Collection Of Plots for A Testbeam Paper. List of Possible Plots R/Phi resolution, charge sharing, noise etc. Noise performance and few Landau distributions.

Collection Of Plots for A

Testbeam Paper

List of Possible Plots

R/Phi resolution, charge sharing, noise etc.

Noise performance and few Landau distributions Testpulse MP/Irradiation fluence vs position MP/Irradiation fluence vs position for different bias voltages For full irradiated area, MP vs HV to extract full depletion voltage Detection efficiency at certain threshold. Charge sharing comparison at full vs at none, and transition region Resolution comparison at full vs at none, and transition region. Ballistic deficit with one pitch bin.

Jianchun Wang 2

R/Phi sensor

Jianchun Wang 3

10/19/09 Jianchun Wang 4

Charge Sharing (I)

Seed threshold 5.4 Ke

Side threshold 2.7 Ke

Strip pitch (40, 50) mm

Nstrip = 1

Nstrip = 2

Nstrip = 3

R sensor of R/ f pair

Range: angle0.5

Cluster SizeP

erce

ntag

geSide threshold ~ 2 × noise


Charge Sharing (II)

Pitch (mm)

40 – 5050 – 6060 – 7070 – 8080 – 90

90 – 100

R/f data is split into 1 of angle & 10 mm of pitch sub-samples.

Sub-samples of 0, 3, 7 and 11 are with reasonable large statistics.

Angle ( )-0.5 – 0.52.5 – 3.56.5 – 7.5

10.5 – 11.5



Charge Sharing With Different Thresholds (I)

Jianchun Wang 6

pitch (40, 50) mm, angle (–0.5, 0.5)

Nstrip = 1

Nstrip = 2

Nstrip = 3

Angle ( )-0.5 – 0.52.5 – 3.56.5 – 7.5

10.5 – 11.5

Seed threshold = 4 ADC

Side threshold = 2 ADCApproximate conversion for R/F

22.5 Ke / 15 ADC = 1.5 Ke/ADC

Charge Sharing With Different Thresholds (II)

Jianchun Wang 7

pitch (40, 50) mm, angle (–0.5, 0.5)

Nstrip = 1 Nstrip = 2

Nstrip = 3

Angle ( )-0.5 – 0.52.5 – 3.56.5 – 7.5

10.5 – 11.5


Side threshold = 3 ADCApproximate conversion for R/F

22.5 Ke / 15 ADC = 1.5 Ke/ADC


The Eta Curve Eta curve plot – the relationship

between charge sharing and track hit position.

It can be generated in two ways.

Method one (Thanks to suggestion from Jan Buytaert):

1) Find track projected hit position on VELO plane.

2) Find the two adjacent strips between the centers of which that the track hits.

3) Calculated charge sharing before applying threshold.

Method two (useful in hit position reconstruction):

1) Applying thresholds and form clusters.

2) Select two- or more-strip clusters that matched with track.

3) Calculate charge sharing.

Tra

ck H

it F

ract

ion

Only Strip N has Charge

Cluster Fraction =

Only Strip N+1 has Charge

( )

i

ADC i N

ADC

Center of Strip N

Center of Strip N+1

Eta Curve Correction


Pitch = 40 – 50 mmAngle = (-0.5, 0.5)Nstrip = 2 only

Fit eta profile and correct RVELO measurement.

Tra

ck H

it F

ract

ion

Cluster Fraction

ProfileFit to pol3

s = 11.7

s = 10.1

RVELO – Rtrack ( mm )


Resolution vs Pitch

R sensor of R/ f pair

Preliminary !

Angle ( )- 0.5 – 0.52.5 – 3.56.5 – 7.5

10.5 – 11.5

Tracking precision is removed from the hit resolution.

Tracking precision is determined at each point (~6 mm).

Error bar includes: Statistic error from fit ~ 0.2–0.5 mm,

except for few points.

Different fitting methods ~ 0.1–0.5 mm.

Guestimated uncertainty on alignment error & tracking precision ~ 0.5 mm. Contribution: ~0.1–0.4 mm.



Comparison


Source Resol @ 40 mm

Pitch/sqrt(12) 11.5 mm

ACDC3 ~ 9.2 mm

TED Fit ~ 10.6 mm

FNAL Fit 8.10.6 mm

Normal Incidence (0.5)

Preliminary !

TED result was produced by Silivia Borghi and presented by Kazu Akiba at the Florence LHCb week

Low momentum track, momentum not measured. Multiple scattering effect is not removed precisely.

If resolution is determined from RMS of residual instead of fit, then the projection to 40 mm is 9.60.6 mm


Resolution vs Track Angle

Effective track angle is determined in plane perpendicular to the strip.

Sub-samples of 0, 3, 7 and 11 are with reasonable large statistics.

Other angles are due to concentric strip, thus with small amount of hits.

