Transcript Update of the ADP system debugging

TC detector
Flavio Gatti - Univ. and INFN of Genoa
on behalf of the TC Group
PSI- Review Meeting, 17 Feb. 2010
Status in RUN 2009


Longitudinal detector (Bars+PMTs): it has operated as in RUN 2008
without HW changes.
Transversal detector (Fiber+APD): the readout electronics has been
modified for achieving faster output pulses.
Total output
80 ns
~20 ns risetime
Diff. out (-)
Diff. out (+)
Longitudinal Detector Performance


Small worsening of timing resolution (see next talk by Cecilia)
Check under way:



Baseline noise/EMI at the Disc input worse than in 2008
Optical coupling (silicone grease, air gap, absorption)
PMT Gain: Re-measure PMT gain of 4 selected bars (8-PMT)
2008
2009
Intrinsic resolution
(jitter of e+ time
of flight not
included)
Most probable origin: higher baseline noise
@ discriminators (DTD) input




Baseline oscillates (50Hz) with amplitude of 15 mV
While a moderate threshold decrease improves timing res., a further decrease
gives worse results (10mV).
But 10 mV threshold was the value used in the beam test!
High level of signal cleanliness is needed for best performances. Some further
improvement must be achieved, taking care (monitoring) the baseline noise
Baseline @ Disc input
Baseline sampled at high freq.
New APD FE electronics


New electronics based on voltage amplification of APD current signal
on noise matched resistor R3
The boost capacitor reduce the quenching effect of HV line resistors
enhancing signal in proximity of breakdown point
RMS noise value of 25mV in the 20MHz BW achieved at the output of
the 8 summing channels
Amplitude [V]

Time [A.U.]
New APD FE electronics



Improved PCB layout (better separation of Analogic and Digital
grounding, lower coupling even if higher bandwidth- about 20 MHz)
New components instead of ones that have shown failures.
Control bus (I2C) improved
New FE Board
Old FE Board
Electronics production and commissioning
actual schedule







Design of the new electronics in Dec.09
Prototype and test: Mar.09
Component procurement and PCB production: Apr. 09
Delay in delivery and bad production. First 32 channels at the end of
July and second 32 channels in a second production at end of august.
Mounting and commissioning:
 TC US 1st week of August,
 TC DS 1st week of September.
Positive stand-alone tests before entering in COBRA.
Due to the tight mounting schedule and the priortary DAQ tuning before
starting stable data taking, we were forced to skip the full integration test
with DCH inside COBRA. Test were made only accessing the electronics
outside COBRA.
First TC Full operating test with DCH:
found a not strict correlation with DCH noise from APDs US/DS




From
From
From
From
Run
Run
Run
Run
50225-50358 APD ON,
50359 APD off for DAQ problem
50506 APD ON
50590-50659 APD ON -> DS OFF-non effect -> US off noise disappear
APD ON
APD ON
APD OFF
DAQ Problem
APD DS OFF
APD US OFF
?
Katia F.
Comparison with Run2008



Effect of APD on the DCH in run 2009 of old electronics
From run 38104 al run 39349 gli APD OFF, in the remaining runs APD ON.
Negligible effect on DCH
APD OFF
APD ON
Katia F.
Negative results acting outside COBRA


A full survey of all possible sources of noise/EMI outside COBRA results
with no improvement: conditions of stable performance were not achieved
Need of working directly on the TC on-Board electronics → wait for the shutdown
APD ON
Katia
APD OFF
Work under way in Jan.-Mar.


Level of EMI/noise signal at the input:


20-50 uV if not coherent,
3-6 uV if coherent.


0.1-1 V signal on output cable--> possible irradiation to DCH
Positive feedback if input couple with outputs
Due to high amplification:

Work planned with the US and DS TC:
1- test an EM shielding on the preamplifier input
2- test the effect of pick-up on HV and LV cable
3–reduce the cable emission inside COBRA
4- eliminate any cause of coupling of the output to the input
5- change the not working boards and a few resistors on I2C bus receiver

Test the effect of these changes with the TC inside the COBRA.

Two phase work:
1-the first just after the Xtmas holidays (jan 11 to 29) in order to find the causes of the
excess noise and study/test the solutions
2-the second period after a while in order to have enough time to build the shielding
tool in our workshop (2nd half of February)
1St Phase done







Balanced the I2C bus from Pacth Panel to F.E. boards (change of pull-up
resistors) → now bus signal in the specifics in any conditions and stable
when idle
Inserted a shield on to APD-Preamplifier input → decoupled input from EMI
and HV or output (digital or analogic cables)
Reduced the analogic output band from 150 Mhz to approx 50 MHz with
further feedback capacitor → reduced the capability to emits high frequency
signals
Decoupled differential amplifier output from twisted pairs cable with small
resistor (60ohm/polarity) → reduced back action from electronics rack
Reduced HV- output coupling by choosing different cable paths and proper
shielding → output signal decoupled from FE input through HV cables
Burst signals generated by HV where filtered ( do be done properly later)
Improved Power supply of F.E boards by changing cables from Patch Panel
→ decoupling of different F.E. Board respect supplied by the same Patch
Panel
Few snapshot of the work under way
Decoupling HV-input-output
Shield of APD-Preamplifier input
Test of improved supply sables
Insert 2 capacitors, change 1 resistor
Insert 2 output matched resistors
Before/After (few examples)
Typical EMI from HV
Out
HV filtered
Out
HV-500MHz BW
HV-20MHz BW
Clean I2C transmission
Pattern and stable after
transmission
I2C with high freq oscillation
after the transmission
Example of self-sustained and/or picked-up
oscillations

100MHz, 20 MHz, 8.3 MHz oscillations were observed but not found a clear origin

Most probable conclusion: mix of picked-up signal that were amplified and
sustained by electronics through output-input coupling inside COBRA.
100 MHz
8 MHz
Effect on the DCH noise:
at the beginning of the 1st phase test
}
In on
Out off
In off
In off
Various Grounding
and filtering configuration
In On full init
In On full not init
In On partial
In On partial init.
In off
Out off
Effect on the DCH noise:
at the end of the 1st phase test
Presently achieved an average noise RMS of 25 mV as in the APD test with standalone
electronics at Genoa and PSI (outside COBRA), Aug 09.
Conclusions
 The goal of the 1st Phase beam-off has bee achieve successfully.
 After working on shielding, filtering and signal matching, we are
now at the target performance of APDs in operating conditions
inside COBRA.
 Negligible effects on DCH noise.
 We need to perform these modifications for all channels and both
TC in the next month
 Most of February needed for preparing these modification
(shields, cables, filters on HV,…)
 Completion of work: end of March 2010