Electronics for Transmission Systems

Laboratory 1 – Synthesis and realization of active filters

Names:

-Pablo Orellana (20225753)

-Asim Ali (20225721)

-Muzzamil waqas (20225694)

The objective of this lab is twofold:

• On the one hand, synthesize and implement basic structures of active filters by
analyzing their behavior.
• On the other hand, automate the measurement of the Bode diagram by controlling
measuring devices. This point will provide basic knowledge and practice regarding
ordering to distance from measuring devices via the IEEE-488 instrumentation bus and
programming graphics.

B. Preparation

I. Automatic measurement of the Bode Diagram

1. How to generate a logarithmic distribution of points with frequency as input variables

minimum, maximum frequency and number of points?

First, we need to know that a logarithmic distribution of frequency has a linear distribution
when the logarithm of the frequency is used. See figure 1.

Figure 1. Linear distribution of logarithmic frequency points

It is observed that the linear rate “r” can be expressed in terms of fmin, fmax and N as:
log(𝑓𝑚𝑎𝑥) − log⁡(𝑓𝑚𝑖𝑛) 1 𝑓𝑚𝑎𝑥
𝑟= = log ( ) … (1)
𝑁−1 𝑁−1 𝑓𝑚𝑖𝑛
The n-th linear term can be calculated as:

𝑡𝑛−𝑙𝑖𝑛𝑒𝑎𝑟 = 𝑟(𝑛 − 1) + log(𝑓𝑚𝑖𝑛) … (2)

So, the n-th logarithmic term will be:

𝑡𝑛−𝑙𝑜𝑔 = 10𝑡𝑛−𝑙𝑖𝑛𝑒𝑎𝑟 … (3)

Replacing (1) in (2) and (2) in (3):

𝑛−1 𝑓𝑚𝑎𝑥
log( )+log⁡(𝑓𝑚𝑖𝑛)
𝑡𝑛−𝑙𝑜𝑔 = 10𝑁−1 𝑓𝑚𝑖𝑛

𝑛−1 𝑓𝑚𝑎𝑥
log( )
𝑡𝑛−𝑙𝑜𝑔 = 10𝑁−1 𝑓𝑚𝑖𝑛 ∗ 10log⁡(𝑓𝑚𝑖𝑛)
𝑓𝑚𝑎𝑥 𝑛−1
log( )
𝑡𝑛−𝑙𝑜𝑔 = (10 𝑓𝑚𝑖𝑛 )𝑁−1 ∗ 10log⁡(𝑓𝑚𝑖𝑛)
𝑛−1
𝑓𝑚𝑎𝑥 𝑁−1
𝑡𝑛−𝑙𝑜𝑔 =( ) ∗ 𝑓𝑚𝑖𝑛
𝑓𝑚𝑖𝑛
𝑛−1 𝑁−𝑛
𝑡𝑛−𝑙𝑜𝑔 = 𝑓𝑚𝑎𝑥 𝑁−1 ∗ 𝑓𝑚𝑖𝑛 𝑁−1

Example:

For fmax=100,fmin=10 and N=5

𝑛−1 5−𝑛
𝑡𝑛−𝑙𝑜𝑔 = (100) 4 ∗ (10) 4

So:
1−1 5−1
𝑡1−𝑙𝑜𝑔 = (100) 4 ∗ (10) 4 = (100)0 ∗ (10)1 = 10
2−1 5−2 1 3
𝑡2−𝑙𝑜𝑔 = (100) 4 ∗ (10) 4 = (100)4 ∗ (10)4 = 17.7828
3−1 5−3 2 2
𝑡3−𝑙𝑜𝑔 = (100) 4 ∗ (10) 4 = (100)4 ∗ (10)4 = 31.6228
4−1 5−4 3 1
𝑡4−𝑙𝑜𝑔 = (100) 4 ∗ (10) 4 = (100)4 ∗ (10)4 = 56.2341
5−1 5−5
𝑡5−𝑙𝑜𝑔 = (100) 4 ∗ (10) 4 = (100)1 ∗ (10)0 = 100

Logarithmic distribution of 5 points between 10 and 100 Hz.

