In this paper novel corrective circuits to avoid multiple operating points in a square-root domain first-order filter are proposed. By employing a DC test it is demonstrated that the filter possesses three operating points (two stable and one unstable) and the corrective circuits enforce the proper operating mode. The corrective circuits and filter are able to operate with very low supply voltages (as low as V GS + 2V DS sat ). Moreover, a detailed analysis concerning the impact that produces the corrective circuits on the filter performance is discussed. Both measurement and simulation results are provided to validate the circuits and analysis employed. circuit simulation packages resort to the well-known Newton-Raphson method, they are able to find only one DC operating point. Recent works are focused on finding MOP by using alternative methods [6, 7], but they are neither general nor easy to be implemented.In fact the existence of MOP is not only a problem for circuit analysis but also for circuit design. Particularly in continuous time implementations this behaviour needs to be avoided in order to enforce the correct operation mode at any time. However, very limited attention has been devoted to explore MOP [8,9] by circuit designers. To solve partially this lack of research, this paper focuses on a circuit designer-oriented treatment of the MOP problem that after achieving a detailed analysis of the circuit behaviour, it provides a solution at circuit level. This methodology is applied to a square-root domain (SRD) system that exhibits MOP, but it can be readily extended to other circuits.The SRD systems [10], a subset of the so called companding systems, are an interesting alternative to develop circuits with very low supply voltage (V DD <2V GS ) required in portable equipment and systems implemented by the most modern CMOS fabrication process [11]. The companding technique exploits the current mode and the large signal behaviour of transistors to achieve a non-linear compression of the voltage swings, allowing an extended dynamic range and less sensitive of the signal/noise ratio to the reduction of supply voltage. Although there is an internally non-linear behaviour in these systems the input-output relationships remain linear [12]. This fact is obtained implementing specific non-linear blocks in such a way that the linearization arises as a whole, i.e. the internal non-linearities cancel out conveniently leading to a linear system. Often these non-linear blocks are implemented employing translinear techniques [13,14]. The first companding systems were implemented employing BJT transistors (or MOS transistors in weak inversion) exploiting their exponential I -V characteristic leading to log-domain systems [15][16][17]. Afterward the systems were realized by the quasi-quadratic law of MOS transistors to avoid the problems of mismatches and poor bandwidth of MOS devices in weak inversion region, leading to SRD systems [18,19].In this paper a SRD first-order filter is designed based on the MOS translinear l...