is equal to K1 1
2cosðaπ
, and the phase at this frequency is
4Þ
2 1ðβ=aÞ
K2
equal to (argfK 1 g þ ð2βa

4
Þπ
). In addition, when ω → ∞ the gain is equal to K1ωβ a.
2.5. Design equations for generalized filters of order a

Because of the absence of commercially available fractional-order elements [30]–[32], an alternative way
for implementing a fractional-order transfer function is its approximation by an integer-order function.
There are several methods for obtaining rational approximations of the variable sq, such as the Carlson’s
method, Matsuda’s method, Oustaloup’s method, and Continuous Fraction Expansion (CFE) method
[33]-[40]. In this paper the second-order approximation of the CFE, given by (12) will be employed.
ða2 þ 3a þ 2Þs2 þ ð8 2a2 Þs þ ða2 3a þ 2Þ a2 s2 þ a1 s þ ao
sa ≅ ≡
ða2 3a þ 2Þs2 þ ð8 2a2 Þs þ ða2 þ 3a þ 2Þ b2 s2 þ b1 s þ bo
2 (12)
β
β þ 3β þ 2 s2 þ 8 2β2 s þ β2 3β þ 2 A2 s2 þ A1 s þ Ao
s ≅ 2 ≡
β 3β þ 2 s2 þ 8 2β2 s þ β2 þ 3β þ 2 B2 s2 þ B1 s þ Bo
Substituting with (12) into (1), (4), (7), and (9), the derived transfer functions are, respectively,
K1 b2 s2 þ b1 s þ bo
H LP ðsÞ ¼
a2 þ K 2 b2 D ðsÞ
K1 a2 s2 þ a1 s þ ao
H HP ðsÞ ¼

a1 K 2 b1 ao K 2 bo
s2 þ s þ (13)
a2 K 2 b2 a2 K 2 b2 a2 K 2 b2
H AP ðsÞ ¼ K 1
2 3

K1 b2 s2 þ b1 s þ bo 6 1 A2 s2 þ A1 s þ Ao 7
H BP ðsÞ ¼ 6
4 7
ða2 þ K 2 b2 Þ D ðsÞ B2 B1 Bo 5
s2 þ sþ
B2 B2
where D(s) is given
DOI: 10.1002/cta

a1 þ K 2 b1 ao þ K 2 bo
DðsÞ ¼ s þ 2
s þ : (14)
a2 þ K 2 b2 a2 þ K 2 b2
Comparing the transfer functions of the FLPF and FHPF, it is easily derived that the numerator in
HHP(s) is derived through the substitution: bj → aj (j = 0, 1, 2) in the numerator of HLP(s).
Inspecting the transfer functions of FLPF, FHPF, and FAPF, it is concluded that all of them have the
same form. Consequently, they could be realized by the same topology just by changing the coefficient
values. A suitable solution for this purpose is the Functional Block Diagram (FBD) of the inverse
follow-the-leader multi-feedback topology (IFLF) depicted in Figure 1.
The realized transfer function using this topology is
G2 s2 þ Gτ11 s þ τG1 τo2

H ðsÞ ¼ : (15)
s2 þ τ11 s þ τ11τ2
Comparing the coefficients of (15) with those of FLPF, FHPF, and FAPF the design equations of all
the filters are
a2 þ K 2 b2 a1 þ K 2 b1
τ1 ¼ and τ 2 ¼ : (16)
a1 þ K 2 b1 ao þ K 2 bo
The corresponding design equation for the scaling factors Gi(i = 0,1,2) are summarized in Table I.
With regards to the FBPF realization, a possible solution is the FBD in Figure 2, where H1(s) and
H2(s) are the filter blocks in Figure 1, which realize the transfer function in (15).
The expressions for time-constants of the stage #1 are as given by (16), while for stage #2 they are
B2 B1
τ 12 ¼ and τ 22 ¼ : (17)
B1 Bo
The corresponding design equation for the scaling factors Gij(i = 0,1,2 and j = 1,2) of both stages are
also summarized in Table I.
Figure 1. FBD for realizing FLPF, FHPF, and FAPF of order a. [Colour figure can be viewed at
wileyonlinelibrary.com]
Table I. Values of scaling factors for realizing the FBDs in Figures 1 and 2.
Filter G2 G1 G0
b2 b1 bo
FLPF K 1 a2 þK 2 b2
K 1 a1 þK 2 b1
K 1 ao þK 2 bo
a2 a1 ao
FHPF K 1 a2 þK 2 b2
K 1 a1 þK 2 b1
K 1 ao þK 2 bo
FAPF K 1 aa22 K 2 b2
þK 2 b2 K 1 aa11 K 2 b1
þK 2 b1 K 1 aaoo K 2 bo
þK 2 bo
b2 b1 bo
FBPF K 1 a2 þK 2 b2
K 1 a2 þK 2 b1
K 1 ao þK 2 bo
A2 A1 Ao
B2 B1 Bo
DOI: 10.1002/cta
Figure 2. FBD for realizing FBPF of order a. [Colour figure can be viewed at wileyonlinelibrary.com]
3. GENERALIZED FILTERS OF ORDER a + β
3.1. Fractional-order lowpass filter (FLPF)

