dark energy survey (des)

DARK ENERGY DARK ENERGY SURVEY (DES)SURVEY (DES)

Francisco Javier Castander Serentill

All material borrowed from DES collaboration

IEEC/CSIC

Announcement of OpportunityAnnouncement of Opportunity

Blanco Instrumentation PartnershipBlanco Instrumentation Partnership• Develop a major instrument for Blanco 4m CTIO

• Submit a science, technical & management plan

• Community instrument

• Up to 30% of Blanco 4m for 5 years commencing in 2007 or 2008

• Letter of intent March 15, 2004

• Proposals August 15 2004

The Science Case for the Dark Energy Survey

James Annis For the DES Collaboration

The Dark Energy Survey

• We propose to make precision measurements of Dark Energy– Cluster counting, weak lensing and

supernovae– Independent measurements

• by mapping the cosmological density field to z=1– Measuring 300 million galaxies– Spread over 5000 sq-degrees

• using new instrumentation of our own design.– 500 Megapixel camera– 2.1 degree field of view corrector– Install on the existing CTIO 4m

Cosmology in 2004

Combine to measure parameters of cosmology to 10%. We enter the era of precision cosmology.

– Confirms dark energy (!)

2003 Science breakthrough of the year

WMAP measures the CMB radiation density field at z=1000

Sloan Digital Sky Survey measures the galaxy density field at z < 0.3

The Big Problems: Dark Energy and Dark Matter

• Dark energy?Who ordered that? (said Rabi about muons)

• Dark energy is the dominant constituent of the Universe

• Dark matter is next

The confirmation of Dark Energy points to major holes in our understanding of fundamental physics

1998 Science breakthrough of the year

95% of the Universe is in forms unknown to us

Dark Energy

1. The Cosmological Constant Problem Particle physics theory currently provides no understanding of why

the vacuum energy density is so small: DE (Theory) /DE (obs) = 10120

2. The Cosmic Coincidence ProblemTheory provides no understanding of why the Dark Energy density

is just now comparable to the matter density.

3. What is it?Is dark energy the vacuum energy? a new, ultra-light particle? a

breakdown of General Relativity on large scales? Evidence for extra dimensions?

The nature of the Dark Energy is one of the outstanding unsolved problems of fundamental physics. Progress requires more precise probes of Dark Energy.

• One measures dark energy through how it affects the universe expansion rate, H(z):

H2(z) = H20 [ M (1+z) 3 + R (1+z) 4 + DE (1+z) 3 (1+w) ]

matter radiation dark energy

• Note the parameter w, which describes the evolution of the density of dark energy with redshift. A cosmological constant has w = 1.

w is currently constrained to ~20% by WMAP, SDSS, and supernovae

• Measurements are usually integrals over H(z) r(z) = dz/H(z)• Standard Candles (e.g., supernova) measure dL(z) = (1+z) r(z)• Standard Rulers measure da(z) = (1+z)1 r(z)• Volume Markers measure dV/dzd = r2(z)/H(z)• The rate of growth of structure is a more complicated function of H(z)

Measuring Dark Energy

DES Dark Energy Measurements• New Probes of Dark Energy

– Galaxy Cluster counting• 20,000 clusters to z=1 with M > 2x1014 M

– Weak lensing• 300 million galaxies with shape measurements

– Spatial clustering of galaxies• 300 million galaxies

• Standard Probes of Dark Energy– Type 1a Supernovae distances

• 2000 supernovae

Supernova

• Type 1a Supernovae magnitudes and redshifts provide a direct means to probe dark energy – Standard candles

• DES will make the next logical step in this program:– Image 40 sq-degree repeatedly

– 2000 supernovae at z < 0.8

– Well measured light curves

SCP EssenceLSST

DES

SNAP

2000 2005 2010 2015 2020

SDSS

Current projects

CFHLSPanStarrs

Proposed projects

New Probes of Dark Energy

• Rely on mapping the cosmological density field

• Up to the decoupling of the radiation, the evolution depends on the interactions of the matter and radiation fields - ‘CMB physics’

• After decoupling, the evolution depends only on the cosmology - ‘large-scale structure in the linear regime’.

