Fibre feed question

As an antidote to the long, normally cloudy, Scottish winter nights I built
a Czerny Turner Spectroscope loosely based on Dale Mais' design in
'Practical Amateur Spectroscopy.



Using an 1800 lpmm grating , resolution is less than an Angstrom in the
orange. Everything went fine until I came to the fibre optic delivery system
when, using the design guidance in PAS, I simply placed the fibre (Ocean
Optics, na about 0.2, 400um fibre with a Y for calibration) next to the
slit.



I was prepared for the inefficiency of this arrangement but not the
horrendous distortion which appeared - the slit was simply not focusable.
Opening the slit allowed me to refocus on the fibre itself but that really
reduces the resolution.



With the help of some modelling software (OpticsLab) I suspect that this
may be related to poor matching of the fibre - slit - collimator mirror
arrangement and I eventually solved the problem with an Ocean Optics fibre
collimator lens



However, while this removed the distortion it is even less efficient (the
collimated beam is about 3mm in diameter) and it merges the image of my two
fibres (one for measurement the other for calibration).



Looking on the web I see a wide variety of possibilities, some quite complex
(and presumably lossy) while others look like the original design in PAS.



I wondered if anyone in the group has any experience building and using
conditioning and delivery optics for a fibre feed spectroscope?



Many Thanks



Iain