Title: GLAST Proposal Review
1GLAST Large Area Telescope Tracker
Subsystem WBS 4.1.4 Highlights in Progress Since
the January Review Robert Johnson Santa Cruz
Institute for Particle Physics University of
California at Santa Cruz Tracker Subsystem
Manager johnson_at_scipp.ucsc.edu
2Outline
- Overview
- Results from January Review
- Action Item Status
- Changes Since January Review
- Hardware update since January
- Schedule and Cost
- Summary
3Tracker Overview
- 16 layers of tungsten converter foils.
- 12 layers of 3 X0 converters
- Followed by four 18 layers
- x-y Si-strip detector pair closely following each
converter foil. - Two additional pairs at the bottom are needed for
triggering. - 19 stiff composite tray panels support SSDs on
both faces with electronics on two sides. - Converters are on the bottom face, just above the
SSD plane - 2-mm gap between trays
- Carbon-fiber sidewalls conduct heat to the base
and stiffen the module. - Electronics are based on 2 ASICs, PC boards, and
custom flex cables. - 31.6 kg mass per module.
- 10.5 W of power per module.
Multi-Chip Electronics Module (MCM)
19 Carbon-Fiber Tray Panels
Carbon-Fiber Wall
2 mm gap
Readout Cable
4Tracker Production Overview
Module Structure (walls, flexures,
thermal-gasket, fasteners) Engineering SLAC,
Hytec Procurement SLAC
SSD Procurement, Testing Japan, Italy, SLAC
SSD Ladder Assembly Italy
10,368
Tracker Module Assembly and Test Italy
2592
18
Tray Assembly and Test Italy
342
342
Electronics Design, Fabrication Test UCSC, SLAC
Composite Panel Converters Engineering SLAC,
Hytec, and Italy Procurement Italy
648
Cable Plant UCSC
5Results from January PDR/Baseline Review
- The Tracker has an Experienced Project Team .
- The concept of the Tracker is well matched to the
science goals and utilizes mature technology. - The design is well thought out and can be
implemented within the available time. - A strong consortium of groups in Italy has taken
the responsibility for assembling and testing all
the ladders, trays, and towers. This is a crucial
contribution to the Tracker which is being
executed in a very effective and competent
manner. - Baseline the Tracker with Increased Contingency
6Action Item Status
Action Item Response
1 Baseline the Tracker with Increased Contingency The current basis for estimating Project contingency needed uses 25 of the Tracker cost to go VS 18 in January
2 Thoroughly evaluate pre-production ICs Agreed Plan in place
3 Refine Assembly and Test Procedures Agreed In process
7Changes Since January 02 Review
- The schedule has been revised to accommodate the
overall LAT schedule extension. - Added time for additional ASIC submissions.
- Incorporated a more conservative schedule for
Tracker module production with the Italian
groups. - Critical Deliveries
- Modules 3, 4 available March 29, 2004, with 11
weeks of float. - Modules 15, 16 available August 4, 2004, with 7
weeks of float. - Budget
- Estimated cost at completion increased 180K
- MOA with Italy
- INFN ready to sign and has been providing the
requested funding and covering for the delay in
ASI approval. - ASI has not yet signed but has recently begun
funding Tracker work with approval of 800k for
FY2002.
8Prototype Tower Module
- Mechanical prototype loaded with dummy SSDs.
- Includes flexure interface to the Grid.
- Used to test and evaluate assembly methods.
- Assembly worked well, but tray tolerances were
not quite to specification. Tray assembly
tooling has since been revised and specifications
met. - Vibration and Thermal testing
- Single panel LAT-TD-759. ?
- Grid-to-tower interface thermal cycle report in
progress. ? - Random vibration of the full module LAT-TD-788.
Problems (see following)
Composite trays and mass models of the prototype
tower, before mounting of the sidewalls.
9Vibration Test Problem 1
- Some screws attaching the sidewalls to the bottom
tray backed out during transverse-axis
qualification-level random vibration. This
problem was solved (as verified in a retest) - Replace hex heads by Torx-Plus to enable
tightening to the full specification without
stripping the heads. - Double the number of screws in the non-MCM sides
of the bottom tray. - Add aluminum washer-inserts to the sidewalls for
these fasteners. - We also plan to use a secondary locking mechanism
for the screws in the final assembly, as extra
insurance.
Full module instrumented for thrust-axis vibration
10Vibration Test Problem 2
- In the first test, a hairline crack formed in the
bottom-tray closeout between the edge and a
corner-flexure fastener. - In the second test suite at transverse-axis
qualification levels the cracking increased and
the problem propagated to all 4 corners, breaking
loose the corner-flexure mount inserts. (The
mid-span flexures all held fast.)
- Remediation in process
- Analysis to make sure we understand the failure
and the cure. - Remove light-weighting cutouts in the bottom tray
and increase its height by 5 mm. - Reinforce each corner with a bonded metal
bracket. - Remediation Plan presented to Project ART.
Cracking as seen after the second test suite.
11Tracker ASICs
- The second run of the GTRC and GTFE ASICs needed
to be resubmitted - GTRC V2
- Missing rule in the Tanner DRC led to a short in
the memory. - Although complete chip was simulated with
Synopsys Timemill, clock skew margins were not
sufficient for the IC to function at nominal
supply voltage (Tanner does not extract the clock
network as a distributed RC network). - GTFE Vf a minor modification was not verified by
simulation due to misunderstanding, missing an
error that removed the baseline restoration. - GTFE Ve the same missing DRC rule led to a
nonfunctional analog chain. - Remediation
- Careful crosscheck of the DRC rule-set against
the Agilent vendor rules and use of the MOSIS DRC
checking service. - Improved pre-submission procedural control.
