Last Post: August 28, 2014:
I have a rather in-depth question which is keeping me busy for more than a week now :-)
As far as I understand a single radio can only transmit on one frequency at a time. It can hop around but in such case it continuously (sequentially) changes the frequency being used. What I don't understand is how more bandwith (like 40 MHz channel bonding instead of a single 20 MHz channel) makes it possible to double the transfer rate?! A single radio can not send data on all these frequencies at once (in parallel). It can only hop around within the 40 MHz band being used. But hopping around means it is only using each frequency for a very slight amount of time so it seems you would not really gain more speed in doing so. I know it helps to avoid interference though...
The only way I can explain it is when the radio is not constantly sending a signal on one single frequency (to leave space between frames, ... ?) and is moving at a much higher speed between frequencies. In such case it doesn't have to wait X milliseconds/nanoseconds to send the next frame, it can commence immediately on a different frequency and as such gain time and those end up transmitting more data.
I read something about OFDM subcarriers...rather difficult :) But still, I suspect the data within these subcarriers is not being sent in parallel ?!
Thanks a lot!
There isn't really any "hopping around" as you state. For Wi-Fi type communications, that only happened with 802.11 frequency hopping (FH) radios. This was deprecated several years ago. That is not to say that you won't still find some companies still using it.
OFDM transmits on orthogonal sub-carriers 312.5 kHz apart. 40 MHz segments are twice as wide as the 20MHz ones and that is how it gets twice as much throughput in the same amount of time. It's actually slightly better than that because of other techniques it can take advantage of.
Bluetooth does use frequency hopping, but the techniques there have also changed over the years.
One of the best descriptions of how OFDM works is in Matthew Gast's 802.11:The Definitive Guide book.
Thanks for your answer. I'll have a look at the book you referred to!
Does this mean a radio can transmit on several frequencies at the same time? If that is the case: is there a limit... do these frequencies need to stay close together or can you e.g. bond the entire 5 GHz band and thus transmit on all those MHz's between 5 and 6 GHz at once/in parallel? I always thought a receiving radio could only listen to one frequency at any moment... however these multiple sub-carriers seem to suggest that is not the case? I know you have a base band and side bands surrounding it but thought these were really close together.
It's probably some kind of misconception at my side :-)
If you look carefully at OFDM and what makes the sub-carriers "orthogonal", you'll see the secret.
Normal 802.11a/g uses 20 MHz wide channels, yet the final 802.11ac designs will be able to use 160MHz wide channels. Eight times as wide = eight+ times as much throughput (under ideal conditions).
The 20/40 widths used in 802.11n are using adjacent channel pairs. With ac it is theoretically possible to use discontinuous slices of the 5 GHz band.
One thing - don't get confused about the difference between channels, frequencies, and channel widths.
You should look at the Shannon–Hartley theorem, it was developed in the '40s and underlies pretty much all telecommunications.
It states that the (maximum) Channel Capacity = Bandwidth of the Channel * log2 (1 + Signal/Noise). Note that doubling the Bandwidth doubles the Capacity -> 40MHz wide channels have twice the theoretical capacity of 20 MHz channels (with equal S/N).
This is the theorem that mathematically explains WHY wider channels have more capacity, WHY channels with stronger signal have more capacity, and WHY channels with less noise have more capacity.
This is completely independent of the modulation technique or the media - it applies to wired communications in exactly the same way that it applies to wireless communications.
I wish I could help more, but it has been some decades since that class...
I had a quick look at some of the material pointed to. When I read about DSSS I noticed I had the same question, it doesn't relate to OFDM (I think).
"Direct Sequence Spread Spectrum (DSSS) uses a coding technique for digital radio transmission on multiple radio channels. DSSS technology actually “spreads” a transmitted radio signal out over a wider frequency band. This coding technique modifies the frequency spectrum to spread the signal across may frequencies from its carrier, thereby increasing it bandwidth. The DSSS receiver does not detect the narrow band signals because it is designed to “listen” to a pseudo-random code sequence generated by the transmitter in this wide channel bandwidth. The disadvantage is that the effective RF power is reduced, reducing the radio range and penetrating capability."
But how does this "spreading" looks like/work? In order to better explain my problem, I made some drawings (attached).
Drawing 1) is described as wide band, 2) as narrowband
Most often when radio waves are explained, they are drawn as in 3) -> a pure sine wave with a certain frequency (the wave is being modulated to carry data... not shown here) If you would transmit this sine wave, it would look like a very tiny vertical line on a spectrum analyser, right?! It's just one frequency, very narrow. Maybe this is only theory and any wave transmitted is not as pure as a sine wave and comes with neighbouring waves with slightly different freq?
How can you get such a broad signal as shown in 1) (e.g. when using DSSS) Is it like shown in 4): sine waves with different frequency being transmitted rapidly after each other? Or as in 5): waves with different freq transmitted in parallel? I'm trying to figure out how this simple sine wave is producing figure 1)
Looking forward to your answers!
"If you would transmit this sine wave, it would look like a very tiny vertical line on a spectrum analyser, right?"
Correct ! Although the vertical line may not be tiny in the vertical direction, just narrow.
When OFDM was first proposed to the FCC for use as a "spread spectrum" technology, they denied it as not being true spread spectrum. Finally, a sufficient number of people convinced them that it has the same effect.
If you were to look at real spectrographic plots of 802.11 (DSSS) signals, you'd see that they are not the smooth curves that are typically drawn, but consist of many points (roughly) evenly spread to each side of the carrier frequency.
Search for terms like DSSS, "Spreading Factor" or "Processing Gain". Concentrate on the difference between "chips" and (data) "bits", and how the bits are converted to chips. Note also the distinction between "coding" and "modulation". See how the Exclusive Or Function (XOR) is used to create the "chipped" sequences.
Remember one of the hallmarks of Spread Spectrum transmission is that they use much more bandwidth than would be required by say a CW signal to transmit the same data.