Compensating for Amplifier Non-Linearity in a SEL Controller

TL;DR

This work tackles the instability arising from amplifier non-linearity in Self-Excited Loop (SEL) controllers used for CW superconducting linacs. It introduces an affine calibration framework that models the SSA response with both a slope and an offset, refines the forward-power-to-drive mapping, and thus yields more accurate operating envelopes for SEL modes. The approach is validated with data from a 32 kW solid-state amplifier in the Fermilab PIP-II HB650 cryomodule context, demonstrating improved alignment between drive commands and actual forward power. Additionally, a prototype piecewise calibration scheme is proposed to further capture non-linear behavior by separating low- and high-power regions, setting the stage for more robust operation in real hardware and potential future comparisons between affine and piecewise methods.

Abstract

The Self-Excited Loop (SEL) architecture, used in some continuous-wave (CW) superconducting linacs, relies on a positive feedback mechanism that requires carefully defined operating limits to ensure stable operation. These limits are typically derived from amplifier calibration, which characterizes the relationship between forward power and DAC drive. However, amplifier non-linearity often prevents a simple linear fit of this characteristic, introducing errors that can compromise stability. To address this, we present a modified calibration procedure that incorporates amplifier non-linearity into the SEL framework. The approach is validated with test data from a 32 kW solid-state amplifier (SSA) and a cavity emulator developed for the Fermilab PIP-II linac.

Figures (6)

• Figure 1: SEL architecture.
• Figure 2: Drive limits.
• Figure 3: SSA calibration.
• Figure 4: Observed SSA non-linearity.
• Figure 5: Non-linear fit of SSA.