Hi Malcolm,
I would personally move the prism a bit further away from the optical axis, so it wouldn't affect/obstruct light falling onto the CCD at all, as in addition to casting shadows, light can be diffracted on the edge of the prism and thus introduce more variations to the data. Yes, it can be corrected with good flats but all in all I feel that not allowing the prism to affect the photons hitting the main CCD could in some cases be superior to later correcting data with flats and hoping it will entirely remove all introduced artefacts.
|