Drift Tube Calibration and synchronization and Cosmic rays data

1
Drift Tube Calibration and synchronization and
Cosmic rays data
I report on - the parameters that define the
performance of the DT chambers and of their
trigger capability - the way we may extract
them assuming we have the drift time and the
trigger information. and on - preliminary
results from cosmic rays data on DT with
autotrigger slides 23-29
Preliminary
2
(No Transcript)
3
DT Local Trigger parameters introduction
Let a DT chamber in a space with a B(x,y,z)
magnetic field, working in standard conditions,
and a track with its time and direction passing
the chamber in a zone where its ionization
charges reach the wire with a mean drift velocity
vd,
The performance of the Local Trigger depends
essentially on two of the parameters which must
be set 1) the parameter ST i.e the mean drift
velocity set in each BTI of the minicrate and its
relation to the mean drift velocity
Vdrift of the chamber ST 12.5
d /Vdrift ns Tmax STdistance of staggered
wires in units of BTI frequency
clock(80MHz). Granularity of ST is 1/2 i.e.
Dtmax6.25 ns Tmax1.6 2) The relative
phase of the time of the track with respect to
the clock in the Mini Crate the alignement
is sampled at the clock frequency and its
efficiency depends on the phase of the clock
of the Mini Crate with respect to the time of the
track in the chamber. 3) Geometrical
parameters that define the relative correlation
between BTIs in the 2 PHI SLs .
4
What do we know about the dependence of the
efficiency Vdrift (ST parameter ) in
the Local Trigger ? From Data ( see slice
30) TB 2003 CMS in 2004/045 Local Trigger
efficiency does not change for Vdrift error of
1.7 TB 2000 (only BTI trigger i.e. trigger
only from one SL ) CMS NOTE 2001/052 ,NIM
1999 Vdrift systematic error 8 with
track angle 00 BTI H trigger efficiency 80
as expected with no error for 200 tracks ,
BTI H trigger efficiency drops at 68 (80
expected with no error) for 350 tracks BTI H
trigger efficiency drops at 30 (75 expected
with no error) From EMULATOR The behavior of
TB 2000 confirmed with old EMULATOR(zotto CMS
note 2001/052) The behavior of TB2003 confirmed
with new EMULATOR (Vanini talk and CMS in
p) Full dependence can be studied for any wanted
Vdrift error. CONCLUSION 1.7 accuracy is good
enough on Vdrift parameter for the efficiency
of the LT
5
vdrift dependence in the DT cell
Drift velocity is constant in the Ar/CO2 gas
mixture 85/15 at the nominal HV cell setting. The
apparent drift velocity varies only due to Lorenz
angle in presence of B field normal to the wire
and with the angle of incidence of the track.
The dependence of the drift velocity on angle
and B field has been measured TB2000 CMS NOTE
2001/041 and computed with Garfield cell
simulation Bologna CMS NOTE 99/064 and Madrid
CMS note in preparation . The simulation results
have been verified with the measured ones. At
trigger level we cant correct for the angle phi
but for high pt tracks it is well within 400,
neither for the magnet field but the expected B
field should be below 0.3 T everywhere except
only in MB1 stations in the wheel 2 and -2
where it may be of some relevance (up to .8 T).
6
Synchronization
There are two Synchronizations processa) inside
each chamber (sampling phase) b) between each
chamber and the Sector Collector (transmission
phase) M. Dalla Valle talk ,
HH efficiency in a given chamber as a function of
phase. For tracks completely out of
sampling_phase, the local trigger is output in
two contiguous bx and the trigger quality is
deteriorated.
7
Dependence of Trigger efficiency
Local Trigger Phase (Sampling Phase)
From TB 2001 data CMS NOTE 2001/51 Single SL
data depending on the phase of the sampling
clock the Htrig efficiency ranges from 70 to
80. From TB 2003 data CMS IN 2004/045 For
Configuration HHHLLL in the Traco, the Local
Trigger efficiency varies from 90 to 65 for
any phase From EMULATOR TB 2003 and 2001
behavior confirmed with EMULATOR . TB 2003 and
2001 behavior confirmed with EMULATOR . Full
dependence can be studied for any wanted error.
8
How best Sampling_phase for bunched particles
can be measured.
The procedure exists ( coded) , has been tested
with TB Data 2003 (CMS IN 2004/042) and with
TB2004 for 2 chambers. The algorithm is
luminosity independent .It is based on the fact
that for in phase tracks the output of correlated
triggers is always in the same bx , for out of
phase events is in two bx. The Mean Time RMS
distribution / or the HL/HH trigger quality
ratio as function of phase are used.
HHHLLL Trigger efficiency as a function of phase
efficiency vs sampling phase


