Skip to main content

July 2025 STAN

View All COS STANs
July 15, 2025
COS NEWSLETTERS

About This Article

In this STAN, we cover major reference file updates improving FUV calibration, two new COS FUV lifetime positions for Cycle 33, Phase II submission best practices, and the COS team’s AAS presentation previewing the new Hubble Spectroscopic Legacy Archive.

Callout

End callout

 

Major Delivery of New Reference Files Following Updates to the Geometric Distortion and Walk Corrections

During the week of July 14, 2025 the COS team delivered 62 new reference files as part of the effort to improve the geometric distortion and walk corrections for the FUV detector. The geometric distortion and walk corrections effectively set the coordinate system in which the analysis of COS FUV data occurs, so changes to these files necessitated updates to all downstream reference files in the calibration process with positional or wavelength dependencies. The new reference files can be thought of as being members to two categories: (1) files that implement the new coordinate frame; (2) files that calibrate data within the new coordinate frame. Note that all of the new reference files should be used together, and these should not be mixed-and-matched with older versions of reference files. This update affects all COS FUV data, and users should re-download data products from MAST regardless of observing date or lifetime position to receive the most up-to-date calibrated data products.

Updated reference files that implement the new coordinate frame are: GEOFILE, DGEOFILE, XWLKFILE, YWLKFILE, and BRFTAB. The delta-geometric correction and x-walk correction reference files (DGEOFILE and XWLKFILE) had previously been dummy files (full of zeros), and both will now have an impact on calibration. The y-walk correction reference file (YWLKFILE) had previously contained arrays of size 16384x32, which have now been reduced to 1024x32. This was done to match the size of the FUV detector segments in the y-direction in pixels to implement a new y-dependent y-walk correction. In order to utilize this new correction, users must install CalCOS v3.5.0 or later. Detailed descriptions of the project to improve the geometric distortion and walk corrections can be found in the following ISRs:

Updated reference files that calibrate data within the new coordinate frame are: SPOTTAB, GSAGTAB, BPIXTAB, TDSTAB, XTRACTAB, TWOZXTAB, PROFTAB, TRACETAB, LAMPTAB, DISPTAB, FLATFILE, and FLUXTAB. These file types have been updated for all lifetime positions and all USEAFTER dates. As part of this process, TWOZONE extraction has been implemented for the G130M/1055 and G130M/1096 settings at LP2, and the special GSAGTAB that had only applied to those two settings has been retired. Detailed descriptions of the project to re-derive the downstream reference files as part of this work will be documented in a series of ISRs released throughout 2025 and 2026. While this major reference file delivery has been pending, the COS team continued monitoring the flux calibration provided by the existing TDSTAB and found that for some recent observations the absolute flux calibration was not operating within the quoted 5% error level. Users with a particular need for accurate flux calibration who have G160M data taken in 2024 or later should re-download their data products from MAST, as the new combination of reference files improves the absolute flux calibration to remain well within the promised accuracy.

Unrelated to the above corrections, but implemented at the same time, is the new high voltage dependent sensitivity correction (HVDSCORR), which utilizes the new HVDSTAB reference file. This correction accounts for small changes in sensitivity when the detector is operated at different high voltage levels. It is applied to individual photon events via the photon weight column EPSILON in corrtag files, and scales values with respect to a reference high voltage level, at which the absolute flux calibration is defined. Although this module was first introduced to CalCOS in v3.5.0, the HVDSTAB has been a dummy file until now. The new HVDSTAB should be used along with the updated FLUXTAB and TDSTAB files to provide proper flux calibration for COS FUV data. 

Together, the new geometric distortion, delta-geometric, x-walk, y-walk, and high voltage dependent sensitivity corrections, along with the updated downstream reference files, have improved the wavelength calibration and flux calibration of COS FUV data. The accuracy of the overall wavelength assignment has improved from about 1/2 of a resolution element (3 pixels) to about 1/4 of a resolution element (1.5 pixels). The maximum systematic offset between wavelengths assigned to features at different locations on the detector has been reduced from about 20 km/s to about 10 km/s for medium resolution modes. An example comparison of the wavelength residuals for data processed with the old and new suites of reference files is shown in Figure 1. Proper flux calibration at the edges of the detector segments has significantly improved with the new reference files. Artificial flux spikes and deficits that had been caused by the old geometric distortion correction have been removed, and errors in the absolute flux calibration for wavelengths recorded at the edges of the detector have been reduced to well within the 5% requirement for the entire history of COS observations, as shown in Figure 2.


