Unexpected behavior of CRL + crystal

[SergeyRa]

(Also reported in oasys-kit/ShadowOui)

When testing schemes with collimating lens installed before asymmetrically cut crystal, I found an unexpected behavior: lens widgets that use PreRefl data, followed by asymmetrically cut crystal produce a strange diagonal strike in beam cross-section, which I cannot explain physically. In addition, when using the 'ideal lens' widget instead (but with the same focal distance), the resulting beam cross-section is OK (see also .ows with example here: Test.zip).

[srio]

Thanks Sergey for your post. I'm looking to it and it is hard to interpret... or to identify what's wrong.
To be continued...

[srio]

There is indeed a bug in shadow3.
It happens that the lens switch polarization of about half of the intensity.
This makes not sense.
I will look at that in detail and fix shadow3, but this will take some time.

For the time being, you can patch the beam after the lens to put all the intensity as s-polarized using a script:

where the script is:

import numpy

out_object = in_object_1.duplicate()
intensity = out_object._beam.getshonecol(23,nolost=0)
print(intensity)

out_object._beam.rays[:,6] = numpy.sqrt(intensity)
out_object._beam.rays[:,7] = 0.0
out_object._beam.rays[:,8] = 0.0

out_object._beam.rays[:,15] = 0.0
out_object._beam.rays[:,16] = 0.0
out_object._beam.rays[:,17] = 0.0

this changes the output from before:

to now:

Thanks a lot Sergey for reporting this problem!