BTOW/BSMD Geometry
Jan Balewski has found an error in the BEMC geometry in STAR.
This effects both data reconstruction AND simulations. Here is his
link on the subject:
http://drupal.star.bnl.gov/STAR/2008/25
Oleg Tsai, John Scheblein and I have investigated this problem
using assembly, construction drawings and comparisons to
previous measurements made during construction. This took a
while because we found discrepancies of 8-20 mm for the radius
of the BEMC among different drawings. These had to be resolved
before I could make a final report.
The BEMC was originally designed to span the eta interval 0-0.99
At a later time, but before construction, it was decided to leave
room at the eta = +/-1 ends to allow for BSMD
electronics inside the magnet pole tips. This space was obtained
by reducing the last eta ring (of 20 in each module) by 38 mm.
This decision was made to avoid changing the integration
envelope of the BEMC, especially wrt to the EEMC which was also
under construction at the time.
There are several other problems with BTOW/BSMD/BPRS geometry:
1) the space between the modules at Z~0 was not modeled
correctly. There is a stainless steel plate (0.477 cm)
and an air gap (0.215 cm) that have been omitted.
2) the stainless steel plate (thickness = 0.79375 cm) at eta~1 has been omitted
3) BSMD electronics (equivalent material 0.5 cm G10 and 0.03 cm Cu) at
eta ~1 have been omitted.
First lets examine the actual dimensions of the detector as
determined from drawings and available parts:
Oleg Tsai, John Scheblein and I have investigated this problem
using assembly, construction drawings and previous measurements
made during construction. We have confirmed the drawings by
comparing previous measurements with the drawing: specifically,
the thickness of the scintillator (0.50 cm) and inner faceplate (1.91
cm). These measurements are accurate to +/-0.05 cm.
Here are our conclusions and a summary of our measurements:
From Drawing EMC051-D-1:
Radial Distance to Bottom Megatile R = 225.467 cm
Thickness of Megatile DR = 0.60 cm
Radial Distance to Center R+DR/2. = 225.767 cm
Distance to Z= 0 Z = 259.745 cm
eta = AsinH(Z/(R+DR)) = ASinH(259.745/(225.967)
eta = 0.9839
We did a separate check of the 9th megatile:
From Assembly Drawing: 341E1A-2K.dwg
http://www.physics.ucla.edu/~trent/BOTTOM-MEGATILE.pdf
http://www.physics.ucla.edu/~trent/MEGA-1.pdf
http://www.physics.ucla.edu/~trent/MEGA-2.pdf
http://www.physics.ucla.edu/~trent/MEGA-END.pdf
Note: there is a discrepancy of 8 mm in these drawings
from the radius to the front plate in the previous
drawing and the BEMC TDR.
Radial Distance to 9th Megatile R
R_0 to Front Plate (TDR) R_0 = 223.5 cm
Thickness of Z=0 SS plate: T_0 = 0.477 cm
Thickness of Z=0 airgap T_A = 0.215 cm
Z=0 to Scintillator: DeltaZ_0 = 0.692 cm (SS plate+air)
Radial Thickness of SS Front Plate: DeltaR = 1.91 cm
Radial Thickness of Scintillator DeltaR_s = 0.50 cm
R = R0 + R_stack + Scintillator Thickness/2
R = 223.50 + 12.65 + 0.50/2 = 236.40 cm
Z = 271.73 cm
eta = ArcSinH(Z/R) = ArcSinH(271.73/236.40)
eta = 0.9832
These two different estimates are in good agreement.
In StEmcGeom.cxx, the eta of a given tower is calculated as
the default, ie, (n_eta+0.5)*0.05. The Z coordinate of
the hits is calculated from the nominal ETA value.
Since the spatial resolution of hits in the towers is on
the order of 3-4 cm, we do not need to change this algorithm
for data analysis. The ETA centroid of the last tower changes
from 0.975 to 0.967. This is also smaller than the geometric
error we make by calculating Z_detector by assuming a
common radius for the shower maximum.