• Eventually the evolution becomes non-linear and complex structures like galaxies and clusters form - ‘non-linear structure formation’.

z = 30 z = 0

Spatial Clustering of Galaxies

• The distribution of galaxy positions on the sky reflects the initial positions of the mass

• Maps of galaxy positions are broken up in photometric redshift bins

• The spatial power spectrum is computed and compared with the CMB fiducial power spectrum.

• The peak and the baryon oscillations provide standard rulers.

• DES will– Image 5000 sq-degrees– Photo-z accuracy of z < 0.1 to z = 1

– 300 million galaxiesCooray, Hu, Huterer, Joffre 2001

LSST

DES

SNAP

2000 2005 2010 2015 2020

PlanckSDSS WMAP

PanStarrs

Weak Lensing

Ds distance to sourceDl distance to lens

Dls distance from lens to source

Light path

Background galaxy shear maps

Lensing galaxies

• Weak lensing is the statistical measurement of shear due to foreground masses

• A shear map is a map of the shapes of background galaxies

Weak Lensing

Shear maps(z)

Galaxy map

z = 1/4z = 1/2

z = 3/4

DeepLens CFHLS

• The strength of weak lensing by the same foreground galaxies varies with the distance to the background galaxies.– Measure amplitude of shear vs. z

– shear-galaxy correlations

– shear-shear correlations

• DES will– Image 5000 sq-degrees

– Photo-z accuracy of z < 0.1 to z = 1

– 10-20 galaxies/sq-arcminute

LSST

DES

SNAP

2000 2005 2010 2015 2020

PlanckSDSS WMAPPanStarrs

Peaks in the Density Field

• Clusters of galaxies are peaks of the density field.

• Dark energy influences the number and distribution of clusters and how they evolve with time.

2 Mpc16 Mpc

Cluster Masses

• Our mass estimators– Galaxy count/luminosity– Weak lensing– Sunyaev-Zeldovich

• The South Pole Telescope project of J. Carlstrom et al.

• DES and SPT cover the same area of sky

• Self calibration– Mass function shape allows

independent checks– Angular power spectrum of clusters– Allows an approach at systematic

error reduction

SZ

OpticalLensing

X-ray

Mass

Cluster Counting

• Locate peaks in the density field using cluster finders– Red sequence methods

– SZ peaks

• DES will– Image 5000 sq-degrees

– Photo-z accuracy z = 0.01 to z = 1

– 20,000 massive clusters

– 200,000 groups and clusters

z = 0 1 3

z

N

LSST

DES

SNAP

2000 2005 2010 2015 2020

PlanckSDSS WMAPPanStarrs

Low mass

High mass

Very massive

13.7 log M < 14.2

14.2 log M

14.5 log M

We aim at ~5% precision on Dark Energy

Cluster Counting Weak Lensing Supernova

The Planck satellite will provide tighter input CMB measurements, and the constraints will improve slightly.

ww w

DEMM

Joint constraints on w and wa are promising: initial results suggest wa ~ 0.5.

w ~ 5% and DE ~ 3%

The Dark Energy Survey

• We propose the Dark Energy Survey– Construct a 500 Megapixel camera

– Use CTIO 4m to image 5000 sq-degrees

– Map the cosmological density field to z=1

– Make precision measurements of the effects of Dark Energy on cosmological expansion:

• Cluster counting

• Weak lensing

• Galaxy clustering

• Supernovae

5000 sq-degrees

Overlapping SPT SZ survey

4 colors for photometric redshifts

300 million galaxies

Brenna Flaugher for the DES Collaboration BIRP Meeting August 12, 2004 TucsonFermilab, U Illinois, U Chicago, LBNL, CTIO/NOAO

Design of the Dark Energy Survey

James Annis


Science Goals to Science Objective

• To achieve our science goals:– Cluster counting to z > 1

– Spatial angular power spectra of galaxies to z = 1

– Weak lensing, shear-galaxy and shear-shear

– 2000 z<0.8 supernova light curves

• We have chosen our science objective:– 5000 sq-degree imaging survey

• Complete cluster catalog to z = 1, photometric redshifts to z=1.3• Overlapping the South Pole Telescope SZ survey• 30% telescope time over 5 years

