Pixel integrated stave concepts

1
Pixel integrated stave concepts

• Valencia 2007 SLHC workshop

2
Introduction

• We now have modules on staves, how is the
  integrated stave approach different?
• What are the schemes that have been proposed?
• Common requirements for the active elements
• How much material or CHF can be saved?

3
Module on stave vs. integrated stave

• Present modules are fully functional detectors
  complete with control chip and cable
• Present stave is purely mechanical
• In an integrated stave approach electrical
  services are combined with the mechanics, and
  what is loaded on this is not a stand-alone item.
  
• Consensus emerging to move control functions to
  end of stave (stave control chip) and to load
  bare modules (all silicon flip chip assemblies
  on stave).

4
Examples from BL workshop
SP stave example (not maximally integrated)
Fully integrated stave
Monolithic structure
5
Single chip vs. multi-chip bare module

• 3D sensor processing has introduced the
  possibility of active edges as a natural
  byproduct.
• But active edges can also be implemented on
  planar sensors
• If active edges are used, multi-flip chip
  assembly is an unnecessary complication at
  non-trivial added cost
• There is no acceptance to be gained by using a
  big sensor tile vs. smaller active edge tiles
• Sensor production and flip chip are cheaper for
  single-chip assemblies than multi-chip
  assemblies, if nothing else because the yield is
  higher in both cases.

6
Shingling along Z

• Mechanical structures proposed at BL workshop all
  showed flat surfaces (maybe not intended as
  final, but nevertheless)gt no shingling along Z
• This helps thermal performance and
  manufacturability
• Active edges (or 2-sided staves) are required to
  avoid shingling and preserve full acceptance.

7


Assembly flow simplification
NOW
SIMPLER
Wafer probe
wafer probe
Bump deposition, Thin and dice
More standard and higher yield 1 to 1 bump
bonding no need for ultra high KGD yield.
Bump deposition, Thin and dice
Chip probe/select
16-to-1 Flip chip
1-to-1 flip-chip
rework
Bare module probe
Toss bad ones
SC module full probe
Module assembly
Higher level of integration at IC stage no
separate module control chip
Module full test
Stave/sector assembly
Stave/sector assembly
Stave/sector test
Stave/sector test
8
Outer layers

• 3-D sensors and therefore active edges apply
  naturally to B-Layer
• But many advantages of active edges are more
  important for outer layers, where cost is
  critical
• Cost savings from using single chip modules and
  no shingling
• Simplified assembly flow (faster and cheaper)
• Integrated stave using SC modules is therefore a
  good model for entire detector
• Readout modularity needs to be studied for outer
  layers.
• Full development of active edges even for planar
  sensors in important for pixel upgrade!

9
The integrated stave conceptual parts(examples
can be replaced by equivalent function)
End of stave card could be built-in Stave
controller chip goes here
Signal cable with 24 flaps before lamination.
Could also be multiple cables.
Al power planes added to signal cable. May
contain DC-DC converters at a few points, or
serial power taps.
Mechanical support with cooling.
Combine all the above and test. Then add SC
modules.
Tab wraps around and is w-bonded to SC module
10
End of stave card I/O
Traces on flex
15m miniature coax to PP2
End of stave card
40MHz clock
24
40MHz clock
24
Clock fanout
1GHz command
40MHz clock bar
24
2GHz data
40MHz command
command De-serializer
24
2GHz data
40MHz command bar
24
160MHz data
2 Data serializers
3
24
160MHz data bar
Muxed NTCs
4-12
4-12
HV bias groups
HV bias groups
NTC mux
LV power (eg)
Power at HV
DC-DC
12-24
Twisted pairs
11
NTCs for independent monitoring AND safety
NTC1
W
interlock
next
NTC2
A
W
Mux Sense IC
NTC3
W
NTCn
V
DCS
W
W
12
Material savings

• Items no longer used
• Module control chip
• Module pigtail connectors
• Shingled carbon surface
• Lighter items
• Power conductors
• Less massive if SP or DC-DC
• Or could tolerate larger Vdrop if DCDC placed at
  ends of stave.
• Mechanical structure
• See mechanics pixel talk Thu. New developments
  not central to integrated stave
• New items needed
• DC-DC (unless placed only at ends) or SP signal
  level shifters control.
• Want even more reductions?
• See monolithic structures tomorrow.

13
Conclusion

• Many ideas, but prototyping is needed to
  understand details.
• Good start on mechanical items (see tomorrow)
• Not much done on passive electrical components.