Skip to main content

Chapter 8 · 6 hours

Design of High-Order Active Filters

IOE past exam questions

Past questions and answers

42 questions set from this chapter, 6 of them more than once. Most asked first.

  • Asked 7 times
  • 2083 Baisakh · 2+4 marks
  • 2075 Asoj · 5 marks
  • 2073 Shrawan · 5 marks
  • 2071 Shrawan · 6 marks
  • 2070 Chaitra · 5 marks
  • 2081 Bhadra · 4 marks
  • 2080 Bhadra · 5 marks

What is GIC (Antoniou's GIC)? How can a GIC be used to simulate a grounded inductor (in the passive filter)? Explain with necessary figures and derivations.

Answer

GIC (Antoniou's GIC)

The GIC (Antoniou's generalized impedance converter) is a two-op-amp circuit with five impedances Z1…Z5Z_1 \dots Z_5 in a chain. Its input impedance is

Zin=Z1Z3Z5Z2Z4Z_{in} = \frac{Z_1Z_3Z_5}{Z_2Z_4}

By choosing which ZZ's are resistors and which are capacitors, it can simulate a grounded inductor or an FDNR.

 Iin -->
 V1 o-----+---------- A1 input
          |
          Z1
          |
          +---------- A1 output (Va)
          Z2
          |
     (3)  +---------- A1 and A2 inputs
          Z3
          |
          +---------- A2 output (Vb)
          Z4
          |
     (5)  +---------- A2 input
          Z5
          |
         GND

Op-amp A1 has its inputs at nodes 1 and 3; A2 has its inputs at nodes 3 and 5.

Derivation (ideal op-amps: virtual short gives V1=V3=V5=VV_1 = V_3 = V_5 = V; no input current):

IZ5=VZ5(flows through Z4 too)Vb=V+Z4VZ5IZ3=V−VbZ3=−Z4Z3Z5V(flows through Z2 too)Va=V+Z2IZ3=V−Z2Z4Z3Z5VIin=V−VaZ1=Z2Z4Z1Z3Z5VZin=VIin=Z1Z3Z5Z2Z4\begin{aligned} I_{Z5} &= \frac{V}{Z_5} \quad\text{(flows through } Z_4 \text{ too)} \\ V_b &= V + Z_4\frac{V}{Z_5} \\ I_{Z3} &= \frac{V - V_b}{Z_3} = -\frac{Z_4}{Z_3Z_5}V \quad\text{(flows through } Z_2 \text{ too)} \\ V_a &= V + Z_2 I_{Z3} = V - \frac{Z_2Z_4}{Z_3Z_5}V \\ I_{in} &= \frac{V - V_a}{Z_1} = \frac{Z_2Z_4}{Z_1Z_3Z_5}V \\ Z_{in} &= \frac{V}{I_{in}} = \frac{Z_1Z_3Z_5}{Z_2Z_4} \end{aligned}

Simulating a grounded inductor

Grounded inductor simulation: make Z4=1sC4Z_4 = \dfrac{1}{sC_4} and Z1,Z2,Z3,Z5Z_1, Z_2, Z_3, Z_5 resistors:

Zin=R1R3R5R2 sC4=sL,L=C4R1R3R5R2Z_{in} = \frac{R_1R_3R_5}{R_2}\,sC_4 = sL, \qquad L = \frac{C_4R_1R_3R_5}{R_2}

(Equally, Z2Z_2 can be the capacitor, giving L=C2R1R3R5/R4L = C_2R_1R_3R_5/R_4.) With all resistors equal to RR, L=R2CL = R^2C. One end of the simulated inductor is ground, so the GIC replaces the shunt (grounded) inductors of an LC ladder, e.g. in a high-pass ladder.

Example: L=0.5L = 0.5 H with C4=0.1 μC_4 = 0.1\ \muF and R1=R2=R3=R5=RR_1 = R_2 = R_3 = R_5 = R: R=L/C=0.5/10−7=2.236 kΩR = \sqrt{L/C} = \sqrt{0.5/10^{-7}} = 2.236\ \text{k}\Omega.

Advantages: the realized inductor has high Q and low sensitivity to op-amp gain; it uses only R, C and op-amps, so the low-sensitivity property of the passive LC ladder is kept.

  • Asked 2 times
  • 2078 Bhadra · 2+3+5 marks
  • 2079 Bhadra · 2+3+5 marks

What is generalized impedance converter (GIC)? How can you simulate the grounded inductor in the passive filter using GIC? Realize the following passive filter to be active simulation of grounded inductors. Use frequency scale factor Kf = 2000 and also perform the magnitude scale to get practically realizable element values in your final circuit. [Figure: high pass ladder - voltage source with series source resistor (value not legible), series 1.618 F capacitor, shunt 618 mH inductor, series 500 mF capacitor, shunt 618 mH inductor, series 1.618 F capacitor, 1 kΩ load resistor]

Answer

Generalized impedance converter (GIC)

The GIC (Antoniou's generalized impedance converter) is a two-op-amp circuit with five impedances Z1…Z5Z_1 \dots Z_5 in a chain. Its input impedance is

Zin=Z1Z3Z5Z2Z4Z_{in} = \frac{Z_1Z_3Z_5}{Z_2Z_4}

By choosing which ZZ's are resistors and which are capacitors, it can simulate a grounded inductor or an FDNR.

 Iin -->
 V1 o-----+---------- A1 input
          |
          Z1
          |
          +---------- A1 output (Va)
          Z2
          |
     (3)  +---------- A1 and A2 inputs
          Z3
          |
          +---------- A2 output (Vb)
          Z4
          |
     (5)  +---------- A2 input
          Z5
          |
         GND

Op-amp A1 has its inputs at nodes 1 and 3; A2 has its inputs at nodes 3 and 5.

Derivation (ideal op-amps: virtual short gives V1=V3=V5=VV_1 = V_3 = V_5 = V; no input current):

IZ5=VZ5(flows through Z4 too)Vb=V+Z4VZ5IZ3=V−VbZ3=−Z4Z3Z5V(flows through Z2 too)Va=V+Z2IZ3=V−Z2Z4Z3Z5VIin=V−VaZ1=Z2Z4Z1Z3Z5VZin=VIin=Z1Z3Z5Z2Z4\begin{aligned} I_{Z5} &= \frac{V}{Z_5} \quad\text{(flows through } Z_4 \text{ too)} \\ V_b &= V + Z_4\frac{V}{Z_5} \\ I_{Z3} &= \frac{V - V_b}{Z_3} = -\frac{Z_4}{Z_3Z_5}V \quad\text{(flows through } Z_2 \text{ too)} \\ V_a &= V + Z_2 I_{Z3} = V - \frac{Z_2Z_4}{Z_3Z_5}V \\ I_{in} &= \frac{V - V_a}{Z_1} = \frac{Z_2Z_4}{Z_1Z_3Z_5}V \\ Z_{in} &= \frac{V}{I_{in}} = \frac{Z_1Z_3Z_5}{Z_2Z_4} \end{aligned}

Simulating a grounded inductor with a GIC

Grounded inductor simulation: make Z4=1sC4Z_4 = \dfrac{1}{sC_4} and Z1,Z2,Z3,Z5Z_1, Z_2, Z_3, Z_5 resistors:

Zin=R1R3R5R2 sC4=sL,L=C4R1R3R5R2Z_{in} = \frac{R_1R_3R_5}{R_2}\,sC_4 = sL, \qquad L = \frac{C_4R_1R_3R_5}{R_2}

(Equally, Z2Z_2 can be the capacitor, giving L=C2R1R3R5/R4L = C_2R_1R_3R_5/R_4.) With all resistors equal to RR, L=R2CL = R^2C. One end of the simulated inductor is ground, so the GIC replaces the shunt (grounded) inductors of an LC ladder, e.g. in a high-pass ladder.

Example: L=0.5L = 0.5 H with C4=0.1 μC_4 = 0.1\ \muF and R1=R2=R3=R5=RR_1 = R_2 = R_3 = R_5 = R: R=L/C=0.5/10−7=2.236 kΩR = \sqrt{L/C} = \sqrt{0.5/10^{-7}} = 2.236\ \text{k}\Omega.

Active simulation of the given high-pass ladder

The values (C 1.618, L 0.618, C 0.5, L 0.618, C 1.618) are the 5th-order Butterworth high-pass ladder obtained from the low-pass prototype 0.618, 1.618, 2, 1.618, 0.618. Assumption: the prototype is normalized with Rs=RL=1 ΩR_s = R_L = 1\ \Omega (the source resistor is not legible); the 1 kΩ in the figure matches the impedance scaling chosen below.

Scaling: kf=2000k_f = 2000 (rad/s), and choose km=1000k_m = 1000:

R′=kmR,L′=kmkfL,C′=CkfkmR' = k_mR,\qquad L' = \frac{k_m}{k_f}L,\qquad C' = \frac{C}{k_fk_m}
ElementNormalizedScaled
RsR_s, RLR_L1 Ω1 kΩ
C1C_1, C5C_5 (series)1.618 F0.809 μF
L2L_2, L4L_4 (shunt)0.618 H0.309 H
C3C_3 (series)0.5 F0.25 μF

Replace each grounded 0.309 H inductor by a GIC with Z4=C4Z_4 = C_4: L=C4R1R3R5/R2L = C_4R_1R_3R_5/R_2. Choose C4=0.1 μC_4 = 0.1\ \muF and R2=R3=R5=1 kΩR_2 = R_3 = R_5 = 1\ \text{k}\Omega:

R1=LR2C4R3R5=0.309×100010−7×106=3.09 kΩR_1 = \frac{L R_2}{C_4R_3R_5} = \frac{0.309 \times 1000}{10^{-7}\times10^{6}} = 3.09\ \text{k}\Omega
Vs-1k-+-0.809uF-+-0.25uF-+-0.809uF-+--+
      |         |        |         |  |
              [GIC]            [GIC] 1k
              0.309H           0.309H |
                |                |    |
GND ------------+----------------+----+

Final circuit: Rs=RL=1 kΩR_s = R_L = 1\ \text{k}\Omega; series capacitors 0.809 μF, 0.25 μF, 0.809 μF; two GICs, each with R1=3.09 kΩR_1 = 3.09\ \text{k}\Omega, R2=R3=R5=1 kΩR_2 = R_3 = R_5 = 1\ \text{k}\Omega, C4=0.1 μC_4 = 0.1\ \muF. The filter is a 5th-order Butterworth high-pass with cutoff ωc=2000\omega_c = 2000 rad/s (318.3 Hz) and no inductors.

  • Asked 2 times
  • 2075 Chaitra · 1+4 marks
  • 2074 Chaitra · 1+4 marks

What is (ideal) gyrator? How can you simulate inductor using gyrator? Explain with necessary derivation.

Answer

Ideal gyrator

An ideal gyrator is a lossless two-port that converts a voltage at one port into a proportional current at the other. With gyration conductance gg (gyration resistance r=1/gr = 1/g):

I1=gV2,I2=−gV1I_1 = g V_2, \qquad I_2 = -g V_1
 I1 ->                    <- I2
 o------+  +---------+  +------o
  V1    |--| gyrator |--|    V2
 o------+  |   g     |  +------o
           +---------+

It is non-reciprocal and inverts impedance: a load ZLZ_L at port 2 appears at port 1 as Zin=1g2ZL=r2ZLZ_{in} = \dfrac{1}{g^2Z_L} = \dfrac{r^2}{Z_L}.

Simulating an inductor

Terminate port 2 with a capacitor CC. The load current is sCV2sCV_2, and it leaves port 2, so I2=−sCV2I_2 = -sCV_2.

−sCV2=−gV1  ⇒  V2=gsCV1I1=gV2=g2sCV1Zin=V1I1=sCg2=s r2C\begin{aligned} -sCV_2 &= -gV_1 \;\Rightarrow\; V_2 = \frac{g}{sC}V_1 \\ I_1 &= gV_2 = \frac{g^2}{sC}V_1 \\ Z_{in} &= \frac{V_1}{I_1} = s\frac{C}{g^2} = s\,r^2C \end{aligned}

So the input looks like an inductor

Leq=Cg2=r2CL_{eq} = \frac{C}{g^2} = r^2C
 o----+  +---------+  +----+
      |--| gyrator |--|   === C   => L = r^2 C
 o----+  +---------+  +----+

Example: r=10 kΩr = 10\ \text{k}\Omega, C=0.1 μC = 0.1\ \muF gives L=108×10−7=10L = 10^8 \times 10^{-7} = 10 H.

Practical realization: a gyrator is built with two voltage-controlled current sources of opposite sign (transconductance ±g\pm g), or with op-amps (e.g. the Riordan circuit, a form of GIC).

Grounded vs floating: one gyrator with a grounded C gives a grounded inductor. A floating inductor needs two gyrators connected back to back with a grounded capacitor between them, giving L=r2CL = r^2C between the two outer terminals.

Simulated inductors are small, have high Q and can be integrated, so a passive LC ladder can be converted to an active RC filter while keeping its low sensitivity.

  • Asked 2 times
  • 2081 Bhadra · 2+4 marks
  • 2083 Baisakh · 6 marks

What is Bruton Transformation? Design (simulate) the 4th order Butterworth low pass filter in resistively terminated lossless network with half power frequency 2,000 rad/sec and practically realizable elements using FDNR. [Table: doubly terminated Butterworth element values normalized to 1 rad/s, 1 Ω terminations: n = 2: 1.414, 1.414; n = 3: 1, 2, 1; n = 4: 0.7654, 1.848, 1.848, 0.7654; n = 5: 0.618, 1.618, 2, 1.618, 0.618]

Answer

Bruton transformation

The Bruton transformation scales every impedance of an LC ladder by 1/s1/s (i.e. by k/sk/s). The transfer function (a ratio of impedances) does not change, but the element types change:

OriginalImpedance ÷ sNew element
Resistor RR/sR/sCapacitor C=1/RC = 1/R
Inductor LLLResistor R=LR = L
Capacitor C1/(s2C)1/(s^2C)FDNR D=CD = C

The FDNR (frequency-dependent negative resistance) has impedance Z=1s2DZ = \dfrac{1}{s^2D}, which is −1ω2D-\dfrac{1}{\omega^2D} on the jωj\omega axis. All inductors become resistors and the shunt capacitors become grounded FDNRs, which are easy to build with a GIC.

Design: 4th-order Butterworth LPF, ωc=2000\omega_c = 2000 rad/s

Prototype (table, Rs=RL=1 ΩR_s = R_L = 1\ \Omega): series L1=0.7654L_1 = 0.7654, shunt C2=1.848C_2 = 1.848, series L3=1.848L_3 = 1.848, shunt C4=0.7654C_4 = 0.7654.

After the Bruton transformation (normalized):

  • Rs=1 Ω→Cs=1R_s = 1\ \Omega \to C_s = 1 F; RL=1 Ω→CL=1R_L = 1\ \Omega \to C_L = 1 F
  • L1→R1=0.7654 ΩL_1 \to R_1 = 0.7654\ \Omega; L3→R3=1.848 ΩL_3 \to R_3 = 1.848\ \Omega
  • C2→D2=1.848C_2 \to D_2 = 1.848; C4→D4=0.7654C_4 \to D_4 = 0.7654

Scaling: kf=2000k_f = 2000, choose km=5000k_m = 5000 (so CsC_s becomes 0.1 μF):

C′=Ckfkm=12000×5000=0.1 μFR1′=0.7654×5000=3.827 kΩ,R3′=1.848×5000=9.24 kΩD′=Dkf2km:D2′=1.8482×1010=9.24×10−11,D4′=3.827×10−11 F⋅s\begin{aligned} C' &= \frac{C}{k_fk_m} = \frac{1}{2000 \times 5000} = 0.1\ \mu\text{F} \\ R_1' &= 0.7654 \times 5000 = 3.827\ \text{k}\Omega, \quad R_3' = 1.848 \times 5000 = 9.24\ \text{k}\Omega \\ D' &= \frac{D}{k_f^2k_m}: \quad D_2' = \frac{1.848}{2\times10^{10}} = 9.24\times10^{-11},\quad D_4' = 3.827\times10^{-11}\ \text{F·s} \end{aligned}

FDNR by GIC: in Zin=Z1Z3Z5Z2Z4Z_{in} = \dfrac{Z_1Z_3Z_5}{Z_2Z_4}, make Z1=1sC1Z_1 = \dfrac{1}{sC_1}, Z3=1sC3Z_3 = \dfrac{1}{sC_3} and the rest resistors:

Zin=R5s2C1C3R2R4  ⇒  D=C1C3R2R4R5Z_{in} = \frac{R_5}{s^2C_1C_3R_2R_4} \;\Rightarrow\; D = \frac{C_1C_3R_2R_4}{R_5}

Take C1=C3=0.1 μC_1 = C_3 = 0.1\ \muF and R2=R5=10 kΩR_2 = R_5 = 10\ \text{k}\Omega, so D=C2R4D = C^2R_4:

R4=DC2:D2′⇒R4=9.24×10−1110−14=9.24 kΩ,D4′⇒R4=3.827 kΩR_4 = \frac{D}{C^2}: \quad D_2' \Rightarrow R_4 = \frac{9.24\times10^{-11}}{10^{-14}} = 9.24\ \text{k}\Omega, \quad D_4' \Rightarrow R_4 = 3.827\ \text{k}\Omega
Vin-+-||-+-3.83k-+--9.24k-+--+-- Vout
    |0.1uF       |        |  |
             [FDNR D2] [FDNR D4] === 0.1uF
                 |        |  |
GND -------------+--------+--+

DC path: the series 0.1 μF source capacitor blocks DC bias for the op-amps, so add large bleeding resistors (e.g. 1 MΩ) across the source capacitor and from output to ground. With equal values they keep the DC gain at about 0.5, as in the original ladder.

ElementFinal value
Source and load capacitors0.1 μF each
Series resistors3.827 kΩ, 9.24 kΩ
FDNR D2D_2 (GIC)C1=C3=0.1 μC_1 = C_3 = 0.1\ \muF, R2=R5=10R_2 = R_5 = 10 kΩ, R4=9.24R_4 = 9.24 kΩ
FDNR D4D_4 (GIC)C1=C3=0.1 μC_1 = C_3 = 0.1\ \muF, R2=R5=10R_2 = R_5 = 10 kΩ, R4=3.83R_4 = 3.83 kΩ
Bleeding resistors1 MΩ each

Answer: the inductorless filter has half-power frequency 2000 rad/s, a 4th-order Butterworth response and only practical R, C and op-amp elements.

  • Asked 2 times
  • 2082 Bhadra · 6 marks
  • 2082 Baisakh · 6 marks

Design the fourth order Butterworth low-pass filter in resistively terminated lossless realization using a leapfrog simulation. Your final design should accommodate half power frequency of 1000 rad/sec and should have practically realizable elements in it. [Table: doubly terminated Butterworth element values normalized to 1 rad/s, 1 Ω terminations: n = 2: 1.414, 1.414; n = 3: 1, 2, 1; n = 4: 0.7654, 1.848, 1.848, 0.7654; n = 5: 0.618, 1.618, 2, 1.618, 0.618]

Answer

Leapfrog (operational) simulation realises the doubly terminated LC ladder by simulating its equations rather than its elements: each reactive element is replaced by an op-amp integrator, and the integrators are coupled exactly as the ladder voltages and currents are coupled. The result keeps the low sensitivity of the passive ladder.

From the table (n = 4), the doubly terminated Butterworth prototype (1 Ω terminations, ω0=1\omega_0 = 1 rad/s) is Rs=1R_s = 1, L1=0.7654L_1 = 0.7654 H, C2=1.848C_2 = 1.848 F, L3=1.848L_3 = 1.848 H, C4=0.7654C_4 = 0.7654 F, RL=1 ΩR_L = 1\ \Omega (T-ladder: series L first, shunt C to ground).

Write the state equations of the doubly terminated ladder (normalised, Rs=RL=1 ΩR_s = R_L = 1\ \Omega). Every current is turned into a voltage by multiplying by a 1 Ω resistance, so each equation becomes an integrator:

I1=Vin−V2sL1+1V2=I1−I3sC2I3=V2−V4sL3V4=I3sC4+1Vo=V4⋅RL\begin{aligned} I_1 &= \frac{V_{in} - V_2}{sL_1 + 1} \\ V_2 &= \frac{I_1 - I_3}{sC_2} \\ I_3 &= \frac{V_2 - V_4}{sL_3} \\ V_4 &= \frac{I_3}{sC_4 + 1} \\ V_o &= V_4\cdot R_L \end{aligned}

The first and last equations are lossy integrators (the terminations give the "+1"); the middle ones are lossless integrators. Neighbouring integrators feed each other in alternate directions, which gives the "leapfrog" structure.

Using inverting op-amp integrators, Vo=−1sRC∑ViV_o = -\frac{1}{sRC}\sum V_i (lossless) or Vo=−∑Vi/RsC+1/RV_o = -\frac{\sum V_i/R}{sC + 1/R} (lossy, resistor RR across CC), the signs are arranged as follows. Inverters supply the signals of opposite sign:

Op-ampTypeInputs (each through RR)OutputNormalised CC (R = 1 Ω)
A1lossy integratorVinV_{in}, −V2-V_2−I1-I_10.7654 F
A2integrator−I1-I_1, I3I_3V2V_21.848 F
A3inverterV2V_2−V2-V_2R / R
A4integrator−V2-V_2, V4V_4I3I_31.848 F
A5lossy integratorI3I_3−V4-V_40.7654 F
A6inverter−V4-V_4V4=VoV_4 = V_oR / R

Check of one stage: A2 gives −(−I1)+I3sCA2=I1−I3s(1.848)-\frac{(-I_1) + I_3}{sC_{A2}} = \frac{I_1 - I_3}{s(1.848)}, which is the ladder equation for V2V_2.

Denormalisation. Frequency scale by ω0=1000 rad/s\omega_0 = 1000\ \text{rad/s} and choose all resistors RR = 10 kΩ (impedance scale kz=104k_z = 10^4). Each capacitor becomes C=g/(ω0R)C = g/(\omega_0 R):

CA1=0.76541000×104=76.54 nFCA2=1.8481000×104=184.8 nFCA4=1.8481000×104=184.8 nFCA5=0.76541000×104=76.54 nF\begin{aligned} C_{A1} &= \frac{0.7654}{1000 \times 10^4} = \text{76.54 nF} \\ C_{A2} &= \frac{1.848}{1000 \times 10^4} = \text{184.8 nF} \\ C_{A4} &= \frac{1.848}{1000 \times 10^4} = \text{184.8 nF} \\ C_{A5} &= \frac{0.7654}{1000 \times 10^4} = \text{76.54 nF} \end{aligned}

Final element values:

Op-ampResistorsCapacitor
A1input(s) 10 kΩ, feedback 10 kΩ across C76.54 nF
A2input(s) 10 kΩ184.8 nF
A4input(s) 10 kΩ184.8 nF
A5input(s) 10 kΩ, feedback 10 kΩ across C76.54 nF
inverters10 kΩ in, 10 kΩ feedbacknone
 Lossy integrator (A1, last stage)
           +----[ R ]----+
           +----| C |----+
 Va--[R]--+|             |
 Vb--[R]--++----(-)      |
                (+) A ---+---- Vo
                 |
                GND
 Lossless integrator: same circuit without the
 feedback resistor (only C from output to (-)).

The passband gain is 1/2 (−6 dB), the same as the doubly terminated prototype; the response is that of the passive ladder but with no inductors, and the low sensitivity of the ladder is kept.

Answer: six op-amps (four integrators, two inverters), all resistors 10 kΩ, integrator capacitors 76.54 nF, 184.8 nF, 184.8 nF and 76.54 nF; half-power frequency 1000 rad/s.

  • Asked 2 times
  • 2081 Baisakh · 8 marks
  • 2081 Bhadra · 7 marks

Design (simulate) the 4th order Butterworth LPF in doubly-terminated (resistively-terminated lossless) network using Leapfrog simulation. The necessary information is listed in the given table: Order (n) = 4 and LPF; R1 = 1; L1 = 0.7654; C2 = 1.848; L3 = 1.848; C4 = 0.7654; R2 = 1

Answer

Leapfrog (operational) simulation realises the doubly terminated LC ladder by simulating its equations rather than its elements: each reactive element is replaced by an op-amp integrator, and the integrators are coupled exactly as the ladder voltages and currents are coupled. The result keeps the low sensitivity of the passive ladder.

Given prototype (normalised to 1 rad/s, 1 Ω terminations): R1=1R_1 = 1, L1=0.7654L_1 = 0.7654 H, C2=1.848C_2 = 1.848 F, L3=1.848L_3 = 1.848 H, C4=0.7654C_4 = 0.7654 F, R2=1 ΩR_2 = 1\ \Omega.

 Vin  1 Ohm  L1=0.7654   L3=1.848
 o---[R1]---(LLL)--+---(LLL)--+------o Vo
                   |          |    |
               C2=1.848  C4=0.7654 [R2=1]
                   |          |    |
 GND --------------+----------+----+

Write the state equations of the doubly terminated ladder (normalised, Rs=RL=1 ΩR_s = R_L = 1\ \Omega). Every current is turned into a voltage by multiplying by a 1 Ω resistance, so each equation becomes an integrator:

I1=Vin−V2sL1+1V2=I1−I3sC2I3=V2−V4sL3V4=I3sC4+1Vo=V4⋅RL\begin{aligned} I_1 &= \frac{V_{in} - V_2}{sL_1 + 1} \\ V_2 &= \frac{I_1 - I_3}{sC_2} \\ I_3 &= \frac{V_2 - V_4}{sL_3} \\ V_4 &= \frac{I_3}{sC_4 + 1} \\ V_o &= V_4\cdot R_L \end{aligned}

The first and last equations are lossy integrators (the terminations give the "+1"); the middle ones are lossless integrators. Neighbouring integrators feed each other in alternate directions, which gives the "leapfrog" structure.

