As mentioned my nemisis is getting the spectro to guide on the star with sufficient accuracy on a 20micron slit. Well, I think using a 70/30 beamsplitter in an OAG body has got me there!!!! ( check back when I can get some clear skies to put it through its paces) At least I can repeatably position the artificial star in/on the slit and have a good image for guiding..... let's see what happens in real life!
Another option is to use a cylinderical lens to form an extended linear image of the star, then present that to the grating.
Fulvio Mete
www.lightfrominfinity.org/ has been doing some excellent work with such a concept, applied to a standard spectro. My idea it to apply it to a transmission grating to give "better" resolution. Finding a suitable FL and position is a challenge..... the benefits are that you don't have to guide on the slit- the star image makes its own one!