𝑡𝑛−𝑙𝑜𝑔 = [10; ⁡17.7828; ⁡31.6228; ⁡56.2341; 100]; ⁡⁡𝑛: 1,2, … , 𝑁

We can generate this distribution in MATLAB by using the logspace(X1,X2,N) function, see
Figure 2. Using as inputs the exponents of the decades 10^X1 and 10^X2 and N the number of
points.
 Figure 2. Logarithmically space vector

2. The active filter uses operational amplifiers. These amplifiers are not ideal (Slew rate and/or
cutoff frequency)! It is therefore necessary to ensure to remain in linear mode. Outside of
bandwidth problem, the limits of a measurements are often linked to the signal/noise ratio!
What precautions should I take?

The signal to noise ratio should be higher, also to ensure the operational amplifier to remain in
linear mode. The input voltage should be lower than the Vs which is the supply Voltage of the
amplifier. See Figure 3

Figure 3. Operational Amplifier Scheme

3. What is the procedure to follow to draw the Bode diagram point by point? In deduce the
approach to be adopted to carry out the programming of the measurement bench. We only
ask the amplitude plot, why?

First of all, the transfer function must be rewritten in the proper form. This means separate the
transfer function into its constitutive parts and be able to recognize each part. Either a
constant, poles and zeros.

For example:
𝑠 𝑠
2(𝑠 + 10) 2(𝑠 + 10) 2 ∗ 10(10 + 1) 5 (10 + 1)
𝐻(𝑠) = 2 = = = ∗
𝑠 + 7𝑠 + 12 (𝑠 + 4)(𝑠 + 3) 4 ∗ 3 ∗ ( 𝑠 + 1)( 𝑠 + 1) 3 ( 𝑠 + 1)( 𝑠 + 1)
4 3 4 3
This function has:

• A constant
• A zero at s=-10
• Two poles at s=-4 and s=-3

Then, we need to replace s by jw:

𝑗𝜔
5 (10 + 1)
𝐻(𝑗𝜔) = ∗
3 (𝑗𝜔 + 1)(𝑗𝜔 + 1)
4 3
We can draw this transfer functions using the rules for drawing bode diagrams.
 Term Magnitude Phase
Constant: K 20log10(|K|) K>0: 0°
K<0: ±180°

Pole at origin -20 dB/decade passing -90°

1 through 0dB at 𝜔 = 1
𝑗𝜔
Zero at origin +20 dB/decade passing +90°
𝑗𝜔 through 0dB at 𝜔 = 1
Real Pole 1. Draw low frequency 1. Draw low frequency
1 asymptote at 0 dB. asymptote at 0°
𝑗𝜔 2. Draw high frequency 2. Draw high frequency
𝜔0 + 1 asymptote at -20 asymptote at -90°
dB/decade. 3. Connect with a straight
3. Connect lines at ω0. line from 0.1·ω0 to
10·ω0
Real Zero 1. Draw low frequency 1. Draw low frequency
𝑗𝜔 asymptote at 0 dB. asymptote at 0°
+1
𝜔0 2. Draw high frequency 2. Draw high frequency
asymptote at +20 asymptote at +90°
dB/decade. 3. Connect with a straight
3. Connect lines at ω0. line from 0.1·ω0 to
10·ω0

Then, for the transfer function H(𝑗𝜔):

 Figure 4. Bode plot of H(jw)

II. Synthesis of a Butterworth low-pass filter

We want to create a Butterworth low-pass filter admitting:

• A passband fp=50kHz with a gain a=-3dB in the passband.

• A maximum gain b=-20dB for frequencies above fa=225kHz.

Draw the template of this filter then calculate its selectivity.

 Figure 5. Template of the Butterworth Low-Pass Filter

The selectivity can be calculated as:

𝑓𝑝 50𝑘𝐻𝑧
𝐾= = = 0.22
𝑓𝑎 225𝑘𝐻𝑧

Recall the approximation function of a Butterworth filter.

1
|𝑇(𝑗Ω)|2 = , 𝐴𝑝𝑟𝑜𝑥𝑖𝑚𝑎𝑡𝑖𝑜𝑛⁡𝑜𝑓⁡𝐵𝑢𝑡𝑡𝑒𝑟𝑤𝑜𝑟𝑡ℎ⁡𝑓𝑖𝑙𝑡𝑒𝑟
1 + 𝜀 2 Ω2𝑛
Determine the expression of 𝜀 as a function of a. Make the digital application with a=-3dB.