______________________________________________________
IF the space near Z=0 was neglected, then it appears all the towers
are shifted outwards in |Z(eta)| by 0.692 cm. It is worth mentioning
that WeiJiang Dong found in his analysis of West Barrel data
(2003) a 5 mm discrepancy between extrapolated TPC tracks
and positions in the SMD. (see pp 101-102 of his thesis):
http://drupal.star.bnl.gov/STAR/theses/ph-d/weijiang-dong
It is unclear from his wording which direction this is in
and this data only covers the West Barrel, but it is of
the correct magnitude.
I should note these comparisons with STAR geometry:
Drawing EMC051-D-1 + TDR: Rmin = 225.467 cm
GSTAR Rmin = 225.405 cm
_______________________________________________________
Checking the positions of the BSMD_eta strips to see if they
cover the same eta range, we measured the
spare boards still available at UCLA: Here is a
calculation of the outer edge of the last eta strip:
(eta<0.5) Strip Pitch = 1.537 cm
(eta<0.5) Strip Width = 1.445 cm
(eta>0.5) Strip Pitch = 1.9613 cm
(eta>0.5) Strip Width = 1.88 cm
Circuit Board Edge to Strip 1 = 1.524 cm
Circuit Board Edge of Strip 76 = 116.7 cm
Circuit Board Edge to Strip 150= 263.80 cm
Calculate:
z_1 = 1.524 + 0.692 = 2.216 cm (inner edge)
z_1c = 1.524 + 0.692 - 1.445/2 = 2.939 cm (center)
z_76c = 116.7 + 0.692 + 1.88/2. = 118.33 cm (center)
z_150 = 263.80 +0.692 = 264.49 cm (outer edge)
Comparing with calculations of StEmcGeom.cxx:
Drawings StEmcGeom Drawing-StEmcGeom
z_1c 2.939 2.689 +0.25
z_76c 118.33 118.146 +0.18
z_150 264.49 264.185 +0.31
There is good agreement between the geometry of the SMDeta
as calculated in StEmcGeom and our detector measurements.
R_smde= 230.705 cm (radius of pad plane)
eta_150 = ASinH(264.476/230.705)
eta_150 = 0.9812
Which is in good agreement with the previous estimates
for the edge of the scintillator towers.
It should be noted that the effective radius of the
BSMDs are ~1 cm different than the positions of the
pad readout planes. Since the ionization occurs
near the wire and drifts in a perpendicular direction
to the pads, we must expect an error ~1 cm in the
position determined by the SMD and TPC at high eta.
Recommendation:
STAR geometry should be fixed. Specifically
0) The radius of the BEMC is in dispute by +/-8 mm. This
should first be resolved by survey. IF it is found to
be R=223.5 mm (and the Z positions are verified):
1) It is NOT necessary to correct the z positions of the BTOW/BSMD in
StEmcGeom.cxx for data analysis. The calculated positions are
within the detector measurement errors of +/-3 mm.
2) Make the following changes to GSTAR:
a) Create a central gap for two two materials
SS plate (0<Z<0.477)
Air Gap (0.477<Z<0.692)
b) Change lengths of BTOW/BSMD (smaller by ~6-7 cm)
c) Add SS plate (0.79375 cm Stainless Steel) at eta~1 (angle = 40.395 degrees)
d) Add BSMD electronics at eta>1 (0.5 cm G10 + 0.03 cm Cu)
e) Add stainless steel plates in phi space between modules (CRACKWD=0.655).
Stephen Trentalange
Oleg Tsai
______________________________________
Update on August 14 2008
Here is an outline of the changes that need to be made to
simulation and analysis software:
The geometry software to be modified is contained
(mostly) in pams/geometry/calbgeo/calbgeo2.g This should be
updated to calbgeo3.g Here is a summary of the necessary
changes:
Changes to existing volumes:
1) calb_EtaCut = 1.00 or 0.99 needs to be changed to 0.9835
This variable is NOT defined in file calbgeo2.g
2) The z_min of each scintillator/lead layer/smd, etc needs to
be changed from 0 to 0.692
Block CBTW
Block CSUP (CSCI,CPBP,CSCI)
Block CSMD (CSMG,CSDA,CSMB,CSMC,CSME,CSHI
Additional volumes to be added:
1) Stainless Steel Plate z={0.0,0.477} Rmin,Rmax
2) Air Gap z=(0.477,0.692) Rmin,Rmax
3) Stainless Steel Plate eta~1
- trent's blog
- Login or register to post comments