CR018 Wideband Noise Model for AMS/RF CMOS Simulation

Analysis of Internally Bandlimited Multistage Cubic-Term Generators for RF Receivers

“…Values for and were extracted from simulated device models and found to be 0.55 and 1.8 V m, respectively. The value of is consistent with other recent reports of submicrometer CMOS devices [26], [27] but a bit lower, likely due to device operation near weak inversion and with relatively low . was simulated and found to be 0.8.…”

“…Starting from (24), the time parameter is removed from the arguments of and , while is relegated to a subscript to condense the expressions (24) The autocorrelation of the error is therefore given by the following, where the delay offset is also placed within the subscripts of and where it exists: (25) Note that terms such as are equal to zero due to the fact that one out of the four terms is a zero-mean bit sequence that is independent of the other three bit sequences. Multiplying out only the terms resulting in and yields the following: (26) (27) A brief discussion on the notation in (26) and (27) is in order. Terms with the same letter can be shown to have the same PSD.…”

Digitally Enhanced Alternate Path Linearization of RF Receivers

“…Values for γ and E sat were extracted from simulated device models and found to be 0.55 and 1.8 V/µm, respectively. Although γ is traditionally thought to be greater than unity in submicron devices, this result is consistent with other recent reports of noise in submicron CMOS devices [39,41]. H (ω LO − ω 2 )/H (ω LO − ω 1 ) was simulated and found to be equal to 0.8.…”