[Halld-pid] tof times

[Alexander Ostrovidov]

On Wednesday, March 04, 2015, Beni Zihlmann wrote:
> Hi All,
> 
> I would like to share another example why I think it is important to look
> at raw data and individual pmts. Attached you find two root files. Each
> shows one 2-d histogram with the horizontal axis the tdc time in ns
> and the vertical axis the PMT number where the first 44 are left/up
> and the second 44 are right/down.
> if you project to the horizontal axis each individual PMT you find that
> some PMTs seem to have some issues like multiple peaks. some more
> than others, some very distinct. This maybe helpful to identify potential
> problems with the PMT or the electronics.
> 
> cheers,
> Beni
> 

Thanks, Beni, Multiple peaks from signal reflections are something
we have to live with unless we decide to get rid of simple passive splitters.
Some splitters have worse reflections than others (I guess it depends on
the deviation of actual parameters of 3 resistors in a splitter). I think, 
this shouldn't affect the quality of TOF data as long as we are using 
only the first-per-channel hit in the 300-500 ns triggered window. The
debatable question is if we should put such rejection of the  "reflection"
hits into JANA TOF hit factories, or let the user code somewhere down
the pipeline deal with them.

Sasha