– 40 sq-degree time domain survey• 5 year, 6 months/year, 1 hour/night, 3 day cadence


DARK ENERGY SURVEY (DES)DARK ENERGY SURVEY (DES)

Science Goal: measure w=p/ρ, the dark energy equation of state, to a precision of δw ≤ 5%, with

• Cluster Survey

• Weak Lensing

• Galaxy Angular Power Spectrum

• Supernovae


DES: Requirements

Science Goals

• Cluster Survey

• Weak Lensing

• Galaxy Angular Power Spectrum

• Supernovae

Science Requirements

redshifts, area, filters, limit mag, red

image quality, area

photometry, area, limit mag

repeat, area, filters, red


Science Requirements

1. 5000 sq-degrees• Significantly overlapping the

SPT SZ survey area• To be completed in 5 years

with a 30% duty cycle

2. 4 bandpasses covering 390 to 1100 nm• SDSS g,r,i,z• z modified with Y cutoff

3. Limiting magnitudes• g,r,i,z = 24,24,24,23.6• 10σ for small galaxies

4. Photometric calibration to 2%• 1% enhanced goal

5. Astrometric calibration to 0.1”

6. Point spread function• Seeing < 1.1” FWHM

• Median seeing <= 0.9”

• g-band PSF can be 10% worse

• Stable to 0.1% over 9 sq-arcminute scales

From chapter 3 of NOAO proposal; version 3 of requirements.

Version 4, under review, will be a formal science requirements document.


Limiting Magnitude

Limiting magnitude (10σ for small galaxies) was set by flow down of science goals:

• ½ L* cluster galaxies at redshift 4000A break leaving blue filter

– g,r,i,z = 22.8,23.4,24.0,23.3– Complete cluster catalog

• Galaxy catalog completeness– g,r,i,z = 22.8,23.4,24.0,23.6– Simple selection function

• Blue galaxy photo-z at faint mags– g,r,i,z = 24.0,24.0,24.0,23.6– Photo-z for angular power spectra

and weak lensing

0 redshift 1.5

0 redshift 1.5

Mag of ½ L* galaxy

photo-z – spectro-z

i = 23-24

Red Galaxy


Galaxy Cluster Redshifts

• the distribution of the number of clusters as a function of redshift is sensitive to the dark energy equation of state parameter, w.

four filters (griz) track 4000 Å break.

Need z band filter to get out to redshift >~ 1

• DES data will enable cluster photometric redshifts with dz~0.02 for clusters out to z~1.3

for M > 2x1014 M

Theory


Photometric Redshifts

Resulting limiting magnitudes give very good photometric redshifts

• Monte Carlo simulations of photometric redshift precision– Evolving old stellar pop. SED– Redshifted and convolved with

filter curves. Noise added.– Polynomial fit to photo-z– For clusters, averaging all

galaxies in the cluster above limiting magnitude.

• Template fit for photo-z

• These are sufficient to achieve our science goals.

½ L* 2 L*

1.0x1014 M0

Clusters

Red galaxies


The Footprint

Requirements• Overlap with SPT SZ survey• Redshift survey overlap

Footprint• -60 <= Dec <= -30• SDSS Stripe 82 + VLT surveys

Overlap target Right Ascension (deg)

Declination (deg)

Area (sq. deg.)

SPT -60 to 105 -75 to -60

-30 to -65-45 to –65

4000

SDSS Stripe 82 -50 to 50 -1.0 to 1.0 200

Connection region

20 to 50 -30 to –1.0 800

DIRBE dust map, galactic coordinates


Survey Strategy I

• Design decision 1: area is more important than depth– Image the entire survey area multiple times

• Design decision 2: tilings are important for calibration– An imaging of the entire area is a tiling– Multiple tilings are a core means of meeting the photometric

calibration requirement: offset tilings, not dithers

• Design decision 3: substantial science with year 2 data– We will aim for substantial science publications jointly with the

public release of the year 2 data.