𝜀 = √10−𝑎/10 − 1

𝜀 = √10−(−3)/10 − 1

𝜀 = √100.30 − 1 = 0.9976 ≈ 1

Determine the expression of the minimum order 𝑛𝑚𝑖𝑛 of the normalized low-pass filter as a
function of b, 𝜀 and k so that the response curve in amplitude passes through the points of
coordinates (Ω=1, a) and (Ω=a, b).

10−𝑏/10 − 1
ln⁡( )
𝑛≥ 𝜀2
1
2ln⁡(𝐾 )
 20
1010 − 1
ln ( )
0.99526
𝑛≥ = 1.53
2 ln(4.5)
𝑛 ≥ 1.53
𝑛𝑚𝑖𝑛 = 2
From the specifications given above, calculate the order of the filter and deduce its transfer
function.
1 1
|𝑇(𝑗Ω)|2 = 2 2𝑛
= , 𝑠 = 𝑗Ω
1+𝜀 Ω 1 + Ω4
1
|𝑇(𝑠)| =
1 + 𝑠4
1 + 𝑠4 = 0

𝑠 4 = −1

(𝑒 𝑗𝜙 )4 = 𝑒 𝑗(𝜋+2𝜋𝑘)
4𝜙 = (2𝑘 + 1)𝜋
𝜋
𝜙= (2𝑘 + 1)
4

Figure 6. Poles draw in the complex plane

3𝜋 5𝜋
From Figure 6, the poles 𝜙 = 4
and 4
are chosen because those make the system stable.
3𝜋 1 1
𝑒𝑗 4 = − +
√2 √2
5𝜋 1 1
𝑒𝑗 4 = − −
√2 √2

1 1
𝑇(𝑠) = =
1 1 1 1 𝑠2 + 1.414𝑠 + 1
(𝑠 + − )(𝑠 + + )
√2 √2 √2 √2
 1
𝑇(𝑝) = 𝑝 𝑝 ,⁡⁡⁡⁡⁡⁡𝜔𝑝 = 2𝜋𝑓𝑝 = 314.16𝑘𝑟𝑎𝑑/𝑠
(𝜔 )2 + 1.414 𝜔 + 1
𝑝 𝑝

III. Synthesis of a Chebychev low-pass filter

Determine the expression of 𝜀 in function of a.

𝜀 = √10−𝑎/10 − 1

𝜀 = √101/10 − 1

𝜀 = √100.1 − 1 = 0.5088

𝑓 1
Knowing that the gain response curve passes through the coordinate point (Ω𝑎 = 𝑓𝑎 = 𝐾 , 𝑏),
𝑏
determine the expression of the minimum order of the filter 𝑛𝑚𝑖𝑛 .

𝑓𝑎 1 225
Ω𝑎 = = = = 4.5
𝑓𝑏 𝐾 50

Figure 7. Abaque permettant de determiner l’ordre n d’un filtre de Tchevyscheff avec K, a,b

1<n
 𝑛𝑚𝑖𝑛 = 2
We want to realize a Chebyshev filter, having a maximum ripple of 1dB in the bandwidth, and
satisfying the following template: a gain a=-1dB in the bandwidth (for frequencies below
fp=50kHz) and a maximum gain b=-20dB for higher frequencies at fa=225kHz. Determine its
transfer function (see table T1 in the appendix giving the expressions of normalized
transmission functions).

Looking the table T1 the expression of the chevyshev transfer function is:
1 𝑝
𝑇(𝑠) = , 𝑠=
1 + 0.9956𝑠 + 0.9070𝑠 2 𝜔𝑝
1
𝑇(𝑝) = 𝑝 𝑝
1 + 0.9956 + 0.9070( )2
𝜔𝑝 𝜔𝑝

IV. Realization of low-pass filters

1. The Sallen Key Structure

Figure 8. Sallen-Key structure

Calculate the transfer function s/e in a general case. We will specify how to achieve
the perfect voltage amplifier of gain K and it will be assumed that the operational
amplifier used is ideal.
 Figure 9. Sallen-Key Circuit in Laplace Domain

