I. Introduction
The performance of continuous-time (CT) delta-sigma modulators is adversely affected by circuit non-idealities such as gain coefficient errors due to process variation and temperature. The gain coefficients are mapped onto capacitor-transconductance ( C=g m ) or resistor-capacitor ( R · C ) time constants. In CMOS technologies, g m -C or R-C time constants are liable to variations of up to ±30% [1] . These variations may have significant impact on the noise transfer function (NTF), the signal transfer function (STF) [2] , [3] , the signal-to-noise ratio (SNR), the in-band noise (IBN) [4] , the maximum stable amplitude (MSA) [5] and, thus, the stability of the modulator in general [3] .