Using inverting op-amp integrators, Vo=−1sRC∑ViV_o = -\frac{1}{sRC}\sum V_i (lossless) or Vo=−∑Vi/RsC+1/RV_o = -\frac{\sum V_i/R}{sC + 1/R} (lossy, resistor RR across CC), the signs are arranged as follows. Inverters supply the signals of opposite sign:

Op-ampTypeInputs (each through RR)OutputNormalised CC (R = 1 Ω)
A1lossy integratorVinV_{in}, −V2-V_2−I1-I_10.7654 F
A2integrator−I1-I_1, I3I_3V2V_21.848 F
A3inverterV2V_2−V2-V_2R / R
A4integrator−V2-V_2, V4V_4I3I_31.848 F
A5lossy integratorI3I_3−V4-V_40.7654 F
A6inverter−V4-V_4V4=VoV_4 = V_oR / R

Check of one stage: A2 gives −(−I1)+I3sCA2=I1−I3s(1.848)-\frac{(-I_1) + I_3}{sC_{A2}} = \frac{I_1 - I_3}{s(1.848)}, which is the ladder equation for V2V_2.

Denormalisation. Frequency scale by ω0=2π(1000)=6283.2 rad/s\omega_0 = 2\pi(1000) = 6283.2\ \text{rad/s} and choose all resistors RR = 10 kΩ (impedance scale kz=104k_z = 10^4). Each capacitor becomes C=g/(ω0R)C = g/(\omega_0 R):

CA1=0.76546283.19×104=12.18 nFCA2=1.8486283.19×104=29.41 nFCA4=1.8486283.19×104=29.41 nFCA5=0.76546283.19×104=12.18 nF\begin{aligned} C_{A1} &= \frac{0.7654}{6283.19 \times 10^4} = \text{12.18 nF} \\ C_{A2} &= \frac{1.848}{6283.19 \times 10^4} = \text{29.41 nF} \\ C_{A4} &= \frac{1.848}{6283.19 \times 10^4} = \text{29.41 nF} \\ C_{A5} &= \frac{0.7654}{6283.19 \times 10^4} = \text{12.18 nF} \end{aligned}

Final element values:

Op-ampResistorsCapacitor
A1input(s) 10 kΩ, feedback 10 kΩ across C12.18 nF
A2input(s) 10 kΩ29.41 nF
A4input(s) 10 kΩ29.41 nF
A5input(s) 10 kΩ, feedback 10 kΩ across C12.18 nF
inverters10 kΩ in, 10 kΩ feedbacknone
 Lossy integrator (A1, last stage)
           +----[ R ]----+
           +----| C |----+
 Va--[R]--+|             |
 Vb--[R]--++----(-)      |
                (+) A ---+---- Vo
                 |
                GND
 Lossless integrator: same circuit without the
 feedback resistor (only C from output to (-)).

The passband gain is 1/2 (−6 dB), the same as the doubly terminated prototype; the response is that of the passive ladder but with no inductors, and the low sensitivity of the ladder is kept.

Since no frequency is given, the normalised design (all R = 1 Ω, C = 0.7654, 1.848, 1.848, 0.7654 F) is the answer; the 1 kHz values above show how it is made practical.

Answer (1 kHz example): R = 10 kΩ everywhere, capacitors 12.18 nF, 29.41 nF, 29.41 nF, 12.18 nF.

  • 2081 Bhadra · 5 marks

What is GIC? How can it be used to avoid shunt inductors in LC ladder circuit?

Answer

A generalized impedance converter (GIC) is an active two-port, usually Antoniou's two-op-amp, five-impedance circuit. Its input impedance is the load impedance Z5Z_5 multiplied by a conversion function k(s)=Z1Z3/(Z2Z4)k(s) = Z_1Z_3/(Z_2Z_4). By choosing which positions hold capacitors, a GIC can make an inductor or an FDNR from only R, C and op-amps.

GIC circuit and input impedance

   Iin  node 1
 o-->-----+--------------(+) A1
 Vin      |
         [Z1]
          +--- node 2 ---- output of A1
         [Z2]
          +--- node 3 ---- (-) A1 and (-) A2
         [Z3]
          +--- node 4 ---- output of A2
         [Z4]
          +--- node 5 ---- (+) A2
         [Z5]
          |
         GND

With ideal op-amps the two input terminals of each op-amp are at the same voltage and draw no current, so V1=V3=V5=VinV_1 = V_3 = V_5 = V_{in}.

Node 5: V4−VinZ4=VinZ5  ⇒  V4=Vin(1+Z4Z5)Node 3: V2−VinZ2=Vin−V4Z3=−Z4Z3Z5Vin  ⇒  V2=Vin(1−Z2Z4Z3Z5)Node 1: Iin=Vin−V2Z1=Z2Z4Z1Z3Z5Vin\begin{aligned} \text{Node 5: } & \frac{V_4 - V_{in}}{Z_4} = \frac{V_{in}}{Z_5} \;\Rightarrow\; V_4 = V_{in}\left(1 + \frac{Z_4}{Z_5}\right) \\ \text{Node 3: } & \frac{V_2 - V_{in}}{Z_2} = \frac{V_{in} - V_4}{Z_3} = -\frac{Z_4}{Z_3 Z_5}V_{in} \;\Rightarrow\; V_2 = V_{in}\left(1 - \frac{Z_2 Z_4}{Z_3 Z_5}\right) \\ \text{Node 1: } & I_{in} = \frac{V_{in} - V_2}{Z_1} = \frac{Z_2 Z_4}{Z_1 Z_3 Z_5}V_{in} \end{aligned} Zin=VinIin=Z1Z3Z5Z2Z4Z_{in} = \frac{V_{in}}{I_{in}} = \frac{Z_1 Z_3 Z_5}{Z_2 Z_4}

Avoiding shunt inductors

Put a capacitor in position 4 (Z4=1/sCZ_4 = 1/sC) and resistors elsewhere:

Zin=R1R3R5R2 sC=sL,L=CR1R3R5R2Z_{in} = \frac{R_1 R_3 R_5}{R_2}\, sC = sL, \qquad L = \frac{C R_1 R_3 R_5}{R_2}

With all resistors equal to RR: L=CR2L = CR^2. One end of this inductor is ground, so it replaces a grounded (shunt) inductor.

In an LC ladder the shunt inductors (for example in a highpass ladder, where every inductor goes from a node to ground) are exactly such grounded inductors. The procedure is:

  1. Obtain the passive ladder (prototype values, then frequency and impedance scaling).
  2. Keep the resistors and capacitors as they are.
  3. Replace each shunt inductor LL by a GIC: choose CC (e.g. 10 nF) and R=L/CR = \sqrt{L/C}.
 o--[Rs]--||--+--||--+--||--o Vo
              |      |      |
            [GIC]  [GIC]   [RL]
              |      |      |
 GND ---------+------+------+

Example: a shunt inductor of 0.1 H with CC = 10 nF needs R=0.1/10−8R = \sqrt{0.1/10^{-8}} = 3.162 kΩ in positions 1, 2, 3 and 5.

Advantages: no bulky, lossy inductors; the ladder's low sensitivity is kept; elements are easy to integrate. Floating (series) inductors cannot be replaced by one grounded GIC; for those, use two GICs or the Bruton/FDNR method.

  • 2079 Baisakh · 5 marks

What is GIC? How GIC can be used to simulate the floating inductor in the passive filter? Explain.

Answer

A generalized impedance converter (GIC) is an active two-port, usually Antoniou's two-op-amp, five-impedance circuit. Its input impedance is the load impedance Z5Z_5 multiplied by a conversion function k(s)=Z1Z3/(Z2Z4)k(s) = Z_1Z_3/(Z_2Z_4). By choosing which positions hold capacitors, a GIC can make an inductor or an FDNR from only R, C and op-amps.

Input impedance of the GIC

   Iin  node 1
 o-->-----+--------------(+) A1
 Vin      |
         [Z1]
          +--- node 2 ---- output of A1
         [Z2]
          +--- node 3 ---- (-) A1 and (-) A2
         [Z3]
          +--- node 4 ---- output of A2
         [Z4]
          +--- node 5 ---- (+) A2
         [Z5]
          |
         GND

With ideal op-amps the two input terminals of each op-amp are at the same voltage and draw no current, so V1=V3=V5=VinV_1 = V_3 = V_5 = V_{in}.

Node 5: V4−VinZ4=VinZ5  ⇒  V4=Vin(1+Z4Z5)Node 3: V2−VinZ2=Vin−V4Z3=−Z4Z3Z5Vin  ⇒  V2=Vin(1−Z2Z4Z3Z5)Node 1: Iin=Vin−V2Z1=Z2Z4Z1Z3Z5Vin\begin{aligned} \text{Node 5: } & \frac{V_4 - V_{in}}{Z_4} = \frac{V_{in}}{Z_5} \;\Rightarrow\; V_4 = V_{in}\left(1 + \frac{Z_4}{Z_5}\right) \\ \text{Node 3: } & \frac{V_2 - V_{in}}{Z_2} = \frac{V_{in} - V_4}{Z_3} = -\frac{Z_4}{Z_3 Z_5}V_{in} \;\Rightarrow\; V_2 = V_{in}\left(1 - \frac{Z_2 Z_4}{Z_3 Z_5}\right) \\ \text{Node 1: } & I_{in} = \frac{V_{in} - V_2}{Z_1} = \frac{Z_2 Z_4}{Z_1 Z_3 Z_5}V_{in} \end{aligned} Zin=VinIin=Z1Z3Z5Z2Z4Z_{in} = \frac{V_{in}}{I_{in}} = \frac{Z_1 Z_3 Z_5}{Z_2 Z_4}

Grounded inductor

With Z4=1/sCZ_4 = 1/sC and all other positions resistors RR: Zin=sCR2Z_{in} = sCR^2, i.e. L=CR2L = CR^2. This works only when one end of the inductor is grounded.

Floating inductor using two GICs

A series (floating) inductor, as in a lowpass ladder, has neither end grounded. It is simulated by two identical GICs connected back to back with a resistor RxR_x between their port-2 terminals:

  a o--[GIC 1]--o x --[ Rx ]-- y o--[GIC 2]--o b
      port1  port2             port2  port1
      (GIC 2 is the same circuit, reversed)

Treat the GIC with Z5Z_5 removed as a two-port (port 2 at node 5). From the analysis above, V1=V2V_1 = V_2 and I1=I2/k(s)I_1 = I_2/k(s), with k(s)=Z1Z3/(Z2Z4)k(s) = Z_1Z_3/(Z_2Z_4). With Z4=1/sCZ_4 = 1/sC and equal resistors, k(s)=sCRk(s) = sCR. In transmission (ABCD) form:

GIC 1: [1001/k],series Rx:[1Rx01],GIC 2 reversed: [100k]\text{GIC 1: } \begin{bmatrix} 1 & 0 \\ 0 & 1/k \end{bmatrix}, \quad \text{series } R_x: \begin{bmatrix} 1 & R_x \\ 0 & 1 \end{bmatrix}, \quad \text{GIC 2 reversed: } \begin{bmatrix} 1 & 0 \\ 0 & k \end{bmatrix} [1001/k][1Rx01][100k]=[1kRx01]\begin{bmatrix} 1 & 0 \\ 0 & 1/k \end{bmatrix}\begin{bmatrix} 1 & R_x \\ 0 & 1 \end{bmatrix}\begin{bmatrix} 1 & 0 \\ 0 & k \end{bmatrix} = \begin{bmatrix} 1 & kR_x \\ 0 & 1 \end{bmatrix}

This is the matrix of a single series (floating) impedance Z=k(s)Rx=s (CRRx)Z = k(s)R_x = s\,(CRR_x), so between a and b we get a floating inductor

L=CRRxL = C R R_x

Example: CC = 10 nF, R=RxR = R_x = 10 kΩ gives L=10−8×108L = 10^{-8} \times 10^8 = 1 H. The two GICs must be well matched; a mismatch leaves a small unwanted grounded element. (The Bruton/FDNR method is the usual alternative for lowpass ladders.)

  • 2072 Chaitra · 5 marks

Draw the circuit diagram of a generalized impedance converter. Derive the relationship between input and output current. How can it be used to simulate a grounded FDNR? Explain.

Answer

GIC circuit

The generalized impedance converter (GIC) (Antoniou circuit) uses two op-amps and five impedances in a chain from the input to ground:

   Iin  node 1
 o-->-----+--------------(+) A1
 Vin      |
         [Z1]
          +--- node 2 ---- output of A1
         [Z2]
          +--- node 3 ---- (-) A1 and (-) A2
         [Z3]
          +--- node 4 ---- output of A2
         [Z4]
          +--- node 5 ---- (+) A2
         [Z5]
          |
         GND

Relation between input and output current

Treat node 1 as port 1 (V1V_1, I1I_1) and node 5 as port 2, where the load Z5Z_5 draws current I2I_2 (V2=I2Z5V_2 = I_2 Z_5). Ideal op-amps force node 1, node 3 and node 5 to the same voltage and draw no input current, so V1=V2V_1 = V_2. Let VN2V_{N2} and VN4V_{N4} be the voltages of nodes 2 and 4 (op-amp outputs).

Node 5: VN4=V2+Z4I2Node 3: VN2−V1Z2=V1−VN4Z3=−Z4Z3I2Node 1: I1=V1−VN2Z1=Z2Z4Z1Z3 I2\begin{aligned} \text{Node 5: } & V_{N4} = V_2 + Z_4 I_2 \\ \text{Node 3: } & \frac{V_{N2} - V_1}{Z_2} = \frac{V_1 - V_{N4}}{Z_3} = -\frac{Z_4}{Z_3}I_2 \\ \text{Node 1: } & I_1 = \frac{V_1 - V_{N2}}{Z_1} = \frac{Z_2 Z_4}{Z_1 Z_3}\, I_2 \end{aligned}

So the GIC passes voltage unchanged and scales current:

V1=V2,I1=Z2Z4Z1Z3 I2,Zin=V1I1=Z1Z3Z2Z4 Z5=k(s) Z5V_1 = V_2, \qquad I_1 = \frac{Z_2Z_4}{Z_1Z_3}\, I_2, \qquad Z_{in} = \frac{V_1}{I_1} = \frac{Z_1Z_3}{Z_2Z_4}\,Z_5 = k(s)\,Z_5

Grounded FDNR

Put capacitors in positions 1 and 5 (Z1=1/sC1Z_1 = 1/sC_1, Z5=1/sC5Z_5 = 1/sC_5) and resistors in 2, 3, 4:

Zin=R3s2C1C5R2R4=1s2D,D=C1C5R2R4R3Z_{in} = \frac{R_3}{s^2 C_1 C_5 R_2 R_4} = \frac{1}{s^2 D}, \qquad D = \frac{C_1 C_5 R_2 R_4}{R_3}

With C1=C5=CC_1 = C_5 = C and R2=R3=R4=RR_2 = R_3 = R_4 = R: D=C2RD = C^2 R. At s=jωs = j\omega, Zin=−1/(ω2D)Z_{in} = -1/(\omega^2 D), a negative resistance whose value depends on frequency: an FDNR (also called a D-element or supercapacitor).

Practical design: choose CC (e.g. 10 nF), then R=D/C2R = D/C^2. Example: D=10−12 F2ΩD = 10^{-12}\ \text{F}^2\Omega with CC = 10 nF gives R=10−12/10−16R = 10^{-12}/10^{-16} = 10 kΩ. Because port 1 is referred to ground, this FDNR is grounded, which is exactly what the Bruton-transformed lowpass ladder needs (shunt capacitors become grounded FDNRs).

  • 2074 Asoj · 4+4 marks

What is generalized impedance converter (GIC)? Explain how inductors can be simulated using GIC? Simulate the following highpass filter by active simulation of grounded inductors such that ω0 is 4000 rad/s and practically realizable elements. [Figure: source V1, 1 Ω source resistor, series 1.618 F, shunt 0.618 H, series 0.5 F, shunt 0.618 H, series 1.618 F, 1 Ω load (output V2)]

Answer

A generalized impedance converter (GIC) is an active two-port, usually Antoniou's two-op-amp, five-impedance circuit. Its input impedance is the load impedance Z5Z_5 multiplied by a conversion function k(s)=Z1Z3/(Z2Z4)k(s) = Z_1Z_3/(Z_2Z_4). By choosing which positions hold capacitors, a GIC can make an inductor or an FDNR from only R, C and op-amps.

Simulating a grounded inductor with a GIC

   Iin  node 1
 o-->-----+--------------(+) A1
 Vin      |
         [Z1]
          +--- node 2 ---- output of A1
         [Z2]
          +--- node 3 ---- (-) A1 and (-) A2
         [Z3]
          +--- node 4 ---- output of A2
         [Z4]
          +--- node 5 ---- (+) A2
         [Z5]
          |
         GND

With ideal op-amps the two input terminals of each op-amp are at the same voltage and draw no current, so V1=V3=V5=VinV_1 = V_3 = V_5 = V_{in}.

Node 5: V4−VinZ4=VinZ5  ⇒  V4=Vin(1+Z4Z5)Node 3: V2−VinZ2=Vin−V4Z3=−Z4Z3Z5Vin  ⇒  V2=Vin(1−Z2Z4Z3Z5)Node 1: Iin=Vin−V2Z1=Z2Z4Z1Z3Z5Vin\begin{aligned} \text{Node 5: } & \frac{V_4 - V_{in}}{Z_4} = \frac{V_{in}}{Z_5} \;\Rightarrow\; V_4 = V_{in}\left(1 + \frac{Z_4}{Z_5}\right) \\ \text{Node 3: } & \frac{V_2 - V_{in}}{Z_2} = \frac{V_{in} - V_4}{Z_3} = -\frac{Z_4}{Z_3 Z_5}V_{in} \;\Rightarrow\; V_2 = V_{in}\left(1 - \frac{Z_2 Z_4}{Z_3 Z_5}\right) \\ \text{Node 1: } & I_{in} = \frac{V_{in} - V_2}{Z_1} = \frac{Z_2 Z_4}{Z_1 Z_3 Z_5}V_{in} \end{aligned} Zin=VinIin=Z1Z3Z5Z2Z4Z_{in} = \frac{V_{in}}{I_{in}} = \frac{Z_1 Z_3 Z_5}{Z_2 Z_4}

Put a capacitor in position 4 (Z4=1/sCZ_4 = 1/sC) and resistors elsewhere:

Zin=R1R3R5R2 sC=sL,L=CR1R3R5R2Z_{in} = \frac{R_1 R_3 R_5}{R_2}\, sC = sL, \qquad L = \frac{C R_1 R_3 R_5}{R_2}

With all resistors equal to RR: L=CR2L = CR^2. One end of this inductor is ground, so it replaces a grounded (shunt) inductor.

Design

The given ladder is a 5th-order Butterworth highpass prototype (ω0=1\omega_0 = 1 rad/s, 1 Ω terminations): series 1.618 F, shunt 0.618 H, series 0.5 F, shunt 0.618 H, series 1.618 F. Both inductors are grounded, so each is replaced by a GIC.

Frequency scale to ω0=4000 rad/s\omega_0 = 4000\ \text{rad/s} and impedance scale by kzk_z (chosen as 10410^4 so that capacitors are in the nF range):

Rnew=kzR,Lnew=kzLω0,Cnew=Ckz ω0R_{new} = k_z R, \qquad L_{new} = \frac{k_z L}{\omega_0}, \qquad C_{new} = \frac{C}{k_z\,\omega_0}

kz=10000 Ωk_z = 10000\ \Omega, so the source and load resistors become Rs=RLR_s = R_L = 10 kΩ.

PrototypeScaledFinal
C1C_1 = 1.618 F (series)C=1.618/(kzω0)C = 1.618/(k_z\omega_0)40.45 nF
L2L_2 = 0.618 H (shunt)kz(0.618)/ω0k_z(0.618)/\omega_0 = 1.545 HGIC: C = 10 nF, R=L/CR = \sqrt{L/C} = 12.43 kΩ
C3C_3 = 0.5 F (series)C=0.5/(kzω0)C = 0.5/(k_z\omega_0)12.5 nF
L4L_4 = 0.618 H (shunt)kz(0.618)/ω0k_z(0.618)/\omega_0 = 1.545 HGIC: C = 10 nF, R=L/CR = \sqrt{L/C} = 12.43 kΩ
C5C_5 = 1.618 F (series)C=1.618/(kzω0)C = 1.618/(k_z\omega_0)40.45 nF
Rs,RLR_s, R_L = 1 Ωkz×1k_z \times 110 kΩ

Each grounded inductor is replaced by a GIC with Z4=CZ_4 = C and R1=R2=R3=R5=RR_1 = R_2 = R_3 = R_5 = R, so L=CR2L = CR^2 and R=L/CR = \sqrt{L/C}.

Final circuit (each GIC realises the grounded inductor in that position):

Vi o--[Rs]--[C1]-+----[C3]-+----[C5]-+----o Vo
                 |         |         |
              [GIC-L2]  [GIC-L4]    [RL]
GND ------------------------------------------

Answer: Rs=RLR_s = R_L = 10 kΩ; series capacitors 40.45 nF, 12.5 nF, 40.45 nF; each simulated inductor 1.545 H = GIC with C = 10 nF and four resistors of 12.43 kΩ; ω0\omega_0 = 4000 rad/s.

  • 2081 Baisakh · 1+3+5 marks

What is a generalized impedance converter (GIC)? How can you simulate the grounded inductor using GIC? From the LC ladder given in figure below, design a highpass filter with a half power frequency of 5 kHz and the largest capacitance of 10nF using inductor simulation. [Figure: source V1, 1 Ω source resistor, series 1.618 F, shunt 0.618 H, series 0.5 F, shunt 0.618 H, series 1.618 F, 1 Ω load (output V2)]

Answer

A generalized impedance converter (GIC) is an active two-port, usually Antoniou's two-op-amp, five-impedance circuit. Its input impedance is the load impedance Z5Z_5 multiplied by a conversion function k(s)=Z1Z3/(Z2Z4)k(s) = Z_1Z_3/(Z_2Z_4). By choosing which positions hold capacitors, a GIC can make an inductor or an FDNR from only R, C and op-amps.

Simulating a grounded inductor with a GIC

   Iin  node 1
 o-->-----+--------------(+) A1
 Vin      |
         [Z1]
          +--- node 2 ---- output of A1
         [Z2]
          +--- node 3 ---- (-) A1 and (-) A2
         [Z3]
          +--- node 4 ---- output of A2
         [Z4]
          +--- node 5 ---- (+) A2
         [Z5]
          |
         GND

With ideal op-amps the two input terminals of each op-amp are at the same voltage and draw no current, so V1=V3=V5=VinV_1 = V_3 = V_5 = V_{in}.

Node 5: V4−VinZ4=VinZ5  ⇒  V4=Vin(1+Z4Z5)Node 3: V2−VinZ2=Vin−V4Z3=−Z4Z3Z5Vin  ⇒  V2=Vin(1−Z2Z4Z3Z5)Node 1: Iin=Vin−V2Z1=Z2Z4Z1Z3Z5Vin\begin{aligned} \text{Node 5: } & \frac{V_4 - V_{in}}{Z_4} = \frac{V_{in}}{Z_5} \;\Rightarrow\; V_4 = V_{in}\left(1 + \frac{Z_4}{Z_5}\right) \\ \text{Node 3: } & \frac{V_2 - V_{in}}{Z_2} = \frac{V_{in} - V_4}{Z_3} = -\frac{Z_4}{Z_3 Z_5}V_{in} \;\Rightarrow\; V_2 = V_{in}\left(1 - \frac{Z_2 Z_4}{Z_3 Z_5}\right) \\ \text{Node 1: } & I_{in} = \frac{V_{in} - V_2}{Z_1} = \frac{Z_2 Z_4}{Z_1 Z_3 Z_5}V_{in} \end{aligned} Zin=VinIin=Z1Z3Z5Z2Z4Z_{in} = \frac{V_{in}}{I_{in}} = \frac{Z_1 Z_3 Z_5}{Z_2 Z_4}

Put a capacitor in position 4 (Z4=1/sCZ_4 = 1/sC) and resistors elsewhere:

Zin=R1R3R5R2 sC=sL,L=CR1R3R5R2Z_{in} = \frac{R_1 R_3 R_5}{R_2}\, sC = sL, \qquad L = \frac{C R_1 R_3 R_5}{R_2}

With all resistors equal to RR: L=CR2L = CR^2. One end of this inductor is ground, so it replaces a grounded (shunt) inductor.

Design

The ladder is the 5th-order Butterworth highpass prototype (1 rad/s, 1 Ω): series 1.618 F, shunt 0.618 H, series 0.5 F, shunt 0.618 H, series 1.618 F. The shunt inductors are grounded and are simulated by GICs.

Frequency scale to ω0=2π(5000)=31,416 rad/s\omega_0 = 2\pi(5000) = 31{,}416\ \text{rad/s} and impedance scale by kzk_z (chosen so that the largest capacitor, 1.618 F, becomes 10 nF: kz=1.618/(ω0×10−8)k_z = 1.618/(\omega_0 \times 10^{-8})):

Rnew=kzR,Lnew=kzLω0,Cnew=Ckz ω0R_{new} = k_z R, \qquad L_{new} = \frac{k_z L}{\omega_0}, \qquad C_{new} = \frac{C}{k_z\,\omega_0}

kz=5150.3 Ωk_z = 5150.3\ \Omega, so the source and load resistors become Rs=RLR_s = R_L = 5.15 kΩ.

PrototypeScaledFinal
C1C_1 = 1.618 F (series)C=1.618/(kzω0)C = 1.618/(k_z\omega_0)10 nF
L2L_2 = 0.618 H (shunt)kz(0.618)/ω0k_z(0.618)/\omega_0 = 101.3 mHGIC: C = 10 nF, R=L/CR = \sqrt{L/C} = 3.183 kΩ
C3C_3 = 0.5 F (series)C=0.5/(kzω0)C = 0.5/(k_z\omega_0)3.09 nF
L4L_4 = 0.618 H (shunt)kz(0.618)/ω0k_z(0.618)/\omega_0 = 101.3 mHGIC: C = 10 nF, R=L/CR = \sqrt{L/C} = 3.183 kΩ
C5C_5 = 1.618 F (series)C=1.618/(kzω0)C = 1.618/(k_z\omega_0)10 nF
Rs,RLR_s, R_L = 1 Ωkz×1k_z \times 15.15 kΩ

Each grounded inductor is replaced by a GIC with Z4=CZ_4 = C and R1=R2=R3=R5=RR_1 = R_2 = R_3 = R_5 = R, so L=CR2L = CR^2 and R=L/CR = \sqrt{L/C}.

