New Theory Leads to Gigahertz Antenna on a Chip | Hacker News Reader

New Theory Leads to Gigahertz Antenna on a Chip

spectrum.ieee.org

69 pointsjonbaer22 comments

I took antenna theory in college for a couple years and then worked in the automotive field designing in-glass antenna systems for cars in the 1990s for a few years.

It has been a few years but reading the article I couldn't help but recall some of my college courses and how I was able to apply some of what I learned in college to the state of the art at the time.

We use to have to drive cars by a local glass company to have the rear window removed and taped back in until I got to the office. Once at the office I would lay the glass on a metal frame and layout an antenna using gold conducting tape. Put the windshield back into the car and blast it with network spectrum analyzer.

Antenna's have both a theoretical wave length and an actual wave length as measured through the medium you're radiating from.

We had loads of conduit in our work shop so one day I managed to grab a couple pieces of conduit and made some rough back of the napkin calculations to convert the theoretical wave length to an actual wave length for metal conduit. My work was for AM and FM radio in the 88-108MHz.

One day a customer comes into the office who knew me well from many visits and saw me hack sawing a piece of conduit with a wire hanging off the side. He asked what I was doing and I told him I was calibrating an antenna.

We were a start up so we had to make due with stuff lying around.

I couldn't help but read the article and wonder how the dielectric mentioned would effectively shorten the wavelength. Wave length in a medium can be thought of the effort to push back and forth the electrons to radiate.

Another analogy is it's easier to wave your hand in air than it is in water; water being akin to the dielectric. A dielectric with 60% efficiency seems more material science than electrical engineering.

Interesting work. Curious what materials they use or perhaps going to higher frequencies.

One basic point to keep in mind here is that in dielectrics the wavelength of radiation can be many times shorter than in freespace - up to 20 or 30 in some cases. So if you have some system that is too small to set up the resonances you would need for an antenna you can put it in a dielectric with a high K value and the resonances will all be shifted to higher frequencies.

It looks like there's an interesting bit of analysis in here, but that basic concept is as as old as Maxwell. As for gigahertz antennas on a chip... well, you all probably have one in your pocket :)

Yeah, looking at the PRL, it sounds like they're advertising a new theoretical framework for _designing_ antennas, not, like, some new, henceforth-undiscovered law of Electromagnetism.

Every cellphone I've disassembled has the 2.4GHz antenna not on-chip, but in the case. I've seen PCB antennas as well, but not on-chip whenever efficiency matters.

After having read the PRL paper [1], I am not really sure why they left out gain (and pattern) measurements for the antennas. In my experience, radiation efficiency is only one part in determining the received signal strength. The impedance match is also crucial in determining the effective gain.

The figures in their paper also seem to demonstrate strong resonant coupling effects, which do not always translate to effective radiation.

[1] http://journals.aps.org/prl/abstract/10.1103/PhysRevLett.114...

Since the antenna is quite small it would be bad at receiving signals for that reason alone. Reception isn't really the important issue here though...

Incredible that things that, at least to me, seem rather close to the fundaments, are still discovered today.