data transmission needs and challenges for frontier
TRANSCRIPT
![Page 1: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/1.jpg)
1
July 14 - 25, 2014Paris Diderot Univesity
2nd International Summer School on:INtelligent Signal Processing for FrontIEr Research and Industry
Data transmission Needs and Challenges for Frontier Particle Physics:Radiation hardness and longevity
S. HouAcademia Sinica
Taipei, Taiwan
![Page 2: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/2.jpg)
2
Outline
− Introductiondata transfer protocols, low loss fiber innovationopto-electronics, industrial development
− Optical links in collider experimentsLHC devices CDF tracker optical link, longevity
− Radiation hardness of optical links predictionstests to CDF transmittertests to VCSELs, PINs
− Radiation hardness of commercial devicesLight-coupling materialVCSEL, Control IC, BER
![Page 3: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/3.jpg)
3
Introduction:data transfer protocolslow loss fiber innovationopto-electronicsindustrial development
![Page 4: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/4.jpg)
4
Introduction: data transfer protocols
− Internet TCP (Transmission Control Protocol) : Duplex data transfer, demanding reliabilityTCP forwards “packets” of Internet Protocol over networks between hostsEncapsulated, Telnet, FTP, SMTP, HTTP, … for applications
− Computer peripherals, e.g. USB (Universal Serial Bus) : Standardize connection of computer peripherals: keyboards, hard drives ..Host centric bus with layers of protocols
− Detector data acquisition:Detector measurements are analog signals,Front-end ADC, multiplexing … real-time data output, no handshapeMay require inputs for slow control for clock, setup,
![Page 5: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/5.jpg)
5
Internet TCP− TCP (Transmission Control Protocol)
Internet communication with three way handshakesReliable with data packets ACKnoleged in SEQuence− Acknowledgements indicate delivery of data− Checksums are used to detect corrupted data.− Sequence numbers detect missing, or mis-sequenced data.− Corrupted data is retransmitted after a timeout.
SYN (synchronize sequence numbers) ISN (initial sequence number)
![Page 6: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/6.jpg)
6
Computer network with USB ─ USB (Universal Serial Bus ) “Single Master + Multiple Slaves”
One interface for many devicesHot pluggable, automatic configuration Hi-Speed – 4.8 Gbps (USB3)
─ USB transaction ,strictly defined frame, error checking and handshaking
- Token Packet, header defining what to follow- Optional Data Packet- Status Packet, to acknowledge transactions
![Page 7: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/7.jpg)
7
NTT Technical Review
Fiber capacity
Speeds and capacities of optical core, access networks are being boosted
![Page 8: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/8.jpg)
8
Fiber optical lossSilica glass fiber:Low-mass, high bandwidth, little loss over long distance− Absorption Scattering losses (Rayleigh, Raman…)− Bending losses (micro bending)
1st generation, ~1975, 0.8 μmfiber-MM, Laser GaAs2nd generation, ~1980, 1.3 μm, fiber MM SM, InGaAsP FP-laser3rd generation, ~1985, 1.55 μm, fiber SM InGaAsP DFB-laser4th generation, 1996, 1.55μm WDM s
![Page 9: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/9.jpg)
9
Opto-electronics
850-nm VCSELs - for short-distance over multimode fiber1300-nm VCSELs - for longer-distance communication over single-mode fiber 1550-nm VCSELs - tunable sources for DWDM multiple transmissions over one fiber
− Transceivers:Laser diodes as light sourcePIN diodes as photo detectorDriver chip convert electrical signals to optical, and vice versaControl IC for the application protocols
![Page 10: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/10.jpg)
10
Light sourcesSpontaneous light emission Laser Diodes (LD), Light Emitting Diodes (LED) LEDLD has an optical cavity for stimulated emission. Light Amplification by Stimulated Emission of Radiation (LASER)
![Page 11: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/11.jpg)
11
Photo-detectors
− Photo-detector (PD)converts incident light pulse into electrical signal- PIN Photo-detector (most common PD in use) - Avalanche Photo-diode (APD)
− PIN diode is operated in the reverse-bias mode- wide depletion region to create electron-hole pairs- low junction capacitance allows for very fast switching
![Page 12: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/12.jpg)
12
Building an eye pattern
Deterministic jitter:− caused by dispersion in fiber, and− inadequate bandwidth of transceiver components
Random jitter:− caused by thermal noise, shot noise in components
Non-Return to Zero (NRZ)
Eye Diagram on oscilloscope triggered by data clock.
