FPGA Central

Hi,

This has been a very informative forum while i have been working on thi
small simulation. I have a couple of questions now.

1) First I am simulating a simple uncoded 16-QAM OFDM system. I trie
evaluate the performance of such a system in an AWGN channel and th
results i have matches results of a single carrier system using 16 QAM i
a AWGN channel.

2) Now I would like to study the effects of cyclic prefix in an OFD
system and do that i have made the necessary modifications to m
transmitter and receiver. Specifically, i add the cyclic prefix at th
transmitter pass it through a multipath channel, remove the cyclic prefi
at the receiver before the FFT operation at the receiver.

3) The multipath channel i consider have static paths ( No fading, N
Dopper ).

My question is : For such an uncoded system what channel impulse respons
can i consider whereby i can see the benefits of the cyclic prefix.
A comparison of channel impulse response vs prefix lengths ( greater an
less than impulse response length) is what i have in mind.

Also, I am not clear on what sort of compensation needs to be done at th
receiver for the channel. If we assume perfect knowledge of the channel a
the receiver an operation of the form FFT(Y)./FFT(h) would suffice ?, wher
Y is the received signal before FFT operation and h is the impuls
response


Thanks

>Hi,
>
>This has been a very informative forum while i have been working on this
>small simulation. I have a couple of questions now.
>
>1) First I am simulating a simple uncoded 16-QAM OFDM system. I tried
>evaluate the performance of such a system in an AWGN channel and the
>results i have matches results of a single carrier system using 16 QA
in
>a AWGN channel.
>
>2) Now I would like to study the effects of cyclic prefix in an OFDM
>system and do that i have made the necessary modifications to my
>transmitter and receiver. Specifically, i add the cyclic prefix at the
>transmitter pass it through a multipath channel, remove the cycli
prefix
>at the receiver before the FFT operation at the receiver.
>
>3) The multipath channel i consider have static paths ( No fading, No
>Dopper ).
>
>My question is : For such an uncoded system what channel impuls
response
>can i consider whereby i can see the benefits of the cyclic prefix.
>A comparison of channel impulse response vs prefix lengths ( greate
and
>less than impulse response length) is what i have in mind.
>
>Also, I am not clear on what sort of compensation needs to be done a
the
>receiver for the channel. If we assume perfect knowledge of the channe
at
>the receiver an operation of the form FFT(Y)./FFT(h) would suffice ?
where
>Y is the received signal before FFT operation and h is the impulse
>response
>
>
>Thanks
>
>

Hi, As long as the delay spread of channel is less than CP length
multipath won't affect performance. But you have to consider if the pea
does not appear at the very beginning, then there might still be some IS
"leaking" from backwards.

As long as the channel impulse response is less than the duration of
the cyclical prefix, there should be no inter-symbol interference. In
this case the receiver need only remove the cyclical prefix samples and
conduct an FFT for demodulation of the symbol.

Should this not be the case as is usually expected, the receiver needs
to compensate by creating a time-domain equalizer (TEQ) at the front
end. In truth, the purpose of the TEQ is to create an overall shortened
impulse response when placed in series with the actual channel that
ordinarily has a impuse response longer than the CP.

There are any number of methods of computing the TEQ, all with various
tradeoffs.

Good luck.

-V

>As long as the channel impulse response is less than the duration of
>the cyclical prefix, there should be no inter-symbol interference. In
>this case the receiver need only remove the cyclical prefix samples and
>conduct an FFT for demodulation of the symbol.
>
>Should this not be the case as is usually expected, the receiver needs
>to compensate by creating a time-domain equalizer (TEQ) at the front
>end. In truth, the purpose of the TEQ is to create an overall shortened
>impulse response when placed in series with the actual channel that
>ordinarily has a impuse response longer than the CP.
>
>There are any number of methods of computing the TEQ, all with various
>tradeoffs.
>
>Good luck.
>
>-V
>
>

Hi,

Thank you for your replies. I don't think I made myself clear in m
earlier posting.

1) What I want to understand through simulation is the effect of cycli
prefix length on performance in a multipath channel.

For example consider that the channel impulse response is [0.7071 0 0 0
0 0 0 0 0.7071], which is a channel with impulse response length 10
static paths with no fading component).
Now if we consider four cases
1)cp=0,
2)cp=5,
3) cp=11
4) cp =15

The first two cases, based on theory should result in ISI. My question is
when i perform the simulations i do not see much of a difference. Will th
difference be seen only at high SNR's ?


2) Presently, I transmit a known OFDM symbol at the beginning of th
simulation, At the receiver after FFT operation I divide the output wit
the known OFDM symbol values and store them as channel state informatio
for the remaining simulation.
For all subsequent transmissions after the FFT operation i divide th
output with the stored channel state information ( Zero forcing i
frequency domain).
My question is even for cyclic prefix lengths greater than the impuls
response, don't we still need to do equalization operation at the receive
( something trivial as i am doing now).
Also can anyone suggest me a reference to other frequency domai
equalization approaches.

Thanks