The transfer function of this FLPF and its magnitude and phase responses are given by (18) and (19),
respectively,
K1
H LP ðsÞ ¼
sαþβ þ K 3 sβ þ K 2
K1 (18)
jH ðjωÞj ¼ sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

aπ ða þ βÞπ βπ
ω2ðaþβÞ þ 2K 3 ωaþ2β cos þ 2K 2 ωaþβ cos þ K 23 ω2β þ 2K 2 K 3 ωβ cos þ K 22
2 2 2
h i 0 1
ωaþβ sin ðaþβ Þπ
þ K 3 ωβ sin βπ
argfH ðjωÞg ¼ argfK 1 g tan1 @ A:
2 2
h i βπ (19)
ðaþβ Þπ β
ω cos 2
aþβ þ K 3 ω cos 2 þ K 2
The half-power frequency can be found as the solution to the following equation, derived taking into
account that the maximum gain of the filter is K1/K2,
aπ
ða þ βÞπ βπ
ωh 2ðaþβÞ þ 2K 3 ωh aþ2β cos þ 2K 2 ωh aþβ cos þ K 23 ωh 2β þ 2K 2 K 3 ωh β cos K 22 ¼ 0: (20)
2 2 2
3.2. Fractional-order highpass filter (FHPF)

In a similar way, the transfer function of a FHPF, with maximum gain equal to K1/K2, is
sαþβ
H HP ðsÞ ¼ K 1 : (21)
The responses of this filter are described by
K 1 ωaþβ

2 2 2
0 1 (22)
ða þ βÞπ βπ
ωaþβ sin þ K 3 ωβ sin
ða þ βÞπ 1 B
B 2 2 C
argfH ðjωÞg ¼ argfK 1 g þ tan @ C
2 ð a þ β Þπ βπ A
ω cos
aþβ þ K 3 ω cos β þ K2
2 2
and the half-power frequency is calculated by solving (23)

aπ
ða þ βÞπ βπ
ωh 2ðaþβÞ 2K 3 ωh aþ2β cos 2K 2 ωh aþβ cos K 23 ωh 2β 2K 2 K 3 ωh β cos K 22 ¼ 0: (23)
2 2 2
DOI: 10.1002/cta
3.3. Fractional-order bandpass filter (FBPF)

The transfer function of a FBPF is
K 3 sβ
H BP ðsÞ ¼ K 1 (24)
with magnitude and phase responses given given by (25)
K 1 K 3 ωβ

2 2 2
0 1 (25)
ða þ βÞπ βπ
ωaþβ sin þ K 3 ωβ sin
βπ B 2 2 C
argfH ðjωÞg ¼ argfK 1 K 3 g þ tan1 B @ C:
A
2 ð a þ β Þπ βπ
ω cos
aþβ þ K 3 ω cos β þ K2
2 2
The peak frequency (ωp) is calculated solving the equation in (26), which is derived from (25)
d
applying the condition: dω jH ðjωÞjω¼ωp ¼ 0.
aπ
ða þ βÞπ βπ
aωp 2ðaþβÞ þ aK 3 ωp aþ2β cos þ ða βÞK 2 ωp aþβ cos βK 2 K 3 ωp β cos βK 22 ¼ 0 (26)
2 2 2
jH ðjωÞjω¼ωp
while the half-power frequencies are calculated from the condition jH ðjωÞjω¼ωh ¼ pffiffi .
2
The quality factor (Q) of the filter is then given by
ωp
Q¼ (27)
ωh2 ωh1
where ωh2 and ωh1 are the upper and lower half-power frequencies, respectively.
3.4. Fractional bandstop filter (FBSF)

Finally, the transfer function of a FBSF is that given in (28)
sαþβ þ K 2
H BS ðsÞ ¼ K 1 : (28)
The magnitude and phase responses are these in (29)
vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

u ða þ βÞπ
u
u ω2ðaþβÞ þ 2K 2 ωaþβ cos þ K 22
u 2
jH ðjωÞj ¼ K 1 u aπ
t ða þ βÞπ βπ
2 2 2
0 1 0 1 (29)
ða þ βÞπ ða þ βÞπ βπ
ωaþβ sin ωaþβ sin þ K 3 ωβ sin
B 2 C B 2 2 C
argfH ðjωÞg ¼ argfK 1 g þ tan1 B @ C tan1 B
A @ C:
A
ða þ βÞπ ða þ βÞπ β
βπ
ω cos
aþβ þ K2 ω cos
aþβ þ K 3 ω cos þ K2
2 2 2
d
The peak frequency (ωp) is calculated from (29) under the condition: dω jH ðjωÞjω¼ωp ¼ 0, while the
half-power frequencies are calculated using the fact that at these frequencies there is a 0.707 drop of
passband gain which is equal to K1. The quality factor of the filter (Q) is calculated from (27).
DOI: 10.1002/cta
3.5. Design equations for generalized filters of order a + β

