Search

Wire bonding is widely used in various radio frequency (RF) devices and is one of the common forms of signal interconnection. In this article, a new geometric model with related physical parameters is used to describe various actual types of the bondwire, and the corresponding calculation method for the total length of the bondwire is obtained. Consequently, a new equivalent circuit modeling method of the single‐bondwire can be developed and the accuracy of model has been improved. On this basis, using the concept of partial inductance, the modeling method of double‐bondwire is proposed, which can be applied to asymmetric double‐bondwire modeling. The model of bondwire obtained by this modeling method is applied to the design of PA, and the special electrical properties of the bondwire are used to verify a novel harmonic control method, which second and third harmonics could be controlled. The results show that in the Sub‐6G, the proposed bondwire model has good accuracy, and its performance is close to the measured results. The designed PA combined with harmonic control network has good performance, achieving drain efficiency of 62.6%–70% and output power of 40–41.5 dBm at 3.2–3.8 GHz. It is proved that the proposed bonding line modeling method can be applied to the design of high‐efficient PA. [ABSTRACT FROM AUTHOR]

To cope with the degradation of digital predistortion when bandwidth changes, a bandwidth adaptive behavioral model with dynamic structures for radio frequency (RF) power amplifiers (PAs) is proposed in this article. By applying the nonlinear post‐compensation technology, PA's nonlinearity is decomposed and analyzed. Short‐time Fourier transform and transfer functions are adopted to study the influence of bandwidth variation on the linear memory effect. Based on PA's memory effect analysis, a bandwidth adaptive model with dynamic structures and a coefficient update scheme with low complexity are proposed. Experiments carried on PAs show that the proposed model achieves similar linearization performance as other conventional models. Simultaneously, the proposed model has fewer coefficients and a reduction in the coefficient update of at least 50%. [ABSTRACT FROM AUTHOR]