[d2n-analysis-talk] TDC Thresholds

[Brad Sawatzky]

Redirecting back to d2n-analysis-talk...

On Wed, 20 Jan 2010, David Flay wrote:

> > Now, can you find out how many TDC hits per event there are per PMT
> > for this run/dataset?
> >
> > Are the data-sets the same for both of your plots?  ADC:TDC seems to
> > have 2.7M events, but the TDC only plots have 0.47M events.  Does
> > that make sense?
> 
> No -- the TDC only plot has less runs chained together -- I was trying
> to get the 2D plot to look like the one I sent you back in August, and
> thought I needed better statistics on the 2D plot, and then realized
> my binning was wrong, and I kept the extra runs chained together.

Alright.  No problem then.

[ . . . ]
> > Also, maybe I'm imagining it, but there seems to be hints of a band
> > around TDC channel 200--300 in the 2D spectra.  That would be ~600ns
> > after the prompt peak.  After-pulsing maybe?  If so, then it looks
> > like the THaCerenkov class is reporting the hit closest to the TDC
> > stop.  The 1877s use a LIFO (last-in first-out) model to store hits
> > in the buffer, so that would be consistent with the code in
> > THaCherenkov.C that uses hit[0] only...
> >
> I can kind of see that -- mainly on the plot labeled with [6], but not
> so prominent anywhere else.

It's hard to see in the 2D plot, the effect (if present) would be
enhanced on a 1D TDC plot with an appropriate x-axis range.

> Concerning the multihit capability -- I don't think it's utilized by
> the analyzer -- the source code mentions the multihit possibility of
> some modules, but when using the MakeSelector, my array of TDCs is
> singly indexed -- wouldn't it be doubly indexed, ranging from 0 to 5
> for each TDC -- like:
> 
> L.cer.t[0][4] -- for the 0th mirror, and the 5th hit?  (if multihit
> capability was 'enabled')

You are correct that the THaCherenkov class does not expose more than
one TDC hit, even from a multi-hit TDC.  All hits are decoded at the low
level though (which is why the THaDecData class, for example, can access
them).

The THaCherenkov class, derived from the THaScintillator class, seems to
be selecting the hit that comes closest to the TDC stop.  For
multiple-hit events, that is the least likely hit to be correct.  My
patch changes the code to select the hit closest in time to the trigger.
(You still only get one TDC datum exposed.)

-- Brad

-- 
Brad Sawatzky, PhD <brads at jlab.org>  -<>-  Jefferson Lab / Hall C / C111
  Ph: 757-269-5947 -<>- Pager: 757-584-5947 -<>- Fax: 757-269-7848
The most exciting phrase to hear in science, the one that heralds new
  discoveries, is not "Eureka!" but "That's funny..."   -- Isaac Asimov