Pitch ( mm)

40 – 5050 – 6060 – 7070 – 8080 – 90

90 – 100

Large statistics

For discussion purpose only

Worse than testbeam 2004 results (ref: lhcb-2007-151)

Different Thresholds

Jianchun Wang 13

Angle ()- 0.5 – 0.52.5 – 3.56.5 – 7.5

10.5 – 11.5


Side threshold = 2 ADC

Seed threshold = 3.6 ADC

Side threshold = 1.8 ADC


Side threshold = 3 ADC

RR sensor

Jianchun Wang 14

N-type Sensor Charge Collection

Jianchun Wang 15

The relative position between irradiation profile and sensor Y is adjusted according to the center of two transition regions ( Position = Yprofile – 39.3 mm).

N-type sensor is flipped (Position = -Y).

XVELO (mm)

YV

ELO

(m

m)

Hit map determined by tracking

All angles

P-type Sensor Charge Collection

Jianchun Wang 16

XVELO (mm)

YV

ELO

(m

m)

Hit map determined by tracking

1 bad beetle chip

The sudden drop after 30 mm is unexpected.

It is very unlikely that the irradiation profile is wrong.

Normal incident track only

Basic on Charge Distributions

The FE electronics were under-powered, resulting in low gain. Most probable charge ~16 ADC instead of ~40.

Constant thresholds (seed=3.6, inclusion=1.8) are used (noise ~ 0.9 ADC counts). Thresholds are low enough to study irradiated sensors.

Gain differences are partially corrected using header heights.

Only hits that match with pixel tracks are looked at, to reduce the influence from uncertainty of noise hits.

Charge distributions are fit to Landau convoluted with Gaussian. The width of Gaussian is fixed to an average value so as to reduce the uncertainty on Landau MP.

In some cases there are shoulders/tails on low side that were not well understood. Fits are at peak areas. Fit range affects MP obtained from fit. MP represents, but not completely, the charge collection efficiency.


Charge (ADC counts)

Sensor Charge Collection

Jianchun Wang 1801/29/10

= – Y

X (

mm

)

N-type

= + Y

X (

mm

)

P-type

?

?

Tracks at 0-8 degrees, detector biased at 500 V.Hit map determined by pixel tracks that matches with VELO hits.

N-type MP Charge At Different HVs

Jianchun Wang 19

HV (V)50040030020010050

Sum of all angles

Some points need further work

P-type MP Charge At Different HVs

Jianchun Wang 20

HV (V)50040030020010050

Some points need further work

Comparison Between N- and P-type Sensor

Jianchun Wang 21

P-type

N-type

Comparing Different Electronics Settings

Jianchun Wang 22

N-typeKazu setting

P-typeKazu setting

N-typeChris setting

P-typeChris setting

optimized for sensors after irradiation.

Optimized for current running in the pit.

biased at 500 V

More on N-type Sensor

Jianchun Wang 23

Artificial parameter from MP so that the shape looks more like the irradiation profile

Slopes in the transition region exhibit small discrepancy.

N-type sensor

MP vs Y for Different X Slices

Jianchun Wang 24

X Slices(–5, )

(–10, –5)

(–20, –10)

( , –20)

X flipped

Phi Value of Sector Borders

Jianchun Wang 25

The VELO alignment wrt pixel tracks has very loose constraint in phi.

Check if this is the source of the shift in MP vs position for different X slices.

Look at phi of matched pixel hits for each sector. Borders are clear.

Fit to error function. The average edge value of the neighboring border is consistent with p (+0.0017 and +0.0016 for N-type and P-type respectively).

Borders between sectors 0&1, 2&3 are consistent with 3/4p and 5/4 p.

The maximum difference is ~ 0.003 corresponding to shift of 0.12 mm at R=42mm.

Ruled out

N-type, Sector 1Edge = 3.1472Sigma = 0.0016

N-type, Sector 2Edge = 3.1395Sigma = 0.0008

F (rad)

Num

ber

of M

atch

ed H

its

Detection Efficiency


Due to the trigger scheme and different DAQ clock frequencies for the two systems, tracks seen by pixel and VELO are not necessarily the same.

Pixel tracks are matched with hits from one sensor (± 200 mm) to ensure this is a real track and seen by VELO.

We then look at the other sensor to see if there is hit that matches the track. The detection efficiencies are thus determined.

Beam profiles are not guaranteed to be the same for different conditions so the weight of dead areas changes for different condition runs.

A dead chip and few dead strips and certain border areas are removed.

In this way, the detection efficiencies reflect more precisely the effect of irradiation fluences and/or bias voltages.

Cleanup of Dead Strip & Borders

Jianchun Wang 27

X (mm)

Y (

mm

)

X (mm)

Y (

mm

)

N-sensor

P-sensor

N-sensor

P-sensor

Remove 6 bad

strips & borders

Remove 4 bad

strips & borders

hit position expectation that are unmatched

01/29/10

!

!


Jianchun Wang 28

N-typeKazu setting

P-typeKazu setting

Normal incident tracks

Biased at 500 V

01/29/10

Not from 0

N-type Sensor Charge Sharing

Jianchun Wang 29

Largest strip ADC value of each cluster

Low tail due to large cluster size

Y = (–42, –32) mm Y = (–32, –18) mm Y = (32, 42) mmY = (18, 32) mm

P-type Sensor Charge Sharing

Jianchun Wang 30

Largest strip ADC value of each cluster

Low tail due to large cluster size

Y = (–42, –32) mm Y = (–32, –18) mm Y = (32, 42) mmY = (18, 32) mm


Jianchun Wang 31

N-typeKazu setting

P-typeKazu setting

All angles

01/29/10

Bias Voltage (V)

500 400 300200 100 50


Jianchun Wang 32

N-typeChris setting

P-typeChris setting

All angles

01/29/10

?

