Look at Digital Communications written by Proakis.  Many
non-constelation depended FEC is shown on BPSK plots because its easier
to compare.  As soon as you add the constellation in on top then you
add many more varibles that make it harder to compare FEC's.

I like to normalize the rate of all the FEC's to capacity (Shannon
bound) to  give a better indication how they perform relative to each
other.

Proakis treats modulation and FEC separate and then both combined.  My
suggestion is to compare FEC's with BPSK and then analize the
modulation type.

Hi,

I was looking at the upper bound on BER for convolutional encoder. 
"High Rate Punctured Convolution Coded for Viterbi and Sequential
Decoding"
by David Haccoun and Guy Begin talks about the BER plots for BPSK and
QPSK
with and without puncturing.
I am not able to find anything which talks about BEr calculation for QAM.
Can someone suggest reference?

thanks in advance!