Let's Talk Frequencies: Can Faster Freq Carry More Data?

[tenacity2986]

Is a faster frequency capable of carrying more data than a lower frequency? i would say yes because the period is shorter in a faster frequency so you can modulate more data (if amplitude modulation).

Does more power get used if the frequency is faster but amplitude is the same?


These arent homework problems, just me trying to connect the dots...
i saw a chart in my vlsi class that points out that microprocessor freequencies have tapered at 3ghz... Why don't we run them faster? is there a physical barrier limitiation to the amount of data we can rate the processors at? like not fast enough transistor/gate switching? Hmmm makes me wonder...

 

[waht]

Is a faster frequency capable of carrying more data than a lower frequency? i would say yes because the period is shorter in a faster frequency so you can modulate more data (if amplitude modulation).

Yes. We use fiber optics because we can cram more data on a beam of light (infrared or visible) which is of much higher frequency than a radio wave.

In short, if the data modulation is faster then the carrier, then the data becomes the carrier. And that could be a mess.

Does more power get used if the frequency is faster but amplitude is the same?

Yes, definitely, it takes more energy to cause more oscillations per unit time.

i saw a chart in my vlsi class that points out that microprocessor freequencies have tapered at 3ghz... Why don't we run them faster? is there a physical barrier limitiation to the amount of data we can rate the processors at? like not fast enough transistor/gate switching? Hmmm makes me wonder...

Generally, the transistors cannot be switched faster, and heat dissipation becomes a major issue, and they become unstable. That doesn't mean that we don't have transistors that can switch faster, we have transistors capable of switching frequencies up to 500 GHz. We just can't squeeze millions of them on a single chip.

 

[tenacity2986]

we can in fact squeeze more transistors onto a single intergrated chip which is why the field of very large scale integration (vlsi) is so big now... The new intel core i7 has something of the sort of 800 million transistors on a single die. They already have started processing at 32nm and these chips are being designed to have 1.9 BILLION transistors on a chip. The amount of power is actually reducing on these chips while the number of transistors are increasing. yet they are operating at the same frequencies. Essentially we are trying to accomadate power prowess with low power long battery life for on the go device users...

 

[ravioli]

Yes. We use fiber optics because we can cram more data on a beam of light (infrared or visible) which is of much higher frequency than a radio wave.

..
Yes, definitely, it takes more energy to cause more oscillations per unit time.
...

We use fiber optics because we have more bandwidth, which means a more possible datarate. Remember that if we are given 100kHz of bandwidth, then we can transfer up to 200k symbols/s without ISI. So really its not due to the faster oscillation of the carrier wave, but due to the increased bandwidth capacity of a fiber optic line that we can transfer more data.

And... it does not take more energy to cause more oscillations per unit time(theoretically of course). Remember that the energy spectral density is $$|X(\omega)|^2$$ and that to get the energy contributed by frequencies between $$\omega_1$$ and $$\omega_2$$ is $$\frac{1}{\pi} \int_{\omega_1}^{\omega_2} \! |X(\omega)|^2 \, d\omega$$. Then from there you have the relationship between energy and power...

So in summary, it's all about the bandwidth of your signal, and not the frequency of your carrier wave. Of course bandwidth will be limited by the capability of the transistor that you use, and the frequency responses of all of the components in your system. So that's where the limitation starts to occur in as you mention a microprocessor, and it's primarily a bandwidth limitation caused by the transistors. Also note that whenever you increase the speed of the processor, you are also decreasing the symbol time. This reduction of symbol time actually spreads your spectrum out (More Bandwidth!). If the transistors cannot transmit this new bandwidth, then your signals going through the system will be misinterpreted which means system instability!

 

[ravioli]

Also I just thought of another thing about frequencies, especially when related to microprocessors(and I should've thought about it earlier...). You will have more power consumption as you increase the clock frequency. This is because in CMOS circuits, the power consumption occurs mainly on the transitions. Faster clock, more transitions, more power...

But in terms of a power or energy in a signal, my prior comment still holds.