Final circuit (each GIC realises the grounded inductor in that position):

Vi o--[Rs]--[C1]-+----[C3]-+----[C5]-+----o Vo
                 |         |         |
              [GIC-L2]  [GIC-L4]    [RL]
GND ------------------------------------------

Answer: Rs=RLR_s = R_L = 5.15 kΩ; capacitors 10 nF, 3.09 nF, 10 nF; each 101.3 mH inductor = GIC with C = 10 nF and R = 3.183 kΩ; half-power frequency 5 kHz.

  • 2069 Chaitra · 2+4+6 marks

What is generalized impedance converter (GIC)? How can you simulate the grounded inductor in the passive filter using GIC? Explain. The following circuit is a high pass filter having half power frequency of 1 rad/sec. Design a high pass filter having half power frequency of 4.5 kHz by active simulation of inductors. In your final circuit the largest capacitance should be 0.1 μF. [Figure: source V1, 1 Ω source resistor, shunt 1.618 H, series 0.618 F, shunt 0.5 H, series 0.618 F, shunt 1.618 H, 1 Ω load (output V2)]

Answer

A generalized impedance converter (GIC) is an active two-port, usually Antoniou's two-op-amp, five-impedance circuit. Its input impedance is the load impedance Z5Z_5 multiplied by a conversion function k(s)=Z1Z3/(Z2Z4)k(s) = Z_1Z_3/(Z_2Z_4). By choosing which positions hold capacitors, a GIC can make an inductor or an FDNR from only R, C and op-amps.

Simulating a grounded inductor with a GIC

   Iin  node 1
 o-->-----+--------------(+) A1
 Vin      |
         [Z1]
          +--- node 2 ---- output of A1
         [Z2]
          +--- node 3 ---- (-) A1 and (-) A2
         [Z3]
          +--- node 4 ---- output of A2
         [Z4]
          +--- node 5 ---- (+) A2
         [Z5]
          |
         GND

With ideal op-amps the two input terminals of each op-amp are at the same voltage and draw no current, so V1=V3=V5=VinV_1 = V_3 = V_5 = V_{in}.

Node 5: V4−VinZ4=VinZ5  ⇒  V4=Vin(1+Z4Z5)Node 3: V2−VinZ2=Vin−V4Z3=−Z4Z3Z5Vin  ⇒  V2=Vin(1−Z2Z4Z3Z5)Node 1: Iin=Vin−V2Z1=Z2Z4Z1Z3Z5Vin\begin{aligned} \text{Node 5: } & \frac{V_4 - V_{in}}{Z_4} = \frac{V_{in}}{Z_5} \;\Rightarrow\; V_4 = V_{in}\left(1 + \frac{Z_4}{Z_5}\right) \\ \text{Node 3: } & \frac{V_2 - V_{in}}{Z_2} = \frac{V_{in} - V_4}{Z_3} = -\frac{Z_4}{Z_3 Z_5}V_{in} \;\Rightarrow\; V_2 = V_{in}\left(1 - \frac{Z_2 Z_4}{Z_3 Z_5}\right) \\ \text{Node 1: } & I_{in} = \frac{V_{in} - V_2}{Z_1} = \frac{Z_2 Z_4}{Z_1 Z_3 Z_5}V_{in} \end{aligned} Zin=VinIin=Z1Z3Z5Z2Z4Z_{in} = \frac{V_{in}}{I_{in}} = \frac{Z_1 Z_3 Z_5}{Z_2 Z_4}

Put a capacitor in position 4 (Z4=1/sCZ_4 = 1/sC) and resistors elsewhere:

Zin=R1R3R5R2 sC=sL,L=CR1R3R5R2Z_{in} = \frac{R_1 R_3 R_5}{R_2}\, sC = sL, \qquad L = \frac{C R_1 R_3 R_5}{R_2}

With all resistors equal to RR: L=CR2L = CR^2. One end of this inductor is ground, so it replaces a grounded (shunt) inductor.

Design

The given ladder is a 5th-order Butterworth highpass (1 rad/s, 1 Ω): shunt 1.618 H, series 0.618 F, shunt 0.5 H, series 0.618 F, shunt 1.618 H. All three inductors are grounded, so three GICs are needed.

Frequency scale to ω0=2π(4500)=28,274 rad/s\omega_0 = 2\pi(4500) = 28{,}274\ \text{rad/s} and impedance scale by kzk_z (chosen so that the largest capacitor, 0.618 F, becomes 0.1 μF: kz=0.618/(ω0×10−7)k_z = 0.618/(\omega_0 \times 10^{-7})):

Rnew=kzR,Lnew=kzLω0,Cnew=Ckz ω0R_{new} = k_z R, \qquad L_{new} = \frac{k_z L}{\omega_0}, \qquad C_{new} = \frac{C}{k_z\,\omega_0}

kz=218.57 Ωk_z = 218.57\ \Omega, so the source and load resistors become Rs=RLR_s = R_L = 218.6 Ω.

PrototypeScaledFinal
L1L_1 = 1.618 H (shunt)kz(1.618)/ω0k_z(1.618)/\omega_0 = 12.51 mHGIC: C = 100 nF, R=L/CR = \sqrt{L/C} = 353.7 Ω
C2C_2 = 0.618 F (series)C=0.618/(kzω0)C = 0.618/(k_z\omega_0)100 nF
L3L_3 = 0.5 H (shunt)kz(0.5)/ω0k_z(0.5)/\omega_0 = 3.865 mHGIC: C = 100 nF, R=L/CR = \sqrt{L/C} = 196.6 Ω
C4C_4 = 0.618 F (series)C=0.618/(kzω0)C = 0.618/(k_z\omega_0)100 nF
L5L_5 = 1.618 H (shunt)kz(1.618)/ω0k_z(1.618)/\omega_0 = 12.51 mHGIC: C = 100 nF, R=L/CR = \sqrt{L/C} = 353.7 Ω
Rs,RLR_s, R_L = 1 Ωkz×1k_z \times 1218.6 Ω

Each grounded inductor is replaced by a GIC with Z4=CZ_4 = C and R1=R2=R3=R5=RR_1 = R_2 = R_3 = R_5 = R, so L=CR2L = CR^2 and R=L/CR = \sqrt{L/C}.

Final circuit (each GIC realises the grounded inductor in that position):

Vi o--[Rs]-+----[C2]-+----[C4]-+-------+----o Vo
           |         |         |       |
        [GIC-L1]  [GIC-L3]  [GIC-L5]  [RL]
GND --------------------------------------------

Answer: Rs=RLR_s = R_L = 218.6 Ω; series capacitors 0.1 μF each; simulated inductors 12.51 mH, 3.865 mH, 12.51 mH, realised by GICs with C = 0.1 μF and R = 353.7 Ω, 196.6 Ω, 353.7 Ω; half-power frequency 4.5 kHz.

  • 2076 Chaitra · 1+3+4 marks

What is GIC? How can you simulate a grounded inductor? Design a fourth order Butterworth highpass filter having ωo = 16,000 rad/s and practically suitable elements using simulated inductors. [Table: doubly terminated Butterworth element values normalized to 1 rad/s, 1 Ω terminations: n = 2: 1.414, 1.414; n = 3: 1, 2, 1; n = 4: 0.7654, 1.848, 1.848, 0.7654; n = 5: 0.618, 1.618, 2, 1.618, 0.618]

Answer

A generalized impedance converter (GIC) is an active two-port, usually Antoniou's two-op-amp, five-impedance circuit. Its input impedance is the load impedance Z5Z_5 multiplied by a conversion function k(s)=Z1Z3/(Z2Z4)k(s) = Z_1Z_3/(Z_2Z_4). By choosing which positions hold capacitors, a GIC can make an inductor or an FDNR from only R, C and op-amps.

Simulating a grounded inductor with a GIC

   Iin  node 1
 o-->-----+--------------(+) A1
 Vin      |
         [Z1]
          +--- node 2 ---- output of A1
         [Z2]
          +--- node 3 ---- (-) A1 and (-) A2
         [Z3]
          +--- node 4 ---- output of A2
         [Z4]
          +--- node 5 ---- (+) A2
         [Z5]
          |
         GND

With ideal op-amps the two input terminals of each op-amp are at the same voltage and draw no current, so V1=V3=V5=VinV_1 = V_3 = V_5 = V_{in}.

Node 5: V4−VinZ4=VinZ5  ⇒  V4=Vin(1+Z4Z5)Node 3: V2−VinZ2=Vin−V4Z3=−Z4Z3Z5Vin  ⇒  V2=Vin(1−Z2Z4Z3Z5)Node 1: Iin=Vin−V2Z1=Z2Z4Z1Z3Z5Vin\begin{aligned} \text{Node 5: } & \frac{V_4 - V_{in}}{Z_4} = \frac{V_{in}}{Z_5} \;\Rightarrow\; V_4 = V_{in}\left(1 + \frac{Z_4}{Z_5}\right) \\ \text{Node 3: } & \frac{V_2 - V_{in}}{Z_2} = \frac{V_{in} - V_4}{Z_3} = -\frac{Z_4}{Z_3 Z_5}V_{in} \;\Rightarrow\; V_2 = V_{in}\left(1 - \frac{Z_2 Z_4}{Z_3 Z_5}\right) \\ \text{Node 1: } & I_{in} = \frac{V_{in} - V_2}{Z_1} = \frac{Z_2 Z_4}{Z_1 Z_3 Z_5}V_{in} \end{aligned} Zin=VinIin=Z1Z3Z5Z2Z4Z_{in} = \frac{V_{in}}{I_{in}} = \frac{Z_1 Z_3 Z_5}{Z_2 Z_4}

Put a capacitor in position 4 (Z4=1/sCZ_4 = 1/sC) and resistors elsewhere:

Zin=R1R3R5R2 sC=sL,L=CR1R3R5R2Z_{in} = \frac{R_1 R_3 R_5}{R_2}\, sC = sL, \qquad L = \frac{C R_1 R_3 R_5}{R_2}

With all resistors equal to RR: L=CR2L = CR^2. One end of this inductor is ground, so it replaces a grounded (shunt) inductor.

Design

Step 1: Lowpass prototype (table, n = 4): series L1=0.7654L_1 = 0.7654 H, shunt C2=1.848C_2 = 1.848 F, series L3=1.848L_3 = 1.848 H, shunt C4=0.7654C_4 = 0.7654 F, 1 Ω terminations.

Step 2: LP to HP transformation (s→1/ss \to 1/s): each series L becomes a series capacitor 1/L1/L, each shunt C becomes a shunt (grounded) inductor 1/C1/C:

LowpassHighpass element
L1=0.7654L_1 = 0.7654 HC1=1/0.7654=1.3065C_1 = 1/0.7654 = 1.3065 F (series)
C2=1.848C_2 = 1.848 FL2=1/1.848=0.5411L_2 = 1/1.848 = 0.5411 H (shunt)
L3=1.848L_3 = 1.848 HC3=0.5411C_3 = 0.5411 F (series)
C4=0.7654C_4 = 0.7654 FL4=1.3065L_4 = 1.3065 H (shunt)

Step 3: Scaling (both inductors are grounded, so they are simulated by GICs).

Frequency scale to ω0=16,000 rad/s\omega_0 = 16{,}000\ \text{rad/s} and impedance scale by kzk_z (chosen as 10410^4):

Rnew=kzR,Lnew=kzLω0,Cnew=Ckz ω0R_{new} = k_z R, \qquad L_{new} = \frac{k_z L}{\omega_0}, \qquad C_{new} = \frac{C}{k_z\,\omega_0}

kz=10000 Ωk_z = 10000\ \Omega, so the source and load resistors become Rs=RLR_s = R_L = 10 kΩ.

PrototypeScaledFinal
C1C_1 = 1.3065 F (series)C=1.3065/(kzω0)C = 1.3065/(k_z\omega_0)8.166 nF
L2L_2 = 0.5411 H (shunt)kz(0.5411)/ω0k_z(0.5411)/\omega_0 = 338.2 mHGIC: C = 10 nF, R=L/CR = \sqrt{L/C} = 5.815 kΩ
C3C_3 = 0.5411 F (series)C=0.5411/(kzω0)C = 0.5411/(k_z\omega_0)3.382 nF
L4L_4 = 1.3065 H (shunt)kz(1.3065)/ω0k_z(1.3065)/\omega_0 = 816.6 mHGIC: C = 10 nF, R=L/CR = \sqrt{L/C} = 9.036 kΩ
Rs,RLR_s, R_L = 1 Ωkz×1k_z \times 110 kΩ

Each grounded inductor is replaced by a GIC with Z4=CZ_4 = C and R1=R2=R3=R5=RR_1 = R_2 = R_3 = R_5 = R, so L=CR2L = CR^2 and R=L/CR = \sqrt{L/C}.

Final circuit:

Vi o--[Rs]--[C1]-+----[C3]-+-------+----o Vo
                 |         |       |
              [GIC-L2]  [GIC-L4]  [RL]
GND ----------------------------------------

Answer: Rs=RLR_s = R_L = 10 kΩ; series capacitors 8.166 nF and 3.382 nF; L2L_2 = 338.2 mH (GIC: C = 10 nF, R = 5.816 kΩ); L4L_4 = 816.6 mH (GIC: C = 10 nF, R = 9.036 kΩ); ω0\omega_0 = 16,000 rad/s.

  • 2082 Baisakh · 1+4 marks

What is ideal gyrator? How can you simulate floating inductor using gyrator? Explain with necessary figure and derivations.

Answer

Ideal gyrator

An ideal gyrator is a lossless two-port that inverts impedance: a load ZLZ_L on port 2 appears at port 1 as Zin=r2/ZLZ_{in} = r^2/Z_L, where rr is the gyration resistance. Its defining equations are

V1=−rI2,V2=rI1⟹Zin=V1I1=r2ZLV_1 = -r I_2, \qquad V_2 = r I_1 \quad\Longrightarrow\quad Z_{in} = \frac{V_1}{I_1} = \frac{r^2}{Z_L}

So a capacitor CC on port 2 gives Zin=r2sC=sLZ_{in} = r^2 sC = sL with L=r2CL = r^2C. A gyrator is built with op-amps (e.g. two voltage-controlled current sources of opposite sign, I1=gV2I_1 = gV_2, I2=−gV1I_2 = -gV_1, r=1/gr = 1/g).

Floating inductor using gyrators

One gyrator with a grounded capacitor gives only a grounded inductor. A floating inductor is made by cascading two identical gyrators with a grounded capacitor between them:

  a o---+---------+----+----+---------+---o b
        | Gyr. 1  |    |    | Gyr. 2  |
        |   (r)   |   === C |   (r)   |
        |         |    |    |         |
 GND ---+---------+----+----+---------+---- GND

The transmission (ABCD) matrix of an ideal gyrator (V1=rI2V_1 = rI_2, I1=V2/rI_1 = V_2/r, with I2I_2 leaving port 2) and of a shunt capacitor are

G=[0r1/r0],YC=[10sC1]G = \begin{bmatrix} 0 & r \\ 1/r & 0 \end{bmatrix}, \qquad Y_C = \begin{bmatrix} 1 & 0 \\ sC & 1 \end{bmatrix}

For the cascade:

G YC G=[0r1/r0][10sC1][0r1/r0]=[rsCr1/r0][0r1/r0]=[1sCr201]\begin{aligned} G\,Y_C\,G &= \begin{bmatrix} 0 & r \\ 1/r & 0 \end{bmatrix}\begin{bmatrix} 1 & 0 \\ sC & 1 \end{bmatrix}\begin{bmatrix} 0 & r \\ 1/r & 0 \end{bmatrix} \\ &= \begin{bmatrix} rsC & r \\ 1/r & 0 \end{bmatrix}\begin{bmatrix} 0 & r \\ 1/r & 0 \end{bmatrix} = \begin{bmatrix} 1 & sCr^2 \\ 0 & 1 \end{bmatrix} \end{aligned}

The matrix [1Z01]\begin{bmatrix} 1 & Z \\ 0 & 1 \end{bmatrix} is that of a single series impedance ZZ between the two terminals. Here Z=s r2CZ = s\,r^2C, so the circuit between a and b is a floating inductor

L=r2CL = r^2 C

Example: rr = 10 kΩ and CC = 10 nF give L=108×10−8L = 10^8 \times 10^{-8} = 1 H. The two gyrators must have equal gyration resistance; any mismatch leaves an unwanted grounded element. The capacitor is grounded, which suits IC realisation.

  • 2079 Bhadra · 3+5 marks

What is Bruton transformation? Design the 4th order Butterworth low pass filter with half power frequency 20,000 rad/s and practically realizable elements using FDNR. [Table: doubly terminated Butterworth element values normalized to 1 rad/s, 1 Ω terminations: n = 2: 1.414, 1.414; n = 3: 1, 2, 1; n = 4: 0.7654, 1.848, 1.848, 0.7654; n = 5: 0.618, 1.618, 2, 1.618, 0.618]

Answer

Bruton transformation

Bruton transformation: divide every impedance of the RLC ladder by ss (more generally by ksks). A voltage transfer function is a ratio of impedances, so it does not change:

ElementImpedanceAfter dividing by ssNew element
Resistor RRRRR/sR/scapacitor C=1/RC = 1/R
Inductor LLsLsLLLresistor R=LR = L
Capacitor CC1/sC1/sC1/(s2C)1/(s^2C)FDNR D=CD = C

In a lowpass ladder the inductors are in the series arms (often floating) and the capacitors are shunt (grounded). After the transformation the inductors become plain resistors and the grounded capacitors become grounded FDNRs, which one GIC can realise. So no inductor is left.

Design: 4th-order Butterworth LPF, ω0\omega_0 = 20,000 rad/s

From the table (n = 4), the doubly terminated Butterworth prototype (1 Ω terminations, ω0=1\omega_0 = 1 rad/s) is Rs=1R_s = 1, L1=0.7654L_1 = 0.7654 H, C2=1.848C_2 = 1.848 F, L3=1.848L_3 = 1.848 H, C4=0.7654C_4 = 0.7654 F, RL=1 ΩR_L = 1\ \Omega (T-ladder: series L first, shunt C to ground).

Step 1: Bruton transformation (divide all impedances by ss): source and load resistors (1 Ω) become 1 F capacitors, each inductor LL becomes a resistor of LL Ω, each shunt capacitor CC becomes a grounded FDNR with D=CD = C.

Step 2: Scaling. Frequency scaling factor ω0=20,000 rad/s\omega_0 = 20{,}000\ \text{rad/s}. Choose the terminating capacitors as CtC_t = 10 nF, so

kz=1ω0Ct=120000×1×10−8=5000 Ωk_z = \frac{1}{\omega_0 C_t} = \frac{1}{20000 \times 1 \times 10^{-8}} = 5000\ \Omega Rnew=kzR,Cnew=Ckz ω0,Dnew=Dkz ω02\begin{aligned} R_{new} &= k_z R, \qquad C_{new} = \frac{C}{k_z\,\omega_0}, \qquad D_{new} = \frac{D}{k_z\,\omega_0^2} \end{aligned}

Step 3: Element values. Each FDNR is a GIC with C1=C5=CC_1 = C_5 = C and R2=R3=R4=RDR_2 = R_3 = R_4 = R_D, so D=C2RDD = C^2R_D. Taking C=CtC = C_t gives RD=Dnew/Ct2=kzDR_D = D_{new}/C_t^2 = k_z D.

PrototypeAfter BrutonScaledFinal element
Rs=1 ΩR_s = 1\ \OmegaCs=1C_s = 1 F1/(kzω0)1/(k_z\omega_0)10 nF
L1=0.7654L_1 = 0.7654 H (series)R1=0.7654 ΩR_1 = 0.7654\ \OmegaR1=0.7654×5000R_1 = 0.7654 \times 50003.827 kΩ
C2=1.848C_2 = 1.848 F (shunt)FDNR D2=1.848D_2 = 1.848D2=1.848/(kzω02)=9.24×10−13D_2 = 1.848/(k_z\omega_0^2) = 9.24 \times 10^{-13}GIC: C = 10 nF, RDR_D = 9.24 kΩ
L3=1.848L_3 = 1.848 H (series)R3=1.848 ΩR_3 = 1.848\ \OmegaR3=1.848×5000R_3 = 1.848 \times 50009.24 kΩ
C4=0.7654C_4 = 0.7654 F (shunt)FDNR D4=0.7654D_4 = 0.7654D4=0.7654/(kzω02)=3.827×10−13D_4 = 0.7654/(k_z\omega_0^2) = 3.827 \times 10^{-13}GIC: C = 10 nF, RDR_D = 3.827 kΩ
RL=1 ΩR_L = 1\ \OmegaCL=1C_L = 1 F1/(kzω0)1/(k_z\omega_0)10 nF

Step 4: DC path. The source and load capacitors block DC, and the op-amps in the GICs need a DC bias path. Put RaR_a across the source capacitor and RbR_b across the load capacitor. They must be much larger than kzk_z (the capacitor impedance at ω0\omega_0). At DC the FDNRs are open, so the DC gain is Rb/(Ra+ΣR+Rb)R_b/(R_a + \Sigma R + R_b). To keep the prototype DC gain of 1/2, set Rb=Ra+ΣRR_b = R_a + \Sigma R:

RaR_a = 100 kΩ, ΣR\Sigma R = 13.07 kΩ, so RbR_b = 113.1 kΩ.

Final circuit (Cs∥RaC_s \| R_a in series from the source, CL∥RbC_L \| R_b as the load, each DkD_k a GIC FDNR to ground):

Vi o--[Cs||Ra]--[R1]-+----[R3]-+-------+----o Vo
                     |         |       |
                    [D2]      [D4]  [CL||Rb]
GND --------------------------------------------

Put capacitors in positions 1 and 5 (Z1=1/sC1Z_1 = 1/sC_1, Z5=1/sC5Z_5 = 1/sC_5) and resistors in 2, 3, 4:

Zin=R3s2C1C5R2R4=1s2D,D=C1C5R2R4R3Z_{in} = \frac{R_3}{s^2 C_1 C_5 R_2 R_4} = \frac{1}{s^2 D}, \qquad D = \frac{C_1 C_5 R_2 R_4}{R_3}

With C1=C5=CC_1 = C_5 = C and R2=R3=R4=RR_2 = R_3 = R_4 = R: D=C2RD = C^2 R. At s=jωs = j\omega, Zin=−1/(ω2D)Z_{in} = -1/(\omega^2 D), a negative resistance whose value depends on frequency: an FDNR (also called a D-element or supercapacitor).

Answer: source and load capacitors 10 nF; series resistors 3.827 kΩ and 9.24 kΩ; FDNRs D2D_2 (GIC with C = 10 nF, RDR_D = 9.24 kΩ) and D4D_4 (C = 10 nF, RDR_D = 3.827 kΩ); RaR_a = 100 kΩ, RbR_b = 113.1 kΩ.

  • 2082 Bhadra · 2+4 marks

What is Bruton Transformation? How can you simulate grounded inductor using FDNR? Explain with necessary figures and derivations.

Answer

Bruton transformation

Bruton transformation: divide every impedance of the RLC ladder by ss (more generally by ksks). A voltage transfer function is a ratio of impedances, so it does not change:

ElementImpedanceAfter dividing by ssNew element
Resistor RRRRR/sR/scapacitor C=1/RC = 1/R
Inductor LLsLsLLLresistor R=LR = L
Capacitor CC1/sC1/sC1/(s2C)1/(s^2C)FDNR D=CD = C

In a lowpass ladder the inductors are in the series arms (often floating) and the capacitors are shunt (grounded). After the transformation the inductors become plain resistors and the grounded capacitors become grounded FDNRs, which one GIC can realise. So no inductor is left.

Simulating the inductor problem with FDNR

A lowpass LC ladder has inductors in the series arms. Inductors are bulky, lossy and cannot be integrated. After the Bruton transformation (all impedances divided by ss):

  • each inductor LL becomes a resistor R=LR = L (no inductor left),
  • each grounded capacitor CC becomes a grounded FDNR D=CD = C,
  • each terminating resistor RR becomes a capacitor 1/R1/R.

So the inductor is "simulated" indirectly: the whole network is converted to R, C and grounded FDNRs with the same voltage transfer function.

 Original:  o-[Rs]-(L1)-+-(L3)-+---o
                        C2     C4 [RL]
 Bruton:    o-[Cs]-[R1]-+-[R3]-+---o
                        D2     D4 [CL]

Realising the FDNR with a GIC

   Iin  node 1
 o-->-----+--------------(+) A1
 Vin      |
         [Z1]
          +--- node 2 ---- output of A1
         [Z2]
          +--- node 3 ---- (-) A1 and (-) A2
         [Z3]
          +--- node 4 ---- output of A2
         [Z4]
          +--- node 5 ---- (+) A2
         [Z5]
          |
         GND

With ideal op-amps the two input terminals of each op-amp are at the same voltage and draw no current, so V1=V3=V5=VinV_1 = V_3 = V_5 = V_{in}.

Node 5: V4−VinZ4=VinZ5  ⇒  V4=Vin(1+Z4Z5)Node 3: V2−VinZ2=Vin−V4Z3=−Z4Z3Z5Vin  ⇒  V2=Vin(1−Z2Z4Z3Z5)Node 1: Iin=Vin−V2Z1=Z2Z4Z1Z3Z5Vin\begin{aligned} \text{Node 5: } & \frac{V_4 - V_{in}}{Z_4} = \frac{V_{in}}{Z_5} \;\Rightarrow\; V_4 = V_{in}\left(1 + \frac{Z_4}{Z_5}\right) \\ \text{Node 3: } & \frac{V_2 - V_{in}}{Z_2} = \frac{V_{in} - V_4}{Z_3} = -\frac{Z_4}{Z_3 Z_5}V_{in} \;\Rightarrow\; V_2 = V_{in}\left(1 - \frac{Z_2 Z_4}{Z_3 Z_5}\right) \\ \text{Node 1: } & I_{in} = \frac{V_{in} - V_2}{Z_1} = \frac{Z_2 Z_4}{Z_1 Z_3 Z_5}V_{in} \end{aligned} Zin=VinIin=Z1Z3Z5Z2Z4Z_{in} = \frac{V_{in}}{I_{in}} = \frac{Z_1 Z_3 Z_5}{Z_2 Z_4}

Put capacitors in positions 1 and 5 (Z1=1/sC1Z_1 = 1/sC_1, Z5=1/sC5Z_5 = 1/sC_5) and resistors in 2, 3, 4:

Zin=R3s2C1C5R2R4=1s2D,D=C1C5R2R4R3Z_{in} = \frac{R_3}{s^2 C_1 C_5 R_2 R_4} = \frac{1}{s^2 D}, \qquad D = \frac{C_1 C_5 R_2 R_4}{R_3}

With C1=C5=CC_1 = C_5 = C and R2=R3=R4=RR_2 = R_3 = R_4 = R: D=C2RD = C^2 R. At s=jωs = j\omega, Zin=−1/(ω2D)Z_{in} = -1/(\omega^2 D), a negative resistance whose value depends on frequency: an FDNR (also called a D-element or supercapacitor).