Survey Strategy II

• Year 2– g,r,i,z 100 sec exposures– g,r,i,z =24.6, 24.1, 23.6, 23.0– Calibration: abs=2.5% rel=1.2%– Clusters to z=0.8– Weak lensing at 12 gals/sq-arcmin

• Year 5– z 400 sec exposures– g,r,i,z =24.6, 24.1, 24.3, 23.9– Calibration: abs=<2% rel=<1%– Clusters to z=1.3– Weak lensing at 28 gals/sq-arcmin

• Two tilings/year/bandpass

• In year 1-2, 100 sec/exp• In year 3, drop g,r and devote

time to i,z: 200 sec/exp• In year 5, drop i and devote

time to z: 400 sec/exp

• If year 1 or 2 include an El Nino event, we lose ~1 tiling, leaving three tilings at the end of year 2. This is sufficient to produce substantial key project science.


DES Time Allocation Model

September: 4 bright+ 4 dark nights 22 nights October: 4 bright+ 5 dark nights 22 nights

November: 4 bright+ 4 dark nights 22 nights

December: 4 bright+ 4 dark nights 21 nights Telescope shut down Dec 25, 31

January: 4 bright+ 5 dark nights 11 nights and the 2nd half of all nights

February: 3 bright+ 3 dark nights 11 nights and the 2nd half of all nights

March – August all none

Total 257 nights 108 nights

Time to the Community and to the Dark Energy Survey


Time Allocation

• Analytic calculation of time available– 30 year CTIO weather statistics– 5 year moving averages– Calculate photometric time– Can complete imaging survey and

time domain survey with 3 sq-degree field of view camera

• Simulations of observing process– Use mean weather year – Survey geometry– Observing overhead– NOAO time allocation model– High probability of completing core

survey area in time allocated

Probability of obtaining 8 tilings per year over survey area. Dark is 100%, light yellow ~50%

=> DES time allocation model just sufficient to achieve science objective.

CTIO mean weather year


Photometric Calibration Strategy

• Calibrate system response– Convolve calibrated spectrum

with system response curves to predict colors to 2%

– Dedicated measurement response system integrated into instrument

• Absolute calibration– Absolute calibration should be

good to 0.5%– Per bandpass: magnitudes,

not colors– Given flat map, the problem

reduces to judiciously spaced standard stars

• Relative calibration– Photometry good to 2%– Per bandpass: mags, not colors– Use offset tilings to do relative

photometry• Multiple observations of same

stars through different parts of the camera allow reduction of systematic errors

• Hexagon tiling:– 3 tilings at 3x30% overlap– 3 more at 2x40% overlap

– Aim is to produce rigid flat map of single bandpass

– Check using colors• Stellar locus principal colors


Survey Simulation

• We plan a full scale simulation effort– Led by Huan Lin– Centered at Fermilab and Chicago– Using analytic, catalog and full

image simulation techniques

• Over 4 years– Underway, starting with photometric

redshift simulations

• Use the simulations in 3 ways:– Check reduction code

• Mock data reduction challenge• Chris Stoughton

– Prepare analysis codes• Mock data analysis challenge• Josh Frieman

– Prepare for science

• Survey simulations– Jim Annis

• Catalog level simulations– Lin, Frieman, students for photo-z and

galaxy distributions – Risa Weschler’s Hubble Volume n-body– Albert Stebbins’s multi-gaussian

approximation– Mike Gladder’s empirical halo model

• Image level simulations– Erin Sheldon for weak lensing– Doug Tucker and Chris Stoughton

• Terapix skyMaker• Massey’s Shapelets code


Survey Planning Summary

• We have well defined science goals and a well defined science objective– A 5000 sq-degree survey substantially overlapping the SPT survey– A time domain survey using 10% of time

• The science requirements are achievable.– A good seeing, 4 bandpass, 2% calibration, i ~ 24 survey

• Multiple tilings of the survey area the core of the survey strategy and photometric calibration.

