Proposed changes to the charge_exchange IDS

[olivhoenen]

This proposal is coming from Lynton Appel and was initially opened in https://jira.iter.org/browse/IMAS-5781

Background

The charge exchange system on JET,MAST-upgrade and other tokamaks have conical lines of sight that are focussed to a finite diameter, d, in the so-called object plane situated in the plasma cavity. At present the charge_exchange IDS doesn't permit the value of d to be specified. In addition, it is assumed that the etendue of the whole system is specified and implicitly each channel is assumed to be identical; also the size of the base diameter of the optical cone is not specified. The direction vectors of the optical cones can be computed from the position of given points and the centroid of the lens. However, this definition would be very awkward to use if refraction is significant. In addition, the given points are explicitly stated to be "measurement points", presumably an approximation as emission is not from discrete points, and this field is unrelated to the conical sections.

Proposal

I propose modifying the charge _exchange IDS so that it is more closely aligned to diagnostic purposes.

(1) To define the following fields per channel.

v_c — the position of the entrance pupil in 3-D xyz global cartesian coordinates.

v_axis — the normal orientation vector of the optical cone in passing through the entrance pupil and perpendicular to the lens plane. This is in 3-D xyz global cartesian coordinates and orientated towards the plasma cavity.

f — focal length of the collector lens (scalar float)

D_cone — diameter of the base of the optical cone at the entrance pupil (scalar float).

d_image — diameter at the image plane (typically the diameter of the optical fibre) (scalar float).

Most of these parameters can be different for each channel, depending on the specific design of the diagnostic. Therefore, I strongly recommend that these parameters are defined separately for each channel.

Note that D_cone may in practice be the same as the lens diameter, but this will depend on the specific design of the optical system. As the optical geometry depends on the base size of the optical cone I propose to use D-cone and not the true lens diameter.

As these parameter do not refer to anything that exists in the plasma cavity, it will be possible to include refractive effects if required (specifically I have specified d_image and not d_object.

The existing field "position of measurements" can be retained if desired, but should be stated to be "approximate" positions assuming point emission.

(2) Optionally, the details of the aperture can be included as they are done at present. However, I recommend it is not included unless you choose to describe the complete diagnostic system which would include details of the both the spectrometer and collection lens systems; these both consist of compound lenses. In addition, the details of the optical fibre bundle and the ccd detector (or other photon detector device) would need to be included.

[lappel2]

Here is a small correction to my original post.

In the documentation the given points for each channel are explicitly stated to be "measurement points" and are time-dependent. These are presumably an approximation as emission is not from discrete points; this field is thus unrelated to the conical sections.

This existing field "position" of measurements can be retained if desired, but should be stated to be "approximate positions assuming point emission".

[olivhoenen]

I've updated the description accordingly @lappel2.

[imbeauf]

Indeed, the charge_exchange IDS has been designed with the ITER diagnostic in mind, in which all channels are going through the same collimating aperture. We can make this more general by moving "aperture" and "etendue" under "channel". Can you confirm that what you call the "entrance pupil" is the same thing as the "collimating aperture" ? Since at least its position can be described by the existing "aperture" structure (moved under channel).
I understand the other fields are completely new and will need to be added.

[lappel2]

Yes, collimating aperture and entrance pupil, same, the centre of the lens of an equiavlent single-lens system
.

[imbeauf]

Ok, so I have prepared a PR with the following extensions:

• Copy the aperture structure (and the etendue and etendue_method nodes) from the top level to the channel level. I kept it also at the top for the sake of backward compatibility, and also it will be simpler if the values are common to all channels. So your requested v_c field is now channel/aperture/centre/r, phi and z. And your requested v_axis is channel/aperture/x3_unit_vector
• Added the other fields as follows: channel/focal_length, cone_diameter, image_diameter