Practical points

  • Because the terminations become capacitors, there is no DC path; resistors RaR_a (across the source capacitor) and RbR_b (across the load capacitor) are added, much larger than the capacitor impedances in the passband, with Rb=Ra+ΣRR_b = R_a + \Sigma R to keep the DC gain at 1/2.
  • Only grounded FDNRs are needed, so one GIC per shunt capacitor is enough, and the low sensitivity of the ladder is kept.
  • 2072 Kartik · 6 marks

What is the importance of Bruton transformation in filter design? How can you simulate FDNR using generalized impedance converter (GIC)? Explain with example.

Answer

Importance of the Bruton transformation

The Bruton transformation divides every impedance of an RLC ladder by ss: R→R \to capacitor 1/R1/R, L→L \to resistor LL, C→C \to FDNR D=CD = C. Because a voltage transfer function is a ratio of impedances, it is unchanged. Its importance:

  1. Removes all inductors, including floating (series) ones in lowpass ladders, which are hard to simulate directly.
  2. The only active elements needed are grounded FDNRs, each made with one GIC; floating elements are avoided.
  3. The design starts from a doubly terminated LC ladder, so the very low sensitivity of the ladder to element changes is kept.
  4. The final circuit has only resistors, capacitors and op-amps, suitable for integration and for low frequencies where inductors would be huge.
  5. Extra care: a DC path must be added (large resistors across the terminating capacitors).

FDNR using a GIC

   Iin  node 1
 o-->-----+--------------(+) A1
 Vin      |
         [Z1]
          +--- node 2 ---- output of A1
         [Z2]
          +--- node 3 ---- (-) A1 and (-) A2
         [Z3]
          +--- node 4 ---- output of A2
         [Z4]
          +--- node 5 ---- (+) A2
         [Z5]
          |
         GND

With ideal op-amps the two input terminals of each op-amp are at the same voltage and draw no current, so V1=V3=V5=VinV_1 = V_3 = V_5 = V_{in}.

Node 5: V4−VinZ4=VinZ5  ⇒  V4=Vin(1+Z4Z5)Node 3: V2−VinZ2=Vin−V4Z3=−Z4Z3Z5Vin  ⇒  V2=Vin(1−Z2Z4Z3Z5)Node 1: Iin=Vin−V2Z1=Z2Z4Z1Z3Z5Vin\begin{aligned} \text{Node 5: } & \frac{V_4 - V_{in}}{Z_4} = \frac{V_{in}}{Z_5} \;\Rightarrow\; V_4 = V_{in}\left(1 + \frac{Z_4}{Z_5}\right) \\ \text{Node 3: } & \frac{V_2 - V_{in}}{Z_2} = \frac{V_{in} - V_4}{Z_3} = -\frac{Z_4}{Z_3 Z_5}V_{in} \;\Rightarrow\; V_2 = V_{in}\left(1 - \frac{Z_2 Z_4}{Z_3 Z_5}\right) \\ \text{Node 1: } & I_{in} = \frac{V_{in} - V_2}{Z_1} = \frac{Z_2 Z_4}{Z_1 Z_3 Z_5}V_{in} \end{aligned} Zin=VinIin=Z1Z3Z5Z2Z4Z_{in} = \frac{V_{in}}{I_{in}} = \frac{Z_1 Z_3 Z_5}{Z_2 Z_4}

Put capacitors in positions 1 and 5 (Z1=1/sC1Z_1 = 1/sC_1, Z5=1/sC5Z_5 = 1/sC_5) and resistors in 2, 3, 4:

Zin=R3s2C1C5R2R4=1s2D,D=C1C5R2R4R3Z_{in} = \frac{R_3}{s^2 C_1 C_5 R_2 R_4} = \frac{1}{s^2 D}, \qquad D = \frac{C_1 C_5 R_2 R_4}{R_3}

With C1=C5=CC_1 = C_5 = C and R2=R3=R4=RR_2 = R_3 = R_4 = R: D=C2RD = C^2 R. At s=jωs = j\omega, Zin=−1/(ω2D)Z_{in} = -1/(\omega^2 D), a negative resistance whose value depends on frequency: an FDNR (also called a D-element or supercapacitor).

Example

Third-order Butterworth LPF (prototype Rs=1R_s = 1, L1=1L_1 = 1 H, C2=2C_2 = 2 F, L3=1L_3 = 1 H, RL=1R_L = 1) at ω0\omega_0 = 10,000 rad/s.

  1. Bruton: Cs=1C_s = 1 F, R1=1 ΩR_1 = 1\ \Omega, D2=2D_2 = 2, R3=1 ΩR_3 = 1\ \Omega, CL=1C_L = 1 F.
  2. Choose CtC_t = 10 nF: kz=1/(104×10−8)=104 Ωk_z = 1/(10^4 \times 10^{-8}) = 10^4\ \Omega.
  3. Resistors: R1=R3R_1 = R_3 = 10 kΩ. Terminating capacitors 10 nF.
  4. D2=2/(104×108)=2×10−12D_2 = 2/(10^4 \times 10^8) = 2 \times 10^{-12}; with GIC capacitors 10 nF, RD=D/C2=2×10−12/10−16R_D = D/C^2 = 2 \times 10^{-12}/10^{-16} = 20 kΩ.
  5. DC path: RaR_a = 1 MΩ, Rb=Ra+20R_b = R_a + 20 kΩ = 1.02 MΩ.
 Vi o--[10nF||Ra]--[10k]--+--[10k]--+---o Vo
                          |         |
                       [GIC D2] [10nF||Rb]
 GND ---------------------+---------+
  • 2080 Bhadra · 4+6 marks

What is FDNR? How can you use FDNR to avoid the inductor in filter design? Explain. Design third order Butterworth low pass filter having half power frequency 4000 rad/s using FDNR. [Table: doubly terminated Butterworth element values normalized to 1 rad/s, 1 Ω terminations: n = 2: 1.414, 1.414; n = 3: 1, 2, 1; n = 4: 0.7654, 1.848, 1.848, 0.7654; n = 5: 0.618, 1.618, 2, 1.618, 0.618]

Answer

FDNR

A frequency dependent negative resistor (FDNR) is a one-port with impedance Z(s)=1s2DZ(s) = \dfrac{1}{s^2D}. At s=jωs = j\omega, Z=−1ω2DZ = -\dfrac{1}{\omega^2 D}: a real, negative resistance whose size falls as 1/ω21/\omega^2. DD (unit F²Ω) is its constant. It cannot be made from passive elements; a GIC with two capacitors realises it.

Using FDNR to avoid inductors

Bruton transformation: divide every impedance of the RLC ladder by ss (more generally by ksks). A voltage transfer function is a ratio of impedances, so it does not change:

ElementImpedanceAfter dividing by ssNew element
Resistor RRRRR/sR/scapacitor C=1/RC = 1/R
Inductor LLsLsLLLresistor R=LR = L
Capacitor CC1/sC1/sC1/(s2C)1/(s^2C)FDNR D=CD = C

In a lowpass ladder the inductors are in the series arms (often floating) and the capacitors are shunt (grounded). After the transformation the inductors become plain resistors and the grounded capacitors become grounded FDNRs, which one GIC can realise. So no inductor is left.

The FDNR itself is a GIC: Put capacitors in positions 1 and 5 (Z1=1/sC1Z_1 = 1/sC_1, Z5=1/sC5Z_5 = 1/sC_5) and resistors in 2, 3, 4:

Zin=R3s2C1C5R2R4=1s2D,D=C1C5R2R4R3Z_{in} = \frac{R_3}{s^2 C_1 C_5 R_2 R_4} = \frac{1}{s^2 D}, \qquad D = \frac{C_1 C_5 R_2 R_4}{R_3}

With C1=C5=CC_1 = C_5 = C and R2=R3=R4=RR_2 = R_3 = R_4 = R: D=C2RD = C^2 R. At s=jωs = j\omega, Zin=−1/(ω2D)Z_{in} = -1/(\omega^2 D), a negative resistance whose value depends on frequency: an FDNR (also called a D-element or supercapacitor).

Design: 3rd-order Butterworth LPF, ω0\omega_0 = 4000 rad/s

From the table (n = 3): Rs=1 ΩR_s = 1\ \Omega, L1=1L_1 = 1 H (series), C2=2C_2 = 2 F (shunt), L3=1L_3 = 1 H (series), RL=1 ΩR_L = 1\ \Omega.

Step 1: Bruton transformation (divide all impedances by ss): source and load resistors (1 Ω) become 1 F capacitors, each inductor LL becomes a resistor of LL Ω, each shunt capacitor CC becomes a grounded FDNR with D=CD = C.

Step 2: Scaling. Frequency scaling factor ω0=4000 rad/s\omega_0 = 4000\ \text{rad/s}. Choose the terminating capacitors as CtC_t = 100 nF, so

kz=1ω0Ct=14000×1×10−7=2500 Ωk_z = \frac{1}{\omega_0 C_t} = \frac{1}{4000 \times 1 \times 10^{-7}} = 2500\ \Omega Rnew=kzR,Cnew=Ckz ω0,Dnew=Dkz ω02\begin{aligned} R_{new} &= k_z R, \qquad C_{new} = \frac{C}{k_z\,\omega_0}, \qquad D_{new} = \frac{D}{k_z\,\omega_0^2} \end{aligned}

Step 3: Element values. Each FDNR is a GIC with C1=C5=CC_1 = C_5 = C and R2=R3=R4=RDR_2 = R_3 = R_4 = R_D, so D=C2RDD = C^2R_D. Taking C=CtC = C_t gives RD=Dnew/Ct2=kzDR_D = D_{new}/C_t^2 = k_z D.

PrototypeAfter BrutonScaledFinal element
Rs=1 ΩR_s = 1\ \OmegaCs=1C_s = 1 F1/(kzω0)1/(k_z\omega_0)100 nF
L1=1L_1 = 1 H (series)R1=1 ΩR_1 = 1\ \OmegaR1=1×2500R_1 = 1 \times 25002.5 kΩ
C2=2C_2 = 2 F (shunt)FDNR D2=2D_2 = 2D2=2/(kzω02)=5×10−11D_2 = 2/(k_z\omega_0^2) = 5 \times 10^{-11}GIC: C = 100 nF, RDR_D = 5 kΩ
L3=1L_3 = 1 H (series)R3=1 ΩR_3 = 1\ \OmegaR3=1×2500R_3 = 1 \times 25002.5 kΩ
RL=1 ΩR_L = 1\ \OmegaCL=1C_L = 1 F1/(kzω0)1/(k_z\omega_0)100 nF

Step 4: DC path. The source and load capacitors block DC, and the op-amps in the GICs need a DC bias path. Put RaR_a across the source capacitor and RbR_b across the load capacitor. They must be much larger than kzk_z (the capacitor impedance at ω0\omega_0). At DC the FDNRs are open, so the DC gain is Rb/(Ra+ΣR+Rb)R_b/(R_a + \Sigma R + R_b). To keep the prototype DC gain of 1/2, set Rb=Ra+ΣRR_b = R_a + \Sigma R:

RaR_a = 100 kΩ, ΣR\Sigma R = 5 kΩ, so RbR_b = 105 kΩ.

Final circuit (Cs∥RaC_s \| R_a in series from the source, CL∥RbC_L \| R_b as the load, each DkD_k a GIC FDNR to ground):

Vi o--[Cs||Ra]--[R1]-+----[R3]-+----o Vo
                     |         |
                    [D2]    [CL||Rb]
GND ------------------------------------

Answer: terminating capacitors 0.1 μF; R1=R3R_1 = R_3 = 2.5 kΩ; FDNR D2D_2 = GIC with C = 0.1 μF and RDR_D = 5 kΩ; RaR_a = 100 kΩ, RbR_b = 105 kΩ.

  • 2080 Baisakh · 1+3+5 marks

What is frequency dependent negative resistor (FDNR)? How can it be realized? Realize the following passive filter using FDNR, having ωo = 25000 rad/s and practical element values in your final circuit. [Figure: Butterworth filter at normalized frequency - source V with 1 Ω series resistor, series 0.618 H, shunt 1.618 F, series 2 H, shunt 1.618 F, series 0.618 H, 1 Ω load]

Answer

FDNR

A frequency dependent negative resistor (FDNR) is a one-port with impedance Z(s)=1s2DZ(s) = \dfrac{1}{s^2D}. At s=jωs = j\omega, Z=−1ω2DZ = -\dfrac{1}{\omega^2 D}: a real, negative resistance whose size falls as 1/ω21/\omega^2. DD (unit F²Ω) is its constant. It cannot be made from passive elements; a GIC with two capacitors realises it.

Realisation

The FDNR is realised with a GIC (Antoniou circuit):

   Iin  node 1
 o-->-----+--------------(+) A1
 Vin      |
         [Z1]
          +--- node 2 ---- output of A1
         [Z2]
          +--- node 3 ---- (-) A1 and (-) A2
         [Z3]
          +--- node 4 ---- output of A2
         [Z4]
          +--- node 5 ---- (+) A2
         [Z5]
          |
         GND

Put capacitors in positions 1 and 5 (Z1=1/sC1Z_1 = 1/sC_1, Z5=1/sC5Z_5 = 1/sC_5) and resistors in 2, 3, 4:

Zin=R3s2C1C5R2R4=1s2D,D=C1C5R2R4R3Z_{in} = \frac{R_3}{s^2 C_1 C_5 R_2 R_4} = \frac{1}{s^2 D}, \qquad D = \frac{C_1 C_5 R_2 R_4}{R_3}

With C1=C5=CC_1 = C_5 = C and R2=R3=R4=RR_2 = R_3 = R_4 = R: D=C2RD = C^2 R. At s=jωs = j\omega, Zin=−1/(ω2D)Z_{in} = -1/(\omega^2 D), a negative resistance whose value depends on frequency: an FDNR (also called a D-element or supercapacitor).

Design: 5th-order Butterworth LPF, ω0\omega_0 = 25,000 rad/s

Given prototype: Rs=1 ΩR_s = 1\ \Omega, L1=0.618L_1 = 0.618 H, C2=1.618C_2 = 1.618 F, L3=2L_3 = 2 H, C4=1.618C_4 = 1.618 F, L5=0.618L_5 = 0.618 H, RL=1 ΩR_L = 1\ \Omega.

Step 1: Bruton transformation (divide all impedances by ss): source and load resistors (1 Ω) become 1 F capacitors, each inductor LL becomes a resistor of LL Ω, each shunt capacitor CC becomes a grounded FDNR with D=CD = C.

Step 2: Scaling. Frequency scaling factor ω0=25,000 rad/s\omega_0 = 25{,}000\ \text{rad/s}. Choose the terminating capacitors as CtC_t = 10 nF, so

kz=1ω0Ct=125000×1×10−8=4000 Ωk_z = \frac{1}{\omega_0 C_t} = \frac{1}{25000 \times 1 \times 10^{-8}} = 4000\ \Omega Rnew=kzR,Cnew=Ckz ω0,Dnew=Dkz ω02\begin{aligned} R_{new} &= k_z R, \qquad C_{new} = \frac{C}{k_z\,\omega_0}, \qquad D_{new} = \frac{D}{k_z\,\omega_0^2} \end{aligned}

Step 3: Element values. Each FDNR is a GIC with C1=C5=CC_1 = C_5 = C and R2=R3=R4=RDR_2 = R_3 = R_4 = R_D, so D=C2RDD = C^2R_D. Taking C=CtC = C_t gives RD=Dnew/Ct2=kzDR_D = D_{new}/C_t^2 = k_z D.

PrototypeAfter BrutonScaledFinal element
Rs=1 ΩR_s = 1\ \OmegaCs=1C_s = 1 F1/(kzω0)1/(k_z\omega_0)10 nF
L1=0.618L_1 = 0.618 H (series)R1=0.618 ΩR_1 = 0.618\ \OmegaR1=0.618×4000R_1 = 0.618 \times 40002.472 kΩ
C2=1.618C_2 = 1.618 F (shunt)FDNR D2=1.618D_2 = 1.618D2=1.618/(kzω02)=6.472×10−13D_2 = 1.618/(k_z\omega_0^2) = 6.472 \times 10^{-13}GIC: C = 10 nF, RDR_D = 6.472 kΩ
L3=2L_3 = 2 H (series)R3=2 ΩR_3 = 2\ \OmegaR3=2×4000R_3 = 2 \times 40008 kΩ
C4=1.618C_4 = 1.618 F (shunt)FDNR D4=1.618D_4 = 1.618D4=1.618/(kzω02)=6.472×10−13D_4 = 1.618/(k_z\omega_0^2) = 6.472 \times 10^{-13}GIC: C = 10 nF, RDR_D = 6.472 kΩ
L5=0.618L_5 = 0.618 H (series)R5=0.618 ΩR_5 = 0.618\ \OmegaR5=0.618×4000R_5 = 0.618 \times 40002.472 kΩ
RL=1 ΩR_L = 1\ \OmegaCL=1C_L = 1 F1/(kzω0)1/(k_z\omega_0)10 nF

Step 4: DC path. The source and load capacitors block DC, and the op-amps in the GICs need a DC bias path. Put RaR_a across the source capacitor and RbR_b across the load capacitor. They must be much larger than kzk_z (the capacitor impedance at ω0\omega_0). At DC the FDNRs are open, so the DC gain is Rb/(Ra+ΣR+Rb)R_b/(R_a + \Sigma R + R_b). To keep the prototype DC gain of 1/2, set Rb=Ra+ΣRR_b = R_a + \Sigma R:

RaR_a = 100 kΩ, ΣR\Sigma R = 12.94 kΩ, so RbR_b = 112.9 kΩ.

Final circuit (Cs∥RaC_s \| R_a in series from the source, CL∥RbC_L \| R_b as the load, each DkD_k a GIC FDNR to ground):

Vi o--[Cs||Ra]--[R1]-+----[R3]-+----[R5]-+----o Vo
                     |         |         |
                    [D2]      [D4]    [CL||Rb]
GND ----------------------------------------------

Answer: terminating capacitors 10 nF; series resistors 2.472 kΩ, 8 kΩ, 2.472 kΩ; two FDNRs, each a GIC with C = 10 nF and RDR_D = 6.472 kΩ; RaR_a = 100 kΩ, RbR_b = 112.9 kΩ.

  • 2079 Baisakh · 5 marks

Design a fourth order Butterworth low pass filter having half power frequency of 16000 rad/s using FDNR. [Table: doubly terminated Butterworth element values normalized to 1 rad/s, 1 Ω terminations: n = 2: 1.414, 1.414; n = 3: 1, 2, 1; n = 4: 0.7654, 1.848, 1.848, 0.7654; n = 5: 0.618, 1.618, 2, 1.618, 0.618]

Answer

The inductors of the doubly terminated ladder are removed with the Bruton transformation, and the shunt capacitors become grounded FDNRs (Z=1/s2DZ = 1/s^2D) made with GICs.

From the table (n = 4), the doubly terminated Butterworth prototype (1 Ω terminations, ω0=1\omega_0 = 1 rad/s) is Rs=1R_s = 1, L1=0.7654L_1 = 0.7654 H, C2=1.848C_2 = 1.848 F, L3=1.848L_3 = 1.848 H, C4=0.7654C_4 = 0.7654 F, RL=1 ΩR_L = 1\ \Omega (T-ladder: series L first, shunt C to ground).

Step 1: Bruton transformation (divide all impedances by ss): source and load resistors (1 Ω) become 1 F capacitors, each inductor LL becomes a resistor of LL Ω, each shunt capacitor CC becomes a grounded FDNR with D=CD = C.

Step 2: Scaling. Frequency scaling factor ω0=16,000 rad/s\omega_0 = 16{,}000\ \text{rad/s}. Choose the terminating capacitors as CtC_t = 10 nF, so

kz=1ω0Ct=116000×1×10−8=6250 Ωk_z = \frac{1}{\omega_0 C_t} = \frac{1}{16000 \times 1 \times 10^{-8}} = 6250\ \Omega Rnew=kzR,Cnew=Ckz ω0,Dnew=Dkz ω02\begin{aligned} R_{new} &= k_z R, \qquad C_{new} = \frac{C}{k_z\,\omega_0}, \qquad D_{new} = \frac{D}{k_z\,\omega_0^2} \end{aligned}

Step 3: Element values. Each FDNR is a GIC with C1=C5=CC_1 = C_5 = C and R2=R3=R4=RDR_2 = R_3 = R_4 = R_D, so D=C2RDD = C^2R_D. Taking C=CtC = C_t gives RD=Dnew/Ct2=kzDR_D = D_{new}/C_t^2 = k_z D.

PrototypeAfter BrutonScaledFinal element
Rs=1 ΩR_s = 1\ \OmegaCs=1C_s = 1 F1/(kzω0)1/(k_z\omega_0)10 nF
L1=0.7654L_1 = 0.7654 H (series)R1=0.7654 ΩR_1 = 0.7654\ \OmegaR1=0.7654×6250R_1 = 0.7654 \times 62504.784 kΩ
C2=1.848C_2 = 1.848 F (shunt)FDNR D2=1.848D_2 = 1.848D2=1.848/(kzω02)=1.155×10−12D_2 = 1.848/(k_z\omega_0^2) = 1.155 \times 10^{-12}GIC: C = 10 nF, RDR_D = 11.55 kΩ
L3=1.848L_3 = 1.848 H (series)R3=1.848 ΩR_3 = 1.848\ \OmegaR3=1.848×6250R_3 = 1.848 \times 625011.55 kΩ
C4=0.7654C_4 = 0.7654 F (shunt)FDNR D4=0.7654D_4 = 0.7654D4=0.7654/(kzω02)=4.784×10−13D_4 = 0.7654/(k_z\omega_0^2) = 4.784 \times 10^{-13}GIC: C = 10 nF, RDR_D = 4.784 kΩ
RL=1 ΩR_L = 1\ \OmegaCL=1C_L = 1 F1/(kzω0)1/(k_z\omega_0)10 nF

Step 4: DC path. The source and load capacitors block DC, and the op-amps in the GICs need a DC bias path. Put RaR_a across the source capacitor and RbR_b across the load capacitor. They must be much larger than kzk_z (the capacitor impedance at ω0\omega_0). At DC the FDNRs are open, so the DC gain is Rb/(Ra+ΣR+Rb)R_b/(R_a + \Sigma R + R_b). To keep the prototype DC gain of 1/2, set Rb=Ra+ΣRR_b = R_a + \Sigma R:

RaR_a = 100 kΩ, ΣR\Sigma R = 16.33 kΩ, so RbR_b = 116.3 kΩ.

Final circuit (Cs∥RaC_s \| R_a in series from the source, CL∥RbC_L \| R_b as the load, each DkD_k a GIC FDNR to ground):

Vi o--[Cs||Ra]--[R1]-+----[R3]-+-------+----o Vo
                     |         |       |
                    [D2]      [D4]  [CL||Rb]
GND --------------------------------------------

Put capacitors in positions 1 and 5 (Z1=1/sC1Z_1 = 1/sC_1, Z5=1/sC5Z_5 = 1/sC_5) and resistors in 2, 3, 4:

Zin=R3s2C1C5R2R4=1s2D,D=C1C5R2R4R3Z_{in} = \frac{R_3}{s^2 C_1 C_5 R_2 R_4} = \frac{1}{s^2 D}, \qquad D = \frac{C_1 C_5 R_2 R_4}{R_3}

With C1=C5=CC_1 = C_5 = C and R2=R3=R4=RR_2 = R_3 = R_4 = R: D=C2RD = C^2 R. At s=jωs = j\omega, Zin=−1/(ω2D)Z_{in} = -1/(\omega^2 D), a negative resistance whose value depends on frequency: an FDNR (also called a D-element or supercapacitor).

Answer: terminating capacitors 10 nF; series resistors 4.784 kΩ and 11.55 kΩ; FDNRs: D2D_2 = GIC (C = 10 nF, RDR_D = 11.55 kΩ), D4D_4 = GIC (C = 10 nF, RDR_D = 4.784 kΩ); RaR_a = 100 kΩ, RbR_b = 116.3 kΩ.

  • 2075 Asoj · 6 marks

Simulate the Butterworth 4th order low pass filter in resistively-terminated lossless network using FDNR. [Table: doubly terminated Butterworth element values normalized to 1 rad/s, 1 Ω terminations: n = 2: 1.414, 1.414; n = 3: 1, 2, 1; n = 4: 0.7654, 1.848, 1.848, 0.7654; n = 5: 0.618, 1.618, 2, 1.618, 0.618]

Answer

The inductors of the resistively terminated LC ladder are removed with the Bruton transformation (divide every impedance by ss), and the grounded capacitors become FDNRs (Z=1/s2DZ = 1/s^2D), each realised by a GIC.

From the table (n = 4), the doubly terminated Butterworth prototype (1 Ω terminations, ω0=1\omega_0 = 1 rad/s) is Rs=1R_s = 1, L1=0.7654L_1 = 0.7654 H, C2=1.848C_2 = 1.848 F, L3=1.848L_3 = 1.848 H, C4=0.7654C_4 = 0.7654 F, RL=1 ΩR_L = 1\ \Omega (T-ladder: series L first, shunt C to ground).

Normalised FDNR network (ω0=1\omega_0 = 1 rad/s): source capacitor 1 F, R1=0.7654 ΩR_1 = 0.7654\ \Omega, FDNR D2=1.848D_2 = 1.848, R3=1.848 ΩR_3 = 1.848\ \Omega, FDNR D4=0.7654D_4 = 0.7654, load capacitor 1 F.

