Max bandwidth [bps] vs frequency spectrum [z]

[julian403]

Max bandwidth [bps] vs frecuency spectrum [z]

Hello All. I allways have a question regarding the max bandwidth that a carrier wave can it modulate or transport. Due to as we all know, for more frecuency has a carrier, more bandwidth is available. Why is that?

I couldnt find a develop in the books. I thinks that Nyquist can have an answer. can we say that the maximum bandwidth that a sinusoidal carrier can carry is proportional to half the frequency of the carrier wave?

if f [Hz] is the carrier frecuency, so the max bps available is f/2.

Why so the IR spectrum can transport more bandwidth that a radio frecuency? With the same modulation and codification (of course)

 

[cariban]

Re: Max bandwidth [bps] vs frecuency spectrum [z]

Hello All. I allways have a question regarding the max bandwidth that a carrier wave can it modulate or transport. Due to as we all know, for more frecuency has a carrier, more bandwidth is available. Why is that?

I couldnt find a develop in the books. I thinks that Nyquist can have an answer. can we say that the maximum bandwidth that a sinusoidal carrier can carry is proportional to half the frequency of the carrier wave?

if f [Hz] is the carrier frecuency, so the max bps available is f/2.

Why so the IR spectrum can transport more bandwidth that a radio freuency? With the same modulation and codification (of course)

Don't mix bandwidth with carrier frequency. For a baseband signal, the carrier frequency is zero. The carrier just shift the baseband signal to be a passband signal. The bandwidth of the signal is completely determined by the baseband signal before modulation.

Also the unit of bandwidth should be Hz, not bps. There could be multiple bps/Hz depending on the channel SNR (Shannon's theory). For example, 16-QAM modulation will get 4bps/Hz.

 

[julian403]

Re: Max bandwidth [bps] vs frecuency spectrum [z]

Thanks for the reply but why we can get more bandwidth taking a frecuency translation to microwave than the radio frecuency?

I think that if there is a baseband signal of 10 Mbps, and using a 16-QAM modulation (will get 4bps/Hz), so it takes 2.5 MHz. With that baseband signal frecuency spectrum we cant make a frecuency translation less than 2.5 MHz. Its the reason that we assume more bandwidth with more frecuency translation?

I means, Ka band gives more bandwidth that C band. Why is that?

 

[volker@muehlhaus]

Re: Max bandwidth [bps] vs frecuency spectrum [z]

I means, Ka band gives more bandwidth that C band. Why is that?

In general, frequency allocation allows wider bandwidth in the upper bands.

 

[julian403]

Re: Max bandwidth [bps] vs frecuency spectrum [z]

In general, frequency allocation allows wider bandwidth in the upper bands.

But, why is that? there is a fundamental physical basis or simply because it is governed by a governmental institution?

 

[filip.amator]

Re: Max bandwidth [bps] vs frecuency spectrum [z]

But, why is that? there is a fundamental physical basis or simply because it is governed by a governmental institution?

Because you can't transfer over communication channel infinite amount of the information using finite bandwidth.

 

[volker@muehlhaus]

Re: Max bandwidth [bps] vs frecuency spectrum [z]

But, why is that? there is a fundamental physical basis or simply because it is governed by a governmental institution?

It is a legal regulation thing that you are not allowed to use 20GHz +/- 10GHz for one signal over the air.

Some colleagues design wonderful microwave radar systems with great measurement accuracy by using >20GHz sweep range at 77GHz. But it's not legal to radiate that bandwidth, so they can use that in closed metal tanks only.

 

[G4BCH]

Re: Max bandwidth [bps] vs frecuency spectrum [z]

But, why is that? there is a fundamental physical basis or simply because it is governed by a governmental institution?

There is a fundamental limit. You can't occupy more spectrum than that between the channel centre, lets say the modulated carrier to keep things simple, and 0Hz, as you said orignally Nyquist has the answer. Try to occupy more and the result is aliasing around 0Hz. This is why as you go higher in frequency the greater the bandwidth you can use. Regulations aside you could easily transmit a signal 1GHz wide with a centre frequency of 1GHz, but you couldn't do it at say 100MHz; you can't occupy bandwdth that doesn't exist.

 

[cariban]

Re: Max bandwidth [bps] vs frecuency spectrum [z]

There is a fundamental limit. You can't occupy more spectrum than that between the channel centre, lets say the modulated carrier to keep things simple, and 0Hz, as you said orignally Nyquist has the answer. Try to occupy more and the result is aliasing around 0Hz. This is why as you go higher in frequency the greater the bandwidth you can use. Regulations aside you could easily transmit a signal 1GHz wide with a centre frequency of 1GHz, but you couldn't do it at say 100MHz; you can't occupy bandwdth that doesn't exist.

