KI V² Sensitivity and Observing parameters

This page contains KI sensitivity and instrument details for proposal and observation planning for V2 observations. In the V2 mode the main data product is the normalized visibility amplitude squared for a given source. Integrations on the target source and calibrators are interleaved in time, where each integration comprises fringe tracking data and all necessary calibrations.

The V2 mode is open for proposals from all Keck users (UC, Caltech, UH, NASA and NOAO TACs). The sensitivity values below are based on past interferometer performance in good weather. The interferometer performance degrades substantially in poor seeing.

Sensitivity numbers updated: August 3, 2009

New for semester 2010A

The sensitivity values given below are based on current performance for these modes and while every effort will be made to perform proposed science observations, no guarantees are made for shared-risk observing.

AO (adaptive optics) sensitivity:

• R < 12 (somewhat dependent on source color, less sensitive than AO alone due to the dichroic which sends light to interferometer)
• See the Keck Observatory AO page for more details.
• Offset from a nearby AO star to the interferometry target is supported but this mode has a lower observing efficiency. The maximum offset is roughly 15 arcsec, but is position dependent, please contact NExScI for details. The minimum offset is ~0.2 arcsec, but see the angle tracking section for other options for closer binaries.
• Note that the image rotators are used in vertical angle mode.

Angle tracking

• V2:
  • J-band, 80 Hz: 5.0 < J < 10.5
  • H-band, 80 Hz: 5.0 < H < 10.5 (for fringe tracking in K-band only)
• V2-SPR:
  • H-band, 80 Hz: 5.0 < H < 9.0
• For bright sources, a neutral density filter can be used to shift the range by up to 4.2 mags.
• The FOV is 0.6 arcsec with a plate scale of 0.04 arcsec/pixel.
• For binaries in which both components in are in the angle tracker FOV, a single component can be selected and sent to the fringe tracker if the separation is greater than ~0.1 arcsec. However, note that this mode results in decreased observing efficiency.

Fringe tracking:

• General notes
  • There are several KI visibility modes covering different wavelengths and spectral resolutions. Only two modes may be run within a single night, the allowed combinations and switching overhead times are given (need link here).
  • V2: The fringe tracker performs science observations in EITHER K (2.2 microns) or H (1.6 microns) band. Note that the frame rate must be the same for an entire cluster (target + calibrators), but that it can be changed between clusters. If an aperture is used for attenuation, it must also remain in place for the entire cluster.
  • V2-SPR: The primary fringe tracker stabilizes the fringes in K-band for the secondary fringe tracker which performs science observations in K-band using a longer integration time (1Hz) to obtain high dispersion (R~1700; 330 channels). The spectral range covers from 2.0 to almost 2.4, including the HeI, Br gamma and at least 3 of the CO bandheads.
  • L band: The L-band primary fringe tracker performs science observations at L-band. The calibration accuracy of this mode is not yet quantified due to the background contribution but will be better for brighter sources.
  • Simultaneous K and L-bands: In this configuration, the light from each Keck are divided as for the nulling mode and half sent to both the K and L-band fringe trackers. This results in a 1.2 mag sensitivity loss compared to the values below. The K-band camera can be set for 5, 10 or 42 channels.
• Sensitivities

  The following table summarizes the best sensitivity for each mode, see the text below for details.

  Wavelength	# Spectral Channels	Spectral Resolution	Magnitude limit
  H	4	22	9.0
  K'	5	27	10.3
  K'	10	54	9.5
  K'	42	230	7.6
  K'	330	1900	7.0
  L	10	63	6.0
  K'/L	5/10	27/63	9.1/4.8

  The limits below and in the table apply to unresolved targets, see table below for corrections for resolved sources.

  • Low resolution modes:
    • 5 pixels across K'
    • 200 Hz: 6.2 < K' < 9.9
    • 100 Hz(**): 6.6 < K' < 10.3

    • 10 pixels across K'
    • 200 Hz: 5.4 < K' < 9.1
    • 100 Hz(**): 5.8 < K' < 9.5

