Lat. -idus

I am trying to make sense of the life-cycles of *es-stems in Latin.
Jens's theory of the *s/*t/*h1 morpho-phonological complex makes one
expect adjectival derivatives in *-eto- (and substantivisations
thereof), such as *wenh1-eto- from *wenh1-os. This is indeed what we
seem to find in some cases (cf. Skt namas- 'homage' vs. Gaul. nemeton,
OSax. nimid- 'holy place'). But we also find derivatives in *-osto- (~
*-esto-), such as Lat. onustus 'burdened' (cf. Gk. onostos 'to be
blamed'), venustus 'lovely' (there was an Alpine tribe called the
Venostes beside the Veneti), <honestus> and other similar ones. These
can be understood as *-eto- formations blended with *-os/*-es-. But by
far the most regular type of Latin adjectives related to verbs in
*-eh1-, *-eh1-sk^e/o- and to *es-stem nouns contains those in <-idus>,
like <fri:gidus> (<fri:geo:>, <fri:ge:sco:>, <fri:gor>, <fri:ge:do:>),
<torpidus>, <calidus>, <lepidus>, <vapidus>, <mu:cidus>, etc. (also with
contraction as in <u:dus>, <cru:dus>). What is the origin of this
formation? Is it *-edo- "substituted" for *-eto-? But why?

One possibility that has just occurred to me is that the PIE alternant
of *s was not *t but a voiceless fricative (of a "thorny" kind) which
was "hardened" into *t in such branches as Celtic, Germanic and Greek,
but which fell together with the fricative reflex of *dH in the ancestor
of Latin, yielding -d- in intervocalic positions: *-eþo-s > *-eðos >
-idus. The combination *-sþ- developed into Lat. -st-. Any comments?

Piotr