UV LED quality test
 All 872 UV LEDs tested
 Selection is about to finish
Jaroslav Zalesak
Institute of Physics ASCR, Prague
May 11, 2006
CALICE Collaboration Meeting,
McGill Uni, Montréal, Québec, CA
1
Measurement setup
•LED Driver pulses – UV LED – opt. fibers – APD – preamp – scope
•Measured by scope of 4 channels/fibers for each LED (semi-automatic procedure)
•@ different positions (diff radius: 0.58, 1.15, 1.76 mm) wrt. central LED point
5
4
3
11
10
6
1
12
7
2
9
8
 Light source asymmetric
wrt. LED central point
Phi response variations
 Zero angle position
uncertainty
reproducibility  5%
May 11, 2006
CALICE Collaboration Meeting,
McGill Uni, Montréal, Québec, CA
2
UV LED (400nm) characteristics
Angle Anisotropy
Zero position setup uncertainty
3.500E-08
Good linearity behavior
3.000E-08
Linearity
2.000E-08
Area [Vs]
Area
2.500E-08
1.500E-08
1.000E-08
2.50E-08
2.00E-08
5.000E-09
0.000E+00
1.50E-08
0
60
120
180
240
300
360
1.00E-08
Theta
fibre 11
fibre 3
fibre 6
fibre 8
5.00E-09
sum
Temperature Dependence
Ampl [V]
3.00E-08
0.00E+00
2.800
0.3500
3.300
3.800
4.300
4.800
LV [V]
0.3300
fiber 11
0.3100
fiber 3
fiber 6
fiber 8
SUM / N_fib
0.2900
0.2700
0.2500
0.2300
0.2100
0.1900
2-3% Temperature correction
change of APG gain (10) w/ T
0.1700
0.1500
299.5
300.0
fiber 11
fiber 3
May 11, 2006
300.5
fiber 6
301.0
fiber 8
301.5
302.0
302.5
303.0
303.5
CALICE Collaboration Meeting,
Temp [K]
McGill Uni, Montréal, Québec, CA
SUM / N_fib
fit 11
fit 3
fit 6
fit 8
fit sum
3
UV LEDs Response – full sample 872 LEDs (I)
 after light calibration and temperature corrections
Area [nVs]
Temp corrected Calibrated Area
45.0
40.0
35.0
30.0
25.0
20.0
15.0
10.0
5.0
0.0
300
400
500
600
700
800
900
1000
1100
LED #
fiber 11
May 11, 2006
fiber 3
fiber 6
fiber 8
Tcorr Calib SUM
CALICE Collaboration Meeting,
McGill Uni, Montréal, Québec, CA
4
UV LEDs Response – full sample 872 LEDs (II)
 response normalized to the average over all LEDs (summed up all fibers)
 visible decrease of response for beginning of data taking
 but not correlated with LEDs themselves 
Relative Area
Different SUMs
CT = 1 - 0.1 * (T-To)
1.50
1.40
1.30
-2 K
1.20
1.10
Temp
1.00
0.90
+2 K
0.80
0.70
0.60
0.50
300
400
500
600
700
800
900
1000
1100
LED #
norm SUM
May 11, 2006
SUM * Tcorr
Tcorr Calib SUM
CALICE Collaboration Meeting,
McGill Uni, Montréal, Québec, CA
Temp
5
Single LED time stability
Norm Area
Time Dependence LED# 230
1.40
• long time stability: over 20% decrease of fiber light response
1.30
1.20
1.10
1.00
0.90
0.80
0.70
0.60
0.0
10.0
20.0
30.0
40.0
50.0
Time [Days]
fiber 11
fiber 3
fiber 6
fiber 8
Tcorr Cal SUM
Norm Area
Time Dependence LED# 639
• middle time stability: RMS  5% (reproducibility)
• short time stability: 1.5-2% LED w/o touching (= statistics)
1.20
1.15
1.10
1.05
1.00
0.95
0.90
0.85
0.80
25.0
30.0
35.0
40.0
45.0
50.0
fiber 11
fiber 3
fiber 6
fiber 8
55.0
60.0
65.0
Time [Days]
May 11, 2006
CALICE Collaboration Meeting,
McGill Uni, Montréal, Québec, CA
Tcorr Cal SUM
6
Study of time stability
to be corrected
Area [nVs]
Time dependence - Temp Corr Calib Area
45.0
40.0
 divide into 2 parts:
Y= A + B * (x-x0)
[slope]
Y= A
for x>x0 [const]
 fit 3 parameters:
A,B and break point x0
35.0
30.0
25.0
20.0
15.0
10.0
5.0
0.0
0.0
10.0
20.0
30.0
40.0
50.0
60.0
Time [Days]
Fiber 11
Fiber 3
Fiber 6
Fiber 8
fiber 3
May 11, 2006
CALICE Collaboration Meeting,
McGill Uni, Montréal, Québec, CA
 vary for each fiber
Preliminary:
 break points are similar
 650±40   day 30
fiber 11
7
Selection criteria
LED# 640-1179 (532pcs)
 Preliminary: only “constant” part
i.e. 532 LEDs
100
90
 = 12.7%
80
70
 within range of ±15% 82% of LEDs
 RMS = 13%
60
50
 additional (redundant) criteria:
variation range for each fiber
40
30
20
10
LED #
all
640--1179
koef
RMS
1.000
13.1%
0.955
12.7%
May 11, 2006
100%
872
100.0%
532
100.0%
± 20 %
1.5
773
88.6%
498
93.6%
1.
50
1.
40
1.
30
1.
20
1.
10
1.
00
0.
90
0.
80
0.
70
0.
60
un
de
r
0.
50
0
± 15 %
1.35
650
74.5%
434
81.6%
± 10 %
1.22
498
57.1%
309
58.1%
±5%
1.11
260
29.8%
176
33.1%
CALICE Collaboration Meeting,
McGill Uni, Montréal, Québec, CA
8
Conclusion
 All UV LEDs (872 in our hands) were tested
 “no name” leds look reasonable!
 in general: LED light emission show same characteristics  similar behavior
 decrease of response @ first measured leds is probably due to instability
of the measured equipment (aging,..) than degradation of light of LEDs themselves
 apply correction for each channel/fiber with function of const + slope line
 main selection criterion based on led light variation being within range ± 15%
 in ‘const’ sub-sample of 532 leds 82% satisfied  1.35 highest/lowest emission
 for entire LED system the PINs watch of LED response important
to keep its time and temperature stability
 distribution of one LED light into single fibers (up to 100% of lowest->highest yield)
rather important than overall variation of light among leds
May 11, 2006
CALICE Collaboration Meeting,
McGill Uni, Montréal, Québec, CA
9