Switching speeds determine how fast (or how densely)data can be encoded on a carrier wave .Carrier wave may be a very high frequency visible light but if the encoding speed is slow,like say turning a light on and off mechanically,data rates will be slow.
So the frequency of carrier waves only becomes a problem when we have encoding technologies that can exceed the frequency of the carrier.
On the other hand we need a large bandwidth for transporting huge amounts of data being generated. This is where visible light shines. Even though most of the telecom carriers use frequencies in the IR the property of glass fibers is such that it can carry broadband signals in both IR and in visible light. So high data densities are achieved simply be encoding multiple streams of information on a single fiber thread.
This is such a big advantage that comparatively slow data encoders(current state of the art ethernets being able to o/p at few gbs only) can be colliminated so that terabits of data can be sent over the fiber. More if multiplexing tech can be improved(as it will be) in the future.
In simple words we can continue increasing capacity of our channel without improving the data encoding speed. If the channel is saturated we can add more channels.
This can continue to a point where essentially limitations will be set by how much of data can be consumed or stored. Not how much of it can be transmitted. As an example we can broadcast live data to tv sets 24/7 in HD but if the data were to be stored ,very soon all of the storage units will be full ,before the transmission channel experiences even the tiniest of strains.
One of the key pieces in data transmission tech is the oscillator which generates the carrier wave & also the data signal which is put on the CW.
For visible & IR CW we have leds & lasers but what do we have for microwaves,mmwaves?
We have crystals. Quartz has served & continues to serve well for lower end of frequencies. But as we move to higer frequencies (30 gigabose,100gigabose,1terabose)we run into problems. Some dielectrics hold promise for generating CW for these frequencies but it gets increasingly difficult to cycle over the frequency ranges that are available to us here. There are no band gap engineering tricks that we can pull off to create CW sources at these frequencies. The gap exists not just at terabose but also for the bands in mmwave region.
The most promising alternative is provided by what are known as optoelectronic oscillators
General principle of operation of optoelectronic oscillators/emitters
short version: if we switch light on and off very rapidly we can generate currents in a photodetector at high frequencies. This can then be amplified(or used directly) to produce a CW.
Longer version
1. A laser light source(imp. for it to be a laser,can't be a regular led. Power is not imp. Can be a miliwatt laser) is made to pass through a delay line.
2. The delay line can be a long optical fiber or a small sphere (technical term whispering gallery) which acts as a resonator & in which light can travel to induce non linear optical effects (changes in frequency,phase etc that are not possible otherwise) & generate what is known as a frequency comb.
3. In a frequency comb the frequencies of light get shifted by tiny amounts. For example:
original laser frequency 400 terabose.
Optical comb frequencies 400.01 terabose,400.02 terabose etc.
In this example we have produced shifts of .01*1000=10 gigabose & .02*1000=20 gigabose respectively. This is just an example actual numbers depend upon construction of the delay line /WG oscillators.
4. Next in this setup we have a photodetector. We feed our shifted frequencies to it & out comes the oscillating current. All we need now is an antenna that can filter the desired frequency.
OEOs can produce extremely pure frequencies with little noise & decently high efificiencies.But there are some problems with OEO
1. It is difficult to miniaturize. Long delay lines wont work at all. WG can work but its still bigger than crystals used otherwise.But more imp. its difficult to fabricate(glass being the choice of material here) shapes are imp. Mm size perfect sphere on a chip is tough.
2. It needs a laser source for the non linear effect to take place which further restricts miniaturization.
Principle of operation of our emitter
1. One of the key principles on which our design is based upon is that non linearity can be induced not just via energy source(like laser) but also inside a material(with photonic crystals for example). In other words optical non linearity can be either material or energy based. So we can use any normal light source if our material is non linear.
2. The other thing is that photodiodes can be patterened & don't have to be a simple transducer. In this way they themselves can become emissive like an antenna & result in better light matter interactions. For this they must be constructed from a patternable semiconducting material.
With this setup we can emit from a light source. Semiconducting patterned photodiodes will generate frequency comb & produce time varying current resulting in a CW. A wavelength difference of just 1 nm can induce an oscillation in terabose range.
In essence what we have done is create an optical equivalent of a klystron. Instead of an electron beam we have a photon beam. Everything else is the same.
This setup gives us widely tunable set of frequencies for our data transmission needs.
What channels can carry the frequencies we will be generating?
Regular wires will work fine,if they can be made thin enough.It is possible to make nano carbon wires <10 micrometer thick (thats about the same thickness as individual strands in optical cables). Their high conductivity & ballistic transport properties can carry low frequency(frequency less than visible light) signals 100 km at a stretch just like optical cables carry light signals for the same distance but with 2 very compelling benefits
1. Tighter integration with existing electronic components. No optoelectronic switching. Wireless & wired communication using the same set of frequencies & oscillators(but on different mediums wire or air so isolation can be maintained).
2. Tighter materials integration. We can combine electrification and communication efforts & become more productive. The same set of wires that do one thing can be made to do another.
Notes
A carrier wave is a communication channel on which data can be sent
At the moment we don't have technologies that can realistically approach encoding/decoding at the speed of light. Photonic computing has major challenges to overcome before it can get there. And one of the biggest one being a fast rewritable optical storage. Right now electronic storage offers the best read/write performance & optical storage tech is good for archicval storage only. Even there it faces stiff competition from magnetic tapes(which are getting better everyday).
1Bose=1hz
Glossary
CW-Carrier Wave
OEO-opto electronic oscillators
WG-Whispering Gallery
Comments
Post a Comment