WMKO Next Generation Adaptive WMKO Next Generation Adaptive Optics Optics Build to Cost Concept Review: Build to Cost Concept Review: Introductions & Introductions & Charge to the Review Committee Charge to the Review Committee Taft Armandroff, Hilton Lewis Taft Armandroff, Hilton Lewis March 18, 2009 March 18, 2009
WMKO Next Generation Adaptive Optics Build to Cost Concept Review: Introductions & Charge to the Review Committee. Taft Armandroff, Hilton Lewis March 18, 2009. Introductions. Reviewers: Brent Ellerbroek (TMT) Mike Liu (UH) Jerry Nelson (UCSC) Directors Taft Armandroff Mike Bolte - PowerPoint PPT Presentation
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
WMKO Next Generation Adaptive OpticsWMKO Next Generation Adaptive OpticsBuild to Cost Concept Review:Build to Cost Concept Review:
Introductions & Introductions & Charge to the Review CommitteeCharge to the Review Committee
Taft Armandroff, Hilton LewisTaft Armandroff, Hilton Lewis
March 18, 2009March 18, 2009
2
Introductions
• Reviewers:– Brent Ellerbroek (TMT)– Mike Liu (UH)– Jerry Nelson (UCSC)
• Directors– Taft Armandroff– Mike Bolte– Tom Soifer for Shri
Kulkarni– Hilton Lewis
• SSC co-chair– Chris Martin
• NGAO Team
3
Review Success Criteria
• The revised science cases & requirements continue to provide a compelling case for building NGAO
• We have a credible technical approach to producing an NGAO facility within the cost cap and in a timely fashion
• We have reserved contingency consistent with the level of programmatic & technical risk
These criteria, plus the deliverables & assumptions (next page), were approved by the Directors & presented at the Nov. 3, 2008 SSC meeting
4
Review Deliverables & Assumptions
• Deliverables include a summary of the:– Revisions to the science cases & requirements, & the scientific impact
– Major design changes
– Major cost changes (cost book updated for design changes)
– Major schedule changes
– Contingency changes
• Assumptions– Starting point will be the SD cost estimate with the addition of the science
instruments & refined by the NFIRAOS cost comparison• Better cost estimates will be produced for the PDR
– No phased implementation options will be provided at this time• Some may be for the PDR to respond to the reviewer concerns
– Major documents will only be updated for the PDR• SCRD, SRD, FRD, SDM, SEMP
– Will take into account the Keck Strategic Planning 2008 results
5
Agenda
9:00 Introductions & Charge
9:15-14:30 Review Presentation
with 10:15 break & 12:30 Lunch
14:45 Review Panel Discussion & Report Drafting
16:45 Draft Report from Panel
17:15 End
WMKO Next Generation Adaptive Optics:WMKO Next Generation Adaptive Optics:Build to Cost Concept ReviewBuild to Cost Concept Review
Peter Wizinowich, Sean Adkins, Rich Dekany, Peter Wizinowich, Sean Adkins, Rich Dekany,
Don Gavel, Claire Max & the NGAO TeamDon Gavel, Claire Max & the NGAO Team
Provided by the Directors & SSC co-chairs in Aug/08• $60M cost cap in then-year dollars
– From start of system design through completion– Includes science instruments– Must include realistic contingency – Cap of $17.1M in Federal + Observatory funds ($4.7M committed)
• An internal review of the build to cost concept to be held and reported on no later than the Apr/09 SSC meeting
10
The Challenge
• Previous estimate ~$80M in then-year dollars– NGAO estimate at SDR, including system design (SD), ~ $50M– Science instrument estimate at proposal ~ $30M– Instrument designs were not part of the NGAO SDR deliverables
11
Cost Reduction Approach
• Review & update the science priorities• Review other changes to the estimate (e.g. NFIRAOS cost
comparison)• Update the cost estimate in then-year $• Present & evaluate the recommended cost reductions
– As architectural changes– As a whole including performance predictions
We believe the criteria have been successfully met
Science PrioritiesScience Priorities
13
Key Science Drivers
Five key science drivers were developed for the NGAO SDR (KAON 455):
1. Galaxy assembly & star formation history
2. Nearby Active Galactic Nuclei
3. Measurements of GR effects in the Galactic Center
4. Imaging & characterization of extrasolar planets around nearby stars
5. Multiplicity of minor planets
• We will discuss how our recommended cost reductions impact this science.
14
Science Priority Input: SDR Report
From the SDR review panel report (KAON 588) executive summary:• The panel supported the science cases
– “The NGAO Science cases are mature, well developed and provide enough confidence that the science … will be unique within the current landscape.”
• The panel was satisfied with the science requirements flow down & error budget– “The science requirements are comprehensive, and sufficiently analyzed to properly
flow-down technical requirements.”– “… high Strehl ratio (or high Ensquared Energy), high sky coverage, moderate
multiplex gain, PSF stability accuracy and PSF knowledge accuracy … These design drivers are well justified by the key science cases which themselves fit well into the scientific landscape.”
• The panel was concerned about complexity & especially the deployable IFS – “However, the review panel believes that the actual cost/complexity to science
benefits of the required IFS multiplex factor of 6 should be reassessed.”– “… recommends that the NGAO team reassess the concept choices with a goal to
reduce the complexity and risk of NGAO while keeping the science objectives.”• The panel had input on the priorities
– “The predicted Sky Coverage for NGAO is essential and should remain a top requirement.”
