On the design and manufacturing of miniaturized microstripline power splitters for driving multicoil transmit arrays with arbitrary ratios at 7 T.

The purpose of the current study was to implement unequal microstrip power splitters for parallel transmission at 7 T that are optimized for size and loss and that can be configured for a wide range of power ratios. The splitters will enable the use of more transmit coils without a corresponding increase in the number of transmit channels or amplifiers to control specific absorption rate, shorten RF pulses, and shim inhomogeneous RF fields. Wilkinson unequal power splitters based on a novel microstrip network design were optimized to minimize their size under 8 cm in length and 9 cm in width, enabling them to be included in coil housing or cascaded in multiple stages. Splitters were designed and constructed for a wide range of output power ratios at 298 MHz. Simulations and bench tests were performed for each ratio, and a methodology was established to adapt the designs to other ratios and frequencies. The designs and code are open source and can be reproduced as is or reconfigured. The single-stage designs achieved good matches and isolations between output ports (worst isolation -15.9 dB, worst match -15.1 dB). A two-stage cascaded (one input to four outputs) power splitter with 1:2.5, 1:10, 1:3, and 1:6 ratio outputs was constructed. The worst isolation between output ports was -19.7 dB in simulation and the worst match of the three ports was -17.8 dB. The measured ratios for one- and two-stage boards were within 10% of the theoretical ratios. The power-handling capability of the smallest trace was approximately 70 W. Power loss for the one- and two-stage boards ranged from 1% to 3% in simulation compared with 5.1% to 7.2% on the bench. It was concluded that Wilkinson unequal microstrip power splitters can be implemented with a small board size (low height) and low loss, and across a wide range of output power ratios. The splitters can be cascaded in multiple stages while maintaining the expected ratios and low loss. This will enable the construction of large fixed transmit array-compression matrices with low loss.

Ratio-adjustable power splitters for array-compressed parallel transmission.

PURPOSE: To implement and validate low-loss ratio-adjustable power splitters (RAPS) for array-compressed parallel transmission (acpTx). METHODS: In acpTx, a small number of transmit channels drive a larger number of transmit coils, which are connected via an array compression network that implements optimized coil-to-channel combinations. Previous networks comprised a bank of power splitters, followed by attenuators to implement the amplitudes of the compression weights for each coil, but this resulted in high power dissipation in the network. Recognizing that an acpTx network need only implement relative attenuations between outputs, a RAPS circuit was developed which combines power splitting and relative attenuation, and has low insertion loss. RAPS circuits were experimentally validated and used to build an array compression network for a one-channel-to-four-coil spiral acpTx excitation experiment. RESULTS: Bench tests showed that the RAPS circuits came within 0.05 dB of the desired output ratios, and power dissipation was approximately 0.5 dB (10%). The spiral excitation experiment showed that the ability to optimally drive four coils with a single channel reduced excitation error by 46% compared to driving one coil, without using attenuators in the array compression network. CONCLUSION: RAPS circuits enable the construction of low-loss array compression networks for parallel transmission. Magn Reson Med 79:2422-2431, 2018. © 2017 International Society for Magnetic Resonance in Medicine.

Experimental implementation of array-compressed parallel transmission at 7 tesla.

PURPOSE: To implement and validate a hardware-based array-compressed parallel transmission (acpTx) system. METHODS: In array-compressed parallel transmission, a small number of transmit channels drive a larger number of transmit coils, which are connected via an array compression network that implements optimized coil-to-channel combinations. A two channel-to-eight coil array compression network was developed using power splitters, attenuators and phase shifters, and a simulation was performed to investigate the effects of coil coupling on power dissipation in a simplified network. An eight coil transmit array was constructed using induced current elimination decoupling, and the coil and network were validated in benchtop measurements, B1+ mapping scans, and an accelerated spiral excitation experiment. RESULTS: The developed attenuators came within 0.08 dB of the desired attenuations, and reflection coefficients were -22 dB or better. The simulation demonstrated that up to 3× more power was dissipated in the network when coils were poorly isolated (-9.6 dB), versus well-isolated (-31 dB). Compared to split circularly-polarized coil combinations, the additional degrees of freedom provided by the array compression network led to 54% lower squared excitation error in the spiral experiment. CONCLUSION: Array-compressed parallel transmission was successfully implemented in a hardware system. Further work is needed to develop remote network tuning and to minimize network power dissipation. Magn Reson Med 75:2545-2552, 2016. © 2016 Wiley Periodicals, Inc.

Array-compressed parallel transmit pulse design.

PURPOSE: To design array-compressed parallel transmit radiofrequency (RF) pulses and compare them to pulses designed with existing transmit array compression strategies. THEORY AND METHODS: Array-compressed parallel RF pulse design is proposed as the joint optimization of a matrix of complex-valued compression weights that relate a full-channel physical array to a reduced-channel virtual array, along with a set of RF pulses for the virtual array. In this way, the physics of the RF pulse application determine the coil combination weights. Array-compressed pulse design algorithms are described for four parallel transmit applications: accelerated two-dimensional spiral excitation, multislice RF shimming, small-tip-angle kT -points excitation, and slice-selective spokes refocusing. Array-compressed designs are compared in simulations and an experiment to pulses designed using four existing array compression strategies. RESULTS: In all cases, array-compressed pulses achieved the lowest root-mean-square excitation error among the array compression approaches. Low errors were generally achieved without increasing root-mean-square RF amplitudes or maximum local 10-gram specific absorption rate. Leave-one-out multisubject shimming simulations demonstrated that array-compressed RF shimming can identify useful fixed coil combination weights that perform well across a population. CONCLUSION: Array-compressed pulse design jointly identifies the transmit coil array compression weights and RF pulses that perform best for a specific parallel excitation application. Magn Reson Med 76:1158-1169, 2016. © 2015 Wiley Periodicals, Inc.

Field programmable gate array compression for large array multispeckle diffuse correlation spectroscopy.

Significance: Diffuse correlation spectroscopy (DCS) is an indispensable tool for quantifying cerebral blood flow noninvasively by measuring the autocorrelation function (ACF) of the diffused light. Recently, a multispeckle DCS approach was proposed to scale up the sensitivity with the number of independent speckle measurements, leveraging the rapid development of single-photon avalanche diode (SPAD) cameras. However, the extremely high data rate from advanced SPAD cameras is beyond the data transfer rate commonly available and requires specialized high-performance computation to calculate large number of autocorrelators (ACs) for real-time measurements. Aim: We aim to demonstrate a data compression scheme in the readout field-programmable gate array (FPGA) of a large-pixel-count SPAD camera. On-FPGA, data compression should democratize SPAD cameras and streamline system integration for multispeckle DCS. Approach: We present a 192×128 SPAD array with 128 linear ACs embedded on an FPGA to calculate 12,288 ACFs in real time. Results: We achieved a signal-to-noise ratio (SNR) gain of 110 over a single-pixel DCS system and more than threefold increase in SNR with respect to the state-of-the-art multispeckle DCS. Conclusions: The FPGA-embedded autocorrelation algorithm offers a scalable data compression method to large SPAD array, which can improve the sensitivity and usability of multispeckle DCS instruments.