No frequency is given, so the design is shown denormalised to a practical example of ω0\omega_0 = 10,000 rad/s:

Step 1: Bruton transformation (divide all impedances by ss): source and load resistors (1 Ω) become 1 F capacitors, each inductor LL becomes a resistor of LL Ω, each shunt capacitor CC becomes a grounded FDNR with D=CD = C.

Step 2: Scaling. Frequency scaling factor ω0=104 rad/s\omega_0 = 10^4\ \text{rad/s}. Choose the terminating capacitors as CtC_t = 10 nF, so

kz=1ω0Ct=110000×1×10−8=10000 Ωk_z = \frac{1}{\omega_0 C_t} = \frac{1}{10000 \times 1 \times 10^{-8}} = 10000\ \Omega Rnew=kzR,Cnew=Ckz ω0,Dnew=Dkz ω02\begin{aligned} R_{new} &= k_z R, \qquad C_{new} = \frac{C}{k_z\,\omega_0}, \qquad D_{new} = \frac{D}{k_z\,\omega_0^2} \end{aligned}

Step 3: Element values. Each FDNR is a GIC with C1=C5=CC_1 = C_5 = C and R2=R3=R4=RDR_2 = R_3 = R_4 = R_D, so D=C2RDD = C^2R_D. Taking C=CtC = C_t gives RD=Dnew/Ct2=kzDR_D = D_{new}/C_t^2 = k_z D.

PrototypeAfter BrutonScaledFinal element
Rs=1 ΩR_s = 1\ \OmegaCs=1C_s = 1 F1/(kzω0)1/(k_z\omega_0)10 nF
L1=0.7654L_1 = 0.7654 H (series)R1=0.7654 ΩR_1 = 0.7654\ \OmegaR1=0.7654×10000R_1 = 0.7654 \times 100007.654 kΩ
C2=1.848C_2 = 1.848 F (shunt)FDNR D2=1.848D_2 = 1.848D2=1.848/(kzω02)=1.848×10−12D_2 = 1.848/(k_z\omega_0^2) = 1.848 \times 10^{-12}GIC: C = 10 nF, RDR_D = 18.48 kΩ
L3=1.848L_3 = 1.848 H (series)R3=1.848 ΩR_3 = 1.848\ \OmegaR3=1.848×10000R_3 = 1.848 \times 1000018.48 kΩ
C4=0.7654C_4 = 0.7654 F (shunt)FDNR D4=0.7654D_4 = 0.7654D4=0.7654/(kzω02)=7.654×10−13D_4 = 0.7654/(k_z\omega_0^2) = 7.654 \times 10^{-13}GIC: C = 10 nF, RDR_D = 7.654 kΩ
RL=1 ΩR_L = 1\ \OmegaCL=1C_L = 1 F1/(kzω0)1/(k_z\omega_0)10 nF

Step 4: DC path. The source and load capacitors block DC, and the op-amps in the GICs need a DC bias path. Put RaR_a across the source capacitor and RbR_b across the load capacitor. They must be much larger than kzk_z (the capacitor impedance at ω0\omega_0). At DC the FDNRs are open, so the DC gain is Rb/(Ra+ΣR+Rb)R_b/(R_a + \Sigma R + R_b). To keep the prototype DC gain of 1/2, set Rb=Ra+ΣRR_b = R_a + \Sigma R:

RaR_a = 1 MΩ, ΣR\Sigma R = 26.13 kΩ, so RbR_b = 1.026 MΩ.

Final circuit (Cs∥RaC_s \| R_a in series from the source, CL∥RbC_L \| R_b as the load, each DkD_k a GIC FDNR to ground):

Vi o--[Cs||Ra]--[R1]-+----[R3]-+-------+----o Vo
                     |         |       |
                    [D2]      [D4]  [CL||Rb]
GND --------------------------------------------

Put capacitors in positions 1 and 5 (Z1=1/sC1Z_1 = 1/sC_1, Z5=1/sC5Z_5 = 1/sC_5) and resistors in 2, 3, 4:

Zin=R3s2C1C5R2R4=1s2D,D=C1C5R2R4R3Z_{in} = \frac{R_3}{s^2 C_1 C_5 R_2 R_4} = \frac{1}{s^2 D}, \qquad D = \frac{C_1 C_5 R_2 R_4}{R_3}

With C1=C5=CC_1 = C_5 = C and R2=R3=R4=RR_2 = R_3 = R_4 = R: D=C2RD = C^2 R. At s=jωs = j\omega, Zin=−1/(ω2D)Z_{in} = -1/(\omega^2 D), a negative resistance whose value depends on frequency: an FDNR (also called a D-element or supercapacitor).

Answer (for ω0\omega_0 = 10⁴ rad/s): terminating capacitors 10 nF; series resistors 7.654 kΩ and 18.48 kΩ; FDNR GICs with C = 10 nF and RDR_D = 18.48 kΩ (D2D_2) and 7.654 kΩ (D4D_4); RaR_a = 1 MΩ, RbR_b = 1.026 MΩ. For any other ω0\omega_0, use kz=1/(ω0Ct)k_z = 1/(\omega_0 C_t) and the same steps.

  • 2073 Chaitra · 5+5 marks

What is frequency dependent negative resistor (FDNR)? How can it be used to avoid inductors in Lowpass LC ladder circuit? Explain. From the circuit given in figure 1 design the lowpass filter having ωo = 10⁴ rad/s and practical element values using FDNR. [Figure 1: source V1, R1 = 1 Ω, series L1 = 2.024 H, shunt C1 = 0.994 F, series L2 = 2.024 H, load R2 = 1 Ω]

Answer

FDNR

A frequency dependent negative resistor (FDNR) is a one-port with impedance Z(s)=1s2DZ(s) = \dfrac{1}{s^2D}. At s=jωs = j\omega, Z=−1ω2DZ = -\dfrac{1}{\omega^2 D}: a real, negative resistance whose size falls as 1/ω21/\omega^2. DD (unit F²Ω) is its constant. It cannot be made from passive elements; a GIC with two capacitors realises it.

Avoiding inductors in a lowpass LC ladder

Bruton transformation: divide every impedance of the RLC ladder by ss (more generally by ksks). A voltage transfer function is a ratio of impedances, so it does not change:

ElementImpedanceAfter dividing by ssNew element
Resistor RRRRR/sR/scapacitor C=1/RC = 1/R
Inductor LLsLsLLLresistor R=LR = L
Capacitor CC1/sC1/sC1/(s2C)1/(s^2C)FDNR D=CD = C

In a lowpass ladder the inductors are in the series arms (often floating) and the capacitors are shunt (grounded). After the transformation the inductors become plain resistors and the grounded capacitors become grounded FDNRs, which one GIC can realise. So no inductor is left.

Put capacitors in positions 1 and 5 (Z1=1/sC1Z_1 = 1/sC_1, Z5=1/sC5Z_5 = 1/sC_5) and resistors in 2, 3, 4:

Zin=R3s2C1C5R2R4=1s2D,D=C1C5R2R4R3Z_{in} = \frac{R_3}{s^2 C_1 C_5 R_2 R_4} = \frac{1}{s^2 D}, \qquad D = \frac{C_1 C_5 R_2 R_4}{R_3}

With C1=C5=CC_1 = C_5 = C and R2=R3=R4=RR_2 = R_3 = R_4 = R: D=C2RD = C^2 R. At s=jωs = j\omega, Zin=−1/(ω2D)Z_{in} = -1/(\omega^2 D), a negative resistance whose value depends on frequency: an FDNR (also called a D-element or supercapacitor).

Design from Figure 1, ω0\omega_0 = 10⁴ rad/s

Prototype (Figure 1): R1=1 ΩR_1 = 1\ \Omega, series L1=2.024L_1 = 2.024 H, shunt C1=0.994C_1 = 0.994 F, series L2=2.024L_2 = 2.024 H, R2=1 ΩR_2 = 1\ \Omega (a 3rd-order Chebyshev ladder). Numbering the elements 1, 2, 3 from the source:

Step 1: Bruton transformation (divide all impedances by ss): source and load resistors (1 Ω) become 1 F capacitors, each inductor LL becomes a resistor of LL Ω, each shunt capacitor CC becomes a grounded FDNR with D=CD = C.

Step 2: Scaling. Frequency scaling factor ω0=104 rad/s\omega_0 = 10^4\ \text{rad/s}. Choose the terminating capacitors as CtC_t = 10 nF, so

kz=1ω0Ct=110000×1×10−8=10000 Ωk_z = \frac{1}{\omega_0 C_t} = \frac{1}{10000 \times 1 \times 10^{-8}} = 10000\ \Omega Rnew=kzR,Cnew=Ckz ω0,Dnew=Dkz ω02\begin{aligned} R_{new} &= k_z R, \qquad C_{new} = \frac{C}{k_z\,\omega_0}, \qquad D_{new} = \frac{D}{k_z\,\omega_0^2} \end{aligned}

Step 3: Element values. Each FDNR is a GIC with C1=C5=CC_1 = C_5 = C and R2=R3=R4=RDR_2 = R_3 = R_4 = R_D, so D=C2RDD = C^2R_D. Taking C=CtC = C_t gives RD=Dnew/Ct2=kzDR_D = D_{new}/C_t^2 = k_z D.

PrototypeAfter BrutonScaledFinal element
Rs=1 ΩR_s = 1\ \OmegaCs=1C_s = 1 F1/(kzω0)1/(k_z\omega_0)10 nF
L1=2.024L_1 = 2.024 H (series)R1=2.024 ΩR_1 = 2.024\ \OmegaR1=2.024×10000R_1 = 2.024 \times 1000020.24 kΩ
C2=0.994C_2 = 0.994 F (shunt)FDNR D2=0.994D_2 = 0.994D2=0.994/(kzω02)=9.94×10−13D_2 = 0.994/(k_z\omega_0^2) = 9.94 \times 10^{-13}GIC: C = 10 nF, RDR_D = 9.94 kΩ
L3=2.024L_3 = 2.024 H (series)R3=2.024 ΩR_3 = 2.024\ \OmegaR3=2.024×10000R_3 = 2.024 \times 1000020.24 kΩ
RL=1 ΩR_L = 1\ \OmegaCL=1C_L = 1 F1/(kzω0)1/(k_z\omega_0)10 nF

Step 4: DC path. The source and load capacitors block DC, and the op-amps in the GICs need a DC bias path. Put RaR_a across the source capacitor and RbR_b across the load capacitor. They must be much larger than kzk_z (the capacitor impedance at ω0\omega_0). At DC the FDNRs are open, so the DC gain is Rb/(Ra+ΣR+Rb)R_b/(R_a + \Sigma R + R_b). To keep the prototype DC gain of 1/2, set Rb=Ra+ΣRR_b = R_a + \Sigma R:

RaR_a = 1 MΩ, ΣR\Sigma R = 40.48 kΩ, so RbR_b = 1.04 MΩ.

Final circuit (Cs∥RaC_s \| R_a in series from the source, CL∥RbC_L \| R_b as the load, each DkD_k a GIC FDNR to ground):

Vi o--[Cs||Ra]--[R1]-+----[R3]-+----o Vo
                     |         |
                    [D2]    [CL||Rb]
GND ------------------------------------

Answer: terminating capacitors 10 nF; series resistors 20.24 kΩ each; FDNR = GIC with C = 10 nF and RDR_D = 9.94 kΩ; RaR_a = 1 MΩ, RbR_b = 1.04 MΩ.

  • 2072 Chaitra · 6 marks

Design a Fourth order Butterworth low pass filter having half power frequency of 4000 rad/s using Frequency dependent negative resistor (FDNR). [Table: doubly terminated Butterworth element values normalized to 1 rad/s, 1 Ω terminations: n = 2: 1.414, 1.414; n = 3: 1, 2, 1; n = 4: 0.7654, 1.848, 1.848, 0.7654; n = 5: 0.618, 1.618, 2, 1.618, 0.618]

Answer

The design uses the Bruton transformation to remove the inductors of the doubly terminated Butterworth ladder; the shunt capacitors become grounded FDNRs (Z=1/s2DZ = 1/s^2D), each realised with a GIC.

From the table (n = 4), the doubly terminated Butterworth prototype (1 Ω terminations, ω0=1\omega_0 = 1 rad/s) is Rs=1R_s = 1, L1=0.7654L_1 = 0.7654 H, C2=1.848C_2 = 1.848 F, L3=1.848L_3 = 1.848 H, C4=0.7654C_4 = 0.7654 F, RL=1 ΩR_L = 1\ \Omega (T-ladder: series L first, shunt C to ground).

Step 1: Bruton transformation (divide all impedances by ss): source and load resistors (1 Ω) become 1 F capacitors, each inductor LL becomes a resistor of LL Ω, each shunt capacitor CC becomes a grounded FDNR with D=CD = C.

Step 2: Scaling. Frequency scaling factor ω0=4000 rad/s\omega_0 = 4000\ \text{rad/s}. Choose the terminating capacitors as CtC_t = 100 nF, so

kz=1ω0Ct=14000×1×10−7=2500 Ωk_z = \frac{1}{\omega_0 C_t} = \frac{1}{4000 \times 1 \times 10^{-7}} = 2500\ \Omega Rnew=kzR,Cnew=Ckz ω0,Dnew=Dkz ω02\begin{aligned} R_{new} &= k_z R, \qquad C_{new} = \frac{C}{k_z\,\omega_0}, \qquad D_{new} = \frac{D}{k_z\,\omega_0^2} \end{aligned}

Step 3: Element values. Each FDNR is a GIC with C1=C5=CC_1 = C_5 = C and R2=R3=R4=RDR_2 = R_3 = R_4 = R_D, so D=C2RDD = C^2R_D. Taking C=CtC = C_t gives RD=Dnew/Ct2=kzDR_D = D_{new}/C_t^2 = k_z D.

PrototypeAfter BrutonScaledFinal element
Rs=1 ΩR_s = 1\ \OmegaCs=1C_s = 1 F1/(kzω0)1/(k_z\omega_0)100 nF
L1=0.7654L_1 = 0.7654 H (series)R1=0.7654 ΩR_1 = 0.7654\ \OmegaR1=0.7654×2500R_1 = 0.7654 \times 25001.913 kΩ
C2=1.848C_2 = 1.848 F (shunt)FDNR D2=1.848D_2 = 1.848D2=1.848/(kzω02)=4.62×10−11D_2 = 1.848/(k_z\omega_0^2) = 4.62 \times 10^{-11}GIC: C = 100 nF, RDR_D = 4.62 kΩ
L3=1.848L_3 = 1.848 H (series)R3=1.848 ΩR_3 = 1.848\ \OmegaR3=1.848×2500R_3 = 1.848 \times 25004.62 kΩ
C4=0.7654C_4 = 0.7654 F (shunt)FDNR D4=0.7654D_4 = 0.7654D4=0.7654/(kzω02)=1.914×10−11D_4 = 0.7654/(k_z\omega_0^2) = 1.914 \times 10^{-11}GIC: C = 100 nF, RDR_D = 1.913 kΩ
RL=1 ΩR_L = 1\ \OmegaCL=1C_L = 1 F1/(kzω0)1/(k_z\omega_0)100 nF

Step 4: DC path. The source and load capacitors block DC, and the op-amps in the GICs need a DC bias path. Put RaR_a across the source capacitor and RbR_b across the load capacitor. They must be much larger than kzk_z (the capacitor impedance at ω0\omega_0). At DC the FDNRs are open, so the DC gain is Rb/(Ra+ΣR+Rb)R_b/(R_a + \Sigma R + R_b). To keep the prototype DC gain of 1/2, set Rb=Ra+ΣRR_b = R_a + \Sigma R:

RaR_a = 100 kΩ, ΣR\Sigma R = 6.534 kΩ, so RbR_b = 106.5 kΩ.

Final circuit (Cs∥RaC_s \| R_a in series from the source, CL∥RbC_L \| R_b as the load, each DkD_k a GIC FDNR to ground):

Vi o--[Cs||Ra]--[R1]-+----[R3]-+-------+----o Vo
                     |         |       |
                    [D2]      [D4]  [CL||Rb]
GND --------------------------------------------

Put capacitors in positions 1 and 5 (Z1=1/sC1Z_1 = 1/sC_1, Z5=1/sC5Z_5 = 1/sC_5) and resistors in 2, 3, 4:

Zin=R3s2C1C5R2R4=1s2D,D=C1C5R2R4R3Z_{in} = \frac{R_3}{s^2 C_1 C_5 R_2 R_4} = \frac{1}{s^2 D}, \qquad D = \frac{C_1 C_5 R_2 R_4}{R_3}

With C1=C5=CC_1 = C_5 = C and R2=R3=R4=RR_2 = R_3 = R_4 = R: D=C2RD = C^2 R. At s=jωs = j\omega, Zin=−1/(ω2D)Z_{in} = -1/(\omega^2 D), a negative resistance whose value depends on frequency: an FDNR (also called a D-element or supercapacitor).

Answer: terminating capacitors 0.1 μF; series resistors 1.913 kΩ and 4.62 kΩ; FDNRs: D2D_2 = GIC (C = 0.1 μF, RDR_D = 4.62 kΩ), D4D_4 = GIC (C = 0.1 μF, RDR_D = 1.913 kΩ); RaR_a = 100 kΩ, RbR_b = 106.5 kΩ.

  • 2071 Chaitra · 5 marks

What is Frequency Dependent Negative Resistor? How can it be used to avoid bulky inductors in the design of your circuits? Explain with suitable examples.

Answer

Frequency dependent negative resistor

A frequency dependent negative resistor (FDNR) is a one-port with impedance Z(s)=1s2DZ(s) = \dfrac{1}{s^2D}. At s=jωs = j\omega, Z=−1ω2DZ = -\dfrac{1}{\omega^2 D}: a real, negative resistance whose size falls as 1/ω21/\omega^2. DD (unit F²Ω) is its constant. It cannot be made from passive elements; a GIC with two capacitors realises it.

Avoiding bulky inductors

Inductors for audio and low frequencies are large, heavy, lossy (low Q), pick up magnetic noise and cannot be put on an IC. The FDNR method removes them:

  1. Start from the doubly terminated LC ladder (it has the lowest sensitivity).
  2. Apply the Bruton transformation: divide every impedance by ss. Then R→C=1/RR \to C = 1/R, L→R=LL \to R = L, C→C \to FDNR D=CD = C. The voltage transfer function does not change.
  3. In a lowpass ladder all inductors are in series arms, so they all become ordinary resistors. The shunt capacitors become grounded FDNRs.
  4. Realise each FDNR with one GIC: capacitors in positions 1 and 5, resistors in 2, 3, 4, giving D=C2RD = C^2R.
  5. Add large resistors across the terminating capacitors for a DC path.
   Iin  node 1
 o-->-----+--------------(+) A1
 Vin      |
         [Z1]
          +--- node 2 ---- output of A1
         [Z2]
          +--- node 3 ---- (-) A1 and (-) A2
         [Z3]
          +--- node 4 ---- output of A2
         [Z4]
          +--- node 5 ---- (+) A2
         [Z5]
          |
         GND

Example

Third-order Butterworth LPF at ω0\omega_0 = 10,000 rad/s. Prototype: Rs=1R_s = 1, L1=1L_1 = 1 H, C2=2C_2 = 2 F, L3=1L_3 = 1 H, RL=1R_L = 1. Without FDNR the scaled inductors (with kz=104k_z = 10^4) would be L=kzL/ω0L = k_zL/\omega_0 = 1 H each, which is bulky.

With FDNR (CtC_t = 10 nF, kz=104 Ωk_z = 10^4\ \Omega):

ElementBrutonFinal
RsR_s1 F10 nF (with RaR_a = 1 MΩ across it)
L1L_1 = 1 H1 Ω10 kΩ
C2C_2 = 2 FDD = 2GIC: C = 10 nF, RDR_D = 20 kΩ
L3L_3 = 1 H1 Ω10 kΩ
RLR_L1 F10 nF (with RbR_b = 1.02 MΩ across it)

The 1 H inductors are replaced by 10 kΩ resistors and one GIC.

  • 2071 Shrawan · 2+4 marks

What is FDNR? Explain how FDNR avoids the use of inductor. Following circuit is a lowpass filter having half power frequency of 1 rad/sec. Obtain a lowpass filter having half power frequency of 5 kHz and largest capacitor of 0.01 μF using FDNR. [Figure: source V1, 1 Ω source resistor, shunt 0.618 F, series 1.618 H, shunt 2.0 F, series 1.618 H (labelled 1.618 F in the figure), shunt 0.618 F, 1 Ω load (output V2)]

Answer

FDNR and how it avoids inductors

A frequency dependent negative resistor (FDNR) is a one-port with impedance Z(s)=1s2DZ(s) = \dfrac{1}{s^2D}. At s=jωs = j\omega, Z=−1ω2DZ = -\dfrac{1}{\omega^2 D}: a real, negative resistance whose size falls as 1/ω21/\omega^2. DD (unit F²Ω) is its constant. It cannot be made from passive elements; a GIC with two capacitors realises it.

Using the Bruton transformation (divide all impedances by ss), each inductor LL becomes a resistor LL, each grounded capacitor CC becomes a grounded FDNR D=CD = C, and each terminating resistor becomes a capacitor. So the lowpass ladder needs no inductor at all; each FDNR is one GIC with D=C2RD = C^2R.

Design: 5 kHz, largest capacitor 0.01 μF

The given ladder is a 5th-order Butterworth LPF (π form): shunt C1=0.618C_1 = 0.618 F, series L2=1.618L_2 = 1.618 H, shunt C3=2C_3 = 2 F, series L4=1.618L_4 = 1.618 H (shown as F in the figure; it must be an inductor in a lowpass ladder), shunt C5=0.618C_5 = 0.618 F, 1 Ω terminations.

Step 1: Bruton transformation (divide all impedances by ss): source and load resistors (1 Ω) become 1 F capacitors, each inductor LL becomes a resistor of LL Ω, each shunt capacitor CC becomes a grounded FDNR with D=CD = C.

Step 2: Scaling. Frequency scaling factor ω0=2π(5000)=31,416 rad/s\omega_0 = 2\pi(5000) = 31{,}416\ \text{rad/s}. Choose the terminating capacitors as CtC_t = 10 nF (the largest capacitor allowed; all capacitors in the final circuit are 0.01 μF), so

kz=1ω0Ct=131415.9×1×10−8=3183.1 Ωk_z = \frac{1}{\omega_0 C_t} = \frac{1}{31415.9 \times 1 \times 10^{-8}} = 3183.1\ \Omega Rnew=kzR,Cnew=Ckz ω0,Dnew=Dkz ω02\begin{aligned} R_{new} &= k_z R, \qquad C_{new} = \frac{C}{k_z\,\omega_0}, \qquad D_{new} = \frac{D}{k_z\,\omega_0^2} \end{aligned}

Step 3: Element values. Each FDNR is a GIC with C1=C5=CC_1 = C_5 = C and R2=R3=R4=RDR_2 = R_3 = R_4 = R_D, so D=C2RDD = C^2R_D. Taking C=CtC = C_t gives RD=Dnew/Ct2=kzDR_D = D_{new}/C_t^2 = k_z D.

PrototypeAfter BrutonScaledFinal element
Rs=1 ΩR_s = 1\ \OmegaCs=1C_s = 1 F1/(kzω0)1/(k_z\omega_0)10 nF
C1=0.618C_1 = 0.618 F (shunt)FDNR D1=0.618D_1 = 0.618D1=0.618/(kzω02)=1.967×10−13D_1 = 0.618/(k_z\omega_0^2) = 1.967 \times 10^{-13}GIC: C = 10 nF, RDR_D = 1.967 kΩ
L2=1.618L_2 = 1.618 H (series)R2=1.618 ΩR_2 = 1.618\ \OmegaR2=1.618×3183.1R_2 = 1.618 \times 3183.15.15 kΩ
C3=2C_3 = 2 F (shunt)FDNR D3=2D_3 = 2D3=2/(kzω02)=6.366×10−13D_3 = 2/(k_z\omega_0^2) = 6.366 \times 10^{-13}GIC: C = 10 nF, RDR_D = 6.366 kΩ
L4=1.618L_4 = 1.618 H (series)R4=1.618 ΩR_4 = 1.618\ \OmegaR4=1.618×3183.1R_4 = 1.618 \times 3183.15.15 kΩ
C5=0.618C_5 = 0.618 F (shunt)FDNR D5=0.618D_5 = 0.618D5=0.618/(kzω02)=1.967×10−13D_5 = 0.618/(k_z\omega_0^2) = 1.967 \times 10^{-13}GIC: C = 10 nF, RDR_D = 1.967 kΩ
RL=1 ΩR_L = 1\ \OmegaCL=1C_L = 1 F1/(kzω0)1/(k_z\omega_0)10 nF

Step 4: DC path. The source and load capacitors block DC, and the op-amps in the GICs need a DC bias path. Put RaR_a across the source capacitor and RbR_b across the load capacitor. They must be much larger than kzk_z (the capacitor impedance at ω0\omega_0). At DC the FDNRs are open, so the DC gain is Rb/(Ra+ΣR+Rb)R_b/(R_a + \Sigma R + R_b). To keep the prototype DC gain of 1/2, set Rb=Ra+ΣRR_b = R_a + \Sigma R:

RaR_a = 100 kΩ, ΣR\Sigma R = 10.3 kΩ, so RbR_b = 110.3 kΩ.

Final circuit (Cs∥RaC_s \| R_a in series from the source, CL∥RbC_L \| R_b as the load, each DkD_k a GIC FDNR to ground):

Vi o--[Cs||Ra]-+----[R2]-+----[R4]-+-------+----o Vo
               |         |         |       |
              [D1]      [D3]      [D5]  [CL||Rb]
GND ------------------------------------------------

Answer: all capacitors 0.01 μF; series resistors 5.15 kΩ each; FDNR GICs with C = 0.01 μF and RDR_D = 1.967 kΩ, 6.366 kΩ, 1.967 kΩ; RaR_a = 100 kΩ, RbR_b = 110.3 kΩ.