It is noticed from Figure 9 that:

Potentials V1 and V2
𝑆 𝑆
𝑉1 = = … (1)
𝑍 𝐾
(1 + 𝑍7 )
6

𝑉2 𝑉1
= … (2)
𝑍4 + 𝑍5 𝑍5
Replacing (1) in (2):
𝑍4 𝑆
𝑉2 = (1 + ) … (3)
𝑍5 𝐾
Currents i1, i2, i3 and i4 in Potential V2

𝑖1 = 𝑖2 + 𝑖3 + 𝑖4 … (4)
Also
𝐸 − 𝑉2
𝑖1 = … (5)
𝑍1
𝑉2 − 𝑆
𝑖2 = … (6)
𝑍2
𝑉2
𝑖3 = … (7)
𝑍3
𝑉1
𝑖4 = … (8)
𝑍5
Replacing (1) and (3) in (5), (6), (7) and (8):

𝐸 𝑍4 𝑆
𝑖1 = − (1 + ) … (9)
𝑍1 𝑍5 𝐾𝑍1
𝑍4 𝑆
𝑖2 = (1 + − 𝐾) … (10)
𝑍5 𝐾𝑍2
𝑍4 𝑆
𝑖3 = (1 + ) … (11)
𝑍5 𝐾𝑍3
 𝑆
𝑖4 = … (12)
𝐾𝑍5
Replacing (9), (10), (11) and (12) in (4):
𝑖1 = 𝑖2 + 𝑖3 + 𝑖4

𝐸 𝑍4 𝑆 𝑍4 𝑆 𝑍4 𝑆 𝑆
− (1 + ) = (1 + − 𝐾) + (1 + ) +
𝑍1 𝑍5 𝐾𝑍1 𝑍5 𝐾𝑍2 𝑍5 𝐾𝑍3 𝐾𝑍5

Multiplying by 𝑍1 𝑍2 𝐾 both sides:

𝑍4 𝑍4 𝑍4 𝑍1 𝑍2 𝑍1 𝑍2
𝑍2 𝐾𝐸 − (1 + ) 𝑍2 𝑆 = (1 + − 𝐾) 𝑍1 𝑆 + (1 + ) 𝑆+ 𝑆
𝑍5 𝑍5 𝑍5 𝑍3 𝑍5

𝑆 𝑍2 𝐾
=
𝐸 [(1 + 4 − 𝐾) 𝑍 + (1 + 4 ) 𝑍1 𝑍2 + 𝑍1 𝑍2 + (1 + 𝑍4 ) 𝑍 ]
𝑍 𝑍
𝑍 1 𝑍 𝑍 𝑍 𝑍 2
5 5 3 5 5
𝑆 𝐾
=
𝐸 [(1 + 𝑍4 − 𝐾) 𝑍1 + (1 + 𝑍4 ) 𝑍1 + 𝑍1 + (1 + 𝑍4 )]
𝑍5 𝑍2 𝑍5 𝑍3 𝑍5 𝑍5

Transfer function for the Sallen-Key general case (a):

𝑆 𝐾 𝑍7
= … (13),⁡⁡⁡𝐾 = 1 +
𝐸 [(1 + 𝑍4 ) (𝑍1 + 𝑍1 + 1) − 𝑍1 𝐾 + 𝑍1 ] 𝑍6
𝑍 5𝑍 𝑍
2 3 𝑍 𝑍 2 5

To achieve a low-pass filter the components can be chosen as shown in (b). Deduce
the expression of the transfer function of these four elements and of the gain K. Put it
𝐴
under the form 𝐻(𝑝) = 𝑝 𝑝 and calculate A, m and 𝜔0 in function of R1, R2,
1+2𝑚 +( )2
𝜔0 𝜔0
C1, C2 and K.

