Note that truly FDA-approged phototherapy devices require a prescription in the US. It's quite likely that device isn't even UVB, much less "FDA approved"

(I have a condition that requires UVb. I wasted so much money, and skin burns, on Amazon's shitty medical grey-market... )

Androv medical makes quality lights. Even though their page seems kind of shady.

Because the study says this is the most effective wavelength.

The study (open-access PDF link in the article) says that in their testing on human skin samples, the 293 nm wavelength was the most efficient, and would require 10 times less exposure to get the same effect as the 305 nm wavelength. They seem to suggest that 30 seconds of exposure with the 293 nm wavelength lamp might be the equivalent of 30 minutes of sunlight.

The PDF also says that D3 is produced from lights all the way up to 320 nm, so the 311 ones sold on Amazon would work. But given the 293 vs 305 effectiveness, how much exposure would you need to actually make a meaningful difference? Probably a lot!

Until we know that 311 nm is rather benign, we must assume that using a stronger UV source is more likely to induce cancer. While I don't know about the size of the effect here, in general you try to use the lowest possible dose when medicating.