Uploaded image for project: 'Data Management'
  1. Data Management
  2. DM-37812

Evaluate flat illuminator flats

    XMLWordPrintable

Details

    • Improvement
    • Status: Done
    • Resolution: Done
    • None
    • None
    • 4
    • Data Release Production

    Description

      From cslage on slack:

      The new dome flat illuminator is installed, but still has some issues. With the monochromator in place, the light intensities are quite low. However, there is an option to put a mirror in the monochromator so the screen is illuminated with white light. The light does not fully illuminate the dome screen (see photo), but flats taken with this condition are still flat within a few %. Maybe this is acceptable and we could replace the current system of manually changing LED lights. We took 20 flats with each of the three filters today.
      czw
      , if you have a minute among the many other things I know you are doing, can you weigh in on whether these flats will be acceptable?
      SDSSr_65mm - 2023020100023-43
      BG40_65mm_1 - 2023020100053-72
      OG550_65mm_1 - 2023020100074-93

      Attachments

        1. bg40_20230207.png
          bg40_20230207.png
          156 kB
        2. bg40.png
          bg40.png
          148 kB
        3. BG40-mono_20230202b.png
          BG40-mono_20230202b.png
          101 kB
        4. cpVerifyFlat-BG40-mono.ipynb
          238 kB
        5. cpVerifyFlat-OG550-mono.ipynb
          212 kB
        6. cpVerifyFlat-SDSSr-mono.ipynb
          192 kB
        7. dm37812.log
          5 kB
        8. monochromatic.log
          6 kB
        9. og550_20230207.png
          og550_20230207.png
          170 kB
        10. og550.png
          og550.png
          160 kB
        11. OG550-mono_20230202c.png
          OG550-mono_20230202c.png
          84 kB
        12. sdssr_20230207.png
          sdssr_20230207.png
          152 kB
        13. sdssr.png
          sdssr.png
          146 kB
        14. SDSSr-mono_20230202a.png
          SDSSr-mono_20230202a.png
          77 kB

        Activity

          I don't think these are acceptable. In the following pictures, I'm plotting the amp means of the verifyFlatProc images (serial overscan, bias, dark, flat corrected using the flat generated from these images) as a function of seqNum, with the prior run on top, this run on the bottom, and no change between left/right (this is an existing plot product I'm reusing). The data from today seem to have a ~5% dropoff, compared to the previous data which was stable to a much greater degree (the amps are offset by a constant relative to C10).

          I haven't directly looked at these images yet, and the cp_verify result does claim they're fine (it's using ~the average of the inputs, so I don't think it rises above the current threshold, which probably isn't set well).

          czw Christopher Waters added a comment - I don't think these are acceptable. In the following pictures, I'm plotting the amp means of the verifyFlatProc images (serial overscan, bias, dark, flat corrected using the flat generated from these images) as a function of seqNum, with the prior run on top, this run on the bottom, and no change between left/right (this is an existing plot product I'm reusing). The data from today seem to have a ~5% dropoff, compared to the previous data which was stable to a much greater degree (the amps are offset by a constant relative to C10). I haven't directly looked at these images yet, and the cp_verify result does claim they're fine (it's using ~the average of the inputs, so I don't think it rises above the current threshold, which probably isn't set well).

          I've added a similar analysis of the monochromatic flats: these have been processed with the white-light flats constructed yesterday.  These do not have the flux drift seen in the white-light data.

          czw Christopher Waters added a comment - I've added a similar analysis of the monochromatic flats: these have been processed with the white-light flats constructed yesterday.  These do not have the flux drift seen in the white-light data.

          Ok, I'm not that surprised by the white light being unstable at all. There is no field lens in the system that basically homogenizes the beam.
          The monochromatic light uses the same system, so I'm a bit surprised that is more stable.... I need to think on that.

          pingraham Patrick Ingraham (Inactive) added a comment - Ok, I'm not that surprised by the white light being unstable at all. There is no field lens in the system that basically homogenizes the beam. The monochromatic light uses the same system, so I'm a bit surprised that is more stable.... I need to think on that.

          pfagrelius pointed out that the variation is in time... so ignore my comment.

          pingraham Patrick Ingraham (Inactive) added a comment - pfagrelius pointed out that the variation is in time... so ignore my comment.
          cslage Craig Lage added a comment -

          I took a new set of white light flats with the new dome flat illuminator. I let the lamp warm up for over an hour and 30 minutes after raising the power from 910 W to the full 1000W. Based on the RubinTV output, these look like the stability is better than 1%. 20 flats with each filter at ~ 20,000 ADU

          SDSSr_65mm_1 - 2023020700022-41
          OG550_65mm_1 - 2023020700042-61
          BG40_65mm_1 - 2023020700062-81

          cslage Craig Lage added a comment - I took a new set of white light flats with the new dome flat illuminator. I let the lamp warm up for over an hour and 30 minutes after raising the power from 910 W to the full 1000W. Based on the RubinTV output, these look like the stability is better than 1%. 20 flats with each filter at ~ 20,000 ADU SDSSr_65mm_1 - 2023020700022-41 OG550_65mm_1 - 2023020700042-61 BG40_65mm_1 - 2023020700062-81

          I've uploaded a comparison of the 20230201 white-light data with these new images, and they definitely look more stable than before.

          czw Christopher Waters added a comment - I've uploaded a comparison of the 20230201 white-light data with these new images, and they definitely look more stable than before.

          I'm marking this as done.  I think this last set showed that letting the lamp warm up helped.  If we need to check this more, I'll open a new ticket.

          czw Christopher Waters added a comment - I'm marking this as done.  I think this last set showed that letting the lamp warm up helped.  If we need to check this more, I'll open a new ticket.

          People

            czw Christopher Waters
            czw Christopher Waters
            Christopher Waters, Craig Lage, Elana Urbach, Parker Fagrelius, Patrick Ingraham (Inactive)
            Votes:
            0 Vote for this issue
            Watchers:
            5 Start watching this issue

            Dates

              Created:
              Updated:
              Resolved:

              Jenkins

                No builds found.