>
>>If there is no noise in your system, how do you define your "SNR
>>range"? Surely, in order to be able to turn the knob on the system
>>SNR, you should have some finite noise?
>
>
>>A higher order modulation implies that the distance between valid
>>constellation points is smaller. Consequently, for a given transmit
>>power, the amount of noise the system can tolerate without receiving a
>>bit in error is smaller. An SNR that is sufficient for receiving a
>>QPSK signal without error might very well prove insufficient for a
>>higher order constellation.
>>
>>Provided that you have a properly implemented SNR knob, you could try
>>increasing the system SNR in steps (say 1dB) and examine the effect on
>>BER.
>
>----------------------------------
>Thank you for your reply.
>
>What I meant was when I've tried simulating the system without noise in
>the system, it gave me 0 BER reading which is correct. I'm just making
sure
>that it the system is correct without any noise. Sorry if there was any
>confusion caused. When I've added in the additive white Gaussian noise,
it
>does gave me the proper BER vs SNR graph for PSK modulation but not so
for
>higher order QAM.
>
>The effect of SNR towards the BER is the higher the SNR, the lower the
BER
>is.
>
>I've tried scaling up the constellation by using:
>scale = modnorm(modulation.Constellation ,'avpow', 100);
>st = scale*modulate(modulation, data);
>but it seems that there is no effect on the results.
>
>What I've noticed is there is some error in the amplitude of the
combiner
>equation (which is not the same for the modulator's constellation) which
>causes the QAM to decode wrongly. This does explain why only QAM has
>problems while PSK does not as PSK only uses phase shift and not
amplitude.
>Do I need to have some compensation on the receiver side to counter the
>Rayleigh fading channel effect on the amplitude or is there any other
way
>to solve it? 
>
>Thank you for your patience and guidance =)
>

If you use QAM modulation you have to do an "amplitude correction". I
don't know exactly why....:).... Example of such Alamouti scheme for QAM
modulation can be find in m-file AlamotiDiversityCoding_ver2.m from hamid
ramezani. It is available on matlab-exchange web page.

>If there is no noise in your system, how do you define your "SNR
>range"? Surely, in order to be able to turn the knob on the system
>SNR, you should have some finite noise?


>A higher order modulation implies that the distance between valid
>constellation points is smaller. Consequently, for a given transmit
>power, the amount of noise the system can tolerate without receiving a
>bit in error is smaller. An SNR that is sufficient for receiving a
>QPSK signal without error might very well prove insufficient for a
>higher order constellation.
>
>Provided that you have a properly implemented SNR knob, you could try
>increasing the system SNR in steps (say 1dB) and examine the effect on
>BER.

----------------------------------
Thank you for your reply.

What I meant was when I've tried simulating the system without noise in
the system, it gave me 0 BER reading which is correct. I'm just making sure
that it the system is correct without any noise. Sorry if there was any
confusion caused. When I've added in the additive white Gaussian noise, it
does gave me the proper BER vs SNR graph for PSK modulation but not so for
higher order QAM.

The effect of SNR towards the BER is the higher the SNR, the lower the BER
is.

I've tried scaling up the constellation by using:
scale = modnorm(modulation.Constellation ,'avpow', 100);
st = scale*modulate(modulation, data);
but it seems that there is no effect on the results.

What I've noticed is there is some error in the amplitude of the combiner
equation (which is not the same for the modulator's constellation) which
causes the QAM to decode wrongly. This does explain why only QAM has
problems while PSK does not as PSK only uses phase shift and not amplitude.
Do I need to have some compensation on the receiver side to counter the
Rayleigh fading channel effect on the amplitude or is there any other way
to solve it? 

Thank you for your patience and guidance =)

On Aug 28, 11:04&#4294967295;pm, "rayzor" <rayzch...@hotmail.com> wrote:
> Dear All,
>
> I'm currently simulating a Alamouti MIMO system to find out the BER for
> the different number of antennas and different modulation schemes over the
> Rayleigh fading channel with AWGN for the selected SNR range.
>
> I've been using the predefined "modem.pskmod" and "modem.qammod" in matlab
> for my modulation, demodulation and maximum likelihood decoding. All is
> working well for the PSK modulation system but not so for the QAM system.
> When there is no noise in the system, the PSK modulation schemes gave me 0
> BER for the entire SNR range (which is correct)

If there is no noise in your system, how do you define your "SNR
range"? Surely, in order to be able to turn the knob on the system
SNR, you should have some finite noise?

> but the QAM modulation
> schemes starting from 8-QAM onwards (same for 16-QAM and 64-QAM) gave me a
> random BER values.

A higher order modulation implies that the distance between valid
constellation points is smaller. Consequently, for a given transmit
power, the amount of noise the system can tolerate without receiving a
bit in error is smaller. An SNR that is sufficient for receiving a
QPSK signal without error might very well prove insufficient for a
higher order constellation.

Provided that you have a properly implemented SNR knob, you could try
increasing the system SNR in steps (say 1dB) and examine the effect on
BER.

>
> I hope someone would enlighten me on this matter and your assistance would
> be greatly appreciated.
>
> Thank You
>
> Regards,
> Rayzor

Dear All, 

I'm currently simulating a Alamouti MIMO system to find out the BER for
the different number of antennas and different modulation schemes over the
Rayleigh fading channel with AWGN for the selected SNR range. 

I've been using the predefined "modem.pskmod" and "modem.qammod" in matlab
for my modulation, demodulation and maximum likelihood decoding. All is
working well for the PSK modulation system but not so for the QAM system.
When there is no noise in the system, the PSK modulation schemes gave me 0
BER for the entire SNR range (which is correct) but the QAM modulation
schemes starting from 8-QAM onwards (same for 16-QAM and 64-QAM) gave me a
random BER values.

I hope someone would enlighten me on this matter and your assistance would
be greatly appreciated.

Thank You

Regards,
Rayzor