Transcript MC Digi for Large Angle Vetoes

MC Digi for Large Angle Vetoes
V. Palladino, T. Spadaro
MC Digi for Large Angle Vetoes
Detailed PMT simulation needed to assess efficiency @ low-energy
Include and correctly treat:
• Gain fluctuations
• Optical g’s path fluctuations
• Signal generation
• Time over threshold FEE 8.0
Signal (mV)
N photo electrons
MC input parameters:
6.4
• Gain at operation point
G = 1.1 ×106
• 1st dynode collection eff. 4.8
e1 = 0.85
• Intra-dynode collection eff.
3.2
ed = 0.98
• 1st dynode time fluctuations
1.6
dT1 = 0.5 ns
• Intra-dynode time fluctuations
dTd = 0.8 ns
Large Angle Veto WG meeting – Mainz – 6/9/2011
4
3
2
1
Time (ns)
2
MC Digi: Data/MC agreement
Data/MC agreement
comparing MIP muons
and using test beam data
Further tests in progress
Data
MC
Integrated charge (pC)
Large Angle Veto WG meeting – Mainz – 6/9/2011
3
MC Digi: effect of cable
Cable effect correctly reproduced in standalone simulation
Whether or not it has to be inserted in official code is under scrutiny
MIP signal
Reminder, total cable length:
small LAV’s: 6.15 m, 7.15 m, intermediate LAV’s: 8.5 m, big LAV’s: >~10 m
Large Angle Veto WG meeting – Mainz – 6/9/2011
4
MC Digi: Data/MC agreement
Data/MC comparing test beam data to MIP muons, varying (nominal)
threshold for data
For MC use: 290 pF PMT capacitance, 10 mV threshold, 3 mV hysteresis
MC
Data
Large Angle Veto WG meeting – Mainz – 6/9/2011
Time over threshold (ns)
5
MC Digi: Data/MC agreement
For MC use: 290 pF PMT capacitance, 10 mV threshold, 3 mV hysteresis
Data/MC agreement: looking in detail, not really satisfying
MC
Data
Large Angle Veto WG meeting – Mainz – 6/9/2011
Time over threshold (ns)
6
Characterization of ToT curve: CPMT
For MC, study dependence of Q(ToT) on PMT capacitance
MC, CPMT = 200 pF
MC, CPMT = 300 pF
MC default, CPMT = 250 pF, Vthr = 10 mV, Vhyst = 3 mV
Below ~ 8pC, leading time contribution significant
For large signals, trailing time contribution dominates
Large Angle Veto WG meeting – Mainz – 6/9/2011
ToT (ns)
7
Characterization of ToT curve:Vthr
For MC, study dependence of Q(ToT) on FEE threshold voltage
MC,Vthr = 11 mV
MC,Vthr = 9 mV
MC default, CPMT = 250 pF, Vthr = 10 mV, Vhyst = 3 mV
Uncertainty on nominal threshold, ~ 2mV
Sizeable variation: dToT/dVthr ~ 3ns/mV, for low charges
Large Angle Veto WG meeting – Mainz – 6/9/2011
ToT (ns)
8
Characterization of ToT curve:Vhyst
For MC, study dependence of Q(ToT) on FEE hysteresis voltage
MC,Vhyst = 2 mV
MC,Vhyst = 4 mV
MC default, CPMT = 250 pF, Vthr = 10 mV, Vhyst = 3 mV
Hysteresis uncertainty ~ several % (to be confirmed)
Acting on trailing time only: dToT/dVhyst ~ 1 ns/mV for low Q
Large Angle Veto WG meeting – Mainz – 6/9/2011
ToT (ns)
9
Direct measurement of CPMT
Equivalent circuit for the direct measurement of CPMT
= 33Ω
Transmission line
Rx
PMT
PMT
V(t)
Divider
33 Ohm resistor introduced to correct for parasitic inductance
(ringing)
Measure signal V(t) for Rx = 50, 100, 200Ω
Fit the time constant t = CPMT (R + Rx)
Large Angle Veto WG meeting – Mainz – 6/9/2011
10
Direct measurement of CPMT
V(t) (mV/0.4 ns)
Rx = 50 Ohm
Fit with p0 e-t/t
time(s)
Rx (Ω)
Time constant t (ns)
CPMT (pF)
50
12.6(2)
152(2)
100
20.4(3)
153(2)
200
36.62(5)
157(2)
Large Angle Veto WG meeting – Mainz – 6/9/2011
11
Data/MC agreement: old...
For MC use: 157 pF PMT capacitance, take into account the 33 Ohm
resistor, 9.5 mV threshold, 2 mV hysteresis
Data/MC agreement: from the old situation....
