Initial /cl-/ and /tl-/ are both found in a number of Welsh words. /Dl-/ appears only as a result of the Soft Mutation.

As you said, there is no "tl-" in Gnomish (I won't speak about Early Noldorin, as my knowledge of it is next to none), so no problem here :)

As for "cl-", its adaptation to Sindarin is part of a larger question of how Tolkien later changed Gnomish words on "c-" (and "cr-", "cl-", "cw-") generally. Which in turn is part of the overall question of how Tolkien later treated his Gnomish creation (that is - those parts of it he didn't simply reject) in total.
So far it looks to me as though there isn't any universal approach. Yes, sometimes he liked one or more of the "modern" words, so he preserved them and just adapted their phonological structure and etymology (including creation of a brand new root, if needed). But in other cases he preserved the "ancient" root (or most of it) and created brand new derivations from it. Etc. etc.
So I don't think this can be just about finding some straightforward pattern like "this Gnomish consonant cluster will be this Sindarin consonant cluster and that's it".

+ܤܡܝ ܦܠܕܢܝܘܤ That’s interesting. If tl- remains in Welsh but not dl-, that hints that dl- > gl- might be the only change. There don’t appear to be any clear examples of primitive dl-, however, so this remains a hypothetical case.

+Lokyt L. I agree that it’s not just a case of simple substitution of consonants. Obviously if Gnomish cl- came from primitive kl- it would remain cl-. I am trying to decide if Gnomish clum should remain or become tlum.

+Paul Strack Why not *telum?

+Ekin Gören My thinking as well.

+Paul Strack I believe you're not asking the right question here (at least in the particular case of "clum").

Gn. "clum" comes from early primitive √TEL, a root that "survived" to the middle and late periods more or less intact. But can any kind of "clum" or "tlum" be derived from it in N./S.?
Doesn't the "u" in "clum" come from Gn. vocalisation of primitive syllabic "l"? Beause if it does, it probably can't exist in Sindarin at all...
And if it doesn't, then what kind of process removed "e" from primitive "*telu-" (if I reconstruct ancient form of Gn. "clu-" in this case correctly). Does this process still exist in middle/late N./S.? And what about the "u" now, can that still be there at all?
Also this derivative suffix "-mbī" > "-mmi" > "-m" (?): does this one have a middle/late N./S. "descendant"? If so, which one is that? If not, what N./S. suffix can we use as a substitute?
These are the questions I would ask :)

+Lokyt L. My theory is that G. clum < t’lumbe < telumbi where the e in the initial syllable was lost because it was unstressed and because the result was a “favored
initial syllable”.

+Ekin Gören I suppose I could use *telum, but when adapting early words I prefer to stick as close to the original form as I can, ideally use the exact same form if possible.

But tlum/clum may be too much of a stretch. The Sindarin syncope I described above only occurs when there are two vowels of the same quality, so tala... > tla... is possible (as in tlaud) but maybe not telu... > tlu...

+Paul Strack Well, intriguing as the topic is, I can neither confirm nor falsify your theory. But if you're right, then I agree that "telu-" > "tlu-" is rather impossible in Sindarin.

As for sticking to the original form, I think it's a question whether one does that better by preserving the exact string of phonemes or rather by preserving the ideas behind that string. I believe the latter is closer to Tolkien's methods.

+Lokyt L. I agree that consistency with later forms and constructions is essential when adapting early words. But if I can figure out a way to be consistent with later constructions AND preserve the original form, I consider it even better.