An Efficient 24-30 GHz GaN-on-Si Driver Amplifier Using Synthesized Matching Networks

This paper presents a broadband GaN microwave monolithic integrated circuit driver amplifier (MMIC DA) with compact dimensions of 1.65 mm x 0.78 mm for 5G millimeter-wave communication. The optimal impedance domain satisfying the preset goals was first acquired using the simplified load-pull procedure and small-signal simulations, followed by a weighted average method to determine the reference center matching point from which the optimal intrinsic load can be deduced. By means of de-embedding load-pull contours, modeling based on theoretical analysis, and simulation fitting for parameter identification, the nonlinear output capacitance and a series RLC model circuit approximating the input impedance response of the stabilized transistor were extracted. Under the design principle of fully absorbing the parasitic parameters of the device, explicit formulas and tabulated methods related to the Chebyshev impedance transformer were applied to construct filter-based synthesized matching networks at each stage and finally convert them into an implementable mixed-element form via the single-frequency equivalence technique. Measured on-wafer pulsed results for the proposed two-stage DA across 24-30 GHz demonstrated up to 31.1 dBm of saturated output power (P-sat) with less than 1 dB total fluctuation, 19.3 +/- 1 dB of small-signal gain, and 39.8% of peak power-added efficiency (PAE) at the mid-frequency.

Automated summary

This paper presents a compact GaN microwave monolithic integrated circuit driver amplifier (MMIC DA) for 5G millimeter-wave communication. The design process involves determining the optimal impedance domain, extracting the nonlinear output capacitance and input impedance response of the stabilized transistor, and constructing filter-based synthesized matching networks using the Chebyshev impedance transformer. The measured results demonstrate high saturated output power, low fluctuation, and good power-added efficiency.