Time-division polynomial pre-distorter for linearisation of 1.5 T MRI power amplifier

Abstract

A time-division polynomial (TDP) model is proposed for modelling and linearising a 1.5 T magnetic resonance imaging (MRI) power amplifier (PA) with strong non-linearity in high input signal dynamic range. In order to demonstrate the merit of this non-linear model, a 64 dBm 1.5 T MRI PA (63.89 MHz) and two different Sinc-pulse signals are used in modelling and linearisation measurements. The TDP is compared with the conventional non-memory polynomial (NMP) and no digital pre-distortion for the 1.5 T MRI PA, which is driven by test signal with 2 ms time length and 2% duty cycle. The proposed TDP leads to up to 9 dB improvement in the normalised mean square error compared with the NMP in two different test signals. More importantly, TDP illustrates significantly better reduction in amplitude modulation/amplitude modulation (AM/AM) and amplitude modulation/phase modulation (AM/PM) conversion compared with the NMP.

Keywords

• polynomials
• radiofrequency power amplifiers
• linearisation techniques
• magnetic resonance imaging
• mean square error methods
• AM-PM conversion
• AM-AM conversion
• normalised mean square error
• digital pre-distortion
• NMP
• nonmemory polynomial
• linearisation measurement
• sinc-pulse signal
• nonlinear model
• magnetic resonance imaging PA
• TDP model
• time-division polynomial model
• MRI power amplifier
• time-division polynomial pre-distorter
• magnetic flux density 1.5 T
• frequency 63.89 MHz
• time 2 ms