15
Science Priority Input: Keck Scientific Strategic Plan
From the Keck SSP 2008:• “NGAO was the unanimous highest priority of the Planetary, Galactic, &
Extragalactic (in high angular resolution astronomy) science groups. NGAO will reinvent Keck and place us decisively in the lead in high-resolution astronomy. However, the timely design, fabrication & deployment of NGAO are essential to maximize the scientific opportunity.”
• “Given the cost and complexity of the multi-object deployable IFU instrument for NGAO, …, the multi-IFU instrument should be the lowest priority part of the NGAO plan.”
• Planetary recommendations in priority order: higher contrast near-IR imaging, extension to optical, large sky coverage.
• Galactic recommendations in priority order: higher Strehl, wider field, more uniform Strehl, astrometric capability, wide field IFU, optical AO
• Extragalactic high angular resolution recommendations a balance between the highest possible angular resolution (high priority) at the science & high sensitivity
16
Science Implications of no Multiplexed d-IFU
• Galaxy Assembly and Star Formation History– Reduced observing efficiency
– Decreases overall statistics for understanding processes of galaxy formation and evolution
• Can be supplemented with complementary HST & JWST results at higher z
• General Relativity in the Galactic Center– Decreased efficiency in radial velocity measurements (fewer stars
observed at once)
• Can gain back some of efficiency hit with a single on-axis IFU that has higher sensitivity (especially for galaxy assembly) & larger FOV (especially for GC)
16
17
Flowdown of Science Priorities(resultant NGAO Perspective)
Based on the SDR science cases, SDR panel report & Keck Strategic Plan:1. High Strehl
• Required directly, plus to achieve high contrast NIR imaging, shorter AO, highest possible angular resolution, high throughput NIR IFU & high SNR
• Required for AGN, GC, exoplanet & minor planet key science cases
2. NIR Imager with low wavefront error, high sensitivity, ≥ 20” FOV & simple coronagraph• Required for all key science cases.
3. Large sky coverage• Priority for all key science cases.
4. NIR IFU with high angular resolution, high sensitivity & larger format• Required for galaxy assembly, AGN, GC & minor planet key science cases
5. Visible imaging capability to ~ 800 nm• Required for higher angular resolution science
6. Visible IFU capability to ~ 800 nm7. Deployable multi-IFS instrument (removed from plan)
– Ranked as low priority by Keck SSP 2008 & represents a significant cost
8. Visible imager & IFU to shorter
Included in B2CExcluded
Cost EstimateCost EstimateStarting PointStarting Point
19
NGAO System ArchitectureKey AO Elements:• Configurable laser Configurable laser tomographytomography• Closed loop LGS AOClosed loop LGS AO for low order correction over a wide field• Narrow field MOAO Narrow field MOAO (open loop) for high Strehl science, NIR TT correction & ensquared energy
X
20
Cost Estimation Methodology (KAON 546)
• Cost estimation spreadsheets– Based on TMT Cost Book approach, simplified for SD phase– Prepared for each WBS element (~75 in all)– Prepared for each of 4 phases
• Preliminary design, detailed design, full scale development, delivery/commissioning
– Prepared by technical experts responsible for deliverables– Process captures
• WBS dictionary• Major deliverables• Estimates of labor hours• Estimates of non-labor dollars (incl. tax & shipping) & travel dollars• Basis of estimate (e.g. vendor quote, CER, engineering judgment)• Contingency risk factors & estimates• Descope options
– Standard labor classes, labor rates & travel costs used
Total = 138 16,045 16,804 1,681 34,531 7,697 42,227 100%
% = 46% 49% 5% 100% 22% 122%
23
SDR Reviewer Comments
• “Based on the cost and schedule of past and planned projects of lower or similar complexity, the review panel believes that the NGAO project cost and schedule are not reliable and may not be realistic.
Contingencies are also too tight. In particular, the time of 18 months allocated for manufacturing and assembly and 6 months for integration and test, is probably optimistic by a large amount.”
• Relevant to this point they also said:– “The review panel believes that Keck Observatory has assembled an
NGAO team with the necessary past experience … needed to develop the Next Generation Adaptive Optics facility for Keck.”
– “The proposed schedule and budget estimate have been carried out with sound methodology”
• Clarification: Reviewers thought our lab and telescope I&T durations were smaller by 2x than our plan (they are 6 & 12 months, respectively).