Detection Efficiency Vs Mp

Jianchun Wang 33

N-typeP-type

Biased at 500 V

Detection efficiency is determined by

1. charge collected (MP)2. charge sharing (cluster size)3. seed threshold (constant ADC)

Detection Efficiency Vs MP for Different HV

Jianchun Wang 34

N-typeKazu setting P-type

Kazu setting

All anglesBias Voltage (V)

500 400 300200 100 50

MP vs HV

Jianchun Wang 35

N-type P-type

Non-irradiated

Vdep = 117±7 V

irradiated

irradiated

Fit with a naïve function

Non-irradiated

From non-irradiated

Vdep = 771±43 VVdep = 1218±96 V

For Resolution Study

Jianchun Wang 36

Track Effective Angle (degree)

Select regions Y< –16 mm & Y > 16 mm.

Angles: 0-2, 2-4, 6-8 degrees

Pitches: 64-70, 70-80, 80-90, 90-100 mm

Y (mm)

Pitc

h ( m

m )

01/29/10

Resolution vs Pitch

Jianchun Wang 37

Normal Incidence (0.5)

R of R/ f pairN-type

0-2 degree

P-type0-2 degree

Fully irradiated (Kazu)

Fully irradiated (Chris)

Non-irradiated (Kazu)



Non-irradiated (Chris)

Error not fully estimated

R of R/f pair (Chris, 0 degree)

01/29/10

Resolutions are obtained through Gaussian fit to residual distributions, not just RMS due to bkg hits.

Tracking errors are removed.

Charge Sharing vs Pitch

Jianchun Wang 38

R of R/ f pair

N-type0-2 degree

P-type0-2 degree


Fully irradiated (Chris)




Non-irradiated (Chris)

Error not estimated

R of R/f pair (Chris)

Angle ( )-0.5 – 0.52.5 – 3.56.5 – 7.5

10.5 – 11.5

Resolution vs Pitch

Jianchun Wang 39

N-type

P-type

Error not fully estimated

R of R/ f pair

Angle ( )- 0.5 – 0.52.5 – 3.56.5 – 7.5

10.5 – 11.5

Irradiated Fully None

Angle (degree)0-22-46-8

01/29/10

Center of Residual vs HV

Jianchun Wang 40

N-typefully-irradiated6-8 degree tracks

64 – 70 mm

90 – 100 mm

80-90 mm

70 – 80 mm

Naïve interpretationMax difference ~150tan(8) = 21 mm

Center of Residual vs HV

Jianchun Wang 41

64 – 70 mm

90 – 100 mm80-90 mm

70 – 80 mm

P-typenon-irradiated6-8 degree tracks

Full depletion voltage ~ 110 V

Inefficiency Issue

Jianchun Wang 42

The window is ± 200mm for reference plane hit with pixel tracks.If noise hit gets in due to this window, the efficiency would be lower. The window is tighten to ± 100mm, for reference plane. The efficiency difference is negligible.200 mm might be too tight for DUT.

Non-perpendicular Beam For Irradiation

Jianchun Wang 43

Irradiation profile offsetOld = 39.3 mmNew = 36.8 mm

X Slices(–5, )

(–10, –5)

(–20, –10)

( , –20)

Angle = 0.251

Before rotationAfter rotation

X

YPosition

Position*

N-type

MP vs Y for Different X Slices

Jianchun Wang 44

X Slices(–5, )

(–10, –5)

(–20, –10)

( , –20)

N-type P-type

Sensor Charge Collection

Jianchun Wang 4501/29/10

= – Y(rotate)

X (

mm

)

N-type

= + Y (rotate)

X (

mm

)

P-type

?

?

Tracks at 0-8 degrees, detector biased at 500 V.X

YPosition

X

Y

Position

PixelVELO

Pixel

YX YX

120 GeV proton beam

Pixel

Y

Scint

RR( )F

X

Z

Y

~ 1 m

Pixel X/YVELOPixel Y

Pixel X/Y

LowFluenceRegion

Transition Region

HighFluenceRegion

Inefficiency vs R

Jianchun Wang 49

90-135135-180

180-225 225-270

Radius (mm)

Inef

ficie

ncy

Rat

e

N type R sensor

Inefficiency vs R

Jianchun Wang 50

90-135135-180

180-225 225-270

Radius (mm)

Inef

ficie

ncy

Rat

e

P type R sensor

Collection Of Plots for A Testbeam Paper. List of Possible Plots R/Phi resolution, charge sharing, noise etc. Noise performance and few Landau distributions.

Documents

n strip

ke strip pitch

ke slide

seed threshold

strip clusters

r velo r track

jianchun wang2 slide

jianchun wang4 charge