From:
owner-project-blm-lhc@listbox.cern.ch on behalf of Gianluca Guaglio
[Gianluca.Guaglio@cern.ch]
Sent: Monday, July 28, 2003 2:38
PM
To: project-blm-lhc (beam loss detection for the
LHC)
Subject: Lauvain data
Categories: CERN SpamKiller
Note: -49
At least, here the Lauvain data
analysis.
In bold my
suggestion for future experiments
GENERAL
SETUP
We pointed the beam on a
collimation copper plate of width 5 mm with an hole of 11 mm of
diameter.
The plate have been fixed with
adhesive tape to the board. Improve the collimator
fixing.
The electronic board consist in
4 channel of integrator+threshold comparator+monostable. Into the
Channels, a constant current of 10 pA has been
injected.
We used a ionization
chamber (IC) kept in vertical position, such as the beam enter diametrically
into it.
The IC was connected to the
channel 4 , the input of this channel has been disconnected and deviated to the
ground.
The beam faces, in the order,
the collimator, the component on the board, the ionization
chamber.
We used the Labview Virtual
Instrument CFC_v11timegraph_lauvin.vi contained in the library
CFC_v11timegraph.llb
Fix current mode
(FIX)
This measurement has been token
injecting a constant current of 1 nA in the CFC.
Initially, we measure the
output signal (almost 5 Hz) without the beam on the 4 channels. Obviously, the
channel 4 didn't see the injected current and so it gave a frequency of around
0.05 Hz. After few points, we switch the beam on and we register the variation
in the answer during the irradiation. After some points, we switch the
beam off and recorded again few points without beam, to see if there is an
integral effect.
Scan mode
This measurement has been token
injecting a current from 1 mA to 30 pA in 7 steps. Precisely 1
mA; 0.5 mA; 0.1 mA, 1 microA; 10 nA; 100 pA; 30 pA.
The scans are recorded
with constant beam.
Procedure
For each component to be
irradiate, we make a scan without the beam, then, in the order, a Fix and
a scan at 5E6 p/cm2/s, 1E7 p/cm2/s and 5E8 p/cm2/s. Finally we make a scan
without the beam and we changed the irradiated component.
We irradiate the following
components: JFETs, Amplifier of the integrator (AMPL), the threshold
comparator (COMP) and the monostable (MONO).
Measurements
The files are saved in the
form: YYYYMMDD_hhmm_cc_note.cfc where
YYYY =year, in our case
2003
MM =month, here
06,
DD =day, here 02 or
03
hhmm = hour and
minutes
cc= number of
channel
note= generally contain the
mode of measurement (fix or scan), the component irradiated and a progressive
number.
The data are summarized and
analyzed in the excel files present in that directory.
All the file recorded the 02
June are testing files. I take the files 20030602_1846_??_zero as reference and
I reported it in every fix*.xls files.
ANALYSIS
General
From the files
20030602_1846_??_zero we can see that there is a temporal drift of the channel
2. We tried, in the files *_2257_??_scan0.cfc To verified if it a proportional
drift or it is a parasitic current in the channel. It looks like the second
possibility, with an extra current of 30 pA. In the files *_2323_??_fix0.cfc we
have tried to change the amplifier (1st point) and also the JFETS (2nd point)
but without success.
In the followings *_fixch2.cfc
we change the resistor for the bias current and we see, obviously, a signal
changing. We reused the previous resistance. Finally we change the
monostable, but always with the same result. So we decide to stay with this
current and we report everything in the original
configuration.
Our machine time stared with
the 20030602_* files, following the above operation mode. There have
been only few exceptions:
- *_0117_* files are an extra
scan after the 5E8 p/cm2/s,
- *_scanAMPL2* are files made
for better realize the radiation effect on the amplifier
- *_scanCOMP4.cfc are two
successive acquisitions to eliminate the errors on channel 3 present in the
first one.
- Sometimes, for the scanning,
thee are only the 3 free channel while the IC channel is keep only before the
scan, to save time.
At the end of the irradiation,
we put the chamber in horizontal (the beam pass it longitudinally) ad we
recorded the *_mono5.cfc file. Then we remove the collimator and the board,
recording the last files with different beam intensity.
