Current Projects

Wearable wireless thermometers for internal body temperature measurements

Radiometer Block Diagram

Funded by: the National Science Foundation
Graduate students: Dr. Robert Scheeler (graduated 2013), Parisa Momenroodaki

Microwave radiometry is an attractive method for internal thermometry, with the possibility of a wearable device which can continuously monitor temperature inside body tissues in different parts of the body, store the data, and transmit it to a digital medical record. Currently, there are a limited number of available device solutions, and they are usually not wearable or wireless. We are working on a possible path to implementing such a thermometer, with some initial results demonstrating about 0.2K measurement sensitivity and a difference between the maximal and minimal error w.r.t. a thermocouple measurement of 0.5K. Several probes for multi-band radiometers have been developed at frequencies of 410MHz, 1.4, 2.7 and 4.9GHz. The main challenges of RF interference, sensitivity, calibration, spatial resolution, miniaturization, and probe design are discussed. The block diagram of a 3-frequency internal body temperature measurement system is shown in the figure. The power received from tissues layers by the narrowband probes is coupled to the Dicke radiometer circuits, which consists of a switch that is required for calibration, a low-noise amplifier (LNA) followed by band-pass filters and a diode detector circuit. The hot and cold loads are used for continuous temperature calibration of the radiometer. The detector output is a DC voltage which can be integrated over time to increase the signal-to-noise ratio (SNR). This output is digitized, processed and transmitted through a wireless unit. A micro-controller unit (MCU) controls the radiometer switches enabling phase-sensitive detection of the very low human black-body power levels.

In a number of disorders, this temperature difference changes and is not easy to measure externally. For example, long duration of exercise in heat conditions, such as in the case of athletes or soldiers under heavy training, can provoke brain heating leading to premature fatigue and even death. Cancer cells can have increased temperatures, as can inflamed tissues such as joints of arthritis patients. Sleeping disorders are accompanied by changes in the circadian cycle, which are in turn related to changes in phase and amplitude of periodic core body temperature variations. Infants suffering from hypoxia-ischemia have an elevated brain temperature, and if detected can be effectively treated by hypothermic neural rescue. In addition to diagnostics, therapy can be assisted by internal temperature monitoring, e.g. in hyperthermia for cancer treatment and clinical high-intensity focused ultrasound (HIFU) for noninvasive therapy, where the knowledge of local temperature increase would be useful. Applications we are working on include circadian cycle monitoring (with collaborator Dr. Gurley, see We are also working with Speag and IT’IS in the modeling area with the hope of using their virtual population models (see

Package Design for Improving Active Device Efficiency

Funded by: Infineon
Graduate students: Sushia Rahimizadeh

Package and parasitics

Active devices, such as transistors and diodes, are often packaged to allow easier integration with circuit boards. At RF and microwave frequencies, the reactances of the package limit the impedance range that can be presented to the intrinsic device. Additionally, this parasitic reactance is largely capacitive and will limit the harmonic content necessary for achieving high-efficiency operation. In cooperation with Infineon Technologies, packages are designed to present desirable harmonic impedances to the device by using a combination of bond-wires, MOS capacitors, and package parasitics. Accurate full-wave modeling of the passive package environment is demonstrated along with a methodology for manipulating package and bond-wire geometry to achieve specific harmonic impedances. The package model is used in conjunction with harmonic balance simulations including a transistor die non-linear model to design harmonically-terminated packages for highly-efficient power amplifiers at S-band.

