Sixth order Low-filter Sallen-Key

Castro Martínez Agustín Ismael, García Torres Johan, Gonzales López Sergio, Méndez Cansino Juan Pablo, Tapia
Saltillo Victor Manuel.
Facultad de Ciencias de la Electrónica.
Benemérita Universidad Autónoma de Puebla.
Puebla, México.
Mailing address:72000.
[email protected], [email protected], [email protected] , [email protected] ,
[email protected]

Abstract. – The following report corresponds to the practice of The transfer function for a low pass filter is the
type Sallen-key filter, this was the design and implementation of following
one sixth-order. For both filters implementation was carried out 𝐸2 𝑮𝑾𝟎 𝟐
by operational amplifier (OPAMP) with registration TL081, = 𝑾𝒐 (1)
𝐸1 𝒔𝟐 + 𝒔+𝑾𝟎 𝟐
considerations which are specified in the theoretical part were 𝑸
taken for the design of both filters. The following equations are obtained from the transfer
Keyword- Transfer function and low pass function of the previous equation:

I. INTRODUCTION 𝑏0 = 𝑊0 2 (2)
The quality factor denominated with the letter (Q) is a 𝑊0
property of the conventional or normal filters, this quality 𝑏1 =
𝑄
(3)
factor tends to be relatively low for the most part, in order
to improve this disadvantage, the Sallen-key arrangement √𝑏0
𝑄= (4)
can be used so that this filter parameter can be optimized. 𝑏1

The Sallen-Key filter transfer function is shown below:

In order to design both filters it was assumed a
corresponding transfer function for each filter so that the 𝒖
𝐸2 𝑹𝟏𝑹𝟐𝑪𝟏𝑪𝟐
transfer function allow the knowledge and to determine the = 𝟏 𝟏 𝟏−𝒖 𝟏 (5)
𝐸1 𝒔𝟐 +( + + )𝒔+
obtained values of earning for the filters, also using the 𝑹𝟏𝑪𝟏 𝑹𝟐𝑪𝟏 𝑹𝟐𝑪𝟐 𝑹𝟏𝑹𝟐𝑪𝟏𝑪𝟐

analysis of electrical circuits it is possible the calculus of the

values implemented in the circuit. Finally after the To prepare the different low pass filters we occupy the
realization of the previous process, a simulation was made following two designs:
for the corresponding filter so that it could be observed that
the design is correct in order to implement each circuit of Design 1
the filter physically. 𝐶1 = 𝐶2 = 1𝐹 (6)
1
Denormalization in Frequency 𝑅1 = 𝑅2 =
𝑊0
(7)
1
Performing a denormalization is very important in signal 𝑢 =3− (8)
𝑄
processing. Recall that a normalization of the circuit is
necessary to change an initial frequency to another desired
frequency. IV. EXPERIMENTAL DEVELOPMENT

II. CONTENT DEVELOPMENT

To start the practice, it was necessary to use
A. Objetive
table 1.1.
Design, simulate and implement a sixth order device, using
the mathematical analysis known as Sallen-Key, make an
analysis of the results obtained.

III. Theoretical
framework

For this practice we make two designs, in the following

description we divide the two low pass filters to report in a
clear way.
Table1. Factors for chebychev
filters b) For the second system of second order, we
have:
From where we can obtain the values to be substituted in 𝑆 2 + 0.339763𝑆 + 0.557720
our transfer function and consequently in our circuit:
(𝑠 2 + 0.124362𝑠 + 0990732)(𝑠 2 + 339763𝑠 𝑏0 = 0.557720 𝑏1 = 0.339763
+ 0.557720)(𝑠 2 + +0.464125𝑠
+ 0.124707) 𝜔0 = √𝑏0 = 0.746806
It should be noted that our filter is a cascade construction 𝑅𝑎 = 4.4113
of 3 standard filters of second order.
a) For the first second order system, we have: √𝑏𝟎
Ø= = 2.19802
𝑆 2 + 0.1243625 + 0.990732 ≡ 𝑆 2 + 𝑏1 + 𝑏𝟏

