m-derived Filters

Simple LC ladder filters provide a gradual transition between passband and stopband, which is insufficient for many communication and signal processing applications. The m-derived filter section was introduced to achieve a sharper cutoff characteristic and simultaneously solve the impedance mismatch problem that arises when multiple filter sections are cascaded. Understanding m-derived filters is essential for GATE aspirants because questions often test the relationship between the m-parameter, the pole of attenuation, and the image impedance behavior.

Core Concept Explanation

A constant-k filter is the simplest LC prototype filter where the product Z1 times Z2 equals R0 squared at every frequency, with R0 being the characteristic impedance. While this gives a useful passband and stopband, the attenuation rises very slowly just beyond the cutoff frequency. For applications requiring sharp frequency discrimination, this gradual rolloff is unacceptable.

The m-derived filter is obtained by multiplying the series impedance of the prototype by a factor m and modifying the shunt arm such that the image impedance remains the same as the original constant-k section. The parameter m satisfies 0 less than m less than 1. This modification introduces a resonant condition in the shunt arm, creating a frequency of infinite attenuation called the pole of attenuation or f-infinity, located just beyond the cutoff frequency fc.

The shunt arm of the m-derived T-section contains the capacitor C divided by m in series with an inductor equal to (1 minus m squared) times L divided by 4m. At a specific frequency, this series LC combination resonates and presents zero impedance to ground, effectively short-circuiting the signal path and producing infinite attenuation. This is the fundamental mechanism that makes m-derived filters far superior to simple constant-k designs near the cutoff region.

A second critical advantage is impedance matching. The image impedance of a constant-k section varies significantly near cutoff, making cascaded sections interact badly. The m-derived half-section, placed at the input and output of a composite filter, is specifically designed with m equal to 0.6 to produce a nearly constant image impedance across the entire passband, matching the terminating resistance R0 and minimizing reflection loss.

Mathematical Expression

For a low-pass prototype constant-k T-section with series inductor L and shunt capacitor C, the m-derived T-section has: series arm inductance of mL/2 on each side, and a shunt arm consisting of C/m in series with (1 minus m squared) times L divided by 4m. The cutoff frequency of the m-derived section remains identical to the prototype: fc equals 1 divided by (pi times square root of LC). The pole of infinite attenuation is given by the expression f-infinity equals fc divided by the square root of (1 minus m squared). As m approaches 1, f-infinity moves far away from fc and the section behaves like a constant-k filter. As m approaches 0, f-infinity moves very close to fc giving an extremely sharp but narrow attenuation peak.

The image impedance of the m-derived T-section in the passband is given by Zi-T equals R0 times the square root of (1 minus (f/fc) squared), which is the same form as the constant-k section. However, the m-derived pi-section has an image impedance Zi-pi that closely approximates R0 over most of the passband when m is chosen as 0.6. This is why 0.6 is the standard value used in composite filter design for terminating half-sections.

Practical Understanding

In practical filter design, a composite filter is constructed by combining one or more constant-k sections for broad attenuation with one or more m-derived sections for the sharp attenuation peak, and terminating with m equals 0.6 half-sections at both ends for impedance matching. This combination gives flat passband response, sharp cutoff, high stopband attenuation, and good impedance characteristics simultaneously.

The value of m is chosen based on where the attenuation pole is needed. If f-infinity is to be located at 1.25 times fc, then m equals the square root of (1 minus (fc/f-infinity) squared) equals the square root of (1 minus 0.64) equals 0.6. This explains why m equals 0.6 appears so frequently in filter design tables. The engineer selects the pole location based on the closest interfering frequency that must be blocked, and computes m accordingly.

For high-pass m-derived filters, the same principle applies with the role of inductors and capacitors interchanged. The series arm uses a capacitor mC/2 on each side, and the shunt arm resonates to create the attenuation pole above the cutoff frequency. The formulas mirror those of the low-pass case with frequency inversion applied.