Using the components chosen in (b):

𝑍1 = 𝑅1 … (14)
1
𝑍2 = … (15)
𝐶2 𝑝
𝑍3 = 𝑅2 … (16)
𝑍4 = 𝑅3 … (17)
1
𝑍5 = … (18)
𝐶1 𝑝
Replacing (14), (15), (16), (17) and (18) in (13):
𝑆 𝐾
=
𝐸 [(1 + 𝑍4 ) (𝑍1 + 𝑍1 + 1) − 𝑍1 𝐾 + 𝑍1 ]
𝑍5 𝑍2 𝑍3 𝑍2 𝑍5
𝑆 𝐾
=
𝐸 [(1 + 𝑅 𝐶 𝑝) (𝑅 𝐶 𝑝 + 𝑅1 + 1) − 𝑅 𝐶 𝐾𝑝 + 𝑅 𝐶 𝑝]
3 1 1 2 𝑅2 1 2 1 1
 𝑆 𝐾 𝑅2
= ∗
𝐸 [𝑅 𝐶 𝑝 + 𝑅 𝑅 𝐶 𝐶 𝑝2 + 𝑅1 + 𝑅1 𝑅3 𝐶1 𝑝 + 1 + 𝑅 𝐶 𝑝 − 𝑅 𝐶 𝐾𝑝 + 𝑅 𝐶 𝑝] 𝑅2
1 2 1 3 1 2 𝑅 𝑅 3 1 1 2 1 1
2 2
𝑆 𝑅2 𝐾
= 2
𝐸 [𝑅1 𝑅2 𝐶2 𝑝 + 𝑅1 𝑅2 𝑅3 𝐶1 𝐶2 𝑝 + 𝑅1 + 𝑅1 𝑅3 𝐶1 𝑝 + 𝑅2 + 𝑅2 𝑅3 𝐶1 𝑝 − 𝑅1 𝑅2 𝐶2 𝐾𝑝 + 𝑅1 𝑅2 𝐶1 𝑝]
𝑆 𝑅2 𝐾
=
𝐸 [𝑅1 + 𝑅2 + (𝑅1 𝑅2 𝐶2 + 𝑅1 𝑅3 𝐶1 + 𝑅2 𝑅3 𝐶1 − 𝑅1 𝑅2 𝐶2 𝐾 + 𝑅1 𝑅2 𝐶1 )𝑝 + 𝑅1 𝑅2 𝑅3 𝐶1 𝐶2 𝑝2 ]

Dividing by 𝑅1 + 𝑅2 :
𝑅2
𝑆 𝑅1 + 𝑅2 𝐾
=
𝐸 (𝑅 𝑅 𝐶 + 𝑅1 𝑅3 𝐶1 + 𝑅2 𝑅3 𝐶1 − 𝑅1 𝑅2 𝐶2 𝐾 + 𝑅1 𝑅2 𝐶1 ) 𝑅 𝑅 𝑅 𝐶𝐶
[1 + 1 2 2 𝑅1 + 𝑅2 𝑝 + 1𝑅 2 +3𝑅1 2 𝑝2 ]
1 2
𝑅2
𝑆 𝐾
𝑅1 + 𝑅2
=
𝐸 1 + 𝑅1 𝑅2 [𝐶 𝑅 ( 1 + 1 + 1 ) + 𝐶 (1 − 𝐾)]𝑝 + 𝑅1 𝑅2 𝑅3 𝐶1 𝐶2 𝑝2
𝑅1 + 𝑅2 1 3 𝑅1 𝑅2 𝑅3 2 𝑅1 + 𝑅2

𝑆(𝑝) 𝐴
= 𝐻(𝑝) = 𝑝 𝑝
𝐸(𝑝) 1 + 2𝑚 𝜔 + (𝜔 )2
0 0
It is deduced:
𝐾
𝐴= … (19)
𝑅1
(1 + )
𝑅2
𝑅1 + 𝑅2
𝜔0 = √ … (20)
𝑅1 𝑅2 𝑅3 𝐶1 𝐶2
1 𝑅1 𝑅2 1 1 1
2𝑚 = [𝐶 𝑅 ( + + ) + 𝐶2 (1 − 𝐾)]
𝜔0 𝑅1 + 𝑅2 1 3 𝑅1 𝑅2 𝑅3

𝜔0 𝑅1 𝑅2 1 1 1
𝑚= [𝐶1 𝑅3 ( + + ) + 𝐶2 (1 − 𝐾)]
2 𝑅1 + 𝑅2 𝑅1 𝑅2 𝑅3

1 𝑅1 + 𝑅2 𝑅1 𝑅2 1 1 1
𝑚= √ [𝐶1 𝑅3 ( + + ) + 𝐶2 (1 − 𝐾)] … (21)
2 𝑅1 𝑅2 𝑅3 𝐶1 𝐶2 𝑅1 + 𝑅2 𝑅1 𝑅2 𝑅3
Show that the two parameters which are the natural frequency and the damping
coefficient can be adjusted independently of each other. Suggest a procedure for
adjusting this cell.