The design equation will be derived following a similar procedure to that in the case of the filters in
Section 2. Thus, using the expressions in (12) and the definitions in Table II, the derived
expressions for the transfer functions of fractional-order filters are summarized in (30):
K 1 N ð sÞ
H LP ðsÞ ¼
D 4 D ðsÞ
K 1 N ðsÞ
H HP ðsÞ ¼
D4 DðsÞ
K 3 N ** ðsÞ
H BP ðsÞ ¼ K 1 (30)
D 4 D ðsÞ
K 1 N ðsÞ þ ðK 2 b2 A2 Þs4 þ ðK 2 b2 B1 þ K 2 b1 B2 Þs3
H BS ðsÞ ¼
D4 D ðsÞ
ðK 2 b2 Bo þ K 2 b1 B1 þ K 2 bo B2 Þs2 þ ðK 2 b1 Bo þ K 2 bo B1 Þs þ K 2 bo Bo
þ
D ðsÞ
The expressions for polynomials N(s) and D(s) in (30) are given by (31) and (32), respectively,
N ðsÞ ≡N 4 s4 þ N 3 s3 þ N 2 s2 þ N 1 s þ N o (31)
D3 3 D2 2 D1 Do
DðsÞ ≡s4 þ s þ s þ s þ (32)
D4 D4 D4 D4
while the polynomials N*(s) and N**(s) in (30) are derived from the corresponding expression in (31)–
(32) by performing the following substitutions: bj → aj, Bj → Aj, (j = 0,..,4) and Bj → Aj, (j = 0,..,4),
respectively.
The transfer functions in (30) all have the same form, and, consequently, they could be realized by
the same topology just by changing the coefficient values. The corresponding FBD of the IFLF filter is
that depicted in Figure 3.
Table II. Definitions of the variables used in the expression in (30).
Variable Expression
Do ao Ao + K3 bo Ao + K2 bo Bo
D1 a1 Ao + ao A1 + K3 b1 Ao + K3 bo A1 + K2 b1 Bo + K2 bo B1
a2 Ao þ a2 A1 þ ao A2 þ K 3 b2 Ao þ K 3 b1 A1 þ K 3 bo A2
D2
þ K 2 b2 Bo þ K 2 b1 B1 þ K 2 bo B2
D3 a2 A1 + a1 A2 + K3 b2 A1 + K3 b1 A2 + K2 b2 B1 + K2 b1 B2
D4 a2 A2 + K3 b2 A2 + K2 b2 B2
No bo Bo
N1 b1 Bo + bo B1
N2 b2 Bo + b1 B1 + bo B2
N3 b2 B1 + b1 B2
N4 b2 B2
Figure 3. FBD for realizing FLPF, FHPF, FBPF, and FBSF of order a + β. [Colour figure can be viewed at
DOI: 10.1002/cta
The realized transfer function of this FBD is that in (33)
G4 s4 þ Gτ 13 s3 þ τG1 τ22 s2 þ τ1 τG21τ3 s þ τ 1 τG2 τo3 τ4

H ðsÞ ¼ : (33)
s4 þ τ11 s3 þ τ11τ2 s2 þ τ 1 τ12 τ3 s þ τ 1 τ21τ3 τ4
Comparing the coefficients in (33) with those in (30), the derived expressions for time-constants are
D4 D3 D2 D1
τ1 ¼ ; τ2 ¼ ; τ3 ¼ ; and τ 4 ¼ : (34)
D3 D2 D1 Do
The corresponding equations for the scaling factors Gi(i = 0,1,2,3,4) are summarized in Table III.
Having available the design equations in (16), (17), and (34) as well as those in Tables I and III
the design of various types of filters is facilitated.
4. REALIZATION OF FRACTIONAL-ORDER GENERALIZED FILTERS USING OTAS
The topology for realizing FLPF, FHP, and FAPFs of order a (i.e. the FBD in Figure 1), using OTAs as
active elements, is demonstrated in Figure 4. The lower left OTA is a multiple output element, where
the non-inverting output is used for realizing the FLPF and FHPF functions, while the inverting output
is used for FAPF function. This modification could be easily performed through an appropriate
switching scheme. The corresponding OTA-C filter suitable for implementing the FBD in Figure 3
is demonstrated in Figure 5.
The features of both topologies are the following:
a. The filter functions are realized by the same core, and, therefore, they can be considered as
generalized filter structures.
Table III. Values of scaling factors for realizing the FBD in Figure 3.
Filter G4 G3 G2 G1 Go
FLPF K 1 ND44 K 1 ND33 K 1 ND22 K 1 ND11 K 1 NDoo

