To return to the main issue: there was a theory about, claiming that
nasal infixion was originally a nasal suffix in IE, like fingo from
*fig-n-o. When JRRT wrote the original Proto-Eldarin background for Q
and S, he might very well have taken account of that theory, though
it is abandoned. So, maybe the nasal suffix theory of nasal infixion
in Q should be abandoned as well - or retained as an optional
explanation in IE too.
As for palatals, they are between the dentals and the velars and the
tyelpetéma becomes dental (telpe) in Telerin, velar in Sindarin
(celeb). I do not know of any historic language having originally any
distinction between palatals and both velars and dentals, but
palatals may come from either. In Rom. languges they come from velars
(compare Church Latin and Italian "Caesar" with Gk "Kaisar") but in
Gaelic they come from dentals: "is" (pron. ish) "teine" (pron.
chayney). So they are between velars and palatals, just as velars are
between palatals and labialised velars: in Satem-languages the
labialised series become velar, in Centum-languages the palatalised
velars (not pure palatals! or?) become pure velars.
The old and abandoned theory held PIE had all three series - and,
once again, JRRT may have used that in Proto-Eldarin "reconstruction".
Hans Georg Lundahl