24
Results of NFIRAOS Cost Comparison (KAON 625)
• Comparison provided increased confidence in NGAO SDR estimate– Methodology largely gave us reasonable system design estimates – NGAO traceably less expensive than NFIRAOS & we
understand why
• Some areas identified that require more work:– Contingency rates need to be re-evaluated
• At minimum should be increased for laser & potentially for RTC
– Laser procurement estimate needs to be more solidly based• Will have ROMs soon & a fixed price quote for PDR through ESO
collaboration
– Minor items: Laser system labor & cost of RTC labor
25
Science Instrument Cost Estimates
• The science instruments are only at a proposal level– Estimate of $3M (FY06 $) each for NIR imager and Visible imager
in NGAO proposal (June 2006)– NIR & visible imager estimates updated by Adkins– Estimate of $14M (FY06 $) for deployable multi-IFS in NGAO
proposal (June 2006)• This is not included in the starting cost estimate
– No estimate available for NIR IFS when the build-to-cost process began• We did have the Nov/08 ATI proposal for the design costs of a near-IR
IFS• Just assumed $5M total for the starting point
26
Contingency
• NGAO budget at SDR included 22% contingency– $7.7M on a base of $34.5M in FY08 $– $9.1M on a base of $40.2M in then-year $
• Increased contingency based on NFIRAOS cost comparison– $0.68M for laser to increase laser contingency from 19 to 30%– Additional $0.45M to increase overall contingency from 22 to 25%
• Instruments only at proposal level– Assume 30% contingency
27
Starting Cost EstimateStart from SDR cost estimate
• Very ambitious spending profile both for finding funds & ramping up effort– Highly desirable to maximize
science competitiveness
– Slow current start-up rate imposed by available funds
– Critical to produce viable funding/management plan during preliminary design
• NGAO system labor profile is flat after initial ramp-up– $19.4M in then-year $ or 47%
of NGAO system budget
– ~ 40,000 hours/year from FY10 to FY14 or ~ 20 FTEs
NGAO Spending Profile
0
2000
4000
6000
8000
10000
12000
14000
16000
18000
20000
FY07 FY08 FY09 FY10 FY11 FY12 FY13 FY14 FY15
Fiscal Year
Th
en-Y
ear
$k
System
Instruments
Total
NGAO System Labor $ Spending Profile (without contingency)
0
500
1000
1500
2000
2500
3000
3500
FY07 FY08 FY09 FY10 FY11 FY12 FY13 FY14 FY15
Fiscal Year
Th
en
-Year
$k
Total cost
NGAO labor only
AO Design Changes AO Design Changes to Support Build-to-Costto Support Build-to-Cost
30
AO Design Changes Summary
A. Architectural changes allowed by no deployable multi-IFS instrument1. LGS asterism & WFS architecture
2. Narrow field relay location
B. New design choices that don’t impact the requirements1. Laser location
2. AO optics cooling enclosure
C. Design choices with modest science implications1. Reduced field of view for the wide field relay (120” vs 150” dia.)
2. Direct pick-off of TT stars
3. Truth wavefront sensor (one visible instead of 1 vis & 1 NIR)
4. Reduced priority on NGS AO science– Fixed sodium dichroic, no ADC for NGS WFS & fewer NGS WFS subaperture
scales (2 vs 3)
5. No ADC implemented for LOWFS (but design for mechanical fit)
6. OSIRIS role replaced by new IFS
31
Science Instrument Design Changes
• NGAO Proposal had three science instruments ($20M in FY06 $)– Deployable multi IFS instrument
– NIR imager
– Visible imager
• For the SDR we included OSIRIS integration with NGAO• Science instrument design changes that impact the science
capabilities– No deployable multi IFS instrument
– Addition of single channel NIR IFS
– Removal of OSIRIS (science capabilities covered by NIR IFS)
– No visible imager
– Extension of NIR imager & IFS to 800 nm
32
Revised NGAO System ArchitectureKey Changes:1. No wide field science instrument • Fixed narrow field tomography• TT sharpening with single LGS AO• 75W instead of 100W• Narrow field relay not reflected2. Cooled AO enclosure smaller3. Lasers on elevation ring4. Combined imager/IFU instrument & no OSIRIS5. Only one TWFS
33
LGS Architecture (A1)• Absence of multiple d-IFS allowed us to rethink the LGS asterism
– 1st architecture result: a fixed, fewer LGS asterism (4 vs 6) to provide tomographic correction over the narrow science field
– 2nd: no tomographic correction is provided over the wide field. • 3 point & shoot LGS used in single beacon AO systems for each tip-tilt NGS
– 3rd: able to reduce the overall laser power from 100W to 75W• Went from ~11W/LGS to 12.5W/LGS in central asterism & 8W/LGS for tip-tilt
– Also performance analysis defined # of subapertures (only 1 lenslet array)
– “3+1” optimized integ. time– PNS optimized integ. Time– 60” radius FoR for PNS
• LOWFS– 0.32 throughput
– 2 TT + 1 TTFA
– Single LGS AO sharpened
– J+H band– No ADC (Design change C5)
– 32x32 MEMS DM– H2RG (4.5 e-, 0.85 QE at J)
– 60” rad FoR (Design change C1)
• Seeing Conditions– 37.5%: r0 = 14 cm, 0 = 2.15”