![Page 13: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/13.jpg)
13
Commercial optical transceiverstelecommunication and computing
10 Gb/s SFP+ transceivers4.8 Gb/s USB3 Active Optical Cable
10 GB/s SFP+
5 Gbps USB3
![Page 14: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/14.jpg)
14
Dense parallel devices− Commercial dense parallel devices− 12 ch Transceivers
MTP/MPO connector Pigtailed
− 4 ch Transceivers, 12 ch transmitter
Pigtailed
![Page 15: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/15.jpg)
15
Chip-on-Board, QSFP of TrueLight− Assembly service to customer
VCSEL, PIN arrays of TrueLightLens, driver, circuit provided by customer,MT ferrule, 4-in 4-out, 40 Gb QSFP optical engine
TX 10 Gb tests
RX 10 Gb tests
![Page 16: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/16.jpg)
16
Light Peak Technology− Light Peak, Intel technology delivers high bandwidth starting at 10 Gb/s to
mainstream computing and consumer electronics
− Lens/Prism : precision PEI molding ..
VCSELs aperturefrom t Ø 5 to 20 µm
Full acceptance angle25.4 degree
Spherical-aberration freePlano-Convex Hyperbolic Lens
Divergence angle26 – 32 degree
(FOCI prism)
![Page 17: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/17.jpg)
17
Optical driver, USB-3, 5 Gbps
![Page 18: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/18.jpg)
18
FOCI optical engine functional diagram
VCSEL/PIN: • 850 nm bare die, • 4.8 Gb/s or 10 Gb/s• >0 dBm (1mW)
Optical IC: • VIA Labs V0510, • TSMC 90nm technology• USB-3 protocol, ~60 mW, • 4.8 Gb/s TX/RX driver
+ regulator/controller
![Page 19: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/19.jpg)
19
Bit Error Rate test
![Page 20: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/20.jpg)
20
Bit Error Rate test
![Page 21: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/21.jpg)
21
Optical links in collider experimentsLHC devices CDF tracker optical link, longevity
![Page 22: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/22.jpg)
22
2. ATLAS SCT transceiver
Opto-flex: Holds VDCs, DORICs, PINs, VCSELsVDC: VCSEL Driver CircuitDORIC: Digital Optical Receiver Integrated Circuit
− Detector data acquisition:Analog signals digitization (front-end IC, ADC)
data transferFront-end INPUT clock, IC configurationFront-end OUTPUT continuous data transfer 1 RX PIN + DORIC for clock+command
2 TX VCSEL + VDC for data45o fibre end, mirror to VCSEL, PIN face
![Page 23: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/23.jpg)
23
ATLAS LAr MTX
TWEPP2012, Liu, SMU
Similar to CERN VL product, TOSA + MM fiber, 850nm Use LOCld, driver of SOS process
Speed 8 Gb/s Different geo/connector configuration
constrain by LAr geometry
![Page 24: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/24.jpg)
24
High throughput array
OSU prototype, ATLAS PixelVCSEL array, driver of 130 nm CMOS. 8 channel + 4 spare channels
TWEPP2012, Gan, OSU
![Page 25: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/25.jpg)
25
CDF optical linkThe 1st optical link in Collider experiment inner tracker
Dense Optical Interface ModuleByte-wide parallel link
8-bits + clock53 Mbyte/sec, BER<10-12
Transmitter on portcard :Laser-diode arrayASIC driver chip
Receiver on FIB crate :PIN-diode arrayASIC receiver chip
Multi-mode fiber ribbon