𝑏2 𝑅𝒃 = 𝑅𝑎 (1.5450)
1
𝜔0 = √𝑏0 = √0.990732 = 0.9953 𝑅= 𝜔0
= 1.3390Ω
𝑅𝑏 = 6.81551
√𝑏𝟎
Ø= = 8.0036
𝑏𝟏 1 𝑅𝑏
𝑀 =3− Ø
= 2.5450 → 2.5450 = 1 + ⁄𝑅
1 𝑎
𝑅= 𝜔0
= 1.0046Ω
→ 𝑅𝑏 = 𝑅𝑎 (1.5450)
1 𝑅𝑏 (IV)
𝑀 =3− = 2.875 → 2.875 = 1 + ⁄𝑅
Ø 𝑎
Si 𝑅𝑎 = 1Ω → 𝑅𝑏 = 1.5450Ω
→ 𝑅𝑏 = 𝑅𝑎 (1.875)
2.678 = 𝑅𝑎 𝑅𝑏 ⁄𝑅𝑎 + 𝑅𝑏
Si 𝑅𝑎 = 1Ω → 𝑅𝑏 = 1.875Ω
𝑅1 + 𝑅2 = 𝑅𝑎 𝑅𝑏 → 𝑅2 = 𝑅𝑎 𝑅𝑏 ⁄𝑅𝑎 + 𝑅𝑏 →
Como 𝑅 = 1.0046 → 𝑅𝑎 𝑅𝑏 ⁄𝑅𝑎 + 𝑅𝑏 1 = 𝑅𝑎 𝑅𝑏 ⁄𝑅𝑎 + 𝑅𝑏 𝑅𝑎 + 𝑅𝑏 = 𝑅𝑎 𝑅𝑏
→ (2.01) 𝑅𝑎 + (2.01) 𝑅𝑏 = 𝑅𝑎 𝑅𝑏 𝑅𝑎 + (1.5450)𝑅𝑎 = 𝑅𝑎 𝑅𝑏
Sustituyendo …… → (2.678) (𝑅𝑎 + 𝑅𝑏 ) = 𝑅𝑎 𝑅𝑏
(2.01)𝑅𝑎 + [2.01(1.875 𝑅𝑎 )] = 𝑅𝑎2 (1.875) → 𝑅𝑏 = 1.5450𝑅𝑎
___________________________________________ ___________________________________________
2.01𝑅𝑎 + 3.768𝑅𝑎 = 𝑅𝑎2 (1.875) _____________________________________

→ 5.7787 = 𝑅𝑎 (1.875) → (2.678) (𝑅𝑎 + 1.545 𝑅𝑎 ) = 𝑅𝑎 (1.545 𝑅𝑎 )

→ 𝟑. 𝟎𝟖𝟐 = 𝑅𝑎 2.678(2.545𝑅𝑎 ) = 𝑅𝑎2 (1.545 𝑅𝑎 )

→ 𝑅𝑏 = (3.082)(1.875) = 𝟓. 𝟕𝟕𝟖 → 6.8155 = 𝑅𝑎 (1.545) → 𝑅𝑎 = 4.4113

Image 1. Circuit of first quadratic factor

Image 2. Circuit of second quadratic factor
c) For the third system of second order, we have:
2
𝑆 + 0.464125𝑆 + 0.124707
𝑏0 = 0.124707 𝑏1 = 0.464125

𝜔0 = √𝑏0 = 0.35313

√𝑏𝟎
Ø= 𝑏𝟏
= 0.76086

1
𝑅= 𝜔0
= 2.83174Ω

1 𝑅𝑏
𝑀 =3− = 1.6856 → 1.6856 = 1 + ⁄𝑅
Ø 𝑎

→ 𝑅𝑏 = 𝑅𝑎 (1.5450)

Si 𝑅𝑎 = 1Ω → 𝑅𝑏 = 0.6856Ω
2𝑅 = 𝑅𝑎 𝑅𝑏 ⁄𝑅𝑎 + 𝑅𝑏
2(2.83174) = 𝑅𝑎 𝑅𝑏 ⁄𝑅𝑎 + 𝑅𝑏 Image 3. Circuit in cascade
→ 5.66348(𝑅𝑎 + 𝑅𝑏 ) = 𝑅𝑎 [𝑅𝑎 (0.6856)] The scaling in frequency and impedance was done using
the formulas already known for these processes.
→ 5.66348(𝑅𝑎 + [𝑅𝑎 (0.6856)]) =
𝑅𝑎 [𝑅𝑎 (0.6856)]
Factor D:
→ (9.54636𝑅𝑎 ) = 𝑅𝑎2 (0.6856) 𝑊𝑑𝑒𝑠𝑒𝑎𝑑𝑎
𝐹𝑎𝑐𝑡𝑜𝑟 𝐷 =
→ 9.54636 = [𝑅𝑎 (0.6856)] 𝑊𝑖𝑛𝑖𝑐𝑖𝑎𝑙
Factor Z:
→ 13.92 = 𝑅𝑎
𝐶𝑖𝑛𝑖𝑐𝑖𝑎𝑙
𝐹𝑎𝑐𝑡𝑜𝑟 𝑍 =
→ 𝑅𝑏 = 9.5463Ω 𝐶𝑑𝑒𝑠𝑒𝑎𝑑𝑜
It was required to make a sixth order chebyshev
filter, which will have a cutoff frequency of 10khz. And
likewise to avoid some undesired impedance values
it was also required to make a scaling in impedance;
what results in table 1.2.
normalized
scaling in scaling in
Frequency impedance
and
(10kh) frequency

R1 1 1 159

R2 1 1 159
3.0695
Ra 3.0695 488
Image 3. Circuit of third quadratic factor

Rb 5.74 5.74 412

The total or cascade union of our system is shown
below.
1.59e-5
C1=C2 1 10n
 R3 Outa 5 1.3390
R3 1.3390 1.3390 212
R4 5 6 1.3390
R4 1.3390 1.3390 212
6.8155 Rc 7 Outb 6.8155
Rc 6.8155 1083