– 50.0%: r0 = 16 cm, 0 = 2.7”
– 62.5%: r0 = 18 cm, 0 = 2.9”
– 87.5%: r0 = 22 cm, 0 = 4.0”
35
Justification for Assumptions• 100 ph/cm2/sec/W in mesosphere
– 150 ph/cm2/sec/W shown at SOR• Power at laser output
– Prediction lower for Hawaii• By sin where = angle between
geo-magnetic field & beam direction (62 at SOR, 37 at HI)
• 3E9 atoms/cm2 Na density– Based on Maui LIDAR
measurements
Measured
Predicted
Median 4.3x109 cm-2
3x109 cm-2
38
Performance Analysis Science Cases• The following parameters were used to define the key science driver
cases for the performance analysis
Galaxy Assembly
Nearby AGNs
Galactic Center
Exo-planets Minor Planets
Zenith angle 30 30 50 30 30
Guide stars Field stars Field stars IRS 7,9,12N Field stars Field starsGuide star color M M M MOff-axis evaluation radius 1" 1" 2" 0" 0"Required sky coverage 30% 30% n/a 30% 30%Galactic latitude 30 30 n/a 10 30
Science filter K Z K H Z60s (image)900 (spectra)
120sMax science exposure 1800s 900s 300s
39
Tomography Error versus Field Position• Many alternative asterisms evaluated• Selected 10”-radius “3+1” fixed asterism with 50W total
– Best performance & considered lowest performance risk option– Remaining 25W in 3 point & shoot lasers
Max. science field radius
40
Wavefront Error versus Laser Power
50W in science asterism
50W +median Na
density
41
Strehl Ratio versus Laser Power
50W in science asterism
42
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0.00
2.00
4.00
6.00
8.00
10.00
12.00
14.00
16.00
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
H-b
and
Ensq
uare
d En
ergy
1-D
Tip
-Tilt
Err
or, R
MS
(mas
)
Sky Fraction
EE70mas and Tip-Tilt Error vs. % Sky Coveragefor Galaxy Assembly case, median seeing
Tip-Tilt Error
EE 70 mas
EE 41 mas
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0.00
2.00
4.00
6.00
8.00
10.00
12.00
14.00
16.00
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Z-ba
nd E
nsqu
ared
Ene
rgy
1-D
Tip
-Tilt
Err
or, R
MS
(mas
)
Sky Fraction
EE70mas and Tip-Tilt Error vs. % Sky Coveragefor Nearby AGN case, median seeing
Tip-Tilt Error
EE 33 mas
EE 17 mas
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0.00
2.00
4.00
6.00
8.00
10.00
12.00
14.00
16.00
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
H-b
and
Stre
hl R
atio
1-D
Tip
-Tilt
Err
or, R
MS
(mas
)
Sky Fraction
Strehl Ratio and Tip-Tilt Error vs. % Sky Coverage for Exoplanets case, median seeing
Tip-Tilt Error
Strehl Ratio
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0.00
2.00
4.00
6.00
8.00
10.00
12.00
14.00
16.00
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Z-ba
nd S
treh
l Rati
o
1-D
Tip
-Tilt
Err
or, R
MS
(mas
)
Sky Fraction
Strehl Ratio and Tip-Tilt Error vs. % Sky Coverage for Minor Planets case, median seeing
Tip-Tilt Error
Strehl Ratio
Performance versus Sky Coverage
1d Tilt Error (mas)
% EE (70 mas)
K-bandb = 30
% EE (41 mas)
45
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0.00
2.00
4.00
6.00
8.00
10.00
12.00
14.00
16.00
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
H-b
and
Ensq
uare
d En
ergy
1-D
Tip
-Tilt
Err
or, R
MS
(mas
)
Sky Fraction
EE70mas and Tip-Tilt Error vs. % Sky Coveragefor Galaxy Assembly case, median seeing
Tip-Tilt Error
EE 70 mas
EE 41 mas
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0.00
2.00
4.00
6.00
8.00
10.00
12.00
14.00
16.00
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Z-ba
nd E
nsqu
ared
Ene
rgy
1-D
Tip
-Tilt
Err
or, R
MS
(mas
)
Sky Fraction
EE70mas and Tip-Tilt Error vs. % Sky Coveragefor Nearby AGN case, median seeing
Tip-Tilt Error
EE 33 mas
EE 17 mas
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0.00
2.00
4.00
6.00
8.00
10.00
12.00
14.00
16.00
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
H-b
and
Stre
hl R
atio
1-D
Tip
-Tilt
Err
or, R
MS
(mas
)
Sky Fraction
Strehl Ratio and Tip-Tilt Error vs. % Sky Coverage for Exoplanets case, median seeing
Tip-Tilt Error
Strehl Ratio
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0.00
2.00
4.00
6.00
8.00
10.00
12.00
14.00
16.00
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Z-ba
nd S
treh
l Rati
o
1-D
Tip
-Tilt
Err
or, R
MS
(mas
)
Sky Fraction
Strehl Ratio and Tip-Tilt Error vs. % Sky Coverage for Minor Planets case, median seeing
Tip-Tilt Error
Strehl Ratio
Performance versus Sky Coverage
Z-bandb = 30
Strehl
46
Performance versus Seeing
Median
37.5%
87.5%
47
Optimum # of Subapertures
49
Optimum # of Subapertures
Conclusion: A single scale across pupil works well
(N = 64 assumed for costing)
3E9 Na, Opt Subaps
3E9 Na, N = 64
1E9 Na, Opt Subaps
1E9 Na, N = 64
50
Off-axis Performance
Median seeing
Max. IFU radius
Max. imager radius
Imaging radius requirement
51
Off-axis Performance
Median seeing
Max. imager radius
52
Performance Analysis Summary• “3+1” science asterism + 3 point & shoot lasers has excellent
performance for narrow field science• Overall performance comparable to estimates at SDR