![Page 26: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/26.jpg)
26
CDF tracker Port Card
− Total 128 portcards5 optical TX each portcardtotal 570 TX
− Transceiver : convert low-voltage differential signals from DAQ to Silicon detector
− Digital Data Receiver :Decode 5-bit commands into 10 SVX3D control and calibration
− Analog DDR: regulates AVDD for clean frontend operation
− Optical transmitter :8 data bit + 1 data valid signal to Laser Diode Array to 9-bit parallel optical link
![Page 27: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/27.jpg)
27
Silicon Detector Installation
![Page 28: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/28.jpg)
28
CDF optical link, transmitter assemblyDie-bond / Wire bondlaser-diode array on BeO submountdriver chip on substratefibers on V-groove
Alignmentfibers to laser emitting facets
![Page 29: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/29.jpg)
29
CDF optical link, transmitter electronics
InGaAs/InP Edge-emitting laser diode array1550 nm wavelength12-ch diode array, 250 µm pitchBare die power, >1 mW/ch @20mAInsertion to fiber: 200 ~ 800 µW/chCustom made by Chunghwa Telecom
biCMOS ASIC driverbipolar transistors, AMS 0.8 µm process9-channelsInputs: Diff. ECL or LVDS
differential >100 mVEnable by TTL lowOutput light: adjustable by
~2mA/0.1VAt Vcc-VLD=3V, 20mA/ch
![Page 30: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/30.jpg)
30
CDF optical link, receiver module
Die-bond / Wire bondPIN-diode array on Al2O3 submountdriver chip on substratefibers on V-groove to PIN diodes
InGaAs/InP PIN diode 12-ch array, 1550 nmby TL, Chunghwa Telecom.Operation condition :light on: 50 ~ 800 µW light off: <10 µW<1.1 W/moduleOutputs :9 independent diff. ECL
![Page 31: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/31.jpg)
31
CDF optical transmitter test
Inputs: − ECL or LVDS signal− TTL-enable
Light by O/E probe
Input ECL
TTL enable
![Page 32: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/32.jpg)
32
CDF edge-emitting laser, characteristics
Laser light at I,V and TemperatureI-V approximately linearDuty cyclestable output to input 50%Linear to temperature
Temp (oC)
![Page 33: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/33.jpg)
33
CDF optical transmitter production uniformity
Quality analysislight measured from pigtails at 30oCwide deviation channel-by-channelmainly due to insertion efficiency
Span within ~400 µWσ ~72 µW to the mean/module
![Page 34: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/34.jpg)
34
CDF optical receiver responses
Receiver connected to a Transmitter Light power chosen forwide distribution
Light pulse width are consistent
Receiver ECL outputsby a Tektronix diff. probe
Consistent duty cycles in favored operation range (2.8~3.2V)
Saturates for high light level
![Page 35: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/35.jpg)
35
CDF optical link, Bit-Error-Rate test
BERT by FermilabPC ISA bus interfaceTbert, Rbert - test boards
At 63 MHz, minimum BER <10 –12
Burn-in 3-days on ASICs, diodes1-day BERT
⇒ reject devices infant mortalitybad components fail quickly
![Page 36: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/36.jpg)
36
Optical transmitter, accelerated ageing
Transmitters biased at 60oC for 330 days
Accelerating factor , Arrhenius eq.
