[bratislav (Bratislav)]

It would be actually possible to make quite decent (in fact more than decent) two simple BK-7 lens coma corrector for a 6" f/2.8 .
Catch is, you have to make it as a system , that is mirror will have to work together with lenses. You wouldn't be able to use it alone, lenses are essential part of the system.
Consequence of catch #1 is, mirror will have to be hypebola. Twice as deep as paraboloid (which is hard enough as it is at f/2.8 ).
So here's the prescription - mirror is ground at f=450mm (that is, f/3). This actually makes it easier to figure, as well as asking for much smaller secondary. The eccentricity would be -2.1 .
I also made lenses easier to manufacture. First lens is plano-concave. Lenses touch at the edges (also to make it easier to manufacture, no need for precision spacers, but make sure you bevel edges well). Second lens is biconvex, BFL is a generous 87 mm .
Performance is excellent (better than diffraction limited) across 6 mm circle which is larger than the field stop of a high power eyepiece (say 2.5mm Nagler Type 6).

BTW, don't even think on skimping on eyepieces. Minimum would be Nagler line (or similar), even better, use Ethos. Most eyepieces will work rather poorly at f/2.8 even on axis.

Having said all that, this is not a project for a beginner. Deep hyperbola, tight control of lens radii and wedge, near zero mechanical tolerances for alignment and collimation ... not a job for someone without an access to a well equipped machine shop.

So, I'm sorry to say, but you will have much better chance of a success if you just buy a Paracorr2. It will be much cheaper in the long run actually, and you will be able to use it in your larger fast mirrored Dobs later.