G.Sguazzoni 1 Tracker Week April ‘03 FE hybrid assembly and FE hybrid assembly and quality assurance quality assurance The hybrid team: Alan Honma, Michael Poettgens, Gabriella Pasztor, Bernd Reinhold, G.S. Bruno Wittmer CERN –EP Division “ “ ...resista ...resista nce nce of hybrids of hybrids is is not not futile…” futile…”
30
Embed
Tracker Week April ‘03G.Sguazzoni1 FE hybrid assembly and quality assurance The hybrid team: Alan Honma, Michael Poettgens, Gabriella Pasztor, Bernd Reinhold,
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
G.Sguazzoni 1Tracker Week April ‘03
FE hybrid assembly andFE hybrid assembly andquality assurancequality assurance
The hybrid team:Alan Honma, Michael Poettgens, Gabriella
...custom patched for: gas/water flow meter NIM-TTL pulser driving
G.Sguazzoni 12Tracker Week April ‘03
Slow control: softwareSlow control: software
Temperature monitor
Humidity monitor
Peltier current
Safety
TCPUDP
Temperature regulation
Water & gas flow
Custom version of Aachen Cooli Devive Controller – ACDC (LabVIEW)
G.Sguazzoni 13Tracker Week April ‘03
Post PA assembly testsPost PA assembly tests
Typical test sequence during a cooling cycle
Hybrid ON* Ped, Noi Pulser Short
Detection Hybrid OFF*
*Added after problem with hybrid in cold (see later)
TROOM
–20oC
~25 min/4 hybrids
G.Sguazzoni 14Tracker Week April ‘03
Long term cyclesLong term cycles Sample test!!! emulates ~10
CMS running-shutdown periods
Each cycle same tests as the single cycle…
TROOM
–20oC
~5 hours/4 hybrids
last cycle: stronger! (TROOM –25oC +35oC)
G.Sguazzoni 15Tracker Week April ‘03
Noise issues for hybrid+PANoise issues for hybrid+PA
No reference plane below PA PA is a weak point wrt noise injection
ampli channel
GND lines connected to the support plate by a wide copper strip
Hybrid
ARC FEARC FE
Hybrid GND plane
Hybrid Al support plate
PA line
ARC FEadaper
G.Sguazzoni 16Tracker Week April ‘03
Noise comparisonNoise comparison
Grounding scheme comparison
PA lines length
structure
BAD grounding
NEW grounding
G.Sguazzoni 17Tracker Week April ‘03
Short detectionShort detection
1 SHORT = 2 channels lost
G.Sguazzoni 18Tracker Week April ‘03
Short detection principleShort detection principle
Calibration signal is sent every 4 channels
PIN HOLE like effect may also contribute
IN OUT
ch. #N
ch. #N+1
CALsignal
G.Sguazzoni 19Tracker Week April ‘03
Short detectionShort detection Simple and effective threshold criterium
SHORTs
Pulse shape
Pulse heights
G.Sguazzoni 20Tracker Week April ‘03
The first (detected) short!The first (detected) short!
Short = bonding problem PA
APV
G.Sguazzoni 21Tracker Week April ‘03
Open detectionOpen detection
OPEN = 1 channel lost
G.Sguazzoni 22Tracker Week April ‘03
Pulser hardwarePulser hardwareA
RC
NIMpulsertr
igge
r NIMto
TTL
start
out
50“antenna” very close (~1mm) to the bond
pads
Peak - InvOFF Pulser ON Peak
Mode Optimal
latency
OPENS
G.Sguazzoni 23Tracker Week April ‘03
Open detection: basicsOpen detection: basics
Removed Hybrid-PA bonds or scratch on PA lines*
Spikes in the pedestal profile!!!
Removed Hybrid-PA bonds or scratch on PA lines*
Spikes in the pedestal profile!!!
On-board CM subtraction
Pedestal ~unchanged
On-board CM subtraction
Pedestal ~unchanged
*even quite close to the bonding pads
G.Sguazzoni 24Tracker Week April ‘03
The open-detection criteriaThe open-detection criteria
Optimal to detect isolated opens
Difference wrt neighbours
Mean Subtracted Pedestal
OPENSin our test
hybrid
G.Sguazzoni 25Tracker Week April ‘03
The TOB outcome so far…The TOB outcome so far… 78 received Reception TEST
72/74 good; 2 not OK 11 FNAL & 5 UCSB 2 ‘ TEC1.1 ’ Aachen
Assembled 48/48 good
Bonded 45/45 good
Thermal cycle test Standard: 20/20 good
20 FNAL;
Long term: 8/9 good, 1 fails in cold
14846/14848 good bonds; 1 short, no opens
G.Sguazzoni 26Tracker Week April ‘03
PedestalPedestal Large non-uniformities
Max-MinMax-MinAPV wise
Pedestal cuts in module-test draft procedure ~ too tight
Pedestal
Pedestal<Pedestal>
Pedestal<Pedestal>
APV wisePre
lim
inary
!!!
G.Sguazzoni 27Tracker Week April ‘03
Hybrid+PA NoiseHybrid+PA Noise
Normalized Noise cuts (20%) in module test draft procedure
-20oC
ROOM Noise increases in cold
Peltier current?
Prel
imin
ary!
!!
Noise<Noise>
Noise
G.Sguazzoni 28Tracker Week April ‘03
The faulty hybrid:i.e. the kind of nasty problems we are supposed to discover
The faulty hybrid:i.e. the kind of nasty problems we are supposed to discover
The hybrid fails only @–200C No APVs output but answers to I2C commands other strange behaviors (current, DCU readings) only a certain command sequence induces the failure;
otherwise the hybrid works perfectly Test procedure modified to foresee hybrid OFF/ON &
initialization before any group of tests
First crude investigation: No visible defect MUX chip???
…further study needed
G.Sguazzoni 29Tracker Week April ‘03
Test Station throughputTest Station throughput Target:
Nominal: 32 hybrids/day UCSB will take over TOB bonding and testing
– they are duplicating the setup 4 hybrids 40 min (preparation+test)
Limiting factors Cooling time Loading/Unloading
(40min/4hyb) * 32hyb ~ 5h Effective production request could be
higher… A second “CERN” teststation is in preparation (Cooling cycle test could be done on sample