4.413 Rd 7 0 4.413
Rd 4.413 701
X2 7 6 Outb OPAMPFI

C3=C4 1 1.59e-5 10n C3 6 0 1

C4 5 Outb 1
R5 2.8317 2.8317 450

R6 2.8317 2.8317 450

9.5463 R5 Outb 8 2.8317
Re 9.5463 1517
R6 8 9 2.8317
Rf 13.92 13.92 2213
Re Outc 10 9.5463
C5=C6 1 1.59e-5 10n
Table 2. Calculated values Rf 10 0 13.92

X3 10 9 Outc OPAMPFI
Accumulated results
C5 9 0 1
From where we can obtain the values to be substituted in C6 Outc 8 1
our transfer function and consequently in our circuit.
It should be noted that our filter is a cascade construction.

.ac dec 1000 .001 1e6

.probe

.INC Practica2BUENA.CMD

.END

The denormalization gave the following result:

R1 1 2 1

R2 2 3 1

Ra 0 4 3.0695

Rb Outa 4 5.74

X1 4 3 Outa OPAMPFI

C1 0 3 1

C2 Outa 2 1

V1 1 0 AC 1
R1 1 2 1

R2 2 3 1

Ra 0 4 3.0695

Rb Outa 4 5.74

X1 4 3 Outa OPAMPFI

C1 0 3 1.59e-5

C2 Outa 2 1.59e-5

V1 1 0 AC 1

R3 Outa 5 1.3390

R4 5 6 1.3390

Rc 7 Outb 6.8155 R1 1 2 159

Rd 7 0 4.413 R2 2 3 159

X2 7 6 Outb OPAMPFI Ra 0 4 488.050

C3 6 0 1.59e-5 Rb Outa 4 912.66

C4 5 Outb 1.59e-5 X1 4 3 Outa OPAMPFI

C1 0 3 100n

R5 Outb 8 2.8317 C2 Outa 2 100n

R6 8 9 2.8317 V1 1 0 AC 1

Re Outc 10 9.5463

Rf 10 0 13.92 R3 Outa 5 212.901

X3 10 9 Outc OPAMPFI R4 5 6 212.901

C5 9 0 1.59e-5 Rc 7 Outb 1083.664

C6 Outc 8 1.59e-5 Rd 7 0 701.66

X2 7 6 Outb OPAMPFI

.ac dec 1000 .001 1e6 C3 6 0 100n

.probe C4 5 Outb 100n

.INC Practica2BUENA.CMD
R5 Outb 8 450.24

.END R6 8 9 450.24

The scaling of impedance gave the following result: Re Outc 10 1517.86

Rf 10 0 2213.28

X3 10 9 Outc OPAMPFI

C5 9 0 100n

C6 Outc 8 100n

.ac dec 1000 .001 1e6 Fig. 5. Frequency 52.41kHz.

.probe

.INC Practica2BUENA.CMD
V. CONCLUSIONS

The design of filters is very important since these can have

.END different applications in signal processing, since most
everything not around or paragraphs electronic or electric-
Data obtained in the oscilloscope powered signals, an example of this is the air conditioning
Frequency required sample or signals of temperature and based on
Channel 1: Input voltage
Channel 2: Output voltage these system is tongues or not.

Filters do not allow as its name indicates it to filter out

signals according to their frequencies, these will be
designed according to the application to which the filter is
oriented, we can design low pass filters, high pass, band
passes, passes all or reject all, these can be passive or
active as that implemented in this practice.

In the case of the pass filter low Sallen-Key, the

relationship between the output impedance of the amplifier
and filter resistors, adjusts transfer function to comply with
the established cut-off frequency. The best choice of
values of components depends on the conditions of the
circuit and what you want to do.

When filters experimentally in the laboratory for

processing signals, the implementation was not adequate
since resistances values were not necessary but
Fig. 3. Frequency 7.599kHz commercial values approximate to the estimates in the
design.

III. REFERENCES
[1] Floyd, T. “Electronic Devices: Conventional - Flow Version”. 4ª
Ed., Prentice Hall, Englewood Cliffs, N. J. 1996.
[2] Sedra, A. S. and Smith, K.C., “Microelectronic Circuits”, Oxford
University Press, Inc., New York, New York, 1998.
[3] Boylestad, Robert L., Nashelsky, Louis. “Theory of circuits and
electronic devices”. 10ª Ed. Pearson Education. Prentice Hill Inc.
Mexico, 2009.
[4] M.E . Van Valkenburg, “Analog Filter Design”, HRW Series in
Electrical and Computer Engineering, Illinois University Inc, New
York, New York, 1982
Fig. 4. Frequency 9.158kHz. [5] Huelsman, L. “Active and Passive Analog Filter Design and
Introduction”. McGraw-Hill Inc. N.Y. USA. 1993.
150kHz