Working through "Guide to practical harmony" of Tchaikovsky in chapter 9 about the inversions of diminished and augmented triads, we read that the first inversion of the diminished triad is consonant because the respective intervals between bass and the upper voices are all consonant.
Indeed in chord: D - B - D - F (major scale) we have major sixth (to B) and minor third (to F) from the bass voice. So we consider it to be a consonant chord, and we use the appropriate rules (and do not comply with the rules of dissonance being supported by consonances, resolution and so on).
The question is: why do we discard the diminished fifth between tenor and soprano (B - F)? Are the dissonances in upper voices considered to be less "dissonant" (sorry) as opposed to dissonances with the bass in common practice music?
An extreme example (+ a small question): What if we have a chord where two of upper voices form a minor and major third with a bass (which will be by a definition from the book a consonant chord, but is strongly dissonant). Is the catch here, that we are not diatonic anymore, or because it is not a triad?