Kelly Byron1, Fraser Robb2, Shreyas Vasanawala3, John Pauly1, and Greig Scott1
1Electrical Engineering, Stanford University, Stanford, CA, United States, 2GE Healthcare, Aurora, OH, United States, 3Radiology, Stanford University, Stanford, CA, United States
Synopsis
Wireless power transfer
(WPT) inside the MRI bore, as part of completely wireless patient coils, should
have minimal RF interactions and be a compact system that allows for flexible
placement of the coils while maintaining power transfer efficiency. Changes in coil positions can drastically
reduce power amplifier efficiency. Here
we present a MEMS based auto-tuning approach to compensate for the varying load
seen by the power amplifier. An array
primary coils also using high-power RF MEMS devices to steer power to a local
secondary coil is also demonstrated.
Introduction
As
the number of channels on patient coils continue to increase, there is a
proportional increase in the number of connectors, cabling, and RF baluns, all leading
to increased patient discomfort. To make
large coil arrays comfortable for the patient, it would be ideal to use lightweight,
flexible coils that are completely wireless, including wireless power delivery
inside the MRI bore. We have previously
shown an MRI wireless power transfer (WPT) system operating with minimal RF
interactions through the use of RF gating1. However, this system uses a large power
amplifier and antenna tuner that cannot be operated inside the magnet. A class-EF power amplifier can be made
compact and efficient to operate inside the bore2, but only for a
single load impedance. To account for
the varying load of the coupled WPT coils an adjustable impedance matching
network needs to be used to preserve power amplifier efficiency. Both systems had a single WPT primary coil for
flexibility of placement of the secondary but at the expense of lower mutual
coupling. Here we present an RF MEMs
based impedance matching system that can function
inside the MRI bore and a MEMs switched array of primary coils to allow
a greater coupling efficiency over a wide area.Methods
Figure 1 shows a block
diagram of our system, as well as a more detailed look at the impedance
matching system. A directional coupler
measures the forward and
reverse voltages as inputs to an AD8302 gain/phase detector. An Arduino microcontroller then samples and
converts the
outputs to reflection
coefficient and impedance which are compared to an impedance lookup table for
auto-tuning to the desired 20 Ω load of a class
EF amplifier. The impedance matching network is a pi-match
with 12 RF MEMS switches, as shown in Figure 2.
The inductor and capacitor values are binary weighted in order to cover
an even distribution of S11 values. Switching
each value individually covers the entire range of S11 values in Figure 4B. For initial testing, capacitors are switched
in pairs as a pi-match of equal capacitor values to limit lookup table memory. The calculated S11 values are shown in Figure
4C, with measured values that are slightly shifted due to parasitics on the
MEMs matching board. The RF switches
themselves need to be able to block the MRI high-power RF levels inside the bore. As a result, we are using the RF MEMS switches
presented in 3 with the simple opto-coupled gate drive circuit shown
in Figure 3 for control from the Arduino digital pins. Figure 4 shows the basic schematic and construction
of the array of primary coils for 10MHz WPT, with the same MEMS switched and
gate drive circuit controlling which coil is switched in series with the
matching network. Results
Figure 3B and C shows
that the falling edge delay is much longer than the rising edge for the MEMS
gate driver. This is a property of optocouplers
and the fall time delay can be reduced by decreasing the value of RL, however,
this leads to a proportional increase in the current and the power drawn from
the 82V supply. The primary coils have a
Q of about 163 without the MEMS switches.
With the addition of the switches and cabling the Q drops to about 44. The impact of adding a secondary coil near
the primary coil that is switched in is shown in Figure 4H, demonstrating much
better coupled impedances and the ability to transfer power with a few
centimeters between coils. With a 20 Ω resistor on the output
of the matching network the tuning process achieves a minimum reflection
coefficient of about 0.17 and with the coupled coils it is about 0.31. Discussion & Conclusions
Having primary and
secondary coils of the same size optimizes the coupling between them and
improves efficiency. Even with greater
separation between coils the majority of power will be transferred to the
secondary coil. Smaller coils will also
pick up less EMF from the scanner. In
our preliminary testing, Figure 5 shows the system inside the MRI bore where we
were able to tune the coils and transfer and receive power. Future work will
include improved organizing of the system cabling and testing the impact of the
WPT system on MRI images. The ability to employ high power RF MEMs devices
allows easy multiplexing of the primary WPT coils to steer power more locally
to a secondary harvesting loop. Similarly,
MEMs based auto-tuning provides more flexibility in matching to
primary/secondary coil position changes.
Acknowledgements
Acknowledgment: We would like to thank GE Healthcare for their
research support. This project is
supported by grants R01EB008108, P01CA159992, and
R01EB019241. References
1.
K. Byron, P. Stang, S. Vasanawala, J. Pauly, and G. Scott, A High Power RF
Gated Wireless Power Transfer System. Proc. Intl. Sco. Mag. Reson. Med. 24
(2016).
2.
K. Byron, F. Robb, S. Vasanawala, J. Pauly, and G. Scott, A MRI Compatible
Class-EF Power Amplifier Designed to Drive a Wireless Power Transfer System.
Proc. Intl. Sco. Mag. Reson. Med. 25 (2017).
3.
D. Spence and M. Aimi, Custom MEMS Switch for MR Surface Coil Decoupling. Proc.
Intl. Sco. Mag. Reson. Med. 23 (2015).