Preliminary results on H8 test beam data Cosenza

1
Preliminary results on H8 test beam data Cosenza
Frascati Pavia Roma1 Roma3

• First results of summer 2001 H8 test beam data.
• all is preliminary, several systematic effects
  need (a probably long!) investigation but most
  results are known with good accuracy already
• many people made an important effort to get a
  quick feedback to chamber constructors on the
  chamber performance
• although most things are preliminary almost all
  seems to be as expected
• all tracking based results were obtained using
  CALIB, a c program developed by several
  off-line people from the Italian community. It
  will be the standard test beam analysis tool for
  Italian groups.

• Basic results of this presentation concerning
  (mostly BIL, BML under way)
• t0, t1, tmax drift characteristics (Ludovico!)
• r-t relation
• tube efficiency
• resolution HV scan threshold scan

2
Note in preparation
3
Drift properties
700 ns width
raw time (ns)
t0 ? start of physical distribution
t1 ? end of physical distribution
Not expected to carry any relevant information,
needed to equalize tdcs
tmax t1 t0 depends on many relevant drift
properties
4
tmax distribution_at_ nominal conditions
Tubes have all the same drift properties
rms of the distribution is 2 ns (compatible with
error in the determination of t0)
average value 665 ns
5
The tdc rise time depends on threshold
Zoom of the first part of the time distribution
and compare the rise time corresponding to
different threshold values
From here on all distributions are equal
20 ns
40 ns
6
tmax depends on HV threshold
Threshold scan
HV scan
t0
t1-t0
background
7
Display events
no way to include this hit in the track without
spoiling the C2. Smaller radius, d ray?
8
Get autocalibration using standard methods
Distribution is clean and well described by a
gaussian around 0
All layers, layer 2 3
Band ? 10 mm
1 mm
9
rt-relations
BIL Multilayer 1

Undistinguishable in this scale!
space (mm)?
BIL Multilayer 2
time (ns)?

space (mm)?
time (ns)?
10
Look at small differences
Pessimistic !
D rt(2)-rt(12)
D rt(1)-rt(12)
Pessimistic !
20 mm band
20 mm band
time (ns)?
time (ns)?
20 mm band
D rt(2)-rt(1)
Several geometrical parameters were found to be
imprecise Present results are expected to improve
Pessimistic !
time (ns)?
11
Studying the tube efficiency
Study the efficiency of this tube
Construct tracks crossing the tube under study
the hit of that tube may be absent. When present
quote hit to be inefficient if it contributes
significantly to the track C2 (at present
5ndof) (More complicated than that!).
12
The MDT tube efficiency
Good hits (efficient hits)
Hits totally absent 0.1/
Radius (mm)
Bad hits (high C2 hits)
C2 cut
Radius (mm)
Radius (mm)
Could be due to d rays, the corresponding radius
has to be smaller than track hit ? negative
residuals
13
MDT hit tube efficiency
96.5 efficiency ( a bit pessimistic)
14
Something which resambles alignment
Same run, independent processing in the 2 BIL
multilayers
nominal QBIL0.017 rad
ltagt-0.01000
Track reconstructed in BIL, ml 1 (rt-rel from ml1)
Da0.00025 25 mm in 10 cm

ltagt-0.01025
Track reconstructed in BIL, ml 2 (rt-rel from ml2)
15
And look at translations too
Same run, independent processing in the 2 BIL
multilayers
Track reconstructed in BIL, ml 1 (rt-rel from ml1)
Track reconstructed in BIL, ml 2 (rt-rel from ml2)
16
How to get the correct point resolution of one
MDT hit
If this hit was included in the track it would
pull the track itself and would bias the
resolution
This hit is not included in the track
drift time of excluded hit
Track parameters using remaining hits (a,b,Da,)?
d (distance of minimum approach) and
extrapolation error
Distribution of (d drift) convolutes
resolution and extrapolation error
17
Getting the resolution / 2

• the only way to get MDT resolution
• select clean events (one track only) good
  tracks (8 hits in BIL, no multiple association,
  acceptable C2)
• remove one MDT-hit at a time and compute new
  track parameters using remaining MDT-hits (say 7
  in BIL)

• compute distance of minimum approach of the
  straight line to the tube center of the remove
  MDT-hit (d) histo drift- d fit resulting
  distribution with gaussian distribution gt get
  sdrift- d(radius)

.15mm/sqrt(7)

• compute extrapolation error gt get
• ltextrapolation errorgt(radius)

signed radius (mm)
Resolution(radius) ? sdrift- d(radius)2-
ltextrapolation errorgt(radius)2
18
Fits at different radii
r-slice 1.5 mm
r-slice 6 mm
r-slice 11.5 mm
r-slice 13.5 mm
All distributions are quite clean, rms very close
to gaussian width. In the results which follow
distribution of residuals in different slices was
fitted with gaussflat level. Width of
distribution was taken to be s of gauss
distribution
19
Around r0.mm things are a bit messy
Tube wall region inefficient
6? track
0? track
Low e points not used to get resolution
20
The resolution on different layers
Run 2011 QBIL 6? Nominal conditions
21
The resolution we get at nominal conditions, run
2011
Run 2011 QBIL 6? Nominal conditions
Dotted line is the input resolution
Each point is the average of 4 layers
Continuous line is the fitted resolution
22
The resolution at another angle
Run 2022 QBIL 1? Nominal conditions
23
The two angles on the same plot
Resolution at 1 6 degrees
Garfield prediction _at_ 25 electrons threshold
24
Threshold scan
Garfield prediction at 25 electrons thr
38 mV ?13 e threshold 60 mV ?20 e threshold 98 mV
?33 e threshold
25
HV scan
Quick dirty method The beam setup allows a
simpler computation of space resolution. If track
is orthogonal to the chamber drift times of tube
in one layer is equal to the one of layer2
sr(r)1/?2 vd(r) st(r) sr(r) required space
resolution vd(r) drift velocity (as computed
from tube time spectrum) st(r) rms width of the
difference 2 time measurements
Quick dirty method compared with CALIB and
found to agree everywhere for at small and large
radii
26
HV scan quick dirty method
HV scan
(threshold)
Increase of the HV corresponds to an effective
lowering of the threshold and gives better
resolution mostly close to the wire
27
Conclusions

• About 10 M events were succesfully collected at
  the H8 test beam this september. Data were taken
  at many different conditions and will allow a
  campaign of study which will cover most aspects
  and functionalities of MDT system.
• Tracking based analysis is needed to validate the
  performances of MDT chambers electronics in
  quasi final configuration.
• First results obtained indicate
• uniform drift properties
• good tube efficiency (d rays)
• space resolution at nominal conditions as
  expected
• hv and threshold scan as expected

• Most of all above is based on CALIB (V3R2) a
  realistic c MDT reconstruction program
  developed by Italian groups.
• It cumulates few years of man work
  documentation in preparation
• can read many different types of inputs
  including MC-Geant4 DAQ-1 it handles most MDT
  problems
• is (always was) available for all groups.
  Documentation and help available as well
• Italian groups plan to use it as a standard
  analysis tool for (test beam) MDT data analysis.
• ) Domizia Orestano, Mauro Iodice, Toni
  Baroncelli, Paolo Bagnaia, Ludovico Pontecorvo,
  Nanda Pastore, Fabrizio Petrucci, Evelin Meoni