Ea= 0.4 eV for GaAsF=15 to T1= +6oC
Wear-out degradation~10% at 60oC, no failurecorresponding to CDF operation at +6oCfor 15 years
![Page 37: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/37.jpg)
37
CDF optical link longevity
− Commissioning was problematic:~ a dozen bit-error channels temperature from Room to 6oC
increased light power wider width NRZ bit locku
loosened connectors permanent loss
− Optical link bit-errorafter commissioning ~ 2%
− 10 years operationageing, radiation damage outer tracker (ISL)
stay at ~2% inner most (SVX-II)
increase to a total ~ 4%
Fraction of silicon ladders
![Page 38: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/38.jpg)
38
Radiation hardnessof optical links PredictionsTests to CDF transmitterTests to VCSELs, PINs
![Page 39: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/39.jpg)
39
Radiation damage− Interaction of radiation with material
Gamma Total Ionizing Dose (TID), SEUCharge hadron, neutron NIEL, degradation, SEU
− Ionization:charge trapped in oxide or at interface
− Displacement: (NIEL, non-ionizing energy loss )atoms can be removed from their initial crystal lattice positions defect in components degradation, annealing
− Single event effect (SEE)transient, corrupted bits,IC functional interrupt
![Page 40: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/40.jpg)
40
NEIL in Si damage function
− Hypothesis: Damage parameters scale with the NIEL1 MeV neutron equivalent damage
RD50
![Page 41: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/41.jpg)
41
NIEL of Proton in GaAs
− Proton damage to GaAs LED, Solar Celldoes not agree with beam tests
J.H. Warner et al.,IEEE TNS 51, 2887 (2004) C. Poivey, EPFL Space Center 9th , Jun 2009
![Page 42: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/42.jpg)
42
Radiation Field in ATLAS
− Exposure to radiation from pp collisionsmainly pionsneutrons (hadrons interact with material)
− 10 years operationTID > 100 kGyNIEL φeq > 1015 1 MeV n/cm-2
![Page 43: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/43.jpg)
43
Radiation fields @LHC, SLHC
Fluence expressed for 1 MeV neutron equivalentof the NIEL (Non-Ionizing Energy Loss) calculations NIEL factors for Neutrons [MeV] 1 20GaAs [keV cm2/g] 0.55 2.3Si [keV cm2/g] 1.8NIEL factors for Protons [MeV] 30 70 200 GaAs [keV cm2/g] 4.03 3.64 3.93 Si [keV cm2/g] 4.78 3.16 1.94
[Summers, IEEE NS. 40 1372 (1993)]
Radiation level @LHC/SLHCRadiation dose after 7.3 years
LHC @1034 cm-2sec-1
1.E+13
1.E+14
1.E+15
1.E+16
1.E+17
0 20 40 60
Radius [cm]
Flue
nce
in 1
MeV
n/c
m2
Si
GaAs
5 years SLHC @ 10^35 cm-2 sec-1 with a safety factor of 2
1.0E+14
1.0E+15
1.0E+16
1.0E+17
1.0E+18
0 20 40 60
Radius [cm]
1 MeV
neut
ron F
luenc
e [n
/cm^2
]SiGaAs
![Page 44: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/44.jpg)
44
CDF Radiation Measurement
Radiation from pp̄ CollisionsTLD measurements + model
r = radius to the beamα ~1.5 in barrel region
Locations of TLDs(Thermal Luminescent Dosimeters)
R.J Tesarek IEEE NSS 2003
αrAdose =
![Page 45: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/45.jpg)
45
CDF Radiation estimation
By MC for distributions of particles inpp̄ Collisions
Dose to fluence conversion:1 Gy = 100 Rad = 3.8x109 MIPs/cm2
(Silicon tracker is specified for 2 MRad)
Radiation tolerance:Optical Transmitter required for 200 kRad 7.6x12 MIPs/cm2
Dose rate at the optical Transmitter r=15 cm, dose rate = 17 Rad/pb-1
200 kRad 11.8 fb-1 in the beam
Team tests with fixed energy hadronsdose scaled by dE/dx1 Gy = 2.1x109 p(200MeV)/cm2
= 5.4x108 p( 30MeV)/cm2
