A New Spectroscopic Facility for the Kiso Schmidt

1
The 6dF Galaxy Survey - The First YearWill
SaundersAnglo-Australian Observatory
2
Salient Dates

• Feb 2001 6dF delivered to UKST
• May 2001 first useful 6dFGRS data
• June 2001 6dF declared commissioned
• Aug 2001 runz useable
• Sept 2001 acquisition improved
• Nov 2001 merged sample adopted
• Dec 2001 6dFDR workable
• Dec 2001 much photography
• Feb 2002 R-arm failure
• Mar 2002 Robot repair and recalibration
• Mar 2002 mirrors coated
• Mar 2002 first fibre repairs
• Apr 2002 adaptive tiling adopted
• Jul 2002? Spectrograph VPH upgrade
• Sep 2002? First data release

3
Data Reduction

• Uses 2dFDR. Many teething problems adapting to
  6dF data
• Now working very reliably, public version issued.
• Automated fibre identification using known gaps
• Outstanding concerns
• sky subtraction imperfect (as per 2dF)
• wavelength calibration could be improved
• need to splice V and R spectra
• only tramline extractions for now

4
Acquisition

• Serious problems throughout first year.
• Guiding uses 4 x 7 fibre guide bundles.
• Fibres too fragile, many (50) broken during
  commissioning and since.
• Usually 2-3 useable, not enough
• Repairs impossible, redesign completed and
  remanufacture in progress.
• May move to 75um fibres from current 100um

5
Fibre breakages

• Very sensitive to accurate robot calibration
• 30 fibres (10) broken during commissioning
• normal breakage rate 1/week
• March/April breakage rate much larger due to
  imperfect calibration after R-arm failure
• Now much better, aim is max 2 fibres/lunation
• Many fibres damaged, poor throughput or fringed.
• Only one field plate in use currently while other
  undergoes repair.
• Repairs very difficult and slow but improving
• Field Plate 1 120 fibres, Field Plate 2 90
  fibres

6
Robot Positioning

• Accuracy mostly excellent. Problems with homing
  of encoders on startup caused instability and
  loss of few nights data.
• Current uncertainty in precise robot plate scale
  at 0.0001 level
• many manual fibering interventions required.
• Configuring time 1 hour (with all fibres!),
  little effort made to tune up.
• New appointment Martin Oestreich will watch over
  mechanics.

7
Quality Control

• Initially very erratic - just trying to sort out
  problems preventing data being taken at all.
• Many small changes to spectrograph alignment,
  focus, wavelength range.
• Very little feedback possible prior to 6dFDR
• Now more stable but still need real-time S/N
  calculator to determine exposure times
• Biggest problem acquisition.
• No astronomer present for bulk of 6dFGS
  observations
• poor feedback loop timescales
• 6dF complex and delicate instrument, must expect
  variable quality at some level.

8
(No Transcript)
9
Spectrograph Upgrade

• Current peak efficiency 8 (cf 2dF 5)
• 6dF camera overfilled by spectra at ends of slit
  and ends of wavelength range.
• Existing reflection gratings good but still only
  65 efficiency at peak.
• Will shortly switch to VPH gratings, 580V
  (3900A-5600A) and 425R (5350A-7650A) at same
  resolution. Should reduce losses for both
  reasons, hope for 50 efficiency gain.
• Mirrors coated March 2002 but no efficiency gain.

10
Survey Progress

• 2001
• May 11
• Jun 17
• Jul 0
• Aug 19
• Sep 19
• Oct 19
• Nov 20
• Dec 2
• 2002
• Jan 13
• Feb 0
• Mar 25
• Apr 47
• May 42
• Total 232

• 6dFGS progress has been erratic
• with many interruptions due to other programs and
  instrumental problems. However, the rate at which
  fields are observed has continually increased.
  The last two months have had excellent weather,
  and have achieved record numbers of fields
  despite only having one field plate for May.
  However, both field plates have large numbers of
  broken fibres.
• 20,000-25,000 galaxies observed.
• Relatively few additional targets due to
  priority weighting

11
Data Quality

• 6dF capable of taking excellent data - average
  S/N gt 10/pixel in both V and R in 1 hour 1/2
  hour, 10-20 objects/frame with gt1000
  electrons/pixel in V.
• Poor acquisition, robot calibration, seeing or
  transparency can easily halve S/N or worse
• No real-time S/N calculator developed yet.
• Large fraction of data compromised by either
  observational or instrumental problems.
• Data quality and consistency generally improving.
• Concern over whether quality is good enough for
  Dn-s.

12
Data Reduction Progress

• reliable 6dFDR package only available Dec 2001.
• 1/3 of data reduced
• 1/5 of data put through runz
• Further work needed to splice V and R data
• Velocity dispersion techniques not yet in place
• Completeness very variable, 60-95.

13
6dFGS prospects

• Minimum 1100 fields to observe in 3 years
• gt 28/lunation gt required average 4/clear night
• Spectrograph upgrade may allow extra field/night,
  or better higher S/N.
• Ways to streamline observing and/or trim the
  sample extremely desirable
• Doing better than 2dFGRS