Beam Secondary Shower Acquisition System: Front-End RF Design (2)
Student MeetingJose Luis Sirvent
PhD. Student26/08/2013
-20dB
40dB
-6dB
pCVD
CividecAmplifier
Attenuator
CividecDiamond Detector
DC-4GHzSplitter
-6dB
-6dBDC-4GHzSplitter
-6dB
34dB
-12dB
-32dB
-6dB
HV
12V
Tunnel
Surface
1. The dynamic range and the three lines• Front-End proposal
Termination50Ω
Fc= 5 Hz
Low Pass Filter
DC
• Not yet clear:1. If DC measurement is necessary then -9dB splitter: Needed lines 34dB, -15dB & -35dB2. ADC or QIE10?ADC Needed filters in lines 200MHz – 2GHzQIE10 Needed Amplif DC-2GHz or no AmplifProblem for Long lines + QIE10??
1.Front-End amplifier selection• Option 1: Commercial from Cividec
Price: 1755 CHF Price: 1755 CHF
1.Front-End amplifier selection• Option 2: Build our own amplifier with Gali + MGA
– Cheaper alternativeDevelopment of a board
– Inside Cividec 40dBGali 52 (Price: 1.2€)Agilent MGA-62563 (Price: 8.2€)
– Radiation Tolerance:Tested in 2005
• http://lhc-expt-radmon.web.cern.ch/lhc-expt-radmon/meetings/2005-03/Gorisek-BCM%20in%20ATLAS.pdf
1.Front-End amplifier selection• Option 3: Evaluation Boards (20dB) from Mini-Circuits
(Price: 52€)
• http://217.34.103.131/pdfs/GALI-52+.pdf• http://217.34.103.131/pcb/WTB-409-52+_P02.pdf
1.Front-End amplifier selection• Option 4: Build our own amplifier • with BJTs
• http://www.intersil.com/content/dam/Intersil/documents/an15/an1503.pdf
2. Noise study in lines (In simulation)
New!
2 4 6 8 10 12 14
x 10-4
-2
-1
0
1
2
x 10-3
Time (s)
Ampli
tude
(V)
104
105
106
107
108
109
0
2
4
6
8
10
x 10-7
Frequency (Hz)
|Y(f)
|
-2.5 -2 -1.5 -1 -0.5 0 0.5 1 1.5 2 2.5
x 10-3
0
1
2
3
4
5
6x 10
5
Volts
Num
ber o
f poin
ts
Noise distribution
Gaussian fit mean:0.28uV sigma:378uV
Measured in SPS BA5
2. Noise study in linesExample of digitalization with noise (Sigma=6.4mV)
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5
x 10-3
0
0.2
0.4
0.6
0.8
1
Time (s)
Am
plitu
de (
v)
Digitalized profile ADC 14 bits
ADCC GFit Sigma:714.98um
ADCB GFit Sigma:566.73um
ADCA GFit Sigma:313.05um
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5
x 10-3
-10
-5
0
5
10
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5
x 10-3
0
0.2
0.4
0.6
0.8
1
Time (s)
Am
plitu
de (
v)
Digitalized profile ADC 14 bits
ADCC GFit Sigma:780.93um
ADCB GFit Sigma:570.95um
ADCA
GFit Sigma:317.58um
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5
x 10-3
-10
-5
0
5
10
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5
x 10-3
0
0.5
1
1.5
2
2.5
3
x 10-10
Position (um)
Cha
rge
(C)
Digitalized profile QIE10
QIE10C GFit Sigma:563.2um
QIE10B GFit Sigma:566.26um
QIE10A
GFit Sigma:358.4um
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5
x 10-3
-10
-5
0
5
10
ADC Shaper Off QIE10 ADC Shaper On
2. Noise study in linesEvolution of the Beam Sigma Error VS Noise
Considerations:1. The noise is white: Distributed in the whole spectrum
ADC Shaper On Still affected but working better than Shaper Off, in practice the results should be better.
2. The 40dB amplifier: Considered in simulations as ideal DC-infinite Hz (In reality RF 1MHz-2GHz)The 34dB line in practice could not be used by the QIE10 since it has a high-pass filter, the value used should
be -6dBObservations:3. QIE10 performs well by integrating , the mean noise value is 0 so in certain way integration filters noise.4. As specked GausFit error increases with noise up to near 10% for these conditions and Noise sigma=102.4mV5. For low noise (sigma < 1.6mV) Best ADC Shaper ON / For high noise (sigma > 1.6mV) Best QIE10
-20dB
pCVD Attenuator
CividecDiamond Detector -6dB
DC-4GHzSplitter
-6dB
-6dB
-20dB
HV
Tunnel Surface
2. Noise study in linesFor QIE10 we cannot amplify…so…• Long lines QIE10 Front-End proposal
Termination50Ω
Long CK50 link (~100m)
Fc= 5 Hz
Low Pass Filter
DC
0.5 1 1.5 2 2.5 3 3.5 4 4.5
x 10-3
0
10
20x 10
-12
Position (um)
Cha
rge
(C)
Digitalized profile QIE10
QIE10C GFit Sigma:562.83um
QIE10B GFit Sigma:566.24um
QIE10A
GFit Sigma:566.14um
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5
x 10-3
-10
-5
0
5
10
QIE10
2. Noise study in linesFor QIE10 we cannot amplify…so…
Considerations:1. Metres of cable considered for the simulations: 100m2. In any case not taken into account impedance mismatching / reflections
Observations:3. For same conditions in average QIE10 performs better4. Obviously in the previous case QIE10 part of the scan was amplified (by simulation) and therefore the SNR was better 5. Tendency shown as previously6. QIE10 performance affected by noise and lines attenuation (SNR)