• The survey can be completed using:– 22 nights a month between September and October– 21 nights in December– 22 half nights a month in January and February


DES Instrument Project

OUTLINE• Science and Technical

Requirements• Instrument Description• Cost and Schedule

Prime Focus Cage of the Blanco Telescope

We plan to replace this and everything inside it


DES Instrument Reference Design

3556 mm

1575 mmCamera

Filters

Optical Lenses

ScrollShutter

The Reference Designrepresents our current design choices and may change with more analysis

1.2.1 CCDs1.2.2 CCD Packaging1.2.3 Front End Electronics1.2.4 CCD Testing1.2.5 Data Aquisition1.2.6 Camera Vessel1.2.7 Cooling1.2.8 Optics1.2.9 Prime Focus Cage1.2.10 Auxiliary Components1.2.11 Assembly and Testing

Instrument Construction Organization


Optics Design

• 2.2 deg. FOV Corrector – 5 powered elements (Fused Silica)– one aspheric surface (C4)

• four filters – griz needed for DES– others can be used

• More details of the design in the next talk (Steve Kent)

• Cost for the glass ~ 660k$• Cost for figuring ~ $1M• ~ 1.5 yr delivery

Corrector


Dark Energy Survey:Optical Design and Issues

2.2 Deg. Field of View Corrector

• Requirements• Performance• Issues

Steve Kent, Fermilab, for the DES CollaborationDark Energy SurveyBIRP, Aug 12, 2004


2.2 Deg. Field of View Corrector

● 14 requirements total

– 0.39 to 1.1 μ (SDSS filter bandpasses)

– Scale 17.7 arcsec/mm

– Field size 450 mm diameter (2.2 degrees)

– D80 < 0.64 arcsec everywhere (FWHM < .4 arcsec)

– No ADC (Atmospheric Dispersion Corrector)

– Minimize ghosting

– Space for filter, shutter

– Design choices should minimize procurement, fabrication schedule.


Gladders may11 design

Features:

•Flat focal plane•Five lenses + Filter

• (including dewar window)•All fused silica•One aspheric surface•Largest diameter 1.1 meters•Flexibility spacing elements•Low distortion (<1%)•Good ghosting properties

• star halos• exit pupil image

Filter

Shutter

C1

C2

C3

C4

C5

Filter


CCDs

• Reference Design: LBNL CCDs– QE> 50% at 1000 nm – 2k x 4k– 15 micron pixels– 250 microns thick– fully depleted (high resistivity)– back illuminated– 4 side buttable– readout 250 kpix/sec– 2 RO channels/device– readout time ~17sec– fringing eliminated– PSF controlled by bias voltage

R&D on LBNL CCDs nearly finished. LBNL CCDs have been used at LICK and on the WIYN Telescopeand on the Mayall


CCD QE and Read noise

1

10

100

0 1 10 100

Sample time (ms)

Noi

se (

elec

tron

s)

Read noise for a recently finished DALSA 2k x 4k

250 kHz → 7e-

To get redshifts of ~1 we spend ~50% of survey time in z-band.

LBNL CCDs are much more efficient in the z band than the current devices in Mosaic II

DECam / Mosaic II QE comparison

0

10

20

30

40

50

60

70

80

90

100

300 400 500 600 700 800 900 1000 1100

Wavelength (nm)

QE, LBNL (%)QE, SITe (%)


CCD Wafers:• Existing masks have 2/wafer• to be cost efficient we will make new masks with 4/wafer

CCD Acquisition Model

Reference Design Acquisition Model• Order CCDs through LBNL – good relationship with commercial foundry• Foundry delivers wafers to LBNL (~650 microns thick)• LBNL

– applies backside coatings for back illuminated operation– oversees thinning (~ 250 microns thick) and dicing– tests all devices on cold probe station

• LBNL delivers all tested, unpackaged devices to FNAL• FNAL packages and tests CCDs• Prepared to package ~ 160 CCDs (spares, yield)


Packaging

• CCD Packaging will be done at Fermilab

• LICK and LBNL have already successfully packaged small quantities.

• We are developing a working relationship with R. Stover at LICK (we visited in July) to learn packaging techniques

CCD Packaging is very similar to building the components of silicon vertex detectors. Fermilab has built many vertex detectors for CDF and D0, and is contributing to CMS

CCD packaged at LBNL

Invar Foot

AlN circuit board

Wirebonds to CCD


Packaging and Testing Process

• Packaging and testing keep up with anticipated CCD delivery rate of 20/month (5 wafers).

