A 1st go at a fully enclosed design was a failure.

The hot section with a PETG base successfully backflowed, despite having a lot of holes. 

It didn't backflow with a PLA nose cone.  Any updraft still caused it all to go up before it ever went into the hot section.

The PLA nose cone also melted.  This part would have to be made of PETG or widened quite a bit.  A PETG version would be welded.

Repositioning the inlets made no difference.

The next idea was to make a big old cutout like the stick burner.  At minimum, an HLS elevator port might reduce the updraft.

This might have gotten slightly more backflow going, but it still melted. 

Then tried completely plugging all the air holes in the base with clay.

This got it to just start looking like a liftoff plume.  1 problem is the diameter of the rocket obstructing most of it, but past experiments showed it could form a much larger plume when it's working.  Another problem is the backflow still getting defeated by the updraft.  The ideal solution is to eliminate the updraft.

Fumes tend to pool inside the base but get taken up by the updraft when they hit the edge.  Given the air tight base not being very easy to achieve & not solving the biggest problem, it was pretty obvious that the hot section should stand on its own & have a single piece rocket slip over it, with an enclosed PETG base just for alignment.

All that can be done for heat resistance is to make it bigger & out of PETG.

There could be  some kind of forced air which directs all the fumes out the bottom, whether it's backflow incense or not.  A fan in the nose could just blow down.  It could just defeat the updraft.  1 original idea was to make the entire cylinder a mesh of air holes.

Backflow cones are no longer sold anywhere locally so it's another snail mail order to do any more on this project.  There are ways to make inferior backflow cones out of normal cones.

Reviewing the 1 reddit post, backflow incense has never been very popular.