![Page 46: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/46.jpg)
46
Radiation tolerance of laser diodes
− Displacement, NEIL Light degradation, partial recovery by annealingIncreasing threshold current
− Edge-emitting Fabry-Peror laserActive optical cavity, typically~500x50x10 μm3
− VCSELvery small active volume~∅20x10 μm3
very high Rad-hardness
![Page 47: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/47.jpg)
47
Radiation damage to PIN photo-detectors
− Displacement, NEIL Increase in leakage Ileakdecrease in responsivity Iphoto
− Damage pattern variesSi Epitaxial PINGaAs PIN
− Faster PIN smaller device,
Higher rad-hardness
![Page 48: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/48.jpg)
48
Proton beam tests on CDF transmitterCyclotron 30 MeV protons at INERMeasuring L, temperature at DC modefiber connection out of beam area
![Page 49: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/49.jpg)
49
Proton irradiation to CDF transmitter
Flux rate dependence not obvious− IUCF 200 MeV protons at 3 flux rate− Pulsed by 50% duty cycle signals
flux- = 3.4 E9, 11.5 E9, 29.2 E9/cm2secfluence = 4.3 E12, 12.8 E12, 30.0 E12/cm2
CDF edge-emitting type laser
Slow annealing after − IUCF 200 MeV protons − Transmitter at DC mode− Total fluence 1.2E13 /cm2
− 8 hours annealing
↔8hrsannealing
![Page 50: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/50.jpg)
50
Proton beam tests to CDF transmitterIUCF 200 MeV protonECL 25MHz inputsO/E probing for biCMOS driver wave form
Before irradiation
@ 3.0x1013
(1.4 Mrad GaAs1.8 Mrad Si)
![Page 51: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/51.jpg)
51
Rad-hardness of VCSELs and PINsVCSEL and PINs tested for LHC− Louvain-la-Neuve CRC:
neutron 20 MeV (av) of deutron on Be target− Tohoku CYRIC: proton 30, 70 MeV− Indianna IUCF: proton 200 MeV− INER: proton 30 MeV− Los Alamos: neutron
![Page 52: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/52.jpg)
52
VCSEL characteristics
L-I of VCSEL array
L-T of a VCSEL array Temp. on circuit board
Two types of GaAs VCSEL (850 nm)− Proton implant VCSEL (ATLAS on detector)− Oxide confined VCSEL (off-detector on ROD)Thin active layer 10 μm, very rad-hardLittle temperature dependence
VCSELDAQ
![Page 53: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/53.jpg)
53
VCSEL degradation, annealing
L-I of VCSEL (oxide)vs. online Fluence
L-I of VCSEL (oxide)vs. Annealing time
VCSEL light degradation linear to fluencerad-hard fiber connected to readout, independent of flux rate
Fast annealing by charge injectionat operation current (10 nA) applied
![Page 54: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/54.jpg)
54
VCSEL annealing in time
Charge injection at at the operation 10 nAf(t)= f∞ - a⋅ exp( -t/ τ )
recovery time τ ~ 5 hours
![Page 55: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/55.jpg)
55
VCSEL rad-hardness summary VCSEL light degradation in proton radiation
(ATLAS Oxide confined)deviates from the NIEL calculation
GaAs solar cellSrour, IEEE TNS 50, 653 (2003)
![Page 56: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/56.jpg)
56
Proton damage to Epi Si PIN, degradation
IUCF 200 MeV proton, to 4x1014 p/cm2
DC Laser light to Epi PINs Online current measurement Responsivity
dropping fast in early fluence
![Page 57: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/57.jpg)
57
Proton damage to PIN, wave-form
Wave form at full depletion, after 4x1014 (200 MeV) p/cm2
rise/fall time < 1 ns (20%-80%)
beforeTruelight PINafter
before Centronic PINafter
![Page 58: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/58.jpg)
58
