8.5 ACQ PEAKXD Acquisition Mode

An ACQ/PEAKXD sequence is used to improve centering in the cross-dispersion (XD) direction. We recommend a minimum S/N of 40 for all dispersed-light target acquisition exposures.

8.5.1 NUV PEAKXD Acquisitions

The steps executed for NUV ACQ/PEAKXD sequences are:

1. A short exposure of the Pt-Ne wavelength calibration lamp through the WCA aperture is obtained. The spectrum is collapsed along the dispersion direction, its centroid is calculated, and the center of the target aperture is computed.
2. A target spectrum is recorded for the user-specified time using a sub-array tailored to each grating and central wavelength (excluding edge effects and airglow lines). The spectrum is collapsed along the dispersion direction.
3. The target XD location is assumed to be the median position of the collapsed spectrum.
4. The slew required to move the target spectrum in the XD direction to the center of the aperture is computed.
5. The telescope is slewed by the calculated offset to center the target in the XD direction.

The user must specify the aperture (PSA or BOA, typically the same as for the science exposure), the grating and central wavelength, and the exposure time. The use of MIRRORA or MIRRORB is not allowed. The stripe (SHORT, MEDIUM, or LONG, corresponding to stripes A, B, or C) to be used in the computation may be specified; however, the default stripe B (MEDIUM) is recommended for most settings, as it achieves the best centering.¹

Note: For extended sources observed with the NUV detector, light from the three spectral stripes may overlap. In this case acquisitions will likely fail and should be avoided.

Due to increased NUV detector background and mechanism position uncertainties, certain NUV central wavelength (cenwave) settings have proven to be unreliable in the NUV ACQ/PEAKXD target acquisition phase. The cenwave settings listed below have been proven reliable. NUV ACQ/PEAKXD target acquisitions should use one of these settings. Users who wish to use other cenwaves should contact the COS team via the Help Desk.

8.5.2 FUV PEAKXD Acquisitions

As described in Section 4.1.7, the COS FUV detector exhibits gain sag. Gain sag effects are alleviated by periodic lifetime moves, which place the FUV spectrum on unsagged regions of the detector (see Section 5.12).

One consequence of the gain sag effect is the mis-registration of photon events in the cross-dispersion (XD) or Y direction, commonly referred to as Y walk. While Y walk does not adversely affect science data, it can reduce the accuracy of target acquisitions obtained in dispersed light. If the target is centered in the aperture, but the Y walk shifts its spectrum in the XD direction, then the FUV ACQ/PEAKXD algorithm at previous lifetime positions (LP1–LP3) would miscalculate its centroid and move the target away from the aperture center.

The FUV Segment B is more affected by Y walk and gain sag than Segment A, mainly due to the intense geocoronal Lyman-α emission that hits the detector during G130M observations. To combat the effects of Y walk, the LP1–LP3 FUV ACQ/PEAKXD algorithm was modified to use only Segment A data.

From Cycle 25 onward (October 2, 2017), the FUV ACQ/PEAKXD sequence differs from the algorithm used in previous cycles. It uses only the total number of counts and is not affected by Y walk. For FUV ACQ/PEAKXD acquisitions, either segment A or B may be used when available, but use of SEGMENT=DEFAULT is recommended. All G140L ACQ/PEAKXD acquisitions and those with G130M cenwaves restricted by the COS 2025 rules use only segment A.  In the case of cenwave G140L/1280, detector segment B is left on by default but counts from segment B are not used for acquisition calculations. The ETC correctly calculates the needed exposure time for cenwave 1280 acquisitions, and also lists the count rate for segment B. Any count rate violation warnings for segment B are real and should not be ignored.

The new ACQ/PEAKXD algorithm works much like ACQ/SEARCH except that, instead of a spiral, the spacecraft is moved linearly along the XD axis between exposures. An array containing the total counts at each dwell point is constructed. Its centroid is computed, and the telescope is moved to center the target in the aperture in the XD direction. The user must specify the aperture, grating, central wavelength, and the exposure time at each dwell point. The use of MIRRORA or MIRRORB is not allowed. The number of steps, called NUM-POS, may be 3, 5, 7, or 9. The STEP-SIZE is given in arcseconds. There are three options for the centering algorithm, CENTER=FLUX-WT, FLUX-WT-FLR, and BRIGHTEST, and they work just as described in Section 8.3. For most applications, we recommend the use of NUM-POS=3, STEP-SIZE=1.3, and CENTER=FLUX-WT, as this combination (which is the default) is the fastest pattern that centers targets to within the requirements. Observers who wish to use NUM-POS=5 are advised to use STEP-SIZE=0.9 and CENTER=FLUX-WT-FLR. The special parameter CENTER=DEFAULT sets CENTER=FLUX-WT if NUM-POS=3 and CENTER=FLUX-WT-FLR if NUM-POS=5, 7, or 9. If NUM-POS is 5, 7, or 9, then the value of STEP-SIZE should be manually entered in APT. While any combination of NUM-POS and STEP-SIZE is allowed, it should be noted that for PEAKXD NUM-POS=3 with STEP-SIZE=1.3 and NUM-POS=5 with STEP-SIZE=0.9 are the only two combinations that are routinely tested.

Note: FUV ACQ/PEAKXDs on extended sources are possible, but they require a NUM-POS and STEP-SIZE combination tuned to the extent of the source.

¹ There are special restrictions with the G230L grating, which positions first-order light on the detector only for certain stripes. The MEDIUM stripe is required for the 2635 central wavelength, the SHORT stripe is required for the 3360 wavelength, and either SHORT or MEDIUM may be used for the 2950 and 3000 wavelengths. The LONG stripe may not be used with G230L at all.