Transcript ECperformance

EC Performance Part 1
1997-2006
C. Smith, UVa
1. Cosmic runs
2. Pedestals
3. Attenuation lengths
4. PMT HV and gain
EC PRIMER
En  energy deposited in scintillator stack
Etot    En / f s
f s  sampling fraction [~ 0.27]
g n  ADC calibration  ~0.1 MeV/channel 
i ,o u ,v , w n
En  g n ( Ansig  Anped ) / exp(  xuvw / n )
Ansig  ADC channel
Anped  ADC pedestal
xuvw  distance from PMT to reconstructed hit
n  effective attenuation length of stack
w
u
v
inner
outer
Energy Calibration: EM shower vs. MIP
Sampling fraction vs. x,y
MIP energy vs. x,y
EM shower: Energy deposition non-uniform
function of position and depth. Difficult to
define calibration benchmark.
e-
μ
Minimum ionizing muon: Uniform and localizable
energy deposition profile (~2 MeV / cm ).
Cosmic ray runs
Integrated energy deposition
PMT Inner 5 strips
PMT Outer 8 strips
Light Guide
Light Guide
x-dependence
Light guide
PMT HV adjusted to achieve desired ADC calibration
(10 channels/MeV)
Linear fit of x-dependence of MIP energy deposition used
to obtain attenuation length and ADCmax
x
μ
μ
Uniformity of MIP response
Since Sept 1997…
• ~ 180 pedestal runs
– Accumulate 1000-2000 events
– PEDMAN used to analyze
– Data stored in $CLON_PARMS/pedman/Tfiles
• ~ 120 cosmic ray runs
– 50-100 million triggers at 2 kHz (need at least 12 hours)
– Only single pixel events retained
– No event recorder – data histogrammed directly from ET
• EC PMT HV adjusted ~ 20 times
– Usually just prior to each run period if needed
Time Dependence of Pedestals
http://www.jlab.org/Hall-B – Forward Calorimeters
ADC swaps
http://www.jlab.org/Hall-B – Forward Calorimeters
Time Dependence of Pedestals
http://www.jlab.org/Hall-B – Forward Calorimeters
Time Dependence of Pedestal Noise
http://www.jlab.org/Hall-B – Forward Calorimeters
Time Dependence of Pedestal Noise
http://www.jlab.org/Hall-B – Forward Calorimeters
Hardware
• EC HV mainframe replacement rate excessive
– Frequent trips during running (radiation ?, cooling ?, firmware?)
• Several ADCs have noisy pedestals
–
–
–
–
Increases event size
Affects accuracy of cosmic or MIP calibration
Note MIP is only 100 channels above pedestal
Future cable routing should be less vulnerable to disturbance and
more securely bundled.
• PMT replacement rate ~ 6 yr -1
– Mostly resistor and FETs in HV divider
– Photonis still carries XP2262 in catalog
Time Dependence of Attentuation Lengths
Sector 2
Sector 3
Time Dependence of Attenuation Lengths
Time Dependence of Attenuation Lengths
Time Dependence of Attenuation Lengths
Attentuation Lengths – Summary and Observation
• Attenuation lengths have decreased 10-20 % in sectors 1-4
over last 10 years of running.
• About ½ this effect seen in sectors 5-6.
• Origin ?
– Possible scintillator surface degradation due to weight of
overlaying layers and lead minimized in vertical CLAS
configuration. Effect of maximal loading (90 g/cm2) studied
using prototype stacks - < 5% effect.
– Estimated 10 yr radiation dose ~ 100 Gy. Prototype studies with
60Co source showed 16 % decrease in attenuation length after
160 Gy dose.
History of PMT HV Changes – Sectors 1-6
•
HV were adjusted as
needed after cosmic runs to
maintain constant PMT gain
•
Run index covers Sept.
1997-Mar. 2004
•
During this period HV have
increased 3-10%
•
Some dependence on theta
seen for U inner strips (PMT
1-36)
History of PMT HV Changes – Sector 1
HV have not been changed
since Mar 2004
Time Dependence of PMT HV – Sector 1
Time Dependence of PMT HV – Sector 2
Time Dependence of PMT HV – Sector 3
Time Dependence of PMT HV – Sector 4
Time Dependence of PMT HV – Sector 5
Time Dependence of PMT HV – Sector 6
Time Dependence of PMT Gains – Sector 3
log G  n log V
G / G  V / V 
n
Empirically for 12-stage XP2262: n = 9-10
PMT Gains – Summary and Observation
• Strongest gain decrease (up to 15% / year) seen for
short (small angle) U inner PMTs
• Larger gain decrease for inner vs. outer PMTs
• Smaller overall gain decrease for sectors 5 and 6
Question: Is 50% gain decrease over 6 years reasonable ?
From Phillips PMT Handbook: 50% decrease in anode sensitivity
expected after 300-1000 C charge collected.
300 C / 6 years / 0.5 data taking / 3x107 s y-1 ~ 3.3 μA
This current is equivalent to energy deposition rate ~ 500 kHz MeV-1
PMT Gains – Recommendations
• Need to independently confirm PMT gain decrease
• Check if implied energy deposition rate is consistent with soft photon
and hadronic rates
• Make in-situ measurement of PMT anode current during electron
beam. Note 3 uA is still only 1% of anode string current so HV
power supply current monitor may not reveal anything.
• Restore operation of laser system and use to monitor further gain
decrease. Comparison to laser runs of 4 years ago may yield some
information.