N N N N N
FHPF K 1 D44 K 1 D33 K 1 D22 K 1 D11 K 1 Doo
N N N N N
FBPF K1 D4
4
K1 D3
3
K1 D2
2
K1 D1
1
K1 Do
o
N 4 þK 2 b2 A2 N 3 þK 2 b2 B1 þK 2 b1 B2 N 2 þK 2 b2 Bo þK 2 b1 B1 þK 2 bo B2 N 1 þK 2 b1 Bo þK 2 bo B1 N o þK 2 bo Bo
FBSF K1 D4 K1 D3 K1 D2 K1 D1 K1 Do
Figure 4. OTA-C realization of FLPF, FHPF, and FAPF functions of order a.
DOI: 10.1002/cta
Figure 5. OTA-C realization of FLPF, FHPF, FBPF, and FBSF functions of order a + β.
b. They offer resistorless realizations because of the employment of the transconductance parameter
(gm) of OTAs for realizing the required time-constant as well as for scaling factors.
c. Their frequency characteristics and fractional order could be electronically adjusted. This origi-
nates from the fact that the transconductance gm is controlled by an appropriate bias current.
d. Only grounded capacitors are utilized.
A suitable OTA structure is demonstrated in Figure 6. Considering that the MOS transistors operate
in the subthreshold region and the aspect ratio of Mn1 and Mn3 is A times the aspect ratio of Mn2 and
Mn4, the realized transconductance is given by the expression
4A
g m ¼ gmo : (35)
ð1 þ A Þ2
Figure 6. OTA structure with enhanced linearity.
Table IV. Bias currents and scaling factors for fractional filters of order a.
fo = 100 Hz C = 50 pF LP (a = 0.5) HP (a = 0.5) AP (α = 0.5) BP (α = 1, β = 0.5)
Io 6.37 nA 6.37 nA 6.37 nA 3.82 nA 19.1 nA
Io1 6.37 nA 6.37 nA 6.37 nA 3.82 nA 19.1 nA
Io2 573.4 pA 573.4 pA 573.4 pA 955 pA 955 pA
G2 0.167 0.833 0.667 0 5
G1 0.5 0.5 0 0.5 1
Go 0.833 0.167 0.667 1 0.2
DOI: 10.1002/cta
Figure 7. Estimation of the accuracy of the employed approximation in the case of a LP filter (a) magnitude
response, (b) phase response, (c) relative error of the magnitude and phase. [Colour figure can be viewed at
DOI: 10.1002/cta
In (35), gmo is the transconductance of OTA in the case that A = 1. For example, gmo = Io/nVT., where
VT is the thermal voltage (26 mV@ 27 °C), Io the bias current, and n the subthreshold slope factor of an
MOS transistor. A benefit of the OTA in Figure 6 is its enhanced linearity [41].
5. SIMULATION RESULTS
The performance of the proposed generalized filter structures will be evaluated using the Analog
Design Environment of the Cadence software and the Design Kit provided by the AMS 0.35-μm
CMOS process. The employed power supply voltages were VDD = 1.5 V and VCM = 0.75 V. The
aspect ratios of the transistors of the OTA in Figure 6 (A = 5) were 60 μm/10 μm for Mb1–Mb3,
25 μm/1 μm for Mn1, Mn4, 5 μm/1 μm for Mn2, Mn3, and 10 μm/15 μm for Mp1–Mp2.
Figure 8. Post-layout simulation of the frequency responses of the filter topology in Figure 4 (a) magnitude,
and (b) phase (the corresponding theoretically predicted responses are given by dashed lines). [Colour figure
can be viewed at wileyonlinelibrary.com]
DOI: 10.1002/cta
As a first design example, the filter topology in Figure 4 will be emulated. Considering that the pole
frequency is fo = 100 Hz and both capacitors have the same value of 50 pF, the values of the dc bias
currents as well as scaling factors for the topology in Figure 3 are summarized in Table IV. The
implementation of the condition G2 = 0 is performed by omitting the rightmost OTA as well as the
OTA biased at current G2Io in Figure 4; the output of the filter is derived at the output of the remaining
rightmost OTA. Also, the condition G1 = 0 is performed by omitting the OTA biased at current G1Io.
In order to estimate the accuracy of the approximation in (12), the derived post-layout magnitude
and phase responses of a LP filter (a = 0.5) are demonstrated in the plots in Figures 7a–b,
respectively. In addition, the corresponding plots of the theoretically predicted responses and those
derived from the original fractional-order filter are also provided. In order to facilitate the reader, the
corresponding relative error plots (i.e. referred to the original filter response) are supplied in
Figure 7c. According to these results, the employed approximation offers accuracy better than 10%
in both magnitude and phase within the range [fo/8, 8fo], where fo, is the pole frequency. This range
could be further enhanced using a higher-order approximation instead of the second-order
approximation employed in this paper at the expense of circuit complexity.
Following the above discussion, the derived magnitude and phase frequency responses are given in the
plots of Figure 8, where the dashed lines corresponding to theoretically predicted responses are also provided.
As a second design example, the behavior of fractional filters of order a + β will be evaluated,
utilizing the same bias conditions and aspect ratios for OTAs as in the previous case. Considering
that fo = 100 Hz and value 50 pF for capacitors, the values of dc bias currents for the filters are
summarized in Table V. The realization of the zero scaling factors is achieved following the same
procedure as in the previous design example.
In order to have evidence about the realistic behavior of the filter, taking into account the parasitic
capacitance and resistances appearing at the nodes of the topology, its layout design has been performed.
The obtained frequency responses with the corresponding theoretically predicted are depicted in Figure 8.
The electronic tuning capability of the proposed generalized fractional-order filters is
demonstrated in the case of an a + β order FBPF. The corresponding values of bias currents and
scaling factors for tuning a and β are given in Table VI, while the obtained frequency responses
Table V. Bias currents and scaling factors for fractional filters of order a + β.
fo = 100 Hz C = 50 pF LP (a = 1, β = 0.5) HP (α = 0.5, β = 1) BS (a = 1, β = 0.5) BP (a = 1, β = 0.5)
Io 8.03 nA 8.6 nA 8.03 nA 8.03 nA
Io1 8.03 nA 8.6 nA 8.03 nA 8.03 nA
Io2 3.37 nA 2.26 nA 3.18 nA 3.37 nA
Io3 1.39 nA 1.25 nA 1.47 nA 1.39 nA
Io4 424.7 pA 455 pA 424.7 pA 424.7 pA
G4 0 0.8333 1 0
G3 0.0476 0.5556 0.7619 0.2381
G2 0.2973 0.3438 0.5946 0.4054
G1 0.5556 0.0476 0.5926 0.4074
Go 0.8333 0 0.8333 0.1667
Table VI. Bias currents and scaling factors for tuning the order a and β of an FBPF.
fo = 100 Hz C = 50 pF BP (a = 0.7, β = 0.5) BP (α = 0.8, β = 0.7) BP (a = 0.9, β = 0.9)
Io 9.89 nA 8.1 nA 6.63 nA