HL/HH
MT RMS
Track- clolck phase (arbitrary offset )
TTCrx delay ns
9
Remarks on vdrift and on sampling_phase setting
in the Local Trigger a)The Local Trigger drift
velocity (ST parameter) could be set
individually in each BTI but drift velocity and
sampling_phase of the Local Trigger are
correlated and there is only 1 sampling phase
value for each Mini Crate i.e. for each DT
chamber so just one setting per chamber for the
phi SL is reasonable. Different values of vdrift
on Theta SL BTIs can be set since in theta the
angle of track is fix but compromise with
sampling_phase have to be found. b) The
apparent mean drift velocity in the CMS DT
chambers at LHC can be measured or can be
computed from simulation. The range of variation
inside a chamber is anyway 2 due to track
angle (TN 2001/041 , TN 2003/052) and that can
be correct for only on HLT and offline. The same
for the error in MB1/-2 due to the B field. c)
The signal propagation along the wire produces an
intrinsic phase error which ranges from 0 up 4.5
ns/meter of wire length (TN 2003/042) 9ns for
phi wires and 14 ns for theta MB3 wires.
d) The intrinsic phase error for cosmic rays on
just one chamber ranges from 0 to 25 ns.
10

• How do we check the value of the mean drift
  velocity with bunched particles (i.e. all events
  with the same phase )
• from track fit minimization with vdrift as free
  parameter CMS NOTE 2001/041,CMS NOTE
  2003/001,etc
• from Mean Time Time box (slide 31)

• Commissioning and magnet test will be performed
  with cosmic rays.
• Can we measure the Drift Velocity with
  autotriggered data on cosmic rays? with which
  accuracy ?
• From tests on cosmic rays and autotrigger
  performed in Legnaro even if the T0 ( necessary
  for a good drift velocity measurement) varies
  event by event from 0 up to 25ns, there are
  hints that , with some easy(?) software
  implementation, a T0 and so Tmax and vdrift
  accuracy of 3 at least can be set .
• Can we autotrigger on cosmic rays ? With which
  purity ?
• ( see second part of the talk, COSMIC RAYS with
  AUTOTRIGGER slides 23-29)

11
Commissioning
verify DT chamber performance and working point (
minimum noise, max efficiency , best drift time
linearity and resolution) the parameters are
HV and Threshold
Operations at commissioning step 0- with TP
signals, FE channels and Mini Crate connections
are checked ( DCS local DAQ ) step 1-
with random triggers (L1A) the noise is measured
as a function of HV and of Threshold up to
values a little above the working voltage ( DCS
local DAQ). step 2- The efficiency of one
SL as a function of HV and Threshold can be
checked - setting the others SLs at
nominal HV and Th, - Trigger L1A from the
autotrigger type H of the chamber () (BTI
SL under test masked. Only H trigger on the
other SLs. Best configuration ?? )
12
Operations at commissioning continue
- scan in HV and Threshold in the SLi ()
- Results from TDC data occupancy in
time window , timebox shape , Mean Time
shape , ( DCS local DAQ), - from track
reconstruction cell resolution , Drift Time /
position linearity and efficiency (?) -
with standard configuration on all BTIs, check
trigger efficiency of the Sli from the ratio of
HL/HH trigger quality -Quasi automatic algorithm
must be implemented ? NB is necessary a quasi
automatic HV scan procedure and analysis
semi automatic analysis programs have been
already prepared and used in the on line program
semi automatic
analysis programs not ready () trigger
operation not done yet , must bedefined in
Legnaro and at ISR
13
What do we know of Single SL Local Trigger
The BTI H trigger efficiency is 80 with
less then 5 wrong bx assignment. NIM A
438(1999) 302 and CMS NOTE 2001/051
Standard acceptance Minimal
acceptance
14

• For HV Threshold efficiency curve we may use
  the
• occupancy in time window
• Track reconstruction efficiency ( even with large
  error 1255mm600mm/hit

• Drift velocity can be measured at 3 (software
  to be implemented) and
• x(t,B0,angle) parameterization can be verified

• Work to be done before commissioning
• test the 1 SL autotrigger capability and define
  the best trigger configuration for that
• (people ? , hardware ?, software implementation
  ? in Legnaro and at ISR)
• define a standard sequence with DCS and DAQ and
  define quality cuts for commissioning likes
  production cuts

15
Synchronization in the cosmic challenge and
magnet tests

• GOALS ( as far as I understood but see Benvenuti
  talk)
• Cabling and signal distribution checks
• Study the cell performance with the real B field
• - Learn how to synchronize a sector as in
  TB2004 but without external scintillators.
  Particles are not bunched so the synchronization
  process is not between the particle generation (
  40MHz clock) and each chamber but between
  different chambers ( see E. Conti and Dalla Valle
  talks).