Key information for users wishing to process data on a local machine:

  • Full list of updated reference files to download from CRDS:
    BPIXTAB, BRFTAB, DGEOFILE, DISPTAB, FLATFILE, FLUXTAB, GEOFILE, GSAGTAB, HVDSTAB, LAMPTAB, PROFTAB, SPOTTAB, TDSTAB, TRACETAB, TWOZXTAB, XTRACTAB, XWLKFILE, YWLKFILE
  • Minimum version of CalCOS required to use y-dependent y-walk correction and high voltage dependent sensitivity correction:
    CalCOS v3.5.0
Residual distortions in the wavelength solution using the old geometric distortion correction (top) and the new geometric distortion, delta-geometric, and walk corrections (bottom).
Figure 1: Residual distortions in the wavelength solution using the old geometric distortion correction (top) and the new geometric distortion, delta-geometric, and walk corrections (bottom). The width of the distribution decreases from 1.71 pixels to 0.89 pixels, and the magnitude of coherent structures visible in the top panel is reduced in the bottom panel. 
Data points show residuals in the absolute flux calibration for the G130M/1291 mode in the 1290-1295 A wavelength range with respect to CALSPEC models of WD0947+857 (before 2012-07-23) and WD0308-565 (after 2012-07-23).
Figure 2: Data points show residuals in the absolute flux calibration for the G130M/1291 mode in the 1290-1295 A wavelength range with respect to CALSPEC models of WD0947+857 (before 2012-07-23) and WD0308-565 (after 2012-07-23). Black circles show flux residuals using the suite of old reference files, while blue circles show residuals using the suite of new reference files. The new results show clear improvements over the old results.

Back to Top

Multiple New Lifetime Positions in Cycle 33

The COS FUV detector experiences gain sag, where the ability of a given location on the detector to record photons diminishes with repeated exposure to light. To mitigate this, every few years a new location in the cross-dispersion direction of the detector is introduced for the spectra to fall on the COS FUV detector. In Cycle 33, we introduce Lifetime Position 7 (LP7) and LP10 in combination with previous LPs, LP3 and LP5. Table 1 describes the lifetime position that corresponds to each COS FUV mode.

  • G130M cenwaves 1055, 1096, 1222, and 1291 will move from LP2, LP4, and LP5 to the new LP7. 
  • G130M cenwaves 1300, 1309, 1318, and 1327 will remain at LP5.
  • G160M cenwaves will move from LP6 to the new LP10.
  • G140L cenwaves will remain at LP3.

Although the spectral resolution and sensitivity vary slightly with LP, these changes are mostly transparent to users. An exception is the use of LP7, which carries increased overheads that are discussed in the next article. The characteristics of lifetime positions through LP6 are discussed in COS documentation (A summary is available in the COS Data Handbook), while the characteristics of LP7 and LP10 will be covered in current and forthcoming instrument science reports.

The Cycle 33 lifetime positions of supported modes.
Table 1: The Cycle 33 lifetime positions of supported modes.

Back to Top

Best Practices for COS Phase II Submissions

Submitting a safe, technically sound Phase II file that adheres to all our policies will reduce the risk of scheduling delays. Users are reminded to:

  • Provide ETC ID numbers for all exposures, including acquisition exposures.
  • Verify that each exposure is safe by running the Bright Object Tool (BOT) in APT, using the GALEX catalog whenever possible. Programs submitted with unexplained BOT warnings may lead to a delay in scheduling the observations.
  • If a target consists of multiple sources, a target field is crowded, or a target is faint, consider executing an offset target acquisition.
  • Specify the buffer time for all TIME-TAG exposures. Correctly calculating the buffer time is important to ensure that no data are lost during readout. In most cases, the buffer time should be 2/3 of the value calculated by the ETC, but there are exceptions for bright targets.
  • Follow the target list and instrument configuration specified in the approved Phase I proposal. Changes of grating, central wavelength, or lifetime position can be requested by a minor change request to your Contact Scientist, provided there is no change to the science goals. More substantial changes (such as target changes, more restrictive scheduling constraints, instrument changes, and anything that alters the science goals) need to be requested by a major change request to the Telescope Time Review Board (TTRB), using the link from the Program Status webpage.
  • Consider alternative observational approaches to achieve your science goals if observations require a string of more than 6 consecutive orbits, as such strings will execute at shared risk (i.e., it will not be eligible for repeat if impacted by observatory problems). 

