This is under active investigation by MPE/MPA. We received the instructions to run the code at Subaru which we have begun to do now.

exec: sum: We found that the particular mentioned bright star was flagged as not visible due it's proximity to a black dot.

When netflow is called

it in turn calls ics.cobraOps.TargetSelector to compute target visibilities.

The attached plot shows a red cross for all non-visible targets. Observe that it falls very close to 

a black dot. The orange color indicates the priority=nonobservationcost. There is actually another

target at x~56, y ~ -70 that gets dropped for the same reason despite large weight (orange halo).

We have scanned the entire focal plane and find not situation where a visible 

high priority target has not been allocated.

The definition of the black dot avoidance of course might require attention, but netflow

seems to be ok.

I'd propose to close this ticket in favor of a new ticket to investigate black dot avoidance criteria.

Slides from Max presented during a discussion on this topic 2022-09-12 added: bright_star.key.pdf

We asked Chi-Hung about xml file at a meeting a few days ago. According to him, xml file was updated in June run based on the measurement during the run, except for the arm length. For arm length, the values measured at test bench (i.e. off-the telescope) which has good repeatability. (The arm length on the telescope has not been confirmed for its repeatability (INSTRM-1671). ) The xml file in the 2021 September is not based on measurement.

I have changed the script to generate pfsDesign according to the discussion in the meeting we had on 2022-09-12. The change is in the ticket branch: https://github.com/Subaru-PFS/ets_target_database/tree/tickets/FIBERALLOC-44 . Basically, we get back to the original function so that we can load the measured geometry information of both cobras and dots.

Just only checks for NGC6633.

If we take a margin fraction of 1.8333 (means the dot radius is 1.375mm, which is the assumption in the previous run), the completeness of bright Gaia sources (in SM1 field) is ~79%, while if we don't take the margin the completeness is ~95%.

An example of losing targets by expanding the dot search radius is below:

The changes adopted in the last run have been merged, so this ticket can be closed. Note that some quick fixes unrelated to this ticket but necessary for the run are also included in the ticket branch unfortunately.