In OFDM, First we modulate our message signal and then we take the IFFT of that signal to load the message on subcarrirers. And after that the cyclic prefix is added to combat the frequency selective nature of channel. What should be the length of the cyclic prefix?
The length of the cyclic prefix should be at least equal to the number of channel taps minus one, so that the linear channel filtering appears as if it is a cyclic convolution.
There's also a tradeoff, right? The length of the cyclic prefix, therefore the guard interval, is determined by the amount of echo tolerance you need, for that application, in that environment. Then comes the design spiral.
Longer cyclic prefix means greater immunity to inter-symbol interference, caused by multipath distortion. With very many OFDM subchannels, long guard intervals (cyclic prefix) are affordable, without eating too much into overall channel capacity. You will have slow symbol rate in each subchannel, so that relatively long guard intervals will not unduly impact the bits/symbol capacity of that subchannel.
Why not always go for long guard intervals, and the necessary many, many subchannels?
Doppler shift! If the subchannels are very close together, then mobility can more easily cause doppler shift to make the receiver lose track of each subchannel. So if we're to expect simple equalizers (convolutional filters) to be adequate in OFDM, we have to keep those subchannels far apart, compared with what is possible in fixed service.
So, I'd say, the length of the cyclic prefix should take into account the amount of echo tolerance required for that application, in that environment. In turn, that determination will tell you how many subchannels to plan on.
In 8 MHz channels, general practice seems to be, for fixed service, 8000 subchannels or more. The guard interval 1/8 or 1/16 is doable. For mobile service, looks like 2000 or 4000 subchannels are the preferred choice. And this would scale up, depending on the total channel width.
I think my colleagues have responded largely to your response.
In the general case, we take CP = 1/4 size of IFFT or 1/8 size of IFFT or 1/16 size of IFFT, depending on the number of paths imprinted by the channel and on the number of useful symbols transmitted.
good luck
Thanks for the valuable answers.
Great. So I conclude that I have understood-
✓The minimum length of Cyclic Prefix should be the number of channel taps minus one.
✓And greater the length of CP greater will be immunity to ISI.
I want to know further that is there any limitation of increasing length of Cycic Prefix so that I should keep on mind that in my Underwater visible light communication application.
Dear Benyarou Mourad Albert Manfredi Emil Björnson Thank you for you answers.
The longer the CP length the shorter the amount of time utilized for useful information. Keep that in mind when designing/implementing your OFDM system.
- Badges
- Science topic
- Similar topics
- Social Psychology
- Attitude