[d2n-analysis-talk] T2 timing corrections

[Brad Sawatzky]

On Mon, 29 Mar 2010, MATTHEW R POSIK wrote:

> I think my code is plotting the correct thing, I think I am just
> poorly explaining what I am doing, so let me try this again. So I get
> the value of Ndata.DBB.t2 from the replayed TTree, in the case of run
> 2060 Ndata.DBB.t2 = 2.
>    Then to plot the T2 trigger, for each event I loop over the hits
>    this is done using the following for loop:
> 
>        for (Int_t hit =0; hit<Ndata.DBB.t2 ; hit++)
>          {
>              t2_histogram->Fill( DBB.t2[hit] );
>          }
> 
> So is this same thing as DBB.t2[Ndata.DBB.t2-1] , since I have T2[ ]
> running from 0 to Ndata.DBB.t2-1, ie( T2[0] and T2[1] ) ?

Not quite... reread the previous email.  If you want to plot the hit
farthest from the common stop in time (ie. the first hit the TDC sees
after the interaction in the target) then you want to plot the *last*
hit in your TDC array, not the first.

>   I also looked at the shoulder events in TDC 04 from run 2060 (
> http://jlab.org/~posik/d2n/BB/trigger/t2_walk/2060_tdc04.png )
> 
>  I plotted the re-timed TDC 04 (Red histo) and the not retimed TDC 04 (blue
> histo) on the same plot in log scale. To make these plots I used the
> following cut on each TDC:
> 
>                   "BB.tr.n==1&&(DBB.evtypebits&1<<2)==(1<<2)"
> 
> I then redid the same plot as above, however this time I did not use the
> BB.tr.n==1 cut So the cut on each TDC is now
> 
>                     "(DBB.evtypebits&1<<2)==(1<<2)"
> 
> You can now clearly see a well defined shoulder in both the re-timed TDC 04
> (Red histo) and the not re-timed TDC 04 (blue histo). The plot is here:

Interesting.

> http://jlab.org/~posik/d2n/BB/trigger/t2_walk/2060_tdc4_1.png
> 
> So it seems that the one track cut greatly reduces this shoulder, but
> that may just be because that cut right now is throwing out a lot of
> statistics.  I also compared the TDC 04 vs the T2 trigger to the
> re-timed TDC 04 vs the T2 trigger, this can be seen here:
> 
> http://jlab.org/~posik/d2n/BB/trigger/t2_walk/2060_tdc4_t2.png

Also a good plot.

> The re-timed TDC looks good against the T2 trigger (right plot),
> except there seems to be a few events that fall at the two horizontal
> bands at DBB.t2 = 560 and DBB.t2 = 565 that do not fall in line with
> the vertical peak at TDC channel ~260. It is these events that
> contribute to the shoulder events in the TDC histogram. I did verify
> this by plotting the re- timed TDC 04 and making a cut on the T2
> trigger, this can be seen here:
> 
> http://jlab.org/~posik/d2n/BB/trigger/t2_walk/2060_tdc04_t2_shoulder_ct.png
> 
> With that said, I still do not know why these events do not seem to shift
> with the T2 re-timing of the TDC. Could they some how be correlated with the
> T1, since the T2 trigger is formed from the shower and cerenkov and the
> shower had two triggeres, the T1 and T6?

Hmm... that's a good thought.  If that's what's happening you should see
it in a plot like this, but with T1.
  http://jlab.org/~posik/d2n/BB/trigger/t2_walk/2060_tdc4_t2.png

> Also I wanted to check where the structure in the T2 at chan. < 400 were
> falling in the TDC.
> 
> http://jlab.org/~posik/d2n/BB/trigger/t2_walk/2060_tdc4_t2selection.png
> 
> So I plotted for run 2060 the re-timed TDC 04 with T2<400 (red histo,
> scaled to the blue histo) and the re-timed TDC 04 with T2>400 (blue
> histo).  From this is seems that what ever is generating the T2
> trigger as chans<400 are events that do not fall into the timing peak
> of the TDC.

Remember that if you cutting on T2 outside of its self-timing peak means
(by definition) that some other trigger is carrying the timing.  A
significant fraction of the T6 triggers are good electrons (and would
have generated a T2 but for prescaling).  Events with a T2 well off its
self timing peak means that the system accepted a trigger pretty far
away in time from whatever fired the Cerenkov component of the T2
trigger -- ie. should be dominated by background.

Just FYI, you can get some weird apparent structure when plotting the
DBB.xxxTnn[0] hit for high rate data.  Things might look more reasonable
if you plot DBB.xxxTnn[Ndata.DBB.xxxTnn-1] instead.

> I am going to take a look at the trigger diagram and see if I can
> figure anything out.

Very good.  All through this email I just about started typing a
suggestion that you do X, Y or Z only to find in the next paragraph that
you were one step ahead of me -- excellent.

-- Brad

-- 
Brad Sawatzky, PhD <brads at jlab.org>  -<>-  Jefferson Lab / Hall C / C111
Ph: 757-269-5947  -<>-  Fax: 757-269-5235  -<>- Pager: brads-page at jlab.org
The most exciting phrase to hear in science, the one that heralds new
  discoveries, is not "Eureka!" but "That's funny..."   -- Isaac Asimov