BPSK Modulation Demodulation via Simulink and Pulse Repetition Catching

Hi to everybody :)

The basic simple BPSK env can be seen below. Sample Time: 1/10000 M-ary number:2

1) If I choose the "Decision Type" to "Log-likelihood ratio" in "BPSK Demodulator" then the output of the "Spectrum Analyzer2" has no peak but the "figure;plot(xcorr(simout2,'unbiased'))" gives a big magnitude value than if the "Decision Type" was chosen to "Hard Decision" same xcorr code written and same time the Spectrum of the "Hard Decision" choice gives a peak. It makes me confusion. If "Hard Decision" was chosen, then "Spectrum Analyzer2" gives a higher peak. This is OK. but if I write "figure;plot(xcorr(simout2,'unbiased'))" for "Hard Decision", then it gives me lower peak value than acc. to "Log-likelihood ratio". I was expecting a bigger magnitude value because the "Spectrum Analyzer2" gave a higher peak.

2) How can I catch the pulse repetition time (to see ?) via ACF or any other tech ? Above, xcorr did not give repeated triangle like shapes for above simulink scenario. Does it means there is no repeated pulse ? don't think so but ?