High Field MRI Probes

Graduate students: Patrick Bluem

MRI Probe Phantom

In clinical 1.5 T and 3 T magnetic resonance imaging (MRI) instruments, the object being imaged is closely coupled to the detector through near fields and detection can be viewed as quasi-static. MRI can also be excited and detected using long-range coupling with traveling waves, demonstrated by several research groups over the past few years. One potential benefit of this approach is more uniform coverage of samples that are larger than the wavelength of the NMR signal. Uniform spatial coverage in MRI is traditionally achieved by tailoring the reactive near field of resonant probes. This approach is valid when the radio-frequency wavelength at the Larmor frequency is substantially larger than the target volume, which does not hold for wide-bore high-field systems (>4 T and >60 cm bore diameter). The motivation for using high DC magnetic flux density (B0 > 3 T) is increased spatial resolution, improved SNR, better parallel imaging performance and potential for improved contrast. However, the proton Larmor frequency for hydrogen increases from 64 MHz at 1.5 T to 447 MHz at 10.5 T, resulting in waveguide effects both in the bore and the imaging volume. To help control the excited modes, structures are placed around the imaging volume to modify the boundary conditions.

In collaboration with Harvard University and the Center for Magnetic Resonance Research (University of Minnesota), circular patch probes have been measured on 16.4 T small-bore, 7 T wide-bore, and 10.5 T wide-bore scanners.

High-efficiency PAs and transmitters for communications, radar and medical applications

PA targeted for the W-CDMA downlink band exhibiting over 84 percent PAE at 2.14 GHz. The projects in this area are continued work in reducing power consumption in analog front ends with new circuit topologies that give higher efficiency. Our first publications in this area were in 1995, with the first demonstrated microwave-frequency class-E power amplifier. Our results in X-band and UHF power amplifiers had record published efficiencies [ref], and we are continuing a strong effort in this direction at lower microwave frequencies with increased output power. In this area, my group collaborates with the power electronics and analog electronics group at the University of Colorado at Boulder. New directions that we are expanding in is in maintaining linearity [ref] with high efficiency at high power levels, scaling to higher frequencies, increasing the level of integration in advanced materials such as GaN, applying the concept to radar waveforms, etc [ref]. Effective contours to high efficiency PA design.

Another related area is intelligent transmitters, which involves sensing, control algorithms, dynamic tuners and dynamic biasing. The tuners can be based on existing electronic technology, or on micro-electromechanical components (RF MEMS) and their packaging and hybrid assembly, and it is a considerable challenge to make these devices practical. An application that would benefit from impedance and supply tuning is in medicine for transmitters used in tumor ablation and blood-vessel sealing.

Previous Research

LabVIEW Open LSNA Toolbox

Global Variable Funded by: National Instruments, CU-Boulder
Developed by: Dr. Tibault Reveyrand

This LabVIEW toolbox enables high level functions on your current instrumentation while presenting high-level abstraction. After defining basic drivers of your instrumentation, the user can conect to power meters, load-pull systems, vectorial receivers and even several calibrated LSNAs with just one polymorphic VI. The toolbox comes with several examples in LabVIEW. One of them is a complete DC power supply sweep and S-parameters measurements dedicated to transistor measurements.

The Large Signal Network Analyzer is an already defined meta-instrument in the toolbox and includes its own calibration procedure. An LSNA can be built without any downconverter, with mixers or sub-samplers.

The theory behind the LabVIEW code can be found in the Special Topics course taught by Dr. Reveyrand in Spring 2016.

A version of the toolbox (with proprietary external software removed) can be downloaded here.

Wireless powering for battery-less sensors

Modular rectenna circuit.An area in which we have promising initial results, as well as a best paper award, is in RF energy harvesting and wireless powering of wireless sensors. This is an area with a strong collaboration with the Colorado Power Electronics Center (CoPEC), with strengths in low-power management design. The work resulted in a comprehensive patent application and licensing of the IP by several companies, e.g. Cymbet. The applications are for low-maintenance batteryless sensors for manufacturing environments, structural monitoring, and healthcare. We have shown that broadband statistically varying randomly polarized background microwave radiation can be efficiently rectified and the stray energy stored over time for useful electronic applications. We have also shown that FCC-compliant low-power transmitters can be strategically placed to enable constant very low power density energy delivery and storage. Our goals related to this research are to improve the integration of our current hybrid demonstrations, and to expand the circuit-antenna library so that we can address many concrete applications with the best-suited architecture. Source pull block diagram; Source pull results

