Status of New TPC( )Ⅱ

Performance Study

Yohei Nakatsugawa

LEPS Meeting in Taiwan

What we have done

TEST experiment in October 2005 Ⅰ Pad Response Function

TEST experiment in December 2005 Ⅱ Position Resolution in X-Y plane

In all tests, only MWPC part was used.

Pad Response Functiones

timat

ed a

vala

nche

pos

ition

(mm

)

central pad position(mm)

estimated avalanche position(mm)

sigma= 6.99×10-2mm

mean = 0

Usually, Gaussian is used.

simulation (intrinsic)

I searched for better function …

avalanche position = 0 (fixed)

Pad Response Function

F(x)=1

(p1+p2 (x-p3)2) a

a=3.5

sigma = 4.431×10-2 mm

mean = 6.62×10-3 mm

p3 = avalanche position

Is this really better function?

Pseudo Lorentzian

better than Gaussian

TESTⅠ 　 (October 2005)

55Fe

pulse motor

anodewire

trigger

To decide PRF,We measured the dependence of pad gain on hit position.

Source spot is on layer4.

TPC was set vertically and moved by 500μm step by pulse motor.(total: -2500~2500μm one pad width)

X-ray source (55Fe, 5.9keV) was used.collimator

One anode wire above the layer4 wasused as trigger.

layer4

32mm

3mm 16mm

x

y

z

y

pad plane 1mm

TEST set upⅠ

This is actually not a photo of TEST set up, but it was like this.Ⅰ

Analysis

Source event selection

Three pads in the source spot are hit.Other layers are not hit.

PRFmeasured(x) =∫ ｄｘ’ PRF(x’)*exp[ -{ x - x’ }2/(2*σ2) ]σ=1.106 (by Monte Carlo ）

Unfolding the dispersion of X-ray

assuming Gaussian

Compare Gaussian and Pseudo Lorentzian

Analysis

measuredPseudo Lorentzian

Pseudo Lorentzian

Gaussian

χ2 Gaussian 10.5 Pseudo Lorentzian 5.2

Pseudo Lorentzian is better.

TEST ( December 2005 )Ⅱ

γ

Pd converter

magnet

e-

e+

up veto scintilator

start up start down scintilatorsolenoid

trigger condition: up start up start down

To see angle dependence of XY resolution, data were taken at0°(figure),±12.5°,±25° TPC angle.

rotation

Sector 1 & 4 could not be read out due to geometry in the hatch. (Cable length was not enough.)

~25cm

(~5cm width)

dipole

Layer9 was also read out.( new FADC)

TEST set upⅡTPC side view

e-

TEST set upⅡTPC side view

e-

TEST set upⅡTPC front view

TEST set upⅡsweep magnet

Flux is closed by iron planes.

TEST set upⅡsweep magnet

Typical Event Display

storage ring

top viewTPC angle　　　 0°

Typical Event Display

TPC angle　 12.5°

storage ring

top view

Typical Event Display

TPC angle　　　 25°

storage ring

top view

Extra Data

γe-

e+

start up standard start counter

We just wanted to see multi-track and to read out all sectors!

trigger condition: Tag up start up STC

Typical Event Display

TPC angle　　　 0°

storage ring

top view

Typical Event Display

TPC angle　　　 0°

storage ring

top view

Analysis

require more than 4 layers are hit.

line fit tracking ( least square ) residual distribution

x position : estimated by PRF (PL)y position : center of layer

TPC angle : 0° sigma : 176μm

12.5°371μm

25°824μm

ADC value of pulse peak is used.

Time

peak

AD

C

Our goal …at least ~300μm too bad..Resolution is getting worse as track angle increase.

Analysis

y position in layer

Anode signal has random fluctuation. Anode which has large signal drags estimated hit position.

true positionestimated

In previous analysis,y position was fixed.(center of layer)

y position also have to be estimated. using anode information

true position

Analysis

using anode signal

y position = center of gravity of charge induced by 5 anode wires = ∑yi * wi * ADCi ( i = 1~5)

w1 = w5 = 0.1098 , w2 = w4 = 0.3776 , w3 = 0.4567

12345

not using layer1

no field cage Electric field may be distorted.

calibration of anode

using cosmic ray data ( taken after beam time ended ).

Actually, anode 1 & 27 (most outside) have lower gain in calibration data.

TPC angle : 0°sigma : 175μm

12.5°196μm

25°382μm

(176μm) (371μm) (824μm)

Analysis

Resolution is improved.

Analysis

Comparison with Gaussian PRF

175μm

196μm

382μm

192μm

197μm

383μm25°

12.5°

0°

Pseudo Lorentzian GaussianTPC angle

Pseudo Lorentzian is slightly better…

Intrinsic resolution

require 5 layers are hit

residual distribution in the middle layer.

track including and excluding the middle layer

σexc =54

σσinc =45

σ σ= σinc * σexc

0° 133μm12.5°163μm25° 331μm

intrinsic resolution

Next

X-Y position resolution was improved by using anode information,but is still over 300μm. further analysis

Study of Z-direction drift velocity, position resolution, dependence on dip angle …

We are waiting for completed new TPC !

DAQ collector has to be arranged for new FADC. ( The number of channel is different. 32ch 16ch )

Convention

1

2

3

45

6

1234567

89

x

y