[Casiopea]

Originally Posted by j4mes_bond25

So, generally speaking, if "in" is used as prefix in examples such as "insane, inception or instant", it would be an exception to this "rule" i.e. NOT having "t" sound between "n" and "s" sounds.

Well, the difference is position. Below, epenthetic [t] sits at the end of the word; more specifically, it sits in the coda of a syllable. (The symbol # represents a syllable boundary)

tense = te[nts]
sense = se[nts]
chance = cha[nts]
finance = fi#na[nts]
pen#sion = pe[nt]#sion
ten#sion = te[nt]#sion

Adding [t] has to do with the ease of articulation. Articulating [n] then moving to [s], [sh] produces a [t] sound, an epenthetic consonant that shares place of articular with [n].

fi#na[ns] => fi#na[nts]
pen#sion => pe[nt]#sion

As for the exceptions, "insane", "inception", and "instant", [n] and [s] are separated by a syllable boundary, but they are also separated by a morpheme boundary (_). [n] belongs to the coda of the first morpheme, and [s] belongs to the onset of the second morpheme:

in_sane
in_ception
in_stant

Now, I don't know if that's the reason speakers don't appear to insert [t] in that environment, but it stands to reason, especially given that speakers know that in-, alone, is a separate morpheme. That is, sane, ception, and stant don't have to be productive morphemes.

Note, and here's a rather surprising twist. Even though I do not pronounce the words below with epenthetic [t] (I speak a North American dialect of English), I do, in fact, pronounce one of the exceptions, "instant", with [t], in[t]#stant. What irony. The reason ... I can explain later if you'd like. This post is too long already. ;-D

tense = te[ns]
sense = se[ns]
chance = cha[ns]
finance = fi#na[ns]
pen#sion = pe[n]#sion
ten#sion = te[n]#sion

All the best.