    • 4 pixels across H
    • 200 Hz: 5.3 < H < 9.0

    • 10 pixels across L
    • 100 Hz: 2.5 < L < 6
  • Medium resolution mode
    • 42 pixels across K'
    • 200 Hz: 3.5 < K' < 7.2
    • 100 Hz(**): 3.9 < K' < 7.6
  • High resolution mode (V2-SPR):
    • Primary fringe tracker (OPD stabilization)
      • 5 pixels across K', fringe stabilization
      • 200Hz: 4.45 < K' < 7.0
    • Secondary fringe tracker
      • 330 pixels across K', long integrations
      • 1Hz: limits set by primary fringe tracker
  • (**) NOTE: The spectral channel relative visibility calibration of 0.02 has only been validated at 200 Hz. Although this is considered low risk, the 100 Hz mode may have higher relative flux errors.
  • For bright sources, an aperture can be used to shift the magnitudes by ~2.4 mags.
  • The DN-to-electron conversion is 0.26.
  • As KI is a fringe-tracking interferometers, all targets and calibrators must exceed a threshold visibility to be observable. The limits given above are appropriate for unresolved (i.e. point-like) sources. If the target is resolved, the instrument sensitivity decreases as shown in the table below. Targets can not be larger than 4.5 mas diameter at K band or 3.3 mas diameter at H band. Please contact NExScI if you have questions about the observability of potential targets.

    The table below gives the reduction in sensitivity (X given in magnitudes) for a given source diameter.

    X K_mag	Diameter (mas)	X H_mag	Diameter (mas)
    0.00	< 1	0.00	< 0.7
    0.25	1.6	0.25	1.2
    0.50	2.2	0.50	1.6
    0.75	2.7	0.75	2.0
    1.00	3.1	1.00	2.2
    1.25	3.4	1.25	2.5
    1.50	3.7	1.50	2.7
    1.75	3.9	1.75	2.8
    2.00	4.2	2.00	3.0

• Absolute and relative calibration levels
  • Visibility mode verification tests have shown a 0.03 systematic bias in the squared visibility at H and K bands. Users should not assume measurements of absolute visibility to better than this level. The spectral channel-to-channel relative visibility systematic has been measured at 0.02 in squared visibility for the 4 and 5 channels modes and 0.01 for the 10 and 42 channels modes. See the flux bias memo for more details.
  • For the V2-SPR mode (330 channels), the channel-to-channel performance is expected to be 0.02. The expected absolute visibility-square accuracy based on preliminary measurements is 0.05. Differential phase measurements have been demonstrated to ~6 deg.

• Spatial resolution and field of view
  • Nominal fringe spacing (wavelength/baseline) for the 85 m Keck-Keck baseline:
    • K: 5.2 mas
    • H: 3.8 mas
  • Fringe tracker field-of-view: ~50 mas FWHM at K' band, scales with wavelength (set by a single-mode fiber matched to the diffraction limit of a single telescope)

  • Keck-Keck baseline is 85 meters and oriented approx. 38 degrees East of North. Plots of the u-v Coverage (as a function of declination for a 50 degree zenith angle range):
    • 50 deg dec
    • 35 deg dec
    • 25 deg dec
    • 15 deg dec
    • 5 deg dec
    • 0 deg dec
    • -5 deg dec
    • -15 deg dec
    • -25 deg dec

Sky coverage:

• Delay Range:
  • continuous (fast delay lines - FDL ): +/- 15 m
  • quasi-static (long delay lines - LDL): +/- 70 m

  The total delay is calculated by summing the FDL and LDL values. Zenith pointing occurs at a delay of approx. -40 meters, which means that targets at high and low declinations can not be observed at all hour angles. We recommend using getCal to see the detailed coverage of specific objects.

  Up to different 4 long delay line positions can be used in a single night. The long delay lines take 15 minutes to re-position and align and should be changed only between groups of targets and calibrators. The timing gui in getCal plots the detailed coverage for a given LDL position.

• Zenith Range:
  • Good performance for zenith angles < 50 deg.
  • Observations can be made at somewhat higher zenith angles, but the performance degrades.
  • The hard limits depend on the azimuth and telescope, and are included in the getCal calculations.
  • 3 deg radius around zenith excluded, degraded performance is sometimes seen within 5 degrees of zenith

Observing efficiency:

For sources where all magnitudes (V,J,K) are > 1 magnitude above the sensitivity limits, KI can achieve 6 scans per hour under good weather and seeing conditions. A scan includes the fringe tracking data and calibrations. A separate scan on a nearby calibrator is required to determine the system visibility. This means that for a source/calibrator pair, there will be 3 integrations on each in an hour.

For sources at the sensitivity limit of one of the sub-systems, the efficiency is generally closer to 4 scans per hour.

The first 30 minutes after the domes open may be necessary for interferometer set-up and no science observations are guaranteed during this time.

Return to the KI Support page