  • 2070 Asar · 3+4 marks

What is FDNR? How FDNR avoids the use of inductor? Explain. Following circuit is a lowpass filter having half power frequency of 1 rad/sec. Obtain a lowpass filter having half power frequency of 4.5 kHz and largest capacitor of 0.1 μF using FDNR. [Figure: source V1, 1 Ω source resistor, series 0.7654 H, shunt 1.848 F, series 1.848 H, shunt 0.7654 F, 1 Ω load (output V2)]

Answer

FDNR and how it avoids inductors

A frequency dependent negative resistor (FDNR) is a one-port with impedance Z(s)=1s2DZ(s) = \dfrac{1}{s^2D}. At s=jωs = j\omega, Z=−1ω2DZ = -\dfrac{1}{\omega^2 D}: a real, negative resistance whose size falls as 1/ω21/\omega^2. DD (unit F²Ω) is its constant. It cannot be made from passive elements; a GIC with two capacitors realises it.

Under the Bruton transformation (all impedances divided by ss) a resistor becomes a capacitor, an inductor becomes a resistor and a capacitor becomes an FDNR. A lowpass ladder therefore turns into a network of R, C and grounded FDNRs with the same transfer function, and each FDNR is a GIC (D=C2RD = C^2R). No inductor remains.

Design: 4.5 kHz, largest capacitor 0.1 μF

Given prototype (4th-order Butterworth, 1 rad/s): series L1=0.7654L_1 = 0.7654 H, shunt C2=1.848C_2 = 1.848 F, series L3=1.848L_3 = 1.848 H, shunt C4=0.7654C_4 = 0.7654 F, 1 Ω terminations.

Step 1: Bruton transformation (divide all impedances by ss): source and load resistors (1 Ω) become 1 F capacitors, each inductor LL becomes a resistor of LL Ω, each shunt capacitor CC becomes a grounded FDNR with D=CD = C.

Step 2: Scaling. Frequency scaling factor ω0=2π(4500)=28,274 rad/s\omega_0 = 2\pi(4500) = 28{,}274\ \text{rad/s}. Choose the terminating capacitors as CtC_t = 100 nF (the largest capacitor allowed; all capacitors are made 0.1 μF), so

kz=1ω0Ct=128274.3×1×10−7=353.68 Ωk_z = \frac{1}{\omega_0 C_t} = \frac{1}{28274.3 \times 1 \times 10^{-7}} = 353.68\ \Omega Rnew=kzR,Cnew=Ckz ω0,Dnew=Dkz ω02\begin{aligned} R_{new} &= k_z R, \qquad C_{new} = \frac{C}{k_z\,\omega_0}, \qquad D_{new} = \frac{D}{k_z\,\omega_0^2} \end{aligned}

Step 3: Element values. Each FDNR is a GIC with C1=C5=CC_1 = C_5 = C and R2=R3=R4=RDR_2 = R_3 = R_4 = R_D, so D=C2RDD = C^2R_D. Taking C=CtC = C_t gives RD=Dnew/Ct2=kzDR_D = D_{new}/C_t^2 = k_z D.

PrototypeAfter BrutonScaledFinal element
Rs=1 ΩR_s = 1\ \OmegaCs=1C_s = 1 F1/(kzω0)1/(k_z\omega_0)100 nF
L1=0.7654L_1 = 0.7654 H (series)R1=0.7654 ΩR_1 = 0.7654\ \OmegaR1=0.7654×353.68R_1 = 0.7654 \times 353.68270.7 Ω
C2=1.848C_2 = 1.848 F (shunt)FDNR D2=1.848D_2 = 1.848D2=1.848/(kzω02)=6.536×10−12D_2 = 1.848/(k_z\omega_0^2) = 6.536 \times 10^{-12}GIC: C = 100 nF, RDR_D = 653.6 Ω
L3=1.848L_3 = 1.848 H (series)R3=1.848 ΩR_3 = 1.848\ \OmegaR3=1.848×353.68R_3 = 1.848 \times 353.68653.6 Ω
C4=0.7654C_4 = 0.7654 F (shunt)FDNR D4=0.7654D_4 = 0.7654D4=0.7654/(kzω02)=2.707×10−12D_4 = 0.7654/(k_z\omega_0^2) = 2.707 \times 10^{-12}GIC: C = 100 nF, RDR_D = 270.7 Ω
RL=1 ΩR_L = 1\ \OmegaCL=1C_L = 1 F1/(kzω0)1/(k_z\omega_0)100 nF

Step 4: DC path. The source and load capacitors block DC, and the op-amps in the GICs need a DC bias path. Put RaR_a across the source capacitor and RbR_b across the load capacitor. They must be much larger than kzk_z (the capacitor impedance at ω0\omega_0). At DC the FDNRs are open, so the DC gain is Rb/(Ra+ΣR+Rb)R_b/(R_a + \Sigma R + R_b). To keep the prototype DC gain of 1/2, set Rb=Ra+ΣRR_b = R_a + \Sigma R:

RaR_a = 10 kΩ, ΣR\Sigma R = 924.3 Ω, so RbR_b = 10.92 kΩ.

Final circuit (Cs∥RaC_s \| R_a in series from the source, CL∥RbC_L \| R_b as the load, each DkD_k a GIC FDNR to ground):

Vi o--[Cs||Ra]--[R1]-+----[R3]-+-------+----o Vo
                     |         |       |
                    [D2]      [D4]  [CL||Rb]
GND --------------------------------------------

Answer: all capacitors 0.1 μF; series resistors 270.7 Ω and 653.6 Ω; FDNR GICs with C = 0.1 μF and RDR_D = 653.6 Ω (D2D_2) and 270.7 Ω (D4D_4); RaR_a = 10 kΩ, RbR_b = 10.92 kΩ.

  • 2080 Bhadra · 2+4 marks

Draw the figure of RLC series circuit using FDNR. Use fourth order Butterworth low pass by FDNR with half power = 1×10³ Hz and capacitor in final design = 0.01μF. [Figure: source V1, 1 Ω source resistor, shunt 0.7654 F, series 1.848 H, shunt 1.848 F, series 0.7654 H, 1 Ω load]

Answer

RLC elements using FDNR (Bruton transformation)

Dividing every impedance by ss keeps the voltage transfer function the same and changes each element as follows:

 Element     Impedance    /s          New element
 --[ R ]--     R          R/s         --| C=1/R |--
 --(LLL)--     sL         L           --[ R=L ]--
 --| C |--     1/sC       1/(s^2 C)   --[FDNR D=C]--

So a series R–L–C branch becomes a series C–R–FDNR branch:

 Before:  o--[ R ]--(LLL)--| C |--o
 After:   o--| 1/R |--[ L ]--[D = C]--o

In a lowpass ladder the capacitors are grounded, so the FDNRs are grounded and each is one GIC (D=C2RDD = C^2R_D).

Design: 4th-order Butterworth LPF, 1 kHz, capacitors 0.01 μF

Given ladder (1 rad/s, 1 Ω): shunt C1=0.7654C_1 = 0.7654 F, series L2=1.848L_2 = 1.848 H, shunt C3=1.848C_3 = 1.848 F, series L4=0.7654L_4 = 0.7654 H, load 1 Ω.

Step 1: Bruton transformation (divide all impedances by ss): source and load resistors (1 Ω) become 1 F capacitors, each inductor LL becomes a resistor of LL Ω, each shunt capacitor CC becomes a grounded FDNR with D=CD = C.

Step 2: Scaling. Frequency scaling factor ω0=2π(1000)=6283.2 rad/s\omega_0 = 2\pi(1000) = 6283.2\ \text{rad/s}. Choose the terminating capacitors as CtC_t = 10 nF (all capacitors in the final design), so

kz=1ω0Ct=16283.19×1×10−8=15915 Ωk_z = \frac{1}{\omega_0 C_t} = \frac{1}{6283.19 \times 1 \times 10^{-8}} = 15915\ \Omega Rnew=kzR,Cnew=Ckz ω0,Dnew=Dkz ω02\begin{aligned} R_{new} &= k_z R, \qquad C_{new} = \frac{C}{k_z\,\omega_0}, \qquad D_{new} = \frac{D}{k_z\,\omega_0^2} \end{aligned}

Step 3: Element values. Each FDNR is a GIC with C1=C5=CC_1 = C_5 = C and R2=R3=R4=RDR_2 = R_3 = R_4 = R_D, so D=C2RDD = C^2R_D. Taking C=CtC = C_t gives RD=Dnew/Ct2=kzDR_D = D_{new}/C_t^2 = k_z D.

PrototypeAfter BrutonScaledFinal element
Rs=1 ΩR_s = 1\ \OmegaCs=1C_s = 1 F1/(kzω0)1/(k_z\omega_0)10 nF
C1=0.7654C_1 = 0.7654 F (shunt)FDNR D1=0.7654D_1 = 0.7654D1=0.7654/(kzω02)=1.218×10−12D_1 = 0.7654/(k_z\omega_0^2) = 1.218 \times 10^{-12}GIC: C = 10 nF, RDR_D = 12.18 kΩ
L2=1.848L_2 = 1.848 H (series)R2=1.848 ΩR_2 = 1.848\ \OmegaR2=1.848×15915R_2 = 1.848 \times 1591529.41 kΩ
C3=1.848C_3 = 1.848 F (shunt)FDNR D3=1.848D_3 = 1.848D3=1.848/(kzω02)=2.941×10−12D_3 = 1.848/(k_z\omega_0^2) = 2.941 \times 10^{-12}GIC: C = 10 nF, RDR_D = 29.41 kΩ
L4=0.7654L_4 = 0.7654 H (series)R4=0.7654 ΩR_4 = 0.7654\ \OmegaR4=0.7654×15915R_4 = 0.7654 \times 1591512.18 kΩ
RL=1 ΩR_L = 1\ \OmegaCL=1C_L = 1 F1/(kzω0)1/(k_z\omega_0)10 nF

Step 4: DC path. The source and load capacitors block DC, and the op-amps in the GICs need a DC bias path. Put RaR_a across the source capacitor and RbR_b across the load capacitor. They must be much larger than kzk_z (the capacitor impedance at ω0\omega_0). At DC the FDNRs are open, so the DC gain is Rb/(Ra+ΣR+Rb)R_b/(R_a + \Sigma R + R_b). To keep the prototype DC gain of 1/2, set Rb=Ra+ΣRR_b = R_a + \Sigma R:

RaR_a = 1 MΩ, ΣR\Sigma R = 41.59 kΩ, so RbR_b = 1.042 MΩ.

Final circuit (Cs∥RaC_s \| R_a in series from the source, CL∥RbC_L \| R_b as the load, each DkD_k a GIC FDNR to ground):

Vi o--[Cs||Ra]-+----[R2]-+----[R4]-+----o Vo
               |         |         |
              [D1]      [D3]    [CL||Rb]
GND ----------------------------------------

Answer: all capacitors 0.01 μF; series resistors 29.41 kΩ and 12.18 kΩ; FDNR GICs with C = 0.01 μF and RDR_D = 12.18 kΩ (D1D_1) and 29.41 kΩ (D3D_3); RaR_a = 1 MΩ, RbR_b = 1.042 MΩ.

  • 2080 Baisakh · 2+3+5 marks

What is FDNR? How can FDNR be used to avoid the inductor in passive filter? Simulate the Butterworth 5th order lowpass filter using FDNR referring table below. Your final design should have ωo = 10,000 rad/s and practically realizable elements. [Table: Order (n) = 5; R1 = R2 = 1; L1 = 0.6180; C2 = 1.618; L3 = 2.000; C4 = 1.618; L5 = 0.6180]

Answer

FDNR

A frequency dependent negative resistor (FDNR) is a one-port with impedance Z(s)=1s2DZ(s) = \dfrac{1}{s^2D}. At s=jωs = j\omega, Z=−1ω2DZ = -\dfrac{1}{\omega^2 D}: a real, negative resistance whose size falls as 1/ω21/\omega^2. DD (unit F²Ω) is its constant. It is realised with a GIC containing two capacitors.

Avoiding the inductor

Bruton transformation: divide every impedance of the RLC ladder by ss (more generally by ksks). A voltage transfer function is a ratio of impedances, so it does not change:

ElementImpedanceAfter dividing by ssNew element
Resistor RRRRR/sR/scapacitor C=1/RC = 1/R
Inductor LLsLsLLLresistor R=LR = L
Capacitor CC1/sC1/sC1/(s2C)1/(s^2C)FDNR D=CD = C

In a lowpass ladder the inductors are in the series arms (often floating) and the capacitors are shunt (grounded). After the transformation the inductors become plain resistors and the grounded capacitors become grounded FDNRs, which one GIC can realise. So no inductor is left.

Put capacitors in positions 1 and 5 (Z1=1/sC1Z_1 = 1/sC_1, Z5=1/sC5Z_5 = 1/sC_5) and resistors in 2, 3, 4:

Zin=R3s2C1C5R2R4=1s2D,D=C1C5R2R4R3Z_{in} = \frac{R_3}{s^2 C_1 C_5 R_2 R_4} = \frac{1}{s^2 D}, \qquad D = \frac{C_1 C_5 R_2 R_4}{R_3}

With C1=C5=CC_1 = C_5 = C and R2=R3=R4=RR_2 = R_3 = R_4 = R: D=C2RD = C^2 R. At s=jωs = j\omega, Zin=−1/(ω2D)Z_{in} = -1/(\omega^2 D), a negative resistance whose value depends on frequency: an FDNR (also called a D-element or supercapacitor).

Design: 5th-order Butterworth LPF, ω0\omega_0 = 10,000 rad/s

Prototype from the table: R1=1R_1 = 1, L1=0.618L_1 = 0.618 H, C2=1.618C_2 = 1.618 F, L3=2L_3 = 2 H, C4=1.618C_4 = 1.618 F, L5=0.618L_5 = 0.618 H, R2=1 ΩR_2 = 1\ \Omega.

Step 1: Bruton transformation (divide all impedances by ss): source and load resistors (1 Ω) become 1 F capacitors, each inductor LL becomes a resistor of LL Ω, each shunt capacitor CC becomes a grounded FDNR with D=CD = C.

Step 2: Scaling. Frequency scaling factor ω0=104 rad/s\omega_0 = 10^4\ \text{rad/s}. Choose the terminating capacitors as CtC_t = 10 nF, so

kz=1ω0Ct=110000×1×10−8=10000 Ωk_z = \frac{1}{\omega_0 C_t} = \frac{1}{10000 \times 1 \times 10^{-8}} = 10000\ \Omega Rnew=kzR,Cnew=Ckz ω0,Dnew=Dkz ω02\begin{aligned} R_{new} &= k_z R, \qquad C_{new} = \frac{C}{k_z\,\omega_0}, \qquad D_{new} = \frac{D}{k_z\,\omega_0^2} \end{aligned}

Step 3: Element values. Each FDNR is a GIC with C1=C5=CC_1 = C_5 = C and R2=R3=R4=RDR_2 = R_3 = R_4 = R_D, so D=C2RDD = C^2R_D. Taking C=CtC = C_t gives RD=Dnew/Ct2=kzDR_D = D_{new}/C_t^2 = k_z D.

PrototypeAfter BrutonScaledFinal element
Rs=1 ΩR_s = 1\ \OmegaCs=1C_s = 1 F1/(kzω0)1/(k_z\omega_0)10 nF
L1=0.618L_1 = 0.618 H (series)R1=0.618 ΩR_1 = 0.618\ \OmegaR1=0.618×10000R_1 = 0.618 \times 100006.18 kΩ
C2=1.618C_2 = 1.618 F (shunt)FDNR D2=1.618D_2 = 1.618D2=1.618/(kzω02)=1.618×10−12D_2 = 1.618/(k_z\omega_0^2) = 1.618 \times 10^{-12}GIC: C = 10 nF, RDR_D = 16.18 kΩ
L3=2L_3 = 2 H (series)R3=2 ΩR_3 = 2\ \OmegaR3=2×10000R_3 = 2 \times 1000020 kΩ
C4=1.618C_4 = 1.618 F (shunt)FDNR D4=1.618D_4 = 1.618D4=1.618/(kzω02)=1.618×10−12D_4 = 1.618/(k_z\omega_0^2) = 1.618 \times 10^{-12}GIC: C = 10 nF, RDR_D = 16.18 kΩ
L5=0.618L_5 = 0.618 H (series)R5=0.618 ΩR_5 = 0.618\ \OmegaR5=0.618×10000R_5 = 0.618 \times 100006.18 kΩ
RL=1 ΩR_L = 1\ \OmegaCL=1C_L = 1 F1/(kzω0)1/(k_z\omega_0)10 nF

Step 4: DC path. The source and load capacitors block DC, and the op-amps in the GICs need a DC bias path. Put RaR_a across the source capacitor and RbR_b across the load capacitor. They must be much larger than kzk_z (the capacitor impedance at ω0\omega_0). At DC the FDNRs are open, so the DC gain is Rb/(Ra+ΣR+Rb)R_b/(R_a + \Sigma R + R_b). To keep the prototype DC gain of 1/2, set Rb=Ra+ΣRR_b = R_a + \Sigma R:

RaR_a = 1 MΩ, ΣR\Sigma R = 32.36 kΩ, so RbR_b = 1.032 MΩ.

Final circuit (Cs∥RaC_s \| R_a in series from the source, CL∥RbC_L \| R_b as the load, each DkD_k a GIC FDNR to ground):

Vi o--[Cs||Ra]--[R1]-+----[R3]-+----[R5]-+----o Vo
                     |         |         |
                    [D2]      [D4]    [CL||Rb]
GND ----------------------------------------------

Answer: terminating capacitors 10 nF; series resistors 6.18 kΩ, 20 kΩ, 6.18 kΩ; two FDNRs, each a GIC with C = 10 nF and RDR_D = 16.18 kΩ; RaR_a = 1 MΩ, RbR_b = 1.032 MΩ.

  • 2078 Bhadra · 2+5 marks

What do you mean by Frequency Dependent Negative Resistor (FDNR)? Design the 4th order Butterworth low pass filter with ωp = 20,000 rad/sec using FDNR. In your final circuit all the elements should be practically realizable. [Table: doubly terminated Butterworth element values normalized to 1 rad/s, 1 Ω terminations: n = 2: 1.414, 1.414; n = 3: 1, 2, 1; n = 4: 0.7654, 1.848, 1.848, 0.7654; n = 5: 0.618, 1.618, 2, 1.618, 0.618]

Answer

FDNR

A frequency dependent negative resistor (FDNR) is a one-port with impedance Z(s)=1s2DZ(s) = \dfrac{1}{s^2D}. At s=jωs = j\omega, Z=−1ω2DZ = -\dfrac{1}{\omega^2 D}: a real, negative resistance whose size falls as 1/ω21/\omega^2. DD (unit F²Ω) is its constant. It is realised with a GIC containing two capacitors.

It is used with the Bruton transformation (divide all impedances by ss): inductors become resistors and grounded capacitors become grounded FDNRs, so the ladder has no inductors.

Design: 4th-order Butterworth LPF, ωp\omega_p = 20,000 rad/s

From the table (n = 4), the doubly terminated Butterworth prototype (1 Ω terminations, ω0=1\omega_0 = 1 rad/s) is Rs=1R_s = 1, L1=0.7654L_1 = 0.7654 H, C2=1.848C_2 = 1.848 F, L3=1.848L_3 = 1.848 H, C4=0.7654C_4 = 0.7654 F, RL=1 ΩR_L = 1\ \Omega (T-ladder: series L first, shunt C to ground).

Step 1: Bruton transformation (divide all impedances by ss): source and load resistors (1 Ω) become 1 F capacitors, each inductor LL becomes a resistor of LL Ω, each shunt capacitor CC becomes a grounded FDNR with D=CD = C.

Step 2: Scaling. Frequency scaling factor ω0=20,000 rad/s\omega_0 = 20{,}000\ \text{rad/s}. Choose the terminating capacitors as CtC_t = 10 nF, so

kz=1ω0Ct=120000×1×10−8=5000 Ωk_z = \frac{1}{\omega_0 C_t} = \frac{1}{20000 \times 1 \times 10^{-8}} = 5000\ \Omega Rnew=kzR,Cnew=Ckz ω0,Dnew=Dkz ω02\begin{aligned} R_{new} &= k_z R, \qquad C_{new} = \frac{C}{k_z\,\omega_0}, \qquad D_{new} = \frac{D}{k_z\,\omega_0^2} \end{aligned}

Step 3: Element values. Each FDNR is a GIC with C1=C5=CC_1 = C_5 = C and R2=R3=R4=RDR_2 = R_3 = R_4 = R_D, so D=C2RDD = C^2R_D. Taking C=CtC = C_t gives RD=Dnew/Ct2=kzDR_D = D_{new}/C_t^2 = k_z D.

PrototypeAfter BrutonScaledFinal element
Rs=1 ΩR_s = 1\ \OmegaCs=1C_s = 1 F1/(kzω0)1/(k_z\omega_0)10 nF
L1=0.7654L_1 = 0.7654 H (series)R1=0.7654 ΩR_1 = 0.7654\ \OmegaR1=0.7654×5000R_1 = 0.7654 \times 50003.827 kΩ
C2=1.848C_2 = 1.848 F (shunt)FDNR D2=1.848D_2 = 1.848D2=1.848/(kzω02)=9.24×10−13D_2 = 1.848/(k_z\omega_0^2) = 9.24 \times 10^{-13}GIC: C = 10 nF, RDR_D = 9.24 kΩ
L3=1.848L_3 = 1.848 H (series)R3=1.848 ΩR_3 = 1.848\ \OmegaR3=1.848×5000R_3 = 1.848 \times 50009.24 kΩ
C4=0.7654C_4 = 0.7654 F (shunt)FDNR D4=0.7654D_4 = 0.7654D4=0.7654/(kzω02)=3.827×10−13D_4 = 0.7654/(k_z\omega_0^2) = 3.827 \times 10^{-13}GIC: C = 10 nF, RDR_D = 3.827 kΩ
RL=1 ΩR_L = 1\ \OmegaCL=1C_L = 1 F1/(kzω0)1/(k_z\omega_0)10 nF

Step 4: DC path. The source and load capacitors block DC, and the op-amps in the GICs need a DC bias path. Put RaR_a across the source capacitor and RbR_b across the load capacitor. They must be much larger than kzk_z (the capacitor impedance at ω0\omega_0). At DC the FDNRs are open, so the DC gain is Rb/(Ra+ΣR+Rb)R_b/(R_a + \Sigma R + R_b). To keep the prototype DC gain of 1/2, set Rb=Ra+ΣRR_b = R_a + \Sigma R:

RaR_a = 100 kΩ, ΣR\Sigma R = 13.07 kΩ, so RbR_b = 113.1 kΩ.

Final circuit (Cs∥RaC_s \| R_a in series from the source, CL∥RbC_L \| R_b as the load, each DkD_k a GIC FDNR to ground):

Vi o--[Cs||Ra]--[R1]-+----[R3]-+-------+----o Vo
                     |         |       |
                    [D2]      [D4]  [CL||Rb]
GND --------------------------------------------

Put capacitors in positions 1 and 5 (Z1=1/sC1Z_1 = 1/sC_1, Z5=1/sC5Z_5 = 1/sC_5) and resistors in 2, 3, 4:

Zin=R3s2C1C5R2R4=1s2D,D=C1C5R2R4R3Z_{in} = \frac{R_3}{s^2 C_1 C_5 R_2 R_4} = \frac{1}{s^2 D}, \qquad D = \frac{C_1 C_5 R_2 R_4}{R_3}

With C1=C5=CC_1 = C_5 = C and R2=R3=R4=RR_2 = R_3 = R_4 = R: D=C2RD = C^2 R. At s=jωs = j\omega, Zin=−1/(ω2D)Z_{in} = -1/(\omega^2 D), a negative resistance whose value depends on frequency: an FDNR (also called a D-element or supercapacitor).

Answer: terminating capacitors 10 nF; series resistors 3.827 kΩ and 9.24 kΩ; FDNRs: D2D_2 = GIC (C = 10 nF, RDR_D = 9.24 kΩ), D4D_4 = GIC (C = 10 nF, RDR_D = 3.827 kΩ); RaR_a = 100 kΩ, RbR_b = 113.1 kΩ.

  • 2076 Asoj · 2+2+4 marks

What is Frequency Dependent Negative Resistor (FDNR)? How can it avoid the use of inductors? Explain. Design a lowpass filter having ω0 of 10⁴ rad/s using FDNR from the filter circuit given below. In your final design, all the elements should be practically realizable. [Figure: 1 Ω source resistor, shunt 1 F capacitor, series 2 H inductor, shunt 1 F capacitor, 1 Ω load]

Answer

FDNR

A frequency dependent negative resistor (FDNR) is a one-port with impedance Z(s)=1s2DZ(s) = \dfrac{1}{s^2D}. At s=jωs = j\omega, Z=−1ω2DZ = -\dfrac{1}{\omega^2 D}: a real, negative resistance whose size falls as 1/ω21/\omega^2. DD (unit F²Ω) is its constant. It is realised with a GIC containing two capacitors.

How it avoids inductors

Under the Bruton transformation every impedance is divided by ss: R→C=1/RR \to C = 1/R, L→R=LL \to R = L, C→C \to FDNR D=CD = C. The voltage ratio is unchanged, the inductor becomes a resistor, and the grounded capacitors become grounded FDNRs, each made with one GIC (capacitors at positions 1 and 5, D=C2RD = C^2R).

Design: ω0\omega_0 = 10⁴ rad/s

Given ladder (3rd-order Butterworth, π form): Rs=1 ΩR_s = 1\ \Omega, shunt C1=1C_1 = 1 F, series L2=2L_2 = 2 H, shunt C3=1C_3 = 1 F, RL=1 ΩR_L = 1\ \Omega.

Step 1: Bruton transformation (divide all impedances by ss): source and load resistors (1 Ω) become 1 F capacitors, each inductor LL becomes a resistor of LL Ω, each shunt capacitor CC becomes a grounded FDNR with D=CD = C.

