A new multi-valued neural network for the extraction of lumped models of analog circuits
Analog Integrated Circuits and Signal Processing
| Hi-index | 0.00 |
The metal-oxide-semiconductor field-effect transistor (MOSFET) alternating-current (AC) behavior with two different parasitic channel series resistance descriptions were studied. An absorbed parasitic series resistances approach was compared to the conventional lumped resistance approach both theoretically and with real device values. The results suggested that absorbing the parasitic channel series resistances into the current description decreases the AC accuracy of the MOSFET model compared to conventional model with lumped resistances. The input admittance, output impedance, gain and backward gain characteristics were studied and the largest differences emerged in the input and output behavior. The theoretical study is confirmed by empirical comparisons of an 80 脳 1.0 μm 脳 90 nm n-channel metal-oxide-semiconductor (NMOS) device characteristics up to 110 GHz.