I am studing some pulses that are about 200-300 ns wide and a rise time of few ns,    which settings would be best for coincidence time measurements?

In some preliminary work I found for 700 MegaS the time measurement is better without time calibration (in -0.05 to 1V) rather than with time calibration in -0.5 to 0.5,  my pulses are about 60 mV.   Is it expected that always with time calibration time accuracy would be better or depends?   

Also I use this code snippet to find time for channel 1 and the same idea for chan. 2.

// find peak in channel 1 above threshold
      for (i=0 ; i<1022 ; i++)
         if (waveform[0][i] < threshold1 && waveform[0][i+1] >= threshold1) {
            tt1 = (threshold1-waveform[0][i])/(waveform[0][i+1]-waveform[0][i])*(time[0][i+1]-time[0][i])+time[0][i];
            break;
         }

Thanks!

Abaz