Step 2: Scaling. Frequency scaling factor ω0=104 rad/s\omega_0 = 10^4\ \text{rad/s}. Choose the terminating capacitors as CtC_t = 10 nF, so

kz=1ω0Ct=110000×1×10−8=10000 Ωk_z = \frac{1}{\omega_0 C_t} = \frac{1}{10000 \times 1 \times 10^{-8}} = 10000\ \Omega Rnew=kzR,Cnew=Ckz ω0,Dnew=Dkz ω02\begin{aligned} R_{new} &= k_z R, \qquad C_{new} = \frac{C}{k_z\,\omega_0}, \qquad D_{new} = \frac{D}{k_z\,\omega_0^2} \end{aligned}

Step 3: Element values. Each FDNR is a GIC with C1=C5=CC_1 = C_5 = C and R2=R3=R4=RDR_2 = R_3 = R_4 = R_D, so D=C2RDD = C^2R_D. Taking C=CtC = C_t gives RD=Dnew/Ct2=kzDR_D = D_{new}/C_t^2 = k_z D.

PrototypeAfter BrutonScaledFinal element
Rs=1 ΩR_s = 1\ \OmegaCs=1C_s = 1 F1/(kzω0)1/(k_z\omega_0)10 nF
C1=1C_1 = 1 F (shunt)FDNR D1=1D_1 = 1D1=1/(kzω02)=1×10−12D_1 = 1/(k_z\omega_0^2) = 1 \times 10^{-12}GIC: C = 10 nF, RDR_D = 10 kΩ
L2=2L_2 = 2 H (series)R2=2 ΩR_2 = 2\ \OmegaR2=2×10000R_2 = 2 \times 1000020 kΩ
C3=1C_3 = 1 F (shunt)FDNR D3=1D_3 = 1D3=1/(kzω02)=1×10−12D_3 = 1/(k_z\omega_0^2) = 1 \times 10^{-12}GIC: C = 10 nF, RDR_D = 10 kΩ
RL=1 ΩR_L = 1\ \OmegaCL=1C_L = 1 F1/(kzω0)1/(k_z\omega_0)10 nF

Step 4: DC path. The source and load capacitors block DC, and the op-amps in the GICs need a DC bias path. Put RaR_a across the source capacitor and RbR_b across the load capacitor. They must be much larger than kzk_z (the capacitor impedance at ω0\omega_0). At DC the FDNRs are open, so the DC gain is Rb/(Ra+ΣR+Rb)R_b/(R_a + \Sigma R + R_b). To keep the prototype DC gain of 1/2, set Rb=Ra+ΣRR_b = R_a + \Sigma R:

RaR_a = 1 MΩ, ΣR\Sigma R = 20 kΩ, so RbR_b = 1.02 MΩ.

Final circuit (Cs∥RaC_s \| R_a in series from the source, CL∥RbC_L \| R_b as the load, each DkD_k a GIC FDNR to ground):

Vi o--[Cs||Ra]-+----[R2]-+-------+----o Vo
               |         |       |
              [D1]      [D3]  [CL||Rb]
GND --------------------------------------

Answer: terminating capacitors 10 nF; series resistor 20 kΩ; two FDNRs, each a GIC with C = 10 nF and RDR_D = 10 kΩ; RaR_a = 1 MΩ, RbR_b = 1.02 MΩ.

  • 2075 Chaitra · 6 marks

Design a fourth order Butterworth low pass filter having half power frequency of 4000 rad/s using Leapfrog simulation. [Table: doubly terminated Butterworth element values normalized to 1 rad/s, 1 Ω terminations: n = 2: 1.414, 1.414; n = 3: 1, 2, 1; n = 4: 0.7654, 1.848, 1.848, 0.7654; n = 5: 0.618, 1.618, 2, 1.618, 0.618]

Answer

Leapfrog (operational) simulation realises the doubly terminated LC ladder by simulating its equations rather than its elements: each reactive element is replaced by an op-amp integrator, and the integrators are coupled exactly as the ladder voltages and currents are coupled. The result keeps the low sensitivity of the passive ladder.

From the table (n = 4), the doubly terminated Butterworth prototype (1 Ω terminations, ω0=1\omega_0 = 1 rad/s) is Rs=1R_s = 1, L1=0.7654L_1 = 0.7654 H, C2=1.848C_2 = 1.848 F, L3=1.848L_3 = 1.848 H, C4=0.7654C_4 = 0.7654 F, RL=1 ΩR_L = 1\ \Omega (T-ladder: series L first, shunt C to ground).

Write the state equations of the doubly terminated ladder (normalised, Rs=RL=1 ΩR_s = R_L = 1\ \Omega). Every current is turned into a voltage by multiplying by a 1 Ω resistance, so each equation becomes an integrator:

I1=Vin−V2sL1+1V2=I1−I3sC2I3=V2−V4sL3V4=I3sC4+1Vo=V4⋅RL\begin{aligned} I_1 &= \frac{V_{in} - V_2}{sL_1 + 1} \\ V_2 &= \frac{I_1 - I_3}{sC_2} \\ I_3 &= \frac{V_2 - V_4}{sL_3} \\ V_4 &= \frac{I_3}{sC_4 + 1} \\ V_o &= V_4\cdot R_L \end{aligned}

The first and last equations are lossy integrators (the terminations give the "+1"); the middle ones are lossless integrators. Neighbouring integrators feed each other in alternate directions, which gives the "leapfrog" structure.

Using inverting op-amp integrators, Vo=−1sRC∑ViV_o = -\frac{1}{sRC}\sum V_i (lossless) or Vo=−∑Vi/RsC+1/RV_o = -\frac{\sum V_i/R}{sC + 1/R} (lossy, resistor RR across CC), the signs are arranged as follows. Inverters supply the signals of opposite sign:

Op-ampTypeInputs (each through RR)OutputNormalised CC (R = 1 Ω)
A1lossy integratorVinV_{in}, −V2-V_2−I1-I_10.7654 F
A2integrator−I1-I_1, I3I_3V2V_21.848 F
A3inverterV2V_2−V2-V_2R / R
A4integrator−V2-V_2, V4V_4I3I_31.848 F
A5lossy integratorI3I_3−V4-V_40.7654 F
A6inverter−V4-V_4V4=VoV_4 = V_oR / R

Check of one stage: A2 gives −(−I1)+I3sCA2=I1−I3s(1.848)-\frac{(-I_1) + I_3}{sC_{A2}} = \frac{I_1 - I_3}{s(1.848)}, which is the ladder equation for V2V_2.

Denormalisation. Frequency scale by ω0=4000 rad/s\omega_0 = 4000\ \text{rad/s} and choose all resistors RR = 10 kΩ (impedance scale kz=104k_z = 10^4). Each capacitor becomes C=g/(ω0R)C = g/(\omega_0 R):

CA1=0.76544000×104=19.13 nFCA2=1.8484000×104=46.2 nFCA4=1.8484000×104=46.2 nFCA5=0.76544000×104=19.13 nF\begin{aligned} C_{A1} &= \frac{0.7654}{4000 \times 10^4} = \text{19.13 nF} \\ C_{A2} &= \frac{1.848}{4000 \times 10^4} = \text{46.2 nF} \\ C_{A4} &= \frac{1.848}{4000 \times 10^4} = \text{46.2 nF} \\ C_{A5} &= \frac{0.7654}{4000 \times 10^4} = \text{19.13 nF} \end{aligned}

Final element values:

Op-ampResistorsCapacitor
A1input(s) 10 kΩ, feedback 10 kΩ across C19.13 nF
A2input(s) 10 kΩ46.2 nF
A4input(s) 10 kΩ46.2 nF
A5input(s) 10 kΩ, feedback 10 kΩ across C19.13 nF
inverters10 kΩ in, 10 kΩ feedbacknone
 Lossy integrator (A1, last stage)
           +----[ R ]----+
           +----| C |----+
 Va--[R]--+|             |
 Vb--[R]--++----(-)      |
                (+) A ---+---- Vo
                 |
                GND
 Lossless integrator: same circuit without the
 feedback resistor (only C from output to (-)).

The passband gain is 1/2 (−6 dB), the same as the doubly terminated prototype; the response is that of the passive ladder but with no inductors, and the low sensitivity of the ladder is kept.

Answer: four integrators and two inverters, all resistors 10 kΩ; integrator capacitors 19.13 nF, 46.2 nF, 46.2 nF, 19.13 nF; half-power frequency 4000 rad/s.

  • 2074 Chaitra · 7 marks

Design the fourth order Butterworth low pass filter using leapfrog simulation. In your final design the half power frequency should be 10000 rad/s and practically realizable elements. [Table: doubly terminated Butterworth element values normalized to 1 rad/s, 1 Ω terminations: n = 2: 1.414, 1.414; n = 3: 1, 2, 1; n = 4: 0.7654, 1.848, 1.848, 0.7654; n = 5: 0.618, 1.618, 2, 1.618, 0.618]

Answer

Leapfrog (operational) simulation realises the doubly terminated LC ladder by simulating its equations rather than its elements: each reactive element is replaced by an op-amp integrator, and the integrators are coupled exactly as the ladder voltages and currents are coupled. The result keeps the low sensitivity of the passive ladder.

From the table (n = 4), the doubly terminated Butterworth prototype (1 Ω terminations, ω0=1\omega_0 = 1 rad/s) is Rs=1R_s = 1, L1=0.7654L_1 = 0.7654 H, C2=1.848C_2 = 1.848 F, L3=1.848L_3 = 1.848 H, C4=0.7654C_4 = 0.7654 F, RL=1 ΩR_L = 1\ \Omega (T-ladder: series L first, shunt C to ground).

Write the state equations of the doubly terminated ladder (normalised, Rs=RL=1 ΩR_s = R_L = 1\ \Omega). Every current is turned into a voltage by multiplying by a 1 Ω resistance, so each equation becomes an integrator:

I1=Vin−V2sL1+1V2=I1−I3sC2I3=V2−V4sL3V4=I3sC4+1Vo=V4⋅RL\begin{aligned} I_1 &= \frac{V_{in} - V_2}{sL_1 + 1} \\ V_2 &= \frac{I_1 - I_3}{sC_2} \\ I_3 &= \frac{V_2 - V_4}{sL_3} \\ V_4 &= \frac{I_3}{sC_4 + 1} \\ V_o &= V_4\cdot R_L \end{aligned}

The first and last equations are lossy integrators (the terminations give the "+1"); the middle ones are lossless integrators. Neighbouring integrators feed each other in alternate directions, which gives the "leapfrog" structure.

Using inverting op-amp integrators, Vo=−1sRC∑ViV_o = -\frac{1}{sRC}\sum V_i (lossless) or Vo=−∑Vi/RsC+1/RV_o = -\frac{\sum V_i/R}{sC + 1/R} (lossy, resistor RR across CC), the signs are arranged as follows. Inverters supply the signals of opposite sign:

Op-ampTypeInputs (each through RR)OutputNormalised CC (R = 1 Ω)
A1lossy integratorVinV_{in}, −V2-V_2−I1-I_10.7654 F
A2integrator−I1-I_1, I3I_3V2V_21.848 F
A3inverterV2V_2−V2-V_2R / R
A4integrator−V2-V_2, V4V_4I3I_31.848 F
A5lossy integratorI3I_3−V4-V_40.7654 F
A6inverter−V4-V_4V4=VoV_4 = V_oR / R

Check of one stage: A2 gives −(−I1)+I3sCA2=I1−I3s(1.848)-\frac{(-I_1) + I_3}{sC_{A2}} = \frac{I_1 - I_3}{s(1.848)}, which is the ladder equation for V2V_2.

Denormalisation. Frequency scale by ω0=104 rad/s\omega_0 = 10^4\ \text{rad/s} and choose all resistors RR = 10 kΩ (impedance scale kz=104k_z = 10^4). Each capacitor becomes C=g/(ω0R)C = g/(\omega_0 R):

CA1=0.765410000×104=7.654 nFCA2=1.84810000×104=18.48 nFCA4=1.84810000×104=18.48 nFCA5=0.765410000×104=7.654 nF\begin{aligned} C_{A1} &= \frac{0.7654}{10000 \times 10^4} = \text{7.654 nF} \\ C_{A2} &= \frac{1.848}{10000 \times 10^4} = \text{18.48 nF} \\ C_{A4} &= \frac{1.848}{10000 \times 10^4} = \text{18.48 nF} \\ C_{A5} &= \frac{0.7654}{10000 \times 10^4} = \text{7.654 nF} \end{aligned}

Final element values:

Op-ampResistorsCapacitor
A1input(s) 10 kΩ, feedback 10 kΩ across C7.654 nF
A2input(s) 10 kΩ18.48 nF
A4input(s) 10 kΩ18.48 nF
A5input(s) 10 kΩ, feedback 10 kΩ across C7.654 nF
inverters10 kΩ in, 10 kΩ feedbacknone
 Lossy integrator (A1, last stage)
           +----[ R ]----+
           +----| C |----+
 Va--[R]--+|             |
 Vb--[R]--++----(-)      |
                (+) A ---+---- Vo
                 |
                GND
 Lossless integrator: same circuit without the
 feedback resistor (only C from output to (-)).

The passband gain is 1/2 (−6 dB), the same as the doubly terminated prototype; the response is that of the passive ladder but with no inductors, and the low sensitivity of the ladder is kept.

Answer: four integrators and two inverters, all resistors 10 kΩ; integrator capacitors 7.654 nF, 18.48 nF, 18.48 nF, 7.654 nF; half-power frequency 10,000 rad/s.

  • 2073 Shrawan · 8 marks

The following circuit is a third-order Chebyshev lowpass filter. Simulate it using the leap-frog method. The final design should have ω0 = 4000 rad/s and practically realizable element values. [Figure: source V1, 1 Ω source resistor, series 2.026 H, shunt 0.9941 F, series 2.026 H, 1 Ω load]

Answer

Leapfrog (operational) simulation realises the doubly terminated LC ladder by simulating its equations rather than its elements: each reactive element is replaced by an op-amp integrator, and the integrators are coupled exactly as the ladder voltages and currents are coupled. The result keeps the low sensitivity of the passive ladder.

Given ladder (3rd-order Chebyshev, normalised so the passband edge is 1 rad/s): Rs=1 ΩR_s = 1\ \Omega, series L1=2.026L_1 = 2.026 H, shunt C2=0.9941C_2 = 0.9941 F, series L3=2.026L_3 = 2.026 H, RL=1 ΩR_L = 1\ \Omega.

Write the state equations of the doubly terminated ladder (normalised, Rs=RL=1 ΩR_s = R_L = 1\ \Omega). Every current is turned into a voltage by multiplying by a 1 Ω resistance, so each equation becomes an integrator:

I1=Vin−V2sL1+1V2=I1−I3sC2I3=V2sL3+1Vo=I3⋅RL\begin{aligned} I_1 &= \frac{V_{in} - V_2}{sL_1 + 1} \\ V_2 &= \frac{I_1 - I_3}{sC_2} \\ I_3 &= \frac{V_2}{sL_3 + 1} \\ V_o &= I_3\cdot R_L \end{aligned}

The first and last equations are lossy integrators (the terminations give the "+1"); the middle ones are lossless integrators. Neighbouring integrators feed each other in alternate directions, which gives the "leapfrog" structure.

Using inverting op-amp integrators, Vo=−1sRC∑ViV_o = -\frac{1}{sRC}\sum V_i (lossless) or Vo=−∑Vi/RsC+1/RV_o = -\frac{\sum V_i/R}{sC + 1/R} (lossy, resistor RR across CC), the signs are arranged as follows. Inverters supply the signals of opposite sign:

Op-ampTypeInputs (each through RR)OutputNormalised CC (R = 1 Ω)
A1lossy integratorVinV_{in}, −V2-V_2−I1-I_12.026 F
A2integrator−I1-I_1, I3I_3V2V_20.9941 F
A3inverterV2V_2−V2-V_2R / R
A4lossy integratorV2V_2−I3-I_32.026 F
A5inverter−I3-I_3I3=VoI_3 = V_oR / R

Check of one stage: A2 gives −(−I1)+I3sCA2=I1−I3s(0.9941)-\frac{(-I_1) + I_3}{sC_{A2}} = \frac{I_1 - I_3}{s(0.9941)}, which is the ladder equation for V2V_2.

Denormalisation. Frequency scale by ω0=4000 rad/s\omega_0 = 4000\ \text{rad/s} and choose all resistors RR = 10 kΩ (impedance scale kz=104k_z = 10^4). Each capacitor becomes C=g/(ω0R)C = g/(\omega_0 R):

CA1=2.0264000×104=50.65 nFCA2=0.99414000×104=24.85 nFCA4=2.0264000×104=50.65 nF\begin{aligned} C_{A1} &= \frac{2.026}{4000 \times 10^4} = \text{50.65 nF} \\ C_{A2} &= \frac{0.9941}{4000 \times 10^4} = \text{24.85 nF} \\ C_{A4} &= \frac{2.026}{4000 \times 10^4} = \text{50.65 nF} \end{aligned}

Final element values:

Op-ampResistorsCapacitor
A1input(s) 10 kΩ, feedback 10 kΩ across C50.65 nF
A2input(s) 10 kΩ24.85 nF
A4input(s) 10 kΩ, feedback 10 kΩ across C50.65 nF
inverters10 kΩ in, 10 kΩ feedbacknone
 Lossy integrator (A1, last stage)
           +----[ R ]----+
           +----| C |----+
 Va--[R]--+|             |
 Vb--[R]--++----(-)      |
                (+) A ---+---- Vo
                 |
                GND
 Lossless integrator: same circuit without the
 feedback resistor (only C from output to (-)).

The passband gain is 1/2 (−6 dB), the same as the doubly terminated prototype; the response is that of the passive ladder but with no inductors, and the low sensitivity of the ladder is kept.

Answer: three integrators and two inverters, all resistors 10 kΩ; integrator capacitors 50.65 nF, 24.85 nF, 50.65 nF; passband edge 4000 rad/s.

  • 2072 Kartik · 6 marks

Design third order Butterworth low pass filter using Leapfrog simulation. Your final design should have half power frequency of 4KHz and practically realizable elements. [Table: doubly terminated Butterworth element values normalized to 1 rad/s, 1 Ω terminations: n = 2: 1.414, 1.414; n = 3: 1, 2, 1; n = 4: 0.7654, 1.848, 1.848, 0.7654; n = 5: 0.618, 1.618, 2, 1.618, 0.618]

Answer

Leapfrog (operational) simulation realises the doubly terminated LC ladder by simulating its equations rather than its elements: each reactive element is replaced by an op-amp integrator, and the integrators are coupled exactly as the ladder voltages and currents are coupled. The result keeps the low sensitivity of the passive ladder.

From the table (n = 3): Rs=1 ΩR_s = 1\ \Omega, series L1=1L_1 = 1 H, shunt C2=2C_2 = 2 F, series L3=1L_3 = 1 H, RL=1 ΩR_L = 1\ \Omega.

Write the state equations of the doubly terminated ladder (normalised, Rs=RL=1 ΩR_s = R_L = 1\ \Omega). Every current is turned into a voltage by multiplying by a 1 Ω resistance, so each equation becomes an integrator:

I1=Vin−V2sL1+1V2=I1−I3sC2I3=V2sL3+1Vo=I3⋅RL\begin{aligned} I_1 &= \frac{V_{in} - V_2}{sL_1 + 1} \\ V_2 &= \frac{I_1 - I_3}{sC_2} \\ I_3 &= \frac{V_2}{sL_3 + 1} \\ V_o &= I_3\cdot R_L \end{aligned}

The first and last equations are lossy integrators (the terminations give the "+1"); the middle ones are lossless integrators. Neighbouring integrators feed each other in alternate directions, which gives the "leapfrog" structure.

Using inverting op-amp integrators, Vo=−1sRC∑ViV_o = -\frac{1}{sRC}\sum V_i (lossless) or Vo=−∑Vi/RsC+1/RV_o = -\frac{\sum V_i/R}{sC + 1/R} (lossy, resistor RR across CC), the signs are arranged as follows. Inverters supply the signals of opposite sign:

Op-ampTypeInputs (each through RR)OutputNormalised CC (R = 1 Ω)
A1lossy integratorVinV_{in}, −V2-V_2−I1-I_11 F
A2integrator−I1-I_1, I3I_3V2V_22 F
A3inverterV2V_2−V2-V_2R / R
A4lossy integratorV2V_2−I3-I_31 F
A5inverter−I3-I_3I3=VoI_3 = V_oR / R

Check of one stage: A2 gives −(−I1)+I3sCA2=I1−I3s(2)-\frac{(-I_1) + I_3}{sC_{A2}} = \frac{I_1 - I_3}{s(2)}, which is the ladder equation for V2V_2.

Denormalisation. Frequency scale by ω0=2π(4000)=25,133 rad/s\omega_0 = 2\pi(4000) = 25{,}133\ \text{rad/s} and choose all resistors RR = 10 kΩ (impedance scale kz=104k_z = 10^4). Each capacitor becomes C=g/(ω0R)C = g/(\omega_0 R):

CA1=125132.7×104=3.979 nFCA2=225132.7×104=7.958 nFCA4=125132.7×104=3.979 nF\begin{aligned} C_{A1} &= \frac{1}{25132.7 \times 10^4} = \text{3.979 nF} \\ C_{A2} &= \frac{2}{25132.7 \times 10^4} = \text{7.958 nF} \\ C_{A4} &= \frac{1}{25132.7 \times 10^4} = \text{3.979 nF} \end{aligned}

Final element values:

Op-ampResistorsCapacitor
A1input(s) 10 kΩ, feedback 10 kΩ across C3.979 nF
A2input(s) 10 kΩ7.958 nF
A4input(s) 10 kΩ, feedback 10 kΩ across C3.979 nF
inverters10 kΩ in, 10 kΩ feedbacknone
 Lossy integrator (A1, last stage)
           +----[ R ]----+
           +----| C |----+
 Va--[R]--+|             |
 Vb--[R]--++----(-)      |
                (+) A ---+---- Vo
                 |
                GND
 Lossless integrator: same circuit without the
 feedback resistor (only C from output to (-)).

The passband gain is 1/2 (−6 dB), the same as the doubly terminated prototype; the response is that of the passive ladder but with no inductors, and the low sensitivity of the ladder is kept.

Answer: three integrators and two inverters, all resistors 10 kΩ; integrator capacitors 3.979 nF, 7.958 nF, 3.979 nF; half-power frequency 4 kHz.

  • 2071 Chaitra · 6 marks

Using leapfrog method simulate the LC ladder circuit given below to obtain a low pass filter having passband of 6KHz and suitable element values. [Figure: source V1, 1 Ω source resistor, shunt 2.0237 F, series 0.9941 H, shunt 2.0237 F, 1 Ω load (output V2)]

Answer

Leapfrog (operational) simulation realises the doubly terminated LC ladder by simulating its equations rather than its elements: each reactive element is replaced by an op-amp integrator, and the integrators are coupled exactly as the ladder voltages and currents are coupled. The result keeps the low sensitivity of the passive ladder.

Given ladder (π form, a 1 dB-ripple Chebyshev prototype with passband edge 1 rad/s): Rs=1 ΩR_s = 1\ \Omega, shunt C1=2.0237C_1 = 2.0237 F, series L2=0.9941L_2 = 0.9941 H, shunt C3=2.0237C_3 = 2.0237 F, RL=1 ΩR_L = 1\ \Omega. Here the state variables are the capacitor voltages V1V_1, V3V_3 and the inductor current I2I_2 (the source is converted to a current Vin/RsV_{in}/R_s).

Write the state equations of the doubly terminated ladder (normalised, Rs=RL=1 ΩR_s = R_L = 1\ \Omega). Every current is turned into a voltage by multiplying by a 1 Ω resistance, so each equation becomes an integrator:

V1=Vin−I2sC1+1I2=V1−V3sL2V3=I2sC3+1Vo=V3\begin{aligned} V_1 &= \frac{V_{in} - I_2}{sC_1 + 1} \\ I_2 &= \frac{V_1 - V_3}{sL_2} \\ V_3 &= \frac{I_2}{sC_3 + 1} \\ V_o &= V_3 \end{aligned}

The first and last equations are lossy integrators (the terminations give the "+1"); the middle ones are lossless integrators. Neighbouring integrators feed each other in alternate directions, which gives the "leapfrog" structure.

Using inverting op-amp integrators, Vo=−1sRC∑ViV_o = -\frac{1}{sRC}\sum V_i (lossless) or Vo=−∑Vi/RsC+1/RV_o = -\frac{\sum V_i/R}{sC + 1/R} (lossy, resistor RR across CC), the signs are arranged as follows. Inverters supply the signals of opposite sign:

Op-ampTypeInputs (each through RR)OutputNormalised CC (R = 1 Ω)
A1lossy integratorVinV_{in}, −I2-I_2−V1-V_12.0237 F
A2integrator−V1-V_1, V3V_3I2I_20.9941 F
A3inverterI2I_2−I2-I_2R / R
A4lossy integratorI2I_2−V3-V_32.0237 F
A5inverter−V3-V_3V3=VoV_3 = V_oR / R

Check of one stage: A2 gives −(−V1)+V3sCA2=V1−V3s(0.9941)-\frac{(-V_1) + V_3}{sC_{A2}} = \frac{V_1 - V_3}{s(0.9941)}, which is the ladder equation for I2I_2.

Denormalisation. Frequency scale by ω0=2π(6000)=37,699 rad/s\omega_0 = 2\pi(6000) = 37{,}699\ \text{rad/s} and choose all resistors RR = 10 kΩ (impedance scale kz=104k_z = 10^4). Each capacitor becomes C=g/(ω0R)C = g/(\omega_0 R):

CA1=2.023737699.1×104=5.368 nFCA2=0.994137699.1×104=2.637 nFCA4=2.023737699.1×104=5.368 nF\begin{aligned} C_{A1} &= \frac{2.0237}{37699.1 \times 10^4} = \text{5.368 nF} \\ C_{A2} &= \frac{0.9941}{37699.1 \times 10^4} = \text{2.637 nF} \\ C_{A4} &= \frac{2.0237}{37699.1 \times 10^4} = \text{5.368 nF} \end{aligned}

Final element values:

Op-ampResistorsCapacitor
A1input(s) 10 kΩ, feedback 10 kΩ across C5.368 nF
A2input(s) 10 kΩ2.637 nF
A4input(s) 10 kΩ, feedback 10 kΩ across C5.368 nF
inverters10 kΩ in, 10 kΩ feedbacknone
 Lossy integrator (A1, last stage)
           +----[ R ]----+
           +----| C |----+
 Va--[R]--+|             |
 Vb--[R]--++----(-)      |
                (+) A ---+---- Vo
                 |
                GND
 Lossless integrator: same circuit without the
 feedback resistor (only C from output to (-)).