16
Synchronization in cosmic challenge and Magnet
test
Sequence of operation a) with TP -Check all
cabling compatibility and TTC signal distribution
timing with TP signals generated in the 4
Minicrates of the sector on a common LDAQ
signal. -Choose the trasmission_phase in order
to have 100 trasmission efficiency from the LT
in the chambers to the Sector Collector. ( How ?
) . b) Set each chamber at the HV and Th working
point defined at commissioning . c) Chose 1
chamber for HHHL autotrigger and set SC so that
it picks the autotrigger signal of the chosen
chamber and sends it back as L1A to all chambers
of the sector. d) Find in each chamber the
sampling_phase for local synchronization to the
triggered events in the other chambers with SC
L1A (see E.Conti). e) Take data with magnetic
field with SC L1A or DTTF L1A (see E.Conti).
17
How can we synchronize two or more chambers in
the cosmic test example
The color gives the bx of the autotrigger in each
chamber for a given time of the cosmic the y
axis gives the efficiency that any cosmic produce
a autotrigger (HH or HL or LL) in that chamber
passing at a the time set in the x axis. In the
example , lets a cosmic at t12 ns. The
probability that chamber MB3 has a autotrigger is
90 , and the trigger output is at bx 102 with
100 probability. The same cosmic in MB1 has a
probability of 70 to trigger and half a
probability to be in bx 101 and half at bx 102.
Setting in phase the chamber MB1 with respect
MB3 means to move the TTCrx delay ( in the
example of 12 ns ) so each chamber gives the same
bx
bx 103 in MB1
18
Can we measure the drift velocity in the magnet
test and check the cell and the trigger behavior
? Yes but with a accuracy do to T0 jitter of 3
better if event by event T0 corrections possible
wheel 2
wheel 1
wheel 0
19
LHC startup

• Operations at startup chambers already set at
  nominal HV and Threshold as from magnet and
  cosmic tests Local Trigger configuration 0
  loaded. The sampling_phase is set as a
  function of mean TOF in the chamber with
  respect to TP syncrhonization.
• Step 1- in technical runs with random triggers
  (L1A) the noise is measured ( ? as a function of
  HV and of Threshold up to the working voltage?
  DCS local DAQ operation).
• Step 2 - data are collected locally in
  autotrigger chamber by chamber ,
• from timebox and MT verify bx assignment and
  Drift velocity,
• Measure/verify sampling phase and adjust TTCrx
  fine delay in each chamber (TOFphase),
• adjust trasmission of any chamber to the SC ,
• DT LT information sent to DTTF for DTTF
  synchronization.

