[G14_run] Questions about information of the initial photon

[Ken Livingston]

Hi Dao,
Here's the photon finding routine from one of the Glasgow KLambda analysis.
That's looping over all the good photons (stat = 7 or 15) and finding 
the one with the closest time to the proton.
It's in the rootbeer (http://nuclear.gla.ac.uk/~kl/rootbeer/) framework, 
but the concept should be fairly clear from the code.
That might be of some help. Some other's use GPID to do the photon / PID.

Cheers,
Ken



   //----------------Do best photon 
stuff----------------------------------------------------------------------------------
   /*Find the best photon by minimizing the time difference between the 
proton
     and photon vertex times. The proton vertex time comes from the TOF 
time,
     extrapolating back to the hadronic vertex. We use the TOF becuase 
it has a
     better resolution than the start counter*/

      for(int photon=0;photon<TAGR_NH;photon++){
        if((TAGR[photon].STAT!=15)&&(TAGR[photon].STAT!=7)) continue; 
//check good photon

        //Calculate the time difference
        
timeDif=(((SCPB[EVNT[ProtonID].SCstat-1].Time)-((SCPB[EVNT[ProtonID].SCstat-1].Path)/(vl*bcp)))-((TAGR[photon].TPHO)+((EVNT[ProtonID].Z)/vl))-0.667);
        //fill up timing plot for all photons
        tvertex_tgamma_allphotons->Fill(timeDif);

        //Fill up an initial photon 4-vector..
        lPhoton1->SetPxPyPzE(0,0,TAGR[photon].ERG,TAGR[photon].ERG);

        lKplus_CM = lab_to_cm(lPhoton,lTarget,lKplus);
        *v3_Kplus_CM =  lKplus_CM->Vect();
        //cosktcm = cos(lKplus_CM->Theta());

        /*Look at prompt and randoms for all photons..
          this will be used for photon random subtraction. These photons
          will then be used to scale the asymmetry distributions for
          PARA and PERP*/
        if(fabs(timeDif)<gPROMPT){                                    
//if in prompt peak
          tagERG_prompt[currentPlane]->Fill(TAGR[photon].ERG);        
//fill appropriate prompt hist
          
tagERG_theta_prompt[currentPlane]->Fill(cos(lKplus_CM->Theta()),TAGR[photon].ERG); 
//Fill 2d distbn
          promptHits++;                                            
//Count the prompt hits
        }
        else 
if((fabs(timeDif)>gRAND_MIN)&&(fabs(timeDif)<gRAND_MAX)){    //or if random
          tagERG_random[currentPlane]->Fill(TAGR[photon].ERG);    //fill 
appropriate rand hist
          
tagERG_theta_random[currentPlane]->Fill(cos(lKplus_CM->Theta()),TAGR[photon].ERG); 
//fill 2d distbn
        }
        promptMult->Fill(promptHits);

        //Minimize the difference between the proton and photon vertex times
        if(pow(timeDif,2)<pow(bestTime,2)){
          bestTime=timeDif;
          bestPhoton=photon;
          
KaonTimeDif=(((SCPB[EVNT[KPlusID].SCstat-1].Time)-((SCPB[EVNT[KPlusID].SCstat-1].Path)/(vl*bck)))-((TAGR[bestPhoton].TPHO)+((EVNT[KPlusID].Z)/vl))-0.667);
        }

      }


On 03/24/2012 02:22 PM, Dao Ho wrote:
> Dear all,
>
> Previously  I did some quick analysis of the K0 Lambda channel for g14. It
> turn out that I had a very low statistics result, so I naively assumed
> something was wrong with my particle selection conditions. Now, I check to
> see how much statistics I have rejected with each of my selection
> conditions. So far, the most important player in my selection conditions
> is the information about the initial photon (from the beam).
>
> I choose only events with None Zero photons in the RF bucket.
> I choose only events with the tagger ID (tagrid) being the same for all
> charges in those events. Consequently, I reject events with zero tagrID
> (found no photon matching).
>
>
> To my surprise, with "good" non-helicity events (at least 1 charged track
> events) about half of the time I see ZERO photon in the RF bucket. In
> other words, my selection removes 50% statistics because the events are
> recorded with zero photon in the RF bucket. So my question is how come
> this number too high? Can we do anything about it?
>
> My second finding is about events with zero tagrID (found no matching
> photons). This condition also removes about 10% of the "good" non-helicity
> events. Any comment?
>
>
>
> Here I would like to ask about the PID when cooking. My understanding is
> that we measure the TOF (flight time), the fight distance, and the
> curvature of a charge track. We use these to determine the beta and the
> momentum of the particle. From these quantities we can derive the mass of
> the particle (the so-call TOF mass). Will there be any chance that a
> particle with a TOF mass of electron (<5MeV) is really the pion?
> My concern is if our measure TOF is smaller than the real TOF by some
> nanoseconds, we would get a bigger beta, and a smaller mass. Is my concern
> a legitimate concern?
>
> This question is because in my selection scheme of +,+, -,- tracks. Half
> of the plus tracks has the mass of the positron, and the same thing
> happens with the negative tracks. Naturally, I suspect that there are
> pairs of e+/- created by photon, so I look for events with both e+/-. The
> number of events (4 tracks, 2+, 2-) with the electron/positron pair is
> about 10%. As a result, I find out there are a lot of events with only ONE
> positron/electron track.
> By the way, I consider the charged particle is electron if it TOF mass is
> less than 5MeV.
>
> Please let me know if you have suggestions or comments.
> Sincerely,
> Dao Ho
>
>
> _______________________________________________
> G14_run mailing list
> G14_run at jlab.org
> https://mailman.jlab.org/mailman/listinfo/g14_run