Backup tiles & EBVFAC

[schlafly]

We should check to figure out if we are observing backup tiles to extinction-corrected depths or to EBV = 0 depths. We intend to do the latter, but it looks like we're actually doing the former. The ETC has code to do the latter, but maybe it's not active?

[schlafly]

Some notes from Anand:

• values for this tileid=42957 exposures (using: speed = efftime_spec * ebvfac * airfac / exptime)
  TILEID NIGHT EXPID EXPTIME EFFTIME_ETC EFFTIME_SPEC SPEED

---

42957 20220112 118314 606.36 1.25 0.0 0.0
42957 20221108 152390 600.53 1.78 3.0 0.13
42957 20230126 164810 601.82 4.15 5.03 0.22
42957 20230127 164945 601.39 10.51 11.5 0.51
42957 20230128 165074 605.67 10.76 11.91 0.53
42957 20230129 165212 606.17 4.23 5.47 0.24
=> the etc values are in agreement with the efftime_spec values, as you noticed.

• the per-exposure efftime_spec is computed here: https://github.com/desihub/desispec/blob/d9f7cd833452088bbeed2791a000516c37e00da9/py/desispec/exposure_qa.py#L308-L335
  => nothing particular done for backup
  (but I m not 100% sure that s the place to look at ...)

[araichoor]

I m not sure at what level this should be treated but, in case relevant for this discussion: in the fiberassign scripts, nothing particular is done for the EBVFAC which is stored in the header.

it s just based on the median value of the 5000 ebv values.

here: https://github.com/desihub/fiberassign/blob/270b4e80c53d94e48c54a9448d32c3e6b7ec90d5/bin/fba_launch#L388-L404
and here: https://github.com/desihub/fiberassign/blob/270b4e80c53d94e48c54a9448d32c3e6b7ec90d5/py/fiberassign/fba_launch_io.py#L2158