UM PPS Lab Activities - PowerPoint Presentation

Slide1

UM PPS Lab ActivitiesMid-size Panel Tests

PPS meeting January 15, 2012

Claudio, Curtis, Dan, Ethan, RileySlide2

MP1 TestsJanuary 15, 2013UM PPS Activities2

Panel re-baked over night at 100⁰C

nd

filled with 90%Ar 10%CF

4

at 600

Torr

December 18, 2012.

13 HV lines at 100M

Ω

+ 20 RO lines (attenuation~9.4), discriminated at -150mV, to the Wiener 20 channels counter

(the other 4 lines of the RO card are masked)

Three types of test:

Voltage Scan, with/without Source at fixed R-quench

Position Scan: (un)collimated source moved across RO lines

Uniformity: un-collimated source at 19 cm above the panel



uniform spray of betas over the instrumented region

Remember: This panel is Ni-SnO

2

with line pitch 1mm, with the high numbered channels toward R-quenchSlide3

Initial Voltage ScansJanuary 15, 2013UM PPS Activities3

Top two plot scans with lights ON

- left HV always ON

- right 1’ off before run

The panel is light sensitive!

From next test on, all runs under

a cardboard box

At 880V Rate(

Bkg

)~Rate(

Src

)!

Working

RegionSlide4

Uniformity Scan – Many VoltagesJanuary 15, 2013UM PPS Activities4

The two position are different by 10 mm toward the low numbered channels

RESULT#1) Very Low background

RESULT#2) Hit map substantially flat

RESULT#3) Rate scale with Voltage

(

if the

bkg

remains low!

)

RESULT#4) The rate on the second

scan is much lower at 870V

865V data bad

(to be fixed!)Slide5

Uniformity Scan – 3 PositionsJanuary 15, 2013UM PPS Activities5

Pos2= Pos1 + 5 mm

Pos 3= Pos1 - 5 mm

In all cases, there is an increases for channel #1 and for the higher

numbered channels (#19 is less responsive of the nearby ones)Slide6

Position Scan Un-collimatedJanuary 15, 2013UM PPS Activities6

Gaussian fit ±8 bins around the max

(First 3 runs eliminated, fit too poor at the edge)

Source moved by 2mm/runSlide7

Position Scan Un-collimated ResultsJanuary 15, 2013UM PPS Activities7

As in the Uniformity Scan, the

number of event for higher

numbered channel is larger

Slope = 1/pitch

Fit result not

compatible with

the hardware

value 1

mmSlide8

Long Uniformity ScanJanuary 15, 2013UM PPS Activities8

This hit map gives channel by channel response



correct ion factors respect to uniform distribution Slide9

Channel ResponseJanuary 15, 2013UM PPS Activities9

Relevant -20% +25%, with a few jumps between nearby channels!Slide10

Hourly Hit Map January 15, 2013UM PPS Activities10

Limited difference



It seems stableSlide11

Normalized Hourly Hit MapJanuary 15, 2013UM PPS Activities11Slide12

Uniformity in TimeJanuary 15, 2013UM PPS Activities12

Is the performance (hit rate) of the panel constant in time?

The quadratic fit works better

 the hit rate is not constant,

it decreases in time!

Is this a measurement of the

panel (SnO

2

) degradation or

is it just an initial effect

?Slide13

Hourly RateJanuary 15, 2013UM PPS Activities13

The rate decrease does

not seems to stabilize

Later as in the first test

Weekend long runSlide14

Hourly Channel RateJanuary 15, 2013UM PPS Activities14

Analysis in progress: it will be interesting to see if the decrease rate trend

is the same for all the channels or higher for the more responsiveSlide15

Hourly Correction VariationJanuary 15, 2013UM PPS Activities15Slide16

Position Scan CollimatedJanuary 15, 2013UM PPS Activities16

Movement every 1 mm

New fit function Gaussian + Linear over then whole range

A few runs eliminated

(edge effect)Slide17

Fit of the Mean FitJanuary 15, 2013UM PPS Activities17

The quadratic fit has a better reduced

χ

2

than the simple linear

function and the slope is close to 1. But why ?

Not OKSlide18

Fine Position Scan Collimated72 runs with source moved by 0.25 mm each timeJanuary 15, 2013UM PPS Activities

18

The correction makes

the peak amplitude

more stable!

Before and after correcting

channel by channel the hit mapsSlide19

Lin Vs Quad Fit of the Mean FitJanuary 15, 2013UM PPS Activities19

Once more the quadratic fit has a much better

χ

2

/#

dof

than the linear function

In both cases the slope is far from 1

by many time the error on the fitSlide20

Lin Vs Quad Fit of the Mean Fit (2)January 15, 2013UM PPS Activities20

Now the pure linear function fits better the data, but the slope is still far from the unity

Now the quadratic function fits quite well the data, and the slope is not too far from the unity.Slide21

ConclusionsA lot of new data on the Mid-size panel. It responded very nicely!Significant photo-sensitivity forced us to cover the panelVoltage Scan: with 90%Ar 10%CF4 at 600

Torr

and quench resistor 100 M

Ω

the working region is limited around 860V

Uniformity Scan: MP1 has channel by channel different response

 correction deduced from a long run

Long Uniformity Scan: the acquisition rate is not exactly linear, but it tends to slowly decrease in time. The very long run over the weekend will allow to better check rate (and corrections) stability

Position Scans: thanks to the corrections the data seems better, but the correlation of the source Vs reconstructed position seems to be more slightly quadratic than linear. Even with small source movements we were not able to see the inter-strip spaces (drop in the acquisition rate, broader peak distribution ...). We have also a new position scan at 870V to be analyzed to see any voltage

effect

We have more results than what has been presented here . . .

January 15, 2013

UM PPS Activities

21Slide22

Conclusion (2)Ethan & Riley contributed on taking the scan runsCabinet with doors to shield the Ni-SnO2 panels from the lights is ready to be installedProgress on the gas mixing station leak fixing and on starting our gas chromatography (Curtis and Riley)The CAEN DT5472 default software is only an example of DAQ, we need to invest time/effort to have a solid code for taking data

January 15, 2013

UM PPS Activities

22