Improving the Efficiency of Digitally Controlled Switching Gradient Amplifiers for Driving Different Gradient Insert Coils

Introduction: High-performance MRI systems demand the use of digitally controlled switching gradient amplifiers (SGAs) that need complex electronic circuits and control algorithms to achieve not only good image quality but also good power efficiency [1-2]. Moreover, the imaging capabilities of clinical MRI system may be improved by the use of gradient insert coils that achieve gradient amplitudes and slew rates two to twenty times larger than conventional whole-body gradient systems. These high-performance gradient insert coils allow the acquisitions of high-resolution MR images in small animal models that cannot be achieved with stock clinical systems. Current waveform errors, voltage oscillations, and poor energy efficiency are some examples of the problems that users and designers of gradient coils may face. Some of these problems may be associated with a mismatch between the characteristic impedance of the gradient insert coils and the normal load or tuning parameters of the gradient amplifiers. Digitally controlled SGAs offer the flexibility to easily change tuning parameters that affect the behaviour of the control algorithm for different loads, gradient waveforms, current amplitudes and gradient switching requirements. In this work, we present a method to tune SGAs on a clinical MRI system with insertable gradient coils attached, with only limited access to information about the underlying electronic circuits, communication protocol, and control algorithms that govern the SGAs. All that is required is a basic understanding of how the tuning parameters are stored on the gradient processor, and how they are related to the frequency response of a gradient coil, i.e. inductance (LC), resistance (RC), and cut-off frequency or roll-off (RO) or to the feedforward and feedback control system, i.e. high-frequency gain (HFG), loop gain (LG), and rate feedback (RFB).