– Assumptions different than at SDR (e.g. we are now using lower Na density & sodium return values)
– Analysis tool/inputs have evolved (e.g. lower tomography error, higher atmospheric transmission off zenith & higher throughput)
– Lower total laser power but smaller tomography volume
– Most importantly performance optimized for on-axis science
Narrow Field Relay Location (A2)• At SDR the location of the multiple deployable IFS & LOWFS required
that the narrow field relay be in reflection off a choice of dichroics• Narrow field relay now in transmission• Allows option of not using a dichroic in front of the LOWFS
– Saves cost of dichroics & switcher
– Higher throughput to LOWFS & science instruments
54
Laser Location (B1)• Likely availability of new lasers allowed a new design choice
– Lasers on elevation moving part of telescope (previously Nasmyth) higher throughput & no need for tracking beam transport system
55
AO Optics Cooling Enclosure (B2)• At SDR assumed that we would cool the entire AO enclosure
including science instruments• New approach: cool as little as possible beyond the science path
– Science instrument front face forms a seal to cooled enclosure
Cooled Volume
SDR New
57
Reduced Wide Field Relay FOV (C1)• 150” dia SDR FOV reduced to 120” with new assumptions• Allows a smaller image rotator + smaller wide field relay optics• Allows a smaller DM – 100 mm instead of 140 mm
higher performance tip-tilt platform Wide field relay scaled down by 100/140 ~70%
OAP1, upper level
K-mirror rotator, upper level
140 mm Woofer DM
LGS WFS focal plane
OAP2
Tweeter DM
OAP3
OAP4
LOWFS/dIFS focal plane
NIR Imager focal plane
NGS WFS TWFS focal
plane
Visible Imager focal plane
FSM
FSM
Fold down K-mirror
LOWFS Boxes
OAP1OAP2
100 mm Woofer DM
25mm tweeter DM
Switchyard mirrorOAP3
OAP4
Science Instrument
NGS WFS
LGS WFS
58
Direct LOWFS Pick-offs (C2)• At SDR pickoffs for TT stars in front of d-IFS & after dichroic that fed
narrow field relay no interference• New design: direct pickoff of each TT star
– no dichroic to split light between LOWFS & science instruments
Pickoffs can vignette science field & can’t use science target for LOWFS
Higher throughput to LOWFS & science instruments
dIFS anddIFS andTip/Tilt sensorsTip/Tilt sensors
Dichroic changer
Narrow field Narrow field science science
instrumentinstrument
Narrow field Narrow field science AO relayscience AO relay
59
One Truth Wavefront Sensor (C3)• At SDR had a NIR Truth WFS (TWFS) in one of the LOWFS units & a
visible TWFS in the narrow field relay• New design: 1 TWFS - a visible TWFS in one of the LOWFS.
Rationale: – Location: low probably of finding a star in the narrow field
– Calibration: Calibrate TWFS for science camera; MEMS impact well defined
– Wavelength: Shouldn’t impact performance
60
Reduced NGS AO Science Priority (C4)
• Fixed sodium dichroic, no ADC & fewer lenslets (2 vs 3)• Rationale (besides need to cut costs):
– NGS vs LGS regime for NGAO• NGS only provides an advantage for science next to very bright NGS• Backup science on nights with > 1 mag cirrus extinction• NGS science has not been a strong driver
– NGS AO regime for NGAO vs Keck I• Higher Strehl NGS AO science on bright targets • Higher sensitivity NGS AO science at K-band on similar magnitude
targets• Other NGS AO science may be better done with K1 NGS AO• K1 NGS AO probably offers more availability
– Reduced capabilities straightforward to implement as future upgrades if motivated by the science
61
OSIRIS role replaced by new IFS (C6)• Carefully reviewed OSIRIS role
– In consultation with Larkin & McLean• Determined that a new IFS was required by science
• Minor science benefit to having both new IFS & OSIRIS– Perhaps some plate scales– Perhaps some multiplexing if new IFS deployable (extra cost)
• More overall science benefit to continuing to use OSIRIS on K1
• NGAO cost savings & design freedom in not having to implement OSIRIS
62
Design Impact in other Areas• Motion control degrees of freedom reduced by 37%
– AO devices reduced from 126 to 77– Laser devices from 89 to 59
• Tomography computation reduced by ~ factor of 10~ ratio of tomography volumes = (120”/40”)2
• Optical switchyard reduced dramatically– Reduced from 7 to 3 mechanisms – Dichroics reduced from 8 to 2
Impact on Science RequirementsImpact on Science Requirements
64
Impact on ability to meet Science Requirements
Key Science Driver SCRD Requirement Performance of B2C
Galaxy Assembly(JHK bands)
EE 50% in 70 mas for sky cov = 30% (JHK)
EE > 70% in 70 mas for sky cov 90% (K band)
Nearby AGNs(Z band for Ca triplet)
EE 50% in 1/2 grav sphere of influence
EE 25% in 33 mas MBH 107 Msun @ Virgo cluster (17.6 Mpc )
General Relativity at the Galactic Center(K band)
100 as astrometric accuracy 5” from GC
Need to quantify. Already very close to meeting this requirement with KII AO.
