METIS WCU blackbody source has too restricted wavelength range

[oczoske]

What happened?

The black-body source in the METIS WCU is instantiated on a wavelength range that is determined by the filter that is in place when the OpticalTrain is built. In wcu_img_lm that is Lp. After FilterWheel.change_filter(), the source is not recomputed and therefore does not necessarily cover the wavelength range of the new filter. The Lp definition reaches into the M-band which is why most M-band filters receive some flux. The bug was probably missed because of that.
The same problem occurs in wcu_img_n.
The fix should occur in WCUSource.get_wavelength(), where the full LM (or N) wavelength range should be used.

Code to reproduce

cmd = sim.UserCommand(use_instrument="METIS", set_modes=['wcu_img_lm'])
metis = sim.OpticalTrain(cmd)
wcu = metis['wcu_source']
wcu.set_fpmask("pinhole_lm")
metis['filter_wheel'].change_filter("Mp")
metis.observe()
assert metis.image_planes[0].data.max() != metis.image_planes[0].data.min()
# -> AssertionError

Exception traceback

ScopeSim bug_report()

scopesim: 0.11.2a2

IRDB installation

Cloned and linked IRDB repository

Use case

Other

[hugobuddel]

I don't particularly like the .change_*() methods, because it makes everything less declarative as witnessed by this issue.

We can perhaps satisfy the desire to have those methods and to have everything declarative by recreating the entire optical train when something changed?

That way, the source is still restricted in wavelength, but always to the right wavelength range.