The natural frequency and the damping coefficient can be adjusted independently of
each other by modifying the gain K, because only the damping coefficient depends of
that variable.

2. Realization of low-pass filters

The filter will be carried out by using the Sallen-Key structure and by considering the
transfer function established previously.
Calculate the components necessary for the realization of both Butterworth and
Chebyshev filters as well as the gain K knowing that R1=R2=R3=R5=R and
C1=C2=C=4.7nF. When making the filter, the overall gain is not necessarily unity. This is
of little importance because it suffices to add an amplifier or attenuator stage. During
handling, the filter will be produced without worrying about the low frequency gain,
the template then having to be translated into gain in order to check that the filter
produced corresponds to the given specifications.

Replacing R and C in (19), (20) and (21):

𝐾 1 𝑅4 1 𝑅4
𝐴= = (1 + ) = (1 + ) … (22)
𝑅1
(1 + 𝑅 ) 2 𝑅5 2 𝑅
2

𝑅1 + 𝑅2 2𝑅 √2
𝜔0 = √ =√ 3 2= … (23)
𝑅1 𝑅2 𝑅3 𝐶1 𝐶2 𝑅 𝐶 𝑅𝐶

1 𝑅1 + 𝑅2 𝑅1 𝑅2 1 1 1
𝑚= √ [𝐶1 𝑅3 ( + + ) + 𝐶2 (1 − 𝐾)]
2 𝑅1 𝑅2 𝑅3 𝐶1 𝐶2 𝑅1 + 𝑅2 𝑅1 𝑅2 𝑅3
1 √2 𝑅 2 3 √2
𝑚= [𝐶𝑅 ( ) + 𝐶(1 − 𝐾)] = [3𝐶 + 𝐶 − 𝐶𝐾]
2 𝑅𝐶⁡ 2𝑅 𝑅 4𝐶⁡
√2 √2 𝑅4
𝑚= [4 − 𝐾] = [4 − 1 − ]
4 4 𝑅
√2 𝑅4
𝑚= [3 − ] … (24)
4 𝑅

Butterworth Case:
1 𝐴
𝑇(𝑝) = 𝑝 2 𝑝 = 𝑝 𝑝
(𝜔 ) + 1.414 𝜔 + 1 1 + 2𝑚 𝜔 + (𝜔 )2
𝑝 𝑝 0 0
It is deduced:
1 𝑅4
𝐴 = (1 + ) = 1 … (25)
2 𝑅
√2
𝜔0 = 𝜔𝑝 = 2𝜋𝑓𝑝 = … (26)
𝑅𝐶

𝑓𝑝 = 50𝑘𝐻𝑧, 𝐶 = 4.7𝑛𝐹
Replacing fp and C in (26)
√2 √2
2𝜋𝑓𝑝 = → 2𝜋 ∗ 50 ∗ 103 =
𝑅𝐶 𝑅 ∗ 4.7 ∗ 10−9
6
√2 ∗ 10
𝑅= = 957.7833Ω
2𝜋 ∗ 50 ∗ 4.7
Replacing R in (25)
1 𝑅4 1 𝑅4
(1 + ) = 1 → ⁡ (1 + )=1
2 𝑅 2 957.7833
𝑅4 = 957.7833Ω

Replacing 𝑅4 in (24)
√2 𝑅4 √2
𝑚= [3 − ] = ∗ 2 = 0.7071
4 𝑅 4
 So, for Butterworth:
𝑅1 = 𝑅2 = 𝑅3 = 𝑅4 = 𝑅5 = 957.7833Ω
𝐶 = 4.7𝑛𝐹

Chevychev case:

