Transcript ye_5-4_LAr_SMU

Optical link for LAr upgrade readout
Objectives:
1.
Evaluate the 0.25 m Silicon on Sapphire (SOS) technology (Peregrine
Semiconductor) for the LOC ASIC (done in 2006-2007)
2.
Design the link-on-chip (LOC) ASIC version 1 (LOC1), a “prove of principle” chip.
(done in 2006 - 2007)
3.
Test LOC1. (in progress for 2007)
4.
Design and test LOC2, a user chip. (2008 - 2009)
5.
Irradiation evaluation on link components: VCSELs, fibers (in progress in 2007)
Manpower:
One FTE, two RAs
in a team of three FTEs, two RAs and two faculty members.
J.Ye @ SMU
Report: R&D work for LAr upgrade
May 4, 2007 @ UCSC
Evaluation of the 0.25 m SOS technology (1)
1.
Three gamma (Co-60) tests for total dose effect (TID) and one proton test
(for SEE and TID) were carried out on a dedicated test chip.
Ring
Osc. #1
Shift
Register #1
Shift
Shift
Register
Register #2
#3
Ring Oscillator #2
Shift Register #4
Resistors
Resistors
Image of the dedicated test chip.
• The transistor array is used for TID test.
• The shift register is used for SEE test.
• PLL components, ring oscillator are used to
measure parameters for LOC1 design.
• Peregrine offered 50% discount of $40k, as
a $20k in-kind support to this project. We
share all results up to 100 krad with them.
Transistors Array
PLL components
2.
3.
4.
J.Ye @ SMU
The gamma TID tests were carried out at BNL (many thanks to Jim Kierstead
and Francesco) on the transistor array which contains 48 NMOS, 48 PMOS,
with 4 types of different layouts, to study the leakage current and threshold
voltage change in ionizing irradiation.
The proton (230 MeV) SEE test was carried out at MGH on the shift
registers.
Results submitted to NSREC, RADECS and NASA symposiums.
Report: R&D work for LAr upgrade
May 4, 2007 @ UCSC
Evaluation of the 0.25 m SOS technology (2)
5.
Results from gamma (Co-60) TID tests.
o
o
Two types of tests were carried out:
•
With the sapphire substrate floating to study leakage current sources.
•
With the substrate grounded during irradiation, a way we’ve found to control
the leakage current and threshold voltage change in radiation.
Leakage current sources:
NMOS
o
PMOS
Back channel leakage has always been a concern in SOS and SOI technology. With
the grounded substrate, the leakage current increase and threshold voltage change
in our test chip during irradiation up to 4 Mrad are very small and not a concern.
NMOS
PMOS
 This 0.25 m SOS technology is very rad-hard.
J.Ye @ SMU
Back channel leakage
contribution is 70% in NMOS,
15% in PMOS. Side leakage can
be mitigated with layout
techniques (like the ELT), Back
channel has been controlled
through expensive fabrication
process.
Report: R&D work for LAr upgrade
May 4, 2007 @ UCSC
With a grounded substrate,
the leakage current increase
is controlled below 250 nA.
Threshold voltage increase is
small and simulated to be
within fabrication variations.
A mechanism is proposed,
under study with GEANT. A
provisional patent has been
filed from SMU.
Evaluation of the 0.25 m SOS technology (3)
6.
Proton test
o
o
o
Source: 230 MeV proton, 1.9×1013 proton/cm2
Tests: TID up to 106 Mrad (Si); SEE cross section (probability).
Results:
•
TID:  survives the TID test of 106 Mrad(Si)
Vcontrol(V)
•
J.Ye @ SMU
SEE:  very small SEE cross section.
Test Element Type
Fluence(proton/cm2)
Error Count #
Cross section(cm 2)
Std Shift Register
18×1012
0
<5.6×10-13
ELT Shift Register
18×1012
0
<5.6×10-13
Res. Hard Shift Register
18×1012
0
<5.6×10-13
SET free logic latch
18×1012
0
<5.6×10-13
Report: R&D work for LAr upgrade
May 4, 2007 @ UCSC
Tests on fiber and VCSEL
The Fiber:
o
o
o
o
o
Infinicor SX+ 50/250m/1.6mm MM
10G fiber from Corning. Germanium doped.
Tested wit Gamma (Co-60) and
Proton (230 MeV, 1.9×1013 proton/cm2).
Very small light loss at low flux (dose rate).
Big loss at high flux but anneals
very quickly (within 1 hour) back.
Very promising for LHC upgrade.
More tests with gamma needed.
The VCSEL:
o
o
o
o
o
Two HFE6192-562 (10G LC w/ 50 ohm flex) from Finisar tested .
Irradiated with 230 MeV proton, 1.9×1013 proton/cm2.
The VCSELs are biased during irradiations.
Eye diagram – see plots and table.
Looks very promising but more tests needed.
