Transcript Calorimeter Group

Choice of action
•
•
•
•
We should choose one solution as the best candidate for a
further test with one full line
Contrary to the previous situation with a clearly malfunctioning
ECAL base. We have two solutions modified HCAL and
modified ECAL which fulfils the minimum basic requirement
i.e. a regulating system which is independent of LV or MV
values.
So the decision has to be taken on more subtle arguments than
does it works at all!
A large number of measurements done on different models by
Pavel and I (and others before) ECALm and HCALm or
HCALmw or better HCALmc
October 1st 2008
1
Jacques Lefrancois
Arguments (I)
•
•
•
•
Noise comparison: essentially the same noise for both solutions,
even in cases where the medium voltage was reduced to be close to
the minimum allowed (Youri's worry) (still a test Wed afternoon
on HCALmc but >99% probability noise as HCALmw or better)
Insensitivity to medium and low voltage variation. Both solutions
OK for 830volts HV, from 78 to 48 MV and +-6 to +-4 LV
essentially no differences.
The 60KHz noise induced in the MV line is about twice as big for
ECALm than HCALm (0.2 to 0.4 volts) but the circuits are not
sensitive to MV noise.
There exists some parasite pulse on the OP-AMP output but very
far from non-linear operation 1.4v p-p for HCALmc about 1v
ECALm if 100 ohms in line
October 1st 2008
2
Jacques Lefrancois
Arguments (II)
•
•
•
A test was done with a 2.5% modulation of the control line at
5Hz : Hcal mc has a very short(<2ms) overshoot while the
ECALm has a bigger overshoot which oscillates for about 50 ms
Test were done switching on-off medium voltage with or without
LV . This has no effect on OP-AMP for HCALm while for
ECALm it produces spikes up to 40 volts at the input of the OP
AMP . The datasheets specifies : "absolute maximum ratings of
input = +-15v or +-supply voltage whichever is less"
Even if we did not destroy any OP AMP and if abrupt turn onoff is not planned a naturally protected circuit is better.
October 1st 2008
3
Jacques Lefrancois
Summary
•
•
•
In view of arguments (II) above and of the fact that the HCAL
design has been tested satisfactorily for > a year on 1500 bases
My advice therefore is to choose the HCALm for the solution to
be tested on the line test and for the 6000 bases if the line test is
OK.
Among the variants of HCALm, HCALmc has performances at
least as good as HCALmw and should be easier for
modifications. Therefore it should be the one chosen.
October 1st 2008
4
Jacques Lefrancois
Measurements on line test (my opinion)
•
•
•
•
Check the noise on the bases
Check the voltage calibration and its independence on LV and
MV
Redo abrupt switch on off
?
October 1st 2008
5
Jacques Lefrancois