TRANSISTOR-LEVEL REPRODUCTION AND BIAS ANALYSIS OF A SINGLE-OTRA MULTIFUNCTIONAL BIQUAD FILTER

Abstract

The Operational Transresistance Amplifier (OTRA) is a current-mode active build ing block whose differential current inputs present a virtual ground and whose out put voltage is the difference of the input currents scaled by a transresistance gain. Because the external resistors and capacitors connect directly to low-impedance virtual-ground nodes, OTRA-based filters are free of the parasitic-input-capacitance limitations that constrain voltage-mode op-amp realisations. Mann, Kumar, and Bhaskar (ICSC 2021) proposed a single-OTRA multifunctional biquad that realises low-pass, band-pass, and high-pass responses from one active element and a canonic set of passive components. This dissertation reports a full reproduction of that biquad in three tiers: an ideal symbolic (MATLAB) model, a behavioural single-pole OTRA macromodel, and a transistor-level CMOS realisation of the Mostafa–Soliman fourteen-MOSFET OTRA ina0.18 mprocesssimulatedinngspice. The symbolic and behavioural tiers reproduce the published responses to within a fraction of a decibel. The transistor level low-pass and band-pass responses also match, but the high-pass response was initially observed to droop, with a passband plateau of −3.56dB against the target 0dB and a −3dB frequency of 1.76MHz against the target 4.25MHz. A systematic investigation traced this droop to the bias voltage VB1, which sets the tail currents of the voltage-biased transistors and therefore the signal-path transcon ductance and the transresistance gain-bandwidth. A focused VB1 sweep showed a narrow functional window: at VB1 = −1.31volt the high-pass passband recovers to approximately 0dB with a −3dB frequency of 4.17MHz, matching the published design, while the low-pass and band-pass responses remain matched at the same global bias. An attempted regulated-cascode fix at the high-impedance internal node is shown to be both unbiasable on a signal-referenced node and unnecessary, since output-resistance boosting does not change the controlling gain-bandwidth product. The principal contribution is the identification of VB1, rather than device sizing, output resistance, or the model card, as the operative lever that recovers the high-pass response.