[bsdubernerd]

I assume this happens on very small threads and/or with thick layers?

I print both depending on size, and a full thread is usually more self-supported, not _less_, while printing faster to booth. Toothed threads tend to suffer from curling more easily when comparing at the same overhang slope, and are weaker structurally.

I'm using toothed threads for larger threads to save on time/material usually, or oddball constraints.

I can see why you would get the effect you're describing, but this would only happen if you're printing with insufficient Z resolution in order to support the overhang of the thread. I cannot see how a toothed thread would improve in such a scenario.

[regularfry]

Yes, small threads in particular are prone to this. Although the first thread I needed it on was 2mm pitch with 0.2mm layer height, so not that small.

[bsdubernerd]

I'm a bit surprised by this. 2mm pitches are well printable at .2, although I usually don't use the stand ANSI/DIN profiles, but limit the overhang profiles manually depending on the material (from 45 up to 70, really) to improve them. Still, at 2mm, this shouldn't be necessary.

Maybe the slicer you're using is doing something odd.

Keep in mind that if the filament string looks pulled in a full thread, a toothed thread with the same settings will be even weaker. In such cases it will change the result to localized sagging areas, where the structural element is just the start/end of each tooth.

A screw, even plastic one, will easily self-thread into the thin amount of plastic remaining, but that doesn't mean that the thread works as intended structurally.

If you have issues with smaller threads my reccomendation is to print an undersized cylinder and use a tap for best results, or just slightly undersize the hole and self-thread instead.