After step 1 we can start to be read out and to
sent LT segments at least with 70 efficiency
for correlated ones. In the worst situation,
being completely out of phase, the good output
trigger of each chambers may be in 2 contiguous
bx with 65 efficiency. The probability to send
at the same bx the 2 LT of 2 completely out of
phase chambers is so 24 but in a sector there
are 6 possible combination of 2 chambers. So we
should trigger with some reasonable efficiency.
From analysis of Drift Time data (MT in the LDAQ)
we can find if the bx assignment was wrong of 1
or -1 bx with respect an expected one . The fine
phase can be adjusted locally and the coarse
phase in different chambers of this doublet can
be checked and adjusted at SC or at SC level.
With a iterative process all chamber of a sector
can be syncrhonized
20
Conclusion DT CALIBRATION AND SYNCHRONIZATION
at LHC
LOCAL TRIGGER CALIBRATION we want to verify
the Drift Velocity i.e. to measure the mean drift
velocity in each chamber from drift time data
(slide31). A sample of 10000 events/chamber is
a good starting statistic the measure can be
performed locally, at LDAQ or/and at the HLT
farms level, the vdrift corrections with respect
the nominal values is expected to be 1.6
which is also the granularity on the hardware.
Local trigger SYNCHRONIZATION we have to
find the best phase of the considered tracks
with respect to the sampling clock. The best
phase is TOF dependent so it must be performed
with pp tracks. The best phase is vdrift
parameter dependent and so few values of vdrift
parameter should be checked. 5000-10000 events
/phase/chamber are enough with 25 steps in phase
of 1 ns. The measure can be done locally in the
LDAQ or/and at the HLT farms level .How can we
trigger in order to have data for checking Vdrift
and for synchronization?Even if chambers are not
synchronized , LT efficiency is above 65 for
any phase value.Once chambers are locally
synchronized , SC synchronization and DTTF
synchronization can be easily performed as in
TB2004 (Marcellini talk). full L1 efficiency
what is the impact of no calibration in the LT ?
We do not have data but in HLT code LT is
perfectly simulated. In HLT there is a perfect LT
synchronization ( mean TOF in a chamber set to 0
) but all drift time errors are included (cell
non linearity, wire propagation, TOF differences)
and there is just one LT Vdrift parameter for all
chambers so HLT L1 simulated results (A.Vanini
and S.Lacaprara talks) refer at a perfect
synchronization but at a minimum calibration.
21
Conclusion DT CALIBRATION AND SYNCHRONIZATION
In HLT and OFFLINE code , once a rough track
angle and position is computed in both
projections, the T0 incertitude for the
propagation along the wire is corrected for, the
B field and the position in the cell and the
angle are known and so x can be computed from
xx(t,B,angle) parameterization. That is what is
done in the current reconstruction code of the
simulated data. Offline calibration is so to
compute -the relative difference in time between
the channels of a chamber (with TP data , done
already at commissioning. The measured
differences are of the order of 1ns in all fully
equipped chambers as expected from cabling
accuracy), -the T0 of each chamber (from
residuals and/or from time box and MT) and
verify that -the parameterization is correct
from residuals and from t t (x) relation. The
other aspect is -alignment (see Matorras )
between chambers. For LHC this can be performed
with alignment data and with well locally
reconstructed pp tracks . -B field map
accuracy.The worst situation for no calibration
in offline other then B field and alignment-
can be a local DT reconstruction without
parameterization, with a uniform constant vdrift
velocity and T0 in each chamber, vdrift velocity
and T0 computed chamber by chamber directly
from data (that is what is done in the cosmic
data reconstruction ). One could have events with
global L1A shifted of 1 bx depending of the kind
of trigger (muon RPC, muon CSC, Muon DT , Ecal,
Hcal and in each one could have a sector or wheel
dependence). That is local (DTTF from sector to
sector for DT) and global trigger synchronization
that can be performed only checking the data
(timebox and MT) as a function of the different
trigger origin.
22
CONCLUSION DT Calibration and synchronization
the parameters are HV and Threshold defined at
commissioning but verification in B field and at
startup needed see commissioning slide.11
Chamber
the parameters are drift velocity (1/Chamber but
no strong variation expected if a fix value in
all detector is used) and sampling_phase
(1/Chamber must be measured with data ) see LT
introduction slide 4


Local Trigger
HLT and offline
(consideration only for DT) the parameters are T0
(1/Chamber), x x( tdrift , B, angle)
parameterization or vdrift(angle,B) and geometry
alignments. The channel by channel T0 variation
is very little by construction i.e. 1 ns, is
constant in time and is measured with TP. see
Amapane Lacaprara Matorras talks and CMS notes

• WORK to DO
• - Verify completeness of calibration and
  synchronization analysis and of foreseen
  processes.
• Prepare semiautomatic procedure interacting with
  trigger , DCS and DAQ people.
• Test procedure.

