I analyzed the following runs that Oleg Tsai took for the BSMD/BPRS timing.

RUN  BSMD  BPSD

    Coarse/Fine     Coarse = 106.6 ns Fine =Coarse/10

13039108 106/4 106/8
09 106/7 107/1
10 107/0 107/4
11 107/3 107/7
12 107/7 108/0
13 108/0 108/3
14 108/3 108/7
15 105/5 105/9
16 105/8 106/2
17 106/1 106/5
19 105/0 105/4
20 104/0 104/4
21 104/5 104/9
 

In the plot below, the graphs represent the integral for 50000 events of (ADC-PED)>300 for each BSMD/PSD crate, normalized to the 4th point. 

The top two rows are the BSMD crates and the bottom row are the BPSD crates.  PSD Crate 10 had bad data for this set of runs so I did not include it when calculating the average timing for the BPSD crates.

The optimum timing for the BSMD is 104 + 176 ns = 105 + (176-106.6)/106.6 ~ 105/7

The optimum timing for the BPSD is 104 + 204 ns = 105 + (204-106.6)/106.6 ~ 105/9