Do we have a document that describes the calibration (and other) data that we need to routinely take? After the July experience, it's clear that we need to be more careful about taking the calibrations, and that stability cannot be assumed – while we think that the problem is the flat field screen, this isn't yet proven (e.g. Kota Hayashi's results don't yet seem to reproduce the effects we see).

So, I think we need a document (pbworks?) which describes the standard calibrations, including how to both take and reduce them. Then these "run" tickets could refer to that general description, and only add extra things that are special. An example would be that in December I think we're going to want to look at how fibres move as we change the flat field screen illumination (e.g. using one ring-lamp at a time).

I think it is good idea. In the recent runs, we referred previous tickets, PBworks page, and obslog.. but a dedicated document should exist.. 

I'm sorry for slow work, but I'm making the procedure here:
https://sumire.pbworks.com/w/page/155335224/Procedure%20to%20take%20test%20and%20calibration%20data%20during%20engineering%20run

And here are a few comments:

I'd suggest to take

• Quartz (and arc) at least every night to trace stability
• (If there is a chance and we can prepare)
  • Some data to to study dot penumbra
  • Quartz at different Cobra phi/theta position

Also, we need to make new maskFiles for mod32/mod16 with all SMs if you need ones. Arnaud has updated the mask for fibrProfile, but not for the other items (INSTRM-2123). 

For the fiberProfiles, thank you for including the dithers on the dot-roach darks. Could I further request that the dot-roach darks be taken immediately before (or after) the observations they are intended to correct? I understand that this may require an additional fiber move, but I think it's worth it to reduce the possibility of the quartz brightness changing.

For the background study, could you also please observe each lamp separately in the mod32 configuration? Data was previously obtained with something like a mod10 configuration, but I think the extra spacing could be helpful.

I second Moritani-san's request for quartzes at different cobra positions; I think this will be important to see how good the extraction is when the fibers aren't at the home position. If you have lots of spare time, maybe we could do a complete fiberProfile run (the four sets of mod4 observations) with the exposed fibers at a position that isn't "home", but a position calculated/suspected to have the worst effect on the fiber profile (without going into the black spot penumbra; maybe this is just 180 degrees from the home position?).

If "dot-roach" dark means quartz taken after dot-roaching, with all fiber hidden (step 5–7 in the PBworks page), maybe we could take some if we take data one group after another?

Previous data were taken during exposure for the PSF modeling... is only mod32 enough?

Yes, a "dot-roach dark" is a quartz exposure with all fibers hidden. The current plan is:

• dot-roach dark brn, dithered (but I don't see dithering mentioned on the PBWorks page)
• dot-roach dark bmn, dithered (but I don't see dithering mentioned on the PBWorks page)
• mod4 traces bmn, dithered
• mod4 traces brn, dithered
  In this case, the brn traces are taken long after the brn dot-roach dark. I would like to minimise the time between taking the "dot-roach dark" and the corresponding exposures with fibers revealed. An example ordering might be:
• dot-roach dark brn, dithered (but I don't see dithering mentioned on the PBWorks page)
• mod4 traces brn, dithered
• dot-roach dark bmn, dithered (but I don't see dithering mentioned on the PBWorks page)
• mod4 traces bmn, dithered
  As I said, this probably means an extra fiber move.

OK, you are requesting to add dithered trace with all fiber hidden behind dot. 

Just out of curiosity... in July the groups 2,4,4 were taken on after another.. and did you see some instability in allHidden traces? (It was not dithered thought..)

If on-telescope time is approved, we should have enough time in this period,, so problably we can separate brn / bmn. We (Arnaud) used this procedure to save time to dot-roaching.

I didn't see instability in the quartz fluxes, but I want to be even more sure that it's not lurking.

Oh, even more important than the quartz fluxes is the possibility of the fibers moving on the detector due to normal flexure in the instrument over time.

We've seen shifts in the July 2023 run (compare dot-roach dark 97774 and quartz 97805 for fiberId=41 on r1). This might be due to the slit not returning to the correct location or optical changes over time, but we can relatively cheaply suppress the latter by taking the dot-roach dark immediately before or after the quartzes they are intended to correct, with no intervening exposures. (I don't think there's a cheap way to suppress the former.)

Cobras that fail to reach their destination may be acceptable for science exposures, but for calibs it means we’re going to get sub-optimal measurements for that fiber that will affect it (and potentially neighboring fibers) for the entire run or even longer. I think that means it’s generally worth spending extra time allowing the cobras to converge for calibs.

Do we need/want to check hexapod repeatability when doing the dithers? I.e do we want to run [ 0.0, -0.4, -0.2, 0.0, 0.2, 0.4, 0.0 ]?

The default sequence is like that. More precisely, (home, -0.4, -0.2, 0.0, 0.2, 0.4, home, and home=0) and if the slit is power off, take 1 frame with power off (so we should slit on before profile) at the beginning and at the end.

Regarding to take arc/trace to check illumination pattern with and without flat screen, could you confirm which lamp to be used? PFIlamp to compare the July.. is my understanding correct?

Regarding to sparse-sky, Jim told that the dark-sky is most important (and moony sky). so exposure time should be longer to obtain enough count (around half of full fell or less in IR).

The run has been finished so I close this.