1 𝐴
𝑇(𝑝) = 𝑝 𝑝 2= 𝑝 𝑝
1 + 0.9956 𝜔 + 0.9070(𝜔 ) 1 + 2𝑚 𝜔 + (𝜔 )2
𝑝 𝑝 0 0

It is deduced:
𝜔𝑝 2𝜋𝑓𝑝 √2
𝜔0 = = = … (27)
√0.907 0.9524 𝑅𝐶

𝜔0
2𝑚 = ∗ 0.9956 … (28)
𝜔𝑝

𝑓𝑝 = 50𝑘𝐻𝑧, 𝐶 = 4.7𝑛𝐹
Replacing fp and C in (27)
√2 √2
2𝜋𝑓𝑝 = 0.9524 ∗ → 2𝜋 ∗ 50 ∗ 103 = 0.9524 ∗
𝑅𝐶 𝑅 ∗ 4.7 ∗ 10−9
6
0.9524 ∗ √2 ∗ 10
𝑅= = 912.1928Ω
2𝜋 ∗ 50 ∗ 4.7

Replacing m and R in (28)

√2 𝑅4 0.9956
2∗ [3 − ] =
4 𝑅 0.9524
√2 𝑅4
[3 − ] = 1.0454
2 912.1928

2
𝑅4 = (3 − 1.0454 ∗ ) ∗ 912.1928
√2
𝑅4 = 1388Ω ≈ 1.4𝑘Ω
Replacing 𝑅4 in (24)
√2 𝑅4 √2 1388
𝑚= [3 − ] = ∗ (3 − ) = 0.5227
4 𝑅 4 912.1928

So, for Chevychev:

𝑅1 = 𝑅2 = 𝑅3 = 𝑅5 = 912.1928Ω
𝑅4 = 1388Ω ≈ 1.4𝑘Ω
𝐶 = 4.7𝑛𝐹

C. Experimentation
I. Study of filters in simulation

Butterworth

Gain:

K=-3dB

Bandwidth:

BW=49.513kHZ

Order of the filter:

We choose the gains of the two following frequencies separated by an octave.

f1=80.509kHz

f2=2*f1=160kHz

We use the relation dB/octaves instead of dB/decades because for the linear region, its ranges
is less than a decade.

6n=G1(f1)-G2(f2)

6n=-9.711-(-21.640)

6n=11.929

n=1.98816
 𝑛≈2
Min Attenuation in Stop-band:

𝐴𝑠𝑡𝑜𝑝⁡𝑏𝑎𝑛𝑑 = −25.976𝑑𝐵

Ripple:

Ripple=0dB

Chevychev
Gain:

K=-3dB

Bandwidth:

BW=60.589kHz

Order of the filter:

We choose the gains of the two following frequencies separated by an octave.

f1=80.509kHz

f2=2*f1=160.157kHz

We use the relation dB/octaves instead of dB/decades because for the linear region, its ranges
is less than a decade.

6n=G2(f2)-G1(f1)

6n=-6.2274-(-19.388)

6n=13.161
 n=2.1935

𝑛≈2

Min Attenuation in Stop-band:

𝐴𝑠𝑡𝑜𝑝⁡𝑏𝑎𝑛𝑑 = −27.814𝑑𝐵

𝑓𝑠𝑡𝑜𝑝⁡𝑏𝑎𝑛𝑑 = −1.2608𝑀𝐻𝑧

Ripple:

Ripple = 1.5394dB

II. Experimental Result

Table of Butterworth Filter

F(KHz) Ve(V) Vs(V) Vs/Ve 20log(Vs/Ve) dB

100Hz 2 2 1 0
 1kHz 2 2 1 0
10kHz 2.08 2.08 1 0
50kHz 2.16 1.32 0.61111 -4.27759
100kHz 2.04 0.52 0.2549 -11.8725
200kHz 2.04 0.28 0.137254 -17.2494
225kHz 2.04 0.24 0.117647 -18.5883
300kHz 2.04 0.24 0.117647 -18.5883

It is seen from the table above, according to the measures, that the circuit tested in the lab
works as a Butterworth Low Pass Filter, at 50kHz the attenuation is -4.27759dB almost -3dB
the bandpass frequency. At 225kHz the attenuation is -17.2794dB almost -20dB the frequency
fa given. After the cut-off frequency the attenuation increases fast.