FPGA Central

Theres something Im confused about regardibg the cyclic prefix. In article
I've read they say that if a refletcion symbol arrives anytime before th
cyclic prefix ends, then that is fine. Why is just the arrival of that
before cyclic prefix ends important, even after the len of cyclic prefi
that delayed symbol still exists in energy to interfere with my symbol
Even if I start sampling at the right OFDM symbol start time, am I no
still receving the reflection symbol energy for the length of that symbol.

Zeph80 said the following on 29/09/2006 20:47:
> Theres something Im confused about regardibg the cyclic prefix. In articles
> I've read they say that if a refletcion symbol arrives anytime before the
> cyclic prefix ends, then that is fine. Why is just the arrival of that
> before cyclic prefix ends important, even after the len of cyclic prefix
> that delayed symbol still exists in energy to interfere with my symbol.
> Even if I start sampling at the right OFDM symbol start time, am I not
> still receving the reflection symbol energy for the length of that symbol.

The cyclic prefix (CP) has two purposes. Firstly, it acts as a
time-domain guard interval, i.e. it prevents intersymbol interference
(ISI) from the previous symbol.

Secondly, it converts the linear convolution of the channel into
circular convolution. Circular convolution in the time domain is
equivalent to multiplication in the frequency domain (when using a DFT;
see
http://en.wikipedia.org/wiki/Discret...tion_theorem);
i.e. it ensures there is no intercarrier interference (ICI) either.

However, these only occur if the channel length (maximum reflection
delay) is shorter than the CP length.

--
Oli

On Fri, 29 Sep 2006 14:47:38 -0500, "Zeph80"
<surabhi_talwar@hotmail.com> wrote:

>Theres something Im confused about regardibg the cyclic prefix. In articles
>I've read they say that if a refletcion symbol arrives anytime before the
>cyclic prefix ends, then that is fine. Why is just the arrival of that
>before cyclic prefix ends important, even after the len of cyclic prefix
>that delayed symbol still exists in energy to interfere with my symbol.
>Even if I start sampling at the right OFDM symbol start time, am I not
>still receving the reflection symbol energy for the length of that symbol.

In the receiver the FFT only looks at the window after the end of the
cyclic prefix, i..e, the energy received during the guard interval is
discarded. This means that any of the energy from the previous OFDM
symbol does not affect the current symbol, since any delayed image of
the previous symbol didn't extend past the guard interval. Energy
from the guard interval that extends into the FFT window looks like a
circular convolution of the channel with the current FFT window.

It's pretty cool, really, IMHO. It does cost the additional
bandwidth and energy required to transmit the cyclic prefix, but the
benefits are substantial.


Eric Jacobsen
Minister of Algorithms, Intel Corp.
My opinions may not be Intel's opinions.
http://www.ericjacobsen.org