February 11, 2014, ISSCC, San Francisco—Advanced biomedical systems are combining integrated sensors and actuators to offer better data and improved comfort and usability at the point of care. These systems not only help to reduce the cost of healthcare, but enable early diagnosis, personal point of care, and therapy outside of traditional clinical settings.
This digitally assisted healthcare delivery increasingly incorporates new technologies to help people achieve health care rather than disease treatment. These integrated systems offer great potential to improve patients' quality of life.
Amre El-Hoiydi from Phonak Communications presented a wireless digital hearing aid. Like an iPhone, hearing aids can be equipped with accessories to connect to other audio inputs such as TV, music player, etc. These accessories can be either wired or wireless but must work within the constraints of extremely limited battery capacity and very stringent requirements for minimizing noise.
One noise source is the voltage ripple across the interface which must be less than 10micro volts. This battery voltage ripple affects the analog microphone input, the PWM output, and internal reset thresholds. These challenges are exacerbated by using a wireless interface. The communications protocol must be developed to ensure high fidelity audio is transferred across interface, so an unconditional audio packet repetition allows low power and low delay while improving interference robustness.
Power management includes a boost –buck DC to DC converter to slow battery current consumption variations and provide greater supply isolation for the hearing aid. The radio runs at 2.4 GHz supplying a 2Mb per second data stream.
Nick Van Helleputte from Imec described a signal acquisition system for personal health monitoring. The increasing number of devices for connected personal health all face similar challenges; low power, small size, low cost, and must deliver high-quality multiple sensor data. This SoC can interface to temperature, ECG, bioimpedance, accelerometers, and other sensors.
The primary inputs go through an instrumentation amplifier, which has very stringent specifications. To overcome the shortcomings of existing instrumentation amplifiers, the DC offset is subtracted at the input in a differential feedback scheme. Other inputs are handled by a bio-impedance signal chain that can support square wave or pseudo-sine wave excitation in the I/O processing stages. The resulting data are processed in a DSP accelerator to offload the Cortex M0.
Po-Tsang Huang from National Chiao Tung University is investigating multi-channel neural sensing with a 2.5-D microsystem. The electrocorticography (ECoG)system uses TSVs to connect micro-probes to the signal acquisition and processing chips, comprised of an 180 nm analog and three 65nm digital functions. The interposer also provides interconnections between the various chips.
The 16-channel analog front end has differential signal conditioning, analog multiplexers, and 4 11-bit successive approximation ADCs. The processed signals are separated into different frequency bands to distinguish the various wave types and feature classification. Power manage includes clock and power gating to the pseudo-multi-master interposer bus and hierarchical header. Overall power is 676.3 microwatts.
Po-Hung Kuo from National Taiwan University researches locomotive ICs for in-vivo monitoring. The detector module uses electrolytic bubbles for propulsive force to allow implantable, wireless powered, remote controlled body function monitoring. Electrolysis electrodes on the 4 sides of the module create O2 and H2 bubbles that drive the module away from the bubbles.
The bubbles are about 20nL, much smaller than normally occurring bubbles in circulation. A 10MHz signal supplies power and a 1Mb/s data and command stream to the chip. The CMOS chip in 0.35micron has no external components and responds to control signals to move and change directions .
