pipe1d-79-image.png
WR_20230710_spike_and_noise.png
[PIPE1D-73] Feedback from the weekly run of July 10 Created: 11/Jul/23 Updated: 13/Jul/23 |
|
| Status: | Open |
| Project: | DRP 1D pipeline |
| Component/s: | None |
| Affects Version/s: | None |
| Fix Version/s: | None |
| Type: | Bug | Priority: | Normal |
| Reporter: | vlebrun | Assignee: | price |
| Resolution: | Unresolved | Votes: | 0 |
| Labels: | None | ||
| Remaining Estimate: | Not Specified | ||
| Time Spent: | Not Specified | ||
| Original Estimate: | Not Specified | ||
| Attachments: |
pipe1d-79-image.png
WR_20230710_spike_and_noise.png
|
| Description |
|
Over the 400 galaxy spectra of the weekly run (PIPE2D 2023.28), the success rate felt from 94% to 64% since the April production (PIPE2D 2023.12). On the very first spectrum
there is an unflagged spike that is mistaken for an emission line. In addition the noise in the IR arm is a factor close to 10 higher than in the two other arms. The situation is similar in all spectra with a wrong redshift measurement. There is no specific redshift value for those values in error, which shows the problem seems to come from CR removal rather than from sky line subtraction. |
| Comments |
| Comment by vlebrun [ 11/Jul/23 ] |
|
no problem for the 13 quasars, the emission lines are strong enough to dominate the residuals that can be seen in spectra |
| Comment by price [ 13/Jul/23 ] |
|
The spurious line is indeed due to an unflagged CR. Here's fiberId=46, visit=1005 at 837.2nm: |
| Comment by price [ 13/Jul/23 ] |
|
Regarding the noise in the NIR, the simulated images were recently regenerated, which involved changes to how the NIR images were constructed. I believe I corrected the gains, but the read noise changed a lot (because it had previously been treated as a CCD, with a CCD-like read noise). I set the read noise to 20, which is similar to the 18 I found in camera.yaml, but there's some ambiguity in the units: the code treats it as electrons, but there's a comment that says it's ADU (and it makes a huge difference, because the gain is 9). We'll try to track down what the value should really be. |