Io1 9.89 nA 8.1 nA 6.63 nA
Io2 3.25 nA 2.95 nA 2.65 nA
Io3 1.24 nA 1.30 nA 1.37 nA
Io4 368.6 pA 443.7 pA 550.4 pA
G4 0.0771 0.0453 0.0196
G3 0.2979 0.3156 0.3333
G2 0.4200 0.4425 0.4759
G1 0.3654 0.3501 0.3333
Go 0.1643 0.0780 0.0196
DOI: 10.1002/cta
are demonstrated in Figure 9. Using these plots, the obtained frequency characteristics of the filters
are summarized in Table VII, where the corresponding theoretically predicted values are given in
parentheses.
(a)
(b)
Figure 9. Post-layout simulation results for demonstrating the electronic tuning capability of the order a + β
of the FBPF through (a) magnitude, and (b) phase response (the corresponding theoretically predicted re-
sponses are given by dashed lines). [Colour figure can be viewed at wileyonlinelibrary.com]
Table VII. Frequency characteristics of the FBPFs.

FBPF a = 0.7 β = 0.5 a = 0.8 β = 0.7 a = 0.9 β = 0.9 a = 1 β = 0.5
fpeak (Hz) 104.7 (100) 104.7 (100) 104.7 (100) 106.8 (100)
Gain(max) (dB) 5.6 (5.7) 4.1 (4.2) 1.8 (1.9) –4.7 (4.9)
fh1 (Hz) 30.3(28.5) 42.19 (39.95) 55.78 (53.13) 33.03 (31.01)
fh2 (Hz) 304.4 (293.8) 250.6 (242.7) 194 (188.2) 257.1 (248.7)
Slope at upper stopband (dB/oct) 3.96 (4) 4.9 (5) 5.8 (5.8) 5.9 (5.9)
Slope at lower stopband (dB/oct) 3 (3) 4.5 (4.4) 5.9 (5.8) 3 (3)
DOI: 10.1002/cta
Table VIII. Bias currents and scaling factors for tuning the peak frequency of the FBPF (a = 1, β = 0.5).
Bias current fo = 10 Hz fo = 100 Hz fo = 1 kHz
Io 802.8pA 8.03nA 80.28nA
Io1 802.8pA 8.03nA 80.28nA
Io2 336.8pA 3.37nA 33.68nA
Io3 139.5pA 1.39nA 13.95nA
Io4 42.47pA 424.7pA 4.25nA
G4 0 0 0
G3 0.2381 0.2381 0.2381
G2 0.4054 0.4054 0.4054
G1 0.4074 0.4074 0.4074
Go 0.1667 0.1667 0.1667
(a)
(b)
Figure 10. Post-layout simulation results for demonstrating the electronic tuning capability of the peak
frequency of the FBPF through (a) magnitude, and (b) phase responses (the corresponding theoretically
predicted responses are given by dashed lines). [Colour figure can be viewed at wileyonlinelibrary.com]
DOI: 10.1002/cta
The electronic tuning capability of the peak frequency of the FBPF is achieved through the
utilization of Table VIII. The obtained responses are plotted in Figure 10, while their most important
characteristics are summarized in Table IX.
The sensitivity performance of the FBPF with regards to the effect of MOS transistor mismatching
as well as process parameters variations has been evaluated using the Monte-Carlo analysis (N = 100
runs). The derived statistical plot for a FBPF (a = 1, β = 0.5, and fo = 100 Hz) about maximum gain,
peak frequency, bandwidth, and slopes in the lower and upper stopband attenuations are
simultaneously given in Figure 11. The values of the standard deviations were 0.65 dB, 20.6 Hz,
29.2 dB, 0.34 dB/oct, and 0.61 dB/oct, respectively, confirming that the proposed topologies have
reasonable sensitivity performance characteristics.
Table IX. Frequency characteristics of the FBPF with tuned peak frequency (a = 1, β = 0.5).
FBPF fo = 10 Hz fo = 100 Hz fo = 1 kHz
fpeak (Hz) 10.96 (10.47) 109.6 (104.7) 1.09k (1.05k)
Gain(max) (dB) 4.9(4.9) 4.74 (4.86) 4.52 (4.86)
fh1 (Hz) 3.25(3.1) 33 (31) 331.3 (310.1)
fh2 (Hz) 26.2 (24.9) 257.1(248.7) 2.45 k (2.49 k)
Slope at upper stopband (dB/oct) 5.8 (5.9) 5.9 (5.9) 6 (5.9)
Slope at lower stopband (dB/oct) 3.35 (3) 3 (3) 3 (3)
Figure 11. Post layout Monte-Carlo simulation results about (a) maximum gain, (b) peak frequency, (c)
bandwidth, (d) slope in the lower stopband, and (e) slope in the upper stopband of the FBPF (a = 1,
β = 0.5). [Colour figure can be viewed at wileyonlinelibrary.com]
DOI: 10.1002/cta
6. CONCLUSION
The proposed fractional-order generalized filters fulfill the demand for realizing low-voltage analog
filters with only grounded capacitors and electronic adjustment of their frequency characteristics.
Another important feature is the electronic tuning capability of the order of the filters. The provided
simulation results for orders 0.5, 1.2, 15, and 1.8 confirmed their correct behavior for various types
of filter functions. Taking into account that their performance has been evaluated at post-layout level