• Packaging:– one CCD takes ~ 1 week to complete– Plan to have capabilities to start 2/day

• Testing: – estimate 2 days/CCD– 3 identical test stands needed to keep up with 5 CCDs/week

• LBNL cold probe test results will guide which CCDs to package 1st

• Assume 60 good devices from production run and up to 18 good devices from preproduction run


CCD Test Stand and Acceptance Criteria

• Testing– linearity, full well depth, QE, CTE,

readnoise, dark current

• Testing and acceptance criteria will be defined as we gain experience with LBNL CCDs

• Will also consider impact of acceptance criteria on community

• Multiple tilings reduces impact of bad regions

• Study with 100 consecutive bad columns found ~1.5% of tiling area was imaged less than 3 times after 5 complete tilings

Broadband, High-Intensity

Light Source(~300 to 1100 nm) S

hu

tte

r

Monochromator(~300 nm to 1100 nm) F

ilte

rs

CCDelectronics Integrating

Sphere CalibratedPhotodiode

Light Tight Box

Fe55 Source

Cryostat withCCD

CCDQuartz Window

DAQ and InstrumentControl Computer


Camera Reference Design

Focal Plane

62 2k x 4k CCDs for main image, 4-side buttable, 15 micron pixels

8 1k x 1k CCDs for guiding and focus

Camera Design

Focal Plane

feed throughboard

Frontendelectronics


Camera Vessel

• Vacuum feed through board brings signals out of cryostat

Camera is separated into two spool pieces: one for signal feed throughsone for cooling and vacuum services

Removal of cooling spool piece allows access to back of focal plane and cables

Cooling/ Vacuum spool piece


Cooling and Integration

• Reference Design has LN2 reservoir inside cryostat

• Fill from recondensing dewars on floor

• investigating alternative: Gifford-McMahon cryo coolers on cooling spool piece which condense N2 directly into reservoir

Will fully assemble prime focus cage at FNAL and test all systems together (corrector, focal plane,cooling, data acquisition, data management....)before shipping to Chile


Front End Electronics and DAQ

• Large focal plane implies long cables between CCD and electronics crates• Reference design has clock drivers and preamps as part of the cable assembly • Goals are noise < 5 e-, linearity <0.25%, support a readout rate of 250 kpix/sec• Reliable operation requires careful consideration of internal and external components• Minimize heat generated in the PF cage by locating DAQ off telescope

off the cage


Data Acquisition

DAQ Parameters

62 (+4)

2k x 4k (1k x 1k)

2

250 kHz (17 s)

971 MB

57 MB/s

4

1k x 1k

2

1 MHz (0.5 s)

8 MB

16 MB/s

# CCDs

Pixels/CCD

Amps/CCD

Digitization rate

Bytes/image

Data rate FEDAQ

Image CCDs Guide CCDs

DES data rates are relatively high by astronomy standards, but not for particle physics.

• We will use the Monsoon data acquisition system, developed by NOAO.

• We will modify it to separate digital and analog functionality.

Using Monsoon shortens development time and enables collaboration with NOAO and other Monsoon users.


DES Modifications:ADCs will reside on the telescope.The rest of the electronics will be off the telescope.

• Save space and power on the telescope.

• Reduce noise (ADCs are closer to the CCDs).

• Save money.

Data Acquisition

LINUX PCPCI FIBER CARD

Ethernet Link100Mb/s

1Gb/s Fiber(50Mpixel/s)