I would like to comment this a little bit. If you are using all-digital transmission, there is no carrier frequency limitation by using complex-valued symbols. The spectrum of complex-valued signal may not be symmetric around 0Hz.
While to transmit the signal in the physical channels, the complex-valued signal must be converted to real-valued by using REAL[.] operator. After being converted to real-valued signal, the spectrum becomes symmetric around 0Hz.
1. If your signal is already real-valued, for example BPSK, the spectrum is thus symmetric around 0 Hz, so you can use 0Hz as carrier frequency. This case is actually widely used in digital circuits to transfer information, for example USB, I2C, DDR...;
2. If you use complex-valued signal like QPSK, 8PSK, 64QAM, the spectrum is not symmetric around 0Hz, then you can't transmit the signal to the channel directly, a carrier with frequency above the bandwidth/2 must be used to avoid frequency overlapping, just like G4BCH has pointed out.

 

[julian403]

Re: Max bandwidth [bps] vs frecuency spectrum [z]

You can't occupy more spectrum than that between the channel centre, lets say the modulated carrier to keep things simple, and 0Hz, as you said orignally Nyquist has the answer

Yes I thought that! I dont care about regulation because If someone use a narrow laser beam to transmit information between two points, it wouldnt bring interference to the EM spectrum due to the little space beam. Or if the signal goes through a fiber, it wont generate interference on air, so the regulation its other subject.

If I take a laser of 100 nm, the carrier is 300 THz. With BPSK the max bit rate would be 300 Tbps, its the physic limit (of cours the sensor and the emitter needs to works with that but thats the information rate limit). With QPSK, it would be 4 times more bandwidth and with 16APSK.... better yet.

Thats right? I think I still do not completely understand it.

2. If you use complex-valued signal like QPSK, 8PSK, 64QAM, the spectrum is not symmetric around 0Hz, then you can't transmit the signal to the channel directly, a carrier with frequency above the bandwidth/2 must be used to avoid frequency overlapping, just like G4BCH has pointed out.

If we call Nyquist there must be in somewhere a /2.

 

[cariban]

Re: Max bandwidth [bps] vs frecuency spectrum [z]

Yes I thought that! I dont care about regulation because If someone use a narrow laser beam to transmit information between two points, it wouldnt bring interference to the EM spectrum due to the little space beam. Or if the signal goes through a fiber, it wont generate interference on air, so the regulation its other subject.

If I take a laser of 100 nm, the carrier is 300 THz. With BPSK the max bit rate would be 300 Tbps, its the physic limit (of cours the sensor and the emitter needs to works with that but thats the information rate limit). With QPSK, it would be 4 times more bandwidth and with 16APSK.... better yet.

Thats right? I think I still do not completely understand it.


If we call Nyquist there must be in somewhere a /2.

The channel capacity is completely determined by the Shannon theorem:
C = W*log2(1+S/N)
If your channel is good, not much noise there, then you will have higher S/N and you can use higher constellation like 1024QAM or even higher. Then you will get higher bit rate.

Note that Nyquist theory is related to sampling rate, not bit rate. The sampling rate should be higher than the bandwidth to avoid spectrum overlap (if the spectrum is continuous, passband may be different).
For a digital communication system:
1. Symbol rate defines the bandwidth;
2. One symbol may carry several bits. BPSK 1 bit, QPSK 2 bits,...,1024QAM 10 bits.
3. Sampling rate must be higher than bandwidth;
4. Digital signal has a spectrum equal to square pulse. So it will be re-shaped by using sqrt filter (square root raised cosine) to limit the bandwidth to be symbol rate * (1+alpha), where alpha is the roll-off factor;
5. The filter signal will be shifted to the desired band by multiply with the carrier. Engineers usually call it as "frequency mixing".

 

[biff44]

Re: Max bandwidth [bps] vs frecuency spectrum [z]

Hello All. I allways have a question regarding the max bandwidth that a carrier wave can it modulate or transport. Due to as we all know, for more frecuency has a carrier, more bandwidth is available. Why is that?

)

Not true at all.
A lot of times a system has a fixed RF bandwidth, and to get more BPS thru, they use a more complex modulation scheme. Like a DOCSIS 3.1 system uses a high order QAM modulation to pump date thru Comcast's aging coaxial lines.

https://www.arris.com/globalassets/resources/white-papers/arris_spectral_efficiency_of_docsis_wp.pdf


There are orthogonal MIMO systems that use stray reflections and multipath to achieve MUCH higher data bandwidth than the RF bandwidth would imply