Front-end electronic status
Walter Bonivento
INFN Cagliari
22/11/2006 Phone conference
Status of CARDIAC production
1)
PCB
9110 produced and delivered to the company
2) BOARD ASSEMBLY
planning ; 4000- and 5110+
(according to P.C. table of 26/7 we need 5040+ 3540- INCLUDING SPARES)
in production the last batch of 3000+ (DIALOG side completed)
a)
b)
c)
d)
e)
boards sent to Rome2 for burn in: 2474+ and 2609sent to LNF: 1196+ 926- (~100 sent back for repair or check)
sent to CERN: 412+ 969boards to be sent soon to burn in: 700
plus some old shipping (small numbers~O(10)/shipping)
the rest under connector assembling or reworking following a test failure
22/11/06 Muon front-end
meeting
Walter Bonivento - INFN Cagliari
2
Test bench
The test bench, managed (and under the responsibility of) by G. Auriemma (INFN
Roma, Univ Potenza), tests:
1)
2)
3)
4)
5)
gain for individual channel with variable Cdet
ENC for individual channel with variable Cdet
cross talk
missing output and input
shorts
This means that if you find broken connections, they broke with DHL…
What it does not test:
1)
2)
currents
addresses
Testing a little late wrt to burn-in: a few HW failures in recent times
22/11/06 Muon front-end
meeting
Walter Bonivento - INFN Cagliari
3
New noise findings
Reminder:
1) no noise test required for test bench of CARDIAC’s ”the noise is measured
at the chip testing company (Microtest)”
2) + chips showed only some more noise at the testing company (but only a factor of
50%)
3) fortunately a noise test was foreseen anyway in the board test bench
4) the boards were (successfully) validated on few chamber types for the PRR
“anomalous” noise found in dressing chambers with positive boardslarge noise
with large Cdet chambers in a large fraction of + boards
the “anomalous” noise is also visible at test bench level, at large Cdet
(N.B.:the reverse was not obvious due to simplified grounding in the test
bench).
SHORT TERM ACTION: a database from Giulio to select good boards to mount on
large Cdet + chambers
22/11/06 Muon front-end
meeting
Walter Bonivento - INFN Cagliari
4
Technical details
The origin of this noise was not found yet (I am in contact with W.Riegler to try
to solve it)
Facts (which might help to understand the cause):
1) single channel response and ENC is as expected at any Cdet. The noise
curves for single channels are well behaved.
2) the observed “anomalous” noise is coherent i.e. when all channels are
switched off but one, no problem. Otherwise the noise in much larger.
3) the effect is also present on the negative board but it is on the left side of
the noise scan curve, so it is not an issue for these boards
On the right side of the – boards there is no coherent noise at all and
therefore the overall noise level is at single channel level
3) the effect is amplified at large Cdet; it grows much faster than linearly
4) not all the boards show it at the same level!
it might depend on a finely tuned parameter…
Missing: deep comparison of a silent + board and a noisy + board
Not many (experienced) people available for deep testing… (me with a recently
born baby…)
22/11/06 Muon front-end
meeting
Walter Bonivento - INFN Cagliari
5
Action plan
If we do not succeed in finding the cause:
1)
2)
3)
4)
5)
use Giulio’s database for immediate need of dressing
test if the noisy boards can fit on the low Cdet chambers TO BE DONE
IMMEDIATELY to see if it works
if yes  fine
if not, we can modify the + noisy boards to become silent – boards (it works,
I tested it switching one chip from + to -; by the way there is no significant
difference in the feed currents with + or -) and modify some – to become +
and test them and so on (this will let us save some money)
…
22/11/06 Muon front-end
meeting
Walter Bonivento - INFN Cagliari
6
Ghosts
There is no cross-talk among chips in the negative boards measurements up
to very high charges and large Cdet do not show any effect; measurements
done with both SPB types
the effect mentioned by A.Vorobyev has to be traced to some other effect
(chamber grounding…)
There is no problem in the DIALOG delay settings DLL must be calibrated
beforehand
There is no significance dependance of the response on the the bias voltage
in the well behaved boards measurement done on the CARDIACGEM
(D.Raspino)
22/11/06 Muon front-end
meeting
Walter Bonivento - INFN Cagliari
7
Offset vs LV Supply
Offset: soglia in mV
a cui è massima la
rate di conteggi, in
funzione della
tensione di
alimentazione
nell’intervallo 2.4÷2.6
V a step di 0.5 V
Ch 0
Ch 1
Ch 4
Ch 5
Ch 2
Ch 3
Ch 6
Ch 7
L’offset aumenta
linearmente in
funzione della
tensione
Il canale 14 non è morto,
si è rotto il pin di
ingresso nella
connessione della board
al test bunch
22/11/06 Muon front-end
meeting
Ch 8
Ch 9
Ch 10
Ch 11
Walter Bonivento - INFN Cagliari
Ch 12
Ch 13
Ch 14
Ch 15
8
Min Charge vs LV Supply
Minimum detectable
charge in funzione
della tensione
nell’intervallo 2.4÷2.6
V a step di 0.5 V.
In questo intervallo
di tensioni la min.
detectable charge
non sembra seguire
alcun trend legato
alla tensione di
alimentazione.
Ch 0
Ch 2
Ch 1
Ch 3
Ch 4
Ch 6
Ch 5
Ch 7
Ch 13
Ch 8
Ch 9
Ch 12
Ch 14
22/11/06 Muon front-end
meeting
Ch 10
Ch 11
Walter Bonivento - INFN Cagliari
Ch 15
9
Sensitivity vs LV Supply
Ch 0
Sensitivity in
funzione della
tensione
nell’intervallo 2.4÷2.6
V a step di 0.5 V
Nella maggior parte
dei canali la
sensitivity sembra
diminuire in funzione
della tensione.
In modo minore
nell’intervallo
2.45÷2.55 V
Ch 1
Ch 4
Ch 5
Ch 3
Ch 2
Ch 6
Ch 7
Ch 8
Ch 9
Ch 12
Ch 13
Ch 10
Ch 14
Ch 11
22/11/06 Muon front-end
meeting
Walter Bonivento - INFN Cagliari
Ch 15
10
Noise vs LV Supply
Ch 0
Ch 5
Ch 4
Noise (a.u.) in funzione
della tensione
nell’intervallo 2.4÷2.6 V
a step di 0.5 V
Ch 1
Ch 3
Anche in questo caso
non è presente alcun
trend in funzione della
tensione di
alimentazione.
Ch 6
Ch 7
Ch 2
Ch 13
Ch 8
Ch 9
Ch 12
Ch 14
Ch 10
22/11/06 Muon front-end
meeting
Ch 11
Walter Bonivento - INFN Cagliari
Ch 15
11