Work done with the radius difference plots in issue #169 shows what appears to be issues with the occulter radii that come out of the L1 pipeline.
The first effect is that at the start of the waves program, there is a quick rise in occulter radii of about .3 pixels in both cameras. This is assumed to be some thermal expansion of the occulter itself.
0.3 pixels out of a radius of 340 pixels, then delta R / R = 8.8e-4. The expansion coefficient of aluminum is 21e-6 / C. 8.8e-4 / 21e-6 = 42C temperature change required to cause this.
This issue is seen on most long days without cloud interruptions. The assumption is that while taking darks and flats directly before the waves program, the occulter has time to cool (possibly as much as 42C), then quickly heats back up the 42C when we start taking data. The trend is also happening in the synoptic program, but here, the rapid rise occurs during the shorter wavelength lines, where the noise is higher, and the data cadence is lower, so we observe the effect.
The second issue is a slow differential change in occulter radii over a long observing day. In the attached plot, we ca see noise in both the rcam and tcam radii, but there is a clear trend of a drop in (rcam_raidus - tcam_radius) vs time. The interpretation here is a change in the RCAM focus position due to the change in the gravity vector on the camera mount. At sunrise, gravity pulls the camera toward the O1 and flips to pulling the camera away from the O1 at the end of the day. With TCAM, the gravity vector is either neutral or pulling away from the O1.
Steps going forward:
-Hardware measurements are used to understand if/how the cameras move over the day, and thermal measurements track how hot the occulter gets when observing.
To fix the occulter thermal radii change, we may need to consider if we are using the correct occulter diameter and possibly change our observing strategy to pre-heat the occulter. It is unlikely we will ever try to cool the ucomp occulter actively.
We can add a stronger spring to the camera mount linear stage to fix the differential change in occulter radii. When the camera stage was designed, we weren't planning to add the camera water cooling so that the stage may be undersized. However, if this doesn't work, or we want to fix the commission data, we can add a correction factor to the distortion correction to address the rcam focus drift.
Work done with the radius difference plots in issue #169 shows what appears to be issues with the occulter radii that come out of the L1 pipeline.
The first effect is that at the start of the waves program, there is a quick rise in occulter radii of about .3 pixels in both cameras. This is assumed to be some thermal expansion of the occulter itself.
This issue is seen on most long days without cloud interruptions. The assumption is that while taking darks and flats directly before the waves program, the occulter has time to cool (possibly as much as 42C), then quickly heats back up the 42C when we start taking data. The trend is also happening in the synoptic program, but here, the rapid rise occurs during the shorter wavelength lines, where the noise is higher, and the data cadence is lower, so we observe the effect.
The second issue is a slow differential change in occulter radii over a long observing day. In the attached plot, we ca see noise in both the rcam and tcam radii, but there is a clear trend of a drop in (rcam_raidus - tcam_radius) vs time. The interpretation here is a change in the RCAM focus position due to the change in the gravity vector on the camera mount. At sunrise, gravity pulls the camera toward the O1 and flips to pulling the camera away from the O1 at the end of the day. With TCAM, the gravity vector is either neutral or pulling away from the O1.
Steps going forward: -Hardware measurements are used to understand if/how the cameras move over the day, and thermal measurements track how hot the occulter gets when observing.
To fix the occulter thermal radii change, we may need to consider if we are using the correct occulter diameter and possibly change our observing strategy to pre-heat the occulter. It is unlikely we will ever try to cool the ucomp occulter actively.
We can add a stronger spring to the camera mount linear stage to fix the differential change in occulter radii. When the camera stage was designed, we weren't planning to add the camera water cooling so that the stage may be undersized. However, if this doesn't work, or we want to fix the commission data, we can add a correction factor to the distortion correction to address the rcam focus drift.