Before irradiation
After irradiation
Rise/fall
time (ps)
O-power
(W)
Rise/fall
time (ps)
O-power
(W)
L1
114/130
431
110/128
133
L2
120/132
450
122/132
295
VCSEL
J.Ye @ SMU
Report: R&D work for LAr upgrade
May 4, 2007 @ UCSC
L1, before irradiation
L1, after irradiation
LOC1 design and submission
1.
Status:
o
o
2.
Prototype chip with the clock
unit (PLL), serializer, laser
driver, designed and
submitted in Feb. 2007. Chip
will be back to SMU May 16,
2007.
One self-biasing PLL
(1.25GHz), one LC PLL (4GHz)
are also placed on this chip
for irradiation tests, for
component level studies for
future higher speed designs.
Goal:
o
o
Evaluate the LOC design,
perform in-lab tests, and
rad-hard tests at system
level.
Provide a test chip to work
on fiber attachment at chip
level.
3.
Spec:
4.
Submission cost:
J.Ye @ SMU
o
2.5 Gbps – 3.125 Gbps.
$34k. Peregrine offered 15%
discount from $40k, as an in-kind
$6k contribution to this program
that they are also interested in.
Report: R&D work for LAr upgrade
May 4, 2007 @ UCSC
Conclusions on the SOS evaluation and the LOC1 design
SOS technology evaluation:
1.
2.
Peregrine 0.25 m silicon-on-sapphire technology is rad-hard for LAr
readout upgrade ASIC development.
SEE cross section based on shift registers is found to be less than 5.6×10-12
for 230 MeV protons.
LOC1 ASIC design and submission:
1.
2.
3.
A prototype serializer ASIC, the LOC1, has been designed and submitted.
The design speed of this ASIC is 2.5 – 3.125 Gbps.
LOC1 is not a user chip, it is only meant for the “proof of principle” for key
components and their integration in the LOC design concept.
LC-based PLL with a central frequency of 4 GHz has been implemented in the
same submission for future higher speed designs.
Other components in the link:
We have preliminary candidates for the VCSEL and the fiber.
J.Ye @ SMU
Report: R&D work for LAr upgrade
May 4, 2007 @ UCSC
Plans for 2007 and 2008
1.
In-lab and irradiation tests on LOC1 (2007)
o
o
2.
3.
May – August 2007.
•
Irradiation tests (gamma, proton) on LOC1, including data analysis:
•
Preliminary reliability studies may also be carried out.
Sept. – December 2007.
Sept – March 2008.
Fiber attachment studies (2007 – 2008):
o
o
o
Three attachment schemes will be studied, including flip-chip bonding.
For serial data speeds around 3 Gbps, industrial standard packaged VCSEL module is
preferred for fiber coupling.
Attracted local small companies to submit SBIR proposals on fiber coupling studies.
Design and evaluation of LOC2 (2008 – 2009):
o
o
o
o
J.Ye @ SMU
Preparation for LOC1 tests:
•
design PCBs to carry LOC1.
•
incorporate the board in our existing test systems already developed to test
the GOL.
•
Change VHDL code to record more information in irradiation tests.
Plan:
•
In-lab function tests on LOC1:
Design of LOC2 will start after we finish testing LOC1.
LOC2 will be a user chip. Close interaction with the “users” will be needed to define
the interface and functionalities. At that time, we will need to address issues like
the reference clock frequency, jitter and data bus width, signal standards (CMOS
or LVDS), etc. We expect a much longer design period for LOC2 than for LOC1
because of these issues.
Chip packaging issue will be addressed with LOC2.
LOC2 will be fully evaluated in a link system, in radiation environment. Reliability
tests will be carried out with packaged chips.
Report: R&D work for LAr upgrade
May 4, 2007 @ UCSC
Budget (for 2008)
Scenario one, LOC2 design and submission in FY 2008:
Manpower (1)
One FTE @ $55k/yr, w/ 25% benefit
$68.75k
Two RA EE students,
Manpower (2) $1,500/mon/student. For two:
$36k/yr, w/ 32% tuition benefit.
$47.5k
M&S (1)
LOC2 submission
$40k
M&S (2)
PCB fab, flip-chip/wire bonding,
components for LOC2 test
$15k
Travel
For irradiation tests only
$5k
Others
Reliability test
$0
Total
With SMU 45.5% overhead
$256.4k
J.Ye @ SMU
Report: R&D work for LAr upgrade
May 4, 2007 @ UCSC
Budget (for 2008)
Scenario two, LOC2 design in 2008 but submit for fabrication in FY 2009:
Manpower (1)
One FTE @ $55k/yr, w/ 25% benefit
$68.75k
Two RA EE students,
Manpower (2) $1,500/mon/student. For two:
$36k/yr, w/ 32% tuition benefit.
$47.5k
Total
$169.1k
J.Ye @ SMU
With SMU 45.5% overhead
Report: R&D work for LAr upgrade
May 4, 2007 @ UCSC