Artificial Intelligence Based High Power Calibration Method for RF Pulse Amplifiers

The development, design, and implementation of a calibration method for a high-power RF pulse amplifier system using an artificial intelligence (AI)-based control architecture is presented. The implementation of the proposed artificial intelligence (AI) method for the RF control system allows the application of the calibration process across a broad frequency range from 1 kHz to 200 MHz in CW mode, to pulse up to 100 kHz within a 35 MHz bandwidth, and across various power ranges. The proposed AI-based system is trained to predict power levels based on the operational frequency. The proposed AI-based calibration method is implemented for a case study using a commercial linear RF amplifier, E&I A075, to deliver RF power up to 75W in CW mode. The system underwent testing for various load conditions ranging from 1:1 to 4:1 VSWR. It has been confirmed that the linear amplifier successfully pulsed up to 100 kHz with stability. Results also show that the proposed method reduces the error by over 80% in a 35 MHz bandwidth and with 75W in CW mode.