and that the obtained Monte-Carlo results show that the filters have reasonable sensitivity
characteristics, they could be considered as attractive candidates in practical applications where high
performance analog signal processing stages are required.
ACKNOWLEDGEMENTS
This work was supported by Grant E.029 from the Research Committee of the University of Patras
(Programme K. Karatheodori).
REFERENCES
1. Elwakil A. Fractional-order circuits and systems: an emerging interdisciplinary research area. IEEE Circuits and
Systems Magazine 2010; 10(4):40–50.
2. Podlubny I, Petráš I, Vinagre BM, O’Leary P, Dorčák Ľ. Analogue realizations of fractional-order controllers.
Nonlinear Dynamics 2002; 29(1–4):281–296.
3. Dorčák L’, Valsa J, Gonzalez E, Terpák J, Petráš I, Pivka L. Analogue realization of fractional-order dynamical
systems. Entropy 2013; 15(10):4199–4214.
4. Freeborn T, Maundy B, Elwakil A. Field programmable analogue array implementation of fractional step filters. IET
Circuits, Devices and Systems 2010; 4(6):514–524.
5. Tepljakov A, Gonzalez E, Petlenkov E, Belikov J,, Monje C, Petráš I. Incorporation of fractional-order dynamics into
an existing PI/PID DC motor control loop. ISA Transactions Jan. 2016; 60:262–273.
6. Das S, Pan I. Fractional order signal processing: introductory concepts and applications. In Springer Briefs in Applied
Sciences and Technology. : Springer, 2012. ISBN:978-3-642-23117-9.
7. G. W. Bohannan, “Analog realization of a fractional control element—revisited”, in Proceedings of the 41st IEEE
International Conference on Decision and Control, Tutorial Workshop 2: Fractional Calculus Applications in
Automatic Control and Robotics, Las Vegas, USA, Dec. 2002.
8. Bohannan GW. Analog fractional order controller in temperature and motor control applications. Journal of
Vibration and Control Sept. 2008; 14(9–10):1487–1498.
9. Caponetto R, Dongola G, Fortuna L, Gallo A. Non integer order operators implementation via switched capacitors
technology in the book titled. In New Trends in Nanotechnology and Fractional Calculus Applications, Baleanu
D, Guvenc Z, Tenreiro Machado JA (eds) ch.7 Editors. Springer, 2010; 87–96.
10. Varshney P, Gupta M, Visweswaran GS. New switched capacitor fractional order integrator. Journal of Active and
Passive Electronic Devices 2007; 2:187–197.
11. Sheng H, Chen Y, Qui T. Fractional processes and fractional-order signal processing—techniques and applications.
In Signals and Communication Technology. Springer, 2012. ISBN:978-1-4471-2233-3.
12. Li M. Approximating ideal filters by systems of fractional order. Computational and Mathematical Methods in
Medicine 2012; 2012 Article ID 375858. doi:10.1155/2012/365054.
13. Chen YQ, Moore KL. Discretization schemes for fractional-order differentiators and integrators. IEEE Transactions
on Circuits and Systems–I: Fundamental Theory and Applications 2002; 49(3):363–367.
14. Tseng CC, Lee SL. Design of fractional order digital differentiator using radial basis function. IEEE Transactions on
Circuits and Systems I: Regular Papers 2010; 57(7):1708–1718.
15. Das S, Pan I. Fractional-order signal processing: introductory concepts and applications. Springer Briefs in Applied
Sciences and Technology Chapter 2012; 2:13–30.
16. Radwan A, Soliman A, Elwakil A. First-order filters generalized to the fractional domain. Journal of Circuits,
Systems, and Computers 2008; 17(1):55–66.
17. Radwan A, Elwakil A, Soliman A. On the generalization of second-order filters to the fractional-order domain.
Journal of Circuits, Systems and Computers 2009; 18(2):361–386.
18. Maundy B, Elwakil A, Freeborn T. On the practical realization of higher-order filters with fractional stepping. Signal
Processing 2011; 91(3):484–491.
19. Ahmadi P, Maundy B, Elwakil A, Belostotski L. High-quality factor asymmetric-slope band-pass filters: a fractional-
order capacitor approach. IET Circuits, Devices and Systems 2012; 6(3):187–197.
20. Soltan A, Radwan A, Soliman A. Fractional-order filter with two fractional elements of dependant orders. Microelec-
