Hi, "J.H. Lee" wrote: > Dear Powel, Thank you for the information.For the tracking efficiency > calculations, how many planes were allowedto be missed (in X-Y and > U-V) for a track in a module? For T3 and T4 we allowed 1 missed hit in X , 1 in Y and 2 missed hits in V+U. For T5 we allowed 1 missed hit in X, 1 in Y and because second module in off 3 missed hits in V+U. > I assume they were notcalculated by inserting MC tracks into real > data. What were the reference trackscompared to for the calculations? For T5 we use BfsFrontTrack (T2-D3-T4), swim it through the D4 magnet and check whether is has a valid hit on Tof2, if so, we use it as a reference track. For T4 we start with a FfsTrack swim in through D3, D4 and check the valid hit on Tof2, and additionally we check whether it match to any of T5 local tracks (confirmation by T5), if so, we use it as a reference track. For T3 we use BfsTrack (T2-D3-T4-D4-T5), and if there is a valid hit on Tof2 we use it as a reference track. > How do ghost tracks play in the efficiencycalculations? I don't think, that we have a many ghosts using tracking parameters as mentioned above. To see it, please, look at the attached figures that shows a correlation between tracks slopes and positions for T2-T3 and T4-T5. The plots were obtained by combining all reconstructed tracks in one DC (od T2)detector with all reconstructed tracks in the another DC detector. In the efficiency calculation we require that a local track match to a reference track simultaneously in x, ax, y, ay (position in x, slop in x, position in y, slop in y), means in a four independent parameters. > The tracking efficiency for T3 at 12 degree for central > events(trigger6 ~ 28% central), which you said 70%, Now I have a better statistic (I have already reduce data files) and the T3 efficiency in run 5650 is 81% +-3% (statistical error) > is much better than I've guessed.If it's the case, do you think we can > use T3 as we initially planed to use for at12 degree, i.e. > reconstructing tracks in T3 instead of relying on external tracksto > find hit association? Is the low efficiency for T5 mainly due to > background orusing 2 modules? Yes, in fact a T5 performance per plane is now better than T4 performance. As you've probably heard, Radek is working on a new T3 tracking. The idea is as you mentioned above. We want to take a Ffs tracks (or just T2 tracks) and use them as a seed for initial tracks in T3. Radek already got a promising results. This method can be use to find a missing T3 tracks, It can be used for T5, as well. I attached a plot that shows a correlation between moment of BfsFront and BfsBack tracks for run 5650 (12 deg.). Pawel. > Too many questions. I'll stop here. JH > > ----- Original Message ----- > From:Pawel Staszel > To: brahms-l@bnl.gov > Sent: Wednesday, November 14, 2001 1:35 PM > Subject: Re: DC and Run Plan (Re: Run plan) > Dear J.H. Lee > The centrality dependence analysis for DCs has yet not been > done > so I can not answer most of your question right now. > The track reconstruction efficiencies for the latest runs at > 12 deg. for DCs > are T3: 70% T4:92% and T5: 70% (only 2 modules for T5 are > used). Is is for the > trigger #6 events. The efficiencies per plane in T4 and T5 > are about 90%. This value is not > corrected for a dead channels. > Lets us continue this discussion tomorrow. > Pawel. > > > "J.H. Lee" wrote: > > > ----- Original Message ----- > > From: "Zbigniew Majka" <ufmajka@cyf-kr.edu.pl> > > > To: <brahms-l@bnl.gov> > > Sent: Tuesday, November 13, 2001 12:47 PM > > Subject: Run plan > > > .... > > > Justification: > > > 1. All DCs are working and we have to try to use them > > > to obtain unique results. > > > .... > > > Zbigniew > > > > > > > Dear Zbigniew and Powel, > > > > It would be very nice if you could add some more > > information on "all DC's are working" for the people > > who are not so familiar with the details of DC > > performances > > to help decide what program we would focus on for the rest > > > > of the run. I understand that still a lot of work is > > required > > to answer some of the questions, but I think it would be > > very > > informative for all of us if you let us know as much as > > possible > > on those questions. Thank you in advance. > > > > 1. Efficiency > > - What are the plane-by-plane efficiencies (hardware > > efficiency) > > with the current HV and threshold settings? I assume the > > numbers > > are quite high (>90%?), otherwise we'll have serious > > difficulties > > to reconstruct tracks. > > - What are the centrality dependent track reconstruction > > efficiencies > > (software efficiency, if you like)? I understand those > > numbers are > > much harder to estimate, but we have to know them for any > > yield > > calculations as we all know. One of the ways to estimate > > the numbers > > is inserting MC tracks into real events or merging two > > reconstructed > > events and try to reconstruct the mixed events. > > > > 2. Track Reconstruction at 12 (and 20 and 3/4) degree > > - What are the mean number of the hits in the planes as a > > function of > > the centrality (Tile ADC Sum, BB Multiplicity, or Trigger > > 6)? > > - If you reconstruct tracks with the current CD > > configurations and > > reconstruction software/geometry/parameters, what are the > > mean numbers > > of tracks in T3, T4, and T5 as a function of the > > centrality? > > - What are the numbers for (T4+T5 matched tracks)/(T4 > > tracks) or > > (T4+T5 matched tracks)/(T5 tracks), also as a function of > > the centrality? > > > > 3. Extrapolating Tracks > > - How does a distribution of > > "position of T2 tracks projected on T3 - position of the > > nearest hit in T3", > > say for central (or trigger6) events at 12 (and 20, 3/4) > > degree? > > This will more or less tell us whether we can utilize T3 > > without > > recontructing > > tracks in it. > > > > Regards, JH > > -- > ---------------------------------------------------------------------- > | Pawel Staszel | > | Niels Bohr Institute Tb 8 email: staszel@nbi.dk | > | Blegdamsvej 17 phone: (+45) 35 32 53 51 | > | København Ø FAX: (+45) 35 32 50 16 | > | Danmark | > ---------------------------------------------------------------------- > > > -- ---------------------------------------------------------------------- | Pawel Staszel | | Niels Bohr Institute Tb 8 email: staszel@nbi.dk | | Blegdamsvej 17 phone: (+45) 35 32 53 51 | | København Ø FAX: (+45) 35 32 50 16 | | Danmark | ----------------------------------------------------------------------
This archive was generated by hypermail 2b30 : Thu Nov 15 2001 - 12:41:20 EST