Hi all, please ignore the message I sent yesterday. I found problems in
the algorithms used when looking at the data today. The conclusions
yesterday are mostly wrong. Sorry
Chi
On Fri, 29 Nov 2002, Chi Zhang wrote:
>
> Hi,
>
> I made a class TBLDetFitEL to fit some of the field off data taken
> during the last few months. I guess this would be my topic on the analysis
> meeting but since I changed BlastLib once again, I want to keep everyone
> updated.
>
> I did some work on runs 2723-2727 which as I know were with field off.
> (Arron and I took them). Few interesting things:
>
> 1st: Tong discovered from graphically display WC data that a lot of tracks
> came from down stream. my ananlysis confirmed it. I recover a lot of
> events that can be fitted into a elastic scattering from Z0>22cm, i.e.,
> off stream. a lot of them gather at Z0=30-40cm.
>
> They can be fitted to a hadron side mass of 4-5 GeV. the fitted hadron
> mass spectra display peaks at this region, a cut of Z0<22cm get rid of
> them with almost no residual.
>
> They display a huge left right aymmetry, Sum(electron sector
> number)/number of events = 0.32, beam is scrapping something on one side?
>
> 2nd: The fitted hadron mass spectra has a peak at M=0.45GeV. A cut of
> hadron side average tdc - electron average tdc > 5 nanosec, get right of
> this peak with almost no residual.
>
> So these events correspond to 'fast fast' coincidence I guess.
>
> They display a huge left right asymmetry too. Sum(electron sector
> number)/number of events = 0.70, more right sector than the left.
>
> 3rd: a cut of:
> -22<Z0<22, (phi_e-phi_p)^2<0.1, average proton tdc - average
> electron tdc > 5 nano second, left me with proton mass of
> 1.2+-0.4.
>
> there is only 1% left right asymmetry in these events:
> Sum(electron sector number)/number of events = 0.506.
>
> I am afraid this is the best resolution I can get from scintillator
> signals. there are 1-2 deg out of 20 deg in electron angle, a few degrees
> out of 70-80 deg on proton side. Z0 resolution is 6-7 cm.
>
> 4th: sum up the numbers:
>
> run 2723-2727 contains: 530917 physics events, 477241 of them I would call
> clean events in the sense that one tof on each sector fired, one or zero
> CC on each sector fired.
>
> 39117 of them can be fit into a elastic scattering model. The fit is done
> in the following way. if I assume the one with smaller tdc is electron,
> assum the particles hit the center of TOF scintillators in X and Z
> directions(see blast.geom), with y calculated from tdc position times
> 50pecosec/channel and 0.0147cm/pecosec. I can then for any position Z0
> along target cell a theta_e, and a theta_p. Then theta_p2 can be
> calculated from theta_e. vary Z0 to find a point where theta_p calculated
> from theta_e is equal to theta_p calculated from geometry.
>
> among them:
> 30905 are close to coplain: (phi_e - phi_p)^2<0.1 rad^2. (pretty loose)
> among these 30905:
> I can fit 11393 to target region: -22<Z0<22; a big loss.
> among these 11393:
> 8867 of them survived the cut: average proton tdc-average electron tdc>5
> nano sec
> among these 8850 have: theta_e< 30 degrees
>
> To the end, I exhausted my wisdom on inventing new cuts. The resulting
> proton mass reconstruction is: 1.2+-0.4 GeV.
>
> The plan is to go on to field on case. But I am not sure how to do it
> yet.
>
> If you are interested in this stuff, please see my directory:
> spud?:~blast/commis/phase1/ANALYSIS/CHI/TestFit.C and DetFit.root
> and the source code in BlastLib2.
>
> The codes are checked into CVS, tagged v2_8.
>
> Again, they are not tested with any online analysis code since test the
> code one macro and another would take a whole day. no copy is made in
> lib/spud.
>
> I think I found the problem that crippled ntuple.C on Monday. It was a
> broken dependence between TBLCut_coda and TBLDetRecon. I think it is a bad
> idea to keep these two classes together. I added book keeping functions
> into TBLDetRecon so TBLCut_coda shall become obsolete. However, I did not
> change TBLCut_coda and the file structure of BlastLib2 anyway.
>
> If you want to compile the library for spuds, first revert to earlier
> versions with:
> cvs up -rv2_3 or 4 or 5
>
>
> Chi
>
>
>
>
>
>
>
>
>
>
>
This archive was generated by hypermail 2.1.2 : Mon Feb 24 2014 - 14:07:29 EST