Note on UMC beamfile and Radial weight function
Note on UMC beamfile and Radial Weight function
Takeshi K. Komatsubara (KEK-JPNS), since 2002.Dec.06
0. Summary
- Use the newest beamfile for E949-02 (710MeV/c):
kstop_kmu21.02ps20.710mev.dat.test03
in ku9:~e949/koma/umc/umcbeam/scr_release.021029/
with the UMC v7 updated on
Dec 04, 2002
UMC simulation could start with K+ beam at the front of the degrader.
However, in the long history of E787,
the stopping distributions
(X,Y and Z stopping positions in the target) had never been
reproduced well by the simulation, presumably because
we have to have the parfect information of
beam momentum and divergence, change of them spill by spill,
materials in the beamline and etc.
We therefore start UMC simulation with the already-stopped K+'s
in the target.
A table of stopping positions,
so-called the UMC beamfile
is prepraed in advance by analyzing
real data in each year
and in each beam momentum setup.
The B4 hits, kaon-fiber hits and accidental hits in the target,
which are produced by the kaons coming to rest,
in the real data are also stored in the beamfile.
A KOFIA analysis code, developed by P.Meyers (Princeton) for E787 and
called "umcbeam", analyzes the Kmu2 muons in the km21 trigger
(because no extra partcle other than K+ and mu+ is in the event)
and produces the beamfile.
The code simply selects muons (without kinematic constraints)
at near normal incidence to the RS (|DIP|<10 degree)
and make a table of X, Y and Z stopping positions
by the B4-SWATH target analysis (with UTC track) in setup_kine.
Any trigger has a trigger bias;
the table of stopping positions from the km21 real data
should be corrected for, whenever a stopped K+ is generated at the target in UMC,
by a proper weight, which is "real function wght" in
/e949/src/umc/ustart.F.
To get the corresponding weight function to a new beamfile,
we generate Kmu2 decays in UMC with the beamfile (without imposing the weight at all,
by putting 'IBMMOD' -5/ in the .par file of the UMC generation)
select the event whose track satisfies |DIP|<10 degree, and then
calculate the km21 trigger acceptance as a fuction of radial distance
of the stopping X and Y positions from the center of the target
( RVTX == sqrt(XVTX**2+YVTX**2) ).
Please remind, the km21 trigger condition was changed from E787 (RS 19ct+20ct+21ct)
to E949 (RS 17ct+18ct+19ct).
In the figures of UMC below,
- Top Left: RVTX distribution after the dip cut, and before(solid) and after(brokeb) the km21 trigger condition.
- Bottom Left: Z (ZVTX) distribution after the dip cut, and before(solid) and after(broken) the km21 trigger condition.
- Top Right: ratio of solid and broken histograms in each RVTX bin.
- Bottom Right: ratio of solid and broken histograms in each ZVTX bin,
and the UMC simulation for E949-02 was made with T-counter thickness of 0.700 cm.
The km21 trigger acceptance as a fuction of ZVTX is not taken into accout to the UMC beamfile.
- Plot of the weight function for E787 (by P.Meyers, from the E787-95 beamfile)
- Plot of the weight function from E949-02 kmu21 beamfile.
- Plot of the weight function from E949-02 kmu21 beamfile;
in this case
for both umcbeam analysis and weight-fuction calcuation, the additional condition that
the track hits 19ct was imposed to make the conditions closer to E787.
The radial fuction is different for each.
The E949 UMC simulation until the UMV-v7 update on 2002.Dec.04 had been performed
with the beamfile of 2. (kstop_kmu21.02ps20.710mev.dat.test03) with the weight function of 1.
(for E787).
Shaomin compared the distributions of real data, UMC with 2. and UMC with 3;
he concluded that the 3. would be better.
We therefore set the weight function of 3. to the updated ustart.F in UMC v7
and will start the beam file of 3. (kstop_kmu21.02ps20.710mev.dat.test03).
- I wondered, why the UMC with 2. and with the beamfle of 2. does NOT work so well.
I suspected a comtamination of Kpi2 pions in the beamfile by E949's km21 trigger (RS 17ct+18ct+19ct), and
simulated Kpi2 decays with the km21 trigger and |DIP|<10 degree
(as the case of simulating Kmu2's).
What I found is, however, that the contamination of
Kpi2's is 1.17% by E787's trigger [19ct+20ct+21Ct],
1.25% by E949's trigger [17ct+18ct+19ct], and
0.58% by E949's trigger with 19ct;
the effect of contaminated pions
to the beamfile seems to be negligible.
- To understand the shape of the radial-weight function to each,
I made distributions of the opening angle between the radial vector
(0.0,0.0)-->(XVTX,YVTX) and the direction vector of the muon at the vertex (UB,VB)
before and after imposing the trigger condition.
- Plot from E787 beamfile
- Plot from E949-02 kmu21 beamfile.
- Plot from E949-02 kmu21 beamfile with the 19ct requirement
From Top Left to Bottom right, the opening-angle distributions for events with the radius
in [0.0,1.0], [1.0,2.0], .., and in [5.0,6.0] are presented.
For 2. the distributions after imposing the trigger condition are still flat.
For 1. and 3., the distributions of the R in [4.0,5.0] and in [5.0,6.0] have a "dip"
at around +1 and -1 (i.e., when the charged track goes to the opposite direction of the radial vector);
those events have a relatively long range in the target and would not reach RS Layer 19.
This effect introduces the km21 trigger bias (and the necessity of the radial-weight function).
- It is still a mystery why the UMC with 2. and with the beamfle of 2. does NOT work so well...
Last modified by T.K.Komatsubara(KEK), Mon Dec 16 13:22:26 JST 2002