Transcript LHCb Vertex Detector and Beetle Chip

LHCb Vertex Detector and Beetle Chip
Outline:
• Introduction VErtex LOcator
• Pile-up trigger
• Radiation hardness of electronics
• Architecture of Beetle chip
• Test beam results
• Conclusions & outlook
18 December 2002
1
Martin van Beuzekom
The LHCb Vertex Locator (VELO)
The VELO provides an accurate
reconstruction of primary &
secondary (displaced) vertices
18 December 2002
2
Martin van Beuzekom
Vertex Locator
• 21 VELO + 2 Pile-Up stations
• R/ Silicon strip detectors
• ~ 180k strips
• Typical hit resolution 8 mm
• CO2 cooling: 2 kW @-10 C
• Radiation dose 100 kGy (4 years)
• Vacuum feedthrough: 20k pins
18 December 2002
3
Martin van Beuzekom
RF shielding
•
•
•
•
•
Secondary vacuum
26 A peak current through box
300 mm Al foil
 11 mm hole for the beam !!
Detectors are retracted during beam injection
18 December 2002
4
Martin van Beuzekom
Pile-up trigger (veto)
RA ZPV - ZA
=k
=
Rb
ZPV - ZB
•
•
•
•
Detection of multiple primary vertices
R-strip detectors only
L0-trigger: prompt binary signals needed
Processing time < 2.2 ms
•
•
•
•
Logic OR of 4 detector signals in Beetle
1024 binary signals @ 80 Mbit/sec
Special (complex) hybrid
Connectivity challenge !
18 December 2002
5
  



Martin van Beuzekom
Front-end chip requirements
• 40 MHz sampling rate
• Peaking time < 20 ns
• Remainder of peak after 25 ns < 30 %
(spill over)
• S/N > 14 (300 mm Si)
• L0-trigger rate: 1.1 MHz =>
Readout time 900 ns
• Trigger latency up to 4 ms
• Buffering of 16 events
• Power consumption < 6 mW / channel
• Radiation hardness >100 kGy
18 December 2002
6
Pulse from charge
sensitive amplifier
time
Martin van Beuzekom
Radiation hardness of CMOS
Problem: Single Event Upsets (SEU) i.e. bit-flips
Cure : Triple redundancy + majority voting for all registers
Problem: Hole trapping in SiO2 (Total Ionizing Dose)
Cure: Use deep submicron process + special layout techniques
MOS transistor characteristics are determined by width (W) and length (L)
V+
Gate oxide
++++
18 December 2002
7
Martin van Beuzekom
Radiation hardness(2)
• Hole trapping => threshold shift
• Deep Sub Micron process => thinner gate oxide
• Tox < 10 nm : tunneling decreases the
amount of trapped holes in the gate-oxide
• Solution to parasitic transistors: enclosed layout
(only for nMOS)
• Guard rings around nMOS
Field
oxide
++
++
Disadvantages of enclosed layout:
• Larger area / capacitance
• No “long” transistors possible
18 December 2002
8
Martin van Beuzekom
Beetle functional diagram
Beetle chip is designed in 0.25 mm technology
Readout (4x)
Pipeline (186 deep)
Front-end (128 x)
comparator
18 December 2002
9
Martin van Beuzekom
Beetle front-end
Pulse shape depends on:
• Shaper feedback resistance (Vfs)
• Shaper current (Isha)
Noise is related to the current
in the pre-amplifier (Ipre)
Pre-amplifier:
Ipre
thermal noise
power
Shaper:
slower pulse: noise
10 ns
20 ns
18 December 2002
spillover
25 + 40 ns
10
Martin van Beuzekom
Layout of beetle
Beetle:
Designed by ASIC lab
Heidelberg and NIKHEF
• 0.25 mm CMOS technology
• Pipeline depth 186 cells
• Size (5.5 x 6.1) mm2
Front-end (128 x)
Pipeline cells
18 December 2002
11
Martin van Beuzekom
Radiation hardness test
Vout [V]
Single chips were tested with X-ray facility at CERN
time [ns]
Beetle showed full functionality up to 300 kGy (12 LHCb years):
• All digital functions worked up to 450 kGy
• Analog performance degradations are small
18 December 2002
12
Martin van Beuzekom
Hybrid & Silicon
• Micron PR02-R detector
– Thickness: 300 mm
– 2048 strips
– pitch: 40 – 92 mm
– length: 6.4 – 66.6 mm
– angular coverage 182°
• 2 layer pitch adapter
• 16 Beetle chips
– High yield
• 4 layer hybrid (NIKHEF design)
– 75 mm kapton / layer
– 17 mm copper / layer
18 December 2002
42 mm
8 mm
13
Martin van Beuzekom
Test beam setup
Trigger scintillators
120 GeV pions / muons
X-Y Si tracking station
Hybrid under test
“Continuous” beam => no fixed relation between
particle and Beetle clock signal (40 MHz)
• Readout 8 time samples of Beetle
• Measure time between trigger and next Beetle clock
time
Collected > 10 Million events in 4 days
18 December 2002
14
Martin van Beuzekom
Pulse shape analysis
Data after baseline correction & common mode noise subtraction
Noise distribution
s=1
Most probable
value
18 December 2002
15
Martin van Beuzekom
Pulse shape analysis (2)
Peak height
(Signal)
Spill over
25ns
Rise time (10 to 90 %)
18 December 2002
16
Martin van Beuzekom
S/N results
Short strips:
• Sstrip/Nstrip = 19
• Spill over = 32 %
Long strips:
• Sstrip/Nstrip = 13
• Spill over = 37 %
• The new design of the silicon sensor will have only 45  strips
=> expected S/N of long strips > 15
18 December 2002
17
Martin van Beuzekom
Conclusions
• Radiation hardness of Beetle demonstrated up to 300 kGy
• Performance degradation due to hybrid is minimal
• Pulse shape characteristics:
• Spill over < 30 %
• S/N > 15
• Beetle is compliant with all LHCb requirements
• Beetle was chosen by VELO group
• Inner tracker will also use Beetles
18 December 2002
18
Martin van Beuzekom
Outlook
• Next version of Beetle (1.2) + next
version of hybrid is planned to be tested
in the test beam of summer 2003
• Small design improvements => new
Beetle (1.3) submission in spring 2003
• The more complex Pile-Up hybrid is
currently under construction (4 times
more signals!), to be tested
18 December 2002
19
Martin van Beuzekom
18 December 2002
20
Martin van Beuzekom
18 December 2002
21
Martin van Beuzekom
Radiation effects
Total Dose effects
Ionising
Non-Ionising
Silicon
Detectors
Bulk damage
Bulk damage
Readout
chips
Hole trapping
(SiO2)
18 December 2002
Single Event effects
Upsets
Damage
Yes
Latch-up
Triple redundancy No problem
+majority voting For LHCb
22
Martin van Beuzekom
18 December 2002
23
Martin van Beuzekom