Extrasolar planets around old field brown dwarfs (H band)
Contrast ratio H > 10 at 0.2” from H=14 star (2 MJ at 4 AU, d* = 20 pc)
Meets requirements (determined by static errors)
Multiplicity of minor planets (Z or J bands)
Contrast ratio J > 5.5 at 0.5” from J < 16 asteroid
Meets requirements: WFE = 170 nm is sufficient
√
√
√
√
√
65
B2C Design Changes: only modest effect on meeting science requirements
• Galaxy Assembly: B2C exceeds SDR requirements
• Nearby AGNs: B2C doesn’t meet EE requirement (didn’t meet at SDR either). Still in interesting regime for BH mass measurements (MBH 107 Msun @ Virgo cluster). Need to review & more clearly define requirement.
• General Relativity at the Galactic Center: Key variables (e.g. differential tilt jitter, geometric distortion in AO & instrument, differential atmospheric refraction) not strongly affected by laser power. Confusion only slightly worse than SDR design.
• Extrasolar planets around old field brown dwarfs: contrast ratio not affected by B2C design changes. Static errors dominate.
• Multiplicity of minor planets: Meets SDR requirements√
√
√
√
√
66
NGAO comparison to JWST & TMT• Higher spatial resolution for imaging & spectroscopy than JWST
– JWST much more sensitive at K. NGAO more sensitive at J & between OH lines at H
• Lots of NGAO science possible in 5 years prior to TMT 1st science– Key community resource in support of TMT science (do at Keck 1st if can)
– Could push to shorter or multi-object IFS or … as TMT arrives on scene
• NGAO could perform long term studies (e.g., synoptic, GC astrometry)WMKO NGAO JWST
Diffraction-limit (mas) at 2 m 41 63Diffraction-limit (mas) at 1 m 20 limited by samplingSensitivity 1x ~200x at 2 mImager NGAO Imager NIRCam IRIS Imager IRMS
4 slit; 3x3 IFU up to 3 120 FORProjected 1st science paper ~2015 ~2014
TMT NFIRAOS147
~2020
~80x
67
NGAO comparison to JWSTEvaluation of key science cases:
Key Science Case JWST & NGAO
Galaxy Assembly (JHK)
JWST much more sensitive at K.NGAO sensitivity higher between OH lines at H.NGAO sensitivity higher for imaging & spectroscopy at J.NGAO wins in spatial resolution at all .NGAO provides higher spectral resolution.
Nearby AGNs (Z) Only NGAO provides needed spatial resolution (especially at Ca triplet).
General Relativity at Galactic Center (K)
Only NGAO provides needed spatial resolution (especially important to reduce confusion limit).Long term monitoring may be inappropriate for JWST.
Extrasolar Planets around old Field Brown Dwarfs (H)
Only NGAO provides needed spatial resolution.JWST coronagraph optimized for 3-5 m, >1"; NGAO competitive ≤2 m, <1".
Multiplicity of Minor Planets (Z or J) Only NGAO provides needed spatial resolution.
68
NGAO comparison to TMT• NGAO & NFIRAOS wavefront errors are ~ the same (162 vs 174 nm rms)
– Similar Strehls but higher spatial resolution for TMT
– Similar spatial resolution for IFU science but higher sensitivity for TMTKey Science Case TMT & NGAO
Galaxy Assembly (JHK)
NGAO & TMT have the same spatial resolution with ~20 & 50 mas IFUs, but TMT has higher sensitivity.NGAO may do most of Z < 2.5-3 targets either before TMT or because of scarce TMT time.
Nearby AGNs (Z)NGAO will screen most important targets. With 3x higher spatial resolution TMT will detect smaller black holes.
General Relativity at Galactic Center (K)
TMT wins in spatial resolution, sensitivity less important. Significant value in continuing NGAO astrometry into TMT era (MCAO field stability concern; Keck access easier).NGAO synoptic advantage.
Extrasolar Planets around old Field Brown Dwarfs (H)
TMT spatial resolution an advantage.Control of static wavefront errors & PSF characterization will be critical (NGAO will have 5 year head start on experience).NGAO synoptic advantage.
Multiplicity of Minor Planets (Z or J)
TMT spatial resolution an advantage; NGAO could move to shorter . Much of this science may be done before TMT?NGAO synoptic advantage.
