View
0
Download
0
Category
Preview:
Citation preview
K.K. Gan Opto-Chips Review 1
Overview of Opto-Link R&D
March 11, 2008
W. Fernando, K.K. Gan, A. Law, H.P. Kagan, R.D. Kass, J. Moore, A. Rau, S. SmithThe Ohio State University
M.R.M. Lebbai, P.L. SkubicUniversity of Oklahoma
B. Abi, F. Rizatdinova Oklahoma State University
K.K. Gan Opto-Chips Review 2
Outline
● Architecture● Transmission on micro-cables● Bandwidth of fiber● Radiation hardness of PIN arrays● Radiation hardness of VCSEL arrays● Compact MT-style opto-pack● Plan for opto-chips● Summary
K.K. Gan Opto-Chips Review 3
● G. Darbo, P. Farthouat, A. Grillo, ATL-P-EN-0001
Read Out Architect
FE
ModuleController
Super-ModuleController
OptoLink
fibersTTC
Data
M links
N links
160 Mb/s 320 MHz
320 Mb/s
3.2 Gb/s
Bypassed forinner pixel layers
160 MHz
K.K. Gan Opto-Chips Review 4
● optical link of current pixel detector is mounted on patch panels:
much reduced radiation level● use micro-twisted pairs for transmission
between pixel and opto modules
simplified the design/production of both types of modules
what is the bandwidth of the micro cables?
Transmission on Micro-Cables
K.K. Gan Opto-Chips Review 5K.K. Gan 5
Bandwidth of Micro Twisted Pairs
(current pixel cable)
● current pixel cable with thick insulation is quite optimum!
~640 Mb/s
~1 Gb/s
~320 Mb/s
K.K. Gan Opto-Chips Review 6K.K. Gan 6
Eye Diagrams100 µm current pixel cable
60 cm
● transmission at 640 Mb/s is adequate● 127 µm cable is slightly better● will investigate the use of TRT cable for longer transmission
640 Mb/s
1280 Mb/s
140 cm127 µm cable
140 cm
K.K. Gan Opto-Chips Review 7
● optical links of current pixel detector use
rad-hard/low-bandwidth SIMM fiber fusion spliced to
rad-tolerant/medium-bandwidth GRIN fiber
what is the limit of the bandwidth?
Bandwidth of Fiber
K.K. Gan Opto-Chips Review 8
Bandwidth of Fiber
● transmission at 3.2 Gb/s is adequate
✔satisfy the requirement of B-layer upgrade
■ current SLHC architecture calls for raw rate of 3.2 Gb/s
plus 20% overhead for 8b/10b encoding
more efficient encoding will improve margin of operation
2 Gb/s 3.2 Gb/s 4.25 Gb/s
11 + 80 m spliced SIMM/GRIN fiber
K.K. Gan Opto-Chips Review 9K.K. Gan 9
Radiation-Hardness of Silicon PIN
● irradiate PIN/VCSEL arrays with 24 GeV protons at CERN● PIN responsivity decreases by 3x at 114 Mrad (SLHC: 69 Mrad)● no degradation of rise/fall time
✔ operation at 160 MHz is OK
0
2
4
6
8
10
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
Pre-irradPost-irrad
Cou
nt
Responsivity (A/W)
K.K. Gan Opto-Chips Review 10
Radiation-Hardness of GaAs PIN
● all arrays are front side illuminated● PIN responsivities decrease
by ~10x at 53 Mrad● should repeat irradiation
to SLHC dosage of 34 Mrad
K.K. Gan Opto-Chips Review 11K.K. Gan 11
VCSEL LIV Characteristics
ULM requires higher voltage to operate● all arrays have very good optical power
Optowell
ULM 5GULM 10G
Pre-irrad
K.K. Gan Opto-Chips Review 12K.K. Gan 12
VCSEL LIV Characteristics
● both arrays have very good optical power
Pre-irrad
K.K. Gan Opto-Chips Review 13K.K. Gan 13
VCSEL Power vs Dosage
● Optowell & ULM (10 Gb/s) survive to SLHC dosage
2007:Two arrays each(2 x 7 channels)
1st irradiationperiod
2nd irradiationperiod
K.K. Gan Opto-Chips Review 14K.K. Gan 14
VCSEL Power vs Dosage
● AOC (5 & 10 Gb/s) survive to SLHC dosage
2007:Two arrays each(2 x 7 channels)
1st irradiationperiod
2nd irradiationperiod
K.K. Gan Opto-Chips Review 15
Annealing of VCSEL Arrays
● recovery is slow● Optowell has the highest annealed power
Two arrays:2 x 7 channels
K.K. Gan Opto-Chips Review 16
Annealing of VCSEL Arrays
● recovery is slow but adequate annealed power
K.K. Gan Opto-Chips Review 1717
Opto-Pack Development● current pixel detector uses Taiwan optical packages
VCSEL mounted on PCB with poor heat conduction
micro soldering of 250 µm leads is difficult● Ohio State develops new opto-pack for SLHC
■uses BeO base with 3D traces for efficient heat removal
■wire bond to driver/receiver chip
OSU
Taiwan
K.K. Gan Opto-Chips Review 18
Results on Opto-Packs● 35 VCSEL & 6 PIN opto-packs have been fabricated
◆ all VCSEL opto-packs except one have good coupled power principle of new opto-pack has been demonstrated
0
1000
2000
3000
1 3 5 7 9 11
Channel
Po
wer
(µW
)
V012 Optowell V013 Optowell
V002 AOC V010 AOC
V004 ULM 5Gb/s V005 ULM 5Gb/s
V006 ULM 10Gb/s V007 ULM 10Gb/s
MT ferrule
Ceramic guide pin
VCSEL array
1 cm
K.K. Gan Opto-Chips Review 19
Opto-Chips● Proposal is to submit 4 mm2 prototype chips:
◆ receiver/decoder operating at 40, 160 and 320 MHz
■ use bi-phase marked encoding due to the low speed
◆ VCSEL driver operating at 640 Mb/s and 3.2 Gb/s
◆ both designs take advantage of LHC experience
◆ SMC block: 3.2 GHz serialization clock ● Goal:
◆ gain experience with IBM 130 nm technology
◆ find out problems/limitations of design
◆test radiation-hardness/SEU in summer 2008
■submission date: Monday, March 24
■cost: $18K
K.K. Gan Opto-Chips Review 20
Relation to GBT/Versatile Link
● GBT: driver/receiver operating at several Gb/s are being designed:
◆ operate with single mode laser (1310 nm)
■ ATLAS SCT/Pixel use multi-mode laser (850 nm)
no duplication of effort● Versatile Link:
◆ proposal has been submitted to ATLAS/CMS to develop
opto-links with single channel devices
◆Versatile Link group is fully aware of the R&D effort here
via the ATLAS/CMS Joint Opto-Link Working Group
K.K. Gan Opto-Chips Review 21
Summary● much progress in last three years on R&D of various components:
◆ micro cables
◆ fusion spliced SIMM/GRIN fibers
◆ Si and GaAs PIN arrays
◆ VCSEL arrays
◆ array packaging✔ these components can be used for the upgrades and bandwidth and radiation-hardness limitations have been measured ● Questions for the Review Committee: ◆ Is it time to fabricate prototype opto-chips to learn the limitations? ◆ Will the design be ready for March 24th submission? ◆ What changes should be implemented before/after submission?
Recommended