Proton damage to PIN, orientation matters PIN face parallel to beam twice damage
![Page 59: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/59.jpg)
59
Forward bias (30 MeV p)
PIN responsivity, dark current Tohoku 30 MeV proton Control beam on/off
LED light on/off Online PIN VR scan
beam on, beam off Each strip is a 0-20V scanBeam off: + LED on + LED offBeam on: + LED on + LED off
ATLAS Truelight Epi Si PIN
LED onLEDoff
Quick recovery expelling dark current
![Page 60: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/60.jpg)
60
PIN proton damage, summary
PIN rad-hardness diameter, thickness A/W, speed & rad-hard
Proton energy dependenceSi PIN : compatible with
30 MeV and 70 MeV protonsGaAs PIN : twice damage by 30 MeV
than 70 MeV protons
![Page 61: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/61.jpg)
61
Radiation hardnessof commercial devicesUSB, SFP+ devices of >5 GbpsLight-coupling materialVCSEL, Control IC, BER
![Page 62: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/62.jpg)
62
Rad-hardness of USB Optical Enginelight-coupling material
− Light coupling prism receptacleMaterial: PEI (polyetherimide), as for the TOSA/ROSA tipoptical quality surface
− Co60 gamma ray, Total-Ionizing-Doseat INER Taiwanflux: 3.5 kGy/hr, total: 117 kGy NO LOSS !!for light transmissionwithin the 2% systematic error
![Page 63: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/63.jpg)
63
Proton to VCSEL, Optical IC− Proton Irradiation
INER 30 MeV proton beamflux of 3.5x1010 p/cm2s, to a total 1.2x1014
equivalent to 8.9x1014 n(1MeV)/cm2s
− Irradiation measurementFOCI module DC biasedVCSEL mid-level DC light
![Page 64: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/64.jpg)
64
LANSCE neutron test
USB transceiver
Fiber SMA cables
− Beam profile similar to ATLAS − USB transceiver in Bit-Err-Rate
Straitix II GX, PRBS 27-1 bit patternTX path, RX path tested separately
neutron flux 2.9×105 n/cm2sover 1.5 days to 3.8×1010 n/cm2
TX: 0 error, upper limit 1.0×10-10 cm2/ch (95% CL)RX: 11 errors SEE cross section 2.9×10-10 cm2/ch
![Page 65: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/65.jpg)
65
Rad-hardness of Transceiver IC
− Sub micron CMOS process is rad-hard− Avoid vulnerable components
− USB optical chip runs well under irradiationpower-cycle failed initialization OTP (one-time programmable memory corrupted)
Configuration corruptedafter irradiation
![Page 66: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/66.jpg)
66
Total Ionizing Dose to Commercial driver − QSFP, miniPOD, PPOD, ONET8501V, ONET1101L
tested with X-ray or γ-ray, none meet the ATLAS LAr radiation requirement. − Kintex 7, ONET8501
tested with a neutrons in Los Alamos SEU rate of Kintex 7 is too high for LAr.
Vendor Part# Gbps # ch Rad type (krad/hr) TID (krad)
QSFP Avago AFBR-79EIDZ 10 4 60Co γ 10 75
miniPOD Avago AFBR-810FN1Z 10 1 x-ray 360 66
PPOD Avago AFBR-810EPZ 10 12 x-ray 360 150
VCSEL driver TI ONET8501V 10 1 x-ray 39 178
F-P laser driver TI ONET1101L 10 1
x-ray 9.6 46460Co γ 10 < 900
Vendor Part# # of ch Flux (n/cm2/s)
Fluence(n/cm2) # errors σ (cm2)
Kintex-7 Xilinx XC7K325TFFG900
16 (2 tested) 4.6E5 2.1E11 4/4
(2 shared) 1.6E-11
VCSEL driver TI ONET8501V 1 4.6E5 2.1E11 0 < 5E-12
SMU, TWEPP2012
![Page 67: Data transmission Needs and Challenges for Frontier](https://reader030.vdocument.in/reader030/viewer/2022020916/61adc99b9a7e9171ba42d2b9/html5/thumbnails/67.jpg)
67
Summary
− Fiber-optics advancing rapidly for IT industry
New commercial devices are compact and faster
Rad-hard drivers to be developed HEP applications
Quality Control, Rad-hard tests for lifetime assurance