Science Instruments Science Instruments to Support Build-to-Costto Support Build-to-Cost
70
NGAO Science Instrumentation
• Background• Approach to design/build to cost• Changes to Instrumentation• Baseline capabilities
71
Background• NGAO science requirements established a need for certain capabilities in
the SD phase– Imaging
• ~700 nm to 2.4 m• high contrast coronagraph
– Integral field spectroscopy in near-IR and visible• spatially resolved spectroscopy for kinematics and radial velocities• high sensitivity• high angular resolution spatial sampling• R ~ 3000 to 5000 (as required for OH suppression and key diagnostic lines)• Improved efficiency
– larger FOV– multi-object capability
– At SDR • two imagers and an integral field spectrograph (IFS) on narrow field high Strehl AO
relay (IFS might be OSIRIS)• 6 channel deployable IFS on the moderate field AO relay with MOAO in each channel
– Build to cost forces a narrowing of scope, significant reduction in number and capabilities for science instruments
– May only be able to afford one science instrument
72
Approach to design/build to cost1. Be sure instrument capabilities are well matched to key science
requirements– Galaxy assembly & star formation history– Nearby Active Galactic Nuclei– Measurements of GR effects in the Galactic Center– Imaging & characterization of extrasolar planets around nearby stars– Multiplicity of minor planets
2. Match instrument capabilities to AO system – maximize benefit of improved capabilities for science gains
3. Understand which requirements drive cost4. Resist the temptation to add features5. Maximize heritage from previous instruments6. Exploit redundancies in compatible platforms – e.g. Near-IR imager and
Near-IR IFS7. Evaluate ways to break the normal visible/near-IR paradigm of using
different detectors
73
Changes to Instrumentation
• No deployable IFS• One broadband imager• One new IFS• Address cost drivers
74
NGAO Imaging Capability
• Broadband– z, Y, J, H, K (0.818 to 2.4 µm)
– photometric filters for each band plus narrowband filters similar to NIRC2
• Single plate scale– selected to optimally sample the diffraction limit, e.g. 2(/D) or 8.5 mas
at 0.818 µm
• FOV– 34.8" x 34.8" with 8.5 mas plate scale
• Simple coronagraph• Throughput ≥ 60% over full wavelength range• Sky background limited performance
– z, Y, J, H, K (0.818 to 2.4 µm)– ~5% band pass per filter, number as required to cover each wave band
• Spectroscopy– R ~4,000– High efficiency e.g. multiple gratings working in a single order
• Spatial sampling (3 scales maximum)• 10 mas e.g. 2(/D) at 1 m • 50 to 75 mas, selected to match 50% ensquared energy of NGAO• Intermediate scale (20 or 35 mas) to balance FOV/sensitivity trade off
• FOV on axis– 4" x 4" at 50 mas sampling– possible rectangular FOV (1" x 3") at a smaller spatial sampling
• Throughput ≥ 40% over full wavelength range• Detector limited performance
77
Narrowband Science
• Extra-galactic– IFS will be used for targets with known redshifts
• Therefore 5% bandpass sufficient?• 5% spans Hα and NII lines for example
– 4 narrowband (5%) filters will cover the K-band
– Excitation temperatures• Need at least 4 lines• Can expect to get 2 or more in each filter• Can optimize center wavelength to maximize this• Practical to use 2 or more exposures to get enough lines
– Imaging spectrograph allows you to detect, and discount image motion for better photometric matching of spectra
– Need to have enough FOV to ensure you cover the whole object in each exposure
• Exoplanet detection– Broadband filters available with narrow FOV ~1" x 1"
78
Narrowband Science• Nearby AGN (Black Holes)
– Galaxy kinematics• CO bandhead 4 to 5% wide (OSIRIS Kn5 filter)• Brackett gamma, H_2 emission lines (OSIRIS Kn3 filter)
– Remain in that passband to z = 0.03
• Same arguments on practicality of non-simultaneous spectra apply
– Central Black Hole• Narrowband adequate for measuring black hole mass (only 1 line) • ~1“ diameter FOV
• Galactic Center (e.g. GR effects)– Narrowband acceptable for RV measurements– Being used now– Want better SNR
• Throughput• Higher angular resolution to reduce stellar confusion, but keep present FOVs
– Could use more FOV
79
IFS/Imager Product Structures
• Some clear commonalities– Single
instrument eliminates having 2 of everything in green
• IFS design based on OSIRIS– 85 x 85 lenslets, 200 m pitch, 17 mm x 17 mm overall
• OSIRIS 64 x 64 lenslets, 250 m pitch, 16 mm x 16 mm overall• Very similar collimator aperture
– Larger camera, Hawaii-4RG with 15 m pixels• OSIRIS Hawaii-2, 18 m pixels• Camera focal plane 1.6 times OSIRIS in each dimension
• Multiple gratings to optimize efficiency– Not a novel approach, SINFONI uses multiple gratings
• Imager very straightforward design– Narrow field AO relay at f/46 with 40" FOV makes imager optics easier
86
Cost Comparisons• OSIRIS
– Full cost in 2005 dollars $5.63M– In 2009 dollars $6.6M– OSIRIS has IFS and imager– New IFS and imager have larger FOVs; FY09 cost estimate $11.8M
• Specific high cost components:– OSIRIS collimator and camera $1M in 2009 dollars
• Budget is $2.1M for NGAO IFS
– OSIRIS lenslet array $70K in 2009 dollars• Budget is $150K for NGAO IFS
• NIRC2– Full cost in 2001 dollars $5.9M– In 2009 dollars $8M– NIRC2 has three plate scales, and spectroscopic capability– Many more features than the NGAO imager
87
MOSFIRE comparison
• MOSFIRE costs are as built costs in 2009 dollars
• NGAO imager cost estimates are in 2009 dollars
• MOSFIRE optics for 6.8' FOV cost ~$1.2M
88
TMT IRIS Cost Comparison
• IRIS estimate = $17.6M in FY09 $, excluding 23% contingency• Major differences from NGAO instrument
– On-instrument WFS $4M
– Materials only costs:• Two kinds of slicer: mirror & lenslet, & 2 scale changer mechanisms ~$1.2M• More difficult TMAs ~$1M• Imager optical path is separate including filters & pupil masks ~$0.6M• Instrument rotator ~$0.3M
– IRIS/TMT interfaces more complex
– NGAO instrument reuses previous designs
• IRIS cost without OIWFS & additional features ~$10.5M versus $10M for NGAO instrument