tronics Journal 2012; 43(11):818–827.
21. Radwan A, Salama K. Fractional-order RC and RL circuits. Circuits Systems and Signal Processing 2012;
31(6):1901–1915.
DOI: 10.1002/cta
22. Radwan A. Resonance and quality factor of the RLaCa fractional circuit. IEEE Journal on Emerging and Selected
Topics in Circuits and Systems 2013; 3(3):377–385.
23. Ali A, Radwan A, Soliman A. Fractional-order Butterworth filter: active and passive realizations. IEEE Journal on
Emerging and Selected Topics in Circuits and Systems 2013; 3(3):346–354.
24. Tripathy M, Biswas K, Sen S. A design example of a fractional-order Kerwin–Huelsman–Newcomb biquad filter
with two fractional capacitors of different order. Circuits Systems and Signal Processing 2013; 32(4):1523–1536.
25. Soltan A, Radwan AG, Soliman AM. Fractional order Sallen–Key and KHN filters: stability and poles allocation.
Circuits Systems and Signal Processing 2015; 34(5):1461–1480.
26. Tripathy M, Mondal D, Biswas K, Sen S. Experimental studies on realization of fractional inductors and fractional-
order bandpass filters. International Journal of Circuit Theory and Applications 2015; 43(9):1183–1195.
27. G. Tsirimokou, and C. Psychalinos, “Ultra-low voltage fractional-order differentiator and integrator topologies: an
application for handling noisy ECGs”, Analog Integrated Circuits and Signal Processing, vo. 81, 2, pp. 393-405,
2014.
28. Tsirimokou G, Laoudias C, Psychalinos C. 0.5 V fractional-order companding filters. International Journal of Circuit
Theory and Applications 2015; 43(9):1105–1126.
29. Tsirimokou G, Psychalinos C. Ultra-low voltage fractional-order circuits using current-mirrors. International Journal
of Circuit Theory and Applications, International Journal of Circuit Theory and Applications 2016; 44(1):109–126.
30. Biswas K, Sen S, Dutta P. Realization of a constant phase element and its performance study in a differentiator cir-
cuit. IEEE Transactions on Circuits and Systems-II: Express Briefs 2006; 53(9):802–806.
31. Krishna MS, Das S, Biswas K, Goswami B. Fabrication of a fractional-order capacitor with desired specifi-
cations: a study on process identification and characterization. IEEE Transactions on Electron Devices 2011;
58(11):4067–4073.
32. Elshurafa AM, Almadhoun MN, Salama KN, Alshareef HN. Microscale electrostatic fractional capacitors using
reduced graphene oxide percolated polymer composites. Applied Physics Letters 2013; 102(23):232901.
33. Carlson GE, Halijak CA. Approximation of fractional capacitors (1/s)^(1/n) by a regular Newton process. IEEE
Transactions on Circuit Theory 1964; 11(2):210–213.
34. Matsuda K, Fujii H. H∞–optimized wave-absorbing control: analytical and experimental results. Journal of
Guidance, Control, and Dynamics 1993; 16(6):1146–1153.
35. Oustaloup A. Systèmes Asservis Linéaires d’Ordre Fractionnaire: Théorie et Pratique”. Masson: Paris, 1983.
36. Oustaloup A. La Dérivation Non Entière, Théorie, Synthèse et Applications. Hermes Science Publications: Paris,
1995.
37. Oustaloup A, Levron F, Mathieu B, Nanot F. Frequency-band complex non integer differentiator: characteriza-
tion and synthesis. IEEE Transactions on Circuits and Systems – I: Fundamental Theory and Applications
2000; 47(1):25–39.
38. Djouambi A, Charef A, Besançon A. Optimal approximation, simulation and analog realization of the fundamen-
tal fractional order transfer function. International Journal of Applied Mathematics and Computer Science 2007;
17(4):455–462.
39. Gruau C, Picart D, Belmas R, Bouton E, Delmaire-Size F, Sabatier J, Trumel H. Ignition of a confined high explosive
under low velocity impact. International Journal of Impact Engineering 2009; 36(4):537–550.
40. Krishna B, Reddy K. Active and passive realization of fractance device of order ½. Active and Passive Electronic
Components 2008; 2008 Article ID 369421. doi:10.1155/2008/369421.
41. Tsirimokou G, Psychalinos C, Elwakil A. Emulation of a constant phase element using operational transconductance
amplifiers. Analog Integrated Circuits and Signal Processing 2015; 85(3):413–423.
DOI: 10.1002/cta