MC
Data
Large Angle Veto WG meeting – Mainz – 6/9/2011
Time over threshold (ns)
12
... and present
For MC use: 157 pF PMT capacitance, take into account the 33 Ohm
resistor, 9.5 mV threshold, 2 mV hysteresis
Data/MC agreement: now much more satisfying
MC
Data
Large Angle Veto WG meeting – Mainz – 6/9/2011
Time over threshold (ns)
13
A first to-do list
ToT vs Q curve well reproduced for MIP’s
PMT capacitance value confirmed by direct measurement
Quantitative check satisfactory
Caveats:
agreement for extremely low values of ToT to be studied
Uncertainty on hysteresis to be determined
Cable modifications treated:
relevant for biggest LAV’s
inserted into digitization code
To-do list:
check with electron data
study of global LAV response
new data acquisition campaign for better assessment of data/MC
agreement to be tested with known threshold and hysteresis values
Large Angle Veto WG meeting – Mainz – 6/9/2011
14
Precise threshold measurement
Precisely assess threshold and hysteresis, mandatory for Q vs T reliability
Sub-mV total uncertainty needed
Tried with a standard approach, such as efficiency profile: measure efficiency
as a function of minimum signal voltage
Above approach would need clean environment: in presence of a 2-3 mV
radiofrequency noise, width of efficiency profile depends on noise
Try to overcome this, by measuring crossing times
Large Angle Veto WG meeting – Mainz – 6/9/2011
15
Accurate mmt of threshold
The idea is to register a signal copy and the LVDS output with a flash ADC
The LVDS transition time, TL, can be correlated with the time TS(Vth)
expected from signal, assuming a trial value Vth for the threshold
As Vth varies, the correct threshold value is found as the one for which the
time difference TL-TS is independent of the signal amplitude
For this to work, the signal time characteristics have to be preserved as
much as possible
Large Angle Veto WG meeting – Mainz – 6/9/2011
16
Setup for threshold mmt
Input to PMT from Hamamatsu C10196, ultrashort light pulser, 70 ps FWHM
Large Angle Veto WG meeting – Mainz – 6/9/2011
17
Setup for threshold mmt
Use LeCroy WaveSurfer 44Xs as 2.5 GS/s, 8-bit flash ADC
Signal from DC 50 Ohm
For the moment, an improper treatment of LVDS output: one polarity
to signal, the other grounded via 100 Ohm resistor
Large Angle Veto WG meeting – Mainz – 6/9/2011
18
Problems for LVDS signal
Due to this treatment, LVDS rise and fall times significantly worsened
time over threshold possibly affected: 10-90% time to 1-1.5 ns
but rise and fall time correlations expected to be maintained
LVDS signal (V)
Time (ns)
Large Angle Veto WG meeting – Mainz – 6/9/2011
19
Time correlation measurements
Leading time correlations as a function of trial threshold values
For the true value of the threshold, dT should be independent from Vmin
Vmin
Thanks to Paolo Valente for the
animated gif preparation
dT (leading) (ns)
Large Angle Veto WG meeting – Mainz – 6/9/2011
20
Time correlation measurements
Rise time correlations as a function of trial threshold values
For the true value of the threshold, dT should be independent from Vmin
Vmin
Thanks to Paolo Valente for the
animated gif preparation
dT (trailing) (ns)
Large Angle Veto WG meeting – Mainz – 6/9/2011
21
Time correlation measurements
Not exactly true... effect of overdrive-dependent delay at the comparator
must be subtracted for dT to be independent on Vmax
Reminder: for old (new) voltage at comparator is amplified x5 (x3) wrt original signal
Large Angle Veto WG meeting – Mainz – 6/9/2011
22
Correlation factor
Correlation factor well suited to evaluate “best” threshold value
For leading time, method sensitivity is on the order of 0.2-mV
Vle ~ -28.3 mV
Correlation factor for leading times
Vth (V)
Large Angle Veto WG meeting – Mainz – 6/9/2011
23
Identification of optimal threshold:
Correlation factor well suited to evaluate “best” threshold value
For trailing time, method sensitivity is on the order of 0.1-mV
Vth(trailing) ~ -30.7
Correlation factor for trailing times
Vth (V)
Large Angle Veto WG meeting – Mainz – 6/9/2011
24
Compare with efficiency profile
Evaluation compared with that from efficiency profile
e
Vmin
Large Angle Veto WG meeting – Mainz – 6/9/2011
25
Proposal for threshold
validation
Work shown able to clarify important points, but for en-mass mmts:
To proceed in a reasonable time, need more channels
TDC treatment of LVDS must be taken into account
One may think:
use a 32-channel, 5-GS/s, 8-bit flash-ADC with 1 Vpp at max
Include in acquisition together with TDC
Above approach would work for old FEE boards, in which each channel
has an independent analog copy in output
For new FEE’s, analog output are given for 4-fold and 16-fold sums only
Proposed solution:
pulse individual channels in a round-robin fashion
can acquire with few channels of flash-ADC
might exploit 4-ch, 12-bit, 2-GS/s V1729 digitizer, presently available
Large Angle Veto WG meeting – Mainz – 6/9/2011
26
Logical scheme for fast simulation
Particle generation and decay
Interactions in LAV volume
Energy release, Cerenkov
G4
Optical g’s: number, direction of emission, position
Transport to the PMT
Mimmo’s transport matrix
Number, energy of optical photons at PMT cathode
Photocathode emission
Dynode amplification
Signal generation
Transmission line
FEE: Threshold discrimination
This work
Leading and trailing times
Large Angle Veto WG meeting – Mainz – 6/9/2011
S1
MC Digi: effect of cable
Cable induces RLC(G) filter
C ~ 100 pF/m, L ~ 2.4 10-7 H/m
Silver copper steel, R ~ 86Ω/km
G varies withω, it is negligible for
ω< 150 MHz, where signal lies
x
x + dx
Attenuation
|F(ω)|
* cable data sheet, typical
values for 30 m length
ω (MHz)
Large Angle Veto WG meeting – Mainz – 6/9/2011
S2
MC Digi: effect of cable
Cable induces RLC(G) filter
C ~ 100 pF/m, L ~ 2.4 10-7 H/m
Silver copper steel, R ~ 0.086Ω/m
G varies withω, it is negligible for
ω< 150 MHz, where signal lies
|F(ω)|, MIP signal Fourier transform
ω (MHz)
Large Angle Veto WG meeting – Mainz – 6/9/2011
S3