Diseño de Filtros Activos Parte 4

Name: Diseño de Filtros Activos Parte 4
Uploaded: 2025-07-07T16:01:20.000Z
Duration: 50 min 42 s

Active Filter Design with Chebyshev Polynomial Approximation

Introduction to Active Filters

The video introduces practical designs of active filters using the Salenki architecture and Chebyshev polynomial approximation, highlighting its complexity and usefulness in understanding polynomial approaches.

Limitations of Salenki Architecture

A key disadvantage of the Salenki architecture is its poor adaptability as filter order increases; from order four onwards, coefficient calculations become cumbersome and less practical.

While simpler architectures exist that better accommodate second-order filters, they have specific advantages making them suitable for certain applications.

Mathematical Foundations

The discussion begins with a normalization assumption for cutoff frequency and gain set to one while designing a low-pass filter to simplify the transfer function analysis.

The transfer function is defined as K/(1 + s²), where K is assumed to be 1, introducing ripple (ε) in the passband which contrasts with Butterworth's maximally flat response.

Characteristics of Chebyshev Filters

Chebyshev filters exhibit a more abrupt cutoff at the passband compared to Butterworth filters, despite having oscillations due to their cosine-based functions.

The family of functions related to filter order (C_n(ω)) varies between -1 and 1, depending on frequency ω. This relationship can be iteratively developed for different orders.

Transfer Function Insights

For direct Chebyshev filters, ripple occurs in the passband; inverse Chebyshev filters show similar ripple but in the stopband—less relevant for this series.

It’s noted that if the filter order (n) is even or odd affects transfer function values at zero frequency (S = 0), impacting design considerations.

Ripple and Attenuation Relationships

Ripple relates directly to maximum attenuation allowed in the passband; it must remain within limited values for effective filtering.

Maximum attenuation can be calculated using a formula involving ripple factor ε: √(10^(A_max/10)-1), linking performance metrics crucially.

Filter Order Considerations

The filter order influences discrimination capabilities; higher differences between maximum attenuations indicate better discrimination quality.

Discrimination factors are tied closely with hyperbolic cosine relationships, emphasizing how selectivity impacts overall filter performance.

Understanding Chebyshev Filters and Their Characteristics

Frequency Response and Selectivity

The frequency response of a filter is characterized by its passband and stopband. A higher value of K (subscripted as K_subs ) indicates greater selectivity in frequency, meaning the transfer function's slope drops sharply.

Chebyshev filters exhibit oscillation within the passband but provide significant selectivity, with their K values approaching one.

Mathematical Foundations

For frequencies above the cutoff ( omega_p ), traditional trigonometric functions are not applicable; instead, hyperbolic functions are used to define the behavior of the filter.

It can be demonstrated that for frequencies greater than omega_p , both C_n(omega) and the transfer function H(omega) decrease monotonically, indicating a consistent decline in output.

Filter Design Considerations

A direct Chebyshev low-pass filter of order n has an expression involving polynomial coefficients in its denominator, while its gain is divided by 2^n-1 .

The design process requires understanding ripple and maximum attenuation parameters. Coefficients vary significantly based on these factors, which can be found in reference tables.

Practical Examples and Coefficient Tables

An example presented involves a direct filter with a maximum attenuation of 1 dB and a ripple factor of 0.5089—common values that influence coefficient calculations.

The resulting coefficient table includes filters up to order nine, emphasizing careful handling of decimal precision during calculations to avoid errors.

Gain Variability and Quality Factor

The gain within the passband varies between specific limits influenced by ripple effects. This variability depends on the filter's order.

Adjustments may be necessary to ensure that independent terms in transfer functions equal one for easier extraction of quality factors and frequency characteristics.

Example Calculation: First Order Filter

Starting with a first-order filter where only one parameter is available (the independent term), normalization leads to specific squared modulus values at defined frequencies.

For normalized conditions at zero frequency, the squared modulus equals one; however, it reaches approximately 0.7943 at higher normalized frequencies.

