Relevant visits:

1. 1B: opt-rear, 2B: opt-front, 3B: IR-front, 4B: IR-rear. The tower connector is attached to the spider on the IR-front side.

The two box plots (one each for `r` and `b`) show the iqr for the median `dx` pixel position of each fiber as well as any significant outliers. The color coding on the row background corresponds to the different ring lamps.

There appears to be no positional difference whatsoever for the different dome flats taken with the various ring lamps. This is reflected in the generated 2D maps as well, which don't show any visual difference.

`r4` has a larger `dx` compared to the other `r` spectrographs but they are still consistent with each other.

I've attached the 2D `dx` detector map residuals for visit 104640 for each detector, which are representative of all the dome visit residuals.