Transcript HCAL_cable_issues

Cable issues at the detector for
HCAL
Peter Göttlicher, DESY-FEB, December 12th, 2008
- Scheme of EUDET
- Direction of ILD
- GND
- power distribution
- signal individual/daisy-chain
Condition as on EUDET
Peter Göttlicher, DESY-FEB, December 12th, 2008
1.8A at DIF – layer end
< 0.1 A in gap
LDA
48 Layer
….90A/half-octant
Signals from LDA to DIF
Individually
DIF
DIF
DIF
No Common GND-reference
….. Service has 2 dm2/half octant
What to do after EUDET?
2
Cabling EUDET
EUDET-prototype: list for 130mm x 140mm
Peter Göttlicher, DESY-FEB, December 12th, 2008
Sorting with CoralDraw easy,
Reality will be much worse
FRNC-cables are normally thicker
Most is flame resistant outer jackets
Cooling
Cooling
What is need for ILD?
+ GND-defining … More space!
- 100A through LDA or
compacter power cables
- daisy-chaining signal would
safe some space.
Can ILD live with
more (or that) space?
signal
140 mm
power
One power+signal per Layer
3
GND in half octant –ILD
Peter Göttlicher, DESY-FEB, December 12th, 2008
Resistivity of connectors: mW, 1.8A
Floating Power supply
mV equal potential would be nice:
- bias regulated in DIF
so not so sensitive as
CERN/FNAL-test beam
GND-bar
effecting performance the
order of magnitude:
What really required – open ?
90A
Voltage drop allowed
45A
45A
To local metal
PE-GND
and connection to supply
Wmm2 length  1m (Current  90 A / 2) / 2
Cross  sec tion  (0.017
)

m
2
allowed voltage  1mV
 210mm2
Reasonable thickness
Factor 4 more, if power at outer end
4
Power in half octant –ILD
Power grouped for few DIF’s,
Peter Göttlicher, DESY-FEB, December 12th, 2008
- power supply of “usual” current
Floating Power supplies
90A
Voltage drop allowed
- voltage drops allowed
local regulators for LV and bias
- Keep loops small:
... GND-voltage
… EMI and
magnet discharge
…voltage and return in
ONE cable
DIF
DIF
DIF
DIF
DIF
DIF
DIF
DIF
DIF
DIF
- With distributing the current-return connections to GND-bar
the current into drops an order down!!! Thinner possible?
My guess: Space for DC/DC (magnet!) not less than power-bars
5
Signal: individual/daisy-chain
Peter Göttlicher, DESY-FEB, December 12th, 2008
LDA
DIF
DIF
DIF
DIF
DIF
EUDET: with all signals star-like to LDA
The local bundle gets thick
Connector space at LDA gets large
DIF
LDA
DIF
DIF
DIF
DIF
Most in bus like cable
- more time while boot
- risk reduction with both ends to LDA or
two connectors to LDA at both ends
… Less cables, less connectors at LDA
DIF
DIF
Individual signals
Keep to minimum
One cable with multiple tails
Reduces space and access-blockage
More time in production….. Better access in ILD
6
Summarizing Comments
Peter Göttlicher, DESY-FEB, December 12th, 2008
Getting a concept conclusive for all crafts:
Need to get basics from all
Need in the early stage basic inclusions in the design
- Concept for EUDET:
Is cable count adequate for scaling to ILD?
- Power/GND:
Bar-like (thick wires) gets better EMI, less space
- Signal: The star-like concept is not the only solution
Daisy chain like reduces space requirements
Other comment: Open issue?
….. Is temperature of cooling water enough to
REGULATE stack temperature for SiPM?
CMS plans Peltier: Other solution additional local heaters
…. Every addition: More power in DIF area
7