Understanding Filter Design and Transfer Functions

Introduction to Chebyshev Approximation

The discussion begins with the application of a general equation for Chebyshev approximation, focusing on the relationship between variables such as hds , k , and s + a_0 .

The coefficient a_0 is specifically addressed, with the variable S' defined as S_par = 1.9652267 .

Frequency Cutoff and Gain Calculation

A cutoff frequency of 1 kHz is proposed, maintaining a gain ( K ) equal to 1.

Resistance ( R ) is set at 10K, emphasizing that values can vary within commercial ranges (e.g., 5K, 20K).

Capacitor Value Determination

The capacitor value ( C ) is calculated using the formula involving frequency and resistance, resulting in approximately 8.1 nanofarads.

The design ensures that the transfer function meets specific criteria at both DC and cutoff frequencies.

Bode Diagram Analysis

A Bode diagram illustrates that at 1 kHz, the gain measures -1 decibel; at DC, it equals 0 decibels.

The simulation results are generated using Analog Devices' Spice software, indicating minimal changes in filter characteristics.

Transitioning to Second Order Filters

The focus shifts to designing a second-order filter under similar conditions (gain = 1; frequency = 1 kHz).

Key coefficients for this second-order filter are introduced: S_1 = 1.0977343 , a_0 = 1.1025103 .

Mathematical Development for Second Order Filters

The general form of the transfer function for second-order filters is presented as dependent on coefficients in both numerator and denominator.

Calculating capacitance leads to an approximate value of 15.2 nanofarads based on previous equations.

Practical Implementation Considerations

Real-world construction involves adjusting gains while ensuring output characteristics remain unchanged through operational amplifiers.

Observations from Simulation Results

After simulating the second-order filter design, oscillation patterns emerge alongside expected gain behavior across different frequencies.

Designing a Third-Order Filter

Overview of Filter Design

The goal is to design a third-order filter with characteristics similar to previous filters, using coefficients from an earlier table. The transfer function at the origin has a squared magnitude of 1 and at frequency omega_p equals 0.7943.

Transfer Function Components

The general expression for the numerator involves k multiplied by 2^-1 , resulting in 4k . The polynomial in the denominator will be of third degree, with coefficients extracted from the provided table. Here, K is set to 1.

Calculation Challenges

Initial calculations yield a numerator value of approximately 0.491. To create a third-order filter, it combines first-order and second-order filters into two transfer functions: HD1 and HD2, which must align correctly in their coefficients during multiplication.

Coefficient Extraction

By manipulating equations and performing detailed calculations, values for coefficients B_0 , C_1 , and C_0 are determined for both filter orders needed to achieve the desired third-order response. This process includes factoring out common terms for clarity in calculations.

Circuit Configuration

Two capacitors are identified as necessary components for completing the circuit—one for each filter order—though adjustments may be required based on practical component availability affecting cutoff frequencies. A gain adjustment between stages is also discussed as part of circuit complexity management.

Simulation Results and Observations

Simulation Outcomes

Upon simulating the designed filter, results show two peaks (crests) due to equal cutoff frequencies at the origin, characteristic of third-order Chebyshev filters that exhibit more complex behavior than lower orders (first or second). This complexity introduces potential challenges in achieving precise filtering outcomes without errors.

Higher Order Filters Complexity

As designs progress towards fourth-order filters, complications increase significantly; maintaining accuracy becomes challenging due to higher probabilities of error during configuration adjustments and simulations involving multiple components like resistors and capacitors that need careful selection based on their values (e.g., 16nF and 30nF capacitors).

Final Adjustments Needed

It’s noted that final adjustments might be necessary between stages to ensure optimal performance; this highlights how intricate tuning can become when dealing with higher order filters while still aiming for effective frequency selectivity despite increased design complexity.

Conclusion on Filter Design Process

Key Takeaways

The discussion concludes by emphasizing that while designing high-selectivity filters can lead to complex configurations requiring patience and meticulousness, successful implementation yields significant benefits in frequency response precision across various applications within electronic circuits.