Main WRK PDF

Uploaded by

Copyright:

Available Formats

You might also like

Main WRK PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Main WRK PDF

Uploaded by

Copyright:

Available Formats

INTERNATIONAL JOURNAL OF CIRCUIT THEORY AND APPLICATIONS

Int. J. Circ. Theor. Appl. 2017; 45:595–612

Fractional-order electronically controlled generalized ﬁlters

Georgia Tsirimokou1, Costas Psychalinos1,*,† and Ahmed S. Elwakil2

Received 25 January 2016; Revised 18 July 2016; Accepted 26 July 2016

Copyright © 2016 John Wiley & Sons, Ltd.

2. GENERALIZED FILTERS OF ORDER a

2.1. Fractional-order lowpass ﬁlter (FLPF)

The magnitude and phase responses are given respectively by

with a corresponding phase angle of (3)

2.2. Fractional-order highpass ﬁlter (FHPF)

2.3. Fractional-order allpass ﬁlter (FAPF)

2.4. Fractional-order bandpass ﬁlter (FBPF)

In the case that α = 2β, then ωpeak = ωo.

equal to (argfK 1 g þ ð2βa

2.5. Design equations for generalized ﬁlters of order a

G2 s2 þ Gτ11 s þ τG1 τo2

3. GENERALIZED FILTERS OF ORDER a + β

3.1. Fractional-order lowpass ﬁlter (FLPF)

3.2. Fractional-order highpass ﬁlter (FHPF)

The responses of this ﬁlter are described by

and the half-power frequency is calculated by solving (23)

3.3. Fractional-order bandpass ﬁlter (FBPF)

with magnitude and phase responses given given by (25)

3.4. Fractional bandstop ﬁlter (FBSF)

The magnitude and phase responses are these in (29)

3.5. Design equations for generalized ﬁlters of order a + β

Table II. Deﬁnitions of the variables used in the expression in (30).

The realized transfer function of this FBD is that in (33)

G4 s4 þ Gτ 13 s3 þ τG1 τ22 s2 þ τ1 τG21τ3 s þ τ 1 τG2 τo3 τ4

4. REALIZATION OF FRACTIONAL-ORDER GENERALIZED FILTERS USING OTAS

FLPF K 1 ND44 K 1 ND33 K 1 ND22 K 1 ND11 K 1 NDoo

Figure 4. OTA-C realization of FLPF, FHPF, and FAPF functions of order a.

Figure 6. OTA structure with enhanced linearity.

Io 9.89 nA 8.1 nA 6.63 nA

Table VII. Frequency characteristics of the FBPFs.

You might also like