The theme of this year's ISSCC is adaptive circuits. What are we adapting to? In the case of the semiconductor industry, we are adapting to a tanking industry. One recently laid-off Texas Instruments engineer--one of the 3600 victims of the carnage two weeks ago--told me he was given a couple of days to pack up his desk and go find himself. He was part of the wireless division in Dallas that TI had been planning to sell. But when there were no good buyers, they just laid everyone off.
But wireless is going to be about much more than just phones, and if the papers here are any indication, that research is building to critical mass. One of the most promising applications for the kind of work engineers once only did for cell phones is in medical implants and body area networks.
The ex-TI engineer pointed to a talk Sunday night by Ali Hajimiri, a high-speed and RF integrated circuits luminary at Caltech. Hajimiri enjoys long, moonlit walks on the beach, high-speed and RF, and low-frequency high-precision circuits. In 1993, he worked on BiCMOS chipset for GSM and cellular units. In 1997, he investigated low-phase-noise integrated oscillators for Lucent Technologies (that's Bell Labs to you). Turn-offs: mean people.
But on Sunday night, our protagonist from TI said Hajimiri was explaining biology to the audience. "DNA, RNA," our frustrated engineer said, "I didn't know what I was doing there."
Wireless transmission and wireless power are adapting to burgeoning medical applications: eye implants like the artificial retina at December's International Electron Devices Meeting; brain implants, and little implanted reservoirs that keep the blood evenly doped with drugs such as insulin. One acronym you can expect to hear a lot more of is MICS -- the medical implants communications standard.
The MICS band operates at 402-405 MHz. So there have been a lot of papers about how to transmit at that wavelength from a tiny, featherweight chip. Jeremy Holleman at the University of Washington created a 500-microwatt neural tag for our favorite cyborg moth with an analog front end frequency multiplying transmitter: a low noise amplifier core built on an op-amp core.
He multiplied an initial 45Hz frequency by a series of dense circuit diagrams that I didn't understand, and before I could drift off into sweet oblivion, the nine-fold multiplier had worked its magic. The interesting thing about this is that Holleman's transmitter can also operate in the 433 MHz ISM (industrial, scientific and medical) band. That's good for the moth--ok, maybe less for the moth than for the moth researchers--and good for all the poor little medical rats who run around with giant cables extending out of their craniums. Hopefully that grisly tether can be replaced before too long with a chip that lets them run around unimpeded. You know, except for the chip digging into their brain.
But back to our protagonist from Texas Instruments: he says that after a couple of days at ISSCC, he's looking at getting into biomedical applications. It's clear from the research presented so far that for medical implants to become a reality--whether they are retinal implants, brain implants, or brain-machine interfaces--the first thing you need to be able to do with those things is communicate with them. The second is to power them indefinitely.
The best practice case for energy harvesting and tiny radios seems to be the DARPA moth. That project is a test bed for coming up with a lot of the next-generation power storage, energy scavenging, and wireless communication devices that will be crucial for next-generation biomedical applications.
Next post: energy harvesting and wireless power. Can it be done? Can it be done without burning your house down?