The passband gain is 1/2 (−6 dB), the same as the doubly terminated prototype; the response is that of the passive ladder but with no inductors, and the low sensitivity of the ladder is kept.

Answer: three integrators and two inverters, all resistors 10 kΩ; integrator capacitors 5.368 nF, 2.637 nF, 5.368 nF; passband edge 6 kHz.

  • 2070 Chaitra · 6 marks

Simulate third order Butterworth low pass filter using Leapfrog simulation. [Table: doubly terminated Butterworth element values normalized to 1 rad/s, 1 Ω terminations: n = 2: 1.414, 1.414; n = 3: 1, 2, 1; n = 4: 0.7654, 1.848, 1.848, 0.7654; n = 5: 0.618, 1.618, 2, 1.618, 0.618]

Answer

Leapfrog (operational) simulation realises the doubly terminated LC ladder by simulating its equations rather than its elements: each reactive element is replaced by an op-amp integrator, and the integrators are coupled exactly as the ladder voltages and currents are coupled. The result keeps the low sensitivity of the passive ladder.

From the table (n = 3): Rs=1 ΩR_s = 1\ \Omega, series L1=1L_1 = 1 H, shunt C2=2C_2 = 2 F, series L3=1L_3 = 1 H, RL=1 ΩR_L = 1\ \Omega.

 Vin  Rs=1   L1=1H        L3=1H
 o---[ R ]--(LLL)--+--(LLL)--+----o Vo
                   |         |
                 C2=2F     [RL=1]
                   |         |
 GND --------------+---------+

Write the state equations of the doubly terminated ladder (normalised, Rs=RL=1 ΩR_s = R_L = 1\ \Omega). Every current is turned into a voltage by multiplying by a 1 Ω resistance, so each equation becomes an integrator:

I1=Vin−V2sL1+1V2=I1−I3sC2I3=V2sL3+1Vo=I3⋅RL\begin{aligned} I_1 &= \frac{V_{in} - V_2}{sL_1 + 1} \\ V_2 &= \frac{I_1 - I_3}{sC_2} \\ I_3 &= \frac{V_2}{sL_3 + 1} \\ V_o &= I_3\cdot R_L \end{aligned}

The first and last equations are lossy integrators (the terminations give the "+1"); the middle ones are lossless integrators. Neighbouring integrators feed each other in alternate directions, which gives the "leapfrog" structure.

Using inverting op-amp integrators, Vo=−1sRC∑ViV_o = -\frac{1}{sRC}\sum V_i (lossless) or Vo=−∑Vi/RsC+1/RV_o = -\frac{\sum V_i/R}{sC + 1/R} (lossy, resistor RR across CC), the signs are arranged as follows. Inverters supply the signals of opposite sign:

Op-ampTypeInputs (each through RR)OutputNormalised CC (R = 1 Ω)
A1lossy integratorVinV_{in}, −V2-V_2−I1-I_11 F
A2integrator−I1-I_1, I3I_3V2V_22 F
A3inverterV2V_2−V2-V_2R / R
A4lossy integratorV2V_2−I3-I_31 F
A5inverter−I3-I_3I3=VoI_3 = V_oR / R

Check of one stage: A2 gives −(−I1)+I3sCA2=I1−I3s(2)-\frac{(-I_1) + I_3}{sC_{A2}} = \frac{I_1 - I_3}{s(2)}, which is the ladder equation for V2V_2.

Denormalisation. Frequency scale by ω0=104 rad/s (example, since no frequency is given)\omega_0 = 10^4\ \text{rad/s (example, since no frequency is given)} and choose all resistors RR = 10 kΩ (impedance scale kz=104k_z = 10^4). Each capacitor becomes C=g/(ω0R)C = g/(\omega_0 R):

CA1=110000×104=10 nFCA2=210000×104=20 nFCA4=110000×104=10 nF\begin{aligned} C_{A1} &= \frac{1}{10000 \times 10^4} = \text{10 nF} \\ C_{A2} &= \frac{2}{10000 \times 10^4} = \text{20 nF} \\ C_{A4} &= \frac{1}{10000 \times 10^4} = \text{10 nF} \end{aligned}

Final element values:

Op-ampResistorsCapacitor
A1input(s) 10 kΩ, feedback 10 kΩ across C10 nF
A2input(s) 10 kΩ20 nF
A4input(s) 10 kΩ, feedback 10 kΩ across C10 nF
inverters10 kΩ in, 10 kΩ feedbacknone
 Lossy integrator (A1, last stage)
           +----[ R ]----+
           +----| C |----+
 Va--[R]--+|             |
 Vb--[R]--++----(-)      |
                (+) A ---+---- Vo
                 |
                GND
 Lossless integrator: same circuit without the
 feedback resistor (only C from output to (-)).

The passband gain is 1/2 (−6 dB), the same as the doubly terminated prototype; the response is that of the passive ladder but with no inductors, and the low sensitivity of the ladder is kept.

The normalised simulation (ω0\omega_0 = 1 rad/s) has all resistors 1 Ω and capacitors 1 F, 2 F, 1 F.

Answer (example, ω0\omega_0 = 10⁴ rad/s): all resistors 10 kΩ; integrator capacitors 10 nF, 20 nF, 10 nF.

  • 2082 Bhadra · 8 marks

Design a 5th order Butterworth LPF using cascade of 1st order active section and MFB biquads with DC gain equal to unity and half power frequency at 1 kHz. Make the largest capacitance 0.1 μF.

Answer

Butterworth poles (n = 5) lie on the unit circle at 0°, ±36° and ±72° from the negative real axis, so

TN(s)=1(s+1)(s2+0.618s+1)(s2+1.618s+1)T_N(s) = \frac{1}{(s+1)(s^2 + 0.618s + 1)(s^2 + 1.618s + 1)}

Each quadratic s2+(1/Q)s+1s^2 + (1/Q)s + 1 gives Q=1/0.618=1.618Q = 1/0.618 = 1.618 and Q=1/1.618=0.618Q = 1/1.618 = 0.618. The filter is one first-order section followed by two MFB biquads; ω0=2π(1000)=6283.2\omega_0 = 2\pi(1000) = 6283.2 rad/s.

First-order section

An inverting lossy integrator: input R1R_1, feedback Rf∥CR_f \parallel C:

T1(s)=−1/(R1C)s+1/(RfC)T_1(s) = -\frac{1/(R_1C)}{s + 1/(R_fC)}

Normalised: R1=Rf=1 ΩR_1 = R_f = 1\ \Omega, C=1C = 1 F (pole at 1, gain 1). Scaling with CC = 0.1 μF: R=1/(ω0C)=1/(6283.2×10−7)R = 1/(\omega_0 C) = 1/(6283.2 \times 10^{-7}) = 1.592 kΩ for both R1R_1 and RfR_f.

          +---[Rf]---+
          +---| C |--+
 Vi-[R1]--+--(-)     |
              A -----+--o
     GND --(+)

MFB biquads

              +----------[R2]-----------+
              |                         |
              |       +----[C2]----+    |
              |       |            |    |
 Vi o--[R1]---+--[R3]-+----(-)     |    |
              |              A ----+----+--o Vo
             [C1]     GND---(+)
              |
             GND

For this multiple-feedback (MFB) lowpass biquad:

T(s)=VoVi=−1R1R3C1C2s2+sC1(1R1+1R2+1R3)+1R2R3C1C2,T(0)=−R2R1T(s) = \frac{V_o}{V_i} = \frac{-\dfrac{1}{R_1R_3C_1C_2}}{s^2 + \dfrac{s}{C_1}\left(\dfrac{1}{R_1} + \dfrac{1}{R_2} + \dfrac{1}{R_3}\right) + \dfrac{1}{R_2R_3C_1C_2}}, \qquad T(0) = -\frac{R_2}{R_1}

Normalised design (ω0=1\omega_0 = 1, unity DC gain): take R1=R2=R3=1 ΩR_1 = R_2 = R_3 = 1\ \Omega. Then ω02=1/(C1C2)=1\omega_0^2 = 1/(C_1C_2) = 1 and ω0/Q=3/C1\omega_0/Q = 3/C_1, so

C1=3Q,C2=13QC_1 = 3Q, \qquad C_2 = \frac{1}{3Q}

Scaling: ω0\omega_0 = 6283.2 rad/s. For each section choose kzk_z so that its largest capacitor C1C_1 becomes 0.1 μF: kz=C1n/(ω0×10−7)k_z = C_{1n}/(\omega_0 \times 10^{-7}). Then all three resistors are kzk_z and C2=C2n/(kzω0)C_2 = C_{2n}/(k_z\omega_0). (Each section is driven by an op-amp output, so sections may be scaled separately.)

SectionQC1nC_{1n}C2nC_{2n}R1=R2=R3R_1 = R_2 = R_3C1C_1C2C_2
s2+0.618s+1s^2 + 0.618s + 11.6184.8541 F0.2060 F7.726 kΩ100 nF4.244 nF
s2+1.618s+1s^2 + 1.618s + 10.6181.8541 F0.5393 F2.951 kΩ100 nF29.09 nF

Final circuit

 Vi -->[1st order]-->[MFB, Q=1.618]-->[MFB, Q=0.618]--> Vo
       R=1.592k        R=7.726k          R=2.951k
       C=0.1uF         C1=0.1uF          C1=0.1uF
                       C2=4.244nF        C2=29.09nF

Each stage has DC gain −1, so the overall DC gain magnitude is 1 (sign −1; add a unity inverter if a positive output is needed). Half-power frequency 1 kHz.

Answer: first-order: R = 1.592 kΩ, C = 0.1 μF; MFB 1 (Q = 1.618): R = 7.726 kΩ, C1 = 0.1 μF, C2 = 4.244 nF; MFB 2 (Q = 0.618): R = 2.951 kΩ, C1 = 0.1 μF, C2 = 29.09 nF.

  • 2081 Bhadra · 8 marks

Design a 4th order Butterworth LPF using cascade two MFB biquads with dc gain equal to unity and half power frequency at 1000rad/sec. Make the largest capacitance equal to 0.1 μF in your final circuit.

Answer

Butterworth poles (n = 4) lie at ±22.5° and ±67.5° from the negative real axis on the unit circle:

TN(s)=1(s2+1.8478s+1)(s2+0.7654s+1)T_N(s) = \frac{1}{(s^2 + 1.8478s + 1)(s^2 + 0.7654s + 1)}

so the two biquads have Q1=1/1.8478=0.5412Q_1 = 1/1.8478 = 0.5412 and Q2=1/0.7654=1.3066Q_2 = 1/0.7654 = 1.3066, with ω0\omega_0 = 1000 rad/s.

MFB lowpass biquad

              +----------[R2]-----------+
              |                         |
              |       +----[C2]----+    |
              |       |            |    |
 Vi o--[R1]---+--[R3]-+----(-)     |    |
              |              A ----+----+--o Vo
             [C1]     GND---(+)
              |
             GND

For this multiple-feedback (MFB) lowpass biquad:

T(s)=VoVi=−1R1R3C1C2s2+sC1(1R1+1R2+1R3)+1R2R3C1C2,T(0)=−R2R1T(s) = \frac{V_o}{V_i} = \frac{-\dfrac{1}{R_1R_3C_1C_2}}{s^2 + \dfrac{s}{C_1}\left(\dfrac{1}{R_1} + \dfrac{1}{R_2} + \dfrac{1}{R_3}\right) + \dfrac{1}{R_2R_3C_1C_2}}, \qquad T(0) = -\frac{R_2}{R_1}

Normalised design (ω0=1\omega_0 = 1, unity DC gain): take R1=R2=R3=1 ΩR_1 = R_2 = R_3 = 1\ \Omega. Then ω02=1/(C1C2)=1\omega_0^2 = 1/(C_1C_2) = 1 and ω0/Q=3/C1\omega_0/Q = 3/C_1, so

C1=3Q,C2=13QC_1 = 3Q, \qquad C_2 = \frac{1}{3Q}

Scaling

ω0\omega_0 = 1000 rad/s. In each section set C1C_1 (the larger capacitor) to 0.1 μF: kz=C1n/(ω0×10−7)k_z = C_{1n}/(\omega_0 \times 10^{-7}); then R1=R2=R3=kzR_1 = R_2 = R_3 = k_z and C2=C2n/(kzω0)C_2 = C_{2n}/(k_z\omega_0).

SectionQC1nC_{1n}C2nC_{2n}R1=R2=R3R_1 = R_2 = R_3C1C_1C2C_2
s2+1.8478s+1s^2 + 1.8478s + 10.54121.6236 F0.6159 F16.24 kΩ100 nF37.94 nF
s2+0.7654s+1s^2 + 0.7654s + 11.30663.9197 F0.2551 F39.2 kΩ100 nF6.509 nF

Check for section 2: ω0=1/R2C1C2=1/(39.2k)2(10−7)(6.509×10−9)≈1000\omega_0 = 1/\sqrt{R^2C_1C_2} = 1/\sqrt{(39.2\text{k})^2(10^{-7})(6.509 \times 10^{-9})} \approx 1000 rad/s.

Final circuit

 Vi -->[MFB 1, Q=0.5412]-->[MFB 2, Q=1.3066]--> Vo
         R = 16.24k             R = 39.2k
         C1 = 0.1uF             C1 = 0.1uF
         C2 = 37.94nF           C2 = 6.509nF

Each section has DC gain −1, so the cascade has DC gain +1 (unity) and half-power frequency 1000 rad/s.

Answer: section 1: R1=R2=R3R_1 = R_2 = R_3 = 16.24 kΩ, C1C_1 = 0.1 μF, C2C_2 = 37.94 nF; section 2: R1=R2=R3R_1 = R_2 = R_3 = 39.2 kΩ, C1C_1 = 0.1 μF, C2C_2 = 6.509 nF.

  • 2079 Bhadra · 7 marks

Design a 4th order Butterworth LPF using cascaded two Sallen-key biquad having half power frequency of 1 kHz and largest capacitor of 0.1 μF.

Answer

Butterworth n = 4 (pole table): p=−0.3827±j0.9239p = -0.3827 \pm j0.9239 and −0.9239±j0.3827-0.9239 \pm j0.3827, so

TN(s)=1(s2+0.7654s+1)(s2+1.8478s+1),Q1=11.8478=0.5412,Q2=10.7654=1.3066T_N(s) = \frac{1}{(s^2 + 0.7654s + 1)(s^2 + 1.8478s + 1)}, \qquad Q_1 = \frac{1}{1.8478} = 0.5412,\quad Q_2 = \frac{1}{0.7654} = 1.3066

Sallen–Key biquad

                 +--------[C1]---------+
                 |                     |
 Vi o--[R1]------+--[R2]--+----(+)     |
                          |       A ---+---o Vo
                         [C2]   +-(-)  |
                          |     |      |
                         GND    +------+

Unity-gain Sallen–Key lowpass biquad (C1C_1 from the middle node to the output, C2C_2 to ground):

T(s)=1R1R2C1C2s2+sC1(1R1+1R2)+1R1R2C1C2T(s) = \frac{\dfrac{1}{R_1R_2C_1C_2}}{s^2 + \dfrac{s}{C_1}\left(\dfrac{1}{R_1} + \dfrac{1}{R_2}\right) + \dfrac{1}{R_1R_2C_1C_2}}

Normalised equal-resistor design (R1=R2=1R_1 = R_2 = 1, ω0=1\omega_0 = 1): C1C2=1C_1C_2 = 1 and ω0/Q=2/C1\omega_0/Q = 2/C_1, so

C1=2Q,C2=12QC_1 = 2Q, \qquad C_2 = \frac{1}{2Q}

Scaling

ω0=2π(1000)=6283.2\omega_0 = 2\pi(1000) = 6283.2 rad/s. In each section the larger capacitor C1C_1 is made 0.1 μF: kz=C1n/(ω0×10−7)k_z = C_{1n}/(\omega_0 \times 10^{-7}), R1=R2=kzR_1 = R_2 = k_z, C2=C2n/(kzω0)C_2 = C_{2n}/(k_z\omega_0).

SectionQC1nC_{1n}C2nC_{2n}R1=R2R_1 = R_2C1C_1C2C_2
s2+1.8478s+1s^2 + 1.8478s + 10.54121.0824 F0.9239 F1.723 kΩ100 nF85.36 nF
s2+0.7654s+1s^2 + 0.7654s + 11.30662.6131 F0.3827 F4.159 kΩ100 nF14.64 nF
 Vi -->[SK 1, Q=0.5412]-->[SK 2, Q=1.3066]--> Vo
        R1=R2=1.723k        R1=R2=4.159k
        C1=0.1uF            C1=0.1uF
        C2=85.36nF          C2=14.64nF

Both sections are unity-gain followers, so the DC gain is 1 and no gain compensation is needed. Half-power frequency 1 kHz.

Answer: section 1: R1=R2R_1 = R_2 = 1.723 kΩ, C1C_1 = 0.1 μF, C2C_2 = 85.36 nF; section 2: R1=R2R_1 = R_2 = 4.159 kΩ, C1C_1 = 0.1 μF, C2C_2 = 14.64 nF.

  • 2079 Baisakh · 8 marks

Design a 4th order Butterworth LPF using cascade of two Sallen-key biquads having half power frequency of 10 kHz, using 0.1 μF capacitors. Perform gain compensation if necessary.

Answer

Butterworth n = 4 (pole table): p=−0.3827±j0.9239p = -0.3827 \pm j0.9239 and −0.9239±j0.3827-0.9239 \pm j0.3827, so

TN(s)=1(s2+0.7654s+1)(s2+1.8478s+1),Q1=11.8478=0.5412,Q2=10.7654=1.3066T_N(s) = \frac{1}{(s^2 + 0.7654s + 1)(s^2 + 1.8478s + 1)}, \qquad Q_1 = \frac{1}{1.8478} = 0.5412,\quad Q_2 = \frac{1}{0.7654} = 1.3066

Equal-component Sallen–Key biquad

                 +--------[C]----------+
                 |                     |
 Vi o--[R]-------+--[R]---+----(+)     |
                          |       A ---+---o Vo
                         [C]  +--(-)   |
                          |   |        |
                         GND  +--[RB]--+
                              |
                             [RA]
                              |
                             GND

Equal-component Sallen–Key lowpass (R1=R2=RR_1 = R_2 = R, C1=C2=CC_1 = C_2 = C, amplifier gain K=1+RB/RAK = 1 + R_B/R_A):

T(s)=Kω02s2+ω0Qs+ω02,ω0=1RC,Q=13−K  ⇒  K=3−1QT(s) = \frac{K\omega_0^2}{s^2 + \dfrac{\omega_0}{Q}s + \omega_0^2}, \qquad \omega_0 = \frac{1}{RC}, \qquad Q = \frac{1}{3 - K} \;\Rightarrow\; K = 3 - \frac{1}{Q}

Element values

ω0=2π(104)=62,832 rad/s\omega_0 = 2\pi(10^4) = 62{,}832\ \text{rad/s}, CC = 0.1 μF (all four capacitors):

R=1ω0C=159.2 ΩR = \frac{1}{\omega_0 C} = 159.2\ \Omega
SectionQK=3−1/QK = 3 - 1/QRAR_ARB=(K−1)RAR_B = (K-1)R_A
10.54121.152210 kΩ1.522 kΩ
21.30662.234610 kΩ12.35 kΩ

Overall passband gain K1K2K_1K_2 = 2.5748 (8.21 dB).

Gain compensation

To bring the DC gain back to 1, replace the input resistor RR of each section by a divider R1aR_{1a} (from ViV_i) and R1bR_{1b} (to ground) that attenuates by 1/K1/K while keeping the Thevenin resistance equal to RR:

R1a=KR,R1b=KRK−1R_{1a} = KR, \qquad R_{1b} = \frac{KR}{K-1}

(Check: R1a∥R1b=RR_{1a} \parallel R_{1b} = R and R1b/(R1a+R1b)=1/KR_{1b}/(R_{1a}+R_{1b}) = 1/K.)

SectionR1aR_{1a}R1bR_{1b}
1183.4 Ω1.205 kΩ
2355.7 Ω288.1 Ω
 Vi -->[SK 1: K=1.1522]-->[SK 2: K=2.2346]--> Vo
       R=159.2, C=0.1uF     R=159.2, C=0.1uF
       RA=10k, RB=1.522k    RA=10k, RB=12.35k

With the forced 0.1 μF capacitors at 10 kHz the resistors are small (159.2 Ω); a good op-amp with low output resistance is needed.

Answer: all C = 0.1 μF, R = 159.2 Ω; section 1: K = 1.1522 (RBR_B = 1.522 kΩ, RAR_A = 10 kΩ); section 2: K = 2.2346 (RBR_B = 12.35 kΩ); uncompensated gain 2.5748; for unity gain use input dividers 183.4 Ω / 1.205 kΩ (section 1) and 355.7 Ω / 288.1 Ω (section 2).

  • 2072 Chaitra · 8 marks

Design Sallen key lowpass filter for fourth order Butterworth filter. The final circuit should have ω0 = 10,000 rad/s and practically realizable elements. (Refer table 1 of Butterworth pole locations)

Answer

Butterworth n = 4 (pole table): p=−0.3827±j0.9239p = -0.3827 \pm j0.9239 and −0.9239±j0.3827-0.9239 \pm j0.3827, so

TN(s)=1(s2+0.7654s+1)(s2+1.8478s+1),Q1=11.8478=0.5412,Q2=10.7654=1.3066T_N(s) = \frac{1}{(s^2 + 0.7654s + 1)(s^2 + 1.8478s + 1)}, \qquad Q_1 = \frac{1}{1.8478} = 0.5412,\quad Q_2 = \frac{1}{0.7654} = 1.3066

Equal-component Sallen–Key biquad

                 +--------[C]----------+
                 |                     |
 Vi o--[R]-------+--[R]---+----(+)     |
                          |       A ---+---o Vo
                         [C]  +--(-)   |
                          |   |        |
                         GND  +--[RB]--+
                              |
                             [RA]
                              |
                             GND

Equal-component Sallen–Key lowpass (R1=R2=RR_1 = R_2 = R, C1=C2=CC_1 = C_2 = C, amplifier gain K=1+RB/RAK = 1 + R_B/R_A):

T(s)=Kω02s2+ω0Qs+ω02,ω0=1RC,Q=13−K  ⇒  K=3−1QT(s) = \frac{K\omega_0^2}{s^2 + \dfrac{\omega_0}{Q}s + \omega_0^2}, \qquad \omega_0 = \frac{1}{RC}, \qquad Q = \frac{1}{3 - K} \;\Rightarrow\; K = 3 - \frac{1}{Q}

Element values

ω0=10,000 rad/s\omega_0 = 10{,}000\ \text{rad/s}, CC = 0.01 μF (chosen):

R=1ω0C=10 kΩR = \frac{1}{\omega_0 C} = 10\ \text{k}\Omega
SectionQK=3−1/QK = 3 - 1/QRAR_ARB=(K−1)RAR_B = (K-1)R_A
10.54121.152210 kΩ1.522 kΩ
21.30662.234610 kΩ12.35 kΩ

Overall passband gain K1K2K_1K_2 = 2.5748 (8.21 dB).

 Vi -->[SK 1: K=1.1522]-->[SK 2: K=2.2346]--> Vo
       R=10k, C=0.01uF      R=10k, C=0.01uF
       RA=10k, RB=1.522k    RA=10k, RB=12.35k

If unity DC gain is required, the input resistor of each section is split into a divider R1a=KRR_{1a} = KR, R1b=KR/(K−1)R_{1b} = KR/(K-1): section 1: 11.52 kΩ and 75.69 kΩ; section 2: 22.35 kΩ and 18.1 kΩ.

Answer: all R = 10 kΩ and C = 0.01 μF; section 1 (Q = 0.5412): K = 1.1522, RAR_A = 10 kΩ, RBR_B = 1.522 kΩ; section 2 (Q = 1.3066): K = 2.2346, RAR_A = 10 kΩ, RBR_B = 12.35 kΩ; ω0\omega_0 = 10,000 rad/s, passband gain 2.5748 (8.21 dB).

  • 2070 Asar · 4 marks

What are the different techniques of designing higher order active filters? Discuss briefly.

Answer

Higher-order active filters (order > 2) are designed by one of these methods:

1. Cascade design

The transfer function is factored into second-order sections (plus one first-order section for odd order), each realised by a biquad (Sallen–Key, MFB, Tow–Thomas, KHN) and connected in cascade. Op-amp outputs have low impedance, so sections do not load each other.

  • Simple to design and tune; each pole pair is set separately.
  • Sensitivity is higher, especially for high-Q sections.
  • Order of sections: usually low Q first, high Q last.

2. Direct (element) simulation of the LC ladder

Start from the doubly terminated LC ladder, which has the lowest sensitivity, and replace the inductors:

  • Gyrator / GIC inductor simulation: grounded inductors (highpass ladders) by GICs, floating ones by two GICs or gyrators.
  • FDNR (Bruton transformation): divide impedances by ss; inductors become resistors and capacitors become grounded FDNRs (lowpass ladders).

3. Operational simulation (leapfrog)

The state equations of the ladder (currents and voltages of L and C) are realised with integrators and summers that feed each other alternately. It keeps the ladder's low sensitivity and suits lowpass and bandpass filters.

4. Multiple-loop feedback (MLF)

Integrator or biquad blocks with several feedback paths, e.g. follow-the-leader feedback (FLF) and primary-resonator block (PRB); the leapfrog is also a multiple-loop structure.

MethodSensitivityEase of design/tuning
Cascadehighereasiest
Ladder simulation (GIC, FDNR)lowmoderate
Leapfrog / MLFlowmore complex

Questions from Old Question Collection (BEI EX 606 and BEX EX 704) (Scanned IOE papers: BEI EX 606 2078–2083 and BEX EX 704 2069–2076), Old Question Collection (EX 704) (IOE BEX EX 704 papers from 2069 to 2081) and 2080 course paper (ENEX 301) (IOE ENEX 301 new-course paper, 2082 Chaitra). Answers are written for this site; check them against your class notes.

Chapter titles and hours from the IOE syllabus ↗