SYNC


SYNC


N NODES

SUPERVISORY NODELINUX PC


CCDor

FPA

10Mb/sEthernet

10Mb/sEthernet

10Mb/sEthernet

CCDor

FPA

CCDor

FPA

CCDor

FPA

CCDor

FPA

CCDor

FPA

CCDor

FPA

CCDor

FPA

CCDor

FPA

CCDor

FPA

CCDor

FPA

CCDor

FPA

SYNC SYNCN NODES

SUPERVISORY NODELINUX PC

CCDor

FPA

10Mb/sEthernet

10Mb/sEthernet

10Mb/sEthernet

CCDor

FPA

CCDor

FPA

CCDor

FPA

CCDor

FPA

CCDor

FPA

CCDor

FPA

CCDor

FPA

CCDor

FPA

CCDor

FPA

CCDor

FPA

CCDor

FPA

PIXEL ACQUISITION NODE 1

DETECTOR HEADELECTRONICSNODE 1

SYNC SYNCN NODES

SUPERVISOR NODELINUX PC

CCDor

FPA

10Mb/sEthernet

10Mb/sEthernet

10Mb/sEthernet

CCDor

FPA

CCDor

FPA

CCDor

FPA

CCDor

FPA

CCDor

FPA

CCDor

FPA

CCDor

FPA

CCDor

FPA

CCDor

FPA

CCDor

FPA

CCDor

FPA

PIXEL ACQUISITION NODE 2 PIXEL ACQUISITION NODE 3



We will modify this

part.

Monsoon architecture:


We Can Do This!

The Silicon facility at Fermilab has experience building the Run 0, I, & II silicon vertex detectors:

° Micron precision assembly ° Wirebonding ° Thermal Management ° CleanroomsBuilding a CCD focal plane uses many of the same skills, but has many fewer devices.

LBNL has extensive experience with CCD development and packaging for SNAP/JDEM

UIUC has experience building large, high rate data acquisition systems at SLAC, Fermilab, and Cornell.

U Chicago has experience with optical design and optical systems on SDSS° DES does not depend on pioneering development work.° The main issues are cost, schedule, and integration.

The DES collaboration has assembled a team of experienced scientists, engineers, designers and technicians


Schedule Milestones

• Optics and CCDs are the most Challenging tasks• CCDs: Preproduction run: FY05, Production run: FY06 and FY07• Optics: Order glass in FY06, Figuring/polishing in FY07

ID Task Name Start

10 Funds available (Fermilab R&D, external) Fri 10/29/04

38 Ready to submit masks to Dalsa Thu 12/2/04

40 Order 24 wafers of devices Thu 12/30/04

43 1st fu lly processed CCDs (8) in hand Thu 10/6/05

46 Submit order for final processing of remaing 1st batch wafers Thu 10/13/05

58 First production CCDs in hand Thu 7/6/06

17 FY07 funding available Tue 10/31/06

363 Order Corrector elements Tue 10/31/06

146 CCDs Ready for mounting on focal p lane Tue 1/15/08

368 Lenses and filters complete Mon 1/21/08

335 Dewar closed, ready for testing Tue 3/11/08

265 Camera testing complete - ready for corrector Tue 8/26/08

423 corrector, fi lters and cage ready for camera Mon 9/1/08

453 Prime focus cage complete Mon 10/27/08

455 Testing complete Mon 1/5/09

457 Ready to Ship to Chile Mon 1/19/09

489 Ready to Mount on Blanco Mon 3/16/09

491 1st light Mon 4/27/09

493 1st useful data Mon 5/25/09

Qtr 3 Qtr 1 Qtr 3 Qtr 1 Qtr 3 Qtr 1 Qtr 3 Qtr 1 Qtr 3 Qtr 1 Qtr 3 Qtr 12004 2005 2006 2007 2008 2009

Fully Commissioned byJune 2009!


Total Cost profile in Then Yr $(excluding institutional overhead)

FY04 FY05 FY06 FY07 FY08 FY09 TotalM&S 0 1,097 3,321 3,362 520 68 8,368M&S Contingency 0 0 379 338 1,580 782 3,079Total M&S 0 1,097 3,700 3,700 2,100 850 11,447

Labor 609 1062 955 1150 560 299 4,635Labor Contingency 304 531 478 575 280 150 2,318Total Labor 913 1593 1433 1724 839 449 6,953

Total (M&S + Labor) 913 2,690 5,133 5,424 2,939 1,299 18,400

The Reference Design represents our current choices for meeting the science goals

Total cost for the Instrument project is $18.4 M excluding institutional overheads and 22.5M$ with overhead in then year $.

We will be ready for observations by June 2009. This schedule is funding limited.


Instrument Project Organization


Conclusions

• We have a strong collaboration with a wide variety of skills that cover all aspects of this project

• With this collaboration we can complete the instrument and start survey operations on the telescope in 2009

dark energy survey (des)

Documents