Re: QAM constellation
Hi m_nasiriyan,
The probability to (erronously) detect a certain symbol when an adjacent symbol is transmitted depends mainly on the distance between them (more distance = less error probability). For that, the error probability of the system is dominated by the minimum distance between symbols.
For a given average symbol energy, and in the case of a 16-QAM like constellation, the minimum distance is maximized when the distances between adjacent symbols are uniform. In this way, the performance of the system (bit error probability vs. Eb/No) is maximized with the uniform constellation, giving a performance that is 4 dB worse than BPSK.
Regards
Z