DARPA ONR Micro-Power Conversion (MPC) Project

MPC Overview Poster

Graduate Researchers: Scott Schafer, Andrew Zai, Michael Litchfield Researchers: Dr. David Sardin and Dr. Tibault Reveyrand Collaborators: CoPEC group at the University of Colorado (

The University of Colorado leads a project in integrated GaN microwave transmitters with dynamic supplies, with TriQuint Semiconductor as a subcontractor. We especially thank Dr. Chuck Campbell for technical advice, as well as Maureen Kalinski and John Hitt from TriQuint. We also thank Dr. Dan Green (DARPA) and Dr. Paul Maki from ONR for support competent and thoughtful encouragement along the way, as well as Dr. John Albrecht (formerly DARPA) for the opportunity to work on this project.

The project goals are to:

  1. Design 10-GHz transmitter that can efficiently amplify high peak-to-average ratio (PAR) signals
  2. Implement transmitter in GaN technology with a high level of integration of RF PA and dynamic power supply (supply modulator)
  3. Enable digitally reconfigurable efficient transmission of broadband signals (500MHz) for communications and spectrally confined radar

Our current results for the GaN MMIC PAs measure state-of-the-art PAE>60% at 10GHz with output power >10W and a large signal gain >20dB from a two-stage power-combined architecture in the 150-nm GaN on SiC TriQuint process. In addition, we have demonstrated 70% efficient single stage PAs with watt-level output, and the PAs are designed to maintain efficiency over varying drain supply. The dynamic supplies (supply modulators) are also implemented as MMICs in the same GaN process and show >90% efficiency with 5W output power and 100MHz switching in a 2mm x 2mm chip.

We are interested in both communications signals and radar signals with PAR>7dB. Our test communication signals are typically multi-carrier (OFDM) signals and we have demonstrated supply modulators that can reproduce envelope bandwidths over 300MHz. The supply-modulated transmitters support varying amplitude radar pulses that are frequency modulated, allowing spectral confinement. We are also exploring radar signals with varying pulse shapes and amplitudes on a pulse-to-pulse basis.

Another aspect of the project have been nonlinear measurements and modeling of the entire system. We have successfully used a time-domain large signal network analyzer to characterize some of the circuits, and have shown that system-level modeling flow can be used to describe the entire system, from base-band to the RF modulated carrier.

For more information, please see summary program review slides:

  1. UCB MPC Review June 2014
  2. UCB MPC Review Feburary 2014

Three-dimensional micro-fabricated microwave and millimeter-wave circuits and antennas

Another active area of research has been in collaboration with Nuvotronics LLC (DAPRA and NASA) in the area of wafer-scale microfabricated coaxial lines and passive and active coaxial-based components. The advantages of these lines, fabricated by Nuvotronics, is extremely low loss into the millimeter-wave range, extremely good isolation of neighboring lines enabling high density circuits, broad bandwidth and low dispersion, and amenability for integration with passive and active surface-mount components. Our research goals are focused on design of completely new components in this technology, in order to push the bandwidth, power handling and flexibility for various communications and sensing applications. Some results include 22:1 bandwidth impedance transformers and 22:1 bandwidth power divider networks which operate up to millimeter-wave frequencies. G-band steerable antenna array; A micro-coaxial broadband wilkinson power divider achieves an 11:1 bandwidth from 2-22 GHz; A 10-dB micro-coaxial directional coupler centered at 26 GHz.

Cryogenic Microwave Radome

Waveguide characterization of FSS for cryogenic radomes.In this project, we are developing a layered anisotropic periodic artificial electromagnetic material that increases isolation between transmit and receive antennas of a bistatic radar at 10GHz, while allowing high gain in the receive direction. The receiving antenna is envisioned to be cryogenic for improved gain with small electrical size, and the radome will be thus cooled internally. This allows for low radome loss and the potential of integrating Josephson Junctions for inductive tuning. This project is a collaboration with Dr. Horst Rogalla (NIST and Research Professor at CU), funded by AFOSR.