Scalable modeling with simple topology for stacked millimeter‐wave transformers

Abstract

SUMMARYScalable modeling with very simple topology for stacked millimeter-wave transformers is presented. Because of high-frequency effect and thick metal effect, the architecture is based on single-π and double-π network for transformers with turn ratio 1:1 and 1:2, respectively. The model parameters are extracted from two factors-the layout and process data. Simple and accurate expressions for the self-inductance, mutual coupling inductance, and oxide capacitance are provided for the model. The proposed model is verified by 65 nm technology transformers. A very close agreement is shown for S-parameter, self-inductances of each coil, and coupling coefficient up to 110 GHz.

An efficient mm‐wave LNA impedance match synthesis methodology with coordinate descent method

“…Furthermore, the electromagnetic (EM) simulation is very time-consuming when evaluating the candidate design solutions, and it occupies the majority of the time for the whole optimization. For this reason, in some mm-wave ICs synthesis methods, surrogate modeling methods such as gaussian process, 19,20 kriging metamodeling, 21 radial basis function substitution model, 22 artificial neural networks 23 and others [24][25][26][27][28][29] have been proposed to reduce the high computational cost of EM simulation. In these methods, some EM simulations in local search are replaced by building and updating surrogate models.…”