Fix for LTIME and DLS NaNs aiming for TS

In issue #620 , the problem was addressed that for some signal processes the LTIME and DLS obtained NaNs, causing (many) signal events to be cut away. Since the TS is near and a fix is preferred, a temporary fix is proposed only for these NaN cases. The other cases are left untouched. For LTIME, the initial ctau is used, and for DLS, the terms related to distance are removed, in case of NaN. I will try to provide more information on the effects of the MR asap. For now, the following has been observed:

• Old total BW (left) vs. new total BW (right):

• The LTIMEs that were NaNs before probably have underestimated values, since comparing the LTIME to the initial LTIME for non-NaN events caused the LTIME value to be lower in case the initial LTIME was used.
• This means that an additional population (the previous NaNs) at low LTIMEs appears:
•