23
Cosmic rays autotriggered by a MB3 chamber in
Legnaro. CMS in 04/045
Cosmic rays set up
LNL MB3 chambers BTI and TRACO ? set quiet
configuration, the so called sync388 set-up i.e.
only very high quality trigger in the TRACO HH,
HL or a H certified by a H_trigger in the theta
view. Autotrigger SELECTION? only H in
CH2 Trigger autotrigger ? L1A of the Mini
Crate
24
Cosmic ray autotrigger with 1 chamber ,
With autotrigger, the gitter of the cosmic time
with respect to the L1A trigger was verified from
the time of the scintillator recorded with
respect to the autotrigger.
The time of the scintillators signals recorded
with respect to the autotrigger are 90 within 25
ns
25
Cosmic ray autotrigger with 1 chamber
Autotriggers Cosmic rays have flat distribution
in time but L1A stops the TDC always in phase
with the eadge of the clock any event has an
error do to the clock of about 25ns/(sqrt 12).
The range of the offset can be seen on the Mean
Time plot. Tmax distribution flat in 50 ns ( 64
tdc units).
But the offset is the same in all the drift times
recorded , event by event
From the correlation a 3 accuracy on Tmax should
be reachable (15ns/500ns ) , correcting To_ev
event by event from the plot above . Or with a
minimization fit ( event by event) of all the
residuals of the reconstructed tracks segments
with To_ev as parameter,
1 TDC count 0.78 ns
26
Cosmic ray autotrigger with chamber CH2
Autotrigger purity. Reconstructing offline the
track segment of the triggering chamber with the
standard Orca package (with no T0 correction
event by event), in 99 of the events a track in
both projections has been found in 0.5 of the
total events the chambers had many hits ( cosmic
shower ). The expected single point error is
around 500 mm as it should with a 12 ns time
error.
MB3 36 no alignments corrections fit performed
- in q with 3 or 4 layers -in j with at least
3 and up to 8 layers
angle distribution of the tracks reconstructed in
the event
Residual 7 point fit 400 mm error
layer 1 450 mm
j
layer 7 550 mm
q
27
Cosmic ray autotrigger with 1 chamber, drift
time data and autotrigger signal from chamber1
and chamber 2
Track segment Reconstruction with standard Orca
package (with no T0 correction event by event)
99 efficiency on the triggering chamber. In the
other chamber , due only to geometrical reason .(
taking in account the cell disconnected ) the
track reconstruction efficiency is arounf 96.
NB the T0 of the two chambers are shifted of
12ns the shift cannot due to TOF (not more
than 2 ns). It is due essentially to a the phase
of the sampling clock in each chamber with
respect to the muon., gtgt
Tbox of the 2 chambers
Occupacy in the 2 chambers
Residuals of the Position of the track
Reconstructed in ch1 extrapolated to ch2
xch2-xch1extrapolated cm
28
Relative autotrigger output time of the 2 chambers
50000 ev
The recorded autotrigger signal of CH2, the
autotriggering chamber, is a spike and 100
efficient. In 3 of the events there was a
double autotrigger but in the 5ms window of tdc,
compatible with the cosmic rate ( 2K Hz).
TDC units Time
The recorded autotrigger signal of CH1 presents
instead two spikes 25 ns apart (32 tdc units) ,
is 68 efficient and the shift in time with
respect to the CH2 autotrigger signal is 10 ns (
the two autotriggers are recorded in 2 channels
of the same TDC). That agrees with the 12 ns
shift of the time boxes and 2 ns of TOF . The
clock of CH1 is displaced of 10 ns with respect
to the clock of CH2 so it is completely out of
phase for the muons recorded by the CH1 chamber.
The experimental results agrees completely.
50000 ev
TDC units Time
29
Conclusions cosmic rays data studies with DT
autotrigger

• The test performed in Legnaro with the
  autotrigger signal as L1A, shows that with
  cosmic rays the Mini Crate can work and data give
  plain information as far as uniformity is
  concerned
• on the chamber performance
• on the trigger performance
• and
• on the role of phase in different chambers.

BACK UP SLIDES gtgtgtgtgt
30
CMS NOTE 2004/045 TB2003 bunched beam Local
Trigger HHHLLL Local Trigger efficiency (vs
phase) for 2 different vdrift in the BTI and 2
track angles
CMS NOTE 2001/052 TB2000 Only 1 SL BTI and
emulator H efficiency (within 6 ns window phase)
for different vdrift in the BTI vs track angle
0.9
Vdrift50mm/s.
Vdrift55.1mm/s
0.6
Angle with respect to normal to the chamber
degrees
NB The efficiency for tracks normal to the
chamber does not depend strongly on the assigned
drift velocity if the error is within 8. It is
relevant as soon as the tracks hits cells not in
the same column.
31
(No Transcript)
32
END
BACK UP SLIDES gtgtgtgtgt
33
(No Transcript)
34

35
Introduction to DT Local Triggr parameters
Glossary Configuration it contains the
parameters for the Read Out and for the Local
Trigger (masks, triggers parameters and
requirements) , see Marina Passaseo
talk. Autotrigger signal that is generated
under specific requirement on the Local Trigger
quality of a chamber. It can be used for the L1A
generation only locally or used by the Sector
Collector that receives the track segments from
4(5) chambers to generate the L1A for all the
chambers of the sector. (see Dalla Valle
talk). Nominal HV Nominal Threshold are the
working point of the chamber defined at TB at
ISR Standard configuration the configuration
foreseen for normal condition LHC
36
(No Transcript)