Numerical Example

Consider designing an m-derived low-pass T-section with a cutoff frequency of 10 kHz, characteristic impedance R0 of 600 ohms, and an attenuation pole at 11 kHz. The prototype constant-k values are first calculated, then the m parameter is determined, and finally the m-derived component values are obtained by scaling.

Given:
fc = 10 kHz = 10,000 Hz
R0 = 600 Ω
f∞ = 11 kHz = 11,000 Hz

Why this formula applies:
m-derived T-section is derived from constant-k by multiplying series arm by m
and adding resonant element in shunt arm. Pole location fixes m.

Formula:
m = √(1 - (fc/f∞)²)

Prototype values:
L = R0/π·fc = 600 / (π × 10000)
C = 1/(π·R0·fc) = 1 / (π × 600 × 10000)

Substitution:
m = √(1 - (10000/11000)²) = √(1 - (0.9091)²) = √(1 - 0.8264) = √0.1736

Calculation:
m = 0.4166 ≈ 0.417

L = 600 / (π × 10000) = 600 / 31415.9 = 19.1 mH
C = 1 / (π × 600 × 10000) = 1 / 18,849,556 = 53.05 nF

m-derived T-section components:
Series arm (each half): mL/2 = 0.417 × 19.1 mH / 2 = 3.98 mH
Shunt capacitor: C/m = 53.05 nF / 0.417 = 127.2 nF
Shunt inductor: (1 - m²)L / 4m = (1 - 0.174) × 19.1 mH / (4 × 0.417)
               = 0.826 × 19.1 / 1.668 = 9.45 mH

Final Answer:
m = 0.417
Series inductors (each): 3.98 mH
Shunt capacitor: 127.2 nF
Shunt inductor: 9.45 mH
Attenuation pole confirmed at f∞ = 11 kHz

Exam Tip: GATE often asks to identify m from a given f∞/fc ratio. Remember: m = √(1 - (fc/f∞)²). Also note that the shunt arm resonates at f∞, not at fc. The value m = 0.6 is used for terminating half-sections to match image impedance, not to place an attenuation pole.

Mechanism of Sharp Cutoff

• The shunt arm of the m-derived T-section contains a series LC resonant circuit. At the resonant frequency f-infinity, the impedance of this shunt arm drops to zero, creating a short circuit across the load and producing theoretically infinite attenuation.
• The parameter m controls the location of f-infinity relative to fc. A smaller m places the pole closer to fc, giving a sharper initial cutoff. However, attenuation may decrease again beyond the pole, which is why constant-k sections are added for sustained stopband attenuation.
• The pi-section form of the m-derived filter has its resonant element in the series arm, and it produces the same attenuation pole. Pi-sections are preferred when impedance matching at the port is the primary concern.
• For a composite filter, the standard design uses one constant-k T-section as the core, flanked by m equals 0.6 half-sections. The m-derived section may also be inserted inside with m chosen for the required attenuation pole location.
• The image impedance match at m equals 0.6 occurs because the frequency-dependent variation of the m-derived pi image impedance nearly cancels out across the passband, keeping it close to R0 and reducing reflection at the filter terminals.

Quick Revision

• m-derived filter is obtained from constant-k by scaling series arm by m and modifying shunt arm. The parameter m satisfies 0 less than m less than 1.
• Pole of attenuation: f∞ = fc / √(1 - m²). This pole gives infinite attenuation just beyond fc, creating sharp cutoff.
• To find m from pole location: m = √(1 - (fc/f∞)²). This is the most frequently asked formula in GATE problems.
• m = 0.6 is used exclusively for terminating half-sections in composite filters to ensure constant image impedance across the passband.
• Composite filter = constant-k sections (broad stopband) + m-derived section (sharp pole) + m=0.6 half-sections at ends (impedance match).
• Common trap: The cutoff frequency of the m-derived section is the same as the prototype constant-k section. Only the pole location changes, not fc.
• The shunt arm resonates at f∞ in T-section; the series arm resonates at f∞ in pi-section. Both produce the same pole of attenuation.