Changes in Overheads for FUV Spectroscopy with the G130M Grating at Lifetime Position 7 and G160M Grating at Lifetime Position 10

At the beginning of Cycle 33, LP7 will become the default for G130M spectroscopy using 1055, 1096, 1222, and 1291 cenwaves, with G130M cenwaves greater than 1291 remaining at LP5. Similarly, LP10 will become the default for G160M spectroscopy for all cenwaves. Similar to observations at LP6, G130M observations at LP7 will experience increased wavelength calibration overheads due to the introduction of SPLIT Wavecals. On the other hand, G160M observations at LP10 will no longer require SPLIT Wavecals. The Cycle 33 COS Instrument Handbook discusses the wavelength calibration overheads at LP6 due to the inclusion of SPLIT Wavecals. Scientists proposing the use of the G130M grating with cenwaves 1055, 1096, 1222, or 1291 in Cycle 33 are strongly encouraged to read this material regarding the overheads incurred at LP7. A key difference for LP7 is that there are no alternative LP options if users have a high SNR requirement that necessitates 4 FP-POS within a single orbit. APT Version 2025.3.4 includes these overheads at LP7 in support of the Cycle 33 Phase II deadline. Users are encouraged to contact the COS team for assistance in advance of the Phase II deadline if they have questions.

COS FUV Lifetime Policy and SNR Guidelines

To help preserve the COS FUV detector’s performance into the 2030s, STScI has introduced a policy starting in Cycle 33 that caps lifetime usage at 2% per program per Lifetime Position (LP). Any usage beyond 1% must have been justified in Phase I, explaining why alternative observing modes or STIS cannot be used instead. Proposers must estimate lifetime usage and ensure total counts per pixel stay within the allowed limits. Optimizing setups — for example, by adjusting FP-POS use, turning off segments, or shortening exposures — is strongly encouraged to stay under the cap. Full details and example calculations are provided in the March 2025 STAN.

Furthermore, observers are reminded that the maximum achievable SNR for COS is limited by fixed pattern noise, which depends on the grating and the number of FP-POS used (see the 50th percentile values in COS ISR 2023-11). Planning exposures that exceed this limit does not yield higher-quality data but does shorten detector life. Observers should keep total exposure times across all orbits consistent with the realistic SNR ceiling for their mode. Anything exceeding that limit must have been justified in the Phase I proposal.

Following these guidelines during Phase II preparation ensures efficient use of COS while maximizing its lifetime and scientific impact. The COS team will review submitted programs for efficiency and may request changes if exposure times exceed practical SNR limits without clear justification. For any questions, please contact your COS Contact Scientist.

Reminder of the COS2025 Policies

Users preparing Cycle 33 Phase II submissions are reminded that the COS2025 policies are still in effect. These policies consist of restrictions on the choice of detector segment and FP-POS positions for the G130M observing modes. The policies are designed to maximize the FUV detector lifetime by minimizing the exposure of the FUVB  segment to geocoronal Ly α emission. Under COS2025, there are four G130M central wavelengths (cenwaves) that can be used with both detector segments on 1055, 1096, 1222, and 1291. For the other G130M cenwaves (1300, 1309, 1318, 1327) only segment FUVA can be on. Observations of the zero-redshift Ly α wavelength range can be performed at LP3. This strategy must be justified in the Phase I proposal and approved by the COS instrument team. Detailed information about the rules is available at the COS2025 policies page.

Back to Top

COS Poster at the 246th Meeting of the American Astronomical Society

A COS poster was presented at the AAS meeting in Anchorage, Alaska, and is available on the AAS website . We invite all COS users to review it to learn about the latest developments for COS. Additionally, all COS posters are archived on the COS poster page.

The HASP and HSLA programs: Legacy UV/Optical spectra from nearly 30 years of HST (R. Sankrit, the HSLA working group, and the COS team)

The Hubble Advanced Spectral Products (HASP) and the new Hubble Spectral Legacy Archive (HSLA) programs make available high-level scientifically validated spectra obtained by the Cosmic Origins Spectrograph (COS) and the Space Telescope Imaging Spectrograph (STIS) instruments on the Hubble Space Telescope (HST) spanning over 28 years. HASP provides automatic co-added spectra for nearly all individual visits and programs, while the HSLA will provide automatic co-added spectra, astrophysical classifications and searchable metadata for more than 6000 unique COS/STIS targets across programs. In this poster we will describe (i) the main features of the data products available from these programs, (ii) the target classification scheme implemented in HSLA, (iii) how the database may be queried and spectra retrieved and (iv) supplementary scripts (implemented as jupyter notebooks) for creating custom coadded spectra in several cases not covered by the standard HASP and HSLA pipelines.

Back to Top

 

 

 

HST Help Desk


Please contact the HST Help Desk with any questions.