For each irradiation we
recorded also the fluence given us by the local operator.
We irradiate each components
until almost 500 Gy.
We have also tried to correlate
the IC frequency with the flux intensity.
JFET
Integral effect (no beam): from the FIX, we don't see any relevant
variation of the channels before and after the irradiation. There is only a
shift (+3%) on the channel 1 during before the last measurement.From the SCAN
files we can observe a residual parasitic current of almost 30
pA.
Cross section effect (with
beam): from the FIX file we can see an increase of 70% (+ 700 pA) during the
highest irradiations, only of 2% during the lowest irradiation (5E6). To mark
the fact that, during irradiation, there is a decreasing trend: could it be a
recovering effect? From the SCAN files we see an increasing during
irradiation at low current. Equivalent parasitic current: almost 200 pA.
Minor increasing in channel 2.
AMPL
Integral effect (no beam): from the FIX, we don't see any relevant
variation of the channels before and after the irradiation. From the SCAN files
too.
Cross section effect (with
beam): from the FIX file we can see a decrease of 25% (- 250pA) during
the 1E7 irradiation, of >-100% during the highest irradiation
(5E8). From the SCAN files we see a decreasing during irradiation: -10 pA
at 5E6, -200pA at 1E7, -800pA at 5E8 . It is also relevant the fact
that the channel 2 increase his equivalent current up to +150 pA during the 5E8
irradiation (both in FIX and SCAN)
COMP
Integral effect (no beam): from both the FIX and SCAN, we don't see any
relevant variation of the channels before and after the irradiation. (Excluded
the probable error in the final scan on channel
3).
Cross section effect (with
beam): from the FIX file we can see an increase up to+10% (+ 100pA)
during the 5E8 irradiation. Minor increasing in channel 2 and 3. From the
SCAN files we see an increasing of +75 pA at 5E8.
MONO
Integral effect (no beam): from both the FIX and SCAN, we don't see any
relevant variation of the channels before and after the irradiation.
(Excluded an error in the FIX on channel 1). Note the decrease in the
last FIX measurement in one point before the switching on. (Alimentation
problem?)
Cross section effect (with
beam): from the FIX file we can see really small increase. From the
scan we can see a probable error in the 5E6 scan on channel
3.
Chamber
We have collected all the
channel 4 average data in the fix_chamber.xls file, "summary" sheet. From this
file we can see that there is not a good linear correlation with the recorder
intensity and the flux values, particularly for the 1E7
values.
What's more, we can see a
general increasing of the values with the time. Activation of some chamber
material? Temperature drift? It is on going the analysis of a long time
measurement to verify this assumption.
Finally we have to note that,
for the horizontal position, the thicker wall reduce the number of counts of
about a factor 300 with 5E8 p/cm2/s.
The values between the
collimated and the not collimated are almost the same, with the exception of the
lowest intensity beam: ratio free/collimation ~300?
CONCLUSIONS
The most sensible components
are the JFET and, secondly, the amplifier.
There are some crosstalk
between the channels, probably through the power
alimentation.
The chamber is not linear in
low proton energy fluxes.
Proposed
actions
Improve collimator
fixing.
Insert the alimentation
measurements in the SW.
Analysis of a long time
measurement.
Better uncoupling
of the channels to avoid crosstalks.
Resolve spikes problem
from the source-switch.
Lessons
learnt
Reproduce entire measurement
set up in house (we lost a lot of time in Lauvain trying to fix the collimator
better)
Radiation could also force the
component to extract current signal (see AMPL)
It is better to calibrate the
amplifier gain after a warm up (see channel 2 drift)
For the moment I think this is
enough.
See you for further
discussions
Gianluca
Guaglio
Nuclear
Engineer
CERN
AB
Division BDI Group BL Section
Site
de Meyrin CH-1211 Genève 23 (Switzerland) or
Site
de Prevessin F-01631 CERN Cedex (France)
Building
865/R-A19, mailbox Z15700
Phone
0041 22 76 78079
Fax
0041 22 76 79560
e-mail
gianluca.guaglio@cern.ch