Revised Cost EstimateRevised Cost Estimate
90
Revised Cost EstimateIncluding all proposed cost reductions & new cost estimates:
• Inflation assumption = 2.0% in FY09 & 3.5%/yr in FY10 to 15
• Cost of MEMS ($425k total)– Estimate has increased from $75 to $150/actuator based on recent
quotes
• Laser cost estimate– Nominally the laser power decrease from 100 to 75W should have
reduced the SDR laser procurement cost estimate by ~ $1M– However, we have not reduced our SDR cost
• We have transferred some $ from labor to non-labor
– Initial rough estimates from the ESO laser preliminary design contracts are consistent with the $5.7M budgeted for laser procurement
– Recall that laser contingency has been increased to 30%
95
Other Post-SDR Changes considered in B2C
• B2C estimate includes NSF MRI proposal budget for K2 center launch telescope– Early phased implementation of NGAO with nearer-term K2
science benefits– Essentially identical launch telescope to one received for K1 LGS
• Evaluated to meet NGAO requirements
– Launch telescope cost based on quote– Reason for FSD dollars in FY10/11
• B2C estimate also includes NSF ATI proposal budget for IFS design study
• Solution for MASS/DIMM implementation– TMT donated equipment being implemented by CFHT/UH
96
AO Contingency & Risk
• Overall contingency has increased from 22.6% to 24.2% – Due to increased laser contingency– Contingency has not been increased on any other WBS– Contingency has not been decreased due to the reduced complexity
• Risk has been significantly reduced in a number of areas– Laser
• Collaboration with ESO, GMT, TMT & AURA on laser preliminary designs• ESO providing 250 kEuros each to 2 companies for preliminary designs• WMKO/GMT/TMT/AURA providing 125 kEuros each to the same companies
for additional risk reduction (using $300k of AURA funding)• All information will be shared with all under NDAs• ESO will procure 4x 25W lasers• WMKO could potentially order with ESO or TMT to reduce costs
– Complexity• All of the design changes move us in the direction of a less complex system• Simpler subsystems (e.g., LGS WFS, launch facility, motion control, RTC, etc.)• Significantly reduced complexity for I&T
97
Approach to NGAO Cost Changes
• Started with SDR cost estimate summary spreadsheet– Summary includes labor, travel, non-labor & contingency for 85 WBS
elements in each of 4 phases (PD, DD, FSD, DC)
• Referenced initial cost sheet to understand cost impact of each design change
• Each cost change is highlighted (red) in cost estimate summary, a comment has been added & a corresponding equation put in the cell– Contingency is automatically updated using the original rate
• Used actual hardware costs from initial cost sheets wherever possible– If available used labor associated with a specific task in a cost sheet
• Performed check with cost sheet estimator in some cases• Tried to be conservative with labor reductions
– Especially conservative in PD phase since PD phase still evolving
98
Cost Changes by WBShrs PY Labor Non-labor Travel Conting Total
4 AO System Development4.1 AO Enclosure 0 0.0 0 0 150 0 27 177
Assessment of Build-to-Cost Review Assessment of Build-to-Cost Review Deliverables & Success CriteriaDeliverables & Success Criteria
+ Conclusions+ Conclusions
103
Review Deliverables Summary (1 of 3)
• Revisions to the science cases & requirements, & the scientific impact– Galaxy assembly science case & requirements need to be
modified for a single IFU instead of multiple deployable IFUs• Scientific impact of no multi d-IFUs viewed as acceptable (low priority
in Keck SSP 2008 & single, higher performance IFU part of B2C)
– Only minor impacts on all other science cases
• Major design changes– Major design changes discussed in this presentation– Design changes documented in KAON 642– Performance impact of design changes documented in KAON 644
104
Review Deliverables Summary (2 of 3)
• Major cost changes– Major cost changes discussed in this presentation– All cost changes documented with comments & equations in cost
book summary spreadsheet by WBS and phase• Viewed as better tool than cost book for tracking changes
– Decision not to update cost book until PDR costing phase• Summary cost spreadsheet will be used as input to the PDR costing
• Major schedule changes– No major schedule changes assumed
• 2 month slip in milestones assumed for cost estimate
– New plan needs to be developed as part of preliminary design• Preliminary design phase replan is a high priority post this review
105
Review Deliverables Summary (3 of 3)
• Contingency changes– Reviewed contingency as part of NFIRAOS cost comparison
• Laser, & potentially RTC, increase identified as needed
– Laser contingency increased to 30%– Other bottom-up contingency estimates viewed as sufficient
especially given reduction in complexity with design changes
106
Review Success Criteria Assessment
• The revised science cases & requirements continue to provide a compelling case for building NGAO– NGAO continues to be compelling scientifically
• We have a credible technical approach to producing an NGAO facility within the cost cap and in a timely fashion– We believe that we have a very credible technical approach to
producing the facility within the cost cap & in a timely fashion– Beyond the criteria for this review we need to work on producing a
• We have reserved contingency consistent with the level of programmatic & technical risk– We believe that we have met this criteria
107
Conclusions
• The build-to-cost guidance has resulted in a simpler & therefore less expensive NGAO facility with similar science performance– This has primarily been achieved at the expense of a significant science
capability (e.g., the multiple deployable IFS)
• Pending the outcome of this review our management priorities will switch to:– Replanning & completing the preliminary design in a timely fashion– Developing a viable funding & management plan for delivering NGAO in a
timely fashion as a preliminary design deliverable
Thanks to all for your participation in this review!