- Controlled row-wise clock routing in the GTRC
core. - Submission of a back-up GTRC version without RAM.
12Tracker ASICs
- The digital section does function in the GTFE Ve
Vf chips, allowing substantial testing to
proceed - Probe-card testing of the GTFE using the full set
of test vectors developed according to
LAT-TD-247. - Testing on the mini-MCM using GTRC V1 chips.
- Several problems were addressed in the versions
in production - Proper response in the GTFE to undefined
op-codes. - Logic errors in GTFE register loading.
- Protection resistors in the I/O pads led to
unreliable operation at the full 20 MHz. - Clock skew issues in the GTRC.
- DRC errors and the GTFE Vf comparator bug.
- Chips presently in fabrication (Due August 22)
- GTRC with custom RAM and GTRC with Flip-Flop RAM
(both usable for flight). - AC and DC coupled GTFE versions.
13Electronics Testing
- Several mini-MCMs are under test at UCSC and
SLAC, using the V1 GTRC and Vf GTFE chips. - Using COM card, interface board, and Python
script at UCSC, to develop the test station for
MCM production. - Using 2 COM cards, TEM, and SCL script at SLAC to
debug the ultimate readout scheme. - Mini-MCM has been interfaced with flight like
Tower Electronics Module(TEM) to verify
functionality and interfaces. Continued use for
hardware and flight software development. - Parts qualification is in progress for
poly-switches and HV caps. - Planning is in progress for radiation testing
(total dose and FEE) after the new chips arrive.
A mini-MCM loaded with 2 prototype front-end
chips and 1 controller chip.
14Electronics Test Systems
- GTFE wafer probing
- Operational on a manual probe station (undiced
wafers will be available end of August). - Thorough set of test vectors for complete
functional testing (see LAT-TD-247). - All of the test vectors have been exercised
through the simulation and actual chips (which
uncovered the bugs fixed in the present
submission). - GTRC wafer probing
- Complete hardware and software for manual
testing. - Thorough set of test vectors (see LAT-TD-248).
- No bare die yet to test it on (probes dont fit
V1). - Test vectors are presently being exercised
through the simulation and on the mini-MCM (with
some restrictions, such as not being able on the
mini-MCM to cycle through the address space).
15Electronics Test Systems
- MCM testing during production at Teledyne
- Second-version interface board is in hand with
enclosure plus a better monitoring interface for
power supplies and LVDS bias levels. - Test plan documented in LAT-TD-153 and
LAT-TD-249. - Test vectors are under development for functional
testing, based on extending the GTRC set. - VME ADC modules were obtained, and software is in
progress to test the power supply voltages and
the LVDS driver bias levels. - Some software has been developed for threshold
scanning (analog performance tests) and tested on
the d-version chip. - Basic system, to be operated by its developers,
will be ready for testing the engineering-model
MCMs in September. - MCM burn-in station layout of special cables is
in progress.
16Electronics Production
- Teledyne has successfully developed and tested
the tooling for bonding the pitch adapter flex
circuit onto the MCM PWB. - All parts and materials are present at Teledyne
to begin assembly of the dummy MCMs for the
mechanical/thermal engineering model tracker
module. These modules will have all parts, but
with dummy or bad (GTFE Ve) ICs and with some
capacitors replaced with resistors to mimic the
IC heat load. - All of the molded plastic MCM carrying cases are
in hand (in fact, in sufficient numbers for the
flight build). These are used for MCM testing
and burn-in following assembly, as well as for
storage and transport. - Assembly of the dummy MCMs is beginning at
Teledyne. - Pisa is testing the MCM-tray interface,
especially with thermal cycling.
17Engineering Model Tooling Production
- Experience with the prototype Tracker module led
to many detailed changes in the assembly tooling
designs for trays. - Plyform (Italy) is now assembling EM composite
panels with the new tooling, and work is
beginning to add the converters and bias
circuits. - GA Engineering (Italy) has assembled all of the
dummy SSD ladders and is now working on the live
ones, expected completion July 30, 2002. The
electronics test system for the ladders is
operational. - Tools for mounting the ladders and for assembly
of the top/bottom trays are in progress.
18Procurement of Long-Lead Items
- Silicon procurement and deliveries on schedule.
- 2314 HPK SSDs have been received and tested in
Italy. - Production of flight SSD ladders will begin as
early as August 5, 2002 pending ladder assembly
procedure review and release. - The carbon-carbon closeout material is on order
from Alcomp (US). - In purchasing some items for the EM sufficient
quantities for the flight build were ordered to
reduce costs - The molded MCM carrying cases.
- All of the screws for sidewall attachment.
19Summary Schedule
20Key Milestones
21Cost Commitments
22Cost Profile
23Workforce Plan
24Summary
- The flexure-mount redesign effort is on schedule
and should not delay the Engineering Model
schedule or the CDR. - All other aspects of the Tracker design are
complete. All lowest-level drawings are in place
as drafts, with 22 released. Good progress on
assembly and tooling drawings. See the drawing
tree on the Tracker web page. - Some details of the production fixtures and
processes are still being tweaked and tested as
the Engineering-Model build progresses. - Work is accelerating on documentation of test
plans and processes and production processes, in
preparation for CDR. - Overall, the Tracker development is progressing
on schedule for the subsystem CDR in January 2003
and the LAT CDR in April 2003.